DLC Boot 2022 4.0 Build 220226

South African Journal of Science Volume 118 Issue 1/2

Published on Jan 27, 2022

The South African Journal of Science is a multidisciplinary science journal published bimonthly by the Academy of Science of South Africa. Our mandate is to publish original research with an interdisciplinary or regional focus and to provide a forum for discussion of news and developments in research and higher education. About the cover: Flower of Hibiscus sabdariffa L. The hibiscus is rich in phytochemicals, particularly polyphenols and anthocyanins. El-Naeem and colleagues investigated xanthine oxidase inhibition by isolated anthocyanins for development of potential novel drugs to treat diseases such as gout and hypertension. Other articles in this issue include: Aligning conservation priorities | Analysing socio-hydrology research | Using waste for wastewater treatment | Debunking the Data & Cloud Policy.

*the number of responses in each group for each task that failed language/information balance by providing too much or too little information

Table 5. Group comparisons on specific discourse tasks

Linguistic formulation level As shown in Table 5, the patients with aphasia received significantly lower ratings than both the Alzheimer group and the normal control group. The patients with Alzheimer's disease did not differ significantly from the normal control group on the linguistic formulation rating for any task. This suggests that the individuals with aphasia exhibited marked difficulty in lexicalizing their ideas as manifested by increased hesitations, revisions, circumlocutions, paraphasic errors, and ambiguous pronouns as compared with normal control subjects. This formulation difficulty disrupted the coherence of the various texts. Responses for a patient with fluent aphasia are illustrated in Table 6 in sample responses on the fable retell, gist, and lesson tasks and are contrasted with the minimal linguistic problems exhibited in the samples for both a normal control and a patient with Alzheimer's disease.

Normal Control Welt, you have this raven, ana you know that ravens are black. He saw a lot of pigeons in a pigeon coop, and he said he wanted to be there with them, but his color made him different, and uh painted his feathers white, and tried to come in there and dilly-dally with the pigeons, but when he started crowing instead of cooing, they knew he was a fake, so they chased him out. So hey, he had to go back home where the other ravens were, but he had this white paint on him. The ravens looked at him and said, "Hey man, what are you? You're white, get outta here." So he was ousted. So he had no place to go. He was persona non grata. | Oh, the gist of the story is that a raven wanted to be other than what he was and he had suffered the consequences.

| Oh, go back and read Shakespeare and some of the things, unto thine own self be true and be natural, don't try to be something that you aren't. Be yourself. Be thyself. Comments | Absence of formulation difficulties. Subject clearly isolates central meaning by integrating textual information with real world knowledge. Subject responds appropriately to task demands, recognizing requirements for condensing the text in gist and lesson form. Pragmatic informativity is normal. Although in the retell the subject gives more information than contained in the original, the balance between amount of language and information is normal since elaboration enriches story and reflects normal stylistic differences.

A raven saw a , saw the patg- (6 seconds) (E: pigeons) pigeon, pigeons in the ka-coop. The-the raven looked and they had of food. The raven where he got th-th-th uh paint to paint hisfeathers white so he can go into the cope-coop and have some food then. Well he got in there, into the coop and as soon as he wanted foo, no he-he wanted to, not talk, well he wanted to crow. The pigeon, "No-ho, that's not a cousin." So they threw him out and he had to run away. And then he went back to the raven-raven family and-and they looked at him, his color isn 't right. It was white so he had to lea- they threw him out.

Uh, you know, the raven uh, like any-any bird has got to eat, and the way they get their food is uh, apparently by guile. By uh, uh, confusing the uh, other...whatever the hell it was. I forgot. Um, it was a raven and what else? (E: The raven and the pigeon.) Pigeon, okay. Uh, the raven uh-uh, was gonna do whatever he's needed to do to accomplish the goal uh, which was to get some food uh, when this uh, rabbit had plenty of food and he didn't, so he uh-uh, used his head and figured the program out and uh...achieved the end that he was after, the more food.

Don't change, or like that? Don't change the color...don't change, don't change your, don't change, don't change yourself, what you are to get some money, *** to get uh food. Don't change, don't change, don't change, don't change things to, don't-don't-don't take, don't take, don't change, don't change your life just to get what you want to.

Uh, well, just the main idea is that uh, any animal or person or any thing else will do whatever they have to do to get the food they need to survive. Survival of the fittest, I guess.

Moderate formulation difficulties manifested in repetitions, revisions, hesitations, and semantic paraphasias. Subject is able to make inferences, but it is less clear as to how these inferences relate to the text. Responses are appropriate to the task demands for each task. While the pragmatic informativity is normal for the retell, there is excessive language in relation to amount of information conveyed for the gist and lesson tasks due to repetitions and revisions.

Minimal formulation difficulties. Some evidence of anomia on the retell task probably due to memory deficits in that he has forgotten the characters and the rabbit character represents an intrusion from the previously presented fable. Also perhaps due to memory deficits, his responses do not convey the central meaning. He appears to be relying more on world knowledge than the context of the story to give a generalized response for the gist and lesson tasks. He clearly recognizes the task requirements. Language information balance is normal.

Table 6. Sample responses for "Raven & Pigeons" fable

Pragmatic level Communicative intentions For the majority of pragmatic measures, there were no significant group differences on the communicative intentions measure. That is, the three groups were able to recognize the intent of the probes and respond appropriately to the task with minimal, if any, difficulty. The only exception of this pattern was the deficit observed in the group with Alzheimer's disease on the story generation task using the single frame picture. A significant number of AD patients exhibited difficulty going beyond the static picture to generate a dynamic narrative, instead producing a picture description. This tendency toward descriptive discourse is exemplified in the response by a patient with Alzheimer's disease presented in Table?. Single frame picture: Story generation

Well, uh there is a man there and there is uh looks like a child and um there's a dog and I think there's an older man there too. I think so. Um I think that's an older man in the picture. Um it looks it's an autumn day out there and it's kind of overcast, kinda like it is today. Um, and the man and the boy are talking and uh I think that the boy, the little guy, I think that they are going fishing or they're thinking about going fishing.

Let's see, it must have been during the 1930's uh a young man going to college, university, state university. And they were sitting on a a kind of a, on a small, on a ru-ru-rubber, rubber, rudder, sitting on a rubber of a car. They used to have that. And uh this uh, there was a, this young man he has a friend, a friend which was a dog, probably sheepshep-shepard. And dad was smoking and he was thinking well, "you've got to say goodbye." And uh I think he was a farmer. And so the man is thinking, "I'm going to leave and going to college or in the university." And dad is gonna wis-miss and so will the dog miss this young man.

The patient produced a description of the explicit information contained in the picture (i.e., there is a man there... looks like a child and um there's a dog...) with few inferences (i.e., they are going fishing.) Moreover, the inference that "the boy and man are fishing" was an atypical interpretation that did not integrate the salient props. Additionally, this patient with AD failed to define the relationship between the two characters. This patient also exhibited minimal linguistic difficulties such as a few hesitation and some repetition of thoughts. The patient had moderate word finding problems as evidenced through hesitation, repetitions and reformulations. Despite these linguistic difficulties, he produced a typical frame of interpretation. He expressed the central meaning, (i.e., young man going to college) conveyed the relationships of the characters (father and son). He also made inferences with his real world knowledge to infer the time period of this picture (i.e., 1930's) as well as inferring the emotions of the characters.

They had their son, and they're waiting for a train that's gonna take the boy away to college. Anyway he has a lunch that his mother has finished for him. His collie dog senses that the boy is gonna be going away. And the boy is really duded up cause he has him a pair of argyle socks, and regular black shoes and a matching sort of tie and white shirt. And he's looking with anticipation on his face, and you see the father's face, "well there goes one of my hands so I'm gonna do all the chores that he's been doing" but they're waiting for the train.

The normal subject produced a narrative with a typical frame of interpretation without any linguistic deficits. The normal subject expressed the central meaning of the story ( the boy is going off to college), conveyed the relationship between the characters (father and son), and gave background information (waiting for a train). His response revealed proficiency I making inferences between the depicted information and real-life knowledge. When he interjected descriptive information (i.e., he has him a pair of argyle socks and regular black shoes and matching sort of tie and white shirt), it was used to support his inferences, e.g., that the boy was dressed up for the big occasion.

Table 7. Single frame picture story generation

S.B. Chapman, A.P. Highley and J.L. Thompson

Drawing inferences The most relevant measure for distinguishing the patients with Alzheimer's disease from the aphasic or normal control groups was the pragmatic inferencing variable. That is, the patients with AD exhibited significant difficulties on this domain due to problems drawing the necessary inferences between the text and real world knowledge. Consequently, the texts of the AD patients were less coherent than the normal control or aphasic groups. This pattern was particularly apparent on the fable tasks (see examples in Table 6) and story generation for the single frame picture (see examples in Table 7). On the single frame picture, half the subjects with AD failed to make the necessary inferences between the information depicted and real world knowledge to establish the typical frame of interpretation (i.e., going to college). With regard to proverb explanation, both the aphasic and Alzheimer's groups performed significantly below the normal control group on the pragmatic aspect of drawing inferences. On this task, both patient groups exhibited significant difficulty formulating correct nonliteral explanations for proverbs. It is interesting to note that even though the two patient populations overlapped on the drawing inferences dimension, the patients with aphasia continued to show greater linguistic difficulties than the AD group or normal control group (Table 5). The multiple choice proverb interpretation task was less difficult than the spontaneous explanation task for the patients with fluent aphasia. That is, most patients with fluent aphasia were able to choose the correct abstract interpretation, suggesting that they can appreciate the abstract meaning of proverbs when linguistic formulation demands are minimized. Conversely, patients with AD continue to exhibit difficulty given choices, often choosing the abstract foil (in the form of another proverb). The following sample responses from a patient with fluent aphasia and a patient with Alzheimer's disease for the unfamiliar proverb "One swallow doesn't make a summer" demonstrates the similar difficulties in drawing the necessary inferences. Patient with fluent aphasia: All right, (laughs) so what does that mean, um, well it takes all summer. Uh, what, uh, well, a swallow doesn't make a summer. Right. Um, it takes time, uh let's see time, uh, autumn could also be just as well, so the swallow, what about it, if he goes in the summer time, well, that's great, but um, how about in the autumn, he goes south. I'm not doing well with this at all. I mean it doesn't um, I don't know, maybe the birds, uh, one swallow, that's a bird isn't it? (E: Can you think of a reason why one swallow wouldn't make a summer?) Right. Nope, I'm sorry. Note the similarities between the above explanation and the following by a patient with Alzheimer's disease: Um...one swallow doesn't make a summer. Mmmm, I don't know what to say to that. Uh (5 sec) Oh, we 're not talking about the swallow as a bird, are we? One swallow doesn 't make a summer. Umm...one swallow doesn't make a summer- one swallow- one bird doesn't make a...one, more than one way to skin a cat, is it? Um...well, just because someone is doing this job, doing it that way doesn't mean you can't do it a different way anduh, andreachthe same conclusion. Neither explanation incorporates the necessary inferences to convey the correct, generalized meaning of the proverb. The patient with aphasia seems unable to move beyond the context of "birds" to the more global meaning of the proverb. That is, he

exhibits difficulty in reformulating the proverb outside the original language of the proverb by accessing and selecting his own words. Notice that the patient with AD attempts to use an unrelated proverb, "There's more than one way to skin a cat," to help explain the proverb "One swallow doesn't make a summer," indicating that he realizes that a proverb should be interpreted abstractly, but fails to make the correct inferences that integrate the textual information with real world knowledge. Language/information balance On this pragmatic domain, both patient groups tended to construct texts in which the distribution between language and information was judged to be impaired. The patients with aphasia and dementia producedless information than normal controls, but did not use less language. The number of instances in which subjects were rated to use more language to convey the information than deemed appropriate for the task is indicated in Table 5. This tendency to use excessive language (verbosity) was particularly interesting for the proverb task in that the aphasic group appeared to have more difficulties on the proverb explanation task than the AD group.

Clinical validation of disparate profiles A secondary goal of the study was to determine whether or not the profiles identified by the results of this study could be used to clinically distinguish the two groups. Clinicians were blinded to the diagnosis and attempted to classify the subjects' discourse responses according to whether the response seemed consistent with that of a normal control, a patient with early AD, or a patient with fluent aphasia. At the time of attempted classification, the clinicians were provided with information related to the expected profile for each diagnostic group. Specifically, the raters were instructed to classify the patients as aphasic if they exhibited linguistic formulation difficulties, but still seemed to be able to draw inferences between the textual and real world knowledge across most tasks. In contrast, classification of AD was recommended for individuals who showed deficits in drawing inferences between the textual and world knowledge with only minimal linguistic formulations problems. If no visible symptoms of linguistic formulation difficulties outside a normal range or deficits in drawing inferences were present, the individual should be classified as normal. The quantitative results revealed that there was no difficulty differentiating the normal controls from the two patient populations. All normal subjects were identified as normal on all tasks. For the aphasic group, three patients were misclassified as producing responses similar to patients with AD for fables; one was misclassified on the single picture task, and five patients were misclassified on the proverb explanation task. Two of the aphasic patients were misclassified on both the fable and proverb tasks. For the patients with AD, two were judged to have aphasia on the basis of their fables responses, two were misclassified as aphasic on the single picture task and none were misclassified on the proverb task. Table 8 presents atypical responses for patients who were misclassified.

(RETELL) This was a story of the raven that-painted his feathers white. And got over, to get over the pigeons and they caught him and made him get away. But he left, let me see, some of them was still there but the pigeons got them all out of the way. (E: Some of who were still there?) The pigeons- the ravens, the ravens was there. But they finally got them all out. They all uh had uh yellow-yellow-no white, the was painted white. (E: And how did the pigeons know. ..that this was a raven and not a pigeon?) Because they s-I don't know, it's, they sounded different (laughs).

(RETELL) Well, you had now- it was a bird, it wasuh, raven uh, who was uh, is a raven and uh I believe he noticed that uh uh the crows uh was a getting something to eat. Uh, so he uh, got uh painted or they painted him, figured he could get more, a better deal on that, that's right— I believe that the uh crows that's what it was, kinda knocked him out of that game. (E: And what happened?) So he went back to his uh (4 seconds), he went back to his own business. (7 seconds) Took care of him it, got his own food.

(LESSON) Uh don 't, don 't, don 't let the ravens get in the chicken coop.

(LESSON) There's no- there's no uh uh, it's no one, you can't get it for nothing, you can't get anything for nothing.

Responses contain mild formulation difficulties manifested by pauses, revisions, and repetition of words. Responses do not convey the central meaning of the fable. However, subject attempts to fulfill task demands by formulating responses in terms of a retell and lesson, respectively.

Mild to moderate formulation difficulties manifested in semantic paraphasias, revisions, hesitations, and ambiguous pronouns. Retell response is vague and must be interpreted by listener, though most key elements are included. Lesson response is a generalized statement which seems to reflect the individual's value system. Subject fulfills task of lesson.

(ONE SWALLOW DOESN'T MAKE A SUMMER) I don't know if that's talking about a bird or (laughs) if-if I got hot. Uh, thatthat means thatuh, they was a lot more birds. Or there's more of anything. Let's see, a swallow, one swallow does not make a summer. Uh-uh (5 sec). Sw-uh...uh...b-uh. (E: One swallow doesn't make a summer.) (10 sec) See (6 sec) swallow. ...(E: Can you think of an example of when you might use a saying like that?) Uh...there's not a uh, let's see, there's not a uh-uh drop of rain, there's not a, there's not any r-uh-rain-uh... let's see. Swallow (5 sec) a... I don't know ***.

(TOO MANY COOKS SPOIL THE BROTH) Too many cooks spoil the broth...too many cooks spoil the broth.. (E: Got a guess on it?) Well, there's bound to be something there. I know it happens (laughs). I know that's happened. Uh...(E: What does it mean in general?) Yeah, well uh...you can't get a whole lot of things done with a whole lot of people. You've got to come up with some kind of a thing that's "doable" and that some people may not be involved with it. Uh...you don't want to have uh-youjust can't extake 'em all. (E: How come?) Well, they got too many eyes-eyes or too-gonna uh, the cooks too as well-too many peole get involved in something ***thatway responsibility had to be uh and uh...you can't have twelve people getting involved with something and somebody's got to do it.

Only mild language formulation difficulties, manifested in incomplete thoughts, pauses, and fillers. Unable to make the inferences necessary to convey the global meaning of the proverb.

Despite formulation difficulties, is able to convey the general meaning of the proverb by making inferences between the text and his real world knowledge.

Table 8. Discourse responses from patients who were misclassified

Discussion This paper presented a framework for examining salient components of discourse function relevant to defining communicative competence and applied the framework to clinical characterization in patients with fluent aphasia and patients with early stage AD. Our premise was that communicative competence could best be illuminated by characterizing discourse abilities at multiple levels of representation, including both linguistic and pragmatic aspects. The results of the study revealed that discourse was equally impaired across a range of tasks for both the mild aphasia group and the early stage Alzheimer's disease group when compared to normal control subjects. At a global level of description, the discourse performance in both groups could be characterized as sharing word finding deficits, reduced content, incoherent responses, and verbosity. The most important finding, however, was the disparate pattern of breakdown in linguistic and pragmatic domains of discourse production between patients with mild fluent aphasia and patients with early stage AD as compared to normal control subjects. The patients with aphasia were significantly impaired in the linguistic domain, whereas the patients with AD showed greater disturbances in the pragmatic domain of drawing inferences. Our data confirm previous evidence that linguistic formulation is disturbed in fluent aphasia and relatively preserved in early AD (Bayles et al., 1989; Blanken et al., 1987; Cardebatet al., 1993; Ehrlichet al., 1997). The evidence of a linguistic formulation deficit in mild fluent aphasics is not surprising given linguistic deficits are criteria in the definition of aphasia. The patients with mild fluent aphasia had difficulty across most verbal tasks varying from retelling and interpreting a fable (in the form of a lesson) to explaining the meaning of proverbs. The aphasic group performed lower on the linguistic formulation domain than the patients with AD and the normal control group. Additionally, the present findings support the conclusion of Blanken et al. (1987) that the linguistic disturbances in AD do not represent an impairment of lexicalization, but rather an inadequacy in accessing the appropriate meaning through impaired inferencing. The AD group received ratings comparable to the normal control group in the linguistic formulation domain. We had anticipated that the AD group would display some mild to moderate formulation difficulties on the conceptually more complex discourse tasks such as constructing a fable lesson or explaining the meaning for a proverb since anomia is a well-documented deficit in the early stages of AD. Perhaps the very mild degree of impairment in our patients with AD contributed to our failure to find a difference. Alternatively, it may be that even normal controls have some formulation difficulties as shown in the normal example and thus more variability in linguistic formulation is acceptable. The most prominent difficulties for the AD group were in the pragmatic domain of drawing inferences. The patients with AD performed lower on the fable retell task, the picture generation story task, abstracting the central meaning in the form of a gist, and deriving the didactic meaning in the form of a lesson. The ability to draw inferences between textual content and real world knowledge tends to be relatively preserved in patients with fluent aphasia and impaired in early AD. Cardebat et al. (1993) reported a similar difficulty with inferencing in AD using a different narrative methodology. Specifically, these researchers found that patients with AD consistently failed to mention the complicating action which represents the most important story event. In another study, Chapman et al. (1995) found that patients with early stage AD had difficulty drawing the necessary inferences between a static picture and real world knowledge to allow them to

S.B. Chapman, A.P. Highley and J.L. Thompson

produce a dynamic narrative. Thus, disturbances in the pragmatic domain of drawing inferences between textual knowledge and real world knowledge were manifested independent of linguistic formulation difficulties in patients in the early stages of AD. In addition to linguistic and pragmatic factors, our patients with AD may have been unable to produce a dynamic narrative from the single frame picture because of cognitive deficits. For example, they may have had memory problems that precluded storing the information long enough to draw the necessary inferences. Another possible explanation is that the subject did not attend to the most salient aspects of the picture. The proverb explanation task was the least informative task in discriminating between the two patient populations. However, when used in combination with the multiple choice task, the informative value increased. Both patients with fluent aphasia and patients with AD had difficulty spontaneously producing the nonliteral meaning for proverbs. The similarities between groups were underscored by the fact that five patients with fluent aphasia were subjectively misclassified as having dementia based on their proverb responses due to the difficulty in providing nonliteral explanations. In contrast, only the patients with AD had difficulty recognizing the nonliteral meaning for proverbs given choices. These findings are consistent with previous data that patients with fluent aphasia had minimal difficulty identifying the nonliteral meaning of proverbs when they did not have to propositionalize the meaning in language form (Chapman et al., 1997; Kempler, Van Lancker, & Read, 1988). Previous evidence suggested that patients with early AD had little difficulty with familiar proverbs, but marked difficulty with unfamiliar ones (Chapman et al., 1997). In the present study, we had too few items to determine differences between familiar and unfamiliar proverb interpretation. Clearly, unfamiliar proverbs require greater inferencing than familiar ones. Familiar proverbs may be retrieved much like the meaning of single words, whereas the meaning of unfamiliar proverbs must be resolved through inferencing between the text and some plausible meaning from real life (Van Lancker, 1990). It is important to note that not all subjects within groups conformed to a diseasedistinct pattern, despite the relatively consistent findings confirming a discrepancy between patient groups in the linguistic domain and the pragmatic domain of drawing inferences. The exceptions were manifested in a number of patients with fluent aphasia who showed marked difficulties in drawing inferences and in some patients with AD who exhibited notable linguistic formulation deficits. The diagnostic category alone will not define the pattern. Clearly, some patients with aphasia show inferencing deficits and some patients with AD show early linguistic formulation deficits (Chapman et al., 1997). Knowledge of an individual's pattern of strengths and weakness in the linguistic and pragmatic domains is the key to adequate diagnosis and treatment for both populations. The ability to interpret the speaker's intentions for communicative purposes may be intact at more severe levels of aphasia and later disease progression in AD (Bond et al., 1983; Hamilton, 1994). Minimal differences were found in the ability to interpret the intentions of the speaker, i.e., communicative intentions, across all three groups despite the disparate linguistic and pragmatic inferencing performance in aphasia and in AD. The ability to recognize a request made for information may be preserved late into the progression of AD (Hamilton, 1994). The only exception to this pattern was identified in the AD group who failed to respond appropriately on the picture story generation. Instead, they tended to give a picture description instead of a narrative with a sequence of events. Cardebat and colleagues (1993) described a similar disturbance in patients with AD.

Clinical implications The conceptual framework described herein offers an informative method for parsing discourse into component domains relevant to functional communication for assessment and treatment purposes in adult neurogenic populations. The most important principle to be derived from our research and others (Blanken et al., 1987; Ehrlich, 1994; Hamilton, 1994; Cardebat et al., 1993) is that discourse tasks are more revealing than isolated language measures in characterizing the full scope of communication breakdown. In particular, discourse tasks that vary in length and conceptual complexity provide a unique window to view both linguistic planning processes and pragmatic phenomena such as communicative intentions, drawing inferences between different types of knowledge, and using the appropriate amount of language to convey information. Such an approach allows us to broaden the focus of our clinical descriptions and to examine co-occurrence of linguistic difficulties and pragmatic impairments regardless of the patient's diagnostic category. If we take a narrow a path in clinical assessment and treatment because of preconceived notions of expected symptomatology, we are likely to misdiagnose and mismanage the communication breakdown in a large number of patients. The present methodology has limitations as to how much performance on story retell and gist construction tasks may actually reflect what truly goes on in a communicative exchange. Some insights can be offered regarding the type of tasks that are the most illuminating when attempting to characterize communicative function in the linguistic and pragmatic domains. Based on our work and others, we propose that the most instructive methods would be tasks that meet some of the following criteria: (a) tasks that require holistic processing rather than isolated observations about pictures. For example, picture descriptions do not require that an individual go beyond a primary analysis of the depicted elements (Cardebat et al., 1993; Chapman & Ulatowska, 1993). A task that requires the individual to construct a narrative based on the pictured information forces drawing inferences. The most effective pictures for eliciting narrative discourse are those that are dynamic and contextually rich, not pictures depicting isolated people performing unrelated activities. (b) tasks that involve reformulating the text information at different levels of generalization. For example, the various discourse tasks used in the present study such as story gist and story lesson place different demands on drawing inferences between the textual information and real world knowledge. Thus, as the subject proceeds from a retell or story generation, to a gist, and finally to a lesson, there exist greater and greater demands to condense the original story information, paraphrase it using his/her own words, and integrate the text information with real world knowledge through drawing inferences. (c) tasks that elicit linguistic formulation for texts of varying lengths. For example, longer texts are required for retells and shorter responses are possible with gist and lesson interpretations. Difficulties in accessing the language to convey one's thoughts may be more transparent when formulating longer responses and spontaneously generated responses rather than retells. (d) tasks that involve encoding the information in short term memory storage rather than viewing the stimulus while producing the response. This is recommended because viewing the stimulus tends to encourage less paraphrasing and less inferencing.

S.B. Chapman, A.P. Highley and J.L. Thompson

Theoretical implications Discourse studies in neurolinguistics help to elucidate the nature of the organization of the complex processes underlying human communication as well as the neurobiology of discourse. This study advances the theoretical notion that linguistic and pragmatic phenomena can be dissociated in patients with primarily linguistic deficits and those with more conceptual impairments. The disparity found between patients with fluent aphasia and patients with AD supports Caramazza's(1984) two postulations about the informative value of contrastive studies of brain-damaged populations. First, he claimed that the information from behaviors in brain-damagedpatients could reveal components of complex behavior that are not apparent in normal behavior (i.e., fractionation). Clearly, in normals the role of linguistic formulation and pragmatic inferencing operates effectively and interdependentlyin discourse production such that two processes are not readily separated We also speculate that in severely impaired patients, both systems breakdown such that it becomes problematic to identify the separate components. In our study, the components of linguistic formulation and pragmatic ability of drawing inferences were dissociated across a variety of discourse production tasks in two different brain damaged groups. Second, Caramazza proposed that brain damage affects certain functions and spares others to reveal potential associations and dissociations (i.e., transparency). This latter pattern was evident in the dissociations found in aphasia and AD at mild/early levels of impairment between the linguistic domain and the pragmatic domain of drawing inferences, even though discourse output in the two clinical populations showed similar changes in reduced information, verbosity, and incoherence. The discourse differences between patients with fluent aphasia and patients with AD have implications for the neurobiological organization of discourse functions. The inferencing deficits in the patients with early AD may be due to involvement of the right hemisphere. Evidence from studies of patients with right hemispheric damage has suggested that these patients have difficulty drawing inferences and deriving nonliteral meanings (Wapner, Hamby and Gardner, 1981). Perhaps the relatively intact right hemisphere in our aphasic population allowed them to draw inferences and interpret proverbial meaning at a nonliteral level when they did not have to simultaneously retrieve the words to express the information. Moreover, the evidence that a number of patients did not conform to the classification schema for their disease type (i.e., APH or AD) warrants further examination into the neurobiology of their brain disease. One might speculate that patients with mild aphasia who show obvious difficulties in the pragmatic domain of drawing inferences may have experienced a more diffuse brain injury due to more widespread ischemia that did not resolve. For patients with early stage AD who show early prominent language formulation symptoms, it would be interesting to determine whether these patients showed more focal brain disease in the left Perisylvian region in the early stage. Additionally, comparing patients with AD who have language deficits as a prominent early symptom may be an intriguing contrastive group to patients with primary progressive aphasia. From a humanitarian sentiment, the differences in discourse components between patients with aphasia and patients with AD indicate that the similarities may well reflect different underlying mechanisms. Identifying the individual's strengths and weaknesses along the component dimensions of discourse function will guide more detailed diagnosis leading to improved intervention. In aphasia, we often fail to go beyond the linguistic deficits even though a number of patients show deficits in pragmatic domains such as drawing inferences. This domain needs to be assessed in aphasia. While intervention is the rule rather than the exception in aphasia, there is considerable controversy as to whether

some form of intervention be implemented with early stage AD patients. At a general level, the speech-language pathologist's role is quite similar to the widely accepted practices of aphasia management. That is, it is necessary to help the patient to be communicatively viable for as long as possible. As evident from this study, our patients with aphasia and dementia have revealed an incredible array of strategies that allow them to achieve some level of successful communication. Thus, the identified strategies provide empirical data to guide professionals to the development of more appropriate methods for enhancing communication in these populations. Acknowledgement—This investigation was supported by grants for the National Institute of Aging/National Institutes of Health (AG09486 and S-P30-AG12300-02) for the Alzheimer's Disease Center. We express our gratitude to Hanna K. Ulatowska who designed the tasks and Jacqueline A. Prince for help in data management.

References Adams, C, Labouvie-Vief, G., Hobart, C, & Corosz, M. (1990). Adult age group differences in story recall style. Journal of Gerontology: Psychological Sciences, 45, 17-27. Adams, C. (1991). Qualitative age differences in memory for text: A life-span developmental perspective. Psychology and Aging, 6,323-336. Adams, C, Smith, M. C., Nyquist, L., & Perlmutter, M. (1997). Adult age-group differences in recall for the literal and interpretive meanings of narrative text. Journal of Gerontology: Psychological Sciences, 52B, 187-195. Bayles, K. A., (1982). Language function in senile dementia. Brain and Language, 16, 265-280. Bayles, K. A., Boone, D. R., Tomoeda, C. K., Slauson, T. J., & Kaszniak, A. W. (1989). Differentiating Alzheimer's patients from the normal elderly and stroke patients with aphasia. Journal of Speech and Hearing Disorders, 54,74-87. Blanken, G., Dittmann, J., Haas, J., & Wallesch, C.-W. (1987). Spontaneous speech in senile dementia and aphasia: Implications for a neurolinguistic model of language production. Cognition, 27, 247-274. Bond, S. L., Ulatowska, H. K., Macaluso-Haynes, S., & May, E. B. (1983). Discourse production in aphasia: Relationship to severity of impairment. In R. H. Brookshire (Ed.), Proceedings of the clinical aphasiology conference (pp. 202-210). Minneapolis: BRK Publishers. Caramazza, A. (1984). The logic of neuropsychological research and the problem of patient classification in aphasia. Brain and Language, 21, 9-20. Cardebat, D., Demonet, J.-F., & Doyon, B. (1993). Narrative discourse in dementia. In H. H. Brownell and Y. Joanette (Eds.), Narrative discourse in neurologically impaired and normal aging adults (pp. 317-332). San Diego: Singular Publishing Group, Inc. Carlomagno, A. (1994). Pragmatics in Aphasia. (G. Hodgkinson, Trans.). San Diego: Singular Publishing Group, Inc. (Original Work published 1989). Chapman, S. B., & Ulatowska, H. K. (1991). Nature of language impairment in dementia: Is it aphasia? Texas Journal ofAudiology and Speech Pathology, 17,3-9. Chapman, S. B., & Ulatowska, H. K. (1992). Methodology for discourse management in the treatment of aphasia. Clinics in Communication Disorders, 2, 64-81. Chapman, S. B., & Ulatowska, H. K. (1994). Differential diagnosis in aphasia. In R. Chapey (Ed.), Language intervention strategies in adult aphasia (pp. 121-131). Baltimore: Williams & Wilkins. Chapman, S. B., Ulatowska, H. K., King, K., Johnson, J., & McIntire, D. D. (1995). Discourse in early Alzheimer's disease versus normal advanced aging. American Journal of Speech-Language Pathology, 4, 125-129. Chapman, S. B., Ulatowska, H. K., Franklin, L. R., Shobe, A. E, Thompson, J. L., & McIntire, D. D. (1997). Proverb interpretation in fluent aphasia and Alzheimer's disease: Implications beyond abstract thinking. Aphasiology, 11, 337-350. Chapman, S. B., & Ulatowska, H. K. (1997). Discourse in dementia: Consideration of consciousness. In M. Stamenov (Ed.), Language structure, discourse and the access to consciousness (pp. 155-188). Amsterdam: John Benjamins Publishing Co. Davis, G. A. (1986). Pragmatics and treatment. In R. Chapey (Ed.), Language interventions strategies in adult aphasia (pp. 169-193). Baltimore: Williams & Wilkins. Delis, D. C, Kramer, J., & Kaplan, E. (1984). The California proverb test. Unpublished protocol. DeSanti, S., Koenig, L., Obler, L. K., & Goldberger, J. (1994). Cohesive devices and conversational discourse in Alzheimer's disease. In R. L. Bloom, L. K. Obler, S. De Santi, and J. S. Ehrlich (Eds.), Discourse analysis and applications: Studies in adult clinical populations (pp. 201-215). Hillsdale: Lawrence Erlbaum Associates.

S.B. Chapman, A.P. Highley and J.L. Thompson

Ehrlich, J. S. (1994). Studies of discourse production in adults with Alzheimer's disease. In R. L. Bloom, L. K. Obler, S. De Santi, and J. S. Ehrlich (Eds.), Discourse analysis and applications: Studies in adult clinical populations (pp. 149-160). Hillsdale: Lawrence Erlbaum Associates. Ehrlich, J. S., Obler, L. K., & Clark, L. (1997). Ideational and semantic contributions to narrative production in adults with dementia of the Alzheimer's type. Journal of Communication Disorders, 30,79-99. Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). Mini-Mental State: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189-198. Garcia, L. J., & Joanette, Y. (1994). Conversational topic-shifting analysis in dementia. In R. L. Bloom, L. K. Obler, S. DeSanti, and J. S. Ehrlich (Eds.), Discourse analysis and applications: Studies in adult clinical populations (pp. 161-183). Hillsdale: Lawrence Erlbaum Associates. Goffman, E. (1974). Frame analysis: An essay on the organization of experience. Cambridge, MA: Harvard University Press. Goodglass, H., & Kaplan, E. (1983). The assessment of aphasia and related disorders (2nd ed.). Philadelphia: Lea & Febiger. Grice, P. (1975). Logic and conversation. In P. Cole and J. Morgan (Eds.), Syntax and semantics 9: Pragmatics (pp. 41-58). New York: Academic Press. Hamilton, H. (1994). Conversations with an Alzheimer's patient. Cambridge: Cambridge University Press. Henderson, V. W. (1996). The investigation of lexical semantic representation in Alzheimer's disease. Brain and Language, 54, 177-178. Hier, D., Hagenlocker, K., & Shindler, A. (1985). Language disintegration in dementia on a picture description task. Brain and Language, 25, 117-133. Hodges, J. R., Patterson, K., Graham, N., & Dawson, K. (19%). Naming and knowing in dementia of Alzheimer's type. Brain and Language, 54, 302-325. Holland, A. (1982). Observing functional communication of aphasic adults. Journal of Speech and Hearing Disorders, 47, 50-56. Hughes, C. P., Berg, L., & Danziger, W. L. (1982). A new clinical scale for the staging of dementia. British Journal of Psychiatry, 140, 566-572. Kempler, D., & Zelinski, E. M. (1994). Language in dementia and normal aging. In F. A. Huppert, C. Brayne, and D. W. O'Connor (Eds.), Dementia and normal aging (pp. 331-365). New York: Cambridge University Press. Kempler, D., Van Lancker, D., & Read, S. (1988). Proverb and idiom comprehension in Alzheimer disease. Alzheimer Disease and Associated Disorders, 2, 38-49. Kontioloa, P., Laaksonen, R., Sulkava, R., & Erkinjuntti, T. (1990). Pattern of language impairment is different in Alzheimer's disease and multi-infarct dementia. Brain and Language, 38, 364-383. Levinson, S. C. (1983). Pragmatics. New York: Cambridge University Press. Martin, A., & Fedia, P. (1983). Word production and comprehension in Alzheimer's disease: The breakdown of semantic knowledge. Brain and Language, 19, 124-141. Matthews, P. J., Obler, L. K., & Albert, M. L. (1994). Wernicke and Alzheimer on the language disturbances of dementia and aphasia. Brain and Language, 46,439-462. McKhann, G., Drachman, D., Folstien,M., Katzman, R., Price, D., & Stadlan, E. M. (1984). Clinical diagnosis of Alzheimer's disease. Neurology, 34,939-944. Nicholas, M., Obler, L. K., Albert, M. L., & Helm-Estabrooks, N. (1985). Empty speech in Alzheimer's disease and fluent aphasia. Journal of Speech and Hearing Research, 28,405-410. Obler, L. K. (1981). Review of Le language des dements. Brain and Language, 72, 375-386. Obler, L. K. (1983). Language and brain dysfunction in dementia. In S. Segalowitz (Ed.), Language Junctions and brain organization (pp. 267-282). New York: Academic Press. Obler, L. K., & Albert, M. L. (1984). Language in aging. In M. L. Albert (Ed.), Clinical neurology of aging (pp. 245-253). New York: Oxford University Press. Ska, B., & Guenard, D. (1993). Narrative schema in dementia of the Alzheimer's type. In H. H. Brownell and Y. Joanette (Eds.), Narrative discourse in neurologically impaired and normal aging adults (pp. 299-316). San Diego: Singular Publishing Group, Inc. Smith, B. R., & Leinonen, E. (1992). Clinical Pragmatics. New York: Chapman & Hall. Ulatowska, H. K., Allard, L., Donnell, A., Bristow, J., Haynes, S. M., Flower, A., & North, A. J. (1988). Discourse performance in subjects with dementia of the Alzheimer's disease type. In H. Whitaker (Ed.), Neuropsychological studies in nonfocal brain damage (Vol. 2, pp. 108-131). New York: Springer-Verlag. Ulatowska, H. K., & Chapman, S. B. (1991). Neurolinguistics and aging. In D. N. Ripich (Ed.), Geriatric Communication Disorders (pp. 21-38). Austin: Pro-Ed. Ulatowska, H. K., & Chapman, S. B. (1994). Discourse macrostructure in aphasia. In R. L. Bloom, L. K. Obler, S. DeSanti, and J. Ehrlich (Eds.), Discourse in adult clinical populations (pp. 19-46). New York: Lawrence Erlbaum Associates. Ulatowska, H. K., & Chapman, S. B. (1995). Discourse studies. In R. Lubinski (Ed.), Dementia and communication (pp. 115-132) San Diego: Singular Publishing Group. Ulatowska, H. K., Chapman, S. B., & Johnson, J. (1995). Processing of proverbs in aphasics and old-elderly. Clinical Aphasiology, 23, 179-193. Ulatowska, H. K, Chapman, S. B., & Johnson, J. (in press). In H. H. Hamilton (Ed.), Inferencing in processing of text in elderly populations. Anthology on old age and language. Dallas: Garland Publishing. Ulatowska, H. K., Chapman, S. B., Highley, A. P., & Prince, J. A., (in press). Discourse in healthy old-elderly adults: A longitudinal study. Aphasiology.

Ulatowska, H. K., Chapman, S. B., Johnson, J., & Branch, M. S., (submitted). Macrostructure and inferential processing in discourse of aphasic patients. Van Lancker, D. (1990). The neurology of proverbs. Behavioral Neurology, 3, 169-187. Wapner, W. A., Hamby, S., & Gardner, H. (1981). The role of the right hemisphere in the apprehension of complex linguistic materials. Brain and Language, 14, 15-33.

Appendix Fables Former and Sons A farmer worked in a vineyard and became rich. He wanted his sons to be just like him. On his deathbed the farmer told his sons that there was a great treasure buried in the vineyard. After the farmer died, the sons went to the vineyard and dug up the soil. They could not find a buried treasure. At harvest time, the vineyard produced the best grapes ever. Now the sons understood the meaning of the treasure.

Raven and Pigeons A hungry raven saw that pigeons in the pigeon coop had a lot of food. He painted his feathers white to look like them. But when he started to crow, they realized that he was a raven and chased him away. So he returned to his own kind. But the other ravens did not recognize him because he had his feathers painted white, so they also chased him a way.

Two Roosters Two roosters were fighting over the chicken yard. The one who was defeated hid in the corner. The other rooster flew to the top of the roost and began crowing and flapping his wings to boast of his victory. Suddenly, an eagle swooped down, grabbed the rooster and carried him away. This was good luck for the defeated rooster. Now he could rule over the roost and have all the hens that he desired.

Single frame picture The picture is a Norman Rockwell print depicting a rural scene around the 1930's. The characters in the picture (going from left to right) are an older man dressed in workclothes, looking down at the ground; a young man dressed in a suit, looking off to the left side of the picture; and a sheep dog with his head on the young man's lap, looking at the ground. The older man has two hats in his hand and is smoking a rolled cigarette. The young man has a suitcase beside him with a State U. sticker on it, books on the suitcase, and a small wrapped package in his lap. They are sitting next to each other on the running board of an old truck. The truck says "Bar W Ranch" on the door and has built up wooden sides on the truck bed. In the foreground of the picture is a bare dirt ground, and the side of a railroad track. In the left hand corner of the picture there is a red lantern and a red flag sitting on a box.

Proverbs Spontaneous Familiar 1. Too many cooks spoil the broth. 2. Don't count your chickens before they are hatched.

Multiple choice 1.

Too many cooks spoil the broth. A) One person can make soup better than ten. B) Too many cooks, the broth is the first course. C) A penny saved is a penny earned.

Unfamiliar 3. One swallow doesn't make a summer. 4. The used key is always bright

S.B. Chapman, A.P. Highley and J.L. Thompson D) A task is at risk when more people are involved than are needed.

2.

Don't count your chickens before they are hatched. A) One shouldn't always assume that things will turn out the way one expects. B) The good is the enemy of the best. C) Chickens don't continue to sit on eggs after they have hatched. D) There may be fewer chicks than there were eggs.

3.

One swallow doesn't make a summer. A) It's not wise to draw a conclusion based on a single example. B) One shouldn't think that winter is over just because the first bird has arrived. C) People who live in glass houses shouldn't throw stones. D) One swallow of a cold drink will only make you want more.

4.

The used key is always bright. A) The key to success is to always use bright ideas. B) Do unto others as you would have others do unto you. C) Regular practice results in the best performance. D) Tools stay shiny when one frequently works with them.

/. Neurolinguistics, Vol. 11, NQS 1-2, p. 79-87, 1998 © 1998 Published by Elsevier Science Ltd. All rights reserved

Coherence and informativeness of discourse in two dementia types Matti Laine*, Minna Laaksof, Elina Vuorinen*, and Juha Rinne* * University of Turku, Finland; University of Helsinki, Finland.

Abstract—We examined the coherence and informativeness of discourse in vascular dementia (VaD) and probable Alzheimer's disease (AD) by analyzing work history interviews in 8 patients with VaD, 11 patients with AD, and 19 age- and education-matched normal controls. The patient groups had comparable levels of cognitive impairment (mild-to-moderate dementia). The results show that both VaD and AD patients exhibited impaired global thematic coherence and reduced informativeness in their discourse. By contrast, the degree of local coherence between two successive utterances did not reliably differentiate the patient groups from the controls. A case-by-case analysis indicated that severely impaired global coherence was found among the ADpatients only. Correlational analyses showed that global coherence was the only discourse variable related to conceptual/semantic impairment.

Introduction This study focuses on certain aspects of discourse in the two most common dementia types, Alzheimer's dementia (AD) and vascular dementia (VaD). Comparison of these dementia types is of interest as the few relevant studies have dealt with AD patients and discourse analyses of VaD patients are lacking, even though VaD accounts for about one third of dementia cases. Studies of language-related functions in AD have revealed widespread changes, particularly in semantically mediated linguistic tasks. Such tasks include language comprehension, picture naming, word-picture matching, and word classification. As regards spontaneous speech, fluent and grammatical but empty output has been considered as most typical in early AD (Obler & Albert, 1981). There are fewer studies on language functions in VaD but some authors at least have suggested a pattern somewhat different from AD: less fluent, dysarthric output is considered more common. Anomia, a cardinal feature of acquiredlanguage disorders, has been noted in VaD as well (Villardita, 1993). Extensive cognitive deterioration in dementia, affecting both linguistic and nonlinguistic functions, should affect the pragmatic aspects of communication as well. In particular, one would expect that semantic breakdown and attentional/memory disorders would impair the maintenance of a coherent discourse. Indeed, it has been reported that in conversations and narratives, AD patients show less cohesion and coherence (Appell, Kertesz& Fisman 1982; Ripich & Terrell, 1988; Glosser & Deser, 1990) and appear less aware of their errors (McNamaraet al., 1992). In addition, noninformative, empty speech output has been connectedto AD (Nicholas et al., 1985). We do not know of any similar analyses with VaD patients.

M. Laine, M. Laakso, E. Vuorinen and J. Rinne

We analyzed two discourse phenomena, coherence and informativeness, in patients with VaD and AD. These were chosen because they are crucial for the integrity of topical content in discourse and previous studies have suggested that the discourse of AD patients is characterizedby decreasedcoherence and'emptiness'. To the best of our knowledge, this is the first study that extends systematic discourse analysis to VaD as well. Moreover, in earlier studies the cognitive correlates of impaired discourse in dementia received only scant attention. Our previous lexical-semantic analyses of the present patient samples (Laine, Vuorinen & Rinne 1997) allow us to determine whether these changes in discourse are related to concomitant deficits in semantic function and lexical retrieval. Materials and Methods Subjects All subjects were tested at the Department of Neurology, University of Turku. Informed consent was obtained prior to testing. Details of the diagnostic procedures are presented in Laine et al. (1997). Background information is summarized in Table 1. The VaD patients were demented and showed clinical and neuroradiological evidence of cerebrovascular disease which was temporally related to the onset of cognitive impairment showing stepwise progression. Thus, they fulfilled the recently proposed criteria for probable vascular dementia (Roman et al., 1993). The patients with AD met the NINCDS-ADRDA clinical criteria for probable AD (McKhann et al., 1984). Severely demented patients were excluded from both groups. In the present study, the number of participants was slightly smaller than in Laine et al. (1997): taped conversation was missing for three subjects, one VaD patient was too dysarthric for discourse analysis, and one AD patient produced too little spontaneous speech. As Table 1 indicates, all groups were matched by age and educational level. In addition, the two patient groups were equated on average Mini-Mental State Examination (MMSE) (Folstein, Folstein & McHugh, 1975) performance and average Clinical Dementia Rating (Hughes et al., 1982). Approximate disease duration was also comparable in the dementia groups. Analysis of conversational speech All subjects were interviewed concerning their work history and this interview was recorded. Tapes were transcribed and the protocols were analyzed. To ensure comparable speech samples for the subjects, their first twenty utterances in the interview were identified. 'Utterance' was defined by syntactic and semantic criteria, as well as by connectives and/or pauses (for comparable criteria, see e.g., Glosser, Wiener & Kaplan (1988)). They ranged from complete clause structures to syntactically incomplete utterances. However, as one of our objectives was to study the topical coherence of discourse, minimal verbal responses and interactional phrases not maintaining the topic were excluded. In addition, occasional unclear utterances (due to poor recording and/or dysarthria) were left out of the analysis.

F=0.0; p=n.s.

F=1.04;p=n.s.

F=1.01;p=n.s.

F=1.40;p=n.s.

Education was defined by a three-level scale (1 = primary school level, 2 = secondary school level, 3 = high school/college level). A modified Clinical Dementia Rating (Hughes et al. 1982) was employed: 0 = no dementia, 1 = questionable dementia, 2 = mild dementia, 3 = moderate dementia.

Group comparisons (Tukey's method) showed that both patient groups differed significant-ly (p < .01) from the controls but not from each other.

Table 1. Background information on the subject groups: means, standard deviations and ranges for age, education, disease duration in months, Mini-Mental State score, and Clinical Dementia Rating Two independent judges scored each utterance for four different features. The first two features relate to coherence. Coherence refers to the appropriate maintenance of topic in discourse (Halliday & Hasan, 1976). Two aspects of coherence were differentiated: (a) local coherence is an estimate of how closely an utterance is thematically related to the immediately preceding utterance (or, in the case of the first utterance, to the question posed by the examiner). The scale ranged from 1 to 5 where 1 = very low and 5 - very high local coherence (cf. Glosser & Deser, 1990). (b) The global coherence measure evaluates the relatedness of the utterance to the general topic of the examiner's question. This scale also ranged from 1 to 5 where 1 = very low and 5 = very high global coherence. The latter two discourse features relate to the informative content conveyed by an utterance: (c) use of non-referential lexical items. These were instances where a referring lexical item was used but its referent was not specified or evident in the immediate context. Each utterance was scoredfor the existence of non-referential items (0 = no, 1 = yes), (d) The informativeness scale evaluates the existence and extent of new information in an

M. Laine, M. Laakso, E. Vuorinen and J. Rinne

utterance. A three-point scale was employed where 0 = no new information, 1 = partially new information, and 2 = new information conveyed. Interrater reliability for the four variables was satisfactory. Product-moment correlations between the two independent scorings were as follows: local coherence .65, global coherence .73, non-referential lexical items .81, and informativeness .65. The final data matrix was preparedby averaging the two ratings. Utterance

Non-ref. items

Examiner: tell me a little about the work you had Patient: now I haven't had for a long time well. I haven't had any permanent ones anymore anymore I sell sold well my son [has] them or I gave them to him and he has just been using them for a longer time Examiner: uhm Patient: so [they] are own indeed Patient with AD

Examiner: so tell me about your work history Patient: well we have been fixing that new building

Subject: well... 34 years I was... working in a prison

first 17 years at the N.N. local prison

and after that another 17 years at the N.N. regional prison

Control subject Examiner: [...] and could you tell a bit about that profession, your profession

Table 2. Discourse extracts from AD, VaD, and control subjects

Examples of the scoring of coherence and informativeness in the subject groups In order to give a better idea of the scoring method we employed, Table 2 includes samples of discourse from an AD patient, a VaD patient, and a normal control. These examples also highlight some of the difficulties our dementia patients had in the interviews. Measures related to lexical retrieval Our previous analysis of the present subjects provided us with several variables that reflect lexical retrieval. These will be only summarized here as they are described in detail in Laine et al. (1997). By correlating these variables with the coherence and informativeness measures, we wished to study to what extent disorders of lexical function are related to discourse impairments (coherence and informativeness) in our dementia patients. (a) Overall naming performance was assessed by the Finnish version of the Boston Naming Test (BNT: Laine et al., 1994). The naming score was the number of items that were spontaneously named correctly within 25 seconds. (b) Semantic abilities related to lexical retrieval were measured by three tasks. Two of them were presented during naming of 30 BNT items (see Laine et al. (1997) for details). The first was a multiple-choice task where the subject was to choose the specific semantic feature related to the BNT item to be named from among four alternatives. For example, the following alternatives were offered for the target 'helicopter': one moves with it in the water, one flies with it, it moves on tracks, it moves in the snow. In the second semantic multiple-choice task, the subject had to find the correct superordinate for the BNT item to be named from among eight written category names (furniture, building, plant, animal, vehicle, tool, musical instrument, and game accessory). The third semantic task was administered separately: the odd-oneout task requiredthe subject to decidewhich picture in a group of five does not go with the others. Because all of the pictures are semantically related, successful performance required fine-grained semantic analysis (e.g., desk, chair, sofa, rocking chair, stool). According to the statistical analysis reported by Laine et al. (1997), the two dementia groups were inferior to the control group on all four of these variables. Results Performance of the subject groups on the discourse-related measures Group comparisons were performedby one-way ANOVAs. (a) Local coherence (Figure 1): For this variable, only a marginally significant effect was present (F(2,35) = 4.68; Levene's test p=.03; Brown-Forsythe equality of means test p = .07). Figure 1 shows that both patient groups tended to have lower local coherence values than the controls did. (b) Global coherence (Figure 2). The one-way ANOVA revealed a significant main effect (F(2,35) = 7.55; p = .002). Group comparisons (Tukey's test) showed that both the VaD and the AD patients were significantly inferior to the controls (p values <. 05. and < .01, respectively).

M, Laine, M, Laakso, E. Vuorinen and J. Rinne

(c) Use of non-referendal nouns/pronouns (Figure 3) also showed a significant main effect in the one-way ANOVA (F(235) = 7.29; Levene's test p < .0001; Brown-Forsythe equality of means test p = .03). In group comparisons, both the VaD and AD groups produced significantly more non-referential items than the controls. (d) Informativeness (Figure 4). A significant main effect was obtained on this variable as well (F(2,35) = 13.26; Levene's test p < .0001; Brown-Forsythe equality of means test p = .003). Group comparisons confirmed that both patient groups produced significantly less informative output than the controls did.

Figure 1. Local coherence ratings for the VaD, AD, and control subjects.

Figure 2. Global coherence ratings for the VaD, AD, and control subjects.

Figure 3. Proportion of utterances including non-referential lexical items for the VaD, AD, and control subjects.

Figure 4. Informativeness ratings for the VaD, AD, and control subjects. In addition to the group analysis, we searched individual protocols for sequences of severely disrupted global coherence. This analysis suggests that severely impaired global coherence may be more common in AD than in VaD: we found three AD patients where both raters had scored a sequence of four or more utterances with the lowest possible global coherence score (= 1). In the VaD group, a maximum of two sequential utterances received this score.

Relationships between lexical-semantic and discourse variables We explored the interrelationships between lexical and discourse variables through correlation coefficients. In this analysis, the two patients groups were combined. Pearson's product-moment correlations were calculated between the four discourse variables (local coherence, global coherence, use of non-referential lexical items, and infonnativeness) and the four lexical-semantic measures (BNT summative score, choice of specific semantic feature, choice of superordinate, odd-one-out task performance). The correlation coefficients are given in Table 3. Local coherence

Table 3. Correlation coefficients between lexical-semantic and discourse-related variables in the combined group of VaD and AD patients (n = 19)

M. Laine, M. Laakso, E. Vuorinen and J. Rinne

Table3 indicates that only global coherence correlates significantly with semantic abilities (choice of superordinate, odd-one-out task performance) in our dementia patients. Both local and global coherence correlate with overall naming but those relationships could be caused in part by the common variance BNT performance has with general cognitive impairment (correlation between BNT andMMSE score = .52; p < .05). Informativeness correlates significantly with BNT score whereas the use of non-referential lexical items does not show significant correlations with any of the variables examined here. Discussion This study had two main purposes. The first was to confirm the existence of certain pragmatic difficulties in dementia patients. We also wanted to see whether the pragmatic difficulties reported in AD patients apply to VaD patients as well. On the basis of our results, both hypotheses were confirmed. Our study replicated previous findings with regard to impaired global but not local coherence (Glosser & Deser, 1990), abnormally frequent use of non-referential lexical items (Ripich & Terrell, 1988), and decreasedinformativeness (Nicholas et al., 1985) in AD patients. Moreover, with our limited patient samples, we were able to show that these findings apply to VaD patients as well. On the basis of the discourse measures used here, the two patient groups behaved quite similarly fashion. The only possible difference, worth exploring in further studies, relates to global coherence where case-by-case analyses suggested that AD patients may be somewhat more impaired than VaD patients. The correlation analyses revealed an interesting clue to the underlying mechanisms of impaired global coherence. Global coherence was the only discourse variable that correlated significantly to performance level at semantic processing tasks. Thus global coherence in discourse may have some of its roots in 'deep', conceptual/semantic structures. As was noted in the introduction, semantic impairment is very often present even at the early stages of dementia. Local coherence, on the other hand, reflects the relationship of an utterance to the preceding one, and it can remain quite good even when the overall topic of conversation becomes lost. In our study, informativeness of discourse correlated significantly with overall BNT performance. It is quite understandable that the word-retrieval difficulties many of our dementia patients suffered from hampered their ability to convey new information during conversation. On the other hand, the use of non-referential lexical items did not show statistically significant correlations to semantic-lexical variables. One might speculate that the use of such non-informative items is a strategic choice which only indirectly reflects word-finding difficulty. Moreover, the strict and dichotomous scoring (existent vs. nonexistent) for non-referential lexical items may have affected the magnitude of the correlation coefficients. In conclusion, the present study shows that both VaD and AD patients exhibit impaired global coherence and informativeness in their discourse. By contrast, the local coherence measure did not reliably differentiate the patient groups from the matched controls. Global coherence appeared to be most impaired in the AD group, but lowered scores for this measure were observed in the VaD group too. Correlational analyses showed that among the discourse variables studied, global thematic coherence was the only one that was significantly related to the degree of conceptual/semantic impairment in dementia.

References Appell, J., Kertesz, A., & Fisman, M. (1982). A study of language functioning in Alzheimer patients. Brain and Language, 17, 73-91. Folstein, M. F., Folstein, S. E, & McHugh, P. R. (1975). "Mini-Mental State": A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189-198. Glosser, G., & Deser, T. (1990). Patterns of discourse production among neurological patients with fluent language disorders. Brain and Language, 40, 67-88. Glosser, G., Wiener, M., & Kaplan, E. (1988). Variations in aphasic language behaviors. Journal of Speech and Hearing Disorders, 53, 115-124. Halliday, M. A. K., & Hasan, R. (1976). Cohesion in English. London: Longmans. Hughes, C. P., Berg, L., Danziger, W. L., Coben, L. A., & Martin, R. L. (1982). A new clinical scale for staging of dementia. British Journal of Psychiatry, 140, 566-572. Laine, M., Goodglass, H., Niemi, J., Koivuselka-Sallinen, P., Tuomainen, J., & Marttila, R. (1994). Adaptation of the Boston Diagnostic Aphasia Examination and the Boston Naming Test into Finnish. Scandinavian Journal ofLogopedics and Phoniatrics, 18, 83-92. Laine, M., Vuorinen, E., & Rinne, J. (1997). Picture naming deficits in vascular dementia and Alzheimer's disease. Journalof Clinical and Experimental Neuropsychology, 19, 126-140. McKhann, G., Drachman, D., Folstein, M., Katzman, R., Price, D., & Stadlan, E. M. (1984). Clinical diagnosis of Alzheimer's disease. Report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer's disease. Neurology, 34, 939-944. McNamara, P., Obler, L. K., Au, R., Durso, R., & Albert, M. L. (1992). Speech monitoring skills in Alzheimer's disease, Parkinson's disease, and normal aging. Brain and Language, 42,38-51. Nicholas M., Obler, L., Albert, M., & Helm-Estabrooks, N. (1985). Empty speech in Alzheimer's disease and fluent aphasia. Journal of Speech and Hearing Research, 28,405-410. Obler, L. K., & Albert, M. L. (1981). Language in the elderly aphasic and dementing patient. In M.T. Sarno (Ed.), Acquired Aphasia (pp. 385-398). New York: Academic Press. Ripich, D. N., & Terrell, B. Y. (1988). Patterns of discourse cohesion and coherence in Alzheimer's disease. Journal of Speech and Hearing Disorders, 53, 8-15. Roman, G. C., Tatemichi, T. K., Erkinjuntti, T., Cummings, J. C., Masdeu, J. C., Garcia, J. H., Amaducci, L., Orgogozo, J.-M., Brun, A., Hofman, A., Moody, D. M., O'Brien, M. D., Yamaguchi, T., Grafman, J., Drayer, B. P., Bennett, D. A., Fisher, M., Ogata, J., Kokmen, E., Bermejo, F., Wolf, P. A., Gorelick, P. B., Bick, K. L., Pajeau, A. K., Bell, M. A., DeCarli, C., Culebras, A., Korczyn, A. D., Bogousslavsky, J., Hartmann, A., & Scheinberg, P. (1993). Vascular dementia: Diagnostic criteria for research studies. Report of the NINDS-AIREN International Workshop. Neurology, 43, 250-260. Villardita, C. (1993). Alzheimer's disease compared with cerebrovascular dementia. Neuropsychological similarities and differences. Acta Neurologica Scandinavica, 87, 299-308.

J. Neurolinguistics, Vol. 11, Nos 1-2, p. 89-102, 1998 © 1998 Published by Elsevier Science Ltd. All rights reserved Printed in Great Britain 0911-6044/98 $19.00 + 0.00

Affective prosodic disturbance subsequent to right hemisphere stroke: A clinical application Robert T. Wertz*f, Constance R. Henschelf, Linda L. Anther*, John R. Ashford*, and Howard S. Kirshnerf 'Veterans Administration Medical Center, Nashville, Tennessee Vanderbilt University School of Medicine, Nashville, Tennessee

Abstract—We examined Ross' (1981) hypothesis regarding the disruption of affective prosody subsequent to right hemisphere brain damage (RHD). Twenty patients who had suffered a right hemisphere stroke were compared with 18 normal, non-brain-damaged subjects for affective prosody in spontaneous speech, gesturing accompanying spontaneous speech, repetition of affective prosody, comprehension of affective prosody, and comprehension of affective gestures. In addition, we attempted to classify our RHD subjects with Ross' prosodic taxonomy, determine the relationship between classification and site of lesion, and explore the contribution of coexisting dysarthria to prosodic disturbance. All RHD subjects displayed affective prosodic disturbance in spontaneous speech. Only one normal subject was judged mildly dysprosodic. The RHD group had significantly more difficulty in repeating affective prosody and comprehending affective gestures. There were no significant group differences in gestures accompanying spontaneous speech or comprehending affective prosody. Eighty percent of the RHD subjects were classified with the Ross taxonomy, however there was no systematic relationship between classification and site of lesion. And, while dysprosody without a coexisting dysarthria was present in three RHD subjects, 17 displayed a coexisting dysarthria. We conclude that affective dysprosody is common subsequent to RHD, however the relationship between classification of dysprosody and site of lesion or the contribution of coexisting dysarthria to dysprosody is not clear.

Introduction The perception and production of linguistic and affective prosody and their disruption subsequent to brain damage have received considerable debate in the literature. Currently, it is not clear whether brain damage does or does not disrupt prosody (Baum & Pell, 1997); whether linguistic prosody is disrupted by a left hemisphere lesion and affective prosody is disrupted by a right hemisphere lesion (Behrens, 1988; Bradvik et al., 1991; Bryan, 1989; Cancelliere & Kertesz, 1990; Emmery, 1987; Gorelick & Ross, 1987; Shapiro & Danley, 1985; Weintraub, Mesulam, & Kramer, 1981); whether right hemisphere lesions disrupt both the perception and production of affective prosody, only perception of affective prosody, or only the production of affective prosody (Gorelick & Ross, 1987; Heilman et al., 1984; Lalandeet al., 1992); or whether prosodic disturbance results from dysarthria (Duffy & Folger, 1996; Kent & Rosenbek, 1982; Ropper, 1987) rather than another process independent from involvement of the muscles utilized in speech. Tompkins (1995) has summarized the data regarding prosodic disturbance subsequent to right hemisphere brain damage. She observes: "... some right hemisphere damaged adults have speech that is less varied intonationally than that of non-brain-damaged adults ...

R.T. Wertz, C.R. Henschel, L.L. Auther, J.R. Ashford and H.S. Kirshner

some right hemispheredamagedadults sound hypermelodic ..." and "whether linguistic or emotional prosody is the subject of concern, the evidence on prosodic production problems after right hemisphere damage is mixed" (pp. 21-22). Prosody defined Several definitions of prosody exist, but all are similar. Monrad-Krohn (1947) defined prosody as the melodic line of speech produced by variations of pitch, rhythm, and stress of pronunciation. For Hargrove and McGarr (1994), prosody represents the linguistic use of vocal aspects of speech without consideration of the segmental aspects (speech sounds or phonemes). They observe that the rules of prosody are systematic, bound by conventions, and convey important information to listeners. Tompkins (1995) says prosody includes elements of speech melody, rate, stress, juncture, and duration; has a pervasive influence on communication; and is critical in speech production and perception. Prosody can convey linguistic meaning—whether the utterance is a statement or a question or whether a person lives in the "white house" or the "White House"—or emotional or affective meaning—whether the speaker is happy, sad, angry, surprised, etc. Affective prosody Ross (1981) has popularized the notion that the right hemisphere is responsible for producing and perceiving affective prosody. Specifically, he suggests that ". . . affective components of language, encompassing prosody and emotional gesturing, are a dominant function of the right hemisphere, and their functional-anatomical organization in the right hemisphere mirrors that of prepositional language in the left hemisphere" (p. 561). Data to support this position are provided by ten patients (Ross, 1981); an additional 14 patients (Gorelick& Ross, 1987); and a case report (Wolfe& Ross, 1987). Explorations of Ross' hypothesis indicate mixed results. There is considerable evidence that right hemisphere brain damage disrupts perception of affect (Blonder, Bowers, & Heilman, 1991; Bryan, 1989; Darby, 1993; Heilman et al., 1984; Lalande et al., 1992). However, Van Lancker and Sidtis (1992) and Cancelliere and Kertesz (1990) report that either a left or right hemisphere lesion may result in problems perceiving prosody. And, Tompkins and Flowers (1985) suggest differences in the perception of emotional intonation between left and right brain damaged people may depend on the difficulty of the task. Support for deficits in the production of affective prosody subsequent to right hemisphere brain damage, other than reports by Ross and colleagues, must be qualified. It ranges from Baum and Pell's (1997) observation that the right hemisphere is not specifically engaged in the production of affective prosody to Weintraub, Mesulam, and Kramer's (1981) suggestion that right hemisphere damage may affect both linguistic and affective prosody. In between is Bradviket al.'s (1991) report that prosodic impairment following a right hemisphere lesion could be linguistic in nature and not secondary to an affective disorder; Cancelliere and Kertesz's (1990) results that indicate both left and right hemisphere lesions disrupt expression of emotional prosody; and Shapiro and Danley's (1985) observation that right anterior and central lesions disrupt emotional and nonemotional prosody, and right posterior lesions result in exaggerated pitch variation and intonational range in producing both emotional and nonemotional prosody.

Nevertheless, Ross and colleagues have provided a testable hypothesis. Specifically, they contend that right hemisphere brain damage results in impaired spontaneous, affective prosody and gesturing; impaired repetition of affective prosody; impaired comprehension of affective prosody; and impaired comprehension of emotional gesturing. Moreover, Ross (1981) has provided a taxonomy for classifying what he calls the "aprosodias" into types similar to those seen in aphasia subsequent to left hemisphere brain damage. These include: motor, sensory, global, conduction, transcortical motor, transcortical sensory, mixed transcortical, and anomic. And, as in the classification of aphasic types, Ross provides an anatomical relationship in the right hemisphere for each aprosodic type. Dysarthria Inpaired speech motor control associated with dysarthria can impair prosody (Darley, Aronson, & Brown, 1975; Duffy, 1995). And, Tompkins (1995) indicates that dysarthria may contribute to the perception of restricted intonational variability in some right hemisphere damagedadults. Thus, we need to ask whether disrupted linguistic or affective prosody subsequent to right hemisphere brain damage results from some prosodic mechanism that resides in the right hemisphere or dysarthria, for example, associated with a unilateral upper motor neuron lesion. Kent and Rosenbek (1982), discussing prosodic disturbance and neurologic lesion, observed, "The right cerebral hemisphere . . . appears to have a privileged role in the processing of prosodic and affective information" (p. 285). Tentatively, they called the speech pattern observed in their patients with a right-cerebral lesion a dysarthria. While they recognized that there may be some dispute about the suitability of this terminology, they point out several similarities between the speech patterns of patients with a rightcerebral lesion and patients who display a hypokinetic dysarthria associated with Parkinson's disease. Similarly, Ropper's (1987) patients with a right hemisphere stroke displayed severe dysarthria characterized by slowness of speech; incomplete pronunciation of each syllable; hypophonic, slightly harsh voice quality; and restricted amplitude that resulted in monotonous speech. Duffy and Folger (1996) report that some of their patients with right upper motor neuron lesions produced prosodic abnormalities that were characterized by slow rate, increased rate in segments, excessive and equal stress, or combinations of these signs. And, Hird and Kirsner's (1993) comparison of patients with right hemisphere lesions, hypokinetic dysarthria, or ataxic dysarthria failed to reveal significant acoustic differences among groups in the production of linguistic or emotional prosody.

Clinical questions Several generations of speech-language pathologists have been taught that a right hemisphere lesion will impair production and perception of affective prosody, and these deficits can be observed by clinical, bedside appraisal. We examined this premise by evaluating spontaneous speech, spontaneous gesturing, repetition of affective prosody, comprehension of affective prosody, and comprehension of affective gesturing in patients who had suffered a right hemisphere stroke and in normal, non-brain-damaged adults. Specifically, we asked: does perception and production of affective prosody and gesturing differ between right hemisphere damaged patients and non-brain-damaged adults, can

R.T. Wertz, C.R. Henschel, L.L. Auther, J.R. Ashford and H.S. Kirshner

prosodic disturbance subsequent to right hemisphere damage be classified according to the Ross (1981) taxonomy, what is the anatomical relationship to prosodic classification in right hemisphere damaged patients, and what is the relationship of dysarthria to prosodic disturbance in right hemisphere damaged patients? Methods Subjects Twenty consecutive patients who had suffered a right hemisphere stroke and who were admitted to an inpatient rehabilitation program were evaluated. All met the following selection criteria: had suffered a first, single, right hemisphere stroke; displayed no significant aphasia; and had no other, coexisting neurological disease. Eighteen normal, non-brain-damaged volunteers from the community were matched with the right hemisphere sample for age and education. All had no history of stroke or other neurological disease. Table 1 shows demographic information for the two groups. Right hemisphere damaged(RHD) subjects,. 11 males and nine females, hada mean age of 64.75 years and a mean education of 13.00 years. Normal subjects, ten males and eight females, had a mean age of 59.94 years and a mean education of 14.72 years. There was no significant difference (p < .05) in age or years of education between groups. Fifteen of the RHD subjects had suffered an ischemic stroke, and five had suffered a hemorrhagic stroke. Seventeen of the 20 RHD subjects displayed right hemiplegia. The RHD group was an average of 5.2 weeks postonset, with a range of one to 12 weeks postonset.

Age in Years Education in Years Weeks Postonset Gender Male Female Stroke Ischemic Hemorrhagic Right Hemiplegia

Table 1. Demographic data for the right hemisphere-damaged (RHD) and normal groups.

Measures We followed the clinical protocol provided by Ross (1981), as closely as possible, to evaluate prosody, dysarthria, and gesturing in spontaneous speech; repetition of affective prosody; comprehension of affective prosody; and comprehension of affective gesturing. All tasks were presented by the same speech-language pathologist to all RHD and normal subjects. Scoring was done "on-line" at the time of evaluation. Audio recordings were made of subjects' spontaneous speech samples and repetitions of affective prosody for later intra- and interjudge reliability. All RHD subjects were evaluated on the rehabilitation ward, and all normal subjects were evaluated in a speech clinic or their homes. Spontaneous prosody, dysarthria, and gesturing To evaluate prosody, dysarthria, and gesturing in spontaneous speech, each RHD subject was asked, "Tell me how you feel about your illness and being in the hospital?" Each normal subject was asked, "Have you ever been really sick or know someone who was? Tell me about it. "A minimum of two minutes of spontaneous speech was obtained from each RHD and normal subject. Prosody in spontaneous speech was rated on an eight-point, equally-appearing interval scale, ranging from "0," no affective prosodic disturbance, to "7," most severe prosodic disturbance. Prosody judgements were based on each subject's speech melody, rate, stress, juncture, and duration and, specifically, whether the prosody was affectively appropriate for the content of the subject's response. Similarly, dysarthria was rated on an eight-point scale, ranging from "0," no dysarthria, to "7," most severe dysarthria. The presence and severity of dysarthria was based on the subject's use of respiration, phonation, articulation, and resonance to produce intelligible and naturally sounding speech (Wertz & Rosenbek, 1992). Gesturing during spontaneous speech was rated on a three-point scale, ranging from "0," normal; through "1," reduced;to "2," absent. Repetition of affective prosody To evaluate repetition of affective prosody, each subject was instructed, "I'll say a sentence. You say the sentence back to me exactly as I said it to you." Six sentences, each conveying a different emotion—happy, sad, anger, surprise, tearful, disinterest—were presented, live voice, with the examiner facing the subject. Sentence stimuli were: "You bought a new car?" "We'll be back tomorrow." "I've got to go to Chicago." "It's fine with me." "We're going to the movies." "It's time to leave now." Each subject's repetition of each sentence was scored"+," appropriate prosody that matched the examiner's production, or "-," inappropriate, reduced, or absent prosody compared with the examiner's production. Thus, performance could range from zero to six correct. Comprehension of affective prosody The same sentences and emotions were used to assess comprehension of affective prosody. Each subject was instructed, "I'll say a sentence, and you tell me how I feel or point to the word that indicates how I feel." A different randomization of words, in large type, indicating the six emotions—happy, sad, anger, surprise, tearful, disinterest—was placed

R.T. Wertz, C.R. Henschel, L.L. Auther, J.R. Ashford and H.S. Kirshner

before the subject for each stimulus presentation. Verbal or gestural responses—pointing to a word—were accepted. The examiner stood behind the patient and spoke each sentence conveying a specific emotion. Plus or minus scoring was used, thus performance could range from zero to six correct. Comprehension of affective gesturing The same emotions were used to assess comprehension of affective gestures. Each subject was instructed, "I'll show you a gesture, and you tell me how I feel or point to the word that indicates how I feel." As in the comprehension of affective prosody task, a different randomization of words, in large type, indicating the six emotions was placed before the subject for each stimulus presentation. Verbal or gestural responses were accepted. The examiner stood in front of the subject and conveyed each emotion with the appropriate gesture and facial expression. Plus or minus scoring was used, thus performance could range from zero to six correct.

Reliability As indicated above, subject's spontaneous speech and repetition of sentences were audio tape recorded for reliability analysis. Table 2 shows intra- and interjudge reliability. Intrajudge reliability results from the examiner listening to the audio tape recordings, approximately one week after the examination, and re-rating the subject's performance. Interjudge reliability results from a second speech-language pathologist, "blinded" to the subject's group assignment (RHD or normal), listening to the audio tape recordings and rating the subject's performance. A 47% sample was used for interjudge reliability. As shown in Table 2, kappa values (Cohen, 1960) indicate significant interjudge reliability (p < .0005) for the presence of dysprosody and dysarthria in spontaneous speech, severity of dysarthria in spontaneous speech, and severity of dysprosody on the affective repetition task. Interjudge agreement was significant (p < .004) for severity of dysprosody in spontaneous speech. An 89% sample was used for intrajudge reliability. As shown in Table 2, intrajudge reliability was significant (p < .0005) for all presence and severity measures in spontaneous speech and for severity of dysprosody on the affective repetition task. Measure

Spontaneous Speech Presence of Dysarthria Presence of Dysprosody Severity of Dysarthria Severity of Dysprosody

Table 2.

Kappa statistic for interjudge (47% sample) and intrajudge (89% sample) scorer reliability.

Classification and localization The Ross (1981) taxonomy, shown in Table 3, was used to classify subjects into an aprosodic type based on their performance in the spontaneous prosody and gesturing, affective repetition, comprehension of affective prosody, and comprehension of affective gesturing tasks. Ratings of 1 through 7 on the spontaneous prosody task; 1 or 2 on the spontaneous gesturing task; or one or more errors on the affective repetition, comprehension of affective prosody, and comprehension of affective gesturing tasks were used to signify impairment. Thus, for example, as shown in Table 3, a subject with a rating of 1-7 on the spontaneous prosody task; 1-2 on the spontaneous gesturing task; one or more errors on the affective repetition task, and no errors on the comprehension of affective prosody or comprehension of affective gesturing tasks was classified as demonstrating "motor" aprosodia. Similarly, we employed Ross' (1981) suggested site of lesion relationships with the aprosodic types to determine brain-behavior relationships in our RHD subjects. For example, a "motor" aprosodia is predicted to result from a lesion in the right frontal or frontal-parietal areas, and a "sensory" aprosodia is predicted to result from a lesion in the right posterosuperior temporal and posteroinferior parietal lobes. Site of lesion data were available from CT or MRI scans for 16 of our RHD subjects. Type

Motor Sensory Global Conduction Transcortical Motor Transcortical Sensory Mixed Transcortical Anomic

Table 3. Taxonomy for classifying the "aprosodias" from spontaneous speech and gesturing (SPG), repetition of affective sentences (R), comprehension of affective prosody (CP), and comprehension of affective gestures (CG). P = poor and G = good. (After Ross, 1981). Results Table 4 shows a comparison of the presence of dysprosody, dysarthria, and reduced or absent gesturing in the spontaneous speech task between the RHD and normal subjects. None of the normal subjects displayed dysarthria. Conversely, 17 of the 20 RHD subjects were rated dysarthric. A Chi Square analysis indicated significantly more RHD subjects were dysarthric (p < .0001). Similarly, only one of the normal subjects was rated dysprosodic, however all RHD subjects were rated dysprosodic. A Chi Square analysis indicated significantly more RHD subjects were dysprosodic (p < .0001). Nine of the

R.T. Wertz, C.R. Henschel, L.L. Auther, J.R. Ashford and H.S. Kirshner

normal subjects and 13 of the RHD subjects displayed reduced or absent gestures accompanying spontaneous speech. A Chi Square analysis indicated no significant difference (p < .05) between groups in gestural performance accompanying spontaneous speech. Measure

* Significant at p <. 0001 Table 4.

Number of subjects in the right hemisphere damaged(RHD) and normal groups with (+) and without (-) dysarthria, dysprosody, and diminished or absent gestures in spontaneous speech.

Rated performance on all tasks for each group is shown in Table 5. Higher ratings on the spontaneous speech task (dysprosody and dysarthria) indicate more impairment. Higher scores on the spontaneous gesturing, affective repetition, comprehension of affective prosody, and comprehension of affective gesturing tasks indicate less impairment. As implied by the Chi Square analyses, t tests indicated the RHD group displayed significantly more severe dysarthria (p < .001) and significantly more severe dysprosody (p < .001) in spontaneous speech than the normal group. There was no significant difference (p < .05) between groups for gestures accompanying spontaneous speech. The RHD group made significantly more errors than the normal group in repetition of affective prosody (p < .001) and comprehension of affective gestures (p < .001). RHD subjects made more errors in comprehension of affective prosody than normal subjects, however this difference approached but did not reach significance (p < .058). Measure

Spontaneous Speech Dysarthria Dysprosody Gestures Affective Repetition Prosodic Comprehension Gestural Comprehension

* Significant at p<.001 Table 5. Comparison of the normal and right hemisphere damaged (RHD) groups for rated dysarthria and dysprosody (0-7) and gestures (0-2) in spontaneous speech and errors (0-6) in affective repetition, comprehension of affective prosody, and comprehension of affective gestures.

Table 6 shows the type of aprosodia for subjects in each group according to the Ross (1981) taxonomy. Two (11%) of the normal subjects were classified as demonstrating a conduction aprosodia, 11 (61%) were unclassified, and five (28%) were normal. Four (20%) of the RHD subjects were classified as demonstrating a motor aprosodia, 11 (55%) were classified as global aprosodia, one (5%) was classified as transcortical motor aprosodia, and four (20%) were unclassifiable. Type

Motor Sensory Global Conduction Transcortical Motor Transcortical Sensory Mixed Transcortical Anomic Unclassified Normal

Table 6. Classification of the number (N) and percent of normal and right hemisphere damaged (RHD) subjects with Ross' (1981) aprosodic taxonomy. Site of lesion data were available for 16 of our 20 RHD subjects. Table 7 shows a comparison of site of lesion with classification by the Ross (1981) taxonomy. Of the four RHD subjects classified as motor aprosodia, two had a temporal-parietal lesion, one had a subcortical lesion, and site of lesion for one patient was not known. Of the 11 patients classified as global aprosodia, two had a parietal lesion, two had a parietal-occipital lesion, one had a frontal-temporal-parietal lesion, one had a frontal-temporal-parietal-occipital lesion, two had a subcortical lesion, and the site of lesion was unknown for three subjects. The one RHD subject classified as transcortical motor aprosodia had a subcortical lesion. And, of the four RHD subjects who were unclassifiable, one had a frontal-parietal lesion, two had a frontal-temporal-parietal lesion, and one had a subcortical lesion. Dysarthria coexisted with dysprosody in 17 of the 20 RHD subjects, and three displayed dysprosody with no coexisting dysarthria. None of the normal subjects was rated dysarthric. Site

Parietal Frontal-Parietal Temporal-Parietal Parietal-Occipital Frontal-Temporal-Parietal Frontal-Temporal-Parietal-Occipital Subcortical Undetermined

Table 7. Relationship between aprosodic type and site of lesion for the right hemisphere damaged subjects.

R.T. Wertz, C.R. Henschel, L.L. Auther, J.R. Ashford and H.S. Kirshner

Discussion Our clinical test of Ross' (1981) hypothesis about affective components of language, encompassing prosody and emotional gesturing, being a dominant function of the right hemisphere and the functional-anatomical organization of affective prosody in the right hemisphere mirroring that of propositional language in the left hemisphere provides mixed support. We asked: does perception and production of affective prosody and gesturing differ between RHD patients and non-brain damaged adults, can prosodic disturbance subsequent to RHD be classified according to the Ross (1981) taxonomy, what is the anatomical relationship to prosodic classification in RHD patients, and what is the relationship of dysarthria to prosodic disturbance in RHD patients? Reliability It appears clinicians can rate prosodic performance in spontaneous speech and repetition of affective prosody reliably. Intra- and interjudge agreement was significant. However, there is a need to determine intra- and interjudge reliability in rating gestures accompanying spontaneous speech, comprehension of affective prosody, and comprehension of emotional gestures. Moreover, there is a need to determine the consistency of subject's performance on all tasks by comparing test and retest performance. Prosodic disturbance subsequent to RHD Dysprosody appears to be a common occurrence subsequent to RHD. All of our RHD subjects displayed mild to severe prosodic disturbance in spontaneous speech. Only one normal subject was rated dysprosodic, and the rating was mild—"1" on the 0-7 rating scale. The influence of RHD on gesturing during spontaneous speech is debatable. Thirteen (65%) RHD subjects displayed reduced or absent gesturing during spontaneous speech. However, nine (50%) of the normal subjects also displayed reduced or absent gesturing during spontaneous speech, and there was no significant difference in the presence or severity, when reduced or absent, between groups. Thus, our results do not indicate reduction in gestures accompanying spontaneous speech can be attributed solely to RHD. RHD does influence repetition of affective prosody and comprehension of affective gesturing. Normal subjects repeated affective prosody significantly more correctly than RHD subjects. However, eight (45%) of the normal subjects failed on one or two of the affective repetition stimuli. Conversely, 19 (95%) of the RHD subjects failed on one to six of the affective repetition stimuli, and mean performance, 2.50, was significandy lower than normal mean performance, 5.44. Of course, poor repetition of affective prosody may result from failure to comprehend the affect to be repeated, failure to produce the requested affect, or both. Of the eight normal subjects with impaired affective repetition, five also had impaired comprehension of affective prosody. Of the 19 RHD subjects with unpaired affective repetition, 12 had impaired comprehension of affective prosody. None of the normals displayed impaired comprehension of affective gestures. Conversely, 14 (70%) of the RHD subjects displayed impaired comprehension of affective gestures, however severity was mild—one or two errors—in 13 RHD subjects. Comprehension of affective prosody approached (p < .058), but did not reach, significance between groups. Eleven (61%) of the normal subjects displayed problems in

comprehending affective prosody, however none missed more than one of the six stimuli. Conversely, 12 (60%) of the RHD subjects displayed problems in comprehending affective prosody, and performance ranged from zero to five correct for the six stimuli. Ross' (1981) taxonomy predicts that problems comprehending affective prosody is not universal across RHD patients. Those with a motor, conduction, transcortical motor, or anomic aprosodia are expected to have good comprehension of emotional prosody. Thus, group comparison of comprehension performance between RHD subjects and normals may be inappropriate. Classification Sixteen (80%) of our RHD subjects could be classified in the Ross (1981) taxonomy. Moreover, two (11%) of the normal subjects could be classified. Both were classified as demonstrating conduction aprosodia. An additional 11 normal subjects who made errors on the prosodic tasks were unclassifiable in the Ross taxonomy, and five were not classified, because they made no errors on any task. For the RHD group, 11 of the classified subjects demonstrated global aprosodia, four classified as motor aprosodia, and one classified as transcortical aprosodia. All of the four unclassified subjects would have demonstrated motor aprosodia, however three had mildly impaired comprehension of affective gestures, and one had mildly impaired comprehension of affective prosody. Classification of brain damaged patients is problematic. In aphasia, two approaches have been employed. The Western Aphasia Battery (Kertesz, 1982) uses discrete cut-off scores for fluency, auditory comprehension, repetition, and naming. This approach results in classifying almost all aphasic people. The Boston Diagnostic Aphasia Examination (Goodglass and Kaplan, 1983) uses a range of performance in specific behavioral characteristics—melodic line, phrase length, articulatory agility, grammatical form, paraphasiain running speech, repetition, word finding, and auditory comprehension. This approach results in classifying 40 to 60 percent of aphasic people. Our application of the Ross taxonomy to classify our sample of RHD subjects employed discrete cut-off scores. Any rating other than "0" (normal) on the spontaneous speech and gesturing task and any error on the affective repetition, affective comprehension, and comprehension of affective gestures tasks signified impairment. This approach resulted in our failure to classify 20% of our RHD subjects and failure to differentiate RHD patients from 72% of the normal subjects with the classification system. Functional-anatomical relationships One of the purposes in classifying brain damaged patients is to explore brain-behavior relationships. This has been the traditional approach in aphasia, and it appears to be Ross' (1981) approachin classifying the aprosodias subsequent to RHD. Specifically, aprosodic types should relate with lesion localization. Our results imply the relationships between lesion localization and type of aprosodia are not precise. For example, none of the three patients for whom we had localization data and who classified as motor aprosodia had a lesion in the right hemisphere analogous to left hemisphere lesions that cause Broca's aphasia. Of the eight patients on whom we had localization data and who classified as global aprosodia, only two had the large supra- and infra-Sylvian infarction predicted by Ross to result in global aprosodia. Our one patient who classified as transcortical motor

R.T. Wertz, C.R. Henschel, L.L. Anther, J.R. Ashford and H.S. Kirshner

aprosodia had a subcortical lesion similar to that of Ross' (1981) transcortical motor aprosodic patient whose lesion involved the anterior limb of the internal capsule, head of the caudate nucleus, and putamen, without evidence of cortical involvement. This localization in the right hemisphere, as Ross observed, is different from the left medial frontal lesions believed to result in transcortical motor aphasia. Thus, lesion localization in our sample of RHD aprosodic patients does not demonstrate that "functional-anatomic organization in the right hemisphere mirrors that of propositional language in the left hemisphere" (Ross, 1981, p. 561). Dysarthria Kent and Rosenbek (1982) elected to call their patients with dysprosody subsequent to RHD dysarthric. Similarly, Ropper (1987) and Duffy and Folger (1996) report prosodic disturbance in dysarthric patients who have sustained RHD. Thus, it is not clear whether dysprosody subsequent to RHD results from dysarthria or some prosodic mechanism that resides in the right hemisphere. All of our 20 RHD patients displayed prosodic disturbance. Seventeen displayed a coexisting dysarthria, and three did not. Prosodic disturbance in the three patients without dysarthria was mild, "1" or "2" on the seven-point rating scale, in spontaneous speech. However repetition of affective prosody, comprehension of affective prosody, and comprehension of affective gestures ranged from mild to severe in these patients. In the 17 patients with coexisting dysarthria, the severity of dysarthria ranged from mild, "1" on the seven-point rating scale, to moderately severe, "5." Dysprosody ratings in these patients' spontaneous speech ranged from mild, "2," to severe, "6." Similarly, their impairment on the repetition of affective prosody, comprehension of affective prosody, and comprehension of affective gestures ranged from mild to severe. Thus, except for dysprosody in spontaneous speech, there was no apparent difference between RHD patients who had coexisting dysarthria and those who were not dysarthric. The type of dysarthria (Duffy, 1995) present in our RHD patients was either unilateral upper motor neuron or hypokinetic. Both disrupt prosody. However, the presence of dysarthria does not explain deficits in comprehension of affective prosody or comprehension of affective gestures. Thus, our three patients who were dysprosodic and not dysarthric imply prosodic disturbance can result from RHD that is not attributable to dysarthria. In the majority of cases, however, it is not clear whether prosodic disturbance in spontaneous speech and repetition of affective prosody results from dysarthria, a "pure" dysprosody, or both. Conclusions Our data provide answers to the questions we posed. First, RHD patients displayed a significant disruption in prosody in spontaneous speech, repetition of affective prosody, and comprehension of affective gestures. RHD patients did not differ significantly from normals in gestures accompanying spontaneous speech or comprehension of affective prosody. Second, dysprosody in RHD patients could be classified by Ross' (1981) taxonomy in 80 percent of the cases. Third, there was no systematic relationship between site of lesion and type of aprosodia in our RHD patients. And, fourth, while three of our RHD patients displayed dysprosody with no coexisting dysarthria, most RHD patients had

coexisting dysarthria. Thus, it is not clear what contributes to the perception of dysprosody in spontaneous speech and repetition of affective prosody in RHD patients who have a coexisting dysarthria. Additional tests of Ross' (1981) hypothesis about dysprosody subsequent to RHD are necessary. These may want to employ additional stimuli or trials to establish the range of normal performance. The number of trials and stimuli we employed resulted in considerable overlap between RHD and normal subjects on the repetition of affective prosody and comprehension of affective prosody tasks. Thus, the sensitivity and specificity of the measures we employed is unacceptable. In addition, Ross' (1981) taxonomy and its relationship to site of lesion requires additional exploration with larger samples. Moreover, there is a need to establish test-retest reliability on all tasks and inter- and intrajudge reliability for observation of gestures accompanying spontaneous speech, comprehension of affective prosody, and comprehension of affective gestures. Finally, there is a need to differentiate dysprosody resulting from dysarthria from dysprosody resulting from RHD. This may require perceptual and acoustic comparison of dysarthric subjects without RHD, subjects with RHD and no dysarthria, and subjects with RHD and coexisting dysarthria. References Baum, S. R., & Pell, M. D. (1997). Production of affective and linguistic prosody by brain-damaged patients. Aphasiology, 11, 177-198. Behrens, S. J. (1988). The role of the right hemisphere in the production of linguistic stress. Brain and Language, 33, 104-127. Blonder, L. X., Bowers, D., & Heilman, K. M. (1991). The role of the right hemisphere in emotional communication. Brain, 114, 1115-1127. Bradvik, B. B., Dravins, C, Hollas, S., Rosen, I., Ryding, E, & Ingvar, D. H. (1991). Disturbance of speech prosody following right hemisphere infarcts. Acta Neurologica Scandanavia, 84, 114-126. Bryan, K. L. (1989). Language prosody and the right hemisphere. Aphasiology, 3, 285-299. Cancelliere, A. E., & Kertesz, A. (1990). Lesion localization in acquired deficits of emotional expression and comprehension. Brain and Cognition, 13, 133-147. Cohen, J. (I960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 10,37-46. Darby, D. G. (1993). Sensory aprosodia: A clinical clue to lesions of the inferior division of the right middle cerebral artery? Neurology, 43, 567-572. Darley, F. L., Aronson, A. E., & Brown, J. R. (1975). Motor speech disorders. Philadelphia: W. B. Saunders. Duffy, J. R. (1995). Motor speech disorders: Substrates, differential diagnosis, and management. St. Louis: Mosby. Duffy, J. R., & Folger, W. N. (1996). Dysarthria associated with unilateral central nervous system lesions: A retrospective study. Journal of Medical Speech-Language Pathology, 4, 57-70. Emmory, K. D. (1987). The neurological substrates for prosodic aspects of speech. Brain and Language, 30, 305-320. Goodglass, H., & Kaplan, E. (1983). Boston diagnostic aphasia examination. Philadelphia: Lea and Febiger. Gorelick, P. B., & Ross, E. D. (1987). The aprosodias: Further functional-anatomical evidence for the organization of affective language in the right hemisphere. Journal of Neurology, Neurosurgery and Psychiatry, 50, 553-560. Hargrove, P. M., & McGarr, N. S. (1994). Prosody management of communication disorders. San Diego: Singular Publishing Group, Inc. Heilman, K. M., Bowers, D., Speedie, L., & Coslett, H. B. (1984). Comprehension of affective and nonaffective prosody. Neurology, 34, 917-921. Hird, K., & Kirsner, K. (1993). Dysprosody following neurogenic impairment. Brain and Language, 45,46-60. Kent, R. D., & Rosenbek, J. C. (1982). Prosodic disturbance and neurologic lesion. Brain and Language, 15, 259-291. Kertesz, A. (1982). Western aphasia battery. New York: Grune & Stratton. Lalande, S., Braun, C. M. J., Charlebois, N., & Whitaker, H. A. (1992). Effects of right and left cerebrovascular lesions on discrimination of prosodic and semantic aspects of affect in sentences. Brain and Language, 42, 165-186. Monrad-Krohn, G. (1947). The prosodic quality of speech and its disorders. Acta Psychologia Scandanavia, 22, 225-265. Ropper, A. H. (1987). Severe dysarthria with right hemisphere stroke. Neurology, 37, 1061-1063.

R.T. Wertz, C.R. Henschel, L.L. Auther, J.R. Ashford and H.S. Kirshner

Ross, E. D. (1981). The aprosodias: Functional-anatomic organization of the affective components of language in the right hemisphere. Archives of Neurology, 38, 561-569. Shapiro, B. E., & Danley, M. (1985). The role of the right hemisphere in the control of speech prosody in prepositional and affective contexts. Brain and Language, 25, 19-36. Tompkins, C. A. (1995). Right hemisphere communication disorders: Theory and management. San Diego: Singular Publishing Group, Inc. Tompkins, C. A., & Flowers, C. R. (1985). Perception of emotional intonation by brain-damaged adults: The influence of task processing levels. Journal of Speech and Hearing Research, 28,527-538. Van Lancker, D., & Sidtis, J. J. (1992). The identification of affective-prosodic stimuli by left- and righthemisphere damaged subjects: All errors are not created equal. Journal of Speech and Hearing Research, 35, 963-970. Weintraub, S., Mesulam, M. M., & Kramer, L. (1981). Disturbances in prosody: A right hemisphere contribution to language. Archives of Neurology, 38, 742-744. Wertz, R. T., & Rosenbek, J. C. (1992). Where the ear fits: Perceptual evaluation of motor speech disorders. Seminars in Speech and Language, 13, 39-54. Wolf, G. I., & Ross, E. D. (1987). Sensory aprosodia with left hemiparesis from subcortical infarction: Right hemisphere analogue of sensory-type aphasia with right hemiplegia? Archives of Neurology, 44, 668-671.

J. Neurolinguistics, Vol. 11, Nos 1-2, p. 103-118, 1998 © 1998 Published by Elsevier Science ience Ltd. All rights reserved Printed in Great Britain 0911-6044/98 $19.00 + 0.00

The role of emotion in the linguistic and pragmatic aspects of aphasic performance Marjorie Perlman Lorch*, Joan C. Borod* and Elissa Koff 'Birkbeck College, University of London, London, UK; Queens College and Mount Sinai Medical Center, City University of New York, New York, NY; *Wellesley College, Wellesley, MA, USA.

Abstract—Considerations of aphasics1 performance typically focus on aspects of linguistic impairment. Similarly, researchers tend to emphasize right braindamaged subjects' relatively poor performance in response to emotional content or context. The spared or heightened emotional abilities of aphasic communication often go unnoticed. Research will be reviewed which suggests that aphasics have the ability to successfully utilize emotion in the comprehension and expression of both linguistic and pragmatic content and contexts. Evidence from a wide range of research on lexical processing, prosody, and discourse will be reviewed which indicates that emotion may play a facilitatory role in the comprehension and production of communication in language-impaired people. A large group study involving 15 left brain-damaged, 12 right brain-damaged and 16 normal controls was carried out to investigate posed and spontaneous emotional expression and perception, including the vocal and verbal, as well as facial, channels for spontaneous expression. Results will be considered with respect to the neuropsychological organization of linguistic and emotional cognitive systems.

Introduction This paper is concerned with the communication of emotion in aphasic left hemisphere brain-damaged sufferers (LBDs) and right hemisphere brain-damaged sufferers (RBDs) compared with non-neurologically impaired normal controls (NC). The aim of this paper is to explore the relationship between linguistic and extralinguistic aspects of communicative behavior which comprise aphasic speech. Underlying this topic is the question of what constitutes pragmatically appropriate affective communication. General awareness of the social context and specific awareness of the interlocutors and their mental state/contents are key factors contributing to successful communication (Lesser & Milroy, 1993). Therefore, expression of emotion is a key component of communicative behavior. The communication of emotion is multidimensional, involving several distinct modes or channels simultaneously—face, language, speech and voice, gesture, and posture. Further, it has been suggested that different channels of emotional expression are subserved by distinct neural systems, based on evidence regarding selective impairment in braindamaged persons (Borod, 1993; Bowers, Bauer, & Heilman, 1993). Considerations of the performance of LBD aphasic speakers typically focus on aspects of their linguistic impairment. Similarly, researchers tend to emphasize RBD persons' relatively poor performance in response to emotional content or pragmatic context. In the modern aphasia literature, there has been little attention paid to the other communicative aspects of affective behavior which occur together with language—i.e., the dynamic interplay of facial expression, vocal quality and body movement. However, in order to

M. Perlman Lorch, J.C. Borod and E. Koff

achieve effective communication, in its broadest sense, the expression of all of these aspects of higher cerebral function must be successfully integrated. A historical perspective on considerations of emotional behavior in aphasic communication will be presented. Subsequently, a re-examination of data collected in an extensive group study on the effect of cortical brain damage on emotional expression will be carried out. These data were collected as part of an NIH-funded research project that tested hypotheses regarding the role of the right hemisphere in emotional cognition. RBDs were the experimental group, and LBDs were included, along with NCs, as a control group. In this study, focus on the performance of the LBDs on various tasks from this project will be provided to shed light on aphasic emotional communicative behavior.

The relation between propositional and emotional speech The consideration of emotional speech alongside propositional aspects of speech has had a variable history. At various periods in aphasiological history, clinicians have been relatively more inclusive about what constitutes communication, including speech and voice, while in other periods, focus has been exclusively on linguistic capacities. Broca was the first to observe that aphasics may retain the ability to express certain automatic and emotional aspects of speech. In 1861, he had noted how certain patients "have, in a certain way, two degrees of articulation. Under ordinary circumstances, they invariably pronounce their favorite word ('mot de prédilection'): but when they experience a fit of anger, they become capable of articulating a second word, most often a crude oath" (Broca, 1861, p. 333; citedin Harrington, 1987, p. 216). Hughlings Jackson (1868) developed the notion that emotional speech can be selectively spared in aphasics with left hemisphere lesions, drawing tile significant distinction between propositional and non-propositional speech. Aphasia was described as a loss of "higher" voluntary speech movements with sparing of "lower" involuntary, automatic speech, such as emotional exclamations: "The movements of speech are educated movements, and thus differ widely from those movements which may be said to be nearly perfect at birth, such as those for respiration, smiling, swallowing, etc. All the movements represented in the corpus striatum unilaterally require long education.... The muscles always acting bilaterally, and chiefly represented bilaterally in the corpus striata, are born with their centres for movements nearly perfect. Thus, then, the term "Intellectual Language" merges in the larger term "Special movements acquired by the Individual," and the term "Emotional Language" in the term "Inherited Movements" (Jackson, 1868, p. 275). The conservation of emotional experience in aphasics was attested to by Pierre Marie (1906). He went further than Jackson in emphasizing the intact affective and pragmatic nature of their social interactions: "I would like to recall the voluntary politeness of these aphasics as additional proof of the conservation in their case of the affective and moral sphere" (translatedin Lebrun, 1994, p. 235). Further discussion of the issues of emotional behavior in the context of aphasia can be found in the writings of Goldstein (1948). Goldstein extended the Jacksonian distinction from propositional and non-propositional speech to what he termed "abstract language"definedas "volitional, propositional and relational;" and "concrete language"—defined as including "the instrumentalities of speech" and emotional utterances. He noted that everyday language is a combination of both types of language and that if one of the two forms of language is particularly disturbed by pathology, the individual tries to overcome

this by an increased use of the other form. Goldstein's notions of the concrete and abstract attitude are crucially involved in the emotional expression of the aphasic: "Because of the difference of the relationship of emotional and nonemotional language to the personality, there is a difference of impairment of both in aphasia"(Goldstein, 1948, p. 25). Goldstein's notion of abstract attitude and the original Jacksonian distinction between propositional and nonpropositional speech, along with discussions of symbolic function developed by Head (1926), were reflected in the discussions of the left hemisphere in its role for language dominance in the second half of this century. This can be seen in a typical passage from Alajouanine and Lhermitte (1964): "The more the disorder affects the systems of symbolic formulation, the more the ability to deal verbally with ideas is impaired and the more prominent are dissociations between automatic and propositional language. In this latter state the ability to express and to understand appears not to depend on linguistic units per se but rather on psychological factors, among which...affective reactions...are noteworthy" (Alajouanine & Lhermitte, 1964, p. 216). Arguments supporting a more holistic and integrated approach to the study of communication were put forth by an increasingly smaller minority with the rise of more componential approaches (e.g., Geschwind (1965) in neurology and Chomsky (1965) in linguistics-cf. Sarles, 1977). At this time, with the rise of the localizationist approach (e.g., Geschwind, 1973) and the notion of modularity of function (Fodor, 1983), neurolinguistic investigators pursued characterizations of aphasia which correlated the functional architecture of the left hemisphere with components of grammar (e.g., Caplan & Hildebrandt, 1988). However, with the growing interest in parallel and distributed processing models and connectionism, there has been a resurgence of interest in a Systems Approach to cognitive processing (e.g., Damasio, 1989). Recently, the emphasis in aphasia research has broadened out from a focus on the clinical pathological correlation of linguistic functions to a more socially driven analysis of communicative interaction. This new interest in the consideration of pragmatic function, alongside linguistic function, is well represented by Lesser and Milroy (1993). Although there is a thorough discussion of gesture, body movement, and other non-verbal components of conversational communication in aphasia in this publication, no mention is made of emotional communication as part of the extralinguistic aspects of social interaction.

Neuropsychological research into emotional communication Pragmatic and affective impairments are typically attributed to right hemisphere pathology. Many investigators have pursued the hypothesis of right hemisphere specialization for emotional processing (Borod, Koff, & Caron, 1983; Bryden& Ley, 1983; Buck, 1984; Heilman, Bowers, & Valenstein, 1985; Ross, 1985). The role of the right hemisphere in the higher-order organization of arousal, attention, intention, and spatial processing has, at various times, been implicated as the source of difficulties with various tasks involving emotional cognition (e.g., Borod, 1992; Gardner,Brownell, Wapner, & Michelow, 1983). The more specific hypothesis—that different emotional valences (positive and negative) are distinctly lateralized with positive emotions being a left hemisphere specialization and negative emotions being a right hemisphere specialization, has also received some support (e.g., Davidson, 1984; Sackeim, Greenberg, Weiman, Gur, Hungerbuhler, & Geschwind,

M. Perlman Lorch, J.C. Borod and E. Koff

1982). Numerous studies have addressed these issues and tried to raise evidence in support of one of these two different formulations (see Borod, Caron, & Koff, 1981; Borod & Koff, 1984). In terms of the neuropsychological organization of emotion and hemispheric processing, specific findings have been reported: a) the left side of the face (controlled predominantly by the right hemisphere) has been found to be more intense than the right side during both posed and spontaneous expressions for neurologically intact subjects (Borod, Koff, & White 1983; Dopson, Beckwith, Tucker, & Bullard-Bates, 1984; for review, see Borod, Santschi-Haywood, & Koff, 1997); b) a right hemisphere dominance for emotional prosody and impairments of these functions in RBD patients (Ross & Mesulam, 1979); c) spontaneous emotional expression occurs less often in RBDs than LBDs or NC (Blonder, Burns, Bowers, Moore, & Heilman, 1993; Buck & Duffy, 1980). In addition to cortical right hemisphere control of emotional cognition, one must also include for consideration evidence of impairments in social or affectively appropriate behavior arising from a range of diffuse, bilateral and/or subcortical lesion sites. For example, pragmatically inappropriate language was found to be common in patients with diffuse or bilateral lesions in a detailed behavioral study of anosagnosic patients (Weinstein & Kahn, 1955). Referring to self in the second or third person, joking, slang or inappropriately informal language, euphemisms, cryptic speech, aphorisms, metaphors, or malapropisms reflecting pragmatic impairment were listed as typical features of these patients' speech. This pattern of language behavior (part of the larger syndrome of denial of illness) was only seen in lesions arising from a cerebral vascular accident if it involved deep subcortical structures or increase in cerebrospinal fluid pressure (which would give rise to diffuse and bilateral effects). Pragmatically, socially or affectively inappropriate language is generally an indication of pathology which extends beyond the language cortex of the dominant hemisphere (Lorch, 1995). Given the current support for the view that there is a special role for the right hemisphere in the cognitive control of the expression and perception of emotion, this raises the question: What are the consequences for extralinguistic communication in aphasic persons with left brain damage and intact right hemisphere systems? When discussing the psychiatric aspects of aphasia, Benson (1973) states that appropriate expressions of frustration, anger and depression are quite common in aphasic sufferers. The notion that left hemisphere patients become depressed has been long-standing in the literature, first developed in depth by Goldstein (1948). Research into the lateral differences in hemispheric valence of emotional expression, mentioned above, has in part been pursued in light of the incidence of depression following left but not right hemisphere stroke (Downhill & Robinson, 1994; Robinson & Szetela, 1981). A more complex picture has emerged in a longitudinal study of the emotional and social reactions of stroke sufferers. Nelson, Cicchetti, Satz, Sowa and Mitrushina (1994) found that differential recovery rates were dependent on side of lesion, with left hemisphere stroke sufferers having increased indifference, inappropriateness and depression as measured at 2 months post-onset. By 6 months post-onset, however, their emotional recovery seemed to stabilize, whereas the emotional functioning of right hemisphere stroke sufferers appeared to worsen at this point. Paralinguistic features of voice Determination of the significant behavioral features which comprise meaningful and appropriate displays of affective content in social communication is still limited. One aspect of

communication which has received considerable attention is intonation. The communicative function of intonation is both emotional and linguistic, being used to convey emotional tone, linguistic (lexical) stress, contrastive stress and syntactic structure. The acoustic correlates of the emotional aspects of intonation in conversational speech were first measured by Lieberman(1967). A number of distinct vocal signal changes were identified with the expression of emotion: raised or lowered average fundamental frequency (F0) when angry; extreme emotion resulted in wider F0; lowered and narrowed F0 range; and breaking sentence into many breath-groups or extended breath-groups. Other paralinguistic vocal devices identified include: filled and unfilled pauses; sound pressure level (loudness) raised or lowered; and diminished pitch perturbations of speakers' F0. Examination of F0 and sentence level pitch perturbations has also been carried out in aphasic speakers (Scherer, 1985). Intonational aspects of aphasic speech have been classified as distinct parts of the major aphasic syndromes: Broca's aphasic speech is characterizedas being dysfluent, and/or intonationally flat, and Wemicke's aphasic speech is characterized as being generally fluent or even hyperprosodic (Goodglass & Kaplan, 1972; 1983). In studies by Danly and colleagues (Danly, Cooper, & Shapiro, 1983; Danly & Shapiro, 1982), they demonstrated that the impairment of intonation in aphasics, as measured by F0 contours, was linked in part to the specific syntactic function of the utterance. The speech of Broca's aphasics gives the impression of being dysprosodic due to the slow rate and numerous interruptions in the production of utterances. However, the aphasics were found to exhibit natural terminal falling F0 and F0 declination in short, simple sentences, and were impaired relative to the syntactic complexity of longer sentences (Danly & Shapiro, 1982). Moreover, the Broca's aphasic speakers appeared to compensate for their halting speech by typically ending phrases with a sizable continuation rise in F0 This aids the listener in demarcating syntactic processing units. This continuation rise is often used as a normal conversational discourse device. Reviewing the evidence In the 1980's, there was a growing shift of interest from the well-explored linguistic functions of the left hemisphere to the less well-defined functions of the right hemisphere (e.g., Brownell, Michelow, Powelson, & Gardner, 1983; Zaidel, 1976). At that time, Borod, Koff, [Perlman] Lorch and Nicholas commenced a large-scale project investigating the neurological organization of emotional expression. In this project, groups of right brain-damaged patients (RBDs) and left brain-damaged patients (LBDs) with focal cortical lesions, and neurologically healthy control subjects (NCs) were tested on an extensive range of experimental tasks which focused on the various components which comprise human emotional cognition. The initial hypotheses were constructed to test the assumption that right brain damage would lead to impairment on tasks of emotional expression and perception. The results were thus analyzed for differences between the RBDs as compared with the LBDs and NCs (Bloom, Borod, Obler, & Koff, 1990; Borod, Koff, Lorch, & Nicholas, 1985; Borod, Koff, Lorch, & Nicholas, 1986a & b; Borod, Koff, Lorch, Nicholas, & Welkowitz, 1988; Borod, Lorch, Koff, & Nicholas, 1987; Kent, Borod, Koff, Welkowitz, & Alpert, 1988). In this section, the data from that project will be re-evaluated specifically with regard to the LBD patients' performance. This review is focused on two issues: 1) What intact abilities of emotional expression are evident in the LBDs' performance? and 2) Does their performance differ in any way from NCs? This re-examination may indicate some features

M. Perlman Lorch, J.C. Borod and E. Koff

of aphasic performance which draw on spared extra-linguistic abilities to compensate in communication. Experiment I Subjects All subjects were drawn from the Neurology Service of the Boston Veterans Administration Medical Center. Subjects were 12 males with unilateral right hemisphere lesions (RBDs) and 15 males with unilateral left hemisphere lesions (LBDs), all resulting from a cerebral vascular accident (CVA), and having an absence of history of psychiatric disorder, psychotropic drug treatment or secondary neurological disorder. Sixteen males without neurological impairment served as controls (NC). All subjects were right-handed by self (or family) report and did not differ on formal assessment of lateral dominance (Coren, Porac, & Duncan, 1979). The subjects did not differ significantly on the demographic variables of age, education and occupational status. Subjects had a mean age of 58 years (S.D. =8.1 yrs), a mean of 13 years of education (S.D. = 2.8), and a mean occupation level (Hollingshead & Redlich, 1958) of 3.9 (S.D. = 1.5), generally classified as "middle-class white collar worker." Table 1 displays subject details of age, months post onset, lesion site, aphasia type and scores on the Weschler Adult Intelligence Scale (WAIS) Performance IQ (Weschler, 1958). ID

Side of lesion Right Right Right Right Right Right Right Right Right Right Right Right

Lesion site P,S P,S P,T P,T,S P, O, S P F.T.S F,P,T,S F,P,T,S F,P,T,S F,P,T,S F,P,O, S

Side of lesion Left Left Left Left Left Left Left Left Left Left Left Left Left Left Left

ID = patient identification number MPO = months post-onset of illness WAIS PIQ = Weschler Adult Intelligence Scale Performance IQ F = Frontal P= Parietal T = Temporal O = Occipital S = Subcortical

Table 1. Characteristics of patients

Lesion site F F,S F,S F, S P,T P,T P,T, S P,T, S P.T.S P, T, O F,P F, P,T F,T, S F, P., T, S F, P, T, S

Method Spontaneous production of emotion was assessed during the viewing of slides designed to elicit expressions of positive and negative emotions (Buck, 1978). There were 16 slides depicting a variety of emotionally laden scenes—e.g., a baby picking flowers, a beautiful sunset, a couple embracing, a surgical procedure, a young victim of starvation, and a photographic double exposure. Subjects were seated facing a one-way mirror, behind which a video-camera recorded their responses. Each slide was presented for 20 seconds; after 6 seconds, the subject was requested to describe his feelings and reactions to the slide. Analysis 1—Spontaneous emotional facial expressions Two judges viewed the videotapes and rated each response for Responsivity—i.e., did a facial response occur, (0 = no, 1 = yes), Appropriateness of the facial expression response (0 = inappropriate, 1 = appropriate) and Intensity of the response (7-point scale, 1 = minimal to 7 = maximal intensity). Mean rating scores were used because inter-rater reliability for a sub-sample of scores was high (Pearson's r = +.74). The mean scores for emotional facial responses to the slides viewed were analyzed by group for each of the 3 ratings described above. The results are displayed in Table 2. (The results are averaged across spontaneous and posed facial expressions.) Measure

Means and standard deviations NC RED LBD 0.72 (0.14) 0.59(0.19) 0.74 (0.14) 0.93 (0.06) 0.87 (0.09) 0.83 (0.14) 3.19 (0.56) 3.38 (0.60) 3.62 (0.65)

Table 2. Mean scores for emotional facial responses The LBDs were slightly more responsive than NCs, and both groups were more responsive than the RBDs. For the measure of Intensity of responses to the slides, the LBDs produced more intense responses than either the NCs and the RBDs. To examine the relationship among the different response variables, the Spearman rank-order procedure was used to correlate scores for Responsivity (Resp.), Appropriateness (Approp.) and Intensity (Intens.). As can be seen in Table 3, the only significant correlation was between Responsivity and Intensity for the LBD group. This can be interpreted as indicating that for the LBD group, when an emotional response occurred, there was a significantly greater chance that it would be more intensely expressed than for the RBD or NC group. Group RBD LBD NC

Resp. vs. approp. 0.02 0.14 -0.44

Resp. vs. intens. 0.67 0.84* 0.31

Intens. vs. approp. 0.37 -0.14 -0.02

* p<0.001 Table 3.

Spearman rank-order correlation rhos among parameters of facial emotional expression by subject group

M. Perlman Lorch, J.C. Borod and E. Koff

Analysis 2—Channels of communication The contribution of the three channels of communication-face, intonation and verbal production-to each subject's response in viewing the 16 slides was rated for "utilization or degree of involvement" on a 4-point scale (0 = noncontributory, 1 = minimal, 2 = moderate, 3 = maximal involvement).

Figure 1. Communication of emotion as a function of group and channel. LBD indicates left brain-damaged patients; RBD, right brain-damaged patients; and NC, normal controls. LBDs' use of facial and intonation channels was higher than it was for NCs and significantly higher than for RBDs; see Figure 1. When analyzed for frontal (F) versus nonfrontal (NF) (i.e., posterior) lesions, FLBDs, NFLBDs and NFRBDs used the facial expression channel significantly more frequently than FRBDs. Clinicians have long observed that extralinguistic channels of communication can supplement or substitute for impaired linguistic expression in spontaneous production. It is not surprising that those with left frontal damage used the speech channel less frequently than the other groups, given their nonfluent aphasia, but it is notable that those with left hemisphere lesions used the intonation channel more than those with right hemisphere lesions regardless of intrahemispheric location of lesion anteriorly or posteriorly.

Analysis 3—Propositional and nonpropositional contributions to speech communication The relative contribution of propositional and nonpropositional speech in the subjects' verbal responses to the emotionally ladened slides was also analyzed. A rating of 1 was given to verbal responses which were predominantly propositional with little nonpropositional contribution. A rating of 2 was given to verbal responses with a balance of the two, as seen in modal speech. A rating of 3 was given to verbal responses which were deemed to be relatively more nonpropositional in its composition, with reliance on intonation contours. The LBDs (x = 2.14) relied more on nonpropositional speech than the NCs (x = 1.92), whereas the RBDs (x = 1.64) relied predominantly on speech without a nonpropositional contribution.

Analysis 4—Verbalization of emotion The linguistic content of the verbal responses to the emotionally-ladened slides was also analyzed. It became apparent that although the participants were all asked specifically to discuss their feelings about the slide being viewed, many, in fact, offered descriptions of the pictorial aspects of the scene represented instead. Two ratings of linguistic (semantic) content were made on a 4-point scale (0 = not present, 1 = minimally present, 2 = moderately present, 3 = maximally present) for the emotional content (Feelings) and (inappropriate) descriptions of pictorial content (Form). When these responses were analyzed with regard to inter- and intra-hemispheric lesion site groupings, it was found that the aphasic subjects with left nonfrontal/posterior lesions-4 Wemicke's and 2 Conductions-(and the subjects with right frontal lesions) produced the greatest number of these inappropriate pictorial rather than emotional verbal responses.

Figure 2. Use of "form" in describing feelings as a function of lesion side and location. LBD indicates left brain-damaged patients and RBD, right brain-damaged patients.

Experiment II The finding of greater intensity in facial emotional expression reported above prompted the analysis of the intensity of emotion expressed in the verbal responses. Bloom, Borod, Obler and Koff (1990) carriedout a detailed analysis of the verbal responses in a subset of the subjects from the original study. Three LBDs, 3 RBDs and 3 NCs were randomly selected from the 43 subjects used in the original study. Responses to 4 of the original 16 slides viewed were analyzed—2 slides designed to elicit positive emotions and 2 slides designed to elicit negative emotions. The 36 transcripts of verbal reactions to the slides (4 responses x 9 subjects) were submitted to discourse analysis. A filtering procedure was applied to control for the artifactual differences caused by agrammatism present in the LBDs. All free grammatical morphemes (function words) which tend to be omitted in agrammatic aphasics were removed from all of the subjects' transcripts, leaving only substantives (content words) for analysis.

M. Perlman Lorch, J.C. Borod and E. Koff

Analysis 1—Category accuracy and valence accuracy Five naive raters judged how accurately subjects used words to express emotions in the written transcripts of their responses to viewing the slides. Responses were rated for emotional category accuracy and valence accuracy. Table 4 displays the percentage correct means for category accuracy and valence accuracy by subject group. As can be seen, the LBDs were less accurate that the NCs in communicating in their discourse the category of emotion but were slightly better, particularly at linguistically expressing negative emotion, than the NCs when only positive or negative valence expression was measured.

Table 4. Category accuracy and valence accuracy by subject group In order to account for the successful communication of emotion in aphasic discourse token/type ratios were calculated. As expected, LBD token/type ratio was higher (2.0) than for the RBD (1.3) and NC (1.5) groups. This reflects an effective pragmatic device in using repetition for emphasis and in expressing intensity for the LBDs. Analysis 2—Intensity Four raters characterized the words in the 36 discourse transcripts for intensity of emotion expressed, using a 5-point Likert scale (1 = least, 5 = most). Slide Type Positive Negative

Table 5. Mean ratings for intensity by subject group The LBDs were found to be more intense in linguistic communication of emotion than NCs (and RBDs) despite their aphasia.

Experiment III In a follow-up study with new subjects, Bloom, Borod, Obler and Gerstman (1993) elicited three types of narratives from pictures: 1) emotional content-a girl whose dog is hit by a

car, 2) visual spatial content-moving a box by climbing on books piled on a chair, and 3) procedural/neutral content-how to fry an egg, from a new group of LBD (n = 12), RBD (n = 9), and NC (n = 12) right-handedadults. The stories produced by the subjects to these 3 pictures were rated by 2 trained judges as "appropriate" or "inappropriate" on 7 pragmatic features adapted from Grice (1975) and Prutting and Kirchner (1987): topic maintenance, conciseness, specificity (i.e., lack of ambiguity), lexical selection, revision and repair, relevancy and quantity of information. When the narratives were assessed for pragmatic features, the LBDs were found to be significantly impaired relative to the NCs for specificity, revision and repair, lexical selection, and quantity of information but not significantly different from the NCs on conciseness, topic maintenance, and relevancy. While their performance was (as expected) significantly more impaired than the NCs overall with respect to the mean number of pragmatic features (n = 7) rated as appropriate, the LBDs were found to be significantly less impaired on the emotional narrative (x = 5.42) than the visual spatial (x = 4.08) (p < .05) and procedural (x = 4.17) (p < .10) narratives. This held true when the data were examined by individual subject rather than pooled group mean scores. The number of individuals in each group that performed better than or equally in the emotional (E) compared to the nonemotional (NE) condition is displayed in Table 6.

Table 6. Number of individuals in each subject group by performance condition This improved performance was true for the emotional narrative production in both the LBDs with anterior and those with posterior lesions. Both the anterior and the posterior subgroup of LBDs performed better on the emotional versus nonemotional conditions for mean pragmatic features produced appropriately (anterior LBD emotional and nonemotional mean scores, 6.34 vs. 5.33; posterior LBD emotional and nonemotional mean scores, 5.00 vs. 3.69). Thus, there was a facilitating effect of emotional context on pragmatic performance for the LBDs (with the reverse found for the RBDs). Emotionally ladenedmaterial has been found to provide facilitatory contexts for aphasic persons in a variety of different modalities and tasks: auditory comprehension (Boiler, Bole, Vrtunski, Patterson, & Kim, 1979; Reuterskiold, 1991); oral reading and writing (Landis, Graves, & Goodglass, 1982); and repetition (Ramsberger, 1996). Experiment IV Facilitation of facial movement by emotional context Jackson (1878) and Geschwind (1975) stress the role of context for the apraxic patient, who may not be able to carry out a specific movement to verbal command or even to imitation, but will perform spontaneously in a more naturalistic setting. Apraxia is considered a disorder of voluntary movement which is under left hemisphere control (Kimura, 1982). The question addressed in this study was whether emotional cueing would improve the performance of apraxic patients by providing a more "automatic" or affective

114

M. Perlman Lorch, J.C. Borod and E. Koff

context. Apraxic patients can carry out spontaneous emotional movements with the same sets of muscles used in praxic skills (Nathan, 1947). "The apraxic...is able to smile, lick his lips, express disgust in a normal manner when he feels thai the occasion demands; but he is unable to act these gestures...when he thinks the occasion demands...He remains well adapted to his environment...he is able to play cards, laugh at a joke, eat and drink in the manner customary in his society" (Nathan, 1947, p. 475). Since the right hemisphere is implicated in the expression of emotion, the possibility that emotional cues could facilitate posed praxic responses in patients with LBD was hypothesized. Such a finding would be interpreted as an instance of compensation for a primary behavioral deficit of the dominant hemisphere by a secondary system. The original set of patients with LBD and RBD, detailed in Table 1, and NCs were videotaped while being examined on tasks of bucco-facial praxis in both non-emotional and emotional contexts. Six movements were selected for having a neutral (non-emotional) form and an emotional analog. Three movements involved the upper face, and three involved the lower face. This was done in order to control for the effect of (central) facial paralysis of the lower face. The six commands used in the facial movement elicitation task were: Non-emotional/neutral 1) close one eye 2) lift your eyebrows 3) lower your eyebrows 4) put out your tongue 5) raise the corners of your mouth 6) pucker your lips

Emotional context analog 1) close one eye like a wink 2) lift your eyebrows like you're surprised 3) lower your eyebrows like a frown 4) stick out your tongue like you're making a face at me 5) raise the corners of your mouth like a smile 6) pucker your lips like a kiss

All of the neutral items were presented first; emotionally cued items followed in a second block. This allowed for the establishment of a baseline measure of apraxic performance for each subject to which the degree of improvement due to emotional facilitation could be compared. It was considered unlikely that any improvement in the Emotional condition, presented second, could be attributed to a rehearsal effect. Apraxic patients do not tend to improve with repeated trials (Mateer, 1978). Two judges rated the execution and accuracy of each videotaped movement with respect to quality. As would be expected, the LBDs demonstrated significant impairment in buccofacial praxis as indicated by a poor performance on the neutral tasks relative to the RBDs and NCs. However, the LBDs' motor performance was more strongly facilitated by provision of the additional emotional cue (which was more pronounced in those movements involving the lower face) than the other two groups (F(2,40) = 2.98; p - .06) as displayed in Table 7 below.

NC LBD RBD

NE 2.80 2.19 2.70

E 2.90 2.58 2.91

Table 7. Execution mean ratings for each group by condition

The role of emotion in aphasic performance

115

The analogy can be drawn between the facilitation effect of emotional context achieved in this study for LBDs with bucco-facial apraxia and therapeutic techniques which draw on other aspects of cognitive behavior attributed to the right hemisphere used to facilitate speech production in aphasic persons, such as Melodic Intonation Therapy (Sparks, Helm, & Albert, 1974),Visual Communication Therapy (Gardner, Zurif, Berry, & Baker, 1976), and Visual Action Therapy (Helm-Estabrooks, Fitzpatrick, & Barresi, 1982). Experiment V Voluntary emotional expression Posed emotional facial expressions were elicited from both verbal and visual input for 8 emotions (3 positive and 5 negative) with the LBD, RBD, and NC subjects described in Table 1. Performances were rated for intensity of expression, category accuracy and valence accuracy. On ratings of intensity (1 = minimal intensity to 7 = maximal intensity), LBDs (x = 4.14) posed positive emotions with significantly more intensity than either NCs (x = 3.69) or RBDs (x = 3.54). Post-hoc tests to examine valence differences within each subject group demonstrated that positive emotions were posed significantly more intensely than negative emotions by both the LBDs (xs = 4.14 vs. 3.63) and NCs (xs = 3.69 vs. 3.35) while there were no differences for the RBDs. Discussion In summary, the 5 experiments investigating emotional, linguistic and pragmatic aspects of communication reviewed here demonstrate the wide range of emotional communicative behaviors which are available to aphasic LBDs in speech and facial expression. This largescale project included a number of other investigations of posed and spontaneous production of emotional facial expressions and the perception of emotional facial expressions reported in detail in Borod et al. (1986a). LBDs were found to be as accurate as NCs for both posed and spontaneous facial expression and face perception, receiving a similar overall mean accuracy score-LBDs (x - 0.69) and NCs (x = 0.72), as compared with the impaired performance of the RBDs (x = 0.51), with no significant differences between conditions or valences. Evidence has been reported that both fluent and non-fluent aphasics produce more gestures than controls in a free conversational setting (Blonder, Burns, Bowers, Moore, & Heilman, 1995; Feyereisen & Seron, 1982). Feyereisen and Seron (1982) suggest that this indicates that "...social competence could be intact in aphasia." The data presented here from the project by Borod, Koff, Lorch and Nicholas are consistent with that conclusion with respect to emotional communication. Our aphasic subjects were found to be equal to or more highly rated than NCs and RBDs on measures of emotional communication: 1) responsivity and intensity of spontaneous emotional facial expression; 2) the use of both face and intonational aspects of voice as channels of expression; 3) the reliance on nonpropositional speech; 4) expressing the valence (positive/negative) of their emotions verbally; 5) intensity in the linguistic communication of their emotions; 6) the use of discourse in emotional contexts; 7) facilitation of emotional context on pragmatic aspects of discourse; 8) facilitation of emotional context on the quality of bucco-facial praxic movements; and 9) facilitation of emotional context on the intensity of posed facial expressions.

116

M. Perlman Lorch, J.C. Borod and E. Koff

For Smith and Kemp-Wheeler (1996), the question of why we need emotions is consideredone of the "unsolved mysteries of the mind." In their discussion, they examine the way emotion may change the way we think. The present review has attempted to explore the way emotion may change the way we communicate, and how aphasic speakers may express emotion more fully than other aspects of language, with speech and voice thus contributing to their communicative success. Acknowledgments—The first author wishes to acknowledge the invaluable input of Dr. Renata Whurr. This work was supported by NIH Grant No. NS06209 to the Aphasia Research Center of Boston University Medical Center, by NIH Biomedical Research Support Grant No. 1-S07RR07186 to Wellesley College, by NIH Grant Nos. MH42172 and MH44889 to Queens College of the City University of New York, and by the Wellcome Trust. Figures 1 and 2 were reproduced by permission of the American Medical Association.

References Alajouanine, T., & Lhermitte, F. (1964). Aphasia and physiology of speech. In D. Rioch and E. Weinstein (Eds.), Disorders of Communication. Baltimore: Williams and Wilkins Co. Benson, D. F. (1973). Psychiatric aspects of aphasia. British Journal of Psychiatry, 123, 555-566. Blonder, L. X., Burns, A. F., Bowers, D., Moore, R. W., & Heilman, K. M. (1993). Right hemisphere facial expressivity during natural conversation. Brain and Cognition, 21,44-56. Blonder, L. X., Burns, A. F., Bowers, D., Moore, R W., & Heilman, K. M. (1995). Spontaneous gestures following right hemisphere infarct. Neuropsychologia, 33, 203-213. Bloom, R. L., Borod, J. C., Obler, L. K., & Koff, E. (1990). A preliminary characterization of lexical emotional expression in right and left brain-damaged patients. International Journal of Neuroscience, 55,71-80. Bloom, R. L., Borod, J. C., Obler, L. K., & Gerstman, L. (1993). Suppression and facilitation of pragmatic performance: Effects of emotional content on discourse following right and left brain damage. Journal of Speech and Hearing Research, 36, 1227-1235. Boiler, F., Cole, M., Vrtunski, B., Patterson, M., & Kim, Y. (1979). Paralinguistic aspects of auditory comprehension in aphasia. Brain and Language, 7, 164-174. Borod, J. (1992). Interhemispheric and intrahemispheric control of emotion: A focus on unilateral brain damage. Journal of Consulting and Clinical Psychology, 60, 339-348 Borod, J. (1993). Emotion and the brain—Anatomy and theory: An introduction to the special section. Neuropsychology, 7,427-432. Borod, J. C., Caron, H., & Koff, E. (1981). Asymmetries in positive and negative facial expressions. Neuropsychologia, 19, 819-824. Borod, J. C, & Koff, E. (1984). Asymmetries of affective facial expression. In N. Fox and R. Davidson (Eds.), The psychobiology of affective development. Hillsdale, NJ: Lawrence Erlbaum Associates. Borod, J. C., Koff, E., & Caron, H. (1983). Right hemisphere specialization for the expression and appreciation of emotion: A focus on the face. In E. Perecman (Ed.), Cognitive processing in the right hemisphere, (pp. 83-110). New York: Academic Press. Borod, J. C., Koff, E., Lorch, M. P., & Nicholas, M. (1985). Channels of emotional expression in patients with unilateral brain damage. Archives of Neurology, 42,342-348. Borod, J. C., Koff, E., Lorch, M. P., & Nicholas, M. (1986a). The expression and perception of facial emotion in brain-damaged patients. Neuropsychologia, 24, 169-180. Borod, J. C., Koff, E., Lorch, M. P., & Nicholas, M. (1986b). Deficits in facial expression and movement as a function of brain damage. In J.-L. Nespoulous, P. Perron, and A. R. Lecours (Eds.), The biological foundations of gestures. Hillsdale, NJ: Lawrence Erlbaum Associates. Borod, J. C, Lorch, M. P., Koff, E., & Nicholas, M. (1987). Effect of emotional context on bucco-facial apraxia. Journal of 'Clinical and Experimental Neuropsychology, 9, 147-153. Borod, J. C., Lorch, M. P., Koff, E., Nicholas, M., & Welkowitz, J. (1988). Emotional and non-emotional facial behaviors in patients with unilateral brain damage. Journal of Neurology, Neurosurgery and Psychiatry, 51, 826-832. Borod, J. C., Koff, E., & White, B. (1983). Facial asymmetry in spontaneous and posed expressions of emotion. Brain and Cognition, 2, 165-175. Borod, J. C., Santschi-Haywood, C., & Koff, E. (1997). Neuropsychological aspects of facial asymmetry during emotional expression: A review of the normal adult literature. Neuropsychological Review. 7, 41 59. Bowers, D., Bauer, R., & Heilman, K. (1993). The nonverbal affect lexicon: Theoretical perspectives from neuropsychological studies of affect perception. Neuropsychology, 7,433-444. Brownell, H., Michelow, D., Powelson, J., & Gardner, H. (1983). Surprise but not coherence: Sensitivity to verbal humor in right-hemisphere patients. Brain and Language, 18, 20-27.

The role of emotion in aphasic performance

117

Bryden, M. P., & Ley, R G. (1983). Right-hemispheric involvement in the perception and expression of emotion in normal humans. In K. M. Heilman and P. Satz (Eds.), Neuropsychology of Human Emotions (pp. 6-44). New York: Guilford Press. Buck, R. (1978). The slide-viewing technique for measuring non-verbal sending accuracy: A guide for replication. Catalog of Selected Documents in Psychology, 8, 63. Buck, R (1984). The Communication of Emotion. New York: Guilford Press. Buck, R, & Duffy, R. J. (1980). Nonverbal communication of affect in brain damaged subjects. Cortex, 16, 351-362. Caplan, D., & Hildebrandt, N. (1988). Disorders of syntactic comprehension. Cambridge, MA: MIT Press. Chomsky, N. (1965). Aspects of the Theory of Syntax. Cambridge, MA: MIT Press. Coren, S., Porac, C, & Duncan, P. (1979). A behaviorally validated self-report inventory to assess four types of lateral preferences. Journal of Clinical Neuropsychology, 1, 55-64. Damasio, A. R. (1989). Time-locked multiregional retroactivation: A systems level proposal for the neural substrates of recall and recognition. Cognition, 33, 25-62. Danly, M., & Shapiro, B. (1982). Speech prosody in Broca's aphasia. Brain and Language, 16, 171-190. Danly, M., Cooper, W., & Shapiro, B. (1983). Fundamental frequency, language processing and linguistic structure in Wernicke's aphasia. Brain and Language, 19, 1-24. Davidson, R. J. (1984). Affect, cognition, and hemispheric specialization. In C. E. Izard, J. Kagan, and R. Zajonc (Eds.), Emotions, Cognition and Behavior. Cambridge: Cambridge University Press. Downhill, J. E., & Robinson, R G. (1994) Longitudinal assessment of depression and cognitive impairment following stroke. Journal of Nervous and Mental Diseases, 182, 425-431. Dopson, W., Beckwith, D., Tucker, D. M., & Bullard-Bates, P. (1984). Asymmetry of facial expression in spontaneous emotion. Cortex, 20, 243-252. Feyereisen, P., & Seron, X. (1982). Nonverbal communication and aphasia: A review. Brain and Language, 16, 191-212, 213-236. Fodor, J. (1983). The modularity of mind. Cambridge, MA: MIT Press. Gardner, H., Brownell, H., Wapner, W., & Michelow, D. (1983). Missing the point: The role of the right hemisphere in the processing of complex linguistic materials. In E. Perecman (Ed.), Cognitive processing in the right hemisphere (pp. 169-191). New York: Academic Press. Gardner, H., Zurif, E., Berry, T., & Baker, E. (1976). Visual communication in aphasia. Neuropsychologia, 14, 275-292. Geschwind, N. (1965). Disconnexion syndromes in animals and man. Brain, 88, 237-294, 585-644. Geschwind, N. (1975). The apraxias: neural mechanisms of disorders of learned movement. American Scientist, 63, 188-195. Geschwind, N. (1973). The brain and language. In G. A. Miller (Ed.), Communication, language and meaning (pp. 61-72). New York: Basic Books. Goldstein, K. (1948). Language and language disturbances. New York: Grune and Stratton. Goodglass, H., & Kaplan, E. (1972). The assessment of aphasia and related disorders. Philadelphia: Lea and Febiger. Grice, H. (1975). Logic and Conversation. In P. Cole and J. Morgan (Eds.), Studies in syntax and semantics: Speech acts. New York: Academic Press. Harrington, A. (1987). Medicine, mind and the double brain. Princeton, NJ: Princeton University Press. Head, H. 1926. Aphasia and kindred disorders of speech. Cambridge: Cambridge University Press. Heilman, K. M., Bowers, D., & Valenstein, E. (1985). Emotional disorders associated with neurological diseases. In K. M. Heilman and E. Valenstein (Eds.), Clinical neuropsychology (pp. 377-402). Oxford: Oxford University Press. Helm-Estabrooks, N., Fitzpatrick, P., & Barresi, B. (1982). Visual action therapy for global aphasics. Journal of Speech and Hearing Disorders, 44,385-389. Hollingshead, A. B., & Redlich, F. C. (1958). Social Class and Mental Illness. New York: Wiley. Jackson, J. H . (1868). On the physiology of language. The Medical Times and Gazette, 2, 275-276. Jackson, J. H. (1878). In J. Taylor (Ed.) (1932), Selected writings of Jackson. London: Hodder and Stoughton. Kent, J., Borod, J. C., Koff, E., Welkowitz, J., & Alpert, M. (1988). Posed facial emotional expression in braindamaged patients. International Journal of Neuroscience, 43, 81-87. Kimura, D. (1982). Left-hemisphere control of oral and brachial movements and their relation to communication. Philosophical Transactions of the Royal Society, London, (B) 298, 135-149. Landis, T., Graves, R, & Goodglass, H. (1982). Aphasic reading and writing: Possible evidence for right hemisphere participation. Cortex, 18, 105-112. Lebrun, Y. (1994) Selections from the work of Pierre Marie. In P. Eling (Ed.), Reader in the history of aphasia. Amsterdam: John Benjamins. Lesser, R, & Milroy, L. (1993). Linguistics and Aphasia: Psycholinguistic and pragmatic aspects of intervention. Harlow: Longmans. Lieberman, P. (1967). Intonation, perception and language. Cambridge, MA: MIT Press. Lorch, M. P. (1995). Some neurolinguistic evidence regarding variation in interlanguage use. In L. Eubank, L. Selinker, and M. Sharwood Smith (Eds.), The current state of interlanguage: Studies in honor of William E. Rutherford. Amsterdam: John Benjamins. Marie, P. (1906) La troisième circonvolution frontale gauche ne joue aucun rôle special dans la fonction du langage. Semaine Médicale 26, 241-247. [Translated in Lebrun, 1994] Mateer, C. (1978). Impairments of nonverbal oral movements after left hemisphere damage: A followup analysis of errors. Brain and Language, 6,334-341.

118

M. Perlman Lorch, J.C. Borod and E. Koff

Nathan, P. W. (1947). Facial apraxia and apraxic dysarthria. Brain, 70,449-478. Nelson, L. D., Cicchetti, D., Satz, P., Sowa, M., & Mitrushina, M. Emotional sequelae of stroke: A longitudinal perspective. Journal of Clinical and Experimental Neuropsychology, 16,796-806. Prutting, C, & Kirchner, D. (1987). A clinical appraisal of the pragmatic aspects of language. Journal of Speech and Hearing Disorders, 52, 105-119. Ramsberger, G. (1996). Repetition of emotional and nonemotional words in aphasia. Journal of Medical Speech-Language Pathology, 4, 1-12. Reuterskiold, C. (1991). The effects of emotionality on auditory comprehension in aphasia. Cortex, 27, 595604. Robinson, R. G., & Szetela, B. (1981). Mood changes following left hemispheric brain injury. Annals of Neurology, 9, 447-453. Ross, E. (1985). The modulation of affect and nonverbal communication by the right hemisphere. In M.-M. Mesulam (Ed.), Principles of Behavioral Neurology (pp. 239-257). F. A. Davis: Philadelphia. Ross, E., & Mesulam, M.-M. (1979). Dominant language functions of the right hemisphere? Prosody and emotional gesturing. Archives of Neurology, 36, 144-148. Sackeim, H. A., Greenberg, M. S, Weiman, A. L., Gur, R. C., Hungerbuhler, J. P., & Geschwind, N. (1982). Hemispheric asymmetry in the expression of positive and negative emotions: Neurological evidence. Archives of Neurology, 39, 210-218. Sarles, H. B. (1977). Language and human nature. Minneapolis: Minnesota Press. Scherer, K. (1985). Methods of research on vocal communication: Paradigms and parameters. In K. Scherer and P. Ekman (Eds.), Handbook of methods in nonverbal behavior research (pp. 136-198). New York: Cambridge University Press. Smith, P. T., & Kemp-Wheeler, S. M. (1996). Why do we need emotions? In V. Bruce (Ed.), Unsolved mysteries of the mind. Hove: Taylor and Francis. Sparks, R., Helms, N., & Albert, M. (1974). Aphasia rehabilitation resulting from melodic intonation therapy. Cortex, 10,303-306. Weinstein, E. A., & Kahn, R. L. (1955). Denial of illness: Symbolic and physiological aspects. Springfield, IL: Charles C. Thomas. Weschler, D. (1958). The Measurement and appraisal of adult intelligence. Baltimore: Williams and Watkins. Zaidel, E. (1976). Auditory vocabulary of the right hemisphere following hemidecortication. Cortex, 12, 191211.

Pergamon

J. Neurolinguistics, Vol. 11, Nos 1-2, p. 119-136, 1998 ® 1998 Published by Elsevier Science Ltd. Allrightsreserved Printed in Great Britain 0911-6044/98 $19.00 + 0.00

PII: S0911-6044(98)00009-8

Processing of lexical ambiguity in patients with traumatic brain injury Karen L. Chobor* and Avraham Schweigerf *Department of Neurology, New York University Medical Center; tDepartment of Speech and Hearing Sciences, City University of New York Graduate Center

Abstract—The goal of this study was to determine the type of cognitive functions required for the interpretation of lexical ambiguity. Brain-injured and normal subjects performed lexical decision (reaction time) and matching tasks, using three categories of ambiguous words: homonymy, polysemy and metaphor. Patients showed much slower reaction times on metaphor than normals, while accuracy rates for matching were comparable, suggesting relative intactness of the semantic interpretations of these words despite a retrieval deficit.

Introduction Most words possess some indeterminacy in their meanings (Burgess & Simpson, 1988), so ambiguity may be a characteristic that pervades natural language processing (Simpson, Burgess, & Peterson, 1987; Swinney, 1982). In English, it has been estimated that over 50% of words have more than one meaning (Ziff, 1967). A significant amount of research has been done in the field of lexical ambiguity comprehension in both normal and brain-injured subjects, using the distinct categories of homonymy, polysemy, and metaphor. Results of these investigations have been interpreted according to models of selective or multiple access, the former referring to the access of a single meaning of an ambiguous term over other meanings according to word frequency and /or context effects, and the latter referring to the access of all meanings of an ambiguous term. In practice, it is not always clear into which category a given ambiguous word falls. The reason is that both etymological and psychological factors are relevant, and these may sometimes contradict each other. For example, words that are readily perceived as having quite distinct senses, such as port (harbor) and port (alcoholic drink) turn out to have a (tortuous but meaningful) link in a common Latin source (Durkin & Manning, 1989). The majority of lexical ambiguity studies in normals have been interpreted according to the multiple, or exhaustive, access model. In studies using brain-injured subjects, interpretation has centered on localizing the neurological substrate for such linguistic processing rather than determining the necessary cognitive mechanisms. Right hemisphere function has often been assumed to be necessary for successful processing of lexical ambiguity (Gardner et al., 1983; Brownell et al., 1984). A comprehensive review of the research in lexical ambiguity using both normal and brain-injured populations does not reveal such clear-cut interpretations. There are, in fact, a substantial number of studies that speak to selective access in favor of multiple access in normals. Further, results of studies using brain-injured subjects reveal much evidence suggesting cognitive difficulties traditionally assigned to the frontal lobes as responsible

120

K.L. Chobor and A. Schweiger

for the handling of multiple meanings. Such cognitive abilities include the conscious manipulation of dual meaning and the ability to shift from one category to another, functions otherwise subsumed under the category of abstraction. The purpose of this study was to clarify the cognitive functions required for the successful interpretation of lexical ambiguity. Brain-injured and normal subjects served as the study populations. Review of the literature Lexical ambiguity is a phenomenon whereby a single word may have more than one meaning. The two most common forms of lexical ambiguity are homonymy and polysemy, followed by metaphor. Homonymy is defined as one word having two or more distinct meanings, while both polysemy and metaphor refer to an instance whereby one word may have two or more related meanings. A finer distinction between the latter two reveals that metaphor is seen as instances of elliptical simile, where "like" or "as" is implied for the purpose of using one thing to stand for another. Thus, polysemy is an aspect of metaphor while remaining a distinct phenomenon. Beardsley (1962) states that a term may have a central meaning (its ordinary designation) and a marginal meaning (its connotation). The standard designation of wolf, for example, might include "mammal", "four legged", and "canine", whereas the marginal meaning would include "fierce", "voracious", "clever", etc. Meanings other than the central, or dominant, meaning are those which constitute polysemy and metaphor. The two divergent meanings of homonymy are, of course, both central meanings, such as the meanings for ball ("toy" or "a formal party for dancing"). Lexical ambiguity and brain function Right hemisphere function and lexical ambiguity The right hemisphere has traditionally been considered to be responsible for the comprehension of meanings secondary to the central meaning of a word (Gardner et al., 1983; Brownell et al., 1984). This includes alternative meanings (Chiarello, 1988) and the appreciation and integration of relationships in verbal discourse and narrative materials (Brownell, Potter, & Michelow, 1984) as well as the understanding of jokes (Gardner, 1994). Patients with right hemisphere brain damage exhibit a striking amount of difficulty in handling complex linguistic material. They exhibit clear difficulties in abstracting morals and in acquiring a sense of the overall gestalt of linguistic entities; they seem unable to appreciate the relations among key points of a story or joke. They may exhibit difficulties in interpreting phrasal metaphor or using verbal context to make linguistic judgments. Confronted with complex linguistic entities, these patients exhibit clear and recurring difficulties relating to the abilities to conceptualize the unit as a whole, to appreciate its purpose and form, and to integrate specific elements appropriately within these forms. Many of these patients seem insensitive to the context in which these linguistic entities are produced and utilized (Burns et al., 1983). The insensitivity to non literal meaning has also been observed at the single word level (Gardner et al., 1983). The right hemisphere

Processing of lexical ambiguity

121

also makes some contribution to language comprehension and reading processes (Krashen, 1976; Searleman, 1977; and Gazzaniga, 1983). Though there is overlap between right hemisphere functions and tasks of lexical ambiguity, it is clear from a summary of the work in this area that dichotomizing the left and right hemisphere will not suffice in explaining brain-damaged patients' difficulty with such material. In analyzing the cognitive processes that are necessary for the interpretation of ambiguous words, particularly the "abstract" demands in the accessing of the multiple meanings of a lexical item, it appears that these may be more localized to the frontal lobes. Thus, the anterior/posterior dichotomy of brain function may be of equal importance or of greater importance than (left/right-) hemispheric asymmetry when investigating lexical ambiguity. Frontal lobe function and lexical ambiguity Traditionally, damage to the frontal lobes has been associated with impairments in abstract thinking (Goldstein, 1948). It has been argued that abstraction and metaphoric thinking depend on a common underlying mechanism of part-whole relationship (Brown, 1997). This study investigated the relationship between the comprehension of lexically ambiguous words and the ability to apprehend an item as belonging to more than one category, the ability to simultaneously hold more than one category (or concept) in one's mind, and the ability to shift from one category (or concept) to another. Such conceptual or categorical thinking skills, first described by Goldstein (1948) as an "abstract attitude", have traditionally been assigned to the frontal lobe (Luria, 1966). In concrete behavior, the category cannot be accessed from the instance, e.g., the color "red" is not abstracted independent of a red apple. If the patient can accomplish this task, he may not arrive at two or more categories from an instance, e.g., that an apple is a member of the category of shape (round), or color, and of food. The impairment of abstraction can be relatively selective, as in defects of color naming or sorting, or it can be generalized and affect a great many perceptual tasks. In the latter cases, the deficiency is the basis for impairments on tests such as the Wisconsin Card Sorting Test (Heaton, 1981), in which the patient is required to sort objects along several dimensions. Given an instance of a category, e.g., shown a round red object and asked to group it with similar objects, the patient cannot derive the target category from the member items, nor can the patient sort along several dimensions (color, shape, etc.). Lezak (1995) describes conceptual difficulties, or problems with abstraction, as involving at least (1) an intact system for organizing perceptions even though specific perceptual modalities may be impaired; (2) a well-stocked and readily available store of remembered learned material; (3) the integrity of the cortical and subcortical interconnections and interaction patterns that underlie thought; and (4) the capacity to process two or more mental events at a time. It is proposed that the basis for difficulty with lexical ambiguity involves problems with mental shift considered a cognitive (conceptual) function associated with the frontal lobe rather than the reducedutility of the right hemisphere with linguistic items. On this view, brain damagedpatients who have difficulty (disproportionate to performance on other standardized measures) on tests which require the ability to appreciate two concepts at a single time, to shift from one interpretation of a word or concept so as to emphasize a different one, would be expected to have difficulty handling lexical ambiguity. In sum, it is proposed that the same mental process is responsible for (a) appreciating lexical ambiguity and for (b) mental shift tasks. The prediction is that a subject who

122

K.L. Chobor and A. Schweiger

performs poorly on (a) will also perform poorly on (b). It is also predicted that subjects who perform well on (a) will perform well on (b). While intact mental shift is necessary for lexical ambiguity tasks, it is a general cognitive function that is not only applied to such tasks. Effects of brain injury on the comprehension of multiple meanings It is commonly known by clinicians that brain damaged patients without aphasia have difficulty with polysemous words. For example, given a target word such as bank, and asked to point to words such as money, river, etc., that go with the target word, they will often select only one meaning. Other cases, asked the color of an orange, may say yellow or red. A similar phenomenon occurs with verbal nouns, e.g., "what do you shovel snow with?" Such patients are unable to deal with more than one meaning or interpretation at a time. There is an inability to retrieve the concept, perhaps due to blocking or persistence of the initial interpretation out of which the ambiguous item develops. There have been relatively few studies that speak to the question of lexical ambiguity in the brain-injured population, and findings among those that do exist have been inconsistent. It is likely that this arises from (1) difficulty in drawing distinctions between homonymous, polysemous, and metaphoric words which results in a lack of uniformity within and across studies of the selection of stimuli, and (2) differences in the manner of presentation of a task and the requiredresponse mode. These issues are illustrated below. Winner and Gardner (1977) presented tasks of metaphoric competence, i.e., matching metaphoric phrases to pictures followed by verbal description, to groups of normal and brain damaged subjects. For the right hemisphere brain damaged (RBD) patients, 43% of the initial responses were metaphoric ones (40% of the time). However, when asked to describe metaphors, these patients did quite well. The left-hemisphere brain damaged (LBD) patients selected the metaphoric picture in their initial choice 58% of the time, with literal pictures chosen 18% of tile time. These patients had difficulty in describing metaphors. This double dissociation of performances on the two metaphor tasks clarifies the contributions made by each hemisphere to linguistic and aesthetic functioning. More importantly, however, competent performance of the RBDs on the verbal description condition invalidates the common assertion that they are insensitive to metaphor. Brownell et al. (1990) examined metaphoric sensitivity to single polysemous words, whereby subjects were asked to judge semantic relatedness along a set of common adjectives: warm, loving, cold and hateful. The results of their study suggest a qualitative difference in the manner in which left- and right- brain-damaged patients process certain aspects of word meaning. The authors point out the striking finding that relative to the LBDs, the RBDs do not appreciate metaphoric meaning fully even at the single word level. This is inconsistent with the study by Winner and Gardner(1977) since in the latter study, the RBDs could use language to interpret metaphor. Tompkins (1990) presented an auditory lexical decision task to LBDs and RBDs, in which they heard a priming phoneme string and then a target phoneme string, either a real word or a non-word. Subjects were asked to indicate whether tile target string was a word or not; accuracy and time data were collected. Results of this study showed that brain damaged subjects' auditory lexical decisions for ambiguous target words can be facilitated by primes related to the metaphoric or literal meanings of those words. Brain damaged subjects' performance in conditions conducive to automatic processing resembled that of controls in all ways, except for absolute speed. Of note, RBDs did not have difficulty accessing

Processing of lexical ambiguity

123

metaphoric aspects of word meaning, perhaps because access proceeded relatively automatically in this task, and perhaps because the anterior-posterior distinction was not made. Using word triad and dyad tasks, Gagnon et al. (1994) did not find a clear disparity between LBD and RBD patients' ability to handle metaphor. The RBDs performed similarly to the LBDs for the metaphoric and neutral triad task, but significantly worse than the controls on the metaphoric triads. The LBD group always performed significantly lower than the controls in the dyad task. However, the RBDs performed similarly to the controls and much better than the LBDs when they had to perceive the presence of a primary meaning and an alternative neutral meaning in the dyad task. The RBD group performed like the LBD aphasic group and were significantly worse than the control group when they had to perceive the presence of an alternative semantic metaphoric meaning in a pair of words. Milberg et al. (1987) studied the processing of lexical ambiguities in Wernicke's and Broca's aphasic patients on performance of a lexical decision task. They were to decide whether the third word of an auditorily presented triplet series of words was "real" or not. The first and third words of each triplet were related to one, both, or neither meaning of the second word which was semantically ambiguous. The performance pattern of the Wernicke's aphasics was similar to that of normals; they showed selective access to different meanings of the ambiguous words, as demonstrated by the fact that the context provided by the first word affected semantic facilitation on the third word. In contrast, Broca's aphasics showed no semantic facilitation in the priming condition. These results are consistent with previous findings, suggesting that semantic representations may be largely spared in Wernicke's aphasics. The authors state that the complete failure of Broca's aphasics to demonstrate semantic facilitation in any condition in the experiment suggests that they may have a deficit either in the underlying representation of words or in accessing this information via automatic processing routines. The latter is consistent with the idea that impaired processing of lexical ambiguity requires intact frontal function. In contrast to these findings, Katz (1988) found that on an auditory lexical decision task with stimulus pairs containing ambiguous (semantically-related) or unambiguous (unrelated) words as primes, Broca's aphasics produced a pattern of results similar to normal subjects; namely, faster reaction times for target words preceded by semantically related than unrelated words (i.e., semantic priming). These results do not support Milberg et al. (1987) but rather suggest that the subjects in the Milberg et al. (1987) study had difficulty with the word triplet paradigm use. It is clear from the studies reviewed here that there is no consensus on the neurological or cognitive localization of lexical ambiguity. However, there is ample evidence that questions the contribution of the right hemisphere as necessary for this type of language processing. While there is controversy concerning most aspects of lexical ambiguity and its neural representation, some points can be made from the literature. First, all known meanings of a word are available to normal subjects (Simpson, 1984). Context, if present, then constrains interpretation (Onifer & Swinney, 1981). There is controversy over how long context effects take, whether meaning frequency or dominance affects the order and strength of meaning selection and whether meaning selection occurs automatically or under the control of the subject. However, once a meaning has been selected, additional but nonrelevant meanings of the ambiguous word are suppressed or eliminated within a third of a second so as not to interfere with the appropriate interpretation (Swinney, 1979). While neuropsychological studies are contradictory, damage to the right hemisphere does not produce a consistent relative sparing. There is, however, evidence in the clinical

124

K.L. Chobor and A. Schweiger

literature that the frontal lobes play an important role in word meaning interpretation or in the underlying ability to scan multiple interpretations serially or simultaneously. Research questions The purpose of the present investigation is to examine the possible contributing factors for the observations documented above, particularly the inconsistency of brain-damaged patients' abilities to access alternate word meanings. This includes the lack of any clear-cut disparity in the abilities of left- and right-hemisphere damaged patients to perform such tasks and the uncertainty of any difference in the automatic and conscious processing of ambiguous words. The following specific questions will be addressed: (A) Do patients with abstraction problems (as measured by neuropsychological tests) show reduced priming for ambiguous words and their multiple meanings, relative to unrelated words, despite their abstraction problems? To answer this question, subjects will participate in a lexical decision task that will measure speed of response using target ambiguous words paired with their multiple meanings, along with unrelated words and non-words. (B) Are problems with abstraction (as measured by neuropsychological tests) related to decreased ability to perceive lexical ambiguity of the three types tested here, i.e., homonymy, polysemy, and metaphor? To answer this question, experimental subjects will be administered a matching task whereby they will be required to match target ambiguous words with their multiple meanings, presented along with unrelated word choices. (C) Comparing brain-injured and normal subjects, are there differences in performance on tasks of lexical ambiguity of the three types tested here, i.e., homonymy, polysemy and metaphor (with specific attention to metaphor, given the greater abstraction demands on this word category)? Methodology Subjects Two groups of subjects were selected for this study. One of these groups consisted of 16 traumatically brain-injured (TBI) patients selected from an area outpatient rehabilitation center according to the criteria listed below. The other group consisted of 22 normal subjects drawn from the same rehabilitation setting (personnel and visitors), also according to the criteria listed below. Ages ranged from 19 to 45 years. Monolingualism was determined by having subjects rate their general experience with languages other than English according to a Likert scale (see Appendix A). Those scoring 2 or below (range: 04) were included in the study. All subjects were right handed; handedness was determined by A. J. Harris' Test of Lateral Dominance (1958). Education level was determined by asking the subject to indicate the highest degree attained by the time of testing, and only those subjects with a minimum of a high school education were included in the study. All subjects had normal or corrected vision and audition. The TBI subjects consisted of patients who had sustained a single traumatic brain injury over six months prior to testing, with no

Processing of lexical ambiguity

125

other history of neurological or psychiatric disorders. The normal subjects had no history of neurological or psychiatric disorders. In order to ensure adequate comprehension of the experimental procedure, subjects were required to achieve 100% yes/no reliability for six personally relevant or general information questions from the Western Aphasia Battery (Kertesz, 1982). Subjects were required to perform at a minimum of 90% accuracy on a 40-item practice run on a computer-generated reaction time task to ensure the ability to perform Task One. They were also required to demonstrate at least 90% accuracy in matching 4 choices to a target synonym arrayed in the manner in which Task Two was presented so as to indicate that visual/reading ability was sufficient to process printed stimuli physically similar to those of the task. The age range for the experimental group was 19-45 with a mean of 36.8 years. Twelve of the 16 patients in this group graduated from college, while the remainder had a high school education. Eight of the 16 were male and 8 were female. Duration post-onset ranged from 8 to 48 months. Test procedures EachTBI subject was requiredto perform Tasks One and Two (see below) in addition to undergoing a neuropsychological evaluation described below, while each normal control was required to perform only Tasks One and Two, in addition to receiving the Raven's Progressive Matrices (Raven, 1960) and the Shipley Institute of Living Scale (Zachary, 1991), both of which can be converted into IQ scores. Half of the subjects underwent Task One followed by Task Two; the other half underwent the reverse. Further, stimuli for both Task One and Task Two were randomized. The following ensured attention to the task: (a) subjects were visually and auditorily alerted to Task One by a large cross presented in the center of the screen, along with an audible tone; (b) only those subjects with accuracy scores of 80% or above on Task One were included; (c) subjects were given 2 ten-minute breaks at equal intervals during Task One; and (d) only those experimental subjects who were able to sustain 40 minutes of therapy were included in the study. Task One (for both groups of subjects) Task One consisted of a lexical decision task. Subjects were asked to indicate whether a letter string was a word or not by pressing a key on the computer's keyboard designated "yes" or "no". Each subject was seatedin front of a Macintosh quadra605 computer with a 14 inch colored monitor which presented stimuli at regular intervals, described below. Stimuli were presented in black on white background via the software package, Mac Probe, which collects reaction time and accuracy data The task was introduced to subjects with the following instructions: "You will hear a sound which will alert you to pay attention to the computer screen. At the same time, you will see a cross on the screen which is also intended to hold your attention. Then you will see two different words on the computer screen, shown one after the other. The first will be shown for only a short while, and may or may not be related to the second word. Do not pay attention to the first word. When the second word comes on the screen, press the "yes" key if you think it is a word in English and the "no" key if you do not think it is a word in English. Press the chosen key as quickly as possible."

126

K.L. Chobor and A. Schweiger

These instructions were followed by: (a) An audible "beep" lasting 500 msec which served as a warning signal intended to direct the subject's attention to the computer screen. (b) A "priming" word (e.g., "beam") was presented for 100 msec. (c) The target word, that is, a word related in meaning to the prime (e.g., "light" or "plank" for "beam") was presented for 300 msec. Interstimulus intervals (ISIs) were 500 msec. There were 60 lexically ambiguous primes, each paired with 1 non-word and 2 related words, which served as targets. Additionally, there were 60 non-ambiguous primes, each paired with 1 non word and 2 real words (chosen from the list of related words which were paired with the ambiguous words above), which also served as targets. Non-words were constructed by combining one syllable from each of two primes into pronounceable letter strings. Words and non words were randomized for presentation. This priming task was chosen to address the automatic levels of processing lexically ambiguous stimuli. The rationale for its use was to explore the pervasiveness of abstraction deficits on handling such material, i.e., levels of processing that are presumed to precede more conscious comprehension such as is required by a matching task. Thus, if patients with abstraction difficulties were found to have longer reaction times for lexically ambiguous material, then this would serve as evidence for impairment beyond the level usually tapped by studies of this topic. Task Two (for both groups of subjects) Task Two consisted of a word matching task that required subjects to determine whether 4 choices matched a target word in meaning. Stimuli consisted of a single target word printed at the top center of a 5 x 8 index card. Four other words were printed underneath this target word, spaced at regular intervals. This task was introduced with the following instructions: "You will see a single target word followed by 4 other words. Circle those that are related to the target word in meaning. There may be more than one correct choice." This matching task was designed to address conscious processing of lexically ambiguous material, and so, to provide converging evidence to data from the priming (automatic processing) task described above. In other words, the combination of the priming task with a matching task aimed to make the expected study findings more robust than if one or the other 'task were used. Materials For this study, the stimuli used to investigate the understanding of alternate meanings of words were of two types, as follows: (a) a priming word followed by target words of the following types: either of two of that (real word) prime's multiple meanings, unrelated words, or non-words, presented on a computer screen, and (b) a single target word or nonword followed by 4 choices, including two or three of that target's multiple meanings and unrelated words. Stimuli for part (b) were presented on 5x8 inch index cards with word sprinted in black lettering.

Processing of lexical ambiguity

127

The stimuli for Task One and Task Two consist of the following types of words: polysemous words, homonymous words, and metaphoric words (see Appendices A, B and C); non-words; and unrelated words. There are a total of 20 primes for each of the first three categories (homonymy, polysemy, metaphor)—totaling 60 ambiguous words. Additionally, there were 120 nonwords and 60 unrelated words which served as targets. Note that for the priming task, the words listed in Appendices A, B, and C served as the primes. For example, the target word "chicken" primed the target word "coward". Most of the polysemous and homonymous words were selected according to a list provided by Durkin and Manning (1989), compiled from a variety of sources, mainly form previous norm-gathering studies (e.g., Gilhooly & Logic, 1989; Gorfein, Viviani, & Leddo, 1982; Nickerson & Cartwright, 1984), with the majority of words chosen from the Gilhooly and Logic (1980) list as this list contains the most extensive normative data about the frequency, imagery, age of acquisition, familiarity, and concreteness of each ambiguous word. Word choices for these were matched for part of speech, length, and frequency. The metaphoric word list was generated by the investigator, along with two assistants (see acknowledgments) by using the concept of elliptical simile, described earlier. Further, two normal subjects (matched for educational level and age) were asked to choose 20 metaphoric words form the total list of 60 lexically ambiguous words, given instructions to "find those words in the "choices" column that can be used to stand for, or in place of, their corresponding targets." There was high interrater reliability: one rater correctly identified 19 items and the other, 17 items. One rater made 1 false positive response and the other, 0, while one rater made 1 omission and the other rater made 3 omissions. Word lists were created according to the following rules for selection: All of the polysemous and homonymous words had at least one noun meaning and verb meaning each. For Task Two, four words in each of the polysemous and homonymous word lists had three choices that were potentially correct (i.e., matching meaning); the remainder had only two potentially correct choices. For those that had a verb meaning as a correct choice, at least one foil was also a verb. Though many words in the total list had more than two meanings, viable synonyms were considered only those which were assigned a minimum relatedness score of 12.17 and a maximum relatedness core of 4.0 as assigned by Durkin and Manning (1989), where available. Eighty-five percent of the words used had this information available. As the features (i.e., frequency, imagery, age of acquisition, familiarity, and concreteness) considered for matching homonymous and polysemous words were not available for the metaphoric words, the latter category was treated as a separate category. Six of the metaphoric words and verb meanings. The metaphoric words as a group had two synonyms that were considered potentially correct choices for task two. Results Task One: Reaction time task Prior to analyses, trials on which reaction time measures fell farther than 2 standard deviations from the group centroid (by computing Mahalanobis d-squaredfor all raw scores within each subject) were eliminated, as these were considered to represent lapses of attention (a common practice in studies using reaction times). Using the two standard

128

K.L. Chobor and A. Schweiger

deviation guide, no more than 2% of the total number of trials for any one subject were eliminated from the analysis. The first step in the analysis of these data was to calculate the difference between the reaction times for the prime-target pairs, e.g., "chicken - coward" (metaphoric pairing) and "chicken - bird" (literal pairing). For the metaphoric word category, the order of subtraction was metaphoric minus literal values. As the two related homonymous and polysemous targets were matched for frequency, etc., the order of subtraction was random. The next step in the analysis was to obtain averages of the above calculations by category (homonymy, polysemy, and metaphor) for each subject, followed by averaged across subjects. These figures were used in all of the following analyses with the exception of the final analysis of variance (ANOVA). Initial simple correlations of all reaction time scores showed an unexpected correlation between homonymy and metaphor. Such a correlation between variable violated the assumption of ANOVA with repeated measures and artificially inflates the chance of finding spurious significance. Therefore, the initial method of analysis was a Multivariate Analysis of Covariance (with age as a covariate), or the Hotelling's T-Squared, which takes into account the dependence between variable when evaluating differences between two groups. This is a stringent measure of the difference between groups (patients and normals) across the reaction time variable (homonymy, polysemy, and metaphor). The overall multivariate results were evaluated using Wilk's Lambda (df = 3,24) which resulted in a value of 0.761 with p = .024. Since the previous analysis demonstrated differences between the groups, the next question of interest concerned which of the variables they differedon. Results of t-tests for each of the three variables (homonymy, polysemy, and metaphor) are depicted in Table 1. Dependent variable

t

df

Homonym Polysemy Metaphor

-.606 .462 2.905

36 36 36

Table 1.

2-tailed P .547 .646 .006

T-Test values for independent samples comparing patients and normals on the three category types. These values correspond to Graph 1.

There was a significant difference between the groups for metaphors, with an apparent difference between mean difference for metaphors and the other two variables in the patient group, i.e., it suggests a significant slowing in responding to metaphoric prime-target pairs when compared to polysemous and homonymous pairs in patients but not in normals. In fact, when compared to the unrelated prime-target pairs, there was no priming of metaphoric prime-target pairs when compared to polysemous and homonymous pairs in patients but not in normals. In fact, when compared to the unrelated prime-target pairs, there was no priming of metaphoric prime-target pairs. In order to statistically examine this relationship between the means, an ANOVA with planned comparison was performed (as it was hypothesized that responses to metaphoric pairs would be slower than the other two word categories): the mean differences for the metaphor category was compared with the combined means for the homonymy and polysemy categories since no difference was found between the latter two. The result of this analysis confirms the significance of the difference between metaphor and the other two variables {F(df = 1,36) = 4.16, p < .05}. Actual reaction times (means andS.D.) for patients and normals are shown in Table 2.

Processing of lexical ambiguity

129

Graph 1. Mean differences in reaction times for three word types. For each prime, reaction times for two related targets were computed and averaged across trials and across subject groups. This depicts a statistically significant difference between metaphor and the other two variables, i.e., it shows significant slowing in responding to metaphoric prime-target pairs when compared to polysemous and homonymous pairs in patients but not in normals.

H P Ml M H P Ml M

Patient mean 721 746 782 720 Normal mean

597 610 639 605

Patient S.D. 269 281 266 271 Normal S.D. 111 86.8 134 125

Table 2. Reaction Time performance (Mean and S.D.) for patients and normals on the following conditions: homonymous pairs (H), polysemous pairs (P), metaphoric pairs (Ml) and literal (M) word pairs.

130

K.L. Chobor and A. Schweiger

The next question for Task One concerned the relationship between performance on the neuropsychological battery and the reaction time data for the three variables (word categories). A canonical correlation showed that they are significantly correlated with the following outcome: R = .987; chi-squared (df = 27) = 43.69, p = .022. A more specific question concerning the relationship between tests of abstraction and reaction time performance was investigated using Backward Stepwise Multiple Regression Analysis. In this analysis, reaction time differences for metaphors was used as the dependent variable with scores on the neuropsychological battery as predictors. The following tests were noted to contribute most to the prediction of reaction time data in the regression equation: Wisconsin Card Sort Test, Trails B, Verbal I.Q., Word Fluency (FAS), Boston Naming Test, and Raven's. The results of this latter analysis with the standardized beta weights ate presented in Table 3. Variable VIQ BNT Trails-B FAS WCST-C Raven's

Standardized beta values -1.836 1.151 .658 .506 .483 .377

Table 3. Results of the multiple regression analysis for neuropsychological tests and the "metaphor effect" (F = 3.887, df = 6,8, p > .05, Multiple R = .863). The metaphoric-prime target pairs minus the same prime-literal meaning pairs. This shows the significance of the combined performance on tests of abstraction in predicting reaction time differences for metaphoric and literal word pairs, i.e., performance on BNT contributed the most weight to reaction time performance. In order to determine the presence of overall priming, an ANOVA was performed and the results showed a main effect for relatedness of prime for the patients (F = 4.92, p = .032), but no interaction, thus demonstrating the priming effect for patients but no difference for normals. As expected, the patients performed significantly slower than the normals with means as follows: TBI group: 764.04 and Normals: 612.72. Task Two: Matching task In contrast to the reaction time findings, ANOVA showed no main effect for group for the matching task (F = .93, p > .05). On the other hand, significant differences were noted among categories for both groups (F = 5.63, p = .01), also unlike the reaction time task. Note that even though the differences are statistically significant, they appear to be relatively small. There was a statistically significant difference among the three categories for performance on the matching task, with performance on homonymous words more accurate than polysemous and metaphoric words, and no significant difference between metaphor and polysemy (See Tables 4 and 5). However, this may be a spurious finding due to unknown variables, since this pattern was not revealed in the reaction time data, and since the actual differences are, indeed, small.

Processing of lexical ambiguity

Group Word Category Interaction

df 1,36 2,72 2.72

MS effect 88.81 29.09 04.82

F 1.88 5.63 0.93

131

p-level 0.18 0.01 0.40

Table 4. Results of the analysis of variance for matching performance

Patients Normals

Homonymy 93.9 95.6

Polysemy 92.1 94.2

Metaphor 93.2 94.2

Table 5. Means of percent correct for patients and normals, by categories Discussion The goal of this study was to determine the type of cognitive functions required for the successful interpretation of lexical ambiguity. The primary contention held that patients exhibiting problems with abstraction would demonstrate impairments in the sensitivity to alternate meanings of words and that this would be evident on both the lexical decision (reaction time) task and the matching task. Further, this was expected for the three categories of ambiguous words used: homonymy, polysemy, and especially for metaphor, given the greater demands on abstraction for this word type. The experimental procedures were developed to evaluate the difference between braininjured and normal subjects on these tasks, and were designed to allow analysis of any differential effect of word type (i.e., homonymy, polysemy, and metaphor) on lexical access and on a comprehension task. There is sufficient evidence that questions the degree of right hemisphere involvement in lexical ambiguity interpretation (Winner & Gardner, 1977; Gagnon et al., 1994) and other evidence that points to frontal lobe involvement in these types of tasks (Milberg et al., 1987). The results of this study support the latter, as is evidencedby the fact that tests of abstraction contributed much more weight than tests thought of as affecting right hemisphere function to predicting performance on lexical ambiguity tasks. Task One WAIS-R Verbal I.Q. scores contributed most heavily to predicting reaction time performance for metaphors, i.e., the higher the Verbal I.Q. score, the smaller the difference between the appreciation of concrete and metaphoric meanings of words in this category, indicating approximately equal facility of access of word meaning. The other tests found to correlate in this direction were WCST category subtest, Trails B, Word Fluency (FAS), BNT, and Raven's, with Ravens showing the lowest weight of this group of tests, which, save for the BNT are widely considered to be tests of abstraction. Note, however, that the standardized weight of the Raven's, while relatively small, was still in the same direction. It is notable that while the Raven's is a test of abstraction, it also requires intact visuospatial processing, and may thus also be considereda test of right hemisphere function.

132

K.L. Chobor and A. Schweiger

Reaction time performance for polysemy was most related to performance on WMS-R Verbal subtests and Trails B, and to a lesser degree with the WCST Perseveration subtest, Stroop, and Raven's. For homonymy, reaction time performance was most related to Word Ruency, Verbal I.Q., and BNT, and to a lesser degree, WCST, Trails B and Stroop. There was the expected finding that patients performed with longer response latencies on the reaction time task than did normals. Another expected finding was slower reaction times for words related to the target in meaning than for words unrelated to the target in meaning, for the patient group. The most striking finding on the reaction time task was that the patients evidenced significantly slower reaction times on metaphor than on polysemy and homonymy both within their own group and as compared to the normals. Task Two On the matching task, which used the same ambiguous words as the lexical decision task, patients performed at accuracy rates which were comparable to the normals. While this finding is surprising, it is important to note that the matching task involved untuned multiple choice. The finding that patients' performance was comparable to normals' performance suggests relative intactness of the semantic representations of these words in spite of retrieval deficits. It is noteworthy that patients and normals showed higher accuracy for responding to homonyms than to polysemous or metaphoric words on the matching task, despite the fact that the homonymous and polysemous words were matched closely in terms of frequency, etc. This suggests relative ease in discerning discrete multiple meanings such as for homonymous words as compared to a relative deficit in fully appreciating the marginal meanings of the polysemous and metaphoric word stimuli. Recall that homonyms consist of only central meanings, while polysemous and metaphoric words consist of both central and marginal meanings, which, within categories, are not mutually exclusive. General issues This study has revealed numerous methodological difficulties confronted in the field of lexical ambiguity. For example, divisions between polysemy and metaphor are somewhat blurred and often arbitrary. Indeed, Brownell et al. (1990) investigated brain damaged patients' appreciation of alternative word meanings using a word list which consisted of "metaphoric polysemous" words. In considering the above findings in light of the widely held definitions of metaphor, it would be prudent for researchers to clearly define the categories of lexically ambiguous words and to control for their individual parameters. In the classical literature on this topic which dates back to Aristotle, metaphor is seen as an instance of novel poetic language in which words are not used in their normal everyday sense (Lakoff, 1989). In contrast to language as the central realm of metaphor, Ortony (1993) considers metaphor to be more in the domain of thought, or the mapping of conceptual domains. Thus, the focus of metaphor is in the way we conceptualize one mental domain in terms of another, i.e., cross-domain mapping. For example, metaphor can be seen in terms of mapping from a source domain (tabloid) to a target domain (rag). Here, there is a conceptual correspondence between one domain of objects (newspapers) and a very different domain of objects (cloth). Such cross-domain mapping requires cognitive operations which are necessary for testes of abstraction, namely, set-shifting (WCST,

Processing of lexical ambiguity

133

Trails B), flexibility of thought (Word Fluency), and conceptualization (Verbal I.Q.)—all of which were strongly related to performance on reaction time for metaphor. Generally, then, the same cognitive functions are necessary for metaphor and tests of abstraction. To date, there is great variability in the methods used in studies on this topic, including subject selection, task requirements, and response mode. Aside from the obvious difficulties with issues such as these, they raise special concern with regard to the intention of this study. While the primary goal was to isolate the cognitive functions responsible for lexical ambiguity interpretation, an extension of this would involve identifying its neurological substrate. Thus, it was presumed that patients who exhibited difficulties in abstraction would also evidence frontal lobe impairment, and thus, difficulty on the study tasks. While most authors claim to use patient populations with discrete right or left lesions, there may exist an inherent bias toward selecting patients with anterior lesions, as a common exclusion criterion for right hemisphere patients is visual neglect—a symptom associated with right posterior damage. Further, most studies do not use the same lexically ambiguous stimuli over a variety of tasks, as this one did A further methodological difference in the literature on this topic concerns the use of sentences vs. words as stimuli. Single words were chosen for this study as this would assist in eliminating the possible influences of impaired attention or memory in the TBI subjects. Additionally, differences exist among priming studies which use sentences. For example, Onifer and Swinney (1981) showed differences in priming ability when word choices were presented immediately following the occurrence of the ambiguity in a sentence, contrary to when these word choices were presented 1.5 seconds following occurrence of the ambiguity. Additionally, metaphoric target phrases require significantly longer processing times than literal target phrases if reading time measures are taken at the end of the target phrase rather than at the end of the target sentences in which the phrases are embedded Given that there were dissimilar findings for both tasks in this study, this calls into question the validity of findings of studies which use one or the other type of task. Indeed, the matching data in this study revealed the relative facility of access even in the braininjured population, while the reaction time data showed this population to be quite below normal levels of performance. As there is some question as to the issue of how the number of meanings for lexically ambiguous words affects their comprehension, a sampling of performance of both patients and normals on the matching task was taken in order to see if accuracy was affected by this parameter. Of the sixty target items, eight had three potentially correct matches, while the remainder had only two. A preliminary analysis of these data indicated that patients and normals performed at approximately equal accuracy rates (48% and 51%, respectively) for those items with three potentially correct matches, with much higher accuracy for those items with only two (92% and 94%, respectively). Though this study was not set up to address the question of selective vs. exhaustive access, this preliminary analysis would point in the direction of supporting selective access, whereby the access of two meanings on an ambiguous word may suppress the third. Certainly, this issue should be addressed in a more rigorous manner, as it raises important questions regarding the possible contributing factors to disambiguation. Thus, in addition to providing support for the cognitive localization of lexical ambiguity to function (such as abstraction) related to the frontal lobe, this study has highlighted some of the concerns related to other investigations of this topic. The literature is replete with studies on lexical ambiguity, particularly with normals, and it is evident that findings

134

K.L. Chobor and A. Schweiger

using the brain-injured population would add new perspective on the topic, when such studies are stringently controlled. Conclusions The most common way of investigating sensitivity to lexical ambiguity is through the use of tasks which require conscious processing, such as the matching of words or phrases. The present study differed in that it used both a matching task and a priming (reaction time) task; the latter was included in an attempt to contribute to the robustness of the expected findings, i.e., worse performance for patients with abstraction difficulties. However, a discrepancy in performance on these tasks was found, with brain-injured subjects performing as well as normals on the matching task, though performing with much slower reaction times on the priming task. This is important in highlighting the stage of cognitive processing (automatic vs. conscious) at which disambiguation takes place, and also points to the necessity of careful design in studies on this topic. Given the large discrepancies found in other studies of the topic of lexical ambiguity using brain-injured subjects, serious consideration was given to other possible confounding variables in the design of this particular study. These include subject selection, task requirements, response mode, and selection of stimuli. While this study sets out to isolate the cognitive processes involved in lexical ambiguity comprehension, it did not rely on a patient population with clear-cut lesion sites. Instead, it used the procedure of evaluation to determine purported lesion site, and provide findings regarding right hemisphere and frontal involvement in lexical ambiguity. The findings provided support for the cognitive mechanisms underlying the comprehension of lexical ambiguity, all of which are traditionally assigned to the frontal lobes. These include the ability to simultaneously apprehend one or more item and to shift from one item (or mental set) to another. A natural progression in the study of this topic would be to attempt to localize the neurological underpinnings for these cognitive mechanisms, a task made simple by the use of a population of subjects with focal brain damage, i.e., left, right, anterior, or posterior. References Beardsley, M. (1962). The metaphorical twist. Philosophy and Phenomenological Research, 22,293-307. Brown, J. W. (1997). Process and creation. In A. Andersson and N.-E. Sahlin (Eds.), The complexity of creativity (pp. 35-50). Dordrecht: Kluwer. Brownell, H. H., Potter, H. H., Michelow, D., & Gardner, H. (1984). Sensitivity to lexical denotation and connotation in brain damaged patients: A double dissociation? Brain and Language, 100,717-729. Brownell, H. H., Simpson, T. L, Bihrle, A. M., Potter, H. H., & Gardner, H. (1990). Appreciation of metaphoric alternative word meanings by left and right brain damaged patients. Neuropsychologia, 28, 375-383. Chiarello, C. (1988). Semantic priming in the intact brain: Separate roles for the right and left hemispheres? In C. Chiarello (Ed.), Right hemisphere contributions to lexical semantics (pp. 59-69). New York: SpringerVerlag. Durkin, L., & Manning, J. (1989). Polysemy and the subjective lexicon: Semantic relatedness and the salience of intraword sentences. Journal of Psycholinguistic Research 18, 577-612. Gagnon, L., Goulet, P., & Joanette, Y. (1994). Semantic processing of ambiguous words after a right hemisphere lesion. Poster presented at the International Neuropsychological Society. Angers, France. Gardner, H., Brownell, H., Wapner, W., & Michelow, D. (1983). Missing the point: The role of the right hemisphere in the processing of complex linguistic materials. In E. Perecman (Ed.), Cognitive processes in the right hemisphere (pp. 169-191). New York: Academic Press. Gardner, H. (1994). The stories of the right hemisphere. In W. Spaulding (Ed.), Forty-first Nebraska symposium on motivation 1992-1993. Lincoln, NE: University of Nebraska Press.

Processing of lexical ambiguity

135

Gazzaniga, M. S. (1983). Right hemisphere language following brain bisection: A 20-year perspective. American Psychologist, 38, 525-537. Gilhooly, K. J., & Logic, R. H. (1980). Meaning-dependent ratings of imagery, age of acquisition, familiarity, and concreteness for 387 ambiguous words. Behavior Research Methods and Implementation, 2, 428450. Goldstein, K. (1948). Language and language disturbances. New York: Grune and Stratton. Harris, A. J. (1958). Harris Tests of Lateral Dominance. Manual of directions for administration and interpretation (3rd ed.). New York: The Psychological Corporation. Heaton, R. K. (1981). Wisconsin Card Sorting Test. Odessa, FL: Psychological Assessment Resources, Inc. Kertesz, A. (1982). Western Aphasia Battery. Ontario: The Psychological Corporation. Krashen, S. D. (1976). Cerebral asymmetry. In H. Whitaker and H. A. Whitaker (Eds.), Studies in Neurolinguistics (Vol. 2, 157-191). New York: Academic Press. Lakoff, G. (1993). The contemporary theory of metaphor. In A. Ortony (Ed.), Metaphor and thought. Cambridge University Press: New York. Lezak, M. D. (1995). Neuropsychological assessment (second ed.). New York: Oxford University Press. Luria, A. R. (1966). Higher cortical functions in man. New York: Basic Books. Milberg, W., Blumstein, S. E., & Dworetzky, B. (1987). Processing of lexical ambiguities in aphasia. Brain and Language, 31, 138-150. Ortony, A., (1993). Metaphor and thought (2nd ed.). Cambridge: Cambridge University Press. Onifer, W., & Swinney, D. A. (1981). Accessing lexical ambiguities during sentence comprehension: Effects of frequency of meaning and contextual bias. Memory and Cognition, 9, 225-236. Raven, J. C. (1960). Guide to the Standard Progressive Matrices. London: H. K. Lewis. Searleman, A. (1977). A review of right hemisphere linguistic capabilities. Psychological Bulletin, 84, 503-528. Simpson, G. B. (1984). Lexical ambiguity and its role in models of word recognition. Psychological Bulletin, 96, 316-340. Simpson, G. B., Burgess, G. B., & Peterson, R. R. (1987). Human comprehension processes and the indeterminacy of meaning. Cognitive Systems, 2, 213-232. Swinney, D. A. (1979). Lexical access during sentence comprehension: (Re)consideration of context effects. Journal of Verbal Learning and Verbal Behavior, 18, 645-659. Swinney , D. A. (1982). The structure and time course of information interation during speech comprehension: Lexical segmentation, access, and interpretation. In J. Mehler, E. C T. Walker, and M. Garrett (Eds.), Perspectives on mental representation: Experimental and theoretical studies of cognitive processes and capacities. Hillsdale, NJ: Lawrence Erlbaum. Winner, E., & Gardner, H. (1977). The comprehension of metaphor in brain-damaged patients. Brain, 100, 719-727. Zachary, R. (1991). Shipley Institute of Living Scale. Manuel, CA: Western Psychological Services. Ziff, P. (1967). Some comments on Mr. Harmon's confabulations. Foundations of Language, J, 403-408.

Appendix A Metaphoric words

Primes (for RT task)

Targets (for RT task)

Targets (matching task)

Choices (matching task)

BUG BLUE PIG SHARK RAG NUTS DOUGH PEANUTS LEMON DOG BUTTERFLIES SHADES EGG FRESH CHICKEN WHEELS BOOT TUBE BANANAS GREEN

INSECT COLOR SLOB BIRD CLOTH VEGETABLES BREAD SNACK FRUIT PURSUE INSECTS CURTAINS INDUCE OPEN BIRD BUTTONS CHAIN STEREO CRAZY COLOR

SNAKE SMELL FRUIT THIEF PILLOW CRAZY MEAT SMALL FENCE ANIMAL FURNITURE SUNGLASSES FOOD NEW REPTILE HOUSE KICK TELEVISION SNEAKY TEXTURE

PESTER SAD COMEDIAN SAILOR TABLOID RELAXED TOWEL WHEAT MINERALS PLANT FUNNY EAT FURNITURE FRAYED COWARD TIRES GLUTTON CYLINDER FRUIT OLD

RESERVE HAPPY ANIMAL FISH CHECK SEEDS MONEY LARGE CAR KNOW NERVOUS WATCH GESTURE BRAT LAZY CAR SHOE STEM WACKY NEW

136

K.L. Chobor and A. Schweiger

Appendix B Homonymous words

Primes (for RT task)

Targets (for RT task)

Targets (Matching task)

Choices (Matching task)

ROCK SCALE PLANT FILE BAT BILL STABLE POST BANK DECK BOLT BALL COUNT SEASON DATE PUNCH SHIP COACH PORT PALM

MINERAL FISH VEAL WATER STOOL TAIL STEADY FLOAT FREEZER ROLL SCREW DANCE ADDITION TIME TIME HIT BOAT BUS HARBOR HAND

GAME LEAF FLOWER TOOL VEGETABLE DOLLAR KITCHEN MAIL VALLEY SHIP SPOON PARADE NOBILITY BURN FRUIT DRINK TRACTOR TEACH WINE CABIN

SWEEP PAINT BUILDING CABINET CLUB SHOE HORSE AFTER MONEY CARDS SWIM CLOTHING SING PEACE CALENDAR READ THROW HELICOPTER CITRUS KNIFE

SWAY BALANCE CASTLE BUSH ANIMAL BIRD FAST WOOD RIVER DOOR RUN TOY NURSE FLAVOR POEM ROSE SEND PAINT SUITCASE TREE

Appendix C Polysemous words

Primes (for RT task)

Targets (for RT task)

Targets (matching task)

Choices (matching task)

HIDE PART MOUTH LAP CHOP STEP BOX LIGHT SINK TRIP NOTE CHARM LETTER PLUG STUMP BEAN BLOCK SHEET FORK MARCH

CONCEAL SECTION PEAK SQUEEZE CLEAN WALK TYPE CABOOSE RECEPTACLE FOUNTAIN MESSAGE QUALITY ALPHABET STOPPER TREE LADDER CUBE PAPER ROAD JUMP

SKIN STAPLE FACE PAGES SKIRT DRIVE CONTAINER STEM SHED FALL SEED FLATTER STAMP PUMP LEG SMILE WOOD UMBRELLA CLOUD PROTEST

VENEER ROLE RIVER KNEES CUT RIBBON STEP IGNITE SUBMERGE JOURNEY MUSIC JEWELRY SPOKE ELECTRIC PUZZLE PLANK STOP BRAID COMB MONTH

RUB COIL ROOF LICK MEAT STAIR FLIGHT LAMP IRON SIT PLAY DRESS OPERA HOSE JOKE LIGHT DIVE BED UTENSIL WALK

Persamon

©

1998

J. Neurolinguistics, Vol. 11, Nos 1-2, p. 137-152, 1998 Published by Elsevier Science Ltd. All rights reserved Printed in Great Britain 0911-6044/98 $19.00 + 0.00

PII: S0911-6044(98)00010-4

Shifting the burden to the interlocutor: Compensation for pragmatic deficits in signers with Parkinson's disease Judy A. Kegl and Howard Poizner Center for Molecular and Behavioral Neuroscience, Rutgers, The State University of New Jersey

Abstract—This paper takes a qualitative look at the behavior of interlocutors in the context of conversations with deaf signers of American Sign Language (ASL) who have Parkinson's disease (PD), a neurodegenerative disease affecting sensori-motor planning and execution, attentional systems, and memory. Deficits in these areas, while non-linguistic, impact pragmatic aspects of conversational interaction. Three signers with PD at varying levels of severity (mild, moderate-to-severe, and severe) are compared and contrasted with attention given to the effects of PD on their use of conversational regulators and turn-taking devices. The behavior of persons engaging in conversation with signers who have PD is shown to corroborate in tangible ways what we have identified as the attentional and sensori-motor deficits of PD. Interlocutors' complex repositioning to maximize visual acuity, increased use of shift regulators, hyperattentiveness to sign initiation, as well as increased reiteration and expansion of information all serve to amplify and reinforce the contributions of signers with PD to ASL conversations. Over the course of the disease, unimpaired interlocutors gradually "become" the attentional system of signers with PD. Increased tapping and attention-getting devices used to explicitly reorient attention compensate for attentional deficits and maintain the participation of signers with PD in multi-participant conversations. Introduction Pragmatics can be thought of as language in context, both the real world context and discourse context. Real world context focuses on using shared knowledge to inform interpretation. Discourse context focuses upon chunks of communication beyond single sentences or phrases that cohere in an interdependent fashion. These chunks can be comprised of independent sentences produced by a single speaker that are strung together in a narrative or comprised of utterances interwoven by several participants in the context of a conversation. Analysis of language in context must take into account not only grammatical competence on the part of producers of language output but also the viability of the message as received by the interlocutor. Sometimes the behavior of the non-neurologically impaired interlocutor can be the best indicator of the nature of a deficit that impacts conversations they are participating in. This paper takes a qualitative look at the behavior of the interlocutor in the context of conversations with Deaf signers of American Sign Language who have Parkinson's disease, a neurodegenerative disease involving motor and attention deficits that impact pragmatic aspects of conversational interaction. We begin by reviewing aspects of the deficits of unilaterally cortically-lesioned signers as a basis of comparison.

138

J. Kegl and H. Poizner

Signers with cortical lesions Damage to posterior portions of left cerebral cortex The sentence has been claimed by linguists such as Noam Chomsky to have a privileged linguistic status. Examination of a left temporo-parietal lesioned signer of ASL (as compared with his neurologically intact identical twin) revealed a dissociation of linguistic versus affective facial expression focused specifically at the sentence level (Kegl & Poizner, 1997). The aphasic signer exhibited preserved abilities to use word-level adverbial facial expressions, discourse level markers of topic, and affective facial expression, while exhibiting specific problems with clause-level grammatical facial markers that are associated with complementizer position (such as wh-question markers, markers of conditional expressions, etc.) as well as syntactically moved topics. This deficit in the use of clause level grammatical facial expressions coexists with reduced production of syntactically possible word orders and construction types in ASL. This interdependence makes sense since permutations in word order are frequently signaled by differences in the occurrence and spread of facial expressions. In addition, this signer tends neither to consistently inflect signs with subject and object agreement nor to consistently use other aspects of spatial grammar.

Compensations Despite the syntactic deficits identified in this left-lesioned signer, interlocutors use a variety of strategies, often unconscious, to maximize communicative success. For example, in signing to him, his wife and brother signed slower and used a much higher proportion of ASL sentences using canonical SVO ordering. Sentences produced by this left-lesioned signer were often repeated in an expanded and inflected form by the interlocutors and correct interpretation was verified. Damage to posterior portions of right cerebral cortex Sentence level processes in both signed and spoken languages seem particularly susceptible to damage that occurs in the Perisylvian regions of the left hemisphere, while discourse level disruptions occur more frequently with right hemisphere damage. Neurological evidence for a distinction between sentence and discourse was found when we examined differential impairment of parallel grammatical processes that occur at both the level of syntax and discourse in a right temporo-parietal lesioned hearing signer of ASL (Kegl & Poizner, 1991; Poizner & Kegl; 1992). This right-lesioned signer was able to establish and consistently maintain both person and locative agreement within the confines of the sentence, but failed to do so between sentences strung together in a discourse. Agreement is a process that links two grammatical entities. In person agreement, noun phrase arguments are linked to morphological markers of subject and object on the verb via association with the same location in the signing space. This process signals which noun phrase functions as the subject and which functions as the object. In locative agreement, oblique arguments are similarly linked to their respective verbs, registering the locative or directional relations they bear in a given clause. From sentence to sentence within a discourse, the person agreement and locative agreement of noun phrases with the same pre-

Compensation for pragmatic deficits in signers with PD

139

established locations in the signing space, serve as a cohesive device that maintains reference throughout a narrative or conversation. This same right-lesioned signer exhibited problems in the use of discourse level markers of role shift (breaking of eyegaze with interlocutor, shift of the body or head orientation toward the referent whose role is being taken on, facial caricature, etc.), while consistently marking role shift at the sentence level via a change of third person to first person grammatical agreement (Loew, Kegl, & Poizner, 1997) and while consistently using eye gaze and head tilt for grammatical agreement. In transitive sentences in ASL, eyegaze to a referent marks object agreement, while head tilt toward a referent is a marker of subject agreement (see Bahan, 1996). Compensation Except for some problems with the linguistic encoding of direct visuo-spatial mapping (using verbs of motion and location and classifiers to describe real world spatial relations) the communication of this right-lesioned signer was highly grammatical and intelligible at the level of the sentence. Beyond the sentence, the inconsistencies of coreference from sentence to sentence are compensated for in two ways. The signer herself compensates for inconsistent discourse coreference by redundantly re-establishing the referential indices of referential noun phrases in each sentence. So, in essence, each sentence becomes its own encapsulated discourse. The interlocutor compensates, almost unwittingly, by reconstructing these redundant sentence sized narratives into a larger cohesive text. In the case of role play, the interlocutor typically uses shared knowledge to reconstruct insufficiently signaled cases of role play. Another cue is to attribute a role shift to situations in which maintaining the same first person referent would be illogical. In any case, the burden of reconstructing either coreference through a narrative or instances of role shift are placed upon the interlocutor rather than the narrator. Summary Cortical lesions tend to affect circumscribed components of linguistic (or other cognitive) capacities, leaving other aspects intact. The compensations made on the part of the interlocutor tend to occur at the level of error correction in the course of interpretation, and are frequently so automatic that the errors and stylistic aberrations on the part of the rightlesioned signer actually go unnoticed. Signers with Parkinson's disease In contrast with cortical lesions, subcortical degeneration of the sort observed in Parkinson's disease (PD) affects sensori-motor planning and execution, attentional systems, and memory. Yet, deficits in these areas, while non-linguistic in nature, can directly impact language production and comprehension.

140

J. Kegl and H. Poizner

Parkinson's disease Patients with Parkinson's disease are thought to decompose high level motor plans, being unable to automatically execute multiple motor tasks either sequentially or simultaneously (Marsden, 1982; Marsden & Obeso, 1994; Harrington & Haaland, 1991; Weiss et al., 1997). They seem unable to program movement parameters for the next motor act (Marsden, 1987; Beneckeet al., 1987), or to scale movement amplitude (force) to fit task requirements (Margolin & Wing, 1983; Hallett & Khosbin, 1980). Manifestations of Parkinson's disease in signers of ASL Manifestations of PD tend to be characterized by across-the-board effects on sign articulation, attention, and memory rather than by circumscribed linguistic errors. Flattening of distinctions Typically, language tends to maintain a delicate balance between ease of articulation and ease of perception. Overall, signers coping with the effects of Parkinson's disease adjust this balance to favor ease of articulation over ease of perception, putting the burden of perception on their addressees. In other publications (Kegl & Poizner, 1993; Brentari & Poizner, 1994; Brentari, Kegl, & Poizner, 1995; Loew, Kegl, & Poizner, 1995) we have discussed the nature of the ease-of-articulation changes that characterize PD. The examples discussed below can be found in those publications. Accommodations favoring ease of articulation include an overall laxing of handshape, movement, orientation and place of articulation distinctions such that each handshape is only minimally distinguished from the others. With handshape, this can involve failure to fully open or close the fist; failure to maximally spread, extend, or contact the fingers; or failure to suppress the involvement of non-selected fingers in a given sign articulation. Movements are laxed by being only minimally realized. For example, a translatory movement that extends from slightly in front of the signer to near the addressee can be reduced to orientation from signer to addressee with the most minimally perceptible movement from one toward the other. Hand internal movements are laxed by minimizing the separation between or bending and closure of the fingers. Whole arm movements are reduced to movement at fewer or even to a single joint. For example, shoulder and elbow movements may be distalized to the wrist or even fingers. Orientation is often shifted to an unmarked position. For example, palm upward orientation of the baseband in a sign like NEW, or READ, may be shifted to a sidewards orientation, forcing the entire articulation of these two-handedsigns to follow suit. Place of articulation is often laxed by failure to make contact with the location on the body with which the sign is associated. In addition, the shortening of sign movements can lead to a shortening of the distance between, and therefore a change in the locus of, sign articulations. While raising can occur as a result of a reduced transition between a sign made at the face and a following sign made at chest level (typically only in signers with mildPD), signers with PD tend to produce signs lower than usual, possibly as a result of lessened resistance to gravity in an attempt to reduce effort. In fact, there is an across-theboard shrinking and lowering of the signing space over the course of the disease. Facial expressions are also compromised. Affective facial expressions are replaced by an almost masklike expression. Grammatical facial expressions, responsible for carrying the bulk of syntactic information in ASL are slower to disappear.

Compensation for pragmatic deficits in signers with PD

141

One accommodation is that maximum articulation of a given grammatical facial gesture is not attained. The eyebrows are only slightly raised; the brow only slightly furrowed. In fact, facial expressions can also become distalized. For example, the raised eyebrows of a yes/no-question are often realized instead by a slight raising of the upper eyelid. Another accommodation involves changes in the way facial expressions are mapped onto the signed sentence. Aarons, Bahan, Kegl, and Neidle (1992) note that grammatical facial expressions are associated with a given position in a sentence and spread from that position over a circumscribed syntactic domain (known as a c-command domain). As a result, the facial expression builds to or wanes from a maximal articulation associated with that position (Bahan, 1996). In signers with PD, this mapping of an independently modulated facial expression to a distinct position on an independently signed manual sequence of signs is the kind of sensory motor act of planning and coordination that poses problems. Instead, the facial expression, reduced but unmodulated for intensity is coupled with the manual string and spread evenly over the clause. (See Kegl, Cohen, & Poizner (in press) for a description of this alternate form of non-manual mapping to a manual sequence of signs and Brentari, Kegl & Poizner (1995) for parallel examples of the coupling of articulatory components solely within manual articulations of signing as produced by signers with PD.) Short of the most severe cases, it should be noted that the laxing and flattening of distinctions noted in all PD signers are not a mandatory result of a compromised motor system. Signers with PD are able to produce good exemplars of signs in isolation. The ease of articulation accommodations mentioned here occur in extended sign sequences such as signing full sentences or chunks of discourse. Thus, requirement for production of extended sign sequences make it much more likely that a signer with PD will show the reductions and laxing discussed above. Attentional problems In terms of attention, signers with PD tend to have a hard time disengaging attention from one person to another, so that in conversations involving more than two participants, signers with PD find it difficult to shift attention from one interlocutor relinquishing a turn to the next interlocutor who has recently taken the floor. In a signed language conversation where visual contact is crucial to receiving the message, failure to shift one's attention to the next contributor to the conversation can leave one lost. Furthermore, failure to give the appropriate addresseeresponse that signals willingness to yield the turn, may leave the signer initiating a turn unsure of what to do. Summary We are working from the premise that the linguistic competence of Parkinsonian signers is left intact, and that they are, in fact, using aspects of this competence to systematically lax and reduce signs, making them easier to produce. However, the modifications made in Parkinsonian signing combined with attentional problems can interfere greatly with the pragmatic aspects of language use, particularly participation in conversations. Flattening of distinctions in sign articulation makes the Parkinsonian signer more difficult to understand, problems with initiating signed utterances and attentional problems make it difficult for the signer with PD to maintain participation in the conversation at all. Moreover, their failure to give the appropriate addressee responses in timely fashion may even lead to their being seen as disruptive of the conversational flow.

142

J. Kegl and H. Poizner

Regulators and turn-taking in ASL conversations Before turning to our study of the impact of Parkinson's disease on ASL conversations, we need to present some basic background on the nature of conversational regulators and turntaking devices in ASL. Basically, in any conversation, participants take on one of two roles: 1. the participant who at any given moment claims the speaking turn (Duncan, 1972) and 2. the participants that at any given moment do not claim the speaking turn (Kendon, 1967). The foundational article in this area was written by Baker (1977), who used a classification system proposed by Weiner and Devoe (1974) to analyze a corpus of conversational data elicited from two dyads of deaf signers. Baker videotaped two conversations and analyzed the initial portions for regulator behaviors. These excerpts were short, about 58-82 signs. The forms Baker (1977:218-219) found for initiation, continuation, and shift regulators are summarized below. These regulators form the basis for our study of Parkinsonian signers, although we will not limit ourselves to dyadic conversation. Initiation regulators According to Baker's very careful analysis, a signer initiates a turn by raising or extending the hands from a full-rest (arms at sides, relaxed), half-rest (waist-height, palms facing signer), or quarter-rest (resting one hand against the body above the waist) position, followed by some attention-getting gesture (pointing, touching, waving the hand), and an optional lean forward toward the addressee. Statements begin with a breaking of eyegaze once the addressee's attention is confirmed, while questions maintain eyegaze with addressee. Addressees signal their permission for the signer to initiate a turn by making eye contact1 with the signer and by assuming a rest position and not signing. Continuation regulators The signer currently in control of a conversational turn signals that they will maintain the turn after a bounded information chunk has been completed or after a brief pause by breaking eyegaze with the addressees, optionally increasing one's signing speed, and by not returning to a rest position. Other methods of holding the floor include the use of filled pauses that indicate thinking or pondering such as looking up, using a quizzical expression, shaking the index finger or palm, or by some postural shift. Another option is holding the final configuration and position of the last sign articulated throughout the pause. 1 While maintaining eye contact may seem necessary for the mundane purpose of visually processing the signed message, it is also an important feature of the addressee's response. One bit of evidence for this comes from conversational interactions between the right-lesioned signer mentioned earlier and her interlocutors. This subject tends to monitor her interlocutor's signing and visually track (using smooth pursuit, which necessitates breaking eyegaze with her addressee) those signs that move into the lower quadrant of her left visual field. This behavior is immediately noticed by her interlocutors and is extremely disruptive. Despite knowing that this behavior is a consequence of her lesion, it is difficult for her interlocutors not to register this behavior as rudely violating conversational protocol. This insuppressible reaction to a pragmatic violation contrasts sharply with the almost unconscious compensations made by these same interlocutors in response to the right-lesioned signer's failure to mark role shifts or to maintain coreference across a discourse.

Compensation for pragmatic deficits in signers with PD

143

The addressee signals continued attention and willingness to remain the observer by maintaining eyegaze and offering various subtle back-channel cues such as head-nodding, smiling, postural shifts, facial activity, uncertainty, or confusion. Addressees optionally point toward the signer after each proposition or repeat the ends of sentences. As might be expected, this continuation regulator feedback is common and more pronounced in the feedback that sign language interpreters give to deaf speakers as they process their signed communication for subsequent voice interpretation. Shift regulators A signer signals willingness to yield the turn to another participant in the conversation by reestablishing eyegaze with the addressee; by an optional decrease in signing speed near the end of the turn, by optionally calling for an addressee response via a "what do you think?" gesture; by pointing to the addressee; by holding or raising the last sign while making a questioning facial expression; or by returning the hands to a rest position. The addressee signals a turn-claim during another's turn by optionally increasing and intensifying head nodding and pointing at the prior claimant; by switching palm orientation from toward oneself to palm upward with the heel of the hand higher than the fingertips, by moving out of rest position, or via a number of options enumerated in the initiating and continuation regulators sections above. To initiate a turn by interrupting the current claimant, the addresseerepeats their first few signs until the current claimant makes eyegaze and yields the floor or until the current claimant suppresses the addressee's turn claim. Summary As can be seen, the mechanics of conversational control are complex, multi-leveled and require constant monitoring of the interlocutor(s). Claiming a turn becomes even more complex as we move from the conversational dyad to larger groups of participants. With increasing demands on the number of participants to be monitored and on the number of competing claims to fend off, we would expect the Parkinsonian signer to become increasingly disadvantaged in the conversational context. Turn-taking in signers with Parkinson's disease We examined conversations involving three ASL-signing2 Deaf men in their seventies, each of whom had been previously diagnosed with Parkinson's disease, one mild, one moderate to severe, and the other severe. Data were drawn from a two-hour first intake interview, where background information concerning cause of deafness, schooling, employment and family history were collected. Each interview incorporated many opportunities for open-ended conversation as well. In the course of these interviews each subject engaged in both dyadic (two-way) conversations and conversations with more than two participants. In these two types of situations, we noted their behavior with respect to 2 All three individuals were prelingually deaf, culturally integrated into the Deaf community, and considered ASL as their primary and preferred form of communication.

144

J. Kegl and H. Poizner

monitoring the participation of others and in claiming and yielding conversational turns. In particular we noted instances of physically touching the subject to get attention, eye or head orientations toward an interlocutor (JL, JK, VJ and JC are all testers serving as interlocutors; all but the author JK are Deaf), cases where the subject failed to observe a turn claim by another participant, signals of turn claiming and turn yielding on the part of the subject, the use of back channel signals like head nods, copying the ends of sentences, and gestures of understanding or disbelief, and finally cases where the subject continued to sign simultaneously with another participant in the conversation (indicating competition for the floor or repetition by the interlocutor of the subject's signing as a means of verifying meaning.) While our analysis is based upon behaviors in the interviews as a whole, in each of the sections below, we present coding results from a few minutes of conversation from each of the three signers with PD to give a flavor of the differences between their performances. Mild Parkinsonian (LN) Table 1. presents a few minutes of conversational turn-taking by mild Parkinsonian subject LN. Roughly each row corresponds to a turn.3 indicating that in dyadic conversation, where his efforts can be focused upon claiming and yielding turns, his performance falls within the normal range. He alternated turns with ease and consistently offered back-channel signals of attentiveness and understanding whenever yielding the floor to his interlocutor. He frequently initiated turn claims and at one point (see the 2.5 second overlap in signing), he actually competed with his interlocutor for the turn. When we move from dyadic to more than two conversational participants, the monitoring of cues from multiple participants puts added attentional demands on LN. Despite fairly natural turn-taking behavior, a stark reduction in his attempts to claim a turn suggest that he may be overburdened by the added attentional demands. In fact, in the one turn he finally negotiates, he asks his wife to make his point for him. LN did miss two turn shifts between tester JL and his wife, but these constituted a side conversation rather than formal turn taking in the course of the main conversation. In the broader context of LN's conversational interaction, we were able to observe instances of all the initiation, continuation and shift regulators noted by Baker (1977). Among these regulators, were simultaneous turns which occur when a conversant claims a turn and another fails to yield. These occurred with LN, the longest overlap lasting a full five seconds. Furthermore, simultaneous sign also occurred in the context of back channel behavior. In contrast with Baker's reports that these overlapping stretches tend to be much longer in ASL (average length 1.5 seconds with the longest stretch she documented lasting 4.3 seconds (Baker, 1977: 216)) than in spoken language conversations (mean duration less than .5 second (Jaffe & Feldstein, 1970)), factoring out simultaneous turns, LN's simultaneous signing was considerably shorter and less frequent than observed in standard ASL conversations. One of LN's preferred forms of feedback is questioning. For example, when the tester describes the make up of her family he responds with the equivalent of "Oh, so you have one child?" These tend to get coded as turn initiations, but if they were coded with back channel signals, they would up the length of simultaneous signing during back channeling considerably. 3

The only exception to this is when the time code at the beginning of a row appears in parentheses (see Table 2.). This indicates a shift of eyegaze or body orientation that, while significant for grammatical agreement or some other function that resembles the gesturing involved in turn taking, does not constitute a conversational turn.

Compensation for pragmatic deficits in signers with PD Dyad 02:40 2:41 2:58 2:59 3:03 3:04 3:11 3:12 3:13 3:21 3:23 3:32 3:35 3:36 3:40 Multiple 00:00 00:02 00:27 1:24 1:25 2:09 2:14 2:18 2:19 2:24 2:30

Orient.

To

Attention touch

Orient.

JL -------------To

1

1 1 1 1

Attention touch

Turn yield

Back channel

1

1

1

1

1

1

1

1 1

1

1

1

1

Simultaneous signing

With

.5

JL

2.5

JL

Simultaneous signing

With

5.0

JL

1 1 1 1 1 1 1 1 Missed shift

Turn claim

Turn yield

Back channel

1 1

1 1

wife

1

JK JL*

1 1

wife*

1

JL* wife*

1

Turn claim

JL JK wife*

1 1

Missed shift

145

JK

1 1

1 1

Table 1. Conversational turntaking in mild Parkinsonian subject LN

Moderate to severe Parkinsonian (JH) In contrast with LN, whose symptoms of PD are milder and more confined to the planning, execution, and sequencing of motor plans than to attentional problems, JH's conversational performance strikes one as more disrupted across the board. Still we see throughout his conversations evidence of his mastery of the rules for turn-taking in ASL. Yet his turntaking is still pragmatically off and jarring to the interlocutor. Table 2. presents the results of coding several minutes of JH's conversational interactions in both a dyad and a larger group. In the dyadic conversation, the lack of back channel signaling on JH's part is notable. He strongly signals a turn yield with very fixed eyegaze on his addressee.but he offers little in the way of nodding, copying of final signs in an utterance or other signals that he is following along with the thread of the conversation. Yet, his contributions to the conversation are timely and appropriate, as well as numerous, demonstrating that he is indeed following just fine. He just fails to give much feedback. When he does it is typically a very reduced headnod or a smile and little more. Most of JH's conversations, even those with more than two participants, tend to become dyadic because once he focuses upon one interlocutor, he has trouble disengaging and shifting to another one. There is an effect of where relative to JH a given interlocutor is positioned. In the conversational stretches coded in Table 2., VJ is on JH's left, while JK is across the room basically right of center. This attentional difficulty is a form of

146

J. Kegl and H. Poizner

neglect and in JH's case seems to be more strongly manifested with individuals situated to his left. Dyad 00:00 00:03 00:07 00:10 00:12 00:14 00:15 00:18 00:20 00:25 00:26 00:30 00:32 00:35 00:36 01:09 Multiple 07:18 07:20 07:30 07:32 07:57 08:04 08:08 08:09 08:14 08:23 08:32 08:39 08:44 08:49 08:52 08:56 08:57 09:06 09:42 09:47 09:57 09:58 09:59 10:23 11:07 (11:19) (11:20) 11:36 11:40 11:44

Attention touch

1 1 1

Attention touch

Orient.

To

1

VJ --VJ

avert 1 avert avert 1 avert 1 1 1 avert 1

Orient.

avert 1

1 1

1 avert avert 1 1

1

1 1 1 1

1

1

Missed shift

Turn claim 1

Turn yield

Back channel

Simult. signing

With

Simult. signing

With

.2

VJ

1 1 1 1 1

VJ

1

smile

1 1 1 1

VJ JK VJ 1 VJ ---To JK ---JK --------VJ

1 1 1

Missed shift

1 Turn claim

Turn yield 1

Back channel 1

1

1

1 1 1 1 1 1 1

1

1 1 1 1 1 1 1

VJ --JK -VJ JK -index(VJ) JK --VJ

1

1

attempt 1 1

attempt 1 1

1

1

1

1 1 1 1 1

Table 2. Conversational turntaking in moderate-to-severe Parkinsonian subject JH In observing JH's conversations, we found that there was a higher than usual occurrence of physical touches on the part of interlocutors to shift his attention to them as they took the

Compensation for pragmatic deficits in signers with PD

147

floor, however, most of these touchings came from VJ, the interlocutor on JH's left. She used them to disengage attention from JK as well as to claim a turn from JH himself. When JH claimed a turn from VJ, he averted his gaze from her (as is typical in turn initiation), but once averted, he did not easily disengage or monitor VJ for signals that she wanted to re-initiate a turn. In addition, JH exhibited very little simultaneous signing and was frequently unsuccessful in claiming a turn away from another interlocutor. He would partially raise his hands from a full-rest position, but not forcefully enough to take the floor. In contrast with LN, much of the communication addressed to JH was in the form of questions that required answers. This means that he was frequently drawn into claiming a turn, rather than left to initiate a turn on his own.

Severe Parkinsonian (JW) JW constitutes one of our most interesting cases. Being in the late stages of PD, JW has a very masklike facial expression and his signing is completely distalized to the wrist and fingers. In fact, in addition to realizing most of the movement of signs with his fingers alone, even using the fingers as articulators, he restricts their movement to the most minimal of bending or separating possible. His signing gives the impression of a person very sporadically and very slowly bending the fingers while keeping them in an outstretched plane. The major movement change in the hand is at the metacarpal joints where the fingers meet the palm of the hand. Table 3. presents a glimpse of JW's conversational interaction. Basically, he arranges himself in a single postural configuration where his head is til ted to the side to allow him to look at an interlocutor and he remains in that position throughout the conversation. We don't see any of the characteristic changes in head orientation or eyegaze that signal claiming a turn. However, it is also the case that JW rarely needs to negotiate for a turn. In general his interlocutors speak to him in questions and wait sometimes for extremely long periods of time for the slightest of answers. Like JH, JW has difficulty disengaging attention. For example, in the coded conversation among multiple conversants in Table 3., JW is unable to disengage attention from a picture of himself in an army uniform that is serving to illustrate his story about faking being hearing in order to join the army, which he did. Over the course of 8-9 turns by his wife and the tester, with repeated touching to gain his attention, they were unable to draw his attention away from the picture. They eventually combined repeated touching of JW with moving their hands right onto the picture to capture his gaze. It is also notable that both the tester and his wife positioned themselves very distinctly on JW's direct right. In fact, they were in a line with him (his wife almost behind him) rather then facing him, which allowed them to physically place their hands/signs between JW and the picture he was fixated upon. A turn initiation by JW was generally not signaled by anything more than beginning to sign or fingerspell a response. Signals of turn claiming, turn yielding and back channel information were totally absent from his signing, but interlocutors were highly attentive to the slightest indications that he was beginning to participate. Simultaneous signing on JW's part was nonexistent. In fact the lag time for a response could be several seconds. However, his interlocutors would frequently sign simultaneously with JW copying and clarifying what he was signing, decoding for themselves and each other the minimal, yet linguistically decodable, articulation of his signs.

148

J. Kegl and H. Poizner Dyad

Attention touch

00:00 00:04 00:08 00:14 00:21 00:28 00:39

To

Missed shift

1

Attention touch 1

1 1 1 1 avert 1 Orient.

no claim 1

17:55

--

wife

17:56

--

JC

JC pix

JC

18:14 18:17 18:23 18:26 18:38

1

1

touch pix touch pix

1 1

18:40 18:44

1

18:51 18:55 19:03

1 1 1

1 1

19:05

1

19:10

JC

Simult. signing

With

1 JC To

JC

1 1

.5 1 1 1 1

--

1

JC

no change 1

17:53

18:06 18:12 18:13

.5

no claim

JC/wife

1

JC

no change

JC ---filmer JC filmer JC

--

17:59 18:02

.5

9

JC

touch pix

With

?

--

17:47

Simult. signing

no change

pix

17:46

Back channel

no change

Missed shift wife

17:45

Turn yield

1

— --

eyes up left avert eyes 1

Turn claim

no change

JC ----

00:40

00:41 00:51 00:56 01:04 01:08 01:12 01:13 01:15 01:16 01:33 Multiple

Orient.

-JC hand

wife

Turn claim

wife

--

JC/wife

Back channel

no response no response no response no response no response no response 1 no response 1 1 no response 1

wife

JC --pix pix

1 Turn yield

1 1 1

wife JC JC pix

wife

1

pix

wife

1

JC

no response no response no response 1 1 no response no response 1

Table 3. Conversational turntaking in severe Parkinsonian subject JW

Compensation for pragmatic deficits in signers with PD

149

Conversational compensation by signing interlocutors

A perusal of Baker's summary of conversational regulators and turn-taking in ASL reveals a complex and orchestrated interweaving of manual and non-manual signals on the part of both the speaker and addressee(s). All participants in the conversation are constantly giving and monitoring eyegaze as well as gestural cues that confirm their status in the conversational process. Participating in a conversation is clearly a multi-tasking operation — something individuals with Parkinson's disease are very bad at. The more components of conversational monitoring were required, the more compromised conversation became— even for mildly affected LN. We need to recognize, however, that conversation is not always a competition for control of the floor. Sometimes other interlocutors expressly seek to include the Parkinsonian participant in a conversation. For example, the three signers with PD studied here were extremely interesting personalities. LN was very active in the Deaf community with lots of stories to tell. JH was a pop artist, a competitive diver and a professional wrestler. He was doing motorcycle acrobatics like standing on his head on the handlebars of a moving motorcycle controlled only by himself at age 65. JW was actually inducted into the army and almost sent overseas in World War II when it was finally discovered that he could not hear and had been fooling everyone. The testers were mesmerized by the stories and anecdotes these subjects shared with them. In such cases, since the burden of making conversational interaction work can no longer be shared equally, the lion's share of responsibility for conversational success falls to the participants without PD. And, they eagerly take it on. In the remainder of this paper, we focus upon the burden placed on the interlocutor when conversing with a Parkinsonian signer, the compensatory strategies recruited to maximize comprehension and maintain the flow of information, and what these compensations reveal about the nature of the deficits associated with PD. We examine how interlocutors compensate for three characteristic deficits associated with Parkinsonian signing: 1. the flattening of distinctions in sign articulation; 2. problems initiating signed utterances; and 3. problems related to disengaging attention. Adjusting the target of foveal vision Sacrificing perceptual saliency exacts a steep price in terms of the demands it places on the addressee. In interactions with LN and JH, we saw interlocutors simply pay closer attention to their signs and reiterate sentences for clarification. But, with JW the processes of easing articulation were taken to their maximal extent. A hearing speaker at this stage would probably be inaudible and unintelligible and certainly would be beyond the technology we have for capturing and analyzing speech. But, JW was still understandablealbeit with great effort. Shifting of attention to multiple interlocutors was no longer a possibility for JW, although we do see one instance of his shifting of gaze to monitor the person behind the camera for possible input to the conversation. Instead, one of the interlocutors (or each in turn) takes on the task of receiving and decoding the communication from JW and transmitting it to the others. We saw this clearly in the interaction between tester JC and JW's wife. But given the maximal reduction of JW's signs, simple monitoring was not enough. The "designated decoder" unconsciously adjusted his position to maximize the perceptibility of JW's signing.

150

J. Kegl and H. Poizner

In ASL, signs are distributed in the signing space which ranges from the top of the head to a semi-circle about waist high, extending a bent elbow's length in front of the signer such that signs with more marked handshapes and movements occur near the chin and neck in the area of the interlocutor's foveal vision (Siple, 1978). Fixating at this point also allows facial expressions (crucial to the grammar of ASL) to fall in the area of highest visual acuity. Typically, the addressee focuses at about the signer's chin and uses peripheral vision, rather than adjusting fixation, to process signs (mandatorily signs with unmarked movements and handshapes) that occur in the periphery. But Parkinsonian signing is typically lowered and reduced such that all signs are in essence marked. Furthermore, facial expression is severely attenuated. While facial masking is characteristic of PD, we see the preservation of minimal eyegaze and grammatical expressions maintained as long as possible. But JW's interlocutors converged on a solution. Over the course of the conversation, JW's tester can be observed to shift his position gradually downward getting JW's waist height hands in the area of foveal vision, but at the expense of facial information. Eventually, we see the full solution. By the end of the conversation, the tester has unknowingly maneuvered himself such that he is kneeling on the floor with his head tilting and looking up at JW's face through JW's hands. The tester moved to the one vantage point where JW's hands and face could both be processed in the interlocutor's area of foveal vision. The tester reestablished the balance between ease of articulation and ease of perception by adjusting his position. Increasing the use of shift regulators Although less so in the case of LN, all the testers maximized the participation of signers with PD in conversations by increasing the use of shift regulators. Their communication became primarily questions and restatements. Testers also elaborated on statements by the signers with PD, fleshing out the story and offering additional opportunities for the signers with PD to chime in with additional information. More and more turns on the part of the tester and family members became questions demanding a response, and that response was patiently waited for, although sometimes prodded out with various means of contacting or starting the signer along his way. Becoming the PD signer's attentional system Finally, in multi-participant conversation, the interlocutors as a group, became the PD signer's attention system, systematically tapping him or working together to draw his attention where needed. In cases like that of JH, where the attention system is asymmetrically affected, we see this asymmetry reflected in the behavior of the interlocutors. VJ (on JH's left), for example, begins to consistently tap JH before every conversational turn, where the other participants continue to use visual cues to reorient him. JW's interlocutors literally touch him and interrupt his line of sight to disengage his attention from one target and reorient it to another.

Compensation for pragmatic deficits in signers with PD

151

Conclusion In contrast with the individualized nature of sentence production or interpretation, the pragmatic aspects of conversation are a mutually negotiated operation. When one participant is not fully up to the task, the situation demands compensation from the remaining members of the conversation. The rules of pragmatics are such that errors cannot just be ignored or corrected covertly in the interlocutor's interpretation. When a signer with PD fails to signal that a turn is being yielded, the person claiming the turn is actually thwarted from continuing. Something about this group endeavor leads pragmatic violations to be more jarring and more salient than a simple slip of the tongue or agreement error that might be encountered from a signer with aphasia. Observing the behavior of persons engaging in conversation with signers who have Parkinson's disease corroborates in very tangible ways what we know to be the attentional as well as sensori-motor planning and execution deficits associated with its progression. Their complex repositionings to maximize perception of the reduced and laxed signing demonstrate clearly the consequences of the PD signer's prioritization of ease of articulation. Their increased use of shift regulators, patient attentiveness to the PD signers' every move, and taking over in many cases responsibility for clarifying or amplifying the PD signers' minimally signaled turntaking claims registers sensitivity to their deficits in quickly planning and initiating signed motor acts. And finally, we see single and multiple interlocutors actually become (via their tapping and attention getting devices) the attentional system of conversational participants with PD, registering the impact that neglect plays in Parkinson's disease. All of these accommodations occur in the absence of any knowledge of the nature of the disease, but rather are the consequence of repairing breakdowns in conversational turntaking to keep the process going. As is the case with their linguistic capacities, we do not see any pragmatic deficit per se in signers with PD. Rather what we see are extralinguistic factors of sensori-motor planning and execution as well as attention that impede the ability of these rather eager and interesting conversationalists to successfully employ the turntaking devices they know to hold their own in conversations, especially those with more than two participants. The elaborate compensations made by interlocutors in conversations with signers who have PD highlights for us the important role pragmatics plays in our communicative lives. Acknowledgements—The authors are grateful to numerous individuals who participated in the collection and transcription of these data including Joanne Lauser, Jimmy Challis, Vicki Joy Sullivan, Patricia Trowbridge, and Toni Fuller. We are also grateful to the participants in the Aphasia Committee meeting (IALP) in Montreal (July, 1997) for insightful questions and constructive comments that contributed greatly to our thoughts on this subject. Portions of this study were partially funded by National Science Foundation grants #SBR-9410562 and #IRI-9528985 to Boston University and Rutgers University, National Science Foundation grants #BNS-9000407 and #SBR-9513762 and grant #DC01656 from the National Institute on Deafness and Other Communication Disorders of the National Institutes of Health to Rutgers University. Correspondence should be addressed to Judy Kegl, Center for Molecular and Behavioral Neuroscience, Rutgers University, 197 University Avenue, Newark, NJ 07102.

References Aarons, D., Bahan, B., Kegl, J., & Neidle, C. (1992). Clausal structure and a tier for grammatical marking in American Sign Language. Nordic Journal of Linguistics, 15, 103-142. Bahan, B. (1996). Non-manual realization of agreement in American Sign Language. Doctoral Dissertation, Boston University. Benecke, R., Rothwell, J. C, Dick, J. P. R., Day, B. L, & Marsden, C. D. (1987). Disturbance of sequential movements in patients with Parkinson's Disease. Brain, 110, 361-379.

152

J. Kegl and H. Poizner

Brentari, D., & Poizner, H. (1994). A phonological analysis of a Deaf Parkinsonian signer. Language and Cognitive Processes, 9, 69-100. Brentari, D., Poizner, H., & Kegl, J. (1995). Aphasia and Parkinsonian signing: Differences in phonological disruption, Brain and Language, 48, 69-105. Duncan, S. (1972). Some signals and rules for taking turns in conversations. Journal of Personality and Social Psychology, 23, 283-292. Hallett, M., & Khoshbin, S. (1980). A physiological mechanism of Bradykinesia. Brain, 103,301-314. Harrington, D. L., & Haaland, K. Y. (1991). Sequencing in Parkinson's Disease. Brain, 114,99-115. Jaffe, J., & Feldstein, S. (1970). Rhythms of dialogue. New York: Academic Press. Kegl, J., Cohen, H., & Poizner, H. (In press). Articulatory consequences of Parkinson's Disease: Perspectives from two modalities. Brain and Cognition. Kegl, J., & Poizner, H. (1991). The interplay between linguistic and spatial processing in a right-lesioned signer. Journal of Clinical and Experimental Neuropsychology, 13,38-39. Kegl, J., & Poizner, H. (1993). Preservation of syntactic distinctions in a moderate to severe Parkinsonian signer. Presented at the Sixth Annual CUNY Sentence Processing Conference. University of Massachusetts at Amherst, March 18-20. Kendon, A. (1967). Some functions of gaze direction in social interaction. Acta Psychologica, 26, 22-63. Loew, R., Kegl, J., & Poizner, H. (1995). Flattening of distinctions in a Parkinsonian signer. Aphasiology, 9, 381-396. Loew, R., Kegl, J., & Poizner, H. (1997). Fractionation of the components of roleplay in a right-lesioned signer. Aphasiology, 11, 263-281. Margolin, D. I., & Wing, A. M. (1983). Agraphia and micrographia: Clinical manifestations of motor programming and performance disorders. Acta Psychologica, 54, 263-283. Marsden, C. D. (1982). The mysterious motor function of the basal ganglia. Neurology, 32, 514-539. Marsden, C. D. (1987). What do the basal ganglia tell premotor areas? In Motor areas of the cerebral cortex (Ciba Foundation Symposium 132). Chichester: Wiley. Marsden, C. D., & Obeso, J. A. (1994). The functions of the basal ganglia and the paradox of stereotaxic surgery in Parkinson's Disease. Brain, 117, 877-897. Poizner, H., & Kegl., J. (1992). The neural basis of language and motor behavior: Perspectives from American Sign Language. Aphasiology, 6, 219-256. Poizner, H., & Kegl, J. (1993). Neural disorders of the linguistic use of space and movement. In P. Tallal, A. M. Galaburda, R. Llinas, and C. von Euler (Eds.), Temporal information processing in the nervous system: Special reference to dyslexia and dysphasia (vol. 682, pp. 192-213). Annals of the New York Academy of Sciences. New York: The New York Academy of Sciences. Siple, P. (1978). Visual constraints for sign language communication. Sign Language Studies, 19, 95-110. Weiner, M., & Devoe, S. (1974). Regulators, channels, and communication disruption, Research Proposal, Clark University. Weiss, P., Stelmach, G. E., & Hefter, H. (1997). Programming of a movement sequence in Parkinson's Disease. Brain, 120, 91-102.

Pergamon

©

1998

J. Neurolinguistics, Vol. 11, Nos 1-2, p. 153-177, 1998 Published by Elsevier Science Ltd. All rights reserved Printed in Great Britain 0911-6044/98 $19.00 + 0.00

PII: S0911-6044(98)00011-6

Pragmatics in frontal lobe dementia and primary progressive aphasia Joseph B. Orange*, Andrew Kertesz1 and Jennifer Peacock* 'School of Communication Sciences and Disorders, University of Western Ontario; 'Department of Clinical Neurological Sciences, Lawson Research Institute, St. Joseph's Health Centre

Abstract—The purpose of this study was to document the pragmatic performance of subjects with frontal lobe dementia (FLD), non-fluent primary progressive aphasia (PPA), and fluent PPA using topic-directed conversationbased interviews. Unique profiles of pragmatic performance emerged for the three subject groups using objective measures of discourse, pragmatics, and scores from the Profile of Communicative Appropriateness (PCA) (Penn, 1985). The differences in performance provide a clearer understanding of the selective influences of frontal and frontotemporal pathology on the pragmatics of patients with FLD and PPA.

Introduction Frontal lobe dementia (FLD) and primary progressive aphasia (PPA) are distinct, yet interrelated clinical syndromes, that are characterized by progressive declines in selected spheres of cognitive and language abilities. In the case of FLD, disturbances in cognition, behaviour, and emotion are prominent early features (Brun, Mann, Englund, Neary, et al., 1994). These disturbances include, but are not limited to, memory deficits, poor judgement and problem solving abilities, lack of insight, inflexibility of thought and behaviour, aggressiveness, impulsivity, denial of problems, changes in personality, decreased initiative, social withdrawal, flattened affect, disinhibition, and emotional lability (Brun, Mann, Englund, Neary et al. 1994; Gustafson, 1993; Snowden, Neary, & Mann, 1996). FLD has been associated with motor neuron diseases (Caselli, Windebank, Petersen, Komori, Parisi, Okasaki et al., 1993; Neary, Snowden, Mann, Northen, Goulding, & Mcdermott, 1990; Strong, Grace, Orange, & Leeper, 1996). First described by Mesulam (1982), PPA is a clinical syndrome characterized by progressive deterioration in language with relative preservation of cognitive abilities such as memory and attention, personality, and insightfulness within the first few years following onset of language problems (Karbe, Kertesz, & Polk, 1993; Weintraub, Rubin, & Mesulam, 1990). Problems with cognition may emerge only after several years of disease progression, sometimes upwards of five to seven years following the onset of the language disturbances (Green, Morris, Sandson, McKeel, & Miller, 1990; Kesler, Artzy, Yaretzky, & Kott, 1995). One of the unifying features of FLD and PPA is the presence of language disturbances. Descriptions of the language characteristics of patients with FLD and patients with PPA have appeared in the literature over the past decade. General patterns of the speech and language characteristics of patients with FLD have been described in small groups of patients. Dysarthric-like speech output and dysprodic disturbances are apparent, with relatively preserved language until the late stage of the disease where echolalia, perseveration, and mutism occur (Neary, 1990). Reduced spoken output, or logopenia, is one of the prominent features in patients with FLD, and is

154

J.B. Orange, A. Kertesz and J. Peacock

reflected in lower verbal fluency scores and reduced conversational initiation (Barber, Snowden, & Craufurd, 1995; Gustafson, 1993; Neary, 1990; Neary, Snowden, Mann, Northen, Goulding, & Mcdermott, 1990). Increased talkativeness and circumlocution also have been described in a select of group of patients with FLD (Gustafson, 1993; Neary, 1990). Patients in the early stage of FLD may show reduced and aspontaneous verbal output, few elaborations of information, and stereotyped and repetitive utterances (Gustafson, 1993; Neary et al., 1990). Others may confabulate and exhibit anomia in the form of semantic paraphasic errors, although these features have been observed less frequently (Johanson & Hagberg, 1989; Neary et al., 1990). With disease progression, language may be digressive and socially inappropriate, reflecting the advancing disinhibition and losses in social awareness (Brun, Mann, Englund, Neary et al., 1994). The language of patients with PPA varies considerably, but primarily includes features observed in non-fluent type aphasias (Karbe, Kertesz, & Polk, 1993; Kempler, Metier, Riege, Jackson, Benson, & Hanson, 1990; Kertesz, Hudson, Mackenzie, & Munoz, 1994). The language of the non-fluent subgroup of patients PPA is characterized by telegraphic utterances, phonemic paraphasias, reduced spoken output, limited phrase length, reduced grammatical complexity, and relatively preserved auditory and reading comprehension (Duffy & Petersen, 1992; Karbe, Kertesz, & Polk, 1993; Kertesz, Hudson, Mackenzie, & Munoz, 1994; Thompson, Ballard, Tail, Weintraub, & Mesulam, 1997; Weintraub, Rubin, & Mesulam, 1990). Anomia may be present and prominent in some cases (Mesulam, 1982; Weintraub et al., 1990), but is usually less severe than the limited spoken output (Duffy & Petersen, 1992). Apraxia of speech also may be evident in patients with more advancedprogression (Karbe, Kertesz, & Polk, 1993). Recent evidence suggests that two other subgroups of PPA patients also exist. One group exhibits fluent-like language and impaired comprehension (Thompson et al., 1997), while the other exhibits dysfluency and impaired comprehension (Snowden, Neary, Mann, Goulding, & Testa, 1992). Thompson et al. (1997) examined the grammatical complexity of four subjects with non-fluent-like PPA from discourse samples collected over a 3-7 year period. They found both non-fluent and fluent profiles of language in their sample. The fluent-like PPA subgroup primarily show word-retrieval difficulties and comprehension problems with relatively preservedmorphosyntactic structures. Snowden and her colleagues (1992) noted that the third subgroup of non-fluent PPA patients show changes in personality and social interaction skills similar to alterations observed in patients with bilateral frontal lobe deterioration. The recent studies of the language of patients with FLD and PPA have described a wide range of linguistic abilities but few have examined their pragmatic performance in spontaneous spoken tasks. The frontal and anterior frontotemporal cortical sites of neuropathology in FLD and PPA (and related subgroups) suggest that disturbances in pragmatics may be prominent. It is clear that pre-frontal symptoms such as disinhibition, distractibility, impersistence, attention problems, among others, may contribute to disruptions in pragmatics and discourse. Horner (1985) noted over a decade ago that analysis of the pragmatics of patients with progressive language deterioration would likely identify the presence of repetition of ideas, stereotyped phrases, self-corrections and revisions, off-topic utterances, and egocentric comments. The purpose of this study was to document the pragmatic performance of subjects with FLD, non-fluent PPA, and fluent PPA. Our goal was to describe unique profiles which would distinguish the three groups. It was anticipated that any observed differences among the FLD and the two PPA subgroups would advance our understanding of the theoretical perspectives of pragmatic performance in the presence or absence of cognitive impairments, help clarify the influence

Pragmatics in frontal lobe dementia and PPA

155

of progressive frontal and frontotemporal pathology on pragmatic abilities, and provide clinical intervention options for patients, family, and professional caregivers. Method Subjects Fourteen subjects participated in the study. All were native English speakers, right handed, and had a minimum of Grade 8 education. There were 7 females and 7 males. Ages ranged from 41 to 81 (see Table 1). None had a history of psychiatric, medical, or neurological conditions prior to their evaluations and diagnosis by a neurologist (AK). Comprehensive medical, neurological, neuroimaging, psychiatric, neuropsychological, and language assessments revealed the presence of FLD and PPA congruent with currently accepted clinical diagnostic criteria (Brun, Mann, Englund, Neary, et al., 1994; Mesulam, 1982). Extensive histories were obtained from patients or family members to document the presenting complaints and to compile an historical profile of problems. Three subjects were diagnosed with FLD, five subjects were diagnosed with non-fluent PPA, and six with fluent PPA by a neurologist (AK).

Cognitive and language testing Neuropsychological testing was conducted to measure performance across a number of cognitive systems (e.g., memory, attention, nonverbal intelligence, etc.). A core battery of neuropsychological tests was administered to the majority of subjects while selected tests and subtests were given to several of the subjects. The core battery of neuropsychological tests included the Mattis Dementia Rating Scale (Mattis, 1976), Wechsler Adult Intelligence Scale - Revised (Wechsler, 1981), Raven's Coloured Progressive Matrices (Raven, 1947), Wechsler Memory Scale - Revised (Wechsler, 1987), and the ReyOsterrieth Complex Figure (Osterrieth, 1944; Rey, 1941). The other frequently administered tests included all of or selected subtests from the California Verbal Learning Test (Delis, Kromer, Kaplan, & Obler, 1987), Benton Visual Retention Test (Benton, 1974), Facial Recognition Test (Benton & Van Allen, 1972), and the Hooper Visual Organization Test (Hooper, 1958). FLD subjects' performances on the neuropsychological battery of tests and subtests showed significant problems with memory, nonverbal intelligence, and frontal lobe-based processes including judgement, insightfulness, and inhibition. The non-fluent and fluent PPA subjects' scores on the neuropsychological tests generally ranged from low-normal to low-average and did not indicate the presence of obvious cognitive impairments. The Western Aphasia Battery (WAB, Kertesz, 1982) was administered to all but one of the subjects (i.e., FLD subject CPY). The most prominent disturbance among all subjects on the WAB was their anomia. WAB Aphasia Quotient scores for all subjects who completed it, except one (Subject CP), were below the cut-off criterion of normal language performance and ranged from 7.2 to 91.4 (see Table 1).

Subject Groups

Features

Non-fluent PPA (n = 5)

FLD(n=3)

Sex Age (years) Education (years) Handedness WABAQ MDRS

Fluent PPA (n = 6)

EB

CPY

CP

RH

VH

FS

AS

BV

VB

JH

IM

RR

KS

JS

M 65 12 R 80.2 100

F 41 19 R NA NA

F 60 14 R 97.2 114

F 78 8 R 38.0 untestable

F 77 N/A R 16.0 untestable

F 55 17 R 7.2 untestable

M 65 10 R 16.7 66

F 65 11 R 77.7 94

F 75 22 R 82.6 109

M 81 N/A R 81.8 82

M 76 15 R 73.9 83

M 77 12 R 65.8 89

M 69 19 R 90.0 129

M 58 16 R 91.4 N/A

Note: N/A = Not available WAB AQ = Western Aphasia Battery Aphasia Quotient (Kertesz, 1982); total possible score is 100 MDRS = Mattis Dementia Rating Scale (Mattis, 1976); total possible score is 144; subjects untestable because of severe language problems

Table 1. Subject demographic information

Pragmatics in frontal lobe dementia and PPA

157

Procedure To obtain an extended sample of spontaneous language, all subjects were invited to participate in a topic-directed interview as part of their comprehensive neuropsychological and language assessment. A single, trained examiner acted as the partner in the topicdirected interviews for all fourteen subjects. Topic-directed interviews have been used previously for subjects with dementia, and have been shown to be useful in generating extended samples of spoken discourse suitable for analyses of discourse and pragmatic features (Garcia & Joanette, 1994; Illes, 1989; Mentis, Whittaker, & Gramigna, 1995; Ripich & Terrell, 1988). Topic-directed interviews were conducted in a quiet, distraction-free room. Interviews were video recorded. The pre-selected topics were introduced by the examiner either during the Spontaneous Speech section of the WAB or separately during the assessment using the open-endedrequest, "Tell me about ". The following five topics were used: (a) your family, (b) your health right now, (c) what you do each day, (d) where you were born and raised, and (e) the jobs you had or the work that you did. The topics were introduced in the same order for all subjects. The topics are similar to those used by Ripich and Terrell (1988) in their study of the discourse of subjects with Alzheimer's disease, and those used by Illes (1989) in her study of the discourse of subjects with various forms of cortical and subcortical dementias. During the interview, the examiner was limited to using a single, simple prompt for extending talk on a topic, and in using only back-channelling responses that indicated her comprehension or agreement (e.g., "Yes", "Umhum", facial expressions and head nods, etc.). When subjects made a concluding statement, stopped speaking for fifteen seconds and where there were no indications of word searching behaviours (such as upward/downward eye gazing, prolongations of sounds, or fillers such as "Uhm", "Uh", "Like", "Just a minute", etc.), or otherwise signalled the temporary completion of the monologue (used downward intonation, made eye contact and paused for turn shift, gestured for turn shift, etc.), the interviewer made the single prompt "Tell me more about .". If subjects indicated (verbally or nonverbally) that there was nothing more to be said on that topic, the interviewer introduced the next topic until all five had been discussed. If subjects shifted topic during their extended monologue, the interviewer did not interrupt but waited until subjects had concluded their remarks. No time limits were imposed on any topic or the interview task. All samples were orthographically transcribed and segmented into turns and utterances according to the criteria established in the Shewan Spontaneous Language Analysis system (Shewan, 1988). The segmentation of utterances was based on morphosyntactic, semantic, and prosodic criteria. Utterances in the samples were coded for several discourse and pragmatic features. These are listed in Tables 2 and 3. Operational definitions for the discourse and conversational topic management measures are presented in Appendix A. The measures were selected based on findings from previous work on the discourse and pragmatic performances of subjects with various forms of dementia and aphasia which showed their effectiveness in differentiating the performance of subjects (Illes, 1989; Nicholas & Brookshire, 1993); Ripich & Terrell, 1988; Tomoeda & Bayles, 1993; Ulatowska, Allard, Reyes, Ford, & Chapman, 1992. In addition to identifying the discourse and pragmatic measures outlined in Appendix A, the transcript of each subject was scored using the definitions and guidelines of the Profile of Communicative Appropriateness (PCA) (Perm, 1985; Penn, 1988). The items

158

J.B. Orange, A. Kertesz and J. Peacock

comprising the PCA are outlined in Appendix B. All items from the PCA, with the exception of Control of Direct Speech in the category Sociolinguistic Sensitivity, were scored by a trained rater (author JP). This single item was not scored and was eliminated from the analyses of PCA ratings as its definition was non-specific and could not be coded uniformly across subjects. The PCA was selected as an outcome measure because it has been used successfully in previous studies to document the pragmatic performances of individuals with various forms of dementia (Penn, Sonnenberg, & Schnaier, 1988), and for profiling pragmatics in individuals with aphasia (Penn, 1988). Agreement studies To assess inter-rater agreement for the scoring and the rating of the outcome measures used in this study, point-by-point percent agreement scores were calculated. Two trained raters re-scoredthe entire transcripts from two randomly selected subjects (i.e., 14% of the data base). Inter-rater agreement for utterance segmentation was 95.5%. The mean percent agreement for coding the discourse and pragmatic outcome measures was 94%. Agreement scores for only three outcome measures were below 80%. These were categories (a) number of could not evaluate utterances/topic (75%), (b) percent of could not evaluate utterances/topic (75%) (this category occurred only for the non-fluent PPA subjects), and (c) percent of utterances in side sequences - external (60%). The most likely reason for the low agreement on these measures was their infrequent occurrence in the transcripts of all subjects. The mean percent inter-rater agreement for PCA ratings was 91.5%. Data analysis The small number of subjects who participated in this preliminary study warranted against using parametric and non-parametric statistical analyses for group comparisons. We undertook a cautious approach in interpreting absolute differences in scores among the groups. For the purposes of data analysis, percentages, rates, and time-based scores were calculated. Time-based measures were calculated using the speaking time of the samples (i.e., total time of the sample minus all inter- and intra-pauses longer than 5 seconds). Means and standard deviations were calculated for the topic management and PCA outcome measures. These values were used as a basis for descriptive comparisons. Results Data derived from the samples obtained in the topic directed interviews for the FLD subjects, the non-fluent PPA subjects, and the fluent PPA subjects are presented in Tables 2 through 6. Individual data points, summary means, and standard deviations are shown.

Subjects Discourse measures

EB

FLD (n = 3) CP CPY

# words

61

813

90

MLU

5.6

14.7

4.7

# utterances

11

61

21

% complete utt.

100

100

95

% incomp. utt.

0

0

5

% stereotype

0

4.9

0

% overt

0

0

0

#self correct/utt.

0.1

0

0

% pronoun

103

14.8

13.3

% pwa

0.1

0.8

0

% demonstrative

20

11.6

5

# words/min.

38.9

157.9 41.3

# comp. utt./min.

6.4

11.8

9.2

# incomp. utt. /min.

0

0

0.5

# reps./utterance

0.1

0

0

X

(SD) 321.3 (426) 8.3 (5.5) 31 (26.5) 98.3 (2.6) 1.7 (2.9) 1.6 (2.8) 0 (0) 0 (0.1) 12.8 (2.3) 0.3 (0.4) 12.2 (7.5) 79.4 (68.0) 9.1 (2.7) 0.2 (0.3) 0 (0.1)

RH

Non-fluent PPA (n = 5) AS PS VH

BV

108

115

22

18

211

5.6

5

1.2

1.3

5

21

26

18

16

42

81.0

96.2

88.9

100

78.6

19.0

3.8

11.1

0

21.4

19.0

46.2

44.4

6.3

9.5

0

0

0

0

11.9

0.2

0

0.2

0

0.2

4.9

18.3

0

0

18

29.2

66.7

0

0

10.5

4.2

66.7

0

0

26.7

52.9

53

16.3

13.5

65.9

6.4

10.2

9.6

4.5

7.5

2.5

0.9

2.2

0.8

5.3

0.6

0.6

0.5

0.6

0.5

X (SD) 65.8 (52.9) 3.62 (2.2) 24.6 (10.4) 88.9 (9.3) 11.1 (9.3) 25.1 (19.1) 2.4 (5.3) 0.1 (0.1) 8.2 (9.3) 21.3 (28.1) 19.5 (28.6) 40.3 (23.8) 7.6 (2.3) 2.3 (1.8) 0.6 (0.1)

VB

JH

Fluent PPA (n = 6) IM KS RR

JS

741

537

176

1451

572

124

9.4

9.2

4.5

7.9

10.3

5.4

92

58

40

176

78

23

91.3

93.1

62.5

83.5

98.7

100

8.7

6.9

37.5

16.5

1.3

0

2.2

5.2

17.5

25

1.1

0

4.3

1.7

2.5

2.8

1.7

0

0.4

0

0.2

0.3

0.1

0.5

19.3

18.6

17

21.8

16.8

14.5

4.2

6

6.7

27.8

1

0

20.2

25

3.3

16.5

13.5

11.1

82.2

120.1

51

96.3

66.8

40.8

8.7

11.6

7.3

9.2

7.1

6.9

1.4

1.3

3.8

2.3

0.6

0.3

0.5

0.2

0.5

0.5

0.2

0.9

Table 2. Summary of discourse measures from topic directed interview

X (SD) 600.2 (480.7) 7.8 (2.3) 77.8 (54.2) 88.2 (13.9) 11.8 (13.9) 8.5 (10.3) 2.2 (1.4) 0.3 (0.2) 18 (2.5) 7.6 (10.2) 14.9 (7.5) 76.2 (29.5) 8.5 (1.8) 1.6 (1.3) 0.5 (0.3)

FLD Subjects (n = 3)

Topic measures

Total # utterances Total # topic units # utterances / topic # on-topic utterances / topic # off-topic utterances / topic # could not evaluate utterances / topic % on-topic utterances % off-topic utterances % utterances in side sequences % utterances in side sequences - metastatements % utterances in side sequences - external % perseverative utterances % intrusive utterances

EB

CPY

CP

X (SD)

11 5 2.2 2.2 1.8 0 82 18 9.1 0 9.1 0 0

61 5 12.2 12.2 0 0 100 0 1.6 1.6 0 0 0

21 10 2.1 2.1 0 0 100 0 4.8 4.8 0 0 0

31.0(26.5) 6.7 (2.9) 5.5 (5.8) 5.5 (5.8) 0.6 (1.0) 0(0) 94(10.4) 6.0 (10.4) 5.2 (3.8) 2.1 (2.4) 3.0 (5.3) 0(0) 0(0)

Table 3. Measures of topic analysis for FLD subjects

Non-fluent PPA subjects (n == 5)

Topic measures

Total # utterances Total # topic units # utterances / topic # on-topic utterances/topic # off -topic utterances/topic # could not evaluate utterances/topic % on-topic utterances % off-topic utterances % utterances in side sequences % utterances in side sequences - metastatements % utterances in side sequences - external % perseverau've utterances % intrusive utterances % could not evaluate utterances/topic

RH

VH

FS

AS

BV

x(SD)

21 10 2.1 0.7 0.8 0.6 33 38 9.5 0 9.5 0 4.8 29

26 11 2.4 0.8 1 0.5 34.6 42.3 0 0 0 3.8 19.2 23.1

18 5 3.6 1.2 1.6 0.8 33 44 5.6 5.6 0 44 0 22

16 6 2.7 0.4 0.1 0.5 37.5 12.5 25 12.5 12.5 0 0 50

42 10 4.2 3.8 0.2 0.2 90.5 4.8 40.5 38.1 2.4 0 0 4.8

24.6 (10.4) 8.4 (2.7) 2.7 (0.7) 1.4(1.3) 0.7 (0.6) 0.5 (0.2) 45.7(25.1) 28.3 (18.3) 16.1(16.5) 11.2(15.9) 4.9 (5.8) 9.7 (19.3) 4.8 (8.3) 25.8 (16.3)

Table 4. Measures of topic analysis for non-fluent PPA subjects

Ruent PPA subjects (n = 6)

Topic measures

Total # utterances Total # topic units # utterances / topic # on-topic utterances/topic # off -topic utterances/ topic # could not evaluate utterances/topic % on-topic utterances % off -topic utterances % utterances in side sequences % utterances in side sequences - metastatements % utterances in side sequences - external % perseverative utterances % intrusive utterances

VB

JH

EM

RR

KS

JS

x(SD)

92 10 8.7 8.1 1.1 0 88 12 9.2 8.7 0 0 0

58 8 8.6 6.8 0.5 0 93 7 7.3 6.9 1.7 0 0

40 12 15 1.5 1.8 0 45 55 3.3 15 0 5 12.5

176 15 7.4 7.7 4.0 0 66 34 11.7 6.3 1.1 0 6.8

78 8 12.8 9.3 9.3 0 0 12.8 9.8 100 0 0 0

23 6 0 3.8 0 0 100 0 3.8 0 0 0 0

77.8 (54.2) 9.8 (3.3) 8.8(5.1) 6.2 (3.0) 2.8 (3.5) 0(0) 65.3 (37.9) 20.1(20.52) 7.5 (3.4) 35.9(45.1) 0.47 (0.75) 0.8 (2.04) 3.2 (5.3)

Table 5. Measures of topic analysis for fluent PPA subjects

Pragmatics in frontal lobe dementia and PPA

163

Percentages for subject groups FLD subjects Nonfluent PPA Fluent PPA subjects subjects (n = 5)

PC A Measures

Response to interlocutor Inappropriate Mostly inappropriate Subtotal

0 0 0

42.9 21.4 64.3

0 8.3 8.3

Some appropriate Mostly appropriate Appropriate Subtotal

0 33.3 66.7 100

7.1 14.3 14.3 35.7

0 33.3 58.3 91.6

Could not evaluate

20.0

44.0

20.0

5.8 5.8 11.7

38.9 50.0 88.9

8.3 33.3 41.6

Some appropriate Mostly appropriate Appropriate Subtotal

0 17.6 70.6 88.2

11.1 0 0 11.1

2.8 16.6 38.9 58.3

Could not evaluate

5.6

40.0

0

Inappropriate Mostly inappropriate Subtotal

0 0 0

18.2 18.2 36.4

0 10.8 10.8

Some appropriate Mostly appropriate Appropriate Subtotal

0 0 100 100

9.0 27.3 27.3 63.6

5.4 8.1 75.7 89.2

Could not evaluate

9.5

68.6

11.9

Semantic content Inappropriate Mosdy inappropriate Subtotal

Cohesion

Table 6.

Summary of percentage scores on the Profile of Communicative Appropriateness for FLD, Nonfluent PPA, and Fluent PPA subjects

164

J.B. Orange, A. Kertesz and J. Peacock

Table 6 (continued) Percentages for subject groups Fluent PPA FLD Subjects Nonfluent PPA Subjects Subjects (n = 5) (n=6)

PCA Measures

Fluency

0 0

57.1 23.8

17.1 19.5

0

80.9

36.6

Some appropriate Mostly appropriate Appropriate Subtotal

0 6.3 93.7 100

4.7 14.3 0 19.0

14.6 21.9 26.8 63.3

Could not evaluate

23.8

40.0

2.4

0 0

50 10

5.9 8.8

0

60

14.7

Some appropriate Mostly appropriate Appropriate Subtotal

27.3

9.1 63.6 100

10 20 10 40

11.8 29.4 44.1 85.3

Could not evaluate

61.9

50.0

29.2

3.6 14.3

4.3 15.2

1.7 1.7

17.9

19.5

3.4

Some appropriate Mostly appropriate Appropriate Subtotal

10.7 10.7 60.7 82.1

8.7 10.9 60.9 80.5

11.8 10.2 74.6 96.6

Could not evaluate

15.2

16.4

10.6

Inappropriate Mostly inappropriate Subtotal

Sociolinguistic sensitivity Inappropriate Mostly inappropriate Subtotal

Non-verbal communication Inappropriate Mostly inappropriate Subtotal

Pragmatics in frontal lobe dementia and PPA

165

Discourse measures As a backdrop to the results obtained from the analyses of the pragmatic performance of the FLD, non-fluent, and fluent PPA subjects, several measures of discourse were first calculated. A summary of the findings are provided in Table 2. These measures reflect traditional measures of language performance (e.g., number of words, number of utterances, mean length of utterance (MLU), as well as discourse measures reflecting intraand inter-sentential relationships of lexical and referential elements (e.g., stereotyped phrases, pronouns without antecedents, and demonstrative pronouns). A review of the mean scores of the discourse measures for all three groups reveal the following. From the standpoint of overall quantity of production, non-fluent PPA subjects generated the fewest words (x =_65.8), fewest utterances (x = 24.6), and the fewest number of words per minute (x = 40.3). The fluent PPA subjects, on the other hand, generated the highest number of words (x = 600.2), the highest number of utterances (x = 77.8), and had a substantially higher rate of words per minute (x = 76.2). The MLU for the non-fluent PPA subjects (x = 3.62) was less than one half the values for the FLD subjects (x = 8.3) and the fluent PPA subjects (x = 7.8). These findings are in keeping with current views on the reduced spoken output of non-fluent PPA patients and the normal or near normal levels of spoken output among fluent PPA subjects (Thompson et al., 1997). The quantity of the spoken output of the FLD subjects was skewed upwards by the scores of one subject (CPY). High verbal output in FLD subjects has been documented previously (Gustafson, 1993; Neary, 1990). The mean number of words and utterances of the other FLD subjects, however, were below those of the highly verbal fluent PPA group. The percentage of incomplete utterances by the non-fluent and fluent PPA groups were substantially higher (x = 11.1 versus x = 11.8) than the percentage for the FLD group (x = 1.7). The use of stereotyped phrases, pronouns with antecedents, and demonstrative pronouns was highest among the non-fluent PPA group and second highest among the fluent PPA group. These measures reflect, in part, the word finding difficulty described in subjects with PPA by other investigators, and particularly those with a fluent-like PPA (Weintraub et al., 1990). Scores for the FLD group were lower than those of the PPA subjects, suggesting that they experienced less word finding difficulty in the topic-directed interviews. The following example from non-fluent PPA subject BV illustrates the incomplete utterances and stereotyped utterances characteristic of the discourse samples of the non-fluent PPA group. Example 1 Examiner: Subject: Subject: Subject: Subject: Subject: Examiner: Subject: Subject: Subject: Subject: Subject: Subject: Subject:

and your address? um it's uh R or... no wait a minute! its um ... that's not right um today. it it's in Clinton. uhhuh. and and um it's uh four no ... oh dear. um let's see. Clinton and um ... oh gosh. my uh address is um ah... oh gosh.

166

J.B. Orange, A. Kertesz and J. Peacock

Topic Differences in pragmatic performance among the FLD, non-fluent, and fluent PPA subjects were more pronounced on measures of conversational topic management. The summary of scores for the FLD subjects, the non-fluent PPA subjects, and the fluent PPA subjects are presented in Tables 3, 4, and 5, respectively. The non-fluent PPA subjects produced the fewest number of utterances per topic (x = 2.7) compared to the fluent PPA subjects (x = 8.8) and the FLD subjects (x = 5.5). The non-fluent PPA subjects also produced the fewest number of on-topic utterances per topic (x = 1.4) and had the lowest percentage of on-topic utterances (x = 45.7). The fluent PPA subjects had only two-thirds of their utterances rated as being on-topic (x = 65.3%), while the FLD subjects talked predominantly on-topic (x = 94%). The non-fluent PPA subjects also exhibited the highest percentage of off-topic utterances (x = 28.3) compared to the fluent PPA subjects (x = 20.1) and the FLD subjects (x = 6.0). The following examples illustrate the off-topic utterances produced by the non-fluent PPA, fluent PPA, and FLD subjects. In Example 2, non-fluent PPA subject RH produces off-topic utterances that are unrelated to the examiner's question about her name. Example 2 Examiner: Subject: Subject: Subject: Subject: Subject:

tell me your full name. R G H (full name spoken). and I don't very much very very much about uh um um day. and my uh my father's dead. my father uh you know he was really... we'd he's awful.

In Example 3, the fluent PPA subject RR talks about unrelated issues following the examiner's questions. Example 3 Examiner: Subject: Examiner: Subject: Subject: Examiner: Subject: Examiner: Subject: Examiner: Subject: Subject:

when did you notice you had language problems? uh the the things that uh you know in in talking. in talking? well when I was in working and and uh I had... I get the people talking about... uhhuh. that me... yeah. okay? yeah. now if they don't do the things that uh I think that's not right then I would you know just give them something and say well "wouldn't it be..." you know things like that.

Although of low occurrence, FLD subjects did produce off-topic comments. In Example 4, FLD subject EB responds off-topic to a question by the examiner: Example 4 Examiner:

what problems have you been having Mr. B.?

Pragmatics in frontal lobe dementia and PPA Subject: Subject: Examiner: Examiner:

167

uh I had an M I R (spelled each letter) and and a and and I had a C A T scan (spelled) and I proper proper name is and the and it's called... but those are the tests that you've had. what is your problem?

Additional detailed analyses were undertaken to examine the percentage of off-topic talk in which the subjects temporarily suspended discussion about the primary topic (i.e., side sequence) (Jefferson, 1972). Side sequences included (a) talk on a topic related to the primary topic (i.e., a subtopic), (b) comments concerning the lexical or emotional difficulty of talking about the primary topic (i.e., side sequence - meta-statement), and (c) comments that addressed matters in the immediate context of the assessment (e.g., video camera, microphone, lighting, etc.) or were otherwise unrelated to the primary topic. The non-fluent PPA subjects had the highest percentage of utterances categorized in side sequences(x = 16.1), with fluent PPA subjects (x = 7.5) and FLD subjects (x = 5.2) exhibiting less than half the percentage of the non-fluent group. The fluent PPA group, however, had the highest percentage of utterances in side sequences that were metastatements (x = 35.9) compared to the non-fluent PPA subjects (x = 11.2) and the FLD subjects (x = 3.0). This latter finding is not surprising considering that the fluent PPA group exhibited the greatest word finding difficulty, as measured by the discourse measures and their ratings on the appropriate use of items in the Semantic Content and Fluency categories of the PCA (see below). The non-fluent PPA subjects had the highest percentage of side sequence utterances that were unrelated to the primary topic (x - 4.8). For all three groups, however, the percentage of utterances categorized as unrelated and external to the primary topic was relatively low. The following examples illustrate the meta statements made by fluent and non-fluent PPA subjects. In Example 5, the fluent PPA subject VB comments on her word finding difficulty in response to the examiner's question. Example 5 Examiner: Subject: Examiner: Subject: Examiner: Subject: Subject: Subject: Subject: Subject: Subject:

have you been here before? where? in this hospital? for mmh something that happen to me? uhhuh. no. yesterday in with professor or doctor oh .... damn. I am thinking of his name. it's coming but uh ... when I when I am tense everything goes (gestures with hands).

In Example 6, non-fluent PPA subject BV comments that she would like to have a relative come in and explain her difficulty. Example 6 Examiner: Subject: Subject: Subject: Examiner:

What kinds of problems have you been having? Well I don't know how you know I'm .... I wish K (name of relative) would be could her to now ... to tell yeah. I know you'd like to ask her yeah.

168

J.B. Orange, A. Kertesz and J. Peacock

Subject:

but um um well uh uh I'm I'm in the um in the ...

The following example illustrates the external utterances produced almost exclusively by the non-fluent PPA subjects. In example 7, non-fluent PPA subject RH makes unrelated comments to the examiner's question before continuing on the topic. Example 7 Examiner: Subject: Subject: Subject: Subject: Examiner: Examiner:

What kind of work did you use to do? oh uh it I don't know and uh ... he had a store for a while you know and it runs and its and it runs and it gobble up. it wasn't in on Owen Owen Sound it was later we have to get home to get an an an I I didn 't. that's okay. you have a little bit longer you daughter in law knows and so does B.

For the last two measures on conversational topic, non-fluent PPA subjects exhibited the highest percentage of perseverative (x = 9.7) and intrusive utterances (x = 4.7). The fluent PPA subjects produced few perseverative and intrusive utterances, while FLD subjects did not produce any at all. The majority of the perseverative utterances produced by the nonfluent PPA subjects were stereotyped phrases including "Like no" and" Okay yeah". Profile of Communicative Appropriateness (PCA) As a final measure of the pragmatic performance of the FLD subjects, the non-fluent PPA subjects, and the fluent PPA subjects, ratings were made (i.e., Inappropriate, Mostly inappropriate, Sometimes appropriate, Mostly appropriate, and Appropriate) for items in each of the six domains of pragmatics comprising the PCA. Domain items of the PCA are listed in Appendix B. Ratings for items within each domain were collapsed for subjects in the FLD, non-fluent, and fluent PPA groups. Percent scores for the ratings were calculated for items that could be evaluated in each domain. These figures are presented in Table 6. For the pragmatic domain of Response to Interlocutor, the non-fluent PPA subjects' performance on items that could be evaluated, such as requesting information, using clarification requests, and replying to questions, were rated frequently as either inappropriate or mostly inappropriate (64.3%). The fluent PPA subjects (91.6%) and the FLD subjects (100%), on the other hand, were rated as appropriately using responses with their conversational partner the majority of the time. For the pragmatic domain of Semantic Content, the non-fluent PPA subjects showed the highest ratings of inappropriate or mostly inappropriate use of elicited items related to topic management, word finding, and idea completion (88.9%). These ratings concur with the objective measures of topic management discussed above wherein the non-fluent PPA subjects exhibited the highest percentage of off-topic utterances. The fluent PPA subjects' ratings also showed that they experienced problems with idea sequencing, idea completion, and maintaining or shifting topics (41.6%). The ratings for the FLD subjects showed that they experienced little difficulty with issues related to topic and idea completion (88.2%). Again, the objective measures of topic maintenance discussed previously for the FLD subjects support these ratings. In the domain of Cohesion, which includes of use of references, tense structure, and other inter- and intra-sentential linguistic elements, the non-fluent PPA subjects showed

Pragmatics in frontal lobe dementia and PPA

169

the poorest ratings for items that could be evaluated (36.4%). It should be noted, however, that a high percentage of items could not be evaluated as either the non-fluent PPA subjects could not produce the items and therefore the items did not appear in the transcript, or because the task was not sufficient in structure, complexity, or length to elicit the items. The ratings for the non-fluent PPA subjects for items that could be evaluated in the Fluency domain of pragmatics, such as repetitions, revisions, incomplete phrases, and word finding difficulties, were either inappropriate or mostly inappropriate (80.9%). For the fluent PPA subjects, nearly two thirds of the items in the Fluency domain that could be evaluated were used appropriately or somewhat appropriately (63.3%). FLD subjects were rated with few problems for revision and word findings abilities. For items that could be evaluated in the pragmatic domain of Sociolinguistic Sensitivity, such as the use of stereotypes, indirect speech acts, and acknowledgements, the non-fluent PPA subjects were rated as using the items inappropriately or mostly inappropriately the majority of the time (60%). The fluent PPA subjects and the FLD subjects experiencedfewer difficulties, as reflected in their high ratings of appropriate use of these complex elements of pragmatic performance. It should be noted, however, that many of the items comprising this domain could not be evaluated for the FLD and nonfluent PPA subjects for reasons similar to those discussed for items in the Cohesion domain. Finally, the FLD subjects, the non-fluent PPA subjects, and the fluent PPA groups were rated quite highly in the appropriate use of speech and gestural components of nonverbal communication. Most of the items in the domain could be evaluated for all three subject groups. Discussion The dataderivedfrom topic-directed interviews and reported herein are the first to address in a systematic manner the pragmatic performances of subjects with FLD, non-fluent PPA, and fluent PPA. The use of three separate sets of measures of discourse and pragmatics show consistent performance among related constituent elements of pragmatics for all three subject groups. Differences in pragmatic performance among the three diagnostic groups, however, reveal three separate, but overlapping profiles of abilities and weaknesses. One of the important findings of our study is that the two subgroups of PPA subjects can be identified and distinguished on the basis of their pragmatic performance. This supports the assertions made by others, on the basis of analyses of language and neuropsychological data, that there are two, and possibly three, subgroups of PPA patients (Snowdenet al., 1992; Thompson et al., 1997). The non-fluent PPA subjects' pragmatic performance, based on the multiple objective and rating measures used in this study, can be characterised best as inappropriate use of topic maintenance skills, production of off-topic comments, poor use of responses to partner's questions and requests, difficulty tracking and using devises of reference, and limited contributions to the semantic development of topics, most likely the result of word finding difficulties and incomplete utterances. The pragmatic performance profile of the fluent PPA subjects in this study, while similar in some respects to that of the non-fluent group, is distinguished by their more subtle difficulties in topic maintenance such that they are able to develop topics more fully and respond more appropriately to requests by their discourse partner. The percent of their talk that is rated as appropriate to the requests and demands of their partners is greatly superior to that of non-fluent PPA subjects. Moreover, the fluent PPA subjects display an

170

J.B. Orange, A. Kertesz and J. Peacock

insightfulness and awareness of the linguistic and pragmatic problems that is not apparent in the non-fluent PPA subjects. The fluent PPA subjects' comments about their difficulty communicating reflects an awareness of problems that is not evident among the non-fluent PPA group. It is fully acknowledged that the extent of differences in the pragmatic performances between the non-fluent and fluent PPA groups is subtle and may only be a matter of the degree of the ratings within the same domain of pragmatics. Further detailed analyses of additional components of pragmatics, such as the correct use and interpretation of figurative language or the appropriate use of paralinguistic phenomena (e.g., affective prosody), is warranted to explore more fully differences in the pragmatic performances of the PPA subgroups. A second important finding is that the pragmatic performances of the FLD subjects in this study are generally well preserved and uniquely different from the PPA subjects. Whereas the PPA subjects exhibited greater difficulty keeping their comments on the topics presented by the examiner, FLD subjects showed greater ability to generate comments that were on-topic and relevant to the theme of discussion. The FLD subjects also were judged as more appropriate in their responses to the requests of the examiner and in their acknowledgements to the responses of the examiner, despite reduced verbal output for two of the three FLD subjects in the discourse sampling task. The pragmatic performance of the FLD subjects in this study can best be described by Grice's (1975) general principles, or maxims, of conversation. That is to say, our FLD subjects' pragmatic performances are appropriate in relevance, manner, and quality, although quantity may be reduced. This finding is at variance with previous descriptions of the language and communicative performance of FLD subjects, such as increased talkativeness, perseveration, irrelevance, reduced conversational initiation, insensitivity, social inappropriateness, and confabulation (Barber et al., 1995; Gustafson, 1993; Neary, 1990; Neary et al., 1990). Previous reports of the language and communication individuals with FLD are generally based on accounts of performance recorded during standardized neuropsychological testing (Miller, Cummings, Villanueva-Meyer, Boone, Mehringer, Lesser, & Mena, 1991; Neary et al., 1990), from interview-based responses of relatives of patients (Barberetal., 1995; Kertesz,Fox, & Davidson, 1997), or from selective tests of language (e.g., verbal fluency) (Pasquier, Lebert, Grymonprez, & Petit, 1995). The features of conversation-elicited stories given by relatives is difficult to capture on formal, standardized tests. The nature of the topic directed interview task used in this study, however, may very well have provided the linguistic and thematic support needed by the FLD subjects to overcome their inertia in the initiation, maintenance, and relevance of their responses. This is a limitation of the topic directed interview. Future studies might explore the limits of the pragmatic performance of FLD subjects under less constraining contexts (spontaneous conversation with a unfamiliar partner). Moreover, studies also need to be undertaken to explore the relationship among general cognitive domains (such as attention and memory), language, frontal lobe-based behaviours (e.g., inhibition, judgement, set-shifting), and multiple measures of pragmatics such as those used in this study. Exploring the nature and strength of the relationships in a group of FLD patients will advance our theoretical understanding of the cognitive and linguistic influences on pragmatics, and how their relative influence changes over the progression of the illnesses. The final important finding of the study is that there was considerable agreement and overlap of findings among the three approaches used in the study to examine the pragmatic performance of the FLD and PPA subjects. The results from the detailed and very timeconsuming transcription, coding, and analyses of discourse and pragmatic features, such as percent pronouns without antecedents, percent stereotyped utterances, and percentage of onand off-topic utterances, among others, were similar to the rating profiles obtained on the

Pragmatics in frontal lobe dementia and PPA

171

PCA. The ability of the PCA to discriminate the pragmatic performances of two individuals with different types of dementia has been shown (Penn et al., 1988). It may very well be the case that the PCA can be used to examine and distinguish the pragmatic performances of subjects suspected of having FLD or PPA rather than undertaking lengthy transcription and micro-analysis procedures. Future studies should examine the potential of the PCA in this regard. Acknowledgements—The authors are grateful for the help of Pat McCabe in collecting the discourse samples. The assistance of Christina Acre in coding the discourse samples also is acknowledged. The support and participation of the subjects and their families in this study are gratefully appreciated and acknowledged.

References Barber, R., Snowden J., & Craufurd, D. (1995). Frontotemporal dementia and Alzheimer's disease: Retrospective differentiation using information from informants. Journal of Neurology, Neurosurgery, and Psychiatry, 59, 61-70. Benton, A. L. (1974). Revised Visual Retention Test (4th ed.). New York: The Psychological Corporation. Benton, A. L., & Van Allen, M. W. (1972). Prosopagnosia and facial discrimination. Journal of the Neurological Sciences, 15, 167-172. Brun, A., Mann, D. M. A., Englund, B., Neary, D. et al. (1994). Consensus on clinical and neuropathological criteria for frontal lobe dementia. Journal of Neurology, Neurosurgery, and Psychiatry, 57, 416-418. Caselli, R. J., Windebank, A. J., Petersen, R. C, Komori, T., Parisi, Okazaki, H., et al. (1993). Rapidly progressive aphasic dementia and motor neuron disease. Annals of Neurology, 33 (2), 200-207. Delis, D. C., Kromer, J. H., Kaplan, E., & Obler, B. A. (1987). California Verbal Learning Test. New York: The Psychological Corporation. Duffy, J. R., & Petersen, R. C. (1992). Primary progressive aphasia. Aphasiology, 6, 1-15. Garcia, L. J., & Joanette, Y. (1994). Conversational topic-shifting analysis in dementia. In R. L. Bloom, L. K. Obler, S. De Santi, & J. S. Ehrlich (Eds.), Discourse analysis and applications: Studies in adult clinical populations (pp. 161-183). Hillsdale, NJ: Erlbaum. Green, J., Morris, J. C., Sandson, J., McKeel, D. W., & Miller, J. W. (1990). Progressive aphasia: A precursor of global dementia? Neurology, 40,423-429. Grice, H. P. (1975). Logic and conversation. In P. Cole and J. L. Morgan (Eds.), Syntax and semantics 3: Speech acts (pp. 41-58). New York: Academic Press. Gustafson, L. (1993). Clinical picture of frontal lobe degeneration of the non-Alzheimer type. Dementia, 4, 143-148. Hooper, H. E. (1958). The Hooper Visual Organization Test: Manual. Beverley Hills, CA: Western Psychological Services. Homer, J. (1985). Language disorder associated with Alzheimer's dementia, left hemisphere stroke, and progressive illness of uncertain aetiology. In R. H. Brookshire (Ed.), Clinical Aphasiology Conference Proceedings (pp. 149-158). Minneapolis, MN: RBK Publishers. Illes, J. (1989). Neurolinguistic features of spontaneous language production dissociate three forms of neurodegenerative disease: Alzheimer's, Huntington's and Parkinson's. Brain and Language, 37, 628642. Jefferson, G. (1972). Side sequences. In D. Sudnow (Ed.), Studies in social interaction (pp. 294-338). New York: Free Press. Johanson, A., & Hagberg, B. (1989). Psychometric characteristics in patients with frontal lobe degeneration of non-Alzheimer type. Archives of Gerontology and Geriatrics, 8, 128-137. Karbe, H., Kertesz, A., & Polk, M. (1993). Profiles of language impairment in primary progressive aphasia. Archives of Neurology, 50, 193-201. Keenan, E. O., & Schieffelen, B. B. (1976). Topic as a discourse notion: A study of topic in the conversations of children and adults. In. C. N. Li (Ed.), Subject and topic: A new typology of language (pp. 335-384). New York: Academic Press. Kempler, D., Metter, E. J, Riege, W. H., Jackson, C. A., Benson, D. F., & Hanson, W. R., (1990). Slowly progressive aphasia: Three cases with language, memory, CT, and PET data. Journal of Neurology, Neurosurgery, and Psychiatry, 53, 987-993. Kertesz, A. (1982). Western Aphasia Battery. New York: Grune & Stratton. Kertesz, A., Fox, H., & Davidson, W. (1997). Frontal Behavioural Inventory: Diagnostic criteria for frontal lobe dementia. Canadian Journal of Neurological Sciences, 24, 29-36. Kertesz, A., Hudson, L., Mackenzie, I. R. A., & Munoz, D. G. (1994). The pathology and nosology of primary progressive aphasia. Neurology, 44, 2065-2072. Kesler, A., Artzy, T., Yaretzky, A., & Kott, E. (1995). Slowly progressive aphasia, a left temporal variant of probable Pick's disease: 15 years of follow-up. Israel Journal of Medical Science, 31, 626-628.

172

J.B. Orange, A. Kertesz and J. Peacock

MacWhinney, B. (1991). The CHILDES project: Tools for analyzing talk. Hillsdale, NJ: Lawrence Erlbaum Associates. Mattis, S. (1976). Mental status examination for organic mental syndrome in the elderly patient. In L. Bellak and T. B. Karasu (Eds.), Geriatric Psychiatry. NY: Grune & Stratton. Mentis, M., Whittaker, J., & Gramigna, G. D. (1995). Discourse topic management in senile dementia of the Alzheimer's type. Journal of Speech and Hearing Research, 35, 1054-1066. Mesulam, M.-M. (1982). Slowly progressive aphasia without generalized dementia. Annals of Neurology, 11, 592-598. Miller, B. L., Cummings, J. L., Villanueva-Meyer, J., Boone, K., Mehringer, C. M., Lesser, I. M., & Mena, I. (1991). Frontal lobe degeneration: Clinical, neuropsychological, and SPECT characteristics. Neurology, 41, 1374-1382. Neary, D. (1990). Dementia of the frontal lobe type. Journal of the American Geriatrics Society, 38, 71-72. Neary, D., Snowden, J. , Mann, D. M. A., Northen, B., Goulding, P. J., & McDermott, N. (1990). Frontal lobe dementia and motor neuron disease. Journal of Neurology, Neurosurgery, and Psychiatry, 53, 23-32. Nicholas, L. E., & Brookshire, R. (1993). A system for quantifying the informativeness and efficiency of the connected speech of adults with aphasia. Journal of Speech and Hearing Research, 36,338-350. Osterrieth, P. A. (1944). Le testde copie d'une figure complexe: Contribution a 1'étude de la perception et de la mémoire. Archives de Psychologie, 30, 286-356. Pasquier, F., Lebert, F., Grymonprez, L., & Petit, H. (1995). Verbal fluency of frontal lobe type and dementia of the Alzheimer type. Journal of Neurology, Neurosurgery, and Psychiatry, 58, 81-84. Penn, C. (1985). The Profile of Communicative Appropriateness: A clinical tool for the assessment of pragmatics. South African Journal of Communication Disorders, 32, 18-23. Penn, C. (1988). The profiling of syntax and pragmatics in aphasia. Clinical Linguistics & Phonetics, 2, 179207. Penn, C., Sonnenberg, B., & Schnaier, Y. (1988). Dementia and communication pathology: Two case examples. South African Journal of Communication Disorders, 35, 65-74. Raven, J. C. (1947). Coloured Progressive Matrices Sets A, Ab, and B. London: H. K. Lewis. Retherford, K. S. (1993). Guide to analysis of language transcripts. Eau Claire, WI: Thinking Publications. Rey, A. (1941). L'examen psychologique dans les cas d'encéphalopathie traumatique. Archives de Psychologie,_28, 286-340. Ripich, D. N., & Terrell, B. (1988). Patterns of discourse cohesion and coherence in Alzheimer's disease. Journal of Speech and Hearing Disorders, 53,8-15. Shewan, C. M. (1988). The 'Shewan Spontaneous Language Analysis (SSLA)' system for aphasic adults: Description, reliability, and validity. Journal of Communication Disorders, 21, 103-138. Snowden, J. S., Neary, D., & Mann, D. M. A. (1996). Pronto-temporal lobar degeneration: Pronto-temporal dementia, progressive aphasia, semantic dementia. New York: Churchill Livingstone. Snowden, J. S., Neary, D., & Mann, D. M. A., Goulding, P. J., & Testa, H. J. (1992). Progressive language disorder due to lobar atrophy. Annals of Neurology, 31, 174-183. Strong, M. J., Grace, G. M., Orange, J. B., & Leeper, H. A. (1996). Cognition, language, and speech in amyotrophic lateral sclerosis: A review. Journal of Clinical and Experimental Neuropsychology, 18, 291303. Thompson, C. K., Ballard, K. J., Tait, M. K, Weintraub, S., & Mesulam, M. (1997). Patterns of language decline in non-fluent primary progressive aphasia. Aphasiology, 11, 297-321. Tomoeda, C. K., & Bayles, K. A., (1993). Longitudinal effects of Alzheimer disease on discourse production. Alzheimer Disease and Associated Disorders, 7, 223-236. Ulatowska, H. K., Allard, L., Reyes, B. A., Ford, J., & Chapman, S. (1992). Conversational discourse in aphasia. Aphasiology, 6,325-331. Wechsler, D. (1981).Wechsler Adult Intelligence Scale-Revised. NY: Psychological Corporation. Wechsler, D. (1987). Wechsler Memory Scale - Revised. NY: Psychological Corporation. Weintraub, S., Rubin, N. P., & Mesulam, M.-M. (1990). Primary progressive aphasia: Longitudinal course, neuropsychological profile, and language features. Archives of Neurology, 47, 1329-1335.

Appendix A Definitions for outcome measures Discourse measures Words Words were identified according to Nicholas and Brookshire's (1993) definitions and guidelines. According to their definition, words are "intelligible in context to someone who knows the ...topic being discussed". They "do not have to be accurate, relevant, or informative relative to the ... topic being discussed (p.348)."

Pragmatics in frontal lobe dementia and PPA

173

Mean length of utterance The average number of morphemes per utterance (Retherford,1993, p. 272). Utterance An utterance, as defined by Shewan (1988), is a "complete thought, usually expressed in a connected grouping of words, which is separated from other utterances on the basis of content, intonation contour, and/or pausing" (p. 124). Utterances may also be expressed a nonverbal actions. Complete utterance A complete utterance was defined as an utterance which expresses the speaker's complete, uninterrupted thought Incomplete utterance Incomplete utterances were identified following the definitions and guidelines of the Childes Project (MacWhinney, 1991). According to this definition, incomplete utterances are "incomplete but not interrupted" utterances in which the speaker "trails off" without completing the thought (p. 43). e.g., Subject: Subject:

The doctor here was in touch with him and told him... Oh the B-12 shot was something that I got for a long time.

Stereotype utterance A stereotype utterance contains an idiosyncratic, non-propositional phrase or phrases, such as fine thank you, pretty good, okay now, etc. e.g., Subject: Examiner: Subject:

I was very good. Tell me about all the jobs that you 've had. I was very good um...

Overt statements regarding anomia Overt statements regarding anomia refer to comments made by the speaker which indicate that he/she is experiencing difficulty recalling an intended word. e.g., Subject: Subject: Subject:

Yesterday with doctor, or professor... I am thinking of his name. I can't seem to recall it.

Self correction Self corrections refer to all instances of "retracing with correction" as defined and described in the Childes Project (MacWhinney, 1991). Self corrections refer to instances where " a speaker starts to say something, stops, repeats the basic phrase, changes the syntax but maintains the same idea" (p.52). e.g., Subject:

It just it was just terrible to try and do it.

Pronoun without antecedent A pronoun without antecedent includes a referent which was not clearly established in the preceding discourse and thus is unknown or ambiguous to the conversational partner. Demonstrative pronoun A pronoun that points out that which it modifies (e.g., this, that, those, these) (Retherford, 1993, p. 272). The pro-forms here and there, which designate location, are also included in this category. e.g., Subject: e.g., Subject:

I think those sleeping pills helped me. I've never been here before.

Repetitions Repetitions refer to all instances of "retracing without correction" described in the Childes Project guidelines and definitions (MacWhinney, 1991). Repetitions refer to instances where "a speaker begins to say something, stops and then repeats the earlier material without change" (p.51). Repetitions may occur on sounds, syllables, words or phrases. e.g., Subject:

I think / think those pills helped me.

II. Topic Analysis Topic unit A topic unit is defined as a sequence of utterances that address the same topic and/or an utterance that addresses a unique topic. The concepts of topic unit, on-topic, and off-topic were drawn from and based on

174

J.B. Orange, A. Kertesz and J. Peacock

Keenan and Schieffelin's (1976) definition of topic, as well as definitions by Garcia and Joanette (1994) and Mentis, Whittaker, and Gramigna (1995). Several aspects of a given topic may be discussed within the scope of a single topic unit. In a topic-directed interview format topic unit boundaries typically coincided with each new question asked by the examiner, however this was not always the case. e.g., Topic unit #1 Topic unit #2

Examiner: Tell me all about your family. Subject: I've got an older son and he's home right now. Subject: My youngest son's up in Guelph studying veterinarian medecine. Subject: Good. Examiner: Tell me about your health right now. Subject: Well it's not too bad.

Side sequence A side sequence refers to an utterance or a series of utterances which temporarily suspend the on-going topic of discussion to address matters which arise during the course of the discussion. The definition was based on the working definition of Jefferson (1972). Two distinct types of side sequence were identified: metastatements and external side sequences. Meta-statement A meta-statement is a specific type of side sequence in which the speaker suspends the on-going conversation to comment on some aspect of the topic being discussed or on some aspect of his/her performance in discussing the immediate topic. Meta-statements may occur as a single utterance or a series of utterances within the boundaries of a topic unit e.g., Examiner: Subject:

So tell me all about your family. Do you mean my personal family or my parents and stuff?

External side sequence An external side sequence is an utterance or series of utterances which interrupt the on-going topic of conversation to address matters in the immediate environment or other matters which are unrelated to the immediate topic of conversation. Examples of external side sequences include: Comments about an event which arises in the interview environment during discussion of a given topic. e.g., Subject: Subject:

I started to read one of those things from a to z and it took me an hour. Is that a computer on your desk?

Comments about matters external to the interview situation. e.g., Subject: Subject: Examiner:

He had a store but that wasn't in Owen Sound it was later. We have to go home soon. That's okay we only have a little while longer to go.

On-topic utterance An on-topic utterance is defined as an utterance that contributes to the advancement or maintenance of the ongoing topic of discussion. On-topic utterances: • demonstrate obvious and appropriate relation to the on-going topic of discussion, • demonstrate no obvious relation to the on-going topic in terms of semantic content, but function to permit the continuation of the topic, • are identified as meta-statements, • are self corrections. Off-topic utterance An off-topic utterance is defined as an utterance that does not contribute to the advancement or maintenance of the on-going topic of discussion and appears inappropriate within the context of the immediate topic. Offtopic utterances include: • those which demonstrate no apparent or appropriate relation to the on-going topic of discussion, • those which do not appear to be functioning to permit the continuation of an on-going topic, • those which convey false or incorrect information with respect to the topic of discussion or are identified as external side sequences, • perseverative or intrusive utterances. Could not evaluate utterance Could not evaluate utterances include those which could not be analyzed for topic measures as they lacked intelligible content. Responses identified as could not evaluate include: • completely unintelligible utterances with no identifiable semantic content. • utterances consisting solely of non-meaningful filler words (e.g., um, uh) or sound repetitions.

Pragmatics in frontal lobe dementia and PPA

175

Perseverative utterance A perseverative utterance is one which continues the preceding topic of conversation despite the introduction of a new topic by the conversational partner. It also includes exact repetition of utterances or repetition of utterances with the same propositional or non-propositional meaning. Intrusive utterance An intrusive utterance is one in which the speaker refers to a previously discussed topic of conversation and it is inappropriate within the context of the current topic of discussion.

Appendix B Profile of Communicative Appropriateness Inapp. R e

I n

Request

s

t

Reply

e P

r

Clarification request

to 1 o

o

Acknowledgment

c n

u

Teaching Probe

s

t

Others

e

o r

S C

Topic initiation

e o

Topic adherence

m n

Topic shift

a t

Lexical choice

n e t n i c

t

Idea completion Idea sequencing Others

Mostly Inapp.

Some App.

Mostly Inapp.

App.

CNE

Comments

176

J.B. Orange, A. Kertesz and J. Peacock

Profile of Communicative Appropriateness (continued) Inapp.

c

Ellipsis

o

Tense use

h

Reference

e

Lexical substitution forms

s

Relative clauses

i

Prenominal adjectives

o

Conjunctions

n

Others

F

Interjections

1

Repetitions

u

Revisions

e

Incomplete phrases

n

False starts

c

Pauses

y

Word-finding diffic. Others

s

Polite forms

o c

S

i

e

o

n

1

s

Reference to interlocutor Placeholders, fillers, stereotypes Acknowledgment

i

i

n

t

g

»

Self correction

u

v

Comment clauses

i

i

Sarcasm/humour

s

t

Control of direct speech

t y i

Indirect speech acts

c

Others

Mostly Inapp.

Some App.

Mostly App.

App.

CNE

Comments

Pragmatics in frontal lobe dementia and PPA

177

Profile of Communicative Appropriateness (continued) Inapp.

c

Vocal aspects:

o

Intensity m

Pitch

N m

Rate

o u

Intonation

n

n

Quality

v

i

Non-verbal aspects:

e

c

Facial expression

r

a

Head movement

b t

Body posture

a

i

Breathing

1

0

Social distance

n

Gesture and pantomime Others TOTALS

Mostly Inapp.

Some App.

Mostly App.

App.

CNE

Comments

This page intentionally left blank

J. Neurolinguistics, Vol. 11, Nos 1-2, p. 179-190, 1998 © 1998 Published by Elsevier Science Ltd. All rights reserved Printed in Great Britain 0911-6044/98 $19.00 + 0.00

PII: S0911-6044(98)00012-8

Pragmatics in the absence of verbal language:

Descriptions of a severe aphasic and a language-deprived adult Nina F. Dronkers , Carl A. Ludy and BrendaB.Redfern

*VA

Northern California Health Care System;

University of California, Davis

Abstract—Two cases with vastly different etiologies are presented to illustrate pragmatic competence in the absence of verbal language. The first is a man with severe Broca's aphasia who lost the ability to use any propositional language after a massive left hemisphere stroke. The second is a congenially-deaf woman with no exposure to language until well into adulthood. Despite their lack of verbal skills, both cases demonstrate a full command of pragmatic abilities and function as competent social actors. This finding reinforces the view of pragmatics as a vital part of social interaction. Introduction Ask any linguist, psycholinguist, speech pathologist, neurolinguist, neurologist, or philosopher what pragmatics is and you may get a cautious answer. Most would venture that pragmatics has something to do with the social aspects of language, i.e., the way in which things are communicated rather than what is communicated. Yet the specifics of how the 'way' and the 'what' relate to each other and how they interact to achieve effective communication are much more difficult to pin down. Most people are just not sure what role pragmatics plays in communication or to what extent it relates to the other components of language. Too frequently, 'language' is considered to consist only of syntax, semantics, morphology and phonology. Yet, we know that the inappropriate use of prosody, vocal intensity, eye contact, turn-taking, maintenance to the topic, or physical proximity to the listener can cause an utterance to be misinterpreted or cause the listener to "tune out". These errors result in poor communication just as surely as grammatical, lexical, or phonological errors cause misinterpretations. It has been known for some time that certain patients (mostly with right hemisphere disease) are shunned for their odd pragmatic behavior, even though their grammar, lexical selection, and pronunciation may be flawless. At the same time, aphasic patients may demonstrate perfect pragmatic skills and not be able to utter a word. Even with this knowledge, pragmatics continues to be low on the totem pole of language, and its role in communication is often overlooked. In this paper, we offer some thoughts on the importance of pragmatics in communication particularly in light of its remarkable perseverance in the absence of verbal language. In our Aphasia Research Lab, we have been struck by the similarities between two patients of vasdy different etiologies. One is a severe Broca's aphasic with a massive left hemisphere lesion. The other is a language-deprived adult, whose profound deafness was misdiagnosed as cognitive impairment, so that language exposure was delayed until well past the developmentally

180

N.F. Dronkers, C.A. Ludy and B.B. Redfern

critical language learning period. Both of these individuals have severe limitations in speaking, but seem to be competent, engaged social actors. We have attempted to document their preserved pragmatic abilities and will discuss them in the context of intercommunication. Some history and perspective The notion of pragmatics first emerged in philosophy in the effort to relate meaning to thought and communication and to understand how the units of meaning were used in real situations. The word itself is attributed to Morris (1938) who divided 'semiotics' (the 'science of signs') into three areas: (1) semantics—the content of signs (what they refer to), (2) syntax—how signs relate to one another (how they can be combined into larger units), and (3) pragmatics—how signs are deployed by one individual for uptake by another (how they are used and interpreted). Consideration of all three realms was considered necessary for a full grasp of meaning. Another philosopher, Austin (1962), launched these ideas to a wider public with an influential lecture series at Harvard in 1955. He sought to help grammarians and philosophers find a common ground in talking about language. The only way to approach a sentence had been to treat it as a verifiable description of a state of affairs, that is, as a true or false statement of fact. Austin showed that more could be made of language by considering it as performing deeds: people use words and sentences to do things. By looking at what they are trying to do, the meaning of the language units they are using becomes clearer. These ideas were further popularized and brought into linguistics by Searle (1969; 1971) andGrice (e.g., 1968; 1975). With then- discussions of 'speech acts', linguists could now talk about the difference between sentences and utterances. Relevance, presupposition, implication, and so on, could be handled in systematic ways. (See Levinson (1983) for description of pragmatics within linguistics.) The revolution promised by Austin was well under way. The tools provided by studying pragmatics have been well used in applied linguistics. Practitioners in speech pathology drew directly on Austin and Searle to develop instruments to help clinicians understand how an individual uses language, with the goal of remediating such deficits in language-disordered patients (Gallagher & Prutting, 1983; Prutting & Kirchner, 1987). In neurolinguistics, it became clear that pragmatics encompassed an area of language for which right-hemisphere involvement is critical (Joanette, Goulet, & Hannequin, 1990; Molloy, Brownell, & Gardner, 1990). Patients with unilateral right-hemisphere damage usually do well with the semantic and syntactic aspects of language, but show deficits in the context-sensitive, socially appropriate uses to which language is put in everyday situations. These applications and discoveries were made possible by defining pragmatics as a language function. Today, we have nurses and other health care providers being schooled in pragmatics as "right brain communication" (Boss, 1996) and cognitive scientists using theories of utterance interpretation to discern the influence of pragmatics on people's understanding of what is communicated by other speakers (Gibbs & Moise, 1997). Though the pragmatics revolution may be firmly launched, there is still room for growth in applying it to our paradigm of language. In particular, much of the work has focused on deficits in pragmatics. This has been useful for defining these language functions and for establishing their validity. But surely there is a positive side to pragmatic abilities, a "glass half full" vantage point for considering their communicative functions.

Pragmatics in the absence of verbal language

181

The tendency to treat pragmatics as an extension or attachment to the traditional functions of linguistics has obscured its independent reality in human interactions, the important role it plays in what could be called "nonverbal social communication". Even without the benefits of semantics or syntax, a great deal of social interaction is still possible. Pragmatics could be a window onto these important competencies. The Pragmatic Protocol (Pruning & Kirchner, 1983; 1987) For this paper, we will use the Pragmatic Protocol to describe the pragmatic abilities of our two individuals. The protocol is well recognized in the rehabilitation community as a measure of pragmatic skills in both children and adults (Sohlberg & Mateer, 1989). It is based on a 15-minute unstructured conversational sample in which the clinician can rate patients on how well they use language. The relationship between the communicative partners must be positive or neutral, so that both expect to engage in cooperative discourse. The clinician judges performance on each of 30 pragmatic behaviors, broken down into two main aspects: verbal and nonverbal. Verbal aspects of pragmatics include speech acts, topic selection and maintenance, turn taking, lexical selection, stylistic variations and paralinguistic aspects. Nonverbal aspects include kenesics andproxemics. After watching the interaction, the clinician evaluates each of the 30 behaviors as appropriate or inappropriate over the course of the episode. The criterion is whether the behavior facilitated or detractedfrom the communicative exchange. Behavior is only judged inappropriate if it could be said to penalize the individual or if it makes a difference in the interaction.

Case descriptions Patient AK (severe Broca 's aphasia) Patient AK is a 63-year-old white male who was pre-morbidly right-handed and a native English-speaker. He had some college education and had retired after a 23-year career in the Navy as a Chief Petty Officer; he was in a second career as a corporate training manager. In 1991, he suffered an extensive left middle cerebral artery infarction (see Figure 1) which left him with a dense right hemiparesis, dysarthria, apraxia of speech, and severe Broca's aphasia. His productive speech consists almost entirely of the recurring utterance /tõnõ tõnõ/ (or slight variations thereof) which permeate every attempt at verbalization, including spontaneous speech, repetition, and naming. (See Appendix A for a sample transcription1.) His comprehension is typical of Broca's aphasia, with good single word and simple sentence understanding, breaking down on complex grammatical constructions. These behaviors are quantified in his Western Aphasia Battery (WAB) scores which can be seen in Table 1.

1

A video sample of this patient's conversational skills is contained in Telerounds Program #9, "Neuroanatomical Correlates of Production Deficits in Aphasia" (Dronkers, 1993), produced by the National Center for Neurogenic Communication Disorders at the University of Arizona, Tucson.

182

N.F. Dronkers, C.A. Ludy and B.B. Redfern

Figure 1. Three-dimensional reconstruction of Patient AK's lesion. This reconstruction depicts the extent of the middle cerebral artery infarction in this patient with a severe Broca's aphasia. The upper left image is a horizontal slice similar to those seen in CT images. The upper right image depicts a coronal section, and the lower left, a sagittal section through the left hemisphere. The lower right template shows a 3-D reconstruction of the brain with the lesion apparent on the lateral surface of the left hemisphere. Date 3-92 9-92 1-97 max

Fluency 1 1 1 10

Aud. comp 5.25 6.1 7.5 10

Repetition .2 .4 0 10

Naming 4 1.4 0 10

WABAQ 17.7 21.8 17.0 100

Table 1. Summary scores obtained on the Aphasia Quotient of the Western Aphasia Battery for Patient AK. The range for Broca's aphasia on the WAB is 0-4 for fluency, 4-10 for auditory comprehension, 0-7.9 for repetition, and 0-8 for naming. AK's writing is limited to imitating samples and to writing his full name. His knowledge of numbers is better preserved, and he often uses numbers traced in the air as part of his communication (e.g., to indicate dates or specific numbers). AK's reading is severely impaired and limited to a few single words and to letters. He also has moderate-to-severe limb and buccofacial apraxia. His basic calculation and constructional praxis skills are

Pragmatics in the absence of verbal language

183

essentially intact. In spite of his limited productive speech, AK can carry a tune perfectly well, andean sing the words if he has a model to mimic. Table 2 summarizes some of the other neuropsychological test scores that reflect his language and cognitive abilities. Peabody Picture Vocabulary Test Date 6-97

Vocabulary age 4-1 Token Test

Date 6-97

Parts A-E (max correct=67) 42

Part F (max correct=96) 53

Wechsler Adult Intelligence Scale—Revised (Scaled Scores) Date

6-97

Picture completion 5

Picture arrangement 7

Block design 8

Object assembly 5

Digit symbol 3

PIQ

86

Raven's Coloured Progressive Matrices Date 3-92 9-92 1-97

Raw score (max correct=36) 22 30 32

Table 2. Sample scores obtained on other standard tests for Patient AK AK lives with his wife and has children and grandchildren nearby. He enjoys following sports, playing computer games, and building elaborate Lego models. His social life includes visits from fellow sports enthusiasts and active involvement in several stroke support groups. Patient AK'spragmatic abilities Patient AK's performance on the Pragmatic Protocol is based on three ratings of a conversation with one of the authors, and is summarized in Table 3. In terms of any inappropriate pragmatic behaviors, the ratings indicate that AK only occasionally initiates directives, queries, and comments, and that the variety of his speech acts is also greatly reduced. He rarely selects or introduces the topic of discussion, and makes little effort to change it. Thus, the partner is left with the burden of carrying the conversation. These shortcomings are not surprising in light of his restricted productive speech. AK has very little verbal language with which to establish new topics and must rely on others to take that role. In other circumstances, AK can, of course, convey that he has wants or needs, but is not always successful in conveying the content of his message.

184

N.F. Dronkers, C.A. Ludy and B.B. Redfern

Communicative act Verbal aspects Speech acts 1. Speech act pair analysis 2. Variety of speech acts Topic 3. Selection 4. Introduction 5. Maintenance 6. Change

Appropriate Responds to directives, queries and requests; acknowledges comments comments, asserts, disagrees

Inappropriate

N/A

Only occasionally initiates directives, queries, comments reduced variety rarely selects topic rarely introduces new topic

contributes to maintenance does not change topic

Turn taking 7. Initiation 8. Response 9. Repair/revision 10. Pause time 11. Interruption/overlap 12. Feedback to speakers 13. Adjacency 14. Contingency 15. Quantity/conciseness

partner carries burden responds as a listener asks for clarification normal pauses no interruptions or overlaps nods, gestures appropriately waits turn stays on topic non-verbal comments

Lexical selection 16. Specificity/accuracy 17. Cohesion Stylistic variations 18. The varying of communicative style Paralinguistic aspects 19. Intelligibility 20. Vocal intensity 21. Vocal quality 22. Prosody 23. Fluency Nonverbal aspects Kenesics and Proxemics 24. Physical proximity 25. Physical contacts 26. Body posture 27. Foot/leg hand/arm move. 28. Gestures 29. Facial expression 30. Eye gaze "not applicable/no opportunity

adjusts speech style with prosodies

normal intensity normal vocal quality normal prosody recurring utterance is at normal rate and smoothness normal distance normal contacts normal body posture normal movements normal gestures normal facial expressions normal eye gaze to observe

Table 3. Summary Pragmatic Protocol ratings for Patient AK Despite his difficulty in initiating new topics, AK does use gestures, facial expressions, and intonational variations on his recurring utterance to respond to directives, queries, and requests, and to acknowledge the comments of others. He also expresses his comments, assertions, and disagreements in this way. He contributes to maintaining the conversation by responding as a listener, and asking for clarification, again by using intonational variations, facial expression, and by nodding and gesturing. He does not interrupt and waits his turn appropriately. AK expresses stylistic variations by adjusting his speech style with prosodies. He uses stress and intonation to modify his recurring utterance. Paralinguistic aspects include

Pragmatics in the absence of verbal language

185

normal intensity of the recurring utterance with normal vocal quality, prosody, rate, and smoothness. All nonverbal aspects of AK's communicative efforts are perfectly normal. He maintains a natural distance between himself and his conversational partner, with normal physical contacts. His body posture, movements, gestures, facial expressions, and eye gaze are all also normal. Thus, AK is an example of an individual with little to no verbal output except for the recurring utterance /tõnõ tõnõ/ which he varies with changes in stress and intonation. This strategy works well for him in terms of maintaining near-normal pragmatics. Though he has difficulty initiating and changing topic, he exhibits excellent pragmatic abilities by using appropriate turn taking, stylistic variations, and paralinguistic and nonverbal skills that keep him part of the conversation. The case of "Chelsea" (language-deprived) "Chelsea" has a very different language history, although she also maintains excellent pragmatic behavior in spite of impoverished verbal language. She is a 49-year-old white female who was born the second of seven children in a rural community in Northern California. She was born with a severe to profound sensori-neural hearing loss but was misdiagnosed as mentally retarded during her childhood. Chelsea's mother knew her to be deaf and ignored professional advice to institutionalize her, raising her at home among siblings. She learned to cook and do housework and helped her mother raise the younger children. She was denied admission to local schools and a school for the deaf. As a result, Chelsea did not acquire any language or receive any formal education until the age of 32 when she was referred to a neurologist and a speech pathologist by a social worker who realized her situation. At that time, she was fitted with bilateral hearing aids and began an intensive program of oral and signed language instruction, as well as education in math and other academic subjects. Chelsea currently lives at home with her parents, and works parttime in a veterinarian's office as an assistant. Her performance on several standardized tests can be found in Table 42. Peabody Picture Vocabulary Test Date 7-80 12-80 4-81 8-81 8-81 10-81 3-82 6-82 1-83

Vocabulary age 2-3 3-2 3-11 4-3 5-5 5-3 6-10 5-11 5-8

With or without signing without without without without with with with with with

2 Numerous neuropsychological and language tests have been administered to Chelsea over the years, and cannot all be represented here. The first author can be contacted for further details.

186

N.F. Dronkers, C.A. Ludy and B.B. Redfern Token Test Date 8-82 10-86 10-87 7-89

Parts A-E (max correct=67) 65 60 52 55

Part F (max correct=96) 53 55 43 57

Wechsler Adult Intelligence Scale (Scaled Scores) Date

10-80 12-81* 10-86* 10-87*

Picture completion 10 6 12 12

Picture arrangement 4 4 4 2

Block design 6 8 8 7

Object assembly 10 16 10 7

Digit symbol 3 4 4 5

PIQ

77 84 89 84

* WAIS-R

Raven's Coloured Progressive Matrices Date 12-81 1-83 10-86 10-87

Raw score (max correct=36) 24 24 29 28

Table 4. Sample scores obtained on standard tests for Chelsea Chelsea's case addresses many interesting questions concerning the critical age for language acquisition, particularly whether it is possible to learn language after long periods of language deprivation in childhood. Her situation is analogous to those of linguistically "feral" children, such as Genie (Curtiss, 1977) or the Wild Boy of Aveyron (Itard, 1801), who did not acquire language because of lack of exposure. For Chelsea, the lack of exposure was due to severe hearing loss, but she was otherwise raised in a normal and loving family environment, contrary to previous cases. Lenneberg (1967) believed that children who did not learn language by the age of puberty would not be able to acquire it normally, while Krashen(1973) lowered this critical age to five years. The answer offered by Chelsea's case is the same one concluded by Curtiss for Genie; the critical age for language acquisition is different for the different components of language. Both Genie and Chelsea continue to develop their vocabulary, years after beginning to learn language as adults. Their knowledge of syntax, however, remains virtually absent. Chelsea's conversational style is to string words together, with no evidence of syntax or morphology. (See Appendix B for a sample transcription.) Chelsea's pragmatic abilities In the realm of pragmatics, Chelsea has developed quite normally. Her social skills are most appropriate, and she, like Patient AK, is very pleasant company. This is reflected in her performance on the Pragmatic Protocol (Table 5).

Pragmatics in the absence of verbal language

Communicative act Verbal aspects Speech acts 1. Speech act pair analysis

Appropriate

2. Variety of speech acts

Responds to directives, initiates queries and comments appropriate use and diversity

Topic 3. Selection 4. Introduction 5. Maintenance 6. Change

introduces topics tries to maintain topic makes some change in topic

Turn taking 7. Initiation 8. Response 9. Repair/revision 10. Pause time 11. Interruption/overlap 12. Feedback to speakers 13. Adjacency 14. Contingency 15. Quantity/conciseness

N/A

initiates questions responds as a listener rarely asks for clarification normal pauses some overlap when signing nods, gestures appropriately waits turn stays on topic

Lexical selection 16. Specificity/accuracy 17. Cohesion

limited, but appropriate

Stylistic variations 18. The varying of communicative style

adjusts speech style

Paralinguistic aspects 19. Intelligibility

signs are intelligible

20. 21. 22. 23.

normal intensity normal vocal quality almost normal prosody normal rate and smoothness

Vocal intensity Vocal quality Prosody Fluency

Inappropriate

187

often repeats, doesn't add not always informative

verbal responses are not always intelligible

Nonverbal aspects Kenesics and proxemics 24. Physical proximity normal distance 25. Physical contacts normal contacts 26. Body posture normal body posture 27. Foot/leg hand/arm move. normal movements 28. Gestures normal gestures 29. Facial expression normal facial expressions 30. Eye gaze normal eye gaze not applicable/no opportunity to observe

Table 5. Summary Pragmatic Protocol ratings for Chelsea The rating was based on a videotaped sample of a conversation between Chelsea and her teachers one and a half years into her training when she had considerably less language than she does now. The rating indicates that she demonstrates a variety of speech acts, responding to directives and initiating questions and comments in a normal fashion. She also introduces topics, and tries to maintain and change them. She initiates questions, responds as a listener, uses normal pauses, nods and gestures appropriately, awaits her turn in the conversation, and stays on topic. Her choice of lexical items, while limited, is appropriate, and her signs, if not her speech, are intelligible. Furthermore, she adjusts her speech style, uses normal intensity, vocal quality, rate, and smoothness, and virtually normal prosody given her hearing loss. In nonverbal aspects, Chelsea positions herself at a

188

N.F. Dronkers, C.A. Ludy and B.B. Redfern

normal distance, with normal physical contacts, body posture, movements, gestures, facial expressions, and eye gaze. The only area in which Chelsea could be rated with inappropriate pragmatics was in the realm of turn taking. Here, she rarely asks for clarification if she does not understand the content. Instead, she will often repeat what the other person has said, and does not add concise, new information on her own. This is not surprising considering her limited vocabulary and lack of experience with language. Discussion We have chosen two very different case studies to illustrate that pragmatic competence is an ability which can persist in the virtual absence of verbal language. Because both cases are largely nonverbal, this ability can be considered as independent of the verbal modality. This, we believe, is a necessary corrective to the tendency to treat pragmatics as an epiphenomenon of speech. In fact, we believe that both pragmatics and speech should be treated as phenomena of social interaction. The history of pragmatics guides us towards considering speech as social action, as accomplishing social deeds. These two cases provide further evidence that nonverbal social skills and abilities can exist in parallel to verbal language. The contrast between our two cases is also instructive. With AK, language breakdown came after the development of full social competence. His pragmatic challenges relate to the loss of propositional communicativeness. He has had to adjust his social engagements to his decreased ability to convey meaning by speaking. Chelsea, on the other hand, did not begin developing verbal language until the age of 32. Until that time, she was only able to develop such social competence as was not dependent on language. As she continues to learn language, she must also learn the attendant conversational skills. Yet her abilities as a nonverbal social actor give her the basis for developing these skills in tandem. Ultimately, her language development may reach a limit determined by the critical learning period, while her social communicative development will continue to the full potential of her personality. We began this paper by outlining the revolution that was wrought in philosophy and linguistics by the introduction of pragmatics as an aspect of understanding human communication. It should be clear that we believe this revolution has not finished running its course. There is still a predilection for installing speech at the apex of human interaction. Yet clinicians have had an insight for years, which has been hard to express, that there is more to communication than words and sentences. Our two cases, so far apart yet so similar, define a point in space of social connection. Other points in this space might be the development of verbal ability without pragmatics, or, preserved language with pragmatic breakdown. Only the total space defined by these points and others will reveal the full universe of human communicative potential. Acknowledgments—We are grateful to Peter Glusker, M.D. and Catherine O'Connor, M.A. for involving the first author in Chelsea's case and for providing some of her history and test scores.

References Austin, J. L. (1962). How to do things with words. Oxford: Oxford University Press. Boss, B. J. (19%). Pragmatics: Right brain communication. Axone, 17(4), 81-85. Curtiss, S. (1977). Genie: A psycholinguistic study of a modern-day "wild child". New York: Academic Press.

Pragmatics in the absence of verbal language

189

Dronkers, N. F. (1993). Neuroanatomical correlates of production deficits in aphasia. Telerounds Production Mr. 9, University of Arizona, Tucson. Gallagher, T., & Prutting, C. (Eds.) (1983). Pragmatic assessment and intervention issues in language. San Diego: College-Hill Press. Gibbs Jr., R. W., & Moise, J. F. (1997). Pragmatics in understanding what is said. Cognition, 62, 51-74. Grice, H. P. (1968). Utterer's meaning, sentence-meaning, and word-meaning. Foundations of Language, 4, 118. Grice, H. P. (1975). Logic and conversation. In P. Cole and J. L. Morgan (Eds.), Syntax and semantics 3: Speech acts (pp. 41-58). New York: Academic Press. Itard, J. M. G. (1801). De l'éducation d'un homme sauvage ou des premiers développements physiques et moraux du jeune sauvage de l'Aveyron. Paris: Gouyon. Joanette, Y., Goulet, P., & Hannequin, D. (1990). Right hemisphere and verbal communication. New York: Springer-Verlag. Krashen, S. (1973). Lateralization, language learning, and the critical period: Some new evidence. Language Learning, 23, 63-74. Lenneberg, E. H. (1967). Biological foundations of language. New York: Wiley. Levinson, S. C. (1983). Pragmatics. Cambridge: Cambridge University Press. Molloy, R., Brownell, H. H., & Gardner, H. (1990). Discourse comprehension by right-hemisphere stroke patients: Deficits in prediction and revision. In Y. Joanette and H. H. Brownell (Eds.), Discourse ability and brain damage: Theoretical and empirical perspectives (pp. 113-130). New York: Springer-Verlag. Morris, C. W. (1938). Foundations of the theory of signs. Chicago: University of Chicago Press. Prutting, C. A., & Kirchner, D. M. (1983). Applied pragmatics. In T. Gallagher and C. Prutting (Eds.), Pragmatic assessment and intervention issues in language (pp. 29-64). San Diego: College-Hill Press. Prutting, C. A., & Kirchner, D. M. (1987). A clinical appraisal of the pragmatic aspects of language. Journal of Speech and Hearing Disorders, 52, 105-119. Searle, J. R. (1969). Speech acts: An essay in the philosophy of language. Cambridge: Cambridge University Press. Searle, J. R. (Ed.) (1971). Philosophy of Language. Oxford: Oxford University Press. Sohlberg, M. M., & Mateer, C. A. (1989). Introduction to cognitive rehabilitation: Theory and practice. New York: The Guilford Press.

Appendix A Transcription of Patient AK's speech in conversation: Interviewer: Patient AK: Interviewer: Patient AK: Interviewer: Patient AK:

Do you like coming to Group? (referring to the weekly support group meeting) (nodding emphatically) Tõ-nõ tõ-nõ. Nõ-tõ. Tõ-tõ. Tõ-nõ tõ. Tõ-tõ. Tõ-tõ. It's a nice group of people, isn't it? Tõ-tõ-tõ-tõ. Does [your wife] enjoy it? [Your wife] enjoys sitting with the other... (shrugs) Tõ-tõ. Tõ-nõ tõ-nõ. (waves hand in air) Õõ tõ-tõ tõ-no tõ-nõ tõ-nõ tõ-nõ. (makes pushing-away gesture with hand, points to self, then waves it off) Tõ-tõ tõ-tõ tõ-tõ. Interviewer: Yeah. (Underlining indicates syllabic stress.)

Appendix B Transcription of Chelsea's speech in conversation: (Both participants sign at the same time they are speaking.) Interviewer: Chelsea: Interviewer: Chelsea: Interviewer: Chelsea: Interviewer: Chelsea: Interviewer: Chelsea: Interviewer:

(addressing second interviewer) I've told Chelsea for the last two days that I had a gift for her. Gift. From Colorado. Colorado. I remembered! (presents gift) Do you want to open it? (accepts wrapped gift, begins to untie ribbon.) Ribbon. Ribbon. What do you think it is? Think? (shakes head) A book. Think it's a book?

190

N.F. Dronkers, C.A. Ludy and B.B. Redfern

Chelsea: Interviewer: Chelsea: Interviewer: Chelsea: Interviewer: Chelsea: Interviewer: Chelsea:

Book? Don't think. Is it a blouse? Blouse? No. You...collar. Collar? Oh, a scarf. Yes, for my birthday.... (continues unwrapping, still unfolding paper) There's nothing....! tricked you! (laughs; takes out small box) Oh! Thank you! (hugs interviewer) Jewelry! That is named 'turquoise', (finger spells 'turquoise') Turquoise.

Pergamon

©

J. Neurolinguistics, Vol. 11, Nos 1-2, p. 191-206, 1998 ience Ltd. All rights reserved 1998 Published by Elsevier Science Printed in Great Britain 0911-6044/98 $19.00 + 0.00

PII: S0911-6044(98)00013-X

The use of gestures as a compensatory strategy in adults with acquired aphasia compared to children with specific language impairment (SLI) Bibi Fex and Ann-Christin Månsson Logopedic Unit, ENT Clinic, Helsingborg Hospital, Helsingborg, Sweden

Abstract—Gestures used in a confrontation naming task by four severely aphasic patients, two nonfluent, two fluent and four controls, were categorized according to Ekman and Friesen (1969). The findings were compared to an earlier study of gestures used by SLI children. There were similarities between the two language-impaired groups in how to compensate for the missing access to the word by using gestures as pragmatic cues.

Introduction Communication among individuals involves a variety of modalities. The spoken word carries the burden of communication but there are also paralinguistic channels such as intonation, facial expression, eye movement and hand and arm gestures that convey meaning and contribute greatly to communication for both speaker and listener. Gestures are natural and more or less frequent in normal conversation by speakers of different languages. Gesturing is a supplementary or alternative means for communication for both children and adults. Ekman and Friesen (1969) categorized nonverbal behaviour and found the following types: emblems, illustrators, affect displays, regulators and adaptors. Emblems have a direct verbal translation, consisting of a word or two or perhaps a phrase. Emblems are well known by all members of a group, class or culture. An emblem may repeat, substitute or contradict some part of the verbal behaviour. Illustrators are movements directly tied to speech, serving to illustrate what is being said verbally (i.e., to repeat, substitute, contradict or augment the information provided verbally); they are socially learned. There are six types of illustrators: Batons are movements which time out, accent or emphasize a particular word or phrase; Ideographs trace the itinerary of a logical journey, the direction of an idea; Deictic movements point to an object in the environment; Spatial movements depict a spatial relationship; Kinetographs are movements which depict a bodily action; Pictographs are movements which draw a picture of their referent. Affect displays have the face as the primary site for indicating emotions such as happiness, surprise, fear, sadness, anger, disgust and interest. Regulators are related to the conversational flow, the pacing of the exchange. The most common regulators include head nods, eyebrow raises, etc. A person can perform a regulator without knowing it, but if asked s/he can easily recall it. Adaptors are movements that were first learned as part of adaptive efforts to satisfy personal needs, perform bodily actions, or manage emotions.

1 Authors are listed alphabetically; both authors share equal responsibility for the paper.

192

B. Fex and A.-C. Månsson

Self-adaptors involve touching one's face, hair or lips to facilitate or inhibit sound production or speech. Hand-to-face adaptors are a rich source of information. Månsson and Lundstrom (1996) investigated 8 specific language impaired (SLI) children and 8 normal language developed children with regard to their nonverbal communication (age 3:5 to 6:7). The children were video-recorded while they were telling the pre-school teacher about their wishes for Christmas gifts. The results showed that the SLI group used 436 gestures while the normals used 318 gestures (a statistically nonsignificant difference). There were individual differences among the groups. Emblems, adaptors and affect displays were used more often by the SLI children, while the normal children used more regulators and illustrators (statistically significant). The authors concluded that children develop gestures in the following order: affect displays—>adaptors—>emblems —> regulators—>illustrators The SLI children used more pictographs and deictic movements while the normal children used more batons and spatial movements (statistically significant). Månsson and Lundstrom (1996) suggested that these subgroups of illustrators are developed in the following order: deictic movements—>kinetographs—>pictographs—>batons—>spatial movements. Observing gestures in children adds knowledge that can be used diagnostically, but also prognostically. Children with more advanced gestures such as, for example, pictographs might have a better prognosis than children who do not use pictographs in their nonverbal behaviour. Soderbergh (1980) described mother-child dialogues during play activity. Verbal, somatic and vocal language was observed. The verbal channel covered all verbal language, the vocal channel included communication that was signalled by voice, e.g., pitch, stress, etc. The somatic channel comprises communication that is signalled via the face, body and posture. The author says that verbal language is developed on the basis of body language, and that adult-child dialogue develops towards normal adult-adult dialogue. Somatic (body language) communication continues to be used by the adult speaker, not only as an accompaniment to speech, but as an independent and necessary part of dialogue. A complete model of dialogue must therefore include the somatic component. A listener must be somatically active. The listener reacts continuously to the speaker by posture and body movements and mimics and responds to the speaker. Similarly, Vygotsky (1978) claimed that children use gestures and words at an early age to get their needs and wishes satisfied. He also stated that children use gestures to regulate others and thereby learn to regulate themselves. In that way, children take an active role in their own cognitive development. McNeill and Levy (1982) divide nonverbal communication into three systems of gestures: iconic (pictographs according to Ekman and Friesen), metaphoric (depicting a more abstract concept) and beats (batons). They are of the opinion that icons depict whole scenes, while extranarrative statements lacking the sequentiality constraint tend to be accompanied by more formless gestures or beats. The theory traces the sensory-motor images of speakers back to the ontogenetically primitive stages of early childhood and shows how they continue to play a role in generating language for adult speakers. At age 9 or 10, deictics and iconics are well developed but beats are still quite rare. Jancovic, Devoe and Wiener (1975) also point out that gestures become more frequent and more complicated as the child's language develops. Wilkinson and Rembold (1981) suggest that, as the child's language improves, gestures complement and enrich the verbal language. Deictics are used to give a message when verbal ability is not well enough developed and these gestures diminish with age. By contrast, pictographs expand a verbal

Gestures as a compensatory strategy

193

message. Depending on age and language complexity, the use of pictographs expands. Sanmarco (1984) observed how mothers and children interacted when they tried to solve a problem in copying a model. SLI and normal language children from 3 to 11 years old were studied. It was noted that the SLI children used deictic gestures more than normal children. Early aphasia therapy focused on re-educating linguistic skills. Dean and Skinner (1995) conclude that, when the therapist profiles the communication performance of aphasic clients in addition to results from assessments, remediation will be more balanced. Lately it has been suggested that the assessment of aphasia not only give the status of the spoken and written language, but also the person's complete communicative ability. A series of papers have provided support for the argument that nonverbal performance should be profiled in aphasia assessment. Some studies have shown that aphasic patients can use nonverbal communication to a greater degree than spoken or written language. Duffy, Duffy and Pearson (1975) claimed that there was one central communication and symbol system in human beings and that damage to this mechanism would result in lowered performance on all symbolic tasks. Aphasic individuals would not be able to use gestures. The observation that non-verbal communication (NVC) is used for structuring verbal production, i.e., for semantic planning, and the finding that this could be seen in Wemicke's aphasics, in that the amount of gesturing increased when they had problems with semantic planning, are also highly relevant for the interpretation of NVC in aphasics, according to Delis et al. (1979). Cicone et al. (1979) studied spontaneous gestural production and spoken language in four aphasic subjects, two Broca's aphasics and two Wemicke's aphasics. In general the gestures of the patients paralleled their speech output. The total amount and nature of gesturing produced by each patient was investigated. Posterior aphasics produced more movements than anteriors. Anterior aphasics produced a higher percentage of head movements as well as movements of the whole body and the torso. No differences were revealed between the aphasic groups in measuring hand configuration and orientation. The difference in the number of movements between the groups seems unremarkable. The results of the Wemicke's aphasics more closely resembled those of the normal controls, but the aphasics' language and gestures lacked clarity. Their gestures were often elaborate but generally unclear and confusing like their speech. In Broca's aphasics, a large percentage of iconic gestures occurred; they produced simple and unelaborate units in both modalities. Their output was sparse, but their gestures were generally informative and clearly intelligible. Wemicke's aphasic patients used many more pointing gestures in their communication. These results do not support the claim that aphasics spontaneously improve their communication by gestures and the authors say that their findings can be interpreted as evidence in favour of a central organizer which controls the critical features of communication, irrespective of the modality of expression. Spontaneous gestures used by 20 aphasic patients during a verbal confrontation naming task were studied by Helm (1979). Broca's aphasics produced at least one gesture in association with the greatest number of the 40 pictured items. The Wemicke's patients rated second to the Broca's group in the frequency of their gesturing but the quality of their gestures exceeded all other groups. The results showed that different groups of aphasic patients have differential capacities for producing self-initiated gestures. According to Helm, an evaluation of gestural skills should include both an apraxia examination and a tool for noting spontaneous use of gestures. A study by Ahlsen (1985) clearly shows that aphasics have a much higher use of NVC than non-aphasics. Even interindividual variation was greater in the aphasic group. The results were summarized as follows: The aphasics used: (a) much more NVC illustrating

194

B. Fex and A.-C. Månsson

factual information; (b) more NVC showing positive and negative feelings; (c) more NVC for turn keeping, turn giving and feedback eliciting; (d) less NVC for turn taking (in fact none). There were no differences between the aphasics and the controls in the amount of use of (a) NVC showing hesitation, b) feedback giving, (c) emphasis, d) emblems for "yes" and "no". The aphasics used a substantial amount of NVC, especially pointing, which was often used without verbal accompaniment. Ahlsén (1991) also studied nonverbal communication in a patient with Wernicke-type aphasia and found that he used pointing gestures and illustrators a lot more than other aphasics and controls. Le May, David and Thomas (1988) found that hand gestures are used most by Broca's aphasics and least by non-aphasic controls. Batons, ideographs and deictic movements were associated with Broca's aphasics, while both Wemicke's and Broca's aphasics used kinetographs more than the controls. These authors raise the question of whether therapy should encourage or teach Broca's aphasics to use more batons, ideographs and deictic movements as compared to their use of pictographs. Would an increased use of pictographs increase the clarity of their communication? Hermann, Reichle and Lucius-Hoene (1988) found that aphasic patients produced almost twice as many verbal utterances as their healthy counterparts. The patients produced repetitive utterances without value, e.g., automatisms and perseverations. The researchers calculated the percentage of nonverbal elements over the total number of verbal and nonverbal communicative actions. Aphasics used fewer speech-focused movements and significantly more codified gestures than their counterparts. There was no difference in the use of descriptive gestures. Holland (1982) points out that aphasics communicate better than they speak. There is little doubt that aphasia is often associated with nonverbal disorders according to Feyereisen (1988). The fact that some cases of aphasia are explained by general conceptual deficits and others by modality-specific disorders may suggest that an intermediate position should be taken with regard to the existence and/or role of a central organizer. Feyereisen (1983) also noted an increase in the ratio of gestures per word in aphasic patients. These observations suggest a functional role for iconic gestures in word retrieval. McNeill (1985) said that gestures dissolve along with speech in aphasia. Butterworth and Hadar (1989) point out that, if gesture is to be linked to one or more computational stages of speech production, the model must include the following stages: Stage 1: Pre-linguistic message construction; Stage 2: Determination of the grammatical form; Stage 3: Selection of the lexical items in abstract form from a semantically organized lexicon; Stage 4: Retrieval of phonological word forms on the basis of Stage 3; Stage 5: Selection of prosodic features including the location of sentence stress points; Stage 6: Phonological stage in which word forms are ordered syntactically and prosodic features marked; Stage 7: Full phonetic specification with all timing parameters specified; Stage 8: Instructions to articulators. In this model, gestures and speech are autonomous. The authors claim there is no reason why one should not occur without the other. Kendon (1986) was of the opinion that gestures can throw light upon the nature of thought. Slama-Cazacu (1976) believed in a "mixed syntax" in which gesture and speech are used in alternation, with gestures sometimes "filling in" when speech had ceased. Glosser, Wiener and Kaplan (1986) found that moderate aphasics produce significantly fewer semantic, modifying and relational gestures and more pantomimic and deictic gestures than control subjects or mild aphasics, who did not differ from each other on these measures. The more severely impaired aphasic subjects used fewer complex communicative gestures while producing proportionally more of die nonspecific, nonconsensually shared, unclear gestures. Geschwind (1975) proposed that the left hemisphere is dominant not only for speech but also for learned movements. He said that many Broca's aphasics are unable to carry out commands with the non-paralysed left extremities. Hanlon, Brown and

Gestures as a compensatory strategy

195

Gerstmann (1990), on the other hand, found that, in their study, functional activation of the hemiplegic right arm in the production of communicative gestures favourably affects the naming capacity of nonfluent aphasics. By activating the motor system of the hemiplegic side, the production of deictic gestures had a facilitative effect on the naming performance of nonfluent aphasics but not fluent aphasics. Duffy and Duffy (1981) stressed the active role of the left hemisphere, particularly in more complex symbolic nonverbal communication such as pantomime. For example, nonfluent subjects tend to present with a low pitch, monotonous intonation, poor intensity and quality control and a slow rate. On the whole, fluent subjects did not do this. Prosodic features, for example, intonation and rate, were thus used in a compensatory fashion on many occasions. Behrmannand Perm (1984) found that gesture, pantomime and facial expression were also frequently actively used as a compensatory strategy by nonfluent subjects. In the fluent patients, however, such actions seemed to be coincidental concommitants of the verbal message. Bosone (1977) found that, following group therapy, 4 of 17 aphasic patients self-initiated gestures for communication outside the clinical setting; 4 did not learn the gestures and 6 learned but did not transfer the use of gestures. Goodglass and Kaplan (1963) studied the gestural performance of 20 aphasics. The aphasic group not only showed impaired gestural performance but failed to improve on imitation trials of gestural tasks. The authors interpreted the gestural abnormality in aphasia as an apraxic disturbance. The aim of this study The aim of this study was to investigate how two nonfluent and two fluent aphasics would use different kinds of spontaneous gestures to initiate, accompany or substitute for missing words on a confrontation naming task where they were shown drawings representing nouns. Could these aphasic adults cue the target word by using gestures? Were their gestures the same or of the same kind as those used by SLI children? Would it be possible to suggest some type of gesturing as a means of therapy parallel to speech therapy by using a paralinguistic channel in order to gain access to lexical semantics? Materials and method Subjects Four Swedish-speaking post-CVA aphasic adults, 3 males and 1 female, were chosen for this investigation. They were two nonfluent and two fluent aphasics. Their ages ranged between 70 and 79 years and they exhibited no gross sensory loss or dysarthria. None of them had any motor disorder except for one nonfluent aphasic who had a paralysed right arm (all of them were right-handed premorbidly) and used her left arm and hand instead. The nonfluent aphasics had anterior lesions in Broca's area, and the fluent aphasics had posterior lesions in Wernicke'sarea. None of the subjects had any visual deficit, but two of them hadpresbyacusis and wore hearing aids.

196

B. Fex and A.-C. Månsson

Language and speech The two nonfluent subjects used only one-word utterances. Their speech rate was low, sentence length was short, and the melodic contour was lost. Their speech production was effortful, with more pauses than actual words. They had word finding problems but good comprehension. The nonfluent male subject had verbal apraxia in combination with his aphasia. However, he had no signs of limb apraxia, and could use both hands with different tools and when making gestures. The nonfluent female subject had a moderate tendency towards perseveration, which did not occur in the nonfluent male subject. Both subjects could read and understand short sentences, but were not capable of reading stories due to their memory deficit. The male subject was rather good at writing, but his spelling was not perfect. He did not use writing as a substitute means of communication. The nonfluent female subject did not like to write with her left hand, but used it in daily living activities. The fluent aphasics had fluent, empty, often neologistic speech, verbal and phonemic paraphasia and poor comprehension. Verbal repetition was impaired. None of the fluent aphasics were helped by phonemic or semantic cues, but written cues such as letters and words did help. When the patient was searching for the target word, the prosodic features were unclear, but when the word was found, prosody became clear and the voice louder. These patients had perfect articulation. The speech of the fluent aphasics was more abundant than normal speech. The amount and rate of word production was higher than in normals. The melodic contour of the sentences approximated normal speech. The four aphasic subjects were tested with the Swedish Screening Test for Aphasia, and were diagnosed as severely aphasic. They had all received language therapy in the interval since onset but no specific training in gestures. Four Swedish-speaking adults without any history of CVA were chosen as controls; their ages ranged between 64 and 78 years. Nonfluent aphasic E, male Nonfluent aphasic F, female Fluent aphasic K, male Fluent aphasic L, male

Time since onset 36 months 34 months 9 months 38 months

Age 77 years 72 years 79 years 70 years

Method All subjects were presented with a confrontation naming task, the SBP Skånes Benamnings Prövning by Apt (1994), which consists of 64 drawn items representing nouns. All subjects were videorecorded. There was no time limit, in order to allow the aphasics to feel relaxed and take the time they needed. The important factors were the types and number of gestures related to all 64 items, and also the number of verbal trials, not the time used for testing. The mean time for testing the controls was only two minutes. Nonfluent aphasic E Nonfluent aphasic F Fluent aphasic K Fluent aphasic L

Time for test 22 minutes 35 minutes 54 minutes 30 minutes

Gestures as a compensatory strategy

197

The authors examined the videotapes several times frame by frame. Gestures were categorized according to Ekman and Friesen's (1969) system. The number of verbal trials, the number and type of gestures, and the intervals when the gestures were used, i.e., prior to, simultaneously or after the word, were examined. Results Nonfluent aphasia E Yes or no are not adequately used. Head nodding and shaking are not adequately used either. Pictographs are usually but not always clear. The patient uses gestures to compensate and substitute for words and also to accompany word trials. Nonfluent aphasic F As her right arm is paretic, the patient points with her left arm and hand. She is generally helped by her pointing (and also by phonetic prompting). She uses her left index finger, holding it up in the air to point out turn keeping. When hesitating, she moves her fingers up to her lips. When she finds the correct word in her lexicon, and while pronouncing it, her voice is more intense and her prosody and intonation, more emphatic. Her gestures are not very clear; sometimes they are difficult to interpret. After she is unable to find a word she shakes her head twice in succession. Her many pointing gestures are not used to compensate or substitute for the target word but to elicit it. The two ideographs she used indicate the direction of her thought (idea). Fluent aphasic K All his gestures were clear and easy to understand. His few gestures were used to compensate for the word in the beginning and thereafter to accompany it. The target word is characterizedby many literal paraphasias and numerous trials. Prosody is emphatic and the intensity of the voice is raised when the patient finally hits the right target word.

Fluent aphasic L This patient's gestures are clear. It is easy to understand the referent. Prosody is emphatic and the intensity of the voice is raised when the patient hits the right target word. The total number of gestures produced by all 4 aphasics was 412, compared to 1 gesture producedby the normals. The percentage of gestures used by all aphasics is shown in Figure 1.

198

B. Fex and A.-C. Månsson

Figure 1. Gestures used by all aphasics The nonfluent aphasics used a total of 240 gestures and the fluent group used 172, compared to the controls' 1. A statistical analysis using chi-square tests was done. The chisquare test for 2 x 4 contingency tables of the distribution of nonfluent versus fluent aphasic subject emblems, illustrators, affect displays and self-adaptors had a significance of p < 0.05. The chi-square test for 2 x 5 contingency tables of the distribution of nonfluent versus fluent aphasic subjects' illustrators, batons, deictics, spatials, pictographs and ideographs had a significance of p < 0.01.

Figure 2.

Verbal and nonverbal trials in 1: Nonfluent E, 2: Nonfluent F, 3: Fluent K, 4: Ruent L, and 5: Controls

The number of verbal trials made by all aphasics was 744, compared to 256 for normals (Figure 2). The chi-squaretest for 3 x 2 contingency tables for the nonfluent versus fluent aphasic subjects and for the aphasics versus the controls with respect to the distribution of verbal and nonverbal trials had a significance of p < 0.0001.

Gestures as a compensatory strategy

199

Pictographs Pictographs were used by the nonfluent and fluent aphasics, prior to, simultaneously and after their verbal trials. The nonfluent aphasics produced 37 pictographs prior to verbal trials, compared to the fluent aphasics who only used 5. The fluent aphasics produced 68 simultaneous pictographs, compared to only 16 in the nonfluent aphasics. Pictographs produced after verbal trials totalled 6 in the nonfluent aphasics and only 1 in the fluent aphasics.

Figure 3. Pictographs used by the nonfluent aphasics

Figure 4. Pictographs used by fluent aphasics The chi-square test for 2 x 3 contingency tables of the distribution of nonfluent versus fluent aphasic subjects' pictographs used prior to, simultaneously and after words (Figure 4) had a significance of p < 0.0001.

Deictics The nonfluent aphasics produced 65 prior deictics (Figure 5) and the fluent aphasics, only 2 (Figure 6). Simultaneously used deictics amounted to 10 in the nonfluent aphasics and 20 in the fluent aphasics. Deictics used after word trials totalled 2 in the nonfluent and only 1 in the fluent aphasics.

200

B. Fex and A.-C. Månsson

Figure 5. Deictics used by the nonfluent aphasics

Figure 6. Deictics used by the fluent aphasics The chi-square test for 2 x 3 contingency tables of the distribution of nonfluent versus fluent aphasic subjects' deictics used prior to, simultaneously and after words had a significance of p < 0.0001. Batons Only one baton was used prior to verbal trials by the nonfluent aphasics and the fluent aphasics did not use any. The nonfluent group used 16 simultaneous batons and the fluent group, 2. After verbal trials, 10 batons were produced by the nonfluent aphasics and only 3 by the fluent aphasics (Figures 7 and 8).

Gestures as a compensatory strategy

201

Figure 7. Batons used by the nonfluent aphasics

Figure 8. Batons used by fluent aphasics The chi-square test for 2 x 3 contingency tables of the distribution of nonfluent versus fluent aphasic subjects' batons used prior to, simultaneously and after words resulted in p = 0.6032, which is not significant.

Spatials Spatials were only used simultaneously with the verbal trials; they were produced twice by the nonfluent aphasics and 4 times by the fluent aphasics. The chi-square test of 2x3 contingency tables of the distribution of nonfluent versus fluent aphasic subjects' spatials used prior to, simultaneously and after words could not be calculated because of the many results of zero obtained in the testing. Discussion This study found that aphasic subjects with acquired brain damage use the same type of gestures as do SLI children. This might indicate that, when people do not have access to a word, they naturally use a paralinguistic channel, in this case, gestures. Illustrators seem to be the gestures used to get access to the word. The two nonfluent aphasics used more

202

B. Fex and A.-C. Månsson

pointing gestures than the fluent aphasics, who used more pictographs. The nonfluent aphasics used more pointing gestures and pictographs prior to uttering words indicating that these gestures play a functional role in word retrieval. The fluent aphasics use the same kinds of gestures simultaneously in order to pick out the right target word from their flow of speech. Batons are more frequently used by the nonfluent aphasics (N=27) than the fluent aphasics (N=5). Subject Nonfluent E used 16 batons simultaneously with his word trials; the gestures seemed to serve as pragmatic cueing to find the right word. A number of authors (e.g., Duffy et al., 1975; Cicone et al. 1979; Duffy & Duffy, 1981; Glosser et al., 1986) argue in favour of a central organizer which controls both language and gesture. Thus, both verbal and gestural communication should be impaired in aphasia. Feyereisen (1988) and Christopoulou and Bonvillian (1985) argue against a central symbolic deficit, stating that the more severe the level of apraxia and aphasia, the more gestures are produced by aphasics. However, Feyereisen (1988) also shared the opinion that aphasia was associated wuli nonverbal disorders. He claimed that some cases of aphasia were explained by conceptual deficits and others by modality-specific disorders, thereby staking out an intermediate position in the discussion concerning a central organizer. He suggested a functional role for iconic gestures in word retrieval. Butterworth and Hadar(1989) considered that gestures and speech are globally different and autonomous and that gesture and speech often occur one without the other. This statement contradicts McNeill's (1985) opinion that gestures dissolve together with speech in aphasia. Based on our data, our opinion with regard to the role of a central organizer favours the middle ground. The central organizer is the concept. The different paralinguistic channels, for example, one that leads to the phonological/phonetical representation and another that leads to gestural behaviour, are autonomous but seem to originate from the same centrally organized concept. Emblems Emblems have a direct verbal translation. In this study, the nonfluent aphasics produced 32 emblems, the fluent aphasics 11 and the controls, none. Head nods and head shakes were used more frequently by the nonfluent aphasics as a substitute for "yes" and "no", but not adequately in the case of the male nonfluent. He used this emblem to substitute for or contradict his faulty speech production. His head nods gave supplementary information, confirming to the listener what was right or wrong. In their study of SLI children compared to normals, Månsson and Lundström (19%) found a significant difference in the distribution of the different categories of gestures. The children used the same types of gestures but interesting differences in how they used them were found. The SLI children produced more emblems, adaptors and affect displays and the normals, more regulators and illustrators. Emblems like headnods and head shakes get the listener's attention and are easy to use for the SLI children. Emblems were used parallel to or instead of verbal expressions. The authors concluded that the SLI children were at a more immature stage than the normal children in their language development and therefore used gestures that also seemed to be more immature. It appears that both SLI children and nonfluent aphasics prefer to use emblems in the absence of semantic access.

Gestures as a compensatory strategy

203

Illustrators Blustrators are socially learned movements that are directly tied to speech. They illustrate what is said and repeat, substitute, contradict or augment the information given verbally. The controls in this study did not use any of these movements since they were administered a confrontation naming task. Normals use illustrators in ordinary conversation. The nonfluent aphasics seem to use illustrators as elicitors, while the fluent aphasics use them as a guideline in picking the right word out. Månsson and Lundstrom (1996) found that SLI children use more pictographs than normal children do to compensate for their language impairment. SLI children need to use illustrators socially as they lack experience and are often misunderstood when communicating; SLI children's parents are uncertain about what messages they are receiving, and thus need more knowledge about their children's NVC. Deictics Deictic movements are pointing gestures. This study found that deictics are used more often by nonfluent aphasics. The nonfluent female aphasic subject frequently used deictics prior to uttering a target word and continuously pointed to the object in question. This seems to be a case of pragmatic cueing, i.e., pointing in order to get access to the word. Hanlon et al. (1990) suggested that pointing with the right hemiplegic limb at the picture to be named would help nonfluent aphasics in a confrontation naming task by functionally activating the proximal motor system to reach early levels in speech production. Our nonfluent aphasic with a paralysed arm did not do this but used her left, non-paralysed, arm only. This kind of stimulation most probably needs specially designed therapy. Deictic gestures are more often used simultaneously by the fluent aphasics. They recognize and point to the picture, and thereby seem to reinforce the word and its phonetic prompting while they make several trials to find the right word. Wilkinson and Rembold (1981) showed that deictics are used when verbal ability is missing in small children. They also indicated that deictics help to develop language ability. Pictographs Pictographs are movements which draw a picture of their referent. In the present study, the fluent aphasic group used more simultaneous pictographs (68) in their attempts to find the target word than the nonfluent aphasics, who used only 16 pictographs, and the controls with 0. This fact is in agreement with Feyereisen's (1988) opinion of iconic gestures. Le May et al. (1988) suggest that Broca's aphasics should be encouraged to use batons, ideographs and deictics more than pictographs to clarify their communication. Our nonfluent aphasics, however, used more pictographs prior to the word (37) than did the fluent aphasics (5); this could indicate that they used the pictographs as a pragmatic cue in order to access the lexicon. In the fluent group, simultaneous pictographs seem to be used as a support and a guide to facilitate word retrieval in their copious flow of speech. By using a pictograph prior to the word and during the latency period in the word finding process, nonfluent aphasics trigger the word and thus they should be encouraged to use such gestures in therapy. Overall, pictographs were used more frequently by the fluent aphasics.

204

B. Fex and A.-C. Månsson

In comparison, SLI children were found to use pictographs more frequently than normal children to substitute for a missing word or to clarify matters for the listener when their phonology was incorrect or incomplete. Children who have a lexical semantic impairment are not likely to use pictographs to the same extent as children with only phonological disorders. Batons Batons accent or emphasize a particular word and phrase. The nonfluent aphasics used more batons than the fluent aphasics. The nonfluent and dyspraxic subject E used most of his batons simultaneously with his verbal trials while the other nonfluent subject, K, used most of her batons after the word as an accentuation feature together with a louder voice and well pronounced prosody. When the word is correctly presented, the baton stresses this fact. This can also be regarded as pragmatic cueing. The controls used none of these gestures. SLI children were not found to use batons at the same rate as normal children who had a certain knowledge about language and prosody. The SLI children, lacking this knowledge, could not use batons, which seem to be more advanced gestures. Ideographs Ideographs are gestures that trace the itinerary of a logical journey. Such gestures were only used twice by one of the nonfluent aphasics, and not by the fluent aphasics or the controls. The SLI children studied by Månsson and Lundström (1996) did not use these gestures either. As our test was a confrontation naming test, it did not encourage the subjects to use ideographs. Even though the nonfluent aphasic who used ideographs did not find the word she sought, she showed that she understood the logic of the process by using this gesture, which appears to be more advanced than the others. Spatial movements Spatial movements depict a spatial relationship. One fluent aphasic used 4 such movements simultaneously with word production and one nonfluent used 2. None of the controls used any as they had access to the word and did not need to indicate any spatial relation to it. The SLI children produced no such gestures but the normally speaking children used a few.

Affect displays Affect displays show emotions (e.g., happiness, surprise, fear, sadness, anger, disgust and interest.) Twelve of these were produced by the fluent subjects and 8 by the nonfluent subjects. These gestures appeared when the subjects were frustrated with their word finding problems; the same was true of the SLI children.

Gestures as a compensatory strategy

205

Self-adaptors Self-adaptors involve touching one's face, hair or lips to facilitate or inhibit speech production. According to Ekman and Friesen (1969), hand-to-face adaptors are a rich source of information. In this study, the nonfluent aphasics produced 33 self-adaptors in comparison to the fluent aphasics' 43 and the controls' 1. The self-adaptors in all three groups consisted of hand-to-mouth or -nose movements, when the subjects were searching for the target word. One of the nonfluent aphasics used the touching of her lips as a pragmatic cue. The two fluent aphasics touched the nose or ear while choosing the right word. Self-adaptors are more frequently used by SLI children than by normally speaking children. These gestures might be interpreted as a support in eliciting the word or as a sign of discomfort when one is not able to verbalize a need. Conclusion In this study and the one conducted by Månsson and Lundström (1996), aphasics and SLI children were found to use gestures to compensate for their language problems. In the clinical setting, when assessing developmental or acquired language disability, we strongly recommend also doing an analysis of the gestures that accompany or replace speech. The nonverbal behaviour research by Ekman and Friesen (1969) provides a good categorization of gestures. Based on this study, clinicians can describe the gestures of their patients in detail and also examine their pragmatic cueing in the absence of the spoken word. To judge from this study, it is advisable to encourage the natural use of gestures in aphasia therapy. Clinicians working with aphasic individuals often see a discrepancy between their abilities as measured by standard assessments or tests, and the communicative possibilities revealed when the patients want to "talk" about matters important to them. More comprehensive studies are required if the relationship between verbal and nonverbal behaviour is to be fully understood. Acknowledgment—The authors are indebted to Osamu Shiromoto, Dept. of Communication Disorders, Hiroshima Prefectural College of Health and Welfare, Japan, for his help with the statistics.

References Ahlsén, E. (1985). Discourse patterns in aphasia. Gothenburg Monographs in Linguistics 5. Department of Linguistics, University of Göteborg, Göteborg, Sweden. Ahlsén, E. (1991). Body communication as compensation for speech in a Wernicke's aphasic—A longitudinal study. Journal of Communication Disorders, 24, 1-12. Apt, P. (1994). SBP Skånes Benämningsprövning. Standardized revision based on Naming. Escape. EU project. Behrmann, M., & Penn, C.1984. Non-verbal communication of aphasic patients. British Journal of Disorders of Communication, 19, 155-168. Bosone, Z. (1977). Paper presented to the Amerind Conference, St. Lewis, cited in M. Skelly, 1979. Amerind gestural code based on universal American Indian hand talk. New York: Elsevier, North-Holland. Butterworth, B., & Hadar, U. (1989). Gesture, speech, and computational stages: A reply to McNeill. Psychological Review, 96, 168-174. Christopoulou, C, & Bonvillian, J. D. (1985). Sign language, pantomime, and gestural processing in aphasic persons: A review. Journal of Communication Disorders, 18, 1-20. Cicone, M., Wapner, W., Foldi, N., Zurif, E., & Gardner, H. (1979). The relation between gesture and language in aphasic communication. Brain and Language, 8, 324-349. Dean, A., & Skinner, C. (1995). Perceptual assessment in aphasia. In S. Wirtz (Ed.), Perceptual Approaches to Communication Disorders (pp. 84-99). London: Whurr Publishers.

206

B. Fex and A.-C. Månsson

Delis, D., Hamby, S., Gardner, H., & Zurif, E. (1979). A note on temporal relations between language and gestures. Brain and Language, 8,350-354. Duffy, R. J., & Duffy, J. R. (1981). Three studies of deficits in pantomimic expression and pantomimic recognition in aphasia. Journal of Speech and Hearing Research, 24, 70-84. Duffy, R. J., Duffy, J. R., & Pearson, K. (1975). Pantomime recognition in aphasics. Journal of Speech and Hearing Research, 18, 115-132. Ekman, P., & Friesen, W. B. (1969). The repertoire of nonverbal behaviour: Categories, origins, usage and coding. Semiotica, 1, 49-98. Feyereisen, P. (1983). Manual activity during speaking in aphasic subjects. International Journal of Psychology, 18, 545-546. Feyereisen, P. (1988). Non-verbal communication. In F. C. Rose, R. Whurr, and M. A. Wyke (Eds.). Aphasia (pp. 46-88). London: Whurr Publishers. Geschwind, N. (1975). Theapraxias: Neural mechanisms of learned movements. American Scientist, 63, 188195. Glosser, G., Wiener, M., & Kaplan, E. (1986). Communicative gestures in aphasia. Brain and Language, 27, 345-359. Goodglass, H., & Kaplan, E. (1963). Disturbances of gesture and pantomimes in aphasia. Brain, 86,703-720. Hanlon, R. E., Brown, J. W., & Gerstman, L. J. (1990). Enhancement of naming in nonfluent aphasia through gesture. Brain and Language, 38, 298-314. Helm, N. A. (1979). The gestural behavior of aphasic patients during confrontation naming. Doctoral dissertation, Boston University. Hermann, M., Reichle, T., & Lucius-Hoene, G. (1988). Nonverbal communication as a compensative strategy for severely nonfluent aphasics. A quantitative approach. Brain and Language, 33,41-54. Holland, A. L. (1982). Observing functional communication in aphasic adults. Journal of Speech and Hearing Disorders, 47, 50-56. Jancovic, M., Devoe, S., & Wiener, M. (1975). Age-related changes in hand and arm movements as nonverbal communication. Some conceptualization and an empirical exploration. Child Development, 46, 922-928. Kendon, A. (1986). Some reasons for studying gestures. Semiotica, 62,3-28. Le May, A., David, R., & Thomas, A. P. (1988). The use of spontaneous gesture by aphasic patients. Aphasiology,2, 137-145. Månsson, A.-C., &Lundström, C. (1996). En Jämförelse mellan normalspråkiga och språkstörda barns ickeverbala kommunikation. Unpublished Master's dissertation, Lund University, Institution of Logopedics and Phoniatrics, Lund, Sweden. McNeill, D. (1985). So you think gestures are nonverbal? Psychological Review, 92,350-371. McNeill, D., & Levy, E. (1982). Conceptual representation in language activity and gesture. In R. J. Jarvella and W. Klein (Eds.), Speech place and action: Studies in deixis and related topics (pp. 271-295). Chichester: John Wiley. Sanmarco, J. G. (1984). Joint problem-solving activity in mother-child dyads: A comparative study of normally achieving and language disordered preschoolers. U-M-I Dissertation Information Service. Slama-Cazacu, T. (1976). Nonverbal components in message sequence: 'Mixed syntax'. In W. C. McCormack and S. A. Wurm (Eds.), Language and man: Anthropological issues (pp. 14-29). The Hague: Mouton. Söderbergh, R. (1980). En modell for beskrivning av dialoger mellan barn och vuxna. 1 (Språkstimulering i förskoleåldern. Stockholm, Child Language Research Institute, Paper 5). Vygotsky, L. (1978). Mind in society. The development of higher psychological processes. Cambridge, MA.: Harvard University Press. Wilkinson, L. C., & Rembold, K. L. (1981). The form and function of children's gestures accompanying verbal directives. In P. S. Dale and D. Ingram (Eds.), Child language: An international perspective (pp. 175190). Baltimore: University Park Press.

J. Neurolinguistics, Vol. 11, Nos 1-2, p. 207-221, 1998 © 1998 Published by Elsevier Science Ltd. AH rights reserved Printed in Great Britain 0911-6044/98 $19.00 + 0.00

PII: S0911-6044(98)00014-1

Relationship between language impairment and pragmatic behavior in aphasic adults Jan R. Avent*, Robert T. Wertz* and Linda L. Auther* 'California State University, Hayward, California; Veterans Administration Medical Center, Nashville, Tennessee

Abstract— We explored the relationship between language impairment and pragmatic performance in aphasic adults during the first 48 weeks postonset. Twenty patients, ten fluent and ten nonfluent, who had suffered a single left hemisphere stroke participated in the study. Half of the patients received 44 weeks of individual treatment and half received 44 weeks of group treatment. Language impairment— auditory comprehension, reading, oral-expressive language, and writing— was determined with the Porch Index of Communicative Ability at foui weeks postonset, pretreatment, and at 15, 26, 37, and 48 weeks postonset. Pragmatic performance— verbal aspects, paralinguistic aspects, and nonverbal aspects — was assessed during conversation with a pragmatic protocol at the same points in time. The patients made significant improvement in both language impairment and pragmatic performance between 4 and 48 weeks postonset, and there were no significant differences in improvement between the fluent and nonfluent groups or between patients receiving individual or group treatment. In our fluent aphasic patients, language impairment and pragmatic performance were significantly related at four weeks postonset but not thereafter. Conversely, in our nonfluent aphasic patients, language impairment and pragmatic performance were not significantly related at four weeks postonset, but they were significantly related at 15, 26, 37, and 48 weeks postonset. The significance of the relationship between language impairment and pragmatic performance was sporadic over time in the individual and group treatment groups. Introduction Traditional assessment of aphasia has focused on language impairment. It is designed to determine whether a language problem exists, document specific deficits, plan treatment, and measure change in performance over time (Lomas, Pickard, Bester et al. 1989). During the past 20 years, however, assessment of aphasia has expanded to include assessment of communicative competence (Holland, 1996; Newhoff & Apel, 1997). The impact of aphasia on communication is evident, particularly, in conversational ability (Prins, Snow, & Wagenaar, 1978). However, measures of improvement in conversational ability have been ignored in most aphasia treatment studies. Because conversation is one of the most important aspects of language use (Levinson, 1983) and basic to communication in a social world (Brinton & Fujiki, 1989; Smith, 1985), its inclusion in the evaluation of treatment effects appears important. One measure of communicative competence is pragmatics; the study of how utterances have meaning in situations (Leech, 1983). Pragmatic assessment of aphasia is appealing, because it is conducted within a social context where linguistic, paralinguistic, and nonverbal skills are observable. Pragmatic performance can be assessed by observational

208

J.R. Avent, R.T. Wertz and L.L. Auther

profiles, communicative efficiency measures, standardized testing in real or simulated life situations, family and significant others questionnaires, and composite assessments (Manochiopinig, Sheard, & Reed, 1992). One striking observation that has resulted from the assessment of communicative competence is that individuals with aphasia, generally, display more intact pragmatic skills than they do in the more formal aspects of syntax and lexical components of language (Avent & Wertz, 1996; Holland, 1996; Roberts & Wertz, 1993). Despite the presence of more intact pragmatic abilities in individuals with aphasia, impairments in pragmatic skills are evident. In an investigation designed to test the usefulness of a descriptive taxonomy of pragmatic behaviors, Prutting and Kirshner (1987) assessed the conversational abilities of six different diagnostic groups—children with language disorders, children with articulation disorders, children with normal language, adults with aphasia, adults with right hemisphere damage, and adults with normal language. They observed distinct profiles for each diagnostic group. The aphasic subjects showed a mean of 82% appropriate pragmatic performance. The majority of the deficits noted related to linguistic constraints, including specificity and accuracy, pause time in turn taking, quantity and conciseness of the message, fluency, and the variety of speech acts produced. Goldblum (1985) compared chronic fluent and nonfluent aphasic subjects' conversational skills using the Prutting and Kirshner Pragmatic Protocol. She found that both groups retained their social competence, but the fluent subjects had fewer inappropriate ratings than the nonfluent subjects. The majority of inappropriate behaviors for both groups were in the categories of specificity/accuracy, fluency, pause time in turntaking, and quantity/conciseness. The nonfluent group had more difficulty in pause time in turntaking and quantity/conciseness than the fluent group. Recent evidence suggests that pragmatic skills improve with treatment during the first year postonset. Roberts and Wertz (1993) evaluated the communicative effectiveness of 20 treated aphasic subjects at one, three, six, nine, and 12 months postonset using the Pragmatic Protocol (Prutting & Kirchner, 1983; 1987) and found that pragmatic performance improved throughout the first year postonset. Mean appropriate performance was 87% at one month, 91% at three and six months, 92% at nine months, and 93% at 12 months. Thus, the Pragmatic Protocol appears useful for documenting pragmatic performance and determining change in performance over time. Avent and Wertz (19%) investigated the influence of type of aphasia and type of treatment on aphasic patients' pragmatic abilities in conversation between 4 and 48 weeks postonset of stroke. Ten patients were fluent and ten were nonfluent. All patients received six to eight hours of treatment each week for 44 weeks. Ten patients received group treatment with no direct manipulation of language deficits, and ten patients received individual, stimulus-response treatment with direct manipulation of language deficits in all communicative modalities. Mean conversational pragmatic ability for the 20 patients was 87% appropriate at four weeks postonset. Mean performance after 44 weeks of treatment was 92% appropriate. There were no significant differences in improvement in pragmatic performance between fluent and nonfluent patients, and overall improvement in pragmatic performance was not influenced by the type of treatment, group or individual. However, group treated patients showed significant change in pragmatic abilities during the first 11 weeks of treatment, whereas the individually treated patients did not. The results of these studies indicate that a pragmatic protocol is useful for describing differences among different disorders (Prutting & Kirchner, 1987); distinguishing differences among subtypes of aphasia (Goldblum, 1985); documenting changes over time (Roberts & Wertz, 1993); and describing treatment effects (Avent & Wertz, 1996).

Language impairment and pragmatic behavior

209

However, there is no widespread use of pragmatic assessment in the management of aphasia. Failure to assess conversational ability may result from a variety of factors. Traditional linguistic-based assessments reveal well recognized impairments in syntax, semantics, and phonology (Gallagher, 1991). Pragmatic assessments, on the other hand, capture individualistic communicative performance that is highly sensitive to interactional contexts and cultural expectations (Gallagher, 1991; Hough & Pierce, 1994). Pragmatic assessments are further hindered by the absence of a general theory encompassing spoken discourse or pragmatic aspects of language (McTear, 1985). This lack of a" accepted theoretical framework has resulted in confusing terminology and no defined procedures for assessment (Gallagher, 1991; Penn, 1993; McTear, 1985). Another factor which may influence the scarcity of conversational assessment is historical reliance on standardized testing. Typically, treatment outcome is determined with standardized tests. However, the sole use of standardized tests raises questions about the purpose of documenting treatment effectiveness. While standardized tests have played an important role in establishing the efficacy of aphasia treatment, from a clinical perspective, discrepancies exist between standardized language test performance and observations of aphasic individuals' conversational skills (Aten, Caligiuri, & Holland, 1982; Binder, 1984; Blomert, 1990; Holland, 1980; Newhoff & Apel, 1997; Penn, 1985; Sarno & Levita, 1971; Ulatowska, Haynes, Hildebrand,& Richardson, 1977). Standardized tests, generally, are based on a linguistic framework that does not provide a clear understanding of how an aphasic individual uses his or her language (Penn, 1993; Chapman & Ulatowska, 1992; Gurland, Chwat, & Wollner, 1982; Holland, 1980; Ulatowska et al., 1977). Therefore, to demonstrate that treatment creates a change in communicative competence, other measures may be more appropriate, for example, a pragmatic analysis (Lund & Duchan, 1993; Green, 1984). Sarno (1993) suggests the goal of rehabilitation is to "restore the person's role as a communicator, regardless of whether certain symptoms have been eradicated or particular linguistic skills have improved" (p. 325). Conversational performance provides an opportunity to study "the communicator role" of individuals with aphasia. Moreover, Holland (1982) observed aphasic people communicate better than they talk, implying a difference between pragmatic performance and language impairment. Because a measure of conversational ability was not used in the majority of aphasia treatment studies, it has not been determined whether aphasic patients' pragmatic abilities are related to their linguistic performance on standardized tests. The purpose of this study was to determine whether there is a relationship between pragmatic performance and severity of language impairment in aphasia subsequent to stroke. We attempted to answer the following questions: What is the relationship between pragmatic performance and language impairment in aphasic people during the first year postonset; does the relationship between pragmatic performance and language impairment differ between fluent and nonfluent aphasic people during the first year postonset; and is the relationship between pragmatic performance and language impairment influenced by the type of treatment, individual or group, aphasic people receive? Methods This investigation was a retrospective study of change in aphasic people who were treated during the first year postonset. The data come from a standardized language measure and pragmatic analysis of videotaped conversational samples by subjects who participated in

210

J.R. Avent, R.T. Wertz and L.L. Auther

the first Veterans Administration Cooperative Study on aphasia (Wertz, Collins, Weiss et al., 1981). Prior to receiving treatment, four weeks postonset, each patient conversed with his clinician about events during the previous week. These conversations were repeated after each 11-week treatment period at 15, 26, 37, and 48 weeks postonset. Similarly, a standardized language measure was administered at the same points in time. Subjects Study patients were aphasic subsequent to a first, left hemisphere thromboembolic infarct; four weeks postonset at entry; under 80 years of age; premorbidly literate in English; and had no present or previous other neurological disease (Wertz et al., 1981). Three criteria were used to select subjects from the pool of 34 subjects who completed the VA Cooperative study. First, videotapes containing a conversational sample had to be available for evaluations conducted at 4, 15, 26, 37, and 48 weeks postonset. Second, patients were selected to include an equal number of fluent and nonfluent individuals. And, third, patients were selected to include an equal number of individuals in each of the two treatments, individual and group. Performance by 20 subjects who satisfied these criteria was analyzed. Table 1 shows age, education, and Porch Index of Communicative Ability (PICA) (Porch, 1967) overall percentile performance at four weeks postonset, pretreatment; type of treatment received; and type of aphasia for all subjects.

Subject

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Age (years)

Education (years)

PICA overall percentile

Treatment

Aphasia

73 63 58 65 45 57 63 78 41 51 68 65 62 50 64 55 78 52 52 61

16 9 6 8 12 12.5 8 8 12 12 10 8 9 12 6 7 3 13 12 12

43 72 50 42 15 16 60 66 73 44 49 55 58 49 35 39 26 47 15 19

Individual Group Group Individual Group Individual Individual Group Individual Group Group Individual Individual Group Individual Group Group Individual Individual Group

Nonfluent Fluent Fluent Nonfluent Nonfluent Nonfluent Fluent Fluent Fluent Fluent Fluent Nonfluent Fluent Fluent Nonfluent Nonfluent Fluent Nonfluent Nonfluent Nonfluent

Table 1. Patient descriptive data at four weeks postonset for age, education, language severity, type of treatment, and type of aphasia. Table 2 shows descriptive data for the subjects classified by type of aphasia, fluent and nonfluent, and Table 3 shows descriptive data for the subjects classified by the type of

Language impairment and pragmatic behavior

211

treatment received, individual or group. The two groups, classified by type of aphasia, did not differ significantly (p < .05) in chronological age or in years of education. However, the groups differed significantly in language severity on the PICA. Nonfluent aphasic subjects scored significantly lower (p < .003) than the fluent aphasic subjects. The two groups, classified by type of treatment received, did not differ significantly (p < .05) in age, education, or language severity on the PICA.

Variable

Age in years Mean Standard Deviation Range Education in years Mean Standard Deviation Range PICA Overall Percentile Mean Standard Deviation Range

Type of aphasia Nonfluent Fluent (n = 10) (n = 10)

61.20 11.25 41-78

58.90 7.82 45-73

8.90 2.74 3-12

10.65 3.03 6-16

54.70 14.24 26-73

32.60 15.03 15-55

Table 2. Descriptive data at four weeks postonset for age, education, and language severity for subjects classified by type of aphasia.

Variable

Age in years Mean Standard Deviation Range Education in years Mean Standard Deviation Range PICA Overall Percentile Mean Standard Deviation Range

Type of treatment Individual Group (n = 10) (n = 10)

59.40 8.62 41-73

60.70 10.73 45-78

10.45 2.94 6-16

9.10 2.95 3-12

42.90 18.73 15-73

44.40 18.66 15-72

Table 3. Descriptive data at four weeks postonset for age, education, and language severity for subjects classified by type of treatment received.

212

J.R. Avent, R.T. Wertz and L.L. Auther

Treatment. Each subject received six to eight hours of treatment each week for 44 weeks. Individual treatment was conducted by a speech-language pathologist and consisted of four hours of stimulus-response type of treatment designed to improve language in all communicative modalities—auditory comprehension, reading, oral-expressive language, and writing—and two to four hours of machine-assisted treatment. Three fluent and seven nonfluent aphasic patients received individual treatment. Group treatment was conducted by a speech-language pathologist in groups of three to seven patients for four hours each week, and an additional two to four hours of group recreational activity was provided. Group treatment was designed to facilitate language use in a social setting, but there was no direct manipulation of speech or language deficits. Seven fluent and three nonfluent aphasic patients received group treatment. Pragmatic analysis The pragmatic aspects of language were analyzed for each patient with the Pragmatic Protocol (Prutting and Kirchner, 1983; 1987). The Pragmatic Protocol assesses verbal, paralinguistic, and nonverbal aspects of language use and is designed to screen overall communicative abilities. It consists of 30 pragmatic aspects of language that were extrapolated from normative data on children and adults. For the purposes of this study, the category of stylistic variations was omitted from the ratings, because it required observations not available in the videotape samples of conversation. Pragmatic aspects assessed in the 29-item protocol we employed included verbal aspects, for example, variety of speech acts, topic selection, specificity/accuracy,etc.; paralinguistic aspects, for example, intelligibility, vocal intensity, prosody, etc.; and nonverbal aspects, for example, physical proximity, body posture, gestures, etc. According to Prutting and Kirchner (1987), the Pragmatic Protocol rating should be completed after viewing an unstructured, spontaneous conversational interaction between the subject and his or her partner. Each of the protocol behaviors is scored as appropriate, inappropriate, or no opportunity to observe. All conversational samples were between the patient and a speech-language pathologist who conducted the periodic evaluations. Procedures Prior to the beginning of the study, the 100 subject videotapes, five for each of the 20 subjects—4, 15, 26, 37, and 48 weeks postonset, were randomized and coded by number to control for bias from knowing the time postonset. A pragmatic protocol for each conversation was completed for each of the 100 conversations. Five percent of the videotapes were selected randomly to determine point-to-point reliability. Interjudge reliability between two judges for the pragmatic ratings was 94 percent. PICA performance at 4, 15, 26, 37, and 48 weeks postonset utilized the data from the VA Cooperative Study. Interjudge reliability between two judges for PICA overall percentile performance, obtained in the VA Cooperative Study, was 98 percent.

Language impairment and pragmatic behavior

213

Results Language impairment and pragmatic performance Table 4 shows language impairment, PICA overall percentile, and pragmatic performance, percent appropriate behaviors, for all subjects at entry, four weeks postonset, and at 15, 26, 37, and 48 weeks postonset. Language impairment improved significantly over time, 44th percentile at four weeks postonset and 69th percentile at 48 weeks postonset. Similarly, pragmatic performance improved significantly over time, 87 percent appropriate at four weeks postonset and 93 percent appropriate at 48 weeks postonset. At all points in time, pragmatic performance was better than language performance.

Time

Measures PICA %ile

Pragmatic % App

Mean

Range

Mean

Range

4 Weeks

43.65

15-73

87.45

72-97

15 Weeks

62.00

15-90

91.15

76-97

26 Weeks

64.70

13-90

91.35

79-97

37 Weeks

67.60

20-93

92.25

79-100

48 Weeks

68.70

32-95

92.90

83-100

Table 4. Mean and range in the PICA overall percentile and percent appropriate pragmatic performance at 4, 15, 26,37, and 48 weeks postonset for all subjects (n = 20). As shown in Table 5, pragmatic performance was more inappropriate in verbal aspects than in paralinguistic or nonverbal aspects at all points in time from 4 to 48 weeks postonset. Subjects' specificity/accuracy was the least appropriate of all verbal aspects, and it remained least appropriate during the 44-week treatment trial. Intelligibility and fluency were the least appropriate paralinguistic aspects. Nonverbal aspects were generally appropriate, and none emergedas more inappropriate than the others. The total number of inappropriate behaviors in the 20 subjects declined over time from 73 at four weeks postonset to 42 at 48 weeks postonset. However, at 48 weeks postonset, half of the 20 patents remained inappropriate in specificity/accuracy.

214

J.R. Avent, R.T. Wertz and L.L. Auther

Behavior

Weeks postonset 15

Verbal Aspects Speech act pair analysis Variety of speech acts Topic Selection Topic Introduction Topic Maintenance Topic Change Turntaking Initiation Turntaking Response Turntaking Repair/Revision Turntaking Pause Time Turntaking Interpretation/Overlap Turntaking Feedback to Speaker Turntaking Adjacency Turntaking Contingency Turntaking Quantity/Conciseness Specificity/Accuracy Cohesion Subtotal Paralinguistic Aspects Intelligibility Vocal Intensity Prosody Fluency Subtotal

2 1 1 4 5

1 1 3

48

5

1

1 1 4

1 1 3

1

6 5 18 6

1 1

1

1

5 2 16 4

7

4

14 2

6 3 15 3

53

32

30

32

28

9 1 2 6

9 1 1 7

8 2 2 6

4 1 1 6

6 1 1 4

18

18

18

12

12

10 6

1 1

Subtotal

Table 5.

37

1 1

Nonverbal Aspects Physical Proximity Physical Contacts Body Posture Foot/Leg and Hand/Arm Movement Gestures Facial Expression Eye Gaze

Total

26

73

52

1

1

2

2

50

46

42

Number of subjects exhibiting specific, inappropriate verbal, paralinguistic, and nonverbal pragmatic behaviors at 4,15,26,37, and 48 weeks postonset.

Language impairment and pragmatic behavior

215

Fluent and nonfluent Table 6 indicates fluent patients displayed less language impairment, indicated by the PICA overall percentile, than nonfluent patients at all points in time. Differences ranged from 23.4 percentile units at 15 weeks postonset to 13.6 percentile units at 48 weeks postonset. Similarly, pragmatic performance was better in fluent patients than in nonfluent patients at all points in time. Differences ranged from 3.1 percent at 26 weeks postonset to 5.0 percent at 48 weeks postonset. While fluent patients performed significantly better on the PICA than nonfluent patients, there was no significant difference between fluent and nonfluent patients in the amount of improvement attained—21 percentile units in the fluent group and 29 percentile units in the nonfluent group—during the 44-week treatment trial. Pragmatic performance did not differ significantly between groups at any point in time, and there was no significant different between groups in the amount of improvement attained—5.7 percent in the fluent group and 5.2 percent in the nonfluent group—during the 44-week treatment trial. Time F

4 Weeks 15 Weeks 26 Weeks 37 Weeks 48 Weeks

54.7 73.7 73.8 75.0 75.5

Comparison PICA %ile NF D

32.6 50.3 55.6 60.2 61.9

+22.1 +23.4 +18.2 +14.8 +13.6

F

Pragmatic °>% D NF

89.7 93.2 92.9 94.4 95.4

85.2 89.1 89.8 90.1 90.4

+4.5 +4.1 +3.1 +4.3 +5.0

Table 6. Mean and mean difference (D) in the PICA overall percentile and percent appropriate pragmatic performance between the fluent (F) and nonfluent (NF) aphasia groups at 4, 15, 26, 37, and 48 weeks postonset. Individual and group treatment Ten patients, three fluent and seven nonfluent, received individual treatment, and ten patients, seven fluent and three nonfluent, received group treatment between 4 and 48 weeks postonset. Table 7 shows comparisons between the individual and group treatment groups at 4, 15, 26, 37, and 48 weeks postonset. There were no significant differences between groups in PICA overall percentile performance or pragmatic performance at any point in time. Both groups attained a similar amount of improvement on the PICA—27 percentile units for the individual treatment group and 23 percentile units for the group treatment group—during the 44-week treatment trial. And, there was no significant difference between groups in the amount of improvement attained. Both groups displayed similar improvement in pragmatic performance—4.2 percent in the individual treatment group and 6.7 percent in the group treatment group—during the 44-week treatment trial. And, there was no significant difference bet ween groups in the amount of improvement attained. However, a significant time by type of treatment interaction emerged. This resulted from significant improvement in group treated patients, 5.3 percent, between 4 and 15 weeks postonset, and nonsignificant improvement in individually treated patients, 2.1 percent, during the same time period.

216

J.R. Avent, R.T. Wertz and LL. Auther Time

Comparison PICA %ile I G

4 Weeks 15 Weeks 26 Weeks 37 Weeks 48 Weeks

46.0 65.5 67.7 71.3 72.7

413 58.5 61.7 63.9 64.7

Pragmatic °k G

D

I

+4.7 +7.0 +6.0 +7.4 +8.0

88.0 90.1 91.5 92.8 92.2

86.9 92.2 91.2 91.7 93.6

D

+1.1 -2.1 + 0.3 +1.1 -1.4

Table 7. Mean and mean difference (D) in PICA overall percentile and percent appropriate pragmatic performance between individual (I) and group (G) treated patients. Relationship between language impairment and pragmatic performance As shown in Table 8, the PICA overall percentile was significantly related with pragmatic performance at all points in time for the total sample of 20 patients. However, when the sample was divided into fluent and nonfluent patients, PICA performance was significantly related with pragmatic performance in the fluent group at four weeks postonset but not thereafter. Conversely, in nonfluent patients, PICA and pragmatic performance was not significantly related at four weeks postonset, but it was significantly related at 15, 26, 37, and 48 weeks postonset. Comparison of PICA and pragmatic performance in individual and group treated patients indicated a significant relationship for individual treated patients at all points in time except 37 weeks postonset. Conversely, in group treated patients, PICA and pragmatic performance was significantly related at only 37 weeks postonset. In addition, we examined the relationship between total improvement, between 4 and 48 weeks postonset, in PICA and pragmatic performance. As shown in Table 8, no significant relationships emerged for all subjects combined, fluent patients, nonfluent patients, individually treated patients, or group treated patients. Time 4 weeks 15 weeks 26 weeks 37 weeks 48 weeks Change between 4 and 48 weeks

A

F

Cohorts NF I

G

.68* .63* .54** .58* .57*

.66** .33 .01 .19 .02

.59 .66** .66** .69** .81*

.86* .84* .69** .57 .69**

.31 .59 .42 .66** .56

.36

.49

.31

.52

.33

* Significant at p < .01 ** Significant at p < .05

TableS.

Relationship (r) between PICA overall percentile and percent appropriate pragmatic performance for all subjects (A), fluent (F) and nonfluent (NF) groups, and individual (I) and group (G) treated patients at 4, 15, 26, 37, and 48 weeks postonset and change in PICA and pragmatic performance between 4 and 48 weeks postonset.

Language impairment and pragmatic behavior

217

Discussion If pragmatic performance in conversation is considered to be a measure of communication and performance on a standardized test of aphasia, for example, the PICA is a measure of language impairment, our results support Holland's (1982) observation. Our aphasic people communicated better than they talked. However, we found no differences in the severity or amount of improvement in pragmatic performance bet ween fluent and nonfluent aphasic people. Moreover, the type of treatment received, individual or group, appeared to have no influence on the severity of or amount of improvement in pragmatic performance. The relationship between pragmatic performance and language impairment is inconsistent over time, and the point in time when this relationship exists appears to differ between fluent and nonfluent aphasic patients. And, while the severity of pragmatic performance and the severity of language impairment appears related in a combined group of fluent and nonfluent aphasic people, improvement in pragmatic performance is not related with improvement in language impairment. Pragmatic performance and language impairment Generally, our sample of aphasic people displayed better pragmatic performance in conversation than language ability on a standardized test for aphasia. This was the case at four weeks postonset, pretreatment—72 to 97 percent pragmatically appropriate compared with 15th to 73rd PIC A overall percentile—and at 48 weeks postonset, post-treatment— 83 to 100 percent pragmatically appropriate compared with 32nd to 95th PICA overall percentile. Performance in both domains improved significantly over time, 4 to 48 weeks postonset. Mean change in the PICA overall percentile was greater, 25 percentile units, than in percent appropriate pragmatic performance, 4.45 percent. Probably, this results from a "ceiling effect" on the pragmatic protocol. Appropriate pragmatic behaviors were high early postonset, 87 percent, and had little room to improve over time. Pragmatic behaviors that were most impaired at four weeks postonset—specificity/ accuracy, cohesion, intelligibility—remained the most impaired at 48 weeks postonset. Prutting and Kirchner (1987) reported 100 percent of their aphasic sample were inappropriate in specificity/accuracy. Ninety percent of our sample demonstrated inappropriate specificity/accuracy at four weeks postonset. Fifty percent continued to display inappropriate specificity/accuracy at 48 weeks postonset. This aspect of pragmatic behavior, along with cohesion and intelligibility, may be most similar to the semantic, syntactic, and phonologic measures contained in standardized measures of language impairment, for example, the PICA. Fluent and nonfluent Difference between pragmatic performance and language impairment became apparent when fluent and nonfluent aphasic people were compared. Our fluent aphasic patients displayed significantly less impairment on the PICA than our nonfluent aphasic patients. However, we found no significant differences in pragmatic performance between our fluent and nonfluent groups. The difference between groups in language impairment is probably an artifact of the dichotomy—fluent and nonfluent. The nonfluent group included the most severely aphasic people, global, and the fluent group included the most mildly aphasic

218

J.R. Avent, R.T. Wertz and L.L. Auther

people, anomic. Thus, measures of language impairment typically result in group differences when aphasic patients are classified into fluent and nonfluent groups. However, pragmatic analysis indicates fluent and nonfluent aphasic people do not differ significantly in their percent of appropriate pragmatic behaviors. This was our observation, and it is in line with Goldblum's (1985) results. When improvement in pragmatic performance and language impairment is compared in fluent and nonfluent groups, the groups appear to change similarly. Both of ours made essentially the same amount of improvement of both measures during a 44-week treatment trial, and there were no significant differences between the fluent and nonfluent groups in the amount of improvement attained on either measure. Thus, the response to time and treatment appears to be similar in fluent and nonfluent aphasic patients.

Individual and group treatment The influence of the type of therapy appeared minimal on both pragmatic performance and language impairment. Individually treated and group treated patients made essentially the same amount of improvement on both measures, and there were no significant differences between groups on either measure at any point in time. The only significant difference between treatments was significant improvement in pragmatic performance in group treated patients during the first 11 weeks of treatment and the absence of significant improvement in pragmatic performance during the same period in individually treated patients. Failure to find a treatment group effect is somewhat surprising. Individually treated patients received stimulus-response treatment designed to improve language impairment in all communicative modalities—auditory comprehension, reading, oral-expressive language, and writing. Group treated patients received methods designed to improve communication in a social setting and no direct manipulation of specific language deficits. One might predict that these different treatments would result in individually treated patients making more improvement in language impairment and group treated patients making more improvement in pragmatic performance. This did not occur. Aten, Caligiuri, and Holland (1982) reported that group treatment with chronic aphasic patients that was designed to improve functional communication resulted in significant pre- to post-treatment improvement on the Communicative Abilities in Daily Living (CADL) (Holland, 1980) but not on the PICA. We found group treatment designed to improve communication in a social setting did not result in significantly better pragmatic performance than individual, traditional, stimulus-response treatment designed to improve language impairment. Our patients entered treatment at four weeks postonset and continued until 48 weeks postonset. Aten et al.'s patients were chronically aphasic when treatment began. Moreover, similarities between what we call functional communication and pragmatic aspects of language are not clear. We observed that pragmatic performance was generally better than performance on a standardized test for aphasia. However, the relationship between pragmatic performance and functional communication remains to be demonstrated.

Language impairment and pragmatic behavior

219

Relationship between language impairment and pragmatic performance In our total sample of 20 aphasic patients, pragmatic performance and language impairment were significantly related pretreatment, four weeks postonset, and at 15, 26, 37, and 48 weeks postonset. However, when nonfluent aphasic patients were compared with fluent aphasic patients, pragmatic performance and language impairment were significantly related in the fluent group only at four weeks postonset. Conversely, in the nonfluent group, pragmatic performance and language impairment were not significantly related at four weeks postonset, but they were significantly related at 15, 26, 37, and 48 weeks postonset. And, when the relationship was examined according to the type of treatment received, pragmatic performance and language impairment in individually treated patients were significantly related at all points in time except at 37 weeks postonset. Conversely, in group treated patients, pragmatic performance and language impairment were significantly related only at 37 weeks postonset. Finally, when improvement in pragmatic performance and improvement in language impairment were compared, we found no significant relationship for all patients combined, fluent patients, nonfluent patients, individually treated patients, or group treated patients. Some (Aten et al., 1982; Binder, 1984; Blomert, 1990; Holland, 1980; Newhoff and Apel, 1997; Penn, 1985; Sarno and Levita, 1971; Ulatowska et al., 1977) suggest discrepancies exist between standardized language test performance and observations of aphasic patients' conversational skills. While discrepancies exist, our results indicate pragmatic performance during conversation and performance on a standardized aphasia test are significantly related in a sample of aphasic people. Thus, performance on one measure predicts performance on the other. However, the strength of the relationship, correlations ranging from +.57 to +.68, is not overwhelming. The relationship between the two measures changes when aphasic people are subdivided into fluent and nonfluent groups. At four weeks postonset, pragmatic performance and language impairment was significantly related in fluent aphasic patients but not in nonfluent aphasic patients. Failure to find a significant relationship in nonfluent patients resulted from good pragmatic performance, 80% appropriate, coexisting with severe language impairment, PICA performance below the 20th percentile, in 30 percent of the nonfluent sample. Conversely, the two measures were significantly related in nonfluent patients between 15 and 48 weeks postonset but not in fluent patients. Scatter plots indicate that nonfluent patients' language impairment, by 15 weeks postonset, had improved sufficiently to coincide with their generally intact pragmatic performance at four weeks postonset. By 15 weeks postonset, both pragmatic performance and language impairment in fluent aphasic patients had improved, and there was a narrow range of severity in both behaviors among members of the nonfluent group. The sporadic relationship over time between pragmatic performance and language impairment in the different treatment groups probably results from unequal assignment of fluent and nonfluent patients to the different treatments. Seven nonfluent and three fluent patients received individual treatment, and three nonfluent and seven fluent patients received group treatment. Finally, while we observed a significant relationship between pragmatic performance and language impairment at all points in time between 4 and 48 weeks postonset for our entire sample, no significant relationship was observed between improvement in pragmatic performance and improvement in language impairment between 4 and 48 weeks postonset. Thus, while performance on one measure appears to predict severity on the other measure, improvement on one measure does not predict improvement on the other measure.

220

J.R. Avent, R.T. Wertz and L.L. Anther

The work to be done is exploration of aphasic patients' pragmatic performance and its relationship to their communicative competence. While Leech (1983) suggests pragmatic performance is a measure of communicative competence, it is not clear how pragmatic performance in aphasia relates to functional communication. Previous reports (Avent and Wertz, 19%; Binder, 1984; Goldblum, 1985; Prutting and Kirchner, 1987; Roberts and Wertz, 1993) have compared pragmatic performance and language impairment in aphasia. Similarly, our results indicate pragmatically appropriate performance can coexist with significant language impairment. We may want to ask what pragmatic appropriateness, as indicated by a pragmatic protocol, implies about an aphasic person's functional communication. References Aten, J. L., Caligiuri, M. P., & Holland, A. L. (1982). The efficacy of functional communication therapy for chronic aphasic patients. Journal of Speech and Hearing Disorders, 47,93-96. Avent, J. R., & Wertz, R. T. (1996). Influence of type of aphasia and type of treatment on aphasic patients' pragmatic performance. Aphasiology, 10, 253-265. Binder, G. M. (1984). A societal and clinical appraisal of pragmatic and linguistic behaviors. Unpublished master's thesis, University of California, Santa Barbara, CA. Blomert, L. (1990). What functional assessment can contribute to setting goals for aphasia therapy. Aphasiology, 4, 307-320. Brinton, B., & Fujiki, M. (1989). Conversational management with language-impaired children. Rockville: Aspen Publishers, Inc. Chapman, S. B., & Ulatowska, H. (1992). Methodology for discourse management in the treatment of aphasia. Clinical Communication Disorders, 2, 64-81. Gallagher, T. M. (1991). Pragmatics of language: Clinical practice issues. San Diego: Singular Publishing Group, Inc. Goldblum, G. (1985). Aphasia: A societal and clinical appraisal of pragmatic and linguistic behaviors. South African Journal of Communication Disorders, 32, 11-18. Green, G. (1984). Communication in aphasia therapy: Some of the procedures and issues involved. British Journal of Disordered Communication, 19,35-46. Gurland, G. B., Chwat, S. E., & Wollner, S. G. (1982). Establishing a communication profile in adult aphasia: Analysis of communicative acts and conversational sequences. In R. H. Brookshire (Ed.), Clinical Aphasiology Conference Proceedings (pp. 18-27). Minneapolis: BRK Publishers. Holland, A. L. (1980). Communicative abilities in daily living. Baltimore: University Park Press. Holland, A. L. (1982). Observing functional communication of aphasic adults. Journal of Speech and Hearing Disorders, 47, 50-56. Holland, A. L. (1996). Pragmatic assessment and treatment for aphasia. In G. L. Wallace (Ed.), Adult aphasia rehabilitation (pp. 161-173). Boston: Butterworth-Heinemann. Hough, M. S., & Pierce, R. S. (1994). Pragmatics and treatment. In R. Chapey (Ed.), Language intervention strategies in adult aphasia (third ed.) (pp. 246-268). Baltimore: Williams and Wilkins. Leech, G. N. (1983). Principles of pragmatics. London: Longman. Levinson, S. C. (1983). Pragmatics. Cambridge: Cambridge University Press. Lomas, J., Pickard, L., Bester, S., Elbard, H., Finlayspn, A., & Zoghaib, C. (1989). The Communicative Effectiveness Index: Development and psychometric evaluation of a functional communication measure for adult aphasia. Journal of Speech and Hearing Disorders, 54, 113-124. Lund, N. J., & Duchan, J. F. (1993). Assessing children's language in naturalistic contexts (third ed.) Englewood Cliffs, N.J.: Prentice Hall. McTear, M. F. (1985). Pragmatic disorders: A question of direction. British Journal of Disorders of Communication, 20,119-127. Manochiopinig, S., Sheard, C., & Reed, V. A. (1992). Pragmatic assessment in adult aphasia: A clinical review. Aphasiology, 6, 519-533. Newhoff, M., & Apel, K. (1997). Impairments in pragmatics. In L. L. LaPointe (Ed.), Aphasia and related neurogenic language disorders, second ed. (pp. 250-264). New York: Thieme Medical Publishers, Inc. Penn, C. (1985). The profile of communicative appropriateness: A clinical tool for the assessment of pragmatics. South African Journal of Communication Disorders, 32,18-23. Penn, C. (1993). Aphasia therapy in South Africa: Some pragmatic and personal perspectives. In A. L. Holland and M. M. Forbes (Eds.), Aphasia treatment: World perspectives (pp. 25-53). San Diego: Singular Publishing Group, Inc. Porch, B. E. (1967). Porch index of communicative ability. Palo Alto, CA: Consulting Psychologists Press. Prins, R. S., Snow, C. E., & Wagenaar, R. (1978). Recovery from aphasia: Spontaneous speech versus language comprehension. Brain and Language, 6, 192-211.

Language impairment and pragmatic behavior

221

Priming, C. A., & Kirchner, D. M. (1983). Applied Pragmatics. In T. M. Gallagher and C. A. Fruiting (Eds.), Pragmatic assessment and intervention issues in language. San Diego: College-Hill Press, Inc. Printing, C. A., & Kirchner, D. M. (1987). A clinical appraisal of the pragmatic aspects of language. Journal of Speech and Hearing Disorders, 52, 105-119. Roberts, J. A., & Wertz, R. T. (1993). Communicative effectiveness in treated aphasic patients during the first post onset year. In M. L. Lemme (Ed.), Clinical Aphasiology, 21, 291-298. Sarno, M. T. (1993). Aphasia rehabilitation: Psychosocial and ethical considerations. Aphasiology, 4,321-334. Sarno, M. T., & Levita, E. (1971). Natural course of recovery in severe aphasia. Archives of Physical Medicine and Rehabilitation, 52, 175-179. Smith, L. (1985). Communicative activities of dysphasic adults: A survey. British Journal of Disordered Communication, 20,31-44. Ulatowska, H. K., Haynes, S. M., Hildebrand, B. H., & Richardson, S. M. (1977). The aphasic individual: A speaker and a listener, not a patient. In R. H. Brookshire (Ed.), Clinical Aphasiology Conference Proceedings (pp. 198-213). Minneapolis: BRK Publishers. Wertz, R. T., Collins, M. J., Weiss, D., Kurtzke, J. R, Friden, T., Brookshire, R. H., Pierce, J, Holtzapple, P., Hubbard, D. J., Porch, B. E., West, J. A., Davis, L., Matovitch, V., Morley, G. K., & Resurreccion, E. (1981). Veterans Administration cooperative study on aphasia: A comparison of individual and group treatment. Journal of Speech and Hearing Research, 24, 580-594.

This page intentionally left blank

Pergamon

©

J. Neurolinguistics, Vol. 11, Nos 1-2, p. 223-241, 1998 1998 Published by Elsevier Science Ltd. All rights reserved Printed in Great Britain 0911-6044/98 $19.00 + 0.00

PII: S0911-6044(98)00015-3

A cross-language analysis of conversation in a trilingual speaker with aphasia Luise Springer*, Nick Miller and Frauke Bürk† *RWTH Aachen, Germany; †University of Newcastle upon Tyne. Abstract—This study explores the relationship between formal and functional language performance of a trilingual speaker with aphasia, focusing on her monolingual use of German and English in free conversations and semi structured interviews with different interlocutors. We use a conversation analysis framework to highlight the contrasting pictures between her formal test scores and sources of trouble across different conversations. In the discussion possible reasons for divergences across settings and languages are examined.

Introduction In aphasiology, as in other fields of language enquiry, it has been recognised for some time now that an account of successes and problems in communication is incomplete unless the full, collaborative nature of conversational exchanges is encompassed in any descriptive or analytic procedure. Achieving mutual understanding is as much a task of the listener as it is of the speaker. Several approaches have been developed aimed at characterising conversational interaction. One of the approaches, developed from ethnomethodological perspectives on language behaviour, is conversation analysis (Schegloff, Jefferson & Sacks, 1977; Heritage, 1989). Essentially, this is a data driven, bottom-up approach to analysing the sequential, collaborative construction of a conversation, that attempts as far as possible to embark on the analytic task with as few prior assumptions as possible regarding the nature of what might cause trouble for participants and how repairs of any trouble spots in conversation might be negotiated. Conversation analysis (CA) and derivatives have been applied to the description and analysis of talk in an increasing range of communication settings—including the field of speech-language pathology. CA has been employed to gain further insights into the nature of communication in conversations between children with semantic-pragmatic disorders and adults (Brinton & Fujiki, 1989; Willcox & Mogford-Bevan, 1995), speakers with aphasia and their partners (Ferguson, 1994; 1996; Milroy & Perkins, 1992; Perkins, 1995), aphasic speakers with each other (Klippi, 1995), speakers with dementia and their partners (Hamilton, 1994; Orange, Lubinski & Higginbotham, 1996), and speakers with head injury (Friedland& Miller, in press). Researchers have examined the nature of breakdowns, the properties of self repairs (Laakso, 1997), and variations across settings, partners and topics (Ferguson, 1994). To our knowledge there are no accounts yet applying CA principles to the comparison of performances across different languages (though see Friedland & Miller, 1997) in bilingual speakers with aphasia, the topic of this report. One claimed advantage of CA over formal language testing is that it does not look at the results of isolated levels of linguistic analysis divorced from communication. Such isolated results may tell one about the potential strengths and weaknesses of a speaker and

224

L. Springer, N. Miller and F. Bürk

underlying impairments they have to face. However, they may give little indication of the speaker's true communicative competence. Aphasia test results also ignore the fact that in a person's conversations there may be formal 'errors' that cause no breakdown in communication and there may be well-formed utterances that do. CA, by contrast, concentrates in assessment on what a speaker is actually able to achieve with her/his (limited) formal language skills, and, just as importantly, what the speaker can achieve in tandem with another speaker. CA lends itself, too, to addressing issues of communication (breakdown) in bilingual settings in general, and in individual speakers. For instance, if a pragmatic disorder is evident while using one language, is it also present and of the same type and degree while speaking the other language? Are differences in overall severity of impairment between the languages concerned reflected in the comparative severity of conversational breakdowns across languages? Do the trouble spots for the conversational partners in the two languages reflect common underlying deficits in language processing, or do language specific trouble spots arise? For speakers who habitually maintain strict language separation, does code switching or mixing become a problem in aphasia and what kind of consequences does this have for their communication with monolingual and bilingual speakers? We examined some of these issues in a German-Enghsh-Italian trilingual speaker with agrammatic aphasia (although the analyses below pertain only to her German and English interactions). Our enquiries were directed at an examination of the possible relationships between the speaker's (HK) formal language profiles in German and English and the trouble sources and repair strategies used in her monolingual conversations in German versus English. For this we employed elements of CA methodology. More specifically we wished to establish whether HK's success at communicating in German and English, in comparable settings, was equal across languages or whether success mirrored, or was influenced by, the significant differences established on formal language assessments. We examined whether the trouble sources in conversation could be directly attributed to the identified formal language impairments or whether for one or both languages there existed a pattern of troubles independent of the formal language profile. We analysed the nature of the trouble and repair instances to see whether a similar pattern was observable across languages or whether HK employed strategies available to her in one language but not the other. A further target of our enquiries concerned the effects of different conversational types (informal versus semi structured) and partners (Ferguson, 1994) on the quantity and quality of trouble spots and repairs. We made the following predictions and hypotheses: 1) Given that HK is a right-handed speaker with a lesion in the left hemisphere, we hypothesized that there would be no pragmatic disorderin either of her languages, i.e., turn taking, topic maintenance, self-monitoring and collaboration with the interlocutor would be normal. 2) From a descriptive point of view, we predicted that results on formal language tests would show a quantitative score in favour of German, but that the profile of the types of impairment would be qualitatively similar for German and English. 3) Consequent on the predicted qualitative similarity, we hypothesized that the sources of trouble in both languages, with all speakers and in all settings would be identical. 4) However, the differences in quantitative scores across languages would lead to a greater number of incomplete or fragmentary utterances in English, a more unfavourable typetoken relationship for closed as well as open class words, a greater presence of interjections and pauses/hesitations and in turn a raised number of trouble sources in English. Furthermore, it was hypothesized that these depressed English scores, in

Cross-language conversation analysis

225

comparison to German, would lead to longer and more complex repair trajectories. On the basis of the quantitative dissimilarity, it was also predicted that there would be more intrusions of German into the English conversations than vice versa. Methods and materials This section gives the medical and language background of HK, details the assessments we carried out with her and explains the measures and analyses we conducted on the data gathered. Case history Background At the time of this study HK was 34 years old, married with three children. She was born in Germany, but at the age of five moved with her family to Italy. There she attended a German school (where she had some instruction in Italian and learned some English) for 7 years before moving back to Germany. After a four month stay in Germany the family moved, with her father's job, to Britain. In London HK went to an English school and passed her school leaving exams (A-levels). HK reported that she had many English friends and spoke English outside of the home. Her command of the English language was very good. At the age of 20 HK and her family returned permanently to Germany where she then passed the German 'Abitur' (equivalent to the English A-levels). Subsequently she trained and worked as a horse-riding instructor. According to HK, before her subarachnoid haemorrhage (SAH) her language skills were equal in German and English whereas her Italian was not of the same standard. Focus in this study was on HK's dominant languages, German and English. Medical history At the age of 33 HK suffered a SAH due to an aneurysm in the left posterior communicating artery (diagnosed by CT-scan). The episode left her with a right hemiplegia and a marked expressive as well as receptive dysphasia with comprehension relatively better preserved. She had a speech apraxia. On the German version of the Wechsler Adult Intelligence Test HK achieved a performance IQ of 98. Initially HK received regular speech therapy in German in hospital and after discharge further treatment in the community. Eight months after her haemorrhage she came for a 7week stay at the specialist Aphasia Ward, at the Aachen University Hospital in Germany, for an intensive rehabilitation programme with daily speech and language therapy. Data gathering At the end of her treatment in Aachen HK was assessed using the Aachen Aphasia Test (AAT) in German and the English adaptation of the AAT (Miller, De Bleser & Willmes, 1997). The AAT commences with a semistructured interview with the therapist, with predetermined questions (about the person's illness, their family, work and hobbies),

226

L. Springer, N. Miller and F. Btirk

designed to gather a spontaneous language sample. The directions for the interview instruct the therapist to give maximum encouragement and time to the speaker but preferably with minimum intervention on the part of the therapist. The German and English versions of this interview (referred to below as the semistructured interview) were video recorded. In addition, two naturally occurring, general conversations without predetermined topics were videotaped, one in each language (i.e. German and English), again with both speakers visible on the screen. These are the conversations referred to below. Three of the conversation partners, who were strangers to HK, were introduced as monolingual speakers. For the English AAT interview HK knew the interlocutor as someone she had received therapy from in German, although the therapist had a native-like command of English. Transcription, data coding and analysis of conversation The conversations were transcribed orthographically from the video recordings by two of the authors working independently. Their versions were compared and any disagreements reviewed on the videos until consensus was reached. Unintelligible utterances were transcribed in a broad phonemic script or recorded as the number of syllables spoken if a clear transcription was not possible (e.g. due to simultaneous talk or laughter by the interlocutor). No utterances, including apparently meaningless vocalisations, laughter, pauses, attempts at written communication and silences were omitted. Notation followed Lesser and Milroy (1993). The transcripts were divided into speaker turns. A turn was taken to be a conversational contribution by one speaker followed by either silence or the start of another speaker's contribution. For the purpose of analysis the length of conversations was matched by taking only those turns in the longer of each conversational pair (German and English interview; German and English free conversation) up to the length of the number of turns contained in the shorter of the pair. Turns were further subdivided into utterances ranging from complete well-formed phrases to interjections. Following Orange et al. (1996), sources of trouble (TS) were divided into phonological, morphological-syntactic, semantic, discourse and 'other' trouble sources. The categories for describing repair types and resolution patterns were also taken from Orange et al. (1996)—viz. repair by repetition, elaboration (interlocutor expands on possible meaning of speaker's utterance), reduction (elliptical responses, confirmation and denial responses, indications of not knowing the answer to a question), substitution (use of alternate but equivalent form of meaning or altered syntax but preserved meaning), and unrelated (interlocutor does not respond to a repair initiator or gives a reply unrelated to the trouble source). Repair resolution was registered as 'most successful' (single trouble source, initiator and repair), 'successful1 (more than one repair initiator and repair used to resolve a single trouble source), or 'unsuccessful' (trouble source not repaired). Repair complexity was coded as simple (single trouble source) or complex (primary TS with embedded TS's). Full definitions are available in Orange et al. (1996). Repair initiation was counted as either self (HK) or other (conversation partner) initiated and repair resolution as self, other or collaboratively achieved. This gave a taxonomy of self-initiated self repairs (SI-SR), self-initiated other repairs (SI-OR), self initiated collaborative repairs (Si-Col R), other initiated self repairs (OI-SR), other initiated other repairs (OI-OR), and other initiated collaborative repairs (OI-Col R). Examples are given in the appendix.

Cross-language conversation analysis

227

We analysed the pattern of topic initiation, maintenance and shift from the point of view of presence and appropriateness. We observed the ability of the speakers to take turns, to appropriately relinquish and resume them and their ability to acknowledge trouble in the conversation and to work together to repair it. Coding and counts were done by hand by two authors independently. Disagreements were analysed again until consensus was reached. In practice, disagreements ranged from 510 percent of the counts. Control speakers Our prime interest concerned HK's own performance in her different languages and across settings, rather than a comparison with 'healthy' (monolingual) speakers of German and English. As such HK acted as her own control. However, we did monitor the communicative behaviour of her conversation partners. Apart from some instances of abrupt topic shifts (discussed later) that were trouble sources for HK, there were no occasions where the phonological, morpho-syntactic or lexical-semantic content of the interlocutor's talk acted as a .TS for HK or led to the need for any repairs. Data processing and analysis Three sets of comparisons were made to address the hypotheses regarding performance across the different languages, conversation types and partners. Firstly we compared the formal AAT results in the two languages; the number and distribution of linguistic units across languages and settings; and the number, length and distribution of interjections and pauses. The second set of comparisons concerned the number and nature of TS, and the pattern of repair initiation sequences. The third set of comparisons examined the repair strategies and the length, complexity and success of repair attempts. Comparisons were made using appropriate exact non-parametric statistical procedures. Fisher's exact test was used for 2x2 tables and its generalisation for general two-way contingency tables employing the StatXact program-package (1989). In addition, permutation tests for two-sample problems (Edgington, 1995) were used whenever counts per turn were analysed, again making use of the StatXact program. Results This section reports first the results of formal tests in German and English and descriptive statistics for the different conversations in terms of turn totals, time taken and distribution of linguistic units and pauses. We go on to examine the sources of trouble, the nature of repair initiation and the strategies used in repair negotiation. Finally we look at the length and success of repair sequences across speakers and settings.

228

L. Springer, N. Miller and F. Bürk

Comparison of the two versions of the AAT According to the AAT diagnostic criteria, aphasia was classified as Broca's aphasia in both languages. Spontaneous speech was characterised by agrammatism (see syntax rating), word retrieval problems (see semantic rating and naming scores), speech apraxia (see articulation ratings and repetition scores). Only in English was there some reliance on automatised language (e.g. idioms, formulaic language). The general trend shows English scores to be more depressed, both in spontaneous speech and on the formal subtests. AAT- subtests (percentile ranks) COM ART AUT SEM PHO SYN TT REP WRIT NAME 99 48 3 German 4 3 4 2 17 119 5 (79) (51) (70) (58) English 46* 33* 2 22 103* 3 3* 3 3 1 (35) (39) (60) (44) * denotes significant differences between English and German. AAT- Spontaneous speech ratings

COMP 112 (98) 100 (85)

Table 1. AAT Performances. Spontaneous speech is rated 0 (absent or unscorable) to 5 (normal range). The rating scales are: COM: communicative abilities; ART: articulation and prosody; AUT: automatized speech; SEM: semantic structure; PHO: phonemic structure; SYN: syntactic structure. The AAT-subtests are: TT: Token Test; REP. repetition; WRIT: written language; NAME: confrontation naming; COMP auditory and written comprehension.

Since the (preliminary) norms for the English AAT (Miller et al., 1997) are very similar to the original German test, as are the reliability (consistency) coefficients, psychometric single case analysis (Willmes, 1985) was employed to compare performances for the five AAT subtests for both languages. Subtests 'naming', 'repetition' and 'written language' were significantly better in German. The profiles of impairment are approximately parallel in the two languages with the exception of 'naming' which is significantly worse in English. Generally the degree of severity was moderate in German and moderate to severe in English. Table 2 presents descriptive information on the two German and two English conversations: namely conversation time, total client turns, the percentage of troublesource-repair-sequences (TSR) and the mean length of trouble-source-repair-units. We extracted the first 212 turns from the German conversation to match the turn-length of the shorter English conversation, while the interviews were matched at the 175 turns of the shorter German interview. Analyses below are based on the 106 and 87 turns respectively that constituted HK's part in these conversations. Conversation times in the informal conversations were shorter than in the interviews, in both languages. In contrast, the number of turns is distinctly higher in the former. Numerically, most of the trouble-sources and repairs appeared in the German interview, i.e. in 34.5% of HK's conversational turns we identified signals where HK or her partner initiated repairs, guesses, repetitions and comments about the troubles. There are, however no significant differences in the relative frequency of TSR (Fisher's exact test, all p >.10, two-tailed, for conversation vs. interviews in each language respectively, and German vs. English for each conversation type). Nevertheless, the mean number of turns to repair troubles is generally short. The highest mean length was required to repair trouble-sources in the English interview. Employing the randomisation test for the independent twosamples problem (Edgington, 1995) a significant difference was found in the mean length

Cross-language conversation analysis

229

of the TSR when comparing the interview in German vs. English (p = .003, two-tailed), and conversation vs. interview in English (p = . 0207, two-tailed).

Conversation time (min.:sec.) Total no. of client turns Percentage of TroubleSource-Repairs (TSR) Mean Length of TSR-units* Range

German conversation

German interview

English conversation

English interview

19:12

23:15

18:22

23:23

106

87

106

87

23.58% 25 2.6 1-10

34.48% 30 1.9 1-5

25.47% 27 2.8 1-9

25.28% 22 4.0 1-18

Table 2. Conversation time, number of turns, percentage and mean length of troublesource-units. *Mean Length of TSR is calculated by averaging the number of consecutive unsuccessful turns that comprise each Trouble-Source-Repair sequence. German conversation n=106 turns Total no. of interjections 187 mean/turn: 1.76 Percentage of client turns with 44.33% just yes/no answers 47 Total length of pauses (sec.) 142 mean/ turn: 1.33 Percentage of 1 sec. pauses 49.05% 52 Percentage of 2-4 sec. pauses 30.18% 32 Percentage of 5-6 sec. pauses 1.88% 2 Micropauses 88

German interview n=87 turns 319 3.66 16.09% 14 146 1.67 52.87% 46 21.83% 19 0 59

English conversation n=106 turns 286 2.69 60.37% 64 249 2.6 39.62% 42 58.49% 62 7.54% 8

English interview n=87 turns 381 4.37 26.43% 23 176 2.02 57.47% 50 45.97% 40 6.89% 6

54

17

Table 3. Number and percentages of interjections and pauses. Mean duration of pauses was longer in the English conversation and interview (Monte-Carlo version of independent two-samples permutation test p = <.001 and <.002 respectively). Long, 5-6 second, pauses are almost exclusive to English. The number of interjections, yes/no answers and pauses and the percentages per turn ate given in Table 3. HK used significantly more interjections in the German and English interview in comparison to the informal conversations. That this difference is not due to just yes/no interjections appearing in elliptic answers is demonstrated in the percentage of yes/no answers. As can be seen, fewer yes/no answers were present in the interviews in contrast to the conversations (Fisher's exact test, p < .0001 for German as well as English). Since we were interested in the consequences of the linguistic impairments, we analysed performances at phrase and word level in the four different conversations. Table 4 presents the mean length of utterances (words), and the number of complete and deviant sentences and the number of one-word utterances.

230

L. Springer, N. Miller and F. Bürk German German conversation interview

Mean length of utterances (MLU) (words/phrase) mean: range: Total no. of one-word phrases (without interjections) Total no. of sentences (without oneword phrases) Percentage of complete sentences (including complete elliptic answers) Percentage of deviant sentences Table 4.

English English conversation interview

3.52 1-8

3.78 1-9

2.92 1-9

2.16 1-7

11

5

12

27

72

201

40

61

22.22% 16 77.77% 56

33.83% 68 66.16% 133

10% 4 90% 36

9.83% 6 90.16% 55

Mean length of utterances (MLU) and number and percentage of complete, deviant sentences and one-word utterances.

The mean length of utterances and the total number of sentences is higher in the German conversations. The ratio of one-word : complete sentences was significantly different between the German conversation and the German interview (p = .0007) and between the German and English interview (p < .0001). The comparison of complete and deviant sentences showed significant differences between the conversation and interview in English (p = .0003). The same tendency was found in German (marginally significant p = .0750). The type of conversation influenced sentence production, with HK producing more sentences in the interviews, particularly for German.

Total no. of open class words Intrusion of other language No. of different words Type-token-ratio: Total no. of nouns Different nouns Type-token-ratio: Total no. of verbs Different verbs Type-token-ratio: Total no. of adjectives Different adjectives Type-token-ratio: Total no. of closed class words Intrusion of other language No. of different words Type-token-ratio:

German conversation n= 106 turns 190 3 119 0.62

German interview n=87 turns 371 0 219 0.59

English conversation n= 106 turns 83 11 57 0.68

English interview n=87 turns 118 14 66 0.55

82 53

161 96

39 28

38 20

0.64

0.59

0.71

0.52

47 37

114 82

13 10

38 20

0.78 58 27 0.46

0.71 96 41 0.42

0.72 20 11 0.55

0.52 25 18 0.72

146

125

54

88

4 37 0.25

0 75 0.60

7 34 0.62

14 32 0.36

Table 5. Number and type-token-ratio of open and closed class words.

Cross-language conversation analysis

231

Table 5 gives the number and type-token-ratio of open and closed class words in the four conversations. Significant differences were found in the distribution of open class : closed class words between the German conversation and German interview (Fisher's Exact Text p < .0001 two-sided). There were no differences between the two English settings. It is clear from these data that despite the shorter turn totals, more content words are produced in the interviews. The type-token ratio for closed class words in the German interview also indicates that a greater range of function words was produced. A comparison with the English interview illustrates that this effect is not simply due to the longer conversation time available. Despite comparable time scales the ratio of open to closed class words is significantly different between the German and English interviews (Fisher's Exact Test p<.0001). As predicted, the direction of intrusions from one language into another was almost wholly unidirectional, from German into English. The only instance of English within an otherwise clearly German utterance was when HK was talking about the problems of keeping her languages separated and she stuck on the word 'und' (and). Table 6 gives the number and percentage of trouble-sources (TS), differentiated according to self-initiated-self repair versus interactive repairs. As HK's spontaneous speech is characterised by agrammatism we would expect that morphological-syntactic TS's would feature prominently. Despite the low ratings on the AAT for syntax, the majority of trouble-sources were lexical-semantic in all four settings. There were only marginally significant differences in the distribution of types of troublesources between the German conversation and the German interview (p = .051). No comparison between any other conversation pairs reachedsignificance.

Trouble-Source (TS) Phonological Morphologicalsyntactic Lexical-semantic Discourse Code-switching Total no. of TroubleSource (TS)

German conversation n= 106 turns SI-SR Other repairs 0 1 0.94% 3 3 2.83% 2.83% 5 7 4.71% 6.6% 0 5 4.71% 0 1 0.94% 8 17 7.54% 16.03%

German interview n=87 turns Other SI-SR repairs 2 0 2.29% 0 1 1.14% 19 9 10.34% 20.68% 0 0 0

0

12 13.79%

19 20.68%

English English interview conversation n=87 turns n=106 turns Other SI-SR Other SI-SR repairs repairs 2 0 2 2 2.29% 1.88% 1.88% 0 1 0 0 0.94% 16 1 8 11 7.54% 10.37% 1.14% 18.39% 0 0 0 2 1.88% 2 1 1 0 2.29% 0.94% 1.14% 20 11 2 16 10.37% 15.09% 2.29% 22.98%

Table 6. Number and percentage of Trouble-Sources (TS) in self-initiated and self-repair sequences and in interactive repairs. Other repairs = interactive TSR: OI-SR, SI-OR, Si-Col R, OI-Col R

Table 7 summarises the general trouble-source-repair-sequences, also broken down according to self initiated self repairs versus interactive repairs. Although TSR totals appear small, the proportion still represents more than one in five turns over a period of 23 minutes. We compared the trouble-source-sequences SI-SR and OI-SR which are most common in normal conversation with other and collaborative repairs (OI-OR, CoR) which are common in clinical settings. There was a marginally significant difference between the

232

L. Springer, N. Miller and F, Bürk

conversation and interview in English (p-value = .056) and the German and English interview (p = .062), otherwise no other comparisons between settings reached significance. German conversation n=106 SI-SR 8 7.54% OI-SR 2 1.88% SI -OR 6 5.66% O1 -OR 5 4.71% SI - Col R 2 1.88% OI - Col R 2 1.88% Total no. of TSR sequences 25 23.58%

Trouble-Source-Repair (TSR) Sequences

German interview n=87 12 13.79% 0 7 8.04% 5 5.74% 5 5.74% 1 1.14% 30 34.48%

English conversation n=106 11 10.37% 1 0.94% 8 (-2) 7.54% 2 1.88% 2 (-2) 1.88% 3 2.83% 27 (-4) 25.47%

English interview n=87 2 2.29% 1 1.14% 9(-l) 10.34% 0 7 8.04% 3 3.44% 22 (-1) 25.28%

Table 7. Number and percentage of Trouble-Source-Repair (TSR)-sequences. SI: self-initiated SR: self-repair - failed repair; OI: other initiated; collaborative repair

OR: other repair;

ColR:

Table 8 shows the types, number and percentage of repair strategies used by HK's conversational partners. The distribution of repair strategies was significantly different between the English conversation and English interview (exact p = .0019). Subsequent comparisons per category revealed that unrelated repairs were significantly more frequent in the English conversation and substitutions were significantly more frequent in the English interview. Unrelated sequences were evidenced solely in the English conversation (mostly topic shifts). Failed repairs were also exclusive to English. Partner Repair Strategy (RP) repetition unrelated elaboration reduction substitution Total no. of RP* (including recursive turns)

German conversation 5 22.72% 0

German interview 6 30% 0

14 63.63% 1 9.09% 2 18.18% 22

12 60% 0 2 10% 20

English conversation 5 15.62% 9 (-4) 28.12% 13 40.62% 3 9.37% 2 6.25% 32

English interview 4 11.76% 1(-1) 2.94% 20 58.82% 0 9 26.47% 34

* interactive repair-sequences (without SI-SR)

Table 8. Number and Percentage of Partner Repair Strategies (RP).

Cross-language conversation analysis

233

As shown in Table 9 (cf. also Table 2) the English interview had, as predicted, the longest repair sequences. The shortest and most successful resolutions occurred in German.

Resolution (RS) most successful and simple successful and simple successful and complex unsuccessful and simple unsuccessful and complex Total no. of RS percentage/client turns

German conversation n=25 TSR SI-SR Other repairs 8 12 32% 48% 0 2 8% 0 3 12% 0 0

German interview n=30TSR SI-SR Other repairs 12 12 40% 40% 0 4 13.33% 0 2 6.66% 0 0

0

0

0

0

8 7.54%

17 16.03%

12 13.79%

18 20.68%

English conversation n=27 TSR Other SI-SR repairs 11 4 40.74% 14.81% 2 0 7.4% 6 0 22.22% 2 0 7.4% 2 0 7.4% 16 11 10.37% 15.09%

English interview n=22 TSR Other SI-SR repairs 8 2 36.36% 0.9% 1 0 4.45% 0 10 45.45% 1 0 4.45% 0 0 2 2.29%

20 22.98%

Table 9. Success and Complexity of Self and Other Repair Resolution. Discussion We have presented a speaker who, despite having attained high levels of proficiency (to senior school leaving level and beyond) in two of her three languages, nevertheless showed some significant differences in performance in these languages after her SAH. In the following we discuss these differences and similarities in relation to our hypotheses and to possible explanations for the pattern of performances found. We hypothesized that following a circumscribed lesion affecting functions of the left anterior cerebral hemisphere HK would not evidence a pragmatic disorder. This indeed seemed to be the case. HK was able to turn-take appropriately, she maintained topics both within and across turns, she relinquished her turn appropriately and took up talk again when the interlocutor signalled a turn end. Even where HK did not or could not signal her turn-taking intentions verbally, it was clear from the video that she appropriately used eyecontact, body posture, hand gestures and prosodic cues (loudness, speed and pitch changes) to hold open her turn or signal her wish to take over a turn. When she experienced problems continuing her turn because of word finding or sentence structuring problems, she was able to hold her turn open through the use of interjections or the nonverbal gestures. In this way she showed intact control of local management systems (Levinson, 1983, Glindemann, 1990) The number of self-initiated repairs and her ability to join in collaborative repairs indicated that she had intact self and other monitoring skills. Her ability to recognise and attempt to repair inappropriate intrusions of one language into the other also testifies to the intactness of her pragmatic skills. The instances where 'discourse' featured as a trouble source (Table 6) actually involved discourse disruptions on the part of her conversation partners. For instance, in the English conversation, some of the failed repairs stem from the interlocutor's failure to recognise that HK was trying to hold open her turn, was engaged in (word or syntax) searching behaviour, or was appealing for assistance in repair. The conversation partner brought about a breakdown by interrupting HK's turn, failing to recognise that HK had more to add; by responding erroneously, based on information from

234

L. Springer, N. Miller and F. Burk

just part of HK's turn; or, because she had failed to understand HK and was unable/unwilling to engage in repair negotiation, inappropriately changed the topic. The profile of impairment across languages was, as predicted, similar (Table 1), though lexical-semantic problems in English did deviate away from the otherwise parallel impairment pattern in other areas. Such a profile might have been expected from the close structural relationship of German and English. It also fits an interpretation of aphasia as being an impairment of central language processes (Bates, Wulfeck & MacWhinney, 1991; Menn, O'Connor, Obler & Holland, 1995), which may nevertheless be affected to differing degrees individually. It also concurs with arguments for the non-separation of languages in the brain, at least as regards neurological localisation (Zatorre, 1989; Paradis, 1992), and the view that if there is apparent differential impairment of languages in an aphasic speaker these relate more to control and resource issues or aspects of language specific structures than to differential impairment of central language processes. The greater facility in German was apparent in the descriptive details of HK's utterances (Tables 4 and 5). In both interactional settings HK produced in German significantly more complete sentences and failed attempts at sentences (deviant sentences). This is reflected in the significant differences in mean length of utterances for the matched settings, though HK does show herself capable on occasions of producing structures in English of equal length to the German sentences. In the analysis of word level units HK's German performance again outranks the English. There are more words and more different words (see type-token relationships) produced in both the German settings compared to the equivalent English ones. If one accepts the presence of pauses and non-specific interjections to indicate searching and planning behaviour, then the comparison of pause times and number of interjections per turn (Table 3) also reveals the greater ease in speaking German. The quantitative scores in favour of German may be interpreted to arise from several sources. German was acquiredlong before English. It has been shown that, at least up to the age of puberty, there is a positive relationship between earlier acquisition and greater proficiency, or resistance to attrition, in a language (Harley & Wang, 1997). Furthermore, HK had not used English for the 13 years before her SAH. Having just completed an intensive course of German speech-language therapy may also have exercised an influence on the relative scores of English and German. We hypothesised that consequent on the comparable pattern of impairments across languages, the profile of trouble sources in both languages would be closely related. This was true, with TS's restricted almost exclusively to lexical-semantic causes (Table 6). Despite the fact that HK was clearly agrammatic, with many instances of morpho-syntactic deviations and heavy reliance on single word utterances (see Tables 1 and 4), these did not feature as TS. This illustrates well the advantages of a conversation analysis approach, which focuses on communicative success and trouble from whatever source rather than making (as it would turn out here) false predictions solely on the basis of formal language scores, structural analyses, or preconceived notions of categories of discourse trouble, divorced from the communicative context and partners. The lack of morpho-syntactic TS's may also reflect the finding (Levinson, 1983; BrintonandFujiki, 1989) that it is generally not considered appropriate for adults in a symmetrical relationship to correct the other's grammar. Regarding the number of TS's arising in the matched settings, our hypothesis that there would be a significantly greater total in English was not upheld (Tables 6 and 7). Nevertheless, if one examines the length of repair trajectories required to solve the troubles (Tables 2 and 9), it is obvious that repair sequences for the English TS's are both significantly longer and more complex. This may be interpreted as a manifestation of HK's

Cross-language conversation analysis

235

poorer formal scores in English. That this is not simply a reflection of less experience at speaking with an aphasic partner on the part of the English interlocutors is seen from a comparison of the English conversation and interview. The partner with less experience (conversation) has the less complex repair trajectories. In this case the greater number of failed repairs in the conversation is more likely to be a manifestation of inexperience. The range of repair interactions (Table 7) was similar across languages, as predicted. There was only one (marginally) significant difference between self initiated self repairs versus interactive repairs, between the English conversation and interview. Three possible accounts for the fewer SI-SR in the interview are possible. It may be a reflection of the greater alertness and support available from the more experienced conversation partner in the interview. It could stem from HK's familiarity with the interlocutor (from whom she had received therapy) and her resultant feeling that either she did not need to monitor herself and/or the knowledge that she would receive support for any TS's. A third possibility may be that HK was less able to self-monitor in the linguistically more demanding setting (see below for arguments concerning this). The range of partner repair strategies used in the two languages was, as hypothesized, similar. Elaboration was the most commonly employed strategy. However, there were significant differences. In particular the 'unrelated' category arose only in English. There were significantly fewer of these in the English interview, while at the same time the number of 'substitutions' employed there was significantly higher. We interpret this pattern of more unrelated and failed repairs in the English conversation as a reflection of the pragmatically less skilled interlocutor. The AAT overall language behaviour ratings show that there was considerably more burden placed on the interlocutor in the English settings than the German ones. The English conversation partner was clearly overburdened at times. By contrast, in the English interview the TS totals point to a similar task facing the interlocutor. On this occasion, though, the partner is able to assist constructively in the negotiation of repairs. The size of the task (compared to in German) this time is discernible in the longer and more complex repair trajectories. Contrasts in turn totals and formal linguistic accuracy between the free conversation and semi-structured interview settings are worthy of consideration. The interviews lasted longer, yet were associated with a decreased number of turns. The interview turns were characterised by more interjections, less turns consisting of just yes/no, and more turns with sentences or attempts at sentences. A possible explanation for this stems from the nature of the setting and is interpretable within the framework of adaptation theory (Kolk & Heeschen, 1990; 1996). Kolk et al. claim that agrammatic aphasic speakers suffer from an underlying computational processing deficit, causing them to have difficulty in processing language in real time. In the fast exchanges of a normal conversation agrammatic speakers are able to compensate for this by reducing computational load in the planning stage, by for instance producing reduced syntactic structures (in German these are elliptical structures similar to context ellipsis in everyday language) or relying on minimal turns. They thereby increase their conversational speed and achieve better maintenance of their place in a conversation. Heeschen and Kolk found within-subject stylistic variations like this in approximately half of the agrammatic speakers they studied. HK fits this pattern in as far as she is able to change speech styles according to situational/task demands. The compensatory aspect of adaptation theory would suggest that HK would produce more minimal turns, elliptical responses and fragmentary structures in the conversation, which demands faster alternation of turns and only partial control (if at all) of the conversation by HK. By contrast, in the interviews, the directions in the AAT manual for gathering the speech sample encourage the therapist to permit the speaker with aphasia as

236

L. Springer, N. Miller and F. Bürk

much time as possible, giving verbal and non-verbal support to carry on, but not intervening as soon as a silence or a TS occurs. There is therefore not the communicative pressure on the speaker to exchange turns rapidly as in normal conversational settings, leading to less demands on processing speed. The aphasic speaker is granted time to formulate and does not have to spend time and attention on holding or regaining her turn. Although there is no explicit instruction for the speaker to aim for formal accuracy over functional success, the very fact that the interview is the prelude to a formal test may engender this attitude. Between the German conversation and interview (Table 3-5) there is a fall in minimal yes/no turns and one word utterances and the percentage of deviant sentences. There is a rise in the number of complete sentences and range of lexical items. To this extent the picture supports Kolk and Heeschen's contentions. However, there are still apparent costs to HK. There is a concomitant rise in the proportion of interjections and mean length of pauses per turn. These higher percentages of interjections within the interviews could be taken as surface symptoms for the underlying lexical and syntactical problems and the increased aim for formal accuracy. Concurrently there is also a rise in TSR's in the German interview. Thus, although HK, given time, can attempt and does produce more words and more complete structures, the setting only serves to expose more her underlying impairments. As before, however, the prevailing TS's are lexical-semantic, even when longer and more complete utterances occur that potentially could unmask her syntactic problems. The other observation when contrasting the conversations and the interviews, is that a significant improvement is seen between the German settings which does not materialise for English. A possible account for this failure to improve in the English interview is offered below. As regards the intrusions of one language into another, several observations might be made. The only intrusions of English into German are when HK was actually talking about the topic of switching and mixing. The occurrence of German words and phrases in the English exchanges was significantly higher. The number of German intrusions into English was highest in the AAT interview in what was meant to be a monolingual English conversation. The fact that HK knew the therapist was able to speak German may have had some bearing on the number of intrusions. Nevertheless HK was aware of the intrusions and either self corrected or appealed for help with finding the word or phrase. As far as accounts for this profile of intrusions go, two related ideas may be considered. As an agrammatic speaker with an anterior lesion, we may assume that a major problem for HK is that there is a raised threshold of activation to produce utterances (Bates, Wulfeck& MacWhinney, 1991), both syntactic frames, morphological modifications and lexical items. Hence the predominance of canonical forms, fragmentary utterances, high frequency words and high type-token relationship, in both languages, but particularly for English. As a language that had been learned first and used exclusively over the preceding 13 years German would have had the advantage in activation both from assumed higher automaticity (Bates et al., 1991) and from being a more high frequency language, and consequently being more primed for activation, or having a higher resting level in preparation for activation. Further, there could be another cost, in terms of brain resources (von Studnitz & Green, 1997), to activating English over German. Not only would English be expected to cost more in resources to activate the lower frequency, less 'automatised' language, but more resources would be needed to keep the more accessible German suppressed. This may help to explain further significant differences between languages and/or between conversation types. There was a significant improvement in performance in the

Cross-language conversation analysis

237

German AAT interview over the conversation, which was not paralleled in the English exchanges. The greater success in the German interview was interpreted above to stem from the added time available, less conversational pressure and the focus on achieving formal accuracy. In English this may not have been achieved because the continued need to devote resources to suppressing German deprived HK of the advantages that may have accrued from the greater time available. Also, given HK knew the interlocutor could speak German, we may surmise that German remained relatively activated (Grosjean, 1989) for HK, hence also the greater susceptibility to intrusions. HK did not feel she was using German as a conscious or unconscious strategy to activate English words. If it was this, then it was unsuccessful, and, if the supposed monolingual English speakers had not been able to understand German, then it would actually have led to more breakdowns. Another pointer to the problem being one of suppression of German is the observation that the majority of intrusions concern either highly automatised phrases (e.g. oh Mensch - oh goodness) or closely related words-e.g. Musik for music, Vater for father, mein for my. Conclusions We set out to explore the contrasts in performance in HK's two main languages, in particular within the context of the relationship between her formal test scores and her success in conversations. The latter was examined in the framework of conversation analysis, which was able to highlight aspects of communication that were not immediately apparent from HK's AAT results. A main observation was that formal test results, especially in German, pointed towards syntax as being a prime area of impairment, and one that might be beneficially tackled in therapy. By contrast the conversation analyses implicated lexical-semantic sources of trouble as the main causes of communication difficulties, even in German where the formal naming scores were relatively (to English and to syntax) good. Despite overall similar patterns of impairment, differences did emerge between settings and languages which were not accounted for merely by the divergence in apparent severity of impairment in German and English. We interpreted these to have arisen in part from varying opportunities afforded by the different settings (e.g. time and greater control of the conversation for HK in the interviews; willingness/ability of the interlocutor to engage in negotiative repairs) and the extent to which HK was able to exploit the potential advantages or cope with the possible drawbacks of the situation. These contrasts emphasize that identical or like (formal) language impairments have different consequences for communication according to the conditions and participants present in a setting and according to how interlocutors handle the situation. This further underscores the contention that assessment of communication in aphasia is incomplete without a consideration of the effects of the conversational partner and setting and that intervention directed at partner behaviour is as important as any remediation targeting the underlying impairments. Indeed, as shown here, it is only in live communicative settings that one can accurately gauge which formal impairments have the greatest consequences for mutual (mis)understanding. For HK a management programme would ideally have included work directed at her lexical difficulties. Simultaneously though, and integrated with this, much would be gained from therapy addressing the communicative style of significant others in her family and social circle (Miller, 1989; Lesser & Algar, 1995). The conversation analysis framework also provided several parameters that could be used as indicators of progress in

238

L. Springer, N. Miller and F. Burk

therapy. These included the proportion of negative (e.g. failed and unrelated) repairs compared to successful repair types; the length and complexity of repair trajectories; the proportion of self-initiated and self-repairs to ones dependent on others. A simple 'time taken to communicate' measure was not predictive of the level of formal and functional success. Further aspects of HK's language picture were discussed from a neurolinguistic perspective. Without having to posit separate anatomical localisations for her different languages or different functional circuits within a psycholinguistic model of language functioning, it was possible to account for the variations (and similarities) seen between German and English. Specifically, the poorer formal scores in English (both on the AAT and structural description), the absence of an improvement in the interview setting parallel to across German settings and the predominantly unidirectional intrusion of German into English could be accounted for in terms of processing capacity and costs. The greater costs for English in turn could be led back to discrepant ages of acquisition, the length of time English had remained unused and the effects of recent concentration on German in therapy. Acknowledgment—We are grateful to Klaus Willmes for guidance in conducting the statistical analyses for this study.

References Bates, R, Wulfeck B., & MacWhinney (1991). Cross linguistic research in aphasia, Brain and Language, 41, 123-148. Brinton, B., & Fujiki, M. (1989). Conversational management with language-impaired children: pragmatic assessment and intervention. Rockville, Maryland: Aspen Publishers. Edgington, E. S. (1995). Randomization tests (3rd ed.). New York: Marcel Dekar. Ferguson, A. (1994). The influence of aphasia, familiarity and activity on conversational repair, Clinical Linguistics and Phonetics, 8, 143-157. Ferguson, A. (1996). Describing competence in aphasic/normal conversation, Clinical Linguistics and Phonetics, 10, 55-63. Friedland D., & Miller, N. (1997). Conversation analysis of language choice and separation in bilingual speakers with dementia, Disorder and Order in Talk, Conference Proceedings (p. 15). London: University College. Friedland, D., & Miller, N. (in press). Conversation analysis of communication breakdown after closed head injury. Brain Injury. Glindemann, R. (1990). Welche Probleme haben Aphasiker beim turn-taking? In R. Mellies, F. Ostermann, and A. Winnecken (Eds.), Beitrage zur interdisvplinaren Aphasieforschung (pp. 1-29). Tubingen: Narr. Grosjean, F. (1989). Neurolinguists beware: The bilingual is not two monolinguals in one brain, Brain and Language, 36, 3-15. Hamilton, H. (1994). Requests for clarification as evidence of pragmatic comprehension difficulty: The case of Alzheimer's Disease. In R. I. Bloom, L. K. Obler, S. De Santi, and J. S. Ehrlich (Eds), Discourse analysis and applications: Studies in adult clinical populations (pp. 185-199). Hillsdale, NJ: LEA. Harley, B., & Wang, W. (1997). The Critical Period Hypothesis: Where are we now? In: A. De Groot and J. Kroll (Eds). Tutorials in bilingualism (pp. 19-51). Mahwah, NJ: Lawrence Erlbaum Associates. Heritage, J. (1989). Current developments in conversation analysis, In D. Roger et al. (Eds.), Conversation: An interdisciplinary perspective (pp. 24-47). Clevedon: Multilingual Matters. Klippi, A. (1996). Conversation as an achievement in aphasics, Studio Fennica Linguistica 6, Helsinki: Suomalaisen Kirjallisuuden Seura. Kolk, H. H. J., & Heeschen, C. (1990). Adaptation symptoms and impairment symptoms in Broca's aphasia. Aphasiology, 4,221-231. Kolk, H. H. J., & Heeschen, C. (1996). The malleability of agrammatic symptoms: A reply to Hesketh and Bishop. Aphasiology, 1, 81-96. Laakso, M. (1997). Self-initiated repair by fluent aphasic speakers in conversation, Studia Fennica Linguistica, 8, Helsinki: Kirjallisuuden Seura. Lesser, R., & Algar, L. (1995). Towards combining the cognitive neuropsychological and the pragmatic in aphasia therapy. Neuropsychological Rehabilitation, 5, 67-92. Lesser, R., Milroy, L. (1993). Linguistics and aphasia: Psycholinguistic and pragmatic aspects of intervention. London: Longman. Levinson, S. (1983). Pragmatics. Cambridge: Cambridge University Press.

Cross-language conversation analysis

239

Menn, L., O'Connor, M., Obler L., & Holland, A. (1995). Non-fluent aphasia in a multilingual world. Amsterdam: Benjamins. Miller, N. (1989). Strategies of language use in assessment and therapy for acquired aphasia. In P. Grunwell et al. (Eds), The functional evaluation of language disorders (pp. 97-124). London: CroomHelm. Miller, N., De Bleser R., & Willmes K. (1997). The English language version of the Aachen Aphasia Test. In W. Ziegler and K. Deger (Eds.), Clinical Phonetics and Linguistics (pp. 257-265). London: Whurr. Milroy, L., & Perkins, L. (1992). Repair strategies in aphasic discourse Towards a collaborative model. Clinical Linguistics and Phonetics, 6, 27-40. Orange, J., Lubinski, R., & Higginbotham, D. (19%). Conversational repair by individuals with dementia of the Alzheimer's type. Journal of Speech and Hearing Research, 39,881-895. Paradis, M. (1992). The Loch Ness Monster approach to language lateralisation. Brain and Language, 43, 534537. Perkins, L. (1995). Applying conversation analysis to aphasia: Clinical implications and analytic issues. European Journal of Disorders of Communication, 30, 372-383. Schegloff, R, Jefferson, G., & Sacks, H. (1977). The preference for self correction in the organisation of repair in conversation. Language, 53,361-382. StatXact (1989). Statistical software for exact nonparametric inference. Cambridge: Cytel Software Corporation (Version 2.0, 1991). von Studnitz, R., & Green, D. (1997). Lexical decision and language switching. International Journal of Bilingualism, 1, 3-24. Wilcox, A., & Mogford-Bevan, B. K. (1995). Assessing conversational disability. Clinical Linguistics and Phonetics, 9, 235-254. Willmes, K. (1985). An approach to analysing a single subject's scores obtained in a standardized test with application to the Aachen Aphasia Test (AAT). Journal of Clinical and Experimental Neuropsychotogy, 7,331-352. Zatorre, R. (1989). On the representation of multiple languages in the brain. Brain and Language, 36, 127-147.

Appendix Examples of Trouble-Source-Repair (TSR) Sequences: 1. Self-Initiated-Self Repair Sequences (SI-SR) Example (English conversation): 67 HK: er (1.0) riding er instructor urn ah in-struct-ress 108 HK: um um my father has um (5.0 writes down) in London Example (German conversation): 14 HK: (laughs) Apotheke (.) ja genau (laughs) danke () Apotheke gefahren (.) und mein Sohn hat hingele NEIN ICH mich hingesetzt und mein Sohn hat ah vorne ah gewartet 2. Other-Initiated - Self-Repair Sequences (OI-SR) Example (English interview): 168 HK: oh er (1.0) erm (2.0) erm er I now (.) is erm (1.0) erm erm (writes down) dress-age and um (simultans.) 169 FB: sorry? (simultans.) 170 HK: dressage 171 FB: umu 3. Self-Initiated - Other-Repair Sequences (SI-OR) Example (English conversation): 4 HK: and mein (German) eh (2.0) father and mother and (.) sister and (2.0) me has erm (1.0) erm (2.0) erm 5 CH: you've moved on ? 6 HK: yeh yes Example 32 HK: 33 FB: 34 HK:

(English interview): bath and um (2.0) [is] (1.0) erm you went to the bathroom or (simultans. yes (simultans.)

240

L. Springer, N. Miller and F. Bürk

4. Other-Initiated - Other-Repair Sequences (OI-OR) Example (English conversation): 55 HK: yen erm (1.0) um (2.0) er (4.0) n..now [a] er all forgotten 56 CH: all your Italian or? Example (German interview): 11 HK: also äh ähm ähm äh (2.0) Harmonie und Smartie habe ich zur Weide gebracht und äh konnt ich mir schwindelig oder äh (1.0) mein Sohn hat gesagt äh setz dich hin und mach das (1.0) meine Mutter is krank äh ähm vier fünf sechs Jahre alt 12 OB: sechs Jahre is das her? 13 HK: ja 5. Self-Initiated - Collaborative Repair Sequence (SI-ColR): Example (English interview) 138 HK: oh er born er no er no (3.0) I. was. born (2.0) I don't know (laughs) 141 KB: umu 142 HK: I erm 143 FB: you don't know? (laughs) 144HK: ermerermuml ne.. ne (German) er be nein (German) ererm ts(l.0) erm ich nein ich weiß nicht (German) (3 sylls. unintelligible) er 145 FB: you're not (.) er tell me about the city where you were born erm 146 HK: erm 147 FB: is that right? 148 HK: erm doch (German) ah 149 FB: umu which country were you born? 150 HK: um (writes down

) Germany 151 FB: oh Germany 6. Other-Initiated - Collaborative Repair Sequence (OI- ColR): Example (English conversation): 82 HK: erm (3.0) um (2.0) my father has erm (.) um (1.0) tube 83 CH: tube? 84 HK: yeh er: (1.0) er er so um(.) [fater] (gloss: German father) erm (2.0) um mein (German) ah father has erm (1.0) own er (1.0) um(2.0) own um (2.0) erm (4.0) oh Mann (German) (2.0) um tube er (writes/draws) [bazer] 85 CH: umu? (2.0) he he was a salesman of that? 86 HK: er yeh je ja (German) Example 121 FB: 122HK: 123 FB: 124 HK: 125 FB: 126 HK: 127 FB: 128 HK: 129 FB: 130 HK: 131FB: 132 HK:

(English interview): and then you stayed for how long in hospital? (3.0)erno a week or longer? no no no no no (1.0) two months oh really in that hospital or yeh no no (simultans.) no er I er er (1.0) erm umu (simultans.) in the first hospital you stayed yeh erm (1.0) this es (writesdown) er day? yeh one day yeh yeh

Examples of Partner Repair Strategies (RP) 1. Repetition (Example: English interview): % HK: no um no um oxy- gen (writes down) 97 FB: um oxygen 2. Unrelated repairs (Example: English conversation): 122 HK: um mover (1.0) um(l.0) um further er down and (.) er (.) I'm (.) [ts] whats this called um 123 CH: do you think you could better write in English than speak in English?

Cross-language conversation analysis

241

3. Elaborations (Example: English interview): 40 HK: um um(l.0) I have um(.) er (2.0) erm (.) taxi erm 41 FB: umu (.) you called a taxi? 4. Reductions (Example: English conversation): 190 HK: yeh well (.) [ei] er (2.0) erm (3.0) NO (2.0) (unintelligible 1 syll.) er (2.0) I (.) could (.) can't (.) wor (.) diese (German) ehm words er (1.0) wordsn..er (1.0) words erm (2.0) erm 191 CH: which words? 5. Substitutions (Example: English conversation): 130 HK: um(l.) urn(3.0) um(l.0) the german is genauso (German) (laughs) 131 CH: yeh (1.0) the same (3.0) um

Examples of German intrusions into English: Example (English conversation): 45 CH: how old was you then? 46 HK: vier (German) eh (2.0) umfive nein (German) er six nein oder (German) erm five (1.0) years was I Example (English interview): 24 HK: ah ah ich (German) yeh er doctor has erm erm (2.0) got me ah tablets [antsjum (.) was is (German) er erm (1.0) ts doctor said er er Migrane (German) 43 HK: je er mein (German) er little girl um ts erm (2.0) um (.) erm um Examples of English intrusions into German: Example (German conversation): 120 HK: na sicher (.) NOW IS eh ja åhm NOW IS (3.0) ähm (1.0) MIST (.) UP (gloss: mixed up) 150 HK: ja no (.) Schrand oder (.) ähm (3.0) es rensiert ein nicht AND zwei ah (.) Wochen ich (laughs) nachholen ne?

This page intentionally left blank

7. Neurolinguistics, Vol. 11, Nos 1-2, p. 243-251, 1998 Published by Elsevier Science Ltd. Printed in Great Britain 0911-6044/98 $19.00 + 0.00

Pergamon

Author Index -AAarons, D., 141, 151 Adams, C., 61, 63, 75 Ahlsen, E, 193, 205 Alajouanine, T., 105, 116 Albert, M. L., 36, 51, 52, 56, 57, 58, 76, 79, 87, 115, 118 Albyn Davis, G., 5, 10 Alexander, M. P., 3, 8 Algar, L., 237, 238 Allard, L., 12, 20, 53, 76, 157, 172 Alpert, M., 107, 117 Amaducci, L., 87 Andelman, R, 3, 8, 16, 19 Anderson, A., 24, 32 Antinucci, F., 4, 10 Apel, K., 20, 207, 209, 219, 220 Appell, J., 36, 51, 79, 87 Apt, P., 196, 205 Ares, G., 26, 27, 31, 32 Aronson, A. E., 91, 101 Artzy, T., 153, 171 A ten. J. L., 209, 218, 219, 220 Atkinson, J. M., 43, 51, 52 Au, R., 87 Austin, J. L., 9, 19, 76, 180, 188 Avent, J. R., 207, 208, 220

-BBahan, B., 139, 141, 151 Baker, E, 115, 117, 142, 144, 149 Ballard, K. J., 154, 172 Bara, B. G., 3, 6, 8 Barber, R., 154, 170, 171 Barresi, B., 115, 117 Bates, E, 7, 10, 106, 117, 234, 236, 238 Bauer, R. M., 3, 8, 103, 116 Baum, S. R., 89, 90, 101 Bayles, K. A., 33, 35, 36, 37, 46, 51, 56, 57, 58, 71, 75, 157, 172 Beardsley, M., 120, 134 Beckwith. D., 106, 117 Beeson, P. M., 34, 51 Behrens, S. J., 3, 8, 89, 101 Behrmann, M., 195, 205 Bell, M. A., 87 Belleville, S., 32

Benecke, R., 140, 152 Bennett, D. A., 87 Benson, D. F., 3, 8, 106, 116, 154, 171 Ben ton, A. L., 155, 171 Berg, L., 61, 76, 87 Bermejo, F., 87 Berry, T., 115, 117 Bester, S., 207, 220 Beukelman, D., 7, 9 Bick, K. L., 87 Bihrle, A. M., 3, 8, 13, 19, 22, 32, 134 Binder, G. M., 209, 219, 220 Blanken, G., 9, 57, 58, 59, 71, 73, 75 Blomert, L., 209, 219, 220 Blonder, L. X., 90, 101, 106, 115, 116 Bloom, R. L., 3, 8, 11, 13, 15, 16, 17, 19, 20, 34, 51, 53, 75, 76, 107, 111, 112, 116, 171, 238 Blumstein, S. E., 135 Boada, R., 3, 10 Bogousslavsky, J., 87 Boiler, F., 113, 116 Bond, S. L., 55, 72, 75 Bonvillian, J. D., 202, 205 Boone, D. R., 75, Boone, K., 170, 172 Borod, J. C, 3, 8, 13, 15, 16, 17, 19, 103, 105, 106, 107, 111, 112, 115, 116, 117 Bosone, Z., 195, 205 Boss, B. J., 180, 188 Bowers, D., 2, 3, 8, 9, 90, 101, 103, 105, 106, 115, 116, 117 Bradvik, B. B., 89, 90, 101 Branch, M. S., 63, 77 Braun, A. R, 52 Braun, C. M. J., 3, 9, 101 Brentari, D., 140, 141, 152 Brinton, B., 207, 220, 223, 234, 238 Bristow, J., 53, 76 Brookshire, R. H., 2, 8, 13, 20, 75, 157, 171, 172, 220, 221 Brown, J. R., 91, 101, 206 Brown, J. W., 121, 134, 206 Brownell, H. R, 2, 3, 5, 8, 9, 10, 13, 19, 20, 22, 32, 75, 76, 105, 107, 116, 117, 119, 120, 122, 132, 134, 180, 189 Brun, A., 87, 153, 154, 155, 171

244

Author Index

Bryan, K. L., 3, 8, 52, 89, 90, 101 Bryden, M. P., 105, 117 Bub. D., 33, 36, 51 Buck, R., 105, 106, 109, 117 Bugbee, J., 16, 20 Burgess, G. B., 119, 135 Burns, A. F., 106, 115, 116, 120 Butler, K., 20 Butters, N., 52 Butterworth, B., 194, 202, 205, 220 Button, G., 51, 52

-CCaffrey, J., 51 Caligiuri, M. P., 209, 218. 220 Campbell, A., 16, 19 Cancelliere, A. E., 3, 8, 89, 90, 101 Cannito M., 5, 9 Canter, G., 5, 10 Caplan, D., 105, 117 Caplan, L. R., 53, Caramazza, A., 57, 74, 75 Cardebat, D., 58, 71, 72, 73, 75 Carlomagno, A., 55, 75 Caron, H., 105, 116 Carroll, J. J., 3, 5, 8 Caselli, R. J., 153, 171 Casey, N., 51 Causino Lamar, M. A., 34, 36, 38, 39, 43, 51 Chantraine, Y., 6, 8, 21, 24, 32 Chapman, S., 12, 20, 55, 56, 57, 58, 59, 60, 61, 62, 63, 71, 72, 73, 75, 76, 77, 157, 172, 209, 220 Charlebois, N., 3, 9, 101 Chertkow, H., 33, 36, 51 Chiarello, C, 8, 120, 134 Chomsky, N., 1, 8, 105, 117, 138 Christopoulou, C, 202, 205 Chwat, S. E., 209, 220 Cicchetti, D., 106, 118 Cicone, M.. 2, 3, 8, 9, 193, 202, 205 Clark, H. H., 21, 32, Clark, L., 59, 76, Coben, L. A., 87 Code, C, 21, 32, 231 Cohen, R, 141, 152 Cohen, J., 94, 101, Cohen, M, 3, 8, Cole, M., 19, 76, 116, 117, 171, 189 Collins, M. J., 210, 221 Cooper, W., 107, 117

Coren, S., 108, 117 Corosz, M., 75 Coslett, H. B., 3, 8, 101 Craufurd, D., 154, 171 Culebras, A., 87 Cummings, J. C., 87, Cummings, J. L., 170, 172 Curtiss, S., 52, 186, 188

-DDamasio, A. R., 105, 117 Danielczyk, W., 51 Danly, M., 3, 10, 107, 117 Danziger, W. L., 61, 76, 87 Daoust, H., 6, 9, 21, 22, 30, 32 Darby, D. G., 90, 101 Darley, F. L., 91, 101 David, R., 194, 206 Davidson, R. J., 105, 116, 117, 170, 171 Davidson, W., 105, 116, 117, 170, 171 Davis, G. A., 5, 7, 8, 10, 16, 19, 55, 75 Davis, R, 4, 8 Davis, L., 221 Dawson, K., 76 Day, B. L., 152 De Bleser, R, 225, 239 De Groot, A., 238 De Santi, S.. 34, 51, 53, 56, 75, 76, 171, 238 Dean, A., 193, 205 DeCarli, C., 87 Deger, K., 239 Deimling, G. T., 48, 52 Delis, D. C., 3, 62, 75, 155, 171, 193, 206 Deloche, G., 5, 8 Demonet, J.-F., 58, 75 DeSanti, S., 19, 20, 75, 76 Deser, T., 79, 81, 86, 87 Devoe, S., 142, 152, 192, 206 Dick, J. P. R., 152 Dipper, L., 3, 8 Dittmann, J., 9, 75 Dodd, B., 37, 51 Donnell, A., 53, 76 Dopson, W., 106, 117 Downhill, J. E., 106, 117 Doyel, A., 12, 20 Doyle, P., 16, 19 Doyon, B., 58, 75 Drachman, D., 76, 87 Dravins, C, 101

Pragmatics in neurogenic communication disorders Drayer, B. P., 87 Dronkers, N. F., 179, 181, 189 Duchan, J. F., 209, 220 Duffy, J. R., 89, 91, 100, 101, 154, 171, 193, 195, 202, 206 Duffy, R. J., 106, 117, 193, 195, 202, 206 Duncan, P., 108, 117, Duncan, S., 142, 152 Durkin, L., 119, 127, 134 Durso, R., 87 Dworetzky, B., 135 Dwyer, J., 2, 8

-EEdgington, E. S., 227, 228, 238 Edmondson, J. A., 3, 10 Ehrlich, J. S., 19, 20, 33, 34, 35, 51, 53, 59, 71, 73, 75, 76, 171, 238 Ekelman, B., 52 Ekman, P., 118, 191, 192. 197, 205, 206 Elbard, H., 220 Ellis, N., 9 Emmery, K. D., 89, 101 Enderby, P., 37, 51 Englund, B., 153, 154, 155, 171 Erkinjuntti, T., 76, 87

-FFederici, A., 7, 10 Fedia, P., 76 Feldstein, S., 144, 152 Ferguson, A., 223, 224, 238 Ferreira, A., 16, 20 Feyereisen, P., 5, 8, 115, 117, 194, 202, 203, 206 Finlayson, A., 220 Fischer, P., 36, 51 Fisher, M., 87, 227, 228, 229, 231 Fisman, M., 51, 79, 87 Fitzpatrick, P., 115, 117 Flamm, L., 3, 9 Flower, A., 53, 76 Flowers, C. R., 3, 10, 90, 102 Fodor, J., 105, 117 Foldi, N. S., 2, 3, 8. 9, 205 Folger, W. N., 89, 91, 100, 101 Folstein, M. F., 61, 76, 80, 87 Folstein, S. E., 61, 76, 80, 87

245

Ford, J., 157, 172 Fox, H., 116, 170. 171 Franklin, L. R., 75 Frederix, M., 5, 9 Friden, T., 221 Friedland, D., 223, 238 Friesen, W. B., 191. 192. 197, 205, 206 Fujiki, M., 207. 220, 223, 234, 238 Fulton, S., 52

-GGagnon, L., 123, 131, 134 Gallagher, T., 7, 9, 180, 189, 209, 220, 221 Garcia, J. H., 87, Garcia, L. J., 34, 43,44, 51, 56, 76, 157, 171, 174 Gardner, H., 2, 3, 8, 9, 10, 13, 19, 22, 32, 74, 77, 105, 107, 115, 116, 117, 119, 120, 122, 131, 134, 135, 180, 189, 205, 206 Garrett, K., 7, 9 Garrett, M., 135 Garrod, S., 24, 32 Gatterer, G., 51 Gazzaniga, M. S., 121, 135 Gerstman, L. J., 3, 8, 10, 13, 15. 17, 19, 112, 116, 206 Geschwind, N., 105, 113, 117, 118, 194, 206 Gibbs Jr., R. W., 189 Gilhooly, K. J., 127, 135 Giroux, F., 7, 10 Glindemann, R., 233, 238 Glosser, G., 79, 80, 81, 86, 87, 194, 202, 206 Goffman, E., 42, 52, 64, 76 Goldberg, T. E.( 52 Goldberger, J., 51, 75 Goldblum, G., 208, 218, 220 Goldstein, H., 16, 19 Goldstein, K., 104, 105, 106, 117, 121, 135 Goodglass, R, 18, 19. 61, 76, 87, 99, 101, 107, 113. 117, 195, 206 Gorelick. P. B., 87, 89, 90, 101 Goulding, P. J., 53, 153, 154, 172 Goulet, P., 2, 6. 9, 13, 19. 21, 22, 30, 31, 32, 134, 180, 189 Grace, G. M., 153, 172 Grace, G. W., 4, 9 Grafman, J., 87

246

Author Index

Graham, N., 76 Gramigna, G. D., 52, 157, 172, 174 Graves, R, 18, 19, 113, 117 Green, D., 236, 239 Green, G., 7, 9, 49, 52, 209, 220, Green, J., 153, 171 Greenberg, M. S., 105, 118 Grice, H. P., 14, 17, 19, 64, 76, 113, 117, 170, 171, 180, 189 Griesel, R., 10 Grieve, K., 10 Grimm, J., 9 Grosjean, R, 237, 238 Grymonprez, L., 170, 172 Guenard, D., 76 Our, R C, 105, 118 Gurland, G. B., 209, 220 Gustafson, L., 153, 154, 165, 170, 171

—H — Haaland, K. Y., 140, 152 Haas, J., 75 Hadar, U., 194, 202, 205 Hagenlocker, K., 52, 57, 76 Hallett, M., 140, 152 Halliday, M. A. K., 81, 87 Hamby, S., 74, 77, 206 Hamilton, H. E., 34, 35, 36, 38, 40, 41, 42, 44, 46, 52, 56, 72, 73, 76, 223, 238 Hanlon, R. E, 194, 203, 206 Hannequin, D., 2, 9, 21, 22, 31, 32, 180, 189 Hanson, W. R, 154, 171 Hargrove, P. M., 90, 101 Harley, B., 234, 238 Harrington, A., 104, 117, Harrington, D. L., 140, 152 Harris, A. J., 124, 135 Hart, S., 36, 52 Hartmann, A., 87 Hasan, R, 81, 87 Hayes, S., 12, 20 Haywood, C, 19, 106, 116 Head, H., 105, 117, 177, 197, 202 Heaton, R K., 121, 135 Hefter, H., 152 Heilman, K. M., 2, 3, 8, 9, 89, 90, 101, 103, 105, 106, 115, 116, 117 Helm, N. A., 9, 76, 87, 115, 117, 193, 206, 239 Helms, N., 118

Henderson, V. W., 57, 76 Heritage, J., 43, 51, 52, 223, 238 Hermann, M., 194, 206 Hickson, L., 51 Hier, D. B., 2, 9, 34, 52, 53, 57, 76 Higginbotham, D., 223, 239 Highley, A. P., 55, 61, 63, 64, 76 Hildebrand, B. R, 209, 221 Hildebrandt, N., 105, 117 Hird, K., 91, 101 Hirst. W., 2, 9 Hobart, C., 75 Hodges, J. R, 36, 52, 57, 76 Hofman, A., 87 Holland, A. L., 7, 9, 12, 19, 55, 76, 194, 205, 206, 207, 208, 209, 217, 218, 219, 220, 234, 239 Hollingshead, A. B., 108, 117 Holtas, S., 101 Holtzapple, P., 221 Homan, R. W., 3, 10 Hooper, H. E., 155, 171 Horner, J., 154, 171 Hough, M. S., 3, 5, 9, 13, 19, 209, 220 Hubbard, D. J., 221 Hudson, L., 154, 171 Hughes, C. P., 61, 76, 80, 81, 87 Hungerbuhler, J. P., 105, 118 Hupet, M., 5, 9, 24, 32 Hutchinson, J., 34, 43, 52 Hymes, D., 11, 14, 19 Hynd, G., 3, 8

-I-

Illes, J., 41, 52, 157, 171 Ingvar, D. R, 101 Itard, J. M. G., 186, 189

-JJackson, C. A., 52, 154, 171 Jackson, J. H., 104, 113, 117 Jacobs, J. R, 3, 9, 10, 32 Jaffe, J., 144, 152 Jancovic, M., 192, 206 Jefferson, G., 25, 32, 42, 52, 53, 167, 171, 174, 223, 239 Jensen, M., 34, 43, 52 Joanette, Y., 2, 6, 7, 8, 9, 10, 13, 19, 20, 21, 22, 26, 30, 31, 32, 34, 43, 44,

Pragmatics in neurogenic communication disorders 51, 56, 75, 76, 134, 157, 171, 174, 180, 189 Johnson, J., 39, 61, 63, 75, 76, 77 Juarez, L., 7, 10

-KKahn, R. L., 106, 118 Kaplan, E., 61, 62, 75, 76, 80, 87, 99, 101, 107, 117, 155, 171, 194, 195, 206 Kaplan, J. A., 2, 3, 9, 10, 22, 32, Karbe, R, 153, 154, 171 Kaszniak, A. W., 33, 35, 36, 37, 51, 75 Katzman, R., 76, 87 Keenan, E. O., 171, 174 Kegl, J., 137, 138, 139, 140, 141, 151, 152 Kelly, B., 3, 10 Kempler, D., 36, 52, 58, 72, 76, 154, 171 Kendon, A., 142, 152, 194, 206 Kent, J., 107, 117 Kent, R. D., 89, 91, 100, 101 Kertesz, A., 3, 8, 51, 79, 87, 89, 90, 99, 101, 125, 135, 153, 154, 155, 170, 171 Kesler, A., 153, 171 Khoshbin, S., 152 Kim, Y., 113, 116 Kimura, D., 113, 117 King. K, 75 Kinney, J. M, 48, 52, 53 Kirchner, D. M., 14, 15, 20, 113, 118, 180, 181, 189, 208, 212, 217, 220, 221 Kirsner, K., 91, 101 Kleiber, G., 25, 32 Klippi, A., 223, 238 Knoefel, J. E., 51 Koenig, L., 51, 75 Koff, E., 18, 19, 103, 105, 106, 107, 111, 115, 116, 117 Kokmen, E., 87 Kolk, H. H. J., 235, 236, 238 Komori, T., 153, 171 Kontioloa, P., 76 Korczyn, A. D., 87 Kott, E., 153, 171 Kramer, J., 3, 10, 62, 75, 89, 90, 102 Krashen, S. D., 121, 135, 186, 189 Kroll, J., 238 Kromer, J. H., 155, 171 Kurtzke, J. R, 221

247

-LLaakso, M., 79, 223, 238 Laaksonen, R., 76 Laine, M., 79, 80, 83, 87 Lakoff. G., 4. 9, 132, 135 Lalande, S., 3, 9, 89, 90, 101 Lamoureux, M., 2, 9 Landis, T., 18, 19, 113. 117 LaPointe, L. L., 20, 34, 53, 220 Le May, A., 194, 203, 206 Lebert, F., 170. 172 Lebrun. Y., 10, 104, 117 Lecours, A.-R., 2, 9, 32, 116 LeDoux, J., 2, 9 Leech, G. N., 11, 19, 207, 220 Leeper, H. A., 153, 172 Leinonen, E., 55, 76 Lenneberg, E. H., 186, 189 Leonard, L. B., 5, 10 Lepage, Y.. 2, 9 Lesser, R., 7, 9, 10, 33, 40, 43, 48, 52, 103, 105, 117, 170. 172, 226, 237, 238 Levinson, S. C. 38, 55, 76, 180, 189, 207, 220, 233, 234, 238 Levita, E., 209, 219, 221 Levy, E, 192, 206 Ley, R. G., 105, 117 Lezak, M. D., 121, 135 Lhermitte, F., 105, 116 Lieberman, P., 107, 117 Linebaugh, C., 3, 9 Ling, P. K., 3, 9 Loew, R., 139, 140, 152 Logic. R. H., 127, 135 Lomas, J., 207, 220 Lorch, M. P., 18, 19, 103, 106, 107, 115, 116, 117 Lubinski, R., 16, 19, 47, 49, 52, 76, 223, 239 Lund, N. J., 206. 209, 220 Lundstrom, C., 192, 202, 203, 204, 205, 206 Luria, A. R., 121, 135 Lyons, J., 16, 19

-MMacaluso-Haynes, S. M., 12, 20, 53, 75, 76, 209, 221 Mace, N., 37, 52 Mackenzie, I. R. A., 154, 171

248

Author Index

MacWhinney, B., 172, 173, 234, 236, 238 Mann, D. M. A., 153, 154, 155, 171, 172, 240 Manning, J., 119, 127, 134 Manochiopinig, S., 208, 220 Mansson, A.-C., 191, 192, 202, 203, 204, 205, 206 Margolin, D. I., 140, 152 Marie, P., 104, 117 Marin, O., 53 Marsden, C. D., 140, 152 Marshall, J. C., 9 Marterer, G., 51 Martin, A., 76 Martin, R. L., 87 Martino, G., 3, 8 Marttila, R., 87 Masdeu, J. C., 87 Mateer, C. A., 2, 3, 10, 114, 117, 181, 189 Matovitch, V., 221 Matthews, P. J., 57. 76 Mattis, S., 155, 172 May, E. B., 75, 194, 203, 206 McCawley, J., 4, 9 McDermott, N., 172 McDonald, S., 3, 6, 9 McGarr, N. S., 90, 101 McGarry, K., 3, 10 McHugh, P. R., 61, 76, 80, 87 Mclntire, D. D., 75 McKeel, D. W., 153, 171 McKeith, I., 52 McKhann, G., 61, 76, 80, 87 McNamara, P., 79, 87 McNeill, D., 192, 194, 202, 205, 206 McTear, M. R, 209, 220 Mehringer, C. M., 170, 172 Mellies, R., 238 Mena, I., 170, 172 Menn, L., 234, 239 Mentis, M., 34, 35, 43, 44, 52, 157, 172, 174 Mesulam, M.-M., 3, 10, 89, 90, 102, 106, 118, 153, 154, 155, 172 Metter, E. J., 154, 171 Michel, D., 1, 2, 8 Michelow, D., 2, 3, 8, 9, 22, 32, 105, 107, 116, 117, 120, 134 Milberg, W., 123, 131, 135 Miller, B. L., 170, Miller, G. A., 117 Miller, J. W., 153, 171, Miller, N., 223, 225, 228, 237, 238, 239

Milroy, L., 7, 9, 40, 41. 43, 52, 103, 105, 117, 223, 226, 238, 239 Mitrushina, M., 106, 118 Moise, J. R, 180, 189 Molloy, R., 3, 9, 180, 189 Moody, D. M., 87 Moore, R. W., 106, 115, 116 Morley, G. K., 221 Morris, C. W., 180, 189 Morris, J. C., 153, 171 Muir, N., 48, 52 Munoz, D. G., 154, 171 Murasugi, K. G., 3, 10 Myers, J. L., 3, 9 Myers, P., 13, 20

-N-

Nathan, P. W., 114, 118 Neary, D., 53. 153, 154, 155, 165, 170, 171, 172 Neidle, C., 141, 151 Nelson, L. D., 106, 118 Nespoulous, J., 13, 19, 22, 32, 116 Newhoff, M., 12, 16. 20. 207, 209, 219, 220 Nicholas, L. E., 2, 8, 157, 172 Nicholas, M., 18, 19, 57, 76, 79, 86, 87, 107, 115, 116 Niemi, J., 87 North, A. J., 12, 20, 53, 76, 205 Northen, B., 153, 154, 172 Nyquist, L., 75

o O'Brien, M. D., 87 O'Connor, M., 234, 239 Obeso, J. A., 140, 152 Obler, L. K., 3, 8, 11, 13, 15, 16, 17, 19, 20, 33, 35, 39, 51, 52, 53, 56, 57, 58, 59, 75, 76, 79, 87, 107, 111, 112, 116, 155, 171, 234, 238, 239 Ogata, J., 87 Ogrocki, P. K., 53 Okazaki, H., 171 Oldfield, O. D., 32 Oleyar, K., 16, 19 Onifer, W., 123, 133, 135 Opie, M., 7, 10

Pragmatics in neurogenic communication disorders Orange, J. B., 37, 48, 52, 153, 172, 223, 226, 239 Orgogozo, J.-M., 87 Ortony, A., 132, 135 Ostermann, F., 238 Osterrieth, P. A., 155, 172 Ostrove, J. M., 3, 9 Owens, R. E., 11, 20

-P-

Pajeau, A. K., 87 Paradis, M., 1, 7, 9, 10, 234, 239 Parisi, D, 4, 10, 153, 171 Pasquier, F., 170, 172 Patterson, K., 76 Patterson, M., 113, 116 Pearce, S., 6, 9 Pell, M. D., 89, 90, 101 Penn, C, 7, 10, 153, 157, 171, 172, 195, 205, 209, 219, 220 Perecman, E., 9, 13, 20, 32, 116, 117, 134 Peretz, I., 32 Perkins, L., 7, 10, 33, 36, 37, 40, 41, 43, 46. 48, 52, 223, 239 Perlmutter, M., 75 Petersen, R. C., 153, 154, 171 Peterson, R. R., 45, 55, 119, 135 Petit, H., 170, 172 Pick, L., 16, 19, 171 Pickard, L., 207, 220 Pierce, J., 221 Pierce, R. S., 2, 5, 9, 10, 209, 220 Pinizzotto, A., 52 Poissant, A., 32 Poizner, H., 137, 138, 139, 140, 141, 152 Polk, M., 153, 154, 171 Porac, C., 108, 117 Porch, B. E., 207, 210, 220, 221 Potter, H. H., 3, 8, 10, 13, 19, 22, 32, 120, 134 Poulshock, S. W., 48, 52 Powelson, J. A., 2, 8, 107, 116 Prather, A., 3, 8 Prather, P., 3, 8 Prescott, T., 19 Price, D., 76, 87 Prince, J. A., 75, 76 Prins, R. S., 207, 220 Prinz, P., 5, 10

249

Pruning, C, 7, 9, 14, 15, 20, 113, 118, 180, 181, 189, 208. 212, 217, 220, 221

-RRabins, P. V., 37, 48, 52 Ramsberger, G., 113, 118 Randolf, C, 36, 52 Rau, M. T., 48, 52 Raven, J. C, 77, 125. 130. 131, 132, 135, 155, 172, 183, 186 Read, S., 72, 76 Redlich, F. C, 108, 117 Reed, V. A., 208, 220 Rehak, A., 3, 5. 8, 10, 22, 32 Reichle, T., 194, 206 Rembold, K. L., 192, 203, 206 Resurreccion, E., 221 Retherford, K. S., 172, 173 Reuterskiold, C., 18, 20, 113, 118 Rey, A., 155, 172 Reyes, B. A., 157, 172 Richardson, S. M., 209, 221 Riege, W. H., 154, 171 Rinn, W., 2, 8 Rinne, J., 79, 80, 87 Ripich, D. N., 34, 35, 52, 76, 79, 86, 87, 157. 172 Roberts, J. A.. 208, 220, 221 Robinson, R. G., 106, 117, 118 Roger, D., 238 Roman, G. C, 80, 87 Roman, M., 3, 10 Ropper, A. H., 89, 91, 100, 101 Rorie, K., 16, 19 Rosen, L, 101 Rosenbek, J. C, 89, 91, 93, 100, 101, 102 Ross, E. D., 2, 3, 10, 89, 90, 91, 92, 93, 95, 97, 98, 99, 100, 101, 102, 105, f 106, 118 Ross, J., 4, 10 Rothwell, J. C, 152 Rubin, N. P., 153. 154, 172 Ryding, E., 101

-sSabat, S., 34, 35, 39, 52 Sacks, R, 25, 32, 38, 52, 53, 223, 239

250

Author Index

Saddy, J. D., 3, 10 Sadock, J. M, 4, 10 Saffran, E, 53 Salmon, D. P., 52 Sandson, J., 153, 171 Sanmarco, J. G., 193, 206 Sarles, H. B., 105, 118 Sarno, M. T., 87, 209, 219, 221 Satz, P., 106, 117, 118 Schegloff, E. A., 25, 32, 37, 41, 52, 53, Schegloff, F, 223, 239 Scheinberg, P., 87 Schieffelen, B. B., 171 Schiffrin, D., 11, 20 Schlanger, B., 3, 5, 10 Schlanger, P., 3, 5, 10 Schneiderman, E. I., 3, 10 Schnitzer, M., 5, 10 Schwartz, M., 36, 53 Searle, J. R., 180, 189 Searleman, A., 121, 135 Seibert, G. B., 3, 10 Semple, J. M., 36, 52 Seron, X., 5, 8, 9, 24, 32. 115. 117 Shallice, T., 26, 32 Shapiro, B. E., 3, 10, 89, 90, 102, 107, 117 Sheard, C., 208, 220 Shekim, L. O., 34, 53 Shewan, C. M., 157, 172, 173 Shindler, A. G., 46, 52, 53, 57, 76 Shindler, S., 46, 52, 53, 57 Shobe, A. E., 75 Sidtis, J. J., 90, 102 Silverman, J., 3, 9 Simpson, G. B., 119, 123, 135 Simpson, T. L., 3, 8, 9, 119, 134 Siple, P., 150, 152 Ska, B., 6, 8, 13, 19. 21, 22, 32, 76 Skinner, C., 193, 205 Slauson, T. J., 51, 75 Smith, B. R., 55, 76, Smith, L., 207, 221 Smith, M. C., 75 Smith, P. T., 118 Snow, C. E., 207, 220 Snowden, J. S., 36, 53, 153, 154, 169, 171, 172 Soderbergh, R., 192, 206 Sohlberg, M., 181, 189 Sowa, M., 106, 118 Sparks, R., 115, 118 Speedie, L. J., 2, 3, 8, 9, 101 Stadlan, E. M., 76, 87 Stein, S., 2, 9

Stelmach, G. E., 152 Stemmer, B., 3, 7, 10 Stephens, M. A., 48, 52, 53 Stern, R., 12, 20 Stickel, G., 4, 10 Strong, M. J., 153, 172 Stuss, D. T., 3, 8 Sulkava, R., 76 Swinney, D. A., 119, 123, 133, 135 Szetela, B., 106, 118

-TTait, M. E., 154, 172 Tatemichi, T. K., 87 Taylor, T., 4, 8, 52, 117, 118 Terrell, B. Y., 34, 35, 52, 79, 86, 87, 157, 172 Thomas, A. P., 118, 194, 206 Thompson, C. K., 154, 165, 169, 172 Thompson, J. L., 55, 75 Tirassa, M., 3, 6, 8 Tomoeda, C. K., 51, 75, 157, 172 Tompkins, C. A., 2, 3, 10, 21, 32, 89, 90, 91, 102, 122 Town, P., 3, 8 Tucker, D. M., 106, 117 Tuomainen, J., 87 Tweedy, J. R., 3, 8, 16, 19 Tyson J., 3, 8

-uUlatowska, H., 12, 13, 20, 34, 53, 55, 56, 57. 58, 59, 61, 63, 73, 75, 76, 77, 157, 172, 209, 219, 220, 221

y Valenstein, E., 105, 117 Van Allen, M. W., 155, 171 van Dijk, T. A., 4, 10 Van Lancker, D., 72, 76, 77, 90, 102 Vertes, D., 52 Villardita, C., 79, 87 von Studnitz, R., 236, 239 Vrtunski, B., 113, 116 Vuorinen, E., 79, 80, 87 Vygotsky, L., 192, 206

Pragmatics in neurogenie communication disorders

-W-

-Y-

Wagenaar, R., 207, 220 Wagner, C. M, 2, 10 Wales, R., 3, 9 Wallesch , C.-W., 9, 75 Wang, W., 234, 238 Wapner, W. A., 2, 3, 8, 9, 22, 32, 74, 77, 105, 117. 134, 205 Wechsler, D., 60, 155, 172, 183, 186, 225 Weiman, A. L., 105, 118 Weiner, M., 142, 152 Weinstein, E. A., 106, 116, 118 Weintraub, S., 3, 10, 89, 90, 102, 153, 154, 165, 172 Weiss, D., 210, 221 Weiss, P., 140, 152 Welkowitz, J., 3, 8, 16, 19, 107, 116, 117 Wertz, R. T., 89, 93, 102, 207, 208, 210, 220, 221 Weschler, D., 108, 118 West, J. A., 221 Weylman, S. T., 3, 10 Whitaker, H. A., 3, 8, 9, 36, 53, 76, 101, 135 White, B., 90, 106, 116 Whitehouse, P., 52 Whittaker, J., 52, 157, 172, 174 Whitworth, A., 33, 48, 52 Wiener, M., 80, 87, 192, 194, 206 Wiggs, C, 52 Wilcox, A., 239 Wilcox, M. J., 5, 7, 8, 10, 16, 19, 20 Wilkinson, L. C., 192, 203, 206 Williams, S., 5, 10, 19, 20, 75, 116, 118, 220 Willmes, K., 225, 228, 238, 239 Windebank. A. J., 153, 171 Wing. A. M., 140, 152 Wingfield, A., 3, 5, 8 Winnecken, A., 238 Winner, E., 3, 10, 122, 131, 135 Wolf, G. L, 102 Wolf, P. A., 87 Wollner, S. G., 209, 220 Worrall, L., 51 Wulfeck, B., 7, 10, 234, 236, 238 Wurm, S. A., 206 Wyke, M. A., 206

Yamaguchi, T., 87 Yaretzky, A., 153, 171

-zZachary, R., 125, 135 Zaidel, E., 107, 118 Zatorre, R, 234, 239 Zelinski, E. M., 58. 76 Zettin, M., 3, 6, 8 Ziegler, W.( 239 Ziff, P., 119, 135 Zoghaib, C, 220 Zurif, E, 7, 10, 115, 117, 205, 206

251

This page intentionally left blank

J. Neurolinguistics, Vol. 11, Nos 1-2, p. 253-257, 1998 Published by Elsevier Science Ltd. Printed in Great Britain 0911-6044/98 $19.00 + 0.00

Pergamon

Subject Index -A-

-B-

Aachen Aphasia Test (AAT), 225, 226, 227, 228, 231, 235, 236, 237, 238, 239 abstract attitude, 105, 121 access, 6, 10, 12, 75, 119, 123, 124, 131, 133, 135, 191, 195, 201, 203, 204 adaptor, 191, 192, 198, 202, 205 affect display, 191, 192, 198, 202, 204 affective facial expression, 116, 138, 140 age of acquisition, 127, 135 agreement, 25, 26, 94, 98, 101, 138, 139, 144, 151, 157, 158, 170, 203 Alzheimer's disease (AD), 51, 52, 53, 55, 56, 57, 58, 59, 60, 61, 65, 67, 68, 69, 71, 72, 74, 79, 80, 81, 82, 83, 84, 85, 86, 87 ambiguity, 6, 7, 40, 64, 113, 119, 120, 121, 122, 123, 124, 131, 132, 133, 134, 135 American Sign Language (ASL), 137, 138, 139, 140, 142, 143, 144, 145, 149, 150, 151, 152 aphasia, 3, 7, 8, 9, 10, 12, 13, 15, 16, 17, 18, 19, 20, 21, 32, 40, 43, 52, 55, 56, 57, 58, 59, 60, 61, 65, 68, 69, 71, 72, 74, 75, 76, 87, 91, 92, 99, 101, 103, 104, 105, 106, 108, 110, 112, 115, 116, 117, 118, 122, 125, 135, 151, 152, 153, 155, 157, 158, 171, 172, 179, 181, 182, 189, 191, 193, 194, 195, 196, 202, 205, 206, 207, 208, 209, 210, 211, 215, 217, 219, 220, 221, 223, 224, 225, 228, 234, 237, 238, 239 assessmenj, 8, 9, 10, 20, 32, 33, 35, 47, 48, 5&, 55, 73, 76, 108, 117, 135, 157, 167. 172, 189, 193, 205, 207. i 208, 209, 220, 221, 224, 237, 238, 239 attention, 1, 7, 22, 25, 33, 35, 36, 37, 38, 43, 47, 48, 55, 57, 58, 60, 80, 103, 105, 107, 124, 125, 126, 127, 133. 137, 140. 141. 142. 143, 144, 145, 146, 147, 148, 149, 150, 151, 153, 154, 155, 170, 202, 236 attentional system, 137, 139, 150, 151

baton, 191, 192, 194, 198, 200, 201, 202, 203, 204 Boston Diagnostic Aphasia Examination (BDAE), 61, 87, 99 Boston Naming Test, 83, 87, 130 Broca's aphasia, 9, 99

-ccaregiver (see also conversational partner, interlocutor), 33, 34, 35, 36, 37, 38, 47, 48, 49, 50, 52, 53, 155 closed head injury, 6, 238 cognition, 8, 32, 51, 75, 101, 104, 105, 106, 107, 116, 117, 135, 152, 153, 172, 189 coherence, 8, 19, 34, 43, 52, 57, 65, 79, 80, 81, 82, 83, 84, 85, 86, 87, 172 common reference, 23, 25, 31 communication environment, 16, 35 communicative competence, 12, 13, 14, 18, 55, 56, 57. 71, 207, 208, 209, 220, 224 communicative intentions, 55, 59, 60, 61, 63, 64, 67, 72, 73 compensatory strategy, 10, 149, 191, 195 comprehension, 2, 3, 5, 7, 8, 9, 10, 14, 18, 19, 20, 29, 36, 38, 42, 56, 61, 76, 89, 91, 93, 94, 95, 96, 98, 99, 100, 101, 103, 113, 116, 117, 118, 119, 121, 122, 125, 126, 131, 133, 134, 135, 149, 154, 157, 181, 182, 189, 196. 207, 212. 218. 225. 228, 238 context, 1, 2, 3, 4, 5, 6, 10, 11, 14, 15, 16, 17, 18, 19, 21. 22, 24, 25, 31, 34, 36, 37, 48, 58, 68, 81, 102, 103, 104, 113, 114, 115, 116, 119, 120, 123, 135, 137, 143, 144, 167, 170, 174, 175, 180, 209, 220, 234, 235, 237 Contron»J Word Association Test (FAS), 130, 131 Conversation Analysis Profile for People with Cognitive Impairment (CAPPCI), 48, 50, 52 Conversational Analysis (CA), 19, 33, 34, 36, 37, 47, 48, 49, 50, 51, 52, 135, 171, 220, 223, 224, 234, 237, 238, 239

254

Subject Index

conversational discourse, 24, 32, 34, 37, 51, 75, 107, 172 conversational partner (see also interlocutor, caregiver), 22, 34, 35, 36, 37, 38, 39, 40, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 168, 173, 175, 185, 224, 232, 237 cooperation, 14, 22, 25, 31

-Ddefinite reference, 23, 25, 26, 27, 31 dementia, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 46, 47, 48, 49, 50, 51, 52, 53, 55, 56, 58, 59, 61, 69, 72, 75, 76, 79, 80, 81, 83, 86, 87, 153, 155, 157, 158, 171, 172, 223, 238, 239 dementia of the Alzheimer type (DAT) (see also Alzheimer's disease), 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 48, 172 dementia with Lewy bodies, 38, 39, 46, 52 differential diagnosis, 53, 75, 101 digression, 22, 25, 31, 34 discourse, 1, 2, 3, 4, 5, 8, 9, 10, 12, 13, 14. 15, 16, 17, 18, 19, 20, 21, 22, 24, 30, 31, 32, 33, 34, 35, 36, 37, 40, 43, 44, 46, 47, 50, 51, 52, 53, 55, 56, 57, 58, 59. 60, 61, 62, 63, 64, 65, 67, 69, 71, 73, 74, 75, 76, 77, 79, 80, 81, 82, 83, 85, 86, 87, 103, 107, 111, 112, 115, 116, 120, 137, 138, 139, 141, 153, 154, 157, 158, 165, 167, 169, 170, 171, 172, 173, 181, 189. 205, 209. 220. 226. 231, 233, 234, 238, 239 discourse analysis, 2, 14, 19, 20, 51, 53, 64, 75, 76, 80, 111, 171, 238 discourse macrostructure, 12, 20, 76 discourse process, 13, 58, 59, 60, 63 double dissociation, 5, 6, 8, 18, 22, 122, 134 dyshyponoia, 3, 4, 5, 7

eye contact, 14, 142, 157, 179 eyegaze, 139, 142, 143, 144, 145, 147, 149, 150

-Ffable, 55, 61, 62, 63, 64, 65, 68, 69, 71, 77 facial expression, 4, 6, 7, 14, 15, 94, 103, 109, 110, 115, 116, 117, 138, 140, 141. 143, 147, 150, 157, 177, 184, 185, 187, 188, 191, 195 fluent aphasia, 55, 56, 57, 58, 59, 60, 61, 65, 68, 69, 71, 72, 74, 75, 76, 87, 239 fluent primary progressive aphasia, 153, 154, 155, 158, 163, 164, 165, 166, 167, 168, 169 frame, 55, 61, 62, 63, 64, 65, 67, 68, 72, 76, 77, 197, 223, 236 Frontal lobe dementia (FLD), 153, 154, 155, 158, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172 functional communication, 19, 55, 73, 76, 206, 218, 220

-Ggesture, 1, 4, 6. 7, 16, 89, 94, 95, 96, 98, 99, 100, 101, 103, 105, 114, 115, 116, 141. 142, 143, 144, 167, 184, 185, 187, 189, 191, 192, 193, 194, 195, 196, 197, 198, 201, 202, 203, 204, 205. 206. 212, 214, 233 grammar, 1, 4, 5, 6, 7, 10, 31, 45, 105, 138, 150, 179, 234

-Hhomonymy, 119, 120. 124, 127, 128, 131, 132

-E-

-I-

emblem, 191, 192, 194, 198, 202 emotion, 8, 10, 93, 94, 103, 106, 107, 109, 110. Ill, 112, 114, 116, 117. 153 emotionality, 17, 20, 118

ideograph, 191, 194, 197, 198, 203, 204 idiom (see idiomatic expression), 3, 5, 10, 76, 228 idiomatic expressions (see idiom), 2, 7, 9

Pragmatics in neurogenic communication disorders illocutionary force, 2, 3, 6 illustrator, 191, 192, 194, 198, 201, 202, 203 indirect speech act, 2, 3, 5, 7, 9, 31, 169 inference, 1, 2, 5, 6, 7, 8, 10, 13, 19, 31, 32, 55, 63, 67, 68, 71, 72, 74, 76, 239 informativeness, 14, 79, 80, 81, 82, 83, 84, 85, 86, 172 initiation, 13, 14, 38, 41, 43, 46, 49, 50, 137, 142, 144, 147, 154, 170, 175, 184, 187, 214, 226, 227 Intellectual Quotient (IQ), 108, 125, 130, 131, 132, 133, 183, 186, 225 interlocutor (see caregiver, conversational partner), 2, 21, 22, 23, 24, 31, 34, 35, 38, 41, 43, 103, 137, 138, 139, 141, 142, 143, 144, 145, 147, 149, 150, 151, 163, 168, 176, 223, 224, 226, 227, 233, 235, 237 intervention, 9, 13, 18, 19, 20, 30, 34, 36, 48, 49, 50, 51, 52, 74, 75, 117, 155, 189, 220, 221, 226, 237, 238 intonation (see also prosody), 1, 4, 7, 8, 10, 90, 102, 107, 110, 115, 117, 118, 157, 173, 177, 185, 191, 195, 197

-K-

255

-Mmacrostructure, 12, 13, 20, 76, 77 maintenance, 14, 17, 44, 46, 50, 79, 81, 113, 169, 170, 174, 179, 181, 184, 187, 214, 224, 227, 235 meaning, 1, 2, 3, 4, 5, 6. 7, 8, 10, 13, 31, 58, 59, 62, 63, 64, 65. 67. 68, 71, 72, 74, 75, 77, 90, 117, 119, 120, 121, 122, 123, 124, 126, 127, 130, 131, 132, 133, 134, 135, 144, 175, 180, 188, 189, 191, 207, 226 memory, 1, 7, 25, 30, 35, 36, 39, 42, 43, 44, 51, 57, 58, 60, 62. 72, 73, 75, 79, 133, 135, 137, 139, 140, 153, 155, 170, 171, 172, 196 metaphor, 2, 3, 5, 10, 31, 106, 119, 120, 122. 123. 124, 127, 128, 129, 130, 131, 132, 135 minimal turn, 40, 45, 46. 235 modularity, 5, 6, 105, 117

-Nnarrative, 8, 9, 13, 17, 19, 20, 21, 22, 32, 35, 37, 59, 67, 71, 72, 73, 75, 76, 79, 112, 113, 120, 137, 139 nonliteral meaning, 2, 4, 5, 6, 63, 72, 74

kinetograph, 191, 192, 194 -P-

language impairment, 22, 75, 76, 171, 191, 203, 207, 209, 213, 215, 216, 217, 218, 219, 220, 224, 237 language processing, 3, 4, 13, 117, 119, 123, 224 languages, 124, 138, 191, 223, 224, 225, 227, 228, 231, 233, 234, 235, 236, 237, 238, 239 learned helplessness, 47, 49, 50, 52 left hemisphere damage, 19, 106, 116 lexical ambiguity, 119, 120, 121, 122, 123, 124, 131, 132, 133, 134, 135 lexical decision, 119, 122, 123, 124, 125, 131, 132, 239 linguistic competence, 1, 3, 4, 5, 7, 141 literal meaning, 1, 3, 4, 5, 31, 122, 130

paralinguistic, 1, 2, 4, 7, 14, 15, 55, 58, 106, 107, 116, 170, 181, 184, 185, 187, 191, 195, 201. 202, 207, 212, 213, 214 Parkinson's disease (PD), 38, 87, 91, 137, 139, 140, 141, 144, 145, 147, 149, 150. 151 pause (see also silence), 27, 38, 41, 142, 184. 187, 208. 214, 234 pauses (see also silence), 14, 25, 39, 80, 107, 142, 158, 176, 184, 187, 196, 226, 227, 229, 234, 236 pictograph, 191. 192, 193, 194, 197, 198, 199, 202, 203, 204 planning, 24, 56, 58, 59, 73, 137, 139, 141, 145, 151, 193, 234, 235 polysemy, 119, 120, 124, 127, 128, 130, 131, 132, 134

256

Subject Index

pragmatic, 1, 3, 4, 5, 6, 7, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 25, 31, 32, 33, 34, 35, 36, 47, 51, 52, 55, 56, 58, 59, 60, 61, 63, 64, 67, 68, 69, 71, 72, 73, 74, 79, 86, 103, 104, 105, 106, 112, 113, 115, 116, 117, 118, 137, 141, 142, 151, 153, 154, 157, 158, 165, 166, 168, 169, 170, 179, 180, 181, 183, 184, 185, 186, 187, 188, 189, 191, 202, 203, 204, 205, 207, 208, 209, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 223, 224, 233, 238 Pragmatic Protocol, 14, 181, 183, 184, 186, 187, 208, 212 pragmatics, 1, 2, 4, 5, 6, 7, 8, 9, 10, 11, 12, 14, 15, 16, 17, 18, 19, 20, 21, 31, 34, 56, 75, 76, 137, 151, 153, 154, 158, 168, 169, 170, 172, 179, 180, 181, 185, 186, 188, 189, 207, 220, 221 presupposition, 1, 4, 5, 9, 180 primary progressive aphasia (PPA), 74, 153, 154, 155, 158, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172 Profile of Communicative Appropriateness (PCA), 153, 157, 158, 163, 164, 167, 168, 171, 172, 175. 176, 177 prosody (see also intonation), 3, 5, 6, 8, 9, 10, 14, 51, 89, 90, 91, 93, 94, 95, 96, 98, 99, 100, 101, 102, 103, 106, 117, 118, 170, 179, 184, 185, 187, 196, 197, 204, 212, 214, 228 proverb, 55, 61, 62, 63, 64, 65, 68, 69, 71, 72, 74, 75, 76, 77

-Rrecovery, 19, 106, 221 reference, 6, 7, 23, 25, 27, 139, 152, 169, 176 referential communication, 21, 22, 23, 25, 26, 31 referential task, 22, 25 regulator, 137, 142, 143, 144, 149, 150, 151, 152, 191, 192, 202 Relevance, 8, 180 repair initiation, 226, 227 repair strategy, 224, 227, 232, 235, 239 right hemisphere, 2, 4, 7, 8, 9, 10, 13, 15, 19, 20, 21, 22, 31, 32, 74, 77, 89, 90, 91, 92, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 110,

114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 131, 133, 134, 135, 138, 179, 189, 208 right hemisphere damage (RHD), 11, 19, 90, 107, 134 right hemisphere function, 119, 120, 121, 131 role shift, 139, 142

-ssarcasm, 2, 5, 9, 176 selective access, 123, 133 semantic, 4, 5, 6, 7, 8, 9, 15, 18, 22, 32, 33, 35, 40, 43, 51, 52, 53, 55, 57, 58, 59, 62, 64, 76, 79, 80, 83, 85, 86, 101, 111, 119, 122, 123, 132, 134, 154, 157, 163, 167, 168, 169, 172, 174, 180, 193, 194, 196, 204, 217, 223, 226, 227, 228, 231, 234, 236, 237 semantic processing, 35, 86, 134 semantics, 1, 3, 5, 8, 19, 31, 53, 76, 117, 134, 171, 179, 180, 181, 189, 195, 209 sentence, 1, 3, 4, 5, 6, 7, 8, 9, 10, 12, 22, 31, 34, 53, 59, 61, 93, 94, 95, 101, 107, 133, 134, 135, 137, 138, 139, 141, 143, 144, 149, 151, 152, 180, 181, 188, 189, 194, 196, 229, 230, 233, 234, 235, 236 sentence grammar, 1, 4, 5, 6, 7 shared knowledge, 16, 26, 137, 139 Shipley Institute of Living Scale, 125, 135 signing space, 138, 140, 150 silence (see also pause), 38, 39, 226, 236 simultaneous signing, 144, 147 situational context, 4, 5, 10, 58 speaking turn, 23, 25, 26, 27, 30, 31, 142 speech, 2, 3, 4, 5, 7, 8, 9, 10, 11, 16, 18, 19. 20, 21, 22. 24, 31, 32, 33, 34, 35, 37, 38, 46, 48, 49, 52, 59, 64, 75, 76, 79, 80, 87, 89, 90, 91, 93, 94, 95, 96, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 110, 115, 116, 117, 118, 119, 127, 135, 149, 153, 154, 157, 158, 169, 171, 172, 176, 179, 180, 181, 183, 184, 185, 187, 188, 189, 191, 192, 193, 194, 195, 196, 202, 203, 205, 206, 208, 212, 214. 220, 221, 223, 225, 228, 231, 234, 235, 239

Pragmatics in neurogenic communication disorders speech act, 2, 3, 4, 5, 7, 9, 10, 16, 19, 31, 52, 59, 64, 117, 169, 171, 180, 181, 183, 184, 187, 189, 208, 212, 214

-Ttherapy, 7, 9, 13, 33, 52, 115, 117, 118, 125, 193, 194, 195, 196, 203, 205, 218, 220, 225, 226, 234. 235, 237, 238, 239 topic, 13, 14, 17, 21, 22, 34, 35, 37, 38, 39, 40, 42, 43, 44, 45, 46, 47, 49, 50, 51, 52, 76, 80, 81, 86, 103, 113, 126, 132, 133, 134, 138, 153, 154, 157, 158, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 179, 181, 183, 184, 185, 187, 206, 212, 214, 223, 224, 226, 227, 232, 233, 234, 236 topic management, 37, 38, 43, 46, 49, 52, 157, 158, 166, 168, 172 traumatic brain injury, 61, 119 treatment, 13, 16, 18, 19, 20, 49, 55, 56, 72, 73, 75, 108, 207, 208, 209, 210, 211, 212, 213, 215, 217, 218, 219, 220, 221, 225 turn-taking, 145, 146, 149, 151, 208, 214

-uutterance, 2, 3, 4, 5, 6, 8, 22, 23, 24, 25, 27, 31, 38, 40, 79, 80, 81, 82, 84, 85, 86, 90, 104, 107, 137, 141, 145, 149, 154, 157, 158, 165, 166, 167, 168, 169, 170, 173, 174, 175, 179, 180, 181, 184, 185, 194, 196, 224, 226, 229, 230, 231, 234, 236

-VVascular dementia, 87 verb of motion, 139 verbal communication, 1, 7, 9, 21, 22, 23, 25, 31, 32, 56, 169, 189, 193, 194, 205, 206 verbosity, 56, 69, 71, 74

257

-wWestern Aphasia Battery, 99, 125, 135, 155, 171, 182 Wisconsin Card Sorting Test, 121, 135

How to create DLC boot 2022 on USB?

What is DLC boot?