Eric Manders
Katholieke Universiteit Leuven (Belgium)
Faculty of Medicine
Dpt ENT/Speech-Language Pathology Kapucijnenvoer 7
B 3000 LEUVEN (Belgique)
Tél. : 0032 16 332364
Télécopie : 0032 16 332335
e-mail : eric.manders@uz.kukeuven.ac.be

The interest in pragmatics in the area of language pathology can be considered as an offshoot of the shifts that were first noticed in general linguistics (Searle 1969, Austin 1962). The accent moved from a formal language approach, concentrating on syntax and semantics, to a functional pragmatic linguistic theory, emphasizing language use in context.

In the study of normal language acquisition a similar change of accent was noticed. For decades structural aspects such as syntax and phonology had been studied. Later on semantic features were the main topic of research. Since the middle of the 1970s the interest in pragmatics was gradually rising. Authors such as Dore (1974) and Halliday (1975) tried to make inventories of the preverbal and early verbal communicative intentions of young children.

The next logical step was to introduce these findings into research on language pathology. Besides problems with structure and content, most children and adults with language disorders also show difficulties with language use. Therefore there is a clear need for pragmatically based assessment and therapy procedures.

A comprehensive assessment procedure should evaluate all the relevant parameters of pragmatics: (1) communicative intention, (2) presupposition, (3) social organization of the conversation or discourse and (4) contextual features (Roth & Spekman 1984). The notion of context is essential in studying pragmatics. When context changes, an utterance with the same sound and sentence structure can suddenly take on another meaning. The sentence ‘Hey these cookies look nice’ can be a pure observation (declarative function) but can also be a indirect request to get one (imperative function). Communicative intentions (or functions) have to do with the meaning a speaker wants to get across to the listener. Even very young children demonstrate a whole range of different communicative functions such as requesting, protesting, greeting and drawing attention. Adequate presuppositional abilities are a necessary component for efficient and constructive communication. Being able to take the listener’s perspective is essential. It is usually referred to as ‘role taking’. Therefore one has to adapt his or her language use on parameters like the age and the social status of the communication partner, the channels available for communication and the environment in which the communication takes place (Roth & Spekman 1984). Social organization of conversation or discourse includes some more or less formal parameters such as initiating or terminating discourse, conversational repair and revision strategies and staying on topic for an appropriate amount of time.

Any assessment of pragmatic abilities that claims to be exhaustive will have to include all aspects mentioned above. Three procedures can be used to achieve this aim: (1)naturalistic observation during free play or discourse, (2)manipulated observation, and (3)structured elicitation tasks.

There is a tradition in functionally orientated communication tests in the field of neurogenic language disorders. Sarno introduced her Functional Communication Profile in the late sixties (1969). We also mention other procedures such as the Communicative Abilities in Daily Living (Holland, 1980), the Edinburgh Functional Communication Profile (Skinner 1984) and the ANELT (Amsterdam Nijmegen Everyday Language Test, Blomert 1991). We used some of the existing tasks, combined with informal procedures, to evaluate the pragmatic (dis)abilities of traumatic brain injured (TBI) subjects (Cools and Manders 1998) and of patients with Alzheimer’s disease (Manders and Van Vreckem 1995) and also to study conversational characteristics of aphasics when talking to communication partners, differing in degree of familiarity with the patient and with the problem of aphasia itself (Ooms and Manders 1995).

For evaluating the communicative abilities of 10 TBI-subjects (Cools & Manders 1998), we used a test procedure, consisting of seven tasks. These included language comprehension, functional language abilities, procedural function, narrative abilities and the ability to make abstractions. General pragmatic function was assessed by means of the pragmatic protocol (Prutting & Kirchner 1983,1987). Our results confirmed recent findings in the literature, showing that TBI-patients have problems with procedural function, descriptive function, appropriate use of cohesion relations and the adequate production of narratives. Our patient group also showed slight problems in functional communication as measured by the ANELT and in some aspects of general pragmatic ability, as shown with the pragmatic protocol.

In our comparative study of language functioning in normal and abnormal ageing as compared to the language of aphasics (Manders & Van Vreckem 1995), we also used the pragmatic protocol which is based on the observation of the spontaneous, unstructured conversation with at least one conversation partner during 15 minutes or more. The results showed that normal elderly persons are quite good communicators from a pragmatic point of view. Only 3,2 % of all pragmatic parameters was scored as inappropriate in this group. In aphasics the percentage of inappropriateness increased to 24,7 % and in Alzheimer patients even to 31,3 %. When considering the relative percentage of inappropriate parameters in the three main categories (Table 1), we could see that nonverbal aspects (e.g. body posture, eye gaze) are least affected in the three groups. In aphasics paraverbal parameters (such as articulation, loudness, prosody) were most disturbed. Verbal aspects (turn taking, topic introduction and maintenance, variety of speech acts) were the least appropriate in the Alzheimer population.

Table 1. Percentage of parameters scored as inappropriate within each category for the 3 subgroups

There was no single pragmatic parameter on which the normal elderly performed worse than the pathological groups. On the other hand, there were few parameters on which AD-patients scored better than aphasics. The exception were paralinguistic aspects such as intelligibility (45 % inappropriateness in AD, 86 % in aphasia), fluency (15 % disturbed in AD versus 33 % in aphasia) and prosody (30 % vs. 46 %). This is another proof for the disharmonic profile observed in dementia: more or less formal aspects of communication are preserved until the late stages of the disease, whereas creative and intentional aspects decline from the first stages on.

In another study we used a conversation analysis procedure to evaluate functional communicative abilities in aphasia (Ooms & Manders 1995). We focused in the first place on general communication style, on the degree of conversational symmetry and on communication breakdown and repair. We were also interested in evaluating eventual differences when aphasics converse with communication partners, differing in degree of familiarity to the aphasic patient and to the problem of aphasia in general. Each of four aphasics had a semi-structured conversation with four different partners: (1) a family member, (2) a person who was unfamiliar to the patient and who did not know what aphasia was, (3) a speech and language therapist and (4) another aphasia patient. All these conversations were videotaped for later analysis. The results can be summarized as follows:

• Most conversations progressed in a smooth way, even those with strangers. The global conversational structure was preserved in most cases.
• The aphasic patients did not adapt their way of communicating to their conversation partner. Their way of talking seemed to depend in the first place on factors such as personal style and strategies learned in therapy
• Communication partners on the other hand adapt their communication style depending on familiarity with the patient and on knowledge of the nature of aphasia. Family members and speech and language therapists tend to ask the patient explicitly to name what they see in pictures or what is present in the room. Speech and language therapists seem to have problems to letting go of their role as an instructor. They correct the patient in an implicit or explicit way, using cueing and other therapeutic techniques, although they were asked to converse with the patient in a natural, spontaneous way.
• In comparison to their communication partners aphasics made more use of nonverbal and alternative communication strategies. Verbal responses were as expected shorter and less complex. A lot of minimal and stereotyped responses were observed. Aphasics also used more and longer pauses.
• In most of the conversations a certain degree of asymmetry was noted. Usually the ‘healthy’ person took the lead in structuring the conversation. He/she asked more questions and introduced more new topics. It was striking that in conversations between aphasics the least number of questions was observed. With only one exception the non-aphasic speaker scored higher on average speaking time and number of utterances used.
• It was clear from our observations that most of the communication breakdowns were caused by the aphasic patient. The nature and the degree of breakdowns did not depend on the characteristics of the interlocutor, with only one exception. When conversing with familiar persons few mistakes with presupposition were noted, whereas in talking to strangers aphasics seem to have problems with appropriate role taking from time to time.
• When considering the kind of repair strategies, it was obvious that speech and language therapists and family members used more specific techniques such as phonological and semantic cueing. Aphasics also made use of nonverbal resources (gestures, …) to restore breakdowns.

An implication from this and other earlier studies (Milroy & Perkins 1992, Gerber & Gurland 1989) is the importance of carrying out a thorough conversation analysis as a base for formulating concrete directives towards the patient and his or her caretakers, in terms of defining the most effective repair strategies, the most efficient communication modes and so on.

In the field of childhood language impairments our literature search did not produce a rich harvest of standardized tests for evaluating pragmatics, except for the Test of Pragmatic Skills (Shulman 1985) and the Test of Pragmatic Language (TOPL, Phelps-Terasaki and Phelps-Gunn 1992). We used the latter to evaluate the differences in pragmatic language abilities between the American norm group and a sample of Dutch speaking children with and without language and learning impairments (Manders & De Groote 1998). The specific purposes of the TOPL, as defined by the authors, are the following: (1) to identify those students who are significantly below their peers in pragmatic language skills, (2) to determine the particular kinds of pragmatic language strengths and weaknesses that individual students might possess and (3) to document students’ progress in pragmatic language as a consequence of special intervention programmes (manual, p. 11).

We found in our study that Dutch speaking children performed significantly better on this TOPL than the American norm group. Only 5 % of the children from our group scored below the 50 th percentile, whereas more than 70 % scores above the 75 th percentile. When considering the TOPL-scores of the children with language and learning disabilities, it was obvious that they scored much poorer compared to the children following regular education and compared to the (American) norm group (Figure 1). These results suggest that language-impaired children seem to have problems not only with formal language aspects (syntax, morphology,...) but also with pragmatic abilities.

Figure 1: distribution of scores for the normal language (NL) and the language disordered (LD) group.

Because of the above mentioned lack of formal instruments one often has to resort to more or less informal procedures like conversation analysis (Stroobant, Sinnaeve & Manders 1998) to study pragmatic parameters such as roletaking abilities, adequacy of repair strategies for communication breakdowns and the application of appropriate cohesion relations in discourse. Using a modification of the conversation analysis procedure described by Adams & Bishop (1989), we found important qualitative and quantitative differences between a child/child (CC) and a child/adult (CA) condition. When conversing with adults children tend to use more utterances (p < 0.05), more turns and more utterances per turn (p < 0.01). More topic changes were observed in child/child conversations and less minimal (verbal and non verbal) responses appeared in this condition. More pauses were noted in the CA condition. Older children seem to be better at recognizing transition relevance places. This can be inferred from the observation that they produced less inappropriate overlaps than younger children. The same was true for 6-year olds with normal language development as compared to age matched language impaired subjects.

As to repair behaviour some differences between the CC and CA condition can be observed. Significantly more (p < 0.01) repair sequences were used in the child/adult interaction and children tended to use more selfcorrections when talking to adults.

Our observation of cohesion factors showed that in the CA condition more referents were recoverable from the linguistic context than in the CC situation. Younger children made more use of referents that could be retrieved from the situational context (p < 0.05).

In addition to more or less general procedures such as conversation analysis, it is also possible to carry out in depth studies of specific intentional categories such as story telling or narrative function (Sinnaeve, Stroobants & Manders 1998). Narratives can be considered as very specific types of discourse. They are less contextualized than conversation. Telling stories demands a macrostructure that organizes the discourse unit, in contrast with conversation wherein topic shifts constantly occur and local negotiation strategies are very important. Furthermore story telling demands adequate presuppositional abilities and adequate decision making as to the most effective way to get the message across. The speaker (story teller) has to convey the information in an organized, coherent and interesting way. He/she has to take responsibility for the continuity, accuracy and completeness of the information provided. This means that narrative function is a complex but at the same time very important communicative ability that emerges step by step in the course of children’s development to reach a more or less ‘adult’ form only at the age of about six (Liles 1993). In our own study we used a procedure consisting of (1)story retelling, (2)story telling and (3)story generating. In the first condition an expanded version of the ‘story test’ (part of the Leidse Diagnostische Test, Schroots & Van Alphen de Veer 1976) was administered. A simple story is first read to the child, he/she has to repeat the sentences and finally he/she has to retell the story in his/her own words. In the story telling condition an existing story, known by the child, was to be told (for instance ‘Little Red Riding Hood’). The story generation part consisted of an initial sentence that was introduced (e.g.. ‘There once was a witch who liked cake very much...’). The child had to generate a story based on that sentence.

The stories were analysed on form and on content. The main results regarding formal aspects were the following: (1) younger children used less words per sentence and less sentences per story, (2) significantly more words and sentences per story were used in the story telling condition than in story retelling and generating.

With respect to content, it is worth mentioning that the structural complexity of the (generated) stories tends to increase with age in children developing language normally. While 5-years old still use a lot of descriptive stories (no chronologically related sequence of the actions), 6-years old children make more use of sequential and even plotted stories. In language disordered 6 and 7-years old children the amount of descriptive and sequential stories stays the same and no significant increase in the number of plotted stories is noted (figure 2).

Figure 2 : structural complexity of the generated stories for the different subgroups (RE = regular education, SE = special education for language and learning disordered children)

Pragmatically based assessment procedures can be used as a complement to test procedures focusing on semantics, syntax and/or phonology. Especially when confronted with children or adults who achieve better than expected on formal aspects of language, but who do not seem to apply their potentially present language capacities sufficiently in everyday language use, pragmatic procedures seem advantageous. Standardized and normed testing procedures as well as more informal procedures based on observation of free play or discourse (e.g. conversation analysis) should be further developed. Some of the existing procedures mentioned in our description above seem very promising for clinical implementation in diagnosis, first of all to detect subjects with pragmatic disabilities, but also to differentiate between subgroups of language disordered patients. As a next step it is very clear that principles of this diagnostic approach concentrating on pragmatics will also be used in planning and executing language therapy. Much effort has been made in this direction during recent years.

Adams C., Bishop D. (1989) Conversational characteristics of children with semantic-pragmatic disorder. In : Exchange structure, turntaking, repairs and cohesion, Br. Journal of Disorders of Communication, 24 ; 211-239.

Blomert L., Koster C., Kean M.L. (1991) Amsterdam-Nijmegen test voor alledaagse taalvaardigheid (1), Logopedie & Foniatrie, 63 ; 368-374.

Cools C., Manders E. (1998) Analysis of language and communication function in traumatic brain injured patients. Int. J. of Rehabilitation Research, 21 ; 323-329.

Holland A. (1980) Communicative Abilities in Daily Living, Baltimore, Univ.Park Press.

Liles B. (1993) Narrative discourse in children with language disorders and children with normal language, Journal of Speech and Hearing Disorders, 30 ; 868-882.

Manders E., De Groote N. (1998) Verkennend onderzoek naar de bruikbaarheid van de Test of Pragmatic Language (TOPL) bij Nederlandstalige kinderen, Tijdschrift voor Logopedie & Audiologie, 28 ; 60-65.

Manders E., Van Vreckem C. (1995) La fonction langagière dans le viellisement normal et anormal et dans l'aphasie : un point de vue pragmatique. In: Actes du 2e Congrès Européen, Athènes, Ellinika Grammata.

Milroy L., Perkins L. (1992) Repair strategies in aphasic discourse; towards a collaborative model, Clinical linguistics and phonetics, 6 ; 27-40.

Ooms A., Manders E. (1995) Conversatie-analyse bij afasiepatiënten: een verkennend onderzoek, Logopedie, 8 (6) ; 22-28.

Prutting C., Kirchner D. (1983) Applied Pragmatics, In: Pragmatic Assessment and Intervention Issues in Language, San Diego, College Hill Press.

Prutting C., Kirchner D. (1987) A clinical appraisal of the pragmatics of language, Journal of Speech and Hearing Disorders, 52 ; 105-119.

Roth F., Spekman N. (1984) Assessing the pragmatic abilities of children : Part 1 : organizational framework and assessment parameters, Journal of Speech and Hearing Disorders, 49 ; 2-11.

Sarno M. (1969) The functional communication profile, New-York, Institute of Rehabilitation Medicine, N-Y University Medical Centre.

Sinnaeve K., Stroobant V., Manders E. (1998) Narratieve vaardigheden bij kinderen uit het gewoon en het type-8 onderwijs. Logopedie, 11(3) ; 19-25.

Stroobant V., Sinnaeve K., Manders E. (1998) Conversationele vaardigheden bij kinderen uit het gewoon en het type-8 onderwijs. Logopedie, 11(2) ; 16-23.