A L I N G U I S T I C A P P R O A C H T O T H E D E T E C T I O N O F M I N I M A L L A N G U A G E D Y S F U N C T I O N IN A P H A S I A CLAIRE P E N N , B.A. (Sp. & H. THERAPY) (WITWATERSRAND) Department of Speech Pathology and Audio logy, University of the Witwatersrand, Johannesburg SUMMARY Linguistic tests, based on an expressive language sample and clinical tests of aphasia, were administered to a minimally impaired aphasic subject in an attempt t o compare their relative utility and sensitivity. Tests based on a linguistic framework were found to be far more sensitive to the errors in the subject's language than traditional clinical tests. Specifically, the linguistic tests isolated difficulties with the processing of d e e p structure relations of c o m p l e x sentences. Various reasons for the results are suggested and s o m e points are made regarding the validation of theoretical constructs through aphasic language data. A suggested m o d e l for the classification, testing and therapy of aphasic disorders, based on linguistic theory, is postulated. OPSOMMING 'n Minimaal afasie pasient is onderwerp aan linguistiese toetse, wat gebaseer is o p 'n ekspressiewe taalmonster en kliniese t o e t s e van afasie. Daar is g e p o o g o m onderskeidelik die bruikbaarheid en sensitiwiteit van die t o e t s e te vergelyk, Daar is gevind dat toetse wat o p 'n linguistiese raamwerk gebaseer is, meer sensitief is t.o.v. van die foute wat in die taal van die proefpersoon v o o r k o m , in vergelyking met die konvensionele kliniese toetse. Die linguistiese toetse het meer probleme t.o.v. die verwerking van diep struktuurverwantskappe van gekompliseerde sinne uitgelig. Verskeie motiverings vir die resuitate, sowel as aspekte betreffende die geldigmaking van teoretiese samestellings deur afatiese taalgegewens is aangevoer. 'n Voorgestelde patroon vir die klassifikasie van afasie, wat o p linguistiese teoriee gebaseer is, is gepostuleer. During t h e past few years, t h e T h e o r y of G e n e r a t i v e G r a m m a r a n d experi- m e n t s based o n this t h e o r y have d o n e m u c h t o increase o u r k n o w l e d g e of t h e s t r u c t u r e o f n o r m a l language. It has b e e n s h o w n t h a t s u c h an a p p r o a c h has had p a r t i c u l a r l y valuable a p p l i c a t i o n in t h e testing of language c a p a b i l i t i e s . O n l y very r e c e n t l y has it b e e n suggested t h a t t h e s e m e t h o d s m a y also p r o v e useful in t h e area of language p a t h o l o g y , n o t a b l y , a p h a s i a . T h e T h e o r y of G e n e r a t i v e G r a m m a r p r o p o s e d originally b y C h o m s k y 2 is based o n t h e fact t h a t t h e s p e a k e r of a language can p o t e n t i a l l y c r e a t e a n d u n d e r s t a n d an u n l i m i t e d n u m b e r of s e n t e n c e s . T h i s a b s t r a c t c a p a b i l i t y is re- ferred t o as t h e c o m p e t e n c e of a s p e a k e r a n d is c h a r a c t e r i z e d as a s y s t e m of rules w h i c h can g e n e r a t e an infinite n u m b e r of s e n t e n c e s in a language a n d assign a s t r u c t u r a l d e s c r i p t i o n t o e a c h s e n t e n c e . This s y s t e m o f rules of lin- guistic d e s c r i p t i o n c o n s i s t s of t h r e e m a i n c o m p o n e n t s - t h e s y n t a c t i c , s e m a n t i c a n d p h o n o l o g i c a l c o m p o n e n t s . 2 · 6 · 8 · 1 4 Tvdskrif van tlic Suid-Afrikuanse Vereniging vir Spruak- cn Gi'lioorhcrlkiinde, Vol. 21, Dcscmber 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) 4 Claire Penn The syntactic component which is viewed as the central device in the gram- mar, generates the sentences of the language, assigning to each both an under- lying phrase-marker (which represents the deep structure of the sentence) and, by means of transformational rules, a derived phrase-marker (which represents the surface structure). The meaning of the sentence is derived mainly from its deep structure by means of semantic rules of interpretation and the phonetic realization (i.e. the sound) of the sentence is derived from its surface structure by means of phonological rules. Fig. 1 provides a schematic representation of the model of generative gram- mar as discussed above and shows how its components are presumably related. SYNTACTIC COMPONENT Deep Structure Surface Structure S e m a n t i c a l l y i n t e r p r e t e d d e e p s t r u c t u r e ( M E A N I N G ) P h o n e t i c a l l y i n t e r p r e t e d surface s t r u c t u r e ( S O U N D ) Fig. 1. The Components of a Linguistic Description (adapted from Fodor et a l . 4 ) Studies on the relationship between generative linguistics and aphasia are still- rather limited in number and seem to be more concerned with the validation of linguistic hypotheses than with the more practical issues of aphasia which would concern the language pathologist. For example, aphasic language data has been used to help confirm the reality of the theoretical distinction between competence and performance, the former being what the speaker-hearer actually knows about the language and the latter referring to how he actually uses the language. Some findings have indicated that in aphasia competence is intact while certain performance features are not. (See for example, the work of Schuell et a l 1 9 , Taylor and A n d e r s o n 2 1 , Weigl and Bierwisch 2 4 ). Other re- searchers feel that aphasia has provided sufficient evidence to warrant a modification of the traditional competence-performance d i s t i n c t i o n 2 6 . Further examples of aphasic language being used to validate linguistic hypo- theses may be found in the work of linguists such as Blumstein 1 , Goodglass7 Journal of e Soth African Speech and Hearing Association, Vol. 21, December 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) A Linguistic Approach to Minimal Language Dysfunction in Aphasia. a n d ' T r a i l l 2 2 who have indicated that aphasic language breakdown is system- atic and logical and provides much proof for the distinctiveness of certain rules or levels of the grammar.* . It is evident from the above discussion that while there is an increasing . amount of current research being undertaken on the more theoretical issues of linguistics, as related to aphasic language, there has been a dearth of studies on the clinical application in this area. Nevertheless, there seems to be suffi- cient evidence to suggest that linguistic methodology may contribute a great deal to the clinician's understanding of aphasic language, and provide, in particular, added insight into the classification, testing and therapy of aphasic disorders. Methodology can be developed within the framework of generative linguistics that not only provides insights into the nature of the langu- age deficit in aphasic patients, but is also sensitive to the. severity of the aphasia deficit. . , Further psycholinguistic studies should be able to develop more precise methods of assessing language deficit in aphasia. (Schuell et al.19) The rationale for.the present study stems from the writer's impression that many speech therapists may not be sufficiently aware of the dynamics of language and the factors involved in its comprehension and use. A number of currently used clinical tests of aphasia are apparently insensitive to certain aspects of language breakdown, particularly when this breakdown is minimal in nature. On the other hand, linguistic methods, since,they are so closely related to a model which attempts to account for the entire knowledge a speaker-hearer has about his language 2 , could enable the clinician to test the patient's language in detail and thus assess his abilities more accurately. The present study was thus designed to test the efficacy of linguistic methods for the detection of aphasic impairments in the hope of providing some prac- tical evidence towards an area of cooperation which is as yet rife with theo- retical facts, but limited in clinical application. METHOD The aim of this study is to determine the value of tests based on current linguistic theory in discovering minimal areas of language difficulty in an aphasic subject. An attempt will be made to compare such tests with some currently popular clinical tests of aphasia in order to determine their relative utility and sensitivity. * It is interesting to note that m a n y o f the aphasic language studies under- taken recently w i t h i n a linguistic framework have served to confirm the work o n normal language which has cast s o m e d o u b t o n the validity of s o m e of the main constructs of C h o m s k y ' s original theory. For e x a m p l e , aphasic language data seems to confirm the ideas of McCawley (cited b y W h i t a k e r 2 6 ) w h o suggests that syntax and semantics cannot easily be separated in reality and that semantics, not syntax is the central point in the grammar. See also, L a n h a m 1 0 for a discussion of recent linguistic theory. Tydskrif van die Suid-Afrikaanse Vi'rt'niging vir Spraak- en Gchoorhcidkunde, Vol. 21, Dcscmbcr 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) 6 Claire Penn SUHJHCTS Experimen tal Subject The subject, H.J., a White male aged 57 years, was selected for the purposes of tins study according to the following criteria: Firstly, he had suffered damage to that part of the central nervous system necessary for language and had displayed evidence of a loss of linguistic skills - two necessary criteria for aphasia, according to Tikofsky 2 2 . Secondly, he had an adequate pre-morbid acquisition of language and damage to the Central · nervous system was stabilized. There were no signs of dysarthria or apraxia and little disintegration of intellectual functioning - features which commonly confuse characterizations of aphasic language. He could express himself fairly well and make himself understood. This was an important criterion of selection since testing relied on the expressive avenue to a large extent. Finally, and most important for the purpose of this study, the subject showed aphasic impairments which were relatively subtle and would thus potentially go unde- tected by a therapist or doctor relying on clinical tests which are linguistically unsophisticated. His therapist in fact described him as having "totally ade- quate speech" with no receptive difficulty. Control Subjects 4 control subjects with normal language and no history of brain damage were selected in order to assess the suitability of the experimental tasks for non- aphasic subjects. A wide age-range (from 12-73 years) was chosen specifically to ensure'that the test items were appropriate for any particular age or in- tellectual level. Till·'. TF.STS Κ Μ Ρ LOVED I.N ΤΗ Κ S T U D Y In testing the language abilities of an adult who supposedly has full linguistic competence, the number of potential point? of breakdown is extremely large. It would appear virtually impossible and certainly uneconomical to test every aspect of the linguistic system. It was therefore decided that potential areas of difficulty could best be deduced from an expressive sample, which may be considered as a starting point in a linguistic analysis. 1. Tests for expressive speech. The tests employed to elicit expressive speech were: / a. Spontaneous narrative speech, where the subject was requested to discuss an incident of interest (adapted from L u r i a 1 2 ) . b. Two sequences of pictures from the picture arrangement subtest of the Wechsler Adult Intelligence Scale where the subject was asked to describe each sequence. c. The description of six pictures chosen from the Thematic Apperception Test — an item used successfully by many authors (e.g. VVepman and J o n e s 2 5 ) to elicit a sample of expressive speech. , Journal of the South African Speech and Hearing Association, Vol. 21, December 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) A Linguistic Approach to Minimal Language Dysfunction in Aphasia 2. Linguistic Tests On the basis of a linguistic analysis of this expressive sample, tests were de- signed to assess the subject's competence with respect to a specific aspect of the grammar, namely .complex sentences. A complex sentence is one, in generative grammar terms, whose initial phrase-marker contains more than one node labelled " S " (i.e. where there is sentence embedding). An example of such a sentence is: By a series of transformations, this sentence appears in the surface structure as "The man who is happy eats fruit". The types of complex sentences tested in this study were relative sentences and complement sentences. Four types of complement sentences and 4 types of relative sentences were tested. Table I provides a brief summary of the linguistic features of each type of complex sentence.9 In addition to the 8 types of complex sentences tested, some items were included containing more than one embedding. Various methods were used to test the subject's competence with regard to these complex sentences. Main Tests Recall of Sentences. In this task, the subject was required to repeat sentences to him orally by the examiner. 25 such sentences were presented, all involving embedded sentences. This method of sentence repetition has been used by several workers as a measure of linguistic competence, (e.g. Goodglass7, Luterman and B a r 1 3 , M e n y u k 1 7 ) . Comprehension of Sentences. This method, adapted from W r i g h t 2 8 and L u r i a 1 2 involved asking the subject questions about specific aspects of the sentence presented to him. 27 such items were compiled. Additional Tests Recall of Paragraphs. In this test, the subject was required to recall two para- graphs containing a large number of embedded sentences. The hypothesis made Tydskrif van die Suid-Afrikaanse Vrniging vir Spraak- en hoorhkunde, Vol. 21, Dsmbr 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) ° Claire Penn RELATIVE SENTENCES COMPLEMENT SENTENCES (Where the embedded sentence is embedded on a NP of the matrix sentence and in some way quali- fies that NP) Types Examples 1. A sentence The man who embedded on a is happy, eats subject NP fruit. where the matrix subject equals the embedded subject. 2. A sentence The event the embedded 011 a paper reports subject NP is untrue, where the matrix : subject does not equal the em- bedded subject. 3. A sentence The man caught embedded 011 an the mole which object.NP where had walked on the matrix object his lawn, equals the em- bedded subject. 4. A .sentence I fed the tabby embedded on an cat the gardener object NP where chased. the matrix object does not equal the embedded subject. Types Examples A. Noun Phrase Complements (where the NP is rewritten as a sentence so that the underlying sentence may function as the subject or object of the matrix sentence) 1. Where the The fact that sentence James was stu- acts as a pid, worried subject. Mary. 2. Where the Mary denied sentence acts that James as an object. was actually stupid. B. Verb Phrase Complements (where the complement sentence is immediately dominated by the constituent VP.) 1. Where the ma- Many students trix sentence tend to avoid equals the em- hard work, bedded sentence. 2. Where the The fat ser- matrix sen- geant ordered tence does not the men to equal the em- march, bedded sentence. / 1 / TABLE 1. Descriptions and examples of complex sentences. Journal o the South African Spe and Hearing Association, Vol. 21, December 974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) A Linguistic Approach to Minimal Language Dysfunction in Aphasia with respect to this test was that the subject would recall significantly fewer embedded sentences than normal control subjects. Use of Words in a sentence. Words were given to the subject which optionally or obligatorily take an embedded s e n t e n c e 2 7 and he was asked to use them in a sentence. Where possible, he was then asked questions about his own sen- tence. It was hypothesized that if the subject showed general difficulty with the use of complex sentences, he would, in this task, either avoid the use of an embedded sentence altogether or show breakdown and confusion. Performance Tests. These tests, which require a non-verbal response were included to ensure a valid interpretation of the subject's verbal responses on the preceding tests. Such tests, since they require more concrete responses (e.g. pointing) from the subject than the other tests, were inappropriate for certain types of complex sentences viz. the complement sentences, since these contain abstract relationships which cannot easily be portrayed by visual means. Thus relative sentences only were investigated in this test by means of two types of performance response. Firstly, where the subject had to carry out 9 commands all containing complex sentences using tokens from the Token T e s t 3 . Secondly, where the subject had to point to one picture out of a series of 4 which correctly portrayed the complex sentence presented to him by the Investigator. Four such items were included. Control of Performance Variables All test sentences were controlled for with respect to certain features of a non-linguistic nature known to affect responses on linguistic tasks. This is a very important consideration when attempting to tap the 'competence' of a subject, since it is necessary to eliminate all performance variables. Sentence length. All test sentences were balanced for syllable length (10 or 11 syllables each) so that memory factors could be controlled for. Control sentences were administered to the subject before testing (i.e. simple sen- tences without embedded structures) to ascertain the number of syllables the subject could recall and comprehend without difficulty. It was presumed that if difficulty should occur in test sentences, this would be a function of sen- tence complexity rather than length. Abstraction level and frequency. In view of the fact that aphasic patients often encounter difficulty with words of an abstract nature and similarly have difficulty with words used i n f r e q u e n t l y 1 9 , an attempt was made to include in the test sentences fairly concrete words which appear frequently in everyday vocabulary and which were noted to be present in the subject's spontaneous speech.* Semantic constraints. In many cases, the comprehension of a sentence is facilitated by the semantic constraints found within that s e n t e n c e . 4 , 1 2 i.e. words which are commonly associated together give clues as to the relation- * Unfortunately, the feature of abstraction could n o t b e c o n t r o l l e d for as successfully in c o m p l e m e n t s e n t e n c e s as in relative sentences, since the semantic structure of the former involves more abstract attitudinal factors. Tydskrif van die Suid-Afrikaanse Vrniging vir Spraak- en hoorhkunde, Vol. 21, Dsmbr 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) 10 Claire Penn •ship between sentence constituents. To control for the possibility of the sub- ject in this study relying on such constraints, these were as far as possible, deliberately avoided so that the subject would rely maximally on grammatical relations within the sentence. An example of such a sentence is "The man who likes squirrels, eats nuts." 3. Clinical Tests The clinical tests for aphasia used in this study were selected on a pragmatic rather than a theoretical basis - that is irrespective of the theories underlying each test and the writer's bias for or against a particular theory of aphasia. Instead, the writer consulted some clinicians experienced in aphasia therapy and asked them what aphasia tests they felt to be currently most popular and useful. A further criterion for the selection of these tests was that they should be well standardized and easy to administer. The four tests selected were: a. The Minnesota Test for the Differential Diagnosis of Aphasia (MTDDA) which is a widely-used test that has been standardized on a large aphasic population and is very comprehensive. b. The Sklar Aphasia Scale (SAS) which is based on a modern theory of aphasia and is quick and easy to a d m i n i s t e r 2 0 . c. The Token Test which is purported to be highly sensitive in the detection of minimal receptive aphasic disorders3. d. Luria's tests for receptive speech which are currently very popular and test many aspects in some d e t a i l 1 2 . Table II serves to summarize the tests administered to the experimental sub- ject. Expressive Tests Linguistic Tests Clinical Tests 1. Narrative 1. Main Tests 1. MTDDA speech. a. Recall b. Comprehension 2. WAIS picture 2. Additional Tests 2. Sklar Aphasia arrangement a. Recall of Scale subtest. , paragraphs b. Performance 3. TAT cards responses. 3. Token Test ' c. Use of words / in a sentence. 4. Luria's tests 3. Control for receptive sentences. 1 speech. TABLE II. Summary of tests administered to the subject. Journal of e South African Speech and Hearing Association, Vol. 21, December 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) A Linguistic Approach to Minimal Language Dysfunction in Aphasia P R O C E D U R E H.J: was tested over a six-week period for a total of about 20 hours, during which time, the expressive, linguistic and clinical tests of aphasia were ad- ministered. Each interview was tape-recorded and transcribed immediately afterwards. While administering the clinical tests of aphasia, the investigator attempted as much as possible to conform to the test procedures laid down by the respective test manuals. However, during administration of the lingu- istic tests, there was less strict application of standard test procedures and instructions or test items were repeated where necessary. This was done in an attempt to assess as closely as possible the subject's underlying linguistic com- petence by eliminating performance variables such as poor attention span and memory, known to affect test responses. The control subjects were tested on linguistic tests only. METHOD OF A N A L Y S I S Responses on both expressive and linguistic tests were analyzed within a framework of generative linguistics. The deep structure relations of each sentence were investigated and the linguistic contexts of irrelevant linguistic features e.g. pauses were noted. RESULTS TESTS FOR EXPRESSIVE SPEECH On a surface level, H.J's response to the expressive tests confirmed his thera- pist's diagnosis of "slow, but totally adequate expressive speech." Though a few word-finding difficulties were evident, very few errors of syntax or sequencing were noted and speech was fairly fluent. The only consistent feature of this expressive sample was the presence of a number of long pauses. A detailed analysis of the linguistic context of these pauses revealed a system- atic breakdown in the subject's expression. It was found that in many cases, these pauses occurred in the position of a sentence boundary within sentences i.e. in complex sentences. For example: • We caught the boat / / which was taking us to Madagascar. It was a job / / to get her onto the boat. In phrase-marker terms, the loci of such pauses were thus: (S1//(S2)) and (SI (S2)//S1) LINGUISTIC TESTS The results of the linguistic tests confirmed the generality of H.J.'s difficulty with complex sentences. Main Tests Recall. In general, the recall of each type of sentence replicated expression. While a number of sentences were repeated correctly, in many cases, for both relative and complement sentences, a pause occurred in the position one would expect from an analysis of the subject's expressive speech viz. he fre- Tydskrif van die Suid-Afrikaanse Vrniging vir Spraak- en Ghoorhkunde, Vol. 21, Dsmbr 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) 12 Claire Penn quently paused at the sentence boundary within a sentence. Furthermore, he frequently requested the repetition of the sentences by the investigator before he was able to recall them. With the control sentences on the other hand, (i.e. those containing no embedded constructs), the subject never requested the repetition of a sentence and, furthermore, showed no evidence of the pausing described above. Since these control sentences were matched with the test sentences for everything except sentence-embedding, it seems valid to assume that the pauses and the need for repetition noted jn test sentences were thus related to the specific linguistic features of these sentences. Comprehension. This task proved to be a highly sensitive detector of lingu- istic difficulty. For all types of complex sentences, H.J. showed a consistent confusion of the deep relations within the sentence. An example of this is: The girl the thin boy saw was athletic. Examiner Subject's response Was the boy athletic? The boy was athletic. Did the girl see the boy? The girl saw the boy. Was the girl thin? The girl was thin. Was the girl athletic? The girl wasn't athletic. Did the boy see the girl? The boy saw the girl. Was the boy thin? The boy was athletic so I think he wasn't thin. Thus it appears that the subject is showing lack of comprehension charac- terized by a gross crossing of sentence boundaries. At the same time, he is showing inconsistency i.e. in the same sentence, he indicates that "the girl sees the b o y " and "the boy sees the girl". Once again, H.J. repeatedly re- quested a repetition of the test items. These factors, together with the fact that questions on control sentences were consistently answered correctly and • never had to be repeated, seem to provide further evidence for the hypothesis that the subject's area of difficulty lies in the processing of complex sentences. A D D I T I O N A L T E S T S Results of additional tests confirmed to a large ex'tent findings on the main tests. In the paragraph recall task, H.J. recalled only 5 out of a total of 23 embedded sentences (as opposed to an average of 16 with the normal subjects). He tended to avoid a complex sentence where possible when asked to use words in a sentence optionally taking an embedded sentence, and long'pauses were evident when he was given words which obligatorily take an embedded sentence. Furthermore, he showed hesitancy and confusion on the perform- ance tasks involving the comprehension of complex sentences. CLINICAL T E S T S The MTDDA. H.J. displayed extremely mild errors on this test. Of chief interest to the writer were his responses to the test items involving oral and aural language abilities. In the two sections involving auditory skills and speech and language skills, he made I6V2 errors out of a total of 266. The Journal, of [lie South African Speech and Hearing Association, Vol. 21, December 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) A Linguistic Approach to Minimal Language Dysfunction in Aphasia majority of these errors seemed to be related to an impairment in auditory retention span. The SAS. H.J.'s score on the SAS was 5,5% indicating "n o impairment" according to the test manual. The Token Test. H.J. made only one error out of a total of 51 items on this test. Though the authors of this test do not specify the number of errors which need to be made before a subject is considered aphasic, it appears most unlikely that a single error would indicate this. Luria's tests for Receptive Speech. H.J. made 3 errors on Luria's tests. One of these errors which occurred in the test for "complex grammatical construc- tions" involved precisely that sentence construction which was examined in linguistic tests viz. the complex sentence, and thus indirectly confirmed the results of these tests. Table 111 summarizes the subject's performance on the linguistic tests and clinical tests for aphasia. LINGUISTIC TESTS RESULTS 1. Recall of complex sentences. 2. Comprehension of complex sentences. 3. Recall of paragraphs. 4. Use of words in a sentence. 5. Performance responses. 6. Control sentences. 1. Errors made on 1 5 out of 25 sentences. 2. Errors made on 20 out of 27 sentences. 3. 5 out of 23 complex sentences recalled. 4. Evidence of delay and confusion. 5. Errors made on 6 out of 13 sentences. 6. No errors made. CLINICAL TESTS RESULTS 1. MTDDA 2. S.A.S. 3. Token Test 4. Luria's tests for receptive speech. 1. "Simple aphasia" with most errors in the visuo-motor section. -· 2. "No impairment". 3. 1 error made out of 51 instructions. 4. 3 errors made. TABLE 111. Summary of subject's responses on linguistic tests and clinical tests for aphasia. Tydskrif van die Suid-Afrikaanse Vrniging vir Spraak- en Ghoorhkunde, Vol. 21, Dcmbr 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) 14 Claire Penn QUALITATIVE ANALYSIS OF THE SUBJECT'S RESPONSES During testing, the ease with'which H.J. performed on the respective tests was noted, since it was felt that such an observation would help to confirm quan- titative results. In general he appeared to find it much easier to respond to clinical tests than to linguistic tests. He completed most items in the former with relative facility and showed little fatigue. On the other hand, he showed consistent difficulty with the linguistic tests — he frequently requested rest periods and many delayed responses were noticed. His comments after testing confirmed the fact that linguistic tests were more difficult. He further reported that he would feel 'insulted' if some of the items from the clinical tests were given to him normally because of their extremely basic nature. RESPONSES OF C O N T R O L SUBJECTS The performance of the control subjects on the linguistic tests was indistin- guishable from what one would expect from a normal speaker of English. Though the group was small, this was taken as evidence for the suitability of the tests for non-aphasic subjects of differing ages and intellectual abilities. DISCUSSION LINGUISTIC T E S T S V E R S U S CLINICAL T E S T S The results of this study seem to have confirmed to a large extent the hypo- thesis that tests based on a linguistic framework are more sensitive in detecting areas of subtle difficulty in-an aphasic patient than some currently used clini- cal tests. The possible reasons for this are manifold: Firstly, one of the most striking features of the linguistic tests in comparison to the clinical tests is that the former were based on a large expressive sample of speech. A linguistic analysis of this sample provided very important guidelines for further testing. The clinical tests on the other hand, provided only limited samples of speech and were thus not adequate for the purposes of a language analysis. The scoring methods on the clinical tests precluded many important features of the subject's performance. For example, some errors on the SAS were noted as being 'grammatically inappropriate' but the pass-fail system of scoring ignored the nature of these errors. This appears to be a major short- coming of many clinical tests if one considers the chief aim of testing to be a detailed description of a patient's language abilities. 1 8 / / The linguistic tests used in this study attempted to assess both the expression and reception of the subject with regard to the grammatical structure in question. This is considered to be of great importance, since only with a knowledge of b o t h avenues, can one hope to gain a near adequate character- ization of language c o m p e t e n c e . 2 · 2 1 A shortcoming of the clinical tests (with the possible exception of Luria's tests) is that very rarely are the same specific language elements examined in equal depth in expression and reception. It would seem that in such tests, reception is regarded as an entity rather than as an important part of language, in fact more "rudimentary" than expressive language. 2 1 Journal of the South African Speech and Hearing Association, Vol. 21, December 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) A Linguistic Approach to Minimal Language Dysfunction in Aphasia Thus, though the writer is not suggesting that the clinical tests used here were totally insensitive to areas of difficulty in the subject's language, the reasons outlined above may possibly account for the fact that they were of limited value in detecting the specific areas of difficulty encountered b y the subject viz. an inability to extract or retain the deep structure relations in complex sentences. It is thus not surprising that he reported little difficulty on clinical tests. A LINGUISTIC E X P L A N A T I O N OF R E S U L T S The question that remains to be answered concerns a linguistic characteriza- tion of the subject's difficulties. Is the confusion he is showing with complex sentences reflective of a basic lack of competence or can it be related to per- formance variables? It might be tempting to suggest, as W h i t a k e r 2 6 does, that since the modalities of both reception and expression are involved, the com- petence is impaired, but such a proposal cannot account for the variability in the subject's responses. He did not show consistent difficulty and his responses were seldom totally incorrect. Furthermore, there was evidence of paraphrasing, which according to Fodor et a l . 4 , is an indication of intact com- petence. Clearly then, H.J.'s difficulties seem to be related to performance variables. It is less easy however, to specify the actual nature of these performance vari- ables. One possible interfering variable, suggested by workers such as Schuell et a l . , 1 9 is a decrease in retention span. According to this theory, H.J. might not be retaining the sentence long enough to recall it or to perceive its deep relations. A further possible interfering variable has been indicated by L e n n e b e r g 1 1 who suggests that faulty 'timing' in aphasia may lead the subject to rely on surface structure rather than deep structure clues for comprehen- sion. Though H.J.'s performance may indeed be related to one or b o t h of these performance variables (or many of the others which were not controlled for and which cannot be specified) it is the writer's belief that these variables are not generalized over the entire language as Schuell et a l . , 1 9 suggest. Rather, they occur within the specific linguistic context of an embedded sentence. H.J. is unable to relate portions of semantic information from one sentence to another when these sentences are in the matrix-embedded relationship. This idea is supported by the fact that he showed no difficulty with the con- trol sentences and lends confirmation to the belief that the " . . . best expla- nations of performance have come from a consideration of features from the underlying grammar." 2 SOME T H E O R E T I C A L C O N S I D E R A T I O N S Some results of this study have been of interest in helping to clarify a con- troversial issue arising out of linguistic theory, namely the Derivational Theory of Complexity (DTC). This theory predicts that sentence comprehension involves the diverse application of linguistic rules necessary t o generate a sen- tence. Thus, for example, a subject would take longer to comprehend a passive sentence than an active one since the former is transformationally more complex. Tydskrif van die Suid-Afrikaanse Vrniging vir Spraak- en hoorhkunde, Vol. 21, Dsmbr 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) 16 Claire Penn If one considers the adjective in transformational grammar terms, the length of the derivational history of a sentence·is greatly increased by the addition of an adjective before a noun. For example, the phrase "the good m a n " is de- rived from 3 basic transformations — " t h e man who is good" " t h e man good" "the good man". Thus in terms of the DTC, recall or comprehension of sentences containing adjectives would be poorer than recall of a simple sentence, or for that matter, one containing a relative clause only, this sup- posedly being the first step in an adjective transformation. The DTC would thus predict that H.J. would have great difficulty with sentences containing adjectives. This however, was found clearly not to be the case. Even in the Token Test where commands involve up to 4 adjectives per sentence (i.e. in transformational terms, 12 additional transformations), he showed no diffi- culty at all. ι In a like manner, H.J.'s performance on test sentences where the relative pronoun or complementizer had been optionally deleted, also proved to be of interest. According to the DTC, a sentence such as "The dog the girl saw was fat." (where the relative pronoun "which" has been optionally deleted) is more complex than a sentence such as "The ball which the boy kicked was blue", (where the relative pronoun is retained). The reason for this is that an additional transformation is involved in the former; in addition deletion of the relative pronoun is felt to eliminate certain surface structure cues which normally facilitate the comprehension of underlying deep structure relations. The DTC would thus predict that H.J. would show more difficulty in sen- tences where these surface structure cues (i.e. the relative pronouns or com- plementizers) had been deleted. Again this was found not to be the case — H.J.'s performance did not differ significantly, in either case. It seems obvious from the above examples, that we cannot explain certain test findings in terms of the DTC and that for adjectives and the optional deletion of relative pronouns and complementizers anyway, this theory should be viewed with suspicion. This has, interestingly enough, been the general conclusion of experiments involving the language performance of nor- mal subjects 5 . Thus the test data, in addition t o helping to confirm that lingu- istic methods are useful for the study of aphasia, have also provided some insight into an area of grammar and have supported the idea that aphasia can legitimately be used as a testing-ground for linguistic theory, (cf. Marshall and N e w c o m b e 1 6 . ) CONCLUSION AND IMPLICATIONS / On the basis of the data obtained in this study, it may be concluded that a linguistic approach to the testing of aphasia seems to be very useful for the characterization of a patient's language abilities. Furthermore it appears that such tests are sensitive to subtle and specific areas of language breakdown and thus seem superior to many clinical tests which often measure a general lan- guage ability rather than specific features. Test findings have also helped to provide evidence for certain issues arising out'· of linguistic theory. However, there are certain shortcomings to the present study which make conclusions highly tentative and which suggest the need for a great deal of Journal of e South African Speech and earing Association, Vol. 21, December 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) A Linguistic Approach to Minimal Language Dysfunction in Aphasia further research. One aphasic subject only was used, and a highly specific aspect of the grammar was investigated. Similarly, tests were limited to the oral and aural language abilities. A linguistic consideration of H.J.'s reading and writing may have helped t o broaden our understanding of his problem. Finally, and probably most important, the hypothesis, tests and results of this study were considered within the framework of a theory which is not, as yet, conclusive and which is constantly changing in the light of empirical facts. As such, all conclusions and implications are only as valid as is the theory. Nevertheless, the present study has served t o highlight an approach to the classification, testing and therapy of aphasic disorders which appears to be very fruitful. The writer wishes to suggest that the following figure might help t o provide a guideline to the clinician concerned with characterizing aphasic language in generative grammar terms; RECEPTION EXPRESSION Competence Performance Competence Performance PHONOLOGY SYNTAX SEMANTICS Fig. 2. Proposed model for the classification and testing of aphasia Such an outline has the following implications for diagnosis, classification and therapy: It suggests that language breakdown may occur on a number of levels, i.e. the three levels of the grammar - phonological, syntactic and semantic - and that each level must be considered in detail when studying aphasic language; an attempt should be made to assess the subject's reception and expression with respect to these levels and, furthermore, diagnosis should aim at discovering whether the subject's problem is one of competence or performance, using linguistic methods of testing. It has been shown that the experimental subject has a defect of a specific syntactic structure, evident in both reception and expression and related t o performance features rather than a competence breakdown. Thus he may be Tydskrif van die Suid-Afrikaanse Vrniging vir Spraak- en Ghoorhkunde, Vol. 21, Dsmbr 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) 18 Claire Penn fitted.into the appropriate categories of the above model. Such a character- ization has helped to provide therapy cues for the subject. Since reception is basic to expression, this aspect should be worked on initially; since the defect is one of performance and not competence, therapy should involve the teach- ing of the successful application of the rules, rather than the rules per se. Finally, since H.J. has difficulty with a particular syntactic structure, therapy should be concerned specifically with the complex sentence — teaching him how to relate accurately features in the embedded sentence tp features in the matrix sentence.* The above model has helped to provide a highly systematic characterization of the subject's language capabilities which is essential for an adequate therapy programme. In addition, it has helped to reduce the 'confusion in terminology' which is so apparent in the field of aphasia t o d a y . 1 5 Although, as mentioned before, there is a great need for further research both in the area of aphasic language and in the area of generative linguistics, the writer feels that cooperation between these disciplines is essential. Not only can generative linguistics provide an extremely exciting and valuable contri- bution to aphasic research, but it can also use the wealth of language data available from aphasia in helping to validate some of the basic theoretical con- structs. It is the writer's belief that such inter-disciplinary 'cross-fertilization' will do much to increase our knowledge of aphasic, and,-indeed, normal language, in the years ahead. REFERENCES 1. Blumstein, S.E. (1973): A Phonological investigation of Aphasic speech. Mouton: The Hague. 2. Chomsky, N. (1965): Aspects of the Theory of Syntax. M.l.T. Press, Cambridge, Massachusetts. 3. DeRenzi, E. and Vignolo, L.A. (1962): The Token Test: A sensitive test to detect receptive disturbances in Aphasics. Brain, 3, 665-678. 4. Fodor, J.Α., Bever, T.G. and Garrett, M. (1968): The Development of Psychological models for speech recognition. M.l.T. Press, Cambridge, Massachusetts. * It is interesting to n o t e that the findings in the present study confirm find- ings in the literature 5 regarding the value of the c o m p l e x sentence as a mea- sure of linguistic c o m p e t e n c e . T h e writer w o u l d suggest that even in cases w h o do not necessarily show obvious b r e a k d o w n in c o m p l e x sentences as H.J. does, the a d o p t i o n of such items in assessing aphasic language (and indeed, other areas of language p a t h o l o g y ) may prove e x t r e m e l y useful. Journal of e South African Speech and Hearing Association, Vol. 21, December 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) A Linguistic Approach to Minimal Language Dysfunction in Aphasia ι 5. Fodor, J.A. and Garrett, M. (1967): Some syntactic determinants of sentential complexity. Perception and Psychophysics, 2(7), 289-296. 6. Garrett, Μ. (1967): Psychological theories and linguistic constructs. M.l.T. Press, Cambridge, Massachusetts. 7. Goodglass, H. (1968): Studies on the grammar of aphasics. Chap. 5 in Developments in applied psycholinguistic research. Rosenberg, S and Koplin, J.H. (Eds.) MacMillan, London. 8. Griffiths, P.D. (1968): A critical Examination of Competence, Un- published Report, Dept., Linguistics, University of Edinburgh, Scotland. 9. Jacobs, R.A. and Rosenbaum, P.S. (1968): English Transformational Grammar. Blaisdell Publishing Co., London. 10. Lanham, L.W. (1973): How do we do linguistics now? A question for academic linguists in the mid-70's. Unpublished Report, Dept., Pho- netics and General Linguistics, University of the Witwatersrand, Johannesburg. 11. Lenneberg, E.H. (1967): Biological Foundations of Language Wiley, New York. 12. Luria, A.R. (1970): Traumatic Aphasia: Its Syndromes, Psychology and Treatment. Mouton Press, The Hague. 13. Luterman, L.B. and Bar, A. (1971): The diagnostic significance of sen- tence repetition for language impaired children. / Speech and Hearing Dis., 36(1), 29-39. 14. Lyons, J. (1971): Generative syntax. Chap 6 in New Horizons in Lingu- istics, Lyons, J. (Ed), Pelican, London. 15. MacMahon, M.K.C. (1972): Modern Linguistics and Aphasia. British J. Dis. Comm., 7(1), 54-63. 16. Marshall, J.C. and Newcombe, F. (1966): Syntactic and semantic errors in paralexia. Neuropsychology, 4, 169-176. 17. Menyuk, P. (1971): The Acquisition and Development of Language. Prentice Hall, New Jersey. 18. Osgood, C.E. and Miron, M.S. (1963): Approaches to the Study of Aphasia, U. of Illinois Press. 19. Schuell, H„ Jenkins, J.J. and Jimenez-Pabon, E. (1964): Aphasia in Adults. Harper and Row, N.Y. 20. Sklar, M. (1966): Sklar Aphasia Scale .Manual. W. Psychological services, California. 21. Taylor, O.L. and Anderson, C.B. (1968): Neuropsycholinguistics and language re-training. Unpublished Report, Indiana U. 22. Tikofsky, R.S. (1966): Language Problems in Adults. Chap. 11 in Speech Pathology. Rieber, R.W. and Brubaker, R.S. (Eds.). N. Holland Publishing Co., Amsterdam. 23. Traill, A. (1970): Transformational Grammar and the case of an Ndebele speaking aphasic. J. of the S.A. Logopedic Soc. 17, 48-66. 24. Weigl, E. and Bierwisch, M. (1970): Neuropsychology and linguistics: Topics of common research. Foundations of Language, 6, 1-18. 25. Wepman, J.M. and Jones, L.V. (1969): Five Aphasias: A commentary on aphasia as a regressive linguistic phenomenon. Chap 13 in Disorders of Communication. McKrioch, D. and Weinstein, E.A. (Eds.), Hafner. Tydskrif van die Suid-Afrikaanse Vrniging vir Spraak- en Ghoorhkund, Vol. 21, Dcmbr 1974 R ep ro du ce d by S ab in et G at ew ay u nd er li ce nc e gr an te d by th e P ub lis he r (d at ed 2 01 2) 20 Claire Penn 26 27 28 PAUDIIQ BOOKSHOP ( P T Y ) LTD. lAMPUo BOEKHANDEL (EDMS) BPK. BOOKS Y O U S H O U L D POSSESS BROWN, R. A FIRST LANGUAGE: THE EARL Y STAGES Price R13,30 Postage 35 cents. B A R - A D O N , Α., & LEOPOLD (EDS.) CHILD LANGUAGE: A BOOK OF READINGS Price R12,05 Postage 35 cents. T R A V I S , L.E. (ED.) HANDBOOK OF SPEECH PA THOLOG Y & < AUDIOLOGY