With respect to Type 4 questions, Wernicke’s aphasics produced some conjunction-initial clauses and some clauses involving the subjunctive (i.e. the mood directly indi-cating subordination).
Broca’s aphasics gave few answers beginning with a subordinating conjunction.
One of them did produce some answers involving the subjunctive. However, the ma-jority of structurally linked and grammatical answers produced by Broca’s aphasics, as well as the rest of the answers given by Wernicke’s aphasics, were statements that assumed the point of view of one of the characters seen in the picture, rather than be-ing purely descriptive. The subjects answered the question as if they were in the “men-tal state” of the characters. These answers are referred to as “situational statements”
with ‘theory of mind’ type reasoning. In them, the verb was inflected in the first, rather than the third, person singular (or second person singular, with reference to the partner in the situation shown in the picture); their meanings differed sharply from descriptive statements, as they directly represented the thought or statement of the characters they
“cited”; most of them did not involve a subordinating conjunction. These answers are supposed to involve syntactic structural recursion but they contain simple statements instead, with ‘theory of mind’ type reasoning. An example of a situational statement:
The picture: A girl is showing her scar to a boy.
Question: Vajon mire gondol a fiú? ‘What may the boy be thinking of?’
S.T.’s answer: Mindjárt rosszul leszek! ‘I’m going to be sick.’
Possible recursive construction: (Ő) arra gondol, hogy mindjárt rosszul lesz. ‘He thinks he is going to be sick.’
In the sense of Takano and Arita (2010), ‘theory of mind’ type reasoning is recursive.
The subjects, in addition to seeing themselves as able to infer other people’s mental states, considered other persons (e.g. ones seen in pictures) to be able to infer further (third) persons’ mental states, thus exhibiting recursive constructions.
23
Table 2 shows that the share of situational statements increased in Broca’s answers to Type 4 questions: Situational statement (43.3) 43.3 (74.0) 60.3 (31.0) 31.0 Sentence with subjunctive
Table 2. With type 4 questions percentage of all structurally linked answers (outside the brackets:
that of grammatical answers) in the grammatical categories, in aphasia types and in normal controls According to Table 2 the majority of the grammatical answers produced by Broca’s aphasic patients were situational statements containing ‘theory of mind’ type reason-ing. The low percentage of subordinating conjunctions in Boca’s aphasics’ answers shows that syntactic structural recursion is impaired. This is also suggested by the fact that 74.0% of Broca’s aphasics’ answers to Type 4 questions were simple situational statements without subordinating conjunctions. On the other hand, 43.3% of the grammatical answers given by Wernicke’s aphasics were also situational statements without subordinating conjunctions, but 30.0% of the answers by Wernicke’s aphasics were descriptive sentences beginning with a subordinating conjunction. Syntactic structural recursion was less impaired in Wernicke’s aphasia.
5. Discussion
Agrammatic (Broca’s) aphasics avoided giving answers based on syntactic-structural recursion; their access to syntactic recursion was found to be severely limited. On the other hand, their performance in recursive theory-of-mind inferences (expressed in simple situative sentences) remained intact. The dissociation of syntactic-structural recursion and theory-of-mind inferences can be observed in Wernicke’s aphasia to a lesser degree; this is in harmony with earlier observations that associate limited syn-tactic abilities primarily with Broca’s aphasia and consider grammatical errors com-mitted by Wernicke’s aphasics as consequences of the impairment of their lexical processes. Agrammatic Broca’s aphasics may use recursive theory-of-mind inferences (and situative sentences carrying them) in their responses as a repair/compensatory strategy in order to avoid syntactic-structural recursion.
This result was corroborated by a case study investigating linguistic aspects of the process of recovery of one of the Broca’s aphasic subjects. During his recovery, the
24
patient gradually came to apply the fundamental grammatical principle of agreement.
Recursive syntactic structures began to occur in the subject’s responses and, at the same time, the ratio of situative sentences expressing theory-of-mind inferences dropped; the latter type of responses completely disappeared by the end of the recov-ery process.
The use of simple situative statements could also be observed in the case of control subjects, but only in 31% of their responses. All other replies they gave involved clausal embedding signaled by a subordinating conjunction. Therefore, structural and theory of mind recursion represent two alternative strategies from which members of the control group were able to choose at will, whereas the aphasics were forced to choose the use of situative statements.
Why were ‘situational statements’ overused in the responses for Type 4 questions?
For two reasons: on the one hand, the content of Type 4 questions requires conclusions to be drawn from the pictures; on the other hand, subjects tried to avoid syntactic-structural recursion and produced simple sentences instead, expressing theory-of-mind reasoning. Subjects were able to project themselves into the state of mind of the char-acter in the picture. This process has to be able to use linguistic devices to integrate and control perspective shifts (MacWhinney 2009). The content of situational state-ments showed that Broca’s aphasic subjects correctly identified themselves with the mental states of the characters in the pictures. In this way complex syntactic structural recursion was avoided.
The share of situational statements increased in Broca’s aphasics’ answers to Type 4 questions. How did aphasic subjects “know” when to substitute a purely de-scriptive perspective for a non-dede-scriptive perspective? On the one hand, a subset of linguistic devices indicating a non-descriptive perspective was available for them to control perspective shift, cf. theory-of -mind statements contain the first person singu-lar feature (instead of the third person), their syntactic structure was very simple, sometimes fragmented correctly, their semantic content referred to simple feelings, emotions. On the other hand, syntactic structural recursion requires complex introduc-tory formulas, subordinating conjunctions, agreement relations between main and embedded clauses, and two propositions to control a descriptive perspective. This complex linguistic subsystem was partially available or was not available at all for aphasics. Syntactic structural recursion was substituted for theory of mind recursion on the basis that the linguistic system and the social cognition system interact with one common recursion module. In agrammatic aphasia syntactic representations are dis-connected from the recursion module but theory-of-mind type reasoning can access the recursion module.
6. AD subjects
The validity of the above observation can be supported by cases exhibiting the con-verse dissociation, if we can find such. The same tests were conducted with the partic-ipation of Hungarian speaking subjects with Alzheimer’s disease. The subjects were classified on the basis of Mini-mental state examination (Folstein, M. F. – Folstein, S.
E. – McHugh 1975, Tariska et al. 1990), ADAS-Cog (Rosen et al. 1984) and DSM-IV
25
(American Psychiatric Association 2000). In persons with Hungarian speaking Alz-heimer’s disease (AD), as opposed to the case of aphasics, the language faculty be-comes limited gradually as the disease progresses. We tried to find out to what extent structural embedding linguistic operations vs. recursive theory-of-mind inferences were involved in their case. Therefore, we administered the above pictures and ques-tions (‘What might X in the picture be thinking of?’; ‘What might Y in the picture be asking Z to do?’) to four persons with mild and two with moderate AD.
In the responses of subjects with mildAD, there was no significant difference be-tween them and normal controls in the proportion of replies involving syntactic-structural recursion vs. situative responses. We can infer that in mild AD, both struc-tural and theory-of-mind recursion are unaffected.
In the case of moderate AD, we found a significant difference in the proportion of responses involving syntactic-structural recursion and simple situative responses re-quiring recursive theory of mind inference: the ratio of situative sentences was signifi-cantly lower than in normal control subjects. On the other hand, in the responses of persons with moderate AD, the share of sentences involving syntactic-structural recur-sion (embedding involving hogy ‘that’ clauses) was not lower than in normal control responses. Additionally, we also received semantically irrelevant descriptive state-ments referring to some aspect of the picture. That is, while syntactic-structural recur-sion may remain unaffected in moderate AD, theory-of-mind inferences seem to be impaired to some extent.
Subjects Relevant responses Irrelevant responses
Mild AD 86.2 13.8
Moderate AD 62.1 37.9
Table 3. AD subjects responses for question Type 4
Mild AD (160) Moderate AD (82) Normal (982)
Simple, descriptive clause 13.0 3.8 20.5 13.5 –
Situative statement 22.0 1.9 2.4 2.4 31.0
Simple clause with subjunctive 3.0 17.0 1.3 –
Conjunction + situative statement 7.0 2.4. 1.2. 24.0
Conjunction + descriptive clause 55.0 8.1 (58.5) 54.8 19.5 45.0 Table 4. Irrelevant content in responses from the point of view of the question and/or the picture In order to support this with additional evidence, we administered a primary (6-sentence-long) and a secondary (8-(6-sentence-long) false belief test to our two moderate
26
AD subjects (following Youmans–Bourgeois 2010). In the primary false belief test, the subjects gave correct answers to all questions. However, the secondary false belief test proved to be more difficult for the moderate AD subjects. Although the situation was facilitated with respect to memory (they could use a test sheet), both persons gave the wrong answers. Results of the secondary false belief test support the limitation of theory-of-mind inference abilities for both moderate AD patients.
In sum: in the mild and moderate stages of Alzheimer’s disease, recursive clausal embedding abilities remain unaffected, but recursive theory-of-mind inferences be-come limited by the medium stage of the disease. Moderate AD subjects tend to avoid utterances in first person singular that assume the state of mind of another person, i.e.
the use of situative sentences. Clausal embeddings (descriptive sentences involving the conjunction hogy ‘that’) suggest, on the other hand, that they are still able to attribute intention indirectly, in a third person singular format. What is missing is the projection of themselves into the state of mind of another person. This is also supported by the results of the secondary false belief test.
Unlimited syntactic-structural recursion and limited theory-of-mind inferences in Alzheimer’s disease as opposed to limited syntactic-structural recursion and unim-paired theory-of-mind inferences in Broca’s aphasia: this is a pattern of double disso-ciation. This finding supports theories (e.g. Siegal–Varley 2006, Zimmerer–Varley 2010) that argue for the mutual independence of these two types of recursion in adults.
7. Arithmetic
With respect to recursion in arithmetical calculations, we found another case of double dissociation. Following Varley et al. (2005), we gave seven different tasks to agram-matic aphasics and persons with moderate AD. With the latter, we found limitation in the recursion of arithmetical operations.
This limitation did not concern the four fundamental operations with one-digit numbers but showed a steep decline in those with two-digit ones. The manipulation of three-digit numbers was unsuccessful in all cases. Our AD subjects did have an idea of numbers, and they were able to transpose numbers from verbal to visual representa-tion, but the idea of the infinity of numbers and any operations based on it were not accessible; neither were the rules of operations involving fractions. Recursive calcula-tions (e.g. arithmetical tasks involving parentheses) were understood without impair-ment, they were able to make differences in the hierarchical order of operations (addi-tion vs. multiplica(addi-tion), but they were unable to produce a recursive structure without help (e.g. inserting parentheses into calculation tasks and figuring out the result).
Subjects
27
Broca’s aphasic 1. 0 0 0 11 0 25 100
Broca’s aphasic 2. 0 0 0 0 0 0 16.7
Table 5. Error rates (in percentages)
The Hungarian speaking Broca’s aphasics did not exhibit any limitation in arithmetical operations, they calculated correctly, they were able to produce potentially infinite sequences of numbers, applied recursive arithmetical operations correctly, inserted parentheses in various combinations, even double ones. They did all that using numer-ical symbols and operation signs; that is to say, they were not necessarily able to ver-balise their otherwise correct operations.
8. Summary
In the case of Hungarian speaking patients with moderate AD, syntactic recursion involving embedding is relatively unimpaired, as opposed to their limited ability to tackle theory-of-mind and arithmetical recursion. Conversely, we found limited syn-tactic recursion but normal theory-of-mind inferences and recursive arithmetical oper-ations in Hungarian speaking agrammatic aphasics.
9. Conclusion
The production differences observed in the tests are explained by the fact that we do not seem to have to do with a single recursive operation whose application may be impaired or remain intact at various levels; rather, we encounter separate recursive operations bound to individual linguistic and non-linguistic subsystems that may be selectively impaired. We have seen that these operations are not independent of one another: the impairment of one may trigger the use of another one as a substitution mechanism or repair strategy. With respect to the relations recursive sentence structure – recursive theory-of-mind inferences, we have found that in cases of a deficit of oper-ations of the left-hand constructions, recursive operoper-ations of the right-hand construc-tions can be used as parts of a repair strategy.
Recursive operations manifested in theory-of-mind inferences may also be dissoci-ated from syntactic recursion. The use of erroneous or impaired theory-of-mind abili-ties is highly probable to determine person, number, and tense features of a clause, but does not necessarily prescribe that the clause has to be a recursively embedded one.
The impaired theory-of-mind inferences were not repaired by syntactic structural re-cursion in the responses by our subjects.
The accessibility of recursive operations is limited in Alzheimer’s disease for the-ory-of-mind and calculation but unlimited with respect to linguistic representations;
whereas in agrammatic aphasia, linguistic representations may be disconnected from the recursion module, theory-of-mind and calculation systems are able to access it (cf. Zimmerer–Varley 2010). These dissociations argue for a theoretical model that
28
posits a module of recursive operations in the human mind that is shared by linguistic, theory-of-mind, and arithmetical performance. This common recursion module is ac-cessible to a limited extent for the theory-of-mind and arithmetical subsystems while it is fully accessible for representations of linguistic constructions in Alzheimer’s dis-ease, whereas in agrammatic aphasia, the representations of linguistic constructions may be detached from the recursion module while theory-of-mind and arithmetical systems may access it at will.
References
American Psychiatric Association 2000. Diagnostic and statistical manual of mental disorders (4th ed., text. rev.) Washington, DC: Author.
Bánréti, Z. 2010. Recursion in aphasia. Clinical Linguistics & Phonetics. 2010, Vol. 24, No. 11:
906–914.
Evans, N. – Levinson, S. 2009. The Myth of Language Universals. Behavioral and Brain Sciences.
32: 429–448.
Everett, D. L. 2009. Piraha culture and grammar: A response to some criticisms. Language. 85:
405–442.
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. 12/3: 189–
198.
Hauser, M. D. – Chomsky, N. – Fitch, T. W. 2002. The Faculty of Language: What Is It, Who Has It, and How Does It Evolve? Science. 298: 1569–1579.
Kertesz, A. 1982. The Western Aphasia Battery. New York: Grune & Stratton.
MacWhinney, B. 2009. The emergence of linguistic complexity. In Givón, T. – Shibatani M. (eds.):
Syntactic Complexity. Amsterdam: John Benjamins. 405–432.
Rosen, W. G. – Mohs, R. C. – Davis, K. L. 1984. A new rating scale for Alzheimer’s disease. Ameri-can Journal of Psychiatry. 1984/Nov.: 356–364.
Siegal, M. – Varley, R. – Want, S. C. 2006. Mind Over Grammar. Reasoning in Aphasia and Devel-opmental Contexts. In Antonietti, A. – Liverta-Sempio, O. – Marchettio, A. (eds.): Theory of mind and language in developmental contexts. Berlin: Springer. 107–119.
Stark, J. 1998. Everyday Life Activities Photo Series. Vienna: Verlag Peter Poech. (Photo Cards:
Druckerei Jentzsch)
Takano, M. – Arita, T. 2010. Asymmetry between Even and Odd Levels of Recursion in a Theory of Mind. In Rocha, L. M. – Yaeger, L. S. – Bedau, M. A. – Floreano, D. – Goldstone, R. L. – Vespignani, A. (eds.): Proceedings of ALife X. www.citeulike.org/user/jasonn/article/7293338 Tariska P. – Kiss É. – Mészáros Á. – Knolmayer J. 1990. A módosított Mini Mental State vizsgálat.
(The modified Mini Mental State examination) Ideggyógyászati Szemle. 43: 443.
Varley, R. A. – Klessinger, N. J. C. – Romanowski, Ch. A. J. – Siegal, M. 2005 Agrammatic but numerate. Psychology PNAS Early Edition. www.pnas.org/cgi/doi/10.1073/pnas
Youmans, G. – Bourgeoi, M. 2010. Theory of mind in individuals with Alzheimer-type dementia.
Aphasiology. 24/4: 515–534.
Zimmerer, V. – Varley, R. 2010. Recursion in severe agrammatism. In van der Hulst, H. (ed.): Re-cursion and Human Language. 393–405. Berlin: De Gruyter Mouton.
29