Verb processing in Basque and French agrammatic aphasia: A “post-lexical access” deficit

Abstract

Background: Verb processing is largely described across languages as being impaired in agrammatic aphasia, i.e., non-fluent aphasia with so-called “telegraphic style”. Although both lexical and morphosyntactic errors have been reported in the literature, this paper questions the claim that impaired access to verbs is a hallmark of this clinical syndrome.

Aims: The present study aims to assess how agrammatic verb processing is impaired in two languages with distinct grammatical properties: Basque and French, and to test hypotheses that suggested an access deficit to verbs on this new database. Moreover, the nature of agrammatic verb errors is analysed from an interdisciplinary neuropsycholinguistic perspective according to which aphasic symptoms should be interpretable at different levels of organisation of language processing involving neural, cognitive and linguistic aspects.

Methods & Procedures: A protocol built on Basque and French grammatical properties was designed in order to assess whether errors are specific to verbs, whether they are lexical or morphosyntactic and whether the verb argument structure complexity increases lexical and morphosyntactic processing difficulties, in both production and comprehension. One Basque-speaking and one French-speaking patient with agrammatism and matched controls were assessed on various oral tasks (object and action naming; sentence production and comprehension; prepositional phrase production), each of them containing 20 pictures displayed on a computer. Data were collected using a digital recorder.

Outcomes & Results: Results show that agrammatic speakers produced many lexical verbs of every argument structure type. Errors were specifically morphosyntactic and increased with the verb argument structure complexity. However, verb errors were different in Basque and French due to their distinct morphosyntactic properties. In addition, whereas errors appeared in French in the use of prepositions, case morphology was well preserved in Basque, raising the issue of considering distinct neurocognitive mechanisms underlying different morphological systems.

Conclusions: This paper supports the view that Basque and French agrammatic data collected from this study do not result from a lexical-access deficit. However, this interpretation depends on how one considers inflected verbs to be processed (endolexicon or exolexicon), as addressed in the discussion. In conclusion, a “post-lexical access” deficit is rather suggested, that is prior to morphophonological encoding, and affects abstract morphosyntactic operations required to implement the verb argument structure.

Keywords:

• Agrammatism
• Basque
• French
• Inflection
• Verbs
• Argument structure
• Cross-linguistic case study
• Cross-disciplinary
• Aphasiology
• Linguistics
• Neurocognitive sciences

Many thanks to Scott, Phaedra, Kristel, Shab and the two anonymous reviewers for their valuable help in the revision of the manuscript.

Notes

2. ¹The first description of a Basque-speaking aphasic case is found in Pitres (1895), cited by Erriondo Korostola (1995). The latter presented a PhD thesis in Psychology under Michel Paradis’s supervision at the University of the Basque Country, in 1993 that was published in 1995. Erriondo assessed 59 Basque-Spanish bilingual aphasic patients’ language performance with the Bilingual Aphasia Test (Paradis, 1987), which had been adapted to Basque/Spanish, Basque/French and Basque/English. Three papers were published based on the corpus Erriondo collected between 1986 and 1989: “Aphasia manifestations in Basque” Erriondo Korostola and Laka (2001); “Agramatismoaren sintomak euskaraz” (Basque agrammatic manifestations), Laka (2003). However, the latter is not a study of agrammatic aphasia but a description and linguistic analysis of grammatical errors found in Basque-speaking aphasics’ speech; “Itzuleraren hipotesia: Jabekuntza eta afasiaren erkaketa” (The regression hypothesis: a comparison between language acquisition and aphasia), Ezeizabarrena and Laka (2006). Munarriz and Ezeizabarrena (2009) wrote a chapter about phonological paraphasias observed in one Basque-Spanish bilingual adult. Finally and more recently, Adrover-Roig et al. (2011) presented a paper about language attrition after left basal ganglia damage in one Basque-Spanish person with aphasia. Note that except Pitre’s description of a Basque/French/Béarnese dialect speaking aphasic patient, all other studies concern Basque/Spanish bilingual aphasic patients.

3. ²Western, Central, Navarrese, Navarro-Labourdin, Eastern Navarrese, Souletin.

4. ³This task can be adapted to French with the use of clitics because third singular and third plural pronouns are homophonous (e.g., il / ils [il]), when followed by a consonant initial verb. Therefore, they do not impede the target of the task, which consists of assessing verb inflection comprehension: il boit [ilbwa]/ils boivent [ilbwav], ‘he drinks/they drink’. Vowel initial verbs create liaison contexts, where the plural sibilant is realised, e.g., atterir “to land” il atterit [ilateri]/ils atterissent [ilzateris] “he lands/they land”. Hence, it impedes the target of the task, which is verb inflection comprehension, as speakers can decode singular versus plural agreement difference through pronouns: il/ilz. Finally, French verbs of the first conjugation group (verbs ending in -er in the infinitive; e.g., parler “to speak”) are not useful for this type of task because third singular and third plural present tense forms are orally homophonous (il parle [ilparlə]/ils parlent [ilparlə], “he speaks/they speak”). Thus, the singular/plural subject agreement difference is not orally perceptible.

5. ⁴The Basque-speaking agrammatic participant was assessed in Basque only because it was his mother tongue and he felt more comfortable speaking Basque than he did French.

6. ⁵The reason why the number of items corresponding to each verb argument structure type is different in Basque and French is that the verb “to kiss” in Basque requires a ditransitive structure (muxu eman “to give a kiss to somebody”)—used twice in the task—whereas, in French, it requires a transitive structure (embrasser “to kiss somebody”). And the verb “to brush one’s hair” is intransitive in Basque (orraztatu) but in French, it is pronominal (se coiffer) and classified as a transitive verb.

7. ⁶The reason why the production of verbs was tested in a sentence picture description task rather than in a sentence completion task, which would be much more constrained is that sentence completion task cannot be realised in the same conditions in Basque as in French. In spoken form, Basque being a SOV language, it would be possible to have the participants produce only the verb (e.g., complete the sentence: [subject] [object] _____). However, in French, as the verb is not final, the participants would be forced to produce verb and object(s) (complete the sentence: [subject] _____ [object]).

8. ⁷Except in the case of question-stimuli that forced the production of noun + absolutive case in Basque or an accusative noun in French, where no grammatical case or preposition is overtly realised; e.g., Zer erosi du Peiok? Qu’a acheté Pierre? What did Peter buy? egunkariaØ / Ø le journal ‘the newspaper’.

9. ⁸The omission of the auxiliary in Basque would be similar to producing—ING verbs in English (e.g. *“The man singing”, instead of “The man is singing.”). This corresponds to a shortened form also found in the Grammar of titles (e.g., Leiçarrague, 1571) as remarked by B. Oyharçabal (personal communication, 2011). However, as in this task, participants were taught to produce inflected verbs (aspectual verb + auxiliary), auxiliary omissions were scored as incorrect.

10. ⁹Au is a contracted form of the preposition à “to” + definite masculine determiner le “the”

11. ¹⁰Regarding three out of the five comprehension errors produced by Agr-BSQ, it is difficult to know whether it is not a matter of dialect (stimuli 3, 15, 20). However, Ctr-BSQ did not produce any error and both have the same dialect. Moreover, since the participants were asked to choose between two pictures, the factor chance can have an impact on the results. Therefore, more data and better controlled data are needed in order to draw further conclusions about agrammatic speakers’ performance in decoding inflected verb structures.

12. ¹¹Goodglass and Geschwind (1976).

13. ¹²Still recently, Boeckx (2012, Chapter 1) writes: “Take the lexicon: is this to be thought of as a pre-syntactic resource component? A post-syntactic morphological component? An encyclopedia? A repository of idiosyncracies? And, if the latter, of which kind? At Word/morpheme level, or at the Construction/idiom/phrase level?”

14. ¹³Different morphological theories exist, e.g., lexical morphology assuming that all morphology take place in the lexicon (e.g., Di Sciullo & Williams, 1987; Lieber, 1981; Selkirk, 1982) versus syntactic morphology assuming that (at least inflectional) morphology takes place in syntax proper (Baker, 1985, 1988) or mixed models such as split morphology (Anderson, 1982; Perlmutter, 1988).