Verb Development Problems in Dutch SLI Rob Zwitserlood (student number: 3030113)

Dutch SLI, boy, 8 years old, TAK Storytelling Task 2 (Van Weerdenburg, 2006): ‘Vader wil, wil kinderen wil karretje maken en dat vader ook meehelpen, met da karretje maken, eh… wagen, dan duwen, heel hard racen, duwen dan duwen….meisje duwen. Dan, eh…duwt en dan snel op… en eh, stukker, mmm, boom enne stukken en jongen koprol en meisje handstan. Daarna stukke kar. Alleen, alleen dragen. Vader weer overnieuw dragen’.

1

1. Introduction Research on SLI focuses to a large extent on English SLI children. The question arises whether the accounts of English SLI also hold for Dutch SLI. In recent years, research on Dutch SLI has received more attention. In this paper the prominent research on verb development in Dutch SLI will be discussed and compared with English SLI. The following three research questions will be answered: 1. What are the problems in verb morphology for Dutch SLI children? 2. What are the problems in verb morphology for English SLI children? 3. To what extent can verb development problems in SLI be seen as universal or language specific? In the second paragraph a definition and general characteristics of SLI are discussed. In the third paragraph the verb morphology of Dutch and English are compared. In paragraph four the prominent research on Dutch SLI is reported. In paragraph five a comparison is made between Dutch and English verb morphology problems. In paragraph six conclusions and suggestions for further research are presented.

2. Specific Language Impairment The definition for SLI is an exclusion definition. SLI is characterized by difficulty with language that is not caused by known neurological, sensory, intellectual or emotional deficit (Leonard, 1998). Often a degree of severity is specified: more than half a year delay in reception and more than one year delay in production (De Jong, 1999). Researchers agree that grammatical morphology is a particular locus of difficulty for children with SLI. These morpho-syntactical problems have drawn much attention in research over the past years. However, language problems of SLI children are not solely restricted to morpho-syntaxis. Although this paper will focus on morpho-syntactic problems in SLI children, and especially on verb morphology, it is important to realize that SLI children suffer from language problems in all language domains. According to De Jong (1999), the symptoms of SLI can take the form of (1) difficulties in the comprehension and production of grammatical morphemes, (2) phonological restrictions, (3) problems in lexical storage and access and (4) pragmatic problems.

3. Typology of Dutch and English Verb Morphology Dutch is generally considered to be a SOV (Verb-second) language, whereas English is a SVO language. The basic word order in Dutch is found in dependent clauses. The infinitival verb is basegenerated in final position. In matrix clauses and after topicalized elements, the finite verb is in second position. Verb second involves a movement of the verb from its base-generated position to a functional 2

head dominating V (C or I). In Dutch, regular verb infinitives have the –en ending, whereas in English infinitives take the bare stem form. Therefore it is easier to identify infinitives in Dutch sentences than in English sentences. The Dutch inflectional paradigm is somewhat richer than the English verb paradigm: Singular

Plural

Present Tense

Dutch

English

Dutch

English

first person

-Ø (= zero affix)

-Ø

-en*

-Ø

second person

-t (-Ø in VS order)

-Ø

-en*

-Ø

third person

-t

-s

-en*

-Ø

Regular Past Tense

Singular

all persons

-de / -te

Plural -ed

-den* / -ten*

-ed

*-en is reduced to schwa: ə Table 1:. Inflectional paradigm for regular present and regular past tense in Dutch and English.

4. Morpho-syntactic development in Dutch SLI In this paragraph three mayor studies concerning Dutch SLI are presented. Especially the results concerning verb morphology and verb argument structure will be highlighted.

4.1 The GRAMAT Study by Bol and Kuiken (1988) The first extensive study on Dutch SLI has been performed by Bol and Kuiken (1988). They developed the GRAMAT model (Grammaticale Analyse van Taalontwikkelingsstoornissen) for a grammatical analysis of language disorders. Their model was based on the LARSP (Language Assessment Remediation and Screening Procedure) for English by Crystal (1979). The authors studied children with a normal language development (ND) in six age groups ranging from 1;0 up to 4;00 years as well as older children with SLI, children with hearing problems and children with Down syndrome. From all children spontaneous speech samples with 100 utterances were obtained and analyzed. The analysis was performed at clause level, phrase level and wordmorphology level. The SLI children varied in age from 4;1 to 8;2 years, with a mean age of 5;11 years. The results for the SLI group are presented here for the three levels of analysis mentioned above. In general, the Dutch SLI children used more Minors when compared to ND and matched for MLU.

3

(1) Minors: utterances like: ja, nee, dag, kijk hier, weet niet, yes, no, bye, look here, dunno Fewer self-repetitions were counted in the SLI samples and more utterances were unintelligible, incomprehensible or considered to be ambiguous. The mean sentence length of the SLI children was lower than in the ND group. At clause level, some structures were used more often by the SLI children (e.g. Adverb-Verb, ObjectVerb, Subject-Complement) whereas other structures were less frequent (e.g. Subject-Adverb) than in ND. Inversion questions (e.g. Verb-Subject-X) were also less used. Longer and more complex utterances with four or more clause elements (e.g. Subject-Verb-Adverb-Object) were far less frequent in the SLI samples. The same holds for coordinate clauses with complementizers en (and), maar (but) and want (because). At phrase level the results indicated that Dutch SLI children used less articles, pronouns and possessives in their noun phrases. Longer phrase combinations of three or four elements (e.g. PrepDet-Adj-N) were seldom used. A prominent characteristic of Dutch SLI was the omission of Dutch er in obligatory contexts: (2) Ik doe er water in. I put

water in (it).

SLI: *Ik doe water in. I put water in.

No significant differences in the use of modals and copulas were observed. At the word-morphological level the use of the diminutive suffix for nouns and the 1Sg Present tense form for verbs was significantly below that of the ND group. Surprisingly, no problems were reported with the 3Sg Present tense –t morpheme. English SLI children are known to make many errors in the 3Sg Present morpheme in the form of omissions (Rice et al, 1995).

4.2 The Dutch SLI Study by De Jong (1999) In 1999 De Jong performed a study on Dutch SLI children.He studied the inflectional morphology and argument structure in 35 Dutch SLI children with an average age of 7;8 years old. De Jong designed two experiments: one where past tense forms were elicited in narratives, and one with a verb argument alternation task. De Jong also critically reviews Bol and Kuiken (1988) and states that the GRAMAT analysis does not take errors (omissions and substitutions) into account, since only correct verb forms were counted. In his view, omissions and substitutions are a characteristic of SLI children. They do learn to master the verb paradigms of the target language, but fail to consistently mark the verbs correctly for tense. 4

In his data, the SLI children produced longer samples with less past tense forms and showed more omissions of past tense inflection. Quite a few SLI children omitted the verb entirely in some utterances. He also found that the SLI children frequently substitute paste tense forms for present tense forms and often tend to use periphrastic verbs with infinitives in past tense contexts: (3) Hij ging het boek lezen he went the book read This high occurrence of the construction gaan (go) + infinitive was observed in present tense as well. De Jong regards this as a ‘strategy’ to avoid inflection of the main verb. This is a phenomenon that can also be observed in younger Dutch ND children. Van Kampen (1997) suggests that in young children the frequent use of this construction is an effect of economy of derivation. Children use these constructions as a temporary ‘least effort’ solution to avoid Verb-second movement in matrix clauses. The construction disappears when the verb-second rule is fully mastered, around the age of 3;6 to 4;0 years. This view is also supported by an experiment with young ND children and agrammatic aphasics on verb movement by Zuckerman et al (2001). In examining his data on subject-verb agreement De Jong (2204) recognizes three types of agreement errors: a) Omission of an inflectional marker (3Sg –t, Plural –en). (4) *dan ga mama (correct form is gaat) then go mummie b) Misrepresentation of the number feature of the verb (mostly a singular verb form with a plural subject). (5) *dat doet altijd mijn vader en moeder (correct form is doen) that does always my father and mother c) Uninflected (infinitive) form in final position (sometimes with omission of the auxiliary in second position). (6) *en dan Ø papa mama wakker maken (correct form is Ø = gaat) and then daddy mummie wake (up)

5

The first two errors are also made by younger ND children and can thus be seen as immature. The third error does not occur in younger Dutch children. Compared to the younger ND group, the number of errors in the SLI group was disproportionally high. Contrary to Bol and Kuiken (1988), he not only found errors in 1Sg, but also in 3Sg and 1, 2, 3Pl. De Jong also researched the verb argument structure in his SLI group. He states that the SLI children show a propensity to select intransitive frames. They also tend to substitute transitive verbs for intransitive verbs, as well as di-transitive verbs for mono-transitive verbs. SLI children possess a restricted set of lexical verbs. They tend to rely on a subset of General All Purpose (GAP) verbs. These verbs have a nonspecific meaning (e.g. English: do, go; Dutch: doen, gaan,). The omission of verb arguments is reported by Rice and Bode (1993) for English SLI.

4.3 The Developmental Dyslexia and SLI study by Wilsenach (2006) Wilsenach studied Dutch ND children, children at risk for dyslexia and SLI children. One of the goals of her study was to see whether developmental dyslexia and SLI can be seen as different disorders or not. She was especially interested in the use of the auxiliaries hebben (have) en zijn (be) and past participles in the SLI group. Previous studies on Dutch SLI suggest that they do not experience problems in this category (Bol and Kuiken, 1988 and Bol and De Jong, 1992). This is discrepant with the findings in German SLI studies, another Verb-second language (Clahsen, 1989). Clahsen found high deletion rates and no systematic progress in German SLI children as they got older. Leonard (1998) reports serious problems for English SLI children with auxiliaries. The results on a morpho-syntactic dependency discrimination task that involved auxiliaries, show that SLI children perform much worse than both other groups. The children had to discriminate between grammatical and ungrammatical versions of two different morpho-syntactic relations, expressed in sentences with a tense modal verb and a past participle and in sentences with a modal and an infinitive. Wilsenach suggests that the SLI children had problems in grasping the linguistic demands of the tasks. Another experiment tested the verb argument structure by means of a sentence imitation task. Three types of sentences were used: intransitive constructions with one argument, ditransitive constructions with two arguments and intransitive constructions with a (not obligatory) adjunct. Even the shortest sentences were poorly imitated by the SLI group, which can be seen as a floor-effect. The results were compared to the results on a short term memory (STM) task. The STM of the SLI children was significantly below that of the ND group. Since her research did not include spontaneous speech analysis or elicitation tasks, conclusions on production of auxiliaries and verb arguments by SLI children can not be drawn.

6

5. A comparison between the verb morphology problems in Dutch and English SLI The first two research questions from the introduction are repeated here:

1. What are the problems in verb morphology for Dutch SLI? Dutch SLI children have problems with 1Sg (Bol and Kuiken, 1988). De Jong (1999) recognizes three types of agreement errors: omission of an inflectional marker (3Sg –t, Plural –en), misrepresentation of the number feature of the verb (mostly a singular verb form with a plural subject) and use of an uninflected (infinitive) form in final position (sometimes with omission of the auxiliary in second position). Past tense is often substituted by present tense, and also overproduction of gaan (go) + infinitive is observed. Wilsenach (2006) showed that Dutch SLI children perform poorly at grammaticality judgment tasks concerning auxiliaries.

2. What are the problems in verb morphology for English SLI children? Leonard (1998) provides an overview of verb morphology problems in English SLI. The children mainly experience problems with 3Sg –s, Past tense –ed, the infinitive participle ‘to’ and with the copula ‘be’ and auxiliary ‘be’ and ‘do’ forms.

6. Universal features of SLI across languages The third research question from the introduction is repeated here:

3. To what extent can verb development problems in SLI be seen as universal or language specific? It is generally accepted that SLI children encounter serious problems in the verb morphology of their target language. Thus, this could be seen as an universal characteristic. The different symptoms of problems with verb morphology in Dutch and English SLI children seem to be related to their different target languages. For instance, English SLI children experience serious problems with auxiliaries (Leonard, 1998) whilst Dutch SLI children do not (Bol and De Jong, 1992). The explanation might be that the English auxiliary system is more complicated than the Dutch system and also less salient because of the contractions common in English spontaneous speech (e.g. he’s, we’ll). For Tense and Agreement, both English and Dutch SLI children perform poorly on Past tense. Agreement errors are made in both languages. According to de Jong (1999), Dutch children also have problems with number (singular / plural). Problems in person and number are hard to identify in English, because all the regular verb forms take a zero affix except for 3Sg –s. 7

When we look at the amount of infinitives used by Dutch and English SLI children, Wexler et al (2004) conclude that Dutch children use less infinitives than English children. They explain this discrepancy as originating from the differences between the morphological systems of the two languages. This can also be seen as an example of SLI taking different forms in different languages. For Dutch SLI children, verb movement is considered to be problematic. Verbs remain in final position (by means of aux-insertion) or are sometimes omitted. Verb movement problems are not reported in the literature on English SLI. These problems can be regarded as characteristic of Verbsecond languages. SLI children have a limited set of lexical verbs at their disposal. Problems for SLI children in verb argument structure have been reported for both languages. Because of the high incidence of GAP verbs, SLI children fail to learn the argument structures of many lexical verbs. Since verb argument can be seen as a universal characteristic of verbs, these problems could be regarded as universal.

7. Conclusions and suggestions for further research The symptoms of SLI certainly have some universal characteristics: phonological problems, lexicalsemantic problems, morpho-syntactical problems and pragmatic problems are broadly recognized by investigators in SLI research. The morpho-syntactical problems seem to vary according to the target language the SLI child uses. To quote Leonard (1998: p.117): ”Although much needs to be learned about SLI across languages, one conclusion seems safe: if there is an universal feature of SLI, apart from generally slow and poor language learning, it is well hidden. In any given language, children with SLI might show areas of extraordinary weakness. But these areas will differ from language type to language type. If we are to understand SLI, we will need to take this fact into account”. All previous research on Dutch SLI has been based on cross-sectional data. A longitudinal approach could provide more insight in verb development of Dutch SLI children over time. Interesting research questions could be: 1. Do Dutch SLI children leave the extended optional infinitive stage at a certain age? 2. Does the marking of gaan in Aux+infinitive precede the marking of lexical verbs in SLI? 3. Does an increase in the number of verbs used lead to more inflection and argument structure errors in SLI? 4. What can be observed about the use of the auxiliaries hebben en zijn? Can the findings of Bol and De Jong (1992) be confirmed? 8

References: Bol, G. & Kuiken, F. (1988). Grammaticale Analyse van Taalontwikkelingsstoornissen, PhD Dissertation, University of Amsterdam Bol, G.W. & De Jong, J. (1992). Auxiliary Verbs in Dutch SLI, Scandinavian Journal of Logopedics and Phoniatrics 17, 17-21 Clahsen, H. (1989). The grammatical characterisation of developmental dysphasia, Linguistics 27, 897-920. Crystal, D. (1979). Working with LARSP, Edward Arnold, London. Jong, J. de (1999). Specific Language Impairment in Dutch: Inflectional Morphology and Argument Structure, Groningen, The Netherlands, Groningen Dissertations in Linguistics 28. Jong, J. de (2004). Grammatical Impairment: An Overview and a Sketch of Dutch. In Verhoeven, L. & Van Balkom, H. (Eds.), Classification of Developmental Language Disorders, 261-281, New Jersey: Lawrence Erlbaum Associates. Leonard, L.B. (1998). Children with specific language impairment. Cambridge, MA: MIT Press. MacWhinney, B. (2000). The CHILDES project: Tools for analyzing talk. Third Edition. Mahwah, NJ: Lawrence Erlbaum Associates. Rice, M.L. & J.V. Bode (1993). GAPs in the lexicon of children with specific language impairment, First Language, 13, 113-132. Rice, L., K. Wexler, P.L. Cleave (1995). Specific Language Impairment as a Period of Extended Optional Infinitive, Journal of Speech and Hearing Research, Vol. 38, 850-863. Wexler, K., Schaeffer, J. , Bol, G. (2004). Verbal Syntax and Morphology in Typically Developing Dutch Children and Children with SLI: How Developmental Data Can Play an Important Role in Morphological Theory, Syntax 7:2, 148-198, Blackwell Publishing, Malden, USA. Van Kampen, J. (1997). First steps in WH-movement, PhD Dissertation Utrecht University. Wilsenach, C. (2006). Syntactic Processing in Developmental Dyslexia and in Specific Language Impairment, PhD Dissertation, LOT Publications 128, Utrecht. Van Weerdenburg, M. (2006). Language and Literacy Development in Children with Specific Language Impairment, Dissertation, EAC, Research Centre on Atypical Communication, Radboud University, Nijmegen. Zuckerman, S., Bastiaanse, R. & van Zonneveld, R. (2001). Verb Movement in Acquisition and Aphasia: Same Problem, Different Solutions—Evidence from Dutch. Graduate School for Behavioral and Cognitive Neurosciences, University of Groningen, The Netherlands, Brain and Language 77, 449–458.

9