Running Head: SLI grammar plurifunctional forms of DO What do children with Specific Language Impairment do with multiple forms of DO?

Purpose: This study was designed to examine the early usage patterns of multiple grammatical functions of DO in children with and without SLI. Children’s use of this plurifunctional form is informative for evaluation of theoretical accounts of the deficit in SLI. Methods: Spontaneous uses of multiple functions of DO were analyzed in language samples from 89 children, 37 children with SLI age 5;0-5;6, 37 age-equivalent children, and 15 language-equivalent children, age 2;8-4;10. Proportion correct as well as the types of errors produced were analyzed for each function of DO. Results: Children with SLI had significantly lower levels of proportion correct Auxiliary DO use compared to both control groups, with omissions of the DO form as the primary error type. Children with SLI had near-ceiling performance on Lexical DO and Elliptical DO, similar to both control groups. Conclusions: Plurifunctionality is not problematic: children acquire each function of DO separately. Grammatical properties of the function, rather than surface properties of the form, dictate whether SLI children will have difficulty using the word. Overall, these results support the Extended Optional Infinitive account of SLI, and the use of Auxiliary DO omissions as part of a clinical marker for SLI. SLI grammar plurifunctional forms of DO 3 Much of what is known about young children’s language acquisition is based on studies of particular linguistic forms or functions. For example, auxiliary DO has been the focus of numerous studies (Stromswold, 1990; Santelmann, Berk, Austin, Somashekar, & Lust, 2002; Rowland, Pine, Lieven, & Theakston, 2005; Rowland, 2007; Paradis, Crago, Rice, & Marquis, 2008). A significant outcome of this area of investigation is identification of auxiliary DO as part of a grammatical clinical marker for Specific Language Impairment (SLI) (Rice, Wexler, & Cleave, 1995; Rice & Wexler, 1996; Rice, Wexler, & Hershberger, 1998; Rice & Wexler, 2001; Rice, Hoffman & Wexler, 2009). At the same time, DO carries multiple grammatical functions, most of which have received relatively little investigation in the child language literature. Multiple grammatical functions for a given form could increase the complexity of a child’s task when acquiring auxiliary DO and possibly could contribute to the documented difficulty in acquiring this grammatical form. The general purpose of this investigation was to explicate five different grammatical functions of DO evident in young children’s utterances and to describe use of the different functions by children with SLI and ageand language-equivalent control groups, with the aim of evaluating theoretical accounts of the grammatical deficits of SLI. Five Functions of DO in the Adult Grammar, from Verb Phrase to Complement Phrase: Lexical Verb, Emphatic Auxiliary, Auxiliary, Tag Question, and Ellipsis DO is a flexible form in English, assuming multiple, and very different, roles in the morphosyntax (cf. Quirk, Greenbaum, Leech, & Svartvik, 1985; Lightfoot, 1999). It can serve as a main verb of the clause, as a special form of auxiliary that precedes the main verb, or as a formal grammatical marker whose sole function is to mark finiteness, an obligatory SLI grammar plurifunctional forms of DO 4 requirement for clauses involving tense and agreement features of the grammar. X’ theory provides a precise model of the grammatical functions of DO (cf. Chomsky, 1993, 1995; Haegeman, 1991). Within this model, morphology is closely related to syntax because morphological elements carry word order and phrasal movement requirements, hence the term morphosyntax (Pollock, 1989). In this model of the adult grammar, simple matrix clauses consist of a noun phrase (NP) and a verb phrase (VP). Verbs carry tense (TNS) and agreement (AGR) features that are essential for clause structure. An additional projection, the inflection phrase (IP), is needed for finiteness marking to meet the requirements of TNS and AGR checking in simple clauses. Figure 1 demonstrates an IP with DO as a lexical main verb. The subject NP is assumed to originate in the “specifier” position to the left of V and move to the “specifier” position to the left of the finite verb in I (head) position. The finite verb originates in the V (“head”) of the VP. (Note that each constituent, VP, IP, CP, has a specifier and a X’ position.) In English, lexical verbs originating in V cannot move overtly to I position of the IP. Instead, the V moves covertly to I position of the IP in order to show TNS and AGR (in Figure 1, third person present –s on the base form of DO in the clause “She does her homework.”). As a finite lexical verb, DO shares an important grammatical property with other lexical finite verbs: It cannot raise to the left of the subject. “*Does she her homework?” and “*Completes she her homework?” are both ungrammatical clauses. In another way lexical DO is different from other lexical verbs: The meaning is under specified and must be determined from the interaction of its direct object with context. “She does her homework” may mean “She completes her homework” (with a more semantically specified main lexical verb) or “She is working on her homework,” (with a different main lexical verb and an auxiliary form of BE that appears in I to carry TNS and AGR marking). Here we are SLI grammar plurifunctional forms of DO 5 interested in the similar grammatical properties of main verb DO and other lexical main verbs, and will refer to this function as Lexical Verb DO in our coding system 1 .We are also interested in whether children could confuse lexical main verb DO with auxiliary DO which does raise to the left of the subject. [Insert Figure 1] In a different grammatical context, DO can appear as a stressed form before the main verb to mark a finite clause as positive rather than negative. As illustrated in Figure 2, in this usage DO carries TNS and AGR in I, the head of IP projection, in communicative situations in which positive DO denies a stated or implied negative. Here we refer to this usage as Emphatic DO. This form of DO also cannot raise to the left of the subject. “*Does she swim?” loses the emphatic positive sense and instead is interpreted as the auxiliary use of DO. It is not clear if emphatic DO raises from V to I, or is inserted directly into I; for our purposes we focus on the fact that it clearly is in I as a verb form carrying finiteness marking. [Insert Figure 2] Auxiliary DO in this study refers to auxiliary DO use in questions, and is illustrated in Figure 3. Pollock (1989) worked out the syntactic properties of auxiliary DO and copula and auxiliary BE in English. Under this model, questions are derived from the matrix clauses via movement of TNS and AGR features to a projection above (i.e., to the left of) the IP, known as the complementizer phrase (CP). This projection also includes two sites: C and the specifier of C’ position. As shown in the illustrated clause, “Where does the girl swim?” the Whelement is moved from the lower VP to the specifier of C’ position. Non-emphatic auxiliary 1 In this text, capitalized names for the forms of DO denote the names of coding categories as implemented in this study. Uncapitalized names are generic linguistic terms for the structures. In this section, first use of coding category names is italicized, as are the Government and Binding theory terms for morphosyntactic phrasal projections. SLI grammar plurifunctional forms of DO 6 DO occupies the C position where TNS and AGR appear via movement from I. The standard account is that auxiliary DO moves from IP in the same way as auxiliary BE (Pollock, 1989) although Chomsky suggests it may be directly inserted in C (1995, p. 164, footnote 20). Figure 3 depicts the standard version. In this position DO does not carry the meanings of either the lexical verb usage or the emphatic DO usage. Because there is no apparent meaning for this use of DO it is sometimes called “Dummy” DO. Lightfoot (1999) argues that the introduction of DO as solely a structural requirement was triggered by historic changes in the verb movement system of the English grammar. Importantly, although lexical verb and emphatic DO carry TNS and AGR they appear in the IP, not CP, projection. [Insert Figure 3] A type of auxiliary DO also appears in reduced clauses that are understood to repeat the content of an earlier clause, sometimes referred to as “stranding.” Elliptical DO is illustrated in Figure 4. This example is a reply to a question such as “Does she swim in the morning?” In the reply, “She does,” the omitted VP is inferred from the previous utterance. Of interest here is the requirement that elliptical DO carries TNS and AGR as the occupant of I. In this example it does not move to C, to the left of the subject. [Insert Figure 4] Figure 5 illustrates the finiteness marking property of DO in tag questions, referred to here as Tag DO. The example clause, “She smiles, doesn’t she?” is similar to elliptical DO in that the tag construction, “doesn’t she” is related to another matrix clause. As such, the choice of DO for both elliptical DO and in tag constructions is linked to lexical verbs in matrix clauses. If the matrix clause has BE copula or auxiliary, as in “She is happy,” DO is not allowed and BE auxiliary must appear in the tag question, “isn’t she?” or elliptical sentence, SLI grammar plurifunctional forms of DO 7 “She is.” There are also important differences between elliptical and tag functions of DO. In tag constructions DO moves to the left of the subject for TNS and AGR marking to C of the CP projection, with interrogative force. Also, as shown in the examples, the tag question must contrast positive and negative polarity, such that a positive clause has a negative tag. In the interest of simplicity, this semantic/pragmatic requirement is not included in Figure 5. [Insert Figure 5] In summary, like other lexical main verbs, lexical DO includes the feature “+verb” and cannot move overtly outside the VP and thus cannot raise around the subject. Emphatic DO precedes the lexical verb in what could appear to be the auxiliary position but it cannot raise to the left of the subject, presumably for semantic reasons. Auxiliary DO is generated in I and moves to C in main clause interrogatives and tag-questions. Elliptical DO stays in I with an omitted VP. DO in tag questions moves to C and coordinates the selection of DO if a lexical verb appears in the matrix clause and contrasts the polarity of the tag (negative vs affirmative) with the polarity of the main clause. Children must sort through these various grammatical requirements as they encounter instances of DO use in the speech they hear and as they employ the forms of DO in their utterances. In short, there is much more to DO than auxiliary use in questions, with many possibilities of confusion about when movement is allowed to I or C and when movement is not allowed. Brief Summary of Children’s Acquisition of DO and DO as Part of a Clinical Marker of