Dativizable or Non-dativizable: That is the question? A syntactic- semantic analysis of English (non)-dativizable constructions in the production of a set of 2L1 English/Spanish simultaneous bilingual twins

Calder´on, S. S. and Fuertes, R. F. (2016).Xjenza Online, 4:44–57. Xjenza Online - Journal of The Malta Chamber of Scientists www.xjenza.org DOI: 10.7423/XJENZA.2016.1.07 Research Article Dativizable or Non-dativizable: That is the question? A syntacticsemantic analysis of English (non)-dativizable constructions in the production of a set of 2L1 English/Spanish simultaneous bilingual twins Silvia S´anchez Calder´on∗1and Raquel Fern´andez Fuertes1 1University of Valladolid, Plaza Campus Universitario, Valladolid 47011, Spain Abstract. This paper analyzes the syntactico-semantic factors which trigger Dative shift in English dativizable verbs, i.e. those verbs that allow alternation between double object and prepositional complement constructions. It also focuses on non-dativizable verbs that restrict their subcategorization framework to either double object or prepositional complement constructions. This syntactico-semantic relation between dativizable and nondativizable structures is addressed in acquisition by examining the incidence for the two verb types in a set of English/Spanish 2L1 bilingual twins. Our results show that the syntactic and semantic features that dativizable and non-dativizables present go hand in hand with the age of first occurrence and the language development of the participants. Hence, dativizable to-dative double object constructions (DOC) are the utterances produced the earliest at the age of 2, as opposed to dativizable to/for-datives and non-dativizable constructions, which begin to emerge at around the age of 3. Finally, our results also suggest that the high adult input frequency explains the twins’ early production of dativizable structures and that, in the same way, the children’s low exposure to non-dativizable utterances correlates with the later occurrence in the twins’ spontaneous production. Keywords: dativizable, non-dativizable, DOC, to/fordative, bilingual acquisition, L1 English 1 Introduction Over the last few decades, the so-called (non)-dativizable constructions have been broadly studied from different syntactico-semantic perspectives. As illustrated in (1), the verbal subcategorization of these structures can project their arguments as double object complements (i.e. indirect object (Oi)-noun phrase (NP) along with a direct object (Od)-NP), or as an object complement followed by a prepositional object preceded by the preposition to or for, as shown in (2). Despite the fact that dativizable verbs have both verb complementation possibilities, there are some others, as shown in (3), that lack this syntactic alternation (non-dativizable verbs, henceforth). (1) a. He bought me a beer (dativizable fordative double object) b. They offered her some food (dativizable to-dative double object) (2) a. He bought a beer for me (dativizable fordative) b. They offered some food to her (dativizable to-dative) (3) a. He told me that story (non-dativizable double object) b. The teacher explained it to his students (non-dativizable to-dative) c. I thanked her for her help (non-dativizable fordative) The possibility of a single verb to project their verbal argument complements both as double objects (DOCs) (examples in (2)) and as object plus prepositional constructions (examples in (1)) has triggered the debate in the literature of dativizable constructions as to which structure is syntactically base-generated and which one is derived (see section 2). Contrarily, the so-called nondativizable verbs (examples in (3)) restrict their sub- *Correspondence to: Silvia S´anchez Calder´on (silvia-sanc[email protected]a.es) c 2016 Xjenza Online 45 Dativizable or Non-Dativizable: That’s the question? The English production of a set of 2L1 bilingual twins categorization framework, as either DOCs or prepositional structures, they lack a dativizable counterpart. In view of the potential argument alternation or dative shift of some verbs, semanticists have also posed some issues concerning the meaning that dativizable and non-dativizable constructions convey. Some linguists claim that there is no difference in the meaning when verbs allow argument alternative choices (Aoun & Audrey Li, 1989; Hale & Keyser, 2002; Krifka, 2003; Levin, 1993; Oehrle, 1976; Pinker, Lebeaux & Frost, 1987; Pylkk¨anen, 2002) unlike others who argue that there are semantic distinctions between dativizable DOCs and dativizable prepositional constructions (Green, 1974; Krifka, 2003; Greenbaum, Leech, Svartvik & Quirk, 1985). Taking these syntactico-semantic issues into account, the aim of this study is to disentangle the syntactic and semantic dichotomies that exist in the literature of English dativizable and non-dativizable constructions by looking at their production in spontaneous speech. The participants of this research are a set of English/Spanish simultaneous bilingual twins, Simon and Leo. As mentioned above, only one of the L1s of these children is examined, namely, English. This paper is organized as follows: section 2 reviews previous works both from the side of linguistic theory discussing the factors governing the syntax and semantics of these constructions, as well as those from the point of view of the acquisition of dativizable and non-dativizable constructions. Section 3 includes the hypotheses that guide this study. Data selection and classification criteria are presented in section 4. The data analysis is developed in section 5 on the basis of the following variables: (a) age, (b) MLUw, and (c) input. Section 6 presents the conclusions and points to directions for further work. 2 Theoretical background and previous acquisition studies 2.1 Syntactic approach to (non)-dativizable constructions The so-called dativizable constructions have led to a dichotomy as far as their syntactic derivation is concerned. Some linguists (Aoun & Audrey Li, 1989; Snyder & Stromswold, 1997) claim that dativizable double object constructions are the basic structure from which dativizable to/for-datives derive. Alternatively, there are those who argue that dativizable DOCs are syntactically generated from to/for-dative structures (M. C. Baker, 1997; Chomsky, 1955; Larson, 1988, 1990). Regardless of the semantic principle of Uniformity of Theta Assignment Hypothesis (M. C. Baker, 1988, UTAH,) which states that “identical thematic relations between items are represented by identical structural relations between these items at D(eep)-structure”, the underlying symmetry of thematic roles between dativizable DOCs and dativizable to/for-datives is not displayed in their S(urface)-structure. According to Larson (1988), to/for-dative constructions are the basic dativizable constructions from which dativizable to/for-dative DOCs derive. The verbal head in dativizable to/fordatives subcategorizes for a prepositional phrase (PP), headed by the preposition to or for. As illustrated in (4), the verbal head offered triggers head-to-head movement and rises to the specifier of the higher V’ in order to meet Case and theta requirements, leading to what Larson terms the non-shifted version. In turn, Larson claims that dativizable DOCs derive from dativizable to/for-datives in a passive-like process, shaping, what he terms, the shifted version. More specifically, and as shown in (5), the preposition to, which is a case assigner in the non-shifted version in (4), absorbs its case being assigned to its prepositional complement her. This absorption of Case triggers NP-movement of her to the specifier of the lower VP. Similarly, the verbal head rises into the specifier of the higher V’ in order to assign Case and theta role to its adjacent argument her. Due to this verbal movement, the Od in the non-shifted version is caseless; hence, it reduces its argument position to a non-thematic adjunct position, analogous to passive byphrases, as depicted in (5). (4) They offered some food to her (dativizable todative) 10.7423/XJENZA.2016.1.07 www.xjenza.org Dativizable or Non-Dativizable: That’s the question? The English production of a set of 2L1 bilingual twins 46 (5) They offered her some food (dativizable DOC) Alternatively, it has been argued that dativizable to/for-datives are derived from dativizable DOCs by a passive-like process, where, as opposed to the previous argument, the Oi occupies an adjunct position via NP-movement (Aoun & Audrey Li, 1989; Snyder & Stromswold, 1997). Hence, example (4) derives from (5). In fact, as will be discussed in section 2.3, Snyder and Stromswold (1997) argue that the lexical item to is the factor delaying the acquisition of dativizable todative constructions. Unlike dativizable verbs whose arguments can be projected as double object complements or as objects followed by prepositional complement constructions headed by the lexical item to/for, non-dativizable DOCs and non-dativizable prepositional verbs do not have the syntactic strength to trigger dative shift. As illustrated in (6a), the non-dativizable verb explain can only subcategorize to-dative complements, hence the ungrammaticality of (6b). (6) a. The teacher explained it to his students (non-dativizable to-dative) b. *The teacher explained his students it (ungrammatical nondativizable to-dative) Larson (1990) accounts for the argument restrictions of non-dativizable to-dative verbs as the impossibility of the oblique preposition to undergo case marking. Consequently, unlike his argument concerning the derivation of dativizable DOCs (Larson, 1988), the absorption of the preposition to in prepositional non-dativizables via a passive-like process would violate the recoverability of deletion or Dative shift. Furthermore, the syntactic alternation of nondativizable constructions is mainly restricted in the lexicon. Thus, verbs of Latin origin, as exemplified in (8) with the verb donate, can only take an [NP to-NP] complement. Furthermore, this factor that prevents the alternation of non-dativizable structures is also linked to verbal semantics. In particular, only morphologically native verbs, as illustrated in (7), are capable of triggering dative shift (Mazurkewich & White, 1984; Green, 1974; Oehrle, 1976). (7) a. He gave me $50 (DOC dativizable native verb) b. He gave $50 to me (to-dative dativizable native verb) (8) a. He donated $50 to me (non-dativizable to-dative Latinate verb) b. *He donated me $50 (ungrammatical nondativizable DOC, Latinate verb) 2.2 Semantic approach to ditransitive constructions Regarding semantic factors in (non)-dativizable ditransitive constructions, different studies have pointed out that there are divergences in meaning between both structures (Aoun & Audrey Li, 1989; Hale & Keyser, 2002; Krifka, 2003; Levin, 1993; Oehrle, 1976; Pinker et al., 1987; Pylkk¨anen, 2002). These denotative differences, as will be shown below, do not go hand in hand with the syntactic derivation of both structures argued in section 2.1. Nevertheless, there are arguments that point to a common underlying semantic ground between dativizable and non-dativizable constructions (Green, 1974; Krifka, 2003; Greenbaum et al., 1985). On the one hand, some linguists assume that there is a transfer of possession relation (i.e., a cause-HAVE relation) hidden between the Oi and the Od in dativizable DOCs (Aoun and Audrey Li, 1989; Hale and Keyser, 2002; Krifka, 2003; Levin, 1993; Oehrle, 1976; Pinker et al., 1987; Pylkk¨anen, 2002). As depicted in (9), the Oi her is narrowly related in a possession relation with the Od some food. In other words, the causal agent they causes the possessor her to have some food (theme). 10.7423/XJENZA.2016.1.07 www.xjenza.org 47 Dativizable or Non-Dativizable: That’s the question? The English production of a set of 2L1 bilingual twins (9) a. They offered her some food (dativizable DOC) b. Casual agent possessor/goal possessum/theme thematic roles c. ‘X causes Y (to HAVE) Z’ semantic structure It should be noted that the possession transfer relationship in dativizable DOCs does not infer that the Oi ends up possessing the Od (Aoun & Audrey Li, 1989; Hale & Keyser, 2002; Krifka, 2003; Levin, 1993; Oehrle, 1976; Pinker et al., 1987; Pylkk¨anen, 2002). As illustrated in (10), the causal agent Isends a letter to Bill but we cannot guarantee that Bill will receive it. Thus, even though dativizable DOCs imply a possession relation, the fact that the possessor ends up possessing the theme is not always accomplished. (10) I sent Bill that letter (but he never got it) On the other hand, dativizable to/for-dative constructions are claimed to denote literal or metaphorical motion towards a goal (for example, they express a mental movement as in the verbs show or tell). In other words, they suggest a cause-GO/cause-GOAL relation between the cause/patient and the path/goal thematic roles. As exemplified in (11), the causal agent they caused the cause/patient some food to go to the path/goal her. (11) a. They offered some food to her (dativizable DOC) b. Casual agent cause/patient path/goal thematic roles c. ‘X causes Y (to GO to) Z’ semantic structure There are also arguments that point to a common underlying semantic ground between dativizable structures (Bruening, n.d.; Green, 1974; Krifka, 2003; Rappaport Hovav & Levin, 2008; Greenbaum et al., 1985). Hence, dativizable DOCs and dativizable to/for-dative constructions are equally regarded as verbs of transfer of possession (cause-HAVE relation) and verbs of motion (cause-GO/cause-GOAL). As depicted in (12), where John gives Mary a book, a twofold explanation can be given: (a) the book is moved from John’s possession into Mary’s possession (transfer of possession meaning, taking into account dativizable DOCs as the basic semantics) or, (b) possession of the book was transferred from John to Mary (cause-Go/cause-Goal relation, taking into account dativizable to/for-datives as the semantic referential point). (12) a. John gave Mary a book b. John gave a book to Mary Non-dativizable semantic restrictions are correlated with their corresponding dativizable counterparts (Comrie, Malchukov & Martin H., 2010; Oehrle, 1976). More specifically, non-dativizable DOCs are rooted in the absence of motion along some path in the same way as non-dativizable to/for-dative constructions lack a transfer of possession meaning. As can be seen in (13), the causal agent John causes the patient theta role Max to metaphorically possess a kick (path/goal). (13) a. John gave Max a kick (non-dativizable DOC) b. Causal agent cause/patient path/goal thematic roles c. ‘X causes Y (to GO to) Z semantic structure 10.7423/XJENZA.2016.1.07 www.xjenza.org Dativizable or Non-Dativizable: That’s the question? The English production of a set of 2L1 bilingual twins 48 Thus, the compositional semantics of the verbal phrase (VP) in non-dativizable DOCs is not compatible with the direction/path feature encoded in dativizable to/for-dative constructions. In other words, the example in (14) cannot be understood, according to Malchukov and Oehrle’s principles, as John’s causing a kick to go to Max; thus its ungrammaticality. (14) *John gave a kick to Max (ungrammatical nondativizable DOC) 2.3 Previous studies on the acquisition of (non)-dativizable constructions A small number of studies have been carried out in order to examine the order of acquisition of dativizable structures. Gropen, Pinker, Hollander, Goldberg and Wilson (1989) analyzed the spontaneous speech of 5 L1 English children in the Brown corpus in CHILDES (MacWhinney, 2000, Child Language Data Exchange System,). They found out that children start producing dativizable DOCs and dativizable prepositional constructions simultaneously in the second year, with neither structure subsequently uttered after the other. However, Snyder and Stromswold (1997) replicated Gropen et al.’s study including more participants and proposed that dativizable DOCs and dativizable todatives depend on two parametric properties: property A allows the grammar to produce dativizable DOCs as opposed to dativizable to-datives which merge the property of dativizable DOCs with another property, termed by Snyder and Stromswold as property B. Taking these observations into account, they implemented a study of the data from L1 English children to analyze the correlation in the acquisition of both structure types. Their results showed that dativizable DOCs are crucially acquired before dativizable to-datives as indicated by a sign test (p=.00098) and a ttest (t(11) = 4.15, p =.002). However, this order effect in acquisition did not correlate with the children’s input. The acquisition of the dative alternation has also been correlated with the presence or the lack of negative evidence of these structures in the child’s input (C. L. Baker, 1979; Gropen et al., 1989). In particular, C. L. Baker (1979) claims that children are conservative in their productions since their output will be affected by the input they have received. Hence, non-dativizable constructions like those in (15a) will be generalized as dativizable structures (see example (15b)) because the child has heard non-dativizable utterances less frequently in his/her input. In other words, children are not aware of non-dativizable verbal constraints in their emerging language development since children apply them to verbs that do not allow dative shift in the adult grammar (Mazurkewich & White, 1984). (15) a. I donated a book to the library (non-dativizable to-dative) b. *I donated the library a book (ungrammatical nondativizable DOC) Viau (2007) carried out a study dealing with the semantics of dativizable structures in L1 English acquisition. He attempted to demonstrate that dativizable DOCs denote transfer of possession by analyzing the correlation in acquisition between dativizable DOCs and have (the verb of possession par excellence), between prepositional constructions and the verb go (the verb of motion par excellence) and between dativizable DOCs and dativizable prepositional constructions and causative verbs. His results showed that dativizable DOCs convey transfer of possession since they are concurrently acquired with the verb have, in the same way as dativizable prepositional constructions denote motion shown in the correlation in acquisition with the verb go. In turn, dativizable DOCs and dativizable prepositional constructions display a correlation in acquisition with causative verbs. Despite the latter results, dativizable prepositional structures were acquired later than dativizable DOCs. It should be noted that there is not much research on the acquisition of non-dativizable constructions. Thus, further investigations should be carried out in order to fill the gap of studies that deal with this type of constructions. 3 Hypotheses Taking into account these previous studies on the linguistic description of the structures under analysis, as well as those on their acquisition, the following possible scenarios might be postulated in the acquisition data of (non)-dativizable constructions: •Hypothesis 1: If dativizable to/for-datives are transformationally derived from dativizable DOCs via a passive-like process, then dativizable to/fordatives are expected to be acquired later than dativizable DOCs. •Hypothesis 2: If, on the contrary, dativizable DOCs are transformationally derived from dativizable to/for-datives via a passive-like process, then dativizable to/for-dative are expected to be produced earlier than dativizable DOCs. •Hypothesis 3: Regardless of the derivational accounts between dativizable to/for-datives and dativizable DOCs, dativizable to-dative and fordative constructions are expected to be concurrently acquired since, despite the syntactic status of the to-PP and for-PP as subcategorized constituents or as adjuncts, respectively, they are considered to be a syntactic block of prepositional ob- 10.7423/XJENZA.2016.1.07 www.xjenza.org 49 Dativizable or Non-Dativizable: That’s the question? The English production of a set of 2L1 bilingual twins ject dativizable constructions as they are able to trigger dative shift resulting in DOCs. •Hypothesis 4: If dativizable to/for-dative constructions require a lexical item to/for in order to meet semantic motion requirements (cause-GO), then it is expected that children acquire them at a later stage, as opposed to dativizable DOCs, which, in contrast, do not require an additional constituent to accomplish the semantic attribution of transfer of possession (cause-HAVE). •Hypothesis 5: If semantics plays a role, despite restrictions in the subcategorization framework of non-dativizable constructions, a correlation in acquisition is expected between those constructions that share a semantic analogy, i.e., between dativizable and non-dativizable DOCs, as they have a cause-HAVE relation, and between dativizable and non-dativizable to/for-datives, as they imply a cause-GO relation. •Hypothesis 6: If non-dativizable constructions are subject to syntactic and semantic constraints, then a later acquisition is expected for nondativizable constructions as opposed to dativizable structures, which do not offer any limitations in their subcategorization framework or in their semantics. •Hypothesis 7: If input is taken into account, the order of acquisition will correlate with the frequency with which a child is exposed to (non)- dativizable constructions. Thus, the acquisition data, which revolve around the above-mentioned hypotheses, aim at shedding some light on the actual analysis of (non)-dativizable structures. That is, the results of this study will reflect how acquisition data can explain the syntactic and semantic properties of those constructions. 4 Methodology 4.1 Data selection In order to provide an answer regarding the relative order of production of English dativizable and nondativizable DOCs and dativizable and non-dativizable to/for-dative constructions, we have extracted child data from the FerFulice corpus available in the CHILDES project (MacWhinney, 2000, Child Language Data Exchange System,). More specifically, we have conducted our study by analyzing the production of a set of Spanish/English simultaneous bilingual twins, who were born and raised in a Spanish geographical background. Parents address children using the oneparent one-language strategy (Melanie, the mother, engages in conversations with the children in English as she is an English native speaker, whereas the father, Ivo, addresses them in Spanish as he is a Spanish native speaker) (Fern´andez Fuertes & Liceras, 2010, more information on the children’s background as well as on the data collection process appears in). It should be noted that the analysis of this study has focused on the spontaneous production of English (non)-dativizable utterances so that only one of the L1s of these children is analyzed, i.e. English. The age range that covers this longitudinal corpus ranges from 1.0 to 6.5 years old and for this study the corpus has been analyzed in its entirety. The search for both dativizable and non-dativizable ditransitive instances has been manually done by taking into account the following patterns of subcategorization: (a) S+V+NP+NP (for (non)-dativizable DOCs) and (b) S+V+NP+to/for-NP (for (non)-dativizable to/fordative constructions). We have implemented the same process for the twins and for the child-directed speech (to analyze the adult input). 4.2 Data classification Data have been classified regarding the type of participant, the children’s age and the Mean Length of Utterance measured in words (Brown, 1973, MLUw;). Besides, concerning the type of construction, we have considered those structures which trigger Dative shift (i.e. dativizable structures), classifying them as those verbs which subcategorize for an Od along with a fordative PP and, which are dativizable as DOCs, as shown in (16). Similarly, we have also taken into account dativizable constructions which allow for a to-dative PP in their subcategorization framework, and which, in turn, trigger Dative shift of their arguments as dativizable DOCs, as depicted in (17). (16) a. He bought a house for me (dativizable fordative) b. He bought me a house (dativizable fordative DOC) (17) a. He gave a present to me (dativizable todative) b. He gave me a present (dativizable to-dative DOC) Apart from those constructions that are dativizable as both to/for-dative and dativizable DOCs, we have also analyzed utterances which do not allow for Dative shift, that is to say, non-dativizable constructions. Examples in (18) and (19) show to/for-dative and non-dativizable structures, respectively. 10.7423/XJENZA.2016.1.07 www.xjenza.org Dativizable or Non-Dativizable: That’s the question? The English production of a set of 2L1 bilingual twins 50 (18) a. I have something for you (non-dativizable fordative) b. He said that to the mice (non-dativizable to-dative) (19) They called her Snow White (non-dativizable DOC) In our data classification, we have discarded the following cases: (a) utterances whose to/for-PP denotes a locative or temporal meaning, as in (20) and (21), respectively, (b) passive idiomatic expressions such as those in (22), since they are fixed expressions that lack an active counterpart, (c) DOCs which are paraphrazable as one single verb, as in (23), where the constituent give it a try can be paraphrased as to try since they have a monotransitive nature, (d) for-PPs with a proxy meaning, i.e., those constructions where the preposition for is paraphrazable as in exchange for, such as those in (24) since the PP lacks a recipient connotation, (e) those constructions which do not show a canonical SVOO or SVO+to/for-PP order, as in example (25), which illustrates a non-canonical word order DOC whose wh-Od has undergone wh-movement for syntactic reasons; (f) ditransitive structures which subcategorize afrom-PP, such as those in (26). These structures have been discarded since the PP denotes a source meaning, contrary to the recipient theta role patterns that (non)- dativizable prepositional constructions present in their PP. (20) *SIM: he drives her to the car (21) *MEL: who came for your birthday? (22) *LEO: I am used to it (23) *MEL: give it a try (24) *MEL: you traded the plane for the little blue pistol (25) *MEL: what did he tell you? (26) *EMM: you learn that from mommy, don’t you? 5 Results and Discussion In order to analyze the relative order of acquisition between dativizable and non-dativizable constructions, we have taken into account three variables, as follows: (a) age of production, (b) language development as measured in terms of MLUw, and (c) the effect of adult input. 5.1 Age of onset of production This section discusses three different analyses: (a) the age of onset production that the participants display regarding dativizable to/for-dative constructions as opposed to dativizable DOCs, (b) the results obtained concerning the first occurrence of non-dativizable structures, and (c) it offers comparative findings between structures that trigger dative shift (i.e., dativizable utterances) and non-dativizable constructions. It should be highlighted that in order to determine the age of onset of production, we have taken into account the first clear use of dativizable and non-dativizable constructions. 5.1.1 Age of onset of production: dativizable constructions As illustrated in Table 1, Simon and Leo start producing dativizable to-dative DOCs at around 2 years old. In particular, Simon begins to utter them at 2;03.26 as opposed to Leo, whose first constructions appear one month later, that is to say, at 2;05.00. Examples in (27) show Simon’s and Leo’s first utterances, both produced with the verb give, the ditransitive verb par excellence. Notice also that the twins have not realized the head of the D(eterminer) P(hrase) encoded in the Od. (27) a. *SIM: Give me tv (Simon, 2;03.26) b. *LEO: Give me farmer (Leo, 2;05.00) By focusing on the first occurrences of dativizable DOCs which trigger for-dative shift, Simon and Leo, and particularly the latter participant, present a divergent age of acquisition. They start producing dativizable for-dative DOCs at 2;11.06 and 3;03, respectively, with a four month difference between them. Examples (28a) and (28b) illustrate the twin’s early production of prepositional ditransitives headed by the preposition for. (28) a. *SIM: do you make me a cake? (Simon, 2;11.06) b. *LEO: who brought me this? (Leo, 3;03.00) Broadly speaking, the twins have shown a subsequent production of prepositional ditransitives in contrast with dativizable to/for-dative DOCs. Particularly, both children start uttering dativizable prepositional utterances at around the age of 3. Nevertheless, an exception has been found in Leo’s early production of dativizable for-dative DOCs where, as shown in Table 1, his first occurrence correlates with his early production of prepositional ditransitives at around the age of 3. More specifically, and as shown in (29a) and (29b), Simon and Leo start uttering dativizable for-datives at 3;01.20 and 3;02.24, respectively, with a month difference between them. In turn, dativizable to-dative constructions, as illustrated in (30), were first produced at 3;02.12 by Simon and 3;01.06 by Leo, also showing a month’s difference between the twins’ output of those structures. 10.7423/XJENZA.2016.1.07 www.xjenza.org 51 Dativizable or Non-Dativizable: That’s the question? The English production of a set of 2L1 bilingual twins (29) a. *SIM: to make the honey just for me (Simon, 3;01.20) b. *LEO: take it and the baby for you (Leo, 3;02.24) (30) a. *SIM: we bought that to you (Simon, 3;02.12) b. *LEO: give it to the zoo (Leo, 3;01.06) When focusing on the age of early utterance of dativizable to-dative and dativizable for-dative prepositional constructions, our findings show that their age of onset production is at around the age of 3. Despite the fact that Simon and Leo start uttering dativizable prepositional structures at around the same age, they show different ages of onset for each construction. As illustrated in (30a), Simon’s early dativizable fordative prepositional ditransitives are initially uttered at 3;01.20, preceding his first production of dativizable todative prepositional constructions at 3;02.12, as illustrated in (31a). Alternatively, Leo shows a reverse order of acquisition between dativizable to/for-dative prepositional utterances since he starts producing dativizable to-dative prepositional constructions at 3;01.06, as exemplified in (31b), one month earlier than his early utterance of dativizable for-dative at 3;02.24, as can be seen in (30b). Thus, the production of dativizable to/for-datives goes hand in hand as syntactic analogous prepositional constituents, confirming our H3, as will be discussed in section 6. (31) a. *SIM: do you make me a cake? (Simon, 2;11.06) b. *SIM: give me tv (Simon, 2;05.00) In particular, Simon’s early production of dativizable to/for-dative DOCs, as illustrated in (29a), repeated as (31a) and (27a) repeated as (31b), respectively, has been observed at around the age of 2, one year earlier than his onset production of non-dativizable DOC at 3;05.12 (example in (32)). It is worth standing out that Simon’s early production of dativizable for-dative DOCs is found one year later than the utterance of his early dativizable to-dative DOC, as discussed in section 5.1.2. (32) *SIM: we call them blue (Simon, 3;05.12) However, Leo’s production of dativizable for-dative DOCs and dativizable to-dative DOCs shows a different time frame since he starts uttering the former constructions at 3;03 (as shown in (28b), repeated below as (33a), one year later than his first occurrence of to-dative DOC at 2;05 (as illustrated in (27b), repeated below as (33b)). (33) a. *LEO: who brought me this? (Leo, 3;03) b. *LEO: give me farmer (Leo, 2;05) All in all, and as illustrated in Table 1, dativizable to/for-dative DOCs are the dativizable structures produced the earliest by the twins at the age of 2, preceded by the first occurrence of dativizable prepositional utterances, displayed at the age of 3. Table 1: Age of onset production of dativizable constructions. Dativizable DOC Prepositional ditransitive to-dative for-dative to-dative for-dative Simon 2;03.26 2;11.06 3;02.12 3;01.20 Leo 2;05.00 3;03.00 3;01.06 3;02.24 5.1.2 Age of onset of production: nondativizable constructions As illustrated in Table 2, non-dativizable constructions, regardless of their type, begin to be produced at the age of 3. Thus, the age of onset production of dativizable prepositional ditransitives correlates with the early production of non-dativizables. Nevertheless, for the twins non-dativizable prepositional ditransitives and non-dativizable DOCs do not begin to emerge in a parallel way. More specifically, Simon’s linear order of the early production of non-dativizables is the following: he starts uttering non-dativizable prepositional todative constructions at 3;04.28, one month earlier than non-dativizable DOCs (i.e., at 3;05.12) and five months earlier than non-dativizable prepositional for-datives (i.e., at 3;09.13). On the other hand, Leo starts producing non-dativizable for-dative prepositional struc- 10.7423/XJENZA.2016.1.07 www.xjenza.org Dativizable or Non-Dativizable: That’s the question? The English production of a set of 2L1 bilingual twins 52 tures at 3;02.00, eight months earlier than his first nondativizable to-dative construction (i.e., at 3;10.05). As displayed in Table 2, there has not been any utterance found concerning non-dativizable DOCs across Leo’s corpus. Table 2: Age of onset production of non-dativizable constructions Non-Dativizable Prepositional ditransitive DOC to-dative for-dative – Simon 3;04.28 3;09.13 3;05.12 Leo 3;10.05 3;02.00 – Despite the fact that dativizable prepositional ditransitives and non-dativizables start being uttered at around the age of 3 (compare Tables 1 and 2), Simon and Leo display subtle differences in their order of first production. As example (34) shows, dativizable for-dative utterances are the first prepositional ditransitives produced by Simon at 3;01.20, as opposed to Leo, who starts uttering dativizable to-dative constructions at 3;02.24, as illustrated in (35). (34) *SIM: to make the honey just for me (dativizable fordative; Simon, 3;01.20) (35) *LEO: give it to the zoo (dativizable todative; Leo, 3;01.06) Likewise, non-dativizable constructions reflect differences in the twins’ order of first occurrence. Thus, as exemplified in (36), Simon begins to produce nondativizable to-datives at 3;04.28, preceding the age of onset production of non-dativizable DOCs and nondativizable for-datives, as can be seen in (32), repeated here as (37) and (38), respectively. (36) *SIM: why is the cat saying that to the mice (non-dativizable to-dative; Simon, 3;04.28) (37) *SIM: we call them blue (non-dativizable DOC; Simon 3;05.12) (38) *SIM: I want one for myself (non-dativizable for-dative; Simon, 3;09.13) On the contrary, as exemplified in (39), nondativizable for-dative constructions start being uttered by Leo at 3;02.00, preceding the first production of nondativizable to-datives at 3;10.05, as shown in (40). (39) *LEO: thank you for my playdough (non-dativizable fordative; Leo, 3;02.00) (40) *LEO: you have something to the toad (non-dativizable todative; Leo, 3;10.05) As shown in Table 2, non-dativizable DOCs are not reflected in Leo’s data, hence comparative results cannot take place regarding these structures in the twins. In view of the above-mentioned results, dativizable constructions (i.e. prepositional ditransitives and DOCs) are produced earlier than constructions which do not trigger dative shift (i.e. non-dativizables). More specifically, the relative order of early occurrence of both structures in Simon and Leo is illustrated in (41a) and (41b), respectively. (41) a. Dativizable to-dative DOC >dativizable for-dative DOC >dativizable for-dative >dativizable to-dative >non-dativizable to-dative >non-dativizable DOC >nondativizable for-dative (Simon’s relative order of first production of (non)-dativizable to/for-datives) b. Dativizable to-dative DOC >dativizable to-dative >non-dativizable for-dative >dativizable for-dative >dativizable for-dative DOC >non-dativizable to-dative (Leo’s relative order of first production of (non)-dativizable to/for-datives) These results evidence differences between the twins. On the one hand, Leo’s early production of dativizable for-dative DOCs correlates with his first occurrences of dativizable to/for-datives and non-dativizable to/fordatives at around the age of 3. On the other hand, Simon displays a correlative age of early occurrence between dativizable to/for-datives and non-dativizable constructions (regardless of the type) at around the age of 3. One year earlier, Simon shows a correlation in his first utterances of the block of dativizable DOCs (i.e., dativizable to/for-dative and DOC), which differs from Leo’s performance since he does not show a parallelism between the production of DOCs. In fact, as discussed in section 5.1.1, he begins to produce dativizable to-dative DOCs at 2;05.00, one year earlier than his first utterance of dativizable for-dative DOCs at 3;03.00. Furthermore, unlike Simon, we have not found evidence of Leo’s production of non-dativizable DOCs; hence, we cannot offer comparative results between the participants regarding these types of utterances. 10.7423/XJENZA.2016.1.07 www.xjenza.org