Second Language Research 2012 Ballester 217 41

http://slr.sagepub.com/Second Language Research

http://slr.sagepub.com/content/28/2/217The online version of this article can be found at:

DOI: 10.1177/0267658312438534

2012 28: 217Second Language ResearchElizabet Pladevall Ballester

contextChild L2 English acquisition of subject properties in an immersion bilingual

Published by:

http://www.sagepublications.com

can be found at:Second Language ResearchAdditional services and information for

http://slr.sagepub.com/cgi/alertsEmail Alerts:

http://slr.sagepub.com/subscriptionsSubscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://slr.sagepub.com/content/28/2/217.refs.htmlCitations:

What is This?

- May 23, 2012Version of Record >>

at Ulle/Biblioteca Edif. Central on April 16, 2014slr.sagepub.comDownloaded from at Ulle/Biblioteca Edif. Central on April 16, 2014slr.sagepub.comDownloaded from

http://slr.sagepub.com/


http://slr.sagepub.com/content/28/2/217

http://slr.sagepub.com/content/28/2/217



http://slr.sagepub.com/cgi/alerts

http://slr.sagepub.com/cgi/alerts

http://slr.sagepub.com/subscriptions

http://slr.sagepub.com/subscriptions

http://www.sagepub.com/journalsReprints.nav

http://www.sagepub.com/journalsReprints.nav

http://www.sagepub.com/journalsPermissions.nav

http://www.sagepub.com/journalsPermissions.nav

http://slr.sagepub.com/content/28/2/217.refs.html

http://slr.sagepub.com/content/28/2/217.refs.html

http://slr.sagepub.com/content/28/2/217.full.pdf

http://slr.sagepub.com/content/28/2/217.full.pdf

http://online.sagepub.com/site/sphelp/vorhelp.xhtml

http://online.sagepub.com/site/sphelp/vorhelp.xhtml





Second Language Research28(2) 217–241

© The Author(s) 2012Reprints and permission: sagepub.

co.uk/journalsPermissions.navDOI: 10.1177/0267658312438534

slr.sagepub.com

secondlanguageresearch

Child L2 English acquisition of subject properties in an immersion bilingual context

Elizabet Pladevall BallesterUniversitat Autònoma de Barcelona, Spain

AbstractAlthough thoroughly analysed in adult second language acquisition (L2A), the acquisition of subject properties in child L2A has not received so much attention and the majority of studies deal with longitudinal data or only with a single subject property. This study contributes new cross-sectional data from 5-year-old Spanish children acquiring second language (L2) English in an immersion bilingual context with age of onset three and analyses the acquisition of all the subject properties traditionally associated as a cluster in comparison to a monolingual control group of the same age. The data were obtained by means of an orally elicited judgement/correction experimental task and enabled us to explore the children’s awareness of grammaticality contrasts in their L2, the influence of their L1 after two years of exposure and the extent to which the cluster of subject properties traditionally associated and studied together develop similarly in acquisitional terms. Results suggest that 5-year-old children show sensitivity to grammaticality contrasts in their L2 English and indicate that even after two years of exposure and although the age of onset of acquisition was only three, their L2 is influenced by their first language (L1) subject properties, although developmental effects also need to be taken into account when explaining the results.

Keywordschild L2A, English and Spanish subject properties, orally elicited judgment task, L2 grammaticality contrasts, L1 influence, bilingual immersion setting

I Introduction

The status of child second language acquisition (L2A) in itself and the importance of age of onset of acquisition in comparison to adult L2A and monolingual and bilingual first language acquisition (L1A) is now a research focus in a variety of languages and linguistic phenomena (see, among others, Blom, 2008; Haznedar and Gavruseva, 2008; Ionin,

Corresponding author:Elizabet Pladevall Ballester, Departament de Filologia Anglesa i Germanística, Facultat de Filosofia i Lletres, Edifici B, Universitat Autònoma de Barcelona, 08193 Bellaterra, Cerdanyola del Vallès, Barcelona, SpainEmail: [email protected]

438534 SLR28210.1177/0267658312438534Pladevall BallesterSecond Language Research2012

at Ulle/Biblioteca Edif. Central on April 16, 2014slr.sagepub.comDownloaded from



218 Second Language Research 28(2)

2008; Meisel, 2007, 2008, 2009; Pladevall-Ballester, 2010; Schwartz, 2003, 2004; Unsworth, 2005a, 2005b). The present study focuses on a group of 5-year-old child sec-ond language (L2) English learners whose first language (L1) is Spanish, who live in Spain and whose age of first exposure to L2 English in an immersion bilingual context was three. More specifically, the study explores the children’s awareness of grammatical-ity and ungrammaticality with respect to L2 English subject properties in comparison to native monolingual controls of the same age. The linguistic items to be analysed include the cluster of properties associated with the traditional pro-drop parameter (Jaeggli, 1982; Jaeggli and Safir, 1989; Rizzi, 1982, 1986) (i.e. null and overt pronominal and expletive subjects, preverbal and postverbal subjects and that-trace effects) and inflectional mor-phology, although results will be discussed according to more recent theoretical propos-als (Platzack, 2004; Rizzi, 2002, 2005). The acquisition of subject properties in child L2 English has not been extensively analysed and the majority of studies dealing with it explore longitudinal data (Hilles, 1991; Lakshmanan, 1991, 1994; Mobaraki et al., 2008; Park, 2004). This study contributes new cross-sectional data from Spanish chil-dren acquiring L2 English in an immersion bilingual context. The data were obtained by means of an orally elicited judgement/correction experimental task, where children were presented with grammatical and ungrammatical sentences uttered by puppets and were prompted to judge whether the sentences sounded strange and whether to change or repeat them as originally said. The data obtained will enable us to delve into the chil-dren’s awareness of grammaticality contrasts in their L2, the influence of the L1 on the children’s L2 after two years of exposure and the extent to which the cluster of subject properties traditionally associated and studied together develop in a similar way. Results suggest that 5-year-old children are indeed sensitive to grammaticality contrasts in their L2 English as they were capable of making correct and incorrect judgements about the sentences, and there is a clear difference in their judgements of grammatical sentences (which were extremely accurate) and their judgements of ungrammatical sentences, whose rejection was significantly lower than in the native control group. Yet, some incor-rect judgements and repetitions of sentences included hesitations, which indicate the children’s awareness of ungrammaticality but incapacity to positively discern and cor-rect the target ungrammatical element. In spite of the children’s linguistic sensitivity in their L2, results point towards the existence of a certain influence of L1 subject proper-ties – but not structural transfer – on child English L2A still after two years of exposure, with that-trace effects patterning differently from the rest of subject properties, as expected. Yet developmental effects are also contemplated when discussing the results.

This article is organized as follows: Section II focuses on the theoretical framework adopted in the study; studies related to subjects in adult and child L2 English are briefly presented in Section III; Section IV deals with the study of child L2A; Section V presents the research questions and hypothesis of the present study and its methodological aspects; Section VI analyses the results, Section VII discusses the results and, finally, Section VIII draws essential concluding remarks.

II English and Spanish subject properties

Early research on the L2A of English subjects adopts the traditional version of the Null Subject Parameter (Chomsky, 1981; Jaeggli, 1982; Jaeggli and Safir, 1989; Rizzi, 1982, 1986)




Pladevall Ballester 219

as originally proposed within the Government and Binding framework, which provided an account of a cluster of phenomena present in null subject languages but absent in overt subject languages, on the basis of parametric variation related to richness of inflec-tion and identification properties and so by which languages with a positive null subject value (i.e. Spanish and Italian but not English or French) share a number of properties which include null subjects, the absence of expletive pronouns, subject-verb inversion, that-trace sequences and rich verbal agreement.

The Null Subject Parameter is accounted for differently within Minimalism and hence the related presumably clustered properties of subjects in English and Spanish are analysed differently. The present research follows Platzack’s (2004) syntactic account of subjects: The Person Phrase Hypothesis, which provides a unified way of accounting for subjects and agreement in both null and non-null subject languages according to a minimalist view of parameters and the EPP.1 Following Ritter (1995), Platzack (2004) argues that agreement (Agr) originates as the head of a Person Phrase (PersP), which can take a DP as a complement and which is merged in a theta position, as the subject in Spec-vP. Depending on the element taken from the lexicon to express Persº several possibilities arise. If the element expressing Persº has grammatical but not phonological features, PersP will appear as a DP. If the element merged as Persº has grammatical as well as phonological features, this element can either be a free morpheme – in English and Spanish – or a bound morpheme – only in Spanish. If it is a free morpheme, it will be a pronoun with or without a DP complement. If it is a bound morpheme (Agr), as it happens in languages with rich agreement, such as Spanish or Italian, it should find a host outside PersP. T will assign Case to PersP and the EPP feature in T will attract Persº (Agr) to T, projecting a new PersP. V will also be attracted to T and then internally merged to PersP without projecting. Finally, morphological merger will take place and Persº+T will be realized as an affix on V. In languages without V-raising, such as English, agreement is not an argumental affix and Persº is either a free morpheme (i.e. a pronoun) or it is not phonologically realized and PersP is represented as an ordinary DP. In those languages, PersP (i.e. the subject) would subsequently raise to Spec-T attracted by the EPP feature in T.

That-trace effects were also traditionally linked to the Null Subject Parameter and more specifically to the ECP and the availability vs. non-availability of postverbal sub-jects in a language. More recent syntactic accounts have dissociated that-trace effects from the Null Subject Parameter, such as that of Pesetsky and Torrego (2001, 2004), which considers that-trace effects an instance of the T-to-C movement asymmetry simi-lar to that found in English matrix wh-questions. Both phenomena involve the absence of an element in C, an auxiliary verb or the complementizers that, when there is subject wh-extraction. The authors propose that, in that-trace effects constructions, that is not C but ‘an instance of T that has moved to C’ (2001: 371). When the wh-phrase is a non-subject, that will move from T to C. Following Pesetsky and Torrego’s (2001, 2004) account of that-trace effects, Gallego (2006) observes that Romance languages cannot make use of subject DPs to delete uT on C and therefore, that-trace effects are not attested. The present study will depart from the traditional concept of clustered null subject properties, which in the light of the results will be discussed under more recent theoretical claims.





III Subjects in L2 English

The study of subjects being omitted and all their related properties in early child lan-guages irrespectively of whether the target grammar allows them or not has been a focus of research within the study of first language acquisition from the traditional Principles and Parameters framework to more recent accounts (Hyams, 1986, 1989, 1992, 1996; Radford, 1990, 1992; Rizzi, 1982, 1986, 1994a, 1994b, 2000, 2005; Roeper and Rohrbacher, 1995, among others) and has inevitably influenced the study of subject properties in L2A. Although the acquisition of subjects in adult L2 English has been extensively studied in a variety of L1s and mainly using cross-sectional experimental methodology in instruction acquisition settings in both English and non-English speaking countries (Phinney, 1987; Tsimpli and Roussou, 1991; Wakabayashi, 2002; White, 1985, 1986, among others) child L2 English subjects have been less widely explored and only using longitudinal data col-lection methods in naturalistic acquisition settings in English-speaking countries and mainly focusing on the production of null/overt subjects and inflection (Hilles, 1991; Lakshmanan, 1991, 1994; Mobaraki et al., 2008; Park, 2004). Early research on the L2A of English subjects adopts the traditional version of the Null Subject Parameter outlined above. This is the case in three of the most relevant accounts of adult English L2 subjects, namely White (1985, 1986), Phinney (1987) and Tsimpli and Roussou (1991), which fol-lowed Rizzi (1982, 1986).

As for subjects in child L2 English, Hilles (1991) and Lakshmanan (1991, 1994) study them on the basis of the Morphological Uniformity Principle (MUP) (Jaeggli and Safir, 1989), which states that null subjects are only licensed in languages with uniform verb paradigms and that the acquisition of the Null Subject Parameter should be related to the emergence of verbal inflection. Wakabayashi (2002), for adult L2 English, and Park (2004) for child L2 English, which compare the L2A of the non-allowed null sub-jects in English by Japanese and Spanish native speakers, adopt the Minimalist Program framework and its claim that language acquisition involves the learning of formal fea-tures and finally, Mobaraki et al. (2008) discuss in relatively traditional terms whether child L2 English subject use indicates presence/absence of functional projections in early child grammars in comparison to child L1A.

Research on child L2 English subjects has mainly dealt with longitudinal data and the observation of null/pronominal subjects and their relationship with inflection. There is thus a clear need for cross-sectional data and the analysis of the ‘traditional’ cluster of subject properties as a whole in child L2 English, as was done in adult L2A earlier studies.

IV The importance of child L2A

The study of child L2A is now a focus of research, particularly regarding its similarities and differences in different grammatical areas with respect to child L1A and adult L2A (Blom, 2008; Ionin, 2008; Meisel, 2007; 2008, 2009; Pladevall-Ballester, 2010; Schwartz, 2004; Unsworth, 2005b; Whong-Barr and Schwartz, 2002, among others). Child L2A shares representational and developmental properties of both L1A and adult L2A and hence it is a valuable tool in the general picture of language acquisition (Lakshmanan, 1995; Schwartz, 1992, 2003, 2004). The child L2 learner is cognitively more mature than





the child L1 acquirer but less mature than the adult L2 learner, she is affected by the existence of the L1 grammar and she is remarkably successful in ultimate attainment studies. It is also important to distinguish child L2A from bilingual L1A. The former is defined as successive child bilingualism, whereas the latter is defined as simultaneous child bilingualism, which refers to the acquisition of two languages from birth. Child L2A implies acquiring an L1 and starting L2 exposure at least when most grammatical principles of the native language are assumed to have been acquired, around the age of four (Guasti, 2002). Yet, and on the basis of maturational effects on successful ultimate attainment, another age of onset needs to be established as the limit after which native-like ultimate attainment is no longer possible, namely around the age of seven (DeKeyser, 2000; Schwartz, 2004; Unsworth, 2005a). Non-native acquirers who are first exposed to an L2 between the ages of 4 and 7 are thus child L2 acquirers. In Haznedar and Gavruseva’s (2008) words, child L2 learners are ‘successive bilinguals who have acquired the basic fundamentals of their native language (L1) and who are exposed to a second language (L2) between the ages of 4 to 8’ (p. 3). Therefore, age of onset of acquisition, as Meisel (2009) points out, is essential in order to distinguish child L2A from monolingual and bilingual L1A. However, as the author indicates, a range of variation with respect to age of onset is to be expected and hence research in this area should also include children with age of onset of three, as is the case of the present study. In fact, other proposals had also been made with regards to a lower cut off point, namely around the age of three (for discussion, see McLaughlin, 1978; Lakshmanan, 1995; Lakshmanan, 2009).

V The study

1 Research questions and hypotheses

This study examines a group of 5-year-old child L2 English learners with L1 Spanish acquiring English in the context of an immersion American school in Spain and whose age of onset of acquisition is three. The main aim of this experimental study is to explore the children’s sensitivity to grammaticality contrasts with respect to L2 English subject properties in comparison to a control group of monolingual English-speaking 5-year-old children. The lack of cross-linguistic studies on child L2 English subjects and the fact that the existing longitudinal studies mainly focused on the presence vs. absence of pro-nominal subjects and their relationship with inflection makes the study of all L1 Spanish-L2 English subject properties (i.e. null and overt pronominal and expletive sub-jects, preverbal and postverbal subjects and that-trace effects) even more relevant, par-ticularly because subject properties have been the target of some studies on L2 English in adult L2A.

Three research questions guide this piece of research:

1. Are 5-year-old children able to discern grammatical and ungrammatical subject properties in their L2 English after two years of exposure?

2. What is the role of the L1 in the children’s L2 (with respect to subject properties) after two years of exposure?

3. Do the subject properties traditionally associated (Null Subject Parameter) and studied together develop in a similar way in child L2 English?





Following Schwartz (2003, 2004), we assume that child L2A grants full access to UG and the principles that guide acquisition. Children in this study should hence show awareness of grammaticality contrasts in their L2 English as their L2 acquisition process not only relies on general learning skills but also on internal mechanisms as in L1A given that maturational constraints are assumed not to have had important effects at the time of first exposure yet. However, L1 Transfer is also assumed to be present in early stages, but since the children under study have been exposed to the L2 for two years in an immersion setting and their age of onset of acquisition is below four, L1 Transfer effects are pre-dicted to be minimal, although we need to bear in mind that the children under study do not live in an English-speaking country and hence their exposure to their L1 is greater and certain L1 influence on the children’s L2 performance is indeed expected. As com-pared to the control group under study, the experimental children will have had fewer years of exposure to English by virtue of starting their exposure at the age of three and not from birth. Hence, developmental effects may be also present in that L2 children may undergo the same developmental stages as L1 children but some years later. As for the cluster of subject properties and except for that-trace effects which, as shown in Section II, have been dissociated from the subject properties (Pesetsky and Torrego, 2001, 2004), they should develop similarly in child L2A, assuming the theoretical framework adopted here (Platzack, 2004) by which the presence/absence of null subjects and the possibility/impossibility of postverbal subjects are accounted for by the way the EPP feature in T is checked and if we assume that UG is at least partly operative. In other words, subject properties should display statistically similar results in the data. However, showing that the children’s levels of non-target L2 performance are similar across the different phenomena tested will not enable us to postulate clustering effects in child L2, since this could be due to other factors and hence our contribution will be largely descrip-tive in this respect.

2 Method

An elicited oral judgement/correction experiment (Johnny Lion and Susie Cow) was designed to elicit judgements and corrections of ungrammaticality from the children. Children were presented with grammatical and ungrammatical sentences uttered by pup-pets and were prompted to judge and indicate whether the sentences sounded right or wrong. However, 5-year-old children are not particularly familiar with the concept of a sentence sounding right or wrong. That is why the experimenter sometimes used the words ‘strange’ or ‘funny’ to exemplify what we meant or paraphrased what they were to do telling children to decide if they would say the sentence as such or not. This might have been problematic in that a given structure can sound strange but still be grammatical, although this would possibly create more problems when judging semantic or pragmatic alternatives and not so much when judging syntactic structures, as is the case in this study. It should be mentioned that the children were not assessed in their L1 to find out how they would have judged the sentences in Catalan/Spanish. This would have given us more information as to how the children were judging the sentences and would have provided more insights into the L1 transfer issue.





Once the sentence was judged, children had to decide whether to change the sentence or repeat it as it was originally said. Twenty ungrammatical and eight grammatical sen-tences were designed according to the properties to be tested and taking vocabulary items from previous class activities. The 28 sentences were divided into two blocks (14 each) which were presented to the child on two different sessions, depending on the children’s attitude and tiredness. The specific items to be tested in the sentences included four pro-nominal null subjects in main (with and without previous reference) and embedded clauses, two null expletive subjects, three postverbal subjects in main and embedded clauses, two instances of that-trace effects, three third person singular morphemes -s, two regular past morphemes -ed, two irregular past and two inflected forms in non-inflected positions. The grammatical sentences were designed to act as counterparts of each ungrammatical type of sentence and at the same time served as fillers and enabled us to see whether children were paying attention to the task. The number of sentences was kept to a minimum considering the linguistic items to be tested and the grammatical sentences were only restricted to eight to represent the number of properties to be tested. In the same vein, and as indicated, the grammatical sentences also served as fillers to ensure that chil-dren were paying attention. The restricted number of sentences was also motivated by the fact that 5-year-old children’s span of attention is relatively short and the experiment was already conducted in two sessions. Yet there were probably too few items relating to verb inflection and including grammatical and ungrammatical that-trace sentences with wh-extraction from the object position would have helped the author identify if wh-extraction, not only subject extraction, is too complex for children at this age. However and as already mentioned, including more sentences in the experiment was not considered given the age and attention capacity of the children.

The presentation of the 28 sentences was ordered in a way that could not create a pat-tern for the child. That is to say, sentences that tested the same properties were not pre-sented together and grammatical counterpart sentences were carefully inserted among the ungrammatical ones. Table 1 illustrates the sentences and their target linguistic items.

a Experimental data: Data from child L2 English were collected in an American immer-sion school in Barcelona. Children from all over the world as well as Spanish children attend the school from nursery class (three years) to 12th grade (seventeen years). Chil-dren are taught entirely in English and only hear Spanish in their Spanish Language classes.3 The school was chosen as it proved to be the school which best suited the requirements of the study, namely it provided the most intensive English immersion programme as compared to other similar schools in the area of Barcelona.4 The study aimed at Kindergarten class (5-year-olds), who had started attending the school, and hence being exposed to L2 English, at the age of 3 in nursery class. About half of the class had Spanish/Catalan parents, whereas the other half had American parents, mixed American-Spanish parents or parents with other nationalities. Though all family situa-tions were worth studying, the study obviously focused on Spanish/Catalan children, whose home languages were both Spanish and Catalan and who were only exposed to English as an L2 at school. At the time of testing, the children had been exposed to Eng-lish for about 2,362 hours during two and a half school years, so the group is clearly beyond the initial state of acquisition.





In order to familiarize with the children, I visited the class two days a week for a month and a half before the actual experimental tasks took place and I took part in all class activities with the children. I was introduced as a visiting American teacher who had to write a book about children. I therefore pretended not to understand any Spanish or Catalan so that target children would only communicate in English with me. The class had a tutor and one assistant, who were both American. Thus all communication between them and the children was expected to be in English, though the tutor could understand

Table 1 Sentences and linguistic items in the experimental task

Linguistic item Sentences

3rd person singular -s (5) My best friend play football twice a week. (9) My sister like rainforests very much.(18) Ms Valerie say that we have to clean up the tables.

Regular past morpheme -ed

(10) Last week we finish our books on rainforests.(24) The two boys fix their problem yesterday.

Irregular past (20) Last Thursday I take a book from the library.(26) The children in Kindergarten go to the gym yesterday morning.

Inflected forms in non-inflected position

(11) He didn’t waited for me!(16) They didn’t saw any Toucans in the rainforest.

Null expletives (1) Is raining a lot today. (28) Seems that Ms Valerie is very happy today.2

Null subject main clause

(7) Elizabeth likes card games. Plays ‘Snap’ every day. (reference) (21) Had fun in ‘Western Night’ and he ate a lot of pizza. (no

reference)Null subject subordinate clause

(1) My sister is always tired because works a lot! (13) We will not go to the patio if don’t clean up the tables.

Postverbal subjects main clause

(7) Has come my granny from America. (25) Cried the baby all night long.

Postverbal subjects subordinate clause

(14) They didn’t know when finished the class.

That-trace effects (4) Who did you say that came late?(23) Who do you think that will win the game?

Grammatical sentences

(3) Last Monday we went for lunch very late. (irregular past) (6) My sister loves apples so she eats one every day. (subject main

clause)(12) Who do you think will arrive first? (that-trace)(15) It is raining very hard these days. (Expletive)(17) We will be late if we don’t take the train. (subject subordinate

clause)(19) He didn’t know when the class started. (preverbal subject

subordinate clause) / (inflected form in non-infl position)(22) The children in Kindergarten finished their drawings very

quickly. (regular past) / (preverbal subject main clause)(27) My friend John likes toucans and monkeys very much. (3rd

person singular)





Spanish and Catalan. All class activities and games were held in English and other school staff such as secretaries, lunchtime teachers or assistants addressed the children in English. Only forty minutes a day of Spanish language class was given but Spanish chil-dren tended to address each other in Spanish in the playground, but not inside the class-room, unless there was any other international student around. No questionnaire could be used to find out the extent to which the children used their L1s and the L2: neither the children, who obviously cannot reflect on this being 5-year-olds nor the parents who rarely bring or pick up their children from this type of school. That is why we spent a month observing the children who took part in the experiment.

Fifteen Spanish/Catalan children were selected to participate in the study. Two stu-dents were left out as permission was not obtained to carry out the study with one of them and the other had severe difficulties in understanding the task and English in general as he had just started attending the school that same year. Children were individually taken to a classroom where a puppet theatre was set up. Two experimenters were involved in the task; Experimenter 1 (EXP1), who sat by the child, explained the task and prompted the child to judge the sentences uttered by the puppets, and Experimenter 2 (EXP2), a volunteer undergraduate student, who did the puppets show behind the theatre and uttered the sentences.

EXP1 told the child a brief story about Johnny Lion and Susie Cow (the two puppets), which acted as a prompt for children to judge the sentences: ‘Johnny Lion and Susie Cow are best friends. They go to the same school but Johnny has had a bit of an ear infection for some days and has stayed at home. Susie Cow is visiting him and is telling him sev-eral things about school. The problem is that Johnny’s ear is not totally recovered and cannot hear very well and on top of that, Susie Cow sometimes speaks funny. And this is why they need your help’. Susie Cow needed the child to repeat what she was saying to Johnny Lion but since she sometimes spoke funny, the child was told to decide whether the sentence uttered sounded right or funny and then repeat it to Johnny Lion or change it so that Johnny Lion could understand better. It was emphasized that what Susie Cow was saying was right and that they had to pay attention to the way she said it. As Susie Cow uttered every sentence, Johnny Lion pretended not to hear or understand well the sentence. EXP1 reminded the child that he/she had to help Susie Cow repeat the sentence for Johnny Lion and Susie Cow talked directly to the child, asked for his/her help and thanked him/her after every sentence. Sentences could be repeated twice, either by the puppet or by EXP1, if the child did not understand or hear them. At the same time, if the child repeated the sentence without paying attention or analysing whether it sounded right or funny, the task was stopped or the response was not given any score and counted as a drop-out. The task was originally designed to take place in two sessions of 14 sen-tences each. However, depending on the child’s attitude and tiredness, the whole task was carried out in one only session or in more than two sessions.

b Control data: English children control data were collected in a community primary school in Milton Keynes, United Kingdom, which offers education to children from Early Years (4–5 years old) to Year 7 (eleven-twelve years old). In order to collect paral-lel data to the experimental data, research focused on the Early Years class (Tadpoles Class: 5-year-olds). The general procedure of the task, the structure of the sentences, the





linguistic items to be tested, the order of presentation and the number of informants (i.e. fifteen) remained exactly the same. Experimenter 2 (EXP2) (i.e. the experimenter who did the puppets at the experimental school) was a Teaching Assistant from the Tadpoles class here, who knew the children well. The prompt story about Johnny Lion and Susie Cow remained the same. As for the timing of the experiment, although it was initially planned that children would carry it out in two different sessions, as it had been the case at the experimental school, all of them proved to be capable of listening and judging the 28 sentences in only one session.

c Data analysis: Experimental and control children’s responses were first transcribed and then coded and analysed according to instances of correct judgements (i.e. accept-ance of grammatical sentences and rejection/correction of ungrammatical sentences) and incorrect judgments. Then, instances of hesitations, irrespectively of whether the judge-ment was analysed as correct or incorrect, were coded. Hesitations were taken to indicate that the child perceived the oddness of the sentence though s/he was not able to produce the target corrected grammatical construction and were considered to be those instances of filled or unfilled pauses with a cut-off point of 0.5 seconds and that could clearly be identified by the experimenter. In order for the analysis of hesitations to be more objec-tive, there should have been two raters noting each instance of hesitation and hence a check of inter-rater reliability, which was at the time of the experiment not possible. Yet hesitations were still included in the analysis since they contribute valuable information on how the children perceived and reacted to the sentences.

Ungrammatical sentences were grouped into six variable conditions representing ‘Verbal Inflection’, ‘Null Subjects in Main Clauses’, ‘Null Subjects in Subordinate Clauses’, ‘Null Expletives’, ‘Postverbal subjects’ and ‘That-trace sequences’. Grammatical sentences were grouped into the same variable conditions representing ‘Verbal Inflection’, ‘Overt Subjects’, grouping null subjects in main and subordinate clauses under the same variable, ‘Overt Expletives’, ‘Preverbal Subjects’ and ‘That-trace sequences’.

In order to provide an explanation to our research questions, the statistical analysis focused on the two following objectives:

1. compare control native and non-native data for each condition examined in the task; and

2. compare the results on the traditionally associated subject properties among them-selves to see if they develop in a similar way.

VI Results

1 Non-native and native control data

Bearing in mind that the sample size was relatively small (i.e. fifteen informants in the experimental and the control groups) and in order to verify that the appropriate statistical tests were applied, the Kolmogorov–Smirnov test of normality of distribution with the Lilliefors significance correction was applied to all variables. Results indicated that the data are not normally distributed (p < .05), which implies that non-parametric statistical tests had to be applied. Mann–Whitney U test and a Wilcoxon Z test were used in this data analysis and therefore medians and not means are compared, although means will





Table 2 Mann–Whitney U test between experimental and control groups: Rejection/correction of ungrammaticality

Variable Age and L1 n Median (percentage)

Mean (percentage)

Mann–Whitney U

Asympomatic signicficance

Rejection of ungrammaticality verb inflection

Spanish 5-year-olds

15 54.16 45.83 4.500 .010

Control 5-year-olds

15 100.00 87.50

Rejection of ungrammaticalitynull expletives

Spanish 5-year-olds

15 50.00 28.57 .000 .001

Control 5-year-olds

15 100.00c 100.00

Rejection of ungrammaticality null subjects in main clauses

Spanish 5-year-olds

15 50.00 42.85 .000 <.001

Control 5-year-olds

15 100.00c 100.00

Rejection of ungrammaticality null subjects in subordinate clauses

Spanish 5-year-olds

15 50.00 71.42 10.500 .023

Control 5-year-olds

15 100.00c 100.00

Rejection of ungrammaticalitypostverbal subjects

Spanish 5-year-olds

15 33.33 28.47 3.000 .005

Control 5-year-olds

15 66.66 76.18

Rejection of ungrammaticalitythat-trace sequences

Spanish 5-year-olds

15 0.00 14.28 14.500 .131

Control 5-year-olds

15 50.00 50.00

Note: c beside a number indicates that it is a constant value.

also be included in the tables for clarity purposes. The level of significance will be α = .05 all throughout the analysis.

Our first comparative analysis contrasts non-native and native control results of rejec-tion/correction of ungrammatical sentences variables. As Table 2 illustrates, native and non-native results differ significantly (p < .05) in all variables, except in ‘That-trace sequences’. Whereas the control group obtains constant rejection medians of 100.00% in ‘Verb inflection’, ‘Null Expletives’ and ‘Null Subjects in Main and Subordinate Clauses’, non-native percentages are significantly lower. In the case of ‘Postverbal Subjects’, the control median is not so high (i.e. 66.66%) but still significantly higher than that of the experimental group (i.e. 33.33%; U 3.000, p = .005). ‘That-trace sequences’ behave dif-ferently from the other variables and prove rather difficult to correct for both groups, whose difference of results is not significant. The two medians seem to differ notably (i.e. 0.00% in the experimental group and 50% in the control group), although the results from the control group show a remarkable deal of variability ranging from 0.00% to 100.00%, hence causing the non-significant difference.





Table 3 Mann–Whitney U test between experimental and control groups: Acceptance of grammaticality


Mean (percentage)

Mann–Whitney U


Acceptance of grammaticality: verb inflection

Spanish 5-year-olds

15 100.00 95.83 21.000 .317

Control 5-year-olds

15 100.00c 100.00

Acceptance of grammaticality overt expletives

Spanish 5-year-olds

15 100.00c 100.00 24.500 1.000

Control 5-year-olds

15 100.00c 100.00

Acceptance of grammaticality overt subjects

Spanish 5-year-olds

15 100.00 91.66 21.000 .317

Control 5-year-olds

15 100.00c 100.00

Acceptance of grammaticality preverbal subjects

Spanish 5-year-olds

15 100.00c 100.00 24.500 1.000

Control 5-year-olds

15 100.00c 100.00

Acceptance of grammaticality: that-trace sequences

Spanish 5-year-olds

15 0.00 33.33 7.000 .013

Control 5-year-olds

15 100.00c 100.00


Regarding percentages of acceptance of grammaticality variables and as Table 3 indicates, non-native judgements are extremely accurate (i.e. with means of over 90% in all properties except one) and very similar to those of the control group. No signifi-cant differences emerge except in ‘That-trace sequences’, whose grammatical exam-ples still prove too difficult for the experimental group to process but not for the control group.

No hesitations in the children’s judgements and corrections occurred with grammati-cal sentences either in the control or the experimental groups, whereas they were (mini-mally) present in ungrammatical sentences, particularly among the non-native children. As Table 4 shows, experimental children mostly hesitated when judging ungrammatical null subjects in main clauses, where a significant difference emerges with respect to the control group (i.e. U 7.000, p = .008), and presented very low percentages of hesita-tions in the other variables, except in ‘That-trace sequences’ where no hesitations occurred. Control children displayed practically no hesitations in their results, except in ungrammatical ‘Postverbal Subjects’, where they hesitated even more than experi-mental children.





2 Non-native data on subject properties

In order to examine whether the different subject properties tested in the task display similar behaviour in acquisitional terms, rejection/correction of ungrammaticality variables were statistically compared to one another by means of paired comparisons of related samples (Wilcoxon Z test with the Bonferroni correction). Acceptance of grammaticality variables were not compared among themselves since, as seen above, they display similarly accurate results. Table 5 shows that all subject variables, including ‘That-trace sequences’, pattern alike in the experimental group, as the null hypothesis of equal medians is maintained for all variables (i.e. p> .05). An almost significant difference is observed between ‘Null Subjects in Subordinate Clauses’ and ‘Null Expletive Subjects’ both with a median of 50.00% but with a very different distribution of results, as Figure 1 illustrates. The discussion section will next deal with the findings in the study.

Table 4 Mann–Whitney U test between experimental and control groups: Hesitation about ungrammaticality


Mean (percentage)

Mann–Whitney U


Hesitation ungrammaticality verb inflection

Spanish 5-year-olds

15 12.50 12.50 12.500 .085

Control 5-year-olds

15 0.00 5.35

Hesitation ungrammaticalitynull expletives

Spanish 5-year-olds

15 0.00 7.14 21.000 .317

Control 5-year-olds

15 0.00c 0.00

Hesitation ungrammaticalitynull subjects in main clauses

Spanish 5-year-olds

15 50.00 42.85 7.000 .008

Control 5-year-olds

15 0.00c 0.00

Hesitation ungrammaticalitynull subjects in subordinate clauses

Spanish 5-year-olds

15 0.00 14.28 17.500 .141

Control 5-year-olds

15 0.00c 0.00

Hesitation ungrammaticalitypostverbal subjects

Spanish 5-year-olds

15 0.00 19.04 21.500 .674

Control 5-year-olds

15 33.33 23.80

Hesitation ungrammaticality: that-trace sequences

Spanish 5-year-olds

15 0.00c 0.00 24.500 1.000

Control 5-year-olds

15 0.00c 0.00






Table 5 Wilcoxon Z test with Bonferroni correction applied to rejection of ungrammaticality/subject properties variables in the experimental group

Null expletives

Null subjects in main clauses

Null subjects in subordinate clauses

Postverbal subjects

that-trace sequences

Null expletives − Z (–1.414)p = .157

Z (–1.857)p = .063

Z (–.106)p = .916

Z (–1.000)p = .317

Null subjects in main clauses

Z (–1.414)p = .157

− Z (–1.633)p = .102

Z (–1.364)p = .172

Z (–1.633)p = .102

Null subjects in subordinate clauses

Z (–1.857)p = .063

Z (–1.633)p = .102

− Z (–2.379)p = .170

Z (–1.994)p = .460

Postverbal subjects Z (–.106)p = .916

Z (–1.364)p = .172

Z (–2.379)p = .170

− Z (–.687)p = .492

that-trace sequences Z (–1.000)p = .317

Z (–1.633)p = .102

Z (–1.994)p = .460

Z (–.687)p = .492

−

VII Discussion

Having presented the results, this section discusses the data obtained in relation to the key issues addressed in the study, namely children’s awareness of grammaticality contrasts in L2 English and the kind of errors they make, influence of the children’s L1 on their L2

Spanish 5-year-olds

Age and L1

100

80

60

40

20

0

1

5

4

Rejection of ungrammaticality-that-trace sequences

Rejection of ungrammaticality-postverbal subjects

Rejection of ungrammaticality-null subjects in subordinate clauses

Rejection of ungrammaticality-subjects in main clauses

Rejection of ungrammaticality-null expletives

Figure 1 Box-and-whiskers plot of the subject properties variables from the experimental group





performance and L1-independent development, and whether the different subject proper-ties tested show similar results.

The fact that children in this study can and indeed do make judgements of grammati-cal and ungrammatical sentences in their L2 English already shows a certain degree of awareness of linguistic properties in their L2. More importantly, children obtained extremely accurate results in their judgement of grammatical sentences where no sig-nificant differences emerged with respect to the control monolingual group in any vari-ables except for ‘That-trace sequences’, whose grammatical examples are still too difficult to judge by non-native 5-year-olds. Yet their judgements of ungrammatical subject properties were not so accurate and the children’s rejection of ungrammaticality was significantly lower than in the native control group in all variables, except for ‘That-trace sequences’, whose ungrammatical examples are still too difficult to be cor-rectly judged even by native monolingual children. This shows that experimental chil-dren in this study are indeed sensitive to grammaticality contrasts in their L2 English, as their judgements of sentences are clearly not randomly made but respond to the well-formedness of sentences. In fact, some incorrect judgements included hesitations, which indicate the children’s awareness of ungrammaticality but incapacity to positively dis-cern and correct the target ungrammatical element. As specified above, hesitations were analysed as such irrespectively of whether the repetitions were coded as correct or incorrect. Thus, when judging and repeating the sentences, some of the children cor-rectly changed the ungrammatical elements but they remained hesitant about it. This was analysed as correct and also as a hesitation. Others hesitated but did not change anything and others changed the ungrammatical elements but in an incorrect way or even changed other elements which were not originally ungrammatical, thus showing that they are conscious of the fact that something is not appropriate even if they cannot work out exactly what it is. These cases were coded as incorrect or correct and also as hesitations. More specifically, where an omitted subject had to be provided or a postver-bal subject had to be preverbal, if the child made errors in correctly choosing the appro-priate lexical item (i.e. ‘he’ instead of ‘she’ or ‘my mother’ instead of ‘my sister’) these instances were counted as correct, as the child has realized a pronoun is needed there even if s/he cannot figure out which one is needed, since what is being tested is the impossibility of having a null element as a subject (see the examples from Lara, Guillermo and Lisa in (1)). However, if the child provided the wrong verb inflection in sentences testing inflection (i.e. ‘plays’ instead of ‘played’) the item was considered incorrect, as in the inflection sentences what is tested is precisely whether the child is able to provide the correct inflection (see the examples from Maria and Alvaro in (2)). It should be mentioned that sometimes the child correctly changes other elements in the sentence to accommodate the sentence to the type of inflection provided hence showing awareness of grammaticality (see the example from Maria in (2)). If the child provides the right pronoun or makes the lexical subject preverbal and changes the inflection – even if the original one was already correct – the item is considered correct as well, since the target is the pronoun/lexical subject (see the examples from Guillermo in (1) and in (4) and the example from Maria in (4)). Cases where the child had to provide a subject pronoun that were solved using other strategies that render the sentence grammatical were also coded as correct (see the example from Maria in (3)





where she creates coordination so as to allow the null subject pronoun). Some examples of hesitations, changes and correct judgements/repetitions and the way they were coded are included below:

(1) Task sentence (3): My sister is always tired because works a lot. Target sentence: My sister is always tired because she works a lot. Alvaro: My sister is very tired because aaa (hesitation) works a

lot (coded as hesitation and incorrect) Lara: My sister is always tired because he works a lot (coded as correct) Guillermo: My sister is always tired because she’s working a lot (coded as correct) Alexandra: My sister is always tired because works (hesitation)a lot (coded as hesitation and incorrect) Lisa: My sister is tired because it works a lot

(coded as correct)(2) Task sentence (5): My best friend play football twice a week. Target sentence: My best friend plays football twice a week. Alba: My best friend plays football twice a week

(coded as correct) Maria: My best friend played football two weeks ago

(coded as incorrect) Alvaro: My best friend is playing football

(coded as incorrect)(3) Task sentence (7): Elizabeth likes card games. Plays ‘Snap’ every day. Target sentence: Elizabeth likes card games. She plays ‘Snap’ every day. Alvaro: Elizabeth likes card games. She can Snap every day (coded as correct) Alexandra: Elizabeth likes card games and (hesitation) she likes to

play every day(coded as a hesitation and correct)

Jordi: Elizabeth likes card games. She plays (hesitation) Snap every day

(coded as a hesitation and correct) Maria: Elizabeth likes card games and plays Snap every day (coded as correct) Ana: Elizabeth likes card games. She plays Snap all the day (coded as correct)(4) Task sentence (8): Has come my granny from America. Target sentence: My granny has come from America. Guillermo: My granny is from America

(coded as correct)





The children’s capacity to make judgements, even if incorrect, and their use of hesita-tions indicate that the children are not merely relying on imitation or general learning mechanisms but that they are linguistically sensitive and use (at least part of) the internal language mechanisms which are operative in L1A, as was predicted by the assumption that true child L2A grants access to UG and the principles that guide acquisition (Schwartz, 2003, 2004). Age of onset of acquisition of the children under study was three, and considering that, in spite of individual variation, maturational constraints are assumed to start affecting language acquisition around age four (Haznedar and Gavruseva, 2008; Meisel, 2008, 2009; Schwartz, 2004; Unsworth, 2005a, 2005b, among others), the fact that these children are linguistically sensitive is well accounted for. Indeed we will see below that their acquisition of subject properties resembles L1A in some respects and that they seem to know that English is not a prop-drop language, particularly when deal-ing with subjects of subordinate clauses. However, we need to bear in mind that there are still significant differences with respect to the control group in ungrammatical sentences and that these children still present some L1 influence on their L2 performance after two years of exposure to L2 English in an immersion environment and with a low age of onset of acquisition. To what extent this influence is due to structural L2 Transfer or is just due to L2 performance-related effects cannot be shown with the present data.

Focusing on the comparison of results between experimental and control 5-year-old children in their judgement of ungrammatical sentences, we can observe that percentages of correct judgements are significantly lower than those of the control group in all sub-ject variables except for ‘That-trace Sequences’ in which the control group obtained lower and hence more similar results to the experimental group (see Table 2). Results on

Alexandra: Have come (hesitation) my granny from America (coded as a hesitation and incorrect) Maria: My granny comes from America

(coded as correct)(5) Task sentence

(26):The children in kindergarden go to the gym yesterday morning.

Target sentence: The children in kindergarden went to the gym yesterday morning.

Alvaro: The children in kindergarden go (hesitation) to the gym yesterday morning

(coded as a hesitation and incorrect) Jordi: The class of kindergarden (hesitation) go (hesitation) to

the gym yesterday(coded as a hesitation and incorrect)

Lisa: The children of kindergarden (hesitation) they going to gym

(coded as a hesitation and incorrect) Martín: The children of kindergarden go yesterday in the gym (coded as incorrect)





‘That-trace sequences’ in both the experimental and control groups were much lower than the other subject variables, thus showing the different developmental path of this property with respect to the other properties.5 Results on ‘Verb Inflection’ were also significantly lower than those of the control group, thus keeping a parallel development to subject variables. After two years of being exposed to L2 English in an immersion context, the children under study allow a statistically significant percentage of mainly null expletive and postverbal subjects and also (but to a lesser extent) null referential subjects in main and subordinate clauses, which are the L1 properties related to the strong pronominal [+interpretable] Agreement features on the verb of the children’s L1 Spanish (as outlined in Platzack, 2004; see also Section II). The influence of the L1 was initially expected to be minimal given that these children have already been exposed to L2 English for two years in an immersion setting, which means that they are not in their initial stages of acquisition any longer, and given that their age of onset of acquisition was three, which means that in theory the first critical period (i.e. around age four) has not started and the children’s L2 should not be so much affected by their L1. In fact, a number of studies have recently explored the relationship between L1 Transfer and age of onset, pointing out that L1 Transfer increases as age of first exposure is delayed (approaching puberty) (Ionin, 2008), and that L1 Transfer is clearly more present in adult L2A than in child L2A (Blom, 2008). In Lakshmanan and Selinker (1994), the development of the complementizers that in two child L2 English learners (one having Spanish as his L1) and two Spanish-speaking adolescent L2 English learners is explored and compared and results suggest that transfer effects are overridden in the case of the child L2 learners but not in the case of the adolescents. It seems to be the case that children in the present study, and as far as subject properties are concerned, are somehow affected by their L1 and thus differ from L1A. However, another factor needs to be taken into account when addressing the presence of L1 influence, which is the extent to which the L1 and the L2 are used in the context under analysis. Although the children were exposed to L2 English in an immersion context, they continued using their L1 at home and in the com-munity, and even at school in some occasions and hence the influence of their L1 is inevitably greater than in other studies where the L2 is the language of the community. Given that the children prove to be linguistically sensitive, that their age of onset of acquisition is three and that they do not seem to be treating all English sentences as if they were Spanish (especially the subjects of embedded clauses) it would be too strong a claim to state that L1 structural Transfer is at work here. As there is not enough evidence to postulate L1 Transfer, the present results could also be explained through L1 influence on the children’s L2 performance, which results from the fact that they live in a Spanish/Catalan community and are hence inevitably influenced by their L1. This ‘borrowing’ (Corder, 1992, cited in Lakshmanan and Selinker, 2001), is a ‘communicative strategy during which speakers use certain aspects of their mother tongue to ‘express meanings’, on-line as it were, because their interlanguage simply lacks the means to do it’ (Lakshmanan and Selinker, 2001) and is a characteristic of language use which does not imply that children are affected by structural transfer.

A further objective of the study was to determine whether all subject properties tradi-tionally associated with the Null Subject Parameter develop in a similar way, although this cannot give us evidence for clustering effects. In other words, we analysed whether all





subject properties display similar results in acquisitional terms, that is to say, if the chil-dren’s results on the different variables are significantly different from each other or not. As mentioned before, the study of the whole cluster of subject properties has not been carried out in child L2 English and was done for adult L2 English, determining that there is no cluster, but without statistical analysis (Tsimpli and Roussou, 1991; White, 1985). Statistical analysis is provided here but no causation can be derived from it, so clustering cannot be posited. As seen in Section V, the null hypothesis of equal medians could not be rejected for any variables, which implies that the children’s results in all subject variables are not significantly different from each other, which, together with the fact that they sig-nificantly diverge from the control group and considering the theoretical syntactic frame-work adopted here (Platzack, 2004) and its predictions, indicate that at least null and postverbal subjects develop similarly. Contrary to what was expected, results of ‘That-trace Sequences’ were also non-significantly different from those of the other properties, given the variability in the results of the other variables. Yet the fact that their percentages are so low compared to the other variables and that they remain similarly problematic in grammatical and ungrammatical sentences and in ungrammatical sentences even for the control native children sets them apart from the occurrence of missing and postverbal subjects. Although clustering cannot be statistically demonstrated here, the notion of clus-ter can be at least theoretically maintained for the subject properties if we assume that missing and postverbal subjects6 are derived from the pronominal/non-pronominal verbal agreement feature and hence the way the EPP is checked (Platzack, 2004). As for ‘That-trace sequences’, although they were long associated with the rest of subject properties and in fact their occurrence or non-occurrence coincides with whether languages accept null and postverbal subjects or not, they have been recently given different syntactic accounts (see Section II; see also Gallego, 2006; Pesetsky and Torrego, 2001, 2004), which clearly set them apart from the other subject properties.

As has already been mentioned, ‘Postverbal Subjects’ and ‘Null Expletives’ obtained lower percentages than ‘Null Subjects’, especially ‘Null Subjects in Subordinate Clauses’, namely mean values of 28.47% and 28.57%, respectively versus 42.85% and 71.42% (see Table 2). This implies that children in this study found it more difficult to judge and cor-rect ungrammatical null expletives and postverbal subjects than missing referential sub-jects. The fact that L1 Spanish children accepted more ungrammatical null expletives than null referential subjects in L2 English confirms the results from Phinney (1987) and Tsimpli and Roussou (1991), although these studies examined adult L2 learners and used different methodologies and hence are hardly comparable. These results also confirm Lakshmanan’s (1991, 1994) longitudinal study on child L2 English, whose L1 Spanish-speaking child, Marta, produced significantly more null expletives than null referential subjects, although they only involved it is contexts. Unfortunately, no comparable cross-sectional (Spanish) child L2 English studies are reported in the literature as far as we know. The low rejection of ungrammatical postverbal subjects observed in this study con-trasts to their low acceptance in White (1985) and rejection in Tsimpli and Roussou (1991), although again these studies deal with adult L2 speakers of English.

Some of the results presented set a parallelism to L1A and might well be explained through developmental effects and not only through L1 influence. Since the two groups under study have not had the same exposure to English, as the control group was exposed





to English from birth, it is plausible that the L2 children are going through the same devel-opmental stages as the L1 children but only some years later. Apart from a similar behav-iour with respect to that-trace effects, which can also be the result of processing difficulties, there are two other aspects which seem parallel to L1A, namely the correctly high rejec-tion of ‘Null Subjects in Subordinate Clauses’ in comparison to the other variables and the high percentage of hesitations in ‘Null Subjects in Main Clauses’. Null subjects in early child language display a different distribution and acquisition pattern in the L1 acquisition of null subject languages and in non-null subject languages (Rizzi, 2002, 2005) and this might indicate the possible existence of two subject UG parameters, namely the tradi-tional Null Subject Parameter and the Root Subject Drop Parameter. The former is quickly fixed on the target value whereas the latter is delayed with respect to the negative value.7 In other words, children rapidly acquire the setting of the Null Subject Parameter. Children acquiring null subject languages like Italian or Spanish produce null subjects in all syntac-tic environments from very early on (Bel, 2001; Guasti, 1996; Rizzi, 1994, 2000) and children acquiring a non-null subject language, such as English or French, never allow null subjects in non-initial positions (i.e. subordinate clauses or wh-environments) but continue dropping subjects in the specifier position of the root for a much longer period (Clahsen et al., 1996; Haegeman, 1996a, 1996b; Rizzi, 2000; Roeper and Rohrbacher, 1995). Children in this study reject many more null subjects in subordinate clauses than in root main clauses and null expletive subjects also in root positions and present a signifi-cant level of hesitation in their judgements of null subjects in main clauses in their L2 English as L1 English children do in early child English. Yet this occurs much later than in L1A, given that we are obviously dealing with a different process of acquisition in L2A, which is mainly affected by the children’s L1 and other external factors inevitably leading to lack of uniformity and slower rate of acquisition. Interestingly enough, in her study of child L2A acquisition of Spanish subject properties, Pladevall Ballester (2010) found that L1 English 5-year-old children acquiring L2 Spanish found it hard to correctly judge null/overt subjects in subordinate clauses in L2 Spanish and obtained much more accurate results in judging null/overt subjects in root positions.8 In order to test whether the differences between L1 and L2 children are due to L1 Transfer or developmental effects a comparison between L2 English children from a null subject L1 language and a non-null subject L1 language would have to be made, which is beyond the aims of the present study. Another possibility would be to control amount of exposure between L1 and L2 children. However, this would imply cognitive differences which might clearly influence the results and the issue of age of onset would not be at play.

A last word should be mentioned about results on the inflection variable, which as mentioned above, are also significantly lower than those of the control native group. L2 English children significantly accepted non-inflected third person singular present tense or regular and irregular past tense verb forms.9 Yet we do not know at this point whether L1 and L2 children follow the same developmental patterns with the only dif-ference that L2 children are three years (i.e. age of onset) behind L1 children or if their process of L2 acquisition of inflectional verbal properties certainly follows a different path from child L1A, although their age of onset was as early as three and contrary to Schwartz (2003, 2004) and the Domain-by-age model where in the domain of inflec-tion, child L2A is claimed to be similar to L1A. Positing different paths is in line with





Meisel (2008, 2009), where the author corroborates that the acquisition of verb mor-phology in L2A children with ages of onset between ages three and four differ signifi-cantly from L1A children.

VIII Concluding remarks

This study contributes new cross-sectional data on child L2A of English subject proper-ties as compared to a monolingual control group in an immersion bilingual setting. All in all, the findings suggest that 5-year-old children show sensitivity to grammatical/ungrammatical contrasts in L2 English subject properties, which does not mean that their judgements are native-like, and that results exhibit L1 influence on the children’s L2 performance for the subject properties traditionally associated as the Null Subject Parameter (except inflectional development and that-trace sequences), although results cannot really provide evidence for structural L1 Transfer.

Parallel development to L1A in the children’s acquisition of subjects in subordinate and main clauses has been observed and it has also been acknowledged that results may also be due to the effects of development and to the fact that L1 children have had more exposure than L2 children at the time of testing and may have already gone through the same stages. That-trace sequences were shown to develop differently from the other prop-erties, as expected, and to present many more difficulties of understanding and acquisition even for the control monolingual group.

In fact, considering that the children under study had been exposed to L2 English for two years in an immersion setting and that their age of onset was three, we would have expected less L1 influence and more accurate results in the ungrammatical sentences. This adds evidence to the distinction between the L1A and L2A processes even in the case of child L2A and that, as Meisel (2009) posits:

evidence strongly supports the claim that fundamental differences are likely to emerge already in early childhood and that the FDH (Fundamental Difference Hypothesis, Bley-Vroman, 1990) can therefore be assumed to apply to cL2 as well as to aL2 acquisition. (p. 13)

We would like to point out, however, that the fact that L1A and child L2A are different processes does not entail that child L2A and adult L2A are alike. Maturational constraints do exist in child L2A but not so strongly as in adult L2A, and we still assume that, at least partly, child L2 language is constrained by internal universal mechanisms that are clearly not available in adult L2A any more and that this is what makes children’s linguistic sen-sitivity to grammaticality and ungrammaticality possible and what makes successful ulti-mate attainment not inevitable, but possible, contrary to adult L2A. Further research should therefore compare data from this study to later age groups whose age of onset of acquisition was the same. This would enable us to see whether successful ultimate attain-ment of the specific constructions tested with age of first exposure being three is possible. Other possibilities of future research could involve a comparison with older learners with the same length of exposure or children of the same age with different lengths of exposure in order to test whether it is age and cognitive maturity or length of exposure that facilitates acquisition.





Notes

1. Similar minimalist accounts of properties of the Null Subject Parameter include, amongst others, Alexiadou and Anagnostopoulou, 1998; Contreras, 1991; Kato, 1999; Manzini and Roussou, 2000; Manzini and Savoia, 2002; Ordóñez and Treviño, 1999).

2. As an anonymous reviewer pointed out the null expletive item with the verb seem in sentence (28) might pose problems in that it is actually quite accepted in spoken English. However, the L1 English children correctly rejected the null expletive and provided ‘it’ in 100% of the cases.

3. Children at Benjamin Franklin International School (BFIS) at the primary school level (i.e. from 1st grade onwards) also attend Catalan classes if they wish to and 12th grade students may choose whether to get preparation for Spanish or American university entrance exams.

4. The fact that the school was American whereas, as will be seen in the next sections, data from control groups and from adults come from British English was not taken into account since the grammatical subject properties do not differ in both varieties. Studies on child L2 English acquisition of subject properties have so far generally been conducted in English-speaking communities.

5. Previous research has shown that that-trace effects in English display dialectal differences (Rizzi, 1990), which could have played a role in the low results obtained by the control group. Had grammatical sentences involving wh-extraction from the object position been included in the study, we could have had a much clearer indication as to whether these constructions are too difficult to process or dialectal differences were at work here.

6. An anonymous reviewer pointed to the fact that pragmatic factors, such as given vs. new infor-mation can make VS orders possible in English (e.g. In the corner of the room is a desk) as well as the lexically restricted V2 constraint that appears in quotatives (e.g. ‘No!’, cried the wizard). This could create contradictory or misleading data for the language learner. Yet the present study focuses on strictly syntactic structures without taking pragmatic-syntactic phenomena into account.

7. Rizzi (2002, 2005) argues that this extends to other linguistic properties and he proposes two groups of parameters, namely those which do not give rise to parametric discontinuity that can be observed in production and which are fixed according to Wexler’s (1998) Very Early Parameter Setting (VEPS) and those which give rise to parametric discontinuity (i.e. target inconsistencies observable in the data). The Null Subject Parameter would belong to the first group, whereas the Root Subject Drop Parameter would belong to the second group.

8. Directionality of acquisition differences are clearly beyond the scope of the present article and will be pursued in future research.

9. As an anonymous reviewer indicated, failure to supply the target inflection could have been due to phonological difficulties in pronouncing final consonant clusters (although this would have been more plausible in pronouncing final -ed morphemes than final -s or irregular past tense forms). The fact that the task was designed to be carried out orally entails this possibility. Yet the possibil-ity of doing such a task in a written form with 5-year-old children was obviously ruled out.

References

Alexiadou A and Anagnostopoulou E (1998) Parametrizing AGR: Word order, V-movement and EPP checking. Natural Language and Linguistic Theory 16: 491–540.

Bel A (2001) Sujetos nulos y sujetos explícitos en las gramáticas iniciales del castellano y el cata-lán [Null subjects and explicit subjects in the early grammars of Spanish and Catalan]. Revista Española de Lingüística 31: 537–62.





Bley-Vroman R (1990) The logical problem of foreign language learning. Linguistic Analysis 20: 3–49.

Blom E (2008) Testing the domain-by-age model: Inflection and placement of Dutch verbs. In: Haznedar B and Gavruseva H (eds) Current trends in child second language acquisition: A generative perspective. Amsterdam: John Benjamins, 271–300.

Chomsky N (1981) Lectures on government and binding. Dordrecht: Foris Publications.Clahsen H, Kursawe C, and Penke M (1996) Introducing CP: wh-questions and subordi-

nate clauses in German child language. In: Koster C and Wijnen F (eds) Proceedings of the Groningen assembly on language acquisition. Groningen: Center for Language and Cognition, 5–22.

Contreras H (1991) On the position of subjects. Syntax and Semantics 25: 63–79.Corder P (1992) A role for the mother tongue. In Gass S and Selinker L (eds) Language transfer

in language learning. Amsterdam: John Benjamins, 18–31.DeKeyser RM (2000) The robustness of critical period effects in second language acquisition.

Studies in Second Language Acquisition 22: 499–533.Gallego AJ (2006) Phase effects in Iberian romance. Cuadernos de Lingüística del Instituto

Universitario Ortega y Gasset 13: 59–78.Guasti MT (1996) The acquisition of Italian interrogatives. In: Clahsen H (ed.) Generative per-

spectives on language acquisition. Amsterdam: John Benjamins, 241–70.Guasti MT (2002) Language acquisition: The growth of grammar. Cambridge, MA/London: The

MIT Press.Haegeman L (1996a) Root infinitives, clitics and truncated structure. In: Clahsen H (ed.) Generative

perspectives on language acquisition. Amsterdam: Benjamins, 271–308.Haegeman L (1996b) Non overt subjects in the core grammars of English and French. Unpublished

manuscript, University of Geneva, Switzerland.Haznedar B and Gavruseva H (2008) Recent perspectives in child second language acquisition.

In: Haznedar B and Gavruseva H (eds) Current trends in child second language acquisition: A generative perspective. Amsterdam: John Benjamins, 3–15.

Hilles S (1991) Access to Universal Grammar in second language acquisition. In: Eubank L (ed.) Point counterpoint. Universal Grammar in the second language. Amsterdam: John Benjamins, 305–38.

Hyams N (1986) Language acquisition and the theory of parameters. Dordrecht: Reidel.Hyams N (1989) The null subject parameter in language acquisition. In: Jaeggli O and Safir KJ

(eds) The null subject parameter. Dordrecht: Kluwer Academic, 215–38.Hyams N (1992) A reanalysis of null subjects in child language. In: Weissenborn J et al. (eds)

Theoretical issues in language acquisition. New Jersey: Lawrence Erlbaum, 249–67.Hyams N (1996) The underspecification of functional categories in early grammar. In: Clahsen H

(ed.) Generative perspectives on language acquisition: Empirical findings, theoretical consid-erations, crosslinguistic comparisons. Amsterdam: John Benjamins, 91–127.

Ionin T (2008) Progressive aspect in child L2-English. In: Haznedar B and Gavruseva H (eds) Current trends in child second language acquisition: A generative perspective. Amsterdam: John Benjamins, 17–55.

Jaeggli O (1982) Topics in romance syntax. Dordrecht: Foris.Jaeggli O and Safir K (1989) The null subject parameter and parametric theory. In: Jaeggli O and

Safir K (eds) The null subject parameter. Dordrecht: Kluwer, 1–44.Kato MA (1999) Strong and weak pronominals in the null subject parameter. Probus 11: 1–37.Lakshmanan U (1991) Morphological uniformity and null subjects in child second language acqui-

sition. In: Eubank L (ed.) Point counterpoint: Universal Grammar in the second language. Amsterdam: John Benjamins, 389–410.

Lakshmanan U (1994) Universal Grammar in child second language acquisition. Amsterdam: John Benjamins.





Lakshmanan U (1995) Child second language acquisition of syntax. Studies in Second Language Acquisition 17: 301–29.

Lakshmanan U (2009) Child second language acquisition. In: Ritchie W and Bhatia T (eds) The new handbook of second language acquisition. Bingley: Emerald, 377–400.

Lakshmanan U and Selinker L (1994) The status of CP and the tensed complementizer that in the developing L2 grammars of English. Second Language Research 10: 25–48.

Lakshmanan U and Selinker L (2001) Analysing interlanguage: How do we know what learners know? Second Language Research 17: 393–420.

Manzini MR and Roussou A (2000) A minimalist theory of A-movement and control. Lingua 110: 409–47.

Manzini MR and Savoia L (2002) Parameters of subject inflection in Italian dialects. In: Svenonius P (ed.) Subjects, expletives and the EPP. Oxford: Oxford University Press, 157–200.

McLaughin B (1978) Second language acquisition in childhood. Hillsdale, NJ: Erlbaum.Meisel J (2007) Age of onset in successive acquisition of bilingualism: Effects on grammatical

development. Working Papers in Multilingualism 80: 3–32. Hamburg: University of Hamburg.Meisel J (2008) Child second language acquisition or successive first language acquisition? In:

Haznedar B and Gavruseva H (eds) Current trends in child second language acquisition: A generative perspective. Amsterdam: John Benjamins, 55–80.

Meisel J (2009) Second language acquisition in early childhood. Zeitschrift für Sprachwissenschaft 28: 5–34.

Mobaraki M, Vainikka A, and Young-Scholten M (2008) The status of subjects in child L2 English. In: Haznedar B and Gavruseva H (eds) Current trends in child second language acquisition: A generative perspective. Amsterdam: John Benjamins, 55–80.

Ordóñez F and Treviño E (1999) Left dislocated subjects and the pro-drop parameter: A case study of Spanish. Lingua 107: 39–68.

Park H (2004) A minimalist approach to null subjects and objects in second language acquisition. Second Language Research 20: 1–32.

Pesetsky D and Torrego E (2001) T-to-C movement: Causes and consequences. In: Kenstowicz M (ed.) Ken Hale: A life in language. Cambridge, MA: MIT Press, 355–26.

Pesetsky D and Torrego E (2004) Tense, case, and the nature of syntactic categories. In: Guéron J and Lecarme J (eds) The syntax of time. Cambridge, MA: MIT Press, 495–538.

Phinney M (1987) The pro-drop parameter in second language acquisition. In: Roeper T and Williams E (eds) Parameter setting. Dordrecht: Reidel, 221–38.

Pladevall Ballester E (2010) Child L2 development of syntactic and discourse properties of Spanish subjects. Bilingualism: Language and Cognition 13: 185–216.

Platzack C (2004) Agreement and the person phrase hypothesis. Working Papers in Scandinavian Syntax 73: 83–112.

Radford A (1990) Syntactic theory and the acquisition of English syntax. Oxford: Basil Blackwell.Radford A (1992) The acquisition of the morphosyntax of finite verbs in English. In: Meisel JM

(ed.) The acquisition of verb placement: functional categories and V2 phenomena in language acquisition. Dordrecht: Kluwer Academic, 23–62.

Ritter E (1995) On the syntactic category of pronouns and agreement. Natural Language and Linguistic Theory 13: 405–43.

Rizzi L (1982) Issues in Italian syntax. Dordrecht: Foris.Rizzi L (1986) Null objects in Italian and the theory of pro. Linguistic Inquiry 17: 501–57.Rizzi L (1990) Relativized minimality. Cambridge, MA: MIT Press.Rizzi L (1994a) Some notes on linguistic theory and language development: The case of root

infinitives. Language Acquisition 3: 371–93.





Rizzi L (1994b) Early null subjects and root null subjects. In: Hoekstra T and Schwartz BD (eds) Language acquisition studies in generative grammar. Amsterdam: John Benjamins, 151–76.

Rizzi L (2000) Remarks on early null subjects. In: Friedemann M-A and Rizzi L (eds) The acquisi-tion of syntax: Studies in comparative developmental linguistics. London: Longman, 269–92.

Rizzi L (2002) On the grammatical basis of language development: A case study. Unpublished manuscript, University of Siena, Italy.

Rizzi L (2005) Grammatically-based target-inconsistencies in child language. In: Deen K, Nomujra J, Schultz B, and Schwartz BD (eds) Proceedings of the inaugural conference on Generative Approaches to Language Acquisition - North America (GALANA). Cambridge, MA, CONN/MIT Working Papers in Linguistics, 19–49.

Roeper T and Rohrbacher B (1995) Null subjects in early child English and the theory of economy of projection. University of Pennsylvania Working Papers in Linguistics 2: 83–119.

Schwartz BD (1992) Testing between UG-based and problem-solving models of L2A: Developmental sequence data. Language Acquisition 2: 1–19.

Schwartz BD (2003) Child L2 acquisition: Paving the way. In: Beachley B, Brown A, and Conlin F (eds) Proceedings of the 27th annual Boston University Conference of Language Development (BUCLD). Somerville, MA: Cascadilla Press, 26–50.

Schwartz BD (2004) Why child L2 acquisition? In: van Kampen J and Baauw S (eds) Proceedings of GALA 2003: Volume 1: LOT Occasional Series 3. Utrecht: LOT Publications, 47–66.

Tsimpli I-M and Roussou A (1991) Parameter resetting in L2? University College London Working Papers in Linguistics 3: 149–69.

Unsworth S (2005a) Child L1, child L2 and adult L2 acquisition: Differences and similarities. In: Brugos A, Micciulla L, and Smith CE (eds) Proceedings of the 28th annual Boston University Conference on Language Development (BUCLD). Somerville, MA: Cascadilla Press, 633–44.

Unsworth S (2005b) Child L2, adult L2, child L1: Differences and similarities: A study on the acquisition of direct object scrambling in Dutch. Unpublished PhD thesis, Utrecht University, The Netherlands.

Wakabayashi S (2002) The acquisition of non-null subjects in English: A minimalist account. Second Language Research 18: 28–71.

Wexler K (1998) Very early parameter setting and the unique checking constraint: A new explana-tion for the optional infinitive stage. Lingua 106: 23–79.

White L (1985) The ‘pro-drop’ parameter in adult second language acquisition. Language Learning 35: 47–62.

White L (1986) Implications of parametric variation for adult second language acquisition: An investigation of the ‘pro-drop’ parameter. In: Cook V (ed.) Experimental approaches to sec-ond language acquisition. Oxford: Pergamon Press, 55–72.

Whong-Barr M and Schwartz BD (2002) Morphological and syntactic transfer in child L2 acquisi-tion of the English dative alternation. Studies in Second Language Acquisition 24: 579–616.




Second Language Research 2012 Ballester 217 41

Documents

Transcript of Second Language Research 2012 Ballester 217 41