Bilingualism across the lifespan: Word Order in
Dutch-Spanish
bilingual children and teenagers
U
NIVERSITEIT VAN
A
MSTERDAM
Thesis Research Master Linguistics June 2017
Supervision: Brechje van Osch
Aafke Hulk Suzanne Aalberse
Name: Elisabet García González
Student
number: 10404503
Abstract
: This exploratory study investigates the knowledge of pre- and post-verbal subjects in children and teenagers acquiring heritage Spanish in the Netherlands. A comparison between bilingual children and adult heritage speakers of Spanish is also established An oral preference judgment task was used to test word order preferences in relation to the factors of verb type, focus and definiteness of the subject. The analysis reveals that children between the ages of 9 and 13 years old accept both pre- and post-verbal subjects in Spanish, but we find a trend towards VS that increases over time. Our results confirm that these bilingual children are sensitive to the distinction of verb types and definiteness. In contrast to monolingual peers, focus does not influence word order preferences at either age. This finding suggests that older bilingual children and teenagers do not entirely resemble monolingual age matched children in the acquisition of subjects in Spanish, nor do they assimilate adult heritage speakers.1. INTRODUCTION ... 3
2. BACKGROUND ... 4
2.1OVERT SUBJECTS IN SPANISH ... 4
2.1.1 Syntax-semantics interface: Verb type ... 5
2.1.2Syntax-discourse interface: focus and definiteness ... 5
2.2 PREVIOUS RESEARCH ... 7
2.2.1 Word order in Spanish monolingual speakers ... 7
2.2.2 Word order in heritage Spanish ... 8
3. THE PRESENT STUDY ... 10
4. METHOD ... 11
4.1PARTICIPANTS ... 11
4.1.1-Spanish bilingual (heritage) ... 11
4.1.2 Spanish monolingual (control) ... 12
4.2STIMULI: ... 13
4.3PROCEDURE: ... 14
4.4ANALYSIS: ... 15
5. RESULTS ... 15
5.1WORD ORDER PREFERENCES (RQ1) ... 16
5.2VERB TYPE, FOCUS AND DEFINITENESS (RQ2) ... 16
5.2.1 Children age 9 ... 17
5.2.2. Children age 13... 19
5.3INTERIM SUMMARY ... 22
6. DISCUSSION... 23
6.1WORD ORDER PREFERENCES OVERALL ... 24
6.2ACQUISITION OF FACTORS ... 25
7. CONCLUDING REMARKS: LIMITATIONS OF THE STUDY AND SUGGESTIONS FOR FURTHER RESEARCH ... 28
ACKNOWLEDGEMENTS ... 29
1. Introduction
Over the last two decades, linguistic research on heritage speakers has generally focused on adult bilinguals. This subtype of bilinguals acquired a language, usually at home, different to the mainstream language of the society (Rothman, 2009). Nevertheless, children acquiring a language in this context can be categorized as heritage speakers, too. By considering them as two different groups, research sometimes overlooks the possibility that childhood bilinguals are the earlier stage of future heritage speakers, and a connection between the two has rarely been investigated (for an exception see Polinsky, 2016; Montrul, 2016).
A major reason to separate these two populations origins in opposite findings. While research on bilingual children generally attests convergence with monolinguals peers1 (De Houwer, 1990; Meisel, 1994), studies on adult bilingual speakers often indicate divergence (Silva-Corvalán, 1994; Montrul, 2004). Nevertheless, an explanation for this disparity might be the differences in input in the heritage language along the stages of the bilingual development. During the first years of life, bilingual children may receive balanced input in both languages, or may be exposed to one of the languages to a larger extent. Regardless of the circumstances, the use of the heritage language often decreases once they enter school and are formally educated in the mainstream language (Rothman, 2009). In some cases, this drop in input leads to language attrition, the loss or degradation of linguistic structures due to the lack of use or transfer from the dominant language2 (Polinsky, 2011).
By studying the acquisition of subject position in Spanish, we hope to find how the heritage language of bilingual children and adolescents in The Netherlands changes over time. Furthermore, we draw a connection with previous research on 5-year-old (García Gonález et al., in prep) and adult heritage speakers (van Osch et al., in prep) of Spanish in this country. In doing so, we aim to explore word order from the early to the late stages of its development in bilinguals.
According to linguistic theory, the position of subjects in intransitive verbs in
1 While bilingual children may undergo delay on the acquisition of certain phenomena, it is widely
accepted that bilingual children follow the development of their monolingual peers (see De Houwer, 1990; Meisel, 1989 for parallelisms in the acquisition of syntax) or difference (Blom et al., 2008, for the acquisition of Dutch grammatical gender by Moroccan children)
2 Polinsky (2011) shows examples of language attrition in the use of relative clauses by heritage speakers
Spanish is rather flexible, and is influenced by multiple factors (Domínguez, 2013). Verb type and focus are two factors that have been described comprehensively. Combined, they determine whether the subject will appear in pre-or post-verbal position (henceforth SV and VS). Unfortunately, experimental studies that have investigated the acquisition of this phenomenon cannot clearly establish how exactly word order in speakers is determined (Montrul, 2004; Zapata, 2005), since other factors that play a role such as definiteness, animacy or tense are not always taken into account. Such studies have followed the perspective of the Interface Hypothesis (IH) (Sorace 2000; Tsimpli et al., 2004; Sorace & Filiaci, 2006; Iverson et al, 2008; Montrul et al., 2008; 2009), which predicts sequential acquisition of each factor according to their level of complexity. Previous research, however, has mostly focused on adults. Therefore, the present study is motivated by the need to enrich our knowledge on bilingual language development and, broadly, aims to answer the following research question:
What factors determine word order in bilingual children, and when are they acquired?
The paper presents previous literature on the theory of word order in Spanish and reports on findings from experimental data in bilingual acquisition. In the next section, we first describe the topic from a theoretical perspective, and summarize previous acquisition research in different language and age groups. Following, we outline our two research questions in Section 3. A description of materials, participants and procedure of the experiment is given in Section 4. Section 5 and 6 include the results of the experiment and our interpretation of the findings, respectively. We conclude the paper with an outline of relevant findings and suggestions for future research.
2. Background
2. 1 Overt subjects in Spanish
Subject position in intransitive clauses in Spanish is influenced by syntactic, prosodic and pragmatic conditions (Dominguez, 2013). Due to the internal (syntax and semantics) and external (syntax and discourse) domains that relate to the different factors involved, the topic has been widely studied within the generative framework of the Interface Hypothesis (IH) (Tsimpli et al., 2004; Sorace & Filiaci, 2006; Iverson et
al, 2008; Montrul et al., 2008; 2009). In the following sections we describe the interface nature of verb type, focus and definiteness, according to the IH.
2.1.1 Syntax-semantics interface: Verb type
The unaccusative or unergative class of the verb influences whether the subject of intransitive clauses appears in SV or VS position: while unaccusative verbs relate to VS orders, as example 1a shows, unergative verbs are more frequent in SV position (example 1b). The distinction between unaccusative and unergative is syntactic as well as semantic (Perlmutter, 1978; Levin & Rappaport Hovav, 1995; Sorace, 2000). Syntactically, unaccusative verbs have an internal argument whereas unergative verbs have an external argument. Semantically, each argument type is associated with distinct characteristics: whereas internal arguments represent patient, non-volitional roles; external arguments symbolize agentive, volitional and willed roles3.
1 a) Vino Juan (UA - VS)
‘Came Juan’
1 b) María lloró (UG - SV)
‘María cried’
Since the unaccusative/unergative distinction is semantically as well as syntactically encoded, verb type is a factor that exhibits an interface between semantics and syntax. The IH establishes predictions in relation to the acquisition of linguistic phenomena that integrate two linguistic domains: phenomena at the syntax-semantics interface entail robust linguistic judgments in both monolingual as well as bilingual speakers (Sorace & Serratrice, 2009). Furthermore, L1 attrition is rare in topics within internal interfaces. Thus according to the IH, verb type in Spanish is expected to be a robust factor that crucially influences word order. Nevertheless, there are other interfaces that entail an even higher level of complexity, which we describe in the next section.
2.1.2 Syntax-discourse interface: focus and definiteness
3 This phenomenon is known as split-intransitivity and is attested cross-linguistically, where the syntactic
behavior of intransitive verbs is language-specific (Levin & Rappaport Hovav, 1999; Sorace, 2000) (see Sorace, 2000, for auxiliary selection in Romance and Germanic languages in intransitive verbs. (Levin & Rappaport Hovav, 1999; Sorace, 2000).
In combination with predicate type, the focus context in which an intransitive verb is used influences the position of the subject (Domínguez, 2013). Informational focus can consist of new as well as given information (Roggia, 2011). Focus is distinguished into broad, which introduces an entire sentence or phrase (2a), and narrow, which highlights the subject4 (2b): broad and narrow focus prompt SV and VS, respectively.
2 a) ¿Qué pasó? ‘What happened?’ María cantó (SV) ‘María sang’ b) ¿Quién cantó? ‘Who sang?’ Cantó María (VS) ‘Sang María’
It is assumed that when predicate type and focus interact, focus overrides the predicate type distinction: VS is preferred in narrow focus regardless of the unaccusative or unergative status of the verb as shown in Table 1 (Hertel, 2003; Lozano, 2006; Dominguez, 2013; Montrul, 2004). interface External Internal interface Narrow focus ‘Who V-ed? Broad focus ‘What happened? Cantó María (VS) Sang María María cantó (SV) María sang Unergative predicates Llegó María (VS) Arrived María Llegó María (VS) Arrived María Unaccusative predicates
Table 1. Expected word order preferences, adapted from Dominguez (2013: 125)
Focus is a factor that integrates syntax, a linguistic domain, and discourse, a
4For the purpose of the topic in question we refer to narrow focus in cases in which the ‘focused’
constituent is the subject. However, any other constituent, e.g., the object, can be highlighted in the narrow context.
linguistic domain, since word order alternates to satisfy specific discourse interpretations. Following the IH, the integration of these two domains requires greater processing demands for bilingual speakers, which ultimately delays the acquisition of syntax-discourse interface phenomena (Sorace & Serratrice 2009).
Apart from verb type and focus, the definiteness of the subject also influences word order. In addition to focus, definiteness concerns the syntax-discourse interface. When a subject is newly presented to the discourse, it is likely to appear in indefinite form, whereas known information (known to both speaker and hearer) appears in the form of definite NPs or proper names (López Meirama, 1997): SV is more frequent with definite subjects (3a), and VS is commonly preferred with indefinite subjects (3b) (Roggia 2011, Tsimpli et al., 2004).
3 a) El amigo de María vino a la fiesta
‘Maria’s friend came to the party’
b) Vino un amigo de María ‘A friend of Maria’s came’
The relation between definiteness and word order is not exclusive to Spanish nor to intransitive verbs and subjects. In other Romance languages such as Italian or Portuguese, definiteness influences the word order of objects and other verbs classes (Sheehan, 2004). This relationship, however, has rarely been addressed in the literature on the acquisition of Spanish word order (for an exception see Roggia, 2011; van Osh et al., in prep). In the following sections, we present experimental and corpus studies in Spanish monolingual and bilingual speakers, and show how verb type, focus and definiteness influence order preferences in each population.
2.2 Previous research
2.2.1 Word order in Spanish monolingual speakers
The literature on word order describes a relationship between verb type and focus: in narrow focus VS is preferred regardless of verb type (Hertel, 2003; Lozano 2006; Dominguez, 2013). Nevertheless, experimental studies do not always find such an interaction, but main effects of verb type and focus instead (van Osch, et al., in prep). A
common finding is that whenever focus is included as a factor, monolingual speakers’ preferences for word order vary to a larger extent than when only verb type is considered (Roggia, 2011; Zapata et al., 2005). Although the evidence showing the influence of definiteness on word order is scarce, its effect seems to parallel with that of focus (Roggia, 2011).
Research on the acquisition of Spanish word order consists of few studies on corpus data of three children, who are sensitive to verb type distinctions at the early age of 2 (Grinstead, 1998; Bel, 2001, 2003; Ortega Santos, 2006; Casielles, 2006). Hitherto, no study has investigated the influence of focus or definiteness in word order of Spanish monolingual children. However, cross-linguistic research indicates late acquisition of other discourse-related phenomena, such as the acquisition of the article system in Dutch (Schaeffer & Mathewson, 2005). Such evidence suggests that children may be aware of the effect of focus and definiteness later in life, as opposed to verb type. For Italian, Vernice & Guasti (2015) found that 4 and 5-year-old children preferred VS with unaccusative verbs and SV with unergative predicates. Importantly, definiteness of the subject was found to affect order choice in unaccusative predicates only: definite subjects prompted SV orders with these verbs, which otherwise appear with VS. However, we cannot assume that Spanish monolingual children may make this distinction. The abovementioned findings do not provide sufficient insight as to how these factors, or other we do not examine in this study, evolve in the acquisition of word order in monolingual children5.
2.2.2 Word order in heritage Spanish
The seminal studies on the acquisition of word order in bilinguals are mostly on heritage speakers of Spanish in the United States, all of which consistently find an SV overgeneralization compared to monolingual speakers (Zapata et al., 2005; De Prada-Pérez &Pascual y Cabo 2012; Montrul 2004)6. In contrast to these findings, van Osch et al., (in prep) found heritage speakers of Spanish in The Netherlands overgeneralize VS, compared to monolinguals. The authors explained this result as an effect of cross-linguistic influence from Dutch. As opposed to English, where generally only SV is
5The few studies available on children and do not provide enough evidence about word order preferences
in early acquisition, although Friedman & Costa (2011) found a VS overgeneralization in a sentence repetition task.
6Note that Zapata et al., (2000) did not include a control group, but their findings on heritage speakers
possible, Dutch allows for VS in main clauses as a consequence of the V2 phenomenon7. When adverbs or other elements appear in first position of the sentence, the verb moves to second position and the subject follows in third place (4b) (Adger, 2003). Therefore, both Dutch and Spanish allow VS. Nevertheless, these orders have a different syntactic nature in each language. In Spanish, subjects of both unaccusative and unergative verbs can remain in their base post-verbal position. In the case of unergative constructions, the subject can optionally move to [Spec, IP], where it receives a different interpretation. SV in unergative predicates is interpreted in broad focus, and VS, in narrow focus (Dominguez, 2013). In contrast, post-verbal subjects in Dutch are consequence of the obligatory nature of the verb to move to second position when other elements are topicalized in root sentences (Riemsdijk & Zwarts, 1997).
4 a) Jan at gisteren een broodje
‘Jan ate yesterday a sandwich’
b) Gisteren at Jan een broodje ‘Yesterday ate Jan a sandwich’
Similar to the acquisition research of monolingual speakers, the studies on American heritage speakers report contradicting results as to what factors influence word order. Generally, the distinction of predicate types does not seem to raise much difficulty for bilinguals, while the acquisition of focus is sometimes reported to be problematic or contested that it does not take place at all8. (Zapata et al., 2005). A major problem is that some of these studies do not always include or control for the same factors, which could influence the various contradicting results in the literature. In contrast, van Osch et al., (in prep) included definiteness of the subject as a possible predictor, and controlled for other factors such as tense or animacy. As opposed to verb type and focus, definiteness was not a significant predictor for word order in these heritage speakers.
7 In most of the Germanic languages, the finite verb appears in second position in main clauses. Because
objects usually precede non-finite verbs in languages like Dutch, they are considered to have SOV underlying order. When the object or elements such as negative adverbs are topicalised to first position, the finite verb must appear in second position, e.g. Ditboek las ikgisteren‘I read this book yesterday’. In Generative Grammar, this phenomenon is explained by strong features [uclause-type] and [top] that need not be checked in the T and C nodes respectively in the syntactic tree (Adger, 2003)
8De Prada-Pérez &Pascual y Cabo (2012) actually find the opposite: verb type was more vulnerable tan
The abovementioned studies investigate word order in adult heritage speakers, but no research is available on when bilingual children acquire knowledge of the different factors, and what their word order preferences might be. However, exploratory research on 5 year-old Spanish-Dutch bilingual children in the Netherlands has examined verb type and focus as possible factors for word order (García González et al., in prep). Although definiteness was not tested as a factor in this study, it was controlled for by including definite subjects only. The authors did not report knowledge of either verb type or focus in these bilingual children. Nevertheless, an SV overgeneralization was found, which could perhaps be explained by the exclusion of indefinite subjects, argued to increase the choice of VS (Roggia, 2011, van Osch et al., in prep).
3. The present study
Since previous research does not provide sufficient insight about the acquisition and development of word order in Spanish, the nature of this study is exploratory. As such, precise hypotheses cannot be formulated. Instead, we establish expectations on the basis of the IH. Although our focus is on children acquiring heritage Spanish in The Netherlands, we include experimental data on monolingual children to establish a comparison between monolingual and bilingual development. Additionally, we connect our results with previous research in adult heritage speakers aiming to follow the bilingual development of this topic across the lifespan. This study is guided by two main research questions, the first concerns word order preferences, the second, considers the knowledge of the three different factors.
Previous research on word order has found differences between monolinguals and heritage speakers. Specifically, heritage speakers of Spanish in The Netherlands overgeneralize VS as compared to monolingual controls, which has been explained by cross-linguistic influence from Dutch (van Osch et al., in prep). In contrast, García González et al., (in prep) found an SV preference in 5-year-olds, although no control group was included. On the basis of these opposing findings, we aim to answer the following research question:
RQ1) Does a drop in input in Spanish of heritage speakers favour a
Although we cannot predict the exact age at which the change may occur, we anticipate that the preference for VS increases as input in Spanish decreases and exposure to Dutch becomes more prominent in the daily lives of bilingual children. Since we assume that the tendency towards VS potentially occurs due to cross-linguistic influence from Dutch, the preference is only expected in bilinguals.
In regard to the factors that influence word order, van Osch et al., (in prep) found that adult heritage speakers of Spanish in the Netherlands exhibit knowledge of verb type and focus, but not of definiteness. Thus bilingual children should show knowledge of verb type and focus at some stage during the development, but they might not necessarily be sensitive to the influence of definiteness. On the basis of these findings, we formulate our second research question:
RQ2) At what age (if at all) are verb type, focus and definiteness acquired
by bilingual children? And are there differences with respect to monolingual peers?
According to the IH, the greater difficulty of factors at the syntax-discourse phenomena results in a relatively late age of acquisition. Thus, for all children, verb type is expected to be acquired earlier than focus and definiteness. In the case of bilingual speakers, acquisition of syntax-interface factors may occur later than in monolinguals, or may not take place at all. Therefore, the differences between monolingual and bilingual children should be greater for focus and definiteness.
4. Method
For the purpose of this study, we adapted van Osch’s et al., (in prep) grammaticality judgment task into an oral preference task suitable for children at the ages of 9 and 139.
This allowed us to connect our results to van Osch’s et al., findings on adult heritage speakers. The cross-sectional data was collected in a period of 4 months.
4.1 Participants
4.1.1-Spanish bilingual (heritage)
Thirty Spanish-Dutch bilingual children divided in two groups of 9 and 13 years old were tested for this study. Mean ages are provided at the end of this section10. The main
criteria to be included in the study were: a) to have been born in the Netherlands or moved to the country before the age of 5; b) to have been exposed to Dutch and Spanish preferably from birth, otherwise before the age of 5; c) to use and be exposed to Spanish on a daily basis.
27 of 30 children were born in The Netherlands, and the remaining three were born in Spain but had moved to the Netherlands before the age of 5. In all cases, the mother is a Spanish native speaker, and the father is a Dutch native speaker. Around 65% of the Spanish-speaking parents are originally from Spain. The background of the remaining 35% varies between Mexico, Colombia and mostly Argentina11. 4 children
have two Spanish-speaking parents but were exposed to Dutch before the age of 5.12 Prior to the experiment, all parents filled in a language questionnaire for the child reporting the daily use and exposure to Spanish, as well as general information about the child’s exposure to the language since birth13. The amount of exposure to Spanish differs across participants since some children, about 60%, attend a Spanish Saturday school, reading groups or language lessons organized by the Spanish consulate in The Netherlands. The child’s network of speakers also varies across participants, especially for those children who go to a Spanish-speaking country two to three times a year to visit their relatives, as opposed to those who have only done so a few times in their lives. Nonetheless, all children in this paper are considered balanced bilinguals as they comprehend and use both languages with ease. In some occasions, participants were siblings, and spoke mostly Dutch but also Spanish between them.
4.1.2 Spanish monolingual (control)
Two control groups including 46 Spanish monolingual children of 9 and 13 years old were tested for this study. All were children born and raised in Madrid. They spoke
10 For the sake of clarity of this paper, we will refer to these groups as 9-year-old and 13-year-old
bilingual group even though the age range is greater in both groups.
11 In no case the Spanish-speaking parent spoke Caribbean Spanish, in which the distribution of null and
overt subjects is different from other varieties of Spanish (Villa García, 2013)
12The slight differences between participants raised from the difficulty to find children that met all ideal conditions, namely, one Spanish-speaking parent, been born in The Netherlands and been exposed to the two languages from birth.
exclusively Spanish and had basic knowledge of English, which they had learned in school. Table 2 displays mean ages for the four groups.
Monolingual: 9yo (N=24) Monolingual: 13yo (N=22) Bilingual: 9yo (N=15) Bilingual: 13yo (N=15) Mean: 9.3 Range: 8.11-9.9 SD: 0.5 Mean: 13.4 Range: 12.1-14.0 SD: 0.5 Mean: 9.3 Range: 7.8-10.4 SD: 0.9 Mean: 12.8 Range: 11.6-15.0 SD: 1.1
Table 2. Mean age for each group.
4.2 Stimuli:
The experiment consisted of an oral preference task adapted from van Osch et al., (in prep.). For the sake of comparison, the verbs from the original task were used to create stories that were suitable for children of different ages. The stimuli targeted the three factors: Unaccusativity (unaccusative or unergative), focus (broad or narrow) and definiteness of the subject (definite or indefinite). Similar to van Osch et al., (in prep.), stories followed by a question were used to contextualize the target items and to create the specific broad or narrow focus. The question ¿Qué pasó? ‘What happened?’ targeted the broad focus condition, while the question ¿Quién V? ‘Who V-ed?’, elicited the narrow focus condition.
A total of 32 stimuli were divided into 8 conditions. Responses consisted of pairs of sentences with SV and VS orders. Three practice items were included but no fillers were used in order to keep the task short and manageable for children. Table 3 displays an overview of the conditions and an example (5) follows below. The complete set of items can be found in Appendix 3.
Subject Focus Unaccusativity Condition DEFINITE BROAD UNACCUSATIVE 1 INDEFINITE BROAD UNACCUSATIVE 2 DEFINITE NARROW UNACCUSATIVE 3 INDEFINITE NARROW UNACCUSATIVE 4 DEFINITE BROAD UNERGATIVE 5 INDEFINITE BROAD UNERGATIVE 6
DEFINITE NARROW UNERGATIVE 7 INDEFINITE NARROW UNERGATIVE 8
Table 3. Summary of conditions in the oral preference task.
5. Condition 1
Tres perritos están jugando en la playa y se están divirtiendo mucho. Mientras juegan, llega su amigo el gato, muy triste porque no lo habían avisado.
[Q]¿Qué pasó? [A1] El gato llegó (SV) [A2] Llegó el gato (VS)
‘Three dogs are playing at the beach and having a lot of fun. While they play, their friend the cat comes and sees them and he gets very sad because they had not asked him to join them.’
[Q]‘What happened?’ [A1]‘The cat arrived’ (SV) [A2] ‘Arrived the cat’ (VS)
4.3 Procedure:
Participants of the monolingual groups performed the task during school hours. The experiment was carried out individually in a quiet room. Previous to the task, the child was briefly asked about her language background and knowledge of other languages. This was followed by a short chat with the experimenter about the motivation of the study. Children of the bilingual groups were mostly tested at their homes, and occasionally, at the facilities within the University of Amsterdam or similar institution, such as the Instituto Cervantes in Utrecht. When the child was bilingual, the experimenter talked to her for at least fifteen minutes to assure she felt comfortable, and to activate her Spanish. During the instruction of the task, participants were told that two puppets (Ana and María) were girls living in the Netherlands who were learning
Spanish and were going on a holiday to Spain. The child was asked to help the puppets with their Spanish by telling them which answer (order) sounded most natural to them, so that the puppets could practice and be understood in Spain. The experimenter stressed that no answers were incorrect and instructed the child to choose “the first natural response”. The stimuli were presented on a computer on Microsoft PowerPoint. For each experimental item, the child would listen to a story followed by a question while she focused on an image that accompanied the story. Two puppets appeared on a screen sequentially and each provided an answer to the question. Each answer displayed either SV or VS order and the child was instructed to choose one of them. The order of presentation of the puppets and the target response were randomized.
4.4 Analysis:
Previous to the statistical analysis, the data was studied and plotted in a bar graph representing the frequency of the chosen order in relation to the factors unaccusativity, focus, and definiteness. This provided a priori understanding of word order preferences per age and language group, and the effect of each factor. For the statistical analysis, response was always the dependent variable. A first mixed-effects model (glmer) on binomial data included the interaction between age and language as fixed factors for the entire data set. Additional models were run for subsets of the data including language, unaccusativity, focus definiteness and their interaction as fixed factors. Fixed effects were significant at α = .05. Additionally, subject and item were modeled as random factors which allowed us to reduce the effect of individual variation. Possible effects such as order of the stimuli presented and preference for a specific puppet were also checked. Neither of them showed an effect and were therefore not included in the final model. All analyses were carried out in the programming language and statistical tool R.
5. Results
The results are organized as follows. First, we performed statistical analysis to find whether bilingual children differ from monolinguals in their overall word order preferences and if there are changes over time. These results, which respond to RQ1 are presented in §5.1. Secondly, we analyzed whether the three target factors unaccusativity, focus and definiteness, influence word order in children to answer RQ2. These results are reported by age groups in 5.2. For all mixed-effects models described
in this section, response was always the independent variable, and subject and item were modeled as random factors to control for within-group variation and differences across items.
5.1 Word order preferences (RQ1)
We performed statistical analysis to find whether bilingual children differ from monolinguals in word order preferences and if there is an age effect. In order to do so, we ran a mixed-effects model for all data including age, language and the interaction between the two as fixed factors. The data set from all four groups was included in the model. Both language (z= 2.29, p<0.05) and age (z=2.02, p<0.05) are main effects, but the interaction is not significant. This result means that, overall, children’s preference for the VS order increases over time. Additionally, VS is relatively more frequent in bilinguals compared to monolinguals. The graph in figure 1 depicts the relative frequency of SV and VS in children over time.
Figure 1. Relative frequency of SV and VS over time
5.2 Verb type, focus and definiteness (RQ2)
We ran two additional models for each age group, 9 years old and 13 years old, including monolinguals as well as bilingual data. The aim was to study the influence of
0 50 100 150 200 250 300 350 400 450 500
9yo 13yo 9yo 13yo
monolingual bilingual
SV VS
the three factors and the differences between language groups at each stage. In both cases, unaccusativity, focus, definiteness and language, as well as the interaction between the four factors, were set as fixed effects. As in the previous model, subject and item were modeled as random effects, and response was the outcome.
5.2.1 Children age 9
The model revealed main effects of unaccusativity (z= 5.47, p<0), and definiteness (z=4.76, p<0), as well as marginally significant effects of language (z=1.76, p= 0.07) and focus (z=1.86, p=0.06). Thus all factors are significant predictors for word order in monolingual and bilingual children. Taking this result overall, these main effects mean that VS is relatively more frequent with unaccusative verbs (Figure 2), narrow focus (Figure 3) and indefinite subjects (Figure 4) for all children at age 9.
Figure 2. Order preference for verb types in 9-year olds
0 50 100 150 200 250 300 350 400 450 unaccusative unergative SV VS
Figure 3. Order preference focus types in 9-year olds
Figure 4. Order preference for definiteness of the subject in 9-year olds
In addition to main effects for each factor, we found a two-way interaction between language and focus (z=-2.03, p< 0.05), and a marginally significant interaction between focus and definiteness (z=1.75, p=0.07). Thus the only difference between language groups relays in children’s sensitivity to how focus influences word order. Whereas all three factors influence word order in 9-year-olds overall, monolinguals are aware of the effect of focus to a larger extent than bilinguals. This can be observed in figure 5, where bilinguals seem to make no distinction for broad or narrow focus, in contrast to monolinguals, in which the difference between broad and narrow focus is greater.
0 50 100 150 200 250 300 350 400 450 broad narrow SV VS 0 50 100 150 200 250 300 350 400 450 definite indefinite SV VS
Figure 5. Interaction between focus and definiteness for monolingual and bilingual children
The almost significant interaction between focus and definiteness indicates that the distinction between definite and indefinite subjects is greater in narrow focus than in broad focus, which we can observe in figure 6 below.
Figure 6. Marginal interaction between focus and definiteness for all 9-year-olds
5.2.2. Children age 13
A third model in which only the data of 13-year olds were included revealed main effects of language (z=1.76, p=0.07), unaccusativity (z=6.09, p<0), focus (z=4.84, p<0), and definiteness (z=4.97, p<0). Similarly to the findings for 9-year olds, all factors are significant predictors for word order for both language groups. Once again, VS is relatively more frequent with unaccusative verbs (Figure 7), narrow focus, (Figure 8)
0 50 100 150 200 250 300
broad narrow broad narrow
monolingual bilingual SV VS 0 50 100 150 200 250
definite indefinite definite indefinite
broad narrow
SV VS
and indefinite subjects (Figure 9) in these 13-year olds.
Figure 7. Order preference for verb types in 13-year olds
Figure 8. Order preference for focus types in 13-year olds
0 50 100 150 200 250 300 350 400 unaccusative unergative SV VS 0 50 100 150 200 250 300 350 400 broad narrow SV VS
Figure 9. Order preference for definiteness of the subject in 13-year olds
In addition to the main effects, the model revealed a two-way interaction between language and focus (z=-3.29, p<0), which indicates that the distinction between focus types is greater in monolinguals than in bilinguals, as depicted in figure 10. This result mimics the finding for 9-year olds, thus, there is no apparent difference in age groups to this respect. This also means that the knowledge of how focus influences word order does not seem to change over time. Additionally, we find a marginally significant (z=-1.9, p=0.06) interaction between unaccusativity and focus. We can observe in figure 11 that the overall difference between broad and narrow focus for unergative verbs is greater than for unaccusative verbs.
0 50 100 150 200 250 300 350 400 definite indefinite SV VS 0 50 100 150 200 250
broad narrow broad narrow
monolingual bilingual
SV VS
Figure 10. Interaction between language and focus in monolingual and bilingual 13-year olds
Figure 11. Marginal interaction between verb and focus in bilingual and monolingual 13-year olds
5.3 Interim summary
The statistical analysis for word order preferences provided us with two conclusions. First, the preference for VS significantly increases with age, and thus we find more instances of this order in adolescents regardless of language group. Second, VS is more frequent in bilinguals compared to monolinguals. The absence of an interaction between language and age does not permit us to conclude a specific behavior for each of the four groups.
When analyzing the interaction between language, unaccusativity, focus and definiteness for each age group, we find differences across language groups. In monolinguals, unaccusativity, focus and definiteness are main effects, whereas all factors but focus are significant in bilinguals. Even though word order in bilinguals might be affected by the influence of focus, it certainly is to a lesser extent than in monolinguals. Table 4 below includes main effects and interactions for the three models described in this section.
0 50 100 150 200 250
broad narrow broad narrow
unaccusative unergative
SV VS
p-value value -z Interface Effect/Condition Group p<0 5.47 INTERNAL UNACCUSATIVITY 9-year-olds p=0.06 1.86 EXTERNAL FOCUS p<0 4.76 DEFINITENESS p=0.07 1.76 - LANGUAGE p<0.05 -2.03 - LANGUAGE*FOCUS p=0.07 1.75 - FOCUS*DEFINITENESS p<0 6.09 INTERNAL UNACCUSATIVITY 13-year-olds p<0 4.84 EXTERNAL FOCUS p<0 4.97 DEFINITENESS p=0.07 1.76 - LANGUAGE p<0 3.29 -LANGUAGE*FOCUS p=0.06 -1.90 -UNACCUSATIVITY*FO CUS p<0.05 2.48 -LANGUAGE ALL p<0.05 2.03 -AGE
Table 4. Summary of main effects and interactions for the three statistical analysis performed
6. Discussion
The purpose of this exploratory study is twofold: we investigate word order preferences in bilingual and monolingual children and examine how they are influenced by verb type, focus and definiteness. Taking a comparative approach, we strive to determine whether bilingual children differ from their monolingual peers in both the word order used and the acquisition of the aforementioned factors. Below, we interpret our results and evaluate our two research questions:
1) Does a drop in input in Spanish of heritage speakers favour a preference towards VS compared to monolingual peers?
2) At what age (if at all) are verb type, focus and definiteness acquired by
bilingual children? And are there differences with respect to monolingual peers?
6.1 Word order preferences overall
As shown in this study, word order preferences gradually develop in all children until the age of 13 years. Since we find an overall increase in the frequency of VS, we interpret this result as the natural process of acquisition of word order in Spanish. It is important to stress that the differences we find for overall preferences are quantitative rather than qualitative, as we analyzed relative preferences based on the frequency of SV and VS. Let us recall that based on the theoretical description of word order, VS is more felicitous than SV in 3 out of 4 cases: unaccusative verbs in broad as well as in narrow focus, and unergative verbs in narrow focus. (Domínguez, 2013; Hertel, 2003). Even though experimental studies on monolinguals do not always confirm this theory, they often find main effects of verb type and focus, VS being the preferred option for unaccusatives and narrow focus (van Osch et al., in prep). Considering that Spanish is an SV(O) language, we could speculate that children are exposed to a great deal of SV structures from early on. In the case of intransitive verbs, children might need larger exposure to this phenomenon to realize that VS is, in fact, salient. If that was the case, the same realization should reflect the acquisition process in bilinguals.
Although we evidenced an increase in the use of VS in both groups, the frequency of this order relatively higher in bilinguals than in monolinguals. This is intriguing, since it contradicts results from the body of research on heritage speakers of Spanish in the United States. While the language combination in the former is English-Spanish, our heritage speakers are Dutch-Spanish bilinguals. Thus, the main difference between these studies and the present research is the society language that heritage speakers are exposed to.
In line with van Osch et al., (in prep), we propose that the V2 phenomenon in Dutch may be the cause for the opposing findings between heritage speakers in the United States and The Netherlands. As explained in Section 2, Dutch, allows VS orders unlike English. The presence of VS in the Dutch input may simply raise bilingual speakers’ awareness of this frequent order in Spanish early on, since V2 in Dutch is in
fact acquired very early by monolingual children (Blom, 2014). Thus these circumstances could enhance the choice of VS. Future research of young bilinguals with other language combinations could aim to prove this argument. If the choice of VS is found in bilinguals whose other language also allows VS orders, our findings might be an indication of cross-linguistic influence from the contact language.
The fact that the VS frequency continues to increase in these bilingual children over time is particularly interesting when contrasted to van Osch et al. (in prep.) results on heritage adult speakers. Those speakers exhibited a VS overgeneralization across conditions, compared to monolinguals. Considering that word order preferences in our bilingual children seem to drift apart from monolinguals already at 9 years old, they may behave like adults in van Osch et al., (in prep.) once they reach adulthood.
Furthermore, if the influence from Dutch causes the diversion between bilingual and monolingual children, we would anticipate that the trend for VS becomes accentuated as input in Dutch increases. Therefore the tendency found in teenagers might not yet occur in younger bilinguals who have not yet entered Dutch school. Only one study has investigated word order in 5-year-old Dutch-Spanish bilinguals, and found an overall preference for SV (García González et al., in prep). Although such study used a similar methodology, the findings must be taken with caution, since the SV preference could have been caused by the use of items with definite subjects only. However, this result could be a further indication that VS increases as the input in Dutch becomes prevalent, since those children have just entered the Dutch education system, and are or have been immersed in Spanish until recently.
6.2 Acquisition of factors
Our results show that at least verb type, focus and definiteness of the subject influence word order in bilingual children already at age 9. Importantly, all factors are main effects at both 9 and 13 years-old, as well as the interaction between language and focus, which demonstrates divergence with respect to monolingual peers. Specifically, bilinguals differ from monolinguals in their sensitivity to how focus influences word order in Spanish. This result, however, does not fully confirm our expectations, which we based on the IH and the findings for adult heritage speakers in van Osch et al., (in prep). On the one hand, the IH predicts that the syntax-discourse factors of focus and definiteness are more vulnerable in bilinguals than in monolinguals, while this is not the
case for verb type. On the other hand, adult heritage speakers exhibited knowledge of verb type and focus but not of definiteness (van Osch et al., in prep). Accordingly, we expected that verb type would be acquired early on, while the acquisition of focus and definiteness would be delayed, or not occur at all. Our predictions are thus borne out with respect to verb type, but not entirely for focus and definiteness. Below, we discuss this finding with respect to two possibilities, which are not mutually exclusive. First, the difference in terms of focus and definiteness in child and adult heritage speakers may be explained as a case of re-analysis or attrition, in line with a proposal by Polinsky (2016). Second, we inquire whether the assessment of same-level of complexity for focus and definiteness according to the IH is indeed accurate.
Polinsky (2016) offers four different scenarios of bilingual development in which child bilinguals and adult heritage speakers may pattern alike, or one of them may assimilate monolinguals (the baseline), while the other deviates. The relevant scenario for our study is that in which a certain structure might be acquired by bilingual children, but undergoes reanalysis or attrition in adult heritage speakers. In other words, how a certain linguistic phenomenon is understood and produced might change at some point of the acquisition process. Nevertheless, Polinsky’s scenario implies that bilingual children parallel monolingual children, while both deviate from adult heritage speakers. However, our results do not exactly correlate with such scenario: 13-year-old bilinguals have not (yet) acquired knowledge of focus, unlike monolingual peers and adult heritage speakers. Moreover, bilingual children are sensitive to the effect of definiteness, like monolinguals, but different from adult heritage speakers. Importantly, Polinsky’s proposal of re-analysis cannot be interpreted for each factor individually, but for the phenomenon of word order itself. Therefore, there are multiple elements that may affect how bilinguals integrate, develop and re-analyze the information they receive with respect to word order, one of them being the continuously changing exposure to both their languages throughout their lives. The fact that these children are not yet fully sensitive to focus distinctions does not imply that they will not acquire this factor eventually. In fact, it seems plausible that these bilingual children represent an intermediate stage in which they have not yet become fully aware of the influence of focus on word order. In that case, the range of ages targeted in this study represents a crucial time lapse in the heritage language development of bilinguals that could explain this potential intermediate stage. Notice that the children in this study were exposed to both their languages during early childhood, but in most cases, their input in Spanish
decreased over the years. Even though some of the 9-year-old children receive additional input in Spanish through reading groups or Spanish lessons, this is rarely the case in adolescents, which we can confirm by the information collected via input questionnaires. We can interpret these changes in the input as a contributing factor for the differences that we find for age groups.
An alternative explanation why our results are not fully explained by the IH might be the presumed level of complexity of the syntax-discourse factors in particular. According to the IH, both focus and definiteness are properties that engage in a linguistic as well as in a cognitive domain. Thus it is the high processing demand of integrating information from separate domains what makes properties at the external interfaces vulnerable in the grammars of bilingual speakers. However, what if there were multiple degrees of complexity within the different interfaces? We wish to acknowledge the possibility that focus and definiteness are not equally complex, but actually represent different levels on a continuum of gradual complexity. This idea has lately been advanced for various studies on heritage languages (Hoot, 2017; Lakeko & Polinsky, 2015; Westergaard & Andersen,2015). Specifically, Lakeko & Polinsky (2015) tackle the problem to establish the structural complexity of interface phenomena, while Hoot (2007) argues that focus in Spanish cannot fully be treated as an internal or an external interface factor. Instead, he proposes that focus represents characteristics of both interfaces, which could explain why some studies find that heritage speakers acquire this property (De Prada Pérez & Pascual y Cabo, 2012), whereas other find the opposite (Zapata et al., 2005). Regardless of the interface nature of focus, Hoot claims that proficiency of the heritage language alone should lead to successful acquisition. If we follow Hoot’s proposal, our bilingual children should eventually acquire knowledge of focus unless critical absence of input in Spanish.
Nevertheless, others have argued that the complexity of a property like focus should not matter once it is acquired, as long as it is frequent in the input (Westergaard & Andersen). This proposal on frequency versus complexity could explain why adult heritage speakers of Spanish (van Osch, et al., in prep) do not undergo attrition of this factor. Yet, the question still remains for the different effects of definiteness in bilingual children and adult heritage speakers. On the one hand, assuming that definiteness is in fact a syntax-discourse factor, the IH does not explain its early acquisition and, presumably, later attrition. On the other hand, in Westergaard & Andersen’s view, definiteness might be less complex than initially established, and its effect less frequent
in the input. Thus, children might be able to become aware of this factor early on, but not rely on it as much when they realize the importance of other factors like focus.
In sum, Westergaard & Andersen’s proposal may add to Polisnky’s scenario of reanalysis. Assessing what degree of complexity corresponds to each of the factors under investigation goes beyond the goal of this study. Nevertheless, it is crucial to highlight that various elements may contribute to the re-analysis of word order in the lifetime of bilinguals. No need to say that the complexity of the phenomenon itself, of each factor individually, may be a major contributor to this re-analysis. Yet, the changes in the input of a bilingual’s two languages cannot go unseen. Furthermore, we have little understanding of how focus and definiteness engage in other linguistic phenomena in Dutch, and whether this might add to the re-evaluation of word order in Spanish, too.
7. Concluding remarks: limitations of the study and suggestions for further research
This study has shown that bilingual children acquiring heritage Spanish in The Netherlands are aware of the influence of verb type and definiteness in word order, but are still developing their knowledge of focus. Moreover, these children choose VS more frequently than monolingual peers. This finding indicates that bilingual children are still developing their knowledge of word order. Yet, they seem to follow a path in which they eventually resemble the characteristics found for adult heritage speakers in van Osch et al., (in prep)
By targeting different stages of acquisition, this study has opened a window to bilingual development across de lifespan. Nevertheless, we have also raised many inquiries with respect to earlier and later stages of the development, as well as to other factors that may play a role throughout the acquisition process. Importantly, this research focused on comprehension, but it is worth exploring whether the same results would be replicable in production tasks. A full view of how word order is understood and produced is fundamental when comparing bilinguals and monolinguals, as the gap between the two domains might be greater in production.
Despite our focus being on bilingual speakers, this project has provided valuable insight in the acquisition process of word order in monolinguals. Considering the number of studies that have found that adult monolinguals do not exactly behave as it has been extensively prescribed, research that tests the acquisition of word order in
young monolingual children is extremely necessary. The present study has shown that the expected adult-like behaviour framed in the literature of word order is only reached in adolescence. Moreover, the two control groups we tested do not unambiguously show consecutive acquisition of all factors, as predicted by the IH. Therefore, future research should aim at investigating the acquisition process in monolinguals from early on.
Lastly, we should add that the complexity of the various factors involved in word order in Spanish leads to great variation among speakers. The studies that report monolinguals’ use of word order often convey contradicting results. Although different methodologies are obvious reasons for these mismatches, it is worth considering that the disparities may arise from considerable variation among speakers. Therefore, it could be expected that such variation would affect bilinguals’ interpretations, too. An issue that is often forgotten is that heritage speakers, or children acquiring a heritage language in this case, may be exposed to largely diverse sources of input. In fact, most of these children’s input comes from speakers whose L1 Spanish is also influenced by L2s and sometimes L3s, and this plausible effect should not go unnoticed when comparing bilingual and monolingual children.
Acknowledgements
I would like to thank all children who participated in this study without whom this project would have not been possible. The findings reported in the study are only a fraction of what I have learn from my time with these bilingual children. I am immensely grateful to their parents, for their help, commitment and enthusiasm.
I thank Rosana Hilara, Carmen Serrano, Chus, and all the staff from Colegio Arcángel and IES Europa for their cooperation in the testing of monolingual children.
Especial thanks to Hannah Lutzenberger for her continuous and indefatigable support throughout this project and very helpful feedback. I want to thank Kyle Snyder, too, for the revision of this paper.
Finally, I am grateful to Aafke Hulk and Suzanne Aalberse for their supervision, encouragement and confidence in me. Last and most importantly, I am eternally grateful to Brechje van Osch for her exceptional supervision throughout this process, help and continuous support.
References
Adger, D. (2003). Core Syntax; A minimalist Approach, Oxford University Press. Al-Kasey, T., & Pérez-Leroux, A.-T., (1998). Second language acquisition of Spanish
null subjects. In: Flynn, S., Matohardjono, G., O'Neil, W. (eds.), The Generative
Study of Second Language Acquisition, 161- 185. Hillsdale, N.J.: Lawrence
Erlbaum.
Bel, A. (2001). Sujetos nulos y sujetos explícitos en las gramáticas iniciales del castellano y el catalán. Revista Española de Lingüística, 31 (2), 537-562.
Bel, A. (2003). The syntax of subjects in the acquisition of Spanish and Catalan.
Probus, 1 (5), 1-26.
Blom, E., Polišenská, D., & Weerman, F. (2008). Articles, adjectives and age of onset: the acquisition of Dutch grammatical gender. Second language research, 24 (3), 297-331.
Blom, E. (2014). From the bird to the frog: A closer look at the acquisition of V2 in Dutch. Linguistic Approaches to Bilingualism, 4 (3), 321-324
Casielles, E., Jean A., Sahyang K., Geoffrey N., and Work, R. (2006). Syntactic and Discourse Features of Subjects in Child Spanish: Evidence from Spanish/English Bilingual Acquisition. BUCLD 30, 72-83.
De Houwer, A. (1990). The acquisition of two languages from birth: A case
study. Cambridge: Cambridge University Press.
De Prada Pérez, A. de & Pascual y Cabo, D. (2012). Interface Heritage Speech across Proficiencies: Unaccusativity, Focus, and Subject Position in Spanish. Refereed Proceedings from the Hispanic Linguistics Symposium 2010. Somerville, MA: Cascadilla Proceedings Project.
Dominguez, L. (2013) Understanding interfaces: second language acquisition and
native language attrition of Spanish subject realization and word order variation, Amsterdam, John Benjamins.
García González, E., van Osch, B., Aalberse, S. Hulk, A. (in prep.) The acquisition of pre- and post-verbal subjects in Dutch-Spanish bilingual children.
Grinstead, J. (1998). Subjects, sentential negation and imperatives in child Spanish and
Catalan. Ph.D. Dissertation, UCLA. 188
Grinstead, J. (2000). Tense, Number and Nominative Case Assignment in Child Catalan and Spanish. Journal of Child Language 27(1), 119-155
Hertel, T. J. (2003). Lexical and discourse factors in the second language acquisition of Spanish word order. Second Language Research, 19 (4), 273-304.
Hoot, B. (2017) Narrow presentational focus in heritage Spanish and the syntax‒discourse interface. Linguistic Approaches to Bilingualism, 7:1, 63-95 Hulk, A., & Müller, N. (2000). Bilingual first language acquisition at the interface
between syntax and pragmatics. Bilingualism: Language and Cognition, 3, 227-244.
Iverson, M., Kempchimsky, P., & Rothman, J. (2008). Interface vulnerability and knowledge of the subjunctive/indicative distinction with negated epistemic predicates in L2 Spanish. EUROSLA Yearbook, 8, 135–163.
Kupisch, T. & Rothman, J. (2016) Terminology matters! Why difference is no incompleteness and how early child bilinguals are heritage speakers.
International Journal of Bilingualism, 1-19.
Laleko O. & Polinsky M. (2016). Between syntax and discourse: Topic and case marking in heritage speakers and L2 learners of Japanese and Korean. Linguistic
Approaches to Bilingualism, 4, 396-439.
Levin, B. and Rappaport Hovav, M. (1995) Unaccusativity: At the Syntax-Lexical
Semantics Interface, Linguistic Inquiry Monograph 26, MIT Press, Cambridge,
MA
López Meirama, B. (1997). La posición del sujeto en la cláusula monoactancial en
español. Santiago de Compostela: Servicio de Publicacións da Universidade de
Santiago de Compostela.
Lozano, C. (2006). Focus and split-intransitivity: the acquisition of word order alternations in non-native Spanish. Second Language Research, 22 (2), 1-43. Meisel, J. (1994). Bilingual first language acquisition: French and German
grammatical development. Amsterdam: John Benjamins
Montrul, S. (2004). The Acquisition of Spanish. Morphosyntactic Development in
Monolingual and Bilingual L1 Acquisition and in Adult L2 Acquisition.
Amsterdam: John Benjamins.
Montrul, S. (2004). Subject and object expression in Spanish heritage speakers: A case of morphosyntactic convergence. Bilingualism: Language and Cognition, 7, 125-142.
Montrul, S., Foote, R. and Perpiñán, S. (2008). Gender agreement in adult second language learners and Spanish heritage speakers: the effects of age and context of acquisition. Language Learning 58, 503–53.
Montrul, S. (2016). Dominance and proficiency in early and late bilingualism. In Silva-Corvalán, C & Treffers-Daller, J. (2016), Language Dominance in Bilinguals.
Issues of Measurement and Operationalization. Cambridge: Cambridge
University Press.
Ortega-Santos, I. (2006). “On New Information Focus, Sentence Stress Assignment Conditions and the Copy Theory: A Spanish Conspiracy.” University of
Maryland Working Papers in Linguistics 14. Kazanina, N., U.Minai, P.
Monahan, and H. Taylor (eds.). College Park, Maryland: UMWPiL, 88-212. Paradis, J., Genesee, G., & Crago, M. (2011). Dual language development and
disorders: A handbook on bilingualism and second language learning, 2nd edn.
Baltimore, MD: Brookes Publishing.
Perlmutter, D. M. (1978). Impersonal Passives and the Unaccusative Hypothesis. In J. J.Jaeger, A. C. Woodbury, F. Ackerman, C. Chiarello, O. D. Gensler, J. Kingston, E. E. Sweetser, H. Thompson & K. W. Whistler (eds.), Proceedings
of the Fourth Annual Meeting of the Berkeley Linguistics Society February 18-20, 1978, pp. 157-189. Berkeley: Berkeley Linguistics Society.
Polinsky, M. (2011). Reanalysis in adult heritage language: New Evidence in Support of Attrition. Studies in Second Language Acquisition, 33(2), 305-328.
Polinsky, M. (2016). Bilingual Children and Adult Heritage Speakers: The Range of Comparison. International Journal of Bilingualism, 1-17.
Riemsdijk, H. & Zwarts. F (1997). Left Disclocation in Dutch and the Status of Copying Rules. In: Anagnostopoulou, E., H.C. van Riemsdijk & F. Zwarts, eds. Materials on Left Dislocation. Amsterdam/Philadelphia: John Benjamins. pp 13-29.
Roggia, A. (2011). Unaccusativity and word order in Mexican Spanish: An examination
of syntactic interfaces and the split intransitivity hierarchy. PhD dissertation.
Pennsylvania State University.
Rothman, J. (2009). Understanding the nature and outcomes of early bilingualism: Romance languages as heritage languages. International Journal of
Sheehan, M. (2004). The syntax and semantics of inversion in Romance. Durham
Working Papers in Linguistics 10: 195-209.
Silva-Corvalán, C. (1994). Language contact and change. Oxford: Oxford University Press.
Silva-Corvalán, C & Treffers-Daller, J. (2016). Language Dominance in Bilinguals.
Issues of Measurement and Operationalization. Cambridge: Cambridge
University Press.
Sorace, A. (2000). Differential effects of attrition in the L1 syntax of near-native L2 speakers. Proceedings of The 24th Boston University Conference on Language Development, pp. 719–725. Somerville, MA: Cascadilla Press.
Sorace, A. & Filiaci, F. (2006). Anaphora resolution in near-native speakers of Italian.
Second Language Research, 1, 22(3), 339-368.
Sorace, A. & Serratrice, L. (2009). Internal and external interfaces in bilingual language development: Beyond structural overlap in International Journal of
Bilingualism. 13, 2, 195-210.
Tsimpli, I., Sorace, A., Heycock, C., & Filiaci, F. (2004). First language attrition and syntactic subjects: A study of Greek and Italian near-native speakers of English.
International Journal of Bilingualism, 8 (3), 257-277.
Unsworth, S., Argyri, F., Cornips, L., Hulk, A., Sorace, A., & Tsimpli, I. (2014). The role of age of onset and input in early child bilingualism in Greek and Dutch.
Applied Psycholinguistics, 35 (04), 765-805.
Unsworth, S. (2016). Amount of exposure as a proxy for dominance in bilingual language acquisition. In Silva-Corvalán, C & Treffers-Daller, J., Language
Dominance in Bilinguals. Issues of Measurement and Operationalization.
Cambridge: Cambridge University Press.
van Osch, B., Aalberse, S. Hulk, A., Sleeman, P. (in prep.) The acquisition of subjects in heritage speakers of Spanish in the Netherlands.
Westergaard, M., & Anderssen, M. (2015). Word order variation in Norwegian possessive constructions. Germanic Heritage Languages in North America, 21.
