The acquisition of copula verbs in Basque by bilinguals

Abstract

This article focuses on the acquisition of copula verbs in Basque by Basque–Spanish bilinguals. Basque and Spanish have two copula verbs: izan and egon and ser and estar, respectively. Basque copulas are similar to their Spanish counterparts in terms of the grammatical contexts in which they are used. However, Basque and Spanish differ in one specific property, the progressive, which is built with izan in Basque. This article assumes the theoretical framework of Zagona for Basque because copula selection can be accounted for by this approach. The present study is the first one to address the issue of the acquisition of copula verbs by Basque–Spanish bilingual children in Basque; we analyze the language used in a storytelling task by 19 Basque-dominant bilinguals compared to that used by 19 child L2 Basque learners, in order to elucidate whether both groups of bilingual children use the copula in a target-like way in Basque. The study shows that no copula choice errors are produced by any of the participants with any predicates. The distribution of the predicates very much resembles the distribution of the predicates in previous studies for Spanish. Adhering to Zagona’s framework, the progressive was included in our study. In this respect, the progressive is used in a target-deviant manner, a finding that can be attributed to crosslinguistic influence. These results are ultimately attributed to children’s parallel knowledge of both interpretable and uninterpretable features, although uninterpretable features seem to be acquired a little earlier.

Keywords

adjectival individual level local predicative stage level temporal

Introduction

Some languages such as Spanish and Basque have two copula verbs. The distribution of these two predicates is a rather intricate issue both in Basque and Spanish, but it is almost identical in both languages. While grammars of Spanish for L2 learners devote plenty of space to these aspects of the Spanish language because it poses difficulties to L2 learners of Spanish whose mother tongue lacks this distinction (Geeslin & Guijarro-Fuentes, 2006, and references therein) and much theoretical work has been achieved in this domain, Basque is an almost unexplored language in theoretical terms. In the domain of language acquisition, this topic has been far less explored in Spanish as will be demonstrated in the studies that are reviewed. With regard to Basque, this subject has not been investigated at all, and hence, the present study is the first one to analyze this issue. This being so, a recent approach by Zagona (2009) for Spanish will be adopted and recent studies on Spanish acquisition will be discussed in order to derive new hypotheses and predictions for the present study.

The article is organized as follows. The first section presents and discusses the relevant studies on copula acquisition in Spanish. Section “Theoretical issues” deals with the theoretical framework adopted in this article. The research questions are spelled out in section “Research questions.” The next section presents the methodology and the results. Finally, section “Discussion and conclusions” is devoted to the results and a discussion closes the article.

Literature review

The present study shows that bilingual children, whose mother tongue possesses copula dichotomy, encounter no difficulties in acquiring a language that also has two copulas. As of today, there are no studies that have analyzed the acquisition of copulas and their predicates, that is, stage-level predicates (SLPs) and individual-level predicates (ILPs) in the Basque language. Let us first provide an impression of the topic under scrutiny. Basque makes a distinction between these two types of predicates, a distinction that seems “reminiscent of the one found in Spanish between ser and estar” (Etxepare, 2003, p. 365) and will be dealt with in the next section of this article. This distinction is attested to in the Western dialects but not in the Eastern ones (Zabala, 2003, p. 426). In order to circumscribe the present piece of research in the current discussion about SLPs and ILPs, we refer to some studies that have analyzed the acquisition of Spanish, a language that has a copula system with two copulas and may have influenced the Basque system.

As already mentioned, teachers and lecturers have long made the observation that this dichotomy poses enormous difficulties to learners whose native languages possess just one copula verb. Hence, it comes as no surprise that there is a wide range and high number of studies on this subject. Alongside the description of the most frequent problems encountered by the L2 learners of Spanish (Guijarro-Fuentes, 2008), pedagogical approaches to their teaching have a long-standing tradition, for instance, Morales and Smith (2008) to quote but just one recent proposal. The topics studied in Spanish L2 acquisition range from purely grammatical issues (Pérez-Leroux, Álvarez, & Battersby, 2010) to more sociolinguistic topics such as copula choice and gender (Geeslin & Guijarro-Fuentes, 2007, 2008).

Studies on first-language acquisition of Spanish are scarce and, hence, the range of phenomena studied regarding ser and estar “to be” are rather limited. The first study to deal with the issue of the early acquisition of ser and estar in Spanish was carried out by Sera (1992). She studied the spontaneous language of two monolingual boys from the Child Language Data Exchange System (CHILDES) database (Linaza corpus)¹ between the ages of 1;06 and 3;06 (n = 2514 utterances, 71 with ser and estar) and elicited data with the Frog Story (n = 500) with 3- (n = 11), 4- (n = 12), 5- (n = 11), and 9-year-old (n =12) monolingual Spanish children. The utterances produced by the participants in the Frog Story are distributed as follows: 3- (n = 150), 4- (n = 132), 5- (n = 82), and 9-year-old (n = 85) children and adults (n = 51). She also studied the use of ser and estar by adult Spanish-speaking monolinguals (2419 utterances, 426 with ser and estar) in the CHILDES database. Figure 1 summarizes Sera’s results for Spanish in which the four categories she discusses are included. The auxiliaries have been included in Figure 1, and they will be discussed in the results section² for Basque. Note that, as Holtheuer (2009) observes, Sera did not tease apart locations and nonlocations when referring to prepositional phrases (PPs).

Figure 1.

Sera’s results (1992).

Sera’s results can be summarized as follows:

Children’s performance in the longitudinal study (CHILDES corpus), regarding the use of nominals, adjectives, and locations, very much resembles the usage of the adults.

The use of locatives remains constant across age groups, although the older children use slightly more locatives than the younger children. Locatives are used exclusively with estar.

Younger children use adjectives with both ser and estar and the younger ones use them more than the older ones.

The use of nominals in the age group of the 9-year-olds more than doubles the use of nominals in younger ages. That is, younger children use fewer nominals than older children and always with the verb ser.

As for the auxiliaries, children use more auxiliaries with estar in the Frog Story than in the CHILDES corpus. Auxiliaries are exclusively used with estar.

Let us now have a closer look at Sera’s data in relation to the different adjectives. Sera’s (1992, p. 415) data in her Table 2 show that most adjectives used can be used with both verbs, although only three of the adjectives have been reported by Sera as being used exclusively with one of the verbs. Size and color adjectives appear exclusively with ser (largo is one exception), although they can also appear with estar.

One of the questions that has been put forward in longitudinal studies is which morphological forms are acquired first by bilingual children and whether grammatical persons are acquired simultaneously with ser and estar. Silva-Corvalán and Montanari (2008, p. 352), who studied an English–Spanish bilingual, show that the third-person singular is the first form to appear both with ser and with estar in the present tense. However, es (ser) is the earliest unambiguous occurrence at 1;6, Silva-Corvalán and Montanari (2008, p. 349), that is to say, 2 months before está (estar). The third-person plural is established next, together with the first-person singular. The first-person plural follows and the second-person plural is not used at all. This hierarchy of acquisition applies to both ser and estar.

In the same vein, Holtheuer (2009) studied the spontaneous speech of 11 Spanish-speaking children living in Santiago de Chile (Chile) for which she audio and video recorded two sessions of spontaneous speech with their caregivers of approximately 1 hr duration. One child was recorded only once. The age of the children ranged from 1;10 to 3;07,³ and the total amount of target utterances ranged from 33 for the youngest at the age of 1;11–121 for the child aged 2;11. Furthermore, she also collected utterances containing ILP and SLP verbs produced by the caretakers, which ranged from 84 to 341 utterances. In order to study the distribution of ser and estar, Holtheuer (2009, p. 93f.) made a more fine-grained classification of the predicates consisting of nominals, adjectives, prepositions, adverbs, auxiliary, omissions, clauses, and others. Utterances containing wh-words were also counted. Her study confirmed some of the findings of Sera (1992) in that nominals are exclusively used with ser and clauses with estar. Estar is used predominantly in the adverbial domain, where 50% of the occurrences of estar belong to this area. With adjectives, the copula that is used the most is ser, as opposed to what was found in the experimental study by Sera (1992).

With regard to target-deviant uses, Holtheuer found four (n = 84, 4.5%) examples of target-deviant use of the copula.⁴ That is, where an estar copula was expected, the children used ser. Interestingly, the children also made errors where estar was expected five times. However, on closer examination of the examples, one may interpret one of them as being a substitute of “hay” there is, as is the case in example 19 (Holtheuer, 2009, p. 124). If this interpretation is correct, the example listed is not a copula error per se, and therefore, it should have been excluded from the analysis. Holtheuer cites a master thesis by López-González (footnote 44) (we have not had access to it), which studied the use of ser and estar by Maria and Koki. Both studied children-made selection errors with ser and estar, whereby Maria makes errors ranging from locatives, noun phrases (NPs), adverbials, and with the progressive, whereas Koki was reported as making only one error, but interestingly, with the progressive as well.

Regarding the use of adjectives with both verbs ser and estar, Sera (1992) also conducted an experiment with 52 Spanish-speaking children of Cuban origin living in Florida. Its objective was to elicit location (of objects and events acting as subjects) and adjectival predicates. The participants were distributed in four age groups ranging from 3;6 to 4;9 (n = 14), 5;4 to 6;11 (n = 10), 7; 1 to 8; 11 (n=18), and 9;6 to 11;1 (n = 10). Sera shows that the youngest children perform over the 60% level with adjectives, and performance improves overall as the children become older. As opposed to this, all children perform at ceiling with estar when objects have to be located, but the performance of all children is well below chance (between 20% and 30%) when it comes to using ser with events. Sera admits that it is not clear at all why children overuse estar with eventive nouns where ser is expected. On the basis of these results, Sera makes the interesting observation that children choose the verb ser or estar according to the semantic or syntactic properties of the predicate and not according to the semantic properties of the subject. In order to ascertain whether adults perform with 100% accuracy, Sera conducted an experiment with adult native speakers of Spanish to find out whether this group of speakers would accurately choose ser with locations of events and estar with location of objects. Indeed, all native speakers of Spanish performed the task with estar at ceiling except with events, since they used ser with locations of events in 81% of the contexts.

Holtheuer (2009, p. 157) criticized the limitations of Sera’s study regarding adjective use and conducted a qualitative analysis on her own data.⁵ She confirmed Sera’s findings in that children use very few adjectives with ser and estar and even fewer adjectives that are allowed with both. However, she argues that this does not represent any evidence that the children have not acquired the correct grammatical representation of ser and estar, as even adults use very few adjectives with both copulas in her study. Open-scale adjectives are used more frequently with ser, while the copula estar shows preference for neither open scale nor absolute adjectives. An interesting result is that the children studied use eventive participles exclusively with estar. The few adjectives used with both copulas are color adjectives and loco “crazy” as well as fea “ugly,” at later ages. This, according to Holtheuer, does not necessarily mean that younger children are not aware of this distinction, since even adults in her study make use of very few adjectives that can be combined with both copulas. The low use of bimodal adjectives is attributed to the methodology used in her study, that is, data collection in a naturalistic setting. One of the issues that Holtheuer raises is whether the input contains enough positive evidence that ser and estar behave differently in a number of ways. A thorough analysis of the caregivers’ input shows that there is no contrastive evidence that would help children to figure out the target-like use of the copula at stake. Holtheuer (2009, p. 162) found one single example of this sort in the entire corpus reproduced here:

(1) Sus patas que eran blancas estaban entonces coloradas.

Their feet that were (ser) white were (estar) then red.

In their study of one bilingual Spanish–English child, Silva-Corvalán and Montanari (2008) investigated whether the child’s input mirrored the parents’ input. Structural and conceptual frames, as well as predicate types, were analyzed. As for the structural frames, they distributed the different uses of ser in copulatives, progressives, and passives in the case of ser and in copulatives, estar + PP and progressives with estar. The frequency of use of each of the structural frames was strikingly similar in child and adult speech. The same goes for the conceptual frames these scholars adopted (identification of location, characterization, identification of entity, aspectual auxiliary, identification of possessor, identification of manner, and identification of goal). Even in the predicate types (Adj. or NP, Q-word, adverb, PP, personal pronoun, zero predicate, Prep + Q-word), adults and children perform similarly. However, children and adults do not demonstrate identical frequencies in all types of structures, which Silva-Corvalán and Montanari (2008) attribute to interactional practices.

Holtheuer (2009) argues that children may exploit other distributional facts in the adults’ input in order to infer that Spanish has two copulas that are not interchangeable, that is, ser is combined with nominals and estar with PPs. Interestingly, she argues that the adjectives ending in -ado/-ada are the ones that help children to acquire the grammatical properties ascribed to ser and estar, since children never use participles with ser. Holtheuer analyzed the utterances of the caregivers and came to the following conclusion: The preposition de “of” is used 88% of the times with ser, and 84% of the occurrences with prepositions with estar are distributed between con “with” and en “in.” Holtheuer (2009, p. 112ff.) concludes that children as young as 4 use prepositions in an adult-like fashion, and therefore, prepositions may act as syntactic bootstraps, but the semantics and frequency information in the adult’s input is also important when acquiring ser and estar. Holtheuer also analyzed the total amount of errors (160 out of 558) the children studied made, and divided it into errors of omission, commission (erroneous copula selection), morphosyntax (agreement), and semantics–pragmatics. The percentage of errors among children ranges from 4.1% to 57%, which means that there is a wide variation as to the amount of errors the individual children make. She makes the observation that the morphosyntactic errors (subject–verb agreement) are by and large the most frequent ones followed by omission errors. We will not focus on this observation any further because subject–verb agreement errors do not pertain to copula selection per se or omissions. They have not been studied in the present article. Moreover, and judging by the examples cited by Holtheuer (2009, p. 197, examples 10a and b), the so-called semantic–pragmatic errors seem to be more discourse errors that can occur with any other verb and, hence, are not typical of copula at all.

Regarding negative evidence, Holtheuer (2009, p. 207ff.) analyzed the corrections of the caregivers. She makes the observation that all parents but one corrected their children, the rate of corrections ranging from 25.7% to 85.7%. Ser and estar are corrected equally frequently 51% versus 49% and that the more children receive corrective feedback, the fewer errors they make.

A number of other scholars have made the observation that children acquiring Spanish omit both copulas to different extents (e.g. Silva-Corvalán & Montanari, 2008), although Liceras, Fernández Fuertes, and Alba de la Fuente (2012) qualify this phenomenon as nonrobust. Hence, some of the available studies on early language acquisition have focused on the issue of whether there is a stage of copula omission. In this line of investigation, Liceras et al. (2012) studied two bilingual Spanish–English twins’ early language development (from the Ferfulice corpus available in CHILDES). Based on a study of monolingual English children by Becker (2000, 2004) who made the observation that her subjects omitted more copulative verbs with ILP than with SLP, the scholars set out to investigate the question of whether the Spanish dual system influenced the acquisition of the English copula system in any significant way. Liceras et al. (2012) showed that copula omission is significantly lower in Spanish–English bilinguals and that no significant difference between ILP and SLP could be found. Also in line with the investigation carried out by Becker (2004), Liceras et al. (2012) discarded utterance length taken as a processing load as a possible explanation for copula omission in the two bilingual Spanish–English twins. They, however, calculated the mean length of utterance (MLU)(measured in words) for the utterances that contain copulas.

As far as comprehension is concerned, Schmitt and Miller (2007) show that children as old as 4–5 years have not yet mastered the comprehension of ser and estar in Spanish. Schmitt and Miller tested 35 children aged from 4;5 to 6;3 (mean age: 5;3) recruited in Mexico City and Punta Arenas (Chile) using an elicitation task. The results show that children at the ages studied treat ser and estar as two different verbs with diverging properties. In a second comprehension experiment, with 24 children from Punta Arenas, Chile, discourse properties were studied using a picture story that depicted a change of state. A cat was born thin but became fat because of magic food. This experiment was set up to investigate the impact of discourse information, that is, the pragmatic implicatures. The results show that children and adults perform equally target like with the verb estar but show statistically significant differences with the verb ser. Schmitt and Miller attribute the differences observed between both experiments to the availability, or not, of world knowledge. If the child made his or her decision based on world knowledge, he or she would perform at ceiling and almost adult like; if, on the contrary, the child relies on pragmatic information for copula choice, he or she seemed to assume that ser can be used in more contexts.

The pragmatic interface has been considered the locus of crosslinguistic influence in bilingual child acquisition by Hulk and Müller (2000) and Müller and Hulk (2001). In particular, these scholars have suggested that crosslinguistic influence from one language to the other needs to meet certain conditions, and interface areas are particularly vulnerable. According to Hulk and Müller (2000) and Müller and Hulk (2001), crosslinguistic influence must be due to internal language conditions provided that the following two requirements are complied with (a) crosslinguistic influence occurs at the interface between two modules of grammar and, more particularly, at the interface between pragmatics and syntax in the so-called C-domain and (b) syntactic crosslinguistic influence occurs only if language A has a syntactic construction that may seem to allow more than one syntactic analysis and, at the same time, language B contains evidence for one of these two possible analyses. That is to say, there has to be a certain overlap of the two systems at the surface level. The topic studied in this article only meets condition (b) as will be shown in section “Theoretical issues” and is, therefore, not a good candidate for crosslinguistic influence. But this topic is worth pursuing, since the present study is the first one to analyze the acquisition of the counterparts of ser izan and estar egon.

The studies reviewed serve as a benchmark since the Basque copula system is almost identical to the Spanish one. Hence, the research questions that have been addressed are the ones that have been discussed in the literature until now for Spanish (see section “Research questions” below). However, the issues of comprehension and negative evidence will not be dealt with because the data were collected in a semidirected naturalistic setting, and the interaction partner took a great deal of care not to correct the participants.

Theoretical issues

All languages, to our knowledge, make the distinction between ILPs and SLPs. The former refers to features that are not restricted in time and the latter to characteristics that have some sort of time restriction, which is a first approximation to the subject at stake. Since most languages only possess one copula verb, this time component is encoded in the predicate (Schmitt & Miller, 2007). However, some languages such as Spanish and Basque possess two copula verbs. It is rather an intricate issue as to how the predicates are distributed across copulas in Basque and Spanish, but the ILPs almost always occur with one copula and the SLPs with the other, so that an almost perfect complementary distribution exists. To start with, we will briefly sketch the theoretical framework adopted in this article based on Spanish data and will then apply it to Basque in the remainder of the section.

Theoretical framework

Traditional grammars of Spanish have assumed that the distribution of ser and estar is constrained by the type of predicate because some adjectives can only be combined with one of the copulas. A number of adjectives, however, are allowed with both verbs. Note, however, that approximately 80% of all adjectives are grammatical with both ser and estar, and additional linguistic and extralinguistic features of the context rule copula choice. As for Basque, we are not aware of any study that has quantified the adjectives.

Basque also possesses two copulas, izan (2b1) and egon (2b2), which, by and large, share the same distribution as ser and estar as will be shown below. In order to simplify the presentation of their features, wherever the distribution of the predicates is identical in Basque and Spanish, examples of Basque will be provided in the remainder of the article, and the adequacy of the adopted approach will be discussed. As it can be seen in (2), some adjectives in Basque such as alai and pozik “happy” are only allowed with one of the two copulas, while their Spanish translation allows both.

(2) a. [Spanish] 1. El presidente no es alegre/gordo

2. El presidente no está alegre/gordo

The president is (ser/estar: 3p.sg.) not happy/fat

b. [Basque] 1. Lehendakaria ez da alai-a⁶/lodi-a

President-art. neg. is (izan: 3p.sg.) happy/fat-art

2. Lehendakaria ez dago poz-ik/lodi-a

President-art neg. is (egon: 3p.sg.) happy-part/fat-art

The examples in (2) show that the adjectives alegre and gordo are allowed both with ser and estar in Spanish and lodi with both egon and izan in Basque. The traditional argument to explain the differing interpretation has been an aspectual one. Although the presence of ser in (2a1) would be understood as a permanent or intrinsic personality trait of such a president, the presence of estar (2a2) would imply a transitory condition/characteristic. Namely, ser denotes an individual-level property (ongoing, imperfect reading), and estar denotes a stage-level property (temporally limited, perfective reading) (Fernández-Leborans, 1999; Luján, 1981; Schmitt, 2005). The same claim would explain the distribution of lodi in Basque. Zagona (2009) argues that the differences in interpretation observed in ser and estar with some adjectives (3a and b) cannot be extended to other predicates and, hence, cannot be explained in terms of aspectual restriction (3a and b) because ser and estar are not temporally constrained as the following examples illustrate (3c). The same argument holds for Basque. As a matter of fact, the copulas ser and izan can take the past and permanent locations, which are also permitted with estar and egon. The same argument goes for Basque.

(3) a. Temporally unrestricted [Spanish] Itziar es [_AP <Itziar > linda ] (AP)

[Basque] Itziar [_AP <Itziar > polita] da

Itziar is beautiful

b. Temporally restricted [Spanish] Itziar está [_AspPAsp[_AP<Itziar> linda]] (AspectP)

[Basque] Itziar [_AP <Itziar > polita] da

Itziar looks beautiful

c.1. [Spanish] Anguita fue profesor

Anguita was a teacher

El Museo Balenciaga está en Guetaria

The Museum Balenciaga is in Getaria

2. [Basque] Anguita irakaslea zen

Anguita was a teacher

Balentziaga Museoa Getarian dago

The Museum Balenciaga is in Getaria

A recent approach by Zagona (2009) goes the opposite way; on the one hand, she claims that the structural architecture of the verbs at stake determines which predicates can occur with one or the other verb and, on the other hand, that the temporal–aspectual interpretation of the utterance is linked to the prepositional feature [P] of the estar verbs. Let us introduce this proposal for Spanish in order to evaluate whether it explains the facts in Basque.

Zagona (2009) makes two important claims in her proposal. First, she claims that the crucial distinction is that verbs of the ser type are atemporal and that verbs of the estar type have a spatial-temporal component that constrains the type of predicate. This is due, she argues, because estar has an uninterpretable P(repositional) feature that ser lacks as illustrated in (4) and adapted from Zagona (2009). Note that Basque expresses temporal and local relations by using postpositions. Zagona’s line of argument applies perfectly to Basque because this structural property is linked to the head parameter and not to the copula selection itself. Other features of the verb ser will be discussed below.

(4) [Spanish] El Guggenheim está en/*a Bilbao

El Guggenheim es un museo

[Basque] Guggenheim Bilbo-n/*Bilbo-ra dago

Guggenheim museoa da

The Guggenheim is in Bilbao

The Guggenheim is a museum

Note that the preposition selected by estar is the stative preposition en in Spanish; in Basque, egon selects a stative postposition -n. Path prepositions can neither be combined with egon nor with estar.

Thus, the selective mechanism for estar is claimed to be an uninterpretable prepositional feature, [uP], as illustrated in (5) and (6a and b), which has to check its features against the complement; Zagona argues that the distribution of a number of predicates can be derived with the structure outlined above, and the two main subclasses of PPs are those in (5), that is, (a) temporal location that comes to represent the aspect phrase and (b) spatial location. Hence, Zagona (2009, p. 5) proposes the following checking constellation:

(5) estar: [ v, uP ]

[uP] checked by:

a. AspectP (temporal P)

b. PP (spatial and other PPs)

Hence, examples in (4) correspond to the checking constellation in (5b). Estar-type copulas also occur with participles (6), which are adjective like since they can be coordinated with them.

(6) [Spanish] Los exámenes están corregidos

[Basque] Azterketak zuzendu-ta daude

The exams are marked

[Spanish] Los exámenes son corregidos (por el profesor)

[Basque] Azterketak (irakasleak) zuzenduak dira

The exams are marked (by the teacher)

The participles with ser have a passive interpretation, and the ones with estar have a result state, in Spanish (Zagona, 2009, p. 11). The state reading applies for Basque as well. A number of words in Basque, that is adjectives or adverbs in De Rijk’s terms, take egon. The most important for the purposes of this article are lo “asleep” and esna, “awake,” gertu “close” and gaixo “sick,” and haserre “angry” and prest “ready.” Note that the word lo is classified as noun “sleep” and as a predicative “asleep” by De Rijk. In any case, lo always takes egon and the same applies to bizi. Bizi is classified either as adjective or as a noun. The common denominator to these words is that they are participle like and denote states or change of state, hence having a temporal dimension. For the purposes of this article, we will refer to them as adjectival predicates.⁷ Some of these temporally marked words are built regularly by adding -ta, such as apurtuta “broken,” erosita “bought,” salduta “sold,” and are known as participles. Other words are marked with the stative suffix -ik, bizirik “alive,” gaixorik “sick.” Note however, that both gaixo “sick” and gaixorik “sick” are allowed in free variation. According to Zagona, adjectival participles are similar to stative locatives and have the following structure in Spanish. This structure explains the Basque participles as well, with the only exception that the head is on the right.

In sum, participles and stative prepositions in Spanish and postpositions in Basque can be explained by the checking constellation in (5). That is to say, the complement is a Preposition whose specifier hosts aspect with the value [-Fin] and checks its features against the participle in the complement. Moreover, Zagona (2009) argues that the progressive in Spanish can be derived with the structure outlined in (5). In order to discuss the progressive, the complex Tense Phrase (TP) (8), which includes three layers of time, utterance time (UT-T), aspect time (AST-T), and event time (EV-T) must be introduced first.

According to Zagona, TP assigns temporal thematic arguments to external and internal arguments. Based on the study by Demirdache and Uribe-Etxebarria (2000), Zagona (2009, p. 5) argues that the inflectional phrase (IP) structure has three temporal layers, where times and “temporal relations (are) determined by tense and aspect” (9a and b).

(9) a. Times:

i. UT-T = utterance time: the external argument of Tense

ii. AST-T = Assertion time: external argument of Asp

iii. EV-T = the event interval

b. Temporal relations determined by Tense and Aspect:

BEFORE, AFTER, COINCIDENCE

UT-T is the time of speaking. The relation between assertion time and EV-T gives rise to different temporal readings. That is, the assertion time can coincide with the UT-T or can lie behind or after it, which is not relevant to the present article. The aspectual relations are more central to this article.

Zagona (2009, p. 5) makes one crucial assumption that “tense and aspect are temporal prepositional heads” whose values are [+finite] for tense and [−finite] for aspect and whose function is to locate the event in the time continuum. Moreover, Zagona (2009, p. 15) argues that “aspectual morphology has the effect of excluding event boundaries from the interpretation.”

The crucial assumption is that, “where the Viewpoint head is imperfective (for example in progressives), the AST-T corresponds to the internal interval of the event, excluding its boundaries” (Zagona, 2009, p. 5). If it is perfective, as in the participles, the boundaries are included. The progressive, as with any of the perfective tenses, merges with the whole event structure according to Zagona, whereas “copular estar merges temporally simple phrases” (Zagona, 2009, p. 15), whereby Viewpoint aspect is the crucial projection. Furthermore, Zagona (2009, p. 15) claims that the progressive has a [Path] argument “involving a change of state,” which is in apparent contradiction with her proposal. Crucially, she argues that the progressive morphology in the structure where it is merged has the effect of excluding the event boundaries understood as temporal prepositions (aspect heads). That is, “the [Path] is not temporally anchored because its onset is checked by the imperfective [-finite] Viewpoint head” (Zagona, 2009, p. 16). The structure outlined above explains the progressive by merging the aspect phrase with the TP. Let us now compare the Spanish and Basque progressives in order to evaluate whether this approach explains Basque data.

While Spanish requires the verb estar for the progressive (10), Basque requires izan (13),⁸ when the progressive is built regularly. The progressive is built with estar, and the main verb is followed by the suffix -ndo in Spanish. Its Basque counterpart is built with the main verb and the suffix -ten + ari + izan (14). Just a few verbs, which are conjugated with both izan (13) and egon (12), that crucially take locatives may have a progressive reading such as to sleep, “lotan egon” (12) and “lanean ari” (13). Similar structures are also known in Spanish (11, taken form Zagona, 2009, p. 14) and will be ignored here.

(10) [Spanish] Itziar está durmiendo

Itziar is sleeping

(11) [Spanish] Su carta está al llegar

The letter is on the way

(12) [Basque] Itziar lo-tan/izardi-tan [dago]_EGON

Itziar is sleeping/sweating

(13) [Basque] Itziar lan-(e)an ari [da] _IZAN

Itziar is working

(14) [Basque] a. Itziar eskutitzak irakur-tzen ari [da]_IZAN

Itziar is reading the letters

b. Itziar eskutitzak irakur-tzen aritu-ko [da]_IZAN

Itziar will be reading the letters

c. Itziar eskutitzak irakur-tzen ari-tzen [da]_IZAN

Itziar usually reads the letters

In fact, Zagona’s approach explains the progressive in (12). The words lo “sleep” or izardi “sweat” (12), which are difficult to classify as verbs or nouns, take a locative suffix -tan and must be combined with egon, which is in line with Zagona’s approach. However, Basque also allows izan with a locative suffix as in (13) with a progressive reading, which is, at first sight, a contradiction. These two “verbs” lan and lo differ in one important point, while lo denotes a state excluding the event boundaries, lan denotes an activity excluding the event boundaries. Hence, the exclusion of the event boundaries would not result in the checking by an imperfective [−finite] viewpoint head and, therefore, in the selection of egon instead of izan in all instances, as seems to be the case. However, the selection of izan seems to be determined by the presence of ari with verbs of activity (14a). Note that ari behaves like a main verb since it can take progressive (14c) and future affixes (14b) and also has some modal properties. The important point is that ari, as with other modal verbs, determines the auxiliary selection irrespective of the main verb involved. Thus, examples as (14a–14c) cannot be explained with a copula approach. Ortiz De Urbina (2003, p. 286) analyzes clauses such as “eskutitza irakurtzen” (14) as clausal complements of ari under the rubric of “locative nominalizations” and, therefore, a modal-like approach is more appropriate. One of the arguments adduced by Ortiz De Urbina (2003, p. 28) is that the arguments of the “lower” verb are marked in the auxiliary. Along the same lines, examples such as (14) are better explained with a modal approach. In sum, the issue of ari is not a copula issue per se. Note, however, that (12) and (13) are superficially very similar.

To sum up, the uninterpretable syntactical [uP] features on estar account for all the uses of estar (locatives, participles, adjectives, and progressive). Such an uninterpretable feature would be checked by features of the complement, even though interpretable (semantic) features may be at work as well.⁹ We have shown above that Zagona’s approach adequately explains the Basque facts on the copula (locatives, so-called participles and adjectives) and progressive egon. Let us now look at some distributional properties of Spanish ser and Basque izan.

Nominal (15–16) and path (17) predicates are only allowed with ser in Spanish and with izan in Basque.

(15) [Spanish] El león es (*está) un animal feroz (Nominal complement)¹⁰

The lion is (ser: 3p.sg.) a animal fierce

(16) [Basque] Leoia abere izua da (*dago)

Lion animal fierce is (izan: 3p.sg.)

Only ser/izan is also permitted with change of location (path) PPs as in (17) and with change of state Participle Phrases (PrtPs) as in (18):

(17) [Spanish] Este paquete es (*está) de/para la escuela (Goal/Origin/Directional PP)

[Basque] Fardel hau eskolakoa/eskolarako da (*dago)

Parcel this from school/to school is (ser/izan: 3p.sg.)

(18) [Spanish] Este libro fue (*estuvo) escrito por Paula

[Basque] Liburu hau Paulak idatzia da¹¹ (*dago)

Book this Paula written is (ser/izan: 3p.sg.)

A number of nouns denoting states must be combined with izan as in (20 and 21) (De Rijk, 2008, p. 154; Salaburu & Kintana, 1984, p. 108; Zabala, 2003, p. 426ff.), for example, bizi “life,” bizi izan “to live,” beldur “fear,” beldur izan “to be afraid,” egarri “thirst,” egarri izan “to be thirsty,” haserre “anger,” haserre izan “to be angry,” mintzo “speech,” mintzo izan “to speak,” logale, “sleepiness,” and logale izan “to be sleepy.” Interestingly, these nouns must appear without the article -a, although all other nouns and adjectives in predicative position or closing an NP require the article. However, Basque has some peculiarities that are worth mentioning here. Basque has other uses of both copulas, that is, izan/egon, that are not attested to in Spanish and, hence, require closer examination. The definite article that is a suffix has a controversial status (Artiagoitia, 2002) and is indicative of whether the predicate has to be considered as stage level (18a) or individual level (19b).

(19) [Basque] a. Lorea sendagile-Ø dago

Lorea doctor is (izan: 3p.sg.)

Lorea is (working as) a doctor

b. Lorea sendagile-a da (izan: 3p.sg.)

Lorea doctor-the/a is

Lorea is a doctor

(20) [Basque] Lehendakaria Gasteizen bizi da (izan: 3p.sg.)

President the Gasteiz-in life is

The president lives in Gasteiz

(21) [Basque] Lehendakaria Legebiltzarrean mintzo da (izan: 3p.sg.)

President the Parliament-in speech is

The president is speaking in the Parliament

Taking Zagona’s (2009) proposal into account described in the previous section, Table 1 captures the distribution of copula choice in Spanish and Basque along the lines of Zagona (2009). As far as adjectives are concerned, both Basque and Spanish have bimodal adjectives, that is, adjectives that can appear with ser/izan and estar/egon. Zagona’s approach nicely explains their diverging readings. The uninterpretable feature [P] in estar/egon, which has an aspectual component, accounts for the temporally restricted interpretation of the estar/izan predicates, that is, for example, the adjective. The adjectival predicates that appear with ser/izan lack [P] and are, therefore, temporally unrestricted.

Table 1.

Distribution of ser and estar in Basque and Spanish.

Complement type	Ser	Izan	Estar	Egon
PPs of location	*	√	√	√
Aspectually delimited AP, PrtP	*	*	√	√
Path (change of location) PPs	√	√	*	*
Change of state PrtP	√	√	*	*
Relational PPs, APs	√	√	*	*
Nominal and clauses	√	√	*	*
Relational APs	*	*	*	*

PP: prepositional phrase; AP: aspect phrase.

In sum, Basque and Spanish share identical distributions and the analysis proposed by Zagona (2009) for copula fits perfectly into Basque with the only exception that Basque is a left-branching language and, hence, all heads have to be represented to the right of the complement. [Path] arguments are combined with ser/izan, while [locative] arguments are combined with estar/egon. However, two differences are worth mentioning:

Basque regular progressive is built with izan ari and not with egon as in Spanish and has been analyzed as a complex construction containing a clausal complement.

However, some progressive readings are built with nouns (verbs) + locative postposition + egon, a structure that very much resembles the Spanish construction (see (12) above).

These two distributional properties of Basque may pose a problem for the acquisition of copula choice in Basque by Spanish-dominant children.

Research questions

Following the linguistic analysis provided above, literature review on the topic, and bearing in mind that this is the first study to deal with the acquisition of copula in Basque, the following research questions are addressed:

Is copula choice target like in Basque-dominant and Spanish-dominant children? From the literature reviewed for Spanish and from the framework we adhere to, we predict that copula choice is target-like in both groups since the early L2 learners of Basque have the dichotomy SL and IL predicate in their mother tongue.

Do Basque-dominant and Spanish-dominant children use the egon progressive in a target-like manner? If not, is this a case of crosslinguistic influence?

Regarding adjectives and more specifically, bimodal adjectives, can the findings of Sera (1992) and Holtheuer (2009) be supported in this study?

Does the theoretical model proposed by Zagona (2009) explain the data? More specifically, are interpretable features (aspectual features) acquired earlier as uninterpretable ([uP]) features?

Methodology

In what follows, we will briefly discuss some methodological issues.

Participants

For the present study, we used the cross-sectional data by Luque available in the CHILDES database (http://childes.psy.cmu.edu/). She interviewed 38 subjects in Basque in a school in the years 1995–1997 in the area of Durango (Bizkaia), an area that belongs to the Western Basque dialects.¹² The data collected included various tasks of which we have used the spontaneous conversation and a storytelling task, which lasted for about 20 min per participant. All the participants attended a school (D curriculum¹³) where the vehicular language was exclusively Basque. Of the total, 19 subjects were either Basque speakers or Basque dominant, whereas 19 subjects were either Spanish speaking or Spanish dominant. Of the five original groups, Luque and Azurmendi grouped the Basque dominant and Spanish dominant together resulting in an even distribution of 19 participants in each group. Unfortunately, the authors do not provide any information as to how they came up with this classification. Neither did they state the age of the children when the recordings were made. The investigators visited the subjects five times and collected data in only four of the sessions. One of their main results reported by Luque and Azurmendi (1997) is that Spanish-dominant children increase their vocabulary proficiency from session 1 to session 4, whereas the Basque-dominant children have a stable profile. Furthermore, Spanish-dominant children who have been schooled from age 2 perform better than those that have been schooled from age 3. In view of these results, we have adhered to the grouping made by Luque and Azurmendi (1997). Below is the list of transcripts that were taken into account for this article (see Table 2).¹⁴

Table 2.

Basque data selection.

	Basque-dominant bilinguals	Spanish-dominant bilinguals
First session	9	12
Second session	17	19
Third session	16	19

Data selection

For our purposes, we have analyzed the Basque data in the Luque corpus in CHILDES, and we have separated copula structures with egon (SL-be) and izan (IL-be). We excluded imitations from the adults and self-repetitions as well as questions, in particular, where questions with egon (SL-be) and who questions with izan (IL-be). We divided the predicates in nominal predicates, locative predicates, gerunds, ad hoc progressives, adjectival predicates, and predicative adjectives. The rationale into this classification was twofold: on the one hand, we wished to know the frequencies of the different predicates; on the other hand, we wanted to compare our results mainly with the results from the literature (Holtheuer, 2009; Liceras et al., 2012; Sera, 1992; Silva-Corvalán & Montanari, 2008).

We will briefly explain how the data were counted and coded. If locative PPs were omitted (dago otsoa “(there) is a wolf”) and it could be inferred from the context that a locative had been intended, the copulative utterance was counted as a PP. Questions referring to the predicates have been excluded from the counts, but total questions have been included. We have added a label with the name of “*progressives.” Under this label, structures that are literal translations of Spanish utterances such as “*dago jaten,” which corresponds to “está comiendo” and are not grammatical in Basque, have been included. Omissions were not counted since there are a number of problems with omissions that have been extensively discussed by Holtheuer (2009). First of all, it is difficult to determine whether a child intended to produce a copula or any other verb even if oral and visual information is available (Holtheuer, 2009, p. 95ff.). In Basque, the missing verb may be a formally identical to a bivalent verb where an allocutive form is incorporated (22) in which case it would not count as copula.

(22) [Basque] Hau gorria duzu

This (A3s) red(A3s) have (have, A3s/E2s)

This is red

Sometimes, coarticulation may pose a challenge for determining whether a copula is missing ése el caballo “That is a horse” versus es el caballo “It is a horse” as noted by Holtheuer (2009, p. 99). And once it is out of the question that the copula is missing, it is impossible to figure out whether an ILP or a SLP was intended and, therefore, whether IL or SL predicates are more prone to omissions. To decide that a copula is missing solely on the basis of a transcript is impossible. More importantly, copula omissions have not been counted since the children are at an age in which copula omission are target like in adult Basque. Hence, we refrained from counting omissions.

Results

In what follows, we will present the results divided into ILPs and SLPs by both the Basque-dominant and the Spanish-dominant groups. Before we present our own data, we will refer to some observations made by the research team who collected the data used in the present article and designed the data collection in order to study vocabulary growth. Although vocabulary growth is not the subject of the present article, some observations made by Luque and Azurmendi (1997) are in order here. Luque and Azurmendi note that bilinguals from birth show no or very little variation in the language they use to answer to the questions. As opposed to this, Spanish-dominant children show a steady increase in the vocabulary used as shown by Figure 2. Moreover, the age of schooling has a positive effect on the proficiency attained such that the earlier the children are exposed to Basque, the better their performance in both groups. Basque-dominant and Spanish-dominant bilinguals show similar patterns of vocabulary use along the entire study, their only difference being quantitative. This observation fits in perfectly with our own results since Spanish-dominant children overall use far fewer copulas as Figures 3 to 6 show.

Figure 2.

Vocabulary growth according to Luque and Azurmendi (1997).

Figure 3.

Individual-level predicates (izan) in three sessions (1, 2, and 3) by the Basque-dominant children.

Figure 4.

Individual-level predicates (izan) in three sessions (1, 2, and 3) by the Spanish-dominant children.

Figure 5.

Stage-level predicates in three sessions (1, 2, and 3) by the bilingual Basque-dominant children.

Figure 6.

Stage-level predicates in three sessions (1, 2, and 3) by the Spanish-dominant children.

Let us first look at the ILPs. The predicates have been classified according to five categories: predicative nominals, predicative adjectives, directional postpositions, other constructions, and other postpositions based on the theoretical framework adopted here. The first observation that comes to mind is that the predicative nominals are the most frequent category followed by the predicative adjectives. Predicate nominals are attested from the very first recording onward (23 and 24) in both groups of children.

(23) [Basque dominant] Hori da bere aitatxo

That is his/her father

(24) [Spanish dominant] Ez da amama

(That) is not granny

Directional postpositions and other postpositions are used but have no overall weight whatsoever. It is noteworthy that [path] postpositions (25) are used exclusively by Basque-dominant children.

(25) [Basque dominant] Zein zenbakia da zure etxekoa?

Which number is from our home

What is your telephone number?

Predicative adjectives (26 and 27) are rather numerous in the Basque-dominant group but not as frequent in the Spanish-dominant group. Predicative adjectives belong to three classes of adjectives, that is, measures, modals, and colors. Interestingly, almost all adjectives can be used with both izan and egon (Basque dominant 31/32, 1 with izan; Spanish dominant 5/6, 1 with egon), although one adjective per group can be used exclusively with izan or egon. That is, no adjective has been found that has been used with both izan and egon.

(26) [Basque dominant] Ze zara txikitxue

Because (you) are small

Makina da polite

The machine is nice

Hau da baltza

This is black

Baina da gaiztoa

But (it) is bad

(27) [Spanish dominant] Ze txikiak dira, begira

Because (they) are small, look

Pastelak gorriak dira

The biscuits are red

Ze da gaiztoa

Because (it) is bad

A few unclassifiable utterances have been grouped under the label “other” and will not be analyzed any further. Children whose home language is Basque do not change their user profile over time. That is to say, they use around the same number of predicative nominals in the first and following sessions, although a slight increase can be observed. The same holds for the other categories.

The category that is used most is the category of locatives (28 and 29) with the verb egon, and its use increases over time, especially in the Spanish-dominant participants.

(28) [Basque dominant] Ze egon da amama otsoaren tripan barruan.

Because Granny was in the wolf’s belly

(29) [Spanish dominant] Basoan dau otsoa

The wolf is in the woods

Predicative adjectives (30 and 31) are not very numerous with egon but are more frequent in the Basque-dominant than the Spanish-dominant group, and their use increases over time in both groups. Interestingly, most adjectives can be used with both copulas. Only gaizki “unwell” and antzekoa “similar” can be used with egon and izan, respectively.

(30) [Basque dominant] Edurre dau hotz hotza

The snow is very cold

(31) [Spanish dominant] Behin joan da amamaren etxera ikustera ze dago gaizki

Once upon a time, (she) went to visit granny because she is sick

Adjectival predicates (32–34) are far more numerous than the predicative adjectives with egon and are characterized by the presence of the stative suffix -(r)ik or -ta/da if they are regular adjectival participles.

(32) [Basque dominant] Gaizorik dago

(He) is sick

(33) [Spanish dominant] Dau pozik

(He) is happy

(34) [Basque dominant] Zergatik otsoa dau hilda

Because the wolf is dead

Pegatute dau hau

(This) is glued

Let us now discuss some qualitative issues that concern our own study. The most remarkable observation concerns the use of progressives; target-like structures as well as target-deviant structures are used by both groups of children. That is, the latter are a more or less direct translation of Spanish constructions. Recall that both Basque and Spanish possess the progressive form.

The target-like progressive (35 and 36) with the locative is not very frequent in any of the groups but far more frequent in the Basque-dominant group.

(35) [Basque dominant] Negarrez dau

(She) is crying

(36) [Spanish dominant] Baina dau lotan

(She) is sleeping

Only two Spanish-dominant children (37) use each of the examples of target-deviant progressive (38a and b). However, one of the examples could be a target-like locative with egon (38b). It is not possible to state which one of the forms is intended.

(37) [Basque dominant] Hartzen loreak dau

(She) is taking flowers

(38) [Spanish dominant] a. Dau jaten

He is eating

b. Lanean dago

(He) is at work/working

Basque-dominant participants are the ones who use target-deviant progressives (37) most. Note that not all Basque-dominant participants use progressives, and target-like progressives are a minority (4, n = 17). Interestingly, those participants who use target-like progressives also use target-deviant progressives. Some participants use only target-deviant progressives. One of the children in the last session uses example (39), which is not a copula construction but is a case of substitution of egon for izan in a progressive context.

(39) aitatxo dau lan eiten. (Target-like “aitatxo lan eiten ari da)

Since all of the children speak Spanish, it is plausible to assume that this progressive construction is a calque of Spanish, especially because none of the children use the target-like progressive with ari. In order to exclude the possibility that the present distribution is determined by the characteristics of the input, we searched the input of the observer. The observer uses the target-like progressive with ari only on two occasions with two children. Interestingly, she also uses the target-deviant progressive with egon with 3 Basque children out of 38. Note that nine Basque-dominant children and one Spanish-dominant child use the target-deviant progressive with egon, and, crucially, three of the children have been exposed to the target-deviant progressive in the recording. Since we do not know whether the investigator interacted with the children outside the recording and if she did, whether she used target-deviant progressives, it cannot be excluded that the use of the target-deviant progressive is a consequence of the data collection itself. It is well known that such progressive forms are rather frequent in second-language learners of Basque that impact the input the younger generation is exposed to.

In sum, the distribution of the predicates of izan and egon very much resembles the distribution of predicates in adult and in child Spanish. That is, the verb izan is combined with nominal predicates and [path] postpositions as well as with a number of adjectives of measure and color as well as modals. Neither locatives nor participles are found with izan. The verb egon, by contrast, is used with locative postpositional phrases and adjectival predicates, including stative -(r)ik and -ta/-da participles as well as target-deviant progressives. Neither [path] postpositional phrases nor nominal predicates are found with egon. However, half of the (9/19) Basque-dominant children use target-deviant progressives in Basque, a result that we had expected to find with Spanish-dominant children.

Discussion and conclusions

This article set out to study the use of two Basque copulas, egon (SL-be) and izan (IL-be), by a Spanish-dominant and a Basque-dominant group of children raised in the Basque Country. Its purpose was to analyze whether both groups choose the copula in a target-like manner, whether they show similar patterns of use and whether the patterns of use change over time.

As to the first issue (namely, is the copula choice in Basque target like in Basque-dominant and Spanish-dominant children?), which has been addressed in this article, it has been clearly shown that copula choice is always target like in Basque for both groups of participants. One possible explanation is that the group of children has already acquired the correct use of the copula, and any intermediate stages where copula choice may have been target deviant have been surmounted. However, Sera (1992) studied 2 children who might be younger than the ones in this study and made the same observation for Spanish. The same goes for Holtheuer’s (2009) study, where 10 Chilean children were studied. According to these studies, children do not make many copula choice errors even at very early stages. If this is so, it seems more plausible to assume that copula choice does not pose major problems to children acquiring Basque in a bilingual setting and that the acquisition of a similar dichotomy (egon/ser for SL predicates and izan/estar for IL predicates for Spanish-dominant children) is error free, presumably, because Spanish has a similar dichotomy. In other words, Spanish would boost the acquisition of the dichotomy in Basque. Moreover, it has been shown that the patterns of use in the Basque-dominant and the Spanish-dominant group are identical. The IL verb izan is primarily used with nominal predicates, and the SL egon is used with locatives. In turn, this finding is in line with those of other studies on monolingual Spanish children (Sera, 1992; Holtheuer, 2009). The only differences found between the Basque-dominant and the Spanish-dominant group is quantitative. That is, Spanish-dominant children use fewer copulas overall, partly because their lexicon is rather restricted.

Regarding research question two, that is whether the progressive is used in a target-like manner, three important observations have to be made. To start with, both groups of participants use target-like and target-deviant progressives with egon. However, the Spanish-dominant group makes use of the progressive on only two occasions. It is not implausible to suggest that this group uses an avoidance strategy in order to circumvent the intricacies of this phenomenon in Basque. The third and most important observation is that it is precisely the Basque-dominant participants who make abundant use of target-deviant progressives. As has been shown, the use of target-deviant progressives cannot for sure be attributed to incomplete acquisition or to crosslinguistic influence, since the observer herself uses this type of construction with 3 out of 19 children. Nevertheless, it would be interesting to pursue this line of investigation with a population whose caregivers are native speakers of Basque and do not make copula selection errors in order to test the issue of crosslinguistic influence in the sense of Hulk and Müller (2000) and Müller and Hulk (2001). Because only condition B is met, that is, the surface overlap, and no pragmatic interface is involved, we leave this question open for future research.

Regarding research question three, the semantic properties of the adjectives and their distribution, it has been shown that predicative adjectives are frequent with izan but not so numerous with egon, though all belong to the class of colors, measures, and modals very much in line with Sera’s (1992) and Holtheuer’s (1999) results for Spanish. Moreover, both copulas are attested with predicative adjectives; however, according to linguistic description adopted in this article, the majority of adjectives belong to the class of adjectives which can be used with both copulas. Nevertheless, no single adjective has been used with a temporally restrictive reading as well as with a temporally unrestricted one by our participants. This is in line with Sera (1992), who made the observation that most adjectives were used with either ser or estar but not with both (Sera, 1992, p. 416). Only bueno “good” largo “long” and guapo “good looking” appeared with both in her corpus. On a closer look at Sera’s (1992, p. 416) Table 2, which provides an overview of the adjectives, it reveals, on the one hand, that most of the adjectives with estar (CHILDES 13/19, Frog Story 18/23) are participles or words derived from verbs that denote states. On the other hand, most adjectives used with ser (CHILDES 17/18,¹⁵ Frog Story 7/8) and estar (CHILDES 5/5,¹⁶ Frog Story 4/6) can be used with both ser and estar. Holtheuer (2009, p. 157) made a similar observation in her study. In relation to the issue of copula selection with adjectives, our results are consistent with previous findings with monolingual Spanish children that copula selection is almost error free. However, Silva-Corvalán and Montanari (2008), who studied one Spanish–English bilingual, have made the observation that their participant indeed makes copula choice errors. They point out that “few errors of copula choice” (eight tokens of commission errors in 97 predicate adjectives) could be attested (Silva-Corvalán & Montanari, 2008, p. 356). However, they argue that “commission errors in the choice of copula with adjectives may not be explained as a consequence of the influence of English” (Silva-Corvalán & Montanari, 2008, p. 357). It may be the case that the commission errors are not genuinely influenced by English, but their observation is indicative of the fact that, in a bilingual setting, if one of the two languages lacks this lexical distinction in copula constructions (e.g. English), this feature may delay the acquisition of such a distinction in the other language. Coming now to the participles, or adjectival predicates, the data show that this type of predicate is exclusively attested to the copula egon. This finding confirms the results for Spanish in previous studies (Sera, 1992).

Linking the last issue with the question of whether Zagona’s model explains our data, our fourth research question, this study has shown that her model with an uninterpretable feature [P] as a lexical property of the verb egon, which lacks izan, adequately explains our data since no copula selection errors could be attested so far. Although P is an uninterpretable feature, one of the subconditions of checking (5), repeated here as (39) has an interpretable feature, namely, aspect.

(39) estar: [ v, uP ]

[uP] checked by:

a. AspectP (temporal P)

b. PP (spatial and other PPs)

Recall that the perfect participles and gerunds all have an aspectual component in Spanish. In the same vein, the Basque participles with -ta and -(r)ik and some of the progressives with egon are aspectually marked, that is, aspect is an interpretable feature. Locatives are, on the contrary, not temporally restricted and are genuinely uninterpretable. In this sense, the question arises as to whether uninterpretable features are acquired earlier than the interpretable ones. It has been shown that locative postpositions are numerous and present from the very first recording onward. On the contrary, adjectival predicates (-ta and -(r)ik participles) are only used by a handful of participants in the first session. Their use increases over time although the Spanish-dominant group, on the whole, does not use this type of adjectival predicate very often. However, the lack of this predicate in the first session is evidence for the fact that this interpretable feature has not yet been acquired. Hence, uninterpretable features such as P are acquired earlier than such features that have been included under the umbrella of interpretable features. However, a longitudinal study should check this issue since the data have been collected in an experimental data setting.

To conclude, this article is the first one to deal with the issue of copula acquisition in Basque and investigates the use of the copula in Basque by a Spanish-dominant and a Basque-dominant group of children. It has been shown that performance patterns in copula choice are identical in both groups. The only difference found is that Spanish-dominant children tend to speak less and have, therefore, fewer utterances. This is a quantitative issue and not a qualitative issue that we leave for a future study. Interestingly, the Basque-dominant group uses calques from Spanish for the progressive form. A longitudinal study with children whose parents are native speakers of Basque could shed light on the issue of whether the target-deviant progressive is a matter of crosslinguistic influence.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Notes

Author biographies

Pilar Larrañaga is an assistant Professor in Spanish. Her main research interests are Spanish and Basque syntax as well as grammar acquisition by bilingual children and L2 learners of Spanish. She has authored and co-authored a number of books, journal articles and book chapters which have been published in a number of international journals such as International Journal of Bilingualism, First Language and Linguistics. Pedro Guijarro-Fuentes is a Reader in Spanish. His main research interests are in the interdisciplinary field of Spanish applied linguistics, sociolinguistics, psycholinguistics, and bilingualism. His main current research projects include the study of spatial prepositions, object marking and word order by a variety of learners, projects funded by the British Academy, the Art and Humanities Research Council (UK) and the German Research Foundation (DFG). He has authored and co-authored numerous articles and book chapters. His research has appeared in different international, refereed journals such as Language Learning, Bilingualism: Language and Cognition, Cognition, Studies in Second Language Acquisition, Applied Linguistics, Applied Psycholinguistics and First Language amongst others.

References

Artiagoitia

(2002). The functional structure of the Basque noun phrase. In Artiagoitia

Lakarra

Goenaga

(Eds.), Erramu Boneta. Festschrift for Rudolf P.G. de Rijk (Supplements of ASJU, pp. 73–90). Bilbao, Spain: Servicio Editorial de la Universidad del País Vasco.

Becker

(2000). The development of the copula in child English: The lightness of be (Unpublished doctoral dissertation). University of California, Los Angeles, CA.

Becker

(2004). Copula omission is a grammatical reflex. Language Acquisition, 12, 157–167.

Demirdache

Uribe-Etxebarria

(2000). The primitives of temporal relations. In Martin

Michaels

Uriagereka

(Eds.), Step by step: Essays on minimalist syntax in honor of Howard Lasnik (pp. 157–186). Cambridge, MA: MIT Press.

De Rijk

(2008). Standard Basque: A progressive grammar. Cambridge, MA: MIT Press.

Etxepare

(2003). Valency and argument structure in the Basque verb. In Hualde

J. I.

De Urbina

J. O.

(Eds.), A grammar of Basque (pp. 363–446). Berlin, Germany: Mouton de Gruyter.

Fernández-Leborans

M. J.

(1999). La predicación: Las oraciones copulativas [Predication: Copulative sentences]. In Bosque

Demonte

(Eds.), Gramática descriptiva de la lengua española (pp. 2354–2460). Madrid, Spain: Espasa.

Geeslin

Guijarro-Fuentes

(2006). The second language acquisition of variable structures in Spanish by Portuguese speakers. Language Learning, 56, 53–107.

Geeslin

Guijarro-Fuentes

(2007). Exploring Copula choice in Spanish: A look at gender. In Holmquist

Lorenzino

Sayahi

(Eds.), Selected proceedings of WSS3 (pp. 92–102). Somerville, MA: Cascadilla Proceedings Project.

10.

Geeslin

Guijarro-Fuentes

(2008). Variation in contemporary Spanish: The role of bilingualism and language contact in the changing status of Estar. Bilingualism: Language and Cognition, 11, 365–380.

11.

Guijarro-Fuentes

(2008). ¿Ser o estar? Cuestiones sin resolver para el español como lengua extranjera [Ser or estar: unresolved questions for Spanish as a foreign language]. Revista Nebrija, 2–3, 1–6.

12.

Holtheuer

(2009). Learning Ser and Estar: A study of production and comprehension in Chilean Spanish (Unpublished doctoral dissertation). University of Canberra, Australia.

13.

Hulk

Müller

N. N.

(2000). Bilingual first language acquisition at the interface between syntax and pragmatics. Bilingualism: Language and Cognition, 3, 227–244.

14.

Liceras

J. M.

Fernández Fuertes

Alba de la Fuente

(2012). Overt subjects and copula omission in the Spanish and the English grammar of English-Spanish bilinguals: On the locus and directionality of interlinguistic influence. First Language

15.

Luján

(1981). The Spanish copulas as aspectual indicators. Lingua, 54, 165–210.

16.

Luque

M. L. L.

Azurmendi

M. J.

(1997). Adquisición léxica del euskara desde un programa de inmersión total precoz: Un estudio longitudinal [Lexical acquisition of Basque from early immersion program: a longitudinal study]. In: Actas do I Simposio Internacional sobre o bilingüismo (pp. 1091–1102), Universidade de Vigo, Spain.

17.

Morales

Smith

D. J.

(2008). Las imágenes mentales en la adquisición de la gramática de la segunda lengua: El caso de “ser’ y “estar’ en español [Mental images in the grammar acquisition of a second language]. In Revista Nebrija de Lingüística Aplicada a la Enseñanza de Lenguas (Vol. 2–3, pp. 1–24), Universidad Nebrija, Spain.

18.

Müller

Hulk

(2001). Crosslinguistic influence in bilingual language acquisition: Italian and French as recipient languages. Bilingualism: Language and Cognition, 4, 1–21.

19.

Ortiz De Urbina

(2003). Periphrastic constructions. In Hualde

J. I.

Ortiz

Urbina

(Eds.), A grammar of Basque. Berlin, Germany: Mouton de Gruyter, pp. 284–300.

20.

Pérez-Leroux

A. T.

Álvarez

Battersby

(2010). Cuando era feliz, e indocumentado [When I was happy and ignorant]: An aspectual approach to copula choice in L2 Spanish. In Borgonovo

Español-Echevarría

Prévost

(Eds.), Selected proceedings of the 12th Hispanic linguistics symposium (pp. 209–220). Somerville, MA: Cascadilla Proceedings Project.

21.

Salaburu

Kintana

(1984). Euskara. Hobean hobe [Basque: from good to better]. Bilbao, Spain: Servicio Editorial de la Universidad de País Vasco.

22.

Schmitt

(2005). Semi-copulas. In Kempchinsky

Slabokova

(Eds.), Aspectual inquiries (pp. 121–145). Dorchercht, Netherlands: Springer.

23.

Schmitt

Miller

(2007). Making discourse-dependent decisions: The case of the copulas ser and estar in Spanish. Lingua, 117, 1907–1929.

24.

Sera

(1992). To be or to be: Use and acquisition of the Spanish copulas. Journal of Memory and Language, 31, 408–427.

25.

Silva-Corvalán

Montanari

(2008). The acquisition of ser, estar (and be)by a Spanish-English bilingual child: The early stages. Bilingualism: Language and Cognition, 11, 341–360.

26.

Zabala

(2003). Nominal predication: Copulative sentences and secondary predication. In Hualde

J. I.

De Urbina

J. O.

(Eds.), A grammar of Basque (pp. 426–447). Berlin, Germany: Mouton de Gruyter.

27.

Zagona

(2009). Ser and Estar: Phrase structure and aspect [Ser and estar cannot be translated. They must remain as they are]. In Nishida

Russi

(Eds.), Proceedings of 2008 Chronos (pp. 1–20). Austin, TX: University of Texas.