The effects of age and experience with input in the acquisition of clitic climbing in heritage and L2 Spanish

Abstract

Aims and objectives:

This study examines the role of age of first exposure and experience with input in the syntax of English–Spanish bilinguals. More specifically, I examine the production of clitic climbing constructions in Spanish (e.g., lo quiero ver “[I] want to see it” [Kayne; Rizzi]).

Design/methodology:

I compare two experimental groups of heritage speakers of Spanish (n = 16) and L2 Spanish learners (n = 17) from the United States (matched in proficiency) against a group of native Spanish speakers from Mexico (n = 20). A sentence completion task was employed to elicit proclitic sentences across four verbal conditions: two in which clitic climbing is possible but with a higher or lower probability of occurrence, and two in which proclitic placement is agrammatical.

Data and analysis:

Results show a strong tendency to avoid clitic climbing constructions across all testing conditions. Two logistic regression analyses report no differences across all groups, who only favored the proclisis in highly grammaticalized verbs; proficiency among the experimental groups was a predictor in the production of these sentences.

Findings/conclusions:

A different time of onset of first exposure to the second language and a different experience with linguistic input (heritage language acquisition vs L2 acquisition) do not appear to affect the production of complex proclitic sentences in Spanish.

Originality:

Previous studies have employed a few selected periphrastic conditions to elicit clitic climbing constructions among English–Spanish bilinguals. This study further expands the range of verbal matrices employed in the four testing conditions and uses a more controlled testing environment.

Significance:

This study adds adult bilingual data to the ongoing debate on whether an early exposure to the second language results in advantages in the morphosyntactic domain.

Keywords

Bilingualism age input heritage L2 Spanish clitic climbing

Age and linguistic experience in L2 acquisition

This study examines the potential role of age and input in the production of sentences with complex object clitic placement in Spanish by English–Spanish bilinguals. In the United States, heritage speakers and second language (L2) learners are two types of bilingual individuals who share English as their primary (dominant) language and a second or less-dominant language (e.g., Spanish, Chinese, Russian), the latter with distinct developmental trajectories for each group. Generally, heritage speakers are second-generation immigrants who begin learning the minority language from an early age at home. L2 learners are already native speakers of English by the time they start developing a second language (normally at around puberty) and their learning usually takes place in the context of a classroom (Montrul, 2010a; Polinsky, 2006; Polinsky & Kagan, 2007; Silva-Corvalán, 1994; Valdés, 2001). The context of acquisition also determines the type of input these learners receive. Heritage speakers learn the minority language informally through oral interactions with their caregivers, whereas L2 learners learn the language in a formal context through large amounts of enriched visual and aural input. An advantage of learning in this context is that learners are also exposed to the formal aspects of the target language, which forces metalinguistic reflection of what a possible sentence is in the L2. Heritage speakers normally do not receive instruction in the languages spoken at home. Instead, they develop a vernacular, or a more limited variety of the language; not surprisingly, their linguistic knowledge of the most complex aspects tends to be more limited. Another dimension that distinguishes heritage language learners from traditional classroom learners is knowledge of the target language culture. The former belong to an ethnolinguistic community that affords them opportunities to develop sociocultural competence in the minority language, and in some cases, access to informal instruction (Carreira & Kagan, 2011; Montrul, 2010a, 2016); L2 learners, on the contrary, have partial knowledge of the cultural norms of the target language community, and opportunities to practice the second language within these enclaves are usually more limited (e.g., short-term study abroad experiences). In sum, these two groups of bilinguals share native-like competence in the majority language, but they differ in how the less-dominant language develops: the weaker language of classroom learners is the L2, whereas for heritage speakers, it is the first language (L1).

During the last decade, a growing body of research has identified similarities and differences between heritage and second language acquisition in adulthood (Au et al., 2002; Montrul, 2006, 2010a; Montrul et al., 2008). The consequences of not having developed academic literacy and being exposed to the formal aspects of the target language greatly resemble the different stages of L2 development in adulthood (Bruhn de Garavito, 2002; J. H. Kim et al., 2009; Montrul, 2006). For language instructors—specifically those on college campuses—it is important to understand the similarities and differences among these learners to develop appropriate materials and/or curricula that address the pedagogical needs of each group (Bowles, 2011; Brinton et al., 2008; Lynch, 2008; Mikulski, 2010a, 2010b; Potowski et al., 2009; Valdés et al., 2006). Studies that compare the outcomes of early and late exposures to the L2 are key for materials selection, program development, and better assessment procedures (Rothman et al., 2016). In this study, I examine the extent to which the age of first exposure to and experience with input affect the production of variable object clitic placement in Spanish, which is the less-dominant language.

It is generally agreed upon that an early age is one of the most important factors for successful language learning. A considerable amount of research has examined the existence (or lack thereof) of a “critical period” for the acquisition of the first language (Lenneberg, 1967; Penfield & Roberts, 1959), a limited “window of opportunity” during which native-like acquisition of a language is possible. Scholarly work supporting the Critical Period Hypothesis in L2 learning has not yet arrived at a consensus as to the onset and offset ages (Abrahamsson & Hyltenstam, 2009; Krashen, 1973; Lenneberg, 1967; Long, 2005; Pinker, 1994; Seliger, 1978), or even its existence (Birdsong, 1992, 2007; Bongaerts et al., 2000). Studies divide between those that state that native-like attainment of a second language is not possible if the onset time of acquisition exceeds the limits of this window of opportunity (Bley-Vroman, 1990; DeKeyser, 2000; Hawkins & Casillas, 2008; Hawkins & Franceschina, 2004; Scovel, 1988), and those that establish softer age boundaries and allow for external factors such as socialization patterns or differences with the input received (Hartshorne et al., 2018; Herschensohn, 2009; Singleton & Ryan, 2004; Unsworth, 2005; White & Genesee, 1996). The effects of age are evident among speakers of heritage languages, whose exposure to the minority language takes place during childhood. Several studies have confirmed that phonology (e.g., speech production, perception) is positively affected by early exposure (Au et al., 2002; Chang et al., 2011; Gor, 2014; J.-Y. Kim, 2015; Knightly et al., 2003; Lukyanchenko & Gor, 2011; Oh et al., 2003; Saadah, 2011; Yeni-Komshian et al., 2000). However, the advantages of early exposure are rather selective in the morphosyntactic domain: early bilinguals show limitations when producing and interpreting overt subjects (Polinsky & Kagan, 2007; Silva-Corvalán, 1994), long-distance dependencies (J. H. Kim et al., 2009), reflexive pronouns (Polinsky & Kagan, 2007), verb placement (Håkansson, 1995), and relative clauses (O’Grady, Lee & Choo, 2001). Other studies have found advantages for heritage speakers over L2 learners in knowledge of unaccusativity (Montrul, 2005) and in double complementizer questions (Cuza & Frank, 2015); or no differences between low-proficient heritage and L2 groups in syntactic structures interfacing other domains (i.e., semantics or pragmatics) such as clitic left dislocations and differential object marking (Keating et al., 2011; Montrul, 2010b).

A possible explanation for these outcomes is that heritage grammars undergo attrition and/or incomplete acquisition of several linguistic properties. This is thought to be a consequence of an early and intense immersion in the dominant language along with a significant reduction of input in the heritage language (Montrul, 2012). At around 3–4 years of age, basic vocabulary, phonology, syntax, and semantics are usually in place, but to attain native-like competency in the language, it is necessary to maintain a substantial and continuous amount of input during childhood and adolescence (Berman, 2004). In addition, certain grammatical properties have different activation and developmental schedules: some are acquired during childhood, while others are mastered later in school. For example, the expression of subjunctive mood in Spanish emerges by age 2 in monolingual children (López Ornat, 1994), but full command and use is not mastered until 12 years of age (Blake, 1983). When children are exposed to two languages, the amount of input in the home language is substantially reduced, and consequently, it is common to find signs of “attrition” in adulthood: structural properties that were acquired or mastered by age 4–6 are lost, resulting in difficulties with lexical production and interpretation (Schmid & Jarvis, 2014), or speakers use code-switching strategies and lexical borrowings to compensate for gaps during speech (J. Paradis et al., 2011). Another related side effect is “incomplete acquisition,” “partial acquisition,” or “acquisition without mastery” (Montrul, 2002, 2008; O’Grady et al., 2011; Polinsky, 2006; Silva-Corvalán, 2003): children may have acquired grammatical properties of the heritage language, but a reduction in native input after 6–18 years of age prevented them from developing those properties in ways similar to monolingual children, a situation that is commonly found in bilinguals who do not receive schooling in the minority language. Overall, studies on the acquisition of heritage languages suggest that these speakers develop core aspects of their family language, but preference for a more restricted word order is a recurring feature (Montrul, 2010a).

But as Montrul (2016) states, age cannot be the only factor for explaining why both types of bilinguals show distinct strengths and weaknesses depending on the linguistic domain under study. If age alone were responsible for selective advantages in certain domains, these advantages should also be present in language comprehension, not just production. Experience with linguistic input understood as “timing, type, modality, frequency, and amount of exposure to relevant input and use of the L2” varies in these two groups of learners (Montrul et al., 2013, p. 88): heritage speakers do not receive formal instruction in the language, but learn to speak it informally through oral interactions. Consequently, their stronger skills are listening and speaking. In contrast, L2 learners are exposed to the formal register of the language through written input and training with the formal aspects (i.e., explicit grammar rules) of the target language. Oral interaction in the classroom is limited, and for that reason, reading and writing are their stronger skills. If amount and quality of input are key factors in second language development, heritage speakers and L2 learners should show similar performance, given that both types of bilinguals undergo reduced input in the target language. However, the type of cognitive knowledge these bilinguals develop as well as the available ways to access such knowledge matter when examining the linguistic competence of these groups of speakers.

Studies in cognitive neuroscience hypothesize that the mechanisms in the brain responsible for language activate different types of cognitive knowledge, which is stored in different parts of the brain, and it is generated by different types of educational experience (R. Ellis, 2005; N. C. Ellis & Larsen-Freeman, 2006; M. Paradis, 2004, 2009) On one hand, explicit knowledge is declarative knowledge of grammar rules, is stored in the procedural memory, and can be explained (i.e., verbalized); implicit knowledge, on the other hand, is knowledge that cannot be articulated, is stored in the procedural memory, and is executed with automaticity and speed (Bialystok & Hakuta, 1994; R. Ellis, 2004; Han & Ellis, 1998; Hu, 2002; Hulstijn, 2005). These two types of knowledge interact with each other and can be activated using specific tasks. Implicit knowledge can be accessed using elicited imitation tests, timed grammaticality judgments, oral narrative tests, and speaking tests; tasks that exploit explicit knowledge are typically untimed grammaticality judgment tests and metalinguistic knowledge tests (Absi, 2014; Akakura, 2011; R. Ellis, 2006; Norris & Ortega, 2001; Spada & Tomita, 2010). Furthermore, the way the stimuli are presented and how the responses are elicited and measured contribute to making the task either more implicit (relying more on intuition) or more explicit (relying on learned and verbalized grammar rules). For example, grammaticality judgments can reveal speakers’ knowledge of explicit grammar rules but may not be appropriate for heritage speaker informants due to their underdeveloped metalinguistic skills (Bowles, 2011; Correa, 2011a, 2011b). Conversely, oral naturalistic production does not require high cognitive control or analytic knowledge, but L2 learners are at a disadvantage here because they have mostly received written input (Bialystok & Ryan, 1985). Several studies comparing early and late bilinguals acknowledge different behavior depending on the type of task (Alarcón, 2011; Montrul et al., 2008). However, these studies used written tasks to measure explicit knowledge and oral tasks for measuring implicit knowledge. When studies control for explicitness of the task and type of register, similar results are reported (Bowles, 2011), finding no differences between heritage speakers and L2 learners in oral explicit tasks (Montrul et al., 2013). Apparently, the explicitness of the task matters for these two types of learners when we control for task modality. If linguistic knowledge elicited through tasks that maximize implicit knowledge is closer to true grammatical competence than knowledge elicited through reading and writing, it is plausible that heritage speakers have advantages not only in phonology but also in morphosyntactic aspects.

In sum, although the literature in heritage and second language acquisition in adulthood acknowledges that an early age of exposure translates into advantages in phonology, it is still unclear whether these benefits also extend to the morphosyntactic domain. Task modality and task explicitness matter in how linguistic knowledge is accessed by early and late bilinguals, and therefore more studies that examine the syntactic knowledge of these groups while also controlling for these variables are needed. This study focuses on knowledge of “clitic climbing” constructions in Spanish, a syntactic phenomenon that results in proclitic placement in verbal periphrases as in lo quiero ver “(I) want to see it” (Rizzi, 1982), in contrast with the enclitic (in situ) alternative quiero verlo, with both options showing no semantic effects. Provided that this operation takes place predominantly in the oral register (Davies, 1995), I examine whether heritage speakers of Spanish show better production of proclitic objects due to an early and oral-based exposure to the target language using an elicited production task. This article is structured as follows: section “The Syntax of Clitic Climbing Constructions in Spanish” describes the phenomenon at hand under the lenses of the generative and usage-based (i.e., sociolinguistic) frameworks; section “Clitic Climbing in L2 and Heritage Spanish” revisits the literature to date on the L2 acquisition of clitic climbing constructions; section “The Study” is devoted to the methodology of the study; and in section “Conclusions and Future Directions,” I present the conclusions with implications for theories of second language development and pedagogical outcomes.

The syntax of clitic climbing constructions in Spanish

Subjects and objects in Spanish pronominalize in two distinct ways. Personal pronouns are autonomous words that can function in isolation; object pronouns, however, do not have prosodic content and thus must lean onto a host to function. For this reason, personal pronouns are considered “strong” forms in the syntax, while object clitics are considered “weak” (Gerlach, 2002; Klavans, 1995; Zwicky, 1994), as in (1). In Spanish grammar textbooks personal pronouns are often referred to as tónicos “stressed,” and object pronouns as átonos “unstressed” (Fernández Soriano, 1993, 1999).

(1) ¿A quién viste? A él/*lo

to who see-2.sg.pret to him/*cl.3.acc

Who did (you) see? Him

Another characteristic of Spanish object clitics, closely related to their prosodically weak nature, is that their placement varies according to the finiteness of the verb: clitics precede single finite verbs (i.e., proclisis) but are attached to non-finite forms (i.e., enclisis) (Fischer & Goldbach, 2016), as in the examples show in (2a, b):

(2) a. Lo compró

cl.3.acc buy-3.sg.pret

(S/he) bought it

b. Leyéndolo

reading-cl.3.acc

Reading it

According to the taxonomy proposed by Zwicky (1994), English also contains clitics. However, English clitics are considered “simple” in the sense that they are the result of phonological reductions of free morphemes, becoming subordinates to a neighbor word (e.g., We gave ’em to ’er). This type of cliticization is associated with casual or rapid speech. In contrast with Spanish object clitics, which are classified as “special,” simple clitics do not observe idiosyncratic syntax: regardless of the finiteness of the verb, they always remain in their associated verb phrase (VP)-argument positions. The special syntax of Spanish object clitics is more clearly evidenced in certain periphrastic constructions, where clitics can be placed either proclitic to the matrix verb or enclitic to the non-finite verb complement (3a, b). The syntactic phenomenon by which a clitic abandons its argument position within the non-finite domain and moves up to the finite (matrix) verb domain is known as “clitic climbing” or “clitic promotion” (Kayne, 1989; Luján, 1980; Rizzi, 1982).

(3) a. Tengo que verlo

have-1.sg to see-inf-cl-3.acc

b. Lo tengo que ver

cl-3.acc have-1.sg to see-inf

(I) have to see it

Formal accounts to cliticization in Romance propose that this phenomenon is only available to certain verbs. Aissen and Perlmutter (1976) divide verbs into “trigger” and “non-trigger” verbs, according to their ability to host dependents (e.g., clitics) of a complement (i.e., embedded, non-finite) verb. These “trigger” verbs may belong to three classes: modals, aspectuals, and aspectuals containing motion verbs. Similar taxonomies have been proposed by Napoli (1981) for Italian, Picallo (1990) for Catalan, and Zagona (2002) for Spanish. For example, factive verbs such as lamentar “lament, be sorry for” disallow clitic climbing, but modals (poder “be able”), aspectuals (ir + a “be going to”), and motion verbs (ir “go” + gerund) allow it. Rizzi (1982) proposes that certain verbs may undergo a process of “restructuring” by which a complex verbal unit consisting of a main verb and an embedded one may merge into a single unit.

However, these approaches show limitations in explaining why a clitic would abandon the complement position where it is generated and appear in the matrix VP domain (Bok-Bennema, 2006). Apparently, clitic climbing is a sort of optional movement which may, but need not, apply (Rizzi, 1982; Roberts, 1997; Rooryck, 2000). Variationist studies point to the properties of the matrix verb as the best predictor for clitic climbing phenomena: the more grammaticalized a verb is in terms of its tense-aspect-mood expression, the more capable it is of attracting clitics from embedded complements, and vice versa. Highly grammaticalized matrices undergo a process of semantic “bleaching” (Myhill, 1989), resulting in merely auxiliary-like forms indicating progressive aspect (e.g., ir “go” + gerund, estar “be” + gerund, ir + a “be going to” as a future marking). Napoli (1981) hypothesizes varying degrees of “conceptual unity” between matrix and embedded verbs, and Rosen (1989) makes a distinction between “light” and “heavy” verbs. Davies (1995) went a step further by analyzing a corpus of oral speech from various dialects of Spanish. He found that this operation was more common in the oral register and identified 32 matrices capable of attracting clitics from non-finite complements in periphrastic contexts, arranged by their likelihood for attracting clitics. His approach has the advantage that it captures those matrices with a certain degree of clitic climbing occurrence, but do not belong to a strict grouping (e.g., auxiliary, modal). For example, saber “to know/to know how” showed higher incidence to clitic climbing (33%) compared to necesitar “need” or deber “must,” modal verbs with a 20% of incidence, or aspectual verbs (e.g., comenzar + a “start to” and dejar + de “stop,” both at 25%). Table 1 shows an adaptation from Davies’ (1995) gradience of matrices with varying degrees of clitic climbing affinity in oral and written Spanish (p. 374). Subsequent studies following up with Davies’ (1995) findings report similar trends (Schwenter & Cacoullos, 2014; Torres Cacoullos, 1999).

Table 1.

Clitic climbing incidence by verb in spoken and written corpora, adapted from Davies (1995, p. 374).

Verb	Average (spoken)	Average (written)
ir + a	0.86	0.66
acabar + de	0.85	0.29
haber + de	0.80	0.33
volver + a	0.81	0.27
soler	0.86	0.05
poder	0.60	0.22
venir + a	0.55	0.18
saber	0.52	0.14
terminar + de	0.55	0.08
querer	0.47	0.15
empezar + a	0.45	0.13
llegar + a	0.42	0.11
dejar + de	0.41	0.10
tener + que	0.38	0.13
comenzar + a	0.45	0.05
aprender + a	0.45	0.00
lograr	0.41	0.02
necesitar	0.33	0.07
deber	0.32	0.08
salir + a	0.35	0.00
pensar	0.24	0.07
desear	0.20	0.04
tratar + de	0.20	0.01
pasar + a	0.00	0.17
preferir	0.15	0.00
procurar	0.14	0.00
intentar	0.11	0.01
resolver	0.05	0.00
esperar	0.00	0.00
insistir + en	0.00	0.00
soñar + con	0.00	0.00
haber + que	0.00	0.00

While variationist (sociolinguistic) studies concentrate on establishing patterns of highly frequent instances of syntactic phenomena in the speech of a community of speakers (Labov, 1996), theoretical studies focus instead on identifying the syntactic conditions that allow for possible sentences in language grammars (Chomsky, 1995). Clitic promotion in Romance has received significant attention in theoretical linguistics (Ordóñez, 2002, 2012; Raposo & Uriagereka, 2005; Roberts, 2010, to cite a few), and two traditional approaches have been proposed to account for cliticization in Romance: on one hand, “movement-based” approaches share the idea that clitics first generate as regular VP arguments and then surface to the left of V via syntactic movement (i.e., Move-alpha) and language-specific rules, leaving a trace or copy (Kayne, 1975, 1989; Quicoli, 1976; Uriagereka, 1995); on the other hand, “base-generation” approaches propose that clitics generate in a position located to the left of the verb as adjuncts (not arguments) leaving the VP-internal argument position free for a possible co-indexed determiner phrase (DP) which receives a thematic role, as it occurs in “clitic doubling” constructions, such as La vi a ella “I (cl) saw her” (Borer, 1984; Jaeggli, 1982; Rivas, 1977; Strozer, 1976; Suñer, 1988). Both approaches complement each other: movement-based accounts cannot explain the co-existence of clitics and DPs in the same sentence, but in base-generation approaches, clitics cannot undergo movement at all.

Blasco Aznar’s (2002) proposal combines the two types of analyses just discussed (as others have, for example, Bleam, 2000; Ormazabal & Romero, 2013; Roca, 1992; Romain, 2015; Torrego, 1998, among others): she proposes that clitics are base-generated as a subpart of a VP-internal DP argument, more specifically as a determiner D whose noun phrase (NP) sister is the empty category pro. In finite clauses, clitics move from the DP complement to the left of an abstract (empty) functional head above the tense phrase (TP): the “clitic holder” (XP) phrase, yielding the [cl V] order as in lo vio “(S/he) saw it.” In verbal periphrases, a clitic that is generated as part of an embedded non-finite VP may adjoin to the left of the aforementioned functional head above T as long as the embedded C(omplementizer) node is empty (e.g., *lo prefiero no ver “[I] prefer not to see it”), and assuming a certain degree of clause union between the matrix and embedded verbs. In addition, for a clitic to be attracted to the matrix domain, the Agr(eement)P must project subject features (i.e., impersonal periphrases such as es mejor verlo “[It] is better to see it” do not allow clitic climbing). Based on Kayne’s (1991, 1994) proposals, a clitic is devoid of any features that need to be checked by the heads it moves through, allowing it to reach the abstract Xº node projected by the TP (see Figure 1).

Figure 1.

Syntactic derivation of lo quiere ver “(S/he) wants to see it” following Blasco Aznar’s (2002) movement-based approach.

In sum, to better understand complex cliticization in Romance, more than one approach is necessary. The production of clitic climbing constructions in Spanish is regulated by the lexical and syntactic properties of the matrices. Sociolinguistic research reports that not all matrices are capable of attracting clitics, and those that attract verbal complements from their embedded non-finite VPs do so in a gradient fashion: some verbs show a higher probability of attracting clitics than others. Theoretical approaches to clitic promotion concentrate on the nature of clitics and on defining the specific place in the syntax to where clitics belong. Two major syntactic conditions must be met for grammatical sentences to surface: the embedded complementizer phrase (CP) node must be free of any intervening material such as negation adverbs, and the matrix Agr(eement)P must project subject features, for example, impersonal periphrases do not allow clitic promotion.

The L2 acquisition of clitic climbing constructions in Spanish by speakers whose first language (L1) is English thus involves the resetting of various parameters that differ between the L1 and the L2. First, English object pronouns are all strong, and contrary to Spanish object clitics, they behave as full DPs and occupy argument positions in the sentence in a strict subject–verb–object (SVO) order, regardless of the finiteness of the verb that selects them. Second, English object pronouns may cliticize and remain attached to a prosodically stronger host due to a phonological reduction that is a characteristic of rapid, oral speech. However, Spanish object clitics are not phonological reductions of independent words, instead they show idiosyncratic (special) syntax: their placement varies according to the finiteness of the verb. Third, due to the syntactic and phonological properties of its object pronouns, variable object clitic placement does not exist in English. L1 English learners of L2 Spanish must incorporate all of the lexical and syntactic features that entail such operation to their interlanguage grammars, and be aware of the fact that this operation is enabled almost invariably in spontaneous, oral speech (Davies, 1995; Torres Cacoullos, 1999).

Clitic climbing in L2 and heritage Spanish

Generative approaches to Romance clitichood hypothesize that clitic promotion is an evidence of the Null Subject Parameter (e.g., Kayne, 1989). Provided that children start using clitics early in childhood (Grohmann & Neokleous, 2014; Larrañaga & Guijarro-Fuentes, 2012), it is possible that clitic climbing constructions could also be among the first parameters set in L1 acquisition (Wexler, 1996, 1998). Rodríguez-Mondoñedo et al. (2005) found instances of clitic climbing constructions in the speech of five children as early as age 3;10. Using an imitation task, Eisenchlas (2003) found that 71 Argentinian children showed adult-like production of proclitic placement with modal verbs by age 4;0. Older groups showed low rates of pre- to post-verbal repositioning (6%), and no children placed clitics incorrectly. Requena (2015) found similar results with children also at age 4;0, and like adult speakers, children are also aware of the lexical and semantic properties that affect variable object clitic placement.

In general, English–Spanish bilingual children show comparable performance to monolingual adults using clitics in restructuring contexts. Pérez-Leroux et al. (2011) replicated Eisenchlas’ (2003) task with 23 Canadian simultaneous and sequential bilingual children, finding that both groups of bilinguals showed an enclitic bias not attested in monolingual children. However, simultaneous bilinguals performed worse than sequential bilinguals, repeating preverbal clitics as post-verbal and using clitic omissions in both placement options. Shin et al. (2017) analyzed placement of third-person accusative clitics in the speech of 17 bilingual children of Mexican descent (6;0–11;9) and 43 children from Mexico (6;3–11;9), finding no differences between groups and clitic placement choices in verbs with varying degrees of grammaticalization. In fact, the patterns found in bilingual children matched those from 21 adults from the same community. Seeking a common ground between Pérez-Leroux et al.’s (2011) and Shin et al.’s (2017) results, Requena and Dracos (2018) examined data from 29 bilingual children (9;0–12;0) in a high-density Hispanic community in Texas using an elicited production task. Consistent with the reported variationist trends (Davies, 1995), 87% of the total responses were proclitic, the rates of clitic climbing instances with ir a and tener verbs differed from each other (96% vs 65%), and proclitic placement was higher with animate referents. In contrast with Pérez-Leroux et al.’s (2011) results, the performance of simultaneous bilinguals resembled monolingual speakers’, at least in high-density communities.

Experimental studies that examine the L2 acquisition of clitic properties by adult classroom learners follow a “parameter resetting” approach that (Ungureanu, 2011; White, 1998, 2003) if the L2 Spanish learner still has access to the Universal Grammar (UG) in adulthood, it would be possible to incorporate the new properties to the interlanguage. For example, English pronouns are prosodically “strong” (tónicos) and observe the behavior of independent words; the syntax of French clitics in [Aux V] contexts is affected by the verb-raising parameter (e.g., Je veux le acheter vs Lo quiero comprar “I want to buy it”) (Kayne, 1991; Rizzi, 1982). Therefore, the task of L2 learners is to incorporate the idiosyncratic syntax of Spanish clitics to their grammar system. Liceras (1985) found that in spite of being exposed to the same classroom input, 60 speakers of L1 French and L1 English incorrectly placed object clitics in [V *cl inf] contexts as in *yo quiero lo leer “I want to read it” when asked to recount a story (or alternatively writing a dialogue) describing a series of images. Bruhn de Garavito and Montrul (1996) examined the written production and judgments of clitics in various periphrastic contexts from 20 L1 French learners, finding that these latter used clitics correctly but rated clitic climbing as impossible in Spanish. The authors concluded that once the parameters associated with clitics and [Aux V] periphrases have been initially activated, it may take a while before these properties are fully set. Similar results were found in Duffield and White’s (1999) study with 55 intermediate- and advanced-level L1 French and L1 English learners. They combined the results of an offline grammaticality judgment task with an online sentence-matching task, finding that in grammatical [cl V inf] and ungrammatical [V *cl inf] contexts the majority of L2 learners opted for lowering object clitics, and only in 18.5% of the corrections, the proclitic choice was used. Similar to Bruhn de Garavito and Montrul’s (1996) results, L2 learners did not see the correct “climbing” option as possible in Spanish, leading the authors to conclude that the acquisition of properties that do not match those from the L1 is still possible. This conclusion is supported by the results from Rossi et al. ’s (2017) study: they examined sensitivity to wrong clitic linearizations [V *cl inf] in the Spanish of 25 highly proficient learners using a self-paced reading task, finding that L2 learners showed native-like sensitivity when parsing “regions of interest” in which clitics were incorrectly placed. Halloran and Rothman (2012) investigated whether L2 Spanish learners showed sensitivity to clitic placement restrictions in periphrases with exceptional case-marking (ECM), which do not allow two clitics to appear together (Strozer, 1976), as in (4):

(4) Juan me escuchó cantarla ayer

John cl-1.dat heard sing-cl-3.f.acc yesterday

John heard me sing it yesterday

Sixty-five L1 English participants performed a scalar grammaticality judgment task, and only the advanced-proficiency and native groups showed awareness of these restrictions: the beginner- and intermediate-level proficiency learners treated ECM and non-ECM contexts similarly, and as in previous studies, the authors concluded that full access to UG is possible in adulthood.

Studies that compare the performance of heritage speakers and L2 learners in the less-dominant language have found in clitic climbing constructions a fertile ground to test hypotheses about possible transfer effects from the dominant language, which may result in divergent use and knowledge from native trends. Montrul (2010a) found that native and heritage speakers use the climbing option at higher rates than L2 learners, and both experimental groups discard ungrammatical [V *cl inf] sentences. However, while the heritage and native groups rated grammatical sentences with climbing similarly, the L2 group assigned low rates to grammatical proclitic sentences, similarly to previous findings with L1 English classroom learners. An early exposure to the target language, the author concluded, results in advantages for certain morphosyntactic structures. Montrul (2010b) examined the performance of the lower proficiency heritage and L2 groups from the previous study, finding that the heritage group showed comparable production to native speakers, whereas the L2 learners did not produce any proclitic instances at all. Similar to the more advanced participants in her previous study, the heritage and L2 groups rated [V *cl inf] sentences as impossible, and just like their more advanced-proficiency counterparts, few L2 learners rated the [cl V inf] as possible in Spanish. These data reinforce the author’s previous conclusion: an early age of onset of exposure (even if minimal) confers advantages in certain aspects of core syntax. In line with previous studies, L2 learners seem to require a certain amount of time to become aware of the full range of options for clitic placement in Spanish. Using a picture elicitation task, Thomas (2012) found that 10 heritage speakers outperformed 16 L2 learners and a group of native speakers in their use of proclitic placement across a selection of verbs with varying degrees of clitic climbing affinity: ir + a “to be going to,” querer “to want,” and preferir “to prefer.” L2 learners and monolingual speakers behaved alike and showed more lexical sensitivity, opting for proclitic placement in more highly grammaticalized (i.e., auxiliary-like) verbs and more enclisis with modal verbs. The results from an occurrence/preference task showed that all groups had sensitivity to the semantic constraints that affect clitic placement in verbs ir + a “to be going to” to express future motion versus the past form fue a “s/he went to” to express physical motion. Some heritage participants showed indeterminate judgments when rating clitics in ungrammatical periphrases with embedded negations as in *Juan lo quisiera no saber “John wanted to not know it.” Native speakers and L2 learners, in contrast with heritage speakers, receive more exposure to written input and develop metalinguistic awareness. Consequently, they are more aware of the placement choices that affect the type of verb, the author concluded. In a sentence recall task, Meijer and Fox Tree (2003) found that heritage speakers of Spanish are aware of the syntactic restrictions of these constructions. Eighteen heritage speakers were asked to repeat 20 sentences with [V inf cl] and [cl V inf] pairs after they were presented English primes. Overall, 93.3% of the responses were recalled accurately, and 25.6% of the proclitic sentences were recalled using enclitic placement, perhaps due to cross-linguistic influence if those participants lacked a unitary system containing the rules of both languages, as the authors state.

In sociolinguistic studies of Spanish in the United States, informants are usually divided into three generational groups, depending on the age of arrival to the country (early or late), or if the participant was born in the country and has at least one ancestor that belongs to those groups. One study found that enclisis was overall the preferred option across three generations (Gutiérrez & Silva-Corvalán, 1993), but in later work, it was found that periphrases with ir a “to be going to” with future value, progressive estar “be” + gerund, and modal poder “may” were the most frequent periphrastic contexts for proclitic placement across the three generations (Silva-Corvalán & Gutiérrez, 1995). More variation was found with modal verbs and certain inchoative expressions such as empezar + a “begin,” deber “must,” tener + que “have to,” querer “want,” tratar + de “try” + infinitive (Gutiérrez, 2008); bilinguals less familiarized with modal matrices favor the enclitic option (Peace, 2012). Gutiérrez (2014) compared data from interviews between the 1990 and 2000 decades, and he noticed that bilingual’s clitic placement preferences change over time, with a recent tendency to favor enclisis in matrices with infinitival complements, and some highly grammaticalized matrices had almost lost their ability to attract clitics (e.g., empezar/comenzar + a and tener + que). In a similar fashion, Limerick (2017) found that age of arrival to the host country and length of residence affect proclitic placement in [V inf] contexts: ir + a, poder and querer.

Overall, the experimental and variatonist studies reviewed so far report that heritage and native Spanish speakers show similar production rates of complex proclitic sentences, although heritage speakers may overextend clitic climbing to less-favorable contexts for native Spanish speakers. L2 learners will instead favor the enclitic option regardless of their proficiency, but more advanced individuals show closer (native-like) knowledge of what a possible sentence with climbing in Spanish is. The behavior of these groups may be a consequence of different exposure onset times, different input experiences, and how this linguistic knowledge is accessed. The next section takes into consideration all these variables and proposes a new experiment to shed more light on how these bilinguals perform with Spanish clitics.

The study

Research questions and hypotheses

The current literature on L2 acquisition reports that speakers of minority languages are granted selective advantages over traditional classroom learners in aspects related to the phonological domain, but these advantages do not necessarily apply to morphosyntax. Factors such as type of experience with input, task modality, and task explicitness matter when accessing the grammar knowledge of these learners. Studies that examine clitic climbing constructions among heritage speakers and L2 learners of Spanish report that heritage and native Spanish speakers show comparable rates of clitic climbing production. However, heritage speakers’ knowledge of which matrices are more likely to attract clitics is more limited, leading heritage speakers to favor proclisis with less grammaticalized verbs. L2 learners show a tendency to favor enclitic constructions, but they seem to display better knowledge than heritage speakers of the lexical and syntactic properties of variable object clitic placement, likely as a result of their metalinguistic awareness.

One limitation of these studies is that few matrices have been employed in elicited production tasks (e.g., ir + a “to be going to,” poder “to be able to,” querer “to want,” and preferir “to prefer”), and consequently, we do not know yet if the reported tendencies apply to verbs with similar degree of grammaticalization, as reported by Davies’ (1995) corpus study with native Spanish. He also states that matrices do not necessarily belong to strict groupings (i.e., auxiliary-like vs modal-like). For example, aspectual verbs (i.e., highly grammaticalized verbs such as comenzar + a “start to” and dejar + de “stop doing”) only show a 25% rate of clitic climbing affinity. In addition, variationist studies report that clitic climbing with estar “to be” + gerund has become categorical in both native and heritage Spanish (Gutiérrez, 2014). Thus, the effects of an early versus late exposure to the target language and type of input (oral vs written) will be more evident in matrices with varying likelihood of attracting clitics, such as highly grammaticalized (higher probability) versus low grammaticalized (lower probability). Therefore, only [V inf] periphrases should be tested. Provided that clitic climbing constructions are also regulated by strict syntactic conditions, it is necessary to test not only whether these two groups of bilinguals use ungrammatical proclisis in periphrases with embedded negations as in Thomas’ (2012) study, but also the production (or lack thereof) of proclisis in impersonal periphrastic constructions, which also disable proclitic placement.

Considering all of this information, the following research questions are formulated:

RQ1. Do L1 English L2 Spanish learners and heritage speakers of Spanish show target production of clitic climbing constructions compared to monolingual Spanish speakers across the four aforementioned verbal conditions?

Based on previous findings, I predict divergent use from the native trends among the experimental groups: heritage speakers will overextend proclitic placement across all verbal contexts (Thomas, 2012), and L2 learners will instead opt for the enclitic placement option (Bruhn de Garavito & Montrul, 1996; Montrul, 2010a, 2010b; Thomas, 2012).

RQ2. To what extent does language proficiency in the less-dominant language affect clitic climbing production in English-dominant L2 Spanish learners and heritage speakers of Spanish?

I predict that language proficiency will affect the rates of clitic climbing use among both groups across the four verbal contexts: more proficient heritage speakers and L2 learners will show higher rates of proclitic placement in matrices that allow it (e.g., matrices with high and low degrees of grammaticalization), and will avoid proclisis in contexts where clitic climbing is forbidden (e.g., impersonal constructions, sentences with embedded negation adverbs).

Participants

To test these predictions, a pool of monolingual Spanish speakers, heritage speakers of Spanish, and L2 Spanish learners born and raised in an English-speaking country were interviewed. Fifty-three informants participated in this study: 20 native speakers from Mexico, 17 L2 learners, and 16 heritage speakers of Spanish were recruited through paper and online advertising (N = 53). The participants signed a consent form and filled out a language background history questionnaire (Cuza & Frank, 2015), comprising two sections. In Section 1, participants were asked to provide sociolinguistic information such as place of birth, languages spoken during childhood, the parents’ first language(s), and language(s) of schooling, as well as other language patterns of use. Participants had to indicate their language uses at home, at school, at work, and in social situations, and also the language in which they were more comfortable, using 7-point scales ranging from English-only (1) to Spanish-only (7). In Section 2, participants completed a self-assessment of the four language skills in English and Spanish (reading and listening comprehension, writing and speaking production) using scales ranging from limited (1) to native (4). In this section, participants were also asked to fill out a written proficiency test in Spanish (Diploma de Español como Lengua Extranjera, or DELE), comprising 30 multiple-choice vocabulary questions and a passage with 20 “fill in the blank” grammar questions, each with three possible answers (Cuza et al., 2013). Following previous work (Montrul & Bowles, 2010; Montrul & Slabakova, 2003), scores were divided into three proficiency groups: 40–50 for “advanced proficiency,” 30–39 for “intermediate proficiency,” and “low proficiency” for scores 0–29.

Sixteen heritage speakers of Spanish (seven men, nine women) comprised the heritage speakers group. All of them were undergraduate and graduate students at a large Midwestern university. Twelve were born and raised in the United States, and four were born abroad (Mexico, Guatemala, Argentina, Puerto Rico) and arrived to the country before age 3. At the time of testing, the mean age was 25 (range, 18–45, SD = 7.24). Their sociolinguistic information showed that 88% of the participants had at least one native Spanish speaker ancestor, and all participants except for two had traveled to Spanish-speaking countries. Their use of languages are as follows: at school, 64% reported that they use English or mostly English, while 17% employ both English and Spanish; at home, 47% used more Spanish; at work, 64% of the heritage participants stated that they used slightly more English or mostly English; in social situations, 76% used mostly English and 17% used both English and Spanish. English was the primary language of instruction at elementary school (70%), high school (76%), and at college (70%). Overall, 52% of the heritage participants stated that they were more comfortable using mostly English, while 47% reported that they were equally comfortable using English and Spanish. Their overall self-assessment scores across the four language skills were 2.7/4 (adequate) in Spanish and 3.7/4 in English. The proficiency scores show that 10 participants (62.5%) were advanced, and that 6 (37.5%) were intermediate. Based on their sociolinguistic background information, these participants were considered to be native speakers of English.

Seventeen college-level learners of Spanish whose native language is English (7 men, 10 women) comprised the L2 learners group. Like in the heritage speakers group, all of them were undergraduate and graduate students at a large Midwestern university. These informants were all born and raised in the United States except for one, who was born and raised in the United Kingdom. Their mean age at the time of testing was 26 (range, 19–68, SD = 11.46), and all of them except five had already visited a Spanish-speaking country in the past. Their use of English and Spanish are as follows: 63% reported using more English or mostly English at school and 26% reported using slightly more Spanish; 84% indicated that they used mostly English at home; 52% reported using more English at work; 84% reported using mostly English in social contexts. The majority of them had English as the primary language of instruction in elementary school (100%), high school (74%), and university (42%); 74% reported to be more comfortable using English and 21% indicated that they feel equally comfortable using English and Spanish. The overall assessment of the four language skills was 3.2/4 (good) for Spanish and 4/4 (native) for English. Their proficiency scores in the target language showed that 11 (64.7%) were at advanced level and 6 (35.2%) were intermediate.

Twenty native Spanish speakers (6 men, 14 women) born and raised in Mexico comprised the native control group. Their mean age at the time of testing was 22 (range, 18–33, SD = 4.34), 16 of them were students at a large western Mexican university and 4 of them were employed. Although most of the participants had already taken courses of English in high school, they did not consider themselves bilinguals. Like the experimental groups, these native speakers filled out a language background history questionnaire and their overall self-assessment of the four language skills in both languages was 3.7/4 (native-like) in Spanish and 2.7/4 (adequate) in English. Participants in this group were not requested to fill out the DELE proficiency test. Table 2 shows a brief summary of the characteristics of the three groups.

Table 2.

Experimental groups: descriptive statistics for the variables age at testing, proficiency self-rating, and proficiency scores.

Groups	n	Proficiency scores (max = 50)		Proficiency self-rating (maximum 4)		Age at testing
Groups	n	M	(SD)	M	Range	M	Range
Native	20	–	–	3.7	3.7–3.8	22.6	18–33
Heritage	16	41.2	(4.3)	3.1	2.6–3.7	25.25	18–45
Advanced	10	43	(2.1)	3	2.4–3.7	27.8	18–45
Intermediate	6	36.5	(2.8)	3.2	3–3.7	20	18–21
L2 learners	17	41	(4.5)	3.2	2.4–3	26.5	19–68
Advanced	11	44.9	(2.7)	3.5	3.4–3.7	29.6	20–68
Intermediate	6	37.1	(2.1)	2.7	2.4–3	21	19–27

Test conditions and task

The goal of this study is to examine whether these two groups of English–Spanish bilinguals’ production of clitic climbing constructions is equally affected by the lexical and syntactic restrictions as native Spanish speakers. A selection of matrix verbs with varying degrees of climbing affinity was employed to examine awareness of those lexical properties among English-dominant participants. Following Davies’ (1995) array of matrix verbs, the high-grammaticalized matrices chosen were ir + a “be going to” (.86), soler “be accustomed to” (.86), acabar + de “have just + participle” (.85), haber + de “have to” (.80), volver + a “do again” (.81), and poder “be able to” (.60), and the selected low-grammaticalized matrices were aprender + a “learn to” (.45), tener + que “have to” (.38), salir + a “leave” (.35), necesitar “need” (.33), deber “must” (.32), and pensar “think of” (.24). The verbal periphrases that were used to test the syntactic restrictions were six impersonal sentences: hay + que “(one) has to,” es mejor “(it) is better to,” es necesario “(it) is necessary to,” es urgente “(it) is urgent to,” es preferible “(it) is preferable to,” and es recomendable “(it) is advisable to.” Six test tokens with embedded negations (e.g., prefiere no pintarlo “s/he prefers not to paint it”) were used to test this syntactic condition. Six test tokens for each of the four verb conditions were designed, 24 test tokens in total with 24 fillers. A summary of the testing conditions with sample test tokens is presented in Table 3.

Table 3.

Testing conditions. The tick symbols represent a higher or lower probability for a clitic to be attracted to the matrix verb domain. (* means that the sentence is considered a grammatical by a native speaker).

Type of matrix verb	Proclitic placement	Enclitic placement
High occurrence: Lo va a cuidar bien Va a cuidarlo bien S/he is going to take care of him	√√	√
Low occurrence: Necesitan comprarla La necesitan comprar They need to buy it	√	√√
Impersonal periphrasis: Es mejor llamarlo luego Lo es mejor llamar luego It is better to call him later*	*	√
Periphrasis with an embedded negation adverb: Prefiere no pintarlo Lo prefiere no pintar S/he prefers not to paint it*	*	√

The method employed to elicit proclitic complex sentences across the four conditions mentioned above was an oral sentence completion task, to force participants to produce forms in a more controlled environment (Cuza et al., 2013; Eisenbeiss, 2010; Gropen, 2000). Participants sat with the researcher and were asked to finish a sentence after being prompted with a contextual situation using recorded audio and pictures, using non-finite verbs in parenthesis. Participants’ responses were first recorded and then coded for analysis. Four possible answers were considered: proclitic placement, for example, Vicente la acaba de enviar hace un momento “Vicente has just sent it a while ago”; enclitic placement as in Vicente acaba de enviarla hace un momento; DP object as in Vicente acaba de enviar la tarea hace un momento “Vicente has just sent the homework a while ago”; or null object as in Vicente acaba de enviar ∅ hace un momento “Vicente has just sent ∅ a while ago.” This latter option is possible in several dialects of Spanish (Clements, 2006; Suñer & Yépez, 1988), while in others are considered ungrammatical (Campos, 1986; Schwenter, 2006). Figure 2 shows a representative test token with four possible answers.

Figure 2.

Sample test token, sentence completion task.

Results

Responses were coded for each of the participants. Incorrect responses were labeled as “other.” Table 4 shows examples of what speakers produced in each category. Table 5 shows the raw counts for each of the five possible response (proclisis, enclisis, DP object, null object, other) across the four verbal conditions (auxiliary, modal, impersonal, embedded negation) and by group (native speakers, heritage speakers, L2 learners).

Table 4.

Sentence completion task, sample responses by object type and verb condition. (* means that the sentence is considered a grammatical by a native speaker).

High-grammaticalized verbs
Proclitic:	HS_05_B: Lo va a cuidar “(S/he) is going to take care of it”
Enclitic:	L2_10_A: Acaba de enviarla “(S/he) has just sent it”
DP:	C_01_A: Habrá de llevar su traje “He will have to bring his suit”
Null:	HS_09_B: Suele recoger “(S/he) usually picks up”
Low-grammaticalized verbs
Proclitic:	L2_01_A: Las aprende a tejer “(S/he) learns to weave them”
Enclitic:	HS_03_B: Piensa cambiarlo “She thinks about changing it”
DP:	C_12_B: Necesitan comprar una guía “They need to buy a guide”
Null:	HS_01_A: Debe entregar “(S/he) has to deliver”
Impersonal sentences
Proclitic:	–
Enclitic:	C_06_A: Hay que hacerlas “(One) has to do them”
DP:	HS_07_B: Es mejor llamar a Nadina “It is better to call Nadina”
Null:	L2_11_B: Es necesario cambiar “It is necessary to change”
W/ embedded negation
Proclitic:	HS_13_B: *Lo prefiere no tomar “(S/he) prefers not to take it”
Enclitic:	L2_12_A: Prefiere no pintarlo “(S/he) prefers not to paint it”
DP:	HS_07_B: Deseará no comprar la camiseta “(S/he) would prefer not to buy the t-shirt”
Null:	L2_04_A: Suele no llevar “(S/he) usually does not wear”

Table 5.

Sentence completion task, raw counts of all response objects by verb and group.

Group	n	Proclitic object		Enclitic object		DP object		Null object		Other
Group	n	Count	%	Count	%	Count	%	Count	%	Count	%
Native speakers:	20	7	5.83% (7/120)	80	66.66% (80/120)	23	19.2% (23/120)	10	8.32% (10/120)	–	–	High grammaticalized
Heritage speakers:	16	13	13.54% (13/96)	41	42.7% (41/96)	33	34.37% (33/96)	11	11.45% (11/96)	–	–
L2 learners:	17	11	10.8% (11/102)	48	47.05% (48/102)	19	18.62% (19/102)	25	24.5% (25/102)	–	–
		–	–	70	58.33% (70/120)	8	6.66% (8/120)	42	35% (42/120)	–	–	Low grammaticalized
		–	–	33	34.37% (33/96)	18	18.75% (18/96)	44	45.83% (44/96)	1	1.04% (1/96)
		4	3.92% (4/102)	43	42.15% (43/102)	14	13.72% (14/102)	41	40.2% (41/102)	–	–
		–	–	94	78.33% (94/120)	9	7.5% (9/120)	11	9.16% (11/120)	6	5% (6/120)	Impersonal sentence
		–	–	57	59.37% (57/96)	8	8.33% (8/96)	15	15.62% (15/96)	17	17.7% (17/96)
		–	–	47	49% (47/102)	9	8.82% (9/102)	16	15.7%(16/102)	30	29.41% (30/102)
		1	.83% (1/120)	82	68.33% (82/120)	14	11.66% (14/120)	22	18.33% (22/120)	1	.83% (1/120)	W/ embedded negation
		1	1.04% (1/96)	47	49% (47/96)	10	10.42% (10/96)	31	32.3% (31/96)	5	5.2% (5/96)
		3	2.94% (3/102)	40	39.21% (40/102)	4	3.92% (4/102)	43	42.15% (43/102)	11	10.8% (11/102)

Contrary to what was expected, all groups favored enclitic placement followed by DP and null object responses in highly grammaticalized verbal periphrases, and heritage speakers and L2 learners behaved alike when they produce the few clitic climbing instances. DP objects were the third preferred option among heritage speakers, but for L2 learners, it was the null object option. There were no incorrect responses in this verbal context, that is, no [V *cl inf] linearizations. In low-grammaticalized periphrases, all groups strongly favored the enclitic option, with zero instances of proclitic placement among the native and heritage participants, and a few L2 learners opted for raising clitics in this verbal condition. All groups showed similar rates of null object responses, and there were few DP objects. In this verbal condition, only one heritage participant produced an incorrect response. The responses with impersonal constructions showed no instances of ungrammatical proclitic placement, with a strong preference for the grammatical enclitic option and few instances of DP and null objects across all groups. L2 learners in general had more trouble with this condition as showed by the number of incorrect responses, but heritage speakers displayed more target use. In periphrases with embedded negation adverbs, enclisis was the favored option across all groups. Heritage speakers and L2 learners behaved alike, with L2 learners producing slightly more null object responses. Again, L2 learners showed more difficulty with this condition compared to the native and heritage groups. Figure 3 shows the percentages of proclitic and enclitic placement across the four periphrastic conditions by group.

Figure 3.

Sentence completion task, rate of proclitic and enclitic responses by group and verb conditions.

To answer the first research question, a statistical analysis was performed to compare the production rates of proclitic placement across the four verb conditions in the control and experimental groups. A mixed logistic regression model with random intercepts was designed to predict whether or not a clitic climbing response was given, with group and verb type as fixed effects and the repeated measures on speaker as a random effect. R (R Core Team, 2019) was used to perform the statistical analysis, with the lme4 function (Bates et al., 2015) to fit the binary logistic mixed regression model and the sjPlot package (Lüdecke, 2018) to display the resulting odds ratios, confidence intervals, and p values in a table. A 5% level of significance was considered for all hypotheses testing and 95% confidence for all interval estimates.

Group and verb type did not have a significant interaction effect, and verb type did not have a significant interaction effect $(χ^{2} = 7.92, d f = 6, p = . 244)$ according to a likelihood ratio test comparing models with and without the interaction. There was a significant effect from verb type $(χ^{2} = 61.6, d f = 3, p < . 001)$ , but not from group $(χ^{2} = 2.62, d f = 2, p = . 270)$ . According to the marginal and conditional Nakagawa R, 93% of the variation in proclitic occurrence is due to the fixed effects in the model and 95.1% is due to both the fixed effects and the random effects due to repeated measures on speakers. Differences in the use of proclitic responses were found when comparing highly grammaticalized verb contexts with low-grammaticalized periphrases and with sentences with embedded negations. The odds of proclitic responses for modal verbs are 90.4% lower than the odds of proclitic responses being chosen for auxiliary verbs (z = −3.91, p < .001). The odds of using proclisis in embedded negations are 87.8% times lower the odds of using a similar response for auxiliary-like verbs $(z = - 3.86, p < . 001)$ . There was no significant difference between impersonal sentences and auxiliary-like verbs $(z = - 0.173, p = . 863)$ . However, this is not particularly meaningful because no proclitic responses were observed for impersonal constructions, resulting in a large standard error for this coefficient, and thus a large p value. Although not statistically significant, the occurrence of proclitic placement for impersonal verbs is lower than for auxiliary-like verbs. Table 6 shows odds ratios, their confidence intervals, and p values for individual terms in the model without interaction. Note that “Group [Control]” and “VerbType [Auxiliary]” are not included because they comprise the comparison baseline reference for each independent variable.

Table 6.

Odd ratios, confidence intervals, and p values for individual variables.

	Proclitic responses
	Odds ratios	95% CI	p
Predictors:
Group [HS]	2.441	[0.677, 8.798]	.172
Group [L2]	2.555	[0.725, 9.009]	.145
VerbType [modal]	0.096	[0.032, 0.289]	<.001
VerbType [impersonal]	0.000	[0.000, 5.9E87]	.863
VerbType [embedded neg]	0.122	[0.045, 0.333]	<.001
Random effects
σ²	3.29
τ_{00 Speaker}	1.41
n _Speaker	53
Observations	1,272
Marginal R²/Conditional R²	.930/.951

To answer the second research question, a second statistical analysis was conducted to measure the effect of the independent variables (i.e., type of verb, group [heritage speakers and L2 learners]) and proficiency on the dependent variable (i.e., proclitic vs enclitic placement). A binary logistic regression was performed using the logistic procedure in SAS v. 9.3 (SAS Institute Inc., 1999) in the Windows 10 environment. Firth’s adjustment (i.e., penalized likelihood estimation) was incorporated due to some very small proportions. Penalized likelihood methods increase the likelihood rates with a penalty function for very small estimated proportions. No significant differences were found across groups (χ² = .093, df = 1, p = .761), but a significant effect was found with highly grammaticalized verbs (p < .0001) and with proficiency (p = .0041). The odds of using proclitic placement in highly grammaticalized verbs are 6.8 times the odds of exerting climbing when the verb is modal-like, assuming that the other variables of the model remain constant. The odds of exerting the climbing option among low-grammaticalized verbs are 8.4 times the odds when the verb is impersonal, but not significantly different if proclitic placement is exerted in sentences with embedded negations. The odds of exerting the climbing option increase by 1.15 times for every unit increase in the proficiency scores, provided all other variables remain constant. Table 7 shows the predicted probabilities of exerting the climbing option by group and verb, in a probability scale between 0 and 1. In the next subsection, I discuss the data in relation with the research questions described in subsection “Research Questions and Hypotheses.”

Table 7.

Predicted probabilities (mean values) of using proclisis by group and verb.

	High occurrence	Low occurrence	Impersonal sentence	W/ embedded negation
Heritage speakers:	.1048	.0187	.0021	.0191
L2 learners:	.1158	.0209	.0023	.0213

Discussion

The first research question was formulated in terms of native language competence. Would it be possible for English–Spanish bilinguals to show similar proclitic placement patterns like native Spanish speakers? I hypothesized that both groups would deviate from the native norm, with heritage speakers overextending proclitic placement across all verbal contexts and L2 leaners using predominantly the enclitic option. The results from the elicitation task show that all groups behaved alike across the four periphrastic contexts; therefore, none of the two hypotheses were confirmed. Although the number of instances of clitic climbing across all groups was very low, those few take place in highly grammaticalized verb matrices (i.e., auxiliary-like), the findings align with the reported trends in native Spanish (Davies, 1995). No group exerted the climbing option in matrices with a low degree of grammaticalization (i.e., modal-like), although it is possible to see some variation in these contexts. In contexts where syntactic movement of functional material was not possible, all groups behaved similarly, resorting to the default in situ clitic placement. These tendencies were confirmed by the statistical analysis: verb type (but not group) was a significant predictor in clitic climbing production across the three groups. The measures, that is, penalized likelihood that were applied to the statistical model to compensate the reduced number of proclitic instances, show that the likelihood of using proclitic placement in modal-like matrices and in sentences with embedded negations is almost 90% lower in both contexts if those odds are compared to the odds of exerting the climbing option in auxiliary-like matrices.

In the second research question, I hypothesized that more proficient bilinguals would show better production of proclitic placement in contexts where it is possible, and would avoid using this option in forbidden periphrastic contexts. The second statistical analysis showed no differences between the two experimental groups, and just as in the first analysis, highly grammaticalized matrices affected the rate of proclitic responses in both groups. Proclisis in these latter contexts takes place at 6.8 times higher than with modal verbs. Regardless of age of exposure and type of experience with input, both groups showed similar odds ratios in the use of proclisis in auxiliary-like contexts: 0.1048 in the case of heritage speakers, 0.1158 for L2 learners. The probabilities of using proclisis with impersonal constructions and with sentences with embedded negations in both groups were very low. Proficiency was found to be a predictor in the use of proclitic responses in both groups. The odds of producing clitic climbing increase at a rate of 1.15 times for every increase unit in the proficiency scores (intermediate level: 30–40/50, advanced level: 40–50/50), assuming all other variables remain constant. However, the probabilities of using proclitic responses in modal-like contexts, that is, in matrices where clitic climbing is less likely to happen—albeit possible—are very low in both experimental groups, with figures close to the odds ratios of using proclisis in contexts where clitic climbing is not possible. In the light of these data, the hypothesis formulated earlier for this research question is partially confirmed: more proficient participants employed proclitic responses in highly grammaticalized contexts, but avoided this option with low-grammaticalized matrices, contexts that allow for some variation in their likelihood to attract clitics. In the next section, I connect these new data with theories of language development and discuss possible pedagogical implications for the teaching of Spanish as a second or heritage language.

Conclusions and future directions

The goal of this study is to examine the potential role of factors such as age of first exposure and experience with input in the Spanish of English–Spanish bilinguals: heritage speakers, who are exposed to the minority language early in childhood through aural interactions at home; L2 learners, who start learning the second language later in life and in formal settings (e.g., classroom), contexts in which written input and limited oral interaction are usually the norm. Provided that clitic climbing constructions are syntactic phenomena that take place predominantly in spontaneous oral speech, it would be reasonable to speculate that adult heritage Spanish speakers show native-like production of complex proclitic sentences in an elicited production task. L2 learners, due to a late start with the second language and by virtue of having received a more limited exposure to oral input, would opt for the default in situ option even when the lexical and syntactic conditions would favor proclitic placement. To test these predictions, a pool of participants performed an elicited production task testing four conditions, in two of which proclitic placement is possible with a high or low degree of likelihood, and in two of which clitic climbing is not possible.

The analysis of the few proclitic instances obtained in this task indicate that both groups of bilinguals show comparable production of clitic climbing sentences across those four verbal conditions, favoring proclitic placement only in highly grammaticalized matrices. A strong tendency to use enclitic placement was found not only in constructions that disallow clitic promotion but also with low-grammaticalized contexts, in which more variation exists among native Spanish speakers according to Davies’ (1995) corpus. These results are in line with those obtained in Liceras’ (1985) study with L1 English and L1 French learners: sometimes object clitics are perceived as free words in the sentence and sometimes as suffixes. The fact that more proficient L2 learners used proclitic placement at rates higher than their intermediate-level counterparts supports Bruhn de Garavito & Montrul’s (1996) and Duffield and White’s (1999) conclusions: it is possible for L2 learners whose languages have or do not have clitics to successfully reset all the parameters associated to this syntactic operation, although a certain amount of time and exposure to input is necessary. Recall that the advanced L2 learners had visited Spanish-speaking countries in the past and the majority were college graduate students who had already received a significant amount of instruction in the formal aspects of the second language. These findings also align with Montrul’s (2010a, 2010b) studies: more proficient L2 learners and heritage speakers show better production rates of proclitic placement in verbal periphrases; however, the intermediate heritage participants from this study did not perform better than their L2 learner counterparts. In Montrul (2010a, 2010b), an oral narrative task was employed, and only a few auxiliary and modal matrices taking infinitival complements were analyzed. Thomas (2012) found that native and heritage speakers outperformed L2 learners in a picture elicitation task, but the author did not control for proficiency. Moreover, her heritage speakers overextended proclitic placement to a low-grammaticalized verbal periphrasis. In contrast to previous studies, the data reported here show that heritage language learners do not show advantages when using complex clitic placement in Spanish.

These data suggest that if a certain proficiency level in the target language is reached, it may be possible for both L2 learners and heritage speakers to achieve clitic climbing patterns of use comparable to native Spanish speakers, at least in contexts that favor movement of functional material from infinitival complements. The Full Access/Full Transfer model proposes that the interlanguage system immediately incorporates the principles and parameters from the L1 grammar when first exposure to the target language happens (Full Transfer); also, if the interlanguage system cannot accommodate representations from the target language input, a restructuring process takes place, drawing from options of UG (full access) (Schwartz & Sprouse, 1996). Even if clitic climbing patterns of use are not explicitly taught, L2 learners of Spanish may still develop a certain awareness of which matrices favor clitic climbing and use this option in their periphrases with clitics.

The most striking finding in this experiment was the overall reduced number of proclitic instances in contexts that favor this operation. A sentence completion task was selected to elicit oral data because it prompted participants to produce instances of proclitic placement in a variety of verbal contexts more efficiently compared to spontaneous speech (Eisenbeiss, 2010; Eisenberg, 2005; Gropen, 2000). Adding fillers, one after each test token, would avoid training effects. However, in the specific case of variable object clitic placement, spontaneous data seem to better capture the nature of this syntactic phenomenon. Furthermore, from a cognitive approach, enclitic forms may be computationally “cheaper” to handle for both native and non-native speakers alike in working memory compared to proclitics, provided that they do not need to abandon their canonical DP object positions and thus no movement operations are required (Grüter & Crago, 2012; Mateu, 2015). For theories of syntax, if semi-spontaneous elicitation tasks have an effect on the production of clitic climbing sentences, movement of certain constituents may be post-syntactic, taking place at the phonological form (PF) level (Aoun & Benmamoun, 1998; Bobaljik, 2002; Embick & Noyer, 2001; Sauerland & Elbourne, 2002). The more recent variationist literature on heritage Spanish in the United States reports that certain matrices have become almost totally grammaticalized while other matrices have become more lexical in nature, and consequently do not favor object clitic movement (Gutiérrez, 2014; Limerick, 2017; Peace, 2012). Future studies examining how English–Spanish bilinguals handle variable object clitic placement in Spanish will benefit from using sociolinguistic elicitation methods. By contrasting corpora from heritage speakers and L2 learners, it will become more clear if the patterns of cliticization are similar to native speakers’, and more specifically in low-grammaticalized matrices, which show more variation.

Finally, these data point to interesting pedagogic outcomes. To explain the L2 learners’ unfamiliarity of clitic climbing constructions, it has been suggested that in the absence of semantic effects, it is difficult for non-native speakers to notice in the input that clitics can be preposed (Bruhn de Garavito, 2002, 2013; Bruhn de Garavito & Montrul, 1996; Liceras, 1985). In other words, if L2 learners do not perceive certain periphrases to comprise a single lexical unit, they will not exert the climbing option. The fact that there are no changes in meaning makes the acquisition of this syntactic operation even more challenging for L2 learners, given that this option is less salient in the available input. If this is the case, the acquisition of the clitic inventory and its properties and the acquisition of clitichood-related phenomena are in fact two separate processes. If clitic climbing is acquired, it happens at later stages of second language development.

Footnotes

Acknowledgements

I would like to thank all the participants in the study, as well as my thesis committee (Alejandro Cuza, Elaine Francis, Mariko Moroishi Wei, and Colleen Neary-Sundquist) for their insightful feedback. I am grateful to the two anonymous reviewers who evaluated this manuscript for their useful criticism and suggestions. Portions of this study were presented at the Third International Conference on Heritage Languages (University of California Los Angeles), and at the 48th Linguistic Symposium on Romance Languages (York University, Toronto). This research received one of the 2018 NFMLTA/MLJ Dissertation Completion Awards. All remaining errors are my own.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Antonio Martín Gómez

Author Biography

Antonio Martín Gómez obtained his PhD in 2020 at Purdue University and is currently serving as an Assistant Professor of Spanish at the University of Wisconsin-La Crosse. His research compares the morphosyntax of traditional classroom learners and heritage speakers who learn their Spanish in the United States.

References

Abrahamsson

Hyltenstam

(2009). Age of onset and nativelikeness in a second language: Listener perception versus linguistic scrutiny. Language Learning, 59(2), 249–306.

Absi

Z. A. A.

(2014). Metalinguistic knowledge and speaking proficiency in Syrian university-level learners of English [Unpublished doctoral dissertation]. University of Essex.

Aissen

Perlmutter

D. M.

(1976). Clause reduction in Spanish. In Thompson

, et al Whistler

Edge

Jaeger

J. J.

Javelin

Patrick

Small

Van Valin

R. D.

Jr . (Eds.), Proceedings of the 2nd Annual Meeting of the Berkeley Linguistics Society (pp. 1–30). Berkeley Linguistics Society.

Akakura

(2011). Evaluating the effectiveness of explicit instruction on implicit and explicit L2 knowledge. Language Teaching Research, 16(1), 9–37.

Alarcón

I. V.

(2011). Spanish gender agreement under complete and incomplete acquisition: Early and late bilinguals’ linguistic behavior within the noun phrase. Bilingualism: Language and Cognition, 14(3), 332–350.

Aoun

Benmamoun

(1998). Minimality, reconstruction, and PF movement. Linguistic Inquiry, 29(4), 569–597.

T. K.

Knightly

L. M.

Jun

S.-A.

J. S.

(2002). Overhearing a language during childhood. Psychological Science, 13(3), 238–243.

Bates

Maechler

Bolker

Walker

(2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67(1), 1–48.

Berman

R. A.

(2004). Language development across childhood and adolescence. John Benjamins.

10.

Bialystok

Hakuta

(1994). In other words: The science and psychology of second-language acquisition. Basic Books.

11.

Bialystok

Ryan

E. B.

(1985). Toward a definition of metalinguistic skill. Merrill-Palmer Quarterly, 31(3), 229–251.

12.

Birdsong

(1992). Ultimate attainment in second language acquisition. Language, 68(4), 706–755.

13.

Birdsong

(2007). Nativelike pronunciation among late learners of French as a second language. In Bohn

O.-S.

Munro

M. J.

(Eds.), Language experience in second language speech learning: In honor of James Emil Flege (pp. 99–116). John Benjamins.

14.

Blake

(1983). Mood selection among Spanish-speaking children, ages 4 to 12. Bilingual Review / La Revista Bilingüe, 10(1), 21–32.

15.

Blasco Aznar

M. V

. (2002). On the acquisition of object pronominal clitics in Spanish (Order No. 3063805) [Doctoral dissertation]. ProQuest Dissertations and Theses Global.

16.

Bleam

(2000). Clitic Climbing and the Power of Tree Adjoining Grammar. In Abeillé

Rambow

(Eds.), Tree Adjoining Grammars: Formalisms, Linguistic Analysis and Processing (pp. 193–220). CSLI Publications.

17.

Bley-Vroman

(1990). The logical problem of Foreign language learning. Linguistic Analysis, 20(1–2), 3–49.

18.

Bobaljik

J. D.

(2002). A-chains at the PF-interface: Copies and “covert” movement. Natural Language & Linguistic Theory, 20(2), 197–267.

19.

Bok-Bennema

(2006). Clitic climbing. In Everaert

Riemsdijk

(Eds.), The Blackwell companion to syntax (pp. 469–518). Blackwell.

20.

Bongaerts

Mennen

van der Slik

(2000). Authenticity of pronunciation in naturalistic second language acquisition: The case of very advanced late learners of Dutch as a second language. Studia Linguistica, 54(2), 298–308.

21.

Borer

(1984). Parametric syntax. Foris.

22.

Bowles

M. A.

(2011). Measuring implicit and explicit linguistic knowledge: What can heritage language learners contribute? Studies in Second Language Acquisition, 33(2), 247–271.

23.

Brinton

D. M.

Kagan

Bauckus

(2008). Heritage language education: A new field emerging. Routledge.

24.

Bruhn de Garavito

(2002). Verb raising in Spanish: A comparison of early and late bilinguals. In Skarabela

Fish

A. H.-J.

(Eds.), Proceedings of the 26th annual Boston University conference on language development (pp. 84–94). Cascadilla Press.

25.

Bruhn de Garavito

(2013). What research can tell us about teaching: The case of pronouns and clitics. In Whong

Gil

K.-H.

Marsden

(Eds.), Universal grammar and the second language classroom (pp. 17–35). Springer.

26.

Bruhn de Garavito

Montrul

(1996). Verb-movement and clitic placement in French and Spanish as a second language. In Stringfellow

Cahana-Amitay

Hughes

Zukowski

(Eds.), Proceedings of the 20th annual Boston University conference on language development (pp. 123–134). Cascadilla Press.

27.

Campos

(1986). Indefinite object drop. Linguistic Inquiry, 17(2), 354–359.

28.

Carreira

Kagan

(2011). The results of the National Heritage Language Survey: Implications for teaching, curriculum design, and professional development. Foreign Language Annals, 44(1), 40–64.

29.

Chang

C. B.

Yao

Haynes

E. F.

Rhodes

(2011). Production of phonetic and phonological contrast by heritage speakers of Mandarin. The Journal of the Acoustical Society of America, 129(6), 3964–3980.

30.

Chomsky

(1995). The minimalist program. The MIT Press.

31.

Clements

J. C.

(2006). Null direct objects in Spanish. In Clements

J. C.

(Ed.), Functional approaches to Spanish syntax (pp. 134–150). Palgrave Macmillan.

32.

Correa

(2011a). Heritage language learners of Spanish: What role does metalinguistic knowledge play in their acquisition of the subjunctive? In Ortiz-López

L. A.

(Ed.), Selected proceedings of the 13th Hispanic linguistics symposium (pp. 128–138). Cascadilla Proceedings Project.

33.

Correa

(2011b). Subjunctive accuracy and metalinguistic knowledge of L2 learners of Spanish. Electronic Journal of Foreign Language Teaching, 8(1), 39–56.

34.

Cuza

Frank

(2015). On the role of experience and age-related effects: Evidence from the Spanish CP. Second Language Research, 31(1), 3–28.

35.

Cuza

Pérez-Leroux

A. T.

Sánchez

(2013). The role of semantic transfer in clitic drop among simultaneous and sequential Chinese-Spanish bilinguals. Studies in Second Language Acquisition, 35(1), 93–125.

36.

Davies

(1995). Analyzing syntactic variation with computer-based corpora: The case of modern Spanish clitic climbing. Hispania, 78(2), 370–380.

37.

DeKeyser

(2000). The robustness of critical period effects in second language acquisition. Studies in Second Language Acquisition, 22(4), 499–533.

38.

Duffield

White

(1999). Assessing L2 knowledge of Spanish clitic placement: Converging methodologies. Second Language Research, 15(2), 133–160.

39.

Eisenbeiss

(2010). Production methods in language acquisition research. In Blom

Unsworth

(Eds.), Experimental methods in language acquisition research (pp. 11–34). John Benjamins.

40.

Eisenberg

(2005). When conversation is not enough: Assessing infinitival complements through elicitation. American Journal of Speech-Language Pathology, 14, 92–106.

41.

Eisenchlas

(2003). Clitics in child Spanish. First Language, 23(2), 193–211.

42.

Ellis

N. C.

Larsen-Freeman

(2006). Language emergence: Implications for applied linguistics—Introduction to the special issue. Applied Linguistics, 27(4), 558–589.

43.

Ellis

(2004). The definition and measurement of L2 explicit knowledge. Language Learning, 54(2), 227–275.

44.

Ellis

(2005). Measuring implicit and explicit knowledge of a second language: A psychometric study. Studies in Second Language Acquisition, 27(2), 141–172.

45.

Ellis

(2006). Modelling learning difficulty and second language proficiency: The differential contributions of implicit and explicit knowledge. Applied Linguistics, 27(3), 431–463.

46.

Embick

Noyer

(2001). Movement operations after syntax. Linguistic Inquiry, 32(4), 555–595.

47.

Fernández Soriano

O. M

. (1993) Los pronombres átonos [Unstressed Pronouns in Spanish]. Taurus.

48.

Fernández Soriano

O. M

. (1999). El Pronombre Personal. Formas y Distribuciones. Pronombres Átonos y Tónicos [The Personal Pronoun in Spanish. Morphology and Distribution. Unstressed and Stressed Pronouns]. In Demonte

Bosque

(Eds.) Gramática Descriptiva de la Lengua Española (pp. 1209–1274). Espasa Calpe.

49.

Fischer

Goldbach

(2016). Object clitics. In Fischer

Gabriel

(Eds.), Manual of grammatical interfaces in Romance (pp. 363–390). Walter de Gruyter GmbH.

50.

Gerlach

(2002). Clitics between syntax and lexicon. John Benjamins.

51.

Gor

(2014). Raspberry, not a car: Context predictability and a phonological advantage in early and late learners’ processing of speech in noise. Frontiers in Psychology, 5(1449), 1–15.

52.

Grohmann

K. K.

Neokleous

(Eds.). (2014). Developments in the acquisition of clitics. Cambridge Scholars Publishing.

53.

Gropen

(2000). Methods for studying the production of argument structure in children and adults. In Menn

Bernstein Ratner

(Eds.), Methods for studying language production (pp. 95–113). Lawrence Erlbaum Associates.

54.

Grüter

Crago

(2012). Object clitics and their omission in child L2 French: The contributions of processing limitations and L1 transfer. Bilingualism: Language & Cognition, 15(3), 531–549.

55.

Gutiérrez

M. J.

(2008). Restringiendo la subida de clíticos: reflexividad, modalidad verbal y contacto lingüístico en el español de Houston [Restricting clitic climbing: reflexivity, verb modality and language contact in the Spanish of Houston]. Hispanic Research Journal, 9(4), 299–313.

56.

Gutiérrez

M. J.

(2014). La subida de clíticos en tiempo real y aparente en el español de Houston [Clitic climbing in real and apparent time in the Spanish of Houston]. In Enrique-Arias

Gutiérrez

M. J.

Landa

Ocampo

(Eds.), Perspectives in the study of Spanish language variation: Papers in honor of Carmen Silva-Corvalán (pp. 301–320). Servizo de Publicacións e Intercambio Científico, Universidade de Santiago de Compostela.

57.

Gutiérrez

M. J.

Silva-Corvalán

(1993). Clíticos en una situación de contacto [Clitics in a language contact situation]. Revista Española de Lingüística, 23(2), 207–220.

58.

Håkansson

(1995). Syntax and morphology in language attrition: A study of five bilingual expatriate Swedes. International Journal of Applied Linguistics, 5(2), 153–171.

59.

Halloran

Rothman

(2012). The acquisition of clitics in L2 Spanish: Examining restrictions on clitic solidarity. In Smith

Ihsane

(Eds.), Selected papers from the 42nd linguistic symposium on Romance languages (pp. 3–16). John Benjamins.

60.

Han

Ellis

(1998). Implicit knowledge, explicit knowledge and general language proficiency. Language Teaching Research, 2(1), 1–23.

61.

Hartshorne

J. K.

Tenenbaum

J. B.

Pinker

(2018). A critical period for second language acquisition: Evidence from 2/3 million English speakers. Cognition, 177, 263–277.

62.

Hawkins

Casillas

(2008). Explaining frequency of verb morphology in early L2 speech. Lingua, 118(4), 595–612.

63.

Hawkins

Franceschina

(2004). Explaining the acquisition and non-acquisition of determiner-noun gender concord in French and Spanish. In Prévost

Paradis

(Eds.), The acquisition of French in different contexts: Focus on functional categories (pp. 175–206). John Benjamins.

64.

Herschensohn

J. R.

(2009). Fundamental and gradient differences in language development. Studies in Second Language Acquisition, 31(2), 259–289.

65.

(2002). Psychological constraints on the utility of metalinguistic knowledge in second language production. Studies in Second Language Acquisition, 24(3), 347–386.

66.

Hulstijn

J. H.

(2005). Theoretical and empirical issues in the study of implicit and explicit second-language learning: Introduction. Studies in Second Language Acquisition, 27(2), 129–140.

67.

Jaeggli

(1982). Topics in Romance syntax. Foris Publications.

68.

Kayne

R. S.

(1975). French syntax: The transformational cycle. The MIT Press.

69.

Kayne

R. S.

(1989). Null subjects and clitic climbing. In Jaeggli

Safir

(Eds.), The null subject parameter (pp. 239–261). Kluwer Academic Publishers.

70.

Kayne

R. S.

(1991). Romance clitics, verb movement, and PRO. Linguistic Inquiry, 22, 647–686.

71.

Keating

G. D.

VanPatten

Jegerski

(2011). Who was walking on the beach? Anaphora resolution in Spanish heritage speakers and adult second language learners. Studies in Second Language Acquisition, 33(2), 193–221.

72.

Kim

J.-Y.

(2015). Perception and production of Spanish lexical stress by Spanish heritage speakers and English L2 learners of Spanish. In Willis

E. W.

Butragueño

P. M.

Zendejas

E. H.

(Eds.), Selected proceedings of the 6th conference on laboratory approaches to Romance phonology (pp. 106–128). Cascadilla Press.

73.

Kim

J. H.

Montrul

Yoon

(2009). Binding interpretations of anaphors by Korean heritage speakers. Language Acquisition, 16(1), 3–35.

74.

Klavans

(1995). On clitics and cliticization: The interaction of morphology, phonology, and syntax. Garland.

75.

Knightly

L. M.

Jun

S.-A.

J. S.

T. K.

(2003). Production benefits of childhood overhearing. The Journal of the Acoustical Society of America, 114(1), 465–474.

76.

Krashen

S. D.

(1973). Lateralization, language learning, and the critical period: Some new evidence. Language Learning, 23(1), 63–74.

77.

Labov

(1996). When intuitions fail. In McNair

Singer

Dobrin

Aucoin

(Eds.), CLS 32: Papers from the parasession on theory and data in linguistics (pp. 77–105). Chicago Linguistics Society.

78.

Larrañaga

Guijarro-Fuentes

(2012). Clitics in L1 bilingual acquisition. First Language, 32(1–2), 151–175.

79.

Lenneberg

E. H.

(1967). The biological foundations of language. Hospital Practice, 2(12), 59–67.

80.

Liceras

(1985). The value of clitics in non-native Spanish. Second Language Research, 1(2), 151–168.

81.

Limerick

P. P.

(2017). Variable clitic placement in US Spanish. In MacDonald

J. E.

(Ed.), Contemporary trends in Hispanic and Lusophone linguistics: Selected papers from the Hispanic linguistic symposium 2015 (pp. 49–70). John Benjamins.

82.

Long

. (2005). Problems with supposed counter-evidence to the critical period hypothesis. International Review of Applied Linguistics in Language Teaching, 43(4), 287–317.

83.

López Ornat

. (1994). La Adquisición del Lenguaje: Talón de Aquiles y Poción Mágica de la Teoría Cognitiva [Language Acquisition: The Achilles Tendon and Magic Potion of Cognitive Theory]. Cognitiva, 6(2), 213–239.

84.

Lüdecke

(2018). sjPlot—Data visualization for statistics in social science (Version 2.6.2). https://CRAN.R-project.org/package=sjPlot

85.

Luján

(1980). Clitic promotion and mood in Spanish verbal complements. Linguistics, 18(5–6), 381–484.

86.

Lukyanchenko

Gor

(2011). Perceptual correlates of phonological representations in heritage speakers and L2 learners. In Danis

Mesh

Sung

(Eds.), Proceedings of the 35th annual Boston University conference on language development (pp. 414–426). Cascadilla Press.

87.

Lynch

(2008). The linguistic similarities of Spanish heritage and second language learners. Foreign Language Annals, 41(2), 252–381.

88.

Mateu

V. E.

(2015). Object clitic omission in child Spanish: Evaluating representational and processing accounts. Language Acquisition, 22(3), 240–284.

89.

Meijer

P. J.

Fox Tree

J. E.

(2003). Building syntactic structures in speaking: A bilingual exploration. Experimental Psychology, 50(3), 184–195.

90.

Mikulski

A. M.

(2010a). Age of onset of bilingualism, language use, and the volitional subjunctive in heritage learners of Spanish. Heritage Language Journal, 7(1), 28–46.

91.

Mikulski

A. M.

(2010b). Receptive volitional subjunctive abilities in heritage and traditional foreign language learners of Spanish. The Modern Language Journal, 94(2), 217–233.

92.

Montrul

. (2002). Incomplete acquisition and attrition of Spanish tense/aspect distinctions in adult bilinguals. Bilingualism, 5(1), 39–68.

93.

Montrul

(2005). On knowledge and development of unaccusativity in Spanish L2 acquisition. Linguistics, 43(6), 1153–1190.

94.

Montrul

(2006). On the bilingual competence of Spanish heritage speakers: Syntax, lexical-semantics and processing. International Journal of Bilingualism, 10(1), 37–69.

95.

Montrul

(2008). Incomplete acquisition in bilingualism: Re-examining the age factor. John Benjamins.

96.

Montrul

(2010a). Dominant language transfer in adult second language learners and heritage speakers. Second Language Research, 26(3), 293–327.

97.

Montrul

(2010b). How similar are adult second language learners and Spanish heritage speakers? Spanish clitics and word order. Applied Psycholinguistics, 31(1), 167–207.

98.

Montrul

(2012). The grammatical competence of Spanish heritage speakers. In Beaudrie

S. M.

Fairclough

(Eds.), Spanish as a heritage language in the US: State of the science (pp. 101–120). Georgetown University Press.

99.

Montrul

(2016). The acquisition of heritage languages. Cambridge University Press.

100.

Montrul

Bowles

(2010). Is grammar instruction beneficial for heritage language learners? Dative case marking in Spanish. Heritage Language Journal, 7(1), 47–73.

101.

Montrul

De La Fuente

Davidson

Foote

(2013). The role of experience in the acquisition and production of diminutives and gender in Spanish: Evidence from L2 learners and heritage speakers. Second Language Research, 29(1), 87–118.

102.

Montrul

Foote

Perpiñán

(2008). Gender agreement in adult second language learners and Spanish heritage speakers: The effects of age and context of acquisition. Language Learning, 58(3), 503–553.

103.

Montrul

Slabakova

(2003). Competence similarities between native and near-native speakers: An investigation of the preterite-imperfect contrast in Spanish. Studies in Second Language Acquisition, 25(3), 351–398.

104.

Myhill

(1989). Variation in Spanish clitic climbing. In Walsh

T. J

(Ed.), Synchronic and diachronic approaches to linguistic variation and change (pp. 227–250). Georgetown University Press.

105.

Napoli

D. J.

(1981). Semantic interpretation vs. lexical governance: Clitic climbing in Italian. Language, 57(4), 47–69.

106.

Norris

J. M.

Ortega

(2001). Does type of instruction make a difference? Substantive findings from a meta-analytic review. Language Learning, 51(s1), 157–213.

107.

O’Grady

Lee

Choo

(2001). The acquisition of relative clauses by heritage and non-heritage learners of Korean as a second language: A comparative study. Journal of Korean Language Education, 12(2), 283–294.

108.

O’Grady

Kwak

H. Y.

Lee

O. S.

Lee

(2011). An emergentist perspective on heritage language acquisition. Studies in Second Language Acquisition, 33(2), 223–245.

109.

J. S.

Jun

S.-A.

Knightly

L. M.

T. K.

(2003). Holding on to childhood language memory. Cognition, 86(3), B53–B64.

110.

Ordóñez

(2002). Some clitic combinations in the syntax of Romance. Catalan Journal of Linguistics, 1, 201–224.

111.

Ordóñez

(2012). Clitics in Spanish. In Hualde

J. A.

Olarrea

O’Rourke

(Eds.), The handbook of Hispanic linguistics (pp. 423–451). Wiley-Blackwell.

112.

Ormazabal

Romero

(2013). Object clitics, agreement and dialectal variation. International Journal of Latin and Romance Linguistics, 25(2), 301–344.

113.

Paradis

(2004). A neurolinguistic theory of bilingualism. John Benjamins.

114.

Paradis

(2009). Declarative and procedural determinants of second languages. John Benjamins.

115.

Paradis

Genesee

Crago

M. B.

(2011). Dual language development & disorders: A handbook on bilingualism and second language learning. Paul H. Brookes Publishing.

116.

Peace

M. M.

(2012). ¿Lo puedo subir o puedo subirlo? La subida del clítico en el español del oeste de Massachusetts [Can I climb or can’t I? Clitic climbing in the Spanish of West Massachusetts]. Southwest Journal of Linguistics, 31(1), 131–160.

117.

Penfield

Roberts

(1959). Speech and brain mechanisms. Princeton University Press.

118.

Pérez-Leroux

A. T.

Cuza

Thomas

(2011). Clitic placement in Spanish–English bilingual children. Bilingualism: Language and Cognition, 14(2), 221–232.

119.

Picallo

(1990). Modal verbs in Catalan. Natural Language and Linguistic Theory, 8(2), 285–312.

120.

Pinker

(1994). The language instinct: How the mind creates language. William Morrow.

121.

Polinsky

(2006). Incomplete acquisition: American Russian. Journal of Slavic Linguistics, 14(2), 191–262.

122.

Polinsky

Kagan

(2007). Heritage languages: In the “wild” and in the classroom. Language and Linguistics Compass, 1(5), 368–395.

123.

Potowski

Jegerski

Morgan-Short

(2009). The effects of instruction on linguistic development in Spanish heritage language speakers. Language Learning, 59(3), 537–579.

124.

Quicoli

(1976). Conditions on clitic movement in Portuguese. Linguistic Analysis, 2(3), 199–223.

125.

R Core Team (2019). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R-project.org/

126.

Raposo

E. P.

Uriagereka

(2005). Clitic placement in Western Iberian: A minimalist view. In Cinque

Kayne

R. S.

(Eds.), The Oxford handbook of comparative syntax (pp. 639–697). Oxford University Press.

127.

Requena

P. E.

(2015). Direct object clitic placement preferences in Argentine child Spanish [Unpublished doctoral dissertation]. Pennsylvania State University.

128.

Requena

P. E.

Dracos

(2018). Impermeability of L1 syntax: Spanish variable clitic placement in bilingual children. In Bertolini

A. B.

Kaplan

M. J.

(Eds.), Proceedings of the 42nd annual Boston University conference on language development (pp. 644–658). Cascadilla Press.

129.

Rivas

A. M.

(1977). A theory of clitics (Order No. 0336009) [Doctoral dissertation]. ProQuest Dissertations and Theses Global.

130.

Rizzi

(1982). Issues in Italian syntax. Foris.

131.

Roberts

(1997). Restructuring, head movement and locality. Linguistics Inquiry, 28, 423–460.

132.

Roberts

(2010). Agreement and head movement: Clitics, incorporation and defective goals. The MIT Press.

133.

Roca

(1992). Object Clitics in Spanish and Catalan. Catalan Working Papers in Linguistics, 2, 245–280

134.

Rodríguez-Mondoñedo

Snyder

Sugisaki

(2005). The parameter of clitic-climbing: The view from child Spanish. In Deem

K. U.

Nomura

Schulz

Schwartz

B. D.

(Eds.), The proceedings of the inaugural conference on generative approaches to language acquisition North America, Honolulu (Vol. 2, pp. 241–248). MIT Working Papers in Linguistics.

135.

Romain

I. J.

(2015). A Phase Approach to Spanish Object Clitics (Order No. 3689735). [Doctoral Dissertation, University of California, Los Angeles]. ProQuest Dissertations & Theses Global.

136.

Rooryck

(2000). Configurations of sentential complementation: Perspectives from Romance languages. Routledge.

137.

Rosen

S. T.

(1989). Argument structure and complex predicates [Unpublished dissertation]. Brandeis University.

138.

Rossi

Diaz

Kroll

J. F.

Dussias

P. E.

(2017). Late bilinguals are sensitive to unique aspects of second language processing: Evidence from clitic pronouns word-order. Frontiers in Psychology, 8(342), 1–13.

139.

Rothman

Tsimpli

I. M.

Pascual

Cabo

(2016). Formal linguistic approaches to heritage language acquisition: Bridges for pedagogically oriented research. In Pascual

Cabo

(Ed.), Advances in Spanish as a heritage language (pp. 13–27). John Benjamins.

140.

Saadah

(2011). The production of Arabic vowels by English L2 learners and heritage speakers of Arabic [Unpublished doctoral dissertation]. University of Illinois at Urbana–Champaign.

141.

SAS Institute Inc. (1999). SAS/ACCESS® Version 9.3 Procedure Guide. SAS Institute Inc.

142.

Sauerland

Elbourne

(2002). Total reconstruction, PF movement, and derivational order. Linguistic Inquiry, 33(2), 283–319.

143.

Schmid

M. S.

Jarvis

(2014). Lexical access and lexical diversity in first language attrition. Bilingualism: Language and Cognition, 17(4), 729–748.

144.

Schwartz

B. D.

Sprouse

R. A.

(1996). L2 cognitive states and the full transfer/full access model. Second Language Research, 12(1), 40–72.

145.

Schwenter

S. A.

(2006). Null objects across America. In Face

T. L.

Klee

C. A.

(Eds.), Selected proceedings of the 8th Hispanic linguistics symposium (pp. 23–36). Cascadilla Proceedings Project.

146.

Schwenter

S. A.

Cacoullos

R. T.

(2014). Competing constraints on the variable placement of direct object clitics in Mexico City Spanish. Revista Española de Lingüística Aplicada/Spanish Journal of Applied Linguistics, 27(2), 514–536.

147.

Scovel

(1988). A time to speak: A psycholinguistic inquiry into the critical period for human speech. Newbury House.

148.

Seliger

(1978). Implications of a multiple critical periods hypothesis for second language learning. In Ritchie

W. C.

(Ed.), Second language acquisition research: Issues and implications (pp. 11–19). Academic Press.

149.

Shin

N. L.

Requena

P. E.

Kemp

(2017). Bilingual and monolingual children’s patterns of syntactic variation: Variable clitic placement in Spanish. In Benavides

A. A.

Schwartz

R. G.

(Eds.), Language development and disorders in Spanish-speaking children (pp. 63–88). Springer.

150.

Silva-Corvalán

(1994). Language contact and change: Spanish in Los Angeles. Clarendon Press.

151.

Silva-Corvalán

(2003). Linguistic consequences of reduced input in bilingual first language acquisition. In Montrul

Ordóñez

(Eds.), Linguistic theory and language development in Hispanic languages: Papers from the 5th Hispanic linguistics symposium and the 4th conference on the acquisition of Spanish and Portuguese (pp. 375–397). Cascadilla Press.

152.

Silva-Corvalán

Gutiérrez

M. J.

(1995). On transfer and simplification: Verbal clitics in Mexican-American Spanish. In Hashemipour

Maldonado Soto

Van Naerssen

Terrell

T. D.

(Eds.), Studies in language learning and Spanish linguistics in honor of Tracy D. Terrell (pp. 302–312). McGraw Hill.

153.

Singleton

D. M.

Ryan

(2004). Language acquisition: The age factor. Multilingual Matters.

154.

Spada

Tomita

(2010). Interactions between type of instruction and type of language feature: A meta-analysis. Language Learning, 60(2), 263–308.

155.

Strozer

J. R.

(1976). Clitics in Spanish [Unpublished doctoral dissertation]. University of California, Los Angeles

156.

Suñer

(1988). The role of agreement in clitic-doubled constructions. Natural Language and Linguistic Theory, 6(3), 391–434.

157.

Suñer

Yépez

(1988). Null definite objects in Quiteño. Linguistic Inquiry, 19(3), 511–519.

158.

Thomas

D. L. T.

(2012). Grammatical optionality and variability in bilingualism: How Spanish-English bilinguals limit clitic-climbing (Order No. NR97479) [Doctoral dissertation]. ProQuest Dissertations and Theses A&I.

159.

Torrego

(1998). Nominative Subjects and Pro-Drop INFL. Syntax, 1(2), 206–219.

160.

Torres Cacoullos

. (1999). Construction frequency and reductive change: Diachronic and register variation in Spanish clitic climbing. Language Variation and Change, 11(2), 143–170.

161.

Ungureanu

M.-L.

(2011). Resetting the functional features of the accusative clitic in the second language acquisition of Spanish. Papers from the Annual Meetings of the Atlantic Provinces Linguistic Association (PAMAPLA), 35, 46–64.

162.

Unsworth

(2005). Investigating subject-specific literacies in school learning. In Unsworth

(Ed.), Researching language in schools and communities: Functional linguistic perspectives (pp. 245–274). Cassel.

163.

Uriagereka

(1995). Aspects of the syntax of clitic placement in Western Romance. Linguistic Inquiry, 26(1), 79–123.

164.

Valdés

(2001). Heritage language students: Profiles and possibilities. In Peyton

J. K.

Ranard

D. A.

McGinnis

(Eds.), Heritage languages in America: Preserving a national resource (pp. 41–49). Delta Systems.

165.

Valdés

Fishman

J. A.

Chávez

Pérez

(2006). Developing minority language resources: The case of Spanish in California. Multilingual Matters.

166.

Wexler

(1996). The development of inflection in a biologically based theory of language acquisition. In Rice

M. L.

(Ed.), Toward a genetics of language (pp. 113–144). Lawrence Erlbaum.

167.

Wexler

(1998). Very early parameter setting and the unique checking constraint: A new explanation of the optional infinitive stage. Lingua, 106(1–4), 23–79.

168.

White

(1998). Clitics in L2 French. In Clahsen

(Ed.), Generative perspectives on language acquisition: Empirical findings, theoretical considerations and crosslinguistic comparisons (pp. 335–369). John Benjamins.

169.

White

(2003). Grammars beyond the initial state: Parameters and functional categories. In White

(Ed.), Second language acquisition and universal grammar (pp. 100–150). Cambridge University Press.

170.

White

Genesee

(1996). How native is near-native? The issue of ultimate attainment in adult second language acquisition. Second Language Research, 12(3), 233–265.

171.

Yeni-Komshian

Flege

J. E.

Liu

(2000). Pronunciation proficiency in the first and second languages of Korean–English bilinguals. Bilingualism: Language & Cognition, 3(2), 131–149.

172.

Zagona

(2002). The syntax of Spanish. Cambridge University Press.

173.

Zwicky

A. M.

(1994). What is a clitic? In Nevis

J. A.

Joseph

B. D.

Wanner

Zwicky

A. M.

(Eds.), Clitics: A comprehensive bibliography 1892-1991 (pp. xii–xx). John Benjamins.