On the grammaticality of passivization in bilingual compound verbs

Abstract

Objectives/research questions:

We examined stative and eventive passive bilingual compound verbs (BCVs) in Spanish/English code-switching. Of particular interest to us was the availability of passivization in bilingual eventive passive hacer “do” constructions, purportedly banned in bilingual speech due to a universal syntactic restriction.

Methodology:

A total of 119 bilinguals from Northern Belize and 36 from Southwest United States completed a two-alternative forced-choice acceptability task and a language background questionnaire.

Data and analysis:

The analysis was conducted using Thurstone’s Law of Comparative Judgment.

Conclusion:

For stative passive BCVs, results revealed that Spanish/English bilinguals from both contexts gave the highest ratings to code-switched constructions without the light verb hacer. For eventive passive BCVs, however, Belize bilinguals gave preferential ratings to passive constructions with the light verb hacer. Conversely, US bilinguals rejected them. Notably, among Belize bilinguals, eventive passive BCVs that were rated as most acceptable were constructions with no gender agreement between the light verb and the feminine antecedent noun.

Originality:

This is the first cross-community analysis that investigates stative and eventive passive BCVs in Spanish/English code-switching.

Implications:

Our findings show that the light verb hacer is compatible with both stative and eventive passive BCVs. Crucially, context-specific linguistic norms and social factors rather than a universal syntactic restriction primarily determine the availability of passivization in eventive passive BCVs. Our theorizing of code-switching grammars, thus, necessitates careful consideration of invariant and variable production patterns that are profoundly shaped by historical and sociolinguistic conditions.

Keywords

Spanish/English code-switching bilingual morphosyntax bilingual compound verbs (BCVs)stative and eventive passives grammatical gender

Introduction

In the last decade, the study of bilingual compound verbs (BCVs) in Spanish/English code-switching (CS) has sparked empirical questions concerning where, how, and why these hybrid structures are used in bi/multilingual speech. While hacer “do” BCVs have been noted for their Creole-like features (Pfaff, 1979), productivity (Balam, 2015; Vergara Wilson, 2013), and linguistic creativity (Balam, 2015, 2016a), estar “be” BCVs have been shown to be rare in oral production despite their congruence with pre-existing templates in monolingual discourse (Balam, Parafita Couto, & Chen, 2021; Guzzardo Tamargo, 2012). Notably, ser “be” BCVs, which are typically attested in passive constructions (Balam & Prada Pérez, 2013), today remain the most poorly understood bilingual verb construction.

The term BCV refers to syntactic combinations generally comprising two verbs, in which a fully inflected “auxiliary” (Backus, 1996), “operator¹” (Romaine, 1986), or light verb from the recipient language co-occurs with a donor language infinitive verb that mainly provides lexical meaning. In Spanish/English CS, for instance, the fully inflected Spanish light verb hacer co-occurs with an English infinitive verb, as in (1). Whereas hacer bears tense, aspect, and mood (TAM) features, the English verb contributes meaning (Balam, 2015, 2016b; Balam et al., 2020; Reyes, 1982; Vergara Wilson, 2013). Importantly, contiguous elements in the BCV refer to a single action or event:

(1) Hacemos get por qué. . .

Do.1PL.PRES get.INF why²

“We understand why . . .”

(Data from Belize: Balam & Prada Pérez, 2016, p. 274)

Note that because of the occurrence of the “be” operator verb in these types of constructions in other CS varieties (e.g., irundadu “be” in Tamil/English CS: Annamalai, 1989; hᴐwa “be” in Bengali/English CS: Chatterjee, 2016; hona “be” in Hindi/English CS: Kachru, 1978), specifically in passive forms, the term BCV in our paper is inclusive of syntactic units in which the operator verb is the Spanish light verb hacer or the auxiliary verbs estar and ser. The primary characteristic that distinguishes estar and ser BCVs from hacer BCVs is the way in which aspectual information is encoded (for relevant discussion, see Balam, Parafita Couto, & Chen, 2021). As (2) and (3) illustrate, in estar and ser BCVs, the Spanish auxiliary verb co-occurs with an English participle that contributes not only lexical but aspectual meaning as well (i.e., -ing encodes imperfective aspect whereas -ed encodes perfective aspect):

(2) O talvez están developing un curso nuevo. . .

Or maybe be.3PL.PRES developing.PROG a course new

“Or maybe (they) are developing a new course. . .”

(Data from Miami, FL: Guzzardo Tamargo, 2012, p. 47)

(3) El package fue delivered esta mañana

The package be.3SG.PRET delivered.PSTPART this morning

“The pack was delivered this morning.”

(Previously unpublished example from Tampa, FL: Balam, 2016a)

In terms of their morphosyntactic structure and geographic distribution, BCVs in Spanish/English CS differ in important ways. Whereas hacer BCVs lack structural equivalents in the component languages, estar and ser BCVs have monolingual templates in Spanish and English (Balam, Parafita Couto, & Chen, 2021; Moyer, 1992; Pfaff, 1979). In naturalistic discourse, hacer is particularly ubiquitous and preferred in Belize as the default operator verb to incorporate English verbs in code-switched speech (Balam, 2015, 2016b; Balam et al., 2020; Balam, Parafita Couto, & Chen, 2021), whereas estar is favored in Spanish/English CS varieties in the United States (Balam et al., 2020; Halberstadt, 2017). In the case of ser BCVs, to our knowledge, no previous study has examined these constructions in Spanish/English CS; thus, much less is known about the cross-community use of these highly infrequent structures in naturalistic discourse. Importantly, differences in the geographic distribution of BCVs raise important concomitant questions regarding speakers’ linguistic competence and oral production patterns, which may vary not only across individuals but communities as well.

Endeavoring to contribute to our understanding of BCVs in Spanish/English CS communities, we investigate the grammaticality of two different types of stative and eventive passives, namely, those without hacer and those with hacer (for further details on passives in monolingual and Spanish/English CS, see section “Passives in monolingual Spanish and Spanish/English code-switching”). As (4) shows, there are stative passive BCVs in which only the auxiliary verb estar co-occurs with the English past participle. Conversely, there are stative passive BCVs, as in (5) and (6), in which the auxiliary verb estar co-occurs with the past participle of hacer (i.e., hecho) and an English past participle (i.e., considered; ostracized):

(4) El pobre hombre está confused.

The poor man be.3SG.PRES confused.PSTPTCP

“The poor man is confused.”

(Data from Gibraltar: Moyer, 1992, p. 204)

(5) En Sixth Form, ‘tas hecho considered un adult

In Sixth Form be.2SG do.2SG.PSTPTCP.M considered.PSTPTCP an adult

“In Sixth Form, you are considered an adult.”

(Data from Belize: Balam et al., 2014, p. 254)

(6) Losotros estamos hechos ostracized from the group. . .

We be.3PL.PRES do.3PL.PSTPART.M ostracized from the group

“We are ostracized from the group. . .”

(Previously unpublished example from Northern Belize: Balam, 2016a)

While stative passive BCVs, as in (4), are well attested in several Spanish/English corpora (e.g., Belize: Balam, Parafita Couto, & Chen, 2021; Gibraltar: Moyer, 1992; Southwest United States: Pfaff, 1979), there is scant documentation of passive BCVs in which hacer is incorporated. Passive BCVs, as in (5) and (6), have only been attested in Belize. These hybrid structures are innovative as the inflected Spanish auxiliary verb carries tense and mood features whereas the hacer past participle functions as an aspectual marker that may be marked for grammatical gender and/or number as (5) and (6) illustrate, respectively (Balam, Parafita Couto, & Chen, 2021). Note that in both cases, there is a gender and/or number agreement dependency between the light verb hacer, the sentential subject, and the inflected auxiliary verb.

Of relevance to our study is that although hacer BCVs have been previously noted for their syntactic productivity both in Belize (Balam, 2015, 2016b; Balam & Prada Pérez, 2016) and Southwest United States (Halberstadt, 2017; Reyes, 1982; Vergara Wilson, 2013), their occurrence is particularly controversial with eventive passives in code-switched speech, an issue we address here. Basing their analysis on interview and judgment data from Spanish/German CS, González-Vilbazo and López (2011) postulate that eventive passive BCVs are incompatible with the light verb hacer due to a syntactic constraint “imposed on the computational system rather than a language-speciﬁc—or code-switching speciﬁc—type of restriction” (p. 843). The ungrammaticality of eventive passive BCVs with hacer, therefore, results from a syntactic restriction that is universal in nature (for details, see section “Passives in monolingual Spanish and Spanish/English code-switching”).

Our goal in the present analysis was twofold. First, we investigated the availability of passivization in stative and eventive BCVs in Spanish/English CS. Second, we examined the status of grammatical gender in passive BCVs with hacer. While previous studies have pointed out that hacer bears TAM features (Balam, 2015, 2016b; González-Vilbazo & López, 2011; Reyes, 1982), this is the first study to examine hacer in bilingual passives as a carrier of grammatical gender. To further shed light on the role that community norms have on the acceptability of passive BCVs, we examine here how speakers’ acceptability judgments differ based on their exposure to Spanish/English CS varieties in Northern Belize and Southwest United States, two sociolinguistic contexts that are well known for the use of hacer BCVs. Before describing our methodology, we first provide an overview of monolingual and bilingual passive constructions. Subsequently, we outline key findings in antecedent work on BCVs in Spanish/English CS.

Passives in monolingual Spanish and Spanish/English code-switching

In Spanish, there are three main types of passives: stative, eventive, and pseudo-passives. Our main concern here is with stative (adjectival) and eventive (verbal or analytical) passives, which are distinguished via the use of the copulas estar and ser, a verbal difference encoded in Spanish but not in English (Bruhn de Garavito & Valenzuela, 2008; Luján, 1981). In stative or adjectival passives, as (7) illustrates, the auxiliary estar co-occurs with a Spanish past participle (i.e., abandonada_fem “abandoned”) that matches in grammatical gender with its antecedent noun (i.e., casa_fem “house”):

(7) La casa estaba abandonada.

The house.3SG.F be.3SG.IMP abandon.PSTPTCP.F

“The house was abandoned.”

This salient gender agreement feature is also present in eventive or verbal passives, as in (8), in which ser co-occurs with a gender-marked past participle that agrees with the embedded noun. Syntactically, adjectival and eventive passives differ, as only the latter involves raising of the theme to subject position (for syntactic analyses of adjectival and eventive passives, see Wasow, 1977, and Levin & Rappaport, 1986, respectively):

(8) La casa fue construida.

The house.3SG.F be.3SG.PRET build.PSTPTCP.F

“The house was built.”

In antecedent work, the frequent use of eventive passives, as in (8), is generally associated with Spanish/English communities in the United States (Lipski, 2008, p. 67). This contrasts the preferential use of pseudo-passive constructions with the impersonal clitic se (e.g., Se contruyó la casa “The house was built”), which is characteristic of monolingual Spanish communities (Bruhn de Garavito & Valenzuela, 2008, p. 326, Mendikoetxea, 2012). Crucially, eventive passive constructions with the Spanish light verb hacer, as in (9), have been reported as strictly banned in code-switched speech:

(9) El army fue hecho defeated.

The.3SG army be.3SG do.PSTPTCP.M defeated.PSTPTCP

“The army was defeated.”

(Data from Belize: Balam, 2016a, p. 196)

The restriction on passivization in eventive passive hacer BCVs was proposed by González-Vilbazo and López (2011), who examined four puzzling syntactic characteristics of “hacer + V” constructions. As it relates to the passive voice, González-Vilbazo and López advance that the Spanish light verb hacer can co-occur with intransitive German verbs, as in (10), and it can also occur with transitive and unaccusative predicates. Importantly, it cannot co-occur with German verbs in eventive passive constructions, as in (11):

(10) Juan hace schlafen

Juan do.3SG.PRES sleep

“Juan sleeps.”

(11) *El libro fue hecho lesen.

The book be.3SG.PRET do.PSTPTCP.M read

“The book was read.”

(Data from Barcelona: González-Vilbazo & López, 2011, p. 843)

In González-Vilbazo and López’s view, when a German verbal root is brought into the computational system, as lauf- “run” in (12), it cannot value or satisfy Spanish little v’s conjugation class feature because it lacks this feature. Thus, the light verb hacer must be inserted to save the bilingual derivation. In (13), the configuration is grammatical as hacer does bear a conjugation feature, so it is able to satisfy the [uConj] feature of v_sp (González-Vilbazo & López, 2011, p. 842). The v_sp acquires a conjugation class feature and becomes a morphological base to which TAM features are attached

In the case of the active verb in (13), the light verb hacer spells out as little v, and it does not incorporate to the lexical verb which it selects. Note that with monolingual Spanish verbs (e.g., cant-ar “to sing,” camin-ar “to walk”), incorporation occurs because the verbal roots have matching features. Thus, the roots value the [uConj] feature in little v and then move up, as (14) shows, to incorporate to little v (González-Vilbazo & López, 2011, p. 841):

In the case of eventive passive BCVs, however, the lexical verb remains unincorporated, as little v is realized as hacer. This is wherein the problem lies. In a grammatical construction, the internal argument would be able to move up to Spec, v or subject position. The lexical verb satisfies the conjugation class feature in a monolingual Spanish verbal construction and moves up to adjoin to little v; hence, creating a space for a constituent to move up in the syntactic tree. In the case of bilingual eventive passive BCVs with hacer, however, this space is not available given that little v is phonetically spelled out as hacer.

The raising of the internal argument to subject position would violate the Relativized Minimality/Minimal Link Condition (for further discussion, see González-Vilbazo & López, 2011, p. 838). As a result of this syntactic restriction, the light verb hacer is incompatible with eventive passives. This restriction, however, does not apply to eventive passives without hacer. It is also not applicable to stative passive BCVs, as in these constructions the subject is an external argument of the verb and not an internal argument that has undergone movement from the object position. Thus, stative passives without hacer (e.g., “tas considered ‘you are considered. . .’”) or with the light verb (e.g., “tas hecho considered ‘you are considered. . .’”) are grammatical forms (for a syntactic analysis that accounts for “estar + English present participle” constructions, see López et al., 2017).

BCVs in Spanish/English CS communities

Within the last decade, several studies have provided valuable insight into the use of hacer and/or estar BCVs in the Spanish/English CS varieties of Northern Belize (e.g., Balam, 2016b; Balam, Parafita Couto, & Chen, 2021; Balam & Prada Pérez, 2016) and the US Hispanophone context (e.g., Guzzardo Tamargo, 2012; Halberstadt, 2017; Vergara Wilson, 2013). Most of these studies, however, have focused on analyzing data from a specific contact situation. For example, Balam (2015) examined 1,750 hacer BCVs extracted from interviews conducted with 62 bi/multilinguals (ages 14–99) from Northern Belize. This analysis revealed not only a remarkably high token frequency of hacer BCVs among younger generations, but it showed that hacer BCVs have evolved across generations, expanding to novel syntactic (e.g., copulative, control) and pronominal (e.g., reflexive, reciprocal) contexts, including the stative and eventive passives that are the focus of the present study.

In an analysis of oral narratives from New Mexican Spanish/English bilinguals, Halberstadt (2017) examined the production of four bilingual verb structures (i.e., estar + gerund; hacer + verb; Spanish noun phrase (NP) + verb; Spanish pronoun + verb). Halberstadt’s study showed that out of the total numbers of tokens extracted, 5% (129/2702) of the data were code-switched. Notably, estar BCVs (e.g., Él estaba looking at her “He was looking at her”) and “NP + verb” constructions (e.g., Dos hombres leave “Two men leave”) were the most frequent verb switches, constituting 45% (n = 58) and 30% (n = 39) of the code-switched data, respectively. Halberstadt’s findings revealed that estar BCVs are the most prevalent verb switch among Spanish/English bilinguals from New Mexico.

More recent research conducted from a comparative lens, however, elucidates the effect that language experience has on the acceptability of BCVs across communities. Balam et al. (2020) investigated the acceptability of canonical and non-canonical hacer and estar BCVs in present progressive constructions. Drawing on intuitional data from Spanish/English bilinguals (n = 106) from Northern Belize and two US contexts (New Mexico and Puerto Rico), they found that speakers’ language experience influenced bilinguals’ judgments of BCVs. Whereas New Mexico and Puerto Rico bilinguals gave the most preferential ratings to canonical estar BCVs (e.g., La secretaria está auditing el report “The secretary is auditing the report”), Northern Belize bilinguals gave the most preferential ratings to canonical hacer BCVs in progressive constructions (e.g., La secretaria está haciendo audit el report “The secretary is auditing the report”).

Importantly, only Northern Belize and New Mexico bilinguals who use hacer constructions accepted hacer constructions, which suggests that exposure to and/or use of hacer BCVs are necessary conditions for speakers to develop intuitional knowledge about these structures. Importantly, estar constructions do not appear to be as constrained by speakers’ exposure to community linguistic norms, as these BCVs were rated, to different degrees, as acceptable by all bilingual groups. These results suggest that context-specific language practices play an important role in determining the employment of BCVs.

Noteworthy is that even though previous studies have shown the productive use of Spanish/English BCVs in Northern Belize and Southwest United States, bilingual stative and eventive passive constructions with hacer are highly infrequent (Balam, 2015, 2016a). A question that arises, therefore, is whether ser BCVs are indeed ungrammatical forms, as González-Vilbazo and López (2011) assert, or whether they constitute an idiolectal or community-specific phenomenon. Another question that arises relates to the status of grammatical gender. As previously pointed out, a distinctive feature of bilingual passive hacer BCVs is that the light verb seemingly functions not only as a morphological marker of aspect but grammatical gender as well. In the present study, we explore whether an invariable masculine-marked light verb hecho is employed in these forms as a result of the application of the masculine default gender, which is well attested in Spanish/English mixed nominal constructions (e.g., Balam, 2016c; Balam, Lakshmanan, & Parafita Couto, 2021; Clegg & Waltermire, 2009, and references therein) but not Spanish/English bilingual verbs.

In the ensuing section, we explain how we examined (1) the availability of passivization in stative and eventive passive BCVs; and (2) speakers’ adherence or not to the maintenance of normative gender agreement patterns in bilingual passive BCVs.

The current study

We collected and analyzed intuitional data from two communities where Spanish is in intense contact with English, namely, Northern Belize and Southwest United States.

Communities under study

Northern Belize and Southwest United States offer fertile ground to examine stative and eventive passive BCVs as estar and hacer BCVs have been attested in both contexts. Belize is an officially English-speaking Central American/Caribbean country where in recent decades Spanish has risen as the language of the majority (for further discussion, see Balam, 2014, 2016b). Importantly, the varieties that thrive today as distinctive linguistic identity markers among younger generations in Northern Belize are Spanish/English CS and Belizean Kriol (Balam & Prada Pérez, 2017). Conversely, in Southwest United States, Spanish and English co-exist on a daily basis with a strong Spanish heritage factor that shapes the unique “blended” culture of the region due to the constant contact between English and different monolingual varieties of Spanish. Although both contexts are characterized by Spanish/English CS practices that are similar, particularly in the nominal domain, there appears to be more variability in the verbal domain (Balam et al., 2020), which we further investigate in this study.

Participants

A total of 149 Spanish/English bilinguals took part in this study: 119 from Northern Belize and 36 from the Southwestern region of the United States (New Mexico: 24; Texas: 9; California: 2; Arizona: 1). Participants’ self-reports were used for our analyses. Recent research on language self-assessment confirms its validity as a complementary assessment approach to language tests when measuring language proficiency (Edele et al., 2015; Li & Zhang, 2021). In addition, research with heritage speakers of Spanish in the Netherlands has shown self-reports to correlate significantly with other measures of proficiency such as the Diploma Español de Lengua Extranjera (DELE) and lexical decision tasks (van Osch, 2019).

All participants declared that they learned to speak both Spanish and English/Belizean Kriol during their childhood, and that they “mix Spanish and English” at least “sometimes.” Overall, both bilingual groups reported a higher proficiency in English than in Spanish. In terms of language use, Northern Belize bi/multilinguals reported more frequent use of Spanish and Spanish/English CS, whereas Southwestern US bilinguals indicated that they more frequently speak English. These differences, however, were not marked. Table 1 shows the demographics of each group, as well as the mean results of their self-assessment in terms of language proficiency and use in English and Spanish, and frequency of use of CS in their everyday communication.

Table 1.

Participant characteristics.

	Northern Belize	United States
Number of participants	119	36
Number of male/female participants	36/83	5/31
Mean age in years (SD; range)	19.9 (5.0; 18–42)	22.9 (5.5; 18–40)
Self-assessment: Spanish proficiency^a	2.8 (SD: 0.8)	3.2 (SD: 0.8)
Self-assessment: English proficiency^a	3.1 (SD: 0.7)	3.7 (SD: 0.6)
Self-report: frequency of use of Spanish^b	3.3 (SD: 0.8)	3.3 (SD: 0.8)
Self-report: frequency of use of English^b	3.1 (SD: 0.8)	3.7 (SD: 0.6)
Self-report: frequency of use of Spanish/ English code-switching^b	3.3 (SD: 0.8)	3.1 (SD: 0.9)

On a scale of 0–4 (0 = not applicable; 1 = poor; 2 = average; 3 = good; 4 = excellent).

On a scale of 0–4 (0 = not applicable; 1 = rarely; 2 = sometimes; 3 = frequently; 4 = all the time).

Materials

We generated six code-switched base sentences: three sentences with stative passive BCVs and three sentences with eventive passive BCVs. Even though González-Vilbazo and López’s (2011, p. 836) restriction applies to eventive passives only, we investigated the acceptability of both stative and eventive passive bilingual BCVs in order to elucidate any significant cross-community similarities or differences in the acceptance of these two BCV forms.

The following participle forms were included in the six base sentences: recognized, fixed, delivered; and paved, charged, involved, respectively. These participles were selected based on exemplars from naturalistic speech data (Balam, 2015, 2016b; Balam, Prada Pérez, & Mayans, 2014) Sentences were controlled for tense, animacy of the lexical subjects, sentence length, and plausibility. Since stative passives are typically expressed in the imperfect form of estar whereas eventive passives are generally expressed in the preterite form of ser when referring to a past action (Bruhn de Garavito & Valenzuela, 2008), all stative passives were presented in the imperfect and all eventive passives in the preterite so that tense would not play a role in participants’ judgments.

In order to determine whether speakers showed sensitivity to the light verb hacer as a carrier of grammatical gender, conditions included both masculine- and corresponding feminine-marked conditions whose referents were feminine nouns. Only commonly used feminine inanimate subjects were included in the test items (e.g., la escuela “school,” la iglesia “church”). This allowed us to analyze speakers’ preferences vis-à-vis the maintenance of gender agreement between past participle form of hacer and the antecedent noun.

Note that the inclusion of masculine nouns presented a problem, as we would not be sure whether gender agreement marking on the masculine-marked light verb (i.e., hecho) was based on the employment of the masculine noun gender or the masculine default gender. By contrast, if speakers overwhelmingly rated masculine-marked bilingual BCVs as more acceptable than feminine-marked bilingual BCVs despite the feminine gender of the Spanish antecedent noun, this could be interpreted as indicative of the application of the masculine default gender in bilingual speech (for relevant discussion on the application of the masculine default gender in Spanish/English CS, see Balam, Lakshmanan, & Parafita Couto, 2021).

We generated four variants for each of the six base sentences as follows:

(15a) Ser + hacer_{Pastpart.masc} + V_Pastpart: Hector se molestó porque la escuela no fue hecho recognized.

(15b) Ser + hacer_Pastpart.fem+ V_Pastpart: Hector se molestó porque la escuela no fue hecha recognized.

(15c) Ser + V_Inf: Hector se molestó porque la escuela no fue recognize.

(15d) Ser + V_PastPart: Hector se molestó porque la escuela no fue recognized.

Note that (15a)–(15d) only exemplify eventive passive conditions, but the patterns were the same for both sets of conditions. These four conditions allowed us to examine speakers’ intuitions not only of bilingual passives, but also the status of the light verb. The exemplars in (15a) and (15b) differ in terms of gender agreement. In the case of (15a), there is no gender agreement between the light verb hecho “done” and the embedded, corresponding feminine NP la escuela “the school,” whereas in (15b), the feminine-marked light verb hecha “done” agrees in gender with the lexical subject. The exemplar in (15c) is an ungrammatical form, in which the auxiliary verb estar “to be” co-occurs with an infinitive rather than the expected past participle form. Finally, the exemplar in (15d) is representative of a stative passive BCV without the incorporation of the light verb hacer.

Procedure

Data were collected via a 2AFC task, which was administered to gain insight into speakers’ well-formedness judgments of BCVs. From previous work, we know that bilingual speakers make consistent well-formedness judgments about the acceptability of code-switched constructions based on the grammatical principles that underlie their linguistic competence (e.g., González-Vilbazo & López, 2011; Parafita Couto & Stadthagen-González, 2019; Stadthagen-González et al., 2018). Given that passives are highly infrequent structures in both monolingual Spanish and Spanish/English naturalistic code-switched discourse (Bruhn de Garavito & Valenzuela, 2008), the present method was adequate to obtain further insight into speakers’ intuitions about the use of these constructions. Furthermore, recent studies have shown compelling evidence that the 2AFC task provides a more detailed, granular insight into bilingual speakers’ intuitions in comparison to the more traditional Likert-type-scale judgment task (Stadthagen-González et al., 2018).

In the 2AFC task, exemplars of each condition were presented in permutational pairs (comparing, for example, 15a vs 15b, 15a vs 15c, 15b vs 15c) until all possible comparisons for the conditions were presented (i.e., six pairs for each code-switched base sentence), thus generating a total of 36 critical paired comparisons for the present study. In addition, participants were presented with 36 filler comparisons, which focused on a different switch type involving the complementizer que “that” (results for those comparisons will be reported elsewhere). In the 2AFC task, which was administered online using Qualtrics, speakers were asked to choose the sentence that sounded more natural and acceptable to them. A choice had to be made before progressing to the following pair of sentences. Both the order of presentation of the paired items and the individual sentences in each pair were randomized for each participant.

Research questions and hypotheses

We sought to test the following hypotheses in relation to the following three research questions.

RQ 1. What are the speakers’ judgments regarding the grammaticality of stative passive BCVs?

Hypotheses. Based on previous findings (Balam et al., 2020), we anticipated both bilingual groups to accept stative passives BCVs. Importantly, we only expected Northern Belize bilinguals to display a preference for innovative conditions with the light verb hacer given the attested predisposition toward hacer BCVs in Northern Belize.

RQ 2. What are the speakers’ judgments regarding the grammaticality of eventive passive BCVs?

Hypotheses. In line with González-Vilbazo and López (2011), we anticipated eventive passives would be either rejected or dispreferred by both bilingual groups.

RQ 3. What is the effect of gender agreement between the subject and the light verb hacer on speakers’ judgments of stative and eventive passive BCVs?

Hypotheses. We anticipated Northern Belize bilinguals to show a preference for masculine-marked conditions (despite gender disagreement), in line with the lack of gender agreement documented in oral production (Balam, 2014, p. 90) and the overwhelming employment of the masculine default gender in Northern Belize code-switched speech (Balam, 2016c). We also anticipated US bilinguals to show a preference for masculine-marked conditions, as the use of the masculine default gender in bilingual discourse has also been attested in many US Spanish/English communities (e.g., Clegg & Waltermire, 2009, and references therein).

Results

Participants’ responses to the 2AFC task were analyzed using Thurstone’s (1927) Law of Comparative Judgment, Case V, which analyzes participants’ pairwise comparison of the stimuli to generate a ranking of preference among conditions as well as a measure of relative comparison between them. These measures can be interpreted as values on an interval scale that represents a psychological continuum (in our case, the acceptability of the bilingual passive BCVs). The unit of measurement along that scale is defined as the standard deviation of the distribution (Brown & Peterson, 2009). Stadthagen-González et al. (2018) provide a detailed explanation of how to perform this type of analysis.

Table 2 summarizes the results of Thurstone’s analysis for Northern Belize and Southwest US bilinguals. The measure values are relative to the pattern with the lowest acceptability for each participant group (which is by convention set to 0). Since Thurstone’s analysis yields an interval scale for each set of participants, it is not possible to draw comparisons between groups of participants, because their scales do not share the same origin or “point 0.” We looked at the relative rank for each condition and the distance between conditions within each group of participants, but without drawing comparisons between participant groups. We calculated confidence intervals using Montag’s (2006) method (which was specifically developed for paired comparison data). The 95% confidence interval for Northern Belize data was ± 0.06, while for US data it was ±0.10.

Table 2.

Thurstone measure values for bilingual groups: stative passive bilingual compound verbs (BCVs).

Pattern	Example	Northern Belize	Southwest United States
A. Estar + hacer_{Pastpart.masc} + V_pastpart	Jessica se molestó porque la batería no estaba hecho charged	1.19	0.00
B. Estar + hacer_Pastpart.fem + V_Pastpart	Jessica se molestó porque la batería no estaba hecha charged	0.23	0.59
C. Estar + V_Inf	Jessica se molestó porque la batería no estaba charge	0.00	0.39
D. Estar + V_Pastpart	Jessica se molestó porque la batería no estaba charged	1.48	2.42

As can be seen in Table 2, Belizean bilinguals showed the highest preference for the stative passive BCV without the incorporation of the light verb hacer (Condition D), followed by the condition without gender agreement (Condition A). They showed the least preference for sentences that evinced gender agreement between the light verb and the antecedent noun (Condition B) and those where the auxiliary verb estar co-occurred with an infinitive verb (Condition C), in that order (all differences between means were outside the corresponding 95% confidence intervals). Southwest US participants preferred Condition D by a wide margin, but they also accepted to some extent Conditions B and C (with only a small difference between those two conditions and a marginal overlap in confidence intervals), which were both largely rejected by the Belize group. Finally, Southwest US participants showed the least preference for Condition A, which was a close second in the preferences of Belizeans.

With eventive passive BCVs, there were more striking differences between the two bilingual groups. As Table 3 shows, Northern Belize bilinguals gave the highest ratings to eventive passive constructions with hacer (Condition A, for example, la escuela no fue hecho recognized), which are purportedly banned in bilingual speech according to González-Vilbazo and López (2011). Eventive passive BCVs without hacer were also rated as acceptable forms (Condition D, for example, la escuela no fue recognized). By contrast, Southwest US bilinguals rated Condition A as their least preferred and rated Condition D as the most acceptable form.

Table 3.

Thurstone measure values for bilingual groups: eventive passive bilingual compound verbs (BCVs).

Pattern	Example	Northern Belize	Southwest United States
A. Ser + hacer_{Pastpart.masc} + V_pastpart	Hector se molestó porque la escuela no fue hecho recognized	1.64	0.00
B. Ser + hacer_Pastpart.fem + V_Pastpart	Hector se molestó porque la escuela no fue hecha recognized	0.83	0.75
C. Ser + V_Inf	Hector se molestó porque la escuela no fue recognize	0.00	0.56
D. Ser + V_Pastpart	Hector se molestó porque la escuela no fue recognized	1.52	2.87

It is important to note that Northern Belize bilinguals rated eventive passive BCVs without hacer (Condition D) as more acceptable than passive constructions in which there was gender agreement between the light verb hacer and the antecedent noun (Condition B). Passive BCVs, in which the auxiliary verb co-occurred with an infinitive verb, were least preferred among Northern Belize bilinguals, whereas this form displayed more acceptability among US bilinguals. Finally, US participants had very similar levels of acceptance for Conditions B and C (with marginal overlap of confidence intervals).

Discussion and conclusion

We investigated the grammaticality of stative and eventive passive BCVs and the status of grammatical gender in these constructions among Spanish/English bilinguals from two contexts where hacer BCVs have been attested. In relation to our first research question, results revealed that both bilingual groups showed a preference for stative passive BCVs without the light verb hacer. This suggests that although constructions with hacer are preferred in certain verbal contexts such as bilingual progressive constructions (Balam et al., 2020), this is not the case with passive BCVs, where Northern Belize bilinguals actually align with their US counterparts in their preference for stative passives without hacer (e.g.,. . .la batería estaba charged “. . .the battery was charged”). Results also revealed variability, however, as the two bilingual groups differed in their preference for stative passive BCVs with hacer: whereas Northern Belize bilinguals rated masculine-marked conditions as acceptable, Southwest US bilinguals rejected them.

With regard to our second research question, we found even more variability between the two bilingual groups. Data from Southwest US Spanish/English bilinguals partially lend support to González-Vilbazo and López’s (2011) contention that eventive passives are incompatible with the light verb hacer, as they rejected these forms, particularly in cases with gender disagreement. On the other hand, bilinguals from Northern Belize showed a preference for eventive passive BCVs that lacked gender agreement between the light verb hacer and the antecedent feminine noun (e.g.,. . .la escuela no fue hecho recognized “. . .the school was not recognized”). These results show, therefore, that in Northern Belize, passivization is actually compatible with the light verb hacer, and this compatibility is more than likely not an idiosyncratic phenomenon restricted only to a sub-group of Spanish/English code-switchers.

In terms of our third research question, results also differed according to the bilingual group. Among Southwest US bilinguals, the least preferred conditions for stative and eventive passives were cases in which there was lack of gender agreement. There was more acceptance of feminine-marked BCVs, (e.g.,. . .la escuela no fue hecha recognized “. . .the school was not recognized”), which suggests speakers were aware that these BCVs are grammatical within a more normative verbal paradigm where gender agreement is maintained between syntactically related constituents (i.e., the light verb and the antecedent noun). By contrast, among Northern Belize bilinguals, masculine-marked BCVs that lacked gender agreement were deemed acceptable, and they were preferred in eventive passive BCVs.

Our results revealed that both stative and eventive passives are compatible with the light verb hacer in Spanish/English CS, but compatibility is inherently tied to the group of code-switchers that is examined. In light of these findings, we posit then that the availability of passivization with the light verb hacer is largely determined by community linguistic norms and social factors rather than a universal syntactic restriction, as proposed by González-Vilbazo and López (2011). Hebblethwaite (2007) aptly highlights that CS is not derived from formal aspects of grammar alone, and consequently, our theorizing of CS “cannot be divorced from the influence of social structure” (p. 262). Our findings highlight the importance of testing syntactic hypotheses in understudied CS contexts where social conditions have been propitious for the emergence of structural innovations.

In the case of Belize, sociohistorical conditions that were conducive to prolific CS practices paved the way for innovative passive constructions to emerge among younger generations (Balam, 2015, 2016a). Crucially, this development has occurred in tandem with broader linguistic trends attested in oral production, namely, (1) the preferential use of hacer BCVs (Balam et al., 2020); and (2) the overwhelming employment of the masculine default gender in code-switched discourse (Balam, 2016c). We know that passive structures have historically been susceptible to an ongoing process of diachronic change and re-structuring (Pountain, 1994, p. 111). Stative and eventive passive BCVs in Northern Belize suggest that, despite their infrequent use in both monolingual and contact Spanish, passives are also vulnerable to syntactic innovation.

In Northern Belize, conditions such as positive attitudes to CS and low levels of linguistic prescriptivism (“relaxed norms” in the sense of Muysken, 2013, p. 174) were instrumental in fostering the ideal environment for hacer BCVs to thrive and further grammaticalize (for relevant discussion, see Balam, 2016b). Although we agree that there are syntactic constraints in CS, it is also important to remember that bi/multilingual speakers are able to “avoid the straitjacket of grammatical rules,” (Gardner-Chloros & Edwards, 2004, p. 108). The marked preference for eventive passive BCVs among Northern Belize code-switchers may appear to violate (morpho)syntactic restrictions. This is a clear example, however, “[of] rule-breaking behavior, which should be seen not in relation to static norms but in terms of language change and convergence” (Gardner-Chloros, 2010, p. 202).

Note that the more preferential ratings given to feminine-marked bilingual passive forms and the rejection of masculine-marked BCVs among Southwest US Spanish/English bilinguals is not unsurprising. This could be partly attributed to the ongoing process of standardization taking place in this region, where increasingly intense contact with monolingual varieties of Spanish is leading to the adoption of more standard Mexican Spanish patterns and forms (for relevant discussion on this phenomenon in New Mexico, for instance, see Waltermire, 2017). It is not surprising, therefore, that speakers from this context give more preferential ratings to passive BCVs that show adherence to gender agreement. This starkly contrasts Northern Belize, where gender disagreement is the linguistic norm across age-groups and where even overt linguistic resistance against the use of “standard” linguistic forms has been attested among younger generations (see Balam & Prada Pérez, 2017). Thus, our results suggest that marked differences vis-à-vis sociolinguistic conditions, degree of prescriptivism, and context-specific linguistic norms can be influential in determining notable differences in morphosyntactic aspects of CS grammars across communities.

Although a syntactic analysis is beyond the purview of this paper, it is worth mentioning that the grammaticality of eventive passive hacer BCVs is in line with non-lexicalist exoskeletal models of CS that have been more recently proposed (e.g., Grimstad et al., 2018; López, 2020). In light of the fact that inflectional elements can be generated independently from verb roots, these models are able to account for code-switched verb structures in which tense and aspectual information can be generated independently from the verb stem, as in the case of eventive passive hacer BCVs which are preferred in Belize. These models are also able to account for syntactic variation in CS outcomes across communities (see, for example, López’s (2020) analysis of masculine- vs feminine-marked determiner phrases in Spanish/English CS).

Overall, our results highlight the crucial role that social factors and community-specific oral production patterns have on code-switchers’ linguistic competence. Patterns in the use of BCVs are intricately connected not only to language-internal factors but to context-specific linguistic norms and the emergence and propagation of novel morphosyntactic forms that may take place only in certain Spanish/English communities. Our findings suggest that González-Vilbazo and López’s (2011) analysis fails to account for these new data and for the structural innovation and variability that we find in the use of BCVs across Spanish/English CS communities (Toribio, 2017). The adoption of a usage-based perspective, in which corpus data are complemented with data from experimental tasks, offers a more promising approach that can better elucidate our understanding of synchronic and diachronic aspects of CS grammars (for further discussion, see Backus, 2015), both within and across bi/multilingual communities.

Footnotes

Acknowledgements

We are thankful to the editor and anonymous reviewers for their insightful comments and suggestions in earlier versions of this paper. We are also very grateful to our colleagues at Muffles Junior College (Arturo Acosta), Corozal Junior College (Yvette Sanker Riverol), and the University of Belize (Lugie Cruz) for helping to facilitate data collection in Belize.

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

Osmer Balam

Notes

Author biographies

Osmer Balam (PhD, University of Florida) is currently an Assistant Professor of Spanish at the College of Wooster in Wooster, Ohio. His scholarly work focuses on sociolinguistics, language contact outcomes, and morphosyntactic variation in code-switching grammars.

Hans Stadthagen-González (PhD, University of Bristol) is currently an Associate Professor of Psychology at the University of Southern Mississippi. The main focus of his research is on the psychology of language, the study of bilingualism and cognition, emotional aspects of language, and language assessment.

Eva Rodríguez-González (PhD, University of Kansas) is currently an Associate Professor of Spanish at the University of New Mexico. Her main areas of research include classroom-based language learning and cognitive approaches to language acquisition. She is particularly interested in foreign language proficiency assessment and program evaluation.

María del Carmen Parafita Couto (PhD, University of Kansas) is a María Zambrano Senior Researcher at the University of Vigo (Spain) and University Lecturer at the Leiden University Center for Linguistics (the Netherlands). She studies the impact of bilingualism and language contact on language structure, particularly the syntax of code-switching.

References

Annamalai

(1989). The language factor in codemixing. International Journal of the Sociology of Language, 75, 47–54.

Backus

(1996). Two in one. Bilingual speech of Turkish immigrants in the Netherlands. Tilburg University Press.

Backus

(2015). A usage-based approach to code-switching: The need for reconciling structure and function. In Stell

Yakpo

(Eds.), Code-switching between structural and sociolinguistic perspectives (pp. 19–37). De Gruyter Mouton.

Balam

(2014). Notes on the history and morphosyntactic characteristics of Spanish in Northern Belize. Kansas Working Papers in Linguistics, 35, 79–94.

Balam

(2015). Codeswitching and linguistic evolution: The case of “Hacer + V” in Orange Walk, Northern Belize. Lengua Y Migración, 7(1), 83–109.

Balam

(2016a). Language use, language change and innovation in Northern Belize contact Spanish [Doctoral dissertation, University of Florida].

Balam

(2016b). Mixed verbs in contact Spanish: Patterns of use among emergent and dynamic bi/multilinguals. Languages, 1(3), 1–21. https://doi.org/10.3390/languages1010003

Balam

(2016c). Semantic categories and gender assignment in contact Spanish: Type of code-switching and its relevance to linguistic outcomes. Journal of Language Contact, 9(3), 1–31. https://doi.org/10.1163/19552629-00903001

Balam

Lakshmanan

Parafita Couto

M. C.

(2021). Gender assignment strategies among Spanish/English simultaneous bilingual children. Studies in Hispanic and Lusophone Linguistics, 14(2), 241–280.

10.

Balam

Parafita Couto

M. C.

Chen

M. A.

(2021). Being in bilingual speech: An analysis of Estar “be” constructions in Spanish/English code-switching. Journal of Monolingual and Bilingual Speech, 3(2), 238–264.

11.

Balam

Parafita Couto

M. C.

Stadthagen-González

(2020). Bilingual verbs in Spanish/English Code-switching. International Journal of Bilingualism, 24(5–6), 952–967.

12.

Balam

Prada Pérez

(2013, October). Light verb constructions in stative and eventive passives in contact Spanish [Paper presentation]. Hispanic Linguistics Symposium, Ottawa, Ontario, Canada.

13.

Balam

Prada Pérez

(2016). On the Productive Use of “Hacer + V” in Northern Belize bilingual/trilingual codeswitching. In Guzzardo Tamargo

R. E.

Mazak

Parafita Couto

M. C.

(Eds.), Spanish-English codeswitching in the Caribbean and the U.S. (pp. 261–279). John Benjamins.

14.

Balam

Prada Pérez

(2017). Attitudes toward Spanish and code-switching in Belize: Stigmatization and innovation in the Spanish classroom. Journal of Language, Identity and Education, 16(1), 17–31. https://doi.org/10.1080/15348458.2016.1260455

15.

Balam

Prada Pérez

Mayans

(2014). A congruence approach to the study of Bilingual Compound Verbs in Northern Belize Contact Spanish. Spanish in Context, 11(2), 243–265. https://doi.org/10.1075/sic.11.2.05bal

16.

Brown

T. C.

Peterson

G. L.

(2009). An enquiry into the method of paired comparison: Reliability, scaling, and Thurstone’s law of comparative judgment. Rocky Mountain Research Station, US Department of Agriculture, Forest Service.

17.

Bruhn de Garavito

Valenzuela

(2008). Eventive and stative passives in Spanish L2 acquisition: A matter of aspect. Bilingualism: Language and Cognition, 11(3), 323–336. https://doi.org/10.1017/S1366728908003556

18.

Chatterjee

(2016). Structural changes in Bengali-English bilingual verbs through the exploration of Bengali films. Languages, 1(1), 5. https://doi.org/10.3390/languages1010005

19.

Clegg

J. H.

Waltermire

(2009). Gender assignment to English-origin nouns in the Spanish of the Southwestern United States. Southwest Journal of Linguistics, 28(1), 1–17.

20.

Edele

Seuring

Kristen

Stanat

(2015). Why bother with testing? The validity of immigrants’ self-assessed language proficiency. Social Science Research, 52, 99–123.

21.

Gardner-Chloros

(2010). Contact and code-switching. In Hickey

(Ed.), The handbook of language contact (pp. 188–207). Wiley-Blackwell.

22.

Gardner-Chloros

Edwards

M. H.

(2004). Assumption behind grammatical approaches to code-switching: When the blueprint is a red herring. Transactions of the Philological Society, 102(1), 103–129.

23.

González-Vilbazo

López

(2011). Some properties of light verbs in codeswitching. Lingua, 121(5), 832–850. https://doi.org/j.lingua.2010.11.011

24.

Grimstad

M. B.

Riksem

B. R.

Lohndal

Åfarli

T. A.

(2018). Lexicalist vs exoskeletal approaches to language mixing. The Linguistic Review, 35(2), 187–218.

25.

Guzzardo Tamargo

. (2012). Linking comprehension costs to production patterns during the processing of mixed language [Unpublished doctoral dissertation, Pennsylvania State University].

26.

Halberstadt

(2017). Investigating community norms and linguistic mechanisms in codeswitching: Bridging linguistic theory and psycholinguistic experimentation [Unpublished doctoral dissertation, Pennsylvania State University].

27.

Hebblethwaite

(2007). Intrasentential code-switching among Miami Haitian Creole–English bilinguals [Unpublished doctoral dissertation, Indiana University].

28.

Kachru

B. B.

(1978). Toward structuring code-mixing: An Indian perspective. International Journal of the Sociology of Language, 16, 24–47.

29.

Levin

Rappaport

(1986). The formation of adjectival passives. Linguistic Inquiry, 17(4), 623–661.

30.

Zhang

(2021). A meta-analysis of self-assessment and language performance in language testing and assessment. Language Testing, 38(2), 189–218.

31.

Lipski

J. M.

(2008). Varieties of Spanish in the United States. Georgetown University Press.

32.

López

(2020). Bilingual grammar: Toward an integrated model. Cambridge University Press. https://doi.org/10.1017/9781108756181

33.

López

Alexiadou

Veenstra

(2017). Codeswitching by phase. Languages, 2, 9. https://doi.org/10.3390/languages2030009

34.

Luján

M. E.

(1981). The Spanish copulas as aspectual indicators. Lingua, 54(2–3), 165–210. https://doi.org/10.1016/0024-3841(81)90068-1

35.

Mendikoetxea

(2012). Passives and se constructions. In Hualde

J. I.

Olarrea

O’Rourke

(Eds.), The handbook of Hispanic linguistics (pp. 477–502). John Wiley & Sons.

36.

Montag

E. D.

(2006). Empirical formula for creating error bars for the method of paired comparisons. Journal of Electronic Imaging, 15(1), 222–230.

37.

Moyer

M. G.

(1992). Analysis of code-switching in Gibraltar [Unpublished doctoral thesis, Universitat Autònoma de Barcelona].

38.

Muysken

(2000). Bilingual speech: A typology of code-mixing. Cambridge University Press.

39.

Muysken

(2013). Language contact outcomes as the result of bilingual optimization strategies. Bilingualism: Language and Cognition, 16(4), 709–730.

40.

Parafita Couto

M. C.

Stadthagen-González

. (2019). El book or the libro? Insights from acceptability judgments into determiner/noun code-switches. International Journal of Bilingualism, 23(1), 349–360. https://doi.org/10.1177/1367006917728392

41.

Pfaff

C. W.

(1979). Constraints on language mixing: Intrasentential code-switching and borrowing in Spanish/English. Language, 55(2), 291–318. https://doi.org/10.2307/412586

42.

Pountain

C. J.

(1994). Syntactic anglicisms in Spanish: Exploitation or innovation. In Mair Parry

Davies-Winifred

(Eds.), The changing voices of Europe. Social and political changes and their linguistic repercussions. Past, present and future (pp. 109–124). University of Wales Press.

43.

Reyes

(1982). Language Mixing in Chicano Spanish. In Amstae

Elías-Olivares

(Eds.), Spanish in the United States: Sociolinguistics aspects (pp. 154–165). Cambridge University Press.

44.

Romaine

(1986). The syntax and semantics of the code-mixed compound verb in Panjabi/English bilingual discourse. In Alatis

D. T.

(Ed.), Language and Linguistics. The interdependence of theory, data and application (pp. 35–49). Georgetown University Press.

45.

Stadthagen-González

López

Parafita Couto

M. C.

Párraga

C. A.

(2018). Using two-alternative forced choice tasks and Thurstone’s law of comparative judgments for code-switching research. Linguistic Approaches to Bilingualism, 8(1), 67–97. https://doi.org/10.1075/lab.16030.sta

46.

Thurstone

(1927). A law of comparative judgment. Psychological Review, 34(4), 273–286.

47.

Toribio

A. J.

(2017). Structural approaches to code-switching: Research then and now. In Lopez

R. E. V.

Ornelas

Avelar

Cyrino

S. M. L.

(Eds.), Romance languages and linguistic theory 12: Selected papers from the 45th linguistics symposium on romance languages (LSRL), Campinas, Brazil (pp. 213–233). John Benjamins.

48.

van Osch

. (2019). Vulnerability in heritage speakers of Spanish in The Netherlands: An interplay between language-internal and language/external factors [Unpublished doctoral dissertation, University of Amsterdam].

49.

Vergara Wilson

. (2013). One construction, two source languages: Hacer with an English infinitive in bilingual discourse. In Carvalho

A. M.

Beaudrie

(Eds.), Selected proceedings of the 6th workshop on Spanish sociolinguistics (pp. 123–134). Cascadilla Proceedings Project.

50.

Waltermire

(2017). At the dialectal crossroads: The Spanish of Albuquerque, New Mexico. Dialectología, 19, 177–197.

51.

Wasow

(1977). Transformations and the Lexicon. In Culicover

Wasow

Bresnan

(Eds.), Formal syntax (pp. 327–360). Academic Press.