I think,therefore digo yo : Variable position of the 1sg subject pronoun in New Mexican Spanish-English code-switching

Abstract

Using the New Mexico Spanish-English Bilingual Corpus, the present paper examines the variable position of the 1sg Spanish subject pronoun yo—pre- versus post-verbal—to consider the effect that code-switching may have on structural change. In an analysis of close to 700 tokens of yo, a rate of 16% post-positioning is found, which is within the range of post-position in non-contact varieties and thus contraindicative of the convergence hypothesis, in accordance with which the almost exclusive use of preverbal subject pronouns in English would predict lower rates of post-verbal yo in a converged contact variety. Moreover, by testing factors hypothesized to account for choice of post-posing yo using multivariate analysis, it is shown that bilinguals display similar constraints on yo post-positioning in New Mexican Spanish as monolingual speakers of Spanish, providing stronger support for an anti-convergence account. Results are discussed in terms of bilingual parallel activation, syntactic priming, and construction grammar.

Keywords

Code-switching syntactic priming construction grammar subject pronouns Spanish

Introduction

Linguistic convergence is a type of contact-induced change whereby two languages become more structurally similar over time through the adoption of features from one language (usually the socially dominant language) into the other (usually the socially subordinate language). Code-switchers are often pointed to in the literature as the cause of such convergence between languages in contact, but is it that simple? A growing body of psycholinguistic processing and production research calls for us to test this assumption. Such studies increasingly indicate that the bilingual brain is quite capable of switching between and inhibiting one or the other language, which makes convergence of grammars in bilinguals anything but a foregone conclusion. While these studies in the psycholinguistic laboratory can inform the work done in other fields of linguistics, experimental paradigms do have limitations that can be overcome with access to corpora of bilingual speech, such as the New Mexico Spanish-English Bilingual (NMSEB) corpus (Torres Cacoullos & Travis, in preparation). This corpus represents the language usage patterns of early bilinguals in a long-standing community that has maintained the Spanish language for many generations, a special kind of speech community among varieties of Spanish in the US, the likes of which have generally not been studied in previous tests of the convergence hypothesis. Taking our data from the NMSEB corpus, the present paper uses the variable placement of the Spanish 1sg subject pronoun yo to shed further light on the debate about the role of code-switching in convergence.

While in English a subject pronoun generally occurs before the verb in any declarative utterance, Spanish displays more flexibility, allowing for subjects to be left unexpressed, as in (1), or expressed in the preverbal (2) or post-verbal (3) position. This difference in the typical structures in English and Spanish provides us with a rich but simple platform on which to test whether the presence of English in the code-switching of New Mexican (NM) bilinguals is leading to the convergence of NM Spanish grammar towards more English-like structures and constraints.

(1) Null subject:

Francisco (H) y allá en los montes en ese lado ‘and there on that side of the mountain I-NULL

Ø miré la lumbre que iba brincando. looked at the light dancing.’

[18 Las minas, 1:09:56–1:10:00]¹

(2) Pre-posed yo:

Rocío yo no podía salir. ‘I-PRE couldn’t go out.’

[18 Las minas, 1:05:04–1:05:05]

(3) Post-posed yo:

Sandra y le sacaba yo el papelito, ‘and I-POST removed the paper for him,’

[03 Dos comadres, 0:25:27–0:25:29]

The paper proceeds as follows. We start by reviewing the existing literature on convergence and code-switching. Then, we introduce the corpus and variable context in greater detail and explain the coding procedures followed and the hypotheses behind them, from which we explore both overall rates of use (in comparison with other varieties) and independent factors co-occurring with a post- versus preverbal 1sg pronoun—including recent use of English—in order to understand the choice to post-pose yo. The analyses presented demonstrate that code-switchers are skilled separators of their two grammars (cf. Poplack, 1980), rather than conflaters of them, as has been proposed (e.g., Backus, 2005).

The question of convergence in bilinguals and bilingual speech communities

It is frequently assumed that contact between two languages results in convergence between the two grammars and specifically that code-switching gives impetus to that change (Backus, 2005; Gumperz & Wilson, 1971, p. 165; Thomason, 2000, p. 311; Thomason & Kauffman, 1988, p. 68; Winford, 2008, p. 135). Considering prior studies of convergence in varieties of Spanish spoken in the US, Risso (2010, pp. 109–110), based on overall rates of use, proposed that subject pronoun placement in Spanish spoken in New York does represent change via contact both with English and across dialects, but that stronger than these contact effects was the region of origin of the speaker (2010, p. 111).

Montrul’s (2004) study, based on an elicited production task where participants were asked to retell the story of Little Red Riding Hood using pictures from a children’s book, demonstrated higher rates of VS order with unaccusative verbs; she found similar rates of post-verbal subject placement across monolingual and bilingual (Mexican-American “heritage” speaker) groups of Spanish speakers. She nevertheless concludes that convergence towards English has taken place in the sense that the “heritage speakers” of her study have lost lexical-semantic and discourse-pragmatic constraints (cf. Silva-Corvalán, 1994, p. 166; Sorace, 2004).

This and many other studies claiming convergence study speakers of middling proficiency, not highly proficient bilinguals, and analyze highly contrived production contexts that are unlikely to elicit natural speech (e.g., Toribio, 2004). In her study of Mexican-American speakers of Spanish across three contact generations in Los Angeles, Silva-Corvalán (1994, p. 16) points out that, generally speaking, within two generations of arrival to the US descendants of Spanish-speaking immigrants no longer speak Spanish fluently in their homes and communities. As Lipski (1996, p. 148) notes, the very existence of middle-range proficiency heritage speakers represents a loss over time of Spanish in these individuals, their families, and potentially their communities. While interesting for discussions of language acquisition and attrition on the individual level, it is, in our view, inapt to consider their speech behaviors as contributors to, or indicators of, convergence of grammars on the community scale because this speech does not persist in Spanish-speaking communities in the US.

For these reasons we use spontaneously produced code-switching data from highly proficient early bilinguals from a community that has maintained Spanish for many more generations than the populations studied in other regions. We use this data to challenge the proposal that code-switching leads to convergence.

Community-based studies of variation and change have offered convincing evidence to counter the convergence hypothesis. Despite a surface pattern that appears to be converging towards the grammar of one language in a contact situation, several studies find that some apparent changes are not actually a result of contact, but rather of language internal factors, or are actually not changes at all. For example, Poplack, Zentz and Dion (2012) compared the patterns of usage of Quebecois French preposition stranding in frequent code-switchers to those of more sparse code-switchers and found no differences, further refuting the idea that frequent code-switchers are the agents of convergence. Likewise, Torres Cacoullos and Travis (2011, 2015b) have found no effect of the presence of code-switching on patterns of New Mexico Spanish subject expression among bilinguals. To evaluate convergence in the present study we must ask not only how often Spanish speakers place yo after the verb, but also in what contexts does this occur.

Corpus and variable context

The data on which this study is based were extracted from the NMSEB corpus (Torres Cacoullos & Travis, in preparation). These data consist of sociolinguistic interviews conducted in northern New Mexico by bilingual members of the community (see Travis & Torres Cacoullos, 2013, for detailed description of the corpus). People have been speaking Spanish in this region since the Spanish settled there by way of present-day Mexico in the 17th century. This means that, crucially, these speakers are neither immigrants nor the children of immigrants and they represent a population of bilinguals whose Spanish variety is well-established in the community. For this study, we used 20 interviews from the corpus with 23 participants, comprising approximately 18 hours of spontaneous, conversational speech, or 60,000 Intonation Units (IUs) and 202,000 words.

All tokens of Spanish conjugated first singular verbs that appeared with an expressed Spanish subject yo were extracted from the corpus (N = 669). Tokens of yo appearing on a separate IU than the main verb were excluded (see Torres Cacoullos & Travis, 2015b). Tokens occurring in interrogative clauses were also excluded, given that the patterns of subject position are under different constraints in questions. It should be noted that no tokens of expressed yo as the subject of an English verb were found. Furthermore, in a sample of 100 tokens of I + English verb declaratives, 1sg I categorically appeared before the verb, indicating that these speakers’ English matches other varieties in the eschewing of post-verbal subjects.

Although rarely studied separately from other grammatical persons in past work, here we choose to focus only on the 1sg subject pronoun, because all persons do not appear to be conditioned by the same constraints. For example, in their study of subject position in wh-questions in Puerto Rican Spanish, Brown and Rivas (2011, pp. 37–38) found that first- and second-person pronouns favor, and third-person disfavors, preverbal position in this context. Silva-Corvalán (2001, p. 165) finds higher rates of post-verbal non-specific subject referents, such as tú or uno, than specific subject pronouns (due, she proposes, to the defocusing function of the post-verbal pronoun position).

While there is a large body of existing literature on subject pronoun expression in Spanish (see references in e.g., Silva-Corvalán, 2001, pp. 154–169), there is somewhat less about subject pronoun position. Tendencies that have been reported in the little work that has been done include an increased likelihood of a post-verbal subject pronoun in the presence of preverbal adverbials (Ocampo, 1995, p. 155; Silva-Corvalán, 1982), with copulas (Bentivoglio, 1988, p. 19; Brown & Rivas, 2011, p. 39) and with new information (Givón, 1983, p. 19; Herring, 1990, p. 169). From a pragmatic perspective, it has been proposed that post-verbal position is associated with a “focus of attention” with 1sg and 2sg pronouns (Posio, 2011, p. 797), as well as with the second-person pronoun usted(es) ‘you-formal’ (Serrano, 2012, p. 116).

Coding and hypotheses

Each token was coded for the position of the pronoun in relation to the verb, either pre- or post-verbal. Of the 669 tokens coded for the present study, 106 appeared following the verb, resulting in a 16% post-verbal rate (see Table 1). Some of the most outspoken proponents of the convergence hypothesis with respect to subject expression have drawn conclusions based solely on differences in rates amongst monolingual and bilingual speakers (e.g., Toribio, 2004, p. 172). A comparison of rates of post-verbal yo in NM Spanish with those of non-contact varieties outside the United States is given in Figure 1, which demonstrates that the rate of post-verbal yo found among NM bilinguals in the present study is slightly higher than the range of rates of post-verbal yo observed in the non-contact varieties, from as low as 5% in Cali, Colombia (Travis & Torres Cacoullos, 2012, p. 713), to 14% in Peninsular Spanish (the Corpus de Referencia de la Lengua Española Contemporánea (CORLEC) (139/999) (Smith, 2011, cf. Posio 2011, p. 778, n.1, who reports a rate of circa 12%)). A higher rate of post-posed yo in NMSEB seems contraindicative of convergence towards English syntactic patterns in this code-switching variety.

Table 1.

Distribution of post-verbal yo (N = 669).

Dependent variable	Number of the tokens	% of data
Post-verbal yo	106	16
Preverbal yo	563	84

Figure 1.

Rates of post-verbal yo in speech corpora: New Mexico Spanish-English Bilingual corpus (NMSEB), Peninsular Spanish (Corpus de Referencia de la Lengua Española Contemporánea (CORLEC)) (Smith, 2011), Corpus of Conversational Colombian Spanish (cf. Travis, 2007; Travis & Torres Cacoullos, 2012), New Mexico-Colorado Spanish Survey (NMCOSS) (cf. Bills & Vigil, 2008; Torres Cacoullos & Travis, 2013), Mexico City (Lastra & Butragueño, 2015), and Miller and Schmitt Chilean Conversation Corpus (2012) (Champi & Perrotti, 2013).

However, the rate of post-verbal yo is lower in these data than in a near-monolingual Spanish sample from the New Mexico-Colorado Spanish Survey (NMCOSS) (cf. Bills & Vigil, 2008) largely devoid of code-switching, unlike the corpus used in the current study: in the speech of 11 Spanish-dominant older speakers from NMCOSS, born 1897–1918, who have minimal use of English of any kind in their recordings, the post-verbal rate was 21% (N = 95/448) (Torres Cacoullos & Travis, 2013). As these NMCOSS data represent an earlier stage of NM Spanish, the difference in post-positioning rate between NMSEB and NMCOSS raises the question of whether the NMSEB speakers have lost the lexical-semantic and discourse-pragmatic constraints used by their less bilingual forebears, a hypothesis that requires empirical testing.

It is clear, then, that rates of post-verbal positioning of yo alone are not sufficient to interpret the variability observed in this structure (cf. Labov, 1969; Torres Cacoullos & Travis, 2015a, 2015b); we must also look at what factors shape patterns of positioning. We conduct multivariate analysis to analyze the factors that may contribute to the choice to pre- or post-pose the 1sg subject pronoun. The factor groups (predictors) and the hypothesis each one operationalizes are outlined below.

Presence of English in the surrounding discourse

Given results in the psycholinguistic research on parallel activation (e.g., Schwartz & Kroll, 2006; Schwartz, Kroll, & Diaz, 2007), we coded for the presence of English in the same or immediately preceding clause as the target verb. The motivation for this factor group was to test the effect—if any—that the presence of English-language material would have on subject position. We focused our attention on the same or immediately preceding clause because we sought to test the effect of maximally proximate code-switching (cf. Torres Cacoullos & Travis, 2015b). Examples (4)–(6) outline the coding used for this factor group. We distinguished between single-word English-origin items (4) (N = 74) or multi-word switches (5) (N = 157) uttered by the same speaker. Tokens that had no English in the same or immediately preceding clause (6) (N = 341) were given a separate code.²

(4) Single-word English item in same clause:

Ivette yo estaba bien happy. ‘I-PRE was very happy.’

[06 El túnico, 0:13:58–0:13:59]

(5) Multi-word switch in same clause:

Ivette .. yo perdí a lot of that. ‘I-PRE missed a lot of that .’

[06 El túnico, 0:37:58–0:37:59]

(6) No English in the same or immediately preceding clause:

Francisco … yo oía más antes de las brujas esas y todo. ‘I-PRE had heard about those witches before and everything.

yo no creía hasta entonces. I-PRE didn’t believe in them until then.’

[18 Las minas, 1:11:27–1:11:32]

The convergence-via-code-switching hypothesis would predict an increase in the use of more English-like syntax (i.e., preverbal yo) in the presence of English because of the stronger activation of English at that moment. However, if we take the point of view that code-switchers are skilled separators of their two grammars, we expect that the presence of English in the preceding discourse will not matter. We made the distinction between single versus multi-word code-switches to allow for the potentially different syntactic implications of a single word, which can easily be inserted in the structure of either language, as compared to a multi-word switch, which necessarily invokes English structural patterns (cf. Poplack, 1978; Sankoff, 1980; Sankoff & Poplack, 1981, for discussion of the Equivalence Constraint). It remains to be established for this community whether single words are (nonce) borrowings, or whether some are truly code switches (cf. Aaron, 2015; Poplack, 2012; Poplack, Sankoff, & Miller, 1988; Sankoff, Poplack, & Vanniarajan, 1990). Thus, in order to be conservative in our analysis, we included in the multivariate analysis only those tokens with unambiguous (i.e., multi-word) switches in the same clause as, or immediately preceding clause to, the target token, and those tokens with no English.

Verb class

The literature on verb class effects has suggested that VS order is favored with verbs whose subject is not the semantic agent (Mayoral Hernández, 2006, p. 227; Rivas, 2008, p. 909). The monolingual Spanish speakers in Montrul’s (2004) study demonstrated higher rates of VS order with unaccusative verbs, that is, intransitive verbs whose syntactic subject is not a thematic agent, such as venir ‘to come’ and irse ‘to go, leave’. In a study of Spanish word order for 3sg human specific subjects in the same NMSEB corpus, Houle (2012) found unaccusative verbs to strongly favor VS order for full NPs and proper names, but she failed to find an effect for 3sg pronouns. To test whether the effects of verb class can be observed for the 1sg subject pronoun, we coded verbs into one of five classes, following those of Houle (2012): unaccusative (7), intransitive, transitive, copula, and quotative. If unaccusative verbs favor post-positioning, as has been proposed for non-contact varieties of Spanish with respect to subject–verb order for full NP subjects, this could be evidence against “loss” of this lexical-semantic constraint. However, the absence of an effect would not necessarily be evidence for such “loss”, as the applicability of this constraint to (1sg) pronoun subjects remains to be determined.

(7) Verb class, unaccusative verb

Sandra …(1.1) pues aquí venía yo , ‘well I-POST came out here,

[03 Dos comadres, 1:05:10–1:05:13]

Referential distance

In accordance with the widely accepted observation that the form of referring expressions depends on the assumed cognitive status of the referent (Gundel, Hedberg, & Zacharski, 1993) and that the proposed function of post-positioning of the subject is to introduce new information (Givón, 1983, p. 19), we coded for the number of intervening clauses between token and previous mention as subject, up to five clauses (see Travis & Torres Cacoullos, 2012, p. 728, for subject continuity effects on yo expression). With this factor group we are interested in investigating whether the speaker’s choice to post-pose yo, in Chafe’s words, “bring[s] the idea of themselves back into the active consciousness of the listeners” (1994, p. 87). We hypothesize that, if such a function applies to post-positioning of the 1sg pronoun, it would be favored when the previous mention was farther away from the target token.

Example (6) above gives an illustration of coreferential contexts where there are no intervening clauses; below, example (10) illustrates where there is one, (8) two, and (9) three intervening clauses.

Structural priming

In language production, priming can be defined as the tendency to repeat a structure similar to one recently used or presented in the preceding discourse. To test for positional priming effects of yo-to-yo priming, as well as cross-language priming effects of I-to-yo priming, we coded for the position of the previous coreferential 1sg pronoun uttered by the speaker. In 30% of the data, the previous mention of the previous coreferential 1sg was within one clause, in an additional 15% it was within two clauses, and for 55% of the data the previous mention occurred at a distance of three or more clauses. Tokens were coded as follows: preverbal I (8), preverbal yo (9), or post-verbal yo³ (10). Priming is often observed in corpus studies of pronoun usage, where it is included as a variable (cf. Cameron & Flores Ferrán, 2004; Travis, 2007; but see Brown & Rivas, 2011 for a counter-example). We therefore hypothesize that a previous post-verbal yo will favor and a previous preverbal yo will disfavor a post-verbal yo. This factor group also enables us to test whether I, which is almost always preverbal, also primes a preverbal yo (as Torres Cacoullos & Travis, 2011, p. 258, found for cross-language priming of 1sg subject pronoun expression).

(8) Previous preverbal I:

Sandra I don’t know how to read. ‘I don’t know how to read

… y nos habían llegado unos libritos que … and some little books had arrived that

se llamaban, were called,

Fun with Dick and Jane. Fun with Dick and Jane.

…(0.9) pues ya yo había aprendido to …(0.9) well I-PRE had already learned to

picture read. picture read.’

[03 Dos comadres, 0:59:27–0:59:35]

(9) Previous preverbal yo:

Fabiola yo [creí que la N]ancy no había venido. ‘I-PRE thought that Nancy hadn’t come.’

Jake [no=]. ‘no=.

… they have four. … they have four.’

Fabiola … vino también la Nancy? ‘Did Nancy come too?’

…(1.3) yo creía que no más tres muchitas. ‘I-PRE thought just three girls ((came)).’

[09 La salvia, 0:43:10–0:43:17]

(10) Previous post-verbal yo:

Fabiola es lo que le dije yo=. ‘that is what I-POST said to him.

… no tengas mucho apuro le dije yo . … don’t trouble yourself I-POST said to him.’

[09 La salvia, 0:48:30–0:48:33]

Other structural elements in the clause

It has been observed for both contact and non-contact varieties that preverbal structural elements in the clause favor post-verbal subjects (e.g., Houle, 2012; Ocampo, 1995, p. 155; Silva-Corvalán, 1982). In order to test for this in NMSEB, tokens were coded for whether there was a complement (direct or indirect object) or an adjunct (a prepositional or adverbial phrase) before the verb (11) or after the verb (12), or whether there was no such element (negation and discourse markers, e.g., pues, were not considered to be a structural element).

(11) Preverbal structural element⁴

Bartolomé …(1.2) !sí.. yotambién firmaba. ‘… yes.. I-PRE too would sign.’

[02 La Marina, 0:43:43–0:43:46]

(12) Post-verbal structural element

Sandra yo me fui con la Lynette. ‘I-PRE left with Lynette.’

[03 Dos comadres, 0:48:25–0:48:26]

Results of multivariate analysis

The results of multivariate analysis using Goldvarb X (Sankoff, Tagliamonte, & Smith, 2005) are displayed in Table 2. Factor groups selected as significant are verb class, other structural elements, and realization of previous coreferential 1sg pronoun.

Table 2.

Factors contributing to the post-positioning of 1sg pronoun yo in the New Mexico Spanish-English Bilingual (NMSEB) corpus (N = 669)^a.

	Input = 0.133	Overall rate: 16%
Factor group	Factor weight	% Post-verbal yo	% of data
Verb class
Quotative—decir	0.90	24/43 (56%)	6%
Copula	0.61	11/48 (23%)	7%
Unaccusative	0.61	12/57 (21%)	9%
Transitive/Intransitive^b	0.43	59/521 (11%)	78%
Other structural elements
Preverbal	0.74	31/107 (29%)	16%
Post-verbal/None^b	0.45	75/562 (13%)	84%
Position of previous expressed 1sg pronoun
Post-verbal yo	0.69	18/60 (30%)	11%
Preverbal I and preverbal yo^b	0.48	84/567 (15%)	89%

Factor groups included in analysis but not selected as significant: Presence of English and Referential distance.

A number of factors within different groups were collapsed in the final analysis as chi square tests for independence showed no significant difference in the distribution: transitive (N = 52/453) and intransitive (N = 7/68) (χ² = 0.083, p = 0.773); post-verbal elements (N = 31/233) and no other elements (44/329) (χ² = 0.001, p = 0.975); preverbal I (N = 33/207) and preverbal yo (N = 51/360) (χ² = 0.328, p = 0.567).

Verb class effects

As can be observed from Table 2, quotatives strongly favor post-position (FW .90) (the closer the FW to 1, the more favorable the factor to post-position). The quotative class is composed exclusively of the verb decir ‘to say.’ Quotative decir represents over one half of all decir tokens (N = 43/64) in the corpus, and the majority of decir tokens with post-verbal yo are quotatives (N = 24/34). Houle (2012) found that, in NMSEB for 3sg, decir favored VS order among full NPs (although not 3sg pronouns). A favoring of post-verbal yo with decir does not appear to be unique to NMSEB, nor to Spanish in contact with English. For example, in the non-contact variety represented in the Corpus of Conversational Colombian Spanish, the post-verbal yo rate with decir is three times as high (N = 11/65) as the overall post-verbal rate (N = 15/422) (Travis, personal communication, based on data presented in Travis, 2007). Likewise, in a study using the CORLEC, which contains conversational data from monolingual Peninsular Spanish speakers, Smith (2011) found that post-verbal yo was preferred with decir (digo yo ‘say I’).

In addition, we observed a strong tendency in relation to the position of the quoted speech: when the quote preceded quotative decir (i.e., “Quote” + decir ), we find a categorical rate of post-verbal yo (13/13), but a post-verbal rate of just one third (11/30) when the quote followed decir (which is nevertheless double the overall rate of 16%). This suggests, then, that not only do we have a strong lexical effect by which decir favors post-verbal yo, but we have a specific construction in the form of “Quote” + decir + yo.

We can also see in Table 2 that unaccusative verbs favor post-position (FW .61), a result that has been observed with respect to full NPs for 3sg specific human subjects within this same corpus (Houle, 2012).

Despite the great attention that these verbs have received in the literature, they represent just 9% of the total (57/669), and only 11% of post-verbal yo (12/106). Furthermore, 75% (43/57) of so-called unaccusative verbs are in fact motion verbs (ir(se) ‘go/leave’, venir ‘come’, salir ‘go out’, llegar ‘arrive’, and entrar ‘enter’), and these motion verbs account for almost all of the post-verbal tokens in this ‘class’ (N = 11/12) (see Table 3).

Table 3.

Yo position according to verb types coded as unaccusative (N = 57).

	Post-verbal yo	Total
Ir(se) ‘go/leave’	4	27
Venir ‘come’	3	8
Nacer ‘be born’	1	5
Salir ‘go out’	1	4
Llegar ‘arrive’	2	3
Entrar ‘enter’	1	1
Total	12	48^a

Other verb types coded as unaccusative but not displaying post-verbal yo tokens are: crecer ‘to grow’ (N = 1), criarse ‘to be raised’ (N = 3), morirse ‘to die’ (N = 1), quedarse ‘to stay/remain’ (N = 3), and sentirse ‘to feel’ (N = 1).

Lastly, we note that the copula favors post-positioning (FW .61), while transitive/intransitives slightly disfavor post-positioning (FW .43), corresponding with results from previous research for both Puerto Rican (Brown & Rivas, 2011, p. 39) and Venezuelan Spanish (Bentivoglio, 1988, p. 19). Also, Houle (2012) found intransitive and transitive verbs to disfavor VS order with 3sg pronouns (although not with full NPs and proper names).

Other structural elements

The results of the multivariate analysis in Table 2 reveal that the presence of a preverbal element (i.e., complement or, adjunct) in the clause strongly favors post-positioning of yo (FW .74). The effect of preverbal adverbials has been investigated by Silva-Corvalán (1982), among others, who found that post-positioning of the subject pronoun was more frequent when a preverbal adverbial was present. Here we find that not only preverbal adverbials (N = 17/31), but also other preverbal adjuncts (e.g., prepositional phrases) or objects (N = 14/31) favor post-positioning almost equally. This result is also found for 3sg pronouns in NMSEB (Houle, 2012). Thus, the results observed here corroborate the findings from other varieties of Spanish on the behavior of subject-pronoun positioning, demonstrating that these highly proficient bilinguals display the same kinds of constraints as monolinguals on subject–verb word order in Spanish, even in a highly frequent code-switching situation.

Structural priming effects

Finally, Table 2 shows a coreferential subject priming effect where, as expected, a preceding post-verbal yo favors post-position (FW .69) more than does a preverbal yo or English I. Importantly we find that preverbal I patterns like preverbal yo, as seen in Table 4 (a chi square test for independence showed no significant difference (χ² = 0.328, p = 0.567) in the distribution of these two groups). Thus, while a preceding, coreferential 1sg subject pronoun disfavors post-verbal yo, there seems to be no difference between the languages of the pronoun: preverbal English I and Spanish yo display the same distributional patterns with respect to disfavoring post-positioning.

Table 4.

Yo position according to previous realization (coreferential subject priming) (N = 627).

	Post-verbal yo		Total
	N	%	N	%
Preverbal I	33	16	207	33
Preverbal yo	51	14	360	57
Post-verbal yo	18	30	60	10
Total	102	16	627	100

Non-significant factor groups

Referential distance

No significant effect for referential distance was found. Table 5 shows the distribution of 1sg tokens according to referential distance, where we can see that the direction is the opposite of what was expected if post-positioning of the 1sg pronoun has a function related to the introduction of new information, with a rate of 21% post-verbal yo when the subject was mentioned in the preceding clause, but 13% when it was mentioned at a distance of five clauses or more. It is worth noting that this is in the opposite direction of unaccusative verbs, which, although numbers are low, do show the expected effect: at a distance of five or more intervening clauses, the post-verbal rate is 36% (5/14), but when the subject is mentioned in the preceding clause, the post-verbal rate is 28% (5/18). Before interpreting the results of Table 5 as supporting loss of discourse-pragmatic constraints, it will have to be established whether and how referential distance affects first-person singular subject pronouns which, as inherently given, may not be sensitive to the same effects as third-person subjects.

Table 5.

Yo position according to referential distance (N = 553).

	Post-verbal yo		Preverbal yo		Total
	N	%	N	%	N	%
Mentioned in preceding clause	34	21%	130	79%	164	30%
1–4 intervening clauses	31	17%	154	83%	185	33%
5+ intervening clauses	26	13%	178	87%	204	38%
Total	91	17%	462	83%	553	100%

Presence of English

The primary question for our study is whether code-switching promotes convergence. We hypothesized that if this is the case, the presence of English should increase the likelihood of a more English-like syntax, that is, preverbal yo. For the most stringent test, we narrowed the window to multi-word English uttered within the same or immediately preceding clause as our target yo tokens. Table 6 displays the distribution of pre- and post-verbal yo tokens for this factor group, and as can be seen, the rate of post-verbal yo is identical when multi-word English was present as when there was no English.

Table 6.

Yo position according to presence of English (N = 498).

	Post-verbal yo		Preverbal yo		Total
	N	%	N	%	N	%
Multi-word	27	17	130	83	157	32
No English	59	17	282	83	341	68
Total	86	17	412	83	498	100

Multi-word English is in the immediate vicinity for approximately a third of the tokens (32%); thus, it is clear that there is an abundant amount of code-switching in this corpus. Despite this, we see no evidence of convergence: 1sg subject pronoun position in NM Spanish is not becoming more English-like, nor is the presence of English in the discourse a factor that influences word order choice.

Discussion and conclusions

We began the analysis of our data looking at rates of post-verbal yo in NMSEB as compared to non-contact varieties of Spanish. We found that the rate of post-verbal yo (16%) observed in the present study falls within the range of, or is higher than, rates of post-verbal yo observed in non-contact varieties. This does not support the convergence hypothesis, which would predict lower rates of post-verbal subjects in contact with English. If we were to solely consider rates in the present analysis we could say that convergence of subject–verb word order in the 1sg has not occurred in this variety of Spanish. However, rates alone give no indication of the factors that contribute to yo post-positioning.

In order to understand the grammar of yo positioning, we considered lexical, structural, and discourse-pragmatic factors. The quantitative analysis applied here finds no evidence that these bilingual speakers have lost the constraints operative in non-contact varieties. Rather, these data elucidate patterns in NM Spanish which are very similar to patterns found in varieties of Spanish devoid of code-switching and show no evidence of influence from English. The position of yo can be accounted for by the frequency of a specific construction (“Quote” + 1sg Decir + yo, a language-internal factor not resulting from contact, accounting for one quarter of post-posed tokens), simple localized priming effects (i.e., previous post-verbal yo favors post-verbal yo), and the presence of a preverbal structural element.

If NM Spanish is adopting more English-like structures, we might expect to see evidence of this with respect to subject–verb word order, since English subject pronouns are nearly categorically preverbal. Importantly, there is a striking lack of an effect for the presence of multi-word English, even when uttered within the same or immediately preceding clause, a result which mirrors that found by Torres Cacoullos and Travis (2015b) for 1sg subject expression in these data.

We believe that previous research on parallel activation and language switching is highly relevant to this discussion of convergence in code-switching, as it allows analysts to form hypotheses about what happens in the bilingual brain during the production of code-switching and provides complementary information to the patterns we observe on the surface through corpus research. Recent psycholinguistic research indicates that bilinguals have both of their languages activated in parallel at all times, even when only one language is apparently necessary based on the language of the task at hand or context (e.g., Libben & Titone, 2009; Marian & Spivey, 2003; Schwartz & Kroll, 2006; Schwartz et al., 2007; van Hell & De Groot, 2008). Despite this, bilinguals are also very good at switching languages, which has important implications for the purported role of code-switchers in convergence. In fact, the case of being good switchers is especially true of bilinguals, such as those studied in our data, who acquired their two languages either simultaneously or sequentially, but very early in life (certainly before the “critical period”, see Lenneberg, 1967; cf. Birdsong, 1999). While some research has found asymmetrical switching costs among L2 learners (Meuter & Allport, 1999), Costa and Santesteban (2004) found that for highly proficient early bilinguals the costs of switching between their L1 (Spanish) and their L2 (Catalan) are not asymmetrical, nor do they manifest asymmetrical switching costs for their much weaker and later acquired L3 (English). The authors take this to indicate that highly proficient bilinguals have mastered the task of language switching in general, and that switching costs have nothing to do with the relative strength of the language pairs, but rather, with the far more vast experience that the highly proficient bilinguals have doing the cognitive work of switching and inhibiting languages.

If we accept that code-switching is not “mixing” languages, but rather highly skilled alternation between languages (cf. Haugen, 1950; Kroll, Dussias, Bogulski, & Valdés Kroff, 2012; Poplack, 1980), then the skill with which a particular group of bilinguals is able to switch to one language while inhibiting the other should be brought to bear on the role these bilinguals play in being the agents of language change. Based on the sheer quantity of spontaneous, un-elicited code-switching present in this corpus and the high level of ability that NM Spanish-English bilinguals have in their two languages, the participants can be said to be skilled switchers. Therefore, it is unlikely that bilingual behavior such as code-switching undertaken by these speakers would lead to change in NM Spanish resulting in a more English-like grammar (convergence). This assumption is supported by the results of the present study.

Taking our results in conjunction with the psycholinguistic literature on bilingual parallel activation and language switching, we conclude that there is no indicator of convergence in NM Spanish. If code-switching is a mechanism of change then, given the abundant amount of code-switching in the present corpus, we should have been able to observe change with these speakers. This leads us to conclude that NM Spanish-English bilinguals may, in fact, be able to maintain two separate and independent grammars on both the individual and community levels, contrary to previous claims that frequent code-switching results in convergence of grammars.

Footnotes

Acknowledgements

We thank Rena Torres Cacoullos and Catherine Travis for permission to use the NMSEB Corpus. Many thanks also to the anonymous reviewers for their invaluable input into the development of this work. This material is based in part upon work supported by the National Science Foundation under Grant No. DGE1255832 to Amelia J Dietrich. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

Funding

This work was partially supported by National Science Foundation (NSF) grant #1019112/1019122 awarded to Rena Torres Cacoullos and Catherine E. Travis to support the development of the NMSEB corpus.

Notes

Author biographies

Nicole M. Benevento is a PhD candidate in Hispanic Linguistics and Language Science at The Pennsylvania State University. Her research to date has focused on utilizing corpus data to investigate lexical stress and prominence in connected speech. Her research interests include phonology, bilingualism, and language contact.

Amelia J. Dietrich earned a PhD in Hispanic Linguistics and Language Science from The Pennsylvania State University. Her research interweaves corpus-based and experimental psycholinguistic methods to answer questions about language usage patterns and the role of frequency-based information in bilingual language production and comprehension with the aim of informing models of bilingual grammar and general models of language processing. She is the recipient of a National Science Foundation Graduate Research Fellowship.

References

Aaron

(2015). Lone English-origin nouns in Spanish: The precedence of community norms. International Journal of Bilingualism, 19(4), 459–480.

Backus

(2005). Codeswitching and language change: One thing leads to another? International Journal of Bilingualism, 9, 307–340.

Bentivoglio

(1988). La posición del sujeto en el español de Caracas: Un análisis de los factores lingüísticos y extralingüísticos. In Hammond

R. M.

Resnick

M. C.

(Eds.), Studies in Caribbean Spanish dialectology (pp. 13–23). Washington, DC: Georgetown University Press.

Bills

G. D.

Vigil

N. A.

(2008). The Spanish language of New Mexico and southern Colorado: A linguistic atlas. Albuquerque, NM: University of New Mexico Press.

Birdsong

(Ed.). (1999). Second language acquisition and the critical period hypothesis. Mahwah, NJ: Lawrence Erlbaum Associates.

Brown

Rivas

(2011). Subject – verb word-order in Spanish interrogatives: A quantitative analysis of Puerto Rican Spanish. Spanish in Context, 8, 23–49.

Cameron

Flores-Ferrán

(2004). Perseverance of subject expression across regional dialects. Spanish in Context, 1, 41–65.

Chafe

(1994). Discourse, consciousness, and time: The flow and displacement of conscious experience in speaking and writing. Chicago, IL: University of Chicago Press.

Champi

Perrotti

(2013). Constraints on yo expression in Chilean Spanish Conversation. Unpublished manuscript, Penn State University.

10.

CORLEC: Corpus de Referencia de la Lengua Española Contemporánea: Corpus Oral Peninsular, director F. Marcos Marín. Retrieved from www.lllf.uam.es/~fmarcos/informes/corpus/corpusix.html (género conversacional).

11.

Costa

Santesteban

(2004). Lexical access in bilingual speech production: Evidence from language switching in highly proficient bilinguals and L2 learners. Journal of Memory and Language, 50, 491–511.

12.

Givón

(1983). Topic continuity in discourse: An introduction. In Givón

(Ed.), Topic continuity in discourse: A quantitative cross-linguistic study (pp. 1–41). Amsterdam, The Netherlands: John Benjamins.

13.

Gumperz

Wilson

(1971). Convergence and creolization: A case from the Indo-Dravidian border. In Hymes

(Ed.), Pidginization and creolization of languages (pp. 151–167). Cambridge, UK: Cambridge University Press.

14.

Gundel

J. K.

Hedberg

Zacharski

(1993). Cognitive status and the form of referring expressions in discourse. Language, 69, 274–307.

15.

Haugen

(1950). The analysis of linguistic borrowing. Language, 26, 211–231.

16.

Herring

S. C.

(1990). Information structure as a consequence of word order type. Proceedings of the Sixteenth Annual Meeting of the Berkeley Linguistics Society (pp. 163–174).

17.

Houle

(2012). Cognitive and construction effects in Spanish word order. Unpublished paper presented at The Australian Linguistics Society, Australian National University.

18.

Kroll

J. F.

Dussias

P. E.

Bogulski

C. A.

Valdés Kroff

J. R.

(2012). Juggling two languages in one mind: What bilinguals tell us about language processing and its consequences for cognition. In Ross

(Ed.), The psychology of learning and motivation, volume 56 (pp. 229–262). San Diego, CA: Academic Press.

19.

Labov

(1969). Contraction, deletion, and inherent variability of the English copula. Language, 45, 715–762.

20.

Lastra

Martín Butragueño

(2015). Subject pronoun expression in oral Mexican Spanish. In Carvalho

A. M.

Orozco

Lapidus Shin

(Eds.), Subject pronoun expression in Spanish: A cross-dialectal perspective (pp. 39–57). Washington, DC: Georgetown University Press.

21.

Lenneberg

E. H.

(1967). Biological foundations of language. New York: Wiley.

22.

Libben

M. R.

Titone

(2009). Bilingual lexical access in context: Evidence from eye movements during reading. Journal of Experimental Psychology. Learning, Memory, and Cognition, 35, 381–390.

23.

Lipski

(1996). [Review of the book Language Contact and Change: Spanish in Los Angeles by C. Silva-Corvalán.] Language, 72, 146.

24.

Marian

Spivey

M. J.

(2003). Competing activation in bilingual language processing: Within- and between-language competition. Bilingualism: Language and Cognition, 6, 97–115.

25.

Mayoral Hernández

(2006). A variation study of verb types and subject position: Verbs of light and sound emission. In Camacho

Flores-Ferrán

Sánchez

Déprez

Cabrera

M. J.

(Eds.), Romance linguistics 2006. Selected papers from the 36th linguistic symposium on romance languages (LSRL) (pp. 213–226). Amsterdam, The Netherlands/Philadelphia, PA: John Benjamins.

26.

Meuter

R. F. I.

Allport

(1999). Bilingual language switching in naming: Asymmetrical costs of language selection. Journal of Memory and Language, 40, 25–40.

27.

Miller

Schmitt

(2012). Chilean Conversation Corpus.

28.

Montrul

(2004). Subject and object expression in Spanish heritage speakers: A case of morphosyntactic convergence. Bilingualism: Language and Cognition, 7, 125–142.

29.

Ocampo

(1995). The word order of two-constituent constructions in spoken Spanish. In Downing

Noonan

(Eds.), Word order in discourse (pp. 425–447). Amsterdam, The Netherlands/Philadelphia, PA: John Benjamins.

30.

Poplack

(1978). Dialect acquisition among Puerto Rican bilinguals. Language Sociology, 7, 89–103.

31.

Poplack

(1980).“Sometimes I’ll start a sentence in Spanish y termino en español”: Toward a typology of code-switching. Linguistics, 18, 581–618.

32.

Poplack

(2012). What does the Nonce Borrowing Hypothesis hypothesize? Bilingualism: Language and Cognition, 15, 644–648.

33.

Poplack

Sankoff

Miller

(1988). The social correlates and linguistic processes of lexical borrowing and assimilation. Linguistics, 26, 47–104.

34.

Poplack

Zentz

Dion

(2012). Phrase-final prepositions in Quebec French: An empirical study of contact, code-switching and resistance to convergence. Bilingualism: Language and Cognition, 15, 203–225.

35.

Posio

(2011). Spanish subject pronoun usage and verb semantics revisited: First and second person singular subject pronouns and focusing of attention in spoken Peninsular Spanish. Journal of Pragmatics, 43, 777–798.

36.

Risso

(2010). Subject pronoun placement as evidence of contact and leveling in Spanish in New York. International Journal of the Sociology of Language, 203, 101–114.

37.

Rivas

(2008).La posición del sujeto en las construcciones monoactanciales del español: Una aproximación funcional. Hispania, 91, 897–912.

38.

Sankoff

(1980). The social life of language. Philadelphia: University of Pennsylvania Press.

39.

Sankoff

Poplack

(1981). A formal grammar for code-switching. Papers in Linguistics: International Journal of Human Communication, 14, 2–45.

40.

Sankoff

Poplack

Vanniarajan

(1990).The case of the nonce loan in Tamil. Language Variation and Change, 2, 71–101.

41.

Sankoff

Tagliamonte

S. A.

Smith

(2005). Goldvarb X: A multivariate analysis application [Computer program]. Retrieved 15 May 2013 from http://individual.utoronto.ca/tagliamonte/goldvarb.htm

42.

Schwartz

A. I.

Kroll

J. F.

(2006). Bilingual lexical activation in sentence context. Journal of Memory and Language, 55, 197–212.

43.

Schwartz

A. I.

Kroll

J. F.

Diaz

(2007). Reading words in Spanish and English: Mapping orthography to phonology in two languages. Language and Cognitive Processes, 22, 106–129.

44.

Serrano

M. J.

(2012). El sujeto pronominal usted/ustedes y su colocación: variación y creación de estilos comunicativos. Spanish in Context, 9, 109–131.

45.

Silva-Corvalán

(1982). Subject expression and placement in Mexican-American Spanish. In Amastae

Elías-Olivares

(Eds.), Spanish in the United States: Sociolinguistic aspects (pp. 93–120). Cambridge, UK: Cambridge University.

46.

Silva-Corvalán

(1994). Language contact and change: Spanish in Los Angeles. Oxford, UK: Oxford University Press.

47.

Silva-Corvalán

(2001). Sociolingüística y pragmática del español. Washington, DC: Georgetown University Press.

48.

Smith

H. L.

(2011). A look at subject pronoun placement: Post verbal “yo” in Spanish. Unpublished manuscript, Penn State University.

49.

Sorace

(2004). Native language attrition and developmental instability at the syntax–discourse interface: Data, interpretations and methods. Bilingualism: Language and Cognition, 7, 143–145.

50.

Thomason

S. G.

(2000). Linguistic areas and language history. In Gilbers

Nerbonne

Schaeken

(Eds.), Languages in contact (pp. 311–327). Amsterdam, The Netherlands: Rodopi.

51.

Thomason

S. G.

Kauffman

(1988). Language contact, creolization and genetic linguistics. Berkeley, CA: University of California Press.

52.

Toribio

A. J.

(2004). Convergence as an optimization strategy in bilingual speech: Evidence from code-switching. Bilingualism: Language and Cognition, 7, 165–173.

53.

Torres Cacoullos

Travis

C. E.

(2011).Testing convergence via code-switching: Priming and the structure of variable subject expression. International Journal of Bilingualism, 15, 241–267.

54.

Torres Cacoullos

Travis

C. E.

(2013). Spanish and English in New Mexico: juntos pero no revueltos. Keynote, Celebrating varieties of New Mexican Spanish. Department of Spanish and Portuguese, University of New Mexico, September 27.

55.

Torres Cacoullos

Travis

C. E.

(2015a). Gauging convergence on the ground: Code-switching in the community. International Journal of Bilingualism, 19(4), 365–386.

56.

Torres Cacoullos

Travis

C. E.

(2015b). Two languages, one effect: Structural priming in code-switching. Bilingualism: Language and Cognition. DOI: https://dx-doi-org.web.bisu.edu.cn/10.1017/S1366728914000406

57.

Torres Cacoullos

Travis

C. E.

(in preparation). New Mexico Spanish-English Bilingual (NMSEB) corpus, National Science Foundation 1019112/1019112, http://nmcode-switching.la.psu.edu/

58.

Travis

C. E.

(2007). Genre effects on subject expression in Spanish: Priming in narrative and conversation. Language Variation and Change, 19, 101–135.

59.

Travis

C. E.

Torres Cacoullos

(2012). What do subject pronouns do in discourse? Cognitive, mechanical and constructional factors in variation. Cognitive Linguistics, 23, 711–748.

60.

Travis

C. E.

Torres Cacoullos

(2013). Making voices count: Corpus compilation in bilingual communities. Australian Journal of Linguistics, 33, 170–194.

61.

van Hell

J. G.

De Groot

A. M. B.

(2008). Sentence context modulates visual word recognition and translation in bilinguals. Acta psychologica, 128, 431–451.

62.

Winford

(2008). Processes of creole formation and related contact-induced language change. Journal of Language Contact, THEMA II, 124–145.