Child heritage speakers’ acquisition of the Spanish subjunctive in volitional and adverbial clauses

ABSTRACT

The Spanish subjunctive mood (SUBJ) is said to be highly vulnerable in heritage language (HL) acquisition. However, there is little controlled research on HL-speaking children acquiring the various Spanish SUBJ contexts, so we do not have a clear picture of when, how, or why heritage speakers (HSs) develop in the SUBJ as they do. This study tests the development of the SUBJ in two of the earliest acquired contexts by monolingual children—SUBJ with volitional clauses and adverbial clauses with future reference. Through an oral sentence-completion task administered to 50 school-aged child HSs, this study observes whether language-internal factors (modality, variability) and speaker factors (age, exposure/use, or morphosyntactic proficiency) influence acquisition of the SUBJ in the examined contexts. Although SUBJ is categorically used in the first-generation input the child HSs receive at home, school-aged HSs exhibit elevated optionality; the majority show a pattern of use typical of very young monolingual children, and there is wide variance among the child HSs across all ages. Overall, they exhibit slightly more optionality within epistemic modality (adverbials) than deontic modality (volition). Crucially, exposure to and use of Spanish and, even more so, a standardized measure of Spanish morphosyntactic proficiency were strongly associated with SUBJ use in both contexts by the child HSs. We argue that the observed vulnerability in these early-acquired SUBJ contexts follows from an interaction between the child HSs’ engagement with the HL environment (including their resulting command of the HL grammar) and linguistic factors common to all SUBJ contexts.

Acknowledgments

We would like to express our sincere thanks to all participants, parents, and school administrators for their participation and support. We are also very thankful to Raquel Nuñez, Russell Miller, Dustin Lyles, Mackenzie Chandler, and Alex Baadsgaard for their help with data collection and transcription. We are grateful to the editor and anonymous reviewers for their insightful comments on earlier versions of this article. Of course, any remaining errors are exclusively ours.

Data availability

The data that support the findings of this study are available from the corresponding author, MD, upon reasonable request.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

1 Throughout the manuscript, the abbreviation HS or HSs will be used as both a noun (e.g., “child HSs” = “child heritage speakers”) and as an adjective (e.g., “HS children” = “heritage-speaking children” and “HS SUBJ use” = “heritage speaker subjunctive use”).

2 perfut means periphrastic future.

3 This example from Gonzalez (1978:38) corresponds to a Spanish-speaking child from a Spanish-dominant home in the Lower Rio Grande Valley of Texas (near the border with Mexico). We selected this example because it clearly illustrates early target SUBJ use with cuando ‘when,’ which we assume to emerge by the same age, if not earlier, among monolingual children.

4 Although Castilla-Earls et al. (2021) examined production of the SUBJ by a sample of typically developing bilingual children ages 4; 00–6; 11 including contexts of volitional and temporal clauses, rates by the context are not reported separately. Overall accurate SUBJ responses amounted to 55%.

5 For the scales on Exposure (a) and Use (b), parents were asked to report information about weekdays and weekends separately, but we then averaged the percentages reported. This resulted in a single score for Exposure and a single score for Use.

6 The variable Current exposure and use of Spanish at school represented the children’s Spanish exposure to and use of Spanish during the academic year prior to the summer when data collection for this study took place (during an optional summer program). Based on the Likert scale created (see Appendix A), the average score for Current exposure and use of Spanish at school was 3.10 (SD = 1.21), but there was a range between 1 and 4, which depended upon whether the children received some Spanish instruction or had a bilingual support aid in the classroom, or if they just used Spanish minimally in social contexts. None of the child participants received instruction in Spanish through any sort of evening or weekend heritage language school or institute. The children’s current exposure and use of Spanish at school correlated negatively with age in months [r = –.86, p < .0001]. This is explained by the fact that the local district where data were collected followed an early exit bilingual model designed to transition students to English-only classrooms by the fourth grade. However, during the academic year before this study, the district had increased the amount of Spanish instruction provided in the pre-kindergarten and kindergarten grades with the goal of moving toward implementing a dual-language model in the elementary school grades in the future. This resulted in even greater Spanish support for the youngest children in this study.

7 We acknowledge that BESA-ME was designed and normed to be used with Spanish-English bilingual children ages 7;0-10;11, yet we tested seven children between ages the ages of 11-14. Since we were not using this assessment to identify language impairment (one of its main uses) but rather as a proficiency measure, for consistency we deemed it appropriate to use the same BESA-ME morphosyntactic subtest with the seven child participants that exceeded the upper age limit. None of these participants scored at ceiling in either the English or Spanish morphosyntactic subtest.

8 There was no significant correlation between Spanish proficiency and age [r = .19, p = .160] for the children tested in this study. As would be expected, though, there was a strong correlation between English proficiency and age [r = .73, p < .0001].

9 To confirm that effects of Age were not obscured by the inclusion of multiple predictors in the model or collinearity with Proficiency (a question raised by an anonymous reviewer), we ran additional analyses with Age. We examined Age in a univariate model and found no significant relationship between Age and SUBJ use. This remains the case if Condition is added into the model. An increase in age did not result in greater SUBJ use by the HS children for either of the two contexts tested. Also, the correlation between Spanish Proficiency and Age was only r = .25, which was too low to cause an issue of collinearity.

10 We acknowledge that the BESA and BESA-ME tests include four cloze items that directly assess children’s knowledge of subjunctive as part of the overall assessment of their morphosyntactic proficiency. Since the BESA and BESA-ME are standardized tests, we could not exclude a nonstandardized count (ranging from 0 to 4 based on accuracy on the subjunctive items) from a standardized score. However, to ensure that the effect of overall morphosyntactic proficiency was not driven by these four questions that directly tapped subjunctive knowledge, we ran a separate model that controlled for participants’ scores on those four items. In this model, Spanish morphosyntactic proficiency was still significant (p < .001), and the OR continued to show a large effect size.

11 For Standardized Proficiency, the cutoff points were –6.04 to –1.47 (Low), –1.09 to 0.25 (Mid), 0.44 to 6.12 (High). With respect to Spanish Exposure and Use, the cutoff points were 2.0 to 3.0 (Low), 3.5 to 4.0 (Mid), 4.5 to 5.0 (High).

12 The contrast estimate is the difference in the probabilities that an event will happen. In other words, in this case, the High-Proficiency group has a .31 greater probability of producing the SUBJ as compared to the Mid-Proficiency group, and the High-Proficiency group has a .70 greater probability of producing the SUBJ as compared to the Low-Proficiency group.

13 In the few cases in which it was difficult to determine between 2 and 3 points, or between 3 and 4 points, a decision was made based on the language spoken with close friends and family members who spend a significant amount of time with the child.