Spanish–English bilingual heritage speakers processing of inanimate sentences
Abstract
This study investigated how heritage Spanish–English bilinguals integrate their cue hierarchies to process simple sentences in both Spanish and English. Sentence interpretation is achieved by weighing the various cues that are present in the sentence against that language's cue hierarchy. The Unified Competition Model (UCM) suggests that bilinguals show a variety of patterns in sentence interpretation strategies depending on language proficiency. Previous research on heritage Spanish–English bilinguals and late bilinguals has demonstrated differences in cue utilization and sentence interpretation compared to monolinguals. However, good-enough processing suggests that when a sentence does not meet certain heuristics, like the first-noun agent heuristic, a semantic representation of the sentence will be processed instead of a syntactic one. Even with reliable sentential-level cues such as word order, a plausible semantic representation of the sentence is favored. This is especially the case with inanimate–inanimate (IA-IA) sentences, like in the present study, in which there is less reversibility of thematic roles without competing with semantic plausibility. For this study, participants (n = 32) read a total of 80 inanimate sentences in English and Spanish with subject–verb–object and object–verb–subject (OVS) word orders, indicated the subject of the sentences, and completed language proficiency and dominance tasks. When reading Spanish sentences, participants read the OVS word order faster. English proficiency was a significant predictor of sentence reading time and choice selection time but did not predict word choice. The results suggest that IA-IA sentences pose challenges for cue utilization and thematic role assignment due to semantic limitations. This study found that participants may prioritize semantic plausibility over syntactic representations in sentence processing, supporting a good-enough processing model.
Study specs
- Institution
- University of California
- Discipline
- Linguistics,Psychology
- Year
- 2024
- Human Data Platform
- Prolific
- Source
- View Source Google Scholar
Peer Review & Critical Discussion
Potential Selection Bias in 2023 Cohort
The participant pool shows a concerning overrepresentation of users from high-income demographics. Looking at Table 3, we can see that 78% of respondents had annual incomes above $75k, which significantly limits the generalizability of these findings to broader populations.
Non-naive Participants Issue
I've noticed a methodological concern regarding participant naivety. Given that Prolific users often complete multiple studies, there's a real risk that participants had prior exposure to similar experimental paradigms, which could confound the results.
RLHF Applicability to This Study Design
The implications for RLHF training pipelines are understated. If we accept the authors' conclusions about preference stability, this has direct consequences for how we should structure reward model training. The temporal decay effect described in Section 4.2 is particularly relevant.
Verify your expertise to join discussion
Create an account and verify your credentials to participate in peer discussions.