Characterizing multilingual speech patterns in Northern Belize: Profiling code-switches of 18 English-Spanish-Kriol speakers in Orange Walk Town, Belize

Objectives/research questions: In the present study, we have two main objectives. First, we aim to characterize the proportions of code-switching in the multilingual spontaneous speech of the community in Orange Walk Town, Northern Belize. We analyze overall rates of unilingual versus multilingual speech production and the distribution of the languages used (Spanish, Kriol, and English). To identify the language asymmetry in intraclausal code-switching, we apply the Matrix Language Frame (MLF) Model (Myers-Scotton, 2002), which posits that the Matrix Language (ML) provides the clause's morphosyntactic structure and that system morphemes align with the ML. Previous studies indicate that the speaker’s choice of ML in multilingual clauses can depend on various extralinguistic factors (e.g., Parafita Couto, Davies, Carter & Deuchar, 2014). Therefore, this study’s second objective is to examine whether a relation exists between the choice in ML and three sociolinguistic factors: (i) self...Show moreObjectives/research questions: In the present study, we have two main objectives. First, we aim to characterize the proportions of code-switching in the multilingual spontaneous speech of the community in Orange Walk Town, Northern Belize. We analyze overall rates of unilingual versus multilingual speech production and the distribution of the languages used (Spanish, Kriol, and English). To identify the language asymmetry in intraclausal code-switching, we apply the Matrix Language Frame (MLF) Model (Myers-Scotton, 2002), which posits that the Matrix Language (ML) provides the clause's morphosyntactic structure and that system morphemes align with the ML. Previous studies indicate that the speaker’s choice of ML in multilingual clauses can depend on various extralinguistic factors (e.g., Parafita Couto, Davies, Carter & Deuchar, 2014). Therefore, this study’s second objective is to examine whether a relation exists between the choice in ML and three sociolinguistic factors: (i) self-rated proficiency, (ii) code-switching attitude, and (iii) language use in the participant’s social network. Methodology: We selected and analyzed a subset of nine video recordings from a dataset comprising 50 video recordings of multilingual spontaneous speech from Orange Walk Town. In this subset, eighteen participants (8 females, 10 males, aged 18–25 years, plus one 76-year-old) engaged in paired conversations lasting approximately 15 minutes each (total: 140 minutes). Participants also completed detailed language background questionnaires, providing information on, i.e., proficiency, code-switching attitudes, and language use in their social networks. Data & analysis: All speech data were transcribed at the clausal level using ELAN transcription software, with further coding and analysis done in Excel. Results: A total of 4010 clauses were identified, with 3214 (80%) being unilingual, and 632 (16%) multilingual. All participants produced both unilingual Anglophone (English or Kriol) and Spanish clauses within their conversations, indicating habitual engagement in multilingual speech and code-switching within and between clauses or full sentences. Among the unilingual clauses, 1921 (60%) were Anglophone (English or Kriol), slightly surpassing the 1291 (40%) Spanish unilingual clauses. Although no considerable language preference was observed in unilingual clauses, there was a marked preference 3 for Spanish as the ML in multilingual clauses; with 539 (87%) being Spanish ML versus 83 (13%) Anglophone ML. Proficiency levels related to language preference for unilingual outcomes but not for multilingual outcomes. The preference for Spanish ML in multilingual clauses may be related to the prestigious status of Kriol. Further exploration of the speaker’s social network data is needed to relate this factor to speech production outcomes. Originality: This study presents new insights by analyzing recently collected spontaneous speech data from a young generation in a relatively unexplored multilingual community. The findings shed light on how proportions of different languages are distributed in multilingual speech, shaped by community norms or individual sociolinguistic factors. Implications: The study reaffirms that code-switching is a common practice in Northern Belize. Despite individual differences in the proportions of unilingual versus multilingual speech, the choice of ML in multilingual speech reflects a community-wide norm. This ML choice, which inherently involves an asymmetry in language dominance within a clause, can be influenced by sociolinguistic factors, such as switching toward a language with higher prestige, as found by Blokzijl, Deuchar and Parafita Couto (2017). Future research should delve deeper into this Northern Belizean community on a larger scale, considering variables such as age and individual language attitudes toward Kriol, Spanish, and English. Additionally, examining code-switching patterns within conflict sites, such as determiner-noun-adjective constructions, could enhance our understanding of how code-switching operates and whether community standards influence these patterns.Show less

Leiden University Student Repository

Documents

In Collections

Characterizing multilingual speech patterns in Northern Belize: Profiling code-switches of 18 English-Spanish-Kriol speakers in Orange Walk Town, Belize