A corpora-based analysis of directionality and part of speech effects on Spanish-English code-switching in Florida and Texas

Code-switching, or the rapid alternation from one language to another, is of particular interest to linguists because bilinguals can rapidly code-switch without causing any difficulties in communication. Analyses of bilingual corpora (collections of spoken or written linguistic data) have highlighted that in Spanish-English code-switching, switching between certain part of speech categories (i.e., determiner-noun; el flag) is more common than others. Similarly, descriptive trends of Spanish-English code-switching have highlighted that broadly speaking, switching from Spanish to English (i.e., el flag) is more common than in the opposite direction (i.e., the bandera). However, these investigations have largely been based on solely the Bangor Miami Corpus, leaving it unknown as to whether similar trends occur in corpora from different regions and/or different modalities (i.e., oral data).

The present study investigated whether Spanish-English bilinguals code-switched between similar parts of speech and in similar directions via oral and written corpora created in Texas and Florida. To accomplish this, we analyzed the following corpora:

  1. The Bangor Miami Corpus: oral corpus, created in Miami, FL, contains 242,475 words (34% Spanish, 63% English)
  2. The Spanish in Texas Corpus: oral corpus, created in Austin, TX, contains approximately 500,000 words (96% Spanish, 4% English)
  3. The LinCe Corpus: written corpus, created in Austin, TX, contains 390,953 words (33% Spanish, 64% English)

This project was guided by the following research questions:

  1. To what degree is the preference for switching from Spanish to English found in the Bangor Miami Corpus attested in other corpora?
  2. To what degree are part of speech effects attested across corpora?
  3. If part of speech/directionality effects are found across corpra, do the appear differently in written/oral corpora?

This project found that interjections and (proper) nouns typically preceded a code-switch. Additionally, coordinating conjunctions (i.e., but), interjections, and nouns typically followed code-switches. Lastly, this project found that although switching from Spanish to English was more common in the Bangor Miami and LinCe corpora, the opposite was true for the Spanish in Texas Corpus.n

These findings simultaneously attest and question the prevalence of Spanish to English code-switching found in the Bangor Miami corpus while also discovering part of speech trends (i.e., the prevalence of nouns preceding a code-switch and coordinating conjunctions following a code-switch) that were previously unattested.

-Project concluded- Co-authored manuscript) published in the proceedings of QUASY hosted at SyntaxFest

I led this project in collaboration with Dr. Zoey Liu and MA Valeria Pagliai