Skip to content

Code-Switching

Prime #
319
Origin domain
Linguistics & Semiotics
Also from
Computer Science & Software Engineering
Aliases
Language Switching, Register Alternation, Dialectal Shift
Related primes
Register (Style) Shifting, Contextual Mode Switching, Variation and Sociolect

Core Idea

Code-Switching is the practice of: (1) alternating between two or more distinct linguistic codes (languages, dialects, jargons, registers) within a single conversation, utterance, or even sentence;[1] (2) triggered by social, pragmatic, or expressive goals — signaling group identity, accommodating an interlocutor's competence, quoting, framing, emphasizing, or achieving stylistic effect;[2] (3) drawing on the speaker's full repertoire of codes as an active resource, not treating any one as the default;[3] (4) operating rule-governed, not haphazard — code-switching has morpho-syntactic constraints on where switches can occur and sociolinguistic constraints on who switches with whom and when.[4]

At its core, code-switching is a structured alternation driven by four components: the linguistic codes, the switching point, the matrix language, and the discourse function. Foundational work by Gumperz (1982),[2] Myers-Scotton (1993),[5] and Poplack (1980)[4] established code-switching as a legitimate linguistic phenomenon with predictable constraints, moving beyond deficit-model framings that treated it as confusion or incompetence.

How would you explain it like I'm…

Changing How You Talk

Imagine you talk one way with your grandma and a different way with your friends on the playground. Sometimes, in the same sentence, you mix the two ways together on purpose. That's code-switching. People do it to fit in, to be understood, or to show who they are.

Switching Languages Mid-Talk

Code-switching means going back and forth between two or more ways of speaking — different languages, dialects, or styles — sometimes inside a single sentence. People do it for good reasons: to show they belong to a group, to match the person they're talking to, to quote someone, or to add emphasis. Speakers use their whole set of languages and styles as a kind of toolbox. It isn't random or sloppy; there are quiet rules about where you can switch and when it makes sense.

Alternating Between Linguistic Codes

Code-switching is the practice of alternating between two or more distinct linguistic codes — languages, dialects, jargons, or registers — within a single conversation, utterance, or even sentence. It's driven by social and expressive goals: signaling identity, accommodating a listener, quoting, framing, or stylistic effect. Bilingual and multilingual speakers draw on their full repertoire as an active resource rather than treating one code as the default. Crucially, code-switching is rule-governed, not random. There are grammatical constraints on where switches can occur within a sentence, and social constraints on who switches with whom and in what context. Earlier views that dismissed it as confusion or incompetence have been replaced by recognition that it's a sophisticated linguistic skill.

 

Code-switching is the practice of alternating between two or more distinct linguistic codes — languages, dialects, jargons, or registers — within a single conversation, utterance, or even sentence. It is driven by social, pragmatic, and expressive goals: signaling group identity, accommodating an interlocutor's competence, quoting, framing, emphasizing, or producing stylistic effect. Bilingual and multilingual speakers draw on their full repertoire of codes as an active resource rather than treating any single code as default. The phenomenon is rule-governed, not haphazard: it obeys morpho-syntactic constraints (the matrix-language frame and equivalence constraint set out by Poplack and Myers-Scotton) on where switches may occur, alongside sociolinguistic constraints on who switches with whom and when. Structurally, code-switching has four components: the linguistic codes involved, the switching point, the matrix language that provides the grammatical frame, and the discourse function the switch performs. Foundational work by Gumperz, Myers-Scotton, and Poplack established code-switching as a legitimate linguistic phenomenon with predictable constraints, supplanting deficit-model framings that treated it as confusion.

Structural Signature

A coordinated deployment of multiple linguistic systems by a single speaker, governed by six italicized role-phrases:

  • The linguistic codes — the distinct varieties (languages, dialects, registers) available to the speaker
  • The switching point — the site of code alternation (clause boundary, phrase boundary, single-word insertion)
  • The matrix language — the base code within which other codes are inserted, subject to morpho-syntactic compatibility
  • The discourse function — the pragmatic work the switch performs (identity, audience accommodation, topic signaling)
  • The syntactic constraint — the grammatical rules governing where switches can occur (Poplack's equivalence constraint, Myers-Scotton's Matrix Language Frame model)
  • The social-identity signal — the marking work a switch performs (foregrounding membership, solidarity, or distance)

Switches occur at well-defined points governed by grammatical compatibility (the Matrix Language Frame model, Myers-Scotton 1993)[5] and by social signaling (the markedness model). The switch is communicative, not lapse — speakers choose codes based on who is present, what topic is active, what identity is being foregrounded. Listeners decode the switch pragmatically (why did the speaker switch here?) alongside the literal content.

What It Is Not

  • Not borrowing — a borrowed word is fully integrated into the borrowing language's phonology and morphology (café in English is no longer a French word). Code-switching preserves the switched material as belonging to the other code.
  • Not interference or errors — interference is unintentional cross-linguistic influence, often treated as an imperfection. Code-switching is deliberate and competent use of multiple codes.
  • Not translation — translation moves content from code A to code B. Code-switching alternates codes while delivering content, often without any translation occurring.
  • Not register/style shifting (#321) — style-shifting moves along a formality or domain dimension within the same code. Code-switching moves between distinct codes. They can co-occur: a speaker can simultaneously code-switch (English→Spanish) and style-shift (formal→casual). Flagged as companion.
  • Not contextual mode switching (#322) — that prime operates at a higher level of abstraction covering any mode/context shift; code-switching is the specific linguistic-code variant. Related but not identical.
  • Not pidgin or creole creation — code-switching is a practice by multilingual individuals; pidgins and creoles are stable linguistic systems emerging from contact-language genesis. Distinct phenomena.
  • Not all bilingual behavior — not every utterance by a bilingual speaker is code-switched; bilinguals also maintain codes monolingually. Code-switching is the selection act, not the mere fact of knowing multiple codes.

Broad Use

  • Bilingual and multilingual communities (core domain): Code-switching between heritage and dominant languages in diasporic communities; between official and vernacular languages; between dialects in polyglossic regions. This is the foundational site of code-switching research.
  • Professional communication: Experts switching between technical jargon and lay language for mixed audiences — doctor-patient, engineer-business-stakeholder, lawyer-client. Code-switching manages audience heterogeneity within a single institutional context.
  • Online communication: Mixing slang and formal writing; switching between platform-specific registers (LinkedIn-formal, Twitter-compressed, Discord-casual) within a single conversation that spans channels. Digital platforms create new code-switching pressures and opportunities.
  • Marketing and branding: Bilingual ad campaigns deliberately mix codes (Spanglish, Chinglish, Franglais) to signal inclusivity or cultural authenticity. Intentional code-switching is now a branded communicative strategy.
  • Literary and artistic use: Authors employ code-switching for voice, atmosphere, or character distinction (Cormac McCarthy's untranslated Spanish, Junot Díaz's Spanglish). Code-switching is an aesthetic tool for representing multilingual consciousness.
  • Organizational speech: Switching between "business code" with clients and "shop code" among peers; front-stage vs. back-stage registers (Goffman's frame analysis). Organizational cultures establish norms for code-switching across role contexts.
  • Programming languages: Polyglot projects where developers switch between languages per layer (SQL for queries, Python for orchestration, JavaScript for UI). Analogous structurally even if the user group is a single brain.
  • Computational linguistics and NLP: Modern multilingual language models (XLM-R)[6] implicitly model code-switched data; code-switching is now a challenge and feature in multilingually trained models.

Clarity

Names the fact that multilingual and multivocal speakers are not using one code at a time but selecting dynamically from a shared repertoire. This clarity pushes back against the monolingual default that treats code-switching as confused or improper. Code-switching is a skilled practice with its own grammar and its own sociolinguistic norms. Recognizing it lets analysts describe what speakers are doing rather than mislabel competent multilinguals as incompetent monolinguals. The semantic shift from "error" to "competence marker" reframes the entire analytic landscape.

Manages Complexity

Lets speakers address heterogeneous audiences, signal nuanced affiliations, and navigate situations where no single code would suffice. Instead of inventing a master-code that covers every need, the speaker maintains multiple codes and switches as situations require. This economy of multiple tools is central to cognitive sociolinguistics: the mind holds several grammars and selects dynamically, rather than forcing all communicative work through a single system. Code-switching is a solution to the problem of variation — how to remain intelligible to audiences with different linguistic competencies and social identities.

Abstract Reasoning

Generalizes to any competence that requires holding multiple rule-sets and alternating between them contextually: a polyglot programmer choosing languages per problem, a diplomat shifting protocols among venues, a skilled leader shifting management styles by team, a cross-functional employee switching vocabulary between engineering and legal conversations. The prime trains the analyst to recognize that the "code" chosen is itself a communicative act. Beyond linguistics, code-switching models code-level decision-making in any domain requiring multiple registers.

Knowledge Transfer

Domain Code A Code B Switch trigger Syntactic constraint
Bilingual household Heritage language Dominant language Addressee, topic, identity signal Matrix language frame (heritage base or dominant base)
Clinical Medical jargon Patient-appropriate lay English Audience competence / rapport Equivalence constraint (switching at phrase boundary to avoid morphological mismatch)
Corporate meeting Engineering speak Finance / marketing speak Stakeholder present Code-switching must preserve syntactic well-formedness in both codes
Social media Formal register Platform slang / emoji Platform context / persona Platform-specific grammar (emoji, hashtag syntax) as additional codes
Programming Python SQL / TypeScript Task domain Language-level syntax must be parseable (switching at API boundary or module boundary)
Academia Technical paper language Conference Q&A informality Setting, audience cognitive load Disciplinary jargon vs. accessible exposition; switching at clause boundary
Diplomacy English lingua franca Local language courtesies Signaling respect / solidarity Code-switching to local language preserves ceremonial protocols

Crossing these rows, practitioners internalize the cost-benefit: maintain multiple competencies (training cost) to earn multi-context reach (communicative benefit). The prime makes explicit the skill that multilingual humans have exercised for millennia. Syntactic constraints are universal; social triggers are context-dependent.

Examples

Formal/Abstract Example: Poplack's Spanish-English Switching

Poplack (1980)[4] documented Spanish-English code-switching in Puerto Rican communities in New York, establishing the foundational empirical pattern. A classic example: "Sometimes I'll start a sentence in English y termino en español." The switch from English to Spanish ("y termino en español" — "and I finish in Spanish") occurs at a clause boundary, respecting what Poplack calls the equivalence constraint: switches occur at points where the grammars of both codes align, permitting surface-level grammatical well-formedness in both. The speaker does not violate either language's morpho-syntactic rules at the switch point. This single example demonstrated that code-switching is not random mixing but governed by grammatical rules.

Mapped back: Poplack's empirical work shows that code-switching is rule-governed. A speaker cannot insert Spanish morphology into an English clause-final position arbitrarily; the switch must respect the syntactic boundaries where both codes "agree." This finding demolished the deficit hypothesis (code-switching as sign of incomplete bilingual competence) and established code-switching as linguistic competence sui generis — a skill requiring fluency in both codes and knowledge of their permissible interaction points.

Applied/Industry Example: Multilingual Customer Service and NLP Models

Modern bilingual customer service centers (e.g., Spanish-English teams in U.S. healthcare, tech support, or banking) operationalize code-switching daily. A customer-service representative might open in English (corporate standard): "Thank you for calling technical support." The customer responds in Spanish ("Hablo español solamente"). The representative code-switches: "No problema. ¿Cuál es el problema técnico?" — then code-switches back to English for a technical term: "Is it a connectivity issue or a device issue?" The switches mark audience accommodation, topic boundaries, and domain expertise.

Simultaneously, modern multilingual NLP models (XLM-R, mBERT)[7] are trained on multilingual corpora that include code-switched data. These models implicitly learn the code-switching patterns of their training data, enabling them to process Spanish-English Tweets, multilingual Slack channels, and diverse global content without explicit code-switching annotation. The system detects illocutionary intent across code boundaries — a request in Spanish followed by a specification in English is still one communicative act. This industrial deployment of code-switching shows how linguistic theory scales into production systems serving multilingual populations.

Mapped back: Code-switching in customer service demonstrates that the practice solves real communicative problems: how to serve audiences with heterogeneous linguistic competencies. Modern NLP models operationalize code-switching as an implicit feature, not a bug. The same three-layer analysis applies: parse the utterance across code boundaries, detect intent regardless of code, and respond appropriately. Code-switching is now an engineering requirement for systems serving multilingual populations.

Structural Tensions

T1 — Syntactic constraints (Poplack) vs. Matrix Language Frame (Myers-Scotton). Poplack's equivalence constraint and free-morpheme constraint suggest that switches can occur only at points where both codes' grammars align (clause boundaries, between free morphemes, never interrupting a word). Myers-Scotton's Matrix Language Frame model proposes that one code (the matrix language) provides the morpho-syntactic frame, and the other code (the embedded language) contributes content morphemes within that frame. These accounts make different predictions. The tension is empirical and theoretical: do real code-switchers follow Poplack's constraints (suggesting universal syntactic principles), or do they follow Myers-Scotton's asymmetric frame (suggesting a psycholinguistic hierarchy)? Multilingual speech often violates both; neither model fully captures the variation. Modern reconciliations suggest that both operate — sometimes code-switching is governed by equivalence, sometimes by frame asymmetry — and context determines which.

T2 — Situational vs. metaphorical switching (Gumperz socio-pragmatic functions). Gumperz (1982)[2] distinguished situational switching (a change in interlocutors or setting triggers a code switch: office context → Spanish; home context → English) from metaphorical switching (a switch within the same situation to invoke associations of a code without a situational trigger: speaking English but switching to Spanish to invoke in-group solidarity). The tension is that this distinction is often blurry. A single utterance can be simultaneously situational (the customer prefers Spanish) and metaphorical (the switch signals solidarity). Empirically, speakers rarely code-switch only for situational or only for metaphorical reasons; the two motivations intertwine. Modern sociolinguistics treats the distinction as analytic rather than categorical — a useful taxonomy that admits real-world blending.

T3 — Prestige, diglossia, and code-choice hierarchies. Code-switching is not neutral; codes carry prestige associations. In many multilingual communities, the dominant language (English in the U.S.) carries prestige, while heritage languages (Spanish, Mandarin) carry identity but lower prestige in institutional contexts. This creates a diglossic hierarchy: speakers code-switch toward prestige codes in formal contexts, toward identity codes in intimate contexts. [8][9]Typological studies of language contact show how code-switching operates within broader patterns of contact-induced language change. The tension is that code-switching can either reinforce these hierarchies (speakers always switch to the prestige language in high-register situations) or subvert them (speakers deliberately switch to heritage languages to claim space and dignity in institutional settings). Code-switching is simultaneously a symptom of language subordination and a form of linguistic resistance.

T4 — Child bilingual acquisition and the mixing question. Young bilingual children often produce utterances that mix codes — is this code-switching (competent practice) or confusion? Linguists now recognize that children are developing code-switching competence; early mixing is not error but scaffolding toward adult norms. The tension is that it's hard to distinguish "the child hasn't learned the code boundary yet" from "the child is developing the skill to strategically blur it." Modern developmental sociolinguistics treats child code-mixing as a window into language acquisition — children learning to code-switch are learning to maintain multiple grammars simultaneously.

T5 — Code-switched data in NLP training and evaluation. As multilingual NLP models scale, they encounter code-switched data (Tweets from multilingual users, diverse diaspora communities, bilingual forums). The tension is methodological: should NLP systems be trained on code-switched data to handle real-world multilingual text, or is code-switched data a corruption to be cleaned? Modern practice treats code-switched data as a legitimate test case — systems must handle code-switching robustly. But evaluation metrics don't always capture this; a model may perform well on monolingual text and fail on code-switched boundaries. This creates tension between accuracy on "standard" text and coverage of real-world linguistic variation.

T6 — Deficit framing vs. asset framing in bilingual education. Historically, code-switching was seen as a problem in bilingual education — teachers attempted to suppress it, treating mixing as a sign of incomplete bilingualism. Modern asset-based frameworks treat code-switching as a legitimate communicative skill and bilingual resource. [3]Contemporary translingual practice scholarship emphasizes that multilingual speakers deploy the full range of their linguistic resources strategically and skillfully, rejecting the notion that they are deficient monolinguals. The tension is institutional: schools still often enforce monolingual norms (English-only policies, Spanish-only classrooms), while sociolinguistic research celebrates code-switching as evidence of bilingual sophistication. Bridging this gap requires retraining educators to recognize code-switching as competence, not deficit. The debate continues in policy: do bilingual programs celebrate code-switching as a skill, or do they treat it as a transitional tool to be replaced by monolingual dominance?

Structural–Framed Character

Code-Switching is a hybrid on the structural–framed spectrum, and the frame it carries is substantial even though a structural core exists. Part of it is a bare pattern — alternating between distinct systems within a single stream; part of it is a vocabulary and set of assumptions inherited from linguistics and semiotics.

The structural seed is simple: a single agent deploys two or more distinct codes, switching between them within one episode. But the concept as it actually travels is thick with its home field's commitments — the codes are languages, dialects, registers, and jargons; the switches are driven by social, pragmatic, and expressive goals like signaling group identity, accommodating an interlocutor, or quoting; and the whole repertoire is treated as an active social resource. Those terms and motives come from the study of human language in social context, so applying the idea to bilingual speech, to register-shifting in the workplace, or to stylistic framing means importing that sociolinguistic lens rather than merely spotting an alternation. Because the interpretive frame carries most of the meaning while a thin relational pattern underlies it, it sits toward the framed side of the middle.

Substrate Independence

Code-Switching is a narrowly substrate-independent prime — composite 2 / 5 on the substrate-independence scale. It originates and primarily lives in linguistic and semiotic substrates, rooted in speaker competence, social signaling, and pragmatic goals. A metaphorical application exists in computing, where one switches between programming languages or protocols, but the phenomenon is fundamentally about language. Transfer to non-linguistic domains is weak, leaving it tethered to the multilingual contexts it came from.

  • Composite substrate independence — 2 / 5
  • Domain breadth — 2 / 5
  • Structural abstraction — 3 / 5
  • Transfer evidence — 1 / 5

Relationships to Other Abstractions

Local relationship map for Code-SwitchingParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.Code-SwitchingPRIMEPrime abstraction: Contextual Mode Switching — is a kind ofContextualMode SwitchingPRIMEPrime abstraction: Remapping — is a kind of, typicalRemappingPRIMEDomain-specific abstraction: Intercultural Communication — is a decomposition of, conditionalInterculturalCommunicationDOMAIN

Current abstraction Code-Switching Prime

Parents (2) — more general patterns this builds on

  • Code-Switching is a kind of Contextual Mode Switching Prime

    Code-switching is a specialization of contextual mode switching in which the alternation is between distinct linguistic codes triggered by social and pragmatic cues.

  • Code-Switching is a kind of, typical Remapping Prime

    Code-Switching is typically a specialization of Remapping, retaining the parent's defining structure while adding the child's specific commitments.

Children (1) — more specific cases that build on this

  • Intercultural Communication Domain-specific is a decomposition of, conditional Code-Switching

    Multilingual participants may repair or manage a cultural gap by switching deliberately among linguistic codes in response to audience and discourse goals.

Hierarchy paths (3) — routes to 3 parentless roots

Neighborhood in Abstraction Space

Code-Switching sits in a moderately populated region (45th percentile for distinctiveness): it has near-neighbors but no dense thicket of synonyms.

Family — Meaning, Signs & Epistemic Stance (31 primes)

Nearest neighbors

Computed from structural-signature embeddings · 2026-07-26

Not to Be Confused With

Code-Switching must be distinguished from Contextual Mode Switching, its nearest neighbor (similarity 0.703), on the basis of scope and substitutability. Contextual Mode Switching is the broader category of adjusting one's mode, intensity, or register in response to context—it encompasses multiple strategies for adapting communication to situation. Mode switching might mean shifting from casual to formal tone, from individual-focused to group-focused attention, from task-oriented to relationship-oriented language, all within a single underlying code or system. Code-switching is the specific instance where the speaker switches between completely distinct linguistic codes—languages, dialects, jargons that have different morphologies, phonologies, and grammars. A speaker can perform contextual mode switching (shift from casual to formal English within English) without code-switching; or can code-switch (Spanish to English) while maintaining a consistent mode (formal to formal, or casual to casual). The distinction matters because mode-switching operates along a continuum within a code (more formal ↔ less formal); code-switching is discrete (you are in one code or another, though you can be at the boundary). A multilingual speaker at a formal business meeting might mode-switch from officecolor to team-talk but remain in English; the same speaker might code-switch to Spanish when greeting a new colleague who prefers it, either formally or casually. Mode is a property of register and style within a code; code is the underlying linguistic system.

Code-Switching is further distinct from Register or Style Shifting, though they are often conflated. Register shifting is the adjustment of formality, vocabulary, and tone along a spectrum within a single language (standard English vs. slang English; legal jargon vs. conversational English). The underlying code remains constant; the register selected from within that code varies. Style shifting is similar: the same language is adapted to context (more flowery vs. more direct; more technical vs. more accessible). Register shifting is smooth and continuous; code-switching is discrete and bounded. A speaker shifting registers in English might gradually shift from formal to casual, with imperceptible boundaries; a speaker code-switching between Spanish and English makes discrete jumps between the codes. A speaker might shift register within English (academic → conversational) or simultaneously code-switch and shift register (academic Spanish → casual English), but the mechanisms are different. Register is a dimension within a code; code is the choice of system itself.

Code-Switching is also distinct from Speech Act Theory and the study of illocution and perlocution (the intended effects and achieved effects of utterances). Speech act theory is about what a speaker does with language—a request, a promise, a threat, a command. The speaker performs an action through the utterance. Code-switching is about which language the speaker uses to perform that action. A request can be performed in Spanish or English or Spanglish; the request is the speech act, and the code choice is independent of it. A Spanish speaker requesting something can stay in Spanish (monolingual request), switch to English (code-switch request), or mix codes within the request (code-switched request). The speech act and the code choice are orthogonal dimensions of language use. Speech act theory analyzes the pragmatic force of utterances; code-switching analyzes the linguistic-code choice and its social/communicative functions.

Finally, Code-Switching is not Metasystem Transition, which is a higher-level phenomenon in which an entire system transitions from one operating logic to another (a economy from planned to market, a politics from autocracy to democracy, a language from synthetic morphology to analytic morphology). Metasystem transition is about the system itself changing its rules, not about an agent choosing between available systems. Code-switching assumes the multilingual speaker has multiple codes available simultaneously and selects between them; metasystem transition is about the system itself changing, affecting all participants. A multilingual person code-switches because they possess multiple languages; a language undergoing metasystem transition (synthetic → analytic morphology) is fundamentally changing what it is, not selecting from options. Code-switching is individual choice within stable codes; metasystem transition is collective evolution of a code itself.

Solution Archetypes

Solution archetypes in the catalog that build on this prime — directly (this prime is a source ingredient) or as a related prime.

Built directly on this prime (1)

  • Code / Register Adaptation: Adapt language, code, and formality to the audience and context without losing meaning or excluding others.

Also a related prime in 8 archetypes

  • Context-Keyed Representation Switching: Maintain several context-specific representations on one substrate, activate the right one from validated context cues, isolate inactive maps from interference, and preserve them for reliable re-entry.
  • Contextual Mode-Switching Protocol: Switch communication or operating mode deliberately when the current context, role, risk, phase, or audience no longer fits the active style of interaction.
  • Contextual Selective Propagation: When a meaning changes in one context, decide where that changed meaning should travel, where it should be translated, and where it should remain bounded.
  • Cross-Language Constraint Check: Check whether communication, interface, policy, or category assumptions survive movement across languages and communities before treating translation or localization as complete.
  • Face-Saving Directness Calibration: Balance clarity with relationship preservation when delivering requests, criticism, refusals, warnings, or conflict.
  • Iconographic Meaning System: Use a coherent set of symbols to communicate meaning quickly and consistently across contexts.
  • Pragmatic Context Alignment: Make intended meaning dependably recoverable by designing the context around the words, not only the words themselves.
  • Symbol-System Coherence in Visual Art: Keep the same visual sign pointing to the same intended meaning unless the work makes the change visible and purposeful.

References

[1] Gardner-Chloros, P. (2009). Code-switching. Cambridge University Press. Comprehensive sociolinguistic synthesis of how, when, and where code-switching occurs across communities and genres, treating it as alternation between distinct codes within a single interaction. Supports the core definition of alternating between distinct codes within a conversation/utterance/sentence (060). ↩

[2] Gumperz, J. J. (1982). Discourse Strategies. Cambridge University Press. Founding sociolinguistic account treating code-switching as a contextualization cue, and distinguishing situational from metaphorical switching driven by social and pragmatic goals. Supports the social/pragmatic/expressive-goals claim (061), the foundational-establishment claim (064), and the situational-vs-metaphorical tension (071). ↩

[3] Canagarajah, S. (2013). Translingual Practice: Global Englishes and Cosmopolitan Relations. Routledge. Develops translingual/codemeshing practice in which speakers deploy their full linguistic repertoire as an active resource rather than treating one code as default, and argues for an asset (not deficit) framing. Supports the full-repertoire-as-resource claim (062) and the asset-vs-deficit framing in bilingual education (073). ↩

[4] Poplack, S. (1980). "Sometimes I'll start a sentence in Spanish y termino en espanol: toward a typology of code-switching." Linguistics, 18(7-8), 581-618. Establishes the equivalence constraint and free-morpheme constraint from a Puerto Rican Spanish-English corpus, showing code-switching is rule-governed and occurs where both grammars align. Supports the rule-governed/morpho-syntactic-constraint claim (063, 066) and the formal Spanish-English example (069). ↩

[5] Myers-Scotton, C. (1993). Duelling Languages: Grammatical Structure in Codeswitching. Clarendon Press. Presents the Matrix Language Frame model: one code (the matrix language) supplies the morpho-syntactic frame into which embedded-language content morphemes are inserted. Supports the matrix-language-frame claims (065, 067). ↩

[6] Conneau, A., Khandelwal, K., Goyal, N., Chaudhary, V., Wenzek, G., Guzman, F., Grave, E., Ott, M., Zettlemoyer, L., & Stoyanov, V. (2020). "Unsupervised Cross-lingual Representation Learning at Scale." In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (pp. 8440-8451). Introduces XLM-R, a multilingual masked-language model trained on 100 languages of CommonCrawl that implicitly ingests code-switched data. Supports the claim that modern multilingual LMs (XLM-R) model code-switched data (068). ↩

[7] Winata, G. I., Cahyawijaya, S., Liu, Z., Lin, Z., Madotto, A., & Fung, P. (2021). "Are Multilingual Models Effective in Code-Switching?" In Proceedings of the Fifth Workshop on Computational Approaches to Linguistic Code-Switching (pp. 142-153). ACL. Directly evaluates mBERT and XLM-R on code-switching tasks, treating code-switched data as a legitimate test case for multilingually pretrained models. Supports the claim that modern multilingual NLP models (XLM-R, mBERT) are applied to / trained on code-switched data (070). ↩

[8] Heller, M. (Ed.). (1988). Codeswitching: Anthropological and Sociolinguistic Perspectives. Mouton de Gruyter. Edited collection on the social meaning of code choice, including the political economy of code choice and prestige/diglossic hierarchies in multilingual communities. Supports the prestige/diglossia/code-choice-hierarchy claim (072). ↩

[9] Thomason, S. G. (2001). Language Contact: An Introduction. Edinburgh University Press. Typological introduction to contact-induced change, borrowing, and bilingual mixed languages, situating code-switching within broader patterns of language contact. Supports the typological-language-contact framing (074). ↩

[10] Muysken, P. (2000). Bilingual Speech: A Typology of Code-Mixing. Cambridge University Press. Three-way typology of intrasentential code-mixing (insertion, alternation, congruent lexicalization). Bibliography-only (cited in Notes provenance).

[11] Auer, P. (Ed.). (1998). Code-Switching in Conversation: Language, Interaction and Identity. Routledge. Conversation-analytic approaches to code-switching, turn-taking, and interactional identity. Bibliography-only.

[12] Zentella, A. C. (1997). Growing Up Bilingual: Puerto Rican Children in New York. Blackwell. Thirteen-year ethnographic study of Spanish-English bilingual acquisition and code-switching among Puerto Rican children in New York. Bibliography-only.

[13] Bullock, B. E., & Toribio, A. J. (Eds.). (2009). The Cambridge Handbook of Linguistic Code-Switching. Cambridge University Press. Comprehensive reference handbook on linguistic and sociolinguistic dimensions of code-switching. Bibliography-only.

[14] MacSwan, J. (1999). A Minimalist Approach to Intrasentential Code Switching. Garland Press (Outstanding Dissertations in Linguistics). Generative/minimalist syntactic account of intrasentential code-switching, grounded in Spanish-Nahuatl data. Bibliography-only. (Original def placed the publisher in the title slot; corrected.)

[15] Solorio, T., & Liu, Y. (2008). "Learning to Predict Code-Switching Points." In Proceedings of the 2008 Conference on Empirical Methods in Natural Language Processing (pp. 973-981). ACL. Early NLP work on predicting code-switch points in Spanish-English speech using surface and syntactic features. Bibliography-only after re-sourcing marker 070. (Original def – "Learning to Identify Languages Automatically, Computational Linguistics 34(4):589-612" – did not match any real Solorio paper and was corrected to the verifiable 2008 EMNLP paper.)

[16] Backus, A. (2005). "Codeswitching and language change: One thing leads to another?" International Journal of Bilingualism, 9(3-4), 307-340. Usage-based account linking code-switching to contact-induced language change. Bibliography-only. (Original def – title "Towards a Typology of Codemixing," IJB 5(2):105-125, 2014 – carried an internally inconsistent volume/year; replaced with the verifiable Backus IJB article on code-switching and language change.)