Cross-linguistic influence is the effect of a language a person already knows on how that person perceives, learns, understands, and produces another language. The first language (L1) supplies established sound categories, word associations, grammatical expectations, and ways of allocating attention; these resources can accelerate new learning where languages overlap and produce systematic learning challenges where they diverge. The same learner may experience facilitation in one feature and interference in another. The central scientific question is how existing linguistic experience interacts with new input, practice, and the developing representations of a second or additional language (L2/L3). Odlin (2003); Jarvis (2000).
A familiar example occurs when someone learning English interprets a new vowel through the vowel categories already established in their first language. Another occurs when a learner recognizes a related word immediately because its form and meaning resemble a known word. A third arises when the learner applies a familiar word order to a new language whose syntax licenses a different pattern. Each example involves prior learning shaping attention and interpretation; the outcome depends on the languages, the specific linguistic construction, the task, and the learner's history. Amengual (2023); Muylle, Hartsuiker, and Bernolet (2023).
What Is Cross-Linguistic Influence?
Cross-linguistic influence (CLI), often called language transfer, is a relationship between linguistic systems within a learner's experience. In second-language acquisition research it includes the effects of L1 knowledge on L2 learning, but the broader concept also includes L2 effects on L1 and interactions involving third, heritage, and additional languages. It may operate during initial encoding, online comprehension, planning an utterance, remembering a word, monitoring an error, or restructuring a developing grammar. Transfer can concern linguistic representations, moment-to-moment processing, or measurable behavior. These levels must be distinguished because similar observable outcomes can arise through different mechanisms. Odlin (2012); Kroll et al. (2012).
L1 refers to a language learned early enough to become an established linguistic system; L2 identifies a language acquired later or in an additional-language context. Those labels describe acquisition histories, not intrinsic ranks of ability. A person's strongest everyday language may change with migration, education, or use. For simultaneous bilinguals, a single directional L1-to-L2 model may fit poorly. Researchers therefore specify age of acquisition, dominance, exposure, literacy, proficiency, and the actual languages involved rather than assuming that a label completely characterizes the learner.
Transfer, interference, facilitation, and cross-language activation
Positive transfer or facilitation occurs when prior language knowledge helps process or acquire a form in the new language. Shared cognates, comparable syntactic configurations, familiar stress patterns, and transferable literacy practices can provide useful starting points. Negative transfer or interference occurs when a previously learned distinction or habit competes with the target-language pattern: learners may miss a phonetic contrast, favor a familiar grammatical option, or assign the wrong meaning to a look-alike word. These names describe a particular result relative to a target task; they do not classify a language or a person. Odlin (2003).
Cross-language activation means that information associated with more than one known language becomes available during processing. Such activation can occur even when only one language is intended. Behavioral evidence from word recognition and structural priming supports interactions between linguistic representations, while models disagree about when activation is restricted, how selection works, and how much is shared across languages. Activation by itself need not produce an error or durable learning. Dijkstra and van Heuven (2002); Hartsuiker et al. (2016).
A descriptive example is not proof of transfer
Suppose speakers of two different first languages make different errors in English articles. The difference is a possible transfer signal, but it could also reflect proficiency, the tasks used, unequal teaching, input frequency, writing experience, or what the learners attended to. An inference of L1 influence becomes more persuasive when researchers establish similarities within an L1 group, differences across L1 groups, and direct links between the behavior and the relevant L1 feature, while controlling plausible alternative explanations. Jarvis's methodological framework treats those converging comparisons as essential to a defensible transfer account. Jarvis (2000).
How a First Language Shapes New Language Learning
Learning a new language starts with an experienced perceptual and cognitive system. Repeated exposure to L1 speech establishes expectations about which acoustic variations distinguish words. Familiar words create networks of forms and meanings. Grammatical experience establishes sensitivity to frequent sequences, cues, and constructions. Writing experience develops particular relationships between symbols and sounds. These systems influence where attention goes and which interpretation initially seems likely when ambiguous L2 input arrives. Ellis (2006); Flege and Bohn (2021).
A learner may initially reuse a familiar category because it efficiently explains incoming data. As additional examples accumulate, competing categories can become more differentiated, connections between forms and meanings can be reorganized, and previously underweighted cues can receive greater attention. The speed and completeness of such changes vary by feature and learner. Transfer is consequently a dynamic aspect of learning rather than a fixed list of first-language mistakes.
Representation, attention, and prediction
Representational explanations emphasize how established sound categories, lexical entries, or grammatical structures overlap with those needed in L2. Attention-based explanations emphasize which cues learners notice and rely upon. Processing explanations emphasize the rapid activation and selection of candidates under time pressure. Prediction-based accounts ask how context and learned expectations shape anticipation of words or structures. These are partially compatible levels of explanation, but an experiment demonstrating a prediction effect does not automatically establish a permanent change in grammatical representation. Ellis (2006); Muylle, Hartsuiker, and Bernolet (2023).
Similarity can help and mislead
A familiar-looking form encourages rapid identification, yet surface resemblance does not guarantee equivalent meaning, pronunciation, usage, or syntax. A shared structural pattern may ease early learning, while near-overlap may encourage a learner to overlook a distinction that matters in the new language. Similarity therefore needs to be specified: shared spelling, related phonemes, analogous grammatical roles, distributional overlap, and learner-perceived resemblance are separable measures. A language pair cannot be labeled globally easy or difficult from family membership alone. Schepens, van der Slik, and van Hout (2016).
Phonological Transfer: Hearing and Producing New Sounds
One of the strongest domains of CLI involves speech perception. Infants and children acquire increasingly language-specific phonetic expectations through exposure. Adults learning additional languages approach unfamiliar sounds through previously organized perceptual spaces. When an L2 contrast maps onto a single familiar category, learners may initially have trouble recognizing the distinction; when the contrast corresponds to established categories, recognition may be easier. Individual results depend on the exact acoustic dimensions, listening conditions, learning history, and phonological contexts. Amengual (2023); Flege, Aoyama, and Bohn (2021).
Sound-category assimilation and perceptual restructuring
A learner need not be deaf to an acoustic difference to have difficulty using it as a reliable linguistic cue. Their L1 may have taught them that variation along that dimension is irrelevant to distinguishing words. For a new L2 contrast, the learner must learn both to perceive the relevant difference and to associate it consistently with words and contexts. The revised Speech Learning Model explains L2 phonetic learning through the evolving relationship between existing L1 categories and emerging L2 categories, emphasizing their perceived proximity and the quality of exposure. This is a theoretical model with productive empirical applications rather than a formula that predicts every learner's accent. Flege and Bohn (2021).
In English learning, contrasts conventionally described through /r/ and /l/ or through particular vowel qualities are useful research examples because their perceptual consequences vary across L1 systems. Their difficulty is never universal to all learners from an entire nation or language community. Dialects, phonetic contexts, prior exposure, and the specific acoustic cues used by a listener affect performance. Work on Japanese-English liquids illustrates how exposure can influence recognition and production, while cautioning against treating a single contrast as representative of all speech learning. Flege, Aoyama, and Bohn (2021).
Phonotactics and syllable structure
Languages differ in which sound sequences can occur at syllable boundaries. If an L2 allows consonant clusters unfamiliar in L1, a learner may insert a vowel, delete a segment, or reinterpret the syllable boundary in speech or listening. Such patterns can reflect perceptual parsing, articulatory planning, phonological constraints, or a combination. Careful research compares perception with production because a speaker may hear a distinction accurately yet find it difficult to articulate under fluent conditions, or may imitate a sound without consistently recognizing lexical contrasts.
Stress, tone, rhythm, and intonation
Transfer operates beyond individual segments. Learners carry expectations about where stress falls, whether pitch contrasts distinguish lexical items, how rhythm groups words, and which contours communicate questions, focus, or stance. An L1 in which pitch functions primarily as sentence-level intonation may support a different learning trajectory for a lexical tone language than an L1 in which lexical pitch contrasts are central. Such comparisons require precise description of the dialect and task; broad categories like tonal and non-tonal conceal substantial variation. Amengual (2023).
Perception and production are related but separable
Second-language speech studies distinguish identification accuracy, discrimination sensitivity, acoustic production, perceived comprehensibility, and accent ratings. A learner can improve intelligibility without eliminating every trace of an accent, and listener judgments depend on rater familiarity and the speech sample. A review and two meta-analyses by Saito emphasize that segmental, prosodic, and temporal variables contribute differently to comprehensibility and perceived nativelikeness. Educational objectives therefore should define intelligible communication rather than assume a single ideal phonetic endpoint. Saito (2021).
Lexical Transfer: Vocabulary, Cognates, and Meaning
A learner's existing word knowledge often supplies a powerful route into another language. Related words can share spelling, sound, roots, or meaning. Recognizing a form such as English information alongside a historically related word in another language may speed learning, while a word pair that looks alike but differs in meaning can encourage a wrong interpretation. These mechanisms involve both linguistic form and semantic mapping. Their strength changes with proficiency, script, frequency, sentence context, and experimental task. Dijkstra and van Heuven (2002).
Cognates and shared lexical structure
Cognates are words with a shared historical origin; in psycholinguistic experiments researchers may also group words by form and meaning overlap, and those operational definitions should be reported explicitly. Cognate facilitation can emerge because related forms activate interconnected lexical candidates or because a familiar mapping reduces the amount of new form learning required. The advantage is conditional, not a guarantee that every cognate will be recognized quickly. Cross-language competition, context, and a learner's script experience can alter the outcome.
False friends and semantic overextension
False friends are words whose resemblance across languages invites a meaning that is inappropriate in the target context. Learning that a look-alike word has an independent meaning involves modifying an initial hypothesis, strengthening the L2 meaning, and noticing contexts in which the two words diverge. A closely related problem occurs when a word's semantic range is broader or narrower than its nearest translation. Learners may overextend familiar categories, choose awkward collocations, or fail to recognize a pragmatic restriction even when the dictionary translation is correct. Transfer at the lexical level concerns meaning, use, and associations as well as isolated word forms.
The bilingual lexicon and nonselective access
The BIA+ model of bilingual word recognition proposes interacting orthographic, phonological, and semantic information along with mechanisms associated with task decisions. It provides a structured explanation for why knowledge from another language can affect target-word recognition, especially with overlapping forms. The model does not claim that all language information is equally active in every circumstance, and it does not describe every spoken, signed, or multilingual situation. Dijkstra and van Heuven (2002).
When learners encounter unfamiliar L2 vocabulary, the current goal and input conditions affect which prior-language candidates become useful. Translation practice, recognition, spontaneous naming, listening under noise, and reading in connected sentences need not give identical transfer patterns. A comprehensive account separates lexical activation from response selection and distinguishes faster performance on a familiar word from durable learning of a new lexical representation.
Derivational morphology and word families
Shared prefixes, suffixes, roots, and morphological patterns can support vocabulary learning. Yet languages differ in which affixes are productive, where stress shifts, which meanings a derivation licenses, and how frequently a construction occurs. Learners may perceive a familiar-looking word family and infer an English word that speakers rarely use. Teaching can connect useful correspondences to real usage, pronunciation, frequency, and context instead of treating every plausible word formation as equally natural.
Grammatical Transfer: Syntax, Morphology, and Interpretation
First-language syntax provides learned expectations about relationships among words, arguments, and sentence structures. Transfer may appear in word order, determiner use, agreement, tense and aspect, case marking, relative clauses, and interpretation of ambiguous sentences. The relevant variables include which constructions exist in each language, how the same meaning is expressed, the distribution of competing alternatives, and how much experience the learner has with L2 input. Comparisons at the construction level are more explanatory than statements that one entire grammar is simple or complex. Muylle, Hartsuiker, and Bernolet (2023); DeKeyser (2005).
Word order and argument structure
If an L1 uses a strongly preferred sequence of subject, verb, and object, the learner may initially parse an L2 sentence using that expectation. But surface word order interacts with case marking, agreement, discourse prominence, and prosody. A learner's interpretation of a sentence can reflect a conflict between familiar order and target-language morphological cues. Investigators therefore examine grammaticality judgments, reading times, picture matching, and spontaneous production rather than assuming that an error identifies a single transferred rule.
Articles, tense, agreement, and grammatical cues
A learner whose L1 lacks a directly comparable article system may need to learn distinctions involving definiteness, discourse familiarity, and reference. Yet missing articles in an L2 writing sample cannot automatically be explained by L1 absence: overall exposure, automaticity, discourse context, and task demands matter. Similarly, languages distribute tense and aspect meanings differently across morphology, adverbs, and context, so familiar conceptual categories may align only partly with available L2 forms. Research on grammatical difficulty identifies salience, complexity, and the form-meaning relationship among relevant influences. DeKeyser (2005).
Cross-language structural priming
Structural priming occurs when encountering or producing a construction changes the probability of using a related construction shortly afterward. In bilingual experiments, a structure presented in one language can facilitate selection of a corresponding structure in another. Such results are evidence that some representations are linked or shared across languages, though the precise interpretation depends on structural equivalence, lexical overlap, proficiency, task design, and the direction of priming. Four multilingual experiments reported substantial within- and between-language priming, supporting a shared-syntax account for the tested materials. Hartsuiker et al. (2016).
Priming can also reveal limits and asymmetries. In research on L2 production, cross-language priming did not simply increase every L1-like word order: an experimental comparison found facilitation for certain grammatical shared orders alongside suppression of incompatible L1-based alternatives. Thus activation is constrained by the learner's knowledge of what the L2 permits. These laboratory changes should not be equated automatically with long-term acquisition. Hopp and Jackson (2023); Shin and Christianson (2012).
Shared syntax and competing developmental theories
Accounts of bilingual syntax differ in whether they predict strongly distinct early representations that grow more integrated, highly shared representations from the beginning, or representation-specific paths that depend on language combination and experience. Cross-linguistic priming contributes to this debate, but a priming effect can reflect temporary activation, learned associations, or more enduring representational overlap. Muylle and colleagues' 2023 synthesis reviews proficiency, word order, case marking, and linguistic distance as conditions that matter for interpreting experimental findings. Muylle, Hartsuiker, and Bernolet (2023).
Transfer in Sentence Comprehension and Real-Time Processing
A learner's L1 can shape how a sentence is interpreted before its final word is heard or read. Listeners and readers integrate lexical plausibility, grammatical marking, discourse information, and expectations about structure. When an L2 uses a less familiar cue, the learner may initially prioritize the one that has been most reliable in L1. The consequence can be a processing slowdown, a temporary misinterpretation, or a tendency to choose one attachment or referent. Online measurements often reveal cross-language effects that are missed by untimed judgments. Ellis (2006); Muylle, Hartsuiker, and Bernolet (2023).
A single reading-time difference is consistent with several explanations, including unfamiliar vocabulary, decoding demands, grammatical prediction, or strategic caution. Strong studies manipulate the cue of interest while matching lexical frequency and sentence plausibility, use multiple measures, and compare how patterns change with proficiency. Processing evidence specifies how a language is used in time; longitudinal learning evidence is required to show how the underlying competence changes.
Prosody and discourse-level expectations
L1 experience also shapes how pauses, pitch accents, intonation, and information structure help identify intended meanings. Some languages rely more heavily on particular grammatical markers, while others distribute equivalent discourse functions across word order, prosody, or contextual inference. Cross-language effects in comprehension can therefore arise even when a sentence has a grammatically correct L2 form. The interpretation depends on context and the interaction of cues, not merely a word-by-word translation.
Pragmatic Transfer and Sociolinguistic Context
Pragmatic transfer concerns how prior experience with communicative conventions influences the interpretation and production of requests, apologies, politeness, indirect meanings, address forms, or discourse organization in another language. Successful speech acts depend on linguistic choices and shared social expectations. Learners may select an expression that is grammatically accurate yet unusually direct, formal, distant, or familiar in a particular setting. A culturally common practice in one community should never be portrayed as a universal property of all its speakers.
Investigating pragmatic transfer requires observing comparable situations, interlocutor relationships, institutional expectations, and learner intentions. Differences in language production may result from identity, deliberate style, unequal access to social settings, or confidence as well as transferred conventions. Unlike a laboratory lexical-decision effect, a pragmatic judgment is often negotiated between speakers whose expectations vary.
Reading and Writing Systems
A learner with literacy in one language already understands that visible marks can encode linguistic meaning, yet orthographic systems differ substantially in how script units relate to sound and morphology. Familiarity with an alphabetic script can ease aspects of another alphabetic writing system, while spelling-to-sound correspondences may be misleading. Transfers can include letter-sound expectations, segmentation practices, punctuation habits, morphological spelling patterns, and ways of organizing written discourse. These are literacy-related processes, and a reader's previous educational opportunities matter alongside L1 phonology.
Reading difficulty in an additional language is not by itself evidence of dyslexia or a language disorder. Researchers and educators should examine instruction, print exposure, spoken-language proficiency, the writing systems involved, and performance in the languages a learner knows. Clinical differential diagnosis and disability identification are separate topics requiring appropriate multilingual assessment; they are outside the present mechanism-focused account.
Positive Transfer, Negative Transfer, and Avoidance
Transfer may be visible as faster recognition or greater accuracy, but it may also affect which forms learners avoid. If a structure requires unfamiliar mapping, the learner may rely on a simpler or more practiced paraphrase. This can reduce the number of observed errors while leaving a knowledge gap. At the opposite extreme, frequent use of a shared construction can reflect genuine facilitation rather than mere reliance on L1. The researcher must examine opportunities to produce each form, not just proportions of errors in the utterances collected. Odlin (2003).
The familiar language-pair prediction that greater structural difference always implies greater difficulty is insufficient. Large differences may be noticed and explicitly taught, while subtle near-similarities may escape attention. Perceived resemblance, linguistic distance, cue reliability, task demands, and instruction jointly shape outcomes. Research on third-language Dutch illustrates that quantitative measures of distance can help explain learner performance while leaving considerable individual variation. Schepens, van der Slik, and van Hout (2016).
Is Transfer Always From L1 to L2?
The direction of influence depends on learning history and the situation. As an additional language becomes well practiced, its vocabulary, pronunciation, or grammar may influence the language learned earlier. The literature calls this reverse or regressive cross-linguistic influence. In multilingual speakers, knowledge from both L1 and L2 can contribute to L3, and the most relevant source may differ by construction. A 2026 review documents studies of regressive grammatical influence even in instructed learning settings, while emphasizing heterogeneity of designs and the need for stronger controls. Olszewska (2026).
Third-language transfer and typological similarity
Third-language learners bring several previously learned systems into the task. The typological primacy model proposes a role for perceived or actual structural proximity in selecting a transfer source. It is an influential theoretical account, not a settled universal rule. Research must distinguish the linguistic distance estimated by the investigator from similarity noticed by the learner, and must compare predictions against alternatives involving recency, proficiency, dominance, and learning context. Rothman (2011); Schepens, van der Slik, and van Hout (2016).
Heritage languages and asymmetric experience
A heritage-language speaker may acquire a family language early and later become more proficient in the surrounding societal language. Amounts and kinds of input change across life stages, creating opportunities for interaction in both directions. Such language histories challenge the assumption that earliest acquisition predicts strongest contemporary representations. Reviews of heritage-language acquisition emphasize exposure, age, and literacy in interpreting differences from comparison groups. Montrul (2023).
Transfer, Language Dominance, and Individual Differences
The strength of transfer varies within nominally identical language groups. Age of acquisition, amount and diversity of input, usage frequency, phonetic aptitude, working memory demands, literacy, formal instruction, social participation, and time spent using each language can all affect performance. These correlates should not be interpreted as interchangeable causes: learners who receive more input may also have greater motivation, different opportunities, or a changed everyday language environment. Flege and Bohn (2021); DeKeyser (2005).
Proficiency and automaticity
At an early learning stage, learners often rely on L1-based analogies because target-language evidence remains sparse. With substantial exposure, some habits weaken, some shared representations strengthen, and some persistent patterns remain. Proficiency and cross-language influence can have nonlinear relationships: a more proficient learner may show less negative transfer in grammatical accuracy but more automatic coactivation in rapid lexical tasks. Studies should therefore interpret specific measures and avoid treating a single language-test score as a complete explanation.
Learning context and social identity
Immersion, classroom instruction, digital practice, work interactions, and home language use supply different distributions of input and feedback. Language choices can also express community membership, family identity, or professional roles. A bilingual speaker's deliberate choice of accent, register, or borrowed expression is not automatically an involuntary processing error. Transfer research is strongest when it incorporates social context without collapsing linguistic mechanism into group stereotypes.
What Evidence Establishes Cross-Linguistic Influence?
Converging evidence is the foundation of a strong scientific inference. Researchers compare learners with different L1 backgrounds acquiring the same L2 target, measure within-group variation, and examine whether the observed difference is predicted by a known property of the prior language. The comparison becomes stronger when groups are appropriately matched on L2 exposure and task experience and when an independently defined L1 measure predicts the result. Jarvis (2000) made these methodological requirements central to contemporary lexical-transfer research. Jarvis (2000).
Cross-sectional learner comparisons
A cross-sectional design compares learners at one point in time. It can identify correlations between L1 background and L2 performance but cannot by itself show developmental change or rule out every historical confound. Sampling should account for the diversity within language groups: bilingual environments, dialects, socioeconomic access to instruction, educational histories, and task familiarity. Randomly equating all speakers of a language to one cognitive profile undermines interpretability.
Longitudinal studies and learning curves
Longitudinal studies observe the same learners over repeated sessions or years, allowing researchers to test whether a hypothesized transfer pattern weakens, strengthens, or changes form. They are valuable for distinguishing an initial strategy from an enduring representation. However, maturation, continuing instruction, attrition, and changes in language use occur alongside learning; longitudinal measurement improves inference but still requires careful comparison conditions.
Training experiments and randomized interventions
Randomized or counterbalanced training designs can test whether explicit contrast, targeted feedback, varied sound examples, or structured practice alters performance on an L2 feature. A transfer hypothesis is strengthened when a manipulation directed at the proposed source of difficulty changes behavior relative to an appropriate control. Immediate improvement does not necessarily generalize to spontaneous speech or survive after training; delayed tests and untrained items provide stronger evidence about retention and generalization. Saito and Plonsky (2019).
Reaction times, eye movements, and priming
Reaction-time, lexical-decision, eye-tracking, and structural-priming tasks reveal how quickly alternatives are activated and how previously encountered structures affect subsequent processing. These methods illuminate stages of comprehension and production, especially when accuracy is near ceiling. Their outcomes are sensitive to stimulus matching, item frequency, response strategies, and statistical choices. Structural priming can be informative about shared representations, but a short-lived shift after a prime should not be automatically interpreted as stable grammar learning. Hartsuiker et al. (2016); Shin and Christianson (2012).
Corpus evidence and spontaneous language
Learner corpora, classroom recordings, and spontaneous conversation samples show how transfer unfolds under realistic communicative demands. They capture choices and avoidance that tightly controlled laboratory tasks may miss. Yet frequencies depend on topic, genre, prompt, editing assistance, and learner selection. Corpus comparisons need denominators: if a construction rarely appears, the absence of errors says little about the learner's ability to use it. Corpus regularities should be triangulated with comprehension, elicitation, and controlled production.
Meta-analysis and the limits of evidence aggregation
Systematic synthesis can estimate an effect across studies, but heterogeneity in language pairs, ages, task types, measures, and exposure complicates pooling. A 2014 meta-analysis of cross-linguistic syntactic priming reported substantial evidence for priming alongside heterogeneous findings in some comparisons; its estimate is specific to the included experiments and cannot be generalized mechanically to all domains of transfer. Jiang and Chen (2014).
Major Theoretical Accounts
Contrastive analysis: a historical starting point
Early contrastive analysis proposed that systematic comparisons between two languages could help anticipate certain learner difficulties. Robert Lado's 1957 work gave detailed attention to sound, grammar, vocabulary, and writing systems. Its historically strong prediction that contrast alone would reliably forecast the complete error pattern proved too simple because learners also create developmental patterns, notice features selectively, and vary in exposure and goals. Modern transfer research retains careful language comparison while asking which differences matter in processing and why. Lado (1957); Odlin (2003).
Learned attention and cue competition
An associative-learning explanation holds that extensive L1 experience makes some cues highly predictive and others easy to ignore. When L2 relies on a cue that L1 rarely uses, the new form may be present in the input yet receive insufficient attention. Concepts such as blocking, overshadowing, contingency, and perceptual learning can explain why repeated exposure sometimes produces surprisingly limited uptake. This is a testable mechanistic proposal with implications for training, not a claim that conscious attention is sufficient for all language learning. Ellis (2006).
Speech learning models
The revised Speech Learning Model describes how phonetic categories respond to new experience over the lifespan. It predicts that forming a sufficiently separate L2 category depends partly on the perceived similarity between L2 input and the nearest available L1 category, together with the precision of prior categories and the nature of new experience. The model is tailored to phonetic learning; its propositions should not be extended without evidence to vocabulary, pragmatics, or every type of syntax. Flege and Bohn (2021).
Shared-syntax and structural-priming accounts
Models of bilingual syntax seek to explain when equivalent constructions in different languages share abstract structure and when linguistic details remain language-specific. Cross-language priming is a central tool because it can reveal facilitation between constructions across language boundaries. The evidence supports meaningful interaction, although experiments differ on the degree of overlap and its development. Hartsuiker et al. (2016); Muylle, Hartsuiker, and Bernolet (2023).
Lexical activation and bilingual interactive models
Interactive models such as BIA+ represent cross-language lexical competition and facilitation during word recognition. Their value lies in explicit claims about the interaction of orthographic, phonological, semantic, and task-related processes. Results from printed words cannot be assumed to apply unchanged to connected spoken language or to every writing system; competing models may fit different tasks or populations. Dijkstra and van Heuven (2002).
Multilingual transfer-selection accounts
Models of L3 acquisition ask which previously known language supplies relevant structure to a newly learned language. Structural proximity, proficiency, acquisition history, and perceived similarity offer competing or complementary explanations. The typological primacy model generates testable predictions, but findings depend on the particular grammatical domains and learner groups under examination. Researchers should avoid presenting one selection mechanism as settled for all multilingual learners. Rothman (2011).
Educational Implications: Using L1 Knowledge Productively
A scientifically informed approach treats prior-language knowledge as a resource for designing practice. Teachers can identify genuine correspondences, make consequential differences salient, and build tasks that require the learner to use the new distinction in meaningful contexts. Instruction that directly compares two structures may be helpful when the learner's habitual interpretation masks a target-language cue. Teaching should remain feature-specific: methods that improve perception of a vowel may not teach article use or pragmatic appropriateness. Ellis (2006); Saito and Plonsky (2019).
Teach contrasts rather than isolated mistakes
For pronunciation, examples that vary by speaker, phonetic environment, and lexical meaning can help a learner identify what difference carries information in the new language. For grammar, compare form, meaning, and distribution in context rather than simply declaring that L1 permits one order and L2 another. For vocabulary, pair cognates with their actual senses and collocations and distinguish false friends explicitly. The instructional objective is a useful new representation supported by repeated retrieval and use.
Build retrieval and generalization into practice
Short-term improvement on practiced sentences provides weaker evidence than successful performance on untrained words, speakers, sentence contexts, or spontaneous conversation. Delayed reassessment can distinguish a temporary task strategy from more durable learning. In pronunciation research, an instructional meta-analysis found important differences according to the measurement target and whether production was monitored or spontaneous; a single overall pronunciation rating can conceal that variation. Saito and Plonsky (2019).
Avoid blanket bans on a learner's first language
Using L1 to explain a contrast or anchor a new lexical meaning can make instruction clearer. Exclusive L2 practice can be useful for particular communicative goals, but a general ban on L1 has no automatic basis in the transfer literature. The informative question is whether a practice helps develop the target-language perception, representation, fluency, and context-sensitive use intended by the lesson. Learners need sufficient opportunities to process the L2 while drawing thoughtfully on the linguistic experience they already possess.
Evaluate progress in multiple conditions
A learner who can identify a word after written study may still need listening practice; someone who succeeds on a multiple-choice grammar test may not yet produce the same form fluently. Assessment should include accuracy, speed, comprehension, and spontaneous use when those outcomes matter. For educational decisions, language background, multilingual competence, and exposure should be documented rather than interpreted through monolingual norms alone.
Common Misconceptions
Does a first language permanently prevent native-like L2 speech?
No single deterministic rule follows from the transfer literature. Early learned phonetic categories can make some distinctions harder to acquire, yet speech learning models emphasize continued experience-dependent adaptation. Accent, intelligibility, phonetic precision, and social identity are separate dimensions. Many learners communicate effectively with clearly recognizable features of their language histories. Flege and Bohn (2021); Saito (2021).
Do language differences predict every error?
A contrast between language systems identifies a plausible locus of transfer but cannot establish that transfer caused a specific mistake. Developmental learning strategies, shared processing constraints, limited exposure, and the wording of a task can produce similar outcomes. Comparisons across L1 groups, longitudinal evidence, and interventions help separate these explanations. Jarvis (2000).
Is using a familiar language in learning another language necessarily harmful?
Prior language knowledge provides existing categories and connections that can support learning. Its influence becomes problematic only relative to a target feature and task when familiar patterns compete with the new ones. Positive and negative effects commonly coexist within the same learner. Odlin (2012).
Is code-switching proof that a learner has confused their languages?
Alternating languages can be deliberate, contextually appropriate, and socially meaningful. Cross-language activation may contribute to some switching patterns, but switching itself does not establish defective language control or a clinical disorder. Explanation requires the conversational setting, interlocutors, speaker preferences, and linguistic competence to be considered.
Can scientists identify a learner's first language from one sentence?
Certain sound or grammatical patterns may be more common in learners with particular linguistic backgrounds, but a single utterance cannot securely identify language history. Dialect, multilingual experience, pronunciation training, accommodation, and individual habits affect the same features. Statistical group-level differences must not be turned into confident claims about an individual.
Cross-Linguistic Influence and Language Disorders
Cross-linguistic influence is a normal consequence of learning and using multiple languages. An accent, code-switching, an unfamiliar word order, or an L2 grammatical error is not by itself a marker of developmental language disorder, aphasia, dyslexia, stuttering, or a speech motor disorder. Clinical judgments require developmentally and linguistically appropriate evidence across a person's languages and relevant functional settings. The present article addresses normal cognitive mechanisms and the methodological separation between language learning and impairment; clinical diagnostic procedures belong to specialized assessments.
Research on multilingual children with and without developmental language disorder can help distinguish effects associated with transfer from difficulties requiring further investigation, but findings from a single learner group do not yield universal diagnostic rules. Cross-language interactions may change with proficiency, input, or educational opportunity. When an individual has significant communication concerns, qualified professionals should evaluate language history, access to instruction, and functioning rather than interpreting English-only performance in isolation. Tribushinina and Boz (2025).
Limits and Open Scientific Questions
Evidence that an L1 predicts an L2 outcome is informative but often correlational. A language background also reflects schooling, social environment, literacy, and patterns of exposure. Randomized training isolates selected mechanisms but usually tests narrower tasks than everyday communication. Corpus studies show ecological behavior while making it harder to infer cause. Consequently, the strength of conclusions depends on the convergence of methods and the specificity of the hypothesis.
Current debates concern how much phonetic, lexical, and syntactic structure is shared across languages, how fast those connections are reorganized, how typological proximity should be measured, whether short-lived processing facilitation produces enduring learning, and how language dominance changes across life. The evidence supports cross-language interactions across many domains; the magnitude, direction, developmental trajectory, and best explanatory theory must be established for particular language pairs and tasks. Amengual (2023); Muylle, Hartsuiker, and Bernolet (2023).
The comparison of statistical language patterns in humans and large language models introduces further research questions, but model performance on multilingual text does not establish equivalence with the perceptual learning, embodied speech production, social experience, or subjective life of human language users. The relevant empirical comparisons require explicit tasks and independently defined mechanisms. Human cross-linguistic influence is studied through learners' histories, perception, behavior, and developing linguistic systems.
Frequently Asked Questions
What is the difference between language transfer and cross-linguistic influence?
Language transfer is a common term for the effect of existing language knowledge on another language. Cross-linguistic influence is often preferred when the account includes multiple directions, online activation, multilingual interactions, and changing relations between languages. Authors use the terms with partially overlapping meanings, so the relevant study's operational definition matters. Odlin (2012).
What is an example of positive transfer?
A learner recognizes a new vocabulary item because its form and meaning align with a well-known word from another language. Comparable grammatical patterns and previously learned literacy conventions can also facilitate early processing. The actual advantage is tested through matched words, tasks, and learner groups.
What is an example of negative transfer?
A learner interprets an L2 sound as belonging to an L1 category and initially fails to use the contrast that distinguishes L2 words. Another example is choosing an L1-like syntactic option where the new language requires a different construction. Such patterns are hypotheses about the learning mechanism until relevant comparisons support them.
Can a second language change the first language?
Yes. Cross-linguistic influence can be bidirectional, and new linguistic experience can affect established lexical, phonetic, or grammatical behavior. The scale and persistence of such changes vary by domain, use, and learner history. A 2026 review synthesizes evidence about reverse grammatical effects in instructed settings. Olszewska (2026).
Does knowing related languages always make another language easier?
Related languages can share useful patterns, but formal similarity can also conceal differences in distribution or meaning. The practical effect depends on the target feature, the learner's proficiency in each known language, and the type of task. Linguistic distance is an informative variable rather than a single universal difficulty score. Schepens, van der Slik, and van Hout (2016).
Why do learners make errors they can correct when given time?
Knowledge that is available during careful reflection may be difficult to retrieve or implement rapidly during speech. Familiar L1-based candidates can compete with less practiced L2 forms, and attentional demands change across tasks. Comparing monitored and spontaneous production helps identify the difference between explicit knowledge and online use.
How can a teacher use transfer research without stereotyping students?
A teacher can compare actual linguistic constructions and test how an individual learner handles them. Useful instruction starts with the student's languages, dialects, literacy, experiences, and immediate goals. Research-based patterns guide hypotheses about which contrasts may deserve practice; the student's observed learning determines how instruction proceeds.
Conclusion
Cross-linguistic influence is a pervasive, measurable feature of how experienced language users learn new linguistic systems. Established knowledge contributes phonetic categories, lexical connections, grammatical expectations, and selective attention. Those contributions can facilitate acquisition, compete with target-language patterns, or change as additional learning accumulates. The scientific account is strongest when it specifies a linguistic feature, distinguishes activation from lasting learning, and tests predictions with convergent comparisons and experiments. A first language shapes the starting conditions of new language learning while continued experience makes those relationships dynamic. Jarvis (2000); Odlin (2003); Muylle, Hartsuiker, and Bernolet (2023).
Related Articles
Cluster Language, Thought and Psycholinguistics — the published scientific chapter connecting language acquisition, bilingual processing, phonology, syntax, and psycholinguistic methods.
Psycholinguistics: How Language Is Processed in the Mind — the general cognitive framework for understanding and producing language.
Speech Perception: How the Mind Turns Sounds Into Words — the sound categorization and recognition processes most relevant to phonological transfer.
Word Recognition: How the Mind Identifies Spoken and Written Words — lexical access, competition, and the recognition of unfamiliar words.
Sentence Processing: How We Understand Grammar in Real Time — online parsing and the interpretation of grammatical cues.
References
Amengual, M. (2023). Cross-language influences in the acquisition of L2 and L3 phonology. In Cross-language Influences in Bilingual Processing and Second Language Acquisition, 74–99. https://doi.org/10.1075/bpa.16.04ame
DeKeyser, R. M. (2005). What makes learning second-language grammar difficult? A review of issues. Language Learning, 55(S1), 1–25. https://doi.org/10.1111/j.0023-8333.2005.00294.x
Dijkstra, T., and van Heuven, W. J. B. (2002). The architecture of the bilingual word recognition system: From identification to decision. Bilingualism: Language and Cognition, 5(3), 175–197. https://doi.org/10.1017/S1366728902003012
Ellis, N. C. (2006). Selective attention and transfer phenomena in L2 acquisition: Contingency, cue competition, salience, interference, overshadowing, blocking, and perceptual learning. Applied Linguistics, 27(2), 164–194. https://doi.org/10.1093/applin/aml015
Flege, J. E., and Bohn, O.-S. (2021). The Revised Speech Learning Model (SLM-r). In Second Language Speech Learning: Theoretical and Empirical Progress. Cambridge University Press
Flege, J. E., Aoyama, K., and Bohn, O.-S. (2021). The Revised Speech Learning Model (SLM-r) applied. In Second Language Speech Learning: Theoretical and Empirical Progress, 84–118. https://doi.org/10.1017/9781108886901.003
Hartsuiker, R. J., Beerts, S., Loncke, M., Desmet, T., and Bernolet, S. (2016). Cross-linguistic structural priming in multilinguals: Further evidence for shared syntax. Journal of Memory and Language, 90, 14–30. https://doi.org/10.1016/j.jml.2016.03.003
Hopp, H., and Jackson, C. N. (2023). Asymmetrical effects of cross-linguistic structural priming on cross-linguistic influence in L2 learners. Applied Psycholinguistics, 44(2), 205–236. https://doi.org/10.1017/S0142716423000164.
Jarvis, S. (2000). Methodological rigor in the study of transfer: Identifying L1 influence in the interlanguage lexicon. Language Learning, 50(2), 245–309. https://doi.org/10.1111/0023-8333.00118
Jiang, X., and Chen, L. (2014). A meta-analysis of cross-linguistic syntactic priming effects. Contemporary Foreign Languages Studies, 14(12), 12–17. https://doi.org/10.3969/j.issn.1674-8921.2014.12.002
Kroll, J. F., Dussias, P. E., Bogulski, C. A., and Kroff, J. R. V. (2012). Juggling two languages in one mind: What bilinguals tell us about language processing and its consequences for cognition. Psychology of Learning and Motivation, 56, 229–262. https://doi.org/10.1016/B978-0-12-394393-4.00007-8
Lado, R. (1957). Linguistics Across Cultures: Applied Linguistics for Language Teachers. University of Michigan Press. Publisher listing
Montrul, S. (2023). Heritage languages: Language acquired, language lost, language regained. Annual Review of Linguistics, 9, 399–418. https://doi.org/10.1146/annurev-linguistics-030521-050236
Muylle, M., Hartsuiker, R., and Bernolet, S. (2023). Cross-language influences in L2 syntactic processing and production in late L2 learners. In Cross-language Influences in Bilingual Processing and Second Language Acquisition, 262–293. https://doi.org/10.1075/bpa.16.11muy
Odlin, T. (2003). Cross-linguistic influence. In The Handbook of Second Language Acquisition. https://doi.org/10.1002/9780470756492.ch15
Odlin, T. (2012). Crosslinguistic influence in second language acquisition. In Encyclopedia of Applied Linguistics. https://doi.org/10.1002/9781405198431.wbeal0292
Olszewska, A. (2026). Instructed language learning and its impact on previously acquired languages: A state-of-the-art review of regressive cross-linguistic influence across grammatical domains. Lingua. https://doi.org/10.1016/j.lingua.2026.104156
Rothman, J. (2011). L3 syntactic transfer selectivity and typological determinacy: The typological primacy model. Second Language Research, 27(1). https://doi.org/10.1177/0267658310386439
Saito, K. (2021). What characterizes comprehensible and native-like pronunciation among English-as-a-second-language speakers? Meta-analyses of phonological, rater, and instructional factors. TESOL Quarterly, 55(3), 866–900. https://doi.org/10.1002/tesq.3027
Saito, K., and Plonsky, L. (2019). Effects of second language pronunciation teaching revisited: A proposed measurement framework and meta-analysis. Language Learning. https://doi.org/10.1111/lang.12345
Schepens, J. J., van der Slik, F., and van Hout, R. (2016). L1 and L2 distance effects in learning L3 Dutch. Language Learning, 66, 224–256. https://doi.org/10.1111/lang.12150
Shin, J.-A., and Christianson, K. (2009). Syntactic processing in Korean-English bilingual production: Evidence from cross-linguistic structural priming. Cognition, 112(1), 175–180. https://doi.org/10.1016/j.cognition.2009.03.011
Shin, J.-A., and Christianson, K. (2012). Structural priming and second language learning. Language Learning, 62(3), 931–964. https://doi.org/10.1111/j.1467-9922.2011.00657.x
Tribushinina, E., and Boz, B. (2025). The effects of heritage multilingualism on foreign language learning: A comparison of children with typical language development and developmental language disorder. Frontiers in Psychology, 15, 1521340. https://doi.org/10.3389/fpsyg.2024.1521340