What Is Language Attrition?
Language attrition is a change or reduction in the accessibility, efficiency, or performance of language abilities that a person previously acquired. It can affect a first language, an additional language, or a particular linguistic skill after everyday use and exposure change. An adult who moves into a new linguistic environment may begin searching longer for familiar words in the language spoken during childhood. A former language student may understand written passages years after lessons have ended while struggling to produce even a short conversation. Another person may speak fluently but acquire subtle features of another language's pronunciation or sentence patterns. These experiences can belong to the broader scientific field of language attrition, although their underlying causes and their implications differ.
The most important principle is that a decline in observed performance does not establish the disappearance of linguistic knowledge. Remembering a word, selecting it among competing alternatives, producing its sounds, understanding it in context, and recognizing it when heard are distinct operations. Their accessibility may change at different rates. Attrition research therefore studies the changing relation among knowledge, retrieval, language use, multilingual experience, and social environment rather than assuming that the entire language is progressively erased. Gallo et al. (2021); Schmid and Köpke (2017).
In its established psycholinguistic usage, first-language attrition concerns non-pathological changes after earlier acquisition, especially when another language becomes dominant. Second-language attrition concerns changes in an additional language after instruction, immersion, or frequent use decreases. Both involve language histories and contextual influences, but they begin from different learning conditions. Language attrition is a research construct describing patterns of change; it is not a clinical diagnosis. Schmid (2016); Schmid (2023).
Why Does Language Become Harder to Use?
A person's linguistic system is shaped by repeated encounters with sounds, words, meanings, and communicative situations. When habitual exposure changes, certain representations are activated less frequently, alternative expressions become more dominant, and the demands placed on attention and retrieval shift. Speech production is particularly sensitive because the speaker must generate a form quickly without being shown the answer. Recognition and comprehension may be preserved even when spontaneous production becomes slow or incomplete. Research on long-term retention of classroom Spanish demonstrates that learning history and type of test strongly affect what looks like retention or loss. Bahrick (1984).
There are several plausible and partially overlapping explanations. Reduced use may lower the current accessibility of representations. A newly dominant language may compete during lexical selection and change habitual activation patterns. Continued bilingual experience may reorganize how languages interact even when knowledge remains substantial. The social environment may also reshape which vocabulary, registers, and discourse practices are relevant. Contemporary research cannot assign one fixed percentage of attrition to disuse, competition, or restructuring across all populations. Schmid and Yilmaz (2021); Gallo et al. (2021).
Accessibility and storage are different questions
A speaker can fail to retrieve a name on command yet recognize it immediately when it appears in a list. This difference is central to cognitive interpretations of attrition. A naming task measures the ability to produce a target rapidly; a recognition task allows the stimulus itself to provide a cue. Poor naming combined with good recognition suggests a retrieval bottleneck, although it does not prove that the stored representation is entirely unchanged. Long-term memory may preserve partial information that supports recognition, relearning, and comprehension even after fluent production weakens.
Language systems contain many kinds of knowledge. A familiar word can retain its meaning while its phonological form becomes difficult to retrieve. A grammatical construction can remain understandable but become unusual in spontaneous speech. A speaker may accurately distinguish native sound categories but produce them with altered timing. Attrition should be described at the level of the measured ability and task rather than as a global score for a whole language.
Frequency, recency, and activation
Frequently encountered words are generally retrieved with less effort than rare ones, but frequency is not a single property. Researchers distinguish lifetime exposure, recent encounters, cumulative proficiency, context diversity, and frequency of actual production. An older adult may retain a rich vocabulary while finding low-frequency words harder to access after decades of using another language. A bilingual professional may remain fluent in everyday family conversation yet hesitate when using specialized terminology never needed in that language.
A theoretical account based on reduced activation predicts that representations used less often become temporarily less available, particularly during rapid naming or demanding production. Its predictions must be tested against alternatives: a highly frequent word can still be influenced by competing forms, and self-reported language use may be an unreliable approximation of actual exposure. Schmid and Yilmaz found evidence of reduced lexical accessibility in migrants, while their measures of individual exposure frequency did not explain the observed group differences. This is an important warning against treating the simple phrase 'use it or lose it' as a complete causal model. Schmid and Yilmaz (2021).
Competition and cross-linguistic influence
Languages acquired by the same person interact. A word in one language may activate another language's phonological, grammatical, or semantic alternatives, depending on the task and speaker. When a new language becomes dominant, selecting an expression from the earlier language can require suppressing or resolving competing responses. Transfer may also affect preferred structures, discourse conventions, lexical associations, and pronunciation. These effects are not uniform across language pairs or tasks. Schmid and Köpke (2017).
For example, a speaker whose daily conversations take place in English may begin choosing English-style phrasing while speaking another language. The change could reflect habitual selection, contact-induced restructuring, or context-sensitive code-switching. Establishing attrition requires evidence about the speaker's earlier and current competence, comparison groups, and specific linguistic measures. Borrowing an English noun in a bilingual conversation is not by itself evidence that the corresponding first-language noun has been forgotten.
A dynamic system rather than a single decline curve
Complex dynamic approaches examine the stability of an established language system alongside variation produced by age, bilingual experience, social environments, and interacting subsystems. A deeply established adult first language can remain robust despite reduced use; particular retrieval operations may still become vulnerable. Researchers have described an acquired first language as a relatively stable attractor state, although the framework is a theoretical interpretation requiring operational tests rather than a literal brain structure. Schmid, Köpke, and de Bot (2013).
A person's dominant language may also change repeatedly across life. Work, schooling, migration, family relationships, and community participation alter which language must perform which function. Consequently, language histories can show decline, stability, partial recovery, and new proficiency in different domains at the same time. The scientifically relevant question is which performance changed, under which conditions, compared with which earlier baseline.
First-Language Attrition: What Happens to a Native Language?
First-language attrition often becomes noticeable after prolonged residence in an environment where another language is required for employment, education, services, and friendships. People may have retained full comprehension of their childhood language while finding word retrieval less automatic. They may experience unfamiliar hesitation, replace specific expressions with more general descriptions, or notice accents and discourse patterns influenced by the socially dominant language. Studies of migrants show that lexical accessibility can differ from that of comparison speakers who remain in the first-language environment. Schmid and Jarvis (2014); Schmid and Yilmaz (2021).
The word native does not imply permanent immunity to change. Every living linguistic system remains sensitive to experience. Still, adult native speakers possess extensive, repeatedly reinforced language knowledge. Their situation is different from that of children whose early language development is interrupted before extended schooling, literacy, and sustained community use. Age of onset, strength of initial acquisition, and sociolinguistic context must therefore be considered before comparing groups.
Lexical retrieval and conversational fluency
Vocabulary access is among the most studied domains of first-language attrition. A speaker may pause longer, use circumlocutions, repeat general-purpose verbs, or produce fewer low-frequency items during a constrained naming task. Researchers compare response times, accuracy, lexical diversity, and spontaneous speech, but these measurements need careful interpretation. A lower type-token ratio, for instance, can result from differences in topic, passage length, task demands, and interactional situation. Schmid and Jarvis used multiple forms of elicitation and free speech to examine precisely these measurement problems. Schmid and Jarvis (2014).
Speech is also an adaptive social activity. A migrant who uses one language mainly for family routines and another for employment may preserve emotionally salient family vocabulary while developing more elaborate occupational vocabulary in the second language. Reduced fluency in a specialist topic may reflect unequal experience rather than overall deterioration. Comparing each language only against an idealized monolingual standard hides this division of communicative labor.
Changes in pronunciation and sound categories
Phonetic attrition has received renewed experimental attention. Speech sounds are learned as categories and production routines shaped by long-term experience. Intensive use of a second language can influence voice onset time, consonant articulation, rhythm, and other measurable features in the first language. A speaker can remain fully intelligible while showing subtle acoustic differences relative to comparable speakers in the first-language environment.
A study published in 2026 compared Polish-English bilinguals living in Poland with long-term Polish-English migrants living in the United Kingdom. It examined Polish stop consonant perception and production, using behavioral measures and electrophysiological assessment. After a short return to Poland, the migrant group's production became more similar to the Poland-based group's production, whereas the corresponding perception difference did not reverse in the same way. The finding demonstrates domain-specific plasticity within that sample; it cannot establish that all phonetic attrition is reversible or that identical changes occur in every language pair. Marecka et al. (2026).
Grammar, morphology, and discourse
Grammatical effects require especially careful interpretation. Morphological agreement, syntactic preferences, and the integration of grammar with discourse may show different sensitivities to language contact and processing demands. A bilingual speaker can produce an unusual construction because it is less accessible at that moment, because of transfer from another language, or because a community uses a different conventional pattern. Different experimental tasks may elicit different outcomes.
Studies of heritage speakers and second-language learners indicate that age and context of acquisition matter for particular morphosyntactic structures. For instance, research comparing Spanish gender agreement across heritage and second-language populations documented how acquisition histories can yield distinct profiles. Such evidence helps define questions about attrition, but incomplete acquisition and later attrition are competing explanations in many heritage-language datasets. One cannot infer individual loss without knowing what the person previously acquired. Montrul, Foote, and Perpiñán (2008); Schmid and Köpke (2017).
Discourse also changes with audience and identity. The rate of story elaboration, use of culturally specific references, and selection of politeness forms may reflect adaptation to a different language community rather than weakened grammar. Research designs therefore separate structural knowledge, processing efficiency, and socially appropriate language choice.
Second-Language Attrition: Why a Learned Language Fades
Second-language attrition occurs when an additional language is no longer regularly practiced or encountered. Its starting conditions matter enormously. People who studied a language for one semester, people who used it daily for years, and people who achieved fluent professional competence have different stores of linguistic knowledge. Their outcomes also depend on whether the retained skill is reading, comprehension, grammar judgment, vocabulary recall, pronunciation, or spontaneous interaction.
Classroom-based second-language attrition has historically received less sustained research than adult first-language attrition in migration. The lack of longitudinal baseline measurements makes it difficult to disentangle what learners once knew, what was only briefly accessible during testing, and what persists for decades. Schmid's review of instructed foreign-language attrition emphasizes this evidence gap. Schmid (2023).
Why beginners and advanced learners differ
Original learning strength is one of the most robust predictors of long-term retained knowledge. In a classic study of Spanish learned in school, Bahrick assessed adults across intervals extending to fifty years after formal instruction. Some knowledge declined relatively early, while portions of previously established language knowledge remained accessible over very long intervals. The strongest patterns were related to level and success of earlier training and to the kind of retrieval or recognition task used. The study does not imply that any specific percentage of proficiency is universally preserved in every language. Bahrick (1984).
A learner who initially memorized a small list of words may lose rapid recall after a prolonged interval. Another who used those words in extensive reading, conversations, and writing may retain multiple retrieval routes. Recognizing a familiar item during reading requires different processing from creating an original sentence under social pressure. Describing attrition as simply proportional to years without practice overlooks this history.
Recognition versus spontaneous production
A former learner may correctly choose the translation of a word from multiple options while failing to produce that translation unaided. Recognition provides external support; production requires internal cue generation, selection, and articulation. Reading can therefore appear surprisingly strong compared with speaking, even when both rely on earlier learning. Tests that measure only one domain cannot establish the status of the others.
Comprehension also depends on task difficulty, context, and redundancy. A familiar topic allows inference from discourse, pictures, and situational knowledge; isolated listening in noise provides fewer aids. Apparent retention in a richly cued situation may coexist with difficulty on decontextualized tests. A valid assessment combines complementary tasks and reports each outcome separately.
Can years without study leave durable knowledge?
Yes. Long-term retention research shows that some foreign-language knowledge can persist for decades. Bahrick referred to exceptionally durable stored material as 'permastore'; the term denotes an empirical pattern in retention research rather than proof that knowledge is literally permanent or resistant to every form of change. Follow-up studies of foreign-language word pairs also found benefits of spaced practice over extended delays. Bahrick (1984); Bahrick et al. (1993).
The practical consequence is that apparently forgotten skills may be easier to reacquire than entirely unfamiliar material. Relearning benefits are consistent with retained information, yet the magnitude of any advantage depends on what was learned and how it is tested. Claims that a language has been retained forever cannot be justified by a single successful vocabulary recognition test.
Which Language Abilities Are Most Vulnerable?
A language is not one indivisible ability. Researchers distinguish phonological categories, pronunciation routines, lexical representations, grammar, semantic organization, pragmatic competence, literacy, and discourse fluency. Each may respond to reduced use differently. The apparent ranking of vulnerability depends strongly on methods, languages, age groups, and baseline proficiency.
Vocabulary and names
Word-finding difficulty is an accessible example because the individual experiences it directly. People can know what they want to say and still fail to generate a precise word quickly. However, naming difficulty can arise from ordinary aging, fatigue, divided attention, low familiarity with the topic, a rarely used register, or bilingual lexical competition. Strong attrition evidence requires more than one anecdotal lapse. Researchers analyze error types, reaction times, and performance across comparable lexical categories. Schmid and Jarvis (2014).
Speech perception and pronunciation
Production and perception are related through long-term language learning but can have distinct plasticity. A person may produce a consonant differently yet identify native-language contrasts reliably, or may show measurable perceptual changes that persist after production adjusts. The 2026 Polish-English re-immersion study supplies a concrete instance of unequal recovery across these components. Marecka et al. (2026).
Morphology, grammar, and sentence processing
Short-term production problems can arise when structures compete or when speakers have become less practiced at selecting a particular grammatical form. Grammatical judgment tests provide complementary evidence but can themselves be influenced by literacy, task strategy, and awareness of normative rules. Long-term structural reorganization is a stronger claim than temporary difficulty and needs longitudinal or convergent evidence. Gallo et al. (2021).
Pragmatics, idioms, registers, and cultural vocabulary
A language may remain grammatically fluent while specific idioms, humor, institutional vocabulary, or culturally situated references become less available. A person who moved abroad twenty years ago may know household vocabulary in a heritage language but not current terminology for digital administration or recent public events. Such vocabulary can be missing because it was never learned in that language, not because it was forgotten. The distinction between loss and absence of new acquisition is critical to valid interpretation.
Reading and writing
Literacy involves learned routines that interact with language processing, schooling, and regular practice. A speaker may converse effectively yet write more slowly in an alphabet or orthography used infrequently. Similarly, reading comprehension can remain strong when written production becomes effortful. Orthographic changes and writing-system experience require separate measurement from oral vocabulary attrition. The presence of a writing difficulty alone is not evidence of a specific learning disorder.
Does Language Disuse Cause Permanent Forgetting?
Current evidence supports a more precise account: reduced exposure and use can be associated with changes in performance and accessibility, and some underlying knowledge may remain available for long periods. Research does not provide a universal biological point at which language knowledge becomes irreversibly lost. The term irreversible is especially demanding because recovery may depend on test type, length and intensity of renewed exposure, and the history of initial learning. Bahrick (1984); Schmid (2023).
Loss also has several meanings. A person can stop using a language as a primary means of social participation while retaining extensive comprehension. A community can shift toward another language across generations even while individual elders remain highly proficient. An individual can show weaker lexical retrieval without measurable decline in grammar. Researchers should specify whether the outcome is reduced social use, decreased processing efficiency, measurable change in representations, or the near absence of functional performance.
Why the idea of a single forgetting curve is misleading
Traditional memory research can track performance across retention intervals, but language is supported by complex networks of use, meaning, context, and continuing new learning. The slope of change depends on what is measured and the interval examined. Different individuals may alternate between periods of active use and neglect. Long-term Bahrick findings show that rates of measured decline are not necessarily constant over decades. Bahrick (1984).
Language may also become more accessible in a familiar context. Hearing voices from an earlier community, returning to familiar routines, or participating in extended conversation can provide cues and reactivate well-practiced associations. These observations motivate experimental tests but should not be treated as proof that every apparent loss is only temporary inhibition.
Initial learning quality and consolidation
Learning a word once in a single classroom exercise is different from encountering it across many meaningful situations. Repeated successful retrieval, varied contexts, and meaningful communicative use strengthen the opportunities for later access. A nine-year longitudinal investigation of foreign-language word pairs found that longer intervals between relearning sessions supported later retention even when acquisition during training was somewhat slower. Because the study involved a small sample and a specific word-pair paradigm, its exact quantitative effects should not be generalized to all spoken language abilities. Bahrick et al. (1993).
Is Language Attrition the Same as Language Shift, Incomplete Acquisition, or Ordinary Forgetfulness?
These concepts describe related but distinguishable phenomena. Language shift usually describes a change in language use or dominance at the level of a person, family, or community. Attrition refers to changes in previously established abilities. A community may shift to the majority language without every bilingual resident losing previously acquired skills. Conversely, an individual may experience measurable phonetic change while continuing to use the original language regularly. Gallo et al. (2021).
Incomplete acquisition is particularly important when studying heritage-language speakers. If a childhood language was used mainly at home without sustained schooling or literacy, some formal registers and structures may not have developed to the same extent as among peers educated through that language. Differences visible in adulthood could reflect developmental history, later reduction in use, influence from another language, or their interaction. Measuring current proficiency alone cannot determine which process occurred. Montrul, Foote, and Perpiñán (2008).
Ordinary word-finding lapses are common during healthy language use, especially under pressure or distraction. A single tip-of-the-tongue episode is insufficient to identify attrition. Attrition research seeks systematic differences over time or between appropriately matched groups, supported by multiple tasks. Language dominance, multilingual processing, and attrition overlap in some measures but should not be collapsed into one causal explanation.
Age, Migration, and the Stability of Earlier Learning
Age is relevant because the strength and organization of a language system develop over time. A child who stops receiving sustained input during an early period of acquisition may show a different trajectory from an adult whose first language was repeatedly used across schooling and working life. Research on age-related differences draws attention to the stabilization of previously acquired language skills, but the exact boundary between developmental vulnerability and adult attrition is not a universal age threshold. Schmid, Köpke, and de Bot (2013).
Migration changes much more than language frequency. Daily interaction partners, literacy practices, employment demands, personal identity, stress, and language attitudes may change simultaneously. A study finding differences between migrants and non-migrants therefore does not isolate one causal mechanism. More persuasive designs measure language use and proficiency before migration, during immersion, and after changes in the environment, while tracking relevant comparison groups.
Adults moving into a second-language environment
In adult migrants, the earlier language frequently remains deeply entrenched while lexical retrieval, phonetic production, and pragmatic conventions adjust. The nature of these changes depends on language pairing and social circumstances. A bilingual may choose the newly dominant language for workplace concepts yet preserve rich first-language knowledge in conversations with close relatives. Monolingual norms are not the only appropriate comparison: comparing bilingual migrants with non-migrant bilinguals can identify effects that accompany bilingual processing independently of migration. Schmid and Yilmaz (2021).
Children and heritage-language environments
The development of a heritage language depends on sustained opportunities for listening, interaction, literacy, and schooling. A child who understands family speech but prefers to respond in the majority language may retain receptive knowledge that exceeds spoken production. Developmental case research documents how changes in input and social context can reactivate previously passive minority-language production, but single cases cannot define population-wide recovery rates. Schmid and Köpke (2017).
Descriptions such as 'a child forgot the native language' can be misleading when baseline development was incomplete or never measured. Researchers should ask which abilities were acquired before environmental change, how they were assessed, and what input remained available. The social value of heritage-language maintenance is a separate consideration from evidence that a specific linguistic representation was lost.
Dominance, identity, and language attitudes
Language dominance describes which language is currently more accessible or useful across contexts. It can change within one lifetime and need not correspond to language first acquired. Identity and attitudes can influence participation in language communities, opportunities for practice, and willingness to produce a language under evaluation. These social variables are important research candidates, but their correlations with measured attrition vary by study; no single emotional or motivational trait predicts every trajectory.
What Can Neuroscience Tell Us About Attrition?
Language relies on distributed neural systems that support auditory and visual processing, lexical selection, semantic interpretation, production, learning, and cognitive control. Researchers use EEG and event-related potentials to measure time-sensitive neural responses, structural and functional MRI to investigate anatomical and activation correlates, and behavioral tasks to connect neural measures with linguistic performance. None of these techniques provides a direct photograph of a forgotten word or a diagnostic scan for everyday language attrition. Gallo et al. (2021).
A major scientific challenge is that increased or decreased brain activation can support different interpretations: increased effort, altered network efficiency, task difficulty, compensatory recruitment, or differences in strategy. The same result cannot establish that a linguistic representation has disappeared. Strong inference requires matched tasks, adequate controls, and converging neural and behavioral measurements.
Distinguishing production and perception experimentally
Phonetic measurements illustrate the value of combining methods. Researchers can measure voice onset time during speech production, phoneme identification and discrimination during listening, and electrophysiological responses to sound contrasts. If production shifts while perception remains stable, a production-specific explanation becomes plausible. If both shift, researchers can test whether the changes share a cause. The recent Polish-English study directly investigated their unequal responsiveness to brief re-immersion. Marecka et al. (2026).
Neural reorganization: promising but not settled
Reviews propose testing whether attrition-related differences reflect reduced representation strength, retrieval difficulty, altered inhibitory control, or adaptive reorganization. These theories make distinct predictions, yet the existence of a neural difference between language groups does not by itself decide among them. Neurobiological mechanisms of non-pathological attrition remain an active research area with substantial methodological uncertainty. Gallo et al. (2021).
How Do Researchers Measure Language Attrition?
The scientific definition requires a temporal claim: ability has changed from an earlier state. Ideally, researchers assess participants before the environmental change and repeat the same or comparable measures later. In practice, long-term studies are difficult, and many investigations compare current migrants with a reference group living in the origin-language environment. This cross-sectional approach can detect group differences but cannot prove that each migrant previously performed like the comparison group.
Longitudinal and cross-sectional designs
Longitudinal designs provide repeated observations, making it possible to estimate within-person change. They still require attention to practice effects from repeated testing, changing motivation, aging, and language contact outside the measured environment. Cross-sectional comparisons are faster and can reach larger populations, but they face selection differences, cohort effects, differences in education, and unequal baseline proficiency. Retrospective reports of 'how fluent I used to be' are valuable personal histories but have limited accuracy as quantitative baselines.
Vocabulary tasks and free speech
Picture naming, verbal fluency, lexical decision, word recognition, and spontaneous narration measure different aspects of word processing. A timed naming task assesses speed and success under production demands. Free speech offers ecological relevance but changes with conversational topic, interlocutor, emotional engagement, and sample length. Analyses of lexical diversity should control for those properties. Schmid and Jarvis explicitly demonstrated why controlled and naturalistic measures should be interpreted together. Schmid and Jarvis (2014).
Grammar and processing tasks
Acceptability judgments can be influenced by metalinguistic knowledge and schooling. Sentence completion and elicited imitation depend on processing load, working memory, and task familiarity. Natural production samples reveal preferred forms but provide fewer observations of rare structures. When these tasks disagree, the divergence can identify a difference between knowledge and deployment; it does not automatically show which test is the true measure of proficiency.
Recording language experience accurately
Researchers often use questionnaires about hours spent speaking, reading, listening, and writing in each language. These are informative but imperfect proxies. Exposure quality, interlocutor proficiency, communicative complexity, and the diversity of contexts also matter. Self-reports can show restricted variation or suffer from recall biases. Evidence that self-reported use does not predict a particular result should not be confused with a demonstration that language experience has no causal role. Schmid and Yilmaz (2021).
Choosing appropriate comparison groups
A person who uses two languages every day should not automatically be compared with someone who has spent a lifetime functioning in only one. Bilinguals can differ from monolinguals in naming speed or language selection even in the absence of sustained loss. Comparisons among monolingual speakers, bilingual non-migrants, migrants, and individuals with different language histories help identify which effects follow from bilingualism and which track prolonged changes in environmental exposure. Schmid and Köpke (2017); Schmid and Yilmaz (2021).
Can a Forgotten Language Be Recovered?
Recovery depends on what changed in the first place. If an ability became harder to retrieve while much underlying knowledge remained, renewed exposure may restore rapid access. If early learning was incomplete, additional acquisition may be necessary. When new pronunciation or grammatical habits have developed, adjustment can require repeated practice over time. Science supports domain-specific recovery and relearning, not a universal guarantee of complete restoration.
What happens after re-immersion?
Re-immersion means returning to an environment where a language becomes highly present in daily life. It can change the frequency and necessity of comprehension and production. In the 2026 experimental study of Polish-English bilingual migrants, a short stay in Poland was associated with more native-like Polish consonant production, while perceptual differences remained. This suggests that performance components can recover on distinct timescales and provides a research framework for separating production from perception. Causal interpretation is still constrained by group comparisons, sample composition, and the particular consonant contrast assessed. Marecka et al. (2026).
Relearning versus learning from scratch
When a former learner recognizes words and rules but cannot use them fluently, relearning can build on residual representations. Productive practice creates opportunities to strengthen routes from intended meaning to linguistic form, while comprehension exposure refreshes familiar input patterns. The benefit of earlier learning varies by proficiency, interval, linguistic domain, and practice schedule. Long-term foreign-language memory research supports the possibility of durable knowledge after substantial disuse. Bahrick (1984).
Why recovery can look uneven
A familiar accent may return quickly during conversation with childhood friends, while formal writing remains difficult because the speaker has not written in that language for years. Everyday vocabulary can reappear before specialist terminology. Receptive grammar may support comprehension before the person produces complex structures with automaticity. Meaningful recovery studies therefore test several abilities across multiple timepoints. A single successful conversation, though personally important, is not a comprehensive proficiency measurement.
How to Maintain Language Skills Through Everyday Use
Research suggests that maintenance is best approached as keeping multiple kinds of linguistic access active. Speaking, listening, reading, and writing engage overlapping but partially different operations. People interested in retaining conversation skills need opportunities for spontaneous speech; those interested in academic literacy also need extensive reading and writing. The goal is to preserve functional access in the situations where the language is needed. Recommendations below are educational applications of learning and attrition research, rather than a clinically validated protocol that guarantees prevention of loss.
Preserve meaningful interaction
Regular conversation requires speakers to retrieve meanings, select words, construct sentences, and respond to unpredictable turns. Interactions with people who genuinely need the language can provide richer retrieval demands than passively hearing familiar background media. The content and relationship matter: talking about current events, family memories, and unfamiliar topics can diversify the language being used. Research directly comparing precise maintenance schedules across all language pairs remains limited.
Practice retrieval, not exposure alone
Recognition and production differ, so relying entirely on passive reading may leave spontaneous recall underpracticed. Producing a summary without a script, explaining a concept, or recalling vocabulary before seeing the answer increases the demands on active retrieval. These strategies align with memory research, while exact claims about their effectiveness for preventing native-language attrition require dedicated studies. Structured practice can combine receptive input with authentic production.
Space practice across time
Spaced relearning has empirical support in foreign-language vocabulary research. Bahrick and colleagues followed word-pair learning over years and observed substantial long-term retention advantages associated with distributed schedules. The evidence provides a principled reason to favor recurring practice over one concentrated burst, but it cannot specify an ideal interval for every language, goal, or learner. Bahrick et al. (1993).
Maintain diverse domains
Someone who speaks a heritage language only at family meals may retain those expressions while lacking current vocabulary for work, healthcare, government services, or technology. Reading contemporary journalism, discussing a wide range of subjects, and writing in different registers can broaden linguistic access. These activities also involve acquiring unfamiliar new terminology. Missing words for new social practices are an acquisition issue as well as a maintenance issue.
Keep realistic goals and measure progress
A practical plan can specify concrete activities and outcomes: comfortable video calls with relatives, reading a novel without constant translation, writing professional correspondence, or participating in spontaneous debates. Track several skills rather than treating speaking speed as the sole measure of competence. Motivation and available time influence sustainability. A learner can reasonably prioritize the contexts that matter most instead of trying to maintain identical proficiency in every register of every language.
Multilingualism, Identity, and the Meaning of Language Change
Languages carry cultural knowledge, memories, relationships, and ways of participating in social life. People who notice changes in a childhood language may experience sadness, disconnection, or frustration, particularly when language was central to belonging. These experiences deserve serious attention as personal realities. Yet an individual's sense that the language has changed is not an objective measurement of the amount or type of psycholinguistic attrition. Scientific assessment and lived significance answer related but different questions.
Multilingual competence is often distributed across contexts. A person may reason fluently about science in a later-acquired language, express family memories most comfortably in an earlier one, and use another language for community life. This functional differentiation is a normal feature of multilingual experience. Evaluating every language against maximal monolingual performance in every domain can distort the interpretation of change.
Language contact can also create new expressive possibilities. Borrowing, switching, and blending resources may serve communicative purposes. A person who alternates between languages or temporarily borrows a word has not thereby demonstrated a failing language system. Identifying genuine attrition requires longitudinal evidence or carefully controlled group comparisons, not a judgment that multilingual speech departs from an imagined pure standard. Schmid and Köpke (2017).
Distinguishing Attrition From Clinical Language Difficulties
Non-pathological language attrition is studied separately from acquired brain injury, neurodegenerative disorders, and developmental language conditions. Aphasia concerns acquired language impairment associated with neurological injury or disease. Developmental language disorder involves persistent developmental language difficulties assessed against appropriate clinical criteria. Dyslexia concerns difficulties in word-level reading and spelling, and stuttering concerns disruptions in speech fluency. These conditions cannot be identified or excluded by observing that a bilingual person has difficulty retrieving a word in an infrequently used language.
A change that follows years of reduced opportunity to use one language and is limited to that language is a different research pattern from sudden, unexplained problems affecting everyday communication across languages. When someone experiences an abrupt new language difficulty, especially alongside facial weakness, confusion, or other neurological symptoms, urgent medical assessment is appropriate. Persistent worsening or functional impairment warrants professional evaluation that accounts for language history and all languages used. This clinical boundary protects against both unnecessary pathologizing of multilingualism and overlooking genuine illness.
Researchers investigating language attrition should not interpret all bilingual differences as deficits. Screening scores obtained in a weaker language may reflect unequal exposure, schooling, cultural familiarity, or test norms. Clinical decisions require qualified professionals and suitable multilingual assessment. The purpose of this article is to explain non-pathological psycholinguistic change and its evidence, not to diagnose individual symptoms.
What the Evidence Establishes and What Remains Uncertain
Established: long-term linguistic performance can change with language experience, and first- and second-language abilities differ in their stability and accessibility. Long-term foreign-language learning studies reveal durable retained information despite years without instruction. Experimental evidence documents measurable differences in lexical accessibility and phonetic production among bilingual groups. Bahrick (1984); Schmid and Jarvis (2014).
Supported but context-dependent: renewed immersion and practice can improve some previously changed performance measures. The magnitude and speed of recovery depend on linguistic domain, population, task, and duration. The 2026 study indicates different recovery profiles for speech production and perception rather than a single uniform reversal process. Marecka et al. (2026).
Limited evidence: a universal ranking of which grammar, vocabulary, pronunciation, and comprehension systems weaken first; a single optimal maintenance schedule; precise individual predictions based only on the number of years abroad; and general conclusions about all heritage speakers derived from one language pair or cohort. Instructed second-language attrition in particular still needs stronger prospective research. Schmid (2023).
Contested hypotheses: the relative causal contribution of disuse, cross-language competition, structural restructuring, and cognitive control; whether particular grammatical differences reflect erosion of representations or difficulties in access; and how age of acquisition shapes different trajectories. These questions require comparisons capable of separating correlated experiences. Gallo et al. (2021); Schmid and Yilmaz (2021).
Theoretical interpretation: dynamic-systems accounts model stable and changing language subsystems across time, while neurocognitive accounts investigate differences in activation and processing. These are explanatory frameworks whose predictions can be investigated; none constitutes a universal clinical test for language attrition. Schmid, Köpke, and de Bot (2013).
Frequently Asked Questions About Language Attrition
Can you forget your first language as an adult?
Adult first-language performance can change substantially, particularly in word retrieval, patterns of use, and pronunciation after prolonged immersion in another language. Deeply established knowledge often remains accessible even when fluent production becomes less automatic. The degree of change varies, and current evidence does not support a universal progression toward complete loss.
Why can I understand my childhood language but struggle to speak it?
Comprehension supplies external linguistic cues, whereas speaking demands internally generated words, sentence organization, and articulation. Receptive and productive abilities can therefore follow different trajectories. Unequal opportunities to hear and produce the language can further widen the gap.
How quickly can language attrition begin?
There is no scientifically established universal timeline. Immediate effects of switching language context can differ from persistent changes across months or years. Research must distinguish short-term changes in activation from enduring differences in proficiency or representation.
Does speaking a second language damage the first?
Bilingual experience changes how languages interact and can influence performance in an earlier language. Calling this 'damage' presupposes a clinical deficit and misrepresents ordinary plasticity. The research examines specific linguistic differences and mechanisms, including competition and reduced accessibility, without treating multilingualism as pathology.
Is a strong foreign accent evidence of language attrition?
Accent differences can reflect cross-language influence, long-term phonetic learning, style, audience, and variation in language exposure. Some phonetic differences fall within attrition research, but accent alone cannot establish that a person has lost native-language knowledge.
Can I recover a language I stopped speaking for twenty years?
Earlier learning can support substantial recognition and relearning even after long disuse. The amount of recovery depends on earlier proficiency, later exposure, linguistic domain, age, and opportunities for practice. The presence of retained comprehension is encouraging evidence of available knowledge, but complete restoration cannot be promised.
Is language attrition related to dementia or aphasia?
Language attrition in psycholinguistic research describes non-pathological change related to language histories and environments. Dementia and aphasia involve different clinical mechanisms. Sudden language loss or progressively worsening daily communication deserves medical attention rather than an assumption of ordinary attrition.
Is it better to read, listen, or speak to maintain a language?
Different activities maintain different aspects of access. Listening and reading provide comprehensible input, while speaking and writing require production and retrieval. A varied routine aligned with the skills a person wants to retain is more defensible than assuming one activity protects all language functions.
Does code-switching mean that two languages are being forgotten?
Code-switching is an established feature of multilingual interaction and may reflect audience, topic, pragmatic goals, and lexical availability. Its occurrence alone does not identify attrition. Researchers need separate measures of earlier and current performance to make a claim about loss.
Are heritage-language speakers necessarily experiencing attrition?
No. Their language profiles may reflect differences in acquisition opportunities, literacy, schooling, language dominance, later use, and possible attrition. Determining attrition requires evidence of an ability previously acquired and subsequently changed. A current difference from monolingual peers does not establish that history.
Related Articles
Cluster Language, Thought and Psycholinguistics — the broader scientific foundation for language processing, acquisition, multilingual cognition, and linguistic experience.
Semantic Memory: Facts, Meaning, and Knowledge — conceptual knowledge and how stored meaning relates to retrieval.
Semantic Priming: How Related Words Influence Processing — automatic and context-sensitive activation of related words.
Speech Perception: How the Mind Turns Sounds Into Words — perception of speech sound categories and spoken-word recognition.
Language Processing in the Brain: Cognitive Networks and Neural Systems — the neural networks that support linguistic representation, access, and processing.
Bilingual Language Processing: How Two Languages Interact in the Mind — cross-language activation, lexical competition, and bilingual language control.
Second Language Acquisition: Cognitive Processes and Learning — learning an additional language and the conditions that shape later maintenance.
Critical Period Hypothesis: Age and Second Language Learning — age, sensitive periods, and the distinction between acquisition and later attrition.
Cross-Linguistic Influence: How a First Language Shapes New Language Learning — transfer and interaction across vocabulary, phonology, and grammar.
References
Bahrick, H. P. (1984). Semantic memory content in permastore: Fifty years of memory for Spanish learned in school. Journal of Experimental Psychology: General, 113(1), 1–29. https://doi.org/10.1037/0096-3445.113.1.1.
Bahrick, H. P., Bahrick, L. E., Bahrick, A. S., and Bahrick, P. E. (1993). Maintenance of foreign language vocabulary and the spacing effect. Psychological Science, 4(5), 316–321. https://doi.org/10.1111/j.1467-9280.1993.tb00571.x.
Gallo, F., Bermudez-Margaretto, B., Shtyrov, Y., Abutalebi, J., Kreiner, H., Chitaya, T., Petrova, A., and Myachykov, A. (2021). First language attrition: What it is, what it isn't, and what it can be. Frontiers in Human Neuroscience, 15, 686388. https://doi.org/10.3389/fnhum.2021.686388.
Marecka, M., Wrembel, M., Casado, A., Sorace, A., and Wodniecka, Z. (2026). Can language attrition be reversed: Evidence from speech perception and production. Cognition, 275, 106571. https://doi.org/10.1016/j.cognition.2026.106571.
Montrul, S., Foote, R., and Perpiñán, S. (2008). Gender agreement in adult second language learners and Spanish heritage speakers: The effects of age and context of acquisition. Language Learning, 58(3), 503–553. https://doi.org/10.1111/j.1467-9922.2008.00449.x.
Schmid, M. S. (2016). First language attrition. Language Teaching, 49(2), 186–212. https://doi.org/10.1017/S0261444815000476.
Schmid, M. S. (2023). The final frontier? Why we have been ignoring second language attrition, and why it is time we stopped. Language Teaching, 56(1), 73–93. https://doi.org/10.1017/S0261444822000301.
Schmid, M. S., and Jarvis, S. (2014). Lexical access and lexical diversity in first language attrition. Bilingualism: Language and Cognition, 17(4), 729–748. https://doi.org/10.1017/S1366728913000771.
Schmid, M. S., and Köpke, B. (2017). The relevance of first language attrition to theories of bilingual development. Linguistic Approaches to Bilingualism, 7(6), 637–667. https://doi.org/10.1075/lab.17058.sch.
Schmid, M. S., Köpke, B., and de Bot, K. (2013). Language attrition as a complex, non-linear development. International Journal of Bilingualism, 17(6), 675–682. https://doi.org/10.1177/1367006912454619.
Schmid, M. S., and Yilmaz, G. (2021). Lexical access in L1 attrition—Competition versus frequency: A comparison of Turkish and Moroccan attriters in the Netherlands. Applied Linguistics, 42(5), 878–904. https://doi.org/10.1093/applin/amab006.