What Is the Critical Period Hypothesis in Second Language Acquisition?
The critical period hypothesis (CPH) proposes that the ease or eventual outcome of language learning changes systematically with biological development. In second language acquisition, the central question is whether beginning substantial exposure to another language during childhood provides opportunities that become less available with increasing age. The evidence establishes a robust relationship between age of acquisition and several long-term outcomes, particularly pronunciation and some aspects of grammar. It does not establish a universal birthday after which a new language cannot be learned. The most useful scientific account specifies which ability is being measured, when meaningful exposure began, how much experience followed, and which social and cognitive conditions accompanied learning. Johnson and Newport (1989); Hartshorne, Tenenbaum, and Pinker (2018); Hartshorne (2022).
Age influences the probability of achieving particular forms of second-language performance. Early starters often attain more native-like long-term pronunciation and grammatical performance in immersion settings. Adolescents and adults can learn rapidly, use sophisticated study strategies, develop rich vocabulary and professional proficiency, and continue improving over many years. These observations describe different aspects of learning: initial rate, long-run attainment, accent, grammatical sensitivity, communicative effectiveness, and the ability to use language in a specific social environment. Treating them as a single score obscures the actual research question. Snow and Hoefnagel-Höhle (1978); Muñoz and Singleton (2011).
Critical periods, sensitive periods, and age effects
A critical period, in its strongest form, is a bounded developmental interval in which a particular experience must occur for a function to develop normally. A sensitive period is a phase of heightened responsiveness to experience; learning remains possible afterward, although its mechanisms, efficiency, or likely outcomes may change. An age effect simply means that a measured outcome varies with age. An age effect by itself does not demonstrate either a critical or a sensitive period. To support a specific biological window, investigators must show an age-related pattern consistent with the proposed mechanism and distinguish it from changing opportunities, accumulated language experience, educational access, and selection effects. Vanhove (2013); Norrman (2024).
Second language (L2) refers here to any additional language learned after the first language (L1), whether learned through immigration, family interaction, school, online communication, or deliberate study. This usage does not imply that only two languages are involved. Age of onset may mean age at first lessons, first meaningful input, arrival in a new linguistic community, or first sustained daily use. These are different events. A ten-year-old who receives one lesson weekly and a ten-year-old who begins living in a language-rich community share a calendar age but have distinct learning environments.
What the hypothesis predicts—and what it does not
A strong maturational version predicts that a particular change in biological learning capacity produces a discontinuity or marked change in the age–outcome relation, potentially with a definable endpoint. A softer sensitive-period version predicts a gradual, domain-dependent change in responsiveness, possibly involving overlapping developmental windows. Experience-based accounts predict substantial age patterns even without a fixed biological cutoff, because age is associated with changing input, prior-language entrenchment, motivation, identity, learning strategies, and everyday demands. Contemporary research considers combinations of these mechanisms. Hartshorne (2022); Norrman (2024).
The CPH is a theoretical explanation of developmental learning patterns. It is not a clinical diagnosis, a screening tool, a statement about intelligence, or a measure of an individual's potential. An adult's pronunciation or grammatical errors are ordinary features of additional-language development unless assessed in a separate clinical context for independent reasons.
Historical Origins and the Development of the Debate
Eric Lenneberg's 1967 book Biological Foundations of Language situated language development within biological maturation and the interaction between organism and environment. Subsequent research often translated this developmental proposal into a sharper question: does puberty mark an endpoint for the effortless acquisition of language? Recent scholarship revisits the assumptions behind that translation. Norrman argues that an epigenetic, developmentally interactive reading of Lenneberg is more faithful to his broader account than treating the critical period as a simple biological switch. This is a conceptual reassessment of the hypothesis, rather than an experiment establishing a particular age threshold. Norrman (2024).
A pivotal empirical test came from Johnson and Newport, who examined English grammaticality judgments in speakers who had immigrated to the United States after growing up with Korean or Chinese. Age of arrival was strongly related to performance among earlier arrivals, and the authors interpreted the age-related profile as evidence consistent with maturational constraints extending into second-language acquisition. The study helped establish the importance of long-term attainment rather than short-term classroom progress, while also revealing why age of arrival and age of meaningful English exposure require careful interpretation. Johnson and Newport (1989).
Later investigations reproduced some age-related trends but differed over their shapes, predictors, and explanations. Birdsong and Molis, using a replication with Spanish-speaking learners, observed continued age-related decline beyond a proposed childhood endpoint and some late learners approaching native-language performance. DeKeyser studied the possibility that learners increasingly rely on explicit analytic resources after childhood, while Flege and colleagues examined phonological and grammatical outcomes alongside use and language background. The historical literature therefore supports a sustained research program with competing explanatory models, rather than one universally accepted breakpoint. Birdsong and Molis (2001); DeKeyser (2000); Flege, Yeni-Komshian, and Liu (1999).
The Most Important Distinction: Learning Rate Versus Ultimate Attainment
The rate of learning describes improvement over a specified interval: for example, how much grammar an adult and a child acquire after the same number of lessons. Ultimate attainment refers to performance after extensive exposure and practice, typically when improvement has slowed, although identifying a genuine plateau is itself difficult. These measurements answer different questions. Older learners may outperform children during early stages because they bring developed attention, literacy, working strategies, world knowledge, and metalinguistic awareness to the task. Younger learners may have an advantage after many years of immersive, socially distributed exposure. Snow and Hoefnagel-Höhle (1978); Muñoz and Singleton (2011).
A child who starts at age five and is tested at age twenty has experienced up to fifteen years of potential contact. An adult who begins at age twenty and is tested at age twenty-five has experienced five. Comparing the two scores confounds starting age with available duration of learning. Matching years of exposure helps but is incomplete if their exposure differs in conversational quality, literacy demands, instruction, professional necessity, or social inclusion. Researchers therefore need longitudinal comparisons, carefully matched cohorts, and models that represent age, duration, intensity, and context jointly. Hartshorne, Tenenbaum, and Pinker (2018); Hartshorne (2022).
A claim such as “children learn languages faster” is consequently underspecified. Faster on which task, under which exposure conditions, and at what interval? A more defensible statement is that children and adults exhibit different profiles of learning advantages across time and linguistic domains.
Landmark Evidence: Age and Long-Term Grammar Performance
Johnson and Newport: age of arrival and English grammaticality
Johnson and Newport's 1989 study asked learners to judge English sentences containing grammatical constructions with differing relationships to their first languages. Performance varied systematically with age of arrival, especially among people who arrived younger. This provided important evidence that age of substantial exposure is associated with the eventual acquisition of morphosyntactic knowledge. The study also raised generalizable methodological issues: immigrant samples are not randomly assigned to ages of arrival, grammaticality judgment is a particular measurement task, and age of arrival may stand in for many connected experiences. Its findings are influential evidence of an age pattern, not by themselves a complete identification of its cause. Johnson and Newport (1989).
Birdsong and Molis: what happens after adolescence?
Birdsong and Molis tested Spanish-speaking learners with methods designed to revisit the earlier study. They observed age-related attainment differences extending into later starting ages, along with some near-native performances. Their results complicate an account that predicts a straightforward early decline followed by a uniform adult plateau. Differences in first-language structure, language use, recruitment, and testing also underline that an age function estimated in one population cannot automatically be generalized to every language pair. Birdsong and Molis (2001).
Hartshorne, Tenenbaum, and Pinker: the 669,498-participant study
A widely discussed analysis examined grammar performance from 669,498 native and nonnative English speakers and fitted a computational model separating age at testing, age of first exposure, and accumulated experience. The authors estimated that the underlying capacity to learn grammar remained relatively strong until approximately age 17.4 and then declined. This number is a model-dependent estimate of a transition in grammatical learning ability in the analyzed data. It is neither a universal developmental deadline nor a claim that someone turning eighteen suddenly loses the capacity to acquire English. The model links an observed performance distribution to a proposed latent learning trajectory, and its interpretation depends on assumptions about experience and the sampled population. Hartshorne, Tenenbaum, and Pinker (2018).
An age near 17.4 describes the estimated onset of decline in a particular statistical model. A learner's ultimate probability of resembling a native speaker depends partly on how much additional learning remains possible and how much exposure follows. The distinction between the age at which learning rate changes and the age beyond which near-native attainment becomes less likely is crucial. Online grammatical assessments also represent a selected form of competence: they do not comprehensively measure conversational skill, accent, writing, comprehension under noise, or meaningful participation in social life.
Meta-analytic evidence and the significance of moderators
Qureshi's meta-analysis combined group-comparison and correlational evidence about second-language morphosyntax. It reported an overall age-related pattern, with effects that varied according to whether learning occurred in foreign-language or second-language contexts, the timing of tests, and whether linguistic performance was measured aurally or in writing. These moderators matter because estimates of age effects change with how researchers operationalize learning and outcome. Meta-analysis strengthens the case that the association is broader than one famous experiment; it does not establish that all included associations derive from the same biological process. Qureshi (2016).
The strongest conclusion across these studies is that starting age and subsequent language outcomes are related across multiple settings. The exact age at which a biologically defined transition occurs, whether there are multiple windows, and how much of the association is attributable to experience remain actively debated.
Pronunciation, Grammar, Vocabulary, and Comprehension Have Different Age Profiles
Speech sounds and foreign accent
Speech perception and production require learners to map variable acoustic input onto categories that are already organized by prior linguistic experience. A new language may distinguish sounds that an existing language treats as variations of one category. Older beginners therefore face the task of modifying well-practiced speech perception and production routines. Flege, Munro, and MacKay reported strong relationships between age of learning and foreign accent judgments in experienced learners. Flege, Yeni-Komshian, and Liu documented age-related differences in English pronunciation and morphosyntax among Korean-speaking immigrants, illustrating why a single critical-period story may fail to account equally for all dimensions. Flege, Munro, and MacKay (1995); Flege, Yeni-Komshian, and Liu (1999).
An accent is a phonetic characteristic of speech. It is not identical to low comprehensibility, and it does not reveal the depth of a person's vocabulary, conceptual reasoning, or social competence. Munro and Derwing demonstrated that heavily accented speech can remain highly intelligible. Language education focused on effective communication should consequently assess intelligibility and comprehensibility directly rather than equating native-like accent with successful language learning. Munro and Derwing (1995).
Morphosyntax and implicit generalization
Grammar learning involves extracting recurring patterns from input, forming representations of word order and inflection, and applying these patterns in unfamiliar contexts. Age-related performance has been detected on tasks involving judgments about grammatical sentences. DeKeyser examined the possibility that younger learners can acquire some complex patterns more implicitly, whereas many adults bring explicit problem-solving and rule-based learning to the same task. This is a mechanistic hypothesis: differences in a test score do not demonstrate that a person used one learning system rather than another unless the study independently measures learning processes. DeKeyser (2000).
An adult may learn a grammatical explanation efficiently and apply it accurately in an untimed exercise while showing more difficulty during spontaneous conversation, when speech planning and comprehension impose processing demands. Conversely, extensive practice can make initially explicit knowledge available more fluently. Timed judgments, spontaneous production, comprehension accuracy, and neural or behavioral measures need not yield identical estimates of learning. Qureshi (2016).
Vocabulary, meaning, and pragmatic participation
Vocabulary continues to be acquired throughout adulthood. Adults can use established concepts, reading experience, professional domains, and deliberate memory strategies to map new words onto existing semantic networks. Yet vocabulary is heterogeneous: learning a definition differs from mastering subtle collocations, frequency-sensitive idioms, social registers, and real-time lexical access. Granena and Long compared phonological, lexical, and morphosyntactic attainment in the same learner groups and reported differing age-related patterns across domains. Their sample and methods support domain differentiation, while the exact timing of proposed sensitive periods requires independent replication and theoretical testing. Granena and Long (2013).
Pragmatic competence adds further complexity. Knowing how to interrupt politely, recognize humor, adapt formality, and negotiate shared meaning depends on extended social participation as well as linguistic structure. Someone can have an accent, highly accurate vocabulary, and excellent pragmatic skill simultaneously. A meaningful account of long-term second-language attainment must respect that multidimensional profile.
Why Age Might Matter: Competing Cognitive and Developmental Mechanisms
Experience-dependent neural plasticity
Developing neural systems are shaped by sensory and linguistic experience. Early exposure can stabilize efficient categories, mappings, and patterns that guide later learning. This provides a biologically plausible rationale for heightened developmental sensitivity without identifying a single switch that closes at puberty. Adult brains also undergo experience-dependent change, and successful adult language learning is direct behavioral evidence that learning remains possible. Neural plasticity and experience influence each other: what is practiced alters processing routines, and established routines alter which distinctions are easy to notice. Mayberry and Lock (2003); Norrman (2024).
To claim that a measured age effect is caused by reduced plasticity, researchers need more than an age-proficiency correlation. They need convergent evidence connecting developmental change in brain organization to task-specific learning performance while considering exposure and preexisting language. Brain imaging differences between early and late bilinguals can reflect years of practice, strategy, task demands, proficiency, and developmental history as well as age itself. Causal attribution requires careful designs and does not follow from a visually striking activation map.
Entrenchment of the first language
A first language provides an established phonological inventory, grammatical expectations, and highly practiced lexical associations. This prior knowledge is useful for learning a related language and can interfere when the new language organizes contrasts differently. Some apparent age disadvantages may therefore arise from the cumulative strength of earlier learning rather than the loss of a general biological capacity. The experience-based account predicts that language pair, similarity, and particular training conditions should moderate outcomes. Observed differences between studies with Korean-, Chinese-, Spanish-, and Italian-speaking learners are relevant to this prediction, although they cannot be attributed to language similarity alone. Birdsong and Molis (2001); Flege, Yeni-Komshian, and Liu (1999).
Implicit and explicit learning resources
Children learn much language through repeated participation in everyday situations, identifying distributional regularities without necessarily explaining them. Adults can also learn implicitly but additionally draw on developed analytical reasoning, literacy, self-monitoring, and instructed practice. A shift in the relative usefulness of these resources can change the shape of learning curves. DeKeyser's fundamental-difference account proposes that adult attainment in particular circumstances is related to the availability of explicit analytic abilities. Empirical patterns are compatible with multiple versions of this idea, and a neat division into “implicit children” and “explicit adults” would go beyond the evidence. DeKeyser (2000).
Social integration, opportunity, and motivation
Children entering school may need to use the community language daily across classes, friendships, play, and informal interaction. Adults may have jobs and family networks organized around their first language, fewer sustained opportunities for low-stakes conversation, or greater consequences for making mistakes. Alternatively, adults studying for professional or personal reasons can show exceptional persistence and highly strategic engagement. These changing conditions are related to starting age and can influence exposure quality and practice intensity. Muñoz and Singleton emphasize the explanatory value of considering age together with a broader constellation of social, psychological, and methodological variables. Muñoz and Singleton (2011).
These mechanisms can operate jointly. A developmental change in perceptual sensitivity may coexist with strong first-language habits and differences in educational opportunity. The scientific aim is to estimate their separate and interacting contributions, rather than assume that the presence of one excludes the others.
First-Language Deprivation and Second-Language Learning Are Different Questions
Studies of delayed access to a first fully accessible language provide a distinctive test of developmental sensitivity. Mayberry and Lock compared adults with different histories of early linguistic experience, including deaf and hearing participants who acquired spoken or signed languages. Adults with early access to a first language often performed strongly when subsequently learning another language, whereas some adults who lacked accessible linguistic input in early childhood showed lasting difficulties. These results suggest that early first-language experience helps organize capacities subsequently available for additional-language learning. Mayberry and Lock (2003).
This evidence should be interpreted with precision. Learning an additional language after an established first language is a different developmental situation from receiving a first accessible language late. The first case involves adapting an existing linguistic system; the second involves building foundational linguistic representations after early developmental opportunities have passed. Studies of sign languages are important because they demonstrate that language development concerns accessible linguistic structure across modalities, not merely the sounds of spoken language. They also remind researchers that educational access and a supportive language environment have enduring consequences. No inference about the intelligence or worth of deaf people follows from these findings.
How Researchers Test a Critical Period
Age-at-onset curves and discontinuity tests
Investigators commonly plot a language outcome against age at first significant exposure. A smooth decrease with age establishes an association but does not uniquely identify a critical period. A sharp change in slope at a particular age is closer to the prediction of some critical-period formulations, yet it requires confidence intervals, robustness checks, and comparison with alternative curves. Vanhove showed that statistical methods used in some influential tests did not always match the discontinuity predicted by the theory; piecewise modeling produced findings that challenged the cross-linguistic robustness of specific breakpoint claims. Methodological disagreement about model choice is therefore part of the scientific substance of the debate. Vanhove (2013).
A plausible age curve can arise from cohort effects, unequal exposure, measurement ceilings, or sampling of unusually motivated adult learners. The strongest analyses test competing model families, examine out-of-sample predictions, avoid treating a fitted breakpoint as a discovered biological constant, and report uncertainty around estimated ages.
Cross-sectional, longitudinal, and natural-experiment designs
Cross-sectional studies compare people of different starting ages at one time. They provide access to broad distributions but blend together differences in cohorts, schooling, motivation, and immigration histories. Longitudinal studies follow learners as they acquire a language; they reveal trajectories and changes in exposure but often require years of research, experience attrition, and cannot randomly assign individuals to grow up at different ages. Natural experiments, migration contexts, and educational policy differences supply valuable contrasts while introducing selection and contextual confounders. Hartshorne (2022); Muñoz and Singleton (2011).
What proficiency tests actually measure
Grammaticality judgments test sensitivity to particular sentence structures and may favor literate participants who are comfortable with explicit reflection. Accent judgments depend on listeners, stimuli, and rating criteria. Conversation tasks assess real-time coordination and interaction; reading tests add literacy requirements; comprehension tasks in noise incorporate auditory and attentional demands. A study that establishes an age effect for written grammatical judgments has not thereby demonstrated the same effect for oral negotiation, meaning construction, or all components of language ability. This is why the question “Can adults become fluent?” cannot be answered solely by a single grammar score.
The difference between accentedness, comprehensibility, and intelligibility is a particularly important measurement lesson. These correlated dimensions can diverge in individual speakers, so a socially meaningful outcome should be measured directly. Munro and Derwing (1995).
Large-scale online testing
Online experiments can recruit enough participants to distinguish subtle age patterns and compare models that smaller samples cannot reliably separate. Hartshorne and colleagues used this opportunity to study grammar across a large range of starting ages. At the same time, online samples are self-selected; testing conditions and language histories are reported with varying precision, and very large samples can provide narrow statistical uncertainty around estimates that still depend on imperfect assumptions. Sample size reduces random error but does not automatically remove selection bias or construct validity problems. Hartshorne, Tenenbaum, and Pinker (2018); Hartshorne (2022).
Is There a Biological Cutoff at Puberty, Age Seventeen, or Age Eighteen?
There is no established universal cutoff that applies to all second-language abilities, learners, and learning environments. Puberty was prominent in historical formulations, but empirical studies have proposed different functional profiles, some showing gradual change and some interpreting model fits as evidence of transitions. The frequently quoted age of 17.4 years is the particular model estimate reported by Hartshorne and colleagues for grammar-learning ability. A pronunciation study, a vocabulary study, an intensive classroom intervention, and a naturalistic migration cohort need not produce the same age profile. Hartshorne, Tenenbaum, and Pinker (2018); Granena and Long (2013).
A biological developmental explanation remains scientifically plausible. Its precise formulation is contested because the data support several interacting mechanisms and because statistical definitions of the proposed boundary differ. Norrman's epigenetic reinterpretation emphasizes that biological development and linguistic experience co-construct the conditions under which learning occurs. Vanhove's statistical critique emphasizes that an age-related trend and a confirmed critical-period discontinuity are different evidential claims. These two interventions address different levels of the argument: developmental theory and quantitative inference. Norrman (2024); Vanhove (2013).
Do Adults Ever Reach Native-Like Proficiency?
Some learners who begin in adolescence or adulthood achieve very high performance on selected linguistic measures, sometimes approaching or falling within the distribution of native-language comparison groups. This challenges an absolute interpretation according to which adult learners are categorically excluded from advanced attainment. At the same time, native-like performance on one task does not imply indistinguishability across pronunciation, spontaneous syntax, lexical processing, sociolinguistic nuance, and every other task. Birdsong and Molis documented variability in late learners; the broader literature emphasizes careful sampling and multidimensional assessment. Birdsong and Molis (2001); Muñoz and Singleton (2011).
“Native-like” is also a research criterion that requires definition. Native-language speakers vary in vocabulary, literacy, pronunciation, register, and experience, and multiple native dialects are legitimate forms of a language. An appropriate comparison identifies the sampled population and measure instead of treating one prestige variety as the natural standard of all human language capacity. For practical purposes, successful language learning may mean participating effectively in work, education, relationships, creative writing, or civic life without being perceptually indistinguishable from a monolingual control group.
What the Evidence Means for Parents, Teachers, and Adult Learners
Early exposure is valuable when it is meaningful and sustained
Early access to rich, comprehensible, interactive language provides opportunities for practice and learning over a longer developmental span. This gives a plausible reason to support language-rich environments for children, including access to an appropriate signed language when necessary. Yet beginning classroom lessons a few years earlier does not automatically deliver the same benefit as sustained immersion. The educational outcome depends on instructional quality, frequency, meaningful interaction, continuity, and the learner's circumstances. In one study of primary foreign-language instruction, Wilden and colleagues found no significant receptive-proficiency advantage for earlier starters at the end of primary schooling, illustrating the importance of separating age from teaching conditions. Wilden, Porsch, and Schurig (2020); Mayberry and Lock (2003).
Adolescent and adult learning deserves ambitious expectations
Adults can draw on explicit explanations, reading, deliberate vocabulary practice, conversation, feedback, and domain-specific motivation. They may progress rapidly through structured instruction while developing pronunciation through targeted listening and production practice. In particular, choosing intelligibility and comprehensibility as speech goals is more relevant to communication than demanding elimination of every detectable accent. None of the evidence supports telling someone that an age threshold has eliminated their ability to become proficient. Snow and Hoefnagel-Höhle (1978); Munro and Derwing (1995).
Learning plans should be based on a diagnostic understanding of goals in the educational sense: current reading and listening ability, pronunciation distinctions that interfere with understanding, grammar needed for the learner's writing, and opportunities for real-world practice. This does not imply a medical diagnosis. A learner preparing for graduate study needs different language experiences from someone developing conversational fluency or interpreting specialized technical material.
Language policies should follow evidence, not an age slogan
Policy proposals for earlier instruction are meaningful only alongside their instructional model, available teacher expertise, continuity through schooling, and opportunities for use. The age-effect literature supports taking early exposure seriously while resisting a simplistic equation between starting earlier and learning better in every setting. Curriculum evaluation should assess outcomes after comparable exposure, consider the distribution of benefits and costs, and recognize that adult language education remains essential. Muñoz and Singleton (2011); Qureshi (2016).
Scientific Evidence: What Is Established, Limited, and Contested?
Established findings
Across a substantial body of observational and experimental research, earlier meaningful exposure is associated with higher average long-term outcomes on several second-language measures. Age effects differ by outcome and testing context. Older learners can outperform younger learners over short learning intervals. Successful adult language learning is common, and accent strength is not interchangeable with intelligibility. These statements are supported by several converging lines of research and quantitative synthesis. Johnson and Newport (1989); Qureshi (2016); Munro and Derwing (1995).
Supported but domain-dependent interpretations
Differential sensitivity in phonology, grammar, and vocabulary is supported by comparative studies, while proposed ages of transition are less stable across research designs. Individual learning profiles reflect interactions among prior-language knowledge, exposure, social context, and potentially maturational mechanisms. Direct estimates of the independent contribution of each factor remain challenging. Granena and Long (2013); Hartshorne (2022).
Contested hypotheses and theoretical interpretation
A single universal post-pubertal cutoff, an exclusively biological explanation of all age effects, and one fixed division between implicit child learning and explicit adult learning are contested formulations. The 17.4-year estimate is a substantive model result that requires domain-appropriate interpretation rather than treatment as a universal law. The balance among maturational change, experience, first-language entrenchment, and social opportunity remains open to empirical refinement. Vanhove (2013); Norrman (2024); Hartshorne, Tenenbaum, and Pinker (2018).
What stronger future research would need to establish
A decisive account should test explicit, domain-specific predictions with multiple language pairs, repeated measurement, consistent operational definitions of onset and exposure, preregistered comparisons of continuous and segmented age curves, and representative samples of late and early learners. It should distinguish cognitive mechanisms by more than self-report or a single test and report uncertainty about estimated developmental transitions. Longitudinal measurements and triangulation across behavioral, educational, and neural evidence would clarify how different mechanisms contribute. This is an assessment of methodological needs, not a claim that any one design can remove all confounding.
Frequently Asked Questions
At what age does the critical period for language learning end?
Research has not identified one universal age for the end of successful second-language learning. Different studies propose different patterns for phonology and grammar. A model based on a very large online grammar dataset estimated a change in learning ability around 17.4 years, while other analyses and domains yield different interpretations. Hartshorne, Tenenbaum, and Pinker (2018); Vanhove (2013).
Is it too late to learn a second language after eighteen, thirty, or sixty?
Age affects typical pathways and some long-term probabilities, but people can learn languages throughout adulthood. The time available, quality of instruction, opportunity to communicate, prior languages, health, motivation, and individual learning experience influence results. The relevant question is which proficiency goals and learning conditions are realistic for a particular person.
Why do young children often acquire a native-like accent?
Early perceptual experience helps organize sound categories before they become highly practiced through another language. Later learners may need to reorganize established categories and articulation routines. Accent outcomes also depend on continued use and input. A noticeable accent can coexist with clear and effective communication. Flege, Munro, and MacKay (1995); Munro and Derwing (1995).
Do adults learn languages faster than children?
Adults and adolescents may make faster initial progress on some tasks, especially when instruction and exposure duration are comparable. Early starters may have an advantage in long-term attainment after sustained immersion. Research must specify whether it is measuring learning rate, ultimate proficiency, pronunciation, or another skill. Snow and Hoefnagel-Höhle (1978).
Does the critical period hypothesis prove that grammar cannot be mastered in adulthood?
No. Studies report age-associated differences in average performance and individual variability. Some late learners achieve exceptionally high performance on selected grammar tasks. A claim of impossibility would require evidence far stronger than observed average declines. Birdsong and Molis (2001).
Are critical periods identical for spoken and signed languages?
Early access to an accessible first language matters across spoken and signed modalities. The developmental consequences of delayed first-language exposure are distinct from the problem of learning an additional language after an established first language. Evidence from sign-language research provides important constraints on theories that equate language with speech alone. Mayberry and Lock (2003).
Should schools always start foreign-language lessons earlier?
The value of an early start depends on sustained, meaningful input, qualified instruction, continuity, and the intended outcome. A few early weekly lessons need not outperform later, more intensive and higher-quality teaching. Starting age is one component of an educational design, not a substitute for that design. Wilden, Porsch, and Schurig (2020).
Conclusion
The critical period hypothesis remains a central framework for studying how language learning changes across development. Research demonstrates consequential associations between age of meaningful exposure and long-term performance, with different profiles across speech sounds, grammar, and other linguistic abilities. The interpretation of these associations requires distinctions between learning rate and attainment, between native-like speech and communicative competence, between first-language development and additional-language learning, and between a statistical breakpoint and a biological mechanism. A scientifically productive account treats maturation, experience, established linguistic knowledge, and social context as interacting influences whose relative contributions must be tested. The practical implication is to value rich early language opportunities while recognizing the substantial and enduring capacities of adolescent and adult learners.
Related Articles
Psycholinguistics: How Language Is Processed in the Mind — the foundational account of real-time human language representation and processing.
Phonological Processing: Speech Sounds and Language Cognition — how speech-sound representations contribute to language cognition and learning.
Speech Perception: How the Mind Turns Sounds Into Words — the perceptual foundations relevant to new sound categories.
Language Processing in the Brain: Cognitive Networks and Neural Systems — networks and mechanisms supporting human language.
Psycholinguistic Research Methods: How Scientists Study Language Processing — experimental approaches to linguistic cognition and evidence quality.
References
Birdsong, D., and Molis, M. (2001). On the evidence for maturational constraints in second-language acquisition. Journal of Memory and Language, 44(2), 235–249. https://doi.org/10.1006/jmla.2000.2750
DeKeyser, R. M. (2000). The robustness of critical period effects in second language acquisition. Studies in Second Language Acquisition, 22(4), 499–533. https://doi.org/10.1017/S0272263100004022
Flege, J. E., Munro, M. J., and MacKay, I. R. A. (1995). Factors affecting strength of perceived foreign accent in a second language. The Journal of the Acoustical Society of America, 97(5), 3125–3134. https://doi.org/10.1121/1.413041
Flege, J. E., Yeni-Komshian, G. H., and Liu, S. (1999). Age constraints on second-language acquisition. Journal of Memory and Language, 41(1), 78–104. https://doi.org/10.1006/jmla.1999.2638
Granena, G., and Long, M. H. (2013). Age of onset, length of residence, language aptitude, and ultimate L2 attainment in three linguistic domains. Second Language Research, 29(3), 311–343. https://doi.org/10.1177/0267658312461497
Hartshorne, J. K. (2022). When do children lose the language instinct? A critical review of the critical periods literature. Annual Review of Linguistics, 8, 143–151. https://doi.org/10.1146/annurev-linguistics-032521-053234
Hartshorne, J. K., Tenenbaum, J. B., and Pinker, S. (2018). A critical period for second language acquisition: Evidence from 2/3 million English speakers. Cognition, 177, 263–277. https://doi.org/10.1016/j.cognition.2018.04.007
Johnson, J. S., and Newport, E. L. (1989). Critical period effects in second language learning: The influence of maturational state on the acquisition of English as a second language. Cognitive Psychology, 21(1), 60–99. https://doi.org/10.1016/0010-0285(89)90003-0
Mayberry, R. I., and Lock, E. (2003). Age constraints on first versus second language acquisition: Evidence for linguistic plasticity and epigenesis. Brain and Language, 87(3), 369–384. https://doi.org/10.1016/S0093-934X(03)00137-8
Munro, M. J., and Derwing, T. M. (1995). Foreign accent, comprehensibility, and intelligibility in the speech of second language learners. Language Learning, 45(1), 73–97. https://doi.org/10.1111/j.1467-1770.1995.tb00963.x
Muñoz, C., and Singleton, D. (2011). A critical review of age-related research on L2 ultimate attainment. Language Teaching, 44(1), 1–35. https://www.cambridge.org/core/journals/language-teaching/article/abs/critical-review-of-agerelated-research-on-l2-ultimate-attainment/6030287858C91932F2A594E805BD23C0
Norrman, G. (2024). Reconceptualizing the critical period hypothesis for second language acquisition: An appraisal of Lenneberg's work on the epigenesis of language. Language Sciences, 105, 101645. https://doi.org/10.1016/j.langsci.2024.101645
Qureshi, M. A. (2016). A meta-analysis: Age and second language grammar acquisition. System, 60, 147–160. https://doi.org/10.1016/j.system.2016.06.001
Snow, C. E., and Hoefnagel-Höhle, M. (1978). The critical period for language acquisition: Evidence from second language learning. Child Development, 49(4), 1114–1128. https://eric.ed.gov/?id=EJ196662
Vanhove, J. (2013). The critical period hypothesis in second language acquisition: A statistical critique and a reanalysis. PLOS ONE, 8(7), e69172. https://doi.org/10.1371/journal.pone.0069172
Wilden, E., Porsch, R., and Schurig, M. (2020). An early start in primary EFL education and the role of teacher qualification and teaching quality. Language Teaching for Young Learners, 2(1), 28–51. https://doi.org/10.1075/ltyl.19002.wil