Vocabulary acquisition is the process through which children discover spoken or signed word forms, infer what they mean, connect them to concepts and communicative functions, and gradually make those connections accessible for understanding and speech or signing. Children learn new words by combining perceptual information, social interaction, statistical regularities, prior linguistic knowledge, memory, and repeated experiences in meaningful contexts. A child who correctly points to a novel object after hearing a new label has made a promising inference; evidence of durable word learning also requires remembering that association, recognizing it in a different situation, and using it appropriately. Research on early comprehension, ambiguous naming, caregiver interaction, and retention supports this multicomponent account. Bergelson and Swingley (2012); Horst and Samuelson (2008); Roembke et al. (2023).
Word learning is a central topic in developmental psycholinguistics because it connects what children hear with what they perceive, remember, categorize, and attempt to communicate. The fundamental scientific puzzle is referential ambiguity: in an ordinary scene, a new expression may refer to an object, a part, a property, an action, a relation, an absent event, or a speaker's intention. Children's achievements become intelligible when individual moments of learning are examined together with the accumulation of information across interactions. This article explains the evidence for those learning processes, how researchers measure them, why immediate success sometimes fails to persist, and how vocabulary develops across different children, languages, and environments.
What Is Vocabulary Acquisition?
A word is more than a sound paired with an object
A lexical entry can connect an auditory or visual language form to semantic information, a grammatical category, morphological relationships, pragmatic expectations, and knowledge about the situations where the expression is useful. The noun puppy is related to dogs as a category, but it also carries information about relative age and the conversational purposes for which a speaker might choose it. Verbs such as borrow require understanding participants and relations. Words for feelings, time, causes, and beliefs depend substantially on linguistic context because their referents are often unavailable for direct visual inspection. Consequently, a theory adequate for a picture-label experiment still needs to explain the acquisition of actions, properties, relations, abstract words, and socially conventional meanings.
Receptive vocabulary describes words a child can understand under particular assessment conditions. Expressive vocabulary describes words a child can produce or intentionally sign. Knowledge is graded: a child can identify a familiar referent without being able to pronounce its name, recognize a word in a familiar routine but fail in a new context, or use an approximation before adult-like production develops. Researchers therefore examine recognition accuracy, reaction time, generalization, spontaneous use, comprehension in context, and retention at different delays. A single vocabulary count summarizes a much richer and changing lexical system.
The mapping problem and the learning problem
Mapping concerns the provisional association between a word form and a possible meaning. The immediate act of choosing a referent is called referent selection. Learning additionally requires encoding and retaining a usable representation, resolving competing meanings, integrating the word with existing knowledge, and extending it to appropriate new instances. Referent selection and retention are empirically separable. In controlled experiments, toddlers can select the unnamed object in response to a new word and then fail to retrieve the association after a brief delay. That observation changed how researchers interpret the phrase fast mapping. Horst and Samuelson (2008); Mather (2025).
Vocabulary acquisition also differs from the broader process of first language acquisition. Children simultaneously learn phonological categories, permissible word sequences, grammatical patterns, discourse routines, and communicative conventions. These domains interact, but this article specifically owns the question of how new words and their meanings are learned, differentiated, retained, and extended. Learning an entire sentence structure or acquiring native phoneme boundaries are independent scientific problems with connections to lexical learning.
How Early Does Word Knowledge Begin?
Understanding precedes many spoken first words
Infants demonstrate sensitivity to familiar word meanings before they can produce many recognizable words. In a looking-while-listening experiment, investigators show two pictures and have an adult name one. If infants systematically look toward the named picture, their behavior supports comprehension under those conditions. Bergelson and Swingley found evidence that infants between six and nine months recognized the referents of several common nouns. This result challenged the simple developmental picture in which conceptual word learning begins only after the first birthday; it did not establish that all young infants possessed adult-like concepts or understood every word they heard. Bergelson and Swingley (2012).
Early receptive understanding is especially relevant to familiar people, body parts, food, and objects repeatedly encountered in recurring routines. Familiarity makes several cues available at once: a relatively stable word form, a restricted set of likely referents, correlated gaze and actions, and a growing history of association. The resulting evidence is cumulative. A child may initially recognize a term only in a narrow context; later exposure supports more flexible understanding and more precise distinctions among neighboring concepts.
Why developmental trajectories vary
Children differ in hearing histories, language exposure, interactions, attention, temperament, linguistic communities, and measurement conditions. Counts obtained from caregiver checklists, eye tracking, or direct elicitation can diverge because the tasks ask different questions. Ages are useful for describing population-level patterns, but a single reported milestone does not reveal all the learning processes or the total resources available to a particular child. Early comprehension and production also develop on partly different schedules.
Longitudinal research links early word-recognition efficiency and later vocabulary growth, although its observational designs do not prove a fixed causal chain. In one small English-learning sample, differences in real-time language-processing efficiency and vocabulary were visible across socioeconomic groups by 18 months. The size, sampling, and cultural specificity of such studies matter when evaluating generality. Fernald, Marchman, and Weisleder (2013).
How Children Recognize and Encode New Word Forms
Discovering a word in continuous language
Speech arrives as a continuous acoustic stream with coarticulation, variable rates, different speakers, and few reliable pauses between words. To learn a word, children must discover which portions of the signal constitute a repeatable expression. Prosody, stress, phonotactic expectations, familiar words, and distributional regularities can all contribute. A learner who repeatedly hears the same sound sequence in different sentences gains evidence that it belongs together. Segmentation is therefore an important precursor to learning the meanings of unfamiliar spoken forms, although segmentation skill alone does not reveal what the word refers to.
Word-form learning involves recognizing meaningful similarity through variation. A familiar word spoken by a grandparent and by a preschool teacher can differ substantially in pitch, timing, and accent. A useful representation must support recognition across these surface changes without confusing phonologically close neighbors. Children's increasing vocabulary also changes the task: the learner must retain enough detail to distinguish a newly heard word from those already known. This is one reason accurate phonological encoding and semantic inference become progressively integrated.
Word forms, lexical neighborhoods, and processing efficiency
When children hear a new label that resembles familiar vocabulary, existing lexical representations can both help and compete with it. Familiar sound patterns support encoding; close neighbors increase the risk of confusion. For an adult, learning the form dax may be easier than learning an extremely unusual sequence, yet retaining dax as a distinct lexical item requires discriminating it from related forms. Consequently, learning experiments should distinguish whether errors arise from recognizing the word form, remembering its meaning, or selecting among objects.
Efficient lexical processing is related to language experience and vocabulary development. A child who can rapidly recognize a familiar word may have more opportunity during an utterance to track subsequent information, although individual differences in processing speed may reflect multiple causes. Observational correlations should not be interpreted as evidence that increasing recognition speed by itself will reproduce all the developmental advantages associated with richer language knowledge. Fernald, Marchman, and Weisleder (2013).
The Referential Ambiguity Problem
Why the visible scene does not specify a word's meaning
Imagine a caregiver saying, “Look at the blicket,” while holding a red wooden toy. The new term could indicate the whole toy, its handle, its color, the material, an action being performed, or membership in a familiar category. A single scene supplies multiple compatible interpretations. The child must combine linguistic grammar, visual attention, the speaker's actions, previous knowledge, and later experiences to narrow them. Controlled word-learning experiments create similar ambiguity on purpose by presenting several unfamiliar objects with a novel spoken label.
The level of ambiguity also depends on what the child already knows. A child who understands that the expression follows an article and occurs where object names often occur may assign greater initial probability to a noun interpretation. A learner with an established word for the whole object might consider a property or part. These are contextual inference strategies; they do not require a perfectly articulated conscious theory of reference.
Multiple cues usually operate together
Proposals emphasizing attention, associative statistics, social intentions, linguistic syntax, or conceptual constraints have sometimes been framed as competing accounts. The more informative question is how the cues are weighted in a particular learning situation and whether the processes implicated in initial choice also support later retention. A clear pointing gesture can direct attention; repeated label-object co-occurrence can increase statistical evidence; sentence structure can suggest a relational meaning; and memory can link these experiences across time. When cues conflict, performance reveals which information children use and under which conditions. Roembke et al. (2023).
This integration helps explain how children learn words in real life without receiving explicit definitions for each expression. Everyday interaction is often partly ambiguous but rich in redundancies. A word may be heard during play, at a meal, in a story, or in another speaker's conversation. The learner can accumulate constraints over many such events even when no single event is decisive.
Fast Mapping: What It Demonstrates and What It Does Not
The classic one-trial inference
Fast mapping usually describes an ability to identify a plausible meaning for a new word after very limited exposure, often by selecting an unfamiliar object when familiar objects already have known names. Such performance shows that children can make useful inferences rapidly. It is scientifically important because vocabulary learning could otherwise seem impossible given the vast number of words and limited number of explicit explanations children receive.
An illustrative laboratory task displays a spoon, a cup, and an unnamed object. Asked to give the experimenter “the toma,” a child may choose the object that has no familiar name. The child has solved a local referential puzzle. That response does not by itself establish that toma will be recognized tomorrow, that its meaning will generalize to a second example, or that its phonological representation is stable enough to support future speech.
Retention can be poor despite correct selection
Horst and Samuelson studied 24-month-olds using a selection task followed by a short delayed test. Across experiments, children were proficient at selecting referents but often retained the new associations poorly. Adding direct naming support improved retention in some conditions. This demonstration matters because success on an immediate pointing task and durable vocabulary acquisition are different dependent variables. Horst and Samuelson (2008).
A 2025 preregistered meta-analysis concentrated specifically on retention following mutual-exclusivity responses. It found a strong initial response tendency alongside a smaller retention effect, which was markedly attenuated when Bayesian publication-bias correction was applied. The synthesis reinforces the distinction between choosing a referent and forming a lasting lexical memory. The precise size and generality of retention effects depend on study designs and statistical assumptions; “children can fast-map” remains too broad a claim unless the outcome is named. Mather (2025).
Repeated exposure and consolidation
Repeated meaningful encounters allow children to refine and retrieve provisional word meanings. A later encounter may confirm an earlier guess, reveal that the word applies to several instances, or expose a mistaken interpretation. Retention involves establishing durable memory traces and access routes, whereas generalization requires deciding which properties are relevant to the category. These are closely linked but distinct accomplishments. Repetitions spaced across contexts may offer opportunities for retrieval and comparison, but optimal schedules and benefits differ with age, material, initial knowledge, and testing method.
The practical consequence is straightforward: an adult should not assume that a word has been learned simply because a child identifies a picture correctly once. Observing comprehension across settings, eliciting spontaneous use without immediate modeling, and revisiting vocabulary in varied contexts provide stronger evidence of an enduring representation than one successful response.
Mutual Exclusivity and Lexical Contrast
How a known name guides interpretation of a new one
Mutual exclusivity refers to the tendency to avoid assigning two basic labels to exactly the same referent when a new, unnamed candidate is present. In the classic Markman and Wachtel experiments, preschool children used known object labels to infer where novel labels likely belonged and could sometimes reinterpret a second term as a part or substance label when it could not plausibly name a new object. This research supports a flexible contrast strategy rather than an absolute rule that objects have only one name. Markman and Wachtel (1988).
Real vocabularies contain synonyms, nicknames, category terms, technical labels, translation equivalents, and words for object parts. A child may know dog, animal, pet, and puppy and use each appropriately. The value of mutual exclusivity is therefore conditional: it helps direct attention when another interpretation is plausible, and it can be overridden by grammar, discourse, speaker intent, and later evidence.
A choice heuristic versus a storage mechanism
Competing explanations for mutual-exclusivity responses include lexical constraints, reasoning about an informative speaker, novelty preference, and learned expectations about labels. These explanations may predict similar performance on a simple two-object pointing task. Tests that vary prior word knowledge, referent familiarity, social context, and delayed recall help distinguish them. Critically, selecting an unnamed referent can occur without constructing a strong long-term association. Mather (2025).
A careful model must specify its predicted outcomes: which object is selected immediately, what meaning is inferred, what is remembered after a delay, and how the word is extended. The same surface behavior can arise from different underlying procedures, making convergent experiments indispensable.
Statistical and Cross-Situational Word Learning
Accumulating evidence across ambiguous experiences
Cross-situational word learning describes acquiring a word-referent association by tracking the statistical relationship between a word and possible referents across multiple encounters. In a sequence of ambiguous scenes, the correct object may recur whenever one particular word is spoken, while unrelated distractors vary. Even without a uniquely identifiable referent on any one trial, the distribution of co-occurrences can identify the likely mapping over time. Smith and Yu demonstrated this capacity in infants aged 12 and 14 months under controlled conditions. Smith and Yu (2008).
Yu and Smith also demonstrated learning from distributed co-occurrence statistics in adult participants, helping establish the paradigm that subsequent infant and child research extended. This does not imply that infants engage in the same explicit calculations or that everyday language exposure is equivalent to a laboratory matrix of pictures and words. A statistical description of successful behavior may be compatible with several cognitive implementations. Yu and Smith (2007).
Associations, hypotheses, and memory
One family of models assumes that learners maintain weighted associations between words and candidate meanings, adjusting them as co-occurrences accumulate. Another emphasizes forming a provisional hypothesis about the referent of a word and checking that hypothesis on subsequent encounters. Trueswell and colleagues used eye tracking in adult experiments to support a propose-and-verify process under their task conditions. The broader literature suggests that attention, memory capacity, task structure, and the availability of corrective evidence can make different strategies more or less likely. Trueswell et al. (2013); Roembke et al. (2023).
A focused 2023 review documented the robustness of the cross-situational learning phenomenon across many populations and paradigms while identifying continuing questions about developmental changes, real-world prevalence, and individual differences. Evidence that a learner can solve a designed task establishes a cognitive capability; it does not quantify how often that process governs ordinary vocabulary growth. Roembke et al. (2023).
When a statistical cue can mislead
Naturalistic environments include multiple labels for one object, words that refer to events rather than entities, speakers who change topics, and occasions when important referents are absent. Correlations between a word and an object can arise because both occur frequently in the same routine without being semantically identical. Learning therefore benefits from syntax, communicative intention, conceptual knowledge, and comparison across contexts. A statistical learner who always maps words to the most frequently co-occurring visible item would routinely mistake background objects for intended meanings.
Statistical evidence is most informative when interpreted as one source of constraint in a developing representational system. The child's knowledge changes both which associations are available and which alternatives are worth considering.
Social-Pragmatic Learning and Joint Attention
How people make reference learnable
Caregivers often accompany labels with gaze, gestures, pointing, object handling, expressive intonation, and timely responses to a child's focus. Joint attention occurs when individuals coordinate attention to an entity or event with some awareness of the other's engagement. In such episodes, an adult's word can be interpreted within a shared activity rather than as an isolated acoustic token. Repeated participation also teaches children how particular speakers use words within conversational routines.
Social information is informative without being infallible. The adult's gaze may be ambiguous, a word may refer to something beyond the child's current focus, and learners may follow perceptual salience even when an adult signals another object. The relative value of social signals changes with development and task demands. A sound developmental account examines how children integrate shared attention with previously acquired lexical and conceptual knowledge.
Contingency and responsive interaction
Responsive conversation connects an utterance to what the child has just done, looked at, or attempted to say. A caregiver who notices a child's interest and provides an appropriate label may make the intended meaning more accessible than an unrelated instruction. Higher-quality symbolic and communicative engagement has been associated with later language outcomes in observational research, although such associations reflect complex family and child factors rather than a guaranteed effect of any single caregiver behavior. Hirsh-Pasek et al. (2015).
This evidence suggests that interaction quality deserves consideration alongside total exposure. A long stream of words directed elsewhere and an exchange in which speakers respond to one another differ in their informational structure. The distinction should be treated scientifically, without ranking caregivers by a raw word count or attributing linguistic outcomes to a single parenting choice.
Learning through several partners and activities
Children may learn from parents, siblings, peers, teachers, grandparents, and members of their broader community. Language is encountered during play, meals, caregiving, work, songs, stories, and shared attention to everyday activities. Some linguistic communities use direct child-addressed speech frequently, while others supply significant learning opportunities through participation and overhearing. The general scientific question is what information a learner can extract from available social and linguistic events, not whether every household reproduces one interaction style.
Different partners can also provide distinct examples of pronunciation, vocabulary, and category use. Such variation may support generalization when the child recognizes a word across voices and contexts. Researchers need sufficiently diverse samples to determine which patterns are general and which depend on specific cultural or language environments.
Shape Bias, Category Formation, and Generalization
Extending a novel noun beyond one object
Learning a new noun requires deciding which other entities belong to the named category. A child shown a new solid artifact may generalize the name to objects with similar shapes even when their color or texture differs. Landau, Smith, and Jones reported a shape emphasis in children's extensions of novel count nouns, with developmental and task-related variation. The finding is an important window onto how lexical categories and object representations interact. Landau, Smith, and Jones (1988).
Shape is not the defining property of every category. Material terms, food names, animals, tools, social roles, and abstract categories require additional knowledge about substance, function, origin, relations, or convention. The relevant dimensions depend on the word, the object domain, prior experiences, and grammatical framing. An account that elevated shape to an unqualified universal rule would fail to explain ordinary lexical categories.
Category learning is richer than resemblance
Children must discover whether a label names an individual, a broad kind, a subcategory, a property, or a part. Their generalizations can be shaped by conceptual distinctions and by the linguistic frames in which labels occur. Learning a category changes subsequent perception and attention because some object similarities become more relevant to future decisions. Likewise, experience recognizing objects supports word learning by furnishing preexisting conceptual structures to which a new form can attach.
This reciprocal relationship matters for theoretical debates. Sometimes language may guide category formation; at other times an established category may make a label easier to learn. Behavioral experiments can test these directions by manipulating prior category experience and subsequent labeling. Broad philosophical claims about whether all concepts require words go beyond the evidence of particular extension experiments.
Learning Verbs, Adjectives, Relations, and Abstract Words
Beyond object names
Noun-learning tasks are experimentally tractable because objects can be displayed simultaneously. Many words cannot be identified by choosing among static objects. Verbs such as chase, share, and rescue specify events, participants, and relations; adjectives depend on a comparison dimension; prepositions and spatial expressions relate multiple objects. A child learning these forms must pay attention to changes over time, speakers' interpretations of events, sentence structure, and repeated contexts where the word appears.
For example, a child hearing “the puppy is daxing the toy” may use the verb's grammatical position and the observed event to infer an action interpretation. Additional occasions are needed to determine whether daxing means pulling, carrying, protecting, or something else. The mapping problem is therefore partly a problem of selecting which aspects of an event are linguistically relevant.
Language provides information about invisible meanings
Words for thinking, remembering, promising, believing, and hoping often refer to events that are not directly visible. Children's comprehension of such expressions can grow through conversation, narratives, contrasting uses, and the structure of the sentences in which they appear. Rich dialogue may supply counterfactual, temporal, and causal information unavailable in a single picture. This is another reason that vocabulary acquisition cannot be reduced to counting object-label pairings.
A vocabulary study restricted to concrete picture nouns should be interpreted within that domain. Generalizing from its effects to all lexical categories requires additional experiments using verbs, modifiers, relational terms, or discourse contexts.
Memory, Retrieval, and the Growth of Lexical Knowledge
Encoding, storage, and retrieval as separate demands
Encoding a new word means forming a representation of its sound or sign pattern and relevant meaning. Retention means the information remains available after the original encounter. Retrieval means it can be accessed when required, whether for recognition or production. These processes are interdependent, but a task can be passed by relying mainly on one of them. Immediate forced-choice performance may be supported by residual attention or short-lived activation; successful recall after a delay provides a different kind of evidence about learning. Horst and Samuelson (2008).
Associations become useful lexical knowledge when the learner can encounter a word in novel surroundings, discriminate it from nearby alternatives, and retrieve relevant meanings without requiring the original training display. Word learning is therefore incremental: a preliminary form-meaning mapping may become more precisely specified as contexts accumulate, phonological forms are differentiated, and relevant semantic properties are revised.
Why retrieval and varied contexts matter
Revisiting a word in more than one informative situation may expose which properties of its referent remain stable. The child who meets the label telescope in a story, sees a toy telescope, and later examines a real telescope receives opportunities to connect a form to a broader category and function. Conversely, repeating a label only while pointing at one picture might produce strong item-specific performance without flexible generalization. The appropriate evidence comes from delayed and transfer tests rather than assumptions about what repetition must accomplish.
Memory constraints also interact with referential ambiguity. If several unfamiliar words and objects are introduced close together, children may choose referents effectively while confusing later associations. Reducing irrelevant competitors can sometimes improve learning, but unusually simplified tasks may no longer resemble everyday linguistic interaction. Experimental control and ecological realism serve different scientific purposes.
Growth produces new opportunities for learning
As vocabulary expands, known words become cues for interpreting unknown ones, established categories enable more efficient generalization, and familiar morphosyntactic constructions help constrain the type of meaning to expect. Vocabulary growth can thus become self-supporting: prior lexical knowledge changes the informational value of the next learning encounter. Still, differences in exposure, interest, task demands, and opportunities make it misleading to expect a uniform trajectory or a single period of explosive growth for all children.
The Language Environment: Quantity, Diversity, and Interaction
Word quantity is only one dimension
The amount of speech children encounter can be associated with vocabulary development because more encounters create more opportunities to notice and practice words. But the same number of utterances can contain widely different lexical diversity, syntactic variety, topics, opportunities to respond, and degrees of contextual support. Researchers therefore distinguish quantity from quality and ask whether different aspects of input predict outcomes at different ages. Rowe (2012).
In a longitudinal study of 50 caregiver-child dyads, Rowe found that richer and more varied vocabulary in speech to toddlers and more decontextualized talk with preschoolers helped explain later vocabulary differences after statistical adjustment for earlier skill and other measured characteristics. Because the study was observational and relatively small, its estimates should be interpreted as developmental associations rather than an experimentally established formula for optimal caregiver speech. Rowe (2012).
Contextual and decontextualized language
Contextualized talk refers to immediately available people, objects, and activities. Such talk gives learners concrete cues and opportunities to test possible meanings. Decontextualized talk involves events, ideas, and relations outside the here and now: telling a story about yesterday, imagining tomorrow, explaining why a machine works, or discussing what a character knows. These modes offer complementary information. Context makes initial reference easier; narrative and explanation extend vocabulary into semantic domains that cannot be taught through immediate pointing.
The useful question is which conversational content fits a child's current knowledge and circumstances. More sophisticated words are valuable when the interaction helps the child infer their meanings. A difficult word placed in an entirely opaque context may offer less usable information than a familiar topic elaborated with a new but interpretable term.
Caution about the language-gap narrative
Vocabulary differences between groups should not be interpreted as a simple measure of caregiver commitment or intelligence. Studies may differ in sampling, transcription methods, which speakers and contexts are counted, and the language varieties being assessed. Socioeconomic correlations are embedded in wider structures of time, access, social conditions, education, and opportunities. Observational group differences cannot be turned directly into a single causal statement about why a particular child knows more or fewer words. Fernald, Marchman, and Weisleder (2013); Rowe (2012).
Scientific and educational work benefits from describing the conditions of measurement, separating within-group variation from group averages, and choosing measures appropriate to the child's actual language experiences. Families are entitled to accurate, respectful information rather than a sensational numerical benchmark.
Shared Book Reading and Word Learning
Why books create distinctive learning opportunities
Shared storybook reading offers repeated exposure to words embedded in a coherent narrative, illustrations, adult commentary, and opportunities for children to ask questions or make predictions. Books can introduce objects, events, and expressions that occur less frequently in daily conversation. A story may use a word several times while making its function increasingly clear. These features make reading an informative context for vocabulary acquisition, although the magnitude of measured benefits varies across intervention design, outcomes, and comparison groups.
In a meta-analysis of 38 studies including 2,455 children, Flack, Field, and Horst examined factors associated with word learning from shared storybooks. Among the moderators were dialogic reading style, word tokens, and the number of words tested. The analysis directly concerned children's word comprehension, and its estimates should not be generalized uncritically to all forms of language development. Flack, Field, and Horst (2018).
Dialogic reading and active participation
Dialogic reading describes interactive practices in which an adult encourages a child to comment, answer, anticipate, or narrate while discussing pictures and events. The important mechanism is a sequence of meaningful contributions and feedback, rather than merely exposing a child to additional printed words. When the child identifies a character's goal, explains an unusual expression, or tells part of a story, the interaction connects vocabulary with retrieval, inference, and conversational purpose.
Evidence from randomized interventions requires careful interpretation. Dowdall and colleagues synthesized 19 randomized trials involving 2,594 children and reported small positive average effects on expressive and receptive language and a larger effect on caregiver book-sharing competence, with dosage moderating child-language outcomes. These are average effects across studied programs, not guaranteed gains for each child or proof that any one reading technique is universally superior. Dowdall et al. (2020).
A separate meta-analysis by Noble and colleagues found more modest average effects on children's language and negligible average effects in the subset using active control groups. The contrast matters because being assigned to any engaging, adult-supported activity may improve measured outcomes relative to receiving no additional activity. Stronger studies must isolate what is specific to shared reading, assess longer-term outcomes, and report meaningful individual variation. Noble et al. (2019).
Association is different from intervention effectiveness
A 2025 systematic review and meta-analyses of home-based shared book reading described positive associations between reading experiences and language or vocabulary outcomes. Much of its quantitative synthesis involved correlations rather than randomized allocation. Such evidence complements experiments by showing how behaviors cluster with development in natural environments; it does not by itself determine whether changing reading frequency will cause the same measured difference. Galea et al. (2025).
For parents and educators, the scientific implication is to treat books as a flexible context for meaningful conversation. Pointing, explaining, relating a story to experience, and following a child's curiosity can create multiple routes into a word's meaning. Research does not support promising a particular vocabulary score from a fixed daily reading duration.
Incidental Learning, Explicit Teaching, and Informative Feedback
Learning words without a formal lesson
Incidental word learning occurs when acquiring vocabulary is not the learner's main stated task. A child may hear the same unusual term during a story and infer enough from the events to understand it later. Many experiments test such learning by exposing children to words in narratives or ambiguous scenes without requesting memorization. Success varies with contextual informativeness, initial knowledge, attention, word properties, and the way comprehension is subsequently measured.
Explicit teaching can supply a clearer initial meaning, especially when the word is abstract, rare, or crucial to understanding a topic. A child may benefit from hearing an explanation, observing an example, comparing near categories, or being asked to use the term in an appropriate context. Instruction and incidental learning can work together: clear explanation provides a starting interpretation, while later exposure allows the child to refine and use it flexibly.
Error correction and conversational expansion
Feedback is informative when it helps the child notice a distinction or provides a more precise formulation without disrupting the communicative exchange. If a child calls a heron a duck, an adult can supply the more specific name while commenting on the bird's long legs and beak. Such a response supplies lexical contrast, descriptive cues, and continued participation. It also creates an opportunity for the child to use the new expression later.
There is no reason to infer that every mistaken use should trigger an explicit correction or that every adult expansion produces durable learning. Both the child's goals and the conversational context matter. Experimental studies separate the contribution of corrective information from additional repetition or attention by controlling what the comparison group hears.
How Children Learn Multiple Meanings and Word Relationships
Polysemy, ambiguity, and contextual selection
Many common words have related meanings: a head can be part of a body, the leader of an organization, or the top of a table. Other words have distinct meanings despite sharing a form, as with bat. To acquire such vocabulary, children must use grammar, contextual expectations, and prior knowledge to infer which sense a speaker intends. Learning a word includes discovering the range of its acceptable uses and the boundaries between related categories.
Lexical knowledge is organized in networks of sound, meaning, category membership, and association. Knowing garden can support interpreting gardener; knowing eat, meal, and hungry helps interpret a new food-related expression. These relationships create opportunities for inference, yet they can also generate errors when a learner overextends a familiar pattern. Word meanings continue to change as the child encounters metaphor, idiom, technical language, and specialized school vocabulary.
Breadth and depth of vocabulary
Vocabulary breadth refers roughly to how many words a learner recognizes or can use according to a particular test. Depth concerns how well words are understood: their senses, typical contexts, category relations, grammatical behavior, and the ability to explain or use them flexibly. A child may recognize dozens of animal names while knowing little about some of them, or understand a smaller set of words in considerable detail. Both dimensions contribute to communication and comprehension, but neither can be read directly from a single checklist total.
Research designs benefit from measuring more than immediate picture selection. Describing a meaning, matching different examples, resolving ambiguous uses, and generating original sentences each probe different aspects of lexical knowledge. A child can improve on one dimension while remaining relatively stable on another.
Vocabulary Acquisition in Bilingual and Multilingual Children
Word learning is distributed across languages
Children exposed to two or more languages may learn different words with different interlocutors, activities, and settings. An English word for an object and a Spanish word for the same object are two lexical forms referring to a shared or overlapping concept; knowledge of one does not imply that the child must already produce the other. Vocabulary acquired at home may differ from vocabulary learned in school, and skills can be distributed unevenly without indicating a deficit in language-learning capacity.
A child who knows the same concept in two languages has acquired two usable words. Counting only one language can miss relevant knowledge. At the same time, adding the counts from each language, counting concepts only once, and assessing the languages separately answer different scientific questions. No one total is equally useful for every research or assessment purpose. Core et al. (2013); Fitton et al. (2023).
Word vocabulary and concept vocabulary
In bilingual research, word vocabulary sums known word forms across languages, including translation equivalents. Concept vocabulary counts the number of distinct meanings represented across the languages, counting translation equivalents only once. The choice changes the interpretation of scores. Knowing dog and perro contributes two word forms but ordinarily one shared concept in a simplified example. Researchers have emphasized that the measures are not interchangeable and that scoring choices can produce distinct patterns across children with varying language balance. Fitton et al. (2023).
Core and colleagues compared vocabulary measures in Spanish–English bilingual toddlers and found evidence supporting total-vocabulary approaches under their studied conditions. Later work with school-aged children reported differential bias in single-language and conceptual scores depending on relative language proficiency. Both findings caution against interpreting an isolated monolingual test as a complete portrait of multilingual lexical knowledge. Core et al. (2013); Fitton et al. (2023).
Exposure patterns, dialects, and learning opportunities
The amount of experience in each language, the situations where it is heard and used, and the child's conversational opportunities influence which vocabulary is most accessible. Repeated use of a word with one family member may make it easier to produce in that setting than in an unfamiliar testing environment. A bilingual child's dominant language can change with schooling, migration, relationships, and available environments. These patterns are central to an accurate scientific account of vocabulary growth.
Dialectal differences and community-specific words are legitimate forms of language knowledge. A standardized assessment based on a different variety may misclassify familiar expressions as unknown. Evaluating word knowledge across languages and varieties is therefore an issue of measurement validity. Clinical decisions, if needed, belong to individualized professional assessment using the child's linguistic history and culturally appropriate methods rather than a vocabulary count obtained in only one unfamiliar variety.
Individual Differences and Developmental Variation
Vocabulary trajectories differ for several reasons
Some children accumulate particular lexical categories earlier than others; some show large receptive vocabularies before expressive growth; others become especially interested in a narrow topic and develop detailed terminology within it. These variations can arise from language experience, cognition, speech production demands, opportunities for interaction, interest, and the sampling method. Statistical association between one child characteristic and vocabulary size does not establish that the characteristic independently causes every later difference.
Measurement age also matters. A word that is acquired easily at one stage may become harder to discriminate when many phonologically similar alternatives are known. Conversely, growing grammatical knowledge can help children interpret an unfamiliar word embedded in a complex sentence. Longitudinal designs help reveal these interacting developments, while cross-sectional designs describe relationships at a single time.
Learning opportunities and structural conditions
Economic and educational inequalities can affect access to books, early childhood services, stable environments, caregiver time, and broader experiences with language. Those conditions are complex, and children within every socioeconomic category vary substantially. The observation that vocabulary scores differ across sampled groups does not imply a uniform family-language pattern, a biological hierarchy, or a deterministic developmental outcome. Fernald, Marchman, and Weisleder (2013); Rowe (2012).
Study design should account for sampling, culture, language variety, and the specific measures being compared. Generalizations based primarily on English-speaking urban populations require verification in other linguistic and social environments. A universally applicable theory of word learning should explain not only what children have in common but also how flexible strategies operate under diverse forms of language experience.
Word Learning and Developmental Language Differences
Typical variation, signs of difficulty, and diagnostic categories
Vocabulary size or delayed production of a few words is a developmental observation. A symptom is a particular observed difficulty. A screening result indicates whether more assessment may be useful. A diagnosis requires a broader professional evaluation grounded in accepted criteria and an adequate account of the child's hearing, development, language experience, and functional communication. These categories serve different purposes and should remain distinct.
Developmental language disorder can involve persistent difficulties with language learning and use that affect communication and daily life. Vocabulary limitations may be one feature, alongside difficulties in grammar, comprehension, narrative, or other domains. A child does not acquire a diagnosis through poor performance on one artificial word-learning task, and a vocabulary measurement alone cannot establish the cause of a difficulty. The specialized assessment, differential diagnosis, and treatment of clinical language disorders belong to clinical speech-language services.
Hearing, speech production, and comprehension
A child who has trouble producing a word may nevertheless understand it. Motor speech, hearing, articulation, fluency, and language representation involve distinct processes that can interact. Researchers therefore separate receptive word learning from expressive production demands and specify the modality of each test. This distinction protects against treating every expressive delay as a failure to learn meaning or every recognition error as proof of a general language disorder.
Hearing and language histories are relevant background for interpreting word-learning performance, but this scientific article is not an individual diagnostic tool. Concerns about persistent difficulties across everyday settings warrant evaluation by qualified professionals familiar with the child's languages and circumstances, using validated assessment approaches.
Neurodiversity and heterogeneous developmental profiles
Children with different developmental profiles can vary in their relative strengths in lexical learning, memory, social attention, and language production. These dimensions should be examined directly rather than attributed automatically to a diagnostic label. A specific study may document differences under controlled conditions without supporting the claim that all members of a broad clinical or neurodevelopmental group learn words through the same mechanism.
Respectful assessment also recognizes that spoken language, signed language, augmentative communication, and different language varieties can convey genuine lexical knowledge. Scientific comparisons should state which abilities are being measured and avoid equating one communication modality with overall cognitive competence.
How Scientists Study Vocabulary Learning
Looking-while-listening and eye tracking
In preferential-looking and looking-while-listening paradigms, researchers present visual displays and play a named word while measuring where a child looks. The difference in looking toward the named target relative to alternatives can provide evidence of word recognition even when overt speech is unavailable. Detailed eye-tracking measures also reveal changes in the timing of decisions, allowing researchers to examine the efficiency of lexical access. Such measures require appropriate baselines, stimulus controls, and caution when interpreting visual attention as semantic knowledge. Bergelson and Swingley (2012); Fernald, Marchman, and Weisleder (2013).
Looking behavior is informative because it offers a relatively low-demand response. Yet gaze reflects several processes: visual salience, familiarity, attention, fatigue, word recognition, and decisions about where to orient. A carefully designed experiment varies target positions, counterbalances stimuli, and includes comparison conditions so that the lexical interpretation is more plausible than a general visual preference.
Referent selection and delayed retention
A referent-selection task asks children to choose which object a new word names when one or more alternatives are present. Researchers can vary the number of familiar and unfamiliar objects, the speaker's gaze, the novelty of items, or whether the label appears in an informative sentence. A delayed retention task later examines whether the child still identifies the new association when some of the original cues are removed. Testing both stages is essential if the scientific claim concerns learning rather than immediate inference. Horst and Samuelson (2008).
A methodologically strong design examines whether the child remembers the form, the referent, or both; whether the choice was made through exclusion; and whether success persists under a new display or with new exemplars. Claims of fast or durable acquisition should always specify the retention interval and outcome type.
Cross-situational training paradigms
Researchers manipulate the number of candidate objects, the consistency of correct associations across trials, and the interference produced by competing pairings. Eye movements, object choices, and later tests can reveal how ambiguous evidence is integrated. These procedures have helped distinguish associative tracking from hypothesis testing and clarify how attention and memory contribute to learning. However, the experimentally created worlds often contain far fewer words and referents than ordinary conversations. Smith and Yu (2008); Roembke et al. (2023).
The scientific value of such paradigms is mechanistic precision. Their ecological validity depends on whether the cognitive operations observed are also deployed in less controlled, socially complex environments. Research combining laboratory experiments with naturalistic observation can narrow that inferential gap.
Caregiver report and vocabulary inventories
Parent or caregiver inventories collect observations about the words children understand or produce during daily life. These tools can sample vocabulary that is difficult to elicit during a short laboratory visit, and repeated administration can describe developmental change. Interpretation depends on the inventory's language and cultural coverage, who completes it, how familiar items are defined, and whether the result concerns comprehension or production.
For bilingual children, deciding how reports from more than one language will be combined is itself a substantive measurement choice. Researchers have documented how single-language and combined scoring methods can change results, especially when languages are unequally developed or some concepts have translation equivalents. Fitton et al. (2023).
Intervention trials, observational cohorts, and evidence synthesis
Longitudinal observational studies examine how early environments or learning skills predict later vocabulary while adjusting for measured covariates. Randomized trials ask whether a specified intervention causes differences relative to an appropriate comparison. Meta-analyses synthesize reported effects but inherit the outcome definitions and design limitations of the included research. These approaches answer complementary questions, and conclusions become stronger when independent methods converge. Rowe (2012); Dowdall et al. (2020); Noble et al. (2019).
Comparisons using passive controls tend to answer whether added activities outperform ordinary routines. Active controls can ask whether a named component offers benefits beyond another engaging activity. Vocabulary interventions should report the number of taught words, training dose, retention interval, whether learners generalized beyond the trained examples, and whether gains extended to broader language ability.
Major Theories of Word Learning
Constraint-based accounts
Constraint-based theories describe expectations that reduce referential ambiguity: a learner may prefer a new label for an unnamed object, interpret some nouns as category terms, or focus on shape in certain object-labeling tasks. Such proposals help explain selective first interpretations without requiring exhaustive consideration of all possible meanings. Their strongest versions predict broad, stable biases; more flexible versions allow context, prior knowledge, and grammatical form to change decisions. Evidence from mutual exclusivity and shape-generalization experiments supports the relevance of constraints while also documenting their limits. Markman and Wachtel (1988); Landau, Smith, and Jones (1988).
Associative and statistical accounts
Associative models treat learning as changes in the strength of connections between forms and candidate meanings. They are particularly suited to explaining how frequency and consistency across multiple uncertain exposures affect eventual mapping. A strong theory must specify how attention determines which candidates are encoded, how associations compete, and how the model handles absent referents or alternative names. Cross-situational experiments support the capacity to learn from distributed statistical evidence but do not by themselves uniquely establish one learning algorithm. Yu and Smith (2007); Roembke et al. (2023).
Social-pragmatic accounts
Social-pragmatic theories focus on speakers' intentions, shared attention, communicative relevance, and the learner's participation in social routines. Their advantage is that reference in ordinary language is purposeful: speakers do not randomly pair labels with visual objects. However, a theory emphasizing interaction must still explain how word forms are stored, generalized, and recognized when the original speaker is absent. Social and statistical information can work together because a shared activity determines which events and associations the child attends to.
Propose-and-verify and hybrid models
Hypothesis-testing accounts posit that a learner provisionally selects a candidate meaning and checks it against later evidence. Associative accounts can sometimes yield similar accuracy but different error patterns or eye movements. The propose-and-verify approach has been investigated in controlled adult tasks; its extension to all ages and natural vocabulary environments remains a testable developmental proposition rather than a universally demonstrated algorithm. Trueswell et al. (2013).
A productive synthesis treats word-form encoding, attention, social interpretation, statistical accumulation, conceptual inference, and memory as interacting components. The precise causal organization remains an empirical question. Theories should be judged by their ability to predict patterns of errors, retention, transfer, and individual differences under independently specified conditions, not solely by their ability to redescribe a successful pointing response.
What the Evidence Establishes, and Where It Remains Limited
Established findings
Children can recognize some familiar word meanings within their first year, infer plausible referents under ambiguity, and learn mappings from repeated statistical information. Controlled work has established that immediate referent selection and longer-term retention can diverge. The word-learning environment includes social cues, linguistic structure, conceptual knowledge, and memory demands. These claims rest on converging experimental programs rather than a single all-purpose theory. Bergelson and Swingley (2012); Smith and Yu (2008); Mather (2025).
Supported but context-dependent findings
Novel count-noun learning can favor shape, unfamiliar labels can guide choices toward unnamed objects, and interactive storybook reading can support target-word learning. These effects depend on stimulus domains, competing cues, definitions of success, and intervention methods. Stronger interpretations require evidence of retention, generalization, and real-life relevance. Landau, Smith, and Jones (1988); Flack, Field, and Horst (2018).
Contested mechanisms and limited generalizations
Researchers continue to debate the relative role of associative accumulation and hypothesis testing in cross-situational learning, the exact mechanisms of mutual-exclusivity behavior, and how much laboratory fast-mapping performance contributes to lasting vocabulary growth. Strong claims about universal optimal input quantity or universal intervention effects also exceed the evidence. Active-control comparisons, sampling limitations, and publication-bias adjustments can materially alter results. Roembke et al. (2023); Noble et al. (2019); Mather (2025).
Methodological priorities
Future-looking scientific questions include which contexts produce durable learning rather than transient selection, how learning strategies change across development, and whether findings replicate across different language communities and modalities. These are descriptions of unresolved research problems, not predictions about particular children. Strong inference requires multiple methods, delayed outcomes, adequate samples, transparent analyses, and clear separation of experimental capabilities from their natural-world prevalence.
Practical Implications for Parents, Educators, and Researchers
Make reference interpretable
When introducing a new term, connect it to a meaningful activity, an informative example, or an explanation the child can interpret. For an object name, a clear demonstration or comparison may be useful. For a verb, act out or describe the relation among participants. For an abstract term, use a short story or contrast. The goal is to provide cues that distinguish the intended meaning from plausible alternatives, so children have evidence for forming an initial mapping.
Support repeated, flexible encounters
Use the word naturally across settings and with different examples. Invite the child to recognize, explain, or use it without assuming that every correct first response demonstrates mastery. Repetition across meaningful contexts can provide opportunities for memory retrieval, contrast, and generalization. A new word embedded in a single rehearsed exchange may be less accessible when the surroundings change.
Follow the conversation and encourage participation
Answer children's questions, elaborate on what interests them, and create opportunities for reciprocal talk. Shared reading can provide particularly good contexts for storytelling and explanations, while play, routines, and everyday tasks are equally legitimate sources of word-learning events. Evidence from intervention studies favors measured expectations: average gains may be modest, and studies with active control groups can yield smaller results than comparisons with no added activity. Noble et al. (2019); Dowdall et al. (2020).
Value all the child's languages
Use the languages and varieties through which meaningful relationships and communication occur. Bilingual word learning may be distributed across speakers and settings; accurate observations consider the words understood and used in each language. An English-only result cannot stand in for a complete estimate of a child's bilingual vocabulary. Fitton et al. (2023).
Interpret progress through several kinds of evidence
For teachers and investigators, a sound evaluation combines appropriately chosen measures of comprehension, production, retention, and generalization. A child may correctly repeat a word while lacking a stable meaning, or understand a word that is not yet available for fluent production. A delayed, novel-context test is particularly informative when the aim is to establish that learning has persisted.
Frequently Asked Questions About Children's Word Learning
Can children learn a word after hearing it once?
Children can sometimes infer a likely meaning from one encounter, particularly when the context is informative and alternatives are constrained. Retention and flexible use usually require separate evidence. The scientific term fast mapping should specify whether the measured outcome was immediate referent selection, memory after a delay, or generalization. Horst and Samuelson (2008); Mather (2025).
How do children know what a new word refers to?
They combine information from visible events, prior word knowledge, category expectations, grammar, speakers' attention and communicative behavior, and statistical regularities across situations. No cue identifies every kind of word equally well. What matters is the accumulation of evidence compatible with a useful meaning.
Does hearing more words always produce a larger vocabulary?
Exposure creates opportunities for learning, but word diversity, conversational context, the child's attention, and earlier knowledge also matter. Observational associations are not a universal causal formula, and no fixed number of words spoken by adults guarantees a particular developmental result. Rowe (2012).
Is shared reading an evidence-based way to support vocabulary?
Studies show that children can learn target words during shared reading, and some randomized interventions improve language outcomes on average. The size of improvement varies; comparisons with active alternatives can show smaller effects. Interactive discussion may be more informative than treating reading as a race to finish pages. Flack, Field, and Horst (2018); Noble et al. (2019); Dowdall et al. (2020).
Are bilingual children learning fewer words?
A single-language vocabulary score can underestimate vocabulary spread across two languages. Researchers distinguish the total number of word forms known from the number of distinct concepts represented, and those measures can yield different conclusions. Development should be interpreted using a child's full linguistic experience rather than a monolingual score treated as a universal measure. Core et al. (2013); Fitton et al. (2023).
Does a child need to speak a word to know it?
No. Receptive comprehension often precedes reliable expressive production, and different tasks place different demands on phonological processing, memory, and motor speech. A child can demonstrate knowledge by looking appropriately, selecting a referent, following a request, or using another communicative modality. How that evidence is interpreted depends on the task.
When does vocabulary difficulty suggest a language disorder?
Persistently limited understanding or use of words across relevant everyday settings may be one reason to seek an individualized developmental or speech-language assessment. Vocabulary alone neither specifies a diagnosis nor identifies a cause. Clinical evaluation considers the child's complete communication profile, hearing and developmental history, languages, and functional context.
Conclusion: Vocabulary Learning as Cumulative Meaning Construction
Children learn words by constructing and revising links between communicative forms and meanings. Perceptual and phonological information makes new forms recognizable; attention and social context make candidate meanings accessible; statistics across encounters help disambiguate reference; category knowledge supports generalization; and memory makes word knowledge available beyond the original occasion. The strongest scientific account links these components to separate measurable outcomes: initial inference, retention, recognition, production, and use in new contexts. Roembke et al. (2023); Mather (2025).
Vocabulary acquisition remains a continuing aspect of human language development, changing as children gain more lexical, grammatical, conceptual, and social knowledge. Research illuminates a flexible learning system that draws on multiple forms of evidence and operates across languages and communicative settings. The result is more than a list of labels: it is an expanding capacity to understand how other people use words to represent objects, actions, relations, events, and ideas.
Related Articles
Mental Lexicon: How Words Are Stored and Retrieved — The lexical representations that support recognition, meaning, competition, and retrieval as vocabulary grows.
Psycholinguistics: How Language Is Processed in the Mind — The broader cognitive framework for acquiring, comprehending, and producing language.
Speech Segmentation: How Listeners Find Words in Continuous Speech — How learners identify possible word forms in the spoken stream before establishing meanings.
Semantic Memory: Facts, Meaning, and Knowledge — The related cognitive-science account of long-term conceptual knowledge as a separate canonical topic.
References
Bergelson, E., and Swingley, D. (2012). At 6–9 months, human infants know the meanings of many common nouns. Proceedings of the National Academy of Sciences, 109(9), 3253–3258. https://doi.org/10.1073/pnas.1113380109.
Core, C., Hoff, E., Rumiche, R., and Señor, M. (2013). Total and conceptual vocabulary in Spanish–English bilinguals from 22 to 30 months: Implications for assessment. Journal of Speech, Language, and Hearing Research, 56(5), 1637–1649. https://doi.org/10.1044/1092-4388(2013/11-0044).
Dowdall, N., Melendez-Torres, G. J., Murray, L., Gardner, F., Hartford, L., and Cooper, P. J. (2020). Shared picture book reading interventions for child language development: A systematic review and meta-analysis. Child Development, 91(2), e383–e399. https://doi.org/10.1111/cdev.13225.
Fernald, A., Marchman, V. A., and Weisleder, A. (2013). SES differences in language processing skill and vocabulary are evident at 18 months. Developmental Science, 16(2), 234–248. https://doi.org/10.1111/desc.12019.
Fitton, L., Goodrich, J. M., Thayer, L., Pratt, A., and Luna, R. (2023). Bilingual vocabulary assessment: Examining single-language, conceptual, and total scoring approaches. Journal of Speech, Language, and Hearing Research, 66(9), 3486–3499. https://doi.org/10.1044/2023_JSLHR-22-00573.
Flack, Z. M., Field, A. P., and Horst, J. S. (2018). The effects of shared storybook reading on word learning: A meta-analysis. Developmental Psychology, 54(7), 1334–1346. https://doi.org/10.1037/dev0000512.
Galea, C., Jones, A., Ko, K., Salins, A., Robidoux, S., Noble, C., and McArthur, G. (2025). Home-based shared book reading and developmental outcomes in young children: A systematic review with meta-analyses. Frontiers in Language Sciences, 4, 1540562. https://doi.org/10.3389/flang.2025.1540562.
Hirsh-Pasek, K., Adamson, L. B., Bakeman, R., Owen, M. T., Golinkoff, R. M., Pace, A., Yust, P. K. S., and Suma, K. (2015). The contribution of early communication quality to low-income children's language success. Psychological Science, 26(7), 1071–1083. https://doi.org/10.1177/0956797615581493.
Horst, J. S., and Samuelson, L. K. (2008). Fast mapping but poor retention by 24-month-old infants. Infancy, 13(2), 128–157. https://doi.org/10.1080/15250000701795598.
Landau, B., Smith, L. B., and Jones, S. S. (1988). The importance of shape in early lexical learning. Cognitive Development, 3(3), 299–321. https://doi.org/10.1016/0885-2014(88)90014-7.
Markman, E. M., and Wachtel, G. F. (1988). Children's use of mutual exclusivity to constrain the meanings of words. Cognitive Psychology, 20(2), 121–157. https://doi.org/10.1016/0010-0285(88)90017-5.
Mather, E. (2025). How well do children remember fast-mapped words? A pre-registered meta-analysis of retention following the mutual exclusivity response. Infant and Child Development, 34(3), e70019. https://doi.org/10.1002/icd.70019.
Noble, C., Sala, G., Peter, M., Lingwood, J., Rowland, C. F., Gobet, F., and Pine, J. M. (2019). The impact of shared book reading on children's language skills: A meta-analysis. Educational Research Review, 28, 100290. https://doi.org/10.1016/j.edurev.2019.100290.
Roembke, T. C., Simonetti, M. E., Koch, I., and Philipp, A. M. (2023). What have we learned from 15 years of research on cross-situational word learning? A focused review. Frontiers in Psychology, 14, 1175272. https://doi.org/10.3389/fpsyg.2023.1175272.
Rowe, M. L. (2012). A longitudinal investigation of the role of quantity and quality of child-directed speech in vocabulary development. Child Development, 83(5), 1762–1774. https://doi.org/10.1111/j.1467-8624.2012.01805.x.
Smith, L., and Yu, C. (2008). Infants rapidly learn word-referent mappings via cross-situational statistics. Cognition, 106(3), 1558–1568. https://doi.org/10.1016/j.cognition.2007.06.010.
Trueswell, J. C., Medina, T. N., Hafri, A., and Gleitman, L. R. (2013). Propose but verify: Fast mapping meets cross-situational word learning. Cognitive Psychology, 66(1), 126–156. https://doi.org/10.1016/j.cogpsych.2012.10.001.
Yu, C., and Smith, L. B. (2007). Rapid word learning under uncertainty via cross-situational statistics. Psychological Science, 18(5), 414–420. https://doi.org/10.1111/j.1467-9280.2007.01915.x.