Understanding language development in infancy

Language does not begin when the first word is spoken. Months before "mama" or "ball" emerge, the infant is laying down the perceptual, social, and cognitive foundations on which language will be built. Understanding how language begins requires looking not at words but at the capacities that make words possible: the ability to perceive speech sounds categorically, to extract statistical regularities from the stream of sound, to attend jointly to objects with a social partner, and to decode the communicative intentions of others.

Patricia Kuhl

b. 1946

Research on perceptual attunement to native phoneme categories across the first year of life, showing that the "universal listener" of early infancy becomes a native-language specialist by 10–12 months. Also demonstrated that live social interaction is necessary for this learning — infants did not narrow their phoneme categories from exposure via video or audio alone.

Michael Tomasello

b. 1950

Emphasised joint attention and the understanding of communicative intentions as the uniquely human foundation of language acquisition. Argued that word learning is not a matter of statistical association but of reading the social intent behind a pointing gesture or gaze.

Jenny Saffran

b. 1967

Demonstrated statistical learning of word segmentation in 8-month-old infants using an artificial language paradigm, establishing that language learning is partly driven by implicit tracking of distributional regularities in the input.

Categorical Speech Perception

The tendency to perceive speech sounds as belonging to discrete categories rather than as points on a continuous physical dimension. Demonstrated in infants using the high-amplitude sucking habituation paradigm: infants treat a shift across a phoneme boundary (e.g., /p/ to /b/) as more salient than an equally large shift within a category. Initially universal, it narrows to native-language contrasts by 10–12 months.

Statistical Learning

The implicit extraction of regularities — particularly transitional probabilities between adjacent syllables — from the input stream. Saffran, Aslin, and Newport (1996) showed that 8-month-old infants segment words from a continuous artificial language after 2 minutes of exposure, using the fact that syllables within words follow each other with higher probability than syllables across word boundaries.

Canonical Babbling

Babbling that conforms to syllabic structure — repeated consonant-vowel sequences such as "bababa" or "mamama." Emerges between 6 and 10 months and is considered a universal milestone of pre-linguistic development. Its onset reflects maturation of the vocal tract and oro-motor control, and its phonological content is increasingly shaped by the ambient language.

Joint Attention

The capacity to co-ordinate visual attention with a social partner toward a shared external object or event, achieved through gaze following, pointing, and referential looking. Emerges reliably around 9–12 months. Functions as the scaffold on which word learning is built: words are typically learned in joint attention episodes in which caregiver and child share a referent.

Child-Directed Speech (Motherese)

The register adults spontaneously adopt when talking to infants: higher pitch, exaggerated prosodic contours, slower pace, simpler grammar, and repetition. Infants prefer it to adult-directed speech. It may facilitate phoneme discrimination, word segmentation, and syntactic learning by making the acoustic properties of the language more salient.

Perceptual Narrowing

The developmental process by which infants' initially universal sensitivity to phoneme contrasts narrows to the contrasts relevant to their native language. By 10–12 months, infants who once discriminated Hindi retroflex consonants or Zulu click consonants no longer do so reliably. Reflects neural tuning driven by statistical exposure to the ambient language.

Last reviewed July 2025
  1. 1.

    Saffran J.R., Aslin R.N., & Newport E.L. (1996). Statistical learning by 8-month-old infants. Science, 274, 1926–1928.

    +About this source

    The landmark demonstration of word segmentation via statistical learning in infants.

  2. 2.

    Kuhl P.K. (2004). Early language acquisition: Cracking the speech code. Nature Reviews Neuroscience, 5, 831–843.

    +About this source

    Overview of perceptual attunement, statistical learning, and the social gating of phoneme learning.

  3. 3.

    Tomasello M. (2003). Constructing a Language. Harvard University Press.

    +About this source

    Argues that joint attention and intention-reading underlie word learning and syntactic development.