My John BabyThe first year, one week at a time

Growing & Milestones

How Babies Narrow Down The Sounds Of Their Own Language

Infants can distinguish speech sounds from any language early on and gradually lose the distinctions their own language does not use, which is a gain rather than a loss.

Adorable baby in overalls sitting on a blanket outdoors, smiling brightly.
Photograph by Helena Lopes via Pexels
General information. This article is journalism, not medical advice, and it cannot know your circumstances. Speak to a qualified professional about anything that concerns you. How we work.

Babies start out able to hear differences between speech sounds that adults around them cannot. Over the first year that ability narrows to the set their own language actually uses.

The starting point is broader than adult hearing

Every language draws its sound categories from a large pool of possible distinctions, and each uses only some. Adults become effectively deaf to contrasts their language ignores.

Young infants show no such limitation and respond to contrasts from languages they have never heard. The capacity is general rather than tuned.

This is measurable through methods that watch whether a baby notices a change in a repeated sound, using looking or sucking as the response. A baby who detects the change reacts, and a baby who does not carries on unchanged.

Narrowing happens over the first year

Through the second half of the first year, sensitivity to unused contrasts declines while sensitivity to used ones sharpens. The system is committing to the language it is hearing.

This looks like loss and functions as gain. Ignoring irrelevant variation makes it far easier to recognise words spoken by different voices at different speeds.

The same narrowing happens in other domains, including face recognition, where early broad sensitivity gives way to expertise in familiar categories.

Exposure has to be social

The narrowing is driven by hearing language in interaction rather than by hearing it in the background. Recorded audio alone has repeatedly been shown to do far less.

The likely reason is that live interaction supplies timing, eye contact and shared attention, which tell a baby which sounds relate to what. Sound without context is noise with structure.

This is the substance behind advice to talk to babies constantly, and why narrating what you are doing is more useful than playing a language recording.

Two languages tune two sets

Babies hearing more than one language maintain the contrasts of both, and the narrowing occurs within each. The process is not divided between them.

Vocabulary in each language may appear at a different pace than in a single-language baby, while total vocabulary across both is generally comparable.

Concerns about hearing or about speech development are worth raising early with a health visitor or GP, since access to audiology and speech services is easier the sooner it is asked for. Waiting to see rarely improves the position.

Questions readers ask

Is a low percentile a problem?

Not in itself. A large number of entirely healthy babies sit on lower lines. What professionals watch is the trend, and any concern should be raised with them.

Should I plot measurements myself between appointments?

It generally adds anxiety rather than information, partly because home measurement is noisy. If you want more frequent checks, discuss it with your health visitor.

Growing & Milestonesgrowth chartspercentilesmeasurement
More in Growing & Milestones
Shalu Prasad
Editor, My John Baby

Shalu edits My John Baby and checks every piece for anything that reads as medical advice.

Also by Shalu Prasad