Why Can You Still Hear Where Someone Grew Up? The Ear Was Trained First

Newborns can distinguish essentially every sound contrast used in any human language, which has been demonstrated by measuring how long they attend to a sound when it changes. Over the first year that ability narrows sharply, and by around ten months an infant reliably distinguishes the contrasts that matter in the language around them while no longer reliably distinguishing ones that do not. The brain is committing to a set of categories, which makes processing the local language enormously more efficient and which makes the discarded distinctions genuinely hard to hear later. An adult who cannot produce a foreign sound frequently cannot hear the difference either, which is the more fundamental problem.

Several factors combine and their relative weight is argued about:

• •Established categories, which pull new sounds towards familiar ones
• •Motor habits, since the tongue and lips have practised one set of movements
• •Rhythm and stress patterns, which differ between languages and transfer
• •Age of first exposure, with outcomes worsening steadily with later starts
• •Quantity of use, since speaking the first language daily reinforces it
• •Identity, since an accent marks belonging and some speakers keep it

The idea of a critical period after which native pronunciation becomes impossible is widely believed and is more complicated than it appears. Large studies find that outcomes decline steadily with age of arrival rather than falling off a cliff at puberty, which argues for a gradual process rather than a window slamming shut. Some adults do achieve pronunciation that native listeners cannot identify as foreign, which a strict cutoff would forbid, although they are a small minority and tend to have unusual motivation, training or musical ability. The current view is that the decline is real, gradual, and driven by accumulated commitment to the first language rather than by a biological deadline.

The effects on the listener are as important as anything happening in the speaker and are better documented than people expect. Studies show that identical recordings are rated as harder to understand when listeners are told the speaker is foreign, and that a face shown alongside audio changes comprehension scores, which indicates that expectation does part of the work. Accented speech is rated as less truthful in experiments, an effect that survives when content is held constant and that appears to come from the slight extra effort of processing rather than from prejudice alone, though prejudice compounds it. Familiarity helps quickly, with listeners adapting to an unfamiliar accent within a minute or two of exposure.

Pronunciation can be improved substantially and the methods that work are specific. Training perception before production is effective, since people who are trained to hear a contrast with feedback, using many different voices rather than one, subsequently produce it better even without practising production. Explicit instruction about where the tongue goes helps, because adults can use conscious knowledge where infants cannot. Working on rhythm and stress frequently improves intelligibility more than working on individual sounds, since listeners rely heavily on those. And intelligibility rather than nativeness is now the usual teaching goal, on the reasonable ground that being understood is the point and sounding native is neither necessary nor, for many speakers, desired.

Infants can hear every contrast used in any language and lose that by around ten months as the brain commits to local categories, which makes discarded distinctions hard to hear afterwards. An adult who cannot produce a foreign sound often cannot hear it either. Outcomes decline steadily with age of first exposure rather than at a fixed cutoff, and training perception with many voices improves production without practising it.