Isochrony is a linguistic analysis or hypothesis assuming that any spoken language's utterances are divisible into equal rhythmic portions of some kind. Under this assumption, languages are proposed to broadly fall into one of two categories based on rhythm or timing: syllable-timed or stress-timed languages (or, in some analyses, a third category: mora-timed languages). However, empirical studies have been unable to directly or fully support the hypothesis, so the concept remains controversial in linguistics.

Rhythm is an aspect of prosody, others being intonation, stress, and tempo of speech. Isochrony refers to rhythmic division of time into equal portions by a language. The idea was first expressed thus by Kenneth L. Pike in 1945, though the concept of language naturally occurring in chronologically and rhythmically equal measures is found at least as early as 1775 (in Prosodia Rationalis). Soames (1889) attributed the idea to Curwen. This has implications for linguistic typology: D. Abercrombie claimed "As far as is known, every language in the world is spoken with one kind of rhythm or with the other ... French, Telugu and Yoruba ... are syllable-timed languages, ... English, Russian and Arabic ... are stress-timed languages."

While many linguists find the idea of different rhythm types appealing, empirical studies have not been able to find acoustic correlates of the postulated types, calling into question the validity of these types. However, when viewed as a matter of degree, relative differences in the variability of syllable duration across languages have been found.

Three alternative ways in which a language can divide time are postulated:

In a syllable-timed language, every syllable is perceived as taking up roughly the same amount of time, though the absolute length of time depends on the prosody. Syllable-timed languages tend to give syllables approximately equal prominence and generally lack reduced vowels.

French, Italian, Spanish, Romanian, Brazilian Portuguese, Icelandic, Singlish, Cantonese, Mandarin Chinese, Armenian, Turkish, Korean and many Indian languages (such as Hindi, Urdu, Punjabi, Tamil, Odia, Bengali, etc.; to name a few) are commonly quoted as examples of syllable-timed languages. This type of rhythm was originally metaphorically referred to as "machine-gun rhythm" because each underlying rhythmical unit is of the same duration, similar to the transient bullet noise of a machine gun.

Since the 1950s, speech scientists have tried to show the existence of equal syllable durations in the acoustic speech signal without success. More recent research claims that the duration of consonantal and vocalic intervals is responsible for syllable-timed perception.

Some languages like Japanese, Gilbertese, Slovak and Ganda also have regular pacing but are mora-timed, rather than syllable-timed.^{[citation needed]} In Japanese, a V or CV syllable takes up one timing unit. Japanese does not have diphthongs but double vowels, so CVV takes roughly twice the time as CV. A final /N/ also takes roughly as much time as a CV syllable, as does the extra length of a geminate consonant.