In phonetics, a vowel is a sound in spoken language, such as English ah! [ɑː] or oh! [oʊ], pronounced with an open vocal tract so that there is no build-up of air pressure at any point above the glottis. This contrasts with consonants, such as English sh! [ʃː], where there is a constriction or closure at some point along the vocal tract. A vowel is also understood to be syllabic: an equivalent open but non-syllabic sound is called a semivowel.

In all verbal languages, vowels form the nucleus or peak of syllables, whereas consonants form the onset and (in languages that have them) coda. However, some languages also allow other sounds to form the nucleus of a syllable, such as the syllabic l in the English word table [ˈteɪ.bl̩] (the stroke under the l indicates that it is syllabic; the dot separates syllables), or the r in Serbo-Croatian vrt [vr̩t] "garden".

There is a conflict between the phonetic definition of "vowel" (a sound produced with no constriction in the vocal tract) and the phonological definition (a sound that forms the peak of a syllable). The approximants [j] and [w] illustrate this conflict: both are produced without much of a constriction in the vocal tract (so phonetically they seem to be vowel-like), but they occur on the edge of syllables, such as at the beginning of the English words "yet" and "wet" (which suggests that phonologically they are consonants). The American linguist Kenneth Pike (1943) suggested the terms 'vocoid' for a phonetic vowel and "vowel" for a phonological vowel, so using this terminology, [j] and [w] are classified as vocoids but not vowels. However, Maddieson and Emmory (1985) demonstrated from a range of languages that semivowels are produced with a narrower constriction of the vocal tract than vowels, and so may be considered consonants on that basis.