Stress-timed - syllable-timed
In the following discussion, no measurements should be regarded as precise; and all claims of linguistic behaviour should be regarded as tendencies rather than absolutes.
Languages in Europe may be grouped, at least in their spoken forms, into two groups by their rhythmic patterns. These are different applications of the general principle of isochrony (~ 'equal times'). In general, the mind likes to generalize spoken output into similarly-sized units. English is a stress-timed language.
- Stress-timed languages, like English, are those in which the time-units that are roughly equal are those marked by the different stresses of the utterance. McArthur gives an example "in the phrase //dozens of/ old / photographs//, dozens of takes about the same time to say as old". (Typical speakers make the old a more drawn-out syllable than the unstressed -ens and of.)
- Syllable-timed languages, for example French and Italian, are those in which the time-units that are roughly equal are those marked by the different syllables. This can lead to the perception of these languages as being 'like machine-guns', having a mechanically regular production of syllables. This may be heard in the recordings of Sarah Bernhardt (1844–1923), such as that at [[1]].
Speakers of a language in one of these groups (and no language is consistently in one alone: all have some features of both) may perceive speakers of languages predominantly in the other group prejudicially: 'she is a machine-gun when she speaks English' was said of an Italian teacher of English; 'he is such a lazy speaker' is often said about speakers of English whose dialect has a habitual lengthening of vowels that their speech is a lazy, or imprecise, drawl - or both.
- It is possible that inaccurate perceptions of this sort contribute to the way in which many American speakers of English realize French words, through a form of hypercorrection: they mis-perceive a French realization of a word which has been naturalized into English. A French speaker pronounces a polysyllabic word, for example café, with little differentiation of stress; an American hearer expects a noticeable stress, and, not hearing it, assumes that the speaker has stressed it in a way the opposite of the English version. That is, the American hears, instead of the (British) anglicized 'CAFF-eh' (IPA: /ˈkæ fe/) or a French neutral 'caff-eh' an apparent 'caff-EH' IPA: /kæ ˈfe/. The American then produces such realizations as 'berr-EH' rather than 'BERR-eh' (IPA: /bə ˈreɪ rather than ˈbɛ reɪ (or ɪ)/, 'ball-EH' rather than 'BALL-eh' (IPA: /bæ ˈleɪ rather than ˈbæl eɪ/), 'clee-SHAY' rather than 'KLEE-shay' IPA: /kliː ˈʃeɪ rather than ˈkliː ʃe (or eɪ)/ and so on
