Introductory Reading on Rhythmic Typology
Snezhina Dimitrova, University of Sofia
The classification of the world’s languages in terms of the type of rhythmic organization they exhibit that is probably still most well-known today is the one proposed by Pike and Abercrombie. According to the theory, there exist two categories or groups of languages: those in which the stresses tend to recur at equal intervals of time (languages with stress-timed rhythm), and those in which each syllable tends to take an approximately equal amount of time (languages with syllable-timed rhythm). In the latter group, the syllable is considered to be the major unit of rhythm, whereas in the former group this role is played by the foot – a stretch of utterance which begins with a stressed syllable and includes all unstressed syllables which follow it. The theory also claims that the two rhythm categories are mutually exclusive, and that every language is characterized by either one or the other of these two types of rhythm.
At the same time, it is surprisingly difficult to decide on a set of criteria for assigning a language to one of the two rhythm types. Attempts have been made to expand the dichotomy by adding to it new rhythm categories (for example, mora-timing), to replace the traditional classification by some other (binary) opposition (for example, the dichotomy “leader-timed” vs. “trailer-timed” languages has been proposed), or to relax the requirement for isochrony of inter-stress and inter-syllable intervals (many researchers prefer to talk about “stress-based and syllable-based“ languages).
However, such attempts haven’t made the task of classifying speech rhythm any easier. On the one hand, there are languages which seem to show features of both stress-timing and syllable-timing. On the other hand, experimental research in support of a dichotomous classification is hard to find, especially as far as the production of rhythm is concerned. Measurements have shown that deviations from strict foot isochrony depend on the number of syllables in a foot, and that deviations from syllable isochrony are related to the number of segments in a syllable. However, scholars have also pointed out that, in speech perception, such deviations often tend to be disregarded and as a result speech tends to be heard as more rhythmical than it really is.
In view of this lack of conclusive empirical support for the stress- / syllable-timing classification, today many scholars prefer to view speech rhythm not as a dichotomy but rather as a scale or continuum, with two hypothetical languages, one perfectly stress-timed and the other perfectly syllable-timed, marking the ends of the scale. The proponents of the traditional rhythm typology say that stress-timed and syllable-timed languages differ in several respects – in terms of the structural types of syllables and their frequency of occurrence, in terms of syllable duration, degree of vowel reduction, etc. But they consider such differences to be “side effects” of the speaker’s attempt to achieve isochrony in production. Alternatively, scholars who treat rhythm in scalar terms claim that the rhythmic differences between languages result from the interaction of a variety of components. According to them, a number of phonetic and phonological, segmental and prosodic phenomena typical of a given language act together, either reinforce or minimize the contrast between stressed and unstressed syllables in that language, and thus bring about our perceptual impression of stress- or syllable-timing. Each component can be broken down into “features” to which a plus or a minus value can be assigned. Thus, each language can be given an overall rhythm “score”, with the help of which the position of that language on the rhythm scale can be determined.
The following components of language rhythm have been suggested by, among others, Dauer and Bertinetto:
1. Vowel quality
(i) Vowels in stressed syllables have full articulation, whereas vowels in unstressed syllables are reduced or centralized.
(ii) Vowels retain their distinct quality in both stressed and unstressed syllables.
(i) Consonants in stressed syllables are articulated more carefully, whereas consonants in unstressed syllables have reduced allophones or are subject to neutralization.
(ii) Consonants have the same articulation in both stressed and unstressed syllables.
(i) If the language is characterized by segmental quantity distinctions, then these are manifested only in stressed syllables; in unstressed syllables they are neutralized.
(ii) Quantity distinctions are independent of stress and are found in both stressed and unstressed syllables.
(i) Stressed syllables are regularly longer than unstressed ones, usually by 1.5 or more, whereas unstressed syllables regularly undergo compression.
(ii) Stressed syllables are only slightly longer, or of the same duration as unstressed ones.
5. Compensatory shortening
(i) Compression of the stressed vowel/syllable as a function of the number of unstressed syllables following it is typical of the language.
(ii) Only a limited amount of compression is observed in the pronunciation of isolated words; alternatively, there is no compensatory shortening in the language.
6. Syllable structure
(i) The language is characterized by a variety of syllable types and relatively uncertain syllable boundaries.
(ii) The language has only a limited number of syllable types: CV and CVC predominate, and processes such as cluster simplification and epenthesis prevent the occurrence of more complex syllable structures; syllable boundaries are relatively well-defined.
7. Relationship between syllable structure and stress
(i) Heavy/strong syllables tend to be stressed, whereas light/weak syllables tend to be unstressed.
(ii) Syllable weight and stress are independent.
8. Nature of stress
(i) Word stress is “free”; it is usually signalled by a combination of length, pitch change, loudness and quality.
(ii) The language has “fixed” word stress, whose phonetic realization usually comes down to only one or two of the features of length, loudness, pitch or quality; alternatively, the language has no lexical stress.
9. Position of stress
(i) The intervals between successive stressed syllables are relatively short, and extra stresses can be inserted to avoid long interstress intervals; stress shift prevents the occurrence of two strong stresses next to each other.
(ii) The intervals between stresses are of variable duration, and the language tolerates relatively long interstress intervals; the occurrence of two consecutive stresses is also tolerated and stress shift is not observed in the flow of speech.
(i) Pitch correlates with stress; stressed syllables are turning points in the intonation contour; emphasis and contrast affect stressed syllables.
(ii) Intonation and accent are independent; emphasis and contrast may affect stressed as well as unstressed syllables.
(i) If the language is characterised by tonal contrasts, they only occur on stressed syllables; unstressed syllables are atonal.
(ii) Tones occur on both stressed and unstressed syllables.
Dauer (1983, 1987)
Wenk and Wioland (1982)