n the general sense, the emphasis or prominence that some syllables in speech bear over others, regardless of how achieved (the intonational means are via pitch, stress, or length); in the specific sense, a synonym for stress, a dimension of speech not reflected in most Western systems of orthography. Technically, “stress” in linguistics denotes intensity of articulatory force, resulting from greater musculatory exertion in forming a sound.
I. TERMINOLOGY
Prominence of a syllable may be produced by stress, but it may also be produced by obtruded (raised or lowered) pitch (fundamental frequency) or increased length (duration) or (most often) by a combination of these factors, which have been found to correlate highly with one another (stressed syllables are often higher in pitch and increased in length as well); together the three phenomena constitute an “intonational contour.” Up until the 1950s, it was common to speak of a “pitch a.” in tonic (pitch-based) längs, such as Ch., Japanese, and ancient Gr., as opposed to the “stress a.” of the mod. Germanic langs., incl. Eng. But subsequent work (e.g. Bolinger), which was in fact in accord with a long line of theorizing by prosodists (e.g. Bright), suggested that pitch was after all the crucial element of “a.” even in stress langs. At the same time, Allen has suggested that, contrary to centuries of trad., stress not pitch was the basis of a. in ancient Gr. At present one may say only that the phonetic phenomena concerning a. are still not fully understood. Since stress is not readily isolated from other factors, it is still not clear whether stress is after all a real entity in itself or simply the resultant of perceived prominence caused mainly by pitch and length, but the latter view looks more likely.
Usage of the terms “stress” and “a.” consequently shows a welter of variation; Crystal reserves “stress” for word-stress and “a.” for prominence at the phrase and sentence levels (“stress belongs to the lexicon, and a. to the utterance”), but many others speak of “sentence stress.” Bolinger, emphasizing pitch, distinguishes (word) stress, (pitch) a. (on stressed syllables), and intonation (pitch changes at the sentence level). Liber-man and Prince, by contrast, posit stress as a relation between syllables in a hierarchy above the lexical level, rather than assigning it as a feature of the segment (so Jespersen 1901). But these are technical problems in linguistics. In nontechnical usage, the simplest thing to say is that “a.” in the sense of emphasis is the more general term, “stress” the more precise, and that “stress” can denote intensity as opposed to pitch or length. Most of the conceptual disputes in metrics over the past two millennia resulted from confusions about the nature of “quantity” vs. “a.,” deeply interrelated acoustic phenomena whose subtlety and complexity were not much penetrated until the advent of modern linguistics.
II. DEGREES AND TYPES
Monosyllables and short polysyllables, particularly substantives such as nouns, normally bear a stress; this is “word-stress.” In longer polysyllables (which form by compounding shorter words, prefixes, and suffixes), one stress dominates all the others, reducing them to unstressed syllables or to a secondary degree of stress. How many degrees of stress are thus created, and perceived, is a disputed question. The Ren. grammarians, following those in Med. Lat., who in turn followed the ancient Greeks, held that there were three, which they called “acute,” “circumflex,” and “grave.” The structural linguists of the mid 20th c. (i.e. Trager-Smith) claimed there were four (primary, secondary, tertiary, weak, denoted 1-2-3-4, respectively, and based on loud-ness), as in the noun phrase elevator operator, stressed 1-4-3-4 2-4-3-4, though there is some evidence that in ordinary speech auditors do not hear that many degrees. The extra degree added by Trager-Smith is for sentence stress and corresponds to Chomsky and Halle’s nucleus; inside a given word there are still only three degrees. As in the formation of polysyllabic words, so in the formation of phrases one stress comes to dominate all the others within its domain. So too at the level of the sentence: this process of dominance, of forming stress hierarchies, is general. All other things being equal, the Nuclear Stress Rule dictates that within any major syntactic constituent, the rightmost major stress will be primary, weakening all others to the left. At the sentence level, its effect is to make the last major element the strongest.
Word-stress is relatively fixed and is coded into the lexicon (it can be found in the dictionary), being controlled by the phonological and morphological rules of the lang. But phrase- and sentence-stress is more variable, depending in part on syntactic rules but also on context and on the emphases that the speaker wishes to make in that particular context. Word-stress thus controls whether a syllable can be stressed, but semantic intent and other factors control whether it will. Stress at the phrasal or sentence level was traditionally called “rhetorical a.”—so Alexander Gil (1619): “accentus est duplex Grammaticus aut Rhetoricus” (ed. Danielsson and Gabrielson [1972], 124). One result of it is that a speaker can say the same word or sentence several different ways by moving the main a. and mean quite different things. “Próject” is a noun, but “projéct” is a verb. One can say “I never said I loved you,” for example, so as to mean (stressing “loved”) “I may have said I liked you but I never said I loved you” or (stressing the second “I”) “I said Jack loved you” or (stressing “you”) “I said I loved Ann, not you, Diane.”
Such cases as this, where mere shift of stress deeply alters meaning, are instances of the phenomenon known as “contrastive stress,” which Shakespeare exploits to effect in Sonnet 130 and which Donne employs frequently (Melton’s “arsis-thesis variation”). This is an important device in poetry. If, to take another example, we alter Ben Jonson’s line and say, “love me only with thine eyes,” which words shall we stress— love me with thine eyes (don’t give me hateful looks)? love me (only me, not anyone else)? only with thine eyes (not with any other part of you)? Semantic clues from the wider context of the poem may help us decide. But a more important determinant in metrical poetry is meter (q.v.), which is not determined by but must align with and so helps us determine the intended or most probable stressings (meanings) of the line. Meter regulates the placement of stresses in sentences according to a fixed scheme via a process called modulation, in which stress values can be demoted or promoted. Knowing the meter will therefore help one to elucidate sense; conversely, grasping the sense of lines in an unfamiliar poem can help one correctly identify its meter. It is one of the great limitations of free verse (q.v.) that it abandons both these hermeneutic instruments.
III. HISTORY
To give a history of a. in the poetries of the West is really to give a history of our progressive understanding, or clarification, of a. in the poetries in the West. In Gr., a. was, according to both the ancient authorities and most modern scholars (though see Allen), determined by pitch. The quantitative meters of Gr. poetry (q.v.), however, were based on syllabic length When Ennius appropriated Gr. meters for Lat. poetry, there may have been an indigenous native trad, of stress-based meters in Italic—the Saturnian (q.v.)—which was thereafter suppressed from Lat. artverse, surviving only in popular verse-forms such as soldiers’ songs; or it may have served a more important role we do not now well understand (see Beare). In any event, sometime about the 3d-4th cs. A.D. , in some massive linguistic transformation still not understood, a. based on pitch was replaced by a. based on intensity or stress in Lat., with the result that a. replaced syllabic length as the basis of verse composition. Subsequent Old Germanic versification is heavily accentual.
With the decline of Rome, understanding of the (artificial) rules of quantity were progressively lost, and Med. Lat. “metrical verse” (based on quantity) begins to be replaced by “rhythmical verse” (based on a.)—first and most importantly in a popular form of poetry, the hymn (q.v.), expressly developed by Augustine, Ambrose, and Hilary for easy singing and memorization by the illiterate masses. Quantitative verse in Lat. continued to be written in quantity at least up to the 12th c., but only as a scholarly exercise.
With the emergence of the European vernaculars, each Romance poetry had to invent a prosody for itself. The general view at present is that the metrical and stanza forms were taken over from Med. Lat., while the rules for accentuation and syllabification varied somewhat in each lang., their prosodies reflecting this fact also (Norberg, Pulgram). Accents do not have the force in Fr. that they do in the Germanic langs, such as Eng. In Eng., word-stress greatly outweighs word-boundary: the stressed syllable is lengthened. But in Fr. the last syllable in the word is characteristically lengthened, while word stress is reduced in favor of phrasal stress. These phrasal stresses at line-end and before the caesura, called accents toniques, are bolstered on occasion by accents d’appui (“prop” as., made by promotion of weakly stressed syllables) and accents contre-toniques (as. on the antepenult of polysyllabic words without mute -e). The Fr. accent oratoire corresponds to Eng. “rhetorical a.” In sum, Fr. prosody (meter) is phrasal in nature, based on rhythmic groupings in verse identical to those in prose. Eng. prosody, by contrast, developed smaller, nonphrasal groups—feet—each based on one stress. Romance words which came into Eng. (esp. in ME) underwent “recession of a.,” by which Eng. speakers shifted left to the beginning of the word the rightward word-stress of the Fr.
In the Ren. reaction against medievalism, the Humanists looked back to Classicism, and esp. the prestigious trad, of Cl. prosody: consequently, quantitative meters were successively attempted in the 16th c. in every vernacular, first It., then Fr., Eng., and Ger. But length is simply not phonemic in the mod. langs., so any metric based on it is doomed to be artificial. Stress, by contrast, is phonemic; and when independent thinking about poetics also emerged in the Ren., many critics and prosodists recognized this fact—though only dimly, and with no accurate terminology with which to describe what they heard and felt when they read poetry. Most of the Eng. Ren. prosodists (collected in Smith) grasped, intuitively, the existence of stress, but they were unable clearly to distinguish stress from pitch (as indeed many still do not), and the terms which they had available—those descended from antiquity—were of little help. The Ren. grammarians often speak of a raising and lowering of the voice; Lily treats a. s.v. “Tonus,” and identifies three varieties: grave, circumflex, and acute. An accurate terminology only begins to appear in the 18th c., with its renewed interest in the origin of lang. It is abetted materially by the rise of Philology in the 19th c., esp. in Germany, though here too most of the terminology continues to be explicitly Classical. In Eng. one finds fresh air in the discussions in the London Philological Society in the latter 19th c. (Ellis, Sweet). But it is chiefly in the 20th c. that linguistics, now fully emerged as an autonomous discipline, begins to develop instruments and techniques for analyzing precisely the phenomena of acoustic phonetics. As for integrating stress within a unified theory, structural linguistics described stress patterns by assigning levels or degrees of stress (see above); generative phonology, following upon (and imitating) the transformational grammar of Chomsky, attempted to find rules which would generate correct stressings but exclude all others and which would mesh with the rules of syntax (see Liberman and Prince, Selkirk). “Generative metrics” (q.v.) sought to apply such rules to scansion (see Beaver) but failed to grasp the nature of convention (esp. metrical) in poetry. Modern linguistics has swept away much of the detritus of two millennia of confusion, error, and received opinion. But if concepts such as pitch and stress have become clearer, they also have become far more complex. The state of linguistic theory at present may well suggest that our understanding of these intricate phenomena is still far from complete.
J. Hart, The Opening of the Unreasonable Writing of Our Inglish Toung (1551); J. Steele, Prosodia rationalised ed. (1779); A. J. Ellis, On Early Eng. Pronunciation, 2 v. (1868-69), “On the Physical Constituents of A. and Emphasis,” TPS (1873-74); Smith; Schipper, History, ch. 8; Thieme, 390—lists Fr. works to 1912; P. Habermann, “Akzent,” “Takt,” Reallexikon I; B. Danielsson, Studies in the Accentuation of Polysyllabic Loan-Words in Eng. (1948); G. L. Trager and H. L. Smith, Jr., An Outline of Eng. Structure (1951); Beare—still essential; Norberg; P. Habermann and W. Mohr, “Hebung und Page 6 Senkung,” Reallexikon 1.623-29; D. L. Bolinger, Forms of Eng. (1965), pt. 1, Intonation and Its Parts (1986), ch. 2; S. Chatman, A Theory of Meter (1965), chs. 3, 4, App.; J. McAuley, Versification (1966); E. J. Dobson, Eng. Pronunciation 1500-1700, 2d ed., 2 v. (1968)—historical phonology; P. Garde, LA. (1968); N. Chomsky and M. Halle, The Sound Pattern of Eng. (1968)—generative phonology; J. C. Beaver, “Contrastive Stress and Metered Verse,” Lang&S 2 (1969), “The Rules of Stress in Eng. Verse,” Lang. 47 (1971); D. Crystal, Prosodic Systems and Intonation in Eng. (1969), Diet, of Linguistics and Phonetics, 2d ed. (1985); I. Lehiste, Suprasegmentals (1970); M. Halle and S. J. Keyser, Eng. Stress (1971)—better avoided; W. S. Allen, A. and Rhythm (1973), esp. chs. 5, 7, 12, 15, 16—stress in Gr.; E. Pulgram, Latin-Romance Phonology: Prosodies and Metrics (1975); S. F. Schmerling, Aspects of Eng. Sentence-Stress (1976); M. Liberman and A. S. Prince, “On Stress and Ling. Rhythm,” LingI 8 (1977)—metrical trees; Studies in Stress and A., ed. L. Hyman (1977); I. Fónagy and P. Léon, LA. français contemporain (1979); D. R. Ladd, The Structure of Intonational Meaning (1980); Scott, 24-28, 39-55; Brogan, 54-57, 110-19; Mazaleyrat, ch. 4; Morier, s.v. “A.” and “Con-tre-a.,” and also “Consonne,” “Voyelle,” and “Syllabe,” some of the latter very long articles; E. Fudge, Eng. Word-Stress (1984); E. O. Selkirk, Phonology and Syntax (1984)—grid theory; M. Halle and J.-R. Vergnaud, An Essay on Stress (1987); A. C. Gimson, Intro, to the Pronunciation of Eng., 4th ed. (1989); D. Oliver, Poetry and Narrative in Performance (1989).
User Comments Add a comment…