COLLEGE ENGLISH

Volume 28 December 1966 Number 3

Chaucerand the Study of Prosody MORRISHALLE AND SAMUELJAY KEYSER

INTRODUCTION writing in the iambic pentametertradi- lin- In this article we propose to character- tion. (It is not, of course, the only utilized the We ize the accentual-syllabic meter known guistic given by poet.) as iambic in the form in shallrestrict our attentionto stressplace- pentameter ment and the number of which it was first used by Geoffrey syllables only Chaucer. We view this meter as an insofar as they participate in the con- struction of a line of verse. abstract pattern which the poet has single We shall not in created or adopted, perhaps only in part attempt any systematicway to the verse line to more consciously. The poet uses this pattern as go beyond single a basis of selection so that he may choose complex structures.' out of the infinite number of sentences of Let us first turn our attention to the a natural those which qualify for facts of English stress placement. In language word like of inclusion in the poem. In the verse of celestial,a speaker modern knows that the stress is interest here the pattern consists in the English primary of 1 regulation two linguistically given on the second syllable, thus celestial and the number of in a 1 properties, syllables not celestial. line and the placement in a line of sylla- bles stress There are other facts about English bearing linguistically given which are than that of adjacent stress also relevant for pur- greater syllables. of meter. Thus nouns Thus we begin with the assumption that poses compound is a stress placement relevant linguistic 1While this article deals with which is the primarily fact utilized by English poet iambic pentameter,we have quoted on occa- sion relevantexamples from certain of Chau- Morris Halle is Professor of Linguistics and cer's works written in iambic tetrameter,in on the staff of the Research Laboratory of Elec- particularthe Romauntof the Rose (RR), the ironics at MassachusettsInstitute of Technolo- Book of the Duchess (BD), and the House of gy. His published work is concentrated pri- Fame (HF). marily in the area of phonology, particularly Abbreviationsof titles throughout follow in the Slavic languages. The Chaucer study is those of A Concordance to the Complete a result of his long standing interest in problems Works of Chaucer and to the Romaunt of the of poetic form. Rose compiled by John S. P. Tatlock and Samuel Jay Keyser is Assistant Professor of Arthur G. Kennedy (The Carnegie Institution English at Brandeis University. His major in- of Washington, 1927). All quotationscome from terests are in linguistics, the history of the En- the Concordance from The Complete Works of glish language,and the intersection of linguistics Geoffrey Chaucer, ed. F. N. Robinson (Boston, and literary study. 1933).

187 188 COLLEGE ENGLISH such as blackbirds are clearly stressed Beginning with the observation that a 1 2 2 1 his has at his and not the poet (like audience) disposal blackbirds blackbirds. Indeed, certain we ask how latter stress en- linguisticgivens, may assignment automatically these are utilized for tails the sequence being understood as the linguistic givens 2 1 purposes of prosody. There are, of noun phrase black birds and not the com- course, two possibilities.The first is that pound noun. Similarly, if a complex noun the poet develops a metrical form which like kitchen towel rack is stressed as is completely independent of the linguis- 2 1 3 1 kitchen towel rack it is understood as tic givens of stress so that celestial is as meaning 'a towel rack for the kitchen' 1 1 3 acceptable as celestial. This hypothesis as kitchen towel while if it is stressed entails that the poet may radically depart 2 from the basis of his his rack it is understood as meaning 'a rack linguistic (and for kitchen towels.' And if it is stressed as audience's) language. The second possi- 1 2 3 is that the does not violate the kitchen towel rack it is clear that it is un- bility poet linguistic givens of his language but English.2 rather incorporates them into a metrical Still other facts about English are ob- pattern which, while extra-linguisticin vious to the native speaker. Thus articles that the patternis not a fact of the spoken such as a, and the, as in a new home, the language, is nonetheless perfectly com- old man, prepositions such as of or in patible with the linguistic givens of the in phrases such as of the people, in the spoken language. house, conjunctions such as and and or, An advocate of the first view is Ten pronouns such as him, his, etc., are with- Brink4 who supposed, for example, that out stress, or with very little stress." while the facts of Chaucer's spoken En- glish demanded a stress pattern of the 1 21In American English the normal stress pat- sort nonetheless Chaucer vio- tern in adjective + noun phrases is 2 1, thus cominge, 2 1 2 1 lated this linguistic given by actually black bird, old man, etc. Commentators on the stress for metrical British observe that in such shifting purposes stress, however, to with words like thus syntactic units the stress distribution is level, rhyme springe, 1 1 1 2 black bird as opposed to blackbird.Thus Daniel cominge.5 Outline 9th ed. Jones, of English Phonetics, On this view (Cambridge, 1960), ?959 observes, "Foreign the present lines which learners should note particularlythe case of one have been word qualifying another. Both the words have as a rule strong stress."In the ensuing examples black bird in a sentence such as "I saw a black of adjective + noun phrases, he makes no bird" has the pattern 2 1. This is not to say distinction between the level of stress on either that the noun phrase could not, in such a sen- constituent. This suggests that the subordination tence, receive a 1 2 stress pattern, prompted, of the adjective stress to the noun stress in for example, by a request for clarification: noun phrases is a peculiarity of American En- "What color bird?" But such an emphatic stress glish. In what follows, then, we shall assume pattern departs from the normal distribution of stress subordination in compounds but level stresses in simple declarative statements. In stress in noun phrases and the like. It will be general, we shall assume neutral stress distribu- seen below that this assumption is consistent tion in the lines of verse which follow unless with the metrical practice of Chaucer and other there are strong contextual reasons for suppos- English poets. The assumption of the (Ameri- ing otherwise. can) 2 1 stress pattern in these phrases renders 4See B. Ten Brink, The Language and Metre a number of regular lines metrically deviant. of Chaucer, 2nd ed. rev. F. Kluge, trans. M. Examples of such lines will be found on p. 200. Bentinck Smith (London, 1901), 5279. 3These facts are intended to convey a pic- 5For a discussion of this point see M. Halle ture of English stress under neutral emphasis. and S. J. Keyser, The Evolution of Stress in Thus in American English the noun phrase English (forthcoming). CHAUCER AND THE STUDY OF PROSODY 189

devisedso thata wordboundary falls 'iambicpentameter', a poetic meterwhich afterthe tenth,twentieth, thirtieth has been used by English poets from etceterasyllable is written Geoffrey Chaucerto the present.In what in perfectiambic pentameter. follows we shall adopt this name without and to characterizein Moreover,on this view the meterof a prejudice attempt line suchas Keats's: as precise a fashion as possible what we intend by it. Silentupon a peakin Darien Suppose, then, that by iambic penta- meter one meant a metrical form which which is cannot be metricallyregular, adheredto the following metricalprinci- distinguishedfrom that of the title of the sonnet ples: Principle 1. On first looking into Chapman's The iambic pentameterline consists Homer of five feet to which may be ap- pended one or two extra-metrical which is not. Givena principleof stress unstressedsyllables. shift for metricalpurposes, both lines Principle 2. mustbe viewedas regular.But this clear- The iambic foot consistsof two syl- ly is not the case.Indeed, it is precisely lables. the differencebetween these two lines Principle 3. thata successfultheory of prosodyought Each even syllable is strongly to characterize. stressed. In what followswe shalladopt the sec- Principle4. ond alternative,namely that basically the Each odd syllable is less strongly poet doesnot violatethe linguisticgivens stressed. of his in languagebut, rather,attempts We shall refer to the above principles generalto utilizethem in actualizingthe as the strict of the iambic metrical interpretation pattern. pentameter line. Notice that (1) above is a line in termsof these Fromthis pointof view, a line like: regular princi- ples. (1) Celestial, whether among the Thus the line may be scannedas fol- thrones,or named lows: 1 1 1 1 from Milton'sParadise Book Celestial,wheth'r among the thrones, Lost, XI, 1 line 296 (hereafterPL. XI. 296) mustbe or viewed as with an unstressed named.6 beginning now be syllablefollowed by a stressedsyllable But notice that (2) abovemay becauseit is a offeredas a counter-exampleto the iam- precisely linguistic1 given bic thus: that celestialis stressedas celestial.Simi- pentametertheory, 1 1 1 1 of doleful larly, becauseregion is so stressed,we Regions sorrow, shades, supposethat a line suchas: wherepeace (2) Regionsof sorrow,doleful shades, 6For elision of -e before an r which separates wherepeace (PL. I. 65) unstressedvowels in Milton, see Robert Bridges, Milton's Prosody (Oxford, 1921), pp. 29ff. See beginswith a stressedsyllable followed also pp. 206ff. below for our treatment of such an unstressed sequences in Chaucer. In the matter of vowel by syllable. elision as well as in certain others, Bridges It has been customary to identify arrived at conclusions that are substantially these lines as examplesof a meter called identical with those presented here. 190 COLLEGE ENGLISH

This is a counter-example precisely be- ing to conform to is strict iambic penta- cause it is a linguistic given that regions meter, that is, doggerel. This fact in itself 1 1 renders his of the is stressed regions and not regions. The principle "legitimate stress shift" (see fn. 4). strict iambic pentameter theory would suspect Prosodists have, of course, been well force us to regard this line as irregular, whereas in fact lines with "inverted first aware of these difficulties with the strict iambic have feet" are perfectly regular in iambic pentameter theory. They verse. tried to shore up the theory by giving a list of allowable deviations. P. Another to the strict Thus, counter-example Baum lists in his Chaucer's Verse (Dur- iambic pentameter is the theory opening ham, N. C., 1961), pp. 13-14, as poetical couplet of XXXth sonnet: Shakespeare's "liberties and licenses," the following de- partures from the strict iambic penta- When to the sessions (3) of sweet, meter: "(1) the weak stress, including silent thought secondary accent in polysyllables, in dif- I summon remembrance up of ferent positions in the line; (2) the heavy things past foot or spondee; (3) the inverted foot or trochaic; (4) the variable use of syllabic In the first line of (3) the strict iambic e; and (5) elision, contraction, and the theory requires that the preposition to slurs which may or may not certify the receive greater stress than when, and of admission or exclusion of tri-syllabic receive greater stress than sweet in viola- feet." tion of the linguistic givens of spoken In very similar terms Wimsatt and English. Moreover, it requires that in the Beardsley in their important article on phrase sweet, silent thought the adjective "The Concept of Meter" (PMLA, sweet receive less stress than silent. But LXXIV, 1959) speak of "the inverted since this stress pattern is in direct viola- first foot, the dropping of the first slack tion of English, the line must be classed syllable, the extra slack syllable internal as deviant. to the line (elided, or not elided in the Similarly, the second line of (3) would anapest) . . ." as of deviations from the also be classed as irregular by the strict strict iambic pentameter that "occur so iambic pentameter theory, for clearly the often as to assume the character of an preposition of has less stress than the accepted complication of the norm." following noun things. Thus the strict The introduction of "allowable devia- iambic theory entails that lines which tions" constitutes a significant modifica- abound in the writings of the best poets tion of the theory. The modified theory are metrically deviant. Evidently such a is clearly an improvement over the strict theory cannot be considered tenable. iambic interpretation in the sense that un- like the latter it is not contradicted a It is obvious on other grounds as well by fraction of the lines it is that this interpretation is untenable. Thus significant sup- to characterize. But even with these the strict interpretation says, in essence, pose modifications the does not recom- that a poet writing in this meter is writ- theory And it little mend itself especially, for it fails to pro- ing doggerel. requires poetic vide to know that any explanation for the fact that sensitivity Chaucer, Milton, certain deviations are tolerated and and Lowell do not write only Pope doggerel. not others. Since the allowed deviations Indeed, when Ten Brink assumes legiti- share only the of included mate of the stress of the property being shifting spoken in a list, why could not other deviations language to conform to the meter, it is also be included in such a list? Rather noteworthy that the meter he is attempt- than look upon headless lines, lines with CHAUCER AND THE STUDY OF PROSODY 191

an inverted first foot etc., as somehow study of prosody is the study of the ab- deviant-though perhaps less deviant than stract patterns-the different arrange- some other lines-we propose below a set ments of linguistic givens-that underlie of principles or rules which by their na- all performances of a given poem; it is ture yield a large variety of metrical pat- not the study of the myriad ways-some terns, in the same way that rules of syntax good, others bad, most indifferent-in yield a large variety of syntactic patterns. which a poem might be recited. With respect to these rules, there will be one of two possible judgments. Either MIDDLE ENGLISH STRESS a line is metrical virtue of by conformity It is obvious that a theory of prosody to the or else a line is unmetrical rules, which takes the linguistically determined by virtue of nonconformity to the rules. stresses of a language as part of the ele- we have it is this dis- said, precisely ments manipulated by the poet assumes tinction.s which a of of a theory prosody an understanding of what these stresses or tradition must given poet poetic are. In other words a theory of make.7 prosody based, in part at least, upon stress The rules we establish an ab- place- propose ment necessarily a of stract that is satisfied presupposes theory pattern by particu- stress placement. In modern such lar of English arrangements linguistic givens. a theory of stress be are not to be with placement may They equated precepts taken for granted since the for a of prosodist, performing particular type the poet and the reader, as speakers of the verse.s The meter of a poem determines same have all internalized the to a extent the manner in which the language, great same system of stress assignment. In dis- poem is to be performed. It never deter- Middle we mines the how- cussing English prosody may performance completely, not make this assumption. We must, ever, more than a score of a sonata any therefore, give a brief summary of the completely determines the way in which Middle stress the sonata should be The English system, especially performed. as our picture of it does not coincide in all with the one found in the 7In this connection we cite the points following handbooks. passage from Robert Bridges' A Letter to a Musician on English Prosody which, in our We begin by considering two separate view, captures precisely the role of prosody in issues: poetry: "What, then, exactly is Prosody? . it denotes the rules for the treatment of sylla- the rules for stress subordination bles in verse, whether they are to be considered i) as long or short, accented or unaccented, eli- in compounds and other se- deable or not, etc., etc. The syllables, which quences larger than a single are the units of rhythmic speech, are by nature word of so indefinite a quality and capable of such different vocal expression, that apart from the Seymour Chatman, "Robert Frost's 'Mowing': desire which every artist must feel to have his An Inquiry into Prosodic Structure", Kenyon work consistent in itself, his appeal to an Review, XVIII, 1956, pp. 421-38, "The present audience would convince him that there is no analysis attempts to describe the verse line as it chance of his elaborate rhythms being rightly is actually 'performed.' It avoids the unfortu- interpreted unless his treatment of syllables is nate assumption that performances involve 'ex- understood. Rules must therefore arise and be ceptions' to some kind of norm. In fact, it agreed upon for the treatment of syllables, and suggests that the poem as document may be this is the first indispensableoffice of Prosody" lifeless until it is actualized into sound pattern. (italics ours MH/JK). The essay from which The analysis starts with the performance, not this passage is drawn is reprinted in a useful the norm; and it suggests that the metrical anthology edited by Harvey Gross entitled The tension of much successful verse lies in the Structure of Verse, Modern Essays on Prosody poet's subtle modification and evasion of the (New York, 1966), pp. 86-101. expected, plus the performer's keenness in in- 8As is frequently done; for example, by terpreting the poet's intentions." 192 COLLEGE ENGLISH

ii) the rules for stress assignment 4. A good Wif was ther of biside in a single word Bathe (A.Prol.445) With to the former we assume regard The traditional view concerning the that in and ad- compound nouns, verbs, rule for stress assignment in simple words the main stress of the second jectives has been that Middle English possessed element is lower than that of the first. two kinds of words, native words and We shall assume that Chaucer stressed Romance words. Similarly, it was sup- 1 2 1 2 1 2 Middle contained brymstoon, greyhound, cartwheel. On posed, English two the other hand, in the constituents that stress rules, one for the native portion and are not nouns, verbs or adjectives and one for the Romance portion. The that contain more than one word, such rule for the native portion was assumed to be identical with that for the bulk of as noun phrases composed of an adjective Germanic Old En- followed by a noun, we shall suppose that languages, including stress subordination does not occur. Thus glish, namely a rule which assigned pri- 1 1 1 mary stress to the initial syllable of words, a phrase such as the first prize or a useful and the like. this 1 excluding prefixes By book we shall treat as "level 1 1 1 1 having rule father, mother, become, coming, re- stress". It is important to note that in this ceived their stress. The rule for the regard we are departing from contem- Romance of the Middle American in which stress portion English porary practice vocabulary was assumed to be the Old subordination does occur in such phrases, French stress rule which was 2 essentially 1 like that for classical yielding patterns like the first prize, a Latin, though with 2 1 certain modifications: useful book (see fn. 2 above). These prin- ciples of stress subordination account for Romance stress rule: the fact that a) the first syllable of com- 1. Assign primary stress to the final pounds consisting of two monosyllabic vowel of a simple word if that vowel words commonly occupies an even posi- is long. tion, whereas b) any constituent of a 2. If the vowel is short and followed by syntactic unit that is not a compound number of consonants, noun, adjective or adverb be found any including may none, then look at the next to the last in an even or an odd position. Examples syllable. illustrating a) are: 3. If the penultimate syllable is strong, 1. Out of the donghil cam that word that is, contains a long vowel, or any ful right! (PF.597) vowel followed by two consonants, 2. Men clepen hym an 'outlawe' or a then assign major stress to the vowel 'theef' (H.Mcp.234) of that syllable. 3. In al the toun nas brewhouse ne 4. If the penultimate syllable is not taverne (A.Mil.3334) strong, then stress the antepenultimate 4. Twelve spokes hath a cartwheel syllable. comunly (D.Sum.2257) Part 1. of this rule is responsible for Examples illustrating b) are: 1 wolde the stress in licour. Part 2. and Part 3. 1. Oold fish and yonge flessh 1 I have fayne (E.Mch. 1418) are responsible for the stress in arrerdge 2. His eyen caste on hire, but in sad 1 rhyming with age. (A.Prol.601). Part 4. wyse (E.Cl.237) of this rule is responsible for the stress 3. The Millere was a stout carl for the nones (A.Prol.545) in words like taffata or Pandarus. CHAUCER AND THE STUDY OF PROSODY 193

It is commonly believed that a speaker vowel. As a consequence, the Romance of Middle English like Chaucer pos- stress rule would assign precisely the sessed both rules and that he systematical- same stress to these words as would be ly applied one rule for the native portion assigned to them by the Germanic stress of the vocabulary and one for the rule, namely initial stress. Thus a word foreign. Since modern English clearly like father, mother or bodi would re- does not have both rules, a change must ceive initial stress by either rule. On the have occurred. And it is believed by other hand, ultimate or penultimate stress many that the direction of that change could only be assigned in words like was toward complete acceptance of the pilgrimdge, pite, merci, nature, etc. by Germanic rule and rejection of the the Romance stress rule. Thus, of the two Romance rule. A typical example of this rules, the Germanic stress was largely dis- view is expressed in the following com- pensable; the Romance stress rule was ment by Henry Sweet (A New English not. Grammar, ?786, Oxford, 1891): It must be recalled that beside final 1 "In Old French the stress generally fell stressed licour, for example, with a long on the same syllable as in Latin, as in final vowel, there was also a second na*ture= Latin nd*turam.Through the variant with initial stress and short final dropping of final Latin syllables many vowel. In the light of this, it becomes French words thus came to have the obvious that the initial stress of words 1 1 1 stress on the last as in o*nour , , , syllable, like licour pite, merci, nature, is not due ho*ndrem,pi*te pie*tdtem.When first in- to these words treated as if troduced into ME French words being they kept were native words, but rather is due to their stress: original nJ*tiire,o*nur, pi*tj; the fact that the final vowel in these but such words afterwards threw the long words was shortened to the stress back on to the first the simply prior syllable by operation of the Romance stress rule. analogy of the native E. words, such as * Indeed, if we suppose that the shortening *fader, bodi becoming* ndtiire, etc." of long final syllable vowels was an op- tional rule in Chaucer's language, then We have argued elsewhere9 against the 1 1- traditional view. We have proposed that both liceur and licour receive stress by at the stage of Middle English when the the Romance rule, the latter form having Romance words first entered the lan- first undergone optional shortening of guage there must obviously have been the long final syllable vowel.1' two stress rules, but that quite soon the Germanic stress rule and dropped the 10The precise environment for shortening of Romance stress rule predominated. We long vowels in final syllables is slightly more shall not present here the details of our complicated. Thus a long final vowel or a vowel before a final consonant argument; rather we shall touch the long single may upon optionally shorten. A long vowel before more major points. than one final consonant will not shorten, how- To begin with, given a stage of En- ever. Thus merci and licour are subject to glish in which both stress rules were shortening; servaunt is not. We operative, it is crucial to note that the argue in Evolution of Stress in English (forthcoming) that this rule was a native En- majority of Middle English words from the glish rule which already existed in the language native portion of the vocabulary were prior to the influx of Romance words. It is either monosyllabic words or else dis- responsible for the short final vowels in words words but syllabic with a short final like wisdom, stirrop and wurthliche (from earlier wisdom, stigrap and wurthlic). Romance 9The Evolution of Stress in English, (forth- words which entered the language simply fell coming). within the domain of this rule. 194 COLLEGE ENGLISH

A similar systematic alternation among postulate the rule which drops final -e certain words in Middle English has and the rule which shortens long vowels been noted by many scholars, among in final syllables in any case, then it is them Ten Brink."1 Specifically, Ten clear that the assumption that the Ger- Brink noted that initial stress in words manic stress rule continued to operate like manire, banJre was often accom- in Middle English after the Romance rule panied by loss of final -e. He assumed entered the language is unnecessary. The that the loss of final -e was due to initial facts of Middle English are equally well stress and that initial stress was due to accounted for without the assumption. these words being stressed according to At this point, however, there are cer- the Germanic stress rule. Let us suppose, tain facts about alternations in Chaucer's however, that the loss of final -e, like line which become quite important. the shortening of final long vowels, was Notice that a word like cominge could an optional rule in Chaucer's Middle only receive initial stress, according to English. Let us further suppose that this our view, if the final -e first dropped. rule, like the final vowel rule, applied to Thus a form like cominge would be a word before the stress rule applied. stressed 1 the Romance stress Then we may account for the initial cominge by 1 rule, but if the final -e is stress in these worsd as well. Thus manJre coming op- 1 tionally dropped prior to the operation and banere receive their stress by Part 2 of the Romance stress rule. Thus we and Part 3 of the Romance stress rule. If, have a systematic alternation between however, the final -e dropping rule were initial stress in a dissyllabic word, and to apply to these forms first, then the penultimate stress in a tri-syllabic word. rule which shortens long vowels in final This puts certain severe constraints on syllables could operate next to yield certain lines in Chaucer. That is to say, if we find a line in which the scansion baner and maner which would then re- requires that a word like cominge be ceive initial stress by Part 2 and Part 3 1 -, tri-syllabic, then it must be the case that of the Romance stress rule, thus baner such a line also requires that cominge and maner. In a similar fashion receive penultimate stress. Otherwise, our nature of stress would lead us would become natdir by the final -e theory placement to lines in Chaucer. dropping rule, then natar by the final suppose exceptional 1 And the supposition of exceptional lines rule and natfr syllable shortening finally must always be the last resort of the the Romance stress rule. by prosodist. that Thus not only is the assumption For example, a line like: 12 the Germanic stress rule operated in late Middle but the as- But now thi coming(e) is to be so English unnecessary, swete sumption that the Romance stress rule (TC.4.507) was in Middle in con- operative English clearly requires that coming be initially with the final -e rule junction dropping stressed and dissyllabic. Thus this line is and the vowel rule now ex- shortening perfectly straightforward. But consider the be- plains systematic correspondence lines such as the following: tween initial stress and short final sylla- bles and between initial stress and lack The cause of his comynge thus of final -e, a correspondence which, in the traditional view, is accidental. 12The enclosing of e in parenthesesindicates Finally, that the rule which final -e has when one considers that it is to drops applied. necessary Thus orthographic coming(e) is is phonetically 11Ten Brink, ?257. comingis. CHAUCER AND THE STUDY OF PROSODY 195

answerde (TC.2.1102) which he puts to use in passages such as The cause v-told of hir comynge, the following: the olde (TC.4.141) The nexte houre of Mars folwynge Gret sweryng is a thyng abhomina- this (A.Kn.2367) ble, Ther maistow seen cominge with / / ' / / And fals swerynge is yet moore Palamon (A.Kn.2128) / reprevable. In these lines the meter demands that The heighe God forbad swerynge comynge be tri-syllabic and penultimate- Iv stressed.13The view that the German- at al, ic stress rule in Middle operated English Witnesse on Mathew, but in special could not possibly account for the neces- sary penultimate stress in these native Of sweryng seith the hooly Jere- words. The view that the Romance stress mye rule operated supplies precisely the de- "Thou shalt swere sooth sired stress to render these lines regular. thyne such as and Rhymes cominge springe othes, and nat lye, further support the view that these words / / / 1 l 1 (4) And swere in doom, and eek in were stressed cominge: springe.'4 Moreover, the existence of stress doub- rightwisnesse" But is a lets such as comyng beside comynge pro- ydel sweryng cursednesse, vides Chaucer with a metrical resource Bihoold and se that in the firste

13In the line Ther maistow seen cominge table, with Palamon the meter also requires that Palamon be treated as occupying two metrical Of heigheGoddes heesteshonurable positions. That this is permissible in Chaucer Hou that the second heeste of will be shown below (pp. 206ff.). The alterna- (e) tive which treats with initial stress cominge hym is this: renders the line exceptional. 14There are two other conceivable explana- "Take nat my name in ydel or tions for rhymes like -inge : springe. One is to deny the assumptionthat in Chaucer only stress amys." vowels bearing may rhyme. But then the ab- rather he forbedeth swich sence of rhymes like the: he, an: man, etc., Lo, would be merely a coincidence. The assumption that only stressed vowels rhyme offers a prin- swerynge cipled explanation for the absence of such Than or a cursed rhymes. homycide many The second possibility is to suppose that thynge; (C.Pard.631-44) -inge bears a 2 stress. This is itself based upon the hypothesis that -inge bore 2 stress in OE. Grammar (Oxford, 1959), ??204.8, p. 383.) It seems to us that the view of 2 stress in -ung Thus the supposition that OE -ung, even if it in OE is questionable. But even if one accepts did bear a 2 stress, became unstressed in ME that position, it is noteworthy that the ME accounts quite naturally for the shift from -ung reflex of OE -ing/-ung is -ing(e). The shift to -ing(e). Otherwise the shift in vowel quality from u -+ i is commonly explained as due to is quite inexplicable. alternations of vowel quality in unstressed syl- Yet a further argument against 2 stress in lables. (See Bruno Borowski, Zum Nebenakzent -inge is that it implies that the metrical behavior beim 1 2 altenglischenNominal-kompositum (Halle, of cominge, etc. would be identical to com- 1921); Joseph Wright, Middle English Gram- 1 2 mar, 3 ed. (Oxford, 1925), ?134; Sievers-Brun- pound nouns like brimstoon. This prediction, ner, Altenglische Grammatik (Halle, 1951), however, is not borne out. (See The Evolution 5142; Anm. and Alistair Campbell, Old English of Stress in English, ch. 1, forthcoming.) 196 COLLEGE ENGLISH

The alternation from one line to the next of prosody to be outlined below contains 1 1 the three rules: of the variants sweryng and swerynge following (for elision see 206ff. below) offers pp. 1. Final -e be a striking example of the way in which may optionally dropped Chaucer makes use of these at the end of a word. linguistically 2. A vowel in a final given variants for poetic effect. Indeed, long syllable may without the of such variants be shortened. recognition 3. The Romance stress rule. operating here, this passage loses both its metrical form and its effect. poetic (Note In this brief excursus we have been also other alternations here, operating informal. Nevertheless the such as that between srwere as a stress necessarily maximum 195 in the preceding pages embody the major points (see p. above) of the of stress seventh line but not in the sixth and the theory placement which accounts for the data Chau- internal off rhyme in the first line of provided by 1 1 cer's poetry. and over the rteryng thyng 1 against Before we leave this topic and proceed 1 to a detailed discussion of the of feminine swerynge : thynge in the theory rhyme which accounts for Chaucer's final lines of the passage.)15 prosody verse, this of stress In summary, then, the theory of stress assuming theory place- ment, an and placement which is entailed by the theory interesting quite indepen- dent result of this theory is worth men- Elsewhere it has been shown 15The passage from the Pardoner'sTale con- tioning. stitutes a variation on Robinson's text based that the rule which governs placement of upon our view of the operation of final -e stress in modern English words is essen- and its effect on stress placement. The reader the Romance stress rule stated notice that in certain in tially will instances, particu- with several modifica- lar thynge, we assume a final -e which was not above, although present in the nominative and accusative in Old tions due in large part to the great influx English. We base our assumption of a final -e of learned words in the 16th and 17th in on the well-known occurrence in thynge centuries. The fact that our assumptions Chaucer of forms with inorganic -e extended lead us to to the nominative and accusative cases from the postulate the Romance stress rule for Middle as well can hard- oblique cases. The evidence for such forms is English cited in Ruth Buchanan McJimsey, Chaucer's ly be a coincidence. Rather it seems that, Irregular -E (New York, 1942). With respect in the course of the development of to thynge, she states that no -e is required in the Germanic stress rule was 374 scansions" and 154 She English, "easy in rhymes. rendered the influx adds, however, that in 16 other scansions a rea- superfluous by large sonable metrical variation would allow a final of Romance words into English after the -e and she also notes the springe : thynge rhyme Conquest. This rule dropped out and the in RR 2627 with springe in the infinitive. She Romance stress rule remained from Mid- further notes the line Thynge that is to reprove as RR 7546 which is deviant without a final -e. dle English into present day English rule for of Because she finds only 16 examples requiring the dominant the assignment final -e, she concludes that the statistical evi- stress in simple words. dence legislates against thynge with an inorganic -e. This seems to us overly hasty in view of two THE THEORY OF CHA UCERIAN facts. The first is the existence of inorganic -e PROSODY in Chaucer in other nouns (for a list of which, see McJimsey, pp. 19-20). The second is the We have characterized above the strict need for an optional -e dropping rule. If it were iambic pentameter interpretation of the case that every example of an -ynge rhyme could only be explained, according to us, by the are paired with infinitives and finite verb forms assumption of an inorganic -e in the rhyme where the -e is historical. Thus the assumption word, then our position would indeed be of an inorganic -e does not seem at all implausi- weakened. But a large number of -ynge rhymes ble. CHAUCER AND THE STUDY OF PROSODY 197

Chaucerian verse by means of a set of Principle 3. principles or rules. These principles taken A stress maximum may only occupy together constitute a theory of prosody, even positions within a verse, but not albeit an inadequate one. The principles every even position need be so occu- assume and arrange in various patterns pied. certain theoretical entities, such as the Definition. foot, which nowhere appear in the actual A stress maximum is constituted line of verse. Prosodists have, in general, by a syllable bearing linguistically recognized the theoretical nature of their determined stress that is greater than entities. Thus, Paull Baum (p. 11) asserts that of the two syllables adjacent to that "Chaucer's line is a series of five it in the same verse. iambs," and then adds in a footnote, "For this flat statement there is to be sure only Each principle embodies several alterna- deductive evidence. If it must be re- tives, and the possible employment of garded as in the first instance an assump- Condition 1 and/or 2 increases further tion or an hypothesis, it can be tested in the number of alternatives. The order in the usual of corroboration and ac- ways which the alternatives are given is sig- counting for apparent exceptions." nificant. Thus the first alternative within The that we for the theory propose each principle represents what we shall Chaucerian iambic pentameter is framed consider the most neutral actualization of in terms of metrical constructs which are, the metrical pattern. The subsequent al- to a certain familiar and relates extent, ternatives yield lines which are perfectly these in a manner to the se- particular regular in that they violate no rule. How- of sounds that make quences speech up ever, they represent a more complex ac- a given line: tualization of the metrical pattern. Thus a line in which each position is occupied by 1. Principle a single syllable and in which each even The iambic verse consists pentameter position is actualized by a stress maximum of ten positions to which may be ap- is deemed neutral. Lines, however, which one or two extra-metrical pended syl- exhibit even positions without stress lables. maximum, or lines which exhibit poly- 2. Principle syllabic (or empty) positions are con- A is a position normally occupied by sidered to be more complex actualiza- but under certain con- single syllable, tions of the metrical pattern.'6 ditions it be more may occupied by Differences in complexity of an anal- than one or none. syllable by ogous sort are found in syntax. For Condition 1. Two vowels may constitute a that 16We have omitted the constraint that Chau- single position provided they lines because it no role in or are a cer's rhyme plays adjoin, separated by liquid what follows, though a complete list of Chau- or nasal or by a word boundary cer's principles must clearly include rhyme, as which may be followed by h-, and well as stanza construction, etc. The one aspect that one of them is a weak- of rhyme in Chaucer which we have called provided attention to fn. is that for two se- stressed or unstressed vowel. (see 14) ly quences to rhyme they must bear some degree Condition 2. of stress (as well as share a sequence of identical An unstressed or weakly stressed segments to the right of the stress bearing word constitute a segment). monosyllabic may It will be seen that the outlined metrical with a principles single position pre- above may be readily formalized; cf. Appendix ceding stressed or unstressed syl- below, where a formal statement of the princi- lable. ples may be found. 198 COLLEGE ENGLISH

example, a sentence such as The man in the first two lines of (5) the tenth bwhocasts the first stone is innocent is as position is not occupied by a stress maxi- regular as The man the first stone is cast mum since the environment to the right by is innocent, though it is obviously less of the tenth syllable contains no syllable. complex than the latter. In the last line of (5), however, the extra- With respect to metrics, differences metrical syllable renders the first syllable in have been / complexity traditionally of rede a stress maximum. The occur- viewed as instances of metrical tension. rence of a stress maximum in the tenth The with this difficulty view, however, will on whether is that it confuses the essential distinction position depend there is an extra-metrical syllable added since between more complex and less complex Chaucer normally places a stress bearing lines with the distinction between regular syllable in the tenth position. This is and irregular lines.17 related to Chaucer's rhyming practice SCANSION which requires that only syllables bear- some of stress As Let us turn, then, to the To ing degree may rhyme. examples. a the stress maximum in the with, the most neutral actualization consequence, begin tenth is without and of the iambic line is, as we position interest, pentameter we shall it in all scan- have stated, one in which each ignore subsequent syllable sions. occupies one metrical position and in It should be obvious, however, that in which each even position is occupied by a stress maximum. the same way that a stress maximum does not occur in the tenth of a line The following lines are typical: position without an extra-metrical syllable, so, too, a stress maximum may never occur Hir heer was un- brighte kembd, in the first syllable of a line. A conse- tressed al (A.Kn.2289) quence of this is that lines in which the / / / / first syllable bears stronger stress than (5) And, sooth to vitaille seyn, greet the second syllable are as metrically reg- plentee (B.ML.443) ular as those in which the stronger stress falls on the Ye shal be deed, by myghty Mars second syllable. / Examples of such lines are: the rede! (A.Kn.1747)

In these lines we use the / to indicate Greyhoundes he hadde, as swift as that the so marked is a stress syllable fowel in flight (A.Prol. 190) maximum; that is, it has greater stress / / / than the Thus this (6) Kepte hir estat, and both of yonge surrounding syllables. and old mark expresses a relationship between (TC.1.130) the syllable and its environment. It says I, your Alceste, whilom quene of nothing whatsoever about the degree of Trace (LGW.432) stress of the syllable, merely that, what- ever it is, it is greater than that of-the two The lines in (6), so-called "inverted neighboring syllables. first feet," are extremely common in Notice that as a consequence of our Chaucer, so common in fact that pro- definition of stress maximum we find that sodists have been forced to allow" for them as being, not wholly iambic as are 17The concept of a more or less neutral lines in of a the (5) but, nonetheless, "per- actualization line is closely related to the missible deviations." In terms of the concept of markedness in linguistics. (Cf. N. Chomsky and M. Halle, The Sound Pattern of theory presented here, the lines require English, forthcoming.) no special comment. They are a natural CHAUCER AND THE STUDY OF PROSODY 199 consequence of a theory which bases its Examples in the fourth position are: on stress maxima and prosodical analysis And weddede the even Indeed, this queene Ypolita position occupancy. (A.Kn.868) theory provides a natural explanation for the occurrence of lines such as these, (7) In armes, with a thousand shippes, while, in terms of a modified strict iambic wente (TC.1.58) pentameter theory, there is no more rea- To for the and do son to to find this devia- preye peple, expect particular servyse (D.Sum. 1897) tion than one, say, in which the second Examples in the sixth position are: and fourth feet are occupied by trochees. It was observed by Jespersen in his Was sent to Corynthe in ful greet "Notes on Meter" that while iambic honour (C.Pard.604) lines have an inverted first foot, may (8) That whilom japedest at loves trochaic lines may not. Thus, for in- stance, Longfellow's peyne (TC.1.507) Life is but an empty dream That I wol lette for to do my is rendered unmetrical if the initial tro- thynges (B.NP.4279) chee is an as in replaced by iamb, Examples in the eighth position are: A life's but an empty dream or And when he cam, hym happede, chaunce (C.Pard.606) To live's but an empty dream par The reason for this between assymmetry (9) Of vyongefolk that haunteden folye iambic and trochaic lines is clear once (C.Pard.464) it is realized that trochaic verses have stress maxima on odd in The droghte of March hath perced only positions to the verse and that an initial iambic foot the roote (A.Prol.2) locates a stress maximum on the second (i.e., on an even) position in the line, The careful reader may, at this point, in direct violation of the trochaic prin- wonder about the first line in (7). Thus ciple. the proper name Ypolita is not shown A further consequence of the rules, in above with a stress maximum on the ante- particular of Principle 3, is that one penultimate syllable so that, in effect, this would expect to find lines in which sev- line is actually an example of a line with- eral even positions, namely, the 2nd, 4th, out stress maxima in the fourth and 6th and 8th, are not occupied by stress eighth positions. Our position on this maxima. Indeed, one ought to find lines question must remain open. It strikes us in which this is true of only one even as reasonable that Ypolita would contain position per line and also of more than a secondary stress on the ante-penulti- one. The only constraint is that a line in mate syllable. Indeed, in polysyllabic which all the even positions are unoccu- words in general one would expect to pied by stress maxima is highly unlikely find in Middle English as in contemporary in view of the natural stress patterns of English secondary stress placement else- the language. (We will see below, how- where in the word. The difficulty is that ever, that such lines do occur.) our only source of stress in words is the The lines in (6) already provide ex- Romance stress rule (see above) which amples of verses without a stress maxi- supplies primary stress only. Therefore, mum in the second position. To turn then it is impossible to say with certainty to examples in which other even posi- (though quite possible to conjecture) tions are not actualized by a stress maxi- that words like citrinacioun, superfluytee, mum, we may cite the following lines: abhominable, Ypolita, etc. had 2 stressed 200 COLLEGE ENGLISH

some rule similar to svllables, perhaps by abhominable the rule which distributes 2 stress in By superfluytee 2 1 (C.Pard.47) modern English in words like Oklahoma, 1 1 1 3 (11) Of thy religioun and of thy bileeve Mississippi, or in words like hurricane, (G.SN.427) 1 3 / advocate (verb). In view of this limita- And of oure silver citrinacioun tion on our knowledge of secondary (G.CY.816) stress placement in Middle English we shall take the conservative (and no doubt (Notice that if superfluytee is stressed that words are 1 artificial) position simple superfluytee, a possible option in Chau- supplied with only a single stressed syl- cer, then the first line in (11) will be re- lable. It should be pointed out, however, classified as exhibiting no stress maxima that if stress is as- secondary placement in the second, fourth, and eighth posi- sumed to have been similar to that in con- tions.) this will not result in temporary English, Lines without stress maxima in rather lines only exceptional lines; supplied two even are far more common. with stress will positions secondary simply require without stress maxima in the reclassification. Examples sixth and eighth positions are: Consider, for example, the following line: Nis nat to yow of reprehencioun (TC. 1.684) So dul of his bestialite (10) vs (12) And how this town come to de- (TC.1.735) struccion (TC.1.141) In terms of our convention this line In so unskilful an oppynyon exhibits a stress maximum in the second (TC. 1.790) position only. If secondary stress were Examples of lines without stress max- 2 2 1 ima in the fourth and the word would be eighth positions supplied, bestialite. are: This would not make (10) a counter- example, but rather a line which exhibits In Southwerk at the Tabard as I lay stress maxima in the second, sixth and (A.Prol.20) In we shall eighth positions. general, then, (13) The to wreken of such alternative ravysshing ignore analyses especially (TC. 1.63) since we have been unable to uncover Eleyne under lines which, when supplied with secon- The smyler(e) with the knyf the cloke dary stress, provide counter-examples to (A.KN.1999) of lines without stress max- our theory. Examples We mentioned earlier that within the ima in the fourth and sixth positions are: space of a verse the likelihood of finding Cucurbites and alambikes eek phrases which contain only one linguistic (14) (G.CY.794) stress is and so with two extremely small, As out of alambik, ful faste stresses. As a lines l(cour consequence, lacking (TC.4.520)is three stress maxima are not common. very 1 The line in (10) above is a possible ex- 1sWe assume a basic form alambik which ample of a line without a stress maximum receives final stress by the Romance rule alone. in the fourth, sixth and eighth positions. Notice that the vowel shorteningrule, if ap- Examples without stress maxima in the plied, would yield alambikwhich would be second, sixth and eighth positions are: stressed alambik by the Romance stress rule. CHAUCER AND THE STUDY OF PROSODY 201

Examples of lines without stress max- secondary and lower stress raised by the ima in the second and eighth positions examples in (15) as well as the last two are: lines in (7), the last line in (9), the last two lines in (11) and the last line in (16). And for the trouthe I demed in his All of these lines contain a preposition or herte (F.Sq.563) pronoun which has not been treated as a stress maximum. It is clear that there is a (15) And by his svde a swerd and a fundamental distinction between major bokeler (A.Prol. 112) categories like Noun, Adjective, Adverb And on the daunce he gooth with and Verb on the one hand, and minor Canacee (F.Sq.277) categories like conjunction, preposition, Examples of lines without stress max- pronoun, article and certain verbs like ima in the second and sixth positions are: to be on the other. It is also clear that words was major category always bear stress. "Nabugadonosor god," seyd (e) But the facts are not so clear when we he (B.Mk.3752) deal with the minor categories. Thus the (16) This Pandarus so desirous to serve rhyme evidence of Chaucer suggests that (TC. 1.1058) certain words like the and a never bear while certain other words Whan that with his shoures stress, like Aprille and he do bear at least soote (A.Prol. 17)19 to, so, by stress, in certain circumstances.21 Since we do (Notice that in the second line of (16) and this that of the two stress one might want to read Pandarus with suggests options for Pandarus,the end stressedone is correctin initial stress, but this reading would spoil 1 1 view of the internal rhyme produced. (For a the internal rhyme Pandarus: desirbus.) 20 discussion of the stress options in proper nouns, There is a second issue with to see Halle and Keyser, forthcoming.) respect 21Consider the three examples in the Con- cordance in which to occurs in rhyme. These 0. "Whan that are: 19Robert Evans, Aprill(e)?," To do al that a man to Notes and Queries, new series, IV (June, 1957), bilongeth the resultsof an extensive (E.Mch.1459) pp. 234-237,presents That on a his maisterto manuscript comparison of the various readings tyme seyde for the line of the Tales. He (G.Cy.1449) frst Canterbury And ful I to (BD.771) points out that 8 mss. support the reading devoutly prayed hym / , The most obvious fact about these three lines Aprillebut that this is by no meansconclusive. is that each ends with an invertedprepositional He says (p. 237), "As there is no conclusive phrase. Thus the normal belongeth to a man textual evidence, I submit that there is a strong has become a man bilongeth to, to his maister possibility, even a probability that Chaucer in- has become his maister to and to hym has be- tended the initial lines of the Prologue to be come hym to. It is perfectly possible, then, that read as a regular decasyllable, perhaps with a to, normally without stress, received stress as a trochaic substitutionin the first foot." Moreover result of this special inversion. Indeed, a phe- in Northern dialects of English the word is nomenon precisely like this occurs in modern pronounced [alprail] (Cf. e.g., W. Grant and English in a special construction in which T. M. Dixon, Manual of Modern Scots [Cam- inversion is quite regular. Consider, first, the bridge, 1921] p. 62) pointing towards a histori- sentence John ate up the apple. In this sentence cal antecedent with a tense i. 2 the normal stress contour is something like John 20Since desirous has a long T (the simplex 43 1 desir rhymes with shire, etc.), there are only ate up the apple; that is, with ate receiving less 1 stress than the adverbial up and up two possible stress patterns, namely desirous receiving less than the object noun apple. But notice that by the Romance rule alone, or dessrous by the it is quitenormal in Englishto invertthe object prior operation of the vowel shortening rule in noun phrase and the adverbial preposition to final syllable before no more than a single seg- yield a sentence like John ate the apple up. ment. Only the former is possible in this line Such an inversion also changes the stress con- 202 COLLEGE ENGLISH

not at the moment see any way in which stress maxima. We have been able to find a definitive judgment can be made with examples for eleven of those possible respect to the relative stresses among combinations. The combinations for words which belong to the minor cate- which we have been unable to find ex- gories, we have assumed that all of them amples are: (a) lines without stress maxi- bear the same degree of stress, whatever mum in positions 2, 4, 6 and 8; (b) lines that may be. without stress maxima%in positions 2, 4 Needless to say, further research may and 6; (c) lines without stress maxima require that our views be modified. It is in positions 2 and 4. The absence of (a) also true that contextual factors may pro- and (b) type lines is probably due to vide a guide as to stress of minor category their rarity in the language. That is, in a words. In all of the examples cited above, space of ten syllables one is likely to find however, the context of the lines seems at least three stresses but rarely less. The to be quite neutral with respect to the absence of (c), however, is suggestive. relevant minor category items. But it is It may be accidental. On the other hand, worth mentioning here that a modifica- it may be that Chaucer consciously tion which demands that to, under normal avoided lines in which the 2 and 4 posi- stress, be given a greater linguistic stress tions were both unoccupied by stress than and or the (as in the last line in (9) maxima, perhaps because of the weak above) does not render the lines in ques- onset imparted to such lines. In this case tion exceptional; instead the lines need the absence of all three types would be be only reclassified. explained as violations of this added In the lines in (6) through (16) we constraint. We shall merely suggest the have not found certain combinations of possibility of incorporating some such even position without stress maxima by stipulation into Principle 3. virtue of linguistically determined weak Up to this point we have dealt with or unstressed syllables being adjacent to even positions occupied by unstressed or one another. There are fourteen possible weakly stressed syllables rather than by combinations of even positions 2 through stress maxima. There is, however, a sec- 8 being occupied and/or unoccupied by ond way in which even positions may be unoccupied by stress maxima; namely, 2 4 when a inter- tour,which is now somethinglike John ate the major syntactic boundary 3 1 between two metrical The is venes positions.22 apple up. importantpoint here that We shall that neutralization of stress inversionhas provided the adverbialpreposi- say tion up with greaterstress than it would have takes place when two main stresses are receivedwithout inversion. Without attempting separated by a major syntactic boundary, to delineatethe rules stressis whereby assigned and we shall attempt to demonstrate that, in these two sentencesin modernEnglish, we like the so-called inverted draw attentionto the fact that a rule which first foot, neu- tralization is also a of would providea word with greaterstress under consequence the inversionthan it would normallyreceive is not withoutparallel. That to in each of its rhyming 22See Jespersen, for example, "Notes on occurrencesin Chauceris the resultof inversion Metre", Linguistica (Copenhagen, 1933), p. 255: is clear. That it receives stress as a result of "This leads us to another important principle: inversionis plausible.Therefore, while we may the effect of a pause: If I hear a syllable after a take to to havea linguisticallydetermined stress pause it is absolutely impossible for me to know in MiddleEnglish, it is by no meansclear that whether it is meant by the speaker as a strong the evidenceof rhymeforces us to do so under or as a weak syllable: I have nothing to compare normaloccurrences. The fact that prepositions it with till I hear what follows. And it is ex- were subject to the Great Vowel Shift lends tremely difficult to say with any degree of cer- furthersupport to the assumptionthat to was tainty what is the reciprocal relation between stressed,since only stressedvowels were subject two syllables separated by a not too short to Vowel Shift. pause." CHAUCER AND THE STUDY OF PROSODY 203 theory presented here while it gives rise In this line the seventh position is occu- to a host of examples which, in terms of pied by a strongly stressed syllable. But the strict (or modified) iambic pentam- our theory states not that odd positions eter theory must be treated as devia- may not be occupied by strongly stressed tions. We shall begin by citing examples syliables, but only that odd positions may of neutralization. not be occupied by stress maxima. And Examples of neutralization between the in the above line we see that the strongly second and third positions are: stressed syllable in the seventh position is not a stress maximum since it is not pre- 1 1 / / ceded a of lesser stress. By God, right in my litel closet by svllable The (17) vonder (TC.3.663) important point here, however, is that the 1 1 / reason that the seventh position is pre- And sevd(e), "Here at this closet ceded by a position containing a syllable dore withoute (TC.3.684) with equal (that is, neutralizing) stress is that there is a major syntactic bound- An example of neutralization between ary intervening between the two posi- the fourth and fifth positions due to a tions. preceding parenthetical phrase is: What Jespersen termed "pauses" (see 1 1 / fn. 22) are, in fact, major syntactic breaks (18) Therfore, as frend, fullich in me within the line. More often than not thev assure (TC.1.680) are represented orthographically by com- mas, semi-colons or colons. Further, it is of neutralization between the Examples a fact about the rules of English stress sixth and seventh positions due, respec- placement that they operate within but tively, to the major break between an not across major syntactic breaks. Thus introductory adverbial clause and the stress subordination will be found within main clause and be- following (elliptical) major categories but not across major tween two coordinate sentences are: syntactic breaks. For this reason, it is / / 1 1 only at "pauses" that one will find two And whan youre prey is lost, woo equal stresses back to back, i.e. absence of (19) and penaunces (TC.1.201) stress subordination. Examples of major / 1 1 Love hath the be of syntactic breaks are the breaks between bvset wel; good an and the cheere! (TC.1.879) interjection following phrase as in (17), or between a phrase of direct Now notice that whereas all of the quotation and the following direct quote in above are treated as "acceptable compli- as (17), or between a parenthetical cations of the norm" in the strict (or expression and the following clause as in modified) view of iambic pentameter, (18), or between the clauses in a com- they are perfectly regular in terms of the plex sentence such as, for example, the view presented here. sentence we have just considered.23 Thus Principle 3 asserts that a stress maximum may only occupy an even 23Wimsattand Beardsley(p. 596) comment within the verse, but that not on the line Wondring upon the word, quaking position for drede (E.Cl.358)as follows: "Here is a every even position need be so occupied. very special relation of phrase to meter. The And in all of the above this requirement double inversion, at the start of the line and is met. Consider, for example, the second again after the caesura,gives the two participial line in (19): verbs a special quiver. But this depends on the fact that thereis a meter;the inversionsother- / 1 1 wise would not be inversions." Love hath byset the wel; be of good In terms of the principles outlined above, cheere! this line is perfectly regular. Indeed, it is 204 COLLEGE ENGLISH

A particularlystriking case of neutrali- often than not, due to initial interjec- zation due to lack of stresssubordination tions are: across breaks are items major syntactic 1 1 / / in a series. Consider, for example, the Lo, he that leet hymselvenso kon- following well-known line in Chaucer: nynge (TC.1.302) 1 1 / / 1 1 1 1 and richest of As (23) 0, Salamon, wvs (20) ook, frr(e), birch,asp(e), alder, richesse (E.Mch.2242) holm, (A.Kn.2921) 1 1 1 1 1 1 popler O wombe! O bely! 0 stynking cod which is reminiscent of the following (C.Pard.534) line from Milton: There is, however, an environmentin 1 1 1 1 1 1 which neutralizationmay come about (21) Rocks, caves, lakes,fens, bogs, dens without the intervention of a syntactic and shades of death. boundary. We noted earlier (see fn. 2) that, unlike AmericanEnglish which ex- In both of these lines the linguistic hibitssubordination in noun phrasescom- stressesin the even syllable positions are posed of an adjective + noun, con- neutralizedby the precedingand follow- temporary British English shows level ing stresses in the first six positions so stress in these constructions. Assuming, that the first six positions contain no then, the absence of stress subordination stress maxima.These lines are precisely in adjective + noun phrasesin Chaucer, like those in (6) and (16). Thus (20) one ought to expect to find lines like all of the preceding lines with in- in which either member of such con- ternalneutralization is a perfectly regular structionsmay occupy an even or an odd thoughmore complex consequenceof the position in the line. (This is so since, theory of prosody presentedhere. given a sequence old man with level Other which be added 1 1 examples may stress, old man, neitherword to (20) are: namely may be a stress maximum.) And, indeed, we 1 1 1 1 1 find this to be the case. That hot, cold, hevy, lyght, moyst, Examples of neutralization in noun (22) and dreye (PF.380) are: 1 1 / / phrases Yong, fressh and strong in armes desirous (F.Sq.23) The Millerewas a stout carl for the nones (A.Prol.545) (The first of these is a headlessline, for 2 2 / / which see below.) A good Wif was ther of biside Bathe Examplesof neutralizationbetween the (24) (A.Prol.445) first and second which / 2 2 positions are,more Alias! a foul thyng is it, by my feith (C.Pard. 524) recisely like the second line in (19) above. 2 2 / texhibits a strongly stressedsyllable in the A wys womman wol bisye hire first and neutralizationbetween the / position evere in oon sixth and seventhpositions. In otherwords, this (D.WB.209) line is a normaliambic line. Moreover,both lines are distinguishablefrom unmetrical,i.e. Lines such as those in (24) are quite exceptionallines, such as, for example,the sen- commonin Chaucer.Further, making the tence on pp. 187-8 above. The differencein natural assumptionthat level stress was complexitybetween the aboveline and neutral characteristic of British from lines like those in (5) is reflectedin the fact English that this line is derivedby the later alternatives Chaucer to the present, it ought to be of Principles1 through3. the case that such lines occur throughout CHAUCER AND THE STUDY OF PROSODY 205

English poetry and this, too, seems to that a position, normally occupied by a be true.24 single syllable, may, under certain condi- We have seen that all of the lines dis- tions, be left vacant or, under yet differ- cussed in (5) through (24) are regular ent conditions, be occupied by more than lines in terms of our principles. We have a single syllable. Let us turn to the condi- also seen that whereas (5) contains the tion under which a position may be most neutral actualization of the metrical occupied by no syllable. pattern, (6) through (24) exhibit more HEADLESS LINE complex actualizations of the pattern. These are lines characterized by even A zero syllable may occur only in the which are not actualized positions by first position in a line. Such a line is called stress either because of neutrali- maxima, a headless line. An example of such a zation or because of the adjacency of line is: syllables with linguistically determined Weakstress. The number of metrical lines (26) Twenty bookes clad in blak or reed of this is We have type extremely large. (A.Prol.294) already seen examples of many of them above. A particularly good example of In this line the first phonetically realized the interaction of these phenomena is the syllable is, in fact, in the second metrical following line which from the point of position. The first position has been real- view of the strict (or modified) iambic ized by zero. Notice also that in this line pentameter theory, would be wildly the second metrical position is realized deviant: by a syllable which does not constitute a stress maximum since it is not surrounded (25) "Knowe ich hire aught? For my by syllables of lesser stress. 1 1 An of a headless line with love, telle me this. (TC.1.864) example the second position occupied by a weakly This example as well as those in (6) stressed syllable in contrast to (26) above through (24) should be sufficient to illus- is: trate the manner in which a variety of complex lines may result from the alter- (27) Of that word took hede Troilus natives our provided by principles. (TC.1.820) METRICAL POSITIONS An often cited line of Chaucer which Up to now we have focussed our at- is relevant here is the well-known Par- tention on the character of the stress son's adage: maxima in Chaucer's iambic pentameter That if what shal iren line. We have in general chosen as our (28) gold ruste, do? examples lines in which there is a one to (A.Prol.500) one between correspondence syllables A is to treat ruste as and But there are other possible reading positions. map- dissyllabic. The stress contour of gold pings possible. Thus Principle 1 states 2 1 that the iambic pentameter line consists ruste would be gold ruste and the line of ten positions, and Principle 2 states would be a straightforward example of the Chaucerian iambic without a stress 24Severalexamples of lines from later poets maximum in the second position. But this arecited in Jespersen's"Notes on Meter"which violates the sense of the line and the indicatelevel stress.To quote just one from clear contrast between and Shakespeare:The courseof truelove never did implied gold run smooth (Mids. 1.1.134).Here, level stress iren. Thus, without dealing here with the in truelove rendersthe line regular. mechanism whereby emphatic stress may 206 COLLEGE ENGLISH

be assigned to any word in a line,25 we weakly stressed position of the iamb must be to consist of two shall assert that gold ruste is what supposed syl- simply lables, with a. But because this the line requires. We shall read (28) as namely follows: violates the canon of one position, one syllable, scholars have either treated the line as or, as did Ten Brink, That if gold rust(e), what shal iren irregular do? emended it to read with thredbare cope. The important point to keep in mind, That is, we take it as a headless line however, is that the one position, one without a stress maximum in the second syllable doctrine is an hypothesis and no position. (We assume ruste to be mono- reasons are given why it should be syllabic by operation of the final -e drop- honored. There is, however, a suspicion ping rule.) that the reason for its widespread accep- tance is due to a prior assumption. The DISSYLLABIC POSITIONS: effect of this doctrine is to CONDITION I essentially allow there to be a one to one mapping No discussion of the Chaucerian line between the theoretical entities proposed can be complete without a discussion of and the phonological entities which ac- the metrical status of several lines which tualize them. But there is no reason why are, in traditional terms, viewed as under- there should be such a relationship be- going various processes of syllable reduc- tween observables and the theoretical tion such as elision, apocope, aphesis, constructs which explain their behavior. svncope, etc. What is essentially at issue In fact, in the spoken language the as- in dealing with these lines is the question sumption of such a relationship is clearly of what constitutes a metrical position for false.26 One consequence of assuming a Chaucer. To begin with, notice that the one to one relationship between con- strict iambic pentameter view described structs and their phonetic realizations earlier is quite explicit about this ques- is to confuse two issues which ought to tion. It states that a metrical position be kept clearly apart. The first issue is must be occupied by a single syllable. what constitutes the metrical structure As a consequence, when faced with lines of a given line. The second issue is how like: a given line should be performed. Thus when Paull Baum Verse With a thredbare as is a (Chaucer's cope, povre [Durham, N. C.] p. 65) comments on a (29) scoler (A.Prol.260) line like: And leyde it above upon the myddeward (G.CY. 1190) (30) Wyd was his parisshe and houses fer asonder (A.Prol.491) it is necessary to recognize them as whether it is better to hypermetrical lines since a mapping of asking "squeeze down to one than to unstressed syllables and stressed syllables parisshe syllable into the iambic foot results in an extra welcome the anapest," he is confusing left over. In the first line syllable being 26A similar aboutthe in the stressed assumption relationship (29) strongly syllable between the sounds of language,the phones, position of the iamb is actualized by the and the theoreticalentities to which they cor- first element of thredbare. Thus the respond,the phonemes,has been assumedby many linguists.For a discussionsee N. Chom- sky, "CurrentIssues in LinguisticTheory," in 25See, for example, N. Chomsky, "A Trans- Fodor and Katz, pp. 93 ff. Note also the state- formational Approach to Syntax," The Struc- ment by Bridgesabove "whether . . . [syllables] ture of Language, ed. Fodor, and Katz (Engle- ... are to be consideredas ... elideableor not wood Cliffs, 1964), pp. 227 ff. ." (seefn. 7). CHAUCER AND THE STUDY OF PROSODY 207

the issues. The question of squeezing But now consider: parisshe down is a question of perfor- / / / / mance. We shall here about And bathed everv vevne in swich say nothing licour how this and similar lines are to be (32) (A.Prol.5) per- 1 1 / / / formed. Rather we shall consider only O, Salamon, wvs and richest of the first issue, namely whether there is richesse (E.Mch.2242) some of systematic way determining Here if we could treat the last two what syllables a sylla- may occupy single posi- bles of and the last two tion from a metrical point of view.27 every syllables We have already seen that the realiza- of Salamnonas occupying a single posi- tion, these lines would be tion of a as zero is possible only quite regular. position Otherwise they would have to be treated when that position is the initial position of a verse. We have also asserted that as exceptions since they would exhibit stress maxima in odd Notice be a positions. :anyposition may occupied by single in syllable. the svmbol to that these lines the relevant sequences Using are and an un- Miarkthe domain of a we see -ery -amon; that is, position, stressed vowel followed an unstressed that the third line of (5) has each of its by syllable. metrical positions actualized Ib a single syllable. thus: Other lines which would be quite regular if there were a principled way of such to a (31) Ye shal be deed, by mvghty Mars assigning sequences single posi- the rede! tion are well-known. Among them are: (A.KN.1747)s- ,C7' / / / 27Notice, then, that it is no accident that the Of Engelond to Caunterburv they processes which scholars have assumed to make wende (A.Prol. 16) unruly lines regular have been phonetic proces- / ' / / ses, such as syncope, apocope, elision and the like. (33) To Caunterburv with ful devout The reason for this is to preserve a one to one corage (A.Prol.22) between the line and the // / mapping phonetic A likerous mouth moste han a metrical structure, and this, as we have sug- gested, results in confusing the performance of a line with the structure of the line. In what likerous tavl (D.WB.466) follows we shall attempt to describe the condi- If the lines could be treated with tions metri- -bury whereby syllables may participate in and -erous cal positions. constituting single positions We shall try to demonstrate that certain then again the lines would be restored principles were used by Chaucer which deter- to regularity. mined the constituency of metrical positions in We mentioned earlier that there were much the same way that analogous principles certain conditions to which a determined metrical in classical according constituency of was to a poetry. The phonological and morphological sequence syllables assigned constraints will be quite different, of course, single metrical position. As a first ap- but the conditions will function similarly. In proximation of one of these conditions, precisely the same fashion as a weak position we formulate the in a classical dactyl can be occupied by one or following: two (- = so a metrical syllables --,), position I in Chaucer's pentameter can be occupied by Condition one or two syllables. This, however, does not Within the same word an unstressed mean that in Chaucer's recitation a two syllable vowel followed a or nasal position was pronounced as a monosyllable any- by liquid more than it means that two syllable positions in classical verse were pronounced as mono- mark the domain of those positions which, by syllables. virtue of the conditions under discussion, are 28In this line we have marked the domain of occupied by more than a single syllable. All each position. In general this will not be neces- positions occupied by a single syllable will be sary. In the lines which follow we shall only left unmarked. 208 COLLEGE ENGLISH

followed by another unstressed vowel that it may only apply within the same may constitute a single metrical posi- word and sequences like many a, glori(e) tion. and, et va be analyzed as many5a and glori(e) and, etcoSince elision may also By Condition 1 all of the in examples occur before an h- initial word, a se- (32) and (33) are rendered regular. like be Thus, and -erous quence contrari ) hire would -ery, -amon, -ury analyzed as that constitute metrical by contrari(e) hire. Indeed, single positions should be so altered is Condition 1. But now consider the foi- Con;dition 11 sug- gested by the following lines which re- lowing set of examples: quire analyses like the above: And from shires / - specially every And manv~aS breem and many/ a ende (A.Prol. 15) luce in stuwe (A.Prol.350) (34) Of thy religioun and of thy bileeve (G.SN.427) (35) With muchel glori(e) and greet / / O hateful harm!/ condicion/ of solempnytee (A.Kn.870) // poverte (B.ML.99) To eschue, and by hire con- In these lines we have in words like trari(e) hire oppresse (G.SN.4)29 and condicion se- specially, religioun Notice that by Condition of unstressed vowels with- extending 11 quences but to across word a line out an operate boundary, intervening liquid or nasal; thus like the is rendered -io- and -ia-. If these could following perfectly sequences regular: be assigned to a single position, (in the fashion indicated in (34) above), then / / Of hire -and eek of his also would be restored to these comynge regularity (TC.3.1675) lines as well. We can do this by making a simple adjustment to Condition 1. In this line a final -e on is Where it now that two un- comynge specifies necessary in order to achieve stressed vowels must be an penultimate separated by stress by the Romance stress rule while or nasal, we intervening liquid simply the fact that the final -e is followed by a stipulate that two unstressed vowels may vowel initial word insures that -e and but need not be an inter- separated by will be assigned to a single position by vening liquid or nasal. Thus the condition Condition now 11.30 appears as: In a line like: Condition 1I Within the same word an unstressed And setten tyme of metyng eft yfeere vowel followed by an optional liquid (TC.3.1712) or nasal followed by another unstressed vowel may constitute a single metrical o29n a count of the various elisions of final position. -e before a vowel or h- initial word in the prologue to the Canterbury Tales, Alexander Ellis, in his There is, however, a Early English Pronunciation,V strong similarity (London, 1869-89), pp. 341 ff., lists 315 as the between Condition 11 as now stated and total number of elisions of final -e before a the traditional concept of vowel elision. following vowel initial word and 147 as the Since the latter deals with final vowels total number of final -e elisions before a follow- h- initial word. Thus it seems that Condi- in words which are followed by vowel ing tion 1 should be altered so as to apply across initial words, all that needs be done to word boundaries. Condition 12 is to drop the requirement 30For discussion of also see fn. 32. CHAUCER AND THE STUDY OF PROSODY 209 the initial stress on metyng indicates that position provided that they adjoin, or the final -e has been dropped prior to the are separated by a liquid or nasal or by operation of the Romance stress rule. a word boundary which may be fol- Thus, in this line, there is no final -e lowed by h- and provided that at which can occupy a single position with least one of the vowels is a weakly eft. Notice that tyme of is ambiguous stressed or unstressed vowel. since it with the may represent tym(e) of As now Condition 1 will not of the rule which stated, only prior operation drops handle the discussed in final -e or else without the sequences (32) tyme'of prior but it will also account for of that rule with -e through (35), operation of assigned those in In en- to a P1. (36). particular, they single position by Condition and to es- are this Condition is almost in final form. single positions by 11 Note also that neither There certain lines which condition."3 posi- are, however, tion is a stress maximum indicate that it must be modified still occupied by since neither of the more. Consider the to eschue svllable dissyllabic sequence is surrounded of in the last of Condition re- positions by syllables line (35). 11 lesser stress. that an unstressed vowel be fol- quires Since Condition allows an another unstressed vowel. But I optional lowed by word and h- to in- we have seen earlier that to may sonorant, boundary already tervene between two the se- indeed bear a certain of stress. vowels, degree -er his constitutes a it is that some condition quence single posi- Thus, possible neither of which is a stress other than lack of stress is sufficient for tion, syllable maximum, in the following line: the assignment of more than one syllable to a Consider, for ex- single position. (37) Withoute bake met(e) was ample, the following lines: /e-.' never his hous (A.Prol.343) And where and of theyVengendred Up to now the examples have dealt (36) what humour (A.Prol.421) with Condition I producing dissyllabic of which ToeFschue, and by hire contra- positions composed syllables were not stress maxima. It is equally ri(e) hire oppresse (G.SN.4) possible, of course, to have such dissylla- In the first line we should like our bic positions occupied by stress maxima. That is to there are lines in Chaucer condition to treat they engendred as say, which illustrate the operation of Condi- and to eschue as to es- they engendred, tion I to yield dissyllabic even positions chue. As Condition now stands it will occupied by a stress maximum. Consider, 1I for the not provide these analyses since they and, example, following: to have some stress. But no- / - / . very likely, And his unwist is to / tice that in each case the sequences to be comynge every analyzed do not involve stress maxima. wight (TC.3.912) if the Thus, we were to drop require- And Phyllis also for Demophoun ment in Condition 11 that both relevant (BD.727) vowels must be unstressed, then the se- / / discussed above will automatic- (38) Trille this pyn and he wol vanysshe quences anon ally be assigned to a single position. We (F.Sq.328) shall reformulate Condition 1I, therefore, The cause y-told of hir comynge, as: 31The between Condition1 and Condition 1 similarity the classicalconditions of elision, echthlipsis Two vowels may constitute a single andsynaloepha is obvious. 210 COLLEGE ENGLISH

acter of the line. But notice that the -e the olde (TC.4.141) is not subject to Condition 1. However, WXvdwas his parissheand houses the sequence the olde is a single position fer asonder (A.Prol.491) and olde is a stressmaximum. In the first line of (38) to ez- constitutes The fifth line in (38) containsthe word a single position by Condition 1 and, parisshewhich may be treatedas a single since ev- containsgreater stress than the position by Condition 1 and the first which a surrounding sy7llables, it constitutes a syllable of constitutes stress stress maximum.Thus, the eighth posi- maximum,regardless of whether the final -e tion is both dissvllabicand occupied by has been droppedor not since a vowel a stress maximum.The line also requires initialword follows. stresson which can One final of a line which con- penultimate coynygyge example only be achieved if the final -e has not tains a dissvllabicposition should suffice. been dropped. But since -e un- consti- Considerthe last line in (36) above: tutes a single position by Condition I which does not contain a stress To eschue, and by hire contrari(e) maximum, hire this causes no difficulty. Finally, the se- oppresse quence -ery in ezery is also a single posi- We have already seen that to es- is a tion by Condition 1, which is not occu- single position by Condition 1. But now pied by a stress maximum so that the consider contrarie hire. We must first first line is quite regularlyiambic. assumethat the final -e has dropped in In the second line in (38) DemophounZ, 1 with bears final stress. order to yield initial stress on contrari rhyming Jasoun, and maintainthe iambic characterof the But Demo- constitutes a single position Condition 1 not a stress line. But the sequence -i hire may also by occupied by constitute a metrical maximum.32 The third line in (38) con- single position by tains a zero first and there- Condition 1. Finally, since hire is sur- position is, rounded But notice that by by completely unstressedsylla- fore, headless. Con- and 1 constitutes a bles, since it very likely bears some ditioil vaIIysshl(e) single of stress, it is a stress position. Moreover.the pusition contains degree probably a stress maximum.(WVAhether the final -e maximum. Thus -i hire constitutes a was dropped by the final -e dropping single metrical position by Condition 1 rule or whether it is to be treated as a and is occupied by a stressmaximum. single position with the following a- does not matterhere.) DISSYLLABICPOSITIONS: The fourth line in (38) requires a CONDITION 2 final -e in in order to the co~mynge yield There is one more condition for posi- appropriatepenultimate stress in comynge tion assignment.The evidence for this which will preserve the iambic char- condition is contained in the following lines: 32Thereis someevidence which suggeststhat also might bear a 2 1 stresspattern in Middle "But we were lever(e) than al this English. Without attemptingto resolve the issue here, note that even assuming a 2 1 stress toun," quod he (A.Mil.3751) the line is since also / / / pattern regular Phyllis may Turne over and / constitute a single position (by Condition 1) the leef chese occupied by a stress maximum. Thus under the another tale assumption of 2 1 stress in also, the line is (A.Mil.3177) scannedas: (39) The cook yscaldedfdor al his longe And Phyllis also for Demophoun ladel (A.Kn.2020) CHAUCER AND THE STUDY OF PROSODY 211

Thow hast translated the Romauns stressed syllable. But notice that the pre- in fact be a mono- of the Rose (LGW.225) ceding syllable may N syllabic word and not necessarily an / /7 / In many places were nyghtyngales unstressed syllable which is part of a (RR.657) word. Thus Condition 21 not only ac- counts for the in and In these lines an unstressed or examples (39) (40) weakly but it also accounts for the following: stressed monosyllabic word is preceded an unstressed in by syllable; particular, Of a and a frater- -er, the past tense -ed and the plural -es. solempn(e) greet nitee (A.Prol.364) We can set up a second condition for position assignment which would say As wel of this asf other thynges something like the following: moore (D.WB.584) Condition 21 For hvm was lever(e) have at his An unstressed or weakly stressed beddes head (A.Prol.293)"" monosyllabic word may constitute a / / single position with a preceding un- (41) And also war him of a significavit stressed or weakly stressed syllable. (A.Prol.662) And saugh his visag(e) al in an So stated, Condition 21 accounts for -er, / than, -er the, -ed for, -ed the, -es of and other kvnde (A.Kn.1401) -es were as a constituting single position. this so a all the are without I ne saugh yeer myri(e) 'lMoreover, positions (A.Prol.764) stress maxima. Without Condition 21 the compaignve lines in (39) would have to be treated as And is also fair and fresh of flour irregular. But there are other lines which (R.R.4333) Condition also renders Thus 21 regular. This is al and som, he heeld virgini- the following illustrate the assignment of tee (D.WB.91) the verbal endings -eth and -est to a single metrical position along with the The sequences of a, as of, at his, of a, following weakly stressed monosyllabic in an, I ne, And is and This is all consti- word: tute single metrical positions by Condi- tion 21 which are not occupied by a stress maximum. Thus all of the lines in And Emelv(e) hvm loveth so ten- (41), as well as those in (39) and (40) drely (A.Kn.3103) / / are regular. (40) No man hateth his flessh, but in his Let us now return to the lines in (29). lyf (E.Mch.1386) Condition 21 will account for the first of these lines since it with a If thou lovest thyself, thou may assign to a single metrical position which is not lovest thy wyf (E.Mch.1385) occupied by a stress maximum. Thus: In these lines -eth so, -eth his and -est With a thredbare cope as is a thy occupy single metrical positions by Condition 21 which do not contain stress povr (e) scoler maxima. is quite regular. (The adjective povr(e) Condition 21 now specifies that any unstressed or weakly stressed monosylla- bic word may constitute a single position 33This reading seems preferable to one which with a preceding unstressed or weakly would treat lever(e) have in this fashion. 212 COLLEGE ENGLISH

is monosyllabic by operation of the final ceding pekke just as was done with leyde -e dropping rule.) it in (29) above. But what of the stress But now consider the second line in maxima assignment? A major syntactic (29). As it now stands Condition 2' will break exists between up and the follow- not apply to leyde it since Condition 21 ing adverb right. Thus neutralization oc- requires that the syllable preceding the curs and the up is not a stress maximum. weakly stressed word, in this case it, be But notice that Pekke also is not a stress unstressed or weakly stressed. But sup- maximum since it is not surrounded by pose we were to drop that requirement. syllables of lesser stress. Thus the line ex- Then leyde it would constitute a single hibits an interesting example of a poly- position and, since ley de is a stress maxi- syllabic first metrical position, occupied mum, the position would be actualized by Pekke hem, and neutralization be- by a stress maximum. Since this altera- tween the second and third positions. tion in Condition 21 does no violence to The line is perfectly regular. the examples discussed up to now, we Now let us consider the second line in shall adopt this modification. Condition (42). The relevant phrase is seven(e) 21 now reads: hennes. By Condition 1 the sequence -en Condition 2 henn- may be assignedas a single metri- cal since the two vowels are An unstressed or weakly stressed position a sonorant, a word bound- monosyllabic word may constitute a separated by metrical with a ary and an h-. Moreover, henn- is a stress single position preced- maximum. Thus the second is ing stressed or unstressed syllable. position occupied by a stress maximum and the By Condition 2 the second line in (29) line is regularly iambic. (We assume is treated as: prior operationof the final -e dropping rule in sevene.) And leyde it above upon the One final line is worth noting. The myddeward. single example of the proper name Attilla occurs in the line: Thus we see that Condition 2, like Condition to 1, may operate produce (43) Looke, Attilla, the grete conquerour which but polysyllabic positions may (C.Pard.579) need not be occupied by stress maxima. Two lines which appear in every dis- If we suppose a single -1-, then the cussion of hypermetrical lines in Chaucer Romance stress rule provides initial stress bear some comment. These are: as Attil(l)a. In this case the line would exhibit neutralization between the first Pekke hem up, r ght as they growe, and second positionsand is quite regular. and (42) ete hymin (B.NP.4157) If, on the other hand, the 1 is geminate, Seven(e) hennes for to doon al his then the Romance stress rule provides pleasaunce (B.NP.4056) penultimate stress as Attilla. In this case In the first line the phrase Pekke hem up Condition 1 may not assign -illa to a is the relevant phrase. Note, first, that up single metrical position since it requires very likely bears greater stress than the that there be a single intervening sonor- words as it does in the preceding just ant. However, Condition 2 will assign -a 63 1 modern English equivalent pick them up. the to a single metrical position and the By Condition 2 hem may be assigned as position will not be occupied by a stress a single metrical position with the pre- maximum. By this latter interpretation, CHAUCER AND THE STUDY OF PROSODY 213

the line is headless. Since we have no way lis (see (38) above) indicate that it of judging whether Attilla contains -1- would be too strong to suppose that or -ll-, we cannot choose. But both al- Condition 1 and Condition 2 were, in ternatives are regular within the frame- fact, phonetic conditions."" We shall not work of our principles.34 attempt to indicate which of the metrical options specified by our conditions were PHONETIC CHARACTER also phonetic options in Middle English. OF THE CONDITIONS Notice that we keep quite distinct At this the reader will have cases of syncope in Middle English point as noticed that Conditions I and 2, while which, indicated by the spelling, entail bonafide variants. Thus we explicitly non-phonetic, nonetheless bear phonetic sup- the doublet a strong resemblance to certain optional pose phonetic comprende beside in view of the fol- phonetic rules of contemporary spoken comprehende lines and their English. Thus the sequence I wol which lowing spellings: a Condi- constitutes single position by As muche joie as herte may com- tion 2 well have been may pronounced (45) prende (TC.3.1687) as I'll a reader of Chau- by contemporary Than they kan in hir lewednesse cer. the in Similarly, sequence -ery every comprehende (F.Sq.223) which constitutes a single metrical posi- the lines tion by Condition 1 is quite normally Similarly, following require doublets: pronounced as a dissyllable in contem- in porary English. Indeed, every example The of love, a benedicitee in the Concordance with the word god every (46) (A.Kn.1785) it is necessary to invoke Condition 1. What! nat ben- On the other there are occurrences lyveth thy lady, hand, diste? (TC.1.780) of forms, such as Canterbury, in which the relevant sequence is sometimes a Further, the well-known bileeve : bleeve single position and sometimes not. Thus and coroune: croune doublets clearly we have the following lines: indicate that we must suppose phonetic doublets in dealing with lines containing these words.36 Of Engelond to Cauntebury they wende (44) (A.Prol. 16) EXCEPTIONS That toward Caunterbury wolden We mentioned earlier that the funda- ryde (A.Prol.27) mental distinction which a theory of prosody must make is that between Examples such as every and -bury in- dicate that it would be wrong to suppose that Condition 1 and Condition 2 had no 35Licklider (p. 56) introduces a principle of "Resolution after development of liquid or phonetic counterparts in Middle English. nasal" to account for the lines in which these On the other hand, examples such as words occur. We have seen that Condition 1 parisshe, vanisshe, Demophoun and Phyl- will account for these lines and a good many other besides. It is interesting to note, however, that Licklider does not treat the reduction in 34A great many of our exampleshave been these words as phonetic. Thus, he is setting up drawn from commentariesby many scholars on conditions for position occupancy implicitly. Chaucerianprosody. The commentariesthat we He says of these forms, "Whether the atonic have examined include A. H. Licklider's Chap- vowel ever completely disappears or not is ters on the Metricof the ChaucerianTradition hard to determine; it probably remains as a (Baltimore, 1910), Baum's Chaucer's Verse very light touch" (p. 56). (Durham, N. C., 1961), as well as the studies by 36For further discussion of such doublets, Ten Brink and A. J. Ellis cited above. see Licklider, pp. 71 ff. 214 COLLEGE ENGLISH

metrical and unmetrical lines. The theory view they may be perfectly regular. The which we have discussed does this. Thus crucial question, of course, is the number in terms of our theory the preceding of exceptions which a theory must pre- sentence is unmetrical. However, one suppose. The inadequacy of the strict need not construct hypothetical excep- iambic theory is that it must suppose a tions. There are lines in Chaucer which, significant portion of lines in Chaucer to in terms of our theory, must be deemed be unmetrical. Indeed, it is precisely for unmetrical. For example, consider the this reason that the theory is modified in following lines: order to avoid this intolerable supposi- tion. But the modification assumes the Ful weel she soong the service form of a list of "permissible deviations" (47) dyvyne (A.Prol.122) which fails to capture the features that / / / these deviations have in common. For this sal and brim- Arsenyk, armonyak reason we have rejected the modified stoon (G.CY.798) iambic pentameter theory in favor of that here. But now we see that this In the first line of (47) the penultimate presented 1 theory also entails exceptions. The ques- syllable of service is stressed by the tion is, whether the entailed Romance stress rule. But observe that exceptions comprise a significant of the lines with this stress the becomes a portion syllable in Chaucer. The answer is that they do stress maximum which occupies the not. Thus a random selection of one seventh position in violation of Principle thousand lines in Chaucer less 3. The alternative of that to yielded supposing than 1.0% exceptional lines. Such a the form service the final -e per- underlying centage of exceptional lines seems toler- rule and the vowel dropping shortening able in view of the exigencies of manu- rule have applied to yield service(e) script transmission, scribal error and, the of which is stressed the Ro- finally, possibility poetic oversight, servic(e) by in the latter seems to mance stress rule does not This though, principle, help. us a last recourse since it fails to do justice alternative that the line be read requires to the craftsmanship of a great as headless, but in this case, too, the first poet.37 syllable of servic(e) is a stress maximum 37The difficulty that recourse to this prin- can have is occupying the seventh position in viola- ciple illustrated by Yvor Winters' tion of 3. comment on the Keats' line Bright star, would Principle I were steadfast as thou art: ". . . however, in In the second line in (47) we find the this line the stressing of would would result in 1 2 an inverted foot in the second and noun Its stress as- position, compound brimstoon. although inversion is possible in this position, signment, in accordance with English it is difficult and generally unlikely, so that we stress subordination in such units, requires naturally expect the stress to fall on 1, which that we a stress maximum in the likewise is the natural recipient of the rhetori- suppose cal stress . . . but as in this the ninth in the line. this is if, line, compari- position Again son is completed, an actual stress should fall a violation of Principle 3. on the second pronoun; but since this pronoun What these lines make clear is that also is coupled with a verb which is mechan- there is a line in terms of ically its equal and on the basis of its inherent sharp drawn, nature could as well take the and since our accent, theory, between metrical and un- the foot ends the line, and a rhymed line at metrical lines. (Indeed, the very concept that, the accent must fall on art. This blunder Keats of an exception to a theory has meaning by could scarcely have occurred as a result of his reading poetry in a dramatic only in terms of the itself.) From fashion theory ... and had he read the line dramatically he our point of view the lines in (47) are would have noticed the error" (emphasis unmetrical. From some other point of added). "The Audible Reading of Poetry," CHAUCER AND THE STUDY OF PROSODY 215

It would exceed the of this arti- scope The creatour of creature cle and also be to to every premature, attempt (48) (G.SN.49) show how various poets throughout the history of English literature have made Is of the secree of secrees, pardee use of the set of principles outlined above. (G.CY. 1447) It is our that these 1 1 / / contention, however, O, Salamon, wvs and richest of principles, first adhered to by Chaucer, have provided the system of prosody for richesse (E.Mch.2242) in of a major portion of English poets. One Examples which he makes use stress of the things which a study of later poets for purposes of internal rhyme are also would show is the way that these princi- found: have been modified a ples by given poet. Thanne spak Alceste, the wor- Thus one might expect to find that some poet used the same schema but allowed (49) thyeste queene (LGW.G.317) a stress maximum in, say, the ninth posi- This Pandarus tho desirous to serve tion, or else modified Principle 2 to ex- (TC. 1.1058) clude an initial zero position. Indeed, In these the one would not be surprised to find a examples rhyming syllables within the line all bear 1 stress poet experimenting with several such regularly modifications of the system. according to the rules of Middle English accentuation. (Notice, by the way, that CHAUCER'S ART POETICAL the first line in (49) is not hypermetrical We have seen that the language of since Condition 1 and Principle 3 allow Chaucer provides him with several stress us to treat -y est as a single position occu- doublets and that he utilized these for pied by a stress maximum. Alternately the poetic effects (see p. 15 above). Addi- line may exhibit a dissyllabic fifth posi- tional examples of stress doublets in tion without a stress maximum.) 1 "1 1 An of the fashion in which Chaucer are divers beside diverse, crea- example 1 1 1 Chaucer manipulates stress, syntax and tour beside creatour, secree beside secree. phonology is contained in the following Chaucer makes use of these variants with- passage: in the same line, balancing one variant, Men reden folk han within the iambic pattern, against the nat that gretter wit other. Examples in which he manipulates the stressed variants within the line are: Than they that han be most with love vnome; In divers art and in divers(e) (50) And strengest folk ben therwith figures (D.Fri. 1486) overcome, The worthiest and of de- reprintedin the Gross volume (see fn. 7), p. grettest 139. gree: But noticethat the "unlikelyinversion" which This was, and is, and yet men shal it Winters rules out is precisely what the line see (TC. 1.241-5) requiressince it allows for the naturalcom- parisonbetween would and art. Notice, also, 1 that a stresson would is not difficultsince it is In the first line the comparative gretter neutralizedby the syntactic boundary that is followed in the next three lines by separatesit from immediatelypreceding main superlative forms, namely most, strengest, stresson star.Thus the line may be scannedas: worthiest and The first two are 1 1 1 / 1 grettest. I 1 1 Brightstar, would were steadfastas thou art stressed most and with neutralizationbetween the first, second clearly strengest. But, and third positionsand with a stressmaximum, as is often the case in Chaucer when he the only one, occurringin the sixth position. uses the same word or syntactic form in 216 COLLEGE ENGLISH adjacent lines, he makes use of stress op- Now, however, we notice a certain sym- tions to vary and make more interesting metry developing in the stanza. Thus in the repetitions. Thus the fourth line in the same way that the seventh line begins (50) may be read as: with an infinitive followed by a preposi- tional of a The and of phrase consisting preposition worthieste grettest and a so, too, we find the fourth degree gerund, line ending in just that fashion, namely with Chaucer opposing stress variants in singe in compleynyng. Now we may ask the same line as in (48) above. Finally, whether the symmetry goes further. notice the repetition of the consonant Thus the first part of the fourth line features of the superlative ending -este begins with a participial adjective wep- in the final line of (50), thus: ynge followed by the infinitive to laughe. But when we look at the end of This was, and is, and yet men shal it see. 38This proposal is, in fact, merely a restora- tion of the The and the obvious between original manuscript readings. play worthiest line as restored appears, for example, in A in the fourth line and was and is in the One-Text Print of Chaucer'sMinor Poems, ed. fifth line. by Frederick J. Furnivall (Chaucer Society, of his use of First Series 24, 61, London, 1868-1880). In 1888 (Another example making the line is in this form in the stressed variants for of printed Chaucer, purposes variety Minor Poems, ed. by W. W. Skeat (Oxford, and interest is contained in the passage Clarendon Press, 1888). In 1890 M. Piaget in his (4) from the Pardoner's Tale cited article "Oton de Grandson et ses Poesies," earlier.) Romania,XIX (1890, p. 414), prints the original consider the stanza Granson ballads from which the Compleynt of Finally, following Venus was freely translated by Chaucer. He from Chaucer's Complaint of Venus in prints Chaucer'srendition alongside the French which he varies syntax and stress not only following "... l'excellente edition du Rev. W. for interest but also as a subtle reinforce- Skeat .. .". In 1892, however, Paget Toynbee ment of sense. The stanza in in his Specimens of Old French Poetry (IX-XV appears centuries) (Oxford, Clarendon Press, 1892) in- Robinson's edition of Chaucer in the fol- cluded the Granson ballads, and in a footnote lowing fashion: on p. 492 he comments, "Plaindre en dormant, the reading, 'Pleye in sleping,' in the printed Now certis, Love, hit is right co- editions of Chaucer's version of this line, is an venable evident error for 'Pleyne in sleping,' the mark over the in MS. = That men ful dere abye thy nobil y ( yn) having been dis- regarded." thing Toynbee's suggestion was duly noted by As wake abedde, and fasten at the Skeat who, in his Complete Works of Geoffrey table, Chaucer, Romaunt of the Rose and Minor (51) Wepynge to laughe, and singe in Poems (Oxford, Clarendon Press, 1894) alters his earlier reading and comments in a footnote compleynyng; on "The French text shows that we And doun to caste and p. 561, visage must read Pleyne, not Pleye; besides, it makes lokyng, better sense. This correction is due to Mr. Often to chaunge hewe and con- Paget Toynbee; see his Specimens of Old tenaunce French, p. 492." in and dremen at Evidently the Toynbee suggestion has been Pleyne slepyng, followed ever since. Indeed, Robinson in his the daunce, edition of Chaucer (p. 1038) lists the manu- Al the revers of any glad felyng. script authorities, of which there are nine, and We propose first to read the seventh line observes that all manuscripts (save one in which as: the first 44 lines are missing) contain the read- ing pleye in slepyng. Nevertheless, he reads Pleye in slepyng, and dremen at the Pleyne in slepyng on the basis of the French daunce.38 Plaindre. CHAUCER AND THE STUDY OF PROSODY 217

the seventh line we find dremen at the Now symmetry (chiasmus) is restored daunce; the symmetry is thus broken. to the verses. Thus the fourth line begins If, however, we look closely at the as the seventh line ends and the fourth manuscript readings we find that one of line ends as the seventh line begins. In- them, MS.Arch.Seld.B.24, does, indeed, deed, the parts of the lines can be inter- have a participial adjective in the second changed along the diagonals of an X. half of the seventh line. It reads: Chaucer has developed, however, an even these lines. and at the deeper symmetry among Pleve dremyng slepyng He has this means of daunce accomplished by infinitives With this as we further balancing against participial authority may and in Thus the seventh line to read: adjectives gerunds -inge. modify both line four and line seven are not in and at only Pleve slepyng, dremvng reversals of one but are the daunce.39 another, they also internally reversed as well. Line four with a and an The of this us for two reasons. begins participial adjective logic escapes infinitive and ends with a reversal of this First, Chaucer'sversion is by no means a literal translation. Thus the fifth line of Granson's sequence, namely an infinitive and ballad is, "Baisser les yeux quant on doit re- another -ing form, this one a gerund. but it in Chaucer as, "And doun garder" appears Schematically it appears as: to caste visage and lokyng". Similiarly Granson's "Tout a rebours de ce vuelt refrain, qu'on adjective -ing plus infinitive *** trouver" is rendered by Chaucer as "Al the infinitive revers of any glad felyng." Thus the argument plus gerund -ing that the French contains Plaindre seems to us In line seven the mirror is also a one when to a image not very strong applied poem retained with an infinitive followed a so freely rendered by its translator. by Secondly, when one considers that every gerund and ends with a gerund followed manuscript reading contains pleye and not by a prepositional phrase. Its schematic Pleyne, it seems impossible not to follow the structure is: manuscript readings. Indeed, since the nine manuscripts are not derived from the same infinitive *** ad- stemmabut in seen as from plus gerund -ing are, fact, descending jective noun three separatestemmata, the probability of sup- -ing plus posing the same error to have appeared in all Thus the which balances is small. to symmetry manuscripts remarkably It seems us, line four line also balances then, that we have no recourse but to restore against seven, the original manuscript readings pleye in each line against itself. slepyng. Moreover, as will be seen below, the Consider now the fifth and sixth lines original manuscript reading makes much the of (51). These lines are identically con- better sense. contains an adverb and 39The basis for this emendation is structed. Each manuscript infinitival considerably weaker than for Pleye in sleping. a compound phrase consisting If the emendation is accepted it will be seen of an infinitive and two nouns. But notice and that this passage illustrates a syntactic that the placing of two identical lines semantic parallelism not unfamiliar to students between four and seven is the of medieval rhetoric. If this emendation is not precisely same sort of device as that of construct- accepted, then the semantic parallel is the only one to be found in the passage. We shall opt ing lines four and seven as mirror images for the full parallelism on the grounds that so of one another. In other words, line four slight an emendation, albeit with only meager and line five are mirrored in this verse an manuscript support, yields extremely com- line six and line in the same which one cannot by seven, plicated pattern reasonably halves of line four suppose is due to chance. fashion that the first 218 COLLEGE ENGLISH

and line seven are mirrored by their Thus, this stanza offers a complicated second halves.41" example of the way in which Chaucer Next let us turn our attention from manipulates the linguistic givens of his the syntactic to the phonetic devices ex- language for a poetic purpose. The ob- hibited in these verses. We have seen vious characteristic of these syntactic and above that Chaucer opposes stressed var- phonological manipulations is that of iants in the same line for the sake of opposition and, indeed, opposition of a variety. Here, too, we find such opposi- special sort, namely reversal. And when tion. Thus in line four the form wepyng one considers the sense of the stanza, is clearly dissvllabic and must therefore Chaucer's intention becomes apparent. " i In these lines he is saying that the price be stressed wepyng. It is how- opposed, one pays for Love is that the normal ever, by the gerund compleynynge world in which one lives is topsy turvy. which, because of its rhyme with thynge Thus because of Love, one lies awake in bed, fasts when he should be cries and lokynge is to be stressed compleyn- eating, when he wants to laugh, sings when he ynge. Thus the phonetic opposition of wants to cry, plays while he is sleeping and, at the dance, dreams instead of wepying versus play- compleynynge comple- as Chaucer sums at the ments the syntax of this line whereby ing. Indeed, up end of the stanza, he suffers: the first half opposes the second. In other words, in precisely the same way that the / / / / Al the reverse of any glad line is a syntactic mirror image, so too felynge is it a phonological mirror image with In other words, in this stanza Chaucer is respect to stress. It is scanned as: manipulating sound, syntax and sense by means of the devices available on their Wepyng to laugh and singe in respective levels, to reinforce the central theme of the stanza, namely the reversal compleynynge of feeling brought about by love, the But now consider line seven in which familiar Petrarchan paradox. the same phenomenon occurs, only this If, then, we were to edit this stanza in time with the -inge forms at the center terms of the theory of prosody, stress of the opposition rather than at the ex- assignment and the interpretation just tremes. It is scanned: presented, the stanza would appear as follows: Pleye in slepynge and dremyng at Now certis, Love, hit is right co- the daunce venable That men ful dere abye thy nobil Not only is the stress pattern identical with line but it also thinge four, opposes pre- As wake and the same the abedde, fasten at the cisely elements, namely table -inge forms. Wepyng to laughe, and singe in worth in order to compleynynge, 40It is noting that, achieve And doun to caste and this effect, Chaucer has seriously altered the visage French original. Thus Granson's fifth line, lokynge "Baisser les yeux quant on doit regarder" ap- Often to chaunge hewe and con- pears in Chaucer as "And doun to caste visage tenaunce and Indeed, it seems reasonable to lokyng." in and at suppose that the resultant symmetry provides a Pleye slepynge dremyng natural explanation for the changes Chaucer in- the daunce, corporated into his translation. Al the reverse of any glad felynge. CHAUCER AND THE STUDY OF PROSODY 219

APPENDIX selection to choose out of the infinitely The stated above be many sentences of natural language those principle may which for inclusion in the formalized in the following fashion: qualify poem. In the light of the above rules we now Branching rule restate this view. The abstract pattern is . V # PP2P3 .... . Pn # (s (s) ) defined by the branching rule. In partic- where V - verse ular there are three possible metrical pat- P = position terns; one with no extra-metrical sylla- s = syllable ble, one with one such syllable, and one # = verse boundary with two such syllables. ( ) = elements enclosed thus The substitution transformations state are optional; that is, the conditions which any sentence or may or may not be part of a sentence must meet in order to present qualify as a realization of one of the n = the total number of patterns and therefore as a possible can- positions in the line; in didate for the iambic pentameter line. the pentameter line n Different branching rule and/or sub- - 10; in the tetrameter stitution transformations define different line n = 8, etc. meters and metrical styles. Iambic tetram- for n= 8 Substitution eter, example, requires that transformations in the rule. A in the SD: X branching change 1. # P1 second substitution transformation will SC: 12 3 -* 1 0 3 the of the line. transformation) change podic composition (optional Thus, for instance, stress 2. SD: Y#Z by allowing XP, maxima to occupy odd, rather than even, rs we obtain a trochaic rather than SC: 1 2 3 4 5 I Sa + (s) 3 4 5 positions - an iambic line. On the other hand, drop- LSbJ the first substitution transformation where n = 6 8 ping 2, 4, and/or eliminates headless lines. This substitution S = stress maximum transformation may, therefore, be said to Sa = neutralized syllable characterize the metrical style of the Sb = weakly stressed sylla- poet. ble equal to adjacent syllables in stress = extra This work was supported in part by the Joint (s) syllable assigned Services Electronics under Contract to under Conditions Program Pn DA36-039-AMC-03200(E): in part by the Na- 1 and 2. tional Science Foundation (Grant GP-2495), t = one of items so enclosed the National Institutes of Health (Grant MH- must be substituted 04737-05), the National Aeronautics and Space Administration (Grant NsG-496), and the U. S. (obligatory transformation) Air Force (ESD Contract AF 19(628)-2487). The authors gratefully acknowledge the helpful (For significance of this formalism see comments of their colleagues, J. V. Cunning- reference in fn. ham, A. Grossman, Paul Kiparsky, and Ha) 25.) of this We mentioned in the of this Ross who have read earlier versions beginning study. We would also like to give special thanks article that a meter was an abstract pat- to M. Bloomfield, Angus McIntosh and James tern which the poet uses as a basis of Sledd.