Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Material Type: Paper; Class: GRECO-ROMAN ARCHTCT; Subject: Classics; University: University of California - Los Angeles; Term: Winter 1996;
Typology: Papers
1 / 184
This page cannot be seen from the preview
Don't miss anything!
Los Angeles
A Statistical Analysis of the Metrics of the Classic French Decasyllable and
Classic Alexandrine
A dissertation submitted in partial satistfaction of the requirements for the degree of Doctor of Philosophy in Romance Linguistics and Literature
by
Henry Parkman Biggs
3.2.1 Stress Profiles
A Statistical Analysis of the Metrics of the Classic French Decasyllable and Classic Alexandrine
0. Introduction
In traditional analyses of the Classic French decasyllable, the line is assumed to obey only a very small number of constraints. This dissertation is an effort to see if there might be a more elaborate pattern underlying these lines. The premise here is that such a pattern may be discoverable if one examines statistical rather than categorical patterns in the verse. There will be four major parts. In the first section, I will discuss the distinction between metrics and generative metrics, reviewing traditional rules of certain meters as well as discussing the additions in generative metrics of Kiparsky (1975), (1977), Verluyten (1989) and Bowers (1982) to the traditional analyses of the line. I will also discuss the particular difficulties to be encountered when analyzing the placement of stress in French and propose a system of rules to be applied to help in objectively assessing stress in French poetry. I will show that this system proves reliable when compared against the assessments of other metricists and also when compared against the speech patterns of poetry readers. In the second part, I will explain the statistical approach that I have taken in this dissertation, detail the works the data was collected from, explain how I collected the data and explain the use of statistical tests to determine the significance of the data that have been collected.
Chapter One: Background
1. Metrics
Generative metrics (Halle & Keyser (1966), Kiparsky (1977), Piera (1980), Hayes (1983) and other work) addresses the question of how rhythmic patterns are realized as phonological strings in poetry. An objective of generative metrics has been to establish the underlying nature of a given metrical pattern (e.g., the French Alexandrine) and from there to create systems for assessing the metrical tension of lines that deviate from the established pattern. This is a significant departure from traditional approaches which make no distinction between the surface form of a line and its underlying rhythmic pattern. The assumption here is that the poet does not ordinarily write unmetrical lines but lines of varying metrical complexity. So, for example, knowing that a line of iambic pentameter is a series of alternating weak and strong syllables, the first example below is a clear reflection of this pattern whereas the second example is significantly more metrically complex:
a. Of life, of crown, of queen at once dispatched w s w s w s w s w s (Shakespeare, Hamlet, I, v, 75)
b. Never, never, never, never, never w s w s w s w s w s (Shakespeare, King Lear, V, iii, 309)
Aligning the realized form in (1b) with the underlying iambic pattern is achieved through an understanding of the correspondence rules that permit variations in the line under specific environments. In the example in (1b) the relevant correspondence rule is that stress inversion is permitted after a syntactic break of some significance. This way of viewing the line differs sharply from the alternative method of viewing certain lines as canonical and all other variants on the form as in violation of the established pattern.
1.1 Principles of Metrics
When we take an overview of the metrical traditions across the world's languages we see diversity and complexity but also common threads. The meters may be based on syllable count, stress patterns, tone combinations, alliteration, heavy and light syllables and so on, yet their rhythmic deployment is often similar. The use of stress within the metrical structure of one language for example may operate similarly to the use of syllable quantity in another, or long and short vowels in one language may function as stressed and unstressed syllables in another. One of the more prevalent tendencies across languages is for the linguistic material of a language to reflect the underlying pattern with particular faithfulness at some point in the line. An apparent metrical universal is that all metrical traditions have some degree of correspondence between the grouping or bracketing of the metrical
Line hemis. hemis. foot foot foot foot foot s w s w || s w s w s w
Jakobson found that “within the line at least one of the boundaries of each word- unit must occur before an odd syllable. Thus word-units with an even number of syllables must begin in an odd syllable. A disyllabic word-unit must cover either the first and second, or third and fourth, or fifth and sixth, or seventh and eighth, or ninth and tenth syllables,... but never the second and third, fourth and fifth, sixth and seventh, eighth and ninth syllables” (Jakobson, 1952:25). That the 'sense-units' within the line were closely aligned with the line's metrical structure further supports the theory that phrasal break placement in poetry tends to reflect its underlying pattern.
Jakobson also found in Serbo-Croatian verse another common pattern that has emerged across many poetic traditions (Kiparsky (1968), Hayes (1983)): a meter's constraints are observed less stringently at the beginning of the metrical units but increasingly towards the end of the line. Jakobson observed almost without exception a ‘bridge’ in the final two syllables of each hemistich requiring that the two syllables belong to the same ‘word-unit’. In other words, the absence of a foot break before hemistich-final syllables was almost categorically echoed
by the absence of a word break in the actual line. That this condition applied to hemistich-final feet with particular constancy supports the metrical tendency observed across languages of patterns being followed loosely at line beginnings and more strictly towards their ends. Chen (1979) found the same phenomenon in traditional pentasyllabic Chinese poetry. He noted that for the line there was often a major pause after the fourth position of the line, and minor pauses were permissible after the second, fifth or sixth position. Chen generalized these tendencies to two hierarchecal archetypes, differing only in their hierarchical branching in the final two feet. Reflecting this, Chen termed these two patterns right-branching and left- branching and formalized them as follows:
(13) a. b. Line Line hemi hemi hemi hemi foot foot foot foot foot foot foot foot | | 1 2 3 4 5 6 7 1 2 3 4 5 6 7 Right-branching Left-branching
Chen noted that within these canonical schemes the first position of the line had the most freedom while the sixth and seventh positions of both the left and right- branching patterns were 'always rigid' (Chen: 397). In short, he found that in the beginnings of the metrical tradition more play was permissible in the line whereas there was little or no freedom at the end of the line.
While there are many metrical patterns that were experimented with by French poets in the sixteenth century, the octosyllable, decasyllable and dodecasyllable (Alexandrine) were the patterns which found particular favor during the Renaissance.
1.2.1 The Octosyllable
The octosyllabic pattern is made up of eight positions which must be filled with a syllable, with an obligatory stress in the eighth position and the possibility of an extrametrical word-final schwa as a ninth syllable. Examples of this pattern are shown below. Notice that there is no fixed caesura within the line or fixed stress placement other than in the eighth position of the line:
(5) 1 2 3 4 5 6 7 8 Las, je n'eusse jamais pensé 1 2 3 4 5 6 7 8 Veu les ennuiz de ma langueur, 1 2 3 4 5 6 7 8 Que tu m'eusses recompensé 1 2 3 4 5 6 7 8 D'une si cruelle rigueur (Ronsard, “Chanson”, CXLI, 1-4)
While within the line there was relatively free deployment of stress, the constraints of syllable count and line-final stress could not be meddled with. Thus, the constructs below would be unmetrical instantiations of octosyllabic verse (in the following constructs, 'ns' means 'no stress' and is marked only in the position under scrutiny):
(6) 1 2 3 4 5 6 7 8 (ns) a. *D'une rigeur impossible (construct) 1 2 3 4 5 6 7 8 9 10 b. *D'une si cruelle méchanceté (construct)
In the first example, an unstressed syllable is occupying the eighth position of the line so the line would not be considered metrical because of the stress constraint, and the second example is in violation of the constraint on syllable count. The octosyllable was commonly applied to song because its eight syllables worked harmoniously with the four beat sequences often used in music. While this pattern was used by many poets, in the sixteenth century it was not employed as commonly as the decasyllable and Alexandrine.
The first fundamental rule of the Classic French decasyllable is that it is composed of ten positions with the option of a stressless syllable following the
1 2 3 4 (ø) 5 6 (ø) 7 8 9 10 Qu'une galère, ou comme on voit en mer (Belleau, Pierres Précieuses, Gagate, 7)
(10) 1 2 3 4 (ø)5 6 7 8 9 10 Et plus d'étrange et forte passion (Labé, Sonets, XXIV, 13)
In the first example, the schwa at the end of 'une' does not elide because it is not followed by a vowel. Thus, it forms a syllable and occupies a metrical position. In contrast, the final schwa of 'galère' does undergo elision and so fails to occupy a position. Indeed, if 'galère' did not elide the line would be unmetrical not only for violating the syllable count but also the caesura. Similarly, in the second example, 'étrange' undergoes elision of schwa while 'forte' later in the line does not.
1.2.2.2 Stress Requirements
The Classic French decasyllable also had obligatory stresses in the fourth and tenth positions, as in the following examples:
(11) 1 2 3 4 5 6 7 8 9 10 (11) Ny le penser de trop penser en elle Ronsard, Les Amours, CLXIX, 1)
(12)
Si Apollo restreint ses rais dorés Scève, Le Délie, CXXIV, 1)
Stress in the fourth position, however, could on occasion be filled with syllables that simply had the potential for stress but might not be stressed in the delivery of the line. Consider the following examples:
Amour avec sa torche accoustumé (Du Bellay, L’Olive, XXII, 2)
(14) 1 2 3 4 5 6 7 8 9 10 Tresjoyeux d’estre arrivé seurement (Scève, Le Délie, XCIV, 8)
Verluyten (1985) found further that these weak stresses in hemistich final position were metrically sound for some poets only in certain genres. For example, Racine, a later Alexandrine poet, allowed them only in his comedies. The following line from Ronsard and its subsequent ill-formed constructs show exactly what constituted a violation of stress placement in the line (here only relevant stresses are marked; again, (ns) indicates 'no stress' and is also only marked in relevant positions):
(15) 1 2 3 4 5 6 7 8 9 10
A stress-site is obligatory in the fourth and tenth positions; a primary stress is preferred.
A fuller account of the syllables considered in this analysis to be eligible for stress will be addressed in 1.4.
The Classic French decasyllable also had a fixed caesura after the fourth position of the line. The first example below is taken from Ronsard, while the subsequently altered line is in violation of the caesural constraint:
a. Las, brusle moy d'un si chaste flambeau 1 2 3 4 || 5 6 7 8 9 10 (Ronsard, CLXVII, 10) 1 2 3 4 5 6 || 7 8 9 10 b. *Las, que tu brusles fort de ton flambeau 1 2 3 4 || 5 6 7 8 9 10 (construct)
Notice that ($b) is fine in terms of stress; it is solely the violation of the caesural constraint that renders the line unmetrical. Thus a stressed-stressless word could not occupy the fourth and fifth positions of the line because it would violate the caesura nor could it occupy the third and fourth positions of the line because it would violate the stress requirement.
A stressed-stressless word could occur in the fourth position provided the stressless syllable was elided to a vowel-initial word in the fifth position. Some later poets did not view the constraints of stress and caesura as inviolable. The nineteenth century poet De Musset sometimes disregarded these constraints in his verse (examples from Grammont (1937)) :
(18) 1 2 3 4 5 6 7 8 9 10 C'est perdre en dé/sir le temps de bonheur De Musset, “Médiocre”
(19) 1 2 3 4 5 6 7 8 9 10 J'ai dit à mon / coeur, a mon faible coeur De Musset, “Médiocre”
Lines such as these will not be addressed in this analysis because they represent a concerted effort to depart from the original metrical scheme of these lines. Indeed, Verluyten (1985) argues that even for these poets the underlying pattern was still the same. I therefore formalize this final constraint of the Classic French decasyllable as follows:
(20) A word cannot occupy both hemistichs of the line.