Consciousness, Literature and the Arts

Archive

Volume 4 Number 2, July 2003

_______________________________________________________________

The Phonostylistics of the Dramatic Character

By

Sébastien Ruffo

Introduction

Guy Nadon, Gérard Depardieu, Jean-Paul Belmondo and Pierre Lebeau[i], four accomplished actors each playing the role of Cyrano de Bergerac, have produced four very different characters: one of them is bitter and acrimonious; one is a poet, lyrical and sad; the third is a clown playing to the audience; the fourth is a strong and witty musketeer. To some extent, Rostand's play contains all of these possibilities: an essay could analyze how the French Romanticist theory of "mélange des genres", skillfully followed by Rostand, has helped him create many tableaus (the kiss, the battle, the death...), each of them calling for a different facet of the hero's versatile persona. But the four videotaped performances cited here are themselves potent interpretations. No explicit scholarly hermeneutics are presented by these performances but, if we focus on the way the actors deliver their lines, it becomes clear that their vocal craftsmanship highlights and moves and informs the text so as to reveal meanings listeners would not have noticed had they read the play in silence. These vocal, although implicit, interpretations are commonly the object of an intuitive cognitive practice: the public does understand the subtle meanings that make for the Musketeer or the Buffoon Cyrano. This lay practice and its vocal object are the starting point of a new breed of performance analysis: the phonostylistics of the dramatic character[1]. This article addresses some of the challenges we encountered in the course of six years of research aimed at the elaboration of phonostyle[2] as an object of its own (hopefully one day at the center of a new discipline of its own), while constantly borrowing concepts and methods provided by related disciplines such as linguistics and literary studies.

Infra-semantic units.

Intonation, or voice pitch, is a phenomenon that has been widely studied at infra-semantic levels, both in French and in English. In this area, for the French language, one important step was finding out that listeners are able to discriminate between five pitch frequencies[3] and many more dynamic patterns (like "rise-fall" for instance). The discreteness of these perceived structures provides the basis on which a linguistic description of intonation can be attempted with concepts parallel to those used in phonology. Where phonemes are found in phonology, intonemes are the first distinctive units of intonation. It should be stressed that, like the phonemes, the intonemes are not defined as carriers of a referential meaning: their coherence remains solely perceptual, gestalt-like organization. For that reason, one should not speak, for instance, of the "intoneme of continuation" or "of interrogation", but only of the "short static medium" intoneme. In the field of intonation studies, such periphrases serve more or less the same purpose as does the International Phonetic Alphabet in phonology and, likewise, there are lists of the most useful intonemes in the two languages considered here.

Phonostyle as connotation

In linguistic theories, conceptual meaning is located at the next level of analysis, where Martinet places the vast category of the monemes. Whether such a thing as an intonative moneme or lexeme exists or not is a hard question and, as we will see, the answer may well depend upon the semantic theory one places behind the units of this level.

The sentence "Cristina, where did you put the milk!", said with an angry voice implying that the speaker has been looking for the beverage and is now accusing Cristina of having displaced it, has a different meaning from the same words used with love instead of anger: "Cristina..." Between the two sentences, the difference is clear: in the first case, anger is more than implied by the voice, it is expressed by it and, in context, it may constitute the most important part of the whole message. Nonetheless, this anger is still technically implicit information and this precise word, "anger", may not quite fit everybody's understanding of the accents performed on the "p" of "put" or on the vowel of "milk": harshness, aggressiveness, intimidation, recrimination are all plausible alternative explicitations. So, if the intonational or, more largely, the phonostylistic pattern manages to work as a signifier for the signified [anger or harshness or...], it does it through connotation, i.e. using an alternative encoding and focusing on the relation between the speaker and the propositional content of the sentence.

According to many theorists[4], this connotative functioning relegates phonostylistic patterns to the margins of the linguistic structure, in an area of the system that is less rigidly organized than syntax or lexicon. The rules followed by the phonostylistical components of expression are elusive and researchers fail to grasp them firmly enough to propose something that would resemble a true grammar or a thesaurus of vocal devices, and thus a hermeneutic approach of phonostyle must describe the semantic aspect of its object using less linguistic references.

Commenting with narratives and impersonations

To figure out the best way to describe formally the cognitive treatment applied by listeners to the voice of an actor, we interviewed people, presenting them with video excerpts from the four Cyranos and asking them to react freely to the voices. Quite commonly, our respondents avoided any sort of reification of the voice patterns and, instead of using visual references to the shape or intensity of the intonational curves, they would rather underline the specifics they heard by using their own voice (and all of their body) to sustain an approximate paraphrase of the excerpts. On other occasions, respondents made up narratives to explain in what type of situations such patterns could be expected. For example, upon listening to a clip where Belmondo's Cyrano says "Je vous délivre" to DeGuiche (whom he then stops withholding on his way to Roxane's), one lady said that Belmondo sounded "like a schoolmaster addressing very young pupils". In all the cases, they struggled to name precisely the attitude or the emotion of the character, and they were more satisfied with their narratives or their impersonations than with any lexical tag they could think of.

What can be learned from these interviews is, first of all, that the linguistic point of view does not seem natural to the respondents: they do not try to articulate their comments by dividing intonation into smaller units or by making references to more abstract functions. In lieu of this, they spontaneously associate voice patterns with connoted pragmatic situations; and, by paraphrasing Rostand, they also reveal that similar patterns can be adjusted to noticeably different phraseologies while keeping their original meaning.[5]

Phonostylistic hermeneutics and objectivity

Using the tags suggested by the first informants, a second and more extensive experiment was conducted, on one hand, to assess the level of objectivity listeners can reach while commenting dramatic voices, and on the other hand to prepare the grounds for a comparative study of the four performances. Twelve groups of approximately 30 students were showed a total of 144 video clips integrated into PowerPoint™ multiple choice questionnaires. Students first listened to the four Cyrano saying the same lines and then had to choose, for each of them, which of a series of tags best described the phonostylistical patterns.

The software processing of the raw data was conducted at the University of Montreal in the Laboratory of the Cours Autodidactique de Français Écrit. Over the past twenty years, its team has developed a methodology to analyze with surveys the language used by targeted groups and to prepare didactical tools aimed at teaching them exactly and exclusively the linguistic rules that they are prepared to understand. What is particular about this methodology is that it does not need a correction grid (provided by the researcher) to determine "by itself" who in a group forms the elite capable of answering questions on the basis of a coherent reflection –and who answers more randomly, showing they do not really understand what they are doing. At the end of the process, not only does every student receive an individual grade, but every possible answer of the questionnaire (choices a, b, c and d of each question, plus rejection and abstention) also receive a similar grade, or level, indicating which students are more likely to choose it. This layering of the class according to the relative difficulty of the demands makes it possible to isolate competing theories, or sub-rules, that are alive in the group. For example, a question about the third person "s" in English present tense verbs could reveal that in a group of francophone speakers, if only the very best students apply the rule perfectly, relatively good students tend to add an "s" also to the third person plural (because they apply their French plural rule), while the weakest students commit this mistake at random. There are two "theories" in the class, plus one layer of random answers.

The 144 questions, answered by 367 students generated close to 9000 statistics. For every possible answer we know the average grade of those who chose it; the standard deviation of their grades; the precise grade a student must have to start having more than 50% of the chances to choose this answer; and, finally, a factor indicating the power this answer has to discriminate between two distinct layers of respondents. All of this without the researchers knowing in advance which answers should be considered good or bad.

Since the questionnaires were based on tags provided by the first inquiry, they were not designed to test a given theory concerning the four performances. In other words, there were no "kinds" of questions, just individual questions asking people whether this line was performed, for example, "with irony, as a lie, as a truth or as an excuse", "with indifference, surprise, disdain or admiration". In general, we had chosen excerpts showing notable contrasts between at least two actors, but outside of this consideration, they came from all parts of the play and they showed Cyrano in interaction with most of the protagonists.

Globally, this "reader-response" methodology made it possible to affirm, statistics in hand, not only that phonostyle is a litigious phenomenon, upon which people often disagree, but also that some individuals are significantly better than others at interpreting it. Knowing that the researcher himself scored 94% when his own answers were corrected with the grid computed according to the group's elite choices, we could even propose that, in this new field, a well-trained listener could practice phonostylistics from disciplinary grounds similar to those supporting the practice of a more traditional literary criticism, without the absolute need to engage in large scale reader-response verification of his hermeneutic insights.

Role as a connoted signified

On a more theoretical level of analysis, the possibility of a connotative process relating phonostylistical patterns to some form of "pragmatic situations" also required further investigation. The first step in this direction is to define the type of pragmatic circumstances that are chiefly involved by vocal connotations. For example, the schoolteacher used above in a comparison with Cyrano could connote a profession, a certain imperiousness, a way to address young people, or large audiences, etc. Theories focused on classical or modern rhetoric, on literary genre, on speech acts, on dialogue or even on the psychology of emotions could all provide useful taxonomies capable of structuring this concept of "situation", but none seemed to fit our object as closely as the theory of role put forward by Robert J. Landy[6]. To him, roles are what personalities are made of, and the same person can have many: lover, professor, daughter, consumer –the list has no end. Each role is organized in such an organic way that the memory of one grin, one gesture, one single intonation belonging to a role is enough to bring it completely back to mind with its many ways, states of mind and relational styles.

Among the interviews of the first investigation, a few cases similar to the schoolteacher cited above led us to search for such roles in the play. Indeed, as mentioned in the introduction, Cyrano de Bergerac is written to provide the actor with a diversity of scenes putting forward many aspects of the hero's personality. We thus hypothesized that an actor could accentuate one or more of four latent roles, which we designated with the names of archetypal occidental personas: Harlequin, D'Artagnan, Werther and Nero. The question now becomes: does a Cyrano who is "more Harlequin than D'Artagnan" use his voice in a specific manner?

A procedure was developed to verify whether the respondents of our questionnaires had or had not interpreted the 144 clips as connoting some cognitive structure resembling these hypothetical roles. But the multiple-choice questions had not been designed to reveal general and lasting tendencies of the character: they solely assessed the punctual decoding of phonostylistical devices. So, to link the many questions to the four roles, the researcher evaluated every tag that the questionnaire had proposed, asking himself if this tag was compatible with each role. For example, if the choice "A" was "Cyrano sounds arrogant", he had to decide if the word "arrogant" suited Harlequin, D'Artagnan, etc. To translate this into statistics, compatible choices were given 1 point, indifferent choices were given 0 and dissonant connotations were given -1.

Comparing the average score the four actors received on these scales allows us to confirm some clear tendencies: for example, the best respondents generally selected tags that the researcher associates with Nero, except for Depardieu who is the only one to receive a negative score for this role. But Depardieu generates also the answers found the most compatible with both Werther and d'Artagnan. As for Belmondo : Harlquin, bien sûr. Of course these four archetypal personas have themselves nothing to do with Cyrano. Their function is heuristic, as their helping us is limited to evoking personality configurations that our culture would be prompt to accept as probable and coherent. Our phonostylistic methodology could do without them and produce a more literary form of critic — if it were not for the last step of our research, a comparison of the response between two culturally distinct groups of students.

Role perception across cultures

The inquiry was conducted mainly in a private French speaking Quebec "collège" providing the equivalent of high school and undergraduate college education. Respondants were found to divide themselves into two statistically relevant groups : those who declared Quebecois French as their mother tongue, and the others (Anglo-Canadians, Vietnamese, Haïtians, Europeans). As required by this elite institution, all spoke French with native or native-like  fluency, but their cultural background was not primarly French-Canadian. The "others" reacted collectively to our test in a surprisingly  homogeneous fashion, thus permitting the emergence of a second set of coherent role structure perceptions. For example, their best answers were found to reflect quite positively the Harlequin structure of Depardieu's Cyrano, whereas the Quebecois' answers contributed negatively to the same structure : listening to him, they consistently avoided answers pertaining to any dimension of laughter. The Nero structured also revealed an important difference : the Quebecois thought he was the less "Nero" of the four characters, but the others show just the opposite : to them, he is by far the most "Nero" Cyrano., and they also respond to his performance with more answers compatible with d'Artagan. On the other hand, the Quebecois found him funny, an idea the others do not share at all.

* * *

We could summarize our research by saying this: what the vocal style of an actor's performance may connote is a group of idiosyncratic characteristics that listeners recognize as making sense together. As we see, the phonostylistics of the drama character reaches very deeply into the human experience, at the limit between person and character, and the concept of style underlying this reflection reunites content and form at an existential level where we can strongly affirm with Buffon that "Le style c'est l'homme".