ONE
Dramatic Text and Mise-en-Scene
Playgoer: Then is all the stage direction of the world's plays worthless? Stage-Director: Not to the reader, but to the stage-director, and to the actor—yes.
Edward Gordon Craig, Art of the Theater
i.i. REASONS FOR A MISUNDERSTANDING
Before going directly to the point of this study—performance, or the theatrical event—I would like to indulge in an initial digression with the aim of élirnmating or trying to eliminate misunderstandings that are still quite common regarding the ways in which the dramatic text can be treated within the field of theater semiotics and, above all, regarding the relationships that are created between the dramatic text (when there is one) and ťhé mise-en-scene, which is to say, the transcoding of the written text into performance. Clearly, we are dealing with two separate problems, the second of which affects the first. In fact, different conceptions of theater semiotics are closely dependent on the way we understand the relationship between these two entities: the dramatic text and its staging (the performance).1
; There is still a widespread tendency among theorists to place the dramatic text in a position of privilege and absolute superiority vis-a-vis its transcoding into performance. The dramatic text supposedly constitutes the "constant" or "deep structure" of these transcodings. (Indeed this bias is found more frequently in theoretical writing, and particularly in semiotic analysis, than in theatrical practice.)
The most important consequences that come from privileging the literary text over performance are (a) the confusion between real staging and virtual staging (it is claimed, in fact, that to reconstruct or analyze the performances virtually inscribed in any given dramatic text—granted that an operation of this .kind is possible and makes sense—is substantially equivalent to, and therefore an adequate substitute for, the reconstruction and analysis of one or more real performances of the text, since these
16
The Semiotics of Performance
performances—according to the same line of reasoning—are nothing othei than realizations of the text); (b) the tendency in performance to privilege the verbal components (texts) over the nonverbal; (c) the actual restrictior of the class of "theatrical performances" to the subclass of "staged perfor mances of written dramatic texts."
As for the third point, I will wait until chapter 2 to provide a full, com prehensive definition of theatrical performance. The first two points will bt discussed here. To begin, however, I would like to stress that real or con crete performances and virtual, ideal, or potential performances are twc completely distinct entities, not correlated bi-univocally, and that the dis tance that separates virtual stagings from real stagings (the only object per tinent to a semiotics of theater strictly speaking) cannot be bridged as lonj as we remain on the level of the written text (regardless of how well it i read or analyzed) without examining its transcodings into a concrete per formance on the stage.3
Clearly, it is not my intention to cast doubt on the legitimacy of a semi otics of the dramatic text. I am criticizing only the erroneous tendency, stU present in the work of many scholars, of confusing the written text with th< performance, or, more precisely, of assuming that the performance is "in eluded" in the text, when, if anything, the converse is true.
Before examining some of the most explicit and symbolic examples of. privileging of the dramatic text above performance in semiotic approache to the theater (it is quite symptomatic that in these cases critics speak aJ most exclusively of the "mise-en-scene"), I would like to list briefly the rea sons that seem to have led to the adoptionof this position, obliging semioti research to reiterate outdated questions already fully resolved elsewhere Twentieth-century theatrical theory and practice has for a long time in fac sanctioned the autonomy and separateness of the mise-en-scene with re spect to the dramatic text from which it almost always takes its initial insp; ration, although this is increasingly less the case in our own time.3 I ca distinguish three separate reasons for the outdated bias in semiotic re search; two are particular, while the third is general.
(1) The written text, when it exists, is generally the only componer of the performance that is present and persistent. In the next chapter, pe: formance components will be called /partial texts/. The written text is usi ally the only part available to the analyst. The other components disappea as we know, with the end of the performance, because their present is ephemeral and non-persistent. They can be retrieved only partially, in varj ing degrees, according to the quality and quantity of "traces" left behin by the performance: the script, director's notes, photographs, document; tion on film or television, descriptions by members of the audieno reviews, and the like. This state of things has undoubtedly favored tr "promotion" of the dramatic text from the status of a single componei that happens to be present and lasting to the status of a unique, significai component, a prioritized element, totally representative of all other con ponents.
Dramatic Text and Mise-en-Scéne
17
(2) The indiscriminate and dangerously metaphoric use of the "linguistic model" that was initiated in theater semiotics toward the end of the 1960s and the beginning of the 1970s4 has regrettably led analysts to focus attention—and usually exclusive attention—on the literary work, the area where the linguistic model exerted its greatest influence. For obvious reasons, analysts neglected performance itself, sometimes considered unapproachable from an analytic perspective only because it is not reducible to the linguistic model.
(3) On a more general level, I would argue that the privileging of the dramatic text with respect to the mise-en-scene or performance as the focus of theater semiotics was influenced to a large extent by the theoretical perspective that views verbal language as the "primary modeling system" {Lotman 1967), the most powerful semiotic device, endowed with total effability and thus capable of translating all the contents expressible by means of nonverbal semiotic devices (this is the position of Hjelmslev 1943; Benveniste 1969; Prieto 1970, and others). I agree with Eco's assertion that:
It is true that every content expressed by a verbal unit can be translated into another verbal unit; it is true that the greater part of the content expressed by nonverbal units can also be translated into verbal units; but it is likewise true that there are many contents expressed by complex nonverbal units which cannot be translated into one or more verbal units (other than by means of a very weak approximation). . . . The conclusion to be drawn . . . will be that without doubt verbal language is the most powerful semiotic device that man has invented, but that nevertheless other devices exist, covering portions of a general semantic space that verbal language does not. (1975: 233-35, emphasis added)
1.2. A CRITIQUE OF THE CONCEPTION OF THE DRAMATIC TEXT AS A ''CONSTANT" OR "DEEP STRUCTURE" OF PERFORMANCE
In an essay written more than a decade ago which later became famous, the Danish scholar Steen Jansen proposed that the relationship between a given dramatic text and its performances should be conceptualized as the relationship between a constant and its variants. Using what I would describe as a purely metaphoric application of Hjelmslev's terminology, he wrote:
The substance of Andromaqué% expression is thus formed by the entire group of different concrete realizations of the play, a group that is distinguished from other concrete realizations of dramatic expression by the fact that all of its components share the dramatic text Andromaque in common; therefore this text functions as a constant. . . . Andromaque's form of expression would be established begirming from the entire group of elements that are common to all the variants constituted by the different concrete productions, a group that en-
=;..,..
18
The Semiotics of Performance
ables us to claim that all these realizations represent the play itself: Andro-macjue. (Jansen 1968a: 72-73)
Ruffini has observed that Jansen's claim, like similar claims by other scholars, amounts to:
a conception of the performance as secondary and (literally) derivative in relation to the literary text. On the one hand the literary text is presented as a constant element in the meaning of the different stagings, and on the other hand as performance in itself (and of itself) since the staging only constitutes a futile or superfluous transposition of the text into form. (1978b: 5-6, emphasis added)
For the moment I shall simply add two observations: first, Jansen's theory of the dramatic text as a constant does not hold up to historical verification and is easily disproved, like all bad generalizations.5 Furthermore, it reveals itself immediately for what it is—the expression of an outmoded theatrical ideology posing as scientific theory.6
Pagnini, whose work (1970) is explicitly rernimscent of Jansen's, takes a position similar to this. Despite his distinction between the "written complex" and the "operative complex" ("operative complex" meaning the "verbal and nonverbal delivery of the written element," necessitating a long series of mediations), and despite his recognition of the "remarkable integration of nonlinguistic levels" that characterizes theater (which would require the use of "diacritical symbols similar to those in sheet music" for the transcription of a segment of the performance [121]), in the end Pagnini also explicitly supports the absolute, hierarchical priority of the theatrical text over the performance and the notion that the performance totally represents the dramatic text. In making this argument, he invokes Chomsky's concepts of "deep structure" and "superficial structure":
The written text can be considered as a kind of basic deep structure, in relation to the superficial structure, or the sound film. In fact it contains what could be called the schematic "dramaticity." It stands as the example that can be retrieved without losing anything, the dynamic chain that we recognize as ''dominant" in the dramatic structure. (Pagnini 125)
Although he indicates a "filmed performance (with sound) of a particularly esteemed production" as the best "dramaturgical text" for semiotic de-coupage,7 Pagnini thus theoretically justifies and actually carries out the expulsion of the "operative complex" in his subsequent tabulation of the dynamic material in a scene from Hamlet, based on Barthes's narrative model (1966), which nonetheless constitutes "only" the (otherwise praiseworthy) analysis of a classical play as a literary text, and not an essay on the "semiology of classical theater," as the title announces.8
As I already mentioned, the views of Jansen and Pagnini are widely shared by scholars in the area of research to which I am referring. Generally, the only element that varies from one author to the next is almost always the linguistic metaphor which is used to "translate" a substantially identical idea: the conception of the dramatic text as a primary, original,
■. '-I
Dramatic Text and Mise-en-Scäm
19
, all-embracing entity, and of performance as a secondary, derivative, superfluous entity. Brandi makes a comparison with the langue/parole pair:
The subordination of the text to the actor does not alter the fundamental fact that the representation of a drama stands in relation to the written play as pa-irole stands to language. In fact, this relationship remains basic. (1974: 222)
Brandi states this in spite of having recognized the "preeminence of the actor and the plot over the text" (I.e.). Later, in the same work, he introduces another analogy using the famous Ogden-Richards triangle in order to sharpen and redefine his conception:
in order to remove any doubt regarding my previous claims, let us say that ,,. between the fact that stands as the basis for the action on stage, the text of this :: action, and the representation of this text, the same relationship exists as be-, tween referent, signified, and signifier, (225)
The signifier-signified pair is used for the same purpose by Kowzan. At the conclusion of a book published in 1970 which represents the first systematic attempt to articulate a semiotics of the theatrical event,9 the Polish scholar affirms his substantially "literary"' conception of theater and performance, in narrower and more decisive terms than the language in the rest of his book:
To take a still more general conclusion from my semiological applications, I ■ would say that the divergence between literature and performance, meaning the relationship between these two different forms from the standpoint of communication and sensory perception, can be explained on the level of the signifier, while the problem of the thematic derivation, meaning the conceptual convergence of different works, can in turn be located on the level of the signified. While inevitably simplifying things, this observation perhaps has the merit of bringing to light the fact that the autonomous character of the art of theater in relationship to literature and its derivative character with respect -to the literary domain only seem to be contradictory. (1975: 219)
I admit that for the sake of the argument I am simplifying perhaps a more cautious and certainly a more complex position, yet Kowzan seems to consider the written play as homogeneous with all its possible transpositions on the stage, containing as its "content-invariable" all the meanings that these transpositions could express only through different signifying systems. Moving from the literary level to the performance level, he essentially suggests that the meanings do not generally change, but the signifiers do, unless there is a difference in affabulation.10 In Kowzan's opinion, semiotics could succeed in accounting for the dual character of theater— that as, formal autonomy juxtaposed with dependency on content (meaning dependency on the literary source)—by means of the signifier/sigrtified pair.11
20
The Semiotics of Performance
1.3. LANGUAGE AND METALANGUAGE, TEXT AND METATEXT
I could continue along the same line, but would prefer to interrupt this for the moment in order to evaluate what has been discussed up to this point. In 1.1.1 already alluded to some of the reasons which give rise to and explain, even if they do not justify, positions like those I have just examined. In the same section I had already anticipated the principal consequences to which these positions lead in the semiotic analysis of the theatrical event. It might be useful to recall here briefly the two consequences we are dealing with in this chapter, (a) The marginalization or elimination of performance as the object of analysis (the field of inquiry is occupied entirely by the dramatic text, which already "contains" all of the performance: the analyst has only to "extract" it from the folds of the text in which it is inscribed or hidden, as the director does, or should do).12 This is a matter of the (deliberate) confusion already alluded to between virtual rnise-en-scene and real mise-en-scene. (b) The privileging of the verbal text (in effect, of the dramatic text delivered on the stage) over nonverbal texts in the analysis of performance/3 I will deal with (a) shortly, examining some studies where these tendencies emerge with particular clarity and, in a certain sense, with great lucidity. I will attempt a critique of the notion of the interchangeabil-ity of the dramatic text and the performance (or the performance text) which the writings already discussed, as well as those I am about to deal with, must necessarily postulate in order to justify the choices mentioned in (a). The refutation of (b) constitutes the real goal of this volume, insofar as I am attempting to construct an analytic model of the performance text in which no component is privileged a priori over the others.14
The concept asserted in (b) requires some immediate observations before I begin discussing (a), which will be the main focus of this chapter. My observations, however, more generally concern all the positions tending to privilege the written dramatic text and/or the verbal text of the performance. As I already observed in 1.1., these positions appear to be linked to the myth of the semiotic omnipotence of language; and I pointed out that language, as the most powerful semiotic artifact known, does not seem to satisfy the principle of total effability. Now I must add that on the basis of the concepts just examined (as well as those to be discussed in the next section), there not only seems to be an implicit faith in the myth of the total effability of language, but also something more: a confusion between the planes of language and metalanguage, text and metatext. Indeed, while I admit that in order to describe the nonverbal codes and texts of a performance we use verbal language as the (principal) descriptive metalanguage15 because of its greater range on the level of content, this fact does not authorize us—when we consider it on the level of the text, as the object-language—to grant verbal language a semiotic status that is superior to the status of the other nonverbal object-languages in performance.
Dramatic Text and Mise-en-Scene
21
And there is still another element. Privileging the dramatic text over performance, or the verbal text in performance over other partial texts, while claiming that they are interchangeable, leads not only to confusing language and metalanguage, but also to confusing possibility and execution, or/more precisely {as I propose to show later), a priori metatext and a posteriori metatext.
1.4. VIRTUAL MISE-EN-SCENE AND REAL MISE-EN-SCENE
A1976 essay by Paola Gulli Pugliatti demonstrates in an exemplary way the confusion between a priori metatext and a posteriori metatext, which is to say between virtual mise-en-scene and real mise-en-scene. I must, however, proceed in the right order. Gulli Pugliatti grounds the primacy and the complete representativity of the dramatic text (or the "language-text," as she terms it) in relation to the "stage-text" on the concept of the former asthe "metalinguistic transcription" of a "pretextual performance project" (9-10). In this way, the written text is no longer seen as the inevitable and fundamental point of departure for the dramatic creation, but is located instead halfway between the "pretextual performance synthesis" S (for an explanation of S see note 16) from which is constituted, as I already mentioned, the "metalinguistic transcription," and the "reception of the transcription of S in its concrete fulfillment as a staged performance" (10).
"Despite the appeal of the different perspective that Gulli Pugliatti's position offers in relation to the hypotheses examined earlier, her introduction of a new, third term ("pretextual performance synthesis")—which appears to be definable only through intuition—in addition to the existing terms (text and performance) is of questionable utility. While gladly acknowledging that her model has a flexibility and adaptability that are lacking in other, similar models (because of the problematic caution in presentation),16 I must point out, however, that my basic reservations are substantially the same as in the cases already mentioned above. The "language-text," meaning the written dramatic text (insofar as it includes the stage-text, because of its prerogative to constitute a "metalinguistic transcription": "what is being affirmed here is the presence of the stage-text in the lines of the language-text" [15]), is selected as the only object of analysis, and the semiotics of dramatic writing is thus viewed as the only valid substitute for a semiotics of performance in the proper sense, which is more or less explicitly pronounced unfeasible ("since the written text is what we possess, it is to the written text that we must constantly refer" [9])-
On the one hand Gulli Pugliatti's approach can be appreciated for its effort to break free from the limitations of an excessively rigid "dramaturgical" concept of the relationship between text and performance. Nevertheless, since it is essentially based on the elevation of the dramatic text to the status of "performance metatext," it actually creates more problems
22
The Semioitcs of Performance
than it solves with respect to the models I have already criticized for view ing the written text as a constant or deep structure. The fundamental flav in Gulli Pugliatti's system is not simply the issue of confusing language am metalanguage. A more important error is her confusion of the a prior metatext (i.e., the dramatic as the metalinguistic transcription of perfoi mance, in Gulli Pugliatti's terms) with the a -posteriori metatext (i.e., the pei formance metatext itself, obtained through the description or transcriptio] of a given performance of that dramatic text). In other words, this is equiv alent to confusing virtual stagings (which are inscribed or hidden in th text) with real stagings of a particular text.17
On this issue complete clarity is crucial. Because the dramatic text in it entirety consists in part of stage directions and in part of a literary elemen (the "lines," which might even be absent),18 when properly analyzed it c& at best reveal to us (or "describe metalinguistically") the staged perfoi mance(s) that it envisions or prescribes.19 This means the "type" of mise en-scene that the dramatist imagined when writing the text, and which, a least according to tradition, is linked to the stage conventions of his time In this regard, we might indeed speak of virtual performance^), but thi would never "contain," even in the virtual sense, the real, concrete perfoi mances of a text, whether past, present, or future. The analysis of th virtual/potential/ideal staging of a given dramatic text can be carried out o] the written text itself, and in this way it is correct to speak of the latter as "metalinguistic transcription of the preverbal performance synthesis, even if it seems legitimate to doubt the utility of such an analysis, since i brings with it the constant danger of sliding into normative pronounce ments.20 In contrast, the real, concrete performances of a given dramatJ text (which, in my opinion, should constitute the only true focus of theate semiotics, strictly speaking) would escape our attention completely if w thought we could capture them through the written texts, carefully read o analyzed.
For example, what can be gained from an analysis like Gulli Pugliatti' study of Shakespeare's King Lair (1976) in the very best of hypothese I (above and beyond the results attainable through an attentive literary stud of Shakespeare's writing or of the dramatic structures of the work in ques tion) is the concept of theater or the type of mise-en-scene that Shakespear had in mind (envisioned? prescribed?) in composing Lear. But this ca never amount to the metatextual description/transcription of an actual pei formance of Lear at the Globe Theater in London around 1610. Even if w only wanted to delineate a supposedly "ordinary" staging of this play i England during the early years of the seventeenth century (let us say, performance reconstituted according to the theatrical conventions of th Elizabethan stage), the play's text alone would be completely inadequate t the task, and lacking records of other partial texts (records I will refer to i 2.4. as "direct and internal"), we would be obliged to turn to cultural texl of the time, whether theatrical (dramaturgical, architectural, etc.) c
Dramatic Text and Mise-en-Scene
23
nontheatrical (literary, pictorial, etc.), using a type of analysis that has been aptlynamed "contextual."21
1.4.1. The Residue-Text: A Particular Case?
In defining the dramatic text as an "a priori metatext," what is being considered is the very common situation where the mise-en-scene, the performance, comes after the written text. It is a subsequent event, both in the logical and temporal sense, toward which the text looks forward, and about,which the dramatist can only formulate hypotheses or forecasts (of a more or less prescriptive type) at the moment of writing.
Nevertheless, there are also less-frequent cases of dramatic texts which, above and beyond their identity as a priori metatexts (always guaranteed by the genre), also constitute at least in part a posteriori metatexts, since they aim at verbally transcribing (especially, though not exclusively, through the stage directions) a previous performance, an actual staging that has already taken place, thus cHstmgvushing themselves from the usual transcription of a future performance. This can occur unintentionally, and hap-pened frequently in Shakespeare's works, as Viola Papetti reminds us in an interesting study of theatrical space in London during the Restoration era (1979) Papetti begins with an examination of the three versions of Shakespeare's The Tempest (1623: Shakespeare; 1670: Shakespeare, Dave-nant-Dryden; 1674: Shakespeare, Davenant-Dryden; Shadwell-Betterton) and ithen extends the focus of her study to other, appropriately selected "contiguous cultural texts," as Papetti herself terms them. These include illustrations, drawings, and other types of documentation on the theatrical sites of London in the seventeenth century. It is very interesting that Papetti considers the three reworkings of The Tempest as residue-texts "where we see signs of a phenomenon (the performance) that was not part of the writing of the play, but which entrusts itself to the text, as an unfaithful though fertile tool of memory" (174). The three versions were published after the respective productions had been performed on stage. According to Papetti, it is therefore not inaccurate to suppose that the often copious revisions published in these versions were not made in a vague, generic way by. the "express will of the authors of the written text," but were rather "the effect of the collective will of those who put the play on stage, functioning on various levels of the performance." An exemplary case is the text "that the copyist Ralph Crane prepared for The Tempest in 1623, com-pletewith a careful set of stage directions, so rich in detail that it evokes the imageof a specific performance, perhaps one given at the Court" (175).
The case of the "residue-text" (under another term) has also been considered by Guarino, who uses it as a starting point for several mteresting observations. One of his examples is Pierre Corneille's AndromHe (1650), which can be considered as a piece a machines, given the quantity and quality of its stage directions (completely out of character with Corneille's practice elsewhere), supposedly demonstrating that the directi&ns are not "a
%-.
24
The Semiotics ov Performance
hypothesis of representation" but constitute "the account of an actual pre auction and the vision expressed in it" (1979: 174). Guarino correctly ot serves that examples of this type;
overturn the theatricality of the dramatic contents, changing it from prescription to reference, by recognizing in the complexity of the representation a previously existing element that is both the condition and the occasion of the text that alludes to it. {174)
Another important example of the residue-text is provided by the "scenari ("outlines" or "plot drafts") of the Commedia dell'Arte tradition. A Marorh recalls (1976) in his observations on the scripts published b Flaminio Scala in 1611 (II Teatro delle Favole Rappresentative), these not onl constituted the basic material for a performance still to be staged but als offered the transcription of a performance that had already taken place ( precise, synthetic transcription, not devoid of literary aspirations):
But are the "outlines," which are later described as "scenari," a synechdoche for the text or the performance? As the pars pro toto of a literary entity that is destined to become theater, they identify on the one hand a mediation in the text between an idea of performance and the performance itself, while at the same time positioning themselves at the very site of the absence of such mediation. (192)
This phenomenology of the residue-text could be developed furthe] Yet even though these cases occur frequently, they are atypical.M Nevei theless we must be careful not to confuse two very different situations. I my opinion, we can speak of the residue-^ext in the strict sense only in th case of a dramatic text that refers to specific previous transcodings of thf text into performance, where the playwright, or his representative, later a tempts to "freeze" the configuration of that event with the completely ir adequate tools of verbal language. A much more frequent and quite differ ent case occurs when the playwright takes into account at the moment t writing the circumstances presented by surrounding theatrical realities This is the situation for all dramatic texts and authors. I might in fact argu without the risk of paradox that, in this broad sense, theater not only repn sents an afterward with respect to the written text, as a logical-tempor; event occurring after the fact, but also a beforehand, an a priori, a previousl existing entity constituted by the complex set of theatrical conventions < the era, by public taste, by performances already seen, by personal comp< tence in the practicalities of staging: in short, by everything within th writer that creates an "idea of theater, or to be more precise, the hypothes of representation that is incorporated, more or less prescriptively, in the ver texture of the dramatic text. We must remember that the playwright n< only takes into account the stage "language" and acting style of his time but also often creates a play while thinking of the specific voice, gesture! and style of a given actor (if the writer and actor are not indeed the sam person): la Champmesle for Racine's Phedre; Burbage for Shakespeare; Fr<
Dramatic Text and Mise~en-Scene
25
derich Lemaitre for Hugo's Kuy Bias; Eleonora Duse and Ermete Zacconi for D'Annunzio; Musco, Ruggeri, and Abba for Pirandello.
I will conclude by observing that every dramatic text (in different degrees depending on the individual case, the author, or the traditions of a given place or period) is to some extent "a -posteriori," as well as being, by nature, "a priori," with respect to its staging. In a certain way, theater "enfolds" the text, placing itself both "upstream" and "downstream" from it, as we can seen in the model recently proposed by Livio in his very interesting work on dramatic writing: language of the mise-en-scene -+ script -> language of the mise-en-scene (1979: 32).23
Resuming a line of argument temporarily interrupted, I must ask if the case of the residue-text in the strict sense is important enough to oblige us to reconsider the terms of the issue we are debating at the moment, and which can be summed up in the following question: Is it possible to claim that a dramatic text "contains" its own staging, the performance that it will become (and the performance that it was given on a previous occasion), in such a way that the critical analyst can recover it solely on this basis, or can designate it as the focus of a semiotic study of theater? The answer is obviously no. The example of the residue-text does not seem compelling enough to undermine the way I have approached the problem. In fact, both in the case of the dramatic text/a priori metatext that alludes prescrip-tively to a future staging (offering the outline of a possible representation virtually inscribed within it) and in the case of the residue-text in the strict sense, which, in addition to offering a project for performance (instructions for staging: see below, 1.8.), also attempts to transcribe, recount, or refer to a previous performance, it does not seem possible to claim that the dramatic text "contains" its performance (a specific, previous performance text) in a way that might enable us to recover the performance from the written text and to make it the focus of analysis. There are two obstacles to this. The first, of a historical nature, concerns the frequent lack of clear proof of the relationship between a residue-text and a specific, previous instance of its performance (as Papetti herself readily admits). The second, decisive obstacle (which I will treat in greater depth in the next section) is theoretical in nature and concerns the incapacity of the kind of verbal description contained in a dramatic text to record the paralinguistic and nonverbal aspects of a mise-en-scene in anything other than an approximate, generalized, and ambiguous manner.24
It scarcely matters whether the mise-en-scene of a dramatic text (one of its specific enactments)25 is viewed as "downstream," meaning subsequent to the dramatic text, or "upstream," meaning prior to it, or whether the dramatic text prescribes the performance "for the future," or refers to it, and describes it, "in the past." Ultimately, it can function both ways. Regardless of the case, the mise-en-scene of a dramatic text is never attainable simply through the analysis of the written text, since the performance does not reside solely within that text. It is nevertheless indispensable, as is already obvious (a) in the case of contemporary performances, to view and
26
The Semiotics of Performance
re-view the concrete theatrical performances of the text in question (followed, when possible, by graphic transcriptions); (b) in the case of past performances ("absent" performances in the strict sense), to consult contextual documentation, both internal and external to the text. These are issues that I will attempt to explore in greater detail in the next chapter. Here I will simply observe in conclusion that the case of the residue-text (a case that is atypical and difficult to substantiate) is not strong enough to cast doubt upon the distinctions I made in the previous pages between virtual and real mise-en-scene, between the a -priori metatext (the dramatic text) and the a posteriori metatext (the hypothetical graphic transcription of one of the theatrical transcodings of that dramatic text). There is no qualitative difference between a supposedly "normal" dramatic text and a residue-text, but only a difference of degree. In the latter case, the documentary value of the text will presumably be higher. Its value will never be very high, however, as Marotti has pointed out with regard to Flarninio Scala's scenari.26
Having thus cleared up what might at first have seemed a troublesome exception, I will return to the main argument. I am now concerned with providing further theoretical foundation for the hypotheses proposed here, through an analysis of the process of transcoding that leads from the dramatic text to the performance. It will thus become clearer why the dramatic text can never in any instance contain/transcribe the performance. It will also become clear that the dramatic text is not the content of the performance, since the content is dissolved into the performance in a completely irreversible way through definitive changes in codes and means of expression.
1.5. THE IRREVERSIBILITY OF THEATRICAL TRANSCODING
The argument developed up to this point enables us to distinguish between two pairs of terms that bear a metalinguistic relationship to each other. Their apparent vagueness (or, rather, the fact that they are not always appropriately differentiated from each other) is to a great extent responsible for the misunderstandings examined in the preceding sections.
expression	content	
	E	C
Dramatic Text Virtual Staging(s)
expression	content	
	E	C
Performance Metatext Real Staging
Dramatic Text and Mise-enScene
27
To repeat, it is only possible to extricate virtual performances from a given dramatic text (an a priori metatext) in a metatextual way. These are the only performances that the dramatic text properly "deals with." On the other hand, the real performances of the work constitute the object of the performance metatext that is a posteriori (one metatext for each staging, obviously). On the subject of the performance metatext, I must repeat that at least for now we are dealing with a purely theoretical postulation, almost a regulatory hypothesis, which up to this point has received only inadequate practical approximations.27
The diversity and the incommensurability of the two metatexts we are considering, the dramatic text and the performance metatext (which from now on will be called "the graphic transcription of performance," anticipating the discussion in 2.4.), are based on and guaranteed by the diversity and the incommensurability of the respective object-texts. This means a virtual performance, a production plan in the former case, and a real stage production, a performance as a realized project in the latter. I have deliberately used the term incommensurability rather than diversity. We must now explore this category more carefully. It will thus be possible to formulate a very important principle regarding the relationship between the dramatic text and the performance.
Let us suppose that we can examine a dramatic text and the graphic transcription, or description, of one of its performances side by side. We will take for granted that the text in question has been fully and faithfully ■utilized, without cuts, manipulations, or interpolations. For the sake of simplicity, we will eliminate all consideration of the prosodic elements always involved in a staged performance (accent, intonation, timbre, volume, and the like).28 It will be clear, as a result, that the literary part (the sum .total of the "lines") is an element common to both texts. The dramatic text will contain the dialogue in written form, and the graphic transcription wilLtranscribe the same lines as they are uttered orally. The case of the par-alinguistic and nonverbal elements, such as gestures, scenography, and the like, is very different. Apart from their mutual metalinguistic status, there is no relationship (nor could such a relationship ever be demonstrated) between the stage directions contained in the dramatic text (for the sake of simplicity we will confine ourselves to stage directions in the narrowest sense) and the linguistic, nonlinguistic, or mixed transcription of the "corresponding'' paraverbal or nonverbal elements of the performance, meaning essentially the stage directions transcoded into the mise-en-scene. These will very probably not only seem qualitatively and quantitatively quite different from each other, but, above all, they will in any case be incommensurable, even incompatible, with each other.
All this may seem obvious, but it is nevertheless useful to focus on the reasons for this situation. To do this we must examine the entire process behind the dramatic text, its transposition onto the stage, and,the eventual transcoding of one of its performances. We can then observe "that the incommensurability I have just noted between the two extremes of this pro-
If
.;.:y,:n
i, '
&
...... ... .
28
The Semiotics op Performance
cess (precisely, between the stage directions in the dramatic text and the transcription of the play's theatrical "execution") can be considered as resulting from the irreversibility of the path that leads from one to the other. This irreversibility in turn implies another, indeed very basic, irreversibility: the irreversibility of the process of transcoding that leads from the stage directions in the dramatic text to their "execution" within the performance itself. The fact that this process is irreversible (deternuning, as we shall see, the total irreversibility of the path between the dramatic text and the performance) results from the non-notational language in which the stage directions are recorded in the dramatic text. According to Goodman (1968), in order for a language to be considered a "notational system" it must possess the following five requirements: absence of ambiguity; syntactical disjoint-edness and differentiation; and semantic disjointedness and differentiation.29 As Goodman demonstrates, in the dramatic text "the dialogue [alone] is in a virtually notational system, with utterances as its compliants. This part of the text is a score; and performances compliant with it constitute the work" (210-11). Leaving out the hypothetical issue of the achievement or nonachievement of "fidelity," we can now see that it is always possible, at least in theory, to compare the written dialogue with the performed dialogue, to move from the former to the latter, and vice versa, and to make the same comparisons between the dialogue of the written text and the graphic transcription of the performance. On the other hand, however:
The stage directions, descriptions of scenery, etc., are scripts in a language that meets none of the semantic requirements for notationality; and a performance does not uniquely determine such a script or class of coextensive scripts. Given a performance, the dialogue can be uravocally transcribed: different correct ways of writing it down will have exactly the same performances as com-plaints.But this is not true of the rest of the text. A given setting, for example, may comply with extensionally divergent descriptions; and its compliance with some descriptions may be theoretically undecidable.30 (211, emphasis added)
My own position is even more radical than Goodman's. I believe, in fact, mat theoretically it is never possible to go "backward" from the theatrical transcoding (or performance) of a given stage direction to the stage direction itself.
written stage direction [ dramatic text 1
transcoded stage direction [ performance ]
A stage direction transcoded into performance can be compliant with a variety of written stage directions, and, conversely, a single stage direction can have several different transcodings as its compliants.31 Stated in simpler terms, this means on the one hand that it is impossible (except in a very general way) to reconstruct the stage directions of a dramatic text by
m
■ t.,
Dramatic Text and Mise-en-Scene
29
starring from the theatrical transcoding. On the other hand, it is equally impossible (except in a very intuitive, subjective, and unascertainable way) to verify if a given stage direction has been transcribed in a way, that is "adequate/faithfulfcorrect.''32 In fact there is no sense in posing this problem. The stage directions within a dramatic text do not constitute a "score," since they are not expressed in a notational language.,Hence they are capable of neither "marking off the performances that belong to the work from those that do not" nor "being uniquely determined" by one of these performances (Goodman 128-29).33 Naturally, as we will see, the irreversibility of the stage directions means that the entire process of theatrical transcoding is completely irreversible, unlike, for example, the performance of a musical score.
Since it is never possible to move "backward" from the performance or theatrical transcoding of a given stage direction to the actual stage direction itself, it is even less possible to move backward from a hypothetical graphic transcription of the theatrical transcoding of stage directions to the original stage directions (the "script," in the strict sense), and to recreate the latter from the former. These two entities also constitute a fortiori the end terms of a completely irreversible process. As such, they are not only different but are also incommensurable with each other.34
written -*■    transcoded -^   transcription of
stage directions       stage directions transcoded directions
I believe that the principle of (dual) irreversibility articulated here with the help of Goodman's model is important. It provides final, definitive proof of the theoretical (and methodological) erroneousness of the claim that the dramatic text can be viewed as a performance metatext; that is, that the dramatic text "deals" with its own staging, or with the performance that is supposedly inscribed within it. In fact, far from "containing" the performance, the dramatic text does not even provide its content. The modalities of its theatrical transcoding-—especially insofar as the stage directions are concerned—render it impossible to recuperate the dramatic text on the basis of the staged performance.
The reversibility-irreversibuity opposition allows us to completely reconsider the models analyzed and criticized above, as well as the hypotheses that I have juxtaposed with them along the way. The models that theorize the primacy and the complete representativity of the dramatic text over the performance and/or the (partial) verbal text over other partial theatrical texts is based on the postulation that dramatic text and performance (performance text) are reversible in relation to each other. In this sense it is correct to call them dual models, as Ruffini does:
The dual nature of the models proposed above is in fact created in relation to the possibility of a ■two-way relationship between the literary text and the performance. . . . From the literary text one can infer the performance, and the perfor-
>'■: <m:^m-^m-:
•        ;;;3:':MH; m:-iU\'
:•.     ■■. ■■■■
-äi^
^::mwmm,r: o-Sv-v
____
30 The Semiotics of Performance
irance can -provide or restore the literary text [emphasis added]. I should note that while the movement between literary text and performance is not a one-way function (several performances, different from each other in form, all correspond to a single written play), the opposite movement is one-way. One could formalize the preceding comments by saying that a two-way relationship is established between two functives which are, respectively, a single literary text and a class of "equivalent" performances (in the sense mat they imply an identical content). (1978a: 155)
In the lines immediately following the passage quoted above, Ruffini shows how, on the basis of a model of dual reversibility, one can (and in fact, must) arrive at the kind of hypothesis that I examined earlier, where the relationship between text and performance is viewed in terms of the relationship between form and content, or signified and signifier, relegating performance to the level of "a secondary, derivative entity when compared to the literary text."
The hypotheses that I have gradually juxtaposed with these models of dual reversibility can nevertheless be brought together to provide a single irreversible model of the relationship between the text and the performance. This model is capable of preserving the primary, autonomous character of the theatrical mise-en-scene. A graphic conception of both models could be expressed as follows:35
Dual Reversible Model Single irreversible Models
DRAMATIC TEXT DRAMATIC TEXT
performance performance
1.6. TOWARD A DEFINITION OF THE DRAMATIC GENRE
Up to now this chapter has focused on the relationship between text and performance. In the following sections I will shift my attention to an examination of some of the hypotheses regarding the classification of the dramatic text as a particular type (or genre) of literary text. I will confirm, through additional arguments, how this classification follows by necessity from the adoption of the single-irreversible model described in the previous section. Far from "absorbing" or "containing" the performance, becoming totally representative of it and deflecting analytical attention away from the mise-en-scene, the dramatic text should instead be defined as such, i.e., as "dramatic." This is because it tends to annihilate itself within the performance, simply because it becomes (although the word enters is more appropriate) the performance through a process of transcoding which sanctions its transformation into something other than itself, its irreversible anriihilation as a literary text.
Before continuing, I would like to make some parenthetical comments.
■■m- ■ -v.
Dramatic Text and Mise-enScene
31
It is quite obvious that the dramatic text can be considered, at least in theory, from two different viewpoints: first as a text/or performance (script), and second as a literary text, independently usable and analyzable (Rossi-Landi 1972). In the abstract, these are both equally legitimate options. In practice, however, to render them really effective, they must be considered in the light of the historical phenomenology of dramatic forms. This demonstrates that in given periods one or the other of these perspectives has taken precedence in relation to the same texts. For examples of the dramatic text as script, we can look to the great tradition of English theater in the sixteenth and seventeenth centuries. As Pagnini has observed, these plays were "generally intended only for the stage, and for this reason the manuscripts were handed over directly to the theatrical companies"36 (1988: 87-88). Similarly, classical Greek plays functioned at first—-around the fifth century b.c. —as scripts, and were usually staged by the authors themselves. Later, from Hellenistic times onward, they became mainly texts, literary masterpieces almost divorced from theatrical practice. Among examples of the play as literary text, the immediate and most prominent case is that of the great romantic writers: Kleist, de Musset, Büchner—writers without a theater, whose plays are situated too far outside contemporary canons of theater and drama to be suitable (or intended) for staging. In other cases, such as Manzoni's tragedies, we see examples of,real, rather than merely contingent, cases of "unstageability." We could also consider the theatrical poetics of the French symbolists, which are marked by a strong "literary temptation." Begirtning with Alfred de Mus-set's "spectacle dans un fauieuil," the symbolist tradition goes as far as to declare the complete self-sufficiency of the dramatic text and the futility, if not the "danger," of theatrical representation if the aim is complete fruition (see Mallarme's "mental theater"). Finally, many interesting examples can be found in contemporary theater where the stage directions, often preponderant and almost always very extensive, are not intended to fulfill the Pirandellian function of prescribing the conditions for staging the play but to encapsulate them descriptively within it, just like a novel, thus also attempting to establish the basis for an independent literary appreciation. We could find examples in the works of Beckett, Arrabal, Handke, Scabia (with his "theater of poetry"), and so on.37 I would like to make it clear, however, that I am offering these comments without any intention of supporting the widely cherished notion that the dramatic text is inevitably linked to the stage as its destination and vocation, where it supposedly achieves its appropriate fulfillment. I agree with Taviani (1978: 22), who claims that this is above all a characteristic desire of the author, a desire that has become "one of the most sentimental cliches of the ideology of theater."38 I would argue that if and when the process of theatrical transcoding of a given play is set in motion (and, as I have already repeatedly stressed, this is the only standpoint from which the dramatic text is of any interest), the parameters according to which its dramatic/theatrical qualities' can be de-
32
The Semiotics of Performance
fined are the same parameters that I have described above. I will attempt to explain and clarify these in greater detail in the remaining pages of this chapter.
As I pointed out in 1.2., Steen Jansen has been the object of sharp criticism for his conception of the relationship between the dramatic text and performance. In spite of this, Jansen's work constitutes one of most rigorous attempts so far (especially in recent times) to articulate a more satisfying definition of the particular kinds of texts that we call /dramatic/ and to build a specific and appropriate theoretical model through which the concepts of "drama" and "dramaticity" might be formulated.
In his 1968 study L'Esquisse Jansen's approach was characterized by a rigid conception of the dramatic text as a constant or deep structure. Since his 1973 essay on the "dramatic situation/' however, Jansen has attempted to modify his earlier rigidity, elaborating a theoretical picture that is somewhat different and clearly more acceptable, and offering some very interesting and useful ideas toward a definition of the dramatic text along the lines of my own single-irreversible model. Central to this framework is the concept of the dramatic work as an abstract entity endowed with an equally abstract dramatic aspect. The two possible concrete manifestations of this dramatic aspect are the written text and the performance, or, more precisely, the dramatic text and the performance text, as Jansen has termed them, thus distmgurshing them from nondramatic texts and performances (1973: 245; 1978a: 22). Though differing in every other aspect (because "they are not composed of phenomena of the same sensory and perceptive order" [1973: 242]), the dramatic text and the performance have in common the fact that they constitute two "equally adequate" manifestations of the dramatic aspect of a given play. This circumstance also places these two entities in "correspondence" with each other. It creates a genuine relationship of complicity between them, since the dramatic quality of each entity will be decided in relation to and on the basis of the other, and vice versa: "A written text .manifests a dramatic aspect if, and only if, a corresponding performance actually exists (that is, a performance that is an equally adequate manifestation) and vice versa" (243). Jansen provides a model of the stmcturation of the "dramatic aspect," proposing two categories of dramatic elements: (1) the lines (replique) and (2) the stage direction (regie). The second category can eventually be divided into two subcategories: (3) character and (4) setting (decor). Therefore,
A set of phenomena, a text or a performance, can be qualified as dramatic if, and only if, all the phenomena can be described as manifesting the elements that belong to one of the two categories.39 (245)
According to Jansen, this definition provides a reliable criterion for dis-tmguishing dramatic texts/performances and nondramatic texts/performances:
Novels in which certain parts (descriptions, or the writer's thoughts, for ex-
Dramatic Text and hAise-enScene
33
ample) cannot also be expressed in an equally adequate way in a performance, whether through the same linguistic phenomena or through theatrical phenomena, will therefore be qualified as nondramatic. The same [can be said] for operas and ballets containing elements of music or dance that cannot be expressed in an equally adequate way in a text.
The theoretical framework which Jansen developed through repeated adjustments and revisions, and which I have summed up here in an inevitably schematic way, cannot fail to provoke a sense of puzzlement. Jan-sen's insistence on the notion of the constant or invariable was eventually thrown out the door—as I already mentioned—only to be allowed back in through the window, more or less deliberately. The only difference (not an essential one, I think) is that now the invariable is no longer the written text but the "dramatic work," an abstract entity whose very postulation seems problematic and ambiguous, serving at bottom simply to change in appearance a condition that basically remains the same. I am obviously alluding to the primacy that Jansen accords to the dramatic text in every aspect. It becomes the fixed, indispensable hinge around which the other components of theatrical performance move, changeable and inessential (see 1978a: 27). The performance thus continues to be conceived of, not as a specific, autonomous text, but only as a manifestation that is physically different from a preexisting-work, form, or structure (where? how?) that can be expressed "in an equally adequate way" in a written text. Clearly, this is not too different from the relationship of signifier to signified, of expression to content, etc., postulated by the dual models examined above, and hence also present in Jansen's own early work. And even if one wishes to accept the (apparently) joint nature of the relationship established in Jansen's later writing between the dramatic text and performance, one must still acknowledge that the "new" model simply states that the dramatic text and performance are identical in one aspect, the dramatic one. We might also consider the objections provoked by the above-mentioned criterion in which Jansen d^oiminates between dramatic texts and performances and nondramatic texts and performances.40
In spite of everything, and staying within the limits of a dual-reversible model that conceives of the dramatic text as primary, and totally representative in relation to the performance, the hypotheses advanced by Jansen constitute a significant step in the right direction toward a definition of the dramatic text as dramatic, that is, in relation to its theatrical destination. It is not by chance that Jansen comments in one of his later works: "I would add that it is impossible, or at least improbable in my view, to describe (analyze) a given dramatic text if one does not take into account its (eventual) performance on the stage" (1978a: '22-23).
The co-implication established between the dramatic text and performance according to the definition of their "dramatic aspect" is undoubtedly the most interesting hypothesis to emerge from these works by Jansen. This idea is taken up and rigorously radicalized by Ruffini (1978a)
34
The Semiotics of Performance
within his single-irreversible model of theatrical transcoding. He thus arrives at the following definition of /dramatic text/ ("literary dramatic text," in Ruffini's own terms):
A script41 can be qualified as dramatic if one can distinguish within it a metatextual part, called "the stage directions/' and if this part (the text, complement of the script) is transcribed (transduced) through the codes in a different medium of expression than the medium of the textual code. (1978A: 126)
As for co-implication, Ruffini himself observes that "having correlated the definition of the genre of the literary dramatic text to retextualization, the resulting types of performance texts are a deductive consequence of the definition itself"(i27).
This brings some advantages over Jansen's model, apart from the already fundamental difference of operating within a single-irreversible model. First, the relationship of co-implication between the literary dramatic text (which I refer to as the dramatic text) and the performance is formalized in such a way as to leave no margin of ambiguity, as is the case with Jansen. For Rufhni:
This means that going in reverse along the path of retextualization, from the performance text to the literary dramatic text, and departing from an assimilation as performance texts of the types that result, one is bound to arrive at my definition of the literary dramatic text. (135)
The reverse claim is similarly made. Second, Ruffini's definition of the "script" is such as to also include those types of text (novels, librettos, choreography) which Jansen, as I mentioned, tends to exclude from the dramatic genre, on the basis of a conception that is really still rather more traditional than Ruffini's.42
I could mention, finally, that Ruffini's definition identifies a very vast class of "theatrical performances/' much vaster in any case than the group composed solely of "performances of dramatic texts," which is what coincides essentially with Jansen's notion of "dramatic performances." But I will return to this second point in greater detail later on.
1.7. DRAMATIC DISCOURSE
The results that follow from Ruffini's study (1978a) may amount to the very best that can be achieved while still operating (like Ruffini) within a "classical" structuralist approach which considers texts in isolation, as complete units in themselves, omitting to a great extent their pragmatic aspects, that is, the concrete circumstances of their production and reception.43 This is, in fact, what I would like to begin to do in this chapter in relation to the dramatic text. I shall bring it back to the concrete process of communication, and study it in relation to the conditions of its utterance and to other speech acts that these conditions imply and produce. It is a matter, first of all, of asking oneself: "Who utters the dramatic text and how? Who is enunciated in it and how? Who speaks to whom and under what conditions?" In discourse
80
Dramatic Text and Mise-en-Scene
35
analysis (at least of the French school), the statement is usually called ./discourse/ when considered from the standpoint of its conditions of enun-ciation. We could speak in the same way of dramatic discourse in the case of the dramatic text when it is investigated from the point of view of the act that makes it a discourse.44
We owe to Anne Ubersfeld's 1977 study the merit of having systematically confronted the issue of enunciation in the dramatic text, and it is best to begin the present discussion of this subject by referring to her work. Having noted as a premise that "theatrical discourse is the finest demonstration of the nonindividual character of enunciation" (1977: 250), Ubers-feld argues that its constitutive characteristic is dual enunciation. How can this characteristic be explained? With two distinct textual levels that are found in the dramatic text:
We know that within the theatrical text we must deal with two distinct textual levels [couches] (two subsets of a textual set). One has the author as immediate subject of utterance, and it includes the sum total of the stage directions (performance instructions, place names, characters' names), while the other involves the "dialogue" level (including "monologues") and has a character as its mediated subject of enunciation. (250-51)
This means then that from the perspective of discourse, the dramatic textds composed of a referring discourse, originated by the writer (the playwright), and a referred discourse, spoken by the character. The character's discourse "is an encapsulated element within an encapsulating element" (251).
But one immediately wonders if—according to Ubersfeld's proposal—it is possible to consider the dual discursive articulation of the dramatic text as a specific and constitutive property of the genre. I think the answer has to be no. As for the constitutive aspect, I would point to the well-known texts that are made up solely of stage directions, meaning that they are completely devoid of the literary part (like certain plays by Arrabal and Beckett; for example, Actes sans paroles). These are still considered as dramatic texts by unanimous consensus, not only by the authors who created them for the theater. Their existence exempts me from further demonstrations of the non-prejudicial nature of dual enunciation in order to qualify a text as dramatic. This, in fact, characterizes a great many dramatic texts, but not dramatic literature in its totality. As for the fact that dual enunciation is not exclusive to the dramatic text, I think that this also does not require too much explanation, if one thinks of the novel, which can include all possible types of discourse, in the first, second, and third person, direct discourse (such as in the novels of Ivy Compton Burnett), or free indirect discourse. To conclude, I would say that dual dramatic enunciation only makes evident on a macroscopic level (in a way, mstitutionalizing it) a characteristic which—as Benveniste has clarified (1970: 18)—may be proper to all written texts (and even, though less frequently, to any oral text). Nor can I accept the objection that the dramatic text, unlike other texts, usually bears the explicit indication of the mediated subjects (or pseudo-subjects)
r-.;:y.
mmmmm^:
36
The Semiotics of Performance
of enunciation, meaning the characters. Even in this case, in fact, my ea lier comments on the nonconstitutive and nonspecific nature of the du enunciation of dramatic discourse also apply.
We must now ask ourselves what type of dramatic text is being referre to when critics attempt to elevate properties that are merely the (nonsp' cific) characteristics of the majority of the dramatic texts written up to now i the level of constitutive and distinctive properties of the genre ("diagno tics" as they are called by Putnam [1970] and Nida [1975]). The answi must be that in all of these cases what is being referred to is the tradition, "classical" dramatic,text, the kind of dramatic text which, according i Szondi's understanding (1956), has "absolute" dialogicity as its specific di; cursive character, and "conflict" as a kind of invariable schematic supe: structure (van Dijk 1979b). Once this important point has been made, aj proaches like that elaborated by Ubersfeld (and other scholars)45 will prov extremely useful.
With regard to the constitutive properties of traditional drama (not th properties of the entire genre), the attempts to differentiate between nam tive and drama will prove equally pertinent on the basis of the kinds of di; course and the choice of grammatical persons that these two genres uriliz in a characteristic and an exclusive way. Naturally, Benveniste's distinctio between discourse and story, and his assignment of the dramatic text to th category of discourse since it belongs to the "body of writings that reprc duce oral discourses or imitate their tone and purpose" (287), are, in a mor or less obvious way, at the basis of these efforts.46 By this route, Cesar Segre recently proposed two models that synthesize the specific charactei istics of drama and narrative on the discursive level:
In narrative works the following scheme is realized:
1 = writer
- I-narrator or character
- HE =
narrated character
-I/YOU
receiver
The subject of the utterance addresses himself to the receiver, reader, or listener; through the possible mediation of an I-narrator or of an I-character-nar-rator, it is he who relates in the third person the vicissitudes of the characters (HE). The first-person utterances (discourses) of the characters are referred from within third-person diegesis realized by the subject of the enunciation, its sender. In a play, however, the following figure is operative:
I-sendBr
\ I-character
	^     HE= „	
	narrated	
■ YOU-chara«er
YOU-receiver
■
Dramatic Text and Mise-en-Scene 37
In other words, any mediation of the I-narrator or character-narrator is eliminated. The text in its substance is made up of the statements by the various I-characters; these-may embrace, in diegetic form (HE-narrated), the narration of events offstage. In short, I is superimposed on HE, whereas in narrative HE is superimposed on I. (Segre 1979b)47
This is substantially what Saraiva argued in his custinction between "literary message" and "dramatic message" as subclasses of the archetypal "oratorical message" (1974).
In conclusion, I would argue that, though statistically frequent and hence distinctive, the discursive properties of dramatic texts that the studies mentioned above have brought to light cannot be elevated to pertinent traits in the definition of the dramatic genre (meaning on the level of prejudicial properties in assigning an individual case to such a class). This would involve narrowing down the genre as a whole, which, though theoretically legitimate, seems of negligible usefulness, especially since it is by now widely contradicted by historical research and contemporary dramatic practices.
1.8. THE DRAMATIC TEXT AS "INSTRUCTIONS FOR USE"
A great number of scholars have stressed the special, characteristic dependence of the dramatic text on its conditions of utterance. I have already ■Cited Ubersfeld (1977: 248), but I should also mention Serpieri, who, while repeating the frequently made comparison between narrative and drama, ; claims that "theater [meaning the dramatic text] is institutionally connected to the process of utterance. It needs a pragmatic context. Its temporal axis is : always based on the present, and its space is the deixis" (1977:95)- Now, an investigation of the dramatic text as utterance inevitably leads the analyst to put speech-act theory into operation (as happens in Serpieri's study). Speech acts—which can be locutionary, illocutionary, or perlocutionary— are acts that are performed for, in, or while saying something, and also for, in, and while enunciating texts. According to Austin:
■ The act of "saying something'' in this full, normal sense [understood as a "doing something/' which includes a phonetic act, a phatic act, and a rhetic act] I call,
: Le., dub, the performance of a locutionary act, and the study of utterances thus far and in these respects the study of locutions, or the full units of speech.... To perform a locutionary act is in general, we may say, also and eo ipso to perform an illocutionary act, as I propose to call it. To determine what illocutionary act is so performed we must determine in what way we are using the locution: asking or answering a question, giving some information or an assurance or a warning, pronouncing sentence, making an appointment or an appeal. . .
I explained the performance of an act in this new and second sense as the performance of an "illocutionary" act, i.e., performance of an act in saying as opposed to performance of an act of saying something. . . . There is yet a further sense in which to perform a locutionary act, and therein an illocutionary act,
mm
wm
38
The Semiotics of Performance
may also be to perform an act of another kind. Saying something will often, even normally, produce certain consequential effects upon the feelings, thoughts, or actions of the audience, or of the speaker, or of other persons: and it may be done with the design, intention, or purpose of producing them; and we may then say, thinking of this, that the speaker has performed an act in the nomenclature of which reference is made either, only obliquely, or even not at all, to the performance of the locutionary or illocutionary act. We shall call the performance of an act of this kind the performance of a perlocutionary act. (Austin 1962: 94-101)48
For some years now the theory of speech acts has been part of debates within literary theory, allowing us to review old issues relating to "genre" and "literary fiction" in a new light. In particular, critics have often tried to apply the concept of illocutionary acts to the analysis of linguistic and semi-otic phenomena of various kinds, including the dramatic text and theatrical performance.491 will return to the speech-act approach to the performance text in chapter 6. On the subject of the dramatic text, however, I would like to investigate immediately what kind of information the theory of speech acts can provide toward the goal of achieving a definition of genre, leaving aside for now other possible and equally interesting questions.
Essentially, we must proceed toward a pragmatic characterization of the kind of text that is called the dramatic text, first of all by asking what kind of speech act it is in comparison with a poem, for example, or a novel, or a work of visual art. More precisely, we must ask: what is its complex illocutionary force of genre? For the purpose of answering this question, we must of necessity take into account the distinction invoked by some theorists between macro-speech acts and micro-speech acts:
Just like actions in general, speech-act sequences require global planning and interpretation. That is, certain sequences of various speech acts may be intended and understood, and hence function socially, as one speech act. Such a speech act performed by a sequence of speech acts will be called a global speech act or a macro-speech act. (van Dijk 1977a: 238)
Van Dijk provided an example, which he constructed as a short telephone conversation. Manetti and Violi subsequently modified this as follows:
A: Hello, this is Luigi. Is Carlo there?
B: Hello. If s me. How are things?
A: Fine, thanks. Listen, Carlo, do you need the car tomorrow?
B: I don't think so. But I have to find out if Maria needs it. Why?
A: I'm supposed to take the children to the doctor, but my car is at the repair shop, I don't want.to walk there with them because of the cold. So if you don't need your car, I'd like to borrow it for the afternoon.
B: Sure. It's no problem as far as I'm concerned and I don't think it's a problem for Maria either.
A: That's really great. Can I come by around two?
Dramatic Text and Mi&e-en-Scene
39
B: Sure. See you tomorrow then. A: See you tomorrow, and thanks.
The entire conversation can be summed up as a complex act of questioning (Luigi asks Carlo if he can borrow his car). (Manetti and Violi 1979: 126)
The request thus constitutes the macro-act in this text. As for all the other speech acts contained in the conversation (greetings, preparatory questions, assertions that serve to justify the main question, thanks, leave-takings), these have subordinate functions in relation to the central act: "Here we observe what can be defined as a hierarchy of acts, linked directly to the perlocutionary goals that are hoped for" (127).50
It is not difficult to assess the importance and usefulness of the notion of the linguistic micro-act, and not only because it "allows us to grasp the overall function of a discourse from the point of view of its illocutionary aspect," thus assuring a level of pragmatic coherence to the discourse.511 think that it can also be very usefully employed within a pragmatic approach to textual typologies, according to the instructions already furnished by van Dijk:
In fact, one of the bases for distmguishing different types of discourse, such as narratives or advertising, is the possibility of assigning one, simple or complex, macro-speech act to the production of such a discourse.52 (van Dijk
At this point I am not interested in discovering whether or to what degree a given dramatic text contains various hierarchies of micro-acts or macro-acts, such as questions, assertions, expressions of advice, promises, or commands, just like any other text. This seems fairly obvious and unnecessary, at least for the purposes of the present discussion. The question that interests me is quite different: Is it possible to define a linguistic micro-act that could characterize in a specific way the texts of the so-called dramatic genre? That is, what kind of macro-act can be associated with the enunciation of a dramatic text, not of a particular dramatic text but of any text insofar as it is recognizable as dramatic? An article by John R. Searle, frequently cited in Ubersfeld's volume (1977), has provided two valuable and substantially compatible, yet separate, observations, suggesting that we view the dramatic text as "a series of directions," and thus grant it the complex illocutionary force of the command or order.
[T]he text of the play will consist of some pseudo-assertions, but it will for the most part consist of a series of serious directions to the actors as to how they are to pretend to make assertions and to perform other actions. . . . The playwright represents the actual and pretended actions and the speeches of the actors, but the playwrighf s performance in writing the text of the play is rather like writing a recipe for pretense than engaging in a form of pretense itself. ... In that sense the author of the play is not in general pretending to make assertions; he is giving directions as to how to enact a pretense which the actors then follow. (1975c: 328, emphasis added)
40
The Semiotics of Performance
I will return in a later chapter to the kind of approach proposed b1 Searle for theatrical fiction (6.5.)- For the moment, I will confine myself t( his conception of the dramatic text as a "series of directions," as a "recipi for pretense," which is not developed in the course of his essay. Indepen dently of Searle's study (but not of the theory of speech acts), Ubersfeld hat put forward a series of substantially similar proposals which enable the i& sue to be further developed and articulated:53
Theatrical discourse is not declarative or informative, but conative (with a prevalence of what Jakobson calls the conative function): its mood is the imperative. . . . The written text (text, outline, screenplay, score, etc.) has the status of commanding the signs of representation (although there are, by necessity, autonomous signs produced without a direct relationship to the text. (255-58)
While this is evident in the case of the stage directions, it is no less evident in the literary part of the text, according to Ubersfeld:
If we encounter the syntagm "a chair" in the stage directions, it is impossible to transform it into "there is a chair." The only transformation that takes into account the way the stage directions function as a text is "put a chair (on the stage, in the theatrical space)." But even the line "Be seated, Cinna!" has the effect of requiring the presence of a seat in the performance space. (258)
The fact that those who execute the theatrical transcoding of the text can "disobey" these orders only serves to confirm their existence. For Ubersfeld, the status of the dramatic text is exactly the same as that of a libretto, a missal, or an infantry manual: "a speech act that supposes and creates its own conditions of utterance." The only important difference is that the dramatic text simultaneously consists of the object + instructions for use. Ubersfeld thus suggests that the schema traditionally underlying the dramatic text—which is the constative hyperphrase "X (author) says that Y (character) says that (statement)"—should be modified in the following manner: "X (author) orders Y (actor) to say that (statement) and X (author) orders Z (director) to do (statement of the stage directions)."54 Her conclusions are absolutely similar to Searle's:
The basic characteristic of theatrical discourse lies in the fact that it can be understood only as a series of commands given with a theatrical production, a performance, in mind, and directed toward an intermediary receiver who has the task of conveying it to the receiver (the public).
The structure of Uberfeld's argument is somewhat schematic, a trait also due to the fact that she is working within the dual-reversible model of the text/performance relationship that I sharply criticized above. Yet, apart from this defect, her work offers many interesting insights which help to develop and refine the "imperative" nature of dramatic texts, even within a framework compatible with the direction I have taken in the previous sections.
We must in fact ask: what kind of command is the dramatic text, of what kind and degree is its illocutionary force, what are the "necessary and suf-
00^^11
Dramatic Text and Mise-en-Scene
41
fident conditions" that would guarantee its success as a macro-speech act, and what are the perlocutionary aims and effects associated with it? We must above all test this theoretical hypothesis against the concrete, varied, historical phenomenology of dramatic forms.
I shall leave aside Austin's still schematic and intuitive taxonomy {1962) as well as Vendler's subsequent updating of Austin's early work (1970:150-51), which is of some interest since it makes a distinction between exerci-tives, properly speaking, and operatives within Austin's category of exerci-tives. I shall proceed instead to Searle, to whom we owe the most sophisticated attempt so far to devise a classification of speech acts which is systematic and explicit in its criteria. In his critical revision of Austin's taxonomy (1975a: 344-69), Searle distinguishes five classes of illocutionary acts: assertives, commissives, expressives, declaratives, and directives. Directives, which form the category that is relevant to the present discussion, are described in the following terms:
Directives: The illocutionary point of these consists in the fact that they are attempts (of varying degrees, and hence more precisely, they are determinates of the determinable which includes attempting) on the part of the speaker to get the hearer to do something. They may be very modest "attempts," as when I invite you to do it or suggest that you do it, or they may be very fierce attempts as when I insist that you do it. (355)
Having specified that the prepositional content of directives "is always that the hearer H does some future action A," Searle gives a few examples of verbs that denote acts belonging to this class: order, command, request, plead, pray, entreat, advise, and the like. It is obvious that we are dealing with a category that contains acts that are quite different from each other both in their illocutionary force as well as their perlocutionary aims and effects.55
It thus seems necessary to refer to a more detailed classification that would allow us to distinguish more clearly between various types of directives and to situate the dramatic text more precisely in relation to them. The taxonomy proposed by Ross on the basis of the theory of deontic logic and action rather than on the basis of speech-act theory offers many useful leads in the right direction. Ross distinguishes two basic types of speech: indicative speech (e.g., /Peter is shutting the door/) and directive speech (/Peter, shut the door/). While utterances of the indicative type can be analyzed semantically as "describing] a state of affairs, that is, a topic thought of as real" (12), in utterances of the other type the topic is not thought of as real. For these, another type of operator must be introduced. The example given earlier should then be rewritten as follows: /(Shutting of the door by Peter) so it ought to be/. The meaning content of this statement is "a directive properly symbolized by 'd(T),' where'd' stands for the specific directive element 'so it ought to be'" (34). Within directive speech, Ross distinguishes between (a) personal directives, "inthe sense that some person is necessarily presupposed as their issuer, and that their affect on
42
The Semiotics of Performance
the hearer depends on the directive's having been issued by this particular speaker" (48), and (b) impersonal directives "which have no definite source and whose motivating force is not therefore dependent on the power, authority, or wisdom of any individual" (49).
Since there is no possible doubt about the fact that, if the dramatic text is a directive, it belongs to the personal directives, I will go on to examine the distinctions that Ross draws within category (a). He defines three subclasses: (al) directives "in the interest of the speaker," which contain in turn (alj sanctioned commands and sanctioned invitations, (al2) authoritative commands and authoritative invitations, and (al3) requests based on solidarity; (all) personal directives in the interest of the hearer (these include advice, warnings, recommendations, and directions for use); and (alll) disinterested directives (exhortations, admonitions, and so on). Leaving aside disinterested directives (alll), the dramatic text might be classified approximately as a personal directive sent in the interest of the speaker and/or the hearer ([aTJ and/or [all]); that is, in the interest of the writer and/or the theatrical operators/readers.56 Let us consider only the theatrical operators (the director, stage manager, and so forth) as the recipients of the directive in question, and then let us take a closer look at the picture. As a set of "directions for use," the dramatic text, according to Ross's classification, should constitute an example of "a directive in the hearer's interest" (all).57 In fact, there is an object to be used "in a certain way" (in this case, a text to be staged), and the author of the object (the writer) "includes" within the object itself the instructions on how to use it, in the interest of possible users (the theatrical operators). It thus seems to completely fulfill the conditions of personal directives (aTJ): it is in the interest of B (hearer = metteur-en-scene)58 to receive directives that concern actions "whose performance is in B's interest" (= the staging of the given text), and which "have the function [to] inform B what course of action A regards as best serving B's interest." The speaker A, the author of the text, insofar as he is the greatest expert of that text, can be plausibly viewed as more capable than B (the director) of deciding what best serves B's interest (44).
Nevertheless, if we attempt a closer analysis of the circumstances in which this particular directive is issued, while aiming above all at historical corifhmation, it becomes clear that things do not happen exactly like this, or, at least, not always like this, or not simply like this. There are two particularly crucial points: first, the interest of A (author) and B (metteur-en-scene) at the point where the directive is issued; and second, the opinions of B about A, about the competence of the dramatic text as directive. As for the first point, it must be said that there are at least two interests at stake here, and they normally do not coincide completely, except in rare cases, where almost always A and B are the same person: Greek tragedy, Elizabethan theater, and so on. A has an interest in seeing that his dramatic text is staged, and that it is carried out "in a certain way," the way he "had in mind" during the act of writing, and which he tried to "imprint" in the
2354
Dramatic Text and Mise-en-Scene
43
text. Next there is B's interest in staging A's dramatic text (it is his job, his profession), but not necessarily according to the manner and intentions of A. In other words, B is interested in acting according to the manner and intentions that B attributes to A reading his text, listening to his observations, etc., or—-when he is dealing with a noncontemporary author-collecting information about him, his work, and the theater of his time. Here, .the situation varies from period to period, according to culture and theatrical conventions. Historically, however, the cases in which B's inter-estis to stage A's text "faithfully," respecting his intentions and directives, are generally quite rare. Except in the case of naturalistic theater of nineteenth-century Europe and a great deal of the "official" theater of the twentieth century, the attitudes and intentions of B throughout the ages (whether he is a great artist, stage manager, or director in the strict sense) are generally rather different from those intending to "serve" the dramatic text in every way by adapting obsequiously to the imagined desire of the author. Some examples immediately come to mind, including the use of the outline by comic actors of the Commedia dell'Arte tradition, the use of the classics by great nineteenth-century authors, or the way the founders of modern stage direction theorized the mise-en-scene as a work of art completely independent of literature and regulated by its own aesthetic principles. And here, clearly, the first problem above is linked to the second in relation to B's opinions of A, of his legitimacy and competence to emanate directives to B in B's interest regarding the uses of the dramatic text, and its staging. Only in a case where B recognizes more or less explicitly that A has this legitimacy and competence can we speak of the dramatic text as an emanation of type (all), that is, "in the hearer's interest" (in Ross's words). But in this case, I must repeat what has already been said about problem (1), that is,apart from a few exceptions, the history of West^ erri theater (and a fortiori Asian theater) does not offer any examples of this kind. '
To conclude on this point, the dramatic text seems mostly to take on the characteristics of "unsolicited advice," like a directive "not in the speaker's interest" disguised as a directive "in the hearer's interest" (Ross 101). Ross's taxonomy can still be of use. I believe it can be argued that, according to historical period and cultural or geographical area, the dramatic text oscillates between two of the three subclasses distinguished by Ross in his "personal directives in the speaker's interest": precisely between al2 (directives based on authority) and al3 (directives based on solidarity). The subclassal^ is the group containing the dramatic texts of theatrical traditions arid periods founded on the authority of the dramatist, on the conception of the dramatic text as a basic and primary aspect of theater, and the resulting emphasis on "fidelity" to the text as an absolute value. The observation that Ross makes on this type of directives seems very significant to me:
A may be considered an authority by B, in the sense that B has an attitude of
44
The Semiotics of Performance
spontaneous obedience toward directives (of at least some kinds) which A addresses to him. B complies, then, not because of any sanctions, but by virtue of a sovereign, disinterested drive which arises out of respect for A's authority [emphasis added]. B feels himself obliged to obey because of A's right to command. A's power or "authority" is in itself, however, nothing but a projection of B's attitude of submission. (41)
The difference between the authority recognized by B in A and the authority that B attributed to A in (all) lies in the fact that in (all) the speaker's authority to issue directives "in the hearer's interest" was recognized. Here we are dealing with a cmtural-ethical authority, so to speak. In (al2>, on the other hand, the hearer recognizes the speaker's right to issue directives in the speaker's own interest. This involves a more or less judicial authority. In order to see the dramatic text as an example of directives in the speaker's interest based on authority, it is sufficient to remember the position of power held by the dramatist in the traditional, institutional wing of European theater from the second half of the nineteenth century up to our own time, which was opposed first by the historical avant-garde movements and later by the neo-avant-garde movements of the postwar era.59 But this is still more the case for certain theatrical "schools" or "sects" in the twentieth century, characterized by "an attitude of spontaneous obedience toward the directives" of the dramatist, by a desire to conform to his wishes that "arises from respect" for his authority (the two phrases quoted are from Ross). An example of this can be found in the case of Pirandello and his heirs (his followers and certain directors), or in the case of the Berliner Ensemble, which has mummified Brechf s work, claiming to have restored and preserved it according to both the letter and the spirit.60
hi other cases we could speak of the dramatic text as a directive (in speaker's interest) based on solidarity (al3).
We often try, in various ways, to influence people to do what we want without having at our disposal either sanctions or authority. Our hope is that without any pressure the other party will act according to our directives purely from sympathy and benevolence. Since compliance depends solely on the kindness of the hearer, such directives will have the form, not of commands, but of (courteous) requests, suggestions, invitations, supplications, or entreaties. (43)
The situation of the dethroned king—which is essentially the dramatist's situation in contemporary theater—seems so effectively described in Ross's text that it needs no further commentary. Here the argument can be linked to speech-act philosophers. In order for a speech act to be "successful," its specific "necessary and sufficient conditions" must be achieved. Especially "in the case of orders and commands... it is fundamental that the speaker should be in the appropriate situation and status for giving an order" (Manetti and Violi: 115). As we have just seen in the history of theater, this requirement for success is rarely satisfied by the dramatic text. In any case, this is becoming less and less the case in contemporary theater, where now the dramatic text represents at most a directive macro-act (rather explicitly
Dramatic Text and Mise-en-Scene
45
in the speaker's interest) based on solidarity rather than authority, and hence denotahle with verbs such as those offered by Searle—request, propose, plead, and so on. Eco (1978) has appropriately taken up Searle's lead (1975c) and suggests that we conceive of the dramatic text as a "set of op-tioml instructions," that is, nonbinding instructions with perlocutionary aims that may or may not be followed by actors and other theatrical personnel.61 There are also some scholars (such as Minervini [1979: 40]) who suggest that we distinguish between the success of the dramatic text's directive macro-act (with related perlocutionary macro-aim and macro-effect) and the success of the rmcro-directives that make up the dramatic text: the individual stage directions, implied instructions, and the like. While the success of the latter always seems guaranteed, at least to some degree, by the theatrical transcoding of the dramatic text (whose outcome, the performance, constitutes the product of the perlocutionary macro-effect),62 a very different result occurs in the case of the micro-directives contained in the text. These are often destined to failure (as happens with increasing frequency today), in part because the individual generating them is almost never granted the authority necessary for their pragmatic appropriateness.63
As implied in my comments in the preceding paragraphs on the relationship between text and performance, the issue is even more radical. The problem—which was once of a historical nature (i.e., to consider the different institutional and power relationships operating in various historical periods between director and other individuals involved in theater) — becomes highly theoretical, and can be synthesized in the distinction between directive and prescriptive (two terms which can no longer be considered strictly synonymous, given the insights of Ross and Searle). The theatrical text contains directives (orders, advice, suggestions, etc., depending on the situation) about the way, or ways, in which it may be staged. Yet it never prescribes nor can it prescribe a single solution for how it should be performed, as "directions for use" in the strict sense actually do. Rather, it suggests a range of more or less equally appropriate possibilities, from which the receiver of the directive can choose. Going back to an analogy .used by Eco, I would argue that the dramatic text bears a closer resemblance to a box of Lego building blocks, which offers a choice of different projects for construction, than to "a box of prefabricated elements, a kit, which puts the user to work simply to produce one and only one kind of final product, with no room for error" (1979: 56).
It should be clear that the dramatist's intent or status is not the issue here (as it was earlier when we examined the complex illocutionary force of the dramatic text as a genre). In fact, even if the dramatist wanted to communicate a definite and very precise way of handling the play on stage (a great deal of modern and contemporary dramatic writing, after all, devotes a significant amount of space to stage directions), and even if the dramatist's voice had enough institutional power to impose new norms (as happened in the case of Pirandello and Brecht), it could not be done in any
46
The Semiotics of Performance
case. The impossibility can obviously be explained in relation to the irreversibility of the path that leads from the dramatic text to the performance. To claim that the process of theatrical transcoding is irreversible means to argue that it is impossible to go "backward" from the performance to the dramatic text (as I have already stated in 1.5.), that is, to verify if a given stage direction has been "executed" in an adequate, faithful, correct way or not, cUstinguishing the actions that belong to the stage directions from those that do not. This as we have seen, is because the part of the play devoted to stage directions is expressed in a language that is not notational.
In conclusion, I would like to affirm that the dramatic text contains and executes illocutionary directives (in varying degrees of force), but it cannot be ascertained whether or not their execution is successfully carried out (with the single, already mentioned exception of the macro-illocution consisting in a simple request for staging). Taking a closer look, one could say that the specificity of the dramatic text as a macro-directive, constituted by a set of directive micro-acts, lies in the unverifiability of the success of these micro-acts (and of the achievement of the perlocutionary aims of their issuance) rather than in the fact that they are optional. The latter characteristic, on the other hand, is common to a great number of other types of directive texts.
This important clarification allows us to accept and utilize the concept of the dramatic text as a "quasi-order" or as a "set of optional instructions" without falling back into the domain of the dual-reversible models criticized earlier. In fact, one could say that (a) the dramatic text "orders" (or requests, or proposes) its staging (in the double sense that it requests to be staged and gives instructions on how to do it). It is another thing, however, (b) to mistake (confuse) the performance(s) "ordered" by the text with the performances to which the text can really lead, will lead, or has already led. Having taken the necessary precautions, (a) in no way implies (b) and is therefore not incompatible with my single-irreversible model.
^00055