In discussing those passages of an opera in which not only the audience but also the characters of the story hear music, musicologists often write of “diegetic” music, adopting well-established terminology from film studies and narratology. The terms “diegesis” and “diegetic” stem from ancient Greek, and owe their long-standing fortune to their presence in seminal writings by Plato and Aristotle; scholars of narrative, drama, film, and music occasionally note, however, that the meaning assumed by “diegesis” and “diegetic” in recent decades is very different from (or even opposite to) the historical one. In fact, the two meanings coexist in current scholarly usage, engendering terminological (and therefore conceptual) confusion. The goal of this article is to explain how this situation came about. Ancient Greece witnessed the emergence of the basic narratological distinction between recounting and enacting, and its gradual (and far from straightforward) association with the terminological distinction between “diegesis” and “mimesis.” A misinterpretation of Aristotle by French filmologues around 1950 gave rise to the modern meaning of “diegesis” (“storyworld,” or even simply “story”), while the misapprehension by which the ancient and modern terminologies are deemed to have arisen entirely independently of each other originates in the influential work of narratologist Gérard Genette. The story reconstructed here involves to-ing and fro-ing between words and concepts: what may appear as a purely lexical study is also, of necessity, a study about ideas. Moreover, this story might serve more generally as an apologue, reminding us of the sometimes peculiar ways in which bits of our collective wisdom emerge and become accepted.
“Chanter pour chanter”
In act 2, scene 2, of Le nozze di Figaro, Susanna urges Cherubino to sing for the Countess a song he has composed himself. Cherubino's hesitant response is shrugged off by Susanna (“manco parole”—“enough with talk”), who proceeds to accompany his performance of “Voi che sapete” on the Countess's guitar. Following the performance, the Countess comments, “Bravo! che bella voce!” (Bravo! what a beautiful voice!). Needless to say, the actors on the stage sing throughout this scene—as they do virtually all the time in Italian opera (and in several other varieties of opera). But anyone familiar with the conventions of opera will be quite clear that (a) no character in the story is singing from the beginning of the scene up to “manco parole” (they are, rather, speaking to each other); (b) Cherubino does sing during his song; and (c) from “Bravo!” onward, all characters resume speaking. Put differently (and perhaps better), it is clear that the characters here do not hear any music until Susanna begins to play the song's “ritornello” on the guitar, and again will not hear any music from “Bravo!” through the remainder of scene 2 (which includes Susanna's aria “Venite, inginocchiatevi,” whose words represent what the characters hear as speech).
The difference between the three moments in the scene exemplifies the distinction that the eighteenth-century composer Grétry referred to as “singing in order to sing” vs. “singing in order to speak.”1 Grétry's “chanter pour chanter” is of course what any singer impersonating Cherubino does during “Voi che sapete.” “Chanter pour parler” is what happens in the rest of the scene—and much of the time in opera more generally: no singing at all is heard by the characters. Grétry's distinction is helpful, and requires only a couple of further specifications. First, as any operagoer knows, the singing that does not mean “singing” is conventionally used to represent not only the characters' “parler” but also their thoughts and feelings (as are the words of spoken drama). Secondly, the distinction we have drawn between varieties of onstage singing can of course be extended to embrace the instrumental music we hear in an opera. Here, too, whereas the audience hears music all the time, the characters do so only in a few, special passages (such as, to remain within Le nozze di Figaro, the march and the fandango in the third-act finale). But the general distinction between the moments in which an opera's characters hear some music and those in which they do not is fundamental, more or less consciously present in the mind of any operagoer, and to a good extent necessary for the understanding of what happens in the situations represented on stage.
For those special operatic passages in which not only we, the audience, but also the characters of the story represented hear some music, opera scholars often use the phrase “diegetic music.” In doing so, they adopt a usage of the adjective “diegetic”—and of the noun it refers to, “diegesis”—that is well established in film studies and in narratology (the branch of literary theory that studies the workings of narrative).2 The terms “diegesis” and “diegetic” are generally understood to have their origins in ancient Greek, and indeed owe their long-standing fortune largely to their presence in seminal writings by Plato and Aristotle. At the same time, scholars of narrative, drama, and film also occasionally note that, in their own specialized use, “diegesis” and “diegetic” have a meaning that is very different from—or even opposite to—the meanings that these terms have had through much of history.3 In fact, the two diverging sets of meanings coexist in current scholarly usage, engendering a form of terminological (and therefore conceptual) confusion. Though a number of scholars have discussed how and why this situation might have come about, I have found their various versions of the story neither as thorough nor as accurate as seems necessary. Understanding and telling that story is what I attempt in the pages that follow. In the process, I hope it will become clear that what may at first appear to be an exercise in philology—a purely lexical study—is also, of necessity, a study about ideas.
Plato, Aristotle, and Us
One of the two main uses of the term “diegesis” in current scholarly English, and of its cognates in other modern languages, relates to a narrative mode: it is generally part of the terminological opposition between “diegesis” and “mimesis,” intended as two basic modalities of discursive enunciation or narrative presentation. This is the much older use of the term, as it can be traced back to ancient Greek. (The other current use of “diegesis”—that by which scholars would refer to Cherubino's song as “diegetic music”—is relatively recent, and we will return to it later in our story.)
As has been described since classical antiquity, and as is still understood today, at the most fundamental level there are two ways of presenting a story. The first is that of recounting (relating, reporting). This recounting is done often (though not always) in the past tense, often (though not always) in the third person, virtually always more or less explicitly after the event that is recounted,4 and always in indirect speech (in that this mode's “pure” form recasts anything said by the persons in the story as indirect speech). The second way of presenting a story is that by which the characters speak in propria persona and in direct speech—as happens most patently in the case of drama, where the story seems to unfold in front of us as though in real time. These two fundamental narrative modes, described at least as early as Plato's Republic (370s BC), eventually came to be associated, respectively, with the Greek terms “diēgēsis” (“διήγησις”) and “mimēsis” (“μίμησις”)—a terminological opposition that began to stabilize in the generation immediately following Aristotle (around 300 BC) and that still survives in much current scholarly practice.5
Two remarks are worth making at this stage. First, the post-Aristotelian use of the words “diēgēsis” and “mimēsis” in this way (in association with the opposition between the two basic modes of narrative presentation) differs somewhat, as we will see, from the practices of both Plato and Aristotle, but this does not necessarily imply a comparable discontinuity in the understanding of the respective narratological concepts between the two philosophers on the one hand and generations of later writers on the other. Secondly, and similarly, that Aristotle's use of these terms differs somewhat from Plato's (as we will also see) does not imply a comparable difference in the two philosophers' understanding of the narrative modes in question.
In the third book of Republic, having discussed the “what,” the contents, of literature, Plato moves on to the “how,” how things are said, his lexis including what we would call “modalities of enunciation” or “discursive modes.” Here he describes the basic types of narrative presentation in a way that he claims originated in Socrates. For the first mode, what I earlier called “recounting,” Plato uses the phrase “haplē diēgēsis”—“simple (or plain) narration.”6 In this case, what we hear is the voice of the storyteller: “the poet himself speaks,” it is narration “through the account [apangelia] of the poet himself.”7 This recounting mode is distinguished from the other basic mode, which Plato calls “diēgēsis dia mimēseōs”— “narration by means of enactment (or impersonation)”:8 in this second mode, the poet conceals himself,9 and we hear only the voices of the characters, who express themselves in direct speech. This happens in drama, which is effected “entirely by means of enactment [mimēsis].”10 Finally, a third, mixed mode of presentation combines the two basic modes, alternating between them. This is Plato's “diēgēsis di' amphoterōn”— “narration by means of both [modes]”:11 in epos (as in the modern novel) the voice of the narrator who does the recounting yields every so often to the voices of the characters, whose utterances are no longer recounted in indirect speech but presented as rhēsis—as direct speech (in quotation marks, we would say). It is only in these passages that the epic poets, taking on the voices of the characters, “effect their narration [diēgēsis] by means of enactment [mimēsis].”12 Conversely, the first mode in its pure form does not entail any impersonation, which is why Plato occasionally refers to it also as “narration [diēgēsis] without enactment [mimēsis].”13 It will be noticed that in these passages Plato uses “diēgēsis” as an overarching term, one that refers to the narrative act (the presentation of a story) in general, so that further designations are needed in order to indicate the narration's specific mode. Indeed, he states expressly that everything in epic poetry is diēgēsis—both the characters' direct speech and what comes between instances of it (that is, the poet's recounting in his own voice).14
An analogous description is found in Aristotle's Poetics, though in a passage whose complexity is compounded by serious problems of textual transmission.15 In fact, the passage has been described as “one of the most difficult” in the entire Poetics.16 There are, however, only two basic ways in which scholars have interpreted (and therefore translated) this crucial passage; both interpretations—in spite of their different emphases—show that, in substance, Aristotle's understanding of the matter is not that different from Plato's. The first family of interpretations, exemplified in the following translation, produces three narrative modes very much along the lines of Plato's: “in the same media one can represent [mimeisthai] the same objects by combining narrative [apangellein] with direct personation, as Homer does; or in an invariable narrative voice; or by direct enactment of all roles”—that is, by using respectively the mixed mode, or the recounting mode, or the enacting mode.17 According to the second family of interpretations, Aristotle distinguishes the two basic ways of presenting a story, recounting and enacting, but specifies that the first of these can be either “pure” or interrupted in places by the direct speech of the characters (as in Homer's epos): “It is possible to imitate [mimeisthai] the same objects in the same medium sometimes by narrating [apangellein] (either using a different persona, as in Homer's poetry, or as the same person without variation), or else with all the imitators as agents and engaged in activity.”18 In other words, Plato's three narrative modes (the two basic ones plus their mixture) essentially recur in Aristotle, but depending on the interpretation of this passage they are presented either as three modes, or as two modes of which one is subdivided—the mode of epos (and the novel) being presented as a subspecies of the recounting mode.19
Many commentators convey more or less explicitly the idea that Aristotle turned Plato's system upside down, by making mimēsis, and no longer diēgēsis, the overarching governing principle of narratology.20 But that idea is predicated upon the assumption that the two terms have a stable meaning; once we understand a number of lexical slippages, the two philosophers' approaches to the narrative modes appear to have much more in common.21 Despite the somewhat different perspectives and goals, both approaches seem compatible with the substance of the following general statement (though I formulate it in modern terms):
Virtually any artistic activity (literature, drama, music, dance, the visual arts) is a form of representation—which implies some kind of correspondence, of analogy, between the representing and the represented.22 In particular, stories can be presented in the medium of words, as they are by both writers and playwrights; this presentation can take place in either of two basic modes (the recounting mode, the enacting mode), or in a mixture of the two (the compound mode).
A first problem is that a single term can be used in different ways. Indeed, even in English the term “imitation” can be used both to refer to representation in the most general sense (as in the statement “for Plato and Aristotle virtually any artistic activity is a form of imitation”) and to denote one specific representational mode, the “enactive” one (“for Plato, drama is the presentation of a story by means of imitation”). Similarly, one can use the English terms “narration” and “narrative” in an extended sense, to designate the presentation of a story in general (whatever the mode), but also more specifically, to designate the presentation of a story in the recounting mode (the more common use in everyday English): for the scholars who understand “narrative” in the first, extended sense (which makes it analogous to Plato's “diēgēsis”), narratology includes the study of drama, whereas those scholars who restrict the meaning of “narrative” to “discourse involving the voice of a narrator” will exclude drama from the ambit of narratology.23
This phenomenon—the possible use of one term to express different concepts—is the converse of that by which a similar concept may be expressed by different words, quasi-synonyms that often introduce a further element of complexity (and confusion). In English, I could refer to the very same mode of presentation as either “recounting,” “relating,” or “reporting.” Finally, and more generally, there is the fact that in everyday language we are not necessarily consistent in the use of our vocabulary—at least not as rigorously consistent as scholars are (or try to be).
Something similar happens in our Greek sources. From the late sixth century BC onward, the noun “mimēsis” is found to be used either in the wider sense of “representation” in general, or in the narrower sense of a specific type of representation, what I earlier referred to as “enactment”:24 we encountered the latter, narrower use of “mimēsis” in the important passage from book 3 of Plato's Republic, while the broader use of “mimēsis” is often found in Aristotle (as in his use of the related verb “mimeisthai” in the passage from Poetics just discussed). (Conversely, Aristotle can be found to use “mimeisthai” for the specific enacting mode,25 while Plato often uses “mimēsis” for “representation” in the more general sense.)26 Analogously, the noun “diēgēsis” and the verb “diēgeisthai” can refer either to narrating in general or to the specific narrative mode I have called “recounting”: the first, general use is exemplified by the same passage from book 3 of Plato's Republic, whereas the occurrences of “diēgēsis” and the adjective “diēgēmatikos” in Aristotle's Poetics can all be made to fall within the specific area of the recounting mode.27
But in the intricate passage from Poetics discussed above, in order to refer to recounting Aristotle uses not the verb “diēgeisthai” but the verb “apangellein.”28 We have thus moved on to the second problematic area alluded to above—that of the use of different terms to express a similar concept. Likewise, elsewhere in Poetics Aristotle uses the noun “apangelia” as a virtual synonym for “diēgēsis” in the specific sense of “recounting.”29 The same is true of Plato, who refers to recounting not just with “haplē diēgēsis” but also with “apangellein” and “apangelia,”30 and to enacting not only with “mimēsis” but also with “dialogos.”31 And again, although the use of “diēgēsis” for recounting and “mimēsis” for enacting seems fairly stable in writers of the generations following Aristotle, those writers, too, knew alternative terms: the adjective referring to the enacting mode, for instance, could be not only “mimētikos” but also “dramatikos,” while for the recounting mode one finds not only “diēgēmatikos” but also other adjectives, such as “apangeltikos” or “amimētos” (amimetic).32
There is, finally, what I earlier referred to as the general problem of inconsistency. Occasionally, Plato deviates from his own practice, using “apangelia” and “apangellein” not for the specific recounting mode but for narrating in general,33 and “diēgēsis” not for narrating in general but for the recounting mode in particular.34
Thus, Plato's basic description of the narrative modes “pervades ancient literary criticism,” but with “a fairly high degree of terminological and conceptual variety.”35 It is inaccurate to claim, as some scholars have done, that Plato termed the basic modal opposition “diēgēsis” vs. “mimēsis”;36 it is, however, equally inaccurate to claim that such a terminological opposition is a modern invention,37 as that opposition is frequently found in post-Aristotelian theorists (while the conceptual opposition is already spelled out clearly in both Plato and Aristotle). As for the abovementioned contention that Aristotle turned Plato's hierarchy of diēgēsis and mimēsis upside down, one could be led to it by noting the ways in which the two terms are combined in specific passages. As we have seen, Plato calls drama “diēgēsis dia mimēseōs” (diegesis by means of mimesis); Aristotle, on the other hand, refers to epic as “diēgēmatikē mimēsis” (diegetic mimesis, mimesis by means of diegesis).38 If one sticks to the words rather than the concepts, this does look like a reversal in the hierarchy (Plato suggesting that mimesis is only one of the ways of effecting diegesis, Aristotle that diegesis is only one of the ways of effecting mimesis). But the alert reader will have worked out by now that Plato's diegesis by mimesis means “presentation of a story effected through enactment,” whereas Aristotle's mimesis by diegesis means “(narrative) representation effected through recounting.” The two concepts are far from incompatible; in fact, they both fit within the theoretical model expressed in the general statement I concocted above in order to claim a degree of compatibility between Plato's and Aristotle's proto-narratological approaches.
It is crucial to note, however, that whatever the fluctuations in the use of “diēgēsis” and its derivatives, historically the term refers either to the act of narrating in general (whatever the mode), or to one of its discursive modes (one particular way of presenting a story), or even to a narrative utterance (the written or oral discourse that is the product of an act of recounting)—but not to anything relating to a story's narrative content. Yet the second, more recent meaning of “diegesis” in current English—that by which scholars write of “diegetic music”—has precisely to do with the specific story being presented and its world. That second usage emerged only in the latter half of the twentieth century.
Filmologues and Narratologists
The term “diegesis”—or better, the French “diégèse”—began to be used in film studies around 1950, to indicate the world in which the story represented in a film takes place. Many accounts attribute this use of the term to the scholar Étienne Souriau. Souriau, however, stated more than once that the term had emerged through the collective work of a group of French scholars of cinema—or filmologues, as they liked to call themselves.39 The group had been meeting since the late 1940s around Souriau and Gilbert Cohen-Séat, and by 1950 the Sorbonne awarded it official acknowledgment by formally establishing its Institut de Filmologie. Anne Souriau—daughter of Étienne and a member of the group—has in more recent years claimed for herself the coinage of the term “diégèse,” as we will see.
“Diégèse,” “diégétique,” and their cognates in other languages have become common currency in studies not only of film and film music, but also of literature, drama, and opera, where they are used to refer to the story told by a work and/or to the world in which that story takes place. A substantial contribution to the diffusion of the recent, “filmological” meaning of the term—and, it must be said, to the ensuing terminological and conceptual confusion—was made by Gérard Genette, one of the leading narratologists of the past fifty years. It is worth tracing the development of Genette's use of the Greek “diēgēsis” and its derivatives before returning to Souriau's “diégèse.”40
In his essay “Frontières du récit” of 1966, reprinted in his book Figures II (1969), Genette refers to Aristotle's Poetics and introduces the Greek word used there for the recounting mode, transliterating it as “diègèsis” (printed as “diégésis” in the later, book version).41 In spite of some elements of confusion in his treatment of the subject (seriously compounded in the English translations), Genette rightly concludes that the difference between Plato's and Aristotle's classifications of the narrative modes boils down to a question of lexical variation.42 Interestingly, in one passage of this essay Genette uses the word “diégèse”: this is clearly meant as a straight morphological adaptation into French of the Greek “diēgēsis” (as “synthèse” is of the Greek “synthesis”), and indeed Genette uses “diégèse” in a context in which the French term has the same meaning as “diēgēsis” in Aristotle's Poetics and in the work of many post-Aristotelian theorists (indicating the recounting mode, or an utterance in that mode).43
At least as early as an essay first published in 1969, however, Genette switched to the more recent, filmological use of the word, as he went on to define “diégèse” as “the spatiotemporal universe to which the primary narrative refers.”44 By 1972, the time of his “Discours du récit,” Genette acknowledged in a footnote that this recent use of “diégèse” “comes to us from the theorists of film narrative,” but he provided no reference for that claim.45 In this essay he fluctuated, however, in his treatment of the modern “diégèse,” using it at times as an equivalent of “story” (“histoire,” the content of a narrative), and at times to denote the world in which the story takes place.46 Genette, as we will see, would later recognize and discuss this problem, but such ambivalence has characterized the term throughout the history of its use in the modern, filmological-narratological sense. Moreover, in the same essay Genette repeatedly referred to the word found in ancient Greek discussions of narrative modes, for which he used “diégésis” (now with acute accents) as opposed to “diégèse”—something that already makes clear the semantic bifurcation in the use of the family of words originating in the Greek “diēgēsis.” Confusingly, Jane E. Lewin's translation of the “Discours” (1980) renders both “diégèse” and “diégésis” with the English “diegesis.”47 In Palimpsestes, of 1982, Genette mentioned that “diégèse” is often treated as a synonym of “action” or “histoire.”48 He admitted to having at least partly contributed to that problematic usage in his “Discours du récit,” but he now distinguished clearly between “histoire” (“a succession of events and/or actions”) and “diégèse,” of which he wrote, “the diegesis, in the meaning suggested by the inventor of the term (Étienne Souriau, if I am not mistaken), which is the meaning I shall be using here, is the world wherein that story occurs.”49 Thus, Genette presented “diégèse” (rendered as “diegesis” in the English translation just quoted) as a recent term, and hesitantly attributed its coinage to Souriau (though without providing bibliographical references). He went on to explain that the semantic slippage, whether intentional or not, by which “diégèse” had come to denote the story was understandable, as it had been facilitated by the “metonymic relation” between the two (the story taking place in the diégèse).50 Furthermore, a by-product of this slippage was the use of the adjective “diégétique” to mean “relating to the story”—a usage that came in handy, in that “historique” would be open to misunderstanding (think of the English “historical”). Finally, Genette observed that this conflation of meaning seemed relatively harmless in ordinary circumstances, which do not require a distinction between the action and its spatiotemporal framework—the world in which it occurs.51 (Note that this conflation of story and storyworld is anyway all internal to the use of the modern term “diégèse,” and keeps it independent of the ancient “diēgēsis”—Genette's “diègèsis/diégésis.”)
Genette sensed the persisting disorder of the situation, so much so that he felt the need to discuss the matter again and at some length in another important book, Nouveau discours du récit of 1983.52 As for the modern term “diégèse,” he now specified that it had been propounded by Souriau in 1948 (though still without providing a reference, or an explanation for the choice of date),53 and he attempted to dispel the misunderstandings surrounding it. He insisted on the distinction between the concepts of diégèse and histoire, but he also repeated his pragmatic defense of the confusing use of the adjective “diégétique” with reference to the histoire. More importantly, Genette tried to resolve the frequent confusion between his “diégésis” and his “diégèse”—a confusion sanctioned by the practice in English-language scholarship, where both terms were often translated as “diegesis,” and both were referred to by the adjective “diegetic.” Unfortunately, he devised a remedy that turned out to be worse than the disease. He claimed that there existed two independent terms, “diégésis,” the ancient Greek for a narrative mode, and “diégèse,” Étienne Souriau's French term for the world of the narrated story. Indeed, he went as far as declaring that “[d]iégésis … has nothing to do with diégèse; or, if one prefers, diégèse (and I had no hand in this) is by no means the French translation of the Greek diégésis” (a version of the facts that appears to have survived as the fable convenue among a number of scholars). Furthermore, he professed that the adjective “diégétique” was used by both Souriau and himself exclusively in relation to the recent, filmological term, and never in relation to the Greek “diēgēsis,” implying that there was no risk of ambiguity.54 One possible implication, then, was that the filmologues had invented the French term while being entirely unaware of the Greek one—an unlikely hypothesis, as one would have to wonder how, in dealing with fundamentally narratological issues, they could have come up with such a word as “diégèse” without knowledge of the Greek “diēgēsis.” The alternative possible implication was that the filmologues shared Genette's awareness of the existence of two different terms, and his professed conviction that they were entirely independent of each other.55
In fact, there is clear evidence to the contrary: when they talked about “diégèse,” the filmologues thought precisely of the Greek word “diēgēsis.” In the work of Étienne Souriau, the term “diégèse” appears for the first time in print in an article of 1951, based on a lecture of 1950.56 There, Souriau discusses a number of technical terms that he claims emerged from teamwork—from a discussion that took place among a group of researchers under his direction. In particular, he states clearly that “diégèse” is based on the Greek “diēgēsis,” meaning “narrative.”57 (In this article, as in much scholarly literature since, the term fluctuates between indicating the story told and indicating the spatiotemporal universe in which that story takes place and which the story implies.) In the preface to a volume he edited for publication in 1953, Souriau insisted that the technical terminology developed by the filmologues—including “diégèse”—was the result of “un travail d'équipe.”58 In a didactic text of 1956 he would again write about “this universe of the work … that we will call the ‘diegetic universe’ (from the Greek word ‘διήγησις,’ ‘the story one tells’), after the manner of the filmologues.”59 And in her entry “Diégèse” for a dictionary of aesthetics, while claiming that she herself had created the term in 1950 within the group of filmologues, Anne Souriau stated that she had derived it “from the Greek ‘διήγησις,’ which designates a narrative and its content.”60 Thus, contrary to Genette's account, the filmologues were in fact thinking of the Greek term when they used “diégèse”; only, they were misconstruing its meaning by claiming that it could have something to do with the “story,” the content of a narrative (Étienne Souriau's “histoire qu'on raconte,” Anne Souriau's “contenu d'un récit”).61 Indeed, when claiming that different works can have the same narrative content, Anne Souriau referred to the latter as “diégèse,”62 whereas Plato, while showing that the same narrative content can be presented in different modes, had used the term “diēgēsis” not to refer to the content, but to discuss those modes.
Genette, on the other hand, must have become aware of the problems produced by the filmologues' terminology, but rather than nipping the terminology (and the problems) in the bud, he made the situation worse by adding to the chain of mistakes. (He seems to have misinterpreted the filmologues if he genuinely believed that in using “diégèse” they intended to coin a new term, deliberately different from and independent of the Greek “diēgēsis.”) The muddle was compounded by the great success and diffusion of Genette's writings on narratology, with their attendant terminology based on filmological practice (“homodiégétique,” etc.). Interestingly, a number of scholars have been content to take Genette's version of the story at face value, continuing to give 1948 as the date for Souriau's coinage of the term “diégèse,” and/or to repeat the assertion that Souriau had not derived “diégèse” from the Greek “diēgēsis.”63
As a coda to this section, I should add that in a fairly late, nonacademic book, Genette appears to have finally conceded that “diégèse” was simply a morphological adaptation of the Greek “diēgēsis,” thus in a sense coming back full circle to his use of the French word in 1966. His Codicille of 2009—a half-ironic miscellany of notes, reflections, and anecdotes presented in the alphabetical form of a dictionary—includes the entry “Néologisme.” Here, a near-octogenarian Genette wrote that he considered his contribution to neologism to have been greatly overestimated, and provided the example of the word “diégèse,” which, after all, “is borrowed from the theorists of film” and “barely gallicizes the Greek ‘diēgēsis.’”64
A Language Well Made
The reader patient enough to have followed me thus far will have sensed my strong disinclination to use the phrase “diegetic music” with the meaning it most frequently has in current studies of film and opera, and even my inclination to proselytize in that sense, which must seem a rather quixotic enterprise.65 I conclude my argument by responding to three possible (and interrelated) kinds of objection to my position, all of which are already voiced more or less explicitly in the scholarly literature.
The first objection is the easiest to dismiss. It is the frequent claim that, however fuzzy or porous the boundary between “diegetic” and “nondiegetic” music, the distinction has proven to be—and still is—of practical use, too much so to be disposed of. This argument amounts to confusingly identifying concepts with terms. Understanding the conceptual distinction between music that is (or could be) heard by the characters of a film or an opera, and music that is not, is important and, as noted above, often necessary for the understanding of the story represented; all I am doing here is contending that the words usually chosen to express that distinction are inaccurate, and that such an infelicitous lexical choice causes problems.
A second objection is that the phrase “diegetic music”—and for that matter the whole lexical area depending on “diegesis” and “diegetic” in the filmological sense—is anyway well understood in current critical practice. And yet, whatever the real extent of such an understanding, the fact remains that the current practice in studies of narrative, drama, film, and music involves the use of the terms “diegesis” and “diegetic” in both senses: both the ancient meaning and the 1950s meaning have survived, and since these meanings are very different (in some contexts, nearly opposite), the two terminologies are not simply living alongside each other, but are actually at (undeclared) war with one another because they are simply incompatible. While many scholars use the English “diegesis” and “diegetic” and their cognates in other languages in the filmological sense, others use these terms in line with the “diēgēsis” of Aristotle and post-Aristotelian writers. (Alain Boillat, for instance, has observed that scholars of literary theory still use the French “diégèse” “in the sense of the ‘diēgēsis’ of the classical tradition.”)66 Thus, when Karol Berger poses the question whether instrumental music of the Western art variety can be “diegetic,” he is referring to the primary meaning that the Greek “diēgēsis” and its derivatives in other languages have had for about twenty-three centuries in Western culture—one of two main modes of narrative presentation.67 The quick answer to that question would be no: such music is incapable of taking on a recounting mode in the way in which words can. In this sense, there is quite simply no such thing as “diegetic music.” One can then imagine the confusion of the reader who moves back and forth between “post-Aristotelian” writers such as Berger and “post-filmological” literature in which film and opera appear to present myriad examples of “diegetic music.” Moreover, the different uses of “diegesis” often coexist within the very same oral or written text.68 But the two lexical systems are, as I have said, incompatible; in fact, if we were to apply the ancient (but still active) opposition between diegesis and mimesis, we would have to conclude that music usually labeled “diegetic,” such as Cherubino's song, is, if anything, rather “mimetic.”69 We should, however, be wary of the risks involved in such applications of literary terminology to music, as they are often based on what Berger would call “inexact analogues.”70
A third and final objection, closely allied with the second, is that we should at this point leave an established practice alone. After all, linguists tell us that when it comes to so-called natural languages such as English or French, we can only register their development, not try to influence it or change it—we can describe, not prescribe. But what I have been discussing here is far from any “natural” language. It is a specialized terminology, very consciously developed and adopted by subgroups of scholars who constitute an infinitesimal fraction of the groups of native speakers of the respective “natural” languages; and it is the practice of scholarly communities to continually scrutinize and refine their terminologies, thus refining the attendant conceptual apparatus.71
It has been suggested to me a number of times that, since the terms whose use I seriously dispute have such wide currency, I should discuss possible alternatives, or at least the difficulties posed by the quest for such alternatives. I will therefore comment briefly on a few other expressions—several of them already in use within the community of opera scholars—but will do so with an obvious proviso: what matters, more than this or that lexical choice on the part of individual scholars, is that the choice be made with full awareness of its implications, and rendered explicit wherever necessary.
While Edward Cone's opposition of “realistic” and “operatic” music will be sufficiently plain if the writer explains the meaning of those terms at the outset, they remain potentially problematic on account of their many other connotations.72 (Furthermore, “realism” is already a sufficiently contentious subject, while in the general sense of the word virtually all music in opera is “operatic.”) The frequently adopted dyad “stage music” vs. “pit music” is also problematic:73 if nothing else, the music representing the sound of Susanna's guitar in Cherubino's “Voi che sapete” comes from the pit, not the stage. Expressions such as the German “drameninhärente Musik” and the Italian “musica intrinseca al film” are good options as far as meaning is concerned, but they are not easily translated into English, where both “inherent” and “intrinsic” no longer suggest “situated within” but rather “essential.”74 In order to refer to music audible as music by the characters of an opera, scholars keen on keeping to ancient Greek terminology may want to consider replacing “diegetic” with “mimetic”; the foregoing discussion should have made it sufficiently clear that in such cases “mimetic music” would be an altogether more apt definition than “diegetic music”—but it should also have made clear that the use of a term with as complex a history as “mimesis” would present its own risks, and would at least require specification that the term was being used in its particular sense of “enactment.” I have occasionally proposed to students the phrase “music as ‘music,’” which suggests that the operatic passages thus defined are the few in which the music we hear actually represents music; I realize, however, that the phrase is rather cumbersome for regular use as standard terminology (and it lacks an obvious opposite term). Finally, if I were to put forward technical neologisms, one possible solution would be that offered by the oppositional couple of “endogenous” vs. “exogenous” music—music “originating from within” vs. “originating outside” (that is, from within or outside the world in which the story presented takes place). Again, scholars should ideally clarify their use of a specific terminology as soon as they introduce it in a piece of writing; and this consideration applies whatever the chosen vocabulary—though I would insist that “diegetic” remains too polysemous a term to be unambiguously functional.
Even readers disinclined to abandon their deeply ingrained linguistic habits relating to “diegesis” and “diegetic” will, I hope, have at least gained by this point a finer sense of the complex history of these terms and their associated ideas. Moreover, the story I have reconstructed here might serve more generally as an apologue, reminding us of the sometimes peculiar ways in which bits of our collective wisdom emerge and become accepted. In that respect, I wish to offer an ironic conclusion by suggesting that one of the very people at the origin of the muddle examined in these pages would probably agree with my interpretation and reconsider his lexical choices. In the article of 1951 that marks the first appearance in print of the modern-day “diégèse,” Étienne Souriau argued that a scholarly field requires what Condillac called “a well-made language”: “refusing to make the effort required for establishing such a language … is condemning ourselves from the outset to a regimen of ill-posed questions, of vague inquiries,” whereas making such an effort is “a small evil necessary for obtaining a great good—clarity, and the posing of problems in a rigorous way.”75
This article began life as one of a series of lectures given as a Visiting Professor at the University of Pavia-Cremona in 2016. I wish to thank Fabrizio Della Seta, whose generous hospitality in Cremona went far beyond the call of academic duty, as well as the colleagues and students who made my stay there both enjoyable and intellectually engaging. For feedback on a draft of the article I am grateful to Karol Berger, Katharine Ellis, Stephen Halliwell, Leofranc Holford-Strevens, and Glenn Most. I also benefitted from conversations with Isabella Gualandri, Michael Reeve, Malcolm Schofield, and Rita Zaninotto.
Grétry, Mémoires, 529: “Il y a chanter pour parler, & chanter pour chanter.” Translations are mine unless otherwise indicated.
Though theoretical approaches to narrative have a longer history, the coinage of the term “narratology” is generally attributed to Tzvetan Todorov, who first used it in French in 1969: Todorov, Grammaire du Décaméron, 10: “narratologie.”
Henry M. Taylor's “The Success Story of a Misnomer” exposes that lexical problem from its very title, but is far from the only piece of writing to mention it.
The lack of synchronism between the “now” of the narrator and that of the narrated is “surely the single most essential feature” of the recounting mode: Berger, “Diegesis and Mimesis,” 412 (Theory of Art, 168).
The scholar who has done most for our understanding of the term “mimesis”—and of its complement, “diegesis”—is Stephen Halliwell. “Halliwell's writings on mimēsis constitute … a mighty effort of synthesis that has no rivals in the current critical literature”: Donini, introduction to Aristotle, Poetica, xxi n39 (“questi lavori di Halliwell sulla μίμησις costituiscono … un poderoso sforzo di sintesi che non ha rivali nella letteratura critica esistente”). The most relevant of Halliwell's publications, to which the present study owes much, include—besides his edition of Aristotle's Poetics in the Loeb Classical Library—Halliwell, Aristotle's Poetics; Halliwell, “Subjection of Muthos to Logos”; Halliwell, Aesthetics of Mimesis; Halliwell, “Theory and Practice of Narrative”; and Halliwell, “Diegesis—Mimesis.” See also Palumbo, Mimēsis. On the post-Aristotelian tendency to the lexical polarization between the nouns “diēgēsis” and “mimēsis” (as between their related adjectives and verbs), see Nünlist, Ancient Critic at Work, and Halliwell, “Diegesis—Mimesis.”
Republic 3.392d, 3.393d, 3.394a. I transliterate from the Greek text in Plato, Republic: Books 1–5.
3.393a: “legei te autos ho poiētēs”; 3.394c: “di' apangelias autou tou poiētou.” All editorial insertions in quoted material are mine.
For instance, 3.393c.
3.394c: “dia mimēseōs holē.”
3.393c: “dia mimēseōs tēn diēgēsin poiountai.”
The expression “aneu mimēseōs … diēgēsis” can be found at 3.393c–d and 3.394a.
Poetics 1448a20–24. My source for the Greek text of Aristotle's Poetics is Halliwell's edition. In order to indicate the precise position of a passage, I adopt the standard practice of referring to the numbering in the 1831 edition by Immanuel Bekker. (Modern editions and translations of Poetics usually provide, in the margins of their texts, references to the matching position in Bekker—though if they provide line numbers they do so only for every fifth line, without specifying further.) The first number refers to the page in Bekker's edition, the “a” or “b” that follows to the left or right column of that page, and the following number(s) to the line(s) within that column. I consulted Bekker's edition of Poetics in the facsimile contained in Aristotelis opera. On the terminology of Aristotle's Poetics, see Wartelle, Lexique de la “Poétique” d'Aristote.
Thus D. W. Lucas in his edition of Aristotle's Poetics, 66.
Translation from Halliwell's edition of Aristotle's Poetics, 35 (Halliwell's emphasis). I have not added in square brackets the original Greek wording for the enactive mode (Halliwell's “direct personation,” “direct enactment”), as in this passage Aristotle does not use a single term for it, but two different periphrases, as can be glimpsed through the translation by Malcolm Heath that I quote as an alternative soon afterward. The “objects” in question are narrative contents, whereas the “media” in this case are words.
Translation from Malcolm Heath's edition of Aristotle's Poetics, 5.
For my present purpose it suffices to highlight the degree of overlap between Plato's and Aristotle's views of the narrative modes. But if one were to accept the second interpretation of Aristotle just discussed, there would be theoretical implications at which I can only hint here. Classing the mode of epos (and the novel) as a subtype of the recounting mode would emphasize the “quotation” aspect of its passages in direct speech, to the detriment of the aspect of immediate enactment that would make these passages kin to the direct speech of the characters of drama. But ancient theorists, differently from some of the modern ones, often do not distinguish between the enactment of drama and that of direct speech in epos: after all, the two types of enactment are implicitly equated both by Plato's expression “diēgēsis di' amphoterōn”— “narrative by means of both [modes]”—and by the later writers' term “miktos” (mixed). See also Nünlist, Ancient Critic at Work, 101. Neither can I discuss here a second issue, that of the differences between drama as text and drama as performance. Here, too, the ancient theorists generally do not distinguish: when it comes to drama, “mimēsis” in its specific sense of “enactment” can be used with reference both to what a poet does in writing the characters' direct speech, and to what an actor does in impersonating a character on stage; see Donini, introduction to Aristotle, Poetica, xxix n60. It seems to be generally acknowledged that Aristotle was interested in tragedy as a literary text more than as a performative enterprise; see Donini, “Poetica e retorica,” 336. On both these issues, see Schaeffer, Qu'est-ce qu'un genre littéraire?, 90–92.
See, for instance, Kirby, “Mimesis and Diegesis,” 118: “For Plato, diegesis is the genus, and mimesis determines the differentiae; the opposite is true in Aristotle.”
In spite of a number of problematic details in his exposition, André Gaudreault comes to a very similar conclusion—that Plato's and Aristotle's systems are largely compatible, once we understand the lexical differences: Gaudreault, Du littéraire au filmique, 57, 64, 68 (From Plato to Lumière, 40, 46, 50).
See Halliwell, Aristotle's Poetics, 14–15. For Plato, Aristotle, and many other classical as well as neoclassical theorists, mimesis (in its broader sense of “representation”) “was the key to the primary question of the relation between works of art and the world”: ibid., 109. On music's representational powers even in the absence of a poetic text, see Rocconi, “Effetti speciali sonori.” See also Schofield, “Music All Pow'rful.”
This point is made in Richardson, “Drama and Narrative,” 142.
See Halliwell, “Diegesis—Mimesis,” 132.
See Poetics 1460a9, where Aristotle writes that epic poets other than Homer “engage in mimesis [mimountai] only briefly and occasionally” (translation from Halliwell's edition, 123). This use has already been noted by Michael W. Haslam in his “Plato, Sophron, and the Dramatic Dialogue,” 22.
For a detailed examination of Plato's uses of “mimēsis,” see Halliwell, Aristotle's Poetics, 116–21. On these oscillations in both Plato and Aristotle, see Nünlist, Ancient Critic at Work, 97n14. Domenico Pesce observed that such fluctuations have often caused confusion in modern interpreters: Pesce, introduction to Aristotle, La poetica, 9.
For “diēgēsis” in Aristotle's Poetics, see 1456b11–12 and 1459b26; for “diēgēmatikos,” 1459a17, 1459b33, and 1459b36–37.
Aristotle uses the verb “apangellein” three times in Poetics, always with the sense of “recounting”: 1448a21, 1460a18, 1460a31–32.
See, for instance, the passage from Republic 3.394c cited above.
See Halliwell, “Diegesis—Mimesis,” 131.
See Haslam, “Plato, Sophron, and the Dramatic Dialogue,” 20, and Nünlist, Ancient Critic at Work, 95, 99. The use of “dramatikos” with reference to the enactive mode has at least one precedent in Aristotle: of the two occurrences of this adjective in Poetics, one has that meaning (1448b35–36).
See Halliwell, “Theory and Practice of Narrative,” 21. In this essay, Halliwell warns against the risk of thinking that Plato's Republic offers a systematic theory of narrative rather than one that is at most embryonic and incomplete.
In a passage from Republic (3.397b), Plato uses “diēgēsis” to mean “haplē diēgēsis.” Cf. Gaudreault, Du littéraire au filmique, 63n24 (From Plato to Lumière, 176n19).
Nünlist, Ancient Critic at Work, 115.
An example among many is Hauch, “Diegesis,” 533.
For Tom Gunning, for instance, Gaudreault has shown that the opposition diegesis/mimesis is “a modern reworking, and indeed distortion, of the original [Greek] texts, in which these two forms had never been systematically opposed”: Gaudreault, From Plato to Lumière, xvii–xxv, here xxi.
The phrase occurs twice in Poetics (1459b33, 1459b36–37).
For the claim that the terminology was the result of teamwork, see below.
For a study of the filmological term in the francophone sphere, see Boillat, “La ‘diégèse.’”
There are two English translations of the essay (both based on the version in Figures II), one by Ann Levonas as “Boundaries of Narrative” (1976), the other by Alan Sheridan as “Frontiers of Narrative” (1982). Genette must have intended the term “diègèsis” (first introduced on page 152 of the article, page 50 in Figures II) as the Greek word in transliteration rather than a French word: it is italicized and has grave accents. He seems to have for a number of years used the system by which, where accents are needed on the letter “e” in French, the Greek letters eta and epsilon are transliterated as “è” and “é” respectively—witness his spelling the genitive of “mimēsis” as “mimèséôs” in Genette, “Genres, ‘types,’ modes,” 392. (This article is better known in its expanded version: Genette, Introduction à l'architexte, here 14.) In a later passage of “Frontières du récit,” Genette makes it explicit that “diègèsis” is his spelling for the Greek term (“en grec, l'emploi du terme commun diègèsis,” 156; Figures II, 56). In the book version of three years later, “diègèsis” becomes “diégésis,” possibly as a result of a decision by the book's editors. Similarly, André Gaudreault, who claims in the French version of his book to have “adopted for the romanization of the Greek terms the most faithful phonetic transcription,” uses the spellings “diègèsis” and “mimèsis” for the Greek terms: Gaudreault, Du littéraire au filmique, 38n14 (“j'ai adopté pour la romanisation des termes grecs la transcription phonétique la plus fidèle possible”).
Genette, “Frontières du récit,” 154 (Figures II, 52; “Boundaries of Narrative,” 3; “Frontiers of Narrative,” 130).
Genette, “Frontières du récit,” 156 (Figures II, 56). The two English translations use “diegesis” to render both Genette's “diègèsis” and his “diégèse,” though when it comes to translating the only instance of the French “diégèse” Sheridan de-italicizes “diegesis”—probably treating it as an English word in order to reflect Genette's suggestion of a modern-language use (“Frontiers of Narrative,” 133).
Genette, “D'un récit baroque,” 211: “l'univers spatio-temporel auquel se réfère la narration première.” The essay's first footnote gives its origin as a communication given at the Journées internationales d'études du baroque, Montauban, 1968.
Genette, “Discours du récit,” 72n1: “le terme diégèse, qui nous vient des théoriciens du récit cinématographique.” “Discours du récit” occupies the major part of the book Figures III, and appeared in English as Narrative Discourse (1980), where this footnote presents a translation different from mine (27n2).
On page 72, note 1, Genette introduces “diégèse” as a term he uses “in the same sense” (“dans le même sens”) as “histoire,” and indeed the annotated index of Figures III has an entry “histoire ou diégèse,” which simply gives a reference to page 72. But under the entry “diégétique,” Genette claims that “in the current usage, the diégèse is the spatiotemporal universe referred to by the narrative discourse” (“dans l'usage courant, la diégèse est l'univers spatio-temporel désigné par le récit”).
Compare “Discours du récit,” 72n1, 75, with Narrative Discourse, 27n2, 30.
Genette, Palimpsestes, 341 (Palimpsests, 294).
Genette, Palimpsestes, 341–42 (translations from Palimpsests, 295).
Genette, Palimpsestes, 342 (translation from Palimpsests, 295).
Genette, Palimpsestes, 342 (Palimpsests, 295). Even among the scholars who use “diégèse” or “diegesis” to refer to the world represented, there are those who include the action (the story) in that world, and those who explicitly rule it out. Film theorist Christian Metz, for instance, claimed that the “diégèse” of the Souriaus comprises everything represented, both the story (including characters and events) and the spatiotemporal world that its fiction implies: Metz, “Quelques points de sémiologie,” 59–60 (Essais, 1:100–101; “Some Points in the Semiotics,” 98). For narratologist John Pier, on the other hand, “diegesis” in the sense of “the spatiotemporal universe in which the story is situated … clearly does not cover actants, actions and events”: Pier, “Diegesis,” 217–18.
The book appeared in English as Narrative Discourse Revisited (1988).
That Genette's references to Étienne Souriau and filmologie were slow to emerge and somewhat reticent is noted in Boillat, “La ‘diégèse,’” 218, 224–26.
Genette, Nouveau discours du récit, 13 (Narrative Discourse Revisited, 17–18, translation from 18). That the adjective “diegetic” refers only to the world of the story (“diégèse”) and not to the narrative mode (“diēgēsis,” “diégésis”) is implied also in Prince, Dictionary of Narratology, s.v. “diegetic.”
David Neumeyer writes that the distinction between the Greek “diēgēsis” and the “diégèse” of filmologists “is lost in English translation”—where both become “diegesis”—“but to little harm if it is remembered that the contemporary meaning does not relate to or depend on the original Greek meanings in Plato and Aristotle”: Neumeyer, “Diegetic/Nondiegetic,” 27. But it is precisely because the two terms have such different meanings that translating them with the same English word is unfortunate and confusing, as Genette himself had lamented: Genette, Nouveau discours du récit, 13 (Narrative Discourse Revisited, 18).
Souriau, “La structure de l'univers filmique.” The author claims this to have been the opening lecture of a course given in the first semester of the academic year 1950–51 at the Institut de Filmologie (231).
Souriau, preface to Souriau, L'univers filmique, 5–6, 9. It is to this piece of writing that David Bordwell must be referring when he writes, “[i]n 1953, the term diegesis was revived by Etienne Souriau to describe the ‘recounted story’ of a film”: Bordwell, Narration in the Fiction Film, 16.
Souriau, Les grands problèmes, 11: “Cet univers de l'œuvre … que nous appellerons à la façon des Filmologues l'univers diégétique (du mot grec διήγησις, histoire qu'on raconte).” This text is a typescript that was mechanically reproduced as material for a course in aesthetics given by Souriau at the Sorbonne; I consulted a reprint of it from 1960.
Souriau, “Diégèse,” 581: “du grec διήγησις qui désigne un récit et le contenu d'un récit.” Anne Souriau took over the editorship of this multi-authored dictionary from her father after his death in 1979.
The same claim survives in recent literature. David Bordwell and Kristin Thompson, for example, write of “diegesis (the Greek word for ‘recounted story’)”: Bordwell and Thompson, Film Art, 76. A fairer summary of the situation is that given by Gaudreault: in using the term “diégèse,” “Souriau retained the French equivalent of the [Greek] word ‘diēgēsis’ rather than the concept: he does not pair it with ‘mimēsis,’ and gives it the sense of ‘story’ rather than ‘narration’”: Gaudreault, Du littéraire au filmique, 25n15 (“Souriau a retenu l'équivalent français du mot diègèsis plutôt que le concept: il ne le couple pas à mimèsis et lui donne le sens d'histoire plutôt que de récit”). My translation here differs from that in From Plato to Lumière, 66n15.
Souriau, “Diégèse,” 583.
In 1984 Paul Ricœur repeated Genette's claims that “diégèse” had been proposed by Étienne Souriau in 1948, and that the term did not make reference to the Greek “diēgēsis”: Ricœur, Temps et récit, 2:121n1 (Time and Narrative, 2:179–80n39). Similarly, the three editions (between 1986 and 2010) of an authoritative dictionary of semiotics refer to Étienne Souriau's “La structure de l'univers filmique” but date it as 1948: see Pier, “Diegesis,” and the related bibliography (Encyclopedic Dictionary of Semiotics, 3:387). Alessandro Cecchi makes a similar claim about Souriau “toward the end of the 1940s,” without giving a reason for such a date: Cecchi, “Diegetic versus Nondiegetic” / “Diegetico vs. extradiegetico”; the temporal reference is made only in the Italian version of the article (2). For Taylor, “Genette (aligning himself with Étienne Souriau) asserts that these terms [“diégèse,” “diégétique”] are not derived from the Greek diegesis”: Taylor, “Success Story of a Misnomer,” 3. (The page numbers I provide for this article refer to the independent pagination of the article in the downloadable PDF version.)
Genette, Codicille, 194: “Diégèse, qui francise à peine le diègèsis grec, est emprunté aux théoriciens du cinéma.”
If one adopts the filmological practice of indicating by “diegesis” the world of the story narrated, it is an obvious step to apply “diegetic” to any sound (including music) that originates within that world. The earliest occurrence I have found in print of such an application to music is in an article from the mid-1970s in which Claudia Gorbman refers to “the opposition of diegetic and extra-diegetic music,” and specifies, “I use ‘diegetic’ to refer, in the semiotic sense, to all that occurs within the apparent world of the narrative”: Gorbman, “Music as Salvation,” 17. Gorbman's book Unheard Melodies (1987) is probably the main influence behind the diffusion among musicologists of the phrase “diegetic music” used in this sense.
Boillat, “La ‘diégèse,’” 241n23: “dans le sens de la diegesis de la tradition classique.”
Berger, “Diegesis and Mimesis.”
The Cambridge Companion to Narrative exemplifies the problematic coexistence of the two very different uses of “diegesis.” Both Heta Pyrhönen (“Genre,” 110) and Brian Richardson (“Drama and Narrative,” 151) use “diegetic” with reference to the recounting mode (in opposition to “mimetic”), whereas for Jason Mittell (“Film and Television Narrative,” 160) diegesis is “the storyworld which the characters experience.” The book's glossary (s.v. “diegesis”) gives as the first meaning the modern one, that referring to the story or storyworld (as already in Prince, Dictionary of Narratology); it has as a second meaning the ancient one (a mode of presentation), but uses here the adjective “diegetic” (differently from Genette, who claimed to reserve the adjective for the modern, filmological meaning). At the risk of sounding anecdotal, I will add that on one occasion I attended an undergraduate lecture by a film-music scholar, who afterward conceded to having used “diegesis” with both its main meanings in the lecture—an ambivalence that is far from unusual in published scholarship.
In the area of film studies, several scholars have suggested quite simply that, in Taylor's words, “what should have been called mimesis and mimetic came to be designated by the terms diegesis and diegetic”: Taylor, “Success Story of a Misnomer,” 2. A similar irony applies to some of the elements that film scholars usually label as “nondiegetic” (superimposed titles, voice-overs): in fact these often fall on the diegetic side of the old narratological opposition; see Fuxjäger, “Diegese, Diegesis, diegetisch,” 29. See also Kessler, “Von der Filmologie zur Narratologie.”
Berger, “Diegesis and Mimesis,” 419, 420 (Theory of Art, 173, 174).
In a sense, there is something of all three of the abovementioned objections in the following statement: “Of course, by now this terminology has been so well established that it would be futile not to use it in its accustomed sense. It has been particularly useful in designating aspects and features of filmic sound as it relates to the relatively closed story-worlds of fiction”: Taylor, “Success Story of a Misnomer,” 3.
Cone, “World of Opera,” passim.
For a recent example, see Berger, Beyond Reason: Berger more than once uses the expression “stage music” for what he defines as “music heard as such by [the] characters” (252), and sets it against “pit music” (245).
Thomas Betzwieser, who uses “drameninhärente Musik,” also defines such music as “diegetic” and “realistic” (using scare quotes only for the second term): Betzwieser, “Verisimilitude,” 304–5. For “musica intrinseca al film,” see Sergio Miceli, Musica per film, 525. I find Carolyn Abbate's distinction between “phenomenal” and “noumenal” music to be a potentially confusing application of philosophical (especially Kantian) terminology, though my objection would require too long an explanation for the present context: Abbate, Unsung Voices, esp. 5, 97, 119, 122. Gary Tomlinson takes over Abbate's terminology cum grano salis, and suggests a partial corrective to her use of “noumenal”: Tomlinson, Metaphysical Song, esp. 88–89.
Souriau, “La structure de l'univers filmique,” 231 (“une langue bien faite”; “Refuser l'effort nécessaire pour établir cette langue … c'est se condamner d'avance au régime des questions mal posées, des recherches vagues”), 234 (“un petit mal indispensable pour un grand bien: celui de la clarté, et d'une position rigoreuse des problèmes”).