Todorov and poetics

Tzvetan Todorov’s Introduction to Poetics—in the French, simply Poétique, though it originally appeared in a collection of essays with the (significant) question there interrogated (Qu’est-ce que le structuralisme?) appended—Todorov’s book is many things: an attempt to give unity to the structuralist investigation of literature, or at least (as the original French title makes clear) show how structuralism bears generally upon literary study; an attempt to give direction to many of the trends in literary study, which (thanks to Todorov’s remarkable breadth of knowledge) it continually takes into account; an attempt to project a future area for criticism. Above all, it is a clearheaded, sensible work: Todorov does not make superficial points, but condenses reflections that must be products of years of work with literature.

Todorov accomplishes all this first and foremost by proposing a field of investigation, poetics (the name is less important--but I explain it somewhat in a comment here), and skillfully making certain distinctions which not only saturate this field with useful concepts—that is, allow it to cull just the right amount from other trends in literary study—but also require the formulation of new concepts, made possible by the general trajectory in which the field is set moving. In short, the result of proposing the field of poetics is not a mashup of all sorts of theories, but a clean framework that is nevertheless flexible, meant to be used and developed in directions that, while keeping it consistent with itself, do not easily launch it into absurdity.

The founding gesture of poetics will demonstrate this clearly. Todorov distinguishes poetics from interpretation on the one hand, and from a science (a word he uses loosely) on the other. Interpretation, which Todorov usefully says may include diverse activities also called exegesis, commentary, explication de texte, close reading, or analyse, is an approach to literature that “sees the literary text itself as a sufficient object of knowledge” (3). In other words, its procedure implies the existence of an empirical object that, if thoroughly known, would give us all we need as critics who study literature. In fact, literature itself, as something apart from these objects, does not even exist for interpretation, to the extent that one could say—though Todorov doesn’t put it this way—that interpretation involves the dissolution of any idea of literature. What Todorov does say is something like the converse: interpretation involves only the consideration of real works, and the establishment of these works’ meanings. Interpretation’s aim “is to name the meaning of the text examined” (4). Its aim is also to fail at this, since it will be “forever incapable of realizing the meaning, but only a meaning” (4). It is important to note, however, that this means the failure is due to how what it specifies is so particular, so real that it demands, along with the effacement of literature itself or the idea of literature, the effacement of the critic who determines or qualifies this meaning, who makes it a meaning “subject to historical and psychological contingencies” (4). The failure is due to the ultimate impossibility of “fidelity to the object,” the inability, at bottom, “to make the text itself speak” (4), and not because at some point, a generalization must be made about how the work is indeed literary which will force the object into abstraction and interrupt interpretation. In short the idea of literature is still in the process of dissolution—it is precisely what never comes back to frustrate interpretation.

This is important because science takes up precisely this task: its aim “is no longer the description of the particular work, the designation of its meaning, but the establishment of general laws of which this particular text is the product” (7). These laws, however, are not laws of literature: a good example is a reading by Freud, who establishes psychoanalytic laws that explain a particular literary text (in other words we don't, without hesitation, call these laws literary, because they aren't). The work is “transpos[ed] into the realm considered fundamental” (7) or translated: its only meaning would be that psychoanalysis can appropriately consider such a text. Remember, however, naming the meaning is precisely foreign to science’s task, and therefore that science is by no means unproductive in this area. Laws are indeed being established which bear upon the idea of literature—historical, psychological, ethnographic, sociological, philosophical laws, which change our notion of what real works of literature are possible.

Poetics, as Todorov puts it, is not a compromise between the two activities here, but what “breaks down the symmetry thus established,” (7) or cuts across all of the features here elaborated. The cut, the disturbance of symmetry and mutual exclusion, is achieved by placing poetics within the field of science, but by making that science into a science of literature—which entails periodic contact with actual works and therefore interpretation.

Let us lay this out as clearly as possible. Todorov says,

In contradistinction to the interpretation of particular works, it [poetics] does not seek to name meaning, but aims at a knowledge of the general laws that preside over the birth of each work. But in contradistinction to such sciences as psychology, sociology, etc., it seeks these laws within literature itself (6).

In other words poetics does not impede the consideration of the idea of literature, but does not seek to make this idea take the form of laws that are not about literature. Thus, “it is not the literary work itself that is the object of poetics” (6). Each work is regarded only as the manifestation of something abstract and general, “of which it is but one of the possible realizations” (7). We thereby put poetics in the category of science—and Todorov appropriately refers to it as “this science:” “this science is no longer concerned with actual literature, but with a possible literature” (7). Put differently, “the goal of this study is no longer to articulate a paraphrase, a descriptive résumé of the concrete work, but to propose a theory” (7).

At the same time, this theory is not external to literature: the theory proposed is “a theory that affords a list of literary possibilities, so that existing literary works appear as achieved particular cases” (7). Thus, interpretation is not excluded from this science: “The relation between poetics and interpretation is one of complementarity par excellence. A theoretical reflection upon poetics that is not sustained by observations of existing works always turns out to be sterile and invalid” (7). It pulls the scientific tendencies of poetics away from other sciences and focuses it upon literature—but in a way that we will soon specify, about which it is important, nevertheless, to say the following: interpretation certainly is not only present to ensure that this tendency is kept in check, as it has been present especially in Anglo-American circles. What we have here in Todorov is less an opposition between “science and poetry” (a famous, tired opposition, which also uses a different notion of “science”) than a suspicion towards interpretation, which, as we said, tends to dissolve the idea of literature—that is, doesn’t oppose itself to anything, or doesn’t stop at opposition, but invades and imposes upon anything that would abstract away from the real text, that would conceive of any possible literature other than the book in front of one’s face. If this means that the emphasis in poetics is indeed put upon science, if poetics indeed tends to establish more laws than suggest meanings, this is only because “a massive imbalance in favor of interpretation characterizes the history of literary studies” (12). But the conclusion Todorov reaches from this is that “it is this disequilibrium that we must oppose, and not the principle of interpretation” (12). And insofar as this is the case, the emphasis on science is not made to the detriment of interpretation: it just means interpretation is made into one of the many things that pulls poetics as a science away from literature. We may state this more robustly as follows: interpretation in poetics remains not primarily as what keeps the science in poetics in check, but which allows science to test its possible literatures against reality, or restrict a little the amount of the possible—that is, make sure the science is not wholly unmotivated, to use (too metaphorically) the semiological term. Insofar as this is the case, interpretation is still a major part of poetics. In fact, one thing can indeed be said for certain, even if science is emphasized: “neither of the two activities takes precedence over the other: both are ‘secondary’” (8). The emphasis on science is not one achieved by a hierarchy: it is the strategic focusing on what is already known to be merely secondary.

We then broach the more specific question of how precisely the science is actually kept in check, or is made into a science of literature—if indeed interpretation is not the thing that brings this about. In an earlier paper entitled “Structural Analysis of Narrative” (translated in Novel 3.1 [Autumn, 1969], p. 70-76), which shares many of the same points and even the same formulations with Poétique, Todorov gives us another distinction that can help us here: he says that we can discern internal approaches to a literary work and external ones. What is normally thought is that science is an external approach, or one of abstraction, and interpretation an internal approach:

For example, when Marxists or psychoanalysts deal with a work of literature, they are not interested in a knowledge of the work itself, but in understanding of an abstract structure, social of psychic, which manifests itself through that work. This attitude is both [that of science] and external. On the other hand a New Critic (imaginary) whose approach is obviously internal, will have no goal other than understanding the work itself; the result of his efforts will be [an interpretation:] a paraphrase of the work, which is supposed to reveal the meaning better than the work itself (70).

In Introduction to Poetics Todorov nicely uses these terms (in a phrase so key I'll set it off) to show precisely how poetics is situated, how it breaks down the symmetry between science and interpretation:

Poetics is therefore an approach to literature at once "abstract" and "internal" (6).

What we are asking then, in asking how poetics as science remains a science of literature, is precisely how this science, which we grant to be abstract, can also be internal if it is not clearly employing interpretation for these purposes.

The answer is that, as a science of literature, its object is not an object, but a discourse. Put differently, if a science just has an object, it is not a structuralist science—and the only science under consideration here is of the structuralist sort. But what, then, is a discourse? Todorov in another essay, parroting Benveniste (see “Subjectivity in Language” in Problems in General Linguistics), explains that discourse is a language in action, a language as it is put in use (in the semiological sense of langue—that is, a finite set of rules making possible all utterances), or a used-language, prior to its being an actual, real moment of language use, or an utterance (parole):

It is necessary to introduce [this concept of discourse] because the rules of language, which are common to all who use it, constitute only a part of the rules which govern our concrete verbal production. They only fix the norm of grammatical combinations within a sentence, a phonology, and a common meaning for words. But between the set of rules common to all utterances [langue] and the exact formulation of a specific utterance [parole] there is a gulf of indeterminacy. The gulf is bridged by the rules of each particular discourse (thus an official letter will not be written in the same way as an intimate one), as well as by the limitations inherent in the context of the speech act (the identity of the speaker and the listener, the time and place of the speech act). The rules of discourse are more restricted than those of [a] language, but less restricted than those of a specific speech act (“The Notion of Literature,” in New Literary History 5.1 [1973], 5-16, p. 14).

If discourse is a language in action, if we specify these rules we are not talking about something external to the utterance, even though what we specify may be abstract: the discursive statement by (Benveniste’s) definition involves (or rather is only “filled” by) the act by which it is uttered (when the utterance is considered as a product of a language: which is why it is not a speech act or performative). This is not necessarily so from the perspective of science: science may discuss structure in general and ignore (for the most part) discourse.

But, again, because what we are discussing is discourse, while we do not consider the literary work itself, we still talk about what is internal to these objects, or what will be realized by them, so we are also not open to the charge that what we talk about is external. At the same time, this is also why they are not pure—that is, not really abstractions, but indeed descriptions (of literary discourse). Todorov puts it this way:

Poetics situates […] abstract notions not within the particular work but in literary discourse; it asserts that they can exist there alone, whereas in the work we always deal with a more or less “mixed” manifestation; poetics is not concerned with this or that fragment with a work, but with those abstract structures which it names “description” or “action” or “narration” (9).

Some skeptics of structuralism might still object that this “there” is really nowhere: literary discourse is still too unreal for them. Unless there is some contact with empirical works, and not a continual interrogation of “the properties of that particular discourse that is literary discourse” (6), we condemn the literary study to the study of pure convention, to the description of literature as something completely unessential. So Christopher Norris in a review of the Introduction disparages Todorov by saying that poetics defines literature “only in terms of its own discursive methodology” (Modern Language Review, 78.3 [July, 1983], 636-637, p. 636). Without requiring contact with the empirical, the definition is “circular” (636).

I would doubt that this is so: there are indeed many questionable elements in Todorov’s account (one of which is that “reading is a trajectory […] that constitutes the text in space and not in linearity,” [5]: the spatialization of the line is precisely self-serving and circular, as much as it is typically structuralist, because it does not thereby abolish the line but, precisely as in Hegel, preserves it within space) but this I find to be pretty solid, if only because Todorov draws the ultimate conclusion from it. This is the following: if the poetics studies is literary discourse, and indeed does not of necessity require contact with real, empirical phenomena (“poetics is not the sum of empirical phenomena,” 10), then poetics must be a finite endeavor, which disperses itself as soon as the discourse which it studies disperses. And Todorov affirms this, and far from closing on what Norris—whose aim can only be to consciously mislead—calls a “grave resignation” (636) in fact seems quite happy that poetics has this sort of pertinence (once more, in the semiotic sense): Todorov says that as studies of other discourses, helped along by the example of the study of literary discourse, may indeed proliferate, poetics’

own role will be reduced to little enough: to the investigation of the reasons that caused us to consider certain texts, at certain periods, as ‘literature.’ No sooner born than poetics finds itself called upon, by t he very power of its results to sacrifice itself on the altar of general knowledge. And it is not certain that this fate may be regretted (72).

This is strong stuff, but it isn’t resigned. It is just announcing that wonderful but (in most of America) still unappreciated property of things structurally considered: they are finite, without ground, and without essence—wholly conventional, never natural. This is why the above argument does not strike me as circular: it is circular only insofar as structuralism as a whole is circular. This, though, is a distinct possibility: we can object to Todorov in another way, which would be that interpretation, perhaps not as the consideration of empirical works but instead as the consideration of works still “in themselves” in one aspect—in their singularity, their irreducibility—will at some point be impeded precisely by the focus on discourse. The irreducibility will precisely be reduced to structure. The only defense we can make for Todorov here would be the one that, again, defends structuralism generally from this “post-structualist” critique: it would be to affirm, with Todorov, the benefits of undoing—even slightly, or perhaps only slightly—the imbalance in favor of interpretation which has characterized the study of literature. This, though, is risky, I would do so only in the following manner, which is not Todorov’s, and which I sketch out much too quickly: unless such focus on singularity or irreducibility is paired with some way to extrapolate, to situate itself and strategically reduce itself, it remains too indeterminate even to be considered impractically, let alone practically, and in fact—this is the highest danger—can be confused precisely with the empirical. Structuralism may help in this process of extrapolation.

Derrida, de Man, materialism

Photo by Hendrick Speck

In Memoires for Paul de Man, Derrida somehow claims that de Man's assertions about the materiality of language constitute "an original materialism:"

There is a theme of "materiality," indeed an original materialism in de Man. It concerns a "matter" which does not fit the classical philosophical definitions of metaphysical materialisms any more than the sensible representations or the images of matter defined by the opposition between the sensible and the intelligible. Matter, a matter without presence and without substance, is what resists these oppositions. We have just placed this resistance on the side of thought, in its strange connivance with materiality.
-Memoires, 52

The last sentence, which summarizes de Man's position, is accurate. But I might claim roundly (despite various problems here: "matter" might also have resonance with Sache; "original" obviously also refers to how this "connivance" is immediately afterward said to relate to history, 53) that Derrida just can't be sure whether de Man's notion of matter "does not fit the classical philosophical definitions of metaphysical materialisms." This is not because de Man's concept of materiality is somehow redolent of a particular metaphysical materialism. Rather it is because the strange connivance with materiality is not an assured connivance in de Man. Or rather, it is only that, a connivance--which takes place only because de Man is indeed so assured about it. And this can't be overlooked merely because de Man's definition of matter, while it might not fit the classical definitions of materialisms, also does not fit "any more than" what we might call one of those definitions (but not necessarily merely one of them--Derrida is reserving the possibility for the sensible/intelligible opposition to be articulated otherwise) which it may or may not resemble. (Derrida is working too hard here to allow himself a way out from de Man's claims. We can't let that act in itself constitute what he also is doing, which is trying to inflect de Man differently, to deconstruct de Man's text.)

In short, we might say that, even if de Man's materialism isn't a materiality of the signifier, it is a materiality of the mark or the letter. And for Derrida, there should be no way that this mark is material, if we're going to be in any degree respectful of, or at least responsible for, the history of that word (which is how deconstruction proceeds--in a manner I might even call too historical for my taste, despite the old inane charge of ahistoricicity brought by the historicists).

For Derrida--just to go over some basics, and in too simple a language--the mark is empirical, and to the extent that the empirical is said to grasp the material (which is already not assured), the mark is therefore material. But any time this mark is taken up, it submits itself to a power of repetition, an iterability, which is ideal and which actually disturbs any notion of materiality or empiricity that we have attributed to it. For when the mark becomes capable of repetition--indeed in proportion to the increase in that power of repetition--it its history is marked in turn as already capable of repetition. And this double marking, this marking prior to it becoming a mark, means that it has already been a trace of itself, and has indeed been capable of repetition otherwise than how it is repeated in being taken up, in being assumed, we might say. As such, the mark remains, in its "materiality," always already participant in some ideality. And the trace itself, the marking of the mark, has itself only to be the possibility of materiality, and is as such "ideal." I put quotes around this because it is, qua ideal, also and only barely ideal, if I can say this: Gayatri Spivak uses the phrase "minimal idealization," and Judith Butler uses the word I've already employed, "ideality"--each of which I think describes this trace in this context of the material/ideal opposition quite well. That is, the trace is ideal to the extent that in this context) we respect the fact that it cannot be empirically determined, and thus is not material, unless of course we conceive empiricism differently--which means, of course, conceiving matter and materialism differently. This, by the way, is how that empiricism can indeed be "radical" in Derrida--in a sense in which I alluded to in my comments to a previous post, and which underlies what I say in another post on Derrida and science (though I critiqued this notion of "radicalism," which I think can be thought differently). Or rather this is how empiricism can knock away metaphysics (presupposing this latter concept is taken in a rough Heideggerian manner). The point is that this "ideal" aspect of the trace, this ideality, can also equally be described by Derrida as a "nonideal exteriority" (Positions, 65): it disrupts the notion of the ideal that we have. But what is important here is that while the trace is a nonideal exteriority, it is also, therefore, not at all material.

In no way then, can the trace, as what disturbs classical philosophical definitions of matter, be said itself to be material in any assured way (nor, then, can the mark--if these are marks of violence, we precisely have to think how far materiality as a concept allows us to think such violence). It has to remain, on some level, minimally ideal (such that it can always also be said to be nonideal). So this means that when we try to think a materialism differently (which, as I've tried to stress, is also is tied to the project of thinking empiricism differently--but which can also involve rethinking dialectical materialism, say, where the material is not so linked to the empirical but rather to the opposition to the ideal qua Geistlich), and come up with a notion of matter that is "without presence and without substance," this matter has ideality, because it is a trace, the possibility or impossibility, of such a matter. In de Man, however, because the trace has ideality we have to say that it is material ("the only word that comes to mind is that of a material vision," "Phenomenality and Materiality in Kant," 82). And in no way should we say that this constitutes a materialism: it is in fact, precisely an idealism which Derrida, by insisting on ideality (a nonideal exteriority which cannot yet be said to be material), resists.

One should also add that Derrida takes up these de Manian issues again in "Typewriter Ribbon: Limited Ink (2)," inflecting them through his attempt to think materiality and spectrality together--that is, to think materiality otherwise, in Specters of Marx. His remarks there (which give body precisely to what he means by "nonideal") are much more adequate and interesting. I would say though, that to pull Derrida's nonideality more in the direction of the ideal can't really be insisted upon enough in America, where the trace is too often thought of as material--given the influence of de Man. In fact, what I'd say is that everything in the following passage on "materialism" and "matter" needs to be replaced with "idealism" and the "ideal," to also understand the particular type of materiality Derrida is talking about and how it would resist a de Manian materiality:

If have not very often used the word "matter," it is not, as you know, because of some idealist or spiritualist kind of reservation. It is that the logic of the phase of overturning this concept has been too often reinvested with "logocentric" values, values associated with those of thing. [...] This is why I will not say that the concept of matter is in and of itself either metaphysical or nonmetaphysical. This depends on the work which it yields, and you know that I have unceasingly insisted, as concerns the nonideal exteriority of writing, the gram, the trace, the text, etc., upon the necessity of never separating them from work, a value itself to be rethought outside its Hegelian affiliation.
-Positions, 64-65

One night in Miami...

We are normally given some twenty or twenty-five minutes of the film, to get us interested in it; then four minutes of commercials, then about fifteen more minutes of the film; some commercials again; and so on to steadily decreasing lengths of the film, with commercials between them, or them between the commercials, since by this time it is assumed that we are interested and will watch the film to the end. Yet even this had not prepared me for the characteristic American sequence. One night in Miami, still dazed from a week on an Atlantic liner, I began watching a film and at first had some difficulty in adjusting to a much greater frequency of commercial "breaks." Yet this was a minor problem compared to what eventually happened. Two other films, which were due to be shown on the same channel on other nights, began to be inserted as trailers. A crime in San Francisco (the subject of the original film) began to operate in an extraordinary counterpoint not only with the deoderant and cereal commercials but with a romance in Paris and the eruption of a prehistoric monster who laid waste New York. [...] I can still not be sure what I took from that whole flow. I believe I registered some incidents as happening in the wrong film, and some characters in the commercials as involved in the film episodes, in what came to seem -- for all the occasional and bizarre disparities -- a single irresponsible flow of images and feelings.
-Raymond Williams, Television: Technology and Cultural Form, 91-2

What is interesting here is the tendency of the Marxist move, which we see I think most concretely in Williams (that is, it is registered most thoroughly in experience, not theory, or rather in the collapse of the thought-work distinction--a factwhich Adorno talks about), though it's present in Marx to a huge degree already (mostly as blistering theoretical irony--something I feel many readers don't fully understand). This is to rely not on clearer insight into the phenomenon in question, but to a fleeting glimpse captured while in an irrational state, which does not disclose that the totality is irrational, but transforms what, in any other state, you would call rational into an irrational moment (which presupposes an irrational totality indeed, but also determination, in the Hegelian sense). I think this has to be distinguished somewhat from (though no doubt it is related to) a hermeneutics of suspicion. I should give, in conclusion, Williams' general theory of "flow" or rather (since it is determined, through institutions) "planned flow," which tries to make sense of this experience:

What is being offered is not, in older terms, a programme of discrete units with particular insertions, but a planned flow, in which the true series is not the published sequence of programme items but this sequence transformed by the inclusion of another kind of sequence, to that these sequences together compose the real flow, the real "broadcasting." Increasingly, in both commercial and public-service television, a further sequence was added: trailers of programmes to be shown at some later time or on some later day.
-Television, 91