The Notion of Literature
Author(s): Tzvetan Todorov, Lynn Moss and Bruno Braunrot
Source: New Literary History , Winter, 2007, Vol. 38, No. 1, What Is Literature Now?
(Winter, 2007), pp. 1-12
Published by: The Johns Hopkins University Press
Stable URL: https://www.jstor.org/stable/20057986
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
The Johns Hopkins University Press is collaborating with JSTOR to digitize, preserve and
extend access to New Literary History
This content downloaded from
83.38.9.201 on Mon, 21 Jul 2025 17:45:11 UTC
All use subject to https://about.jstor.org/terms
The Notion of Literature
Tzvetan Todorov
Before launching into the awesome question of the nature of
literature, let us begin, as a precautionary measure, by examining
not literature itself but rather the kind of discourse which, like
this very study, takes literature as its object. The difference will be one of
approach rather than of objective; but who can say whether the course
of the inquiry is not of greater interest than its final results?
We must first cast a doubt upon the legitimacy of the very notion of
literature; neither the mere existence of the term, nor the fact that a whole
university system is based upon it, can of itself justify its acceptance.
The first grounds for doubt are of an empirical nature. No complete
history of the word and its equivalents, in all languages and throughout
time, has yet been undertaken; yet even a superficial inquiry into the
question suffices to convince us that the term has not always been with
us. In the European languages the word literature, in its present usage,
is quite recent: it barely dates back to the nineteenth century. Could
it be that we are dealing with an historical and not at all an "eternal"
phenomenon? Moreover, many languages (those of Africa, for example)
still have no generic term to designate literature as a whole; and while
L?vy-Bruhl would have sought to explain this absence by the so-called
primitive nature of these languages allegedly incapable of abstraction
and hence devoid of any words designating the general rather than
the specific, the time when we could accept such an explanation has
long passed. Finally, we must also take into account the diversification
of literature in our own countries; who would dare decide today what
is literature and what is not, given the irreducible diversity of all the
written works which, from infinitely different perspectives, tend to be
regarded as literature?
This argument is not decisive: a notion may be legitimate even though
no corresponding word may as yet exist to designate it; it does, however,
create a first element of doubt as to the "natural" character of literature.
Nor is a theoretical examination of the problem more reassuring. Whence
* This essay was originally published in "What Is Literature?" New Literary History 5, no.
1 (1973).
New Literary History, 2007, 38: 1-12
This content downloaded from
83.38.9.201 on Mon, 21 Jul 2025 17:45:11 UTC
All use subject to https://about.jstor.org/terms
2 NEW literary history
do we derive the certainty that an entity such as literature really ex
From experience: we meet literary works in school, then in college
find them in certain specialized stores; references to "literary auth
crop up constantly in our everyday conversations. That an entity
"literature" does function on an intersubjective and social level see
indeed unquestionable. Agreed. But what does this prove? That
larger system?a society, a civilization?there exists an identifiab
ment referred to as literature. But does it prove that all the indivi
works grouped under this heading partake of a common nature wh
we can identify with equal justification?
Let us call "functional" our first definition of this entity?the d?fin
which identifies it in terms of what it "does" as an element in a la
system; and "structural" our second, whereby we seek to test wheth
the individual works collectively regarded as literature in the funct
meaning of the word partake of the same characteristics. The distin
between the functional and the structural points of view should be
ously kept in mind, even though to pass from one to the other is perfec
permissible. In order to illustrate this distinction, let us take the exa
of advertising: its precise function within our society is undeniably c
but what of its structural identity? It can express itself though the
and auditory, as well as other media; it may or may not have a dur
in time; it may be continuous or discontinuous; it may use techniqu
varied as direct inducement, description, allusion, antiphrasis, and s
The unquestionable functional entity?assuming for the moment
it is indeed unquestionable?does not necessarily have a correspon
structural entity. The one need not necessarily imply the other, alth
the affinities between them can be easily observed. The difference is mo
in the point of view than in the thing itself: if literature (or advertisin
discovered to be a structural notion, then the function of its consti
elements will have to be determined; conversely, the functional no
"advertising" can be seen to belong to a structure which is, let us
that of society. Structure is made up of functions, and functions cr
structure; but since it is the point of view which determines the ob
of knowledge, the difference remains an essential one.
Thus the existence of a functional entity "literature" in no way im
that of a structural entity (although it does make us want to find
whether such is not the case). Now functional definitions of literat
(in terms of what it does rather than what it is) are very numerou
though they are not necessarily sociological: when a metaphysician
Heidegger considers the essence of poetry, he too arrives at a functi
definition. To say that "art is the concretization of truth," or that "p
is the creation of being through words," is to say what the aim of
should be, without defining the specific mechanisms which would
This content downloaded from
83.38.9.201 on Mon, 21 Jul 2025 17:45:11 UTC
All use subject to https://about.jstor.org/terms
THE NOTION OF LITERATURE 3
it to fulfill that aim. The function is here of an ontological nature, but
a function nonetheless. Heidegger himself affirms that there is no struc
tural entity corresponding to the functional one when he says elsewhere
that his search is "concerned only with great art." This definition does
not contain any internal criteria which would allow us to identify a work
of art (or literature). It is a mere statement of what a certain kind of
art?the best?ought to do.
It is possible that literature is only a functional entity. But we shall not
pursue this line of thought. We shall assume instead?at the risk of be
ing proved wrong in the end?that it also has a structural identity, and
attempt to find out what it is. Many other optimists have preceded us in
this search: their suggested answers will serve as a point of departure.
Without concern for historical detail, we shall attempt to examine the
two kinds of solution most frequently proposed.
From antiquity to the mid-eighteenth century, roughly speaking, the
same definition recurs, whether explicitly or not, in the writings of the
theoreticians of Western art. Upon close examination, this definition is
seen to consist of two distinct elements. Generically, art is an imitation
which varies according to the medium used; literature is imitation through
language, painting through visual images. Specifically, it is not just any
imitation; for what is imitated is not real but fictional, and need not have
existed. Literature is fiction: this is its first structural definition.
This definition was formulated over several centuries and expressed in
very different terms. It was probably this property of literature which led
Aristotle to note that "poetry is more concerned with the general, history
with the particular" (Poetics, 1451b; the remark has another meaning as
well): a literary sentence does not refer to specific actions, which alone
could occur in reality. Later generations will view literature as essentially
deceptive, false; Northrop Frye has drawn our attention to the ambiguity
of terms such as "fable," "fiction," and "myth," which apply with equal
ease to "literature" and to "falsehood." But this is misleading: a literary
statement is no more "false" than it is "true." The earliest modern logi
cians (Frege, for example) have already pointed out that a literary text
cannot be submitted to the test of truth. It is neither true nor false but
simply fictional?an observation which today has become a commonplace
of literary criticism.
But is such a definition really satisfactory? Aren't we guilty of substi
tuting for a true definition what is merely one of the consequences of
literature? There is nothing to prevent a story relating a real event from
being viewed as literature; no changes in composition are needed, only
the determination to disregard its truth and read it "as if it were litera
ture. Any text whatever can be given a "literary" reading: the question
of truth will not arise precisely because the text is literature, and not the
other way around.
This content downloaded from
83.38.9.201 on Mon, 21 Jul 2025 17:45:11 UTC
All use subject to https://about.jstor.org/terms
4 NEW LITERARY HISTORY
Obviously, what is being offered indirectly here is one of the properties
of literature rather than its definition. But can this property be observed
in every literary text? Is it by chance that we freely apply the word "fiction"
to some types of literature (novels, short stories, plays) but find it much
more difficult to apply the same word to poetry?if indeed we ever do?
It is tempting to suggest that just as the novel is neither true nor false
even though it may describe an event, so poetry in turn is neither fiction
nor nonfiction. The question does not even arise, inasmuch as poetry
does not relate any events, but is very often limited to the formulation
of a meditation or an impression. The specific term "fiction" is not appli
cable to poetry because the generic term "imitation" can remain relevant
only by losing any precise meaning it may have; poetry is often not at
all representational of an external reality, but sufficient unto itself. The
question becomes even more difficult when we consider the so-called
minor genres which are nevertheless present in all the "literatures" of
the world: prayers, exhortations, proverbs, riddles, nursery rhymes (each
involving of course its own special difficulties). Shall we claim that they
too "imitate," or shall we separate them completely from the body of
what we call "literature"?
If all that is usually regarded as literature is not necessarily fictional,
conversely, all that is fictional is not automatically literature. In Freud's
"case histories," for example, the question whether all the misadventures
of "little Hans" or the "wolf man" are true or not is irrelevant: their
status is exactly that of fiction: all one is entitled to say of them is that
they support or contradict Freud's thesis. Or, to take a very different
example: should all myths be viewed as literature, even if their fictional
character is unmistakable?
We are certainly not the first to criticize the notion of imitation in
literature or in art. European classicism continually attempted to modify
the concept in order to be able to retain it. For the term had to be
given a very general meaning if it were to remain applicable to all the
activities with which it had been associated. But this in turn rendered
it applicable to other things as well, so that it became necessary to add
a further specification: imitation must be "artistic," which amounts to
using the term to be defined as part of its definition. At some point in
the eighteenth century the trend reversed itself; instead of yet another
rearrangement of the original definition, an entirely new one was pro
posed. Nothing illustrates the transition better than the titles of two
books which mark the limits of two distinct periods. In 1746, a work on
aesthetics appeared which sums up the opinion of the time: The Arts Re
duced to a Single Principle, by the Abb? Batteux; the principle in question
is the imitation of beautiful nature ("la belle nature"). In 1785, another
work followed: An Attempt to Unite the Arts and Letters through the Concept
This content downloaded from
83.38.9.201 on Mon, 21 Jul 2025 17:45:11 UTC
All use subject to https://about.jstor.org/terms
THE NOTION OF LITERATURE 5
of Self-Sufficiency, by Karl Philipp Moritz. The arts are again united, but
this time in the name of beauty, understood as self-sufficiency ( "in sich
selbst Vollendetes").
The second major definition of literature will thus be based upon
the concept of beauty; it will become more important to "please" than
to "instruct." Near the end of the eighteenth century, the definition of
beauty will center around the belief in the intransitive (as opposed to the
instrumental) nature of a work of art. After having been equated with
the useful, beauty is now defined by its nonutilitarian character. "True
beauty requires that a thing signify nothing but itself, that it be a unity
complete in itself," writes Moritz. And art is defined in terms of beauty:
"If a work of art existed only to point to something other than itself, it
would become nothing more than an accessory; whereas in the case of
beauty, it alone must always be the principal thing." Painting is images
perceived for themselves and not meant to serve any outside purpose,
music is sounds appreciated for themselves. Literature is a noninstru
mental language whose value resides in itself alone, or as Novalis has
said, "an expression for the sake of expression."
Elaborated upon by the German Romanticists, this notion was to
dominate all the Symbolist and post-Symbolist movements in Europe.
What is more, it was to become the basis for the first modern attempts
to create a science of literature. Be it Russian Formalism or American
New Criticism, the point of departure is always the same. The function of
poetry is essentially to emphasize the "message" itself. Even today this is
the dominant definition, although its formulation may vary somewhat.
To be exact, such a definition of literature does not deserve to be
called structural: we are told what poetry ought to achieve, and not
how it proceeds to do so. But soon the functional perspective was to
be complemented by a structural point of view: more than any other
of its aspects, it is the systematic character of a work which allows us to
perceive it in itself. Such was already Diderot's definition of beauty; later
the term "beauty" was to be replaced by "form," which in turn was to
give way to "structure." Formalist studies of literature will have the merit
of being studies of a literary system (of the whole of literature or of the
individual work), thus creating the new science of poetics. Literature,
then, is a system, a systematic language which draws attention to itself,
which becomes autotelic. This is its second structural definition.
Let us examine this hypothesis in its turn. Is literary language alone in
being systematic? Of course not. A rigorous organization can be found
not only in areas usually associated with literature?some, such as ad
vertising, even use the same techniques of rhyme, polysemy, etc.?but
also in some regions far removed from its province. Who could deny
that a judicial or political discourse is organized and obeys certain strict
This content downloaded from
83.38.9.201 on Mon, 21 Jul 2025 17:45:11 UTC
All use subject to https://about.jstor.org/terms
6 NEW LITERARY HISTORY
rules? It is not by chance that until the Renaissance, and above all in
Greek and Latin antiquity, Rhetoric was set alongside of Poetics and as
signed the task of codifying those rules which were concerned with all
extraliterary discourse. One could even go further and question the very
validity of a notion such as that of "the system of a literary work," given
the ease with which any such system could be contrived. Language has
only a limited number of phonemes, and even fewer distinctive features;
nor are the grammatical categories in each paradigm very numerous;
thus repetition, far from being difficult, is inevitable. Saussure is known
to have formulated a hypothesis concerning Latin poetry, according
to which the poet allegedly concealed within the body of the work the
name of the person for whom, or about whom, the poem was written. If
the hypothesis leads nowhere, it is from excess rather than from lack of
proof: any name can be found in a poem of sufficient length. Besides,
why limit the theory to poetry? "This practice was second nature to all
educated Romans, who put everything, no matter how insignificant, into
writing." And why the Romans? Saussure will go so far as to discover
the name of Eton in a Latin text used at that college in the nineteenth
century; unfortunately for him the author was a seventeenth-century
scholar from King's College, Cambridge, and the work was not used at
Eton until a hundred years later!
A system so easily discovered is not a real system. Let us now consider
the complementary test: is every literary text so systematic that it may be
said to be autotelic, intransitive, opaque? The sense of such a statement
is relatively clear when it is applied to poetry which, as Moritz would
have said, is a self-sufficient object; but what of the novel? Not that we
wish to reduce the novel to a mere "slice of life," devoid of conventions
and hence of system; but the presence of this system does not render
the language of the novel in any way "opaque."
On the contrary, such language serves (in the classical European novel
at least) to represent objects, events, actions, characters. Nor can it be
said that the essence of a novel is not in its language but in its narrative
technique; we are only amused by Shklovsky's remark that the sole pur
pose of the philosophical discussions in Dostoevsky's novels is to slow the
pace of the narrative. Perhaps what is opaque in this case is the world
represented; but could not such a conception of opacity (of in transi tive
ness, of autotelism) apply just as well to any everyday conversation?
Many attempts have been made in our time to bring together the two
definitions of literature. But since neither of them, when taken alone,
is really satisfactory, there is little advantage in combining them; in or
der to remedy their weakness, they should be fully articulated and not
merely connected, or worse still, treated as if they were not different. A
This content downloaded from
83.38.9.201 on Mon, 21 Jul 2025 17:45:11 UTC
All use subject to https://about.jstor.org/terms
THE NOTION OF LITERATURE 7
few examples will show, however, that such is unfortunately most often
the case.
In the chapter of Wellek and Warren's Theory of Literature dealing
with "the nature of literature," Ren? Wellek seeks to define "literature"
by distinguishing the particular use made of language in literature, by
contrasting it with the two other main uses: the everyday and the scien
tific. Unlike the scientific, the literary use of language is "connotative,"
that is, rich in associations and ambiguities; it is opaque (whereas in the
scientific use the sign is "transparent; that is, without drawing attention
to itself, it directs us unequivocally to its referent") ; and it is multifunc
tional: not only referential but also expressive and pragmatic (conative).
Unlike the language of everyday use, it is systematic ("poetic language
organizes, tightens the resources of everyday language"); and autotelic,
in that its sole justification is within itself.
So far Wellek seems to be a partisan of our second definition of
literature. Emphasis placed on any kind of function (be it referential,
expressive, or pragmatic) has the effect of drawing us away from lit
erature, where the text derives its value from itself (this is what will be
called the aesthetic function, after a theory propounded by Jakobson and
Mukarovsky in the 1930s). The structural consequences of this functional
approach are a trend towards systematization and an emphasis on all the
symbolic resources of the linguistic sign.
There follows, however, another distinction, apparently expanding on
the opposition between the everyday and the literary use of language.
"But the nature of literature emerges most clearly under the referential
aspect," Wellek states, for in the most "literary" works, "the reference
is to a world of fiction, of imagination. The statements in a novel, in a
poem, or in a drama are not literally true; they are not logical proposi
tions." This, he concludes, is the "distinguishing trait of literature": that
is, its "fictionality. "
In other words, we have passed, even without realizing it, from the
second definition of literature to the first. The literary use of language is
no longer defined by its systematic (and consequently autotelic) charac
ter, but by its fictionality, by statements which are neither true nor false.
Are the two definitions then the same? One would expect such a thesis
to be at least explicitly formulated (to say nothing of proof). Wellek's
conclusion, namely, that all these terms (systematic organization, recogni
tion of the sign, and fiction) are necessary to characterize a work of art,
brings us no closer to a solution. For the question which we are raising
is precisely this: what are the relationships among these terms?
The situation is much the same with Northrop Frye, who deals with the
question in the chapter entitled "Literal and Descriptive Phases: Symbol
as Motif and as Sign" in his Anatomy of Criticism. He too begins by making
This content downloaded from
83.38.9.201 on Mon, 21 Jul 2025 17:45:11 UTC
All use subject to https://about.jstor.org/terms
8 NEW LITERARY HISTORY
a distinction between literary and nonliterary use of language (thereby
combining Wellek's "scientific" and "everyday" uses into one category).
The implicit opposition is between an outward and an inward orienta
tion (toward what lies beyond the signs and toward the signs themselves,
or toward other signs, respectively). The oppositions between centrifu
gal and centripetal, between descriptive and literal phases, between
sign-symbols and motif-symbols are all related to the first distinction. It
is inward direction which characterizes the literary use of language. In
passing, it should be noted that Frye is no more willing than Wellek to
affirm the exclusive presence of this orientation in literature; he merely
affirms its predominance.
Once again we are faced with a version of our second definition of
literature and once again, before knowing it, we have slid back towards
the first. Frye writes: "In all literary verbal structures the final direction
of meaning is inward. In literature, the standards of outward meaning
are secondary, for literary works do not pretend to describe or assert, and
are hence not true, not false. ... In literature, questions of fact or truth
are subordinated to the primary literary aim of producing a structure of
words for its own sake, and the sign-values of symbols are subordinated
to their importance as a structure of interconnected motifs." In this last
sentence, it is no longer transparence but nonfictionality (that is, adher
ence to a true-false system) which is opposed to opacity.
It is no doubt the ambivalence of the word "inward" which accounts
for this passage from one definition to the other, for the term is present
in both, being in turn synonymous with "opaque" as well as "fictional."
The literary use of language is "inward" both because it emphasizes the
signs themselves and because the reality evoked by the signs is fictional.
But perhaps beyond mere polysemy (and thus beyond the elementary
confusion) there is a mutual implication between the two meanings of
the word "inward": could it be that all "fiction" is "opaque," and that
"opacity" is "fictional"? Frye seems to suggest precisely this when he
asserts that if an historical work were to comply with the principles of
symmetry (indicative of a system, and thus of an autotelic structure), it
would enter thereby the realm of literature, and hence of fiction. But
do the two meanings of "inward" really imply each other? An examina
tion of this question will perhaps help to elucidate the nature of the
relationship between the two definitions of literature.
Supposing that a history book does follow the rules of symmetry (and
is thus literature according to our second definition), does it thereby also
become fictional (and thus literature by the first definition)? It clearly
does not. It may become a bad history book; but the change is from
"true" to "false," and not from "true-false" to "fictional." Similarly, a po
litical discourse can be highly systematic without automatically becoming
This content downloaded from
83.38.9.201 on Mon, 21 Jul 2025 17:45:11 UTC
All use subject to https://about.jstor.org/terms
THE NOTION OF LITERATURE 9
fictional. In terms of systematization, is there a radical difference in the
account of a real journey and that of an imaginary one?even though
one is fictional and the other not? Neither the tendency toward system
atization nor the emphasis on inward direction is sufficient to render
a text fictional. Thus at least one of the consequences of the would-be
mutual implication proves to be untenable.
What about the other? Does fictionality necessarily imply a contextual
orientation? Everything depends on the meaning given to the latter ex
pression. Some of Frye's remarks seem to indicate that it is a matter of
simple recurrence, or of syntagmatic rather than paradigmatic orienta
tion. If this be the case, it goes without saying that there are texts which
are clearly devoid of such properties: a narrative can be governed by the
logic of succession and causality alone (although examples of this type
are admittedly rare). If we understand the term in its broadest sense of
the "presence of some sort of organization," then it is equally clear that
all fictional texts possess this "inward orientation"; but what text?of
whatever nature?does not? The second implication is thus no more
rigorously true than the first, and there is no justification for postulat
ing that the two meanings of the word "inward" are in fact one and the
same. Once again the two distinctions (and the two definitions) have
been merely combined without being fully articulated.
All that can be said is that each definition can account for a great
number of works usually called literary, but by no means for all; and
that they are admittedly linked by mutual affinities, but not by mutual
implication. The discussion has not proceeded beyond the stage of
vagueness and imprecision.
The relative failure of our investigation might perhaps be explained
by the nature of the question itself. We have constantly asked ourselves:
how can we distinguish what is literature from what is not? What is the
difference between the literary and the nonliterary use of language?
But to ask such questions about the notion of literature is to assume the
existence of another coherent notion, that of "nonliterature." Perhaps
the time has come to examine this notion as well.
Whether one speaks of descriptive writing (Frye) or of everyday usage
(Wellek), of practical or normal language, a unity is postulated which
upon examination turns out to be highly problematical. It seems obvi
ous that this "usage," which includes jokes as well as practical conversa
tion, the ritualistic language of administration and law as well as that
of the journalist or the politician, scientific as well as philosophical and
religious writing, is not a single entity at all. No one knows exactly how
many types of discourse there are, but it is easy to see that there are
more than one.
This content downloaded from
83.38.9.201 on Mon, 21 Jul 2025 17:45:11 UTC
All use subject to https://about.jstor.org/terms
10 NEW LITERARY HISTORY
Another notion, generic in relation to the notion of literature, must
at this point be introduced: that of discourse. The latter is the structural
correlative of the functional concept of "use" (of a language). It is neces
sary to introduce it 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 and the
exact formulation of a specific utterance there is a gulf of indeterminacy.
This 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 language,
but less restricted than those of a specific speech act.
A particular type of discourse is in turn defined by the list of rules
which it must obey. The sonnet, for example, is characterized by extra
limitations on its meter and rhyme. Scientific discourse in principle ex
cludes any reference to the first or second person of the verb, as well as
to any but the present tense. Jokes have semantic rules lacking in other
discourse, while their metric construction is determined in the course
of the individual speech act. Certain discursive rules consist, paradoxi
cally, of abolishing various rules of language; Samuel Levin in the United
States and Jean Cohen in France have shown how certain grammatical
or semantic rules are put aside in modern poetry. But from the point
of view of the structure of a discourse, rules are always added, never
subtracted. The proof of this is that even in "deviant" poetic utterances
the broken linguistic rule is easily reconstructed; rather than being
abolished, it was merely contradicted by a new rule. In literary studies
the rules of discourse are discussed under the heading of "genres" (or
sometimes "styles," "modes," etc.).
If we admit the existence of various types of discourse we must refor
mulate our question on literary specificity in the following terms: are
there rules which apply to all forms of literature (intuitively identified as
such), and only to those forms? The only possible answer, it seems to me,
is a negative one. Numerous examples have already been noted of the
occurrence of "literary" characteristics outside literature (from puns and
nursery rhymes, journalism and travelogues to philosophical meditations).
It has become equally obvious that there is no common denominator for
all "literary" productions, unless it be the use of language.
The situation changes radically if we turn away from "literature" itself
and focus on its subcategories. There is no difficulty in defining the rules
of certain types of discourse (this has always been the function of the
This content downloaded from
83.38.9.201 on Mon, 21 Jul 2025 17:45:11 UTC
All use subject to https://about.jstor.org/terms
THE NOTION OF LITERATURE 11
various Artes Poeticae, admittedly with much confusion between descrip
tion and prescription). In other types the formulation of rules is more
difficult, although our sense of "discursive competence" convinces us
that they do exist. In fact we have already seen that the first definition of
literature applies particularly well to narrative prose, and the second to
poetry. The origin of these utterly independent definitions can perhaps be
found in the type of literature which was considered in their formation.
The first is based on narrative (Aristotle discusses epic and tragedy, not
poetry), the second on poetry (as is apparent from Jakobson's analyses
of particular poems) : in each case one of the two major literary genres
has been defined as if it were the whole of literature.
The rules of the so-called nonliterary types of discourse can be identi
fied in much the same way. I thus propose the following hypothesis: from
a structural point of view, each type of discourse usually referred to as
literary has nonliterary relatives which resemble it more than do other
types of literary discourse. For example, a certain type of lyric poetry
has more rules in common with prayer than with a historical novel of
the War and Peace variety. Thus the opposition between literature and
nonliterature is replaced by a typology of the various types of discourse.
Frye can be quoted once again, this time without reservation: "our liter
ary universe has expanded into a verbal universe."
The results of this inquiry might at first appear negative, since it es
sentially denies the legitimacy of a structural notion of "literature," and
contests the existence of a homogeneous "literary discourse." Whether
or not the functional notion is legitimate, the structural notion definitely
is not. But the result is only seemingly negative, since instead of the
simple notion of literature we now have a number of different types of
discourse, each equally deserving of attention. If the choice of our object
of study is not dictated by purely ideological reasons (which would then
have to be spelled out), we no longer have the right to limit ourselves to
purely literary subspecies, even if we are employed by the "Department
of Literature" (be it French, English, or Russian). A coherent field of
study demanding recognition is at present hopelessly fragmented among
semanticists and litt?rateurs, socio- and ethnolinguists, linguistic philoso
phers, and psychologists.
An explanation might be found, by the same token, for the dominance
of these two definitions, rather than any other, throughout the history of
literary theory. Viewed in their most general sense, which alone confers
upon them their validity, they become affirmations of the significant
nature of literary texts and of their systematic organization. But isn't
the definition of any discourse?at once system and signification? At
tempting to define literature, the theoretician defines instead a logically
superior notion, the "genus proximum." These are indeed its two essen
This content downloaded from
83.38.9.201 on Mon, 21 Jul 2025 17:45:11 UTC
All use subject to https://about.jstor.org/terms
12 NEW LITERARY HISTORY
tial and complementary aspects, whatever they are called: pleasure and
instruction, beauty and truth, gratuitous play and imitation, syntax and
semantics (although the variation in terminology is in no way unimport
ant: although they refer to the same thing, the various terms signify it
in different ways). What the theoreticians have failed to do, however, is
to indicate the "specific difference" which characterizes literature within
the "genus proximum." Could it be that no such difference is in any way
perceptible? In other words, that literature does not exist?
Centre National de la Recherche Scientifique?Paris
Translated by Lynn Moss and Bruno Braunrot
This content downloaded from
83.38.9.201 on Mon, 21 Jul 2025 17:45:11 UTC
All use subject to https://about.jstor.org/terms
