Levels of Meaning in Literature

Author(s): Northrop Frye

Source: The Kenyon Review, Vol. 12, No. 2 (Spring, 1950), pp. 246-262
Published by: Kenyon College
Stable URL: http://www.jstor.org/stable/4333140 .
Accessed: 01/07/2011 11:14

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at .
http://www.jstor.org/action/showPublisher?publisherCode=kenyon. .

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

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 support@jstor.org.

Kenyon College is collaborating with JSTOR to digitize, preserve and extend access to The Kenyon Review.

http://www.jstor.org
Northrop Frye

LEVELSOF MEANING IN LITERATURE

THE longer one has been familiar with a great work of liter-
ature, the more one's understandingof it grows. It would be
hard to formulate a more elementary principle of literary ex-
perience. Its plain implication, that literature has different levels
of meaning, was made the basis of a systematic development of
criticism in the Middle Ages, and a precise scheme of four levels
of meaning-the literal, the allegorical, the tropological or moral,
and the anagogic-was worked out and adopted by many great
medieval writers, notably Dante. Modern criticism has not only
ignored this, but seems to regard the problem of meaning in
literature as merely an offshoot of the corresponding semantic
problem in current philosophy. In offering a few suggestions
about the possibility of a modern restatement of the medieval
theory, I propose to by-pass the philosophical questions involved,
on the ground that the obvious place to start looking for a theory
of literary meaning is in literature.

The First Level.

Let us start with the word "symbol,"confining ourselves to verbal
symbols. A symbol ordinarily implies at least two things, A, the
symbol proper, and B, the thing representedor symbolized by it.
With verbal symbols particularly, however, there seems to be a
locus or path of signification which passes through a number of
B's. The verbal symbol "cat"is a group of black marks on a page
representing a noise with the mouth which represents an idea
called up from experience of an animal that says meow. A verbal
NORTHROP
FRYE 247
symbol may be of any size: a word, a letter, a phrase, a sentence,
or even larger word-groups may be symbols. The total verbal
structure,the poem or the book, may be regarded also as a single
symbol-complex.But whenever we attempt to answer the question,
what does this symbol or symbol-complex symbolize? we find
ourselves travelling a centrifugal path from the verbal structureto
a realm of experience outside it.
We find this more difficult at some times than at others. It is
easy enough to say, up to a point, what "cat" symbolizes, even
what each of the three letters in it symbolizes. It is harder to say
what the word "of" symbolizes, or the final letter of the word
"lamb." Here we have to enlarge our unit of symbolism to give
an intelligible answer, and in the process we become aware of
another direction of meaning, a direction not centrifugal this
time but centripetal, not running outward into experience but
inward into the total meaning of the verbal structure.The word
"of," we say, has not, like a noun, a direct one-to-one correspond-
ence with a thing symbolized: one has first to relate it to other
words. But it is clear that the same syntactic aspect of meaning
is relevant even to nouns.
If a writer uses the word "cat," that verbal symbol represents,
in addition to its centrifugal locus of meaning, a portion of the
author's total intention in putting it there. Its meaning in the
verbal structure cannot be understood without relating it to the
structure as a whole. Nor can we ultimately interpose even the
"author's intention" between the verbal symbol and the verbal
structure, for the author's intention ceases to exist as a separate
factor as soon as his verbal structureis fixed. Centripetally, then,
the verbal symbol does not represent anything except its own
place in the verbal structure. We said that the latter may also
be regarded as a single symbol-complex.But from the centripetal
point of view, in which the unit of symbolism is to be considered
only in terms of its place in the total verbal structure, the total
verbal structureitself does not "mean" anything except itself.
248 KENYON
REVIEW
Most uses of the term "literal," whether medieval or modern,
fail to make this distinction between the syntactic and the repre-
sentative relations of a unit of symbolism. I do not understand
the common assertion that the verbal symbol "cat" "means liter-
ally" a cat, the animal that says meow. It is surelyobvious that the
verbal symbol stands in a descriptive and representativerelation-
ship to actual cats. Dante says, in commentingon the verse in the
Psalms, "When Israel came out of Egypt": "For should we con-
sider the letter only, the exit of the children of Israel from
Egypt in the time of Moses is what is signified to us." But an
historicalevent cannot be literally anything but an historicalevent;
a prose narrative describing it cannot be literally anything but a
prose narrative. In taking his example from the Bible, Dante per-
haps felt the necessity of making first of all a respectful genuflec-
tion to theological rationalism:in any case there is no such pseudo-
literal basis to his own Commedia.
The literal level of meaning, though it takes precedenceover
all other meanings, lies outside the province of criticism. Under-
standing a verbal structureliterally is the incommunicableact of
total apprehension which precedes criticism. The preliminary
effort to unite the symbols in a verbal structure,and the Gestalt
perception of the unity of the structure which results, are the
closest we can come to describingthe literal level. Everygenuine
response to art, whether critically formulated or not, must begin
in the same way, in a complete surrenderof the mind and senses
to the impact of the work of art as a whole. This occupies the
same place in criticismthat observation,the direct exposure of the
mind to nature, has in the scientific method. "Every poem must
necessarily be a perfect unity," says Blake: this, as the wording
implies, is not a judgmentof value on existing poems, but a defini-
tion of the hypothesiswhich every readermust adopt in first trying
to comprehend even the most chaotic poem ever written. In the
theory outlined at the end of Joyce's Portrait,the first of the three
attributesof beauty, integritas, correspondsto our literal level.
 FRYE
NORTHROP 249
The Second Level.
When we say that such a complex verbal structurehas "mean-
ing," we usually refer to the vast disordered tangle of centrifugal
meanings running in all directions from its words and phrases. By
the time we have apprehendedthe integritas or literal significance,
we have recapitulatedour whole education in centrifugal meaning,
back to our earliest attempts to read. But after we have under-
stood a verbal structure literally, we have then to relate it as a
whole to the body of data which it represents.
This process introduces the conception "literary" to our dis-
cussion of verbal structures. Verbal structureswhich are not "lit-
erary" are primarily descriptions of facts or truths external to
themselves. What interests, say, a scientist or an historian about
words is their accuracyin reproducingscientific or historical data.
But one of the most familiar phenomenain the "literary"group is
the absence of this controlling aim of descriptive accuracy. We
should prefer to feel that an historical dramatist was capable of
reading and using his sourcesaccurately,and would not alter them
without good reason. But that such good reasons may exist in
literature is not denied by anyone, although they seem to exist
only there. Literature,poetryespecially,may always be recognized
by the negative test of the possibility of departing from facts.
Hence the words denoting literarystructure,"fable," "myth," and
"fiction,"have acquireda secondarysense of untruth,like the Nor-
wegian word for poet, digter, which also means liar.
Sir Philip Sidney remarkedthat "the poet never affirmeth,"and
therefore cannot be said to lie. Literaturepresents, not an affirm-
ation or repudiation of facts, but a series of hypothetical pos-
sibilities. The appearanceof a ghost in Hamzletdoes not owe its
dramatic appeal to the question whether ghosts exist or not, or
whether Shakespeare or his audience thought they did. Shake-
speare'sonly postulate is, "let there be a ghost in Hamlet." In this
the poet resembles the mathematicianrather than his verbal col-
leagues in history and science.
250 KENYON
REVIEW
The relation of literatureto factual verbal structureshas to be
established from within one of the latter. Literaturemust be ap-
proached centrifugally, from the outside, if we are to get any fact-
ual significanceout of it. Thus an historiancould learn much from
a realistic novel written in the period he is studying, if he knows
how to allow for its hypotheticalstructure. It would not do much
violence to customarylanguage to use the term "allegorical" for
this whole descriptive level of meaning, and say, for instance,
that a realistic novel was an allegory of the life of its time. In
literary criticism itself, the second level of meaning allows the
critic to employ himself in that routine but indispensable activity
which the master-paintersof the Renaissanceassigned to their ap-
prentices: the activity of filling in background. One begins talk-
ing about "Lycidas,"for instance, by itemizing all the things that
"Lycidas"illustrates in the non-literaryverbal world: English his-
tory in 1637, the Church and Milton's view of it, the position of
Milton as a young poet planning an epic and a political career,the
literaryconventionof the pastoralelegy, Christianteachingson the
subjectof death and resurrection,and so on. It would be quite pos-
sible to spend a whole criticallife in this allegorical limbo of back-
ground, without ever getting to the poem at all, or even feeling
the need of doing so.
We do not ordinarily use the word allegory as we have just
used it, however; we usually restrictit to the one exception to our
rule that the relation of literatureto fact must be establishedfrom
outside literature. A writer is being explicitly allegorical when he
himself indicates a continuous relationship of his central hypo-
thetical structureto a set of external facts, or what he assumes to
be facts. This continuous counterpoint between the saying and
the centrifugal meaning is called allegory only when the relation
is direct. If the relation is one of contrast,we call it irony. The
purpose of allegory is to emphasize the connection of poetry with
affirmativetruth; the purpose of irony is to emphasize its with-
drawal from it.
 FRYE
NORTHROP 251
What position, then, does literatureoccupy in relation to fact-
ual verbal structures? We may get a hint here from another argu-
ment of Sidney, which follows a general Aristotelian line. Sidney
suggests that poetry is a kind of synthesis of history and philoso-
phy. History gives the example of the hero without the precept;
philosophy the precept without the example, and poetry gives us
the poetic image of the hero which combines the two. Or, as we
may say, literature, being hypothetical, unites the temporal event
with the idea in conceptual space. On one side, it develops a nar-
rative interest which borders on history; on the other, a discursive
interest which borders on philosophy, and in between them is its
central interest of imagery.
We may thus distinguish three main rhythmsof literature and
three main areas of it, one in which narrativecontrols the rhythm,
one in which a discursiveinterest controls it, and a central area in
which the image controls it. This central area is the area of poetry;
the parietal ones belong to prose, which is used for both hypotheti-
cal and descriptive purposes. If we look at the word image
closely, we shall see that it really means symbol in its centripetal
aspect, so that imagery in this sense is figuration, the arranging
and patterning of verbal symbols. In medieval and Renaissance
times this formed part of the study of rhetoric, and so we may at-
tempt a tentative definition of poetry as the form of verbal ex-
pression which is organized on rhetoricalprinciples. Of these the
chief is of course recurrentmetre; the auxiliary principles, alliter-
ation, rhyme, quantity or parallelism, are also rhetoricalschemata.
A commonerword for rhetoricalfiguration is style, but this word
is too often used merely as a metaphor for the inscrutablemystery
of genius.
It would be surprising to find any sharp boundary separating
narrativeprose from history or discursive prose from philosophy.
We can only say that whatever is clearly hypothetical is clearly
literary. On the discursive side, the question arises whether the
whole section of philosophy called "metaphysics"should be an-
252 REVIEW
KENYON
nexed to literature. The logical positivists claim that metaphysical
systemsare not descriptiveof anything,but are hypotheticalverbal
structuresdepending for their integrity on propositions which are
neither true nor false. A literary critic would certainly lose no
respect for metaphysicsif all this were true. And even if it is not,
one has only to refer to Plato's dialogues to show how useless any
patented formula of classificationwould be.
In any case we may isolate prose fiction as the form of prose
which is organized on narrativeprinciples, discursive prose being
based ratheron the propositionas its rhythmicunit. One may note
in the history of discursivewriting a recurringeffort to isolate the
propositional rhythm. Hence we have the aphorisms of Bacon,
the quasi-Euclideanform of Spinoza's Ethics, the thesis form of
scholasticism,and, more recently,the tabulatedaphorismsof Witt-
genstein. As a rule such attempts defeat their original purpose by
giving the readerthe impressionof a rhetoricaldevice. Neverthe-
less, the organizing rhythm of discursivewriting is logical rather
than rhetorical. As for narrativeprose, it is clear that we cannot
restrictthe conception of narrativeto the gross events: the basis of
narrativeis the temporal order of symbols; in particular,the word-
order which is the movement of literature. We may, then, suggest
a link between narrative and grammar which would enable us to
associate our three areas of literature with the three areas of the
trivium into which the study of literaturewas formerlydivided. It
goes without saying, of course, that all three literaryelements are
simultaneously present in all literary works.
Thus we may see how, for instance, a strong narrativeor di-
dactic interest in poetry tends to infuse poetry with the word-order
of prose; and, conversely, how euphuism or elaborately figured
prose tends to become "poetic." Criticism was late in under-
standing the importanceof prose, and the subject is bedeviled by
two linguistic difficulties. "Prosaic"is not, as it ought to be, the
exact equivalent of "poetic"; and there is no short word corre-
sponding to "poem" for a literarywork in prose, nor for a literary
NORTHROPFRYE 253
work in general: hence the use of such periphrasticcacophony as
"hypotheticalverbal structure"in the present article. Much more
could be said on these points, but the general shape of the second
level, or the external relations of literature with other verbal dis-
ciplines, should by now be clear enough.

The Third Level.

The composing of a factual verbal structure is a "critical" op-
eration; the composing of a hypothetical structureis a "creative"
one, not that the two are ever separable. If there are three general
aspects of hypothetical writing and a single creative process, we
may best study the latter at the joining points of grammarand logic,
of grammarand rhetoric,and of rhetoric and logic. The link be-
tween grammar and logic is generally recognized: we need only
refer to Aristotle's subject and predicate, and the metaphysical
structuresbased on the fact that the verb to be implies both exist-
ence and identity. Again, a factual verbal structure cannot be
descriptivelycorrect unless it is verbally correct, and the accuracy
of one's meaning is inseparable from the order of one's words.
But this road is under construction: it is the other two that need
surveying.
The link between grammar and rhetoric appears to be a sub-
conscious paronomasia, or free association among words, from
which there arise not only semantic connections, but the more
arbitrary resemblances in sound out of which the schemata of
rhyme and assonance evolve. Finnegans Wake is an attempt to
write a whole book on this level, and it draws heavily on the re-
searches of Freud and Jung into subconsciousverbal association.
Uncontrolled association is often a literary way of representing
insanity, and Smart's Jubilate Agno, which is usually considered
a mentally unbalanced poem, shows the creative process in an in-
teresting formative stage:
For the power of some animalis predominantin every language.
Forthepowerandspiritof a CATis in the Greek.
254 REVIEW
KENYON
The sound of a cat is in the most useful preposition
Kat' euchen....
For the Mouse (Mus) prevailsin the Latin.
For edi-mus,bibi-mus,vivi-mus- ore-mus.. .
For two creaturesthe Bull & the Dog prevail in the English,
For all the wordsending in ble are in the creature.
Invisi-ble,Incomprehensi-ble,ineffa-ble,A-ble. . .
For there are manywordsunderBull. . .
For Brook is under Bull. God be graciousto LordBolinbroke.
It is possible that similar sputters and sparks of the fusing in-
tellect take place in all poetic thinking. The puns in this pas-
sage impress the reader as both outrageous and humorous, which
is consistent with Freud's view of wit as the escape of impulse
from the control of the censor. In creation the impulse appearsto
be the creative energy itself, and the censor the force which adapts
that impulse to outward expression,a force which might be called
the "plausibility-principle."
The final cause of all this paronomasiais the single interlock-
ing verbal structurewhich is the literal work of art. When this
has been developed to the point at which the author's conscious-
ness would normally accept it, it is still easy to see the links that
hold it together. All symbols in a verbal structureare, to use a
term now well established in criticism, ambiguous, both in sound
and in meaning. The factual verbal structure reduces this am-
biguity in two ways: first, by establishing a literal meaning, or
context, and second, by aligning the verbal symbolswith the things
they describe. The hypothetical structurehas deliberatelydiscard-
ed the latter: hence a poet's words, for instance, are limited in
their meaning by the context alone, and thus preserve a good deal
of their original variety of connotation. The repetition of a word
in poetry does not necessarily involve a repetition of the same
meaning, for the context may be different. Pope's Essay on Criti-
cism uses the word "wit" in nine or ten differentsenses. If Pope's
emphasis were on centrifugal or descriptive meaning, such a se-
mantic theme with variations could produce nothing but inextric-
able muddle. But in a poem the different senses all help to build
 FRYE
NORTHROP 255
up the word "wit" as a linguistic network of connotative meaning.
The poet, in short, does not equate a word with a meaning: he
establishes the powers or functions of words.
As for assonance, there is clearly room in semantics for a re-
newed study of what may be called rhetorical etymology, the ver-
bal associations that underlie thinking. The original is always
the unexpected, and the dialogue of Plato's that seems to me most
prophetic of new developments in thought is the one that is gen-
erally regarded as an irresponsible jeu d'esprit. I refer to the
Cratylus, which is clearly concerned with the relation between
thought and verbal association. Free play with words passed itself
off for centuries as real etymology, and when the latter was de-
veloped the former came to be regarded as fantastic nonsense. So
it is from one point of view, but it still remainsa datum for literary
critics of inescapableimportance.
The link between rhetoric and logic, between the image and
the concept, is in the diagrammatic structures underneath our
thoughts, which appear in the spatial metaphors we use. "Be-
side," "on the other hand," "upon," "outside": nobody could con-
nect thoughts at all without such words, yet every one is a geomet-
rical image, and suggests that every concept has its graphic form-
ula. I do not know that psychology has seriously examined the
way in which the arrangementof ideas in thinking is revealed in
the images unconsciouslyemployed to illustrate it; and of course
literary critics are only just beginning to realize that the figures,
illustrations, analogies and epithets - in short, the rhetoric - of
discursive writing form an essential part of its meaning. But
surely if someone says that science needs to be complemented by
poetry or religion or personal emotion because it is a mere cold
and dry approach to experience- a very common type of obser-
vation - he implies that his contrasting principle is warm and
moist, and hence the old myth of the four elements of chaos, or
perhaps an archaic creation myth like the one in the second chap-
ter of Genesis, is the graphic formula of his argument.
256 KENYON
REVIEW
Furtherstudy along such lines would tend, not to minimize or
obliterate the distinction between hypothetical and factual writ-
ing, but to show how the literaryor creative process makes factual
verbal structurespossible, in the same way that the hypothetical
structures of mathematics make the natural sciences possible.
And if the literary critic once understands the ambiguous na-
ture of literal verbal meaning, he need never again be caught
in the rat-trapof identifying all meaning with descriptive mean-
ing. As descriptive meaning is objective and intelligible, litera-
ture, in terms of this theory, must be either meaningless or de-
scriptive of something subjective or emotional - suggestive or
evocative of it, rather,as the subjectiveis too vague to be described.
This implies that literature is an elaboration of the lyrical cri de
coeur, and implies many other things which the critic well knows
to be absurd.
The establishing of the powers of words in literaturetakes us
much further than a mere recognition of ambiguity. The under-
standing of metaphysicsseems to depend on a techniqueof medi-
tation based on the connotative aspect of meaning. One normally
starts with a key word or concept, nature in Aristotle, form in Pla-
to, noumenon in Kant, duration in Bergson, and considers the
term in its centripetal relationships. The same is even more ob-
viously true of theology. This power of comprehension is, of
course, transferableand expansive: we may pass from one philo-
sophical structureto another until we become aware of a larger
verbal form called philosophy. The key words, nature, form, sub-
stance, time, being and the rest, thus expand into conceptualarche-
types, the linguistic elements or principles of this larger verbal
form. The same thing happens with images. All criticism of
poetry that gets beyond the second level of "'background"begins
with structural analysis, which identifies the recurring symbols
and themes in the verbal structure,and separates them into their
elements: this is the consonantiastage of Joyce'stheory. But these
elements are the elements of literature as a whole, and are not
 FRYE
NORTHROP 257
confined to the structurein which they appear. Structuralanalysis
thus expands into functional analysis. Moby Dick cannot remain
within Melville's novel: he is bound to be incorporatedinto our
total verbal experienceof leviathans and dragons of the deep from
the Old Testament onward. This is not a mere process of associa-
tion: the associationsconsolidate into archetypesof imagery. The
archetypalfeatures of narrativeare of equal importance, and may
be perceived in the different types of resolution: the quest-resolu-
tion of romance, the festival-resolutionof comedy, the death-reso-
lution of tragedy.
This conception of archetypesis based on the fact that literary
education is possible, and that the understanding of individual
works of art does expand into an understandingof literature as a
whole. Individual works of art lose nothing of their individuality
when we realize that they are not a series of bottled feelings, to be
uncorkedand resmelt like perfumes. The person who has attained
a matureunderstandingof literature,beyond both dilettantismand
pedantry, understandsit archetypally,whether he himself realizes
this or not. I add this last clause because of certain features in
modern literature that have, until very recently, discouraged
critics from trying to understand it on the third level. One of
these is the law of copyright, which prevents a writer from using
anotherman's work as the basis of his own, as Chaucerdid. This,
by exaggerating the uniqueness of the work of art, has developed
a criticism of connoisseurship, which talks less about literature
than about the pleasures of possessing books. Hence a division
grows between the creative and the critical functions which could
hardly get started in an age which understoodthe real meaning of
literary convention.
The importanceof convention in literatureis in facilitating the
comprehensionof it on the archetypallevel. For instance, when
Milton sat down to write a poem in memory of a friend of his
named Edward King, the question he asked himself was not,
"What can I find to say about Edward King ?" but, "How do the
258 REVIEW
KENYON
conventions and traditions of poetry demand that this sort of
situation should be handled?" Poetry demands, as Milton saw it,
that the elements of his theme should be assimilatedto their arche-
types. EdwardKing is, first, a dead man, who accordingto Chris-
tian doctrine will rise again. Hence the poem will not be about
King, but about his archetype,Adonis, the dying and rising god,
called Lycidas in Milton's poem. This archetype prescribes the
convention of the pastoral elegy, which historicallydeveloped out
of the Adonis lament. King was also a poet and a priest, and is
thus similarly linked with the appropriatearchetypesOrpheus and
Peter. All of these are contained in the figure of Christ, the
archetypeof King as an immortal soul. The poem urgently de-
mands the kind of criticism that will absorb it into the study of
literatureas a whole, and this critical activity is expected to begin
immediately,with the cultivated reader. This gives us a situation
more like that of mathematicsor science today, where the work of
creative genius is critically assimilated to the whole subject so
quickly that one hardly notices the difference between the two
kinds of activity. An even closer connection between creation
and criticismmay be seen in Dante's Conlivio.
The tropological level, therefore, is the archetypallevel, or the
mythical level, for I do not see any way of distinguishingarchetype
in this sense from myth. In all the kicking around that this latter
term has had in currentcriticism, one may notice, as usual, three
main types of what are generally called myths: the narrativemyth
(creation myth, death and resurrectionmyth), the image myth,
including the myth of the god, or archetypalhuman character,and
the conceptualmyth. But the third level is traditionallythe moral
level as well, and we may inquire here what sense can be given to
this word in criticism.
There is clearly no use looking for direct correlationsof aes-
thetic and ethical standards: one of the first laws of literatureis
that morally the lion lies down with the lamb. Bunyanand Roches-
ter, Jane Austen and Huysmans, Shakespeare's sublimity and
 FRYE
NORTHROP 259
Shakespeare'sobscenity, all belong together. Morality, like truth,
is not within literatureat all, and to derive moral values from it we
must again approachit from outside. So far from being "moral"
in any direct sense, the moral value of art seems actually to have
something to do with the breaking down of customarymoral re-
actions. This arises from the very nature of art as hypothetical.
Morality is constantly tending to incorporate itself in a series of
implied or expressed affirmations. But as soon as morality has de-
cided one thing, the poet is apt to hypothesize another; and, as
with truth, the affirmationlimits, and the hypothesis seems to have
something to do with emancipationor deliverance from the affirm-
ation even if we believe the affirmationto be true or good.
The moral value of art is connected with the fact that it forms
part of a "liberal" education, and the axiom underlying a liberal
education is that something does get liberated, even from the
knowledge of good and evil. This something is not the liberating
of the individual from the social imperatives of truth and good-
ness, but his introductioninto the free world of verbal hypothesis.
If we compare tragedy in art with suffering in life, for instance,
we can see that the containing hypothetical form of the art makes
tragedy pleasurable, even when there is no denial whatsoever of
the reality of suffering. Liberation here is not escape, but an in-
crease of intelligibility, a release of the powers of words.
The moral level is the social level, for it is by virtue of its arche-
types or myths that the work of art becomes the focus of a com-
munity. Factual verbal structureshelp to emancipate the human
mind from the pragmatic and compulsory rituals of animal exist-
ence by giving it a conscious vision of what it has to do and see.
Literaturetakes its place in a second effort of emancipation,which
gives man a vision of the total range of his creative powers, and of
his own world in relation to that total power. We express one-
third of the freedom of this vision, in which all the delight and
instruction of art are fulfilled, when we speak of freedom of
thought.
260 KENYON
REVIEW
The Fourth Level.
It is a simple axiom in the sciences that a "new" discoverymerely
articulateswhat was already latent in the order of nature. What
we have just said should illustrate something of the vulgarity of
the notion that a poet sits down with some blank paper and pro-
duces a new poem in a special act of creationex nihilo. The poet's
new poem merely articulateswhat was already latent in the order
of words, and the assumption of a single order of words is as
fundamental to the poet as the assumption of a single order of
nature is to the natural scientist. The difficultyin understanding
this point arises from the confusion of language with dictionary
language, and of literature with the bibliography of literature.
Language in a human mind is not a list of words with their cus-
tomarymeanings attached,but a single interlockingstructure,one's
total power of expressing oneself. Literature is the objective
counterpartof this, a total form of verbal expression which is re-
created in miniature whenever a new poem is written. Literary
education, which assimilates separate works into archetypes or
myths, leads us toward an intuition of this total form. And just
as physics can be looked at from one point of view as a set of in-
ferences from the assumption of a physical universe, so literature
may similarly be regarded as a set of inferences from the assump-
tion of a verbal universe. This verbal universe is that total vision
of creative power which we met at the end of the third level, and
which is not a diffusedbut a single vision, and to which every work
of literature in the world owes everything it has of wonder and
of glory.
The assumption in the word "universe,"whether applied to
physics or to literature,is not that these subjectsare descriptiveof
total existence, but simply that they are in themselves totally in-
telligible. No one can know the whole of physics at once, but
physics would not be a coherent subjectunless this were theoretic-
ally possible. The argument of Aristotle's Physics, which treats
physics as the study of motion in nature, leads inexorably to the
NORTHROP
FRYE 261
conception of an unmoved mover at the circumference of the
world. In itself this is merely the postulate that the total form of
physicsis the physicaluniverse. If Christiantheology takes physics
to be descriptive of an ultra-physicalreality or activity, and pro-
ceeds to identify this ummovedfirst mover with an existent God,
that is the business of Christian theology: physics as physics will
be unaffectedby it. The assumptionof a verbal universe similarly
leads to the conception of an unspeakablefirst word at its circum-
ference. This in itself is merely the postulate that literature is
totally intelligible. If Christiantheology identifies this first word
with the Word of God or person of Christ, and says that the vision
of total human creative power is divine as well as human, the lit-
erarycritic, as such, is not concernedeither to support or to refute
the identification.
In Dante's day the case was different. The whole idea of four
levels had originally come from theology, and had been worked out
in connection with the link between the Word of God and the
Bible. For Dante, therefore, the anagogic level of total intelligi-
bility was identical with the unfallen world of Christianity. Even
Sidney, when he says "nature'sworld is brazen; the poets only de-
liver a golden," is probablythinking of two existent worlds, as he
obviously does not mean that poetry gilds nature. In our own
day, when Joyce speaks of the final claritasor intuition of the total
form of art, he uses the theological term "epiphany,"though with-
out committing himself to the theological affirmationsinvolved.
Still, the religious annexationof the anagogic level of literature
is a historical fact. For us the immediate problem is to study the
archetypesor myths of literatureas parts of a whole, which we can
hardly do without the help of the integrationsof myth which have
been made in the higher religions and incorporatedin their scrip-
tures and sacred books. Thus the Bible becomes, for the literary
critic, an example of the literary form of the scriptures, which
unites narrative mytis (creation, redemption, etc.), image-myths
(the city, the garden, the personal God) and conceptual myths.
262 KENYON
REVIEW
The study of scripture as a certain type of hypothetical verbal
structurewould give point and direction to the currentinterest in
myths and archetypes,which latter, it should be said, cannot be
studied as separable content, but only as part of an analysis of
literary form. This in turn would give point and direction to the
allegorical criticismwhich is the main concernof the learned jour-
nals. And while the final insights into literatureare unspeakable,
and the fourth level is perhaps, like the first, largely outside the
direct scope of criticism,the climbing of this four-storeymountain
of meaning does not lead simply to an 0 altitudo! but to a pan-
oramic view of the surroundingfields of cultivation.