Cognitive Metaphor

Author(s): Charles O. Hartman

Source: New Literary History, Vol. 13, No. 2, Narrative Analysis and Interpretation (Winter,
1982), pp. 327-339
Published by: The Johns Hopkins University Press
Stable URL: http://www.jstor.org/stable/468915
Accessed: 11/06/2010 06:53

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=jhup.

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.

The Johns Hopkins University Press is collaborating with JSTOR to digitize, preserve and extend access to
New Literary History.

http://www.jstor.org
 Cognitive Metaphor
Charles 0. Hartman

ME ETAPHORfascinates twentieth-century critics. It is the city to

which all our roads lead. Traditionally, we feel, it languished
as one among the many figures of rhetoric. Now, perhaps
unjustly distrusting the very word rhetoric, we refuse to relegate
metaphor to the role of ornament. We ask it to transcend the old
categories, to serve functions foreign to a "mere trope." Most impor-
tant, we expect it to function cognitively. In many ways and many
terminologies, modern critics propose that metaphor, somehow, is or
creates or communicates knowledge.
Yet this claim has troubled some philosophers and some critics. It
encounters three major difficulties, each stemming from an implied
definition. First, a metaphor may be a condensed simile, "A is like B"
having become "A is B." Aristotle, whose various mentions of metaphor
anticipate or provide most subsequent definitions of the term, says in
the Rhetoricthat "the simile ... is a metaphor, differing from it only in
the way it is put"-the addition, in English, of "like" or "as."1 Both
tropes state similarities. This is unobjectionable but not very exciting.
Everything in the universe more or less resembles everything else.
(Or, alternatively, nothing is like anything else.) The writer who notes
a similarity between objects says little about them, especially when he
leaves their common characteristic-the basis of comparison-
unspecified. Howard Nemerov distinguishes metaphor from simile
on just this ground, that only the latter "isolates for you the likeness in
virtue of which the comparison is made."2 But in that case it is hard to
see what metaphor adds to simile besides confusion, or why we should
ever prefer it. Aristotle's answer-that "just because it is longer
[simile] is less attractive"-seems trivial; at least, it evades the question
of cognition. Furthermore, when the two things differ most, the basis
of comparison is likely to be the least characteristic attribute of
both-a lowest common denominator. Why, then, do we especially
prize-so warmly favoring Metaphysical juxtapositions-those
metaphors that "yoke by violence together" the most dissimilar
things?
We know ways around this objection-"surprise," the "harmony of
opposites"-but they work by distinguishing poetic effect from ra-
0028-6087/82/130327-13$ 1.00/0
Copyright? 1982 by New LiteraryHistory, The University of Virginia
328 NEW LITERARY HISTORY

tional knowledge. Such a distinction, though surely valid, begs the

question. Nothing in this view of metaphor encourages the critic who
wants to ally it with cognition. We resort, next, to a more sophisticated
definition: metaphor does not blandly remark a similarity but boldly
declares identity. A really is B. Yet the problem here is obvious. If this
is what a metaphor does, then a metaphor is a lie. It takes little philo-
sophical training to see that a false statement will lead to a true one
only by accident or revelation. If a rational or logical process begins
from false premises, its conclusions-even if they happen somehow
to be true-cannot reliably be called knowledge. (This suspicion
underlies Plato's rejection of the poet.) At this point the critic may
retreat into claims for a special kind of poetic knowledge. To do so,
however, seems anachronistic. These claims are essentially mystical,
relics of the Victorian attempt (made explicit by Arnold) to replace
religion with art, to seek revelation in the books of men. The modern
critic, whose defense of cognitive metaphor often suggests envy of the
natural sciences, weakens his position by adopting this line of argu-
ment. Trying to save cognitive metaphor by insisting on a new (and
vague) meaning of cognitive, he shares the fate of those whose fervor
outstrips their rhetoric. He may convert the gullible and gratify the
sympathetic, but he does not convince the skeptic. We do not con-
struct new definitions of knowledgefor psychoanalysis or our senses or
experimental science; what would warrant our doing so for poetry?
The objects of knowledge vary, and the ways of getting or giving it;
but knowledge remains single in kind.
Finally, from natural science itself (and the empiricist philosophy
that serves it, and the psychologies that derive their principles from it)
comes the objection that metaphor produces not knowledge but de-
lusion. The usual term is "category mistake." Though the jargon and
its philosophical basis are relatively modern, the accusation burdens
many traditional definitions of metaphor, as when Aristotle says that
metaphor is "giving the thing a name that belongs to something else."3
Though he approved of metaphor as a "sign of genius"-and though,
as we shall see, he provides elsewhere a way to cut through these
tangles-this definition resembles that with which I began, in disap-
pointing any hope that metaphor could produce real knowledge. If
(as the definition assumes, and as both physics and psychology assure
us) we perceive the world in terms of categories we have learned or
inherited, and if rational action depends on our correctly applying
these categories, then the habit of kidnapping a word and cuffing it to
an alien object endangers sanity itself. Despite a tradition of imagery,
if I treat my guitar as a woman I am insane. And if I realize that my
guitar is not a woman, how can I learn anything from speaking as
COGNITIVE METAPHOR 329

though it were? I think I do learn something, but this view of

metaphor mystifies the process. Why, in short, in questing for
knowledge (or attempting to communicate it), should we begin by
pretending that the object is something it clearly is not? Perhaps, some
critics suggest, comparison with things we know allows us to com-
prehend what we do not. Metaphors sometimes do this; but I know as
well what a guitar is, and what a woman is, before I compare them as
after.
Nor do the confusions generated by this third definition of
metaphor end here. Definitions stemming from Aristotle's imply that
a term from one context, serving as the vehicle, carries across to that
of the tenor.4 Thus in "the ship plows the waves," a term from ag-
riculture invades a nautical description. This might be confusing
enough, though one assumes that the larger context of discourse-the
story of an ocean voyage, for instance-would clarify relations be-
tween the categories involved, the centrality of one and irrelevance of
the other. But many metaphors reverse the transference. In "A Val-
ediction: Forbidding Mourning"-that primary text for modern
treatments of metaphor-the stanzas depicting the compass (the ve-
hicle of a metaphor whose tenor comprises the pair of lovers) bulge
with terms that normally inhabit a context more closely neighboring
on lovers than on compasses. The compass is said to "hearken" and
come "home." These words describe a compass only indirectly,
fancifully-in fact, metaphorically. Somehow, while the compass
stands for the lovers, the lovers stand for the compass; the juxtaposed
contexts grow ever more entangled, and the cognitive value of the
metaphor becomes harder and harder to explain. Even the categories
of metaphor itself, tenor and vehicle, start to confound themselves.
Again, we recognize the phrase "a sea of faces" as a standard
metaphor: a crowd of faces resembles a sea. But when, in "The
Yachts," William Carlos Williams says, "It is a sea of faces," he inverts
the metaphor: the sea has become a mass of faces. Once more, if the
function of metaphor is to communicate knowledge, these poets'
choice of method seems perverse.
Compressed simile, forged equation, and category mistake are the
three most common definitions of metaphor. By specifying certain
characteristics of metaphor, each accounts for part of its effect in our
experience. Yet each jeopardizes the claim that metaphor generates
knowledge. By any of these definitions, "cognitive metaphor" be-
comes an oxymoron. I want to suggest that a fourth definition ob-
viates these difficulties. It is neither new nor mine; I owe its rediscov-
ery to Hugh Kenner.5 In the Poetics, Aristotle lists the kinds of
metaphor: "Metaphor consists in giving the thing a name that belongs
330 NEW LITERARY HISTORY

to something else; the transference being either from genus to

species, or from species to genus, or from species to species, or on
grounds of analogy ... analogy is possible whenever there are four
terms so related that the second (B) is to the first (A), as the fourth (D)
to the third (C)."6We have already seen that the formula "giving the
thing a name that belongs to something else" undermines the concept
of "cognitive metaphor." The last of Aristotle's list, however, repays
attention.
Briefly, it means that one properly generalizes a metaphor, not as a
simple equation (A = B), but as a proportion: A/B = C/D. This defi-
nition, which appears to luxuriate irresponsibly in mathematical sym-
bols and to tempt the critic to trifle with diagrams, has been largely
ignored. Many metaphors seem to offer simpler reductions as pairs of
terms: "my love is a rose" reduces to "lover = rose" and abstracts to "A
= B." Yet as Kenner points out, even so casual a metaphor as "the ship
plows the waves" loses something vital when we reduce it to "ship =
plow." (That "my love is a rose" suffers similarly we will see later.)
More accurately, the metaphor implies that "a ship does to (or is to)
the waves what a plow does to ground": ship/waves = plow/ground.
Though we have obviously abandoned elegance of expression, we
have saved the essential information; and it is information that con-
cerns us when we speak of metaphor as cognitive. Returning to "The
Yachts," we can observe the same kind of metaphor emerging in the
last three stanzas. The yachts in their race, contending, in the poet's
vision, not against each other but against the sea through which they
cut their way, become an emblem of those who own them. We might
reduce this part of the poem, as a metonymy, to an equation of two
terms: yachts = their owners, "the rich." But in those visionary stanzas
Williams turns away from the yachts, attending instead to the "watery
bodies" of the waves, the "sea of faces" which the yachts "pass
over"-their indifference rescinding the celebrative imagery the first
eight stanzas lavished on them. If the poem is a metaphor (it becomes
impossible to read literally), it exceeds the complexities of the metony-
my. A four-term analogy, however, allows us to include the sea on
whose suffering the poet concentrates:

yachts _ rich
sea poor
At least in describing what a metaphor contains, the four-term
analogy surpasses the two-term equation.
Before asking how this improves the case for metaphor as a cogni-
tive process, we should clear a little more ground. Aristotle himself
clearly intends his definition quite narrowly. For him analogy is only
COGNITIVE METAPHOR 331

one kind of metaphor, presumably to allow for "simple" ones in two

terms. But we have seen how the definitions available for two-term
metaphor exclude it from the realm of the cognitive, and we might
turn back to see if we can rescue it. Mathematics of a trivial sort will
help. If we can reduce "my love is a rose" to "A = B," we can also
express that equation as a four-term proportion: A/C = B/C. It stands
essentially unaltered. And this corresponds more closely to what we
understand in reading or hearing such a metaphor. By itself, "my love
is a rose" is merely odd. Instead, what affects us is the attitude it
embodies-though the speaker asserts himself only in "my." The
woman and the rose are not equated in a vacuum; someone compares
them, and he is the matrix in which they become comparable. "My
love is a rose" means something like "I feel about my love as I do
about a rose" or "my love does to me what a rose does to me." (Com-
pare "the ship does to the waves what a plow does to the ground.") The
structure of "my love is a rose," then, is
my love _ rose
me me

Though this diagram obviously does not replace "my love is a rose," it
preserves more information than "lover = rose."7
If "my love is a rose" represents a reduction in the number of terms
of the analogy-though only to three, not to two-many of the most
interesting metaphors in literature require the proportion to expand
beyond four terms. An example from Yeats:

Memory
One had a lovely face,
And two or three had charm,
But charm and face were in vain
Because the mountain grass
Cannot but keep the form
Where the mountain hare has lain.8

This remarkable poem deserves more attention than I can give it

here.9 But it unfolds from a central metaphor. The first two lines list
various (noticeably vague) women; the third dismisses their virtues as
futile (not only "vain" but "in vain," implying attempts-at what, we
do not yet know); and the last three lines speak of a hare, its nest in
the grass, the grass itself, and the mountain on which it grows. Incon-
gruity makes us-probably before we consider consciously-read the
final explanation ("Because") as metaphor. We intuit the basic equa-
tion easily enough: the hare represents some one special woman who
332 NEW LITERARY HISTORY

affected the speaker too strongly for other women's charms to attract
him later, after she left. In short, "hare = woman." But clearly this
ignores most of the metaphor. The hare (within the context of the
vehicle) associates with the form, the grass, and the mountain. By a
process of deduction which is, again, probably unconscious but which
operates on the relations among the terms of the vehicle and transfers
them to the tenor, we fill in the remaining terms:
hare woman
mountain _ speaker
form memory (as in "a memory")
grass memory (the faculty of memory)
The complex process by which we arrive at these terms will concern us
later; but it is clear that we must acknowledge eight terms here and
not four. This example and that of "simple" metaphor suggest that
four terms are only a kind of model or norm and that in practice a
metaphor can reduce the number of different terms by one or ex-
pand it perhaps indefinitely.
We ought also to notice in these examples the presence of implicit
terms. Although in "my love is a rose" the speaker who acts as the
third term of the metaphor declares himself in the possessive pro-
noun, "the ship plows the waves" does not mention "ground" or
"earth," the necessary fourth term. In "The Yachts," although the sea
is said to be full of "watery bodies," what leads us to conclude that it
represents "the poor" is no linguistic evidence in the poem itself, but
the structure of the metaphor. In "Memory" we had to supply the
woman, insert the speaker who never mentions himself, and perceive
the double significance of the title as naming both a faculty and what
it contains or manipulates. In most cases the terms of the vehicle are
explicit, but usually we must deduce at least one term of the tenor.
The logical extension of this technique is what Hi Simons, speaking of
Wallace Stevens, calls "radical metaphor"-metaphor that announces
only its vehicle, the whole tenor being left to implication. Simons
proposes this as the distinctively modern use of metaphor. As Ste-
vens's short poems show, it generates metaphors that yield multiple
interpretations. One could defend any reading in which the relation
among elements in the proposed tenor matches the relation among
explicit elements in the vehicle.10
I hope I will not be thought to claim that by reducing these
metaphors to proportions I replace the originals. More happens in
the metaphor of "Memory" than a diagram even of eight terms iden-
tifies. The "hare" connotes swiftness and wildness, suggesting certain
things about the absent woman it represents. For the speaker to figure
COGNITIVE METAPHOR 333

himself as a mountain is to imply solidity and (since we see the moun-

tain from its side, not from a majestic vista) isolation. Such connota-
tions are indispensable to the effects metaphors have on us. It would
be silly to deny them. But cognition is a word that implies knowledge as
distinct from emotion.11 Obviously, in poetry (as, ideally, anywhere
else in our lives) the two cannot be separated. Yet when we ask of
metaphor how and whether it produces knowledge, we plausibly look
to the information it contains. Proportional analogy proves itself to be
a credible way of representing the solid core of information in any
metaphor. We need only remember, while searching for cognitive
value, that a metaphor has other values as well, and other ways of
working.
It seems possible now to answer the question with which I began. I
suggested one fairly simple way in which metaphor is cognitive when I
noted that each analogy involves implicit terms. These terms are, after
all, bits of information. In "Memory," one of them allows us to replace
the speaker at the center of a situation from which he has apparently
removed himself to ensure a tone of resignation or indifference. The
poem pretends to reflect in detachment on the workings of a mental
faculty. Behind that facade, it nurses a personal wound. Told of the
"mountain hare" that has (unnaturally, says the adjective) left its
"mountain," we infer the speaker's suppressed pain, and perhaps
anger, that the woman has deserted him. We need to discover him in
the poem to understand this. But we accept his reticence as a neces-
sary condition of his accuracy in feeling and comprehending the ex-
perience. His very silence affects us. Only the metaphor, in which he
can remain an implicit term, combines this silence with the informa-
tion we need to perceive its significance. In "The Yachts" the implicit
terms of the tenor, which the structure of the metaphor bids us call
social classes, associate the poem with a point of view closer to Marx's
than Jay Gatsby's. And so on. The "information" in each case may
involve or underwrite tone; but if the objectsof knowledge in a poem
are, like those in psychoanalysis, emotive, this makes it none the less
knowledge. Furthermore, as Aristotle points out, "For some of the
terms of the proportion there is at times no word in existence."12 We
noticed this justification for metaphor earlier-that it makes known
what was unknown.
But we also recalled then that this breaking of new ground does not
account for all or even most metaphors. Often the information these
implicit terms convey seems trivial, as in "the ship plows the waves."
Sometimes, as in "my love is a rose," the metaphor eschews them
entirely. And in any case they depend for their significance on the
metaphor's structure, which leads us to posit them. It is hard to see
334 NEW LITERARY HISTORY

how they could detach themselves from that structure to operate as

knowledge in any wider area. If they contribute to a cognitive func-
tion of metaphor, it must be by the way in which we deduce them.
Cognition must finally arise from metaphor as a process of knowing
rather than reside as a thing known; and we had best turn from the
implicit terms to the metaphorical structure itself.
And there, in fact, we find what we need. The very possibility of
deducing from a metaphor terms which it does not explicitly present
implies that the metaphor is a structure of relations. (The syllogism
and the sorites are similar deductive structures.) This is why increas-
ing the percentage of verbs-which carry the weight of action and
relation in a sentence-almost inevitably renders a passage more
metaphorical. The mathematical form analogies take confirms the
same point. A metaphor equates, not objects, but relations or sets of
relations. If we reduce it to A = B, we are hard put to make sense of
the doubled hyphen between the two letters. But in A/B = C/D, the
sign of equation means what it says. (It is hard to resist the frivolity of
remarking that this holds even for the abstract equation when we take
it "literally": A precedes B in the alphabet-a definable relation be-
tween the two-just as C precedes D.) Metaphor does not state a
similarity of objects, as the first objection to calling it cognitive would
have us think, but an identity of relations.13 Nor does the metaphor
lie, as the second objection complains. What a ship does to the waves
is, literally and exactly, what a plow does to the ground. An action is
one kind of relation that connects two things, an agent and an object.
The metaphor detaches the action from the things and applies it to or
embodies it in new things. If its precision is limited, so is the accuracy
of all means of knowing. Its limits are those of language. We need not
charge metaphor with falsity. The statement in which it begins is as
true as the knowledge we hope to see it communicate. It might be
possible-though beyond both my scope and that of this essay-to
show that all accretions of knowledge take this form. This is how we
learn "what things are," how we form categories.
This returns us to the third objection, that a metaphor confuses
categories. A category mistake is dangerous because it betrays its vic-
tim into taking one thing to be another. But metaphor, once we see it
as analogy, does not threaten this at all. Rather than confuse one
category with another, it demonstrates the identical relations obtain-
ing within both. It does not invite us to act irrationally-as I would if I
believed my love to be a rose-unless we misread it. In fact, to define
metaphor as category mistake is to confuse two categories, poetry and
insanity. In "Creative Artists and Daydreaming," Freud commits this
error. Hugh Ormsby-Lennon has eloquently sketched its history in an
COGNITIVE METAPHOR 335

essay entitled "Metaphor and Madness."14It results from mistaking a

metaphor, "poetry is a divine (or other) madness," for a simple equa-
tion. The real statement to be verified is more like this: that the
relations among the objects of poetry are identical to those among the
contents of consciousness that typify a madman. One could explicate
the details of these relations, like those in the similar metaphors "a
poem is a dream" and "a poem is a joke." But the language distin-
guishes poems from dreams and jokes; nor need we confuse the
poet-as a poet-with the madman.
Though Aristotle, conceiving of metaphor as analogy, has helped
us refute the objections that first encumbered our search for cognitive
metaphor, examples of the striking similarity between understanding
a metaphor and reasoning cognitively in general may clarify the
question. Following the modern tradition, I take as my first text a
stanza from Donne's "Valediction":

Our two soules, therefore, which are one,

Though I must goe, endure not yet
A breach, but an expansion,
Like gold to ayery thinnesse beate.

One could reduce this metaphor to A = B: "our souls are gold" or,
more fully, "our souls are like gold." The simile suggests that the
lovers share with gold some common property, and the obvious can-
didate is "high value." But Donne's treatment of the metaphor in-
trigues us in just this regard, that he bypasses the obvious meaning,
taking up instead a more arcane property of gold, its ductility. Re-
ducing the metaphor to simile, we lose this.
Taking the stanza as a whole, we discover several sets of mutually
opposing terms. Even before the last line, Donne establishes a kind of
analogy:
two soules breach
one expansion
(He also links the second and fourth terms by rhyme.) When we carry
this information into the last line, we find three terms of the
metaphor already given, and the fourth easy to deduce:
breach _(lead)
expansion gold
(In metallurgy as well as in commerce, gold is the opposite of "dull"
lead, which fractures when "beate.")15But we have not finished. We
have moved from an analogy that straddles the line of metaphor to
one that confines itself within the vehicle. To move back into
336 NEW LITERARY HISTORY

metaphor, we must incorporate an antithesis that pervades the whole

poem: "our soules" stand against others, those of "dull sublunary
lovers." From this (as it cooperates with our first analogy) we conclude
that
two soules others' souls
one ours

This analogy resides within the tenor. But in such a chain of

analogies, substitution becomes possible. Things equaling the same
thing equal each other, and from these three proportional equations
we can derive (among others) this fourth:
others' souls (lead)
ours gold
Here we ought to notice two things. First, we have ended with a
version of the initial equation (our soules = gold), but have enriched
it. The value of gold is complicated by the various contexts through
which the metaphor has passed it. The juxtaposition of commerce
and metallurgy suggests a third context, that of alchemy. Alchemists
sought to derive gold, the noblest metal, from base lead. In that sys-
tem, again, gold has high value as compared with lead, but on less
materialistic grounds than we could originally have seen. That a re-
fined or spiritual value can replace a more obvious one is, of course, a
theme that unifies the "Valediction." This is how the line's metaphor
fits the poem in which it appears. Second, the way we arrive at this
conclusion or comprehend the several implications of the line-a way
which our chain of analogies has imitated formally-resembles the
cognitive process of following a chain of syllogisms, a sorites. We
should, however, probably distinguish Donne's perpetual play among
the forms of logic from the real knowledge that underlies it. The
cognitive process of the metaphor is not precisely the argumentative
one it enacts and continues, but the one we perform in understand-
ing it.
"The Yachts," at which we have already glanced, yields a more
direct view of metaphor functioning cognitively. What strikes us most
in the way the poem develops during its first eight stanzas is the
density of metaphors-metaphors which are nevertheless standard
and seem to be offered casually, ornamentally. The pervasive trope is
personification. Williams populates his lines with figures half-realized
by the language. He calls the "lesser and greater craft" on which
spectators crowd "sycophant." "The sea is moody," the ocean beyond
the "well-guarded arena" "ungoverned." The waves in the protected
bay where the race takes place "lap" the yachts' "glossy sides, as if
COGNITIVE METAPHOR 337

feeling / for some slightest flaw." When, in the last three stanzas, "the
horror of the race dawns staggering the mind" of the speaker, we see
these half-hidden metaphors burst out into full light:

Arms with hands grasping seek to clutch at the prows.

Bodies thrown recklesslyin the way are cut aside.
It is a sea of faces about them in agony, in despair ...6

As we follow the poem, we are watching metaphor blossom into

grotesque vision, its cognitive potential realized as knowledge of the
"horror" a social system can breed.
I want to offer a final example to suggest a direction these specula-
tions might take. It comes again from Williams, a short piece he calls
simply "Poem":

The rose fades

and is renewed again
by its seed, naturally
but where
save in the poem
shall it go
to suffer no diminution
of its splendor17

Though more could be said about the poem, I want to concentrate on

one small question: According to what principle does Williams make
the apparently flat and innocuous phrase "shall it go" a line by itself?
So much attention seems odd even when we see that syntactically it
rules the poem's second clause. But the reason for this emphasis be-
comes clear as soon as we realize that the line necessarily embodies a
metaphor. Sessile roses can "go" nowhere. If they could move, they
could not "go" to the poem, which is not a place in the same way as a
garden or a vase. What "fades / and is renewed again / by its seed" is a
flower; but the "rose" that inhabits the poem is a word. In retrospect,
the first line discloses a sort of pun (not the last in the poem) or
paradox. But the pun reveals the heart of poetry, and of language
itself. The word rose, says the poem, serves as a metaphor for the
flower by naming it.
This recalls a statement that critics and others often make: that all
language is metaphorical. While this has become nearly an article of
faith, what we have seen of analogy suggests a logical foundation. To
retrieve the other terms of the analogy, we need only examine (as with
any vehicle and tenor) the contexts to which Williams's two "roses"
belong. One is an element in the domain of words, the other dwells in
338 NEW LITERARY HISTORY

the realm of things. Language becomes the vehicle of a vast

metaphor. In English it comprises not two or four but about half a
million terms. On the other side we find the tenor of the metaphor in
the system of things that constitutes the world. Once again, the
operative word is "system" or "structure." Ever since de Saussure,
linguists have assumed that a word does not mean directly or inher-
ently but by relating to other words within a structure of differences.
Saying that "language is metaphor," we imply that the relation a cer-
tain word has to the system of all words is precisely the same relation a
certain thing (the one the word "names") has to the system of all
things. This raises a further question, which I hinted at before only to
decline it, as I must do again: How does the bric-a-brac of things we
call "the world" become a world, a cosmos, a system? A possible an-
swer is that our minds, our senses, and our language create that
system by a process one ought to call metaphorical. This answer
would take its first authority from the Critiqueof Pure Reason, and it is
the one suggested by the work of gestalt psychologists. But its testing
must remain their work, however valuable the critic's evidence; it is
not our central concern.
Cognitive metaphor is analogical, and so far as we can express a
metaphor as an analogy, it is cognitive. Still, cognition does not ac-
count for all the functions and effects of metaphors. It may not even
be the primary one. One of the least satisfactory defenses of metaphor
holds that it reveals the true nature of a particular thing, a nature
ordinary language cannot reach. But if all language is metaphor,
there is no "ordinary language" in this sense. Metaphor tells us the
truth of things no better and no worse than other uses of language.
The way it does seem to exceed the ordinary capabilities of language
lies in a service that partakes at least as much of emotion as of cogni-
tion, and above all unites the two. Some call this service "defamiliari-
zation"; the original force of "realization" aims in the same direction.
Metaphor not only lets us know, it helps us to know what we know: to
understand.

UNIVERSITY OF WASHINGTON

NOTES

1 Aristotle Rhetoric(tr. W. Rhys Roberts) 3. 10. 1410b16. Is this sentence of Aristotle's

a metaphor? Are all propositional statements metaphors? Northrop Frye's distinction
between "identification as" and "identification with," though it does not answer these
questions, can inform thought about them. See Anatomy of Criticism (1957; rpt. New
York, 1968), p. 123; and, more fully, "The Realistic Oriole: A Study of Wallace Ste-
COGNITIVE METAPHOR 339

vens," in Wallace Stevens: A Collection of Critical Essays, ed. Marie Boroff, Twentieth
Century Views (Englewood Cliffs, N.J., 1963), pp. 170 ff.
2 Howard Nemerov, "On Metaphor," Virginia QuarterlyReview, 45, No. 4 (Autumn
1969), 623. Also reprinted in Nemerov, Reflexionson Poetryand Poetics (New Brunswick,
N.J., 1972).
3 Aristotle Poetics (tr. Ingram Bywater) 21. 1457b6. This has become the most com-
mon definition of metaphor, and covers the subject as far as many dictionaries, glos-
saries, and encyclopedias are concerned.
4 I use the terms tenor and vehicle for lack of better ones. They have a use in distin-
guishing the two contexts that are undeniably brought together, somehow, in a
metaphor. That the view of metaphor they imply needs refinement and qualification is
part of this essay's contention.
5 See The Poetryof Ezra Pound (Norfolk, Conn., 1951), esp. pp. 87-88, but also chs. 7
through 12. Owen Barfield, in Poetic Diction (Middletown, Conn., 1973), seems to start
from the same point, as do other critics. The reader interested in the history of thought
about metaphor cannot do better than to consult Warren A. Shibles, Metaphor: An
Annotated Bibliographyand History (Whitewater, Wisc., 1971). A reference list of those
who have, one way or another, taken metaphor to be analogical can be found on p. 372.
6 Aristotle Poetics (tr. Bywater) 21. 1457b7- 17.
7 I am indebted to Michael Mesic for much of this paragraph's substance.
8 W. B. Yeats, "Memory," in CollectedPoems of W. B. Yeats (New York, 1956), p. 147.
9 But I should say two things. First, "form" has a technical meaning-the nest a hare
makes in grass-which can be missed. Second, if one misreads "Cannot but keep" as
equivalent to "Cannot keep"-as Jon Stallworthy does in Between the Lines: Yeats'Poetry
in the Making (London, 1963, p. 18)-the poem adamantly refuses to make sense.
10 Some valid interpretations exceed others in plausibility for reasons external to the
metaphor itself-considerations of an author's habitual themes, for instance. A review
of critical comment (by Edward Kessler, Joseph N. Riddel, and Lucy Beckett) on Ste-
vens's "Valley Candle" demonstrates both the multiplicity and the convergence of
plausible interpretations. Note that critics other than Simons use the phrase "radical
metaphor" in a very different way; see Shibles, Metaphor, p. 402.
11 See, for instance, "Cognition," EncyclopediaBritannica, 11th ed.
12 Aristotle Poetics (tr. S. H. Butcher) 21. 7. In this instance Butcher is more helpful
than Bywater, whom I have used elsewhere. His translation may be found in Hazard
Adams, Critical TheorySince Plato (New York, 1971).
13 Just so, we may now be able to rescue simile as well as "simple" metaphor by
including the former in the latter. Many similes deal, like metaphors, with relations,
disguising their equations with a modest "like." The example from Donne's "Valedic-
tion" given below supports this point. So does Aristotle; see Rhetoric (tr. Roberts) 3. 11.
1412b34.
14 Hugh Ormsby-Lennon, "Metaphor and Madness," Etc.: A Review of GeneralSeman-
tics, 33, No. 3 (Sept. 1976), 307-18.
15 According to the Encyclopedia Britannica, 11th ed. (articles under the metals'
names), gold shares with lead a high specific gravity and "is nearly as soft as lead." But
"it is also extremely ductile; a single grain may be drawn into a wire 500 ft. in length,
and an ounce of gold covering a silver wire is capable of being extended more than
1300 m." As for lead, "Its breaking strain is very small: a wire 1/10th in. thick is
ruptured by a charge of about 30 lb."
16 William Carlos Williams, "The Yachts," in SelectedPoems (New York, 1969), p. 72.
17 Williams, "Poem," in Picturesfrom Brueghel (New York, 1962), p. 39.