copyright © Donald Freeman, 1991.
Poetics Today 12 (spring 1991) 145-164.

Mark Johnson, The Body in the Mind: The Bodily Basis of Meaning, Imagination, and Reason. Chicago: University of Chicago Press, 1987. 227 pp.

George Lakoff and Mark Turner, More than Cool Reason: A Field Guide to Poetic Metaphor. Chicago: University of Chicago Press, 1989. 230 pp.

Samuel R. Levin, Metaphoric Worlds: Conceptions of a Romantic Nature. New Haven: Yale University Press, 1988. 251 pp.

Phillip Stambovsky, The Depictive Image: Metaphor and Literary Experience. Amherst: University of Massachusetts Press, 1988. 156 pp.

Ann and John O. Thompson, Shakespeare: Meaning and Metaphor. Iowa City: University of Iowa Press, 1987. 228 pp.

Mark Turner, Death is the Mother of Beauty: Mind, Metaphor, Criticism. Chicago: University of Chicago Press, 1987. 208 pp.

"The Nature of Visionary Fancy, or Imagination, is very little Known, & the Eternal nature & permanence of its ever Existent Images is consider'd as less permanent than the things of Vegetative & Generative Nature; yet the Oak dies as well as the Lettuce, but Its Eternal Image & Individuality never dies, but renews by its seed; just so the Imaginative Image returns by the seed of Contemplative Thought; the Writings of the Prophets illustrate these conceptions of the Visionary Fancy by their various sublime & Divine Images as seen in the Worlds of Vision." -- William Blake, A Vision of the Last Judgment (1810)

"Reason is embodied in the sense that the very structures on which reason is based emerge from our bodily experiences.... Imagination is not mere fancy, for it is imagination, especially metaphor and metonymy, that transforms the general schemas defined by our animal experience into forms of reason -- forms even richer than the objectivists' transcendental reason has been taken to be." -- George Lakoff, Women, Fire, and Dangerous Things (1987)

Reviewing these books is something of an exercise in theology. Either one does or one does not believe, as classical theories of metaphor and meaning require, that there is a language "out there" based on classical categories, necessary and sufficient conditions, and predicate calculus. With that belief come certain necessary principles for a theory of metaphor: that there is a "literal," non-figurative language in contrast to the figurative language of metaphor; that this "ordinary" language is unmarked while metaphor is language that is deviant, foregrounded, highlighted, made strange; that there is a difference between literary and everyday metaphor, and a difference between literary and "ordinary" language (a position that had begun to be undermined in more general literary theory before the "linguistic turn" in the study of metaphor examined here, to be sure, but which was stated least equivocally about metaphor).

These principles dominated discussion of metaphor for two thousand years, from Aristotle until a time when, at the turn of the last decade, there appeared a cloud no bigger than a man's hand: a little book by George Lakoff and Mark Johnson, Metaphors We Live By (1980). They argued that, far from being decorative or parasitic upon "ordinary" language, metaphor figures in most if not all language, and is fundamental to our very understanding; that, further, metaphor is not an intellectual abstraction but is based in what Johnson was to call our "embodied human understanding"-- in short, that meaning and metaphor derive from our bodily experience. In a style strongly reminiscent of the English Romantic poet, painter, and revolutionary William Blake (as far as I am aware, the Romantic period was the only epoch in which a similar theory of metaphor was propounded), they asserted at the end of this volume that "[m]etaphor is as much a part of our functioning as our sense of touch, and as precious." (Lakoff & Johnson 1980:239)

Whatever one's view of this work, it has set the agenda for the ensuing debate. Books (to say nothing of scholarly articles) on metaphor have appeared in such profusion that it is nearly a full-time job to keep track of them. Of the six studies of literary metaphor reviewed here (all of which have appeared since late 1987, and these are not the only ones), one lacks a coherent position of its own; three -- unsurprisingly -- espouse the Lakoff-Johnson view; and two stay within the traditional paradigm. All of the authors under review here agree to some extent that metaphor is not merely a matter of words but a matter of thoughts, or concepts (or, for Samuel R. Levin, conceptions). Levin and Phillip Stambovsky postulate a poetic or metaphoric world -- for Levin, readers of poetry construct a world in which the purportedly deviant expression of a metaphor could be said to be true.

In Levin's Metaphoric Worlds, literary metaphors are "expressions that evince a degree of linguistic deviance in their composition." (1) Levin implies that some, perhaps most, language is non-metaphorical and that there is a significant statable distinction between "deviant" and "non-deviant" language richer than the distinction between grammatical and ungrammatical utterances (a distinction that itself has been breaking down in linguistics for more than twenty years). What distinguishes metaphors from ordinary expressions is that while ordinary expressions need only be grasped, metaphors, Levin argues, must be construed.

Levin holds that when we construe literary metaphors, we modify our conception of the world to fit the "deviant" expression -- we conceive of, in his terms, a world in which the "deviant" metaphorical expression could be true. "The reader must start from the actual utterance but then, in awareness of the poet's linguistic straits, he must negotiate through that utterance to the poet's original insight. This negotiation on the part of the reader represents not so much a semantic construal as it does a phenomenological or conceptual construal."[1] (141-2)

Levin distinguishes the mental acts of "conceiving" (concepts) and "conceiving of" (conceptions), on the one hand, from imagining, on the other. "Conceiving of" is an intellection about states of affairs that do not exist (?I conceive of a calm sea/I conceive of a laughing sea), whereas imagining is an intellection about states of affairs that do exist but are not present (?I imagine a laughing sea/I imagine a calm sea). Acts of imagining can be validated by an "empirical sequel" (we may find a calm sea one day); acts of conceiving and "conceiving of" have no empirical sequel (we will never find a laughing sea).

Levin further distinguishes "conceiving" (which produces concepts) from "conceiving of" (which produces conceptions). Of the two mental actions, conceiving has more "epistemological weight," on this account. The conceiver has a clear image of what he conceives, and conception belongs to semantics: one's concept of a horse, for example, is a dictionary entry. But "conceiving of" a laughing sea is, Levin holds, preparing a mental space into which a laughing sea might be placed, but which cannot be filled with a concept. We can conceive of a laughing sea but not conceive one, because a laughing sea does not exist.

This "conceiving of" is, for Levin, the metaphoric impulse. Metaphoric construal, and hence reading poetry, consists in substantial part in "conceiving of the states of affairs that, taken literally, [the `deviant sequences' of metaphorical language] describe." (80) In "conceiving of," a reader of poetry does not interpret the "deviant" language of a metaphorical sequence; rather, he conceives of a world in which that language might not be deviant and the conceptions it embodies might be true. "[I]t is only by the conceiving of such possibilities that a reader can approximate to the insight or vision that the poet has achieved and (imperfectly) expressed." (142), a vision creating a metaphoric world "whose nature, in its abrogation of the canons that govern existential relations in our world, is estranged from common notions of reality and may rightly be termed metaphoric." (237)

That, in brief, is Levin's theory of literary metaphor. It is in essence a theory of reading certain kinds of metaphorical discourse. Metaphoric Worlds is immensely learned, but it depends for its theory to work on a special status for literary metaphor in particular and a deviant status for metaphoric expression in general. I am attracted to the notion that reading poetry like The Prelude is a world-creating experience in which the metaphors of that world are true[2], but the attraction is more to this idea's analogical appeal than to its probable accuracy. Little new about Wordsworth's poetry arises from it.

Levin's theory of metaphoric worlds implies that literary metaphors are different in kind from the metaphors that occur outside of literary artworks, and he gives no systematic accounting for this difference. If we recreate within us the world of a poem by conceiving of a world in which its metaphors can be literally true, then it must also be the case that we recreate within us the force of a non-literary but metaphorical expression by conceiving of a world in which its metaphors are literally true. But then how do we account for Levin's notion of lexicalization, the point at which metaphors can be said to "die," to become backgrounded, to become, in Levin's terms, "non-deviant"? Every theory of poetic language that depends on a variety of deviance -- aktualisace, ostranie, Entfremdung -- has entailed some theory of backgrounding or automatization.

Levin seeks a way out of this dilemma by appealing to our knowledge about our own language. Attacking the Lakoff-Johnson account in Metaphors We Live By, he observes: "When I say `I spent three hours on this problem' or `This theory is weak', I am not aware that these statements are conditioned in any way by concepts like TIME IS MONEY or THEORIES ARE BUILDINGS." (10-11) Levin is presumably likewise unaware that when he writes of a reader "negotiat[ing a] poet's linguistic straits," (141) he is expressing his understanding of the reading process through two fundamental and well-attested metaphors, THOUGHTS FOLLOW PATHS (a subset of LIFE IS A JOURNEY), and POETIC MEANING IS DEEP (from LANGUAGE IS A CONTAINER). But Levin's conscious awareness or unawareness of his own understanding is not evidence bearing on the correctness of this metaphorical analysis, any more than a native English speaker's lack of conscious awareness that he or she cannot question parts of conjoined pairs (*Who did I see you out last night with Bill and?) is evidence bearing on the correctness of that linguistic analysis. Explicit awareness of our human understanding has very little if anything to do with how we represent and characterize that understanding. Hence the issue for competing analyses of metaphor, as of all linguistic phenomena, is not whether we are more aware of one reading than another, but which analysis is supported by better arguments.

Levin insists that "My time is precious" is not metaphorical because that usage, originally catachretic (he asserts), is now completely lexicalized (in his sense) and hence is no longer deviant -- i.e., it does not "result in the need to project novel conceptions," (12) and a figurative expression produces no tension between itself and a term "normally used in that context." (10) But how do we determine what is novel? How do we determine "normally used"? How do we determine the tension Levin takes to be crucial to metaphor? How do we determine when we need no longer "project novel conceptions" for an expression?[3]

In the end, then, Levin's theory of literary metaphor as deviant sequences, like all such theories, is subject to the same problems that Lakoff has documented (1987) for the classical categories of modal logic and compositional semantics. The canons distinguishing "deviant" from "non-deviant" sequences and literary metaphor from other kinds of metaphor, are Out There, foundational, determined by necessary and sufficient conditions, based upon academic literary experience rather than situated in everyday bodily experience. The necessary consequence of this approach, like the more specifically phenomenological approach of Stambovsky's The Depictive Image, is a theory of metaphor that is closed and hermetic.

Stambovsky purports to offer a phenomenological theory of metaphor. For him, the "literary experience" in which metaphor is situated is perceptual not conceptual (66); metaphor arises from "prereflective apprehension" (44) characterized by the "presentational immediacy" (8) that in his view informs all of literary art. "Depictive literary imagery," Stambovsky writes, "...articulat[es] perceived relationships (between objects, people, ideas, or feelings) in a presentationally immediate form that permits the literary reader to assimilate them `whole'...." (8)

Stambovsky is at pains to differentiate literary experience from

literary analysis, which he regards as secondary. Too many studies of metaphor, he implies, confuse the analysis of metaphor, a discursive process, with the experience of metaphor, a "prediscursive," immediate experience that is part of an integrated field of meaning, the "world" of the poem ( hardly a revolutionary formulation). In particular, he opposes "atomistic grammatical, linguistic, or semiotic" (or rhetorical) analysis of literary meaning because, he claims, the "rules of construal" used in these analyses are "internal to grammatical or linguistic or semiotic frames of meaning and not necessarily to those frames of meaning inherent in any given piece of literary art." (47) What these inherent frames of literary meaning really are, Stambovsky does not say. He insists that "aesthetic perception in its primitive, living immediacy is the primary (if not the only) epistemic modality of literary experience." (48)

The few actual literary analyses in The Depictive Image (particularly

the discussion of Emily Dickinson's "The Bird Came Down the Walk") are acute. But they occupy only twenty pages or so of a short book. The rest is for the most part an often opaque account of the philosophical literature and other secondary sources. Stambovsky implies a disdain for studies of metaphor that are merely "concerned with what metaphor is and what it does" (6); this view may account for the book's silence on many issues central to a theoretical treatment of metaphor: What is metaphor? What is metaphor's role in language? What is its role in semantic change? How do we typologize metaphors? What are the constraints on metaphor? Moreover, in insisting upon a particularly literary experience for literary metaphor, Stambovsky, like Levin, implies that metaphor in literature is different from nonliterary metaphor, a position that resurrects the ancient and unsustainable claim for a special literary language. Stambovsky's appeal to the elusive notion of the "literary experience" as justification for literary metaphor's independent status weakens the force of his claims for and about metaphor. The important context for literary metaphor and any theory of it is not "narrative or poetic context" but the "living experience" that, on Stambovsky's own account, metaphor communicates.

Only one of these books, Ann and John Thompson's Shakespeare: Meaning & Metaphor, focuses upon the metaphorical habits of a single writer. More such studies are needed. But this one seems to fall between two -- or among several -- stools. Each chapter of this series of five more-or-less free-standing essays (plus an introduction, an afterword, and a valuable bibliography) takes a different topic from the Shakespearean canon and applies to it a different theoretical approach: the Lakoff-Johnson paradigm to time metaphors in Troilus and Cressida; the little-known semantic field theory of Eva Kittay and Adrienne Lehrer to animal metaphors in King Lear; the metaphorical theory of Group m to body metaphors in Hamlet; J.F. Ross's analogy theory to Sonnet 63, and the views of Donald Davidson on metaphorical meaning to the metaphors of books and printing that abound in Shakespeare's work.[4]

The result is something of a hodge-podge. Because the Thompsons do not give a critical or comparative account of any of these theories -- which vary widely in quality and appropriateness -- the book cannot be read as a metatheoretical treatise on metaphor. Likewise, the Thompsons' theoretical eclecticism prevents them from tracing particular metaphorical habits through the plays and poetry (a project that might produce very interesting and novel results, in light of the recent theoretical advances in metaphor chronicled here). This book contains some new and occasionally penetrating readings of particular passages (the authors are quite good on the PEOPLE ARE BOOKS metaphor, pp. 165-70). It has value as a somewhat idiosyncratic survey of certain modern theories of metaphor and their applications to one author. But the overall effect is more that of a collection of essays by different hands (the seams of authorship frequently show) than a sustained theoretical or literary argument. Precisely this kind of fully articulated, consistent theoretical discussion typifies the three books that center on cognitive metaphor.

Of these, the clearest and most comprehensive introduction to the range of issues for cognitive metaphor is Johnson's The Body in the Mind, a book written with elegance, learning, and passion. For Johnson, the theory of metaphor entails not only a theory of language but a theory of understanding in its deepest sense. "[M]eaning and rationality," he concludes, "are grounded in recurring structures of embodied human understanding" (209). On this account, metaphor is not deviant or "special" language; metaphors are not derived from or parasitic on "ordinary" language; they cannot be explained by reduction to a propositional literal core; metaphor is not decorative (arguably, on this theory, it is not even a "figure of speech," but speech -- that is to say, language -- itself); metaphor does not depict or express understanding or experience. Rather, Johnson argues compellingly, metaphor is the basis of understanding, and enables us to structure experience.

For Johnson, the source of metaphor is a range of physical experiences from which we extract image schemata, preconceptual gestalt structures that are "recurring, dynamic pattern[s] of our perceptual interactions and motor programs that give... coherence and structure to our experience" (xiv). Metaphorical projection consists in our mapping these nonpropositional schemata, their structure and their components from our physical experience onto our nonphysical, abstract experience. The understanding achieved through metaphor is shared -- everyone in a particular culture has roughly the same bodily experiences -- and the image schemata that construct that understanding "constitute a large part of what we mean by form itself in our experience." (206)

This account, then, characterizes, for example, the immense number of BALANCE metaphors in our culture as follows. From an early age we learn physically about bodily balance: we learn to walk, we carry things, we lean over and pick up objects, we fall down, etc. We understand these repeatable and structurally similar bodily experiences, and their meaning for us emerges, Johnson argues, as we construct a preconceptual (and in this case non-visual) image schema that contains the general concomitants of balancing (equilibrium, symmetry, bodily homeostasis, etc.). We then map this image schema and its structure from the source domain of this repeatable physical experience onto the target domains of, for example, our ethical beliefs (notions of fairness and equity), our thought patterns (the mathematical notion of equivalence), even our aesthetic experience (the phenomena of symmetry in the visual arts, or resolution in such disparate artistic structures as literary plots, sonnets and sonata form). Precisely the structuring of experience entailed by this theory of cognitive semantics is, for Johnson, the creative function of metaphor.

The centrality of bodily experience in this theory explains such facts as the dependence on the horizontal plane of the balance metaphors in our language about law and justice. For example, we say that " the evidence weighs in favor of the plaintiff," that a witness's testimony is "biased," or that a jury is "leaning" one way or the other in its deliberations. Each of these metaphors (and the many others like them) depends upon the image schema associated with our folk theory of justice, which includes our cultural icon for justice, a blindfolded woman holding an old-fashioned two-pan scale. In these metaphors, the scale is out of balance in the horizontal plane.[5] Although we do have metaphors of verticalness (MORE IS UP, exemplified in "the stock market rose 5.17 points today"), we have almost no metaphors that depend upon vertical balance, upon some kind of stasis between flying and falling. This gap exists because we have no bodily experience of self-propelled flight from which to project an image-schematic structure onto more abstract, non-physical events. Indeed the very notion of vertical balance is hard to characterize -- hard to understand -- because we have no "geography of human experience," in Johnson's formulation (xxxvii) within which to comprehend that balance. Precisely this lacuna in our metaphorical lexicon is what gives such peculiar force and mystery to Gerard Manley Hopkins's picture, in "The Windhover," of a kestrel hovering on a windy day: "in his riding / Of the rolling level underneath him steady air...." The kestrel is balanced between flying and falling, but Hopkins must comprehend it within a source domain, horseback riding, of which we have real or vicarious bodily experience. If birds could write poetry or humans could fly, the structure of our metaphors would be different, a formulation comprehensible only within an experiential theory of metaphor.

This synchronic evidence for a cognitive theory of metaphor is supported as well by evidence from the history of language. Johnson cites a rich array of research by Eve Sweetser showing that an important source for semantic change is metaphorical projection from bodily experience to more abstract domains. Vision, Sweetser points out, is our primary source of data from early infancy; with the human sense of sight (unlike, for example, our sense of smell) we can focus at will on various features of what we see; what different people see in a given situation is more or less identical if they have the same point of view, so we have "shared public knowledge" of what we see. Hence it is no accident that words having to do with physical sight came to be used metaphorically for knowledge ("I see your point"; "He has no insight into his own problems," etc.) and equally no accident that we have few if any words having to do with knowledge or understanding based on the human faculty of hearing. As an image schema, hearing lacks the structure and components of vision that can be mapped onto understanding: our hearing is much harder to focus and while it can select, it does so with greater difficulty than vision. Both abilities are elements crucial to our understanding.[6] Even the modern locution "I hear ya," meaning roughly "I understand," lacks intellectual conviction and connotes at most a fuzzy emotional sympathy precisely because the faculty of hearing appeals chiefly to intercommunication. Hence the sense is of "I hear ya" is almost that of a powerless phatic communion -- "I sympathize, but there really isn't much I can do except listen."

The paucity of metaphors projected from vertical equilibrium and hearing to more abstract domains points up two related and important issues in metaphor that this theory elegantly solves: groundedness and objectivity. Because metaphor (and hence human understanding) is grounded in repeatable bodily experience, we can have no metaphors by which we understand abstract concepts that depend on a bodily experience that we have not had, directly or indirectly, or that posit a connection between a physical experience with a particular structure and an abstract experience that lacks the crucial elements of that structure. "[A]t the very least," Johnson asserts, "image-schematic structures and their metaphorical projections have a shared, public character which gives them a central role in the objectivity of meaning. By `objectivity,' of course, I mean that there are meaning gestalts connected to structures of bodily experience that we all can share." (175)

Thus for Johnson not only metaphor, but understanding, reason, and imagination -- intellectual activities that perceive new experiential structures and generate new understanding from them -- "must be understood ... as an interaction of a human organism with its environment (which includes its language, cultural traditions, values, institutions, and the history of its social community)." (209) The objectivity of cognitive metaphor is not guaranteed sub specie aeternitatis, after the manner of classical category theory, but in the flux of the human condition.

The Body in the Mind contains much else that is of interest primarily to philosophers -- Johnson's theory of imagination, and its derivation from Kant; his disputes with Objectivism; the details of his theory of meaning and understanding. Its importance for students of poetics resides in the broader philosophical context within which Johnson embeds the theory of metaphor that he and George Lakoff first articulated in Metaphors We Live By[7]. This context is crucial to the theory of literary metaphor that underlies Mark Turner's deep and provocative study of kinship metaphors, Death is the Mother of Beauty: Mind, Metaphor, Criticism.

Turner seeks to account for the ubiquity and potency of the metaphors typified in such passages as (from Keats's Endymion, I, 783-6):

hist, when the airy stress
Of music's kiss impregnates the free winds
And with a sympathetic touch unbinds
Æolian magic from their lucid wombs....

Likewise, Turner's theory seeks to explain why we do not produce kinship metaphors like "*Sable night, father of dread and fear" and "*God of our mothers, Whose Almighty hand Leads forth in glory All the starry band." Finally, Turner seeks to account for the deeply explanatory power of metaphor in our culture. He best formulates this power in discussing science:

Science models systems so that we can recognize, explain, and predict them. The conceptual metaphors implicit in our language are a kind of science. It might be said that genetic laws of transmission or the sociobiological principle of inclusive fitness make sense out of metaphors in language based on metaphoric inferences like inheritance or functional property transfer, but the influence really works in the reverse order: it is the earlier formulation of these sciences in the metaphors of thought and language that made their scientific formulations seem conceptually so natural. (194)

Turner accomplishes these goals in a study that is exhaustive, richly documented, finely articulated, and extraordinarily broad in the range of knowledge and literary examples that it brings to bear on these questions. He is at his best and most relaxed when he discusses literary texts, in particular an obscure passage from Blake and Book 2 of Milton's Paradise Lost ; the highly structured theoretical discussion elsewhere in the book can be hard to follow.

The fundamental task of this book is to explain how we understand a rich domain of literary expression. In that understanding, as in our command of language, we make infinite use of finite means. Those finite means, Turner argues persuasively, consist of seven basic kinship metaphors coupled with ten ways of interpreting those metaphors he calls "metaphoric inference patterns." "Some combination of these conceptual inference patterns," Turner concludes, "accounts for every one of the indefinitely many specific kinship metaphors in our language." (195)

An example of a basic kinship metaphor is THE WHOLE IS THE MOTHER OF THE PARTS, which explains why Russians speak of Russia as the "motherland" (as long as there has been a Russia it has had restive parts) and why Germans speak of a united Germany as the Fatherland (for Germans, whether or not the nation has been divided at a particular historical moment, its unity has always been both a goal and an instrument, and father-metaphors focus on instrumentality). An example of a metaphorical inference pattern is place and time as parent, which gives us an interpretational template for "Babylon is the mother of harlots and abominations," a metaphor that derives from the basic kinship metaphor AN ABSTRACT PROPERTY IS THE PARENT OF SOMETHING HAVING THAT PROPERTY. Because we have a folk theory that parents and children share traits of character, we infer from the assertion that Babylon's children are harlots and abominations that Babylon is a harlot and an abomination; the metaphor shares this metaphorical inference pattern with the very different line, "Deep with the first dead lies London's daughter."

Of these metaphorical inference patterns, the two most important, for Turner, are causation as progeneration (confusingly, a key conceptual metaphor on this account is CAUSATION IS [note: not as] PROGENERATION), to which he devotes an entire chapter, and lineage in the world, mind, and behavior, by which, for example, poets often depict feeling as the parent of behavior ("Fear, father of cruelty"). These two patterns working together, Turner shows, "account for the principal use of kinship metaphor, namely to express paths by which things in the world, the mind, and behavior can spring from one another" (143).

Some metaphors depend upon what cognitive semanticists call Idealized Cognitive Models of kinship, They include metaphors like "the true child of vanity is violence" and "stench, diseases, and old filth, their mother," which are interpretable using the inference pattern causation as progeneration. In metaphors that partake of these Idealized Cognitive Models, the cause corresponds to the parent and the effect to the child; for a metaphor to be interpretable under this inference pattern, a number of conditions must hold: cause and effect must be able to be personified, for example; they must be durative, cohesive, and individuated. These requirements explain, for example, the fact that for particular phenomena we have metaphorical parents/causes that are individuated notions like night or age rather than the diffuse, unknowable, unpersonifiable biological processes of reproduction, and that we use as source domains for metaphors of anger, relatively simple notions like heat and pressure rather than Freudian concepts of displacement[8].

In contrast to the rich, creative kinship metaphors that involve the causation as progeneration inference pattern are those interpreted through mere similarity. In creative metaphors, Turner argues, we reconceive the target domain, with the result that these metaphors create meaning. In similarity metaphors we merely seek to match concepts. Turner's treatment of this process, though as highly articulated as his discussion of causation, is much clearer and gives greater insight into the real virtues of his method.

When we interpret metaphors depending on similarity, the question is "which connections to look for and which to ignore." (185) Our first move is to seek to match the stereotypical behaviors or operations of the two concepts in, for example, "Music is twin-sister to poetry." These concepts have the same "stereotypical components" and differ only in material. In metaphors where the concepts are not equally behavioral, such as (written of the linnet) "My dazzled sight he oft deceives, A brother of the dancing leaves," we map from the more articulated behavior of the glancing light and rhythmic movement of the leaves to the less fully articulated concept of the bird -- like the leaves in the wind, the linnet darts and moves in and out of the sunlight.

What this chapter amounts to is a handbook on the process of interpreting metaphors within Turner's theoretical framework. The book is much more successful here and in the practical criticism of individual literary texts in Chapter 3, particularly in Turner's treatment of the crucial kinship relations and the meaning their metaphors create for Milton's account of Satan, Sin and Death in Paradise Lost. The theoretical chapters show flashes of brilliance and conceptual breadth, but the complexity of the theoretical framework often exceeds the benefit of the insights that framework produces. A more open terminological network might have captured fewer generalizations, but would have enhanced the comprehensibility of what it characterized.

Perhaps the occasionally fierce density of Death is the Mother of Beauty is the fate of highly theoretical books (twenty-five years ago, Chomsky's Aspects of the Theory of Syntax wasn't nightstand reading either); Turner's more recent collaborative volume with George Lakoff, More than Cool Reason, is broader in scope, more practical, less intensely theoretical, and -- possibly as a result -- likely to be the standard work in metaphor for some time to come.

As might be expected from a book subtitled A Field Guide to Poetic Metaphor, More than Cool Reason sets out as theoretical background much of the apparatus in The Body in the Mind and Death is the Mother of Beauty. What seem crucial in this later book are four points: clear accounts of specific metaphors; the theoretical derivation of what Lakoff and Turner call "conceptual compositions"; how the theory of cognitive metaphor makes possible "second-order global readings" (what, fifteen years ago, I called "iconic syntax") of particular poems (they produce a brilliant analysis of William Carlos Williams' "To a Solitary Disciple"); and a powerful demonstration of how an important cultural model, the Great Chain of Being, explains an important quasi-literary form, the proverb, and, more important for our purposes, proverb-like poems and passages.

Lakoff and Turner give a particularly clear and rigorous account of the metaphors centering on the central concepts of life, death, and time. They show, for example, how the LIFE IS A JOURNEY metaphor entails particular correspondences between the source domain of the journey and the target domain of life: the person leading a life is a traveler; his purposes are destinations; the means for achieving purposes are routes, etc. (3). This metaphor is shown to entail metaphors such as BIRTH IS ARRIVAL and DEATH IS DEPARTURE and to give more satisfying explanations of poems like Emily Dickinson's "Because I could not stop for Death" and Frost's "The Road Not Taken." Other life-related metaphors (PEOPLE ARE PLANTS, A LIFETIME IS A YEAR, DEATH IS REST, and several others) are invoked to give an exceptionally rich, suggestive reading of the complex metaphorical structure in Shakespeare's Sonnet 73, "That time of year thou mayst in me behold."

This reading shows, among other things, that the concept of "imagery" is insufficient to explain the peculiar power of one quatrain in this much-studied sonnet:

In me thou seest the glowing of such fire,
That on the ashes of his youth doth lie,
As the deathbed whereon it must expire,
Consumed with that which it was nourished by.

Our experience with hearth-fires structures our understanding of this passage. If we have had no such experience, we need the explanatory footnote that editors of this sonnet usually provide. Non-imagistic bodily experience of a highly articulated sort thus is projected onto the abstract notions of love in late life. If the structure of that source domain (fires) lacks the structure of the target domain (death, and the nature of love-relationships as death approaches), the analysis fails. The theory of cognitive metaphor thus has a built-in constraint -- the requirement that there be a degree of structural isomorphism between source and target domains. This fact is important in a time when the term "theory" is loosely used of any form of literary analysis. Cognitive metaphor and the theory of embodied human understanding produce analyses that are capable in principle of being wrong (and hence of being right); the theory provides a basis for choosing among the characterizations it produces.[9]

Similar theoretical consequences are entailed by the Lakoff-Turner derivation of "conceptual compositions" (72) like personification. Personification, they show, is not just another figure of speech, and not identical to simplex metaphor. Rather, various kinds of personification (a type of which its various actualizations are tokens) are derived by independently motivated "everyday" simplex metaphors that combine in particular patterns. For example, not all personifications depend on persons in the source domain. While the LIFE IS A JOURNEY metaphor maps from a traveler -- a person -- in the source domain of the journey to a person in the target domain, in the phrase "the Grim Reaper," we understand as a person (the reaper) something that is not a person (Death). We compose the personification from two basic metaphors: PEOPLE ARE PLANTS (independently motivated because it is required in order to explain such lines of poetry as "My way of life is fallen into the sere, the yellow leaf") and EVENTS ARE ACTIONS (independently required to explain -- or begin to explain -- lines like "Do not go gentle into that good night", where the event of death becomes an action with an agent). By PEOPLE ARE PLANTS, people are understood as crops that grow and someone must harvest; by EVENTS ARE ACTIONS, the event of death is understood as an action caused by an agent; by their combination, the action is harvesting and the agent is the reaper. Analysis of this sort can be falsified not only if the structure of the proposed source domains is not to be found in the structure of the target domains, but also if the metaphors thought to form the composition are unique to this particular derivation -- if, in short, they are not independently motivated.

This concern with independently motivated metaphors as the basis for analysis is the most convincing aspect of the Lakoff-Turner "second-order global reading" of syntactic patterning in Williams' "To a Solitary Disciple." What they mean by "second-order global reading" arises where Williams' speaker instructs the disciple of the poem's title to observe that the lines of the church steeple being contemplated in the poem rise above the abstract outline of the church building itself to surround the moon:

Rather grasp
how the dark converging lines
of the steeple
meet at the pinnacle--
perceive how its little ornament
tries to stop them--

See how it fails!

The value of this notion of a syntactic metaphor is evident in a variety of other contexts. It seems, for example, a better way to characterize the iconic syntax in poetry about which I have published a number of papers.[10] The most fully developed of these readings (Freeman 1978), traces a strongly causative-inchoative pattern of verbs in the first stanza of Keats's "To Autumn." But merely tracing the pattern does not suffice to establish the meaning claimed for it (that Keats sought to make the Sun the ultimate causer of natural process in the poem's world). What does establish a meaning for this syntactic pattern is to see it as constituting the source domain for the independently motivated metaphor CLOSENESS IS STRENGTH OF EFFECT (discussed in Lakoff & Johnson 1980:128-132; in effect, they argue that a sentence like "John is unhappy" depicts stronger feeling than "John is not happy" because the syntax of the first puts the negative unit closer to the negated adjective. See Freeman 1989 and, for a fuller account of iconic syntax from a purely linguistic point of view, Haiman 1985). The causes and effects of Autumn's agency are built into the verbs ("Conspiring with him How to load and bless with fruit the vines that round the thatch-eves run" -- i.e., how to CAUSE the vines to become laden [effect ---> load] with fruit). There is a mapping between the structure of the stanza's images (repeated intense burgeonings) and the structure of its language (repeated expression of causative-inchoatives in one word). The CLOSENESS IS STRENGTH OF EFFECT metaphor bridges the epistemological gap between pattern and meaning that was problematic in this earlier work.

The same kind of explanatory power is convincingly demonstrated in Lakoff and Turner's extended discussion of how proverbs and proverb-like poetry can be explained by the interaction of one pervasive metaphor, GENERIC IS SPECIFIC, the conversational Maxim of Quantity ("be as informative as is required and not more so"), a folk-theory of forms of being (that their attributes cause their behavior), and a nearly universal cultural model, the Great Chain of Being. These principles applied to a famous proverb-like line from King Lear, "How sharper than a serpent's tooth it is To have a thankless child" ( I, IV, 279-80), can yield a particularly rich reading.

The folk-theory we have of snakes is that they bite (in fact, most don't strike and those who do don't bite and have fangs, not teeth, but folk theory has little to do with herpetology), and that they slither and do not walk (for that reason, they are at the bottom of the Animal segment of the Great Chain, for they are in that respect least like animals). We have some highly specific folk-knowledge about the role of the serpent in the Garden of Eden. We also know that snakes have scales, that they shed their skins, etc. This irrelevant information is ignored by the Maxim of Quantity, "be as informative as need be, but not more so," which picks out from our folk-theory about snakes only the highest-level information about them, that they bite and that one of them caused Eve to eat the the fruit of the Tree of Knowledge of Good and Evil. Our snake-lore is part of the source domain. Our folk-theory of the Nature of Things says that an attribute of snakes, their teeth, causes them to bite and inflict pain.

The GENERIC IS SPECIFIC metaphor picks out the relevant information from the Great Chain cultural model as it structures the target domain, King Lear, his feelings, and his daughter Goneril. Lear is a man, a father, and a king. His daughter, below him on the Great Chain in all three of these respects, owes him deference as a man, gratitude as her father, and allegiance as her king. Her ingratitude in seeking to strip him of his retainers makes it clear that she will give him none of these things. She has, in three ways, violated the Great Chain and caused her father emotional pain. Intense physical pain is caused when serpents bite humans (to say nothing of when a serpent frustrates God's will for man by causing his fall). By the independently motivated metaphor EMOTIONAL PAIN IS PHYSICAL PAIN and the cultural model of the Great Chain, we understand Goneril's ingratitude to Lear as the bite of a snake, as a triple violation of the moral order of the universe as epitomized in the Great Chain,[11] and as a reenactment of the Fall of Man.

It will be contended that when we read that line we extract the Biblical allusion and the Great Chain information at once, and hence do not require all the machinery of the GENERIC IS SPECIFIC metaphor, the folk-theory of the Nature of Things, and the Maxim of Quantity. This objection partakes of the same mistake that Levin makes when he insists that because he is not aware that a particular metaphor structures his understanding of a particular formulation, the analysis positing that metaphor is incorrect. Lakoff and Turner are not proposing a production model but an explanatory model -- that is, a theory of metaphor. And a good theory must be explicit about its structural descriptions and the operations that are based on those descriptions -- here, what particular pieces of information are relevant to the operation of the Great Chain cultural model and how these are established. The rigor with which Lakoff and Turner present this kind of analysis is, again, the book's greatest strength. It is readable and contains many useful insights. But above all, it demands in its analyses a theoretical explicitness that in my experience is unprecedented in literary study.

When we assess a theory of metaphor, the final judgment depends on more than just that theory's elegance and force. Metaphor is crucial not only in our verbal artworks, but in our lives. Hence the deeper question is whether a particular theory is intuitively satisfying, whether it makes the kind of judgments that seem both right and relevant not just to literature but to all our uses of metaphor. The analyses yielded by the cognitive approach to metaphor reflected in these books seem to me to speak precisely to our intuitions about language and poetry in the first instance and, more deeply, to our sense of what it means to be human. Because it is grounded in embodied human understanding, cognitive metaphor as a theory explains in the same way that tonal music pleases and delights. Because tonality is grounded in the physics of the natural universe, it embodies the music of the spheres. Construed as human understanding that is constructed from our physical experience, metaphor arises, as the composer-conductor Leonard Bernstein once remarked about our reception of tonal music, from das Lied von der Erde[12].

Davidson, Donald, 1978. "What Metaphors Mean," Critical Inquiry 5, 31-47.

Ellis, John M., 1989. Against Deconstruction (Princeton: Princeton UP).

Freeman, Donald C., 1975. "The Strategy of Fusion: Dylan Thomas's Syntax," in Style and Structure in Literature, ed. Roger Fowler (Oxford: Blackwell), 19-39.

1976 "Iconic Syntax in Poetry: A Note on Blake's `Ah! Sun-Flower!'," University of Massachusetts Occasional Papers in Linguistics, 2, 51-7.

1978 "Keats's `To Autumn': Poetry as Pattern and Process," Language and Style, 11, 3-17.

1986 "Syntax, Agency, and the Imagination: Keats's `Ode to Psyche' and `Ode on a Grecian Urn'," in Linguistics and the Study of Literature, ed. Theo D'haen (Amsterdam: Noordhoff), 70-88.

1989 "Linguistics and Cognitive Metaphor," Conference on Linguistics and Literary Study, University of Winnipeg, October, 1989.

Group u, 1970. Rhétorique générale (Paris: Larousse).

Haiman, John, 1985. Natural Syntax: Iconicity and Erosion (Cambridge: Cambridge UP).

Kittay, Eva & Adrienne Lehrer, 1981. "Semantic Fields and the Structure of Metaphor," Studies in Language 5, 31-63.

Kövesces, Zoltán, 1986. Metaphors of Anger, Pride, and Love: A Lexical Approach to the Structure of Concepts (Amsterdam: John Benjamins).

Lakoff, George, 1987. Women, Fire, and Dangerous Things (Chicago: University of Chicago Press).

Lakoff, George & Mark Johnson, 1980. Metaphors We Live By (Chicago: University of Chicago Press).

Ross, J. F., 1981. Portraying Analogy (Cambridge: Cambridge UP)

Sweetser, Eve, 1990. From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure and Semantic Change (Cambridge: Cambridge UP).

[1]Levin limits this characterization of the "metaphoric worlds" of the book's title to what one might call "sublime" poetry (he focuses on Wordsworth's The Prelude -- see p. 3). but it seems to me that on his theory this restriction does not matter: the theory stands or falls with respect to poetic metaphor as a whole.

[2]In some respects this facet of Levin's theory is reminiscent of Lakoff's proposal during the generative semantics era of "world-creating" verbs, in which the operation of certain syntactic rules is suspended, e.g. "I dreamed that I was Brigitte Bardot and that I kissed me [but not *myself]". I have been unable to trace any publication of this example; I know it only in the oral tradition of that era.

[3]I believe that Levin also misreads Lakoff-Johnson on the ancient literal-figurative language dispute. He quotes some of their discussion of the conceptual metaphor ARGUMENT IS WAR, and appears to catch them in an inconsistency: "the authors say: `the language [of argument] is metaphorically structured'. [Yet in the very next paragraph] they say that the language of argument is literal. The only apparent way to achieve consistency between these two characterizations is to assume that by the first formulation they mean not that the language used in argument is metaphorical but that lying behind and `structuring' that language is the metaphorical concept. This leaves us with the claim that the actual language in which argument is conducted is in fact literal." (7) Lakoff and Johnson's point there (amplified by Johnson in The Body in the Mind, and by Lakoff and Turner in More than Cool Reason, both discussed below) is simply to reject the distinction between figurative and "literal" language, and they do so in the very passage Levin quotes: "the metaphor [ARGUMENT IS WAR] is not merely in the words we use -- it is in our very concept of an argument. The language of argument is not poetic, fanciful, or rhetorical; it is literal." (Lakoff & Johnson 1980: 5)

[4] See Kittay & Lehrer 1981, Group u1970, Ross 1981, and Davidson 1978.

[5]While in the image schema evoked by "biased," for example, one pan of the scale is higher than the other, the structural aspect of this source domain that is mapped into the target domain is that the plane created by the imbalance is horizontal but not parallel to the "just" plane of the icon.

[6] This discussion is a summary of part of Chapter 2 of Sweetser:in press.

[7]Lakoff 1987, not under review here, provides a similar context for the theory of metaphor in linguistics and cognitive psychology. Whether by coincidence or design, these books are companion pieces.

[8]For the most recent full treatment of particular metaphors, see the treatment of anger, pride, and love in Kövesces 1986.

[9]For discussion of the best-known modern literary "theory" and how it fails these (and many other) tests of a theory, see Ellis:1989.

[10] See Freeman 1975, 1978, 1986.

[11] A point made again in Act IV, when, after Albany discovers that Gloucester's eyes have been gouged out, he reproaches Goneril:

If that the heavens do not their visible spirits
Send quickly down to tame these vile offenses,
It will come,
Humanity must perforce prey on itself
Like monsters of the deep.