Turner Metaphor Invention

8/2/2019 Turner Metaphor Invention

1/21

Porter Institute for Poetics and Semiotics

Poetry: Metaphor and the Conceptual Context of InventionAuthor(s): Mark TurnerReviewed work(s):Source: Poetics Today, Vol. 11, No. 3 (Autumn, 1990), pp. 463-482Published by: Duke University PressStable URL: http://www.jstor.org/stable/1772822 .

Accessed: 17/04/2012 23:33

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

Duke University Press and Porter Institute for Poetics and Semiotics are collaborating with JSTOR to digitize,

preserve and extend access to Poetics Today.

http://www.jstor.org
http://www.jstor.org/action/showPublisher?publisherCode=dukehttp://www.jstor.org/stable/1772822?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/1772822?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=duke


2/21

Poetry:Metaphorandthe ConceptualContext of InventionMark Turner

English, Maryland

Poetic invention is an amorphous subject, not much discussed in con-temporary criticism. Invention seems to be either precritical or beyondanalysis, yet our practice of making critical distinctions about inven-tion, as when we judge a poet to be strikingly inventive, implies thatpoetic invention should be susceptible of analysis. What do classic andsuccessful inventions tell us about particular formal and public acts ofliterary representation and the tacit and private preliterary way weunderstand our selves and our experience?Invention is not originality. Only since the Romantics have thesetwo cognitive phenomena been typically confused. The structure ofinvention, and of particular classic and successful literary inventions,may be wholly unoriginal or may have a dominant unoriginal aspectthat serves as the ground upon which contingent originality plays. Weare vigilant for the new and the variable and concentrate on it. Con-sequently, our consciousness is blind to the unoriginal, which we taketo be merely background.

I would like to thank Frank Thomas for assistance with this essay, Mark Johnsonfor developing in his publications the theory of image-schemas which I use here,and Gilles Fauconnier, Ronald Langacker, Kathy Harris, Jeff Elman, Liz Bates,David McNeill, John Goldsmith, Jim McCawley, Nancy Dray, Peter Dembowski,and David Palermo for conversations about this essay. The rudiments of the con-straint I give in this essay are implicit but not analyzed in Lakoff and Turner (1989)and Turner (1987).Poetics Today 11:3 (Fall 1990). Copyright ? 1990 by The Porter Institute for Poeticsand Semiotics. ccc 0333-5372/90/$2.50.


3/21

464 PoeticsToday 11:3Of course, the unoriginal is not background, at least not in the senseusually conveyed by that description. The unoriginal is normally the

dominant active matrix in any original achievement. Originality is nomore than the exploitation of what is unoriginal. Originality, far frombeing autonomous, is contingent at every point upon the unoriginalstructures that inform it. When we step into a room, or into a poem,we do not have to think consciously "this is a room" or "this is a poem."The automatic nature of our interactions with the unoriginal leads usto think of the unoriginal as simple in itself and too simple for analy-sis, but this apparent simplicity is false. Relative to the complexity ofthe unoriginal conceptual context of invention, it is the original ininvention that is simple. The concept of a "room" or a "poem" is im-measurably more complex than the original aspects of any one roomor any one poem. Explaining the structure of an original momentis relatively easy, once we can explain the underbrush of unoriginalstructures it exploits, but that is hard. This essay, for example, willconsider one small, automatic constraint on invention. This constraintis a profoundly unoriginal component of invention and is routinelyexploited to achieve originality. But, far from being obvious, this con-straint seems to have gone unnoticed in poetic theory. Far from beingsimple, it shows a complexity that I can do no more than indicate here.The purpose of calling attention to an unoriginal aspect of inven-tion in poetic expression is not to homogenize nonliterary and literarypractice. It is rather to tease out what belongs to the poet, againstthe background of what belongs to poetry, against the background ofwhat belongs to language. The imagination must operate in a knownspace; it must work with unoriginal structures of invention. These arethe conditions that the imagination must meet in order to be intel-ligible. Originality is just a step away from pedestrian thought, andpedestrian thought accounts for most of the invention in any poem.A room is more pedestrian than the Sistine Chapel, but the inven-tion of the "room," which belongs to no individual, is beyond theoriginal inventive range of any individual architect, even Brunelleschior Michelangelo. This unoriginal concept informs every exceptionalroom, as it does every pedestrian room.Let us begin a demonstration of these propositions by turning toJohn Bunyan's Pilgrim'sProgress.In one component of its literary in-vention, Pilgrim'sProgress is unswervingly unoriginal: it is structuredby a conventional metaphoric understanding of life as ajourney. Thatconventional and wholly unoriginal metaphoric understanding is con-1. One room can differ from another in an almost unlimited number of unorigi-nal and pedestrian ways, all of which we know without need of reflection. This isan indication of the complexity in a full knowledgeof the concept "room."


4/21

Turner* Metaphorand Invention 465strained in certain ways which are derived from a general constraint,one that, I claim, applies to all metaphor, original and unoriginal.Pilgrim's Progress never requires of itself or its readers any work tomeet those constraints. The conventional metaphoric understandingof life as a journey already conforms to those constraints, and Pil-grim'sProgresssimply inherits those prefabricated and quite unoriginalsatisfactions.2Let us look at a few of these unoriginal satisfactions. A metaphoris a mapping of a source conceptual schema3 (such as our concep-tual schema for journey) onto a target4 conceptual schema (such asour conceptual schema for life). For example, in the conceptual meta-phoric understanding of life as a journey, components of the schema2. Pilgrim's Progress fleshes out this conventional metaphor elaborately, and ex-ploits in concert with it many other conventional metaphors, such as THE MIND ISA BODY, STATES ARE LOCATIONS, and GENERIC IS SPECIFIC. (For a discussion of thesemetaphors, see Lakoff and Turner [1989].) Many of the allegorical moments inPilgrim's Progress do not appear to be extensions of any conventionalized metaphor.See, for example, the Interpreter's presentation of a parlor as the heart of a mannever sanctified by the grace of the gospel, the dust in the parlor as his original sinand inward corruptions, the man who sweeps the dust as the law, water as grace,and the damsel that sprinkles water on the dust as gospel. The effect of law on sinin the heart is to be understood in terms of the effect of the sweeper on dust inthe house: he merely stirs it up to worse effect, unless water is first sprinkled uponit. Since neither Christian nor the reader possesses a conventionalized conceptualmetaphor with this set of mappings, both must rely upon the Interpreter to indi-cate it to them, or hazard what they know to be guesses at the significance of thesweeping scene.3. The notion of a schema will be used throughout this article. Schemata are skele-tal organizations of conceptual knowledge. They have variables, can embed, andcan represent knowledge at all levels. For a history of the development of the ideaof a schema, see Rumelhart (1980). Rumelhart discusses the history of the conceptof a schema, from its original sense in Kant (1787) through its use in psychol-ogy by Head (1920) and Bartlett (1932). The idea of a schema, or frame, will formost readers be familiar from a range of now-classic books and articles from the1960s and 1970s. See, for example, Fillmore (1968); Minsky (1975); Rumelhartand Ortony (1977); and Schank and Abelson (1977). See also Goffman (1974)and Goodman (1976 [1968]). The idea of a schema or frame or script is now acommonplace in work on cognition. See Fauconnier (1985); Fillmore (1985); andRumelhart, McClelland, and the PDP Research Group (1986: 20-22).4. "Source" and "target" are commonly used technical terms in discussions ofmetaphor. See, for example, the entry on "metaphor" in the Oxford International

Encyclopedia of Linguistics (forthcoming), in which the terminology is explained.The target conceptual domain is the domain to be understood metaphorically.The source conceptual domain is the domain in terms of which the target is to beunderstood metaphorically. A conceptual metaphor consists of a target, a source,and a mapping between them. Conventional metaphoric expressions draw theirvocabulary from the source but are taken to refer to the target, as when we take"I'm getting nowhere"-whose vocabulary is drawn from the source schema forjourneys-to refer to the state of someone's life, which is the target.


5/21

466 PoeticsToday 11:3for journey are mapped onto components of the schema for life: Theperson leading a life is a traveler; his purposes are destinations; themeans for achieving purposes are routes; difficulties in life are im-pediments to travel; counselors are guides; progress is the distancetraveled; and so on. The target schema, life, is understood in terms ofthe source schema, journey.In any such conceptual metaphor, the mapping is constrained tobe one-to-one: two distinct senses in the source are not mapped ontoone sense in the target, for that would destroy the identity of thatone sense in the target. In the conventional metaphoric understand-ing of life as a journey, this constraint against violating identity inthe target is automatically satisfied by a fixed mapping. For example,in the source schema, the traveler is distinct from the destination; inthe target schema, the person leading the life is distinct from a goal;and the conventional metaphor maps those two distinct senses in thesource onto those two distinct senses in the target. Pilgrim'sProgresssimply inherits this prefabricated and quite unoriginal satisfaction ofthe constraint against violating identity in the target.If we look at an extension of the conventional metaphoric under-standing of life as ajourney, we find that the constraint against violat-ing identity in the target still applies. For example, if we say, "I am atraveler in life and I am the destination," then we find ourselves con-strained to take the first and second "I" as pointing to two differentsenses (such as the knowing self and the self to be known) so that twodifferent senses in the source (traveler and destination) do not maponto one sense in the target.Now let us consider not the identity of senses in the target but ratherorder relations in the target. Here, too, we will find our metaphoricinvention constrained. First, let us consider some order relations in thetarget and in the source. We conventionally conceive of the momentsof life as being temporally ordered:5 for any two moments in life, onemust precede the other; no moment precedes itself; and precedenceis transitive. We also conventionally conceive of the points on a path,such as the path of ajourney, as spatially ordered: for any two spatialpoints on the path, one must precede the other; no point precedesitself; and precedence is transitive. And we conventionally conceive ofa traveler on ajourney as encountering the spatial points on the pathin the order of their physical succession. Now let us consider whathappens when we try to map ordered components of the source ontoordered components of the target.5. Technically, a relation < on a set A is called an order relation (or a simple order,or a linear order) if it has the properties of comparability (for every x and y in Afor which x does not equal y, either x


6/21

Turner* Metaphorand Invention 467When we understand life metaphorically as a journey, we are con-strained not to violate the order of moments in our concept of life.

We cannot without provoking remark say, for example, "First I wasgetting somewhere in life and then I got off to a good start," becausewe take this as asking us to violate the original temporal order of twomoments in the target: the prior moment in the target ("First I was")is forced to correspond to the later moment in the source ("gettingsomewhere"), and the later moment in the target ("and then I") isforced to correspond to the prior moment in the source ("got off toa good start"). This reverses the order of moments in our schema oflife-which is the target-and disturbs us badly.The conventional metaphoric understanding of life as a journeyautomatically and fixedly satisfies this constraint by mapping spatialpriority in the source onto temporal priority in the target, and Pil-grim'sProgress inherits and deploys this prefabricated and unoriginalsatisfaction of the constraint against violating order in the target.But there are texts that ask for originality in satisfying these unorigi-nal constraints. The Farewell Discourse (chapters thirteen throughseventeen) in the Fourth Gospel is a locus classicus whose exception-ally rich history of commentary wrestles with many of these demands.Before we turn to this text, it will be helpful to pause to make someframing and theoretical remarks.First, we must remark that it is worthy of wonder that any such con-straints even exist upon metaphoric invention, original or unoriginal.These constraints, once stated, may seem obvious, but that is onlybecause their automatic transmission fools us into thinking that theyare inevitable and require no explanation. There is nothing inevitableabout them, or rather, we have no explanation as yet of how they couldbe inevitable. They are constraints concerned with preventing the vio-lation of the target, but notice that they apply only to particular partsof the target; other parts of the target can be violated with impunity.Why are some parts of the target protected while others are not? Thatrequires an explanation, and the shape of the explanation is neitherobvious nor simple. Metaphoric thought is notoriously cavalier, andmetaphoric language is notoriously slippery: When we understandsome target concept metaphorically, we frequently violate or discardindispensable parts of it. Through metaphor, inanimate objects canbecome people ("Mycar refused to start this morning"), moods can be-come colors ("I'm feeling blue"), thoughts can become physical objectswhile minds become bodies ("I cannot grasp this notion"). The widelatitude of metaphor in thought and language often makes it seemunconstrained in its operation, a species of pure free play. Consider,for example, "Trees climb the hills toward the Golan and descendto test their resolve near the desert." Here, a static configuration ap-pears to be understood metaphorically as a dynamic movement, and


7/21

468 PoeticsToday 11:3the agentless event of dynamic movement appears to be understoodmetaphorically as an action by an intentional agent. We are not both-ered that dynamism is mapped onto stasis, that an action is mappedonto an event, or that intentional animate agents are mapped ontoplants. Why then are we bothered when, as in the case we previouslyconsidered, the ordering of a sequence is violated?Aristotle originally noticed that a metaphor is constrained not toviolate various things in the target. He expressed this by saying thatthe source must fit the target in certain ways, including what appearto be conceptual ways.6 Considering that the problem of fitness wasraised by Aristotle, and that metaphor theory has in many ways beena series of responses to and developments of Aristotle's few commentson metaphor, it is odd how little inquiry has been made into the actualdetails of what makes a metaphor conceptually fit. This topic has beenovershadowed by others in the theory of metaphor. There has beenvoluminous work on what constitutes metaphor and how metaphoriclanguage is demarked from other forms of language, on how it is that ametaphor can mean, and on whether or not metaphor can have truth-value, but relatively little work on the conceptual details of metaphor.76. Aristotle writes in The Rhetoric (1405a), "Metaphors, like epithets, must be fit-ting, which means that they must fairly correspond to the thing signified: failingthis, their inappropriateness will be conspicuous: the want of harmony betweentwo things is emphasized by their being placed side by side."There are many grounds on which a metaphor might be judged fitting or unfit-ting. As Wayne Booth points out (1979: 47-70), any rhetoric text from Aristotle'sto Whately's would comment that a metaphor might be judged appropriate orinappropriate in its grandeur or triviality to what is being presented, in its con-tribution to the ethos the speaker desires to establish, in its level of difficulty orinterest for the audience, in the style of its expression, and so on. I am concernednot with these forms of fitness but rather with the conceptual fitness of a metaphor.7. For a survey of such work, see Cooper (1987). Recent work on the concep-tual details of metaphor includes most prominently Dedre Gentner's (1983, 1986)work on structure-mapping, which I will discuss in footnote 13. Older work on theconceptual details of metaphor centers on the chestnut that metaphors are con-ceptually constrained not to be "mixed," but this claim is in general false; often,mixed metaphors do not bother us, and often, different metaphors can cohere (seeLakoff and Turner 1989: "Composing": 70-71, and "Coherence among meta-phors": 86-89). Wayne Booth (1979: 50) gives the following example of a "mixed"metaphor that does not disturb us at all, and that may be all the stronger for beingmixed. A lawyer expresses the struggle between a large utility company and asmall utility company in a figure: "So now we see what it is. They [the large utilitycompany] got us [the small utility company] where they want us. They holdingus up with one hand, their good sharp fishin' knife in the other, and they sayin','you jes set still, little catfish, we're jes going to gut ya."' Booth observes later in thearticle that consistency would require that the catfish be told to "hang still" ratherthan "set still," but that this conceptual inconsistency does not bother us. I suspectthat we would be rather bothered indeed if the metaphor were instead: "You jesvanish, little catfish, we're jes going to gut ya"; or "Youjes jump back in the water,little catfish, we're jes going to gut ya"; or "You jes digest, little catfish, we're jes


8/21


9/21

470 PoeticsToday 11:3Let us turn to a famous passage in the Fourth Gospel to demonstratehow an original aspect of invention can be constituted as an exploita-

tion of an unoriginal constraint in metaphoric invention. Jesus speaksto the Apostles:"Set your troubled hearts at rest. Trust in God always; trust also in me.There are many dwelling-places in my Father's house; if it were not so Ishould have told you; for I am going on purpose to prepare a place for you.And if I go and prepare a place for you, I shall come again and receive youto myself, so that where I am you may be also; and my way there is knownto you." Thomas said, "Lord, we do not know where you are going, so howcan we know the way?"Jesus replied, "I am the way; I am the truth and Iam the life; no one comes to the Father except by me." (John 13:36-14:6,New EnglishBible)The subject of this passage is trust. Jesus will take a journey andthen return, he implies, to take the Apostles on the same journey tothe same destination. This may be read literally or metaphorically orboth. The metaphoric reading calls upon the conventional metaphoric

understanding of life as ajourney and upon its conventional extensionto the metaphoric understanding of death as a departure (see furtherLakoff and Turner [1989]). To this point, there is no felt violationof the target. The metaphoric reading simply inherits prefabricatedsatisfactions of constraints from conventionalized metaphoric under-standings, as we have seen before in the case of Pilgrim's Progress.At this point Thomas asks a perfectly sensible question that makessense either literally or metaphorically. It concerns the state of knowl-edge of the person being told about the journey. The first readingof his question is literal: How can we know the physical path to adestination whose location is unknown to us? The second reading ismetaphoric: How can we know the "way" that will lead us to a goalwhen the goal is unknown to us? This is a question about concep-what they might be, as in, "A schema may be transported almost anywhere. Thechoice of territory for invasion is arbitrary; but the operation within that terri-tory is almost never completely so. We may at will apply temperature-predicates tosounds or hues or personalities or to degrees of nearness to a correct answer; butwhich elements in the chosen realm are warm, or are warmer than others, is thenvery largely determinate. Even where a schema is imposed upon a most unlikelyand uncongenial realm, antecedent practice channels the application of the labels.When a label has not only literal but prior metaphorical uses, these too may serveas part of the precedent for a later metaphorical application; perhaps, for instancethe way we apply 'high' to sounds was guided by the earlier metaphorical appli-cation to numbers (via number of vibrations per second) rather than directly bythe literal application according to altitude" (1976 [1968]: 74-75). What is missingfrom this enticing passage is any account of why certain choices are determinate,how antecedent practice channels the application of labels, or how prior uses serveas part of the precedent for later metaphorical application.


10/21

Turner* Metaphorand Invention 471tual structure and applies equally to the literal and the metaphoricreadings.

When Jesus answers that he is the way,both the literal and the meta-phoric readings fall apart. Literally, his statement violates our schemaof a journey because a person cannot be a way, much less both a trav-eler and a way. Metaphorically, his statement asks us to map two sensesin the source, namely, both the traveler and the way, onto one sensein the target, Jesus. We feel this construal to be a flagrant violation ofprotected structure in the target.It is this dissonance that signals to us that we must perform someoriginal work to arrive at a different construal that satisfies constraints.One strategy is to attribute to the passage a double reading: Jesus istaking one journey, whether literal or metaphoric, in which he is theliteral or metaphoric traveler; and we, or the Apostles, are to takeanother journey, a metaphoric journey, in which we, or the Apostles,are the metaphoric travelers and Jesus is the metaphoric way or con-duit to a state of being that is metaphorically both a location and thedestination of this particular journey.There are other ways to attempt to satisfy constraints. We might ob-serve that divinity in the Fourth Gospel is marked not by iconographicattributes or miracle stories but rather by discourse that violates whatwe take to be reliable conceptions. In this text, divinity talks like this.The divine, unlike the mortal or the everyday, can be a traveler and away, an agent and a path. The divine can violate identity, as when, intrinitarian doctrine, we are told that three are one.11It is notjust that the divine can violate constraints, but that the divinefrequently is signaled exactly through such violation. The constrainton preserving identity is violated, and its violation is a carrier notjustof some significance, but of a particular significance. We would notrecognize the attribution of divinity in this passage in the absence ofknowing constraints.We have seen a few specific manifestations of what may be a gen-eral constraint on metaphor and demonstrations of the unoriginal andoriginal exploitations of that general constraint in literary invention.Let us now try to express that general constraint. This constraint hasto do with the forms of our experience and how these forms structureour thoughts. We experience images in various modalities: a visualimage of a road, an auditory image of a scream, a kinesthetic imageof a pinch, an olfactory image of the smell of roses, and so on. Norich image is wholly unique; rather, it shares skeletal structure with11. We have what I will call in a moment an "image-schema" for one and a dif-ferent "image-schema" for three. In quotidian experience and in our metaphoricconception of a target, three cannot be one. The divine is marked as transcendingsuch constraints.


11/21

472 PoeticsToday 11:3other images. We have a skeletal image of a scream that inheres withinour rich images of particular screams. We have a skeletal image of aflat, bounded planar space that inheres within our rich images of indi-vidual tables, individual floors, individual plateaus. We have a skeletalimage of verticality that inheres within our rich images of individualtrees, individual buildings, individual people. Following Mark John-son, I will use the technical term "image-schema" for such skeletalforms that structure our images.12As I conceive of them, image-schemas are extremely skeletal imageswhich we use in cognitive operations. We have many such image-schemas: of bounded space, of a path, of contact, and of such orien-tations as up/down, front/back, and center/periphery. We have manyimage-schemas of part/whole relational structure. We also have dy-namic image-schemas, such as the image-schema for a rising motion,or a dip, or an expansion, and so on. When we understand a scene,we naturally structure it in terms of such elementary image-schemas.Let us turn now to a first approximation of the general constrainton metaphor by taking up just images. It appears to be the case thatwhen we map one image metaphorically onto another, we are con-strained not to violate the schematic structure of the target image. Forexample, a verticality schema in the target cannot without provokingremark have mapped onto it its inverse; a bounded interior in the tar-get cannot have mapped onto it both bits of an interior and bits of anexterior; and so on.Consider, for example, Auden's lines from "1929":

But thinkingso I cameat onceWheresolitaryman satweepingon a bench,Hanginghis head down,with his mouthdistortedHelplessand uglyas an embryochicken.(1976:50, 11.9-12)

The hanging head of the solitary man is a bounded interior, with anexterior; it has an internal up/down structure (for example, the top of12. An "image-schema,"accordingtoJohnson, "isa recurring,dynamic pattern ofour perceptual interactions and motor programs that gives coherence and struc-ture to our experience. The VERTICALITYschema, for instance, emerges from ourtendency to employ an UP/DOWN rientation in picking out meaningful structuresof our experience. We grasp this structure of verticality repeatedly in thousandsof perceptions and activities we experience every day, such as perceiving a tree,our felt sense of standing upright, the activity of climbing stairs, forming a mentalimage of a flagpole, measuring our children's heights, and experiencing the levelof water rising in the bathtub" (Johnson 1987: xiv). The notion that we use image-schemas to structure our perceptions and conceptions is implicit in Palermo (1987,1988a, 1988b). Langacker (1987, 1988a, 1988b) has been articulating since 1974the ways in which semantic structure is based on what he calls "images," which re-semble Johnson's image-schemata. Technically, Langacker views Johnson's "image-schemata" as a subset of Langacker's "images" (personal communication).


12/21

Turner* Metaphorand Invention 473the head and the bottom of the head); its direction is roughly down-ward (looking down); its open mouth is a concavity in the boundary;its parts (mouth, eyes, top of head, and so on) have relational struc-ture, such as adjacency. Although our rich image of the hanging headmay include all sorts of detail, that detail is structured by these image-schemas. I refer to this structure as the "image-schematic structure"of the target image. We are constrained not to violate it when we mapthe image of the embryo chicken onto it: the interior of the chickenhead maps to the interior of the human head, the boundary to theboundary, the verticality to the verticality, and so on. A similar ex-ample might be Blake's personification of a sunflower, in which theimage of a human body is mapped onto the image of a sunflower, pre-serving the part/whole relations and the orientation of the sunflower.The schematic structure of the target image is not violated.The next consideration to bring to bear in formulating this con-straint is that many things other than images appear to be structuredby image-schemas. Our concepts of time, of events in time, and ofcausal relations seem to be structured by these image-schemas. We liketo think of time, which has no shape, as having a shape such as linearor circular, and of that shape as having skeletal structure. We like tothink of events in time, which have no shape, as having shape such ascontinuity, extension, discreteness, completion, open-endedness, cir-cularity, part/whole relations, and so on. We like to think of causalrelations as having such skeletal shapes as links and paths. Theseshapes, these image-schemas, need not be static. We have a dynamicimage-schema of one thing coming out of another, and we use it tostructure one of our concepts of causation.With this addition we can reformulate the general unoriginal con-straint on metaphoric invention thus: In metaphor,weareconstrainednotto violate the image-schematictructureof the target;thisentails that we areconstrainednot to violatewhatevermage-schematictructuremaybepossessedbynonimagecomponents f thetarget.The formulation of this constraintrequires many clarifications and comments.First, the constraint says nothing about what can or cannot, orshould or should not, be mapped from the source to the target, noth-ing about what components of the target can or should be involvedin the metaphoric invention, and nothing whatever about strategiesof mapping or of reconception that might be used in the service ofsatisfying the constraint.Second, the constraint is not inviolable; however, if it is violated, theviolation is to be taken as a carrier of significance. Usually, we assumethat violations are to be avoided as we construct a meaning for theutterance. The constraint thus guides our understanding by blockingcertain possibilities. But when we conclude that the utterance is to betaken as violating the constraint, then we must look for some signifi-


13/21

474 PoeticsToday 11:3cance in the violation. For example, we might take the violation asan aggressive and intentional request that we change our conceptionof the target in exactly the way indicated by the violation. We thentake the violation as urging us to form some different conception ofthe target, that is, one not violated by the utterance. To put the samepoint differently, one drastic way to satisfy the constraint is to builda new conception of the target. But we assume that a violation of theconstraint is never insignificant. If ultimately we find no significancein the violation, we will find the metaphoric invention either faulty orbeyond our powers.Third, this formulation is summary, and does no more than hint atthe complexity of the subject. I have nothing approaching a definitionor taxonomy of image-schemas, a theory of how they arise and work,or an explanation of their role in structuring concepts. My formula-tion of this constraint does offer a challenge to anyone who thinks thatwhat is automatic and unoriginal in invention is simple in itself and toosimple to need analysis. It is manifest that our unoriginal understand-ing is informed by something like image-schemas, that metaphoricinvention is constrained in ways having to do with image-schemas, andthat originality in metaphoric invention can be constituted as exploita-tions of these constraints.13Yet finding a fair characterization of these13. The image-schematicconstraint hypothesized here bears a complicated rela-tionship to Dedre Gentner's structure-mappingtheory of analogy. See Gentner(1983, 1986). Gentner writes:

The centralideain structure-mappings that ananalogy s a mappingof knowledgefromone domain(the base)into another(thetarget)whichconveys hata systemofrelations hatholdsamongthebaseobjectsalsoholdsamongthe targetobjects.Thusan analogyis a wayof focusingon relationalcommonalitiesndependentlyof theobjectsin whichthose relationsare embedded.In interpretingan analogy, peopleseek to put the objectsof the basein one-to-onecorrespondencewiththe objects nthe targetso asto obtainthe maximum tructuralmatch.Objectsareplaced n corre-spondencebyvirtueof theirlikeroles in the commonrelational tructure; heredoesnot need to be anyresemblancebetweenthe targetobjectsand theircorrespondingbaseobjects.Centralto the mappingprocess s the principleof systematicity: eopleprefer to map connected systemsof relationsgoverned by higher-order relations withinferential mport,ratherthan isolatedpredicates. (1986:3-4)

Gentner here hypothesizes a heuristic used in analogical understanding. It bearsthe following relationships to the image-schematicconstrainton metaphor that Ipropose.If relational structure is image-schematic, as it would seem to be, then the image-schematic constraint I propose entails that we are constrained not to violate therelational structure of the target. This relational corollary to the image-schematicconstraint is compatible with Gentner's heuristic just in those cases of metaphorwhere the relational structure of the target is not violated and is also maximallyinvolved in the mapping: such a case satisfies both the image-schematic constraintand Gentner's heuristic. In one way, the relational corollary to the image-schematicconstraint on metaphor is stronger than Gentner's heuristic because Gentner's


14/21

Turner* Metaphorand Invention 475phenomena would require a theory of image-schemas, of their rela-tion to images, of their origin, of their relation across modalities (aswhen we talk of a "screaming red" or a "sharp tartness"), of their usein structuring concepts, and, beyond all this, a larger theory that couldaccount for why they seem to have privilege in metaphoric invention.14Let us return for just a moment to the lines we briefly inspectedearlier, "Trees climb the hills toward the Golan and descend to testtheir resolve near the desert." We asked earlier why we are not both-ered by the violations in this poetic metaphor, and now we can answer.heuristic of seeking maximal structural match does not imply a constraint againstviolating relational structure in the target; indeed, Gentner's heuristic apparentlywould allow us to violate some relational structure in the target if doing so enabledus to involve more of the relational structure in an analogical match than wouldotherwise be possible. Violating some relational structure in the target should bewelcomed, according to Gentner's heuristic, if it permits greater final structuralmatch. In another way, the relational corollary to the image-schematic constrainton metaphor is weaker than Gentner's heuristic because it is merely a constraintagainst violating image-schematic structure in the target and says nothing aboutwhat should be involved in the mapping, whereas Gentner's heuristic concernsexactly that. Although the image-schematic constraint and Gentner's heuristic canin some cases be compatible, the image-schematic constraint concerns many as-pects of metaphor that Gentner's heuristic apparently does not. The full image-schematic constraint itself concerns all image-schematic structure, including manythings (like slowness as part of an event shape) that are crucial to metaphor but thatdo not appear to fall under what Gentner would describe as relational structure(see further Turner [1988]).14. The rudiments of the image-schematic constraint conjectured here are pre-sented implicitly but not analyzed in Lakoff and Turner (1989). It began as thehypothesis that conceptual metaphor preserves image-schematic structure and wascalled "The Invariance Hypothesis." But this strong version of the InvarianceHypothesis was an overstatement in two ways. The first way is trivial: Many com-ponents of image-schematic structure in the source are simply not involved in themapping. For example, when we understand our boss as a crab, the image ofthe crab, and consequently its image-schematic structure, are simply not part ofthe mapping. Accordingly, that image-schematic structure in the source is not pre-served by the mapping; it is not carried over to the target. (Lakoff and I implicitlyexplain this in chapter 4 when we discuss how the Maxim of Quantity guides usto exclude various components of the source and target from the mapping.) Thesecond way in which the strong version is too strong is substantive: Componentsof the source that are indeed involved in the mapping often have image-schematicstructure that is not mapped onto the target. Consider LIFE IS A JOURNEY. Thereis a path in the source domain, and it is mapped onto the target. That path in thesource has image-schematic structure. But much of this image-schematic structureis simply not mapped onto the target. For example, it is part of the image-schematicstructure of the path to be fixed, to be independent of our traversal of it. Travers-ing the path neither creates nor destroys it. Consequently, we can meet a fork inthe path, choose one fork, take a step, change our mind, step back, and take theother fork. Metaphorically, meeting a fork corresponds to coming upon alterna-tives. But the fixity of the fork does not map onto the fixity of the alternatives.


15/21

476 PoeticsToday 11:3We can understand the form of the line of trees15 as the trace of amovement (or "summary scan," to use Langacker's [1987] term): Thetrace of a climb that crests and then descends has the same image-schematic structure as the line of the trees. Consequently, when one ismapped onto the other, there is no violation of the image-schematicstructure of the target. So much for the images. Now consider theevents and actions in this passage. The target is an event that involvesthe trees and the desert: the trees occupy a position (a literal position)and are opposed in their occupation of it by desert forces that maydislodge them from that position. The source for understanding thistarget event metaphorically is an action: testing one's resolve. Suchtesting has an event shape: we occupy a position (metaphoric or lit-eral) and are opposed in that occupation by some force that would(metaphorically or literally) dislodge us; if we abandon that position,we say our resolve failed. The event shape of the source correspondsto the event shape of the target. Both are structured as a positionedentity that has been moved to that position by one force exerted uponit, and that encounters in that position a countervailing force. Theoutcome is either the stasis or movement of that entity, and this out-Many of our decisions are irrevocable: Shall we boil this egg or scramble it? Shallwe marry Tom or Harry? In these cases, the rejected alternative disappears themoment we engage in the chosen alternative. If we boil the egg, we cannot thenscramble it, and if we scramble it, we cannot then decide to boil it. Metaphorically,one of the forks is destroyed the moment we step down the other. We cannot take astep back and be again at the metaphoric fork in the road because the fork doesn'texist anymore. The metaphoric path, unlike the source path, changes as a resultof being traversed. The fixity of the path in the source, its independence of ourtraversal, is not mapped onto the target. The reason it cannot be so mapped inthese cases is that to do so would violate the image-schematic structure of the tar-get. In the source, there is preservation, which is image-schematic structure. In thetarget, there is destruction, which is image-schematic structure. To map the sourcepreservation onto the target destruction would be to violate the image-schematicstructure of the target, and so we do not map that part of the image-schematicstructure of the source. We see from this example that the strongest version ofthe constraint and the one I present here are incompatible for a range of cases.For these cases, obeying the strong version of the constraint violates the weakerversion, and obeying the weaker version of the constraint violates the stronger ver-sion. It appears that the strongest compromise version of the constraint for whichthere are no clear counterexamples would be: In metaphoricmapping, for thosecom-ponents of the source and target domainsdetermined to be involved in the mapping, preservethe image-schematicstructure of the target, and importas much image-schematicstructurefrom the source as is consistentwith thatpreservation.15. Ambiguity allows us to imagine the image of the trees in various ways. Here,for the sake of discussion, I assume that an unbroken line of trees first climbs,then crests, then descends. Alternatively, one arm of a forest might climb in onedirection while another arm descends in a different direction. In any case, we areconstrained not to violate the schematic structure of the target image.


16/21

Turner * Metaphor and Invention 477come is determined by the size of the vector forces. Again, the image-schematic structure of the target (here, the image-schematic structureof the event shape) is not violated. So much for the events and actions.Finally, consider the causal structure of the target: there is a causal linkbetween the desert and the stasis or recession of the endpoint of theline of trees in the target. In the source, there is a causal link betweenthe desert and the intentional holding or abandoning of the occupiedliteral or metaphoric position. Mapping one onto the other does notviolate the causal structure of the target: a link is mapped onto a link.It is important to observe that many things in the target have beenviolated, such as intentionality, but not image-schematic structure.Let us take up, as a last example, a poem that draws upon thesame conventional metaphoric understanding of life as ajourney thatwe saw in Pilgrim's Progressand the Fourth Gospel, but one that ismore difficult than either of them. Pilgrim'sProgresssimply inheritedthe conventional and unoriginal satisfaction of the image-schema con-straint from the conventional metaphoric understanding of life as ajourney. The passage from the Fourth Gospel, however, asked us to dosome original work in order to satisfy that unoriginal constraint. "AtNorth Farm" byJohn Ashbery moves further out along that gradient:Somewheresomeone is traveling uriously owardyou,At incrediblespeed, travelingdayand night,Throughblizzardsand desertheat,acrosstorrents,throughnarrowpasses.But willhe knowwhere to findyou,Recognize you whenhe sees you,Giveyou the thinghe hasfor you?

Hardlyanything growshere,Yet the granariesareburstingwithmeal,The sacksof mealpiled to the rafters.The streams runwithsweetness, atteningfish;Birdsdarkenthe sky.Is it enoughThat the dish of milkis set out at night,That we thinkof himsometimes,Sometimesand always,withmixedfeelings? (1984: 1)

Pilgrim'sProgressvirtually compels the reader with multiple linguis-tic cues to understand it as structured by the conventional metaphoricunderstanding of life as ajourney and inherits from that conventionalmetaphoric understanding a prefabricated and unoriginal satisfactionof the image-schematic constraint. The passage from the Fourth Gos-pel virtually compels the reader to understand it as structured by theconventional metaphoric understanding of life as a journey, but re-quires of the reader (and apparently of the Apostle Thomas) some


17/21

478 PoeticsToday 11:3original conceptual work in order to satisfy that constraint. The viola-tion of the constraint is indeed the spur to perform that work, whichwe take as having been accomplished exactly when the constraint hasbeen respected."At North Farm,"by contrast, only suggests that we take parts of it asstructured by the metaphor of life as ajourney, and leaves quite widelyopen to us the range of possible metaphoric construals if we chooseto do so. Yet as we go about those construals, we are guided by theexpectation that they are to respect that unoriginal image-schematicconstraint.In "At North Farm," there are no or few overt indications thatthe expressions must be understood metaphorically. The entire sec-ond stanza, for example, can be taken as a literal description. Thequestions of the first stanza can be equally literal. Only the first sen-tence, with its description of incredible speeds and travel without restthrough extreme conditions, offers extravagances that seem unlikelyto be literal. If we do bring our capacities for metaphoric construal tobear on this first sentence, and, in the interests of consistency and rich-ness, extend them to the rest of the poem, we find ample warrant totake the schema of travel and journeying as the source: in the expres-sions "somewhere someone is traveling," "toward,""speed," "travelingday and night," "Through blizzards and desert heat," and so on, wecan find unproblematic expression of our conventional metaphoricunderstanding of life experiences as journeys, especially of progresstoward goals asjourneys toward destinations, where impediments andhardships in the journey correspond metaphorically to difficulties inreaching the goal.Beyond that, the possibilities for metaphoric construal open in manydirections. Some of the possible metaphoric correspondences mightbe the correspondence of apparent physical capacities of the travelerto personal capacities of the person with the goal (specifically, physi-cal stamina may correspond to psychological stamina); the correspon-dence of physical relentlessness of the traveler to the psychologicalor social relentlessness of the person with the goal; and the corre-spondence of reaching the physical location of the person addressed("you") to beginning a psychological and social interaction with thatperson.

To take a complex example, one reminiscent of the passage from theFourth Gospel, consider that there appears to be doubt about whetherthe traveler knows the location of the person addressed ("you"). Thislocation constitutes the traveler's destination, but how can the travelerknow the path to that destination without knowing where it is? As theApostle Thomas said, "Lord, we do not know where you are going, sohow can we know the way?" Moreover, there is doubt about whether


18/21

Turner* Metaphorand Invention 479the traveler can even recognize the person sought. We can take thisfirst as just referring to the source domain of travel: We may construethese doubts as indicating that the traveler knows or hopes that theperson sought exists, but is not certain where, and consequently doesnot know the path, but is exploring possible avenues. The traveler maynot even recognize the person sought. Alternatively, given the ambi-guity of "will he . . . recognize [and] give," we may construe that evenif the traveler finds the person sought, the traveler may decline to rec-ognize that person socially and give whatever he has to give. To whatmight these construals correspond metaphorically? Being ignorant ofthe physical path to the location may correspond metaphorically tobeing ignorant of how to engage "you."These are only a few of the possible metaphoric construals, and Iwill not pursue others but simply observe that as we go about theseconstruals, we are guided and empowered by the unoriginal constraintthat we not violate the image-schematic structure of the target. For ex-ample, we are not to destroy the image-schematic "one-ness" of a sensein the target by mapping onto it two different senses in the source.Nor are we to destroy the order of temporal moments in the target.Nor can a path in the source that may be a dead-end be mapped ontoan interaction in the target that certainly achieves the goal, for thatwould violate the image-schematic relation of "leading to": if in thetarget the interaction certainly leads to the goal, but in the sourcethe path does not certainly lead to the destination, then mapping thedestination onto the goal and the path onto the interaction wouldviolate the image-schema of "leading to" in the target, and we areconstrained against doing that. In all cases, we are guided by an image-schematic constraint which applies, I claim, to the myriad complex,inventive, and original possible metaphoric construals that I do notconsider here.There is a system to imagination. Though infinitely variable and un-predictable, imagination is grounded in structures of invention eitherwholly unoriginal or with an originality that consists of exploitationswithin a known and unoriginal space. Were imagination free, we wouldtake its products to be unintelligible, meritless caprices rather thansignificant, valuable achievements. Metaphoric imagination, includingmetaphoric imagination in those poems we regard as most original,suggestive, and demanding, appears to be guided by an utterly un-original constraint so unrecognized in criticism and so daunting inits complexity that it cannot yet be truly formulated but must be ges-tured toward, with a heavy reliance upon the reader's intuitive senseof what it means: the image-schematic structure of the target is not tobe violated.It is predictable and unobjectionable that literary criticism would


19/21

480 PoeticsToday 11:3attend in the main to those aspects of imagination that might be ex-pected to vary from age to age, or country to country, or author toauthor, or even passage to passage. We are vigilant for the new and thevariable. But attending to what varies and not to what abides meansthat we see only a contingent aspect, when we believe ourselves tobe seeing the whole. In one respect, to understand the sophisticatedways in which the Ashbery poem can work we must understand theunoriginal structure of invention it shares with both Pilgrim'sProgressand the Farewell Discourse from the Fourth Gospel. To understand itsoriginality is in one respect to understand its movement to satisfy anabiding unoriginal constraint on metaphor that Pilgrim'sProgress sat-isfies in a prefabricated way and the Farewell Discourse satisfies in itsown original way, a constraint that governs, guides, and empowers allthese poems. It is not possible to understand such a novel and arrest-ing poem as "At North Farm" without considering those unoriginalstructures of invention that are the known space in which imaginationmoves, and whose constrained exploitation is poetic originality.

ReferencesAristotle [384-322 B.c.]1984 [1954] The Rhetoric nd thePoeticsofAristotle,ntroduction by Edward P.J.Corbett (New York:Modern Library).Ashbery, John1984 A Wave New York:Penguin).Auden, W. H.1976 Collected oems, dited by EdwardMendelson(New York:Random House).Black, Max1979 "More About Metaphor," n Metaphor nd Thought, dited by A. Ortony,19-43 (Cambridge:CambridgeUniversityPress).Booth, Wayne1979 "Metaphoras Rhetoric,"in On Metaphor,dited by S. Sacks, 47-70 (Chi-cago: Universityof Chicago Press).Cooper, D.1987 MetaphorOxford: Blackwell).Fauconnier, Gilles1985 MentalSpaces:Aspectsof MeaningConstructionn NaturalLanguage(Cam-bridge, MA:Bradford Books/MIT Press.Fillmore, Charles1968 "The Case for Case,"in Universalsn LinguisticTheory, dited by E. Bachand R. Harms, 1-90 (New York:Holt, Rinehartand Winston).1985 "Framesand the Semanticsof Understanding,"Quadernidi Semantica (2):222-53.Gentner, Dedre1983 "Structure-Mapping:A Theoretical Frameworkfor Analogy," CognitiveScience7(2): 155-70.1986 "The Mechanisms of Analogical Learning."Paper presented at the ARIConference on Analogical Similarity, o appearin SimilarityndAnalogicalRea-soning,edited by S. Vosniadouet al. (London: CambridgeUniversityPress).


20/21

Turner * Metaphor and Invention 481

Goffman, Erving1974 Frame Analysis (New York: Harper and Row).Goodman, Nelson1976 [1968] Languages of Art (Indianapolis: Hackett).1979 "Metaphor as Moonlighting," in On Metaphor, edited by S. Sacks, 175-80(Chicago: University of Chicago Press).Johnson, Mark1987 The Body in the Mind (Chicago: University of Chicago Press).Kittay, Eva1987 Metaphor: Its Cognitive Force and Linguistic Structure (Oxford: Clarendon).Kittay, Eva, and Adrienne Lehrer1981 "Semantic Fields and the Structure of Metaphor," Studies in Language 5:31-63.Lakoff, George, and Mark Turner1989 More than Cool Reason: A Field Guide to Poetic Metaphor (Chicago: Universityof Chicago Press).Langacker, Ronald1987 Foundations of Cognitive Grammar. Vol. 1, TheoreticalPrerequisites (Stanford:Stanford University Press).1988a "An Overview of Cognitive Grammar," in Rudzka-Ostyn 1988: 3-48.1988b "A View of Linguistic Semantics," in Rudzka-Ostyn 1988: 49-90.Minsky, Marvin1975 "A Framework for Representing Knowledge," in The Psychology of ComputerVision, edited by P. H. Winston, 211-77 (New York: McGraw-Hill).Palermo, David1987 "The Transfer Dilemma: From Cross-Modal Perception to Metaphor."Paper presented at the Third International Conference on Thinking, Hono-lulu, Hawaii.1988a "Metaphor: A Portal for Viewing the Child's Mind," in Essays and Ex-periments in Honor of Charles C. Spiker, edited by L. P. Lipsitt et al., 111-36(Hillsdale, NJ:Lawrence Erlbaum Associates).1988b "Knowledge and the Child's Developing Theory of the World," in Ad-vances in ChildDevelopmentand Behavior, Vol. 22, edited by H. W. Reese, 269-95(New York: Academic Press).Ricoeur, Paul1977 The Rule of Metaphor: Multi-Disciplinary Studies of the Creation of Meaningin Language. Translated by Robert Czerny (Toronto: University of TorontoPress).Rudzka-Ostyn, Brygida1988 Topicsin Cognitive Linguistics (Amsterdam/Philadelphia: John Benjamins).Rumelhart, David E.1980 "Schemata: The Building Blocks of Cognition," in Theoretical Issues inReading Comprehension: Perspectivesfrom Cognitive Psychology, Linguistics, Arti-ficial Intelligence, and Education, edited by Rand J. Spiro et al. (Hillsdale, NJ:Lawrence Erlbaum Associates).Rumelhart, D. E., J. L. McClelland, and the PDP Research Group1986 Parallel Distributed Processing: Explorations in the Microstructure of Cognition.Vol. 2, Psychological and Biological Models (Cambridge, MA: Bradford Books/MIT Press).Rumelhart, David, and Andrew Ortony1977 "The Representation of Knowledge in Memory," in Schooling and theAcqui-sition of Knowledge, edited by R. C. Anderson et al., 99-135 (Hillsdale, NJ:Lawrence Erlbaum Associates).


21/21

482 PoeticsToday 11:3Schank, Robert, and Robert Abelson1977 Scripts, Plans, Goals, and Understanding (Hillsdale, NJ: Lawrence Erlbaum

Associates).Turner, Mark1987 Death is the Mother of Beauty: Mind, Metaphor, Criticism (Chicago: Universityof Chicago Press).1988 "Categories and Analogies," in Analogical Reasoning: Perspectives of ArtificialIntelligence, Cognitive Science, and Philosophy, edited by David H. Helman, 3-24(Dordrecht: Kluwer).

Documents

Turner Metaphor Invention