Reasonable and Unreasonable Worlds:

Some Expectations of Coherence in Culture

Implied by the Prohibition of Mixed Metaphor





Dale Pesmen







from Beyond Metaphor: The Theory of Tropes in Anthropology (1991)

edited by James W. Fernandez

reproduced by permission of Stanford University Press















Sir, you may fight these battles as long as you will, but when you

come to balance the account you will find that you have been fish-

ing in troubled waters, and that an ignis fatuus has bewildered

you, and that indeed you have built upon a sandy foundation, and

brought your hogs to a fair market.... I am, sir, yours, etc.


                                                              Joseph Addison


There is an etiquette of metaphor and mixing metaphor is not allowed. From Aristotle through the Enlightenment to the present time, textbooks on rhetoric and style forbid it. Why?

            This essay suggests that certain expectations of reality are implied by the interdiction of mixed metaphor expectations that also inform current discussions of consistency and eclecticism in anthropology, philosophy, and aesthetics. I will argue that the prohibition against "taking up cross-scents" stems from a wider ideology of coherence inseparable from our intuitions of the "rightness" or "goodness" of theories, thoughts, and ways of operating in the world. This notion of coherence governs our judgments of the "truth," "validity," and "realism" of pictures of that world and distinguishes things we can think about from things we find objectionable and/or impossible to think.

            A typical example of the prohibition of mixed metaphor is that of Thomas Gibbons, in his 1767 treatise on rhetorical tropes. Gibbons calls mixed metaphors disgustful, spoiled, repugnant, absurd, and miserable incongruities, and writes that "though the Metaphor is so excellent and lovely a Trope, there is nothing so disgustful as a Metaphor ill-chosen and ill-conducted, according to the old maxim, corruptio optimi est pessima," and "after we have begun a Metaphor, we are to beware lest we spoil it, by introducing something repugnant and dissimilar to the first image" (1969:28). Gibbons agrees with Quintilian's comment that to "set out with a tempest [and end with] a conflagration" has an effect of "most shameful inconsistency" (1891:136). These writers are not alone in finding mixed metaphor "offensive and repulsive"; it is generally considered "one of the most intolerable faults of composition" (Gibbons 1969: 34). Having been shown a mixed metaphor in Paradise Lost, Coleridge told Wordsworth that he could not sleep all the next night for thinking of it (Bett 1932: 242). This is the sort of language that writers on rhetoric, style, and poetics have used to describe their responses to the "vicious combination of metaphors" (Constable 1731:104). The other response (besides this one, formulated often in terms of nausea) seems to be that mixed metaphor is intolerably absurd, literally laughable, as in the example of the New Yorker section "Block that Metaphor." But is ridicule "the test of propriety"? ("T.D." 1825: 723). I would like to argue that this laughter in response to mixed metaphor is nervous laughter, laughter in response to a threat to "the familiar landmarks of . . . our thought, the thought that bears the stamp of our age and our geography breaking up all the ordered surfaces and all the planes with which we are accustomed to tame the wild profusion of existing things." Thus Michel Foucault describes his experience of reading Jorge Luis Borges's "Chinese encyclopedia" taxonomy. Foucault describes a wonderment in which "the thing we apprehend . . . that by means of the fable, is demonstrated as the exotic charm of another system of thought, is the limitation of our own, the stark impossibility of thinking that. But what is it impossible to think, and what kind of impossibility are we faced with here?" (1970: XV). Whatever kind of impossibility it is, the mixing of metaphors produces the same sort.

            Because mixed metaphors suggest impossible worlds, I will examine the various attempts made over the years to explain why metaphors ought not to be mixed. I believe that these attempts describe implicitly what worlds we consider "possible." However, although all style texts forbid and/or ridicule mixed metaphor, there are in effect very few explanations, and those that do exist can be seen as the same explanation invoked in various ways. It is not for the purpose of discussing metaphor or style that I am using this material; I would like rather to look at whatever implicit assumptions are made explicit in the course of prohibiting mixed metaphor. I shall then move from these assumptions and from the various arguments against, to use Foucault's word, the Heterotopias (as opposed to Utopias) (ibid.: xviii) implied or generated by mixed metaphors to examine the notions of order that these "impure unions” as mixed metaphors are so often called in the style texts imply.



The Problem


It is forbidden diet. In prose writers, unclean; in poets, abominable.

"The reason why" is another matter; for [The laws of critics and

grammarians], like those of the comets, are inscrutable to common

intellect.... What anathema of Nature's is there against changing

a metaphor as often as a man pleases?


                                       "T D., " On the Use of Metaphors "


            The anonymous author of the above quotation is one of the few staunch defenders of mixed metaphor. Writers as a rule do not deny that things can be successively compared to any number of things "if considered in several distinct lights"; the problem is one of when and how this may be done. For example, Joseph Addison writes that "an unskillful author shall run these metaphors so absurdly into one another, that there shall be no simile, no agreeable picture, no apt resemblance, but confusion, obscurity, and noise." My question about these objections begins much like that of "T.D.," who speaks of critics and grammarians as


members of a class of persons, who, like certain secondary animals, live and are fed upon those of greater importance than themselves . . . [And] have put [”mixed" or "broken" metaphor], as an Otaheitan would say, "under Taboo."


            Why is it of necessity to fill up a sentence or a period, or a paragraph, or any assignable space neither more nor less, like Shylock’s pound of flesh, on peril of cancelling the bond? What jurisdiction has Matthews or Hoyle to compel us to play out the suit thus before we try another? Are we bound to run down the first similitude we start, like beagles, or be lashed by some whipper-in of a pedagogue for taking up a cross-scent? What law is there to compel us to let our first metaphor, like our first wife, die a natural death before we take a second?--or what canon is there thus coupling a metaphor and matrimony, and insisting that our comparisons, like our wives, if we will have them, shall be pure, and only by one at a time? Has any critic shown the existence of such laws, either in reason or out of it? (1825: 721-22)

            Where I do not agree with "T.D.," however, is in the finality of his conclusion: "The forbiddance of mixed metaphor is founded then upon assertion." The prohibition is not without foundation but rests on certain notions of commitment and fidelity. As we are told, "What God therefore hath joined together, let not man put asunder" (Matthew 19: 6).

            The marriage metaphor for metaphor use is actually quite popular:”T.D." jokes about "sinful flirtations with first one metaphor and then another," "being happily converted . . . to the blessed communion of 'one at a time, for better for worse,"' and "the great licentiousness of [Shakespeare] in joining heterogeneous metaphors" (ibid.: 723), but mixed metaphor is often called "monstrous union" and similar names by most of its critics. This is an indictment with sensual over- (or under-! tones which I shall come back to throughout this essay. There is a sense in which mixed metaphor is an infidelity: a metaphoric predication implies an association not only of two terms, but also of two worlds, in what is felt as a sort of promise of union. Perhaps the initial observation of an analogy is not yet perceived to be "satisfying" or "substantial," and so immediate movement to an entirely different domain is felt as leaving a set of potentialities behind. However, I would like to follow the rhetoric and style texts through some descriptions of the experience of this phenomenon.

            There are several descriptions of mistakes that lead to the generation of mixed metaphors. Broken metaphors are perceived as being impure unions, "the mythical mermaid; what begins as a human being ends up otherwise" (Kellogg 1895: 132). That is, "incompatible" images are set too close together for comfort. The territory of a faithful metaphor is variously described as a sentence, a period, an idea, or "the same connection or clause"; we also find vaguer warnings against "sowing them too thickly" (e.g., Addison 1859: V. 1,41 1 ). In general, mixed metaphor is when we "shall find Metaphors . . . whose meanings can no more accord with one another, than the iron and clay in the feet of the image in NEBUCHADNEZZAR'S dream ' (Gibbons 1969: 28).

            The same error can be made when mixing metaphorical with literal language, such as (from Hunt 1891: 104):


He was cured of his pride and his leprosy.

The orator drew forth their sympathy and their purses.


            The distinction between literal and figurative language leads to the issue of dead metaphor. In the words of "T.D.," "To be sure, like hares, they are so like the ground they sit on, that it requires a sharp eye now and then to make them out" (1825: 720).

            If, as it has been argued, language itself is "a dictionary of faded metaphors" (Tompkins 1897: 310) that have lost their identity as such, then metaphors inevitably mix in any sentence. This is not experientially disturbing, however, because conventional metaphor does not call itself to our attention as figurative language. A metaphor, used as such, refers to something outside the discourse: an image. Dead metaphors have lost most or all of the richness of their imagery. Thus we can elaborate on T.D.'s "taboo": "you may mix old metaphors as you will, but of young ones beware" (1825: 722). Thus it is in our creative or spontaneous use of metaphor that we run the risk of offensively mixing metaphor. Gertrude Buck, in her book on the psychology of rhetoric, has a chapter on "Pathological Forms of Metaphor." She claims that metaphor is only mixed "because it has ceased to be a metaphor. Though maintaining the figurative form, it has become, in the mind of the speaker, a literal statement." When, for example, honor demands that one be "firm as a rock in blotting something from one's soul," Buck contends, one does not really experience the image of a rock blotting something; “The image of the rock obtrudes itself for no appreciable time into the consciousness" (1899: 65). As there is no incongruity to the writer, there is none for a reader who, persuaded, reduces the original figure to a literal statement (or accepts it) with the same rapidity as the author. Mixed metaphor has to do with being conscious of the shifting of images. For the author, "the second figure does not jar with the first, for the first has ceased to be figurative" (ibid.: 64). The implication is, of course, that no one could want to mix images and that the experience of mixed metaphor can cause us to become "disenchanted"; that is, conscious that figurative language is at work on us.



Deception and Claims on Reality: Visualization and Validity


It is hoped that some of the seed will not fall on deaf ears.


            Nelson Goodman (1976: 34-35) considers the extent to which deception is the measure of realism, reducing the notion to absurdity. Since I am looking at the experience of mixed metaphors as unrealistic, the issue of deception persists as relevant. If the consciousness that a rhetorical device is being used can effectively shatter our belief in the "reality" of what we are experiencing, then this element of everyday belief and persuasion is important.

            Wallace Stevens makes an argument related to Buck's when he says, "There is no such thing as a metaphor of a metaphor. One does not progress through metaphors. Thus reality is the indispensable element of each metaphor. When I say that man is a god it is very easy to see that I also say that a god is something else, god has become reality" (1969: 179; emphasis mine). This sense of "reality" apparently refers to a constant relationship that should ideally be maintained between the metaphorical domain and that of which it is predicated. The "man" we describe as a god is a "real" man, even if he is a fictional character. Not only is it assumed that the domain the "man" "lives" in is a "coherent," "real world," but the same very particular assumptions are brought to bear upon the "world" the "god" inhabits. These are assumptions that take the form of expectations about what relations are "natural" within the metaphorical world. In mixed metaphor, these expectations are violated.

            In the course of this essay I hope to show that this traditionally demanded internal "order" is not just an order expected of metaphoric domains in relation to specific domains "in the service of which" they are created; for the remedies offered by the writers of style texts to correct the tendency to mix metaphors show that their criteria or organizing principles are related to those used to judge "realism" in visual art and pictures in general. In addition, these same criteria are much in evidence in the "coherence" and "unity" and "autonomy" we expect of formal structures, systems, and scientific paradigms.

            Metaphoric-literal shifting and mixed metaphor are often said to occur, if not exactly because of forgetfulness, as Buck proposes, then because of "loose and careless speech": John Constable explains that authors, too easily following the heat of imagination, cannot fix it sufficiently "for the finishing out of the true sequel of notions" (1731: 104; emphasis mine). Speaking in an impassioned way can have the same effect, and I shall speak of passionate speech later. In these cases, one is assumed to have become, to borrow an example from a style text, swamped in the meshes of one's argument. This is considered a fault, which there is only one developed way to avoid, a technique suggested by most of our writers, who quote it in each other back to the Greeks. This technique is to "surrender one's thoughts to the picture suggested until it is wrought out as far as needed" (Tompkins 1897: 31); emphasis mine). This technique is seen as bringing metaphor "to the test of nature" by visualizing the image suggested to us! Thus Buck recommends "picturing" as a way to assure that consciousness of earlier stages of the metaphor is retained as the continuation of the idea is being written. Gibbons offers a critique of lines written by "Dr. Young,"


One eye on Death, and one full fix'd on Heav'n,

Becomes a mortal, and immortal man.





       Together some unhappy rivals seize,

       And rend abundance into poverty;

       Loud croaks the Raven of the Law, and smiles


            He protests, "who, but he who has a disorder in his sight, can at the same time have one eye full fixed on one object, and the other eye upon another? . . . Who ever heard of a raven's smiling? And how unfortunate is it that what cannot agree with a raven in its original, should be made to agree with it in a metaphorical state? " (1969: 37-38; emphasis mine). This kind of argumentation (by ridicule) is the verbal counterpart to the "picturing method." It implies that between certain conceptions and characteristics of sense-perception, thought-experiments based on it (visual imagery in particular), and the correct use of metaphor there is a logical relation that allows one to speak of truth. Addison makes this an explicit principle: he, like Gibbons, writes that it is improper to make any being in nature or art do things in its metaphorical state

which it could not do in its original. Now, given that metaphor does indeed "serve to convey the thoughts of the mind under resemblances and images which affect the senses" (Addison 1859: V.2,387-88) and that mixed metaphor is perceived as being unclear, confusing, disgusting, and funny to some degree by virtually all writers, we still have not seen reason to consider mixed metaphor invalid, "a mighty deformity by the inconsequent combination of objects" (Constable 1731: 103; translating Quintilian); yet this is exactly the argument against the mixing of metaphors, an argument that implies that mixed metaphor is an untrue thing to do, proof being that the image created in the mind is not, in some particular sense, "realistic." As Arnold Tompkins has it, "If the parts, when pictured out by a painter, be incongruous, put your Metaphor in the fire, lest there should stand before you a goddess, horse, and ship, all in one" (1897: 312).

            Herbert Spencer, A. R. Radcliffe-Brown, and other social scientists could logically justify the use of metaphor in the development of their theories only as useful "illustrations" and "scaffolding" without which inductions would nevertheless stand by themselves. Yet in the same breath as they thus defend the use of metaphor, they call their analogies (for example, between society and an organism) "real and significant," "essential," and "true." "In the absence of physiological science . . . it was impossible to discern the real parallelisms" (Spencer 1981: 388). These "essential parallelisms" are "disclosed by modern science" (ibid.: 391-92) and are still appearing; Spencer anticipated that further scientific and technological developments would allow for further "revelations" of parallels between society and the functioning of an organism (ibid.: 433) and subsequently had to energetically refute the charge that he "based Sociology upon Biology" (1967: xl-xli), by allowing the metaphor he claimed to use only as scaffolding to inform the directions of his inquiry. Radcliffe-Brown (1952) also repeatedly uses the word "real" to refer to the organic analogy and to social structures, thereby validating social science as an empirical branch of the natural sciences. This same application of the idea of "reality" or "realism" to the use of figures appears in discussions of literary style.

            Theodore Hunt writes that "just as from a good outline of an historical painting we would be able to fill out the scene, so from the metaphorical terms used we should be able clearly and vividly to see the truth," and quotes James Russell Lowell's remark that "A Metaphor, if the correspondence be perfect in all its parts, is one of the safest guides through the labyrinth of truth" (1891:103). And J. H. Gilmore gives, alongside the "picturing" method, the logical "test for analogical comparison" from Aristotle, advising us to "strip it of its verbiage and reduce it to an equation of ratios. E.g., X: Y:: A: B. Thus, 'the ship ploughs the sea' becomes: the ship is to the sea, as the plough is to the land. The application of this test frequently shows that the resemblance is fanciful or remote; or that it involves more than four terms; that is, it is a mixed metaphor" (1891: 157).

            Yet the fact that the requisite consistent relationship of the metaphorical domain to the domain of the matter at hand can be put in logical form changes nothing of its mystery. It still remains to be discussed what criteria are invoked in a judgment of "resemblances" as inconsistent or invalid.

            When Addison writes "I bridle in my struggling Muse with pain," he controls his muse as he would bridle a horse. Once he has begun using an equestrian image to speak of the difficult subject of creative process, by continuing to say that this Muse-horse, itself "no very delicate idea," as Dr. Johnson (Bett 1932: 243) observes, "longs to launch into a bolder strain," Addison is charged by Henry Bett with committing the following "fault" or “inaccuracy": "She is in the first line a horse, in the second a boat; and the care of the poet is to keep his horse or his boat from singing" (ibid.). This still seems to equate a certain surrealism in imagery with low truth value or logical incoherence. Foucault might say that "what is impossible is not the propinquity of the things listed, but the very site on which their propinquity would be possible . . . where could they ever meet, except in the immaterial sound of the voice pronouncing their enumeration, or on the page transcribing it?" (1970: xvi). If by "site" we understand something so wide as to include assumptions about categories, expectations of coherence, etc., I agree; yet what the second part of this quotation from Foucault implies is what our style texts respond to: if the only possible locus for the described phenomenon is a text or discourse, the phenomenon is not "real," is not an acceptable use of figures. But how many styles in art have been labeled "realistic" that could only exist in the medium in which they were created! 1 And again, it is not "just style," but explicitly the "truth-value" of the metaphor that is at issue. The "clearness and safety" of a given metaphor as a way to negotiate intellectual and expressive labyrinths "truly" is what is imperiled by mixing images. None of the explanations I found in the literature escape or go beyond that assumption, based on a prescribed identification between domains. What kind of coherence and autonomy must "worlds" exhibit to merit the authority we grant them?2 The complexity of this problem is summarized by Goodman:


A version is taken to be true when it offends no unyielding beliefs and none of its own precepts. Among beliefs unyielding at a given time may be long-lived reflections of laws of logic, short-lived reflections of recent observations, and other convictions and prejudices ingrained with varying degrees of firmness. Among precepts, for example, may be choices among alternative frames of reference, weightings, and derivational bases. But the line between beliefs and precepts is neither sharp nor stable. (1978: 17)


I would like to cite two sources as to what constitutes "realism"; one is George Marcus and Dick Cushman on ethnographic realism, and the other is Karsten Harries on realism in artistic representation. I choose these two from many such descriptions relating to diverse media, all of which have a striking "family resemblance." "Realist ethnographies," observe Marcus and Cushman, "are written to allude to a whole by means of parts or foci of analytical attention which constantly evoke a social and cultural totality" (1982: 29). (From this, with a few alterations, we can also derive some descriptions of successful "realist" projects in fiction writing. )

            It is, of course, an expectation (as described by James Clifford [1983] and others) that ethnographers find systematicity in cultures and represent these cultures as systematic in the ethnographies they write. An examination of how we expect coherence in worlds of many sorts can help us build up a context in which to examine the particular case of systematicity in both the object of anthropology and anthropological endeavors. As Harries writes, "the presentation of interpretation and analysis is inseparably bound up with the systematic and vivid representation of a world that seems total and real to the reader" (1968: 29).

            Harries traces this "wholeness" to the Platonic notion of a state of timeless "being" from which man perceives himself to have fallen and toward which he strives. This realm of forms or ideas, says Harries, has represented for man "completeness," satisfaction, "true being"; he contrasts it to Edmund Burke's notion of the "sublime." Burke, relying on the Platonic definition of beauty, associates the latter with pleasure and the experience of being integrated in a community, and distinguishes this from the experience of the sublime, a new image of "reality" and "nature," associated with pain and danger, which "confronts man as an alien power" (Harries 1968: 39). The resulting paradox of "how to reconcile the immensity of the sublime with the beginning, middle, and end which Aristotle demands of a work of art" (ibid.: 41) is one of many related paradoxes in descriptions of the relation of engaged people (and participant observers) to what they consider or call "real," "true," etc.

            What does the concept "realism" imply when applied in the case of metaphor? In what does picturing the metaphor ground it? If "realism" is the criterion from which truth emerges, it would be good to know what its authority is like. Goodman (1976) examines some possible parameters of "realism," including resemblance, deception, the amount of information presented, ease of access to this information, and faithfulness. These all relate to realism in different ways, but ultimately, he says,


reality in a world, like realism in a picture, is largely a matter of habit.... For the man-in-the-street, most versions from science, art, and perception depart in some ways from the familiar serviceable world he has jerry-built from fragments of scientific and artistic tradition and from his own struggle for survival. The world, indeed, is the one most often taken as real.... Ironically, then, our passion for one world is satisfied, at different times and for different purposes, in many different ways. (1978: 20)


Yet this constant psychological shiftiness, this eclecticism, this bricolage of realities, is not often acknowledged in the idealized, commonsense notion of reality.




Changing Points of View:

Spatial Perspectives


It will be for you to say whether this defendant shall be allowed to

come into court with unblushing footsteps, with the cloak of hy-

pocrisy in his mouth, and to withdraw three bullocks out of my cli-

ent’s pocket with impunity.


W. Crafts and H. Fisk, Rhetoric Made Racy or Laugh and Learn


            Satisfying that passion for one world in many different ways, "at different times and for different purposes," is something we do not usually notice ourselves doing. However, there are unaccustomed kinds of juxtapositions and "shifts" in "point of view" that do alert us. We notice shifting in Foucault's example of Borges's encyclopedia. Foucault describes the experience of reading that animals are divided into (a) belonging to the Emperor, (b) embalmed, (c) tame, (d) suckling pigs . . . (n) those that from a long way off look like flies, as taking place, by virtue of language, in "an unthinkable space," as "fragments of a large number of possible orders glittering separately . . . in sites so very different from one another that it is impossible . . . to define a common locus beneath them all," a "mythical homeland Borges assigns . . . to that picture that lacks all spatial coherence" (1970: xvii-xviii [emphasis mine], xix). "Lack" and "impossibility" are the judgments we have seen in response to mixed metaphor; but the rhythm of the classification a, b, c, d, . . . also suggests to me a process of shifting, a radical shifting in the case of the Chinese encyclopedia, as in mixed metaphor.

            What kind of shifting? How can we describe what happens when animals are one moment instantiated as innumerable, the next as drawn with a very fine camelhair brush, the next as "et cetera," and the next as "having just broken the water pitcher"? If each of these implies a different "possible order," and a locus within it, the tendency is to read this classification as asking us to move very quickly, impossibly quickly, from picturing one kind of action or position in a particular context (or "domain”) to another action or position, perhaps in a different setting. What's more, we are asked to do this not only in space, yet somehow spatially nevertheless, or at least that is how we describe it. If this is the kind of "shifting" mixed metaphor asks us to do, the language traditionally used to describe mixed metaphor, "repulsive," "nauseous," "disgustful," "spoiled," "repugnant," etc. seems less gratuitous; what is being described is a kind of metaphorical motion sickness.

            Shifting of "points of view" and the associated idea of "space" are fundamental heuristic devices in the examination of literature in the critical writing of Boris Ouspensky in A Poetics of Composition (1973) and of Wayne Booth in The Rhetoric of Fiction (1983). The changes in point of view in the narratives that Ouspensky and Booth discuss are sometimes abrupt or shocking; the shock, I suggest, is akin to that of a mixed metaphor, for reasons I have discussed above. Along what lines, then, can "point of view" shift?

            Ouspensky specifies four "planes" or "basic semantic spheres in which viewpoint may generally be manifested . . . planes of investigation in terms of which point of view may be fixed" (1973:6), only one of which is the "spatial and temporal plane." Foucault writes that the Chinese encyclopedia order is "spatially incoherent," but a comparison with Ouspensky shows in what a complex way Foucault means "spatial." Ouspensky cites (using Yuri Lotman's description of ) the character of the Nose in Gogol's story of the same name. The Nose has a face; it walks, stooping; it runs up stairs, wears a full dress coat, and prays with an expression of devout piety. All of these attributes of the Nose "completely destroy the possibility of imagining it in three-dimensional space" (ibid.: 78). Gogol's description of the Nose is indeed primarily spatially distressing, whereas Borges's list of definitions of "animal" moves us in a much more complicated way in terms of our implied involvement with the "animal" in question and the whole domain or context of the encyclopedia entry. In this way, the exotic and foreign "Chinese" encyclopedia demands a flexibility we do not consciously have in much more complex and conceptual ways than does the Nose, which merely defies visualization. Yet Foucault calls Borges's incoherence "spatial."

            Benjamin Lee Whorf and many other writers have situated this spatializing tendency deeply within the European "relation of habitual thought and behavior to language." What Whorf calls "metaphors of spatial extension" for the expression of the experiences of duration, intensity, and tendencies are just one "part of our whole scheme of OBJECTIFYING' (1956: 145). This conceptualization of various aspects of life in spatial terms and the resulting treatment of them as such entails the bleeding over of certain ways of knowing space to these other realms. Because of this conflation of many dimensions of existence into the notion of space, Ouspensky's and Booth's works (although this is far from their intended purpose) can help us examine the range of experiences and conceptions we automatically treat in spatial terms.

            Ouspensky's work shows that the "point-of-view" metaphor condenses "ideological," "phraseological," "psychological," and "spatial and temporal" planes. Because of the richness and complexity of this condensation, we can afford to lose sight of the fact that "point of view" is a metaphor. We forget about the sense of formalism that the adamantly spatial idea of "composition" brings; "point of view" implies that position is divorced from what is being observed from that position, a form-content dichotomy in which we recognize the split between "spatial form and spatial formless continuum," which Whorf (1956: 147) observed was inseparable from our tendency to deal with various aspects of life as images, as spatial.

            Harries discusses the manifestations of point of view in art, showing its movement from a "pre-Cartesian" form, where there is point of view, but no orderly Academic perspective -- “every part of the painting stands in a relationship of tension to other parts: its different parts cannot be synthesized from one point of view. We cannot take its measure" (1968: 42) -- to the development of the Renaissance unity we have come to identify and expect from art; a unity that "has its foundation in the demand that the work of art be designed with respect to a particular point of view. This results in a new homogeneity; all objects are located in one space and are enveloped by the same light and atmosphere . . . the world . . . in the image of reason" (ibid.: 16- 17; emphasis mine).

            But can worlds really be constructed in the image of reason? They might be reduced to it, yet seem to be constructed more poetically or opportunistically than reasonably. James Fernandez (1982), in his studies of revitalization movements in which new worlds are attempted, finds in them a shiftiness in this regard. Indeed, he goes on to argue (1986 and unpublished lecture, 1987) that this "shape-shifting" is characteristically human and that in everyday life we constantly shift from metaphor to metaphor in our quest for understanding of situations. (Yet, as we have noted, if we were truly committed to the self-consciousness that strives to make assumptions explicit and to be theoretically consistent and truthful to what we start, we ought not to shift.) The implications of any particular predication we make on the world when we think it and act in it never get completely worked out. Shifts occur in our understanding, in the "image" we have of the situation, and suggest different ways of performing. This way of operating, which is very practical, differs from our idea of systematic scientific examination. Yet we see that the methods of art and of science are, in practice, not so absolute: throughout the unfolding of his organic metaphor for society, Spencer shifts, using any number of discrete versions of organic models, supported by a constant sliding from literal to figurative senses. It is a creative mode of thought, and it behaves like one. What I would like to call attention to, however, is that the definition of reality or truth taken from science and philosophy is permeated by ideological notions or norms of consistency and parsimony which are related to the desire for "objectivity." These norms and expectations are brought to bear upon the social sciences and the arts, resulting in a formalism or separation of form and content.

            Certainly figurative language has a dizzying reservoir of possibility. Since in such language "all nature opens her stores, and allows us to collect without restraint," style texts give us rules to guide our choices, invoking the golden mean, some ideal of a safe, prudent way between extremes. Metaphors "must be suited to the nature of the subject; neither too numerous, nor too gay, nor too elevated for it," adds Richard Parker (1852: 117; he paraphrases Hugh Blair). Metaphors of a "discreet Author" are not too long: one writer, with a tendency to "run it on into the tiresome lengths of childishness and affectation," is recalled by Addison as "having said by chance that his mistress had a world of charms, thereupon took occasion to consider her as one possessed of frigid and torrid zones, and pursued her from one pole to the other" (1859: 388). "The reader is wearied, and discourse becomes obscured" (Parker 1852:118). This moderation points again to a valuation of the invisible rhetorical artifice, and to how mixing metaphor, by calling our attention to the device, results in failure to produce persuasion. Our notion of reality, that by which we are persuaded, is of a state of mind from within one model. It was traditionally a tenet of art that the means of creating the illusion must never be noticeable, lest the illusion fail. We have seen this ideal in both art and science.



Arguments and Scientific Models as Spatially Grounded:

The Notion of Structure


Mr. Speaker, I smell a rat; I see him floating in the air; I hear him

rustling in the breeze; but I shall nip him in the bud.


                                    W. Crafts and H. Fisk, Rhetoric Made Racy or Laugh and Learn


The charm dissolves apace,

And as the morning steals upon the night,

Melting the darkness, so their rising senses

Begin to chase the ignorant fumes that mantle

Their clearer reason.


                                       Shakespeare, The Tempest (Act 5, Scene 1)


            The issues of persuasion and convincing without shiftiness bring along with them nuances of illusion and deceiving, which seem to be both necessary (as Aristotle implies) and dangerous, as we shall now see. The "confounding of the imagination" from which mixed metaphor, that "unnatural medley" as Parker calls it, is said to emerge land which it is said to produce) is often experienced as a suspicion that the author of the mixed metaphor is not quite honest. He or she means to pull the wool over our eyes. Of course, the question of whether these metaphors are in the context of a poem, a political speech, a scientific treatise, a sermon, or an ethnography matters; but, surprisingly, suspicion about the professional integrity and/or moral character of the author appears in all of these contexts. Adam Smith, for example, evaluates metaphors in terms of their "justice" (1971: 26-27). In the case of rhetoric, one argument concerning the danger of using metaphor in general is that


the unthinking majority are content with the crudest generalizations and will accept a picture for an argument. Counting on this, the mob orator . . . will tacitly assume that such and such a resemblance or analogy is complete in all points and will begin to draw arguments therefrom. All opposing facts and differences are quietly ignored. Moreover, the appeal is not really to reason, even unreasonable reason, but rather to passion. (Brown 1966: 238; emphasis mine)


This idea of the kinds of metaphoric pictures that make for acceptable argument raises philosophical questions. No one has addressed these more directly than Stephen Pepper.

            In World Hypotheses, Pepper explicitly proposes that, at certain points in theoretical thought, just such an assumption that "the analogy is complete" is necessary, exactly because of the incorrectness of mixing root metaphors. Since, Pepper argues, all facts are not explicable by any one hypothesis, some major facts about the world are tested according to a potentially useful hypothesis, and then if it seems to be a powerful one, the assumption is made for the rest of the facts (1942: 97-98). Each "world hypothesis" (based on one "root metaphor") is autonomous, and, Pepper says, the interpretations of each make such a convincing picture that, if one hasn't compared them with parallel interpretations of rival hypotheses, one will inevitably accept them as indubitable and self-evident (ibid.: 101). There are two ways of acknowledging other theories, in Pepper's view. On the one hand, "we need all world hypotheses, so far as they are adequate, for mutual comparison and correction of interpretive bias" (ibid.: 101; emphasis mine). On the other hand, "eclecticism is confusing"; "if world hypotheses are autonomous, they are mutually exclusive" (ibid.: 104).

            Although Pepper says that this last is a cognitive distinction, rather than a practical one, and that practical matters "often entail other than cognitive considerations" (ibid.), he does not follow up on this important distinction. He raises the question of whether a more adequate world theory can be developed by selecting "the best" in each theory and organizing these bits with a synthetic set of categories, as he claims Alfred North Whitehead and other philosophers do. But he answers that "the eclectic method is mistaken in principle in that it adds no factual content and confuses the structures of fact which are clearly spread out in the pure root-metaphor theories; such mixing is almost inevitably sterile and confusing" (ibid.: 106; emphasis mine). We are deposited back where we began, with the usual objections: that mixed metaphor is confusing and, in addition, sterile (as mating snakes and doves, in Horace's words, must be). Monstrous imagery is invoked:


But why do it? As a flight of fancy it may be amusing, as men have fancied fauns, centaurs, angels, and dragons, but it can hardly be a genuinely creative cognitive achievement. If a man is to be creative in his construction of a new world theory, he must dig among the crevices of common sense. There he may find the pupa of a new moth or butterfly. This will be alive and grow, and propagate. But no synthetic combination of the legs of one specimen and the wings of another will ever move except as their fabricator pushes them about with his tweezers. Moreover, what happens at the joints? What happens under the skin between the centaur’s neck and body? How do the wings of angels fit into their shoulders? Either the eclectic glosses these difficulties over or we perceive confusion. (ibid.: 112; emphasis mine)


            This argument is based on an ideal of an organic sense or life felt to reside only in whole, unbroken models. Thus the trace of the organic metaphor in concepts of culture and society implied by words like “system.” The definitive “picturing” test for mixed metaphor shows quite clearly what sort of image is considered incorrect; we as a culture seem to need to deny that we have use for these mythological beasts, except to classify them as marginal beings which, like medieval and age-of-exploration conceptions of creatures at the unimaginable fringes of the world, are in need of civilization. Not only “the unthinking majority” accepts pictures as arguments.

            Here Pepper treats the deliberate mixing of models, yet the monsters, yet the monsters he uses to speak of contrivance are the same as the products of impassioned or sloppy, inept mixtures, and the same sort as Spencer’s organic metaphor, outlined in note 3, suggests. Pepper says that ultimately, the strength of world hypotheses comes from structural corroboration and that we find them credible because their “scope is unlimited”; eclectic efforts limit the scopes of the contributing metaphors, have no root metaphor, and therefore do not move cognition forward at all (ibid.:112). I am not clear on what kind of cognition this is. It seems to be the same kind that Thomas Kuhn assumes to be involved in the business of “normal science,” which precludes the mixing of paradigms.

            This part of Pepper’s argument is quite similar to Kuhn’s descriptions of the incompatibility and actual impossibility of two or more paradigms coexisting during any period of productive “Normal science”; both writers take this into account as an inevitable narrowing-down of the world to fit into the conceptual categories indicated or established by the reigning metaphor or paradigm. Kuhn writes that "the enterprise seems an attempt to force nature into the preformed and relatively inflexible box that the (learned) paradigm supplies. No part of the aim of normal science is to call forth new sorts of phenomena" (1970: 24). Novelties subversive of the basic commitments of the present paradigm are suppressed. Changing paradigms is considered a waste of scientific energy: "As in manufacture so in science; retooling is an extravagance to be reserved for the occasion that demands it." Kuhn's relativist claim is that once students learn their paradigm's language, that paradigm is a gestalt-like absolute and that a shift "must occur all at once . . . or not at all" (ibid.: 150). Thus, as in Pepper's model, worldviews are incommensurable and do not mix.



Shiftiness and the "Reality"

of Social Science


When a man says he promises anything, he in effect expresses a

resolution of performing it; and along with that by making use of

this form of words, subjects himself to the penalty of never being

trusted again in case of failure.


               David Hume, A Treatise of Human Nature


According to both Pepper's and Kuhn's theories, mixed metaphor inevitably appears, but it is not normal. In general, the fear is that the "black art of metaphor" not only can confine our outlook and conceal evidence, but can also seduce us by changing models midstream, thereby convincing us of virtually anything (assuming that we are completely seduced).

            The shiftinesses of metaphor and particularly of mixed metaphor slide inevitably into shiftiness of the author's character; anyone who displays more than one single "own vision of things" must be shifty, the type liable to "engender snakes with doves." "The experience seems quite impossible. The metaphor does not ring true. The writer is convicted of falsehood to his own vision of things" (Buck 1899: 64). Again, this points to the conclusion that one image used in an argument is experienced as a promise, a commitment to control the future, to synchrony; two images violate that promise, for they suggest the imminence of other "more promising" arrangements.

            If certain relations (or even communications) between people are constituted by a promise (as in a sense marriage is, going back to our metaphor), then, as Peter Wilson observes, breaking such a promise "results in the self-destruction of the relations.... If the failure of promises amounts to the disintegration of structure, then this failure must be regulated" (1980: 101) -- that is, there must be a taboo.

            An additional "problem" with eclecticism is revealed by another class of major complaints about mixed metaphor. On the one hand, mixed metaphor can cause unclarity and muddled confusion: "If the terms of the comparison become too soft and indefinite, the poetry disappears into a shimmering incoherence, . . . the relation between terms dissolves into a vague emotional blur" Brooks 1965: 321-23). On the other hand, this confusion from mixed metaphor can be portrayed as a much more abrupt thing: an enigma, riddle, or mystery Aristotle 1951: 83). Either way, perusal of it becomes "fatiguing" (Quintilian 1891: 127). Work is needed -- and one may feel that no amount of work or "words supplied by the mind of the reader" (Bett 1932: 243) will help attain the desired end point, which seems to be coherence, integration, a unitary "finished understanding." If closure is the goal, mixed metaphors only "tease."

            As Elliott Coleman has noted, Aristotle and Quintilian also said that metaphor itself "names the nameless" (1965: 151) and so fills semantic gaps, indeed says more than single words. Yet by thus pointing to complex relationships and images rather than to already known "kinds," metaphor must necessarily tend to ask the reader to do some work. Keith Basso writes that "metaphorical concepts must be attained on one's own through private acts of discovery and recognition.... Such acts are creative in the fullest and most genuine sense, for they presuppose and exemplify an ability to arrange familiar semantic features into unfamiliar combinations, to form fresh categorizations" (1976:110). "T.D. " writes that "moderating the intellect" may be even more dangerous than to attempt the more challenging broken metaphor: "Nothing can be more nauseous than a long drawn-out simile, crawling upon as many legs as a centipede" (1825: 722). There is an implied ideal in many of the "confusion" complaints, also implicit in the stipulation that rhetoric should not be noticed. This is the ideal that the truth of a scientific) matter should be brought to us and deposited at our feet, so that we need do nothing but passively absorb it, as opposed to some supposed license of "art" to impressionistically allude to things or to "refer" to them "symbolically."

            Tompkins writes that "On the one hand, the Metaphor, though shorter than the Simile, usually makes the mind do more work; on the other hand, the mind is rendered more able to do work, not, however, because it is gratified, but because it is stimulated to exertion" (1897: 310; emphasis mine). Let us then look at the way social scientists, especially anthropologists, work with metaphor, for they cannot, given the complexity of human relationships and human expressivity, simply lay the truth at our feet.

            Gilbert Lewis, in his ethnography of the Gnau of New Guinea, uses concepts from Ernst Gombrich in his discussion of the "framing" of ritual symbols and of the way in which ordinary images and actions may thus "gain mystery" and ultimately additional meaning. Informants may explain a detail; there may be a straight-forward explanation or an ordinary object involved; yet, in situations that strike us as "mysterious," we are nevertheless strangely "alerted" to the presence of the symbolic. Something is inappropriate, out of place; it must be a symbol. "Thus we become confused, speculation is invited" (Lewis 1980: 30). By violating convention, ordinary practice, or common knowledge, one can call attention to mystery, to complicated issues in all their complexity, by isolating elements usually known in other contexts. Alternatively, one may, if clarity is the goal, set the object or action round with clues to its meaning. The anthropologist, like the connoisseur, becomes sensitized to such stylistic cues and "learns to seek or judge whether what is aimed at is clear, unambiguous representation; or whether he, just like those within the society, is being asked to speculate and perceive something familiar in a new light. Isolation and cognitive dissonance may be recognizable as techniques at work in ritual" (ibid.; emphasis mine).

            Of course, Lewis continues, the anthropologist is not free to assume that everything he does not understand aims at mystery in this way -- yet "there are suggestive indications when the spectator is invited to 'ungate' through the isolation and situational oddity of some object or action or ritual" (ibid.: 31).

            One marginally appropriate use of mixed metaphor which several sources grudgingly allow is when it is expressive of a mood we are to attribute to the utterer or the situation. This, of course, in a poem or a sermon, not in a scientific context, where it is expected that what cannot be given explicitly ought not be given. George Whalley argues against "common-sense 18th-century advice" prohibiting the promiscuity of "making two inconsistent metaphors on one object." He asserts: "A successful 'mixed metaphor' seems not so much to clarify the single image as to establish a certain 'tone.' ... Hamlet's violent desperation is conveyed by a wretched incongruity" (1967: 150) . And Bett writes that in a speech of Macbeth, "the metaphors are outrageous.... But despite this there is here a real justification of a psychological kind.... We feel that [the murderer's] insincerity is reflected in these strained and artificial metaphors" (1932: 244-45). Note that these sanctions of the mixing of metaphor when it expresses or generates a mood imply that the internal organization of the metaphorical domain may be "unrealistic" if (and only if) it is to be experienced as incongruously organized, that is, experienced formally, as representative of the kind of order or disorder it exists in, or, as Foucault says, the sort of "site" it could occur on.

            The discovery of mixed metaphors in the work of "learned writers" also gets a different, more approving response. Those mixtures found in Shakespeare and Milton caused a great deal of debate and searching for justification over the years. Although Hamlet's line on taking arms against a sea of troubles was more than once, in the eighteenth century, actually pronounced some sort of misprint, originally written by Shakespeare as "assail of troubles" or other inoffensive version, John Ruskin found what he called "compressed significance" (Bett 1932: 243) in a mixed metaphor from Milton's Lycidas; and Gibbons finds some strings of metaphors to be "actually distinct sentences" (1969: 32). This interpretation of mixed metaphor as an acceptable or at least comprehensible "ellipsis," which I mentioned earlier (Bett 1932: 243), is invoked in such circumstances.

            In "Edification by Puzzlement," Fernandez discusses the sermons of the Bwiti syncretist religious movement of the Fang. By using many "likenesses" in the sermons, Fernandez says, the Bwiti leader tries to "knit the world together," a world that is supposed to be "one thing, but the witches try to isolate men from each other so they can eat them" (1986: 179). Above, I quoted Wallace Stevens that "reality is the indispensable element of each metaphor." The Bwiti sermon with its endlessly mixed metaphors, says Fernandez,


sends us elsewhere to obtain our answers. They are rich in images which must, however, be contextualized by extension into the various domains of Fang culture. The interpretive task is therefore to move back and forth between text and context.... These sermons are examples of what Vygotsky . . . has called "thinking in complex.” The sequence of images (the body images, the forest images, the vital liquids images, the suspended things images, the food images) put forth are not dominated by any overall conceived and stated purpose or by any dominant image.... New materials from various domains of Fang experience are introduced on the basis of association by similarity or contiguity, contrast, or complementarity with this sequence. But then again, abruptly, new elements with all their alternatives are allowed to enter the thought process and raise new thematic preoccupations and to suggest new possible nuclei of attention. By any standard of administered intellectuality, such sermons seem diffuse and spontaneous in the extreme. (ibid.: 181-82)


            This use of mixed metaphor in preaching in a distinctly different culture reminds us that there are text writers who admit the uses of mixed metaphors; yet this has been done close to home as well. Gibbons quotes Longinus and Demosthenes in making a general exception to the rule "no mixing": he says "we may on the same subject and in a manner in the same breath, introduce very different Metaphors ... two or three at the most.... The time of using them is when the passions rush like a torrent, and bear along with them a multitude of Metaphors as necessary for the occasion" (1969: 30). Gibbons describes the way


the sacred indignation of the Apostle blazes out and ceases, blazes out and ceases again, till he has finished his account of those most profligate wretches whose characters he was representing.... I aver that seasonable and vehement passions, and a noble sublimity, are a sufficient apology for the number and boldness of Metaphors; for it is natural for the passions . . . by their own impetuous violence, to seize and carry all before them, and therefore as by an absolute necessity they challenge the boldest Metaphors; nor will they give leisure for the hearer to cavil against their number, as they inspire him with all the ardor of the Speaker. (ibid.: 33; emphasis mine)


            Mixing metaphor for the creation of "mood" calls attention to incongruity and so is excused. Passionate, seductive mixing moves us through not calling attention to the device. The above arguments concerning the passionate mixing of metaphor go well with what Fernandez writes concerning the "mission of metaphor" in Bwiti sermons, "to turn inchoate objects of religious attention into something graspable and then to turn them into each other, to reconcile them, . . . suggesting an overarching integrity of things, a cosmology, by a convincing form of argument: the argument of images" (1982: 546). In general, it seems that the rhetoric of mixed metaphor has a home in religious contexts because of the performative implications, as well as because of a general acceptance that it is a legitimate purpose of sermons to seduce their listeners. It is a Durkheimian vision of a profane and broken-up world made sacred through a contagion of collective, often mixed-up imagery. "By forcing contextualization on the auditor, the cultural experience he is obliged to extend his interpretations to and consult is revitalized . . . cosmogony of an important kind" (Fernandez 1986: 181).

            The important point is that cosmogony emerges out of mixing or "ill-formedness." Basso argues against Noam Chomsky's limitation of creative possibility in language use to "the unfolding of existing structures." "Wise words are prime examples of appropriately ill-formed utterances," he writes, citing Clifford Geertz in attributing the power of metaphor to generate what Basso calls novel semantic categories to "the interplay between discordant meanings" (Basso 1976: 111). The gifted maker of metaphor, who speaks in semantic contradictions, says Basso, is not "consistently obedient to the strictly grammatical rules of language" and violates rules rather than unfolding them when he produces and understands novel utterances (1976: 116). Harries notes how


Challenging Aristotle's claim that "a good metaphor implies an intuitive perception of the similarity in dissimilars," recent discussions have often insisted that, in poetry at least, metaphor joins dissimilars not so much to let us perceive in them some previously hidden similarity but to create something altogether new. As C. Day Lewis claims, "we find poetic truth struck out by the collision rather than the collusion of images." (1978: 73)


            We have seen that the eclectic, colliding, mixing use of metaphor is excused if it happens to be done by great writers or speakers who are given the benefit of this doubt; it is often seen as a side effect of some higher level of functioning. Even Pepper writes that though eclecticism "does not move cognition forward at all, . . . all (or nearly all) the great philosophers were in various degrees eclectic." His three reasons for the occurrence of eclecticism are: (1) "undue faith in self-evidence and the indubitability of fact"; (2) "the desire to give credit to all good intuitions with the idea that these all have to be put inside one theory"; and (3) the recognition that "the great philosophers were not so much systematizers as seekers of fact, men who were working their way into new root metaphors and had not yet worked their way out of old ones. The eclecticism of these writers is, therefore, cognitively accidental and not deliberate, though psychologically unavoidable" (1942: 106; emphasis mine).

            All three of these situations are familiar from our discussions of mixed metaphor and its perpetrators, the inspired, the impassioned, the unifiers of the world, the people in confused states of mind. "Confused" has negative connotations, whereas Pepper describes "exceptional" people in transitional states, looking around, in process. Are these "great philosophers" related to the "learned writers" who, we assume, have some "higher" reason to break metaphors, that we should try to interpret? Kuhn says that a multiplicity of paradigms appears at two stages in the scientific process. The first is in early developmental stages of most sciences, when many views of nature are competing and being tried out and there is no common body of belief (1970: 4). He adds, by the way, that "It remains an open question" what parts of social science have yet gotten beyond this stage and reached the point where one paradigm, "which emphasizes only some part of the too sizable and inchoate pool of information," triumphs (ibid.: 14, 20). The second stage -- at which we see the rules for normal research loosened and "numerous partial solutions" appearing, along with a "proliferation of competing articulations, a willingness to try anything, the expression of explicit discontent, the recourse of philosophy and debate over fundamentals" (ibid.: 91) is during a transitional crisis or "scientific revolution." These crises are characterized by ad hoc adjustments of theories with other theories, blurring of models, and in general, "a world out of joint," as described by Copernicus: "It is as though an artist were to gather the hands, feet, head and other members for his images from diverse models, each part excellently drawn, but not related to a single body, and since they in no way match each other, the result would be monster rather than man"; and by Einstein: "It was as if the ground had been pulled out from under one, with no firm foundation to be seen anywhere, upon which one could have built" (both from Kuhn 1970: 83).

            Pepper insists that the "great [eclectic] philosophers" intuit a new root metaphor, but the intuitions are yet primitive, lack a technical vocabulary, and so are enmeshed in the categories of a former metaphor. We should honor their "keen sense for the facts," he says, not their "confusing eclecticism"; the cognitive value of the work of these men "comes not from the eclectic factor (which is entirely obstructive), but from the creative factor" (1942: 107).

            Despite this, or perhaps because of it, what we are seeing emerge is a general, rigorous, and absolute sanction against mixed metaphor and an indictment of its creators in all domains of our culture, yet at the same time a general sanctioning of it, or at least a certain forgiveness, in the case of either great men, authorities, innovators, and geniuses or the subjects of anthropological inquiry. We allow these (however we determine who they are) to break the very rules which, according to Kuhn, are necessary for normal science to move on.4

            The struggle over understanding how worlds are to be constructed is seen in microcosm in the struggle to understand how the "small world" of the person is to be constructed. Here too we find the same authority of ideas of unity, coherence, wholeness. And the matter is not irrelevant to our topic, particularly if we understand that we "build" ourselves metaphorically by predicating various metaphors on ourselves or on others (Coleman 1965; Fernandez 1972).

            Robert Barrett describes schizophrenia as a concept that pertains to failure to achieve our culture's ideal of personhood, as the "quintessential disease category which absorbed and epitomized" the images of a central nineteenth-century debate within psychiatric thought, a debate that "revolved around two concepts of personhood in tension -- on the one hand the person as a unified indivisible ego and on the other as a divided ego" (1987: 21). As the description of failure, in effect, to be a social person, the concept of schizophrenia reveals negatively valued images (Barrett says "symbols") that imply by opposition Western ideological expectations of individualistic "personhood."

            The central features of this individualistic concept, Barrett writes, quoting Paul Hirst and Penny Wooley, are


those of boundedness, autonomy and indivisibility. It is the conception of the person as centered in a unitary consciousness, a subject "self-possessed" in its conducts, experiences and thoughts, all of which flow from a single and continuous origin. Consistency in conduct, selforiginated action, and unity of memory and experience are all criteria of normality of this subject as conscious origin of itself. (ibid.: 7)


This "autonomous, integrated, atomistic, indivisible individual characterized by reflective self-consciousness" (ibid.: 3) is the ideal type, says Barrett, of "person" in our culture, "person" defined (following Donald Pollock) as "a being to whom members of a culture attribute the capacity to be an agent of meaningful action" (ibid.: 6). In this "undivided totality in relation to God" (ibid.: 21), of course, we recognize the Platonic ideal.

            The concept of "ego splitting," on the other hand, appears


in the empirical ego psychology of the early nineteenth century, encapsulated most clearly in the realist philosophy of Herbart. . . In common with anthropological psychiatry this tradition located personhood in consciousness.... However the commitment to realism also led to a view of the person as an aggregate of multiple egos. Since consciousness changes depending on changing situations and relations, the ego is constantly changing -- a heterogeneous ego. (ibid.: 22)


            The ideology of split personhood that derived from this tradition "found clearest enunciation by Janet, whose concept of 'dissociation' referred to a weakening of the highest integrative functions, leading to a loss of the reality functions" (ibid.: 23). Thus we see here again the same clustering of the same terms -- reality, unity, integration, autonomy, etc.



   Conclusion: Purity and Opportunity


      We never arrive intellectually. But emotionally we arrive constantly.


                                               Wallace Stevens, Opus Posthumous


            Mixed metaphor, besides being "untrue" and "unreal," is an impure union: Rudolf Flesch says that "clearly, if we want to avoid any misunderstanding, the best thing is not to use any rhetoric whatsoever. In other words, we must try not to play any games with our words or ideas" (1946: 104). Flesch offers a list of "plain talk" rules such as "Do not use metaphors without an explanation"; "Do not use periodic sentences"; and "Do not use irony (half the people won't get it)." Well. Enough said. Except that, as Rudolf Arnheim notes, "If simplicity were the one overriding goal of art, evenly stained canvases or perfect cubes would be the most desirable art objects." In art, "the simplicity principle rules unopposed only in closed systems" (1974: 410-11). One aspect of such closed systems is their autonomy, which I have touched on above. Autonomy is apparently an important part of the ideal that I am discussing and that mixed metaphor challenges. Another aspect is what Goodman calls the "touchstone of realism," the ease with which information issues from a work of art. This ease depends, he says, on "how stereotyped the mode of representation is” (1976: 36). The nature of the "realism" of metaphorical domains is too shifty -- it ultimately refers back to "what is conventional" or persuasive, an ideal of a "living" coherence. As we have seen, however, neither the etiquette of metaphor nor the ideology behind it is consistent; mixed metaphor is criticized for being unpersuasive, yet feared and allowed for its persuasive power.

            Harries writes that there is a tension in our expectations of art in that, although we judge works of art by their unity, if metaphor opens a work of art to a dimension that transcends it, refers to things outside of it, then "it destroys our experience of the work of art as a self-sufficient whole." But unity is not necessarily as simple a thing as Flesch might maintain. Harries says that a demand for unity has not traditionally


denied complexity, tension and incongruity, but order should triumph so that what may at first appear to be discordant elements are in the end recognized to have been absolutely necessary. In a successful work of art [traditionally] nothing is superfluous, while it is impossible to add anything without weakening or destroying the aesthetic whole. The task of the observer or the reader is only to open himself to the artist's creation. Emphasis on the unity of the aesthetic object thus implies aesthetic distance, separation of art from life (1968: 74; bracketed comment mine).


            As we have seen, truth seems to imply knowledge of some whole, knowledge only possible from outside that integrated, structured, distanced whole. Thus the "pure knowing subject," as Harries calls him, must be imagined in the place of God -- that is, sharing with God the quality of being one's own foundation. The autonomy of the ideal objective observer and the autonomy of which he has this objective knowledge seem inherently linked --  which presents paradoxes. We expect works of art and other "wholes" both to point beyond themselves, have meaning, be open, and also to be closed, finished, with the presence of things of nature.


More and more we tend to make our ability to comprehend the measure of reality. Our commitment to objectivity and transparence forces us to see the world that moves us with its sights and sounds as no more than the perspectival appearance of a reality that yields its secrets only to the dislocated spirit. So understood, reality leaves no room for genuine mysteries. (Harries 1978: 87-88; emphasis mine)


            Such is the predicament involved in one debate over mixed metaphor which satisfies emotionally but irritates intellectually. But as the epigraph from Stevens makes clear, it is one thing to arrive emotionally and quite another to arrive intellectually. And it may just be that it is this emotional world of constant arrivals that anthropologists and poets are obliged to live in -- as at least their first world. And it may be for that reason that they – unlike their more formal colleagues in the academy -- are prepared to admit the creative power for both world and person of mixed metaphor.

            Perhaps consciousness cannot leave room for "genuine mysteries." What is called mixed metaphor, as we have seen above, is the coming into consciousness of a mixing that goes on all the time, a consciousness that offends our sensibilities because it "calls attention to the device" and perhaps might reveal the inexplicable bases of our worldview.






I am indebted to James Fernandez for his patient readings of drafts of this article and for his helpful suggestions at every stage of the writing. For comments that sparked my imagination as well as watered it, I am grateful also to Steve Coleman, Paul Friedrich, Alaina Lemon, and Bob Peters.


1. Goodman 11976: 10-191 illustrates how eminently conventional and distanced from everyday physical experience is the "realism" of Renaissance perspective Below, I discuss Ouspensky's work with the point-of-view metaphor, and Harries' discussion of point of view and perspective in art.


2. In Needham's (1985) article on Psalmanaazaar, the authority of a wholly fabricated account of an invented culture is attributed to the articulation of details into systems (as opposed to the mere accumulation of a mass of ethnographic data).


3. For example, a potentially troubling difference that Spencer notices between the organic and social organizations is that "while the ultimate living elements of an individual organism are mostly fixed in their relative positions, those of the social organism are capable of moving from place to place." He makes good of this, though, realizing that "while citizens are locomotive in their private capacities, they are fixed in their public capacities” (1981: 396); a shift from a literal to a metaphorical sense of "place." In another such case, where the "reality" of the metaphor is challenged, Spencer juggles several senses of the word "feeling": physical sensibility, emotional and intellectual sensibility, and consciousness (ibid.: 397). On one page, a person is a liver cell, passing away and being replaced; the next moment, however, blood cells are coins, and humans the active agents who propel the currents of commodities

(ibid.: 418-19).


4. Mary Douglas, in her 1966 work on taboo, posits a structural analogy between pollution and sacredness. Punty and Danger could be read as an extended essay on the abhorrence of mixed metaphor: "though we seek to create order, we do not simply condemn disorder. We recognize that it is destructive to existing patterns; also that it has potentiality. It symbolizes both danger and power (1966: 114).