it's due in 24 hours

2

Two Modes ofThought

Let me begin by setting out my argument as baldly as possible, better to examine its basis and its consequences. It is this. There are two modes ofcognitive functioning, two modes of thought, each provid- ing distinctive ways of ordering experience, of constructing reality. The two (though complementary) are irreducible to one another. Ef- forts to reduce one mode to the other or to ignore one at the expense of the other inevitably fail to capture the rich diversity of thought. Each ofthe ways ofknowing, moreover, has operating principles of

its own and its own criteria ofwell-formedness. They differ radically in their procedures for verification. A good story and awell-formed argu- ment are different natural kinds. Both can be used as means for con- vincing another. Yet what they convince ofis fundamentally different: arguments convince one oftheir truth, stories oftheir lifelikeness. The one verifies by eventual appeal to procedures for establishing formal and empirical proof. The other establishes not truth but verisimilitude. It has been claimed that the one is a refinement of or an abstraction from the other. But this must be either false or true only in the most unenlightening way. They function differently, as already noted, and the structure of a

well-formed logical argument differs radically from that of a well- wrought story. Each, perhaps, is a specialization or transformation of simple exposition, by which statements offact are converted into state- ments implying causality. But the types ofcausality implied in the two modes are palpably different. The term then functions differently in the

11

12 · Two Natural Kinds

logical proposition "if x, then y" and in the narrative recit "The king died, and then the queen died." One leads to a search for universal truth conditions, the other for likely particular connections between two events-mortal grief, suicide, foul play. While it is true that the world ofa story (to achieve verisimilitude) must conform to canons of logical consistency, it can use violations of such consistency as a basis ofdrama-as in the novels ofKafka, where nonlogical arbitrariness in the social order provides the engine ofdrama, or in the plays ofPiran- dello or Beckett, where the identity operator, a = a, is cunningly violated to create multiple perspectives. And by the same token, the arts of rhetoric include the use ofdramatic instantiation as a means of clinching an argument whose basis is principally logical. But for all that, a story (allegedly true or allegedly fictional) is judged

for its goodness as a story by criteria that are of a different kind from those used to judge a logical argument as adequate or correct. We all know by now that many scientific and mathematical hypotheses start their lives as little stories or metaphors, but they reach their scientific maturity by a process ofconversion into verifiability, formal or empir- ical, and their power at maturity does not rest upon their dramatic origins. Hypothesis creation (in contrast to hypothesis testing) re- mains a tantalizing mystery-so much so that sober philosophers of science, like Karl Popper, characterize science as consisting principally of the falsification of hypotheses, no matter the source whence the hypothesis has come. Perhaps Richard Rorty is right in characterizing the mainstream ofAnglo-American philosophy (which, on the whole, he rejects) as preoccupied with the epistemological question ofhow to know truth-which he contrasts with the broader question ofhow we come to endow experience with meaning, which is the question that preoccupies the poet and the storyteller. Let me quickly and lightly characterize the two modes so that I may

get on more precisely with the matter. One mode, the paradigmatic or logico-scientific one, attempts to fulfill the idealofa formal, mathemat- ical systemofdescription and explanation. It employs categorization or conceptualization and the operations by which categories are estab- lished, instantiated, idealized, and related one to the other to form a system. Its armamentarium ofconnectives includes on the formal side such ideas as conjunction and disjunction, hyperonymy and hy- ponymy, strict implication, and the devices by which general proposi-

Two Modes ofThought · 13

tions are extracted from statements in their particular contexts. At a gross level, the logico-scientific mode (I shall call it paradigmatic hereafter) deals in general causes, and in their establishment, and makes use of procedures to assure verifiable reference and to test for empirical truth. Its language is regulated by requirements of consis- tency and noncontradiction. Its domain is defined not only by observ- abIes to which its basic statements relate, but also by the set ofpossible worlds that can be logically generated and tested against observables- that is, it is driven by principled hypotheses. We know avery great deal about the paradigmatic mode ofthinking,

and there have been developed over the millennia powerful prosthetic devices for helping us carry on with its work: logic, mathematics, sciences, and automata for operating in these fields as painlessly and swiftly as possible. We also know a fair amount about how children who are weak initially at the paradigmatic mode grow up to be fairly good at it when they can be induced to use it. The imaginative applica- tion of the paradigmatic mode leads to good theory, tight analysis, logical proof, sound argument, and empirical discovery guided by rea- soned hypothesis. But paradigmatic "imagination" (or intuition) is not the same as the imagination of the novelist or poet. Rather, it is the ability to see possible formal connections before one is able to prove them in any formal way. The imaginative application of the narrative mode leads instead to

good stories, gripping drama, believable (though not necessarily "true") historical accounts. It deals in human or human-like intention and action and the vicissitudes and consequences that mark their course. It strives to put its timeless miracles into the particulars of experience, and to locate the experience in time and place. Joyce thoughtofthe particularities ofthe story as epiphanies ofthe ordinary. The paradigmatic mode, by contrast, seeks to transcend the particular by higher and higher reaching for abstraction, and in the end disclaims in principle any explanatory value at all where the particular is con- cerned. There is aheartlessness to logic: one goes where one's premises and conclusions and observations take one, give or take some of the blindnesses that even logicians are prone to. Scientists, perhaps be- cause they rely on familiar stories to fill in the gaps oftheir knowledge, have a harder time in practice. But their salvation is to wash the stories away when causes can be substituted for them. Paul Ricoeur argues

14 · Two Natural Kinds

that narrative is built upon concern for the human condition: stories reach sad or comic or absurd denouements, while theoretical argu- ments are simply conclusive or inconclusive. In contrast to our vast knowledge of how science and logical reasoning proceed, we know precious little in any formal sense about how to make good stories. Perhaps one of the reasons for this is that story must construct two

landscapes simultaneously. One is the landscape of action, where the constituents are the arguments of action: agent, intention or goal, situation, instrument, something corresponding to a "story grammar." The other landscape is the landscape of consciousness: what those involved in the action know, think, or feel, or do not know, think, or feel. The two landscapes are essential and distinct: it is the difference between Oedipus sharing Jocasta's bed before and after he learns from the messenger that she is his mother. In this sense, psychic reality dominates narrative and any reality that

exists beyond the awareness ofthose involved in the story is put there by the author with the object ofcreating dramatic effect. Indeed, it is an invention of modern novelists and playwrights to create a world made up entirely of the psychic realities of the protagonists, leaving knowledge ofthe "real" world in the realmofthe implicit. So writers as different as Joyce and Melville share the characteristic ofnot "disclos- ing" aboriginal realities but leaving them at the horizon ofthe story as matters of supposition-or, as we shall see, ofpresupposition. Science-particularly theoretical physics-also proceeds by con-

structing worlds in a comparable way, by "inventing" the" facts (or world) against which the theory must be tested. But the striking differ- ence is that, from time to time, there are moments oftesting when, for example, light can be shown to be bent or neutrinos must be shown to leave marks in a cloud chamber. Itmay indeed be the case, as Quine has urged, that physics is 99 percent speculation and 1 percent observa- tion. But the world making involved in its speculations is ofa different order from what story making does. Physics must eventuate in predict- ing something that is testably right, however much it may speculate. Stories have no such need for testability. Believability in a story is ofa different order than the believability of even the speculative parts of physical theory. Ifwe apply Popper's criterionoffalsifiability to a story as a test of its goodness, we are guilty ofmisplaced verification.

Two Modes ofThought · 15

Having said that much about how the two modes can be distinguished one from the other, let me now focus almost entirely on the less understood ofthe pair: on narrative. And as I remarked in the preced- ing chapter, I shall want to concentrate on narrative, so to speak, at its far reach: as an art form. William James comments in his Gifford Lectures, The Varieties ofReligiousExperience, that to study religion one should study the most religious man at his most religious moment. I shall try to follow his advice with respect to narrative but, perhaps, with a Platonic twist. The great works offiction that transform narra- tive into an art form come closest to revealing "purely" the deep struc- ture ofthe narrative mode in expression. The same claim can be made for science and mathematics: they reveal most plainly (and purely) the deep structure ofparadigmatic thought. And perhaps James intended his dictum in the same sense, in spite ofhis anti-Platonism. There is another reason, aside from the Platonic, for pursuing this

course. If one takes the view (as I shall in Chapter 5) that human mental activity depends for its full expression upon being linked to a cultural tool kit-a set ofprosthetic devices, so to speak-then we are well advised when studying mental activity to take into account the tools employed in that activity. As primatologists tell us, this amplifica- tion by cultural tools is the hallmark ofhuman skills, and we overlook it in our research with peril. And so, ifone wishes to study the psychol- ogy ofmathematics (as, say, G. Polya did), one studies the works of trained and gifted mathematicians, with particular emphasis on the heuristics and the formalisms they use to give form to their mathemat- ical intuitions. By the same token, one does well to study the work of trained and

gifted writers ifone is to understand what it is that makes good stories powerful or compelling. Anybody (at almost any age) can tell astory- and it is altogether good that story grammarians, so called, are study- ing the minimal structure needed to create a story. And anybody (again, at almost any age) can "do" some mathematics. But great fiction, like great mathematics, requires the transformation of intu- itions into expressions in a symbolic system-natural language or some artificialized form ofit. The forms ofexpression that emerge, the

16 · Two Natural Kinds

discourse that carries the story, or the calculus that depicts amathemat- ical relation-these are crucial for understanding the differences be- tween an inchoate account of a bad marriage and Madame Bovary, between a clumsily argued justification and an elegant and powerful derivation ofa logical proof. I think I have said all that needs saying on this point, a point addressed more to psychologists than to literary theorists. The former, perhaps, will quarrel with the pointoutofdefer- ence to the reductionismofscience. The latterwill almost certainly find the point almost bizarrely obvious.

Narrative deals with the vicissitudes of human intentions. And since there are myriad intentions and endless ways for them to run into trouble-or so it would seem-there should be endless kinds of stories. But, surprisingly, this seems not to be the case. One view has it that lifelike narratives start with a canonical or "legitimate" steady state, which is breached, resulting in a crisis, which is terminated by a redress, with recurrence of the cycle an open possibility. Literary theorists as various as Victor Turner (an anthropologist), Tzvetan Todorov, Hayden White (an historian), and Vladimir Propp (a folk- lorist) suggest that there is some such constraining deep structure to narrative, and that good stories are well-formed particular realizations of it. Not all literary scholars take this view-Barbara Herrnstein- Smith being a notable dissenting voice. Ifitwere the case that there are limits on the kinds ofstories, it could

mean either that the limits are inherent in the minds ofwriters and/or readers (what one is able to tell or to understand), or that the limits are amatter ofconvention. Ifit were the former, ifthe limits on storywere innate, then it would be difficult to explain the eruptions ofinnovation that illuminate the course of literary history. And if it were the latter, the heavy hand ofconvention, that limited the nature ofstory, then it would be just as difficult to explain why there is so much recognizable similarity in tales from all lands, and so much historical continuity within any particular language whose literatures have gone through changes as dramatic as, say, the French or English or Russian. The arguments pro and con are, somehow, more interesting than

conclusive. Their conclusiveness is flawed not only by literary innova- tion but, I suspect, by the impossibility of deciding whether, say,

Two Modes ofThought · 17

Joyce's Ulysses or Beckett's Molloy trilogy fits a particular formula or not. Aside from all that, what level ofinterpretation ofa story shall we take to represent its "deep structure"-litera, moralis, allegoria, or anagogia? And whose interpretation: Jung's, Foucault's, Northrop Frye's? And when, as with antinovel novels, awriter (like Calvino, say) exploits his reader's story expectations by flouting them artfully, does that count as violating or conforming to the canonical form? And as ifthis were not enough, there is the question ofthe discourse

into which the story is woven and the two aspects ofstory (to which we have already alluded): thefabula and the sjuzet) the timeless and the sequenced. Which is constrained, and in whatways? That there may be astructure to time-worn folktales or to myths, amatter to which I shall revert later, nobody will deny. But do these narratives provide a uni- versal structure for all fictions? For Alain Robbe-Grillet or, to take an instance where it is even difficult to decide whether the book is a novel or an exercise in criticism, for Julian Barnes's Flaubert)s Parrot? I think we would do well with as loose fitting a constraint as we can

manage concerning what a story must "be" to be a story. And the one that strikes me as most serviceable is the one with which we began: narrative deals with the vicissitudes of intention. I propose this not only because it leaves the theorist with a certain

flexibility but because it has a "primitiveness" that is appealing. By primitive I mean simply that one can make a strong argument for the irreducible nature of the concept of intention (much as Kant did for the concept ofcausation). That is to say, intention is immediately and intuitively recognizable: it seems to require for its recognition no com- plex or sophisticated interpretive act on the part of the beholder. The evidence for such a claim is compelling. There is a celebrated monograph, little known outside academic

psychology, written ageneration ago by the Belgian studentofpercep- tion, BaronMichotte. By cinematic means, he demonstrated that when objects move with respect to one another within highly limited con- straints, we see causality. An object moves toward another, makes con- tact with it, and the second object is seen to move in a compatible direction: we see one object "launching" another. Time-space relations can variously be arranged so that one object can be seen as "dragging" another, or "deflecting" it, and so on. These are "primitive" percep- tions, and they are quite irresistible: we see cause.

18 · Two Natural [(intis

To answer Hume's objection that such causal experiences derive from association, Alan Leslie repeated the Michotte demonstrations with six-month-old babies. His procedure Ineasured signs of surprise in the infant, which expresses itself in a variety of registerable ways from facial expression to changes in heart rate and blood pressure. Leslie showed the infants a sequence ofcinematic presentations that in their space-time arrangement were seen by adults as caused. He would then intersperse one noncausal presentation that was outside the pre- scribed Michotte space-time limits-and the babywould show starded surprise. The same effect could be achieved by following a noncausal sequence of presentations with a causal one. In each case, Leslie ar- gued, there was some qualitative change in the experience ofthe infant that led to "dishabituation" and surprise. Note that a change in space- time arrangement of the displays that was as large as the one used to shift categoryproducedno effect ifitwas within the categoryofcausal- ity. Michotte's work and Leslie's follow-up provide powerful argu- ments for the irreducibility of causality as a "mental category" in the Kantian sense. Can intentionality as a concept be shown to be as primitive? Fritz

Heider and Marianne Simmel have also used a"bare" animated film to demonstrate the irresistibility of"perceived intention" in the form ofa scenario involving a small moving triangle, a small moving circle, a large moving square, and a box-like empty rectangle-whose move- ments are irresistibly seen as two lovers being pursued by a large bully who, upon being thwarted, breaks up the house in which he has tried to find them. Judith Ann Stewart, more recently, has shown that it is possible to arrange the space-time relationship of simple figures to produce apparent intention or "animacy." We plainly see "search," "goal seeking," "persistence in overcoming obstacles"-see them as intention-driven. Interestingly, from the point of view of Propp's pioneering work on the structure offolktales (to which we shall come presendy), the perception of animacy is induced by varying direction and speed ofmotion of an object with respect to an obstacle. Unfortunately, we do not yet have the analogue experiment on ap-

parent intention for Leslie's baby experiments on apparent causality. It will come soon enough. If it should yield positive results, then we would haveto conclude that "intention and its vicissitudes" constitute a primitive category system in terms ofwhich experience is organized,

Two Modes ofThought · 19

at least as primitive as the category system ofcausality. I say "at least," for the fact remains that the evidence of children's animism suggests that their more primitive category is intention-physically caused events being seen as psychically intended, as in the early experiments that earned Piaget his first worldwide acclaim.

But such experiments, while they tell us about the primitiveness ofthe idea of intention, tell us nothing about the discourse that converts an unworded narrative into powerful and haunting stories. What is it in the telling or writingofa tale that produces Jakobson's literaturnost? In the telling there must be "triggers" that release responses in the reader's mind, that transform a banal fabula into amasterpiece ofliterary narra- tive. Obviously, the language of the discourse is critical, but even before that there is plot, plot and its structure. Whatever the medium-whether words, cinema, abstract animation, theater-one can always distinguish between the fabula or basic story stuff, the events to be related in the narrative, and the "plot" or sjuzet, the story as told by linking the events together. The plot is how and in what order the reader becomes aware of what happened. And the "same" story can be told in different sequence. This means, of course, that there must be transformations of some kind that permit a common base structure of story to be handled in different meaning-preserving sequences. What can we say about the deep structure ofstories-the story stuff,

or fabula, that lends itself to different orders ofpresentation? Could it be the kind of structure that I examined a moment ago and earlier attributed to Victor Turner, Hayden White, Vladimir Propp, and Tzvetan Todorov? That is to say, one "primitive" fabula involves the breach of a legitimate state of affairs, the break then creating a crisis that is nipped in the bud or that persists until there is redress? If there were a corresponding structure in the minds of readers, cinema view- ers, and playgoers, then such a fabula could be plotted in linear order, in flashbacks, or even in medias res) starting virtually anywhere (as Robbe-Grillet succeeds in doing for film and novel, and as, say, Michel Leiris does in his "experimental" antinarrative autobiography)?We do not have to take a stand on how many such fabula there are (as many, for example, as Jung's archetypes?), only that they have some sort of

20 · Two Natural Kinds

being in the beholder's mind that permits him to recognize them in whatever expression encountered. But there is something more to it than that. Kenneth Burke argues

that "story stuff" involves characters in action with intentions orgoals in settings using particular means. Drama is generated, he claims, when there is an imbalance in the "ratio" ofthese constituents. That is to say, a character (say Nora inA DolPs House) is in an inappropriate setting, or an action does not warrant the goal to which it is leading a character. Yet, neither breach, crisis, and redress, nor imbalances in a Burkeian

pentad, are sufficient descriptions of "story stuff." For there are ele- ments of story that rest not upon action and interaction but upon character as such. Conrad's novels provide a good example. Jim's in- scrutability (even to the narrator who "tells" his story) is central to the drama ofLordJim. In The Secret Sharer, the young captain's fascinated obsession with Leggatt drives the story. Some readers actually propose that Leggatt is an imaginary Doppelganger who exists only in the cap- tain's mind. Perhaps, as with Aristotle's recipe for tragedy in the Poet- ics, drama is a working out ofcharacter in action in a plot constrained by a setting. Yet this too cannot be a full account ifwe heed Propp's argument

that, in the folktale, character is afunction ofa highly constrained plot, the chief role ofa character being to play out a plot role as hero, false hero, helper, villain, and so on. Forwhile it may be the case that in the time-smoothed folktale story-stuffdetermines character (and therefore character cannot be central), it is equally true that in the "modern" novel plot is derived from the working out ofcharacter in a particular setting (one of the earliest theorists of modernism, therefore, being Aristotle on tragedy!). Greimas's view is that a primitive or irreducible feature of story

(whatever else it may include) is that it occurs jointly on the plane of action and in the subjectivity of the protagonists. And perhaps this is why deceit, guile, and misunderstanding are to be found so often in myths and folktales from "Little Red Riding Hood" to "Perseus and the Gorgon" and, at the same time, lie at the heart ofso many modern novels and plays. Psychologically, the "dual landscape" view is appealing in suggesting

how the reader is helped to enter the life and mind ofthe protagonists:

Two Modes ofThought · 21

their consciousnesses are the magnets for empathy. The matching of "inner" vision and "outer" reality is, moreover, a classic human plight. It grips the child hearing how the Big Bad Wolf tries to deceive and then is unmasked by Red Riding Hood, or the adult reading Joyce's "Araby," suffering the humiliation of the young boy when his dreams of a gift for the neighbor girl fade in the tawdry atmosphere of the fairground closing. In any case, the fabula of story-its timeless underlying theme-

seems to be a unity that incorporates at least three constituents. It contains a plight into which characters have fallen as a result of inten- tions that have gone awry either because of circumstances, of the "character ofcharacters," or most likely ofthe interaction between the two. And it requires an uneven distribution ofunderlying conscious- ness among the characters with respect to the plight. What gives the story its unity is the manner in which plight, characters, and conscious- ness interact to yield a structure that has a start, a development, and a "senseofan ending." Whether it is sufficient to characterize this unified structure as steady state) breach) crisis) redress is difficult to know. It is certainly not necessary to do so, for what one seeks in story structure is precisely how plight, character, and consciousness are integrated. Bet- ter to leave the issue open and to approach the matter with an open mind.

Language, to whatever use it may be put, has the design feature of being organized on different levels, each level providing constituents for the level above which dominates it. As Jakobson noted in his classic analysis of the sound system of speech, the distinctive features of speech sound are determined by the phonemes that they constitute at the next level up; phonemes are combined according to rules at the next level up, the morpheme, and so on. So too at the levels above sound, for morphemes, lexemes, sentences,

speech acts, and discourse. Each level has its form of order, but that order is controlled and modified by the level above it. Since each level is dominated by the level above it, efforts to understand anylevel on its own have inevitably led to failure. The structure of language is such that it permits us to go from speech sounds through the intermediate levels to the intentions ofspeech acts and discourse. The path by which

22 · Two Natural [(intis

we travel that route varies with our objective, and storytelling is a special objective. In putting any particular expression together, one selects words and

one combines them. How one selects and combines will depend on the uses to which one wishes to put an utterance. Jakobson calls these two primitive language-forming acts, selecting and combining, the vertical and the horizontal axes of language. The vertical axis of selection is dominated by the requirement ofpreserving or modifying meaning by substituting appropriate words or expressions for one another: boy) immature male) lad) and so on. But the rule of substitution goes be- yond synonymy to metaphor. What ofcolt) lamb) fawn? Do they fit boy? We say it depends on context and objective. And \vhat of larger-order substitutions? Which does better for New York: "the biggest city in North America" or "the harbor at the mouth of the Hudson"? Again, it depends. And what of substituting for depression: black mood or "ragged claws scuttling across the floors ofsilent seas"? There is forever amatter ofchoice about the vertical axis: whether to preserve reference as literally as possible, whether to create an atmospheric change by metaphor, whether (as Jakobson and the Prague School urged upon poets) to "make it strange" so as to overcome automatic reading. It is probably the case that scientific or logical writing-or, rather,

writing governed by requirements of a scientific argument-tends to choose words with the object of assuring clear and definite reference and literal sense. It is required by the felicity conditions ofspeech acts of this kind. Litera dominates over moralis and the others. In the telling of a story, one has the selection restriction of representing a referent in the eye of a protagonist-beholder, with a perspective that fits the subjective landscape on which the story is being unfolded, and yet with due regard for the action that is going on. So from the start, the selection ofexpressions must meet the special requirement of that special form of speech act that is a story-of which more presently, when I consider a crucial idea proposed by Wolfgang Iser. The second axis, the horizontal axis of combination, is inherent in

the generative power ofsyntax to combine words and phrases. Its most elementary expression is predication or, even more primitively, the juxtaposition of a comment on a topic, when the topic is "given" or taken for granted and the comment is something new added to it. I see a new species of bird and say to my partner: "Some bird. Fantastic." The first element is the topic; the second the comment. Predication is a

Two Modes ofThought · 23

more evolved form ofmaking comments on topics that permits us to assign a "truth function" to the expression, as in such ordinary sen- tences as The boy has a ball. The boy has a secret. The boy has a burning ambition. The boy has a bee in his bonnet.

The boy is the given; the predicate is new. The sentence can now be translated into a formal or logical proposition and tested for its truth value in the context in which the utterance was made. To the degree that a subject and predicate are "transparent," they can

easily be converted into verifiable propositional form; indeed, one common theory ofmeaning, the verificationist theory, equates mean- ing with the set of verifiable propositions a predicational statement generates. But there are statements or utterances that combine given and new in a manner that is "strange" or that, in Henry James's sense, contains gaps, or where there is a difficult distance between the two. A good case in point is Eliot's lines

I should have been a pair of ragged claws Scuttling across the floors of silent seas.

To render these lines literally as "I am depressed with aging" (taking into account the context ofthe whole of"Prufrock," from which they are extracted) fails to capture the horizontal given-new combination of the poem. Yet, on one interpretation, that may be what they mean- noting that in the vertical axis we have translated "ragged claws ..." into "depression over aging." To be sure, as Jakobson also insisted, meaning always involves translation. But there is some sense in which neither the literal translation ofthe new term nor the resulting combi- nation ofit with the given term succeeds as a poetic translation. And if we take predicate-like utterances in which both the subject and the predicate are nonliteral, the failure is even more evident, as in these lines from MacNeice:

The sunlight on the garden Hardens and grows cold. We cannot cage the minute Within its nets ofgold; When all is told We cannot beg for pardon.

24 · Two Natural I(inds

It is not only "unclear" how to manage the vertical axis-to what does "sunlight on the garden" refer, and "harden" in this context? "Cage"? And then, "cage the minute," etcetera. The language of poetry, or perhaps I should say the language of

evocation, substitutes metaphors for both given and new, leaving it somewhat ambiguous what they are substitutes for. When the terms are combined, the resulting given-new combination is no longer amen- able to being converted into ordinary truth functional propositions. Indeed, at crucial moments it even departs from the "contract" that specifies a clear distinction between given and new in predicative combinations. So neither vertically nor horizontally does the evocative language of

poetry and story conform to the requirements ofplain reference or of verifiable predication. Stories of literary merit, to be sure, are about events in a "real" world, but they render that world newly strange, rescue it from obviousness, fill it with gaps that call upon the reader, in Barthes's sense, to become a writer, a composer of a virtual text in response to the actual. In the end, it is the reader who must write for himselfwhat he intends to do with the actual text. How, for example, to read these lines from Yeats:

The brawling of a sparrow in the eaves, The brilliant moon and all the milky sky, And all that famous harmony of leaves, Had blotted out man's image and his cry.

Which brings us directly to Wolfgang Iser's reflections in TheAct of Reading on what manner ofspeech act is a narrative. I want to touch on only one part ofhis argument, one that is central to my own. With respect to narrative, he says, "the reader receives it by composing it." The text itselfhas structures that are "two-sided": a verbal aspect that guides reaction and prevents it from being arbitrary, and an affective aspect that is triggered or "prestructured by the language ofthe text." But the prestructure is underdetermined: fictional texts are inherently "indeterminate." fictional texts constitute their own objects and do not copy something already in existence. For this reason they cannot have the full determinacy ofreal objects, and indeed, it is the element ofindeterminacy that evokes the text to "communicate" with the reader, in the sense that they induce him to participate both in the production and the comprehension ofthis work's intention.

Two Modes ofThought · 25

It is this "relative indeterminacy ofa text" that "allows a spectrum of actualizations." And so, "literary texts initiate 'performances' ofmean- ing rather than actually formulating meanings themselves." And that is what is at the core ofliterary narrative as a speech act: an

utterance or a text whose intention is to initiate and guide a search for meanings among a spectrum of possible meanings. Storytelling, be- sides, is a speech act whose felicity conditions are unique. The speech act is initiated by giving some indication to a listener or reader, first, that a story is to be recounted; second, that it is true or fictional; and third (optionally), that it fits some genre-a sad story, a moral fable, a comeuppance tale, a particular scandal, a happening in one's life. Be- yond that, there is aconditionofstyle: that the form ofthe discourse in which the story is actualized leaves open the "performanceofmeaning" in Iser's sense. It is this last condition that brings us directly to the discourse properties of stories, to which I turn now.

Discourse, if Iser is right about narrative speech acts, must depend upon forms of discourse that recruit the reader's imagination-that enlist him in the "performance ofmeaning under the guidance of the text." Discourse must make it possible for the reader to "write" his own virtual text. And there are three features ofdiscourse that seem to me to be crucial in this enlistment process. The first is the triggering ofpresupposition) the creation of implicit

rather than explicit meanings. For with explicitness, the reader's de- grees of interpretive freedom are annulled. Examples abound, but Primo Levi's recent The Periodic Table provides a particularly striking case. His subtle setting forth ofthe properties ofaparticular element in each "story"-argon, hydrogen, zinc, and so on-provide a presup- positional background in terms of which the stories may be "inter- preted." How the presuppositional background triggers interpretation is a matter I shall come to shortly. The second is what I shall call subjectiJication: the depiction ofreality

not through an omniscient eye that views a timeless reality, but through the filter of the consciousness of protagonists in the story. Joyce, in the stories ofDubliners, rarely even hints at how the world really is.We see only the realities ofthe characters themselves-leaving us like the prisoners in Plato's cave, viewing only the shadows ofevents we can never know directly.

26 · Two Natural ](inds

The third is multiple perspective: beholding the world not univocally but simultaneously through a set ofprisms each ofwhich catches some part of it. Auden's poem on the death ofYeats is a brilliant example: the poet's death is seen in the instruments ofwinter airports, on the floor ofthe Bourse, in the sickroom, in the "guts ofthe living." Roland Barthes argues in S/Z that without multiple codes ofmeaning astory is merely "readerly," not "writerly." There are doubtless other means by which discourse keeps meaning

open or "performable" by the reader-metaphor among them. But the three mentioned suffice for illustration. Together they succeed in sub- junctivizing reality) which is my way ofrendering what Iser means by a narrative speech act. I take my meaning of "subjunctive" from the second one offered by the OED: "Designating a mood (L. modus subjunctivus) the forms ofwhich are employed to denote an action or state as conceived (and not as a fact) and therefore used to express a wish, command, exhortation, or a contingent, hypothetical, or pro- spective event." To be in the subjunctive mode is, then, to be traf- ficking in human possibilities rather than in settled certainties. An "achieved" or "uptaken" narrative speech act, then, produces a sub- junctive world. When I use the termsubjunctivize) I shall mean it in this sense. What then can we say in any technical way about the means whereby discourse portrays a "subjunctive reality"? For surely that is the key to the issueofdiscourse in great fiction. Let me turn to some of the more systematic ways in which this is accomplished. Begin with the familiar case ofspeech acts and Paul Grice's extension

of the idea to what he calls the Cooperative Principle governing ordi- nary conversation. He proposes maxims of quantity (saying only as much as is necessary), ofquality (saying only the truth, and saying it with perspicuousness), and of relevance (saying only what is to the point). However needed such maxims may be for regulating conversa- tional cooperation, in fact they are guides to banality: to be brief, perspicuous, truthful, and relevant is to be drab and literal. But the existence of such maxims (however implicit our awareness of them), Grice argues, provides us with the means of violating them for pur- poses of meaning more than we say or for meaning something other than what we say (as in irony, for example) or for meaning less than we say. To mean in this way, by the use of such intended violations or "conversational implicatures," is to create gaps and to recruit presup- positions to fill them. As in

Two Modes ofThought · 27

Where's Jack? Well, I saw a yellow VW outside Susan's.

The reader-hearer, if he is to stay on the narrative scene, must fill in, and under the circumstances he is made complicitous with the charac- ters in the exchange. Why doesn't the respondent say outright (per- spicuously) that Jack is visiting Susan? Is it an illicit visit? Is Jack "going the rounds"? Cookbooks on story writing urge the use of implicatures to increase "narrative tension," and they can easily lose their effect when overused. Yet they provide the means for the kind of indirect talk that forces "meaning performance" upon the reader. Presupposition is an ancient and complex topic in logic and linguis-

tics, and one that deserves closer study by the student ofnarrative. A presupposition, formally defined, is an implied proposition whose force remains invariant whether the explicit proposition in which it is embedded is true or false. Their nature and operations have been set forth brilliantly by Stephen Levinson, by L. Karttunen and Richard Peters, and by Gerald Gazdar, and their discussions of presupposi- tional triggers, filters, plugs, and holes are richly suggestive for literary text analysis. They deal with what are called "heritage expressions" and with how a presupposition is built up over discourse in order to pro- ject itself into later statements. Triggers effect such projection. Four simple examples will serve to illustrate their manner ofoperating.

Trigger

Definite descriptions: John saw/didn't see the chimera. Factive verbs: John realized/didn't realize he was broke.

Implicative verbs: John managed/didn't manage to open the door.

Iteratives: You can't get buggy whips any- more.

Presupposition

There exists a chimera.

John was broke.

John tried to open the door.

You used to be able to get buggy whips.

There are many other triggers. I think it is plain (though the details are not easy) that triggering presuppositions, like intentionally violating

28 · Two Natural Kinds

conversational maxims, provides a powerful way of "meaning more than you are saying," or going beyond surface text, or packing the text with meaning for narrative purposes. The use ofpresupposition is greatly facilitated by an informal "con-

tract" that governs language exchanges. As Dan Sperber and Deirdre Wilson have noted, we characteristically assume that what somebody says must make sense, and we will, when in doubt about what sense it makes, search for or invent an interpretation ofthe utterance to give it sense. Example on a London street (after Sperber and Wilson):

Will you buy a raffle ticket for the Royal Naval Lifeboat Insti- tution?

No thanks, I spend summers near Manchester. Ah yes, of course.

Obviously, you cannot press a reader (or a listener) to make endless interpretations ofyour obscure remarks. But you can go a surprisingly long way-provided only that you start with something approximat- ing what Joseph Campbell called a"mythologically instructed commu- nity." And, in fact, most of the devices and tropes that we use in the telling and writing ofstories are not substantively as demanding as the one in Sperber and Wilson's example. To revert to the beginning discussion ofparadigmatic and narrative

modes ofthought, both ofthem surely trade on presupposition, ifonly for the sake ofbrevity. If the scientist or analytic philosopher or logi- cian should be found to be triggering presuppositions in a covert way, he will become the butt of jokes about making a hard sell rather than letting things speak for themselves. His presuppositions should be unpackable, easily so. The writer of fiction who does not use such triggering will simply fail. His story will be "flat." What ofsubjectification, the rendering ofthe world ofthe story into

the consciousness ofits protagonists? Freud remarks in "The Poet and the Daydream" that the act of composition is, after all, an act of de- composition: the artist's separation ofhis own internal cast ofcharac- ters into the characters of the story or play. The plot then becomes a hypothetical actualization ofthe reader's own internal "psychodynam- ics." Freud the psychologist thought, ofcourse, that this was achieved unconsciously, and Milosz the poet agrees:

Two Modes ofThought · 29

In the very essence of poetry there is something indecent: A thing is brought forth that we didn't know we had in us, So we blink our eyes, as if a tiger had sprung out And stood in the light, lashing his tail.

Freud had it in mind that the "internal drama made external" aids the reader to identify not only with characters but with the human plights in which they find themselves. But this kind of theorizing does not help us much in our understanding of discourse. Is there something more precise that can be said about the language by which subjective landscapes and multiple perspectives are evoked in stories? For that is the issue I am addressing-how is reality rendered subjunctive by language? An idea ofTodorov's serves well as a point ofdeparture. The argu-

ment runs somewhat as follows-I say "somewhat" because I am add- ing some elaborations that are not part of his analysis. Suppose one posits first a "way ofsaying" that is as simple, expository, and nonsub- junctive as possible: x commits a crime. In effect it depicts a "product" or event. It asserts. Todorov proposes that there are six simple transfor- mations that transform the action ofthe verb from being a fait accom- pli to being psychologically in process, and as such contingent or subjunctive in our sense. His six simple transformations are as follows: Mode. Modality, literally a modal auxiliary for the verb, subjectifies

the action: must) might) could) would) and so on. Modals are ordinarily classified as epistemic and deontic, the first having to do with matters ofwhat could or must be, the second with value obligations: x must commit a crime and x should commit a crime. And within each class there is a further subdivision between necessity and contingency: x must commit a crime and x might commit a crime) for example, both ofwhich are "perspectival" triggers. Modal transformations also have the effect of implying a context for an act: x must or x should for some reason, implied but not stated, do what the verb requires. Intention. Here, the act is directly embedded in its intention: x plans

to commit a crime (or hopes to) intends to) and so on). Result is a transformation-as in x succeeds in committing a crime-

whose effect is both to presuppose intent and to raise but leave open the question ofhow it all came about. Manner-as in x is keen to commit a crime-subjectifies the act and

creates an attitude that modifies the action's intention.

30 · Two Natural [(inds

Aspect refers to a form of time marking that is related not to an abstract time marker like tense but to the progress ofthe task in which the action is occurring: for example, x is beginning to commit a crime (is in the midst oj; and so on). Paul Ricoeur's Time and Narrative contains an interesting discussion of the way the abstract emptiness of time, defined by tense, must be embodied in a concrete and progressing activity in order for it to constitute narrative time. Aspect transforma- tions are probably the most direct way of providing or evoking such concreteness. Status-as in x is not committing a crime-is a transformation that

opens the possibility that there was a wish to, a set of circumstances that, a possibility that, an accusation that could have led to a crime. Negation is a powerful trigger ofpresuppositions about the possible. "I do not commit crimes" opens a world ofalternative perspectives. Todorov also proposes a half-dozen complex transformations that,

in effect, alter a sentence by adding to it averb phrase that modifies the original or main verb phrase. All ofhis complex verb phrases have the function ofadding "factivity" to the original-that is, a state ofmental activity to accompany the main verb phrase. They place the activity in a landscape ofconsciousness. They are:

Appearance: Knowledge: Supposition: Description: Subjectification: Attitude:

x pretends he has committed a crime x learns y has committed . . . x foresees he will commit . . . x reports he has committed . . . x thinks he has committed . . . x enjoys committing . . .

To put it in Todorov's words, such a transformation, simple or complex, "permits discourse to acquire a meaning without this mean- ing becoming pure information." I assume that "pure information" means for him a form of exposition that minimizes presupposition, that keeps the reader from going too far beyond the information given. The use ofsuch transformations, on the other hand, should thicken the connective web that holds a narrative together in its depiction ofboth action and consciousness. Can Todorov's system oftransformations distinguish good narrative

from, say, good exposition? Our research group tried it, comparing one of the stories in Joyce's Dubliners with a piece of fine expository

Two Modes ofThought · 31

writing by the anthropologist Martha Weigel. "Clay" was our story- one on which we had been working intensively. It is a story laced with ritual-Maria laying out the barmbrack for the other girls at the laun- dry, her tram ride from Ballsbridge to the Pillar and then to Drum- condra, the All Hallows' Eve party and its ritual game ofblind man's bluff. This inspired the choice, for comparison, ofan expository text to which the same analysis could be applied and that dealt with ritual action. Martha Weigel is an anthropologist and a writer of consider- able grace. Her subject is the Southwest and her specialty is the Penitentes, about whom she has written an acclaimed book, Brothers of Light) Brothers ofBlood. It contains a chapter on rituals. That chapter was our choice. Gwyneth Lewis and I set out to compare Joyce's "Clay" (113 sen-

tences long) with Weigel's chapter on Penitente rituals-at least its first 113 sentences. The results ofthe trial run, though they may not be typical ofanything save these two pieces, were so striking that I may be forgiven for reporting them here. Consider, for example, the number ofTodorovian transformations per 100 sentences of text in the Joyce story and in Weigel's exposition:

Todorovian Joyce's Weigel's transformation "Clay" "Rituals"

Simple 117.5 34.6 Complex 84.9 16.0 Total 202.4 50.6

Or, in barest summary, the story contains on average two transforma- tions per sentence; the anthropological account, only one every other sentence. This, admittedly, is the most grossly unadorned word counting-

however much it may be inspired by an hypothesis about how sub- junctivizing is achieved. It tells nothing about the contexts in which these transformations are used or about the uses to which they are put. Why do one in three of Joyce's sentences contain transformations of manner, while only one in ten ofMartha Weigel's do? Or why are a quarter ofJoyce's constructions timed by aspect, while only one in fifty ofWeigel's are? A more subtle analysis is for the future. Rather, I want to say something about "reader response" to the

32 · Two Natural I(inds

Joyce story. In our research, we ask our readers to tell us back the story in their own words: to create, so to speak, a virtual text. Again, I can make no claim for the representativeness ofwhat we are finding, but we did subject to analysis the "told back" version ofone ofour readers, an experienced reader of fiction in his late teens who was reading the story for the first time. He told it back to us a day later. His version of "Clay" was only 24 sentences long (typically shorter than the story), in contrast to Joyce's 113. Compare Joyce and the reader, the numbers standing again for frequency of transformations per 100 sentences.

Todorovian Joyce's Reader's transformation "Clay" virtual text

Simple 117.5 235.3 Complex 84.9 91.1 Total 202.4 326.4

Is the reader picking up the subjunctivized speech of the story? Well, there are twice as many simple transformations in the reader's "story" as in Joyce's, and at least as many complex ones as Joyce used. Our reader is plainly resonating to the story and to its discourse as well. Indeed, the two texts, actual and virtual, even agree closely in terms of the frequency ranking of the transformations used. The simple trans- formations first:

Todorovian Joyce's Reader's transformation "Clay" Rank virtual text Rank

Manner 33.6 1 83.0 1 Aspect 24.7 2 38.0 3.5 Status 23.8 3 50.0 2 Mode 18.6 4 38.0 3.5 Result 10.6 5 25.0 5 Intention 6.2 6 1.3 6

And the match in the complex transformation was just as close:

Todorovian Joyce's Reader's transformation "Clay" Rank virtual text Rank

Description 41.6 1 46.0 1 Subjectification 17.7 2 13.8 2

Two Modes ofThought · 33

Todorovian Joyce's Reader's transformation "Clay" Rank virtual text Rank

Attitude 11.5 3 8.0 4.5 Knowledge 9.7 4 11.3 3 Appearance 2.6 5 8.0 4.5 Supposition 1.8 6 4.0 6

What is vividly interesting is that our young reader provided us with a virtual text that, I think, Joyce would not have minded. (It is to be found in the Appendix, placed side-by-side with Joyce's story.) One does not want to make too much of this particular concordance be- tween the "retell" discourse ofa reader and the text ofa story. But the "results" ofthis first experiment do suggest some hypotheses. The first is that the "mood"-the modus subjunctivus-of the story is preserved in the reading, as well as the substance of the story itself, in the sense both of fabula and sjuzet. There are transformations, to be sure, and (as one can see by comparing the reader's story with Joyce's in the Appendix) these are principally in the form of deletions. Doubtless these deletions serve to "sharpen, level, and assimilate" elements ofthe story (to use Sir Frederic Bartlett's terms from his classic, Remember- ing). In the retelling, turn-of-the-century Dublin seems a bit more like the New York of today; the episode with the military-looking gentle- man on the tram is forefronted in the virtual text more than in the actual one; the doings in the laundry are somewhat flattened. But perhaps the most interesting qualitative transformation in the

retelling is the reader's management of subjunctivity. At first, he tells the story in a way suggesting omniscience about what was happening. This is then modified by peppering the account with "he says" and "he said," where "he" is the author. Then subjunctivizing language begins to take over the virtual text. The reader now says ofMaria that "she is going to x," "she wants to x," "she remembers when x," "she thinks what else she wants to x," "she's forced to x," "she becomes used to (accustomed to) x." Or "they start being merry," or "Maria said to Joe that he should make up with his brother Aify," or "Maria says she's sorry." Or, to take a striking instance ofmood preservation, "and Joe says, you know, since its such a nice night I won't get angry about it, but you know, he doesn't; he's not really happy that she brought it up." The subjective landscape is richly constructed in the virtual text,

34 • Two Natural Kinds

though untransformed "matters of fact" are interspersed ("she goes there and gives the kids their little penny cakes") but only enough to keep a line of action going concurrently with the subjective line. We also asked our reader a good many questions after he had told

back the story so that we might dig a little more deeply into his interpretive activity. For the analysis ofvirtual text (the "retelling") is only one way offinding out what a story like "Clay" means to a reader. Asked about what had particularly struck him in the story, he picks up the witch theme: "her nose nearly touched her chin." He wonders whether her witchlike appearance clashes with the almost saintly qual- ity she is pictured as possessing. Then he asks, does she think she is saintly while others are really sorry for her. His search for a timeless fabula has begun: "1 did kind ofget like

some kind ofevil coming from her . . . even though she was so nice to everybody that she had some hidden evil building up in her, or some- thing." And then, "like artificially nice, almost, like she had no real enemies, she had, you know, she was just nice to everybody, and she wanted everybody, you know, to be nice to her and respect her, which is what she got. But there was that, that, it is not possible for a human to be like that. You know, except, you know, we only saw part ofher; we don't know what the other part ofher is like." And later he adds, "1 was almost happy that he (the old man on the bus), that he had stolen her plum cake, because it's almost like never ... she was so naive that she'd never experienced anything like that, and 1 was happy that she had at least had some negative experience, 'cause not everything was always just, you know, hunky-dory and everything. Bad things do happen, when you're so trusting ofeverybody." From this interpretation, he then raises a series ofquestions about

symbolism, such as why were they "celebrating Halloween in that ritual, Christmas-like way?" Is it a story about the fall ofinnocence? He finally decides that it is. Iser remarks in The Act ofReading that readers have both a strategy

and a repertoire that they bring to bear on a text. This reader's principal strategy seemed to consist in trying to reconcile the "stuff' ofthe story with his repertoireofconceptions about human plights-his collection ofpossible fabulae. He says early on in so many words that he is "not sure what the story is trying to tell us" but admits that he is caught up in it. His interpretation of"Clay" as a story about "the cost of inno-

Two Modes ofThought · 35

cence protected by self-deception" is, so to speak, his personal thumbprint imposed on the story; but it is not entirely idiosyncratic. To begin with, it is not a culturally atypical interpretation (we know from other readers), particularly for a literate New York boy in his late teens. Nor does it do violence to the text: ifwe had asked other readers to "rate" the cultural appropriateness ofhis interpretation (which we are now doing in our research in progress), it would have been rated well. As for capturing the author's intent, what can one say? If it were possible to call up the shade of Joyce, he would doubtless turn the question into a pun for Finnegan's wake! Obviously, itwill always be amoot question whether and how well a

reader's interpretation "maps" on an actual story, does justice to the writer's intention in telling the story, or conforms to the repertory ofa culture. But in any case, the author's act of creating a narrative of a particular kind and in a particular form is not to evoke a standard reaction but to recruit whatever is most appropriate and emotionally lively in the reader's repertory. So "great" storytelling, inevitably, is about compelling human plights that are "accessible" to readers. But at the same time, the plights must be set forth with sufficient subjunctiv- ity to allow them to be rewritten by the reader, rewritten so as to allow play for the reader's imagination. One cannot hope to "explain" the processes involved in such rewriting in any but an interpretive way, surely no more precisely, say, than an anthropologist "explains" what the Balinese cockfight means to those who bet on it (to take an ex- ample from Clifford Geertz's classic paper on that subject). All that one can hope for is to interpret a reader's interpretation in as detailed and rich a way as psychologically possible. In the end, one is asking how a reader makes a strange text his own.

On this point, there is an instructive exchange between Marco Polo and Kublai Khan in Italo Calvino's Invisible Cities. It begins when Marco says:

"Sire, now I have told you about all the cities I know." "There is still one ofwhich you never speak." Marco Polo bowed his head. "Venice," the Khan said. Marco smiled. "What else do you believe I have been talking to you

about?"

36 · Two Natural [(inds

The emperor did not turn a hair. "And yet I have never heard you mention that name." And Polo said: "Every time I describe a city I am saying something

about Venice." ''When I ask about other cities I want to hear about them. And about

Venice, when I ask you about Venice." "To distinguish the other cities' qualities, I must speak of a first city

that remains implicit. For me it is Venice."

Yet, there is something more than assimilating strange tales into the familiar dramas ofour own lives, even more than transmuting our own dramas in the process. It is not just strange tales and familiar dramas that are implicated, but something at a level of interpretation beyond story. It is that form oftimeless meaning which the story "contains" or instantiates though it is not "in" the story: it is the gist, the plight, perhaps what the Russian Formalists called thefabula. There is another exchange between Marco and Kublai that begins to catch the sense of it, of this meaning beyond the details. Marco describes a bridge stone by stone.

"But which is the stone that supports the bridge?" Kublai Khan asks. "The bridge is not supported by one stone or another," Marco an-

swers, "but by the line of the arch that they form." Kublai Khan remains silent, reflecting. Then he adds: ''Why do you

speak to me of the stones? It is only the arch that matters to me." Polo answers: "Without stones there is no arch."

But still, it is not quite the arch. It is, rather, what arches arefor in all the senses in which an arch is for something-for their beautiful form, for the chasms they safely bridge, for coming out on the other side of crossings, for a chance to see oneself reflected upside down yet right side up. So a reader goes from stones to arches to the significance of arches is some broader reality-goes back and forth between them in attempting finally to construct a sense of the story, its form, its mean- Ing. As our readers read, as they begin to construct a virtual text of their

own, it is as if they were embarking on a journey without maps-and yet, they possess a stock of maps that might give hints, and besides, they know a lot about journeys and about mapmaking. First impres- sions ofthe new terrain are, ofcourse, based on older journeys already taken. In time, the new journey becomes a thing in itself, however

Two Modes ofThought · 37

much its initial shape was borrowed from the past. The virtual text becomes astoryofits own, its very strangeness only acontrastwith the reader's sense ofthe ordinary. The fictional landscape, finally, must be given a "reality" of its own-the ontological step. It is then that the reader asks that crucial interpretive question, "What's it all about?" But what "it" is, ofcourse, is not the actual text-however great its literary power-but the text that the reader has constructed under its sway. And that is why the actual text needs the subjunctivity that makes it possible for a reader to create a world of his own. Like Barthes, I believe that the writer's greatest gift to a reader is to help him become a writer. IfI have, then, made much ofthe contingent and subjunctive not so

much in storytelling as in story comprehending, it is because the narra- tive mode leads to conclusions not about certainties in an aboriginal world, but about the varying perspectives that can be constructed to make experience comprehensible. Beyond Barthes, I believe that the great writer's gift to a reader is to make him a better writer.

Perhaps the greatest feat in the history of narrative art was the leap from the folktale to the psychological novel that places the engine of action in the characters rather than in the plot. What makes "Clay" a powerful story is not events, but Maria. Without her, the paltry events of the story (and even these are seen only through the eyes of the protagonists) would make no sense. As it is, they are vivid little epiphanies of ordinariness-her ordinariness, and through her, our ordinariness. What is at the heart of the psychological story is the notion of a

"character" or a "cast ofcharacters." Our young reader of"Clay" ends with "It's actually a depressing story when you get down to it ... like what's it all about for Maria, like what's it all leading to? She works, she's an old lady ... she's done you know probably nothing." He has converted the story into a tale of character-character and circum- stance. Character is an extraordinarily elusive literary idea. Perhaps it is

elusive for reasons beyond the literary. For even in "real life," it is always a moot question whether the actions of persons should be attributed to circumstances or to their "enduring dispositions"-their

38 · Two Natural Kinds

character. Aristotle in the Poetics conveniently distinguishes between "agent" (pratton) and "character" (ethos), the former being a figure in a drama whose actions merely fit the requirements of the plot, and no more, while the latter has traits beyond those required. But this is by no means clear for, as Ricoeur reminds us in Time and Narrative, Aristotle's idea of mimesis includes the notion that drama reflects "character in action" and action surely involves plot and its setting. Besides, can there ever be a figure in a drama who does just what is required by the plot without giving some inkling of what he or she would be like in more general terms? As Seymour Chatman puts it, "If one trait is assigned to an action, why isn't the floodgate thereby opened?" Ask a reader whether he would be comfortable buying a second-hand car from a "false hero" in a Proppian fairytale, or what kind of relationship that false hero might have had with his father. It will soon be plain, as Solomon Asch demonstrated a generation ago, that character (or perhaps we should call it apparent character) is not a bundle of autonomous traits but an organized conception, however much we may construct it from such scraps and clues as we can find. Asch made his point by demonstrating how differently the trait

intelligent was interpreted depending on whether the character to whom it was attributed was also described as cold or as warm. In the first case, intelligent meant "crafty," while in the second it was taken to mean "wise." Apparent character is perceived as a Gestalt) not as a list oftraits that account for particular actions. And the Gestalt seems to be constructed according to some sort of theory about how people are. For example, they have some sort of core characteristic that directs their behavior from within. But if the person in question behaves in a way that violates that core characteristic, we easily explain it away by invoking circumstances. My colleague Henri Zukier and I tried out some typical college-aged readers on a variant ofthe Asch experiment. To begin with, we gave them a short list of consistent trait names characterizing an imaginary person, like spiritual) introverted) religious) to which they would respond by describing him as "a saintly kind of person." Then we added to the list practical and money-minded. One subject: "Sure. A good man, but he's probably in one of those cut- throat businesses." Another: "I've known them like that-like one of those Amish or Mennonite farmers where I grew up, good in his own group and drives a hard bargain outside." (Interestingly enough, when

Two Modes ofThought · 39

subjects begin telling about "circumstances," their language quickly becomes drenched in Todorov transformations.) The inseparability of character, setting, and action must be deeply

rooted in the nature ofnarrative thought. It is only with difficulty that we can conceive ofeach ofthem in isolation. There are different ways ofcombining the three in constructing the dramatis personae offiction (or of life, for that matter). And those constructions are by no means arbitrary. They reflect psychological processes such as those noted by Asch and other psychologists. They also reflect our beliefs about how people fit into society. The alternate ways in which we can construe people, moreover, often run into conflict with each other, and the conflict leaves us puzzled. Indeed, the act ofconstruing another person is almost inevitably problematic. For all that, the choice of one con- strual rather than another virtually always has real consequences for how we deal with others. Our construal of character, indeed, is our first and perhaps most important step in dealing with another. It is this that makes the very act ofinterpreting a person-whether in fiction or in life-inherently dramatic. It is what makes the narrative ofcharacter so much more subjunctive than the folktale or the myth. How characterize the different ways in which we construe "person-

hood" in literature? We could, of course, adopt the character types offered by theories of"personality" from Galen to Freud andJung, and see whether readers offiction use the same categories. But that is too specialized. We already know that even the most ordinary readers go beyondmere characterdepictions to considerationofcircumstance and setting. We need, rather, a "morphology" of persons that captures common sense, that takes into account the range of concerns I have mentioned. Then we can explore how in fact readers fit character, plot, and action together in making the virtual text. Amelie Rorty offers an analysis that, I think, is to the point. It

distinguishes characters, figures, persons, selves, and individuals. She begins with a sketch: "Charaaers are delineated; their traits are sketched; they are not presumed to be strictly unified. They appear in novels by Dickens, not those by Kafka. Figures appear in cautionary tales, exemplary novels and hagiography. They present narratives of types of lives to be imitated. Selves are possessors of their properties. Individuals are centers of integrity; their rights are inalienable." The use ofthese variant construals is, for Rorty, fraught with human conse-

40 · Two Natural [(inds

quences: "we are different entities as we conceive ourselves enlightened by these various views. Our powers of action are different, our rela- tions to one another, our properties and proprieties, our characteristic successes or defeats, our conception of society's proper strictures and freedoms will vary with our conceptions of ourselves as characters, persons, selves, individuals." Let me very briefly sketch Rotty's views and then return more di-

rectly to the general point. She sees characters as evolved from their origin in the Greek concept of the hero. The hero is known by his deeds. "As the hero's distance from the gods increases, his heroism comes to be exemplified in his character rather than in the sheer glory ofhis action." Characters do not have identity crises, since there is no presupposition about their unity; but disharmony among their charac- teristics breeds trouble-in their action, not in their selfhood. To know what sort of character a person is is to know the circumstances that suit him best, for not all characters are suited to the same life. A character's tragedy is to be in circumstances where his disposition is no longer needed, no longer suited. "Characters in time of great social change ... are likely to be tragic." And then, "In fiction, characters are dear to us because they are predictable, because they entitle us to the superiority of gods who can lovingly foresee and thus more readily forgive what is fixed." Figures "are defined by their place in an unfolding drama; they are

not assigned roles because of their traits, but rather have the traits of their prototypes in myth or sacred script. Figures are characters writ large, become figureheads ... Both their roles and their traits emerge from their place in an ancient narrative. The narration, the plot, comes first ..." Whatever else figures are doing, they are filling their roles. A confidante may have gone to buy fish, but her real role is the sharing of confidences. "A figure is neither formed by nor owns experience." They are MaryorMartha, Peteror Paul, Che Guevara or Paul Bunyan. The idea ofpersons) Rotty proposes, comes from two sources: the

dramatis personae of the stage, and the law. "A person's roles and his place in the narrative devolve from the choices that place him in a structural system, related to others." Cent.ral to it is the ideaofa unified center of action and choice-the unit of both legal and theological responsibility. Interest in persons, then, centers upon locating liability.

Two Modes ofThought · 41

The scope ofa person lies in his powers to affect those around him, a scope for which he bears responsibility. When we conceive ofpersons exclusively as sources ofresponsibility,

we think of them as souls or minds, engaged with res cogitans. When we think ofthem as possessing rights and powers, we think ofthem as selves. "When a society has changed so that individuals acquire their rights by virtue of their powers, rather than having their powers defined by their rights, the conceptofperson has been transformed to a concept ofself." Jane Austen describes a world ofpersons on the verge of becoming selves, Trollope one that has already become a world of selves, one in which the property required for stature is no longer land but an assured income due one by virtue ofone's qualities. Finally, individuality) born out of the corruption in societies of

selves: "It begins with conscience and ends with consciousness." At its core is a contrast ofindividual versus society: "an individual transcends and resists what is binding and oppressive in society and does so from an original natural position . . .The rights ofpersons are formulated in society, while the rights of individuals are demanded ofsociety." And so Molloy and Malone, the zaniness of the individual soldier in the midst of an insane war, rip-off as the redistribution ofproperty. Each is a mode ofinterpreting as well as a mode ofdepiction, and in

both, the lines are not clear. Depictions achieve drama by embodying a conflict: is Leggatt in The Secret Sharer a "figure" or an "individual" in Rorty's sense? And as writers alter their "presentation" of person- hood-from the figures ofHomer to the characters ofEuripides, from Jane Austen's persons to Trollope's selves, from Conrad's selves to Beckett's individuals-so too readers change in the approach to per- sonhood. In life, is it crusading senator or macho lover of Marilyn Monroe, a teenage offender in the light of love or the light of justice, which Roger Casement, which of the two Parnells. In literature, is Roth's Zuckerman a character who searches for the setting that will uncork his gifts, the figure in a morality drama, or the individual in revolt? Lionel Trilling, reviewing David Riesman's The Lonely Crowd) con-

jecturedwhether modern sociologywas coming to take the place ofthe novel as a window on the lives of those who live in "other" social classes. But that cannot be right. For the anomaly ofpersonhood-its

42 · Two Natural Kinds

consequential alternativeness-cannot be caught save through the ve- hicle of narrative. And it is this alternativeness-this inherent rest- lessness in deciding on the right depiction ofpersonhood-that gives the novel ofcharacter, the psychological novel, its force, its subjunctiv- ity, and its power to disturb.

One final point and I am done. It is about narrative and history. In a recent book on historiography, Dale Porter raises some extremely in- teresting questions about the strengths and shortcomings ofnarrative history. I do not want to evaluate his arguments, but to comment on one point that recurs in his account (as it has in earlier accounts by Bryce Gallie and by Isaiah Berlin). There is an assumption, implicit to be sure, that a narrative account leaves one open to "errors" that are departures from an aboriginal reality that is better discerned by a more systematic, "logico-scientific" method. Mter all, what we know, the annates) so to speak, is that on Christmas Day at the·Vatican in the year 800, Pope Leo III crowned Charlemagne Emperor of the Holy Ro- man Empire. When an historian of the stature ofLouis Halphen sets these bare "facts" into a web of imperial and papal intentions and of changing "world views," does he risk errors that are more egregious and fanciful than, say, the errors in wait for a sober economic historian who eschews narratives? That one does somethingmore verifiable than the other, few would doubt. Trade and commerce, the flow ofcapital, and so on are documentable in a way that motives and a growing "sense of Europeanness" are not. So should Halphen's account be treated as a form offiction (or "faction") or as fictionalized history? The economist Robert Heilbroner once remarked that when fore-

casts based on economic theory fail, he and his colleagues take to telling stories-about Japanese managers, about the Zurich "snake," about the Bank of England's "determination" to keep sterling from falling. There is a curious anomaly here: businessmen and bankers today (like men ofaffairs ofall ages) guide their decisions by just such stories-even when a workable theory is available. These narratives, once acted out, "make" events and "make" history. They contribute to the reality ofthe participants. For an economist (or an economic histo- rian) to ignore them, even on grounds that "general economic forces" shape the world ofeconomics, would be to don blinders. Can anyone

Two Modes ofThought · 43

say a priori that history is completely independent ofwhat goes on in the minds of its participants? Narratives may be the last resort of economic theorists. But they are probably the life stuffofthose whose behavior they study. So we embellish our hard-core annates) convert them into chroniques

and finally into narrative histoires (to borrow Hayden White's way of putting it). And thereby we constitute the psychological and cultural reality in which the participants in history actually live. In the end, then, the narrative and the paradigmatic come to live side by side. All the more reason for us to move toward an understanding of what is involved in telling and understanding great stories, and how it is that stories create a reality of their own-in life as in art.