00001.tif

PDF

Download

Open document

Transcript (if available)

Content AN INQUIRY INTO THE PROBLEM OF STYLE: Barbara Bennison Gray A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (English) A NEGATIVE EXPERIMENT by UMI Number: DP23030 All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. Dissertation FiMshsng UMI DP23030 Published by ProQuest LLC (2014). Copyright in the Dissertation held by the Author. Microform Edition © ProQuest LLC. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106-1346 UNIVERSITY OF SOUTHERN CALIFORNIA G R A D U A T E S C H O O L U N IV E R S IT Y P A R K L O S A N G E L E S 7, C A L IF O R N IA ph.O £ G 77% T h is dissertation, w ritte n by ................... Earbs,r.a..B .e.niuaoji..G 3:a.y..................... u n d e r the d ire c tio n o f hsx....D issertation C o m m ittee, and a p p ro ve d by a ll its members, has been presented to and accepted by the G raduate S chool, in p a rtia l fu lfillm e n t o f requirem ents f o r the degree of D O C T O R O F P H I L O S O P H Y Dean Date--.. August.,.. 1. 9. 6-4 DISSERTATION COMMITTEE Chairm an TABLE OF CONTENTS Chapter I. THE PROBLEM OF STYLE .... II. AIMS AND HYPOTHESES .... III. STYLE AS BEHAVIOR ........ IV. STYLE AS THE SPEAKER .... V. STYLE AS THE LATENT .... VI. STYLE AS THE INDIVIDUAL . . VII. STYLE AS THE IMPLICIT SPEAKER VIII. STYLE AS LANGUAGE ........ APPENDIX ............................ BIBLIOGRAPHY . . ................... Ai. CHAPTER I THE PROBLEM OF STYLE In the past sixty years the crucial problems of liter ary scholarship have come more and more to be seen as prob lems of definition. Thus, not only has the terminology of literary study come under question but also the subject of literary study— the nature of literature itself. Less and less do theorists and critics ask of literature What is or should be its use. More and more the crucial, the critical, question has come to be: What is literature? The concerns of literary theory have gradually shifted from the teleo- logical— catharsis and imitation, pleasure and instruction, and even self-expression— through Kant's ZweckmSssigkeit ohne Zweck, to disregard of purpose, to emphasis instead on nature and function, on the ontological. And more and more the result of such concerns, the answer to such a question has come to be, "Literature is language which ..." Elder Olson, who as a Chicago Neo-Aristotelian dissents 2 •from such definitions, nevertheless recognizes their preva lence . Nowadays when the nature of poetry has become so uncer tain that everyone is trying to define it, definitions usually begin: "Poetry is words which, or language which, or discourse which," and so forth.^ "As a matter of fact," he adds dogmatically, "it is nothing of the kind." Yet logically as well as empirically it seems 'impossible to agree with such a conclusion, to go along, for example, with Olson's most recent statement, that "the greater, and the chief part, of playwriting has nothing to 2 do with words." As a matter of fact, Olson's critical practice actually refutes these theoretical pronouncements. One example from his analysis of Dylan Thomas should suffice, We depend upon diction first of all to get at the poem; unless we can grasp its meaning, we can hardly penetrate to character or activity or situation or anything else. SFo New Critic would disagree with this. If we cannot under stand a poem or a novel or a play without understanding the words which constitute it, words must have something ■*""An Outline of Poetic Theory," in R. S. Crane, ed. , Critics and Criticism: Ancient and Modern (Chicago, 1952), ;p. 564. ^Tragedy and the Theory of Drama (Detroit, 1961), p. 9. _____^The Poetry of _Dylan .Thomas (Chicago., 1954)_,__p_. 53..____ 3 important to do with the writing, and surely the reading, of a poem or any other piece of literature. It may be that 'literature cannot be defined as "language which," but there is no getting around the fact that literature is linguistic — ■ whatever else it might be as well. On the other hand, to accept a definition of literature as 1 1 language which— " seems equally impossible, simply be cause no one has yet been able to fill in the blank with a predication which actually does comprise the distinctive function or attribute necessary to an adequate definition. Literature may indeed be "language which— " but which does or is or has what? Murray Krieger, in describing the theo ries of the critics whom he calls the new apologists for poetry, admits it is true that from their science-poetry distinction onward, our critics have tried to show, even in their most esoteric interpretations of complexities, that the language of poetry is the key to everything else.^ These new apologists include the major British and American critics of the past half century. Yet if language is the key, these men of keen intelligence and vast experience with literature have yet to unlock an adequate definition with fit. The language of literature often is ambiguous or ^The New Apologists for Poetry (Minneapolis, 1956), p. 94.__________________________________________________________ ---------------------------------------------------------------------------— ~~ 4 metaphorical or elliptical or deviational simply because language often is such, whether literary or otherwise. No difference of degree can be consistently distinguished, and no difference of kind has been demonstrated. Take, for example, the current contention that litera ture may somehow be defined as metaphor. "The essence of poetry," declares Cleanth Brooks, "is metaphor," and W. K. Wimsatt, Jr., in his epilogue to Literary Criticism; A Short 'History, written with some assistance from Brooks, also pro nounces metaphor "the principle of all poetry."^ Such a conception of the defining feature of poetry is the pivotal point of the work of a critic like Philip Wheelwright, and it inevitably involves, as in Wheelwright, a distinction be tween literary and other "kinds" of language— notably "scientific"— and something of a retreat towards mysticism.^ 8 That Brooks' most recent work is entitled The Hidden God comes finally as no surprise. Yet such contentions must yield to the fact that metaphor is a prevalent feature of ^The We 11-Wrought Urn (New York, 1947) , p. 223. 6(New York, 1957), p. 750. ^The Burning Fountain (Bloomington, 1954), p. 61. Cf. also Metaphor and Reality (Bloomington, 1962). ?New_Hayen ,_19.6 3-______________ : __________________________ - - — much non-literary discourse, and even of so-called non- poetic instances of language, the scientific and the philo sophical. The density of metaphors in various parts of a work which is generally not classified as literature, Kant's .Critique of Pure Reason, has been successfully used to help determine the irregular sequence of the composition of the .Critique. The use of metaphor and other "figures" was one way in which Kant apparently hoped to make his work clear to the general public.9 Recognition of the fact that such fea tures as metaphor are an integral part of much more than strictly poetic or literary discourse has led some critics to assert that "metaphor is the omnipresent principle of language,"'*'9 in which case it can scarcely be the essence or the principle of literature. On the other hand, some critics are forced to admit that there are genuine poems, as well as parts of genuine poems, which are not metaphorical at all or even, more broadly speaking, "figurative."'*''*' q . . . S. Morris Engel, "On the Composition of the Critique, ' forthcoming in Ratio. Cf. also Max Black's analysis of metaphor as a tool of philosophical inquiry in Models and Metaphors (Cornell, 1962), pp. 25-47. ■ * " 9I. A. Richards, The Philosophy of Rhetoric (New York, 1936), p. 92. -^Monroe C. Beardsley, Aesthetics; Problems in the 'Philosophy—of—Criticism—(New—York,—1958)-,—p— 154------------ 6 A peculiar consequence, however, of this urge for defi nition, and particularly of these resultant definitions of literature, or poetry, as metaphor is the insistence that criticism, both in theory and practice, is and should be metaphoric too. "It is true," Wimsatt admits, that metaphor in poetry is not the same thing as meta phor in poetic theory. Yet a metaphorical theory of poetry is almost necessarily a theory of multiple fo cuses and hence a historic theory and a perspective theory. . . . It seems to us, finally, that metaphor is not only in a broad sense the principle of all poetry but is also inevitable in practical criticism and will be active there in proportion as criticism moves beyond the historical report or the academic exercise. In The Mirror and the Lamp M. H. Abrams distinguishes "the role in the history of criticism of certain more or less submerged conceptual models— what we may call 'archetypal analogies'— in helping to select, interpret, systematize, 1 ^ and evaluate the facts of art." There is no doubt that metaphors and analogies do form an integral part of the 'language of criticism. But whether or not they can be taken as the defining feature of all practical criticism that moves beyond the historical report or the academic exercise, 1 9 • ^Literary Criticism, p. 750. Cf. also Rxchard Foster, l"Criticism as Poetry," The New Romantics: A Reappraisal of the New Criticism (Bloomington, 1962). ^(Hew York, 1953), p. 31. any more than they can be taken as the defining feature of the language of literature, is the crucial question. What is important to decide is whether or not criticism ought to be metaphorical and analogical. Does there not come a time in literary criticism and scholarship when "the exploration of serviceable analogues, whose properties" are "by meta phorical transfer, predicated of a work of art" ceases to be helpful, when the study of literature falls victim to what Abrams has further called "the endemic disease of analogical thinking"— "hardening of the categories" (pp. vi, 34-35)? Such times must always come in philosophy and science, if these disciplines are to maintain intimate and precise understanding and control of their speculations, and, even ■more importantly, of the actual objects of their investiga tion. Metaphors and analogies are not arguments or proofs? it is not possible to agree or disagree with them in the agreed upon ways of intellectual discourse. The insistence that criticism should be metaphorical would inevitably re- 14 suit, perhaps already has resulted, in a Tower of Babel. •^This epithet is a recurrent one in Rene Wellek's Con cepts of Criticism (New Haven, 1963). Cf. pp. ix, 2, 54, with p. 311. Olson also uses the epithet and expands upon it in "An Outline of Poetic Theory," p. 546. But as a Neo- Aristotelian, who denies a basic unity in literature and literary studies, Olson finds no cause for alarm in such a situation._____ _______________________________ _______________ '8' •Cant does not attempt to prove anything with analogy and metaphor— only to explain and illuminate. Literature may be analogical and metaphorical, but surely criticism, if it is really going to be about literature, and not flights of •fancy inspired by it, ought to be subject to the same re quirements of rigor and responsibility that attend upon Dther scholarly discourse. Of late some effort has been made to retrieve the vo cabulary of literary criticism and scholarship from the ethereal heights of metaphor and analogy and to reduce the confusion of tongues by rigorous terminological analysis and evaluation. Wellek's latest work, Concepts of Criticism, is one man's effort in this direction. Moreover, interdisci plinary groups have shown increasing interest in trying to clarify terminology used across disciplines. A notable jrecent example was the conference held under the joint aus pices of the Social Science Research Council and Indiana I Unxversity to explore the possibility of finding a common basis for discussing and, hopefully, understanding, particularly among linguists, psychologists, and literary critics, the characteristics of style in language. Out of such discussions, it was hoped, might come a clearer percep tion of what literature is and what the constituent ele ments of style are. If literature is an aspect of be havior, is there any way in which these groups can reach a meeting of minds on the nature of this behavior and its_place_in_human_cul.ture? How_can_the_understandings___ 9 of one group be used to shed light on those of other groups and on the whole problem of style in literature?^ Attendance at the conference comprised a cross-section of leaders in a wide variety of disciplines, literary scholars and critics such as Wellek, I. A. Richards, Wimsatt; the philosopher Monroe C. Beardsley and the philosopher and linguist Rulon Wells; folklore scholars such as Richard M. Dorson; psychologists such as George A. Miller of Harvard and C. E. Osgood of Illinois; the secretary of the Linguis tic Society of America, Archibald A. Hill; the internation ally famous scholar with a record of significant contribu tions in general linguistics, poetics, Slavic philology, literary history, folklore, and Paleosiberian languages, Roman Jakobson. Yet despite the diversity and erudition of this array and the laudable aims of the conference, one would have, on reading the published record, to agree with Wellek's closing statement that the conference has not been a success . . . if its purpose was to estab lish a common language and to throw light on its pro fessed central topic, the problem of style and particu larly of style in literature and methods of analyzing style. l^John W. Ashton, Foreword, in Thomas A. Sebeok, ed., Style in Language (New York, 1960), pp. v-vi. •^In Sebeok, ed., p. 408. 10 Anything which the conference did succeed in accomplishing must be considered incidental to this central failure. No doubt a number of things might account for this ■failure. Wellek was particularly struck by the fact that the question of style has not been discussed at all in terms of the enormous labor which has gone into it for centuries or in terms even of the theories and methods of the many contemporary practitioners of stylistics who come to the mind of every student of literature.(p. 408) Only one passing, and disparaging, reference was made to 'the studies produced by the adherents of the so-called Sleo-Idealistic school (of the Croce, Vossler, Spitzer brand)," by the linguist Edward Stankiewicz, who claimed that this "school" has not "contributed significantly to the exploration of style problems because of their programmatic disinterest in theoretical concepts and in a strict method- 17 ology." No mention at all was made of Erich Auerbach, Damaso Alonso, Amado Alonso, for example, or of the very ac tive French school of stylistics inspired by Ferdinand de Saussure's pupil, Charles Bally, which supposedly models its ■investigations on the " structural" conception of language developed by Saussure and promulgated by Leonard Bloomfield. Another notable omission was that of detailed analyses of I. literary texts, something that might reasonably be expected 17 _______In Sebeok, ed. , p. 96._________________________________ ------------------------------------------------------------------------------------------rr •from a conference proposing to examine the concept of style in literature. Only Fred H. Higginson's paper on "Style in Finnegans Wake" could qualify as one, and it is primarily a chronological comparison of numerous manuscript versions, a tind of analysis not often available to students of style, who are usually confronted with the one and final text. Ihus it was in fact not a practical application of any par ticular conception of style, and Higginson need not have used the term at all. Finally, one of the chief inhibitions to communication and agreement which arose during the con ference and which seems to arise nowadays whenever scholars in the "social sciences" and the "humanities" meet was the peculiar and disturbing issue that can only be called "scientism." In his foreword Ashton noted that there had emerged during the conference a clear difference in method and understanding in the treatment of the questions whether problems of style (however it might be defined) might be treated quanti tatively like problems of social behavior, whether lit erature is amenable to truly scientific analysis, whether the scientific analyses proposed were really meaningful to the literary critic. (p. vi) When the poltergeist of the quantitative, "the truly scien tific," is called up, the social "scientist" becomes aggres sively statistical and the literary scholar is thrown into . r 2 1 f t the defensive posture of unscientific humanist. Yet surely this whole difference over "scientific anal ysis" is a smoke screen obscuring the real aims and needs in the study of language and literature. Insistence upon quan tification is about as useful as the critic's insistence upon metaphor. The assertion that nothing is amenable to ■rigorous investigation which cannot be counted makes about as much sense as the assertion that literature is not amen able to rigorous investigation. Both assertions ignore what should be, must be, the first and fundamental question: What is it? What is it you are counting? What is it you are being metaphorical about? Critics of the metaphorical persuasion such as Wimsatt insist upon the reality of style as dogmatically as the psychologists who think it can be quantified. "There are certain kinds of contentual [sic] meaning," says Wimsatt, "which can scarcely be discussed except under the aspect of technique, style, 'form.' These 19 meanings are pre-eminently the ironic-metaphoric." 'Form," he further explains, "is technique and style" (p. 748). The psychologist John B. Carroll in his paper on 18 Cf. e.g. Miller, pp. 392-393, and Wellek, p. 409. ■^Literary Criticism, p. 747. r3 "Vectors of Prose Style" would seem to be in agreement with such an assertion, for he begins It takes little argument or evidence to secure agreement that there are different manners of writing, and that these differ among themselves not only by virtue of the content or the subject matter treated but also by virtue of a host of "stylistic" elements which are present in varying degrees of samples of prose. The improbability of ever coming to a "clearer perception of what literature is and what the iconstituent elements of style are," indeed, of ever coming to a clearer perception of whether or not there even is such an entity or quality as "style," seems patent in the glare of these refusals to de fine terms, to submit conceptions to the test of practical application, and to benefit from the experience of other scholars who have come to grips with the problem of style and learned something from the encounter. The irony of this "scientistic" discussion and dis agreement about how style can be studied is threefold. In the first place, beneath the disputes over approaches and the proliferation of secondary terminology— "ironic- metaphoric," "vector"— lurks an unacknowledged agreement. The psychologist, the linguist, the aesthetician, and the literary scholar all share basic preconceptions about the 20 In Sebeok, ed., p. 283. 14 nature of "style." In the second place, this irony is com pounded by the fact that, no matter who uses it, the term is and continues to function, regardless of the coining of the scientific-sounding term "stylistics," as a metaphor. Fin ally, it must be recognized that insofar as the conception of style entailed by the psychologist's, the linguist's, the aesthetician1s, and the literary scholar's use of the term has been applied to literary analysis it has shown itself in practice unworkable. It has proven unworkable precisely be cause it is a metaphor, a term and conception borrowed from another area of human experience than the one to which it has now been transferred, and one to which the term and the conception are not only alien but also severely distorting. That the term "style" was originally a metaphor every one seems to know, for almost every book on it begins with a reference to its metaphorical etymology. "Style is primar ily a quality of writing," begins a recent work on Style in the French Novel; "it comes from the Latin stilus, the name of the writing rod, and it is only by metaphor that it came 21 to be applied to other activities." But such an appeal to ^Stephen Ullmann (Cambridge, 1957), p. 1. Cf. for example the beginnings of works on style as far apart in time as Sir Walter Raleigh, Style (London, 2nd ed., 1897), pp. 1-2, and F. L. Lucas, Style (London, 1955), pp. 3-4. It 15 etymology is extremely misleading. A great many changes have occurred in the use of the word since it originally applied to a quality of writing, for the writing then re ferred to was not composition but handwriting, script. "Style" is still used to refer to a quality of writing, but writing in a considerably different sense than how letters are shaped by the pen or stilus— to writing as language, as literature. In the meantime, as early as Cicero, the term was adopted— not by poets— but by orators and rhetoricians, and it is in their sense, as a quality or attribute of speech that the term still functions. Thus, "style" is ap plied to literature only by extension. For literature dif fers sharply in several important ways from oratory and even from speech. Insofar as style in literature is still con ceived of as an attribute of language which can be analyzed and treated in relation to but distinguishable from the meaning of the individual work of literature, "style" func tions as a metaphor. Furthermore, it is used to designate is important, however, to remember that the word "style" came from Latin, not from Greek. Aristotle, for example, uses lexis (as opposed to taxis) for what is usually trans lated as "style" or "diction." The Greek stylos means "pil lar," as in the name of the ascetic Simon Stylites. The spelling "styl-" instead of "stil-" is based upon false preek etymologizing. Cf. Oxford English Dictionary under i " Style."________________________________________________________ 16 in literary works an entity or attribute which does not and cannot even metaphorically be said to exist. Etymology, of course, is not proof, any more than meta phor and analogy are. What is necessary is to show, by a logical analysis of its use in current literary discussion, that the term is inappropriate and inadaptable to literature "Style" could serve someone some day as an interesting sub ject for a study in the history of ideas, but such a study will not help clarify the terminological and practical dif ficulties of its present use. The participants in the Indiana conference had expected that a solution to the prob lem of style would yield "a clearer perception of what lit erature is" and in what way it can be considered, to use Ashton's phrase, "an aspect of behavior." Yet, it is my contention, that social scientists and literary scholars who conceive of literature as behavior seriously misunderstand its nature, so that, on the contrary, a solution to the problem of style actually waits upon an adequate conception of what literature is. If one continues to think of litera ture itself as a kind of human behavior— that is, as speech- then one can scarcely achieve an adequate conception of lit erature, an adequate comprehension of the term "style" and what it entails, and quite possibly not even a workable . - 1 7 - conception of the nature of language, whether in literary or any other kind of discourse. A final reason for investigat ing the logical function in current discussion of a term such as "style" is that use of the term frequently entails the use of a number of other terms which are also metaphor ical when applied to literature and sometimes even when ap plied to language. The logical function of these terms would thus also be clarified through an examination of "style." What do such terms as "expression," "choice," "saying," "rhythm," "individual," "trait," mean when applied to style in language and to literature as a linguistic phe nomenon? If literature is not speech, then perhaps such terms as "tone" and "irony" also need analysis and recon sideration for their value in describing the individual work of literature. This should not be a reconsideration of what in rhetoric and other literary theories they have meant, but what in view of the nature of literature they could possibly mean now. What I should like to do in this dissertation is to see how the concept of style functions in relation to the indi vidual literary work. Can it be said that a work has style or a style? If so, what is this attribute, how does it work, can it be studied and analyzed in itself? By doing so rs I hope to throw some light on the problem of critical termi nology, on the nature and function of literature, and per haps even on the nature of language, at least insofar as literary works are linguistic. But in doing so I have to begin with some conception of the nature of literature in order to see whether or not the concept of style is appli cable to it. This procedure is not really circular, any more than the scientist's knowing what he hopes to prove by an experiment before he has conducted it is circular. Some conception of the nature of literature is a fundamental pre supposition from which all literary studies actually are conducted. What I want to do here is to translate my pre supposition into an hypothesis, to bring it out in the open as a tool of my analysis rather than leave it as a hidden determinant. One of the chief contributions of the Chicago Neo-Aristotelians has been the stressing of the need for scrutinizing the assumptions of a critic in order to assess 22 the value of his criticism. Perhaps it might also be stressed that the critic should examine his own assumptions before presuming to offer criticism or terminology derived 22See R. S. Crane, The Languages of Criticism and the Structure of Poetry (Toronto, 1953), and Critics and Criti cism, especially Richard McKeon, "The Philosophic Bases of Art and Criticism." _______ 19 from them. One may then disagree with this analysis of style and related terms because the conception of literature on which it is based is faulty; one may disagree with it be cause, although the conception of literature is sound, the application of "style" to it is faulty; one may agree with the analysis of "style" and yet disagree with the conception of literature. But one will then at least be able to pin point his agreement or disagreement. My premises, hopefully, are on the table. I contend that the application of "style" to literature is based on a false analogy of literature with speech; I contend that the language of criticism ought not to be metaphorical and analogical; I contend that literature is language, not "language which— ," however, but "state ments which— ," statements which function not metaphorically but analogously, and that it is to the work of literature and not to critical and scholarly statements about litera ture that the concept of analogy most fruitfully applies. Thus I will begin by presenting my hypothesis of the nature and function of literature. Secondly, by analyzing a social scientist's conception of the kind of entity style is and the kind of entity to which it applies, I hope to re veal not only the metaphorical and analogical assumptions of even the outspokenly "scientific" proponents of style 20 analysis, but also the basic metaphorical nature of the con cept itself when applied to literature. Next I should like to sketch briefly the background of the confusion between rhetoric and poetic, writing and speech, out of which the concept of style as a quality of literature has arisen and to point out the identity of the earliest and the most cur rent assumptions about the existence of style. Next I should like to examine the practical applications of the concept of style to literary analysis as exhibited in one of the most prolific and erudite students of literary style in the twentieth century, the linguist and Romance philologist, ’ Leo Spitzer. Next I should like to show that this concept of style used and rejected by Spitzer flourishes in contem porary European and American stylistics because of a mis understanding of the concept of individuality and what it logically entails. Then I should like to show that even when applied by current important American critics such as Wimsatt and analyzed at length by contemporary philosophers such as Beardsley, the concept of style still requires the erroneous conception of literature revealed in its use by the contemporary psychologist, the Renaissance rhetorician, the Romantic poet and critic, and the Romance philologist. Thus I hope to have exhibited the basic though unrecognized . — 21 identity of conception among those present at the Indiana conference as well as their identity with those whom they did not deign to mention but from whose experiences with the concept they might have profited. Finally, I should like to show that because of the basic analogical error of the con cept of style when applied to the individual work of litera ture, the about-to-be-established "new science of style," 23 . stylistics, envisioned by American and French linguists alike is a chimera. But some idea of what literature actu ally is must of course come first. 0 ^ Cf. e.g. Ullman's preface to Style in the French 'Novel, p. vii, and the text, p. 10._____________________ CHAPTER II AIMS AND HYPOTHESES The ensuing definition of literature, here to be pre sented ex cathedra, is being treated in full with the com plete apparatus of argument, exposition, and substantiation elsewhere."*" Here it is to be taken as a working hypothesis rather than a theory; the latter calls for argument, the former for explanation. This is the first preliminary point needing mention: the explanation of the definition will lack examples, for it is simply at the moment a way to start. The second preliminary point, to be developed more fully after the definition has been presented, but which is mentioned here as a guide, is that the definition contains nothing really new. Critics heretofore have either stopped short of propounding a^ theory of literature— like Wellek and Warren or Northrop Frye— or they have pursued a theory as far as they could and— like Murray Krieger— ended in a "*"In a projected doctoral dissertation by J. M. Gray, University of Southern California. 22 23 paradox.^ The following definition derives from what we all recognize about literature and the way it functions. What is new about it is that all that is commonly recognized about literature has been combined into an explicit state ment that is neither paradox nor metaphor. To take an example which will lead directly into the business of defining: as was pointed out earlier, litera ture has for some time been considered definable as a spe cial kind of language or at least according to a special in tensity of linguistic devices. As we have seen, even those who would deny the importance of language in literature run aground on the brute fact that, to begin with at any rate, all we have of a literary work is its language. Until we break that code we cannot decide whether or not any given piece of discourse is literary. On the other hand, no set of linguistic elements yields a definitively literary com plex which would furnish us with even a degree, much less a kind, of literary discourse. This difficulty can be solved not by distinguishing between literature and other kinds of language but between literature and other kinds of meaning ful utterances. Language, considered not as la langue— the ^Theory of Literature (New York, 2nd ed., 1956); Anat omy of Criticism (Princeton, 1957); The New Apologists for Poetry._________________________________________________________ — — 2 ~ 4 linguist's abstraction of words, phonemes, morphemes, and sentence patterns— but as la parole— speakers 1 and writers 1 unique and meaningful utterances— yields most importantly differences not of kinds of linguistic elements or devices but of kinds of statements. Every meaningful human utter ance can be designated a statement, as opposed to non language utterances, such as cries, laughter, nonsense. A work of literature is a meaningful human utterance; there fore a work of literature is a statement. The propadeutic question then becomes not how the language of literature differs from other language but how literary statements dif fer from other statements. Students of language have for some time been able to distinguish a good many kinds of statements, such as ques tions, commands, propositions. Perhaps the way to begin differentiating between literary works and other statements is by seeing which kind of statement a literary work most resembles. A specific proposition, as opposed to a general proposition or a question or a command, makes a referential statement about a space-time event. "The cat is on the table in the kitchen." "The salt dissolved in the glass of water." "Sigmund Freud died in 1939." These statements are referential, and we can verify them by looking at the table, 25 tasting the water, checking the records. Sometimes, how ever, a statement purporting to be about an event (i.e. a specific proposition) cannot be verified. "There is human life on a planet in the orbit of the star Sirius." Even if a speculative astronomer were to make this statement, we would have to shrug our shoulders. Such a statement, al though it clearly resembles that of the-cat-is-on-the-table sort, differs in the important respect that it is not veri fiable; that is, it cannot be proven true (or false) in any way that we know. A proposition purports to be true; a ref erential proposition purports to be about an event; a refer ential proposition which is verifiable ijs about an event. A referential proposition which is not verifiable is simply an unverifiable proposition, for its reference is inaccessible to proof. What is important about unverifiable propositions here is that they are the kind of statement most closely re sembling works of literature. Literary works are like unverifiable specific proposi tions. They purport to concern space-time events, but in fact do not. The following constitutes the opening lines of a literary statement: Dans une de ces planetes qui tournent autour de l'etoile nommee Sirius il y avait un jeune homme de beaucoup d'es- prit, que j'ai eu l'honneur de connaitre dans le dernier voyage qu'il fit sur notre petite fourmilifere; il s'ap- pelait Micromegas. . . . Like the sample unverifiable proposition stated above, the beginning of this statement concerns an event, the existence of human life ("un jeune homme") on a planet in the orbit of Sirius. And also like the sample proposition, it is unveri fiable according to the canon of inaccessibility. But here similarities end. To be sure, the quotation is just the be ginning of a literary statement. A proposition is re stricted to a single sentence, but a statement can, like H. C. Lea's A History of the Inquisition of the Middle Ages, consist of a series of related propositions. Both Lea's history and Micromegas concern unique events. Lea's state ment is about the ascendancy and decline of the Inquisition in the Middle Ages. It can be proven true or false, ade quate or inadequate, repetitive or original, by reference to historical documents, artifacts, and other histories. Micromegas concerns the experience of a young man from another planet on his visit to earth. The entire statement is unverifiable not only because we have no way of knowing whether or not there are inhabited planets in the orbit of Sirius, but also because we have no way of knowing whether there is or ever was such a young man as Micromegas either ,in_hea v.en_o r_on_ear.th_who_exper.ienc ed_pr.ec i s ely_the_ady.en=— 27 tures Micromegas experienced and with whom Voltaire himself became acquainted when Micromegas last visited "notre petite fourmiliere." A work of literature is like a referential statement in that it purports to be about an event. A work of literature is like an unverifiable referential statement in that it is not accessible to proof. A work of literature ■may contain some propositions about events or refer to ac tual historical figures, as Micromegas refers to "le docteur Swift" and Maupertuis1 expedition, but the work as a whole, the entire statement, is not a proposition and not subject to verification. For although works of literature, histor ies, and propositions all concern events, they differ from each other in important ways. Though unverifiable, a literary work is not an unveri fiable proposition, because it is never a proposition— that is, an assertion that consists of a subject and predicate and can be judged true or false. The difference between the two becomes readily apparent when we compare a proposition such as "There is human life on a planet in the orbit of the star Sirius" and a one-sentence poem such as Ezra Pound's "In a Station of the Metro." The apparition of these faces in the crowd; Petals on a wet, black bough. 28 "These faces in the crowd are petals on a wet, black bough" is a metaphor but not a work of literature. Pound's poem, however, does not state a metaphor; it presents an event— a unique act of perception. At a specific instant these faces in the crowd are an apparition, that is, are to someone the sudden vision of petals on a wet, black bough. The use of the definite ("the") rather than indefinite ("a/an") noun determiners and the absence of predication signal the speci- ficity rather than generality of the statement which is ■Pound's poem. In addition, the preposition "in" and the noun-determiner "the" of the title serve to specify a par ticular place as the scene— that is, to make an apparition. This poem is not the statement then of a generally appropri ate metaphor for crowds in railway stations or a specifi cally appropriate metaphor for a particular crowd in a par ticular station. It is the presentation of a uniquely oc- curing reaction which constitutes a specific situation. Al though it is an unverifiable statement, it is not a proposi tion but a work of literature. A literary work does not .predicate something; it i_s something. Lea's and other histories, while not propositions, also resemble literary works in being statements about events. Yet a work of literature is never a history, because a 2^ history is a verifiable statement about an event, and a lit erary work is an unverifiable statement about an event. Micromegas, for example, is simply not subject to proof. We have no way of ascertaining whether or not the event de picted in it actually occurred. The existence of the insti tution and proceedings of the Inquisition, however, is com mon knowledge. Lea's specific details and general interpre tation may be disputed, but the occurrence of the Inquisi tion in the Middle Ages readily lends itself to corrobora tion. There are at least four ways in which we can recog nize that a particular work is unverifiable. First of all, the statement may be by nature unverifiable by being a sub jective reaction, such as the cry of longing which consti tutes "O Western Wind" or the unique perception presented in Pound's "In a Station of the Metro." Second, the statement may be accompanied by or contain incidental signals that it is not subject to verification: for instance, labels such as "novel" or "romance"; difference between the name of the author and that of the ostensible narrator— e.g. Daniel Defoe and Moll Flanders; reminders in the text that one is just reading a story— e.g. the famous "intrusions" of Trol lope and Thackeray and the indefiniteness about names and details of Cervantes. This latter device leads to consider 3 C » ation of a third kind of indication of unverifiability, the obvious lack of any precise information to aid verification. Moll Flanders, for instance, appears to be documentary with out in fact presenting any documentation, and the narrator at once declares her pseudonymity. Finally, the obvious fantasticality of a work will indicate its unverifiability. .Gulliver's Travels, like Moll Flanders, appears to be docu- Lentary, but Gulliver's "facts" resist any incorporation into the established body of verifiable human knowledge. In order to distinguish such non-propositional, unverifiable statements about events from any other kinds of statements about events, we will call them statements of events. Works of literature, as distinct from all other kinds of state ments, are statements of_ events. This defining character istic of literary works constitutes the first half of my definition of literature. "Event" as I use it here simply designates any space time occurrence, whether the mere existence of an object or situation, or a change, a transaction. As a matter of fact, literary works can be distinguished among themselves accord ing to the two kinds of events they may present: events which are utterances and events which are actions. A liter ary work which is an utterance is usually, of course, a 31 subjective reaction, such as "O Western Wind" or "Let Me Not to the Marriage of True Minds." "0 Western Wind" is the statement of longing for a lover; "Let Me Not to the Marri age of True Minds" is the protestation of steadfastness in true love. Auden's "Musee des Beaux Arts" is also an utter ance, the musings of someone at an art exhibit as he moves from picture to picture by the "old masters." But an event which is an act of presentation is not confined to lyric poetry. Moll Flanders is not as an event a story about the adventures of a female rake in seventeenth century England; as an event it is an utterance, Moll Flanders in the process of relating her memoirs. Often an utterance is also the mode of narration. The mode of narration in Conrad's "Youth," for example, is "first-person," though this first- person can be identified only by inference as a man of some success in the world who had been present when Marlow told his story. The particular occasion of his recounting of Marlow's recounting of his adventure is not given. The event in "Youth" is actually someone telling about Marlow's telling of his adventure and is an utterance. A literary work in which the event is an action is often a statement in the third-person-omniscient mode of narration: e.g. An American Tragedy, where the event is not an utterance but 32 the social rise and fall of the protagonist, its causes and its effects on the people around him. A literary work which is the statement of an action may also be one written in the "objective" stream-of-consciousness technique: e.g. Ulysses, where the action "takes place" often in the minds of various characters, but the work itself is not an utterance. Now it may be noticed that in all the examples of events I have given, the event consists of the depiction of a human experience, either an utterance or an action, always involving a person, persons, or human-like creatures, how ever anonymous they may be. This derives from the fact that, although an event is defined as any space-time occur rence, the literary event is by the nature of its "literari ness" more restricted than that. The nature of a work of literature is by definition the statement of an event; but the function of works of literature is reflected in and re stricts the kind of event a work depicts. Events in litera ture turn out always to be human experiences, because a work of literature always functions as an analogue in an implied analogy with an aspect of human experience. Analogy is the traditional term for a mode of reasoning in which from the similarity of two things in certain par ticulars or relationships their similarity in other particu 33 lars or relationships is inferred. "As the acorn eventually grows into the oak, so the boy eventually grows into the ■man." Analogy differs from simple comparison, simile, or metaphor in that more than one particular characteristic is compared and further similarities are inferred. The acorn and the boy resemble each other in their initial smallness and in their potential for growth; oaks and men resemble each other in being considered large and full-grown. By recognizing the relationship between the acorn and the oak, a boy may vividly recognize the difference between himself now and what he will become and thus will have inferred the relationship between boy and man. "Big oaks from little acorns grow," on the other hand, is simply a proposition, which may, however, in a given situation be considered one analogue in an implied analogy. In an implied analogy one relationship is stated, and this becomes an analogue only when a listener or reader provides another relationship which is analogous to the first. Although a work of literature is not a general proposi tion like "Big oaks from little acorns grow" but the state ment of an event, it too can and does function as an ana logue, an analogue in an implied analogy with an aspect of human experience. As the character or characters of a 34 literary work are to the event which it states, so are a person or persons to some aspect of human experience. Prom the number of similarities we detect between literary char acters and events and human beings and occurrences in " life, r we infer the relationship that holds for both the literary work and that aspect of human experience to which it is analogous. That relationship between the character or char acters and the event of a literary work and therefore that human experience to which the work is analogous we will call the theme. As differentiated from the utterance or action which is the event, the theme is the pattern of action of the work, what has often been called the controlling idea. The theme is the statement about that aspect of human expe rience to which the literary work is analogous. For each literary work there is one most adequate statement of the theme, although on any given reading we may not be able to state it. The Oxford philosopher Gilbert Ryle has empha sized the useful distinction between "knowing how" and "knowing that." We may understand and appreciate a work long before we are able to articulate its analogous meaning, i.e. its theme. We know how to read the work— to recognize that it is a statement of an event which functions analo g s ______The Concept of Mind (Oxford, 1949), esp. pp. 25-61. 35 gously with an aspect of human experience. But this is dif ferent from knowing that the theme of the work is such-and- such. The knowing how is an absolute preliminary require ment for every reader; the knowing that— the stating of the theme— is the task of literary criticism. Perhaps I can clarify the nature of the theme and therefore the analogousness of the literary work by distin guishing between themes, paraphrases, and implications. A paraphrase is a summary, with more or less detail, of the event depicted in the literary work; for example, a para phrase of Pride and Prejudice would cover the action of 'Elizabeth Bennet's changing relationship with Darcy and how it affects all the other man-woman relationships in the novel— Charlotte and Mr. Collins, Jane and Bingley, Lydia and Wickham, and even to some extent Mr. and Mrs. Bennet. A paraphrase would include specific characters, incidents, how the novel began and how it ended. There are any number of possible paraphrases, which would contain the same kind of information, including some statement about the event, but differ in length and abundance of detail. When we ask for a paraphrase we want to know specifically what happens in the work. The event is conceived as simply the basic unit of action. There is only one event per work of literature, 36 though any number of paraphrases. The implications of a literary work are also practically limitless. The statement of an implication is not an account of the specific literary work but a generalization about life, drawn from the work though not necessarily stated in it, which the work can be taken to corroborate. Implications may be of many different kinds— sociological, psychological, ethical, political, etc. An implication of Pride and Prejudice could concern the in tense social pressures on eighteenth century English women to marry. Another implication could be that women, in any age, ought not to sacrifice their principles to this pres sure by marrying the only and perhaps quite undesirable man who asks. Such generalizations are implied by the action of the novel, though not necessarily stated in it. The theme, finally, is a kind of middle ground between implications and paraphrases. There is only one most adequate conception of the theme, and it will be stated in terms of persons, rather than of the characters, and of life, rather than of the spe cific incidents of the work. The theme of Pride and Preju dice might be tentatively stated as the effect of the im pulse of pride and prejudice, for better and for worse, on the social relationships between men and women. Here the title itself furnishes a useful expression of the theme. 37 The theme is not "the moral" but the subject. It is not an ethical proposition but a statement which identifies what the work is about. It is prior to any implications which the work may have for morality. The simplicity of the theme derives from the attempt to reduce the complex world of a work of literature and of human experience to a concise statement that will cover both. An important implication of 'Pride and Prejudice is the interdependence and inter relatedness of all social behavior, an implication common to several of Jane Austen's novels. But this is too broad an idea to distinguish each of those novels from the others. The event of which Pride and Prejudice is a statement deals chiefly and consistently with only one specific kind of af fective human behavior, pride and its negative side preju dice, largely as it affects heterosexual relationships. Of course the one most adequate statement of the theme of any given work is contingent for adequate substantiation upon a thorough analysis of that particular work, a task for which we will not have the opportunity here. But as a con venient reference during the ensuing discussion of "style" I have added a brief appendix defining the four critical terms which arise out of the preceding definition of the nature and function of literature. Sample critical statements . . 3 8 - based on a single literary work will be given after each to illustrate that stage of criticism which it covers. Now it must suffice to summarize briefly some of the advantages of this theory of literature and its implications .for criticism. The definition stated in full is: A work of literature is a statement of an event which functions as an /analogue in an implied analogy with an aspect of human expe rience . First, as was suggested earlier, it is to be pointed out that the terms in and auxiliary to the defini tion are common coin. They have been systematized and made perhaps more precise than they were formerly, but they are terms already firmly established in the critical vocabulary. The definition of event used here, for example, was adapted 'from Whitehead by Susanne Langer and applied to literature in Feeling and Form.^ "Analogy" and "analogousness," impor tant concepts in Scholasticism, have been widely, although not systematically, applied to literature, perhaps most not ably in A. C. Bradley's famous lecture on "Poetry for Po- C etry's Sake." Francis Fergusson uses them, too, in The ■Idea of a Theater,^ but to refer to relations between the 4(New York, 1953), esp. pp. 208-325. ^Oxford Lectures on Poetry (New York, 1909) , pp. 4-32. 6 (Princeton, 1949), esp. appendix.________ __ 3 9 parts of the work rather than of the entire work to 1 1 life." And of course the essential 1 1 dramatic" character of all lit erature, what I have treated as its "eventness," has in the last twenty years become a critical commonplace and well- established pedagogical principle, as we see in this state ment from an enormously popular textbook and anthology of poetry: . . . all poetry, including even short lyrics or descrip tive pieces, involves a dramatic organization. This is clear when we reflect that every poem implies a speaker of the poem . . . and that the poem represents the reac tion of such a person to a situation, a scene, or an idea.7 YLy disagreement with this observation will become apparent later. Nevertheless, it is important to point out here that my conception of the nature of literature arises out of con temporary critical observations and not from some prior and personal metaphysics. Brooks and Warren's use of the term "poetry" throws into relief the fact that the preceding definition has been of literature rather than of poetry and has involved no dis tinctions of poetry, prose, drama, and other generic and quasi-generic terms. "Poetry" seems to have such a variety of kinds of meanings that the term "literature," which 7 'Cleanth Brooks and Robert Penn Warren, Understanding Poetry— (New_York,_2nd_ed...,_195.0)_,_p liv._____________________ alternates only between "anything written" and "imaginative literature," is preferable. Conversely, "poetry" is some times laudative, sometimes generic, sometimes formal. It seems ultimately definable only as something which purports, usually according to typographical conventions, to be poetry. Many non-literary works are certainly poetry ac cording to this description, and of course prose would ex tend far beyond literature. The term "non-fiction prose" does not really convey much, since it would include all writing which is not literature and not poetry, a father large area of human communication. Drama, on the other hand, refers essentially to produced or producible plays, works "acted out" in the theater. But it fulfills all the requirements for definition as literature, as the phenomenon of closet drama would indicate, and, at least in its written form, must be considered as literature. While the definition of literature I have presented does not cover all things that have been called literature, it does cover all the major kinds of things— lyric, drama, epic, novel— and almost all of the individual works that have been traditionally considered literature. A definition by common characteristic, as this one is, does not permit designating a work as literature before examining it? one _ 41 has to examine each object for the presence of the common characteristic which will classify it as literature. Some poems by great poets will not contain this characteristic, and a great deal of literature not studied academically will, ■But a definition is not a value judgment. It is assumed here that we will ultimately find out more about literature as a human phenomenon if we study instances of it before ■rather than after evaluating them. As for the mode of ex istence of the literary work, this must be considered to be the same as that of any statement, whether a father's extem poraneous , never-repeated, and unrecorded bed-time story or a written work which can be duplicated verbatim again and again. In the case of varying versions of, say, a ballad, the point at which one ballad text differs sufficiently as a statement from another to constitute a different literary work must always be a matter of personal judgment. In the case of folk tales, although different tales treat of the same event, insofar as they are different statements of it, they are different literary works, even though the action, characters, details are very similar. The folklorist's con cern is with recurring motifs and events; the student of •literature's is with uniquely occurring statements of events, .That is the way in which language is a prime of literature, ~ 42 and that is how literature differs from and is not reducible to myth. This definition of literature enables us to slip freely through the terminological wilderness, even to avoid onto logical swamps, such as that of the mode of existence of literary works. Finally, it helps us to escape the paradox of autonomy-versus-referentiality that has haunted the New Critics and was Krieger's despair. The notion of autonomy is found to be essentially a misconception, because by defi nition no meaningful human utterance can be autonomous. A meaningful utterance means— but not necessarily "referenti- ally." Referentiality is a rather narrow, positivistic, and inadequate notion of meaningfulness. Commands, questions, general propositions are not referential, nor is literature. Literature is neither autonomous nor referential; it is meaningful analogously. The language of a work of litera ture, that is, the particular statement, is all that we have jof a work from which to construe the event that will be meaningful to us as an analogue of an aspect of human expe rience. The statement is all that we have, so that the in terpretation of it becomes the primary task of reading and of criticism. Only after this interpreting of the statement can we talk about literature and life, for only then will we toe in possession of both analogues. The question now will be: How does the concept of style (and stylistics) function in the interpretation of literary statements? CHAPTER I I I STYLE AS BEHAVIOR The question of quantification has become a key issue in stylistics, one over which the literary scholars and the social scientists, particularly the psychologists and to some extent the linguists, continually clash. The literary scholars do not refuse to admit statistics, but they in ef fect refuse to accept them, or rather they simply tend to ignore them. 1 1 It is obvious that such a method," one liter ary scholar has concluded, "will be too crude to catch the finer nuances of style. Statistics can never be more than a strictly ancillary technique in style studies."^ In giving the "Closing Statement" from the viewpoint of psychology at the Indiana conference on style, George A. Miller, professor of psychology at Harvard, complained that the four statisti cal papers presented were by and large ignored by the con- ference. "This attitude puzzles me," he said, "because ^Ullmann, Style in the French Novel, p. 30. ^In Sebeok, ed., p. 392. 44 45 statistics is an old, traditional approach to problems of style" (p. 392). He added, however, that "perhaps psycholo gists tend to be a little more sophisticated about statis tics than either linguists or critics." And, as if this were not a highly questionable proposition in itself, he went on to ask, What are we trying to achieve with the statistical ap proach? I could imagine that there is room for both a science of style and an art of style, much as we have a science of pigments and an art of applying pigments to canvas. But I think it is not the intent of the present work that the statistical analysis should be so periph eral to the artistic essence as the chemistry of paint is to great art. Thus the psychologist, as well as the literary scholar, takes back with the left hand what he grants with the right. Statistics may be compared with chemistry, but it is ex pected that, after all, it should have more to do with the "artistic essence" than chemistry does. Furthermore, it is not quite clear what Miller intends should comprise the art of style, what the writer of litera ture or what the literary critic does. It does seem clear that he intends that statistics should comprise at least the major part of any science of style. Yet it is not clear that statistics has, thus far at least, contributed anything positive to establishing the existence of such an entity. The__most__ambitious_and__exhaustiye .study to date of the______ 4 _ 6 . application of statistical methods to the identification of authorship through analysis of style seems to have proved nothing more than that whatever individual style is, it is not something subject to statistical proof. Miller admitted that in an earlier discussion of individual style in his O 'Language and Communication^ he had regarded style only as "a problem in statistical inference" (p. 392), but that now he 'realized there is "vastly more to style" than counting. Still he would argue that counting has many positive virtues, It would be interesting to see what counting has so far proven, and so we will in Chapter IV, but it may be noted now that the results have not so far been positive contribu tions . The reason that they have not may even be guessed •beforehand. The psychologist's complete reliance on quanti fication as a guarantee of objectivity suggests a naive mis understanding of scientific, indeed, of all intellectual problem-solving. Quantification cannot in itself prove any thing. The important matters, the matters that will lead to objective proof or disproof, are the matters of deciding what is to be counted, what is the nature of that which is counted, and why it should be counted— what is intended to 3(New York, 1951), pp. 119-139.__________________________ 47 (be established by the statistics compiled. Miller was fur ther puzzled that no one seemed to have commented on Car roll's paper on "Vectors of Prose Style." Miller was par ticularly interested in what the literary scholars might have thought of Carroll's use of a spatial coordinate system In thinking about style Carroll assumed that he was ana lyzing a set of points in a four-dimensional space. These four dimensions characterize the stylistic space. If we state the value that any particular passage has on these four dimensions, we have placed it in the space of different styles and we have stated its distance from all other points in that space. This is a very powerful tool, if it is appropriate. (p. 393) •Did the other members of the conference, Miller wanted to know, think the spatial analogy is appropriate? No one said and, indeed, what could be said? What is the spatial anal ogy supposed to be appropriate to? It may be remembered that, as was mentioned in the introduction, Carroll did not attempt a definition of style, even as a working hypothesis for his experiment, but assumed "its" existence, although under such circumstances, there is no way of knowing what it, style, will be. Wellek's verdict on the paper was some thing more to the point. "The laborious calculations based on the opinions of 8 different judges about 150 passages of 300 words, according to 29 different criteria," Wellek pointed out, led only "to such obvious results as that the Ihumorous=seriousJ dis.tinc.tion_is_mor.e_reliable_than_the____ 48 'good-bad' or 'weak-strong' distinction" (p. 408). Whether such distinctions and such conclusions, statistically com puted or not, are appropriate to the problem of style, is the basic question to ask, not whether some spatial analogy is appropriate to the problem but whether prior assumptions and subsequent conclusions are. The use of statistics does ■not obviate the need for sound reasoning, for purposeful planning, for, in short, relevant hypotheses. To be sure, the literary scholar is often not familiar with the actual techniques of statistical calculation. Yet this should scarcely prevent him from being able to judge whether or not the theoretical bases, the practical aims, and the possible conclusions to be drawn from so-called "statistical experiments" are sound, valid, and, furthermore applicable to the problem at hand. The question need not be whether or not a statistical approach can establish facts about art but whether or not it in fact has. The literary scholar should, indeed, by virtue of his profession, be able to recognize when terms used to designate or explain phenom ena are not scientifically established at all but actually metaphorical or analogical. This should be especially the case if the term is one like "expression," a term which has had in literary study a long and prominent history. From the Romantics through Croce to a number of contemporary philosophers such as Susanne Langer, the term "expression" has had a lively career in the arts and particularly in lit erature- Now when the psychologist introduces such a term into the discussion of stylistics, it ought to be possible to decide how precisely it is being used and whether or not it applies to the literary fact. Perhaps the psychologist's conception of style is not applicable to literature, no mat ter whether quantified or not. For example, Miller conceives of his field, psychology, as the study of behavior, but not just adult human behavior. The proper area of psychology is all behavior, whether of rats, children, or psychotics, a very large area, as Miller admits. Now, he says, When psychologists try to talk about stylistics, they feel they should narrow it down to "expressive behavior," behavior which says something to another organism about the state of the behaving organism. (p. 387) \ Human language Miller then designates as "a subdivision of this broad area of expressive behavior" (p. 387), and "with in the range of language behavior there is a domain that we call literature" (p. 388). Is it not possible, however, without recourse to or escape from questions of quantifica tion to show that such a conception of language and particu lar ly_as_appl ieja_to_literature is not only irre 1 evant_b ut___ also misleading? Is language always behavior, and if so, whose? Is it always expressive behavior? Does it always "say" something to someone else about the state of the being behaving? Does literature in fact ever do this? The liter ary scholar cannot be considered scientific when speaking in metaphors and analogies. How much less so can those who claim to be working toward the establishment of an approach which, if it does not replace that of the critic, will sup posedly be the scientific as opposed to the "intuitive" (p. 395) analysis of style. By examining Miller's conception of "expressive behavior" it may be possible to demonstrate its irrelevance to most questions of style in language and to all questions of style in literature and to indicate the fundamental misunderstanding of the nature of language and literature that is implicit in the entire concept of style. The psychologist has on this point more in common with the linguist and the literary scholar than he realizes, although the "scientistic" dispute over statistics has obscured both the similarity and the fact that it is a misconception. Miller has defined expressive behavior as "behavior which says something to another organism about the state of the behaving organism." The first question which arises from such a definition is: What does the term "says" mean 51 when applied to a linguistic utterance, i.e. to a statement, as opposed to a cry or babbling? Or, to put the problem more precisely: Can that aspect of a linguistic utterance which "reveals" something about a speaker properly be said to say something about him? Miller's definition of expres sive behavior reveals an ambiguity which is a principal source of confusion among students who apply the tools of psychology to literature. It suggests that any linguistic utterance which can be classified as expressive behavior actually says two things, at once. It says (1) what the statement "means," and (2) what, intentionally or uninten tionally (consciously or unconsciously) the speaker "means." The first conception refers to the way in which language is meaningful; the second refers— not to the way in which lan guage is meaningful— but to the way in which an act of utterance is meaningful. Now, were someone to cry "Ouch!" he might be revealing that he had been hurt. The cry would be symptomatic of his condition, i.e. express his condition. But "ouch" is also meaningful as a linguistic convention and as such differs from other interjections, such as "alas," "hurray," and "hello," and all other words. Although it can be used again and again by many different speakers in many different _ . . — 52 situations and may each time be symptomatic of each speak er’s condition in each situation, it is a meaningful lin guistic element, a word. In a sentence such as "Ouch is an interjection" the word "expresses" no symptom, gives no clue whatsoever to the state of a speaker. The non-linguistic "meaning"— what the speaker is expressing about his own state— is dependent upon the act of utterance. In that sense "ouch" may mean in a given instance something quite different from or even contrary to its conventional use, its meaning as an utterance interpretable approximately as "I am in pain." Interjections are a class of utterances which quite frequently function as a speaker's "expression." But when we turn to more complex utterances, the difference between the meaning of the act of utterance and the meaning of the utterance becomes readily apparent. If we were to say now, "God's in his heaven,/All's right with the world," we might "mean" it ironically, i.e. we might mean there is no God or we think he is out; we might mean it seriously, i.e. that an omniscient, omnipotent being is watching over us; or we might simply be quoting a line of poetry, in which sense the utterance would be quite different in meaning if the lis tener knew that it was two lines from a particular poem. 53 Each one of these meanings, in order to qualify as expres sive behavior, behavior which "says" something about the speaker, would require knowledge of the act of utterance— of the context, suprasegmental phonemes, gestures, facial ex pressions, what the listener knew from past experience about the speaker's beliefs, sincerity, intentions, etc. In short the expressive "meaning" would depend almost entirely on the non-linguistic aspect of the statement as uttered in a spe cific situation. Thus, if the psychologist defines "expres sive behavior" as "behavior which says something to another organism about the state of the behaving organism," he is defining expressive behavior as symptomatic behavior. A symptom is that which is indicative of the presence of a particular condition. It differs from a sign in that the entire object signified by a symptom is the entire condition of which the symptom is a proper part; e.g., red spots are a symptom of measles, and measles is the entire condition begetting and including the red spots. A sign, on the other hand, may be one part of a total condition, which we associate with another sepa rate part. Thus, a ring around the moon is part of a weather condition, but what it signifies is rain— another proper part— and not the entire state of 'low pressure' weather.^ Inasmuch as a linguistic utterance reveals the condition of the speaker by being part of and deriving from it, the ^Susanne K. Langer, Philosophy in a New Key (Cambridge, Mass.. 1942) , p. 57 . ___________________________________________ utterance is a symptom of his condition. If the utterance does not derive from the condition of the speaker, it cannot •reveal anything about his condition. Unless an individual sries "ouch" because he is in pain, his act of utterance |vill not reveal that he is in pain. The word "ouch" will still mean, roughly, "I am in pain," but the act of utter ance will not mean that; it will "mean," or be symptomatic of, whatever condition in the speaker gave rise to it. And whether this "meaning" of the act of utterance is "under stood" or not will, of course, depend upon the "other organ ism" who is listening, and probably watching, at the moment. The behavioral scientist, then, when he is talking about language as expressive behavior is not talking about language as language— a conventional system of human inter course— but about language as symptomatic of psychological or physiological conditions. In Miller's sense of the term | " expression," all utterances, linguistic and otherwise, as well as facial "expressions," gestures, clothes, posture— ■in short, all behavior and results of behavior— can be ■interpreted as expressive. And what they will mean depends solely upon who is listening— or watching. In his sense of the term, "says" is really metaphorical, since what the be having organism actually says and what his behavior, his . 55 utterance, says to the other organism depends upon how the other organism interprets it. This does not even simply re- jsult in the age-old conception .of style as a difference be- jtween what is said (or done) and how it is said (or done) , but a confusion of what is said (or done) with how it is taken. This concept of expressive behavior seems to include so much; yet, it actually includes relatively little that is language. An utterance, as opposed to an act of utterance, can be considered expressive only when the utterance means the same thing that the act of utterance, of which it is a part, reveals. Thus an utterance is expressive when someone makes a statement about his own condition that does in fact correspond to it. Such statements as "I am in pain" or "I feel terrible" are expressive when they can be correlated by an observer with other symptoms of pain or feeling bad on jthe part of the speaker, so that the observer is relatively Isure the speaker does feel terrible or in pain. In this cind of instance, what the statement "says" and what the act of utterance— with the facial expressions, groans, and ges tures of pain accompanying it— "says" are one. If the speaker is not in pain, the statement "I am in pain," even as a part of an act of utterance, is not an instance of . 5 £ expressive behavior. When someone says, "I'm not prejudiced against negroes, but it is a known fact they can't hold a ijob," this act of utterance can readily be interpreted as expressive. Unlike "I am in pain," however, what this act of utterance expresses and what this statement means are not the same. The statement means that the speaker disavows any prejudice, but his act of saying it seems to express a rather obvious prejudice against negroes. The psychoana lyst's casebook abounds with examples of acts of utterance in which what a patient says and what his "saying" it "says" to the analyst are not only distinct but diametrically oppo site. In common parlance we often speak of a person's hav ing "expressed" his interest in something, poetry for exam ple. But this does not necessarily mean that this person has actually said, "I am interested in poetry," but only that his behavior— his eager discussion of various poets, his constant reading, etc.— can be interpreted as sympto matic of his interest. Of course he might be doing this for other reasons, in which case our diagnosis would be incor rect. Interpretation of symptoms, as medical doctors are well aware, is a very conjectural, complicated, and fallible practice. Thus we see that, while an instance of language, i.e. a 57 statement, may be symptomatic, it is always, by definition, meaningful, whether or not it is in addition symptomatic. jWhether or not it is symptomatic and what it is symptomatic i |of depend solely upon its occurrence and interpretation as an act of utterance. A symptom is a symptom only by virtue of its being interpreted as part of a condition which is its cause. A quavery voice may be symptomatic of the speaker's nervousness, but what he is saying in that voice— unless it is "I'm nervous"— is not? while what the quaveriness is ac tually symptomatic of, whether nervousness, deep grief, ex citement, or strangulation, can be determined only from the circumstances of the act of utterance. Language, then, is really a rather small "subdivision of expressive behavior." While acts of utterance with all their accompanying phenom ena often "say" a great deal about us, what we are actually saying is as likely as not to be about something quite other than our own condition at the instant of utterance. Lin guistic utterances are in themselves not behavior but only the products of behavior. What utterances mean and what be havior expresses are two different concepts and two differ ent things. It should clearly follow from the fact that only cer tain kinds of utterances can be symptomatic and those only --------------------------------------------------------------------------------------------------------------------- _ — 5 ' 8 in certain kinds of acts of utterance that the entire con ception of expressive behavior with its concommitant con ception of style as symptom is really irrelevant to any con ception of style in literature. For what does the behav ioral psychologist require a language utterance to be if it is to be an instance of expressive behavior? First, as we have seen, there must be a speaker, a behaving organism mak ing the utterance; second, there must be "another organism" to whom the utterance is symptomatic of the speaker. But is there, in fact, such a state of affairs, such an act, in a literary work or even in the reading or writing of a liter ary work? In Chapter II we defined a literary work as a statement of an event and distinguished the kinds of events of which it may be a statement. Now the question is: Are either or both of these kinds of events acts of utterance according to the requirements of the psychological analyst of style? Are they, in short, expressive behavior? Those works which are statements of actions can scarcely be acts of utterance, unless from the principle that every statement necessarily presupposes the existence of a speaker we deduce ones who must then be in these in stances the authors. The "speaker" would have to be the author: first, because the behaving organisms in the work 59 'have no referents; what there is of them is in the work and that is all there is. The work is not their "behavior" but the result of someone else's behavior. Whatever we may pos tulate of their behavior prior to the work or beyond it must necessarily result from our own private imaginative behavior-. Second, the work simply cannot express something, unless we take it to be the expression of the author, For "express" is properly a relational term? it requires an X that does the expressing and a Y that is expressed, and X and Y must be distinct.^ The X— the expresser— in the case of literature must be the author and the Y— the expressed— the literary work, because the literary work cannot express itself. If we say that a rose is red, we do not mean that the rose expresses redness but that it possesses the quality of redness. Similarly, if we say that a work is ironic, we do not mean that the work expresses irony but that it possesses the quality of irony. Redness and irony are not distinct from the rose and the literary work; they are in them. If we say that a literary work expresses irony and mean that it is ironic, i.e. that it possesses the quality of irony and do not mean that the irony of the work is an expression of someone, then we ought to say so. _____ ^Beardslev, _ Aesthetics., _ p . , . . . . 3.31,__________________________ 60 It is not really a question of whether or not the work was in fact expressive but only that the work is no evidence in itself for establishing this. The work may purport to be a writer's emotional response on a certain occasion. That it was is unverifiable. Hence the event, even when it is an emotional response, is literature. The student of biography will of necessity be much interested in the works of his subject. His subject may be a lyric poet, and there might ■be a good deal of extra-literary evidence to indicate that some personal experiences have been extensively used in the poetry. But in any case, the event which is the poem is not the poet in the act of composition. It can only purport to be. All it is is a statement he has offered to the world. As we have seen, expressive behavior requires an actor, an act, and a witness to whom it expresses something. Ac cording to the behavioral definition, the term "express" not only relates an expresser to a statement but also to a sec ond party to whom the statement must be expressive of the expresser. Thus, no statement, literary or otherwise, can properly be said to be expressive unless an individual states it in the presence of another individual, unless, in short, there is an act of utterance. A statement exists in and of itself and is not expressive unless it occurs in that 61 triangular situation we have described as an act of utter ance. If we read any statement divorced from an act of ut terance, that statement is not, strictly speaking, expres sive. In order to make it expressive we would have to pos tulate a speaker; that speaker would have to be whoever wrote the statement; and we would have to know something about him and the situation— act of composition— in which he made the statement. That is, we would have to determine whom the statement was expressive of, so that we could de termine what it expressed. In addition, we would have to postulate ourselves, the readers, as the individuals in whose presence the act became an expression. Yet, even if we found out everything we possibly could about the author, still we would never be able actually to determine what he was doing, feeling, or thinking in the act of utterance, of composition. The only evidence we have that there even was such an act is the work itself and that was only one third of such an act. The readers of the statement or even of the 'literary work cannot be considered the other organism in whose presence the statement is made because they are not, quite literally, in the presence of the act but only of the utterance. We must distinguish then between (1) an event which is the existence of a statement and (2) an event which 62 is an act of utterance. A work of literature is always a statement; it is never an act of utterance. It may contain one or more such acts which are expressive of the characters presented by the statement that is the work. But that is quite a different matter. Furthermore, this distinction holds for that first kind of event which the literary work may be a statement of: — an utterance. Those who advocate the definition of style as expression frequently regard a literary work which is a statement of an utterance— for example, the lyric poem "0 Western Wind"— as an expressive act simply because it im plies a speaker. We do and must, of course, infer a "speaker" from certain attributes of the poem; indeed, if we did not, the poem would not be the statement of an event. And surely it is. O western wind, when wilt thou blow That the small rain down can rain? Christ, that my love were in my arms And I in my bed again! In this case, there is apparent the apostrophe of the first line, the exclamation of the third, the personal pronouns "my" and "I" in the third and fourth lines. All of these attributes give the poem the character of an utterance. But still there is no speaker, in this instance not even a spe cif ic_author_whom_we_co.uld_po.stulate_as_the_speaker__becausei_ this poem is anonymous. There is simply the statement, analogous perhaps to any and all lovers longing for a be loved, but not expressive of any one lover. Regardless of jwhether or not the poem has style, it certainly does not have expression, nor is it an expression. We infer a speaker because the statement resembles in numerous respects expressions which frequently occur in acts of utterance, yet "0 Western Wind" is no more an expressive behavior than "ouch" would be if recorded on an otherwise blank sheet of paper. It will tell us nothing about who "said" it and "re veals" only that it is the statement of an event which is an utterance, but not an act of utterance. When we infer that the poem is an utterance, we are inferring something about the statement, i.e. it is an utterance. This is not the same as accepting it as an utterance and inferring something about its speaker. The psychologist then if he is going to analyze a poem as an expression, has to make two inferences. First, he has to infer that the poem is an utterance, i.e. that there is a speaker. This is where literary analysis begins. Second, he has to infer a personality for the speaker, a personality of whom this poem is just a symptom. Only here does psycho logical analysis begin. Confronted with a "behaving 64 organism" in the laboratory or consulting room he has to make only one inference, that the utterance is somehow ex pressive of the organism, and he can verify this inference 'by comparison with other actions and other acts of utter ance. Confronted with a poem that is a statement of an event which is an act of presentation, the psychologically- oriented stylistician has to move away from the poem into speculations about its origins: about the physical and mental behavior that precipitated the poem and the physical and mental state that precipitated the behavior that pre cipitated the poem from which he is inferring that there was such behavior and such a state. Such speculations, twice removed from the object which stimulates them, can scarcely be verified unless the stylistician reduces himself to the absurd postulate that the poem is a verification of the poem. That is indeed "the philological circle," from which there is no way out and back to the literary work. To think that "O Western Wind" has style because it is expressive is as absurd as thinking that it does not have style because it cannot be identified as the expression of a particular per sonality.® Style is not the man— neither the implicit ®Cf. Wolfgang Kayser: "Man hat zum Beispiel mittel- alterlicher Lyrik ieglichen "Stir1 absprechen wollen, weil 65 speaker nor the poet nor the poet's complexes. Style, if it 'is to be anything at all, must be an attribute of the liter ary work, something which it possesses, not what it ex presses . Miller himself recognized that his conception of ex pressive behavior did not seem to coordinate very well with ’highly stylized, extremely noncasual utterances of the sort that go into literature" (p. 387). But he failed to under stand why. Rather he thought of the conflict between psy chological and literary and linguistic conceptions as aris ing out of the fact that, on the one hand, psychologists in clude so much more behavior in their notion of style than others want to include, and that, on the other hand,, they prefer to concentrate their attention on "spontaneous ex pressive behavior" (p. 387), as opposed to "noncasual utter ances" like literature. This is not, however, quite to the point. For language is not primarily but only incidentally and occasionally a "subdivision of expressive behavior," as we have seen, and literature not at all. A literary work is not a symptomatic instance of expressive behavior, spontan eous or otherwise; it is not behavior at all. Lhr der Ausdruck der Personlichkeit mangele," in Das sprach- liche Kunstwerk (Bern. 7th.ed., 1961). p. 289.._______________ 66 The psychologists, perhaps unfortunately, have ipade of style a concept so sweeping as to be scarcely distinguish able from the general concern of psychology. Another psy chologist speaking at the Indiana conference, James J. Jenkins, described style in psychology as a hierarchical concept which might extend from an individual performance of some common activity to the entire behavior pattern of an individual. In the most extreme view we might speak of the "life style" of an individual— a personal mode of responding which typifies all the behaviors of the person concerned. In a more limited view we might regard style as a trait that manifests itself in some particular set of behav iors (e.g., social behaviors, athletic behaviors, prob lem-solving behaviors). In a still more limited view style might be thought of as a particular personal mod ification of a single narrow behavior (e.g., handwriting, dress, articulation), etc.^ ■Furthermore, the personal mode of responding, i.e. the "life style," is assumed to be ultimately deducible from any "par ticular modification of a single narrow behavior." Jenkins 'attempted, for instance, in his paper on "Commonality of jAssociation as an Indicator of More General Patterns of Verbal Behavior" to determine if what he knew about an in dividual 's reactions in word-association tests would enable 7"Commonality of Association as an Indicator of More General Patterns of Verbal Behavior," in Sebeok, ed., pp. RCL8j =309._____________________________________________________ 67 him to predict the individual's behavior and attitude in general and vice versa. In fact, "Much of the work of dif ferential psychologists," as Jenkins sees it, "is aimed at finding specific, readily observable behaviors which index, or predict, more general behaviors of greater importance" (p. 309) . Now one may wonder why the psychologist needs the term 'style" for all this. It would seem that terms more common to psychology, such as "character" or "personality" or even simply "individual" might do as well or better as technical terms for what the psychologist wishes to study here. After all, the idea of "a personal mode of responding" would not appear to be a concept readily distinguishable from "person ality," nor the "life style" of the individual from his "in dividuality." Be that as it may, one may grant the psychol ogist his need for such a term and such a distinction. What one cannot grant is that such a distinction and the aims and methods contingent upon its application have anything what soever to do with style in literature. Simply take, for example, the notion of prediction, which is fundamentally involved with the psychologist's con ception of style. Psychology, whether behavioral or other wise, is in many respects a science much like medicine; the i 68 cinds of studies it undertakes and their achievements are oased largely on principles of diagnosis and prognosis. Developing means of detecting symptoms, whether of disease, •frustration, patterns of thinking, or hidden motives, is a diagnostic function, and this function often involves at tempts to predict, even to modify "general behaviors of jgreater importance" on the basis of such understanding. This predictive aim of the psychologist leads him into deal ing with concepts of probability. That fifty per cent, say, of a certain kind of person will react in a certain predict able way in a certain situation is a kind of conclusion often sought by the psychologist. Yet all this is certainly alien |to the study of literature. A cultural "science" such as the study of literature is by the very nature of its subject matter an a posteriori science, devoted to describing, ana lyzing, classifying objects. While Jenkins may talk about j'behaviors," pluralizing it as if it were an object, behav ior is a process, and a process is not analyzable apart from ■its occurring. Miller defined psychology as the study of human behavior. But the study of literature has for its data only a certain kind of product of human behavior. jTherefore, the concept of prediction does not apply. There can be no prediction of products but only of occurrences. ; 69 One may be able some day to predict that, say, one tenth of one per cent of a given population will be authors, but one can never predict even the kind of thing they will write— even by means of probability statistics derived from what has been written. More than one person may behave the same j^ay more than once, but it does not follow that the products of that behavior will be the same. Many people in the course of history have written many poems many times. But there are no two or more identical poems. The essential difference between the discipline of psychology and that of literary study is not a difference of "scientific" versus ’unscientific" but of behavior versus objects, a crucial difference, that is, of kinds of subject matter. The study °f writing literature is a psychological study, but the study of literature is not. Linguistics, too, as we shall see, tries to impose predictive procedures upon the study of products of human behavior— under the guise of stylistics. t I I maintain and, hopefully, shall demonstrate that this sim ply cannot be done. The study of literature is the study of what is; in no way is it a study of what under given condi tions will or even might be. Use of the term "style" has served on this point as elsewhere only to obscure the essen tial differences in the aims, methods, and subject matter of 70 the various disciplines. If literary works are not symptoms or sets of symptoms then there is no psychological study of style in literature. This whole notion of style as symptom, as an index or indices to "more general behaviors of greater importance," phether these are conceived of as personality or soul or id or Weltanschauung, always leads away from, in fact can never be the study of, literary works. Insofar as the style of an object, whether or not a literary work, is conceived of as an index or indices to any entity beyond the object, that study cannot logically be considered a study of the object— but of something else. Literature seems in retrospect quite obviously not behavior or anything like it, and yet theories of style which propose that the soul or id or Weltanschauung or categories of perception of an author can be determined through analysis of his literary works are just as misled jand precisely in the same way as the psychologists who think that literary works can be analyzed in the same terms as be havior. By exploring the psychologist's notion of expres sive behavior we have shown at the simplest level that lit- srary works are not symptomatic of their authors, that lit- jerature in fact is not expressive and not behavior but father the product of a behavior which can only be guessed 71 'at from the existence of the literary work itself. Now, with this understanding well in mind, I should like to take the concept of style as symptom out of the test tube and ap ,ply it to the practice of stylistics. By doing so I should like to demonstrate that no theory of style which conceives of it as caused by and therefore symptomatic of some entity beyond the literary work can function adequately as a theory of literary style simply because in practice such a theory cannot be applied. Not only is it in theory invalid? it is in practice unworkable. CHAPTER IV STYLE AS THE SPEAKER Certainly there is no doubt that a relationship exists between a literary work and its author, and certainly there is no doubt that, given the work and some information about the author, we can make some very interesting correlations between the two, which may tell us a good deal about the author. But it does not follow from these axioms that we can go to the life and character of an author for an explan ation of his literary works, nor that we can reconstruct the life of the author through his works, nor that the attri butes of a literary work are the attributes of the mind and temper of the author. Yet, perhaps more keenly now than ever, scholars and critics are promoting this kind of study, often calling it now stylistics and talking about the "new science of style," with no essential change in premises from those of the sixteenth century rhetorician. The historical sources of this widespread notion among literary scholars, particularly those interested in lin- ________________________________ 7.2______________________________ 73 guistics and stylistics, that they can with scientific as surance interpret the author— whether his Weltanschauung or id or psyche or background or categories of perception or moral vision— through his work are doubtless manifold and largely beyond our scope. The assumption is in part an in heritance from Romanticism, as Abrams has so ably demon- 1 strated m The Mirror and the Lamp. But Romanticism, as Abrams also points out, brought to flower a plant already growing at least as long ago as the Renaissance. George Puttenham was well within the Renaissance critical tradition when he defined style in his Arte of English Poesie (1589) as the ornament and dress of poetry. According to the prin ciple of decorum, style, as ornament, must be suited to the "matter" and subject of the work and can be judged, by it self, as appropriate or inappropriate. Yet Puttenham was contradicting this notion of style as ornament when he also held that, beyond this style adapted to the subject of the work, is an individuality of language which expresses char acter . And because this continuall course and manner of writing or speech sheweth the matter and disposition of the writ ers minde more than one or few words or sentences can ■*"Esp. pp. 226-262. Cf. also Rene Wellek, A History of Modern_Cr.iticism: 1 7 . 5 . Q = : 1 . 9 . 5 . 0 . , — II (New_Hay.en ,_19.55.)____________ 74 shew, therefore, there be that have called stile the image of man, mentis character. . . . For if the man be grave, his speech and stile is grave; if lightheaded, his stile and language also light; . . . if it be humble, or base and meeke, so is also the language and stile.3 Puttenham never acknowledged the inherent contradiction in a theory which defines style in terms of both the author and the subject. His further suggestion that "Men doo chuse their subjects according to the mettal of their minds . . contained the germ of a solution, not, however, to the prob lem of style, but to that of how much and in what ways an author reveals himself in his work. These ideas were, of course, not entirely new in the Renaissance. There are several ancient statements of them,3 as well as the unique instance of Longinus' Peri Hupsous, which finds the main source of the sublime style in the thoughts and emotions of the speaker.^ No doubt the popu larity of Longinus during the eighteenth century attests to the fact that it was during this period that the conception 3In G. Gregory Smith, ed., Elizabethan Critical Essays, II (Oxford, 1937), 142-143, 153-154. JSee A. Otto, Die SprichwSrter und sprichwortlichen Redensarten der R5mer (Leipzig, 1890), p. 257. ^See Dionysus or Longinus, "On Literary Excellence," in Allan H. Gilbert, ed., Literary Criticism; Plato to Dryden (New York, 1940), esp. pp. 170-171. Cf., however, Samuel H. Monk, On the Sublime (2nd ed., Ann Arbor, 1960), pp. 12-14. 75 of style as ornament began to lose place to that of style as "mentis character." The history of this change is an inte gral part of the history of that Romantic revolution which 'left us so much great literature, though not a very satis factory theoretical framework for understanding and appreci ating it, and as such it is part of a longer story which has often been treated. What is important to show here is that these two contradictory concepts have remained extant, often side by side, throughout the last three hundred and fifty years without any essential modification and that, despite terminological and psychological sophistication, the concept of style as the man, which gained ascendancy in the early nineteenth century, retains even today in so-called scien tific stylistics the same inconsistent and unprovable impli cations it has always had. In analyzing the premises implicit in Puttenham's dis cussion as it epitomizes the ideas about style elaborated in I the eighteenth century, Abrams points out that Puttenham's concept contains two implicit assertions: (1) There is an indi viduality about a man's writing which distinguishes his work from that of other authors; we recognize a 'Virgil- ian quality' or a 'Miltonic quality.' (2) This literary trait is correlated with the character of the man him self; the Virgilian quality of style is the equivalent of some aspect of Virgil as he lived. (p. 230) 76 Now, it is illuminating to see the similarity between Puttenham's premises and those of one of the two main branches of contemporary stylistics. According to Ullmann, stylistics itself is not "a branch of linguistics" but "a parallel science which examines the same problems from a different point of view" (p. 10). And this different point of view may itself, according to Ullmann, be divided further into two different approaches to style. The first approach is that of investigating "the expressive qualities of style" and defines style itself as "the means of formulating our thoughts with the maximum effectiveness" (p. 2). This ap proach, as we shall see later, can be most fruitfully treated as a variant of the concept of style as choice, al though it does have some faulty assumptions in common with what Ullmann designates as the "second main branch of con temporary stylistics" and considers as starting "from an en tirely different point of view" (p. 25). The similarity be tween the premises as Ullmann describes them of this second branch and Abrams' description of those of Puttenham and eighteenth century students of style should serve to under score the continuity which exists between the earlier and the contemporary concept of style as symptom. This second branch, as Ullmann sees it, 77 limits its attention to the literary language and is mainly concerned with individual style. At the root of this approach there are two basic assumptions: (1) that there is such a thing as 1 individual style', a set of linguistic habits peculiar to a given writer; and (2) that this individual style is closely bound up with the writer's mind and experience and bears the stamp of his personality. (pp. 25-26) Both of these assumptions we have met before, because both of them are the foundation of the conception of style as symptom and are based themselves on the false notion that literature is expressive behavior, or, more simply, on the age-old confusion of written language with speech. And this confusion of two distinct orders of things— of utterances and of behavior which includes utterances— is implicit even in Puttenham*s premises. Note that Puttenham makes no dis tinction between speech and writing, rather lumps them to gether, so that his comments on style refer to and are exem plified by both. On the other hand, the crucial distinction between speech and writing was recognized as long ago as Aristotle's distinction between rhetoric and poetics. Aris totle did not share Plato's notorious contempt for writing in comparison with speech. As Gerald F. Else explains it in his commentary and translation of the Poetics, On the contrary, as we . . . see repeatedly in the Poetics itself, he insists that what the poet writes can be judged just as well as it is, through reading, as through hearing the words spoken. The poetic then, is something prior to and essentially independent of the voice; being recited is accidental to it. Aristotle was, in many, though not all, respects, the first and for many centuries the last of the objective theorists of literature, especially with this emphasis upon the text as prior to and independent of performer or reciter. Rhet oric and poetics, however, do touch, or, more precisely, the poet does use rhetoric for one purpose: for presenting the speeches of his characters.® But there is nothing in Aris totle about the poet's depiction of himself through his work. Rhetorical considerations are not relevant to the work as a whole but only to certain kinds of behavior of the characters depicted within the work. It seems to me a seri ous mistake to do as Wayne C. Booth, for example, does in The Rhetoric of Fiction: apply the categories and consider ations of rhetoric, particularly Aristotelian rhetoric, to 7 literary works. Booth is thus forced to ignore this 5 Aristotle's PoeticsThe Argument (Cambridge, Mass., 1957), p. 25. 6Cf. Else, pp. 265-270, 562-566. ^Chicago, 1961. Although Booth is a disciple of the Chicago Neo-Aristotelians, it is interesting to note that the mentor of this group, the Aristotelian scholar Richard McKeon, particularly emphasizes in his Introduction to Aris totle (New York, 1947), pp. 620-621, Aristotle's deliberate separation of the Rhetoric and the Poetics.__________________ ’------------------------------------------------------------------------------------------------------------------------------------------------------------------- — — -79 srucial distinction between utterances and acts of utter ance, between literature, even in the etymological sense of whatever is written, and speech, and to have to postulate mythical entities such as implied speakers with implied per sonalities. The rhetoric of Aristotle is still very defi nitely the craft of making speeches. The poet, too, "makes" speeches, but not speeches alone and not his own. The distinctions preserved by Aristotle were consider ably blurred nevertheless in Roman antiquity. In the Middle Ages, partly because the two most important rhetorical genres, the judicial and political, disappeared from politi cal reality with the extinction of the Greek city-states and Roman Republic, rhetoric according to Ernst Robert Curtius, lost its original meaning and purpose. Hence it pene trated into all literary genres. Its elaborately devel oped system became the common denominator of literature in general. This is the most influential development in the history of antique rhetoric.® Furthermore, the whole apparatus of rhetoric developed under the rubric of style— figures of speech and of thought, the three levels and the four virtues of style, and the entire Q notion of style as ornament— went through this adaptation ^European Literature and the Latin Middle Ages, trans. Willard R. Trask (New York, 1953), p. 70. ®.See. Curtius.,_p_._7JL. 80 to literature and continued in full force as a literary theory through the eighteenth century. Indeed, long before the eighteenth century, style, although only one of five parts into which rhetoric was traditionally divided, became the dominant concern of the rhetorician. "No other aspect of rhetoric or poetry in antiquity received so much atten tion in surviving treatises," D. L. Clark points out, and "by the late Middle Ages rhetoric had come to mean style alone." But the full import of this grafting of rhetorical jtheory onto works of literature was not really apparent until the advent of Romanticism. For up until approximately the beginning of the nineteenth century the conception of style as the ornament and dress of thought, as the "flowers of rhetoric" adorning the "matter and subject" of poesy, had firmly maintained the center of interest. But, as we have seen, the full implications of the rhetorical "theory" of l literature were already apparent in such a standard Renais sance handbook of rhetoric as Puttenham's. For if rhetori cal theory has to account, according to its traditional threefold sphere of application, for audience, subject, and "^Rhetoric in Greco-Roman Education (New York, 1957), 38, and Rhetoric and Poetry in the Renaissance (New York, 1922),, p. 47.___________________________________________________ 81 speaker, and if as Ben Jonson said, "Language most showeth a man," then where for a literary work is one to find the audience but in the reader, the matter but in the work as its "content," and the speaker but "in" the work as the source of its "form" or style? As the emphasis fell more and more upon originality and individuality, on fresh and unhackneyed diction rather than traditional matter and fig ures, the latent and previously uninfluential assumption that literature is the same as speech, which derived from the imposition of rhetoric on literature, emerged full-blown as the conception that literature is the expression of its author. Upon the basis of a rhetorical conception of lit erature what could be the relation of an author to his work but that of "speaker" to "spoken"? And if every man "ex presses" his true self when he speaks, if every man's speech ■is his shibboleth, then certainly literature must express its author and reveal, consciously or unconsciously, what most showeth him. Of course what a man expresses in his speech, as we lave seen, is rarely what he says but all that his looks, lis facial expression, dress, intonation, and circumstances express— in short, all that is evident about the act of ut terance. But no one since Aristotle has paid much attention 82 to the fact that literary works are not orations, and not until quite recently have even the crucial structural dif ferences between written and spoken language received their proper attention. The scientific study of language was seriously hampered until this distinction between writing and language was recognized and its implications fully ex plored. For linguistics the distinction was crucial insofar as it enabled the linguist to give speech its logical and genetic priority. For the study of literature, however, it has now become crucial to recognize that literary works are not acts of utterance and that to study their attributes as the attributes of the author is to treat them as if they were acts of utterance and thereby to prevent the under standing, enjoyment, and use of literature for what it is. The Romantics did not invent the idea that style is the man. Indeed, it is the eighteenth century naturalist Buffon who made that assertion famous, though, to be sure, he did not mean precisely what the Romantics made of it— that style is the attributes of the man. What he did mean was that style is the individual contribution of the author to the common property of knowledge and fact about which he chose to write.^ This was simply an ingenious variation of XI * _______Dis,co.urs_sur_le_s.ty.le (1.7-53_,_Paris_,_1.8.7.5)_,_p,._25_. c f-. 83 the traditional thesis that style is something added to the matter, but certainly it was easy enough to mistake "le style est l'homme iri^me" as the original contribution of the author for the contribution of the author as himself. The Romantics did not invent even the conception; it was already inherent in the medieval adaptation of rhetoric to litera ture. As far as literary theory went, the Romantics did not cast off the chains of a traditional but never adequate con ception of literature. Instead of propounding a new, non- ■rhetorical theory of literature, they emphasized an undevel oped aspect of the old conception. In line with their whole sale espousal of originality they found the uniqueness of a poem in the individual personality of its speaker. And in doing so, as we have demonstrated by examining the faulty basic assumption of the whole post-Aristotelian literary tradition, the Romantics were simply pushing to its ultimate conclusion one of the implications of the traditional theory tfhat was new was the fervor with which the Romantic poets and critics adopted this conception to the exclusion of all others and exploited it in its most extreme formulations. One of these extreme formulations was and still is that the also Wellek, A History of Modern Criticism, I, 63-64. 84 author's personality is the only real attribute of and source of interest in the work of literature.^ And the greatest complication of it has been the introduction of the idea of latency, which since Freud has of course taken most often the form of the concept of the unconscious. But, as we shall see, it need not. Almost anything may be proposed as latent, for there really is no way of proving it is not, nor indeed that it is. As may be recalled, Ullmann stated that the psychologi cal approach to style is based on two assumptions: one, that a given writer, no matter what he writes, exhibits a set of linguistic traits peculiar to himself, and, two, that this "individual style" is, to use Ullmann's words, "bound up" with the writer's life and personality. The proof of the first assumption, oddly enough, depends upon what one proposes as the meaning of the "bound up" of the second as sumption. For at no time has the first assumption been argued and demonstrated by itself; it has always been as sumed. That is, the investigator begins with a work or set of works he knows to be by one or the same author and then 12 Cf. Edmund Wilson: "The real elements, of course, of any work of fiction, are the elements of the author's per- sonality . . ." Axel's Castle (New York, 1936) , p. 176._____ 85 seeks to find the distinguishing traits of the author's "style." In those cases where authorship is unknown or dis puted or the possibility of interpolations has been proposed the existence of unique linguistic "traits" has never been satisfactorily proven, and language alone has itself never established the existence of interpolations or the identifi cation of an author. As Kayser declares, bei einem Werk, dessen Autor umstritten ist, hat es noch nie einwandfrei gelingen wollen, aus rein stilistischen Beobachtungen endgiiltige Schlusse auf die Autorschaft zu ziehen.^ And he substantiates this declaration with a number of ex amples of authorial collaborations which have never been successfully distinguished: the common work of Goethe and Schiller, the brothers Goncourt, and particularly the end less collaborations in Elizabethan drama, Beaumont-Fletcher, Massinger-Fletcher, and even Shakespeare-Fletcher, even though we have available for comparison any number of works by each of the authors alone. Furthermore, in cases of unknown authorship, even when the field has been narrowed down by numerous other means to two possibilities and even when elaborate statistical meth ods for determining only one linguistic element have been _____ Das sprachliche Kunstwerk, p. 287.______ _____________ 86 employed, nothing sufficiently conclusive has been estab lished. Starting with a study of the frequency distribu tions of nouns in the De Imitatione Christi, G. U. Yule tried by using this index to determine which of the two can didates for authorship, Thomas a Kempis or Jean Gerson, was 14 the likeliest. He was unable, however, to yield a deci sive index, although he had available other works of both authors to which to apply the same criteria. For the prob lem lies as always in the selection and evaluation of the particular "trait" to be used as an index and the possibil ity of finding samples which will be sufficiently comparable in length and subject matter to make the results decisive. Yule did feel that he might have found a statistic which would be independent of sample size— for homogeneous mate rial. But, as Yule modestly admits himself and as one sym pathetic commentator explained it, Its chief drawback appears to lie in its [the statis tic's] sensitivity to variations from work to work in the style of a single author, which in some cases is nearly as great as its sensitivity to the variations among different authors. D 14 The Statistical Study of Literary Vocabulary (Cam bridge, 1944). 15 barren Plath, "Mathematical Linguistics," in Chris tine Mohrmann, et al., eds., Trends in European and American Linguistics:_1930-1960-(Utrecht and Antwerp, 1961), p. 28. I 87 Even this statistic is applicable only to works "involving very similar ranges of subject matter." In addition, it cannot yet really be considered established because of the lack of a sufficient number of control calculations. And in any case it would only be valid for the number of nouns, while the decision as to what constitutes a noun will need to be based on more rigorous criteria— such as that avail able to structural linguistics— than have heretofore been applied even in Yule's experiments. Distinctions of style which literary scholars and crit ics make with great aplomb evaporate when touched by the cold dead hand of statistics. No author has yet been shown to yield "a set of linguistic habits peculiar" to him alone. Whatever stylistic criteria are or may be, they are not nor may they be solely linguistic. The number of nouns used is simply the number of nouns used, in a language that uses nouns. In order for it to be considered anything more than an instance of language meaningful as such, the linguistic datum must be related to, shown to indicate something beyond its occurrence and meaningfulness as an instance of language, whether this be the possible authorship of another work or something more elaborate, such as the author's personality, soul, Weltanschauung, etc. Thus the second assumption 88 stated by Ullmann and derived from the pronouncements of several of the major European literary scholars devoted to investigating style is the really basic one. Whether as a matter of logic or as a matter of evidence, there simply cannot be nor is there any such attribute as style nor any stylistic trait as distinguishable from facts of language, unless the facts of language are "bound up" with something else. The relating of the author's language to his mind, experience, personality, is in short, one kind of attempt to establish the existence of "style" and "stylistic facts." And the hypostatization of the "latent" or "unconscious" is one means of relating the language of a work to something— in this case and in a very special sense, to the author. The kind of literature which is easiest to interpret as "expressive" is of course lyric poetry. Here the frequently recurring unspecified "I" can readily be identified with the author whose name appears above the text, or the poem, though not mentioning an "I," resembles an utterance in such a way as to require the reader to infer a speaker, as we saw was the case with "0 Western Wind." Furthermore, the sub jects and themes of much lyric poetry are often the states of mind or immediate emotions of the unspecified speaker and can therefore easily be thought of as the emotion or state 89 of mind of the nearest identifiable person, the author. Finally, the poetic theory of several predominantly lyric poets has fostered the notion of poetry as a "spontaneous Joverflow of powerful feelings" on the part of the poet. And although Wordsworth, for example, very definitely specifies that this is a remembered emotion— "recollected in tranquil lity"— and not an "expression" of the moment at all, still it is easy to forget this qualification in the light of the first and more powerful statement, which Wordsworth thought well enough of to use twice in the same essay. It is inter esting to note that the rise of expressive theories of lit erature was in fact accompanied by a shifting of the center of the "poetical" from what had always been either tragedy or epic to lyric poetry. Abrams points out this significant fact repeatedly, although he does not recognize that the 'Romantic theory is, as we have tried to show, simply the pltimate development of a rhetorical conception of poetry in which the "speaker" is an attribute of the utterance. Abrams discusses John Stuart Mill's essays "What Is Poetry?" and "The Two Kinds of Poetry" (1833) as particularly exemplary evidence of this shift in the status of the literary genres."^ -^For his discussion of Mill see esp. pp. 23-26, 148- 149, and for "The Lyric as Poetic Norm." p p . 84-88. __________ 9 . 0 Such a shift was a necessary concommitant of the development of expressive theories because the lyric was the only genre which could be interpreted directly as expression. Mill, as well as John Keble, as Abrams also notes, separated "poetry from oratory on the grounds that a poet pours out his feel ings without reference to an audience" (p. 148, italics ■mine). So it seems that after rhetoric had dominated poetic theory for centuries, a poetic theory evolved which found it necessary to dispense with the audience— the major concern and justification of rhetoric itself and a major concern and justification of poetry in all post-Aristotelian "poetics." The justification for poetry as the means of delight or in struction or catharsis for the reader or audience gave way to a theory of poetry for the poet alone— not poetry for the sake of the audience or for the sake of the poem, but poetry for the poet1s sake. Nevertheless, a major portion of the world's litera ture, including a great deal that is indisputably its best, is not lyrical at all. In the case of epics or romances or novels the authorial or minstrel "I" might obviously inter polate his commentary on the action, but this rarely amounts to anything that can be taken as very revealing of the author. We may have our own opinions about people who interrupt their stories in order to moralize or to tell the ■reader not to worry because everything comes out all right in the end. But these are opinions about the author or min strel as storyteller pure and simple, while dramatists offer jus nothing directly except perhaps a stock prologue or epi logue. The amount and kind of information directly "ex pressed," even in the simple sense of "uttered," by authors in these genres is small and singularly unexciting and unre- vealing of any "true self." On the other hand, a metaphor ical use of the term "express" for these genres— that an author expresses himself through his characters and action— is enormously difficult in practice to apply. The differ ence between what a work means and what the author means or ^expresses, above and beyond the work, even in a metaphorical sense, is a distinction apparently almost impossible to make, if we are dealing with the work as literature. What I jthe author may be expressing through the actions and charac ters and, indeed, which characters might really be the spokesmen for his "true self" must remain in the realm of the strictly conjectural and unprovable. A work of litera ture says what it says, which in itself may in any given case be difficult enough to determine. But if we postulate that, in addition, the author is "saying" something himself, 92 the difficulties of interpreting this are enormous. We may obviate some of these difficulties of interpret ing the literary work as "saying" something of the author by averring that there is a latent meaning in the work, so la tent as even perhaps to be hidden from the author. We might even go so far as to say that this latent or unconscious meaning is the real one, that all which is obvious in the |vork is mere surface and disguise. We need not, of course, but if one will only admit the concept of latency, anything may be admitted under its aegis. The idea of the uncon scious, hidden, or latent stratum of the literary work need not be based on and, indeed, historically precedes the structural hypothesis of the mind proposed by Freud. It may be only the simple assumption, for instance, that the theme jof Paradise Lost is not what Milton declared it— the justi fication of the ways of God to man— but some other theme, such as the conflict in man's soul between the principles of 1 7 •reason and unreason. A. J. A. Waldock m his penetrating 1 p essay on Paradise Lost and Its Critics^0 takes to task scholars as far apart in apparent aims and methods as Edwin 1 7 Cf. Edwin Greenlaw, "A Better Teacher than Aquinas," Studies in Philology, 14:200-214, April 1917. 18C a m b rid ae . 1 9 4 7 . 93 Greenlaw and Maud Bodkin for their application to literature of the hypothesis of unstated or unconscious meanings. The desire to produce such "real meanings" may stem from the hope that the declining popularity of a work might be helped if a meaning more congenial to the current temper were found •for it, while contemporary Jungian and other versions of archetypal criticism seem to stem from a desire to add to the "truth-value" of literature by finding in it strata more profound than the obvious and specific meaning. Critics vary too in the literalness with which they apply such an assumption and the kind of proof they deem necessary to es tablish this latent meaning. E. M. W. Tillyard, for example asserts outright that the real meaning of a poem is the state of mind of the poet at the time he composed the poem and that, if we are ever really to understand the poem, we must bend every effort to reconstruct the specific histori- * 1 Q cal circumstances of that state. But this approach, even though it uses the concept of "state of mind," involves nothing really psychological. What Tillyard is actually in sisting on is the absolute necessity of exhaustive historical 19 . Tillyard and C. S. Lewis, The Personal Heresy; A Con troversy (London, 1934); and Tillyard's Milton (London, 1930) . For a rigorous critique of this position see Waldock pp. 11.9-1.2.9.,_________________________ 1 94 scholarship, the function of which is to find out- everything that the poet could have had in his mind at the time he com posed the poem, so that investigation leads, not to infer ring from the poem the psychological processes which "caused1 1 it, hut to studying the poet's biography, his geographical and political situation, the "climate of ideas" by which he must have been affected, his reading, etc. On the other hand, those whom Ullmann designates as ’ scholars mainly concerned with investigating individual style seek to discern the state of mind of the author on the basis of a completely a-historical conception: that the literary work is itself the datum and the only one for what must have been in the mind of the author when he composed ■it. Specifically how this state of mind can be discovered •in the work and precisely what sort of phenomenon it will prove to be— temperament, soul, psyche, Weltanschauung— will jvary with which of two somewhat different but indiscrimi nately mixed hypotheses is applied. The earlier and more psychological assumption, which we see echoed in Puttenham's kind of reasoning from humble style to humble man, is that the language of a man bears a physiognomic relationship to the man himself. The second assumption is that the author 'is discernible in the work as its creator and that by 95 determining the creative principle which must have motivated the artist one can discern his soul. This second assumption is based of course on the age-old analogy between God as the creator of the universe and the artist as the creator of the literary work. The work functions as a sort of dual symbol- system which points, on the one hand, as literature to the world by means of story, characters, scenes, etc., and on the other hand, to the act of creation and the existence of the state of mind, soul, psyche, Weltanschauung, id, etc., of the author. This latter concept is, to be sure, simply a more sophisticated development of the physiognomic hypoth esis by such theorists as the Romantic Friedrich Schlegel on the very theological analogy which we have noted and serves especially to explain how non-lyrical works can be expres sive of their authors. But particularly as a result of the application of Freudian principles to stylistics both hy potheses are often used indiscriminately even in a single i study to explain specifically how the by now "superficial" literary work is symptomatic of its author. Whether the creator or the man, style does reveal, so the assumption goes, the individual. CHAPTER V STYLE AS THE LATENT Particularly among Romance philologists the concept of style as the latent has become the principal, and sometimes the sole, basis of their definition of stylistics. As the Spanish scholar Amado Alonso defines it, el nombre estilxstica denuncia que se quiere llegar al conocimiento xntimo de una obra literaria o de un cre- ador de literatura por el estudio de su estilo. El principio en que se basa es que a toda particularidad idiomatica en el estilo corresponde una particularidad psxquica. Ya le adelanto que una mera lista de particu- laridades estilxsticas no nos hace conocer y gozar la xndole de una obra ni de un autor: los rasgos difer- entes tienen que componer una fisonomxa.1 In order to explain why stylistic particularities have to compose a physiognomy Alonso quotes the philologist who has by now come to be the outstanding representative of this mode of stylistic investigation, Leo Spitzer, Ha de haber . . . en el escritor una como armonxa pre- establecida entre la expresion verbal y el todo de la obra, una misteriosa correspondencia entre ambas. Nues- tro sistema de investigacion se basa por entero en ese axioma. (p. 96) • ^ Materia y forma en poesxa (Madrid, 1955) , pp. 95-96. 96 971 tfow Spitzer himself with great ingenuity in literally hun dreds of papers and monographs published over a period of ■fifty years has analyzed the "linguistic structure" of lit erary works in half-a-dozen different languages on the basis of the hypotheses described above. Would not the results of such a vast amount of investigation prove a decisive test as to whether or not there is such a relationship? Did Spitzer in fact find the "mysterious correspondence" which he as sumed must exist? Over the years Spitzer developed a technique of style analysis he called the "philological circle" by which he worked from the surface to the "inward life-center" of the work of art; first observing details about the superficial appearance of the particular work (and the "ideas" ex pressed by the poet are, also, only one of the superfi cial traits in a work of art); then, grouping these de tails and seeking to integrate them into a creative principle which may have been present in the soul of the artist; and, finally, making the return trip to all the other groups of observations in order to find whether the "inward form" one has tentatively constructed gives an account of the whole.2 Spitzer justified this technique as the basic procedure of all the humanities, linked it with the Zirkel im Verstehen of Dilthey, the tracing of etymologies, and the development 9 • ^Linguistics and Literary History: Essays m Stylistics (Princeton, 1948) , p. 19 (italics mine) . _____________________ 98 of the concept of language families. But it has come in for the major part of the criticism leveled at Spitzer. The complaint is that the method is not scientific but intuitive and its results unverifiable.^ If, however, there is a cre ative principle manifested in literary works, the procedure ought to be capable of being duplicated and the results ought to be capable of being verified by reference to' the works. Surely relating the parts of a work to each other and the whole is what we must always do in literary analysis. •It little matters if we talk about the creative principle, as it "may have been present in the soul of the artist" or even about the inward form and the superficial appearance, if we are working from the details of the work to a view of the whole work. Now Spitzer defended himself against his critics, and specifically Bruneau, by insisting that his method was sci entific. Yet in restating it in defense of its "scientific" nature he actually omitted the questionable phrases about the literary work— the superficial appearance, the inward Lorm, the creative principle in the mind of the author— and -3 JSee especially Charles Bruneau, "La stylxstique," Romance Philology, 5:1-14, August 1951. Cf. also Raphael Levy, "A New Credo of Stylistics," Symposium, 3:321-334, November 1949; and Jean Hytier, "La methode de M. L. Spitzer" Romanic Review, 41:42-59, February 1950.______________________ 99 offered instead what amounts to a technique of diagnosis, which he did in fact liken to the procedure of the physician. His method is now described as partir dans l1etude'du style d’une oeuvre, d'un auteur ou d'un epoque, d'un detail bien observe, ensuite en in- duire uruuvue d'ensemble hypothetique (d'ordre psycholo- gique), qui ensuite devra §tre contr8lee par d'autres observations de detail— au fond le procede de tout homme de science (etymologiste, physicien— et, depuis dei^lfrjn- memoriaux, du bon critique litteraire).4 What could be more logical, more systematic, more reason able, Spitzer asked, and, to be sure, as a scientific method it now sounds irreproachable. Men of science, of course, at least etymologists and physicians, are students of causes, of sources, the origins of words, the origins of symptoms, of diseases. Thus the style of a literary work now becomes a symptom of something— "d'ordre psychologique." For is it not inevitable that when we think to seek in a work for the "inward life-center" or "form" we are actually looking through the work at something else? Does a work of litera ture really consist of a surface and some concealed inner core? If the surface is all the traits of the work, what could the inner core be but its source, in one or another shape or form, the author? Indeed, in the essay "Linguistics 4"Les theories de la stylistique," Le francais moderne, 20;165-168, July 1952._________________________________________ , ro- 0 and Literary History," already cited, Spitzer seemed to be unable to decide what his method is designed to find. "Lan guage is only one outward crystallization of the 'inward form,1" he declared, "the life-blood of the poetic creation is everywhere the same, whether we tap the organism at 'lan guage' or 'ideas,' at 'plot' or at 'composition" (p. 18). ©ut what is this "inward form," this "life-blood" which "is everywhere the same"? In the same essay he designates it by half a dozen different and by no means synonymous labels: "soul of a particular writer," "individual style," "common spiritual etymon," "psychological root" (p. 11); "clue to the Weltanschauung" (p. 13); "psychogram" (p. 15); "inward life-center," "inward form," "creative principle which may have been present in the soul of the artist" (p. 19) ; and "ideological patterns" (p. 32). Finally, he concludes that since the author lends to an outward phenomenon of language an inner significance . . . the reader must seek to place himself in the creative center of the artist himself— and recreate the artistic organism. (p. 29) The "fisonomxa" of the work must after all be related some how to the author. That pre-established harmony between verbal expression and the whole of a work seems accounted fpr only when one returns to the "creative center of the artist himself." Otherwise, how could verbal expression and 101 work be distinguishable? Unless the verbal expression is •related to something else, it is. the work, and all there is of it. Yet the charge that Spitzer's method is unscientific does not really touch the basic issue. If Spitzer does demonstrate correspondences between the language of a work and the nature of the author, the means by which he himself was initially able to discern these correspondences is not so important. The basic issue, as Ullmann has realized, is that As long as the demonstration is conclusive it surely does not matter in what order the various steps were taken? the main point is that a link has been established be tween a stylistic peculiarity, its root in the author's psyche, and other manifestations of the same mental fac tor. The great merit of Spitzer's procedure is indeed that it has lifted stylistic facts out of isolation and has related them to other aspects of the writer's experi ence and activity. (p. 29) But what does Ullmann, indeed, what does Spitzer mean by re lated? Is the "stylistic fact" caused by the author's psy che or is it imitative of it, or is it imitative of the sub ject of the work, and if so, how do we know this? Spitzer claimed to demonstrate, for example, that a rhythmic pattern occurring in the writings of Diderot "was conditioned by a certain nervous temperament" grounded on the author's 102 5 "erotic ErlebnisYet he also claimed that this rhythmic .pattern is an effect which was achieved by Diderot, albeit "half-consciously" (p. 162), and which is undeniably an ef fect because it can be seen as appropriate to the subjects of the works in which it appears. On the one hand, Diderot's "style is an irruption of the physiological rhythm of speech into writing" (p. 166, Spitzer1s italics), and on the other hand, this "feverish staccato style was invented by" Diderot and made it possible for him to describe in a manner unequaled before him "the sexual mechanism" and "the bodily, physiological, mechanical side of thought," "a new achieve ment" (p. 168) . Now this rhythmic pattern of an Encyclo- .pedie article on the relationship between the sexes is "imi tating” the sexual rhythm; then it is imitating the sexual automatism of characters behaving automatically in sexual situations; now simply "automatic" behavior; and, finally Diderot is also able "to translate into language" rhythmi cally "the 'vibrations' of the imagination— though here, too, the sensuous origin, even the sexual origin, of the emotional pattern, may be discerned" (p. 151). Furthermore, that it is not the reverse process which has taken place The Style of Diderot," Linguistics and Literary His tory, pp. 135-191.______________________________________________ 103 Spitzer infers not only from the well-known sensualistic approach of Diderot's philosophy, but also from the more general consideration that, in linguistics, the concrete pre cedes the abstract: the "etymology" of a stylistic pat tern must be found in that situation which is closest to concrete, to sensuous reality . . . whatever of similar ity we may find in . . . less sensuous passages must be considered as a secondary application of patterns based primarily on the sexual. (p. 151) The specific quality of a style seems to be derived here from whatever was the most concrete subject the author wrote about; in Diderot's case this subject seems to have been sex. Therefore, so this line of reasoning appears to pro ceed, the origin, nature, and significance of the "rhythmic pattern" in the author's works must be sexual. Here is where the genuine circularity of Spitzer's method appears and of all methods which propose to relate the attributes of an author to his writing or vice versa. Consider what we know and what Spitzer, too, must have jknown about Diderot's writings when he sat down to read the I . . ■Encyclopedie article— on the relation of the sexes— m which he claims to have discovered the origin of Diderot's style. We know that Diderot was an author who wrote very few strictly literary works. We know that in one of the best known of these few works, La Reliqieuse, he depicted the compulsive behavior (sexual automatism) of a lesbian abbess. 104 We know that one of his early works, Les Bijoux indiscrets, was considered so indecent that he later disavowed it. We know that most of these few literary works are either plays or largely in dialogue, dialogue in which one of the charac ters is often a "moi." We know, too, that the major part of Diderot's writings were devoted to the Encyclop^die and to journalism in which he espoused a philosophy which— while not "sensualistic" (the wrong word)— was based on the sole validity of the senses. Finally, we know that a consider able part of the non-literary writing which he left behind also consists of an extensive correspondence with one of a series of mistresses. Diderot’s case hardly offers a para digm for penetrating the soul of an author through his style, regardless of the fact that he is a modern author, who "en joys the freedom permitted the conception of 'original genius,'" which is how Spitzer justifies this physiognomic approach. Spitzer claimed to be inferring, according to a strictly scientific procedure, a linguistic trait from the work or works in which it appeared and establishing it as a "stylistic trait" only by correlating it with other passages in the same work or other works by the same author. Yet not only has he not done this, it is quite obvious that what he has done is deduce this essentially Diderotian stylistic 105 trait not from Diderot's language, nor only from what he al ready knew about Diderot as a historical personage and that not so much as a writer of literature but as a philosophe and an amoreux, but from the very subject matter of the works themselves. No wonder that "in this writer, nervous system, philosophic system, and 'stylistic system' are ex ceptionally well-attuned" (p. 135). It is not surprising to learn, then, that Spitzer has shown that Peguy's style, for example, is derived from, or based upon, or related to his Bergsonism, or Jules Romains', author of La Vie unanime, to his Unanimism, or Rabelais' word formations of known roots with fantastic suffixes to a tension in him between the real and the unreal, Utopia (think of the Abbey of Theleme) and naturalism (Rabelais' fundamental details). ® Spitzer's article on Diderot is in troduced by the statement: "I had often been struck, in reading Diderot, by a rhythmic pattern in which I seemed to fhear the echo of Diderot's speaking voice . . (p. 135). Style again is the speaker in the work, and this speaker is ^Cf. "Zu Charles Peguys Stil," Stilstudien (Munich, 1928), II, 301-364: "Der Unanimismus Jules Romains im Spie gel seiner Sprache," Stilstudien, II, 208-300; and Die Wort- bildung als stilistisches Mittel exemplifiziert an Rabelais, Beihefte zur Zeitschrift fur romanische Philologie, No. 19 (Halle, 1910_K__________________________________________________ 106 the author himself. But Diderot's best known literary works are, we will remember, dialogues, and dialogues in which Diderot often appears as himself, while his non-literary works are often letters, including a great many personal ones. Yet Spitzer does not relate the particular tremolo which he hears in this "speaker's speech" to the author's actual speech, which would have been a reasonable, though not particularly significant, relation of linguistic facts "to other aspects of the writer's experience and activity." Instead, he relates it to what the author has all along been "speaking" about. The sexual quality which is attributed to the rhythmic pattern is really being derived from the sexual subject matter of the works. It may be true that Diderot was a particularly sensual man. Certainly he was interested in the subject, for, as we see, he frequently wrote about it. Spitzer has not established either a root or an ana logue of a "stylistic pattern" in the author's psyche; he has simply transformed the subject matter--sensuality— into a "stylistic trait"— a sensual rhythm. The content has been as it were "stylized," and not by the writer but in the per ceptions of the critic. Perhaps Spitzer already recognized in this very paper the real circularity of his procedure, for the passage in which he postulates Diderot's "erotic Erlebnis" as the source of this "rhythmic pattern" follows close upon the analysis pf a section of La Religieuse to which he found it necessary to attribute two rhythmical patterns, the one "mocking" the other (p. 150). Of course it is not that the rhythms mock each other, but that the sentences, or better still, the ac tions described in them, contradict or are incongruous with each other. And the automatism which he finds rhythmically reproduced in the passage is the automatism which is in fact being depicted in it. "In both examples above," he goes on to say, "the emotion has been one of sexual passion; our discussion so far may have given the impression that Diderot is given mainly to such descriptions" (p. 151). But then it is that he proposes that we recognize this rhythm as primar ily sexual in quality— because that is the most concrete "source" of it in Diderot's works and because in that case it can also be attributed to his still more concrete nervous system. For without the substantiation of the author's ner vous system Spitzer would have been without a means of dis tinguishing, of separating, the "rhythm" of the passage from what it was about. By relating it to soul or Weltanschauung or id or psyche of the author he is able to postulate as a constant, a quality or attribute which recurs and is there- 108 fore distinguishable from the text in which it appears. The sexual quality of the rhythm then has to be an attribute of the text and an attribute of the author; unless it is both, the "rhythm" is simply a linguistic feature, and linguistic features do not reproduce the meaning— they are the source of it. The existence of an attribute of style, a particular style, or a stylistic "trait"— depends on whether or not the linguistic feature— the particular sentence, paragraph, or work, the recurrence of specific words or certain parts of speech, etc.— can be made to indicate anything beyond it self, beyond its meaning as a particular instance of lan guage. If the stylistic "trait" is simply a "reproduction" of the meaning of each passage in which it "occurs," then it is not distinguishable from each meaning of each passage which it supposedly "reproduces." Like the term "express," which we discussed earlier, "style" is used as a relational term. In any given instance it means, points to, indicates something which includes but goes beyond the specific meaning of the language. For X, language or an instance of language, is proposed a specific quality or attribute of X' style or a style, Y. But if we say that a rose (X) is red (Y), we mean that the rose pos sesses the quality or attribute of redness, not that it is 109 redness. There are roses which are not red and there are red things which are not roses. If we say that every lit erary work has a particular meaning and also has style, if style is not the particular meaning, what is it? We could disregard the meaning of an utterance by analyzing it into, or rather abstracting, its linguistic units, putting all the words of the utterance in a list, which we would call its diction, and all the sentence patterns we have abstracted into another list, which we would call the syntax of the ut terance. Have we now isolated the style of the utterance? But these words and these sentence patterns occur in dozens of other utterances, which mean differently according to their particular combination. That is what we mean by words and sentence patterns; they are the abstractable elements of any utterance, the recurrent features which give us to know that the particular utterance functions by virtue of belong ing to a particular system of conventions known as language (le lanqage) and a particular language (la lanque). ^ The linguistic features— the aggregate of words and syntax by themselves— may be abstracted again and again from infi-: nitely diverse combinations of infinitely diverse meaning. Ferdinand de Saussure, Cours de linquistigue generale, (Paris, 1916; 4th ed., 1949), pp. 23-39. Cf. above, p. 20. We may call them style if we like, although we already have terms for them, syntax, grammar, words, or, if we make our abstractions finer, morphemes, phonemes, phones. Style is supposedly that and something more than that, however. It is supposedly a particular quality of a partic ular utterance or group of utterances (la parole) which may be co-extensive with but is not the same as their meaning; yet as we see, it is meaning which is the particularity of the particular utterance or utterances. Diderot's La Reli- gieuse, although it could be analyzed linguistically in terms of the language of eighteenth century France (la lanque), is a unique, non-recurring meaning, which derives from the unique, non-recurring combination of its "linguis tic features." When La Reliqieuse recurs, it recurs as the same work. Thus, if what we mean by "style" is the linguis tic features of the work, so that an account of style is simply a grammatical analysis, or again if what we mean by style is the ostensible subject, theme, or more inclusively the fullest and most adequate interpretation of the work, which we can consider as its meaning, then the term is re dundant and misleading. If we mean by style some attribute which inheres in the particular work or works, then the term must comprise more than the linguistic features, which of Ill themselves are not particulars but abstractions and some thing more than the individual interpretation to which the work is subject. Otherwise, style is simply the individual work or works as they are being analyzed or explicated. There is, indeed, one "mysterious correspondence between the verbal expression and the whole of a work"— it is often called the meaning. In order to avoid such circularity in the application of the concept of style the critic, in this case Spitzer, [proposed that the linguistic feature, in this case certain syntactic structures, is a stylistic feature, a rhythm which, while related to— "reproducing," "reinforcing"— the ■meaning which the text obviously has, is distinguishable from it as a symptom of the author, in this case of his "erotic Erlebnis." The whole question of "rhythm" is of course in itself highly complicated, especially in prose, for which we have neither the conventional practice of ■meter, nor the conventional signal of lines which do not ex tend entirely across the page, nor other conventions which often coordinate with meter, such as rhyme and alliteration. One well-known book on English prose rhythm defines rhythm as a series of units or elements or groups which are simi- lar not necessarily in themselves or necessarily in their 1121 duration, but the more alike they are in both character istics the more obvious is the rhythm, and the more un like they are in one characteristic or the other, pro vided the impression of similarity is maintained or in duced, the more interesting the rhythm is . . . for the impression of similarity the idea of expectancy is para mount. For if the expectancy is strong enough we may readily assimilate or organize into similarity things which are palpably unlike either in themselves or in duration or even in both.^ Certainly this tortuous definition would suggest that the concept of rhythm in prose is a highly complicated one. We see that the author concludes after all that expectancy is the decisive criterion, and expectancy does seem the likeli est determinant of whether or not the observer will perceive rhythm. Susanne Langer reminds us that it is we who hear in' the equal ticks of the clock the rhythm of the alternating "ticks" and "tocks."^ And surely when we read poetry, as ®Paull F. Baum, The Other Harmony of Prose (Durham, N.C., 1952), p. 24. Cf. also I. A. Richards, The Principles of Literary Criticism (London, 1926), pp. 134-146. ^Feeling and Form, p. 126. Cf. also Marshall Stearns, The Story of Jazz (Oxford, 1956), p. 13. "The outstanding student of the subject, Richard A. Waterman of Northwestern University, has a phrase for what it takes: 'metronome sense' ["African Influence on the Music of the Americas," in Acculturation in the Americas, ed. Sol Tax (Chicago, 1952), pp. 207-218]. "If your metronome sense is highly developed, you can feel a foundation rhythm when all you hear is a shower of accents being superimposed upon it. The story of the Congo natives being thrilled by the intermittent explo sions of a one-cylinder gasoline engine may well be true. Their highly conditioned ears supplied a rhythmic common denominator." _________ ______ ________________________________ 113 opposed to hearing it recited, we expect at least some re currence, if not a regular meter, in our own "reading," be cause of the conventions, such as those mentioned above, which are already in force when we recognize that we are reading a poem. No doubt rhythm, or more precisely, recur rence, had as meter and as recurring words and phrases a very important function as an aid to composition and memory when poetry went unrecorded and was handed on in a strictly oral tradition.'*'® On the other hand, "free verse" and other kinds of condensation and elimination of repetition seem reasonable developments now that rhythm is no longer func tional. No doubt recurrence will remain a frequent feature of poetry because it is enjoyable, like childhood nonsense rhymes. But it is not a defining feature of literature, as i the existence of a literate audience and written dissemina tion have revealed.11 In French, the language, we will re member, of Diderot, "rhythm" in poetry would be unthinkable 1®Cf. Albert Lord, The Singer of Tales (Cambridge, Mass., 1960) . 11It seems likely that the "free rhythms" which Benja min Hrushovski sees as constituting much of modern poetry and as needing analysis and description are, by the nature of their being free, not rhythms at all, at least not in any sense of recurrence, and therefore not analyzable like meter, which is recurrent and conventional. See Hrushovski, "On Free Rhythms in Modern Poetry," in Sebeok, ed., pp. 173-190. H4j without rhyme because there is so little difference in the 12 French language between accented and unaccented syllables. Thus the idea of a rhythmic French prose— if it does not rhyme, and Diderot's does not, of course— seems equally un thinkable. And thus one would have to conclude that at n o least in modern "unrhymed prose," whether French or Eng lish, rhythm if it means recurrence is not only not expected but, indeed, probably non-existent. One student of style has concluded that insofar as prose is concerned rhythm can only be a metaphor.^ Certainly Spitzer, as an erudite and imaginative lin guist in Romance languages, must have known all the problems involved with the concept of rhythm as a linguistic feature, 1 9 "Rhyme; Rime. Nature and functxon," xn Joseph T. Shipley, ed., Dictionary of World Literature (New York, 1953), p. 344. Cf. also William Beare, Latin Verse and European Song; A Study in Accent and Rhythm (London, 1957). l^Curtius points out that during the Middle Ages poetry was not recognized as an art separate from prose composition. There was no word for it, and the conventions were freely interchanged and combined, so that there existed such phe nomena as rhymed prose, texts in which verse and prose al ternate, etc. See European Literature and the Latin Middle Ages, pp. 151-154. This would further support my contention in Chapter II that the distinction between poetry and prose is one of convention— not definition. Poetry is whatever is presented as poetry. k. Wimsatt, Jr., The Prose Style of Samuel Johnson (New Haven, 1941) , p. 8 . _________________ _____________________ 115. particularly in French. But Spitzer as literary critic, looking for the substratum of the literary work buried be neath the language and ideas, plot and composition, could be struck in Diderot's prose by a "rhythmic pattern" which seemed to him "a self-accentuating rhythm, suggesting that the 'speaker' is swept away by a wave of passion which tends to flood all limits" (p. 135), while his very description of the "pattern" is a hypostatization of the subject, theme, meaning of the work in which it appears. At the same time he is being struck by this rhythm o_f passion the critic is reading about passion, and since Diderot wrote a good deal about passion, the critic attributes the rhythm to passion in Diderot himself. Diderot quite possibly was a sensual man, but his style is not symptomatic of this sensuality, this "erotic Erlebnis." The meaning, themes, subject matter of the works themselves are evidence of Diderot's "erotic Erlebnis," and almost the only evidence. Wellek notes that although Spitzer strongly emphasized the influence of Freud on his own work, his use of Freud seems simply "as a justi fication of a search for 'latency,' for a hidden key, a re current motif, a basic Erlebnis, and even the world view of an author."15 Yet there is no "latency" in a literary work, 15"Leo Spitzer (1887-1960)," Comparative Literature, ___ H e ] no hidden key. If there is a recurrent motif, a world view, it is right there for everyone to see; it is what the author is writing about. Peguy writes about his Bergsonism, Romains about his unanimism, Rabelais about the real and the unreal, Diderot about the erotic. If they did not, there would be no way of finding them in the authors' psyches, and we would have to relate the "stylistic facts" to what ever else these authors wrote about. For when we propose to be relating the attributes of a style to an author we are still actually relating them to the work, although stat ing the relation indirectly. This must always be the prob lem which arises when the critic thinks to find the author's state of mind through his works, whether the critic thinks to use the key of style or the elaborate historical recon structions proposed by Tillyard. The problem of conceiving of style as symptom is not that "the assumption of a necessary relationship between certain stylistic devices and certain states of mind would appear fallacious,""^ if we mean by this that "the whole re lationship between psyche and word is looser and more 12:310-334, Fall 1960. _____"^Wellek and Warren, Theory of Literature, p. 173. 117 oblique than is usually assumed" (p. 173). What makes it fallacious is that no relationship at all can be established between the psyche and the word, or perhaps we should say that the relationship between the psyche and the word is ab solute? that is, it is a "relationship" of identity. One does not demonstrate the existence of a psyche or a state of mind; one assumes that such a phenomenon exists because such things as works of literature exist. Works of literature are necessarily the products of human minds. The critic in fers from the existence of a literary work or works that there must have existed a mind of which this work or works are the product. Then in order to establish what must have been in the mind when it produced the work the critic postu lates the content of the work, sometimes of course under the rubric of style, in which case the attributes of the mind which produced the work are derived from the attributes of the style, the existence of which style is inferred from the assumption that the style of a work is due to the attributes of the mind which produced it. We see now how Puttenham could maintain both that style is the language in relation to the subject of the work and that at the same time style is also the language of the work in relation to the mind of the author. Of course the more a person speaks or writes the more he "sheweth the matter and disposition" of his mind. It is from his speaking or writ ing that we infer he has a mind, and it is from the content and quality of his speech and writing that we postulate the content and quality of his mind. The "mentis character" is a causal hypothesis like that of the Unmoved Mover or the First Cause, simply a postulate designed to halt the infi nite regress of causal explanation. The problem of style, in this phase at least, involves the same sort of reasoning as the problem of intention, which is a problem because, as Beardsley puts it, a general principle of philosophy "is often not kept steadily in mind." If two things are distinct, that is, if they are indeed two, and not one thing under two names (like the Vice President of the United States and the Presiding Officer of the Senate), then the evidence for the existence and nature of the one cannot be exactly the same as the evi dence for the existence and nature of the other. Any evidence, for example, that the Vice President is tall will automatically be evidence that the Presiding Offi cer of the Senate is tall, and vice versa. But evidence that the Vice President is tall will have no bearing on the height of the P r e s i d e n t . - ^ "This point is obscured," Beardsley goes on to show, "where two things, though distinct, are causally connected," as intention and art object are supposed to be. If the inten tion is in the work, then how can it be distinguished from 17 Aesthetics.,_p_ 1 , 8 . . 119 the work itself? Only when we have the artist's statement of intention outside of the work can we compare two things, and then we must still decide if in fact the intention is "in" the work insofar as the work fulfills the announced in tention . The same principle applies to style and the man, whether the "man" in the case is a psyche, an id, a Weltan schauung , or simply an individual "speaker." If the style of a work reveals the character of its author or speaker, and the character of its author or speaker is what consti tutes, what enables us to distinguish, the style of the work then we do not have two distinct things, but one thing— the work— under several nom de plumes, "style," "author," "speaker." Only if the attribute of the style exists also outside the work can it be distinguished from it, like the >rose and redness. Defining the style of a writer is not like diagnosing a disease, nor is it like tracing etymolo gies. Presumably a disease is explicable apart from the person who has it. If we are all ill, then none of us is ill; illness is no longer a state distinguishable from health, from existence. Presumably, too, words change, de velop from other words or from situations requiring new words, but etymological research presumes also that words I2C? are distinguishable from each other, that we can tell at a given point that we have a different word from the preceding one. Literary works are also distinct from one another; their language— words and combinations of words— are dis tinct. Otherwise they are the same work. Authors as human beings are distinct from their utterances, but their states of mind, intentions, psyches, Weltanschauungs are not; the utterances themselves are our evidence for the existence of all these "mental phenomena." A lightheaded man, as Putten- ham would say, whom he knew, might write what we would char acterize as a "lightheaded" work. But if this same light headed man wrote a work we thought grimly serious, we would change our minds, not about the grimness of the work, but about the lightheadedness of the man. Wellek and Warren complain that, for example, in the discussion of the baroque, most German scholars assume an inevitable correspondence between dense, obscure, twisted language and a turbulent, divided, and tormented soul. "But," they counter, "an obscure twisted style can certainly 18 be cultivated by craftsmen and technicians." Yet the counter here should not be an appeal to what the artist can or cannot do but to what is actually being claimed by these 1 8 _____ Theory of Literature, p. 173.__________________________ 121 scholars. They are simply characterizing works as twisted and obscure, and then reasoning: twisted and obscure works must have been produced by twisted and obscure souls. But what evidence do we have for such a correlation; we have no souls, only works. This fallacy is, precisely speaking, the strict form of petitio principii: the proposition to be in ferred is derived from itself or from an equivalent proposi tion. A typical illustration of this reasoning, often given in logic texts, is: "The order and design which we observe in nature prove that God created the world; for there could be no order and design in nature if God had not created the 19 world." And is this same theological argument not the source of that whole Romantic modification of the rhetorical theory of the speaker as the form of the work? One histor ian thus describes how, proceeding from the proportion, "as God is to His creation, so the great modern artist is to his literary creation," Schlegel could conclude— and it is evident that he did so — that just as God, despite his transcendence, is imma nent in the world, showing "the invisible things of him . . . by the things that are made," so also the typical ( modern writer . . . despite his transcendence of his works by virtue of his objectivity, is plainly immanent 19 In W. H. Werkmeister, An Introduction to Critical Thinking (Lincoln, Nebraska, rev, ed., 1957), p. 40._________ 122 in them and reveals his invisible presence by the things he has made. The reductio ad absurdum of this argument appears in Spit- zer1 s essay on "Linguistic Perspectivism in the Don Quijote,1 where, after showing that the novel "reveals" both in plot and language a continual shifting of perspective, he con cludes: . . . let us not be mistaken: the real protagonist of this novel is not Quijote, with his continual misrepre sentation of reality . . . the hero is Cervantes, the artist himself, who combines a critical and illusion- istic art according to his free will.... High above this world-wide cosmos of his making, in which hundreds of characters, situations, vistas, themes, plots, and subplots are merged, Cervantes' artistic self is en throned, an all-embracing creative self, Nature-like, Godlike, almighty, all-wise, all-good— and benign: this visibly omnipresent Maker reveals to us the secrets of his creation. . . .2^ 2®A. E. Lussky, Tieck's Romantic Irony (Chapel Hill, sf.C. , 1932), p. 69. Cf. also Abrams, pp. 235-241. For the •final redundancy cf. A. Alonso's rhythm of aesthetic pleas- pre in creation, "El artista nos transmite con su criatura ( una palida sombra del placer estetico que el va teniendo al lacerlo," Materia y forma en poesia, p. 105. 2- * - In Linguistics and Literary History, pp. 69, 72-73. 2f. Dorothy Van Ghent, The English Novel: Form and Function (New York, 1953), pp. 9-19, who is able to treat the per spectivism of Don Quixote without arriving at such conclu sions about the artist. See also Angel Flores and M. J. Benardete, eds., Cervantes Across the Centuries (New York, ;1947) . __________________________________________________________ CHAPTER V I STYLE AS THE INDIVIDUAL Leo Spitzer was, however, in many respects his own best critic. During the last decade of his life, the years fol lowing the publication of the essays on Diderot and Cervan tes, he gradually abandoned all attempts to study style or language or the author and concentrated on what had actually been the object of his attention all along, the literary work. In a review of Ullmann's book, which praises his es say on Diderot, he declared that this essay was his "last in the Freudian vein." Precisely the insight that "psychological stylistics" is not valid for earlier writers (Montaigne being one glar ing exception) has reinforced in me another tendency which was present in my work from the beginning, that of applying to works of literary art a structural method that seeks to define their unity without recourse to the personality of the author. •It is true, as Wellek points out^ and as this statement -*-(Rev. of Style in the French Novel) , Comparative Lit- arature, 10:368-371, Fall 1958. ^"Leo Spitzer," pp. 318-319. .123 124 indicates, that Spitzer remained ambivalent. Perhaps he never quite realized why "psychological stylistics" did not work, that what he had been doing all along was interpreting literary works, and that the reference of his interpreta tions to the author was what had really made his previous approach "circular," but he ceased to practice psychological approaches in general. In 1951 he delivered a telling blow against his own previous theories by refuting George Boas' contention that meaning in literature is problematic and that a "personal" poem such as Milton's sonnet, "Methought I saw my late espoused Saint," is not intelligible to anyone who has not had, as Boas puts it, "a blind-man's-experience- of-dreaming-of-his-dead-wife-whom-he-had-never-seen."^ Spitzer brilliantly demonstrates that this is not the mean ing of Milton's sonnet and that such biographical reference has distorted for Boas the actual and obvious interpretation of this poem, and he asks in summing up: . . . does Boas not confuse . . . the understanding of the whole empirical psyche of the poet (as he recon structs it) when he wrote the lines with the understand ing of the lines themselves . . . ? The appreciation of the work of art seems here to be drowned out by the 3"The Problem of Meaning in the Arts," in George P. Adams, et al., eds., Meaning and Interpretation, University of California Publications in Philosophy, Vol. 25 (Berkeley, 1950) , p. 319.____________________________ 125 interests of the psychologist interested in the biology of the artist. An autonomous and self-sufficient artis tic structure is subjected to an extraneous causal chain^ .This is not a formal renunciation of course, but it could certainly serve as one. Other evidence that the study of style strictly as an attribute of the language of the literary work and of the author cannot in practice be conducted is the phenomenon Wellek terms "the imperialism of modern stylistics."® This tendency to extend style to the entire work and, indeed, to the whole of literary study had appeared as early as 1915 in D. W. Rannie's little book, The Elements of Style: An Intro duction to Literary Criticism. "Style," Rannie announced, "is the essential part of literature" 7 therefore, "criticism of style is the essential part of literary criticism."® In his little book in 1942 on La poesia de San Juan de la Cruz iDamaso Alonso declared, El estilo es el unico objeto de la critica literaria. Y la mision verdadera de la historia de la literatura— esa lamentable necropolis de nombres y de fechas— 4"Understanding Milton," Essays on English and American ■Literature, ed. Anna Hatcher (Princeton, 1962), p. 130. 5,1 Leo Spitzer," p. 3 20. ®(London, 1915), pp. 8-9. consiste en diferenciar, valorar, concatenar y seriar los estilos particulares.7 And in 1950 Alonso repeated, "El estilo es el unico objeto S ^ i Q t de la investigacion cientifica de lo literario."° In this country the German emigre scholar Helmut Hatzfeld has been particularly vigorous in acclaiming and promoting the advent of what he calls "the new stylistics" and defines as "art- minded philology."® Hatzfeld has established a sort of school of stylistics at the Catholic University of America in Washington, D.C., which produces dissertations annually in the field, and he has compiled and is continually revis ing an extensive bibliography of stylistics applied to Romance literatures.-*-® Hatzfeld's definition of style is of course familiar enough: 7(Madrid, 1942; 3rd ed., 1958), p. 124. ^Poesia espanola: Ensayo de metodos y limites estilis- ticos (Madrid, 1950; 3rd ed., 1957), p. 482. ®"Stylistic Criticism as Art-Minded Philology," Yale French Studies, 2:62-70, Spring-Summer 1949. -*- ® A Critical Bibliography of the New Stylistics: Ap plied to the Romance Literatures, 1900-1952 (Chapel Hill, 'N.C., 1953). Significantly, Hatzfeld seems to have included in his bibliography all critical works, regardless of jwhether or not they use the term "stylistics" or even "style1 , ' .which deal with the text or "language" of the literary work. t Cf. his preface, p. xi._______________________________________ 127 . . . in any work of literature the author is expressing an attitude in a personal language within the general language, a procedure which everybody instinctively calls style. . . . This language, in the widest sense compris ing the whole structure of a work, is literary style. It is not surprising then to find that Hatzfeld finds "The conclusion drawn by Karl Vossler, Amado Alonso, Ulrich Leo, and others that style analysis coincides with literary crit icism in its objective form is, therefore, absolutely con vincing" (p. 63). If literary criticism is the interpreta tion of literary works and if style is the whole work, so the reasoning seems to go, then stylistics obviously coin cides with literary criticism, at least insofar as it is interpretation and not judgment. Yet one may ask, why call this literary interpretation stylistics? Use of the term "style" suggests, as I have at tempted to show, that the language of the work has attri butes above and beyond its meaning, that it does not coin cide with "the whole structure of the work." If it does, then it ought to be called "the whole structure of the work.' But Hatzfeld seems to wish to have it all ways, as the in clusiveness of his bibliography might indicate and the in clusiveness of his definition certainly does. For Hatzfeld _____J-1"Stylistic Criticism as Art-Minded Philology," p. 62. 128 has also defined style as the author's personal language within the general language. The author's language is of course the language of his works, too, the only evidence, in fact, that we usually have of his "language." The crux of the problem here lies, I think, in the question of the "per sonal" or as Damaso Alonso puts it in his definition, the "individual": "Para mi, estilo es todo lo que individualiza un ente literario: una obra, un escritor, una epoca, una 12 literatura." Even Wellek insists upon it. Although we may know hardly anything about the private individual Shakespeare, we do know, Wellek maintains, a great deal about his "poetic personality." We cannot dismiss this problem of personality in litera ture even where there is no biographical evidence of any kind. There is a quality which may be called "Shakes pearian" or "Miltonic" or "Keatsian" in the work of these authors, to be determined on the basis of the works them selves, although it may not be ascertainable in their re corded lives. There are, no doubt, connecting links, parallelisms, oblique resemblances between life and art. . . . Thus we can speak of "personal" style. Perhaps if we now pass in review some of the similar opinions about individual or personal style which we have encountered, and questioned, during the course of our 12La poesia de San Juan de la Cruz, p. 124. Cf. also Poesia espatlola, p. 482. _____-*-3"Closing statement," in Sebeok, ed., pp. 414-415.____ 129 discussion, we may be able to discern not only the basic as sumption that lies behind Wellek's contention, but also just how it is that life and art, in his sense of author and works, do resemble each other. To begin with we saw how contemporary psychologists, conceiving of style as an index to personality, confused the products of behavior, utter ances, with what their behavior or speaking says, that is, indicates about them. It became clear that, at least in the case of language and literature, products of human behavior could in themselves be indices only to the behavior— speak ing, writing, thinking— necessary to produce them, to behav ior, in short, which could only be inferred from the exist ence of its products. Since psychology is not our field, we did not pursue, although we did raise, the question of whe ther "style" was not in psychology a redundant term for what was already and traditionally distinguished as "personality" or "the individual." i Then, tracing the source of a typical Renaissance ■rhetorician's conception of style as both appropriate to the 'subject and indicative of the individual, we saw that it stemmed from the confusion in late Latin antiquity and the ytiddle Ages of writing with speech, owing to or at least co incident with the superimposition of rhetorical theory upon 130 literature. We saw then too how Romantic poetic theory was ■not actually a repudiation of traditional rhetoric but an exaggeration of its confusion between writing and speech, rhetoric and poetry, with the exaggeration taking more and more the form that, while the speaker was in the work, the audience could and should be dispensed with. That the older theory of rhetoric and the newer theory of expression were closely allied appeared quite evident in the easy misinter pretation of a typical earlier statement like Buffon's "Le style est l'homme meme." The idea of a common subject mat ter adorned by the individual writer easily became the idea of a common subject matter adorned with the individual. We found, however, that the more prominent the speaker or poet became in theory, the more difficult it was in prac tice to find an unquestionable symptom or index of him. His only "obvious" appearances were perhaps in an occasional lyric poem or as authorial commentary in a narrative. Since this included only a relatively small amount of literary works and only the most uninteresting parts of some of them, the critic in search of the style which was the man went underground. We saw that the assumptions of a Puttenham and of contemporary students of style remained identical, even despite the contemporary experimental evidence which has gone far to disprove or at least not to validate the assump tion that each individual has a set of identifiable "lin guistic habits." We saw that, on the contrary, the primary modern modification of the Puttenham thesis was an absolute exaggeration of it— the dichotomizing of the individual work into a top layer that is the "matter" and the real, the solid substratum or core that is the author. Thus the top layer, the actual literary work became merely a set of symp toms or indices to its character, the author. The whole work itself was now style— the style of the individual author who was himself the real work. Fortunately no doubt for literary criticism, we found that at this point a general intellectual "law" intervened: that when one is examining one object, even under two dif ferent names, one is still examining only one object. The contemporary student of style can no more than the contempo rary psychologist infer from a product of human behavior anything more than that there must have been a human being behaving to produce it. The qualities of mind and tempera ment, the innermost secrets of the buried psyche, the invol untary but inevitable manifestations of individuality which close study of the literary work was supposed to reveal, re vealed them indeed; they were the work, going as it were r 3 -2 , under another name, but there for all to see, if examined. 'No wonder stylistics came to be considered by these same scholars the sole science of literature. If stylistics con sists in differentiating, evaluating, arranging, and classi fying particular styles, and particular styles are whatever is individual about a work, and whatever is individual about a work is the work itself, then the history of literature is indeed the evaluation and classification of "the individual," but there is no need to call it style or this procedure stylistics. The concept of the individual, we find, is absolute, still functioning conceptually with the force of its origi nal meaning as that which is indivisible and thus distin guishable from a group. One statement differs from another statement if and because it is not the same statement. What is individual about a work or a writer or a period is that we can distinguish it from other works, other writers, other periods. Its individuality is not a matter of style but a matter of individuality, and this consists of everything which makes the work what it is. This individuality is_ its language, that is, the meaning of this particular instance of language, call it what you will— subject matter, content, theme, referent. Comparison and differentiation, we find, 133 are acts of the human mind, not something implicit in things themselves- Individuality and similarity are like rhythm, they depend upon expectation. We find them because we look for them. If we did not think that Shakespeare and not Marlowe wrote Richard II, we would not compare Richard II with Edward II for its respective "Shakespearian" as opposed to "Marlovian" qualities of language. Late Marlowe may be more similar to early Shakespeare than early Shakespeare is similar to late Shakespeare. Richard II and Edward II are two different plays; any other differences depend on differ ences of attribution— not of attributes. Two odes by Keats do not have some intrinsic Keatsian quality in common, al though they may be quite similar in many respects. What determines our attitudes in comparing them is that they have Keats in common or rather our knowledge that they were both written by Keats. The tautology of "individual style" becomes abundantly evident when one is forced, as Wellek is, to talk on the one hand about "uniform styles" and "prevalent styles" caused by imitation and vogue, and, on the other hand, about the vast differences in the works of one individual writing in dif ferent genres or about the diverse "styles" of an author such as Goethe who had a long and varied career. If Elizabethan dramatists "use a uniform style,is it not ultimately because they all wrote Elizabethan drama— to which they did not always sign their respective individual names? Conversely, what reason is there for grouping to gether, say, The Sorrows of Young Werther, Faust, and the Elective Affinities? An intrinsic Goethean quality? Or the fact that they were all avowedly written by the same author? Establishing, for example, Chaucer's authentic "canon" is extremely difficult because, while Chaucer may not have written like his contemporaries, his contemporaries and many of his successors wrote like him and apparently had no scru ples about contributing their own or someone else's work to his "canon." If the "style" of Paradise Lost appears to be uniquely "Miltonic," we might remember that not only do we know that Milton wrote it, we also know that no one else wrote a Biblical epic about Satan, Adam and Eve, God and Christ, the creation and expulsion, in blank verse in the English language of the seventeenth century. At the Indiana conference on style Robert Hoopes of fered the familiar argument that style is an individual quality of literary works, isolable in the works themselves, 14 Theory of Literature, pp. 170-171. Cf. "Closing Statement," in Sebeok, ed., pp. 415-417.______________________ 135 and that even selected passages reveal symptomatic differ ences, because works in the same genre but by different writers can be distinguished from one another and because students can identify the authors of passages cited in exam inations . If we were to place dramas, for example, by Shakespeare, Marlowe, Webster, Tourneur, Etherege, Congreve, and Ibsen before.a person who had not studied them, it would be quite possible to educate him to their stylistic dif ferences. He might bring himself to such an awareness without instruction. Teachers of literature must believe that "The style is the man," he declared, or else they would not request their stu dents to identify passages on examinations, . . . and not only passages that the students have read but others that they have not read from the same authors. The whole point is to see whether the students have de veloped some sensitivity to the particularities of the author's style. (p. 428) It is true that teachers do give students passages to identify, and it is true that students can identify them; it is also true that the works by the various dramatists ■mentioned could be distinguished from one another. But it does not follow that what readers and students are distin guishing, what they are developing a sensitivity to, are "the particularities of an author's style." Hoopes even l^In Sebeok, ed., p. 428. _ _ admitted that some authors, such as Pinero and Henry Arthur Jones, might be indistinguishable, but this is because they are so banal, "because they are so bad . . . they are not worth comparing" (pp. 429-430). The argument can readily be seen to fail here. And in fact students of literature and teachers alike have not developed a sensitivity to particu larities of style but a knowledge of particularities of lan guage and literature. Thus they will find it simple to dis tinguish Shakespeare and Ibsen, even if they have not read the works in question, because even if Ibsen were in English translation, his work and Shakespeare's would be neither in the same language nor in the same dramatic conventions. This would hold also for differences between Etherege or Congreve on the one hand and Ibsen (in English) or Shake speare on the other. Elizabethan English is not precisely the same language as that of the late seventeenth century or of the late nineteenth or early twentieth century (Ibsen, for example, in William Archer's translations), just as Elizabethan or post-Restoration dramaturgical conventions are not the same as the dramaturgical conventions of the nineteenth and twentieth centuries. On the other hand, it is quite likely that Etherege and Congreve would be con siderably more difficult both in language and dramaturgy to 137 distinguish from each other. Webster and Tourneur might well be indistinguishable, just as certain parts of Shake speare might well be indistinguishable from certain parts of Marlowe, unless of course we had read all the works before. For it is our knowledge of the possibilities, based on our familiarity with English linguistic and literary history which enables us to identify passages whose authorship is not specified. Differences in language and subject matter and convention will be immediately apparent, differences which we can coordinate with our knowledge of various writ ers and genres and when they flourished. A satire in heroic couplets offered for identification in a course in English literature immediately narrows the possibilities. But where two authors are working in the same language, that is, a specific language as it is spoken or written at a certain period in history, and in the same literary genre, for exam ple, the drama, as opposed to the lyric or the novel, and in i the same literary convention, such as the rhymed couplet or blank verse or prose, and dealing with extremely similar subjects, be it the witty rake, the avenger, the weak king, or the bourgeois family, it will be impossible to conclu sively distinguish them, as has been illustrated time and time again with the unsuccessful efforts to conclusively 138 separate collaborators or, for example, in the case of Chaucer, unidentified imitators. Now when it is admitted, as in the case of Pinero and Jones, that some authors, writing at the same time in the same genre and in the same language, cannot easily be dis tinguished, we need not conclude that it is because they are "bad." Authenticity, uniqueness, and merit are concepts which have been confused since the advent of Romanticism. A pseudo-Chaucerian poem may be just as fine a literary work as the work or works of Chaucer on which it was modeled. Indeed, owing to the lack of specific dating and identifica tion of much of Chaucer's work, we cannot be very sure that he is really the author of all that we now attribute to him. Concepts of identification, like individuality and style, slip so easily in literary study into criteria of evaluation. Good works should be good works regardless of whether or not they are imitative. Bourgeois dramas which were beginning to be written in the seventeenth century, such as Heywood's "A Woman Killed with Kindness," were considerably more un usual than another revenge tragedy by Tourneur or Webster. Yet they hardly survive as the better plays or better even than those by Pinero or Jones, while the bourgeois dramas of Ibsen are distinguishable from those of Pinero and Jones 139 only as other dramas. They are also better, but that is a mark of distinction, not a sign of distinctiveness. It is hard to see how they would differ stylistically, as opposed to totally. Can Pinero and Jones be distinguished from Galsworthy? We may prefer Galsworthy's plays, but that is because we like what is in them as opposed to what is in a play by Pinero. The development of the idea of style as the individual man and, at the same time, the exaltation of originality in literature did not coincide historically by accident. They are so intimately involved that, as we see, even a contempo rary student of literature cannot invoke the one without im mediately bringing up the other. Yet both conceptions are based entirely on matters external to literary works them selves. Neither refers to qualities or attributes in the work. Thus it happens that, even when the work is supposed [to derive its individuality from the "individualness" of its author, this individuality is actually being determined by its distinctiveness from the literary tradition and histori cal situation in which it appears. The first conception is, as we have seen, neither demonstrable nor applicable. But the second is only applicable on the basis of extensive his torical knowledge and always ends up being an evaluative 140 criterion, which in turn is entirely inadequate, for it would make Marlowe a greater writer than Shakespeare, Dorothy Richardson a greater writer than Tolstoy, and Finnegans Wake a greater work than Ulysses. Having no pred ecessors may be meritorious but having no successors is scarcely so. Although it guarantees individuality, it miti gates seriously against comprehension or enjoyment. Such a conception fosters the logical error that individuality is a quality in the work and obscures the fact that it can only be recognized or determined by comparisons outside, the fact that individuality is actually and only a historical con cept . "Individual" and "personal" are only a posteriori con cepts; they are labels, not qualities. This difference be tween classification and discovery, between description and expectation enters the problem of style at a variety of points. I have limited my attention in this work to the ap- i [plication of "style" to literary works, but this is not of course the only application of style to literature. The term is also used to distinguish and describe genres, peri ods, movements. But such generalizations usually proceed in a hierarchical fashion with the style of the work forming the primary, the pivotal point on which all other conclu 141 sions about style rest. "If we can describe the style of a work or of an author," Wellek has said, "there is no doubt that we can also describe the style of a group of works or a genre. . . . One can generalize even further and describe the style of a period or movement."^ The "if" of course is crucial. But the question of these larger generalizations of "style" is one too vast and complicated to be treated here; it is rather a part of the whole problem of literary classification and literary history. Briefly it must be noted, however, that such phrases as "the style of a genre," "the style of a period," "the style of a movement" may designate one of two entirely different concepts. On the one hand, the phrase, for example, "ba roque style," may designate collectively the entire array of characteristics which is meant by "baroque," so that "ba roque," "baroque period," and "baroque style" are for all practical purposes synonymous. In other words, "style"— with its modifier— in this case simply serves as a classi- ficatory term for a number of characteristics common to the majority of art works of a given date or kind in history. And these characteristics cover the entire range of the works— subject matter, conventions, ideas, language, etc. _____-^Theory of Literature, p. 174.__________________________ The phrase "baroque style" or "ballad style" or "Romantic style" may then seem somewhat tautological and misleading. But insofar as it is synonymous with "the ballad" or "the baroque" or "the Romantic" that is all the term "style" is in such a case— tautological and misleading. On the other hand, "style" plus a modifier may be used to designate a_ characteristic of the genre or period or movement which is exhibited in the works along with other characteristics such as subject matter, conventions, ideas, language, etc. In this case "style" supposedly classifies certain traits within the various works, and then the style of various works constitutes one of the characteristics of the genre, period or movement being delimited. When such a term as "baroque style" or "ballad style" or "Romantic style" is used in this second sense, it is subject to the same criti cisms that have been directed against the use of the term |"style" throughout this dissertation. What is it? Can it I be found, pointed to, in the given work? Can it actually be distinguished from the other characteristics supposedly com prising the baroque, the ballad, or Romanticism? This use of the term seems to be just another way of talking about the style of the work. There are no doubt, as Wellek asserted, "connecting 143 links, oblique resemblances between life and art" at the personal level. For as Puttenham had declared almost four centuries ago, "Men doo chuse their subjects according to the mettal of their minds." Nothing in the language of a literary work could be symptomatic of the author but what the work is about, its subject, its matter, its theme, its •meaning. Those who would analyze the language of literature under the rubric of individual style as if it were a surface scum along with all the other superficial traits, like plot and ideas, do not understand the nature of language: the nature of language is to mean, to be about something. What ever an author writes or says "reveals" whatever we can ever know about what was in his mind at the time he wrote or said it. Whatever the literary work may be about, it does not say two different things at once or say the same thing in two ways. Whatever it "reveals" about the mind of the au thor is whatever it means as a literary work. What the lan guage of the work means is what the work means, and what the work means, as far as we can ever know, is what the author ■means. The language of the work comprises its subject, its plot, its ideas. If something is not in the language, it is not in the work. Hamlet may have an Oedipus complex but Hamlet does not. And whether Shakespeare did or not no one 144 can ever know, for no one knows of any instances of Shake speare's behavior from which to infer that he had an Oedipus complex. No doubt he had a father. But his depiction of a character whose behavior provides grounds for the inference that the character "has" an Oedipus complex only reveals that Shakespeare, or as Freud would have had it, the Earl of Oxford, knew of and could depict a certain kind of behavior. All that the language of a work "reveals" of its author is the work. The work may be said to "reveal" what the au thor knows about and is interested in, if we assume, reason ably enough, that authors write about what they know and are interested in. But how they came to know what they know and why they are interested in what they are interested in is another and considerably more difficult question to answer. Given the existence of a number of works known to be by one person and given extensive detailed information about that (person aside from his works and given works of sufficient i ^similarity to indicate recurrent interests and knowledge, we can perhaps make some correlations between the two, and these parallels may corroborate our speculations about the author. But such correlations will tell us nothing about the works as literature. Biographical data may be able to explain the causes and sources of an interest or knowledge 145 on the part of the author for which the works are evidence of the existence of such an interest or knowledge. But we must have been able to understand, to interpret the work before we can possibly say what the meaning of the work " re veals" about the author— what must have been in his mind when he wrote it— for what was in his mind was the work, the impetus to write it, whatever he knew that he put into it. The subjects and themes authors do write about do reveal something about them, that they were sufficiently interested in and knew enough about something to write about it. We are not surprised to learn that Conrad had been to sea or that Lawrence had a possessive and ambitious mother. We are surprised to learn, however, that Conrad never lived or even traveled in South America, although that is the detailed setting of his longest and finest novel. We might be sur prised too, if we did not already know before we began to read The Rainbow or The Fox that Lawrence was not a woman, yet Shakespeare was not a woman either, though Lady Macbeth, Cordelia, Cleopatra are, and by the same token neither was he Hamlet, though he was a man. CHAPTER V I I STYLE AS THE IMPLICIT SPEAKER The conception of the relationship between author and work stated in the previous chapter immediately raises the problem of meaning. It is indeed a problem which we can scarcely avoid in talking of literature. Literature is lan guage, whatever else it might be as well, and the nature of language is to mean, whatever that might be. What we decide about meaning in language will thus considerably affect, it would seem, what we decide about meaning in literature. Though it is a favorite comparison with critics, aesthe- ticians, and linguists, language is not a medium;^ it is not like stone or oils or wood or even "C" sharp or red pigment. It is not a neutral material out of which we carve our sta tue or paint our picture; it does not in fact exist apart ^Cf. not only avowed Aristotelians like Elder Olson, "William Empson, Contemporary Criticism, and Poetic Diction,' Critics and Criticism, p. 69; but also Wellek and Warren, Theory of Literature, p. 163; Susanne K. Langer, Problems of Art (New York, 1957), p. 148; and Charles F. Hockett, A Course in Modern Linguistics (New York, 1958), p. 553. 146 147 from its use, whether that use is in a work of literature, a road sign, a matchbook cover, or the U. S. Constitution. To be sure, language exists physically as sounds and some times as marks on stone or paper or magnetic recording tape, but there are sounds and marks which are not language. Lan guage is not just sounds and marks but a systematic, conven tional use of sounds and marks which mean. The problem of meaning, however, is actually a basic, perhaps the only, philosophical problem, and one which seems as far as ever from being solved. That language is meaningful we all know, or at least we all act as if we think it is. This meaning fulness is not simply referentiality, for so many instances of language are not referential at all, instances such as commands and instances such as literary works. Must liter ary studies then wait upon philosophy to solve the problem of meaning? Are we prevented from saying what the meaning of a literary work is because no one has quite agreed upon what meaning is? On the contrary, the very activities of criticism and scholarship presuppose, as Spitzer so con vincingly argued in his refutation of Boas described above, that literature is meaningful and can be interpreted as such, that a thorough, fully adequate interpretation of every text is possible, can be agreed upon, and constitutes 1^8 its meaning. The precise way in which we get from those little sounds or marks on paper to understanding, comprehen sion, communication of and with the world around us may never be determined. But we can and do read literary works and quarrel and agree about their meaning, because literary works consist of language, of particular and non-recurring statements in a particular system or convention of language. And these statements, although they cannot be said to refer directly to some state of affairs in the physical world, present events analogous to human experience and are thereby comprehensible to human beings. Given a text, a knowledge of the language in which it was composed, be that Hungarian or the English of the eighteenth century, and a common stock of. human experiences, the scholar or critic not only under stands but also interprets and explains the literary work. And that interpretation or explanation which appears to be the most accurate and adequate in relation to the text which is its object is what I have meant and shall continue to ■mean in this work by the meaning. Indeed, it might be the case that the scholars to whom we have referred in the course of this work have so often concluded that stylistics is the sole science of literature simply because what they mean by "style" is what I have 149 called the particular meaning or interpretation of the work of literature. In that case stylistics certainly would co incide with at least a large part of what is considered lit erary history and criticism. Although "style" would seem a rather injudicious choice of a term, leading to considerable confusion with older or different senses, it would then scarcely be an erroneous conception. I might contend that use of the term "style" to designate what I call the inter pretation or meaning of literary works obscures the nature of such studies. But I could not contend that literature itself was being mistaken and misunderstood for something else or that qualities which it could not possess were being attributed to it. Nevertheless, that style is the meaning of a work of literature does not seem to be what these scholars have meant, for they have not said so and their critical practice has not revealed it. On the contrary, my whole purpose has been to show that what literary scholars have shown themselves to mean by style is something differ ent from, though perhaps co-extensive with, the meaning of the work, that for them style functions as a relational term, and that what it relates to the meaning of the work is its author. Since students of style conceive of it as some thing more than the meaning or interpretation of the 150 literary work, there seems nothing else for it to be but the author. The whole issue of style as ornament raised in the nineteenth century against the previous rhetorical theory has been in many respects a false one. The older rhetorical theory sinned not so much in proposing a distinction between form and content as in being unable to offer anything which could be the form but the author. If we propose that every literary work may be divided into form and content, we have to say what the form as opposed to the content is. We may say that the content is the common property of subject mat ter, as Buffon did, to which the author contributes his style, which is his particular version of the subject, but then that is simply saying that the particular style or form is the particular work as opposed to some other work on the subject. That may not be a felicitous way of putting the notion that the style is the meaning of that particular work as opposed to some other work on the same subject, but it is ■not crucial. The problem with such a conception arises, of course, when one ceases to talk about style in natural his tory and talks about it in literature, literature which not only may not be about "a common subject matter," such as natural history or even the fall of Troy, but also may not 151 'be about anything which has been treated in literature be fore or anything to which one can refer for external verifi cation. The exaltation of originality in poetry, coupled with the recognition that poetry does not seem to be about things in precisely the same way as natural history, per sonal letters, philosophy, sermons, geometry, or tracts are about things makes the distinction between form and content even more difficult to specify. If there is neither a com mon subject matter nor a verifiable referent, and if the en tire work is recognized as an original and non-recurrent object, how can it be distinguished into form and content? The only way that has yet been devised is by distinguishing the author in the work as its form and what the work says as its content, by converting the work, in short, into an act of utterance. It may be wondered, of course, why it seems necessary to critics to preserve a distinction between form and con tent, and one often hears it declared now that form and con tent are inseparable. This inseparability of form and con tent, however, one finds if he looks more closely at such declarations, is only contingent, contingent on whether or not the literary work is "successful." It is only necessary to talk about the inseparability of things or qualities when 152 it is thought that on occasion they can be or are separated. Such declarations in fact usually come to something like T. S. Eliot's statement that, "In the perfect poet form and content fit and are the same thing; it is always true to say that form and content are the same thing, and always true to say that they are different things."^ In practice it be comes necessary to maintain, if there is to be any practical criticism, that form and content may be separated, in order that the work may be discussed and in order that it may be evaluated in terms of whether or not the form "cooperates" with or is "harmonious" with or "achieves a union" with the content. For it is feared that if one really identifies form and content, or their synonyms, as Croce actually did, criticism will cease to be possible. Wellek, for example, rejects Croce's identifications on these very grounds and insists instead that It now seems clear that process and work, form and con tent, expression and style, must be kept apart, provision ally and in precarious suspense, till the final unity; only thus are possible the whole translation and ration alization which constitute the process of criticism.-^ This contention recalls earlier Romantic attempts to heal 2Introd.f Ezra Pound, Selected Poems (London, 1928), ,p. x (italics mine) . Theory of Literature, p. 174. the breach between language and thought, style and truth, implicit in the classical notion of style as ornament, by distinguishing between "a mechanical device of style" and a sincere expression.4 Yet the crucial questions are not whether style is ornament or whether or not form and content, style and expression, process and work are separable, but that if we maintain that they are, what are they, how are they distinguished, can we point to them, can we explain them? Inevitably students of literature who want to deal with literary works but also want to deal with something more than their particular meaning or interpretation end up talking about the "speaker" or author. And no scholars and critics are more subject to this consequence than those who think they are talking about the language of the work. We have already seen this occur time and again. Every literary theory which divides the work into a part which is the meaning (or content) and a part which is the style (or form) ends up talking about speakers and authors. If style is not the meaning of the work but is still something in it, 4Cf. Abrams's discussion of Wordsworth's attempt, esp. pp. 290-291; also Wellek’s discussion and evaluation of A. W. Schlegel1s position on the inseparability of form and content, A History of Modern Criticism, II, 49; and Spitzer on when the term "style" would be meaningless, in a review of Herbert Seidler's Allgemeine Stilistik, Comparative Lit erature, 81146-147, Spring 1956.______________________________ 154 there is nothing else for it to be but the "speaker" or au thor. And this is precisely the case even with those crit ics who contend that style jjs meaning. Wimsatt has declared in the chapter on style as meaning appended to his study of The Prose Style of Samuel Johnson that It is hardly necessary to adduce proof that the doctrine of identity of style and meaning is today firmly estab lished. This doctrine is, I take it, one from which a modern theorist hardly can escape, or hardly wishes to. (p. 2) This may be all very well, but does it solve the problem of style, or does it not simply remove it one step backward? If style and meaning are identical, then the question be comes what is meaning. And, should we decide upon meaning, there would still be the question of the need for a term like "style." It is to be at once suspected that whatever Wimsatt means by meaning he does not mean the particular interpretation of the literary work, nor that he means that style and meaning are really identical, at least not in the logical sense of being indistinguishable from one another. And so it turns out when he attempts to describe "bad style." "Bad style is not a deviation of words from meaning, but a deviation of meaning from meaning. . . . Of the actually conveyed meaning (what a reader receives) from the meaning an author intended or ought to have intended" (p. 10). 155 Once this is admitted it follows quite readily that "The question what the author ought to have said is the true dif ficulty in judging style." It is the only difficulty, for it is the only question, and it is one we implicitly answer every time we judge style. We do it by our sense, more or less definite, of what the author intends to say as a whole, of his cen tral and presiding purpose. (pp. 10-11) Four years, however, after writing his book on Johnson's style, Wimsatt collaborated with Monroe C. Beardsley in an article on "The Intentional Fallacy,"’ ’ in which they stated: One must ask how a critic expects to get an answer to the question about intention. How is he to find out what the poet tried to do? If the poet succeeded in do ing it, then the poem itself shows what he was trying to do. And if the poet did not succeed, then the poem is not adequate evidence of an intention that did not become effective in the poem. . . . A poem can be only through its meaning— since its medium is words— yet it is, in the sense that we have no excuse for inquiring what part is intended or meant. (p. 4) Wimsatt continues to discuss and define style or styles, but at the same time now trying to avoid the intentional fallacy. And his principal method of doing so, as we might have guessed by now, is by converting the poem into an act and the poet into a "dramatic speaker." And he accomplishes this by resorting to Aristotle, and, as might be expected, ^Originally published in the Sewanee Review, 14:468- 488, Summer 1946, but reprinted in Wimsatt's The Verbal Icon (Lexington, Ky., 1954), pp. 3-18.____________________________ 156 ■not to his Poetics but to his Rhetoric. ^ Aristotle conceives verbal discourse as an act, compli cated in itself, and having a personal context of two main dimensions, the speaker and the audience. But he looks on all the features of the verbal act in a prag matic light. . . . That is what rhetoric in the full classical sense means, a pragmatic act of discourse. But there is nothing in the nature of the verbal act to prevent us from looking at the same features not in a pragmatic, but in a dramatic, light, and if we do this we are looking at a given discourse as a literary work. (p- xv) yet how is this speaker that is supposed to be "assimilated into the implicit structure of the poem's meaning" (p. xvi) to be distinguished from the author? For not only impli citly is the poem an act, but also it is an act most expli citly. The poem conceived as a thing in between the poet and the audience is of course an abstraction. The poem is an act. The only substantive entities are the poet and the audience. But if we are to lay hold of the poetic act to comprehend and evaluate it, and if it is to pass current as a critical object, it must be hypostatized. (p. xvii) What indeed is being hypostatized here? Surely it is the poet and his "audience," and surely it is naive to continue to maintain in a literate culture that literary works are acts rather than products of acts. A literary work is no more an act than is a railroad timetable or a "no passing" sign. We infer, to be sure, ^The Verbal Icon, pp. xiv-xvii. ___________ 157 from all three that they must have been produced somewhere along the line by human beings in an act which is uniquely human, that of producing language. Not even IBM machines can compose poems, although they may be able to print, en code, or count them. The question of whether or not the poem, the timetable, or the "no passing" sign continues to exist while no one is reading it could lead us, of course, into some heady metaphysical speculations. But if we have to hypostatize the poem anyway, we might just as well do so and leave the speaker and the audience out of it, since they cannot really be said to be in the poem until we have hypos tatized it and then assimilated them into its implicit structure. A literary work is a statement. Naturally it is not a statement like a bank statement where there is, or is supposed to be, a one-to-one correspondence between certain signs and certain objects, but it is language, and it is re cordable, and if its language is not preserved, whether in the memory or on paper, stone, clay tablets, tape, or disc, neither is it. A literary work presents events, events which can be interpreted, ignored, disliked, rejected, understood, or misunderstood, as analogous, to a greater or lesser degree, to the human experience within the comprehen sion of its readers. But the literary work itself is not an act; it is a product of an act. It may be about actions and sometimes presents acts, and yet in themselves so are a great many other things. We know that decorations on a pot are the result of a human being's behavior, even though the process of producing the pot and its decorations may have been mechanized considerably, because nothing in the non human world designs pots and decorates them. But scarcely anyone maintains that we find potters implicit in their 'pots. We may find their workmanship, a design they have used before, or even their "trademark"— but not them. But it is actually Wimsatt's collaborator, Beardsley, who has worked out most rigorously the implications of the concept of style as meaning or what he calls "the Semantic Definition of Style," which is a corollary to his "Semantic Definition of Literature." By following his presentation of this definition we can see most vividly how the conception of style as meaning leads inevitably to the conception of style as symptom and literature as act of utterance. In his Aesthetics Beardsley presents the fullest statement of this 7 semantic definition of style, which he develops in three versions. Each version apparently incorporates the ^The following discussion will be based primarily on pp. 114-164, 220-266.__________________________________________ .preceding one until we arrive at the final definition. In order to understand this final definition, then, we must follow its development in each of the versions, for the final version is supposedly not clear without a knowledge of the first and second. The first tentative definition is: "The style of a literary work consists of the recurrent fea tures of its texture of meaning" (p. 222). Naturally, it is first necessary to know what the texture of a literary work is. The texture of meaning is "the meanings of certain paragraphs, sentences, phrases, or words" in contrast to the structure of meaning, which consists of the "meanings that depend upon, or are a function of, the whole discourse, or a large section of it" (p. 221). There are, however, not just textures and structures of meaning, but also two kinds of textures, phonetic and semantic, or the sound-texture and the meaning-texture. These two need not coincide, says Beardsley, but in literature they generally do. Furthermore, this sound-texture is divided into two kinds, but these can be covered later. Already there appears to be something disturbing about this definition. It is not really so suc cinct as it appears, and this is, indeed, partly what is wrong with it: there are so many distinctions, and they seem to grow in direct proportion to the number of things 160' to be defined. Perhaps it is not really a definition at all, since it furnishes no defining characteristic but only characteristics; that is, perhaps it is actually a descrip tion. But at any rate, it certainly does not function ac cording to the Principle of Occam's Razor, which Beardsley discusses elsewhere in his Aesthetics. Entities certainly have been multiplied here and, I think, unnecessarily. What virtue is there, for example, in the distinction between structural and textural meaning? Beardsley admits that the grounds for this distinction may not be entirely justified, but he continues with it because he finds it "illuminating" (p. 221). This so-called structural meaning, however, inevitably consists of the entire so-called tex tural meaning. If there is not this explicit relationship between the parts— even words and phrases, not to mention sentences and paragraphs— and the whole, there is not likely to be any whole. The whole of a literary work, indeed of anything, is whatever its parts and their relationship to each other constitute. To discuss the meaning of a word or phrase without relating it to its context, the sentence, anc to discuss the sentence without relating it to the para graph, and the paragraph without relating it to the chapter, and the chapter without relating it to "the whole discourse,1 . ' 161 would surely be to invite serious misunderstanding of each element and thus inevitably of the whole. Literary works do not have some one-to-one, word-to-object correlation with a verifiable and referential physical reality. Nor would Beardsley maintain that they do. Therefore, the relation ship of part to part and part to the whole statement be comes in literary works the crux of understanding and inter pretation. If one wants to talk about the way words or phrases are being conventionally used in the language in which the work is written, one needs no particular context larger than a single sentence. But if one wants to do more than that, in short, if one wants to know what the particu lar instance of language which is this particular literary work means, then he has to interpret this particular mean ing from every part of the work as it functions in the work, and thus the meaning of the particular whole as it is de rived from the relation of the particular parts. If he does not do this, quite likely he is not analyzing the work but only language samples which happen to have appeared in a literary work. The meaning of a literary work is not the sum of its parts but the complex relationship of the parts which constitutes the whole. Only by seeing the parts of a literary work in relation to other parts and to the whole 162 can we understand that the work is a particular statement of a particular event and thus that it is meaningful— not ref erent ially— but analogously. But this procedure, it might be objected, is explica- Q tion; Beardsley is defining style. And indeed this objec tion leads us to Beardsley's second set of unnecessarily multiplied entities. What is the difference between the study of style as the study of textural meanings and expli cation? "To explicate a linguistic expression," says Beardsley, "is to declare its meaning" (p. 129). Thus, both stylistics and explication deal with meaning. Furthermore, the meanings they both deal with are textural: "A critic is explicating when he talks about relatively localized parts of a poem, the meaning of a metaphor, the connotations of a word, the implications of a fragment of ambiguous syntax" ®The two terms are frequently interchanged, especially in European scholarship, where they coincide historically and where the results of employing one "method" as opposed to the other are indistinguishable. Such coincidence of terminology and practice further suggests that there is sim ply no applicable distinction between stylistics and the interpretation of literary texts which is generally consid ered the primary function of literary criticism. Cf. Leo Spitzer, Essays on English and American Literature, pp. 193- 247, for an example of interchangeable use of the terms; Erich Auerbach, Introduction aux Etudes de philologie ro man e (Frankfurt am Main, 1949), pp. 33-37; Helmut Hatzfeld, "Stylistic Criticism as Art-Minded Philology," for the his- torical and practical relationships. _____________________ 163 (p. 130). These "localized parts" actually appear to coin cide with the texture of meaning which Beardsley has dele gated to stylistics. "The features of discourse that make up style are conveniently divided into two parts, diction and syntax (p. 225). Stylistics is co-extensive with the study of figures, and as such is the study of "any devia tion" of a discourse "from what is taken to be the norm of diction and syntactical construction"; that is, it is "the study of all identifiable textural features of a discourse" (p. 222). Beardsley includes under this study metaphors, connotations, and all other aspects of diction and, natur ally, syntax, ambiguous or otherwise. Thus, both stylistics and explication deal with the meanings of localized parts or textures, that is, with diction and syntax, and furthermore with deviations of such--ambiguities, obscurities, etc., .purposeful or not (cf. pp. 226-227). The only discernible differences between the two, although Beardsley never dis cusses their differences or similarities at all, are (1) that stylistics treats the recurrent features, while expli cation treats one feature at a time; and (2) that stylistics treats these as they appear in prose, while explication deals with poetry. But the first distinction of a "local ized meaning" has to take into consideration whether or not I64j it recurs and what its recurrence "means." Furthermore, how do we decide when a given feature ceases being 1 1 localized" and can be considered recurrent, i.e. how can we identify it, and how can we decide that its recurrence is signifi cant, much less what it signifies? Apparently the answer is, according to the criterion of deviation. Yet a good many features which recur in a work are not necessarily devia tions, e.g. sentence structures, words, phrases, etc. In deed, the whole problem is how we can decide that some fea ture is a deviation from the norm— what is the norm? We shall come back to that problem shortly. But the point to note here is that the criterion of recurrence is a purely arbitrary distinction. And this leads to a second objection to Beardsley's multiplied categories: they are purely arbi trary. What is the virtue in having one term for the study of meanings of a prose literary work, style study; and another term for the study of the meanings of a poetic literary work, explication? If one defines literature, as Beardsley does, according to the criteria that: "A literary work is a discourse in which an important part of the meaning is im plicit" (p. 126) and that this is what certain prose works and poems which are called literature have in common, then 165 there seems to be no virtue. And if it is a question of poems having a greater ratio of implicit meanings than prose— more figures, more ambiguous syntactical construc tions, and other such "deviations"— there seems to be even less virtue in this distinction. For that is simply not the case. Joyce's Ulysses and Virginia Woolf's To the Light house have much greater ratios of "deviations" than "O West ern Wind." In fact, as Beardsley himself admits, "O Western Wind," among other "nonfigurative poems" as he calls them, does not have any (p. 154). These confusions well exemplify those which arise when one tries to define literature ac cording to degrees or levels of meaning. Where does one category, that of non-literary language, end and another, that of literary language, begin? Beardsley himself admits that it is not easy to decide (p. 126). How much less easy and more arbitrary is it to decide, on the basis of degree of language meaning, between poetry and prose. One of ■Beardsley's problems is that he continually tries to retain critical terminology merely because it has been used by critics to say something significant about literary works. STet if critics are using different headings under which to talk about the same phenomena, is it really fruitful to re tain the headings and to try to distinguish the phenomena? 166 There often comes a time when it is necessary to ignore the terminology, to find for one's self what the phenomena have in common, and to define them according to this common char acteristic. This is the principle of economy in literature, in science, and in almost any other human activity? it is the principle, in short, of Occam's Razor: do not multiply entities unnecessarily. The distinction between structure and texture is equally arbitrary, as has already been indicated. Where does texture leave off and structure begin? Take, for in stance, the question of sound-texture. It merges impercept ibly into sound-structure by virtue of recurrence: "When certain types of texture are combined and maintained throughout a poem, however, we get large patterns that may be called sound-structures" (p. 221). One would suppose that this happens, too, when certain types of meaning-tex- ture are combined and maintained throughout a poem or a prose literary work— that a larger pattern, a "meaning structure," if you will, occurs, or in short, the work. Then both stylistics and explication disappear and we get what Beardsley calls "elucidation," determining the implicit meanings of the explicit structure of the work. But to return to Beardsley's definition of style, his 167 second version of this definition is: "Style is detail of meaning or small-scale meaning" (p. 223). The first ques tion is when or how is a detail meaningful? That question has already been answered to some extent. A detail is im plicitly meaningful if it deviates "from what is taken to be the norm of diction and syntactical construction." The sec ond question is what is deviation, indeed, what is normal? "'I am here/ Here I am' differ in style," he says, "and this difference is a difference in meaning— in what they suggest about the situation and the speaker's relation to it" (p. 223). All of his examples consist, like this one, of small differences of syntax or diction in two sentences. Thus these differences are, in fact, not differences between lit erary works but differences between two statements having a high degree of referentiality and differing slightly in their referents. The difference between "I am here" and J "Here I am," between whether the speaker says the one or the other, depends not so much on his attitude necessarily as on where he is, in short, upon the referent. Beardsley must inevitably conclude, and he does, that Where there is either no difference in meaning at all, or else a gross difference, we do not say there is a dif^ ference in style; where the difference in meaning is rel atively subtle and is present along with some basic sim ilarity on the primary level, we call the difference in meaning a difference in style. (p. 224) ____________________ 168 What this all amounts to is that statements differ in style if they have the same referent but a somewhat differ ent form. Yet this is precisely what never occurs in liter ature. The literary statement as a whole does not have a referent. Therefore, how can some basic similarity on the "primary level" be compared with some subtle difference on the "secondary level"? How in fact can a literary work be divided into primary and secondary levels at all? If by primary level is meant the "subject" of the work, what does the subject constitute? Wordsworth's twelve-^line poem, "To a Sky-lark," surely dif fers in more than style from Shelley's hundred-and-five-line poem, "To a Skylark." It is no doubt interesting to compare these two works, written in the same language about the same time (1825 and 1820 respectively) by poets familiar with each other's work. Yet each poem consists of different words in different syntactical constructions in different literary conventions of meter and rhyme and with a different meaning, or particular interpretation (what I have chosen to call in its most condensed form the theme). Is either of these poems really about a skylark, in the sense that a de scription of this bird is the interpretation of the particu lar poem, and if not, in what way then are they basically 169 similar? Even when the ostensible "subjects" are exactly the same object, not even differing as much as one skylark from another, a basic similarity on some "primary level" is very difficult to find. Conrad Ferdinand Meyer and Rainer Maria Rilke have, for example, both written poems on the famous fountain in the park of the Villa Borghese in Rome.® Yet the differences between the two poems are clearly the differences between two poems which are not the same poem. They differ in language and therefore they differ in meaning. In fact, although both are supposed to be describing the same object, scarcely any of the same words, except those such as "und," the definite articles, and a few other struc ture words, only one noun, and no verbs are shared. Even the titles, which ostensibly have the same meaning, "Der romische Brunnen" of Meyer and "Romische Fontane" of Rilke, differ in the presence of the definite article and the word used for fountain. Furthermore, Rilke's poem has precisely twice as many lines as Meyer's. Are the eight extra lines to be accounted for as a difference of style or as a differ ence of meaning? Even the poems of a writer like William Blake who often uses the same or similar subjects and titles over again differ not in "style" or "small scale meaning" _____^Cf. Kayser, pp. 285-286.________________________________ 170 but in total significance. The difference between "Chimney Sweeper" in Songs of Innocence and "The Chimney Sweeper" in Songs of Experience is scarcely a "relatively subtle" dif ference in meaning which constitutes style but a severe and explicit difference in meaning, a difference between the meaning of innocence and that of experience. Beardsley admits that with such examples as "I am here/ Here I am" and Go home/ Return to your abode" "We are not able to explore the richness of larger stylistic differences . . (p. 224), differences which, he suggests, exist be tween the styles of such writers as Bertrand Russell and William Faulkner, or Sir Thomas Browne and George Santayana, or Karl Marx and Carlyle. Now three points are to be noted about this list. The first point is that only one writer out of the six mentioned is a writer of non-referential, not to mention non-philosophical prose? or, more specifically, only one is a writer of literature in the strict sense, where basic similarities of meaning do not exist as they perhaps may in economics, symbolic logic, history, or even urn burials. Second, the differences between, for example, the writings of Bertrand Russell and those of William Faulkner simply cannot be reduced to comparisons between small-scale meaning. On what grounds is it possible to ------------------------------_____ i'7l compare them? If we cannot even find similarities on the primary level between two literary works on the same "sub ject," how will it be possible to find them between works which not only are not on the same ostensible subjects nor either referentially or thematically about the same kind of thing, but also are not even the same kind of works? The significant and the obvious differences will occur precisely at the primary level or at whatever level it is that exists the difference between literature and philosophy, a differ ence which can be safely called a major difference in mean ing. Differences between the prose of two "philosophers," such as Browne and Santayana, would doubtless have a great deal to do primarily with the fact that one of them wrote in the English of the seventeenth, the other in that of the twentieth, century, while the difference between Marx and Carlyle must first of all be considered as a difference be tween German and English, no matter how much Carlyle may have tried to make his English like German or how much Marx is translated into English. Finally, and most importantly, it is to be pointed out that even if such works, not to mention literary works, do have similar referents, we could, according to Beardsley's definition, only determine the style of one work by comparing 172 it with another work which had the same referent. We could never discuss the style of a work in and of itself, apart from other works. Style in this case might perhaps function as a relational concept— but only between works, and non- literary works at that. It could only be an applicable con cept if and when there were two or more works referentially and verifiably alike. Like individuality, "style" in this case cannot be considered as an attribute at all, but simply the comparison of two or more works for similarities and differences. The similarities and differences are entirely the result of the act of comparison, not of the objects com pared. To be sure, "I am here" and "Here I am" do differ, as Beardsley claims, and the difference between them is a difference of meaning. Agreed. But where does the term "style" become relevant? It simply becomes a word for dif ference of meaning, a difference which occurs on only a small scale only in small examples— like "I am here/ Here I am"— examples which, furthermore, cannot be fully inter preted in isolation but only according to the particular, the referential situation to which they are being applied. The meaning Beardsley interprets from "Here I am"— that it "suggests that I have been long awaited or searched for" (p. 223)— is not an adequate interpretation for every 173 occurrence of this statement. It is not applicable, for example, to Here I am, an old man in a dry month, Being read to by a boy, waiting for rain . . . One might even say that this instance "meant" exactly the opposite of Beardsley's interpretation. Trying to abstract a definition of style which will apply to the differences between Russell and Faulkner, Browne and Santayana, Marx and Carlyle, from such minute examples is reminiscent of the psychologist's attempt to define and explain verbal behavior in terms of the behavior of rats, who, as everyone knows, do not behave verbally at all.^® Beardsley recognizes these difficulties, I think, al though he never mentions them, for he bypasses this second definition in a third "version" of his conception of style, which modifies the second definition out of existence. Beardsley's third and final definition of style is: "Style is detail, or texture, of secondary meaning plus general purport" (p. 2 24). Yet secondary meaning is in essence much the same as general purport. Both convey something about the speaker. Secondary meaning is what a sentence suggests, 10Cf. e.g. B. F. Skinner, Verbal Behavior (New York, 1957) and John B. Carroll, "The Critical Need in the Study of Language," CCC, 13:23-26, October 1962.___________________ 174 "what we can infer that the speaker probably believes, be yond what it states" (p. 123). General purport also conveys something about the speaker beyond what he states, whatever else there is about him besides his beliefs and feelings. 'For the secondary meaning may also have emotive purport, i.e. it can convey something about the speaker's feelings over and above what he states. Also, what a statement states explicitly gives "utterance to beliefs" (p. 123); that is its primary meaning. The secondary meaning just gives more information about the speaker's beliefs. The meaning of a statement then is whatever it is capable of conveying about the person who makes it, his beliefs, his feelings, and any other of his "characteristics." The study of the style of a statement then is the study of what it conveys about the speaker, his beliefs, his feelings, and anything else about him. Where the primary meaning ceases and the secondary meaning or general purport begins is a question left unanswered. But in any case, no matter where we turn we are back at the speaker's beliefs, feelings, or characteristics, explicit or otherwise. This system of meaning which Beardsley has evolved is rather long and com plicated, scarcely refutable in a single paragraph or two. I (What it all amounts to, however, according to Beardsley 175 himself, is that "The meaning of a linguistic expression, then, is its capacity to formulate, to give evidence of, beliefs" (p. 118), and then he adds feelings and personal characteristics as the secondary meaning and general purport. Supposedly, however, "after we learn the meaning of a sentence, we can speak of its meaning as independent of what any particular speaker does with it" (p. 118). In practice though and, it may be suspected, even in theory, we cannot do this; at least in practice Beardsley does not. For when discussing literary works, he makes this funda mental distinction: "Whatever else it may be, a discourse is a connected utterance in which something is being said by somebody about something" (pp. 237-238). Therefore, every literary work has "first of all an implicit speaker, or voice: he whose words the work purports to be" (p. 238). Apparently we cannot speak of the meaning of a literary work apart from what any particular speaker does with it. The speaker we have always with us. Yet a discourse need not be an utterance; it can be a statement. And a work of litera ture is a statement. A statement is a use of language which can be abstracte from any particular space-time event, i.e. any act of utterance; Beardsley, too, thinks that some statements can be separated from their "pragmatic context ofj 176 utterance" (p. 239) and that, furthermore, a literary work is separable from its pragmatic context, whatever gave rise to it and whoever did. Then why does a literary work have to have an implicit speaker, a speaker, furthermore, who 1 1 reveals, advertises, or betrays himself partly through those very features of purport and meaning that we call 'style'" (p. 225)? A literary work exists both apart from and without reference to any actual space-time event, any pragmatic context; therefore it should be a statement, even according to Beardsley. But because the meaning of a liter ary work is derived, according to Beardsley, from what it says about the speaker, his attitude, his beliefs, his feel ings, and his characteristics, it has to have a speaker. But since the co-author of "The Intentional Fallacy" cannot very well attribute these attitudes, beliefs, feelings, and characteristics to the author of the literary work, he has to invent a wholly imaginary, a non-existent, an unproved and unprovable speaker. Does every literary work have an implicit speaker? Well, if it does, then some works have one implicit and one explicit speaker, and some works have an implicit speaker who "does not show any feelings or make any moral judgment about the event" (p. 241), one, in short, who is rather 17 7 difficult to find. The one genre which enabled the Roman tics to attest to the participation of the speaker in the work and the one which no doubt caused Beardsley to incor porate this speaker— plus a situation, and sometimes even a receiver— into his definition of all literature is the lyric. Except for quotations, there is only one kind of statement which is like an utterance, and that is a lyric. Even indirect discourse is not an utterance but about an act of utterance. Yet a lyric, as we have seen, is not an ut terance either. It frequently possesses characteristics which are common to utterances made by particular persons in particular situations— vagueness about a situation, indefi niteness of speaker and receiver, i.e. use of personal pro nouns and such space and time locutions as "here" and "now." Yet it is not an utterance, only the statement oj? one; it can be and is abstracted from any particular space-time event, i.e. any act of utterance. It has no speaker and it has no referent. When we describe it as having characteris tics in common with utterances, we infer a speaker and we infer a situation and sometimes we even infer a receiver, but they are not there, neither implicitly nor explicitly. And our inference contributes nothing except the recognition that the poem resembles an utterance. In addition, some 178 literary works are statements of an act of narration; a par ticular person tells about a particular action which he wit nessed or in which he was involved, e.g. Lord Jim or The Quiet American. In such a case the speaker is quite expli cit. It is a Marlow or a Fowler or a bishop ordering his tomb. Thus, if every work were to have an implicit speaker, and we wanted to know about him, we would have to make in ferences about him from what we infer about the explicit speaker, and by that time we would be back at the author. Are there two speakers in Lord Jim, Marlow and whoever speaks the first five chapters of straight narrative and then presents Marlow to tell the rest of the story? The jwork has enough speakers without our finding another. More over, some literary works are statements oj: an action, and sometimes these actions are presented without any overt com mentary. "But even here there is someone," says Beardsley, j"telling the story to us, and we know something about the teller from the telling" (p. 238). For despite the admitted •fact that "sometimes the speaker withdraws almost completely (p. 241), he is still there, Beardsley insists, and appar ently he determines that the speaker must be there, not by ■inferences from the work, but by definition, the definition that all discourses have speakers. '"Style is the man* gets 179 us nowhere" (p. 225), he says, the style is that of the speaker— but his discussion of style takes us remarkably close sometimes to the author, that is, to the man. For in stance, he says, regarding speakers who withdraw almost com pletely, "The great French realists, for example, Balzac and Flaubert, set their people and events before us like an im presario, and leave the rest to us" (p. 241). That fine line between the author's and the speaker's "styles" has disappeared altogether here, and we are left with no dis tinction at all. The "intentional fallacy" is not avoided by calling the actual author the implicit speaker. Wimsatt and Beardsley's "implicit speaker" is no different from Booth's "implied author," both of which differ not at all from Spitzer's "unconscious creator" in betraying or reveal ing themselves through their style or, as Booth would have it, their rhetoric. ^ C f . above, p. 80. CHAPTER V I I I STYLE AS LANGUAGE Once again we have seen that what are actually meta phorical transfers from the act of speech to the utterance itself are used to define and explain style; once again that style is defined as a feature of discourse which is sympto matic of the speaker; and once again that such definitions always involve dichotomizing the literary work. In Wimsatt and Beardsley this dichotomizing takes the form of levels of meaning— in Beardsley the "primary" and "secondary" levels, while Wimsatt generally uses the "substantive level" and the "strxctly verbal level," to make the same dxstxnctions. But the term "level," which must certainly be considered a metaphor when applied to language meaning, does not conceal the fact that what is actually involved is a distinction be tween, on the "primary" or "solid substantive level," the literary work, and, on the "secondary" or "strictly verbal level," the implicit speaker, who in this case, as in all ^The Verbal Icon, esp. pp. 133-151, 201-217. 180 18lJ others we have examined, could be no one but the author, and not the real but an inaccessible one at that. To define style as meaning is scarcely to solve the problem of style; it is only to complicate it with another term, upon the def inition of which will then have to wait the meaning of "style." Despite the fact that Beardsley has in the course of this dissertation furnished us with some very cogent terminological distinctions, when it comes to "style," his cogency escapes him. For even if we define meaning, and then define style as meaning, we still have to define and to justify the additional term "style." And even if we define style as the implicit speaker or the implied author, we still have to have some means for distinguishing "him" from either the explicit work or the explicit author. Perhaps the explanation of Hardy's rustic constable (in "The Three Strangers") cited by Wimsatt and Beardsley in their article on the intentional fallacy is more appropriate here. "He's the man we were in search of, that's true," says the con stable, "and yet he's not the man we were in search of. For the man we were in search of was not the man we wanted."2 During the course of this attempt at a radical analysis of the concept of style we have seen, as was asserted in the _____2Cf. The Verbal Icon, p. 5.______________________________ 182 introduction, that "style," despite the addition of the sci entific-sounding suffix "-istics," continues to function metaphorically. In fact the metaphorical nature of the term has become patent in the attempt to stretch it to do duty as a scientific concept in the study of language and litera ture. And other terms clustered about "style" have, so we have seen, suffered the same fate. "Style," as well as such terms as "expression" and "trait," are metaphorical trans fers from the area of human behavior to products of that be havior. We saw that these metaphorical transfers not only occurred historically in the application of rhetorical theory to literature, but also recur contemporarily in the study of style in literature and even in the use of the con cept of style by the social scientist. We saw that such transfers seem indeed to occur whenever "style" is applied pot just to literature but to language as well. The psy chologist, for example, considers language a subdivision of the area of specifically expressive behavior, whereas we saw that not only is literature neither expressive nor behavior, •but also that language itself is only restrictedly expres sive. "Expression," we saw, is a term applicable to lan guage only in acts of utterance and then only in certain •kinds of these. "Express," in the sense of "utter" or 183 "say," has been applied by transfer not to what utterances mean as language but to what they say or, more precisely, to what they indicate, about the speaker to an observer. We saw that, beginning with a basic confusion between written language and speech, these metaphorical transfers occur at every point, whether one is attributing the traits of the speaker to what is spoken, the surface of the work to the state of mind that must have produced it, the similarities of a genre or period to the works or authors which collec tively comprise those very similarities, or the secondary ■meaning of the work to the implicit speaker. These transfers, we saw, color the critic's entire ex pectations not only as to what he will find in literary works but even as to what sort of object he will think lit erary works are. Not only will he endow the work with lin guistic traits (like "character traits") which it cannot be shown objectively to possess or with attributes which are themselves metaphorical, such as rhythm, but also he will see the work itself, whether he does this in full awareness or not, as a sort of metaphor, as an entity primarily refer ring to something beyond it. This "something" beyond it usually turns out not to be the human experience which the work can be said to be about, however. Rather the work 184 becomes an index of, a substitution for, a designation of, a symbol of, the author, in one of his various manifestations, state of mind, creative process, Weltanschauung, personality unconscious, mentis character, implicit speaker. The work itself as a unique, meaningful utterance is seen as at least a surface and in extreme instances of criticism as a blind which covers and conceals its true meaning, the artist him self. The term I have used to designate this interpretation of the language of the literary work— "symptom"— is not, •however, in itself a metaphor but hopefully an accurate de scription of the basic premise which the various approaches to style I have treated have in common: the premise that a work of literature functions as an indicator of a condition of which it is a proper part. This is a more precise term for the way in which these scholars are interpreting the work than, for example, "sign," because a sign is not neces sarily caused by, but only associated with a condition which it may indicate. Then again such a term as "symbol," be sides being much too ambiguous and widely applied, is asso ciated with other than linguistic analyses of the work as an index of the unconscious; for example, archetypal or ortho- doxly Freudian interpretations, which raise interesting but 185 different problems than the problem of style. Critics do themselves occasionally use the terms "symptom" or "sympto matic" and in the way in which I have defined them, as for jexample in the statement by Cleanth Brooks and Robert Penn Warren that, "Individuality in style is important, then, not because it is valuable in itself, but as a symptom of the I . presence of somethxnq else; genuineness." Or they may, as Lpitzer did, describe their method of interpretation in diagnostic terms and liken it to the procedure of the physi cian.^ But such instances do not even suggest that critics realize the implications of treating style as symptom, nor that they recognize precisely that they are doing so. Iron ically enough, "symptom" and "diagnosis" in these instances appear to be simply explanatory metaphors. A literal grasp of the term and its precise application to their interpre tative procedures on the part of these critics and scholars would have obviated much of their literary criticism. Where does this leave us? Of the major groups of scholars whose ideas about style have been discussed thus ^Fundamentals of Good Writing (New York, 1950), p. 439, author's italics; Fundamentals of Good Writing was revised and amplified as A Modern Rhetoric (New York, 1958). ^See above, pp. 99-101, and Spitzer, "Les theories de la stylistique," pp. 166-167._________________________________ 186' far— the social scientists, the philologists, the critics, and the philosophers— only the philologists have done any extensive practical application of the concept. Wimsatt, for example, whom we may classify as a critic, has analyzed extensively the style of the works of only one writer, Samuel Johnson, and he a writer who composed only a handful of works which are strictly definable as literature. On the other hand, a linguist has chastised the philologists for "their programmatic disinterest in theoretical concepts € \ and a strict methodology." An interest in theoretical con cepts and a strict methodology seems to be present, however, in direct ratio to the absence of detailed application to the supposed object of the concepts and methodology— litera ture . Turning now to the linguists, who are supposedly devel oping "a new science of style," we could expect and we do find a predictable lack of any extensive practical applica tion. The linguists are very much concerned with theory and ■method but conduct few experiments or empirical verifica tions of their theories or methods. In relation to 5The Prose Style of Samuel Johnson (New Haven, 1941) and Philosophic Words: A Study of Style and Meaning in the ■Rambler and Dictionary of Samuel Johnson (New Haven, 1948). _____^See above, p. 10, and Sebeok, ed., p. 96.______________ 187 stylistics the linguists seem always to be working in the realm of the about-to-be-established. For example, the edi tor of a widely recommended anthology in linguistics con fessed in his introduction to a section of essays on "Lin guistics and the Study of Literature" that this section had 7 been "included three times and dropped twice." It was ul timately included only because the editor . . . yielded to the argument that since significant work is surely to be done within a short time the stu dent should acquire some awareness of the potential val ue of linguistics in the teaching and understanding of literature. (p. 393, italics mine) The section alluded to contains "A Report on the Language- Literature Seminar" at Indiana University in 1953 in which the authors of the report, Harold Whitehall and Archibald Hill, state their belief that literature cannot be studied . . . with the fullest fruitfulness unless the student is deeply versed in scientific linguistics, and is pre pared to focus this knowledge on both the external and internal characteristics of literature, in the hope of finding, eventually, those particularly significant char acteristics which define, in each culture, the differ ence between literature and mere everyday use of lan guage. (pp. 394-395, italics mine) That was in 1953. Prior to 1953 Hill had tendered two essays in the field of linguistic analysis of literature bearing such modest titles as "Towards a Literary Analysis" n 'Harold B. Allen, ed., Readings in Applied English Lin guistics (New York, 1958), p. 393.____________________________ 188 Q and "A Sample Literary Analysis." Then in an article in 1955, "Linguistics since Bloomfield," after suggesting that linguistics holds great promise for the solution of the problem of meaning, Hill concluded that it also holds great promise for . . . the identification of the metalinguistic components of literature which are peculiar to it, and which contri bute to its total meaning in a way similar to the contri bution of a smile to the total meaning of a non-literary utterance.® Precisely what is meant here is very difficult to determine. I Earlier in the article Hill had defined metalinguistics as "the area beyond, consisting ultimately of the non-linguis- tic objects and behavior with which language corresponds" (p. 18). The examples of metalinguistic entities in poetry which he gives are not meanings, however— what one would ex pect from the definition of metalinguistics— but meter and Lhyme. Yet Hill admits that meter and rhyme are not in themselves meaningful, but "do, however, contribute to the total aesthetic impression, which is the meaning of the poem" (p. 20). O English Studies in Honor of James Southall Wilson (Charlottesville, Va., 1951); Fourth Annual Round Table Meeting on Linguistics and Language Teaching (Washington, D.C., 1953). ^Quarterly Journal of Speech, 41:253-260, October 1955; reprinted in Allen, ed., p. 19._______________________________ 189 Hill goes on to say, I am extremely hesitant to dogmatize at this point, but such work as I have lately been doing in the analysis of poetry leads me to hazard the guess that in lyric poetry words, phrases, and sentences are typically cast into statements of four basic shapes (there may, of course, be others). . . . (p. 21) These four basic shapes are equations, analogies, sums, and affects and can contain and be contained in one another. Why these are shapes and in what way they are the same or differ from traditional rhetorical categories Hill does not say, but he concludes that Investigation of this type is certainly in its infancy, and you may not think it is very promising. Yet since literature is the use of language most characterized by special structural characteristics of its own, it prom ises a way of investigating structure which should be extremely repaying for linguistics. At the same time we can say that linguistics, which has been amply rewarding in the investigation of other types of utterance, holds some hope of illuminating the area of language in which man's values are most deeply embedded, the literature which has been the constant creation, companion, and model of his spirit. (p. 23, italics mine) The dependent clause, "since literature is the use of lan guage most characterized by special structural characteris tics of its own," seems to be an instance of question-beg ging, since what the investigator is supposed to be investi gating is whether or not literature i_s such a use of lan guage so characterized. Then in the concluding sentence Hill seems to be defining literature in still another way— 190 as "the area of language in which man's values are most deeply embedded," while this definition seems furthermore to be restricted to a special kind of literature, that kind which has been "the constant creation, companion, and model" of man's spirit. Does the structural definition bear any ‘ relation whatsoever to the moral one? Certainly one is not derivable from the other. Yet after two more tentative and inconclusive essays in "structural analysis"Hill presented "A Program for the 'Definition of Literature" in which he declared, The burden of this paper has been to insist that litera ture has its being in the area of stylistics and the definition of literature must be sought in stylistics. ^ Be that as it may, Hill's conception is only a program and not, as the title states, a program for the definition of literature, but a program for the definition of the various species of literature, later to be classified into genera, and eventually into orders. Furthermore, this or rather these definitions will be derived from "a corpus for study" established on the basis of permanence, a basis which has no -*-®"An Analysis of 'The Windhover': An Experiment in Structural Method," PMLA, 70:968-978, December 1955; and "'Pippa's Song1: Two Attempts at Structural Criticism," Uni versity of Texas Studies in English, 35, 1956, pp. 51-56. -^University of Texas Studies in English, 37, 1958, p. 52. An abstract appears in Sebeok, ed., pp. 94-95.__________ 'bearing on the kind of defining characteristics to be sought, for these will be formal or "stylistic" characteristics. On such a basis Hill has no hesitation in saying that "Thirty days hath September" is an example of both poetry and liter ature. But this is scarcely much of an admission, since ■Hill's program would also stipulate as literature all the— in his phrase— "institutionalized great books" of Western Europe. Thus literature would presumably include Aris totle's Physics and, because modern works would be included on the basis of formal similarity to the great books, modern works on physics insofar as they formally resemble Aris totle's. The program does not, after all, seem very promis ing. Not only do we still have all the things which have been called literature, we have a great many more which have ■not. And no defining characteristic of literature has yet been unearthed by Hill, much less a conception of style, which is supposed to be the key to all this. It may be that "the definition of literature must be sought in stylistics," but such an assumption does not help at all unless one knows what stylistics is. In any case, whatever it is, it still exists only in the realm of "must be," of the "promising." There must be such stylistic features, it is argued, al though they have not yet been found. 192 Hill's promissory notes are characteristic of the American linguist's position on stylistics. His program matic statements almost always include, for support, refer ence to the as yet only partially published work of Zellig S. Harris in stylistics. Harris prefers to call his work "discourse analysis" but offers it as a method for determin ing style. Yet Harris's contributions scarcely constitute support, for they too are largely programmatic. In 1952 he declared, It remains to be shown as a matter of empirical fact that such formal correlations do exist, that the discourses of a particular person, social group, style, or subject- matter exhibit not only particular meanings (in their selection of morphemes) but also characteristic formal 10 features. * Twelve years later this matter of empirical fact is still wanting. These promissory notes are not, however, restricted to American linguists interested in style. About the same time that Harris and Hill began making their programmatic pro nouncements, a disciple of Charles Bally's at the Sorbonne, Charles Bruneau, was calling in Romance Philology for a 12"Discourse Analysis," Language, 28:1-30, January- March 1952. Cf. "Discourse Analysis: A Sample Text," Lan guage , 28:474-494, October-December 1952, and "Co-occurrence and Transformations in Linguistic Structure," Language, 33: 283-340, July-September 1957._________________________________ 193 clarification of the concept of stylistics so that work in the field might be carried on. He ominously noted at the beginning, however, that Si l'on se rappelle ce qu'il advint jadis de la Tour de Babel, qui ne put @tre achevee par suite de la confusion des langues, on peut concevoir des craintes serieuses au sujet de l'avenir de cette science encore au berceau.-^ Moreover, not only is this "science" still in the cradle, it has been there for all of its relatively long existence. II semble done qu'aprfes un demi-siecle environ d'exist ence le mot de stylistique ne presente plus un sens aussi precis que les mots de phonetique, de morphologie et de syntaxe. (p. 1) It did not occur to Bruneau, however, that possibly the term "stylistics" does not have as precise a meaning as "phonet ics," "morphology," and "syntax" precisely because there is nothing for it to mean. That utterances may be fruitfully analyzed phonetically, morphologically, or syntactically but not stylistically never occurs to him. Rather than discard a term for which there seems to be no meaning, he chooses to "clarify" the term by giving it a definition, equally impre cise, of his own. Apparently, if there exists a term, there must exist a precise concept which it designates and 13»La stylistique," Romance Philology, 5:1-14, August 1951. For other instances of the use of the epithet "Tower of Babel" in describing the state of current terminology see above, p. 7 . ____________________________________________________j 194 furthermore a "science" based on this concept, even though during the course of fifty years none has been found. In view of the history of this term Bruneau has very good rea sons for worrying about its future. Nevertheless, it is to the future stylistics, the fu ture science, that Bruneau appeals in his efforts at clari fication. And he admits from the very beginning that his motives are directed toward it. J'essaierai, en me placant a un point de vue strictement linguistique, de definir exactement le champ d'etudes de la stylistique. Avouerai-je qu'en me risquant a tracer un programme de recherches, j'esp^re recruter une equipe de jeunes "stylisticiens"; cette science recente, dont les possibilites sont immenses, a un besoin urgent de nombreux travailleurs. (p. 1) Why a science which has yet to be established should be in urgent need of workers is a puzzle. The best thing to at tract recruits with, of course, would be not a program but a successful achievement, a demonstration— not an announce ment— of the possibilities. But there seem to exist no dem onstrations. During his twenty-odd years at the Sorbonne Bruneau has accepted or stimulated a number of studies of "the language and style" of individual authors. Yet Bruneau has a justifiably quite modest view of these studies: "Dans ma pens^e, ces travaux n'etaient que des travaux d'approche" (p. 8). This "trench-work" as Bruneau calls it, 195 II est des sciences de "ramassage," si je puis dire, telles que la botanique et la zoologie; elles observent et classent des faits sans etablir de lois. Sous sa forme premibre et modeste, la stylistique ne pouvait guere §tre qu'une science descriptive. (p. 8) •Like most non-scientists who are fond of making analogies .with science, Bruneau seems to have misunderstood the nature of botany and zoology and perhaps of science in general. There exists no science of which the sole purpose is to classify; rather, classification is a relevant aspect, but toy no means the only one, of every kind of study which is called a science. The periodic table in chemistry is as much a system of classification as anything the zoologist makes, and the nature of plant life can be explained in terms of "laws" just as much as chemical reactions can. Un deniably, "II ne sera possible de construire des syntheses solides qu'en s'appuyant sur un grand nombre de faits bien ^tablis, soigneusement dates et pr^cisement localises" (p. 8). But it does not follow from this that it is- possible to construct a solid synthesis of any particular great number of well-established facts. The facts selected may or may ■not have any bearing on one another. That a "science" is only in the programmatic stage, of course, scarcely constitutes conclusive evidence of its in adequate conception, but it is indicative. The science of 196 biochemistry was established only after, not before, consid erable success had been achieved in the study of biochemical phenomena. Alchemy may have led to the science of chemis try, but it never led to the transmutation of "baser" ele ments into gold; that was a misconception that had to be abandoned. Sciences are not raised upon platforms but upon foundations. One important criticism that can be made of the stylistically-minded linguists on both sides of the Atlantic is that they have never clearly demonstrated that style exists and is something that can be studied. They have not studied it; they have only talked about studying it. And such behavior is even more indicative when it is noted that the result of still another "clarification" of the concepts of style and stylistics is the vague conclusion of Bruneau's that Le style serait alors 1'ensemble des faits particuliers que caracterisent l'individu, ^crivant ou parlant, par rapport au materiel que la langue (la societe) lui four- nit. (p. 14). "Par rapport" here tells us nothing. Whether there is such a relationship and what it is are the crucial questions and the ones which Bruneau ignores. What arouses even more apprehension about the still programmatic nature of linguistic stylistics, however, is 197 the extravagance of its claims and expectations. In the first place, the stylistically-minded linguists claim that they are the only students of style who are truly "scien tific," and they are very critical on this basis of all other approaches. "La nouvelle linguistique ou 'criticisme stylistique,'" says Bruneau, referring to the work of Spit- zer, the Alonsos, and others of the intuitive school of ap proach to style, "n'est pas une science" (p. 13). This is perhaps indisputable. What is puzzling though is the im mediately succeeding observation, "Que la stylistique soit indispensable au critique litteraire, le linguiste en est convaincu," which furthermore is followed immediately by the footnote that "Le linguiste constate seulement-— avec re gret— que la stylistique scientifique n'est encore qu'une pauvre chose balbutiante et qui cherche sa voie" (p. 13). Bruneau declares himself convinced that the stylistics which he conceives is scientific, although he admits that he has no grounds whatsoever for the conviction. His conception of the scientific, like that of his American counterparts, seems to rest on a faith in the indisputability of tabula tion. Bruneau feels that stylistics could provide psychol ogy with "documents d'une precision vraiment scientifique" (p. 10), and to illustrate this claim he describes a method 198 of demonstrating the contrasting imaginations of two authors 'by . . . d'aligner, sur deux colonnes ou sur deux pages, des series d1images choisies parmi les plus caracteris- tiques de deux auteurs. Ces images auraient de prefer ence un caract^re commun, soit qu'elles portent sur les mimes objets ou qu'elles expriment des sentiments ana logues. Le contraste des deux 'potentiels' imaginatifs serait sensible. (p. 10) Spitzer is quite right, of course, in pointing out that the decision as to which images are the most characteristic of the authors is an entirely subjective procedure, regard less of what is demonstrated by their juxtaposition. In the second place, these exclusive claims to "scien tificness" tend to merge into even more unfounded and ex travagant claims about the probable achievements of this new scientific approach to style. Most notably, stylistics will, . . . greice a une analyse extr§mement precise et minuti- euse, fondee, k 1'occasion, sur des statistiques, etablir, mieux que 1'introspection ou les autres moyens utilises par les psychologues, les diverses attitudes de 1'esprit humain dans l'infinie variete de leurs nuances. (p. 7) And not only this, Tant au point de vue psychologique qu'au point de vue linguistique, cette enqu^te approfondie peut aboutir a des resultats objectifs? elle permet de "savoir" et, dans une certaine mesure, de "prevoir." (p. 7) Bruneau's expectations here, however, are rather modest in 14 ✓ "Les theories de la stylistique," p. 167. 199 comparison to those of some American linguists. Hill says that the function of stylistics is to reduce the area of lin guistic arbitrariness by explaining as much as possible of linguistic variation. It is not, of course, believed that all variation is controlled— it will never be pos sible to predict all that a man might say before he says it. Nonetheless the area of scientific stylistics is exactly the area in which explanation, and therefore pre dictions, of linguistic choices can be made.-1 - - 5 This is a remarkable claim, but built into it are a number of assumptions, not the least of which is that what can be explained can therefore be predicted. Nevertheless, if stylistics could enable us to predict, if only partially, what a man might say before he says it, it would constitute the most extraordinary and significant achievement of all linguistic science. The issuance of such a claim, there fore, even in the form of a promissory note, merits con siderable attention. The particular kind of linguistics which has been developed in the last thirty years has often and rightly been called "descriptive linguistics," that is, the analysis and description of given data— groups of utter ances, languages, dialects. But the tendency has been grow ing to think of linguistics as a predictive science. This is mirrored in such concepts as "generative grammar," but • ^Introduction to Linguistic Structures (New York, 1958)__p_._408___________________________________________________ 200 also in such notions as that of C. F. Voegelin's that . . . the very interesting possibility exists that lin guistic selection may throw a sharp but narrowly focused light on the large but diffuse problem of free will.*-® If this possibility does in fact exist and if stylistics is in fact that area of linguistics in which it does exist, then the concept of style propounded by the linguists merits the utmost attention and promises to be of the greatest metaphysical importance because it will "shed light on" one of the most ancient problems of human existence. Unfortunately, not only is this possibility purely hypothetical, it is also obviously derived from a purely linguistic mistake, a mistake, that is, in the "choice" and application of certain terms. If style, as the linguists define it, is choice, and if, as such, it is studiable, then one is only being reasonable in supposing that style or "linguistic selection" might throw some light on the nature of choice and therefore on the existence of free will. Un fortunately, style as choice is not a workable concept. Neither is it entirely new— nor are the claims made for it. Thirty years ago I. A. Richards, writing under the heading, not of linguistics or stylistics, but of rhetoric, proposed Casual and Noncasual Utterances within a Unified Structure," in Sebeok, p. 58._____________________________ 201 that A discussion of the reasons for the choice of words— which too often seems a trivial exchange of whimsies— can become an introduction to the theory of all choices.^ Almost everyone at the Indiana conference, as we claimed earlier, had been talking about the same thing— literary critics like Richards and anthropologists like Voegelin in cluded. What had created the difficulties and disagree ments, of course, was that at bottom there was nothing there to work on. Any attempt to apply the concept of style to literature must inevitably end in frustration. A quotation from almost any linguist will serve to show that the definition of style predominant among linguists is that of style as choice. Hill says It is possible to define the sum total of style [not "style" itself] as all the choices of equivalent items which the language offers the user in each linguistic situation. Stylistics under this . . . view, is the col lection and tabulation of these alternatives, sentence by sentence throughout the corpus under study. To be sure, Hill also offers another definition, but "Dif ferent as the two definitions might appear to be," he claims, "they can be reconciled and are no more than obverse and reverse of the same coin" (p. 407). This other -^The Philosophy of Rhetoric, p. 86. -^Introduction to Linguistic Structures, pp. 406-407. 202 definition, it should be noted, however, is not even a defi nition of the sum total of style but of stylistics. The definition of stylistics . . . which has gained con siderable currency among linguists . . . is that stylis tics concerns all those relations among linguistic enti ties which are statable, or may be statable, in terms of wider spans than those which fall within the limits of the sentence. (p. 406) The question which this definition prompts is a simple one: 'Haven't the linguists been doing this all along? To be sure, the linguist occasionally likes to think that he ana lyzes one utterance (i.e., loosely defined, one sentence) at a time, regardless of the utterances which come before or after it.-*-® But this is, at least in English, not an en tirely feasible activity. For in English the linguist deals with a large group of words which are analyzable only in terms of preceding utterances; these are the various substi tution classes— pronouns, auxiliaries, adjective substi tutes, adverb substitutes. Noun-determiners, function nouns, and function verbs quite often function and are ana lyzable only as words which relate one sentence to another. Coordinators, sentence modifiers, and sentence-linkers, as their names indicate, function similarly. ideally the 19 Cf. Harris, Methods m Structural Linguistics (Chicago, 1951), pp. 11-12. 20 _______Cf. W. Nelson Francis, The Structure of American_____ 203 linguist may think he works only with single utterances, but in point of fact, he cannot do linguistics without consider ing relations between and sometimes among utterances. In this sense "stylistics" seems to be a needless addition to an already abundant terminology. In any case this other definition presents us with no concept of style whatsoever. It is therefore difficult to see in what way it is the other side of the coin to style as choice. Thus the relationship between the two definitions and thus the clarity of the one have been obscured. "Stylisti cally considered," says Hill in attempting to combine the two definitions, "what was an unpredictable choice within the sentence becomes something explainable in terms of the wider span" (p. 407). Certainly the use of "thus" as the first word of the first sentence in this paragraph would be inexplicable were it not known that a sentence— and a spe cific kind of sentence at that— had preceded it. Its ap pearance as the first word of an essay or book would not be a seemingly "unpredictable choice," however; it would be in comprehensible. "Thus" can be understood here only as a ■English (New York, 1958), esp. pp. 413-417; and Charles C. ■Fries, The Structure of English (New York, 1952), esp. pp. 1240-255. ____ ______________ __________________ 204 word which relates one sentence or more to others. As an example of predictable choice Hill offers the fact that Linguistically, spherophore and ball-carrier are differ ent items, but under conditions in which the stylistic situation is fully known, the occurrence of one or the other can be predicted. (p. 407) But this seems simply to say that if one knows the entire discourse in which a word appears except the word itself, one can predict which of two possibilities will occur. And that seems indeed to be an unpredictable use of the term "prediction." About the choice of "spherophore" or "ball carrier" Hill explains. We would ordinarily say they "mean the same thing" or "have the same reference," which are loose ways of saying that the difference is without linguistic function— the two forms do not keep separate phrases and sentences apart. They are synonyms, in short. (p. 408) "Spherophore" is in fact a word coined by Hill from Latin roots to be synonymous with "ball-carrier." What Hill is actually offering here, despite the linguistic jargon and the misleading use of "prediction," is simply another state ment of the by now familiar, indeed, hackneyed concept of style as choice, as choice among at least two ways of saying the same thing. "Prediction" here is being used, idiosyn- cratically, to mean knowing why the author chose the words or constructions he did choose from what alternatives were available to him, and seems to have nothing. at all to do____ 205 with knowing what a man might say before he says it. The idea of restricting stylistics to spans wider than the sin gle sentence is relevant only if one is under the mistaken assumption that grammar ends with a period or rather a "sen tence-final intonation pattern." What Hill offers as a concept of style is what a host of linguistically-minded students of the new science of style offer: the ancient concept of style as the manner at tached to the matter. And the concept of choice is just another attempt, like the mind or Weltanschauung or id of the author or the attitude of the implied speaker, to halt the infinite regress of causal explanation by talking about the behavior of the author. Here is James Sledd in A Short /Introduction to English Grammar, which utilizes the tech- Liques of contemporary descriptive linguistics, talking about style: Style, that is, will be for us the manner of saying what is said. . . . The manner of the utterance is as essen tial to its effect as the carefully chosen details which are its matter. . . . If style is the manner of saying what is said, then it follows that style is possible only because there are more ways of saying a thing than one. . . . It should be plain, by this point, that style in language is itself synonymous with linguistic choice. . . . We must recognize that if we want to talk about a man's style, we must know both how he said things and how he might have said them but chose not to. . . . ^ Here is UlLmann presenting and supporting the concept of 206 style of Saussure's disciple Bally and Bally*s disciples such as Bruneau: The pivot of the whole theory of expressiveness is the concept of choice. There can be no question of style un less the speaker or writer has the possibility of choos ing between alternative forms of expression. Synonymy, in the widest sense of the term, lies at the root of the whole problem of style. . . . At the risk of oversimpli fication, one might say that everything which, in lan guage, transcends pure communication belongs to the prov ince of style.22 'Here is H. A. Gleason, author of the standard text, An In troduction to Descriptive Linguistics (New York, 1955), as serting that . . . there are usually several ways in which a sentence, a clause, a phrase, or even a single word can convey the required meaning and still be grammatical. The author is, therefore, presented at every point with options. He must choose, and the choice is one of the elements of style. . . . Style is . . . the patterning of these choices over long stretches.23! Charles Hockett in A Modern Course in Linguistics (New York, 1958), says that . . . two utterances in the same language which convey approximately the same information, but which are differ ent in their linguistic structure, can be said to differ in style. (p. 556) 21 (Chicago, 1959), pp. 261, 263-265. 22Style in the French Novel, p. 6. See also his Seman tics: An Introduction to the Science of Meaning (Oxford, 1962), p. 151. _____23,1 What Is English?" CCC, 13:1-10, October 1962._______ 207 Joshua Whatmough in Language; A Modern Synthesis (New York, 1956) declares, "That is what style is— the selection of particular linguistic units and the variation in their ar rangements" (p. 88). And as a final example of how all- pervading this concept of style is, Charles E. Osgood in "Some Effects of Motivation on Style of Encoding" says, The study of style concerns the variable features of the code. However, the variability need not be completely "free" (i.e., chance, unpredictable); rather, the student of style is interested in the statistical properties of choice where there is some degree of freedom in selection. . . . Stylistics is generally more concerned with struc tural choices than with lexical choices, that is, in how a person talks about something rather than what he talks about. (Sebeok, p. 293) : Such unanimity might make the cautious soul apprehen sive. Might there not be "something" to a concept which so many seem agreed upon and which has, indeed, been agreed upon for centuries? The answer to this question is three fold. One, agreement is more likely to be found in theories or programs than in practices. Two, style would not be the first prevalent and enduring concept which has been found to be fallacious. Three, it is fallacious. When we have said that style is choice, we have not as yet said anything specific. Whose choice is it? Most lin guists would seem to agree that it is the author or speak er's choice. But even this is not so clear cut as it might 208 seem. A work of literature, indeed, any writing, does not seem obviously to be a pattern of choices. How do we know what the author or speaker decided among? We know in a sense what he did choose, that is, what he did say, but this hardly constitutes "choice" in any standard sense of the term, unless, as Sledd says, we know also how the author or speaker might have said what he said but chose not to. In short, we have to know what his alternatives were, and those are not in the text. In some instances, of course, we have alternative texts, earlier versions of the author's which contain words inserted and others crossed out. This does not guarantee, however, that we have in these cases only a difference of how the author said what he said. Perhaps the author's ideas changed during the course of composition. There is no reason to suppose that he began with the liter ary work which he ended up with and that composition is only deciding how to say what one wants to say. And in any case, the number of variant manuscript versions of literary works which are available is relatively slight. Therefore, if the study of style were confined to them, it would constitute a very small and hardly very significant area of study needing numerous new investigators. Of course the linguists are not talking about this. 209 What they want to ask is: How, given the language in which the author was writing, could he have said the same thing differently? This might be expressed in Hill's terms as: What are the synonyms available in the language for every thing which the author said? Actually, this is not choice in any conventional sense of the term since we can never, except in the case of variant manuscripts, really know of any choices which were available to that particular author. No one person has available to him all the elements of the "language" of which he constitutes one user. This may not be obvious in the case of sentence patterns and inflections, but it is certainly so in the case of vocabulary. Any given language consists of an enormous variety of utterances (la parole) of which the linguists’ and lexicographers' abstrac tions (la langue) are of a small number of recurrent fea tures. We know simply what the author said; we do not know what he knew that he might have said. Even though we cannot know what choices were actually available to a particular author writing at a particular time, however, let us grant for purposes of discussion that the entire language was available to him for choice of those linguistic elements which would best convey his meaning. Even though these were not the actual alternatives available 210 to the author, let us assume that there is something— we do not yet know what— to be gained from examining the hypothet ical alternatives to what the author wrote offered by the language in which he wrote. How are we going to do this? What alternatives, even hypothetical ones, present them selves? These alternatives, in order to be considered le gitimately possible alternatives, have, according to Hockett, to convey the same information, though they will, of course, differ in linguistic structure. They have to be, that is, synonyms, words or expressions which are distinct but "mean the same thing." Ullmann's statement that "Synonymy, in the widest sense of the term, lies at the root of the whole problem of style" is therefore no exaggeration. Unfortunately, the concept of synonymy does not clarify the problem at all. It merely pushes it one step backwards. And this happens regardless, and linguists disagree, of whether there are such things as synonyms or not. If syno nyms or synonymous expressions do exist— that is, if one can say exactly the same thing differently— then the concept of choice ceases to be relevant. If two or more words or syn tactical constructions do mean exactly the same thing, then it makes no difference whatsoever which one the author chose to use. The choice was aimless, unpredictable, meaningless. 211 Choice here exists only in the sense that there are two pos sibilities, but there exist no grounds for making a decision between them. It is interesting to note that Ullmann in a chapter on synonymy in his Semantics points out that com plete synonymy does seem to exist "where one would least ex pect it: in technical nomenclatures." The fact that scientific terms are precisely delimited and emotionally neutral enables us to find out quite def initely whether two of them are completely interchange able, and absolute synonymy is by no means infrequent. (p- 141) Later he uses as examples a number of technical terms in German which differ only in that one of each pair has been formed from native roots and the other from Graeco-Latin roots. "And as these synonyms are used in the same con texts," says Ullmann, "and sometimes even in the title of the same book, one can hardly speak even of stylistic dif ferences between them" (p. 142). The inference should be clear, though Ullmann does not draw it. If two words are synonyms, that is, truly alternative choices, then the choice of the one or the other makes no difference, is, in fact, meaningless. Therefore, "the collection and tabulation of these alternatives, sentence by sentence throughout the cor pus under study" would be as pointless, as meaningless, in short, as the choice among the synonyms was for the author. 212 On the other hand, if it is agreed that there are no actual synonyms, then a difference between two words or phrases constitutes a difference— not in style— but in mean ing. If there are no words which have the exact same mean ing, then the difference between any two words, regardless of how similar or dissimilar they are in meaning, is by definition a difference in meaning. That is why they are not synonyms. "The best method for the delimitation of synonyms," says Ullmann in his Semantics, "is the substitu tion test." This, it will be remembered, is one of the fundamental procedures of modern linguistics, and in the case of synonyms it reveals at once whether, and how far, they are interchangeable. If the difference is predominantly objective, one will often find a certain overlap in mean ing; the terms involved may be interchanged in some con texts but not in others. . . .If, on the other hand, the difference between synonyms is mainly emotive or stylistic, there may be no overlap at all: however close in objective meaning, they belong to totally different registers or levels of style and cannot normally be inter changed. (p. 143) If the best method for the delimitation of synonyms is the substitution test, and if certain kinds of words, no matter how similar in meaning, are not interchangeable, then pre sumably they are not synonyms. They do not pass the synon ymy test. Therefore, they do not offer themselves as alter native choices to the writer, even hypothetically. The Lwriter_co.uld_have_used_the_one_if_he_had_meant_the_one.;_he__ 213 used the other because he meant the other. The writer chose to say what he meant. The concept of choice in such a con text evaporates and with it the concept of style. What is the difference between two different words with similar ■meanings? They do not have the same meaning. Although Ullmann does not realize it, he has just refuted his own conception of style, and in doing so has refuted that of all the linguists who define style as choice. The linguists have, in fact, fallen into the same pit as all the others who have tried to talk about style. They have seen that there is nothing in the literary work itself which can be pointed out and designated as style. On the one hand there is the specific meaning or interpretation of the particular instance of language (la parole), and on the other hand there are the features of the work which render it an intelligible instance of a particular language— its jrammar and vocabulary insofar as it belongs to la langue. Where then is style? It is not in the work. Perhaps it is something in the work insofar as the work exists in relation to something else? To be sure, the linguists did not set up the author as the entity to which the work is related by virtue of its having style or a style. But in effect they did the same thing. Style as choice is not precisely an attribute of the author, but then neither is it an attribute <af the work. What the author wrote reveals, by comparison with the language in which he wrote, what he might have written had he chosen to do so. This notion of choice makes it a purely hypothetical behavior just as the author's Weltanschauung, id, psyche, or the implied speaker's atti tude are purely hypothetical entities, entities established in order to enable one to talk about the style of a work since style itself cannot be shown to exist. In every case of the use of the word "style" which we have examined, the user has found it necessary to go outside the work to estab lish the existence of style, and in every case he has had to go to something for which there exists no evidence but the particular work whose style he wishes to discuss. There exists no evidence for the choices of an author except inso far as he chose to write the work which he did in fact jwrite, just as the only thing we know, of what must have I been in the mind of an author at the time he wrote a partic ular work, was the work which he did in fact write at that time. There will be no new science of style. The reason the pronouncements of the linguists on both sides of the Atlan tic have remained in the programmatic stage is that they 215 cannot progress beyond that stage. Were the linguists actu ally to attempt to apply the concept of style which they ad vocate they would see at once that it embodies a simple logical error that renders it inoperable. It seems only fitting that an analysis of the concept of style should conclude with an analogy drawn from an area of human endeavor which has been much invoked in modern stylistics— science. Throughout these pages we have heard time and again numerous contemporary students of style de fend the scientificness of their own conceptions and methods and accuse those with different conceptions and methods of being unscientific. Of course, as we have seen, the charges and countercharges have ultimately proven irrelevant. Nevertheless, it is appropriate to point out one way in which the concept of style during the course of its entire history has resembled a famous and crucial scientific con cept— the concept of ether. Like the concept of style, the concept of ether proved to be at best a stop-gap hypothesis in lieu of adequate ex planations and at worst an inadequate but comfortable ex planation which blinded scientists to the need for a re- evaluation of their conclusions. Although sometimes vehe mently insisted on— for metaphysical reasons— ether had no 216 properties of its own and thus "presented certain problems, not the least of which was that its actual existence had 24 never been proven." To eighteenth and nineteenth century physicists it was obvious that if light consisted of waves, there must be some medium to support them. . . . Hence when experiments showed that light can travel in a vacuum, scientists evolved a hypothetical substance called "ether" which they decided must pervade all space and matter. Later on Faraday propounded another kind of ether as the carrier of electric and magnetic forces. When Maxwell finally identified light as an electromagnetic disturbance the case for the ether seemed assured. (pp. 40-41) But because every theoretician gave ether the properties that were required by his particular theory, "by 1880 the properties that had to be assigned to ether were so contra dictory that physicists began to doubt its existence alto- 9 R gether." Furthermore, as Whitehead points out m Science and the Modern World, Whereas quite a simple sort of elastic ether sufficed for light when taken by itself, the electromagnetic ether has to be endowed with just those properties necessary for the production of the electromagnetic occurrences. In fact, it becomes a mere name for the material which is postulated to underlie these occurrences. If you do not happen to hold the metaphysical theory which makes you 24 Lincoln Barnett, The Universe and Doctor Einstein (New York, rev. ed., 1957), p. 41. ^^Morris Kline, Mathematics in Western Culture (New York, 1953) , p. 435._____________________________________ 217 postulate such an ether, you can discard it. For it has no independent vitality.2° The parallel with style here should be patent: an entity jWith no attributes of its own, whose existence has never been proven, which is assigned the properties of whatever system it happens to be applied to, and therefore has numer ous contradictory properties, and which, finally, can be discarded as a mere duplicate name "if you do not happen to ■hold the metaphysical theory which makes you postulate" such an entity, is open to serious doubt. It was at this point, 1881 to be exact, that two Ameri can physicists, A. A. Michelson and E. W. Morley, performed their classic experiment to discover once and for all .whether there really was any such thing as ether. The spe cific details and purpose of the experiment need not concern us here. Suffice it to say that if ether does exist, "a light ray projected in the direction of the earth's movement should be slightly retarded by the ether flow."^7 There fore, Michelson and Morley "constructed an instrument of such great delicacy that it could detect a variation of even ^®In Alfred North Whitehead: An Anthology, ed. F. S. C. Northrop and Mason W. Gross (New York, 1953), p. 453. 27Barnett, p. 42. 218 a fraction of a mile per second in the enormous velocity of light" (p. 43) . The whole experiment was planned and executed with such painstaking precision that the result could not be doubted. And the result was simply this: there was no difference whatsoever in the velocity of the light beams regardless of their direction. (p. 44) In short, so far as man could determine with the most pre cise tools and methods imaginable, ether did not exist. Then, after a quarter of a century of debate among scien tists, a twenty-six-year-old patent officer published his classic paper on relativity in which he began by rejecting the ether theory. Needless to say, this radical analysis of the concept of style does not constitute a theory of relativity, but it may not be presumptuous to compare it to a Michelson-Morley experiment. Its results are purely negative, but a negative experiment which is precisely conducted, so that its results are conclusive, can be a positive contribution, if for no other reason than that it disposes of concepts which ob struct the development of more adequate conceptions. To be able to discard the concept of an entity which is not neces sary and whose existence can be neither empirically estab lished nor logically deduced— this is to be truly scien tific . APPENDIX POUR CRITICAL TERMS A work of literature has been defined in this disserta tion as a statement o^f an event which functions as an ana logue in an implied analogy with an aspect of human experi ence. Given this definition, four distinctions emerge as basic tools of criticism. In any developed critical discus sion of a literary work the event, that is, what the work is about in the primary sense, ought to be stated. The con ception of the event will be derived from inferences made from the text, which will in turn provide the basis for the major inference, the theme, and for conceptions of human ex perience implicit in the work, its implications. Defini tions of these four terms are provided below. Sample state ments illustrating what phases of the critical procedure the term covers are based on Katherine Anne Porter's novelle, "Old Mortality," from the collection Pale Horse, Pale Rider (New York, 1939), pp. 3-89. A great deal more can be said about any work than these terms cover. They are the minimum 220 221 'basic distinctions for pursuing a critical discussion, not the pursuit itself. EVENT: That single inclusive imaginary space-time oc currence which the work is about (i.e. its subject) and the statement of which is the work. This may be co-extensive with the mode of narration, as in Moll Flanders and "O West ern Wind," but not necessarily so. For example, in Lord Jim the event, i.e. Marlow's recounting of his knowledge of Jim, is not introduced until Chapter Five. The event may include more than the work; e.g. in "O Western Wind" a speaker whose cry it is is inferred from the fact that the statement is a cry. The event may comprise less than the work, as in Lord jjim. The event is not the words or sentences on the page but the total occurrence which the reader infers from the statement which is the work. The event of "Old Mortality" is the relationship of a Southern girl to her family as she grows up to the age of eighteen. INFERENCE: An interpretation of an aspect of the event based upon the information given in the work. "X does Y be cause of Z" is an inference from the work if X, Y, and ! Z are each a part of the statement which is the work. The infer ence may be from any part of the work and on any scale. 222 When Maria asks about the reference to the dead Amy as a "singing angel" in Uncle Gabriel's poem ("Did she really sing?"), we can infer that Maria has not understood the meaning of the poem as her elders understand it (p. 17). This is an inference from just a few lines of the work. On a larger scale we can infer that Miranda's primary motiva tion in marrying suddenly, whether or not she recognized it at the time, was to achieve freedom from her family. THEME: The most inclusive yet concise inference that can be made about the event as a delimiting of the pattern of the event by omission of its particulars and relation of it to an aspect of human experience. It is that statement about the work which covers both it and the aspect of human experience to which the particular work is analogous. The theme of "Old Mortality" is the way in which children embody the traits of the older generation even in rebelling against it. "Old Mortality” is a statement of how this happens in a particular instance. That statement of_ is what is unique about the work, but themes are recurrent; they are what en able us to recognize the work as analogous even though it is unique. Individual themes appear again and again as the themes of particular works. Indeed, the title of "Old Mor- _talityi_sugg.es.ts_this_r.ecurr.ence_of_i.ts_theme_elsewhere.,---- 223 since it is also the title of a novel by Scott. IMPLICATION: A generalization about an aspect of human experience which the event or part of it can be taken to corroborate. There is only one correct delimitation of the theme, but there are any number of implications. The theme is about the work; an implication, however, is about life. Like an inference, an implication may be from any part of the work and on any scale. An implication of Maria's mis understanding inferred above is that children are literal- ists and find it difficult to grasp the metaphoric thinking of adults. An implication of the motivation inferred for Miranda1s hasty marriage is that women often marry early anc suddenly as a rebellion against their families. Another im plication is that Southern society encourages a romantic fatalism. BIBLIOGRAPHY Abrams, M. H. The Mirror and the Lamp. New York, 1953. Ackerman, James S. "A Theory of Style," Journal of Aes thetics and Art Criticism, 20:227-237, Spring 1962. Allen, Don Cameron. "Style and Certitude," ELH, 15:167-175, September 1948. 'Allen, Harold B., ed. Readings in Applied English Linguis tics . New York, 1958. Alonso, Amado. Materia y forma en poesia. Madrid, 1955. . Poesia y estilo de Pablo Neruda. Buenos Aires, 1940; 2nd ed., 1951. . "The Stylistic Interpretation of Literary Texts," Modern Language Notes, 57:489-496, November 1942. Alonso, Amado, and Raimondo Lida, eds. Introduccion a la estilistica romance. Buenos Aires, 1932; 2nd ed., 1942. Alonso, Damaso. La lengua poetica de Gongora. Madrid, 1935. _. Poesia espanola: Ensayo de metodos y limi- tes estilisticas. Madrid, 1950; 3rd ed., 1957. _. La poesia de San Juan de la Cruz. Madrid, 1942; 3rd ed., 1958. Arnheim, Rudolf, Fred Attneave, David Kraehenbuehl, and Edgar Coons. "Information Theory and the Arts: A Sym posium," Journal of Aesthetics and Art Criticism, 17: 501-524, June 1959. . 2 . 2 . 5 . 226 Arnold, Aerol. "Why Structure in Fiction: A Note to Social Scientists," American Quarterly, 10:325-337, Fall 1958. Aronstein, Philip. "On Style and Styles in Language," American Speech, 9:343-351, December 1934. Auerbach, Erich. Introduction aux etudes de philologie romane. Frankfurt am Main, 1949. ________________. Literatursprache und Publikum in der lateinischen Spatantike und im Mittelalter. Bern, 1958 ________________. Mimesis: The Representation of Reality in Western Literature, trans. Willard R. Trask. Prince ton, 1953. Baldwin, Charles S. Ancient Rhetoric and Poetic: Inter preted from Representative Works. Gloucester, Mass., 1959. ___________________. Medieval Rhetoric and Poetic. New York, 1928. ___________________. Renaissance Literary Theory and Prac tice . New York, 1939. Bally, Charles. Precis stylistique. Geneva, 1905. . Traite de stylistigue franpaise. 2 vols. Heidelberg, 2nd ed., 1919. Barnett, Lincoln. The Universe and Doctor Einstein. New York, rev. ed., 1957. Bate, Walter Jackson. The Stylistic Development of Keats. New York, 1945. 'Bateson, F. W. English Poetry and the English Language. Oxford, 1934. Baum, Pauli F. The Other Harmony of Prose. Durham, N.C., 1952. jBeardsley, Monroe C. Aesthetics: Problems in the Philosophy of Criticism. New York, 1958. 227 ■Beare, William. Latin Verse and European Song: A Study in Accent and Rhythm. London/ 1957. Black, Max. Models and Metaphors. Ithaca, 1962. Bloch, Bernard. "Linguistic Structure and Linguistic Analy sis." In Archibald A. Hill, ed., Fourth Annual Round Table on Linguistics and Language Teaching. Washing ton, D.C., 1953. Bloomfield, Leonard. Language. New York, 1933. ____________________ . "Secondary and Tertiary Responses to Language," Language, 20:45-55, April-June 1944. Boas, George. "The Problem of Meaning in the Arts." In George P. Adams, et al., eds.. Meaning and Interpreta tion . Berkeley, 1950. (University of California Pub lications in Philosophy, vol. 25.) Bockmann, Paul, ed. Stil- und Formprobleme in der Literatur. Heidelberg, 1959. (Vortrage des VII Kongresses der Internationalen Vereinigung fiir moderne Sprachen.) Booth, Wayne C. The Rhetoric of Fiction. Chicago, 1961. Bradley, A. C. "Poetry for Poetry's Sake," Oxford Lectures on Poetry. New York, 1909. Bray, J. W. A History of Critical Terms. Boston, 1898. Brewster, W. T., ed. Representative Essays on the Theory of Style. New York, 1905. Brooke-Rose, Christine. A Grammar of Metaphor. London, 1958. Brooks, Cleanth. The Hidden God. New Haven, 1963. ________________ . The Well-Wrought Urn. New York, 1947. Brooks, Cleanth, and Robert Penn Warren. Fundamentals of Good Writing. New York, 1950. 228 Brooks, Cleanth, and Robert Penn Warren. A Modern Rhetoric. New York, 2nd ed., 1958. _________________________________ . Under s tand ing Poetry. New York, 2nd ed., 1950. Bruneau, Charles. "La stylistique," Romance Philology, 5:1- 14, August 1951. Buffon, Comte de. Discours sur le style. Paris, 1875. Carroll, John B. "The Critical Need in the Study of Lan guage," CCC, 13:23-26, October 1962. ________________. The Study of Language: A Survey of Lin- guistics and Related Disciplines. Cambridge, Mass., 1953. Ceccato, Silvio. "Operational Linguistics and Translation," Methodos, 12:11-80, 1960. Chatman, Seymour. "Linguistics, Poetics, and Interpreta tion: the Phonemic Dimension," Quarterly Journal of Speech, 43:248-256, October 1957. _________________. "Robert Frost's Mowing: An Enquiry into Prosodic Structure," Kenyon Review, 18:421-438, Summer 1956. Cherry, Colin. On Human Communication. New York, 1957. Chomsky, Noam. Syntactic Structures. 1s-Gravenhage, 1957. Christensen, Francis. "A Generative Rhetoric of the Sen tence," CCC, 14:155-161, October 1963. "Notes Toward a New Rhetoric: I. Sentence Openers; II. A Lesson from Hemingway," Col lege English, 25:7-18, October 1963. Clark, Donald Lemen. Rhetoric in Greco-Roman Education. New York, 1957. _______________. Rhetoric and Poetry in the Renais sance: A Study of Rhetorical Terms in English Renais sance Literary Criticism. New York, 1922.____________ 229 Clarke, M. L. Rhetoric at Rome: A Historical Survey. Lon don, 1953. Classe, Andre. The Rhythm of English Prose. Oxford, 1939. Cooper, Lane, ed. Theories of Style. New York, 1907. Craddock, C. E. Style Theories as Found in Stylistic Stud ies of Romance Scholars: 1900-1950. Washington, D.C., 1952. Crane, R. S., ed. Critics and Criticism; Ancient and Modern. Chicago, 1952. __________ . The Languages of Criticism and the Structure of Poetry. Toronto, 1953. Cressot, M. Le style et ses techniques. Paris, 1947. Croce, Benedetto. Aesthetic; As Science of Expression and General Linguistic, trans. Douglas Ainslie. London, 2nd ed., 1922. Croll, Morris W. "The Baroque Style in Prose." In Kemp Malone and M. B. Ruud, eds., Studies in English Philol ogy; A Miscellany in Honor of F. Klaeber. Minneapolis, 1929. ________________. Introduction, Euphues, ed. Harry Clemons. London, 1916. Curtius, Ernst Robert. European Literature and the Latin Middle Ages, trans. Willard R. Trask. New York, 1953. I Daiches, David. "The Literary Use of Language," A Study of Literature. Ithaca, 1948. D'Alton, John F. Roman Literary Theory and Criticism. New York, 1931. Davie, Donald. Articulate Energy; An Inquiry into the Syn tax of English Poetry. London, 1955. _____________. Purity of Diction in English Verse. New York, 1953. ' 2 TO' Devoto, Giacomo. "Introduction a la stylistique," Melanges Marouzeau. Paris, 1948. . Linguistics and Literary Criticism, trans. M. F. Edgerton, Jr. New York, 1963. Dionysus or Longinus. "On Literary Excellence." In Allan H. Gilbert, ed., Literary Criticism: Plato to Dryden. New York, 1940. Dobree, Bonamy. Modern Prose Style. Oxford, 1934. Dresden, S. "Stylistique et science de la litterature," Neophilologus, 36:193-205, October 1952. Duckworth, George E. Structural Proportions in Vergil's Aeneid. Ann Arbor, 1962. Eliot, T. S. Introduction, Ezra Pound Selected Poems. Lon don, 1928. Else, Gerald F. Aristotle's Poetics: The Argument. Cam bridge, Mass., 1957. Elton, William, ed. Aesthetics and Language. New York, 1954. i Empson, William. Seven Types of Ambiguity. London, 1930; rev. ed., 1947. ________________. The Structure of Complex Words. New York, 1951. Entwistle, William J. Aspects of Language. London, 1953. Erlich, Victor. Russian Formalism: History-Doctrine. 's-Gravenhage, 1955. Fergusson, Francis. The Idea of a Theater. Princeton, 1949. Firth, J. R. Papers in Linguistics. London, 1957. ____________. "A Synopsis of Linguistic Theory: 1930-1955." In Studies in Linguistic Analysis. Oxford, 1957. 231 Fisher, Solomon. "Meaning and Structure in Poetry," Journal of Aesthetics and Art Criticism, 13:453-461, June 1956. Flores, Angel, and M. J. Benardete, eds. Cervantes Across the Centuries. New York, 1947. Foster, Richard. The New Romantics: A Reappraisal of the New Criticism. Bloomington, 1962. Francis, W. Nelson. The Structure of American English. New York, 1958. Frank, Paul L. "Historical or Stylistic Periods?" Journal of Aesthetics and Art Criticism, 13:451-457, June 1955. Fries, Charles C. "The Bloomfield 'School.'" In Christine Mohrmann, et al., eds., Trends in European and American Linguistics: 1930-1960. Utrecht and Antwerp, 1961. _________________. The Structure of English. New York, 1952. _________________. "Meaning and Linguistic Analysis," Lan guage , 30:57-68, January-March 1954. Frye, Northrop. Anatomy of Criticism. Princeton, 1957. _____________ed. Sound and Poetry. New York, 1957. _. The Well-Tempered Critic. Bloomington, 1963. Fucks, Wilhelm. "On Mathematical Analysis of Style," Bio met rika , 39:122-129, May 1952. _. "Mathematische Analyse von Sprachelementen, Sprachstil und Sprachen," Arbeitsgemeinschaft fur For- schung des Landes Nordrhein-Westfalen. KSln and Opladen, 1955. Gardiner, Sir Alan. Theory of Speech and Language. Oxford, 1951. Gibson, Walker. "Authors, Speakers, Readers, Mock Readers," College English, 11:265-269, February 1950. 232 Glasgow, W. D. "The Concept of Choosing," Analysis, 20:63- 67, January 1960. Gleason, H. A., Jr. An Introduction to Descriptive Lin guistics . New York, 1955; rev. ed., 1961. . "What Is English?" CCC, 13:1-10, Octo ber 1962. Goethe, Johann Wolfgang. "Einfache Nachahmung der Natur, Manier, Stil." In Goethes Werke, ed. Werner Weber and Hans Joachim Schrimpf, vol. 12. Hamburg, 1953. Goodman, Paul. The Structure of Literature. Chicago, 1954. Gourmont, Remy de. Le probl&me du style. Paris, 1902. Greenlaw, Edwin. "A Better Teacher than Aquinas," Studies in Philology, 14:200-214, April 1917. Guberina, P. "Proced^s stylistiques et stylographiques: analyse scientifique et litteraire," Proceedings of the Sixth Triennial Congress of the International Federa tion for Modern Languages and Literatures. Oxford, 1954. Guiraud, P. La stylistigue. Paris, 1954. _. "Stylistiques," Neophilologus, 38:1-11, January 1954. Hall, Robert A., Jr. "American Linguistics: 1925-1950," I, II, Archivum Linguisticum, 3:101-125, 1951; 4:1-16, 1952. ____________________ . Idealism in Romance Linguistics. Ithaca, 1963. Hamm, Victor M. "Meter and Meaning," PMLA, 69:695-710, September 1954. Harris, Zellig s. "Co-occurrence and Transformations in Linguistic Structure," Language, 33:283-340, July- September 1957. 233 Harris, Zellig S. "Discourse Analysis," Language, 28:1-30, January-March 1952; 28:474-494, October-December 1952. . "Distributional Structure," Word, 10:146- 162, August 1954. . Methods in Structural Linguistics. Chicago, 1951. Hattori, R. S. "The Analysis of Meaning," For Roman Jakob- son: Essays on the Occasion of His Sixtieth Birthday. 's-Gravenhage, 1956. Hatzfeld, Helmut. Bibliografia critica de la nueva estilis- tica aplicada a las literaturas romanicas. Madrid, 1955. _________________ . A Critical Bibliography of the New Styl istics: Applied to the Romance Literatures, 1900-1952. Chapel Hill, N.C., 1953. _________________ . Don Quijote als Wortkunstwerk: Die ein- zelnen Stilmittel und ihr Sinn. Leipzig, 1927. _________________ . Literature Through Art: A New Approach to French Literature. New York, 1952. _________________ . "Methods of Stylistic Investigation," Literature and Science. Oxford, 1954. (Proceedings of the Sixth Triennial Congress of the International Fed eration for Modern Languages and Literatures.) _________________ . "Recent Italian Stylistic Theory and Stylistic Criticism." In A. G. Hatcher and K. L. Selig, eds., Studia philologica et litteraria in hon- orem L. Spitzer. Bern, 1958. _________________ . "Stylistic Criticism as Art-Minded Phil ology," Yale French Studies, 2:62-70, Spring-Summer 1949. Heath, P. L. "Intentions II," Aristotelian Society Supple mentary Volume, 29:147-164, 1955. 234 Heilman, Robert B. "Poetic and Prosaic: Program Notes on Opposite Numbers," Pacific Spectator, 5:454-463, Autumn 1951. 'Hendrickson, G. L. "The Origin and Meaning of the Ancient Characters of Style," American Journal of Philology, 26:249-290, July-September 1905. 'Henning, Edward B. "Patronage and Style in the Arts," Journal of Aesthetics and Art Criticism, 18:464-471, June 1960. Herdan, G. Language as Choice and Chance. Groningen, 1956. __________. Type-Token Mathematics. 1s-Gravenhage, 1960. Hill, Archibald A. "An Analysis of 'The Windhover1: An Ex periment in Structural Method," PMLA, 70:968-978, December 1955. __________________ . "Correctness and Style in English Compo sition," College English, 12:280-285, February 1951. __________________ . Introduction to Linguistic Structures. New York, 1958. __________________ . "Linguistics since Bloomfield," Quar terly Journal of Speech, 41:253-260, October 1955. . "'Pippa's Song': Two Attempts at Struc tural Criticism," University of Texas Studies in Eng lish, vol. 35, 1956. __________________ . "Prescriptivism and Linguistics in Eng lish Teaching," College English, 15:395-399, April 1954. __________________ . "A Program for the Definition of Litera ture," University of Texas Studies in English, vol. 37, 1958. __________________ . "A Sample Literary Analysis," Fourth Annual Round Table on Linguistics and Language Teach ing . Washington, D.C., 1953. 235 Hill, Archibald A. "Towards a Literary Analysis," English Studies in Honor of James Southall Wilson. Charlottes ville, Va., 1951. Hockett, Charles F. A Course in Modern Linguistics. New York, 1958. Hofstadter, Albert. "The Scientific and Literary Uses of Language," Symbols and Society. New York, 1955. ■Hulme, T. E. Notes on Language and Style, ed. Herbert Read. Seattle, 1929. (University of Washington Chapbooks, vol. 25.) Hungerland, Isabel. Poetic Discourse. Berkeley, 1958. (University of California Publications in Philosophy, vol. 33.) Hytier, Jean. "La methode de M. L. Spitzer," Romanic Re view, 41:42-59, February 1950. Jack, Ian. Augustan Satire; Intention and Idiom in English Poetry: 1661-1750. Oxford, 1952. Jacobs, Melville. The Content and Style of Oral Literature. Chicago, 1959. (Viking Foundation Publications in Anthropology, vol. 26.) Jakobson, Roman. "The Metaphoric and Metonymic Poles." In Roman Jakobson and Morris Halle, Fundamentals of Lan guage . 's-Gravenhage, 1956. Joos, Martin. "Homeostasis in English Usage," CCC, 13:18- 22, October 1962. _____________. The Five Clocks. Bloomington, 1962. (Sup plement, International Journal of American Linguistics, vol. 28, no. 2.) _____________. "Linguistic Prospects in the United States." In Christine Mohrmann, et al., eds., Trends in European and American Linguistics: 1930-1960. Utrecht and Ant werp, 1961. Kayser, Wolfgang. Das sprachliche Kunstwerk. Bern, 7th ed. 1961.______________________________________________________ 236 'Kenyon, John S. "Cultural Levels and Functional Varieties of English," College English, 10:31-36, October 1948. Ker, W. P. Form and Style in Poetry, ed. R. W. Chambers. London, 1928. Kline, Morris. Mathematics in Western Culture. New York, 1953. Krieger, Murray. The New Apologists for Poetry. Minne apolis, 1956. Kroeber, A. L. Style and Civilizations. Ithaca, 1957. Lambert, Mildred. "Studies in Stylistics," American Speech, 3:173-181, February 1928r 3:326-333, April 1928; 4:28- 39, October 1929; 4:137-144, December 1929; 4:228-243, February 1929; 4:395-402, June 1929; 4:477-489, August 1929. Danger, Susanne K. Feeling and Form. New York, 1953. __________________ . Philosophy in a New Key. Cambridge, 1942. __________________ . Problems of Art. New York, 1957. Lausberg, Heinrich. Handbuch der literarischen Khetorik: eine Grundlegung der Literaturwissenschaft. Munich, 1960. Levy, Raphael. "A New Credo of Stylistics," Symposium, 3: 321-334, November 1949. Lord, Albert B. The Singer of Tales. Cambridge, Mass., 1960. Ducas, F. L. Style. London, 1955. Dussky, A. E. Tieck's Romantic Irony. Chapel Hill, N.C., 1932. Dutwack, Leonard. "Mixed and Uniform Prose Styles in the Novel," Journal of Aesthetics and Art Criticism, 18: 350-357, March 1960. 237 MacDonald, Margaret. "Language and Fiction I," Aristotelian Society Supplementary Volume, 28:165-184, 1954. Malmstrom, Jean. "The Workshop on Generative Grammar," CCC, 13:26-30, October 1962. Mapes, E. K. "Implications of Some Recent Studies on Style," Revue de litterature comparee, 18:514-533, 1938. Marouzeau, J. "Comment aborder 11 etude du style," Le fran ca is moderne, 11:1-6, January 1943. _. Precis de stylistique frangaise. Paris, 3rd ed., 1950. _______. "Quelques aspects de la question du style," Conferences de 11Institut de Linguistigue de l'Univer- site de Paris, vol. 7, 1939. Martin, Harold C., ed. Style in Prose Fiction. New York, 1959. Martinet, Andre. A Functional View of Language. Oxford, 1962. McDavid, Raven I., Jr. (Rev. of The Five Clocks), College English, 25:233-234, December 1963. McKeon, Richard, ed. Introduction to Aristotle. New York, 1947. ________________. "Rhetoric in the Middle Ages," Speculum, 17:1-32, January 1942. Messing, G. M. "Structuralism and the Literary Tradition," Language, 27:1-12, January-March 1951. Miles, Josephine. The Continuity of Poetic Language. Berkeley, 1951. _________________. Eras and Modes in English Poetry. Berke ley, 1957. ____________. The Vocabulary of Poetry. Berkeley, 1946. 238 Miller, George A. Language and Communication. New York, 1951. Miriam Joseph, C.S.S., Sister. Shakespeare's Use of the Arts of Language. New York, 1947. Monk, Samuel H. The Sublime. Ann Arbor, 2nd ed., 1960. Morgan, Douglas N. "Psychology and Art: A Summary and Cri tique," Journal of Aesthetics and Art Criticism, 9:81- 96, December 1950. Morris, E. P. "A Science of Style," Transactions and Pro ceedings of the American Philological Association, vol. 46, 1915. Muller, Gunther. "Morphological Poetics." In Susanne K. Langer, ed., Reflections on Art. Baltimore, 1958. Munro, Thomas. "Style in the Arts: A Method of Stylistic Analysis," Journal of Aesthetics and Art Criticism, 5:128-153, December 1946. Murry, John Middleton. The Problem of Style. Oxford, 1922. Myers, Henry Alonso. "Style and the Man," South Atlantic J Quarterly, 40:259-268, July 1941. •Nowell-Smith, P. H. "Choosing, Deciding and Doing," Analy sis , 18:63-69, January 1958. D'Connor, D. J. "Possibility and Choice," Aristotelian Society Supplementary Volume, 34:1-24, 1960. O'Connor, William V. An Age of Criticism: 1900-1950. Chicago, 1952. Olson, Elder. "An Outline of Poetic Theory," In R. S. Crane, ed., Critics and Criticism: Ancient and Modern. Chicago, 1952. _____. The Poetry of Dylan Thomas. Chicago, 1954. _________ . Tragedy and the.Theory of Drama. Detroit, 1961. 239 Olson, Elder. "William Empson, Contemporary Criticism, and Poetic Diction." In R. S. Crane, ed., Critics and Criticism: Ancient and Modern. Chicago, 1952. Otto, A . Die Sprichworter und sprichwdrtlichen Redensarten der Rdmer. Leipzig, 1890. Parkin, Rebecca P. "Alexander Pope's Use of the Implied Dramatic Speaker," College English, 11:137-141, Decem ber 1949. Passmore, J. A. "Intentions I," Aristotelian Society Sup plementary Volume, 29:131-146, 1955. Perrin, Porter G. "Style," Writer's Guide and Index to English. Chicago, 3rd ed., 1959. Plath, Warren. "Mathematical Linguistics." In Christine Mohrmann, et al., eds., Trends in European and American Linguistics: 1930-1960. Utrecht and Antwerp, 1961. Potter, Simeon. Modern Linguistics. London, 1957. Raleigh, Sir Walter. Style. London, 1897. Rannie, D. W. The Elements of Style: An Introduction to Literary Criticism. London, 1915. Ray, Punya Sloka. "The Formation of Prose," Word, 18:313- 325, December 1962. _________________. "The Logic of Linguistics," Methodos, 13: 239-254, 1961. Read, Herbert. English Prose Style. Boston, 1928; rev. ed., 1955. Reid, T. B, W. "Linguistics, Structuralism, and Philology," Archivum Linguisticum, 8:28-37, 1956. Retamar, R. F. Idea de la estilistica. Havana, 1958. Richards, I. A. The Philosophy of Rhetoric. New York, 1936. _______________. Practical Criticism. London, 1929.________ 240 Richards, I. A. Principles of Literary Criticism. London, 1924; 2nd ed., 1926. Rickert, Edith. New Methods for the Study of Literature. Chicago, 1927. Rieser, Max. "The Semantic Theory of Art in America," Journal of Aesthetics and Art Criticism, 15:12-26, September 1956. Riffaterre, Michael. "Criteria for Style Analysis," Word, 15:154-174, April 1959. ____________________ . "Vers la definition linguistique du style," Word, 17:318-344, December 1961. Roberts, Murat H. "The Science of Idiom: A Method of In quiry into the Cognitive Design of Language," PMLA, 59: 291-306, March 1944. Roberts, William Rhys. Greek Rhetoric and Literary Theory. New York, 1928. Rothwell, Kenneth S. "Structure in Literature," College English, 24:602-607, May 1963. Rouse, W. H. D. "Style." In N. C. Smith, ed., Essays and Studies by Members of the English Association. Oxford, 1942. Rus, Louis C. "Structural Ambiguity: A Note on Meaning and the Linguistic Analysis of Literature," Language Learn ing, 6:62-67, 1955. Ryle, Gilbert. The Concept of Mind. Oxford, 1949. Saidla, L. E. A. Essays for the Study of Structure and Style. New York, 1939. Saisselin, Remy G. "Buffon, Style, and Gentlemen," Journal of Aesthetics and Art Criticism, 16:357-361, March 1958. Sanford, F. H. "Speech and Personality: A Comparative Case Study," Character and Personality, 10:169-198, December- 1942. 243} Sapir, Edward. Language. New York, 1921. Saussure, Ferdinand de. Cours de linguistigue generale. Paris, 1916; 4th ed., 1949. Sayce, Richard A. "The Definition of the Term 'Style,'" Proceedings of the Illrd Congress of the International Comparative Literature Association. 's-Gravenhage, 1962. . "Literature and Language," Essays in Criticism, 7:119-133, April 1957. _________________ . Style in French Prose. Oxford, 2nd ed., 1957. _________________ . "Style in Language," Archivum Linguisti- cum, 14:77-83, 1962. Schapiro, Meyer. "Style." In A. L. Kroeber, et al., eds., Anthropology Today. Chicago, 1953. Schlauch, Margaret. The Gift of Language. New York, 1955. Schorer, Mark. "Fiction and the 'Analogical Matrix.'" In John W. Aldridge, ed., Critigues and Essays in Modern Fiction. New York, 1952. Scriven, Michael. "Language in Fiction II," Aristotelian Society Supplementary Volume, 28:185-196, 1954. Sebeok, Thomas A., ed. Style in Language. New York, 1960. Seidler, Herbert. Allgemeine Stilistik. Gottingen, 1953. Shipley, Joseph T., ed. Dictionary of World Literature. New York, 1953. Skinner, B. F. Verbal Behavior. New York, 1957. Sledd, James. "Coordination (Faulty) and Subordination (Up side-Down)," CCC, 7:181-187, December 1956. ____________ _ . A Short Introduction to English Grammar. Chicago, 1959. 242 Smith, G. Gregory. Elizabethan Critical Essays. 2 vols. Oxford, 1937. Solmsen, Friedrich, ed. The Rhetoric and the Poetics of Aristotle. New York, 1954. Spitzer, Leo. "Answer to Mr. Bloomfield (Language 20, 45)," Language 20:245-251, October-December 1944. . Aufsatze zur romanische Syntax und Stilistik. Halle, 1918. . (Rev. of Allgemeine Stilistik), Comparative Literature, 8:146-149, Spring 1956. _____________. (Rev. of Style in the French Novel) , Compara tive Literature, 10:368-371, Fall 1958. _____________. Essays on English and American Literature, ed. Anna Hatcher. Princeton, 1962. ___________ . "Les etudes de style et les different pays," Langue et litterature. Paris, 1960. (Actes du VIIIe Congres de la Federation Internationale des Langues et Litteratures Moderne, vol. 31.) _____________. Linguistics and Literary History: Essays in Stylistics. Princeton, 1948. ______ _____ . A Method of Interpreting Literature. North ampton, Mass., 1949. ___________ _. Romanische Literaturstudien. Ttibingen, 1959. I ____________ . Romanische Stil- und Literaturstudien. 2 vols. Marburg, 1931. _____________. Stilstudien. 2 vols. Munich, 1928. _____________. "Les theories de la stylistique," Le frangais moderne, 20:165-168, July 1952. _____________. Die Wortbildunq als stilistisches Mittel ex- emplifiziert an Rabelais. Halle, 1910. (Beihefte zur Zeitschrift fur romanische Philologie, No. 19.) . . ^ Spoerri, Theophil. "Die stilkritische Methode," Die Form- werdung des Menschen. Berlin, 1938. Stearns, Marshall. The Story of Jazz. Oxford, 1956. Straumann, Heinrich. "Between Literary Criticism and Seman tics," English Studies, 36:254-262, October 1955. Sturtevant, Edgar H. An Introduction to Linguistic Science. New Haven, 1947. Stutterheim, C. F. P. "Modern Stylistics," Lingua, 1:410- 426, September 1948; 3:52-68, February 1952. Tate, Allen, ed. The Language of Poetry. Princeton, 1942. Tillyard, E. M. W., and C. S. Lewis. The Personal Heresy: A Controversy. London, 1934. Thompson, John. The Founding of English Meter. New York, 1961. Tuve, Rosemond. Elizabethan and Metaphysical Imagery. Chicago, 1947. Ullmann, Stephen. The Image in the Modern French Novel. Cambridge, 1960. _________________ . The Principles of Semantics. Glasgow and Oxford, 2nd ed., 1957. _________________. "Psychologie et stylistique," Journal de psychologie, 46:133-156, April-June 1953. _________________ . Semantics: An Introduction to the Science of Meaning. Oxford, 1962. _________________ . Style in the French Novel. Cambridge, 1957. Utley, Francis Lee. "Structural Linguistics and the Liter ary Critic," Journal of Aesthetics and Art Criticism, 18:319-328, March 1960. Van Ghent, Dorothy. The English Novel: Form and Function. _____New York. 1953___________ ' ______________________________ 244 Verschoor, J. "Parole, langue et les deux stylistiques: Questions de terminologie," Neophiloloqus, 39:184-191, July 1955. Voegelin, C. F., and J. Yegerlehner. "Toward a Definition of Formal Style, with Examples from Shawnee." In W. E. Richmond, ed., Studies in Folklore. Bloomington, 1957. Vossler, Karl. Gesammelte Aufscttze zur Sprachphilosophie. Munich, 1923. ______________. Introduccion a la estilistica romance. Buenos Aires, 1932; 2nd ed., 1942. ______________. Positivismus und Idealismus in der Sprach- wissenschaft. Heidelberg, 1904. Waldock, A. J. A. Paradise Lost and Its Critics. Cambridge. 1947. Walsh, Dorothy. "The Poetic Use of Language," Journal of Philosophy, 35:73-81, January 1938. Waterman, John T. Perspectives in Linguistics. Chicago, 1963 . Watson, George. The Literary Critics. Baltimore, 1962. Weaver, Richard M. The Ethics of Rhetoric. Chicago, 1953. Wellek, Rene. Concepts of Criticism, ed. Stephen G. Nichols, Jr. New Haven, 1962. _____________. A History of Modern Criticism, 1750-1950: Vol. I, The Later Eighteenth Century. New Haven, 1955. ____________ . A History of Modern Criticism, 1750-1950: Vol. II, The Romantic Age. New Haven, 1955. ____________ . "Leo Spitzer (1887-1960)," Comparative Liter ature, 12:310-334, Fall 1960. Wellek, Rene, and Austin Warren. Theory of Literature. New York, 1949; 2nd ed., 1956. 245 Werkmeister, W. H. An Introduction to Critical Thinking. Lincoln, Nebraska, rev. ed., 1957. Werner, Heinz, ed. On Expressive Language. Worcester, Mass., 1955. Whatmough, Joshua. Language; A Modern Synthesis. New York, 1956. . Poetic, Scientific, and Other Forms of Discourse. Berkeley, 1956. Wheelwright, Philip. The Burning Fountain. Bloomington, 1954. ______________________ . Metaphor and Reality. Bloomington, 1962. Whitehall, Harold. "English Verse and What It Sounds Like; From Linguistics to Criticism," Kenyon Review, 18:411- 421, Summer 1956. - ____■ ___________ . "From Linguistics to Criticism; A Book Review" (rev. of An Outline of English Structure), Kenyon Review, 13;710-714, Fall 1951. Whitehall, Harold, and Archibald A. Hill. "A Report on the Language-Literature Seminar." In Harold B. Allen, ed., Readings in Applied English Linguistics. New York, 1958. Whitehead, Alfred North. Alfred North Whitehead; An Anthol ogy, ed., F. S. C. Northrop and Mason W. Gross. New York, 1953. Williams, C. B. "A Note on the Statistical Analysis of Sen- tence-Length as a Criterion of Literary Style," Bio- metrika, 31:356-361, March 1940. Williamson, George. The Senecan Amble: A Study of Prose Form from Bacon to Collier. Chicago, 1951. Wilson, Edmund. Axel's Castle. New York, 1936. 2461 Wimsatt, W. K., Jr. Philosophic Words: A Study of Style and Meaning in the Rambler and Dictionary of Samuel John son. New Haven, 1948. _. The Prose Style of Samuel Johnson. New Haven, 1941. - _______________. The Verbal Icon. Lexington, Ky., 1954. Wimsatt, W. K., Jr., and Cleanth Brooks. Literary Criti cism: A Short History. New York, 1957. Winkler, Emil. Grundlegung der Stilistik. Leipzig, 1929. Yule, G. U. The Statistical Study of Literary Vocabulary. Cambridge, 1944.

Linked assets