Version classiqueVersion mobile

Letters of Blood and other works in English

Göran Printz-Påhlson

The Words of the Tribe: Primitivism, Reductionism, and Materialism in Modern Poetics

Part Two. Linguistic Reductionism in Poetry Criticism

Texte intégral


1Taking my cue from the two quotes from Wordsworth at the end of my first essay—quotes about the dangers of words as instruments for good or evil and about language as a counterspirit—I would like to consider another pervasive attitude towards language: the rejection of words as a means of communication and the postulating of a pre-verbal stage of mankind. Such a stage may be imagined as Adamic and set in Eden, or as something brutish, set in the woods where once the Noble Savage ran and communicated without recourse to words. This pre-verbal stage is a commonplace in the Western mystical tradition (and no doubt elsewhere) from Jacob Boehme to Swedenborg. This attitude has been admirably illuminated by Gerald Bruns in his book Modern Poetry and the Idea of Language, one of the few investigations of the area that seems to follow a path more or less parallel to my own.

2That language can be considered a veil to be lifted or rent asunder in order that man may glimpse a reality behind the words is a notion as old and as un-provable a hypothesis as is its opposite, the idea that words have some special or privileged relation to reality. ‘Language was given to us in order to disguise our thoughts’ is an old cynical apophthegm, probably coined by Voltaire but popularly attributed to Talleyrand (Edwards 50), who clearly had some experience in the practice of it. The idea was employed by Edward Young, and later by Kierkegaard. In a different, weaker form it can be traced back to Plutarch. The warning that Plato offers to Theaitetos, about the misológoi—those who mistrust words—is clearly intended to stave off criticisms from this kind of radically anti-verbal position.

3It is not clear whether this anti-verbalism should be referred to as linguistic primitivism, since it involves a negation of language in primitive times or extreme conditions (as in, for example, the ecstasy of the mystic). Is the language of the angels or the animals to be equated with human language? Or, to convert the question into more modern terms, are the analogical extensions of natural language to be taken seriously? Or are they to be regarded as merely metaphorical descriptions of secondary communication systems, organised according to entirely different principles? Once Sigurd (or Siegfried) had slain the dragon Fafnir in the Norse version of the Niebelungen saga, he sat down to roast the dragon’s heart over the fire. He burnt a finger when testing the temperature and put it in his mouth. Instantly, he could understand the language of the birds. What dragon do we have to slay, what heart do we have to devour, in order to learn to decipher the cryptic messages of poetry and art? Some believe that we can move beyond the various languages and means of understanding used to interpret poetry, and find the single master-discourse—that we can eat the heart of the dragon.

4This vaguely primitivist notion of a more primary language, or pre-language, has another side to it, of equally venerable age. From ancient times, man has tried to decipher the messages inscribed in nature herself, has tried to read the Book of Nature. In particular, starry skies have for millennia fascinated observers who attempted, in order to form higher unities in the seeming chaos, to discern discrete elements (constellations) either to elucidate the physical/historical structure of the universe (cosmology) or to demarcate the influence of heavenly bodies on the individual fates of men (astrology). Like the Book of the World, the script of the stars became an inordinately popular commonplace in Western literature—we see this, for example, in E.R. Curtius’s European Literature and the Latin Middle Ages, in Gabriel Josipovici’s The World and the Book, as well as countless quotes from Edward Young, Ralph Waldo Emerson, and the romantic poets.

5When Galileo, in his important methodological dialogue ‘Sidereus Nuntius’ (‘The Messenger from the Stars’) brings this ancient metaphor into the focus of his experimental and quantitative methodology, he can claim with enormous authority that he has found the key to the decipherment of this sidereal text in a language: in the language of circles and triangles, of numbers and functions, in the abstract languages of geometry and mathematics.

6The position reached here by Galileo can perhaps be described as a kind of primitivism, but an ahistorical and firmly methodological or procedural one. It presupposes that every order of the phenomenal world can be ultimately and accurately described, and also predicted, in a language that has been purified of all its contingent qualities. It is a special kind of primitivism, though: Galileo’s method is an enormously authoritative statement of reductionism in the form of science.


7In our time ‘reductionism’ is mainly used as a term of criticism or abuse, so it may be difficult to remember what a powerful idea in the history of Western civilisation the principle of reducibility has been. The idea of a unified science, where every phenomenon could be studied and explained and classified, so that humanistic study and the social sciences could be reduced to biology, biology to chemistry and chemistry to physics in a hierarchical scale, eventually leaving no discernable gaps at all, is one of the most grandiose schemes for the advancement of human knowledge ever conceived. Ever since the mid-19th century, with the all-embracing explanatory aspirations of Comte and Marx, it has been a burning issue in the social sciences, closely connected with the problem of positivism—more so perhaps on the European continent than in the English-speaking world.

8Of course the anti-reductionist doctrine, which states that every science or area of knowledge is dependant on its own language and terminology and can explain phenomena only within its boundaries, is as old if not older than reductionism itself. Aristotle makes a very clear and strong case for it in a chapter of the Posterior Analytics. Here, one presumes, he is wary of the claims of the Pythagoreans and Plato on behalf of the sciences of mathematics and geometry, claims not reductionist in the modern sense, but which aspired nonetheless to elevate these forms of knowledge to a privileged position in the Greek paideía or educational system:

Nor can a proposition of one science be proved by another science except when the relation is such that the propositions of the one are subordinate to those of the other, as the propositions of optics are subordinate to geometry and those of harmonics to arithmetic. (63)

9Although Aristotle’s term episteme may not exactly correspond to what we mean by a separate science today, the same principle is clearly applicable to any non-reductionist standpoint.

10The best statement I know of the problem of reduction, or alternatively, of irreducibility, in the natural sciences from recent times comes is from Sir Peter Medawar’s Induction and Intuition in Scientific Thought:

Reducibility; emergence: If we choose to see a hierarchical structure in Nature—if societies are composed of individuals, individuals of cells, and cells in their turn of molecules, then it makes sense to ask whether we may not ‘interpret’ sociology in terms of the biology of individuals or ‘reduce’ biology to physics and chemistry. This is a living methodological problem, but it does not seem to have been satisfactorily resolved. At first sight the ambition embodied in the idea of reducibility seem hopeless of achievement. Each tier of the natural hierarchy makes use of notions peculiar to itself. The ideas of democracy, credit, crime or political constitution are no part of biology, nor shall we expect to find in physics the concepts of memory, infection, sexuality, or fear. No sensible usage can bring the foreign exchange deficit into the biology syllabus, already grievously overcrowded, or nest building into the syllabus of physics. In each plane or tier of the hierarchy new notions or ideas seem to emerge that are inexplicable in the language or with the conceptual resources of the tier below. But if in fact we cannot ‘interpret’ sociology in terms of biology or biology in terms of physics, how is it that so many of the triumphs of modern science seem to be founded upon a repudiation of the doctrine of irreducibility? There is a problem here to which methodologists can and do make valuable and illuminating contributions. (15–16)

11This admirably clear and succinct account is followed by an even longer footnote where Medawar, like Aristotle, uses geometry as a prime example of reductionism.


12I hope to be excused for such a lengthy opening gambit, but I think these reflections by an eminent scientist are well worth our attention. The importance of Medawar’s insights—in particular in the notion of the emergence of theoretical concepts—can be brought to bear on the problems of modern poetics on several different levels. Present-day discussions of reductionism have most often been kept within the confines of Althusser’s theory of science, which I consider to be one of the weakest links in his often-convincing criticism of empiricism. Althusser’s attitude to the separate sciences is conventionalist—like that of the early Foucault— and as his criteria for truth conditions are coherence criteria rather than correspondence criteria, his theory does not seem to be powerful enough to exclude spurious areas of investigation. It cannot cope with the emergence of new science or new sub-fields for science.

13Such spurious fields have no doubt emerged in profusion in the area of poetics, both recently and in former times, and one might be able to see a Wordsworthian counter-spirit at work in their language, in order to ‘lay waste, vitiate and dissolve.’ To lay waste the reader’s imagination, to vitiate his judgement and dissolve his patience, that is. As my intentions are not polemical, I shall refrain from engagement in these areas but only address myself to works by scholars whose important contributions are serious and beyond doubt. I will, however, reiterate the point made about irreducibility, by both Aristotle and Medawar—the point about the way notions or ideas or demonstrations of proof are transferable from level to level, but only hierarchically. ‘Thus the genus must be the same, either absolutely, or in some respect, if the demonstration is to be transferable [metabaínein],’ Aristotle says (63). Later he gives an example, brief and to the point, culled from the not too adjacent fields of geometry and poetics: ‘Is every circle a figure?—if one draws a circle the answer is obvious. Well, are the epic poems (epe kuklos) a circle? Evidently they are not’ (81).

14There are some facts about poetry that seem unassailable and incontrovertible, and which have been repeated by generations of poets; for example, that poems are made up of words and not emotions, an idea respected by Eliot and Pound after Valéry.

15Eliot would probably have agreed with Michael Riffaterre, the French-American semiotician and stylistic scholar, when he writes in his Semiotics of Poetry:

The language of poetry differs from common linguistic usage—this much the unsophisticated reader senses instinctively. Yet, while it is true that poetry often employs words excluded from common usage and has its own special grammar, even a grammar not valid beyond the narrow compass of a given poem, it may also happen that poetry uses the same words and the same grammar as everyday language. (1)

16Dryden, the young Wordsworth, and the youngish Eliot would have been surprised. But Riffaterre goes on to say:

In all literatures with a long enough history, we observe that poetry keeps swinging back and forth, tending first one way, then the other. The choice between alternatives is dictated by the evolution of taste and by continually changing aesthetic concepts. But whichever of the two trends prevails, one factor remains constant: poetry expresses concepts and things by indirection. To put it simply, a poem says one thing and means another. (ibid.)

17The later Eliot would indubitably have agreed with the historicist point made in the beginning of the paragraph, but I am not so sure about the conclusion. Riffaterre describes the poet as someone who says one thing and means another. But there is a much simpler term for that: a liar (pseudomenos). So, is poetry a lying activity, perhaps the art of lying? This has been maintained by very respectable people, such as Xenophanes and Plato. But another interpretation of saying one thing and meaning another would be irony, the poet as ironist—the first among the Theophrastian characters, and not much better than a liar.


18Let me return to a previous point before reverting to Riffaterre whose book, I must insist, I find a very substantial contribution to the understanding of one type of poetry. The idea that poetry is a specific and specialised language, and that all the arts are best described as languages, is very widespread at the present time. That the arts are languages has been maintained by Susanne Langer, and from a very different philosophical standpoint by Nelson Goodman and his disciples. Roland Barthes sees the fashion of dress as language, Susan Sontag sees illness as a metaphor, Jacques Lacan sees the subconscious as a language. Examples could easily be multiplied. I shall not ask the question whether this is sound theoretical practice or not; I shall just look at the consequences of some cases closely connected with poetics or the systematic view of poetics in modernism.

19I am not here so much concerned about the practical linguistic reductionism of methodology as in explication de texte or close reading and the like. Clearly, the poem, qua isolated poem, exists on a phenomenological level, a level that is in a very obvious sense linguistic. But in order to maintain that poetry is a specific language over and above its linguistic constituents, perhaps a universal language, one needs much more on which to go. Firstly, one needs a separate grammar of some complexity, consisting of, say, one syntactical, one morphological, and one phonological component, each with a separate set of rules—be they phrase structure rules or transformational rules. Furthermore, one needs a lexicon, and in addition one needs separate components for converting prose into poetry and re-converting poetry into prose. Nobody has claimed to have even outlined such an unwieldy grammar with any consistency.

20The terminology is familiar to anybody who has attended an elementary course in modern linguistics—the framework I cited was one associated with generative grammar at a fairly early stage of its development. There is no evidence that any of the many rival theories we now have would allow any easier access for a sub-theory of poetic language. However the terminology is also familiar to anybody who has dipped into modern critical theory or poetics where these terms abound, used most often in a loose and completely irresponsible way, together with many other, sometimes completely incompatible terms. It is no secret that the structuralist and poststructuralist methods favoured in recent years have had their most convincing successes in the analysis of narrative—where a respectable terminology has been engendered from within the discipline—but seem very inadequate for the analysis of poetry.

21One of the most consistent theoretical accounts of poetic language has been the one given by the Soviet semiotician Jurij Lotman. His sense of the theoretical problems involved in regarding poetry as a language are quite clear, and he has tried to solve these problems by treating the languages of art as secondary modelling systems, that is, as systems which, although modelled on natural languages, at the same time tend to generate further sub-languages. Lotman is far from clear on this point, in spite of his fairly simplistic views on his languages as codes. His translator writes in the preface to the English translation:

Is literature a language, or cinema, or sculpture? Is a given period or school (Realism, the Baroque) a language? Is a specific genre within these art forms a language? Is a specific text a language? At various points in his work Lotman responds affirmatively to all these questions, and the reader must himself determine what the author means by language in each case. (Vroon, x)

22One of Lotman’s most interesting points comes in his attempt to integrate a phonological and a semantic component for his grammar of poetic language. In the course of this attempt he comes to regard the poetic line as unit of both sound and meaning, as a separate and unique word. The view has its venerable antecedents, obviously quite independently, in a passage from Mallarmé’s Crise de vers (Crisis of Verse), where the latter speaks of ‘le vers qui de plusieurs vocables refait un mot total, neuf, étranger à la langue et comme incantatoire’ [‘the verse that from its constituents makes up a total word, new, strange to the language and like an incantation’] (213).

23Yvor Winters quotes Mallarmé’s passage in the beginning of Primitivism and Decadence, with certain disapproving comments. With his customary shrewdness and consistency, he goes on to say:

The poem, to be perfect, should likewise be a new word in the same sense, a word of which the line, as we have defined it, is merely a syllable. Such a word is, of course, composed of much more than the sum of its words (as one normally uses the term) and its syntax. It is composed of an almost fluid complex, if the adjective and the noun are not too nearly contradictory, of relationships between words (in the normal sense of the term), a relationship involving rational content, cadences, rhymes, juxtapositions, literary and other connotations, inversions, and so on, almost indefinitely. These relationships, it should be obvious, extend the poet’s vocabulary incalculably. (3)

24This should be the logical conclusion of Jurij Lotman’s position as well, although, sadly, he never descends from high abstraction to make such a conclusion explicit.

25This position can indeed be regarded as a first step toward a vocabulary, or lexicon, of poetic utterances, a list of hapax legomena or nonce-words that, barring formulaic repetitions and cases of plagiarism, is infinite and, as Winters put it, incalculable. The position could also be regarded as a step towards formulating some morphological rules for the poetic language, or subordinate parts of it, like alliteration, assonance, rhyme, and the like.


26If we leave the phonological aspect aside (as we have to do for quite respectable ‘languages’ like logic and mathematics) in an account of poetic language, we can still secure a syntactic component for the grammar, which would order the elements according to certain rules of selection and combination. This is the well-known principle enunciated by Roman Jakobson for his method in poetics, expounded in many ruthlessly multilingual articles.

27We all have reason to be grateful that Roman Jakobson, arguably the greatest and most versatile linguist of this century, devoted so much time and energy to questions of poetics. His painstakingly detailed linguistic analyses of poems from many languages unearth many real insights into the workings of language in a poem. But these analyses shed light on the poem by entirely conventional means. In his theoretical statements Jakobson’s methodology is as reductionist on the micro-stylistic level as is Lotman on a macro-stylistic level. To maintain that every element in a poem is either opposed to or in parallel with any other element does not give much scope for systematic observation on how, and to what purpose, poets actually work. This is what is ultimately left out of any linguistically reductionist poetics.

28I shall not labour this point here any more, but I do wish to emphasise that I find reductionist claims honourable, and see them as partaking of a long tradition of scientific aspirations. I do not believe that linguistic reductionist methods destroy the poetic element, nor that there is a mystical essence in poetry (any more than in any other activity) that cannot be described and analysed. I also believe that advances in textual grammar and pragmatics—what was once known as discourse analysis— will eventually shed much light on the details of poetic composition and appreciation. However, I cannot be convinced that poetry is a language in the sense required for this kind of analysis to be successful on a theoretical level. Nor do I think that description can be used in a reductive sense about poetry and the arts.

29What, though, about the middle ground between the microscopic and the macroscopic levels? Let us return to Riffaterre. His very impressive ability to read poems in a sensitive way, along with his intimate acquaintance with French poetry and colloquial French, leads to a methodology that is far from absurd, but is nevertheless curiously limiting. His main view of French poetry—a healthy corrective to the idea that French and continental modernism tends towards the lofty, elevated, and prophetic—is that it is in many ways related to riddles, jokes, and conundrums. He uses an elaborate terminology that I am not convinced is always necessary, and which I am not going to reproduce here. Taking a hint from Saussure’s anagrams and paragrams, he postulates for his chosen poems the existence of a hypogram which is hidden from the reader at first reading, and may consist in a cliché, proverbial phrase, or perhaps most often another previous line in a poem or other text. It is signalled by a certain deviation of language that he calls ungrammatical (in obvious contradiction of current usage in linguistics). He achieves some very impressive results in detailed and revealing poems by Eluard, Desnos, early Breton, and prose and poetry by Chateaubriand, Hugo, Rimbaud, and Lautréamont. His remarks on the prose poem as a genre are acute (although not, I think, exhaustive) and his explications of Ponge are among the best I have ever read. Still, the model implied for the making of poetry, as opposed to the reading of poetry, is very curious. Riffaterre sees the poem as emerging from a matrix that is ‘a minimal and literal sentence,’ then transforming into a longer, complex and non-literal periphrasis. Thus his primary category of text-production is expansion, which is a variant of the old amplificatio technique: more and more detailed information is added recursively to the matrix until the poem is complete (19).

30The objection that can be immediately raised is that the recursivity of this technique cannot provide any criteria for its completeness or closure: in theory the poem could be amplified ad infinitum. And, furthermore, the technique of expansion seems to be contrary to the insistence in modernist poetics on the opposite technique, contraction or concentration, as defended by Pound, for instance in his neat ‘dichten = condensare’ (97).

31The only important rhetorical category that can be easily handled by ordinary rules of transformation, however, is the transformation of simile to metaphor. This operation, which is part of a general strategy of concentration, has always been regarded as an important step in the conversion of poetic diction of the rhetorical variety to a more unforced and natural language, associated in various instances with romanticism and modernism.

32The position of Riffaterre is not the deliberately absurd one of many present-day critics of poetry. He has a firm belief in the accessibility of poetic discourse to rational investigation. Yet his methodology and (perhaps, as importantly) his choice of poems have forced him into an overall view of poetry that is curiously antipathetic towards the claims of romantic and modernist poets of the last two centuries. In spite of his strictly objective and scientific attitude, he seems to be feeding his theory with definitions that are persuasive, but favour a specific kind of poetry seen from a limited point of view. Here we can see how linguistic reductionism, however legitimate from a certain point of view, can result in the trivialisation of poetry by reducing it to a game or a conundrum. There is evidence that this idea of ludic poetry is a serious contender for a conspicuous place in postmodernist poetics: I am thinking, for example, of the so-called ‘Martian School’ of British poetry.


33It would be demonstratively unfair to include Leo Spitzer, with his truly modest attitude towards his life-long service to philology (beautifully expressed in the Princeton lecture of 1948, ‘Linguistics and Literary History’) among the linguistic reductionists. He shows himself to be, in this statement of hermeneutic principles, a sworn enemy of reductionism. I will, however, mention him here because I want to refer, however briefly, to his favourite example of fickleness and unstableness, in the words conundrum, quandary, ‘calembour,’ ‘calembredaine,’ etc., which are all derived from similar roots. In this connexion he makes a most ingenious excursion into ‘Cratylisme’—in Genette’s mimological sense: ‘Thus we must conclude that the instability is also connected with a semantic content: a word meaning ‘whim,’ ‘pun’ easily behaves whimsically—just as in all languages throughout the world, the words for ‘butterfly’ present a kaleidoscopic instability’ (7). Spitzer takes this as a hint of the instability that threatens the work of art, and he concludes:

The reason that the clues to understanding can not be mechanically transferred from one work of art to another lies in the fact of artistic expressivity itself. […] To overcome the impression of an arbitrary association in the work of art, the reader must seek to place himself in the creative center of the artist himself—and re-create the artistic organism. (28)

34This activating element in art and poetry, and the problems involved in applying organic models to it, will be, with the benevolent blessing of Spitzer, a leitmotif for my next two essays.


Works Cited

Aristotle. Posterior Analytics. Trans. by Hugh Tredennick. Cambridge, MA: Harvard University Press, 1930, II.

Curtius, E.R. European Literature and the Latin Middle Ages. Trans. by W.R. Trask. London: Routledge, 1953.

Edwards, H. Sutherland. ‘Historic Phrases,’ MacMillan’s Magazine, 35 (1876–1877): 48–54.

Galilei, Galileo. Sidereus Nuntius. Trans. by Albert van Helden. Chicago: University of Chicago Press, 1989.

Josipovici, Gabriel. The World and the Book. Basingstoke: Macmillan, 1971.

Mallarmé, Stéphane. Oeuvres complètes, 2 vols. Paris: Pleiade, 1945, II.

Medawar, Sir Peter. Induction and Intuition in Scientific Thought. Philadelphia: American Philosophical Society, 1969.

Pound, Ezra. ABC of Reading. New York: New Directions: 1960.

Riffaterre, Michael. Semiotics of Poetry. Bloomington, Indiana: Indiana University Press, 1978.

Spitzer, Leo. Linguistics and Literary History: Essays in Stylistics. New York: Russell and Russell, 1948.

Vroon, Ronald. ‘Preface’ to Jurij Lotman, The Structure of the Artistic Text. Trans. by Ronald Vroon. Ann Arbor: University of Michigan Press, 1977.

Winters, Yvor. Primitivism and Decadence: A Study of American Experimental Poetry New York: Haskell House, 1969.


Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search