LANGUAGE AND GNOSIS
1
TRANSLATION exists because men speak different languages. This truism is, in fact, founded on a situation which can be regarded as enigmatic and as posing problems of extreme psychological and socio-historical difficulty. Why should human beings speak thousands of different, mutually incomprehensible tongues? We live in this pluralist framework, have done so since the inception of recorded history, and take the ensuing farrago for granted. It is only when we reflect on it, when we lift the facts from the misleading context of the obvious, that the possible strangeness, the possible ‘unnaturalness’ of the human linguistic order strikes us. Conceivably there is here one of the more central questions in the study of man’s cerebral and social evolution. Yet even the pertinent queries, the statements of astonishment which would put the facts into relief, are formulated only sporadically. Divisions between formal ‘hard-edged’ linguistics on the one hand and contrastive, anthropological investigations of actual language on the other, have further relegated the issue into the shadow of futile, metaphysical speculation.
We ought not, perhaps, to regard as either formally or substantively coherent, as responsible to verification or falsification, any model of verbal behaviour, any theory of how language is generated and acquired, which does not recognize as crucial the matter of the bewildering multiplicity and variousness of languages spoken on this crowded planet. In his foreword to Morris Swadesh’s posthumous The Origin and Diversification of Language, Dell Hymes states: ‘The diversity of languages, as they have developed and been adapted, is a patent fact of life that cries out for theoretical attention. It becomes increasingly difficult for theorists of language to persist in confounding potential equivalence with actual diversity.’ This should have been a commonplace and respectable exigence among linguists long before 1972. Theories of semantics, constructs of universal and transformational grammar that have nothing of substance to say about the prodigality of the language atlas—more than a thousand different languages are spoken in New Guinea—could well be deceptive. It is here, rather than in the problem of the invention and understanding of melody (though the two issues may be congruent), that I would place what Lévi-Strauss calls le mystère suprême of anthropology.
Why does homo sapiens, whose digestive tract has evolved and functions in precisely the same complicated ways the world over, whose biochemical fabric and genetic potential are, orthodox science assures us, essentially common, the delicate runnels of whose cortex are wholly akin in all peoples and at every stage of social evolution—why does this unified, though individually unique mammalian species not use one common language? It inhales, for its life processes, one chemical element and dies if deprived of it. It makes do with the same number of teeth and vertebrae. To grasp how notable the situation is, we must make a modest leap of imagination, asking, as it were, from outside. In the light of anatomical and neurophysiological universals, a unitary language solution would be readily understandable. Indeed, if we lived inside one common language-skin, any other situation would appear very odd. It would have the status of a recondite fantasy, like the anaerobic or anti-gravitational creatures in science-fiction. But there is also another ‘natural’ model. A deaf, non-literate observer approaching the planet from outside and reporting on crucial aspects of human appearance and physiological behaviour, would conclude with some confidence that men speak a small number of different, though probably related, tongues. He would guess at a figure of the order of half a dozen with perhaps a cluster of dependent but plainly recognizable dialects. This number would be persuasively concordant with other major parameters of human diversity. Depending on which classification they adopt, ethnographers divide the human species into four or seven races (though the term is, of course, an unsatisfactory shorthand). The comparative anatomy of bone structures and sizes leads to the use of three main typologies. The analysis of human blood-types, itself a topic of great intricacy and historical consequence, suggests that there are approximately half a dozen varieties. Such would seem to be the cardinal numbers of salient differentiation within the species though the individual, obviously, is genetically unique. The development on earth of five or six major languages, together with a spectrum of derivative, intermediary dialects and pidgins, analogous to the gamut and blendings of skin-colour, would strike our imaginary observer as a profoundly natural, indeed inevitable pattern. If we lived within this pattern, we should experience it as inherently logical and take for granted the supporting or at least powerfully analogous evidence of comparative anatomy, physiology, and the classification of races. Under pressure of time and historical circumstance, the half dozen principal languages might well have bent quite far apart. Speakers would nevertheless be conscious of underlying uniformities and would expect to find that degree of mutual comprehension shared, for example, within the Romance-language family.
The actual situation is, of course, totally different.
We do not speak one language, nor half a dozen, nor twenty or thirty. Four to five thousand languages are thought to be in current use. This figure is almost certainly on the low side. We have, until now, no language atlas which can claim to be anywhere near exhaustive. Furthermore, the four to five thousand living languages are themselves the remnant of a much larger number spoken in the past. Each year so-called rare languages, tongues spoken by isolated or moribund ethnic communities, become extinct. Today entire families of language survive only in the halting remembrance of aged, individual informants (who, by virtue of their singularity are difficult to cross-check) or in the limbo of tape-recordings. Almost at every moment in time, notably in the sphere of American Indian speech, some ancient and rich expression of articulate being is lapsing into irretrievable silence. One can only guess at the extent of lost languages. It seems reasonable to assert that the human species developed and made use of at least twice the number we can record today. A genuine philosophy of language and socio-psychology of verbal acts must grapple with the phenomenon and rationale of the human ‘invention’ and retention of anywhere between five and ten thousand distinct tongues. However difficult and generalizing the detour, a study of translation ought to put forward some view of the evolutionary, psychic needs or opportunities which have made translation necessary. To speak seriously of translation one must first consider the possible meanings of Babel, their inherence in language and mind.
Even a cursory look at Meillet’s standard compendium1 or at more recent listings in progress under the direction of Professor Thomas Sebeok of Indiana University, shows a situation of utter intricacy and division. In many parts of the earth, the language-map is a mosaic each of whose stones, some of them minuscule, is entirely or partially distinct from all others in colour and texture. Despite decades of comparative philological study and taxonomy, no linguist is certain of the language atlas of the Caucasus, stretching from Bžedux in the north-west to Rut’ul and Küri in the Tatar regions of Azerbeidjan. Dido, Xwarši, and Qapuči, three languages spoken between the Andi and the Koissou rivers, have been tentatively identified and distinguished, but are scarcely known to any but native users. Arči, a language with a distinctive phonetic and morphological structure, was, in the 1970s, spoken by only one village of approximately 850 inhabitants. Oubykh, once a flourishing tongue on the shores of the Black Sea, survives today in a handful of Turkish localities near Ada Pazar. A comparable multiplicity and diversity marks the so-called Palaeosiberian language families. Eroded by Russian during the nineteenth century, Kamtchadal, a language of undeniable resource and antiquity, survives in only eight hamlets in the maritime province of Koriak. In 1909, one old man was still conversant in the eastern branch of Kamtchadal. In 1845, a traveller came across five speakers of Kot (or Kotu). Today no living trace can be found. The history of Palaeosiberian cultures and migrations before the Russian conquest is largely obscure. But evidence of great linguistic variance and sophistication is unmistakable. With regard to nuances of action—possibility, probability, confirmation, necessity—Palaeosiberian languages possess a grammar of obvious precision. But we know little of the genesis of these tongues and of their affinities, if any, with other major linguistic groupings.
The Black Sea region and even Russian Siberia are well known; both have been involved in recorded history and in the spread of technology. By comparison, the language-map extending from the south-western United States to Tierra del Fuego is full of blanks and mere guesses. The fundamental divisions are uncertain: what, for instance, are the relations between the enormously ramified Uto-Aztec tree of languages and the great Mayan cluster? For Mexico and Central America alone, current listings reckon 190 distinct tongues. But the roll is incomplete, and entire language groups are designated as unclassified, as possibly extinct, or as identifiable only through hearsay and through their intrusions, in the guise of quotations and borrowings, into other idioms. The mind must be complacent to regard this situation without a radical sense of perplexity.
Tubatulabal was spoken by something like a thousand Indians at the southern spur of the Sierra Nevada as recently as the 1770s. All we know today is that this language was strikingly different from all neighbouring tongues. Kupeño survived into the late eighteenth century, but already then it was dwindling to a small patch of territory at the sources of the San Luis Rey. What may have been its wider past? What models of human similitude and cultural determination will account for the fact that Huite (or Yecarome), still spoken on the Rio Fuerte in the sixteenth century, should have been sharply different from the Cahita languages, themselves a branch of the Hopi family, which literally surrounded it? Mid-sixteenth-century travellers reported the currency of Matagalpa throughout north-west Nicaragua and in parts of present-day Honduras. Now only a handful of families living near the modern towns of Matagalpa and Esteli are thought to know the speech. In northern Mexico and along the Pacific coast, Nawa and then Spanish submerged a score of ancient, separate human tongues. Tomateka, Kakoma, Kučarete—these are now ghost names. Again, an intimation of enigmatic needs and energies crowds upon one.
Blank spaces and question marks cover immense tracts of the linguistic geography of the Amazon basin and the savannah. At latest count, ethno-linguists discriminate between 109 families, many with multiple sub-classes. But scores of Indian tongues remain unidentified or resist inclusion in any agreed category. Thus a recently discovered tongue spoken by Brazilian Indians of the Itapucuru river territory seems to be related to no previously defined set. Puelče, Guenoa, Atakama, and a dozen others are names designating languages and dialects spoken, perhaps over millions of square miles, by migrant and vanishing peoples. Their history and morphological structure are barely charted. Many will dim into oblivion before rudimentary grammars or word-lists can be salvaged. Each takes with it a storehouse of consciousness.
The language catalogue begins with Aba, an Altaic idiom spoken by Tatars, and ends with Zyriene, a Finno-Ugric speech in use between the Urals and the Arctic shore. It conveys an image of man as a language animal of implausible variety and waste. By comparison, the classification of different types of stars, planets, and asteroids runs to a mere handful.
What can possibly explain this crazy quilt? How are we to rationalize the fact that human beings of identical ethnic provenance, living on the same terrain, under equal climatic and ecological conditions, often organized in the same types of communal structure, sharing kinship systems and beliefs, speak entirely different languages? What sense can be read into a situation in which villages a few miles apart or valleys divided by low, long-eroded hills use tongues incomprehensible to each other and morphologically unrelated? I put the question repetitively because, for a long time, obviousness has disguised its extreme importance and difficulty.
A Darwinian scheme of gradual evolution and ramification, of adaptive variation and selective survival, may look credible. Consciously or not, many linguists seem to have worked with some such analogy. But it only masks the problem. Though many details of the actual evolutionary process remain obscure, the strength of Darwin’s argument lies in the demonstrable economy and specificity of the adaptive mechanism; living forms mutate with seemingly random profusion, but their survival depends on adjustment to natural circumstance. It can be shown, over a wide range of species, that extinction does relate to a failure or inexactitude of vital response. The language manifold offers no genuine counterpart to these visible, verifiable criteria. We have no standards (or only the most conjectural) by which to assert that any human language is intrinsically superior to any other, that it survives because it meshes more efficiently than any other with the demands of sensibility and physical existence. We have no sound basis on which to argue that extinct languages failed their speakers, that only the most comprehensive or those with the greatest wealth of grammatical means have endured. On the contrary: a number of dead languages are among the obvious splendours of human intelligence. Many a linguistic mastodon is a more finely articulated, more ‘advanced’ piece of life than its descendants. There appears to be no correlation, moreover, between linguistic wealth and other resources of a community. Idioms of fantastic elaboration and refinement coexist with utterly primitive, economically harsh modes of subsistence. Often, cultures seem to expend on their vocabulary and syntax acquisitive energies and ostentations entirely lacking in their material lives. Linguistic riches seem to act as a compensatory mechanism. Starving bands of Amazonian Indians may lavish on their condition more verb tenses than could Plato.
The Darwinian parallel also breaks down on the crucial point of large numbers. The multiplicity of fauna and flora does not represent randomness or waste. It is an immediate factor of the dynamics of evolutionary breeding, cross-fertilization, and competitive selection which Darwin set out. Given the range of ecological possibilities, the multiplication of species is, quite conceivably, economical. No language is demonstrably adaptive in this sense. None is concordant with any particular geophysical environment. With the simple addition of neologisms and borrowed words, any language can be used fairly efficiently anywhere; Eskimo syntax is appropriate to the Sahara. Far from being economic and demonstrably advantageous, the immense number and variety of human idioms, together with the fact of mutual incomprehensibility, is a powerful obstacle to the material and social progress of the species. We will come back to the key question of whether or not linguistic differentiations may provide certain psychic, poetic benefits. But the many ways in which they have impeded human progress are clear to see. No conceivable gain can have accrued to the crowded, economically harried Philippine islands from their division by the Bikol, Chabokano, Ermitano, Tagalog, and Wray-waray languages (to name only the most prominent of some thirty tongues), or from the related fact that for four of these five idioms the United States Employment Service can list only one qualified translator. Numerous cultures and communities have passed out of history as linguistic ‘drop-outs.’ Not because their own particular speech was in any way inadequate, but because it prevented communication with the principal currents of intellectual and political force. Countless tribal societies have withered inward, isolated by language barriers even from their near neighbours. Time and again, linguistic differences and the profoundly exasperating inability of human beings to understand each other have bred hatred and reciprocal contempt. To the baffled ear, the incomprehensible parley of neighbouring peoples is gibberish or suspected insult. Linguistically atomized, large areas of Africa, India, and South America have never gathered their common energies either against foreign predators or economic stagnation. Though sometimes sharing a lingua franca, such as Swahili, their consciousness of kinship and common need has remained artificial. The deeper springs of action stay rooted in linguistic separateness. Robbed of their own language by conquerors and modern civilization, many underdeveloped cultures have never recovered a vital identity. In short: languages have been, throughout human history, zones of silence to other men and razor-edges of division.
Why this destructive prodigality?
Few modern linguists, with the exception of Swadesh and Pei, have shown the curiosity which this situation ought to arouse. Where an answer is given at all, it is put in casually evolutionary terms: there are many different tongues because, over long stretches of time, societies and cultures split apart and, through accretion of particular experience, evolved their own local speech habits. The facile nature of such an explanation is worrying: it fails to engage precisely those central philosophical and logical dilemmas which spring from the admitted uniformities of human mental structures and from the economically and historically negative, often drastically damaging, role of linguistic isolation. Turn the argument around: let reasons be given why the adoption by the human race of a single language or a small number of related languages would have been natural and beneficial. It appears at once that post hoc justifications for the facts as we know them are wholly unconvincing. The problem lies deeper. And few linguists since Wilhelm von Humboldt, in the early decades of the nineteenth century, have thought about it at the required level of psychological insistence and historical sensibility. It was before Humboldt that the mystery of many tongues on which a view of translation hinges fascinated the religious and philosophic imagination.
No civilization but has its version of Babel, its mythology of the primal scattering of languages.1 There are two main conjectures, two great attempts at solving the riddle via metaphor. Some awful error was committed, an accidental release of linguistic chaos, in the mode of Pandora’s box. Or, more commonly, man’s language condition, the incommunicados that so absurdly divide him are a punishment. A lunatic tower was launched at the stars; Titans savaged one another and of their broken bones came the splinters of isolated speech; eavesdropping, like Tantalus, on the gossip of the gods, mortal man was struck moronic and lost all remembrance of his native, universal parlance. This corpus of myth, springing from a very ancient, obstinate bewilderment, modulates gradually into philosophic and hermetic speculation. The history of such speculation, of the endeavours of philosophers, logicians, and illuminati to explain the confusion of human idioms, is itself a compelling chapter in the annals of the imagination. Much of it is turgid stuff. The argument is shot through with fantastications and baroque torsions. Stemming, as it must, from a meditation on its own shell of being, words focused on the mirror and echoing surface of words, the metaphoric and esoteric tradition of philology often loses touch with common sense. But via arcane images, Kabbalistic and emblematic constructs, through occult etymologies and bizarre decodings, the argument on Babel will feel its way—as did the partially astrological, Pythagorean hypotheses of celestial motion in Copernicus and Kepler—towards cardinal insights. More justly amazed than modern linguistics at the whole business of man’s estrangement from the speech of his fellow man, the tradition of language mysticism and philosophic grammar reaches out to intuitions, to deeps of inquiry, which are, I think, often lacking from current debate. Today we move on drier but shallower ground.
Key images and lines of conjecture recur in the philosophy of language from the Pythagoreans to Leibniz and J. G. Hamann. We are told that the substance of man is bound up with language; the mystery of speech characterizes his being, his mediate place in the sequence leading from the inanimate to the transcendent order of creation. Language is assuredly material in that it requires the play of muscle and vocal cords; but it is also impalpable and, by virtue of inscription and remembrance, free of time, though moving in temporal flow. These antinomies or dialectical relations, which I want to look at systematically in the next chapter, confirm the dual mode of human existence, the interactions of physical with spiritual agencies. The occult tradition holds that a single primal language, an Ur-Sprache lies behind our present discord, behind the abrupt tumult of warring tongues which followed on the collapse of Nimrod’s ziggurat. This Adamic vernacular not only enabled all men to understand one another, to communicate with perfect ease. It bodied forth, to a greater or lesser degree, the original Logos, the act of immediate calling into being whereby God had literally ‘spoken the world.’ The vulgate of Eden contained, though perhaps in a muted key, a divine syntax—powers of statement and designation analogous to God’s own diction, in which the mere naming of a thing was the necessary and sufficient cause of its leap into reality. Each time man spoke he re-enacted, he mimed, the nominalist mechanism of creation. Hence the allegoric significance of Adam’s naming of all living forms: ‘and whatsoever Adam called every living creature, that was the name thereof.’ Hence also the ability of all men to understand God’s language and to give it intelligible answer.
Being of direct divine etymology, moreover, the Ur-Sprache had a congruence with reality such as no tongue has had after Babel or the dismemberment of the great, enfolding serpent of the world as it is recounted in the mythology of the Carib Indians. Words and objects dovetailed perfectly. As the modern epistemologist might put it, there was a complete, point-to-point mapping of language onto the true substance and shape of things. Each name, each proposition was an equation, with uniquely and perfectly defined roots, between human perception and the facts of the case. Our speech interposes itself between apprehension and truth like a dusty pane or warped mirror. The tongue of Eden was like a flawless glass; a light of total understanding streamed through it. Thus Babel was a second Fall, in some regards as desolate as the first. Adam had been driven from the garden; now men were harried, like yelping dogs, out of the single family of man. And they were exiled from the assurance of being able to grasp and communicate reality.
Theologians and metaphysicians of language strove to attenuate this second banishment. Had there not been a partial redemption at Pentecost, when the gift of tongues descended on the Apostles? Was not the whole of man’s linguistic history, as certain Kabbalists supposed, a laborious swing of the pendulum between Babel and a return to unison in some messianic moment of restored understanding? Above all, what of the Ur-Sprache itself: had it been irretrievably lost? Here speculation hinged on the question of the veritable nature of Adam’s tongue. Had it been Hebrew or some even earlier version of Chaldaean whose far lineaments could be made out in the names of stars and fabled rivers? Jewish gnostics argued that the Hebrew of the Torah was God’s undoubted idiom, though man no longer understood its full, esoteric meaning. Other inquirers, from Paracelsus to the seventeenth-century Pietists, were prepared to view Hebrew as a uniquely privileged language, but itself corrupted by the Fall and only obscurely revelatory of the Divine presence. Almost all linguistic mythologies, from Brahmin wisdom to Celtic and North African lore, concurred in believing that original speech had shivered into seventy-two shards, or into a number which was a simple multiple of seventy-two.1 Which were the primal fragments? Surely if these could be identified, diligent search would discover in them lexical and syntactic traces of the lost language of Paradise, remnants equitably scattered by an incensed God and whose reconstruction, like that of a broken mosaic, would lead men back to the universal grammar of Adam. If they did indeed exist, these clues would be deep-hidden. They ought to be ferreted out, as Kabbalists and adepts of Hermes Trismegistus sought to do, by scrutinizing the hidden configurations of letters and syllables, by inverting words and applying to ancient names, particularly to the diverse nominations of the Creator, a calculus as intricate as that of chiromancers and astrologers. The stakes were very high. If man could break down the prison walls of scattered and polluted speech (the rubble of the smashed tower), he would again have access to the inner penetralia of reality. He would know the truth as he spoke it. Moreover, his alienation from other peoples, his ostracism into gibberish and ambiguity, would be over. The name of Esperanto has in it, undisguised, the root for an ancient and compelling hope.
Starting with Genesis 11:11 and continuing to Wittgenstein’s Investigations or Noam Chomsky’s earliest, unpublished paper on morphophonemics in Hebrew, Jewish thought has played a pronounced role in linguistic mystique, scholarship, and philosophy. To both Jew and gentile, the text of the Books of Moses had a revealed character unlike that of any later body of language. Thus Hebrew has served time and again as the diamond edge of the cutter’s tool. In Jewish tradition we find those rubrics that will largely organize the main directions of Western argument about the essence and enigmatic dismemberment of human tongues. Each element of the received text has generated its own traditions of study in Jewish mysticism and rabbinical scholarship.1 There is a philology and gnosis of the individual Hebrew letter as there is of the word and grammatical unit. In Merkabah mysticism, each written character may be regarded as embodying a fragment of the universal design of creation; all human experience, no less than all human discourse unto the end of time, is graphically latent in the letters of the alphabet. Those numinous letters whose combinations make up the seventy-two names of God may, if they are probed to the hidden core of meaning, reveal the cipher, the configurations of the cosmos. Accordingly, prophetic Kabbalism developed its ‘science of the combination of letters.’ Through self-hypnotic meditation on groupings of individual characters, groupings which need not in themselves be meaningful, the initiate may come to glimpse the great Name of God, manifest throughout the lineaments of nature, but enveloped, as it were, in the muffling layers of vulgate speech. But although Hebrew may have a privileged immediacy, the Kabbalist knows that all languages are a mystery and ultimately related to the holy tongue.
In German Hasidism, it is the word rather than the alphabetic sign whose hidden sense and unaltered preservation are of extreme importance. To mutilate a single word in the Torah, to set it in the wrong order, might be to imperil the tenuous links between fallen man and the Divine presence. Already the Talmud had said: ‘the omission or the addition of one letter might mean the destruction of the whole world.’ Certain illuminati went so far as to suppose that it was some error of transcription, however minute, made by the scribe to whom God had dictated holy writ, that brought on the darkness and turbulence of the world. Theosophy, as expressed in the Zohar and in the commentaries which followed, made use of mystical puns and word-games to prove some of its crucial doctrines. Elohim, the name of God, unites Mi, the hidden subject, with Eloha, the hidden object. The dissociation of subject from object is the very infirmity of the temporal world. Only in His name do we discern the promise of ultimate unity, the assurance of man’s release from the dialectic of history. In brief: God’s actual speech, the idiom of immediacy known to Adam and common to men until Babel, can still be decoded, partially at least, in the inner layers of Hebrew and, perhaps, in other languages of the original scattering.
The habits of feeling shown in these occult semantics are remote and often bizarre. But at several points, linguistic gnosis touches on decisive issues of a rational theory of language and of translation. There is a deceptively modern ring to the discriminations between deep structures of meaning, structures buried by time or masked by colloquialism, and the surface structures of spoken idiom. There is an acute understanding, essential to any treatment of communication within and between languages, of the ways in which a text may conceal more than it conveys. There is, above all, a clear sense, persistent in Spinoza as it is in Wittgenstein, of the numinous as well as problematic nature of man’s life in language.
Numerous elements of gnostic speculation, often with reference to Hebrew, are evident in the great tradition of European linguistic philosophy. This sequence of visionary belief and conjecture extends unbroken from Meister Eckhart in the early fourteenth century to the teachings of Angelus Silesius during the 1660s and 1670s. Here also we find a stubborn wonder about the multiplicity and splintering of vernaculars. For Paracelsus, writing in the 1530s, there is little doubt that Divine providence shall one day restore the unity of human tongues. His contemporary, the Kabbalist Agrippa of Nettesheim, spun an arcane web around the figure seventy-two; in Hebrew, and particularly in Exodus with its seventy-two designations of the Divine name, magic forces were compacted. One day other languages would return to this fount of being. In the meantime, the very need for translation was like the mark of Cain, a witness to man’s exile from harmonia mundi. There was, as Coleridge knew, no deeper dreamer on language, no sensibility more haunted by the alchemy of speech, than Jakob Böhme (1575–1624).1 Like Nicholas of Cusa long before him, Böhme supposed that the primal tongue had not been Hebrew, but an idiom brushed from men’s lips in the instant of the catastrophe at Babel and now irretrievably disjected among all living speech (Nettesheim had, at one point, argued that Adam’s true vernacular was Aramaic). Being erratic blocs, all languages share in a common myopia; none can articulate the whole truth of God or give its speakers a key to the meaning of existence. Translators are men groping towards each other in a common mist. Religious wars and the persecution of supposed heresies arise inevitably from the babel of tongues: men misconstrue and pervert each other’s meanings. But there is a way out of darkness: what Böhme calls ‘sensualistic speech’—the speech of instinctual, untutored immediacy, the language of Nature and of natural man as it was bestowed on the Apostles, themselves humble folk, at Pentecost. God’s grammar sounds through echoing Nature, if only we will listen.
Kepler agreed that primal speech lay scattered. But it was not in the rough parlance of the primitive and uneducated that the sparks of Divine significance could be found. It was in the immaculate logic of mathematics and in the harmonics, also mathematical in essence, of instrumental and celestial music. The music of the spheres and of Pythagorean accords proclaimed, as they will in the Prologue to Goethe’s Faust, the hidden architecture of Divine speech. In the visionary musings of Angelus Silesius (Johann Scheffler), Böhme’s intimations are carried to extremes. Reaching back to the mysticism of Eckhart, Angelus Silesius asserts that God has, from the beginning of time, uttered only a single word. In that single utterance all reality is contained. The cosmic Word cannot be found in any known tongue; language after Babel cannot lead back to it. The bruit of human voices, so mysteriously diverse and mutually baffling, shuts out the sound of the Logos. There is no access except silence. Thus, for Silesius, the deaf and dumb are nearest of all living men to the lost vulgate of Eden.
In the climate of the eighteenth century these gnostic reveries faded. But we find them again, changed into model and metaphor, in the work of three modern writers. It is these writers who seem to tell us most of the inward springs of language and translation.
Walter Benjamin’s ‘Die Aufgabe des Übersetzers’ dates from 1923.1 Though influenced by Goethe’s comments on translation in his famous notes to the Divan, and by Hölderlin’s treatment of Sophocles, Benjamin’s essay derives from the gnostic tradition. Benjamin posits, as he will throughout his extraordinarily refined, recreative work as exegetist, as ‘secret sharer’ of the poet’s intent, that those who ‘understand’ a text have largely missed its essential significance. Bad translations communicate too much. Their seeming accuracy is limited to what is non-essential in the fabric of the original. Benjamin’s approach to the question of translatability—can the work be translated at all? if so, for whom?—is Kabbalistic:
one might speak of a life or a moment as ‘unforgettable’ even if all men had forgotten it. If its essence required that it not be forgotten, then that assertion would not be false: it would only point to a requirement not satisfied by man and, simultaneously, to a realm in which it could be satisfied: the memory of God. By the same token, the question of the translatability of certain works would remain open even if they were untranslatable for man. And indeed, given an exacting concept of translation, should this not be the case to some extent? It is in the light of such an analysis that one can ask whether a given work of literature requires translation. The relevant proposition is this: if translation is a form, then the condition of translatability must be ontologically necessary to certain works.
Echoing Mallarmé, but in terms obviously derived from the Kabbalistic and gnostic tradition, Benjamin founds his metaphysic of translation on the concept of ’universal language.’ Translation is both possible and impossible—a dialectical antinomy characteristic of esoteric argument. This antinomy arises from the fact that all known tongues are fragments, whose roots, in a sense which is both algebraic and etymological, can only be found in and validated by ‘die reine Sprache.’ This ‘pure language’—at other points in his work Benjamin will refer to it as the Logos which makes speech meaningful but which is contained in no single spoken idiom—is like a hidden spring seeking to force its way through the silted channels of our differing tongues. At the ‘messianic end of their history’ (again a Kabbalistic or Hasidic formulation), all separate languages will return to their source of common life. In the interim, translation has a task of profound philosophic, ethical, and magical import.
A translation from language A into language B will make tangible the implication of a third, active presence. It will show the lineaments of that ‘pure speech’ which precedes and underlies both languages. A genuine translation evokes the shadowy yet unmistakable contours of the coherent design from which, after Babel, the jagged fragments of human speech broke off. Certain of Luther’s versions of the Psalms, Hölderlin’s recasting of Pindar’s Third Pythian Ode, point by their strangeness of evocatory inference to the reality of an Ur-Sprache in which German and Hebrew or German and ancient Greek are somehow fused. That such fusion can exist, that it must, is proved by the fact that human beings mean the same things, that the human voice springs from the same hopes and fears, though different words are said. Or to put it another way: a poor translation is full of apparently similar saying, but misses the bond of meaning. Philo-logy is love of the Logos before it is a science of differing stems. Luther and Hölderlin move German some distance ‘back’ towards its universal origin. But to accomplish this alchemy, a translation must, in regard to its own language, retain a vital strangeness and ‘otherness.’ Very little in Hölderlin’s Antigone is ‘like’ ordinary German; Marianne Moore’s readings of La Fontaine are thorn-hedges apart from colloquial American English. The translator enriches his tongue by allowing the source language to penetrate and modify it. But he does far more: he extends his native idiom towards the hidden absolute of meaning. ‘If there is a language of truth, in which the final secrets that draw the effort of all thinking are held in silent repose, then this language of truth is—true language. And it is precisely this language—to glimpse or describe it is the only perfection the philosopher can hope for—that is concealed, intensively, in translations.’ As the Kabbalist seeks the forms of God’s occult design in the groupings of letters and words, so the philosopher of language will seek in translations—in what they omit as much as in their content—the far light of original meaning. Walter Benjamin’s summation derives directly from the mystic tradition: ‘For in some degree, all great writings, but the Scriptures in the highest degree, contain between the lines their virtual translation. The interlinear version of the Scriptures is the archetype or ideal of all translation.’
His loyalties divided between Czech and German, his sensibility drawn as it was, at moments, to Hebrew and to Yiddish, Kafka developed an obsessive awareness of the opaqueness of language. His work can be construed as a continuous parable on the impossibility of genuine human communication, or, as he put it to Max Brod in 1921, on ‘the impossibility of not writing, the impossibility of writing in German, the impossibility of writing differently. One could almost add a fourth impossibility: the impossibility of writing.’ Kafka often extended the latter to include the illusions of speech. ‘Is it her singing that enchants us,’ asks the narrator in ‘Josephine the Singer, or the Mouse Folk,’ ‘or is it not rather the solemn stillness enclosing her frail little voice?’ And ‘In the Penal Colony,’ perhaps the most desperate of his metaphoric reflections on the ultimately inhuman nature of the written word, Kafka makes of the printing press an instrument of torture. The theme of Babel haunted him: there are references to it in almost every one of his major tales. Twice he offered specific commentaries, in a style modelled on that of Hasidic and Talmudic exegesis.
The first occurs in his allegory on the building of the Great Wall of China, written in the spring of 1917. The narrative relates the two structures, though ‘according to human reckoning’ the purposes of the Wall were the very contrary to those of the insolent Tower. A scholar has written a strange book asserting that the destruction of Babel did not result from the causes generally alleged. Nimrod’s edifice had fallen simply because its foundations had been defective. The sage argues that the Great Wall shall, itself, serve as plinth for a new Tower. The narrator confesses that he is bewildered. How can the Wall, being at most a semicircle, become a foundation for a Tower? Yet there must be some truth to the bizarre suggestion: architectural drawings for the Tower, albeit shadowy, are included among those for the Wall. And there are detailed proposals regarding the required labour force and gathering of nations. That gathering figures in ‘Das Stadtwappen’ (‘The City Arms’), a brief parable which Kafka wrote in the autumn of 1920. This is among his most riddling texts. The first sentence refers to the presence of interpreters (Dolmetscher) on the building site. As no generation of men can hope to complete the high edifice, as engineering skills are constantly growing, there is time to spare. More and more energies are diverted to the erection and embellishment of the workers’ housing. Fierce broils occur between different nations assembled on the site. ‘Added to which was the fact that already the second or third generation recognized the meaninglessness, the futility (die Sinnlosigkeit) of building a Tower unto Heaven—but all had become too involved with each other to quit the city.’ Legends and ballads have come down to us telling of a fierce longing for a predestined day on which a gigantic fist will smash the builders’ city with five blows. ‘That is why the city has a fist in its coat of arms.’
It would be fatuous to propose any single decoding or equivalence of meaning for Kafka’s uses of Babel. That is not how his method of anagogic and allegoric anecdote works. The Talmud, which is often Kafka’s archetype, refers to the forty-nine levels of meaning which must be discerned in a revealed text. But it is evident that Kafka saw in the Tower and its ruin a dramatic shorthand through which to convey certain exact, though not wholly articulate, intimations about man’s linguistic condition and the relations of that condition to God. The Tower is a necessary move: it arises from some undeniable surge of human will and intelligence. The word Himmelsturmbau embodies a puzzling duality: the Tower is, as Genesis proclaims, an assault on Heaven (Sturm), but it is also a vast Jacob’s ladder of stone (Turm) on which man would ascend towards his Creator. Rebellion and worship are inextricably mixed, as are the impulses of speech to lead towards and away from the truth. The foundations of the Tower preoccupy Kafka even more than the edifice itself. ‘The Burrow,’ his last story, and an unmistakable comment on the relation of the writer to language and reality, shows how the Tower may be seen from its interior, spiralling galleries. Hence the uncanny remark in one of Kafka’s notebooks: ‘We are digging the pit of Babel.’ But what are the concordances between the Tower and the Great Wall, which is usually in Kafka a symbol of the Mosaic Law? What are we to make of the precise shift in verb tenses in the final lines of ‘Das Stadtwappen:’ sagas ‘came from the city,’ presumably long ago, but ‘the city has a fist in its coat of arms’? That of Prague happens not to have a fist but two towers. In all these allusions the menace of language and the mystery of its divided state are present. Another notebook entry may come nearest to being a summary of the range of paradox and tragic dialectic which Kafka concentrated in the emblem of the Tower: ‘Had it been possible to build the Tower of Babel without ascending it, that would have been allowed.’ If man could use language without pursuing meaning to the forbidden edge of the absolute, he might still be speaking a veritable and undivided tongue. Yet to use language without translations, without seeking out the hidden springs of the Law is also impossible, and perhaps prohibited. In Kafka speech is the paradoxical circumstance of man’s incomprehension. He moves in it as in an inner labyrinth.
Labyrinths, circular ruins, galleries, Babel (or Babylon) are constants in the art of our third modern Kabbalist.1 We can locate in the poetry and fictions of Borges every motif present in the language mystique of Kabbalists and gnostics: the image of the world as a concatenation of secret syllables, the notion of an absolute idiom or cosmic letter—alpha and aleph—which underlies the rent fabric of human tongues, the supposition that the entirety of knowledge and experience is prefigured in a final tome containing all conceivable permutations of the alphabet. Borges advances the occult belief that the structure of ordinary sensate time and space interpenetrates with alternative cosmologies, with consistent, manifold realities born of our speech and of the fathomless free energies of thought. The logic of his fables turns on a refusal of normal causality. Gnostic and Manichaean speculation (the word has in it an action of mirrors)1 provide Borges with the crucial trope of a ‘counter-world.’ Contrary streams of time and relation blow like high, silent winds through our unstable, itself perhaps conjectural, habitat. No poet has imaged with more density of life the possibility that our existence is being ‘dreamt elsewhere,’ that we are the mere figure of another’s speech, hurtling towards the close of that single, inconceivably vast utterance in which Jakob Böhme heard the sound of the Logos. As Borges writes in ‘Compass’:
All things are words of some strange tongue, in thrall
To Someone, Something, who both day and night
Proceeds in endless gibberish to write
The history of the world. In that dark scrawl
Rome is set down, and Carthage, I, you, all,
And this my being which escapes me quite,
My anguished life that’s cryptic, recondite,
And garbled as the tongues of Babel’s fall.
(Richard Wilbur’s translation)
There were times when Kafka felt the multiplicity of languages to be a gag in his throat. Borges moves with a cat’s sinewy confidence and fun between Spanish, ancestral Portuguese, English, French, and German. He has a poet’s grip on the fibre of each. He has rendered a Northumbrian bard’s farewell to Saxon English, ‘a language of the dawn.’ The ‘harsh and arduous words’ of Beowulf were his before he ‘became a Borges.’ ‘Deutsches Requiem’ is not only as near as we get to a metamorphic realization of the murderous need which bound Nazi to Jew; in voice and narrative gist the story is also as German as those black woods. Though Borges’s Spanish is often private and Argentine, he is possessed of the specific grain of the language, of the invariants which relate his own poetry to ‘Seneca’s black Latin.’ But keen as is Borges’s sense of the irreducible quality of each particular tongue, his linguistic experience is essentially simultaneous and, to use a Coleridgean notion, reticulative. Half a dozen languages and literatures interweave. Borges uses citations and literary–historical references, often invented, to establish the key, the singular locale of his verse and fables. Close-woven, these diverse idioms and legacies—the Kabbala, the Anglo-Saxon epic, Cervantes, the French symbolists, the dreams of Blake and De Quincey—constitute a mapping, a landscape of recognitions unique to Borges but also, somehow, familiar as sleep. Quick with interchange and mutation, Borges’s several languages move towards a unified, occult truth (the Aleph glimpsed on the nineteenth step in the cellar of Carlos Argentino’s house) as do the individual letters of the alphabet in the ‘cosmic library’ of one of the most secret of his ficciones.
‘The Library of Babel’ dates from 1941. Every element in the fantasia has its sources in the ‘literalism’ of the Kabbala and in gnostic and Rosicrucian images, familiar also to Mallarmé, of the world as a single, immense tome. ‘The universe (which others call the Library) is composed of an indefinite, perhaps an infinite number of hexagonal galleries.’ It is a beehive out of Piranesi but also, as the title indicates, an interior view of the Tower. ‘The Library is total and … its shelves contain all the possible combinations of the twenty-odd orthographic symbols (whose number, though vast, is not infinite); that is, everything which can be expressed, in all languages. Everything is there: the minute history of the future, the autobiographies of the archangels, the faithful catalogue of the Library, thousands and thousands of false catalogues, a demonstration of the falsehood of the true catalogue, the Gnostic gospel of Basilides, the commentary on this gospel, the commentary on the commentary of this gospel, the veridical account of your death, a version of each book in all languages, the interpolation of every book in all books.’ Any conceivable combination of letters has already been foreseen in the Library and is certain to ‘encompass some terrible meaning’ in one of its secret languages. No act of speech is without meaning: ‘No one can articulate a syllable which is not full of tenderness and fear, and which is not, in one of those languages, the powerful name of some god.’ Inside the burrow or circular ruins men jabber in mutual bewilderment; yet all their myriad words are tautologies making up, in a manner unknown to the speakers, the lost cosmic syllable or Name of God. This is the formally boundless unity that underlies the fragmentation of tongues.
Arguably, ‘Pierre Menard, Author of the Quixote’ (1939) is the most acute, most concentrated commentary anyone has offered on the business of translation. What studies of translation there are, including this book, could, in Borges’s style, be termed a commentary on his commentary. This concise fiction has been widely recognized for the device of genius which it obviously is. But—and again one sounds like a pastiche of Borges’s fastidious pedantry—certain details have been missed. Menard’s bibliography is arresting: the monographs on ‘a poetic vocabulary of concepts’ and on ‘connections or affinities’ between the thought of Descartes, Leibniz, and John Wilkins point towards the labours of the seventeenth century to construe an ars signorum, a universal ideogram-matic language system. Leibniz’s Characteristica universalis, to which Menard addresses himself, is one such design; Bishop Wilkins’s Essay towards a real character and a philosophical language of 1668 another. Both are attempts to reverse the disaster at Babel. Menard’s ‘work sheets of a monograph on George Boole’s symbolic logic’ show his (and Borges’s) awareness of the connections between the seventeenth century pursuit of an inter-lingua for philosophic discourse and the ‘universalism’ of modern symbolic and mathematical logic. Menard’s transposition of the deca-syllables of Valéry’s Le Cimetière marin into alexandrines is a powerful, if eccentric, extension of the concept of translation. And pace the suave authority of the memorialist, I incline to believe that ‘a literal translation of Quevedo’s literal translation’ of Saint François de Sales was, indeed, to be found among Menard’s papers.
The latter’s masterpiece, of course, was to consist ‘of the ninth and thirty-eighth chapters of the first part of Don Quixote and a fragment of chapter twenty-two.’ (How many readers of Borges have observed that Chapter IX turns on a translation from Arabic into Castilian, that there is a labyrinth in XXXVIII, and that Chapter XXII contains a literalist equivocation, in the purest Kabbalistic vein, on the fact that the word no has the same number of letters as the word sí?) Menard did not want to compose another Quixote ‘which is easy—but the Quixote itself. Needless to say, he never contemplated a mechanical transcription of the original; he did not propose to copy it. His admirable intention was to produce a few pages which would coincide—word for word and line for line—with those of Miguel de Cervantes.’ (So in James E. Irby’s version. Anthony Bonner reads ‘which would be so easy’ and omits ‘a few’ before ‘pages,’ striking what is surely a false note of prolixity.)1
Pierre Menard’s first approach to the task of total translation or, one might more rigorously say, transubstantiation, was one of utter mimesis. But to become Cervantes by merely fighting Moors, recovering the Catholic faith, and forgetting the history of Europe between 1602 and 1918 was really too facile a métier. Far more interesting was ‘to go on being Pierre Menard and reach the Quixote through the experiences of Pierre Menard,’ i.e. to put oneself so deeply in tune with Cervantes’s being, with his onto-logical form, as to re-enact, inevitably, the exact sum of his realizations and statements. The arduousness of the game is dizzying. Menard assumes ‘the mysterious duty’—Bonner, rightly I feel, invokes the notion of ‘contract’—of recreating deliberately and explicitly what was in Cervantes a spontaneous process. But although Cervantes composed freely, the shape and substance of the Quixote had a local ‘naturalness’ and, indeed, necessity now dissipated. Hence a second fierce difficulty for Menard: to write ‘the Quixote at the beginning of the seventeenth century was a reasonable undertaking, necessary and perhaps even unavoidable; at the beginning of the twentieth, it is almost impossible. It is not in vain that three hundred years have gone by, filled with exceedingly complex events. Amongst them, to mention only one, is the Quixote itself’ (Bonner’s ‘that same Don Quixote both complicates and flattens Borges’s intimation). In other words, any genuine act of translation is, in one regard at least, a transparent absurdity, an endeavour to go backwards up the escalator of time and to re-enact voluntarily what was a contingent motion of spirit. Yet Menard’s fragmentary Quixote ‘is more subtle than Cervantes’s.’ How wondrous is Menard’s ability to articulate feelings, thoughts, counsels so eccentric to his own time, to find uniquely appropriate words for sentiments notoriously at variance with those he usually held:
Cervantes’ text and Menard’s are verbally identical, but the second is almost infinitely richer. (More ambiguous, his detractors will say, but ambiguity is richness.)
It is a revelation to compare Menard’s Don Quixote with Cervantes.’ The latter, for example, wrote (part one, chapter nine):
… truth, whose mother is history, rival of time, depository of deeds, witness of the past, exemplar and adviser to the present, and the future’s counsellor.
Written in the seventeenth century, written by the ‘lay genius’ Cervantes, this enumeration is a mere rhetorical praise of history. Menard, on the other hand, writes:
… truth, whose mother is history, rival of time, depository of deeds, witness of the past, exemplar and adviser to the present, and the future’s counsellor.
History, the mother of truth: the idea is astounding. Menard, a contemporary of William James, does not define history as an inquiry into reality, but as its origin. Historical truth, for him, is not what has happened; it is what we judge to have happened. The final phrases—exemplar and adviser to the present, and the future’s counsellor—are brazenly pragmatic.
The contrast in style is also vivid. The archaic style of Menard—quite foreign, after all—suffers from a certain affectation. Not so that of his forerunner, who handles with ease the current Spanish of his time.
Menard’s labours were Herculean. ‘He dedicated his scruples and his sleepless nights to repeating an already extant book in an alien tongue. He multiplied draft upon draft, revised tenaciously and tore up thousands of manuscript pages.’ To repeat an already extant book in an alien tongue is the translator’s ‘mysterious duty’ and job of work. It cannot and must be done. ‘Repetition’ is, as Kierkegaard argued, a notion so puzzling that it puts in doubt causality and the stream of time. To produce a text verbally identical with the original (to make of translation a perfect transcription), is difficult past human imagining. When the translator, negator of time and rebuilder at Babel, comes near succeeding, he passes into that state of mirrors which is described in ‘Borges and I.’ The translator too ‘must live on in Borges’—or in any other author he chooses—‘not in myself—if indeed I am anyone—though I recognize myself less in his books than in many others, or than in the laborious strumming of a guitar.’ A true translator knows that his labour belongs ‘to oblivion’ (inevitably, each generation retranslates), or ‘to the other one,’ his occasion, begetter, and precedent shadow. He does not know ‘which of us two is writing this page.’ In that ‘transubstantial ignorance’—I find no simpler, less unwieldy term—lies the misery of this whole business of translation, but also what repair we can make of the broken Tower.
We shall return to the Kabbalistic motifs and diverse models of translation inferred in the memoir written on the late Pierre Menard of Nîmes by his erudite friend. Irby qualifies the bonfire in which Menard burned his papers as ‘merry;’ Bonner as ‘gay.’ There are two psychologies here, two Christmases, two visions of heresy and of the phoenix.
2
It is via Leibniz and J. G. Hamann that language mysticism enters the current of modern, rational linguistic study. Both men were in active contact with Kabbalistic and Pietist thought.
Linguistic theory bears decisively on the question of whether or not translation, particularly between different languages, is in fact possible. In the philosophy of language two radically opposed points of view can be, and have been asserted. The one declares that the underlying structure of language is universal and common to all men. Dissimilarities between human tongues are essentially of the surface. Translation is realizable precisely because those deep-seated universals, genetic, historical, social, from which all grammars derive can be located and recognized as operative in every human idiom, however singular or bizarre its superficial forms. To translate is to descend beneath the exterior disparities of two languages in order to bring into vital play their analogous and, at the final depths, common principles of being. Here the universalist position touches closely on the mystical intuition of a lost primal or paradigmatic speech.
The contrary view can be termed ‘monadist.’ It holds that universal deep structures are either fathomless to logical and psychological investigation or of an order so abstract, so generalized as to be well-nigh trivial. That all men known to man use language in some form, that all languages of which we have apprehension are able to name perceived objects or to signify action—these are undoubted truths. But being of the class ‘all members of the species require oxygen to sustain life,’ they do not illuminate, except in the most abstract, formal sense, the actual workings of human speech. These workings are so diverse, they manifest so bewilderingly complicated a history of centrifugal development, they pose such stubborn questions as to economic and social function, that universalist models are at best irrelevant and at worst misleading. The extreme ‘monadist’ position—we shall find great poets holding it—leads logically to the belief that real translation is impossible. What passes for translation is a convention of approximate analogies, a rough-cast similitude, just tolerable when the two relevant languages or cultures are cognate, but altogether spurious when remote tongues and far-removed sensibilities are in question.
Between these two poles of argument, there can be numerous intermediary and qualified attitudes. Neither position is maintained often with absolute rigour. There are relativist shadings in the universalist grammars of Roger Bacon, and the grammarians of Port Royal, and even in the transformational generative grammar of Chomsky. Nabokov, who regards all but the most rudimentary of interlinear translations as a fraud, as a facile evasion of radical impossibilities, is himself a master mover between languages. In their modern guise, moreover, both lines of argument can be traced to a common source.
In 1697, in his tract on the amelioration and correction of German, Leibniz put forward the all-important suggestion that language is not the vehicle of thought but its determining medium. Thought is language internalized, and we think and feel as our particular language impels and allows us to do. But tongues differ as profoundly as do nations. They too are monads, ‘perpetual living mirrors of the universe’ each of which reflects or, as we would now put it, structures experience according to its own particular sight-lines and habits of cognition. Yet at the same time, Leibniz had universalist ideals and hopes. Like George Dalgarno, whose Ars Signorum appeared in 1661, and Bishop Wilkins, who published his remarkable Essay towards a real character and a philosophical language in 1668, Leibniz was profoundly interested in the possibilities of a universal semantic system, immediately legible to all men. Such a system would be analogous to mathematical symbolism, so efficacious precisely because the conventions of mathematical operation seem to be grounded in the very architecture of human reason and appear independent of all local variation. It would be analogous also to Chinese ideograms. Once a lexicon of ideograms had been agreed to, all messages could be read instantaneously, whatever the language of the recipient, and the disaster at Babel would, on the graphic level at least, be mended. As we shall see, mathematical symbolism and Chinese writing are, to this day, implied models in almost all discussions of universal grammar and translation.
In Vico’s ‘philology,’ as in Leibniz’s, universalist and ‘monadist’ strains coexist. Philology is the quintessential historical science, the key to the Scienza nuova, because the study of the evolution of language is the study of the evolution of the human mind itself. Vico knows, this is one of his great clairvoyances, that man enters into active possession of consciousness, into active cognizance of reality, through the ordering, shaping powers of language. All men do so, and in that sense language, and metaphor in particular, are a universal fact and a universal mode of being. In the genesis of the human spirit, all nations traverse the same stages of linguistic usage, from the immediate and sensory to the abstract. Simultaneously, however, Vico’s opposition to Descartes and to the extensions of Aristotelian logic in Cartesian rationalism made of him the first true ‘linguistic historicist’ or relativist. He was acutely perceptive of the autonomous genius and historical coloration of different languages. All primitive men sought expression through ‘imaginative universals’ (generi fantastici), but in diverse tongues these universals rapidly acquired very different configurations. ‘Almost infinite particulars’ constitute both the syntactic and lexical corpus of different languages. These particulars both engender and reflect the differing world-views of races and cultures. The degree of ‘infinite particularity’ reaches so deep, that a universal logic of language, on the Aristotelian or Cartesian—mathematical model, is falsely reductionist. It is only by means of a scrupulous, essentially poetic recreation or translation of a given language-world, such as that of Homeric Greek and of Biblical Hebrew, that the ‘new science’ of myth and history can hope to retrace the growth of consciousness (and growths would be more accurate).1
That Goethe, in a remark dated March 1787, compared Hamann to Vico is well known, as is the fact that Hamann had, ten years before, obtained a copy of the Scienza nuova. It remains unlikely, nevertheless, that there was any direct influence. Hamann’s theories on language and culture go back to the very early 1760s. They spring both from the pregnant muddle of his extraordinary intellect and from his intimacy with theosophic and Kabbalistic speculations. Hamann’s notions are usually fragmentary; they are veiled in a diction as ‘radiantly dark’ as was Blake’s. But the originality and foresight of his conjectures on language are, particularly today, uncanny.
From the 1750s onward, the problem of’l’influence réciproque du langage sur les opinions et des opinions sur le langage’ was very much in vogue. Hamann addressed himself to the theme in his Versuch über eine akademische Frage (1760). He affirms that there is a determining concordance between the directions of thought and feeling in a community and ‘the lineaments of its speech.’ Nature has provided different races with different pigmentation and shapes of the eye. Similarly, it has caused in men imperceptible but decisive variations in the formation of lip, tongue, and palate. These variations are the source of the proliferation and diversity of languages. (This physiological hypothesis was not new, and Hamann himself draws on the English anatomist Thomas Willis.) Languages are as figurative of the particular nature of a civilization as are its garb and social rites. Each language is an ‘epiphany’ or articulate revelation of a specific historical–cultural landscape. Hebrew verb forms are inseparable from the niceties and strict punctualities that mark Jewish ritual. But that which a language reveals as being the specific genius of a community, the language itself has shaped and determined. The process is dialectical, with the formative energies of language moving both inward and outward in a civilization.
In 1761, Hamann applied these views to a comparative examination of the grammatical and lexical resources of French and German. Turgid, erratic as they are, the Vermischte Anmerkungen contain premonitions of genius. Though referring itself to Leibniz, Hamann’s opening statement about the close kinship of linguistic and monetary exchanges, and his confident dictum that theories of language and of economics will prove mutually explanatory, are not only strikingly original but set out in nuce much of Lévi-Strauss’s structural anthropology. Hamann is able to argue in this fashion because he is already working towards a general theory of significant signs, towards a semiology in the modern sense. Mystical exegesis underwrote Hamann’s and Leibniz’s belief that a nerve fabric of secret meanings and revelations lies below the surface structure of all languages. To read is to decipher. To speak ‘is to translate (metapherein).’ Both skills constitute the decoding of the signs or vital hieroglyphs through which life acts on consciousness. In a usage which anticipates the whole of Kenneth Burke’s ‘grammar of motives,’ Hamann identifies ‘action’ (Handlung) with ‘dynamic linguistic posture or structure’ (Sprach-gestaltung). Hamann opposes Kantian categories of universal, mental a priori in the name of those local, determinant energies inherent in a given language. Out of diverse tongues men will necessarily construe diverse mental and even sensory frameworks. Language generates specific cognition. Despite their rhapsodic, mystical format, the Philologische Einfälle und Zweifel of 1772 repay serious attention. Hamann throws out suggestions which anticipate the linguistic relativism of Sapir and Whorf. He seems to be saying that it is different languages that cause the different selections made by men from among that ‘ocean of sensations’ which tides, indiscriminately, through human sensibility. Hamann is arguing that neither Cartesian co-ordinates of general, deductive reasoning, nor Kantian mentalism will serve to account for the creative, irrational, and manifold proceedings through which language—unique to the species but so varied among nations—shapes reality and is, in turn, acted upon by local human experience.
It is one of the achievements of Romanticism to have sharpened the sense of locale, to have given specific density to our grasp of geographical and historical particularity. Herder was possessed of a sense of place. His ‘Sprachphilosophie’ marks a translation from the inspired fantastications of Hamann to the development of genuine comparative linguistics in the early nineteenth century. Herder’s quality can, I think, be overrated. He never shook himself free of the enigma of the natural or divine origin of language as he posed it in his famous essay of 1772. All the evidence seemed to point to an instinctual and evolutionary genesis of human speech, exactly as Lucretius and Vico had supposed. Yet the gap between spontaneous, mimetic speech-sounds and the wonder of mature language seemed too great. Thus the theory of a divine act of special bestowal was never far from Herder’s thoughts. Like Leibniz, Herder had a vivid realization of the atomic quality of human experience, each culture, each idiom being a particular crystal reflecting the world in a particular way. The new nationalism and vocabulary of race provided Herder with a ready focus. He called for ‘a general physiognomy of the nations from their languages.’ He was convinced of the irreducible spiritual individuality of each language, and particularly of German, whose antique expressive strengths had lain dormant but were now armed for the light of a new age and for the creation of a literature of world rank. National character is ‘imprinted on language’ and, reciprocally, bears the stamp of language. Hence the supreme importance of the health of language to that of a people; where language is corrupted or bastardized, there will be a corresponding decline in the character and fortunes of the body politic. Herder carried this belief to curious lengths. He stated in the Fragmente that a language would derive great benefits by guarding ‘itself from all translations.’ The notion is very similar to that of mystical grammarians seeking to protect the holy text from traduction. An untranslated language, urges Herder, will retain its vital innocence, it will not suffer the debilitating admixture of alien blood. To keep the Originalund Nationalsprache unsullied and alive is the eminent task of the poet.
The short years between Herder’s writings and those of Wilhelm von Humboldt were among the most productive in the history of linguistic thought. Sir William Jones’s celebrated Third Anniversary Discourse on the Hindus of 1786 had, as Friedrich von Schlegel put it, ‘first brought light into the knowledge of language through the relationship and derivation he demonstrated of Roman, Greek, Germanic and Persian from Indic, and through this into the ancient history of peoples, where previously everything had been dark and confused.’ Schlegel’s own Ueber die Sprache und Weisheit der Indier of 1808, which contains this tribute to Jones, itself contributed largely to the foundations of modern linguistics. It is with Schlegel that the notion of ‘comparative grammar’ takes on clear definition and currency. Not much read today, Mme de Staël’s De L’Allemagne (1813) exercised tremendous influence. In her impressionistic but often acutely intelligent portrayal of a waking nation, Mme de Staël argued that there were crucial reciprocities between the German language and the character and history of the German people. Expanding on suggestions already made by Hamann, she sought to correlate the metaphysical ambience, internal divisions, and lyric bias of the German national spirit with the gnarled weave and ‘suspensions of action’ in German syntax. She saw Napoleonic French as antithetical to German, and found its systematic directness and rhetoric clearly expressive of the virtues and vices of the French nation.
All these lines of debate and conjecture anticipate Humboldt’s work. But to enter on that work is to enter on an entirely different order of intellectual achievement. The play of intelligence, the delicacy of particular notation, the great front of argument which Humboldt exhibits, give his writings on language, incomplete though they are, a unique stature. Humboldt is one of the very short list of writers and thinkers on language—it would include Plato, Vico, Coleridge, Saussure, Roman Jakobson—who have said anything that is new and comprehensive.
Humboldt was fortunate. An extraordinary linguistic and psychological process was occurring all around him: a major literature was being created. It brought to bear on language and national sensibility a concentration of individual genius together with a common vision for which there are few parallels in history. Goethe, Schiller, Wieland, Voss, Hölderlin, and a score of others were doing more than composing, editing, translating masterpieces. With a high degree of policy and proclaimed intent, they were making of the German language an exemplar, a deliberate inventory of new possibilities of personal and social life. Werther, Don Carlos, Faust are supreme works of the individual imagination, but also intensely pragmatic forms. In them, through them, the hitherto divided provinces and principalities of the German-speaking lands could test a new common identity. Goethe and Schiller’s theatre at Weimar, Wieland’s gathering of German ballads and folk poetry, the historical narratives and plays of Kleist set out to create in the German mind and in the language a shared echo. As Vico had imagined it would, a body of poetry gave a bond of remembrance (partially fictive) to a new national community. As he studied the relations of language and society, Humboldt could witness how a literature, produced largely by men whom he knew personally, was able to give Germany a living past, and how it could project into the future great shadow-forms of idealism and ambition.
During his working years, Indo-European linguistics and the comparative study of classical, Hebraic, and Celtic antiquities, according to new criteria of philological and textual rigour, were laying the foundations for a genuine science of language. That such a science would have to enlist history, psychology, poetics, ethnography, and even various branches of biology, was clear to Humboldt. Like Goethe, he held the individual fact to be, as it were, shone through by the constant energies of universal, organic unity. It is the great weave and pulse of life itself that gives to each isolated phenomenon (isolated only because we may not yet have perceived the surrounding field of force) its meaning. To Humboldt and his brother, this intimation of universality was no empty metaphor. The Humboldts were among the last Europeans of whom it may be said with fair confidence that they had direct professional or imaginative notions of very nearly the whole of extant knowledge. Ethnographers, anthropologists, linguists, statesmen, educators, the two brothers were a nerve-centre for humanistic and scientific inquiry. Their active interests, like Leibniz’s, ranged with authority and passionate curiosity from mineralogy to metaphysics, from the study of Amerindian antiquities to modern technology. When he posited language as the centre of man, Wilhelm von Humboldt was in a position to feel what such a pivot must inform and relate. Yet being in natural touch with the later eighteenth century, Humboldt still possessed a certain receptivity to those traditions of occult linguistic speculation which, as we have seen, led back unbroken to Nicholas of Cusa and Paracelsus. Both the very old and the newest were active in Humboldt’s great enterprise.
That enterprise has come down to us in an incomplete, edited form.1 It includes the lecture ‘Ueber das Entstehen der gram-matischen Formen und ihren Einfluss auf die Ideenentwicklung’ (the title is itself a manifesto) of January 1822, and the magnum opus on which Humboldt was engaged from the 1820s until his death in 1835, and which was posthumously put together and published: Ueber die Verschiedenheit des menschlichen Sprachbaues und ihren Einfluss auf die geistige Entwicklung des Menschengeschlechts. Even translated, that title retains its proud scope: On the Differentiation of the Structure of Human Language, and its Influence on the Spiritual Evolution of the Human Race. Humboldt aims at nothing less than an analytic correlation of language and human experience. He would lay bare the concordance between the Weltanschauung of a given language and the history and culture of those who speak it. Essential to this analysis is the belief that language is the true or the only verifiable a priori framework of cognition. Perception is organized by the imposition of that framework on the total flux of sensations. ‘Die Sprache ist das bildende Organ des Gedankens,’ says Humboldt, using bildend in its forceful, twofold connotation of ‘image’ (Bild) and ‘culture’ (Bildung). Different linguistic frameworks will divide and channel the sensory flux differently: ‘Jede Sprache ist eine Form und trägt ein Form-Princip in sich. Jede hat eine Einheit als Folge eines in ihr waltenden Princips.’ This organic evolutionism goes well beyond and, indeed, against Kant. In so doing, Humboldt arrives at a key notion: language is a ‘third universe’ midway between the phenomenal reality of the ‘empirical world’ and the internalized structures of consciousness. It is this median quality, this material and spiritual simultaneity, that makes of language the defining pivot of man and the determinant of his place in reality. Seen thus, language is a universal. But so far as each human tongue differs from every other, the resulting shape of the world is subtly or drastically altered. In this way, Humboldt conjoins the environ-mentalism of Montesquieu and the nationalism of Herder with an essentially post-Kantian model of human consciousness as the active and diverse shaper of the perceived world.
The shaping agencies of intellect, Coleridge called them ‘esemplastic powers,’ do not, as it were, perform via language. They are inherent in language. Speech is poiesis and human linguistic articulation is centrally creative. It may be that Humboldt derived from Schiller his emphasis on language as being itself the most comprehensive work of art. His own contribution is to insist, in a way that strikes a very modern note, on language as a total generative process. Language does not convey a pre-established or separately extant content, as a cable conveys telegraphic messages. The content is created in and through the dynamics of statement. The entelechy, the purposeful flow of speech—we find in Humboldt a kind of romantic Aristotelianism—is the communication of ordered, perceived experience. But experience only assumes order and cognizance in the language-matrix. Ultimately, but inexplicably, language, die Sprache, is identical with ‘the ideal totality of spirit’ or Geist. As we shall see, the fact that this radical identity cannot be explained will undermine Humboldt’s actual linguistic analyses.
Under pressure of his extraordinary vision and emotional awareness of the life-giving, life-determining powers of language, Humboldt advances the idea that language can be adverse to man. So far as I am aware, no one before him had seen this point, and even now we have hardly grasped its implications. Humboldt’s statement is arresting: ‘Denn so innerlich auch die Sprache durchaus ist, so hat sie dennoch zugleich ein unabhängiges, äusseres, gegen den Menschen selbst Gewalt ausübendes Dasein’ (‘Albeit language is wholly inward, it nevertheless possesses at the same time an autonomous, external identity and being which does violence to man himself’). Language makes man at home in the world, ‘but it also has the power to alienate.’ Informed by energies proper to itself, more comprehensive and timeless than any who make use of it, human speech can raise barriers between man and nature. It can bend the mirrors of consciousness and of dreams. There is a phenomenon of linguistic Entfremdung inseparable from the creative genius of the word. The term is Humboldt’s, and the insight it expresses is of vital relevance to a theory of translation.
Ueber die Verschiedenheit des menschlichen Sprachbaues (particularly sections 19 and 20) is crowded with linguistic conjectures of prophetic brilliance. Man walks erect not because of some ancestral reaching out towards fruit or branch, but because discourse, die Rede, ‘would not be muffled and made dumb by the ground.’ More than a century before the modern structuralists, Humboldt notes the distinctive binary character of the linguistic process: it shares, it mediates between, the crucial antinomies of inner and outer, subjective and objective, past and future, private and public. Language is far more than communication between speakers. It is dynamic mediation between those poles of cognition which give human experience its underlying dual and dialectical form. Here Humboldt clearly anticipates both C. K. Ogden’s theory of opposition and the binary structuralism of Lévi-Strauss.
From this wide range of argument, I want to select those points which are immediate to our theme: the multiplicity of human tongues and the relations between Weltansicht and Wort.
‘The bringing forth of language is an inner necessity for mankind.’ It is, moreover, in the nature of ‘spirit’ to seek to realize, to energize into conscious being, all modes of possible experience. This is the true cause of the immense variety of speech forms. Each is a foray into the total potentiality of the world. ‘Jede Sprache,’ writes Humboldt, ‘ist ein Versuch.’ It is a trial, an assay. It generates a complex structure of human understanding and response and tests the vitality, the discriminatory range, the inventive resources of that structure against the limitless potential of being. Even the noblest language is only ein Versuch and will remain ontologically incomplete. On the other hand, no language however primitive will fail to actualize, up to a point, the inner needs of a community. Humboldt is convinced that different tongues provide very different intensities of response to life; he is certain that different languages penetrate to different depths. He takes over Schlegel’s classification of ‘higher’ and ‘lower’ grammars. Inflection is far superior to agglutination. The latter is the more rudimentary mode, a Naturlaut. Inflection allows and compels a far subtler, more dynamic treatment of action. It makes qualitative perception more acute and conduces necessarily to a more developed articulation (i.e. realization) of abstract relations. To pass from an agglutinate to an inflected tongue is to translate experience ‘upward’.
Humboldt now sets out to perform the crucial experiment. He applies his theory of the reciprocal determinations of language and world-view to specific cases. He seeks to show how Greek and Latin respectively determine particular ethnic, national aggregates of feeling. He would demonstrate that these two great idioms produced contrasting structures of civilization and social reflex. The argument is intelligently set out and gives proof of Humboldt’s at-homeness in classical philology and literature. But it falls unquestionably short of its theoretic aim and promise.
The Greek tone is light, delicate, nuancé. Attic civilization is incomparably inventive of intellectual and plastic forms. These virtues are engendered by and reflected in the precisions and shadings of Greek grammar. Few other languages have cast so finely-woven a net over the currents of life. At the same time, there is that in Greek syntax which helps explain the divisive quality of Greek politics, the excessive trust in rhetoric, the virtuosities of falsehood which sophisticate and corrode the affairs of the polis. Latin offers a grave contrast. The stern, masculine, laconic tenor of Roman culture is exactly correlate with the Latin language, with its sobriety, even paucity, of syntactic invention and Lautformung. The lettering of a Latin inscription is perfectly expressive of the linear, monumental weight of the language. Both are the active mould of the Roman way of life.
Humboldt’s argument is circular. Civilization is uniquely and specifically informed by its language; the language is the unique and specific matrix of its civilization. The one proposition is used to demonstrate the other and vice versa. Knowing the Greeks to have been one thing and the Romans another, we argue back to linguistic differences. In what way do aorist and optative help or fail to account for the indiscriminate bluntness of Spartan life? Can we discern modulations in the ablative absolute as Rome passes from Republican to Augustan Latin? Post hoc and propter hoc are inevitably blurred. Humboldt’s summarizing statement is eloquent, but also self-betraying in its lofty indistinction. Different languages engender different spiritual constructs of reality: ‘der dadurch hervorgebrachte verschiedene Geist schwebt, wie ein leiser Hauch, über dem Ganzen’ (‘the differing Spirit thus produced hovers, like a silent breath, over the whole’). Having identified Sprache with Geist (Hegel’s vocabulary is exactly contemporary with his own), Humboldt must conclude in this way. But having stated, at the outset, that this identification is, in the final analysis, inexplicable, he cannot use it to enforce demonstrable proof. His conviction remains fundamentally intuitive. For all its philosophic reach and sensibility to linguistic values, moreover, Humboldt’s position is not fully worked out. The essential argument is ‘monadist’ or relativist, but a universalist tendency can also be found. Hence the lack of final incisiveness in Humboldt’s key terms, ‘structure of language’ and ‘structures determined by a particular language.’ There is no doubt that these terms infer a wide range of example and historical evidence. But pressed home, they turn into metaphors, into shorthand formulations of the romantic criterion of organic life, rather than into verifiable concepts. Given the mystery at the core of the relations between ‘Language’ and ‘Spirit,’ it could hardly be otherwise.
It has been said that the line from Herder and Humboldt to Benjamin Lee Whorf is unbroken.1 Intellectually this is so. The actual history of linguistic relativity leads via the work of Steinthal (the editor of Humboldt’s fragmentary texts) to the anthropology of Franz Boas. From there it reaches the ethno-linguistics of Sapir and Whorf. One can summarize that history as being an attempt to provide Humboldt’s intuitions with a solid basis of semantic and anthropological fact. Much of the argument is developed in Germany. Nor is this surprising. The first true Germany was that of Luther’s vernacular. Gradually the German language created those modes of shared sensibility from which the nation-state could evolve. When that state entered modern history, a late arrival burdened with myths and surrounded by an alien, partially hostile Europe, it carried with it a sharpened, defensive sense of unique perspective. To the German temper, its own Weltansicht seemed a special vision, whose foundations and expressive genius lay in the language. Reflecting on the drastic extremes of German history, on the apparently fatal attempts of the German nation to break out of the ring of more urbane or, in the east, more primitive and menacing cultures, German philosophers of history thought of their language as a peculiarly isolating yet also numinous factor. Other nations could not feel their way into its arcane depths. But great springs of renewal and metaphysical discovery would surge from what Schiller called die verborgenen Tiefen.
Cassirer’s Philosophy of Symbolic Forms gave fresh impetus to Humboldt’s ideas. Cassirer was in agreement with the theory that the different conceptual categories into which different languages place the same sensory phenomena must reflect linguistically determined differences of perception. The stimuli are demonstrably identical; the responses are often strikingly disparate. Between the ‘physiological universal’ of consciousness and the specific cultural–conventional process of identification and response lies the membrane of a particular language or, as Cassirer put it, the unique ‘inner form’ which distinguishes it from all other languages. In a series of books ranging from Mutterprache und Geistesbildung (1929) to Vom Weltbilder Deutschen Sprache in 1950, Leo Weisger-ber sought to apply the ‘monadic’ or relativity principle to the actual, detailed features of German syntax and, correspondingly, to the history of German attitudes. It was his central affirmation that ‘our understanding is under the spell of the language which it utilizes.’ A very similar formulation was put forward by the linguist Jost Trier. Every language structures and organizes reality in its own manner and thereby determines the components of reality that are peculiar to this given language. This determination constitutes what Trier, in the early 1930s, called das sprachliche Feld. Thus, in a distinctly Leibnizian way, each tongue or language-monad constructs and operates within a total conceptual field (the imagistic correlation with quantum physics is obvious). This field may be understood as a Gestalt. Being linguistically diverse, different cultures impose a different Gestalt on the same raw material and total aggregate of experience. In each case, the linguistic ‘feedback’ from experience is a particular one. Speakers of different languages therefore inhabit different ‘mediary worlds’ (Zwischenwelten). The linguistic world-view of a given community shapes and gives life to the entire landscape of psychological and communal behaviour. It is language which decides how different conceptual groupings and contours are to be ‘read’ and related within the whole. Often a language will ‘filter out’ from the field of potential recognition even more information than it includes in that field. The gauchos of the Argentine know some 200 expressions for the colours of horses’ hides, and such discrimination is obviously vital to their economy. But their normal speech finds room for only four plant names.
In American linguistics, relativism drew both on the legacy of Humboldt and on anthropological field-work. Though treated with reservations, Levy-Bruhl’s concept of a ‘primitive mind,’ in which the ethnographer could observe pre-rational or non-Cartesian linguistic–logical processes, had its influence. Anthropological study of American Indian-cultures seemed to bear out Humboldt’s conjectures on linguistic determinism and Trier’s notion of the ‘semantic field.’ The whole approach is summarized by Edward Sapir in an article dated 1929:1
The fact of the matter is that the ‘real world’ is to a large extent unconsciously built up on the language habits of the group. No two languages are ever sufficiently similar to be considered as representing the same social reality. The worlds in which different societies live are distinct worlds, not merely the same world with different labels attached.
The emphasis on ‘group’ is worth noting. The ‘semantic field’ of a given culture is a dynamic, socially motivated construct. The particular ‘language and reality game’ played by the community depends, in a way very similar to that argued by Wittgenstein in the Philosophical Investigations, on the actions, on the historically evolved and agreed-to customs of the particular society. What we find here is a ‘dynamic mentalism:’ language organizes experience, but that organization is constantly acted upon by the collective behaviour of the particular group of speakers. Thus there occurs a cumulative dialectic of differentiation: languages generate different social modes, different social modes further divide languages.
The ‘monadist’ case has philosophic origins of great distinction in the work of Leibniz and of Humboldt. Its crowning statement is also of great intellectual fascination. The ‘metalinguistics’ of Whorf have for some time been under severe attack by both linguists and ethnographers. It looks as if a good deal of his work cannot be verified. But the papers gathered in Language, Thought and Reality (1956) constitute a model which has extraordinary intellectual elegance and philosophic tact. They are a statement of vital possibility, an exploration of consciousness relevant not only to the linguist but also to the poet and, decisively, to the translator. Whorf was an outsider. He brought to ethno-linguistics a sense of the larger issues, of the poetic and metaphysical implications of language study such as is rare among professionals. He had something of Vico’s philosophic curiosity, but was a chemical engineer with a distinctively modern awareness of scientific detail. The years in which Roman Jakobson, I. A. Richards and Benjamin Lee Whorf were active simultaneously must count among decisive moments in the history of the investigation of the human mind.
Whorf’s theses are well known. Linguistic patterns determine what the individual perceives in his world and how he thinks about it. Since these patterns—observable in the syntax and lexical means of the language—vary widely, the modes of perception, thought, and response in human groups using different language systems will be very different. World-views that are basically unlike will result. Whorf designates these as ‘thought worlds.’ They make up the ‘microcosm that each man carries about within himself, by which he measures and understands what he can of the macrocosm.’ There is, so far as human consciousness goes, no such entity as a universally objective physical reality. ‘We dissect nature along lines laid down by our native language.’ Or to be more exact: there is a fundamental duality in the exercise of human perception (Whorf is drawing on Gestalt psychology). There is a universal but also rudimentary neuro-physiological apprehension of space that may have preceded language in the evolution of the species and that may still precede articulate speech in the growth of the infant. But once a particular language is used, a particular conceptualization, of space follows (Whorf is not altogether clear as to whether language determines that conceptualization or only conditions it). Spatialization, and the space–time matrix in which we locate our lives, are made manifest in and by every element of grammar. There is a distinctive Indo-European time-sense and a corresponding system of verb tenses. Different ‘semantic fields’ exhibit different techniques of numeration, different treatments of nouns denoting physical quantity. They divide the total spectrum of colours, sounds, and scents in very diverse ways. Again, Wittgenstein’s use of ‘mapping’ offers an instructive parallel: different linguistic communities literally inhabit and traverse different landscapes of conscious being. In one of his very last papers, Whorf summarized his entire vision:1
Actually, thinking is most mysterious, and by far the greatest light upon it that we have is thrown by the study of language. This study shows that the forms of a person’s thoughts are controlled by inexorable laws of pattern of which he is unconscious. These patterns are the unperceived intricate systematizations of his own language—shown readily enough by a candid comparison and contrast with other languages, especially those of a different linguistic family. His thinking itself is in a language—in English, in Sanskrit, in Chinese. And every language is a vast pattern-system, different from others, in which are culturally ordained the forms and categories by which the personality not only communicates, but also analyses nature, notices or neglects types of relationship and phenomena, channels his reasoning, and builds the house of his consciousness.
To show that this doctrine ‘stands on unimpeachable evidence,’ Whorf was prepared to apply comparative semantic analyses to a wide range of languages: Latin, Greek, Hebrew (there are important links between his own work and the eccentric Kabbalism of Fabre d’Olivet), Kota, Aztec, Shawnee, Russian, Chinese and Japanese. Unlike many universalists, Whorf had an obvious linguistic ear. But it is his work on the languages of the Hopis of Arizona that carries the weight of evidence. It is here that the notion of distinct ‘pattern-systems’ of life and consciousness is argued by force of specific example. The key papers on ‘an American Indian model of the universe’ date from circa 1936 to 1939, at which point Whorf extended his analyses to the Shawnee language.
Examining the punctual and segmentative aspects of verbs in Hopi, Whorf concludes that the language maps a certain terrain ‘of what might be termed primitive physics.’ As it happens, Hopi is better equipped to deal with wave processes and vibrations than is modern English. ‘According to the conception of modern physics, the contrast of particle and field of vibrations is more fundamental in the world of nature than such contrasts as space and time, or past, present, and future, which are the sort of contrasts our own language imposes upon us. The Hopi aspect-contrast… being obligatory upon their verb forms, practically forces the Hopi to notice and observe vibratory phenomena, and furthermore encourages them to find names for and to classify such phenomena.’ Whorf finds that the Hopi language contains no words, grammatical forms or idiomatic constructions referring directly to what we call ‘time,’ or to the vectors of time and motion as we use them. The ‘metaphysics underlying our own language, thinking, and modern culture’ necessarily imposes a static three-dimensional infinite space, but also a perpetual time-flow. These two ‘cosmic co-ordinates’ could be harmoniously conjoined in the physics of Newton and the physics and psychology of Kant. They confront us with profound internal contradictions in the world of quantum mechanics and four-dimensional relativity. The metaphysical framework which informs Hopi syntax is, according to Whorf, far better suited to the world-picture of modern science. Hopi verb tenses and phrasings articulate the existence of events ‘in a dynamic state, yet not a state of motion.’ The semantic organization of ‘eventuating and manifesting’ phenomena allows—indeed enforces—precisely those modulations from subjective perceptions or ‘ideal mappings’ of events to objective status, which Indo-European grammar finds it so difficult to accommodate or must express wholly in mathematical terms.
In translating into English, the Hopi will say that these entities in process of causation ‘will come’ or that they—the Hopi—‘will come to’ them, but in their own language, there are no verbs corresponding to our ‘come’ and ‘go’ that mean simple and abstract motion, our purely kinematic concept. The words in this case translated ‘come’ refer to the process of eventuating without calling it motion—they are ‘eventuates to here’ (pew’i) or ‘eventuates from it’ (angqö) or ‘arrived’ (pitu, pl. öki) which refers only to the terminal manifestation, the actual arrival at a given point, not to any motion preceding it.1
Thus the entire Hopi treatment of happenings, inferential reasoning, and distant events is delicate and susceptible of provisional postures in just the way so often required by twentieth-century astrophysics or wave-particle theory. The shaping influence of the observer on the process observed, the statistics of indeterminacy, are inherent in Hopi as they are not, or only by virtue of explanatory metaphor, in English.
Crucial to Whorfian semantics is the notion of the cryptotype. He defines it ‘as a submerged, subtle, and elusive meaning, corresponding to no actual word, yet shown by linguistic analysis to be functionally important in the grammar.’ It is these ‘cryptotypes’ or ‘categories of semantic organization’—dispersion without boundaries, oscillation without agitation, impact without duration, directed motion—which translate the underlying metaphysics of a language into its overt or surface grammar. It is the study of such ‘cryptotypes’ in different languages, urges Whorf, that will lead anthropology and psychology to an understanding of those deep-seated dynamics of meaning, of chosen and significant form, that make up a culture. It is, no doubt, exceedingly difficult for an outsider, operating inevitably within the world-frame of his own tongue, to penetrate to the active symbolic deeps of a foreign tongue. We reach for the bottom and stir up further darkness. ‘Cryptotypes,’ moreover, are ‘so nearly at or below the threshold of conscious thinking’ that even the native speaker cannot put them into adequate words. Patently, they elude translation (we shall return to this point). Yet careful, philosophically and poetically disciplined observation does allow the linguist and anthropologist to enter, in some degree at least, into the ‘pattern-system’ of an alien tongue. Particularly if he acts on the principles of ironic self-awareness which underlie a genuine relativist view.
Whorf was tireless in emphasizing the built-in bias, the axiomatic arrogance of traditional and universalist philology, with its scarcely veiled presumption that Sanskrit and Latin constitute the natural, optimal model of all human speech or, at the least, a model manifestly preferable to all others. Whorf’s revaluation of ‘thinking in primitive communities’ coincides in date and spirit with Lévi-Strauss’s early studies of the genius of La Pensée sauvage. Lévi-Strauss would fully endorse Whorf’s assertion that ‘many American Indian and African languages abound in finely wrought, beautifully logical discriminations about causation, action, result, dynamic or energic quality, directness of experience, etc., all matters of the function of thinking, indeed the quintessence of the rational. In this respect they far outdistance the European languages.’ Whorf offers telling instances: the four forms of the pronoun in the Algonkian languages allow compact notations of intricate social situations; the distinction between a tense for past events with present result or influence, and for those with none, in Chichewa, ‘a language related to Zulu, spoken by a tribe of unlettered Negroes in East Africa;’ the three causal verb forms in the Cœur d’Alène language, spoken by a small Indian tribe in Idaho. Here again, Whorf finds the paradox that the ‘semantic field’ of numerous so-called primitive communities segments experience into a phenomenology which is closer than that of the Indo-European language family to the data of twentieth-century physics and Gestalt psychology. Equally fascinating are Whorf’s hints—any theory of translation will want to explore and extend them—that different languages show different degrees of accord between phonetics (which must, in some measure, be universal) and the ‘inner music of meaning.’ German zart, meaning ‘tender,’ calls up tonal associations of bright hardness. English deep ought to go with such sounds of quick, sharp lightness as ‘peep.’ Meaning in a given tongue may go against the grain of apparently universal aural associations. This clash between ‘mental’ and ‘psychic’ codes of recognition may be crucial to the evolution of a particular language and will assume very different forms in different tongues.
A picture of language, mind, and reality based almost exclusively on Cartesian–Kantian logic and on the ‘semantic field’ of Standard Average European (SAE) is a hubristic simplification. The close of ‘Science and Linguistics,’ a paper published in 1940, is worth quoting in full—especially at a time when the study of language is so largely dominated by a theory of dogmatic generality and mathematical aspect:
A fair realization of the incredible degree of the diversity of linguistic system that ranges over the globe leaves one with an inescapable feeling that the human spirit is inconceivably old; that the few thousand years of history covered by our written records are no more than the thickness of a pencil mark on the scale that measures our past experience on this planet; that the events of these recent millenniums spell nothing in any evolutionary wise, that the race has taken no sudden spurt, achieved no commanding synthesis during recent millenniums, but has only played a little with a few of the linguistic formulations and views of nature bequeathed from an inexpressibly longer past. Yet neither this feeling nor the sense of precarious dependence of all we know upon linguistic tools which themselves are largely unknown need be discouraging to science but should, rather, foster that humility which accompanies the true scientific spirit, and thus forbid that arrogance of the mind which hinders real scientific curiosity and detachment.
Whatever may be the future status of Whorf’s theories of language and mind, this text will stand.
3
Such are the distinction and consequence of Whorf’s metalin-guistics, that, even of themselves, critiques of Whorf constitute a fair statement of the universalist case. These critiques bear on the circularity of Whorf’s evidence. Seeing a dripping spring, an Apache will describe it as ‘whiteness moving downward.’ The verbal formulation is clearly different from that in current English. But what direct insight does it afford into Apache thinking? It is tautological to argue that a native speaker perceives experience differently from us because he talks about it differently, and then infer differences of cognition from those of speech. Behind such inference lies a rudimentary, untested scheme of mental action. In ‘A Note on Cassirer’s Philosophy of Language,’ E. H. Lenneberg summarizes a whole range of philosophic doubts: ‘There is no cogent reason to assume that the grammarian’s articulation of the stream of speech is coterminous with an articulation of knowledge or the intellect.’ Words are not the embodiments of invariant mental operations and fixed meanings. The idea that conventional syntactic patterns incorporate uniquely determined and determinant acts of perception is itself the reflection of a primitive dualism. It corresponds to the mind-body image of early psychology. Any operational model of the linguistic process, e.g. Wittgenstein’s proposal that ‘the meaning of a word is its use in the language,’ will refute Whorf’s deterministic parallelism of thought and speech.
Moreover, if the Humboldt–Sapir–Whorf hypothesis were right, if languages were monads with essentially discordant mappings of reality, how then could we communicate interlingually? How could we acquire a second tongue or traverse into another language-world by means of translation? Yet, manifestly, these transfers do occur continually.
The empirical conviction that the human mind actually does communicate across linguistic barriers is the pivot of universalism. To the twelfth-century relativism of Pierre Hélie, with his belief that the disaster at Babel had generated as many kinds of irreconcilable grammar as there are languages, Roger Bacon opposed his famous axiom of unity: ‘Grammatica una et eadem est se-cundum substantiam in omnibus linguis, licet accidentaliter varietur.’ Without a grammatica universalis, there could be no hope of genuine discourse among men, nor any rational science of language. The accidental, historically moulded differences between tongues are, no doubt, formidable. But underlying these there are principles of unity, of invariance, of organized form, which determine the specific genius of human speech. Amid immense diversities of exterior shape, all languages are ‘cut from the same pattern’.
We have met this intuitive certitude in Leibniz and even among the relativistic arguments of Humboldt. The successes obtained by nineteenth-century Indo-European philology in formalizing, in giving a normative and predictive account of the great mass of discrete phonological and grammatical facts, strengthened the universalist bias. Today, the working vision of a universal grammar is shared by almost all linguists. Indeed, it is because it deals with phenomena of a universal, deep-seated character, with the general ground rules of human cognitive processes, that current linguistic theory advances claims to philosophic and psychological authority. ‘The main task of linguistic theory must be to develop an account of linguistic universals that, on the one hand, will not be falsified by the actual diversity of languages and, on the other, will be sufficiently rich and explicit to account for the rapidity and uniformity of language learning, and the remarkable complexity and range of the generative grammars that are the product of language learning.’1
The axiom of universality and the aim of comprehensive description are clear. What remains of great difficulty is the question of levels (it already perplexed such universalists of the late eighteenth century as James Beattie). At what level of the structure of language can ‘universals’ be accurately located and described? How deep must we go below the live, obstinately diverse layers of linguistic usage? During the past forty years, the direction of universalist argument has been one of ever-deepening formalization and abstraction. In turn, each level of proposed universality has been found to be contingent or subverted by anomalies. Singularities have cropped up in what looked like the most general of assumptions. Instead of being rigorous and exhaustive, the description of ‘universal linguistic traits’ has often proved to be no more than an open-ended catalogue.
There are three obvious planes of language on which to seek out universals: the phonological, the grammatical, and the semantic.
All human beings possess the same neurophysiological equipment with which to emit and receive sounds. There are notes whose pitch lies outside the range of the human ear; there are tones which our vocal cords cannot produce. All languages, therefore, fall within certain definable material bounds. All are combinations of a limited set of physical phenomena. It is an obvious move to seek to identify and enumerate the physiological or phonological universals of which each and every spoken tongue is a selective aggregate. One of the most influential of such enumerations is N. S. Trubetskoy’s Grundzüge der Phonologie published in Prague in 1939. Comparing some 200 phonological systems, Trubetskoy set out those acoustic structures without which there cannot be a language and which all languages exhibit. Roman Jakobson’s theory of ‘distinctive features’ is a refinement of Trubetskoy’s universals. Jakobson identifies some twenty universal phonetic elements, each of which can be rigorously characterized according to articulatory and acoustic criteria (e.g. every language must contain at least one vowel). In different combinations, these features make up the phonology, the physical presence and transmission of all languages. Using these crucial markers, a science-fiction writer or computer could devise a new tongue, and one could affirm in advance that it would fall within the set limits of human expressive potentiality. A signal-system lacking these ‘distinctive universals’ would lie, literally, outside the human octave.
In practice, the analysis of phonological universals turns out to be a rather simple-minded and blunt enterprise. A good many conclusions are, again, of the order of unsurprising generality implicit in the statement that all human beings require oxygen. Where the argument becomes prescriptive, problems of rigorous description arise. It seems safe enough to assert that all languages on this earth have a vowel system. In fact, the proposition is true only if we take it to include segmented phonemes which occur as syllabic peaks—and even in that case, at least one known tongue, Wishram, poses problems. There is a Bushman dialect called Kung, spoken by a few thousand natives of the Kalahari. It belongs to the Khoisan group of languages, but is made up of a series of clicking and breathing sounds which, so far as is known, occur nowhere else, and which have only recently been transcribed. Obviously, these sounds lie within the physiological bounds of human possibility. But why should this anomaly have developed at all, or why, if efficacious, should it be found in no other phonological system? A primary nasal consonant ‘is a phoneme of which the most characteristic allophone is a voiced nasal stop, that is, a sound produced by a complete oral stoppage (e.g. apical, labial), velic opening, and vibration of the vocal cords’.1 Having thus defined a PNC, phonologists can identify the conditions under which it occurs in all languages and the determined ways in which it affects the position and stress of other phonemes. But the plain statement that every human tongue has at least one primary nasal consonant in its inventory requires modification. Hockett’s Manual of Phonology (1955) reports a complete absence of nasal consonants from Quileute and two neighbouring Salishan languages. Whether such nasals once existed and have, in the course of history, become voiced stops, or whether, through some arresting eccentricity, Salishan speech never included nasal phonemes at all, remains undecided. Such examples can be multiplied.
Consequently, the universalist case proceeds beyond the somewhat rudimentary and ‘soft-edged’ material of phonology to that of grammar. If all languages are indeed cut from the same pattern, a comparative analysis of syntactic systems will reveal those elements that truly constitute a grammatica universalis.
The pursuit of such a ‘fundamental grammar’ is itself a fascinating chapter in the history of analytic thought. A considerable distance has been covered since Humboldt’s hope that a generalized treatment of syntactic forms would be devised to include all languages, ‘from the rawest’ to the most accomplished. The notion that certain fixed syntactic categories—noun, verb, gender—can be found in every tongue, and that all languages share certain primary rules of relation, became well established in nineteenth-century philology. That ‘same basic mould’ in which all languages are cast came to be understood quite precisely: as a set of grammatical units, of markers which themselves denote nothing but make a difference in composite forms, and of rules of combination.
Some of these rules are of very great generality. No language has been found to lack a first- and second-person singular pronoun. The distinctions between ‘I,’ ‘thou,’ and ‘he’ and the associated network of relations (so vital to kinship terms) exist in every human idiom. Every language in use among men has a class of proper names. No language has a vocabulary that is grammatically entirely homogeneous. A type of clause in which a ‘subject’ is talked about or modified in some manner, is observable in every linguistic system. All speech operates with subject–verb–object combinations. Among these, the sequences ‘verb–object–subject,’ ‘object–subject–verb,’ and ‘object–verb–subject’ are exceedingly rare. So rare, as to suggest an almost deliberate violation of a deep-rooted ordering of perception. Other ‘grammatical universals’ are points of detail: for example ‘when the adjective follows the noun, the adjective expresses all the inflectional categories of the noun. In such cases the noun may lack overt expression of one or all of these categories.’ The most ambitious list of syntactic universals to have been established ‘on the basis of the empirical linguistic evidence’ was that of J. H. Greenberg.1 It enumerates forty-five fundamental grammatical relations, and leads to the conclusion that ‘the order of elements in language parallels that in physical experience or the order of knowledge.’ The underlying grammar of all human speech forms is a mapping of the world. It emphasizes those features of the landscape and of bio-social experience which are common to all men. Differences of stress, organized sequence, relations of hierarchy as between the general and the particular or the sum and the part, these are the counters of reason from which all languages develop. If a language ‘has the category of gender, it always has the category of number.’ Otherwise, there would be human aggregates trapped in eccentric chaos.
Again, the scheme looks more impressive than it actually is. Compared to the total of languages in current use, the number whose grammar has been formalized and thoroughly examined is absurdly small (Greenberg’s empirical evidence is drawn almost exclusively from thirty languages). In syntax, moreover, no less than in phonology, tenacious singularities occur. One would expect all languages with a distinction of gender in the second-person singular to show this distinction in the third person as well. In nearly every known instance, this holds. But not in a very small cluster of tongues spoken in central Nigeria. The Nootka language provides an often-cited example of a grammatical system in which it is very difficult to draw any normal distinction between noun and verb. The alignment of genitive constructions looks like a primal typological marker according to which all languages can be classified into a small number of major groups. Araucanian, an Indian tongue spoken in Chile, and some Daghestan languages of the Caucasus do not fit the scheme. Such anomalies cannot be dismissed as mere curios. A single genuine exception, in any language whether living or dead, can invalidate the whole concept of a grammatical universal. Indeed, this whole approach has since been largely abandoned.
It is, in part, because the statistical, ethno-linguistic approach to syntactic universals has proved unsatisfactory or merely descriptive, that generative transformational grammars propose to argue at much greater phenomenological depths. In doing so, they have sought to drive the very notion of grammar inward, to a specifically linguistic innate faculty of human consciousness.
Chomskian grammar is emphatically universalist (but what other theory of grammar—structural, stratificational, tagmemic, comparative—has not been so?). No theory of mental life since that of Descartes and the seventeenth-century grammarians of Port Royal has drawn more explicitly on a generalized and unified picture of innate human capacities, though Chomsky and Descartes mean very different things by ‘innateness.’ In Descartes this ‘innateness,’ specifically underwritten by a transcendental wager on God, on the congruence between word and world, entails a social context. It anticipates some of the very ‘stimulus and response’ configurations which Chomsky will rebuke. Chomsky’s starting-point was the rejection of behaviourism. No simple pattern of stimulus and mimetic response could account for the extreme rapidity and complexity of the way in which human beings acquire language. All human beings. Any language. A child will be able to construct and understand utterances which are new and which are, at the same time, acceptable sentences in his language. At every moment of our lives we formulate and, understand a host of sentences different from any that we have heard before. These abilities indicate that there must be fundamental processes at work quite independently of ‘feedback from the environment’.1 Such processes are innate to all men: ‘human beings are somehow specially designed to do this, with data-handling or “hypothesis-formulating” ability of unknown character and complexity.’ Each individual on earth has somehow and in some form internalized a grammar from which his, but also any other language is generated. (‘Generation’ translates Humboldt’s erzeugen. Here, as in the shared axiom that language ‘makes infinite use of finite means,’ Chomskian universalism is congruent with the relativism of Humboldt.)
Differences between languages represent differences of ‘surface structure’ only. They are accidents of terrain which impress the eye but tell us scarcely anything of the underlying ‘deep structure.’ Via a set of rules, of which ‘rewriting rules’ are fundamental, ‘deep structures’ generate, i.e. bring to the phonetic surface, the sentences we actually use and hear. We are then able to work back from the actual physical sentence, together with the derivation tree or ‘phrase marker’ constructed for it, to obtain some insight into the underlying ‘deep structure.’ More complex sentences are, in turn, generated by a second class of rules, the ‘rules of transformation.’ These rules—for which the theory of recursive functions offers the best analogy—must be applied in an ordered sequence. Some of them are not ‘context-free;’ their correct application depends on the surrounding linguistic material. It is at this point, presumably, that a universal system modulates into a particular language or family of languages. But any ‘real progress in linguistics consists in the discovery that certain features of given languages can be reduced to universal properties of language, and explained in terms of these deeper aspects of linguistic form’.1
Chomsky contends that a search for universals at the phonological or ordinary syntactic level is wholly inadequate. The shaping centres of language lie much deeper. In fact, surface analogies of the kind cited by Greenberg may be entirely misleading: it is probable that the deep structures for which universality is claimed are quite distinct from the surface structure of sentences as they actually appear. The geological strata are not reflected in the local landscape.
But what are these ‘universal deep structures’ like?
It turns out that it is exceedingly difficult to say anything about them. In the vocabulary of Wittgenstein, the transition from ‘surface grammar’ to ‘depth grammar’ is a step towards clarity, towards a resolution of those philosophic muddles which spring from a confusion of linguistic planes. Chomskian ‘deep structures,’ on the other hand, are located ‘far beyond the level of actual or even potential consciousness.’ We may think of them as relational patterns or strings of an order of abstraction far greater than even the simplest of grammatical rules. Even this is too concrete a representation. ‘Deep structures’ are those innate components of the human mind that enable it to carry out ‘certain formal kinds of operations on strings.’ These operations have no a priori justification. They are of the category of essential arbitrariness inherent in the fact that the world exists. Thus ‘there is no reason to expect that reliable operational criteria for the deeper and more important theoretical notions of linguistics … will ever be forthcoming.’ Try to draw up the creature from the deeps of the sea, and it will disintegrate or change form grotesquely.
Yet ‘only descriptions concerned with deep structure will have serious import for proposals concerning linguistic universals.’ Since descriptions of this sort are rare, rather like cores from the great marine trenches, ‘any such proposals are hazardous, but are clearly no less interesting or important for being hazardous.’ Chomsky then proceeds to offer one example of a genuine formal universal. It concerns the rules which govern the operations and legitimacy of deletion in the underlying structure of sentences of the type ‘I know several more successful lawyers than Bill.’ These rules or ‘erasure transformations’ may be proposed ‘for consideration as a linguistic universal, admittedly on rather slender evidence’.1
Some grammarians would go even ‘deeper’ than Chomsky in locating the universal base of all languages. The sequential order of rules of transformation may itself lie near the surface and be specific to different languages. The whole notion of sequence may have to be modified when it is applied to ‘the rules of a universal base.’ Professor Emmon Bach suggests that ‘deep structures are much more abstract than had been thought’.1 It may be erroneous to think of them, even by analogy, as linguistic units or ‘atomic facts’ of grammatical relation. At this final level of mental organization, we may be dealing with ‘abstract kinds of pro-verbs which receive only indirect phonological representation’ (I take ‘proverbs’ to signify potentialities of meaning ‘anterior to’ even the most rudimentary verbal units). At one level such a scheme of ‘universal base rules’ resembles the logical systems of Carnap and Reichenbach. At another level, most probably metaphoric, it suggests the actual patterning of the cortex, with its immensely ramified yet, at the same time, bounded or ‘programmed’ network of electro-chemical and neurophysiological channels. A system of variables, the set of all names, ‘general predicates,’ and certain rules of constraint and relation between these, would, as it were, be imprinted on the fabric of human consciousness.
This imprint may never be susceptible of direct observation. But the ‘selectional constraints and transformational possibilities’ which we can discern at the surface of language give undeniable proof of its existence, efficacy, and universality. ‘Such a system expresses directly the idea that it is possible to convey any conceptual content in any language, even though the particular lexical items available will vary widely from one language to another—a direct denial of the Humboldt–Sapir–Whorf hypothesis in its strongest form.’2
Whether it is indeed ‘possible to convey any conceptual content in any language’ is what I seek to investigate.
Granted the extreme difficulty of defining universals of grammar, many linguists feel that it is far too early to identify any ‘semantic universals.’ Nevertheless, such identifications have been proposed, certainly since Vico’s suggestion that all languages contain key anthropomorphic metaphors. One of these, the comparison of the pupil of the eye to a small child (pupilla), has been traced in all Indo-European languages, but also in Hebrew (bath-‘ayin), Swahili, Lapp, Chinese, and Samoan.1 Every language contains both ‘opaque’ and ‘transparent’ words, i.e. words in which the relation between sound and sense is purely arbitrary (German Enkel) and those in which it is obviously figurative (French petit-fils). The existence and statistical distribution of these two types of words ‘is in all probability a semantic universal’.2 The presence in every known tongue of certain taboo words, of expressions circumscribed by a zone of prohibition or sacred power, may well be a universal though also context-bound semantic feature. The thought that onomatopoeic patterns, sibilants, lateral consonants, may be rooted in specific modes of human perception—that there are universal ways of ‘sounding the world’—is very ancient. It underlies a number of Plato’s conjectural etymologies. And indeed, i carries values of smallness in almost every Indo-European and Finno-Ugrian language. But English big and Russian velikij suffice to show that we are not dealing with anything like a universal semantic reflex. Lévi-Strauss and several psycho-linguists agree in finding ‘universal binomials’ or contrast-pairs which tend to divide reality for us, and whose polarization is reflected in metaphors and stress patterns throughout all languages (white/black, straight/crooked, rising/falling, sweet/sour). The white/black dichotomy is of particular interest, as it appears to convey a positive/negative valuation in all cultures, regardless of skin-colour. It is as if all men, since the beginning of speech, had set the light above the dark.
Chomsky puts forward a number of semantic universals of a very broad but suggestive type: ‘proper names in any language, must designate objects meeting a condition of spatiotemporal contiguity, and that the same is true of other terms designating objects; or the condition that color words of any language must subdivide the color spectrum into continuous segments; or the condition that artifacts are defined in terms of certain human goals, needs, and functions instead of solely in terms of physical qualities.’1 Again, the problem is one of the degree of precision which can be attached to such generalizations. All languages do subdivide the colour spectrum into continuous segments (though ‘continuous’ begs difficult issues in the neurophysiology and psychology of perception), but, as R. W. Brown and E. H. Lenneberg have shown, they go about their segmentation in ways which can be startlingly different. Indeed, basic questions about the relations between physical perception and linguistic coding remain far more open than Chomsky’s statement suggests.
The history of Chomsky’s own thought and of the modulation from transformational generative grammar to generative grammar has been one of the erosion of semantics. Chomsky’s Remarks on Nominalization emphasize that ‘meaning’ is not at issue. The connections between generative grammatical investigations and semantic studies are all but severed. The true issue is now one of ‘grammaticality.’ The question asked is this: ‘How would a natural-language speaker know that what he is enunciating is part of an organized grammar, how would he know, when encountering a new sentence, whether it is grammatically plausible or not?’ A formal universal is defined as the operational rule whereby grammars generate sentences. The exploration does not touch on stylistics or constructs of meaning in the semantic range. What are to be elicited and formalized are rule-constrained natural usages. When we say ‘He said John laughed,’ we are obeying and enacting the rule whereby ‘he’ and ‘John’ cannot co-refer. Generative grammarians are today not studying words and their manifold, possibly indeterminate, semantic fields. Nor are they mapping the neuro-physiological foundations of human speech. They are positing ‘abstract formalizations in human psychology’ (where ‘psychology’ continues to evade closer definition). The results obtained are not those which Chomsky announced in his earlier works. They are formal, meta-mathematical paradigms of the generation of grammaticality in speech.
To summarize: the evidence for the universality of those linguistic structures of which there is phenomenal evidence is, until now, provisional and putative. It oscillates between postulated levels of extreme formal abstraction at which the language-model becomes meta-mathematical and is divorced more or less completely from the phonetic fact, and levels which are crudely statistical (for example, Charles Osgood’s proposal that the ratio of the number of phonemes to the number of distinctive features in any and every language will vary around an efficiency value of 50 per cent). The guarded conclusion of at least one linguist opposed to facile universalism may prove justified: ‘Linguistic structures do differ, very widely indeed, among all the attested languages of the earth, and so do the semantic relationships which are associated with linguistic structures. The search for linguistic universals… has recently come to the fore again, but it is still premature to expect that we can make any except the most elementary observations concerning linguistic universals and expect them to be permanently valid. Our knowledge of two-thirds or more of the world’s languages is still too scanty (or, in many instances, non-existent).’1 It may be that too many linguists have assumed that the ‘deep structures’ of all languages are identical because they have equated universal criteria of constraint and possibility with what could be in truth aspects only of the grammar of their own tongue or language group.
None the less, the belief that ‘all languages are cut to the same pattern’ is, currently, widespread. Few grammarians would hold with Osgood that eleven-twelfths of any language consist of universals and only one-twelfth of specific, arbitrary conventions, but the majority would agree that the bulk and organizing principles of the iceberg belong to the subsurface category of universals. To most professional linguists today the question is less whether there are ‘formal and substantive universals of language’ but precisely what they are, and to what extent the depths at which they lie will ever be accessible to either philosophic or neurophysiological investigation.
The postulate of linguistic universals or, to be exact, of substantive universals, should lead by direct inference to a working theory of interlingual translation. Proof that mutual transfer between languages is possible should follow immediately on the principle of substantive universality. Translation ought, in effect, to supply that principle with its most palpable evidence. The very possibility of motion of meaning between languages would seem to be firmly rooted in the underlying templet or common architecture of all human speech. But how is one to distinguish substantive from formal universals? How, except by theoretical fiat at one end or local intuition at the other, can one determine whether perfect translation should be possible because formal universals underlie all speech, or whether actual untranslatabilities persist because universals are only rarely or obscurely substantive? The discrimination is cogent in theory but has not been shown to be so in practice. It shares implicit ambiguities with the related distinction between ‘deep’ and ‘surface’ structures. Formal universals can be postulated at remote depths beyond concrete investigation or possible paraphrase. Substantive universals will, inevitably, overlap with the pragmatic, obstinately particularized realities of natural language. Translation is, plainly, the acid test. But the uncertainties of relation between formal and substantive universality have an obscuring effect on the relations between translation and universality as such. Only if we bear this in mind can we understand a decisive hiatus or shift in terms of reference in Chomsky’s Aspects of the Theory of Syntax:
The existence of deep-seated formal universals… implies that all languages are cut to the same pattern, but does not imply that there is any point by point correspondence between particular languages. It does not, for example, imply that there must be some reasonable procedure for translating between languages.
A footnote reinforces the sense of a fundamental uncertainty or non sequitur: ‘The possibility of a reasonable procedure for translation between arbitrary languages depends on the sufficiency of substantive universals. In fact, although there is much reason to believe that languages are to a significant extent cast in the same mould, there is little reason to suppose that reasonable procedures of translation are in general possible.’1
How can the two suppositions be separated? ‘Point by point’ merely obscures the logical and substantive issue. The ‘topological mapping’ in which linguistic universals can be transferred from language to language—note the curious evasion in the phrase ‘between arbitrary languages’—may lie very deep, but if it exists at all, a ‘point by point correspondence’ must be demonstrable. If translation can be achieved, is it not precisely because of the underlying ‘sufficiency of substantive universals’? If, on the contrary, there is little reason to suppose that reasonable procedures of translation are ‘in general’ possible (and what does ‘in general’ really signify?), what genuine evidence have we of a universal structure? Are we not back in a Whorfian hypothesis of autonomous language-monads? Could Hall be right when he polemicizes against the whole notion of ‘deep structures,’ calling them ‘nothing but a paraphrase of a given construction, concocted ad hoc to enable the grammarian to derive the latter from the former by one kind of manipulation or another’?1 Might it be that the transformational generative method is forcing all languages into the mould of English, as much seventeenth-century grammar endeavoured to enclose all speech within the framework of classical Latin?
Once more, the problem of the nature of translation appears to be central to that of language itself. The lacuna between a system of ‘universal deep structures’ and an adequate model of translation suggests that the ancient controversy between relativist and universalist philosophies of language is not yet over. It also suggests that the theory whereby transformational rules map semantically interpreted ‘deep structures’ into phonetically interpreted ‘surface structures’ may be a meta-mathematical ideal of considerable intellectual elegance, but not a true picture of natural language. ‘No set of rules, however complete, is sufficient to describe… the utterances possible in any living language.’2 By placing the active nodes of linguistic life so ‘deep’ as to defy all sensory observation and pragmatic depiction, generative grammar, certainly in its ‘strong’ pristine versions, may have put the ghost out of all reach of the machine.
There is room, I submit, for an approach whose bias of interest focuses on languages rather than Language; whose evidence will derive from semantics (with all the implicit stress on meaning) rather than from ‘pure syntax;’ and which will begin with words, difficult as these are to define, rather than with imaginary strings or ‘pro-verbs’ of which there can never be any direct presentation. Investigation has shown that even the most formal rules of grammar must take into account those aspects of semantics and performance which Chomsky would exclude. Even individual sounds are concept-bound and act in a particular semantic field. It is doubtful, as well, whether a real grammar can start from and allow pre- or ungrammatical sentences as transformational generative grammar must. ‘Grammaticality is, in any case, not a phenomenon that can be measured in terms of simple binary opposition, declaring any linguistic phenomenon to be either grammatical or ungrammatical. There is an infinite gradation between something which every member of a speech-community would use and recognize unhesitatingly as completely normal, to the opposite extreme of something that every speaker would declare was never used … new formations resulting from analogy or blending are taking place all the time, and are being recognized and understood without difficulty.’1
Or to put it in summary fashion: a meta-mathematical view of language, working principally with pre- or pseudo-linguistic atomic units, will fail to account for the nature and possibility of relations between languages as they actually exist and differ.2
Noam Chomsky’s Theory of Government and Binding, his The Science of Language, show a sharpening distance from the universalist innocence and sovereign claims of earlier theories. The promises initially offered to non-technical readers of literature, of philosophy, of translation, have been relinquished. What few attempts have been made at systematic applications of generative grammatical methods to matters and texts in the ‘poetical’ or ‘rhetorical’ modes, to actualities of speech and writing other than rudimentarily formalized, have departed widely from any strict Chomskyan position.
Hence the need of looking in directions which are, I fully admit, more impressionistic and far less amenable to formal codification. But language itself is ‘open-ended’ and charged with energies of the utmost diversity and intricacy. ‘The really deep results of transformational grammar,’ writes George Lakoff, ‘are, in my opinion, the negative ones, the hosts of cases where transformational grammar fell apart for a deep reason: it tried to study the structure of language without taking into account the fact that language is used by human beings to communicate in a social context.’1 Time moves through every feature of language as a shaping force. No true understanding can arise from synchronic abstraction. Even more than the linguists, and long before them, poets and translators have worked inside the time-shaped skin of human speech and sought to elucidate its deepest springs of being. Men and women who have in fact grown up in a multilingual condition will have something to contribute towards the problem of a universal base and a specific world-image. Translators have left not only a great legacy of empirical evidence, but a good deal of philosophic and psychological reflection on whether or not authentic transfers of meaning between languages can take place.
Much of current linguistics would have things neater than they are. Before conceding that the deeper, more important proceedings of language lie far beyond the level of actual or potential consciousness (Chomsky’s postulate), we must look to the vital disorders of literature in which that consciousness is most incisively at work. To know more of language and of translation, we must pass from the ‘deep structures’ of transformational grammar to the deeper structures of the poet. ‘Man weiss nicht, von wannen er kommt und braust,’ wrote Schiller of the surge of language from the depths to the light. No man knows from whence it comes:
Wie der Quell aus verborgenen Tiefen,
So des Sängers Lied aus dem Innern schallt
Und wecket der dunkeln Gefühle Gewalt,
Die im Herzen wunderbar schliefen.
1 A. Meillet and M. Cohen, Les Langues du monde (Paris, 1952).
1 The great work on this subject, and one of the most fascinating of intellectual histories, is Arno Borst, Der Turmbau von Babel: Geschiche der Meinungen über Ursprung und Vielfalt der Sprachen und Volker (Stuttgart, 1957–63).
1 Despite Arno Borst’s exhaustive inquiries, the origins of this particular number remain obscure. The 6 × 12 component suggests an astronomical or seasonal correlation.
1 Here, of course, I am drawing heavily on Gershom Scholem, Major Trends in Jewish Mysticism (Jerusalem, 1941 and New York, 1946).
1 Cf. Alexandra Koyré, La Philosophic de Jacob Boehme (2nd edition, Paris, 1971), pp. 456–62.
1 An English translation of this essay, by James Hynd and E. M. Valk, may be found in Delos, A Journal on and of Translation, 2 (1968).
1 See ch. 1 (pp. 3–54) of Jaime Alazraki’s Borges and the Kabbala, and Other Essays on his Fiction and Poetry (Cambridge, 1988).
1 Borges’s ‘The Mirror of Enigmas’ (in Labyrinths, New York, 1962) argues the specific interactions of gnostic philosophy and the speculum in aenigmate.
1 Cf. ‘Pierre Menard, Author of Don Quixote,’ translated by Anthony Bonner in Fictions (New York, 1962) with James E. Irby’s version of the same story in Labyrinths.
1 Cf. Stuart Hampshire, ‘Vico and the Contemporary Philosophy of Language,’ in G. Tagliacozzo (ed.), Giambattista Vico, An International Symposium (Baltimore, 1969).
1 Edited by H. Steinthal (Berlin, 1883).
1. Cf. R. L. Brown, Wilhelm von Humboldt’s Conception of Linguistic Relativity (The Hague, 1967) and Robert L. Miller, The Linguistic Relativity Principle and Humboldlian Ethno-linguistics (The Hague, 1968).
1 In D. Mandelbaum (ed.), Selected Writings in Language, Culture and Personality by Edward Sapir (Berkeley and Los Angeles, 1949).
1 Language, Thought, and Reality: Selected Writings of Benjamin Lee Whorf, ed. John B. Carroll (Cambridge, Mass., 1956), p. 252.
1 Language, Thought, and Reality, p. 60.
1 N. Chomsky, Aspects of the Theory of Syntax (Cambridge, Mass., 1965), pp. 27–8.
1 Charles A. Ferguson, ‘Assumptions about Nasals: A Sample Study in Phonological Universals,’ in J. H. Greenberg (ed.), Univenals of Language (Cambridge, Mass., 1963), p. 56.
1 Joseph H. Greenberg, ‘Some Universals of Grammar with Particular Reference to the Order of Meaningful Elements’ in op. cit., pp. 73–113.
1 These and the immediately following quotations are taken from N. Chomsky’s review of B. F. Skinner’s Verbal Behavior. First published in Language, 35 (1959), the article is reprinted in John P. De Cecco (ed.), The Psychology of Language, Thought, and Instruction (New York and London, 1967).
1 N. Chomsky, Aspects of the Theory of Syntax, p. 35.
1 Ibid., pp. 180 ff.
1 E. Bach and R. T. Harms (eds.), Universals in Linguistic Theory (New York, 1968), p. 121.
2 Aspects of the Theory of Syntax, pp. 121–2. In Problems of Knowledge and Freedom (New York, 1971), Chomsky puts forward a more cautious view: ‘It is reasonable to formulate the hypothesis that such principles are language universals. Quite probably the hypothesis will have to be qualified as research into the variety of languages continues.’
1 Cf. C. Tagliavini, ‘Di alcune denominazioni della pupilla’ in Annali dell Istituto Universitario di Napoli (1949).
2 Stephen Ullmann, ‘Semantic Universals,’ in J. H. Greenberg (ed.), Universals of Language, p. 221.
1 N. Chomsky, Aspects, p. 29.
1 Robert A. Hall, Jr., An Essay on Language (Philadelphia, 1968), pp. 53–4. For a sanely balanced discussion of the respective, ultimately collaborative claims and merits of Whorfian and universalist linguistics, cf. Helmut Gipper: ‘Der Beitrag der inhaltlich orientierten Sprachwissenschaft zur Kritik der historischen Vernunft,’ in Das Problem der Sprache, ed. Hans-Georg Gadamer (Munich, 1967), pp. 420–5; also, in the same symposium, Wilhelm Luther, ‘Sprachphilosophie und geistige Grundlagenbildung,’ pp. 528–31. Johannes Lohmann’s Philosophic und Sprachwissenschaft (Berlin, 1965) contains a fascinating but idiosyncratic argument for a division of world languages into six fundamental structural types, each correlated with certain ways of experiencing the world, and each corresponding to certain phonetic and alphabetic features. A careful survey of the evidence, and further bibliography, may be found in Helmut Gipper, Bausteine zur Sprachinhaltsforschung (Düsseldorf, 1963), pp. 215 ff. Cf. also the important debate on the linguistic determination of Greek philosophic terms between E. Benveniste in Problèmes de linguislique générate (Paris, 1966), pp. 63 ff., and P. Auberique, ‘Aristote et le langage, note annexe sur les catégories d’Aristote. A propos d’un article de M. Benveniste’ (Annales de la faculté des lettres d’Aix, 43 (1965)). This debate and its implications are in turn reviewed by Jacques Derrida in Marges de la philosophic (Paris, 1972), pp. 214–46.
1 N. Chomsky, Aspects, p. 30, and the relevant footnote on pp. 201–2.
1 Robert A. Hall, Jr., An Essay, p. 53.
2 Ibid., p. 77.
1 Ibid., p. 72.
2 The case is put succinctly by I. A. Richards in ‘Why Generative Grammar does not Help’ (English Language Teaching, 22, i.and ii (1967–8)). An expanded version of this critique forms Chapter IV of Richards’s So Much Nearer: Essays Towards a World English (New York, 1970).
1 New York Review of Books (8 February 1973), p. 34.