Through the Language Glass: Why the World Looks Different in Other Languages

by Guy Deutscher

  The language of legal proceedings, with its zealous insistence on accurate, explicit, and context-independent statements, is an extreme example of the type of elaborate communicative patterns that are more likely to arise in a complex society. But it is not the only example. As I mentioned earlier, in a large society of strangers there will be many more occasions where elaborate information has to be conveyed without reliance on shared background and knowledge. Finite complements are better equipped to convey such information than alternative constructions, so it is plausible that finite complements are more likely to emerge under the communicative pressures of a more complex society. Of course, as no statistical surveys about subordination have been conducted yet, speculations about correlations between subordination and the complexity of a society necessarily have to remain on an impressionistic level. But there are signs that things might be changing.

  For decades, linguists have elevated the hollow slogan that “all languages are equally complex” to a fundamental tenet of their discipline, zealously suppressing as heresy any suggestion that the complexity of any areas of grammar could reflect aspects of society. As a consequence, relatively little work has been done on the subject. But a flurry of publications from the last couple of years shows that more linguists are now daring to explore such connections.

  The results of this research have already revealed some significant statistical correlations. Some of these, such as the tendency of smaller societies to have more complex word structure, may seem surprising at first sight, but look plausible on closer examination. Other connections, such as the greater reliance on subordination in complex societies, still require detailed statistical surveys, but nevertheless seem intuitively convincing. And finally, the relation between the complexity of the sound system and the structure of society awaits a satisfactory explanation. But now that the taboo is lifting and more research is being done, there are undoubtedly more insights in store. So watch this space.

  We have come a long way from the Aristotelian view of how nature and culture are reflected in language. Our starting point was that only the labels (or, as Aristotle called them, the “sounds of speech”) are cultural conventions, while everything behind those labels is a reflection of nature. By now culture has emerged as a considerable force whose influence extends far beyond merely bestowing labels on a preordained list of concepts and a preordained system of grammatical rules.

  In the second part of the book, we move on to a question that may seem a fairly innocuous corollary to the conclusions of the first part: does our mother tongue influence the way we think? Since the conventions of the culture we were born into affect the way we carve up the world into concepts and the way we organize these concepts into elaborate ideas, it seems only natural to ask whether our culture can affect our thoughts through the linguistic idiosyncrasies it has imposed on us. But while raising the question appears harmless enough in theory, among serious researchers the subject has become a pariah. The following chapter explains why.

  * There has been a lot of brouhaha in the last few years about Pirahã, a language from the Brazilian Amazon, and its alleged lack of subordination. But a few Pirahã subordinate clauses have recently managed to escape from the jungle and telegraph reliable linguists to say that reports of their death have been greatly exaggerated. (See notes for more information.)

  PART II

  The LANGUAGE LENS

  6

  Crying Whorf

  In 1924, Edward Sapir, the leading light of American linguistics, was entertaining no illusions about the attitude of outsiders toward his field: “The normal man of intelligence has something of a contempt for linguistic studies, convinced as he is that nothing can well be more useless. Such minor usefulness as he concedes to them is of a purely instrumental nature. French is worth studying because there are French books which are worth reading. Greek is worth studying—if it is—because a few plays and a few passages of verse, written in that curious and extinct vernacular, have still the power to disturb our hearts—if indeed they have. For the rest, there are excellent translations. . . . But when Achilles has bewailed the death of his beloved Patroclus and Clytaemnestra has done her worst, what are we to do with the Greek aorists that are left on our hands? There is a traditional mode of procedure which arranges them into patterns. It is called grammar. The man who is in charge of grammar and is called a grammarian is regarded by all plain men as a frigid and dehumanized pedant.”

  In Sapir’s own eyes, however, nothing could be further from the truth. What he and his colleagues were doing did not remotely resemble the pedantic sifting of subjunctives from aorists, moldy ablatives from rusty instrumentals. Linguists were making dramatic, even worldview-changing discoveries. A vast unexplored terrain was being opened up, the languages of the American Indians, and what was revealed there had the power to turn on its head millennia’s wisdom about the natural ways of organizing thoughts and ideas. For the Indians expressed themselves in unimaginably strange ways and thus demonstrated that many aspects of familiar languages, which had previously been assumed to be simply natural and universal, were in fact merely accidental traits of European tongues. The close study of Navajo, Nootka, Paiute, and a panorama of other native languages catapulted Sapir and his colleagues to vertiginous heights, from where they could now gaze down on the languages of the Old World like people who see their home patch from the air for the first time and suddenly recognize it as just one little spot in a vast and varied landscape. The experience was exhilarating. Sapir described it as the liberation from “what fetters the mind and benumbs the spirit . . . the dogged acceptance of absolutes.” And his student at Yale Benjamin Lee Whorf enthused: “We shall no longer be able to see a few recent dialects of the Indo-European family . . . as the apex of the evolution of the human mind. They, and our own thought processes with them, can no longer be envisioned as spanning the gamut of reason and knowledge but only as one constellation in a galactic expanse.”

  It was difficult not to get carried away by the view. Sapir and Whorf became convinced that the profound differences between languages must have consequences that go far beyond mere grammatical organization and must be related to profound divergence in modes of thought. And so in this heady atmosphere of discovery, a daring idea about the power of language shot to prominence: the claim that our mother tongue determines the way we think and perceive the world. The idea itself was not new—it had been lying around in a raw state for more than a century—but it was distilled in the 1930s into a powerful concoction that then intoxicated a whole generation. Sapir branded this idea the principle of “linguistic relativity,” equating it with nothing less than Einstein’s world-shaking theory. The observer’s perceptions of the world—so ran Sapir’s emendation of Einstein—depend not only on his inertial frame of reference but also on his mother tongue.

  The following pages tell the story of linguistic relativity—a history of an idea in disgrace. For as loftily as it had once soared, so precipitously did the theory then crash, when it transpired that Sapir and especially his student Whorf had attributed far-fetched cognitive consequences to what were in fact mere differences in grammatical organization. Today, any mention of linguistic relativity will make most linguists shift uneasily in their chairs, and “Whorfianism” has largely become an intellectual tax haven for mystical philosophers, fantasists, and postmodern charlatans.

  Why then should one bother telling the story of a disgraced idea? The reason is not (just) to be smug with hindsight and show how even very clever people can sometimes be silly. Although there is undeniable pleasure in such an exercise, the real reason for exposing the sins of the past is this: although Whorf’s wild claims were largely bogus, I will try to convince you later that the notion that language can influence thoughts should not be dismissed out of hand. But if I am to make a plausible case that some aspects of the underlying idea are worth salvaging and that language may after all function as a lens through which we perceive the world, then
this salvaging mission must steer clear of previous errors. It is only by understanding where linguistic relativity went astray that we can turn a different way.

  WILHELM VON HUMBOLDT

  The idea of linguistic relativity did not emerge in the twentieth century entirely out of the blue. In fact, what happened at Yale—the overreaction of those dazzled by a breathtaking linguistic landscape—was a close rerun of an episode from the early 1800s, during the high noon of German Romanticism.

  The prevailing prejudice toward the study of non-European languages that Edward Sapir gently mocked in 1924 was nothing to poke fun at a century earlier. It was simply accepted wisdom—not just for the “ordinary man of intelligence” but also among philologists themselves—that the only languages worthy of serious study were Latin and Greek. The Semitic languages Hebrew and Aramaic were occasionally thrown into the bargain because of their theological significance, and Sanskrit was grudgingly gaining acceptance into the club of classical worthies, but only because it was so similar to Greek and Latin. But even the modern languages of Europe were still widely viewed as merely degenerate forms of the classical languages. Needless to say, the languages of illiterate tribes, without great works of literature or any other redeeming features, were seen as devoid of any interest, primitive jargons just as worthless as the primitive peoples who spoke them.

  It was not that scholars at the time were unconcerned about the question of what is common to all languages. In fact, from the seventeenth century onward, the writing of learned treatises on “universal grammar” was very much in vogue. But the universe of these universal grammars was rather limited. Around 1720, for instance, John Henley published in London a series of grammars called The Compleat Linguist; or, An Universal Grammar of All the Considerable Tongues in Being. All the considerable tongues in being amounted to nine: Latin, Greek, Italian, Spanish, French, Hebrew, Chaldee (Aramaic), Syriac (a later dialect of Aramaic), and Arabic. This exclusive universe offered a somewhat distorted perspective, for—as we know today—the variations among European languages pale in significance compared with the otherness of more exotic tongues. Just imagine what misleading ideas one would get on “universal religion” or on “universal food” if one limited one’s universe to the stretch between the Mediterranean and the North Sea. One would travel in the different European countries and be impressed by the great divergences between them: the architecture of the churches is entirely different, the bread and cheese do not taste at all the same. But if one never ventured to places farther afield, where there were no churches, cheese, or bread, one would never realize that these intra-European differences are ultimately minor variations in essentially the same religion and the same culinary culture.

  In the second half of the eighteenth century, the view was beginning to widen slightly, as various attempts were made to compile “universal dictionaries”—lists of equivalent words in languages from different continents. But although the scope and ambition of these catalogs gradually grew, they didn’t go much beyond a linguistic cabinet of curiosities showcasing weird and wonderful words. In particular, the dictionaries revealed little of value about the grammar of exotic languages. Indeed, for most philologists at the time, the notion that the grammar of a barbarian language could be a worthwhile subject of study seemed perverse. Studying grammar meant the study of Greek and Latin, because “grammar” was the grammar of Greek and Latin. So when remote languages were described (not by philologists but by missionaries who needed them for practical purposes), the descriptions usually consisted of a list of Latin paradigms on one side and the allegedly corresponding forms in the native language on the other side. The nouns in an American Indian language, for example, would be shown in six forms, corresponding to the six cases of the Latin noun. Whether or not the language in question made any case distinctions was irrelevant—the noun would still be duly frogmarched into nominative, genitive, dative, accusative, vocative, and ablative. The French writer Simon-Philibert de La Salle de l’Étang demonstrates this frame of mind in his 1763 dictionary of Galibi, a now extinct language of the Caribbean, when he complains that “the Galibis have nothing in their language that makes distinctions of case, for which there should be six in the declension of each word.” Such descriptions seem to us today like clumsy parodies, but they were conceived in complete earnestness. The notion that the grammar of an American Indian language might be organized on fundamentally different principles from those of Latin was simply beyond the intellectual horizon of the writers. The problem was much deeper than the failure to understand a particular feature of the grammar of a particular New World language. It was that many of the missionaries didn’t even understand that there was something there to understand.

  Enter Wilhelm von Humboldt (1767–1835), linguist, philosopher, diplomat, educational reformer, founder of the University of Berlin, and one of the stellar figures of the early nineteenth century. His education—the best of what the Berlin Enlightenment scene had to offer—imbued him with unbounded admiration for classical culture and for the classical languages. And until he reached the age of thirty-three, there was little to show that he would one day break out of the mold or that his linguistic interests would ever extend beyond the revered Latin and Greek. His first publication, at the age of nineteen, was about Socrates and Plato; he then wrote about Homer and translated Aeschylus and Pindar. A happy lifetime of classical scholarship seemed to stretch in front of him.

  Wilhelm von Humboldt, 1767–1835

  His linguistic road to Damascus led through the Pyrenees. In 1799, he traveled to Spain and was greatly taken with the Basque people, their culture, and their landscape. But above all, it was their language that aroused his curiosity. Here was a language spoken on European soil but entirely unlike all other European tongues and clearly from a different stock. Back from the journey, Humboldt spent months reading through everything he could find about the Basques, but as there wasn’t very much in the way of reliable information, he returned to the Pyrenees to do serious fieldwork and learn the language firsthand. As his knowledge deepened, he realized the extent to which the structure of this language—rather than merely its vocabulary—diverged from everything else he knew and from what he had previously taken as the only natural form of grammar. The revelation gradually dawned on him that not all languages were made in the image of Latin.

  Once Humboldt’s curiosity was aroused, he tried to find descriptions of even more remote tongues. There was almost nothing published at the time, but the opportunity to discover more presented itself when he became the Prussian envoy to the Vatican in 1802. Rome was teeming with Jesuit missionaries who had been expelled from their missions in Spanish South America, and the Vatican library contained many manuscripts with descriptions of South and Central American languages that these missionaries had brought with them or written once back in Rome. Humboldt trawled through such grammars, and with his eyes now wide open after his experience with Basque, he could make out how distorted a picture they presented: structures that deviated from the European type had either passed unnoticed or been coerced to fit the European mold. “It is sad to see,” he wrote, “what violence these missionaries exerted both on themselves and on the languages, in order to force them into the narrow rules of Latin grammar.” In his determination to understand how the American languages actually worked, Humboldt completely rewrote many of these grammars, and gradually the real structure of the languages emerged from behind the facade of Latin paradigms.

  Humboldt set linguists on a steep learning curve. Of course, the secondhand information that he was able to glean about American Indian languages was nothing like the deep firsthand knowledge that Sapir developed a century later. And considering what we know today about how the grammars of different languages are organized, Humboldt was barely scratching the surface. But the dim ray of light that shone from his materials felt dazzling nonetheless because of the utter darkness in which he and his contemporaries had languished.

  For
Humboldt, the elation of breaking new ground was mixed with frustration at the need to impress the value of his discoveries upon an uncomprehending world, which persisted in regarding the study of primitive tongues as an activity fit only for butterfly collectors. Humboldt went to great lengths to explain why the profound dissimilarities among grammars were in fact a window into far greater things. “The difference between languages,” he argued, “is not only in sounds and signs but in worldview. Herein is found the reason and ultimate goal of all the study of language.” But this was not all. Humboldt also claimed that grammatical differences not only reflect preexisting differences in thought but are responsible for shaping these differences in the first place. The mother tongue “is not just the means for representing a truth already recognized but much more to discover the truth that had not been recognized previously.” Since “language is the forming organ of thought,” there must be an intimate relation between the laws of grammar and the laws of thinking. “Thinking,” he concluded, “is dependent not just on language in general but to a certain extent on each individual language.”