Is That a Fish in Your Ear?

EIGHT Words Are Even Worse

In Russian, there are two words, голубой and синий, that mean “blue,” but they do not have the same meaning. The first is used for light or pale blue hues, the second for darker, navy or ultramarine shades. So both can be translated into English, subject to the addition of words that specify the quality of blueness involved. But you can’t translate plain English blue back into Russian, because whatever you say—whichever of the two adjectives you use—you can’t avoid saying more than the English said. The conventions that hold sway among publishers and the general public do not allow translators to add something that is not in the original text. So if you accept those terms of the trade, you could quickly arrive with impeccable logic at the conclusion that translation is completely impossible.

Observations of this kind have been used by many eminent scholars to put translation outside of the field of serious thought. Roman Jakobson, a major figure in the history of linguistics, pointed out that сыр, the Russian word for “cheese,” cannot be used to refer to cottage cheese, which has another name, творог, in Russian. As he puts it, “the English word ‘cheese’ cannot be completely identified with its Russian heteronym.”[39] As a result, there is no fully adequate Russian translation of something as apparently simple as the word cheese.

It’s an indisputable fact about languages that the sets of words that each possesses divide up the features of the world in slightly and sometimes radically different ways. Color terms never match up completely, and it’s always a problem for a French speaker to know what an English speaker means by “brown shoes,” since the footwear in question may be marron, bordeaux, even rouge foncé. The names of fishes and birds often come in nonmatching sets of labyrinthine complexity; similarly, fixed formulae for signing-off letters come in graded levels of politeness and servility that have no possible application outside of the culture in which they exist.

These well-known examples of the “imperfect matching,” or anisomorphism, of languages do not really support the conclusion that translation is impossible. If the translator can see the sky that’s being called blue—either the real one or a representation of it in a painting, for example—then it’s perfectly obvious which Russian color term is appropriate; similarly, if the cheese being bought at the shop is not cottage cheese, the choice of the Russian term is not an issue. If, on the other hand, what’s being translated is a sentence in a novel, then it really doesn’t matter which kind of Russian blue is used to qualify a dress that exists only in the reader’s mental image of it. If the specific shade of blue becomes relevant to some part or level of the story later on, the translator can always go back and adjust the term to fit the later development. The lack of exactly matching terms is not as big a problem for translation as many people think it is.

Pocket dictionaries contain common, frequently used words, and their larger brethren are fattened up with words used less often. Most of those additional words are nouns with relatively precise and sometimes recondite meanings, such as polyester, recitative, or crankset. It’s trivially easy to translate words of that sort into the language of any community that has occasion to refer to synthetic fibers, Italian opera, or bicycle maintenance. Large authoritative dictionaries thus create the curious illusion that most of the words in a language are automatically translatable by slotting in the matching term from the dictionary. But there’s a huge difference between most of the headwords in a dictionary and the words that occur most often in the use of a language. In fact, just two or three thousand items account for the vast majority of word occurrences in all utterances in any language—and they aren’t words like crankset, recitative, or polyester at all.[40]

If translation were a matter of slotting in matching terms, then translation would clearly be impossible for almost everything we say except for our fairly infrequent references to a very large range of specific material things. Conversely, those many people who come up with the false truism that translation is impossible certainly wish that all words were like that. A desire to believe (despite all evidence to the contrary) that words are at bottom the names of things is what makes the translator’s mission seem so impossible.

The idea that a language is a list of names for the things that exist runs through Western thought from the Hebrews and Greeks to the man in the street by way of many distinguished minds. Leonard Bloomfield, a professor of linguistics who dominated the field in the United States for more than twenty-five years, tackled the problem of meaning in the textbook he wrote in the following way. Let us take the word salt. What does it mean? In Bloomfield’s book, the token salt is said to be the label of sodium chloride, more accurately (or at least, more scientifically) designated by the symbol NaCl. But Bloomfield was obviously aware that not many words of a language are amenable to such simple analysis. You can’t get at the meaning of words such as love or anguish in the same way. And so he concludes:

Indeed it is, if you go about it that way.

I still find it bewildering that a man of Bloomfield’s vast knowledge and intelligence should ever have thought that “NaCl” or “sodium chloride” constitute the meaning of the word salt. What they are is obvious: they are translations of salt into different registers of language. But even if we revise Bloomfieldian naïveté in this way, we are still trapped inside the idea that words (translated into whatever other register or language you like) are the names of things.

One well-known reason so many people believe words to be the names of things is because that’s what they’ve been told by the Hebrew Bible:

This short verse has had long-lasting effects on the way language has been imagined in Western cultures. It says that language was, to begin with, and in principle still is, a list of words; and that words are the names of things (more particularly, the names of living things). Also, it says very succinctly that language is not among the things that God created but an arbitrary invention of humankind, sanctioned by divine assent.

Nomenclaturism—the notion that words are essentially names—has thus had a long history; surreptitiously it still pervades much of the discourse about the nature of translation between languages, which have words that “name” different things or that name the same things in different ways. The problem, however, doesn’t really lie in translation but in nomenclaturism itself, for it provides a very unsatisfactory account of how a language works. A simple term such as head, for example, can’t be counted as the “name” of any particular thing. It figures in all kinds of expressions. It can be used to refer to a rocky promontory (“Beachy Head,” in Sussex), a layer of froth (“a nice head of beer”), or a particular role in a bureaucratic hierarchy (“head of department”). What connects these disparate things? How do we know which meaning head has in these different contexts? What does it mean, in fact, to say that we know the meaning of the word head? That we know all the different things that it means? Or that we know its real meaning but can also cope with it when it means something else?

One solution proposed to the conundrum of words and meanings is to tell a story about how a word has come to mean all the things for which it serves. The story of the word head, for example, as told in many dictionaries, is that once upon a time it had a central, basic, or original reference to that part of the anatomy which sits on top of the neck. Its meaning was subsequently extended to cover other kinds of things that sit on top of something else—a head of beer and a head of department would represent extensions of that kind. But as familiar animals with four feet instead of two have their anatomical heads not at the top but at the front, head was extended in a different direction to cover things that stick out (Beachy Head; the head of a procession).

Some such stories can be supported with historical evidence, from written texts representing an earlier state of the same language. The study of how words have in fact or must be supposed to have altered or extended their meanings is the field of historical semantics. But however elaborate the story, however subtle the storyteller, and however copious the documentary evidence, historical semantics can never tell you how any ordinary user of English just knows (a) that head is a word and (b) all of the things that head means.

From this it follows that the word head cannot be translated as a word into any other language. But the meaning it has in any particular usage can easily be represented in another language. In French, for example, you would use cap for “Beachy Head,” mousse for “head of beer,” and chef, patron, or supérieur hiérarchique to say “head of department.” Translation is in fact a very handy way of solving the conundrum of words and meanings. That’s not to say that anyone can tell you what the word means in French or any other language. But what you can say by means of translation is what the word means in the context in which it occurs. That’s a very significant fact. It demonstrates a wonderful capacity of human minds. Translation is meaning.

Linguists and philosophers have nonetheless devised Houdini-like ways of extricating themselves from the self-imposed dilemma of having to account for what words mean qua words. Head is considered a single word with a range of transferred or figurative meanings and can serve as an example of polysemy. Yet equally common words such as light are treated as a pair of homonyms—two different words having the same form in speech and writing—one of them referring to weight (as in “a light suitcase”), the other to luminosity (as in “the light of day”). The distinction between polysemy and homonymy is completely arbitrary from the point of view of language use.[42] Where the spelling is different but the sound the same, as in beat and beet, linguists switch terms and give them as examples of homophony. Yet more subdivisions can be made in the tendencies of words to drift from one meaning to another. A part can stand for a whole when you have fifty head in a flock, or the whole can stand for the part, as when you refer to a sailor walking into a bar as the arrival of the fleet. Sometimes there is or is said to be a visual analogy between the central meaning of a word and one of its extensions, as when you nose your car into a parking slot, and this is called metaphor; sometimes the extension of meaning is the supposed fruit of contiguity or physical connection, as when you knock on doors in your attempt to get a job, and this is called metonymy. The machinery of “figures of meaning,” taught for centuries as part of the now-lost tradition of rhetoric, is fun to play with, but at bottom it’s eyewash. Polysemy, homonymy, homophony, metaphor, and metonymy aren’t terms that help to understand how words mean, they’re just fuzzy ways of holding down the irresistible desire of words to mean something else. It would take a very imaginative language maven indeed to explain satisfactorily why the part of a car that covers either the engine or the luggage compartment is called a “bonnet” in the U.K. and a “hood” in the United States. Despite the enthusiasm of the large throng of hobbyists who contribute to it, the semantics of words is an intellectual mess.

All the same, most languages have words for the same kinds of things and don’t bother with words for things they don’t have or need. They tend to have separate expressions for basic orientation—(up, down, left, right—but see chapter 14 for languages that do not), for ways of moving (run, walk, jump, swim) and for directional movement (come, go, leave, arrive), for family relations (son, daughter, brother’s wife, and so on), for feelings and sensations (hot, cold, love, hate), for life events (birth, marriage, death, sickness, and health), for types of clothing, food, and animals, for physical features of the landscape, and for the cardinal numbers (up to five, ten, twelve, or sixteen). Some have words for fractional numbers, such as the German anderthalb (one and a half) or the Hindi sawa (one and a quarter)—but I don’t believe any has a separate item for the number 2.375. All languages used in societies that have wheeled vehicles have words for wheeled vehicles of various kinds, but none, as far as I know, has a single lexical item with the meaning “wheeled vehicles with chrome handlebars,” so as to refer collectively to bicycles, tricycles, tandems, mopeds, motorcycles, strollers, and lawn mowers. French may have single words for “the whole contents of a deceased sailor’s sea chest” (hardes) and “gravelly soil suitable for growing vines” (grou), but in practice all sorts of real and possible things, classes of things, actions, and feelings don’t have names in most languages. English, for instance, does not possess a designated term for the half-eaten pita bread placed in perilous balance on the top of a garden fence by an overfed squirrel that I can see right now out of my study window, but this deficiency in my vocabulary doesn’t prevent me from observing, describing, or referring to it. Conversely, the existence in Arabic of ghanam, a word that means “sheep” and “goats” without distinction, does not prevent speakers of Arabic from sorting the sheep from the goats when they need to. Just because English does not have a one-word or phrasal counterpart to the French je ne sais quoi or German Zeitgeist, it in no way prevents me from knowing how to say what these words mean. Far from providing labels for “all the things in the world,” languages restrict their word lists to an ultimately arbitrary range of states and actions, while also having means to talk about anything that comes up. The peculiar flexibility of human languages to bend themselves to new meanings is part of what makes translation not only possible but a basic aspect of language use. Using one word for another isn’t special; it’s what we do all the time. Translators just do it in two languages.

One formerly fashionable way of avoiding the insoluble problem of fixing the meaning of a word was to imagine it as the compound product of sublinguistic mental units or “features” of meaning. Take the three words house, hut, and tent. They can all be used to refer to dwellings of some sort, but they refer to three different kinds of dwelling. The task of distinctive feature analysis was to find the minimal semantic constituents that would account for the meaning relations among these three semantically related terms. All three are “marked” with the feature [+dwelling], but only house also has the two features [+permanent] and [+brick]. Tent would be marked [–perma-nent] [–brick] and hut could be marked [+permanent] [–brick]. How wonderful it would be if all words in the language could be decomposed into atoms of meaning in this way. The meaning of a word would then be fully specified through the list of the distinctive features that mark it. If you could show that it was possible to account for the differences in the meanings of all the words in a language by the distribution of a finite set of semantic features, then you could go further still. You would be in a position to build a great Legoland of the mind, in which all possible meanings could be constructed out of irreducible, binary building blocks of sense.

To map some area of vocabulary (let alone a whole language) using only such elementary features of meaning is an enticing prospect, but it runs up against a fundamental problem: what criterion to use to establish the list of the elementary semantic features themselves. Common sense no doubt dictates that [±animate] and [±female] are among the distinctive features relevant to the meaning of the term woman and that [±chrome-plated] is not. But common sense appeals to our total experience of the nonlinguistic world as well as to our ability to find a way through the language maze: it is precisely the kind of fuzzy, vague, and informal knowledge that distinctive feature analysis seeks to overcome and replace. Despite the usefulness of binary decomposition for some kinds of linguistic description and (in far more complex form) in the “natural language processing” that computers can now perform, word meanings can never be fully specified by atomic distinctions alone. People are just too adept at using words to mean something else.

Such quasi-mathematical computation of “meaning” is equally unable to solve an even more basic problem, which is how to identify the very units whose meaning is to be specified. To ask what a word means (and translators often are asked to say what this or that word means) is to suppose that you know what word you are asking about, and that in turn requires you to know what a word is. The word word is certainly a familiar, convenient, and effective tool in the mental toolbox we use to talk about language. But it is uncommonly hard to say what it means.

Computers must know the answer, because they count words. That’s no consolation to us, however. What computers know about words is what they’ve been told, which comes down to this: a word is a string of alphabetic characters bounded on left and right by a space or one of these typographical symbols:—/ ? ! : ; , .[43] Computers don’t need to know what a word means to carry out the operations we ask of them. But we do! And if in some instance we really don’t, then we try to find out from a dictionary, from an acquaintance, or from listening to how other people speak. But all kinds of problems remain.

In languages such as English the identification of words is more art than science. Publishers have their own style sheets with rules for deciding whether couples have break-ups or break ups or breakups; but ordinary people also want to know if “to break up” should be counted as one, two, or three words. Yet nobody can really say.[44]

English prepositional verbs provide unending employment for language experts who want to determine what a word is. They come in three or four parts. Sometimes they stay together—“Did you remember to take out the trash?”—and sometimes they don’t: “I promised to take my daughter out to see a film.” Does that mean that “to take out” is a word (or three) or two different words—“to take out” and “to take … out”—(or six) that look the same? Compilers of alphabetical dictionaries adopt practical solutions, but not the same ones, leaving the underlying question—what word is this?—unresolved. Teachers of English as a second language know the best answer to the question of how many words there are in a prepositional verb. If you want to know how to use the language properly, don’t ask.

Given the labyrinthine complexity of the variable terminologies and conflicting expert solutions to the conundrum of establishing what the word units are in perfectly ordinary English expressions, it seems fairly obvious that an ordinary user of a language such as English doesn’t need to know what a word is—or what word it is—in order to make sense. Wordhood is often a useful notion, but it is not a hard-edged thing.

Other languages undermine the wordness of words in a variety of different ways. German runs them together to make new ones. Lastkraftwagenfahrer (truck driver) is of course a single word in ordinary use, but it can easily be seen as two words written next to each other (Lastkraftwagen plus Fahrer, “truck” plus “driver”), or as three words run together (Last plus Kraftwagen plus Fahrer, “freight” plus “motor vehicle” plus “driver”), or as four (Last plus Kraft plus Wagen plus Fahrer, “freight” plus “power” plus “vehicle” plus “driver”). Hungarian also melds what we think of as many separate words, but in a different and equally elegant way. What a computer would count as a three-word expression, Annáékkal voltunk moziban, for example, would be expressed in English by around a dozen words: “We were [voltunk] at the cinema [moziban] with Anna and her folk” (that’s to say, friends or relatives or hangers-on, without distinction). The modest suffix -ék is all that is needed to turn Anna into a whole group, and the “glued-on” or agglutinated addition -kal says that you were part of it, too. Indeed, at my younger daughter’s wedding in London in 2003, in honor of her Hungarian grandparents I was able (after doing my homework) to raise a toast édeslányaméknak, which is to say in one word “to my dear daughter’s husband, in-laws, and friends.”

Classical Greek has no proper word for word; moreover, in manuscripts and monuments from the earlier period, Greek is written without spaces between words. But that does not automatically mean that Greek thinkers had no concept of a basic unit of language smaller than the utterance. There is evidence of word dividers in Greek written in Linear B and Cyprian, ancient scripts that predate the Greek alphabet, and in various other ways a notion of “basic unit” does seem to emerge even in a language that supposedly has no “word” for the unit thus distinguished.[45] Even Hungarians recognize that some “words” are more basic than others, that beneath the practically infinite welter of possible agglutinated and compounded forms lie nuggets that are the elementary building blocks of sense. Gyerek is Hungarian for “child,” and though it may almost never occur in that form in any actual expression, it is nonetheless the “root” or “stem” corresponding to the English stem word child. Without an operative concept of the meaning units of which a language is made, it would be hard to imagine how a dictionary could be constructed. And without a dictionary, how would anyone ever learn a foreign tongue, let alone be able to translate it?