Babel No More

by Michael Erard

  But their fractiousness is superficial. Neither scholar doubted Mezzofanti’s achievement or its glory, and, interestingly, neither one invoked the divine (angels, tongues of fire) to explain his gifts, not even Russell, for whom Mezzofanti’s figure carried a religious charge. More significant, they both agreed that Mezzofanti had mastered thirty languages. Mastered them.

  Despite Russell’s careful accounting and Watts’s surgical follow-up, disentangling Mezzofanti from folk legend and the preoccupations of his contemporaries is far from straightforward. It’s not only a matter of how you define what it means to speak a language; it also matters who does the listening.

  We commonly assume that native speakers are qualified to judge how well someone uses their language. But their opinions of another’s mastery can take a cultural coloring. In Bologna, Italians welcomed my small attempts with praise, as have Mexicans and Colombians with Spanish. In Taiwan and China, people responded to my elementary abilities with polite enthusiasm: “Oh, you speak Chinese very well!” “Really, you don’t have to be so polite,” I’d reply in Chinese. I took the titters of shocked delight to mean that the sophistication of the reply (which I’d learned from a friend) had outstripped their true opinion of my skill.

  By contrast, the French are vehemently uninterested in having a foreigner mangle their language—in stereotype, anyway. Theoretically, French natives would rate you lower in French than Spaniards would in Spanish. In Japan and Korea, a lower-skilled non-native will be highly regarded and praised, while someone with better skills will be viewed as a threat. Former US ambassador William Rugh, who was posted to Yemen and the United Arab Emirates, once counseled that if you’re a non-native Arabic speaker appearing on Arabic-language media, it’s better to keep a conversation going than to worry about being grammatically correct; the effort, which is appreciated by Arab audiences, looks good diplomatically. For cultural reasons, Arab audiences also prefer truthful speech, so they are more tolerant of imperfect Arabic than of an interpreter, who may alter what a person means. So various are all these responses that you’d have to conclude that a native speaker’s opinion isn’t necessarily a legitimate criterion.

  An even bigger problem is that one person’s “nativeness” in language X isn’t necessarily the same as another’s. Any close look at a speech community demonstrates that pronunciations, vocabularies, and grammars are heterogeneous across social divides, genders, and geographical areas. Two speakers of two dialects would both be considered native speakers, even though they might not be able to describe how their version differs from the standard—and though they might judge the other speaker’s variety as somehow wrong or incorrect. In sign languages, the problem is profound. If you define a native speaker as someone who uses a language from birth at home, then there are hardly any native signers, since the majority of deaf children are born to hearing, nonsigning parents. (I’m speculating, but perhaps this is why sign language courses are so popular in the United States, where college enrollments went up more than 16 percent from 2006 to 2009: though there’s a robust deaf culture, there’s no native signing community to which one, by definition, can’t ever belong.) Russell, the biographer of Mezzofanti, never says what sort of native speaker he’s gathering evidence from.

  Even if you could presume that each native speaker knows the same things that other natives know, it is a shallow standard, because one can pronounce words like a native speaker and also lack linguistic creativity. Confused with a real speaker, the mere mimic gets labeled as a “master” simply because she’s able to string together words uninterruptedly as if she knows what she’s doing. What’s often called “fluency” might be no more than confidence (or blitheness). It’s possible that Mezzofanti only ever aspired to “passing” in most of his languages. One piece of evidence for this is his clear phonetic enthusiasm, which Russell stingingly noted when describing Mezzofanti’s English (the italics are his): “If I were disposed to criticize it very strictly,” the Irishman wrote, “I might say (paradoxical as this may seem,) that, compared with the enunciation of a native, it was almost too correct to appear completely natural.”

  Less obviously, Mezzofanti’s social rank would have restricted an accurate read of his abilities. His meetings were probably fairly formal, which would have reduced unexpected or intimate topics. He could have controlled the meetings, too, so that none of them would endanger his linguistic reputation. Who would dare report that a person of such status couldn’t actually do what he claimed?

  And the Mezzofanti of legend grew in other directions. Travelers to Italy who embarked on the so-called Grand Tour sought out Catholic excesses that fascinated and disgusted them. Their accounts would have drawn Mezzofanti as a Romantic figure, a symbol of Catholicism’s vivacious ruin. In response, the Church would have asserted the opposite. Mezzofanti represented the Church’s ideal view of itself—conservative, theologically pure, and world-encircling—an image embedded in hagiographies like Russell’s.

  It’s also easy to overlook the fact that judgments of “mastery” vary from era to era and to assume that the “fluency” and “mastery” of the eighteenth century would mean the same now that they meant then. Only a small bit of digging turns this on its head. In 1875, for instance, knowing French to the satisfaction of Harvard College meant you could translate at sight “easy” French prose. You didn’t have to orate or converse in it or demonstrate your understanding with a French speaker.

  Another example comes from the life of the adventuring hyperpolyglot Sir Richard Francis Burton (1821–1890). To show his British military superiors that he knew Hindustani, he had to translate two Hindustani books into English, translate a handwritten text, write a short essay, and have a conversation.* (He passed.) Nevertheless, over his lifetime, many of his linguistic achievements were in the spoken mode—in 1853 he became one of only a dozen Christians to have sneaked into the holy city of Mecca by passing as an Indian Muslim—one possessing fluent, though accented, Arabic. Yet to his military superiors, “knowing” a language was likely to mean knowing its grammatical particulars and its life in texts.

  If time gets in the way of really knowing what someone like Mezzofanti or Burton was capable of, it also helps to explain why the hyperpolyglots of yesteryear seem to be bursting with languages while a modern educated person with a grasp of more than four is a rarity. One can talk about active language skills (talking, writing) and receptive skills (reading, listening); the receptive ones—which even monolinguals may have surprisingly a lot of in other languages—are generally easier to acquire and use. In the era of Mezzofanti and Burton, scholars spent far more time reading and translating texts—in receptive activities, in other words—than they spent communicating with people. I’m not saying that no one talked to other people in foreign languages; I’m saying that for the people who were going to go around saying they knew language X or Y, one could assume that their legitimate language activities were reading and translating, which are less taxing and stressful to the brain. You can get a lot of support for reading and translating through dictionaries and grammars. To converse without embarrassing yourself, you have to monitor what you hear and what you say in real time, and not only that, but voices in real life come with accents (which add social information) and environmental noises (which require focus); it’s also a very pragmatics-heavy activity. Thus, in Mezzofanti’s time, it would have been relatively easier to rack up languages, and to do so legitimately, than it would be today, when we seem to treat oral communication as the hallmark of “knowing” a language.

  Given the variety of historical lenses through which one can view the criteria for speaking or knowing a language, it’s simply impossible to assess definitively the claims about Mezzofanti’s ability in all of his, based on the evidence that Russell provides, anyway.

  An unavoidable conclusion is that a count of one’s languages is, at best, an imperfect convenience for talking about someone’s capabilities in them. A language isn’t a unit of measu
re like a kilo or an inch. What is the thing that one has when one has more than one language? Six languages with closely shared vocabulary and grammar don’t burden one’s memory or mental processing as much as six unrelated languages would. Likewise, six languages in which one can speak, read, and write don’t represent the same sort of cognitive investment for an adult learner as six languages in which one has varying degrees of proficiency across a variety of tasks.

  So what’s another way for us to grasp the scope of someone’s “cognitive investment”? Here are some possibilities.

  A folk notion is that when you dream in a language, you’ve crossed some threshold on a path to fluency. In the 1980s, Canadian psychologist Joseph De Koninck found that students of French who made the fastest progress were those who reported speaking French in their dreams sooner than fellow students. Among another group of students studying French, those who had more REM (rapid eye movement) sleep over the course of a six-week immersion program improved the most. For people who have more experience in their languages, perhaps this isn’t a workable measurement alternative, since bilinguals report that they speak, think, and hear in both of their languages. Often, what determines their dream language is the one they used right before they slept, not the one they know the best.

  A related notion is that when you really know a language, you think in it. In fact, the brain doesn’t think in any language. What people refer to as “thinking in a language” comes from being able to speak more immediately in a language without rehearsing it or translating it from a language one might know better; the spoken thought feels as if it’s closer to its source in the brain.

  Does speaking in a different language alter one’s perception? Can the structure of a language and the way its vocabulary maps meanings make the world more colorful, your friends more friendly, the trees wilder? The “linguistic relativity hypothesis,” or Sapir-Whorf hypothesis, as it’s alternately called, proposes that the language you speak actually molds your perception of the world. Quite literally, if two languages have a different range of color words, the person speaking both languages fluently will assign his perception of colors to two different names and perhaps categories. If that’s true, then the hyperpolyglot’s world must appear kaleidoscopic. Indeed, scientists have observed monolingual speakers of Korean using a term, paran sekj, or “blue,” to refer to a greener, less purple color than Korean-English bilinguals think of as “blue.” Other scientists have since seen how bilinguals categorize common containers and even conceptualize time differently from monolinguals. But this evidence is controversial, and the effects of language on cognition haven’t been isolated precisely.

  One question that polyglots don’t get asked is, “When you go crazy, what language do you go crazy in?” Which is too bad, because it’s been demonstrated that psychotic polyglots, it turns out, aren’t equally disordered in each of their languages. In one case recorded by British psychiatrist Felicity de Zulueta, her psychotic patient, a native English speaker, switched into Spanish because he knew that Zulueta also spoke the language. Both were then surprised that his hallucinations and disordered thoughts disappeared. “In Spanish . . . he felt he was ‘sane,’ but when he spoke in English, he went ‘mad,’” Zulueta wrote. In three other cases, Zulueta’s patients had disordered thoughts or heard voices in the language they had learned first and used most. Using a language that they spoke less frequently overall and learned later dismissed their delusions. In another case, a patient was equally psychotic in Italian and English, but heard voices only in Italian, her mother tongue. Not only that—in English she denied that she heard voices at all, whereas in Italian, she readily admitted hearing them. Other patients hear friendly voices in their native languages, hostile ones in their second languages. A subsequent researcher quipped that the more competent an insane person was in a language, the higher their degree of psychosis.

  Some scientists have suggested that the extra effort of using a second language jolts people out of a deluded state into reality. Others suggest that the deeper relationship to your first language makes you less inhibited, and so more likely to express what’s troubling you. In a language learned later, you can hide from your true self.

  People are unlikely to tell potential employers that they can be mentally unbalanced in two languages and unstable in a third, or that they dream in three languages but never in a fourth. Which underscores the convenience, in comparison, of counting hyperpolyglots’ languages and what they can actually use them for.

  Your opinion of Mezzofanti may also depend on how valuable you think a less-than-complete knowledge of a language is. I’m talking about more than a snippet or a bit, more than an exchange of pleasantries or asking for bus directions. A good working knowledge of the core of the language, one that allows you to have real interactions to achieve some purpose, albeit in a limited domain, is what’s at issue here.

  In the Western conception of what it means to know a language, these circumscribed abilities don’t seem to count for much. The dominant view seems to be that language is a discrete object out in the world. Learning it involves shuttling its pieces into your body; once you know it, it’s inside you. In the 1960s, linguists developed a twist on this by arguing that children became so good at language so quickly because when they were born, they already had pieces of language inside them. Just as you know you aren’t hungry anymore by some digestive instinct, you know a language when it reaches some predetermined mark inside you, nourishing and enlightening you. If you’ve gathered only a few pieces of a language, a snack, it can’t change you. It doesn’t count. Call this the “all or nothing” view.

  Nor does a bit of language matter as long as possessing a sole language is the political foundation of the nation-state, “a community imagined by language,” as Benedict Anderson has written. This notion emerged in Europe during Mezzofanti’s day. In the nationalist’s view, a citizen demonstrates her affiliations to the homeland by speaking and writing the national language fully. She preserves her affiliations by eschewing other languages, regional dialects, and nonstandard ways of speaking and writing. In this way, “nativeness” becomes as much a political project as a linguistic one. Speak like a citizen, speak like a native—it amounted to basically the same thing. In France, for instance, spoken and regional dialects were looked down upon in favor of the cultivated Parisian dialect; the French Revolution brought with it the unification of the language. Until the 1960s, very few Indonesians spoke bahasa Indonesia as a first or mother tongue; now millions of them speak what is, in fact, the country’s official language. Schools taught the standard language and governments created exams to test ability in that language. Soon, private companies began accepting the results of those exams for their own determinations of a person’s proficiency, his ability to serve an institution’s goals.

  Things get trickier when more than one language is involved. Here, to “know” a language means—at least in the folk view of languages—that you keep it separate from the others you might know. For a long time, bilinguals were criticized for speaking sentences that contained both their languages. This “code switching” is very rule-governed. Yet it was viewed as a person’s inability to keep things straight, and marked, therefore, their failure to know either language. Bilinguals were seen as abundantly imperfect or overburdened, another unfortunate implication of the “all or nothing” view.

  A bit of language matters more in parts of the world where language isn’t viewed as a discrete object, but something more diffuse and external, like clothes. You don’t put it inside yourself. Instead, you wrap yourself in it. Neither does it create some lens by which the essences of things take different forms. It’s a tool. A tool you use when you need it and as often as you need to—as I needed French and Italian in the Archiginnasio. Call this the “something and something” model.

  A bit of language matters in places where the language isn’t written down, or where not many people are literate, where fewer resources for making l
anguage a fixed, bounded thing exist. Also, a bit of language matters more in countries that have built nationhood around many languages than in those whose national identity is founded on just one. In southern India, for instance, languages appear to be more like uniforms or badges; you wear them to tell people your social identity—the class or caste you belong to, the region you come from, your religion, family, profession, and significantly, your gender. When they treat you like someone who speaks that language, then that’s who you are. But Mezzofanti didn’t come from any places like these. So what was he doing?

  One way to resolve these views is through the idea of “multicompetence.” This has gained some traction in the twenty years since British linguist Vivian Cook proposed it to describe the “language supersystem” or the “total language system” that multilinguals possess. Cook’s goal was to help to see second-language speakers as “successful multicompetent speakers, not failed native speakers,” as he wrote. He meant to replace the “all or nothing” view with the “something and something” view. This means that a multilingual person doesn’t carry two or three monolingual speakers’ worth of language in his or her head. And not only that, you can’t say they know “less than” monolinguals or any other comparisons that make them seem like half-empty glasses. It’s not correct—and it’s unfair—to compare language learners to native speakers who know only their one language; it’s an apples to oranges comparison. The persistent mark of one’s status as a linguistic outsider is not a failure, but a difference. Take a native Mandarin speaker with no other languages and a businessman learning Chinese at the age of fifty. The native speaker can’t be more “successful” in her language, because she never set out to achieve it, while the businessman can’t be less “successful” than she is. He’s bitten off a harder task than she has.