Authors: Michael Erard
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 language 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.
Certainly, Cook wrote to me in an email, humans tend to leave their multicompetence unexplored. “I think that the human ability to
learn
languages in natural environments is mostly untapped,” he wrote. But what are the upper limits of multicompetence, however you want to define that? He was unable to say.
Without being able to observe Mezzofanti directly, it’s hard to nail down the scope or depth of his multicompetence. You can, if you indulge in a bit of anachronism, imagine what he (or anyone) would have to do to gain
the same accolades now. For instance, according to the current European Commission standard for language proficiency, a person at the highest rating—someone we would say had “mastered” a language—has to “have a good command of idiomatic expressions and colloquialisms,” “convey finer shades of meaning precisely,” and “backtrack and restructure around a difficulty so smoothly the interlocutor is hardly
aware of it.” I can’t say in how many languages Mezzofanti really had such abilities.
To be sure, there’s a big difference between a truly multicompetent person and someone with lots of “bits of language,” who would have a hard time succeeding, for instance, at a set of language competitions that the German government has held since 1985 for students between seventeen and nineteen years old.
They undergo a yearlong battery of tests, including discussing a cartoon and a text; submitting a written exam that involves translating, writing, and summarizing; writing a 3,000-word essay; and participating in oral exams as well as a multilingual debating session. Faking abilities, much less in the required minimum of two languages (they can work in as many as four), would be impossible.
Several years ago, I interviewed a member of the US intelligence community who went to job fairs recruiting foreign-language experts and, every so often, met a person who claimed to speak forty languages. Inevitably, the claim was accompanied with the word “fluently.”
He greeted these with exasperated disbelief. I’m a non-native learner of six languages, he told me. I know what I need to do to
learn a language, I know what I need to do to get good at it. Not only are the claims implausible, they’re unsupported. No one who says he or she has forty languages has ever tested out in all of them.
He then described to me what he hired people to do. They’d have to know how to distinguish a prayer from a coded transmission. How to
parse the speech of a non-native speaker, mispronunciations,
errors, and all—someone who may or may not be nervous, who may be speaking in a dialect, who might be using a cell phone connection with lots of static and environmental noise. You don’t get skills like that by watching a few movies. It takes hundreds of hours of practice. It takes firsthand experience in the culture. Given the stakes, analysts should have them.
Could anyone do this in thirty
languages? Probably not. But one’s ultra competence in one language could certainly be informed by a sizable multicompetence in dozens. Indeed, my contact said he works with people who are very good in ten to fifteen languages, people who say, I’ve got to go learn Georgian, and off they go, and then return and say, I’m going to Estonian school, all while they’re studying Turkish at home. In multilingual
environments like this, amazing feats of language prowess are everyday affairs. He told me about a former colleague who wanted to learn languages well enough to quote proverbs in them. It became a game between him and his hosts to sit and quote obscure proverbs to each other. “This guy was
untouchable
as far as languages go,” my contact marveled.
Even for a person with a good working knowledge
of a language that falls short of nativeness, not all professional avenues are closed. Mezzofanti might fare better by the standards of the aviation industry and its expectations for English proficiency among pilots and air traffic controllers. In 2008, the International Civil Aviation Organization (or ICAO, a United Nations–mandated agency) introduced requirements that they be able to speak and
understand English to a certain level of proficiency by March 2011. The goal was to make the Babel of the skies clearer and safer.
Implementing the standards butted against some linguistic realities, though. One was the diversity of accented Englishes that pilots and controllers encounter every single day. In one nine-hour observation, Turkish air traffic controllers interacted with 160 pilots
from Turkish airlines, 14 from German airlines, and 104 pilots from airlines from 26 other countries—all speaking English. Yet only 2 of the 104 pilots worked for airlines based in an English-speaking country and presumably were native English speakers.
Rather than focus on producing native-like English speakers, ICAO focused on flying and landing airplanes safely. You didn’t need to discourse
about squids or default swaps or the existence of God, as you might have to do in other tests of language proficiency. Instrumentation and weather terminology are more relevant topics. You don’t need to speak English like an American or a Brit; you have to be intelligible. This is no easy feat, but it’s one that, unlike native pronunciation, is achievable by adults. And you don’t have to be error-free.
After all, even native speakers make errors.