Till death do us part
As regular readers of the Bremer Sprachblog know, the languages of the world are disappearing at an alarming rate (see for example here, here, here, here, here, and here). According to the most conservative estimates, at least half of the 6,500 languages currently spoken will become extinct by the end of the century (by the way, if you’re wondering why I am addressing you in English today, please bear with me — I have a point to make).
When linguists draw attention to this mass extinction, they naturally portray it as something bad. This negative evaluation seems so natural to us, that we are often surprised when others disagree.
Last week, a story from the forthcoming issue of National Geographic on the topic of language death was taken up in the American press, for example in the New York Times, the Los Angeles Times and the Washington Post. While the specifics of that story have not met with the wholehearted approval of all linguists, in the end we probably all agree that there is no such thing as bad publicity when it comes to raising public awareness of language death.
Then again, it is not as simple as that. Because while the story raised public awareness to some extent, not all members of the public were automatically sympathetic to the cause of saving languages from extinction. For example, the science journalist and blogger Razib Khan took up the story in his blog Gene Expression. He acknowledges that when a language is lost, so is part of the history of the people who spoke it. But, he argues, in the end “the fundamental truths that we value as human beings is [sic] extant in stories which transcend the particular speech in which they are transmitted”. In other words, language loss may not be desirable, but it is nothing to lose sleep over.
His post is fairly moderate and carefully argued and so his position may seem quite plausible. However, it is based on at least two implicit assumptions: first, that language is a neutral carrier of information, and second, that language has no value beyond the information that it is used to express.
Khan and two other commenters then come up with a theory that is supposed to explain why languages become extinct. Khan himself claims that “most local languages die because people choose no longer to speak them (his emphasis). Commenter pconroy agrees and says that “[f]or most people the smart thing to do is to learn the language of the elite in their region or country - especially when one is impoverished or fighting to survive at all”. And jim goes one step further and hails language loss as a mark of progress.
The extinction of languages is a good thing. Becoming fluent in one of the dominant languages opens up the world of knowledge, making people more efficient and productive. You can’t become a (real) doctor or engineer knowing only Navajo, for example.
Leaving aside the question, why one cannot become a doctor or engineer as a monoligual speaker of Navajo, and leaving aside the question whether giving up a language is ever a “choice” that people make, all three statements make sense as theories of language death only if we make an additional assumption: that people can only speak one language, and therefore must abandon their original language if they want to learn English (or some other “dominant language”).
The last several hundred years have seen a great uptick in language extinctions. At the same time mankind’s material condition has improved immensely. The fact that trade and specialization is harder in a Tower of Babel world divided up into tiny language communities is a plausible factor. The regions of the world with the greatest linguistic diversity (say, Papua New Guinea) tend to also be the poorest.
Apart from the obvious fact that the people whose material condition has improved are not the same as the people who were forced to give up their languages (look at the Australian Aborigines if you want a text-book example), there is another assumption underlying this statement: that speaking a single language guarantees successful communication.
Let us briefly look at each of these four assumptions in turn. We will find that none of them holds water.
1. A language is a neutral carrier of information
People often assume that a communicative act begins with a particular piece of information that exists in some speaker’s head, independently of language. The speaker then takes that piece of information, encodes it in language, and communicates it to another speaker. The other speaker hears or reads the language, decodes it, and ends up with the original, language-independent piece of information.
This model is overly simplistic in many ways, of which I will only discuss one: languages are structured in ways that encourage speakers to pay attention to particular aspects of a particular piece of information. As an example, take an event where something moves to some location in some way. The act of moving is naturally encoded by a verb. Interestingly, however, languages differ in the kind of motion verbs they have.
In the Germanic languages (like German or Dutch) and the Slavic languages (like Russian or Polish), most motion verbs encode a particular way of moving (take German laufen “walk”, fahren “drive”, schwimmen “swim”, rennen “run”, kriechen “crawl”, etc.). In contrast, in the Romance languages (like French or Spanish), most motion verbs encode a particular direction of motion (take Spanish entrar “enter”, salir “exit”, subir “go up”, bajar “go down”, etc.).
This means that when speakers of a Germanic language encode a motion event, they must pay attention to the manner of motion, since otherwise they would not know which verb to choose. The direction, in contrast, is encoded in a prepositional phrase (like in, into the room, or to Bremen). This prepositional phrase is often optional (it can be left out), which means that speakers can, but do not have to, pay attention to the direction of motion. For speakers of a Romance language, the situation is reversed. Since Romance motion verbs encode directionality, a speaker of a Romance language must pay attention to direction in order to choose a verb. The manner, in contrast, is encoded by a participle, which, like the prepositional phrases just mentioned, can be left out. Thus, speakers of a Romance language can, but do not have to, pay attention to manner of motion.
And indeed, the psycholinguist Dan Slobin from the University of California at Berkeley and his colleagues and collaborators have shown that speakers of Romance languages consider it more important to talk about direction of motion while speakers of Germanic languages consider it more important to talk about manner of motion. For example, even when they are translating from a Germanic language, Spanish translators often leave out manner information and add information about directionality, and German translators leave out directions and add manner information when translating from a Romance language. Clearly, then, language is not a neutral carrier of information.
Some linguists and anthropologist are willing to go one step further: they argue that such linguistic differences can influence not just what speakers consider more important in communication, but also what speakers consider more important in perception. In other words, the fact that speakers of Romance languages must pay attention to the direction, but not to the manner in which something moves when they want to talk about it may shape their brain in such a way that they are more likely to pay attention to direction and disregard manner even when they are simply watching a motion event with no intention of talking about it (this is referred to as “linguistic relativity”, associated with the early 20th century linguist Benjamin Lee Whorf and more recently with the developmental psychologist John A. Lucy from the University of Chicago). If this turns out to be true (and there is some evidence that suggests that it is, at least in some areas of thought), then this would of course be an even stronger argument against the idea that languages are neutral carriers of information.
Let us turn to the second assumption.
2. A language has no value beyond the information that it is used to express
Obviously, the speakers of a language are typically emotionally attached to that language. On this blog, we often ridicule German language purists who take the existence of a few English loanwords as evidence for the impending extinction of German. This ridicule is well deserved and we will continue to point out that (a) many of these people have a hidden agenda that has nothing to do with language but everything to do with isolationism or even nationalism and that (b) German is alive and kicking and could easily swallow ten times the current number of English loanwords without choking on them. However, the very irrationality of their fear of extincton is evidence for their strong emotional attachment to the German language.
But this emotional attachment, important as it is to the speakers, is not even the main reason why scientists value language. Their attempts to maintain linguistic diversity is motivated by the fact that each language is a unique source of information.
First, the vocabulary of a language typically encodes a complex taxonomy of objects in the world. For example, indigenous peoples have a very exhaustive knowledge of plant and animal species in their natural environment — knowledge that may not precisely match a biologist’s taxonomy of these species but that is optimally geared towards survival in that environment and that would take decades to replicate independently. Anthropologists are interested in these taxonomies because of the cultural assumptions that are embedded in them. Evolutionary psychologists are interested in these taxonomies because they provide evidence of general, species-specific psychological principles that operate in our perception of the world around us. And even biologists are often interested in these taxonomies because they provide them with a massive headstart when cataloguing animals or plants in a particular part of the world.
Second, as was pointed out above, many anthropologists and linguists now believe that the structure of a language can have an influence on the way in which its speakers perceive the world. But language structure clearly does not determine our perception of the world, and if we want to understand how the human mind works, we must find out which aspects of language structure influence our perception and which don’t and to what extent such an influence is possible. The more languages become extinct, the fewer potential data points we have for studying these questions and the more incomplete the final picture will be.
Third, linguistic diversity provides important data even for linguists who deny that language structure has any influence whatsoever on our perception of the world. There is good evidence by now that there are linguistic universals — properties that are necessarily shared by all human languages and that must therefore reflect general psychological or even neurological principles. In order to uncover these universals, we need as much linguistic diversity as possible, because the fewer languages we can look at, the greater the danger that we will mistakenly classify some property as a universal. For example, if the only languages left were Chinese, English, Spanish and Russian, we might be tricked into believing that all languages must have Subject-Verb-Object as their basic word order. But in fact, all six possible orders are found among the languages of the world and Subject-Verb-Object is not even the most frequent one (it is found in 41 percent of the world’s languages while Subject-Object-Verb is found in 47 percent).
One could argue, of course, that we have no right to expect people to speak a language that hinders their economic success simply because anthropologists, linguists and psychologists need these languages as data. And of course one would be right in arguing so. However, no scientist wants to stop anyone from learning a new language that opens up new possibilities. This brings us to the next assumption.
3. People can only speak one language, and therefore they must abandon their original language if they want to learn an economically dominant language
If this assumption were true, this post could not exist. The fact that I was able to write Monday’s post in German and that I will write Sunday’s post in German, but that I wrote today’s post in English, and the fact that all regular readers of this blog are able to understand all of these posts, shows that people can quite easily communicate in one language without giving up the other (in fact, that was the entire point of writing this in English). Now, you might argue that this is because we are all highly educated people. But in fact, the majority of the world’s population is multilingual. In many regions of the world, people could not even communicate with their neigbhors in the next village if they were not multilingual, and this has nothing to do with education. Children who grow up in contact with several languages will learn all of these languages as easily as they would learn a single language.
4. Speaking a single language guarantees successful communication
This assumption is so obviously false that there is litte to say about it. I will merely repeat, in slightly more detail, something that commenter John Emerson already pointed out in the Gene Expression thread: there is no correlation at all between speaking a common language and communicating successfully. Protestants and Catholics in Northern Ireland have both spoken English throughout the thirty-year conflict that cost more than 3,500 lives. Serbs and Croats fought viciously against each other in the last decade of the 20th century despite the fact that they speak closely related dialects of the same language. Switzerland, on the other hand, has four official languages but has not seen a civil war since 1848 (and that war — surprise, surprise — was about religion, not about language). The European nations have fought against each other throughout their entire recorded history, and they were finally able to break this destructive pattern not by adopting a common language but by creating what is, in essence, a peaceful and wealthy, but stubbornly multilingual super-state based on common values — a super-state whose governing institutions spend up to a third of their budget on translation services.
In sum, then, each language is a carrier of a cultural identity, culture-specific knowledge, and a particular way of paying attention to the world around us. It may even turn out to be a filter through which we perceive the world in the first place. And it is evidence for the possible forms that human languages are made of. As such, each language is a fundamental part of its speakers’ identities as well as an irreplaceable piece of data for anthropologists, psychologists, linguists, and many other scientists interested in language, culture and cognition.
Since people can learn more than one language, speakers of small languages need not be in any way disadvantaged as compared to speakers of one of the dominant world languages. Therefore, nothing can possibly be gained by encouraging, enticing or forcing speakers to give up their language in favor of English, Russian, Spanish, Arabic or Mandarin Chinese. But everything can be lost.
[Note: I will add some references to this post over the next few days, but for now I just want to get it out there].