Luke’s Languages: Between plausible and pseudoscientific


Oh, hello there! Hope you had a relaxing week. (What? It was even more packed than a regular school week? Tell me about it…)

Those of us with lives outside this campus might happen to read the BBC website. Okay, probably not. But if you do, you’ll have seen an article in the business section last Friday called “Why speaking English can make you poor when you retire”.

The study was carried out by Dr. Keith Chen, a professor at Yale. He separated a number of languages into two classes according to whether they have weak or strong “future-time reference”. That basically means the degree to which a language explicitly refers to the future. According to Chen, English has strong future-time reference, because we say “I will, I am going to, I have to” for something happening tomorrow, but Mandarin has weak future-time reference, because verbs have no tense and the time is instead inferred from the context or the addition of a word like “tomorrow”.

Chen then took some data from the lives of speakers of languages in both classes, and apparently found that speakers of languages with strong future-time reference make poorer choices when planning for their future. They “save less, smoke more, and exercise less”, reports the BBC article.

Why would this be? Well, “The act of saving is fundamentally about understanding that your future self—the person you’re saving for—is in some sense equivalent to your present self,” says Chen. “If your language separates the future and the present in its grammar, that seems to lead you to slightly disassociate the future from the present every time you speak.”

The article is in a reputable paper, it uses words like “future-time reference”, and it cites a Yale professor. So it’s got to be right, right?

I’m inclined to believe otherwise.

It comes down to a popular but often overestimated linguistic theory called the Sapir-Whorf hypothesis, or linguistic relativity. It’s a huge topic that can’t really be accurately summarized, but it essentially says that your language determines your cognition. It’s had that name for about half a century and been explicitly expressed for a century and a half, but the truth is, there have been theories about the relation between the language you speak and the way you think for as long as people have been aware of language, and those theories always been just teetering between plausible and pseudoscientific.

One of the most popular ideas is that some languages are inherently logical. Beginning in the Renaissance and continuing to the present day, argument after argument has been laid out for why certain languages adhere to the universal laws of reason. But, as you might already be guessing, apparently logic has a strong tendency to belong to the language of the person doing the analyzing. In 1784 one Antoine de Rivarol wrote, “What distinguishes [French] from ancient and modern languages is the order of the construction of the sentence. This order must always be directly and necessarily clear. French first names the subject of the discourse, then the verb, which is the action, and finally the object of the action: this is the logic natural to all men. […] What is unclear is not French; what is unclear is on the contrary English, Italian, Greek, or Latin.” He was far from alone in this belief.

But two questions immediately arise. One, how does subject-verb-object adhere to natural logic better than the different orders found in other languages? No reason is given. Two, if this is true then why have English, Italian, Greek, and Latin writers so often said the same things about their own languages?

The hypothesis can be found absolutely everywhere. My uncle, a computational linguist who has spent much of his life in India and Nepal, said someone told him that “Hindi speakers are less materialistic than others, because they don’t have a verb for ‘to have’ or ‘to possess’. ” But, he explained, in Hindi “I have a car” is expressed “My car is”; the possessive pronoun “my” indicates who it belongs to. They do have a possessive, even if it’s not formed like English’s—and, he added, tongue in cheek, “they have more lawsuits over property rights than you can shake a stick at!”

The Sapir-Whorf hypothesis is also applied in retrospect. In his book The Language and Imagery of the Bible, G.B. Caird mentions a traditional school that believed the mind of the Hebrew speaker was primitive. He cited a scholar who said that whereas English has “go up”, “go down”, “go out”, “go in”, and so on, all modifying the common element, Hebrew has a separate word for each. The scholar thought this was evidence that the Hebrew speaker could not abstract the notion of pure movement from all of these directional movements. Sounds plausible, no?

But Caird pointed out that in the first place, other supposedly sophisticated languages also work this way, including none other than French with monter, descendre, sortir, and entrer. And in the second place, Hebrew has huge ranges of words where the abstraction is transparent. For example, the same root is used for “learn” and “teach”, “die” and “kill”, “eat” and “feed”; you just have a single marker that means “cause to”. English, on the other hand, has separate words for those things. But does that mean we can’t abstract? Is our cognition too primitive to consider learning and teaching related processes? Not at all!

This is not to throw out the Sapir-Whorf hypothesis; it may not be totally missing the mark. Linguistic habits presumably have some an influence of some degree, however small, on cognition. There are some fascinating phenomena that suggest that speakers of different languages may conceptualize their worlds differently. To take a popular example, many languages differ in what they consider the basic colours and what they consider shades of those colours. As I saw on a forum once, an English speaker who hears the claim “crimson, maroon, and pink are shades of red” might reply, “Hey, pink is not red, it’s pink!”—and yet speakers of other languages do consider pink a shade of red, and the reason, say the proponents of linguistic relativity, is because of how the languages name and recognize the colours.

Here’s the key: languages encode things differently—some through vocabulary, some through grammar. And what’s not explicit is often inferred from context or understood just under the surface. Even in the colour example, it’s a simple step for speakers of either type of language to see it the other way. There may indeed be true cognitive differences that correlate with language, but there’s not a lot of hard evidence for them. In practice the conclusions have often been a stretch, and more ethnocentric than anything.

In any case, if you are going to find sound conclusions on the topic, you won’t find them in the business section. Going back to Chen, his explanation is iffy, his results seem highly susceptible to other factors, and worse, I’m not sure I even agree with his assumptions about how to classify languages. Is English really limited to using tenses the way he thinks? Or isn’t it perfectly natural to say “It’s raining tomorrow; we can’t go to the zoo”, without a future tense in sight? Chen thinks I’ll only say that once I cease to dissociate the future from the present. Me, I call shenanigans.