The Limits of Language Are the Limits of Our World

Language and Thought.pdf

Download PDF • 2.21MB

My name is Artem. Today I want to talk about a topic that I find both fascinating and deeply personal. The title of this lecture is “The Limits of Language Are the Limits of Our World.” More broadly, I want to talk about the hypothesis of linguistic relativity—what it is, how to understand it, and why it matters.

The idea for this presentation came up when Roma once talked about doing a presentation on the alphabet. That discussion made me think: instead of letters, why not talk about languages themselves? So Roma, thank you for that idea, and thank you for taking part in this presentation.

I want to begin by explaining why I personally wanted to talk about this. Like many people here, I’ve lived through several languages, and at some point language stops being just a tool and starts feeling like something that shapes how you think. I was born in the Soviet Union, in the Ukrainian SSR, in the city of Odessa. The first language I spoke was Russian. When I was still very young, my family moved to Moldova. I studied in a Russian school there, but at the same time I learned some Moldovan, simply because it was part of the environment.

When I was ten years old, we moved to Ukraine. Again, I studied in a Russian-speaking school, but Ukrainian became an important subject, because Ukraine was already independent at that point. Later, when I was sixteen, we moved to Israel. There I did not attend a Russian school. I studied Hebrew, and I also studied English—because although English formally existed as a subject in my Ukrainian school, in practice it was barely taught.

Then, when I was about twenty-two, my wife and I decided to move to Canada. That involved interviews and processes in French, so I was learning French as well. There’s a long story in the middle about documents, getting married, and having a child—but that’s not important here. What matters is that languages were never abstract for me. They were part of my everyday life.

At the same time, I was also learning programming languages. At university, I studied communications—things like data transmission protocols. Later, already in Canada, I worked in a startup, and we were dealing with early natural language processing. At that time, there were basically two approaches. One was very simple statistical text processing—trying to make computers guess what a text is about. That approach was crude and often worked very poorly. The second approach was more linguistic: breaking sentences down into subjects and predicates, identifying who is the subject, who is the object, what adjectives are used, and so on. And it was during that time that I really started thinking seriously about language itself—what language is, how it works, and why it matters.

To understand why this question is important, let’s take a short look at history. Humans—Homo sapiens—have existed for a very long time. In the lecture I originally mentioned a figure around two hundred thousand years, but to be precise, the oldest securely dated Homo sapiens fossils are closer to three hundred thousand years old. What matters here is that spoken language is extremely old, but writing is not.

Writing—the thing I like to call “external memory”—appeared only about five thousand years ago. That’s when people began recording ideas, when writing systems like cuneiform appeared in Mesopotamia. In a sense, humans started storing information on external media—literally on hard drives made of clay tablets. And without spoken language and writing, we as a civilization would almost certainly not exist.

Today, linguists estimate that there are roughly seven thousand languages on Earth. Most of us, if asked, would probably think of twenty or thirty languages at most. It’s hard to imagine that there are thousands. And yet there are. One of the most striking examples is Papua New Guinea. On a territory roughly comparable in size to California, people speak over eight hundred languages. That means that more than ten percent of the world’s languages exist in one relatively small region.

At the same time, languages are disappearing. On average, about one language disappears every two weeks. This happens because the last speakers die, or because the language is no longer passed on to children. These are often languages that exist only in spoken form, with no written or digital record. When the last speaker dies, the language disappears forever. We cannot restore it. And this matters because language is not just vocabulary—it is a kind of mental firmware. It shapes how we think and how we organize the world.

This brings us to the main idea of today’s talk: the hypothesis of linguistic relativity. This idea was proposed in the early twentieth century, around the 1930s, by linguists associated with Edward Sapir and Benjamin Lee Whorf. Sometimes it’s called the Sapir–Whorf hypothesis. The idea, in simple terms, is that language affects how we see and think about the world.

There are two main interpretations of this hypothesis. The strong interpretation says that language determines how we think—that language completely dictates our perception of reality. The weak interpretation says that language influences how we think—that it shapes habits of attention, perception, and memory. Today, most scientists reject the strong version but broadly accept the weak version. For a long time, especially toward the end of the twentieth century, there were intense debates about this. Some scholars, like Noam Chomsky, argued strongly against these ideas. But in the last few decades, many experiments have provided evidence that language does, in fact, influence cognition in measurable ways.

Let me give some examples.

Let’s start with color. What color is this? And what about this one? Most of you would say “blue.” In Russian, however, there are two basic color categories where English has one: siniy (dark blue) and goluboy (light blue). In English, these are just shades of “blue,” but in Russian they are treated as separate basic colors. If you ask a Russian-speaking five-year-old child, they will easily tell you which one is siniy and which one is goluboy.

There was an experiment where Russian-speaking and English-speaking participants were shown many colored squares and asked to quickly identify which colors matched. Russian speakers were about ten percent faster when the colors crossed the siniy–goluboy boundary. The original lecture phrased this as “different parts of the brain were involved.” To be precise, what was demonstrated is that language categories affect processing speed and that verbal processing plays a role in this effect. The important point is that language can influence perception—not by changing what we see biologically, but by changing how quickly and efficiently we process what we see.

Similar effects appear in other languages that divide the color spectrum differently—for example, languages where orange and yellow are treated as one category, or where green and blue are grouped together. Again, this is not about naming colors. People were not asked to name colors. They were asked to match and discriminate them. And language influenced how quickly they could do that.

Another example comes from grammatical gender. In many languages, including Russian, German, Hebrew, and Spanish, inanimate objects are assigned grammatical gender. A table might be masculine, a plate feminine, even though there is no biological reason for this. Researchers conducted experiments comparing German and Spanish speakers. In German, the word “bridge” is feminine; in Spanish, it is masculine. When participants were asked to describe a bridge, German speakers tended to use adjectives like “elegant” or “beautiful,” while Spanish speakers tended to use adjectives like “strong” or “long.” When the same experiment was done with the word “key,” whose grammatical gender is reversed in the two languages, the pattern reversed as well.

This does not mean people literally believe objects have gender. It means that grammatical structure can subtly influence the associations and metaphors that feel natural. I can say this from personal experience as well. When I first learned Hebrew, many objects had a different grammatical gender than in Russian. I often made mistakes because my brain automatically transferred the gender from Russian into Hebrew. In a sense, I was thinking incorrectly because of the linguistic habits I had already formed.

Another important area is working memory. Our short-term verbal memory can hold information for roughly two seconds. This means that the length of words matters. In English and Russian, digit words like “seven” or “восемь” are relatively long. In many Asian languages—Chinese, Japanese, Korean—digit words are much shorter. Because of this, speakers of those languages can often remember and reproduce two or three more digits in memory tests.

In the lecture I mentioned that this might be one small reason why Asian students often perform well in mathematics. I also explicitly said that this is not the main reason. Education systems and training are far more important. Shorter digit words simply provide a small advantage in certain memory tasks.

Now imagine something even more extreme. Imagine a language that has no numbers at all. In the Amazon rainforest, there is a tribe called the Pirahã. In their language, there are no words for one, two, three, and so on. They have only approximate terms like “a little,” “more,” and “many.” Researchers have tried for decades to teach them exact numbers, without success. When you place five objects on a table, remove them, and ask them to reproduce the same quantity, they cannot do it. The idea here is not that they are incapable of thought, but that without numerical concepts encoded in language, certain kinds of exact reproduction become impossible.

Let me give another example. Everyone, please close your eyes for a second. Now point to where you think north is. You see how uncertain most of us are. In Australia, there are languages in which people do not use “left” and “right” at all. They use only cardinal directions—north, south, east, west. Even five-year-old children in those communities can accurately point north at any time. Their language forces them to constantly track orientation. They don’t say “there’s a spider on your left leg.” They say “there’s a spider on your northwest leg.”

We often joke about “topographic cretinism”—being bad at directions. In those languages, that concept doesn’t exist. People develop an internal compass because their language requires it.

This also affects how people think about time. If you give most of us pictures showing an egg, a chick, and a chicken, we will arrange them from left to right. Hebrew speakers often arrange them from right to left. But speakers of those Australian languages arrange them based on cardinal directions—from east to west, following the path of the sun. If they are facing a different direction, the order changes accordingly. For them, time and space are tightly linked through coordinates.

Another example involves causality. In English, we often say “He broke the vase,” even if it was an accident. In Spanish, it is common to say something equivalent to “The vase broke,” without mentioning the agent. Experiments show that speakers of different languages remember different aspects of such events. English speakers are more likely to remember who caused the accident, while Spanish speakers are more likely to remember the event itself. Again, language shapes what we pay attention to and what we remember.

At this point, people often ask about bilingualism. I was asked whether language changed me or whether I changed because of language. My answer is that bilingualism clearly matters. I think in two languages—Russian and English. When I work, I think in English. When I deal with everyday life, I often think in Russian. When I argue with my wife, I sometimes switch to Hebrew. Each language activates a different mental mode. This doesn’t mean I become a different person. It means that language context influences how thoughts come.

Research on bilingualism shows that speaking multiple languages affects brain development. Early childhood is especially important, because neural pathways are built more easily then, but learning languages later still has significant effects. Studies show that bilinguals tend to have denser neural connections in language-related areas. This does not prevent aging or dementia, but it appears to delay symptoms by several years on average—around four to five years according to many studies.

So what can we do with all this? The answer is simple: use languages actively. Speak them. Read in them. Watch films. Listen to music. Don’t let them fade. I understand Ukrainian very well. I can read and watch movies. But speaking it is extremely hard for me, because I left thirty years ago and never used it actively. Understanding is passive. Speaking requires building and maintaining active neural connections.

For children, this is even more important. Speak to them in your native language. Speak to them in multiple languages. Read books to them. This is not about forcing something artificial—it’s about giving them tools that expand their mental world.

I want to end with the main idea. The limits of language are the limits of our world. Language does not determine us completely. This is not a one-way street. Thought influences language, and language influences thought. They support each other. By learning and using languages, we expand our horizons and our ways of understanding the world.