This Just In: Kids Raised on Sesame Street Have No Concept of the Number 21


According to New Scientist, an experiment in Brazil may offer evidence for linguistic determinism—"the controversial hypothesis that the language available to humans defines our thoughts."

Hunter-gatherers from the Pirahã tribe, whose language only contains words for the numbers one and two, were unable to reliably tell the difference between four objects placed in a row and five in the same configuration, revealed the study….

In order to test if this prevented members of the tribe from perceiving higher numbers, Gordon set seven Pirahã a variety of tasks. In the simplest, he sat opposite an individual and laid out a random number of familiar objects, including batteries, sticks and nuts, in a row. The Pirahã were supposed to respond by laying out the same number of objects from their own pile.

For one, two and three objects, members of the tribe consistently matched Gordon's pile correctly. But for four and five and up to ten, they could only match it approximately, deviating more from the correct number as the row got longer….

Gordon says this is the first convincing evidence that a language lacking words for certain concepts could actually prevent speakers of the language from understanding those concepts.

The article acknowledges that there could be other reasons for the Indians' poor results, "including not being used to dealing with large numbers." There is also the possibility, left unexplored, that tribal people simply enjoy screwing with researchers' heads.

NEXT: Another Election Prediction from the Dismal Science

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

  1. “Language isn’t used for thinking. it is used for communication”

    so you think entirely in pictures, unrelated to any words?

    that would make you quite unique.

    i think people think in a combination of concepts, images and language.

  2. “What’s the Arabic word for tolerance?” TasamuH; what’s the English word for ‘gahel’?

  3. mr. hogan-

    “what’s the English word for ‘gahel’?”
    I’m almost certain it is turnip, right?

  4. Dialect error it should be “jahel”. It’s an abstract noun, something we actually have alot of. “Jahel” I mean, not abstract nouns.

  5. Hasn’t everyone had the experience of being unable to find the words that express your thoughts? How is that possibile if language determinism is valid?

  6. Well, I guess I’ll be cancelling my sex tour to Samoa then.

    One would tend to think that the only number they needed for these tests was a 1. If someone puts 10 batteries in a row, you simply proceed down the row, providing 1 object for each opposite object. A 1-to-1 correspondence as we say in the math field.

  7. I’m a believer in a certain amount of linguistic determinism myself, at least at the margin, but it’s clumsy attempts at demonstrating it like this that tend to put the whole idea into general disrepute.

    Incidentally, there was an argument somewhere else recently on the blogs (Samizdata possibly but I really don’t remember now) about colour terminology in different languages and the effect of this on perception. Interesting but, again, nearly impossible to prove.

  8. Given language determinism, just how was language created?

  9. That study was doubleplusungood.

  10. wellfellow, the ability to think and the ability to use language evolved in parallel.

    Language wasn’t “invented,” any more than the opposable thumb.

  11. Joe,

    Perhaps language is like dexterity? But dexterity, for a pianist, might not limit his ability to compose.

  12. Pinker’s “The Language Instinct” covers this debate quite well. One of the more compelling arguments in that book was presented (briefly) above: if concepts are determined by language, then how does the word for a concept come about?
    Pinker debunks several studies, including one similar to this one. Ultimately, it makes no sense to say that thought is entirely in language, or entirely in pictures and concepts-it’s both.

  13. You’ve lost me.

  14. I wonder if you did the same thing with Eskimos and non-native North Americans, you could get similar results trying to identify different types of snow.

  15. dhex,

    There is considerable evidence that nobody “thinks” in words at all.

    Brain imaging has shown that the language areas are the last the activate when a person is solving logical problem. It appears that other part of the brain actually process the concepts then forward the results to the language centers to be communicated to the outside world.

    People with temporary aphasia report that during their aphasia they could not understand their own internal voice yet they could still solve complex problems, even those relating to language, as long as they had an alternate means of expressing the results. Aphasics in general show little general cognitive impairment. This suggest that language is not central to cognition.

    Human languages is not analogous to programming languages where the form and structure of a particular sequences causes the computer to behave in certain ways. The “Strong” approach to artificial intelligence, popular in the 70’s and 80’s demonstrated the fallacy of this conception of language as a tool for actual cognition.

    Humans routinely deal with novel situations for which we have no vocabulary. It happens in creative fields like science, technology and the arts all the time.

  16. Shannon, this debate is getting muddied by inexact terms. “Language” is being used to mean at least two things, 1) the artifact of words and symbols, 2) the ability to used that artifact.

    In a sense, the entire brain operates using a language, with a set group of symbols (tiny currents, cells, spaces arranged in a certain way) being able to express an infinite number of concepts.

  17. shannon – surely we do come across many things for which we have no words, both good and ill.

    language may or may not be central to cognition, but it’s certainly not going to be isolated from it, either, in most cases. it is virtually impossible to read a sign in english without hearing the word in one’s head, i would think.

    i may be biased in this respect, because my own thought processes is extremely language-centered. an inability to remember street maps vs. sequential directions, etc.

  18. Francis Poulenc has a Sonata for One or Two Pianos.

  19. As usual, Blackadder got there first:

    Edmund: Right Baldrick, let’s try again shall we? This is called adding. If I have two beans, and then I add two more beans, what do I have?
    Baldrick: Some beans.
    Edmund: Yes…and no. Let’s try again shall we? I have two beans, then I add two more beans. What does that make?
    Baldrick: A very small casserole.
    Edmund: Baldrick, the ape creatures of the Indus have mastered this. Now try again. One, two, three, four. So how many are there?
    Baldrick: Three
    Edmund: What?
    Baldrick: And that one.
    Edmund: Three and that one. So if I add that one to the three what will I have?
    Baldrick: Oh! Some beans.
    Edmund: Yes. To you Baldrick, the renaissance was just something that happened to other people wasn’t it?

  20. Maybe there was a miscommunication, the anthropologists thinking they were testing counting, the natives creating feng shui.

  21. Or, on the other hand, the natives really may have been trying to mess with the anthropologist’s heads. Wouldn’t you?

  22. “Or, on the other hand, the natives really may have been trying to mess with the anthropologist’s heads. Wouldn’t you?”

    I hear that’s the only real fun to be had in the Amazon these days.

  23. *sigh* OK, as an Officially Trained Anthropologist(TM), I feel compelled to defend my compatriots. First, let’s consider the “pulling our leg” hypothesis: yes, this sort of thing does happen; Napoleon Chagnon reported it happening to him among the Yanomamo, and there are other cases. But it doesn’t tend to go undiscovered for very long. People seem to think that anthopology consists of dropping in and asking a couple questions: anthropologists (good ones) live with their hosts for years at a stretch, often returning many times over decades, talking to everyone they can, and generally trying to get to the point where they think like a native. Of course, there are bad and lazy anthropologists, but the idea that a simple joke would generally go completely undetected is improbable.

    As for whether it’s possible, yes, there are known examples of this kind of thing in other cultures. Exactly what it means is still open to debate. Imagine if an alien anthropologist came to Earth and asked you, “What’s the name for the number represented by a one with seventeen zeros after it?” Would you know? Should the AA conclude you don’t know how to count that high or have a name for that number, and would he be right? You hardly need to know, of course, because you almost never have anything to do with a number that huge. And guess what? That appears to be pretty much the case in many “primitive” cultures: there just isn’t much need for specifically named very large numbers. Check out this example: metasyntactic variables. Notice that most of them only go to about four, and none beyond six!

    Anyway, I’m surprised that no one has mentioned the possibility that the whole story was badly reported. Given that the American press generally can’t report that 2+2=4 without messing it up, I take the whole thing cum grano salis anyway.

  24. Messing with researcher’s heads had long been a hobby of mine, most recently on the anonymous “ethnic self-identification surveys” that are undertaken at the office. On the last one I got all of my coworkers to indicate their ancestry as “Polynesian Islander”.

    I’m sure that the 30 of us in our department had a statistically significant impact on the survey results.

  25. JD: As I indicated above, the test as reported doesn’t appear to be a counting problem. It’s a “can you make a 1-to-1 correspondence” problem.

    A counting problem would be to show them ten batteries and have them go get the appropriate number of flashlights.

  26. “Language wasn’t “invented,” any more than the opposable thumb.”-joe

    The opposable thumb is a biological structure, the only possible inventor of such is God, or a large black monolith, whichever strikes your fancy. Language, however, is a tool, and while no single individual can claim to invented language out of whole cloth, some group of people over time invented the first words and strung them together to make themselves understood to others of their kind. If you want to say the development of language reinforced the development of intelligence, fine, but the same can be said of the stone axe.

  27. Reminds me of Planet of the Apes. You know, where Zera builds the stairs out of various blocks, and then just sits there staring at the banana. The zoologist is puzzled as to why she does not eat it, and says so out loud. Zera responds, speaking in the presence of humans for the first time, “Because I loath bananas!” Maybe the natives just don’t like math.

  28. maybe they happened to study in an American public school…

  29. What is it with you people and public schools? Around here it’s the private schools that people are taking their kids out of, because they are being run to accomodate the lowest common denominator, not the towns’.

  30. I thought even primitive people could count to 10, by using their fingers. Is it just not having a word for “numbers” above “4” that makes the difference? Makes me subscribe to the screwing around with their heads theory. (And if they know what a battery is, how primitive can they be?)

  31. cdunlea-

    Then is looks like the market is working. Can you say as much for public schools? When I was in school everyone received the same crappy instruction. Perhaps things have changed, but I doubt it.

  32. CDunlea- If things are as you describe in your city, then the situation is opposite that found in the majority of cases.
    I live near Kansas City, home of the notorious, and unacredited, Kansas City Missouri School District. I can’t imagine anyone claiming that private schools–or any schools–are inferior to the KCMSDs.
    It is certainly true that there are good public school districts. The problem is that they tend to be in areas where the kids already have every advantage. The worst schools “serve” those who are starting from a disadvantage, and do nothing to negate it.
    That, in short, is the practical objection to public schools. They reinforce, rather than ameliorate, situations that are already desperate.

  33. In The Fateful Hoaxing of Margaret Mead, Derek Freeman conducts a detailed historical analysis of Margaret Mead’s Samoan researches…By examining hitherto unpublished correspondence between Mead, her mentor Boas, and others–as well as the sworn testimony of Fa’apua’a Fa’am, one of Mead’s traveling companions of 1926–Freeman provides compelling evidence that one of the most influential anthropological studies of the twentieth century was unwittingly based on the mischievous joking of the investigator’s informants. The Fateful Hoaxing of Margaret Mead: A Historical Analysis of Her Samoan Research

  34. This study doesn’t show support for “language determinism” instead it shows support for “concept determinism” The Pirah? not only lack language for numbers but also lack the concept of counting in general. A culture must have the concept before they can have the language elements to describe it. The Western world generates hundreds if not thousands of new symbols every day in order to label all the new things and processes we create but the idea always exist a prior to the creation of the symbol.

    Language isn’t used for thinking. it is used for communication. It is a thin gloss on top of the real computation that our brains preform to accomplish even the most trivial task. Language is the last step in thought not the the first.

  35. I wonder how well it’s working for the members of the last class to graduate before the school gets shut down?

  36. I remember Carl Sagan wrote about an anthropological study of some primitive tribe that supposedly had no idea how babies were made; the tribesmen, when asked, mumbled a bunch of mumbo-jumbo about spirits and whatnot. But, as Sagan pointed out, if some weirdly-dressed stranger came to my town and asked me insulting questions about whether I knew where babies came from, I’d be quite tempted to start talking about storks and cabbages. And if that same someone asked me if I knew how to count to ten, I’d make a point of messiing up somewhere around three.

  37. What’s the Arabic word for tolerance?

  38. Rick,
    The word praxis existed before Nimzowitch used it in the title of his book. Chess practice, if you will, improved when Nimzowitch intorduced the words for certain concepts, allowing chess players to think about the concepts.
    If your grasp of English were poor, your ability to follow this argument–namely, your freedom–would be limited. Do you follow that?

  39. Alan:

    “…your ability to follow this argument–namely, your freedom…”

    The cognitive ability to pursue an argument is a very different thing than the freedom (liberty) to do so.

    Now, what is this “broader definition” of freedom that you think libertarians need to “get their minds around”?

  40. Shannon Love at August 23,12:51 PM:

    “Language is the last step in thought not the first.”

    But not in all thought because surely the cited result does give credence to “linguistic determinism”, at least in the numerical realm, because linguistic determinism means that that the language available enables our thoughts. Perhaps just complex thoughts, since there must be some thought before language.

    Also, perhaps language enables some concepts which then enable more language and so on. Are there any other positive experimental results for non-numerical concepts?

    Shannon Love at August 23, 02:43 PM:

    “Humans routinely deal with novel situations for which we have no vocabulary. It happens in creative fields like science, technology and the arts all the time.”

    Hmmm. But, don’t we often think of those situations in terms of the vocabulary of related or somehow similar situations? But on the other hand, maybe the reason why mathematics is necessary to describe some really exotic situations is because words fail us (there aren’t any) and the application of quantitative relationships is universal. I’m thinking that Quantum Mechanics is an example of this.

    joe at August 23, 03:00 PM:

    “Language” is being used to mean at least two things, 1) the artifact of words and symbols, 2) the ability to used that artifact.

    Those two things together seem like a strong definition of “Language”.

    Another thing that fascinated me about this experiment (I was the one who suggested the article as a blog thread) was the fact that the Pirah? are bartering folks, and you would think that the ability discern numbers of things greater then three would be quite useful to them.

  41. CharlesWT:

    Derek Freeman is famous as the person who “revealed” that Margaret Mead had been hoodwinked by Samoans who amused themselves by telling her tall tales about their sexual promiscuity to see if the stupid white woman would believe it. But I distinctly recall reading something to much the same effect a few years earlier in “Tales from the Margaret Mead Taproom,” by Nicholas Von Hoffman (illustrated by Garry Trudeau). Freeman’s book was obviously more of a scholarly work than Von Hoffman’s, but apparently its thesis was no secret to anyone who (like Von Hoffman) hung around in Pago Pago for a while talking with the Samoans.

  42. Having read the article on the New Scientist site, I’m amazed that anyone arrived at any conclusions at all.

    1. In which language did the researchers tell the test subjects what was expected of them?

    10. What was the nature of the reward given for “correct” answers?

    11. Depending on the frame of reference, 6 things CAN equal 3 things. For example,

    2 oranges + 1 peach + 3 tennis balls = 3 apples.


    You can eat oranges and peaches and apples. If you’re hungry, the tennis balls are meaningless.

    100. Mathematics is a language.

    101. Has this experiment ever been done with people from cultures which have any number of numbers in their languages? What was the point at which people started getting the counting wrong?

    110. What does language have to do with it anyway? People don’t need a common language to play checkers.

    111. I wonder if people whose mother tongue is ancient Greek or Sanskrit think about numbers differently from people who have only singular and plural.

    1000. I think I think mostly in words.

  43. Good discussion of this over on (sorry… I’m too stupid to learn HTML and this nasty head cold is not helping).

  44. The explosion of chess praxis after Aaron Nimzowitch enriched chess theory with new words and terms like “passed pawn” is evidence of the link between language and thought. But the words alone are not enough. Only when libertarians get their minds around a broader definition of freedom will they move beyond primitive dogmatism.

  45. Alan wrote: “Chess practice, if you will, improved when Nimzowitch intorduced the words for certain concepts, allowing chess players to think about the concepts.”

    I’d buy it stated as ‘…allowing chess players to TALK about the concepts (with each other.)’

    Language influences communication.

    Then again I don’t think in words – except when/as I’m trying to put thought into words to communicate – so maybe others’d disagree (freaks!)

  46. speedwell.

    By posting the link you have robbed me of hours of precious time.

    (I’ve got my mind around a broader definition of the word “robbed”.)

  47. Alan,

    BTW, Chess Praxis is the name of one of Nimzowitch’s books.

    Just what “broader definition” of freedom do you suggest? It’s coercive collectivism that truly has primitive roots.

  48. …whereas chasing others off your turf is clearly a late blooming, higher-order development, eh Rich?

  49. Alan said, “Only when libertarians get their minds around a broader definition of freedom will they move beyond primitive dogmatism.”

    What, you mean a broader definition like, “freedom is slavery,” perhaps?

  50. >Malak: “Maybe the natives just don’t like math.”

    I bet the natives just prefer to do all their work using Zermelo-Fraenkel notation.

    >Alan: “The explosion of chess praxis after Aaron Nimzowitch enriched chess theory with new words and terms like “passed pawn” is evidence of the link between language and thought.”

    Interesting point. Often a new term (or new notation) directly precedes (and, um, maybe facilitates) an explosion in a field. Perhaps by encapsulating a difficult or novel idea it makes it easier to pick it up, so to speak, and play around with it and extend it by combining it with other ideas. Quantum mechanics got easier and more fecund after bracket notation, f’rinstance.

    You need glass beads to play the glass bead game.

  51. The article on this subject in this week’s The Economist is much clearer and more informative than the articles linked to here.

    What a great magazine, The Economist.

Please to post comments

Comments are closed.