Does Language Shape What We Think?

A new study looks at what happens when a language doesn't have words for numbers














Share on Tumblr



Image: Vladimir Vladimirov

  • The Wisdom of Psychopaths

    In this engrossing journey into the lives of psychopaths and their infamously crafty behaviors, the renowned psychologist Kevin Dutton reveals that there is a...

    Read More »

My seventh-grade English teacher exhorted us to study vocabulary with the following: "We think in words. The more words you know, the more thoughts you can have." This compound notion that language allows you to have ideas otherwise un-haveable, and that by extension people who own different words live in different conceptual worlds -- called "Whorfianism" after its academic evangelist, Benjamin Lee Whorf -- is so pervasive in modern thought as to be unremarkable.

Eskimos, as is commonly reported, have myriads of words for snow, affecting how they perceive frozen percipitation. A popular book on English notes that, unlike English, "French and German can distinguish between knowledge that results from recognition ... and knowledge that results from understanding." Politicians try to win the rhetorical battle ("pro-life" vs. "anti-abortion"; "estate tax" vs. "death tax") in order to gain the political advantage.

For all its social success, Whorfianism has fared less well scientifically. Careful consideration of the examples above shows why. Try calling dry snow "dax" and wet snow "blicket," and see if you notice a change in how you think about snow. I didn't. The English book's statement assumes that if you don't have a word for something, you can't talk about it ... a claim that the sentence proves false. Finally, calling the law of October 26, 2001 the "USA Patriot Act" may have done as much to stain the word "patriot" as increase enthusiasm for the law.

Oh, and Eskimos don't have all that many words for snow.

In fact, scientists have had so much difficulty demonstrating that language affects thought that in 1994 renown psychologist Steven Pinker called Whorfianism dead. Since then, Whorfianism has undergone a small resurgence. For instance, Lera Boroditsky and colleagues found that speakers of Russian, which treats light blue and dark blue as primary colors, are faster to categorize shades of blue.

While fascinating and important work, these and other similar results are a bit short of showing that "the more words you know, the more thoughts you can have." The recent study that comes closest is an investigation of number.

Although number words and counting are a fixture of life in most cultures from the time we are old enough to play hide-and-go-seek, some languages have only a handful of number words. In a paper published in 2008, MIT cognitive neuroscientist Michael Frank and colleagues demonstrated that Pirahã, a language spoken by a small Amazonian community, has no number words at all. The research team simply asked Pirahã speakers to count different numbers of batteries, nuts and other common objects. Rather than having a word consistently used to describe "one X" a different word for "two Xs" and yet another word for "three Xs," the Pirahã used hói to describe a small number of objects, hoí to describe a slightly larger number, and baágiso for an even larger number. Basically, these words mean "around one," "some" and "many."

The lack of number words had a profound and surprising effect on what the Pirahã could do. In a series of experiments, the researchers presented Pirahã participants with some number of spools of thread. The participants' task was simply to give the researcher the same number of balloons. If the participants were allowed to line up the balloons next to the spools of thread one-by-one, they did fine. But if they weren't allowed this crutch -- for instance, if the spools of thread were dropped into a bucket one at a time, and then the participant had to produce the same number of balloons -- they failed. Although they were generally able to stay in the ballpark -- if a lot of spools went into the bucket, they produced a lot of balloons; a small number of spools, a small number of balloons -- their responses were basically educated guesses.

Could it be that the Pirahã not understand the concept of "same amount"? That's unlikely. When allowed to match the balloons to spools one-by-one, they succeeded in the task. Instead, it seems that they failed to give the same number of balloons only when they had to rely on memory.

This actually makes a lot of sense. Try to imagine exactly seventeen balloons in your head, but without counting them. It's impossible. Decades of research have shown that people can tell the difference between one object and two or between three objects and four without counting, but such fine distinctions with larger numbers like seventeen versus eighteen requires counting. You wouldn't match seventeen balloons to seventeen spools by sight alone. You would count the spools and then count out the same number of balloons.

But the Pirahã can't count. They don't have number words.

This suggests a different way of thinking about the influence of language on thought: words are very handy mnemonics. We may not be able to remember what seventeen spools looks like, but we can remember the word seventeen. In his landmark The Language of Thought, philosopher Jerry Fodor argued that many words work like acronyms. French students use the acronym bans to remember which adjectives go before nouns ("Beauty, Age, Number, Goodneess, and Size"). Similarly, sometimes its easier to remember a word (calculus, Estonia) than what the word stands for. We use the word, knowing that should it becomes necessary, we can search through our minds -- or an encyclopedia -- and pull up the relevant information (how to calculate an integral; Estonia's population, capital and location on a map). Numbers, it seems, work the same way.

I don't know whether my seventh-grade English teacher would be disappointed. Do more words mean more thoughts? Probably not. But more words do make it easier to remember those thoughts -- and sometimes that's just as important.


ABOUT THE AUTHOR(S)

Joshua Hartshorne is a PhD student in the psychology of language at Harvard University. Participate in his experiments at coglanglab.org.


46 Comments

Add Comment
View
  1. 1. AndrewJayPollack 02:02 PM 8/18/09

    I think this is missing the point of how language impacts thought because it's too specific.

    Language and culture impact the way things are considered. English is very unstructured. We make up new words constantly, change the meanings of words as we go, and really do what ever we want to the language to get across an idea. English speaking nations are also generally good at coming up with new things, re-inventing technologies and ideas. We're not terribly good when it comes to the detail of those ideas.

    German is a very structured language. It's quite flexible, but the way new things are done is methodical. Words get longer and more specific and refined meanings are tacked on the end. That same structure is evident in the way Germans approach problems and solutions. There's an essential desire to be specific in a result. Most of the German engineers I know would rather not do something then do it in an inelegant way -- where the American engineers I know will solve a problem any way it can be solved, and worry about refining it later.

    The Japanese language is highly structured and very finely detailed. The smallest nuance can have a big impact on meaning in spoken, written, or gestural language. It is not, however, very good at re-inventing itself for new ideas or concepts. In fact, most technology words end up being from other languages even if they come from Japan. This shows up in their engineering as well. They're amazing at taking a new concept and refining it, detailing it, and pulling out its essentials to put in an extremely efficient net package. I believe the 'exactness' of their language helps them communicate in a way that complements these traits. There's no vagueness at all.

    Obviously these are archetypal social traits and examples will be freely found that don't match -- but there is a reason archetypes ring true on a broad basis.

    One question I have though, is which comes first? Does language impact culture as I've described, or did language differences as described result from those cultural traits?

    Reply | Report Abuse | Link to this
  2. 2. HFB 02:05 PM 8/18/09

    <<A popular book on English notes that, unlike English, "French and German can distinguish between knowledge that results from recognition ... and knowledge that results from understanding." >> And the Turkish language distinguishes between knowledge acquired first-hand, and knowledge acquired second hand (reported, or hearsay).

    Reply | Report Abuse | Link to this
  3. 3. rshoff 02:44 PM 8/18/09

    I don't think language shapes how we think, because I don't have the experience of thinking in words. I do however believe that language as a communication tool affects our interpersonal and cultural dialogue. Those in turn may affect how we think....

    Reply | Report Abuse | Link to this
  4. 4. rshoff in reply to HFB 02:54 PM 8/18/09

    We can do have the same concepts in English though. You and the article just expressed them using English! To blame the language for the ineptitude and ignorance of the speakers (most of us including myself) isn't right. Perhaps these particular words in French, or Turkish, or any other language are more accessible to the common speaker because of cultural dictates. What came first, the chicken or the egg?

    Reply | Report Abuse | Link to this
  5. 5. bitflung 03:12 PM 8/18/09

    when I was a child I held a firm but seldom mentioned belief that language was interrupting my thoughts. I would sit in the morning, gnawing on my sugary breakfast cereal, just thinking. I found one day that the cadence of my thoughts was far slower than just a moment before and noted that I was thinking in "words" rather than "ideas".

    I had a very modest speech impediment (I was partially tongue tied, causing S's and T's to be difficult and slow to pronounce). In my "mind's voice" my words were still too slow, as they were if I spoke out loud.

    I believe the opposite of that quoted from the English teacher. I believe that the more words you know, the fewer ideas you may have. The chaotic static that forms our ideas prior to be realized enough to put into words becomes pigeon-holed into our existing vocabulary the moment we recognize similarities.

    Learn more words and the similarities are noticed sooner, resulting in this pigeon-holing. Fewer words and you can't communicate your thoughts effectively.

    The brightest of us all work very hard through the years to provide a mental environment for our minds to work in. Work with language and that environment is based on words (making the english teacher's statement potentially true, at least for him/her). A Physicist, or Mathematician, or Engineer would, however, have to process more abstract thoughts. Until a conclusion is drawn (be it statement or question) language only taints the process of working with these thoughts.

    Many academics use kinesthetic mental aids, refer here: http://tinyurl.com/ko8l4m . Einstein was known to use visual and kinesthetic mental aids mostly, but also aural. Richard Feynman (if unknown to you, this physicist was one of the greatest yet to live) also used these mental aids as well as verbal tools (paradoxically I'm including this in my overall argument against language).

    Overall, language is often not an efficient tool for processing thoughts. Communicating thoughts, yes. Creating and working with thoughts, not so much. Unless of course the very thing you want to digest mentally are in fact words, in which case they might fit quite well indeed with the processes you go through in your efforts.

    Reply | Report Abuse | Link to this
  6. 6. rightly 04:22 PM 8/18/09

    Language is a learned behavior process. It is not different from any belief process that requires certainty. It is not rational. It is based upon feelings and emotional connections to socially acceptable behavior, rightness of meaning within a peer group, mush as the same process of development of any belief system and the conviction that speech relates to a reality.
    We could not learn to believe a religious dogma or the reality of numbers in any other way.

    Reply | Report Abuse | Link to this
  7. 7. ralphskinner@hotmail.com in reply to AndrewJayPollack 05:12 PM 8/18/09

    Excellent reply by AndrewJayPollock. Your comments are more interesting than the article itself.

    Reply | Report Abuse | Link to this
  8. 8. Oji in reply to AndrewJayPollack 07:07 PM 8/18/09

    The trouble with this analysis is that it is completely purely subjective. It isn't clear that you aren't just taking the aspects of the language and culture that fit your thesis.
    For example, Italian is just as structured as German (at least, relative to English) but the common stereotype of Italian culture is that it is very "laid back".

    Also, it could be argued that Japanese is a very vague language in use because so much is implied. The subject is nearly always ommited from a sentence, and sometimes the object and even the verb. Or all three. Also, contrary to what you suggest it is "very good at re-inventing itself for new ideas or concepts"; it is as creative as any other langauge and is a particularly rich medium for puns and wordplay. How does that fit with your view of the culture?

    Personally, I think there is probably a some effect of language on thought but it is very, very small. As shown by the difficulty of detecting it in any objective way.

    Your final question is very good: I suspect the indirectness of Japanese is more diven by cultural factors than the other way round. I'm not sure how that could ever be demonstrated in anything like an objective way, though.

    Reply | Report Abuse | Link to this
  9. 9. Sirlene 09:15 PM 8/18/09

    When you say "Try calling dry snow "dax" and wet snow "blicket," and see if you notice a change in how you think about snow. I didn't." You forgot something very important: those words do have different meanings for those people because it's part of their life, their culture, the place where they live in. I suggest reading Vygotsky and what he writes about language. The importance of it in terms of our consciousness formation. The important question in relation to what your English teacher said is that words without context and living experience ... have no meaning! That's why so many students around the world are not able to read and understand, interpret and make inferences...
    Just some other comment... The difference between sense and meaning, which may be something that can help.

    Reply | Report Abuse | Link to this
  10. 10. HKcynic 11:10 PM 8/18/09

    I think the relationship between language & culture plays an important role.

    I became convinced when living in Buenos Aires that we really don't have an equivalent in English for whatever "felicidad" means to Argentineans - or at least Portenos.

    The obvious translation - "felicity" is not used any more: is it because we don't really have that feeling / concept?

    Reply | Report Abuse | Link to this
  11. 11. HKcynic 11:11 PM 8/18/09

    I think the relationship between language & culture plays an important role.

    I became convinced when living in Buenos Aires that we really don't have an equivalent in English for whatever "felicidad" means to Argentineans - or at least Portenos.

    The obvious translation - "felicity" is not used any more: is it because we don't really have that feeling / concept?

    Reply | Report Abuse | Link to this
  12. 12. linguist 07:56 AM 8/19/09

    The article overstates what Frank and his colleagues showed about the Piraha. They did *not* show that the Piraha lack number words, though they did indeed make this claim. In fact, when asked to "count up", the Piraha acted just like speakers who have precise words for "one" and "two". Only when asked to "count down" did they start behaving differently. The authors give no explanation for this difference, and do not even express any curiosity about it (strangely enough) -- but simply pool the two separate results in order to support the conclusion that the Piraha have no number words at all, not even "one".

    Reply | Report Abuse | Link to this
  13. 13. galaxy_man 08:24 AM 8/19/09

    It is not words in a language that affect how we think. It's the structure. Each language has a certain form that shifts our thought process into certain channels. For instance, when my father learned to speak spanish, he told me that the majority of conversation in that language was conducted in the past tense. In Japanese, there is no explicit future tense; statements of future action or intent are brought into the present.

    These are very simple concepts but already one notices how they can impact a person's cognitive flow. I believe this is the main reason why young children are most adept at learning new languages - their own thought processes are not yet hard-wired in the form of their primary language.

    Reply | Report Abuse | Link to this
  14. 14. LloydChiro in reply to galaxy_man 09:57 AM 8/19/09

    "These are very simple concepts but already one notices how they can impact a person's cognitive flow."

    How can one's cognitive flow be "impacted?" Do you mean "affect," "alter" or "influence?

    Reply | Report Abuse | Link to this
  15. 15. Stuy Reject 02:20 PM 8/19/09

    The crucial point is whether the question is strictly empirical or to some extent philosophical; in other words, how we can obtain evidence, and further, how much depends on how thought can be defined.

    Reply | Report Abuse | Link to this
  16. 16. MikeB 02:24 PM 8/19/09

    How timely. I was just re-reading the collection of Whorf's essays entitled "Language, Thought and Reality."

    I find Whorf's hypothesis pretty convincing, vastly more so than Chomsky's transformational grammar approach. It seems, though, that brain mapping of speakers of various languages could be helpful in addressing the underlying concepts of Whorf's theory. If different languages -- 0r specific concepts unique to particular languages (for example, the "sanuk" emotional state in Thai) -- are found to be generated in parts of the brain distinct from speakers of other languages that would go far to demonstrate that the thoughts themselves are distinct and bound to the language. This research would be even more interesting if performed on compound bilinguals.

    Reply | Report Abuse | Link to this
  17. 17. Stuy Reject 02:25 PM 8/19/09

    The crucial point is how well the link between language and thought might be empirically studied, and how much the matter is subject to philosophical considerations

    Reply | Report Abuse | Link to this
  18. 18. garance 03:00 PM 8/19/09

    I root for the teacher. There is an old saying warning us that different languages imply different mindsets: "traduttore, traditore" (translator = traitor). And indeed I must confess that I do not quite think the same way in English and in French. Kids in my very poor neighborhood are not good at math because they do not understand the words, not by lack of math ability. Their reasoning is limited and rigid even for the brightest kids. So I am surprised that the paper does not talk about threshold. Living with 300 words is limiting your thought process compared to living with a million, that is easy to demonstrate, just as you can demonstrate that people who never heard of compound interest are not as good at business: they do not have the words, what are the chances that they will reinvent the concept?

    Reply | Report Abuse | Link to this
  19. 19. jkhartshorne in reply to linguist 08:32 PM 8/19/09

    'Linguist' refers to the fact that the authors did the counting task two ways: either by asking the Piraha to name count a collection of 1 object, then 2 objects, then 3 objects, etc., or by starting with 10 objects, then 9, then 8, etc.

    It is true that when starting from a small number and working up, the Piraha used the word baagiso 100% of the time for the single object, but not for any larger numbers. However, when working backwards, they started using baagiso for collections of 6 objects.

    'Linguist' makes a factual error is saying that this is true for 'hoI,' the word previously thought to mean 'two.' When counting forwards, they using 'hoI' for collections of 4, 5 or 6 objects around half the time, and about 1/5 of the time, they use it to describe 10 objects.

    Also, the authors do actually discuss why there would be a distinction between starting small and working up or starting large and working down, though it's a very short paper so the discussion goes by quickly. Bascially, it's an issue of grounding. Imagine you are using the words 'few' and 'many' to do this task. I show you one object, and you say 'few.' I show you 2 objects, and you probably still say 'few'. But by the time I get to 4 or 5, you might switch to 'many,' because at some point you want to make a contrast. But if we start with 10 objects, you say 'many.' Then with 9, you probably still say 'many.' But maybe by 6 or 7 you already switch to 'few'.

    In any case, it's clear that in English, you aren't more likely to call 5 apples "2 apples" depending on whether you previously saw 4 apples or 6 apples.

    Reply | Report Abuse | Link to this
  20. 20. linguist in reply to jkhartshorne 09:22 PM 8/19/09

    No, there was no factual error in my comment. In the counting up task, the Piraha used "hOi" (not "baagiso") exclusively when the quantity was 1. When the quantity was 2, they used only "hoI".

    Yes, they did use "hoI" for some quantities greater than 2 as well, but that's exactly what we expect if the word means "two" but there are no dedicated words for higher numbers. It's what children do, for example, when they have learned the word for number "x" but not "x+1" or higher -- they use "x" to answer "how many" questions for which the answer is "x+n". The reasons for this are slightly too involved to go into on this comments page, but here's one recent paper on the topic: http://ladlab.ucsd.edu/pdfs/BB-exactness.pdf.

    So if we restrict our attention to the counting-up tasks, the Piraha in the Frank et al. study did act just as expected for a people with words for "one" and "two", but not for higher numbers, just as I wrote.

    So the question remains: why do behave differently in the counting-down task? About this, you write: "Also, the authors do actually discuss why there would be a distinction between starting small and working up or starting large and working down, though it's a very short paper so the discussion goes by quickly." I don't see this discussion anywhere in the paper -- did I miss it somewhere? You included a link to it. What page is this discussion on, and where? The only mention I can find is in the paragraph that overlaps pages 820-821, but this paragraph just punts.

    I don't understand what you wrote about "grounding", unfortunately.

    Reply | Report Abuse | Link to this
  21. 21. robject 04:34 AM 8/20/09

    I was skimming the comments hoping to find one that mentions the fact that Whorf did also lay great importance with the structure of language, specifically the subject-predicate format of "western" languages, opposed to, say, American Indian languages which he studied extensively. This is an aspect that has profound impact on how we see the world. In a subject-predicate language it is possible to "label" the world: "John is stupid" "It is cold". In many languages this structure is simply not possible, and this may have a fundamental influence on one's world view, and may even be related to the development of western science in which the object and the subject of research are so much divided.

    Reply | Report Abuse | Link to this
  22. 22. dkom 07:54 AM 8/20/09

    I happened to speak in 2 languages fluently and one partially. Your way of thinking is always influenced by the language most used which doesn't have to be your native one. Going back few years, I found myself in the transition of changing my thinking from native language to learned one and in a state where I forced myself to think only in a new one. Because of limited amount of words and conscious blocking access to a native one's vocabulary my way of thinking became less complicated and therefore (whoever it may be interpreted) life in general easier, throwing out all the unnecessary over-complicated thinking.

    Reply | Report Abuse | Link to this
  23. 23. AJV 08:36 AM 8/20/09

    "Try calling dry snow "dax" and wet snow "blicket," and see if you notice a change in how you think about snow. I didn't. The English book's statement assumes that if you don't have a word for something, you can't talk about it ... a claim that the sentence proves false."

    but didn't you just change the names of dry and wet snow, so they still have words attached to the concept? I always thought the point was that there is no word for the concept, not that the word was different to what I already know.

    I have never seen snow (really) and could not tell you what 'dry' or 'wet' snow is - if i saw wet or dry snow i would call it 'snow', whether it was really 'dax' or 'blicket'.

    Reply | Report Abuse | Link to this
  24. 24. jkhartshorne in reply to linguist 01:14 PM 8/20/09

    Thank you for your correction on which word means which. I always forget. The distinction between the "counting-up" and "counting-down" tasks is on the last paragraph of 820, but I agree that the point is not clear if you aren't looking for it. By "grounding," I mean fixing the use of a relative term. "Tall" means one thing when you are talking about basketball players and a different thing when talking about preschoolers. You would expect the words "tall" and "short" to be used very much the way the Piraha used their "counting" words.

    As far as whether 2-knowers perform in the way you suggest. You cite recent work by Dave Barner. I'm a big fan of Dave's, am very sympathetic to the direction he is going in that work, and have collaborated with him on a related project, but I certainly wouldn't say that his view is by any means established. There is a lot of data to collect first.

    Even then, one would need to show similar results between Piraha and "one-knowers". I personally doubt they'll come out the same, but it's a testable question. For one thing, on Dave's and Asaf's account, I don't see any reason that 'one' would be non-exact when counting down rather than counting up. Actually, it would be very problematic for the theory if this is the case, since under their theory, "one-knowers" should use "one" to refer to one object regardless of context (maybe not "downward-entailing" contexts, but that's outside this discussion).

    So on the standard view, at least, Piraha do not behave as if they have words for "exactly 1" and "exatly 2". Even the Barner & Bachrach account probably predicts something else.

    Reply | Report Abuse | Link to this
  25. 25. jkhartshorne in reply to AJV 01:27 PM 8/20/09

    The point I was trying to make is that having a word for something doesn't mean you can't talk or think about it. You may not know a word for 'wet snow,' but it hasn't stopped us from talking about it!

    It's a general phenomenon in language that the more often something is talked about, the fewer syllables are needed (Zipf's law). So -- on average, anyway -- common objects have short names (cat, dog), and uncommon objects have long names (elephant, salamander). You can extend this to phrases: things rarely talked about may not have words at all. And this makes sense: who need a word for something you never talk about?

    So whether or not a language has a word for something probably tells us nothing more than what they do or don't talk about.

    Reply | Report Abuse | Link to this
  26. 26. linguist in reply to jkhartshorne 03:33 PM 8/20/09

    I'm sorry, no. Though the last paragraph of page 820 of Frank et al's paper mentions "the distinction between the 'counting-up' and "counting-down" tasks", it gives no explanation for it at all and, as I wrote in my first message, doesn't even treat it as a puzzle. Instead, they propose to pool the counting-up and counting-down data, which of course eliminates from consideration the entire "counting-up" effect. But that's not a solution to the problem, it's merely a way of hiding it.

    Fortunately, the paper is on-line so we don't have to bicker about it. Your readers can read the paragraph in question and see for themselves.

    Finally, regardless of whether we accept Barner & Bachrach's *explanation* for the behavior of children who know "one" and "two" when asked "how many" questions, the Piraha are acting -- in the counting-up task, that is -- just like such children.

    Which, yes, leaves the question of their counting-down behavior wide open, and a big puzzle: for everyone.

    As long as the puzzle remains unsolved, the Frank et al. paper has not "demonstrated that Pirah�, a language spoken by a small Amazonian community, has no number words at all," as you claim. Only by disregarding half of their data can you reach that conclusion.

    Reply | Report Abuse | Link to this
  27. 27. jkhartshorne in reply to linguist 01:47 AM 8/21/09

    It's not necessary to disregard any of the data. The claim that the Piraha words mean something like "few," "some" and "many" appears to predict all of the data. I say "appears" because the experiment hasn't been run, but I'd bet money that it would come out roughly as expected.

    Again, imagine you are describing people as tall or short. If you've just been talking about a bunch of 5'3'', you'll probably call someone who is 6'1'' tall. But if you've just been talking about a bunch of 7'3'' NBA players, you are unlikely to call the 6'1'' tall anymore. That's because the word "tall" is relative; it doesn't refer to a specific height.

    So in the "count-up" task, you start with a small number of objects, so you use your smallest quantity word. Add an object, and you move on to the next quantity word in order to make a distinction. Pretty soon, you're out of quantity words (having only 3). In the "count-down" task, you start with a large number of objects and so you use your largest quantity word. Remove an object, and you might want a smaller quantity word in order to make a distinction. So you end up using your smallest quantity word earlier, for larger numbers.

    This would be easier to describe if I could use pictures, but hopefully that made sense.

    To be fair, it's no doubt possible to come up with a theory on which the Piraha really do have words for "one" and "two" and which describes the data. But I don't know of such an account. It would require a lot of additional hypotheses.

    Reply | Report Abuse | Link to this
  28. 28. linguist in reply to jkhartshorne 08:17 AM 8/21/09

    Your proposal is interesting, and it should be tested, as you say. The first point to stress, however, is that it is your proposal, not Frank et al's, Once again, I make the point that Frank et al. offer no account at all for the difference between counting up and counting down, and that consequently their paper is absolutely not a "demonstration" that the Piraha lack words for "one" and "two".

    Even your story, furthermore, cannot in principle explain the actual data reported in the Frank et al. paper, because *someone* must have switched back from "many" to "a somewhat larger quantity" on the way down from 8 to 7. (Look at the graph.) We do not have data from the individual subjects, so it's possible that there were other switches as well -- but this one definitely happened.

    Bear in mind also that the counting-up task involved 6 subjects, while the counting-down task involved only 4 (as the paper makes clear). Also, the question asked of the subjects is translated in the paper as "how much/many is this", but the Piraha do not actually have a dedicated expression for "how much" or "how many" (and the authors do not tell their readers in the article what was actually asked by the experimenters). According to Everett's earlier grammar of the language, they ask "how many is this" by asking "what is this", which might open up an entire other can of worms for interpreting the data. Just a "hoI" more reasons for caution and skepticism in considering this paper a "demonstration" that "Piraha has no number words".

    Reply | Report Abuse | Link to this
  29. 29. linguist in reply to linguist 08:30 AM 8/21/09

    A typo in my previous message. I wrote:

    "Someone must have switched back from 'many' to 'a somewhat larger quantity' on the way down from 8 to 7. "

    I meant:

    "Someone must have switched back from 'a somewhat larger quantity' to 'many' on the way down from 8 to 7. "

    Reply | Report Abuse | Link to this
  30. 30. Nitpicker 01:21 PM 8/21/09

    Goodness lost.

    All this talk about words and nobody noticed the loss of Goodness from the acronym bans. Should it not be banGs?

    Reply | Report Abuse | Link to this
  31. 31. swmuruga 06:09 PM 8/24/09

    well i certainly can visualize 17 balloons or any other object. I just see 3 rows of five plus 2. mind is flexable and can solve such problems. Yes Feynman was one of the all time great minds of the 20th cent.

    Reply | Report Abuse | Link to this
  32. 32. umbahli 10:44 PM 8/24/09

    Language is culture? Culture is language?

    Reply | Report Abuse | Link to this
  33. 33. umbahli 10:47 PM 8/24/09

    Language is culture? Culture is language?

    Reply | Report Abuse | Link to this
  34. 34. yyzhou in reply to AndrewJayPollack 01:47 AM 8/25/09

    I do agree that there is a missing point in this passage, especially that it mentions language associates with our memory more than it does with the thought. Yet, how would thinking be launched without memory? I doubt that memory is the most essencial component of our thought, in which case langugage does affect our thought more than the author indicates.

    Reply | Report Abuse | Link to this
  35. 35. ph0enix 06:05 PM 9/5/09

    1984 anyone? In that book a suppressive socialist government grossly simplifies English, called 'Newspeak', in order to give the language less words with stricter means to keep people in check. The idea was that less words meant less thought and communication. The less communication part would work to a degree I suppose... but less thoughts? Doubt that, as I'm sure most people don't think strictly in words!

    Reply | Report Abuse | Link to this
  36. 36. mcswell 08:42 PM 9/9/09

    Maybe the learners of French would do better with the acronym 'bangs'.

    BTW, I see no sense in which German or Japanese is more structured than English. Different structures, yes; more structured, no evidence.

    Reply | Report Abuse | Link to this
  37. 37. mdarr 04:59 PM 9/28/09

    The Whorf hypothesis has been around for a long time and I'm suprised at any controversy. The study cited is no news at all. The same memory effect was discovered in the 70's with color chips. For short term memory we don't need a word. For long term memory we need a word. Fortunately for humans, if we need to remember something bad enough we make up a word.

    Reply | Report Abuse | Link to this
  38. 38. clydemckendrick 05:15 AM 10/8/09

    Facinating: I'm no linguist, but it charges the question of whether language influences cultural behaviour or vice versa, or perhaps both.

    Things like innovation, engineering advancement, creativity all seem closely linked to the way we use language flexibility.

    We wrote an article in our MILK magazine recent issue questioning the future of swearing and profanity in English and how new composite words and phrases are constantly being invented in pop culture to sharpen the impact of a word.

    Reply | Report Abuse | Link to this
  39. 39. jderwin 09:52 AM 10/11/09

    Of course, language and culture are tied. Languages with female vs male articles, and those with different sentences and even language itself for male and female, manifest the culture of each. Even in English vs German, both from common origins, differ in articles, English having one and German having three major articles (der, die and das) specifying male, female and neutral. Has this affected the cultural differences and the roles of each gender? Think about it.

    Reply | Report Abuse | Link to this
  40. 40. JamesSmith in reply to Nitpicker 12:49 PM 10/18/09

    In my years of French class it was "BAGS" rather than "BANS."
    Beauty, Age, Goodness and Size.
    I don't recall numbers, even when describing quantity, as being referred to as "adjective(s)" as the author does in the article.

    Reply | Report Abuse | Link to this
  41. 41. eversostrong 09:19 PM 10/21/09

    You obviously felt compelled to include examples of your your political beliefs in this article.
    You are a piece of work, mister.

    Reply | Report Abuse | Link to this
  42. 42. GusGus 08:27 AM 10/26/09

    In English we have a word for light red - Pink! How does that make it any easier for us to differentiate shades of red?

    Reply | Report Abuse | Link to this
  43. 43. mmdodol 07:53 AM 3/14/10

    I speak four languages ( Persian Dutch English and French)
    I was born in Iran and went to school there for 3 years. At the age 9 we moved to Belgium where I live now and I learned Dutch. When I was 15 I took a job at a restaurant where everyone spoke French, I had my basics from school and the fact that I live 25 km from the French border helped a lot too but it was at this restaurant where I got really fluent in French. And finally thanks to Hollywood and the Internet I learned English. That and I went to live with my uncle in England for a year when I turned 16.
    So Persian and Dutch are my primarily languages and 75% of what I think is in either one.
    I mostly think in Persian for emotional thoughts like love anger joy frustration etc. and I think because I was educated in dutch my more rational thoughts are in Dutch. This is of course a simplification
    I don't know about how other people think but I have constant conversation in my head when I'm really thinking ( when i'm alone). Like on an exam when I really get in to it and its really hard I even start mumbling sometimes.

    So what s my point? I know 3 of these languages as they were my mothers language ( Persian Dutch and English) and French well I hate French it's a very hard language to learn in my experience. And still I don't have more ideas or thoughts than anyone else hell I'm even bottom of my year and I'm in a class with kids 2 years younger than me so maybe maybe maybe nothing there is no point nothing but I had a point when I started writing and all the background info got me distracted oh well it still is a good read and is still on topic?

    Reply | Report Abuse | Link to this
  44. 44. rbaum in reply to rshoff 04:41 PM 6/4/11

    which*

    Reply | Report Abuse | Link to this
  45. 45. Captainllama 09:01 PM 8/20/11

    In his excellent book "What Do You Care What Other People Think?" Richard Feynman relates an exchange that caused him to change his mind about thought and words. ""Thinking is nothing but talking to yourself," said Feynman. His friend referred to the crazy shape of a crankshaft in a car and asked "How did you describe it when you were talking to yourself?"

    Reply | Report Abuse | Link to this
  46. 46. Cultures Connection 04:14 PM 10/23/12

    Very interesting article. Check out our contribution to the conversation: http://www.culturesconnection.com/does-language-shape-the-way-we-think/

    Reply | Report Abuse | Link to this
Leave this field empty

Add a Comment

You must sign in or register as a ScientificAmerican.com member to submit a comment.
Click one of the buttons below to register using an existing Social Account.

More from Scientific American

Follow Us:

See what we're tweeting about

Scientific American MIND

More »

Free Newsletters


Get the best from Scientific American in your inbox

Solve Innovation Challenges

Powered By: Innocentive

  SA Digital
  SA Digital

Science Jobs of the Week

Email this Article

Does Language Shape What We Think?

X
Scientific American Mind

Subscribe Today

Save 66% off the cover price and get a free gift!

Learn More >>

X

Please Log In

Forgot: Password

X

Account Linking

Welcome, . Do you have an existing ScientificAmerican.com account?

Yes, please link my existing account with for quick, secure access.



Forgot Password?

No, I would like to create a new account with my profile information.

Create Account
X

Report Abuse

Are you sure?

X

Institutional Access

It has been identified that the institution you are trying to access this article from has institutional site license access to Scientific American on nature.com. To access this article in its entirety through site license access, click below.

Site license access
X

Error

X

Share this Article

X