The Changing Tsimané Spectrum.

Elise Cutts reports for Scientific American on an interesting form of borrowing:

Like the ancient Greek of Homer’s time, the Tsimane’ language has no set word for the parts of the color spectrum English speakers call “blue.” Although Tsimane’ does name a number of more subjective hues (think “aquamarine” or “mauve” in English), its speakers—the Tsimane’ people of Bolivia—reliably agree on just three main color categories: blackish, reddish and whitish.

But bilingualism is reworking the Tsimane’ tricolor rainbow, researchers recently reported in Psychological Science—offering a rare, real-time glimpse into how learning a second language can change how people think about abstract concepts and fuel language evolution. The data show Tsimane’ speakers who also speak Spanish are borrowing the concepts of—but not the Spanish words for—new color categories such as blue, green and yellow.

“You could have imagined that they could have just started calling things amarillo and azul” (the Spanish words for yellow and blue), says lead author Saima Malik-Moraleda, a cognitive neuroscientist at Harvard University and the Massachusetts Institute of Technology. But instead “they’re repurposing their own Tsimane’ color words.”

Malik-Moraleda and her colleagues asked 152 people who spoke Tsimane’ or Spanish, or both, to name and sort a set of 84 differently colored chips. Bilingual participants sorted the colors into narrower categories in both languages. For example, to describe blue and green chips in Tsimane’, they chose two hazy Tsimane’ color terms, yushñus and shandyes, and repurposed them as distinct, consistent labels for their freshly acquired color concepts. But monolingual Tsimane’ speakers used these words interchangeably for bluish and greenish colors.

These results add to the evidence that the languages we speak affect how we slice up the rainbow into color categories. The new study documented this effect in real time as a language evolved; other work in languages such as Greek and Welsh has relied on historical evidence to study this transition after the fact, says psycholinguist Panos Athanasopoulos of Lund University in Sweden. “What they’re showing is the end process of this conceptual restructuring” associated with bilingualism, he says.

Although the results reflect changes in how bilingual people talk about the world, they do not reveal whether these speakers actually perceive colors differently than monolingual people do. Psycholinguistics evidence suggests the language we speak can subtly influence how our brains process what we see. Testing whether bilingual Tsimane’ speakers’ use of color terms reflects a difference in how they experience color—that is, whether their brains react to color differently than those of monolingual people—would be a fascinating follow-up, Athanasopoulos says.

Chimané (Tsimané) is a South American language isolate; I don’t know why SciAm uses a weird vertical prime symbol (which shows up in my post as a single quote) rather than a standard acute. More on color categories at LH: 2007, 2010. Thanks, Eric!


  1. uses a weird vertical prime symbol (which shows up in my post as a single quote) rather than a standard acute

    I think the autonym ends in glottal stop, represented by ⟨ ’ ⟩ in a commonly used orthography. See section 2.1.2 on the consonantal orthography and also example (56), section 2.4.2, in Jeanette Sakel, “El Mosetén y Chimane (Tsimane’)”, available here:

    (56) Anic muju’cha’ tsun-tyi’ chätidye’-in Tsimane’-in.
    seguro más.M 1PL-NEX.M relativo-PL chimane-PL
    ‘De verdad, había más de nosotros, chimanes (que otros pueblos tribales).’

    (And unfortunately the typeface used for the text on that SciAm page apparently has no curly apostrophe or single quotation mark.)

  2. Thanks!

  3. David Eddyshaw says

    This is a pretty familiar process: it’s basically just calquing.

    There’s nothing special about it happening with colour words: the idea that it all has some deep neurolinguistic significance comes from the ever-popular usual sub-Sapir-Whorfian blether regarding cross-linguistic colour words.

    Kusaal too has just three basic colour terms, but this does not mean that Kusaasi perceive the world differently: there are plenty of ways of describing different colours; just not by using underived simple adjectives. Some of these ways are every bit as stereotyped and fixed as English colour terms, others are more opportunistic coinages.

    No Oti-Volta language (as far as I know) has a “word for yellow”; but all the ones I have relevant data for call yellow things “dawadawa-coloured.” It’s a fixed expression.

    More generally, it is not the case that in a language with only three basic colour terms those terms “mean” red, white and black. That’s a translation artefact. Canonical red, white and black things will be given those labels, but “white”, for example, really means “all lighter hues, relevant to the colour comparisons we are considering at present.” This stuff about the Tsimane’ “only reliably agreeing on” three colours is a culture-bound misunderstanding of how the system actually works. It’s like concluding from the fact that “few” and “many” do not have fixed numerical meanings in English that English speakers have no concept of exact numbers.

  4. The colour space needs 3 axes to specify a colour. An RGB value is one way to do this. Another method is HSI (hue, saturation, intensity). And there are several other methods in use.

    Just because your own language uses an RGB model, that does not mean that speakers of languages that use an HSI or other colour model “do not perceive colour” or “lack colour concepts”.

  5. David Eddyshaw says

    Presumably all human beings who are not actually colour-blind have the same kind of colour inputs from their eyes (indeed, it is this presumption which underlies all this work that has been done on colour terms across languages.)

    The actual mechanism of colour vision is in one sense well understood, with the human retina having three different types of “cone” photoreceptors, each maximally sensitive to a different wavelength.

    However, there is (astonishingly) no currently adequate theory which explains how this relates to our actual perception of colour; some completely different combinations of primary colours are indistinguishable to all human beings, for example, and nobody has worked out how this functions. A lot of it seems to be in post-processing in the retina itself, which is very far from being a mere passive receptor.

    I get the feeling that all the enthusiasm that has gone into the work on cross-linguistic differences in colour terms is at least partly due to a serious underestimation of how complicated colour vision is in purely physiological terms and how little we understand it. It’s been mistaken for a nice conceptually simple case, when it’s anything but.

    We do not see wavelengths; we do not directly perceive “the spectrum” (in the sense that physicists use the term.)

  6. Uh huh. Light (unsaturated) red is regularly called by “pink” in English – never “light red”. Light black we call “grey”. How backward we anglophones are, lacking the equivalents for “light blue” and “light green”. How our reception of the visual world would be enriched, if we had basic colour terms for those!


    … three different types of “cone” photoreceptors, each maximally sensitive to a different wavelength

    Yes, but let’s not forget that rods can also contribute to our experience of colour, in certain lighting conditions.

    We do not see wavelengths; we do not directly perceive “the spectrum” (in the sense that physicists use the term[).]

    Several times I’ve had to correct misconceptions akin to that, in the writing of academics. One was a renowned philosopher who published often on colour perception.

    People typically assume that because both sound and light are wave phenomena our auditory and visual senses must be pretty similar. How wrong they are!

    And another thing: find me a luminary who can explain (as we can, eh DE?) why it is impossible for most observers to distinguish a pure green dot from a pure blue dot flashed to central vision on a computer screen.

  7. There are ‘vestiges’ of the earlier light-dark system in other languages. As it’s the grape season here in Australia, I’ll use this as an example:

    In English, “white” still means “lighter-coloured” when it describes white wine or white grapes. Similarly in Croatian – “bijelo vino” & “bijelo grožđe”.

    Though the darker coloured wine & grapes are “red” in English (are they really red though?). But in Croatian they are “black”.

    Colour names really do cover a spectrum of meanings (pun intended).

  8. David Eddyshaw says

    In principle, there might be women who can perceive four distinct primary colours, but unfortunately I know of no properly documented real examples.

    I imagine you could easily go your whole life never actually realising that you could see more colours than everyone else, particularly as your language would presumably not give you any way of labelling the differences. “But that one’s a different colour! Can’t you see? It’s more … different …”

  9. J.W. Brewer says

    Hmm. Can those qualia-lacking “philosophical zombies” who recently stumbled through another thread talk about the colors they seem to see?

    @zyxt: see also English “white” (and parallel words in other languages) as applied in a racial sense to human beings whose skin is in reality typically sort of pink-to-beige.

    In Iberian languages, the opposite of “white” wine is “tinted” wine – vin[h]o tinto.

  10. David Eddyshaw says

    One of the things that has always bothered me about a lot of philosophising about qualia is that ignores known physiological fact. It’s all very well, for example, for the Oxford Analytic Mafia to claim that we do too perceive yer actual things

    but as a matter of plain physical fact – we don’t. We don’t possess the necessary apparatus. Even our experience of the present moment is a helpful physiological illusion, doubtless of considerable value in helping our forebears to evade sabre-toothed tigers and what have you.

    My supposed tetrachromat woman is OK though. She needn’t get driven to despair by Wittgenstein’s private language argument, because she can just go ahead and attach labels to objects saying “Colour A” and “Colour B” and come back the next day to check that the labels still look right, and get her friends to confirm that she’s doing it consistently. She can go all behaviourist on herself, if she wants to, with a little help from her friends. She doesn’t have to try to attach labels to qualia.

  11. DE:

    In principle, there might be women who can perceive four distinct primary colours

    Heh, and how would they go with the green-and-blue-dot uncertainty I mentioned? We’d need to know about the specific distributions of their cone types across regions of the retina.

    Of course people with so-called normal vision do perceive four “primary colours”: red and green (in one post-cone opponent process) and blue and yellow (in another). Um … it’s not so black and white (d’oh! a further complication).

  12. John Cowan says

    In Iberian languages, the opposite of “white” wine is “tinted” wine – vin[h]o tinto.

    In Colombian Spanish, quoth en.Wikt, tinto means ‘black coffee’. Furthermore, en tinta means ‘in ink’. Neither coffee or ink is understood to be dark red.

  13. PlasticPaddy says

    Re tinto, some of the wine grapes are called x negro and red wine in Northern Italy is nero, not rosso. I don’t know if tinto/rosso replaced earlier “negro” for wine, or if there was a need to distinguish different styles. For me grapes used in red wine are typically purple with a strong dark blue component and the lighter grapes are “green grapes”, not “white grapes”.

  14. January First-of-May says

    Light (unsaturated) red is regularly called by “pink” in English – never “light red”.

    …apparently the shade known as “coral pink” is indeed considered to be a kind of pink in English. To my perception this cannot be anything but light red.

    (This came as quite a surprise to me, as I also thought that light red was called “pink” until I encountered objects that were legitimately light red. Pink proper is slightly more purplish, AFAICT.)

  15. As to what “pink” et al mean to the average anglophone, the survey at is interesting. It literally only scratches the surface of the colour space, but the pink region stretches markedly far from the “light/pale” edge towards the dark one.

  16. David Eddyshaw says

    its speakers — the Ang’lofone people — reliably agree on just five main color categories …

  17. Amanda Adams says

    One of the things that pleases me about Catalan is that wine, grapes, & olives come in black & white.
    I feel I should mention (I expect I’ve mentioned it before) that many years ago on some BBC radio program about…scientific inventions?…I heard the inventor speaking about an “anomaloscope” which somehow allowed the user to match a colour being observed with a spectral colour being generated. I can’t intuitively grasp how this worked, but the contention was that it could demonstrate whether or not one person’s “green” was the same as another person’s “green”. I wish the BBC would follow up on this.

  18. Amanda Adams says

    Well, I’ll be gobsmacked! One can now get something by googling the word!

  19. David Marjanović says

    Black tea is red in China. Only actually fermented tea is black there.

    In principle, there might be women who can perceive four distinct primary colours, but unfortunately I know of no properly documented real examples.

    Not even Concetta Antico?

  20. David Eddyshaw says

    That’s the most plausible case I’ve heard of, but even that is not completely cut and dried.

    Of course, it’s no use just having the right kinds of photoreceptors if they’re not connected up right; even then, your brain might only be trichromat. You’d be like one of those Apple watches where they’ve disabled one of the sensors in the software, or a laptop lacking the right drivers for some peripheral.

    Of course, you could also be objectively a tetrachromat and subjectively only a trichromat. I’m now trying to make up a greekoid word for “blindsight, but only with respect to a colour that nobody, including you, can actually see.”

  21. CuConnacht says

    Tea without milk is red in the Arab world as well.

  22. PlasticPaddy says

    Just keep taking those hallucinogenics….

  23. DE:

    Of course, it’s no use just having the right kinds of photoreceptors if they’re not connected up right; even then, your brain might only be trichromat.

    But as I pretty well observe above, the standardly wired brain is not trichromat but tetrachromat. What all this implies for supposed four-cone-type cases is a complication upon that fact – not upon the spurious supposition that the typical brain (as opposed to the typical retina) is in any useful sense merely trichromat.

  24. David Eddyshaw says

    That is not the standard use of the words “trichromat” and “tetrachromat”, though I respect your determination that when you use a word, it means just what you choose it to mean, neither more nor less.

  25. But it was you, DE, who initiated a non-standard use of the word trichromat: “your brain might only be trichromat”.

    It’s a case in which attributive and predicative work rather differently. No brain is trichromat in any standard sense, though one could reasonably speak of a “trichromat brain” meaning “the brain of one who is a trichromat” – just as one speaks of the “vertebrate brain” without importing that any brain is itself a vertebrate (with all the knobs and bits).

    I simply worked with your transformed wording, which might easily be taken to suggest that the “normal” brain itself does colour perception and cognition in some trichromatic fashion. I pointed out, mainly so no one here would get the wrong impression, that the brain doesn’t work like that. If we had to choose between characterising the typical human brain as intrinsically trichromat or intrinsically tetrachromat, we’d do better to choose the latter. Agree? If not, why not?

  26. David Eddyshaw says

    I have made my peace with the fact that you evidently know more about this than I can ever hope to. Though this be but a humble sort of wisdom, yet I feel I may claim it for my own. I shall now go and cultivate my garden.

  27. “The thought is that the retina is trichromatic and that the brain is tetrachromatic, with the paired incompatibilities of the Hering theory (no reddish greens) holding, not at the retina, but in terms of central mutual inhibitions.”
    – Boring, EG (1939). [Review of Experimental Psychology, by RS Woodworth]. American Journal of Psychology, 52(1), 131–138.

    A good thought it was, and a good thought it remains.

    How could I claim to know more about human vision than an accomplished ophthalmologist, coruscatingly witty and articulate in many domains – including linguistic realms of which I ken zilch? I only say it as I see it, as a philosopher (trained first through a biologically oriented major in psychology, and grad work in the same vein) who has studied and researched a good deal in matters arising from psychology and neuroscience.

  28. David Marjanović says

    no reddish greens


    Artists have known that for centuries and have been mixing red pigment into their green pigments for painting plants.

    Also… brown.

  29. Of course, red algae, which have a lot of phycobiliproteins as light-absorbing components of their photosystems in addition to chlorophylls, mostly (but not universally) appear red or purple to human eyes.

  30. DM:

    Artists have known that for centuries and have been mixing red pigment into their green pigments for painting plants.

    But the result would not be a reddish green. It should (depending on the exact nature of the pigments) be as if they had added some black.

    Also… brown.

    What, about brown?

  31. You can hear Chimane in use in the community in this clip (‘Enseñanza y aprendizaje de la escritura en lengua Tsimane’) on YouTube. The class discusses the palm species (apparently Attalea phalerata) known in Chimane as manaʼi and in Bolivian Spanish as motacú, as in the subtitles. (I wonder where this latter word comes from. Is it of Jabutian or kindred origin? See reconstructions number 50 ‘caroço de aricuri’ *mətaj and 37 ‘cacho de côco de aricuri’ *mətajtʃu on page 160 of Hein van der Voort (2007) ‘Proto-Jabutí: um primeiro passo na reconstrução da língua ancestral dos Arikapú e Djeoromitxí’, in Boletim do Museu Paraense Emílio Goeldi. Ciências Humanas 2 (2), available here.)

Speak Your Mind