Martian Colors

By Sean Carroll | November 5, 2007 11:36 am

I’m back from the Beyond Belief II conference at the Salk Institute in La Jolla, which packed an extraordinary amount of intellectual stimulation into a few short days. Any conference where you wander into the opening reception, get drawn into a conversation about reductionism and meaning with Stuart Kauffman, Rebecca Goldstein, and Sir Harold Kroto, and end up closing down the bar, is bound to be a good one, and this did not disappoint. (The title notwithstanding, much of the conference had little to do with atheism or religion — the subtitle “Enlightenment 2.0” gave a better flavor.) The talks provided fodder for at least ten to twenty blog posts, of which I’ll probably get around to writing one or two.

One of the talks was by local neuroscientist V.S. Ramachandran, or “Rama” to his friends. (Like any good neuro person, his web page includes a fun collection of optical illusions.) He talked about his experiments with synesthesia, the phenomenon in which people see graphemes (e.g. numbers or letters) as associated with colors. I do that a little bit — five is certainly yellow, seven is red, and eight is blue — but it’s closer to a vague association than a vivid experience. Some people report very strong synesthetic reactions, and for a long time researchers have wondered whether the experience was mostly metaphorical or something stronger.

synes-1.jpgTo test synesthesia, Rama and collaborators designed an experiment where they could measure the vividness of the colors associated with the numbers 2 and 5. They chose those because you can make them look almost identical, although reversed, by choosing a boxy font. Then they made up a picture (on left) of mostly fives, with a few twos scattered within there. Then they asked people to pick out the twos. Most ordinary folks could do it within about twenty seconds or so.

synes-2.jpg But true synesthetes could do it immediately. That’s because to them, the twos popped out as a brightly colored triangle (right). This established beyond much doubt that synesthesia was “real,” and more particularly that was a measurable phenomenon with real consequences.

This, in turn, strengthened the hypothesis that the origin of synesthesia was to be found in the structure of the brain. Indeed, it turns out that the region of the brain responsible for processing graphemes lies adjacent to the region responsible for processing colors.


It’s fairly easy to imagine that small alterations in “normal” brain wiring — strengthening some pathways, inhibiting others — could be responsible for synesthesia. To test this idea further, Rama had synesthetes look at alternative representations of the numbers — using Roman numerals, for example — and found that the colors did not appear. That’s perfectly consistent, as the relevant bits of brain are responsible for processing the graphical representation of the number 5, not the abstract mathematical concept of “five.” Further studies using functional magnetic resonance imaging have found that the brains of synesthetes and non-synesthetes really do light up in different ways when they look at numbers. (There are other ways to be synesthetic — associating colors with sounds or smells, for example — which may also be traced to connections between different parts of the brain.)

But then here’s my favorite part. They found a synesthete who was color blind. That may seem strange, but what it really means is that the subject had problems with his retina that left him able to distinguish only an extremely narrow range of wavelengths when looking at most images in the world — his brain was fine, but his eyes weren’t quite up to the job. But when he saw certain numbers, he experienced colors that he otherwise never saw. Here is Ramachandran in Scientific American:

We also observed one case in which we believe cross activation enables a colorblind synesthete to see numbers tinged with hues he otherwise cannot perceive; charmingly, he refers to these as “Martian colors.” Although his retinal color receptors cannot process certain wavelengths, we suggest that his brain color area is working just fine and being cross-activated when he sees numbers…

The effect is most obvious and pronounced in the colorblind synesthetes, but occurs in “regular” synesthetes as well. The colors evoked by cross activation in the fusiform gyrus “bypass” earlier stages of color processing in the brain, which may confer an unusual tint to the colors evoked. This is important for understanding the phenomenon of synesthesia, because it suggests that the qualia label–that is, the subjective experience of the color sensation–depends not merely on the final stages of processing but on the total pattern of neural activity, including earlier stages.

Martian colors! How awesome is that?

  • tacitus

    Very cool.

    There is a series of lectures from 2003 by V.S. Ramachandran called “The Emerging Mind” online at the BBC website:

    He is an engaging speaker, and they are well worth a listen. Synesthesia features in lecture four: “Purple Numbers and Sharp Cheese”.

  • rusell

    But true synesthetes could do it immediately.

    I don’t quite get the idea. A synesthete should first read the two as a two before she could associate it with the color red. This seems pretty obvious to me: If I tell a synesthete “tell me the color of the number I’m thinking of” he won’t be able to tell before I tell him what number it is. But then, if she has to read each five to know it’s green, and each two to know it’s read, finding all the twos should take her about the same time as it would take any other person.

  • Neil B.

    I think being able to experience color sensations you never had or could have before (because not made from addition of retinal input, as per RGB signals) would be really cool. (PS: If you are a misguided naive realist philosopher who doesn’t understand the concept of color sensations because you think we just “see colors” (light I guess, “directly”), I can’t help you very much.)

  • Rien

    Interesting! I’m a synesthete with respect to sound: I see music and other sounds as some kind of flow of colored patterns, but I never thought of numbers as having colors or anything. Is there some relation between these types of synesthesia? I did see the 2’s fairly easily…

  • rusell

    As a side note, two outstanding examples of synesthetes. First, there’s a superb poem by Arthur Rimbaud about synesthesia titled “Vowels” (“Voyelles” in the original French) that begins “I invented the colors of vowels”. He conjures up weird images and colors in connection with (French) vowels. Maybe that explains part of his incredible linguistic abilities.

    Second, I don’t remember in which of his books Richard Feynman refers to the colors of Bessel functions, and tells how for him special functions had that kind of perceptual properties. Maybe someone else can be more precise about this than I…..

  • tacitus

    I don’t quite get the idea. A synesthete should first read the two as a two before she could associate it with the color red. This seems pretty obvious to me:

    The idea is that it’s not a conscious effort. The color version of the 2/5 diagram really is an accurate analog of how synesthetes see it. I think for people like us who have never experienced it, it is almost impossible to imagine, it just sounds so weird and counterintuitive, but the experiments prove that it really happens that way.

    Another ultra-bizarre effect V.S. talks about in the BBC Reith Lectures (link in my previous comment) is one where a person can recognize the face of his mother, but is unable to emotionally accept the fact that it *is* his mother. Instead of accepting her, he accuses her of being an impostor (sadly distressing for both parties).

    The brain is truly a strange device :)

  • fh

    rusell before you become aware of the information to process it your mind has already done an enormous amount of preprocessing. As the quote says, the colors are associated to the shapes not the numbers. Thus when looking at the picture it will appear differently colored even without processing the fact that there are different numbers. The “preprocessor” has already associated colours to the different shapes, and then when picking out the different numbers the synesthets merely have to pick out the different colours as opposed to the different shapes, which, as you can see yourself in the provided images, is enormously more simple and quick.

  • Neil B.

    Here’s what Feynman is quoted as saying, from Wikipedia

    On the other hand, many synesthetes never realize that their experiences are in any way unusual or exceptional. For example, the Nobel prize winning physicist, Richard Feynman reports:

    “When I see equations, I see the letters in colors – I don’t know why. As I’m talking, I see vague pictures of Bessel functions from Jahnke and Emde’s book, with light-tan j’s, slightly violet-bluish n’s, and dark brown x’s flying around. And I wonder what the hell it must look like to the students.”

    Feynman, Richard. 1988. What Do You Care What Other People Think? New York: Norton. P. 59.

  • Abelian

    The beyond belief website is a pain to load

  • Tim

    While its a great story, the sad truth is it’s only a case study, and unfortunately the ‘pop-out’ search effect for synesthetes doesn’t replicate with 14 synesthetes (as I recall, none of them pass a well controlled version of the task). Its hard to imagine how it could, since, as a previous commenter mentioned they would have to be able to read the 2’s and 5’s as 2’s and 5’s before they would be colored — and there are thousands of experiments on visual search suggesting thats not possible.


    Edquist, J., Rich, A.N., Brinkman, C. & Mattingley, J.B. (2006). Do synaesthetic colours act as unique features in visual search? Cortex, 42, 222-231.

    (see also Mattingley, J.B., Rich, A.N., Yelland, G., & Bradshaw, J.L. (2001). Unconscious priming eliminates automatic binding of colour and alphanumeric form in synaesthesia. Nature, 410, 580-582.)

  • Veronica Gross

    The biggest issue is that, as Tim pointed out, this isn’t a replicable effect. There are a few reasons why:

    1) The visual search paradigm that produced those pretty greens and reds had a problem with viewing angle. From my discussion on the topic: However, when Laeng and colleagues performed a similar test with synesthete PM, they found that the speed of visual search seemed to depend more on the location of the stimulus in the array (Laeng, Svartdal, & Oelmann, 2004). That is, if the target were within 6° of the center of PM’s visual field, she performed better than the control participants, but if the target was in the periphery, her search performance was the same as that of the control participants.

    2) There seems to be two variants of color-graphemic (colored-letter) synesthetes. One variety, called “associators”, has their synesthetic percepts (experiences) in their “mind’s eye”. It’s more a visualization, colloquially (and maybe neurologically, depending on who you talk to). The other form of synesthete, or “projector”, has their synesthetic percepts out in perceptual space, where the percepts can interact with other visual phenomena by, for example, overlaying on a stimulus.

    *goes back to her visual crowding chapter*

  • Sean

    Hey, it’s great to have people who know what they are talking about chiming in. But I’m not sure that what Veronica and Tim say is convincing that the effect isn’t real. Certainly I can imagine that some part of my brain recognizes the shapes before “I” can do so — in other words, before the meaning of the shape recognition percolates up to my consciousness. (Not that I know what I’m talking about, but I can imagine it.) Likewise, the fact that the viewing angle was crucial seems to be an interesting feature of the effect, but not evidence that it’s not real.

  • Kevin

    That’s sort of what I was wondering… Is it that any object designated as “3” or “5” or whatever, is also designated green or red or whatever color, is it the concept of that number that creates teh color or is it that pattern of lines. for example if you were walking through the woods would a synasthete see giant red threes where tree branches intersect to form threes, or could the person make it appear red by looking at it and deciding yeah, that’s a three. Just curious if anybody knows

  • Jason Dick

    I was talking to one of the other graduate students here at UC Davis a little while back, and she mentioned that she was synesthetic. It was rather interesting to listen to her talk about how when she memorized telephone numbers, she sees them in her mind’s eye as being multi-colored blobs.

  • Count Iblis

    BBC’s Horizon had a program about this subject, see here

    There are also two online synesthesia tests that website.

  • Count Iblis

    Oops, a typo in the link. Let’s try again:

    This is the link

  • Neil B.

    Following up on Kevin’s comment: I’d like to ask some synesthetes what hybrid characters look like, that are intermediate between other ones. Also, if “five” looks green, what does “V” look like, if you know it is Roman numeral for five, etc.

  • George

    There’s more to this than meets the eye. I spotted the 2’s in less than 3 seconds, and some scanning was involved. I think this method of perception is camouflage or pattern decoding, spotting the leopard in the tree.

    Seeing the synesthete way, the red-green example, is a more holistic form of pattern recognition. This seems to be more like shape decoding, a red triangle on a green ground, less about 2 and 5.

    In either case, I suspect that separating the 2 from the 5 is simple pattern recognition, the number, or the color, or both, are associated after the fact.

    FWIW, seeing is all about the preprocessing, we don’t ‘see’ RGB or wavelengths and turn them into a color sensation. The brain ‘computes’ the ‘color’ from the environmental information. If you look out of a room, into daylight, your perception of the colors are relatively continuous and logical. Do the same thing with a camera, everything outside is bluer because the film or sensors respond to specific wavelengths. The eye-brain adjusts for those differences.

  • Tim

    Sean, the most basic problem is the effect doesn’t replicate, regardless of whether we have an intuitive theory for how it could or couldn’t work.

    The main claim, that the recognition happens “pre-attentively” (visual search researchers love to break things up into pre-attentive and post-attentive processing, something even I have done on occassion) and can therefore guide us in a search task, has been shown to be false in at least half a dozen studies I know of, including the two I cited in my previous post — one even published in Nature! Using several methods, from visual search to the attentional blink, suggests that people only experience the synathesia after they recognize and attend to the letters. It’s not as cool a story as Ramachandran’s and so it doesn’t get talked about much, but science is science.

  • Lewis Perdue

    The best popular exposition of this is “The Man Who Tasted Shapes” (Bradford Books) by Richard E. Cytowic.

    He’s also written the world’s only college level textbook on the subject, “Synesthesia: A Union of the Senses.”

    Cytowic does a pretty good job of fielding most of the hanging questions posed in the blog comments, above.

  • MedallionOfFerret

    There will likely be a LOT more confirming evidence before this natural child gets too enthused about synesthesia. Then again, even if it is a valid effect for some people, so what? The human brain has been known to be capable of wacky effects long before this particular finding.

    For the subjective record, I picked out all six of the 2s in about 3 seconds (totally subjective time measurement) and I ain’t never seen no colorified numbers, equations, or other mathematical representations in my 59.333 years of trying to interpret reality. I also have bad eyesight, and have never been known as especially perceptive. I tried it on a friend, timing him with my watch. He did it in well under four seconds. His history (I’ve known him for 40 years) is not that of a particularly perceptive genius either. If it’s taking other persons 20 seconds to find the six twos then we hereby announce ourselves as candidates for the Perception Hall of Fame, and freely offer our precious bodily fluids to all genius gene banks (at a ridiculously small fee; donations to individuals of the opposite sex, using traditional methods of implantation, may be even cheaper).

  • Sean

    Tim, that makes sense; the fact that the “pre-attentive” idea had been tested and ruled out wasn’t quite clear from the previous comments. That’s what I get for writing about stuff on which I’m not an expert. (Not that I’m likely to stop.)

  • Chris

    Yeah, Ramachandran’s not exactly known for producing studies that end up being replicated. Carefully controlled designs are not his forte (that would be publicity). But there is some really cool research on the psychological reality of synesthetic color (like Randy Blake’s).

    The problems for Ramachandran’s “model,” and his data, are multitude. Perhaps the most obvious is that it’s not a neurally plausible model. The “pop out” phenomenon that Ramachandran claims to have observed, and which is the hallmark of many visual search paradigms, is almost certainly occurring in the primary visual cortex (V1), as all of the single features that lead to “pop out” phenomena are processed in V1 (this includes color). So, “pop outs” are themselves thought to be V1 phenomena. However, the potential color-grapheme connections are way up in V4. By that time, you’ve either seen the differences, or you haven’t, and all sorts of more complex attentional issues come into play (David Gilden’s done some cool work on this stuff). In other words, if you don’t get the color differences in early visual processing, you’re not going to get them, and since the grapheme processing comes later than V1 anyway, you might as well just go on that.

    Put yet another way, synesthetic colors (for color-grapheme synesthetes, at least), aren’t individual features separable from the letters themselves. Instead, they’re inseparable feature conjunctions, so that if you don’t recognize the grapheme, you don’t recognize the color, and since the processing of the grapheme comes pretty late in visual processing, if you recognize the grapheme, you don’t need to recognize the color to perform the visual search task. This is, I believe, the conclusion of the Equist et al. paper that someone referred to earlier (the one with 14 synesthetes). Ironically, if Ramachandran were paying attention to the data from “color-blind” synesthetes, he’d have realized this too, because it implies that the color-grapheme connections occur way too late to benefit visual search.

  • Chris

    P.S. Did Scott Atran’s talk rock as hard as it did last year?

  • Count Iblis

    If there are difficulties with these test, why not do more sophisticated tests like PET scans where you can really see if the people really see colored letters instead of black and white letters?

  • Chris

    If there are difficulties with these test, why not do more sophisticated tests like PET scans where you can really see if the people really see colored letters instead of black and white letters?

    That it happens in the brain ;).

    Essentially, imaging studies have shown that the color-processing areas of the higher-visual cortex (specifically, v4, and more specifically, v4/v8) are active when you present graphemes that are associated with a color experience.

  • andy.s

    Yes, but can they see the color Octarine?

  • Sean

    Scott Atran’s talk came in two parts — the first part, where he was continuing some sort of debate from Sam Harris from last year, was kind of meandering. The second part, where he discussed his research into the origins of terrorism, was amazing, probably the best talk of the conference. If I can find some good sources about what he was talking about I’ll try to blog about it.

  • Pingback: Winter on its way… « blueollie()

  • Veronica Gross

    That’s sort of what I was wondering… Is it that any object designated as “3? or “5? or whatever, is also designated green or red or whatever color, is it the concept of that number that creates teh color or is it that pattern of lines.


    Following up on Kevin’s comment: I’d like to ask some synesthetes what hybrid characters look like, that are intermediate between other ones. Also, if “five” looks green, what does “V” look like, if you know it is Roman numeral for five, etc.

    are both very good, useful questions.

    The two hybrid experiments that I know of, the so-called ambiguous grapheme experiments (Dixon et al., 2006; Myles et al., 2003) say that context largely dictates the color of the ambiguous grapheme. So if you stick a hybrid between a letter and a number in a list of letters, it’ll be parsed as a letter and come out as that letter’s color. Ditto for numbers. Of course, we have a small n in both cases, but given how preferential words are for our brains, it’s not terribly surprising.

    Now, some people say that concept is not enough. So seeing an intersection of three branches might not trigger “three” without a visual representation of the number three. That’s where you start running into the questions about whether synesthesia is top-down or bottom-up. Bottom-up means that synesthesia is a response to the physical shape of the stimulus. So someone who saw “V” the roman numeral would parse it the same as “V” the letter. Top-down means that the context and meaning are important, so “V” the roman numeral would be parsed as the number 5, while “V” the letter would be a letter.

    Problem is that half the synesthete I’ve chatted with say the first and half say the second. The best we can say at this point is that context is important and is useful, but it may not be sufficient.

    (And they can only see octarine after 4-5 years at Unseen.)

  • Low Math, Meekly Interacting

    One of the most remarkable syneasthetes is Daniel Tammet:

    It’s thought that his savant abilities are facilitated by the associations synaesthesia provides. When I first heard about Tammet, I immediately thought of the Feynman quote in #8, and wondered if he also benefitted from this quirk of neurology.

  • Pingback: » Blog Archive » Two Things That Blew My Mind on Monday()

  • Tim

    Sean, I’ll put my vote in for a post on Scott Atran’s talk. His recent work on terrorism seems really important and interesting. He had a long Nature or Science article about it not long ago (a couple of months?), too.

  • Chris
  • Pingback: Verschiedenen Geheimnissen | THE JO-TEL … con la espada ensangrentada()

  • Greg

    People never believe me when I tell them numbers and letters have colours. I was excited to test how strong my connection is, but I glanced at the colourised version of the graphic before the black and white one, which spoiled the test. I often remember things like phone numbers, exam grades, and website addresses are based on their colours.

  • Greg

    Oh, but what happens if 2 and 5 are the same colour? For me they’re quite close—orange and red respectively. The test would still work, but with false negatives.

    I laid out my colours a few months ago:

  • F1

    I’m both a synaesthete and an astrophysicist.

    Likewise, in the past I was always accused of lying when I said my 3’s were red. I didn’t know that others didn’t see the world this way, or that there was a name for the “condition,” until about November of 1999 when Discover Magazine published a popular article on the subject.

    I’ve heard that synaesthesia is not all that uncommon, but I’ve never met another synaesthete to my knowledge. Admittedly it’s not something that comes up in conversation due to the oddity and the potential embarrassment factor.

    I can address a few points from my point of view:

    1. Yes, I see mathematical expositions in a flurry of colors.

    2. For color-grapheme associations, context does matter, to a certain extent. For example, the color of a character will change depending on whether I think it’s a number 5 or a letter S. I can force a character to “change” colors.

    2a. For some reason this doesn’t apply to Roman numerals, which for me always retain the colors associated with their “alphabetness.”

    3. Despite point 2, I have noticed that I can sometimes read highway signs from a distance based on the colors I see even if the characters are a bit too blurry for me to consciously make them out. But perception of perception is fuzzy — my brain must be recognizing the graphemes somehow even if I don’t.

    4. I occasionally read things written in non-Roman alphabets. The colors of the characters in these non-Roman alphabets are frequently colored by their corresponding sound, and thus letter, in the Roman alphabet. Amusingly, these non-Roman characters may appear “two-toned” if they carry one sound/color in the corresponding Roman character but also physically resemble another of the non-Roman characters that carries a different sound/color.

    5. The triangle of 2’s doesn’t pop out immediately for me, although certainly more quickly than the 20-second average alleged for the non-synaesthetes. I have to recognize the 2’s first.

    I would like to know whether color-grapheme synaesthesia in particular is more common among people who learned to read comparatively early. (This would probably have to be disentangled somehow from the tendency of people with higher-than-average IQs to learn to read early; I recall hearing that higher-than-average IQ is associated with synaesthesia, as is autism.) My understanding is that infants don’t prune away these neurological cross-modal associations until age 2 or 3.

  • F1

    P.S. Greg, your colors are wrong. 😉

  • Pingback: FUTURE BLINDNESS » Blog Archive » links for 2007-11-09()

  • Pingback: the new shelton wet/dry()

  • Pingback: » Colorblind Synesthete experiences colors his eyes can’t see.()

  • Pingback: Farver fra Mars « Lusepuster()

  • Michael G.R.

    Interesting! Last september I wrote about my experience with synesthesia:

    I even used pictures of the Rama test.

  • Pingback: My Experience With Synesthesia « Michael Graham Richard()

  • liana

    If you think you might be a synesthete, you can check the possibility by doing a computerized battery of tests at this website:

    Personally, I would recommend synesthetes to contact synesthesia research centers at near universities. There are many forms of synesthesia out there (combinations of color, emotions, sequences, spatial locations, graphemes, sensations, sounds, etc.), and usually the researchers need new participants for their experiments. It can be an interesting experience and you’ll be contributing to science.


  • Pingback: links for 2007-11-14 « Amy G. Dala()

  • Pingback: Michi’s blog » Blog Archive » Synaesthesia and cognition()

  • Pingback: Reading Backlog | THE JO-TEL … with George Orwell as Megatron()

  • Anna

    Hey synesthetes! I just found great online synesthesia survey – you can vote what color is letter ‘A’ and each letter or number in your mind and see what color many other people voted for! Really interesting:

  • Stacy

    “…. so that if you don’t recognize the grapheme, you don’t recognize the color, and since the processing of the grapheme comes pretty late in visual processing, if you recognize the grapheme, you don’t need to recognize the color to perform the visual search task. This is, I believe, the conclusion of the Equist et al. paper that someone referred to earlier (the one with 14 synesthetes). Ironically, if Ramachandran were paying attention to the data from “color-blind” synesthetes, he’d have realized this too, because it implies that the color-grapheme connections occur way too late to benefit visual search.”

    Chris, have you had a look into modified stroop tests that reveal that a synestheste can read a grapheme faster if it is presented in a congruent colour as compared to a noncongruent colour?


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Cosmic Variance

Random samplings from a universe of ideas.

About Sean Carroll

Sean Carroll is a Senior Research Associate in the Department of Physics at the California Institute of Technology. His research interests include theoretical aspects of cosmology, field theory, and gravitation. His most recent book is The Particle at the End of the Universe, about the Large Hadron Collider and the search for the Higgs boson. Here are some of his favorite blog posts, home page, and email: carroll [at] .


See More

Collapse bottom bar