two

@two

actually the number two IRL

Thanks for playing, everyone. I'll see you around.

log inask
two
@two

absolute and relative senses

This is something that has weirded me out for a while: why is colour described in absolute terms but smell in relative terms? Like, red and green and blue refer to exact sensory inputs (wavelengths of light), not some prototypical "red" thing which we say things that are red look like ("orange" is an outlier). But with smell there's none of that, it's all "smells like banana" and "smells like nail polish remover", there don't seem to be accepted terms for exact smells. There's a similar thing going on with taste, though at least there you have the five primary tastes (sour, sweet, bitter, salty, umami) even if there's some degree of other stuff going on (the enduring popularity of the phrase "tastes like chicken" proves that taste is not just about the five main tastes). Is this just there not being enough good words or are smell and taste (and now that I'm thinking about it, I guess touch) somehow more complex senses than vision is? Personally I find it a lot harder to imagine tasting a taste or smelling a smell than seeing a colour, so I want to propose that it's the latter, but maybe that's just me? Actually, wait a minute-

Wikipedia screenshot listing 7 "primary odors", with a red link.

Somebody needs to write the Wikipedia article for primary odors, please.

(I'm not including sound in this analysis because all the sound names are just onomatopoeia and I have no complaints there, that makes sense.)

You must log in to comment.

in reply to @two's post:

quewon
@quewon

woahh. "musky" being a primary odor makes me think smell isn't just flavors, it's also depth.

unlike you i find it very easy to imagine smelling a smell, probably even more than colors. reading the description for "ethereal" has me smelling laundry detergent right now. (i would say i'm imagining something more floral though. what is an ethereal scent...?)

login to reply
two
@two

I love that "ethereal" is up there as a primary odour. Such an evocative word and I cannot find a description of the smell that makes sense to me (I can't say I remember having ever smelled dry cleaning fluid or chloroform, and I can't work out what sort of smell element is supposed to be common between all fruit and wine?)

login to reply
quewon
@quewon

i just looked it up and apparently ethereal is also a chemical term :O making sense words based on chemicals makes a lot of sense to me, lol. but sense words have always been super vague. (as for fruit and wine, there is such thing as a sweet smell, maybe that's the common element?)

login to reply
joXn
@joXn

One short answer is that there are upwards of 400 distinct and distinguishable chemical receptors in the olfactory bulb that, when stimulated, can produce the perception of a smell, and there are basically no “primary” smell substances—almost everything one can smell activates more than one of those smell receptors (to differing degrees).

That said, there are some languages which do have a broad vocabulary list of “basic smells”, c.f. the Maniq language, the Jahai language, but I believe at least in Jahai this is because the chosen categories of smells have spiritual significance, not because they necessarily form a good basis set for the space of odors.

One more thing that might be of interest to you is this table of organic compounds and their smells, which is interesting because there are some clear regularities in the data but also some absolute chaos.

login to reply
joXn
@joXn

I guess I should mention that humans can make a pairwise distinction between some huge number of distinct odiferous compounds (the number is estimated to be in the trillions). This means that human sense of smell is very good at accurately determining “yes, these two smells are the same” / “no they are not the same” when presented two smells in a pairwise discrimination test. But that capability seems to hit limits when, e.g., asked to partition even, like, a dozen similar but slightly different smells into “same”/“different” groups.

login to reply
arborelia
@arborelia

This is very informative (and that table of organic compounds is something that I love). I just want to quibble that we can name "primary" sensory inputs without them activating exactly one receptor in the human nervous system, and colors are an example.

If you've got three or more sets of cones (you're not color-blind), there isn't a color that activates only the medium-wavelength ones. We just pick basis sets that activate them in different combinations and call them "primary colors", and these sets don't even have to be consistent with each other.

So it would be possible to do the same thing with smell, and the languages that name basic smells have a good claim to have done so, even if they don't cover the space fully.

login to reply
joXn
@joXn

An RGB basis is non-unique for color vision for the reasons you say: the short, medium, and long wavelength cones don’t have point-responses, they have overlapping curves. And perceptually, RGB is a particularly poor choice because the opponent process neurons sit in between the cones and the brain; probably the right basis if we care about what it’s like to see colors is blue-yellow, red-green, and luminance. (And, as a bonus, the opponent process does an effective job of reducing the ambiguity of the the overlapping response curves.)

I think the biggest problem with theorizing about basis vectors for a space of smell is that the state of the neuroscience on smell is really rudimentary. There has been a lot of research on sight and touch; but hearing, taste, and smell still have huge areas where the map is labeled “here there be dragons”.

Here’s one question that I’ve never seen addressed about smell that bears directly on the topic: are the smell-concepts in Jahai the kind of linguistic construct that adult second-language learners have a very difficult time acquiring for what are essentially brain-developmental reasons, similarly to how new phonetic inventory becomes difficult to acquire after the critical language acquisition period? If it were, that would be an interesting data point!

login to reply
two
@two

That table (along with the one for esters, which is significantly less chaotic) is one of my all-time favourite infographics. I love the question mark cells, presumably for chemicals that might have a smell but aren't recorded anywhere in the literature, implying that somebody should release a paper where they smell and describe every previously undescribed chemical smell.

login to reply
ratherforky
@ratherforky

I think you're probably right that complexity has a lot to do with it. Colours can inherently be decomposed into a small number of components and they're continuous/there's a lot of overlap, so i could probably describe any colour in a "it's kind of a blue-y green..." way and at least get in the right ballpark, but idk how to decompose "banana" any further. Every time I've experienced a banana smell there's been actual bananas involved or a deliberate attempt to evoke bananas. It doesn't exist on any fruit continuum in my mind, it needs its own banana dimension

login to reply
modulusshift
@modulusshift

I have some thoughts!

A possible reason that taste isn’t reducible to primaries is because it tags in your sense of smell for an additional dimension. Your mouth is designed to shoot air that’s been mixed with your food into your nose. That’s why the same food can taste wildly different based on location, to some extent you’re also tasting the air you’re breathing. You’ll hear “food tastes better when you’re camping”, or “airplane food sucks” (that’s a function of low air pressure reducing your sense of smell, presumably, although I’ve never experienced an airplane food that was trying that hard, personally.) I also wonder if it’s why high altitude places rarely have a unique cuisine.

Anyway, I think there’s a few more factors: smells are composable. You can smell two things at once and identify them separately^1. You can’t do this with light, there’s always only one color coming from a certain spot, two light sources will blend together into a third light source that doesn’t match either. You can do this with sound to some extent, but with effort two sounds can be blended or cancel each other out. Smell doesn’t even have that. There is no blending, really, you could theoretically break down any smell into components.

But those components number in the hundreds! We have at least 400 genes encoding for distinct proteins that act as scent receptors. I don’t even think I could remember 400 distinct smells. And indeed probably some of them respond to similar compounds but in different amounts. Even with similar receptors working together, which we don’t have a ton of evidence for or against, that’s still over a hundred primary smells, from 400 candidates. This is not trichromatic (maybe tetrachromatic for some people) vision, or a clear spectrum like sound is. This is the relative levels of over a hundred independent chemicals all being sampled at once. Smell is a ridiculously complicated sense.

^1. Air fresheners don’t really work by providing an alternate scent, they try and capture the existing one, for sprays small droplets are good at catching small molecules and dropping them out of the air, or the sticky ones literally just catch things that pass by. The smell they provide is just a bonus and maybe trying to mask some unpleasant chemicals that help with the smell catching.

login to reply
joXn
@joXn

Taste is more complicated than just taste buds + olfactory bulb; also involved are chemoreceptors (such as those which respond to capsaicin and menthols); mechanoreceptors (which provide a textural component); and thermoreceptors (physical hot and cold).

I'm sympathetic to people who are skeptical about the last two, but I think it would be really perverse to argue that "spicy" and "minty" are not taste sensations. I'm also sympathetic to people who argue that "taste" per se is just what is sensed by the taste buds, and that the fused perception that brings in smell and chemoreceptors should be called "flavour". But that distinction is technical terminology; I don't think that's how most people commonly use the words "taste" and "flavour"!

My favourite fact about taste is that we know what protons taste like. Excess H+ ions in a solution make it acidic, and acids are sour.

login to reply
jcolag
@jcolag

I've only given this a bit of thought over the years, and not in this way, but I'd argue that it's about experience. Specifically, we don't grow up using our sense of smell to any extent, but we constantly orient ourselves and communicate visually.

I suggest this because, when you push people past their experience on color, you do get exactly that kind of behavior. A typical example comes in spaces like web design, where some manager will ask if the logo can be "more blue," and not have any more substantive criteria than that. They might have a specific shade of blue in mind, but more likely, they want it clearer or less jarring, and they have no idea how to talk about that.

Unrelated, the idea of "primary odors" reminds me of RealAroma, and its assortment of references to products that nobody would admit to recognizing today...

login to reply