name that smell

January 3, 2014 5:22 PM

Smells can be very hard to identify and name, unless you are given some prompting - or you speak Jahai, the language of an indigenous group in the Malay peninsula.
posted by divabat (23 comments total)

This comes to mind.
posted by thewumpusisdead at 5:40 PM on January 3, 2014 [1 favorite]

I wonder if that would change one's fundamental relationship to the experience of smell? I wonder if our sense that smell is uniquely powerful as an evoker of memory, for example, has to do with our inability to think of smells in abstract terms. If you smell something and think "that's just like grandma's attic" you're taken to a concrete place and time(s), but if you think "that's a blippid smell" you're moving up to a level of abstraction which places all the blippid smells in the universe in the same box.
posted by yoink at 5:43 PM on January 3, 2014 [7 favorites]

English speakers used mostly source-based descriptions (like a banana).

I don't see how saying "that smells like wet concrete" or "that smells floral" renders us less capable of describing a smell than saying "that smells brillig" or "that smells slithy." "Like wet concrete" might be three different words but they only have one meaning, same as a single word.
posted by BuddhaInABucket at 5:50 PM on January 3, 2014 [3 favorites]

I don't see how saying "that smells like wet concrete" or "that smells floral" renders us less capable of describing a smell than saying "that smells brillig" or "that smells slithy."

The color of the sunset.
The color of a firetruck.
The color of a Spartan apple.
Red.

Generic terminology is powerful. Anybody know where we could find a Jahai lexicon?
posted by justsomebodythatyouusedtoknow at 5:56 PM on January 3, 2014 [2 favorites]

I guess I don't think having abstract words to describe smell is particularly important. Minty, orangey, like concrete, skunky- good enough for me! Somebody get a purfumer in here for an expert opinion!
posted by BuddhaInABucket at 6:00 PM on January 3, 2014

I suppose the question when it comes to using metaphors to describe smells is "what sort of X? The same X I'm thinking of?"

Resources about the Jahai/Jehai language.
posted by divabat at 6:02 PM on January 3, 2014 [1 favorite]

Floral, spicy, sweet, putrid, smokey, rotten, fruity -- we have direct words for describing smells.
posted by Chocolate Pickle at 6:12 PM on January 3, 2014 [1 favorite]

That was fascinating. Naming things helps make them more concrete, and helps us understand things better.

Majid and Burenhult found that Jahai speakers could name odors with the same conciseness and level of agreement as colors, but English speakers struggled to name odors.

We can say that we have words in English, or that abstract words are good enough ... but the studies seem to indicate that the Jahai have a common nomenclature that we really do lack; otherwise there wouldn't have been that much of a difference in the tests.
posted by kanewai at 6:23 PM on January 3, 2014 [3 favorites]

I'm a coffee roaster, and since we evaluate coffee a great deal by aroma, I can attest to how frustrating it is that this vocabulary and agreement is totally lacking in English.

This would make my job much easier. What a fantastic read.
posted by furnace.heart at 6:31 PM on January 3, 2014 [2 favorites]

I don't see how saying "that smells like wet concrete" or "that smells floral" renders us less capable of describing a smell than saying "that smells brillig" or "that smells slithy."

The question isn't whether the English language is capable of describing smells. It's whether English speakers asked to identify an odor can, and whether there is an influence of culture or language on the ability.

I can't access the article, because it's behind a paywall and for some reason my institution doesn't have access to it yet, but here's the abstract:

From Plato to Pinker there has been the common belief that the experience of a smell is impossible to put into words. Decades of studies have confirmed this observation. But the studies to date have focused on participants from urbanized Western societies. Cross-cultural research suggests that there may be other cultures where odors play a larger role. The Jahai of the Malay Peninsula are one such group. We tested whether Jahai speakers could name smells as easily as colors in comparison to a matched English group. Using a free naming task we show on three different measures that Jahai speakers find it as easy to name odors as colors, whereas English speakers struggle with odor naming. Our findings show that the long-held assumption that people are bad at naming smells is not universally true. Odors are expressible in language, as long as you speak the right language.

It is impossible to evaluate the research without being able to read the article and it would be irresponsible to try based on a skimpy popular science article. However, it is clear that the authors aren't saying that the English language lacks the ability to describe the source or sensory experience of smells, although there are important differences in the English and Jehai olfactory lexicons.

One thing that is frequently a problem with such research is that it's often impossible to untangle culture and language, yet people (not always the researchers themselves) often want to conclude that the results show language influence cognition.
posted by Kutsuwamushi at 6:33 PM on January 3, 2014 [3 favorites]

I've found a copy of the original paper. From the paper:

So although a verb like ltpɨt is prototypically used to describe the fragrant odors of flowers (e.g. Globba, Lantana spp.) and perfumes, any source whose odor approximates such a quality can be described with the same verb, e.g. the fur of the bearcat (a hunted civet species, Arctitis binturong, whose musk glands emit an odor reminiscent of popcorn).

What sets Jahai color terms apart from English seems to be abstraction. The English "floral" applies to flowers and things that smell like flowers. ltpɨt applies to some things which do not smell at all like flowers. Apparently there's something that flowers and civet pelts have in common, but I don't even know how to imagine what that might be.
posted by justsomebodythatyouusedtoknow at 7:03 PM on January 3, 2014 [1 favorite]

I wonder if we would make more progress by abstracting this in the direction of chemistry. We could identify or group sets of smells by chemical family or by functional group, and be reasonably sure we're talking about the same thing when we learn the name of an odour. Then we wouldn't have to use the word "like", or worry about whether we're talking about flowers or civet pelts. Of course the set of standard smells would be kept in a special exhibit where children would go on field trips to be introduced to the lexicon. "To the Smellatorium!" [/Farnsworth]

In fact we already do something similar with taste and odor panels for detecting odorants in drinking water (Strategies for Taste-and-Odor Testing Methods). Interestingly, this can be more reliable than gas chromatography for detecting the exact contaminant that's causing a problem - partly because the GC can't tell you what a particular chemical tastes or smells like, and because sensitive humans are able to distinguish a variety of chemicals by taste and smell. Not only that, but you can earn money by demonstrating your ability to distinguish tastes and smells.

Here's another article about "Flavor's Language Problem".
posted by sneebler at 8:51 PM on January 3, 2014 [1 favorite]

One possible issue with this research is that we don't know that the task that subjects are engaging in is the same across languages. In English, for example, the question "what smell is this?" is not felicitously answered by "acrid" or "fruity" or whatever -- the only proper answer is the source of the smell. Inversely, "what color is this?" is not felicitously answered by "fire truck" -- the answer is red. In English, these are different tasks (odor = what object exactly produces this stimulus; color = what classificatory group does this stimulus belong to, among a more-or-less canonical set).

This is, evidently, not true of whatever Jahai translation of these questions were used; both are answerable by names. There is no evidence that Jahai speakers are better able to distinguish perceptually similar odors (like lemon/orange) without supplemental cues, which is what English speakers try to do in this task.

In fact, the assumed success criteria for the English speakers is a bit odd. From sec. 4 of the original paper, here are some examples of stimulus odors and incorrect responses from prior studies of English speakers; I've underlined the ones that are in fact arguably correct:

• lemon: air freshener, bathroom freshener, berry, magic marker, candy, citron / citrus, cleaner, Lemon-fresh, Pledge, hard candy, like pine, orange, some kind of fruit
• cinnamon: spicy, sweet, bayberry, candy, Red Hot, smoky, edible, wine, potpourri

The odors in the test are artificial (almost certainly; I don't know about any particular instance, but the most common format for this test is a scratch-and-sniff booklet). So there's an argument to be made that "lemon Pledge" and "potpourri" are actually better answers than "lemon (fruit)" and "cinnamon"!

English speakers have created the wine aroma wheel to categorize odors (within the domain of wine tasting). This research could suggest that Jahai culture has made such a taxonomy easily accessible to everyone, not just wine specialists, but not more.
posted by dendrochronologizer at 9:03 PM on January 3, 2014 [2 favorites]

sneebler: In Chandler Burr's The Emperor of Scent, Luca Turin does what you're after - he abstracts scent by chemistry. In the book he is pretty adamant that there are very specific ways to name scents: specific molecules can only smell like very specific things. Later on in the book he encounters people who question him with "how do we know they're smelling the same thing we are" but he refutes those challenges.

There's also the fragrance wheel though I'm not sure how that applies to anything that's not perfume.
posted by divabat at 9:13 PM on January 3, 2014 [1 favorite]

The way we describe a color, as a point in a 3D color space, seems to me directly connected to the three kinds of cones. It makes sense to say that one color is halfway between two others because, in this continuous 3D space, it is. The three cones give a basis for describing the whole space.

I don't think we can do that with smell. It doesn't make sense. Smells aren't a continuous space. "Cedar" is the sensation of one kind of molecule hitting your olfactory receptor neurons, and "ozone" is a different molecule, and if we wanted to describe the space of possible smells, we would need an axis for every kind of molecule. So maybe that's why we see blue and red together and call it purple, but when we smell snow and pine we call it "snow and pine". We could create smell-specific words and use them instead of "snow" and "pine", but it fundamentally isn't going to be the same as how we describe color.
posted by qxntpqbbbqxl at 9:46 PM on January 3, 2014

qxntpqbbbqxl: there's been a lot of scientific debate about how humans smell - there's Shape theory, then Luca Turin comes in with the controversial Vibration theory, and there's weak shape theory.
posted by divabat at 12:24 AM on January 4, 2014 [1 favorite]

Thank you for posting this divabat.
I think Kutsuwamushi nails a big part of the problem in that The question isn't whether the English language is capable of describing smells. It's whether English speakers asked to identify an odor can, and whether there is an influence of culture or language on the ability.
qxntpqbbbqxl goes on to state ''Smells aren't a continuous space''. I and I believe, people in the perfumary trade disagree. The smells blend; our descriptive powers of language here do not or, if they do only within a very small subset.
posted by adamvasco at 6:34 AM on January 4, 2014

I'm not able to find it at the moment, but there was an article linked on the Blue that described an Amazonian tribe (I think) that did not have the adjectives for colours that most other languages have. So they described colours "like the sky", "like berries" etc. In other words the opposite of this language.
posted by Harald74 at 7:42 AM on January 4, 2014

when we smell snow and pine we call it "snow and pine".

Then again, it's not uncommon to see blue and green together and call it blue green.
posted by scrowdid at 8:03 AM on January 4, 2014

Blue green might be common but it is imprecise.
See here these myriad shades of green all named; and this is what is lacking in the English language in relation to smell; a codification.
posted by adamvasco at 8:17 AM on January 4, 2014

Thanks, a fascinating read.

> I don't see how saying "that smells like wet concrete" or "that smells floral" renders us less capable of describing a smell than saying "that smells brillig" or "that smells slithy."

> I guess I don't think having abstract words to describe smell is particularly important. Minty, orangey, like concrete, skunky- good enough for me!

> Floral, spicy, sweet, putrid, smokey, rotten, fruity -- we have direct words for describing smells.

It's depressing to me that so many MeFites can't read an article (if in fact they even bother to RTFA) without immediately assuming it's full of shit because it's saying something they're not already used to thinking. "Screw that, man, I can describe smells!" Would it hurt to take a minute to assume for the sake of argument that the authors might actually have a useful point rathr than jumping straight to the snark?
posted by languagehat at 8:39 AM on January 4, 2014 [5 favorites]

This measures how much agreement there is in the descriptions people give, but it doesn't measure how good people are actually communicating differences using those descriptions.

For example, let's say in one language ('web') people named colors as rgb triplets, and in another ('English') people named them as color words like we know. If web speakers were prompted to name a color, you would see a large number of labels for the same color (one person might put [235,10,20] and another [233,8,22] for the same color), while English speakers would probably just call it 'red'. So you would see a lot more apparent agreement between English speakers despite the fact that experienced web speakers are likely better at communicating colors using rgb triplets.

If you want to measure actual descriptive ability, you need a different experiment: give a color/smell (call it sample A) to one subject and have them describe it, then give their description to another subject and have them choose which color/smell sample A or B best matches the description.
posted by Pyry at 10:27 AM on January 4, 2014 [1 favorite]

I'd like to point out that if you go back to old chemistry manuals English actually *has* a bunch of terms to describe smells fairly precisely, they've just fallen out out of favour, even with chemists, as taking large sniffs of your chemicals is not considered a good idea anymore.
posted by Canageek at 2:15 PM on January 4, 2014