Why we suck at writing tasting notes: new research

An article in the January 26 issue of the Economist (may require subscription) may go a long way to explaining what goes on here at WB day in, day out. It’s about a study on humans’ ability to describe smells. The bottom line: Most of us suck at it.

The study, by Asifa Majid at Radboud University in the Netherlands and Nicole Kruspe at Lund University in Sweden, compared members of the Jahai tribe in Malaysia to a group of Americans.

[When Majir] asked some Jahai, and also a comparable group of American volunteers, to name colours and odours they were presented with, the Americans generally agreed with one another when it came to naming colours but agreed much less when putting names to odours. When presented with cinnamon, for example, they described it variously as sweet, spicy, wine, candy, edible and potpourri. When presented with baby powder they offered vanilla, wax, baby oil, toilet paper, dentist office, hand lotion, rose and bubble gum as descriptions. Jahai answers, in contrast, were in equal agreement about both odours and colours.

Sound familiar? (FYI, a Dutch control group in an earlier study did badly, too.)

The study suggested the disparity in the two groups might be due to the vocabularies of their respective languages, which in turn may reflect the difference in ways of life:

… because the Jahai have a dozen words dedicated to describing different sorts of smells in the abstract (the equivalent of colour-words such as red, blue, black and white, of which there are generally reckoned to be 11 in English). For example, the Jahai use the word “cŋεs” for stinging sorts of smells associated with petrol, smoke and various insects, and “plʔeŋ” for bloody, fishy and meaty sorts of smells. According to Dr Majid, only “musty” is able to act in this way in English without drawing on analogy (banana-like, gooseberry-noted, and even earthy and sweet-smelling, are all analogies of some sort).

The speculation is that hunter-gatherers have to be more attuned to smells in order to survive, and thus have a richer, more differentiated way of describing them than those of us who are more “civilized.” As The Economist put it in a 2015 story about a prior study by Majid:

This finding challenges the long-standing idea that something about the way human brains are wired limits their ability to put words to smells swiftly. Dr Majid says that in the West, “nothing should have a smell unless you put it there,” and that there are many taboos surrounding talking about odours. The smell-rich environment, scent-centric cultural practices and evident lack of taboos enjoyed by the Jahai and Maniq are, she believes, correlated with their recall and use of smell-related words.

It squares with my experience that people in the wine trade who have tasted around other, experienced people are often much better at describing things precisely (they say things to which I say, “That’s it! I couldn’t put my finger on it”) and also tend to concur more on what wines taste and smell like. In other words, you can with training/experience improve your ability to describe smells in a way that will agree with other experienced people.

This just goes to show how divided we are as Americans: [cheers.gif]
Can’t even agree on what ‘cinnamon’ is!

Wait, if they are more attuned to smells, then why do they use just one word for ‘smells associated with petrol, smoke and various insects’ or ‘bloody, fishy and meaty?’

Our language is very redundant, so I would expect more variation in vocabulary.

Although, ‘candy’ is a terrible descriptor. For cinnamon, unless candy - Big Red or Hot Tamales, it is not too great a word choice.

It seems we loose specificity via having too many words, and they lose specificity by having too few.

As someone who sucks at writing tasting notes, I’m glad to hear I can offer a scientific excuse for it.

i think vagueness of language poses a larger problem

It also would have a lot to do with artificial flavourings and aromas being the prime experience in the first world, such that sometimes caricatured artificial examples become our benchmark.

Vanilla is a great example of being oft-caricatured, and I suspect our exposure to the supposed flavour or scent of vanilla, far outweighs our exposure to vanilla pods themselves.

I agree, I think that this is a big part of it. Offer a test with a 200+ multiple choice answers, results will most likely over the map. Constrain it to 10, as in we only had 10 words to describe something, the variance becomes quite a narrow swim lane.

This highlights the healthy redundancy within our language.

Biology may or may not play a part in it, however it still has to filter through to a matter of verbal and/or written expression and that is a product of their social environment.

As I read the article, I don’t think it’s that this tribe is incapable of giving specific descriptions. It’s just that, in addition to specifics, they have abstract smell terms (e.g., “musty” in English). You can have those as well as terms that refer to a specific smells based on objects/substances (cinnamon, apple, oak).

That difference can explain in part why two tasters describe the same wine so differently. Apart from individual differences (and general lameness), people are reaching for specifics (charred alder wood versus charred apple wood, per James Molesworth, pain grille, per Parker), that pretty much ensure there will be no correlation between tasters’ terms.

In a way, I guess the Davis Aroma Wheel attempts to create abstract terms by grouping specifics:

This is why we rely on points from paid (sometimes compromised) experts.

I don’t think it’s that we have too many words to choose from. That doesn’t explain why the American control group couldn’t peg cinnamon, which is mind-boggling to me. (While English has a very large vocabulary measured by dictionary size, the actually range of words that people use isn’t that different from other languages spoken by literate populations, as I recall.)

Smelling/tasting blindly, I would have thought pretty much anyone could identify, say, a lemon versus a lime, an orange versus a grapefruit, or lamb and beef, or tomato versus a cucumber, or a peanut and an almond. But maybe not. Maybe you have to be attuned to tastes and smells, as wine lovers and good cooks generally are.

Of course, task of description with wine is much harder because the fermentation and aging processes make it very complex chemically, and few people learn to describe it from others who are really knowledgeable.

I don’t think the people who frequent this board could, as a group, possibly qualify as a random pool from which to run such a test. Beyond natural ability and anatomical differences, which of course has a role to play, I think there is something to a hobby where we make an effort to associate aromas and flavors in wine with comparative aromas and flavors. We are a product of the environment we have chosen, to be wine drinkers who are focused over years on what we smell and taste. So it’s hard to know how to relate a study on a random sample to this community. Maybe Nathan can help as a professional statistician.

I agree were not normal [sic]. But the lack of correlation in descriptors among critics, and between WBers, certainly is consistent with the researchers.

I make people blind taste Skittles.

If they can’t identify Skittles color, we’re through.

Man, you’re rigorous!

It is shocking what happens when people put a blind fold on and taste. A guilty pleasure was Hell’s Kitchen for the Gordon Ramsey rants. And when people on that show were blind and tasting the simplest of things, I would be florred by their answers. Obviously, there’s production value and they aren’t exactly world class, classically trained chefs per say, but still shocking. You’d think they’d be more attuned to your point.

I probably could have been a bit more specific there myself there. While I still hold we do burden our language with too many words, there’s also how we associate those words to what we perceive in our environment. People generally develop their own loose applications of them as descriptors, and then cognitive and/or socially apply them. Or at least that was my first impression from what was quoted…

When presented with cinnamon, for example, they described it variously as > sweet, spicy, wine, candy, edible and potpourri> . When presented with baby powder they offered > vanilla, wax, baby oil, toilet paper, dentist office, hand lotion, rose and bubble gum as descriptions

Potpourri? Hand Lotion? Dentist office has to be my favorite up there. There may be a general construct however its hardly a precise description lol. The taste, or smell just triggered those words as the strongest association for what they perceived. And perhaps, if we had a narrower scope of terms, it would have conformed more perhaps. I doubt the tribe has a Dentist Office either as we know it :wink:

Many ways to get there, however your last point is kinda where this ends for me with wine. People learn and people respond to normative influences. Just being with a part of group of experts will influence their choice of descriptors while honing their technique (or perception) for coming to those descriptors.


Then throw them off with an M&M here and there.

Or find a skittles ringer in Ian’s Ringer thread.

Skittles weren’t invented yet when I was a kid, but we used to try to guess the color of Good 'N Plenty tasted blind. Much better odds with only 2 colors.

John, I think you are really on to something. You only truly learn something in minute detail if it really matters intensely to you in some way. if your life depended upon your sense of smell, or if you would starve to death, or freeze to death, or poison yourself without a reliable sense of smell, you bet you’d learn to differentiate key things in a reliable and reproducible way.

I remember reading about the native tribe in northern Canada that has how many different words in their language for “snow” depending upon what type it is? (16, 32, can’t remember, but it’s because it really, really matters to them to be able to tell the difference.)

For us, who cares if we can’t tell if it’s raspberry or strawberry? We have the luxury to be sloppy, and generally are. If we got a sparkplug-like shock or got tasered or got caught by an angry bear, or fell into a semi-frozen lake every time we were wrong, you bet we’d get a lot better quickly.

Eskimos. Or, to be more precise, Inuits and Yupiks. And to be even more precise, that hypothesis has been partially discredited: Eskimo words for snow - Wikipedia .

I remember hearing that thing when I was a kid and was like “whoah, no way”. Then I started to think and quickly came up with a few dozen Finnish words for different kinds of snow, so I guess the claim was somehow plausible.

To be even more precise, it is the hypothesis that the richness of a language’s vocabulary affects thought processes that is discredited. And probably even that statement is not totally accurate.

As I understand it though, the point being made in the research reported in The Economist was how a good smell vocabulary helps communication (rather than thought), and that leads to more agreement when describing smells.

Note though that just because a common language is used, that does not necessarily mean those guys actually perceive smells the same way, any more than our agreeing that something is red means that we experience red in the same way. A colour-blind person might agree with someone with good colour vision that an object is red, but the colour blind person might perceive it as being very similar to green. The same must apply to the wide range of specific anosmias that humans have.