We can describe colors. That is, we can splop down samples of color on paper, and label them. Not only will most people be able to learn the labels and then agree on the colors, but most people will agree that those are representative colors. A pure saturated blue looks pure and saturated to everyone. Other blues are perceived as impure, partial, or kind-of blues.
(Obviously I'm excepting people with vision or color-vision impairments -- that's a small minority, and we consider it impairment, right?)
So we can describe colors with conviction. But when you want to describe a visual image, of course, there's a lot more going on than color. We fall back on analogies, references, and partial description: it's a giant black three-headed dog with slavering fangs.
Describing sound is similar. We think of sounds as being classified by pitch, because pitch is easy to measure and has an objective physical basis (frequency). But pitch isn't actually that important in everyday life. Most people don't have perfect pitch. Even relative pitch (the ability to identify intervals) is not important, unless you're a musician or a birdwatcher. We describe sounds by analogy or reference, just like scenes: it's a loud metallic honking, like a pissed-off brass goose.
(Okay, maybe not a brass goose.)
Describing tastes has always been a fuzzier proposition, and it's not because taste has more dimensions than color or pitch. As I say above, that's a false analogy: vision and hearing have more dimensions than color and pitch! I think taste is fuzzier because we can discriminate lots of tastes -- they vary much more than sounds, I'd say.
(Or perhaps sound varies more, but we're willing to tolerate more slop in describing sounds. There are many brassy honks, but if you imagine something even remotely brassy and honking, then my description was good enough. Whereas people are very picky about flavors.)
Images vary even more, but we are exposed to lots and lots of images, so we have more exemplars to make descriptions from. We give children books full of animal pictures. But it's perfectly possible to grow up without ever tasting asafoetida, and if you do, what am I going to compare it to? It isn't like anything but itself.
However, when you're describing food and drink, taste and smell are always going to be intermixed. So the fact that taste is simple doesn't get you anywhere. I'll try to talk about "flavor", meaning the joint effect of a food's taste and smell. If I forget, don't stress about it.
Fortunately, the weakness of the analogy doesn't stop people from trying. The results are interesting. Even if they aren't wheels in the sense of the color wheel, they're organized hierarchical categories of flavor. They're usually labelled by reference, so you can either identify the categories from your own experience, or go out and find some examples.
On the other hand, none of these wheels are complete. A color wheel encompasses all pure colors (and can be extended to a 3-D solid that encompasses all colors). A pitch wheel encompasses all pitches in an octave (and if you categorize all D-flats together, the wheel encompasses all pitches). But I haven't seen a wheel which claims to describe all flavors.
Which is fine, really. The ones I've seen are for specific purposes: a chocolate flavor wheel, a coffee flavor wheel, a wine flavor wheel. The idea is to cover the flavors you're likely to find in a particular food. And each one was built by an expert in that food -- which is nice, because you can assume that the expert has tried many different varieties of that food. So you have some reason to believe that the wheel is complete, in its domain.
My secret agenda is to see if there's an objective ordering, or even a hint of one, which can be applied to flavors. I figure that if a bunch of different wheels agree in some way, we'll have discovered something.
Dirty |
Fecal |
Gassy |
Farty |
Eggy |
Acrid |
Sickly |
Ammoniacal |
Urine |
Burnt |
Biscuity |
Acidic |
Meaty |
Smoky |
Sulphurous |
Cheesy |
Fishy |
Cabbage |
Sweet |
Onion, Garlic |
Reference:
You can see that context is important. Taken out of context, we don't think of "cabbage" or "meaty" or "sweet" as bad aromas -- unless you dislike cabbage, meat, or cotton candy. This article is written from the point of view of the aroma chemicals industry. (Don't blame me, they call themselves that.) They're trying to make things smell nice, and most people don't want cabbage shampoo.
The organization of the wheel brings some similar aromas together. ("Fecal" is near "farty"; "urine" is adjacent to "ammonia", which is a component of urine. The aromas that can convey appetizing foods are clustered together.)
On the other hand, I don't think I've severed an important link by cutting the wheel between "onion" and "dirty". In fact, you may think I've improved its accuracy: the appetizing flavors are far away from shit. In the original wheel, "onion" was right next to "dirty" and "fecal".
The article also had a wheel for positive aromas, but I think I'll get back to it at the end.
The flavor wheel beer-drinkers use was developed by Morton Meilgaard. It includes all of the above qualities, which I guess makes it an aroma/taste/tactile wheel.
Aromatic, Fragrant, Fruity, Floral |
Alcoholic | odor | feel | |
Solvent-like (plastic, can-liner, lacquer) | ||||
Estery (banana, apple, fruity) | ||||
Fruity (citrus, berry, melon, ...) | ||||
Acetaldehyde (fresh cut green apples) | ||||
Floral (roses, geranium, perfume) | ||||
Hoppy | ||||
Resinous, Nutty, Green, Grassy |
Resinous (sawdust, resin, cedar, pine, spruce) | |||
Nutty (sherry-like: walnut, coconut, beans, almond) | ||||
Grassy (fresh-cut grass, straw) | ||||
Cereal | Grainy (raw grain: husk-like, corn grits, flour) | |||
Malty | ||||
Worty (fresh-wort aroma) | ||||
Caramelized, Roasted | Caramel (burnt-sugar: toffee, molasses, licorice) | |||
Burnt (burnt-bread) | feel | |||
Phenolic | Phenolic (tarry, bakelite, carbolic, pharmaceutical) | |||
Soapy, Fatty, Diacetyl, Oily, Rancid |
Fatty Acid (tallowy, goaty, cheesy, rancid butter) | |||
Diacetyl (buttermilk, butterscotch) | ||||
Rancid (rancid oil) | ||||
Oily (vegetable oil, gasoline, machine oil) | feel | |||
Sulfury | Sulfury (rotten egg) | |||
Sulfitic (burnt-match, choking, burnt rubber) | ||||
Sulfidic (sewage, natural gas) | ||||
Cooked Veg. (overcooked greens, cooked corn) | ||||
Yeasty (fresh yeast, meaty) | ||||
Oxidized, Stale, Musty | Stale (old beer) | feel | ||
Catty (skunky beer) | ||||
Papery (cardboard) | ||||
Leathery | ||||
Moldy (earthy, musty) | ||||
Sour, Acidic | Acidic (pungent, sharp) | taste | ||
Acetic (vinegar) | ||||
Sour (lactic, sour milk) | ||||
Sweet | Sweet (syrupy) | |||
Salty | Salty | |||
Bitter | Bitter (harsh, dry) | |||
Mouthfeel | Alkaline | feel | ||
Mouthcoating (creamy) | ||||
Metallic (coins, iron, rusty water, tinny) | odor | |||
Astringent (mouth-puckering, tannin-like, tart) | ||||
Powdery (dusty, chalky, particulate) | odor | |||
Carbonation (flat/undercarbonated, gassy/overcarbonated) | ||||
Warming (alcoholic) | ||||
Fullness | Body (watery, characterless, satiating, thick) | odor |
References:
As you might expect, the chart can plausibly be divided into "smell", "taste", and "texture" -- although there's an overlap between smell and taste, and no pure textures. The aromas are arranged roughly from pleasant, light odors through darker ones to the nasty ones. However, there are still a lot of arbitrary choices.
Meilgaard has tried to make it more of a cycle by having the smells run into the tastes, and putting the "alkaline" end of the texture group next to the "bitter" end of the taste group. (Alkalis taste bitter and feel soapy.) Nonetheless, I'd say this chart has more value as a hierarchy than as a wheel.
Floral | Floral | Linalool | |
Orange blossom | |||
Rose | |||
Violet | |||
Geranium | |||
Spicy | Spicy | Cloves | bouquet |
Black pepper | |||
Licorice, anise | |||
Fruity | Citrus | Grapefruit | |
Lemon | |||
Berry | Blackberry | ||
Raspberry | |||
Strawberry | |||
Black currant/cassis | |||
Tree Fruit | Cherry | ||
Apricot | |||
Peach | |||
Apple | |||
Tropical Fruit | Pineapple | bouquet | |
Melon | |||
Banana | |||
Dried Fruit | Strawberry jam | ||
Raisin | |||
Prune | |||
Fig | |||
Other | Artificial fruit | ||
Methyl anthranilate | |||
Vegetative | Fresh | Stemmy | |
Cut green grass | |||
Bell pepper | |||
Eucalyptus | |||
Mint | |||
Canned/Cooked | Green beans | ||
Asparagus | |||
Green olive | |||
Black olive | |||
Artichoke | |||
Dried | Hay/straw | ||
Tea | |||
Tobacco | |||
Nutty | Nutty | Walnut | bouquet |
Hazelnut | |||
Almond | |||
Caramelized | Caramelized | Honey | |
Butterscotch | |||
Butter/diacetyl | |||
Soy sauce | |||
Chocolate | |||
Molasses | |||
Woody | Phenolic | Phenolic | |
Vanilla | |||
Resinous | Cedar | ||
Oak | |||
Burned | Smoky | ||
Burnt toast/charred | |||
Coffee | |||
Earthy | Earthy | Dusty | |
Mushroom | |||
Moldy | Musty/mildew | bad | |
Moldy cork | |||
Chemical | Petroleum | Tar | |
Plastic | |||
Kerosene | |||
Diesel | |||
Sulfur | Rubbery | ||
Hydrogen sulfide | |||
Mercaptan | |||
Garlic | |||
Skunk | |||
Cabbage | |||
Burnt match | |||
Sulfur dioxide | |||
Wet dog | |||
Papery | Filter pad | ||
Cardboard | |||
Pungent | Ethyl acetate | ||
Acetic acid | |||
Ethanol | |||
Sulfur dioxide | |||
Other | Fishy | ||
Soapy | |||
Sorbate | |||
Fusel alcohol | |||
Pungent | Hot | Alcohol | |
Cool | Menthol | ||
Oxidized | Oxidized | Acetylaldehyde/sherry | |
Microbiological | Yeasty | Flor yeast | |
Leesy | |||
Lactic | Sauerkraut | ||
Butyric acid | |||
Sweaty | |||
Lactic acid | |||
Other | Horsey | ||
Mousey |
(Late update: I am told that the beer wheel dates back to the late 70's, if not earlier.)
This wheel solely considers aroma -- in fact, the creator's web page talks about making samples that are only sniffed, not tasted. The ranges marked "bouquet" are apparently the ones usually associated with wine; wacky varietals can span the entire non-nasty side of the wheel.
The chart has the light/dark/nasty sequence which we saw in the beer chart, and which we will see again. I couldn't find a point on the wheel which was unarguably the "top", so I split it between the nasty aromas and the light (floral/spicy) ones. "Microbiological" and "Floral" don't seem to be related, so I don't think I did the chart any harm. The fact that "Microbiological" (and "Chemical") have "Other" subcategories indicates that the cyclical nature of the chart is not very complete.
Yes, sulfur dioxide is on there twice. I don't know why. But you can't argue with an aroma wheel that include "Wet dog".
Aromas | Enzymatic | Flowery | Floral | Jasmine |
Wintergreen | ||||
Fragrant | Cardamon, Caraway | |||
Sweet Basil, Anise | ||||
Fruity | Citrus | Lemon | ||
Tangerine | ||||
Berry-like | Raspberry | |||
Blackberry | ||||
Herby | Alliaceous | Onion | ||
Garlic | ||||
Leguminous | Cabbage | |||
Alfalfa | ||||
Sugar Browning | Nutty | Nut-like | Peanut | |
Almond | ||||
Malt-like | Corn | |||
Barley | ||||
Carmelly | Candy-like | Toffee | ||
Pralines | ||||
Syrup-like | Honey | |||
Molasses | ||||
Chocolatey | Chocolate-like | Bakers | ||
Dutch | ||||
Vanilla-like | Swiss | |||
Custard | ||||
Dry Distillation | Resinous | Turpeny | Piney | |
Balsamic | ||||
Medicinal | Camphoric | |||
Cineolic | ||||
Spicy | Warming | Nutmeg | ||
Pepper | ||||
Pungent | Clove | |||
Thyme | ||||
Carbony | Smoky | Tarry | ||
Tobaccoey | ||||
Ashy | Burnt | |||
Charred | ||||
Tastes | Bitter | Pungent | Creosol | |
Phenolic | ||||
Harsh | Caustic | |||
Alkaline | ||||
Salt | Sharp | Astringent | ||
Rough | ||||
Bland | Neutral | |||
Soft | ||||
Sweet | Mellow | Delicate | ||
Mild | ||||
Acidy | Nippy | |||
Piquant | ||||
Sour | Winey | Tangy | ||
Tart | ||||
Soury | Hard | |||
Acrid |
References:
I'd be more impressed with this massive chart if I didn't think it was massively warped in favor of symmetry. They've divided everything into equal numbers of examples, no matter how much variety a given region might have. I mean, "vanilla/swiss" versus "vanilla/custard"? In the same amount of space that they've spent on all the nuts?
They've made a token effort to put the herbs oppsite the dark burnt flavors, but this chart has essentially no wheel-nature. "Tastes" and "aromas" take up equal sides of the wheel (although my chart representation doesn't do that), apparently for no reason but visual balance.
The "Tastes" side is ordered completely at random, and even the subcategories make no sense. (Why are there "bland" and "sharp" varieties of "salt", but not of the other tastes? Why have a section for "astringent", which apparently means "salty and bitter", but not for any other taste combination involving salt? The four -- or five -- tastes combine equally, not in a sequence.)
Many words, but no interesting organization.
Strangely, when I ordered the poster from Sweet Maria's, it differed from the wheel linked on their web site. The categories were the same, but the last (detailed) column on the "Aromas" side was very different:
Aromas | Enzymatic | Flowery | Floral | Coffee Blossom |
Tea Rose | ||||
Fragrant | Cardamon, Caraway | |||
Coriander Seeds | ||||
Fruity | Citrus | Lemon | ||
Apple | ||||
Berry-like | Apricot | |||
Blackberry | ||||
Herby | Alliaceous | Onion | ||
Garlic | ||||
Leguminous | Cucumber | |||
Garden Peas | ||||
Sugar Browning | Nutty | Nut-like | Roasted Peanuts | |
Walnuts | ||||
Malt-like | Balsamic Rice | |||
Toast | ||||
Carmelly | Candy-like | Roasted Hazelnut | ||
Roasted Almond | ||||
Syrup-like | Honey | |||
Maple Syrup | ||||
Chocolatey | Chocolate-like | Bakers | ||
Dark Chocolate | ||||
Vanilla-like | Swiss | |||
Butter | ||||
Dry Distillation | Resinous | Turpeny | Piney | |
Black Currant-Like | ||||
Medicinal | Camphoric | |||
Cineolic | ||||
Spicy | Warming | Cedar | ||
Pepper | ||||
Pungent | Clove | |||
Thyme | ||||
Carbony | Smoky | Tarry | ||
Pipe Tobacco | ||||
Ashy | Burnt | |||
Charred |
I think this latter version is obsolete -- the current SCAA web store shows an image of the first chart. (I suppose I should order another chart from the SCAA and see what I get. I haven't done this yet.)
The first one is certainly more regular: it has "lemon, tangerine" as examples of "citrus", instead of "lemon, apple". And it's more familiar: I don't have a clue what "balsamic rice" is.
But I have to wonder if the regularity and familiarity, like the symmetry, was put in form's sake rather than to express actual sensory qualities. Whoever put "apple" under "citrus" the first time must have been thinking something. It might have been a mistake, but at least I know they weren't pulling examples from a preconceived "citrus fruit" category.
Vegetable | Mushroom |
Raw Coffee | |
Green Tomatoes | |
Truffles | |
Wood | |
Flowery | Jasmine |
Orange Blossom | |
Rose | |
Fruity | Apricot Preserves |
Currant Preserves | |
Red Berries | |
Orange | |
Dried Fruit | |
Dried Plums | |
Dried Bananas | |
Wild Berries | |
Roasted | Roasted Almonds |
Cocoa | |
Caramelized Sugar | |
Caramel | |
Marzipan | |
Espresso/Coffee | |
Black Tea | |
Nutty | Cashews |
Almonds | |
Hazelnut | |
Macadamia | |
Spicy | Liquorice |
Cloves | |
Cinnamon | |
Oriental Spices | |
Vanilla | |
Miscellaneous | Beeswax |
Honey | |
Bread | |
Cream/Milk | |
Butter | |
Tobacco |
References:
Not much to report here; it's a simple hierarchy. Not a wheel in any sense. (You don't get to call it a wheel if you have "Miscellaneous" as one point on your cycle.) The categories are very broad, and while the examples are presumably appropriate for chocolate, they come off as pretty arbitrary. (Marzipan is just almonds and sugar. Caramel is caramelized sugar with cream or butter. Why do these get separate entries? If those combinations are supposed to form new unique aromas, doesn't that throw into question the whole idea of a spectral breakdown of flavors? Oh well.)
Spicy | Cloves | |
Cinnamon | ||
Anise, black liquorice | ||
Fruity | Nuts (Bitter almond, hazelnut) | |
Fruits with pits or seeds | Peach, Apple | |
Baked Apple | ||
Citrus fruits (orange, orange peel) | ||
Floral | Flowers | |
Honey | ||
Empyreumatic | Light | Golden sugar |
Chicory, toast | ||
Medium | Caramelized sugar | |
Burnt wood, ground brown coffee | ||
Brown coffee bean, chocolate | ||
Strong | Burnt sugar | |
Ground black coffee | ||
Black coffee bean | ||
Smoked | ||
Milky | Fresh | Butter, cream, milk |
Heated | Butter, milk | |
Vanilla | Vanilla | Marshmallow |
Vanilla pod | ||
Confectionery | Light | White sugar |
Medium | Corn syrup | |
Light brown sugar | ||
Strong | Dark brown sugar | |
Molasses | ||
Sponge toffee | ||
Maple | Maple | Maple |
Roasted dandelion root | ||
Plant (Woody) | Ligneous | Firewood, wet wood |
Sawdust | ||
Softwood (pine, juniper, cedar, etc) | ||
Plant (Vegetable) | Humus, Forest | Mushroom |
Mold, Potato | ||
Cereals | Malt, Oats, Wheat, Rye | |
Plant (Herbaceous) | Fresh herbs | Stem, grassy |
Shoot, bud | ||
Dry herbs | Crushed leaves | |
Nutshells | ||
Dry herbs, hay | ||
Foreign (Environment) | Solvents | |
Soap and detergents | ||
Plastics and wrapping | ||
Petroleum and derivatives | ||
Mineral, water | ||
Enclosed (dry) dust | ||
Cardboard | ||
Metal | ||
Drugs and drugstore | ||
Foreign (Deterioration) | Sulphured (burnt sulphur) | |
Rancidity (rancid grease) | ||
Confined humidity (soiled mop) | ||
Carbon dioxide | ||
Fermentation (vinegar, yeast) |
References:
Everyone has to get into the act. Now it's maple syrup.
When I transcribed this chart, I started on the right side, instead of the top (as I do for most of the wheels on this page). If I'd started at the top, I would have split "Vanilla" from "Milky", which seemed wrong. Again, most of the wheels I've found start with herb/fruit/spice, so I used that as the default cut point.
I also went counterclockwise, since that's the way the wheel was printed, and it was the only way to preserve the light/medium/dark orderings. Interestingly, the text chart printed below the wheel (on the AAFC web page) shows the categories clockwise, but the entries in each categories counter-clockwise! This implies that even the creators don't think of the wheel as a true circle -- jumping back and forth around the perimeter doesn't bother them.
On the other hand, the categories do cluster to some extent. "Vanilla" does belong near "Milky"; "Floral" and "Fruity" and "Spicy" belong together. And of course several of the categories ("Empyreumatic", "Confectionery") have a logical internal order. So it's not just a totally unordered hierarchy.
Back on the first hand, I'll note that "Foreign Environmental" and "Foreign Deterioration" are together -- but this is really a matter of grouping everything bad and dividing the badness by origin. It doesn't imply that the bad flavors are similar to each other.
You actually get more of a wheel if you snip out the nasty flavors entirely. Then "Herbaceous" is adjacent to "Spicy", and you have a decent light/dark/light progression.
(In case you were wondering, "Empyreumatic" refers to substances produced when organic matter is broken down by heat.)
Oh, and what is it the maple industry isn't telling me about roasted dandelion root?
Green, Grassy |
Fruity, Ester-like |
Citrus, Terpenic |
Minty, Camphoraceous |
Floral, Sweet |
Spicy, Herbaceous |
Woody, Smoky |
Roasty, Burnt |
Caramel, Nutty |
Bouillon, HVP |
Meaty, Animalic |
Fatty, Rancid |
Sulfurous, Alliaceous |
Mushroom, Earthy |
Celery, Soupy |
Dairy, Buttery |
This is given as a "traditional" flavor wheel -- their quotes. Unfortunately, the article promptly drops this chart and focusses instead on a "high-impact" flavor wheel. The article is about "high-impact aromas": Rowe defines these as chemicals which are easy to smell in small quantities, distinctive, attractive, easy to synthesize, not too expensive, and stable in products. These distinctions are obviously important to flavor chemists, but not very significant to us mere sniffers. Some of the categories above don't have high-impact chemicals, and others ("Mint") are apparently too easy for chemists to bother with. So he stretches and shrinks and cuts and tucks and comes up with this:
Green, Grassy | sweet | fresh | ||||
Herbaceous | ||||||
Spicy | ||||||
Fruity, Ester-like | ||||||
Tropical | ||||||
Blackcurrant, Cassis | coffee | |||||
Vegetable | savory | |||||
Nutty | cooked | roast beef | coffee | |||
Caramel | ||||||
Smoky | ||||||
Burnt, Roasted | ||||||
Beefy | ||||||
Pork, Lamb, Chicken | ||||||
Savory, Bouillion | ||||||
Fatty | ||||||
Rancid, Cheesy | ||||||
Mushroom, Earthy | fresh | coffee | ||||
Truffle | ||||||
Garlic | ||||||
Onion |
Why do I bother with the "high-impact" wheel, if its categories are only interesting to flavor chemists? Well, because Rowe talks more about it. Also, he seems to have reordered the wheel into a more consistent form, and he actually provides some evidence for that order.
The chart has a reasonable progression around the wheel. As the article notes, the early range is "sweet", and then it hits "savory" -- but those ranges overlap somewhat. The wheel can also be divided into "fresh" flavors (those produced by living plants) and "cooked" ones (those produced by Maillard reactions when foods are heated. Not just boiled, mind you, but golden-brown-and-delicious stuff.)
So we have at least a four-node cycle: sweet/fresh, sweet/savory/cooked, savory/cooked, savory/fresh. The elements in each range are ordered with some logic. For example, the savory/cooked entries go from "burnt" (which is near "smoky" in the previous range) down through "beefy" (a heavy aroma), through lighter meats, through "fatty" (which is a component of meat) to "cheesy" (which has fat, but is not meaty).
The article also provides spectrographs, as it were: a range of flavors touched by a particular food. Roast beef covers a solid arc of the wheel, which is a good sign -- it indicates that the wheel is clustering flavors which belong together. Coffee also covers an arc, but has some outlying elements.
(The article also mentions chocolate, but the on-line version I am looking at is missing that page! The Mystery of Chocolate remains opaque.)
Right. Looking at all of these wheels, can we come up with a "true" overall spectrum for flavors?
Everyone agrees that flowers, fruits, and herbs go together. Everyone agrees that nuts, fats, meats, and cooked foods (the products of heat reactions) go together. Bad smells (which are generally either the products of decay, or completely non-food chemicals) can safely be put together, although more because they're all bad than because they're similar. Decay and earthy flavors are often put near the dark roasted ones.
When you try to zoom in, the details get fuzzy very quickly. Flowers, fruits, and spices get listed in every conceivable order. When people add green (vegetable) aromas and nutty aromas, those too show up in many different relative positions.
Within a category -- say, fruits, or nuts -- ordering is essentially arbitrary. People like to list the same examples (citrus, apricot, blackcurrant, berries), so these are obviously distinctive flavors, but they get arranged any way you can think of.
I can't say much about meaty and cheesy aromas, because most of these wheels are for products that don't have them. They're associated with fats and cooked aromas, which is no surprise (what with the whole "cooking meat" idea that we've become so attached to the past ten thousand years.)
Nobody even wants to talk about fish.
And yet, despite the non-cyclicalness of all these charts, they're much cooler as wheels than as vertical charts. My copies on this page are boring. I've got a couple of the wheel posters hanging in my kitchen, and they look great.
Therefore, people will keep making them, whether they're justified or not.
In the end, the savory/sweet fresh/cooked cycle -- the last one shown above -- is the closest thing we've got to a consistent, universal organization. Treasure it and enjoy it. I've spent three thousand words and a fair amount of table-generation code demonstrating that nobody has anything else to say about flavor wheels. I hope you've enjoyed it as much as I have. If not, then at least you have references to all these wheels in once place -- which was my other goal.
My Periodic Table of Dessert, a silly and not-really-cyclical chart