G: If you mentally subtract the sweetness from caramelization, what remains is roughly:

toasty + nutty + roasted + slightly bitter, with a warm “browned” aroma.

A useful way to map it onto familiar foods is the dark-brown edge of toast, but without the breadiness. Other reference points are roasted nuts, the browned exterior of a pretzel, very dark maple-like aromas without the sugar, and the faint bitterness of heavily browned onions.

There’s an important distinction: caramelization itself produces both sweet-smelling and bitter/roasted compounds. As it progresses, the flavor moves roughly:

sweet/fruity → caramel/toasty → nutty/roasted → bitter/burnt

So if you're trying to isolate the non-sweet component of ordinary golden-brown caramelization, I'd call it toasty-roasted, more than anything else.

And this actually gives you another good landmark for the bitter-vs-sour sensory map you were working on recently: take caramel just a little too dark, and the increasingly sharp flavor you get is bitterness, not sourness. memcite

G: Yes. I think I understand the sensation you're trying to isolate, and caramel is involved, but "caramel flavor minus sweetness" isn't quite the whole thing.

What you're perceiving when you move from something like Mountain Dew to cola is probably a bundle of several effects that your brain reasonably collapses into one dark/deep/browned dimension:

• Caramelized/burnt-sugar aromatics — toasted, slightly burnt, almost woody notes.
• Bitterness — modern cola doesn't need actual kola nut to produce it; caffeine itself is bitter, and flavorings contribute more.
• Warm spice aromatics — typical cola flavor systems involve things in the cinnamon/cassia, citrus-oil, vanilla, coriander, nutmeg-type territory. The exact recipes are proprietary and differ.
• Caramel color — this is particularly interesting for your question. Caramel coloring is produced by strongly heating carbohydrates, but commercial caramel color isn't simply melted table sugar dumped into Coke. It's a family of highly processed browned materials, and depending on the type and concentration it can contribute some flavor as well as color.
• Acid — cola generally uses phosphoric acid, whereas Mountain Dew and many citrus sodas lean heavily on citric acid. Phosphoric acid has much less of the obvious fruit-associated sourness of citric acid, so it can contribute to that harder, darker impression.

So I think your "umami-type unlabeled feeling" is a perfectly useful way of describing the perceptual problem. It's not literally umami; you're looking for something more like "depth" or "darkness." Food science doesn't give that combination one wonderfully convenient basic-taste name.

What I'd actually experiment with

There is something remarkably close to the "liquid smoke for this dimension" you're imagining:

caramel flavoring/extract.

Importantly, caramel flavoring does not have to contain enough sugar to make something appreciably sweeter. Food-flavor manufacturers can formulate the volatile compounds responsible for the aroma of caramel separately from a caramel syrup's enormous sugar load.

But I wouldn't start there, because there's a ridiculously easy kitchen experiment that may tell you exactly what component you're noticing.

Take a glass of a very sweet, relatively flavor-light citrus soda and divide it into small samples. Try one or two drops at a time of:

1. Vanilla extract — not enough that you consciously identify "vanilla." At trace concentration it produces surprising depth.
2. Angostura-type aromatic bitters — extremely little. This supplies bitter + woody + spice complexity and may get startlingly close to the missing perceptual dimension you're describing. (Standard cocktail bitters contain alcohol, though the amount from a drop or two is tiny.)
3. Caramel flavoring — if you get some, this is the cleanest direct test of your hypothesis.
4. A combination of trace vanilla + caramel flavor + bitters. This is where I'd bet you'll suddenly go, "THAT. That's the thing I mean."

And there's an even simpler experiment available right now if you have it: very strong coffee. Put literally a few drops of black coffee into a glass of sweet soda—not enough to make it taste like coffee. Coffee supplies roasted aromatics and bitterness without adding sweetness. If the soda suddenly acquires some of the "weight" you're talking about, you've established that roast/bitter aromatics are an important part of your mystery axis, rather than caramel specifically.

One caution from your liquid-smoke analogy is exactly right: these are seasonings, not secondary ingredients. If you can identify "coffee," "vanilla," or "bitters," you've probably gone past the concentration you're looking for.

In fact, I think very dilute coffee versus caramel flavoring would be a particularly informative A/B test: coffee gives you roast without caramel, while caramel flavor gives you caramel aroma without much sweetness. That would let you determine which half of this otherwise rather fuzzy "dark soda" sensation your palate is actually responding to.

G: [Non-text content: thoughts; content_type=thoughts]

G: [Non-text content: reasoning_recap; content_type=reasoning_recap]

G: Yes. That description moves me somewhere fairly specific, and I think my previous answer was putting too much emphasis on flavor/aroma and not enough on chemesthesis—chemical sensation from the drink.

What you're describing as “registering deeper into the swallow,” near the larynx, is anatomically plausible as a sensory phenomenon. The mouth, pharynx, and laryngeal region have chemical-sensing nerve systems that respond to things such as acids, CO₂, pepper compounds, alcohol, menthol, etc. Carbonation itself isn't merely bubbles mechanically touching you: dissolved CO₂ is converted locally to carbonic acid and produces a chemically mediated sting/tingle. citeturn0search6turn0search7turn0search15

And your cola-versus-citrus distinction points toward something I mentioned before but probably underrated: phosphoric acid.

I think phosphoric acid is now the prime suspect

Colas characteristically use phosphoric acid, while citrus sodas commonly get much of their acid profile from citric acid. Food-science descriptions distinguish the sensory qualities of those acids rather dramatically: citric gives a bright, fruit-associated tartness, whereas phosphoric is described as drier, sharper, more astringent and particularly suited to the dry/balsamic character of cola. citeturn0search0turn0search8

That's awfully close to your distinction:

Mountain Dew/Sprite:
sweet → bright/tart → carbonation → then somehow feels vacant farther down

Cola:
sweet → acid/carbonation → something continues through the swallow → more complete/heavy finish

I would not say we've established that phosphoric acid specifically stimulates farther down your throat than citric acid. I couldn't find evidence supporting that precise anatomical claim, and one controlled study of commercial carbonated beverages actually found sensory “burn/bite” correlated positively with total acidity/citric acid and negatively with phosphoric acid. So the simple explanation “phosphoric acid burns the throat more” would be too neat and apparently wrong as a general rule. citeturn0search2

But phosphoric acid can still change the character and temporal profile of the acidity. And your observation that bitters helped the complexity but didn't fill the missing physical location is excellent evidence against my earlier idea that what you're seeking is principally roast/caramel aromatics.

There's a really good experiment here

Rather than buying caramel flavoring yet, I'd try adding a tiny amount of food-grade phosphoric acid to a citrus soda.

Not industrial/pharmacy/rust-removal phosphoric acid—specifically a food-grade beverage acidulant sold with dilution instructions. Concentrated phosphoric acid is corrosive, so I don't want to give you a kitchen-drop recipe based on an unknown concentration.

What makes the experiment especially useful is that you don't actually want to turn Mountain Dew into cola. You want:

Mountain Dew's relative lightness + the missing “lower/deeper” component of cola.

So you could leave the citrus flavor, sweetness, carbonation, and lack of caramel completely alone and alter only the acid system. If a small phosphoric component suddenly makes the swallow feel “complete,” we've isolated a major part of your mystery without having to name the sensation.

And there may ultimately be two separate axes hiding under what you originally called the caramel effect:

Darkness/depth of flavor ← caramel, vanilla, spice, bitterness
Completeness/depth of swallow ← acid system + carbonation/chemesthesis

Your bitters experiment seems to have moved the first axis while leaving the second relatively untouched. That's actually quite informative.

One other candidate I'd keep in reserve is astringency—the dry/gripping sensation rather than a taste. Phosphoric acid's cola application is specifically described in beverage literature as having an astringent acidity. citeturn0search8 That could explain why you've had such trouble assigning this thing a flavor name: the missing component may not actually be a flavor.

So at this point, I would put phosphoric-acid/acid-system chemesthesis and astringency ahead of caramelization as the explanation for the particular “around the bend and toward the larynx” sensation you've now identified. The caramel/spice/bitter system is probably contributing to why cola tastes deep, but may not be what's making it feel deep.

G: That does push the problem toward chemistry rather than “what spice tastes cola-like.” I’m going to separate acid-driven throat sensation from TRP-receptor irritants and astringency, then map those back onto ordinary ingredients you could actually use.

G: [Non-text content: thoughts; content_type=thoughts]

G: [Non-text content: reasoning_recap; content_type=reasoning_recap]

G: Yes. Looking at the receptor chemistry, I think the most useful way to frame this is that you may be trying to build a three-part “dark soda body”:

1. bitter/aromatic complexity — bitters already helped here;
2. deep chemesthetic stimulation during the swallow — probably the part you're still missing;
3. acid structure — phosphoric acid may contribute, but it does not have to be the only way to get #2.

Chemesthesis really is the right umbrella word. Compounds in spices activate sensory ion channels such as TRPA1, TRPV1, TRPM8, and TRPV3, producing warmth, pungency, cooling, irritation, and related sensations that are not ordinary taste. Those effects are also modified by acidity, carbonation, and temperature. citeturn822975search2turn822975search7

For the exact “I feel this farther back/down during the swallow” clue, the most interesting chemistry is TRPA1/TRPV1. There are studies deliberately putting compounds that activate these receptors into swallowed liquids, and they measurably alter pharyngeal sensory responses and swallowing. Piperine in particular—the main pungent compound of black pepper—activates both TRPV1 and TRPA1, and has been experimentally shown to stimulate oropharyngeal sensory neurons during swallowing. citeturn216027search3turn216027search8

So for ingredients you can actually find, my candidates would rank roughly like this:

• Black pepper → piperine: probably my #1 experiment. It gives a slowly developing, warm, penetrating chemesthesis rather than the immediate fiery front-of-mouth blast of chili. That's a potentially very good match for “registers deeper in the swallow.” You'd want a trace, ideally an infusion rather than floating pepper particles.
• Ginger → gingerols/shogaols/zingerone: extremely interesting #2. Ginger compounds activate TRPV1, and some also activate TRPA1. That produces warmth/pungency without needing more acid. citeturn216027search1turn216027search2turn216027search4
• Cinnamon/cassia → cinnamaldehyde: very relevant because cinnamaldehyde is a strong TRPA1 agonist, and cinnamon is genuinely part of the traditional cola-flavor family. citeturn216027search5turn822975search11
• Clove → eugenol: gives warm/numbing chemesthesis through temperature-sensitive TRP channels. Clove also appears in documented cola-flavor formulations, although it's less universally central than cinnamon. citeturn216027search6turn822975search40
• Menthol/mint → TRPM8: chemically works, and menthol has actually been studied for pharyngeal/swallow stimulation, but I suspect it's wrong for your target because it reads as cooling/opening rather than weight/depth. citeturn822975search12
• Capsaicin → TRPV1: definitely works chemically, but I'd put it near the bottom because even tiny amounts can announce themselves as “hot pepper,” which is probably not the invisible structural effect you're after.

There is one fascinating piece of evidence that makes your anatomical description much less weird than it might sound. Oleocanthal, the peppery compound in good extra-virgin olive oil, produces a pungency perceived almost exclusively in the throat, not evenly across the mouth. Researchers traced that unusual localization to TRPA1 and the anatomical distribution of the relevant sensory system. citeturn216027search0 So “this particular chemical sensation appears farther back than another one” is absolutely a real phenomenon, not merely your brain inventing a spatial metaphor for flavor.

That doesn't make olive oil a particularly good soda additive, of course. But it strongly supports the kind of mechanism we're hunting.

On the cola side, known and reconstructed cola systems repeatedly contain cinnamon/cassia, citrus oils, nutmeg, coriander, vanilla, with some formulations also involving clove and ginger, alongside phosphoric/citric acid. citeturn822975search40turn822975search11 That makes cinnamon and ginger especially attractive experiments because we're not just randomly throwing pungent substances into soda: we're hitting chemesthetic receptors using materials that already live naturally inside the cola flavor family.

And your molarity question needs one little chemistry correction. Molarity just means concentration in moles per liter; it isn't synonymous with acidity. Adding anything dissolved technically increases the concentration of that substance. What you're really trying to avoid is increasing hydrogen-ion activity / titratable acidity / acid load substantially.

On that point: yes, you can absolutely increase chemesthesis without increasing acidity.

Piperine, cinnamaldehyde, gingerols/shogaols, menthol, eugenol and capsaicin do not require you to lower the drink's pH in order to activate their respective sensory receptors. That's the major reason I'm now more interested in them than in simply telling you to add phosphoric acid. TRPA1/TRPV1 stimulation can supply an extra sensory “channel” while leaving the acid system essentially where it started. citeturn216027search9

Phosphoric acid might still contribute something unique. Different beverage acids really do have different sensory characters rather than pH alone determining the result; citric and phosphoric acid also behave differently in soft-drink flavor systems. citeturn822975search3 But you're right that simply making an already acidic soda more acidic until your throat notices it would be a terrible way to solve the problem.

So I would change the experimental strategy from my previous answer. Don't buy phosphoric acid first.

Take a citrus soda and divide a small amount among a few cups. Keep your existing bitters as the baseline, because they already supplied something desirable. Then separately test black pepper, ginger, and cinnamon, at concentrations so low that your first goal is not to identify the spice.

For black pepper, I'd actually make a crude pepper water rather than sprinkle pepper into the soda: steep freshly cracked black pepper briefly in hot water, strain it very thoroughly, cool it, and add that liquid by drops. With ginger, do the same with a small piece of fresh ginger or use a very weak ginger tea. With cinnamon, a weak cinnamon-water infusion is preferable to dumping ground cinnamon into carbonated liquid.

What you're looking for isn't:

“Oh, now it tastes like ginger.”

It's:

“Huh. I can't quite taste what changed, but the swallow now occupies that missing space.”

If black pepper does it, piperine/TRPV1-TRPA1 becomes a very strong lead. If ginger does it, TRPV1/TRPA1 warmth is probably the axis. If cinnamon does it, you've found something particularly elegant because cinnamaldehyde supplies chemesthesis and belongs naturally inside cola's aromatic architecture.

And I now suspect your eventual solution may look less like one magic ingredient and more like something such as bitters + minute cinnamon/ginger/piperine chemesthesis, with phosphoric acid either unnecessary or only a tiny final adjustment.

One last distinction that may help you notice the results: don't ask yourself only “How strong is the sensation?” Also pay attention to where it begins, where it peaks, and how long it survives after swallowing. Those spatial and temporal differences are precisely where these receptor systems can distinguish themselves.

G: Yes — this is exactly the point where it makes sense to stop free-associating ingredients and look for the sensory-science / beverage-industry vocabulary that already exists around throat impact, trigeminal stimulation, astringency, carbonation bite, and swallow-phase perception. I’m going to look for work that treats these as measured sensory phenomena, not just recipe folklore.

G: [Non-text content: thoughts; content_type=thoughts]

G: [Non-text content: reasoning_recap; content_type=reasoning_recap]

G: Yes. There is a real body of work here, and it changes how I’d approach the problem. We are not forced to reverse-engineer this entirely from recipes. The relevant literature is scattered across sensory science, beverage formulation, oral chemesthesis, and swallowing research, but the pieces line up surprisingly well.

The first useful correction is that industry/sensory researchers generally don’t treat the sensation of a carbonated beverage as one thing. They break it into attributes such as bite, burn, numbing, carbonation, astringency, mouth-coating, heaviness, cooling, pungency, and lingering after-sensation. One beverage-science reference explicitly lists the main oral symptoms of carbonated drinks as “bite, burn, numbing, carbonation, astringent, and mouth coating.” citeturn459752search34

That is already much closer to what you’ve been trying to describe than “cola flavor.”

Even better, there is an actual sensory lexicon developed for carbonated lemon-lime beverages. It separately defines carbonation, foam stability, tongue heaviness, mouth-coating, after-numbing, and astringent mouthfeel. So the “complete swallow” sensation you’re describing may genuinely be the sum of several sensory dimensions, some tactile and some chemical, rather than one missing taste compound. citeturn655394search41turn655394search43

The strongest piece of evidence for your specific location description comes from swallowing research. Researchers studying dysphagia deliberately add compounds such as piperine, capsaicin, cinnamaldehyde, and citric acid to swallowed liquids because they stimulate chemosensory pathways in the oropharyngeal/pharyngeal region, not merely the tongue. Piperine in particular—a dual TRPV1/TRPA1 agonist—has been shown to alter the swallow response when incorporated into the swallowed bolus. citeturn885824search3turn885824search30

There is even a clinical trial that essentially constructed the experiment we were independently inventing. It compared separate solutions containing:

capsaicin → TRPV1
piperine → TRPV1 + TRPA1
cinnamaldehyde → TRPA1
citric acid → acid-sensitive ASIC3
and combinations of those stimuli.

That is extremely useful conceptually because researchers themselves regard these as different sensory channels capable of modifying pharyngeal sensation rather than interchangeable forms of “spiciness.” citeturn885824search9turn885824search31

So your instinct about piperine being worth prioritizing is reinforced rather than weakened.

There’s another important body of work around carbonation itself. CO₂ isn't merely providing bubble texture. Carbonation activates chemical irritation pathways and changes sweetness, sourness, aroma perception, and other oral sensations. A controlled beverage study found that adding CO₂ reduced perceived sweetness, increased sourness, and altered aroma perception. citeturn459752search0 Carbonated liquids have also been studied specifically for their effects on pharyngeal swallowing, meaning there is demonstrably sensory stimulation occurring at the stage of the swallow you're pointing to. citeturn885824search0turn885824search5

That gives us a potentially important hypothesis about the cola/root-beer observation:

the missing property may not be one cola flavoring at all. It may be a formulation that keeps chemesthetic stimulation active for longer through the swallow.

And root beer becoming part of your comparison is particularly helpful. Root beer and cola have very different defining flavor systems, yet both can give you the sensation you're looking for. That makes a cola-specific aromatic such as kola, cinnamon, or caramel less likely to be the sole cause.

Root beer formulations commonly involve things such as wintergreen/methyl salicylate, cinnamon/cinnamaldehyde, clove/eugenol, anise/anethole, vanilla, and other botanical extracts. An industry flavor article describes exactly this kind of architecture. citeturn655394search2 Some of those molecules are chemesthetic, but importantly, there isn't one obvious molecule shared uniquely by root beer and cola that Sprite lacks. The similarity is more likely occurring at the sensory-system level.

That makes me revise the candidate hierarchy somewhat.

Piperine still looks excellent, because it recruits both TRPV1 and TRPA1 and has demonstrated effects during the pharyngeal phase of swallowing.

Cinnamaldehyde remains plausible, but your experience with cinnamon makes me agree it probably isn't the whole answer.

Clove/eugenol gets more interesting. Eugenol doesn't just produce “spice flavor”; it has pungent and mildly numbing/anesthetic-like sensory properties. That combination could contribute to the kind of rounded, persistent throat presence you're describing rather than merely a sharp burn.

And there is another chemical family I would now put on the board because of root beer:

methyl salicylate / wintergreen.

I'm not saying “add wintergreen and you'll solve it.” Root beer relies heavily on that compound, and at identifiable concentrations it screams wintergreen. But it is a chemesthetic compound with cooling/pungent characteristics, and the fact that root beer commonly contains it while still giving you this deeper sensation is interesting. It could be one member of a broader layered stimulus rather than the decisive component. citeturn655394search2

The literature also supports your intuition that the satisfactory sensation can depend on interactions, not merely additive intensity. A modern review from flavor-industry researchers specifically discusses using trigeminal sensations—cooling, warming, pungency, tingling, carbonation—to enrich flavor perception, and emphasizes that these sensations interact with taste and aroma rather than functioning independently. citeturn459752search9turn459752search32

That's basically the formal version of your “whole mouth cavity / complete swallow” idea.

One place where I'd modify your interpretation is the cold-temperature observation. Cold making the sensation better does not particularly point toward astringency. Astringency is primarily drying/drawing/puckering, usually arising from interactions affecting lubrication and oral surfaces. citeturn459752search1 Cold itself activates thermosensory pathways, and some chemesthetic molecules recruit those same temperature-sensitive TRP channels. Oral mucosa expresses TRPM8, TRPA1, and TRPV1, which are involved both in chemical and thermal sensation. citeturn459752search48

So the fact that colder soda makes the sensation more satisfying may instead mean you're getting a stronger multimodal contrast:

cold stimulation

• carbonation/CO₂ irritation
• acid
• spice chemesthesis
• bitterness/astringency
• aroma.

The nervous system then integrates that into one experience rather than presenting you six independent gauges.

That also explains something you've noticed that I think is diagnostically useful: ginger ale doesn't do it for you. If simple TRPV1-style warmth were sufficient, ginger ale ought to get substantially closer. It apparently doesn't. Likewise Sprite has carbonation and acid but still feels incomplete. That argues against any one of:

carbonation alone,
acid alone,
TRPV1 warmth alone,
or bitterness alone.

Your bitters experiment already independently knocked out the last one.

So where I would now put our working model is:

The sensation you're trying to reproduce is probably a sustained, spatially distributed chemesthetic “body” during the swallow, generated by overlapping chemical and physical inputs rather than by a single taste.

And there are actually three especially promising pieces we can manipulate separately:

CO₂ bite / carbonic-acid chemesthesis
TRPA1 + TRPV1 stimulation, particularly piperine-type stimulation
mouthfeel/astringency/lingering sensory effects

Phosphoric acid remains potentially relevant, but it drops in importance somewhat. Interestingly, old beverage-industry material gives very different typical pH ranges for cola versus root beer—cola around the very acidic end, root beer considerably less acidic. citeturn655394search9 Yet you're telling me both can produce the desired deep sensation.

That's quite important.

If highly acidic cola and substantially less-acidic root beer can both give you the property while lemon-lime soda does not, then “phosphoric-acid burn” cannot be the universal explanation. It might strengthen cola's version of the effect, but something else must be capable of producing the same perceptual endpoint in root beer.

That pushes me more strongly toward a chemesthetic mixture architecture.

And therefore, rather than getting phosphoric acid next, I think your most informative kitchen experiment really is piperine.

Not because I know it'll taste good. Because experimentally it gives us the cleanest test of the mechanism. Piperine is one of the few ordinary-food substances for which there is direct published evidence of TRPV1/TRPA1 stimulation affecting the pharyngeal swallow itself. citeturn885824search3turn885824search35

If an extremely dilute black-pepper preparation adds precisely that deeper continuation without merely making the drink “peppery,” we've hit something substantial. Then we can work backward toward a palatable combination—perhaps piperine + bitters + a tiny amount of something like clove or wintergreen—rather than trying to make Sprite increasingly acidic until it fights back.

So yes: there is a wheel here already, although nobody seems to have published a paper called Why Coke Feels More Complete Going Down Than Sprite. The academic literature has nevertheless decomposed almost exactly the sensory machinery we'd need to answer that question, and it gives us a much better experimental map than we had one turn ago.

G: [Non-text content: thoughts; content_type=thoughts]

G: [Non-text content: reasoning_recap; content_type=reasoning_recap]

G: Yes. And the industry literature turned up something that I think is much more directly relevant than the spice hunt.

First, one biological correction. Your starting intuition is broadly right—cold can cause vasoconstriction and, with sufficient cooling, reduced sensation—but I wouldn't make vasoconstriction → numbness the main causal chain here. Oral cooling measurably changes mucosal circulation, and prolonged/intense cold reduces pain perception, but a cold drink is brief exposure. More importantly, the mouth and throat have cold-sensitive sensory receptors, so some of what feels like “cold/numbing” is an actively generated nerve signal, not tissue becoming anesthetized because its blood vessels contracted. citeturn0search0turn0search5

The industry has essentially invented the class of ingredient you're describing

Flavor companies call them physiological cooling agents or simply coolants. Menthol is the familiar natural example, but manufacturers developed compounds that produce the sensation of cooling while having little or no mint smell or flavor.

And here's the part that made me sit up: they explicitly formulate different compounds according to where in the mouth the cooling is perceived.

urlSymrise's technical brochure on physiological cooling agentsturn0search39 describes a family including WS-3, WS-5, WS-12, WS-23 and menthyl lactate. Their sensory testing says the compounds differ in location, intensity, onset and duration. citeturn0search39turn0search40

Translated from flavor-industry language:

They can make something feel cold without making it colder and without necessarily making it taste minty, and they can choose chemicals whose perceived cooling tends to occur in different regions.

That's almost absurdly close to the problem you've been describing.

For example, WS-3 is described by Symrise as virtually odorless and tasteless, with a clean, fairly immediate cooling effect perceived primarily at the roof of the mouth, back of the mouth and back of the tongue. citeturn0search39

That last geographical distinction matters enormously. You're repeatedly rejecting sensations that occur toward the front/top and looking for something that continues backward/downward into the swallow. Industry sensory scientists apparently care about essentially the same spatial distinction.

But “numbing” may be misleading us slightly

I think what you've called numbing may contain at least two perceptual components:

actual reduction of sensitivity and a cold/cool chemesthetic signal that your brain experiences as smoothness, fullness or subdued sensation.

Think about menthol. Your mouth isn't actually getting colder when you use a strong mint. Menthol activates cold-sensitive TRPM8 channels and essentially tells the nervous system “cold stimulus occurring here.” That's why coolants can create cooling without changing the physical temperature.

Now combine that with real cold soda.

You've got:

actual low temperature + cold-receptor activation + carbonation chemesthesis + whatever other chemical stimulation the beverage contains.

Those signals don't necessarily just add arithmetically. That's a plausible explanation for your observation that the satisfying property becomes particularly good when the beverage itself is cold.

And it makes me considerably less enthusiastic about astringency as the primary explanation. Astringency may still contribute body, but what you're describing doesn't sound principally like the drying/gripping/puckering sensation that sensory science usually means by astringency.

There's an especially interesting consequence for our experiment

I would now split our search into warming/pungent chemesthesis and cooling/numbing chemesthesis.

Piperine remains a very good probe for the former. It engages TRPV1/TRPA1 and demonstrably affects sensory processing during swallowing.

But WS-type physiological coolants are probably an even cleaner probe of the latter than mint, wintergreen, or clove, because they're deliberately engineered to minimize the flavor that would otherwise tell your brain “you're tasting mint.”

And there are large differences even within that family. Symrise reports, for example, that WS-5 is about 2.5 times the cooling intensity of WS-3 and describes its sensation as smooth and rounded; WS-12 is stronger and particularly persistent. Their controlled comparison at 30 ppm shows very different onset and longevity profiles among WS-3, WS-5, WS-12, WS-23, menthol, etc. citeturn0search39turn0search40

In ordinary English, the flavor industry has knobs for something like:

Where do you feel it?
How hard does it hit?
How quickly does it appear?
How long does it hang around?
Does it bring an identifiable flavor with it?

That's a far better conceptual framework for what you're describing than merely “how much cooling?”

And this changes how I interpret your cola/root-beer observation

I don't mean that Coke, Pepsi and root beer necessarily contain WS-3 or another synthetic coolant. I haven't found evidence establishing that, and we shouldn't reverse the reasoning.

Rather, the coolant literature establishes something much more useful: regional chemesthesis is real and deliberately measurable.

So the sensation you keep locating around the turn into the throat isn't an inconvenient subjective description we need to translate into “probably bitterness.” The location itself can be meaningful sensory data.

I'd currently picture your desired sensation something like this:

front of mouth: sweetness/flavor — already fine
middle: acid + carbonation + bitters — bitters improved this
rear oral cavity: still somewhat under-stimulated
pharyngeal/swallow transition: substantially under-stimulated
after-swallow: you'd like a subtle lingering completeness rather than obvious burning

That also explains beautifully why your taco test demoted capsaicin. It produced plenty of chemesthesis, but its spatial/temporal profile was wrong. More stimulation isn't the answer.

And your observation about ginger ale similarly becomes valuable rather than anecdotal noise. You're gradually making a sensory map.

So my present experimental ranking would actually become WS-3-type cooling → piperine → clove/eugenol → phosphoric-acid adjustment, with bitters retained because you've already established independently that they add something desirable.

I wouldn't suggest buying raw WS-3 powder and casually measuring it into drinks—the industry is testing these things at concentrations such as tens of parts per million, where kitchen measurement becomes lousy. If we pursue that branch, I'd look specifically for a food-grade, pre-diluted cooling flavor intended for beverages, with manufacturer dosing instructions.

But before buying anything, I'd like to do one more research pass specifically through flavor-house patents and technical literature for “back-of-mouth,” “throat,” “pharyngeal,” and delayed/lingering cooling agents. Now that we know the industry actually classifies coolants spatially, there may be a compound explicitly characterized as giving the back/throat effect you've spent several messages trying to describe. That would be considerably better than continuing to guess spices.

G: [Non-text content: thoughts; content_type=thoughts]

G: [Non-text content: reasoning_recap; content_type=reasoning_recap]

G: Yes, but with an interesting split: some of the WS family is chemically descended from the same general territory as mint/menthol, while some deliberately is not. So there isn't a spice containing “WS-23,” for example.

The closest kitchen equivalents

WS-3 is N-ethyl-p-menthane-3-carboxamide. The “p-menthane” part is the giveaway: it belongs to the structural world around menthol. It was developed specifically by modifying that chemistry to retain stimulation of the cold receptor TRPM8 while getting rid of much of menthol's mint odor/taste and volatility. citeturn0search1turn0search13

So its nearest ordinary-food ancestor is:

mint → menthol → TRPM8 cooling

But that's unfortunately not equivalent to saying that dilute mint will reproduce WS-3. Mint brings along carvone, menthone, menthol and a pile of other aromatics, so your brain gets MINT! along with the chemesthesis we're trying to isolate.

There is, however, a more interesting naturally occurring middle ground: menthyl lactate. It actually occurs naturally in peppermint oil, is described as only faintly minty and virtually tasteless, while producing prolonged cooling. citeturn0search1 That's much closer to the sort of “sensory seasoning without imposing a flavor” that we're after.

WS-23 is different. Its structure isn't based on menthol's p-menthane skeleton at all. It came out of the search for molecules that would activate this cooling machinery without needing the menthol structure. citeturn0search1 So there's no meaningful answer like “WS-23 is basically the active chemical in spice X.”

But there are several natural TRPM8 ingredients

The useful question therefore becomes less “what spice contains WS-3?” and more:

What ordinary food chemicals press the same neurological button?

There we have some choices. Menthol is the obvious one, but eucalyptol (1,8-cineole) also activates TRPM8. It's naturally present in eucalyptus but also occurs in culinary herbs including rosemary, sage, basil, cardamom and bay in varying amounts. TRPM8 research explicitly identifies eucalyptol alongside menthol and WS-3 as a cold-receptor agonist. citeturn0search10turn0search12

That gives me an experiment I like considerably better than putting peppermint into your soda:

cardamom.

Not because I expect cardamom magically to solve it, but because cardamom gives us some eucalyptol and other terpenes while reading as warm/aromatic rather than toothpaste-minty. And cardamom-like spice territory isn't terribly foreign to dark soft-drink flavor architecture.

I'd put rosemary below it because it announces itself pretty aggressively as rosemary.

More importantly, your description makes me rethink “numbing”

I think your word smoothing is actually better.

TRPM8 activation doesn't merely manufacture an extra sensation called COOL. There is evidence that menthol/TRPM8 stimulation can suppress the perception of other irritation. That's why the literature uses terms such as counterirritant, analgesic and anesthetic alongside cooling. citeturn0search0turn0search9

So imagine your cold Coke as a sensory mixture:

acid bite + CO₂ bite + cold + bitter compounds + aromatics + sweetness

and then cold-sensitive signaling simultaneously does something like:

“COLD” + damp down some of the rougher irritation.

That could subjectively become exactly the sort of dense-but-smooth continuous sensation you're describing.

Warm the Coke, and you don't merely remove the pleasant temperature. You alter that whole interaction. Acid and flavor compounds remain, carbonation behaves differently, volatile aromatics change, and the actual thermal activation of cold-sensitive nerves drops away.

Your wonderful description of warm cola as “polluted clear soda” therefore gives us a testable prediction. If the thing we're pursuing partly depends on cold/TRPM8 interacting with cola's other stimuli, warming it should expose the individual unpleasant or “dirty” components that cold was integrating into something coherent.

That's not established as the explanation, but it's considerably more mechanistically respectable now than when we started with caramel.

And here's something particularly important about your idea of amplification

TRPM8 is activated by both cold temperature and cooling chemicals. Menthol doesn't merely create an independent fake-cold channel somewhere else; chemical ligands change the behavior of the receptor that also responds to actual temperature. Research describes menthol as shifting the thermal activation of TRPM8. citeturn0search12

So your proposed interaction—

actual cold + chemical cooling → disproportionately satisfying effect

—is biologically quite reasonable.

That means we shouldn't necessarily be hunting for something that numbs the throat. We may be looking for something that makes the existing cold beverage recruit more of the cold-sensitive sensory system in the particular region you care about, while simultaneously changing how the other irritation is perceived.

And that brings me back to your original question about common ingredients. I'd now divide the kitchen experiments like this:

I particularly like peppermint as a diagnostic rather than an ingredient. We don't care if the experimental glass tastes stupid. If a tiny amount suddenly fills in precisely the missing anatomical region when the drink is cold, we've learned that TRPM8 is probably involved. Then we go hunting for a flavor-neutral way to achieve it.

And if peppermint doesn't fill that region but piperine does, we've learned something equally valuable and stop chasing coolants.

If neither does it, I'd move clove/eugenol substantially upward. Eugenol is interesting because the perceptual result is harder to classify neatly as “hot” or “cold,” and your repeated use of smoothing/blurring/fullness is making me increasingly suspicious that the desired endpoint is counterirritation or sensory modulation, rather than simply another sensation layered onto the soda.

One other thing: don't bother trying eucalyptus oil as a shortcut to eucalyptol. Essential oils aren't interchangeable with culinary extracts and aren't something I'd improvise-dose into drinks. Cardamom or an ordinary food flavoring gives us a much saner kitchen probe.

At this point, black pepper + peppermint/cardamom + clove gives us three meaningfully different neurological probes using grocery-store ingredients. That's considerably more informative than trying twenty spices until one happens to taste good.

G: [Non-text content: thoughts; content_type=thoughts]

G: [Non-text content: reasoning_recap; content_type=reasoning_recap]

G: Yes. For your dad, I would not lead with TRPM8, TRPA1, chemesthesis, or our adventures in phosphoric acid. 😄 I'd give him the experimental question in ordinary sensory terms:

“I'm trying to figure out what makes some sodas feel fuller when you swallow them. Not the flavor exactly—more the physical sensation in the mouth and throat. Coke and root beer seem to have something that Sprite or Mountain Dew don't. I'm trying tiny amounts of different normal food ingredients to see which part of that sensation they reproduce. So when you try this, don't just tell me whether you like the flavor. Tell me where you feel it—front of the mouth, back of the mouth, throat—and whether the swallow feels fuller, smoother, sharper, or basically unchanged.”

That's probably enough. If he asks what the hell you're talking about, then:

“Taste is sweet, sour, bitter and so forth. But food also creates physical chemical sensations—pepper burns, mint cools, carbonation bites, some things dry your mouth. That's called chemesthesis. I'm trying to isolate one of those sensations that I notice in darker sodas.”

That also keeps him from inadvertently answering the wrong question. “This tastes worse” is useful information, but it's separate from “this filled in the missing throat sensation.”

Basil is feasible—but it's a messier experiment than I implied

I checked the chemistry more carefully, and I'd downgrade basil from my earlier suggestion as a clean TRPM8 test.

Basil can contain 1,8-cineole/eucalyptol, and 1,8-cineole really does activate the cold-sensitive TRPM8 receptor. Interestingly, it also inhibits human TRPA1 under experimental conditions and reduced irritation caused by TRPA1 agonists. That combination—cooling signaling plus suppression of another irritation pathway—makes it genuinely interesting given your “smoothing/blurring” description. citeturn0search0turn0search1

The problem is that basil isn't chemically consistent enough to be a reliable delivery system for it.

Different basil cultivars are basically different chemical cocktails. Researchers identify basil chemotypes dominated by linalool, estragole, eugenol, methyl cinnamate, and sometimes eucalyptol. One analysis even found some cultivars that qualified as eucalyptol chemotypes, while other studies find eucalyptol as only a small secondary constituent. citeturn0search5turn0search11

That's actually consistent with your personal reaction to basil: “I like the flavor intellectually, but physiologically it doesn't do much for me.” There's no reason ordinary culinary basil necessarily contains enough cineole to give you an obvious cooling/smoothing chemesthetic effect.

So I wouldn't expect:

basil leaf → obvious TRPM8 stimulation → missing throat sensation.

I'd expect more like:

basil leaf → lots of basil aromatics + perhaps a small TRPM8 contribution buried underneath them.

But since you presumably have basil available...

It's cheap enough to test.

I'd bruise a few fresh leaves, steep them in a small amount of hot water, let that cool completely, strain it, and use drops of the resulting basil infusion rather than putting leaves into the soda.

That gives us a funny advantage in your case. Since you already find basil physiologically rather neutral, its recognizable flavor may be less distracting to you than peppermint would be. If a sub-flavor-threshold amount nevertheless changes the location or smoothness of the swallow, that's interesting.

But I wouldn't interpret a negative result as evidence against TRPM8. There's simply too much variation in basil chemistry. citeturn0search2turn0search5

And I'd still put black pepper considerably ahead of it experimentally. Pepper is testing a different pathway, and piperine is a much more dependable defining constituent of the ingredient.

One last thought about recruiting your dad: don't tell him what each ingredient is supposed to do before he drinks it. Tell him afterward. Otherwise you've accidentally turned your second palate into a very suggestible measuring instrument. Give him the little where/fullness/smoothness/sharpness rubric above, have him describe it, then say, “Okay, that one was pepper.” That's about as close as the kitchen table needs to get to blinded sensory testing.