日本語版

Research note 023 — 2026-09-02

A Map of Techniques That Add Bitterness — Heat, Smoke, Char

In A World Map of Debittering, we laid out how kitchens around the world have removed bitterness. This note is the mirror image of that one. Does cooking thatadds bitterness form a single technique family? That is the question posed in the hypothesis collection box (H9), and this note looks for an answer by holding it up against material on roasting, smoking, okoge (rice crust), and caramel. To state the conclusion first:evidence stands for three routes — heat, smoke, and char — but the three do not work in the same direction, and bundling them together into a single "family" is not yet possible.

What this page establishes

• Coffee has been measured, by trained assessors, to have its bitterness sensory score rise as roasting gets deeper.
• Smoke can either strengthen or weaken bitterness: separate studies have measured it correlating with bitterness in wine and reducing bitterness in tea.
• Okoge and char are described in most sources using words for savory aroma rather than bitterness, and some accounts explicitly deny bitterness altogether.
• The conditions under which burnt sugar (caramel) produces bitterness are discussed in modern confectionery explanations, but a 19th-century recipe makes no mention of bitterness.

Route 1 — Heat: the deeper the roast, the more bitter

Of the three routes, heat is the one most straightforwardly described as「足す」 (in translation: "adding"). In a study where trained assessors rated coffee by roast degree, the bitterness score rose significantly as the bean color deepened (as the Agtron value fell).

But heat does not only increase the 「量」 (in translation: "amount") of bitterness. In the roasting of hojicha, gallate-type catechins that activate the bitter receptor (such as EGCG) decrease monotonically as roasting deepens, while caffeine barely decreases — heatrewrites the makeup of bitterness.

The claim that 「炭火焙煎はまろやか」 (in translation: "charcoal roasting is mellow") can also run opposite to what is actually measured. In a comparison of yellow tea roasted by three methods — charcoal, electric, and drum — the electronic tongue's bitterness output was highest for charcoal.

Route 2 — Smoke: it can strengthen bitterness, or weaken it

Smoke is the hardest to handle of the three routes. In wine, in samples made from grapes exposed to wildfire smoke, the concentration of phenolic glycoconjugates correlated with scores for bitterness, ashiness, sourness, and other attributes.

Yet in lapsang souchong (a tea withered with pine smoke), the pine-smoke process is reported tosignificantly reduce the tea's bitterness, with smoke-derived compounds such as guaiacol antagonistically inhibiting the bitter receptor TAS2R14. The same「煙」 (in translation: "smoke") adds bitterness on one side and takes it away on the other.

One more thing needs care. A sensory study that organized the vocabulary of smoke flavor explicitly states that bitter is included among the original proposed terms, yet a bitterness score never once appears in the actual evaluation table. The vocabulary of smoke is a vocabulary ofsmell, entered in a different ledger from the bitterness of taste.

Route 3 — Char: okoge is spoken of as aroma, not bitterness

This is the place where this note's assumptions were most overturned. Guoba, bao zai fan, pasta all'assassina, socarrat, tahdig, nurungji, okoge, and yuto — material on「おこげ」 (in translation: "scorched-rice crust") gathered across six languages — speaks, in almost every source, using words for savory aroma (焦香, roasted, برشته, 구수한),words of smell, and some sources even explicitly deny bitterness.

For the socarrat at the bottom of paella, even when measured by a trained panel, "roasted" rose significantly while "burnt" was kept lower than for white rice.

Among the Japanese-language sources, the only one that used the word「苦味」 (in translation: "bitterness") was a single personal blog explaining word meanings — and even there, it appeared in the context of describing theboundary between savoriness and bitterness:「焦がしすぎると苦味が出て不快になる」 (in translation: "if it is over-charred, bitterness appears and becomes unpleasant").

Is the technique of charring aiming at bitterness?

The gastrique — deliberately charring sugar — can be traced back to the classics of French cuisine. Escoffier's (1903) sauce bigarade describes the process as「砂糖をカラメル状にして酢に溶かす」 (in translation: "caramelize the sugar and dissolve it in vinegar"), but it never mentions bitterness even once, and does not even use the word "gastrique".

It is only in modern sources, such as explanations from culinary schools, that bitterness gets put into words as a goal of charring.

On the industrial side of caramel coloring, by contrast, bitterness has been identified and quantified as a 「副産物」 (in translation: "by-product"). From Class III coloring made by the ammonia process, bitterness-guided fractionation isolated four major bitter compounds, and their thresholds were reported.

What the map lets us read

• Heat adds bitterness while also rewriting its makeup (gallate-type catechins decrease, while caffeine remains).
• Smoke reverses direction depending on the food. The vocabulary of smoking sits in the ledger of smell, and a bitterness score is not taken.
• Char, on the cooking side, is spoken of as savoriness; bitterness is the word for the boundary (over-charring). The classics do not write bitterness as the goal of charring.

So, within the range of material that can currently be confirmed,「苦味を付与する調理」 (in translation: "cooking that imparts bitterness") is less a single technique family than three separate routes. Whether there is a shared design principle that binds them together — for instance, whether the goal is not to「苦味を足す」 (in translation: "add bitterness") but to「苦味の後ろに香りを足す」 (in translation: "add aroma behind the bitterness") — is left as a question in the hypothesis collection box. The discipline of writing astringency and bitterness, and savoriness and bitterness, as separate things — a discipline of this research lab — is what decided the outline of this map.

Frequently asked

Does coffee get more bitter with a darker roast?

In an experiment using Ethiopian arabica beans, 12 trained assessors rated cold brew by roast degree (the paper's methods do not state the range of the scale), and it is reported that the bitterness score rose significantly as the bean color deepened (as roasting got darker). This result held in the same direction for both ultra-high-pressure and conventional extraction. However, this was an experiment using a single origin and a single lot, and generalizing to extraction methods other than cold brew falls outside the scope of the original study. Reproduction with other beans or roasting machines has not been confirmed within the range of this material.

Does smoking make food more bitter?

Which direction smoke pushes bitterness in splits by the target food. In wine made from grapes exposed to wildfire smoke, the concentration of phenolic glycoconjugates was reported to correlate with scores for unpleasant tastes including bitterness. In tea withered with pine smoke (lapsang souchong), by contrast, there is a report that the same smoke process significantly reduced bitterness. In a sensory study that organized the vocabulary of smoke flavor, bitter was included among the original proposed terms, yet a bitterness score was never once recorded in the actual evaluation table — the vocabulary of smoking is built around smell. So whether smoking strengthens or weakens bitterness needs to be checked separately for each food.

Is okoge (scorched rice) bitter?

Most sources describe okoge using words for savoriness rather than bitterness. An article on Chinese guoba describes the desirable char as "aroma, not bitterness," distinguishing it from the failed form charred to black carbon. English-language articles likewise describe the charred crust of Italy's Assassina as "Never bitter, always bold." In an experiment that measured the socarrat forming at the bottom of paella with a trained panel, what was measured was "roasted" and "burnt" — bitterness itself was not among the evaluation items. Among Japanese-language sources, only a single personal blog used the word for bitterness, and even there it was written about the case of over-charring.

What is not known

• Whether there is a shared mechanism that binds the three routes of heat, smoke, and char together as a single technique family has not yet been confirmed.
• How smoke-derived phenolics act on the human bitter receptor could not be confirmed from any source beyond the abstract-level report on lapsang souchong.
• The thresholds of bitter compounds produced by caramelization are limited to a single report on industrial coloring, and no measurement under conditions closer to home cooking was found.
• No experiment directly measuring okoge's bitterness with a trained panel was found within the range of this material.
• Whether the saying 「炭火はまろやか」 (in translation: "charcoal is mellow") holds for other foods besides yellow tea has not been confirmed.