Salt mostly works by taking something away

Everyone who cooks knows the moment. The pot has been simmering for an hour, it tastes flat and slightly wrong, you add salt, and suddenly it tastes like the thing it was supposed to be. The tomato is more tomato. The stock has depth it did not have thirty seconds ago. Nothing was added except sodium chloride, and yet the change is not in the saltiness.

In 1997 Paul Breslin and Gary Beauchamp at the Monell Chemical Senses Center published a one-page paper in Nature with a title that gives the whole answer away: salt enhances flavour by suppressing bitterness. ¹ The enhancement people describe is largely a removal. Take the bitterness out of a mixture and everything else in it gets louder without anything else changing.

Two years earlier the same pair had done the careful version of the experiment. They took six salts and six bitter compounds, mixed them in every combination at four concentrations, and had panellists rate bitterness and saltiness separately by magnitude estimation. In most cases the bitterness was suppressed by salt, though how much varied by which bitter compound was involved. Urea gave way especially easily. ²

The part that makes it useful

The finding that matters at the stove is the asymmetry. Bitterness suppression, they reported, was not accompanied by an equivalent reciprocal suppression of saltiness, and at high concentrations of sodium chloride no bitter compound affected the saltiness at all. ²

Read that again as a cook rather than a psychophysicist. Salt beats bitter, and bitter does not beat back. That is why salt is the correction for an over-reduced sauce, a scorched onion, a bitter green, a coffee-dark stock, and why the reverse operation does not exist. There is no ingredient you add to make something less salty. You dilute, or you start again.

It is also the whole argument for salting in stages instead of at the end. The bitterness you are suppressing is being generated throughout the cook, so a single correction at the finish is asking one addition to do work that was needed an hour ago.

Where the suppression actually happens

For twenty-five years the mechanism was an open question: is sodium interfering with the bitter receptors themselves, or is the brain doing the mixing further up? In 2024 a group tested it directly, expressing four of the roughly twenty-five human bitter taste receptors in cells and running them against their usual agonists with and without sodium chloride. ³

The answer was untidy in an interesting way. Sodium affected only particular receptor and compound pairs. TAS2R16 showed reduced signalling in the presence of salt; other bitter responses were unchanged whatever the sodium concentration. So part of the suppression happens at the receptor, and for the rest the authors conclude that central effects probably dominate. ³ Some of what salt does to bitterness happens on your tongue and some of it happens in your head, and the split depends on which bitter thing you are tasting.

There is no salty receptor

The other assumption worth losing is that salt has a receptor the way sweet and bitter do. It has two systems, and they want opposite things.

Low concentrations of sodium are attractive, especially to an animal that is short of sodium. High concentrations are aversive. In 2010 a team at Columbia showed these are not two ends of one scale but two separate detectors. The attractive one is selective for sodium specifically and is blocked by amiloride; the aversive one is a non-selective detector that responds to a wide range of salts. They found the attractive pathway has its own dedicated population of taste cells expressing the epithelial sodium channel, ENaC, and when they engineered mice lacking ENaC in those cells the animals showed a complete loss of salt attraction and of sodium taste responses. ⁴

This is the cleanest explanation of why salt substitutes disappoint. Potassium chloride cannot reach the sodium-specific channel that produces the pleasant half of salt taste, and it still registers on the non-selective aversive one. It is telling that in the Monell bitterness experiments potassium chloride appears twice: once in the list of salts being tested, and once in the list of bitter compounds. ² The literature had it filed under both headings, because it belongs under both.

The savoury half

Salt is only half of what people mean by savoury. The other half was named in Tokyo in 1908 by Kikunae Ikeda, who had been working on the broth made from kombu and dried bonito and had become convinced that what he was tasting was not covered by sweet, sour, salty and bitter.

It took the better part of a century for the receptor work to catch up with him, and the note of apology in that phrasing has always struck me. For convenience. He is proposing a fifth basic taste and asking to be allowed a word for it.

In a kitchen the two halves are usually delivered together, and that is what a cook means by savoury: sodium arriving with free glutamate. Parmesan, anchovy, soy sauce, fish sauce, reduced tomato, aged and fermented things generally. None of that is salt doing the work alone, though salt is doing the part nobody notices, which is clearing the bitterness out of the way so the glutamate has a clean run.

The practical version of all this is unglamorous. Salt early and in layers rather than once at the end. Expect it to fix bitterness and expect it not to fix much else. And accept that you cannot take it back out, because the tongue has a dedicated channel for noticing sodium and no channel at all for forgiving too much of it.