[dyauspitr]

On a case by case basis that generally aligns with common sense. Most people can instantly recognize a hook and knife are very different from adding a solvent to food. Drug laws are a good analog for this. I don’t think the universe of possibilities in this space is prohibitively large.

You can probably begin by broadly calling petroleum and petroleum products artificial.

[SPICLK2]

There's a rich heritage and history of solvent-based foods. Vanilla essence, sloe gin, etc.

[dyauspitr]

Ethanol would probably be classified as a natural solvent. The edge cases fall off very quickly, this can definitely be done on case by case basis without introducing onerous bureaucracy.

[SPICLK2]

Even if it were distilled from petroleum?

[dyauspitr]

Ethanol is not distilled from petroleum. Industrially it is produced by distilling plant sugars and starches.

[SPICLK2]

It can be distilled from petroleum, and there is the key distinction that wasn't answered - are "natural" ingredients ones that could be made by "natural" (I'm assuming that "biochemical" is meant here) processes, or are they ones that are made by "natural" processes? Or is it just petroleum that is the problem?

Where does salt fall here?

Why isn't petroleum natural, when it is plant-based?

[dyauspitr]

A good starting point would be to broadly classify petroleum and its products as artificial.