Some interesting experiments have been reported by Yanyun Zhou and Chi-Shing Tse (The Taste of Emotion: Metaphoric Association Between Taste Words and Emotion/Emotion-Laden Words). They were conditioning test subjects with this or that taste and tested how that influenced their attitudes in various aspects of behavior associated with the "linguistic" end of the taste-emotion correlation.
The article also contains some references for the experimental support of the opposite direction of the correlation.
Fortunately, they were looking not just at words denoting emotions, but also at various words strongly emotionally laden denoting all kinds of other things. It turns out that the "emotional charge" of a word is often more important, for the purpose of its ability to summon a certain taste, or to be summoned by a certain taste, than what the word means. For an English based example, the word "sweet" can be pretty sweet (in either its emotional or taste sense), but these researchers found out that "honey" is way sweeter, at least on the taste end - and no, they weren't using any actual honey in their experiments.
It might appear from such literature that the taste defines or stimulates the emotion, while the opposite "causation" is (well possible but) significantly harder to demonstrate. Does it mean that it's taste which causes emotions in individuals rather than vice versa? The authors don't think so and neither do I. The imbalance of evidence probably reflects just the practicalities of test setup. It is way more difficult to condition test subjects with emotions (while controlling for any additional linguistic and situational context of the experiment!) than to condition them for a particular taste.
(The article summarizes and references some findings even on the taste-trait axis (mentioned in the OP), for example between particular tastes and tendency to perceive others as "agreeable", and/or being perceived by others as "agreeable".)
This excursion to psychology was supposed to suggest that within linguistics, the right subfield to analyze the taste-emotion correlation might need to be pragmatics rather than semantics.
Changing gears from psychology to neurology, I want to get one phenomenon out of our way and that's various synesthesiae, most of all the lexical-gustatory synesthesia. That's a rare condition when a person reliably feels a certain taste when stimulated by a particular word. It's not interesting to us, because the specific connections tend to be way more arbitrary than a word like "sweet" causing a sensation of sweet taste.
To further delimit the question, I also want to draw the line so that temperature, texture, viscosity, etc. (of food) do not count as "taste" factors, not even when connected to the perceived taste through a physiological or pathological mechanism.
Being more of a linguist than a gourmet, I may be biased now, but I suspect that the gustatory vocabulary is woefully limited with its just five to seven basic tastes: bitter, sour, salty, sweet, savory, tanniny, and punguent. (We could come up with an additional "taste word" just by naming anything edible and claiming that its taste is distinct from other tastes, but at that point we risk bringing in lots of associations of that particular edible thing beyond its taste. Also, we would run into a problem called the dual patterning of the world, namely that different languages would be expected to split the space of various taste combinations along arbitrary, very different boundaries into their own "taste concepts" - especially if the languages are geographically and culturally separated from each other, so we couldn't even ask the original question in a language independent way; tastes would no longer be a language independent concept.)
Those up to seven basic tastes each have various synonyms and I might need some help from a native English speaker to see which ones of the seven basic tastes have a well defined emotional correlate in English. I suppose that it's just these:
Of which, the first two or three, although entirely distinct as tastes, correlate to roughly the same emotion in English. (OK, not exactly the same. We could associate the sour taste with envy, bitter taste with regret, salty taste with anger. But the semantic overlaps on the emotion side seem to be strong.)
It seems that for some speakers of English there might be just a single axis: sweet versus anti-sweet, agreeable vs. disagreeable, positive emotion vs. a negative emotion, or a slightly larger vocabulary of up to seven items.
Other languages can have their differences in the detail. In Slovak, for example, "horký" might correspond to the English "bitter" in the taste space, and to "painful" in the emotion space. In closely related Czech, there might rather be a connection between "punguent" and "painful" instead ("pálivý/palčivý" - I'm cheating with a pair of related words providing the connection rather than a single word), although even all this is debatable because more than one basic taste could be used to characterize the same emotion even in the same language - and of course vice versa.
There's obviously some variation across languages, including closely related languages, despite the combinatorially small playground of basic tastes, and also despite the fact that some of the basic tastes, are "emotionally unoccupied" in some languages. I'm afraid that this means that the details of the mapping aren't stable across millennia and that we don't really know how those metaphoric mappings first arose in prehistory.
The most universal mapping seems to be the one between "sweet" and "agreeableness". I can find it wherever I look. (Counterexample languages, anyone?) I am inclined to think that this could have originated from glucose based regulation of human metabolism long before the neolithic agricultural revolution when safe food was scarce. (English "glucose" actually comes from Greek γλυκoς, meaning "sweet".) I wonder whether an obesity pandemic might amass the power eventually to reverse the fundamental axis, leading to taste-emotion associations like "sweet" - "regret" in some languages.