Your taste feels like the most personal thing about you, an inner verdict on what is good. Much of it is made rather than found: shaped by what you have been exposed to, by the social position you occupy, and by the way a choice happens to be put to you. What you like is a lot less yours than it feels.
Two research traditions come at what is 'taste' from opposite ends, and both land on the same uncomfortable place: it is less personal than it feels.
The sociological tradition treats taste as social positioning. Pierre Bourdieu's argument was that taste is not innocent aesthetics but a marker of class: the preferences you acquire growing up in a particular place in society, your habitus, quietly encode and reproduce that position, so that the line between highbrow, middlebrow, and popular taste is really a class boundary in disguise (Bourdieu, 1984). Liking opera rather than reality television is, on this view, partly a way of signalling and securing where you stand. The framework has been updated in one important way. The high-status signal is no longer narrow snobbery but breadth: the modern marker of cultural privilege is the omnivore, the person who moves comfortably from classical music to street food to hip-hop, and studies across many countries find that this omnivorousness has itself become a sign of status (Peterson and Kern, 1996).
The psychological tradition looks at how a preference forms in the first place, and its central finding is unsettling in its simplicity. Merely being exposed to something, repeatedly, makes you like it more, and this happens even when you cannot remember having seen it and even when the exposures are too brief to register consciously (Zajonc, 1968). Familiarity itself, with no reward, no argument, no reason attached, breeds liking. It is one of the more reliable effects in the field, confirmed across a large meta-analysis of hundreds of studies (Bornstein, 1989). But it is not limitless. Liking rises with familiarity and then falls with too much of it, tracing an inverted U: the sweet spot is moderate novelty, familiar enough to feel safe, fresh enough not to bore, and past that point exposure starts to breed tedium rather than fondness (Berlyne, 1970).
You can see the same mechanism at the scale of a whole culture, in who becomes famous. Celebrity is largely exposure made public. When a handful of broadcasters, film, television, and radio, pointed everyone at the same faces, the country got the same repeated exposure and shared the same stars; now that attention has splintered across personalized feeds (Webster and Ksiazek, 2012), fame has fragmented with it, and there are people with tens of millions of followers whom much of the public has never heard of. Change the pattern of exposure and you change the map of who is liked. It is an illustration rather than proof, since real fame also rides on algorithms and network effects, but it shows the mechanism working far beyond the lab.
The deepest finding is that preferences are often constructed in the moment rather than revealed. Economics has long assumed that when you choose, you are reading off a stable internal ranking. But people reliably reverse themselves depending on how the question is framed: offered a safe small-payoff gamble and a risky large-payoff one, most people choose the safe bet, yet when asked what each is worth, they price the risky one higher, a straight contradiction of any fixed underlying preference (Lichtenstein and Slovic, 1971). The preference was not retrieved. It was assembled on the spot, and a different question built a different answer, which is why later work speaks of the construction of preference rather than its discovery.
Each tradition has a live fault line worth stating plainly, and the one about taste is no different.
Bourdieu's class-determinism is widely judged overdrawn for the present. His picture came from 1960s and 1970s France, and taste cultures since have grown far more fluid: the omnivore finding is itself partly a rebuttal, since the neat highbrow-equals-elite mapping broke down (Peterson and Kern, 1996). Class still shapes taste, but far less rigidly than Distinction implied, and the mechanism keeps shifting as culture fragments.
Mere-exposure is robust but bounded, and the boundaries matter for anyone tempted to treat it as a magic lever. The effect is strongest for novel stimuli, brief exposures, and low prior familiarity, and it shrinks or reverses with overexposure, with stimuli people already disliked, and over long delays (Bornstein, 1989; Berlyne, 1970). "Familiarity breeds liking" is true within a window and false outside it.
The construction-of-preference finding raises a genuinely hard question: if choices reverse with framing, is there any stable "true" preference underneath, or is it construction all the way down? Reasonable researchers disagree. The reversals are solid and replicated; what they imply about whether authentic preferences exist at all is not settled.
Much of the preference-reversal work uses laboratory gambles, and how far it generalises to weightier real-world choices is debated, though framing effects show up well beyond the lab. Bourdieu's evidence is tied to a particular time and country. And most of this research is Western, so the specific taste hierarchies and even the strength of the omnivore effect may not travel.
Is there a stable preference beneath the constructed one, or not? How do status-through-taste dynamics evolve as culture splinters into countless online niches, each with its own hierarchy? And how far do these patterns, built largely on Western samples, hold across cultures?
The usable core: taste is substantially manufactured, by exposure, by social position, and by framing, which makes it movable and therefore both a tool and a temptation. Familiarity can be earned honestly or manufactured invisibly; a choice can be framed to help someone decide or to steer them past their own judgement. The line, as ever in this series, is whether the person would feel respected if they saw what you were doing.
Mere-exposure is the quiet engine under most advertising: repetition builds familiarity, familiarity builds liking, and the customer rarely traces the warmth back to the ads (Zajonc, 1968). That is real and usable, and it comes with a built-in ethical check that also happens to be a practical one, the inverted U. Push exposure too hard and liking curdles into irritation and wear-out (Berlyne, 1970), so the same evidence that says "repeat" also says "vary the execution and do not saturate." Taste-as-signal is the other lever: people choose brands that fit the taste-identity they want to project, so for high-status positioning, breadth and cultural range now signal more than exclusivity (Peterson and Kern, 1996). And because preference is constructed, how you present options, the default, the order, the framing, often shapes the choice more than any pre-existing preference does. The manipulation line is the invisible one: manufacturing a preference the person could never detect, rather than earning a liking they would endorse.
The machinery of name recognition is mere-exposure in plain sight: repeated exposure to a name, a face, a slogan builds familiarity and, with it, a warmth voters mistake for judgement (Zajonc, 1968). The inverted U is the warning label, over-saturate and the familiar candidate becomes the tiresome one (Berlyne, 1970). Framing does the rest, since the same position draws different support depending on how it is put, which is the constructed-preference finding meeting the party-cue effects of L3-12. All of this works, and the responsible reading is the usual one: familiarity and framing that help people engage with a real choice are fair game, while saturation and framings designed to slip past judgement spend trust once they are seen.
Two implications follow. Because preferences are constructed, the presentation of a public choice is never neutral, defaults, ordering, and framing will shape decisions whether or not anyone intended them to, so the honest move is to design that presentation deliberately and transparently rather than pretend a "neutral" option exists (this is the ethical heart of choice architecture, L5). And because familiarity itself builds acceptance, public-information campaigns can shift attitudes through repetition, but they hit the same inverted-U ceiling, where repetition tips into fatigue and backlash. Design for the sweet spot, not the saturation.
Three habits. First, notice how much of what you "like" is simply what you have been exposed to, and treat a brand-new dislike with some suspicion, it may just be unfamiliarity, which time can turn into fondness. Second, aim for the moderate-novelty sweet spot in what you make and what you consume: too strange to grip, too familiar to bore, and the good stuff usually sits between. Third, when your preference flips depending on how a choice is framed, take it as a signal that there may be no fixed "true" answer waiting inside you, and decide from your values instead of from whichever framing you happened to meet first.
.png)