Taste Machine
If a machine could adjust your aesthetic responses to match what the most refined critics converge on, would you use it? Should you?
This thought experiment, developed in the philosophy of aesthetics through the 1980s and 1990s, puts pressure on the tension between aesthetic subjectivism and aesthetic realism. It asks whether taste is a personal response to be respected or a faculty that can be trained, calibrated, and improved.
Goldman, A. H. (1995). Aesthetic Value. Westview Press.
The scenario
The machine works by mapping your current aesthetic responses and comparing them against the convergent judgments of the most experienced, informed, and sensitive critics across history. Where your responses diverge, it adjusts your sensibility. After treatment, you find Coltrane's late work revelatory rather than difficult. You notice structural complexity in novels you previously found cold. You stop finding certain things beautiful that expert consensus considers meretricious.
Your preferences have changed. But they've changed to match what experts converge on, not to match anything you discovered yourself.
The question isn't whether the machine is possible. It's whether you'd want it, and whether wanting it would be rational.
Subjectivism versus realism
Aesthetic subjectivism holds that taste is constitutively personal. There is no fact about which music is better, only facts about which music you prefer. On this view, calibrating your responses to expert consensus isn't improving your taste; it's replacing it. The machine doesn't give you better preferences; it gives you different ones, and different isn't better if better has no meaning here.
Aesthetic realism holds that some aesthetic responses are more accurate than others. Artworks have properties, depth, coherence, expressive power, and some people are better at perceiving them than others. On this view, the machine would be doing something genuinely useful: training your perception the way a music teacher trains your ear, except faster and more reliably.
The difficulty is that neither view sits comfortably with how we actually talk about taste. We say "you'd love this if you gave it a chance" and mean it as more than prediction. We think some people are better judges of wine than others. But we also recoil from the idea that one person's aesthetic preferences are simply wrong in the way that a factual belief is wrong.
What the machine reveals
The machine is useful not as a real proposal but as a diagnostic. How you react to it tells you something about your underlying commitments.
If you'd use it without hesitation, you probably think of aesthetic taste as a skill, something to be developed toward a target. The target might be expert consensus, or the responses of an idealized critic, or your own future more-informed self. Improvement is possible and worth pursuing.
If you'd refuse, you probably think your current responses are yours in a way that calibrated responses wouldn't be. The point of having aesthetic preferences isn't to be correct; it's to engage with the world from your own perspective. A borrowed sensibility, even a better one by external standards, isn't an improvement. It's a substitution.
Most people's reactions are mixed. They'd want to appreciate things they currently can't appreciate. They wouldn't want to stop appreciating things they currently love. The machine forces a question about which of those is the real goal of aesthetic development.
Discussion questions
- Would a machine that perfectly replicated the taste of your favorite meal satisfy you?
- Is the value of a great meal in the experience or in what produces it?
- Does your answer change depending on whether you know the machine is real?
Take it to the dinner table.
Get 3 thought experiments for memorable conversations, designed for dinner, with friends, at events, or anywhere small talk has gone on too long.