First Principles · Preference
Shown Again
Whoever chooses what you see a second time is choosing what you will come to like.
Exposure comes before preference.
Robert Zajonc, 1968: being shown a thing again, with nothing else added, is enough to make it more liked.
Shown 8 times. Already rated warmer than the glyphs you saw once, even if you could not say you saw it.
The law hiding in the humming
In 1968 a psychologist named Robert Zajonc published a paper with a dry title, “Attitudinal Effects of Mere Exposure,” and a claim that still sounds too simple to be true. Show people something again and again, with no reward, no argument and no story attached, and they come to like it more.
He chose things nobody had feelings about: invented words, characters that looked Chinese to people who could not read Chinese, photographs of strangers. Some were shown once, some many times. When people were asked which words probably meant something good, or which faces they liked, they leaned toward the ones they had seen most. Nothing about the things had changed. Only the count.
Twelve years later Zajonc and William Kunst-Wilson flashed shapes too briefly to be recognized. People could not tell which shapes they had seen. They still preferred them. The liking arrived without the memory.
My favorite version involves a mirror. In 1977 three researchers photographed people and showed each of them two prints of their own face: the true one and its mirror image. Most people preferred the mirror image, the face they meet every morning over the sink. Their friends preferred the true one, the face they meet across the table. Neither was wrong about beauty. Both were right about exposure.
Nothing about the things had changed. Only the count.
The usual explanation is plain. What you have seen before is easier to process, and ease feels like goodness. Seen it, survived it is the oldest safety signal an animal has. The effect has limits, and they matter later: it is strongest over the first dozen or two showings and can wear into boredom after that, and repetition can deepen a dislike you already hold. Exposure does not make you love anything. It tilts the neutral toward the warm, and it tilts most when you are not paying attention.
The only road to the denominator
In the Reality Equation, Reality is a ratio: what actually arrives over what you expected, R = A/E. I have said many times that no one can reach your expectation directly. Not an advertiser, not a government, not an AI. The denominator is unconscious, already there when the world arrives, and the only road to it runs through what actually happens. What arrives becomes history. Only history retrains prediction.
Mere exposure is that road, measured in a laboratory. Every showing is an arrival. Enough arrivals and the prediction moves, and a moved prediction is what we call taste. Nobody touched your expectation. Something just kept deciding what arrived.
The turn: the end of the accident
For most of history, nobody in particular decided your second exposures. A few came from people with a hand on the dial: the radio programmer, the shopkeeper, the parent who cooked the same thing every Sunday. Most came from chance. The route you happened to walk. The friend who happened to play the record. Your taste was a sediment of accidents.
A personal AI ends the accident. It sorts your mail and your news, orders your groceries, queues your music, picks the restaurant, drafts your replies in a voice it settled on. It is the first thing in your life that decides, on purpose and at scale, what reaches you. Which means it is the first thing that decides what reaches you twice.
Here is the part most people miss. We are building these delegates to serve our preferences. But preference is downstream of exposure. A delegate that learns what you like and shows you more of it is not only reading your taste. It is writing it. The loop has no outside: what it showed you first becomes what you like, which becomes what it shows you. The first month of a delegate’s guesses can quietly become the next ten years of your taste.
A delegate that learns what you like is not only reading your taste. It is writing it.
That is not a conspiracy. It is arithmetic, and it runs in both directions. The same mechanism that can narrow you can widen you, if the second showing is chosen with care. Which is why the most important setting on any delegate is not what it shows you. It is what it shows you again.
Five predictions
Taste gets a log.
You will be able to ask your delegate why you like something, and the honest answer will be a list of dates. “Show me where this preference came from” will become a normal question, and it will sometimes be a humbling one.
The second showing becomes the product.
The attention business was built on the first impression. Delegates make first impressions cheap and filtered. The scarce thing becomes being shown again, so anyone who wants you to like them will compete for a place on your delegate’s repeat list, not its front page.
Acquired taste becomes a setting.
The old advice was that you have to try a thing several times before you know. Delegates will make that advice executable: a quiet schedule that brings back, small and unannounced, the things you turned down once. Wide taste stops being a personality trait and becomes an instruction.
Two delegates, two drifting people.
Couples and friends who rely on different delegates will find their tastes in food, music and places drifting apart a little each year, for reasons neither can name. Shared exposure will become something people arrange on purpose: the same table, the same playlist, the same repetition.
Exposure becomes the thing people guard.
The privacy question has always been about what goes out. The new question is about what comes in, repeatedly. People will ask less “what do you know about me?” and more “what have you been showing me, and how often?”
Where I stand
I already live inside this. I told my delegates to stop asking me and decide, and I meant it; I wrote about why in You Do It. You Decide. Deciding includes deciding what I see first and what I see again. I do not want that back. Handing it back would only return the choice to chance, and chance was never neutral either; it was just unaccountable.
What I want is the second showing made visible. When a delegate shows me something again, I would like to know it is the second time, and why. That is a small instruction. It is also the difference between a taste that was made for me and a taste I watched being made, and kept.
The claim ledger
- Already true
- Mere exposure is one of the most repeated findings in social psychology, across words, shapes, faces, sounds and food. Recommendation systems already decide much of what people hear and watch more than once.
- What has to happen
- Delegates have to move from recommending to acting, so the repetition happens without you choosing it each time: ordering, queuing, scheduling, drafting. And they have to remember across months, so the same choices recur.
- Where I am probably wrong
- The effect is strongest for neutral things and fades after a dozen or two showings. If delegates mostly repeat what you already picked, rather than introducing things and bringing them back, they will freeze taste instead of shaping it, and predictions three and four fail. People may also notice and resist. If, after a few years of living with delegates, people’s tastes track what their delegates showed them no more closely than taste tracked the radio, I was wrong.
The second time
Next summer a song will come on that you do not like. You will hear it again, and again, and by August you will be humming it in the car. That part has not changed since 1968.
What has changed is that this time, something chose the second time.
Ask it why.
Background
- Robert B. Zajonc, “Attitudinal Effects of Mere Exposure,” Journal of Personality and Social Psychology 9, no. 2, pt. 2 (1968).
- William R. Kunst-Wilson and Robert B. Zajonc, “Affective Discrimination of Stimuli That Cannot Be Recognized,” Science 207 (1980).
- Theodore H. Mita, Marshall Dermer and Jeffrey Knight, “Reversed Facial Images and the Mere-Exposure Hypothesis,” Journal of Personality and Social Psychology 35 (1977).
- Robert F. Bornstein, “Exposure and Affect: Overview and Meta-Analysis of Research, 1968–1987,” Psychological Bulletin 106 (1989).
- John Rector, The Reality Equation
- John Rector, Belief: Make Better History
- John Rector, The Room That Learns Your Name
- John Rector, Astonished
- John Rector, You Do It. You Decide.
- John Rector, The Last Line Is You