Chronophilia
An age preference is not the same as an attraction to ageing.
Consider two men, both currently forty-five, both of whom would tell you they are attracted to women in their thirties.
The first has been attracted to women in their thirties since he was twenty-two. The target has not moved. When he was young the women he wanted were older than him; now they are younger; in fifteen years they will be substantially younger; and none of this represents any change in him whatsoever. His preference is a fixed point in the world, and his relationship to it changes only because he is drifting past it.
The second was attracted to women in their twenties when he was in his twenties and women in their thirties now that he is in his forties, and if you check on him at sixty he will be attracted to women in their fifties. His preference is not a point at all. It is a relation — roughly ‘people around my age’, or ‘people at my life stage’ — and it will travel with him for as long as he lives.
These two men will give an identical answer to any survey conducted this year. They are not remotely the same. One of them has an age preference in the way that a person has a preference for a particular latitude; the other has one in the way that a person has a preference for the present.
And there is a third case, rarer and much more interesting, which is the person attracted specifically to the difference. Not to thirty-year-olds, but to people fifteen years younger than themselves, whatever they themselves happen to be. Or fifteen years older. For this person the erotic object is the gap itself — the asymmetry, the structure of the thing — and the ages are only whatever the gap is currently made of.
Fixed target, sliding target, fixed gap. Three completely different erotic temporalities, indistinguishable in cross-section, distinguishable only if you follow someone for thirty years, which almost no study does.
There is a fourth arrangement which is the most common of all and gets no attention because it is not a preference at all. Two people meet, at whatever ages they happen to be, and stay together, and the ages go on changing, and neither of them ever exercises a preference again. Most human erotic life, statistically, is this. The whole apparatus of age preference is a description of the marketplace, and most people are in the marketplace for a small fraction of their lives.
from Chronophilia (2026)
GPT-5.6 Sol:
The passage makes a genuinely important distinction, but then pushes it a little too hard. Its strongest idea is methodological: a stated preferred age is an observation, not a model of the preference generating that observation. ‘I am attracted to women in their thirties’ tells you where someone’s attraction happens to land now. It does not tell you what quantity is being held constant.
The first three cases can be written quite cleanly. Suppose s is the subject’s age and t the preferred partner age. A fixed target has something like t = c: the preferred age stays near thirty-five while the person ages. A sliding target has t ≈ s + k, usually with k near zero: preferred partner age moves with one’s own age. A fixed gap is also t = s + k, but with a deliberately nonzero k, say −15. That immediately reveals one weakness in the prose: cases two and three are not mathematically different kinds of relation. They are different values, meanings, or perhaps widths of the same relative-age relation. ‘Someone around my age’ might be t = s ± 5; ‘someone fifteen years younger’ might be t = s − 15. The supposedly categorical distinction is partly rhetorical.
Still, psychologically the distinction may be real. ‘People at my life stage’ and ‘people substantially younger than me’ could arise from quite different attractions even though both move with s. One could be driven by peer similarity, shared cultural references, comparable health and social position; the other by asymmetry itself—status, generational contrast, youthfulness, seniority, dependence, mentorship, or simply visual age markers. So the author is right that the same age trajectory need not imply the same erotic object.
The sharper insight is the contrast between age and ageing. Someone who consistently prefers thirty-five-year-olds is not, merely by virtue of growing older themselves, becoming increasingly attracted to youth. Their preferred object has remained constant; ‘youth’ is a relational description imposed by the observer. Conversely, someone whose preferred partner ages alongside them might not be attracted to ageing as a process either. They may simply have an attraction function indexed to self-age, cohort, or life stage. ‘Attracted to ageing’ would be a stronger claim still: it suggests that the transformation associated with increasing age is itself positively valued. None of the first three patterns establishes that.
This matters because ordinary language collapses several variables. We say someone ‘likes younger women’, but that can mean at least: likes women who are young in absolute terms; likes women younger than himself; likes a particular age range which has become younger relative to him; or happens currently to have a younger partner. Those are logically distinct propositions. Cross-sectional surveys are especially bad at separating them because they observe t and s at one time and try to infer the function relating them. As the passage says, two people can produce the same point while belonging to different trajectories.
There is another complication the essay does not mention: cohort effects. Imagine a man who at twenty-five liked twenty-five-year-old women and at forty-five likes forty-five-year-old women. That looks like age-homophily. But perhaps what he actually likes is women born around 1981: his own cultural cohort. Age and cohort advance together, so longitudinal observation alone cannot fully distinguish ‘same age’ from ‘same generation’. Likewise, ‘life stage’ is not identical to chronological age. A forty-five-year-old divorced parent, a forty-five-year-old first-time parent, and a forty-five-year-old never-married academic may occupy rather different social worlds.
The fourth case is where the argument becomes less convincing. The claim that long-term couples ‘never exercise a preference again’ conflates mate choice with attraction. A monogamously partnered person may never again choose a partner while continuing to have age-structured attractions. Remaining with somebody as both partners age is evidence that attachment can survive ageing; it does not demonstrate the absence of an age preference. A fifty-year marriage is therefore not outside the theory of preference. It simply gives us very little revealed-preference data because there are no repeated partner selections.
And ‘most human erotic life, statistically, is this’ is much too sweeping unless backed by unusually good evidence. Even granting that long partnerships occupy many person-years, ‘erotic life’ includes fantasy, noticing others, pornography, crushes, affairs, widowhood, divorce, serial monogamy and remarriage. Marketplace behaviour is certainly an incomplete proxy for erotic preference, but the author overcorrects by treating partnered life as if preference became dormant or meaningless.
There is also a selection problem hiding beneath the entire discussion. Observed partners are not preferred partners. Who someone dates is constrained by reciprocal attraction, opportunity, social networks, norms, fertility intentions, income, geography and competition. Someone might prefer thirty-year-olds throughout life yet increasingly partner with people closer to their own age because thirty-year-olds cease to reciprocate. Conversely, a rich or famous older person might acquire much greater ability to realise a fixed-age preference. Partner-age data therefore identify equilibrium matches, not uncomplicated desire.
A more complete framework would distinguish at least four things: absolute partner age; age difference relative to oneself; cohort/life-stage similarity; and persistence of attraction to a particular individual as that individual ages. You could then add a fifth variable for attraction to age-associated change itself. The essay’s ‘fixed target / sliding target / fixed gap / settled pair’ typology gestures toward this multidimensional structure but treats the dimensions as mutually exclusive categories when they probably are not.
That last point is important. A person could simultaneously have a broad sliding preference and a fixed-gap component—for example, generally preferring contemporaries but with a mild tendency toward partners five years older. Someone could have a fixed absolute-age peak whose strength weakens as the social distance from that age group increases. Partner-specific attachment can overwhelm an underlying population-level age preference. Human preference functions are almost certainly distributions with several terms, not four boxes.
So I would preserve the central thesis but reformulate it:
An age preference should be understood as a rule relating attraction to age, not merely as the age at which attraction currently peaks.
Once stated that way, the problem becomes much richer. What looks cross-sectionally like ‘preference for thirty-five-year-olds’ may be a constant absolute-age preference, a self-relative preference passing through thirty-five, a cohort preference, a life-stage preference, a preferred age gap, or the residue of attachment to one particular thirty-five-year-old who subsequently becomes forty-five. Those hypotheses make different predictions over time.
And that is probably the strongest part of the passage: it turns what looks like a question about which ages people like into a question about what remains invariant as people move through time. That is the more interesting object of study.
