The wish behind the phrase is to catch it early: to read, from a few small cues, that someone is quietly pulling away, so you can brace for it or head it off. It is an understandable instinct, and a mostly futile one. Reading whether another person wants distance is among the hardest things we try to do, and the research suggests that the more anxiously you scan for the signs, the worse your reading gets, and the more likely you are to bring about the very thing you dread. Two findings explain why, and point at what is actually worth paying attention to.
We are poor mind-readers
Start with a humbling body of work on what psychologists call empathic accuracy, the study of how well people infer each other’s thoughts and feelings. The findings, gathered over decades by William Ickes and others, are sobering. Everyday mind-reading is only modestly accurate, around 20 per cent correct with strangers, and although it improves with people we know well, it remains far from reliable, with even close partners frequently misjudging what the other is thinking.
Apply that to signs of distancing and the problem is obvious. The short reply, the missed message, the slightly flat tone are exactly the kind of ambiguous signals we read least accurately, and the trouble is that a bad day, tiredness, preoccupation, or plain distraction produces the same surface as genuine withdrawal. The behaviours are real enough; it is their meaning we are guessing at, and the evidence says our guesses are not good. A cue that feels like proof is usually just noise wearing the costume of a clue.
Anxiety makes it worse, and self-fulfilling
The reading gets less reliable still when insecurity is driving it. In an influential 1996 paper, Geraldine Downey and Scott Feldman described rejection sensitivity, the tendency of some people to anxiously expect rejection and, as a result, readily perceive it in the ambiguous or merely insensitive behaviour of others. Their studies found something more troubling than mere misperception. The expectation was self-fulfilling: primed to see rejection, people reacted in guarded, testing, or wounded ways that pushed the other person away for real, manufacturing the distance they feared.
This is the sharp edge of the whole exercise. Scanning hard for signs of withdrawal, especially when it comes from a place of anxiety, is not neutral observation. It tilts your interpretation toward the bleakest available reading, and then tilts your behaviour to match, so that you grow cooler, or clingier, or start setting little tests, none of which invites anyone closer. The harder you search for the exit, the more likely you are to open the door yourself.
What actually counts as a sign
If single cues are unreliable and scanning backfires, is there anything solid to go on? There is, but it is duller and slower than a tell. What carries real information is not a moment but a pattern: a consistent, unmistakable drop in investment sustained over time. Someone who used to initiate and now never does, whose effort and responsiveness have fallen away steadily across weeks rather than in a single quiet evening, is telling you something. One ambiguous cue is noise. A durable change in the baseline is signal.
And even then, the only reliable way to learn what the change means is the thing anxiety most wants to avoid: a low-key, direct, non-accusing question. Something as plain as “hey, are we okay? You’ve seemed a little distant and I wanted to check in.” That single move does what no amount of decoding can, because it swaps a guess for an actual answer. Just as importantly, asked warmly rather than as an accusation or a test, it is the opposite of the self-fulfilling spiral. It reaches toward the person instead of quietly bracing against them.
What the evidence does and does not settle
A few honest points belong here. Sometimes the distance is real, and a clear, sustained pullback deserves to be respected rather than chased. Not every withdrawal is a misreading to be corrected; people are allowed to want space, and if someone has plainly asked for it or stopped wanting contact, the kind and correct response is to let them have it rather than to intensify the search for reassurance. Reading cues is also genuinely harder with new acquaintances than with people you know, so hold early impressions especially loosely.
It is worth saying gently, too, that if you find yourself scanning constantly for signs of rejection in the people you care about, that pattern is itself worth attending to, and it often eases with support, because it tends to cost more than the rejections it is trying to pre-empt. Within those limits, the reassurance is real. You are a worse detector of other people’s pulling-away than it feels like you are, most of all when you are anxious. So trust sustained patterns over single cues, trust a kind and direct question over both, and remember that the surest way to find out whether someone wants distance is not to read them more closely, but to ask them more openly.
