[ RESACH · FIELD NOTES ]
Five notes from “the papers you cite but never read” to “your next hypothesis is already in the gaps.” Research-on-research, sourced; the product stays out of the way until the end.
Your reference list is a set of promises: each entry says 'I read this, I understood it, and it supports the sentence in front of it.' Do the arithmetic on how many of those promises you could actually keep, and the field starts to look less like a body of knowledge and more like a very long game of telephone.
A citation is a pointer, and a pointer can point at anything — a paper that supports you, a paper that contradicts you, or a paper that merely gestured at a third paper which never actually made the claim at all. Follow enough chains and you find they bottom out in nothing. This is how a field ends up certain of things no experiment ever showed.
We reach for citation counts because they are the only number lying around — a proxy for quality that is really a proxy for attention. The moment a proxy becomes a target, it stops even doing that. Here is what the count measures, what it doesn't, and what an honest read of a paper's intrinsic quality would have to look at instead.
The best literature tools of the last few years gave us something genuinely new: a map of how papers relate. Connected Papers, Research Rabbit and their kin turn a seed paper into a constellation of neighbours. But a map of relatedness answers 'what else is near this?' — not 'is any of it true?' Those are different questions, and only one of them decides your next experiment.
Every argument in a field stops somewhere — a premise nobody tested, a boundary nobody checked, a chain that dead-ended in an opinion. Those stopping points are not embarrassments to route around. Written as falsifiable claims, they are your research programme. Here is the discipline that turns a gap into a testable hypothesis — and where you can watch it run on a real field.