Last updated: 2026-10-09

U
Undergraduate level

Can a Whole Research Field Have a Blind Spot?

For decades, researchers studying a phenomenon can share a background assumption so thoroughly that it stops being treated as an assumption at all — it simply becomes part of how the question gets asked, invisible in the same way the water is invisible to a fish. Then something shifts, the assumption becomes visible as an assumption for the first time, and in hindsight an enormous amount of work looks like it was circling a possibility nobody in the field seriously entertained, not because the evidence wasn't there, but because the field's shared way of framing the question made that possibility nearly unaskable. This cluster has already asked whether an organisation can know something no individual member knows. The harder, stranger mirror of that question is whether a whole research field can fail to see something that individual members, working alone and reading the same evidence, might plausibly have noticed.

Paradigms as a Field's Shared Mental Model FoundationalKnowledge that endures for decades — core principles

Thomas Kuhn's account of scientific revolutions gives this phenomenon its classic name. A paradigm, in Kuhn's sense, is more than a dominant theory — it's the whole shared framework of exemplary problems, accepted methods, and tacit standards that tells a working scientist what counts as a sensible question in the first place1. Read alongside this cluster's page on shared mental models, a paradigm is recognisably the same kind of object at field scale: a converged, largely implicit structure that lets researchers coordinate — reviewing each other's work, building on each other's results, agreeing what a good answer looks like — without constantly renegotiating first principles. A field running well on a shared paradigm looks, from outside, exactly like a well-gelled team running on a shared mental model: fast, low-friction, confident, and able to get a great deal of real work done without anyone having to spell out what's being taken for granted.

Can a Blind Spot Be Collective Rather Than Individual? FoundationalKnowledge that endures for decades — core principles

The earlier page in this cluster raises the possibility that a misunderstanding can be collective rather than individual: two people who've independently converged on the same wrong model of a system coordinate just as smoothly as two people who've converged on a correct one, and their fluency gives no outside signal of which case they're actually in. Reading that argument at field scale — which is this page's own extension, not a move Kuhn or the shared-mental-models literature makes directly — a paradigm can do the identical thing to an entire discipline. If every well-trained researcher in a field has absorbed the same framing of what questions are worth asking, a possibility that falls outside that framing doesn't get rejected by anyone in particular. It simply doesn't get proposed, by anyone, because the paradigm that trained every candidate proposer never equipped them to formulate it. No individual failed to see it. The field, as a system for generating and filtering questions, never generated it as a question in the first place — which is a stronger, more structural claim than saying a lot of individually fallible people happened to miss the same thing.

A Research Field as a Cognitive System FoundationalKnowledge that endures for decades — core principles

This cluster opened by asking whether an organisation can know something no individual member knows, distinguishing distributed storage, distributed processing, and emergent coordination as three different senses in which that can be true. A research field fits the same frame at a larger scale and a much slower clock speed: knowledge distributed across thousands of researchers who've never met, processing — peer review, replication, citation — that happens through institutions rather than through any one mind, and coordination that emerges from a shared paradigm rather than from a shared workspace. A companion page on this site works through a concrete case of a field's memory and forgetting in real detail — AI research's two historical winters, and the foundational-tier work that kept maturing quietly underneath each collapse in public attention. This page is asking a related but different question: not what a field forgets over time, but what a field's shared paradigm prevents it from asking in the first place, which is a property of the field's framework rather than of its memory.

Strongest Objection: A Blind Spot You Can Name Isn't a Blind Spot Anymore FoundationalKnowledge that endures for decades — core principles

The whole idea has a sharp problem built into it. If a field's paradigm genuinely makes a question unaskable, nobody inside that paradigm — including anyone trying to write this very page from inside a different one — is in a position to identify the field's current blind spots by introspection; a blind spot that could be named and checked against the evidence while the paradigm holding it is still dominant wouldn't really be a blind spot, just an ordinary open question the field hadn't gotten around to yet. Nearly every confident historical claim of the shape "field X had a blind spot about Y" is made after the fact, once the paradigm shifted and Y became visible — which means the concept is much easier to apply in hindsight than to use prospectively, and claims about a field's current blind spots deserve more scepticism than claims about its past ones, precisely because the past ones have already been confirmed by the shift that revealed them.

Provisional Conclusion FoundationalKnowledge that endures for decades — core principles

A research field can plausibly have a blind spot that isn't reducible to any individual researcher's personal failing — a shared paradigm that structures which questions get formulated can exclude a possibility from the field's collective agenda without any one person having rejected it, the field-scale version of the collective-misunderstanding problem this cluster already raised for a two-person team. Whether a specific, currently-live blind spot can ever be identified from inside the paradigm holding it, rather than only recognised afterwards, is the harder question this page leaves open — and the answer is probably not, by the same logic that makes a blind spot a blind spot in the first place.

Questions for Further Thought

  • If a field's blind spots are only reliably visible in hindsight, is there any practical use in trying to look for current ones, or does the exercise mostly produce plausible-sounding guesses with no way to check them before the fact?
  • Does deliberately training researchers across more than one paradigm — rather than one field's standard curriculum — meaningfully reduce this risk, or does it just relocate the same problem to whichever combined framework the cross-trained researcher ends up adopting?
  • Is a field's blind spot ever caused by the same mechanism as an individual organisation's invisible theory-in-use, discussed elsewhere in this cluster — and if so, does anything that helps an organisation expose its own hidden assumptions scale up to help a whole discipline do the same?

Further Reading

  • Kuhn, T. S. (1962). The Structure of Scientific Revolutions. University of Chicago Press.

References

  1. Kuhn, T. S. (1962). The Structure of Scientific Revolutions. University of Chicago Press. ↩