When Does Missing Evidence Support Absence?

“Absence of evidence is not evidence of absence” is often used to prevent premature rejection. It has a point, but it is not true without qualification. The central question is not merely whether anything was found. It is how likely we were to find it if the thing really existed.

Suppose someone claims that an adult cat is inside a desk drawer. We open it, illuminate every corner and find nothing. A cat would almost certainly produce visible, audible or tangible signs, so the lack of signs counts against the claim. If the claim concerns a microscopic mite invisible to the naked eye, the same inspection has almost no power to exclude it. The observation is identical, but its evidential meaning changes with our ability and expectation to detect the object.

In epistemology and philosophy of science, evidence does not carry its weight independently of hypotheses. Its force comes from what competing hypotheses lead us to expect. Missing evidence becomes evidence of absence only when the contrary hypothesis should have produced detectable evidence. That is why the scope of a search, instrument sensitivity, sample size and alternative explanations matter.

My view is that we should neither treat silence as automatic disproof nor use “you have not proved it absent” to keep every claim alive indefinitely. The honest questions are: how did we look; how likely was that method to detect the thing if it existed; and what possibilities remain unchecked? Confidence should change with those answers, not jump between “there” and “not there” on the strength of a slogan.

https://plato.stanford.edu/entries/evidence/