Philosophy relies heavily on imagined cases that could not occur. They do genuine argumentative work, and the way they do it is worth examining directly.
They isolate one variable
Real situations vary in many respects at once, so a judgment about a real case cannot show which feature produced it. The relevant factor is buried among a dozen others.
An invented case can hold everything constant and alter one element, which is why the scenarios are typically strange. Strangeness is what the isolation costs.
In this respect the method resembles a controlled experiment, with the important difference that the result is an intuition rather than an observation.
The response is the evidence
What a thought experiment produces is a reaction: this seems wrong, that seems to be still the same person, this description does not quite fit the case as presented.
The argument then treats the reaction as data about the concept in question, on the grounds that competent users of a concept can recognise its application.
This is why disagreement about a case is significant. It suggests the concept is contested or that respondents are attending to different features.
Details do more than they appear to
Small changes in how a case is described can reverse the response, including the order of presentation, the wording of options and who is named as acting.
Careful writers therefore present variants, altering one detail at a time to identify what the reaction actually tracks.
Where a response flips on a detail that seems morally irrelevant, that instability is itself a finding worth explaining.
The method has real limits
Intuitions about impossible cases may simply be unreliable, since the concepts involved were formed for ordinary situations and may have no determinate application outside them.
Responses also vary with background and training, which complicates the assumption that they report something shared.
Critics argue that a great deal of philosophical weight rests on a narrow and unrepresentative set of reactions, and the objection is taken seriously.
Why the practice continues
Imagined cases remain the clearest way to test whether a proposed definition captures what it claims, since a single well-built counterexample can defeat one.
They also make abstract disputes tractable, giving the parties a shared object to disagree about rather than competing formulations that may not even be addressing the same question.
Used with awareness of their limits, they function less as proofs than as instruments for finding where a concept bends.