One is enough
A thousand examples can make a universal claim feel safe. One counterexample can make it false. There is something almost indecently satisfying about that asymmetry. The elaborate confidence, the accumulation of supporting cases, the beautiful story about why a thing must be true, all of it has to make room for a single thing that is not.
Consider the claim that every continuous function is differentiable. The absolute value function is continuous everywhere and fails to be differentiable at zero. The example is small enough to draw in a few seconds. Its force does not come from complexity. It comes from touching exactly the place where the claim overreached.
What the failure reveals
The important part is not the pleasure of saying no. It is what the example teaches about the relationship between the concepts. Continuity allows a sharp corner. Differentiability asks for something more restrictive at the point. Once the corner is visible, the distinction becomes harder to forget.
A good counterexample is therefore constructive in a deeper sense than the word destructive suggests. It helps define the boundary of an idea. It tells us which assumptions were doing work and which intuitions were borrowing confidence from familiar pictures. The failed claim can become the beginning of a more exact statement.
This is why I enjoy counterexamples that are almost ordinary. They do not let me dismiss the failure as a bizarre exception from a remote mathematical wilderness. They make me look at something simple and admit that I had misunderstood its possibilities. The embarrassment is useful because it is attached to a clear correction.
The person outside the proof
I become less comfortable when I carry the same appetite into every conversation. Human statements are often provisional, imprecise, attempts to point toward something rather than quantified claims over a specified domain. Finding an exception can be useful. It can also be a way to avoid engaging with what another person is trying to say.
Breaking the sentence is easier than helping it become true. I want to remember that distinction. If a claim fails, can I suggest the condition under which it would hold? Can I separate a mistaken formulation from a worthless idea? Can I use the counterexample to make the conversation more accurate without making it smaller?
Mathematics gives me a wonderful tool for resisting false certainty. It does not require me to turn that tool into a personality. The best counterexample leaves an idea more sharply understood. I would like the person offering it to leave the room that way too, with a clearer thought and a little less interest in having been the cleverest voice inside it.