People occasionally describe someone as having strong opinions. The phrase is usually delivered as a neutral observation, though it often carries the faint scent of a warning: this person may be stubborn, loud or difficult. Those are failures of conduct. They have little to do with the strength of an opinion.

By strong, I mean a conclusion that can bear weight. Its owner can account for where it came from, has exposed it to resistance and knows what would change it. Strength is not certainty. It is confidence earned by the quality of the work underneath it.

I am not talking about tastes, direct observations, urgent warnings, questions or brainstorming. “I dislike coriander,” “I smell smoke” and “Could this fail under load?” do not pretend to be distilled judgments. I am talking about a conclusion presented for someone else’s serious consideration — something offered as a view of how the world is or what should be done.

On that meaning, only strong opinions are worth sharing.

An opinion is compressed work

An opinion worth having is a compression of work. Observations, facts, experience, competing explanations and trade-offs go in; a conclusion small enough to carry around comes out. The sentence one eventually says may take ten seconds. Producing a sentence entitled to those ten seconds may have taken years.

Time alone does not make a conclusion good. Ten years spent protecting an assumption from evidence produces a very old weak opinion. Five minutes may be enough when those minutes draw on deep experience. What matters is the work: testing the premises, looking for disconfirming evidence, meeting the strongest counterargument one can find and revising the conclusion when it fails.

Better work should produce better-calibrated confidence, not greater attachment. A conclusion that has survived serious attempts to break it deserves more confidence than a reflex. Neither deserves protection from a decisive fact. The value of the effort lies in producing a model that tracks reality, not an heirloom to be defended from it.

Attention is the listener’s price

Presenting a conclusion creates an asymmetric little transaction. I provide the conclusion; the listener pays first, with attention. They usually cannot inspect the work before deciding whether to listen. People rarely begin with an epistemic nutrition label: “I read one headline, misunderstood half of it and have been thinking about this for forty seconds.” They simply begin talking.

Suppose a headline says a council has banned cars. Someone announces that the council has lost its mind. The first question reveals that the restriction applies to one street on Sunday mornings, and the opinion disappears. Dropping it is correct. Presenting it as a considered conclusion was not. The speaker asked the listener to perform the elementary inspection they had skipped.

There is nothing wrong with saying, “This headline surprised me; what am I missing?” That is an invitation to investigate together. Nor is there anything wrong with thinking aloud when it is labelled as such. The problem is not uncertainty or incomplete thought. It is passing off an unexamined reaction as finished work and charging the listener the difference.

Resistance is part of the exchange

An opinion offered to another person becomes available for inspection. That is not an unfortunate side effect of sharing it. It is the exchange.

A well-formed opinion should have met the obvious objections before it arrives in conversation. Its owner should be able to expose the premises, identify what the conclusion depends on and distinguish a genuine contradiction from a difference in preference. As the quality of that prior work rises, decisive rebuttals should become rarer for the simple reason that the strongest ones have already been sought out.

Still, careful people miss simple things. A new fact or conclusive argument can collapse a view in seconds. Changing immediately may then demonstrate excellent judgment. The speed of the change proves nothing; the relationship between the change and the evidence does. Effort already spent is not a debt reality owes us, and defending a position merely to honour that effort is a sunk-cost error.

Expecting no pushback makes the transaction entirely one-sided. My thought, however little went into it, is entitled to attention and deference; yours, however carefully formed, is not permitted to return in the other direction. What I am requesting is not conversation but a temporary suspension of your judgment.

That is incompatible with the reason I asked to be heard. If my opinion deserves attention because it contains the product of my mental effort, then consistency requires me to treat disagreement as the possible product of yours. I need not accept it. I do need to let it arrive and assess it on its merits.

Conditionally held

Stubbornness is not strength. Volume is not strength. Certainty is not strength. These are often ways of concealing that an opinion has no visible support.

The useful form of strength is conditional. I believe this because of these facts and this reasoning. I have considered these objections. This is the evidence that would change my mind. Until that evidence appears, I will not pretend all positions are equally plausible merely to seem agreeable.

That produces opinions which are simultaneously firm and revisable. Firm, because they have survived contact with alternatives. Revisable, because their owner’s loyalty is to reality rather than to the social cost of being wrong. The familiar “strong opinions, weakly held” is reaching for this update rule, though weakly can sound like readiness to fold under friction. I prefer strong opinions, conditionally held: difficult to move without good reason and easy to move once good reason arrives.

So yes, I have strong opinions. At least, those are the only ones I intend to present as considered conclusions. Show enough of the structure that others can test it. Let them push. Change it when a load-bearing part fails. And when the thought is still young, say that plainly. A good question is more valuable than a counterfeit conclusion. The distinction respects your own judgment, the other person’s judgment and the finite attention between you.