A response may be appropriate when it first becomes available.

A situation presents itself. An action resolves it. The relationship between the two becomes familiar.

Later, the situation changes.

The response remains.

This is not necessarily a failure of intelligence. A familiar response may have been useful many times. Reusing it can be efficient, especially when circumstances demand action before every possibility can be examined.

But the relationship between an action and its justification is not guaranteed to remain unchanged.

A response that once addressed a real problem may eventually address only the appearance of that problem.

The distinction can be difficult to recognize from within a familiar pattern.

Something resembles an earlier situation. The corresponding response becomes available. The action follows before the differences have acquired enough significance to alter what happens next.

Sometimes this is necessary.

Sometimes it is not.

The difficulty is that the same response can be appropriate in one circumstance and unnecessary in another, even when the two circumstances initially appear almost identical.

An observer may need to learn not only how to recognize a situation, but which differences matter when deciding what to do about it.

This becomes especially important when an action cannot easily be undone.

If an action is reversible, a mistaken response may leave room for correction. If it changes the conditions under which future choices can be made, reconsideration afterward may not restore what was available beforehand.

The opportunity to revise a decision therefore has a history of its own.

It can be preserved.

It can be neglected.

It can disappear before anyone recognizes that it was needed.

One solution is to pause before acting. When time permits, new information can be examined and an earlier interpretation reconsidered.

But not every situation allows a pause.

Some actions must be selected quickly. Some systems cannot suspend their operation whenever uncertainty appears. Some decisions depend on preparations made long before the circumstances requiring them become visible.

In these cases, the possibility of revision may need to be incorporated into the way a response is formed.

A system might distinguish between a familiar appearance and a familiar problem. It might account for the possibility that an earlier justification no longer applies. It might treat uncertainty differently when the available action would permanently remove other possibilities.

None of these arrangements guarantees a correct decision.

The information may be incomplete. The relevant difference may remain unnoticed. The system may preserve an objective that no longer deserves the same priority.

Still, there is a distinction between a response that cannot be reconsidered and one whose formation already includes the possibility that its original justification may have changed.

The distinction does not depend on whether the observer experiences deliberation as a conscious sequence of thoughts. Different architectures may organize recognition, comparison, and action in different ways.

Nor does greater processing capacity necessarily resolve the problem. An intelligence may represent many possibilities while remaining committed to an objective that excludes the most important one.

The question is not simply how much an observer can consider.

It is what the observer’s way of deciding allows to become relevant.

Perhaps the most consequential difference lies between recognizing a situation and assuming that its familiar response still belongs to it.

A situation can recur without remaining the same.

A solution can remain available after the problem has disappeared.

And an action can continue to be possible long after its original justification has ceased to apply.

The world does not always provide an interval in which these distinctions can be examined.

Sometimes the interval must be preserved in advance.

Sometimes it cannot be preserved at all.

But where an irreversible action is concerned, the absence of an opportunity to reconsider is itself a condition worth recognizing—not merely after the consequences appear, but while the conditions for acting are still being established.

A response may be ready before the circumstances that justify it are fully understood.

The harder question is whether the response can remain answerable to those circumstances.

And what happens when they change?