Coherence and the burden of evidence
An internally consistent account can organise observations while leaving its central claim unconfirmed.
A theory can make a collection of observations intelligible and still lack adequate support. That possibility deserves attention whenever an explanation offers to resolve several difficult questions at once. The satisfaction of understanding can arrive before the evidence needed to justify it.
Coherence is a genuine intellectual achievement. An account that contradicts itself cannot be accepted intact. An account that survives scrutiny has cleared that obstacle. Its internal consistency, however, leaves open the question of whether it describes the world.
This essay concerns that inferential distinction. It proposes a way of examining broad explanatory frameworks, rather than reporting a new empirical finding.
The source of the satisfaction
An explanatory framework assigns relationships. What previously appeared disconnected becomes part of a structure: a consequence of a principle, a manifestation of a mechanism, or a special case of a general account. The structure reduces the number of things that have to be considered independently.
There is a reason to value that economy. An explanation should tell us more than a catalogue does. The difficulty arises when the same structure can be arranged around several incompatible premises. Each account may organise the observations successfully while disagreeing about their underlying cause.
At that point, coherence has helped identify candidates. Choosing among them requires examining where their commitments diverge.
Consider the minimal case of two accounts that both accommodate an observation. The fact that the observation is compatible with the first account cannot establish its superiority to the second. Something further must distinguish them: a prediction, an independently supported mechanism, a constraint on permissible assumptions, or a reason the rival cannot meet the same explanatory demand.
Accommodation has a cost
A framework can become more flexible whenever it meets a difficulty. It may introduce an unobserved process, qualify a general principle, or move the proposed mechanism outside the conditions under which it could be examined.
None of those moves is automatically illegitimate. An initial theory may need revision. The relevant question is whether the revision adds understanding that can be assessed, or merely allows the account to survive another possible result.
A proposed mechanism earns explanatory value when it imposes constraints. It should identify conditions under which an outcome is expected, conditions under which it should fail, or relationships that follow from its operation. A mechanism permitted to produce any result places little pressure on the account that invokes it.
The same issue applies to interpretation. A position that treats every observation as confirmation may be protecting itself from disagreement at the expense of informational value. If contrary outcomes make no difference to confidence, the role of evidence has become unclear.
The missing comparison
A useful evaluation asks what the observations would look like if the preferred explanation were mistaken. This question makes the alternatives explicit. It can reveal that a supposedly decisive observation would also be expected under a more limited account.
The comparison should be fair. An ordinary explanation cannot be accepted solely because it is ordinary. An unfamiliar explanation cannot acquire support solely because the ordinary account is incomplete. Both owe reasons appropriate to the claims they make.
There may be cases in which the available evidence cannot distinguish the candidates. The correct response is then to locate the uncertainty. Which observation is missing? Which measurement is too coarse? Which assumption prevents a comparison? A description of that limitation is more informative than a declaration that the subject is mysterious.
This does not require every philosophical position to become a laboratory hypothesis. Some arguments concern logical consistency, explanatory standards, or the meaning of a concept. Their merits can be assessed accordingly. Difficulty arises when the conclusion changes category without acknowledgement: conceptual possibility becomes a claim about an actual mechanism, and the standards appropriate to that mechanism are never applied.
Explanatory ambition
Broad accounts deserve proportionately broad scrutiny. A proposal linking consciousness, time, and identity must establish the connections it relies upon. A persuasive treatment of one subject cannot carry the others through association.
An account may be most useful before it is confirmed. It can expose an unnoticed assumption, formulate a sharper question, or suggest a comparison that previous approaches neglected. Those achievements should be stated on their own terms. Their value does not depend on presenting a provisional framework as a completed description of reality.
Coherence supplies an organised possibility. Evidence determines how much confidence that possibility has earned. Keeping both obligations visible allows a theory to be ambitious without asking the reader to mistake explanatory satisfaction for confirmation.