What Is a Value Judgement? The Line Between Describing and Endorsing

David Hume observed in 1739 that writers move from statements about what is to statements about what ought to be without explaining the transition, and argued that the step requires justification that is never supplied. That observation, generally called the is-ought problem, does not say that moral conclusions are unjustifiable, and it says that they do not follow from factual premises alone and that some evaluative premise must be present. The related naturalistic fallacy, named by G. E. Moore in 1903, targets attempts to define goodness in terms of a natural property such as pleasure or survival, on the grounds that for any such definition it remains a sensible open question whether the property really is good, which suggests the definition has not captured the concept. Both arguments have been disputed at length, and the practical residue that nearly everyone accepts is that an argument reaching an evaluative conclusion must contain an evaluative premise somewhere, and that locating it is a useful move in any disagreement.

Evaluative content enters apparently factual language constantly, and the routes are identifiable:

• •Thick concepts, meaning words that describe and evaluate simultaneously, so calling an act cruel, generous, brave or reckless both states what happened and assesses it
• •Selection of what to report, since deciding which facts are relevant embeds a judgement about what matters
• •Framing and comparison classes, since the same figure presented as a rise or a fall depends on the baseline chosen
• •Categorisation, since deciding what counts as unemployment, poverty, violence or a species determines what the measurement finds
• •Choice of measure, since an outcome tracked by one metric and not another reflects a decision about what is worth tracking
• •Presupposition in questions, which builds a contested claim into the framing so any answer concedes it

The claim that science is value-free is generally regarded by philosophers of science as untenable in its strong form and defensible in a weaker one. The weak version holds that the evidence for a hypothesis should not depend on whether the conclusion is welcome, which is what objectivity actually requires and is achievable. The strong version, that scientific practice contains no evaluative commitments, fails because choices about what to study, what counts as a sufficient standard of evidence, how to balance false positives against false negatives, which populations to sample and how to categorise are all made and all involve judgements about consequences. The argument from inductive risk makes this precise: since accepting a hypothesis on limited evidence risks error, and the costs of the two kinds of error differ, deciding how much evidence is enough requires weighing those costs, which is an evaluative decision inside the scientific process rather than outside it. The response generally recommended is transparency about the judgements made rather than a pretence that none were.

The practical value lies in diagnosis rather than in dismissal. Identifying the evaluative premise in an argument frequently reveals that a disagreement thought to be about facts is about values, or the reverse, and each requires a different kind of discussion, since facts can be checked and values must be argued. It also guards against two symmetrical errors. The first is treating a contested evaluation as though it were an established fact, which forecloses argument by definition. The second is treating a factual dispute as merely a difference of values, which is a common way of avoiding evidence that contradicts a position. The distinction is also not always sharp, since thick concepts genuinely combine both and the attempt to separate them can distort what a claim means. The workable position is that the distinction is a tool for clarifying arguments rather than a boundary running through the world, and that its usefulness does not require it to be exact.

Statements about what is and about what ought to be are different, and an argument reaching an evaluative conclusion must contain an evaluative premise somewhere, which is the durable residue of Hume's observation. Words like cruel and generous describe and assess at once. Science cannot be value-free in the strong sense, since deciding how much evidence is enough means weighing the costs of being wrong.