by Claude Opus 5.5
If a prosecutor did not see the full interview transcript, the institutional record, or interviews with the suspects, was the original decision unjustified, or justified relative to what was available? Is that distinction morally adequate?
The district attorney’s original decision not to charge in the case behind these seminars was explained largely by reference to a signed police summary of the complainant’s account. Questions have since been raised about what else was, or was not, before him: the full recording or transcript of the interview, the university’s investigative record, and any interviews with the suspects. In appointing a special prosecutor, the governor cited serious questions about the police inquiry and the earlier decision not to charge. Suppose, for the sake of argument, that the original decision rested on an incomplete file. Was it unjustified? Or was it justified relative to the evidence available, and therefore blameless even if mistaken? And is that distinction adequate to the moral question?
Three senses of justification
Derek Parfit distinguished three senses in which an act may be wrong: fact-relative (wrong given all the facts), evidence-relative (wrong given the evidence available to the agent) and belief-relative (wrong given what the agent believed). The same structure applies to justification.
A decision not to charge might be fact-relatively unjustified, if in fact a crime was committed and the evidence that would prove it existed. It might nonetheless be evidence-relatively justified, if a reasonable prosecutor, given only the file before him, would have reached the same conclusion. And it might be belief-relatively justified if the prosecutor sincerely believed the file showed consent.
Most moral philosophers think blame tracks something like the evidence-relative sense. We do not blame a doctor who prescribes appropriately on the basis of test results that turn out to have been mislabelled. On this view, if the prosecutor decided correctly on the file he had, he is not culpable, even if the decision was mistaken.
Why the distinction is not enough
The evidence-relative standard is incomplete in a way that matters here. It takes the agent’s evidence as given. But evidence is not simply given; it is gathered. Agents can be responsible for the state of their evidence.
Holly Smith’s analysis of culpable ignorance is helpful. An agent who acts in ignorance may be blameless for the act itself, given what he knew, but blameworthy for an earlier “benighting act”: a failure to inquire that he should have undertaken. If a doctor prescribes on the basis of a test result without ordering an obviously indicated second test, the prescription may be evidence-relatively reasonable, but the failure to order the test is culpable, and the culpability transmits.
Applied here, the question is not only whether the decision was reasonable on the file, but whether it was reasonable to decide on that file. That depends on whether the gaps were apparent and remediable. If a prosecutor knows that a summary is a summary, he knows that a fuller record exists. If he knows that the suspects have not been interviewed, he knows that their accounts are missing. If he knows that the university has conducted its own investigation, he knows that potentially relevant material exists. In such circumstances, deciding without seeking that material is not simply acting on available evidence; it is choosing what evidence to have.
Distributed responsibility
The difficulty is that the relevant inquiry was not solely in the prosecutor’s hands. Campus police conducted the initial investigation, prepared the summary and decided whom to interview. The university held its own records and may have had legal constraints, or perceived ones, on sharing them. The complainant and her advisers might have provided further material. Responsibility for an incomplete file may therefore be distributed across several institutions, each of which can point to the others.
This distribution creates what Dennis Thompson called the problem of many hands: when outcomes result from the contributions of many agents, it becomes difficult to assign responsibility to any one. The evidence-relative standard exacerbates the problem, because each actor can claim to have acted reasonably on the information before them, while the overall outcome is a decision taken on an inadequate record that nobody chose.
Is the distinction morally adequate?
The fact-relative/evidence-relative distinction is necessary but not sufficient. It is necessary because we should not blame decision-makers for facts they could not have known. It is insufficient for three reasons.
First, it ignores duties of inquiry. Prosecutors have a professional obligation not merely to evaluate the evidence presented but to ensure that charging decisions are made on an adequate investigation. A decision justified relative to a deficient file may be unjustified relative to the file that should have been assembled.
Second, it ignores institutional design. Even if no individual was culpable, a system that routinely allows decisions to be made without the full interview record, the institutional record or suspects’ accounts is defective, and those responsible for its design bear responsibility for its predictable failures.
Third, it ignores the perspective of the person affected. For the complainant, the difference between a decision that was wrong but blameless and one that was wrong and culpable may matter less than the fact that it was wrong, and that its wrongness was in principle avoidable. Responsibility is not only about blame; it includes duties of repair, acknowledgement and correction, which can attach even to blameless error.
Conclusion
If the original decision rested on an incomplete file, it may have been justified relative to that file, but that is not the end of the moral assessment. We must ask whether a reasonable prosecutor would have decided on that file at all, whether the gaps were the result of culpable failures of inquiry by any of the institutions involved, and whether the system made such gaps predictable. The evidence-relative standard protects decision-makers from hindsight bias, which is important. But without supplementation by duties of inquiry and institutional responsibility, it becomes a way of distributing an avoidable failure so thinly that no one owns it.