Passage A: Literary-prize juries are often told to deliberate until they reach a single collective choice. The procedure sounds rigorous: discussion should correct private blind spots and produce a verdict more considered than any juror's first impression. Yet the demand for agreement can distort judgment. The distortion need not involve intimidation. A forceful chair may merely ask which title the group can defend together, shifting attention from the merits each juror perceives to the anticipated difficulty of composing a common statement. Even courteous discussion can thereby reward positions that are easy to merge. Because the pressure operates before any formal vote, a final ballot cannot reveal how much accommodation has already occurred. Jurors anticipating negotiation may abandon demanding but defensible candidates and rally around the book least objectionable to everyone. The winner then records the geometry of compromise, not a shared assessment of excellence. A better procedure would preserve disagreement rather than conceal it. Each juror should rank the finalists independently and publish a brief rationale before any meeting. Deliberation could still expose factual errors or neglected features, after which jurors could revise their rankings. But the prize would be determined by aggregating those final rankings, and the published rationales would disclose whether the result reflected broad admiration or merely divided preferences. Such a procedure would not make aesthetic judgment objective. It would make the institution's uncertainty legible and reduce the pressure to manufacture unanimity.

Passage B: Critics of jury deliberation treat conversation chiefly as a source of conformity, as though jurors entered the room with fully formed judgments that social pressure could only corrupt. But aesthetic evaluation is often inchoate until reasons are exchanged. A juror may admire a novel's fragmented structure without seeing that another juror experiences the same feature as evasive; explaining the disagreement forces both to specify the standards they are using. Moreover, an objection voiced in discussion is answerable in a way that a numerical ranking is not. A juror can narrow a claim, distinguish two cases, or concede that a favored feature has been overvalued. This process can create a reasoned consensus that is not reducible to compromise. Still, unanimity should not be the measure of successful deliberation. A panel may clarify the grounds of its disagreement and remain divided. Prize institutions should therefore publish a majority statement alongside signed dissents when consensus fails. This arrangement retains conversation's capacity to refine judgment while preventing an official verdict from masquerading as the only conclusion a reasonable reader could reach. Independent ballots alone cannot do this work: they record preferences, but without sustained exchange they may leave the principles behind those preferences untested. A ranking shows where a juror ended, not whether the route to that position can withstand a serious objection. The proper alternative to coerced unanimity is thus accountable deliberation, not the replacement of deliberation by arithmetic.