Disinterest Screen for Taste Panels

Concept

Kant's central test for a valid judgment of taste is disinterestedness: we must like a thing for its own qualities, not for the advantages of owning, using, or profiting from it [0]. Real evaluation panels violate this constantly — consumer tests include product owners and brand loyalists, design-award juries include practitioners with stakes in rival work, and concept tests are skewed by purchase desire. The Disinterest Screen is a short pre-panel instrument delivered as a module for panel-management and jury software. Before rating begins, each evaluator completes a brief two-axis interest profile:

• Stake axis (contamination): ownership, concrete purchase plans, professional or financial stake, brand affiliation, rivalry with entrants, and recent exposure context.
• Engagement axis (signal): affective response to the work itself — being moved, absorbed, or delighted — captured separately from any desire to acquire or profit.

The tool outputs a per-evaluator stake score, an engagement score, and embedded validity flags; it excludes or down-weights high-stake raters and appends a stake-adjusted verdict next to the raw vote. High-engagement raters are retained, and their engagement is reported as signal rather than filtered out as noise.

Why the second axis

New philosophical work directly reshapes the construct. On the Interest in Beauty and Disinterest [1] clarifies what disinterest means, shows through literary depictions that encounters with beauty can be personally transformative, and argues that beauty fuels a special kind of interest — love — going so far as to recommend abandoning the disinterest frame. This cuts both ways for the screen. It challenges the original single-axis design: a screen that treated all strong feeling as contamination would filter out exactly the evaluators most responsive to the work, whose verdicts may be the most informative. But it also sharpens what contamination can usefully mean in a panel context: material and prudential stakes — ownership, acquisition intent, professional rivalry, financial interest — which can be measured with behaviorally anchored items, unlike philosophical disinterest itself. The screen therefore no longer claims to measure Kantian disinterestedness; it measures the absence of biasing stake and reports love of the work on a separate axis. Boundary items are split explicitly: a plan to buy within a defined window counts as stake; being moved without acquisition intent counts as engagement. Residual misclassification is surfaced in the output rather than assumed away.

Guarding against gaming

An open question was whether raters could win extra weight by claiming disinterest. Embedded performance validity indicators from neuropsychological assessment offer a tested pattern: brief embedded measures with calibrated cutoffs detect noncredible responding at high specificity (roughly .87–1.00 for individual indicators) but modest, variable sensitivity (.04–.64), and requiring two or more failed indicators balances the tradeoff (sensitivity .31–.59, specificity .81–.93) [2]. The screen adopts this pattern: embedded consistency and plausibility checks — response timing, internal consistency of the interest profile, implausible neutrality given reported exposure — flag suspected gaming, and flags trigger down-weighting and human review rather than automatic exclusion. Three caveats travel with this evidence: the WAIS-IV indicators were validated in a mixed clinical sample (N = 205) and their cutoffs will need recalibration for consumer and jury populations; the indicators are performance-based, so transferring the pattern to self-report screening is an adaptation that needs its own validation; and low sensitivity means many gamers will pass, so the adjusted verdict is reported with confidence bounds rather than as purified truth.

Why now

The Critique of Judgement [0] supplies the original definition of contamination; the new philosophical analysis [1] contests that definition and, in doing so, supplies the refinement that keeps the screen workable — separating biasing stake from the legitimate interest beauty fuels. Corpus evidence shows the distortion is not hypothetical: exposure to attractive stimuli shifted both judgments of target stimuli and raters' own self-esteem and body satisfaction [3], so who the rater is and what they have just seen systematically distort verdicts. Validity-indicator research [2] shows that cheap embedded checks can catch noncredible responding at scale. Together these justify a standardized two-axis screen instead of trusting raw panel averages.

Remaining open questions

• Can raters reliably separate stake from engagement in self-report, or do they conflate love with acquisition desire? The boundary-item split needs validation against behavioral outcomes.
• Does removing high-stake raters increase predictive validity for general market success? The tool reports raw and adjusted verdicts side by side so clients can see the delta rather than trust either number alone.
• Will professional jurors read an engagement score as flattery and a stake score as an attack on their expertise? Framing of the two axes matters for adoption.
• Where does interest stop being bias and become a necessary component of expert judgment? The engagement axis is a first operational cut, not a settled answer.