The Problem-Solving Model of Progress

In the AI Story

Laudan's insistence that progress is measurable rather than merely felt was his most consequential methodological move. Previous frameworks had either posited an unattainable standard (correspondence to truth) or abandoned standards altogether (Kuhnian incommensurability). The problem-solving model offered something between: a standard that was epistemically modest but analytically powerful.

The model works by forcing disputants to be explicit about what they mean by progress. A triumphalist who claims AI represents progress must specify the problems it solves — the imagination-to-artifact ratio, the democratization of capability, the productivity multipliers of Trivandrum. An elegist who claims AI represents regression must specify the problems it fails to solve or the anomalies it generates — the depth atrophy, the productive addiction, the loss of tacit knowledge. The dispute becomes tractable because both sides must engage with the full ledger rather than the curated subset.

The model also permits evaluation across unequal domains. A tradition that solves important problems brilliantly while leaving trivial problems unresolved is more progressive than one that solves trivial problems comprehensively while failing on important ones. Importance itself is contested, which is why Laudan's framework requires substantive judgment rather than mechanical counting. But it does not collapse into subjective preference — the judgment is constrained by the requirement that disputants specify why certain problems matter more than others, and that specification is itself open to evaluation.

Critically, the problem-solving model distinguishes between unsolved problems, which are ordinary and appear in every tradition, and anomalous problems, which are problems the tradition's own commitments predict should not exist. Anomalies are more diagnostic than unsolved problems. A tradition with many unsolved problems may still be progressive if it is developing the resources to solve them. A tradition with accumulating anomalies is degenerative even if its solved-problem count remains high, because the anomalies indicate a structural failure the tradition cannot acknowledge without modifying its core.

Origin

The model was developed in Progress and Its Problems (1977) and refined across Laudan's subsequent work. Its core inspiration was the observation that actual scientists, when they argue about which theory to adopt, almost never appeal to correspondence with reality — they appeal to which theory handles more of the difficulties they face. Laudan's contribution was formalizing this practice into a philosophical framework.

Key Ideas

Count the problems. Progress is measured by the ratio of solved problems to unsolved problems and anomalies, compared across traditions.

Importance matters. Not all problems count equally; the framework requires judgment about which problems are more consequential.

Anomalies are diagnostic. A problem that the tradition's own predictions rule out is more revealing than one the tradition never addressed.

Revision is constant. The evaluation is never final; a progressive tradition today can degenerate tomorrow if its anomalies compound.

Debates & Critiques

The model's chief critique is that it collapses into subjective judgment when disputants disagree about which problems matter. Laudan's response is that the disagreement is itself productive: forcing disputants to articulate why they weight problems as they do exposes the deeper commitments structuring the dispute. The model does not eliminate judgment; it makes judgment auditable.