The distinction draws on Lakatos's concept of progressive and degenerative research programs but refines it for Laudan's problem-solving framework. Where Lakatos emphasized predictive novelty as the mark of progress, Laudan emphasized problem-solving capacity more broadly — including the capacity to resolve conceptual tensions, address anomalies, and integrate findings from neighboring domains.
A progressive tradition demonstrates specific behaviors. It names its anomalies rather than hiding them. It develops theoretical resources to address them rather than dismissing them. It integrates findings from competing traditions rather than dismissing them as category errors. It modifies its commitments when the evidence warrants, and does so in ways that preserve rather than destroy its problem-solving capacity in other domains. The modification is difficult to execute well — expand too little and anomalies compound; expand too much and the tradition loses the coherence that made it productive.
A degenerative tradition exhibits the opposite pattern. It frames its anomalies as temporary exceptions, or as failures of the observers rather than of the theory. It increases the distance between its predictions and its observations by introducing ad hoc modifications that solve no new problems and explain no new phenomena — the move Lakatos called adding epicycles. It retreats to its core commitments when challenged, treating those commitments as self-evident rather than defensible.
Applied to the AI discourse, the distinction identifies two failure modes for each tradition. The triumphalist tradition degenerates when it dismisses productive addiction, depth atrophy, or task seepage as adjustment costs that will resolve themselves. It progresses when it acknowledges these anomalies and develops frameworks — AI Practice, structured pauses, organizational dams — to address them. The elegist tradition degenerates when it treats any defense of AI tools as capitulation. It progresses when it distinguishes between formative friction and exclusionary friction, acknowledging that not all barriers are pedagogically valuable.
The distinction was developed in Progress and Its Problems (1977), building on Imre Lakatos's work on progressive and degenerating research programs. Laudan's version differed from Lakatos's in two key respects: it measured progress by problem-solving rather than predictive novelty, and it treated degeneration as a gradient rather than a binary condition.
Growth versus contraction. Progressive traditions expand to meet new problems; degenerative traditions contract to exclude them.
Anomaly response as diagnostic. How a tradition handles its anomalies is the most reliable indicator of its trajectory.
Both traditions can degenerate. Neither the triumphalist nor the elegist tradition is inherently progressive; each can fail by suppressing what it cannot explain.
Revision without collapse. Progressive modification preserves the tradition's problem-solving capacity; degenerative modification preserves only its appearance.