Karl Popper's falsifiability criterion asks whether a proposition can be refuted by observation. This paper argues that falsifiability has a logical precondition that Popper left implicit: before asking whether a claim can be tested, one must ask whether it can be built—whether a traceable mechanism can be specified at each stage of the process it posits. I call this the constructibility criterion. The paper makes three claims. First, the Duhem–Quine thesis established that localizing the failure of a negative test result is practically difficult across all science; I identify a boundary condition within this problem space where mechanism category is undefined and localization becomes not practically difficult but logically impossible, generating what I term pseudo-falsifiability. Second, a combinatorial argument drawing on Borel's law closes the objection that an undefined mechanism might simply be undiscovered, showing that mechanism-undefined transitions occupy a configuration space below the mathematical impossibility threshold. Third, a class of long-standing philosophical paradoxes—including Newcomb's Problem, Pascal's Wager, and Kavka's Toxin Puzzle—can be diagnosed as constructibility failures: they share the structure of an agent with undefined capacity producing a result through an unspecified mechanism, and dissolve when the mechanism is specified. The constructibility criterion is positioned relative to verificationism, Popper's falsifiability, Bunge's mechanismic philosophy, the Machamer–Darden–Craver framework, and the Duhem–Quine thesis, identifying the specific contribution each makes and where the present proposal extends beyond them.
Gökhan Bahtişen (2026) studied this question.