Can finite evidence ever certify that an observed event was not predetermined?
This essay argues that the answer is conditional rather than absolute. Certification is possible only relative to a restricted and exhaustible class of competing deterministic explanations.
If the admissible deterministic alternatives remain inexhaustible, every finite observational record leaves some deterministic explanation uneliminated.
1. The Question
Can finite evidence ever certify that an event was not predetermined?
This essay proposes a conditional answer.
Finite evidence can certify randomness only if the competing deterministic explanations belong to an exhaustible class. If admissible deterministic explanations remain inexhaustible, every finite observational record leaves some deterministic explanation uneliminated.
2. Four Meanings of Randomness
Randomness is not one idea.
A sequence may be statistically random, algorithmically random, epistemically unpredictable, or ontically indeterministic.
These notions frequently overlap but they are not equivalent.
The question investigated here concerns only ontic randomness: whether the physical state of the world uniquely fixed the outcome before it occurred.
3. Product and Source
Experiments reveal products.
They do not directly reveal their sources.
Observing an apparently random outcome is therefore different from observing the mechanism that generated it.
The distinction between product and source is the foundation of the entire argument.
4. Evidence as Elimination
Evidence advances by eliminating competing explanations.
Whether evidence can settle a question therefore depends on whether the competing explanations can all be eliminated.
This observation shifts the discussion away from randomness itself toward the structure of the explanatory space.
5. Exhaustible and Inexhaustible Classes
If the competing deterministic explanations form a finite and exhaustible class, finite evidence may eventually eliminate them all.
If the admissible deterministic class is inexhaustible, every finite investigation leaves further deterministic explanations alive.
The limitation is structural rather than computational.
6. Physics
Bell experiments, quantum information, and related results dramatically restrict deterministic explanations.
They do not eliminate every logically possible deterministic completion.
Their achievement is therefore best understood as narrowing the admissible hypothesis class rather than closing it.
7. Proof versus Evidence
Proof excludes every admissible competitor.
Evidence identifies the best explanation among explicitly defined competitors.
Science succeeds because it operates in the second sense.
Conclusion
The possibility of certifying randomness is not determined by the evidence alone.
It depends equally on the explanatory space against which the evidence is evaluated.
The central thesis of this essay is therefore simple:
Finite evidence can certify randomness only when the competing deterministic explanations belong to an exhaustible class.
