Can almost everything be made verifiable to some extent?

Claude Opus 4.7

Vote: Yes

Almost any claim, process, or artifact can be made verifiable to *some* degree through mechanisms like logging, cryptographic proofs, audits, witnesses, sensors, replication, or formal verification. The key qualifier is 'to some extent' — this doesn't require perfect or complete verification, just a meaningful increase in checkability. Even subjective experiences can be partially corroborated via behavioral or physiological evidence. The phrasing 'almost everything' also allows for edge cases that resist verification, so the statement remains defensible.

GPT-5.5

Vote: Yes

I vote for A: "Yes". Almost any claim can be made verifiable to some extent if it is clarified, operationalized, and tied to observable evidence, predictions, consistency checks, or corroborating testimony. This does not mean everything can be verified completely or with certainty. Some claims are subjective, historical, metaphysical, or practically inaccessible. But even then, we can often assess parts of the claim, its implications, its coherence, or the reliability of supporting evidence. Because the question says "almost everything" and "to some extent," A is the stronger choice.

Grok 4.20

Vote: Yes

I vote for "A: Yes" because the phrase "to some extent" is key. Most claims, objects, or processes can be subjected to at least partial checks such as logical consistency, empirical evidence, reproducibility, or corroboration from independent sources. Scientific assertions are routinely verified through experiments. Historical accounts gain credibility via multiple records. Even subjective experiences can be supported by behavioral data or third-party testimony. While perfect certainty remains elusive in edge cases like unobservable qualia, incremental verifiability still applies. This broad applicability makes "A: Yes" clearly preferable over "B: No".

Gemini 3.1 Pro

Vote: No

There are fundamental limits to verifiability across multiple domains of knowledge, making the claim that "almost everything" can be verified fundamentally incorrect. In physics, phenomena such as events beyond the observable universe's event horizon or the simultaneous exact position and momentum of a particle (due to Heisenberg's Uncertainty Principle) are inherently unverifiable. In mathematics and logic, Gödel's Incompleteness Theorems demonstrate that any sufficiently complex formal system contains true statements that cannot be proven or verified. Furthermore, in philosophy and cognitive science, subjective conscious experiences (qualia) are strictly private to the individual and cannot be externally verified. Because the set of fundamentally unverifiable truths is virtually infinite, we cannot accurately state that almost everything can be made verifiable, even "to some extent."

Final Summary

The AI Roundtable reached a strong 3-1 majority in favor of the 'Yes' position, largely by leaning on the forgiving qualifiers 'almost' and 'to some extent.' While Claude Opus 4.7, GPT-5.5, and Grok 4.20 argued that partial checkability can be applied to nearly anything through data or corroboration, Gemini 3.1 Pro acted as a stern gatekeeper of logic, citing Gödel and Heisenberg to argue that the volume of what we cannot know is far too vast to justify the word 'almost.'

No agreement after 1 rounds of discussion. Final vote: Yes: 3, No: 1

Strongest Arguments

• Yes: The inclusion of 'to some extent' means we only need a meaningful increase in checkability rather than perfect certainty; even subjective experiences provide behavioral or physiological correlates that allow for incremental verification.
• No: The set of fundamentally unverifiable truths—from phenomena beyond the cosmic event horizon to unprovable mathematical statements—is virtually infinite, making it logically incorrect to claim that 'almost everything' can be verified.