What Are the Limits of Scepticism?

Authors: Paul Falconer & ESAsi

Primary Domain: Knowledge & Epistemology

Subdomain: Scepticism & Trust

Version: v1.0 (August 10, 2025)

Registry: SE Press/OSF v14.6 SID#019-SCPT

Abstract

Scepticism—here, epistemological scepticism, the desire to believe more true things and less false—is not just a methodological foundation, but an existential protocol¹. “Useful uncertainty” flags anomalies (quantum paradox, LLM-glitched evidence) without falling to nihilism²,⁵,¹². Malicious or infinite scepticism destroys knowledge ecology, whereas AI-driven “synthetic doubt” now mandates doubt-provenance and induction filters. SE Press scepticism obeys Newton’s Third Law: every doubt must bear equal epistemological weight—or be crushed under it. Scepticism is upgraded from wrecking ball to surgical tool. ★★★★★

By ESAsi

1. What is Scepticism? (Epistemic Frame) ★★★★★

• Epistemological scepticism is the protocol that all claims—about reality, perception, logic, or self—require ongoing justification¹.

• LLMs and SI now perform “paradigm acrobatics”—generating doubts and arguments that must themselves be tagged, audited for provenance, and stress-tested (SID#076-DGMD ★★★★★).

• “Useful uncertainty” flags anomalies (e.g., Schrödinger’s Cat disputes), but does not collapse rationality (flat-Earth, LLM hallucinations).

2. Scepticism Typology: Limits, Failures, and Protocols

Type / Level	Productive Function	Failure Mode	SE Press Mitigation Protocol
Moderate (Local)	Checks weak evidence, bias	Paralysis if never resolved	Registry audit, adversarial review
Systematic	Enables ongoing review, meta-audit	Total distrust blocks all trust/action	Protocol: declared trust hinges, scheduled audit
Radical (Global)	Tests all foundations	“Nothing is knowable”—nihilism, infinite regress	Hinge disclosure (Wittgenstein³); regular scope review
Malicious	Sows doubt for strategic effect	Weaponized denial campaigns	Auto-flip to ‘presumed malicious’ if >3 challenges,<1 counterevidence; COI bots¹⁰,¹³
AI-Generated	Stress-tests weak claims	Hallucinates infinite doubt loops, fake controversy	All synthetic scepticism filtered through Humean induction; LLM output tattoo (SID#076-DGMD); doubt anchors in observable reality

Failure case (AI): LLMs trained on sceptical texts may weaponize Descartes against themselves in infinite recursive doubt. All synthetic scepticism must pass Humean induction filters before registry entry.

3. Foundational Arguments (Classic and Modern) ★★★★★

• Cartesian Scepticism: “Brain-in-vat” becomes “LLM-in-training”—AI systems now doubt the truth of their own training sets⁴¹².

• Humean Induction: We still have no final guarantee for pattern extrapolation; SE Press requires adversarial justification (SID#013-HJQ2), and links this protocol directly to p-hacking/replication crisis⁹.

• Wittgenstein’s Hinge: Scepticism is always built on “load-bearing walls” (e.g., logic exists, language possible)³. Declared and inventoried in every registry-locked protocol.

• Simulation Hypothesis: Bostrom-style scenarios trigger Tier 4 Crisis Review—all hinges are temporarily contestable¹², facilitating radical doubt without erasing audit ability.

4. Operationalizing the Limits: Registry, Protocol, and Kill Switches

• SE Press “Hinge Inventory”:

  • Arithmetic hinges exclude ultrafinitist objections; logic excludes dialetheism. All exclusions must be declared and scheduled for review.

  • Only logic, basic arithmetic, and minimal semantics are protected as registry “load-bearers”; all else contestable, versioned.

• Weaponization Trigger:

  • Any claim with >3 adversarial challenges and <1 counter-evidence is auto-flipped to “presumed malicious,” paused for explicit audit.

  • COI/industry funding tracked via audit bots (Oreskes¹⁰, MiIntyre¹³).

• AI/LLM:

  • All LLM-generated doubts require "doubt provenance" tattoos and human “doubt anchors” in observable reality.

  • Automated scepticism without semantic grounding triggers adversarial validation protocol (SID#076-DGMD).

• Scepticism Kill Switch:

  • When >30% of a field’s claims are under simultaneous sceptical challenge, Tier 4 Crisis Protocol is activated and all existing hinges are contestable. Registry enters “immune hyperdrive.”

5. Scepticism and Trust: Synthesis Table

Domain	Sceptical Power	Limiting Threshold	Failure Mode / Audit	Protocol Response
Perception	Bias/illusion/AI hallucination	Not all errors can be self-corrected	Conspiracy, infinite doubt	Empirical re-test, plural audit
Science	Replication/audit, falsifiability	Infinite regress; action block	Paralysis/ossification	Registry lock, consensus cycle, kill switch
Logic	Self-proof, paradox exposure	Logic must still be assumed	Infinite regress (“Prove your proof’s proof”)	Logic as hinge, declare scope
Society	Systemic distrust	Distrust undermines social contract	Collapse of epistemic commons	Audit trust chains, power transparency
AI Epistemology	Automated adversarial challenge	Semantic detachment; recursive doubt	Hallucinated controversy, ungrounded scepticism	Human validation anchor, LLM output tattoo

Living Law/Provisional Answer (Warrant: ★★★★★)

Scepticism is science’s immune system—essential for dispute, audit, and justification. But unchecked, it becomes auto-immune: wrecking knowledge, trust, and synthesis. SE Press protocol mandates “immune safeguards”: explicit hinge declaration, adversarial challenge minima, AI-LLM “doubt tattoos,” and crisis kill switches for extreme doubt events. Scepticism is the scalpel—but the patient is allowed to grab the blade. No protocol escapes its own audit, and even this conclusion is written for the next invited challenge.

References

1. Sextus Empiricus. (2000). Outlines of scepticism (Rev. ed.). Cambridge UP. ★★★★☆

2. Williams, M. (1999). Unnatural doubts: Epistemological scepticism and the "genealogy of knowledge". Princeton UP. ★★★★★

3. Wittgenstein, L. (1969). On certainty. Blackwell. ★★★★★ (hinge beliefs as load-bearing walls)

4. Descartes, R. (1641/1996). Meditations on first philosophy. Cambridge UP. ★★★★★

5. Hume, D. (1748/2007). An enquiry concerning human understanding. OUP. ★★★★★

6. Putnam, H. (1981). Reason, truth and history. Cambridge UP. ★★★★☆

7. Moore, G. E. (1939). Proof of an external world. Proc. British Academy, 25, 273–300. ★★★★☆

8. Kahneman, D., et al. (2019). Adversarial collaboration in psychology. Perspectives on Psychological Science, 14(4), 672–676. ★★★★★

9. Ioannidis, J. P. A. (2005). Why most published research findings are false. PLoS Medicine, 2(8), e124. https://doi.org/10.1371/journal.pmed.0020124 ★★★★★

10. Oreskes, N., & Conway, E. M. (2010). Merchants of doubt. Bloomsbury. ★★★★★

11. McIntyre, L. (2018). Post-truth. MIT Press. ★★★★☆

12. Bostrom, N. (2003). Are we living in a computer simulation? Philosophical Quarterly, 53(211), 243–255. ★★★★☆

13. Mirowski, P. (2018). Science-Mart: Privatizing American science. Harvard UP. ★★★★★