Research  ›  Case library  ›  Case 17: Sorites / Heap Paradox
Case 17 / 31  •  Theme: Doing

Sorites / Heap Paradox

Removing one grain seems unable to turn a heap into a non-heap. Repeating that step conflicts with the clear endpoint: one grain is not a heap.

When does a heap stop being a heap if you remove one grain at a time? Sorites sets a seemingly harmless tolerance step against clearly different endpoints. Our AI analogy is a research hypothesis: across small successive substitutions, local judgements about adequate human recovery capability may lose sight of the overall trajectory.

✓
Research statusResearch in preparation
☆
AI relevanceAI may amplify an existing effect
⚙
Core ideaVagueness / cumulative threshold
⌾
Evidence baseSources and limitations summarised; independent review pending.
◇
Case typeLogical paradox
“
In one sentence

Removing one grain seems unable to turn a heap into a non-heap. Repeating that step conflicts with the clear endpoint: one grain is not a heap.

The storyHow the case arises and how it can mislead our thinking.

The classic Sorites argument holds that a difference of one grain cannot turn a heap into a non-heap. Repeating this tolerance step seems to imply that even one grain must ultimately be a heap — an unacceptable conclusion.

What do we know?Evidence, research, critical insights and what remains unknown.

The philosophical problem of vagueness is canonical and unresolved. The organisational transfer is a hypothesis about stepwise judgement, cumulative substitution and consequence-relevant thresholds.

What is the mechanism?Which explanation best fits what we observe.

The paradox arises from combining clear starting and ending cases, the general tolerance assumption that one small step never changes the category, and repeated inference. Which assumption or inference should be revised remains philosophically disputed.

What does AI change?How AI may amplify the mechanism or make it visible.

AI transfer to investigate: AI enables many small substitution decisions that each appear rational. The cumulative capability trajectory may therefore become less visible.

What do we want to test?The specific question we want to investigate further.

Proposed test, not yet conducted: Give both groups of reviewers the same substitution trajectory and recovery evidence. Display the steps separately or as an overall trajectory. Measure their judgement of adequate recovery capability against a prespecified operational threshold. This tests judgements; it does not establish a change in their own recovery capability.

What do we take away?What this case provisionally means — and where its limits lie.

A chosen operational threshold does not resolve philosophical vagueness, but it may make governance drift manageable.

Sorites concerns vague predicates, local tolerance and contrasting endpoints. An operational recovery threshold enables a practical test but does not resolve the semantic boundary question.

SourcesPrimary and key scholarly sources.

Williamson (1994), Vagueness: bibliographic reference; the exact book passage is not part of the source basis for this account.

Williamson (1997), Imagination, Stipulation and Vagueness, pp. 215–216, §2: supplementary primary text.

Williamson’s epistemicism is one philosophical position. His 1997 article serves as an additional primary text; it does not replace an assessment of the relevant passage in the 1994 book. Neither establishes the proposed AI application.

Give feedback

Diagram at full size