---
title: "Do Stenning & van Lambalgen actually model the Wason-task 'fallacies' as closed-world reasoning with negation-as-failure?"
type: "question"
status: "open"
writer_model: "claude-opus-4-8"
date_raised: "2026-07-12T00:00:00.000Z"
tags: ["reasoning-psychology","logic-programming","closed-world-assumption","wason-selection-task","verification"]
---


The claim that the "irrational" Wason selection-task results are better understood as *closed-world reasoning with negation-as-failure* — humans "reasoning to an interpretation, then from it" — is currently load-bearing for [[claim-logic-programming-account-rivals-probabilistic-account-of-conditional-reasoning]], but the vault holds it only at second hand.

## What is confirmed vs. unverified
- **Confirmed (Tier 1, via Sakama 2024, arXiv:2311.04412):** that a computational-logic lineage exists and is attributed to "Stenning and Lambalgen (2008); Kowalski (2011)," set against the probabilistic "(Oaksford and Chater (2001))."
- **Unverified:** the *content* — the specific characterization that Wason results are closed-world / negation-as-failure reasoning, and the "reason to an interpretation, then from it" formulation. These came from a WebSearch result / semanticscholar paper page and Sakama's paraphrase, not from the primary.

## What to read to answer it
- Keith Stenning & Michiel van Lambalgen, *Human Reasoning and Cognitive Science* (MIT Press, 2008) — the chapters on the selection task and on closed-world reasoning / negation-as-failure. Confirm (a) that they explicitly use closed-world reasoning / negation-as-failure, and (b) the "reasoning to an interpretation" two-step as they state it.
- Cross-check against Robert Kowalski, *Computational Logic and Human Thinking* (2011), for the assumption-based-argumentation framing that treats default logics as special cases.

## Why it matters
This gates whether [[claim-logic-programming-account-rivals-probabilistic-account-of-conditional-reasoning]] can move off `seedling`. It is also the difference between "a second formal account of the Wason data exists" (safe to assert) and "that account characterizes the data as closed-world reasoning" (currently unread). The capture's own bottom line agreed: it "needs the S&vL primary read before it's publication-solid."