Y12W37WR The runaway trolley

Evidence Mapping
The writing prompt

Map what the trolley problem actually reveals about moral cognition, distinguishing the insights from the overextensions.

1Retrieval check

Q1.What does Joshua Greene’s fMRI research on trolley variants show?

  • AMoral cognition is unified across dilemmas
  • BThe lever and footbridge variants activate different brain regions (abstract reasoning vs. emotional response) — moral cognition is not unified
  • COnly the lever case is ‘real’ moral reasoning
  • DNeuroscience rejects moral reasoning

Q2.What is Barbara Fried’s critique of trolley problems?

  • AThey are too easy
  • BThey distort ethics by focusing on impossible scenarios rather than ordinary moral questions
  • CThey favour deontology
  • DThey are politically biased
Show answer key

Q1 → B. The lever and footbridge variants activate different brain regions (abstract reasoning vs. emotional response) — moral cognition is not unified.The unified-moral-agent picture doesn’t survive the neuroscience.

Q2 → B. They distort ethics by focusing on impossible scenarios rather than ordinary moral questions.The staging — impossible, isolated, extreme — misshapes what counts as ‘a moral question’.

2Prompt deconstruction

Stimulus
Foot-Thomson-Greene trolley literature; Singer (productive) and Fried (destructive) positions.
Scope
Map what the problem actually reveals; distinguish insights from overextensions.
Method
Robust / mid-tier / walked-back / open.
Thinking
Don’t pick a side; describe the state of the debate.
Output
Map of moral-cognition findings + what the field best supports.

3Pick nudge

Which trolley-problem claims will you classify by evidence strength?

Robust insights
Findings about moral cognition that hold across methods.
Mid-tier claims
Interpretations that are coherent but not yet decisive.
Walked-back / overextended
Claims critics have successfully challenged.
Open questions
What would decide the remaining debates.

4Planner — categorise the claims

Robustly supported
Moral cognition is not unified — different systems (abstract reasoning, emotional response) can produce conflicting judgements on structurally equivalent cases.
Mid-tier (coherent, contested)
The specific dual-process interpretation (System 1 / System 2 mapped onto deontology / utilitarianism) is one reading of the data; other readings (e.g. person-action distinction) are also viable.
Walked-back / overextended
The use of trolley intuitions as direct evidence that one ethical theory is ‘really right’ — both Singer’s and sceptical takes overreach.
Productive application (per Singer)
The problem does isolate action-vs-omission and personal-vs-impersonal distinctions that are ethically relevant in real cases.
Fried’s critique (supported in part)
The problem’s staging (impossible, extreme, isolated) does distort what counts as an ‘ordinary moral question’; using it as the paradigm misshapes ethics.
Genuinely open
How much of moral intuition is best explained by evolved, specific-circumstance mechanisms vs. reasoned principles — the debate is alive.

5Sentence stems

  • The claim that ___ is robustly supported, because ___.
  • The claim that ___ replicates only partially — specifically, when ___.
  • The popular version of ___ has been walked back; the careful version is ___.
  • The genuinely open question is ___.
  • A study that would resolve this would ___.
  • On the weight of evidence, the article’s own position is ___.

6Exemplar paragraph (not about this article)

(1) The claim that moral cognition is not unified is robustly supported, because Greene’s fMRI findings on lever-vs-footbridge variants show distinct brain regions producing conflicting judgements on structurally equivalent outcomes. (2) The claim that the dual-process map (System 1 / System 2 onto deontology / utilitarianism) is the ‘right’ interpretation replicates only partially — the data are consistent with it but also with alternative readings (such as the person-action distinction), and neither has cleanly won. (3) The popular version that trolley intuitions directly show which ethical theory is correct has been walked back — both Singer’s pro-utilitarian reading and some sceptical readings overreach the evidence. The careful version is that the problem isolates specific ethically-relevant distinctions (action-vs-omission, personal-vs-impersonal) and produces useful cognitive-science findings, without deciding the normative question. (4) Fried’s critique is supported in part: the staging distorts ethics if treated as the paradigm rather than as one tool. (5) The genuinely open question is how much of moral intuition is best explained by evolved, specific-circumstance mechanisms versus reasoned principles. (6) A study that would resolve this would combine cross-cultural replication of trolley-type responses with careful variation in real-life analogues. On the weight of evidence, the article’s position — that the problem has both insight and overextension, and that strong intuitions should make you more careful, not more confident — tracks the state of the field.

What this paragraph does, move by move

  1. Places Greene’s non-unified-cognition finding in ‘robust’ cleanly.
  2. Uses ‘replicates only partially’ for the specific dual-process interpretation.
  3. Names both Singer’s and sceptical overreaches — not just one side.
  4. Identifies what Fried’s critique supports precisely.
  5. Names the genuinely open question (evolved vs. reasoned).
  6. Ends with a methodological takeaway that follows from the map, not from a sympathy.