Contribution
A survey that sorts the manipulation literature by case type and by kind of reply. For the record it supplies Pereboom’s Case 1 in his words (2024 wording), a description of all four cases and of how his argument concludes, and the case types: mid-life, original design, natural force and parallel. Each type is matched to the reply it defeats.
Key insight
The manipulation argument is two premises. No-FR: the manipulated agent is not responsible. No-Diff: she does not differ relevantly from an ordinary determined agent. Each soft-line condition that tries to deny No-Diff is met by some type of case. A bypassing condition is met by original design, where nothing developed is bypassed. A manipulator condition is met by a natural force, where there is no manipulator. Each hard line has to swallow No-FR’s intuition, or explain it away.
Assumptions
- Direct responsibility only. The epistemic condition is “not typically in question”, so a victim’s lack of responsibility is put down to the freedom condition (§3).
- Determinism in the cases. The classic cases are “set in deterministic universes” (§2), except the parallel cases of §6.
Key results
- Case 1, quoted (§2, from Fischer et al. 2024, 98). “A team of neuroscientists has the ability to manipulate Professor Plum’s neural states at any time by remote optogenetic neural stimulation. In this particular case, they do so by pressing a button just before he begins to reason about his situation, which they know will produce in him a neural state that realizes a strongly egoistic reasoning process, which they know will deterministically result in his decision to kill White. Plum would not have killed White had the neuroscientists not intervened.” He is “stipulated to satisfy various compatibilist conditions on responsibility”, for example “a process that is appropriately responsive to reasons, and conforms to higher-order attitudes”. Endnote 2 lists versions in Pereboom 1995, 2001 (pp. 112–113), 2014 (pp. 76–77), 2017 and 2021.
- Cases 2 to 4, described (§5). In Case 2 Plum is “pre-programmed to behave in the way that his manipulators intend”, an original-design case like Ernie (n. 4). In Case 3 he “grows up in an egoistic society that instills in him the very same tendencies”. In Case 4, “an ordinary deterministic world”, he “comes to have the same tendencies and to perform the same action without any manipulators in the picture”. “Pereboom begins with Case 1 and attempts to carry the No-FR judgment over to Case 4 by arguing that there is no relevant difference between any two adjacent cases.”
- Pereboom’s conclusion is abductive (n. 23). “Having presented the four cases, and argued that there is no relevant difference between cases 1–4, Pereboom argues that the best explanation of the Plums’ lack of responsibility for their decisions is that the decisions were causally determined by factors outside of the Plums’ control.”
- Mele’s cases (§2). In Chuck and Sweet Sally, Sally’s values are replaced overnight with Chuck’s, and many judge Chuck responsible and Sally not. In Ernie, Diana makes a zygote so that Ernie does A thirty years on. Ernie is “covert non-constraining control” in Kane’s sense (n. 5).
- The schema and its variants (§3). No-FR, No-Diff, therefore compatibilism is false. A stronger conclusion, that determinism explains why, needs further premises (n. 8). Weaker ones target existing views only.
- The soft-line worry (§4). “For any soft-line view one offers … the proponent of a manipulation argument will simply be able to produce a new case in which Victim does meet the proposed necessary condition.” With Ernie, “bypassing views cannot account for a difference … since there is no difference in terms of whether their capacities, once developed, were bypassed”.
- Hard lines (§5). Frankfurt: freedom does “not depend on the origin of the conditions”. Fischer: intuitions track autonomy, not responsibility. Natalie, a natural-force case, answers the worry that reactions to a manipulator distort judgements about the victim. McKenna reverses the argument: start agnostic at Case 4 and carry agnosticism back to Case 1. McKenna 2004: manipulated agents are responsible for fewer things, not for the character they act from. Cyr, and Montminy and Tinney: blame is mitigated.
- Parallel cases (§6). Paru is manipulated in an indeterministic world where she could have done otherwise. If she is still not responsible, libertarians are pressed too. Only one original-design parallel case is known (Cyr 2020b).
- Dilemmas (§7). King: the libertarian needs a contentious sourcehood thesis. Kearns: either manipulation itself undermines (so No-Diff fails), or it is idle. Vihvelin: the “tool” explanation either justifies a difference or discredits the intuition.
- Todd (§8). The modified argument needs only that Ernie’s blame be mitigated. A second argument asks why God lacks standing to blame Ernie.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Pereboom’s Case 1 agent meets reasons-responsiveness and hierarchical conditions yet is intuitively not responsible | strong | quoted, §2 |
| C2 | Pereboom’s argument concludes by best explanation, not only by no-difference | moderate | reported, n. 23 |
| C3 | Original-design cases defeat bypassing soft-line views | moderate | argued with Ernie, §4 |
| C4 | Natural-force cases defeat manipulator-focused views and the distortion worry | moderate | argued with Natalie, §5 and n. 21 |
| C5 | Parallel indeterministic manipulation cases press libertarians as well | moderate | Paru, §6 |
Concepts
- No-FR and No-Diff. The two premises above.
- Mid-life and original-design manipulation. Intervention on an existing agent, against design before the agent exists.
- Natural-force case. The same change produced with no manipulator.
- Parallel case. A manipulation case in which the victim meets libertarian conditions.
- Hard-line and soft-line. Relative to a case, not to a theorist (§4).
Connections
It is Part 1 of three. Parts 2 and 3, on manipulator-focused and bypassing views, were not read. It builds on McKenna 2008 for the dialectic and on Mele 2006 and 2019 for the cases. Fischer and Ravizza 1998, chapters 7–8, are listed among bypassing views (n. 15), which is where their taking-responsibility condition belongs. SEP Compatibilism (LIT-291) and Caruso’s SEP entry (NOTE-862) cover the same ground from the compatibilist and the skeptic’s side.
Bearing on the record
- THEORY-040. Two points, the first stronger.
- First half (a structural condition is met by a manipulated agent). Case 1 meets reasons-responsiveness and “conforms to higher-order attitudes” by stipulation. Frankfurt’s hard line (“do not depend on the origin of the conditions”, 1988, 45–6) is a second first-hand quotation, cited in a survey, of the bullet THEORY-040 reports him biting.
- Second half (no historical condition escapes). Ernie defeats any bypassing condition, since nothing he has developed is bypassed. Fischer and Ravizza’s condition is listed as a bypassing view (n. 15). That is the strongest reason the record now holds for the second half, though it is reported, not read in Pereboom or Mele.
- The promote_when’s “first-hand reading of Pereboom’s four-case argument or Mele’s zygote argument” is not met. Mele’s Ernie is quoted (2006: 188–9), but not his argument.
- THEORY-121. Its refuter, a four-case version that “manipulates the taking of responsibility”, is not described here. The Ernie point bears on it. If taking responsibility is a bypassing-type condition, an original-design agent takes responsibility in the ordinary way, and the account must either count Ernie responsible (a hard line) or find another difference.
- QUESTION-071. Two things here bear on it.
- Its description of Pereboom’s cases is now confirmed in outline, and its qualification holds: Case 4 has no manipulator. Case 3, an egoistic society, is between the two. Someone else produces the character, but by upbringing, not intervention.
- Part 2 asks what history would tell a manipulated will from one that is merely effective. Original design is the hard case. A history that records only how capacities were formed, or whether they were bypassed, records nothing that separates Ernie from an ordinary agent.
- THEORY-119. Not addressed. Frankfurt cases appear only as the source of “hard line” Frankfurt.
- THEORY-328 is filed from this reading and NOTE-862.
- Nothing here is ML practice.
Limitations
- A survey. It quotes one of Pereboom’s four cases, in the 2024 wording with optogenetics, which differs from the 2001 text.
- It does not report a version of the cases built to defeat Fischer and Ravizza’s taking-responsibility condition.
- Parts 2 and 3, where soft-line views are assessed, were not read.
Open questions
- Whether any soft-line condition survives both original-design and natural-force cases. Parts 2 and 3 take this up.
- Whether Case 3, manipulation by upbringing, belongs with the manipulated cases or with ordinary determination. Most of the soft-line debate turns on it.