Contribution
It argues that the distinction between leeway and source incompatibilism, which Pereboom, Kane, McKenna and others draw, does not survive reflection on what each needs. Leeway incompatibilism must appeal to the agent’s being the source of how her alternatives resolve; source incompatibilism, in its best form, entails that some alternative exists. It is an argument about which incompatibilism to hold, not for incompatibilism (p. 145).
Key insight
Ultimacy and alternatives are tied by indeterminism. If nothing outside the agent determines her action, more than one future is compossible with the past and the laws, so she has some alternative, however poor. The alternative does no work in explaining her responsibility, but its absence is enough to show she lacks ultimacy (p. 150).
Assumptions
- Incompatibilism is assumed throughout. Causal determinism is taken to rule out responsibility; the question is only why.
- Ultimacy requires that the agent “not be causally determined by anything outside of herself” (p. 150), allowing for self-imposed determinism traced to earlier undetermined acts (Ekstrom’s “at each moment”, the tracing condition).
- Robustness is set aside. Timpe brackets the debate over how robust an alternative must be (p. 148) and argues the entailed alternatives matter whatever its outcome (p. 150).
- Blockage cases are not refuted here. A full case for wide over narrow source incompatibilism would need to argue against them; the paper takes the “positive route” instead (p. 148).
Key results
- Three claims (pp. 143–145). The Taxonomy Claim: there are two kinds of incompatibilism, leeway and source. Pereboom states it: “We might call those incompatibilists who incline towards the view that an alternative possibilities condition has the more important role in explaining an agent’s moral responsibility leeway incompatibilists, and those who are predisposed to maintain that an incompatibilist condition on the causal history of the action plays the more significant part source incompatibilists” (Pereboom 2003, 186). The Refutation Claim: Frankfurt-style counterexamples show alternatives are not required, so leeway incompatibilism is false (Stump 1996, 88: “Frankfurt-style counterexamples are successful against PAP; but … the libertarian has nothing to fear from them”). The Inconsistency Claim: the ultimacy source incompatibilism needs is inconsistent with truths about the actual world. “Pereboom’s Living Without Free Will is an exemplar”: agent-causal libertarianism “involves no detected logical inconsistency. But this does not mean that it is in the clear… The most significant empirical objections to agent-causal libertarianism challenge its capacity to accommodate our best natural scientific theories” (Pereboom 2001, 68–69), and reconciling it with physics “involves coincidences so wild as to make it incredible” (Pereboom 2001, 85). If all three claims hold, responsibility exists only if compatibilism is true.
- Against naive leeway incompatibilism (pp. 145–147). In a “Fischer scenario”, Kelvin either decides to go jogging or is smitten dead by God at that instant; he has an alternative, but it would be “ludicrous” to ground his praise in it. Mere alternatives fall to the luck objection. So a leeway incompatibilist must make the resolution of the leeway the agent’s own, not determined by anything outside her, which is the language of source, origination and ultimacy, “the realm of Source Incompatibilism”. Fischer 1982 is quoted for the earliest source-style position (pp. 31–34): Frankfurt’s examples involve “alternate-sequence compulsion”, determinism “actual-sequence compulsion”.
- Narrow and wide (pp. 147–148). McKenna: “[t]he source incompatibilist agrees with Frankfurt as to the unimportance of alternative possibilities, but … [determinism rules out free will] because, if true, the sources of an agent’s actions do not originate in the agent” (McKenna 2003, 201–202). Some source incompatibilists still insist on an alternatives condition, holding the source condition “more ‘fundamental’ or ‘important’”: “This is true of both Robert Kane’s and Derk Pereboom’s views” (p. 147). Others (Stump, Zagzebski, Hunt) require no alternatives at all. Narrow source incompatibilists hold being the proper source needs no alternatives; wide ones hold that ultimacy is fundamental but that some PAP-like condition is also true.
- Ultimacy entails alternatives (pp. 148–151). With O’Connor: what Frankfurt cases show is only that freedom “is not to be identified with the existence of … alternatives” (O’Connor 2005, 209), and alternatives matter as “indicators of the self-determination manifested by one’s actions” (209–210). With Della Rocca: even a non-robust alternative shows “that the action was not determined by external factors”. The alternatives are not “some magical element” that makes for responsibility; they are a necessary, not sufficient, condition of what does, namely ultimacy. If the agent is “the difference-maker”, “there must be more than one future that is compossible with the conjunction of the past … and the laws of nature”. So “the Source Incompatibilist’s emphasis on the ultimacy condition will entail that the agent does have alternative possibilities available to her, even if the explanatory weight of those alternative possibilities does not extend beyond merely securing the falsity of causal determinism”, and “their absence is sufficient for the agent’s not satisfying the ultimacy condition” (p. 150).
- Not voluntary alternatives (pp. 150–151). Kane (PAP-V: could have done otherwise “voluntarily, intentionally and rationally”) and Fischer (any alternatives “must contain voluntary behavior”, 2002, 287–288) ask too much. In the revised Kelvin case, Kelvin meets the ultimacy condition only in the actual sequence and the only alternative is being smitten; he is still the difference-maker, and responsible.
- Conclusion (pp. 151–152). “Insofar as the heart of the Source Incompatibilist’s position is some sort of ultimacy condition, she must also embrace an alternative possibilities condition.” Incompatibilism may not require PAP but requires some alternatives condition, so wide source incompatibilism is preferable to narrow source and to leeway incompatibilism, and the Taxonomy Claim should be rejected “because ultimacy and alternative possibilities are intrinsically related”. The remaining alternatives are morally relevant, “even if it is only in this derivative way”.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | No acceptable leeway incompatibilism rests on mere alternatives; it must appeal to source or ultimacy | moderate | the Kelvin case and the luck objection (pp. 145–146) |
| C2 | The ultimacy condition entails that the agent has some alternative possibility | moderate | from ultimacy to indeterminism to more than one compossible future (p. 150); holds for incompatibilist ultimacy only (n. 44) |
| C3 | The entailed alternatives need not be robust or voluntary | moderate | the revised Kelvin case (pp. 150–151) |
| C4 | Wide source incompatibilism is preferable to narrow source and to leeway incompatibilism | weak to moderate | C1–C3; the case against blockage cases, needed for narrow, is not made here (p. 148) |
| C5 | The taxonomy of leeway against source incompatibilism should be rejected | weak to moderate | follows from C1 and C2 |
| C6 | Kane’s and Pereboom’s views hold an alternatives condition while making the source condition more fundamental | as asserted | p. 147, with no page of Pereboom cited (n. 25) |
Concepts
- Leeway incompatibilism — alternatives play the more important role in explaining responsibility (Pereboom 2003, 186, as quoted).
- Source incompatibilism — an incompatibilist condition on the causal history of the action plays the more significant part (ibid.).
- Narrow source incompatibilism — being the proper source needs no alternatives at all.
- Wide source incompatibilism — ultimacy is fundamental, and some alternatives condition, entailed by it, is also true.
- Fischer scenario — the agent dies in every alternative sequence open to him but one (Fischer 2002, 288, as cited).
- Ultimacy — the agent is not causally determined by anything outside herself to act as she does, allowing self-imposed determination traced to her own earlier undetermined acts (p. 150).
Connections
THEORY-316 states Pereboom’s causal-history incompatibilism from a partial reading (NOTE-823). Timpe takes Pereboom’s own taxonomy as his target and places Pereboom on the wide side, keeping an alternatives condition, which agrees with NOTE-823’s report that Pereboom leaves open alternatives conditions “necessary for moral responsibility but nevertheless irrelevant to explaining why” (Pereboom p. 2). Fischer 2010 (LIT-1868) answers the Dilemma Defense that Timpe’s n. 31 points to through Goetz, and concedes that determinism may undermine “sourcehood” (p. 331).
Bearing on the record
- THEORY-316, not met; it sharpens a caveat rather than confirming or refuting. Its promote_when asks for a first-hand reading of Pereboom’s Chapter 1, pp. 10–37, and a checked leeway reply. Timpe quotes Pereboom’s book only at pp. 68–69 and 85, on agent causation, and cites p. 18 on blockage cases, so he supplies neither. On the account’s content, Timpe agrees that the source condition is the fundamental one, and argues that it entails a weak alternatives condition. That does not touch Step 2. It qualifies Step 1: Frankfurt-style cases do not defeat every alternatives condition, since the incompatibilist must keep the one ultimacy entails, which is what THEORY-316’s “Not that every alternative-possibilities condition is false” already allows. Timpe’s argument runs from ultimacy to alternatives, not the other way, so it is not the refuting argument the promote_when names, that condition (3) matters only because it secures alternatives.
- THEORY-119, that the principle of alternate possibilities is false by Frankfurt-style cases: Timpe accepts PAP may fail and holds that a weaker alternatives condition, entailed by ultimacy, survives for the incompatibilist.
- No instruction for machine-learning practice; nothing for the anthology.
Limitations
- The case for wide over narrow is incomplete by Timpe’s own account: blockage cases, which narrow source incompatibilists rely on, are not answered here (p. 148, n. 31).
- The entailment depends on an incompatibilist ultimacy condition (n. 44); compatibilist accounts of ultimacy are set aside for a later work.
- Second hand for Pereboom. Timpe’s classification of Pereboom as keeping an alternatives condition cites the book as a whole (n. 25).
- The scan’s text layer is poor. Quotations were read from column-wise extraction and are reported as they read there, with spacing restored.
Open questions
- Does Pereboom, in Chapter 1 or 4, accept an alternatives condition entailed by his causal-history condition, as Timpe says?
- Do blockage cases, which make the action unavoidable without determinism, show narrow source incompatibilism coherent after all?