LIT-694: Alternate Possibilities and Moral Responsibility

Harry G. Frankfurt, “Alternate Possibilities and Moral Responsibility”, The Journal of Philosophy 66(23), 4 December 1969, pp. 829–839 — DOI-10.2307/2023833. Reprinted as chapter 1 of Frankfurt, The Importance of What We Care About (Cambridge University Press, 1988), pp. 1–10, “with minor stylistic changes” (LIT-693).

Key takeaways

• PAP is false, and its plausibility is borrowed from coercion. The paper opens by calling PAP “an a priori truth” to some and denies it: “A person may well be morally responsible for what he has done even though he could not have done otherwise” (pp. 829–830). The doctrine that coercion excludes responsibility is not a special case of PAP. Of the three ways Jones may face a threat to do what he has already decided on, only the stampeded Jones2, who acts “entirely because he was terrified”, is excused; Jones3, who acts on his prior decision though the threat would have moved him, keeps his responsibility “even though … he could not have avoided doing what he did” (pp. 831–833).
• The case. Black wants Jones4 to perform “a certain action”, watches, and will intervene by threat, potion, hypnosis or direct manipulation of the brain only if Jones4 is about to decide otherwise. Jones4 decides “for reasons of his own” to do just that, and Black never acts. Jones4 “will bear precisely the same moral responsibility” as without Black, and it would be “quite unreasonable to excuse Jones4 … or to withhold the praise to which it would normally entitle him” (pp. 835–836). The case is credited to Nozick’s lectures (n. 2). Black’s foreknowledge comes from a twitch that precedes one decision and not the other, which “does not beg the question of determinism” (n. 3). Black can be replaced by a machine or by “natural forces involving no will or design” (n. 4).
• Why PAP fails. A fact that makes an action unavoidable may play “no role whatever in the explanation of why he did it”; then it “could have been subtracted from the situation” with nothing changed, and “it seems quite gratuitous to assign it any weight in the assessment of his moral responsibility” (pp. 836–837). We accept “I could not have done otherwise” as an excuse only because we take it to explain the action.
• The replacement, and the determinism question. Frankfurt considers and rejects “not responsible if he did it because he could not have done otherwise”, noting that with the incompatibilist reading of “could have done otherwise” it would make every determined act excused (p. 838). He proposes instead “not morally responsible … if he did it only because he could not have done otherwise”, that is, not because “that was what he really wanted to do”, and says this “does not appear to conflict with the view that moral responsibility is compatible with determinism” (pp. 838–839).

Standing in the record

Filed on 2026-10-04 at the owner’s request, to file Harry Frankfurt “to the best of your ability”. It is moral philosophy and the metaphysics of responsibility, with no instruction for machine-learning practice, so no anthology topic holds it.

Active from a full reading (NOTE-546). Until now the record held this paper only through Robb’s SEP entry (LIT-294, NOTE-269), which quotes it through Fischer’s retelling, and McKenna and Coates’s SEP Compatibilism (LIT-291), which gives “a close approximation”. Three things in those readings change:

• The original case is schematic. It has no vote and no shooting: Black and Jones4 and “a certain action”. The voting version is Fischer’s (1982) and the shooting version is McKenna and Coates’s.
• The prior sign is there, in a note. Robb says the sign is “not always explicit in Frankfurt’s own FSCs” (§4.3.2). In this paper it is explicit in n. 3, the twitch. The dilemma defence that later turns on it is therefore aimed at a feature the original case has.
• Robb’s §5.1 citation of this paper needs its context. Robb cites it for the worry that “determinism may still preclude responsibility due to the way it blocks alternatives”. The paper raises that worry against the intermediate revision of PAP and then answers it with “only because”. That it closes in favour of compatibilism is the paper’s own view, not a report of an opponent’s.

Its successor. “Freedom of the Will and the Concept of a Person” (LIT-692, read) cites this paper (n. 9) for its claim that a responsible agent need not have been able to do otherwise, and supplies what “what he really wanted to do” lacks here: a willing addict who acts on a desire he wants to be his will. That extends is declared on that paper’s side.

The responsibility literature built on it, which the record has not read. Fischer’s “Responsiveness and Moral Responsibility” (LIT-687) and Fischer and Ravizza’s Responsibility and Control (LIT-684) are reported by the SEP to start from these cases: they move reasons- responsiveness from the agent to the mechanism that actually issues in the action. Both are Deferred and unread, so no relation is declared from here. A reading of either should add extends to this paper on its own side if it confirms the dependence.

How it bears on the record’s theories. THEORY-119 is the account filed from this paper: responsibility does not require the ability to have done otherwise. Fischer and Ravizza’s guidance control (THEORY-121) is its best-known development. Wolf’s asymmetry (THEORY-118) holds that PAP is false for praise and true for blame; this paper’s case is stated for both (“excuse Jones4 … or … withhold the praise”, p. 836), and its action is not said to be good or bad, so it claims the symmetric result Wolf denies. Wolf’s “Asymmetrical Freedom” (LIT-688) is unread, and whether she answers a Frankfurt case of wrongdoing is still open.

Access: Crossref and OpenAlex record no open-access copy (oa_status closed). The read copy is the instructor-posted course file named in the history. Cambridge’s publisher preview of the 1988 collection (api.pageplace.de) shows the reprint’s pp. 1–9, which served as a check on the OCR.