The Free Will Defense: A Libertarian Perspective

Category: Free Will

  • Frankfurt’s Gambit: More Hole-in-One Than Airtight Argument?

    Frankfurt’s Gambit: More Hole-in-One Than Airtight Argument?

    The world of philosophy often grapples with questions that seem to have straightforward answers, until a clever thought experiment throws a wrench in the works. One such wrench is the “Frankfurt-Style Case” (FSC), a type of scenario designed to challenge a deeply intuitive idea: that to be morally responsible for an action, you must have been able to do otherwise. This is often called the Principle of Alternative Possibilities (PAP).

    FSCs try to show that someone can be responsible even if they had no other option. Imagine a scenario where an agent decides to do X, and is morally responsible for it. Unbeknownst to them, a nefarious character, let’s call him Black, was waiting in the background. If the agent had shown any sign of not doing X, Black would have intervened and forced them to do X anyway. But, as it happens, the agent does X “on their own,” so Black does nothing. The argument? The agent is responsible for X, even though, due to Black’s presence, they couldn’t have done otherwise.

    Sounds compelling, right? But what if successfully constructing such a case – one that truly and fairly demonstrates moral responsibility in the actual sequence where Black remains idle – is less like a logical checkmate and more like a golfer hitting a hole-in-one? The National Hole-in-One Registry states the odds for an average player are about 12,000 to 1. My research suggests that for an FSC to convincingly work, it might rely on a similar stroke of incredible, compounded luck.

    The Double Whammy: Not Just Luck, but Layers of It

    The core issue isn’t merely that a bit of chance might creep into these scenarios. It’s that FSCs, particularly the more sophisticated indeterministic versions (which don’t assume a clockwork, deterministic universe), are fundamentally riddled with layers of contingency. For an FSC to “succeed” in its aim – showing the agent is responsible for their action without Black needing to intervene – a whole series of fortunate events must align perfectly.

    Let’s break down these layers of luck:

    Layer 1: The Agent’s “Lucky Swing” – Processual Chance

    First, there’s Processual Chance. This refers to all the contingent factors that influence the agent’s own decision-making process and the subsequent action, including how things actually turn out. Think of it as the golfer’s swing and the initial flight of the ball:

    • Circumstantial Luck: The specific situation the agent finds themselves in, the information available (or not available) to them, sudden distractions, or influences. (Is the wind just right?)
    • Constitutive Luck: Who the agent is – their character, dispositions, ingrained habits, even their current emotional state, all shaped by factors largely beyond their ultimate control. This is where the Principle of Deliberative Contingency (PDC), which we’ll explore shortly, starts to play a crucial role. (Is the golfer naturally gifted or having a “good day”?)
    • Resultant Luck: This is the luck in how the agent’s actions actually turn out. Even if they decide to do X, actually accomplishing X as intended, and it being the specific X that Black desires, involves luck in the outcome. (Does the ball, having been struck, actually head towards the green as intended?)
    • Causal Luck (in indeterministic FSCs): If the universe, or at least the agent’s mental processes, aren’t strictly determined, then the very unfolding of their thoughts and the formation of their intention can have an element of irreducible chance. (Did a random neural firing nudge the decision one way or another?)

    For the agent in an FSC to just happen to choose and successfully perform the action Black desires, without any nudging from Black, this multifaceted processual chance (including the crucial resultant luck of the outcome) already needs to break in a very specific, “lucky” way.

    Layer 2: The “Miraculous Alignment” – Composite Structural Chance

    But that’s not all. Even if the agent’s own process “luckily” heads in Black’s preferred direction and “luckily” results in the desired action, there’s another, more encompassing layer: Composite Structural Chance.

    This is the luck of the entire scenario unfolding with such perfect serendipity that the agent’s action (already a product of processual chance, including resultant luck) precisely matches Black’s wishes and does so in a way that doesn’t trigger Black’s intervention mechanism. Black has set certain conditions (a “sign”) for when to intervene. The agent must not only do what Black wants (a matter of resultant luck) but also avoid tripping any of these wires.

    This is the true “hole-in-one” moment. It’s not just that the golfer hit the ball well and it went in the intended direction (processual and resultant luck); it’s that the ball, after its flight, also navigates the unpredictable contours of the green and drops into the tiny cup, all without any further interference. The entire structure of the situation has to align in an incredibly fortunate way. This “structural chance” is composite because it relies on the favorable convergence of all the underlying processual chances – circumstantial, constitutive/PDC, causal, and critically, the resultant luck of the action’s outcome aligning perfectly with Black’s desire and non-intervention criteria.

    The Compounding Factor: It’s a Chain of Lucky Breaks

    The critical point is that these aren’t independent lucky events. They are compounded. For Black to remain idle, the agent’s “lucky” internal process, leading to a “lucky” outcome (resultant luck), must also “luckily” align with the very specific, and often narrow, path that avoids Black’s intervention. It’s like needing a series of coin flips to all land heads. The odds get very long, very quickly.

    The Clincher: Argument 1.1 and the Principle of Deliberative Contingency (PDC)

    My work highlights a crucial dilemma (let’s call it Argument 1.1) that exposes this deep reliance on compounded luck, and this is where the Principle of Deliberative Contingency (PDC) becomes particularly illuminating.

    The PDC states that an agent is subject to deliberative contingency when factors related to their constitutive luck (their character, ingrained beliefs, cognitive habits) or circumstantial luck (the immediate context, salient information, or lack thereof) lead them to perform an action (often a morally questionable one) without the relevant moral alternative even emerging as a significant, influential consideration during their deliberation. Essentially, they don’t seriously “see” or “feel” the pull of doing otherwise, thanks to a lucky (for the FSC designer) shaping of their deliberative landscape.

    Now, consider an agent in an indeterministic FSC. Are they (locally, at the moment of choice) determined or indetermined to perform the action (let’s call it ‘B’) that Black wants?

    1. If the agent is (locally) determined to do B:
    • They might seem to avoid some processual luck (like the randomness of an undetermined choice itself).
    • However, the success of the FSC scenario (Black staying idle) now rests almost entirely on composite structural chance. For Black not to intervene, the agent’s determined path must fortuitously result in action B and align with Black’s non-intervention conditions.
    • And here’s where the PDC bites: Why is the agent so determined to do B? Their “determination” or strong inclination could itself be a product of the PDC. Their constitutive and circumstantial luck might have shaped their deliberation (or lack thereof) such that the alternative to B (the one that would trigger Black) never seriously enters their mind or gains traction. Their “determined” path, leading to the “lucky” outcome of B, is thus “luckily” clear of any thoughts that would make Black step in. The FSC works, but only because the agent was “lucky” enough not to even properly consider the problematic alternative, and for their determined action to be the right one.
    1. If the agent is (locally) indetermined when choosing B:
    • Now, the agent is directly subject to processual chance, especially causal luck and the resultant luck of that undetermined process actually producing B. Their choice and successful execution of B is, to some extent, a lottery.
    • And, crucially, composite structural chance is still required in the background. This “lucky” undetermined choice, resulting in B, must still happen to avoid triggering Black.
    • In this case, their moral responsibility for B seems shaky, as it’s partly down to the luck of an undetermined process and its fortunate outcome.

    The takeaway from Argument 1.1 is stark: Whether the agent is portrayed as determined or undetermined to perform the action Black desires, the FSC is mired in significant, problematic luck. If determined, it’s the structural luck (often facilitated by the PDC ensuring a “smooth” path to the “lucky” outcome B). If undetermined, it’s processual luck (including resultant luck) compounded by the ever-present need for structural luck.

    The PDC is a key part of this “lucky sequence” because it explains how an agent might, without external coercion in the actual sequence, so perfectly align with Black’s wishes. Their deliberation itself is fortuitously constrained or directed by their background and circumstances, leading to the specific outcome Black desires.

    Conclusion: FSCs – A Philosophical Long Shot

    When we dissect Frankfurt-Style Cases through the lens of compounded chance, their power to definitively refute the Principle of Alternative Possibilities diminishes significantly. The seemingly robust demonstration of moral responsibility without alternatives begins to look more like a carefully orchestrated scenario that can only “succeed” if an improbable series of chance events – including the crucial luck of the action’s outcome – align perfectly.

    The alignment required – where the agent’s own (luck-infused) process leads them to do exactly what an intervener wants, resulting in the precise action desired, and in exactly the way that avoids triggering the intervention – is not a common occurrence. It’s a philosophical hole-in-one.

    While FSCs are invaluable for pushing us to think critically about responsibility, freedom, and alternatives, recognizing their profound dependence on compounded processual (including resultant), and structural luck (with the Principle of Deliberative Contingency often playing a silent, facilitating role) suggests they might not be the ace up the sleeve many believe them to be. The intuitive link between being able to do otherwise and being morally responsible is not so easily broken by a game that seems rigged by chance from the start.

  • The Dilemma for Non-Causal Accounts of LFW

    The Dilemma for Non-Causal Accounts of LFW

    In Episode 6: The Problem of Luck with Alfred Mele of The Free Will Show (2020), one of the host, Matt Flummer, asks (6m29):

    • “Some people complain that libertarianism requires that our actions be uncaused, so if they are uncaused (…) so what’s the problem, that people point out, with are actions being uncaused?”   

    Alfred Mele then responds (6m44) : 

    • “You know frankly, I can’t even make sense of the notion of an uncaused action, that is, I think uncaused actions are impossible. I think of actions as events caused in a certain way.”

    Mele is a renowned philosopher, expert in libertarian free-will (LFW), so if he doesn’t grasp the idea of an uncaused action, it seems fair to assume that a wide number of philosophers have the same confusion about an uncaused action. This becomes even more evident when we consider that many of them still find the notion of libertarian free will itself mysterious, which includes agent and event causation, so we can just imagine the lack of comprehension when it comes to the Non Causal (NC) account. But I think, by compelling the NC proponent to clarify their position regarding the causal nature of their actions, there is a clear path forward. 

    The argument can be structured as follows:

    Premise 1: Non-Causal Accounts of LFW Rely on Difference-Making (CDM A/O).

    Non-causal theories of LFW, while rejecting that free actions are determined by prior events or possess a traditional internal causal structure, must still explain how an agent exercises control in bringing about an action rather than another, or an action rather than an omission. This explanation, it is argued, implicitly or explicitly relies on the agent being a difference-maker for the action’s occurrence. The Causes as Difference-Makers principle “CDM : If C caused E, then, had C not occurred, the absence of C wouldn’t have caused E.” (Sartorio 2005)

    Indeed, the notion of agent control seems to presuppose at least a minimal level of difference-making. Even passive background conditions necessary for an event (such as the presence of oxygen for a fire [1]) can be seen as difference-makers in a broad sense – without them, the event would not occur that way. Non-causal theorists, who attribute to the agent a far more direct and active role in ‘performing’ or ‘settling’ an action than that of a mere background condition, must surely be committed to the agent making at least this kind of difference, if not a more robust one. To deny that the agent’s involvement constitutes difference-making would be to render their contribution to the action even more tenuous than that of a static environmental factor, thereby undermining the very notion of control they seek to establish.

    Non-causalists often describe the agent as directly “settling” what happens or “making it the case” that an action occurs. For instance, Carl Ginet speaks of an “actish phenomenal quality” where it seems to the agent as if she is directly bringing about the event (Clarke, Capes et Swenson, 2021, sect. 1.1). Hugh McCann describes free actions as an intentional, spontaneous and “creative undertaking on the agent’s part” (Clarke 2003, p. 20). This direct involvement implies that the agent’s performance of action A is what makes A occur; if the agent had not so performed, A (through that specific exercise of agency) would not have occurred.

    If control is exercised “in” or “by” acting, as some non-causalists like Palmer, and Ginet, suggest (2021, p.10050), then the agent’s very performance of the action is what makes the crucial difference. The action happens because the agent performs it. As Palmer describes : 

    Assuming that no-one else and nothing else has control over whether her action or decision occurs, the person can exercise control over whether her action or decision occurs simply by performing that action or by making that decision, where her performing that action or making that decision constitutes her exercise of control over whether that action or decision occurs.” (2021, p.10052)

    This aligns with the core of Carolina Sartorio’s CDM (A/O) [2] principle: 

    If an agent’s acting in a certain way caused E, then, had the agent failed to act that way, the agent’s failing to act that way wouldn’t have caused E. Conversely, if an agent’s failing to act in a certain way caused E, then, had the agent acted that way, the agent’s acting that way wouldn’t have caused E. (2005, p. 80)

    The agent’s specific action brings about the action’s occurrence (or its being settled), while the corresponding specific omission (failing to perform that very action) would not bring about that same action’s occurrence. This establishes the asymmetry central to difference-making.

    Without such a difference-making role, it’s unclear how the agent could be said to control the action. If what the agent does (or doesn’t do) makes no difference to whether the action occurs, the notion of control seems to evaporate.

    Premise 2: The CDM (A/O) Principle Describes a Causal Relationship of Dependence.

    The Causes as Differences-Makers (A/O) principle is articulated in terms of counterfactuals: what would (or would not) have happened if the agent had acted differently (e.g., omitted an action they performed, or performed an action they omitted). This reliance on counterfactuals links it directly to theories of causation based on dependence.

    Sartorio’s work aims to capture David Lewis’s insight that “We think of a cause as something that makes a difference, and the difference it makes must be a difference from what would have happened without it” (Lewis, 1973, as cited in Sartorio, 2005, p. 71). While she critiques Lewis’s specific theory, her CDM principle itself is built upon assessing the difference an event (or its absence) makes, which is a counterfactual notion.

    Ned Hall, in “Two Concepts of Causation,” (2004) explicitly identifies “dependence” as one of the two fundamental varieties of causation. He defines it as “counterfactual dependence between wholly distinct events” (p. 1). Critically, he states, “Dependence: Counterfactual dependence between wholly distinct events is sufficient for causation” (p.1). If CDM (A/O) embodies such counterfactual dependence, then, according to Hall’s framework, it describes a genuinely causal relation.

    Thus, if an agent’s control is grounded in their being a difference-maker in the sense captured by CDM (A/O), and this principle articulates a relationship of counterfactual dependence, then this aspect of control is, by these lights, causal.

    Conclusion: Non-Causal Accounts of LFW Are Therefore (Dependence-)Causal.

    If both Premise 1 and Premise 2 hold, then non-causal accounts of LFW, by relying on a difference-making principle like CDM (A/O) to ground agent control, inherently incorporate a causal relationship (specifically, causal dependence).

    This leads to the Dilemma for Non-Causal Theorists:

    • Horn 1 : If non-causal theorists accept that their account of control relies on a principle like CDM (A/O) and that this principle describes a form of causal dependence, then their theories are not “non-causal” simpliciter. This admission might not be a defeat but an opportunity for clarification. They could distinguish the type of causation they are employing (i.e., difference-making, counterfactual dependence, perhaps akin to an “enabling” or “structuring” cause) from the types of causation they reject (e.g., “productive” causation in Hall’s sense, or deterministic event-causation). This would mean their theories are “non-causal” in a productive way or “non-event- causally-determined” rather than entirely devoid of any causal relations. Such a move could provide a more robust and less mysterious grounding for control than appeals to purely intrinsic features of actions or subjective experiences alone.
    • Horn 2 : If non-causal theorists deny any reliance on a difference-making principle like CDM (A/O) for grounding control, they face the significant challenge of explaining how an agent can be in control of an action if their acting or not acting makes no difference to whether the action occurs or is settled. Given that even passive background conditions (like the presence of oxygen for a fire) can be understood as difference-makers in the broad sense that without them the effect would not occur, for an agent’s active role in “performing” or “settling” an action to constitute control, it must surely involve at least this minimal level of difference-making. To deny that the agent’s involvement makes such a difference would be to render their contribution to the action even more tenuous than that of a static environmental factor, thereby undermining the very notion of control. As Randolph Clarke (2003) notes, purely non-causal accounts (like those of Ginet or McCann, if interpreted as devoid of such difference-making) “are found not to offer satisfactory views of action and reason-explanation” (p. 1), and their accounts of the “exercise of active control” (p.3) can appear mysterious or insufficient. Without the agent as a difference-maker, the connection between the agent and the action may become too tenuous to support robust control and moral responsibility.

    In essence, our argument pushes non-causal theorists to clarify the nature of the agent’s contribution to action. If that contribution is understood as making a difference in a way captured by counterfactual dependence (CDM A/O), then it implies a causal relationship. If it is not understood as difference-making, the basis of control becomes more obscure than ever.

    [1] Here are two more examples of background conditions: 1) For a plant to grow, planting a seed is a direct action, but this only leads to growth if certain background conditions are met. These include the presence of water in the soil, a suitable temperature range, and available light. Without these enabling environmental factors, the seed will not sprout or thrive, regardless of being planted. 2) Similarly, for an electrical appliance to operate, flipping its switch to the “on” position enables the flow of electricity, the proximate cause. However, this action is futile without crucial background conditions such as a connected power source, an intact and closed electrical circuit, and operational internal components within the appliance itself.

    [2]  Causes‑as‑difference‑makers is a general principle. CDM (A/O)—its formulation for agents’ actions and omissions—has been developed most fully by Sartorio (2016). The underlying idea can be traced back to John Stuart Mill’s System of Logic (1843), David Lewis’s “Causation” (1973), and Ned Hall’s “Two Concepts of Causation” (2004), among others.

    References

    Clarke, R. (2003). Libertarian Accounts of Free Will. New York, NY, US : Oxford University Press USA. https://doi.org/10.1093/019515987X.001.0001

    Clarke, R., Capes, J., & Swenson, P. (Fall 2021 Edition). Incompatibilist (Nondeterministic) Theories of Free Will. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy. Retrieved from https://plato.stanford.edu/archives/fall2021/entries/incompatibilism-theories/

    Cyr, T., & Flummer, M. (Hosts). (2020, September 28). The Problem of Luck with Alfred Mele (No. 6) [Audio podcast episode]. In The Free Will Show. Buzzsprout. https://www.buzzsprout.com/1244627/episodes/5575288-episod-6-the-problem-of-luck-with-alfred-mele

    Hall, N. (2004). Two concepts of causation. In J. Collins, N. Hall, & L. Paul (Eds.), Causation and Counterfactuals (pp. 225-276). MIT Press. https://doi.org/10.7551/mitpress/1752.003.0010

    Palmer, D. (2021). Free will and control: a noncausal approach. Synthese, 198(10), 10043–10062. https://doi.org/10.1007/s11098-021-01701-5

    Sartorio, C. (2005). Causes As Difference-Makers. Philosophical Studies, 123(1-2), 71-96. https://doi.org/10.1007/s11098-004-5217-y