Three Accounts of Doxastic Responsibility

Introduction

“Sorry I can’t be there, hope you understand. I love you all and inshallah [God willing] will meet you the best of places Jannah [heaven]” [1]. That was the message Mohammad Sidique Khan left to his pregnant wife before traveling to the UK and setting off a bomb on 7 July 2005 that killed himself and seven others. His reference to Jannah suggests that he believed this act was justified, glorious, and even triumphant. Holding him accountable, then, seems to involve showing that he is responsible for the political, religious, and moral beliefs that motivated this act.

However, those who defend this attribution of responsibility must answer to the following argument (henceforth “the Basic Argument”):

(Premise 1) We are only responsible for what we can voluntarily control.
(Premise 2) We cannot voluntarily control our beliefs.
(Conclusion) We are not responsible for our beliefs [2].

Using the Basic Argument as a starting point, this paper is organized around three accounts of doxastic responsibility. Accounts 1 and 2 represent standard responses to the above argument, as they reject Premise 1 by disengaging responsibility from voluntary control. However, they ultimately fail in light of a hypothetical case involving Martians given below. In their place, I put forth Account 3, which also rejects Premise 1 but appeals to the notion of voluntary control indirectly, mediated by the influence relation. Account 3 is then modified in light of a vicious regress that afflicts it. I conclude that the resultant account adequately answers the Basic Argument. It additionally furnishes us with a tidy argument that we are responsible for our beliefs because we have voluntary control over belief-influencing factors such as our evidence base and intellectual virtues and vices.

Account 1: Reactive Attitudes

This view contends that to be morally responsible for a thing is to be the proper object of negative “reactive attitudes” on account of it [3]. Reactive attitudes are feelings and attitudes we may have for responsible agents, including positive ones like love and respect and negative ones like resentment and indignation [4]. The argument begins by noting that the primary objects of reactive attitudes are “moral personalities,” or the quality of one’s will. In particular, bad moral personalities are the proper objects of negative reactive attitudes. For instance, we commonly hold those cantankerous or jealous blameworthy, while those generous and compassionate praiseworthy.

If we adopt this account, it becomes clear that we are responsible for our beliefs even if we have no voluntary control over them, as long as we are proper objects of reactive attitudes. Moreover, that for which we are fundamentally responsible—our moral personality—must be involuntary as it embodies our take on the world. For instance, it seems just as impossible for us to condone terrorism voluntarily as for a terrorist to condemn it voluntarily. But that does not mean we’re not responsible for these views since they are precisely what constitute the apt objects of reactive attitudes. In other words, our beliefs reveal and partially constitute the kind of person we aspire to be, rendering us open to certain reactions that amount to responsibility. Throughout this account, there is no room for any notion of control, yet there is responsibility. Hence, Premise 1 is false.

Account 2: Attitudinal Control

The second view admits that responsibility presupposes control but holds us responsible for our beliefs because we have attitudinal control over them. This view begins by conceding that the sort of control we have over action involves a “governance by the will,” which is lacking in the case of belief. However, it contends that responding to reasons constitutes an alternative type of control, namely attitudinal control or “governance with the will” [5]. Attitudinal control occurs when we properly reason with available evidence and thereby form doxastic attitudes.

This account also introduces a hierarchy of control, in which beliefs and intentions are subject to attitudinal control via reasoning, while intentions or will, in turn, control our acts. As it turns out, more faculties belong to this higher category of attitudinal control, such as evaluation, judgment, and emotion. This hierarchical structure also reveals why “voluntary control alone is shallow” [6]. Only when acts are informed by intentions and beliefs formed via rational and comprehensive reasoning (which exerts attitudinal control in the process) are we responsible for them. In this sense, attitudinal control may be, in many ways, more fundamental than voluntary control. And because attitudinal control needn’t be voluntary, Premise 1 is false.

Against Accounts 1 and 2: The Martian Case

While Accounts 1 and 2 may initially seem convincing, they face a powerful yet simple objection that can be illustrated by considering a hypothetical group of Martians [7]. These Martians form beliefs just like us by reasoning about and weighing pieces of evidence. However, they differ from us in a crucial way: they cannot go about collecting new evidence and simply output beliefs deterministically. This process resembles how a function takes in a predetermined set of arguments and rigidly returns values. It is evident that these Martians are not responsible for their beliefs. Yet these Martians are reason-responsive and have attitudinal control insofar as they rationally reason about their available evidence. Thus, both Account 1 and Account 2 fail to vindicate the correct intuition in the Martian Case. Something more is needed to ground responsibility.

Account 3: Derivative Responsibility

Account 3 begins by acknowledging voluntary control’s central role in determining responsibility. However, it recovers our intuition about doxastic responsibility by appealing to the idea of influence. In this account, there are two types of responsibility: non-derivative responsibility, grounded in voluntary control, and derivative responsibility, grounded in control by influence. To elucidate this account, I first define influence . Then, I elaborate on its formal qualities and implications for responsibility. Finally, I apply the regress problem for intentions to beliefs and, on account of this regress, conclude that we sometimes have direct voluntary control over beliefs.

A. Definition

I define influence as follows: a person P has influence over some consequence C if and only if P has voluntary control over some action A, and A leads to C via a non-deviant causal chain, wherein non-deviant should be interpreted as requiring the agent to understand that the causal chain would lead to C before initiating it. For instance, while Bill Gates has no voluntary control (i.e., control at will) over the value of certain stocks, he has influence over them because his buying them would boost their values via a non-deviant causal chain. Similarly, while we have no voluntary control over our beliefs, we may still influence them by certain acts, such as seeking evidence or opening an inquiry.

B. Implications

Influence grounds derivative responsibility, the type of moral responsibility that stems from indirect control by influence. In conditional form, if P has influence over C, then P is derivatively responsible for C. Derivative responsibility is a type of responsibility, so we may also infer that P is responsible for C. We should accept this conditional because it is intelligible, simple, and consistent with our intuitions. It also explains why we hold people accountable for consequences over which they do not have voluntary control (e.g., Gates for meddling with stocks). Specifically, it delivers the correct verdict in the Martian Case. While the Martians have proper reactive attitudes and attitudinal control, they do not have influence over their beliefs since none of the voluntary acts available to them leads to a change in their beliefs via a non-deviant causal chain.

However, humans differ from Martians precisely in our limited capacity to act in ways that foreseeably lead to changes in beliefs. This capacity manifests in two ways. First, we can enlarge or improve our evidence base by browsing online, consulting experts, or conducting research. We can voluntarily undertake or refrain from undertaking these epistemic projects, which enables us to influence beliefs. Second, given the same knowledge base, we can sometimes voluntarily practice intellectual virtues (such as curiosity, humility, and open-mindedness) or vices (such as prejudice, arrogance, and conformity). These also led to a change in beliefs, for better or worse, via a foreseeable chain of causes and effects.

The modifier “sometimes” is significant as this capability is limited and differs by person and situation. For instance, a student at Harvard University has a greater capability to voluntarily expand their knowledge base and practice intellectual virtues than Mohammad Sidique Khan, who, in this respect, seems much like a Martian. Account 3 rules that Khan is responsible for his beliefs to the extent to which he may influence them. While those who want to hold Khan fully responsible for his beliefs may find this conclusion repugnant, it is not counterintuitive. We usually hold individuals in abnormal states (e.g., under extreme distress) or with limited physical or intellectual resources less responsible for their beliefs. These cases are, to some extent, analogous to Khan’s situation. By parity of reason, only partial doxastic responsibility should be attributed to Khan.

C. Beyond Beliefs

The notion that doxastic responsibility can be grounded by appealing to belief-influencing factors is not new. For instance, in his rebuttal against doxastic compatibilism, a position embraced by Accounts 1 and 2, Peels (2014) suggests the centrality of influence for doxastic compatibilism [8]. However, his account is limited to beliefs, and he sought a different grounding of our responsibility for intentions. This is unnecessary. A full-blown account of influence justifies the correct responsibility assignment to all noetic states.

Peels’s chief reason for limiting Account 3 to beliefs is the dissimilarity between beliefs and other noetic states, such as intention [9]. For instance, he pointed out that when the reasons for and against an intention balance out, we are free to choose between intend and refrain. However, when reasons for and against a belief balance out, we are not free to choose but must suspend judgment. Hence, Peels argues that we sometimes have direct voluntary control over intentions but can only influence beliefs. To suggest otherwise would lead to an infinite regress, for if intentions are only influenced, each intention would assume voluntary control over a previous intention-influencing action, which requires another influenced intention to initiate, and so on.

Against Peels, rather than showing the limitations of Account 3, the problem of infinite regress implies the existence of directly controlled beliefs. His argument from intentional and doxastic equilibriums fails because it rests on a false analogy. While Peels is correct in stating that one does not deliberate over beliefs when reasons are balanced, he fails to consider situations where the reasons are balanced between disbelief and suspension of judgment or between suspension of judgment and belief. In such cases, we are faced with a genuine choice. Therefore, our doxastic responsibility is not based solely on influence and may be partially grounded in voluntary control. This recovers the analogy between beliefs and intentions, showing influence as a ground for responsibility that is nonexclusive and continuous across domains of noetic structure.

Conclusion

In this essay, I examined three accounts of doxastic responsibility. Account 1 attempts to dislodge control from responsibility altogether by appealing to the notion of reactive attitudes; Account 2 proposes attitudinal control as an alternative route to responsibility; and Account 3 appeals to the idea of influence, which grounds derivative responsibility. While all three accounts answer the Basic Argument with initial plausibility, only Account 3 coheres with our intuition in the Martian Case. Thus, I am partially and derivatively responsible for a belief to the extent to which I have voluntary control over the relevant belief-influencing factors. And, though overlooked in the literature, I am also fully and non-derivatively responsible for beliefs formed via direct voluntary control.

Notes

[1] BBC News.
[2] Modified from Alston 257–260.
[3] Hieronymi 361.
[4] Fischer and Ravizza 5.
[5] McHugh 2748.
[6] McHugh 2761.
[7] See Peels 693–694.
[8] Also see Clarke 1986; Dretske 2000; Kornblith 1983; Stocker 1982.
[9] Peels 696–698.

Works Cited

Alston, William P. “The Deontological Conception of Epistemic Justification.” Philosophical Perspectives, vol. 2, 1988, pp. 257–299.

BBC News. “Profile: Mohammad Sidique Khan.” March 2, 2011, https://www.bbc.com/news/uk-12621381. Accessed 20 May 2024.

Clarke, Murray. “Doxastic Voluntarism and Forced Belief.” Philosophical Studies, vol. 50, no. 1, 1986, pp. 39–51.

Dretske, Fred. “Entitlement: Epistemic Rights without Epistemic Duties?” Philosophy and Phenomenological Research, vol. 60, no. 3, 2000, pp. 591–606.

Fischer, John Martin, Mark Ravizza. Responsibility and Control: A Theory of Moral Responsibility. Cambridge UP, 1998.

Hieronymi, Pamela. “Responsibility for Believing.” Synthese, vol. 161, no. 3, 2008, pp. 357–373.

Kornblith, Hilary. “The Psychological Turn.” Australasian Journal of Philosophy, vol. 60, no. 3, 1982, pp. 238–253.

McHugh, Conor. “Attitudinal Control.” Synthese, vol. 194, no. 8, 2017, pp. 2745–2762.

Peels, Rik. “Against Doxastic Compatibilism.” Philosophy and Phenomenological Research, vol. 89, no. 3, 2014, pp. 679–702.

Stocker, Michael. “Responsibility Especially for Beliefs.” Mind, vol. 91, no. 363, 1982, pp. 398–417.