The aim of this paper is to argue that doxastic responsibility, i.e., responsibility for holding a certain doxastic attitude, is not based on direct doxastic control. There are two different kinds of direct doxastic control to be found in the literature, intentional doxastic control and evaluative doxastic control. Although many epistemologists agree that we do not have intentional doxastic control over our doxastic attitudes, it has been argued that we have evaluative doxastic control over the majority of our doxastic attitudes. This has led to the assumption that doxastic responsibility is based on evaluative doxastic control. In the first part of this paper I will introduce the notion of doxastic responsibility and the framework of doxastic guidance control as well as the approaches to direct and indirect doxastic control. I will then argue that doxastic responsibility is not based on direct doxastic control by showing that doxastic responsibility is neither based on intentional nor on evaluative doxastic control.
This is a preview of subscription content, log in to check access.
Buy single article
Instant access to the full article PDF.
Price includes VAT for USA
Subscribe to journal
Immediate online access to all issues from 2019. Subscription will auto renew annually.
This is the net price. Taxes to be calculated in checkout.
Although I assume that under some conditions doxastic responsibility assessment is of epistemic significance, I will not argue for it in this paper. This is why I leave open whether the different manifestations of doxastic responsibility assessment are epistemically significant or not.
The focus of guidance control is on the actual sequence of events in which an agent brought about a certain state of affairs \(\sigma \) rather than on the alternative possibilities available to the agent in the very same situation (cf. Fischer 2012, p. 186). That is why guidance control approaches and approaches to responsibility based on them do not fall prey to Frankfurt-type cases.
McHugh (2013) calls the doxastic analogue to practical guidance control epistemic guidance control (cf. McHugh 2013, p. 143). My notion of doxastic guidance control and his notion of epistemic guidance control have a different meaning. McHugh’s epistemic guidance control only refers to what I will call evaluative doxastic control. However, McHugh’s epistemic guidance control can be presented within the framework of doxastic guidance control.
Below, I will distinguish between different kinds of reasons-responsiveness when I introduce strong, weak and moderate kinds of evaluative doxastic control.
Throughout this paper the term “actual world” refers to a specific sequence of events in the actual world in which the considered process/mechanism operates and results in the \(\sigma \) in question. My usage of “actual world” is thus equivalent to Fischer and Ravizza’s usage of “actual sequence of events” (cf. Fischer and Ravizza 1998, p. 44).
Actually Fischer and Ravizza’s definition of practical reasons-responsiveness contains three conditions. According to them a process/mechanism of type M of S is practically reasons-responsive iff in all relevant counterfactual worlds in which M operates and in which there are sufficient reasons to bring about an alternative state of affairs \(\sigma '\), S recognizes these reasons (I.), chooses in accordance with these reasons (II.) and acts in accordance with the choice, i.e. brings about \(\sigma '\) (III.) (cf. Fischer and Ravizza 1998, p. 41). The reactivity condition comprises condition (II.) and condition (III.). It is plausible to assume that there is always at least one relevant counterfactual world in which the agent has sufficient reasons to bring about an alternative state of affairs.
This kind of control is similar to what Hieronymi has called “manipulative” or “managerial” control (cf. Hieronymi 2006, p. 53).
In what follows I use the notion of a belief-forming process broadly, such that all processes with which one can revise one’s belief-system fall under it. Thus, belief-forming processes encompass the processes with which one forms beliefs, the processes with which one sustains beliefs, the processes with which one rejects beliefs as well as processes that result in suspensions of judgment. Moreover, I use the notion of a belief-forming process and the notion of a cognitive process interchangeably.
Note, in this paper I will not specify what belief-influencing actions/omissions are. Of course any approach to doxastic responsibility based on indirect doxastic control has to have a specified notion of belief-influencing action and omission. Since I will not argue for an approach to doxastic responsibility based on indirect doxastic control in this paper, I will leave this difficult task for future research. However, to get a clearer understanding of what a notion of doxastic responsibility based on indirect doxastic control looks like, see the work of Peels (2012), Meylan (2013) and Nottelmann (2007).
One might think that in analogy to the characterization of practical reasons one can characterize epistemic reasons as facts or considerations that bear on the question of what one ought to believe. Thanks to Heinrich Wansing for pointing this out to me. However, the question of what one ought to believe can be posed from different perspectives, including the prudential or the moral perspective. Unless one specifies the perspective from which one considers the question of what one ought to believe, we cannot characterize epistemic reasons as considerations of facts that bear on that question.
Intentional doxastic control is sometimes referred to as voluntary doxastic control in the literature.
Alston refers to direct doxastic control as “basic voluntary control” (cf. Alston 1988, p. 263 ff.). He takes what he calls “non-basic immediate doxastic control” to be another kind of intentional doxastic control, but an indirect kind. This kind of control is indirect because the intention to form a certain doxastic attitude causes agent to perform a certain action (or omission) which ensures that the agent gets the evidential basis which provokes the intended doxastic response. We can refer to this kind of control as indirect intentional doxastic control. I take it that for an agent to have direct intentional doxastic control, the agent’s belief-forming process itself has to be directly responsive to practical reasons or an intention to form a certain belief. This is not satisfied when it comes to the exercise of “non-basic immediate voluntary control”. In this paper I am only concerned with direct kinds of doxastic control and so I will not discuss indirect intentional doxastic control. Note, my use of “intentional doxastic control” refers to the direct kind of intentional doxastic control throughout this paper.
Such an argument against epistemic responsibility is discussed by Levy (2007).
The name “evaluative doxastic control” is inspired by Hieronymi (2006, p. 53).
Golman takes the properties of reliability and conditional reliability to be properties of process types (cf. Goldman 1999, p. 312). Moreover, the property of being a conditional reliable cognitive process depends on the property of being a belief-dependent process, in the way that all conditional reliable cognitive processes are belief dependent cognitive processes. Also the property of being an (unconditional) reliable cognitive process depends on the property of being a belief-independent cognitive process, such that all (unconditional) reliable cognitive processes are belief-independent cognitive processes. Since reliability and conditional reliability are properties of process types, it follows that belief-dependency and belief-independency are properties of process types as well.
Note, the process/mechanism level as well as the agent level refer to the respective level of description.
I chose strong reasons-responsiveness as an example. The considerations of this chapter apply to all approaches to reasons-responsiveness that come with a receptivity condition and a reactivity condition, independent of whether they are strong, moderate or weak approaches.
This refers to the receptivity condition necessary for strong reasons-responsiveness.
I choose strong epistemic reasons-responsiveness as an example. The consideration of this section apply to all approaches to epistemic reasons-responsiveness which come with a receptivity condition and a refined reactivity condition independent of whether they are strong, weak or moderate approaches.
The notion of sufficient epistemic reason can have a strong and a weak normative reading. According to the strong normative reading, an agent has sufficient epistemic reasons to bring about a doxastic attitude D toward p iff the epistemic reasons justify S to have D toward p. According to the weak normative reading, an agent has sufficient epistemic reasons to bring about a doxastic attitude D toward p iff the agent takes her epistemic reasons to justify her to have D toward p.
I suppose that to “to hold or to form an alternative doxastic attitude toward the considered proposition” means the same as “to believe otherwise”. Thus, doxastic responsibility that requires epistemic reasons-responsiveness requires the ability to believe otherwise of some sort.
Note that these worlds are not relevant counterfactual worlds because Albert does not have reasons to form an alternative doxastic attitude toward the proposition that his hands are full of dangerous germs, but instead he has sufficient reasons to disbelieve that his hands are full of dangerous germs, which is the same disbelief as in the actual world.
The refinement made in this section applies to all reasons-responsiveness approaches that come with a receptivity and a reactivity condition independent of whether they are strong, moderate or weak approaches and independent of the domain of reasons to which they apply (i.e., practical, epistemic, prudential etc.).
These reasons judgments are in crucial respects similar to what Steup has called doxastic decisions, see Steup (2000).
One might object that to satisfy the refined reactivity condition it suffices that the process operates on the reasons and on the tacit or the dispositional belief that the agent has sufficient reasons to believe otherwise. Thanks to an anonymous referee. This might indeed be sufficient for a cognitive process to satisfy the reactivity condition. However, since belief-independent cognitive processes cannot operate on doxastic states, they cannot operate on tacit or dispositional beliefs. Thus, belief-independent cognitive processes cannot satisfy the refined reactivity condition.
One might object to that argument that only a few (if any) doxastic attitudes are the result of belief-independent cognitive processes, for even most of our perceptual beliefs depend on certain background beliefs and are thus doxastic outcomes of belief-dependent cognitive processes. Thus, the assumption that the majority of our doxastic attitudes is formed/sustained by belief-independent cognitive processes would not be true. And so we might be responsible for the majority of our doxastic attitudes after all. Thanks to Shane Ryan for pointing this out to me. However, I doubt that most of our perceptual beliefs are formed by processes that operate on background beliefs and thus I doubt that most of our perceptual beliefs are outcomes of belief-dependent cognitive processes. There is a distinction that has to be kept in mind between the empirical claims that belief-independent cognitive processes exist and that the majority of our doxastic attitudes are the doxastic outcomes of belief-independent cognitive processes, and the normative claim that a belief can have positive epistemic status (e.g., epistemic justification) just by being the doxastic outcome of a belief-independent cognitive process.
Of course epistemic agents sometimes refer to background beliefs when asked to give reasons for why they take doxastic attitudes (including perceptual doxastic attitudes) to be justified. These background beliefs might indeed be significant for the epistemic status of the belief in question. Epistemologists such as Susan Haack (cf. 1993, Chaps. 2, 4) have argued that belief-independent processes cannot confer a justificatory status to their doxastic outcomes independent of certain background beliefs. Haack argue for the claim that the epistemic status (i.e., its justificatory status) of a belief depends on certain background beliefs even if the belief is the doxastic outcome of a (reliable) belief-independent cognitive process. As far as I can see, Hack’s argument for what she calls foundherentism (cf. Haack 1993, Chap. 4) and the arguments which she presents against foundationalism, (cf. Haack 1993, Chap. 2) especially her arguments against process reliabilism (cf. Haack 1993, Chap. 7), do not raise any doubts about the existence of belief-independent cognitive processes. These arguments only raise problems for a certain normative assumption often made in some foundationalist theories like process reliabilism, namely that reliable belief-independent cognitive processes can confer prima façie epistemic justification to their doxastic outcomes.
My argument does not depend on any assumptions about the normative function which belief-independent cognitive processes have in a theory of epistemic justification. However, my argument for the claim that we cannot be responsible for the majority of our doxastic attitudes depends on two empirical claims. The first empirical claim is that belief-independent cognitive processes exist. The second empirical claim is that the majority of our doxastic attitudes are formed/sustained by belief-independent cognitive processes. These empirical claims have to be distinguished from the normative claim that some belief-independent cognitive processes can justify (some of) their doxastic outcomes independent of background beliefs. For, even if the epistemic status of a perceptual belief depends on certain backgrounds beliefs, from this assumption it does not follow that the perceptual process, which has caused the belief in question, has operated on these background beliefs.
Of course both empirical claims are claims against which one can argue, and I admit that I do not have any reason to support these claims except my intuition. However, I take the assumption that belief-independent cognitive processes exist to be less controversial than the assumption that belief-independent cognitive processes are able to justify their doxastic outcomes independent of any background beliefs. Note that the existence of belief-independent cognitive processes does not conflict with the claim that perception is cognitively penetrable Stokes (cf. 2013). The empirical assumption that the majority of our doxastic attitudes are formed/sustained via belief-independent cognitive processes is perhaps more controversial than the assumption that such processes exist. This is why I am going to present another argument for the claim that doxastic responsibility is not based on evaluative doxastic control, which does only rely on the empirical claim that belief-independent cognitive processes exist.
To be a proper subject to responsibility assessment allows for both negative as well as positive evaluations. If cognitive processes such as guessing or wishful thinking would be excluded from the realm of epistemically reasons-responsive processes, then agents cannot be proper subjects to responsibility assessment with respect to the doxastic outcomes of these processes. Thus, epistemic reasons-responsiveness should not require responsiveness to normative epistemic reasons, for otherwise an epistemic agent cannot be a proper subject to responsibility assessment with respect to the doxastic outcomes of processes like guessing or wishful thinking.
This kind of sensitivity bears some similarity to the kind of sensitivity employed by approaches to local reliability. “To be [locally] reliable, a cognitive mechanism must enable a person to discriminate and differentiate between incompatible states of affairs. It must operate in such a way that incompatible states of the world would generate different cognitive responses.” (Goldman 1976, p. 771, parenthesis A.K.). To put it differently, a cognitive process that results in a true belief that p in the actual world is locally reliable iff in all relevant counterfactual worlds in which the process operates on a different input, which is incompatible with the truth of p, the process does not result in the belief that p (cf. Goldman 1986, p. 778). However, local reliability requires that a cognitive process is sensitive to the truth of the considered proposition in some way, epistemic reasons-responsiveness* requires the cognitive process in question to be sensitive toward its input, such that the operation of the cognitive process on a different input results in a different doxastic attitude toward the considered proposition in the relevant counterfactual worlds.
My perceptual belief that there is a red wall in front of me is the doxastic outcome of a belief-independent cognitive process. Thus, this belief is not the doxastic outcome of an epistemic reasons-responsive process as introduced in the previous section.
The famous Barn Façade case has been published in Goldman (1976). Goldman attributes this case to Carl Ginet.
This example is discussed at length in the Gettier debate. I am not concerned with the analysis of knowledge in this paper. The example is used because Barney’s perceptual process among other things is not strongly epistemically reasons-responsive* when employed in a barn façade environment.
At least proponents of the assumption that we can be responsible for doxastic attitudes which are the doxastic outcomes of belief-independent cognitive processes, would agree that Barney is a proper subject to responsibility assessment with respect to his gettiered belief.
Thanks to Heinrich Wansing for pointing this out to me.
The original Albert case stems from Steup (2008, pp. 376, 380). In the original case the cognitive process underlying Albert’s psychological disorder is irresponsive to any epistemic reason.
Of course, one might claim that Albert does not own the process underlying his psychological disorder and explain that this is the reason why he is not responsible for his compulsive belief. Even if one could come up with some cases in which a doxastic attitude is formed by a psychological disorder that is not owned by the agent, I do not think that the cognitive processes which underly psychological disorders are in general such that they violate the ownership condition. At least, philosophers as Steup (2008) and McHugh (2013) assume that doxastic attitudes, which are the causal outcomes of psychological disorders, are violating the condition of epistemic reasons-responsiveness. Also Fischer and Ravizza assume that actions which are caused by psychological disorders are not practically reasons-responsive (cf. Fischer and Ravizza 1998, p. 48).
A world in which Albert has washed his hands for exactly 10 minutes can plausibly be taken to be a relevant counterfactual world. So there is at least one relevant counterfactual world in which Albert has a different input, his cognitive process which underlies his paranoia operates on this input and results in a different doxastic attitude toward the proposition that his hands are full of dangerous germs, namely a disbelief.
For the argument, see the discussion of the Barn Façade case at the beginning of this section.
Recall that the kind of doxastic control upon which doxastic responsibility is based has to be necessary and sufficient for doxastic responsibility.
Thanks to an anonymous referee, who asked me to clarify this.
I will argue below that this assumption is wrong.
For a critical discussion of such an environment-relative approach to cognitive ability, see Kallestrup and Pritchard (2014).
Note, this might be the point at which Breyer’s approach to doxastic control and the approach that I have introduced as moderate evaluative doxastic control* in this section deviate from each other.
Since I am concerned here with the connection between reliability and epistemic reasons-responsiveness in general, I omit the * in the following discussion.
In what follows I will use the term “reliability” to refer solely to global reliability.
Recall that the process type which operates in the actual world, the agent, as well as the proposition that the agent considers in the actual world, are held fixed across relevant counterfactual worlds for evaluating whether the process is epistemically reasons-responsive.
Local reliability is a situation-specific property because it is dependent on the specific situation in which the process actually operates (except for the input upon which the process is operating) and the proposition that the agent considers in the actual world.
The truth-ratio of a cognitive process refers to the ratio of correct direct doxastic outcomes of the cognitive process (cf. Goldman 1986, p. 103).
Compare this to the relation between global and local reliability. Global reliability and local reliability are two independent properties, because local reliability is situation-specific and global reliability is not.
Except for certain agent reliabilists such as Greco (2009), Barney’s perceptual belief-forming process is (globally) reliable but not epistemically reasons-responsive in the barn-façade environment. Moreover, let us assume that an agent guesses that \(2+2=4\). The cognitive process of guessing is globally unreliable, but in the context of a belief in a necessarily true proposition, the cognitive process of guessing turns out to be (trivially) locally reliable. Since epistemic reasons-responsiveness is a situation-specific property just like local reliability, it seems plausible to assume that epistemic reasons-responsiveness and global reliability are independent properties. Note, although I cannot argue for this here, local reliability and epistemic reasons-responsiveness are also two independent properties.
Note that the cognitive process which underlies Albert’s psychological disorder is not a reliable process with respect to a lot of different approaches to reliability. For example, the cognitive process that underlies Albert’s psychological disorder would neither count as a reliable cognitive process according to a normal world chauvinistic approach to reliability, nor would it count as a reliable cognitive process according to a two-stage approach to reliability.
Alston, W. P. (1988). The deontological conception of epistemic justification. Philosophical Perspectives, 2, 257–299.
Breyer, D., & Greco, J. (2008). Cognitive integration and the ownership of belief: Response to Bernecker. Philosophy and Phenomenological Research, 76(1), 173–184.
Breyer, D. S. (2013). Knowledge, credit, and cognitive agency. Pacific Philosophical Quarterly, 94(4), 503–528.
Buckareff, A. (2006). Doxastic decisions and controlling belief. Acta Analytica, 21, 102–114.
Burge, T. (2003). Perceptual entitlement. Philosophy and Phenomenological Research, 67(3), 503–548.
Feldman, R., & Conee, E. (1985). Evidentialism. Philosophical Studies, 48, 15–34.
Fischer, J. M. (2012). Deep control: Essays on free will and value. New York: Oxford University Press.
Fischer, J. M., & Ravizza, M. (1998). Responsibility and control: A theory of moral responsibility. New York: Cambridge University Press.
Goldman, A. I. (1976). Discrimination and perceptual knowledge. The Journal of Philosophy, 73(20), 771–791.
Goldman, A. I. (1986). Epistemology and cognition. Cambridge: Harvard University Press.
Goldman, A. I. (1993). Epistemic folkways and scientific epistemology. Philosophical Issues, 3, 271–285.
Goldman, A. I. (1999). Reliabilism: What is justified belief? In L. Pojman (Ed.), The theory of knowledge: Classical and contemporary readings (2nd ed., pp. 306–319). Belmont: Wadsworth Publishing.
Greco, J. (2009). Knowledge and success from ability. Philosophical Studies, 142(1), 17–26.
Haack, S. (1993). Evidence and inquiry. Towards a reconstruction in epistemology. Oxford: Blackwell Publishers.
Hieronymi, P. (2006). Controlling attitudes. Pacific Philosophical Quarterly, 87(1), 45–74.
Kallestrup, J., & Pritchard, D. (2014). Virtue epistemology and epistemic twin earth. European Journal of Philosophy, 22(3), 335–357.
Kornblith, H. (2015). The role of reasons in epistemology. Episteme, 1(1), 1–15.
Levy, N. (2007). Doxastic responsibility. Synthese, 155(1), 127–155.
McCormick, M. (2011). Taking control of belief. Philosophical Explorations, 14(2), 169–183.
McHugh, C. (2013). Epistemic responsibility and doxastic agency. Philosophical Issues, 23(1), 132–157.
Meylan, A. (2013). Foundations of an ethics of belief (Vol. 15). Berlin: Walter De Gruyter.
Nottelmann, N. (2007). Blameworthy belief: A study in epistemic deontologism. Dordrecht: Springer.
Peels, R. (2012). Believing responsibly. Intellectual obligations and doxastic excuses. Ph. D. thesis, Utrecht University.
Steup, M. (2000). Doxastic voluntarism and epistemic deontology. Acta Analytica, 15, 25–56.
Steup, M. (2008). Doxastic freedom. Synthese, 161, 375–392.
Stokes, D. (2013). Cognitive penetrability of perception. Philosophy Compass, 8(7), 646–663.
Wedgwood, R. (2002). The aim of belief. Philosophical Perspectives, 16, 267–297.
Special thanks to Amy Flowerree, Shane Ryan, Heinrich Wansing and to two anonymous referees for giving me extensive comments on earlier versions of this paper.
About this article
Cite this article
Kruse, A. Why doxastic responsibility is not based on direct doxastic control. Synthese 194, 2811–2842 (2017). https://doi.org/10.1007/s11229-015-0951-6
- Doxastic responsibility
- Doxastic agency
- Intentional doxastic control
- Evaluative doxastic control
- Epistemic reasons-responsiveness