Abstract
It is plausible to think that part of what it is to be an autonomous agent is to adequately respond to important changes in one’s circumstances. The agent who has set her own course in life, but is unable to recognize and respond appropriately when evidence arises indicating the need to reconsider and perhaps adjust her plan, lacks an important form of personal autonomy. However, this “evidence-responsiveness” aspect of autonomy has not yet been adequately analyzed. Most autonomy theorists ignore it altogether and the few who have addressed it have failed to give a satisfactory account. In this paper, I first examine an evidence-responsiveness condition proposed by Arneson. I argue there that while Arneson’s condition provides a valuable framework in which to examine evidence-responsiveness, there are several crucial issues that it either fails to address at all or else fails to adequately resolve. That condition is therefore in need of further elaboration and refinement. I then examine a recent article in this journal by Blöser, Schöpf, and Willaschek which develops an account of autonomy that I argue can usefully be understood as employing and elaborating upon the general framework offered by Arneson. I argue that while the elaboration Blöser and her co-authors provide Arneson’s condition is instructive, it is inadequate in several important ways which indicate the form a more satisfactory evidence-responsiveness condition will take. I go on to develop such a condition and conclude by highlighting the advantages to be gained by including that condition in a complete theory of autonomy.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
My use of these terms differs somewhat from Dworkin’s. He in fact identifies six criteria for a satisfactory theory of autonomy: logical consistency, empirical possibility, value conditions, ideological neutrality, normative relevance, and judgmental relevance (Dworkin 1988: 7–9). The “value conditions” criterion holds that a theory should be able to explain why autonomy has been seen as a desirable character trait, and “ideological neutrality” holds that different ideologies should be able to see autonomy as valuable. Both can be subsumed under “normative relevance.” If it is a mark against a theory that it violates Dworkin’s “empirical possibility” criterion by implying that it is “impossible or extremely unlikely that anybody ever has been, or could be, autonomous,” (Dworkin 1988: 7) this will be because it conflicts with our conceptual and empirical judgments about autonomy – judgments to the effect that there is a meaningful form of autonomy that is possible for at least some human agents. As such, for sake of simplicity, this criterion can be subsumed under “judgmental relevance,” which holds that a satisfactory theory of autonomy must “in general accord with particular judgments we make about autonomy” (Dworkin 1988: 9). Dworkin includes under “judgmental relevance” normative as well as conceptual and empirical judgments, but that removes the distinction between this criterion and his several normative criteria and I will thus understand this criterion to instead exclude normative judgments. That leaves us with logical consistency, which I take for granted any plausible theory will satisfy, normative relevance, and judgmental relevance. Dworkin makes no claim as to the relative importance of his criteria. My reasons for placing greater weight on (non-normative) judgmental relevance are explained below.
See Frankfurt 1988. Frankfurt does not himself address autonomy explicitly, but his account of freedom of the will has been fruitfully applied to that context, perhaps most notably by Dworkin, who defined autonomy as “a second-order capacity of persons to reflect critically upon their first-order preferences, desires … and the capacity to accept or attempt to change these in light of higher-order preferences and values” (Dworkin 1988: 20, building upon an account first laid out Dworkin 1970). In response to manipulation cases like that described below, Dworkin went on to add a historically externalist condition to his account. For an example of a purely internalist hierarchical theory of autonomy, see Ekstrom 1993. The hierarchical is certainly the most prominent form of internalism about autonomy, but not the only possible form. For example, Wolf offers an internalist theory in the related context of moral responsibility according to which “responsibility depends on the ability to act in accordance with the True and the Good … to do the right thing for the right reasons” (Wolf 1990: 79–81). Whether an agent is responsible thus depends entirely upon whether she presently possesses that ability. We could perhaps use this as a theory of autonomy as well. This approach to autonomy would be problematic, however. As Mele argues, if an agent intentionally develops evil values and an insensitivity to the Good, only to then be brainwashed with the very ability he worked to eradicate, it is implausible to regard the results of that brainwashing as autonomous, for they are in no way his own (Mele 2001: 164). The possibility of using Wolf’s theory as the basis for an evidence-responsiveness condition of autonomy, rather than as a complete theory thereof, is discussed in note 12.
This historical is thus the most common variety of externalism with respect to autonomy, but it is not the only such variety. As explained in note 8, Christman offers a counterfactually externalist theory of autonomy which, among other problems, also neglects evidence-responsiveness and is thus also in need of a condition like that I will propose.
See, for example, Mele 2001: 145–146.
As this assumes, autonomy is best understood as a matter of degree. It is unlikely that the un-brainwashed agent is fully autonomous; the historically externalist approach need only claim that she is substantially more autonomous than her brainwashed counterpart.
And as I explain in note 19, despite Blöser, Schöpf, and Willaschek’s claim to the contrary, such a condition is internalist in nature and thus compatible with both approaches.
Barbara’s Pluzu desires would also qualify as autonomous on Christman’s counterfactually, rather than historically, externalist theory. According to that theory, an agent is autonomous with respect to some characteristic C if she has the (internalist) capacities to critically reflect upon C and to effectively form intentions to act on the basis of C, and she would not be alienated from C were she to reflect upon C in light of an adequate description of the historical processes that gave rise to it (Christman 2009: 155). The history of the characteristic is therefore not directly relevant; what matters is how the agent would respond to a description of that history. Weimer has argued, rightly in my view, that because the agent’s response to that description is dictated by her present mental structure, Christman’s counterfactual account will have difficulty handling psychological twins cases of the sort that motivate the externalist approach (Weimer 2009b). Regardless, Christman’s theory also requires an evidence-responsiveness condition. Barbara is competent to act on her Pluzu desires and to critically reflect on them. She also satisfies Christman’s counterfactual condition, for she in fact regularly does reflect on those desires in light of their history and is not alienated from them. Owing to the fact that it does not require that she be able to reflect on her desires in light of the new and relevant evidence available to her, Christman’s counterfactual version of externalism joins Mele’s historical version in (problematically, as we shall see) deeming those desires autonomous.
This sort of disconnect between perceptions and corresponding beliefs is not at all uncommon, and not usually problematic. As Audi explains, “there is reason to doubt that perceiving must produce any belief at all. … there seems to be there seems to be a natural economy of nature – perhaps explainable on an evolutionary basis – that prevents our minds being cluttered with the innumerable beliefs we would have if we formed one for each fact we can see to be the case” (Audi 1998: 22). What is unusual and unfortunate about Barbara’s situation is that she fails to form beliefs about perceptions that normally would, and I want to suggest, from the perspective of autonomy should, draw her attention and ground beliefs of crucial importance to her.
The relevant disposition is thus the behavioral disposition to employ one’s (yet to be fully specified) abilities to recognize and adequately respond to evidence if and when one’s circumstances indicate the need to do so. An agent who possesses the relevant abilities, but lacks the disposition to employ them when necessary, is clearly not appropriately responsive to evidence. As we shall see, in developing an adequate evidence-responsiveness condition, much therefore turns on how the relevant abilities and circumstances are spelled out. Underlying that disposition is presumably some sort of monitoring mechanism which detects the relevant circumstances and then activates the relevant abilities. It is because such monitoring does not necessarily carve its mark on conscious experience (fortunately, we need not consciously search for unexpected changes in our circumstances in order to recognize them as such), but rather prompts reflective reexamination if and when the relevant sort of circumstances arise that the disposition in question is more usefully identified by its influence on the more conscious processes involved in such reexamination. As is discussed in more detail below, an agent therefore might be properly disposed to employ her relevant abilities without actually doing so, since the relevant sort of circumstances might not arise.
While I believe that judgments about the nature and requirements of autonomy itself provide the most important source of support for this claim, it is worth pointing out that it also would seem to yield several desirable normative implications. I mention just two such implications here; a more thorough discussion of this secondary criterion is provided in the concluding section. Compare Barbara’s case as I have described it to one in which she retains evidence-responsiveness with respect to her Pluzu habit. In the latter case, she likewise repeatedly reaffirms her initial decision to undertake the experiment, but not because she lacks the ability and disposition to recognize and reconsider in light of the important changes in her circumstances, as in the original. On the contrary, she recognizes that her husband has become a Pluzu user, that her children are being neglected, and that these indicate the need to reexamine the experiment. She critically and rationally engages in such reexamination in light of this new evidence and decides to continue it. As in the original case, if such rational reexamination issued instead in a judgment that she should stop using Pluzu, she would be able to implement that judgment. My claim is that there is an important form of autonomy, evidence-responsiveness, which Barbara lacks in the original case, but not in this adjusted version. If, as may be thought, there is a connection between the autonomy of an agent’s desires and her moral responsibility for actions motivated by those desires, this would imply that she is more responsible and more blameworthy for her continued drug use in the adjusted version of the case, in which she retains an important form of control over that part of her life which she lacks in the original. Similarly, to the extent that it is autonomy that protects against undue paternalistic interferences, this claim implies that interference with Barbara’s experiment for paternalistic reasons would be less objectionable in the original case than in the adjusted version where she has rationally reaffirmed her decision to undertake the experiment in light of all the relevant evidence. Both implications seem to me plausible. Exactly what Barbara’s lack of autonomy would imply for these normative issues and whether those implications are indeed satisfactory of course depends on our understandings of moral responsibility, paternalism, and their relations to autonomy, topics I cannot adequately address here. The above are thus meant merely as suggestions as to some possible and seemingly satisfactory normative implications of my claim that evidence-responsiveness represents an important aspect of autonomy.
And, I should add, a more promising foundation than would the most obvious alternatives. Given that the failings of agents like Barbara can be seen as failures to properly respond to the good reasons they have to make adjustments to their chosen courses, we might instead attempt to account for this aspect of autonomy by applying to this context a reasons-responsiveness theory of moral responsibility. Wolf’s is one such theory. As argued in note 2, Wolf’s account does not provide a plausible complete theory of autonomy due to its lack of an authenticity requirement. Nor does it provide the basis of a promising evidence-responsiveness condition of autonomy, for it addresses only the ability to respond to “the right reasons,” those flowing from the Good. The self-direction constitutive of autonomy, however, is not always directed at the Good. Just as an agent with morally good projects may or may not remain appropriately responsive to evidence bearing on those projects, so too might an agent whose projects are not in line with the Good. Wolf’s theory cannot make sense of this latter claim, however. Fischer and Ravizza’s reasons-responsiveness theory is somewhat more promising, but Arneson’s condition ultimately offers us a more useful starting point. According to their theory, responsibility for an action requires the ability to respond appropriately to reasons not to perform that action (Fischer and Ravizza 1998). (Although only this present-directed condition is relevant to evidence-responsiveness, it should be noted that their theory also requires that that ability be the agent’s own, which they spell out in a historically externalist manner.) While I believe that we perhaps could use Fischer and Ravizza’s reasons-responsiveness condition of responsibility as a basis from which to develop an adequate evidence-responsiveness condition of autonomy, several significant changes would need to be made to that theory. First, the reasons relevant on their account are sufficient reasons to do other than perform a specific action. This focus makes good sense in the context of an account of moral responsibility for a given action. Here, though, we are concerned with a more general ability to respond to reasons to reconsider and perhaps adjust one’s course in life. Second, their account requires responsiveness to too broad a range of reasons for it to serve as an evidence-responsiveness condition of autonomy. As argued below, it is not the capacity to respond to all the reasons that exist which is necessary for autonomy, but only a subset of those reasons – namely, those that are available and unanticipated. Their account includes no such qualifiers. Finally, the asymmetry of Fischer and Ravizza’s account, whereby a robust capacity for reasons-receptivity is required but only a very weak capacity for reasons-reactivity is necessary, is out of place in this context. It may be that, as Fischer and Ravizza propose, an agent who can regularly recognize the reasons there are not to perform an action, but who could translate recognition of such reasons into action in only one possible scenario is morally responsible for her action. In the context of an evidence-responsiveness condition of autonomy, however, this asymmetry is problematic. An agent who can recognize that there are sufficient reasons to reconsider and revise some aspect of her life, but is unable to execute the change indicated by those reasons, is not properly seen as autonomous with respect to that part of her life, but is instead “stuck with” and “victimized by” her pro-attitudes in that domain (Mele 2001: 138–39). To make these changes would be to alter their account rather considerably such that it more closely approximates the condition proposed by Arneson. Arneson’s condition has the advantages of A) focusing on reasons, or evidence, indicating the need to reconsider one’s course in life (or part thereof), rather than reasons not to perform some specific action, and B) restricting the relevant evidence to that which is “disturbing or unanticipated” and “becomes available.” As such, that condition provides a more useful starting point from which to examine evidence-responsiveness as a necessary condition for autonomy. As we shall see, however, Arneson’s account does not anticipate the third adjustment needed on Fischer and Ravizza’s theory, for it too has an insufficiently robust treatment of the reactivity portion of evidence-responsiveness.
For instance, the evidence provided by the Sirens’ song does not indicate that Odysseus needs to reconsider his plan to render himself incapable of steering towards the Sirens because the plan was formed in anticipation of that very evidence.
Weimer 2009a and Rocha 2011 also propose evidence-responsiveness conditions employing Arneson’s framework (explicitly in Weimer’s case, not in Rocha’s). I focus on Blöser, Schöpf, and Willaschek’s condition here because it offers the most complete and most promising set of responses to our unresolved issues. Weimer addresses only the third of the issues in need of resolution and offers the same response to that issue that it receives from Blöser, Schöpf, and Willaschek, a response that we shall see below to be too narrow. Rocha addresses the second and third issues, but offers rather implausible responses to both. On Rocha’s “welfare” theory, autonomy requires that when obstacles to meeting whatever welfare standards an agent has set for herself arise, she must see them as such and find a way to achieve happiness despite them (Rocha 2011: 323). The evidence to which agents must be capable of properly responding thus consists on Rocha’s account of evidence indicating that they are failing to meet their welfare standards, and an autonomy-maintaining response to such evidence is any happiness-securing response. Both responses are implausible. The first because it not only when an agent is failing to meet her welfare standards that autonomy might require that she be able to reconsider and perhaps alter her course in life. An agent might be meeting all such standards and nevertheless lack autonomy because she lacks the capacity to recognize and properly respond to the fact that unanticipated evidence available to her indicates the need to review those very standards. The second because responses to new evidence that secure happiness might nevertheless indicate a lack of autonomy. An agent who responds to evidence indicating a failure to meet her welfare standards might maintain her well-being (at least, on Rocha’s subjective understanding thereof) by simply ignoring or irrationally discounting that evidence. Rocha implausibly takes such a response to maintain autonomy as well as happiness.
Although she expected her husband to pick up the slack with respect to housework and childcare during the experiment, let us assume that a few unopened letters are not enough of a failure in either respect to disconfirm that expectation.
While I will not offer a complete account of availability here, it would seem that such an account would have to view information-acquisition abilities and dispositions in a general way, allowing for the possibility that an agent lacking the ability and/or disposition to acquire information through certain channels might nevertheless satisfy the “minimally adequate” condition due to her ability and disposition to acquire information through other channels. A blind person, for example, might lack one specific information-acquisition ability, but still possess a set of information-acquisition abilities that would qualify as minimally adequate. Further, given that the evidence-responsiveness condition I propose will require that agents be able and disposed to recognize available evidence indicating the need to review as such, and then respond properly to such evidence, we must not set the threshold for minimally adequate abilities and dispositions too high, lest we place our conception of autonomy out of the reach of most agents, which would render the condition both incompatible with our judgments about autonomy itself and of little normative relevance. Of course, we must not set that threshold so low as to render autonomous even those agents intuitively lacking this important aspect of autonomy. This is by no means a unique problem. The necessity of navigating between these extremes manifests itself in several ways in theories of autonomy. Any plausible theory will require that agents are and/or previously were at least at least minimally rational and self-reflective, their lower-order pro-attitudes and actions at least minimally under their control, etc.
The primary problem with Blöser, Schöpf, and Willaschek’s understanding of relevant evidence is that is in these respects too narrow. There is, though, one respect in which it may be too broad, namely its suggestion that confirmatory and disconfirmatory evidence both indicate the need to reconsider. It is not clear why evidence confirming an expectation implied by a value need trigger a review of that value. Such evidence is relevant to the value and should be recognized as such and available for recall should the agent later have reason to reconsider the value, but does not seem to itself provide such a reason.
As should be clear, my proposals are consistent with, but expand upon, those of Blöser, Schöpf, and Willaschek. They propose that the capacity to reconsider, and then reaffirm or abandon, one’s values in light of experiences disconfirming expectations implied by those values is a necessary condition for autonomy. I agree, but have argued that there is more to the evidence-responsiveness aspect of autonomy than this: other forms of responsiveness to other sorts of evidence are also necessary.
It is important to note that the evidence-responsiveness condition I have proposed is therefore internalist in nature. Whether an agent satisfies that condition at a given time depends entirely on her mental structure at that time – specifically, whether she possesses the relevant abilities and dispositions. The necessity of such a condition therefore has no impact on the internalism/externalist debate, which concerns whether an agent’s autonomy depends entirely on factors internal to her present mental structure. As was saw with Mele’s theory, externalist accounts typically include internalist conditions alongside at least one condition addressing some external factor. My claim is that an evidence-responsiveness condition like that I have outlined represents one more internalist condition such accounts must include. Blöser, Schöpf, and Willaschek disagree. They claim that their ERCR condition is externalist because although its possession at a given time depends entirely on the agent’s psychology at that time, an agent’s exercise of the relevant capacities requires factors external to her psychology – namely, external evidence (Blöser, Schöpf, and Willaschek, 246–47). The same is true of my condition. As there are two distinct senses of “external” here, whether we place the internalist or externalist label on such conditions depends upon which is the more relevant. In my view, it is the first. As argued below, autonomy is not properly seen as requiring the exercise, but merely the possession, of the (disposition and) capacity for evidence-responsiveness. Since it is with the possession of the capacity that we should be primarily concerned, the sense of internal addressing possession should play the primary role in determining whether a condition is labeled internalist or externalist. In that sense, both ERCR and the evidence-responsiveness condition I have proposed are internalist.
If nothing else, an explication of minimally adequate information-acquisition and belief-formation and –recognition abilities and dispositions is necessary. In addition, while I do not believe that an exhaustive list of the sorts of self-adjustments evidence might call for is possible, a more detailed examination than that offered here would surely be helpful.
References
Arneson R (1994) Autonomy and preference formation. In: Coleman J, Buchanan A (eds) In harm’s way: essays in honor of Joel Feinberg. Cambridge University Press, New York, pp 42–75
Audi R (1998) Epistemology. Routledge, New York
Blöser C, Schöpf A, Willaschek M (2010) Autonomy, experience, and reflection. On a neglected aspect of personal autonomy. Ethical Theory Moral Pract 13:239–253
Christman J (2009) The politics of persons: individual autonomy and socio-historical selves. Cambridge University Press, Cambridge
Dworkin G (1970) Acting freely. Nous 4:367–383
Dworkin G (1988) The theory and practice of autonomy. Cambridge University Press, Cambridge
Ekstrom LW (1993) A coherence theory of autonomy. Philos Phenomenol Res 53:599–616
Fischer JM, Ravizza M (1998) Responsibility and control: a theory of moral responsibility. Cambridge University Press, Cambridge
Frankfurt H (1988) Freedom of the will and the concept of a person. In: Frankfurt H (ed) The importance of what we care about. Cambridge University Press, Cambridge, pp 11–25
Mele A (2001) Autonomous agents. Oxford University Press, New York
Rocha J (2011) Autonomy within subservient careers. Ethical Theory Moral Pract 14:313–328
Weimer S (2009a) Beyond history: the ongoing aspects of autonomy. J Ethics Soc Philos 4:1–31
Weimer S (2009b) Externalist autonomy and availability of alternatives. Soc Theory Pract 35:169–200
Wolf S (1990) Freedom within reason. Oxford University Press, New York
Acknowledgments
I would like to thank Claudia Blöser, Eric Cave, Marcus Willaschek, and two anonymous referees for their comments on an earlier version of this paper. My thanks, also, to John Martin Fischer and Aron Schöpf for their helpful responses to questions I posed about their work.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Weimer, S. Evidence-Responsiveness and Autonomy. Ethic Theory Moral Prac 16, 621–642 (2013). https://doi.org/10.1007/s10677-012-9381-4
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10677-012-9381-4