Abstract
I argue that advocates of moderate epistemic idealization need some standards against which they can determine whether a particular individual P has a responsibility to acquire some specific piece of information α. Such a specification is necessary for the purpose of determining whether a reason R, the recognition of which depends on accounting for α, can legitimately be ascribed to P. To this end, I propose an initial sketch of a criterion that may be helpful in illuminating the conditions in which P might be said to have a responsibility to take α into account when searching for reasons to act or believe. I worry that reason ascriptions that are based on information acquisition expectations which are not captured by this criterion (and, hence, not legitimate reason ascriptions) may be open to charges of authoritarianism, especially when such a reason ascription prompts one person to demand that another act in a way that is responsive to the ascribed reason(s).
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
I use the term ‘act’ in its broadest sense. I take it to include non-physical acts such as changing what one believes or values.
This point is nicely expressed by Kevin Vallier in his discussion of what he calls the ‘standard conception of idealization’ (Vallier 2014, p. 145).
I acknowledge that an account of moderate idealization might significantly restrain what it means to idealize the information component. For example, one might argue that the information that one obtains by merely deliberating or drawing inferences from one’s existing information base counts as new information. Such information is just a by-product of further ‘internal’ deliberation. That is, such idealization of the information component is just a by-product of idealization of the rationality component. However, when I speak of moderate idealization throughout this paper, I have in mind only those accounts which permit the idealization of an agent’s knowledge base through supplementation of information which is ‘external’ to the agent—i.e., information which cannot be acquired by the agent through purely deliberative means, such as simply drawing inferences from her existing information base.
As I use the term here, to act in an ‘authoritarian’ manner is to demand that a person adhere to some principle, policy, or guiding belief despite the fact that the person in question does not and cannot rationally endorse whatever principle, policy, or guiding belief is in question.
Some deny that such a distinction exists. For example, Maria Alvarez, in Kinds of Reasons, asserts that ‘I take the claim that there is a reason for A to φ to be equivalent to the claim that A has a reason to φ’ (Alvarez 2010, p. 22). However, once the distinction between full and moderate idealization is considered, it becomes apparent that the two phrases are not equivalent.
We of course need an account of what it means to legitimately ascribe a reason to an individual. Such an account will be given in Gaussian terms in the ‘Expectations and Information Acquisition’ section.
As part of his argument against full rationality, Gaus argues that, contrary to popular belief, full rationality does not guarantee convergence between fully rational persons, nor does it guarantee determinacy regarding what any one individual has reason to do (Gaus 2011, pp. 239–243).
It is not uncommon for theorists working in value theory and epistemology to discourage over-idealization (or at least full idealization). Robert Pasnau, in a discussion of ideal theory and epistemology, has said that ‘[i]dealized epistemology… does not involve abstracting away from all human cognitive limitations, but instead focuses on what would count as perfection for beings such as us, in a world such as ours’ (Pasnau 2013, p. 1005). Russell Hardin has argued that moral theorists have failed to adequately appreciate how ‘fundamentally important’ the fact of human cognitive limitations is to the construction of moral theories (Hardin 1988, p. 1, 26). With respect to the notion of full idealization, Hardin does not mince words: ‘any argument that turns on perfect information, perfect calculations, and perfect theory is a house of cards’ (Hardin 1988, p. 17). Christopher Cherniak, in Minimal Rationality, argues that ‘the pervasively and tacitly assumed conception of rationality in philosophy is so idealized that it cannot apply in an interesting way to actual human beings’ (Cherniak 1986, p. 5).
For an extended critique of full idealization, see chapter 5 of Vallier (2014).
As Gaus frames it, ‘if people have crazy beliefs, they do not have reasons to act as the crazy beliefs indicate, though they may take themselves to have such reasons’ (Gaus 2011, p. 236).
While the false belief might help explain the action, it would not justify it.
Or what Gaus refers to as ‘a purely subjective understanding of one’s reasons…’ (Gaus 2011, p. 236, emphasis added).
Even if one does not share Williams’ intuition regarding the gin/petrol case, the need for some idealization is motivated by the idea that there are at least some situations in which agents are mistaken (sometimes deeply) about the reasons they think they have.
For more on the notion of ‘real rationality,’ see Pollock (2006, pp. 3–6). With regard to the notion of human cognitive limitations, Pollock notes that ‘[h]uman beings, and any real cognitive agents, are subject to cognitive resource constraints. They have limited reasoning power, in the form of limited computational capacity and limited computational speed. This makes it impossible, for example, for them to survey all of the logical consequences of their beliefs, or to compare infinitely many alternatives. This is a fundamental computational fact about real agents in the real world…’ (Pollock 2006, p. 3) For a more elaborate discussion of some of these limitations, see chapters two and three of Cherniak (1986) and Simon (1976).
Gaus takes these rules to be universal, though he does remark that they may be more easily retrieved via some languages rather than others (Gaus 2011, p. 244).
Pollock has argued that since reasoning is generally non-terminating (i.e., ‘[t]here will never be a point at which the agent has completed all the reasoning that could possibly be relevant to a decision’) the most we can expect from a ‘real agent’ is that ‘the agent perform a “respectable amount” of reasoning, and then base [his or her] choice on that’ (Pollock 2006, pp. 4–5).
So long as the norms of good reasoning can be determined (an issue I will not take up here).
In Justificatory Liberalism, Gaus advocates for a modest form of open justification in which ‘we treat [one’s system of beliefs and reasons] as open to new information and arguments and, from this external perspective, make judgments about what would then be justified in [that system of beliefs and reasons]… The core idea of open justification is that, at any given time, a justified belief system is, ideally, stable in the face of acute and sustained criticism by others and of new information’ (Gaus 1996, p. 31). Achieving stability in one’s belief system means coming to terms (probably slowly) with the reasons that there are. But P non-judgmentally informing Q of information α, which will help to stabilize Q’s system of beliefs, is a distinct matter from P demanding Q to act in a particular way—or judging Q to have acted wrongly—because P believes that Q ought to have taken α into account. It is the latter issue that I am most concerned with. That said, an agent’s system of beliefs and reasons may generate certain information acquisition responsibilities that would indeed be captured by Gaus’ account of open justification. But, as will hopefully become clear below, not all information acquisition responsibilities can be generated by one’s own system of beliefs and reasons. Some information acquisition responsibilities are foundational—they are meant to help form one’s system of beliefs and reasons. If this is right, the system itself cannot fully set one’s information acquisition responsibilities.
The idea here is that the reason itself was accessible in virtue of the necessary information being accessible.
With regard to the issue of deliberation, it is unclear to what extent it might be possible to point to particular reasoning processes in the same way that we might point to a particular piece of information. In some situations, we might be able to assert ‘If only P had completed inference I, then P would have recognized reason R.’ It is difficult, however, to know how widespread such assertions might be. Moreover, it is unclear exactly what sorts of inferences we might be able to expect others to make. As Cherniak argues, very little of any particular ordering of deductive inferences from ‘easy’ to ‘difficult’ ‘can be shown a priori to apply to all agents’ (Cherniak 1986, p. 31). Considerations like this suggest that the most imposing obstacle to eliminating the vagueness issue from the deliberation component is developing a set of standards (along the lines of those summarized at the end of the ‘On the Practice of Idealization’ section with respect to the information component) by which a third party might be able to non-arbitrarily assess moderately idealized P’s reasoning in a particular situation.
This is not to argue that non-arbitrary determinations about what counts as the right amount of deliberation are in principle impossible. It is only to recognize the challenges that confront such a view. Importantly, if it is possible to remove the vagueness that I take to be inherent in the deliberation component, this does not undermine my claim that the elements of deliberation and information acquisition need to be treated separately in developing an account of moderate idealization. After all, the standards by which we would judge whether a person has deliberated responsibly will diverge from the standards by which we would judge whether he or she has acquired information in a responsible manner.
One commentator on this paper has suggested that I have gone too far in considering a case in which P might be expected to obtain a civil engineering degree since P could simply hire a civil engineer in order to obtain all of the relevant facts. This comment illustrates the very issue at hand: why think it is reasonable to expect that P hire a civil engineer in order to retrieve all of the relevant facts? While I do not dispute that doing so would likely be more cost and time efficient for P, I rely on the hypothetical situation (as indicated by the use of a conditional), in which a degree would be necessary to obtain the relevant facts, in order to demonstrate how some relevant facts might be considered reasonably inaccessible.
This point assumes the reason under consideration cannot be arrived at through purely deliberative means alone.
We are not always aware of the deliberative situations which require our attention. For example, while Vance may in fact have dangerously high cholesterol, it may be the case he is not aware of that fact. Yet, a deliberative situation exists with regard to Vance’s cholesterol problem—he needs to decide on a management plan. The point is that troubling cases of reason ascriptions are cases where an external party recognizes a deliberative situation in which the agent in question does not.
When paired with the direct-confrontation requirement, the effort view may generate a further problem. Namely, how much effort is an agent expected to dedicate to ensuring that he or she is directly confronted with the ‘right’ question? I thank an anonymous referee for this point.
The effort view cannot be salvaged by arguing that one has put in enough effort only if one has discovered all of the relevant and reasonably accessible facts. Such an account just leaves us where we are now—trying to clarify the notions of relevancy and reasonable accessibility.
To my knowledge, not much has been said about what grounds the validity of expectations regarding the acquisition of knowledge of certain epistemic norms. I suspect, however, that such grounds may at least be found in some social contract accounts. Nevertheless, this is not something I take up in this paper.
I acknowledge that there is plenty of room for debate about which exact skills would be required for one to achieve societal literacy. As one reviewer has noted, familiarity with home buying standards might extend beyond what we would consider to be the most common skills of survival and risk aversion in our society (since, for example, not every person will be so situated (e.g., economically) that such information is needed). It is perhaps the case then that this set of skills will be even smaller than I have anticipated.
Smith does consider a case in which the doctor’s nurse fails to inform him that the journal had arrived (Smith 1983, p. 548). On account of this, the doctor’s failure to read the journal is presumably excused. This, however, still assumes that the doctor had a standing obligation to read the journal.
It is unclear just how common these sorts of cases might be. After all, with regard to some particular (hard) case c, while a community may not prescribe any standards for how to deliberate about c in particular, it may prescribe standards which advise a general set of cases of which c is a member. If so, expectations about the agent’s deliberation on c may be captured (at least partially) by considerations for community standards.
Requiring no more than a reasonable level of deliberation.
The generally-reliable means may be partially determined by the community standards.
References
Alvarez, Maria. 2010. Kinds of reasons. New York, NY: Oxford University Press.
Cherniak, Christopher. 1986. Minimal rationality. Cambridge, MA: MIT Press.
Gaus, Gerald. 1996. Justificatory liberalism. New York, NY: Oxford University Press.
Gaus, Gerald. 2011. The order of public reason. New York, NY: Cambridge University Press.
Hardin, Russell. 1988. Morality within the limits of reason. Chicago, IL: University of Chicago Press.
Pasnau, Robert. 2013. Epistemology idealized. Mind 122: 987–1021.
Pollock, John L. 2006. Thinking about acting. New York, NY: Oxford University Press.
Simon, Herbert A. 1976. From substantive to procedural rationality. In 25 Years of economic theory, ed. T.J. Kastelein, S.K. Kuipers, W.A. Nijenhuis, and G.R. Wagenaar, 65–86. New York, NY: Springer.
Smith, Holly. 1983. Culpable ignorance. The Philosophical Review 92: 543–571.
Vallier, Kevin. 2014. Liberal politics and public faith. New York, NY: Routledge.
Williams, Bernard. 1981. Moral Luck: Philosophical papers 1973–1980. New York, NY: Cambridge University Press.
Acknowledgments
I wish to thank Bruce Brower, Gerald Gaus, Eric Mack, David Shoemaker, and the two anonymous referees for Res Publica for their many helpful comments on earlier versions of this article.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Tyndal, J. Moderate Idealization and Information Acquisition Responsibilities. Res Publica 22, 445–462 (2016). https://doi.org/10.1007/s11158-015-9295-9
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11158-015-9295-9