Pragmatic vs. Dialectical Strategies for Resisting Epistemic Relativism: A Reply to Richard Fumerton, Steven Bland

Author Information: Steven Bland, Huron University College,

Bland, Steven.“Pragmatic vs. Dialectical Strategies for Resisting Epistemic Relativism: A Reply to Richard Fumerton.” Social Epistemology Review and Reply Collective 5, no. 5 (2016): 37-40.

The PDF of the article gives specific page numbers. Shortlink:

Please refer to:


Image credit: Kevin Dooley, via flickr

Richard Fumerton has raised several important questions regarding my dialectical strategy for resisting epistemic relativism.[1] My hope is that by answering these questions, I can both clarify and bolster my position.

On Epistemic Circularity

My target in “Circularity, Scepticism and Epistemic Relativism” is the problem of epistemic circularity, according to which an argument for the truth-conduciveness of any epistemic framework must itself be an application of the framework at issue.[2] This problem presents both a sceptical and a relativist threat. The sceptic uses this fact to support her conclusion that no epistemic framework is trustworthy, while the relativist uses it to support her conclusion that no framework can be any more trustworthy than another. My overarching aim is to show that these are distinct threats that must be answered by means of distinct arguments. The argument I present is meant to disarm the relativist threat while leaving the sceptical threat unanswered (though not unanswerable). 

An epistemic framework consists in a set of basic principles and methods. Fumerton correctly notes that I distinguish two senses in which a method can be basic by distinguishing two senses in which one method may depend on another.[3] Method A strongly depends on method B when A’s trustworthiness is an outcome of B, i.e., when A’s justification depends on B. So, for example, my practice of consulting a newspaper’s weather forecast strongly depends on inductive reasoning because I defend my practice by appealing to the newspaper’s successful track record of predicting the weather.

Method A weakly depends on method B when A cannot yield outcomes without an application of B, i.e., when A’s use depends on B.[4] I cannot consult the newspaper’s forecast without seeing it in the paper, so this practice weakly depends on visual perception. Now we can characterize a strongly basic method as one that does not depend on any other method for its justification, and a weakly basic method as one that does not depend on any other method for its use.

Epistemic frameworks may diverge in their strongly basic methods, but they have the same core of weakly basic methods. Scriptural revelation is strongly basic for the Biblical literalist, but not for the naturalist, yet perception, memory, and deductive reasoning are weakly basic for them both. My central claim is that non-naturalistic methods, such as scriptural revelation, weakly depend on strongly basic naturalistic methods, but naturalistic methods do not weakly depend on non-naturalistic methods. To consult scripture, one must appeal to the deliverances of perception, memory, and logical reasoning, but one need not appeal to scripture in order to perceive, remember, or reason.

I attempt to use this asymmetrical relation of weak dependence between naturalistic and non-naturalistic methods in an argument against epistemic relativism and for a rational preference for naturalistic frameworks.[5] It is at this point that Fumerton expresses some puzzlement. In the remainder of this paper, I will present his reconstruction of my argument and his resultant concerns; I will then diagnose what I take to be his misunderstandings of my position; once these have been corrected, I believe that Fumerton’s concerns will be dispelled.

Forming Beliefs

Fumerton reconstructs my argument as follows:

The argument, as I understand it, is that even if one were to embrace nonnaturalist strongly basic ways of forming beliefs, they would be useless without our naturalistic ways of forming belief. They would be useless because they would be weakly basic, and would have nothing to process absent the deliverances of “natural” ways of forming beliefs. So crystal ball gazers, those who rely on scripture, those who read the entrails of birds to predict the outcomes of battles, all need to rely on memory, perception, valid deductive reasoning, and perhaps induction, if they are going to generate conclusions. And those conclusions will have no more claim on truth than the data they processed—garbage in; garbage out. We need to rely on naturalistic methods to get anywhere. By contrast, we don’t need to rely on nonnaturalistic methods to have hope of arriving at truth. So we have a kind of pragmatic argument for preferring the [naturalistic] methods of forming beliefs.[6]

According to this pragmatic argument, naturalistic methods are preferable because they provide non-naturalistic methods with their fundamental data, data about the contents of scripture, the images in crystal balls, the configurations of entrails, and the like. Without these data, non-naturalistic methods have nothing to process, and consequently, remain idle. On the other hand, naturalistic methods do not depend on any other methods for their fundamental data, which makes them self-sufficient in a way that other methods are not. There is, then, a pragmatic impetus to use naturalistic methods rather than their non-naturalistic counterparts: the latter get us nowhere without the former, but the former can get us to the truth without having to rely on the latter.

Having presented the pragmatic argument, Fumerton interjects: “… I’m not sure how we are dividing methods of forming belief into those that are naturalist and those that are not.”[7] Given the way he has understood my position, I take him to be unsure about which methods I regard as providing the fundamental data required for the exercise of our epistemic practices. Thus, he asks whether or not phenomenal conservativists, epistemic conservativists, and radical empiricists endorse naturalistic frameworks.[8] Are the fundamental data restricted to the objects of introspection, perception, and memory? Can we have fundamental data about physical objects? Without answers to these sorts of questions, there is no way of determining which epistemic frameworks are naturalistic; this renders the pragmatic argument for such frameworks ambiguous at best and useless at worst.

Fumerton’s worry rests on a misunderstanding of my notion of weak dependence. He says that some methods weakly depend on others insofar as they “… depend on other information gained from other sources for their successful use in arriving at truth.”[9] This leads him to the conclusion that weakly basic methods provide fundamental information required by other methods. However, this is not an apt characterization of the constitutive function of such methods. Though weak dependence is sometimes a matter of one method providing inputs that are essential to another—such as the case of perception and scriptural revelation—this is not always the case, and therefore, is not what weak dependence amounts to.

Consider probabilistic reasoning, which Fumerton suspects is weakly basic.[10] It is not weakly basic—though it may well be strongly basic—because it makes essential use of arithmetical reasoning. Yet, arithmetical reasoning does not generate the data about which we reason probabilistically. Rather, arithmetic is part of the formal apparatus that is necessarily presupposed in the formulation and application of probability theories. Furthermore, deductive reasoning is weakly basic, yet non-ampliative; our use of epistemic methods relies on such reasoning, not for their fundamental data, but for the truth-preserving inferences that facilitate the transition from evidence to beliefs. I could not use the newspaper’s forecast to make a prediction about tomorrow’s weather without reasoning logically, but logic alone provides me with no information whatsoever.

My claim about the distinctive status of naturalistic methods, then, is less specific than the one Fumerton attributes to me. I am saying that all epistemic practices depend on naturalistic methods for their use, not that all epistemic practices depend on naturalistic methods for their fundamental data. Making this claim, I think, requires only a pre-theoretical notion of naturalistic methods that includes those methods that operate in empirical and scientific inquiries: perception, memory, deductive, inductive, and abductive reasoning, and the like. For this reason, I do not see myself as faced with the unenviable task of moderating theoretical debates about the nature of fundamental knowledge. As far as I can see, my position is compatible with all of the philosophical positions that Fumerton mentions.

Dialectical and Pragmatic Strategies

With this clarification out of the way, I would like to distinguish my dialectical strategy from the pragmatic one that Fumerton attributes to me. The key claim of the pragmatic argument is that non-naturalistic methods would be useless without strongly basic naturalistic methods. But, of course, a commitment to non-naturalistic methods does not preclude the use of naturalistic methods. Biblical literalists do not deny that perception, memory, and logical reasoning confer warrant on beliefs; they diverge from naturalists in thinking that scriptural revelation is a more reliable source of belief than naturalistic methods. Thus, scriptural revelation need not be weakly basic for the Biblical literalist. Indeed, it cannot be weakly basic, and Biblical literalists are perfectly aware of this fact; stressing this point is what makes my strategy dialectical. What they fail to realize is that scriptural revelation cannot be more reliable than perception, memory, and logical reasoning because it weakly depends on these methods.

The problem with non-naturalistic frameworks, then, is not that they are useless for the purpose of tracking the truth, but that whatever success they have is parasitic on the naturalistic methods on which they weakly depend. This being the case, no non-naturalistic framework can be more truth-conducive than a naturalistic framework whose strongly basic methods it must presuppose. I argue that this constitutes a non-circular rationale for preferring naturalistic frameworks, but that it does not establish, or seek to establish, that naturalistic methods are truth-conducive. As such, the dialectical strategy yields a vindication of naturalistic frameworks that undermines the principal argument for epistemic relativism, rather than a justification that addresses any sceptical argument.


Bland, Steven. “Circularity, Scepticism and Epistemic Relativism.” Social Epistemology 30,no. 2 (2016): 150-162.

Bland, Steven. “Incommensurability, Relativism, and the Epistemic Authority of Science.” Episteme 11, no. 4 (2014): 463-473.

Fumerton, Richard. “Fundamental vs. Derivative; Useful vs. Useless: A Response to Bland.” Social Epistemology Review and Reply Collective 5, no. 4 (2016): 56-59.

[1] Fumerton 2016.

[2] Bland 2016.

[3] Fumerton talks exclusively about epistemic principles, rather than methods. While I do discuss epistemic principles in (Bland 2016), my view can be more clearly expressed by focussing only on methods. I don’t think anything important turns on this point.

[4] In (Bland 2016), I say that method A weakly depends on method B when applications of A depend on B. This characterization may well have contributed to what I regard below as Fumerton’s misunderstanding of weak dependence. For this reason, I have adopted this new characterization as a clearer expression of what I mean by weak dependence.

[5] I also make this argument in (Bland 2014).

[6] Fumerton 2016, 57.

[7] Fumerton 2016, 57.

[8] Fumerton 2016, 57-58.

[9] Fumerton 2016, 57.

[10] Fumerton 2016, 58.

Categories: Critical Replies

Tags: , , , , ,

Leave a Reply