Introduction

In this paper, I address Steven Bland’s (Epistemic Relativism) recent attempt to refute the kind of epistemic relativism that is motivated from the premise that epistemic justification is circular. Before getting into it, I will first sketch the argumentative landscape in which his attempt takes place.

Epistemic relativism -the view that the justification of beliefs and epistemic systems (ES)¹ is relative to contingent, non-neutral norms and principles- is often motivated by arguments from epistemic incommensurability. Roughly, an argument from epistemic incommensurability stipulates that, in the context of a dispute between two different ES, there is no neutral or rational way of determining which one of them is epistemically better without arguing circularly. Additionally, given that every attempt to justify an ES in the context of a dispute uses the very system at issue, relativism claims that all epistemic justification is relative and that there is no neutral or absolute way of preferring one ES over another.

Two ES are incommensurable when there is no neutral, systemindependent, epistemic standard to assess and justify one system over the other. Thus, the thesis of epistemic relativism that emerges from incommensurability arguments is supported by these three claims:

Pluralism: There are radically different ES.

Non-neutrality: There is no common, neutral, epistemic principle to justify one ES over another.

Circularity: Justifications in support of ES are circular.

In the literature, the standard example of incommensurability and relativism has been, since Rorty (Philosophy and the Mirror of Nature), the dispute between Bellarmine and Galileo regarding the position and movements of astronomical objects. As the dispute is normally described, Galileo and Bellarmine subscribed to different ES (the Copernican and Ptolemaic systems, respectively) with different methods to justify beliefs about the positions and movements of planets.

Purportedly, Bellarmine endorsed the epistemic principle or method that Paul Boghossian (Fear of Knowledge) calls “Revelation”:

Revelation: For certain propositions p, including propositions about the heavens, believing p is prima facie justified if p is the revealed word of God as claimed by the Bible (69).

Galileo, on the contrary, rejected Revelation (or at least its validity concerning the movements of planets) and held:

Observation: For any observational proposition p, if it visually seems to S that p and circumstantial condition D obtain, then S is prima facie justified in believing p (^{Boghossian, Fear of Knowledge} 64),

as well as epistemic methods such as

Deduction: If S is justified in believing p and p fairly obviously entails q, then S is justified in believing q (^{Boghossian, Fear of Knowledge} 66).

Induction: If S has often enough observed that an event of type A has been followed by an event of type B, then S is justified in believing that all events of type A will be followed by events of type B (^{Boghossian, Fear of Knowledge} 67).

Bellarmine’s and Galileo’s ES are incommensurable in the sense that there is no neutral procedure to compare and evaluate the beliefs and methods of one ES over the other. As it will be clear below, incommensurability states that the justificatory work within each ES is done by different and irreducible principles and methods. Bellarmine considers it unjustified to use Observation to form beliefs about the Earth’s movement, while Galileo takes as unjustified the relying on Scripture to decide what is true about celestial bodies.

Additionally, any attempt by either party to justify the validity of his own methods will either apply the deliverances of the methods at issue or make use of other parts of the ES whose justification is in doubt. Finally, as this means that justification in favor of either Bellarmine or Galileo will be circular, there is no non-relative way of preferring one ES over the other.

In the last few years, however, anti-relativists have tried to undermine relativism by arguing against incommensurability. A common antirelativist strategy has consisted in denying Pluralism.² If it is not true that there are radically different ES, and if every dispute or radical disagreement is merely apparent, then incommensurability does not arise in the first place and relativism cannot be true (at least not from the thesis of incommensurability).³ Related strategies have attempted to show that the Non-neutrality claim is false, since there are epistemic methods or principles which are common to every ES and could settle any radical dispute.⁴

Boghossian, for instance, attacks relativism by denying that Bellarmine andGalileo’s situation is one of incommensurability. According to Boghossian, both parties share the same ES rooted in basic methods such as Observation, Induction and Deduction. Their dispute, then, can be solved once one of the parties comes to apply and interpret those basic methods in the right way. But the problem with these strategies to undermine relativism is that they use circular justification.⁵ Roughly, a justification for an ES is circular if it makes use of the system or its deliverances in proving its superiority.⁶ Not every circular justification of an ES system is vicious, though. However, circular justification is not appropriate when an ES is called into question: if Bellarmine doubts that Galileo’s methods are right, he will not be convinced by a justification that appeals to the deliverances of Galileo’s methods.

To grant to the relativist that Circularity is unavoidable is to grant too much. As long as Circularity is true, the relativist might offer slightly revised relativistic theses. One relativistic strategy consists in arguing that incommensurability is not refuted by pointing out to epistemic methods shared by both Bellarmine’s and Galileo’s ES. For the relativist, incommensurability is not about whether Bellarmine uses Observation and inductive reasoning when reading the Bible. As it was mentioned, incommensurability states that the justificatory work within each ES is done by different and irreducible principles and methods. As it will be clear in the next section, if Bellarmine relies on his eyes when reading the Bible, this does not mean that he justifies what the Bible claims by using his eyes. In short: Galileo and Bellarmine both rely on Observation, but only the former uses it as a source of justification. Therefore, the relativist can say that the incommensurability remains because, (a) the work of Observation in both ES is different, and (b) any attempt to convince Bellarmine to use Observation instead of Revelation to justify beliefs about the planets will make use of the reliability of Observation -that is, it will take for granted that Observation is justified.

A second relativistic strategy consists in arguing that if Circularity is true, relativism is true. Pluralism might be false, but if any attempt to justify an ES should take its reliability for granted, then all epistemic justification is relative. This version of relativism is weaker in that the relativist cannot claim that there is no reason to prefer one ES over another. Since we might have no evidence of there being radically different ES, the claim that epistemic justification is relative leaves open whether there will be objective criteria to prefer one ES over another. In this way, the relativist’s thesis that comes from Circularity amounts to a challenge: insofar as a non-circular justification for an ES has not been provided, relativism remains a live option.⁷

As far as it goes, it seems that attempting to refute relativism by merely denying Pluralism or Non-neutrality is not totally successful, since from the inevitability of circular arguments the relativist might still claim that relativism has not been refuted. In the context of this debate, Bland (“Circularity”, Epistemic Relativism) advances an interpretation of ES and a dialectical argument to meet the relativist’s challenge. In particular, Bland defends that there is a non- circular argument that makes naturalistic (such as Galileo’s) ES preferable over non-naturalistic (such as Bellarmine’s) ones. Before assessing Bland’s proposal, I first reconstruct the steps of his argument.

Bland’s dialectical argument against Circularity

Bland starts by noticing that ES are structured by methods with two kinds of dependence relations: justification and application (or presupposition). On the one hand, Method A depends on method B in case A’s justification depends on B. This is the kind of dependence between the use of instruments (such as telescopes) and perception: to justify the reliability of telescopes I appeal to the reliability of Observation. On the other hand, method A depends on method B in case A’s application depends on B, that is, when A’s reliability needs B’s reliability. This kind of dependence occurs when the use of A is presupposed by method B. An example is the dependence of Induction on Observation: to use inductive reasoning I must apply Observation and presuppose its reliability. Both kinds of dependence are, respectively, strong dependence and weak dependence: if A needs B to be justified, A strongly depends on B; whereas if A needs B to be used, A weakly depends on B (cf. ^{Bland, “Circularity”} 158).

Given these two kinds of dependence, Bland argues that any given method within an ES can be basic in two ways. A method can be strongly basic, case in which that method is justified without strongly depending on any other method. Induction, Deduction, and Revelation are examples of strongly basic methods: they do not require other methods to justify their reliability. Secondly, a method can be weakly basic, such that its application does not weakly depend on any other method. Observation is a weakly basic method, since to form a perceptual belief no other method needs to be presupposed.

Which methods are strongly or weakly basic may vary according to different ES. This allows to make sense of incommensurability. When it comes to the subject of their dispute, the incommensurability between Galileo and Bellarmine can be expressed in terms of which methods they consider as strongly basic, that is, as methods whose justification does not depend on any other method. In the case of Galileo, Observation, Induction, and Deduction are strongly basic methods. Bland calls these naturalistic methods: any ES that consider them as strongly basic is a naturalistic ES. Also, these methods are in a strong dependence relation to other methods that Galileo uses, such as telescope observation.

Unlike Galileo, Bellarmine does not hold the naturalist’s strongly basic methods as strongly basic for his ES. Although Bellarmine can rely on observation and induction in daily, non-astronomical matters, he considers the method of Revelation as strongly basic: the truth revealed in Scripture does not depend on the truth or the reliability of any other methods. For Bland, Bellarmine’s is a case of a non-naturalistic ES. These differences between Galileo and Bellarmine suggest that two ES are incommensurable due to the kind of dependence relations between their respective basic and non-basic methods. To repeat: if Bellarmine relies on his eyes when reading the Bible, this does not mean that he justifies what the Bible claims by using his eyes; Galileo and Bellarmine both rely on Observation, but only the former uses it as a source of justification. So, the dispute between Galileo and Bellarmine is incommensurable because they have two different ES that take different kinds of methods as strongly basic.

Importantly, although Bellarmine’s Revelation does not strongly depend on any other method, it does weakly depend on naturalistic basic methods. That is, to read, comprehend and interpret what Scripture reveals, Bellarmine must presuppose that he can reliably read, engage in inductive and deductive reasoning, etc. Additionally, the opposite is not true of Galileo: his use of naturalistic methods does not weakly depend on Revelation. The difference between both ES is that Bellarmine’s should rely on what Galileo considers as strongly basic, while Galileo is not required -in order to use his naturalistic methods- to appeal to what Bellarmine considers as strongly basic.

In light of what has been said, Bland defends two claims related to the two kinds of epistemic dependence. First: “The reliability of a framework’s strongly basic methods is a sufficient condition for the justification of the non-basic methods that they support” (“Circularity” 159). This means that non-basic methods cannot be less reliable than strongly basic methods, since non-basic methods are reliable insofar as the basic methods they strongly depend on are reliable. Secondly, Bland offers the weak dependence principle (WDP): “Epistemic methods are no more truth-conducive than the methods on which they weakly depend” (Epistemic Relativism 190). WDP states that weak dependence does not transmit reliability: the reliability of a non-basic method is not greater than the reliability of the methods that must be presupposed to apply that method.

WDP needs some qualification, though, since it is not generally true. Inductive reasoning weakly depends on memory, but sometimes the former can be more reliable than the latter (for example, in cases of false memories or memory distortions). This is a reason to restrict the application of WDP. According to Bland, WDP does not hold only in the context of naturalistic ES, that is, when basic epistemic methods are capable of being corrected when they are unreliable, or when the weak dependence on a basic method is minimal in the sense that it is not the only basic method that is presupposed. Conducting experiments in a laboratory is a case in which WDP does not hold: even though causal reasoning weakly depends on memory or observation (and so the latter’s deliverances can be distorted by biases that might affect their reliability), causal reasoning in the context of an experiment is more reliable, since scientists possess mechanisms to correct or diminish the potential unreliability of these basic methods (cf. ^{Bland, Epistemic Relativism} 191). Now, given that Bellarmine’s strongly basic method of Revelation does not possess these mechanisms, WDP applies to it: Revelation cannot be more reliable than the basic methods it weakly depends on.

From this, Bland has the tools to offer a non-circular argument against relativism. Remember that the relativist says that the dispute between Galileo and Bellarmine is a case of Pluralism, Non-neutrality and Circularity. Bland denies Circularity: it is possible to justify an ES without using it. In particular, he intends to argue, without using the deliverances and methods of naturalistic ES, that these are preferable to non-naturalistic ES. This is the argument:

Bland considers that the argument is dialectical because the premises are accepted for both naturalistic and non-naturalistic ES and make explicit the structure (justification and presupposition dependences) of every ES. The argument is non-circular because the naturalistic ES is not used in the reasoning from the premises to the conclusion. In the case of Bellarmine and Galileo’s dispute, the argument proves that Galileo’s ES is non-circularly preferable than Bellarmine’s. Bellarmine must accept the argument’s premises and conclusion, without first being convinced of the reliability of Galileo’s ES. However, in the next section I argue that the argument, at least from Bellarmine’s ES, is not compelling.

Circularity strikes again

I think that Bland’s argument does not entirely avoid Circularity. Even though Bellarmine himself can recognize that he must rely on and presuppose Observation, Induction, and Deduction, this still does not give him a non-circular reason to disregard Revelation as a strongly basic and reliable source of information.

Additionally, even if it were true that Bellarmine can see that he has a reason to abandon Revelation as a strongly basic method, this is not a non-circular reason to accept Galileo’s ES. My evaluation of Bland’s dialectical argument suggests that denying Circularity should explain why making explicit the dependence commitments of one’s ES will lead to suspect of, or reinforce, its reliability.

First, consider the hypothetical case in which Bellarmine is confronted with Bland’s dialectical argument. He grants premises a-d above: he accepts that the procedures and techniques of natural philosophy are reliable, though he takes Scripture as basic in the sense that its reliability comes from the presumption that it was inspired by God. Additionally, he accepts premise e: he needs to use his eyes to read the Bible and engages in inductive and deductive reasoning when interpreting the book. But when it comes to premise f, what reason does Bellarmine have to accept that this dependence prevents Revelation from being more reliable than these other methods? Bellarmine takes for granted the reliability of the Bible, but the dialectical argument is telling him that the reliability of using that book cannot be greater than the reliability of Observation, Induction, and Deduction. Does Bellarmine have a reason to accept f?

Bellarmine might say that the fact that he needs his eyes to read the Bible does not affect the reliability of forming beliefs according to Revelation. He might grant that the reliability of his eyes affects the reliability of his understanding the Bible. For example, if he reads the Bible while drunk, it is his reliability, and not Revelation’s reliability, which gets affected. The reliability of Revelation, for Bellarmine, does not depend on whether he needs his eyes to read it, but on interpreting it in the right way. Thus, Bellarmine might argue that the methods of his ES system are not fairly depicted in the dialectical argument. If Bellarmine can resist the dialectical argument in this way, then a further non-circular premise is needed to convince him that the method of Revelation, even when carried out in the right way, cannot be more reliable than Observation, Induction, and Deduction. In other words, either premise f is circularly true, or it lacks a further non-circular justification.

Bland grants premise f by means of WDP. The validity of this principle must compel Bellarmine to accept that Revelation cannot be more reliable than Observation, Induction, and Deduction. One reason is that if Bellarmine’s understanding of the Bible presupposes the use of his eyes, then if his eyes are unreliable, what he gets from the Bible will not correct the unreliability of his eyes. Now, Bellarmine can argue that Revelation, carried out in the right way, can correct the unreliability of Observation. Additionally, he might say that the fact that Revelation does not correct Observation in the same way that scientific experimentation corrects Observation is not a reason to consider Revelation as not being a corrective method at all -maybe the potential unreliability of our senses is corrected, in the long term, by a constant and consistent use of Revelation. Thus, if Bellarmine is not irrational in holding this position, Bland’s justification of premise f is either circular or it lacks a further non-circular justification.

Second, suppose Bellarmine accepts that he must abandon Revelation. According to the thesis of relativism under discussion, there is no noncircular reason or criterion to prefer one ES over another. Does Bland’s dialectic argument offer Bellarmine a criterion to embrace Galileo’s ES instead of, say, suspending judgement? According to Bland, Galileo’s non-basic methods (such as telescopic observation) strongly depend on Observation, Induction and Deduction. And, unlike Bellarmine’s, according to premise c, Galileo’s non- basic methods cannot be less reliable than the basic methods they strongly depend on. In other words, Galileo’s ES is preferable because it is at least as reliable as Observation, Induction and Deduction. This, in addition to premise H, must force Bellarmine to adopt Galileo’s ES. However, premise h does not establish that Galileo’s ES is more reliable than Bellarmine’s.

The argument has established that, regarding basic methods (Observation, Induction, and Deduction), Galileo’s ES is at least as reliable as them, and Bellarmine’s cannot be more reliable than them. But this still does not say that an ES that strongly depends on those methods is necessarily more reliable, all things considered. Even if Bellarmine accepts premise H, he has no reason to adopt Galileo’s ES, since the reliability of the latter has not been established. What has been established is that an ES is reliable by its dependence on strongly basic methods such as Observation, Induction, and Deduction. It is true that Galileo takes these methods as strongly basic. But this, though, is compatible with denying that Galileo’s ES is more reliable. Put differently: the dialectical argument shows that Galileo’s ES is neither worse nor better than Observation, Induction, and Deduction. Bellarmine can accept that his own ES might not be as reliable as other ES that takes as strongly basic Observation, Induction, and Deduction. However, he can doubt that Galileo’s ES is more reliable, even if it is naturalistic -what makes it naturalistic is that it takes certain methods as strongly basic, not that it is reliable. To go from, “it takes as strongly basic naturalistic methods”, to, “it is more reliable”, it has to be proved that naturalistic basic methods are more reliable, all things considered. So, again, either Bellarmine has to adopt Galileo’s ES on the basis of a circular reason relative to Galileo’s ES or an additional non-circular justification as to why Bellarmine must adopt Galileo’s ES is needed.

Conclusion

Some purport to refute relativism from incommensurability by denying that there exist radically different ES. Resisting this refutation, some friends of relativism have argued that as long as Circularity is true, some form of relativism is also true. Bland’s refutation of relativism by means of a dialectical argument is intended to meet the latest relativistic challenge. However, in order to refute relativism by denying Circularity, it must be explained why recognizing the dependence commitments of one’s ES will lead one to see its reliability; that is, it has to be shown in a non-circular way that dependence relations between epistemic methods carry objective notions of reliability.