Can We Defend Normative Error Theory?

Introduction

An error is a mistake. According to normative error theory, we make a systematic mistake when making normative judgments such as “murder is wrong” because these judgments ascribe normative properties such as the property of being wrong, and such properties are never instantiated.1 Normative error theory (hereafter, NET) is a global error theory about all normative properties, not just the moral kind.2 Its proponents (hereafter, error theorists) deny the instantiation of both moral and epistemic normative properties, but they are split on whether reasons for belief carry normative content and consequently whether we can rationally believe NET.

Many philosophers have tried to undermine this theory by arguing that its defenders, in believing it, argue from a self-defeating position.3 From my view, these objections all rely on talk about our ability to believe NET. I think this allows error theorists to escape self-defeat by adopting the cognitive attitude of non-belief toward the theory they defend. However, I want to argue that we could never even successfully defend NET, and so it won’t matter whether we can believe it. I qualify “defend” with “successfully” to leave open various ways we might attempt to defend what we could never successfully defend and still call that “defense”. You might think, for example, that a poorly constructed theory defense, even if doomed to fail, still fulfils the action description “defending a theory”. My aim is to eliminate the possibility of ever finding success in defending NET. Showing NET indefensible by any plausible metric of success would be a significant and surprising result in its own right. However, it could be that some theories we cannot successfully defend are nonetheless true. I close, therefore, by offering some initial reasons for thinking that our inability to defend NET is very bad for normative error theorists since it gives us good reason to think NET is false. A full defense of these consequences, however, I leave for future work. My principal aim in this paper is to show that we could never successfully defend NET.

Before advancing my argument, we should first consider what we require to successfully defend a theory. In lieu of a complete theory of theory defense and conditions for its success, I will propose a working definition here and a necessary condition for any successful theory defense later. It seems to me that what we mean when we say that someone has defended a theory T (implying minimal success) is that, they have (at least) provided an epistemic reason, relevant to the question of T’s being true or false, which counts as a consideration against rejecting T.4 Let’s stipulate, then, that to successfully defend a theory minimally requires offering a reason which counts in favor of believing that T is true and works against believing that it’s false. This definition means to exclude arbitrary, merely pragmatic, preferential, or crazy “reasons” for belief. One of my opponents aiming to successfully defend NET also excludes such “reasons”. On this, more later.

In offering my working definition, I don’t arbitrarily assign normative status to reasons for belief which count as reasons relevant to successfully defending a theory (hereafter, theory defense reasons); that is, I leave open whether theory defense reasons weigh normatively on belief. I think they do; but I arrive at that conclusion only on consideration of consequences following its denial. I will argue that without epistemic norms, one’s theory defending position is self-defeating, that is, it provides opponents no theory defense reasons which is the aim of a successful theory defense.

To illustrate in a general way what I have in mind, consider Socrates defending some theory T. He first considers various arguments for and against T. He then offers reasons which constitute considerations in favor of T. Finally, he considers objections to his argument and devises replies which undermine these objections. In all this, I understand Socrates to be successfully defending T, where “defending” minimally involves providing reasons favoring the truth of T. As such, it seems to me that to successfully defend T, we should be able to perform at least one of the following actions:

Therefore, in this paper I will understand ability to perform at least one of these actions as constituting a necessary condition for successful theory defense. I assume that all error theorists, whether the believing or unbelieving type, have attempted to perform at least some of these actions while defending NET. However, they must successfully perform at least some of these actions to successfully defend this theory.5

Surprisingly, I think that we cannot successfully perform these actions relative to the successful defense of NET. I think this because I think that theory defense requires that theory defense reasons weigh normatively on belief. Again, I do not assume in advance that theory defense reasons weigh normatively on belief and therefore theory defense is by definition a normativity-discharging enterprise. However, I do think that on reflection theory defense reasons turn out to weigh normatively on belief. And in this paper, I aim to show that NET strips its advocates (but not the rest of us) of access to normativity even if they do not believe this theory. If I’m right, we cannot successfully defend NET. And if we cannot successfully defend it, I think this a serious problem for it.6

This paper consists of five sections. In section I, I consider a recent objection from unavoidable self-defeat levied against error theorists. I do this to introduce error theorists to an escape route from self-defeat objections but also because my project will build and improve on this argumentative strategy. In section II, I analyze an argument for error theory’s unbelievability to show how error theorists might use it to escape self-defeat but also to showcase the trouble with defending NET. With these two arguments considered, I shift in section III to constructing my own argument that we could never successfully defend NET. In IV, I suggest some initial reasons for thinking that that consequently we should reject this theory, which in turn would prove it false.

1. Why error theorists face self-defeat in arguing for NET

If reasons for belief carry normative content, then were NET true, there would be no reasons for belief, including reasons to believe this theory. Many philosophers, including some error theorists, have noted the paradoxical position of believing a theory according to which there are no reasons for belief. For example, Terence Cuneo (2007), opposing this theory, argues that

If they [error theorists] say that there are reasons to believe NET, their view is self-defeating. For the property of being a reason is a normative property, which does not exist if NET is true. But if error theorists say that there is no reason to believe NET, their view is polemically toothless. For if there is no reason to believe NET, it is not a rational mistake to reject this theory. (Cuneo 2007, 117–18; As quoted by Streumer 2013a, 203–4)

Cuneo takes reasons for belief to carry normative content. This allows him to formulate an objection from self-defeat in the sense of a performative contradiction error theorists commit while arguing for their position.7 Error theorists are giving reasons to believe their view according to which there are no reasons for belief. Stan Husi, an error theorist, concedes this worry observing that skepticism about all normative reasons “appears to be cutting off the very justificatory branch it sits upon, seeking to engage [in] a dialectical enterprise while denying its currency” (2013, 429). Husi, along with Jonas Olson and Chris Cowie, instead proposes reasons of a different sort for believing NET which don’t smuggle in normative content.8 If their strategy succeeds, then in supplying these non-normative reasons for their position, error theorists are free from Cuneo’s self-defeat objection.

However, Spencer Case (2020) argues that no matter how error theorists construe reasons for belief, they cannot avoid self-defeat.9 Even indicator evidence—evidence for a proposition which does not count as a consideration in favor of believing it—such as premises logically entailing their conclusion won’t, the following argument shows, save error theorists from self-defeat.10 Case takes this to be sufficient reason to reject this theory. I disagree. I think error theorists can avoid self-defeat while defending NET. However, my project to show that NET cannot be successfully defended crucially adopts elements of Case’s argument. So, his argument is worth reproducing at the start. Here it is in two steps.

Self-Defeat Argument

Step 1

(1) Error theorists are committed to the self-defeating proposition, “NET is true, but I have no reason to believe that”.

(2) If adopting any philosophical position commits us to a self-defeating proposition, then we should reject that position.

___________________________________________________________

We should reject NET.

Step 2

(3) If we should reject NET, then NET is false.

___________________________________________________________

Therefore, NET is false.

Self-Defeat Argument doesn’t stop at Step 1 because NET could still be true even if we should reject it. For example, a utilitarian might have practical reasons to reject an epistemically justified philosophical position (Case 2020, 3). However, Step 2 capitalizes on the normative property ascribed by (2), namely, the property of being obligatory to reject theories entailing self-defeating positions. If we should reject NET, then there is at least one instantiated normative property, but NET eliminates normative properties, so it’s false.

(1) and (2) need support. In support of (2), Case argues that if error theorists are willing to bite the bullet and accept that their position is self-defeating, they should be willing in principle to accept other equally counterintuitive positions that, say, reject a proscription against killing and eating our own children or a proscription against holding contradictory beliefs, provided that such positions are less counter-intuitive than accepting a self-defeating position. After all, that a self-defeating theory correctly represents the world is already highly counterintuitive, so there is no reason, in principle, that the proponent of such a theory should reject comparably counterintuitive commitments.

The whole argument turns on (1). If there are reasons for belief of a sort which do not carry any normative content implicit or otherwise, (1) is false. In support of (1), Case offers the following:

Weak Normativity Argument

(4) The normative error theorist’s partisanship toward the epistemic domain either makes a normative difference to him or it does not.

(5) If it does not, then the error theorist remains committed to self-defeating propositions.

(6) If it does, then NET is false.

___________________________________________________________

Therefore, NET is either self-defeating or false.

It is here that I find a resource for my own case against NET. The dichotomy between reasons which make a normative difference to us and those that don’t is, by my lights, crucial to seeing the problem for NET. How does Case put this distinction to use? Case contends that if error theorists can offer only reasons for believing NET which make no normative difference to them, then they can offer only reasons which need not make any difference to opponents in the debate. If reasons for belief are not considerations in favor of belief, considerations, that is, which obligate one to at least refrain from unreflectively dismissing them before rational deliberation, error theorists are once again polemically toothless. With Cuneo (2007), Case thinks that without considerations which weigh normatively on believing NET, error theorists are polemically toothless, which is to say that they’re in a self-defeating position (premise 5). Alternatively, if error theorists can offer reasons for believing NET which do make a normative difference to them, they now re-introduce normativity into discussion, which is inconsistent with NET (premise 6). Either way, error theorists cannot avoid self-defeat.

In this respect, Case notes that if error theorists want to insist on entitlement to reasons for believing NET—where “reasons” are understood non-normatively—self-defeat persists. Error theorists are here committed to saying, “Error theory is true, but there is no reason—of a kind that anyone need take the least bit seriously, all things considered — for anyone to believe it” (2020, 8; emphasis mine). Stripping reasons for belief from any kind of binding authority might save our ability to believe NET (contra Cuneo 2007), but it will not save error theorists from self-defeat.

However, the problem with Weak-Normativity Argument is that it leaves open an escape route for error theorists. Both Cuneo’s objection and Weak-Normativity Argument assume that error theorists are committed to believing NET, that is, they assume that error theorists always believe the theory they defend. The “self” in “self-defeat” refers to a problematic relationship to believing NET. But what if we cannot believe this theory? If we cannot believe NET, then error theorists do not defend it from a place of commitment to it. Error theorists can then avoid self-defeat altogether by adopting a cognitive attitude of non-belief in this theory. No performative contradiction arises from defending a theory which eliminates reasons for belief if I don’t believe what I’m defending.

This is how I understand Bart Streumer’s recent arguments for NET. Streumer (2013a; 2017a) argues that we cannot believe NET, but that, surprisingly, our inability to believe it fortifies error theorists against self-defeat and other reductio ad absurdum objections. After all, our inability to believe a theory does not make it false. Case reads Streumer’s position as biting the self-defeat bullet, but we don’t need to construe Streumer’s position this way. If I’m right, and if Streumer’s argument for NET’s unbelievability is successful, error theorists can ward off self-defeat objections without appealing to alternative reasons for believing it. Instead, they can claim to successfully defend it by insisting that we cannot believe it. To show all attempts at NET defense futile, I must therefore show that not even NET’s unbelievability can restore its polemical force in the debate. I next introduce Streumer’s argument as a potential escape from self-defeat objections, but in so doing, I observe what I consider a worse problem for error theorists.

2. How error theorists escape self-defeat only to face theory defense futility

Streumer’s argument for NET’s unbelievability (hereafter, Unbelievability Argument) provides an escape from self-defeat and many other objections. But I contend that this argument also betrays the necessity of normativity for successfully defending a theory. In this section, I analyze Unbelievability Argument and demonstrate its force in blocking objections to NET. However, I end by observing its proponents’ unintended application of normative reasons for belief.

The relevant terms in Unbelievability Argument are reasons for belief, belief, normative judgments, and normative properties.11 Streumer qualifies belief to mean full, confident, non-compulsory, rational belief which excludes partial, somewhat confident, compulsory, or crazy belief (Streumer 2013a, 197; 2017a, 7).12 By rational, Streumer only means closed under believed entailment (believing what I believe is entailed by my beliefs), which he takes to be a descriptive property with no normative bearing on belief.13 By reasons for belief, Streumer means any consideration in favor of a belief, and he takes considerations in favor of a belief to weigh normatively on belief.14 In support of this, he says that “reasons for belief are considerations that we base our beliefs on, and we cannot base a belief on a consideration without making at least an implicit normative judgment” (2013a, 198). Normative judgments are beliefs which aim to represent the world. So, when NET says that “normative judgments are beliefs which ascribe normative properties”, this is a cognitivist position about normativity such that our normative judgments aim to represent instantiated normative properties (Streumer 2013b). An example of such a judgment may simply be that we ought to believe in light of the supporting evidence. In what follows, I take these terms just in the sense Streumer takes them.

NET can be construed as the conjunction of the following two propositions:

(J) Normative judgments are beliefs which ascribe normative properties.

(P) Normative properties are never instantiated.

Streumer argues for three claims about this theory. He argues that NET is unbelievable, that NET’s unbelievability undermines objections which have been made against it, and that we can come close to believing this theory and it may be a rational mistake not to. With these claims in hand, error theorists can argue that though NET cannot be believed, this does not make it false, and competing normative theories such as normative realism (including reductive realism) and normative non-cognitivism are false, which makes NET more likely true.

Streumer begins his argument by proposing two claims about belief:

(B1) We cannot fail to believe what we believe is entailed by our own beliefs.

(B2) We cannot have a belief while believing that there is no reason for this belief.

If these claims are true, error theorists can then argue as follows. Anyone who believes NET believes that there are no normative properties. Reasons for belief are normative properties, so if NET is true, there are no reasons for belief.15 So, by (B1), anyone who believes NET believes that there are no reasons for belief. But by (B2), we cannot have a belief while believing that there is no reason for this belief. Therefore, NET is unbelievable.

As it stands, Unbelievability Argument does not show that there are no reasons to believe NET. The conditional claim “if NET is true, there are no reasons for belief” does not (alone) entail that there are no reasons to believe NET. So instead, Streumer (2013a, 199–200) offers the following two claims about reasons:

(R1) There cannot be a reason for someone to do x if this person cannot do x.

(R2) There cannot be a reason for someone to believe that p if this person cannot believe that p.16

If these claims are true, error theorists can then argue as follows. We take reasons for a belief to count in favor of that belief just as reasons for an action count in favor of that action. So, if (R1) is true of actions, then it follows that (R2) is true of beliefs. But if (B1) and (B2) about beliefs are true, then we cannot believe NET. By (R2), we cannot have a reason to believe what we cannot believe. Therefore, there are no reasons to believe NET.

In summary, we can construct Unbelievability Argument as follows:

(P1) According to NET, normative properties are never instantiated.

(P2) Reasons for belief are normative properties.

C1 Therefore, if NET is true, there are no reasons for belief.

(P3) Anyone who believes NET believes C1.

(P4) We cannot have a belief while believing that there is no reason for this belief.

C2 Therefore, we cannot believe NET.

(P5) We cannot have a reason to believe what we cannot believe.

C3 Therefore, there are no reasons to believe NET.

The argument does not stop at C2 because knowing that <if the error theory is true, there are no reasons to believe it> cannot make us believe that there are no reasons to believe it if we cannot believe the antecedent of this conditional claim. So instead, we need another reason to believe there are no reasons to believe NET. (P5) provides this reason.

If Unbelievability Argument is successful, error theorists are in an improved position in the dialectic. Normally, demonstrating a theory’s unbelievability would count against that theory, but in this case, it supports normative error by protecting error theorists from objections directed at believing error theorists (e.g., Olson 2014). If error theorists cannot believe NET, these objections miss the mark.

Finally, Streumer contends that we can come close to believing NET so long as coming close to believing a theory is less than full belief in that theory, meaning that this claim does not contradict (B2). But coming close to belief is not merely partial or weak belief in NET; rather, it is to be convinced that these arguments together seem to show that NET is true (2013a, 203).17 Streumer argues that we can come close to believing this theory by believing arguments in favor of (J) (that normative judgments are cognitive) without explicitly believing (P) (that normative properties do not exist) and by, at a later time, believing arguments in favor of (P) without explicitly believing (J). We can also come close to believing NET by believing arguments against alternative theories; for example, we can believe that, contra irrealist, theories normative judgments really do aim to represent the world, and we can believe that contra realist theories there really are no normative properties.

The strength of Unbelievability Argument lies in its ability to block objections. First, recall Cuneo’s (2007) observation that error theorists are either arguing from a self-defeating position if there are reasons to believe their theory or are polemically toothless in the debate if there are no reasons for belief. Says Streumer in reply:

[This] only shows that if NET is true, there is no reason to believe NET. And the belief that this conditional claim is true will only make us believe that there is no reason to believe NET if we already believe NET, which I have argued we cannot do. (Streumer 2013a, 204)

NET’s unbelievability blunts the force of Cuneo’s (2007) objection.18 Error theorists here argue for NET without believing it, thus retaining polemical teeth in the fight and avoiding self-defeat.

Does this mean error theorists are guilty of a form of bad faith in defending a theory they don’t believe there is any reason to defend? No. Our ability to come close to believing NET at least partially returns error theorists’ dog to the fight. Says Streumer:

Since we can come close to believing the [normative] error theory in these ways, there can be reasons for us to come close to believing it in these ways, and it can be a rational mistake if we do not come close to believing it in these ways. (Streumer 2013a, 204; emphasis mine)

If there are reasons for coming close to believing NET, then error theorists have reason to argue for this theory. And if so, they are saved from bad faith objections.19 In this way, Unbelievability Argument is a powerful strategy for error theorists: it provides an escape from the self-defeat which afflicts every card-carrying error theorist by denying everyone a card, but it also preserves reasons for taking NET seriously since it may be true and it may be a mistake to fail to come close to believing it.

However, we are now beginning to see the trouble for error theorists with attempting to successfully defend NET.20 NET eliminates normative properties. But talk about the strength of Unbelievability Argument in blocking objections seems to be an appeal to this argument’s polemical force in the debate; that is, it seems to appeal to its normative difference to us. Just so, talk about “reasons for coming close to believing” and the “rational mistake” we commit in failing to do so pack no punch in the dialectic if such talk is stripped of anything which weighs normatively on belief. In what follows, I develop this observation into an argument for the indefensibility of NET.

3. Why we could never succeed in defending normative error theory

Can we successfully defend a theory without presenting any reason for believing it? What about a theory according to which there are no reasons for believing it? Can we sincerely do these things? If you are a normative error theorist, you might think that we can. After all, if we cannot believe NET then we don’t, and if we don’t believe it, perhaps we are entitled to marshal normative reasons for belief in its defense.

However, even if we concede that error theorists’ belief in reasons for belief remains safe21 while defending NET (and I will make this concession), this concession won’t do enough to ensure the possibility of a successful defense. This is because to successfully defend NET, we must meet the abovementioned necessary conditions for successful theory defense which we can consolidate into the following:

Theory Defense Condition: We can successfully defend a theory T only if it is possible for us to offer at least one theory defense reason which counts as a consideration in favor of T.22

And I contend that with respect to NET we cannot meet this condition (hereafter, TDC). Before examining this claim, first notice how intuitive this condition is for successful theory defense. Theory defense is a communicative act wherein we express theory defense reasons to interlocutors. And it won’t be just any reasons which count toward success but reasons which actually favor T’s being true. So, if I cannot even in principle offer at least one theory defense reason which favors T, any reason communicated will be no reason against rejecting it, and in that case, I will never have successfully defended T, no matter how many alternative (theory defense irrelevant) reasons I offer. Now suppose S thinks it is possible that someone could successfully defend T. Even if S remains unaware or unable to express a theory defense reason favoring T herself, if she thinks that T could be successfully defended by someone, surely we should take her to think that at least one such reason is available to some other potential defender of T. By contrast, if S correctly believes that no theory defense reason favors T such that no one could ever advocate for T by providing reason against believing it’s false, we will naturally say that S correctly believes defending T a necessarily futile exercise.

Even so, it might look to error theorists like I’m smuggling normativity into a necessary condition for successful theory defense. After all, being correct to believe looks like a normative property.23 If it is, then it looks like I beg the question against error theorists by introducing in advance a condition for theory defense which requires that we correctly believe it is impossible to offer at least one theory defense reason in favor of T. The same is true if I assume that theory defense reasons, here required for theory defense, are themselves normative properties.

As before, I am here only considering theory defense reasons in the sense of reasons which motivate, on pain of being irrational were they thoughtlessly dismissed, to avoid rejecting T. And I continue to remain neutral about their normative status while advancing my argument. The same goes for being correct to believe. If it is correct to believe that two and two make four, and I believe this, then it is irrational for me to reject this claim, whether or not I believe being correct weighs normatively on belief. Later I will propose that we actually do have independent reason to accept the normativity of theory defense reasons, but here I rely only on what I take to be acceptable to error theorists. So, I do not beg the question against NET.

Would NET fail to meet TDC, Unbelievability Argument would be of no use for a successful defense of it. There would be no epistemically relevant reason for opponents to believe the premises of the argument nor to reconsider NET in light of it. Unbelievability Argument would give us no such reason to even come close to believing NET nor help us see why it would be a rational mistake to fail to do so since a theory’s supporting evidence just is a theory defense reason favoring it, and we cannot provide evidence for a theory which no theory defense reason supports. Theory defense reasons here extend far beyond reasons for believing the theory itself. They extend to reasons for believing at least one reason favors accepting or disfavors rejecting a theory T, reasons for believing the premises of arguments whose conclusion advances T in some way, and reasons for believing that objections marshalled against T fail. If belief comes in degrees, then reasons for belief extend even to those reasons which raise our credence level in T to any degree.

Yet NET does fail to meet TDC. Clearly, if theory defense reasons are normative properties, then NET is false. But as before, if they are epistemically non-normative, then opponents can safely ignore them. Earlier, I introduced Case’s Weak-Normativity Argument which puts a dilemma to believing error theorists who in support of NET either offer reasons for belief which render them inconsistent or reasons for belief which make no difference to opponents. We saw how error theorists can escape this dilemma. However, I have now re-directed this dilemma toward theory defense reasons, and I no longer see an escape route for error theorists. Epistemic norms are what provide polemical force to an argument. What we ought to believe compels us on pain of epistemic vice to follow the imperative. We ought not hold contradictory beliefs, for example, on pain of being irrational. Redefining “irrational” in merely psychologically descriptive terms strips it of its argumentative force in philosophical discussion. Just so, error theorists might offer alternative weapons of defense such as reasons of personal preference or pragmatic reasons for advancing NET. But if such reasons are normatively bankrupt, unless I share this preference or those practical goals which render NET useful to me, I can safely ignore these reasons in the debate. Reasons which we can safely ignore fail to count as considerations in favor of a theory’s being true. So, NET fails to meet a necessary condition for rendering even possible a successful defense of it.

It might be objected that so long as error theorists present arguments whose premises, if true, guarantee the truth of NET, and evidence that makes these premises likely to be true, they adequately defend NET. TDC appears to problematically sunder truth from normativity, however, since if NET is true, there are no theory defense reasons, so defendants here fail to meet TDC. Yet, there is nothing stopping error theorists from offering evidence that supports the truth of NET by way of premises and a conclusion or by evidence against objections to NET. So, if NET is true, error theorists—despite offering valid arguments with likely premises which conclude that NET is true—have not successfully defended NET which seems absurd.

In response, we should first note that, as before, whether theory defense reasons are normative properties will be a matter of disagreement between error theorists. Error theorists who reject their normative status will, therefore, read TDC as void of normative commitments in which case NET’s defendant has not failed to meet TDC in the above objection. However, this view faces the Weak-Normativity Argument as we’ve already seen since opponents can safely ignore normatively bankrupt reasons offered in defense of NET.

On the other hand, error theorists who accept the normativity of theory defense reasons which include reasons for belief remain consistent only if they also accept the normative status of the relevant notions of evidence and truth which, as before, are also theory defense reasons. After all, what epistemic value would these notions have in relation to theory defense if we have no epistemic obligation to prefer rational, evidentially supported, true beliefs over irrational, evidentially unsupported, false ones? And to say that we should prefer the former over the latter is to say that these notions carry normative content. To be sure, if NET is true, all judgments deploying theory defense reasons would here be false since there would be no epistemic norms. But can we believe the antecedent of that conditional claim? After all, it was the normative status of considerations which count as favoring NET which Unbelievability Argument granted so as to show that we cannot believe the antecedent of these conditional claims which suppose the truth of NET. The purpose was to ward off reductio ad absurdum objections. By that line of reasoning, NET’s being true cannot make us believe that no theory has ever been successfully defended since we cannot believe NET. Yet all of this shows only that even if NET is true, error theorists despite all appearances have not successfully defended it—not because my arguments divorce truth from normativity but because NET entails the unbelievable consequence that no one has ever successfully defended a theory.

Finally, you might still think that we simply do observe successful defenses of what the defender believes is indefensible. Suppose, for example, that a professor is teaching Kantian ethics to undergraduates. Suppose the professor presents Kant’s main reasons in support of his ethical theory and subsequently answers every students’ objection just as she thinks Kant would (or should) answer it. Suppose further that this professor doesn’t find her students’ objections convincing; rather, she thinks that a Kantian could easily dispense with them. And yet, let’s imagine this professor to be strictly committed to a form of act consequentialism, and that she believes that there simply are no theory defense relevant reasons in favor of being a Kantian at all. Since she has replied to her students’ objections, can we not say that she has defended what she personally believes there are no good reasons to believe, and is therefore a counterexample to TDC?24

The objection clarifies the difference between failing TDC and less disastrous ways a theory might lack support. TDC does not speak to theories we personally believe lack even one theory defense reason favoring it. Theory defenders in such cases can remain open to the possibility of being surprised by an objection not considered. By contrast, if, in the above case, the professor correctly believes it impossible for any potential Kantian ethics theory defender to offer any theory defense reason favoring it, she could not also believe (and remain consistent) that responding to objections counts as a reason in favor of this theory’s being true; that second belief of hers would just be false. If the professor correctly believed that no theory defense reason could ever become available to any Kantian ethics defender, she should therefore say that not even her replies to her students’ easily dispensable objections give her or them any reason favoring Kantian ethics; otherwise, her replies themselves would work against Kantian ethics (if indefensible) by instantiating those very properties (theory defense reasons) she denies are available to any defender of Kantian ethics. If she claims to be defending Kantian ethics, she surely is not, on these suppositions, successfully defending it. This is precisely what makes NET so unusual. We do not normally rule out in advance the possibility of any theory defense reason supporting belief in a theory. Yet, unlike Kantian or consequentialist moral theories, NET itself rules out the possibility of any attempts (including responding to weak objections) counting as considerations in favor of its being true. If I, for example, were to defend NET against my students’ easily dispensable objections, while correctly believing that there are no theory defense reasons in favor of NET—correctly believing that NET fails to meet TDC—I would have to concede to these students that insofar as responding to objections counts toward successfully defending NET, my replies were, in fact, utterly futile toward its defense, even while successfully responding to their objections.

4. Theory defense failure is not safe

If my arguments preventing successful defense of NET are sound, what does this mean for error theorists? I take our inability to successfully defend NET a serious concern for error theorists, but I leave a complete exploration of the problems for later work. Instead, in this section I offer some initial suggestions highlighting the sort of problems which lurk behind NET’s failure to meet a necessary condition for successfully defending theories. I think that the arguments sketched below give us, at the very least, good reason to warn error theorists not to completely ignore our inability to successfully defend NET.

You might think that our inability to successfully defend NET is no problem for error theorists. Recall that many objections to NET target error theorists who appear to argue from a self-defeating position. As was shown, this does not entail that NET is false. Like global skeptics, error theorists could insist that NET might be true and that, in support, good arguments show competing normativity theories to be false. If so, then we might think of error theorists as normativity messengers with a skeptical message. Shooting the messenger won’t absolve us of the skeptical problem. You might think, for example, that some claims are not successfully defensible yet just as conceivably true as conceivably false. Consider the claim that the number of stars in the Andromeda galaxy is an even number. Suppose we lack sufficient information to favor odd or even. If we cannot present information which counts as evidence favoring an even number, it looks like we cannot successfully defend this claim. However, this does not mean that we should reject the claim. After all, the number of stars may in fact be even. Since theories consist of claims, it might be that some theories are true, yet, we cannot successfully defend them on grounds of insufficient information. After all, some philosophers think that skeptical arguments are valuable not because anyone believes their conclusions but instead because they teach us important lessons for our epistemologies and because it is not obvious where these arguments go wrong.25 If that’s right, then can we not provide along these same lines some safety for NET from my objections?

Before responding, it is worth noting that error theorists do not take themselves to be offering a skeptical puzzle for us to solve collectively.26 Streumer only tells us we can’t believe NET enroute to defending it, and he wants us to join him in coming close to believing it by rejecting opposing views. All the same, while arguments for NET are still valuable for our moral epistemologies, and we shouldn’t reject a theory on rhetorical grounds alone, I think our inability to successfully defend NET gives us philosophical reasons to reject it, even while conceding the rhetorical point that problems for the messenger don’t disprove the message. That is, I think that if we know in advance about a theory, T, not merely that we lack sufficient information to mount a convincing case in favor of T (as in the odd or even case above), but that T rules out tout court the possibility of any theory defense reason ever becoming available to anyone, then it seems like the rational thing to do is to reject that theory. What I am suggesting here is that if a theory fails to meet TDC, that fact alone seems to give us good reason to reject it. The problem for NET in failing to meet TDC, isn’t just that it lacks favoring evidence “in hand” to present to opponents (evidence which may yet come), but that nothing could ever count as evidence since the theory itself either guts favoring relations of normative force or flatly eliminates them. TDC failure means no evidence is possibly available to us to offer on a theory’s behalf not that we currently suffer some access limitation which may one day be overcome. As such, it is hard to see any serious reason to consider it a possibility any longer even if we fall short of disproving it. Is this not an ad hoc move against error theorists? To be sure, it is difficult to think of any theory like this other than NET (as far as I know, no other theory eliminates all normative properties or at least all instantiated ones). All the same, it seems to me an independently plausible principle that demand at least one theory defense reason be possibly accessible to us to offer in favor of some theory to ensure the possibility for us of successfully defending it. If my intuition is correct, we won’t need any other theory to justify application of the principle toward rejecting NET.

In suggesting that we should reject NET, I am not antecedently ruling out the possibility that NET is true. Yet, if theory defense reasons carry normative weight, then we are closer to knowing that it’s false. And it looks like they are. As before, denying defense reasons’ normativity results in a self-defeating position for error theorists, and while accepting their normativity is a problem for defenders of NET, it is no problem for anyone else, who, like me, thinks that we are justified in rejecting it. Consider that if I’m wrong, and it would be a rational mistake to reject NET, then despite no theory defense reason possibly counting in favor of it, we would still not enjoy justification in rejecting it. This would mean that, even were it true, the truth of NET would itself be no reason to believe it! This is absurd.27 Likewise, any amount of evidence—such as reasons for belief, arguments, responding to objections (etc.)—marshalled against NET would not, by supposition, justify rejecting it.

Still, you might recall that Streumer advocates adopting the cognitive stance of coming close to belief . You might therefore think, following Streumer, that error theorists could offer reasons to “celieve” NET sufficient for theory defense where “celieving” is between rejecting and believing a theory. After all, we have seen that Streumer thinks there are enough reasons favoring NET such that it is a rational mistake to reject it. To hold that defending NET requires giving reasons for belief and not celief in NET is question-begging. Therefore, NET remains defensible. Note, in response, that Unbelievability Argument might run by parity just as well on “reasons for celief” as “reasons for belief”. If so, then we cannot celieve NET either. Of course, if the parity argument fails, then we can celieve NET. But the determining factor remains the same: are reasons for celief normative on celief or not? Our Case-style dilemma returns: if no, NET is indefensible for weak-normativity reasons; if yes, NET is false. The same problem faces error theorists’ definition of “theory defense”. If error theorists insist that they “defend” NET—where “defend” smuggles no normativity into play—opponents can safely ignore whatever “theory defense reasons” they offer, and otherwise, error theorists rely on the instantiation of what NET denies is instantiated. In either case, the claim that we’re justified in rejecting NET is not affected by introducing normatively deflated definitions of these terms.

The plausibility of my rejection proposal might come to light in the following analogy. Suppose you’re told about a product called MoneySucker©. The only function of this product is to suck money and give nothing in return. It would be silly for a consumer to buy this product. However, suppose it turns out that  no one can buy this product. It’s not for sale and never will be. Perhaps we’re not justified in rejecting the product out of hand. After all, we can’t buy it, so perhaps we can’t ever be sure that it would be a bad purchase. Now suppose you encounter a street salesman promoting MoneySucker©. He yells to passersby: “End all spending”, “Purchases are evil”, “MoneySucker© is the only worthy product remaining because we have no reason to buy things!” You ask him why he’s selling MoneySucker© if it can’t be purchased and we have no reason to buy things? He replies: “For a small sum, I’ll tell you why”. But why should you spend to learn why spending is evil and that a product which is not for sale whose function is to suck dry all our spending power is the only worthy product remaining? You should reject that offer. Similarly, you should not “buy” the arguments of proponents of a theory that “sucks dry” the “currency” of theory defense reasons needed to successfully defend it, which places some of its “sellers” without opponent “purchase”, and which cannot be “bought into” leaving its sellers unable to “make the sale”, that is, unable to successfully defend it. In other words, the reasonable response when presented with a theory (even if NET is the only one) according to which no theory defense reasons could ever count as considerations against rejecting it is to reject it.

If we should reject NET, then it would follow straightforwardly that NET is false. To be sure, showing that we should reject a theory does not always mean that this theory is false. Plausibly, there are claims which we should reject without knowing that they are false. You might think, for example, that we cannot know that our reason is reliable. Even so, it would be rational to reject the claim that our reason is entirely unreliable even if we cannot be certain that this claim is false. However, things are different for NET. Recall that if any normative property is instantiated, NET is false. Now consider the following argument.

Indefensibility Argument

Step 1

(7) We cannot successfully defend NET because it fails to meet Theory Defense Condition (TDC).

(8) If a theory fails to meet TDC, we should reject it.

___________________________________________________________

We should reject NET.

Step 2

(9) If we should reject NET, NET is false.

___________________________________________________________

Therefore, NET is false.

If (8) is true, it’s easy to see how this argument would succeed. I’ve already shown that (7) is true. In this section, I’ve begun to motivate (8) by offering reasons for thinking that we are justified in rejecting NET on grounds that it fails TDC. Unlike the self-defeat argument, (8) does not depend on anyone adopting the attitude of belief in NET while defending it. And as in the Self-Defeat Argument, rejecting (8) looks at first blush more problematic than rejecting other widely held intuitions such as the Law of Non-Contradiction or the claim that it is impermissible to torture innocent children for fun. Rejecting the requirement for possibly offering even one theory defense reason in favor of the theory we defend looks close to saying that in philosophical discourse, there is no difference between good arguments and bad ones. In other words, the intuitive costs of rejecting (8) seem to me so high that it clearly looks like a rational mistake to do so.

You may disagree. Yet even if you’re right, supposing (8) is true, what clearly follows is that we should reject NET because there would be at least one normative property instantiated, the property of being right to reject a theory we know we could never possibly successfully defend; and if so, then NET is false. The argument would succeed regardless of whether or not error theorists are committed to a self-defeating proposition. According to NET, no normative properties are ever instantiated, so (9) would be true by definition. The prima facie plausibility of such an argument, even if not yet completely convincing, I think, already shows that the consequences for error theorists following from our inability to successfully defend NET are not benign, that is, they are the kind that (never-successful) defenders of NET cannot safely ignore.

5. Conclusion

I have argued that an argument for NET’s unbelievability provides an escape to a self-defeat objection to this theory. But it’s a pyrrhic victory, since from these arguments, we can now clearly see that any attempt to defend NET is futile. At first, it might seem crazy to argue that there is a theory which we cannot successfully defend. But when we consider how strange it is to try to defend a theory which entails that there are no normative reasons to believe it, we realize that our inability to succeed in defending this theory is no less strange. I concluded by suggesting that as a result we should reject NET. And if we should reject NET, then there is at least one normative property instantiated, and NET is false. I have not here provided a complete defense of these two consequences following from my central objection to NET, but their initial plausibility strikes me sufficient to make theory defense failure a significant concern in this case.

Acknowledgments

Thanks especially to Bart Streumer, Spencer Case, Christian Lee, Hud Hudson, Joe Milburn, John Greco, Logan Gage, and Mark K. Spencer for comments on early drafts. Thanks also to Anthony Sloan, Lukas Rowland, and Jonathan Taccolini for helpful discussion, and to two anonymous referees for helpful comments on earlier drafts.