Epistemic & Moral Irrealisms:
On GIBBARD & FIELD
3000 words; 11 pages
Abstract (90 words):
While moral irrealisms are common, epistemic irrealisms are rare. Allen Gibbard and Hartry Field, however, are self-proclaimed epistemic irrealists: Gibbard is an epistemic expressivist and Field a kind of epistemic relativist. They both hold that epistemic judgments – about what’s epistemically (ir)rational, or (un)justified, or ought (not) to be believed, and so on – are never literally, stance-independently true. I argue that their positions are false and their arguments unsound. This importantly implies that their arguments for moral irrealism are unsound also; a similar objection applies to all other common arguments for moral irrealism.
Introduction
Most arguments for moral irrealisms can be presented as instances of a schema like this:
(1)Moral judgments have features Φ.
(2)Any judgments with features Φ are never true.
(3)Therefore, moral judgments are never true.[1]
These arguments’ advocates typically presume that an appreciation of their premises and their relation to the conclusion can provide epistemic reason to accept moral irrealism: this understanding can justify someone in accepting a morally irrealistic conclusion, make it reasonable or rational, and such that it should be believed and its negation denied.
Call these kinds of judgments epistemic or intellectual judgments. Most philosophers, including most moral irrealists and even traditional epistemological skeptics, think that such judgments are sometimes true, and true literally and stance-independently.[2]
This presumption, however, suggests meta-epistemological implications. First, a semantic presumption that epistemic judgments are descriptive, truth-apt, and propositional. Second, a metaphysical or ontological presumption that there are truth-makers for epistemic judgments, i.e., epistemic properties or facts, that typically carries with it the presumption that these truth-makers are “stance independent” or “objective”: epistemic judgments, when true, are not made true by anyone’s (or any community’s) attitudes toward them.[3] Third, that there are (not surprisingly) epistemic beliefs and other attitudes towards epistemic propositions (and so an epistemic psychology). And the presumption raises the related question an epistemic epistemology.
Call this set of meta-epistemological presumptions “epistemic realism,” as they are analogous to the set of meta-ethical presumptions called “moral realism.” Although it seems that most philosophers are epistemic realists, Allen Gibbard and Hartry Field, however, are not. As self-proclaimed epistemic expressivists and relativists, they deny that epistemic judgments are ever literally or stance-independently true.
Here I argue that their epistemic irrealisms are false and that the arguments for their positions unsound. These results are important not so much because it’s worthwhile to criticize rare views like epistemic irrealisms (although that can be worthwhile), but because of the more-pressing consequences for moral realism: if Gibbard’s and Field’s arguments for epistemic irrealism are unsound, then their analogous arguments for moral irrealism are also unsound. This because, for Gibbard and Field, their arguments against both kinds of realism share a major premise, viz. instances of premise (2) in the schema above, that I argue is false and unreasonable to accept.
Thus, undercutting arguments for epistemic irrealism has the important consequence that arguments for moral irrealism are undercut. This is true not just for Gibbard’s and Field’s arguments for moral irrealism, but, I suggest, for all the common arguments for moral irrealisms. This is because these arguments’ major premises – instances of (2) – also suggest epistemic irrealisms (although, unlike Gibbard and Field, their advocates did not recognize this). Thus, I aim to raise a general, but powerful, objection to all common arguments against moral realism: each has a premise that we have good reason to regard as false.
- Brief Statements of Gibbard’s and Field’s Views
Let us first understand the views in question, before we consider the arguments in their favor.
Gibbard has a general theory of what it is to judge or consider something to be “rational” that has implications for both moral and epistemic evaluations. On his view, the notion of “rationality” is a part of a cluster of what he calls “normative” notions, and a “notion is normative if we can paraphrase it in terms of what it makes sense to do, to think, or to feel.”[4] His sense of “rationality” pertains to moral evaluations (he relates moral evaluations to the “rationality” of various sentiments, like guilt and resentment in response to an act[5]), to beliefs, and to other items of evaluation. For belief, the notion of rationality has close connection to the epistemic notions that a belief “ought” to be held, is “warranted, and is “well grounded,” and is “justified.”[6]
Gibbard calls his theory the “norm expressivist analysis,” and according to it “to call something rational is to express one’s acceptance of norms that permit it.”[7] He explains that a norm is a “rule or prescription, expressible by an imperative,”[8] which is neither true nor false. Thus, to put these notions together, to call something rational to express one’s acceptance of rules, prescriptions, or imperatives that permit that thing, e.g., doing some action, holding some belief, and so on.
Gibbard elaborates on what it is to express a norm:
Normative talk is part of nature, but it does not describe nature. In particular, a person who calls something rational or irrational is not describing his own state of mind; he is expressing it. To call something rational is not to attribute some kind of particular property to that thing—not even the property of being permitted by accepted norms. . . . The analysis is non-cognitivist in the narrow sense that, according to it, to call a thing rational is not to state a matter of fact, either truly or falsely.[9]
Thus, Gibbard advocates an expressivism or non-cognitivism about both moral and epistemic judgments: none are literally true (although we might say they are, but only in a “minimalist” sense of “true”).[10]
Now let us briefly try to understand Field’s view. Although he claims to take inspiration from Gibbard’s norm expressivism in developing his own ethical and epistemological “evaluationism” or “non-factualism,”[11] his view is actually quite similar to later versions of Harman’s moral relativism.[12]
Consider two sentences that Field considers “evaluative”: first, that “voting for the lesser of two evils is a bad idea,” and, second, that “belief in quarks is justified on current evidence.”[13] He claims that these claims have a “not-fully-factual” status; he thinks they are true, but that they are true or factual only relative to a “norm,”[14] not true norm- or stance-independently. And these particular judgments will be true relative to some norms, but false on other norms.
This idea is common for moral judgments (it’s just moral relativism), but it’s less common for epistemic judgments. Field gives scientific example – that “belief in quarks is justified on current evidence” – but any epistemic judgment can be used to illustrate the theory. Here’s one, that, “My belief that I have hands is justified or reasonable.” On some theories, yes, the proposition that “my belief that I have hands is justified or reasonable” is true. But on other theories – more demanding, often skepticism-inducing ones – that proposition is not true.
This much everyone should agree with: different principles can have different implications for cases. Field adds to this the further, more controversial, claim that no evaluative norms are true or correct, what he calls “objectively correct,” or true in a non-relative manner.[15] A consequence of this is that there are no (particular or general) moral or epistemic truths, norms, or principles that are true necessarily.
For Field, moral or epistemic judgments are never just plain true. However, he suggests that, “we can say that an evaluative utterance is disquotationally true for me iff it is true relative to the norms I regard appropriate to associate with the evaluative terms.”[16] On this view two believers who accept different norms can disagree about, e.g., particular evaluative judgments, even when they accept differing norms (and even when these norms imply that each of their judgments are “true for them”), but this is a disagreement in “attitude,” not a factual disagreement.[17]
Thus, it should be clear why Field’s views are well described as relativistic. A token evaluative (moral or epistemic) judgment’s truth-value is determined by, or relative to, the speaker’s norms. This contrasts with more common, and evaluatively realistic, views on which, e.g., someone’s accepting Nazi moral norms would not make that person’s anti-Semitic moral judgments “true for her,” or an epistemological skeptics’ accepting very high standards for justification would not, in itself, make her belief that her beliefs are unjustified “true for her.” So, Field denies that there are epistemic and moral properties in an ordinary sense (in his view, “reasonableness” is not a “factual property”[18]) and goes well beyond what even contextualists typically maintain about how one’s epistemic principles can influence the epistemic status of one’s beliefs.
2. The Arguments for Gibbard’s and Field’s Views
Gibbard and Field accept these epistemological irrealisms because of the arguments that they accept for their moral irrealisms. They accept moral irrealisms because they accepting a version of the argument above: they think moral judgments have various features (e.g., Φ) and that judgments with these features Φ are never literally true, so moral judgments are never true. Unlike nearly all other moral irrealists, they just simply make the further observation that epistemic judgments have these features Φ also, so they come to the same conclusions about their nature and, thus, an irrealist meta-epistemology.
A concise statement of the arguments given for norm expressivism is not easy to provide. Gibbard’s discussion is dense and the considerations for his position and against others are developed in subtle, not very explicit, ways. But Gibbard thinks his norm-expressive analysis of what it is to call something rational “strains the concept less than do the alternatives:”[19] evidently thinking that we should accept understandings of concepts that are “less straining.” He thinks that to call something rational is to “endorse” it, but to say that something has some natural property is not to endorse it, so naturalisms are mistaken in their understanding of the notion of rationality.[20] This seems true for both ethical and epistemological naturalisms and related notions of rationality: e.g., saying of some action will produce the most pleasure, or saying that some belief is produced by a mechanism belief-forming reliable is does not seem to be the same as saying that an action is right or that the belief is justified. The later seem to have an endorsing function that the natural description lacks.
As a self-proclaimed “naturalist,” Gibbard rejects non-naturalistic theories of rationality: He writes, “Nothing in a plausible, naturalistic picture of our place in the universe requires these non-natural facts and these powers of non-sensory apprehension.”[21] If naturalistic and non-naturalistic theories of what it is to judge or consider something to be rational are inadequate, and yet the notion of rationality “makes sense,” as Gibbard puts it, we likely are left with some kind of non-descriptivism, like his norm expressivism. This conclusion is especially plausible if judgments about rationality have this “endorsement” function that cannot be captured by mere property attributions, whether natural or non-natural.
Field’s main argument for his view is that “we can accommodate all the [relevant moral and epistemic] phenomena using only the norm-relative notions, together with both preference among norms and norm-relative beliefs.”[22] So, basically, Field argues that evaluationism is the simpler hypothesis, simpler hypotheses ought to be accepted, and so evaluationism ought to be accepted. And by denying that there are epistemic and moral properties he can also avoid challenging epistemological questions for how we might detect them. In response, a factualist, i.e., a moral and epistemic realist, claims that Field’s evaluationism yields an “impoverished caricature of evaluative discourse,” but Field’s opinion is that “the nonfactualist will win” the debate.[23]
Whatever the considerations Gibbard and Field offer in favor of their positions and whatever their view’s exact nature, they clearly have epistemically irrealistic implications for judgments about the rationality of beliefs: there are no epistemic truths, facts, propositions or properties, and so no particular epistemic judgments are literally true or stance-independently true.[24]
There are interesting consequences for this kind of position, including for the nature of argumentation and reasoning. On Gibbard’s theory, when someone says that norm expressivism is a reasonable view or one that ought to be held, that person is not saying something that’s true. If they say that the arguments for norm expressivism are strong, or that they justify belief in the theory, or that those who accept these premises ought to accept this conclusion (and people ought to accept these premises) they are not saying something that’s true either. These are all logical implications of the theory, and the theory also implies that any (refined) claim like this is not true either: if you see that something is a logical consequence of your theory, you should accept that consequence. That is not true either.
For all these judgments, if Gibbard is right, only norms have been expressed; no truths stated. For those who do not share Gibbard’s norms (and even for those who do, for this would seem to be not true, if Gibbard is right: if you share Gibbard’s norms, then you ought to accept Gibbard’s conclusions, once you understand and accept his case), it’s not at all clear how or why we should respond to his expressions in any particular way, especially by agreeing with him.
Similar epistemically non-standard consequences follow from Field’s theory, but on his view an epistemic judgment is true when, and only when, one’s epistemic judgments jibe with one’s own norms that are, at best, only “true for the speaker,” never true in relation to any epistemic principles that are necessarily true. So, if we do not accept Field’s norms, or any other norms that would make it such that we ought to accept evaluationism, it is not true that we ought to accept it.
So not only do Gibbard and Field’s views have non-standard implications for a wide range of epistemological judgments, they also have implications for whether one ought to respond to their arguments and whether such a response would be justified or rational.
- What To Make Of This?
At this point, Gibbard and Field might simply reply that I have accurately reported on their theories’ implications and ask what the problem is.
While they might not see a problem, for most of the rest of us, however, their arguments lead to something that we regard as false – viz. that epistemic judgments are never literally or stance-independently true because they are either expressions or only relatively true – and we think have good reason to reject. We could develop these reasons; make a case for why it makes most sense to think that epistemic judgments are, first, either true or false and, second, why they are sometimes true and so why it makes most sense to think that judgments about what makes sense are sometimes literally, stance-independently true.
Reasons can be developed from, among many other sources, the (a) that epistemic error seems possible (if not actual), (b) that (to respond to norm-expressivism) a-epistemicists – people who lack an emotional motivation or affective “oomph” regarding their epistemic judgments – seem possible (if not actual), that (c) epistemic judgments seem descriptive, in light of their logical behavior, as well as (d) that, to many, epistemic judgments just seem true: they don’t seem neither true nor false, and they don’t seem always false either (and they also don’t seem to be true only “relative” to norms that are themselves either not true or only “true” relative to themselves).
Perhaps expressivism or relativism is true for epistemic judgments: it’s not obvious why epistemic judgments must be descriptive and sometimes true because of stance-independent epistemic properties. But I suspect that most of us – given what we believe now – could develop better arguments against epistemic expressivism than for it. So, all things considered, we have more evidence to think epistemic judgments are sometimes literally true and we can see that this implies that epistemic expressivism and relativism is false.
These arguments, however, will typically imply that various premises in arguments for moral irrealisms have at least one false premise – an instance of a general premise like the instances of (2) above that Gibbard and Field appeal to in making their cases for their moral and epistemic irrealisms – and so are unsound arguments. This, of course, undercuts many of the cases for moral irrealisms, including expressivisms and relativisms. So, if it’s reasonable to reject Gibbard’s and Field’s epistemic “norm expressivism” and “evaluationisms,” then it’s reasonable to reject at least one premise in their arguments. Since the ideal candidate is a general premise, linking features of epistemic and moral judgments to their lack of literal truth, rejecting their epistemic irrealisms gives reason to reject their moral irrealisms.