A forum for VE lucubration

Thursday, August 17, 2006

Simon Blackburn, conceptual priority, and the problem of explaining knowledge and truth

In his paper “Reason, Virtue and Knowledge” (2001), Simon Blackburn sets the bar high for what he takes to be required for a virtue-theoretic account of knowledge, truth and justification to amount to more than a mere “fig leaf for reliabilism.” On Blackburn’s view, for an account of VE to be properly VE, then epistemic virtue must have explanatory priority within the account; this is to say, the central epistemic concepts of knowledge, justification and truth must be explained in terms of intellectual virtue, and not the other way around.
The task of upholding the priority of virtue would require a “right-to-left” reading of the following equivalencies, the “right-to-left” upholding of which are in “ascending order of ambitiousness. ”

(1) A proposition is probable (justified) in a circumstance C if and only if an epistemically virtuous agent in C would have confidence in it.

(2) A true proposition is known to be true by an agent S in circumstance C if and only if S in C exhibits epistemic virtues in accepting it.

(3) A proposition is true if and only if an epistemically virtuous agent would accept it, if he exercised the virtues appropriately.

Blackburn suggests these equivalencies as candidates for comprising a proper epistemic model analogous to the manner in which central ethical concepts (right action, justified action, etc.) are cashed out in virtue ethics.
For example:
1. An action produces (or tends to produce, or is such to produce) the greatest balance of benefit over harm or any alternative if and only if it is the action that would be performed by a virtuous agent.
2. An action is the right action to perform in the circumstances if and only if a virtuous agent would perform it in the circumstances.

While it is tempting to read the former equivalencies in such a way that we find ourselves learning a bit about what virtue requires given antecedent conceptions of knowledge, probability and truth, this is exactly what must not be done, Blackburn thinks, for any virtue theory worthy of calling itself such.

I think that Blackburn’s suggestion has quite a few problems. For starters, it is not clear to me what the motivation would be for pursuing this project (within virtue epistemology) other than for the sake of pursuing a project that analogously models virtue ethics. I think that, as a general rule, such a project should raise a cautionary flag: pursuing a theoretical project in epistemology should never have a goal which might be incompatible with the project of discovering what is really the case. Blackburn thinks that a virtue-theoretic account of epistemic concepts must have a particular structure of conceptual priority to be “worthy of calling itself such.” Blackburn appears to not even consider the possibility that an epistemic theory could be successful whilst employing the use of intellectual virtues, and whilst not defending their conceptual priority within an account. I see no good reason to rule out, prima facie, that such an account could successfully come to grips with the nature of epistemic concepts, even if it isn’t “worthy” of being called a virtue theory (by virtue of modeling virtue ethics in such a way that conceptual priority is given to the virtues within the account).

This aside, I want to raise a problem that I think is probably irresoluble if Blackburn’s strategy is used. The problem has to do specifically with how he thinks that knowledge and truth are to be explained within his account. Recall Blackburn’s equivalencies:

(2) A true proposition is known to be true by an agent S in circumstance C if and only if S in C exhibits epistemic virtues in accepting it.

(3) A proposition is true if and only if an epistemically virtuous agent would accept it, if he exercised the virtues appropriately.

Whilst there is much to be said about Blackburn’s own solution for how these priorities can be defended by incorporating a deflationist account of truth (in conjunction with a “use” theory of meaning), what I think stands out as particularly crucial is a problem Blackburn brings up in his analysis of (2) and (3). Blackburn says: “The difficulty is that if truth is described in terms of what a virtuous agent would accept, knowledge cannot be similarly defined on pain of eliminating the distinction between the two. ” Certainly, collapsing knowledge into truth is a move that must be avoided. Blackburn surmises that there is some space between what a virtuous person does accept, and what a virtuous person would accept, and that this distinction might be able to adequately capture the difference between what truth is and what knowledge is. Blackburn writes: “[This distinction would] deliver the idea that truth is what you would get to by investigating virtuously, whereas knowledge is what you have got when you have investigated virtuously. ” I will address this claim shortly. But first, it should be noted that Blackburn thinks that, although this proposed definitional distinction is disputable, there is a positive upshot in it, in that it “does at least reflect the idea that there is normally no gap between aiming at knowledge and aiming at truth. ”
Three important questions emerge here: (1) Is it plausible to think that a virtue account could define knowledge and truth in such a way that they differ only in the respect that one “has been” and the other “would be” accepted by an intellectually virtuous agent? (2) Is Blackburn correct to think that this distinction reflects the idea that there’s normally no gap between aiming at knowledge and truth? (3) Is it the case that there is normally no difference between aiming at knowledge and aiming at truth?
Blackburn seems to answer each of these questions with a “yes”, and my inclination is to respond to each with a “no.” For the sake of keeping this post short(er) than it would be otherwise, I’ll address the first:

Is it plausible to think that a virtue account could define knowledge and truth in such a way that they differ only in the respect that one “has been” and the other “would be” accepted by an intellectually virtuous agent?

One way to pursue this line of inquiry would be to consider an inverting the terms and seeing how it looks. If an inversion of the terms with the definitions looks at all plausible, then this would amount to, prima facie, a strike against thinking that the relevant respect by which knowledge and truth differ is captured by the distinction Blackburn suggests. And so, let us ask whether (invertedly) knowledge could be what you would get to by investigating virtuously, and that conversely, truth is what you have got when you have investigated virtuously. It’s not false that you would be said to have knowledge if you would have investigated virtuously, even though (it would follow) you’d also have truth; similarly, it wouldn’t be false that once you have investigated thoroughly, you would have truth, even though you what you’d also have would be knowledge. Additionally, it’s not the case even that most cases of what you would get by investigating virtuously would preclude knowledge, nor that most cases (or any!) of what you do have when investigating virtuously would preclude truth.
What could be motivating Blackburn’s distinction here, then? The reasonable way to interpret his motivation, I think, is to suppose that Blackburn thinks that something “happens” once one has virtuously inquired, such that the truth that he only would have attained (in the counterfactual case) now becomes knowledge, by virtue of being the output of virtuous investigation. I’m guessing this to be the motivation because, surely, it isn’t rooted in his thinking that truth isn’t what you’d get when you have investigated virtuously (given that knowledge implies truth). And so, for Blackburn to be recognizing these terms as interestingly different, he must be thinking that there is something awry about claiming knowledge to be what you would get if you investigated virtuously (given that this is how he defines truth and presents it as differing in definition from knowledge).
At the end of the day, I think the prospects of adequately explaining what is relevantly different between truth and knowledge can’t be done if the conceptual priority of virtue is to be defended as Blackburn wants to do. And so, maybe we should conclude that a virtue account “worthy of calling itself such” should be, perhaps, abandoned for a virtue theory that’s not as worthy in that respect (but which can clearly discern how knowledge differs from truth.)

Intuition Check: Is There Room for Wisdom in the Demon World?

I’ve been thinking lately about epistemic value pluralism (as Riggs defends it), and particularly, what the prospects are for defending some non-alethic epistemic values relative to which some disposition could qualify as an epistemic virtue by promoting that value. The recognition of a non-alethic epistemic value has at least one obvious advantage: it allows us to avoid the dilemma of either: (i) denying an intuitive platitude, namely, that (not-clearly) alethic dispositions such as insight and openmindedness qualify as intellectual virtues; or (ii) maintaining that they are, but being burdened with the task of showing how these dispositions qualify as virtues relative to an alethic epistemic value (or values). Because it is not at all clear that insight and openmindedness aim at, or reliably produce, any alethic end (such as knowledge or truth), the dilemma of selecting between (i) or (ii) seems rather unpromising. To be clear, (i) entails denying the intuitive, and (ii) requires a defense of a position that is wide open to counterexample.

Epistemic value pluralism, then, appears to be a promising alternative given that the identification of some non-alethic epistemic value (of course, along side some alethic epistemic values) relative to which a disposition could qualify as an epistemic virtue circumvents the problem of having to select between (i) and (ii). Riggs thinks that “intelligibility” is a proper epistemic value, and so, his position allows for not-clearly-alethic dispositions to qualify as epistemic virtues.

I am quite sympathetic to Riggs’ position. This is because I am more prepared to abandon the idea that all epistemic values must be alethic than I am to reject that (for example) insight and openmindedness are intellectual virtues. And, to add, I see no way that either could qualify as an intellectual virtue so long as the epistemic value relative to which they would qualify is alethic.

I don’t, however, think that Riggs’ succeeds in defending (at least) insight as an intellectual virtue by linking it to the epistemic value of intelligibility. (I’m currently working on a paper that argues just this). My sympathy, though, to the prospect of recognizing a non-alethic epistemic value has led me to wonder what might qualify.

Against this background, I want to (tentatively) toss wisdom into the hat of examination. Wisdom certainly falls hand-in-hand with whatever we should call the epistemic good life, and as such, appears to be a candidate epistemic values relative to which dispositions could be said to be epistemic virtues for their promotion of it.

Intuition check: is wisdom alethic? Put more broadly, is there a necessary relation between promoting wisdom, and promoting “a coming to grips” with how things are?

I had always just granted this without much consideration; upon reflection, though, I think that I might have been too hasty.

One thought experiment (actually, a litmus test) I thought of in trying to determine whether an epistemic value is alethic is to ask: could it be reliably promoted in the demon world? (i.e. a radical demon world, in which we deceived in such a way that all our beliefs are false).

Alethic epistemic values such as knowledge and truth and understanding are such that no disposition could promote them in the demon world. Non-alethic values (such as intelligibility) could be promoted in the demon world.

What about wisdom, then? Before rushing to an answer, consider this clarification. I am not asking whether one could be wise in the demon world (although this is a fascinating question in its own right! I’ll return to this.) That’s not the question, though, because such a question would pertain to wisdom as a disposition and not an epistemic value. Because I’m interested in the prospects of wisdom as an epistemic value, the question must pertain to whether it could be reliably promoted by some disposition or dispositions in such a world.

At this point, I am going to crawl so far out on a limb that I can feel it breaking.

I propose (obviously tentatively) the following:

If there is an x such that x is both a disposition and an epistemic value, then if the disposition x can be achieved in the demon world, then the epistemic value x can be promoted in the demon world.

The above conditional seems at least somewhat plausible to me. Of course, it relies on the somewhat controversial assumption that there must be some reliability condition associated with the possession of a disposition. Montmarquet (and I think Fairweather) reject reliability conditions and so would reject my conditional.

If we suppose, for the heck of it, that it holds, then the next interesting question is to ask: can one attain wisdom in the demon world? If the answer to that question is yes, then granting my conditional, the epistemic value of wisdom could be promoted in the demon world. If it could be promoted in the demon world, then (if my litmus test is right) it is a non-alethic epistemic value.

And so, to the question: can wisdom be achieved in the demon world?

There are several ways to think about this. One way is to ask: Can Bob hold all false beliefs and be wise? To this, most would probably say “no.” But the question can be put another way: “If Jonas is wise in our world, then isn’t his demon-world counterpart wise?” I think we might be a bit hesitant to say “no.” Perhaps this is because we think of wisdom as something for which one is responsible. And as such, it might strike us odd to strip this from Jonas’ counterpart for something for which he is not responsible (i.e. that he is, unluckily, being systematically deceived). If we tread this (probably dangerous) road much farther, we can see what comes next: the stripping of a reliability condition from the disposition of wisdom, the stripping of which would run counter to our initial intuition that “Bob can’t hold all false beliefs and be wise.”
I’m not sure where to go from here, as both sides look about the same when you’re sitting on the fence. Suggestions welcome!

Wednesday, August 09, 2006

A Recipe for Counterexamples to Slote's "Sentimentalist Virtue and Moral Judgment"

In his 2001 book “Morals From Motives,” Michael Slote undertook a project to defend a virtue-based ethics of caring within the domain of decisively normative ethics; since then, and particularly in his essay “Sentimentalist Virtue and Moral Judgment”, he has attempted a broader project, which is to defend ethical sentimentalism (with an emphasis on the role of empathy) within a metaethical--and not solely a normative--domain.

In this essay, Slote highlights what I think are some accurate and insightful facts about human empathy as it is manifest in moral evaluation. (See, for example, his comments on temporal immediacy and causal immediacy of potential sentiment-invoking states of affairs, and their relationships to our empathic responses to such states of affairs with these properties of immediacy; his points on this score seems to me correct and, descriptively, helpful). This aside, I want to focus on a concern I have with Slote’s project, which pertains to an amendment Slote makes to Adam Smith’s response to Humean sentimentalism. Hume had, rather famously, maintained the descriptive claim that we tend to approve of traits in proportion to the beneficial effects (we believe) that they produce. Hume holds, as Slote puts it, that we have “empathetic pleasure or displeasure at those (likely) effects.” (Epistemic and Moral Virtues, Pritchard and Brady, p. 123).

Adam Smith had criticized this position by arguing that it cannot account for the fact that we do not approve or disapprove of inanimate objects “for their predictable effects on human happiness” (123). Smith, in “The Theory of Moral Sentiments” offers an agent-based, rather than an effect-based, alternative:

“We approve of someone’s motives if, when we put ourselves in their position, we find that we would have the same sort of motivation” (123, Slote’s précis of Smith’s position).

Smith’s descriptive position is failure, and Slote rightly points this out. This is because it cannot account for our approval of the motives of agents who act supererogatorily. Slote writes: “After all, I might view someone else as, say, more forgiving than I would be about some matter, yet this need not prevent me from approving of the person whom I view as different from myself (and perhaps as more admirable than myself” (123).

Slote thinks that Smith is correct focusing on properties of agents rather than of potential effects as being that which engenders moral approval or disapproval, and offers a variation of Smith’s position that (Slote) thinks is immune from counterexamples like the one he mentions. Slote (importantly) writes:

“Approval and disapproval of actions and their agents, I want to say, involve (the sentiment and mechanism of) empathy, but to put the matter somewhat too baldly, the empathy involved here is empathy with the agent’s empathy or lack of it.” (124).”

Slote explicates this suggestion more carefully in the subsequent paragraph:

“When I, as judge or non-agent observer, empathetically feel the warmth of an agent as displayed in a given action, then the derivative or reflecting warmth that I feel is a (morally non-judging) feeling of approval toward the action or its agent qua doer of that action; and, similarly, when the agent’s actions display an absence of warmth/tenderness, my observer empathy will register or reflect with the contrast with agentive warmth as a cold feeling or (as we say) “chill” of disapproval (124).

Slote’s own position is, I think, subject to just as serious problems as is the view he rejected of Smith’s. I’ll propose a counterexample to Slote:

What form would such a counterexample take?

Consider that Slote holds the following two conditionals:

(i) If an observing agent empathetically feels the warmth of (another) agent as displayed in an action A, then the felt warmth is a feeling of approval toward the action or its agent qua doer of that action.
(ii) When the (observed) agent’s action A displays an absence of warmth/tenderness, my observer empathy will register or reflect a cold feeling or “chill” of disapproval.


A counterexample case to the first conditional would be one in which the antecedent is granted, and neither of the disjuncts in the consequent hold. And so, a counterexample would be one in which an observing agent empathetically feels the warmth of another agent as displayed in an action A, and yet, that felt warmth is neither a feeling of approval toward the action or its agent qua doer of that action.

A counterexample case to the second conditional would be one in which the antecedent is granted (i.e. the observed agent’s action A displays an absence of warmth/tenderness), and yet, it is not the case that (as the consequent claims), my observer empathy will register or reflect a cold feeling or “chill” of disapproval.

For the sake of clarity, I will offer a counterexample first to the second of Slote’s two conditionals. I’ll then use a similar strategy to (more briefly) offer a counterexample to the first conditional. I then offer what I think is a general “cookbook” recipe for generating counterexamples to the conditionals Slote is defending.

And so, first, a counterexample to Slote’s second conditional.

Stabber Hans and the Closing Elevator

Hans, upon careful reflection, determines the following: “I shall attempt, with my knife, to stab the face of an innocent human being.” He determines, also, that he will attempt this plan only once, and that regardless of the result of his attempt, he will never attempt to execute it again.

Hans informs you of his plan, and when you decry it as “heinous and shameful,” he ties you up and stuffs you into a janitorial closet at a nearby University. He says he will not kill you, and that he will release you in three days.

Unbeknownst to Hans, there is a peephole in the closet, which provides you with a view of a portion of a the hallway, and an elevator. Hours later, you see a particularly virtuous classmate, Sissy, walk into the elevator. Sissy, out of routine, presses a button and continues to read a book she is holding. You notice, in horror, that Sissy’s reading her book has left her unaware of an approaching Hans, who is sneaking toward her with a butcher’s knife. He swipes at her, but the elevator door closes just in time to deflect his knife. “Oh well,” Hans says, tossing his knife in a nearby garbage can. “Like I said, that was my one attempt; I’ll never try that again.”

This would be a case, I suspect, in which the observed agent’s action involves a “lack of warmth/tenderness” and yet, it’s not the case that this lack of warmth/tenderness leaves you (the observer) with cold feelings of disapproval. Your sentimental response is not chilled, but rather, a blissful relief. And in fact, Hans’ action (the swiping and missing of Sissy), leaves you delighted to see that (for example) an injustice was not done, and that the goal of his malicious plan did not come to fruition. No matter how greatly an agent’s actions might lack warmth/tenderness, conclusions about an how an observer will respond sentimentally to an that agent’s action cannot be generated without consideration of the effects the observer accepts to be a result of the observed agent’s action.

We can generate a parallel counterexample to the first of Slote’s two conditionals:

Just suppose that Hans, who has undergone counseling and dedicated his life to altruism, decides that he wishes to make amends for his evil-intentioned, unsuccessful attempt on Sissy’s life. Hans decides to give Sissy a bottle of wine and a note of apology. Hans purchases the wine from a man name Franz, who (believing it will be consumed by Hans, whom he loathes) has poisoned it. Franz, tipsy from unpoisoned wine, tells you that he sold Hans poisoned wine. You then learn of Hans’ plans innocent plans to give the wine to Sissy. Franz, worried that you might foil his plan, ties you up and puts you back in the closet with the peep-hole. From it you witness yet another elevator scene: From your peeping view, you witness Hans handing Sissy that bottle of wine and an apology note. Sissy accepts the gift and disappears behind the closing elevator doors.

This revised case, I think, serves as a counterexample to the first of Slote’s conditionals for the reason that the observing agent (you) empathetically feels the warmth of another agent (Hans) as displayed in an action A (for certainly, you are empathetic to what you take to be his empathetic motives), and yet, that felt warmth is neither a feeling of approval toward the action or its agent qua doer of that action. Your approval, rather, is approval of agent qua the agent’s empathic motives in his action, and not approval of the agent qua doer of the action. (And hence, the antecedent in the conditional is granted, and neither disjunct of the consequent holds). To be clear, the only approval felt here is approval of poor Hans’ empathetic intentions in his gift to Sissy; you do not approve, however, of either the act of giving an innocent girl poison wine, nor Hans qua his giving an innocent girl poison wine.

What do these two counterexamples have in common? Both examples are ones in which what has a hand in generating in an observer a feeling of approval or disapproval is, at least partly, outside the domain of control of the observed agent.

In the first counterexample, that Hans missed Sissy with the knife was outside the control of Hans and his non-empathetically warm motive. In the second counterexample, that the wine was poison was beyond the control of Hans and his empathetically warm motive.

Sometimes, the empathy of our motives is not reflected in the outcome of the actions of these motives. We, as observers, are susceptible to approving or disapproving of others empathetic sentiments, and we are also susceptible to approving or disapproving of the results of others’ empathetic (or non-empathetic) sentiments. Some actions (like those in the counterexamples) incorporate mismatchings of agent’s empathetic motive (which invoke approval in an observer) in action coupled with effects that tend to invoke disapproval. Other actions, conversely, incorporate mismatchings of agent’s non-empathetic motives (which invoke disapproval in an observer) in action coupled with effects that tend to invoke approval. In such mismatchings, we have our recipe for counterexample to the position Slote avers.

One thing, I think, that can be learned from this is that no sentimentalist ethic should, when making descriptive claims about human psychology, fail to realize that properties of agents (such as whether we identify them as empathetic) as well as properties of actions (i.e. whether they produce effects of which we approve) are capable of invoking sentiments of approval or disapproval in an observer. Excluding either, I think, is a mistake.