CC BY
16
Эпистемология и философия науки 2021. Т. 58. № 4. С. 38-45 УДК 167.7
Epistemology & Philosophy of Science 2021, vol. 58, no. 4, pp. 38-45 DOI: https://doi.org/10.5840/eps202158457
O
N VERITISM. PRITCHARD'S DEFENSE
Ernest Sosa - PhD, Board of Governors Professor of Philosophy. Rutgers University. 57 US Highway 1 New Brunswick, NJ 08901-8554, USA; e-mail: ernest.sosa@ rutgers.edu
This time Pritchard is on a rescue mission. Veritism is besieged and he rises to defend it. I do agree with much in his Veritism, but I demur when he adds: "So, the goodness of all epistemic goods is understood instrumental^ with regard to whether they promote truth". If Big Brother brainwashes us to believe the full contents of The Encyclopedia Britannica, then even if we suppose those contents to be true without exception, that would not make what they do an unalloyed good thing, not even epistemi-cally. But it does seem to promote truth. What might then diminish Big Brother's action so much, so as to make it so deplorable epistemically after all, despite how powerfully it does instrumen-tally promote truth. At a minimum we need to say more about the relations between epistemic goods and truth, so as to better understand how it is that the epistemic good is made so good by what specific relation to the truth. I lay out a way to understand Veritism so that it can say more about the relations between epistemic goods and truth, thus enhancing our understanding of epistemic normativity. And in a second part I lay out a solution to Linda Zagzebski's Swamping problem for reliabilism. I argue that it is a problem for process reliabilism, but not for a virtue episte-mology that accepts a kind of reliabilism, but in an agential telic framework, and not in a process framework. So, I lay out one way to be a "veritist", by defending explicitly its Axiological side, and by implication its Conceptual side as well. I have raised questions for Pritchard's own defense and have offered in each case an alternative defense that I believe fits the words of his formulations, and is in their spirit as well.
Keywords: reliabilism, virtue epistemology, telic normativity, the Swamping problem, epistemic goods
О
ВЕРИТИЗМЕ: ЗАЩИТА Д. ПРИТЧАРДА
Эрнест Соса - доктор философии, профессор. Ратгерский университет. 57 US Highway 1 New Brunswick, NJ 08901-8554, США; e-mail: ernest.sosa@ rutgers.edu
В этот раз Д. Притчард занят спасательной операцией: вери-тизм в осаде, и он встает на его защиту. Я во многом согласен с его версией веритизма, но меня смущает утверждение, что ценность всех эпистемических благ трактуется в инструментальном ключе. Так, если Большой Брат внушит безоговорочную веру в полное содержание Британской энциклопедии, то даже если мы считаем это содержание истинным без исключений, результат не будет безусловным благом с эпи-стемической точки зрения. В чем же тогда проблема с внушением Большого Брата? Почему его ценность так низка с эпистемической точки зрения, несмотря на то, что оно приближает нас к истине? Я утверждаю, что, хотя в гносеологии ценность истинного суждения является фундаментальной, можно выделить аспекты утверждений, которые оцениваются независимо. В статье ценность попытки вообще сделать что-то оценивается с точки зрения трех соображений: 1) успех или неудача; 2) компетентность лучше, чем некомпетентность; 3) успех благодаря компетентности лучше, чем успех
38
© Ernest Sosa
благодаря удаче. Подобную структуру можно распространить на оценку убеждений. Например, для участника телевизионной викторины удачная догадка будет приемлемой, а для пациента офтальмолога, которому нужно прочитать буквы на оптометрической таблице, удачная догадка не приемлема. В последнем случае скорее будет важным то, насколько правильное суждение зависит от зрительных способностей пациента. Я отмечаю, что алетическое утверждение (т.е. истинность благодаря способности) есть разновидность истины. Во второй части статьи я предлагаю собственное решение проблемы поглощения (swamping problem), которая состоит в том, что ценность истины поглощает всякую иную эпистеми-ческую ценность. Я считаю, что эта проблема имеет значение только для процессуального релайбилизма, а не для агентного телеологического релайбилизма, который защищается в статье. С точки зрения процессуального релайбилизма убеждение рассматривается как продукт, который отличается от производящего его процесса. Так же, как ценность чашки кофе не зависит от того, каким образом она изготовлена, ценность убеждения не зависит от того, как оно сформировано. В статье утверждается, что аналогия «процесс - продукт» в отношении убеждений субъекта является ложной. Убеждения, представления, суждения субъекта неотделимы от самого субъекта. Их следует скорее рассматривать как эпистемиче-ские акты, проявления агентности познающего. Таким образом, в статье критически рассмотрены аргументы Притчарда и предложена альтернативная версия веритизма. Ключевые слова: релайбилизм, эпистемология добродетелей, телеологическая нормативность, проблема поглощения, эпистеми-ческое благо
Duncan Pritchard's crisp insights pose questions of epistemic value well worth pondering. The main theses of his paper are made clear already in its opening paragraphs:
It used to be taken as obvious that truth is central to epistemology. The kind of centrality in question is a form of fundamentally that is both conceptual and axiological. The former, in that we understand what it is for something to be epistemic in terms of its relationship to the truth. So what makes doxastic justification an epistemic property of a belief, for example, is because it is truth - conducive. Justified beliefs have a propensity to be true (in contrast to unjustified beliefs), and thereby stand in a positive relationship to the truth, even when the token belief is false. The latter, in the sense that the truth is the fundamental good of epistemic appraisal, in the way that the beautiful might be thought to be the fundamental good of aesthetic appraisal. So the goodness of all epistemic goods is understood instrumentally with regard to whether they promote truth. This axiological claim is intertwined with the conceptual point. What makes epistemic justification epistemic is that it is truth-conducive, which is also what makes epistemic justification valuable, from a purely epistemic point of view.
In principle, at least, one could endorse the axiological claim without thereby endorsing the conceptual claim, and vice versa. But it is hard to see what the motivation for such a detachment of these twin theses could be. How could truth be the fundamental epistemic good without it being at the same time conceptually fundamental to the epistemic realm as well? Conversely, why would one hold that truth is conceptually fundamental to the epistemic realm if one didn't also maintain that truth is the fundamental epistemic good? Accordingly, in what follows we will treat these two theses as being two sides of the same coin.
Pritchard's Defense of Veritism
1. This time Pritchard is on a rescue mission. Veritism is besieged and he rises to defend both of its components, the conceptual and the axio-logical.
Veritism
Axiological Veritism "Truth is the fundamental good of epistemic appraisal".
Conceptual Veristism "Truth is conceptually fundamental to the epis-temic realm".
I do agree with such Veritism, but I demur when Pritchard adds: "So, the goodness of all epistemic goods is understood instrumentally with regard to whether they promote truth".
If Big Brother brainwashes us to believe the full contents of The Encyclopedia Britannica, then even if we suppose those contents to be true without exception, that would not make what they do an unalloyed good thing, not even epistemically. But it does seem to promote truth. What might then diminish Big Brother's action so much, then, so as to make it so deplorable epistemically after all, despite how powerfully it does in-strumentally promote truth.
At a minimum we need to say more about the relations between epis-temic goods and truth, so as to better understand how it is that the epis-temic good is made so good by what specific relation to the truth.
In what follows I hope to lay out a way to understand Veritism so that it can say more about the relations between epistemic goods and truth, thus enhancing our understanding of epistemic normativity.
2. There is, I suggest, a normativity of attempts as attempts, such that Oswald's attempt to shoot JFK counts as excellent and apt, as a shot, despite how deplorable an action it constitutes, that awful assas -sination. This is a distinctive sort of normativity, with simple principles as follows.
3. Normativity of Attempts as Attempts a. Success is better than failure.
b. Competence is better than incompetence. (Both for attempts and for agents.)
c. Success through competence is better than success by sheer luck. (Success through competence C is success that manifests competence C.)
4. Such telic normativity is applicable to epistemic attempts in particular if we consider the nature of judgment.
a. In the domain of epistemology, and more specifically of gnoseology (theory of knowledge), the notion of alethic affirmation has a fundamental place. But alethic affirmation is compatible with guessing. The sheer guess of the contestant in a prize show is an alethic affirmation. The contestant affirms aiming to get it right, as that is how to win the prize. As eye-exam patients we are also guessing at the row way down on the Snellen Chart, even while trying to get it right based on whatever competence we can muster through our visual input. As we ascend up the rows of the eye-exam chart, we rise above such guessing. What is relevantly added, I suggest, is an increase in confidence that supports the guiding thought that one would attain success if one affirmed alethically. So, this, I say, is what judgment requires.
b. That being so, then judgment aims not just for truth, as does the guess of the contestant, and even above the more substantial vision-test guess at the lower rows of the Snellen Chart. Judgment is rather an alethic affirmation that aims at aptness of alethic affirmation. At the low rows you think you're just guessing because you have very little confidence. But your confidence, on a question whether p, aims to track your epis-temic qualification to get it right on that question, such as the strength of your total pertinent evidence. So, low confidence aligns with one's view that one lacks sufficient reliability, sufficient likelihood of getting it right. As one ascends through the rows, confidence increases, along with increasing willingness to adopt the guiding thought that one would be likely enough to get it right. This rises until one is willing to actually attempt to get it right in the way recommended by the guiding thought, by affirming alethically in the endeavor to get it right aptly. Interestingly, such judgment aligns with the Hippocratic oath that one will treat one's patients ".. .to the best of one's judgment and ability".
5. Judgment thus turns out to be an attempt, an attempt to get it right aptly through an alethic affirmation, or in other words an attempt to get it right through competence rather than luck. For an alethic affirmation, getting it right amounts to truth. So, judgment turns out to be an attempt to attain truth through affirmation, but not just truth. Bare truth is enough
for the guess of the contestant, but not for the judgment of an oncologist. Judgment aims for more than mere truth. It aims also for aptness of affirmation. But aptness of alethic affirmation is a kind of truth; it is truth through competence.
6. Given that judgment turns out to be an attempt, it thereby falls under the telic normativity of attempts with its three principles laid out above as 3a-3c. And this yields the desired defense of three epistemic values: that of truth, that of epistemic competence (adroitness, justification), and that of knowledge (both that of animal knowledge, or bare aptness of alethic affirmation, and also that of reflective affirmation full well, or aptness of judgment).
Thus we can also see a clear way in which truth has a fundamental role in epistemic "axiology", or in epistemic "normativity". (I prefer the latter term because it avoids the misleading overtones of supra-telic value that attach to 'axiology'.)
On Swamping
7. If the operation of a coffee machine is to be prized, by baristas and consumers, just in virtue of its "instrumental" promotion of good coffee, it seems plausible that delicious coffee will have no more value if produced by an efficient, reliable machine than if produced by an inefficient, unreliable machine. Delicious coffee is the fundamental value of the realm of coffee and so it is plausible that coffee is not made better when produced reliably than when produced unreliably. Analogously, if the operation of a coffee machine is prized by baristas and consumers just in virtue of its "instrumental" promotion of good coffee, it seems plausible that delicious coffee will have no more value if produced by an efficient, reliable machine than if produced by an inefficient, unreliable machine. if epistemic justification has distinctively epistemic value just in virtue of its "instrumental" promotion of truth, this makes it plausible that a true belief has no more value if justified than if unjustified. This is Linda Zagzebski's swamping problem for reliabilism.
Pritchard replies in two ways. He argues, first, that the justification of a true belief will automatically import more true beliefs in support of that true belief, and so when justified a true belief will import more epistemic value thereby, by importing more truth. Unfortunately, it does not seem correct that the justification of a true belief need thus import more justified true beliefs. Foundational justification does not seem to require any basis of true justified basis. Moreover, a BIV belief that is highly justified on an elaborate perceptual basis may be true accidentally, no thanks to the justified basis, so that its justification imports no further true
beliefs. (These are at least prima facie issues that would need to be addressed.)
Pritchard's second way of addressing the swamping problem appeals to extra-epistemic value. But the objection can be focused on epistemic value specifically, and would remain about as problematic.
So, here now is a response to the swamping problem that defends Veritism in a different way, while drawing once again on telic virtue epis-temology. And it will also emerge how reliabilist virtue epistemology contrasts with process reliabilism in its ability to deal with this problem.
8. According to generic process reliabilism, a belief is justified to the extent that its source process reliably enough yields belief that is true. A belief is here viewed as a product quite distinct from the process that produces that product. Thus, a belief can be viewed as a sort of map, one separable from the "cartographic" process that yields it.
The problem for such reliabilism is raised by Zagzebski through her analogy with a good espresso produced by an espresso machine. The quality of that coffee is determined by how tasty it is, which is independent from the reliability of the process that produces it. A terrible espresso machine that normally produces undrinkable coffee may on occasion produce a delicious espresso, whose evaluation is hence unaffected by the reliability score of that process.
Similarly, goes the objection, the relevant epistemic quality of a belief is unaffected by the truth-reliability score of the epistemic process productive of that belief. The coffee-pertinent quality of a cup of coffee does not depend on the reliability of whatever process produces it. By analogy, then, the knowledge-pertinent quality of a belief does not depend on the reliability of whatever process produces that belief. Reliabil-ism is thus said to fail as an account of the knowledge-pertinent epistemic quality of beliefs.
Suppose we think of beliefs as maps that help us steer. If we wish to go to Larissa, an accurate map will serve us well. And the accuracy of a map could reach its highest level independently of the quality of the cartographic process that produces it. A map could reach the highest accuracy accidentally if the cartographer is just guessing. So, reliability seems as inessential to the accuracy of a map as it is to the quality of a cup of espresso. We can get to Larissa just as well by means of an accurate map that is unreliably produced as by an accurate map that is reliably produced. All that matters to the quality of a map as a guide to action is its degree of accuracy. What mainly matters is that it be accurate enough to guide us well enough to our destination.
Plato's Meno problem can be put thus in terms of beliefs as maps, but the problem clearly extends to beliefs as representations more generally.
And so the problem extends to affirmations generally, whether these take the form of vocalizations or that of inscriptions, or that of subconscious representations. Considered simply as maps to guide action, our
affirmations are assessable without regard to our reliability as cartographers. If accurate enough, a map is a good-enough map, regardless of how reliably it came to be that way, in line with Plato's point in the Meno.
Accuracy is the pertinent epistemic quality of maps, and the same would then seem to be true of beliefs considered simply as guides to steer by. The problem remains if we take the map to be a complex set of dispositions to guide behavior given a set of desires. If we think of that set of dispositions as just lodged in the agent however it may have got there, we will have another version of the same problem.
We thus have an ostensible problem for any account of a belief's justification in terms of how truth-reliable is the process that produces that belief. This is still a problem if we think of the justification of that belief in terms of how reliably it is put in place so as to guide action well. And it remains a problem even if we think of the pertinent desideratum for that belief as its degree of accuracy, never mind its navigational value. We can focus on that quality of a map, even when it is a map that locates Timbuktu, and we have no plans to go there, nor will or would we ever go there, nor to any other place located on that map. It can still count as a fine map because of its degree of accuracy. But this too is compatible with its having been produced by a terrible cartographic process. So, the analogy with the good espresso remains, and Zagzebski's objection is sustained.
9. In contrast to process reliabilism, it bears emphasis here that telic virtue epistemology takes a different view of judgment and representation, and of belief more generally, whether judgmental or functional, conscious or subconscious. Our virtue epistemology takes assertive representations, judgments, and beliefs to be not maps but attempts1. So, the right analogy is not to a product separable from the agent. In our view, the right analogy is rather to the producing, to the agent's attempt to get it right on a given question. Because of that, the reliabilism of telic virtue epistemology is not the reliabilism of process reliabilism. The telic focus
1 The postulation of subconscious mental acts just comes along with the postulation of implicit reasoning in the explanation of much human and other animal conduct. Suppose we postulate such reasoning in explaining why someone reaches for a glass of water even when they engage consciously in no such reasoning, when instead they just "automatically" reach for the glass that is obviously in their field of vision. Such explanation would seem to require the agent's occurrent acceptance of premises that they do not accept consciously, and this acceptance would seem to count as an act, even while remaining subconscious. Out of the plethora of declarative "sentences" or representations stored in their "belief box", that particular one is then "activated" in an act-like way, so as to function as a premise of their (subconscious) reasoning on that occasion. And the like would seem to be required for folk explanation in animal ethology. (But this is just to hint briefly at a way of thinking about subconscious mental acts and their place in our rational economy.)
is rather on agency, not on process. This enables us to reject the damaging analogies to maps and cups of coffee and their respective sorts of quality. The quality that matters to us is the quality of action, and thereby the quality of a very distinctive sort of products of agency.
Our focus is on actions, whether of praxis or of episteme. This imports a normativity that is telic, and undetachable from the agent and their agency. For virtue epistemology, the epistemic normativity of belief is not detachable from its adroitness, from the degree of reliability of the competence that it manifests. Such normativity of belief is hence not detachable from reliability, in contrast to how the quality of a cup of coffee is detachable from the coffee machine and its operation. Our focus is the aimed 0ing of the agent who 0s, the competence that is exercised in that 0ing, the success of the 0ing, the aptness of that success, etc.
So, the relevant analogues of the judgments and beliefs of our telic epistemology are not cups of coffee, nor maps. The relevant analogues would be performances of the barista or the cartographer, or even performances of the map-user who plans an itinerary. The focus is not just on the map but on the cartographer's epistemic performance, and on the epistemic use of the map, whether consciously intentional and deliberative, or subconscious and functional.
The swamping problem turns out accordingly to be a problem not for reliabilism but for process reliabilism. The telic normativity of virtue reliabilism invokes not just processes but exercises of agency, with a focus on epistemic competence and its manifestations, which makes the swamping objection inapplicable to virtue epistemology.
So, I have laid out one way to be a "veritist", by defending explicitly its Axiological side, and by implication its Conceptual side as well. I have raised questions for Pritchard's own defense and have offered in each case an alternative defense that I believe fits the words of his formulations, and is in their spirit as well.
