Augustine’s Cogito

The one thing people are sure to know about Descartes—who know anything about him at all—is that he said (approximately), “I think, therefore I am.”

Therefore, it is ironic that Descartes was not the first to say this. Consider the following:

…I am most certain that I am and that I know and delight in this. In respect of these truths, I am not at all afraid of the arguments of the Academicians [i.e., skeptic philosophers], who say, “What if you are deceived?” For if I am deceived, I am. For he who is not, cannot be deceived; and if I am deceived, by this same token, I am. And since I am if I am deceived, how am I deceived in believing that I am? for it is certain that I am if I am deceived. Since, therefore, I, the person deceived, should be, even if I were deceived, certainly I am not deceived in this knowledge that I am.

This passage is from St. Augustine’s City of God (XI.26). The reference to the Academics is significant. Augustine was evidently aware of philosophical skepticism and concerned to answer it at least in some respects—just as was Descartes. Moreover, Augustine contrasts the frailty of sense-perception as a means of knowledge with the certainty of “another and far superior sense, belonging to the inner man” (XI.27)—just as does Descartes. Immediately preceding the passage quoted above, Augustine asserts that our knowledge of our own existence is not subject to illusion just because it is not derived from the bodily senses, which only produce in us images resembling sensible objects, not the objects themselves. Intellectual knowledge, such as the knowledge we may demonstrate of our own existence, is free of “any delusive representation of images or phantasms.”

Clearly, Augustine’s cogito was not merely a transient thought which he failed to fully recognize or develop. According to Roy Sorensen (A Brief History of the Paradox, Oxford U.P., 2003, p. 170), Augustine presents the argument no less than seven times in assorted works. (Some other references I’ve found are The Trinity (10.10.14) and the Enchiridion (7.20).) Of course, Augustine does not employ his cogito as the foundational argument for an elaborate epistemology and system of knowledge. Still, in view of the fame the cogito argument acquired with Descartes, it seems odd that Augustine’s prior claim is so little known.

Did Descartes rip off Augustine? He claims not. However, Sorensen is not convinced, pointing out the ubiquity of the cogito in Augustine’s writings and the popularity of Augustine with Descartes’s Jesuit teachers at La Flèche.

In any event, Descartes employed his cogito in one important way that Augustine did not. Specifically, he thought it demonstrated the prior certainty of consciousness, by which I mean loosely that the mind is better known than the body. For example, a visual experience as of an object—a cup, say—in front of me may or may not really be of a cup, but it is unquestionably an experience. Descartes realized that only “thought,” by which he meant to include any conscious event, could support his argument. It wouldn’t do, for instance, to say, “I walk, therefore I am,” even though walking supports the logic of the argument, as presented by Augustine above, just as well as being deceived. Try substituting “walk” for “be deceived” in the passage above. Except where being deceived refers specifically to being deceived about one’s own existence, as opposed to being deceived in general, the logic of Augustine’s argument remains undisturbed. Thus, the fact that the event in question is being deceived appears to be almost incidental in Augustine’s cogito. But it isn’t; only an event involving conscious “thought” will support the argument. And this implies that the reality of thought is more securely knowable than the reality of the objects of thought. This is a point that Augustine may have missed, even though, with his talk of images and phantasms, he had the conceptual tools that would have enabled him to do so.

The prior certainty of consciousness, after going more or less unchallenged for nearly 300 years after Descartes, finally came to be controversial and to be rejected by many philosophers. But I am inclined to think it is true and that the cogito provides a pretty good argument for it. (Of course, the further uses Descartes made of the prior certainty of consciousness, for instance as support for mind–body dualism, are another matter entirely.) In this respect, Descartes saw more deeply into the cogito, even if he did not invent it.

34 thoughts on “Augustine’s Cogito

  1. Nice post. I suspect Descartes’ claim is better known because it played a more important role in his thought and he has become “the founder” of “modern” philosophy in many narratives of its history. Augustine, by contrast, has only recently come to be taken seriously again as a philosopher outside of Christian circles, and his version of the argument does not have the same foundational role in his thinking that Descartes’ does in his. It may be worth noting, though, that Augustine’s anticipation of Descartes has been well known by scholars, at least, since soon after he wrote the Discourse on Method; Mersenne noted it in his correspondence, and Descartes downplayed its significance. There is a nice, brief treatment of this in Brian Stock’s Augustine’s Inner Dialogue: the Philosophical Soliloquy in Late Antiquity, chapter 2, “Soliloquy and Self-Existence.”

    To my mind, one of the virtues of Augustine’s version of the argument over Descartes’ is that Augustine does not attempt to treat it as a foundational principle on which to ground all of our knowledge. He accepts that we cannot have the same degree of certainty about the objects of our experience or the truth of our beliefs about the world as we can about the fact that we have experiences and beliefs. We don’t find Augustine trying to argue from certainty of his own existence to certainty of the existence of God to certainty of the external world, or anything like that. Hence he doesn’t end up in the egocentric predicament, supposing that he never experiences external objects directly but only the ideas that those external objects cause him to have. I’m no Augustine scholar, but in general my sense is that he would affirm that the objects of consciousness are prior to our consciousness of them, even if we can be wildly mistaken about what those objects are. So perhaps it would be better to say not that Augustine affirms “the prior certainty of consciousness,” but “the greater certainty of consciousness”?

    As I see it, the trouble with Descartes’ whole approach, and the reason why Augustine is often more philosophically interesting and insightful, is summarized well by Alasdair MacIntyre:

    “Without sufficient grounds for belief in God Descartes takes himself to lack sufficient grounds for belief in anything at all outside his mind, including his own body. The victory over the skeptics accomplished by the Cogito is too limited a victory, and Descartes has not discovered a foundation for the natural sciences that is immune to skeptical doubt. Should this matter to us as well as to Descartes?

    The answer is “No,” and this for two reasons. The first is that skepticism does not require an answer. The skeptic’s claim is that our beliefs and judgments are open to doubt, because it is always possible that we may be mistaken in believing as we do and in judging as we do, because there is always the possibility of error. But to point out — quite correctly — that it is possible that we are in error gives us no reason whatsoever to believe that we are in fact in error. And, until we are given such a reason, we have no reason whatsoever to doubt what we otherwise have good reason to believe and judge. The skeptic’s claim that we can only truly say, “I know that such and such,” or “I am certain that such and such,” if there is no possibility at all of our being in error about how things are, is a claim that the skeptic has given us no good reason to accept. It follows that the natural sciences do not need the kind of foundation that Descartes aspired to provide for them.

    That this is so becomes even clearer when we consider what kind of foundation a natural science does need and can have. Every science aims at the achievement of a perfected understanding of its particular subject matter. To achieve such a perfected understanding is to make some set of phenomena — tides, thunderstorms, glaciers, the firing of neurons in the brain, the production of chemical reactions in the bloodstream, falling rates of economic growth, or rising rates of violent crime — intelligible and explicable as the outcome of whatever are the fundamental determinants of those phenomena. The concepts and the generalizations through which we identify those fundamental determinants and the complexities of their relationships to particular phenomena are what provide each science with its foundations.

    However, what we have learned from the history of science — something that neither Aristotle nor Descartes were in a position to learn — is that over time in the course of our scientific enquiries our conception of what it would be to achieve a perfected understanding of this or that set of phenomena changes. We find that we have good reason to reject or revise, sometimes radically, our earlier accounts of what the foundations, the first principles, of this particular science are. But, until and unless we find that we have good reason to do so, we have no reason to put in question our present understanding of those foundations. The knowledge that we may later on need to reject or to revise — the knowledge, that is, of our own fallibility — of itself gives us no reason to reject or revise.” (God, Philosophy, Universities, 117-8.

    I would add that it is very hard to see what reasons anyone could offer us for global skepticism that are in fact consistent with skepticism. In this respect, it’s interesting that the ancient Greek skeptics generally didn’t argue for global skepticism as such. They took up various philosophical issues in turn and tried to show that arguments of equal strength could be given on each side. Of course, famously one of the questions they took up was whether there are any sense impressions that are immune to doubt, and they gave some powerful arguments to think not (presented in the extant portions of Cicero’s Academica). But they didn’t quite embrace skepticism, at least if skepticism is the claim that we do not know anything. The early Academic skeptics, and (probably) the Pyrrhonists later on, claimed simply to find no adequate grounds for believing that we know anything, and hence suspended judgment; later Academics (again, probably — there’s some interpretive dispute about this) instead adopted something like fallibilism, and so allowed themselves to hold what seemed to be the most rationally defensible beliefs so long as they were held with due recognition of one’s fallibility and an openness to further inquiry.

    Sorry if this is all familiar territory to you, but I figured it might be of some interest to other readers, who are all no doubt starving for content while Irfan is off eating hummus.

    In any case, nice post again, as usual!

    Liked by 1 person

    • Oh, so you’re saying that hummus isn’t content? Hummusphobe. Will get back to discussing the prior certainty of hummus to consciousness and other deep thoughts after I catch up on some unfinished (unembarked-upon) work.


    • Hi David. Thanks for your interesting reply. I pretty much agree with everything you say here. I just have a question and a remark.

      The question concerns ancient skepticism. My knowledge of it is probably, if not rusty, at least not up to the minute. I was introduced to it in a course on hellenistic philosophy taught by Alex Mourelatos at UT Austin, by the way. Since you’ve mentioned UT in a comment someplace or other, you may know him. Anyway, my understanding is that the Pyrrhonians did advocate what could be called global skepticism inasmuch as they held that no beliefs are justifiable and that one should accordingly abjure all belief. I think what you meant by “global skepticism” was probably something a little different, namely the view that all knowledge claims can be brought down in one fell swoop by a single argument of some kind, such as the appeal to an evil demon. And this seems right to me. The Pyrrhonians—or Sextus, anyway—argues piecemeal, patiently going through claim after claim and arguing that it is no better supported than other, opposed claims. Nevertheless, on the reading I’m talking about, the Pyrrhonians wouldn’t be just fallibilists. They would be saying that the epochê means that one does not hold any proposition to be true or false at all, even provisionally. One reason in favor of this reading is that it makes better sense of the claim that the epochê leads to ataraxia. Such a claim seems to require something more radical than mere fallibilism. Most people—outside Stoicism—make no pretensions to certainty about much of anything. So if giving up certainty were all there were to it, there would be nothing distinctive about Pyrrhonism. But Pyrrhonism is supposed to be distinctive way of life.

      So the question is, does this match your own understanding? In reading the Pyrrhonians the way I do, I admit to being under the influence of Burnyeat’s old paper (which I read in Mourelatos’s class), “Can the Sceptic Live His Scepticism?” That paper is 35 years old now. So maybe there has been new thinking on the topic.

      The remark concerns MacIntyre’s discussion of skepticism. It is good to point out, as MacIntyre does, that the bare possibility of error is not a positive reason to suppose that there has actually been one. And this is relevant because a great deal of modern discussion of skepticism argues from the bare possibility that S is false directly to the conclusion that one does not know that S. So if there is any possible world in which S is false, then S is not known. Those who argue this way rely on the “intuition” that our concept of knowledge requires the exclusion of any possible way that a known belief could turn out to be false. There is an enormous literature that takes this intuition for granted. David Lewis’s paper “Elusive Knowledge,” for example, responds to and participates in this literature. But the intuition has always struck me as ridiculous. Just because if S is known, then S is true, it doesn’t follow that S must necessarily be true, which is what the intuition would require. What knowledge requires is a cognitive channel between subject and object, such that the subject is aware of what is known. Thus what is known is true. But this does not require certainty—it does not require that cognition employ methods that never fail to open a channel of awareness or never appear to have opened such a channel when they really haven’t. Compare seeing: What seeing, say a pen on the desk before me, requires is a visual channel between me and the pen, such that I am visually aware of the pen. And for me to see the pen, it must be there. But this does not require that that my visual system never fail to open a visual channel or never appear to have opened such a channel when it really hasn’t (as in certain sorts of visual illusion or hallucination). The situation with knowledge is precisely analogous.

      But that’s not the remark. The remark is that I don’t think it is legitimate to dismiss a claim on the ground that nothing positive has been said for it, as MacIntyre implies we should do. It might seem to be a sound logical principle that “the burden of proof is on him who asserts the positive,” and as a debating point I would agree. But the enterprise of knowledge isn’t, or shouldn’t be, a debate club. The individual knower, in my view, has a responsibility to assess and rate the rational credibility of any proposition that is brought forward. This includes noxious skeptic warhorses such as that one is actually dreaming right now, that one is a brain in a vat, that one is being deceived by an evil demon, etc. etc. It’s not as though such claims are unintelligible or that we have no knowledge that bears on their epistemic probability of being true.

      Take the evil demon. We can and should evaluate the epistemic status of this hypothesis using the same tools of scientific reason that we would use for any other hypothesis. What are the chances that there are any demons, evil or otherwise? Empirical evidence for them is very slight. Then again, this demon would have to take peculiar interest in me, and what are the chances of that? How did I get to be special? And why would the demon care, about me or anybody else for that matter? Moreover, the idea that my experiences or judgments are really the product of a demon is a “piggyback” theory—a “theory” that says there is a special, hidden mechanism, for which there is no independent evidence, that makes our conventional theory of the world seem to be true. Put it this way: Let all our conventional beliefs about reality be conjoined as T; a piggyback theory says that T is radically false (all or nearly all of its conjuncts are false), but it nevertheless seems true because of some special entity G, for which there is no independent evidence, that creates the illusion. Global skeptical hypotheses like the evil demon or the brain in a vat are piggyback theories. (They aren’t the only kind; but that’s another topic.) But there is an infinite number of such theories, which can be generated by inventing new candidates for G: vat, demon, flying spaghetti monster, etc. And each one has the same epistemic probability, since it has the same evidence (and makes precisely the same predictions)—namely, just the same evidence (and predictions) as for T itself. Now, since there is no independent reason to believe in any of the G, G’, G”, G”’, …, all of the T+G, T+G’, …, must be less probable than T alone. For, the stronger hypothesis is less well supported than the weaker by the same evidence, all else being equal. And so in simple Bayesian terms, in a short time all the T+G, T+G’, …, will quickly have their epistemic probabilities driven to near 0.

      I could go on, but the point is that by the standards of ordinary scientific reasoning, skeptic hypotheses have quite low epistemic probabilities. And in general my approach to skepticism is to examine skeptical hypotheses on their merits, not to look for self-referential contradictions, “genetic” discrepancies (e.g., use of or dependence on concepts that would be invalidated if the skeptic’s thesis were true), or other means of dismissing skeptic hypotheses rather than evaluating them. There is a temptation to think that skepticism must somehow be guilty of some sort of tail-swallowing incoherence, and no doubt it is possible to find statements of skepticism that are indeed thus guilty. But I don’t think that global skepticism is necessarily guilty of such errors. A carefully formulated skepticism will always avoid them. The best defense against skepticism is a rational epistemology, which will show that skeptic hypotheses have low epistemic probabilities, and global skeptic hypotheses have vanishingly small epistemic probabilities.


      • Well, I’m really late to this exceptionally interesting discussion, but I think I owe you a brief response. I haven’t read the most recent stuff on Pyrrhonism, and maybe if I did I’d change my mind, but I’m pretty much with you on your understanding of it here. The Pyrrhonists really are global skeptics in the sense you use that term; it’s the Academics who, in at least some of their guises, are closer to fallibilists. But even the Pyrrhonists seem to me to differ somewhat from more familiar modern skeptics by virtue of distinguishing between belief (doxa) and appearance (phantasia). However we answer the question posed in Burnyeat’s title, it’s at least plausible to think that we can get around in the world without beliefs if there are such things as appearances that are (i) distinct from beliefs and (ii) such that we can act in accordance with them without thereby forming beliefs. I don’t know the contemporary literature on skepticism well at all, but I don’t think Descartes or Hume operate with a distinction of this sort. Hume, for instance, seems to think that we do have to form beliefs in order to get around in the world, though we might not need to form any distinctively philosophical beliefs; for Hume, skepticism is (at least sometimes) something that we adopt in our reflective moments and that might impact our action insofar as it prevents us from acting on the basis of dogmatic philosophical beliefs, but it’s not something that leads us to walk around in the world without believing that the chairs we are trying to avoid bumping into are there to be bumped into. I don’t really find Pyrrhonism much more interesting than Humean skepticism, but the different notions of belief operative in each case do seem to make for an interesting difference.

        As for your more general points about skepticism, I can’t do them justice now and it’s awfully late for a real response anyway, but it strikes me that the considerations you adduce in favor of thinking that skeptical hypotheses have low epistemic probabilities are considerations that genuine skeptical doubt wouldn’t license. If we begin by taking seriously the possibility that all our experience and ordinary beliefs are radically and systematically mistaken, then we have no reason to trust our assessment of the likelihood of things like demons existing or being interested in deceiving us, since that assessment is based on empirical considerations, and those are, ex hypothesi, systematically false. In other words, skepticism isn’t an empirical hypothesis, and so it doesn’t make sense to assess it as though it were one. There might be some local skeptical problems that can be treated that way, but global skepticism of the Cartesian variety isn’t one of them.

        Or so it seems to me. I might be mistaken. In any case don’t let my month-long silence or my pathetically inadequate response suggest that I took your thoughts to be not worth responding to!

        Oh, and I do indeed know Alex Mourelatos. He was one of my mentors in graduate school. I didn’t end up writing anything with him, but I certainly learned a whole lot from him.


        • Hi David. Don’t sweat the “lateness” of your reply. I’m glad to have discussion of this sort of thing any time. I think the main point you raise is interesting:

          it strikes me that the considerations you adduce in favor of thinking that skeptical hypotheses have low epistemic probabilities are considerations that genuine skeptical doubt wouldn’t license. If we begin by taking seriously the possibility that all our experience and ordinary beliefs are radically and systematically mistaken, then we have no reason to trust our assessment of the likelihood of things like demons existing or being interested in deceiving us, since that assessment is based on empirical considerations, and those are, ex hypothesi, systematically false. In other words, skepticism isn’t an empirical hypothesis, and so it doesn’t make sense to assess it as though it were one.

          In other words, couldn’t one reply to my argument against skepticism that it begs the question? Someone proposes that I am being radically deceived by an evil demon, and I reply by citing evidence from my nonskeptic frame of reference, according to which my being radically deceived has a nugatory probability. But what is the value of that? If I am being radically deceived, that evidence is false. What we need to decide between competing hypotheses is evidence or argument that is neutral between them. And since global skepticism encompasses all evidence, there is no such thing as neutral evidence between skeptic and nonskeptic hypotheses. This would be why you suggest that skepticism isn’t an empirical hypothesis.

          However, I think the evidence is neutral. Think about the “piggyback theory” argument. Right now I am having sensory experiences as of sitting at a desk, typing at a computer, drinking coffee, listening to music, and so forth. One “theory” of all this is that I am sitting at a desk, typing at a computer, drinking coffee, and listening to music. Another “theory”—the piggyback theory—is that an evil demon is making it seem as though I am sitting at a desk, typing at a computer, drinking coffee, and listening to music, even though none of these things is really happening. The two theories thus offer competing explanations and interpretations of my experiences, which can be neutrally described as between them. And I have argued that the piggyback theory is very poorly supported for several reasons: it contains major elements, such as an evil demon, for which there is no independent evidence; it raises questions about the demon, such as why it should care about me, that have no independently supported answers; and it is suspiciously ad hoc, since no predictive power accrues to the hypothesis of the evil demon, but all predictions stem from hypotheses about the supposedly illusory external world. These problems do not only arise from the nonskeptic perspective. The skeptic’s piggyback theory confronts them even on its own terms.

          My view is that more or less everything is epistemically possible, so no hypothesis should be dismissed out of hand or ruled to be beyond the pale of rational consideration, as some antiskeptics attempt to do. Every hypothesis must be evaluated on its merits in light of whatever evidence may be supposed to support it. The real trouble with skepticism is that by this standard its hypotheses just don’t fare well.


          • Another good reply. I take your point here, but I’m not convinced that this is how skeptics generally do or always must think of their skepticism. Descartes does not propose the evil demon as a satisfactory explanatory theory of his experience, nor do people who employ the dream argument typically propose that the hypothesis that you are dreaming right now makes better sense of your experience than any alternative. If they did that, they wouldn’t be skeptics, they’d just be very weird theorists. Instead, skeptics pose these as examples to show that it is conceivable that all of our experience can be what it is and yet we be radically deceived. We aren’t supposed to think that the scenarios are likely or explanatorily powerful, but only that they are logically possible. The challenge is not, then, to show that, given our experience, something other than the demon or the dream scenarios is more warranted. The challenge is, rather, to show that we are not radically deceived. I don’t think that can be shown, but I don’t think we have any need to show it (I think, in fact, that it is a trivial consequence of metaphysical realism that most of our beliefs could be false, since for metaphysical realists truth is a property of propositions that transcends our evidence for them).

            So I’d agree with you that, construed as you construe it, skepticism fares badly. But your arguments are, in effect, just a more articulate version of my MacIntyrean refusal to take skepticism seriously. You consider the skeptical hypothesis and you conclude that we have no reason to take it seriously; so do I, I just don’t usually bother to explain why an evil demon hypothesis doesn’t deserve to be taken seriously.

            I might be able to get behind your claim that a response like yours is not question-begging, but neutral between skepticism and non-skepticism. But I worry that this still depends on taking skepticism as a positive claim, when in fact skeptics of at least some varieties are rather posing questions that they claim can’t be answered satisfactorily. Granted, they do have to make the positive claim that it really is conceivable that we are radically deceived. Some have tried to deny that (e.g., Putnam as I understand him), but I think that strategy fails. But I think it really is conceivable that we are radically deceived, and yet I am not a skeptic, so that positive claim is not the skeptical one. What makes a skeptic rather is the negative claim that no attempt to show that we are not radically deceived can be (or, more weakly, has been) successful.

            So I suppose my questions for you would be: do you think I am wrong to construe skepticism this way, or that skeptics can’t construe skepticism this way? If so, why must the skeptic treat skeptical scenarios as empirical explanatory hypotheses rather than merely as illustrations of the conceivability that we are radically deceived?


            • So I suppose my questions for you would be: do you think I am wrong to construe skepticism this way, or that skeptics can’t construe skepticism this way? If so, why must the skeptic treat skeptical scenarios as empirical explanatory hypotheses rather than merely as illustrations of the conceivability that we are radically deceived?

              No, I think you’re quite right about how the modern skeptical argument tends to go. The tendency is to argue that there is some alternative way the world might be that would account for all our sense-experiences and yet render false nearly all of our beliefs about the world. For example, one might be a brain in a vat and still have a train of experiences that is subjectively indistinguishable from the experiences one actually has, as of having a body, breathing air, knowing other people, etc. In such a case, one’s beliefs in having a body, breathing air, knowing other people, etc., would all be false. Therefore, since it’s possible that one is wrong about these things, one does not know them.

              So I agree that modern skepticism typically argues from the premise that one might be wrong straight to the conclusion that one does not know. But notice an unstated premise here, namely that the bare possibility of error blocks knowledge. This premise seems to be more or less taken for granted on all sides, but I think it is pretty clearly false. In fact, I argued explicitly that it is false in my first reply to your original comment (see the fourth paragraph). What knowledge requires is not the elimination of every epistemic possibility of being wrong but the existence of a cognitive channel between knower and known, such that the knower is aware of the thing known. Knowledge is in this way analogous to visual perception. To see something on some occasion does not require the elimination of every epistemic possibility that one is suffering from an illusion; it only requires that an appropriate channel of visual awareness be established between the one who sees and the thing seen.

              So I don’t think I should have said, in my last reply, that, “The real trouble with skepticism is that … [skeptical] hypotheses don’t fare well.” That’s only part of the trouble. The other part is that the mere possibility of error is not sufficient to block knowledge. Both points are required to defeat skepticism. My focus on the first point to the neglect of the latter is due to my concern with the sort of antiskeptics who think they can defeat skepticism by producing some reason why skeptical hypotheses are per se illegitimate and can thus be dismissed without consideration. I think this is the wrong way out and want to make clear why. Maybe what I should have said is that, “The real trouble with skeptical hypotheses is…”

              But with regard to skepticism, not just skeptical hypotheses, the second point is essential and cannot be neglected. Without it, the antiskeptic has no option but to look for some way to rule skeptical hypotheses to be per se illegitimate somehow. And unless it is given adequate emphasis, even a relatively healthy antiskeptic will come off sounding like the kind who is looking to dismiss skeptical hypotheses out of hand. Consider MacIntyre, who starts off with the claims that, “skepticism does not require an answer,” and that, “[just because] it is possible that we are in error gives us no reason whatsoever to believe that we are in fact in error.” His discussion mostly has the flavor of “arbitrary claims can be dismissed out of hand.” That is, he seems to say that the reason we can reject skepticism is that its claims about the possibility of error are not worthy of consideration, not that its claims, when considered, are epistemically extremely weak (and the mere possibility that they are nonetheless true does not block knowledge). This is the way I originally read him. Upon reading the passage more closely, I’m no longer so sure how it should be read. But it is at least a bit confusing, and the reason I’d say is the failure to make both points explicit.


              • I think I see my confusion now. You were initially criticizing MacIntyre’s response to skepticism, but my interpretation of that response makes it, I think, basically equivalent to what you say in your second paragraph here: insofar as skeptical arguments infer that we do not have knowledge from the bare possibility of error, we can dismiss them, because that inference is not warranted; we need to be given some reason to think that a hypothesis is true before we have reason to take it seriously, and skeptical arguments grounded simply in the conceivability of radical error do not give us any such reason. (Is that an accurate characterization of what you’ve said here?) I think I was confused because I took you to be objecting to MacIntyre on that reading; that lead me to confusion about how your initial articulation of the point about the conceivability of error squared with the rest of what you wrote. So let’s see whether I’m sufficiently unconfused now.

                I read MacIntyre as claiming that radical skeptical arguments are per se illegitimate and that we do not need to bother considering the likelihood that we are being radically deceived by an evil demon, or are brains in vats, or are plugged in to the Matrix, or whatever. Insofar as skeptical arguments employ these scenarios only as conceivable in order to show that we cannot know that we are not systematically deceived — rather than arguing, as the ancient skeptics often did, by putting forth an alternative hypothesis as sufficiently likely to show that we should suspend judgment — I’m not sure why you disagree (if you do).

                If I understand you correctly, you want to accept MacIntyre’s claim (as I read it) but you think we also need to show that particular skeptical hypotheses are epistemically weak. My trouble with this is just that, if the skeptical scenario is being put forth not as something that is supposed to be likely, but merely as something that is conceivable, then to disarm the argument we need do nothing more than point out, as you and MacIntyre have, that the bare possibility of error gives us no reason to think that our beliefs are false.

                Whatever MacIntyre would say, I agree with you that if a skeptical argument appeals to some hypothesis as sufficiently likely to warrant suspension of judgment, we need to give some additional reasons to reject it. But of course, as you’ve shown, that’s pretty easy when we’re talking about evil demons, brains in vats, and so on. This is, at the very least, consistent with MacIntyre’s response. It amount to saying: you need to give us some good reason to think that we’re mistaken, and look, the alternative hypothesis you’ve introduced is not something that we have any good reason to think is true.


                • insofar as skeptical arguments infer that we do not have knowledge from the bare possibility of error, we can dismiss them, because that inference is not warranted; we need to be given some reason to think that a hypothesis is true before we have reason to take it seriously, and skeptical arguments grounded simply in the conceivability of radical error do not give us any such reason. (Is that an accurate characterization of what you’ve said here?)

                  No, this is not quite what I would say. The skeptic’s move from the possibility of error to lack of knowledge is not really a question of inference or reasons. (You also say, giving your own view this time, “the bare possibility of error gives us no reason to think that our beliefs are false.”) It’s a matter of supposed criteria of knowledge. I think there is an “intuition” about the land, sometimes in popular usage and often in philosophers’ usage, that if you might be wrong, then you don’t know. For any thing that a person claims to know, if there is some way he might be mistaken that he has not ruled out, then he cannot really claim to know that thing. This is regarded as simply one of the criteria of knowledge, like being justified and true. In effect, it is a matter of definition.

                  I regard this “no possibility of error” criterion as illegitimate, and I have argued for this. But this is not the same thing as regarding skeptical hypotheses as illegitimate. I regard skeptical hypotheses as perfectly legitimate! They are just epistemically weak. I do not regard any factual claims as illegitimate or capable of being dismissed out of hand. People who try to dismiss factual claims out of hand in this context of skepticism generally do so because they don’t see (as I would put it) that the core problem is with the criteria of knowledge, not with the skeptical hypotheses. Skeptical hypotheses, like all hypotheses, must be evaluated on their merits.

                  So my view is really quite different from MacIntyre’s. At least so it seems to me. It seems to me that if we asked MacIntyre why he dismisses skeptical hypotheses, he would say that the skeptic has provided no reason to believe in them. And where there is no reason given in support of a hypothesis, we may dismiss it. But I wouldn’t agree that no reason has been given in support of a skeptical hypothesis. For one thing, every new observation confirms it! (Think grue.)

                  And I also don’t think the empirical rejection of skeptical hypotheses is necessarily so trivial as you imply. Although I think that in contemporary discussion the no-possibility-of-error criterion does most of the work (while getting little recognition for it), it is not at all clear that skeptical hypotheses are so easy to dismiss. That depends on your epistemology. For much of the 20th century at least, many of the most prominent epistemologists would have a big problem with skeptical hypotheses. Consider Quine. He basically acknowledges only one epistemic (alethic) principle of belief acceptance or rejection, namely deductive consistency. All other principles for him are merely pragmatic. So for Quine, the evil demon and brain in a vat hypotheses are “empirically equivalent systems of the world,” and also empirically equivalent with our everyday beliefs about the world around us. So for him there can only be pragmatic reasons to believe in the world around us. We believe in the physical world basically because that’s the most conservative, least upsetting thing to believe. That’s pretty close to saying we don’t know there is a physical world. Or again, take Popper. He has no grounds for ruling out skeptical hypotheses except that their skeptical elements are not falsifiable. But this is again a merely pragmatic consideration. Popper can offer no reason why more falsifiable hypotheses are more likely to be true, and he doesn’t. So I think people have also been scared of the skeptical hypotheses themselves. They have not tended to argue that skeptical hypotheses are empirically rejectable in the way I have done.


                • Well, once again I think I understand your position on this better now than before. But I’m still not entirely sure where our differences lie, though there seem to be some. One thought that comes to mind in response to what you’ve written here is that I’d agree that many epistemological views don’t give us good grounds for rejecting skeptical hypotheses, and I take this as a serious problem for them, such that we’d be better off adopting an alternative epistemology. But I guess I’m still just not seeing why I should regard Cartesian-style skeptical problems as requiring an answer. The skeptical challenge is is: here is a conceivable scenario on which we are radically deceived; now show that this scenario is false and that we are not radically deceived. But the mere conceivability of the scenario gives me no reason to take it seriously, at least so long as I — perhaps unlike many others — accept that it is genuinely conceivable that we are radically deceived. In fact, the conceivability of some such scenario is, I think, a trivial consequence of realism. But since there is no contradiction between the conceivability of such a scenario and my possession of knowledge that, say, I am typing this note to you right now, its conceivability casts zero doubt on whether I really am typing this note to you right now. And since it casts zero doubt, it doesn’t require an answer.

                  Now, maybe I am operating with different assumptions about what it is to require an answer. So let’s see. I am a direct realist about perception, at least in some form or other (there are lots of tricky questions that arise for direct realists and I’m not entirely sure how we should answer them, but I’m convinced that some form of direct realism is right). Hence I take it that you and I have reliable, cognitively direct access to some mind-independent objects. As I see it, this is epistemologically crucial, since even though our direct perceptual access to mind-independent objects does not entail that any of our beliefs about them are correct, it does entail that we have some cognitive content to work with in forming beliefs about mind-independent reality, and lots of these beliefs will be true by virtue of involving very little in the way of theory about their essential nature (‘there is a computer on my lap’ involves no theory about what the real nature of computers or laps may be). I also accept, in outline at least, some empirical theories about how our senses operate to put us in cognitive contact with mind-independent objects — light, retinas, vibrations, ear drums, etc. — and I also accept certain sorts of evolutionary arguments to the effect that we have direct cognitive access to mind-independent objects (since otherwise our chances of survival would be pretty slim) — though I hasten to add that I do not think that direct realism depends on any particular empirical theory about these things.

                  Now, insofar as direct realism about perception yields some cognitive contact with the mind-independent world, I have at least a rudimentary epistemology that explains why skepticism is false. Presented with the question of how I know that I am not radically deceived, I can say, ‘because I am directly aware of mind-independent objects.’ Pressed with the question of why I think I am directly aware of mind-independent objects, I can say, ‘because that makes the best sense of my experience, and arguments that purport to show that I am not directly aware of mind-independent objects fail.’ Aside from showing that standard arguments against direct realism fail, I could rehearse evolutionary arguments or the little that I know about scientific theories about how my sense organs work. So in one sense, I suppose this gives me a rudimentary epistemology, and hence an ‘answer’ to skeptical arguments.

                  It seems important to me, though, that none of these arguments do anything to answer the kind of skepticism that says, “hey, it’s conceivable that you’re radically deceived because you might be a brain in a vat or duped by an evil demon or plugged into the Matrix or whatever.” If we take that sort of hypothesis seriously, then none of our empirical evidence or the theories we formulate to make sense of it can defeat the hypothesis, since they’re all consistent with the truth of the hypothesis. But that’s not a problem, because the hypothesis isn’t based on anything other than its sheer conceivability, which is a trivial consequence of realism. So if the argument boils down to ‘it is conceivable that you are mistaken, therefore you do not have knowledge,’ then I need do nothing more than point out that this is a non-sequitur. If the skeptical argument is supposed instead to appeal to some hypotheses that make equally good sense of my experience, then I can respond in the ways you’ve suggested, but then I’m not responding to the same kind of skeptical argument anymore; I’m responding to a very weird sort of empirical hypothesis instead. I’m inclined to say that even taken as an empirical hypothesis, the mere fact that there are no positive grounds for thinking that I am a brain in a vat or the victim of an evil demon or what not is enough to dismiss the argument, but I’d of course also accept the sorts of considerations you’ve adduced against these hypotheses. The trouble is just that skeptical arguments as I know them aren’t in the business of offering us hypotheses intended to make good sense given our experience and our ordinary assumptions about it, but rather are out to show us that we have no way of defending that experience or those assumptions without presupposing the denial of skepticism.

                  I can see that I’m just talking in circles now, though, so I’m not sure I’m making any progress here. You’ve said enough at this point to convince me that you know what you’re talking about, so I’m not satisfied with my inability to come to terms with what you’re saying. But I’m not sure what else I can say.


                • Yes, I’m afraid this has become repetitive, and I can’t really think of anything new to say. So maybe we should quit for now.

                  I would only add two things, as food for thought. First, you say:

                  So if the argument boils down to ‘it is conceivable that you are mistaken, therefore you do not have knowledge,’ then I need do nothing more than point out that this is a non-sequitur.

                  But whether it is a nonsequitur depends on the criteria of knowledge. Since the skeptic of the kind we’re thinking about regards any possibility of being wrong as incompatible with claiming to know, it is not a nonsequitur for the skeptic. If you want to defeat him, therefore, you need to do more. You need to address yourself to the criteria of knowledge.

                  Second, you say skeptic hypotheses are “conceivable.” Maybe you should try “possible” instead. “Conceivable” has an abstract sound, as though we were talking about some uninterpreted symbolic formalism, for example, that may be making it easy to forget that skeptic hypotheses are live possibilities.


                • Indeed, though there may be some interesting issues lurking beneath the surface. I think I agree that defeating skeptical arguments requires addressing the criteria of knowledge. I just think that we’re doing precisely that when we challenge the move from the conceivability of systematic error to the absence of knowledge.

                  I use ‘conceivable’ as something of a term of art to avoid the ambiguities of ‘possible.’ It may in fact not be at all possible that we are deceived by evil demons because there are no demons, evil or not, but it is, prima facie at least, conceivable that there are evil demons who deceive us — conceivable because there are no logical contradictions involved in supposing that there are. In just the same way, it isn’t possible for a human being to lift up his arms and fly through the air like Superman, but there is, prima facie at least, no logical contradiction involved in thinking so. Some might want to resist the idea that these things are really conceivable if they are in fact impossible given the laws of nature, and I’m somewhat sympathetic to that line of thought, but I find it easier to think about skepticism (and some other things) in terms of conceivability rather than possibility simply to avoid problems concerning possibility. I’ve also been convinced by the work of Stephen Boulter and others that the conceivability criterion of possibility lies behind a lot of dead ends in modern philosophy, so that conflating the two invites unnecessary problems. (See Boulter’s ‘The Medieval Origins of Conceivability Arguments’ and ‘Hume on Induction,’ and, somewhat less satisfying from my point of view, Douglas Rasmussen’s old ‘Logical Possibility: An Aristotelian Essentialist Critique’).

                  In any case, I think it’s clear that we mostly agree about skepticism, and I understand your ideas better as a result of all this repetitive rambling I’ve been doing, so it hasn’t all been in vain! Thanks!


    Zbigniew Janowski

    Ayn Rand mentioned some Augustine-Descartes connection in her “For the New Intellectual” (1960): “Descartes began with . . . (a premise he shared explicitly with Augustine) ‘the prior certainty of consciousness’, the belief that the existence of an external world is not self-evident, but must be proved by deduction from the contents of one’s consciousness.”

    In 1945 she had read (at least some of) Etienne Gilson’s THE UNITY OF PHILOSOPHICAL EXPERIENCE (1937). Therein:

    “When he reached that point, Descartes—or was it only a young boy of sixteen who had heard of it at La Fleche?—remembered that long ago, another man had found himself in a similar difficulty, and had discovered a way out. . . . Like Descartes then, and before him, St. Augustine had become a sceptic in spite of himself, but he had also succeeded in his effort to discover a decisive answer to skepticism. It is to be found in his SOLILOQUIES, book II, chapter 1. Reason is leading the discussion with Augustine: ‘You, who wish to know yourself, do you know at least that you are?—I know it.—How do you know it?—I don’t know— . . . —But you know that you think?—Yes, I know that.—Consequently, that you think at least is true.—It is true.—You know therefore that you are, . . . and that you think’

    “[Descartes]: ‘Then without doubt I exist, also if he deceives me, . . . he can never cause me to be nothing so long as I think that I am something’. Even in this, Descartes was repeating what Augustine had said in another text ON FREE WILL, book II, chapter 3: ‘First, I ask you in order to begin with what is the most evident, whether you are, or not? And in this you cannot fear to be deceived in you answer, because in case you did not exist, you could not possibly be deceived’. And again, in his CITY OF GOD, book XI, chapter 26: ‘If I am wrong, I am, for he who does not exist, cannot be deceived; thus, from the very fact that I am deceived, it follows that I am. How then could I possibly be deceived in believing that I am, since it is an obvious thing that I am so long as I am deceived’?

    “In 1641, when Descartes restated his first principle in his MEDITATIONS IN FIRST PHILOSOPHY, . . . Arnauld . . . was not slow in pointing out the fact that St. Augustine had already said the same thing many centuries ago. Descartes did not seem to relish the remark . . . .”

    Liked by 1 person

    • Augustine also comes up in the Appendix to Rand’s Introduction to Objectivist Epistemology, in a conversation involving Rand and Professors B and E (Allan Gotthelf and Leonard Peikoff, respectively). The topic is “the explicit formulation of axiomatic concepts,” and Rand more or less agrees with Peikoff’s claim that Augustine devised/invented the concept of “consciousness” (pp. 262-63).

      Rand: You mean Augustine was the first to isolate “consciousness” as a concept in the Cartesian sense?

      Peikoff: Yes. “Si fallor, sum.”


  3. For whatever it’s worth, Descartes has often been alleged to have ripped off al Ghazali. I don’t know Ghazali well enough to evaluate the claim.

    I’m not one for global skepticism, but skepticism about the existence of grades, students, registrars, and possibly universities as such (though not their Payroll Depts) would play a very useful epistemic-ethical function right about now.


  4. I’m not convinced that the prior certainty of consciousness is true, whether construed along Augustinian or Cartesian lines. Here I’m echoing part of David Riesbeck’s comment, but taking it further:

    I’m no Augustine scholar, but in general my sense is that he would affirm that the objects of consciousness are prior to our consciousness of them, even if we can be wildly mistaken about what those objects are. So perhaps it would be better to say not that Augustine affirms “the prior certainty of consciousness,” but “the greater certainty of consciousness”?

    I don’t remember what Augustine says, but as a purely philosophical matter, I agree with the point DJR is attributing to him. At best, it seems to me that the cogito underscores the fact that mental acts presuppose a mental agent. It doesn’t prove that knowledge of consciousness is epistemically prior to knowledge of the objects of consciousness.

    Here’s a somewhat Randian way of making the point: the cogito presupposes a grasp of the concept of “deception,” but I don’t think it’s plausible to assert that we can grasp that concept by reflection on our selves, prior to and independently of interaction with other people. We grasp the concept of deception by deceiving and being deceived by other people, where those people are objects existing independently of us. Whereas self-deception seems to lead to puzzles and paradoxes, interpersonal deception is relatively straightforward. In other words, self-deception is the more opaque, less knowable phenomenon that we model on interpersonal deception.

    I also don’t think our knowledge of perceptual “fallibility” is possible without presupposing knowledge of the objects of perception. The supposed “frailty” of perceptual knowledge (if one wants to call it that) would be unintelligible unless we first had knowledge of the objects of perception. But that knowledge seems incompatible with the prior certainty of consciousness. I know Descartes’s Meditations a lot better than I do any part of Augustine, but Meditation 1’s set-up of the issue of perceptual relativity is, to my mind, incoherent in ways that don’t seem to have registered in the literature (at least not that I’ve read). As I see it, that incoherence is a result of Descartes’s presupposing the existence of the consciousness-independent objects of consciousness while simultaneously trying to affirm the prior certainty of consciousness.


    • Looking back, I’m afraid I was a bit slack in my discussion of Descartes’s “prior certainty of consciousness.” I’m sorry I used the phrase “prior certainty of consciousness,” which is vague and also turns out to have Randian connotations I didn’t realize it had. In particular, I don’t mean to be endorsing the Cartesian claims about consciousness that Rand meant to be denying. I’m also unhappy with the phrase “the mind is better known than the body,” which is also vague and can easily seem to imply things I don’t mean to endorse.

      What I had in mind is really pretty limited. It is simply that what we think or perceive about external objects is more securely knowable than the objects themselves. I gave the example of a visual experience of a cup versus the cup itself, and the point it illustrates holds in general. You can be more certain of how warm you think the ambient air is in the room where you are than you can of how warm it actually is; more certain of how you remember spending last Fourth of July than of how you actually spent it; more certain of what you think about Donald Trump than of the facts about The Donald. Thus, I like you am happier with the phrase djr suggested, “greater certainty of consciousness,” than with the phrase “prior certainty of consciousness.”

      In advancing this thesis, obviously I do not mean to be saying that we have direct awareness of all our own mental contents or processes. Nor do I mean to be saying that we have infallible awareness of the contents and processes we do have direct awareness of. And I certainly do not mean to be saying that we begin epistemically from knowledge or awareness of internal states from which we infer the existence of external things. In fact, if anything we do the opposite. How do you know what you think of Trump? Not by consulting your thoughts as such, but by asking what are the facts about Trump. To know what you think about Trump, you ask yourself what is the case about Trump, and however you answer that question also tells you what you think.

      This point might seem less than obvious in the case of sense-perception. For instance, don’t I visually estimate a person’s height by consulting the visual experience I have of him? Don’t I think he is six feet tall because he looks six feet tall? Yes, but—usually—all we mean by such expressions is that our visual system is the source of our height estimate. Just as I might assert some fact and cite the newspaper or a research report as my source, so I say he is six feet tall and cite my visual experience as my source or authority. I mean I looked and my visual system reported that he is (about) six feet tall. I do not mean that I looked and my visual system confronted me with an “appearance” which I then examined qua appearance after assuming some sort of “transcendental epochê.”

      Of course, we do sometimes examine our visual appearances qua appearances. For example, when I look at a disc from an oblique angle and report that it “looks elliptical,” I am not saying that my visual system has reported the disc to be an ellipse. Obviously (at least it should be obvious) my visual system has told me no such thing. What it has told me is that a circular disc is at an oblique angle to my line of sight. But, phenomenologically as it were, a disc at that angle has an elliptical appearance. We typically have no trouble identifying such subjective, phenomenological facts about appearances and no trouble distinguishing them from the objective facts about objects reported by the visual system. Nor do we typically talk about or even notice them. Our usual concentration is not on appearances qua appearances but on “appearances” qua visually experienced objects and their properties. And in this sense if I want to know how tall a person visually “appears,” I must consult how tall I would say he is on the basis of looking. In this latter sense of appearance, there is no access to how something appears other than the objective properties reported by the visual system. Here as elsewhere (e.g., Trump), to know what my mind—in this case my visual system—represents, I must consult what seems to be the case after looking.

      The reason for this is that the representational mind presents us with the contents of mental representations, not with the representations per se. When you think about Trump, you think about Trump, not about your idea of Trump. When you think about last Fourth of July, you think about that time, not about your memory per se. When I think about the pen in my hand, I think about the pen, not about my visual and tactile experiences of it. Not that we can’t be aware of our representations as such. But to be aware of representations is a sophisticated, conceptual achievement. Cognitive psychologists have studied the development of such “metaknowledge” and found that young children have a great deal of trouble distinguishing facts from their (and other people’s) knowledge, memories, sense-perceptions, etc. of facts. Very young children do not make such distinctions at all.

      Since our first awareness of the world is of the world, not of our representations of the world, it is perhaps ironic that we should be more certain about our representations than about the world. Ironic perhaps, but not paradoxical. It is simply a question of the accuracy of our representations versus the accuracy of our knowledge of our representations. With regard to the former, we have to be right about the way the world is. With regard to the latter, we only have to be right about how we represent the world to be. It is easier to be right about the latter because, as it were, we are closer to our representations than to the world. We have, indeed, practically indubitable access to how we represent things to be. Unfortunately, the same cannot be said for our access to how things are.

      I have been using the language of representations here, but the points I’ve been making (that we know the world earlier than we know our minds, but we know our minds with greater certainty) don’t require it. They are a natural consequence of the recognition that our knowledge of the world is a cognitive achievement. We know, perceive, etc. with our minds. The total brain state we have when we know or perceive the world might also arise when we fail to know or perceive simply because the world fails to match our “knowledge” or “perceptions.” If this is right, then all perception and knowledge could be radically false and in that sense deceptive. Even when tamed and converted to “the greater certainty of consciousness,” this remains a consequence of the view.


      • Thanks, and I’m grateful for the clarification. Amusingly, I thought you’d used “prior certainty of consciousness” precisely for its Randian connotations!

        I’m going to have to think about the view you’re laying out in the comment. I think I understand it, but I’m not sure I do, and if I do, I’m not sure whether I agree or disagree. What I am pretty sure about is that at some point you’ve got to transform some of your PoT blogging into journal articles.


    • Another tension, in the case of Descartes:

      A contemporary Aristotelian moderate-realist barrier to demonic skepticism concerning conceptual knowledge would be that of Gyula Klima. This Thomist bar to radical skepticism of concepts is by necessary identity of form in thing and in mind for at least some things.[1] The Randian bar, in 1957 and beyond, is by necessary latch of identity and conceptual identification for at least some things.[2] Concerning the “ideas” that are our perceptions of the so-called secondary sensory qualities, such as scent, heat, and color, Descartes could be reasonably said to be a nominalist.[3] He is, however, not nominalistic when it comes to the ideas that are concepts. I observe that although he is not much occupied with abstraction of species classes from individuals and the alternatives thereon between realism and nominalism, his thinking about definitions and essences and about abstractness in our thought, such as in proofs in geometry, set him rather with realism than with nominalism concerning abstractness. His crafting of demonic skepticism, unlike related crafts in the fourteenth century (Ockham, Wodeham, and Buridan concerning limits of God’s power with respect to human cognition), is not facilitated by an extensive nominalism. Descartes’ background conception of the physical world’s existence and our knowledge of it as contingent on God’s power leaves plenty of imagined room for radical skeptical posit of their possible total falseness, even with a prima facie realist epistemology for an ideal human thinker. Perception and conception, in the final account of Descartes, are certified as truly realist once the necessary existence and necessary goodness of God is proven.[4]

      [1] Klima 2011a; 2011b; see also Gilson 1939, 187–93.
      [2] AS 1016.
      [3] Nolan 2011.
      [4] Descartes on abstraction: 1628, 417–25, 441, 445–52; 1642b, 474–77. Descartes on definition, essence, nature, and identity: Med. 1st Replies 115–19; 1648a, 347–49.

      Descartes, R. 1628. Rules for the Direction of the Mind. In Cottingham, Stoothoff, and Murdoch 1985 (CSM-I)
      ——. 1641a. Meditations on First Philosophy. In Cottingham, Stoothoff, and Murdoch 1984 (CMS-II).
      ——. 1642a. Letter to Regius, January. CSMK.
      ——. 1648a. Comment on a Certain Broadsheet. CSM-I.
      Gilson, E. 1939. Thomist Realism and the Critique of Knowledge. M. A. Wauk, translator. 1986. San Francisco: Ignatius Press.
      Klima, G. 2011a. Demon Skepticism and Concept Identity in a Nominalist v. a Realist Framework. In Klima and Hall 2011.
      ——. 2011b. Demon Skepticism and Non-Veridical Concepts. In Klima and Hall 2011.
      Klima, G., and A. W. Hall, editors, 2011. The Demonic Temptations of Medieval Nominalism. Newcastle upon Tyne: Cambridge Scholars Publishing.
      Nolan, L. 2011. Descartes on “What We Call Color.” In Primary and Secondary Qualities – The Historical and Ongoing Debate. L. Nolan, editor. New York: Oxford University Press.


  5. I wonder if the creation story in the Bible and the neoplatonism of the Gospel of John and other neoplatonic influences on Augustine don’t set him in a prior certainty of consciousness, with the mysterious consciousness of God as best ground of our truth, whatever the senses and world may suggest. On the other hand, it doesn’t seem he appeals to existence of a good, undeceiving God to give him knowledge that he has living existence. God, pure spirit, is also awarded living existence by that tradition, but I doubt Augustine meant OUR living existence known under whatever our cognitive errors was without any of its physical aspect.

    David, Augustine part of your post aside, it does not seem sensible to say “I talk” and know what one is saying without knowing of communication with other people and co-reference (which we observe in infants beginning around one year). “I think, therefore (probably quiet a bit if I know what “think” means at least in its primary sense, but at least the following bit) I talk.”

    I don’t recall just now having seen that wedge against Descartes before, but I haven’t looked through my materials. I have another question for you if you have a minute. I might find this in your dissertation, if you don’t. Understanding concepts like ‘the senses’ or ‘experience’ or ‘systematic error’ are pretty advanced in developmental history. If I don’t first learn ideas of ball, hand, dirt, and owie, how will I come to have those more advanced concepts and how can the more advanced ones be more rightly secure than the earlier ones? Naturally, too, I’d agree with American pragmatists in denying one really has these particular more advanced ones if they’re not cashed out with those earlier, more accessible ones. (Similarly, no support of symbolic, linguistic representations without indexical and iconic representations—Peirce and Deacon. Then too, no understanding of any numbers beyond the naturals without understanding of the naturals.)

    If I pose the possibility, as had Descartes, that my body and its surround is illusory tout court, with what justification do I get to keep on talking with the more advanced concepts? If I call into question any and all history, how am I justified in talking over intervals of talking and knowing? Also, if we get to come to the meditation, as he says, with the assured knowledge that what thinks exists, why not with the assured knowledge that what thinks in talk is a real physical object among others and among other such talkers?


    • Hi Stephen. I have outlined in my replies to David R. and Irfan my approach to global skepticism and the way we can be more certain of our mental contents than of external objects and facts. In view of the sheer size of this topic, those remarks were brief, but still I hope they’re enough to convey a general idea of my view and why I am not a fan of “genetic” arguments of the sort you seem to be raising; that is, arguments to the effect that certain claims cannot be made because they would undercut their own conceptual roots. Thus, for example, global skepticism can’t be formulated without concepts such as the senses, experience, and systematic error, but these concepts depend for their essential content on concepts of physical reality, which would be invalidated if global skepticism were true.

      Rather than repeat what I’ve said in other replies, I’ll just make two comments.

      First, regarding the question how we can keep talking about the senses, error, and such if there’s no physical reality to support these notions, I agree that the dependencies exist and that such concepts can’t continue to be used in quite the same way if global skepticism is true. But of course that goes for most of our concepts if global skepticism is true! Accepting global skepticism would represent a “paradigm shift” of titanic proportions. How we understand the phenomena we formerly took for physical reality would change radically, and so correspondingly would how we understand what we formerly regarded as our senses and our experiences of those phenomena. But it doesn’t follow that there needs to be any conceptual incoherence in the new “paradigm.” (I can’t believe I’m actually using this Kuhnian term; hence the scare quotes.)

      Imagine you’re the character Neo in the movie The Matrix. You thought you were a mild mannered computer programmer with an apartment in the city, a sports car, and a mother who loved you, but then the veil of perfect virtual reality is pulled briefly back and you are forced to recognize that you’re a brain in a vat and almost nothing you thought about your life is true. Now you have to carve out a new understanding of all the phenomena around you. What you previously thought was your sports car now becomes a systematic pattern of neural impulses more or less reliably stimulated by electrodes programmed by some super alien intelligence. (I realize that electrodes, brain, aliens, and so forth are physical reality concepts, if alternative ones, but this is not essential to the story, as the example of Berkeley shows.) Similarly for your experiences of your car. Whereas the car itself is the consistent, invariant pattern in car phenomena (the “permanent possibility of perception” in Mill’s famous phrase), your experiences of the car are the particular occasions on which the phenomena arise. Your experiences will be veridical when they satisfy the invariant pattern, illusory when they don’t in certain explicable ways (seen through a “distorting lens” or in unusual “lighting,” for example). Dreams violate invariant patterns more radically but also explicably, being followed by “waking,” being often related in intelligible ways to your recent thoughts and concerns, and so forth. I don’t see any reason in principle why a conceptually coherent picture of life as a brain in a vat can’t be painted along these lines.

      Second, I think it is odd that Objectivists, who consider it important to uphold the primacy of existence, are so keen to employ genetic arguments—such as the Stolen Concept argument—against global skepticism. These arguments consist essentially in saying that global skepticism must be false because its truth is incompatible with our existing concepts. This makes our concepts of reality the criterion for how reality is. And this gives primacy to consciousness, not to existence. Therefore, such arguments ought to be incompatible with Objectivism.


      • David, I noticed that Bolzano, notwithstanding his many fine realist features, thought we are better acquainted with the events in our inner selves than with events outside us. Would you agree with that? That seems a relative of the inner being more rightly certain than the outer, though I haven’t sorted out the relation. Anyway, I’m better acquainted with the fire in the fireplace than with my reception and processing of the fire’s input to sight, hearing, smell, and sense for heat flow rate across skin surfaces. Attending to whether the burning logs need adjustment is much less labored than attending to and discriminating what are all my sensations from the fire. That’s not to say that’s the way it stands for everyone with fireplace fires. Presumably with my earliest (forgotten) encounters, I noticed the flame motions and colors and felt the heat and was acquainted not at all with what is now so automatic concerning saving the day with the fire.

        You said “Imagine you are Neo . . . .” Now you know I’ll decline those more extravagant imaginations in metaphysics and epistemology—you and Nozick 2001. But to tell the whole truth, I’d have to read a synopsis of The Matrix to get the story. I tried to watch it once maybe ten years ago, but failed. I could not develop any interest in any characters, and there was some sort of annoying sweeping to new angles of view to boot. Anyway, I fell asleep. But Berkeley, now that would be an old acquaintance a pleasure to renew. My first philosophy course in college was with a Thomist. My second was history of modern philosophy. That’s where I first met Descartes and Berkeley. Our Prof. was a young man from Columbia, Monte Cook. (And he is still teaching there at Oklahoma today, nearly 50 years later!) He would argue with the turns of Berkeley’s dialogue, and G. E. Moore came significantly into view in the bout as well. As he lectured, he chain-smoked, and we students were lighting up as well. One of our books of readings for that class was edited by Richard Popkin (The Philosophy of the Sixteenth and Seventeenth Centuries). In recent years, I have gotten Popkin’s second edition of The History of Skepticism—From Savonarola to Bayle. I see here on my shelf that the paper you mentioned by Burnyeat “Can the Skeptic Live His Scepticism?” is collected in the volume he edited The Skeptical Tradition.


        • Hi Stephen.

          I noticed that Bolzano, notwithstanding his many fine realist features, thought we are better acquainted with the events in our inner selves than with events outside us. Would you agree with that? That seems a relative of the inner being more rightly certain than the outer, though I haven’t sorted out the relation. Anyway, I’m better acquainted with the fire in the fireplace than with my reception and processing of the fire’s input to sight, hearing, smell, and sense for heat flow rate across skin surfaces. Attending to whether the burning logs need adjustment is much less labored than attending to and discriminating what are all my sensations from the fire.

          No, I don’t think I would agree with Bolzano. I think I would agree with you. See my reply to Irfan of Dec. 11 above. By “acquaintance” I assume we mean some sort of direct awareness. Knowledge by acquaintance is as opposed to knowledge by “description”—famously—and more broadly it is as opposed to knowledge that requires inference or reasoning. It used to be, in the bad old days, that people (like Russell) thought we are only acquainted with sense-data, so that all knowledge of the external world is descriptive. Those days are in the past, however, and in the new world order direct realism (in one form or another) is the fashion leader. This is a fashion I’m happy to follow. But this does not affect the prior certainty of consciousness. Again, see my reply to Irfan.

          You said “Imagine you are Neo . . . .” Now you know I’ll decline those more extravagant imaginations in metaphysics and epistemology—you and Nozick 2001. But to tell the whole truth, I’d have to read a synopsis of The Matrix to get the story. I tried to watch it once maybe ten years ago, but failed.

          I sympathize with you about The Matrix. I watched it in the theater when it was new and didn’t particularly like it. However, I don’t think it is legitimate to decline to consider what you would think and how you would have to proceed cognitively if you were to be confronted with empirical evidence of the sort I describe; that is, empirical evidence that the world is radically different than we think, so radically different that you have no body, there is no planet earth, your local environment contains none of the things you seem to perceive (furniture, books, rooms, pets, loved ones), maybe even that there is no local environment because space itself is a perceptual error. It is really not that hard to imagine having a series of experiences that would force one to such conclusions. Imagine that your visual field suddenly dissolved into random pixelations which proceeded to form themselves into bizarre configurations that corresponded to nothing and to no world. And suppose that this was accompanied by a series of thoughts introduced involuntarily which described in advance exactly what you would experience next. And suppose you could “think back” at the alien thoughts and challenge them what images or feelings to present, and it could do whatever you suggested. What should you make of this train of events? That you had gone mad, obviously, or were having a nightmare or were under the power of a hallucinogenic drug. But as time wore on, these hypotheses would lose their plausibility, it seems to me. It would be plain that your reason is intact, and the evidence of control of your experiences by the “alien” would be too consistent and coherent to be much like a dream or hallucination. And perhaps moreover the distinction would become irrelevant. Like it or not, your “world” would have become a series of sensations arbitrarily controlled by an alien. And then you would have a considerable amount of conceptual rebuilding to do.

          I don’t see why such future evidence is impossible. The fact that it is inconsistent with our current concepts and theories is irrelevant. Perhaps the alien has given us a train of experience up to now that supports our current concepts and theories to test our fortitude when the shock of the new revelations hits us. Whatever the reason, the alien that gives us the new experiences could as easily have given us the old as well, so that our current concepts and theories—based on the old experiences—are (and were) radically false, however useful they may have been up to now. The larger point is that our concepts and theories are empirical, not a priori. They are based on experience and they are determined by our experiences, not vice versa. Therefore, they can’t dictate what our experiences must be or cannot be. As I mentioned before, especially on a primacy of existence view, I don’t see any way around this conclusion.

          I would strongly recommend the Burnyeat paper, if you can fit it into your workload. He brings great clarity to what exactly is going on with Pyrrhonism.


      • No. We introduce new concepts replacing old in mathematics and in science when there’s specific reason, specific solution, to do so. We don’t lose or aspire to lose validity of all earlier concepts in the discipline. Philosophy following suit in that way is worthwhile to me.

        Talk of a plane with no incidence character of its points or no neighborhoods in the plane goes empty, and mathematicians have plenty of frontier to work on without such wildness. The latter is contrary fidelity to mathematical truth and mathematical existence. Similarly with physical science, and as I deal in it, philosophy.

        Super-alien-deceivers or God-but-a-deceiver who are themselves undeceived look suspiciously like us. Better to start with natural systematic error (you’ve helpfully mentioned in earlier communication) run to infinity, I’d think. But under any form of such extravagant release from known constraints, I don’t think purposing or meaning-seeking or caring is sensible; those would be driftwood from here, concepts and patterns to be let go of, I should think.

        I took a quick look at the Burnyeat paper, I may have before, but anyway these elements of Pyrrhonism I pick up in the scan were in the early part of that first course in modern philosophy I mentioned upstream, and I guess I’d have to dig in further to see what is novel from what I was taught. (More on interpretation of Pyrrhonism—Burnyeat/Frede/Kant/Hegel—is here:
        Moving in mind with appearances alone, recognized as that, gets pretty easy for Kant I’d say (leaving aside the less theoretical rubs of moral law, freedom, God, and immortality), but then, a rose by any other name . . . .


        • Super-alien-deceivers or God-but-a-deceiver who are themselves undeceived look suspiciously like us. Better to start with natural systematic error (you’ve helpfully mentioned in earlier communication) run to infinity, I’d think. But under any form of such extravagant release from known constraints, I don’t think purposing or meaning-seeking or caring is sensible; those would be driftwood from here, concepts and patterns to be let go of, I should think.

          This seems to be the key paragraph, the one where the pertinent arguments occur. But I’m not quite sure what they’re supposed to be. For instance, in your first sentence you say the alien deceiver of my example looks suspiciously human. But so what, exactly? Do you mean that such an alien wouldn’t be cognitively powerful enough to pull off the technical feat of the deception? I agree, that could be a reason for doubting initially that the alien controls your visual experiences. But the alien might explain that he is dumbing down his self-presentation to be able to communicate with you. That is how God has always operated, after all, and understandably. And now, having created “this world” to test you morally, he is moving on to the next phase, in which the illusion of a physical world can be dispensed with. Is there some reason why there could be no amount of empirical evidence sufficient to rationally convince you of this? If there is, I have yet to hear of it.

          The question is one of evidence. I am not saying we have any reason to take such hypotheses seriously given the evidence as it is. In fact, I have argued in my discussion of “piggyback” theories that it is rational to reject such hypotheses on the evidence we actually have. But now the question has changed. At least, if I understand you correctly, you are suggesting that evidence scenarios such as I have described should be dismissed out of hand, that there is some rational basis for refusing to consider them. But what rational basis could there be?

          My argument essentially is that if the world as you experience it looks nonspatial and nonphysical, and looks that way long enough and consistently enough, you must eventually conclude that it is nonspatial and nonphysical. To do otherwise is to deny reality, to stick your head in the sand and refuse to consider the evidence. I don’t see how, from an empiricist or primacy of existence point of view, one can defend such a refusal.

          To deny my argument, I think you must deny either that the sort of evidence scenario I’m proposing is possible or that it could warrant the conclusion that the world is nonphysical.

          It seems particularly difficult to deny the very possibility of the sort of evidence scenario I propose, at least if one is an empiricist. The whole idea of empiricism is that the evidence comes first and theory must respond to that. We do not dictate the evidence, the evidence dictates to us. Any idea we could have that could move us to doubt or reject any particular piece of evidence can only be as strong as whatever evidence supports it. So the question is always one of weighing the assorted evidence and theoretical proposals and finding the most consistent and best-supported account. Therefore, there can never be any support for the idea that some sort of evidence is impossible. The only way to maintain that some evidence is absolutely impossible is to claim that you have some knowledge a priori that is absolutely certain and that forbids such evidence. I doubt you wish to claim this.

          To deny that the proposed evidence scenario could warrant the conclusion that there is no physical world, there are again two possible paths you could take, or so it seems to me. One would be to deny that the proposed evidence is evidence that reality is nonphysical, or in other words to deny that on the proposed evidence scenario the world “looks nonspatial and nonphysical.” For instance, one might try to maintain that our senses and especially the visual system simply always present to us a 3-D world of concrete objects (“entities” in your jargon) bearing attributes and arrayed in space. There is even a sophisticated real life version of this sort of claim in P. F. Strawson’s disinfected and fumigated version of Kant (in Individuals and The Bounds of Sense). This would require an argument showing that “any experience which could be intelligible to us” (to quote Strawson) must be experience as of spatially arrayed concrete objects. Note the primacy of consciousness premise here, by the way. Intelligibility is made a criterion of the real.

          But regardless of this, the claim does not seem to be true, nor intuitively does it seem impossible for experience to present an alternative sort of world. In fact, we don’t have to imagine it, since we routinely experience it. Close your eyes and attend to the phosphene “light show” that appears on the back of your eyelids. This is not a particularly interesting or organized train of experience, but it is not unintelligible or incomprehensible, and it is certainly not impossible, since it exists. It does not present a 3-D world of concrete objects! Nor does the much better organized scenario I am proposing, which might consist of a series of 2-D geometric shapes in psychedelic colors perhaps, organized in recognizable patterns that might be predictive of each other and of other sensations in other sense modalities (including pains and pleasures), but none of it organized so as to suggest a 3-D world or concrete objects with any degree of permanence. It seems to me that on such an evidence scenario the world would appear to be a series of sensations whose esse is percipi. If so, then it is not true that an evidence scenario cannot present the appearance of a nonspatial, nonphysical world.

          The remaining possibility is to argue that although experience might conceivably present us with a nonspatial, nonphysical world, it could still never be rational to conclude that the world is nonspatial and nonphysical. This might be what you have in mind when you say that it is “better to start with systematic error run to infinity” than accept that there is no physical world and that “purposing or meaning-seeking or caring” would cease to be sensible without a physical world. But neither of these, if I’m understanding them correctly (probably not!), is a very strong argument. It is indeed better to start from the hypothesis that one is delusional, as I myself said in my last reply. But it is not necessarily rational to end there. If the pattern of nonspatial, nonphysical appearances goes on long enough to become the new normal, and there doesn’t seem to be anything otherwise wrong with you, and you have an alternative hypothesis that explains your previous “delusion” that you lived in a spatial, physical world, then excuses not to accept the appearance for the reality will come to seem just that. And if this argument still doesn’t convince you epistemologically, think about the case psychologically. After a while, you will inevitably begin to take this new normal at face value. The pattern of red blotches followed by blue squiggles that signals that you are about to feel uncomfortably warm you will think of as just that, the red blotches following by blue squiggles, not “my delusion of red blotches followed by blue squiggles.” You will gradually, inevitably cease to have the rebellious feeling that this is all wrong and must have some physiological explanation and simply accept it as the new reality, as evidence in its own right, the bulk of evidence you have, to be explained on its own terms. And it seems obvious to me—and what I am saying is—that it will be rational for you to do so. If so, then the thesis that we live in a spatial, physical world of concrete objects is not incontrovertible.

          With regard to Burnyeat and Kant, I think Kantian appearances are quite different from what Burnyeat is attributing to the Pyrrhonians. A Kantian appearance, an experience of the phenomenal world, is an object of knowledge. Kantian appearances can be scientifically, rationally known. Pyrrhonian appearances, by contrast, are not knowable. If you ask the Pyrrhonian, “don’t you at least know how things appear?”, he will say, “no.” The Pyrrhonian simply refuses to make any commitments of any kind. A Pyrrhonian appearance is what you have left when you take the content of a belief and withdraw the attitude of belief. Whether it’s psychologically realistic to think we could actually do this—much less achieve ataraxia thereby—is Burnyeat’s question.


    • Fixed. Incidentally, I ran into a technical glitch while trying to upgrade the site today so as to get better functionality, like authors being able to edit their posts (oh the irony). Hopefully will be able to resolve that soon.

      Liked by 1 person

  6. This is a pretty belated comment, but two other pre-Cartesian philosophers not influenced by Augustine who articulated some version of the cogito, albeit not quite so clearly, are Shankara in 8th-century(ish?) India and Ibn Sina (Avicenna) in 10th-11th century Persia. And Ibn Sina actually uses it to argue for a mind-body distinction; see his thought-experiment variously referred to as the “flying man,” “floating man,” and “suspended man” (essentially a person under sensory deprivation):

    Another Cartesian argument, that God would be a deceiver if our cognitive faculties were hopelessly defective, also shows up in the 16th-century Spanish philosopher Francisco de Vitoria.

    Whether they influenced Descartes I don’t know; but it’s not impossible in the case of Ibn Sina or Vitoria. Shankara belongs to a pretty separate stream, however.

    Liked by 1 person

    • Thanks for your reply, however belated. It was informative, since I didn’t know about the Floating Man argument, and I’d never heard of Shankara or Francisco de Vitoria, much less their arguments.

      However, after reading about the Floating Man argument, I’m surprised that it’s thought to have much in common with Descartes’s Cogito. For one thing, whereas the Floating Man imagines a scenario of complete sensory deprivation, the Cogito imagines a scenario of totally deceptive sensory experiences. The question of what we can know in the absence of all outer experience seems to be quite different from the question of whether there is anything we can know for sure we aren’t being deceived about. More importantly, unlike with the Cogito, there doesn’t seem to be a genuine argument given by the Floating Man. The Floating Man says, basically, “Imagine that you had just popped into existence and had been ever since continually in a state of complete sensory deprivation. Isn’t it clear that you would affirm your own existence? Yes, it is clear!” This seems to be more of an intuition pump than an argument. I mean “argument” in the sense of having premises, conclusion, and logical relations between them. In this sense, no reason is given why the person in the Floating Man scenario would affirm his own existence. You’re just supposed to have the intuition that he would. But I don’t myself feel any strong intuitions about the case—nor would I think they deserved much credence if I had them. (But this alludes to my well-known (around here) prejudice against the reliance on intuitions in philosophy.) Perhaps, in the absence of any outer experience, this individual’s mind would be a complete blank and unable to affirm the existence of himself or anything else, and even lacking any such concept as existence.

      By contrast, the Cogito argues on the basis of a simple logical relation: If I am deceived, then there is an I that undergoes the deception. This argument can be attacked, of course. For instance, its premise could be challenged, as Russell does with his claim that Descartes should only be granted that deception is happening, not that he is being deceived. But this only goes to confirm the way in which the Cogito is a true argument and, it seems to me, radically different from the Floating Man.

      I don’t mean to suggest that you were endorsing any strong similarity between these arguments. Thanks for enlightening me about them. It was news to me.

      Liked by 1 person

  7. Pingback: Nightcap | Notes On Liberty

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s