Extended Ignorance

For Embodied, Extended, Ignorant Minds: New Studies on the Nature of Not-Knowing (eds.) S. Arfini & L. Magnani, (Springer). EXTENDED IGNORANCE DUNCAN PRITCHARD University of California, Irvine dhpritch@uci.edu ABSTRACT. According to extended cognition, a subject’s biological cognitive processes can become integrated with features of the subject’s cognitive environment (and thereby ‘extended’), such as technology. When such extended cognitive processes lead to knowledge, the knowledge that results is extended knowledge. This paper explores how the phenomenon of extended cognition relates to the epistemology of ignorance. In particular, our concern will be with not only forms of extended cognition that are epistemically problematic (and which one might be expected to have negative epistemic ramifications) but also forms of extended cognition that are generally epistemically sound. As we will see, some prima facie plausible arguments for how extended cognition might lead to ignorance are not compelling on closer inspection. In addition, some other ways in which extended cognition might lead to ignorance depend on the account of ignorance that one endorses, and in particular whether one opts for a normative or non-normative account of ignorance. 1. EXTENDED COGNITION AND IGNORNACE According to the extended cognition thesis, a subject’s cognitive processes can extend beyond her biological cognitive processes to take in elements of the subject’s environment as proper parts.1 So, for example, in the right conditions one’s cognitive processes, including one’s biological memory, can be integrated with technology that performs a memorial function such that the subject’s 2 memorial cognitive processes are extended, in that they involve both unextended biological components and extended technological components.2 In this way, a technological device, such as a smart phone, might potentially serve a memorial cognitive function, as when one compensates for one’s failing memory by using this device to perform certain kinds of memory tasks (like keeping track of phone numbers).3 The extended cognition thesis represents a radical departure from traditional ways of thinking about cognition. It also raises lots of further questions. For example, if there are extended cognitive processes, then does it follow that there are also extended mental states (this is the closely related extended mind thesis)?4 Also, given that not just any employment of external resources like technology amounts to extended cognition, what does it take to integrate these external resources with a subject’s on-board cognitive processes to lead to extended cognition? Relatedly, if extended cognition is possible, then just how widespread is it? For our purposes we can bracket some of these foundational questions. In what follows, I will take it as given that extended cognition is not only possible, but also that it is increasingly widespread as technology becomes ever more seamlessly integrated into our lives. Our interest is instead in the epistemic ramifications of extended cognition, especially with regard to ignorance (i.e., its negative epistemic ramifications). In particular, does the growth in extended cognition lead to new ways in which we might be ignorant, or at least suggest that we are likely to be more ignorant than before?5 In what follows, I will take it uncontentious that extended cognitive processes can lead to epistemic states like knowledge. That is, one might be sceptical that extended cognitive processes can ever result in bona fide knowledge, but I will be setting this kind of concern to one side for our purposes. Call an extended cognitive process that leads to a true extended belief that satisfies the conditions for knowledge extended knowledge.6 Note that it doesn’t follow from that fact that extended cognitive processes can lead to extended knowledge that they do. In particular, there might be specific reasons why an extended cognitive process might in practice lead to beliefs that have a dubious epistemic pedigree. Accordingly, it will be useful to our discussion to divide our concerns into two categories. On the one hand, there are natural concerns about how epistemically problematic extended cognitive processes (i.e., those that regularly result in false beliefs or at least true beliefs that have a poor epistemic pedigree) can generate ignorance. On the other hand, there are also concerns about whether generally epistemically sound extended cognitive processes (i.e., those that 3 regularly result in true beliefs and extended knowledge) can nonetheless generate interesting forms of ignorance. 2. EPISTEMICALLY SOUND EXTENDED COGNITION I will start by considering how epistemically sound extended cognitive processes might nonetheless lead to ignorance. Let’s imagine that extended cognitive processes develop in such a way that they are generally reliable at producing extended knowledge. It follows that as we become increasingly integrated into highly technological environments, so our acquisition of knowledge (and true belief) also becomes increasingly concerned with extended knowledge (and true belief) rather than knowledge (and true belief) simpliciter. In particular, we will have an ever-expanding body of information available to us which can be directly (albeit extendedly) known via our extended cognitive processes.7 How might this growth of extended knowledge have a bearing on ignorance? On the face of it, after all, any growth in knowledge ought to mitigate against ignorance. I will be considering two possible implications. The first is that the growth of extended knowledge leads to a withering of one’s body of on-board cognitive processes. If so, then could this indicate the emergence of a new kind of ignorance? The second concerns the implications of having a greater body of information epistemically accessible to one via extended cognitive processes. Does this entail that one is more susceptible to a charge of ignorance insofar as one does not epistemically exploit these extended cognitive processes? Let’s take these issues in turn. The first issue can be expressed in terms of the following argument: P1) The more we are reliant on extended cognitive processes, the less we employ our purely on-board cognitive processes. P2) The less we employ our purely on-board cognitive processes, the less epistemically effective they are. P3) The less epistemically effective our purely on-board cognitive processes are, the less we acquire epistemic standings like knowledge and true belief from these sources. C1) The more we are reliant on extended cognitive processes, the less we acquire epistemic standings like knowledge and true belief from our purely on-board cognitive processes. (From P1, P2, P3) P4) Ignorance is to be understood as the absence of an epistemic standing, such as the lack of knowledge or the lack of true belief. C2) The less we acquire epistemic standings like knowledge and true belief from our purely on-board cognitive processes, the more susceptible we are to being ignorant. (From C1, P4) 4 The first premise is contentious because it is not at all clear why a technological cognitive augmentation should thereby lead to a reduction in use of one’s corresponding on-board cognitive processes. One can see the rationale in play: if one can use one’s smart phone to, say, remember phone numbers, then why would one commit phone numbers to one’s biological memory? But that rationale doesn’t really stand up to closer inspection. Sure, one might not need to use one’s biological memory to recall phone numbers anymore, but that doesn’t mean that one doesn’t continue to employ one’s biological memory for other tasks. Accordingly, there’s no inherent reason why one’s use of one’s biological memory as a whole should subside just because certain kinds of memorial tasks have been off-loaded to the extended cognitive process. Of course, (P1) is an empirical claim, so ultimately whether it is true is something that cognitive science will determine. The second premise looks to be in better shape. Again, the rationale in play here is fairly clear: if one doesn’t use one’s biological memory, for example, then this cognitive capacity will atrophy and hence will no longer be as epistemically effective. That certainly looks plausible, at least given that we grant that there is a skill-like dimension to this cognitive capacity, such that it can be trained to be more effective through practice, and hence can be weakened by lack of use. Of course, (P2) is also an empirical claim, and so its truth is something that we would expect cognitive science to determine, but unlike (P1) it at least seems to be antecedently plausible that it is true. In contrast to (P1) and (P2), (P3) looks to be the kind of claim that can assessed a priori. So construed, it also looks straightforwardly true, in that it is hard to see how an epistemically weakened cognitive process could not thereby result in less knowledge or true belief on this whole. Since (P2) also looks compelling, and since the inference to (C1) is clearly valid, this means that what is really carrying the weight of the argument to (C1) is (P1). What about the plausibility of (P4)? It is common in the literature to understand ignorance in terms of the absence of an epistemic standing, with the two most dominant proposals in this regard concerned with either the lack of knowledge or the lack of true belief.8 Accordingly, on standard views of ignorance, (P4) would straightforwardly follow. Interestingly, we are going to be considering an alternative to these popular views of ignorance below. This alternative proposal won’t make any substantive difference to the plausibility of (P4), however, since it effectively imposes a further restriction on conventional accounts of ignorance that understand ignorance in terms of the absence of an epistemic standing (and in that sense is compatible with such accounts), so we can reasonably set this to one side for our current purposes. 5 Perhaps the weakest point in the argument concerns the transition from (P4) to (C2). It might look superficially appealing, in that the lack of epistemic standing at issue in (C1) seems to directly translate, via (P4), into ignorance. But look closer and one can see that there is a lacuna here. For all that really follows from the combination of (C1) and (P4) is that one isn’t gaining epistemic standings to the extent that one did previously from one’s unextended cognitive processes. It doesn’t follow from that claim, however, that one is thereby ignorant of the target propositions, for the simple reason that one might well be gaining (say) knowledge of these propositions from one’s extended cognitive processes. Indeed, this is exactly what we would expect. Take the example of remembering phone numbers that we gave earlier. The shift from using one’s biological memory to employing the technology that forms part of one’s extended cognitive process is precisely a way in which one tends to have more knowledge and true beliefs overall than before, albeit where this expansion in epistemic success is now largely attributable to one’s extended cognitive processes rather than one’s on-board cognitive processes. Accordingly, even if one accepts the contentious (P1), and thus the argument to (C1), there is still a way of blocking the further inference to (C2). The upshot is that one is, if anything, likely to be less ignorant overall than hitherto, rather than more so. Is there a way to rework this argument so that it survives these objections? One possible thought in this regard is that while extended cognition doesn’t necessarily lead to an overall increase in ignorance, it does lead to a specific kind of ignorance, in that one is now more likely to have less knowledge and true belief from one’s on-board cognitive processes than before. In other words, while the growth in extended cognition might lead to more extended true belief and knowledge (and thus less ignorance overall), it leads to less unextended true belief and knowledge, and hence more ignorance of a specific kind that is concerned with one’s unextended epistemic states. The problem, however, is that it is hard to make sense of what this claim amounts to. Take the case of a particular phone number where the subject has extended knowledge (and thus true belief). Is there any coherent way of claiming that the subject is thereby ignorant of this proposition, albeit in a manner that is indexed to a particular type of cognitive process? It’s hard to see how this would work. In general, two subjects who both have knowledge of a proposition are both equally non-ignorant—the manner in which they acquired that knowledge simply doesn’t come into it. For example, consider two subjects who each know a specific proposition, albeit via different routes. So, for instance, they each know that Moriarty is the murderer, but while the first subject (Sherlock) knows this because he witnessed the crime, the second subject (Watson) knows this because he was told this by Sherlock, who he knows to be a reliable informant. Is there any 6 sense in which Watson is ignorant of this proposition that is somehow indexed to first-person experience, or in which Sherlock is ignorant of this proposition that is somehow indexed to testimony? I think the implausibility of this suggestion reveals the oddity of thinking that there is a specific kind of ignorance that is indexed to non-extended cognitive processes.9 3. NORMATIVE AND NON-NORMATIVE CONCEPTIONS OF IGNORANCE This brings us to the second issue we have raised regarding extended cognition and ignorance, which is whether the growth in extended cognition might make one more susceptible to ignorance, in that there is now a wider body of information that one ought to be aware of. What drives this issue is the thought that there is a normative dimension to ignorance, in that to describe someone as ignorant is to do more than merely claim that they lack a certain epistemic standing—such as knowledge or true belief, on the two standard accounts of ignorance (though for convenience I will henceforth focus on the former). The thought is that an ascription of ignorance also entails that this person has exhibited an intellectual failing in this regard, such that to be ignorant is to be unaware of a fact that one ought to be aware of.10 This normative dimension to ignorance would explain why, for example, we wouldn’t describe someone as ignorant for failing to know something that they couldn’t possibly know, either because it is in principle unknowable or because it is practically unknowable, at least for them. For example, it would be odd to treat someone as being ignorant about distant unrecorded facts about the past, such as what Caesar had for breakfast on the day that he crossed the Rubicon. This simply isn’t something that one could realistically know, and if one had a true belief about it, it could only be by virtue of a lucky guess.11 There is thus no intellectual failing on display with regard to one lacking knowledge in this regard. Indeed, not only would one not manifest an intellectual failing by lacking knowledge of this truth, but one would manifest an intellectual failing by trying to find it out (or by groundlessly forming a belief about it). That an attribution of ignorance would be inappropriate in this case, even despite the lack of knowledge, suggests that ignorance has a normative dimension. In particular, where a truth is epistemically inaccessible—i.e., such that there is no feasible route to knowing it¾then one doesn’t count as being ignorant for failing to know it. There are thus truths that nobody knows which aren’t thereby truths that we are ignorant of. The more interesting kind of case, for our purposes, however, are truths where a lack of knowledge 7 for one person might count as ignorance even though that same lack of knowledge in another person wouldn’t count as ignorance. I think there are a number of factors that can lead to this kind of relativity of ignorance. I will here consider three: the practical importance of the truth in question for that person, the epistemic availability of that truth for that person, and whether that person elects to form a belief about the proposition in question. Consider first how the practical importance of a truth can be relevant to whether lacking knowledge of it amounts to ignorance. Notice that we don’t attribute ignorance to subjects who fail to know truths that are completely trivial, such that there would be no good reason for them to know them, even when those facts are epistemically available to the subject. For example, one could determine what every café in town has on their menu if one were so inclined—there is no great practical barrier to this knowledge, beyond the time involved—but what would be the point? There are far better things to be spending one’s time doing, including better things to be devoted to knowing. This is why we wouldn’t ordinarily attribute ignorance to someone for failing to know what all the cafes in town have on their menus, as there is no intellectual failing involved in such a lack of knowledge (indeed, there would be an intellectual failing involved in them seeking out such knowledge). They are unaware of the relevant facts, but they are not thereby ignorant of them, as these are not facts that that they ought to be aware of. What would be pointless knowledge for most people wouldn’t necessarily be pointless knowledge for everyone, however. We can imagine jobs that require someone to know things such as what is on café menus (perhaps they are tasked with compiling a central list for the tourism board). In this case, however, a failure to possess the relevant knowledge could be deemed ignorant—one could imagine our subject’s boss, for example, exclaiming that she can’t believe that her employee is ignorant of something so important to their position. Thus the practical importance of a truth can be such that one person’s failure to know it doesn’t amount to ignorance but another person’s failure to know it does. Now consider how the epistemic availability of truths can have a bearing on ignorance. We’ve already noted that some truths are epistemically inaccessible, which means that they aren’t epistemically available to anyone (such as what Caesar had for breakfast on a certain day). Accordingly, we don’t regard someone’s failure to know them as ignorance. Interestingly, however, there is a kind of epistemic availability that can be variable depending on the individual at issue. For example, we don’t tend to call small children ignorant for failing to know such things as who the President of France is, even though we would call their parents ignorant for being in the 8 corresponding situation. The crux of the matter is that small children aren’t in a position to know these things, and so there is no intellectual failing on their part in their not knowing them. What goes for small children applies more generally to anyone whose epistemic access to the relevant facts is similarly restricted. We do not normally describe distant civilizations as being ignorant of truths known only to the contemporary world, such as claims about quantum physics, for example. To do so would imply an intellectual failing on their part, but that would be improper, given that they simply weren’t in a position to know these things. Interestingly, there is an asymmetry in how we treat ancient civilizations on this score that is relevant here. We do not treat them as being ignorant of the important reference points of our world (our science, for example), but we are far more inclined to treat us as being ignorant of the important reference points of their world. The normative dimension to ignorance can account for this, in that it is far more plausible that our lack of knowledge reveals an intellectual failing on our part (lack of curiosity, for instance), than that their corresponding lack of knowledge reveals an intellectual failing on their part.12 The third kind of variability concerns whether a subject elects to form a belief about a proposition. At any one time, there will be an unlimited range of truths that are, quite rightly, of no great practical importance to one, and as a result one doesn’t form any belief about them. Consider current showbiz trivia, for example. If this isn’t important to me (in contrast, say, to a showbiz gossip columnist), then there is no reason why I should form any belief about whether, say, Jo-Lo and Ben Affleck are back together (which is a live showbiz issue at the time of writing). If I do become aware, from a reliable source, that they are now an item again, and so form a true belief to that effect, then I can obviously come to know what I believe. Crucially, however, if I never come to have a belief in this regard, and so lack this knowledge, then that wouldn’t make me ignorant, as there is no reasonable expectation that I should form any belief about this subject matter. A wide range of truths are like that: they are epistemically accessible to one, but since there is no reasonable expectation that one should know them (even though there might be an expectation that others know them), one doesn’t count as being ignorant as a result of failing to form a belief about them. Interestingly, however, if one does form a belief about this subject matter, even in circumstances where there is no reasonable expectation that one should, and one’s belief is false, then a charge of being ignorant of this fact would become appropriate. This may not be a truth that one ought to know simpliciter, but it is a truth that one ought to know insofar as one forms a belief about it. One can thus count as ignorant by forming a false belief it even though one wouldn’t count as 9 ignorant if one simply opted to form no view about the target subject matter. Whether one counts as ignorant can thus depend on whether one forms a belief about the target proposition. The upshot of the foregoing is that whether a subject’s lack of knowledge amounts to ignorance depends on whether they have exhibited the relevant intellectual failing, and that this can depend, in turn, on several variable factors in play, such as the practical importance of the proposition at issue, the epistemic availability of the knowledge that is lacking, and whether the subject elects to form a belief about the target proposition (and possibly other factors too). It is the relationship between the epistemic availability of the truths in question and ignorance that is most relevant for our current purposes. It is relevant because of how the proliferation of extended cognitive processes makes a wider range of knowledge epistemically accessible. On the plus side, insofar as extended knowledge is bona fide knowledge, then we are in a position to know a great deal more than we did previously, as the technology enables to us to immediately know facts that might have hitherto required a lengthy inquiry (indeed, which may have been, practically speaking, beyond one’s grasp). On the negative side, however, this wider epistemic availability of truths entails that we are exposed to a far greater scope of ignorance, in that in lacking knowledge of truths we can now be charged with being ignorant of them even though previously a lack of knowledge would not have entailed ignorance. We can express the argument in play here as follows: P1*) P2*) C*) The proliferation of extended cognitive processes makes a much wider range of facts epistemically accessible to subjects than would have been epistemically accessible previously. In order for a subject to count as ignorant for lacking knowledge, the relevant truth must be epistemically accessible to that subject. The proliferation of extended cognitive processes doesn’t only lead to an expansion of (extended) knowledge, but also to an expansion of (extended) ignorance. (From P1*, P2*) The first premise, (P1*), seem relatively uncontroversial. One could argue that the contemporary expansion of epistemically accessible facts has more to do with the general availability of technology—in particular, access to the internet—than it does with extended cognitive processes specifically. Nonetheless, it is surely the case that the proliferation of extended cognitive processes would have a significant bearing of the range of truths epistemically accessible to the subject. The second premise, (P2*), is just a statement of the normative account of ignorance. The weak point in the argument doesn’t lie in the premises, however, but rather concerns the fact that the conclusion doesn’t follow from these premises. 10 All that actually follows from these premises is that there is a greater range of facts that we can be held to be ignorant of as a result of the proliferation of extended cognitive processes. But this doesn’t entail that we are more ignorant, or even that we are likely to be more ignorant, as it is consistent with this claim that this wider class of epistemically accessible facts is not only epistemically accessible, but also epistemically accessed, in the sense of being known. Indeed, we would expect this to be the case, given that the extended cognitive processes are doing their job. In particular, there is a dilemma in play here. Either one’s extended cognitive processes are functioning as they ought—and remember in this regard that we are currently working on the assumption that extended cognitive processes are generally epistemically sound¾in which case it is not only the case that a wider range of facts is epistemically accessible, but also that these facts are known. But then the subject isn’t ignorant of these facts. Or else they are not functioning as they ought, and hence not leading to extended knowledge, but then the facts weren’t being made epistemically accessible via the extended cognitive processes. In that case, however, then insofar as a charge of ignorance would have been inappropriate previously due to a lack of epistemic access to the facts in question, then it would still be inappropriate. Either way, then, there is no growth in ignorance as a result of a proliferation of extended cognitive processes. 4. EPISTEMICALLY PROBLEMATIC EXTENDED COGNITION Thus far we have been focusing our attention on whether generally epistemically sound extended cognitive processes—i.e., extended cognitive processes that regularly result in true beliefs and extended knowledge¾might lead to ignorance. The arguments we have considered for how this might occur have been found to be inconclusive. But what about epistemically problematic extended cognitive processes—i.e., extended cognitive processes that don’t generally lead to true beliefs and extended knowledge? Here the prospects are, unsurprisingly, more promising, though as we will see the details tend to depend on the account of ignorance in play. One straightforward kind of case where this might occur would be concerning propositions that we would have hitherto tended to form true beliefs in, and thereby gain knowledge of, but where the introduction of an extended cognitive process leads one to form false beliefs. Perhaps, for example, one’s biological memory was generally reliable, and so a good source of memorial knowledge, but the extended memorial cognitive process that has been introduced is in fact highly 11 unreliable (it is designed to manufacture false memories, say), and hence is not a good source of memorial knowledge. Now, it seems, we can be ignorant of truths that we would have hitherto known (and so not been ignorant of). Notice that if one subscribes to the normative account of ignorance described above, then this might potentially insulate subjects from a charge of ignorance of this kind. It rather depends on how the extended cognitive process is brought about. If, for example, the subject is deceived about the unreliability of the extended cognitive process, and has no reasonable way of spotting that it is epistemically defective, then that should absolve the subject of a charge of ignorance. For while they now lack knowledge they would have had previously, the truths in question are no longer epistemically accessible to the subject. In short, the epistemically problematic nature of the extended cognitive process in play has the reverse effect to the one noted above, in that it is reducing rather than expanding the scope of truths that one can count as being ignorant of. Whether this kind of case is to be understood as extending ignorance thus depends on whether one opts for either the normative account of ignorance or one of the conventional, non-normative accounts of ignorance. A second kind of scenario that is interesting for our purposes is when subjects are led, via the extended cognitive process, to forming beliefs about topics that they would never have (quite rightly) previously considered. Imagine now that those extended cognitive processes are also epistemically defective, in that, for example, they are leading us to systematically form false beliefs in this regard. Perhaps, for example, advertisers are able to influence our extended cognitive processes to prompt us to form false beliefs that are useful to them, such as regarding how a particular toothpaste is far superior to all its competitors (a topic that most reasonable subjects would have previously been unlikely to form any beliefs about).13 Now there are truths that one fails to know because one’s beliefs are epistemically defective (in this case because they are false) where previously one failed to know them because, quite reasonably, one didn’t form any belief about them at all. According to conventional, non-normative accounts of ignorance, this kind of case doesn’t concern an increase in ignorance, as the fact that the subject didn’t have knowledge (say) of the target proposition previously would entail that she was already ignorant of it. On the normative account of ignorance, in contrast, this could potentially be a case where ignorance is being created by the extended cognitive process, given that the agent wasn’t previously ignorant in failing to believe, and thus failing to know the target proposition. Everything now depends on whether, on this account, the unreliable extended cognitive processes is making us more ignorant, by ensuring that our lack of knowledge now qualifies as ignorance even though it didn’t previously. 12 Whether this is so will again depend on the details of the case and, in particular, on whether one has sufficient responsibility for the error. For example, on one version of this scenario, the extended cognitive processes lead us to form false beliefs in much the same way that, say, advertisements in general can lead us to form false beliefs—that is, they apply techniques that will generally lead to these beliefs being formed, but such techniques are resistible if one is sufficiently alert to them. If that’s how we are to think of the extended cognitive processes in play, however, then their proliferation should make responsible believers be more cautious about forming beliefs in certain subject matters, just as they would be in ordinary cases involving the testimony of advertisers. Accordingly, if they fail to exercise caution, and so are led to form false beliefs as a result, then they will now count as ignorant where they would previously have only counted as lacking knowledge (because they didn’t even form the target belief). In contrast, if the extended cognitive process is not resistible in this fashion—if it functions more like a kind of indoctrination, say—then I think that would tend to exclude the agent from the charge of being ignorant, as they would not be culpable for their lack of knowledge. All that would have changed is the reason why they fail to have knowledge, in that it is now because they have a false belief in the target proposition rather than no belief at all. This discussion is relevant to one specific way in which extended cognitive processes might generate ignorance, which is in terms of making us be more confident in our beliefs about certain subject matters than we should be. This has been an observed consequence of our internet access to widespread information, as people tend to be misled by their superficial grasp of a domain into making confident judgments about it.14 If internet access has this effect, then we would expect something similar to occur, perhaps even in a stronger form (as the informational input would feel more like their own knowledge), when subjects have access to the same information via extended cognitive processes. Such overconfidence is, however, clearly epistemically problematic, in that it will lead subjects to form false beliefs about topics that they would previously have demurred from forming a judgement about (since they were aware of their lack of expertise). Moreover, unlike the example given above concerning the advertisers, there need be no deceit or manipulation involved, as what is causing the overconfidence is rather just human fallibility and cognitive bias in the context of extended cognitive processes. This kind of case looks to be quite a straightforward way that extended cognition might lead to ignorance. Notice, however, that the conventional, non-normative accounts of ignorance would struggle to accommodate this claim. After all, if the agent previously lacked a belief in the target 13 propositions (and hence lacked knowledge of them), then it follows that they were already ignorant of them. In contrast, on the normative account of ignorance at least, it is possible that there was no ignorance involved in the subject lacking a belief in this regard—perhaps this simply is not something that they ought to know, and thus have a belief about. Hence, in cases in which that condition is met, such that the subject wasn’t previously ignorant of the target propositions, then this could constitute a scenario where ignorance is being generated via extended cognitive processes. Whether this is so would depend, on the normative account of ignorance, on whether the subject is culpable for the false beliefs that she is forming. Crucially, however, given that such overconfidence is a cognitive flaw on the part of the individual, then we would tend to hold the subject responsible in this regard. They ought to be more careful in how they form their beliefs to correct for this overconfidence. Accordingly, a charge of ignorance is appropriate, just as we treat subjects who, as a result of exposure to the information on the internet, overconfidently form false beliefs. 5. CONCLUDING REMARKS We have examined a range of ways in which extended cognitive processes, both epistemically sound and epistemically problematic, can lead to ignorance. More specifically, we have considered two arguments for how epistemically sound extended cognitive processes might generate ignorance, and found that they are not promising on closer inspection. Moreover, while we have seen that epistemically problematic cognitive processes can lead to ignorance, the exact manner in which this occurs can depend on the account of ignorance in play, and in particular on whether one opts for a conventional, non-normative account of ignorance in terms of the absence of a target epistemic standing (such as knowledge or true belief) or a normative account that in addition regards ignorance as involving an intellectual failing. Of course, we have not considered all the myriad ways that extended cognitive processes might lead to ignorance, but the upshot of the foregoing is to make at least a prima face case for thinking that whether the widespread use of extended cognitive processes leads to the proliferation of ignorance turns on some quite specific details about how these processes are employed (e.g., where they have a tendency to make us overconfident in our judgements, and hence prone to error).15 14 REFERENCES Adams, F., & Aizawa, K. (2001). ‘The Bounds of Cognition’, Philosophical Psychology 14, 43-64. —— (2008). The Bounds of Cognition, Blackwell, Oxford. Allen-Hermanson, S. (2013). ‘Superdupersizing the Mind: Extended Cognition and the Persistence of Cognitive Bloat’, Philosophical Studies 164, 791-806. Arfini, S. (2021). ‘Situated Ignorance: The Distribution and Extension of Ignorance in Cognitive Niches’, Synthese 198, 4079-95. Bjerring, J. C., & Pedersen, N. J. L. L. (2014). ‘All the (Many, Many) Things we Know: Extended Knowledge’, Philosophical Issues 24, 24-38. Carter, J. A., & Gordon, E. (2020). ‘Is Searching the Internet Making us Intellectually Arrogant?’, Polarisation, Arrogance, and Dogmatism: Philosophical Perspectives, (eds.) A. Tanesini & M. P. Lynch, 88-103, London: Routledge. Carter, J. A., Kallestrup, J., Palermos, S. O., & Pritchard, D. H. (2014). ‘Varieties of Externalism’, Philosophical Issues 24, 63-109. ¾¾ (eds.), (2018a). Extended Epistemology, Oxford: Oxford University Press. ¾¾ (eds.), (2018b). Socially Extended Epistemology, Oxford: Oxford University Press. Carter, J. A., & Pritchard, D. H. (2019). ‘The Epistemology of Cognitive Enhancement’, Journal of Medicine & Philosophy 44, 220-42. Chemero, T. (2009). Radical Embodied Cognitive Science, Cambridge, MA: MIT Press. Clark, A., (2008). Supersizing the Mind: Embodiment, Action, and Cognitive Extension, Oxford: Oxford University Press. —— (2013). ‘Gesture as Thought’, The Hand, an Organ of the Mind: What the Manual Tells the Mental, (ed.) Z. Radman, 255-68, Cambridge, MA: MIT Press. ¾¾ (2015). ‘What ‘Extended Me’ Knows’, Synthese 192, 3757-75. Clark, A., & Chalmers, D. (1998). ‘The Extended Mind’, Analysis 58, 7-19. DeNicola, D. R. (2018). Understanding Ignorance: The Surprising Impact of What We Don’t Know, Cambridge, MA: MIT Press. El Kassar, N. (2018). ‘What Ignorance Really Is: Examining the Foundations of Epistemology of Ignorance’, Social Epistemology 32, 300-10. Fisher, M., Goddu, M. K., & Keil, F. C. (2015). ‘Searching for Explanations: How the Internet Inflates Estimates of Internal Knowledge’, Journal of Experimental Psychology: General 144, 67487. Fricker, M. (2016). ‘Epistemic Injustice and the Preservation of Ignorance’, The Epistemic Dimensions of Ignorance, (eds.) R. Peels & M. Blaauw, 144-59, Cambridge: Cambridge University Press. Gallagher, S. (2013). ‘The Socially Extended Mind’, Cognitive Systems Research 2, 4-12. Giere, R. (2010). Scientific Perspectivism, Chicago IL: University of Chicago Press. Heersmink, R., & Knight, S. (2018). ‘Distributed Learning: Educating and Assessing Extended Minds’, Philosophical Psychology 31, 969-90. Huebner, B. (2016). ‘Transactive Memory Reconstructed: Rethinking Wegner’s Research Program’, Southern Journal of Philosophy 54, 48-69. Hutchins, E. (1995). Cognition in the Wild, Cambridge, MA: MIT Press. Kotzee, B. (2018). ‘Cyborgs, Knowledge and Credit for Learning’, Extended Epistemology, (eds.) J. A. Carter, A. Clark, J. Kallestrup, S. O. Palermos, & D. H. Pritchard, 221-38, Oxford: Oxford University Press. Krueger, J. (2011). ‘Extended Cognition and the Space of Social Interaction’, Consciousness and Cognition 20, 643-57. 15 Le Morvan, P. (2011a). ‘Knowledge, Ignorance and True Belief’, Theoria 77, 32-41. ¾¾ (2011b). ‘On Ignorance: A Reply to Peels’, Philosophia 39, 335-44. ¾¾ (2012). ‘On Ignorance: A Vindication of the Standard View’, Philosophia 40, 379-93. ¾¾ (2013). ‘Why the Standard Conception of Ignorance Prevails’, Philosophia 41, 239-56. Le Morvan, P., & Peels, R. (2016). ‘The Nature of Ignorance: Two Views’, The Epistemic Dimensions of Ignorance, (eds.) R. Peels & M. Blaauw, 12-32, Cambridge: Cambridge University Press. Ludwig, D. (2015). ‘Extended Cognition and the Explosion of Knowledge’, Philosophical Psychology 28, 355-68. Lynch, M. P. (2014). ‘Neuromedia, Extended Knowledge, and Understanding’, Philosophical Issues 24, 299-313. Medina, J. (2017). ‘Epistemic Injustice and Epistemologies of Ignorance’, Routledge Companion to the Philosophy of Race, (eds.) P. C. Taylor, L. M. Alcoff & L. Anderson, 247-60, London: Routledge. Mills, C. (2007). ‘White Ignorance’, Race and Epistemologies of Ignorance, (eds.) S. Sullivan & N. Tuana, 13-38, Albany, NY: SUNY Press. Noë, A. (2004). Action in Perception, Cambridge, MA: MIT Press. Nottelmann, N. (2015). ‘Ignorance’, Cambridge Dictionary of Philosophy, 3rd edn., (ed.) R. Audi, 497-98, Cambridge: Cambridge University Press. Palermos, S. O. (2014). ‘Knowledge and Cognitive Integration’, Synthese 191, 1931-51. Palermos, S. O., & Pritchard, D. H. (2016). ‘The Distribution of Epistemic Agency’, Social Epistemology and Epistemic Agency: De-Centralizing Epistemic Agency, (ed.) P. Reider, 109-26, Lanham, MD: Rowman & Littlefield. Pauw, W. T. J. L., de Nooijer, J. A., & Paas, F. (2014). ‘Toward a More Embedded/Extended Perspective on the Cognitive Function of Gestures’, Frontiers in Psychology 5, 359. Peels, R. (2010). ‘What is Ignorance?’, Philosophia 38, 57-67. ¾¾ (2011). ‘Ignorance Is Lack of True Belief: A Rejoinder to Le Morvan’, Philosophia 39, 344-55. ¾¾ (2012). ‘The New View on Ignorance Undefeated’, Philosophia 40, 741-50. Peels, R., & Pritchard, D. H. (2020). ‘Educating For Ignorance’, Synthese (Online First, DOI: https://doi.org/10.1007/s11229-020-02544-z]. Pritchard, D. H (2010). ‘Cognitive Ability and the Extended Cognition Thesis’, Synthese 175, 133-51. —— (2014). ‘Virtue Epistemology, Extended Cognition, and the Epistemology of Education’, Universitas: Monthly Review of Philosophy and Culture 478 (2014), 47-66. ¾¾ (2015). Epistemic Angst: Radical Skepticism and the Groundlessness of Our Believing, Princeton, NJ: Princeton University Press. —— (2016). ‘Intellectual Virtue, Extended Cognition, and the Epistemology of Education’, Intellectual Virtues and Education: Essays in Applied Virtue Epistemology, (ed.) J. Baehr, 113-27, London: Routledge. —— (2018a). ‘Extended Virtue Epistemology’, Inquiry 61, 632-47. —— (2018b). ‘Neuromedia and the Epistemology of Education’, Metaphilosophy 49, 328-49. ¾¾ (2019). ‘Extended Knowledge’, Extended Epistemology, (eds.) J. A. Carter, A. Clark, J. Kallestrup, S. O. Palermos & D. H. Pritchard, 90-104, Oxford: Oxford University Press. ¾¾ (2021a). ‘Good News, Bad News, Fake News’, Epistemology of Fake News, (eds.) S. Bernecker, A. Flowerre & T. Grundman, 46-67, Oxford: Oxford University Press ¾¾ (2021b). ‘Ignorance and Inquiry’, American Philosophical Quarterly 58, 111-23. ¾¾ (Forthcominga). ‘Extended Social Epistemology’, Oxford Handbook of Social Epistemology, (eds.) J. Lackey & A. McGlynn, Oxford: Oxford University Press. 16 ¾¾ (Forthcomingb). ‘Ignorance and Normativity’, Philosophical Topics. ¾¾ (Forthcomingc). ‘Omniscience and Ignorance’, Veritas. Pritchard, D. H., English, A., & Ravenscroft, J. (2021). ‘Extended Cognition, Assistive Technology and Education’, Synthese (Online First, DOI: 10.1007/s11229-021-03166-9). Rowlands, M. (2009). The Body in Mind, New York: Cambridge University Press. Rupert, R. D. (2004). ‘Challenges to the Hypothesis of Extended Cognition’, Journal of Philosophy 101, 389-428. —— (2009). Cognitive Systems and the Extended Mind, Oxford: Oxford University Press. Shapiro, L. (2011). Embodied Cognition, New York: Routledge. Smart, P. (2017). ‘Extended Cognition and the Internet: A Review of Current Issues and Controversies’, Philosophy & Technology 30, 357-90. Sprevak, M. (2009). ‘Extended Cognition and Functionalism’, Journal of Philosophy 106, 503-27. —— (2019). ‘Extended Cognition’, Routledge Encyclopedia of Philosophy Online, (ed.) T. Crane, Routledge: London. Tollefsen, D. (2006). ‘From Extended Mind to Collective Mind’, Cognitive Systems Research 7, 140-50. van Woudenberg, R. (2009). ‘Ignorance and Force: Two Excusing Conditions for False Beliefs’, American Philosophical Quarterly 46, 373-86. Wheeler M. (2013). ‘Is Cognition Embedded or Extended? The Case of Gestures’, The Hand, an Organ of the Mind: What the Manual Tells the Mental, (ed.) Z. Radman, 269-301, Cambridge, MA: MIT Press. Zimmerman, M. J. (2008). Living with Uncertainty: The Moral Significance of Ignorance, Cambridge: Cambridge University Press. 17 NOTES For the canonical defence of the extended cognition thesis, see Clark & Chalmers (1998). See also Clark (2008). Note that we are here focusing on cases of extended cognition that involve the contribution of technological factors. There are two other forms of extended cognition of note in this regard. The first concerns the possible contribution of a subject’s embodied processes to cognition, what is known as embodied cognition. For some recent discussion of this thesis, see Noë (2004), Chemero (2009), Rowlands (2009), and Shapiro (2011). Consider, for example, how the use of gestures can have a bearing on a subject’s cognition processes, as discussed in Clark (2013), Wheeler (2013), and Pauw et al (2014). The second concerns the possible contribution of other cognitive subjects within a subject’s cognitive processes, what is known as socially extended cognition (this is also known as distributed cognition). For discussion, see Hutchins (1995), Tollefsen (2006), Krueger (2011), and Gallagher (2013). Finally, note that technological extended cognition needn’t involve technology that is external to the skin of the cognitive subject, as one can imagine cases where the technology in question is directly embedded within the biological subject (and hence is to this extent ‘internal’ to the subject, even though it is ‘external’ to her on-board cognitive processes). This is so-called neuromedia. For discussion, see Lynch (2014) and Pritchard (2018b). See also Carter & Pritchard (2019), which discusses the general question of the epistemology of cognitive augmentation. 3 The famous case of extended cognition (and extended mind—see endnote 4) given in Clark & Chalmers (1998) is precisely of this kind, whereby a subject (‘Otto’) with memory-loss caused by Alzheimer’s disease uses a notebook to supplement his biological memorial cognitive processes. 4 Interestingly, it was the extended mind thesis that was originally proposed in the famous paper by Clark & Chalmers (1998), with the extended cognition thesis merely noted as part of the supporting case. There are at least arguably distinct theses, however, in that in endorsing the latter one doesn’t seem to thereby be committed to endorsing the former (even if these two theses might naturally be thought to go together). See Giere (2010) and Goldberg (2012) for discussion of why one might accept the extended cognition thesis while rejecting the extended mind thesis. For a very useful recent overview of the extended mind/cognition theses, and how best to understand them, see Sprevak (2019). 5 Despite the extensive discussion of the epistemological ramifications of extended cognition in the theoretical literature, this is, surprisingly, an under-explored topic, at least to the extent that it engages with contemporary work on the epistemology of ignorance. See, however, Arfini (2001) for a recent discussion of a slightly different question, which is the extent to which we can understand ignorance itself as an extended epistemic standing. 6 For further discussion of extended knowledge (including the socially extended knowledge that arises from socially extended cognition—see endnote 2), see Pritchard (2010; 2018a; 2019; forthcominga), Carter et al (2014; 2018a; 2018b), Clark (2015), Palermos (2014), and Palermos & Pritchard (2016). See also the burgeoning literature examining the relevance of extended knowledge (and thus extended cognition) to educational contexts: Pritchard (2014; 2016; 2018b), Heersmink & Knight (2018), Kotzee (2018), and Pritchard, English & Ravenscroft (2021). 7 This has been viewed as a problem for extended cognition and, thereby, for extended knowledge too, in that it entails that one’s cognitive (and thus epistemic) reach goes far beyond what we might otherwise have expected—this is what is known as cognitive (and epistemic) ‘bloat’. For discussion of the problem of cognitive bloat, see Rupert (2004) and Allen-Hermanson (2013). For discussion of epistemic bloat, specifically, see Bjerring & Pedersen (2014), Ludwig (2015), and Smart (2017). 8 For some key defences of the so-called ‘standard view’ (in virtue of its widespread adoption) of ignorance in terms of lack of knowledge, see Zimmerman (2008), Le Morvan (2011; 2012; 2013), and DeNicola (2017). For some of the main defences of the so-called ‘new view’ of ignorance as the lack of true belief, see Goldman & Olsson (2009), van Woudenberg (2009), and Peels (2010; cf. Peels 2011; 2012). Of course, one can imagine distinct accounts of ignorance in terms of a lack of an epistemic standing that diverge from the standard or new views (such as ignorance as the absence of justified true belief, say, or the absence of an awareness of the target fact), but for convenience we will henceforth follow convention and focus on these two epistemic standings. These two main contemporary accounts of ignorance are usefully surveyed in Le Morvan & Peels (2016). Note that we are here focusing on propositional forms of ignorance¾see Nottelman (2015) and El Kassar (2018) for two recent discussions of non-propositional forms of ignorance and how they relate to the propositional variety. There is a much broader contemporary literature on ignorance that extends beyond the question of how we should define ignorance, such as the political implications of ignorance (e.g., Medina 2017), though usually such work, if it takes a stance on the nature of ignorance at all, commits to a particular definition of the term that conforms to the idea that it picks out the absence of an epistemic standing. See, for example, Fricker (2016, 144), who endorses the standard view of ignorance, and Mills (2007, 16), who endorses the new view of ignorance. 9 One might think that there are other kinds of case that are more amenable to the claim in play. For example, consider two subjects who know a mathematical theorem in very different ways: the first subject knows it because she proved it, while the second subject knows it only via testimony (and lacks the means of proving it). Isn’t there something that the 1 2 18 second subject is ignorant of that the first subject is not ignorant of? Yes, but notice that this ignorance doesn’t concern the target proposition, as neither of them are ignorant of that. It rather concerns further propositions that are concerned with the proof in question (such as what inferential principles are involved in the proof). In the same way, there might be propositions related to the target proposition that Watson fails to know (/truly believe), but which Sherlock knows (/truly believes), on account of the fact that his knowledge of the target proposition is testimonial. But that would be irrelevant to the point in hand, as it doesn’t show that he exbibits a kind of ignorance as regards that specific proposition. In any case, these scenarios where there is a possible wider ignorance in play are not relevant to extended knowledge/true belief anyway, as we’ve been given no reason for supposing that knowledge/true belief acquired via extended cognitive processes is likely to be associated with this form of wider ignorance when compared with the corresponding knowledge/true belief acquired via non-extended cognitive processes. 10 See Pritchard (2021b; forthcomingb) for a development of a view of this sort. See also Pritchard (forthcomingc) for the application of this proposal to debates regarding God’s omniscience. Notice that this proposal is in a certain way orthogonal to the conventional accounts of ignorance in the literature that understand it in terms of the absence of a particular epistemic standing, as one could supplement this normative component to whichever epistemic standing is at issue in a conventional account of ignorance (one would just need to fill in the placeholder notion of what it is to ‘aware’ of the target fact accordingly). 11 Indeed, it is far from clear that one can form a belief via guesswork—at least in the sense of belief that is relevant to epistemology anyway (i.e., that propositional attitude that is a constituent part of rationally grounded knowledge)—as to believe a proposition in this sense is to believe that proposition to be true. If that’s right, then one cannot form a belief (in the specific sense just described at least) about a proposition where one is aware that one has no basis for regarding the target proposition as true. 12 An interesting case is that of contemporaneous people who deliberately elect, perhaps with good reason, to epistemically cut themselves off from the wider world. For example, the Sentinelese tribe who inhabit North Sentinel Island in the Bay of Bengal are the most isolated group of people in the world, who avoid almost all contact with the outside world. As a consequence, there will be a wide range of propositions that are epistemically inaccessible to them, such as much of modern science. But would we say that these tribespeople are ignorant of these facts? That sounds improper to my ear. Even so, one might argue that this case is disanalogous with the ones just considered, in that the people in question have chosen to epistemically cut themselves off. In general, after all, one cannot avoid the charge of ignorance on the grounds that one chooses to not know something. I think it matters, however, what kinds of reasons one has for choosing to not know certain facts. As I’ve argued elsewhere, for example¾see Pritchard (forthcomingb; forthcomingc)¾if one has good reasons for not knowing something (as, for example, when a lawyer refuses to consider trial evidence that has been released them in error), then this would not count as ignorance. My instinct is that ‘good reasons’ here could include such considerations as a shared cultural objective to preserve one’s traditional customs by shunning contact with the outside world, though obviously this issue extends well beyond our current concerns. 13 Another more disturbing kind of case could be fake news—that is, roughly, information that is presented as news, and yet designed to be systematically misleading (e.g., Pritchard 2021a)¾that is delivered via an extended cognitive process. 14 In particular, it seems that subjects conflate the material that they read online with information that they have previously learnt and understood, and this leads them to be overconfident in their judgements. See, Fisher, Goddu & Keil (2015) and Carter & Gordon (2020). 15 I am grateful to Selene Arfini and Lorenzo Magnani who provided detailed comments on an earlier version of this paper.