1. Introduction
I aim here to revive and refine an old but enduringly relevant idea. Like an heirloom plant properly tended and pruned, it flowers into a plausible contemporary theory on which knowledge is secure true belief.
We find its seed in Plato's Meno. There Socrates argued that knowledge is more valuable than true belief (or true opinion, orthê doxa). Deploying a memorable analogy, he compared true beliefs or opinions to the statues of Daedelus reputed to run away unless fastened or secured. These, Socrates averred, illustrated ‘the nature of true opinions: while they abide with us they are beautiful and fruitful, but they run away out of the human soul, and do not remain long’.Footnote 1 Knowledge, by contrast, ‘is more honourable and excellent than true opinion, because fastened by a chain’.Footnote 2 Plato depicted Socrates as entertaining the notion that knowledge is true belief ‘chained or secured’ by a logos (account or explanation) – orthê doxa meta logou. Footnote 3
That knowledge is true belief secured by a logos is not the thesis I defend here. The one I defend requires disentangling two Platonic (or at least inspired by Plato) ideas: (i) that knowledge is secure true belief, and (ii) that knowledge is true belief secured by a logos. Arguably (ii) counts as a particular species of (i). My plan is to articulate and defend (i), but not (ii).
I proceed as follows. In section 2, I make some preliminary remarks concerning the theory to be articulated and defended here which we may call ‘the Security Theory of Knowledge’ (hereafter simply ‘the Security Theory’). I explain it in section 3, argue for it in section 4, defend it against some important objections in section 5, and conclude with some retrospective and prospective remarks in section 6.
2. Some Preliminaries
Though inspired by Plato, this paper will not provide a general defense of Platonic views concerning knowledge and reality. Amicus Plato, sed magis amica veritas. For instance, accepting the Security Theory in no way commits one to the Platonic doctrines that (a) all knowledge is anamnesis or recollection, and (b) knowledge and belief have different objects: knowledge is only of the Forms, whereas belief is only of illusory appearances. I reject both (a) and (b), but do not argue for that here.
I also make two key presuppositions regarding factive knowledge (hereafter simply ‘knowledge’).Footnote 4 These presuppositions square with the views of the overwhelming majority of contemporary epistemologists as evidenced by the literature.Footnote 5 The first is that knowledge is at least true belief: knowledge that p is a species of the genus believing that p, and one knows that p only if p is true. The second is that knowledge is more than true belief: belief and truth are necessary but not sufficient for knowledge. In virtue of these presuppositions, the Security Theory can, despite its venerable Platonic inspiration, be aptly classified as a contemporary theory of knowledge.
3. The Theory Explained
In a nutshell, the Security Theory holds that knowledge is secure true belief. This requires some explanation. We may begin with the relatively more straightforward notions of belief and truth, and continue onto the more metaphorical notion of security.
3.1. Belief
Beliefs represent the world as being a certain way, and differ from guesses in requiring conviction in a proposition's truth. Since conviction comes in degrees, so too can belief: although one can both believe that p and that q, one can believe that p more than one believes that q. In any case, to believe something, one must have more conviction in its truth than in its falsehood. Once acquired, beliefs exist across time, sometimes rising to consciousness, but usually existing dispositionally. Call the time over which a belief exists its lifespan.
3.2. Truth
Truth will be understood here in terms of the Aristotelian notion of things being as one believes them to be, or not being as one believes them not to be. For instance, the belief that the Caspian Sea is larger than the Sea of Galilee is true if and only if the Caspian Sea is larger than the Sea of Galilee. The Caspian Sea's being larger than the Sea of Galilee is the truth-maker of this belief.
3.3. Security
A true belief's security will be understood in terms of two interrelated notions: being firmly fixed, and being protected from danger. While we will consider both notions, it will be helpful to first do so independently of their specific application to true belief.
Security as being firmly fixed may be illustrated as follows. Think of a pen intended for public use that is affixed to a desk or counter by being tied with a string. Its being so tied generally protects the pen from being taken. Presumably, someone who really wanted that pen could try to rip it away or cut the string, but its being tied this way provides it some protection from theft. As a second example, consider anchors. A temporary anchor or a permanent one (as in a mooring) helps fixate a vessel and protect it from drifting away or being blown away from a certain expanse of water. Depending on the size and type of the vessel, the weather, and the (sea, lake or river) bed in question, some anchor designs are more effective than others. At any rate, security as fixation can be understood as a species of security as protection, in this case as protection from something's moving (or being moved) from where one wants it to remain.
We turn now to the broader notion of security as protection. It can be understood as preservation from, and relative to, dangers or threats. Take home security. Installing deadbolts on the doors of your house may protect against someone's forcing them open, but not against someone's entering through a window. Installing iron bars across the windows may protect against this, but not against someone boring through a wall. Digging a moat around your house may protect against this, but not against attack by aircraft. An anti-aircraft battery on your roof may protect against this, but not against a ballistic missile attack. You get the idea: there is no absolute home security in the sense of security against any and all dangers. Whatever security something enjoys is relative to the dangers in question, and what is secure against x might not be against y. However, just because something (e.g., deadbolts on a door) does not confer security against any and all dangers (e.g., ballistic missile attacks), it does not follow that it offers no security at all.
With these notions in mind, consider how a true belief may be secure. Suppose you believe that the Earth is warming because of carbon dioxide released by humans. Suppose this belief – call it ‘B1’ – is true. B1 ceases to be qua true belief if it ceases to be true and/or ceases to be believed. That is, B1 ceases to be qua true belief if (a) things cease to be as B1 represents them to be and/or (b) you lose conviction in its truth. In cases like (a), beliefs lose their truth-making connection to the world. In cases like (b), they lose their conviction-sustaining connection to the believer.
True beliefs thus face dangers of two main kinds: qua beliefs to their conviction-sustaining connection to the believer, and qua true to their truth-making connection to the world. We may call the latter ‘truth-defeaters’ and the former ‘belief-defeaters’. Each may be illustrated as follows. Suppose you believe (correctly) that Burj Khalifa tower in Dubai is now the tallest human-made structure in the world. Call this belief ‘B2’. Truth-defeaters of B2 include anything that would render it false; for instance, the Burj Khalifa's destruction by an earthquake, another taller structure being erected, etc. Belief-defeaters of B2 include anything that would sap your conviction that B2 correctly represents the world. Learning that the Burj Khalifa has been destroyed or has been surpassed by another structure might do the trick. But even if all the human-made structures in the world remained the same, you could still cease to believe B2 as a result of (say) a brain injury or being persuaded by a trusted but mistaken friend that the Tokyo Skytree is the world's tallest human-made structure.
Corresponding to the distinction between belief-defeaters and truth-defeaters is the distinction between doxastic security and alethic security. Doxastic security of a belief is a function of how strongly it is protected against belief-defeaters. Alethic security of a belief is a function of how strongly it is protected against truth-defeaters. Consider each in turn.
Doxastic security must be understood in relation to belief-defeaters. None of our beliefs is absolutely secure from belief-defeaters. As a matter of empirical fact, so long as looms the risk of damage to certain regions of the brain so too looms the danger of losing beliefs. Moreover, new evidence or information or challenges can sap the conviction necessary for belief. Take a student who believes correctly that vanadium is the element with atomic number 23; upon pointed questioning from his chemistry professor, however, his conviction in its truth so dissipates that he no longer believes it. Sceptical hypotheses may have a similar corrosive effect on belief.
But even if we cannot achieve absolute doxastic security for all our beliefs, we can still achieve some degree of relative doxastic security for many of them. Take, in contrast with our previous student, another well versed in the properties of vanadium; upon intense grilling from her chemistry professor, her belief that vanadium is the element with atomic number 23 does not waver, for she is able to demonstrate why vanadium does have this atomic structure. The belief of our second student enjoys a stronger degree of relative doxastic security – relative, namely, to the threat of dissipating upon questioning or challenges and even perhaps to certain forms of scepticism – than that enjoyed by the first student.
This relative doxastic security does not protect against any kind of belief-defeater – it does not protect against loss of belief as a result of (say) brain damage – but it does protect against an important kind of belief-defeater. In general, a belief's relative doxastic security is a function of its resistance to belief-defeaters that threaten to sap the conviction sustaining it. To redeploy a metaphor used by Plato in Book IV of the Republic, the more a belief is firmly fixed in the mind like dye indelibly set into wool, the stronger is its relative doxastic security. Or, to use another Platonic metaphor, the more strongly a belief is tethered in the mind, the stronger is its relative doxastic security.
Consider now alethic security in relation to truth-defeaters. A belief is subject to truth-defeaters over its lifespan only if it is subject to being false; hence, only beliefs that cannot be false over their lifespans are absolutely secure from truth-defeaters. Such beliefs presumably include necessarily true beliefs and true beliefs about the past insofar as the past cannot be altered. All other beliefs (including contingent beliefs about the present and about the future) are presumably not secure from truth-defeaters insofar as they may be false during some or all periods of their lifespan.Footnote 6
Even if such beliefs do not enjoy absolute alethic security, however, they may still enjoy some degree of relative alethic security. This may be seen by contrasting beliefs generated by unreliable doxastic processes with those generated by reliable ones.Footnote 7 Insofar as the latter have a greater probability of being true over their lifespans than the former, they enjoy greater relative alethic security as a function of their more limited exposure to truth-defeaters.
To put the point in a related way, true contingent beliefs about the present or the future generated by reliable processes are more firmly connected (or ‘secured’ or ‘tethered’) to their truth-makers and hence less vulnerable to truth-defeaters than beliefs generated by unreliable processes. Thus, while necessarily true beliefs and true beliefs about the past enjoy an absolute alethic security not achievable for true contingent beliefs about the present or the future, the latter beliefs may still enjoy a relative alethic security over their lifespans in virtue of the reliability of the doxastic process from which they were generated and the firmness of their connection to their truth-makers.Footnote 8
3.4. Tying it all together: Knowledge as Secure True Belief
With these conceptions of belief, truth, and security now clarified, we may return to the thesis that knowledge is secure true belief. On the Security Theory, absolute knowledge is true belief with absolute alethic and doxastic security; to wit: belief which, over its lifespan, is never subject to truth-defeaters or belief-defeaters – belief that could only be true and never cease to be held once acquired. Unfortunately, such absolute knowledge is not attainable by us mortals. For even if we suppose that some of our beliefs (such as necessarily true beliefs) may enjoy absolute alethic security, given the human doxastic condition, none of our beliefs enjoys absolute doxastic security, and our contingently true beliefs do not enjoy absolute alethic security either.
But, pace some sceptics, not all knowledge is absolute knowledge. Even if we are not capable of attaining absolute knowledge, we are of relative knowledge. Think of a spectrum. At one end lie mere true beliefs lacking alethic and doxastic security. Such beliefs fail to count as knowledge. At the other end lies absolute knowledge unattainable by us. In between lies relative knowledge: true beliefs that enjoy various degrees of alethic and doxastic security. The more alethic and doxastic security a true belief enjoys, the higher the grade of knowledge it counts as being.
On the Security Theory, then, instances of relative knowledge come in various grades of alethic and doxastic security. Some instances of knowledge may have a higher grade of alethic security than others. Some may have a higher grade of doxastic security. And so on. Ideally, though, the closer our true beliefs approximate absolute knowledge in alethic and doxastic security, the higher the grade of knowledge. Accordingly, while we may know that p and know that q, our knowledge that p may be of higher grade than our knowledge that q. This is so when our believing correctly that p is more alethically and/or doxastically secure than our correctly believing that q.
4. The Theory Justified
I have explained the Security Theory, but explaining is not justifying. A full-dress defense of the theory would require demonstrating its superiority to its main rivals, and space does not permit that here. I will advance, however, eight grounds in support of it. Together they constitute what I take to be compelling preliminary case for the theory.
4.1. The Theory Solves the Extra-Value Problem
The Extra-Value Problem may be stated as follows: if a belief is true and so correctly represents the world, why should we suppose that knowledge is any more valuable or desirable than it? As Socrates pointed out in Meno (97a–c), for instance, whether you have knowledge or true belief about the way to Larissa, both will help you find your way there. Since knowledge is presumably more valuable or desirable than mere true belief, the challenge is to explain how. The Security Theory straightforwardly solves this problem: true beliefs that count as knowledge – in virtue of the alethic and doxastic security they enjoy – are more valuable or desirable than mere true beliefs, because the former are more protected than the latter against the dangers posed by alethic and doxastic defeaters over the course of their lifespans. Knowledge thus has an epistemic value or desirability lacked by mere true belief. Hence knowledge about the way to Larissa is more valuable or desirable than mere true belief about the way.
4.2. The Theory Helps Dissolve Gettier-Type Problems
In general, Gettier-Type Problems for knowledge can be generated by following this formula: find a true belief with some justification (or also with additional conditions), present a situation that elicits intuitions according to which it fails to count as knowledge, and conclude that true justified belief (even with additional conditions) must therefore not suffice for knowledge. Despite all the ink and paper and time and effort devoted to the subject, what has yet to be noticed in the epistemological literature on Gettier-Type problems for knowledge is that they are a species of a more general type of problem for any form of security.
Take once again our example of home security. Suppose someone has installed deadbolts on the doors and an alarm system for his house. The house thus enjoys a fair amount of security. But short of it having absolute security against any and all dangers which cannot possibly be achieved, a ‘Gettier-type’ philosopher can raise putative counter-examples pointing out limitations to whatever level of security is achieved: Are there iron bars on the windows? Is the house safe from aerial attack? And so on. Short of absolute security, there will typically be ways to call into question whatever security it has.
Do such counter-examples establish that there is no home security at all? No. Even if we concede that they establish that the security in question is not absolute, we need not also concede that they establish that there is no home security at all. A home does not have to be absolutely secure to be secure at all, for it can still be secure relative to some dangers. Of course, the more dangers against which a home security system protects, the higher the grade of security it provides. But this does not mean that failure to protect against any and all dangers means that it fails to provide any security at all.
The Security Theory allows us to respond to Gettier-type challenges without piling on epicycle upon epicycle of conditions to parry them, and without narrowly restricting knowledge to nothing but true belief that cannot be false. What Gettier-type problems show is that, with the exception of beliefs that cannot be false, all our other true beliefs do not enjoy absolute alethic security.
Does it follow from this that no such beliefs count as knowledge? No. Just because a true belief does not count as absolute knowledge, it may still, on the Security Theory at least, count as relative knowledge as a function of the alethic and doxastic security it does enjoy. The theory thus does not ‘solve’ Gettier-type problems through the use of various Gettier-parrying conditions or by narrowly restricting the scope of knowledge. The theory rather ‘dissolves’ such problems by explaining how they arise generally and how they fail to show there is no knowledge.
4.3. The Theory Subsumes Important Insights from Internalism and Externalism
A particularly strong form of Knowledge Internalism holds that in order for S to know that p, S must be able to justify her belief that p by adducing reasons or arguments for it. More moderate versions of Knowledge Internalism do not require that S be able to justify her belief that p, but require S to have grounds for this belief that are cognitively accessible to her or at least that supervene on her mental conditions, processes, or events. By contrast, Knowledge Externalism rejects such strictures on knowledge, and (at least on one of its main forms) takes knowledge to be true belief generated by reliably truth-conducive doxastic processes.
The Security Theory does not force us to choose between Internalism and Externalism, but allows us to subsume significant insights from each in light of our notions of doxastic and alethic security. We can see this even if we put aside the complications arising from (a) whether the desiderata posited by versions of Internalism (such as the ability to justify beliefs or have grounds that are cognitively accessible) confer alethic security on beliefs, and (b) whether the desiderata posited by versions of Externalism (such as being generated by a reliable doxastic process or being a reliable indicator of truth) confer doxastic security on beliefs. Even if we assume a negative answer to (a) and (b), our theory allows us to acknowledge that Internalism provides a valuable insight into doxastic security and Externalism into alethic security, two insights worth affirming and preserving.
Take some desiderata we may associate with Internalism. If someone is able to justify her beliefs with reasons or arguments, her belief is more liable than otherwise to withstand the potentially conviction-sapping effect of challenges or questioning. If someone has cognitive access to the grounds for his beliefs, he is more liable to sustain his conviction than someone who does not. Thus does Internalism provide a valuable insight into doxastic security. Now take some desiderata we may associate with Externalism. If someone's beliefs are generated by reliably truth-conducive processes, they are more liable to be true over their lifespans than otherwise. Thus does Externalism provide a valuable insight into alethic security.
On the Security Theory, both alethic and doxastic security are valuable, and the more a true belief that p enjoys such forms of security, the higher the grade of knowledge it counts as being. The theory subsumes valuable insights from Internalism and Externalism without forcing us to choose one over the other.
4.4. The Theory Takes Some of the Sting Out of Scepticism
An important way of arguing for a global scepticism concerning knowledge is to set an extremely high standard for what counts as knowledge and then to argue that nothing satisfies this standard. One way of doing this is to presuppose that for a belief to count as knowledge it must not only be true but must also have absolute alethic and/or doxastic security. To accept the Security Theory is in effect to concede that such scepticism cannot be defeated on its own terms on the alethic security dimension for all contingent beliefs, and for all beliefs on the doxastic security dimension.
Accepting the theory, however, involves rejecting the terms a global sceptic would set for all knowledge. Just because not all true beliefs count as absolute knowledge (which is impossible for us mortals anyway), it does not follow that none of them counts as relative knowledge in virtue of the alethic and doxastic security they do enjoy. The Security Theory can thus take at least some of the sting out of sceptical challenges by showing how a sceptical case against absolute knowledge does not rule out relative knowledge.
Moreover, even if the sceptic does not presuppose that knowledge must be absolute knowledge, and proposes some other standard that must be met for true belief to count as knowledge (e.g., the putative knower must be able to rule out sceptical hypotheses such as whether she is a Brain in a Vat or being misled by an Evil Demon, and the like), the Security Theory can take some of the sting of this weaker scepticism as well.
Here is how: if, as on this theory, relative knowledge is scalar, with some forms of knowledge being of higher grade than others as a function of their alethic and doxastic security, then even if we concede to the sceptic that some true beliefs do not qualify as a certain kind of high grade knowledge (e.g., where the putative knower in question is able to rule out various sceptical hypotheses), it does not follow that we must also concede the sceptical conclusion that these true beliefs do not count as knowledge at all. Instead, we can hold that they count as relative knowledge of a lower grade.
The Security Theory thus subsumes an insight behind the Relevant Alternatives (RA) approach to understanding knowledge, an approach defended by Fred DretskeFootnote 9 and Alvin GoldmanFootnote 10 and others. The key idea of this approach is that knowing that p requires the ruling out of the relevant not-p alternatives. While defenders of this approach differ on what counts as a relevant alternative and what counts as ruling out, in general it provides a way of giving scepticism its due while limiting its threat, since sceptical possibilities (e.g., whether one is dreaming, whether the barn-appearing object in front of one is really only a cleverly-placed barn façade, whether one's department chair is really a zombie, and the like) may not be relevant to one's knowing something in ordinary cases, but may be relevant in others (e.g., in a philosophy seminar). While the Security Theory does not entail that all forms of knowledge require having the capacity to rule out relevant alternatives as does the RA approach, it nonetheless subsumes this approach, for understanding knowledge in terms of the ruling out of relevant alternatives can itself be understood in terms of alethic and doxastic security. That is, the more someone who has a true belief that p is able to rule out sceptical possibilities, the stronger will presumably be the alethic and/or doxastic security of her true belief, and hence the higher the grade of her knowledge that p.
4.5. The Theory Subsumes an Important Insight of Epistemic Contextualism
The Security Theory is a theory of knowledge, not of knowledge attribution as Epistemic Contextualism is standardly understood. Nonetheless, our theory subsumes an important insight of the latter.
Hector-Neri Castañeda,Footnote 11 Gail Stine,Footnote 12 and many others have argued that what we are prepared to count as knowing may depend on context. For instance, what we are prepared to count as someone's knowing that p, where (say) p = Caesar Crossed the Rubicon in 49 BC, may depend on whether we are considering her belief expressed on (a) a television quiz show, (b) a grade school student's essay, or (c) a work defending the accuracy of this dating in the face of a case made by an illustrious scholar that Caesar made his crossing in 48 BC. As Stine pointed out, moreover, our criteria for attributing knowledge might be tighter in some contexts (e.g., a courtroom) than in others (e.g., a grade school classroom).
The Security Theory subsumes (and qualifies) this contextualist insight as follows: if knowledge is scalar along the dimensions of alethic and doxastic security, then the context of a putative knower may very well affect the grade of knowledge we expect of her in a given situation. Take, for instance, one's knowledge that q, where q = blood flows through the left atrium into the left ventricle of the heart. The standard of alethic and/or doxastic security we expect to be met when the stakes are very high (e.g., what we would expect from a heart surgeon when someone's life is on the line on an operating table) will generally be much higher than what we expect when the stakes are much lower (e.g., what we would expect from a freshman anatomy student on a pop quiz). While the surgeon and the freshman may both know that q, the former's knowledge will be of higher grade than the latter's insofar as it enjoys greater doxastic and/or alethic security. Since, on the Security Theory, two persons A and B can know that q, but A's knowledge can be of higher grade than B's, there may indeed be contexts where we do not regard B's knowledge as being of the grade needed when the standards (or stakes) are high. On the Security Theory in contrast with Contextualism, however, this does not mean that we should not attribute knowledge to B in the context in question; it means rather that B's knowledge is not of the grade expected relative to the interests or concerns in play in the context in question.
In sum, the Security Theory of Knowledge subsumes an important contextualist insight about the role of context and the plurality of standards in the attribution of knowledge, without the (more controversial) contextualist implication that we should not attribute knowledge to someone at all just because we do not regard her as having the grade of knowledge needed when the standards (or stakes) are high.
4.6. The Theory Can Explain Why Judgments About Knowledge May Differ
The epistemological literature bears ample witness to how judgments (or intuitions) differ concerning what counts as knowledge. An attractive feature of the Security Theory is that it can provide an explanation for why such differences arise. Here is how: If knowledge is secure true belief, then what people may count as knowledge may differ depending on the kind of doxastic and alethic security they value or regard as salient.
Consider an analogy with national security understood as relative to actual and potential threats to a country. Different individuals (or groups) may differ in their assessment of the risks posed by various threats, and differ concerning the strategies or methods used to secure against them. They may also differ in the tradeoffs they think are worth making.
Similarly, if knowledge is secure true belief, then many individual (or even group) differences concerning what counts as knowledge may be explainable as being at root differences in assessment of the risks posed by alethic and doxastic defeaters, of the strategies or methods to secure against them, and of the tradeoffs worth making.
That people differ over what counts as knowledge does not show that there is no knowledge any more than people differing over national security shows there is no national security. What is explainable (and predictable) on the Security Theory, however, is that inter-individual (and inter-group) disagreement over what counts as cases of knowledge is likely to occur.
4.7. The Theory Bears the Promise of Being Empirically Fruitful
One of the stock criticisms of standard philosophical epistemology from those of a naturalistic bent is that typical accounts of knowledge bear little fruit in terms of generating an empirical research program. The Security Theory by contrast does bear the promise of such fruit. For if we conceptualize knowledge as secure true belief, and understand it in terms of alethic and doxastic security, an empirical research program that naturally ensues involves exploring ways in which such forms of security may be attained.
The following issues for instance invite empirical investigation: What doxastic processes, what reasoning strategies, what methods of inquiry, are most effective in attaining doxastic and alethic security? How do such processes, strategies, and methods compare to one another in terms of such security? Are there tradeoffs to be made between doxastic and alethic security? If so, what are they?
The theory thus bears the promise of actually being refreshingly relevant to cognitive science, cognitive psychology, and related fields, and they in turn become refreshingly relevant to philosophical epistemology.
4.8. The Theory Has Useful Pedagogical Applications
One of the key goals of education is for teachers to impart knowledge to their students. A noteworthy feature of many contemporary theories of knowledge, however, is how typically barren they appear in terms of pedagogical applications. As a striking example, one of a number of others that could be culled from the literature, consider how Moser offers (he says for simplicity) the following analysis of propositional knowledge (PK):
PK. A person, S, has propositional knowledge that P if and only if: P is true; S has justifying evidence E for P that is truth-resistant in the way specified by TR; and S believes or assents to P on the basis of E. Footnote 13
What then is TR? Moser defines it as follows:
TR. S's justifying evidence E for P is truth-resistant if and only if for every true proposition T that, when conjoined with E, contravenes S's justification for P on E, there is a true proposition, T’, that, when conjoined with E & T, restores the justification of P for S in a way that S is actually justified in believing that P. Footnote 14
In fairness to Moser, it should be noted that I have not gone into all the details of how his account of knowledge works and in particular how it is designed to parry Gettier-type challenges. It seems fairly clear, however, that, whatever its virtues, it's not particularly helpful in elucidating what we as educators are trying to do in imparting knowledge to our students.
The situation differs importantly with the Security Theory. If knowledge is secure true belief, our aim as educators should not just be to help our students believe what is true, but rather to help them develop techniques, skills, processes, and strategies that will foster the alethic and doxastic security of their true beliefs, and to help them understand what tradeoffs (if any) have to be made between such forms of security. Conceiving of this goal in terms of such security provides a clearer objective than what is provided by many contemporary theories of knowledge, including Moser's and others. Thus does the theory have the potential for useful pedagogical applications.
5. Objections and Replies
As with any philosophical theory, a number of objections can be lodged against the Security Theory, and space does not permit answering them all here. I will, however, address five telling objections.
5.1. The Brainwashing Objection
Consider the notion of doxastic security being required for knowledge. Suppose that through brain-washing, we were able to induce a very strong degree of doxastic security equivalent to (say) a kind of fanaticism. Now compare two individuals A and B for whom there is the same degree of alethic security regarding some true proposition p, but A believes that p because A has been brainwashed to believe that p, and B believes that p on the basis of B's sensory experience. On the Security Theory, A's belief may count as being of a higher grade of knowledge than B's belief insofar as the former's doxastic security is stronger than the latter's. But isn't this patently absurd?
Reply. It follows on the Security Theory that if two true beliefs have the same degree of alethic security but differ in doxastic security, then the belief that enjoys more doxastic security counts as being knowledge of a higher grade. So, if we assume that brainwashing leads to A's belief that p enjoying more doxastic security than B's belief that p which did not result from brainwashing, then in this case it follows on the Security Theory that A's knowledge is of a higher grade than B's. It is understandable why some may find this counter-intuitive, for it does seem odd at first blush that, in a case of equivalent alethic security, someone who was brainwashed to believe that p would have knowledge of a higher grade than someone who believed that p as a result of sensory experience.
I think that the source of this intuition is that we are inclined, rightly, to hold that brainwashing in general compares poorly to (say) sensory experience as a process leading to knowledge. Why is this? Well, in general, even if brainwashing were successful in generating strong conviction and hence doxastic security for some beliefs, it is liable to come at tremendous costs in alethic security for one's belief system as a whole, including the rigidity of the system and its disconnect from, or tenuous link to, reality.
Thus, even if we suppose that brainwashing might induce more doxastic security than sensory experience over some period of time for some true beliefs, it is likely that it will result in much less knowledge over time relative to one's belief system as a whole. Accordingly, in cases of equivalent alethic security, while brainwashing may, for some true beliefs, lead to higher doxastic security than is the case with (say) sensory experience, and hence lead to a higher grade of knowledge in these cases, it does not follow that accepting the Security Theory requires favoring brainwashing as a process for generating knowledge, for brainwashing is liable to result in a significant net overall loss of knowledge.
5.2. The Truetemp Objection
Keith LehrerFootnote 15 gave the example of Truetemp who, via the implantation of a device in his brain, is given an unusual cognitive faculty of being able to tell the ambient temperature in a highly reliable manner. Unaware of its existence, Truetemp does not know what to make of the beliefs generated by this new faculty. Lehrer concluded that Truetemp would not have justified beliefs or knowledge generated this way, and hence Externalism is mistaken.
Consider now a variation on this example: suppose the device not only enables Truetemp to tell the temperature in a very reliable manner and hence his beliefs thereby generated enjoy a high degree of alethic security, but the device also induces in him a strong degree of doxastic security. On the Security Theory, such beliefs would count as being of a very high grade of knowledge. But isn't this counter-intuitive?
Reply. The Security Theory counts such beliefs as being of a high grade of knowledge, and this may strike some as counter-intuitive. Notice, however, how the intuition behind this variation of the Truetemp example can be explained in terms of the internalist emphasis on being able to justify one's beliefs, or argue for them, or give reasons for them. One of the reasons this epistemic desideratum proves attractive is that, in normal cases, having such a capacity is liable to result in greater doxastic security than otherwise – recall my earlier examples of the two chemistry students. The capacity to justify one's beliefs is instrumentally valuable or conducive to doxastic security.
In the objector's strange example, however, a high degree of doxastic security is already assumed, so the instrumental role played in normal cases by the capacity to justify is rendered otiose. Consequently, I think it is a mistake to apply this intuition generated in normal cases to an abnormal case like the Truetemp variation. The latter asks us to imagine very atypical conditions, and so why shouldn't the kind of knowledge instantiated in such conditions be correspondingly atypical? Moreover, it is worth remembering that we may still suppose that the capacity to justify one's beliefs (or argue for them, or give reasons for them) is a valuable desideratum in terms of one's overall rationality even if we suppose that it is not essential to all forms of knowledge.
5.3. The Alethic vs. Doxastic Security Objection
Consider all the possible cases that would be classified as knowledge on the Security Theory. Suppose one true belief has slightly more alethic security than another but the other has slightly more doxastic security. Which putative form of knowledge is better? Suppose one true belief enjoys a high degree of doxastic security but little alethic security, and another enjoys a high degree of alethic security but a low degree of doxastic security. Which again is better? The Security Theory provides us with little practical guidance concerning how to sort out these matters.
Reply. On this theory, there may be numerous cases of knowledge differing along the dimensions of alethic and doxastic security. While the theory does answer why knowledge is more valuable than mere true belief, it does by not itself sort out which forms of knowledge are better than others, for what is better will be relative to what our interests and values are concerning knowledge, and what doxastic and alethic dangers our true beliefs are liable to face. No theory of the nature of knowledge by itself can answer questions concerning which forms of knowledge are better than others.
Consider an analogy with home security: are deadbolts on the door better than iron bars on the windows? Well, this depends on a number of factors including what dangers we are liable to face, what tradeoffs we are prepared to make in terms of costs and benefits, and so on. Moreover, while iron bars on the windows may, in addition to deadbolts on the doors, be a good idea on the first floor of a house of a crime infested neighborhood, in a much safer one we may be unwilling to install them because of their monetary, aesthetic, or other costs.
Similarly, depending on the context in question, we may prefer some forms of knowledge over others depending on the tradeoffs that must be made concerning their attendant doxastic and alethic security. For instance, knowledge with a lower degree of doxastic and/or alethic security may be better in some contexts than knowledge with a higher degree of doxastic and/or alethic security if the former is available more readily and conveniently and is good enough for the purposes at hand. In other contexts, (say) where someone's life is at stake during surgery, knowledge with a higher degree of alethic and/or doxastic security may be preferable to lower grade knowledge. In short, just because the Security Theory does not sort out by itself which forms of knowledge are more valuable than others, this does not mean it ought to be rejected as a theory of the nature of knowledge.
5.4. The Relativism Objection
Given the previous response, doesn't it follow that the Security Theory entails an implausible cognitive relativism insofar as it takes knowledge to be relative to our interests and values?
Reply. The Security Theory does not take knowledge to be relative to our interests and values. On the theory, knowledge that p entails p's truth, and nothing in the theory entails relativism about truth. Supposing that some forms of knowledge are better than others is relative to our interests and values does not entail that knowledge itself is relative to our interests and values.
Consider again an analogy: while oak and mahogany are both kinds of wood, in some contexts oak is a better wood (relative to our interests) than mahogany and vice-versa. It does not follow from this, however, that oak and mahogany have no objective non-relative properties. In fact, it is because of their objective non-relative properties that some forms of wood may serve our interests better than others. Mutatis mutandis for knowledge.
5.5. The Externalist Bias Objection
You claim that the Security Theory captures important insights of Internalism. But is it not inherently externalistic? For does not the theory, as does Externalism, deny that such internalist requirements as the capacity to justify one's beliefs or have access to the grounds of one's beliefs are required for knowledge?
Reply. An important insight we may garner from Internalism is that knowledge requires doxastic security. Conversely, an important insight we may garner from Externalism is that knowledge requires alethic security. As noted earlier, the Security Theory affirms both insights. It is neither straightforwardly externalist nor internalist, however.
Here is why. The Security Theory neither denies (as does Externalism) nor affirms (as does Internalism) that the internalist requirements mentioned above are required for knowledge tout court, for, on the theory, knowledge may come in variegated forms: forms that meet such internalist requirements, and forms that do not. Knowledge as a genus is true belief with doxastic and alethic security, but its species may instantiate various degrees along both dimensions of such security. The Security Theory, then, while affirming important insights garnered from both Internalism and Externalism, cannot be neatly classified as a form of either.
6. Conclusion
The central aim of this paper has been to revive and refine an old idea inspired by Plato, and to show how it can be developed into a plausible contemporary theory of knowledge. A compelling case can be made for the Security Theory, and it bears the promise of providing a fruitful framework for epistemology, one that subsumes valuable insights from several other epistemological theories. To be sure, much more needs to be done in articulating and defending the theory. I have begun but not ended that task here.
