Here's a paper defending Alvin Plantinga's orginal, epistemically internalist version of reformed epistemology from my first year of grad school.
Plantinga, Sennett, and Reformed Epistemology
In “Reason and Belief in God” (hereafter ‘RBG’), as well as in many other contexts, Alvin Plantinga argues that belief in God is “properly basic” That is, belief in God is rational, or justified, or warranted without being based on propositional evidence. This thesis constitutes the core of what is known as ‘Reformed Epistemology’. Call this thesis ‘RET’. James Sennett, among others, denies RET. For example, in his Modality, Probability, and Rationality, he argues that Plantinga fails to show that it is even epistemically possible that RET is true. In this paper, I will give a sketch of Plantinga’s case for RET, as well as Sennett’s critique of it. Finally, I will argue that although Sennett’s arguments have some force, he ultimately fails to make his case against RET. For he fails to adequately appreciate some features of Plantinga’s epistemology, features that he borrows from the epistemological writings of Roderick Chisholm. Once these features are fully appreciated, it is clear that Sennett’s objections are wide of the mark.
1. Plantinga’s Case for RET
1.1 The inadequacy of Classical Foundationalism
In RBG, Plantinga sets out to demolish the grounds for thinking that belief in God is not properly basic. There, he points out that the grounds for this claim stem from an epistemology he calls ‘Classical Foundationalism’ (hereafter ‘CF’). CF is the view that all known or justifiably believed propositions are either (i) self-evident, incorrigible, or evident to the senses, or (ii) based on (i.e., entailed or sufficiently inductively supported by) propositions of the type mentioned in (i). Now beliefs that satisfy (i) have typically been called ‘foundational beliefs’, and these are among the ones that Plantinga labels ‘properly basic’. But if CF is true, then it would appear that belief in God is not properly basic. For theistic belief doesn’t seem to be self-evident, nor does it appear to be incorrigible or evident to the senses. But if not, then belief in God fails to satisfy the conditions spelled out in (i). So we have good reason to think that if CF is true, then belief in God is not properly basic.
Plantinga has two main objections to CF. First of all, he points out that if it’s true, then it entails the absurdity that a large host of our beliefs that we take to be rationally held, such as the beliefs that there is an external world, that our memories are often reliable, that there is a past, that objects continue to exist when no one is looking at them, etc. (call such beliefs ‘the Commonsense Beliefs’) are in fact irrational, unjustified, or unwarranted. Plantinga doesn’t spell out his reasoning for this conclusion in much detail, but it seems pretty clear what he has in mind. One can suppose the argument for this point has two stages. In the first stage, one would argue that the Commonsense Beliefs aren’t properly basic. For it’s pretty clear that, for example, our belief that the world has existed for more than ten minutes isn’t self-evident: it isn’t certain for one, just on the basis of understanding the meaning of the statement. Furthermore, it’s definitely not an incorrigible belief: it’s not impossible to doubt it in the way that it’s impossible to doubt beliefs about certain kinds of one’s occurrent mental states. Thus, for example, it’s (at least epistemically) possible that the world came into being ten seconds ago, with all of our “memories” created with it. Finally, the belief isn’t evident to the senses, for only things in the present can be evident to the senses. Thus, belief in a past isn’t properly basic. And presumably this line of reasoning can be applied to the other kinds of Commonsense Beliefs mentioned above.
One can imagine that the burden of the second step of the argument is to show that the Commonsense Beliefs aren’t sufficiently based on properly basic beliefs. For recall that according to CF, only perceptual beliefs, beliefs about one’s immediate sense impressions, certain other occurrent mental states (such as what I am feeling or thinking about (or at least what I think I am feeling and thinking about)), and logical and mathematical truths are the only kinds of beliefs that could be properly basic. But then it seems impossible to deduce the commonsense beliefs from the properly basic ones. For example, one can’t deduce the belief that objects continue to exist when one isn’t looking at them from statements about what is evident to the senses. For it seems logically possible that I could have all of the same perceptions of (say) a tree in my back yard, yet it frequently pops in and out of existence while I’m at school. Furthermore, it seems clear that the Commonsense Beliefs can’t be sufficiently inductively supported by properly basic beliefs. Take my belief that I have existed for over five minutes. What properly basic beliefs of mine make it probable that this belief is true? And how are we to interpret probability here? Surely I can’t use the relative frequency interpretation of probability in this case without begging the question. Presumably, similar problems apply to the other Commonsense Beliefs. Therefore, the Commonsense Beliefs aren’t properly basic, nor are they sufficiently based on such beliefs. But if so, then if CF is true, then the commonsense beliefs are irrational and/or unjustified. But surely this is implausible: such beliefs are obviously rational and justified. And since CF is the view that leads to this absurdity, we should reject it.
Second, Plantinga argues that CF is self-refuting. To see this, let’s state the definition of CF more explicitly:
(CF) A belief B is justified for a person S iff B is either (i) self-evident, incorrigible, or evident to the senses, or (ii) based on (i.e., entailed or sufficiently inductively supported by) propositions of the type mentioned in (i).
But then it is clear that CF itself isn’t self-evident, incorrigible, or evident to the senses, nor is it based on propositions that are. Thus, CF is necessarily false, since this is the lot of all self-refuting statements. For these reasons, then, we should reject CF. But CF is the basis for thinking that RET is false. Therefore, the central grounds for rejecting RET are undercut.
1.2 Toward A More Plausible Foundationalism
In the place of CF, Plantinga suggests that we adopt a modified version of foundationalism (hereafter ‘PF’). In RBG, he suggests that we follow Roderick Chisholm in generating the criteria for foundational – or properly basic – beliefs by a sort of inductive process. As Plantinga puts it,
"We must assemble examples of beliefs and conditions such that the former are obviously properly basic in the latter, and examples of beliefs and conditions such that the former are obviously not properly basic in the latter. We must then frame hypotheses as to the necessary and sufficient conditions of proper basicality and test these hypotheses by reference to those examples."
How does it work? Again, Plantinga doesn’t provide much detail, though I think we can spell out what he has in mind. The process can be construed as having several steps. In the first step, one examines one’s beliefs that one takes to be rational. For example, suppose I reflect on the belief that I had a fried egg sandwich for breakfast this morning. Suppose further that, upon reflection, I find that I truly believe this. Finally, suppose that it seems to me that the belief is justified, or rationally held. If so, then one can go on to the second step. In this step, I consider whether I believe this proposition on the basis of other propositions, or whether I believe it in the basic way. As it turns out, I see that the belief is basic for me: I don’t believe it one the basis of other beliefs of mine. I can then go on to the third step. Here, I examine a number of other memory beliefs that seem to be rationally held by me, and I see that they are also held by me in the basic way. I then go on to the fourth step. At this point, I try to surface the kind of circumstances they have in common. For example, I am not on drugs, I’m sufficiently rested and relaxed, etc. Then I have the materials for some preliminary premises in an inductive argument: Belief b1 is justified for S in circumstance c1, belief b2 is justified for S in circumstance c2, etc. From here I go on to the step: I construct the argument that since memory beliefs b1, b2, etc. of k are properly basic in circumstances c1, c2, etc., respectively, then probably all beliefs bi∈B (i.e., all memory beliefs) in all relevant ci∈C (i.e., all relevant circumstances) are
properly basic for me. Thus, we now have our criterion for the proper basicality of memory beliefs, inductively supported by the process of the previous steps: There is a set of circumstances of kind C, and a set of (possible) beliefs of kind B, such that for any person, S, if S forms some bi∈B in some ci∈C. then bi is properly basic for S. One can go on to apply this process to every kind of one’s beliefs, until one has criteria for all basic kinds of beliefs. Using such criteria, one can then sort out all of one’s beliefs into two categories: properly basic and properly based (where properly based beliefs are sufficiently inductively or deductively supported by one’s properly basic beliefs).
Two things should be said about the procedure mentioned above. First of all, it is crucial to notice that PF doesn’t commit one to the absurd view that just any belief can be properly basic. For the beliefs must have the right kind of grounds. To see this, recall that the inductive procedure of forming criteria for proper basicality makes crucial reference to beliefs formed in certain circumstances. These circumstances crucially involve certain kinds of experiences. So, for example, consider my belief that I see a tree in my backyard. This belief isn’t properly basic under just any circumstances. Rather, it is properly basic only if I am in certain kinds of circumstances, e.g., standing in my backyard and attentively looking in a certain direction when there is a certain amount of sunlight, etc. Furthermore, I have to have a perceptual experience of the tree. So if a belief isn’t formed in the right kind of circumstance, and if it doesn’t involve the right kind of experience, then it can’t be properly basic. Thus, it’s not the case that just anything goes with respect to proper basicality. Second, just because a belief, or even a kind of belief, is considered to be properly basic after such a procedure, it does not follow that it is henceforth indefeasible. Other evidence or relevant considerations could turn up which may undermine the proper basicality of a belief (or kind of belief). Plantinga -following John Pollock - calls those things that undermine beliefs ‘defeaters’. There are two basic kinds of defeaters: undercutting and rebutting. An undercutting defeater is a defeater that removes the grounds that one may have for believing a proposition. Thus, if one has an undercutting defeater for some belief B, then one has sufficient justification to withhold judgement with respect to B. A rebutting defeater, however, has more force. If one has a rebutting defeater for some belief B, then one has sufficient justification to believe not-B. With this in mind, we may say that if a belief B is taken to be properly basic by some person S, then B may lose its basic status for S if S comes to have an undercutting or rebutting defeater for B. So, clearly, basic beliefs are defeasible. These two points remove (at least some of) the air of arbitrariness and dogmatism that may seem to attach to the approach.
1.3 The theism-friendly implications of the new foundationalism
Plantinga thinks the same inductive process mentioned above can be used to show that belief in God is properly basic for theists. For example, the theist may examine her beliefs that, say, God made the world, that God disapproves of what I’ve done, etc. If she then reflects on whether such beliefs are rational for her, and concludes that they are, then she may look to see what they have in common. For example, perhaps her belief that God made the world arises within her when she looks at the sky on clear, starry nights, and her belief that God disapproves of what she’s done spontaneously arises within her when she performs certain actions, say. Suppose she thus sees that she doesn’t base these beliefs on other beliefs, such as (say) arguments from natural theology. From this, she concludes that she believes such things in the basic way, and properly so. If all of this is so, then she may formulate criteria of proper basicality accordingly: There is a set of possible circumstances of kind K (e.g., K could be looking reflectively at the starry heavens, or performing a certain putatively immoral action, etc.) and set of (possible) theistic beliefs of kind T such that for any person, S, and theistic belief ti∈T, if S forms ti in some ki∈K, then ti is properly basic for S. In this way, then (thinks Plantinga), belief in God can be properly basic for the theist, in which case we have good grounds for accepting RET.
2. Sennett’s Critique
2.1 Community-relative rationality
Sennett thinks that Plantinga’s case for RET fails. For one thing, he argues that PF entails that rationality and justification are community-relative. This is because Plantinga’s method of generating criteria of proper basicality is too “loose”: such a method permits the possibility (and probability) that different people, and different communities of people, will come up with different and incompatible beliefs that they take to be properly basic. One person will go through Plantinga’s inductive procedure and conclude that belief in God is properly held in the basic way, and another will not. But then, according to PF, the incompatible basic beliefs of the two people are equally justified. Furthermore, it is difficult to see how there can be a resolution as to which person is right about the proper basicality of belief in God. For clearly it won’t help if each of the two people simply report what beliefs are properly basic for them. But this goes against our intuitions about rationality and justification. We believe that a necessary condition for any adequate epistemology is that it ensures the ability to resolve disparity among the beliefs of different people. But if rationality is what Plantinga takes it to be, then it seems that there is no way to resolve such disparity, and this is unacceptable.
2.2 Not a genuine epistemology
Furthermore, PF is not a genuine, normative epistemology, but rather one that merely generates descriptions of what certain people believe in certain kinds of circumstances. Sennett thus takes Plantinga’s supposed epistemological methodology to be nothing more than a way to discover what one believes naturally without propositional grounds. But left out of such an account is the crucial normative element that determines the epistemic propriety or impropriety of such beliefs. One also needs to give an account for why or how such beliefs are justified: what is the justification-making property for such beliefs?
2.3 An adequate foundationalism exists that precludes RET
Finally, a more plausible epistemology can be constructed that avoids the problems Plantinga raises against CF, yet still prohibits belief in God from being properly basic. Recall that Plantinga argues that CF (i) entails that many of our obviously justified/rational beliefs are in fact unjustified/irrational, and (ii) is self-refuting. But Sennett suggests a modified version of foundationalism (henceforth ‘SF’) that avoids objections (i) and (ii), yet prohibits theistic belief from being properly basic. Sennett argues that the obviously rational Commonsense Beliefs precluded by CF all share a common feature: Universal Sanction. He stipulates that a belief B enjoys universal sanction just in case B is a member of a belief-kind K, such that the members of K (i) are affirmed by virtually all people under normal circumstances, and (ii) can’t conceivably be countenanced as unjustified by healthy human beings. With the criterion of universal sanction, SF confers proper basicality on the obviously rational Commonsense Beliefs that CF prohibits. Indeed, Sennett’s criterion of Universal Sanction is little more than an expression of the conditions that make a belief commonsensical. Furthermore, all of the beliefs that CF renders rational are also considered rational under SF. Finally, SF is well motivated, in that its criterion of universal sanction is intuitively plausible. But if so, then SF can be used to reject PF. Sennett offers the following argument for this for this point:
CF fails because it does not allow for the proper basicality of the Commonsense Beliefs.
5. All Commonsense Beliefs bear Universal Sanction.
6. Theistic belief does not bear Universal Sanction.
7. Therefore, there is good reason for SF, which allows for the proper basicality of the Commonsense Beliefs yet reject the proper basicality of theistic beliefs.
8. Therefore, there is reason to reject CF yet also reject RET.
Sennett thinks that the inference from (6) and (7) to (8) is justified if we think that if a belief has Universal Sanction, then this is a good reason to think that it is properly basic; and if we that that if a belief lacks Universal Sanction, then this is a good reason to think that it is not properly basic. Therefore, Sennett argues, PF is not to be preferred to SF, and so Plantinga’s case for RET fails.
3. Evaluation: A Defense of RET
In my view, Sennett’s arguments appear to have considerable force. However, it’s not at all clear to me that Sennett has refuted RET. For he has failed to appreciate some key features of Plantinga’s epistemology: features which, when adequately understood, show that Sennett’s objections are misguided. The features I have in mind are derived from the epistemology of Roderick Chisholm. To gain an adequate understanding of these features, it will be helpful to look briefly at how they arise in the context of Chisholm’s discussion of the Problem of the Criterion.
3.1 Clarifying PF: a particularist theory
Chisholm argues that the Problem of the Criterion arises from two questions: (1) What do I know/justifiedly believe? And (2) how do I decide, in any given case, whether I know/justifiedly believe something? Question (1) pertains to the extent of our knowledge/justified belief, and (2) pertains to criteria for knowledge/justified belief. It appears that one can answer either question just in case one can answer the other. Consequently, failure to answer either question seems to imply that one can’t answer either of them.
Chisholm points out that, barring agnosticism and skepticism, there are two options in solving the problem generated by (1) and (2). One can either (a) try to formulate criteria for knowledge/justified belief without appeal to particular instances of knowledge, or (b) try to identify particular instances of knowledge without appealing to criteria. Chisholm calls option (a) ‘methodism’, and he calls option (b) ‘particularism’. However, he argues that if one is to know anything, then (b) is the only genuine option. For to opt for (a) is to demand, prior to the knowledge of any true proposition, that a criterion for how one knows a given proposition, P, must be in hand. And of course if this were so, then one would need a criterion for knowing whether the previous criterion is a good one. This leads to a vicious infinite regress of required criteria, and so leads to skepticism. If this is right, then it appears that if there is knowledge, then particularism is true: we can know/justifiedly believe at least some things without knowing how we know/justifiedly believe them. This last statement constitutes the essence of particularism. Call this thesis ‘the Particularist Thesis’ (‘PT’ for short):
(PT): For any (properly functioning) human being S, there is at least one belief B, such that (i) S has B, (ii) S sees that B is justified for S, and (iii) it is not the case that S sees the justification for B.
No doubt many have worries about particularism. Perhaps the main worry is that PT seems to be false: it doesn’t seem possible to see that a belief is justified without seeing its justification (where ‘justification’ denotes the propositional evidence for a given belief). A detailed defense of PT would take up a whole paper (or, as is more likely, a book), so I will briefly gesture to two considerations. First, this conclusion seems to be an inevitable consequence of the reasoning above, if one is convinced that we have at least some knowledge. Second, consider the following analogy: Consider Ralph. Ralph is a man with lots of native intelligence. However, he couldn’t afford to go to college, nor has he had time to develop his skills in articulating and defending his opinions to others. Now suppose that Ralph has been called to jury duty. He attends, and carefully listens to the complex and subtle arguments of the lawyers on both sides. At the end of the trial, and after much thought, Ralph makes his decision: he thinks the defendant is innocent. However, after the trial, when a friend asks him why he thinks the defendant is innocent, he is unable to articulate his reasons to his friend, at least not in any satisfying way. Frustrated by this, Ralph goes home and writes out his reasons for his verdict. As it turns out, however, he can’t spell out his reasons in terms of good inductive or deductive arguments. It’s not that he now sees that his reasoning behind his verdict is faulty. Rather, it is just that he can’t put it into words, nor can he think out the reasoning in his head in terms of cogent, persuasive arguments. In this case, Ralph can see that his verdict is justified, but he can’t see its justification. I therefore submit that, with respect to at least some of our beliefs, we are like Ralph: we can see that they are justified, but can’t see their justification. Thus, for these two reasons, I think that PT is true.
With the above brief sketch of Chisholmian particularism in hand, we are now in a position to defend RET against Sennett’s attack. For suppose we now understand Plantinga’s basic theistic beliefs, which are generated via the inductive procedure mentioned earlier, in such a way that they are justified or known in the Chisholmian particularist sense. Then it appears that we may conclude that they are immune to all of the objections raised by Sennett. To see this, let’s examine each of his objections in light of this particularist construal of properly basic theistic belief.
3.2 Against the charge of community-relativity
First, on such a construal, our method of determining such beliefs doesn’t entail that they are community-relative in any damaging sense. To see this, consider what it is that makes one worry about the charge of community-relativity. Intuitively, the worry stems from the fact that we want an epistemological theory that implies that we have (or can get) a reliable indicator of when a given proposition is true. Let’s call the thesis that is expressed in this want, ‘The Reliable Indicator Thesis’ (RIT):
(RIT): (i) There is a reliable indicator I of when a given proposition p is true, and (ii) we have access to I.
We usually call such an indicator ‘justification’. Now the worry seems to be that epistemological theories that imply that justification is community-relative undercuts RIT. For according to such theories, two different people can be justified in holding their respective beliefs, even though the beliefs of one are incompatible with those of the other. But if so, then it would seem that clause (i) of RIT is false. For then it would appear that justification is not a reliable indicator of when a proposition is true. But RIT seems true. Therefore, we should reject community-relative epistemological theories.
This worry is understandable. However, I think it is ultimately an exaggerated worry. For according to Plantinga’s version of “community-relative rationality”, the fact that people can rationally hold beliefs that conflict with beliefs rationally held by others does not rebut clause (i) of RIT. Rather, it merely rebuts the thesis that justification is an infallible indicator of when a given proposition is true. But then no plausible theory includes that thesis. For, ever since Gettier, we have learned to be content with (less than perfect) reliability from justification. If it ever became clear that particularist justification of basic beliefs (i.e., seeing that a basic belief is justified) was unreliable (below 51% of the time, say) in indicating the truth, then of course Plantinga’s theory would be damagingly community-relative. But this hasn’t been shown, so we don’t yet have a damaging objection to PF.
That the worry under consideration is exaggerated can also be seen by considering Plantinga’s epistemology from another angle. Recall that according to his theory, beliefs taken to be properly basic (at least the one’s Plantinga is concerned with, viz., theistic beliefs) are only prima facie justified: they are always open to the threat of defeat. Now, intuitively, if a belief is true, then it cannot be ultimately defeated. And if a given belief repeatedly survives criticism and potential defeat, then this fact is an indication (though, admittedly, not an extremely strong one) that such a belief is true. So it isn’t merely justification per se that is a reliable indicator of truth. Rather, it is (so far) undefeated justification that fills this role. So consider our two people with their prima facie justified beliefs, each of whose beliefs contradict those of the other. In defense of Plantinga, we might say that, true enough, both are prima facie justified in their beliefs. However, since reality is one determinate way, at least one of the beliefs of one of them is false. But if so, then the belief(s) of one of them is merely prima facie justified, while the conflicting belief(s) of the other are ultimately justified. And the system of defeaters utilized in Plantinga’s system will (ideally, at least) weed out the merely prima facie justified beliefs. If his epistemology were damagingly community relative, then it would deny these last two points, asserting instead that the beliefs of both people are ultimately justified.
From the above discussion, we have surfaced a criterion for deciding whether an epistemological theory is damagingly community-relative:
(DCR): For any epistemological theory E, if (i) E implies that one’s basic beliefs can’t be seen to be known/justified, (ii) E implies that one’s basic beliefs are indefeasible, or (iii) E implies that logically incompatible basic beliefs are ultimately justified, then E is damagingly community-relative.
But we have seen that Plantinga’s epistemology doesn’t fall prey to DCR. So, unless Sennett can give us a more plausible version of DCR, such that PF falls prey it, I submit that RET is in the clear on this score.
An analogy might help to make clear the above point. Consider two courthouses: GPI and IPG. In GPI, a person is considered guilty until proven innocent, and in IPG, one is considered innocent until proven guilty. Now, clearly, it is usually much harder to establish guilt in IPG than it is in GPI. Similarly, it is usually much harder to establish innocence in GPI than it is in IPG. However, it seems that it is possible that in both courthouses, the jury may come to a reasonable verdict as to the guilt or innocence of a given defendant. Analogously, there are two types of epistemological theories: particularist and methodist. The former theories are similar to GPI (call such theories ‘GPI-theories’), and the latter theories are similar to IPG (call such theories ‘IPG-theories’). Clearly, Plantinga’s theory is an example of a theory similar to IPG. Now basic beliefs are like the defendants in the courthouses. For it is usually more difficult to determine the proper basicality of a given belief with GPI-theories than it is with IPG-theories. Similarly, it is usually more difficult to determine the improper basicality of a given belief with IPG-theories than it is with GPI-theories. However, it seems possible that both types of theories can make the correct determination of such facts. With that said, let’s move on to Sennett’s next objection.
3.3 PF and normativity
In light of our discussion of the particularist features of Plantinga’s reformed epistemology (namely, its inclusion of PT), I think it is clear that Sennett is mistaken in claiming that PF is merely a descriptive theory of what some people naturally find themselves believing. For according to PF, beliefs are not appraised as properly basic merely because every member of the relevant epistemic community happens to form them in certain circumstances. To the contrary, a necessary condition for a belief’s being properly basic is that it be seen to be justified or known for the person in question. On this account, then, it is not enough that one finds oneself having a certain kind of belief in a certain kind of circumstance. They must also be able to see that the belief is justified. And it is this feature of PF that makes it a normative epistemological theory. Therefore, Sennett’s objection here is a failure.
3.4 SF is not rationally preferable to PF
Finally, even granting that SF can, in principle, be preferable to PF for persons that have never been in the relevant circumstances that generate properly basic theistic belief, SF is not preferable to PF for people that have been in such circumstances. For, given that no epistemology is justified for a person if it doesn’t account for all of their justified beliefs (i.e., all of their beliefs that they can see to be justified), then if a person has a properly basic theistic belief, then PF is preferable for that person.
Furthermore, and quite apart from the last point, Sennett’s argument for the objection under discussion is problematic in itself. For recall that, according to Sennett, the property of Universal Sanction is just the property of being a belief (i) that is affirmed by virtually all people under normal circumstances, and (ii) that can’t conceivably be countenanced as unjustified by healthy human beings. But if so, then Universal Sanction is subject to the same criticism with which he charged PF: it doesn’t speak to whether such beliefs are justified; rather it merely tells us that we need beliefs that have Universal Sanction to live in a way recognizably similar to how we now live. But if this is what Universal Sanction amounts to, then it is clear that it doesn’t do any normative work with respect to basic beliefs. That is, having Universal Sanction doesn’t make a belief likely to be true. Of course, it’s clear that beliefs that bear Universal Sanction are justified. But according to PF, what justifies such beliefs is not their having Universal Sanction. Rather, they have Universal Sanction because they are seen to be justified (in the particularist sense). Sennett has the reasoning backwards. Thus, Sennett’s justification for his inference from (6) and (7) to (8) is faulty, and so his argument is unsound. Therefore, I conclude that Sennett’s objections are unsuccessful, and so he has not shown that theists aren’t justified in taking belief in God as properly basic.
3.5 A final worry
I suspect that many are still not satisfied with Plantinga’s particularism. For it seems that it allows too many beliefs to be prima facie properly basic. For recall that Plantinga’s is an “innocent until proven guilty” epistemology: Many of one’s beliefs are considered to be justified unless or until one has reason to think otherwise. Of course, as we have seen, not all basically held beliefs have this status. Rather, they must first pass our (inductively generated) criteria of proper basicality. However, it must be admitted that it is plausible to think that some beliefs that are in fact false will be considered to be properly basic, and further that these will never, in fact, be defeated. But isn’t this an unacceptable conclusion?
No. To see this, consider GPI and IPG from our “courtrooms” analogy. Now it seems that while both GPI and IPG have their strengths (e.g., they can both be used to get the right verdict), they also have their distinctive weaknesses. For while GPI convicts nearly all of the people that should be convicted, it also, in all probability, convicts many that should not be convicted. Similarly, while IPG acquits nearly all of the people that it should acquit, it also, in all probability, acquits many people that it should not acquit. Which courthouse is better? It depends upon what you think is more important: convicting the guilty or acquitting the innocent. If you think the former consideration is most important, then you will think that GPI is best. But if you think the latter consideration is most important, then you will think that IPG is the best. Similarly, while GPI-theories and IPG-theories each have their strengths, they also have their distinctive weaknesses. For while the former theories prevent many false beliefs from being reckoned as justified, they also, in all probability, prevent many beliefs from being considered justified that ought to be so considered. Similarly, while the latter theories consider many beliefs to be justified that ought to be so considered, they also, in all probability, consider justified many beliefs that ought not be so considered. Which type of theory is better? Again, it depends upon what you think is more important: believing as many truths as possible or disbelieving as many falsehoods as possible. If you think the former consideration is most important, then you will think that GPI-theories are best. But if you think the latter consideration is most important, then you will think that some IPG-theory is best (and, of course, PF is an IPG-theory). While I think the latter is the more important of the two, in this paper I just want to make the weaker point that it isn’t clear that the latter isn’t the more important. Thus, I think that this objection isn’t in any way fatal.
In summary, then, we have taken a brief look at Plantinga’s version of reformed epistemology, as well as Sennett’s critique of it. We then clarified it in light of Sennett’s criticisms, concluding that PF can adequately deal with them. But if so, then unless we can think of other grounds that would defeat PF, then we may conclude that RET is in the clear: belief in God is properly basic for theists.