Notes on Plantinga

By Vern Crisler

Copyright, 2002


1.  Deontologism in Epistemology

Alvin Plantinga speaks of going back to the “fountainheads of Western theory of knowledge, those twin towers of Western epistemology, Descartes and Locke.”  (Warrant: The Current Debate, hereafter, WTCD,  p. 11).  He speaks of this “tradition” as what he calls “classical foundationalism,” and connects up this “tradition” with what he calls the “tradition of classical internalism.”  In his discussion, he wants to “get a proper understanding of justification and internalism, to understand the basic internalist insight.…”


In expressing these views, Plantinga claims that for Descartes and Locke, “deontological notions enter crucially.”  He quotes from Descartes, who had said,

“But if I abstain from giving my judgment on any thing when I do not perceive it with sufficient clearness and distinctness, it is plain that I act rightly….” [from Meditation 4].

He also quotes Locke to similar effect, who had said that:

“Faith…cannot be afforded to anything, but upon good reason….He that believes, without having any reason for believing, may be in love with his own fancies; but neither seeks truth as he ought, nor pays the obedience due his maker…” [from, An Essay Concerning Human Understanding].

On the basis of these quotations, Plantinga claims that:

“According to Descartes, being justified is being within our rights, flouting no epistemic duties, doing no more than what is permitted.”

And that for Locke,

“To act in accord with these duties or obligations is to be within one’s rights;….it is, in a word, to be justified.”

In my opinion, it is historically inaccurate and highly unfair for Plantinga to place Descartes and Locke in any proximity to what he calls “deontological epistemic justification” (WTCD, p. 14).

The fact is, Descartes and Locke are not defining justification, or what it means to be justified.  They are giving a reason for why one should be rational in one’s beliefs.  In other words, they are giving a justification for justification.  To hold that they are defining the concept of justification—as Plantinga appears to think—would be to make them guilty of defining a concept by the very same concept that needed defining.  But clearly, Descartes and Locke are giving a reason why cognizers should be rational.  It seems equally clear that Descartes and Locke could have easily given other reasons for being rational.  An economist, for instance, might say that one should be rational because it’s the most economical means for achieving a cessation of felt-uneasiness.  A businessman might say that rationality is the best way of making a profit that could then be used to further capitalize one’s business.  An accountant would say that being rational is the best way to equalize one’s assets with one’s liabilities and business equity.  There are a whole host of reasons that one could come up with—some more or less better than others perhaps, and many that could be conjoined—but none of these reasons is itself a definition of justification.  Giving reasons for justification in these cases isn’t normally construed as talk about “economic” epistemic justification, or “entrepreneurial” epistemic justification, or “debit-credit” epistemic justification
¾and it would be silly to think that’s what was meant.  For that reason, it’s also silly to talk about “deontological” epistemic justification in Descartes and Locke simply because they were giving reasons for justification.

Plantinga goes on to say that Locke’s “central thought” is that,

“being justified in holding a belief is having fulfilled one’s epistemic duties in forming or continuing to hold that belief.”


No, this can’t be right.  It is incorrect for Plantinga to refer to a Cartesian-Lockean deontological conception of justification.  This is based on a fallacy of equivocation over the meaning of the term “justification.”  It is one thing to be morally justified in being rational; it is quite another to be epistemically justified in one’s beliefs—but Plantinga’s discussion slides back and forth between these two types of justification.  It is true that for Descartes and Locke, one is morally justified in being rational, but being morally justified in being rational is not what it means to be evidentially justified (having a “sufficiently clear perception” of the truth of anything pertaining to a judgment). 

It is unfortunate that Plantinga failed to keep these distinct concepts of justification apart, for then he would have recognized the erroneous nature of his claim that Descartes and Locke are remote ancestors of the “deontological” view of epistemic justification.



2.  Evidentialism


Evidentialism in epistemology appears to be the view that evidence is that state of affairs which confers the property of rightness on a claim about how things are.  There are some who give evidentialism a positivistic twist (broadly construed) so that it’s about a moral or professional test as to what claims are to be accorded prestige and honor, and which claims are to be the subject of loud and haughty disdain or even of sub-voce snickering.

W. K. Clifford famously expressed the latter view when he said, “it is wrong, always, everywhere, and for anyone to believe anything upon insufficient evidence” (Quoted in Plantinga, WTCD, p. 26).  But why accept such a view of evidentialism?  Indeed, by his own criterion Clifford would be wrong to believe in such a claim since it doesn’t appear to be supported by sufficient evidence.  Unfortunately, Plantinga does not spend enough time addressing the differences between Clifford’s peculiar brand of evidentialism and the sounder views of evidentialists such as Conee and Feldman, whom Plantinga quotes as saying:

“Doxastic attitude D toward proposition p is epistemically justified for S at t if and only if having D toward p fits the evidence S has at t.”

The position of Conee-Feldman is that one must have access to the evidence for one’s belief, and if one’s belief fits the evidence, then one is epistemically justified—i.e., one has knowledge.  Plantinga says that Conee and Feldman “do not make the deontological connection” in the way that Clifford did, but I wish that Plantinga had zeroed in on this difference, for it makes it impossible to equate epistemic justification with the narrower and narrow-minded notion of evidentialism as held by Clifford and other positivists (again construing positivism in a broad sense).

One of Plantinga’s interpreters, Michael Sudduth, appears to be making this equation between evidentialism and Cliffordian moral evaluation when he argues that for evidentialism, “Justification is an evaluative term.  It indicates the right sort of cognitive response to evidence” (paper on “Evidentialism”).

But the notion of justification as an evaluative term is descriptive primarily of Clifford’s views (and perhaps of others).  I do not see this in Conee and Feldman’s actual definition of justification, and neither, apparently, did Plantinga, as noted above.  They use the term epistemically justified, not morally or professionally or technically or pragmatically justified, the latter adjectives conveying the notion, more or less, of an “ought.”


Sudduth clarified his position further in response to my argument by agreeing that Conee and Feldman “retain the idea of obligation” and said, “If this entails deontologism, which it surely does, then their view has that element.”  Sudduth, however, stated that he could pronounce no verdict on their views because he was unable to determine how important such a notion of obligation was to them or how it was to be understood.


Conee and Feldman had said,


 “One ought to have the doxastic attitude that fits one’s evidence.  We think that being epistemically obligatory is equivalent to being epistemically justified” (Quoted in Sudduth, p. 2).


On the basis of this, Sudduth classifies them as having a deontological view of justification, though even in this case it doesn’t really seem a large matter for them, for Sudduth points out that for them, ”what matters is not trying your best but your belief actually fitting your evidence.”


Now the important thing to note about the “deontological” quotation from Conee and Feldman—“One ought to have the doxastic attitude that fits one’s evidence” [my emphasis]—is that it is not itself a part of their original definition of justification, which only speaks about a belief’s being justified if it fits the evidence.  Thus, while Conee and Feldman can be classified as possible members of the Cliffordian wing of evidentialism—because of their deontological notion—their original definition of justification is in accord with a broader form of evidentialism, which finds deontologism quite irrelevant to justification.  This fits with their view that “trying your best” isn’t sufficient for justification.


[Note: Since writing the above, Conee and Feldman have written: “According to deontological conceptions of epistemic justification, one has a justified belief in a proposition when one deserves praise (or does not deserve blame) for having the belief…(or believing it violates no duty or obligations)…. We deny that internalism depends on a deontological concept of justification” (Hilary Kornblith, ed., Epistemology: Internalism and Externalism, p. 239).  Since Conee and Feldman make it clear that they are internalists, it follows that they also likely reject deontologism as a proper definition of justification.  Their use of the phrase “ought to have” a certain attitude that fits one’s evidence is then not a definition of justification, but very much in the Descartes-Locke tradition of providing a reason for being reasonable.  And this in no way can be described as epistemic deontologism.]

A basic mistake made by some is to define justification in terms of its results.  Thus, to say that justification is that process whereby one comes out with mostly true beliefs rather than false beliefs, is to substitute a result for a definition.  This mistake was made by Robert Nozick in his “tracking” theory of justification, which defines conditions for knowledge as the view, among other things, that someone would believe a claim if it were true, and wouldn’t believe a claim if it weren’t true” (See Alvin Goldman, Epistemology & Cognition, p. 45ff. for further criticism).

Another mistake is to define knowledge in terms of ”reliable” processes.  I will argue, first, that an unreliable belief-forming process can result in knowledge, and secondly, that a reliable belief forming process, even with conditions of proper function, appropriate environment, and good design plan, and where the truths under consideration are revealed by God himself, is not sufficient for knowledge.



3. Reliabilism & Proper Function

Plantinga’s book, Warrant & Proper Function is puzzling, not because Plantinga lacks skill or knowledge or professional competence, but because he is trying to repair a theory of justification—reliabilism—that fails to be a theory of justification at all.  I don’t mean to be harsh in my assessment, but you really just can’t repair the preposterous.


Commenting on one of the main purveyors of reliabilism, Alvin Goldman, Susan Haack says:

“[M]y thesis is not only that none of Goldman’s reliabilist accounts of justification is defensible; it is also that the shifts and changes in Goldman’s position are such as very strongly to support the conclusion that no reliabilist explication is defensible” (Evidence & Inquiry, hereafter E & I,  p. 152).

I don’t want to repeat all of Ms. Haack’s criticisms of reliabilism, so I’ll tackle it from another point of view.  The basic contention of reliabilism, even a repaired version, is that justification needs an “external” condition, a cause of belief that is reliable or truth-conducive.  Plantinga adds further qualifications to these external causes, namely that of proper function, appropriate environment, good design plan, etc.—call them Plantinga-conditions.  This is supposed to be over against the “traditional” view, which sees justification as ”internal,” consisting of a fit between beliefs and evidence.

From a scholastic perspective, P. Coffey makes an important point about knowledge:

“Our judgments are intellectual acts which are caused, both as to their actual happening (quoad exercitium actus) and as to their affirmative or negative quality (quoad specificationem actus) by a variety of influences.  These may be all described as causes of assent or belief.  Some of them, however, are subjective or psychological.…They are non-intellectual in character….From these subjective influences we can distinguish other causes or motives of our assent to any judgment as true, causes which we describe as grounds or reasons of our assent, motives which are directly intellectual….and they may be collectively catalogued under the comprehensive title of evidence”  (Epistemology, p. 257).

The thing to note here is the distinction between intellectual and non-intellectual causes for belief.  While Coffey believes that non-intellectual causes have some value as causes of our assent, they are valuable only insofar as intellect “can recognize in them indications, indexes, evidences, of the truth of the judgments to which they prompt our assent” (p. 258). Thus, non-intellectual causes are valuable not because they are truth-conducive, but only insofar as they are truth-indicative.  Of course, to recognize such processes as indicators of truth is already to give up the externalist notion of justification.

Non-intellectual causes are physical, psychological, or metaphysical.  It is the contention of “evidentialists” that such causes are not relevant to the justification of a belief.  On the one hand, a cognizer could have unreliable belief-forming processes
and still have knowledge, and on the other hand, could have a completely reliable belief-forming process and still not have knowledge.

Take the former instance. Say that S’s belief-forming processes are unreliable.  He has either a Cartesian demon or a nefarious neuro-scientist who is causally prompting his knowledge episodes.  Now suppose that just by accident, S believes that proposition p is true and knows the reason why it is true (because it is based on truth-indicative evidence a, b, and c).  In this case, S has knowledge, even though his knowledge episodes are casually prompted by a completely unreliable source.  S would still have knowledge even if he believed in his heart that he was not justified in holding his beliefs.  Because S knows the reason why p is true, the knowledge he has obtained is not itself accidental (i.e., opinion).  What is accidental is that S would have a knowledge episode at this particular place at this particular time, for it was no part of the demon or neuro-scientist’s plan to produce knowledge in the cognizer at the time and place it happened.

Take now the next instance.  Say that S’s belief-forming processes are completely reliable, that Plantinga-conditions have been met.  Add further the condition that God reveals the truth of say p, q, and r to S.  Does S know p, q, and r if all these conditions are met?  The answer is no, and this is because despite the complete reliability of his belief-forming processes, and his meeting all the Plantinga-conditions, and the cause of his belief being metaphysically solid, S still does not know why p, q, and r are true.  Even if S believed he was entirely justified in his knowledge, and even if his knowledge was causally prompted by a wildly reliable metaphysical Cause, all that S can do in this case is believe that p, q, and r are true, and he does so on the basis of authority.  We could say at most that he has virtual knowledge, but we couldn’t say that he has actual knowledge per se.

The key to both of these cases is that in the first case, the cause of S’s belief is intellectual whereas in the second case, the cause of S’s belief is non-intellectual.  In the first case, S knows why and in the second case he doesn’t know why.


Coffey says,

“To say that the truth of the judgment is evident is to say that in the objective reality which is being interpreted…the mind clearly sees or apprehends adequate ground for the relation whereby the reality is interpreted or represented in the act of judgment: and not only sees this ground, but is conscious that it asserts and assents to the relation because the ground of it is really there and really apprehended—[so that this evidence] is not only the reason of the truth of the judgment…but is also and eo ipso the “motive” of the mind’s assent to the judgment” (Epistemology, p. 261).

This is the traditional position in epistemology, namely that a person has knowledge when he sees the truth of a belief claim
¾that is, when he has evidence for it¾and, furthermore, that the cognizer is cognitively prompted to believe in the truth of a belief-claim by such evidence.


Is there no value to reliabilism, or to proper function accounts of knowledge?  It is difficult to see what value they could have unless they could become truth-indicative in some way.  They appear to have more value as examples of “metaphysics” rather
than epistemology.  Or better, they have value on that fuzzy borderline between metaphysics and epistemology.  For while proper function, etc., do not enter into the question of one’s knowledge, they do appear to have some relevance to the question of one’s being a knower.  Being a knower, of course, is a necessary condition for one’s having knowledge, so reliabilism and proper function accounts do seem to have some value in countering certain gross forms of skepticism that would seek to undermine knowledge by undermining our status as knowers.  Otherwise, their application to epistemology must be deemed a failure.  However, insofar as reliabilism has increased philosophers’ interest in science, specifically cognitive science, then it has been beneficial to that extent.



4.  Naturalized Epistemology


The main motivation for reliabilism, or repaired versions of reliabilism, is the desire to do what W. Quine called “naturalized epistemology.”  Plantinga has admitted that his views on proper function, etc., are a form of naturalized epistemology.


Quine describes naturalized epistemology as the abandonment of “first philosophy” and claims that science is “not in need of any justification” beyond its own methodology.  He says, “Naturalism does not repudiate epistemology, but assimilates it to empirical psychology….[T]he epistemological question is in turn a question within science….”  Here we have the fundamental orientation of naturalized epistemology: it holds that epistemology is continuous with science, and from this we can determine the motivation for reliabilism or externalism in justification, namely that it opens the door for epistemology to become continuous with cognitive psychology.


H. Putnam speaks of Quine’s views as “puzzling.”  Susan Haack finds an unmistakable scientistic tone in Quine’s notion of epistemology:


“Because the traditional problems of epistemology do not lend themselves readily or obviously to resolution within the psychological or biological sciences of cognition, however, Quine then finds himself, in his scientistic mood, under pressure to shift and narrow the questions with which he is concerned¾to such a point that continuity with the familiar questions of epistemology is broken, and Quine finds himself tempted to cast doubt on the legitimacy of the old projects” (E & I, p. 122).


This has a very familiar ring to it, and almost takes us back to the days of Kant, and even earlier philosophers, who wanted to determine the limits of knowledge, and dismiss any claims that went beyond those (their chosen) limits.


Quine further claims that skepticism is an offshoot of science, i.e., that science prompts our skeptical doubts about the possibility of knowledge.  On the basis of this claim, he recommends the conclusion that “Epistemology is best looked on, then, as an enterprise within natural science.”  But skepticism wasn’t born out of science.  Rather, it was born out of sophistry.  In other words, skeptical doubts were raised not by scientists but by philosophers.  Therefore, Quine’s reason for epistemologists to become “liberated” and end up as “empirical psychologists” is false, and one need not adopt the view that epistemology is a “chapter of psychology.”  Haack points out that Quine went from the concept of the cogency of evidence to the concept of the reliability of evidence, and here we can see the main distinction between evidentialist and reliabilist accounts of justification.


Now, as I’ve said, Plantinga admits that his epistemology is a version of naturalized epistemology (see his WTCD, p. 45ff.), but he thinks he is repairing a defect in it.  In addition to reliable processes as being the thing that turns true belief into knowledge, Plantinga offers up proper functioning in an appropriate environment, a design plan aimed at truth, and a high-truth ratio for the resultant beliefs.


When Goldman first formulated his reliabilist theory he spoke of a truth-ratio that gave results better than 50 percent (“true upshots more often than not”), but he qualified this by stating that these results must be obtained only in normal worlds.  Here, Goldman was trying to exclude processes such as benevolent demons or clairvoyance as sources of high truth-ratios in our beliefs.  This led him to speak of “non-manipulated” environments.


Plantinga was faced with a similar problem when philosopher William Alston asked about whether one needed to know not only the grounds for one’s beliefs but whether one had also to believe that such grounds were a reliable indicator for thinking one’s beliefs counted as knowledge (see, WTCD, p. 44).  This appears to be a question about not only the reasons you have for a belief, but whether you also believe that those reasons were prompted by a reliable process.  For instance, if you have grounds or reasons for believing propositions p, q, r, and s, but also knew that you were suffering from a cognitive malfunction, you’d probably not have reason to think your beliefs p, q, r, and s counted as knowledge.  Alston’s question is obviously an attempt to tease out the answer whether one could still have knowledge, even if one’s grounds were prompted by unreliable sources.


In response, Plantinga says, inter alia, that in some cases the question of reliability needs to be known, and in other cases that it doesn’t.  He allows room for someone to believe he has grounds for a belief, without also believing those grounds are a reliable indicator of knowledge.  Goldman spoke of “normal worlds” but Plantinga doesn’t actually give a reason why someone could have knowledge even without knowing whether the grounds were a reliable indicator of truth.  He says nothing of normal worlds but merely concludes, after dancing around the problem for a page, with the somewhat lame comment, “And the point is that it is the complex, highly articulated nature of the human design plan that makes impossible simple generalizations of these sorts about rationality and warrant” (WTCD, p. 45).  Obviously, he cannot say that reliability must always be known, for then he would have to sacrifice the reliabilist character of his epistemology, and this is the same criticism Haack brings against Goldman, the main proponent of reliabilism. (See her chapter 7, “The Evidence Against Reliabilism” in I&T, pp. 139ff.)




5.  Warranted Christian Belief


And this brings us to the basic fallacy in Plantinga’s approach, one that Goldman couldn’t really solve in his own approach, and it is this: just because a process of belief formation is reliable so that we can say our beliefs are rational because prompted by a probabilistically high belief-forming process, it doesn’t follow from this at all that our resulting beliefs from this process count as knowledge.  You can be as reasonable and as rational as you want to be in your quest for knowledge¾and that’s commendable¾but good intentions just aren’t good enough as a response to skepticism.  Thus, if someone challenges your belief in God, it’s not good enough just to say you are rational in your belief or that you are being reasonable (cf., Locke’s the Reasonableness of Christianity).  Nor is it good enough to claim that the skeptic believes in things he can’t prove.  This is what W. W. Bartley called the “tu quoque” argument for religious belief (see his, The Retreat to Commitment, p. 72).  Plantinga’s argument for religious belief seems to be a variation on the same theme.


Bartley had argued that Protestants were answering the charge of irrationalism directed against them by pointing out that their critics were just as irrational in their beliefs.  Plantinga, inversely, argues against those same types of critics by saying that Christianity is just as rational as any of the critics’ dearly held beliefs (e.g., belief in other minds, etc.)  The idea in both cases is that the rationalist has no right to criticize the religious believer; he has, to use Plantinga’s term, no successful de jure objection to basic religious belief.  Despite the fact that Bartley did not succeed in his own defense of rationality (it turned out to be self-referentially incoherent), his point about the tu quoque argument is relevant.  All that a de jure argument of the tu quoque sort does is present a negative case for religious belief, not a positive one.  This is why Plantinga’s tu quoque is easily derailed by objections of the sort that he himself termed the “Great Pumpkin Objection.”


We’ve seen that reliabilism has undermined the notion of evidential justification for a belief, so that what’s left is a sort of idea that we’re fully warranted in our beliefs if we’ve gotten them through a properly functioning, well-planned, environmentally appropriate process that’s at least aimed at high truth-ratios.  Questions of whether such beliefs are actually true or not are not that important in this view.  One can have a basic religious belief (say that God exists) without having to prove it, because proof is not relevant anymore; only the reliability of the process of obtaining beliefs is the relevant thing.  Hence, the Great Pumpkin Objection arises: Why can’t a believer in the Great Pumpkin claim that he’s fully warranted in believing the Great Pumpkin will return at Halloween and that such a belief is properly basic for him?  This isn’t just a problem for Plantinga’s approach.  It’s a problem for all reliabilists, and for any belief.


Plantinga’s approach is not a “Retreat to Commitment” in Bartley’s sense.  He doesn’t intentionally opt for irrationalism (as say a Wittgensteinian-influenced religious believer might do), but the reliabilist character of his epistemology leaves him without a means for excluding irrationalism.  If he responds that we just know that the Great Pumpkin doesn’t exist, or that it’s irrational to believe that he does, then he has sidestepped the skeptical question.  The (friendly) skeptic does not want to know whether such a belief is irrational or should properly be excluded as a case of knowledge.  He could very well agree with Plantinga that they should be excluded.  Moreover, the skeptic is not arguing about whether irrational beliefs can be properly basic, nor is he arguing about the  number or population of beliefs that can be properly basic.  The skeptic’s worries are not about proper basicality as such, nor are they about what counts as irrational belief.  Rather, what the skeptic wants to know is why Plantinga would exclude some beliefs as irrational (and not properly basic), and why he would include others as rational (and hence properly basic).  Surely this is not an unreasonable question to ask.


Would Plantinga not have to say, in the long run, that some beliefs are irrational and not properly basic simply because they are produced by a process that is unreliable (such as clairvoyance), and wouldn’t he also need to give us the reasons why such processes are unreliable?  But if he does that, he falls under Haack’s criticism, namely that in order convincingly to answer the skeptic he would have to give up the reliabilist character of his epistemology.


The other avenue would be for Plantinga to claim that standards of rationality or epistemic aptness are relative to the speech community or noetic tribe that happens to set the standards of epistemic obligation within a particular epistemic region.  This is the road already scouted by Wittgenstein and Rorty, but then Plantinga would fall under the criticisms of Bartley that such framework relativism is ultimately a tribalistic outlook and has no principial way of distinguishing rational action from irrational action.  Because framework relativism reduces rationality to rationality-within-a-framework, a rationalist who attempts to make use of it can provide no intellectually plausible grounds for a rational individual or society to criticize terrorism:  “Indeed, the rationalist has no reply to it that is effective even from within his own point of view” (RtC, p. 75-76).


It would be better for Plantinga simply to give up reliabilism and proper function accounts of epistemology and instead adopt evidentialism (broadly construed).