Kai-man Kwan
SB=Swinburne                                SE=sense experience 
RE=religious experience                 TE=theistic experience

        In his book The Existence of God, Richard Swinburne presents an argument from RE which utilizes an epistemological principle which is dubbed the Principle of Credulity:
(PC)     If it seems to me that x is present, then it is rational to believe that x is present unless there are special considerations to the contrary.

SB argues that it is a fundamental principle of rationality apart from which we cannot provide any noncircular justification of either ordinary perception or memory.  Then using this principle, SB formulates the following argument for the existence of God:
A) It seems to me that God is present.
B) There is no good reason to think either God is non-existent or not present.
C) Hence it is rational (at least for me) to believe that God is present.

        In his book On the Nature & Existence of God, Richard Gale spent a long chapter on the argument from RE (chapter 8) & offered criticisms of versions of this argument which are defended by Wainwright, Gutting, Alston & SB.  In this essay I intend to look at Gale's discussions in some details & try to respond to him.

        Gale declares that his concern is with "nonsensory experiences that their subjects take to be "direct" or noninferential experiences of God, the eternal one, & the like.     ... those that are taken by their subjects to be of some objective reality that transcends consciousness & exists independently of being experienced."(p.286)  He believes that for these experiences to have "evidential status, these experiences must not only be veridical but cognitive as well. ... It could qualify as veridical in virtue of the apparent object of the experience causing the experience in the "right way," but not count as cognitive since it does not constitute evidence or warrant for believing that this object exists & is as it appears to be in the experience. ... A cognitive faculty should stand on its own two epistemological feet.  ... REs, although possibly veridical, could not be cognitive.  Even if it were possible that their apparent object exist & be the right sort of cause of the experience, we could never know on the basis of these experiences either that this object exists or that the experience is caused in the "right way" by it. ... a RE also could not qualify as a veridical perception of an objective reality, even if its apparent object were to exist & be the cause of the experience."(p.287)  The reasons for these claims are then elaborated.
        Gale believes that "If a type of experience E is cognitive, then the occurrence of an E-type experience counts as evidence for the existence of its intentional accusative, even for those who have not had this experience or even any E-type experiences."(p.287)  The reason is: "That a cognitive experience's evidential status is not person-relative is a fundamental conceptual requirement for an experience having cognitive status".  It means that Gale believes that if REs are to have evidential status for some people, they would also be evidence for all.  Gale further says that "Every argument (for the cognitivity of REs) that I know is based on an analogy with SE" (p.288) whose general form is:
  1. REs are analogous to SEs;
  2. SEs are cognitive; therefore,
  3. REs are cognitive. 
        His general verdict on this kind of argument is 'failure' because he thinks that "the various analogies are very thin & that the disanalogies are very extensive & deep, thereby destroying the different versions of the analogical argument."  He also thinks that "The fate of all these nonsensory faculties of "perception" (e.g. intellectual intuition, moral intuition) are intertwined such that if my polemic vs the cognitivity of RE succeeds, it works equally well mutatis mutandis these other nonsensory "perceptual" faculties." (p.290)  He discusses three forms of this analogical argument (AA) which differ in strength of the analogy with SE asserted. 

Weak Form: LG Fideism
        This form of AA points out that "both (SE & RE) serve as the basis on which objective existential claims are made within different ongoing language games, such a game being a normatively rule-governed human practice of using language." (p.291)  "A language game is individuated by the part of its ontology- what it is about or refers to- that has its categoreal nature defined by the rules of the game.  Each language game has epistemological autonomy in that it has its own principles of rationality & justification, especially in regard to the unique part of its ontology.  Language games can overlap in their ontologies, categories, & epistemology.  Whether a type of experience counts as cognitive is relative to a language game."(p.292) 
        An of-O-type experience- an experience that has O as its apparent object or intentional accusative- counts as cognitive relative to language game L just in case:
(a) The vast majority of the L-players upon having an of-O-type experience assert or believe that they are having an experience of an objectively existent O or that there exists an objective O & agree that it is correct to do so; &
(b) L contains a noninductively derived rule of presumptive inference that if one has an of-O-type experience, then one probably has an experience of an objective O or there exists an objective O.
        Gale's major criticism of this form is that the analogy of this language game with SE is too thin: "whereas the existential claims based upon SE have only a prima facie warrant, since subject to a battery of interconnected tests & checks, there are no defeating conditions for existential claims based upon RE ... (therefore) there is no basis for drawing the distinction ./. his actually perceiving God & it just seeming to him as if he is perceiving God."(p.294)   Hence the veridical-unveridical distinction does not hold in this language game & this disanalogy with SE is fatal to the argument.   He also points out that this is not the actual religious language game which allows for the possibility of error.

LG Fideism with any Old Tests
        In this form of AA, the further analogy drawn upon is that "in both games there are checks & tests or defeating conditions for these games"(p.291).  Furthermore the argument asserts "considerable epistemological autonomy to these games, each having its own distinctive criteria of rationality & epistemic justification".  However, Gale thought it is wrong because "This game is part of the larger religious language game, which employs many of the same principles of deductive & inductive logic as does the sensory-based physical object language"; for the "intellectually sophisticated adults", "God exists" is not basic because "they are willing to give various sorts of epistemic justifications for it."(p.296)    "There is only one big language game of which different so-called language games are interconnected parts.  The RE language game ... is part & parcel of the physical world one"(p.296)  Another objection of Gale is that this argument justifies far too much!  He asks us to consider a RE language game "in which the only test or check is based upon what the cult leader says."(p.297)  He thinks though it is logically possible that this language game is not falsified by other factors, yet it is still obvious that this language game is not cognitive. 

LG Fideism with Analogous Tests
        The third form of AA asserts further that both SE & RE have similar tests for veridical experiences.  Alston, Gutting, SB, &  Wainwright are listed as defenders of this argument.  "All 4 begin with an a priori presumptive inference rule, subject to defeating conditions, linking how things perceptually seem with how they really are that applies to SEs & then argue by analogy that it should also hold for REs.  The rule must be a priori because in both cases any attempt to justify its adoption or establish the reliability of the doxastic practice of making existential claims on the basis of the type of experience in question will suffer from vicious circularity. ... If we are willing to accept a circular justification in one case so as to escape scepticism, we should be willing to do so in the other.  It would not only violate the principle of parity but also be an unwarranted display of chauvinism for us to demand that the reliability of the RE doxastic practice be established by appeal to SE."(pp.298-9)        "That the two doxastic practices of belief formation are to be accorded parity is justified by showing that they are analogous in cognitively relevant aspects, especially in regard to both containing their own checks & tests." (p.299)    In order to evaluate the above argument, Gale lists eleven tests of SE:
  "1. Logical consistency.  Are the sensory-based claims of the subject logically consistent?
  2. Empirical consistency.  Are they consistent with well-established empirical facts?
  3. Existence of apparent object.  Are there good arguments either for or against the existence of the apparent object of the SE?
  4. Reliability of subject.  Does the subject have a good track record in his past perceptual claims?
  5. Agreement.  Do the SEs had by others under normal or standard conditions agree with the subject's?
  6. Continuity ./. contents.  Does the content of the experience stand in the right sort of lawlike relations to those of the vast majority of the subject's preceding & succeeding experiences?
  7. Prediction.  Can future SEs of the subject & others be predicted on the basis of the assumption that the experience is veridical?
  8. Proper position.  Was the subject properly located in space & time so as to have been able to perceive the apparent object of his SEs?
  9. Physiological state of subject.  Was the subject's sensory faculty in good working order?
  10. Psychological state of subject.  Did the subject have any psychological impediments to his being able to perceive adequately, for instance, phobias, prejudices, or the like?
  11. Causal requirement.  Did the apparent object cause the experiences in the right way?
        The analogical arguer takes these tests to form a "cluster" in that no one of them is either necessary or sufficient for veridicality, but satisfaction of a sufficient number is such." (pp.302-3) 
        After listing these tests, Gale then proceeds to argue that the tests of RE are actually gravely disanalogous with the above tests for SE.  Tests 1-3 are applicable but they only specify the bare minimum & passing them does not make any positive case for the cognitivity of the type of experience.  (Though he doubts that mystical experience would pass tests 1-3.)  Test 4 is basically parasitic on other tests.  So he starts to discuss from test 5 onwards.
        On test 5: agreement.  "The agreement or disagreement of other observers is relevant only if their observations are made under so-called normal or standard conditions.  We have a pretty good idea of what they are for SE.  Their mystical or religious analogues vary from one mystical or religious tradition to another but usually involve asceticism, meditation, breathing exercises, & the like."(p.307)  But "mystics (i) count as confirmatory mystical experiences that occur without following the prescribed "mystical way" & (ii) do not count failure to have a mystical experience under the prescribed conditions as disconfirmatory."  Naturally, Gale's complaint is that it looks like a "heads I win, tails you lose" sort of con game.  
        On tests 8-10.  Furthermore  "since the notion of a normal observer & normal conditions has no application to RE, neither will tests 8-10, since their requirements of a physiologically normal observer properly stationed in space & time are parts of the normal or proper conditions requirement.  And since these tests, along with the agreement test, lack religious analogues, this greatly weakens the analogical case for cognitivity."(p.308)  If it is still argued that we have some idea of the psychological & moral conditions that facilitate having a RE, then Gale would reply that "Such causal knowledge as we have pertains to the occurrence of a psychological occurrence or state- the having of an of-God-type experience- & not to its veridicality."(p.308)
        On test 6: "Since the apparent object of a RE ... is not one object among others in a nomically determined system, the continuity-./.-contents test cannot apply to it."(p.309)
        On test 7.  For once, Gale thinks that a test for RE is acceptable: the moral fruits test: "Since God is essentially good, it is probable that those who have a veridical experience of him will benefit morally & spiritually." (p.309)   As for other predictions Gale is more contemptuous.  For example, Gutting suggests that if a person has a veridical RE, he tends to have some more REs later.  To this Gale replies that: "The atheist is far more in need of God's grace than is the person who is already launched along the path of faith.  Thus, the atheist would seem more likely to have an of-God experience if God exists than would the believer, especially since the atheist's disbelief would help to offset the potentially freedom-canceling impact of the experience."(pp.310-1)  Gale admits that these are not sincere judgments.  This retort just serves to show "how little basis there is for making any probabilistic estimates at all in this area.  This radical unpredictability stands in marked contrast with the familiar sort of probabilistic predictions that are made within quantum & thermodynamic theory, as well as of the free actions of persons"  & he finds it a cognitively relevant disanalogy.
        Gale then argues that counter-examples can be produced to Gutting's criteria which, according to Gale, are:
"  (1) those who have had such experiences once would be likely to have them again;
   (2) other individuals will be found to have had similar experiences;
   (3) those having such experiences will find themselves aided in their endeavours to lead morally better lives." (p.311) 
        He suggests that experience of orgasm passes analogues of Gutting's tests & hence "Gutting's criteria for an objective or cognitive experience let in too much"!   He voices a similar objection to Alston.  According to Gale, Alston proposes these criteria for a language game which is objective:
        "(A) The language game is played successfully in that the form of life it involves has proven viable, (B) no assertion warranted by the rules of the game contradicts anything that is warrantedly assertible within other language games we play, & (C) the game is internally consistent because things happen by & large in a manner that is explicable in terms of the categoreal nature of the entities within the game's ontology." 
        Gale again objects by saying that the language game of pain satisfies Alston's criteria also! (p.312f.)  "Thus, the fact that the religious LG satisfies conditions (A)-(C) does not establish that of-God experiences are objective."(p.313)   "One might try to block my counterexamples to the Gutting & Alston criteria for the objectivity or cognitivity of a kind of experience by restricting these criteria to kinds of experiences that are either putatively perceptual or such that the apparent object of the experience would be objective if real.  The latter restriction would be no help, since pains & other types of subjective experiences have apparent objects that are real ... but not objective in the sense of being mind-independent & a common object of the experiences of different persons.  Such restrictions, furthermore, are viciously regressive, since people, especially philosophers, often disagree about whether a certain type of experience is putatively perceptual or such that its apparent object is objective if real at all."(p.313)  Gale realizes that  there are also philosophers who think SE is not objective.  So "What is needed is a criterion for distinguishing ./. cognate & objective accusatives so that these disputes can be settled. ... Surface grammar & ordinary language are inadequate"(p.314) 
        On test 11: caused-in-the right-way test. He notes SB's argument that if there is a God, then every experience of God is veridical & replies: "God, however, is essentially benevolent, & thereby would not cause an of-God experience in a devious or ignoble way.  SB confounds the true proposition that it is impossible for God to cause an of-God experience in the wrong way with the false proposition that it is impossible for an of-God experience to be caused in the wrong way, that is, a way that would preclude it having God as its cause."(pp.314-5)  His examples are REs caused by ingesting LSD or unconscious sexual desires.  He thinks it false that veridical REs can't have a sufficient natural cause because "an event can be causally necessary for a given event though there intervenes ./. them a causally sufficient event.  ... a veridical RE can have a sufficient natural cause, provided it is consistent with the nature of the experience's apparent object to allow such a cause."(p.315)  However, the problem is: "how it could be determined that a RE is caused in the right way by its apparent object. ... Whereas we can determine on the basis of SE that a given SE is caused in the right way by its apparent object, we cannot determine on the basis of RE alone that a given RE is caused in the right way by its apparent object.  ...  Why can't we have an argument based upon REs for the existence of the apparent object of a given RE & its bearing the right sort of causal relation to the experience?  There can be such argument only if REs count as cognitive.  But they are cognitive only if they are subject to similar tests to those that SEs are.  And it has already been shown that this is not the case."(p.316) (Italics mine)   As a summary, tests 6, 8-10 in SE have no RE analogues.  Tests 5, 7 have at best quite weak analogues.  Test 11 is not applicable. 
        Gale further notes the problem of religious diversity.  There is "no analogous diversity of doxastic practices for basing claims about physical objects on SE that differ among themselves as to what counts as confirmatory & disconfirmatory of a given SE being veridical."  "This is a cognitively invidious disanalogy that should destroy the requirement to extend the presumptive inference rule from the SE to the RE doxastic practice in the name of the principle of parity .  There should be parity in their treatment only if they are sufficiently analogous."(p.317) 
        Gale then turns to the responses of the defenders & criticizes them.  SB admits that in RE we lack the knowledge of the perceptual conditions & hence tests 8-10 are not applicable.  However SB believes that only somewhat lessens the evidential force of REs.  Gale complains that "The basis of SB's "somewhat lessens the evidential force of" is obscure.  ... The in-principle impossibility of disconfirming the veridicality of a RE by appeal to the testimony of other observers ... calls into question the very applicability of the agreement test to it.  ... SB seems to have forgotten that he is supposed to be justifying the extension of the PC to REs on the ground that they are sufficiently analogous to SEs"(p.318) 
        Alston has also noted a lot of disanalogies but he offers plausible explanation of them.  Gale's criticism is that "to explain why the tests for sense & REs are not analogous is not to explain away the disnanlogies; & these disanalogies are devastating to the analogical argument."(p.319)  It is because Alston's demand for parity  is based on analogy.  (p.321)  He labels it  'Alston fallacy': "thinking that if he can give a categoreally based explanation for a disanalogy ./. the religious- & SE doxastic practices, it renders the disanalogy harmless."(p.322)   So Gale's conclusion is that the analogical arguers' treatment of the disanalogies is a failure.

Crucial Disanalogy
        However, Gale thinks that he has yet to produce his trump card.  He argues that "A veridical sense perception must have an object that is able to exist when not actually perceived & be the common object of different sense perceptions.  For this to be possible, the object must be housed in a space & time that includes both the object & perceiver.  ... there is no RE analogue to this cpt of objective existence ... Because of this big disanalogy, God is categoreally unsuited to serve as the object of a veridical perception, whether sensory or nonsensory."(pp.326-7)  This view enables us to explain why in SE we have the above tests: "Tests 5-11 are an organic unity of conceptually interrelated tests that psp the same cpt of a veridical sense perception of an objective reality, namely, that of a common space-time in which the subjects & objects of veridical SEs causally interact in a mostly deterministic manner & that permits those objects to exist when not actually perceived & be common objects of different perceptions, as well as to be individuated by their positions in space-time."(p.328)  The common worldview is that of "a common space-time receptacle in which the objective accusatives of veridical SEs are the common causes of the nomic-type coherence among the SEs of the differently positioned perceivers." (p.328)  He also thinks that it is a necessary truth: "no two empirical individuals of the same kind spatiotemporally coincide."(p.328) 
        He then embarks on a lengthy criticism of Strawson who tries to argue for the possibility of a sonic universe.  His aim is to argue that there is no alternative to the above cpt of an objective empirical particular which is a suitable object of a veridical SE. (p.340) 
        In this light he considers REs which have radically different grounds of individuation:- by properties.  This grave disanalogy would "disqualify experiences of them as veridical perceptions of objective particulars."  Furthermore "no analogous explanation can be given for how they can exist unperceived & be common objects of different perceptions to that which was previously given for empirical particulars."(p.341)  It also provides no explanation why some fail to perceive God or why there is just only one God as the object of REs.  He concludes thus: "whereas the worldview into which empirical particulars enter offers an explanation for the organic unity of the sensory tests, there is no analogous view into which the apparent object of REs enters that explains how the various RE analogues to these tests ... are interconnected. ... They form a mere heap, with no rhyme or reason why they should go together & serve to mutually reinforce each other."(pp.342-3) 

Gale's Misrepresentation of his Opponents
        Gale's discussion is in general lively, clear & incisive.  He also throws light on the organic unity of the tests of SE.  However, despite these merits, his case against the cognitivity of REs or his rebuttal of the case for the cognitivity of REs is seriously flawed.  Some minor points frist.  His understanding of LG fideism is superficial.  He regards it as a sort of AA but LG fideism is typically argued for on more conceptual & abstract grounds.  His claim that the RE LG belongs as a part to the physical world LG is also astonishing.  The reverse perhaps is true.  But his major fault is his serious misunderstanding of the more sophisticated defenders of the argument from REs.  It is astounding that he says that the only form of argument for the cognitivity of REs he knows is an analogical argument (AA).  I also believe it is not right to attribute this AA to Alston, Gutting or SB[1].  The AA starts by assuming the unproblematical cognitivity of SE & then bases the case on the analogy of RE with SE.  However, as I have elaborated in chapter 10, SB's argument from RE is better understood as a sort of transcendental argument (TA).  It starts by querying & investigating how we can establish the cognitivity of any sort of experience, including SE.  It then proposes an a priori epistemological principle like SB's PC or Alston's principle of presumptive reliability of a viable epistemological practice.  Since this principle is supposed to be a priori, it is not based on any contingent features of SE or RE nor on the analogy of RE with SE.  This interpretation is especially strained when he says SB is presenting the analogy in the "manner of rearguard action"!  (For SB, I think, that they are all epistemic seemings is the only crucial analogy.) It does not mean the TA cannot find further support from the AA: the two arguments are not mutually exclusive because the defender of TA usually also affirms the cognitivity of SE & some analogy of RE with SE.  Perhaps this partly accounts for Gale's misunderstanding.  If the above interpretation is correct, then Gale's incisive criticisms seem to have hit only a straw man.

Distinction Between the Analogical Argument & the Transcendental Argument for Religious Experience

        We have seen earlier the formulation of the AA.  Let us formulate the Transcendental Argument for RE (TA) as follows:
i) Any justification from experience is possible only when the PC is presupposed.[2]
ii) Hence the PC should be applied to all kinds of experience.
iii) Hence the PC should be applied to RE.
        Actually the two arguments are not mutually exclusive.  The analogical argument can be seen as a supporting argument to the transcendental one.  The considerations of analogy can also help to dispel doubts about the transcendental argument.   But it is important to see that the TA is not based on considerations of analogy & to refute the TA, either (i) has to be denied or the inference from (i) to (ii) impugned[3].  So it is not relevant to elaborate the disanalogies of RE with SE as a response to the TA.  Such considerations, if cogent, can only refute the AA but not the TA.  It is surprising that many critics still cannot see this point.  (It is shown, for example, by the response paper of Daniels to Alston.) 

        Usually people who dispute the application of the PC to REs would suggest that the PC is only applicable when certain set of conditions C is satisfied.  Then he claims that C obtains in the case of SE but not in that of RE.  So they argue that though the PC can apply to SEs, it does not actually apply to REs.  However we have to distinguish two claims:
(D) PC is applicable if C' obtains.  (C' as sufficient condition)
(E) PC is applicable only if C obtains.  (C as necessary condition)
        To support his argument, the critic has to claim (E) rather than (D).  Suppose C includes universality, checkability & so on.  Grant that these are sufficient conditions for PC being applicable.  Yet if they are not claimed as necessary conditions, there is no reason why their lack shows that PC is not applicable.  Now "eating arsenic" is sufficient for death but a man can still die of other causes, e.g. suffocation.  So the critic has to use (E) in his argument:
(E) PC is applicable only if C obtains.
(*) C does not obtain in the case of REs.
(F) So PC is not applicable to REs.
        Now the obvious trouble is how to establish (E) except by fiat or stipulative definition.  The critic has to rely on an analogical argument[4] himself:
- C obtains for sensory experience when PC is applicable to it.
- hence if PC is applicable to REs, C obtains for REs.
- C does not obtain for REs.  So PC is not applicable to REs.
        This argument on its own is exceedingly weak.  To carry conviction it has to be shown further that C is the reason why PC is applicable to SE, i.e. were C not to obtain, PC would then not be applicable to SE.  It further implies that to justify our belief that PC is applicable to SE we have to justifiedly believe that C obtains for SE.  To avoid circularity,  we cannot assume the applicability of PC to SE when we attempt to justify the latter belief.  This, I would argue, pose enormous difficulties for the critics. 

        The above argument seeks to find the necessary condition C for applicability of PC.  If this search is successful & C shown not to obtain in RE, then the TA would be effectively rebutted.  But the above discussion should convince us that to achieve this is a tall order.  However there is another way to undercut the TA.   We may ask why PC is applicable to sensory experience & the following answer can be given:
(D) PC is applicable if C' obtains.
(#) C' obtains for sensory experience.
(G) So PC is applicable to sensory experience.
        The idea is that we can justify the applicability of PC to SE by finding some sufficient condition C'.  Now notice that it can't be argued that since C' does not obtain for RE, PC is inapplicable to RE.  It is because C' is only the sufficient condition & not necessarily also the necessary condition for the applicability of PC.  So this argument can't rebut the TA directly.   However this can refute (i) above by showing that applicability of PC need not be presupposed; it can be inferred from C'.  But then we have to know or justifiedly believe that (D) & (#).  Again we cannot assume PC in justifying (D) & (#).  How it can be done is not easily conceivable.   For any likely candidate for C', both (D) & (#) seem unknowable if we do not assume PC.  Actually the above argument may be valid & sound but we can never use it to justify our trust in sensory experience.  If this line of argument is correct, it gives some substance to the claim that the PC is fundamental.  In another words, the PC seems to be an a priori rather than a posteriori principle.  If so, there seems no reason why it shouldn't be applied to REs. 
        Again the above argument is independent of considerations of analogy or disanalogy which heavily beset the AA.  So the answer given by the PC to the Disanalogy Objection is more radical: the disanalogy is not strictly relevant.  However the TA is not incompatible with AA.  So the latter, if persuasive, can strengthen the former.  The discussion of this section is rather abstract.  Let's now examine the analogical considerations & the above logical considerations will be borne out in more concrete details.

The Logic of Comparison of Kinds of Experience

Disanalogy & Unreliability
        Suppose it has been shown that REs & SEs are greatly disanalogous.  Does it show that REs are then unreliable?  It does not seem to follow.  This argument can only go through if analogy[5] with SE is a necessary condition of reliability of a kind of experience.  This premise is extremely dubious & I can see no way to establish it at all.  Perhaps the critic can press a more modest objection: although the disanalogy does not show that RE is unreliable, it does show that the presumption of its unreliability is justified.  Again I can see no basis except "epistemic chauvinism" for this line of reasoning.  Consider a parallel argument  against induction:
- Deduction is the only reliable inferential method we know.
- Induction is disanalogous with deduction.
- Hence there is a presumption of unreliability of induction.
        How good is this argument?  The second premise seems to be relatively secure.  Deduction is truth-preserving while induction is not.  The former deals with the necessary connection between propositions while the latter the "probabilistic" connection between contingent truths.  Some also say that deduction is non-ampliative[6] while the latter ampliative.   The differences are quite obvious.  However there are problems.   First of all the first premise begs the question by assuming that we do no know that induction is reliable while we know that deduction is reliable. This raises the question how deduction itself is to be justified.  Secondly why do we regard analogy with deduction is essential for a kind of inference to be reliable?  Is it not implicitly assuming that there is & can only be one kind of reliable inference?  There does not seem to be any justification for this assumption.  Exactly because the two methods are concerned with different kinds of epistemological connection, isn't it to be expected that the two methods should also show differences accordingly?  The fact that induction is disanalogous from deduction only shows that the two cannot be reliable in the same way or to the same degree.  It is still a live possibility that induction can be as reliable as it can be in the realm of contingent truths.  So this parallel argument against induction is not sound[7]

        Similarly the above argument against RE should also be rejected.   Why should we expect that there is only one kind of reliable experience?  Isn't it possible that our experience can have some access to different kinds of contingent truths?  If it is the case, isn't it to be expected that there will be several kinds of experience accordingly?  Each may be reliable in its own way & to different degrees.  How can we know a priori that it is not the situation we are in?  It seems to me the contrary expectation reflects the a priori nature of the presupposition of a narrow empiricism.  It is laid down before hand that only SE can be reliable & then it is used as the yardstick to measure all other kinds of experience.   Those who are disanalogous with SE are then deemed unreliable.  It should be clear that there is no obvious reason that we should follow this procedure.

        Perhaps it can be denied that the assumption that only SE & its like are reliable  is an a priori one.  It might be argued that this assumption is justified by the facts of the matter. (Daniels, p.489)  One SE can be confirmed by another & the whole past record of SE is impeccable.  Hence it justifies the assumption that only SE is the  reliable experience & the yardstick of other kinds of experience.  This defense, however, won't do.  This justification of the reliability of SE is dependent upon the reliability of memory.  Otherwise how can we be sure that the past record of SE is good?   But then memory is clearly disanalogous from SE in many respects.  The realm of SE is that of the physical world while that of memory is the past.  SE is mediated by sensations while images do not seem to be essential for memory & so on.  Hence if SE is the only yardstick, memory  does not seem to satisfy it.  In itself this is hardly a consequence that the "empiricist" will accept.  Not to mention that it undermines the earlier justification of SE.
        The appeal of narrow empiricism also derives from the success of science & is reinforced by the resulting scientific worldview.  It seems to show that SE is the only foundation we can rely upon.  But this is not correct[8]: the actual success of science depends on a lot of other epistemic capacities, e.g rational judgment, introspection[9].  These are all very disanalogous from SE.  So even on its own terms, a narrow empiricism can't sustain the idea that disanalogy with SE is enough for discrediting a epistemic capacity.  Our reply is quite general: it works with whatever amount of disanalogy.  This may seem counterintuitive.  Surely in cases when a kind of experience is very disanalogous with SE, e.g. producing massively inconsistent beliefs, we should discredit it?  This may be true but the reason for discreditation is not the disanalogy but the amount of contradictions. 
        In general disanalogy with a well-established principle is not a sufficient reason to discredit another principle unless there is reason to think that the known principle & its like are the only possible principles.  In the case of RE, it amounts to the assumption that "we were entitled to assume that, if it is veridical, religious cognition will be like other kinds of cognition which are veridical.  Prima facie we should however expect it on the contrary to be very different, since its object is so different." (Ewing, p.115)  This point has also been amplified by Alston.  He has delineated clearly several assumptions which , if granted, render the disanalogy intelligible:
"(A) God is too different from created beings, too "wholly other", for us to be able to grasp any regularities in His behaviour.
(B) for the same reason we can only attain the faintest, sketchiest, & most insecure grasp of what God is like.
(C) God has decreed that a human being will be aware of His presence in any clear & unmistakable fashion only when certain special & difficult conditions are satisfied." (Alston 1983, p.129)  Alston thinks that these assumptions seem to be central in many religions & they are in no way incompatible with the substantial reliability of REs.  Given them, we should positively expect REs to be different from sensory experiences.  We do not need to fully agree with Alston's theological explanation of the disanalogy.  It just serves to illustrate the fact that the disanalogy of RE with SE as such is not an adequate reason for asserting the unreliability of RE. 

Analogy & Reliability
        On the other hand some argue from the analogy with SE to prima facie reliability.  The argument can be formulated in this way:
A) SEs are prima facie reliable.
B) REs are sufficiently analogous with SEs.
C) Hence REs are also prima facie reliable.
        I'll call this "the Analogical Argument for RE" (hereafter abbreviated AA).  It has several characteristics.  Firstly, It takes the reliability of SE as the unproblematic starting point.  Secondly, the form of the argument is analogical & hence not deductively tight.  This, however, does not matter.  In many cases a good argument from analogy is as good as anything we can hope for.  To reject the AA just because it is an argument from analogy is clearly unjustifiable.  It is especially so when we have no foolproof argument to establish premise (A) in the first place[10].  In that case our confidence in the reliability of SE is necessarily tied to some characteristics of SE which we regard as marks of reliability, e.g. universality, testability.  If REs share all of these marks, I can see no way to resist the AA.  Similarly, the  argument can be very persuasive  if the analogy is strong.

        So the crucial premise is (B) which is unfortunately quite vague.  The proponent of the AA never says that REs are wholly analogous with SEs.  To say this is to invite immediate defeat.  So the analogy is only argued to be "sufficient" or "in the relevant aspects".  "What are the relevant aspects" is a controversial question & the debates revolve around this issue.  The critics say that the analogy is weak, insufficient, & not obtaining in the relevant aspects.  The defenders contrarily argue that the analogy, though incomplete, is significant, sufficient, & obtaining in the relevant aspects.  How are we going to resolve this dispute?  There seems to be no other way than by patiently looking at the various features of analogy & disanalogy,  & by evaluating the relevance or irrelevance of these features.  This we try to do later.
        Another point should be noted here.  The proponent of the AA contends that a sufficient analogy of RE with SE is a strong reason to suppose that RE is also PF reliable.  This contention, however, does not commit him to the claim that the lack of a sufficient analogy of RE with SE is a good reason for thinking RE unreliable.   Consider a parallel: the claim that a sufficient dosage of arsenic would cause your death does not entail that if you do not take arsenic, then you will not die.  This deduction is clearly fallacious because there is a million ways to cause death.  Similarly, although the proponent of the AA assumes that SE can be unproblematically taken to be a reliable source of beliefs, there is no reason why he cannot also think that there are several different reliable sources which are very different from one another.  Gale seems to think that if one adopts the AA, then one is also committed to the logic of the Disanalogy Objection.  I think it is just a confusion.
Analogy & Unreliability
        There can be another kind of argument: by pointing out the analogy of RE with some kinds of discredited experiences (widely regarded as unreliable) the critic may try to show the unreliability of RE.  For example it can be pointed out that REs are as a rule as "subjective" as aesthetic experiences are.  Then it may be argued that since the latter is cognitively unreliable (in the sense that they do not tell us any objective features of the world) the former probably also is.   As a response, it might be argued that the experience alleged to be discredited is not really so.  Otherwise, the evaluation process seems to be similar to the above.  The points of analogy or disanalogy have to be examined in details & then evaluated as to their relevance. 

Analogy & Rebuttal of charge of unreliability
        Sometimes the critics use various arguments to support the claim that REs are unreliable (which are not arguments in terms of analogy).  These arguments can be rebutted if analogous arguments can be constructed to discredit SEs.  Namely the features that are supposed to discredit REs are also present in SEs in an analogous way.  For example the deductive gap objection can be rebutted in this way.  In the past philosophers make heavy water out of this gap & it is extraordinary that the same gap in SEs is not mentioned at all.  It is no use to reply that SEs can be shown to be reliable by other means.  This only strengthens the impression that that gap is far from fatal.  If REs are still suspected this is not because of the logical gap, but because of the lack of other confirming features or alleged marks of reliability.  So considerations of analogy can still be used for a negative defense of RE even though such considerations are not strong enough to mount a positive case for RE. 

Marks of Reliability
       The analogy argument seems to be operated with the assumption that some marks of reliability can be identified, e.g. universality, intersubjectivity, "testability", etc.  But actually how do we justify this assumption?  It is not possible to establish this by deduction.  It just seems obvious that, say, universality is no guarantee of truth & reliability.  It seems logically possible that a great number of men can be massively & systematically deceived by an evil genius (natural or supernatural).  Is it possible to establish this by induction then?  The prospect is also very dim.  One way to do that is by enumerative induction: to inspect many kinds of reliable experience & to see whether universality is almost always present as a mark.  If so presumably universality can be regarded as a mark of reliability & its presence allows us to argue for the reliability of a certain kind of experience.  However it is clear that such process is hardly practicable unless there is an a priori way to determine the reliability of a given kind of experience.  This assumption is unlikely to be granted by critics of REs.  Perhaps there are other principles of induction, e.g. principle of simplicity, inference to the best explanation, which can be appealed to here.  Of course the problem here is the justification of these principles themselves.  We may argue that these principles are just intuitively correct.  Alternatively we may justify these principles with our best judgments of what are the good inductive arguments.  The former a priori road is not likely to be adopted by the critics of RE since most of them are sort of empiricists.  The latter road obviously presupposes a lot of empirical knowledge.  The whole thing again turns into a circle then.
        How can any mark of reliable experience be established by experiences then?  It seems to be the case that many critics are prepared to assert the reliability of SEs & then use this to support the claim that universality is a mark of reliability.  But then when it is asked that why we regard SEs as reliable, universality is often cited as a reason.  This is painfully circular but sometimes circular arguments work if supported by unshakable confidence.  Then the circularity is declared to be "virtuous & not vicious".  The grounds for this is just opaque to me. 

        Here I am not saying this confidence should be undermined.  Far from it.  The credulity approach exactly acknowledges the necessity of this trust & argues that this element of trust should not be masked by spurious arguments & rhetoric.  Then the natural question is "If this trust is so fundamental, i.e. this is prior to identification of any mark of reliability, why the opposite attitude should be adopted in other cases?"  This is the deeper question which the defender of PC wants to ask.  Hence the PC should be distinguished from the analogical approach to experiences.  It may be called the transcendental approach.  Let's explore the difference further

Evaluation of Gale's Criticisms
        Gale is further confounded by his failure to distinguish the two questions of veridicality:
1) The radical question: what is the case for the cognitivity of a kind of experience as a whole, i.e., that it yields knowledge of reality roughly as that kind of experience indicates?
2) The less radical question: assuming that a kind of experience is in general cognitive, what are the criteria to distinguish the token experiences which are veridical from those which are not?

        Let us consider how are we to answer the first question.  If there is another kind of experience whose reliability or cognitivity we can assume, perhaps the case can be made by using the latter to vindicate the former.  Suppose there is a kind of telepathic experience.  To ascertain whether it is reliable, we can use SE to check its outputs if we can assume that SE is reliable.  However, if a general question is raised as to why we think any kind of experience, including SE, is reliable, the above procedure is of no help.  In general, it is easy to see that empirical considerations cannot be adduced for the case since it clearly begs the question against the sceptic.  Especially in this context, it can be seen that it is no answer to question 1 by pointing out that we have tests to distinguish the veridical from the illusory experiences such as Gale's tests.  Several tests are just negative tests, e.g. consistency test.  The others all seem to presuppose the reliability of some SEs.  How else can we determine what are the normal conditions for reliable SEs & how else can we ascertain whether they obtain in certain situations?  Even if we adopt the organic view of the tests espoused by Gale, the problem is far from being solved.  That view points out that the tests are as we should expect if the ontology of SE is correct.  It is not clear this fact in what sense helps to discredit the sceptical alternatives.  Presumably the evil scientist who manipulates our brains do not just generate the perceptions at random.  Perhaps all perceptions are generated as if  the ontology of SE were true & hence it is also to be expected that the apparent perceptions of the brains-in-vats conform to Gale's test.
        So it seems that Gale is not aware that SB & Alston are pressing question 1 rather than question 2.  Both in the end admit that no empirical argument can be given as a convincing answer.  They posit the PC or the presumptive inference rule as an a priori or transcendental principle.  The adoption of this rule is not dependent on the sort of tests available in SE or the sensory practice (Alston).  So their arguments are not  arguments from analogy.  Gale misunderstands completely their positions.  Perhaps they also appeal to analogical considerations.  These, however, are not the foundation of their arguments.  The analogy is only used to reinforce the conclusion & to forestall some objections.  It is not incompatible with the transcendental argument because the analogical argument only needs to assert that the analogy with SE is a sufficient condition for cognitivity.  It doesn't need to say that the analogy with SE is a necessary condition for cognitivity.  Certainly Gale holds to the latter: for him, REs "are cognitive only if they are subject to similar tests to those that SEs are."(p.316).  Actually he uses this to argue for the non-cognitivity of RE:
A1)      A kind of experience is cognitive only if it is analogous with SE, especially in the matter of having analogous tests.
A2)      RE is not analogous with SE.
A3)      Hence RE is not cognitive.
        It is a pity that he never gives any reason for (A1).  He does not even make it explicit & unquestioningly relies on it as an undisputed premise.  Perhaps he thinks that the analogical arguers should also commit to (A2) & therefore treats as a shared premise ./. Gale & his opponents. This is a mistake.  They need subscribe to this only:
A4)      If a kind of experience is analogous with SE, then it is cognitive.
It would be a poor mistake to confuse (A1) with (A4).  I think that Alston & SB would both reject Gale's assumption: "If REs are cognitive, they would be subject to similar tests to those SEs are."  They rightly point out that if any RE is veridical & there is a being like God, the above assumption is probably false.  So to assert (A1) without argument is to beg the question against SB & the like.  I wonder what reasons are there for believing in this premise?  Despite disclaimers to be a chauvinist, he turns out to be one after all!
        Similar confusions lie behind his strictures of Gutting's & Alston's criteria.   First of all, I think from the context it is clear that Gutting & Alston are confining their discussions to intentional or outer-directed experiences.  The examples of pain or orgasm are totally irrelevant.  Gale is aware of this reply but his reasons for rejecting this reply is only this: "people, especially philosophers, often disagree about whether a certain type of experience is putatively perceptual or such that its apparent object is objective if real at all."(p.313)  Is Gale really serious that there are genuine significant disagreements over the question: "If pain or orgasm is real, are they then objective?"  I think sufficient attention to the phenomenology would generate an unambiguous 'no'.  I also deny that philosophers' disagreement should take primacy over the ordinary people's awareness of the nature of their experiences.  Secondly, these criteria are supposed to be internal criteria, i.e. they are answers to the question: "assuming a kind of experience or epistemic practice is generally reliable, how to distinguish the veridical ones from the unveridical?"  They are not supposed to be answers to the more radical question.   In fact Alston explicitly says that the reliability of any LG can't be established in the strong sense!   Hence Alston can't be claiming that if a LG satisfies his criteria, then it entails the reliability of that LG.  Similarly, it is not clear that experiences which satisfy tests 1-11 would thereby be proved to be veridical.  If it is true, Gale's stricture of Gutting or Alston's criteria of experience is not fair.  He interprets them as logically tight conditions for cognitivity but it is hardly what Gutting or Alston would mean by them.  Hence his 'counterexamples' of pain & orgasm are really ridiculous! 
        The merit of Gale's discussion is that he does press his fingers on some disanalogies.  Perhaps it is the grounds for the strong 'intuition' among the critics that SE is relaible while the others are not.  However, if we disentangle the above two questions, it seems that such disanalogies are not relevant to question 1 unless one can find a test which is non-circular & assumption-free.  The disanalogies are indeed relevant to question 2.  The availability of such tests show that if we assume the general reliability of SE, we are then able to distinguish the veridical SEs from the illusory one by relatively clear criteria.  It is not the same in the case of RE.  It does make us more sceptical of the individual REs.  This does not imply we should then think that RE as a whole is unreliable or completely illusory!  (A kind of experience can be highly unreliable without being completely illusory.  It is all too easy to confuse the two.  The latter claim is astounding: that all through the ages all religious people on the earth are utterly deceived or deluded.  If only one of them is not deluded, RE is not completely illusory even though it were highly unreliable.)  Reliability is a matter of degree & all of Gale's discussions is to show that SE is more reliable than RE.  SE also exhibits a much higher degree of intra-coherence & worldview coherence than RE.  But this is a far cry from saying that REs are then unreliable simpliciter.  Furthermore, part of the problem is generated by taking all REs into a group.  The problem would be alleviated if we confine to some sub-type of RE, e.g. TE.
        Gale's discussion of the 'crucial' disanalogy again displays his epistemic chauvinism.  Suppose he is correct about the necessary conditions for a veridical sense perception.  It is only one concept of an objective empirical particular.  To argue that it is the only viable concept would have presupposed that physical object is the only possible kind of objective existent & that SE is the only viable kind of experience.  It is just arguing in a circle.  Gale also claims that his framework can explain why objects can exist unperceived & the like.  I wonder what sort of explanation has been provided here.  What is space-time & how it can house the object?  Myriad more questions can be asked & I wonder whether the cpt of 'objective existence' is more fundamental to 'spatio-temporal existence'.
      Gale also thinks that his argument would commit him to reject intellectual & moral intuition & the like as non-cognitive.  I wonder whether this shouldn't be taken as a reason to reject his argument.  It is surprising that he also appeals to necessary truth & conceptual truth: "no two empirical individuals of the same kind spatiotemporally coincide." (p.328).  I would like to see how he is going to justify the above claim apart from any appeal to intuitions.  He has not even hinted at the answer.
        I also thinks he exaggerates the disanalogy & underrate the analogy between SE & RE.  The detailed arguments for this claim has been given in chapter 10.  Only some brief points here.  Concerning test 5, Gale asserts that "The agreement or disagreement of other observers is relevant only if their observations are made under so-called normal or standard conditions."  This is incorrect.  Agreement can't be wholly dependent on concept of 'normal observer' because I can't see how we can ascertain the latter apart from some agreement.   Besides we also cherish the agreement in logical intuitions & rational judgment.  The notion of 'normal observer' is also not applicable in these cases.  So we can't dismiss the reports of similar experiences as a sort of important agreement.  On tests 8-10: Gale thinks that if it is still argued that we have some idea of the psychological & moral conditions that facilitate having a RE, then he would reply that "Such causal knowledge as we have pertains to the occurrence of a psychological occurrence or state- the having of an of-God-type experience- & not to its veridicality."(p.308)  This claim is deeply misleading as if we can have direct test of the presence of a physical object apart from any experience.  If it were the case, the sceptical questions can be easily dismissed.  It is also surprising to find Gale claiming that "whereas the worldview into which empirical particulars enter offers an explanation for the organic unity of the sensory tests, there is no analogous view into which the apparent object of REs enters that explains how the various RE analogues to these tests ... are interconnected. ... They form a mere heap, with no rhyme or reason why they should go together & serve to mutually reinforce each other."(pp.342-3)  He himself admits of the plausibility of Alston's explanation of the disanalogies in terms of the categoreal nature of God.  Isn't it a sort of coherence & explanation?  I have also argued in chapters 8 & 10 that many coherence tests are available for TEs which are as we should expect from the nature of the apparent object of this kind of experience.  Isn't it also a coherent relationship ./. the tests & the ontology in cases of TE?

        In general Gale suffers from a preoccupation with SE & he ignores the possibility of a plurality of basic cognitive[11] experiences.  So Gale's criteria of a cognitive experience seem too strong.  Consider again memory, introspection & rational intuition.  In a recent book, Audi argues that they are all basic sources of justification & in this sense cognitive experiences.  However, probably Gale's criteria will rule them all out!  Firstly, they are not of any objective empirical particular in space-time.  Secondly, how can we show, say, by memory alone that the memory experience is caused by the past event?  Even if Gale is prepared to uphold his criteria & to reject all these experiences, there is no reason why we should follow him.  I conclude that Gale's case vs REs has not been made.

Coherence of the Tests of Theistic Experience
        We need to analyze the concept of 'reality'.  An experience is veridical if it is in contact with reality.  But what are the implications of this?  It depends on the kind of reality we are supposed to be in contact.  This is the basic principle.  The epistemic implications of a cognitive experience is in accordance with the categoreal nature of the object confronted in experience.  No a priori restriction is possible.  Of course, if we conceive of reality as very transient & isolated from other kinds of reality, though I would not say it is incoherent, I do think this type of experience is epistemically inferior.  cf. AE Taylor's elucidation of reality in terms of experience.  But there are several dimensions of a 'knowable' reality:
1) It persists in existence => chance of further experiences.
2) It exists independently => sense of reality.
3) It has properties which are not exhausted by a finite stream of experiences => it sustains a process of discovery & reciprocity (questioning & modification).
4) It has causal power => it has effects on our lives. (Not applicable to abstract objects even if they are real: but they can affect us through our grasp of them.  cf. Popper's world 3)
5) It has a coherent description: reality cannot be self-contradictory => experiences of the same object have to be coherent as well.
6) It is part of the Reality & Reality can't be self-contradictory (pace Stace) => so experiences of one reality should cohere with experience of another reality or well-established inferential knowledge of other realities.  (This comports with theistic assumptions.  See Holmes, Contours for a Worldview, p.51f., All Truth is God's Truth, Ch.7.  It is questionable whether the nontheist can justify the adoption of coherentist criteria: why the world must be coherent/ intelligible?  Search for coherence is clearly a human imperative.)  Also see "An Epistemological Argument" in Craig & McLeod. 
7) It has some determinate characteristics => more determinate effects.  E.g. physical object => sensations.  Persons => union of will with will is the highest confirmation of the experience => also highest & fullest kind of knowledge. 
        Personal encounter with God => aiming at union of human will with Divine will (Farrer) => moral progress, self-discovery.  (More detailed models needed.  Circularity?  Yes, but unavoidable.)  God as Creator: providential experience, miracles.  God as ground of community => God among us.  etc.  So the tests of TE are clearly relevant & well-motivated.  See SB's account in Revelation.

Family Resemblance Argument
        Now I want to elaborate a point I have made above.  This is the point that the usual procedure of weighing RE against SE is very dubious because the implicit assumption is that SE is the only yardstick.  Now I have argued in earlier chapters that SE can not be the only basic source of justification.  This claim would only land us in scepticism.  To rise above the sceptical bog, memory must be regarded as another basic source.  Furthermore, I think that introspection & rational intuition are also indispensable basic sources. Now if this is granted, the whole picture of drawing analogy between different types of experience would be radically different from that of the narrow empiricist, whether the PCT is granted or not.  Suppose we can identify the prominent epistemological characteristics of SE as a type as c1, c2, ... cn.  If SE is regarded as the sole yardstick, then insofar as RE lacks some of them, we are tempted to say that RE is discredited to some extent.  This temptation, as I have argued earlier, is by no means irresistible if we can provide a reasonable explanation of this lack.  However, we may still remain a bit dissatisfied.  The situation would be changed if we keep in mind that our basic sources are more numerous & they are also very different from one another.  I have argued that the best way of justifying RE  should not be by way of analogy at all: it should be by the appeal to the PCT as a fundamental principle.  However for the sake of argument let us suppose that this way is out & ultimately we have to rely on analogy.  How should we proceed then?  We should first of all decide what are the basic sources of justification & I suggest that we should include at least SE, memory, introspection & reason among them.  Then we should identify, with regard to each, the set of epistemological characteristics which characterizes the source.  Now suppose a new source of beliefs is alleged to be another basic source of justification & we want to evaluate this claim.  If we proceed to do this by way of analogy, we should compare this source with all the four basic sources.  The analysis is admittedly more complicated but in this way the more simplistic errors can be avoided.  Suppose there is a ci which characterizes SE but not RE.  This might look to be disconfirming RE earlier on but if ci is also not exemplified by the other basic source(s), then this conclusion shouldn't be drawn.  In this way, many of the alleged crucial disanalogies can be rendered harmless, e.g. unavailability of checking procedure, poverty of details, lack of mutual confirmation of several sensory modalities, & impossibility of empirical study of the faculty of perception.
        The other point is that we can identify the really crucial mark of a source of justification with more confidence.  If we want to find out the genuine characteristics of a human being, this is scarcely possible if the only samples available are adult males.  If we select a sample of different sexes, ages & races, the results would be less prone to be errors.  Now this is also the case with the process of deciding what are the marks of a basic source of justification: we should decide what are the characteristics which are shared by all the basic sources we can start with.  The answers seem to be that they are all cases of epistemic seeming & that the source of beliefs possesses some sort of intra-coherence & inter-coherence.  In this way these marks can be seen to be much more important than before.
        In case that some characteristics are present in most but not all the basic sources we start with, then we can still use the notion of family resemblance to evaluate an alleged new source of basic justification.  We treat the original sources as members of a family.  Perhaps there is no one characteristic which is shared by all of them.  Nonetheless they may show family resemblance with one another.  (cf. Wittgenstein)  So if a new source is evaluated, it should be checked with respect to its family resemblance with the whole family.  Hence a multi-dimensional testing is needed.  To carry this out, we need more detailed comparisons of RE with SE, memory, introspection, & reason.  If the result is favorable to RE, then the above argument can be called a family resemblance argument for RE.  I think it is much better than the AA & its chance of being successful is higher.

        Gale’s criticisms of the argument from RE are not convincing. For other criticisms of Gale, see Plantinga’s Warranted Christian Belief (Plantinga 2000, 335-342).

[1] In fact in his book Perceiving God Alston explicitly says, in response to Gale, that his argument is not an analogical argument.  His argument is instead based on his epistemology of doxastic practice which is structurally similar to the Transcendental Argument.
[2] This is effectively the claim that the narrower PCT is a fundamental principle.  The first part of this thesis is an extended argument for this claim.
[3] These could be shown directly: arguments for the implausibility of (i) or the gap between (i) & (ii); or indirectly: arguments for unacceptability of either (ii) or (iii) & hence either (i) is unacceptable or (ii) can't be inferred from (i).
[4] Another possibility is to have a list of prima facie reliable experiences & then search for the common characteristics of these experiences which are plausible marks of reliability.  Problems arise of course from the very beginning: whether RE should be included in that list.  Let us grant that RE should not be included in the initial list.  An acceptable initial list for the "empiricist", it seems to me, should include at least sense experience, moral experience, introspection, logical intuition, and induction.  The reason is that these capacities are essential, either epistemologically or pragmatically, for the scientific enterprise which is held in such esteem by the "empiricist".  If we look at the list, the variety should strike us & it is not obvious that RE would then fail this inductive test.  Indeed Plantinga seems to espouse the inductive route & he arrives at a proper functionalism which is not damaging to RE.
[5] For the time being, we ignore the issue that analogy is not an all-or-nothing matter.  X can be analogous with Y to a greater or lesser extent & in this or that aspect.  This simplification will not affect the argument at this stage but the complications will be dealt with later.
[6] This is not entirely agreed upon by philosophers.  There are some who think that necessary truths can give us new information, e.g. Brand Blanshard.
[7] Here I do not need to assume the reliability of induction.  (I'll leave the question what kind of reliability it is.  This is not essential for the above argument.)  The point is that even if induction were unreliable, this can't be "inferred" from its disanalogy with deduction. 
[8] I have already argued for this claim in some details in chapter 6.
[9] It may be surprising to some.  But introspective reports like "I have seen a flick of the pointer" are essential for intersubjective confirmation of experimental results.  Moreover in more theoretical science we probably need to rely on such statement: "I have proved carefully that this physical theory contains an inconsistency."  This of course relies on the personal memory of working out the steps of the proof carefully.  This is essential, at least in the beginning, for the justified confidence in the impersonal memory that "this inconsistency has been proved".
[10] Even if we had other means to establish (A), this fact does not clearly endanger the AA.
[11] For Gale, a cognitive experience seems to be equivalent to a knowledge-yielding experience.(p.324)  I don't want to go into the controversies over the nature of knowledge & my central claim is only that religious experience, especially theistic experience, can be a basic source of justification.  The claims in the text concerning knowledge can be read in a weaker sense as concerning 'justified belief'.