John Mackie's Naturalistic Explanation Objection to the Argument from Religious Experience

Kai-man Kwan

        John Mackie thinks that the availability of naturalistic explanations defeats the argument from religious experience.  He draws heavily on William James who seems to admit that "So far ... as this analysis goes, the experiences are only psychological phenomena" (quoted by Mackie on p.183).  James did want to assert that there is the Something More & the Beyond.  Mackie comments, "The issue is whether the hypothesis that there objectively is a something more gives a better explanation of the whole range of phenomena than can be given without it.  James himself thinks that it does; yet he gives no real argument to support this opinion.  This is, obviously, a less economical hypothesis than its naturalistic rival" (p.183).  This type of naturalistic explanation objection regards religious experiences as ordinary events to be explained & does not entertain the possibility that they may have PFEF at all.  The further assumptions are that the decisive criterion here is the Principle of Simplicity & that NEs are simpler than the theistic explanation.

Mackie's argument can be formulated as follows:
R1       Ceteris paribus, among competing explanations, adopt the explanation of an event which postulates fewer or fewer kinds of entities.
P1        Explanation of a theistic experience does not invoke any special considerations which violate the 'ceteris paribus' clause in R1.
C1       Hence, among the different competing explanations of a theistic experience which are equally good & adequate, adopt the explanation which postulates fewer or fewer kinds of entities.
P2        For each theistic experience, there are competing theistic explanation & naturalistic explanation which are equally good & adequate.
P3        A naturalistic explanation always postulates fewer or fewer kinds of entities than a theistic explanation of a theistic experience.
C2       Hence, adopt the naturalistic explanation for each theistic experience.

Some comments are in order.  R1 is a methodological rule which underlies Mackie's  contention that theism is a "less economical hypothesis than its naturalistic rival".  It is important to spell it out because simplicity can be measured in different ways and not every way is as epistemically significant as another.  I indeed agree with Mackie that R1 is acceptable.  Further note that R1 only applies to competing explanations.  If two explanations can't be simultaneously true, then surely they are competing.  If two explanations are each specifying only necessary conditions, then they are not competing & there is nothing to choose.  So for two explanations to be competing, they at least each specify causally sufficient conditions for the experience.  P1, P2, & P3 are the three premises that Mackie needs for his argument to go through.  No doubt he thinks P3 is true & essential to his argument.  P1 is an unstated assumption.  It is needed because R1, in any plausible version, is only a ceteris paribus rule.  It may be the case that there is something special about theistic experiences or noetic experiences in general which override the application of R1 to them.  Mackie does not even consider this possibility & simply assumes P1 to be true.  This is essential to his argument.  C1 is a corollary from R1 & P1.  P2 is the second premise which Mackie needs to remove all the evidential force of theistic experiences.   A theist can grant that many theistic experiences are open to NEs.  But even on Mackie's principles, this only removes the weight of those  & only those theistic experiences which are thus explainable.  As long as one theistic experience is veridical, it entails the existence of God.  So similarly if only for one theistic experience the theistic explanation is superior to the NE, there is some evidence for theism.[1]  Since Mackie does not admit that theistic experiences can provide any evidence for theism, he must hold to P2.  C2, which declares the victory of the NE enterprise, follows from C1, P2 & P3.  Since I accept R1, I have to challenge either one of P1, P2 & P3.  In fact, if my defense of the PCT is successful, then P1 should be rejected.  However, for the sake of the argument, I would not assume the PCT in the following discussions & suggest independent considerations which challenge all three premises. 

Is Naturalistic Explanation Really Simpler?
        I have suggested that R1 only captures one notion of simplicity.  In this section I would compare theistic and naturalistic accounts of theistic experiences with respect to different types of simplicity, and then explore the epistemological implications.  Firstly, let me point to some counterintuitive implications of R1: it would also make phenomenalism (indeed, the solipsistic type) intrinsically preferable to realism about external world, and instrumentalism preferable to scientific realism.   In general, all non-epistemic explanations of noetic experience would be simpler according to R1, and hence preferable.  These are not results that my critics would be prone to accept, e.g. Mackie.   It is because it would put the burden of proof on the realists about the external world and about theoretical entities.  If the battle is not then irretrievably lost, it would be an immensely harder one for the realists to fight.  Of course, if we are sure that R1 is the only rule relevant here, then the hard battle has to be fought.  However, I suggest that many would feel that such a distribution of burden of proof is not a fair one.  If they are not to reject R1, then they can suggest that there are other rules which override R1 in those cases.  This is the line I am adopting here.  However, before going on, let me define OI of a worldview or ontology as the ontological inventory of the entities considered real in that worldview.  For example, when I write OI(phenomenalism) = {red qualia, green qualia, ... , kinaesthetic sensations, auditory sensations, ...}, it means that the ontological inventory of phenomenalism is the set of qualia, sensations, etc.  So according to R1, one worldview is simpler than another if the inventory of the former is a proper subset of the latter.  For example, OI(realism about external world) = {red qualia, green qualia, ... , kinaesthetic sensations, auditory sensations, ..., tables, rocks, cats, ...} = OI(phenomenalism) U {tables, rocks, cats, ...}.  Similarly, OI(scientific realism) =OI(instrumentalism) U {protons, electrons, neutrons, etc.}.  (Remember that what is included in any OI is considered real rather than fictional.)
        Now let us consider the debate between the phenomenalists and realists.  Regardless of the detailed merits or demerits of each position, I have suggested that R1 should not be the basic rule here since it imposes an unfair burden of proof on the realists.  I think although realists generally grant R1's applicability to events, they can point out that noetic experiences are a special kind of events which seem to provide direct access to reality.  So this is the contention:
R2       R1 is not applicable to noetic experiences.

This would mean at least that in the case of noetic experiences, the burden of proof is not on the side of realists.  However, in this thesis, I have contended that we should take the force of our noetic experiences even more seriously, i.e., the burden of proof is on the side of sceptics about noetic experiences.  If this is true, then a third rule is true:
R3       Ceteris paribus, adopt an explanation of noetic experiences which preserve their PFEF.

Note that R3 and R1 are not contradicting one another.  Since R1 is only a ceteris paribus rule, we can view R3 as a rule which refines R1 and singles out the special considerations which go with noetic experiences, a special kind of events.  I have already argued that R3 is indeed justified.  Anyway, it seems to be the rule which is accepted by many philosophers when they arbitrate the debate between phenomenalists and realists.  Isn't it very arbitrary not to use it when we come to evaluate another kind of noetic experience, theistic experience?  At the very least, I contend that R2 is right and hence the assumption that simplicity of NE in accordance with R1 is the most significant consideration in this issue is unwarranted.  In other words, P1 is false.

        Actually R3 does not seem to be the only constraint on the application of R1.  If this is the case, then the burden of proof still seems to be on the scientific realists because the existence of theoretical entities cannot be directly supported by noetic experiences.  So if the intuition that it is unfair to put the burden on the scientific realists is correct, then there should be another constraint on R1.  Let us define the fundamental inventory, FI, of a worldview as the set of the entity types considered as fundamental and irreducible in that worldview, symbolised by FI(worldview).  Further define the function n[FI(worldview)] as the number of entries in that FI.  We can then construct another notion of simplicity:
Worldview1 is simpler than worldview2 iff n[FI(worldview1)] is smaller than n[FI(worldview2)].
To distinguish the earlier notion of simplicity from this one, they are called 'OI-simplicity' and 'FI-simplicity' respectively.  My suggestion is that considerations of FI-simplicity are at least as important, if not more important, than considerations of OI-simplicity especially when we evaluate worldviews.  This would explain why instrumentalism is not intrinsically preferable to scientific realism.  Consider their fundamental inventories:
FI(instrumentalism) = {tables, rocks, cats, dogs, moon, sun, stars, ...}.  Obviously, the list is almost endless.  However:
FI(scientific realism before 1900) = {hydrogen, helium, ... uranium}.  n[FI] here is only around 100 << n[FI(instrumentalism)].  This would explain why although postulation of a hundred types of atoms actually increases the OI-complexity, the scientific worldview is still considered simpler.  It is because the myriad of objects and substances seem to be reducible to combinations of these atoms, or, in other words, the scientific worldview is FI-simpler than the instrumentalist worldview.  At one stage, it is thought that FI(scientific worldview) = {protons, neutrons, electrons} with a n[FI]=3.  People started to talk about the end of physics: the world became for them immensely simpler- of course it is FI-simplicity that they are considering.[2]  So I suggest another rule is governing our explanation:
R4       Ceteris paribus, adopt an explanation offered by a worldview which is FI-simpler.

As the progress in science shows, when we consider the question of worldview and the fundamental entities, FI-simplicity seems more important than OI-simplicity.  But now the debate between naturalists and theists over theistic experiences is a question of worldview.  So FI-simplicity is more epistemically relevant here than OI-simplicity.  So let us now evaluate the debate from this perspective.[3]
        We need to distinguish between reductive naturalism and non-reductive naturalism.  The former position asserts that the fundamental entities in physics are the only irreducible entities in reality.  Sensations, minds, selves, etc. are either eliminated or reduced.  However, non-reductive naturalism does not think all of these are either eliminable or reducible: some of them are irreducible.  Consider their fundamental inventories:
FI(reductive naturalism) = {several kinds of quark, electrons, ...}
FI(non-reductive naturalism) = {several kinds of quark, electrons, ..., sensations, minds(?), ...}
FI(theism) = {God}
        It seems clear that theism is FI-simpler than reductive naturalism,[4] which is FI-simpler than non-reductive naturalism.  So if we use R4 instead of R1, as is more appropriate here, a theistic explanation is actually simpler than the NEs.  The tension between reductive naturalism and non-reductive type is also instructive.  True, the reductive variant is FI-simpler but the non-reductive naturalists argue that the forces of our experience cannot be denied.  So the reductive naturalist charges that non-reductive naturalism harbours 'mysterious' entities & urges it to return to the pristine simplicity of reductive naturalism.  On the other hand, the non-reductive naturalist accuses the reductive naturalists of reductionism, scientism & inadequacy to our experiences.  It seems to me both complaints have some weight.  However, theism seems to be the worldview which can accommodate nicely the awkward facts for a reductive naturalist while achieving unsurpassable FI-simplicity.  So this consideration in the particular case of theistic experiences further overrides application of R1.  This is the second reason why P1 has to be rejected.  Indeed if FI-simplicity is more significant than OI-simplicity, then we can formulate this rule:
R5       Ceteris paribus, adopt an explanation offered by a worldview which is FI-simpler even if the explanation is less OI-simple.

Because theism is FI-simpler than both types of naturalism, R5, which is argued on independent grounds, will entail that theistic explanation of theistic experiences is intrinsically preferable to NEs, ceteris paribus.  Anyway, we have sufficient reasons to throw doubt on P1.

        Now P2 (see p.26) is also plagued with difficulties.  First of all, we have to be sure that a NE is really in competition with the theistic explanation before we have anything to choose from.  If a NE just offers a partial causal account of a theistic experience & specifies some necessary conditions for that theistic experience, then the theist can happily accept both its truth & the truth of the theistic explanation.  So for a NE to compete with the theistic one, it has to specify causally sufficient conditions for the theistic experience & we have seen that it is not a easy thing to do.  We have also seen many drawbacks of the prominent NEs.  For all these reasons, the truth of P2 is very dubious.  Furthermore, we should note the actual complexities of the NEs so far offered.  Firstly, there are various sub-types of theistic experience: a plausible general NE of theistic experience has to explain all sub-types.  However, since the sub-types of theistic experience are phenomenologically different and occurring in different situations, presumably quite different psychological explanations or NEs are needed for the different sub-types of theistic experience.  Moreover, experiences of God are numerous and occur spontaneously in most varied situations to people of almost all ages, cultures, religious traditions, personalities, etc.  My earlier discussions suggest that a general NE which can cover all these theistic experiences is indeed difficult to come by.  So in the end, the naturalist needs a set of disjunctive NEs.  As this list grows in length, its simplicity decreases.  So it is doubtful that such a kind of NE postulates few or fewer kinds of entities.  In such a situation, it seems plausible to say that the simplest explanation is actually that all these theistic experiences have the same cause, God.  So even P3 is doubtful if we take the whole class of theistic experiences into account.  We have also noted that the various sub-types of theistic experience exhibit intra-coherence on the theistic explanation.  Especially when we consider the various types of theistic experience which occurred to a single person, telling a coherent, meaningful and beautiful story, isn't it simpler to explain the coherent story by the person's relationship with God?  Whether these happen is an empirical matter and there is plenty of anecdotal evidence for it.  One only needs to read biographies of religious person, e.g. that of Jackie Pullinger.  If all these experiences are delusory, these persons must be insane, which they appear not to be.
        If the above is not sufficient to establish the falsity of P2 or P3, we should also note that simplicity is not the only consideration.  Theism has another explanatory merit: the explanans of the NEs include psychophysical laws, the structure of the human psyche, social nature and experience of human beings, etc. but these in turn need to be explained.  I have already argued that the naturalistic explanation of some of these is itself very complicated and ad hoc.[5]  So the postulation of God at the same time satisfactorily explains the structure of human psyche and the numerous theistic experiences.  This is on the whole simpler than the naturalistic scheme.  My conclusion is that on various counts theism is simpler than the NEs, and when NE is simpler than theism in one respect, that kind of simplicity is overridden by other considerations.  Mackie's argument can be safely rejected in the case of theistic experiences.

Naturalistic Explanations of Non-Theistic Religious Experience
        It is important to distinguish different types of religious experience in response to the naturalistic explanation challenge.  Let us consider the experiences of lesser deities, visions of saints and so on.  It seems to me even if my criterion for non-epistemic explanation is correct, these religious experiences are still more vulnerable:
1) Firstly, if a causally sufficient NE is produced, it seems enough to be evidence-canceling because it is much more plausible to say these NEs would pre-empt the causal action of these deities.
2) Secondly, my arguments for the simplicity of theism do not apply to these experiences.  These deities are included only in the fundamental inventory of polytheism, which has only a low degree of FI-simplicity.
3) It is less plausible to appeal to these deities to explain the structure of human psyche, society, etc. because they are not the Creator.
       Of course, it doesn't imply that NEs are automatically successful unless one revert to the a priori approach again.  The conclusion is that successful NEs of such religious experiences are conceivable and hence the evidential force of these religious experiences more easily defeated than that of theistic experiences.  It further justifies the assimilation of these experiences into a theistic worldview instead of changing a theistic one into a polytheistic or henotheistic worldview.  In this aspect, NEs may well serve the purpose of theists!
       I think similar things can be said for nature mysticism or the PCE: they are more vulnerable to the attack of NEs.  The relation of monistic experience with the project of NE is hard to say.  If monistic experience really implies that the natural world is illusory, then it has already rejected the validity of any NE & it would be difficult for monism to accommodate the insights of many (partial) NEs of religious experience, including monistic experience.  But since the view of monists towards this natural world can be very obscure & 'paradoxical', I do not declare upon this matter with any confidence.  I must ask the monists themselves to clarify their stand on this issue.

        I have so far argued that if we assume that theistic experiences have PFEF, then the naturalistic explanation objection is virtually hopeless.  However, even if we reject this assumption, I have argued that Mackie's argument is still far from being successful.  There are other grounds for rejecting the naturalist's view of burden of proof; at least it is not incumbent on the theist to accept this view.   Hence even the naturalist, unless he has prior strong arguments for `naturalism, should be open to the possibility that a theistic experience may well have PFEF.  This means the critic can no longer complacently assume that a NE is automatically evidence-canceling.  To show that he has to justify the initial scepticism towards theistic experiences.  The burden is also on his side.  The a priori approach to NE should be abandoned and any NE should be subjected to empirical and philosophical scrutiny.  It is at least not clear that powerful NEs have been produced & so the naturalistic explanation objection is yet to be shown to defeat my argument for theistic experience.

[1] Of course the actual assessment is more complicated.  Assuming Mackie's principles, if the vast majority of TEs are readily explained by NEs, then the naturalist can mount a second order inductive argument for the explicability of those TEs which have not yet been explained.  How powerful is this argument is difficult to say.
[2] Of course, even FI-simplicity is not the only constraint on theory construction.  Physicists are later forced to postulate many more 'elementary' particles by experimental results, etc.  This shows that empirical adequacy is all along a very important constraint: we should seek the simplest theory adequate to the empirical data.  However, the concern for FI-simplicity is never lost sight of as is shown by the relief many scientists feel when the quark theory reduces the n[FI(scientific worldview)] again.
[3] The number of fundamental types of explanation also influences the estimate of simplicity.  For example, if we need several fundamental forces which operate according to different, unrelated equations, then we have several fundamental types of scientific explanation at least.  It seems to me we also favour a worldview which has fewer types of fundamental explanation.  This is the drive towards the search for a Grand Unified Theory.  We can also evaluate the relative simplicity of theism from this perspective.  If personal or intentional explanation is irreducible to scientific explanation, then theism is arguably simpler in this aspect.  It is because it only countenances one fundamental type of explanation: personal explanation.  See Swinburne 1979, chapter ?.
[4] Swinburne also argues that the description of God is very simple because God possesses attributes to infinite degree.  Others actually argue that God can be characterised by a single predicate like 'perfect' (Morris) or 'super-eminent' (Schlesinger).  Another aspect concerning the number of instantiations of a fundamental type.  In FI(theism), there is only one type & also only one individual.  In reductive naturalism, there are several fundamental types & each type has an enormous number of instantiations.  This would raise the question why so many individual things do belong to the same type, sharing identical properties.
[5] Nancy Clasby raises the same point: if "the structural elements of the psyche are found to be inherited, independent  of personal experience, and if the development of the psyche expressed in dreams, myths, symbols, occurs as a natural growth process, the relationship between God and the psyche becomes a much more complicated issue.  ... the striking parallels between man's inherited psychological structures and the homologous theological structures raises fundamental questions of causality" (In Robinson and Mitchell, p.86).  She asks "why these particular patterns emerged.  Where did this development code come from?  Man did not make up his own psyche.  As Jung said, it is ridiculous to think that "in prehistoric times the basic mythological ideas were 'invented' by a clever old philosopher or prophet, and ever afterward 'believed' by a credulous and uncritical people" " (p.88).  Furthermore, "If the psyche were structured in some straightforward way which simply extended the animal instinct to flee from danger, to reproduce when possible, to acquire food, the origin of the psyche in survival needs would be evident.  But the psyche is an endlessly convoluted form, full of phenomena which have little obvious relation to survival.  The positive function of religion in primitive societies is often pointed to as the reason for its development.  Religious beliefs, it is said, comforted man, stabilized society and offered primitive explanations for mysterious phenomena.  Since the contrary may also be asserted with equal force (religion frightened man, caused stagnation in society and spread superstitious misinformation), the argument for the utility of religion fails to explain the extraordinary energy which has poured into the evolution of the human psyche and its various theophanies.  Far from having developed along clean, functional lines, the psyche seems a singularly extravagant anomaly" (p.88).  That there is an academic discipline as depth psychology is a surprising thing in a naturalistic world.  Is then a depth psychological NE truly naturalistic?