International Journal for Philosophy of Religion 49 (2001), 111–120

Copyright © 2001 Kluwer Academic Publishers




Śaṃkara’s Principle and
Two Ontomystical Arguments



Alexander R. Pruss


Department of Philosophy

University of Pittsburgh

Pittsburgh, PA 15260





Śaṃkara’s Principle and
Two Ontomystical Arguments



July 24, 2000

I. Introduction

As part of his critique of Buddhist illusionism that claimed that external reality was impossible as a reality but was a mere seeming, the Indian philosopher Śaṃkara (788–820 A.D.) argued:

If a thing outside awareness is as impossible as a barren woman’s son how can we even feel as if something is outside?  Nothing even appears to be like an impossibility.[1]

Śaṃkara himself apparently used his principle that impossible things do not even appear to argue the hyperidealistic claim that it does not even appear to us that there is an external world.  But one can more plausibly use it to argue against the idealist who claims that an external world is impossible.  Evidently, there appears to be an external world, and hence by Śaṃkara’s principle and modus tollens it is possible that there be an external world.  The purpose of this paper is to examine how this anti-sceptical use of Śaṃkara’s principle can be used as part of a new argument for the existence of God.

Two types of arguments for the existence of God that have received much attention are the ontological arguments and the arguments from religious experience.  Both types of arguments have their peculiar weaknesses: the ontological arguments require a possibility premiss, while the argument from religious experience requires that the veridicality of the experience be proved.  Using Śaṃkara’s principle I will show that the two types of arguments can be combined in such a way that each compensates for the weakness of the other, and in combination produces a new argument for the existence of God.  The particular kind of argument from religious experience that will concern me here will be the argument from high mystical experiences, of which the experiences of St. John of the Cross are a paradigm.

The ontological arguments I shall consider will be Alvin Plantinga’s modal maximally-great-being argument and an apparently new argument from radical dependence.  Plantinga’s argument[2] is:

(1)   Necessarily, a maximally great being exists in all possible worlds and is perfectly good, omnipotent and omniscient in them all.  (Premiss.)

(2)   It is possible that there exists a maximally great being.  (Premiss.)

(3)   Therefore, there exists a maximally great being.  (From (1) and (2) by S5).

The argument from radical dependence is as follows:

(4)   A being, x, is radically dependent on another being, y, if and only if it is an essential property of x that y’s activity enters into a causal explanation of x’s existence.  (Definition.)

(5)   There is an actually existent person such that it is possible that this person is radically dependent on an essentially numinous and loving being.  (Premiss.)

(6)   Therefore, there actually exists an essentially numinous and loving being on whom at least one actually existent person is radically dependent.

The step from (5) to (6) here goes as follows.  I shall use proper names as rigid designators.  By existential instantiation of (5), suppose that Smith exists and it is possible that Smith is radically dependent on an essentially numinous being.  Then, there is a possible world w at which it is true that Smith is radically dependent on some essentially numinous and loving being, say Y*WH.  But then it is an essential property of Smith in w that Y*WH’s activity enters into a causal explanation of Smith’s existence and an essential property in w of Y*WH to be numinous.  Therefore, in all worlds in which Smith exists, Y*WH’s activity enters into a causal explanation of Smith’s existence and Y*WH is numinous.[3]  Consequently, (6) holds.

As is well known, the difficulty in ontological arguments is with justifying their possibility premisses, in this case (2) and (5).  But there is reason to believe, or at least a presumption, that some of the high mystics had an experience as of a maximally great being and as of being radically dependent on an essentially numinous and loving being.  If Śaṃkara’s principle were true, it would follow that it is possible there is a maximally great being and that these mystics were radically dependent on an essentially numinous being, and it would follow that (2) and (5) hold, thereby supplying the most controversial premisses of the two ontological arguments.  The rest of this paper sketches how this argument that there is at least a presumption in favor of the existence of God is to be filled out

II. Śaṃkara’s what seems, could be principle

There looks to be in front of me a big shimmering sphere out of which climb various green-skinned beings.  I do not know whether this perception is veridical or not.  Maybe I am hallucinating or dreaming, or maybe some visual illusion afflicts me.  If I knew that my seeing was veridical, I could conclude that there was a big shimmering sphere out of which various green-skinned beings climb.  However, I cannot responsibly conclude this without examining the question of veridicality.  But what I am justified in concluding is that the existence of a big shimmering sphere out of which various green skinned creatures climb is possible—such a thing could be.  On Kantian grounds, it is a condition of the unity of experience that the experience be logically consistent and hence possible.  While it is admitted by most (leaving aside Protagorean relativists) that what seems might not be, nonetheless what seems is possible.  This is Śaṃkara’s principle in a positive formulation.

The principle would indeed be a useful one.  Parmenides and Zeno have argued that there cannot be movement or a multiplicity of things.  Berkeley has argued that there cannot be material objects outside of our minds.  Yet, neither Parmenides (who discusses the deceitful Way of Seeming in his poem) nor Berkeley (who is willing to talk with the vulgar) can deny that there seems to be movement and a multiplicity of physical objects around us.  While Moore might want to refute Parmenides, Zeno and Berkeley by saying that I have two physical hands that move, all we need is the weaker and quite undeniable premiss that there seem to be two physical hands that move.  If Śaṃkara’s principle were to hold, then it would follow that it is possible for there to be two physical hands that move, and in particular it would be possible for there to be a multiplicity of changing physical entities, which would refute Parmenides, Zeno and Berkeley[4].  It would follow that it is not possible that Parmenides’, Zeno’s and Berkley’s arguments for the impossibility of the existence of those external objects that these philosophers deny be sound.

In the same way all dogmatic idealisms that claim the impossibility of various features of our experience are refuted by Śaṃkara’s principle.  Furthermore, once one sees that Parmenides, Zeno and Berkeley cannot have proved that certain seeming realities are impossible, then their idealisms, even when not dogmatic but merely problematic (to use Kant’s phraseology), are no longer very attractive.  If there could be such a thing as movement, then given that there seems to be such a thing (and there is agreement among people about it, etc.), it is plausible to suppose that there is movement.  (To allay readers’ suspicions it must be emphasized that the argument for the actual existence of God to which all this is leading up is not like this argument for the actual existence of movement; there are serious disanalogies here, e.g., because not everyone has mystical experiences while everyone perceives movement.)  Nobody really believes they are being globally deceived by a Cartesian demon or that they are a brain in a vat, and so once the dogmatic idealism has been refuted, the problematic kind is not very gripping on actual persons.

We see thus that Śaṃkara’s principle would be quite useful.  But is it true?  Firstly, I have already suggested a Kantian type argument for it.  The unity of experience requires that it be the case that experiences be coherent and hence possible.  One cannot simultaneously and in the same way (i.e., in a single mental experiencing) behold two contradictory appearances. 

Secondly, imagine, if you can, that you have an experience as of a square circle.  What would it be about this experience that would make it an experience of a square circle, as opposed to, say, a square square, or a circular circle, or, for that matter, a horse?  (Why a horse?  Because that x is a square circle entails that x is a horse, since in every possible world in which there is a square circle—i.e., none—that square circle is a horse.)  One can use a Sellarsian method to account for what makes the experience as of an ordinary circle what it is, namely that it is the same phenomenal experience as one would be having were one veridically perceiving an ordinary circle.  But suppose one were to apply a similar account in the case of a logically impossible “object”, so that an experience is as of such an “object” providing that the experience is the kind that one would have were one veridically perceiving such an “object”.  Now, that one is veridically perceiving a logically impossible “object” entails that a logically impossible “object” exists which in turn entails all propositions.  Hence it is true that were one to be veridically perceiving an impossible “object”, one would be having a phenomenal experience of a horse, so that by the definition a phenomenal experience of a horse is also a phenomenal experience as of a square circle and as of any other impossibilia.  But this is absurd, and hence a definition of what it would mean for a phenomenal experience to be as of some impossible “object” along the lines of the definition for possible objects fails.

Thus, it is unintelligible what it would mean to say that someone is having an experience as of an impossible object.  Thirdly, an implicit reliance on Śaṃkara’s principle as a conceptual truth provides a good explanation for why human beings apt to take the imaginability of some state of affairs as evidence for the possibility of that state of affairs.  For imagination can be considered to be a mental simulation of an experience.  That we think we can imagine something gives us some reason to think that we could have an experience of that something, and if we could have an experience of it then it is possible by Śaṃkara’s principle considered as a conceptual truth along with S5.  This reason is quite fallible, (a) because our imagination may not do justice to our experience and may lack detail that would be needed were the imagination to become an experience but which detail might be impossible to supply (e.g., one might think one can imagine what it is like to be a bat, but not be able to supply in one’s imagination crucial elements of the experience that would be necessary to make it an as of being a being a bat experience), and (b) because of issues of misidentification and mislabeling that will be discussed in the context of experiences, below, which issues are exacerbated in the case of imagination given its vagueness and unclarity.  An implicit reliance on Śaṃkara’s principle thus provides an explanation for some of our ways of judging of possibility, and this implicit use of the principle provides a presumption for its truth.

These considerations, along with the not-to-be-ignored intuitive plausibility of Śaṃkara’s principle, give one good reason to believe that the principle is true, provided that one can formulate it in a way that avoids various apparent counterexamples.  For instance, it once happened to me in a dream that it seemed to me that I had a proof of a mathematical proposition p.  By the Śaṃkara’s principle, it seems to follow that it is possible that the mathematical proposition has a proof and hence that the proposition is possible.  But if a mathematical proposition is possible, it is true.  Now, after I awoke, I realized that what had in the dream appeared to be a proof was essentially nonsense.  By Śaṃkara’s principle it would seem to nonetheless follow that p is true.  But surely it would be possible to have such a dream in a case where the mathematical proposition is false (in the case that actually happened to me, I still do not know if the proposition is true).

However, this is not really a counterexample to the principle.  For in the dream it was not the case that I was phenomenally presented with a seeming proof of p.  Rather, I was presented with a sequence that I misidentified as “a proof of p”.  Were it the case that I was really presented with the mental representation—howsoever unveridical—of a proof of p, then the truth of p would follow.[5]  It is indisputable that we can mislabel and misidentify what we are presented with.  It is important for the defense of the Śaṃkara principle that the word “seem” or “appear” have a very broadly perceptual connotation.  There are other counterexamples that can be answered in a similar way.  For instance, suppose on a blackboard there is a formula of logic, and “it seems to me that” I am seeing a theorem of logic.  One cannot conclude from this that it is possible that the formula is a theorem, since then it would follow that it necessarily is a theorem by S5.  However, the “seems” in “it seems to me that” is not the “seems” of a very broadly perceptual presentation, of an experience, but rather a cognitive “seems”.

We thus have to characterize the sort of seeming to which the principle applies while avoiding the problem of possible misidentification of one’s experience.  Consider another more difficult case.  Suppose that it is an essential property of living elephants to be have heads, but that it appears to me that I am faced with a living headless elephant.  If Śaṃkara’s principle applies, it would follow that headless living elephants are possible, which is false.  To take care of this, we introduce the more technical locution “really seems”.  “An x really seems to s” is true if and only if s would be correctly identifying the content of a single phenomenal experience of hers if she were identifying it to be an x.  In the case of the apparent perception of a living headless elephant, I am presumably misidentifying the object of my experience as an elephant, since anything that is living and headless cannot be an elephant and anything that appears living and headless should not be identified as an elephant.  I am making a mistake about what it is that I seem to perceive.  I should instead say “An elephant-like living headless animal really seems to me”, and of course an elephant-like living headless animal is possible.

What more we say about “really seems” depends on our theory of perception.  If we are sense-data theorists, then we can clarify the notion of “really seems” further by saying that “an x really seems to s” means that s is experiencing sense data that would be correctly identifiable as x.  If, on the other hand, we are Sellarsians, then we cay say that “an x really seems to s” means that a veridical experience of an x is possible and s is having the kind of experience that she would be having were she to in fact be veridically perceiving an x.

Śaṃkara’s principle in its final form then says that whatever really seems to a subject, could be (note that this is analytic on the Sellarsian definition above).  The principle allows for the possibility that the subject misidentifies the content of his experience.  In the new form, the principle is difficult to deny.  A paradigmatic counterexample to it would be a case where a person experiences two logically contradictory properties, P and Q, as simultaneously present (in the same sense) in a single subject x, and correctly identifies the two properties.  But it certainly does not appear possible to imagine such an example.  Could one have an experience of a euclidean triangle whose angles do not add up to two right angles?  Surely the closest to such an experience would be one that is correctly identifiable as an experience of something that is not a euclidean triangle.

III. The existence of God

By the arguments in Section I together with Śaṃkara’s principle, to prove the existence of a perfectly good, omnipotent and omniscient being, it suffices to argue that a maximally great being really seemed to the high mystics.  Certainly, the maximalistic language of mysticism about the greatness of God gives prima facie support to the latter claim.  Alvin Plantinga has argued that the possibility premiss of the ontological argument creates an epistemic tie: it is just as likely that this premiss be true as that it be false.  The data from high mysticism breaks this tie.  For there is a presumption that the phenomenal content of the mystics’ experience is as the mystics have described it, and by Śaṃkara’s principle there is thus a presumption in favor of the possibility premiss.

One could try to escape the argument by claiming that the high mystics lied about the phenomenal content of their experience.  But it is on the whole improbable that they all did given the moral character of many of them.  A much more serious objection is that perhaps the maximalistic language of the mystics is just a matter of their interpretation of their experience, an interpretation that might in fact be a misidentification so that it would be false to say that a maximally great being really seemed to the high mystics, in the sense of “really seems” in the previous section.  Admittedly, a number of the high mystics such as St. John of the Cross were very sophisticated persons capable of carefully analyzing mystical experience, but that theological sophistication could have led them into attributing a content to their experiences that was not phenomenally present. 

However there is, arguably, an agreement between mystics of various religions that they have had an experience of a maximally great reality, a reality than which a greater cannot be conceived.  Perhaps on this point one might even adduce not only Western but also Eastern mystics.  If there is a wide-spread agreement among mystics of different theological persuasions about the apparent object of high mystical experiences having maximal greatness, we have reason to think that this agreement is grounded in the phenomenology of the experiences themselves, i.e., that these mystics are indeed having experiences as of a maximally great being, and perhaps even that their maximalistic theologies/theories-of-ultimate-reality are derived from this.  But if this is right, then by Śaṃkara’s principle it is possible that there exists a maximally great being.  And if necessary existence is a great-making property, i.e., a property that a being or reality must have if it is to count as maximally great, then by S5 there in fact exists a being or reality that in some possible world has the property of maximal greatness.  If, further, we admit that maximal greatness is the sort of property that an entity can only have essentially[6], since otherwise the greatness is “too fragile” to be maximal, it follows that there in fact in this world is a being that has maximal greatness.  And if we can argue the very reasonable claim that a maximally great entity is essentially perfectly good, omnipotent and omniscient, our argument for the existence of God is complete.

Note how the argument of the previous paragraph provides an answer to an objection that Saikat Guha has made in response to a draft of this paper.  Guha had objected that the Eastern mystics had the experience as of a necessarily existent essentially non-personal ultimate reality, which by my argument would then have to exist.  But the existence of such an ultimate reality is logically incompatible with the existence of the God of theism.  Thus, if Śaṃkara’s principle is true, then it is logically impossible that both the Western mystics have experienced what they claim to have experienced and that the Eastern ones have done so.  However perhaps one can identify a common core to their claims: both claim to have had experiences of a reality that has maximal greatness.  It is not implausible to suppose that some mystics may have misinterpreted their experiences on account of the properties they think a maximally great entity should have, while the phenomenal content of their experience was, at least in part, that of the presence of a maximally great entity.  By Śaṃkara’s principle, then, we can derive the possible existence of a maximally great entity from the experiences of both groups of mystics.  But now we need to ask what properties are in fact great-making properties.  I take it that it is highly plausible that omniscience, omnipotence and perfect goodness are all great-making properties.  Indeed, I can hardly see how a non-personal reality can be said to be very great, since such a reality cannot have moral qualities, and it is surely the moral ones that are most important for greatness.  Obviously more would need to be said here.  But as long as at least necessary existence and essential maximal greatness are accepted as great-making properties, we have shown that it is probable (as probable as the claim that someone has had an experience as of a maximally great entity) that a maximally great entity exists.[7]

Note that none of this entails that the high mystical experiences alluded to are in fact veridical or not.  The conclusion follows even if they are non-veridical.

If we are to use the second ontological argument, we need to argue that there is a person such that it is possible that she is radically dependent on some being.  Many of the Western mystics talk of having experienced an utter dependence on God.  This gives us a solid presumption in favor of supposing that at least one of them has had an experience as of her being utterly dependent on an essentially numinous being, and hence by Śaṃkara’s principle that it is possible that she in fact was utterly dependent on an essentially numinous being.  If we take the concept of “utter dependence” to entail the technical notion of “radical dependence”, i.e., of the impossibility of existence without the causal activity of another being, then we have an argument for the possibility of her being radically dependent on some essentially numinous being, from which it follows, as we have seen in Section I, that she in fact is radically dependent on some essentially numinous being.  Given what Western mystics say about God being Love itself, one can also add that a part of the same experience is that this numinous being is also essentially loving, so that we can conclude that in fact some mystic is radically dependent on an essentially numinous and loving being.

In conversation, Richard M. Gale has offered a general a priori objection to both of these ontomystical arguments.  It is just impossible to have a phenomenal experience of a modal property like “maximal greatness” or “essential dependence.”  But I do not see why this should be impossible.  Admittedly, there are serious disanalogies between these kinds of properties and many of those that we experience through the senses.  But I do not see a justification for a priori specifying certain kinds of experiences as impossible, unless of course one can argue that the apparent objects of these putative experiences are impossible.  Arguably, it is a part of theism that in some way God is always directly aware of the truth of all true propositions.  If this is right, it might well beg the question against theism to claim without argument that an experiential awareness of certain kinds of states of affairs is impossible.  Indeed, I take there to be a presumption that if an experience is claimed by an intelligent and careful observer and the object of the experience cannot be shown to be an impossible object, then the experience did occur much as described.

Moreover, arguably, we do have modally-charged experiences.  It seems reasonable to suppose that when I experience myself as responsible for some event, I experience the truth of the counterfactual that were I not to have made some choice that event would not have happened, and of course counterfactuals are modal beasts.  Likewise, there is no reason to suppose that we might not sometimes have direct broadly perceptual access to dispositional properties, which of course are modally charged.  Being tired entails having a dispositional property having to do with how one is disposed to react to the imposition of addition labor;  but being tired is also a property we have direct, though fallible, perceptual access to.

How strong are these arguments?  The most controversial two premisses will be Śaṃkara’s principle and claims about what “really seemed” to the mystics.  To make further progress on the latter issue will require a careful and in-depth examination of the writings of many mystics across religious traditions, and asking about how far these writings support the claim that a maximally great being or radical dependence were the apparent objects of the mystics’ phenomenal experiences.[8]

[1] Quoted in Arindam Chakrabarti, “Metaphysics in India”, in: Jaegwon Kim and Ernest Sosa (eds.), Companion to Metaphysics, Oxford / Cambridge, MA: Blackwell, 1995, 318–323, p. 319.  I am most grateful to Mr. Saikat Guha for the attribution of this principle to Śaṃkara and for the source of this quote.

[2] See The Nature of Necessity, Oxford: Clarendon, 1982, chapter 10.

[3] This last step uses S5, i.e., the transitivity and symmetry of the accessibility relation.

[4] Berkeley will use his account of speaking with the vulgar to argue that I mischaracterize the content of our phenomenal experience, and so the argument does not end here.  But I take it that the “literal” account of our phenomenal experience is at least much more plausible.

[5] Cf. Descartes’ famous remarks in his Meditations on the validity of mathematical proofs done in dreams.

[6] Cf. Plantinga, The Nature of Necessity, p. 214.

[7] A much less ecumenical reply to Guha would be that the Eastern mystics prove themselves to be inaccurate observers of experiences by also claiming to experience their own non-existence, which claim is surely a conceptual impossibility.  However, I shall not press this point, not just for ecumenical reasons, but also because highly paradoxical language is a feature of all mystical traditions.  I do not, however, agree with those who would claim that this paradoxical language is in general genuinely self-contradictory.  If it were, then it could not describe a unified experience.  Rather, I take the apparently self-contradictory parts of it to describe aspects of the experience that transcend, but not contradict, our conceptual powers, much in the way that many theologians think that faith transcends but does not contradict reason.

[8] I am very grateful to Richard M. Gale for his encouragement and a number of insightful comments on drafts of the main argument of this paper, and to Saikat Guha for interesting discussions.