Clarifications Regarding Ontological Arguments and Rationalism (Column 364)

In this column I wish to clarify several aspects that were not sufficiently sharpened in the previous column, all of which concern the relation between rationalism and ontological arguments.

Another Look at the Rationalist–Empiricist Debate

In the previous column I discussed, among other things, rationalism, which holds that it is possible to know the world by means of the intellect (and not only by means of the instruments of cognition). I noted that empiricism attacks rationalism at this point, claiming that it is impossible to rely on the intellect in coming to know the world. The principles of our thinking are founded on the structure of our thought and intellect, but we have no guarantee that this corresponds to what happens in the world itself.

But the claim that we rely on principles of thought in our knowledge of the world is nowadays agreed upon by any reasonable person. In this sense empiricism has lost decisively (contrary to what many people think, and in particular scientists). Beyond observations, it is clear that within the framework of science we also use generalizations (induction) and the principle of causality, as well as other assumptions—all of which are principles not based on observation. From this it follows, as David Hume already showed, that empiricists necessarily assume a correspondence between the form of our thinking (the intellect) and the world, even though he himself proposed an extreme thinning in order to preserve pure empiricism (to cling only to the direct data of the senses, and see everything else as statements about us and not about the world—what Zeev Bechler calls “actualism”). Kant tried to present a conceptual framework that would allow us to return to a sane empiricism, that is, one that also accepts the general laws of science even though we arrive at them by means of a priori conceptual principles. In my judgment, Kant failed in his heroic attempt to save science (by means of a thesis that is itself largely rationalist).

Intuition

Here I will only add that I, in my small way, argue in several places that the only way to philosophically ground science in its non-anorexic versions (the anorexic ones do not hold water in any philosophical sense) is to assume the existence of a non-sensory cognitive faculty—what we usually call intuition. My claim is that scientific generalizations are the fruit of observation by the eyes of the intellect; that is, they have an empirical basis, but not a sensory one. Therefore, one may perhaps treat them as synthetic a priori claims (prior to observation) if one insists that observation is exclusively the use of the senses. But from a substantive point of view these are synthetic a posteriori claims, since intuition too is a cognitive instrument (more precisely: an instrument that has cognitive components in addition to conceptual ones).

In several of my books I have noted that this is the only possible solution to Kant’s problem of the synthetic a priori. In the ninth chapter of his book, Introduction to Epistemology, Hugo Bergmann surveys all the solutions that were proposed to the problem—which he calls “the rationality of the world” (that is, the correspondence that allows us to understand the world with our intellect)—throughout the history of philosophy. In the end he concludes that to this day we stand before a broken cistern in this matter. Interestingly, at the end of the discussion he mentions, in passing and by way of negation, the only possibility that can indeed offer a solution to the problem (p. 182):

But if all these attempts have failed, Hume’s contention remains in full force. The principle of the lawfulness of nature, the supreme principle that underlies science in general and natural science in particular, the principle that lies at the basis of every search (induction) in science, remains without proof.[1]

If we do not wish to cut the Gordian knot of proving the rationality of the world by means of a metaphysical assumption, such as the guaranty of God (Descartes) or a pre-established harmony (Leibniz9) or the action of the intellect upon sensation prior to consciousness (subjective idealism), no path remains before us but to see the rationality of the world as a directive (regulative) assumption guiding the course of science.

In other words, one who wishes to uphold a non-anorexic science like that of Hume, and also not to see science as a collection of our subjective statements about ourselves, has no choice but to adopt metaphysical justifications, such as the guaranty of God. As I showed in the fourth conversation in The First Being and elsewhere (in particular in Two Wagons), the solution I proposed above—namely, our trust in the correspondence between intellect and world through validating intuition—can be grounded only on faith in God (a metaphysical factor that coordinates between us and the world). Bergmann mentions this possibility in passing and by way of negation (“if we do not wish…”), even though it is the only possible solution. For some reason he chooses not to adopt it, perhaps because this is not supposed to be the way of a contemporary philosopher of science.[2]

Incidentally, the solution he ultimately reaches—that the correspondence between intellect and world is a methodological directive (regulative) assumption and not a factual claim—of course does not hold water, for precisely the same reasons that David Hume’s diet does not hold. In brief, neither can explain the fact that our science works, beyond being a useful hypothesis. I elaborated on this in my article on Occam’s razor (see there the proof against actualism).

On the Futility of Philosophy: The Current State of the Rationalism–Empiricism Debate

The problem of the correspondence between intellect and world is perhaps the most fundamental problem of philosophy in its branches, for it is the cornerstone of the philosophical assumption that we can at all think about ourselves and the world with philosophical tools, and thereby also with scientific tools. And to our surprise it turns out that it has only one possible solution. Let this teach you that the common feeling that there is no point to philosophical engagement, since there are many opinions and systems and we have no way to adjudicate among them, is baseless. In this topic you can see a clear philosophical decision on one of the most painful and fundamental questions of philosophy. There are not many opinions here but many mistakes that have arisen throughout history. In the bottom line, only one possibility truly holds water.

In light of what I have described, it seems to me that the debate between empiricism and rationalism is already behind us. There is nowadays no doubt that science assumes a correspondence between intellect and world, and if we do not want (and in fact cannot) see it as merely a regulative assumption, there is no choice but to accept the possibility of deriving synthetic insights (containing information about the world) by a priori means (that is, by our intellect). That is to say, contrary to the prevalent beliefs today, rationalism has defeated empiricism. But as we shall immediately see, this is not the end of the discussion. The very distinction between the two requires further clarification.

Between Rationalism and Ontological Arguments

Within rationalism our conclusions are based on assumptions, except that these are assumptions grounded in intellect and not in observation, and for this rationalism is criticized by empiricists. But ‘ontological’ arguments, such as Descartes’ cogito (discussed in the previous column) or Anselm’s ontological proof, arouse sharp opposition—even from staunch rationalists like Kant. The reason is that the difficulty philosophers see in ontological arguments is deeper than the difficulty in rationalism (the assumption of correspondence between intellect and world).

In ontological arguments the conclusion is generated without any assumptions at all—neither observational nor intellectual. We define a concept, and a logical analysis of that concept (or of its definition) yields a conclusion that is a claim about the world (e.g., that God exists, or that I exist). This looks like sheer hocus-pocus. After all, definitions have no limitations, certainly not limitations stemming from reality; and if from a definition as such one can learn something about the world, that is already a miracle. It is no wonder that even rationalists are unwilling to accept the possibility of such arguments.

An ontological argument derives a claim about the world from purely conceptual analysis, without any assumptions—neither assumptions learned from observation nor rationalist assumptions. This is of course far more surprising than rationalism as such. It is no wonder that such a procedure met with opposition even from Kant (in his critique of Anselm’s ontological proof), whose principal epistemological project was to ground rationalism or at least show its possibility. I clarified this point well in the first conversation of my book The First Being.

Contradictions as Ontological Arguments

As I noted in the previous column, Yuval Steinitz presented an argument in favor of the admissibility of ontological arguments. He claimed that we commonly draw conclusions about the world (usually claims of non-existence) by virtue of logical contradictions. When a concept or some principle contains a contradiction, we assume this means that that principle or concept will not be realized in the world itself. Thus, for example, we deny the possibility that the domed roof of some structure, being square, would have the shape of a circle. This is of course a logical contradiction, and hence the concept of a square object that has a circular shape is self-contradictory and does not exist. Can we infer from this that such an object will also not exist in the world itself? This seemingly assumes a correspondence between our intellect and the world itself. And yet, almost all of us will agree that such an object will indeed not be found in the world. Behold that conceptual-logical analysis can serve as a basis from which a claim about the world can be derived. True, this is a non-existence claim (that such-and-such an object does not exist), but still it is a claim that says something about the world; and if one cannot infer anything about the world from conceptual analysis, then such a claim should not be acceptable either.

However, as I remarked there, one can perhaps say that the concept “a circular square dome” is empty of content, and therefore the claim that such a dome will not be found in the world says nothing. This claim contains an undefined (meaningless) concept, and hence it is itself devoid of content and meaning. But what about the claim that it is impossible that there be in the world a wall that stops every shell and at the same time also a shell that penetrates every wall? Here the two claims have a well-defined meaning, and yet there is a contradiction between them; therefore we infer that in the world itself these two objects will not both be found. Here we have an example of conceptual analysis that yields a meaningful claim which says something about the world, or that contains information about the world (a synthetic claim).

Of course, one can also argue here that the conjunction of these two sentences creates a combination that is meaningless. Given a wall that stops every shell, the phrase “a shell that penetrates every wall” is meaningless, since it concerns all walls, and one of those walls is precisely that wall that stops every shell. If so, that general sentence implicitly also claims that this shell penetrates that wall (which stops every shell). In other words, the combination “a shell that penetrates a wall that stops every shell” is meaningless. But this still feels like pilpul, since at the end of the day we do use conceptual analysis to derive conclusions about things that will or will not be found in the world.

What Is Philosophy?

In the previous column I cited Ron Aharoni’s words arguing that all of philosophy is based on conflating two perspectives on the human being: as subject and as object. Ontological arguments based on this contradiction are in fact the entire framework of philosophical discussion. The most salient example is Descartes’ cogito, discussed in the previous column, which rests on the fact that a person thinks about his thought—that is, he uses the intellect to discuss the existence of the intellect. Note that all of philosophy and logic are exposed to the same fallacy, for these are domains that use the instruments of thought to think about those very instruments. Aharoni proposed to sever this identity and thereby dismantle all of philosophy with a single sword-stroke. He claims that if we do not identify the I that thinks the cogito (the I as subject) with the thought it thinks (the I as object), the contradiction disappears and the conclusion collapses of itself. Put differently, Aharoni claims that a procedure of self-observation (in which I, as subject, observe myself as object) is empiricist (for the person is observing something, even if it is himself), and not rationalist (a pure thought procedure). That is, we have here an empirical claim and not a philosophical one.

If one accepts Aharoni’s claim, two possibilities remain before us regarding ontological arguments: 1) the conclusion is not correct; 2) the conclusion is correct but is not proven from conceptual analysis (it is not an ontological argument) but from observation. True, this is a person’s observation of himself, but according to Aharoni’s claim this too should be treated as observation, like any other observation. Aharoni is not claiming that the conclusions of philosophy are incorrect, but that they are the result of observation and not of thought. In so doing he abolishes the philosophical discipline and sets it upon an observational science. Thus, for example, regarding the cogito, Aharoni of course does not claim that I do not exist, but that my existence is not the result of conceptual analysis but of observation.

In the previous column I showed why I agree with this distinction regarding the cogito, but here I must add that I do not agree with his basic claim that abolishes the possibility of philosophy (see in detail in the series of columns 155–160). To understand this I will recall what I wrote above: I propose a different solution than Kant’s to the problem of the synthetic a priori. My claim is that our intuition is a cognitive instrument and not pure thought (or “pure reason,” in Kant’s terminology). True, this cognition does not take place by the senses but by the intellect (“the eyes of the intellect,” in Maimonides’ terminology at the beginning of The Guide of the Perplexed), and therefore there is perhaps room to treat it as a priori; but in essence it is an empirical observation (non-sensory). If so, I agree with the pre-Kantian claim that one cannot learn anything about the world without observation, but I maintain that intuition is an observational instrument. The generalizations we make, or the adoption of the principle of causality and the identification of causal relations between different events—all these are done by means of our intuition, and this is (the only) justification we have for them.

In columns 155–160 I showed that this is the right way to define philosophy: it is essentially a kind of observational science that deals with “facts” that we observe by the intellect and not by the senses. Science deals with sensory observations (although above we saw that this is not precise, since science too has rationalist components), and philosophy deals with intuitive observations. Note that the solution I propose is similar in a certain sense to Aharoni’s. I agree with his words that philosophy is observational, but my claim is that these are non-sensory observations. The upshot is that he is not correct in saying it can be reduced to science. It remains an independent discipline, and, as noted, it is also the discipline that undergirds science itself.

And Despite It All: More on the Possibility of Ontological Arguments

I now wish to draw attention to the fact that if indeed we view intuition as a cognitive instrument, there is no impediment to accepting even the possibility of ontological arguments. In several places (such as in Two Wagons, in the second gate) I pointed out that at least some of our concepts are not a conventional creation (the conventionalist approach), but the result of our (non-sensory) observation of the world—particularly the world of ideas (the essentialist approach). We observe with our intellect the world of ideas and identify different features of the concept in question and the relations it has with other concepts.

Accordingly, a logical and rational analysis of a concept’s definition can indeed yield insights about the world, for it is the product of observation and not of pure thought. The analysis of the claim “I think” is nothing but an examination of the concepts involved in it (the processes of thought); hence from this analysis I can in principle arrive at insights that contain information about the world. In my view, however, it is unlikely that I would thereby arrive at existential statements about the world (such as my own existence—the cogito—or the existence of God—the ontological proof), for the observation of an idea cannot prove to us its realization in the world. And indeed I showed that both the cogito (in the previous column) and the ontological proof (in the first conversation of The First Being and in the first notebook) do not hold water. Those are not, in fact, true ontological arguments. But arguments such as those brought by Steinitz, whose goal is to deny the existence of objects or principles by virtue of contradictions they contain, can, on their face, certainly be admissible. I examine the world of ideas and see that there is no such idea, and infer from this that it cannot be realized either (there is nothing to be realized). Thus, for example, I see a contradiction between the idea of a wall that stops every shell and the idea of a shell that penetrates every wall, and they cannot coexist in the world of ideas and therefore also cannot both be realized in our world. By contrast, the fact that there is a well-defined idea does not necessarily mean that it is realized, and therefore observing it will not yield an existential statement. It seems to me there is an asymmetry between existential statements and non-existence statements. Perhaps this is what Tolginos meant in the discussion we held (in the first comments to the previous column) regarding Steinitz and ontological arguments, though I assume he would not agree with the rationale I have presented here as a justification for his claim.

In a marginal note I will add that when discussing an object that is necessary-of-existence, there may perhaps be a possibility to prove by an ontological argument the existence of such a concrete object (to derive an existential statement from an ontological argument). The existence of such an object is the result of the necessity of the realization of the idea. If observing the idea shows me that its existence (more precisely: its realization in the world, not its existence as an idea) is necessary, I can infer from this that it is realized in the world. Even here there is much room for pilpul, and this is not the place.

[1] In my opinion he employs a minimizing formulation here: it remains without justification, not merely without proof. It is important to understand that we have here an assumption which, lacking justification, there is no reason whatsoever to hold. The speculation that posits a correspondence between our intellect and the world is so far from reason and from a priori plausibility that a real justification is required in order to hold it. See the fourth conversation in my book The First Being (and the fourth notebook here).

[2] I have already remarked several times that Bergmann himself was a believing man who grew very close to Judaism at the end of his life. Moreover, in the preface to an earlier (!) book of his, Thinkers of the Generation, he explains that faith in God is the only possible explanation for this problem. It is strange that here, in a later book, he is unwilling to adopt it. Perhaps because this is not a philosophical solution, and in a book whose aim is a survey of philosophical problems and their solutions he did not wish to resort to such a solution. This still requires investigation.