From Ch. 5, "Some Metaphysical Implications of Ethics and Religion", of A. E. Taylor's The Elements of Metaphysics (1903).

Arguments for the Existence of God

      § 9. These reflections may naturally lead to some remarks, which shall be made as brief as possible, about the so-called philosophical "arguments for the existence of God" which played a prominent part in Metaphysics before their discrediting at the hands of Kant and Hume.1 Kant's great achievement lies in having demonstrated that the whole force of the "proofs" depends upon the famous ontological argument, best known in modern Philosophy in the form adopted by Descartes in the fifth Meditation. Descartes there argues thus: -- By "God" I mean a completely perfect being. Now, existence is a perfection, and non-existence an imperfection. Hence I cannot think of a non-existing perfect being without self-contradiction. Hence God, because by hypothesis perfect, must exist, and is the only being whose existence logically follows from its definition.

      Kant's even more famous criticism of this famous inference turns upon the principle which he had learned from his study of Hume, that logical necessity is "subjective." If I think of a logical subject as defined by certain properties, he argues in effect, I am necessitated to ascribe to it all the predicates implied in that definition. That is, I must affirm them or contradict myself. Hence, if "existence" is originally included among the perfections by which the subject "God" is defined, the proposition God exists is certainly necessary, but is also tautological, and amounts, in fact, to the mere assertion that "an existing perfect being is an existing perfect being." But if the "existence" spoken of in the predicate is something not included in the definition of the subject, then you cannot infer it from that definition. Now "real existence" is not a predicate which can be included in the definition of a concept. The predicates by which an imaginary hundred dollars are defined are the same as those of a real hundred dollars. It is not by the possession of a new predicate, but by being actually given in a concrete experience, that the real coins differ from the imaginary. Hence all propositions asserting real existence are synthetic, (i.e. assert of their subject something which is not contained in the concept of it), and the real existence of God or any other object can never be deduced from its definition.2

      This Kantian criticism has itself been subjected to much criticism, principally at the hands of Hegel and those subsequent philosophers who have been specially affected by the Hegelian influence. What appears to be the general principle of the Hegelian criticism has been most clearly expressed in English philosophy by Mr. Bradley,3 upon whose discussion the following remarks are chiefly founded.

      In estimating the worth of the ontological proof, we must distinguish between the general principle implied in it and the particular form in which it presents that principle. It is manifest that Kant is perfectly right when he contends that, taking existence to mean presence in the space and time order, you cannot reason from my possession of any idea to the existence of a corresponding object. You cannot say whatever I conceive must exist as I conceive it. But the principle of the ontological proof is perhaps not necessarily condemned by its failure to be thus universally applicable. The principle involved appears to be simply this. The idea and the reality outside its own existence as a fact in the time order which it "means" or "stands for" are mutually complementary aspects of a whole Reality which include them both. For there is, on the one side, no "idea" so poor and untrue as not to have some meaning or objective reference beyond its own present existence.4 And, on the other, what has no significance for any subject of experience is nothing. Hence in its most general form the ontological argument is simply a statement that reality and meaning for a subject mutually imply one another. But it does not follow that all thoughts are equally true and significant. In other words, though every thought means something beyond its own existence, different thoughts may represent the structure of that which they mean with very different grades of adequacy. That which my thought means may be far from being real in the form in which I think it.

      Now, we may surely say that the more internally harmonious and systematic my thought is, the more adequately it represents the true nature of that which it means. If thoroughly systematic coherent thought may be mere misrepresentation, our whole criterion of scientific truth is worthless. How freely we use this ontological argument in practice will be readily seen by considering the way in which, e.g., in the interpretation and reconstruction of historical facts, the internal coherency of a systematic and comprehensive interpretation is taken as itself the evidence of its truth.5 Hence it may be argued that if there is a systematic way of thinking about Reality which is absolutely and entirely internally coherent, and from its own nature must remain so, however the detailed content of our ideas should grow in complexity, we may confidently say that such a scheme of thought faithfully represents the Reality for which it stands, so far as any thought can represent Reality. That is, while the thought would not be the Reality because it still remains thought, which means something beyond its own existence, it would require no modification of structure but only supplementation in detail to make it the truth.

      But if we have anywhere thought which is thus internally coherent, and from its own nature must remain so, however knowledge may extend, we have it surely in our metaphysical conception of the real as the absolutely individual. Thus the ontological proof appears, in any sense in which it is not fallacious, to amount merely to the principle that significant thought gives us genuine knowledge; and therefore, since the thoroughgoing individuality of structure of its object is presupposed in all significant thought, Reality must be a perfect individual. That this perfect individual must further be "God," i.e. must have the special character ascribed to it by beliefs based upon specifically religious emotions, does not follow. How far the "God" of religion is a correct conception of the metaphysical Absolute, we can only learn from the analysis of typical expressions of the religious experience itself. And it is obvious that if by "God" we mean anything less than the Absolute whole, the ontological proof ceases to have any cogency. It is impossible to show that the possibility of significant thought implies the presence of a special finite being, not empirically known to us, within the Absolute.

      The "cosmological" proof, or "argument from the contingency of the world," unlike the ontological, has the appearance, at first sight, of starting with given empirical fact. As summarised by Kant for purposes of criticism, it runs thus: If anything at all exists, there must be also an absolutely necessary being. Now, I exist myself; ergo, the absolutely necessary being exists." To make the proof quite complete, it would be necessary to show that the being whose existence is affirmed in the minor premisses, to wit, myself, is not itself the "absolutely necessary being," and the argument thus completed would become in principle identical with the second of the "proofs" given by Descartes in the third Meditation, where it is inferred that if I, a dependent being, exist, there must be a God on whom I and all things depend.6 As Kant has pointed out, the whole force of this inference rests upon the previous admission of the ontological argument. By itself the cosmological proof only establishes the conclusion that if any dependent existence is real, independent existence of some kind must be real also. To convert this into a "proof of the existence of God," you must further go on to identify the "independent existence" thus reached with the "most real" or "most perfect" being of the ontological proof. For otherwise it might be suggested, as is done by one of the speakers in Hume's dialogue, that the series of phenomenal events itself, taken as an aggregate, is the "necessary existence" upon which the "contingent existence" of each several event depends. "Did I show you the particular causes of each individual in a collection of twenty particles of matter, I should think it very unreasonable should you afterwards ask me what was the cause of the whole twenty. This is sufficiently explained in explaining the cause of the parts."

      To avoid this objection, we must go on to maintain that only the "most perfect being" can be an ultimately necessary being, and that its "necessary existence" is a consequence of its character. This, as we have seen, is the very assertion made in the ontological proof. Hence our criticisms of the ontological proof will be equally applicable to the cosmological. If we combine the two, restating them in accord with our previous remodelling of the former, the argument will take the following form. All propositions directly or indirectly refer to real existence. Hence it would be self-contradictory to assert that nothing exists. But existence itself is only conceivable as individual. Hence the absolutely individual must be really existent. And this is identical with the general principle of our own reasoning in Book II. of the present work. Clearly, if valid, it is valid simply as an argument for a metaphysical Absolute; it neither proves that Absolute itself to be what we mean in religion by God, nor affords any ground for asserting the existence of God as a finite individual within the Absolute.7

      The physico-theological argument, also known as the argument from design, or the teleological proof, differs from the preceding two in being in its current forms honestly empirical. In the shape of an inference from the apparent presence of order and a regard for human good in the structure of nature to the existence of a wise and benevolent being or beings as the author or authors of nature, it has been the most popular of all theistic arguments both in the ancient world, where, according to Xenophon, it was specially insisted upon by Socrates, and in the modern defences of theological beliefs against rationalistic criticism. It must, however, be observed that the criticisms of Hume and Kant are absolutely fatal. to the "argument from design," when it is put forward as a proof of the existence of a God of infinite goodness and wisdom. At best, as Kant says, the observed order and harmony of Nature would enable us to infer a finite degree of wisdom and goodness in its author. The assertion of the absolute harmoniousness and goodness of Nature, which we require to justify the inference to infinite wisdom and goodness in its author, goes far beyond the limits of the empirically verifiable, and can itself only be upheld by some form of the "ontological proof." Hence the "argument from design" could at best prove a God whose wisdom and goodness are, so far as knowable, limited. As Hume forcibly puts the same point, if the empirically known facts of the partial adaptation of Nature to human purposes are valid, as they stand, to prove a wise and good intelligence, are not the equally well-ascertained facts of the partial want of adaptation equally valid to prove defective goodness or defective wisdom?8

      There is a deeper metaphysical reason for this difference between the results of the physico-theological and of the other "proofs," which may be briefly pointed out. The whole conception of the order and systematic unity of the world as due to preconceived "design" is only intelligible if we suppose the author of that "design" to be finite, and subject, like ourselves to temporal mutability. For in the notion of design itself are implied the severance of the mentally conceived ideal from the actuality which waits to be brought into accord with it, and consequently also the time-process which we have already found to be characteristic of all finitude. Hence the physico-theological proof, by itself, can at best be used to establish the reality of finite "gods," not of "God" because it works throughout with the categories of finitude.

      Upon the logical force of the argument, as thus limited by its initial assumptions, only one observation need be made. What the reasoning asserts is not merely that "Nature" is in reality a system exhibiting individuality and purposive interest or even "design" but that it reveals the particular design of assisting and fostering human progress. Now, whether this is so or not would appear to be a question of empirical fact only capable of determination by the methods applicable to other problems of the same empirical kind. Probably the lines along which it will have to be decided in the future are of the following general kind. Evolutionary science seems clearly to have shown that in the influences it knows, e.g,., as "natural" and again as "sexual selection," we have processes which lead to beneficial results without being, so far as we can see, in the least directed by the conscious "design" of establishing those results.9 We should have to ask, then, whether there is actual ground for holding that such influences are not of themselves sufficient to account for the development of human civilisation, so far as it is due to factors belonging to the "environment." If they are so sufficient, the "physico-theological" argument for benevolent super-human agency in moulding the course of human development, becomes superfluous; if they are not, their failure is, so far, good ground for the recognition of finite "designing" intelligences of a non-human kind as forming a factor in our environment. In either case the question appears to be one of empirical fact, and to be incapable of determination in advance on general metaphysical grounds.10 Nor are we justified in assuming that "design in nature," supposing it to exist, must always be directed to securing ends which are either intelligible to us, or, if intelligible, "benevolent," in the sense of furthering our own special human interests. And here I must be content to leave the subject.

      Consult further: -- F. H. Bradley, Appearance and Reality, chaps. 25, 26; J. E. McTaggart, Studies in Hegelian Cosmology, chaps. 6, 8; J. Royce, The World and the Individual, Second Series, lects. 9, 10.


      1. Kant's famous onslaught will be found in the Kritik der Reinen Vernunft, Transcendental Dialectic, bk. ii. div. 3 ("The Ideal of Pure Reason " §§ 3-7. Hume's criticisms are contained in his posthumous Dialogues concerning Natural Religion.

      2. Kant's criticism had been in part anticipated on the first circulation of the Meditations by both Mersenne and Gassendi. See particularly Gassendi's strictures on Descartes confusion of existence with properties in the "Fifth Objections," with Descartes' unsatisfactory reply. Leibnitz repeated the same objection, and proposed to amend the Cartesian proof by a formal demonstration that God's existence is possible, i.e. does not imply a formal contradiction. He then argues -- If God's existence is possible, He exists (by the Cartesian proof). But God's existence is possible, therefore God exists. See, e.g., Leibnitz, Works, ed. Erdmann, p. 177 ; and Latta, Monadology of Leibniz, p. 274. Hume's comments are even more akin to Kant's. "Whatever we conceive as existent we can also conceive as non-existent. There is no being, therefore, whose non-existence implies a contradiction. Consequently there is no being whose existence is demonstrable." (Dialogues concerning Natural Religion, part 9.)

      3. Appearance and Reality, chap. 24.

      4. No thought can be merely and absolutely false, any more than any act can be merely and without qualification bad. Though words may be entirely meaningless, thoughts cannot be.

      5. The appeal to experiment is no objection to the principle. For in making the experiment we do not, of course, get out of the circle of our thoughts, and the experiment only affords a criterion of truth in so far as it leaves us with a new thought which can only be brought into systematic harmony with our old ideas in one determinate way. Except as interpreted by thought, the experiment has no bearing on our knowledge.

      6. This was also a favourite argument with Leibnitz, as Kant notes. For an acute examination of Leibnitz's use of it and the other "proofs," see B. Russell's Philosophy of Leibniz, chap. 15. For Hume's objections to it, see the already quoted part 9 of the Dialogue concerning Natural Religion. The other "proof " of the Third Meditation, namely, that my possession of an idea of God, which I could not have derived from empirical sources, proves the reality of the idea's object, is only a special form of the ontological argument from idea to existence.

      7. As thus remodelled, the double ontologico-cosmological argument might be attacked on two grounds -- (1) That it only proves, once more, that if we admit that all propositions are concerned with real existence, either directly or remotely, we must admit the existence of the Absolute, but does not demonstrate that all propositions are so concerned. (2) That in saying that existence is only conceivable as individual we fall back into the Cartesian misconception of existence as a predicate. I should reply, (1) that the validity of the premiss in question cannot be denied without being confirmed in the act of denial. I.e. unless the suggested proposition that "some propositions at least have no reference to a reality beyond their own presence as psychical facts in my mind," itself has the very objective reference in question, it has no meaning, and is therefore no genuine proposition; (2) that we must distinguish between the what and the that of existence. The "that" of existence is not conceivable at all, but our position is that this inconceivable that is only logically, not really, separable from a what, and that it is precisely this inseparability of the that and the what which we mean by "individuality."

      8. Dialogues concerning Natural Religion, part II.

9. This is quite consistent with our own view that all real processes are teleological in the sense of being marked by subjective interest. For (a) not by any means all teleological process is actual "design" or "volition" (impulse organic craving, habit, etc., are all cases in point); and (b) actual volition need not, always be volition for the result it actually produces. Sexual selection in man would be an instance of a process which may take the form of actual volition, but in that case is rarely, if ever, volition for that improvement of the stock which de facto issues from it.

      10. Cf. Bradley, Appearance and Reality, pp. 200, 496-497 (1st ed.). Professor Flint's attempted reply to the Humian and Kantian criticism of the theistic "proofs" (Agnosticism, chap. 4) has not induced me to modify any of the opinions expressed in this chapter.