An Orthodox Critique of Natural Theology

In one sense metaphysics is prior to knowledge, since it would be the ontological structures of the world that determine whether knowledge is possible, how it is acquired, and under what conditions knowledge is obtained. However, the difficulty is that one has to “know” which metaphysical account is correct in order to determine the right epistemology. Hence, we have something similar to Chisholm’s problem of criteria.[1] In addition, it is important that we make a distinction between 1^st order knowledge (e.g., S knows that p) and 2^nd order knowledge (e.g., S asserts that he knows that p; S knows that he knows that p). The latter (2^nd order knowledge) is what we are concerned with in this paper, since Natural Theology’s claim is “that we can know God.” This is a 2^nd order knowledge claim that requires justification. It may be possible that a person S knows that p, while at the same time that they or we are unable to justify the statement that “S knows that p.” In other words, 1^st order knowledge is not contingent upon 2^nd order knowledge; however, Natural Theology deals with 2^nd order knowledge statements (e.g. “man can know God by the natural light of reason apart from revelation”) and, therefore, we must see whether there is justification for such assertions from the various assumed epistemic starting points. Moreover, we are not just concerned with the justification for particular propositions and whether they qualify as knowledge, but generally speaking, whether it is possible to be justified in asserting that knowledge exists at all (i.e., “I know that knowledge in general exists”) from certain assumptions, since without a justification for this 2^nd order statement about knowledge in general, there is no justification for more specific 2^nd order statements like, “S knows that God exists.” This has particular relevance to our considerations about Natural Theology as a philosophical and epistemological project. I will argue that something that the Natural Theologian, from their particular assumptions and prior premises, will not be able to answer without falling into question-begging or arbitrariness. For example, can a person be justified in saying that they can ‘know’ God exists from the natural light of reason alone before they can justify whether knowledge exists or before justifying that logical or deductive inferences is a legitimate cognitive/epistemic process that leads to knowledge? It seems that the Natural Theologian has to assume this is true, which leads to the absurdity that you can be justified in asserting that you know something even if knowledge is impossible or you are unable to justify your assertion. In other words, the Natural Theologian must be committed to believing that you can be justified without any justifications, a clear contradiction.

A common response to the 1^st order and 2^nd order knowledge problem outlined above is an appeal to internalism. Internalism is the theory of knowledge that is set in contrast to externalism and it asserts that if an agent has an internal a priori and immediate 1^st person access to the reasons/justification of a belief, then they have knowledge.[2] If this is what constitutes knowledge, then it follows that if S knows that p, then S knows that they know that p, since by the internalist’s definition, to have knowledge is to have access to the justifications or reasons why they have knowledge. The problem with this, however, is that it begs the question. How do we know that internalism is true? One will inevitably have to answer, “well if internalism is true, then we can know internalism is true.” Clearly, this is viciously circular. And if we cannot determine whether internalism is true without begging the question, then we cannot justify the claim that “if person S knows that p, then S knows that they know that p.” Second, there seems to be some intuitive and common-sense examples of knowledge that if true, would make internalism false. These objections are known as the examples from epistemically unsophisticated subjects (e.g., highly intelligent animals and young children). For example, young children certainly do not have an internal 1^st person access to the reasons a justifications for a variety of true beliefs they hold, but common-sense would suggest that they nevertheless have knowledge about these things. Furthermore, as Lawrence Bonjour points out, “if internalism is correct, only at best a few epistemologist and students of epistemology will have access to good reasons for the vast majority of beliefs that common sense regards as justified and as constituting knowledge.”[3] The problem becomes an insurmountable epistemic problem if it can be shown (e.g., via skepticism, et. al.) that no epistemologist is in a position to provide a justification for their justification criteria without falling into the vicious circularity of epistemic bootstrapping, that is to say, the epistemologist from their own autonomous assumptions and starting points cannot justify their belief that knowledge is possible or that S knows that p.

Returning to the question of 1^st and 2^nd order knowledge, it seems that 2^nd order knowledge (i.e., knowing that you know, or asserting “S knows that p”) is not necessary to have 1st order knowledge; nevertheless, as discussed above, the question is not about particular propositions of knowledge. The focus of our concern is about being able to justify possibility of knowledge in general. In other words, whether it is the individual subject knowing or another person making  claims about people knowing, someone must be able to first justify that knowledge in general exists, and justify how knowledge is possible, if any one individual is going to be said to have some particular knowledge. One possible way to solve this problem is to frame it in a conditional: if knowledge turns out to be possible, then my true belief (under certain specified justification conditions) would amount to knowledge. The problem with this suggestion, however, is that it either results in a problematic tautology, “knowledge exists if knowledge exists,” or it suggests we could have knowledge before having a justification criteria for obtaining true beliefs, making our true beliefs accidental. It is universally agreed that accidently arriving at true beliefs never amounts to knowledge.[4] Moreover, it doesn’t seem that the externalism, the other epistemic theory, fares any better.[5] Nevertheless, the real issue is that if the justification criteria for 2^nd order statements like “S knows God exists” is not grounded in divine revelation, then they are grounded in autonomous epistemic starting points that fail to do the epistemic work of providing a justification that knowledge exists; therefore, the Natural Theologian’s claim that one can know the existence of God by the natural light of reason alone cannot be justified. As a result, Natural Theology cannot make any claims that its conclusions are justified and rational.

A second problem with Natural Theology concerns metaphysics and epistemology in general. The Natural Theologian must admit that someone can start from different metaphysical and epistemological assumptions and nevertheless be justified in reasoning to a correct conclusion – that God exists. The problem, however, is that admittedly these various assumptions cannot all be true yet different at the same time. Inferences from false premises or starting points to true conclusions do not amount to good reasoning. Since making such inferences is not a sufficient justification for accepting what would be a true conclusion, then it is not clear how Natural Theology can claim that its conclusions are justified and rational. Hence, getting one’s metaphysics and epistemology correct is essential; however, how can one, from the “natural light of reason alone,” determine what the correct metaphysics and epistemology are without falling into vicious circularity?

Having outlined the general problems with Natural Theology’s inability to establish second-order claims about the existence of knowledge without epistemic-bootstrapping, which in turn threatens its claims that its conclusions are justified and rational, let us now turn to more specific problems. Much of the following argument relies on what György Geréby presents in his paper, “Theistic Fallacies,” which will be briefly outlined in this section.

First, Natural Theology assumes there is a “greatest common denominator”[6] among the “monotheistic” religions, and therefore, because it does not identify with any particular religion, maintaining that theistic arguments establish an “essential core” [7] of monotheistic religious beliefs, it can argue for the existence of a transcendent god who is creator and governor of the world. If this is true, then it can justify those shared essential religious tenants (i.e., “the core doctrine”) as rational, consistent, and logically provable. However, as I will attempt to show, since there are fatal philosophical and logical flaws with the idea of a “greatest common denominator” or what Geréby calls this the “rational core doctrine,” Natural Theology will not be able to prove or rationally argue for the existence of God.[8]

Geréby states that the theist[9] position can be characterized by the following main points (1-8) with point 9 being what he calls the theist corollary. There is a god who is:

1. omniscient
2. omnipotent
3. eternal
4. free
5. good
6. creator and sustainer of the universe
7. a person
8. worthy of worship and obedience
9. the theist god is the same for all monotheistic religions

However, since the term ‘god’ – even if considered a proper name – is not a ‘rigid designator, referring to one and only one entity, there can be no “rational core doctrine.” Consequently, since Natural Theology’s arguments for the existence of God depend on a “rational core doctrine,” they will not work for rationally establishing the existence of god. As Geréby correctly points out: “A proper name has to behave like a ‘rigid designator,’ always referring to the same individual in all possible situations.[10] But the term ‘god’ is not ‘rigid,’ since it can mean different things in different possible worlds, like for Aristotle, Averroës, or for Aquinas.”[11] The failure of Natural Theology here is in neglecting the special semantic and logical behavior of the term ‘god.’  This results in what is called the “Theist Fallacy,” what Geréby identifies as the consequence of the non-substitutability of identicals in belief contexts. 

The “Theist Fallacy” is illustrated with several examples. Since the Natural Theologian assumes the theist position and believes that ‘god’ refers to the same entity in all three major monotheistic religions, we can summarize their position with the following argument:

1. Judaism believes in one and only one god.
2. Christianity believes in one and only one god.
3. Islam believes in one and only one god.
4. There is one and only one god.

Therefore, Judaism, Christianity and Islam believe in the same one god.[12]

First, it does not follow that from the fact that Judaism, Christianity and Islam believe that there is one and only one god, that they believe in the same one god. This is known as the quantifier-shift fallacy. For example, just because every baby has a mother, it does not follow that therefore every baby has the same one mother.[13] Second, if we construct a similar argument, we can see how the theist argument of the Natural Theologian clearly fails. Geréby constructs a similar argument from mathematics:

1. Ahmed believes that there is one and only one solution of the problem.
2. Isaac believes that there is one and only one solution of the problem.
3. John believes that there is one and only one solution of the problem.
4. There is one and only one solution of the problem.

Conclusion: Ahmed, Isaac and John believe in the same solution of the problem.[14]

Clearly the conclusion does not follow, which shows that the “theist corollary” that Judaism, Christianity, and Islam believe in the same one god does not follow either. The theist corollary of Natural Theology is refuted by another example, which Geréby cites from Albert of Saxony to demonstrate the fallacious inference:

[1] Aristotle believed in god. [2] God is the Trinity. Therefore, Aristotle believed in the Trinity.[15] 

This is connected to the logical problem concerning intentionality that is illustrated in Eubulides’s famous Hooded man paradox[16]:

[1] John does not know the hooded man coming. [2] The hooded man coming is John’s father. Conclusion: Therefore, John does not know and does know his father.[17]

Again, the general problem here is the non-substitutivity of identicals in epistemic (or intensional) contexts. In other words, “expressions with identical referents cannot be replaced in propositions containing belief-operators, and retain the same truth conditions.”[18]  Consequently, we can derive a principle of the non-substitutivity of identicals, which states that in “epistemic (or intensional) contexts identicals cannot be replaced salva veritate.”[19]

Since the Natural Theologian erroneously assumes there is a shared “core belief” among monotheistic religions, which has been disproven in the above analysis, it follows that Natural Theology’s arguments for the existence of god do not establish the existence of a common referent we can call god. Simply having common divine properties (1-8) is not sufficient for establishing the same referent nor does it justify that there is a “core doctrine” among the three religions.  This is a general problem concerning intentionality and difficulties with determining a common referent. For example, we may talk about the place “the Promised Land” and describe the common properties as being a place where one found Jesus, where one gets all their theology, and where one hopes to return, only to find out that one person is talking about their local Christian bookstore and the other is talking about the place in Israel.[20] The fact of sharing common properties does not imply that there are “core properties” so as to establish a common referent[21] is even more pronounced with things that are considered “mental objects”[22] as opposed to physical objects or places. Unlike physical objects whose identity is established by laws of nature independent of the mind, God is not a natural entity and operates much more like a “mental object” that is identified only by its content, the semantic and epistemic rules of application. In this case, it is the religious use that specifies the meaning of the term god. And because there are three different proper names for god among the three monotheistic religions (with no cross-identifiable referents), this leads Geréby to conclude that indeed the meaning of the three different names is dependent upon the context of rules and instructions of each religion, “forming different god-images, which is sanctioned by the logical character of these beliefs.”[23] In fact, religious “doctrines are similar to long conjunctions of religious (intentional) statements,” meaning that if “the length or the content of such conjunctions gets changed, the truth conditions of the conjunct will also change.”[24] And since each religion has different additions (contents) when defining God in those conjunctions, different truth values for its conjunctions, the term “god, despite having commonly defined properties, cannot serve as a common denominator[25] among religions such that the “existence of God” can be rationally deduced and justified. [26]

As we have seen, Natural Theology has been unable to justify it claims that the human mind by the “light of natural reason alone” can know that God exists apart from revelation, since it cannot make 2^nd order knowledge claims “that one can know God exists” without falling into epistemic-bootstrapping. Due to Natural Theology’s commitment to autonomous epistemology, it also fails to resolve the epistemic difficulties that result from the dialectical interplay between metaphysics and epistemology, making its philosophic project problematic and unjustified. Moreover, natural theology admits that one can start from various metaphysical and epistemological assumptions (various false starting points) and still be justified in reasoning to correct conclusions. Lastly, in assuming there is a “rational core” among the monotheistic religions, Natural Theology is lead to the theist fallacy that occurs when failing to recognize the non-substitutivity of identicals in epistemic or intensional contexts. Since the veracity of Natural Theology depends on there being a common “rational core” doctrine, the inability to establish a “rational core” doctrine demonstrates that Natural Theology itself fails. Since this paper is primarily an epistemological analysis to discover a proper foundation for knowledge and reasoning that doesn’t rely on unjustified or arbitrary assumptions, a critique of natural theology does not amount to abandoning reason in favor of fideism by simply presupposing God and His Divine Revelation.[27] I argue elsewhere that if the epistemic difficulties arise precisely from the nature of autonomous epistemology, then by locating our justification criteria in divine revelation (theonomous epistemology), we avoid the aforementioned problems and will be able to provide the necessary condition for the possibility of knowledge.

[1] Metaphysics will determine epistemology (how and what we can know); however, I have to be able to “know” (i.e., epistemology) what the correct metaphysics is first; therefore,  I cannot answer the first part, without establishing a correct epistemology (the second part), but I cannot establish a correct epistemology (the second part) without establishing the first part, and so on.

[2] “The internalist monist maintains that all warrants are of an internal character in the sense mentioned earlier, i.e. in the sense that the reasons underwriting all warrants must be accessible to the warranted individual through reflection alone.” (Christoph Kelp, “Second Order Knowledge,” 2)

[3] Lawrence Bonjour, Epistemology: Classic Problems and Contemporary Responses, 208.

[4] See the Gettier problem.

[5] For suppose a child is given a complicated algebraic math problem that he does not understand and employs some sort of nonsensical method (which he also does not understand), but that method ends up being reliable for obtaining the correct answer each and every time. The child, however, does not understand the method nor does he understand why it gets him to the correct answer. Would we say this child has knowledge when he employs a correct/reliable method that gets him a correct answer? Most likely we would say no. There is something more that is required than simply having a correct criteria or reliable method for getting true beliefs. We can always ask why reliability (something externalism asserts) constitutes as being a justification. Attempts to answer this question will inevitably either be circular (because reliability is a justification) or they will simply be incapable of providing an answer.

Bonjour states something very similar, asking us to imagine a situation “in which there is a cognitive process that is in fact highly reliable, but which we have no reason to regard as reliable and perhaps even good reason to regard as unreliable. Thus suppose that clairvoyance, the alleged cognitive ability to have knowledge of distant occurrence in a way that does not depend on sensory perceptions and involves a process of some unknown sort that is in fact highly reliable for certain specific people under certain specific conditions (which might include a limitation to a certain range of subject matter). And suppose that some person who in fact has this ability arrives at a belief on this basis and that the requisite conditions for reliability, whatever they may be, are satisfied. Such a belief seems to satisfy the reliabilist requirement for justification, but is it in fact genuinely justified?” (Bonjour, Epistemology, 212)

[6] Geréby calls this the “rational core doctrine.”

[7] “I would like to contend that since it cannot be proved that the “monotheistic” religions worship the same god, and that there is no independently, or philosophically demonstrable “essential core” of religious beliefs, albeit the theistic program implies that. Therefore the initial assumptions of the theistic program are inadequate for the task it sets for itself.” (Geréby, “Theistic Fallacies,” 2)

[8] As Geréby points out, this specific critique of the “rational core doctrine” proceeds as a Modus Tollens argument (p ⸧q / – p // – q): If there is a rational core doctrine, then theistic arguments for the existence of God can work. There is no rational core doctrine. Therefore, theistic arguments for the existence of God do not work.

[9] Geréby cites Richard Swinburne in defining the term ‘theist’: “By a theist I understand a man who believes that there is a God. By a “God” he understands something like a person without a body (i.e. a spirit) who is eternal, free, able to do anything, knows everything, is perfectly good,… who is the proper object of human worship and obedience, the creator and the sustainer of the universe.” (Swinburne, The Coherence of Theism (rev. ed., Oxford 1993, 1)

[10] Furthermore, “there is no such identifiable individual, that would qualify for the single referent of the term (or there are as many, as there are describing a certain identity). The term “god” is therefore a “flaccid designator” even for the monotheists, which could pick up different values in different religious frameworks.” (Geréby, “Theistic Fallacies,” 8)

[11] Ibid, 8.

[12] Ibid., 9.

[13] Geréby uses the example: “Every girl is in love with one boy. Therefore, there is one boy every girl is in love with.” (Geréby, “Theistic Fallacies,” 9)

[14] Ibid., 10.

[15] Ibid., 11.

[16] “Albert recognized that this problem is related to the so-called cognosco venientem paradox (“I know the person coming,” also known as “the Hooded man”), an old problem of meaning and reference of the Megarian Eubulides, which can be found in Aristotle’s Sophistical refutations [Aristotle, Soph. El. 179a26-179b33 and Lucian, Vit. Auct. 22-23].” (Geréby, “Theistic Fallacies,” 11)

[17] Ibid.

[18] Geréby, “Theistic Fallacies,” 12.

[19] Ibid.

[20] As Geréby points out: “sharing certain properties does not imply that these shared properties are core properties. That certain things happen to fly does not necessarily imply that flying is their core property. One flying thing might be a bird, the other an aeroplane while the third a bat, or a plastic bag blown away by the wind.” (Geréby, “Theistic Fallacies,” 15)

[21] Geréby uses the following example: “if people share the property of having legs, hands and head, this fact does not justify the assumption that that there is a real human being who shares only these properties. A real human being must have infinitely more real properties, then this restricted number of abstract forms. Since if these abstract properties would be instantiated, would be real legs, hands and heads, they would be of an individual person, who has many more additional properties, too. But if these properties become instantiated in a concrete form, that will be a real human being. But then it would not be general enough for the ‘core doctrine.’” (Geréby, “Theistic Fallacies,” 16)

[22] “A mental object is not necessarily an empty term, but it functions without identifiable external referents. Mental objects are not necessarily like the “fountain of youth” or the “yeti,” or other entirely fictitious objects. It is simply indefinite as to the existence of its external referent or extension, like in the case of the Hooded man. The important character trait of mental objects is their being regulated by meaning or content rules.” (Geréby, “Theistic Fallacies,” 15)

[23] “The meaning of the term “god” cannot be established independently from a religion, that is, independently of the context of rules and instructions in which it is used. In fact “god” is defined by a particular body of textual, liturgical and legal and aesthetic use bestowed by a tradition.” (Geréby, “Theistic Fallacies,” 17)

[24] Geréby, “Theistic Fallacies,” 16.

[25] “The assumption of theism is, to use a mathematical analogy that religions are like numbers that have a greatest common denominator. If the above analysis is correct, then it is rather the case that religions do not have a greatest common denominator, but rather relate to each other as relative primes.” (Geréby, “Theistic Fallacies,” 17)

[26] To illustrate this point further Geréby defines “religion X as the conjunct of the following members: X = {p & q & r & s} and the religion Y as Y = {p & q & t & u}, where the letters stand for propositions of relevant dogmatic content. As it can be seen, p and q are common elements to both X and Y. (Their intersection is not empty.)” He then argues that this “shared content” would not be sufficient for establishing a rational core: “Can it be said that a conjunct of {p & q} would be a part of both X or Y? Unfortunately, such a “shared content” would not be sufficient to establish a common ground, a ‘core doctrine.’” (Geréby, “Theistic Fallacies,” 22)

[27] Orthodoxy has a conception of natural theology, which I would argue is the correct sense of natural theology – to see and know God through His creation, not through problematic or erroneous epistemic theories or philosophies, or even through syllogistic inference. Orthodoxy holds that God is imminently present in the world through His divine energies (“in Him we move and have our being”). There is not as Staniloae explains a “separation between natural and supernatural revelation.” He states: “Natural revelation is known and understood fully in the light of supernatural revelation, or we might say that natural revelation is given and maintained by God continuously through his own divine act which is above nature. This is why Saint Maximos the Confessor does not posit an essential distinction between natural revelation and the supernatural or biblical one [as we see with Aquinas and other western theologians].”