Deconstructing Aquinas’ Fifth Way: The First Fallacies

Deconstructing Aquinas’ Fifth Way: The First Fallacies January 6, 2019

Ficino (commenter here at ATP with a great interest in Thomas Aquinas) has kindly produced a critique of Aquinas’ Fifth Way, as famously exhibited in Summa Theologiae. This will be split into four parts. This is the third part, the first part (on the preliminaries) can be found here, the second, on formulating the argument and initial issues, can be found here.

Let me set out Ficino’s previous formulation of Aquinas’ argument:

  1. There exists an x such that x lacks cognition and x operates for the sake of an end.
  2. For all x, if x lacks cognition, x is a natural body.
  3. For all x, if x is a natural body and x operates for the sake of an end, x operates by intentionality.
  4. For all x, if x operates by intentionality, x is directed by an intelligence. [I don’t think this premise has to retain “if x is a natural body,” since all cognizant things also are directed by intelligence]
  5. Therefore, for all x, if x is a natural thing [sc. non-cognizant], then x is directed by an intelligence.
  6. Therefore, there exists some y, such that, for all x, if x is a natural thing, x is directed by y, and y is intelligent.

Now we look further at how the argument is logically unsound. The last post ended with the question: Will the Fifth Way go through? Back over to Ficino:

Not yet. If we grant Aquinas a universal major premise to repair Fallacy #1, he falls even sooner into Fallacy #2 (see previous post). His Aristotelian logic presupposes that classes of which universals are affirmed have members when we are not told explicitly that they do. In Boolean logic, universal propositions make no claims about existence. For us to retain Aristotelian logic overturns the advances of the last century and a half, and we have no independent reason to revert to Aristotelian logic (see here). Aquinas could introduce an evidential claim, that there exist intelligent directors of inanimate things’ natural operations. But then, he would be asserting that which he is trying to prove. A defender might say that an evidential claim is contained in the argument’s first step, “we see that some things lacking cognition operate for the sake of an end,” so that we can infer that there exist members of the class of inanimate things whose natural inclinations are guided by an intelligence. But as I shall argue in 4), the Fifth fails to demonstrate that “operates always or frequently in the same way” entails “is directed by an intelligence.” Faults of content in the Fifth Way render its conclusion unsound.

4) The Fifth Way is Unsound

Even if faulty logical relations among premises could be repaired, informal fallacies of content and, as it looks, faulty evidence vitiate the Fifth. As far as I can see, the Fifth contains a) a bifurcation fallacy and b) question-begging steps. Moreover, there is c) counterevidence against its claim that non-cognizant things operate always or for the most part so as to attain that which is best.

a) Bifurcation fallacy/false dilemma.

It is not clear that “by coincidence” (a casu) and “from intention” are exhaustive alternatives. Can we not have a kind of natural necessity that does not aim at a definite end? Natural selection in evolution is not random, but we cannot say that it is “for the sake of” (propter) particular species. As a famous analogy puts it, the pothole does not develop for the sake of the puddle that will fit it. Or in systems of things in nature, it is not clear that the behavior of the constituents aims at an end when the system is open-ended—e.g. the behavior of subatomic particles, or of the earth’s plates in plate tectonics. Even Aristotle had concluded that some products of nature are necessary but do not have proper final causes, such as some residues or excretions (On Parts of Animals 2 676b15‒19) or features like eye colors or curly vs. straight hair, which arise by necessity but are not part of the organism’s essential form (On the Generation of Animals V.1 778a31‒b6, V.3 782b19-24).

b) Begging the Question.

In a question-begging argument, the conclusion may be true, but we are under no logical obligation to believe it. It is generally held that an argument begs the question, first, when its conclusion is contained in one of its premises, so that the conclusion merely repeats the premise, making the argument an assertion not a demonstration. Second, an argument can beg the question when it uses a suspect principle as grounds for drawing a conclusion, without giving grounds for that principle’s truth. In a twist on the second type of question-begging, the principle may be even more suspect than the conclusion for which it is offered as grounds. (Good on ways of begging the question is Morris Engel, “Understanding, Finally, What it is to ‘Beg the Question’” Metaphilosophy 22.3 (1991) 251-264.) I identify three steps in the Fifth where work is done by a question-begging move: i) “attain that which is best;” ii) “not unless directed by something intelligent;” iii) a single “God.”

i). As we saw, Aquinas bases his conclusion, viz. non-cognizant natural bodies operate for the sake of an end, on their tendency to act regularly “so as to attain that which is best.” What is this “best”? This phrase in Aristotelian-Thomistic thinking is code for final cause: “the end (telos) should not be just any last thing, but the best” (Phys. I.2 194a 32–33, tr. Charlton); “the end for the sake of which each has been composed or has come to be has taken the place of the nobly beautiful” (καλόν, PA I.5 645a23‒25, tr. Reeve); nature crafts generation through eggs by necessity and for the sake of what is better, balancing quantity against size etc. (GA III.4 755 a21‒35), and nature always makes the best of the things that are possible, On the Heaven II.5 288a3‒4. But then, the first step of the Fifth resolves to “natural bodies operate for the sake of an end because they operate for the sake of an end.” The conclusion of the first step is contained in the premise that purports to support it. The steps that follow, then, being deduced from a question-begging premise, are likewise vitiated by that fallacy.

ii). What I call the Intelligence Requirement is asserted, not argued for. The IR is asserted in the step, If the operation is ex intentione, then it’s directed by intelligence. I have not seen this claim actually demonstrated in Aquinas’ various teleological arguments. I have only seen Aquinas rely on the example of an arrow that exhibits finality of movement toward a target only if shot by an archer. That is a false analogy, however, because Aquinas needs to prove that the natural inclinations, or directedness, of beings in nature are directed by an intelligent director. But the arrow qua arrow is an artefact, not a natural thing. Second, insofar as its matter makes it a natural thing, its wood and metal or stone make it heavy. In Aristotelianism, heavy bodies have a natural inclination toward the center of the universe, the earth. The arrow’s movement toward a target is not its natural movement but is imposed, as Aristotle would say, “by violence.” Faulty analogy. So far, not so good.

Edward Feser argues for the IR with what I call his “Where is the final cause?” argument, which he repeats in several writings. Feser notes that when a thing begins to move or develop toward an end, the end or final cause does not exist in the thing. But the final cause qua cause, “to have any efficacy, must in some sense exist,” on the principle that a cause must exist prior to an effect. Since the final cause does not exist already in the developing thing, it must exist somewhere else, and Feser eliminates all candidates except intellect. The final cause of the artefact exists in the artisan’s intellect; the final cause of natural things exists in God’s mind (cf. e.g. Aquinas [Oxford: Oneworld Publ. 2009] 116‒118).

Feser joins other Thomist commentators in trying in this way to defend Aquinas’ seemingly tacit assumption that final causality in nature must be the product of a mind. This defense, however, begs the question. That’s because it rests Aquinas’ conclusion on the even more suspect claim that the end “must in some sense exist” already in order to be the cause of the efficient cause. Aquinas speaks of a cause’s existing naturaliter prior to the effect, but this is consistent with his insistence elsewhere that a final cause is prior according to reason or explanation, not in reality.

Cf. “Sed quia omnis causa, inquantum est causa, naturaliter prior est causato … Dicitur enim aliquid prius altero generatione et tempore, et iterum in substantia et complemento… imperfectum est prius perfecto, secundum generationem et tempus, sed perfectum est prius in complemento,” De principiis naturae 4.31.

 “But because every cause, insofar as it is a cause, naturally is prior to the thing that is caused … something is said to be prior to something else in generation and time, and again, in substance and a state of completion … the imperfect is prior to the perfected/completed, according to generation and time, but the perfected is prior in its state of completion,” De principiis naturae 4.31.

This claim ignores the fact that in Aristotle and Aquinas, “nothing prohibits something from existing prior and posterior to something else according to different reasons/explanations (rationes): for the end is prior according to reason (rationem), but posterior in existence (esse); the agent however is the converse” (Aquinas, Comm. in II Phys. l. 5 C182). Aquinas insisted on the same in his own name: “just as the end is prior in intention, so it is posterior in existence” (Comm. on Sentence of Peter Lombard III. d. 23 q. 2 a. 5 ad 4; same at On Truth 22.12 c, On Power 5.5 ad 3, On Evil 2.3 c). And note the converse: “the efficient cause cannot be posterior in existence” (ST 3a 62.6 c).

Feser often argues from the example of an acorn, in which the oak does not exist, to the claim that the oak as final cause must exist somewhere else—in God’s mind (e.g. “Teleology. A Shopper’s Guide,” Philosophia Christi 12.1 (2010) 142-159 at 157, repr. in Neo-Scholastic Essays, St. Augustine’s Press 2015. For much of the content of this article outside a paywall, see the following blog post, with links to earlier posts: “Final Causality and Aristotle’s Unmoved Mover,”). But the end is in the acorn potentially. Seed has soul and is soul potentially (Arist. GA II.1 735a8-9, II.3 737a17-18), and the seed and the fruit are this body potentially (On the Soul II.1, 412b26; “the seed is [this body] potentially,” PA I.1 641b36). Thus, the end, the mature organism, is posterior in existence but prior in explanation; pneuma or soul-heat, the agent of generation, is prior in existence in seed but posterior according to explanation. One can say that the end is ontologically prior in that it is prior in account, and the account defines the substance, cf. Mete. IV.12 390a6 οὐσία οὐδὲν ἄλλο ἢ λόγος, cf. a19, b18 we can say on account of what and what a thing is if we have either the matter or the account.

Feser is right to join mainstream Aristotle scholars in denying that Aristotle believed in an intelligent governor of nature. Feser is right that Aquinas departed from Aristotle on this point. Feser and other modern Thomists, however, have not shown that A-T principles allow them to carve out a way for the end in nature to exist prior to the agent, in the face of Aquinas’ explicit denial that this is so. To insist that the final cause must exist prior to the effect is, I think, an inference from Aquinas’ insistence that “the end is called the cause of the efficient cause, since the efficient cause does not operate except by the intention of the end. Hence the efficient cause is the cause of that which is the end … However, the efficient cause does not cause the end to be the end” (On Principles of Nature 4.28). But the “where is the final cause?” argument treats the final cause in nature as something that makes the effect happen through the instrumentality of the agent. That is to confuse final with efficient and formal cause. Aquinas makes clear that the formal cause exists potentially already in the matter, De Potentia 3.7 c, and to the formal cause pertain a substance’s eventual powers. The final cause explains; it does not make. In fact, final causality became problematized in late Scholasticism precisely because emphasis on the efficient nature of final causality seemed to leave final cause without proper explanatory power (cf. Monte R. Johnson, Aristotle on Teleology [Oxford 2005] 23). I conclude that the “where is the final cause?” argument begs the question as an instance of the “twist” sketched in b) above.

The only other defense I have seen of the Intelligence Requirement is a kind of argumentum ad incredulitatem. An example is by Fr. Leo J. Elders: “ … in order to attain precisely this end the factors involved, which dispose to such an activity, must have been made or arranged in such a way that the effect to be attained was taken into prior account. Since it is the intellect which compares and connects things, this particular disposition and arrangement must be the work of an intellect” (The Philosophical Theology of St. Thomas Aquinas [Leiden and New York 1990] 123). Since we normally suppose that a “disposition” is the work of an intellect, Elders’ description of natural phenomena as “dispositions” begs the question in the first way sketched in b) above.

Aquinas continually speaks of God as the artisan of nature; cf. e.g. De Potentia 3.7 c. In his commentary on Aristotle’s Physics 199b26ff, Aquinas actually adds “divine,” not present in Aristotle’s text: “Hence it is clear that nature is nothing but a certain kind of art, i.e. the divine art, impressed upon things, by which these things are moved to a determinate end,” Comm. in Phys. II.8 l. 14, C268. I get the sense that Aquinas’ faith that God is the artisan of nature leads him to consider unthinkable any teleology that does not end up at a mind.

iii) The Fifth lacks steps to establish that the intelligent y (see 2) previously), even if it exists, is numerically one. A defender may reply that Aquinas is not guilty of a misstep here, since “something” need not denote numerically one thing. It is true that Aquinas, like Aristotle, often reduces a plurality of causes to one cause (e.g. ST 1a 11.3 c, SCG II.30.12), but it is not obvious that Aquinas’ pileup of four singular Latin terms in the Fifth Way can cash out as a plurality. In any case, the Fifth Way is consistent with polytheism. Why not allow different divine governors over different spheres of nature? Edward Feser allows that the Fifth Way does not establish that there is but one governing intelligence. Feser appeals to cosmological arguments for this (cf. Aquinas 118‒119), but it is not obvious that the Unmoved Mover/Necessary Being is identical with the intelligent governor of all natural entities. Aristotle’s Unmoved Mover, one recalls, thinks itself, not particulars, and when final causality does not result from “intelligent choice,” it results “from nature” (Phys. II.4 196b22). Or the governor might be one or more lesser beings like the Demiurge and lower gods of Plato’s Timaeus.

Even Aquinas’ arguments for a single governor of creation in Summa Contra Gentiles I.13.35 and Summa Theologiae 1a 11.3 beg the question, for from the premise that all things in nature operate as part of a single order, it does not follow that a single governor governs that order. Again, a plurality of (divine) governors could govern it in harmony. To pull a single governor out of the Fifth, the Thomist defender has to turn to arguments that the first cause is Pure Act, or that contingent things’ existence must be conferred by a necessary being. But one of the jobs we usually think Aquinas gave the Fifth was to demonstrate that the Unmoved Mover/First Cause/Necessary Being is also the intelligent governor of the world. It’s not obvious that all of nature constitutes one order, or that if it’s a natural order, it is directed by a mind, etc. That there should be one director appears to be an assumption even in Aristotle; cf. Arist. PA III.4 665b14-15 “wherever possible, one [ἀρχή] is better than many.” In my view, the Fifth fails to do that job.

iv) Equivocation and Amphiboly? A fallacy of equivocation occurs when a term bears different senses in different steps in an argument. A fallacy of amphiboly arises from a grammatical construction that can be construed in more than one way. Aquinas is often criticized for insufficient rigor in his use of terms in arguments, and we find instances of this fault in the Fifth. One, the sense of “natural things,” res naturales, in the argument’s conclusion is unclear. In some places in Aquinas, that term cashes out as the equivalent of “natural bodies” composed of a single dominant one of the four ancient elements (e.g. De Veritate 2). But elsewhere, res naturales extends to all orders of things in nature, from minerals up through plants and animals (ST 1a 108.3 c) even to include humans (SCG II.83.25, IV.84.5, ST 3a 5.2 c). Since, however, Aquinas’ wording leaves it unclear how widely he intends res naturales in his conclusion to extend, I do not conclude that we have a fallacy of equivocation. Similarly, the function of the Latin ut, which I translated as “so as to,” is ambiguous. That particle with a verb in the subjunctive (consequantur) can signal either purpose (“so that they may attain”) or result (“with the result that they in fact attain”). I don’t think this ambiguity affects the whole argument, however, since the argument’s flaws are flaws on either construal.


Ficino will continue to look at the Fifth Way in wrapping up matters in the fourth and final piece.

Stay in touch! Like A Tippling Philosopher on Facebook:

Please also donate to my MS treatment appeal (I have just been diagnosed with Primary Progressive Multiple Sclerosis) so I can keep doing what I do!

"I think philosophizing whether group X deserves equal treatment is inherently bigoted."

Freedom of Speech and Offence: The ..."
"Having enough to not die immediately is not enough to live a good life. Poor ..."

Skewed Data from Massive Inequality in ..."
"Related: There is a huge difference in life expectancy between rich people and poor ..."

Skewed Data from Massive Inequality in ..."
"Speaking of knife crime, did you know that in the U.S. murders by knife/cutting instruments ..."

Making Moral Claims Without Doing the ..."

Browse Our Archives

Follow Us!

What Are Your Thoughts?leave a comment