Tennant on Aquinas’s Second Way

Every now and then I point out examples of professional academic philosophers badly misinterpreting some traditional argument for God’s existence, or mind-body dualism, or a natural law conclusion in ethics, or some other idea that is out of step with the naturalistic zeitgeist.  This is, I think, a useful exercise, because it shows how the conventional wisdom on these matters does not rest on actual knowledge or understanding of the ideas criticized.  It is instead held in place by errors and unexamined prejudices that mainstream academics who do not specialize in these matters tend to repeat back to one another and pass on to each new generation of graduate students, who then enter the profession and also repeat them back to each other and to their own students.  This has, over time, created a general assumption that “everyone knows” that the arguments are no good and not worth investigating further, even though a little effort would show that what “everyone knows” is demonstrably false.

Here’s another example, from Neil Tennant’s book Introducing Philosophy: God, Mind, World, and Logic, which contains a very poor treatment of Aquinas’s Second Way.  I don’t mean to bash Tennant, whose book is otherwise very interesting and who is a prominent and serious philosopher.  He is in no way condescending toward Aquinas’s argument, but simply rehearsing stock objections that he honestly takes to be compelling.  But that’s precisely the point.  The fact that even a thinker of Tennant’s stature can make the sorts of mistakes he does reinforces the point that the confidence with which mainstream academics dismiss such arguments is massively out of proportion to their actual understanding of them.

Here’s how Tennant glosses the Second Way:

[Aquinas] thinks that at the point ‘in’ time ‘at which’ the physical universe came into being, there was an act of creation – an event – involving the Deity… From this alleged initial event – the ‘prime moving’ – flow all causal chains through subsequent time.  Take any of these causal chains.  According to Aquinas, if we follow it back through time, we shall (in an atemporal sense) eventually arrive at a terminus in the past: the act of creation itself.  For Aquinas, there can be no back-tracking infinitely far, without end, into the past, along a causal chain of events. (p. 227)

Readers of my book on Aquinas and longtime readers of this blog will be groaning already.  But let me explain what is wrong with this summary in the course of addressing the four objections Tennant raises against Aquinas – all of which rest on misunderstandings that are obvious to anyone familiar with Aquinas’s general metaphysical views. 

First objection

Tennant’s first objection is directed against these lines from the Second Way:

But the series of efficient causes cannot possibly go back to infinity.  In all such series of causes, a first thing causes one or more intermediaries, and the intermediaries cause the last thing; when a cause is taken out of this series, so is the effect.

Tennant’s response to this is to suggest that Aquinas fails to consider the possibility of “causal over-determination” (p. 227).  He gives the example of an outlaw shot with deadly accuracy by two bounty hunters at the same time.  Even if the first shot to reach him hadn’t occurred, the second shot would still have killed him.  Hence, Tennant concludes, Aquinas is mistaken to suppose that the removal of a cause entails that the effect will not occur.

The problem with this objection is that it overlooks Aquinas’s distinction between causal series ordered per se and causal series ordered per accidens (sometimes described by later writers as the distinction between hierarchical and linear series of causes), and wrongly assumes that it is the latter sort of series that Aquinas has in mind in the Second Way.  A linear or per accidens causal series characteristically extends over time and is made up of members each of which has independent or built-in causal power.  One of the examples Aquinas gives is that of a father who begets a son who in turn begets another (Summa Theologiae I.46.2).  If the first member dies, the series can still carry on, because the son retains power to beget a son of his own whether or not his own father is still in the picture. 

Indeed, in many per accidens or linear series of causes, it is not even necessary that some particular preceding cause have been the one that produced the later effect.  Aquinas gives, in the passage just cited, the example of “an artificer [who] acts by means of many hammers accidentally, because one after the other may be broken.  It is accidental, therefore, that one particular hammer acts after the action of another.”  That the artificer used some particular hammer to produce the effect is “accidental” or non-essential to the continuance of the series, because some other hammer could have done just as well.  In other words, Aquinas himself makes more or less the same point Tennant does: In causal series ordered per accidens, the same effect could arise via alternative causal pathways.

The reason this acknowledgement is not fatal to the Second Way is that that argument is not talking about causal series ordered per accidens in the first place, but rather about causal series ordered per se.  Now, in a per se or hierarchical series of causes, the members are typically acting simultaneously rather than over time, and that is because the members other than the first have only derivative or borrowed causal power rather than built-in causal power.  Aquinas’s example in the passage from the Summa just cited is that of a stone which is moved by a stick which is moved by a hand.  The causes and effects in this series are simultaneous.  The stone is being moved by the stick at the same moment that the stick is in turn moved by the hand.  More to the point, the stick moves the stone only insofar as it is itself being moved by the hand, for the stick has no independent or built-in causal power.  It has causal power only qua instrument of the hand.

Now, the difference between the “later” causes in a per se or hierarchical series and the “first” cause of the series has nothing essentially to do with order in time and nothing essentially to do with numerical order either.  Rather, it is the difference between that which has only derivative, borrowed, or secondary causal power and that which has intrinsic, built-in, or primary causal power.  It is the difference between an instrument and that which acts through the instrument.  When, in various places, Aquinas rules out an infinite regress of causes ordered per se or hierarchically, what he is talking about is the impossibility of there being a series of causes having merely instrumental or derivative causal power without there also being something with built-in causal power which acts through the instrumental causes.  For example, he writes:

That which moves as an instrumental cause cannot move unless there be a principal moving cause.  But, if we proceed to infinity among movers and things moved, all movers will be as instrumental causes, because they will be moved movers and there will be nothing as a principal mover.  Therefore, nothing will be moved.  (Summa Contra Gentiles I.13.15)

Similarly, elsewhere he says:

For everything that is moved by another is a sort of instrument of the first mover.  Therefore, if a first mover is lacking, all things that move will be instruments.  But if the series of movers and things moved is infinite, there can be no first mover.  In such a case, these infinitely many movers and things moved will all be instruments.  But even the unlearned perceive how ridiculous it is to suppose that instruments are moved, unless they are set in motion by some principal agent.  This would be like fancying that, when a chest or a bed is being built, the saw or the hatchet performs its functions without the carpenter. (Compendium Theologiae I.3)

Now, this is what Aquinas is talking about in the lines from the Second Way that Tennant is responding to.  When he says there that if you remove the first cause or the intermediaries, the effect will not follow, he is not denying that in a temporal regress of causes ordered per accidens, some other cause might have produced the same effect.  He isn’t even talking about that sort of series at all.  Rather, he is saying that in the kind of series where every member other than the first acts merely as an instrument, the effect will not follow if the first cause or intermediate instruments are removed.  For example, the stone won’t be moved by the stick if the hand isn’t using the stick to move it. 

(Of course, some other stick or some other person could have moved the stone, but that’s irrelevant to Aquinas’s point.  His point is that, whichever stick and whoever’s hand we’re talking about, when someone uses a stick to move a stone, the stick acts only instrumentally rather than with built-in causal power.  Hence, if you remove anything which might act through the stick in order to push the stone, the stick won’t push it.)

Second objection

Tennant next claims that “Aquinas’s central inference has the form: Every event has a (distinct) cause; therefore, Some event caused all (other) events” (p. 228).  He then objects that Aquinas is here guilty of a quantifier-switch fallacy.  (This is the sort of fallacy committed, for example, by someone who reasons that if every reader of this blog is reading it on a computer, it follows that there is some one computer they are all reading it on.)

Now, Tennant does not pretend to be quoting Aquinas here – which, of course, he couldn’t be, since Aquinas never actually explicitly says what Tennant attributes to him.  But neither does Aquinas implicitly say any such thing, and once again, Tennant is misled because he is evidently unfamiliar with Aquinas’s metaphysics of causality, which forms the crucial background context for the Five Ways.  For again, when Aquinas reasons to a first cause, what he is doing is pointing out that something can’t serve as an instrument unless there is some non-instrumental cause acting through it.  Fallacious quantifier-switch style reasoning has nothing at all to do with that.

Now, someone might claim that Aquinas must still be committing such a fallacy in supposing that there is only one such non-instrumental or primary cause operating through all the instrumental or secondary causes.  But to raise such an objection would show only that one hasn’t actually read Aquinas beyond the brief excerpt from the Summa that contains the Five Ways.  For that there is only one ultimate first cause is not something Aquinas is claiming in the first place to have established in the Second Way.  He addresses that question later, in Summa Theologiae I.11.3 (and says a lot about it elsewhere too). 

Third objection

Even more egregious is this remark from Tennant:

[Aquinas] dogmatically rules out the perfectly consistent and imaginable scenario of a class of events extending infinitely far back into the past, without there being any temporally ‘first’ point.  Each event could be caused by a strictly earlier event, while yet no event is initial within the temporal ordering. (pp. 228-9)

In his third objection, Tennant rejects the suggestion that Big Bang cosmology might repair this alleged oversight of Aquinas’s.

Now, what I said in response to Tennant’s first objection already indicates what is wrong with this one, but let’s make the point more explicit.  Tennant is assuming, as so many unwary readers do, that Aquinas is, in the Five Ways, trying to establish that the universe had a temporal beginning and that God was the cause of that beginning.  But nothing could be further from the truth.  For one thing, and to repeat, when Aquinas rules out an infinite regress of causes, he is not talking about temporal regresses, which would involve per accidens or linear causal series.  He is talking about per se or hierarchical causal series the members of which are all operating simultaneously, here and now.  He is saying that there must be a primary or non-instrumental cause operating here and now through the secondary or instrumental causes that are operating here and now.  Whatever one thinks of his position, the issue of whether the universe had a beginning is completely irrelevant to it.

For another thing, Aquinas explicitly rejects in several places any suggestion that it can be proved philosophically that the world had a beginning.  He thinks that that is something we can know only via special divine revelation rather than through natural reason, and thus he consistently avoids getting into the issue when he argues for God’s existence.  For again, to reason causally back to a temporal beginning would be to reason about causal series ordered per accidens or “accidentally,” and when arguing for the existence of God, Aquinas does not appeal to that sort of series.  Indeed, he agrees with Tennant that we can’t prove that such series have a beginning.  Instead, he appeals to per se or hierarchical causal series.

Hence, later in the Summa itself, Aquinas insists that “by no demonstration can it be proved, that the world did not always exist,” because “it is not impossible to proceed to infinity ‘accidentally’ as regards efficient causes”; for example, “it is not impossible for a man to be generated by man to infinity” (Summa Theologiae I.46.2).  He makes the same point in other places, and indeed devoted one of his shorter works to arguing against those who claimed that it could be proved philosophically that the world had a beginning.  In other words, Aquinas firmly, explicitly, and repeatedly acknowledges what Tennant accuses him of “dogmatically” denying!

Fourth objection

Tennant’s final objection is to suggest that Aquinas’s argument would fail even if he were to establish a first efficient cause.  For “why should that be taken, without further ado, to be God, or even simply to be God's doing?  Why shouldn't it simply have happened, in a Godless universe?” (p. 229).

The problem with this objection is that Tennant is simply mistaken in assuming that Aquinas draws this conclusion “without further ado.”  On the contrary, there is much ado about it, beginning precisely in the pages that immediately follow the presentation of the Five Ways in the Summa.  There Aquinas argues at length that a first cause would, on analysis, have to have the divine attributes of simplicity, goodness, perfection, infinity, omnipresence, immutability, eternity, unity, knowledge, power, will, love, and so on.  Indeed, he devotes what in a modern edition comes to about 250 pages of dense argumentation to the topic just in the Summa Theologiae, and he says even more in yet other works.  In short, Aquinas does not simply assume “without further ado” that a first cause would have the attributes definitive of God.  Quite the opposite.

Again, I don’t mean to be too hard on Tennant, specifically.  There is nothing unique about his objections.  On the contrary, variations on them are constantly raised against Aquinas by mainstream academic philosophers and by mainstream academics and intellectuals from other fields (not to mention countless amateurs).  And yet they are all demonstrably based on egregious errors and misunderstandings.  Which, while it tells you nothing about Aquinas, says much about what you should think of mainstream academic and intellectual opinion. 

Related posts:

Clarke on the stock caricature of First Cause arguments

The straw man that will not die

An exchange with Keith Parsons, Part III

This is philosophy?

Warburton on the First Cause argument

The less Rey knows, the less he knows it

Straw men and terracotta armies

So you think you understand the cosmological argument?