An obvious reaction one might have on passing from the First Way to the Second Way, is that the First Way is a particularly clear or paradigmatic case of the same phenomenon investigated in the Second Way; or, if you prefer, that the Second Way is a generalisation of the First. St Thomas’s own words make it clear that he regards the First Way as especially clear and obvious, presumably because the phenomenon from which it starts, that of the processes of change in the world, is particularly clear and obvious. (This claim perhaps helps to cast some doubt on Kenny’s reading of the First Way, which makes it depend on a quite high level of training in Aristotelian physics, not to say Ptolemaic astronomy. Even among well-educated people in the thirteenth century it would surely have been stretching things a little to call an argument which depends on detailed knowledge of physics or astronomy one which is clearer and more obvious.) Be that as it may, there is a good case for making out that the Second Way is a generalisation of the First. The First Way has to do with process of change, the Second Way has to do with efficient causality in general. In St Thomas’s book, not every example of efficient causality is an example of a process of change, or even of the initiation of a process of change. A beginning of existence, for example, which certainly requires an efficient cause, is not, strictly speaking, a change in that which begins to exist. As we commented in the last chapter, for the analysis of change one of the most important points to establish is the subject of change, and that which begins to exist is not strictly speaking a subject of change, a subject which undergoes some modification. Before it begins to exist, it isn’t there to undergo any modification at all. A beginning of existence would thus fall under the Second Way, but not under the First Way.
So far so good: but unfortunately one could make out just as good a case that the First Way is in some sense a generalisation of the Second Way. St Thomas would hold that not every process of change is to be explained in terms of efficient causality, or, at least, he would claim that not every process of change is to be explained in terms of efficient causality alone. The most famous example in St Thomas’s thought is one which is unlikely to appeal to us as scientifically serious: St Thomas thought that the rotation of at least the outermost, and perhaps of others, of the heavenly spheres was to be explained in terms of final causality. These crystalline spheres of Aristotle’s astronomy, which St Thomas seems to have taken perfectly seriously, are intelligent, or are at least guided by intelligences. They know what God is like, and are moved by love of God to try to be like him. God is always active and never changing: hence the ceaseless and invariable circular movements of the heavens, which represent the nearest approach a material being can make to being both always active and never changing. The picture is a nice one, though to our way of thinking it is also bizarre: but what is important is that these spheres move as they do in order to be like God. Nothing pushes them along except desire, and that desire is based on knowledge of the good. That process in particular is initiated and continued by one particular type of final causality, and not by efficient causality at all. So there will be examples of changes which illustrate the First Way which do not illustrate the Second.
If the example just given is too alien to be of any help to the reader, then one can perhaps remember the dog that starts barking in the night-time. The dog starts barking, perhaps, in order to scare away the intruder, or in order to warn the household, or in order to affirm its own territory, or in order to achieve a reward or avoid a punishment from its master. Any one of these explanations will do: which one you choose will depend on what you think is within the limits of canine intelligence. But all of the explanations are in terms of final causality.
This would seem to be as good a point as any to outline the Aristotelian doctrine of the four modes of explanation or causality, to which reference has already been made. Aristotle, and, following him, St Thomas, recognised four different modes of causality or explanation. "Efficient causality", or explanation in the efficient mode, is the closest to what we nowadays generally mean by "causality": an explanation in terms of how a thing came about. "Material causality", or explanation in the material mode, is an explanation in terms of what a thing is made of. "Formal causality", or explanation in the formal mode, is an explanation in terms of in virtue of what the stuff that a thing is made of is the thing that it is. Lastly, "final causality", explanation in the final mode, is an explanation in terms of what a thing is for. We have already seen that for Aristotle and St Thomas "science" is "definite knowledge through explanations", and we have seen them using both efficient and final causality in proving the existence of a cause from the effect. Science may be structured in terms of any of the four different modes of explanation. It is therefore worth being sure that we know what we mean by them. A good example, one given by Aristotle, is that of a faggot, a bundle of sticks tied together in order to be carried more easily. The example is one of an artefact, and indeed most of the straightforward examples are of artefacts. This is because the question of the explanation of natural objects in the final mode is highly disputable nowadays. We will examine it when we come to look at the Fifth Way, but I do not wish to prejudge here the discussion which I will be carrying out there, about the possibility and legitimacy of looking for final modes of explanation of natural phenomena.
If we are looking for an explanation of the faggot, then, the first thing we have to say is what it’s made of, its matter: we have to give an explanation in the material mode. The answer is clear: sticks and string. If instead of sticks and string it consisted of flowers and ribbon, it would not be a faggot but a bouquet. But the same matter, the same sticks and string could be scattered half-way across the forest, or put together in a different way to make a rudimentary birdcage. What makes the sticks and string to be a faggot and not e.g., a heap of sticks with a piece of string on top, or a rudimentary birdcage, is the way they are tied together. This is the form of the faggot, the explanation of what the faggot is in the formal mode, what distinguishes it from other things which could be made of the same matter. The explanation in the efficient mode is clear and simple: it is yonder peasant, gathering winter fuel. Yonder peasant is the one who effects or brings about the existence of the faggot. The explanation in the final mode is also clear: for ease of transportation. That is what faggots are for, the end of a faggot: if you live a good league hence, it’s the devil’s own job to carry a heap of sticks through the snow if you haven’t tied them up in a faggot.
Possibly the notion of explanation in the formal mode is the most obscure of the four. Aristotle gives other examples, in the same context. The difference between a lintel and a threshold is not one of their matter, he would say: they are both beams of wood wide enough to stretch between the uprights of the doorway. One could be substituted for the other and no harm would be done: the former threshold would have become the lintel and the former lintel would have become the threshold. The difference is in their position: it is thus the position at the top of the doorway which is the form of the lintel, what makes this matter, this beam of wood to be a lintel, which constitutes the formal explanation of what it is to be a lintel. Equally, it is the position at the foot of the doorway which is the explanation in the formal mode of what a threshold is. Or again: the difference between breakfast and dinner is not one of matter but of time: the form or formal explanation of breakfast, what makes breakfast breakfast as opposed to dinner, is the time of day; and likewise for dinner. We should be able to accept this even though as a matter of fact we don’t normally have cornflakes for dinner or soup at breakfast. For us there may well be a culturally determined, but purely accidental, difference of matter between breakfast and dinner. The ancient Greeks, less dietetically fortunate than we, had a breakfast that consisted of a sort of porridge, and a dinner that consisted of a sort of porridge. If they were better off they might have had two courses at dinner, the first a sort of porridge and the second a sort of porridge. Plato went on record as saying that no-one could do philosophy in a place as rich as Sicily where people ate three times a day.
Leaving aside the gastronomical questions, though, the different modes of explanation should be fairly clear. Since a science progresses towards explanations in one or another of these modes, we could expect that the Five Ways could be classified according to this system. But it does not seem so easy. The Second Way is explicitly said to be based on efficient causality, and the word "finis", end, that in terms of which explanation in the final mode is given, is also explicitly used in the Fifth Way. But the First Way, as we have just seen, may relate to any kind of explanation which may be given of processes of change, whether efficient or final. Kenny claims that the Third Way relates to explanation in the material mode, and it is certainly true, as we shall see, that one of the "derivatively everlasting" things which the Third Way speaks of is matter. But so are the forms of e.g. animal species: so the Third Way may relate as much to the formal mode of explanation as it does to the material mode. As for the Fourth Way, it is not the easiest task on earth to identify any conceivable or intelligible mode of explanation to which it relates. I have seen it claimed that the Fourth Way relates to "transcendental causality", which I sometimes suspect to be an academically more respectable way of saying "it is not the easiest task on earth to identify any conceivable or intelligible mode of explanation to which it relates".
Thus while one might wish to claim that the First Way relates to final causality at least as much as it does to efficient causality, there is no doubt that the explanation we gave of the First Way in the last chapter in fact related to efficient causality alone. This is partly because final causality is both a difficult and an unfamiliar notion, one which arouses fairly justifiable suspicions in the contemporary reader; and partly because to give an adequate account of final causality would have dragged out the chapter too much, and it seems better to leave that account until we come to deal with the fifth way, where it cannot be avoided. But also, in part, the reason was that I could not see how to give an adequate account of the initiation of processes of change in terms of explanation in the final mode.
This had the perhaps unfortunate consequence that many of the considerations adduced in the account of the First Way have at least as much right to be given in the account of the Second Way. In order to justify Aquinas’s principle "Everything that is in process of change has that change initiated in it by something else", I made use of the principle that every beginning of existence has a cause. This principle relates most obviously to efficient causality, and, indeed, to cases of efficient causality (beginnings of existence) which are not strictly speaking initiations of processes of change at all.
It is not clear what should be said here. It may be that the understanding of the First Way we have been striving to achieve conflates it unduly with the Second Way, and the critic may suspect that the apparent effectiveness of either of the two ways rests on an unnoticed confusion of the distinct but related notion used in the other. There seems to be no a priori way of determining an answer to this: what one must do is await a definite challenge in this line and hope to find a way to refute it. I suspect that any such challenge would have to centre on an allegation that there has been a confusion in the way in which subjects of change in the First Way were identified rather loosely, and draw attention to the mis-match between the account given there and any likely identification that could be made of the efficient causes and effects mentioned in the Second Way.
The actual text of the Second Way clearly follows a structure similar to that of the First Way, and equally clearly exemplifies the overall structure which was suggested for all of the Five Ways.
"The second way is from the notion of efficient cause.
We find, in things around us that we sense, an order of efficient causes.
For if anything were, it would have to be prior to itself, and this is impossible.
This is so whether there is just one intermediate cause or more than one.
If a cause is eliminated, the effect is eliminated too: so if there were no first efficient cause, there would be no last or intermediate cause either. But if the series of causes is infinitely long, then there will be no first efficient cause, and hence no last effect and no intermediate causes. This is obviously not the case.
Again the indentation indicates the structure of the argument, to some extent. The main structure is clear and apparently valid:
2. But we do not find — nor could there be — anything that is the efficient cause of itself.
3. It is not possible that there should be an infinite series of efficient causes.
4. Hence we must suppose some first efficient cause, and this is what all call ‘God’
The next step is taken tacitly: it is the lumping-together move. Anything which is a system of parts related by efficient causality, in which parts are effects, is itself an effect: something which cannot be the efficient cause of itself, and which requires an explanation in terms of a relation to something else. Put this way, the step looks dubious. One might want to hold that there might be parts of the world which are outside the system of efficient causality. One would want evidence for this, but it seems to me that in this case feature X does not generalise sufficiently. It is true that a whole which contains any parts which are in process of change is itself in process of change, as we saw in the First Way: but it does not seem so obvious that a whole which contains parts which are effects must itself be an effect. Perhaps, though, the lumping-together move need not be so widely generalisable. Whatever we may say of the world as a whole, it is clear that a great part of the world is a system of efficient causality, and, indeed, it is worth saying that so far as any of us can tell the whole world is a system of efficient causality. It is hard to see what could be meant by saying that there are parts of the world which fall outside the system of efficient causality which the rest of the world belongs to. In what sense would these causally ineffective and unaffected bits of the world form part of the same world at all? What, we might ask, is a world supposed to be except a unified system of efficient causality?
A system of efficient causes, like the world, is eo ipso a system of effects, and here the lumping-together move appears valid: a system of effects is an effect. Feature X is sufficiently generalisable, in that it extends to a great part of the world at least, if not necessarily to the world as a whole, though in fact it seems to, and it also seems difficult to understand what it would be for this feature not to extend to the whole world. This system of effects, of efficient causality, which is at the very least a very great part of the world, cannot cause itself: it requires an explanation in terms of a relation to something different.
St Thomas goes on to make, in his third premiss, "It is not possible that there should be an infinite series of efficient causes", in slightly different terms, the point made in the First Way, that we cannot go on to infinity in this line; or, as we claimed above, more accurately, going on to infinity in this line fails to be an explanation. "Now, it is not possible that there should be an infinite series of efficient causes. This is because in any ordered series of efficient causes, the first cause is the cause of the intermediate causes, and the intermediate cause is the cause of the last cause. This is so whether there is just one intermediate cause or more than one. If a cause is eliminated, the effect is eliminated too: so if there were no first efficient cause, there would be no last or intermediate cause either. But if the series of causes is infinitely long, then there will be no first efficient cause, and hence no last effect and no intermediate causes. This is obviously not the case." This, I think, given the formulation of Aquinas’s claim which I have used, that going on to infinity in this line fails to be an explanation, and given the account I have given of this claim in the last chapter, needs no detailed defence here, though Aquinas’s detailed defence has some interesting features which might be worth investigating elsewhere.
And thus Aquinas reaches his conclusion: "Hence we must suppose some first efficient cause, and this is what all call ‘God’". The account of the signification of the word "God" contained in this conclusion seems to be unquestionable. Objections to this Way will have to rest on the notion of efficient causality being employed, or on the possible identifications of the subjects of efficient causality, and on the doubt about the extent to which "being a subject of efficient causality" is appropriately generalisable in the lumping-together move. Otherwise the argument appears unassailable.