Cosmological argument

Summary (for Phil U1) of the remainder of JL Mackie’s chapter on Cosmological arguments as promised. This is rather scrappy as it stands, and if I had more time I would polish it – something I mean to do at some future point. But as it is I am stuck with very limited access to use of computer and wi-fi in the New Forest Inn.


I hope it is of some use as it is. (Numbers in brackets are page references to the text, which you should read alongside this summary.)

Part (b) The regress of causes

Mackie first mentions (87) a simple, ‘popular’ first cause argument:

(1) Things must be caused
(2) The chain of causes cannot go back infinitely
(3) There must be a first cause
(4) This first cause is God.

As he says, there are three simple and obvious objections:

(a) Why should we accept (2)? Why can’t there be an infinite regress of causes?
(b) Even if the chain must terminate, why do we have to believe that it terminates in a single first cause? Why might it not be that there was a plurality of first causes at the beginning of things?
(c) Even if there is just one first cause, what reason do we have to identify such a first cause with God?

These questions are temporarily shelved (till 91 bottom), while Mackie goes on to consider Aquinas’s third way (87 bottom).

Third way summarised in a form familiar from your notes and the textbook summary (88 top).

Main criticisms of first stage of the third way (again, we have already covered these in class)

(a) (88 middle) We can challenge the claim that contingent beings must be impermanent (have expiry dates): Why can’t there just be a thing which is contingent, but which happens to exist at all times? (This objection is compelling, unless what Aquinas means by a thing that is ‘able not to be’ something like ‘impermanent entity’.)

(b) Even if contingent things are impermanent, if doesn’t follow that if everything were contingent there would have to have been a time at which they had all already ceased to exist; for:

(i) first (88 bottom), even if there must at some point be such a time, there is not reason to thing that it would already have had to occur. Perhaps the impermanence of currently existing things is going to reveal itself in the future; perhaps we just haven’t yet reached the expiry dates.
(ii) Secondly (89 top), there is no compelling reason for thinking that there must at any point be such a time: surely there could in principle be an infinite, overlapping, series of impermanent things. (Though it must be admitted that it would be improbable for such a series to be maintained over infinite time.)

(c) (89 middle) Even if (a) and (b) could be answered adequately, why think that the principle that nothing can come from nothing is correct? It has an intuitive appeal, admittedly, being supported by experience (all of the ordinary things that we observe do seem to be produced by other things; but it does not seem to be known a priori: we can, it seems, at least conceive of an entity that began to exist without being caused.

Conclusion on first stage of third way: Certainly not watertight proof; but some lesser support has been provided for the view that there must be something permanent.

Second stage analysis (90 top)
(a) Aquinas argues in the second stage that even an infinite series of causes must have a first cause. The ground for this claim is that, in any series of causes and effects, if you take away the first cause, the others will not occur.

Criticism: the argument is sound as to finite series of causes; but it is a fallacy to think that the same applies to infinite causal series.

As Mackie says, however, though the argument is not sound, there is an intuitively appealing thought behind it, namely that where there is a relation of dependence linking events in a series, there must be an end to the regress. This is backed by the thought that even an infinite series of train carriages will not move without an engine, or that an infinite chain of links cannot be held up without a hook, or that an infinite series of gear-wheels in a watch could not turn without a mainspring.

(b) Even if that thought is allowed, however, there’s a further problem (91 middle): Aquinas assumes that the ultimate permanent item must be God, on the ground that any permanent permanent item must either derive its permanence from something else, or derive its permanence from itself in some way that means that its essence involves existence, and must ‘therefore’ be God. But we have no ground for thinking that the ultimate permanent things must be God: surely there might, instead, be a permanent stock of matter that was not God in any sense – whose essence did not involve existence – but which also did not derive its existence from anything else.

Return to the popular first cause argument (91 middle)

The discussion of the second stage of Aquinas’s argument has usefully informed our thinking about the simple, popular argument, especially as regards the first objection to it. The idea that there cannot be an infinite regress of causes does deserve at least some respect: given that effects depend on earlier causes, it is plausible that a causal chain cannot go back infinitely.

So the simple popular version can be defended to a degree. But:
(i) it is still obviously not compelling, even as to the view that there must be a first cause.
(ii) it remains decidedly weak as to the claim that the first cause must be God.

Part (c) Finite past time

Various philosophers have argued (either as a distinct cosmological argument, or as part of the support for the premiss (common to a number of versions of the cosmological argument) that there cannot be an infinite regress of causes) that there cannot have been an infinite past time.

Some of the supposed proofs trade on mathematical arguments about infinity (92 bottom to 93):

(i) there could not have been infinite past time, because we would have had to traverse it to reach the present; but an infinitely long period of time cannot be traversed.

(ii) If there had been infinite past time, then both Saturn and Jupiter would have completed an infinite number of orbits. So the number of orbits completed by each would be the same number (namely, infinity); but since Jupiter orbits more than twice as fast as Saturn (12 years versus 30 years), it ought to have completed more than twice as many orbits. So the supposition that there has been infinite past time leads to a contradiction: that S and J have completed both (a) the same number of orbits and (b) different numbers of orbits. Therefore we should reject that supposition as absurd.

Comment (93 middle):
These arguments trading on apparently paradoxical aspects of the notion of infinity should be rejected:

(i) assumes that infinite past time must nonetheless have had a starting point – one that is infinitely remote in the past. But this is misguided: taking serious the notion of an infinite past means saying that there is literally no starting point; and that from any actual point in the past there is only a finite distance in time from that point to the present. In which case the problem is resolved.

(ii) the paradox is resolved by seeing that the ‘contradiction’ arises only because the argument uses ordinary criteria for counting sets as equal/unequal in size which, when applied to finite sets, yield the result that a set that is smaller than another cannot also be equal to it. But when the criteria are applied to infinite sets, it is possible for a set (using those criteria) to be both smaller than and equal to another. The apparent contradiction is not real.

So there is no a priori disproof of the possibility of infinite past time.

Other arguments against an infinite past (93 bottom) trade on empirical/scientific support – such as the evidence for the big bang. And these, though they are not proofs (for we cannot know with certainty that the big bang was not (in some sense) ‘preceded’ by something else), do yield some significant support for the idea that past time has not been infinite.

The fundamental problem, though (94), with using this in an argument for the existence of God, is that any such argument has to establish that God’s origination and/or existence is somehow self-explanatory and unproblematic, whereas the origination/existence of the universe would be mysterious and problematic in the absence of God. But no such claim can be established: we are not entitled to help ourselves to the notion of God as a being whose existence is self-explanatory because it is not possible that it should not have existed. And if the notion is coherent at all, why should it not apply to the universe itself, rather than God?

(d) Swinburne’s Inductive argument

No cosmological argument can succeed as a deductive proof from known premisses. That’s unsurprising, since the conclusion of any such argument goes far beyond the empirical data available for the premisses.

The question is whether an inductive proof (as offered by Swinburne) fares any better.

(95 bottom to 97 bottom) Summary of Swinburne’s account of principles of inductive reasoning (not critical for what you are doing).

(97 bottom to 99 middle) – discussion of whether a reasonable C-inductive argument (as defined 96 top) for God can be created; again not critical for you.

The essential argument in Swinburne (99 middle) is that the existence of the complex universe that we have is better explained – less puzzling – on the hypothesis that there is a God than on the hypothesis that there is no God.


Swinburne claims that the existence of the universe in the absence of God would be ‘puzzling’ or ‘unlikely’. But this claim is without merit. It is not at all like the thesis of otherwise unexplained similarities between a set of manuscripts without a common ancestor that would explain the similarities.

Swinburne massively underestimates the difficulties and improbabilities attending his notion of an uncaused God who has created everything (100 top half): it is true that if there were such a God as Swinburne imagines (with the power to create a universe through direct fulfilment of His intentions, without any physical or causal mediation), then of course it would be possible for Him to have created the universe as it is; and there might even be some reason to think that He might have wanted to do so. But the inherent probability that there should be such a God at all has to be enter the equation, and this must be rated as very low. We have no real concept, and certainly no experience, of such direct and unmediated creative power; our only experience of creation is of embodied intentions’ being fulfilled indirectly, via bodily movements and causal processes.

Swinburne’s claim that God’s ‘simplicity’ – supposedly manifested in His being a single entity of infinite power, knowledge, and goodness (100 bottom half) makes it easier to believe in God as creator than to believe that the universe in all its complexity should exist in the absence of God is equally unacceptable:

(i) it requires postulating a whole series of infinite powers and attributes and assuming that they are unproblematic,
(ii) it involves ignoring certain particularities that we may not ignore: we shall still have to attribute particularities in God, if we are to be able to explain why He has made a universe like this one rather than another.


A ‘simpler’ hypothesis?

A rather lazy post, in that it requires no new philosophical writing on my part… BUT I should point out that now would be a very good time for F1 and F2 students to read the following three older posts here:

A reasonable hypothesis? (1)

A reasonable hypothesis? (2)

A reasonable hypothesis? (3)

The point is that they are attempts to flesh out the sort of inductive argument for the existence of mind-independent objects that we have been examining, by saying more (in particular) to elucidate the notion of ‘simplicity’ on which that argument trades.

‘Direct’ perception

One of the key messages that I have been aiming to get across this week to F1 and F2 is that, in the arguments against Direct Realism and for Indirect Realism, the notion of ‘direct’ or ‘immediate’ awareness is very specific.

‘Direct’ perception in Russell, as I have been saying, seems to incorporate two notions:

(i) the absence of intermediaries – I am not perceiving anything directly if I perceive it only by perceiving some other object (thus, I am not perceiving Judith Holofernes directly if I am only looking at a photograph of her).

(ii) veridicality – I am not perceiving anything directly unless the way it looks is the way that it really is.

The first of these elements is (I think) pretty obvious and intuitive. That the second of the two notions is involved in the notion of ‘direct’ perception is, however, rather less obvious.

The need for this second element becomes clear when you review the arguments against Direct Realism that we have been considering.

Take, for example, the Argument from Illusion that we considered last week: this runs roughly as follows:

(1) The stick looks bent.
(2) The stick isn’t bent.
(3) What I am directly aware of cannot be the stick, but must be something else – a sense datum or sense data.

Notice that this argument would be hopeless unless ‘directly aware of’ meant something like ‘aware of in such a way that the way it looks must be the way it is’. Otherwise, you could reject the argument by saying, simply: ‘The premisses are true; but they don’t force me to accept the conclusion; after all, maybe the object of which I am directly aware is the external stick – it’s just that (as we all know) when you put a stick in water it looks bent, even though it is straight.’

As I let slip to F1 this morning, the sense data theory (at least in one of the forms in which it is standardly presented) is nowadays in fact pretty unpopular among professional philosophers, even though the A-level syllabus seems to want you to believe that it is by far the best of the Realist theories – one to which the objections that you are obliged to consider are pretty simple to deal with. (The reason for this is that they seem to think that A-level students are too dim to be able to understand a more sophisticated version of representative realism. I know that you are not so dim; so here in outline is what is being concealed from you:)

Why don’t people like sense data? Several reasons, including:

(1) Sense data are supposed to be objects of awareness as to whose appearance we cannot be mistaken. But what is this mysterious faculty by which we are infallibly aware of the nature of sense data? If it gives us infallibly veridical awareness of the nature of its objects, then it is quite unlike any other sensory faculty of which we have any understanding.

(2) Where are sense data located? They are supposed to be private objects – so mine must somehow be in me in a way that makes them inaccessible by anyone else; and they are supposed to be mental items. So the obvious answer is that they must somehow be ‘in the mind’. Now, the mind is either material or immaterial in nature. Either way, though, there seem to be insuperable problems ((3) below).

(3) Sense data are supposed to be invariably and necessarily the way they appear. One cannot be mistaken about them. But then sense data seem to have to have natures that no internal object could have: for example, if the gorilla looks hairy, then I must have a hairy sense datum; if the gorilla looks about eight feet tall, I must have an 8-ft high sense datum; but (even if sense data are material) how can I have such a thing within me? I’m not big enough! And (if sense data are immaterial in nature) how can an immaterial thing be hairy, or eight feet tall?

Where the theory seems to go wrong is in suggesting that, when external things look otherwise than they are, this is to be accounted for by saying that there are internal things that we perceive that really are the way things look. The sense data theory ‘reifies‘ looks – that is to say, it accounts for the appearances – the ‘looks’ of things – by positing the existence of things (objects, entities) called sense data which allegedly really possess the properties that the external things appear to have (like being bent) even if the external things (like the straight stick) lack them. The main problem with this is one of explaining how there could be such internal things, and how it could be that we have infallible, veridical, awareness of the properties of such internal things.

What we really want to be able to say is the following (and this is, in outline, the version of ‘direct’ but ‘representative’ realism that most modern realist philosophers accept):

  • Perception represents the world as containing external, mind-independent objects.
  • Perception sometimes misrepresents the world – as in the case of illusions (it makes things look otherwise than they really are).
  • There is no need to posit the existence of an additional set of internal objects of awareness in order to account for this kind of mismatch between appearance and reality; instead, we can explain it by saying that experience has a representational content – it appears to us that the stick is bent – but without saying that this content involves the existence of any object of awareness that really is bent. This notion allows us to have representation without reification.
  • The only objects of perception are the external, mind-independent, ones. And these are therefore the objects that we perceive as directly as we perceive anything.
  • We are sometimes subject to illusions when we perceive these objects. So in saying that we perceive them directly we should be understood as saying only that we perceive them without perceiving any intermediate object; and we should not be understood as saying that our perception of them is always veridical.

Philosophy Tea misnomer

I completely forgot the tea and the other tea things. So no one had tea in first week. Sorry.

One person emerged shining and guilt-free from the week, namely Savannah, who yet again produced delicious and beautiful cupcakes as pictured. Thank you. I will try harder from now on.