I will be reposting a previous article today because ideas of induction and deduction have come up in another post with a few asking for clarification. Here, I will look at deduction and induction in the context of how they relate to the Kalam Cosmological Argument. Alan Duval has also written about induction and deduction recently here in “Answering Questions about Empiricism as Foundational: Our Rational Nature”.
As some of you will know, my most recent book was on the Kalam Cosmological Argument, the cosmological argument that deals with the beginning of the universe, concluding in favour of it having a cause (i.e. God). the argument is favoured by apologists the world over. It has obsessed me for years to the point that I wrote a dissertation on it that later became my book (Did God Create the Universe from Nothing? Countering William Lane Craig’s Kalam Cosmological Argument). William Lane Craig, famous Christian philosopher and apologist, uses it in almost all of his debates, and it is one of his claims that I will now deal with, taking an excerpt from the book.
The arguments in the book vary from some logically technical (but no more difficult than that below) to the scientifically technical (yet couched in easy to understand terms) to philosophically beard-scratching (talking about time and infinites). Throughout the book, I try to steer the discussion through calm waters of easy-to-digest prose. Whether I have succeeded or not is up to you. Please grab a copy!
2.1 The Form
1) Whatever begins to exist has a cause.
2) The universe began to exist.
3) Therefore, the universe has a cause.
The above argument is a logical syllogism. A syllogism is an argument where the conclusion is inferred from the premises. As Craig likes to set out[i], it follows the same form as:
1) All men are mortal;
2) Socrates is a man;
3) Therefore, Socrates is mortal.
These two arguments are examples of modus ponens deductive arguments. Modus ponens syllogisms take on the symbolism:
P → Q
P
\ Q
Which means that P implies Q; P is asserted to be true, and it follows, using the rule of inference, that Q is true. P and Q are categorical propositions. Both premises in both arguments contain a subject and a predicate. For example, in the Socrates syllogism, 2) contains the subject (Socrates) and the predicate (is a man) in the same way The Kalam Cosmological Argument second premise contains a subject (the universe) and a predicate (began to exist). These premises are essentially assertions which may or may not be true. The conclusion follows necessarily from these categorical propositions (premises). Thus the argument may be valid such that the conclusion does follow necessarily from the premises, but that the premises, as assertions, may be called into question.
This form of syllogism (modus ponens) is, as mentioned, a type of deductive reasoning. This means that, again as mentioned, the conclusion follows necessarily from the premises. As long as one agrees with the premises, then one cannot get away from the fact that the conclusion must be true. If, indeed, all men are mortal, and if, indeed, Socrates is a man, it follows without fail that Socrates is mortal. This is different to inductive logic, which entails garnering a conclusion based on premises that are built on patterns of individual instances, broadly speaking. In fact, it is a probabilistic method. Here is an example of inductive reasoning (as opposed to deductive reasoning):
- All dogs that have ever been seen have ears;
- Therefore, all dogs have ears.
The conclusion follows from the premise probabilistically and not necessarily. Describing the difference here is very important for looking at my first critique of the KCA. In the Socrates example, and as with all strong deductive arguments, the power of the logic is definitional. This is the key to deductive reasoning—the very definition of words determines whether or not the argument can be held to be true. As Nancy Cavender, Professor at the College of Marin, says of tautologies[ii]:
We can determine the truth of… a tautology by logical—deductive—means alone, without the need for any empirical investigation or inductive reasoning, But determining the truth or falsity of a contingent statement requires observation by ourselves and others and, usually, the employment of inductive reasoning.
Now, deductive arguments are not tautologies, not least because they are not coextensive such that the statements are not identical per se, but they do have an equivalence, if not in both directions. Built into the definitions of the words “man” and “mortal” is an equivalence (all men are mortals, even if all mortals are not necessarily men). A tautology would be like saying 3 = 3, whereas (though this analogy has its obvious limitations) a deductive argument might say 3 = 2 + 1. We accept the semantic meaning of both “man” and “mortal” so that, as Cavendar says, “without the need for any empirical investigation or inductive reasoning” we can deduce the conclusion. This is the strongest, most ironclad form of logic, resting on semantics as opposed to empirical observation and probability.
Let us, then, look at the KCA again. The first premise is “whatever begins to exist has a cause”. This, I posit, is a categorical proposition which amounts to an inductive piece of reasoning akin to “all dogs that have ever been seen have ears” as opposed to a categorical proposition based upon definition such that “all men are mortal”. The definition of “man” infers mortality. However, “whatever begins to exist has a cause” is not definitional and is dependent on empirical observation. This premise actually amounts to, at best, “everything which we have observed to begin to exist has appeared to have had a cause”. It is an assertion, a mere assertion at that, that everything which begins to exist has a cause, in the same manner that the premise “all dogs that have ever been seen have ears”. In the same way that it doesn’t necessarily follow that “therefore, all dogs have ears”, it doesn’t necessarily follow that “everything that begins to exist has a cause”. This can be called a simple induction by enumeration to a generalisation. Generalising a rule from a number of observations. Now, Craig might counter that this is not merely based on observation, but is also a metaphysical intuition. It just feels that this is true, that every effect has a cause. More on that later.
It is also worth bringing up here how deductive arguments can mask uncertainty. This is summed up by Jeffery Jay Lowder in a piece “How the Distinction between Deductive vs. Inductive Arguments Can Mask Uncertainty”[iii]. Let’s take a simple syllogism:
1) If A, then B
2) A
3) Therefore, B
What is the probability of B? Well, it is conditional upon A, so if A is true, then B has a probability of 1. It is certain. However, the unconditional probability of B is dependent on the probability of A. A deductive argument does a good job of seeming concrete, valid and sound, but it masks the weakness of the opening premise. Lowder states this in logical terms:
But we want to know the unconditional probability of B, Pr(B), not the contribution made to Pr(B) by the probability of B conditional upon A, Pr(B|A). So, again, what is the contribution made to Pr(B) by A itself? Answer: Pr(A). Suppose A is true by definition. In that case, its probability is 1 and so Pr(B)=1. Now suppose the probability of A is 50%. In that case, Pr(B)=0.5. The probability calculus implies, when Pr(B|A)=1, that the contribution to Pr(B) made by A alone will always equal Pr(A), so long as Pr(A) is not zero, which in “real world” problems is usually the case.
Before I get on to talking at some depth about the idea of induction, let me say a little about acausality. There is a position that could be held, similar to Jung’s synchronicity[iv], that events might appear to be causally connected, appearing one before the other, but not actually having a causal relationship. This would be damaging to the position of adherents to the KCA. Strictly proving causality is notoriously difficult, and the theory of synchronicity is often seen as not being particularly useful.[v] Whilst humans cognitively over-determine causality in some situations, mistaking causality is not equivalent to claiming there is no causal relationship at all. For example, if I am in the jungle and see the undergrowth move, I may conclude that this is a tiger, and run away. Let’s say I am wrong; it is not a tiger. This misattribution does not negate causality. It turns out, let us say, that it was the wind. Causality was still at play, just in a different form. As an aside, we can see why we may have evolved over-determining causality as it can work in our favour. Mistaking a wind for a tiger, and running away, I still live. That said, the idea that there must be causality is indeed arguable, and that certain events could be acausal yet synchronous is worth considering.
What my main objection to the form of the argument represents is the notion that within the deductive form of the KCA, the author smuggles in a premise which amounts to an inductively garnered conclusion. In other words, the KCA should be set out as follows:
- Everything that we have observed to have begun seems to have had a cause
- Therefore, everything that begins to exist has a cause
- The universe began to exist
- Therefore, the universe has a cause
However, this argument is only as good as its inductive premises (see my earlier point and Lowder’s quote with regard to deductive arguments, too), and does not have the deductive qualities that many proponents proclaim. Of the form of the Socrates modus ponens, Craig claims, “This is one of the most basic and important logically valid argument forms”[vi]. But his argument rests upon the idea that the KCA and the Socrates modus ponens arguments are synonymous in form when in fact they are ever so slightly, yet crucially, different.
This has a twofold implication. Firstly, as stated, its strength is only really inductive, allowing weakness into that first premise. Furthermore, the second premise is also not a premise dependent upon semantic logic, but upon empirically based observations and theorising. Secondly, Craig is using induction[vii] to make a claim about things beginning to exist requiring a cause but he is inducing that information from things in the world, rather than inducing the premise from examples of worlds/universes coming into existence. Of that he has no other experience, and so even inductive knowledge can get Craig nowhere here. As philosopher Wes Morriston states in “A Critique of the Kalam Cosmological Argument”[viii]:
Is premise 1 of the kalam argument true? Must everything that begins to exist—even the very first event in the history of time—have a cause? Craig thinks it is unnecessary to give a lengthy defense of this claim. “Does anyone in his right mind”, he asks, “really believe that, say, a raging tiger could suddenly come into existence uncaused, out of nothing, in this room right now?” [Craig 2002][ix] Probably no one does. Craig then invites us to apply this “intuition” to the beginning of the universe, and the case for the first premise of the kalam argument is about as complete as Craig ever makes it.
But surely this is much too quick? Of course, no one thinks a tiger could just spring into existence “in this room right now.” But before we jump to conclusions, we need to ask why this is so. What makes this so obvious? Is it, as Craig seems to suppose, that all normal persons believe the first premise of the kalam argument, and then apply it to the case of the tiger? Call that the top-down explanation. Or is it rather that we have a lot of experience of animals (and other middle-sized material objects), and we know that popping up like that is just not the way such things come into existence? Call that the bottom-up explanation.
The bottom-up explanation takes note of the fact that we are dealing with a familiar context – one provided by our collective experience of the world in which we live and of the way it operates. It is our background knowledge of that context—our empirical knowledge of the natural order—that makes it so preposterous to suppose that a tiger might pop into existence uncaused. We know where tigers and such come from, and that just isn’t the way it happens.
Now contrast the situation with regard to the beginning of time and the universe. There is no familiar law-governed context for it, precisely because there is nothing (read, “there is not anything”) prior to such a beginning. We have no experience of the origin of worlds to tell us that worlds don’t come into existence like that. We don’t even have experience of the coming into being of anything remotely analogous to the “initial singularity” that figures in the big bang theory of the origin of the universe. The intuitive absurdity of tigers and the like popping into existence out of nowhere does not entitle us to draw quick and easy inferences about the beginning of the whole natural order.
Consequently, we have multiple weaknesses of the argument based on its logical form. I don’t really want to go into detail here about the content of premise 1 within this context. Suffice it to say Craig is drawing inductive probability upon a behaviour, in premise 1, over which he has no probabilistic knowledge. Craig simply has no knowledge through experience of worlds beginning to exist (I will expand on this in section 3.4). As such, he has no epistemic right to claim the categorical proposition of premise 1 as being sound. Craig, in this case, merely appeals to intuition (but without the relevant evidence, since we have seen him use the ontology, or the nature of being, of one thing to assume the ontology of another) in a rather dismissive manner:
First and foremost, the causal premiss is rooted in the metaphysical intuition that something cannot come into being from nothing. To suggest that things could just pop into being uncaused out of nothing is to quit doing serious metaphysics and to resort to magic. (Craig, 2007)
Thus we have an inductive premise which is based not on probabilistic empirical evidence (seeing many universes begin to exist with a cause) but on intuition. At the very least, this admits a weakness in the claims of deductive power imbued in the argument. Craig is relying on mere human intuition, and this argument is at best, then, an intuitive argument (albeit one that he uses observation to support). That does not necessarily invalidate it, but takes it down a step or two from the logical pedestal upon which Craig has placed it.
[i] In debates and talks such as his 2010 talk to Biola University entitled “The World’s Ten Worst Objections to the Kalam Cosmological Argument”, https://www.youtube.com/watch?v=sfsYhWNMYr4 (Accessed 09/12/2015)
[ii] Cavender and Kaahane (2009: 37)
[iii] Lowder (2013)
[iv] Jung (2010)
[v] Bishop (2000: 59-62)
[vi] From his 2010 talk to Biola University entitled “The World’s Ten Worst Objections to the Kalam Cosmological Argument”, linked in [3].
[vii] I will be using Craig as a synonym for “the author of the KCA”.
[viii] Martin and Bernard (2002: 95-108)
[ix] Morriston gives the reference as quoted in the bibliography here, but the link is no longer valid. I found virtually the same quote in Craig and Moreland (2009: 189)