But this may be a mere technical objection. The heart of what Craig says seems right. He might add that the conjunction of the premises should seem more plausible than not. Better: the proposition that the conjunction of the premises is true seems more plausible than not.”

If wrote-off my car, the compensation required should be proportional to the extent to which I rely on it. If I had many cars, I could still get to work in one of the others.

In the same sense, a body is an essential vehicle required to experience life. it would be far more grievious to have ones body taken away than to experience some other run of the mill property crime.

To accept the conclusion of an argument, we need to accept not that any given premise is probably true, but that all premises are probably true. If the conjunction of premises is not probably true, then the conclusion isn’t probably true, because it is probable at least one premise is wrong, thus the argument (given what we know) likely isn’t sound.

A concrete example (I owe to Johnsonav)

1) Dice A, when rolled, comes up 1, 2, 3 or 4.
2) Dice B, when rolled, comes up 1, 2, 3 or 4.
C) Dice A and B come up 1-4

This is formally valid by conjunction introduction (or can be made formally valid easily if I’ve slipped up). Now suppose I tell you I have rolled two dice, and argue in that manner. You should accept both 1 and 2: for a six sided fair dice, both Premise 1 and Premise 2 have a probability of 4/6. Yet you plainly shouldn’t accept the conclusion of my argument, valid though it is. Because the probability of both A AND B coming up 1-4 is 4/9. Add as many dice as you care to strengthen the point further.

Another easy counter-example. Take all beliefs you hold about the truth or falsity about certain propositions – all questions you answer with a yes or a no (is there a Table in front of me? Does God exist, etc.) As you hold these beliefs, obviously you consider your beliefs more plausible than their negations. It follows by the same rule that if each of these beliefs 1…N is true, then your set of these beliefs are all true. Yet I shouldn’t say that. I’m fairly sure that at least one of these sorts of beliefs I hold is in error.

From what I can follow of Craig’s argument, it doesn’t seem to make sense. Talking about the argument being sound or whether it in fact follows is simply irrelevant – we can concede all his stuff about arguments being in fact sound or not, about bivalence over truth, and all the rest. If someone sets out a series of concerns (premises, here) to demonstrate a particular conclusion, the argument is only sound if, IN FACT, all premises are true. So whether you should accept the argument depends upon whether, insofar as your epistemic spotlight can tell, that it seems that all premises are true. If it’s more probable that they collectively aren’t, then you shouldn’t accept the conclusion, even if, individually, each is more plausible than not. Attempting to beat an argument by introducing ‘reasonable doubt’ into its premises is entirely acceptable, at it makes confirmation leak out – there are lots of ways the argument could plausibly ‘go wrong’.

So Craig is surely wrong here (although he’s right to note that the original question implies a rebutting defeater in the dog example). The error, insofar as I can tell, springs from an obsession with the P=0.5 threshold. Arguments with premises P much greater than 0.5 can fail to force their conclusions. Likewise, arguments with premises that aren’t more likely than their negations may also have some role (consider the Kalam example above with Ps of 0.6 – that still can be ‘useful’ if discussing with someone who sets P(God) below 0.36).

For what makes a good argument, I think it is simply one which, when heard, exerts pressure to change a reasonable persons probability assignment. I suspect for deductive arguments, it is simply whether the conjunction of the premises is more likely true than not. If you want to show that God exists by the lights of the Kalam argument, that’s the sort of standard you need. For abductive arguments, not so much: you just need to show instead of P(conjunction of premises)>0.5, something milder like P(E|God)>P(E|¬god), where E is some evidence. Our assignment of P(E) itself doesn’t really matter (although obviously if it is low the argumentative force will be pathetic either way).

Why are the rules different? In the former case, we are after the actual probability of God’s existence, whereas in the latter we are considering the likelihood ratio. A deductive argument sets a lower bound on P(conclusion) equal to P(conjunctofpremises). If this is higher than our prior assignment, we’ve made progress (and if it’s higher than P=0.5, it commands belief. If not, it’s feckless. For an abductive or Bayesian argument, we just need to show that it confirms our conclusion over its negation. The persuasive for this will have will be prior sensitive.

Yet I don’t think you’re right here. It seems simple for a substance dualist to say something like: “Sure, you aren’t your body, but your body (in fact) is the only route through which your soul interacts with the world, and body and soul interact. So abuses to your body can harm your soul too (e.g. pain from bodily torture). Murder is bad not because it ‘kills’ your immaterial self, but because it precludes that self from enjoying the great goods of bodily life.” Wouldn’t this get Descartes out of jail?

Many of the Muslims that move to the West end up adopting large elements of western culture and values, those that don’t often have children that do. This is a good thing.

Videos like this only cause trouble, by encouraging racism and intolerance. It is irresponsible of Christians to spread this sort of nonsense.