**3. The basic shape of the argument**

*3.1 Definitions*

Let F be the proposition that the laws of nature, the constants of physics and the initial conditions of the universe must have a very precise form or value for the universe to permit the existence of embodied moral agents (EMAs).

Let EPU be the proposition there exists a material, spatiotemporal domain which permits the existence of embodied moral agents.

Let NSU be the proposition that there exists one (and only one) universe in which the laws and constants of physics are roughly similar throughout.

Let T be the proposition that there exists an omnipotent, omnibenevolent, personal being who is modally prior to the universe and thus transcends it.

*3.2 The argument*

P1) The restricted likelihood principle is sound.

P2) EPU obtains.

P3) T was advocated prior to the discovery of F, and has independent motivation.

P4) F obtains.

P5) If F, then the conditional epistemic probability P(EPU|NSU & k’) << 1.

P6) If F, then the conditional epistemic probability P(EPU|T & k’) is not much, much less than 1. That is, ¬[(EPU|T & k’) << 1].

C) Therefore, necessarily, EPU constitutes evidence for T over NSU.

Now, the first 3 premises seem uncontroversial. I have already explicated and defended the restricted Likelihood Principle in section 2, justifying premise 1. Premise 2 simply defends the truth of proposition EPU, viz. that there exists a material, spatiotemporal domain which permits the existence of embodied moral agents. It would be hard to deny this – our universe is a material, spatiotemporal domain, and we are both embodied and moral agents. Similarly, premise 3 is palpably true: theism, including the type considered here, was advocated long before F came to light (T was proposed millennia ago, F discovered in the last few decades) and has other, sincere putative arguments in support, even if the arguments turn out not to be sound. So the conclusion of the argument rests on the latter 3 premises. I will attempt to justify each of these in turn, but first, a word about the conclusion.

This is a relatively limited conclusion, of course, though it is still a significant one. After I have defended premises 4, 5 and 6, I will turn to the question of whether we can make additional arguments to the conclusion that P(EPU|T & k’) > P(EPU|NM & k’), where NM represents the naturalistic multiverse hypothesis. After that, I will argue that, whether or not the above argument is sound, we will still be able to conclude from the core argument that P(EPU|T & k’) > P(EPU|¬T & k’), and therefore that there is evidence for T over ¬T in this argument. Finally, I will discuss how this relates to the question of P(T|EPU & k’ & F).

**Contents**

2. Some preliminaries on formulation