Christian historian Dr. Wallace Marshall and I are debating whether or not enough evidence points to the existence of a god. For background and format, and Dr. Wallace’s opening statement, see entry one; for the rest, our index.
That the Evidence Points to Atheism (VIII)
by Richard Carrier, Ph.D.
Dr. Marshall claims “for a particular fact to count as evidence for theism it isn’t necessary that” it “should be more probable on theism than on atheism, but only that it make theism more probable” than without it. But I already noted that’s impossible. Unless the fact is more likely on theism than atheism. Marshall doesn’t get the math.
If you start with ignorance (a total absence of evidence), at most there’s a 50/50 chance a god exists or not. The only way to increase that probability is to find some evidence e such that P(e|h) > P(e|~h). Because that’s the only way to increase P(h). Otherwise if P(e|h) = P(e|~h), then you have a multiplier of 1, and anything multiplied by 1 is itself, not a higher number; which evidence won’t move the probability in either direction. And if P(e|h) < P(e|~h) then e decreases P(h), not increases it. As these three options exhaust all possible states of the evidence, clearly P(h) can only increase when P(e|h) > P(e|~h). Marshall is simply wrong to claim otherwise. 
My gremlins analogy illustrates this, showing why no one believes in gremlins. If people had an irrational investment in the existence of gremlins (as they do in God), they would believe in them despite the absence of evidence. Marshall weirdly confuses “people believe x” with “there must be evidence for x.” To the contrary, most false beliefs are believed in error, not on a rational evaluation of evidence. 
Marshall keeps incorrectly claiming natural evils are less than moral evils.
Per capita, more people, especially children, have been maimed, disabled, tormented, and murdered by diseases, accidents, and disasters than by “people.” Until very recently in human history, 50% of all children didn’t survive into adulthood (a body count even the Holocaust came nowhere near to globally achieving) and natural disasters all combined (like famines, floods, earthquakes, and tsunamis) killed, ruined, or maimed more people than war. And then there’s the vast realm of pointless miseries suffered by animals for five hundred million years. Natural evil vastly outnumbers moral. Indeed moral evils have been on steady decline ever since secularism began to take hold, not god or religion. 
But it’s also illogical to say “God has caused less evils than people, therefore God is good.” Even if natural evils were less frequent than moral, God is still responsible for them. And a good God would not allow that; because there’s no plausible (much less probable) reason he would. Therefore a moral God is unlikely to exist. Again, Marshall doesn’t understand the math. If we must assume an improbable excuse exists for God to act exactly like an utterly, morally indifferent and irresponsible person without actually being one, then the probability God exists goes down, not up.
Take an excuse that’s half as likely as not—and Marshall has presented not even one excuse that comes anywhere remotely near as likely as that—that excuse will reduce the probability God exists by 50%. If P(God) starts at 50% and evidence contradicts “God,” but we can get that evidence to fit by adding a supposition not in evidence that has a 50% chance of being true, P(God) becomes the conjunction of P(God) and P(Excuse), and that means, by the laws of dependent probability, P(God&Excuse) = 50% x 50% = 25%. So rather than getting God to be likely, we just made it less likely. And since excuses are not even 50% likely but, as I said, millions to one against, P(God) is millions to one against (unless God is fundamentally devoid of moral character, but there’s no reason to believe in that god either).
Excuses therefore can never rescue belief in God. Unless and until you have evidence those excuses probably really obtain for God.  Marshall has produced no such evidence. Therefore his God probably does not exist.
That’s an unavoidable mathematical fact.
Dr. Marshall falsely claims humans have “always” had right moral ideas. His own Bible refutes this; likewise world over. Good ideas were always mixed with bad, and no divine guidance reformed or corrected anyone. Moralities only improved painfully slowly from trial and error over thousands of years. Exactly as would happen without God; not at all likely with one.
I’ve referenced many examples.  Including God’s vile command that we murder anyone who exercises their freedom of speech, religion, or sexual autonomy (Deuteronomy 12:1-13:16, Leviticus 24:11-16, Numbers 15:32-36; Deuteronomy 22:13-30, Leviticus 20:13), the Lord’s recommendation to mutilate yourself for having disapproved thoughts (Matthew 18:8-9), and God’s failure to correct the New Testament commandment to subjugate women (1 Timothy 2:11-15). And more.
Just take God’s commandment to engage in chattel slavery (Leviticus 25:44-46). That God would allow anyone to claim he said this, rather than that he damned it, is itself immoral. Yet nothing can have stopped God from correcting it. Therefore there can be no moral God.
Dr. Marshall gets the science of moral character wrong, incorrectly claiming disciplined environments impair growth of moral character, thus impugning all effective models of parenting and policing. 
Just because the Danish far better police violence than the United States or Honduras doesn’t mean they’re more undeveloped morally; they’ve the same or greater moral conscience (e.g. they’re much more charitable in taking care of their poor and downtrodden). Survey nation by nation: the more poorly policed breed the greatest number of the criminal, violent, selfish, and corrupt. Ergo God’s helping us police ourselves won’t damage our character.
Instead Dr. Marshall obsesses over trivial things (like gossip) and absurdly disproportionate responses (like inflicting disease), thus not answering anything I’ve actually been talking about.
Marshall also incorrectly claims:
- I didn’t cite evidence that aesthetic responses actually improve differential reproductive success. 
- I didn’t allow God might have a reason for fine-tuning, missing my point, that such excuses still reduce, not increase P(God), just as in the “moral excuses” case.
- That there’s a “preponderance of evidence” against past eternality, when I just got done showing there wasn’t. 
- That I argued past eternality evinces atheism. 
And yet, Dr. Marshall concedes we don’t “expect God to allow evil and suffering” and “this constitutes…evidence for atheism.” So. The current status of this debate: he’s presented no evidence for theism; and concedes I’ve presented evidence against it.
Such is my latest response to Marshall.
Continue now to read Marshall’s next answer.
 See the sources I cited on this fact in my last entry, in endnotes 1 and 2; and my gremlins analogy under the heading “The Logic of Evidence.” Note that P(h) is the probability of the proposal h before considering any evidence; P(e|h) is the probability the evidence e would exist if h is true; and P(e|~h) is the probability the evidence e would exist if h is false.
 See e.g. Satoshi Kanazawa, “Why Do We Believe in God? II” Psychology Today (28 March 2008); Malcolm Murray, “Error Theory,” in The Atheist’s Primer (Broadview 2010), pp. 218-26; J.D. Eller, Introducing Anthropology of Religion, 2nd ed. (Routledge 2014), pp. 1-49 (esp. pp. 15-26); Stuart Vyse, Believing in Magic (Oxford University 2013); and my bibliography in Sense and Goodness without God, p. 272. Also see “List of Cognitive Biases” at Wikipedia for the whole range of how humans regularly draw logically invalid conclusions from available evidence.
 See Steven Pinker, The Better Angels of Our Nature: Why Violence Has Declined (Penguin 2012); Michael Shermer, The Moral Arc (Henry Holt 2015); and Phil Zuckerman, Society without God: What the Least Religious Nations Can Tell Us about Contentment (New York: New York University Press, 2008). As well as: Gregory Paul, “The Chronic Dependence of Popular Religiosity upon Dysfunctional Psychosociological Conditions,” Evolutionary Psychology 7, no. 3 (2009): 398–441, and “Cross-National Correlations of Quantifiable Societal Health with Popular Religiosity and Secularism in the Prosperous Democracies: A First Look,” Journal of Religion and Society 7 (2005).
 See, again, Richard Carrier, The Will of God: 24 Evil Old Testament Verses and “That Christian Nation Nonsense (Gods Bless Our Pagan Nation)” (12 March 2013) as well as “The Real War on Christmas: The Fact That Christmas Is Better Than Christ” (23 December 2016).
 See, e.g., Walter Sinnott-Armstrong, ed., Moral Psychology, 5 vols. (MIT Press 2007-), 1:209–67 and 3:297–370 and Darcia Narvaez and Daniel Lapsley, Personality, Identity, and Character: Explorations in Moral Psychology (Cambridge University 2009); likewise, e.g., Monika Keller, Wolfgang Edelstein, Christine Schmid, Fu-xi Fang, and Ge Fang, “Reasoning about Responsibilities and Obligations in Close Relationships: A Comparison across Two Cultures,” Developmental Psychology 34.4 (1998): 731–41 and Nancy Eisenberg, Klaus Boehnke, Petra Schuler, and Rainer Silbereisen, “The Development of Prosocial Behavior and Cognitions in German Children,” Journal of Cross-Cultural Psychology 16.1 (March 1985): 69–82.
 Richard Carrier, Final Reply on the KCA (31 May 2019).
 “Marshall thus confuses the invalidity of one argument for God (the KCA) as evidence against God” and “we don’t know for sure if existence ultimately began, and agree that if it did, its uncaused first cause was not probably a god,” Richard Carrier, “My Seventh Reply” (6 June 2019).