| Does Objective Morality Exist If God Does Not Exist? |
| < Prev | 1 | 2 | 3 | Next > |
The Argument from the Unbiased Atheist
To support the claim that the evidence suggests the first premise of the moral argument is probably true, consider what I’ll call the Argument from the Unbiased Atheist (AftUA), where the Unbiased Atheist is someone who has no intuitions of moral oughtness existing or not existing, including not intuiting propositions that have moral oughtness’s existence implicit in them (e.g. the Unbiased Atheist wouldn’t have the intuition that Something is morally wrong is true or probably true), and she has no preconceived opinions about whether morality exists.[2] With the following symbolization key:
| M = | Moral oughtness exists. |
| A = | Atheism is true. |
| K = | The relevant background data. |
| Pr(not-M|A&K) = | The probability of M being false given A&K, i.e. the probability of moral oughtness not existing given atheism and the relevant background data. |
Presumably, there are objective probability facts regarding the evidential relationship between data and belief, e.g. “On the basis of the scientific data, quarks probably exist” is objectively true, and this is the sort of probability I have in mind when I say, “Given atheism, moral oughtness probably doesn’t exist” (viz. the objective evidential relationship between “atheism + background data” and “moral oughtness exists”). Because of the Unbiased Atheist’s neutral and unbiased nature, I think she provides a good vantage point to evaluate the objective evidential relation between the data (A&K) and the belief (M). The Unbiased Atheist starts out believing A&K but because of her neutral and unbiased nature, K doesn’t itself include propositions like M is probably true or M is false.
My overall argument can summarized thusly:
| (P1) | If the Unbiased Atheist would be justified in believing Pr(not-M|A&K) is high, then Pr(not-M|A&K) is high. |
| (P2) | The Unbiased Atheist would be justified in believing Pr(not-M|A&K) is high. |
| (C1) | Therefore, Pr(not-M|A&K) is high. |
| (P3) | If Pr(not-M|A&K) is high, then Given atheism, moral oughtness probably doesn't exist is true. |
| (P4) | If Given atheism, moral oughtness probably doesn't exist, then the first premise of the moral argument is probably true. |
| (C2) | Therefore, the first premise of the moral argument is probably true. |
Lines (P1), (P2), and (C1) are more or less the Argument from the Unbiased Atheist (AftUA). The main purpose of the AftUA is to show that on atheism, people’s intuitions of moral oughtness existing are probably delusory and not veridical, which of course implies Pr(not-M|A&K).
(P1) seems true due to what K can and can’t include, and the fact that Pr(not-M|A&K) is about the objective evidential relation between not-M and A&K. To see my point of view on why I think the Unbiased Atheist provides a good vantage point for this sort of probability, suppose an atheist proposes a probabilistic argument from evil, arguing that Pr(not-God|Evil&K) is high where K is some appropriate background knowledge (including e.g. free will for the theist who attempts a free will theodicy).[3] Suppose I argue that since belief in God is a rational intuition, we should include it in K, and when we do we effectively get Pr(not-God|Evil&God) which of course is extremely low, and thus Pr(not-God|Evil&K) is low. Therefore, I argue, the probabilistic argument from evil fails. Even though I think the intuition that God exists is a rational intuition, this maneuver does not seem philosophically appropriate. The intuition of God’s existence, while perhaps rational, is a subjectively experienced intuition, and the objective evidence would still be strongly against theism if Pr(not-M|Evil&K) were high when propositions like God exists are excluded from K. Similarly, I’m offering a sort of “probabilistic argument from atheism” against moral objectivism, arguing that atheism makes moral objectivism unlikely, and that Pr(not-M|A&K) is high when K doesn’t include propositions like Morality exists, though K would include propositions like People have intuitions of morality existing.
Line (P3) is true by definition, since what I mean by Given atheism, moral oughtness probably doesn't exist just is Pr(not-M|A&K) is high. Line (P4) is supported by a mathematical theorem.
My support for (P2) is the following inductive argument (with each premise predicated with, “If atheism is true, this is true:”), where objective moral oughtness is abbreviated as OMO.
- There is zero empirical evidence for objective moral properties. This supports the idea that the Unbiased Atheist has no good reason to accept M.
- It’d be a remarkable coincidence if moral intuitions happened to line up with what these invisible, causally inert moral properties are really like. Such reliance on remarkable coincidence suggests that we wouldn’t have real knowledge of objective moral truths; at best we’d have coincidentally true beliefs.
- OMO properties are suspiciously queer, akin to invisible and nonphysical gods. To illustrate the general idea behind the Argument from Queerness (AfQ), suppose someone claims there is an invisible unicorn floating above my head. This claim is possible, but not plausible. I would be justified in disbelieving in this unicorn. The unicorn is “queer” enough to be prima facia implausible, and we are prima facia justified in rejecting its existence. Moral oughtness is invisible, nonphysical, empirically detectable, and causally inert. To the Unbiased Atheist, moral oughtness likewise seems “queer,” giving her prima facia justification for disbelieving its existence.
- Evolution occasionally gives false beliefs (e.g. gods). So there’s precedent for evolution giving humans delusory intuitions for invisible nonphysical things. And belief in gods potentially serves some evolutionary purpose: “Don’t do stuff that harms the group even if we’re not watching because the gods are watching and they’ll punish you for doing bad stuff.”
- Moral oughtness beliefs have evolutionary value whether true or not. Suspiciously enough, belief in moral oughtness is kind of like the false belief in gods in potentially serving some evolutionary purpose: to get us to behave in the right ways. Such beliefs have evolutionary value regardless of whether moral oughtness exists.
- Our best theory for why we believe in moral oughtness doesn’t require its existence. Moral oughtness is causally inert; its presence or absence would have no effect on whether we’d get moral intuitions, and so the processes that gave us moral intuitions would do so regardless of whether morality existed. Consequently, we don’t need to posit something so extravagant as these invisible and highly metaphysical moral properties to explain moral beliefs; we can just say it’s a trick of evolution to get us to behave in certain ways.
Note: I am not claiming that any single premise by itself is enough to justify the conclusion, but I do think the combination of all six premises justifies the conclusion.
One of my key claims that the AftUA illustrates is that on atheism the objective evidence suggests that people’s intuitions of M are probably delusory. In some cases our justification for beliefs comes from non-evidential grounds. For example, consider for example the following brain-in-vat (BIV) hypothesis: you are recently created (say, within the past five years) brain in a vat hooked up to a supercomputer feeding you all the memories, sense experiences, and intuitions you know have. We can’t justify our belief that BIV is false based on evidence, because there can’t be any evidence against it.[4] Instead it’s our intuition of its falsity that justifies our belief that BIV is false and that this perceived reality is real. The atheist moral objectivist (AMO) could similarly argue that he’s justified in believing OMO based on non-evidential intuition grounds, even if Pr(not-M|A&K) is high. But a crucial disanalogy is that in the case of BIV we don’t have any objective evidence suggesting BIV is true, whereas on atheism the objective evidence suggests that people’s intuitions of moral oughtness existing are probably delusory. This makes the a priori intuition option much less viable for the atheist.
Objections
Objection: moral intuition should affect our assessment of Pr(not-M|A&K)
One could say that although we shouldn’t include a proposition like Morality exists in K, moral intuition gives us nonpropositional support for morality’s existence, i.e. actually experiencing the intuition justifies our belief more than the mere fact that People have the intuition that morality exists (similar to how our intuition of the external world justifies our rejection of the BIV hypothesis). This a priori moral intuition makes us inclined to believe in morality, and this intuition needs to, for lack of a better term, bias our evaluation Pr(not-M|A&K) compared to a more neutral vantage point.
The problem is while that while a priori subjectively experienced intuitions about whether a theory is true might be appropriate for subjective probability assessments, the type of probability the AftUA is using is the objective evidential relation between the belief (moral objectivism) and the data (A&K). To illustrate, suppose that prior to examining the relevant scientific data, Smith has an overpowering intuition that creationism is true, and Jones has an overpowering intuition that evolution is true. Smith and Jones might come away with different subjective probability assessments of how likely evolution is after examining the scientific data, but to correctly evaluate the objective evidential relation between data and theory a more neutral vantage point seems appropriate. Hence the Unbiased Atheist.
In response, one could say that when a person’s intuition is a rational one, that person should use the intuition when evaluating the probability of a theory given the data if she is to assess the overall rationality of the belief. The problem is that even a rational subjectively experienced intuition is still subjectively experienced, and the type of probability the argument is using is nothing more than the objective evidential relation between theory and data, and not the overall rationality of a person’s belief.
To illustrate the difference between the two, suppose Smith and Jones know that in a particular factory 2% of all widgets shaped like a cube contain metal. Let T represent The widget from the factory contains metal and let D be The widget from the factory is shaped like a cube. The objective evidential relation between T and D is such that Pr(T|D) = 0.02, and Pr(not-T|D) = 0.98. Smith is a cyborg who knows she has a metal-detecting implant that gives her the intuition that a widget contains metal whenever a metal-containing widget is in her hand. Unbeknownst to Jones, Smith’s intuition tells her the cube-shaped widget contains metal and she is rational to believe it contains metal, but this does not change the objective evidential relation between T and D, which is still Pr(T|D) = 0.02, since Pr(T|D) represents the probability of the widget containing metal given just the shape of the widget. The overall rationality of cyborg Smith’s belief is such that she is rational in thinking the widget contains metal despite Pr(not-T|D) being high. In contrast, Jones (who is akin to the Unbiased Atheist) would be justified in thinking the widget does not contain metal because he doesn't have Smith's intuition, and is wholly reliant on the objective evidential relation between T and D.
In some cases it is possible for rational intuition to be relevant in evaluating the objective evidential relation between data and theory, e.g. intuiting principles of inductive inference. But the a priori intuition that moral oughtness is associated with certain natural properties is no more a rule of inductive inference than the a priori intuition that God’s existence is currently associated with the universe’s existence. The subjectively experienced intuition of morality’s existence, while perhaps a rational intuition, does not seem like it should affect our evaluation of Pr(not-M|A&K) when it’s the objective evidential relation between not-M and A&K that we’re trying to determine, and not the overall rationality of the belief.
Objection: even with Pr(not-M|A&K) being high, moral intuition still justifies moral objectivism for the atheist
A more promising approach is to bite the bullet and say that Pr(not-M|A&K) is high but only in the sense of an objective evidential relationship between not-M and A&K, and not the overall rationality of the belief. Consider the case of cyborg Smith. Although Pr(not-T|D) is high in the objective evidential relation sense, cyborg Smith’s intuition nonetheless makes it rational for her to believe the widget probably contains metal. Similarly, even though Pr(not-M|A&K) is high in the objective evidential relation sense, moral intuition still justifies an atheist’s belief that objective morality probably exists. Pr(not-M|A&K) as an objective evidential relation between theory and data ultimately becomes irrelevant, and the atheist moral objectivist is still justified in rejecting the moral argument’s first premise.
The problem comes when A&K encompasses all the relevant data, including the existence of moral intuition, and when this objective evidence suggests that the moral intuition is likely delusory. To illustrate, let’s modify the cyborg Smith scenario as follows: Jones shows Smith objective evidence suggesting that Smith’s metal-detecting implant would give Smith the intuition that the widget contains metal even if the widget did not contain metal, making Pr(not-T|D&K) high. Cyborg Smith’s intuition would no longer provide adequate justification, and Pr(not-T|D&K) being high gives her strong rational grounds to believe the widget does not contain metal. Similarly, on atheism the objective evidence suggests we would have the moral intuition even if morality did not exist. As mentioned earlier, moral oughtness is causally inert; its presence or absence would have no effect on what intuitions our brains would give us, nor would it have any effect on the evolutionary and environmental processes that gave us our brains. On atheism, this would seem to undercut our moral intuition as a source of adequate justification, and Pr(not-M|A&K) being high does seem to strongly justify the moral argument’s first premise.
Conclusion
Moral oughtness is invisible, nonphysical, causally inert, and empirically undetectable. So if atheism is true, why shouldn’t the consistent atheist reject the existence of this invisible nonphysical thing that cannot be empirically detected, if the atheist is to reject the existence of invisible nonphysical deities that have (allegedly) not been empirically detected? Given atheism, it seems more likely that people’s belief in moral oughtness is a delusion brought about by evolution to get us to behave in certain ways and help our species survive. Given atheism, objective morality probably doesn’t exist, which suggests that the moral argument’s first premise is probably true.
In providing a neutral, unbiased vantage point from which to evaluate the objective evidential relation between not-M and A&K, the Argument from the Unbiased Atheist helps show that given atheism, objective moral oughtness probably doesn’t exist, thereby suggesting that the first premise of the moral argument is probably true. We have a sort of “probabilistic argument from atheism” against moral objectivism to support the moral argument’s first premise, and I don’t see good way to refute it.
< Prev | 1 | 2 | 3 | Next >
[2] Wade A. Tisthammer created the Argument from the Unbiased Atheist (I mention this because I have been asked who came up with it).
[3] For the record, I do not believe that Pr(not-God|Evil&K) is high, but the hypothetical scenario of it being high does serve as a useful illustration.
[4] This can be shown with the help of mathematics (given that the sort of “evidence” we have in mind is something that makes a hypothesis more or less likely); see p. 2 of Why evidentialism sucks.
[3] For the record, I do not believe that Pr(not-God|Evil&K) is high, but the hypothetical scenario of it being high does serve as a useful illustration.
[4] This can be shown with the help of mathematics (given that the sort of “evidence” we have in mind is something that makes a hypothesis more or less likely); see p. 2 of Why evidentialism sucks.
Comment on this article
