Saturday, July 1, 2017

The Argument from Moral Knowledge

Home  >  Philosophy  >  Atheism/Theism

Moral epistemology asks questions like, “How do we know our moral beliefs are true?” and it’s a question that helps theism. What I’ll call “the argument from moral knowledge” tries to use moral knowledge as an argument for God’s existence. Thus a variety of the argument from moral knowledge goes like this:
  1. If God does not exist, then moral knowledge does not exist.
  2. Moral knowledge does exist (e.g. we know it’s morally wrong to torture infants just for fun).
  3. Therefore, God exists.
Where warrant is that stuff we add to true belief to get knowledge (for example, the warrant for belief might be some type of evidence or justification for that belief), this argument claims that if God does not exist then our belief in morality’s existence is not warranted.

Defining Morality



Since I’ve seen some nontheists redefine morality to avoid some of the problems atheism has with morality as it’s ordinarily defined, I’ll briefly sketch what I mean by “morality” in this argument from moral knowledge. Morality has an “ought” component to it (e.g. moral obligations) but there are a couple different senses of the word “ought.” By “descriptive ought” I mean that type of ought that is nothing more than some purely descriptive state of affairs, e.g. “If you want to do well in school, you ought to study” meaning something like “As a matter of practical necessity, you need to study to do well in school.” A descriptive ought is any ought that has no properties besides purely descriptive ones. By “prescriptive ought” I mean that type of ought that is not a descriptive ought; it prescribes and is not purely descriptive, e.g. “You should not to torture infants just for fun.” The type of morality being referred to here involves the prescriptive ought, e.g. moral wrongness. An action is morally wrong for someone only if they ought not to do it in the “prescriptive ought” sense.

Moral oughts being prescriptive oughts is important because it means that moral facts and properties are, if not supernatural, at least non-natural. Natural facts are facts that can be expressed entirely in the language of psychology and the natural sciences, e.g. “The atomic number of gold is 79” and “Bob needs an anesthetic to not feel pain.” (Caveat: don’t confuse the natural fact with the statement used to express what the natural fact is.) The language of psychology and the natural sciences is purely descriptive language, and thus what natural facts are can be stated entirely in descriptive language. But moral oughtness is not purely descriptive, i.e. what moral oughtness is cannot be stated entirely in descriptive language (it is false that moral oughts have no properties besides purely descriptive ones). Since natural facts are purely descriptive whereas moral ought facts are not, moral ought facts are not natural facts, and thus however morality exists, it does not exist as solely part of the natural realm; e.g. moral wrongness is nonphysical and non-natural. We can summarize the reasoning as follows:
  1. All natural facts are purely descriptive (they can be stated entirely in descriptive language).
  2. Moral ought facts are not purely descriptive (it is false that moral oughts have no properties besides purely descriptive ones).
  3. Therefore, moral ought facts are not natural facts.
The above argument is deductively valid, i.e. the conclusion follows necessarily from the premises by the rules of logic, and symbolic logic proves this.[1]

Morality is Empirically Undetectable and Causally Inert



Another indication that moral properties are non-natural is their empirical undetectability and causal inertness, at least barring the supernatural. To illustrate what I mean by moral oughtness being empirically undetectable, imagine a moral nihilist (who disbelieves in moral oughtness) and a moral realist (who believes in moral oughtness) observe some jerk kicking a dog just for fun; the dog whimpers in pain and runs away. Both agree on all physiological and psychological facts, e.g. that the dog felt pain and suffered minor injury. The moral nihilist says, “I don’t think moral oughtness (like moral wrongness) is attached to that action.” The moral realist says, “I think moral oughtness (moral wrongness) is attached to that action.”

There is no empirical way to determine who is right here. Both views agree on all the same empirically observable facts, and moral oughtness being associated with certain physical facts (such as the dog’s injury) doesn’t explain any physical fact. Moral realism and moral nihilism are empirically indistinguishable from each other, though since moral oughtness is non-natural and nonphysical, it isn’t surprising that both the moral realist and the moral nihilist would agree on all the same physical facts.

The dog kicker scenario also illustrates that moral oughtness is causally inert, at least barring the supernatural. Moral oughtness is non-natural and nonphysical, and something non-natural and nonphysical causally influencing the natural, physical realm would be supernatural. Notice also that whether moral oughtness is associated with natural facts (e.g. the dog’s pain and injury) or not associated with those same natural facts, the physical conditions are the same in both cases, and so barring the supernatural, the presence or absence of moral oughtness would make no causal difference as to what those physical conditions will result in.

Moral Awareness



But if moral properties like moral wrongness are empirically undetectable and causally inert (barring the supernatural) how in the world are we aware of morality? Since morality is empirically undetectable, in practice we have to rely on intuition, i.e. it just seems true to us that morality exists[2], similar to how it just seems true to us that the external world is real (as opposed to being something like a dream). But if we have to rely on moral intuition to justify our belief in morality, how does moral intuition deliver warrant for morality’s existence?

Theism has a straightforward explanation: God, wanting humans to have moral knowledge, designed our cognitive faculties (via evolution or otherwise) in such a way that when they are functioning properly we intuitively apprehend basic moral truths, just as we intuitively apprehend elementary truths of logic and arithmetic. Our moral awareness thus becomes a kind of God-given intuitive knowledge.

In contrast, atheism faces real trouble. Moral properties are causally inert in the absence of the supernatural, and because on atheism it is unlikely that supernatural intervention is responsible for our moral knowledge (e.g. supernatural clairvoyance seems far-fetched), it is likely that moral oughtness is causally inert if atheism is true, at least with respect to moral knowledge. On atheism, we have the intuition of morality’s existence because that’s what our brains give us, and moral oughtness being causally inert means its presence or absence would make no causal difference as to whether our brains would give us this moral intuition (nor would it make any causal difference to the evolutionary and environmental processes that gave us our brains). But if on atheism we’d have the intuition of morality existing even if morality did not exist, then it seems we wouldn’t know that morality exists.

To illustrate why, suppose a cyborg knows she has a metal-detecting implant installed in her brain that’s designed so that when a widget is in her hand, the implant delivers a strong intuition that the widget contains metal if and only if it contains metal. Suppose however the metal-detecting implant malfunctions such that it would deliver the intuition that the widget contains metal regardless of whether the widget contained metal. Then even if the widget in her hand did contain metal and she believed it contained metal on the basis of her intuition, her belief wouldn’t count as knowledge. Moreover, if she learned the metal-detecting implant would give her the intuition that the widget contains metal regardless of whether the widget contained metal, she would no longer have adequate grounds to believe the widget contains metal.

In the cyborg scenario (whereby the implant supplies the intuition that the widget contains metal regardless of whether such metal existed), the cyborg’s intuition clearly does not deliver warrant for the metallic widget belief. Similarly, on the “atheism scenario,” whereby atheism is true and our brains would supply moral intuition regardless of whether morality existed, our moral intuition would not deliver warrant for morality’s existence. But since we rely on moral intuition to justify our belief in morality’s existence, it seems that if atheism is true we are not warranted in believing that morality exists. We can summarize the reasoning in two steps. The first:
  1. In the cyborg scenario, the cyborg’s intuition does not deliver warrant for the metallic widget belief.
  2. If the cyborg’s intuition does not deliver warrant in the cyborg scenario, then if atheism is true moral intuition does not deliver warrant for morality’s existence. (No relevant difference seems to exist between the cyborg scenario and the atheism scenario whereby the cyborg’s intuition would not deliver warrant but moral intuition would in the atheism scenario.)
  3. Therefore, if atheism is true, moral intuition does not deliver warrant for morality’s existence.
With If atheism is true, moral intuition does not deliver warrant for morality’s existence argued for, we are ready for the second argument:
  1. If atheism is true, moral intuition does not deliver warrant for morality’s existence.
  2. We rely on moral intuition to believe that morality exists, such that If moral intuition does not deliver warrant for morality’s existence, we are not warranted in believing that morality exists is true.
  3. Therefore, if atheism is true, we are not warranted in believing that morality exists.
It seems then we have good grounds for affirming the first premise of the argument from moral knowledge:
  1. If God does not exist, then moral knowledge does not exist.
  2. Moral knowledge does exist (e.g. we know it’s morally wrong to torture infants just for fun).
  3. Therefore, God exists.
Our awareness of morality’s existence thus seems to provide evidence for theism.





[1] For philosophy nerds (and all philosophy nerds should be sufficiently versed in symbolic logic) here’s the symbolization key:
  • Nx = x is a natural fact
  • Dx = x is purely descriptive
  • Mx = x is a moral ought fact
The goal is to prove ¬∃x[Mx ∧ Nx], i.e. that there does not exist a moral ought fact that is also a natural fact, from premises (1) and (2) in the argument below:
  1. ∀x[Nx → Dx]
  2. ∀x[Mx → ¬Dx]

  1. ∃x[Mx ∧ Nx] indirect proof assumption
    1. Mf ∧ Nf 3, existential instantiation
    2. Mf 4, conjunction elimination
    3. Nf 4, conjunction elimination
    4. Nf → Df 1, universal instantiation
    5. Df 6, 7, modus ponens
    6. Mf → ¬Df 2, universal instantiation
    7. ¬Df 5, 9, modus ponens
    8. Df ∧ ¬Df 8, 10, conjunction introduction
  1. ¬∃x[Mx ∧ Nx] 3-11, indirect proof

[2] I don’t mean to imply that moral intuition necessarily gives us knowledge of morality’s existence by directly telling us that morality exists (though I don’t rule it out either). It could also be, for example, that we intuit a moral truth like, “It is morally wrong for a man to torture innocent sentient life just for fun” and from there we conclude that morality exists.

6 comments:

  1. Okay, so we can spell out the error in Wade's reason pretty clearly, I think, and it boils down to how he justifies premise 2: the claim that he does indeed have moral knowledge.

    Wade's argument rests on the claim that justification for moral knowledge comes from moral intuition.

    However, he also claims that, if atheism is true, moral intuition doesn't provide justification for moral knowledge.

    When we come to premise 2, then, Wade is stuck in an intractable position:

    On the one hand, he must claim that his moral intuition is a sufficient justification for p2.

    On the other hand, he must claim that his moral intuition *isn't* a sufficient justification for p2, *if atheism is true.*

    In order to have sufficient justification for p2, then, Wade must both appeal to his moral intuition *and* claim, in the step of justifying p2, that atheism is false.

    But, wait: that atheism is false is the conclusion of his argument. Hence, his justification for p2 requires him to beg the question.

    ReplyDelete
    Replies
    1. That's not what begging the question is. Begging the question is when the only reason to believe the premise is if one already accepts the conclusion. Moral intuition justifying premise (2) does not fit that category.

      However, he also claims that, if atheism is true, moral intuition doesn't provide justification for moral knowledge.

      Right, so justifying the first premise and justifying the second premise means one has to accept the conclusion, but that's not question begging; that's deductive validity (i.e. the conclusion following from the premises).

      Delete
    2. The problem is that pointing to your moral intuition is not, on its own, a sufficient justification. If it were, your first premise would be undermined by that fact.

      So, you can't just say, "the fact that I have this moral intuition justifies p2." You have to add something else to that--some qualifier which prevents your justification for p2 from undermining your p1.

      And that's where you end up begging the question.

      Delete
  2. Wade is reticent to spell out his justification for p2, but he has given us enough that we can go though his evaluation for him.

    First, he stares in the article that our justification for moral knowledge comes from our moral intuition, so we can safely assume that he would rest his defense of p2 on this intuition.

    However, his first premise calls this very intuition into question. Specifically, he states that, if atheism is true (or, equivalently, if this is false or if God does not exist) then moral intuition does not provide justification for moral knowledge.

    So, does Wade's moral intuition provide justification for p2? Since he has called the reliability of moral intuition into question with his first premise, this is a question we must consider when we get to his second premise.

    Let M be the proposition that moral intuition does indeed provide justification for moral knowledge.

    Let A be atheism (as used above, which is how Wade has used it in this argument)

    Wade's first premise requires that P(M|A) < 0.5

    The total probability of M is

    P(M|A)*P(A)+P(M|T)*P(T)

    As far as Wade's argument goes, the most charitable values to use for P(M|A) and P(M|T) are 0.5 and 1, respectively.

    He needs both to be high, but if P(M|A) > 0.5 then his first premise fails. So, these are the highest values.

    Since Wade doesn't want to beg the question by assuming that theism is probably true, we'll put the priors for A and T equivalent to each other.

    P(M) = P(A) = 0.5

    So, now, the total probability of M--under the best circumstances for Wade's argument --is P(M) = 0.75

    Seems good, right?

    Unfortunately, no. See, the probability that Wade's argument is sound is equal to the probability of P1 plus the probability of P2, minus one.

    The probability of P2, under these maximally charitable assumptions, is 0.75, and the probability of P1 is also 0.75.

    The probability that Wade's argument is sound, then (again, under these assumptions) is 0.5, exactly.

    The problem for Wade is that we could easily establish that these assumptions are far too generous to Wade's argument.

    For example, there is some chance that This is true, but that our moral into are still not reliable. Hence, P(M|T) = 1 is an overestimate, and if we reduce this value, the probability that Wade's argument is sound falls with it.
    (Note that reducing the value of P(M|A) doesn't help Wade either*)

    In fact, the only way that Wade can salvage his argument from itself is to insist, in the step where he justifies his second premise, that Theism is probably true.

    However, the fact that he has to insist on the truth of theism in order to justify his second premise with the confidence the argument requires means that he must beg the question in order to prosecute his argument successfully.

    *If, for instance, P(M|A) = 0.1, then P(M) = 0.55. the probability of premise 1 would then be 0.95, and the probability that the argument is sound would still be 0.5.

    In the end, it is simply impossible for Wade to justify his second premise without contradicting his first premise or begging the question.

    ReplyDelete
    Replies
    1. I disagree with how you've assigned the probabilities. I prefer a more direct approach. Let A be atheism (the negation of theism), ¬A be the truth of theism, M be "moral knowledge exists" and ¬M be "moral knowledge does not exist." Personally, I'd put P(~M|A) ≈ 0.92 and P(M) ≈ 0.98. Given that, the probability that both premises are true has a lower limit of 90%. I see no reason why one should favor your approach over mine.

      Delete
    2. You're not actually offering a different approach, here. This "approach" produces exactly the same result that I point to above.

      Delete