I’m going to briefly probe some of the logic behind Christian morality. I’d like to extend an invitation to anyone interested to contribute their thoughts by commenting below; I will try to periodically update this post to incorporate any progress we’ve made on this question.

I’m going to simulate a conversation between a Christian C and an atheist A, not unlike a few conversations I’ve had.

AHow do you know what’s right?

[Skipping lots of theology here]

CBecause God tells us what’s right through revelation. 

A: How do we know that God isn’t lying about what is right?

[Notice that this question makes sense regardless of C‘s position on the Euthyphro Dilemma]

C: Because God wouldn’t lie to us.

A: How do you know that God wouldn’t lie?

C: Lying isn’t in God’s nature.

A: Why not?

C: Because God’s nature is good.

A: How do you know that it’s not good to lie?

C: Because God tells us so through revelation.

[Now we have found ourselves back where we started]

A: But we haven’t yet established that God’s revelation is trustworthy. You have fallen into circular reasoning. 

Please note that it is not my intention to suggest that the atheist has better responses to these questions than the theist. Rather, my intention is to disprove the claim that the Christian has better responses to these questions than the atheist. It is nearly universally thought (at least outside of formal philosophy circles) that Christianity has a ‘better’ explanation of morality than what atheism can offer. I think this is not the case. Building a coherent moral theory is a fundamentally difficult philosophical problem, and Christianity doesn’t offer any workarounds for this difficulty, as far as I can tell.

I’d like to add that there is an obvious alternate ending to the above conversation, but it, too, demonstrates my point:

A: How do you know that it’s not good to lie?

C: Because [insert secular moral rationale]
It’s not at all my intention to caricature either side of the conversation. If you find me misrepresenting the views of either side, please offer a correction via the comments. Also, if you would like to propose an extension of the conversation, like, for example, a potential response from the Christian side, please do. If this problem has a solution I’m not aware of, I’d very much like to know.

18 thoughts on “How do we know what’s right? An examination from inside the Christian worldview.

  1. Derry says:

    Marko, the argument you’ve presented is indeed a clear case of circular reasoning! And the point you landed on, that a case was not made for the trustworthiness of God’s revelation, is accurate.

    To avoid this circular argument and provide a basis for the reliability of scripture, we can turn to a number of ancient, non-Christian, independent sources. These are historical eyewitnesses to the events related to Jesus’ life: Tacitus (Roman historian), Josephus (first century Jewish historian), Lucian (second century Greek satirist), Seutonius (Roman historian), and Pliny the Younger (Roman governor in Asia Minor) among others. From their accounts alone, we can reconstruct the life of Jesus: that he lived in the early first century, was from Nazareth, had a brother named James, was considered a “wise king”, was executed by Pontius Pilate before 70AD, was considered a ‘law-breaker’ and a ‘miracle worker’, and that after his death, that many claimed to have seen him alive and were willing to die for what they believed they had seen. The New Testament writings corroborate these external, nonbiblical accounts.

    These accounts also show us that the early church formed around the person of Jesus, a rapidly growing religion based on the worship of a man who had been crucified as a criminal! The best explanation for this is if Jesus actually rose from the dead like he said he would. His resurrection authenticates his ministry, including the works he did as well as his words. He repeatedly affirmed the truths and enduring quality of scripture, saying, “until heaven and earth pass away, not the smallest letter or stroke shall pass from the Law until all is accomplished.” His endorsement of the whole of scripture (both Old and New Testaments) means that he would concur with the Old Testament statement that God does not lie. We’ve arrived at this conclusion not by circular argument, but by a foundational belief -an event in history- the resurrection.

    1. trivialtruths says:

      The problem with this response, though, is that the argument above is independent of whether scripture is reliable. Consider the following:
      Let’s grant that everything in scripture is an accurate representation of what actually occurred a few millennia ago. This doesn’t tell us anything about the reliability of the moral claims in scripture. For example, under these conditions, the Ten Commandments as presented in Exodus really were the Ten Commandments dictated by God. But once again we don’t know whether God was being truthful. To put it in very simple terms, if X is a liar, and some book documents what X says, how accurately the book documents X’s statements doesn’t tell you anything about how accurate X’s statements are. The book itself could be an exact, 100% truthful copy of X’s statements even if the statements themselves are not truthful.

  2. Derry says:

    Your objection in the first post was that the Christian was using circular reasoning to authenticate “truths” in the Bible. It is important to establish that the Bible is reliable because it is a large part of the basis for Christian belief. That was what was attempted in the previous post by using nonbiblical sources and intentionally not using the Bible to authenticate itself. (Other sources of authentication are from the over five thousand manuscript copies and archaeological finds that verify people and places mentioned in the New Testament accounts).

    However, the Bible is not merely reliable as a historical document. It is not like any other book. It’s author makes intentional claims that are meant to apply to every person who has ever lived. It illustrates a picture of a world created good, but corrupted by humankind’s attempt to be autonomous from God, and God’s solution to humanity’s problem.

    Christians have confidence that what the Bible says and claims is trustworthy. This is based on Jesus’ demonstration by his life, death, and resurrection, that he was God (again, facts about Jesus were established in the external evidence). Jesus claimed to be the “exact representation” of the Father (Hebrews 1:3, John 14:9), and the “image of the invisible God” (Colossians 1:15), who performs the exact will of the Father (John 6:38). If Jesus did not lie, at least that we are aware of, then we are justified in believing that God did not lie, which can be further supported by scripture (Numbers 23:19, Hebrews 6:18, Titus 1:2).

    If Jesus was indeed God in the flesh as He claimed and demonstrated, the words he uttered would be significant. In the gospel accounts, he repeatedly authenticated the words of the Old Testament. One notable example that applies here is in Matthew 5:17 where Jesus said, “Do not think that I have come to abolish the Law or the Prophets; I have not come to abolish them but to fulfill them,” meaning that Jesus believed and trusted in the Old Testament “Law” and “Prophets.” He upheld the authority of the moral law as given by God in the Old Testament.

    What evidence is there that God has lied about morality? What good reasons are there for entertaining such an idea? Until we have a reason to think it is probable that God has lied, there is no reason to think that he has.

    1. trivialtruths says:

      I think the point you make is interesting and worth inspecting closely—in particular, the last sentence: “Until we have a reason to think it is probable that God has lied, there is no reason to think that he has.” Indeed, I agree. But there is also no reason to think that he hasn’t. Here the adage occasionally used in apologetics “absence of evidence is not evidence of absence” applies, but doing work for the opposite side of the argument. Given that we don’t know whether God is lying about morality, what is it reasonable to conclude? We are stuck with making a probabilistic conclusion—that perhaps it is 70% likely that God is telling the truth and 30% likely that he is not, for example. But what information can we use to justify granting a higher probability to honesty than to deception? In order to do that, we have to already know something about God’s nature—and the source of that information cannot be God, because we have yet to establish that source as credible. To put this point in vivid terms, suppose for a moment that the ultimate being described in scripture is actually the devil; that the devil is the creator of the universe and thereby the author (indirectly) of all scripture. In all his unconscionable maliciousness, the devil (who describes himself as “God” in scripture) has filled scripture with false moral prescriptions, and sent his son to Earth as a human being to fulfill scriptural prophecies, thereby ensuring that a large number of humans believe the contents scripture, including the false moral prescriptions, dooming the believers to hell, to the devil’s delight. This scenario is indistinguishable from the Christian worldview, is it not? There is absolutely no way for you to tell the difference between the two, since all of scripture, and all of the observable universe for that matter, is exactly identical in the two scenarios. You may quickly brush off this scenario as unlikely, but are you actually justified in doing so? In order for you to state that this scenario is unlikely, you have to already know something about what God’s nature is likely to be. Why is an honest God more plausible than a deceptive God? We must appeal to an inherent property that God himself could not override—for example, if we claim that God cannot be unethical by definition, and then claim that lying is unethical, then we can say for certain that God cannot lie. But first we need a reason to think that lying is unethical, and that reason can’t come from God. Now we find ourselves back in the context of my main argument.

  3. Brent says:

    Hey Marko,

    Fun stuff! A few thoughts:

    Along the lines of Derry’s reasoning, why not flip the whole thing around and, instead of arguing for an honest God on a priori reasons, argue for an honest God on a posteriori reasons? Aka, if we grant that Jesus was God-as-man (a hefty “grant” I understand, but we’re hypothetical here still), then he begins his life at 0 lies and 0 truths. Why not go from there? God’s nature is on display, and we don’t know yet what it is. If we examine Jesus’ words and deeds, counting each time he followed through on a promise, or uttered something we know to be true, and counting each time he deceived another, or uttered something we know to be false, we end with something like 9,000 truths and 0 lies, then we’ve got pretty good a posteriori reasons for thinking the nature of God is to not lie. And the probability that God would lie is much closer to 0 than .3, which gives an answer to your question – “what information can we use to justify granting a higher probability to honesty than to deception?”
    And perhaps if we want to argue on a priori grounds – if we accept the Law of Noncontradiction (LoN) as a necessary truth, then it must be grounded, or sourced, in truth. Or, whatever it is grounded in, cannot be a thing that can even possibly err — otherwise, the LoN cannot be a necessary truth, and even worse, the LoN, which deals exclusively in matters of truth, would not be grounded in truth, which is absurd. This was, IIRC, one of Plato’s arguments against polytheism (the gods are warring, and therefore disagreeing with each other; therefore they cannot be the ground of the LoN — or truth). Obviously, the Christian merely affirms that God is truth (John 17:17, John 14:6), and that He grounds necessary truths (cf. the “logos” in John 1).
    A lot of this has the same ring to it as Descartes’ evil demon, of wanting some sort of Cartesian certainty of things. Sure, it’s possible we’re all being duped (including us Christians). It’s possible that nothing is real, that what is up is down, what is true is really false, etc. But again, a thing’s possibility does not entail its probability. At the end of the day the prudent philosopher (IMO) rests well on having practical certainty of things without having philosophical certainty.

    1. trivialtruths says:

      Practical certainty is fine, and of course is all we can really hope for. The problem is that you haven’t established practical certainty (or even likelihood of any kind) with your argument. Your argument doesn’t apply statistical inference correctly—there are hidden assumptions that mean your conclusion doesn’t actually respond to my argument. For the sake of argument, you grant that Jesus was God-as-man. I am happy to use hypotheticals for the sake of argument, as long as we keep track of what we have assumed. Notice that in your argument, it immediately follows that “God’s nature is on display“. But does this actually follow from Jesus being God-as-man? Jesus being God-as-man implies that Jesus’s behavior is equivalent to God’s behavior (you might dispute this, but notice that disputing it won’t help your argument…if Jesus can behave differently than God, then it becomes even harder to argue that Jesus’s behavior tells you about God’s nature…so I’m giving the charitable interpretation of your statement here. The most work that it can do for the argument is to provide a direct link between Jesus’s behavior and God’s behavior). So now, under your assumption, we can conclude that Jesus’s behavior = God’s behavior. But how do we get to Jesus’s behavior = God’s nature on display? In order to do this, you must assume that God’s behavior = God’s nature on display. Then it follows that Jesus’s behavior = God’s behavior = God’s nature on display. But notice what the second assumption we have introduced here really is. To assume that God’s behavior reflects God’s nature is to assume that God is not deceptive (since a deceptive God would by definition behave contrary to his nature). But my whole argument is that without this exact assumption, a Christian cannot know what is right and wrong. Hence you have used the very assumption that I argued the Christian cannot do without. So the fact that you can proceed to a conclusion with practical certainty isn’t relevant to my argument—because I claim that you can’t do it without using the assumption; using the assumption doesn’t address my claim.

      Before you go on to saying “well we need certain assumptions for everything” (which I think was perhaps part of your intention in the second part of your comment), consider that I could use that reasoning to justify assuming out of hand that God doesn’t exist, as an axiom. I know that we have to use axioms, but we have to have reasons to think that our axioms are extremely likely to be true. I think if anyone were to compare the assumption of the law of noncontradiction with the assumption that an unimaginably powerful being also happens to be honest, they would say that the first requires no justification—it is obviously true, and hence can be used as an axiom without objection, while the second is not obviously true, and is thus worth taking a closer look at. I’m claiming that not only is it not obviously true, but also that we don’t even have any reason to believe it is true with >50% probability.

      Lastly, I’d like to just briefly notice a flaw in the reasoning of the second part of your comment. You state “if we accept _________ as necessarily true, then ______ must be grounded, or sourced, in truth”. I don’t think you actually believe this statement, since you must then be able to insert “God” into the blank space and get a true statement. Without this statement, then you must acknowledge that some necessarily true things do not have to be grounded or sourced in truth. Then what stops me from claiming that those things are just the usual axioms (noncontradiction, excluded middle, yadda yadda yadda)?

      P.S. Sorry if this seems overly forceful. It’s hard to write a counterargument while sounding friendly, but know that this is all in the best intention. Just trying to figure out the nature of reality, as always.

      1. Brent says:

        You are a gentleman and a scholar, sir! This doesn’t seem overly forceful. Ok, since there are several threads starting here, let’s focus on just the first for now. I didn’t intend to assume what we were trying to prove, but really was just quoting the author of Hebrews, “[Jesus] is the radiance of God’s glory and the exact representation of His nature” (Heb 1:3). So it wasn’t intended to be a snuck-in, ungrounded assumption, but rather, I really intended it to be an explicit ground (I meant to note the scripture ref as well but failed — sorry!) My fault. Anyway, I understand the nature of the divine is exactly what is at question. So let’s ignore Heb 1:3, and focus instead solely on Jesus’ behavior.

        Again, if we grant that Jesus is God’s behavior on display, do we not still gain good reasons for thinking that God’s behavior is truthful? Are there examples where Jesus deceived anyone, or told an un-truth? Once again, if we have, say, 1,000 propositions uttered by Jesus, and none of them have been shown to have been deceitful, or false, then does that not lend empirical credibility to the idea that Jesus’s behavior strongly tended toward truth-telling?

        Perhaps you will say “Yes, I can grant for the sake of argument that his record was good. But how do we know he didn’t act truthful during his life just to fake us into thinking his nature is truthful, when all along his nature is deceitful?” This is where I will reiterate that while such a thing is possible, I don’t see a single reason to think it’s probable. I am open to evidence which might suggest otherwise.

        The remaining question is, can we make an inferential leap from behavior to nature? Once again, I don’t see why not. We seem to do such things all the time. If I am acquainted with a person who I know to behave in deceitful ways, I often make the leap — and prudentially I might add — that such a person’s character, or nature, is poor, or deceitful. Likewise, if I am acquainted with a person who I know to consistently behave in virtuous ways, I often make the leap — and typically prudentially so — that such a person’s character is good. This doesn’t mean I’m right 100% of the time, but usually, I think, I end up being a pretty good judge of character. And I think most of us know what it means to be good “judges of character”: we infer from a person’s behavior to their nature.

        So why does the same rule not apply to Jesus, if he actually was God-as-man?

        1. trivialtruths says:

          Apologies for these increasingly long responses, but we are trying to get to the bottom of this so we really have to get in the weeds here.

          There is in fact a good reason why that inferential leap is reasonable for use in the judgement of the character of other human beings and unreasonable for use in the judgement of the character of God. It has to do with the logic of inference itself—so let me offer a technical explanation using Bayes’ Theorem, and then I will summarize in plain language afterward, in case you find one explanation more convincing than the other.

          Bayes’ Theorem relates the probability that a hypothesis H is true before evidence E is observed and after evidence E is observed, which is why it is essential for inference. The theorem states that the probability that H is true given that E is observed, written Pr(H|E), is equal to the probability of observing E given that H is true, written Pr(E|H), multiplied by the a-priori probability that H is true, written Pr(H), and divided by the a-priori probability of observing E, written Pr(E). Putting this all together, we have:

          Pr(H|E) = Pr(E|H) x Pr(H) / Pr(E).

          Intuitively, this theorem makes sense: if E is very likely to occur under hypothesis H, that increases our confidence that H is true if we see E; if the hypothesis is a-priori very likely to be true, that also increases our confidence that H is true; and if we are very likely to observe E regardless of whether H is true, that decreases our confidence that H is true if we see E.

          Now suppose we make an observation E, and we want to compare two hypotheses H1 and H2, in order to see which one is more likely to be true given that we have observed E. Then we can use Bayes’ Theorem to express the ratio of the probability that H1 is true to the probability that H2 is true (given that E was observed):

          Pr(H1|E) / Pr(H2|E) = Pr(E|H1) x Pr(H1) / (Pr(E|H2) x Pr(H2)).

          Let’s apply this to our question—which hypothesis is more likely to be true: my hypothesis that God is deceptive, or your hypothesis that God is truthful (and we are talking about God’s moral teachings specifically here). We have to be specific in defining our hypotheses, so I’m going to define them the following way (this may not match exactly what you had in mind, but it should still demonstrate my point regardless): let D be the hypothesis that God is deceptive—specifically, that God has intentionally conceived of the Christian religion to mislead people morally—so that they go to hell, thinking that they were doing what was required in order to go to heaven. Under D, God has intentionally sent Jesus to Earth to convince humans to become Christians (and hence made Jesus as convincing as possible using miracles, prophecy-fulfilling, etc), among other actions meant to deceive humans. Now let T be the hypothesis that God is truthful—essentially the opposite of D; specifically, that God has been accurately conveying the true nature of morality to humans with all of his actions (including but not limited to the actions of Jesus and the related miracles & fulfilled prophecies). I’m going to use ‘J’ to denote our observations that Jesus has never told a lie as far as we can tell (again, you may prefer slightly different definitions, but my argument will easily generalize and the point will be communicated nonetheless).

          Now that we have our definitions, I’m going to write out the ratio of the probability that my hypothesis D is true to the probability that your hypothesis T is true:

          Pr(D|J) / Pr(T|J) = Pr(J|D) x Pr(D) / (Pr(J|T) x Pr(T))

          So if the quantity on the right is greater than 1, this means that my hypothesis is more likely to be true than your hypothesis, given our observation J; similarly, if it is less than 1, then your hypothesis is more likely to be true.
          Observe that Pr(J|D) = 1, since my hypothesis is specifically that God has sent Jesus to Earth to be as convincing as possible, and that entails never saying anything that a human could tell was a lie. Also observe that Pr(J|T) = 1, since your hypothesis also entails Jesus never saying anything that a human could tell was lie (indeed, under your hypothesis he [presumably] only tells the truth). Hence we can simplify our expression to:

          Pr(D|J) / Pr(T|J) = Pr(D) / Pr(T)

          This result implies that your hypothesis is more likely to be true if and only if your hypothesis has higher a-priori probability than mine—meaning it is inherently more likely to be true, regardless of what we observe. Therefore, to argue that your hypothesis is more likely to be true given what we observe about Jesus’s behavior, you have to argue that your hypothesis is more likely to be true a-priori—without referencing our observation. Fundamentally, this is because both of our hypotheses account equally well for Jesus’s observed behavior.

          Now, from what can we base our arguments about the a-priori probability that our hypotheses are true? Absolutely nothing—we have no idea what the probabilities are, so the most prudent conclusion is that the two hypotheses are equally likely to be true (because a uniform probability distribution is the only information-free way to distribute the probabilities). Indeed, you at least partially recognized this when you said “I don’t see a single reason to think it’s probable”. You’re right, we don’t have a single reason to think it’s probable…but I can say the exact same thing about your hypothesis…we don’t have a single reason to think it’s probable! So we have no reason to choose one over the other—they are on equal footing.

          Of course, the hypotheses above (D and T) are not the only hypotheses about God’s moral truthfulness we can come up with; for example, we could hypothesize that God lies in 1% of his moral statements and tells the truth in 99% of them. But my argument easily extends to these cases as well.

          Now I’m going to argue that for any hypothesis about how fundamentally truthful about morality God is, there is an equally probable hypothesis in which God’s moral truthfulness is exactly opposite. Basically each step of my argument will still hold when the hypotheses are replaced with hypotheses of this kind. For example, if we hypothesize that God lies in 1% of his moral statements, denote this hypothesis “T” and the hypothesis that he lies in 99% of his moral statements “D”. Now follow the same steps of my argument—you will always be able to argue that both hypotheses account equally well for the observed evidence.

          Fundamentally, the issue here is that whatever you propose, I can always propose a hypothesis that explains our observations equally well. This eliminates the observations from consideration, and now you are left arguing that your hypothesis has a higher a-priori probability than mine. Since there is nothing you can say about the a-priori probability (unless you make some sort of secular argument about the morality of lying), there is no reason for you to treat your hypothesis as more likely than mine. So although you may say “It’s possible, but why would I think that it’s probable?”, I can always say [with an equal level of justification] “Why would you think your hypothesis is probable? It’s like knowing nothing about a coin, but asserting that ‘heads’ is more probable. You have no reason to prefer heads over tails.”

          So the inferential leap of preferring hypothesis T over hypothesis D is not justified. Then why are we justified in making an inferential leap when judging whether another human being is being truthful?
          The crucial difference here that makes the inferential leap justified is that we do know something about the a-prior likelihood of a human being truthful or deceptive.

          Suppose a friend tells me something important, and that this friend has told only truths to me in the past (at least as far as I can tell). Denote this observation ‘R’ (for Record). Now let TRUTH be the hypothesis that my friend is truthful, meaning that my friend always tells the truth. Let LIE be the hypothesis that my friend is actually a liar, and has been truthful to me in the past only to gain my trust so that it would be more devastating when they lie to me for the first time. The ratio of the probability of TRUTH to the probability of LIE given that I have observed R, that my friend has only told truths to me in the past, is given by:

          Pr(TRUTH | R) / Pr(LIE | R) = Pr(R | TRUTH) x Pr(TRUTH) / (Pr(R | LIE) x Pr(LIE))

          As before, both hypotheses account for the observation R equally well, since the friend would tell only truths to me up to this point under both hypotheses. Thus the ratio reduces to:

          Pr(TRUTH | R) / Pr(LIE | R) = Pr(TRUTH) / PR(LIE)

          So again we are stuck with comparing the a-priori probabilities of our hypotheses. However, this time, we actually do know something about the coin we are flipping. We know that it takes a human a massive amount of effort to appear to be truthful without actually being truthful—i.e. that good liars are very rare—people with bad intentions are usually revealed to have bad intentions pretty quickly, just because of the massive cognitive overhead required to continuously plan out how you are going to deceive people to achieve your ends, without accidentally revealing them. Liars skilled enough to tell 1000 truths in a row while planning the whole time to eventually trick their friend are hard to come by. This is fundamentally the reason why if a person has a good truthfulness track record, we can safely assume that they don’t secretly intend to betray us…because planning to do so and executing that plan perfectly is something that only a very small percentage of humans would be able to do without making a mistake.

          However, we can’t say the same about an infinitely powerful and intelligent God…this God would also be infinitely good at deception, if it wanted to deceive people. So there is by definition nothing that could stop it from pulling off an infinitely sinister deception. This is the critical difference here. Humans are unlikely to pull off such great deceptions without making a mistake and breaking their perfect track record, while an infinitely powerful and intelligent God could tell 1,000,000,000 truths in a row while all along planning to betray you at the end.

            1. trivialtruths says:

              It is, but it’s a direct consequence of God being omnipotent, which I imported as part of the Christian worldview. You are welcome to relax that assumption, but I don’t think it’ll help us reach the conclusion that God’s moral teachings are probably true. I’m up for exploring it. But then are we in agreement that there is no reason to believe in and follow Christian moral teachings if God is omnipotent? If we have reached agreement here, we can see if we can rescue Christian morality by allowing God to be less than omnipotent.

              1. Brent says:

                Oh, I thought importing the nature of God from the Christian worldview was out of bounds (“to argue that your hypothesis is more likely to be true … you have to argue that your hypothesis is more likely to be true a-priori—without referencing our observation”)? It’s hard for me to ignore the fact that if we’re playing fair, that neither hypothesis gets an a-priori, then there is no longer a “critical difference” between human character and divine character, and we are then justified in inferring divine quality from divine behavior. (For the record, I also disagree that both hypotheses explain the behavior of Jesus equally well — I think the deceptive god hypothesis is going to get really adhoc really quick, esp as it’s applied to the rest of the New Testament, but I digress..)

                Re: your reason for wanting to import this specific omni, “But then are we in agreement that there is no reason to believe in and follow Christian moral teachings if God is omnipotent?” Did you mean to say “if God is [not] omnipotent”? Assuming so, how does God’s non-omnipotence suggest that we should no longer follow His prescriptions? What if one is a Platonist (I’m not!) and believes “the good” transcends the divine, such that the divine is limited by “the good”, and therefore not omnipotent, yet the divine commands “the good”.. Should we then disobey what has been relayed to us? I don’t see why. More importantly, we need to get at what omnipotence means: ie, if it’s “the ability to do anything”, then I deny that.

                But nonetheless, I’m willing to indulge you for the sake of mental exercise (mostly to let you continue the rest of the argument). All cards on the table though, the deceptive God hypothesis truly does not strike me as particularly worthy of concern at this point, esp since I’m not sure the argument is even valid to begin with (ie, if one premise is “no importing a-prioris”, then stating “omnipotence” in any following premise will invalidate the argument). But, if we suppose that the argument is valid & sound (and fair), that Christian morality needs rescuing.. What is your proposal?

                1. trivialtruths says:

                  There’s some confusion here that we should clear up. I never said that a-priori reasoning isn’t allowed; indeed in the quote you referenced at the beginning of this last comment, I say “…you have to argue that your hypothesis is more likely to be true a-priori…“, in other words, you must use an a-priori argument. My claim is that there is no reason to think that the a-priori probability of hypothesis T is any higher than the a-priori probability of hypothesis D. An example of such a reason would be “lying is inherently wrong”, but to use that argument, you have to reference some secular moral argument [this puts us back in the context of my original post]. It was my intention to argue all of this from inside the Christian worldview (visible in the hypothetical tone of character A in my dialogue, and in the title of the post), so I was taking as a starting point basic components of the Christian worldview. Implicitly I had the following assumptions (among others, but these are the most relevant) in mind while writing the post and the subsequent comments:

                  • Assume that there is a sound argument for God’s omnibenevolence.
                  • Assume that there is a sound argument for God’s omnipotence.

                  Here I take omnipotence to mean “being as powerful as possible; able to do anything that is not logically impossible” (this is generally what Christians mean by omnipotence in order to defend against all the “could God create a stone that he couldn’t lift”-style arguments). Notice the above implies that God is limited by the good, and that God doesn’t have the ability to do anything, so I think we are in agreement there.

                  So now let me try to clarify what I’m trying to do with this argument.

                  Basically I observed that, when making moral arguments, Christians usually sneak in another assumption: “God cannot lie”. In contrast to the two assumptions above, which (let’s grant) have sound arguments behind them, it seemed to me that a circular argument was being used to justify this one. Then you and Derry rightly pointed out that if we modify the claim to “God is unlikely to lie” then it seems quite reasonable. At this point we dove headfirst into the depths of inference, in which I used Bayes’ Theorem to show that hypothesis T is more likely than hypothesis D if and only if T is a-priori more likely than D. The reason for this is that both D and T imply J, which makes the observation of J irrelevant for the comparison of the likelihood of the two hypotheses. This specific point is what I meant when I said that both hypotheses explain the observed evidence equally well—I was specifically referring to the fact that Pr(J|D) = Pr(J|T). This is indisputable because every possible observation is identical in T and D. The only difference between T and D is the truthfulness of God’s moral teachings, which we can’t measure empirically. In other words, everything looks exactly the same if T is true relative to if D is true. Hence we can’t look at anything empirical to compare these two specific hypotheses, which is why (as Bayes’ Theorem suggests) all we can do is compare the a-priori probabilities of the two hypotheses. When you said “…I also disagree that both hypotheses explain the behavior of Jesus equally well — I think the deceptive god hypothesis is going to get really adhoc really quick…”, I think what you are really disputing is that the a-priori probabilities are equal (I suspect this because you said ad-hoc, which usually refers to a-priori likelihood). I want to make sure this point is transmitted and received properly, so can you be sure to tell me in your reply if you agree to this statement: “hypothesis T is more likely than hypothesis D if and only if T is a-priori more likely than D”. If so then we can move on to then trying to find a-priori differences between the two hypotheses, but if not I need to communicate this point better. Maybe you can think of it like this: do you agree that, given the way I have defined the hypotheses D and T, both hypotheses imply J? That is, do you agree to both of the conditional statements “T implies J” and ” D implies J”. If not, please know that it was my intention to define T and D so that these conditional statements would be true, so append those to my definitions if that’s not clear.

                  Note that this is a general procedure in reasoning about potential explanations of past events. Generally we posit two hypotheses which both imply the events we are looking to explain, which (by Bayes’ Theorem) allows us to move to comparing their a-priori probabilities. So, for example, we could compare the hypotheses that 9/11 was a real terrorist act vs. the hypotheses that 9/11 was a false flag operation staged by the US government. Since both hypotheses obviously imply that 9/11 happened, we can then move to comparing the a-priori probabilities of the hypotheses. This is exactly what I intended to do here. Both of the hypotheses are supposed to obviously imply that J is true.

                  Ultimately, what I’d like to do is establish a valid argument for the claim “Christian moral teachings are probably true”, based on three premises: that God is omnibenevolent, omnipotent, and everything verifiable in the Bible is true. These are three big premises, but I want to grant them as true so that I can explore the justification of morality from inside the Christian worldview. I think we are going to get there but it may have to be one step at a time.

  4. Brent says:

    I’m not familiar with adhoc explanations referring to a-priori likelihood, although maybe you’re a step ahead of me. But I take an adhoc explanation to be one that results from not thinking a core hypothesis through, such that as difficulties being to arise from the application of the core hypothesis’, more and more contrived explanations are required to sustain it. If you’re familiar with Imre Lakatos’ ‘methodology of science’, in which a refuted hypothesis is one which evidences a growing number of adhoc (perhaps you could use the word “contrived” here) explanations, that’s what I was after. I suspect the deceptive god hypothesis, when used as a working hypothesis (to explain the New Testament, the experience of the early Christians, much of what we know of life, my own personal experiences, etc.,), will begin to amass more and more adhoc explanations in order to sustain it. So I really see it as more of an a-posteriori thought than the converse. But maybe I’m wrong!

    In any case, I’m willing to forego my objections, not because I’m convinced they’re wrong, but mostly because your final paragraph very much intrigued me. Let’s suppose that we need some other argument for thinking that “Christian moral teachings are probably true”.. Where do you think we should go from there? Or are you waiting on me? :)

  5. trivialtruths says:

    What I meant was that the complaint that a hypothesis is ad-hoc is one type of a complaint that the hypothesis is a-priori unlikely to be true, specifically due to its complexity—consider Bayes’ theorem. It implies that a hypothesis can be unlikely either because what we observe is unlikely to occur under the hypothesis or the hypothesis has low a-priori probability. Since ad-hoc hypotheses always imply the observations we see, it is their low a-priori probability that makes them unlikely to be true. One of the related claims I’m making is that we have no reason to think the deceptive God hypothesis has lower a-priori probability than the truthful God hypothesis. Note that under the deceptive God hypothesis, the New Testament is exactly identical, the experiences of the early Christians is exactly identical, what we know of life is exactly identical, and even your personal experiences are exactly identical. You don’t need any more ad-hoc explanations to sustain it than you do for the truthful God hypothesis—you don’t need to import any more information, simply because all of those facts are explained by God having malicious intentions. I would also add that justification for preferring simple hypotheses over complex hypotheses vaporizes in a theological worldview, since we have no reason to expect God to have preferred a simple universe over a complex one, and no reason to think that God himself is simple rather than complex. Anyway, if we are not making progress here, we can move on to the more interesting question. I’m searching for an argument for the statement “Christian moral teachings are probably true”. For reasons including those I’ve laid out in this post and these comment threads, I don’t think that such an argument can be made—or I just haven’t seen one yet. So I’m very interested in hearing such arguments, if you’re up for it.

  6. Brent says:

    Wait. Ad-hocness has nothing to do with anything a-priori. Consult whichever philosophical reference you like, “adhoc” is an a-posteriori measurement.. That’s the whole point of what adhoc explations are, and why they are to be avoided: having no other background information to go on (esp a-priori data), adhoc explanations are seen as indicators of a hypothesis thought should be abandoned. Two hypotheses can be equally probable in all other measures, but an increase in adhoc explanations for one hypothesis indicates preference for the other. It truly has nothing to do with axioms or a-prioris..

    Here’s what I mean about the deceitful god hypothesis (D) probably requiring adhoc explanations. My point is exegetical. When doing exegesis, one has some sort of systematic framework for interpreting passages of a given text. Overall coherence is paramount. The Marcionites, for example, were an early sect of Christians who believed ‘the Yahweh of the Old Testament’ was in fact an evil deity. One of the problems that Tertullian and Origen had with Marcion was that, when Marcion’s exegesis was applied to the rest of the scriptural corpus, Marcion contrived more and more strange explanations, making his overall exegesis less and less cohesive.

    Today, there are Calvinist systematic theologies, Arminian ones, Molinist ones, and a thousand others, each of which have far less adhoc (contrived) explanations for difficult parts of scripture than Marcion’s did. The same problem of exegesis plagues other strange theories like the Valentinian view that God had a child through a cosmic goddess, or the Phibionites’ view that there are 365 heavens, each requiring a secret password to get into. Sure, the New Testament is ‘identical’ for each view, but when each view is applied to the same text, each view begins evincing some measure of adhocness. Note, again, there is nothing a-priori about this process. One can presume equal likelihoods for each view and slowly vet out which view is exegetically stronger by noting which view amasses more contrivances.

    With regards to (D), the deceitful god hypothesis, the only way to truly measure its adhocness against other theological frameworks would literally require one to produce a ‘Systematic Theology of the Deceitful God’, which covers all of theology from prolegomena, to anthropology, to Chrstology, to pneumatology, to eschatology, and so on. Obviously that’s not going to be done, but I strongly suspect its results would play out in a manner similar to the way Marcion’s did. That’s the point I’m trying to make. Perhaps you already understand all this, and perhaps the point you’re trying to make to me is “No, it won’t. A systematic theology of (D) will have just as many adhoc hypotheses as the Calvinist one.” If that’s the point you’re trying to make, let me know. I’d at least want to know if you think (D) would have as few contrivances as (1) Calvinists, (2) the Arminians, (3) the Open Theists (they’ll all have a different measure, as they do each have contrivances), and why. Then I’d bring up some scripture that would be embarrassing for a deceitful god’s mission (to name one example, why he would have foreknowledge throughout most of scripture, but fail to do so at times during his earthly mission).

    I’m a little confused on something else.. Two posts ago you said “Ultimately, what I’d like to do is establish a valid argument for the claim “Christian moral teachings are probably true”… I think we are going to get there but it may have to be one step at a time,” but then last post you said “I don’t think that such an argument can be made.” So, do you think we’re going to get there or not? To be frank, I’m trying to gauge whether or not to continue this exchange. Are you truly concerned that, if a moral deity exists, and you had to place a bet on whether that deity was truthful or deceitful, that you’d bet on deceitful? This isn’t Pascal’s Wager, btw — I’m just trying to understand what’s really motivating you. I hope this doesn’t come off as harsh..

  7. Brent says:

    Btw, adhoc<>complex. When you say “ad-hoc is one type of a complaint that the hypothesis is a-priori unlikely to be true, specifically due to its complexity”, it sounds to me like you’re confusing the virtue of parsimony with the vice of adhocness. Adhocness has very little to do with complexity. As far as adhocness is concerned, a scientific theory can be as explanitorily complex as it wants to be, so long as those complex explanations cohere with each other, and do not evidence clear contrivance. And this is measured by how related an explanation is to the ‘core’ hypothesis vs auxillary hypotheses, how nested such hypotheses are from the core, etc. A hypothesis with only two explanations (and therefore horribly simple), yet with one contrived explanation, is simple, yet adhoc. So, adhoc<>complex. Do read Lakatos on this, if you haven’t yet (it would be fruitful for your work in the sciences too).

  8. trivialtruths says:

    Apologies for the delay—just finished my final exams. Thanks for your patience (and perseverance, in this quite long discussion). I find these discussions incredibly valuable intellectually so I very much appreciate your participation.

    I think we are running into a significant amount of confusion with the evaluation of hypotheses due to some imprecision in the way that we are using the terms a-priori and a-posteriori.

    The fact that you said “ad-hocness has nothing to do with anything a-priori” and “‘adhoc’ is an a-posteriori measurement” make me think that you and I mean different things when we say a-priori and a-posteriori. In my previous comments, I have been (trying to) use a-priori and a-posteriori precisely: when I say “a-priori likelihood”, I’m referring to how likely a hypothesis is without considering any observations; when I say “a-posteriori likelihood”, I’m referring to how likely a hypothesis is when our observations are considered. Bayes’ Theorem relates these two quantities in a precise way. One other thing I think may potentially be obfuscating things here is that when I say “likelihood”, I am referring to how likely something is, not implying that it is likely. I’m not sure if this caused confusion but it seems possible that it did.

    I’m going to direct your attention back to a specific sentence in my previous comment: “Since ad-hoc hypotheses always imply the observations we see, it is their low a-priori probability that makes them unlikely to be true”. This sentence is an explanation for why ad-hocness is a vice—because ad-hoc hypotheses generally have low a-priori probability. Bayes’ Theorem implies that a-priori likelihood is the only criterion by which we can compare two hypotheses which both imply whatever we observe. Pointing out ad-hocness is one way to argue that a hypothesis has low a-priori likelihood. Do this make sense to you? If so, we can continue along this thread. I worry that the differences in how you and I think and communicate about probability, inference, adhocness, and scientific epistemology are so fundamental that we will have to go very very deep to resolve them. Because it may be helpful, I offer a high-level summary/paraphrase of the debate we have had so far as I understand it (particularly this thread of comparing the likelihood of various hypotheses about God’s moral, since we have touched a few other things as well):

    Me: Isn’t it possible that God has deceived us into believing that what is truly wrong is right?

    You: Yes, but this is unlikely because God has been truthful in the past for every statement which we can verify.

    Me: Suppose that God has been truthful in the past intentionally, because he is malicious. He does this to gain our trust, so that when he gives us a false moral statement—which of course we cannot verify—we believe it, and that causes us to go to hell.

    You: It is possible, I suppose, but this is an ad-hoc hypothesis, so it is less likely to be true (than the standard hypothesis).

    Me: Why is an ad-hoc hypothesis about God less likely to be true?

    It is this last question which I believe you have not yet answered. You have discussed the nature of ad-hocness as relating to how contrived a hypothesis is—I said ad-hocness relates to how complex a hypothesis is, but contrived and complex have such similar meanings that it is plausible that we have come to nominal agreement there. I was using complexity to refer to information content, which is also what contrived-ness refers to. I could explain this in detail, but like I said, I think we would have to go very, very deep to nullify our differences how we understand these terms (which I think are simply due to us being introduced to the terms in different contexts—I do research related to information theory, so many of these terms have gained very specific meanings in my mind that may stray from the meanings they have for other people). So for the sake of moving forward, let’s work with what you mean by adhocness. Let’s work on the last question above.

    This is sort of a separate thread, but just to answer your questions about my intentions here: I have a deep interest in the nature of morality, as well as a deep interest in theology. Naturally, then, much of what I think about is the relationship between morality and theology. In this post, I wanted to explore why Christians are justified in believing/following Christian morality, and it is currently not clear to me why they are. So I want to (if possible) derive the statement “Christian moral teachings are probably true” from the assumptions “God is omnibenevolent, omnipotent, and everything verifiable in the Bible is true”. To answer your question of whether I would bet that God is deceitful, I reiterate something I said in the comment I made on February 25th: “we have no reason to choose one over the other”. Your response, of course, is that the deceitful hypothesis is ad-hoc. Which is why I think we should now work on the last question in my summary above.

Leave a comment

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s