Discussion Grounds

Tom,

You write:

We are [discussing] whether Christianity or naturalism is a better explanation [something]. But now it’s becoming apparent that we don’t agree on what the term “Christianity” signifies.

You persist in insisting that Christianity believes in an “imaginary magical friend who grants wishes”… [but] neither of us believes that.

If you do not present Christianity as a belief in an invisible, magical, wish-granting friend, then I’m certainly not going to debate such a view. I’m going to debate whatever set of hypotheses you present as the best explanation for the world as we know it.

But yes, I do think most Christians believe they have an invisible, magical, wish-granting friend. I don’t think that’s a metaphor. I think that’s literally true of what most Christians believe.

You disagree.

For example, you write: “God is invisible and a friend, but he is not an ‘invisible friend’ as the term is generally understood.” I don’t know what to make of that. You seem to be saying that “God is A and also B but God is not A and B.”

You say the problem is that the phrase “invisible friend” conjures up notions of a child’s invisible friend. I’m not used to hearing that phrase. I’m used to the phrase “imaginary friend.” But when I searched wikipedia for “invisible friend” it redirected to “imaginary friend,” so apparently some people use the term that way.

I’m not very familiar with the concept of an imaginary childhood friend. I never had one, and I still have never met anyone who did (except Jesus, whom you say is not an imaginary friend).

The Wikipedia article on imaginary friends has several warnings at the top, so I didn’t want to trust it as even a first-step source of information. I was able to track down a few research papers, though. Apparently the research term for this phenomenon is “imaginary companions.”

Children’s imaginary companions seem to often function as both guardians and playmates. They have personalities. And even though they appear real to the child, the child usually knows “deep down” that their imaginary companion is not real.

Now, who do Christians think Jesus is? A guardian, yes, but probably not a playmate. They certainly think God has a personality, whatever philosophers try to tell them about God. And believers do not appear to “know deep down” that Jesus is not actually real.

So there is a big difference between the idea of a childhood imaginary friend and the believer’s conception of Jesus. And I don’t think anybody would say the two are identical.

But I never said the two were the same, or even implied it. I used the term “invisible friend,” I did not invoke the notion of a childhood imaginary friend, and as you admitted, the term “invisible friend” is literally true of what you believe.

Tom, I know that the phrase “invisible friend” has a negative connotation. That is part of the point. If you can say “I believe in an invisible friend, and here are my reasons for thinking so…” then more power to ya. But I think that forcing believers (in anything) out of their own euphemisms for things helps us biased humans to see things outside of our own bubble, which is important.

For example, I am quite happy to admit that I believe that “Consciousness and morality evolved from the unguided bouncing around of invisible particles.”

Those aren’t quite the words I would use to describe what I believe, but I am happy to say “Yes, that’s what I believe, and here are my reasons…”

Phrasing my beliefs in terms I’d prefer not to use allows me to see them as an outsider might see them, and to see them as how “crazy” they potentially are. For example, it reminds me that it’s far from obvious that consciousness or morality could arise from subatomic particles.

But there is a parallel between a child’s invisible friend and the Christian conception of Jesus. Namely, they are both invisible and they are both a friend. And that is exactly the parallel I’m trying to make, if any.

Okay, as for Christians believing in magic. Jesus sometimes invokes “the Father” to supernaturally affect the natural world. Now if you believe they are the same people, I’m not sure what to say. The concept of the Trinity is incoherent to me.

But either way, Christians themselves certainly invoke the supernatural all the time. They pray to God to affect the natural world in their favor. They believe there is a particular “art” to it, as Jesus taught it (the Lord’s prayer, etc.). Most Christians believe there are specific techniques that are more effective than others, whether it be candles or holy water or drawing a cross on one’s forehead with oil. Most Christians use specific phrases in their prayers that they believe to have special power, like “Amen” or “in the name of Jesus.” Again, you may not believe in all of this but most Christians do, and that’s what I’m talking about.

(But I think you do believe in magic, for you are quite explicitly invoking the supernatural to explain natural phenomenon as the basis of your argument that God exists.)

As for granting wishes, I never said that God’s only role was in granting wishes, just as I never said that his only attributes were invisibility, magicality, and friendliness. In any case, it sounds like you’ve agreed that your idea of God sometimes grants wishes (if not, then I assume you reject the idea of intercessory prayer, as some Christians do).

So yeah, it’s literally true that most Christians believe Jesus is “an invisible, magical, wish-granting friend.” He is many other things as well, but that’s one of them. And it’s no less fair to say it that way than to say he’s “our Creator, Sustainer, and Savior” (or whatever), because that “unfairly” leaves out the part about him being invisible, magical, and wish-granting.

Tom, I understand why this upsets you and other Christians. Perhaps it’s embarrassing to literally believe in an invisible, magical friend who grants wishes. But if you’ve got good reasons for believing all that is true, then it shouldn’t be embarrassing at all! Likewise, I could be embarrassed that I believe “consciousness evolved from invisible particles bouncing around.” But I’m not embarrassed to admit that, because I believe I have good reasons for believing that is true.

The reason I use this language from time to time – while at other times dealing very somberly with the details of various arguments for theism – is because this language made a big difference in my life.

As a Christian, when I realized that I literally believed in an invisible magical friend who grants me wishes, I ranted and raved against the atheist who said it. I told him he was being unfair and disingenuous. I really let him have it.

But later, when I was “off the stage” and not in competition mode, I considered what he had said. And I realized that whatever connotations came with the phrase, it was literally true of what I believed. And I said to myself, “Woah, Luke, you really believe you have an invisible magical friend who grants you wishes. That might be true, but you’d better at least look into that.”

That wasn’t the year I lost my faith. But it was the year that things changed and I was able to look at my faith from the outside and try to examine it as objectively as humanly possible.

So that’s why I use such language. But I don’t just use it on religious believers. I use it on atheists, too. We all need a shock every now and then to step outside the protective bubble of our familiar worldview and examine it from the outside.

And I use this technique on myself. All the time. It’s been very, very useful. It’s just one more tool to combat my own prejudices and biases, which are always barking at the door and sometimes smash through the window and take over.

So I’m not sure what your objections is. It sounds like you’re admitting that your concept of Jesus is that of “an invisible, magical, wish-granting friend,” though of course your concept of Jesus is many other things also (cosmic savior, creator of the universe, etc.). Maybe your objection is not with the literal truth of what I said but with my choice of words. But I’ve just explained the utility of my choice of words.

I’m happy to continue discussing this with you, but again, I don’t think it’s relevant to our debate. When you put forward “Christianity” as an explanation for y, I’ll need you to lay out the hypotheses that you intend by offering “Christianity” as an explanation for y, anyway. And if you don’t include the part about how Jesus grants wishes or whatever, then that’s fine. I’m going to argue against the explanatory merits of whatever hypothesis or set of hypotheses you put forward as an explanation.

By the way, if you want to add more punch to the part of your upcoming book where you write that “What others hear when we [Christians] speak is that we are asking them… to order their… lives according to advice given by a legendary miracle worker who (maybe?) lived two thousand years ago,” you could instead (quite accurately) write that “What others hear when we Christians speak is that we are asking them to order their lives according to the advice of their invisible, magical, wish-granting friend.” Because that’s not just what we hear, it’s what seems to be literally true of what most Christians believe.

Finally, allow me to propose one more point of agreement:

For x to be a successful explanation of y, we do not need to also have an explanation of x.

Richard Dawkins has asserted that for x to be an explanation of y, we must also have an explanation of x.¹ But this is nonsense. This requirement immediately leads to an infinite regress of “Why?” questions. Because I’m lazy, I’ll just quote William Lane Craig:

Dawkins says that you cannot infer a Designer of the universe [from] the complexity of the universe because this raises a further question: namely, “Who designed the Designer?” [But] this argument is quite inept, because philosophers of science [know] that in order to recognize an explanation as the best explanation, you don’t have to have an explanation of the explanation…

Let me give you an example. Suppose archaeologists digging in the earth were to come across artifacts looking like arrowheads and pottery shards… it would obviously be justifiable to infer that these artifacts were the product of some lost tribe of people, even if the archaeologists have no idea whatsoever who these people were or how they came to be there.

Similarly, if astronauts were to discover a pile of machinery on the back side of the moon, they would be justified in inferring that these were the products of intelligent design, even if they had no idea whatsoever where this machinery came from or who put it there…

In fact… if in order to recognize an explanation as the best you have to have an explanation of the explanation, that leads immediately to an infinite regress. You’d need to have an explanation of the explanation of the explanation, and so on… to infinity. You would never have explanation of anything, which would destroy science. [So] Dawkins’ principle, if adopted, would actually be completely destructive of science. That’s how inept this argument is.

So I hope we can agree on that.

Cheers,

Luke

1. Well, he didn’t make that algebraic statement, but he has said that theists cannot offer God as an explanation for something because it leaves God himself unexplained. [↩]

Luke,

Based on recent discussion in the comments, we have one more point of agreement we need to arrive at before we can proceed with this. Quite literally it is a matter of agreeing on the topic of the debate. We are entering into discussions whether Christianity or naturalism is a better explanation of x, where x could over time be many different things. But now it’s becoming apparent that we don’t agree on what the term “Christianity” signifies. Without that agreement, we might think we’re debating one thing but really be talking about two different things.

You persist in insisting that Christianity believes in an “imaginary magical friend who grants wishes” (see here, here, here, here, here, and finally your continued affirmation of that opinion just last night, even after I had explained in some detail the inaccuracy of that view). As I have said twice now (here and here), if that is what you are disputing, then we have nothing to debate, because we agree that’s false. Neither of us believes in that.

Because it’s so important that we work this out, I am going to quote my own comment, in which I explained the inaccuracy of that depiction.

God is not an “invisible friend” in any sense comparable to the usual childhood understanding of the term: a playmate whose participation is under the control and at the whim of the child who invents this friend. You know this is true.

Surely you know that melding words into a phrase can change the words’ meanings compared to their meanings taken individually. Consider “Hopeful,” “monster,” and “hopeful monster,” for example. If one were to argue against that aspect of evolutionary theory by saying, “Hold on, you blithering idiot: those early creatures could never have understood hopefulness!” that arguer would himself or herself be the actual blithering idiot.

You insert “magical” between “invisible” and “friend,” but in my mind that does not erase the allusion to the imaginary childhood playmate, it only adds a touch of fantasy to the image. Maybe (taking a gracious view of it) you’re not making that allusion intentionally, and maybe if pressed on it you would even disavow that connection; but I’m quite sure that’s what many of your readers would take as your intent. If that allusion is not what you want to communicate, you would be well advised to quit saying things that do communicate it.

Again: God is invisible and a friend, but he is not an “invisible friend” as the term is generally understood. To suggest there’s a parallelism there is to do considerable violence to both “God” and “invisible friend” as concepts, and it is irresponsible argumentation. I’m calling you on it.

Now for “magical.” God is decidedly not “magical” in the sense stated in answers.com. He does not “invoke the supernatural.” He is himself the being behind all other being, and his essence is of course supernatural, but he does not use some “art” to “invoke” himself! And he certainly does not (definition 2) use “charms, spells, or rituals to attempt to produce supernatural effects or control events in nature.” The term “magical” as applied to God is just wrong. God acts out of his own being, not out of some magical art.

And then, as for “granting wishes.” I thought you said you had a grip on Christian theology. Don’t you see how you have chopped off 98% of what it means to be in relationship with God? Don’t you see that our relationship with him is not one where he “grants wishes,” but where he builds character, molds our desires to be in tune with his good and ultimate moral nature (and thus molds our very wishes), calls on us to sacrifice ourselves and yield to him, calls on us to give him the worship that is due him?

An “invisible magical friend who grants wishes” is not what I believe in, and not what I’m arguing for. It is a red herring of amazing triviality. If integrity is your intention (and I believe it is), I would expect you would drop the term, both here and on your own blog.

Roman congratulated you for being able to handle your views being stated in a manner not entirely to your liking. You are not the only one who can do that, I assure you. Download the pdf here and search the document for “evangelistic efforts.” There’s one example of self-criticism. The difference in this case, Luke, is that you are repeatedly stating tendentious views that do not accurately apply to Christian belief, and expecting us to accept it not only with magnanimity but with agreement!

So let’s go back to working on common definitions. What is Christianity, and what do Christians believe? We have to come to better agreement on that before we can proceed with other matters.

Tom,

Despite the fact that inference to the best explanation is far less developed than deductive logic,¹ it looks like we’re honing in on some agreement over what it means to say that some hypothesis or theory is the best explanation of some phenomenon. We might say that:

x is the best explanation of y if it is the case that:

(A) if x were true, then by knowing x we would better understand x‘s causal background than by not knowing x [i.e. x is a potential explanation of y],

and if it is also the case that

(B) x possesses the following explanatory virtues to a greater degree than any other known potential explanations of y: testability, consistency with background knowledge, past explanatory success, simplicity, ontological economy,² informativeness, predictive novelty,³ explanatory scope, and explanatory power.

Is that agreeable?

There is much left to say about the value of each of our explanatory virtues, but we can address those points as they come up in our debate.

You’re right to guess I only accept material causes and efficient causes. I don’t think it will be very interesting for you to argue that Christianity offers a better explanation for the telos of humanity than naturalism does. In the same way, it wouldn’t be very interesting for you to argue that Christianity offers a better explanation of souls than naturalism does. As much as possible, each of us are going to have to appeal to evidence from our shared ontology if we are going to have a chance at persuading the other. That is usually the project of natural theology and natural atheology, anyway.

Tom, in order to keep our debate manageable, I hope you can start with one particular phenomenon of the human condition – one on which we agree – and then argue that Christianity provides a better explanation for that phenomenon than naturalism does. You can do the same with other phenomenon of the human condition in the future if you like.

1. For example, see Peter Lipton, “Inference to the Best Explanation“ (2000). Lipton’s “What Good is an Explanation?” (2004) is also useful. [↩]
2. I changed “ontological simplicity” to “ontological economy” just so that it was better distinguished from the explanatory virtue simply named “simplicity.” [↩]
3. I added this one because I doubt you’ll disagree with it. I’m using the term as Richard Swinburne uses it in Is There a God? (1997). Predictive novelty refers to a theory’s ability to predict previously unknown facts. For example, Einstein’s theory of gravity predicted as-yet unobserved phenomena that Eddington later observed to be true. [↩]

Luke,

Greetings. I think you have described the situation well. Neither of our definitions is fully adequate. I see now that if one takes my (B) and (C) as conjunctive with (A), one ends up with heuristic explanation, which is adequate in many cases; but what we’re after is what is true, not what’s merely useful.

Your (A*) revision of my (A) is an improvement, and I suppose we could improve it still further with (A**):

(A**, Heuristic) By knowing x, where x is relevant to y’‘s causal background, we understand y better than by not knowing x.
[(B) and (C) must conjunctively accompany (A**)]

or

(A***, Ultimate) By knowing x, where x is true and x is relevant to y‘s causal background, we understand y better than by not knowing x

The second of these obviates the need to include either (B) or (C). What I like about it is that it emphasizes that explanation is more about increasing causal (as you rightly pointed out) understanding than it is about anything else. For the record, I accept Aristotle’s four-fold approach to understanding causation: material, formal, efficient, and final. This clashes with some scientific reductionists’ view that for many phenomena there are only material and efficient causes. I am not predicting we will have a disagreement over that, and I suggest we not try to sort this out in advance (unless you happen already to agree with Aristotle on this). If the disagreement arises, let’s deal with it then.

I think we have it sorted out well enough to progress, though I want to add these two remaining items to the list of explanatory virtues:

7. Explanatory scope: a good hypothesis will explain a wide range of data (a wider range than rival hypotheses)

8. Explanatory power: a good hypothesis is one that supports the epistemic probability of the evidence

In many cases the test (epistemic virtue 1) of hypotheses must be in the form that’s suitable to inference to the best explanation (IBE). With IBE, in the absence of other tests, one looks at the other explanatory virtues and determines which of two or more rival hypotheses better meets those virtues. There is, for example, no “test” for either the existence or non-existence of God as “test” is commonly understood. There is no “test” for the actuality of claimed historical events. There is, however, evidence in the world, in experience, and in documents, and there are rival hypotheses; and the test must be in the form, which of the rival hypotheses better explains the evidence?

In Letters 9, 10, and 11, I presented a set of phenomena, evidences relating to the human condition, and I suggested that Christianity explains them better than does atheism. I did not formalize it according to these explanatory virtues, since we had not developed the list yet. Would you like me to do that now, or is the case I’ve made on this point already in a form that you are ready to respond to?

Regards,

Tom

Tom,

You proposed an approach to explanation such that “x contributes to the explanation of y if” it meets three criteria. Your first criterion was:

(A) by knowing x we understand y better than by not knowing x

But this is problematic because it requires that any proposed explanation also be a successful explanation. Consider the phlogiston theory of combustion. I think this theory should at least be “in the running.” We reject it not because it cannot be an explanation for combustion, but because it fares more poorly on our list of explanatory virtues than the modern theory of combustion. Your criteria (A) requires that we only ever consider theories that are correct, but of course we can’t know which theory is correct until we can compare them!

We could modify your criterion like so:

(A*) if x were true, then by knowing x we would better understand y than by not knowing x

This works better, I think, but there is still a problem. Explanation is generally thought to be positing the cause of an event. And yet your criterion would be fulfilled by “explanations” that merely help us understand the effects of x. Suppose that y = “The murder of John Dominic.” And I offer x = “Upon his murder, John Dominic’s body did not decay because the court ordered that his body be preserved.” Now this x certainly helps us understand y better, and it also could score extremely well on your other criteria and explanatory virtues, but it would do nothing to “explain” our x, the murder of John Dominic.

And there is another problem with your approach to explanation. You say that “we assess whether x fulfills (A), (B), and (C) by reference to a set of explanatory virtues…” But this is too vague. That’s why I prefer my criterion of:

(B) H possesses the following explanatory virtues to a greater degree than any competing potential explanations of E…

But my account of explanation has problems, too. I said that for H to be even a potential explanation of E, it must be true that “if H were true, then E would be a matter of course.” But as commenter Richard Wein points out, this account doesn’t allow for explanations involving chance.

Suppose we want to explain why E = “Person P won the lottery.” It seems the best explanation we can offer would be something like H = “P bought a ticket. The choice of the winning ticket was random, but according to the rules of the game some ticket-holder had to win. And P was a ticket-holder.” That sounds like not just a potential explanation of E, but a good explanation of E. And yet according to my definition it is not even a potential explanation, for E does not follow as a matter of course from H. In fact, it is extremely unlikely that E would follow from H, assuming we are talking about a lottery involving millions of tickets.

The issues of explanation and chance have been discussed at length by Railton, Batterman, Strevens, and others. Also see sections 2.3 and 3.2 of the Stanford Encyclopedia of Philosophy article on Scientific Explanation.

I don’t think you or I will be able to solve the problems of explanation or even understanding - both of which are difficult – in the course of our debate. I just hope we can settle on some agreeable account of explanation so that we can then argue whether naturalism or Christianity offers the superior explanation of the human condition and other phenomena. Even still, offering explanations may not be the end of it. Explanations are like salted peanuts: “Getting one doesn’t make you stop asking for one; usually just the reverse is true.”

What do you think?

Luke,

I’m not sure why my other conditions, especially (B), all of which are conjunctive, do not solve the problem posed in your Zeus examples. Is the word “adequate” in (B) not (pardon me) adequate?

Tom,

I should have asked about this last time, but I’m still a little unclear about what you mean by offering the following criterion for x as an explanation for y:

(A) by knowing x we understand y better than by not knowing x

To “understand” something means “to perceive and comprehend the nature and significance of” something. So I guess that positing Zeus as the cause of lightning would help me understand that the nature of lightning is such that it has a supernatural cause hiding behind its proximate natural cause, and that each lightning bolt has a telos, as determined by the will of Zeus. Zeus sends each lightning bolt for a reason. So knowing that Zeus causes lightning helps me understand lightning better than I would if I did not know that Zeus causes lightning.

Have I got that right?

There’s also the problem of the word “know,” which assumes that x is true (since “knowledge” is considered to be “justified true belief,” ignoring Gettier problems). So maybe we could use:

(A) by positing x we understand y better than by not knowing x

But even the word “understand” presents a problem. Let’s say that the Ancient Greeks had good reason to believe that Zeus was the cause of lightning, and there were no competing hypotheses. But let’s say that Zeus did not, in fact, exist. Could we then say that by positing Zeus these fictional people understood lightning better than by not positing Zeus? I think it would be more proper to say that by positing Zeus they misunderstood lightning. The reason for this is that “understand” is a success term, at least to my ears.

I might be able to propose a less problematic account of explanation by making use of C.S. Peirce’s abductive schema:

(1) The surprising fact, E, is observed.

(2) But if H were true, E would be a matter of course,

(3) Hence, there is reason to suspect that H is true.¹

Any H that fits in (2) I will call a potential explanation. Given that, may I suggest:

H is a good explanation of E if:

(A) H is a potential explanation of E; and if

(B) H possesses the following explanatory virtues to a greater degree than any competing potential explanations of E:

1. Testability – a good explanatory hypothesis should be testable.
2. Consistency with background knowledge – a good explanatory hypothesis should not contradict our background knowledge.
3. Past explanatory success – a good explanatory hypothesis should fit in a tradition with much past explanatory success.
4. Simplicity – a good explanatory hypothesis should be simple, not making a lot of ad-hoc assumptions.
5. Ontological simplicity – a good explanatory hypothesis should not add more unknown things to our ontology than necessary.
6. Informativeness – a good explanatory hypothesis should allow us to deduce precise details of its effects.

How does that sound?

Luke

1. Peirce: Collected Papers 5.189. Peirce used letters C and A, but I use E for evidence and H for hypothesis. My approach to explanation here mirrors that of Gregory Dawes in Theism and Explanation. [↩]

Greetings once again, Luke,

Thank you for sharing more clearly what you meant by things like “werewolf” and “party tricks,” and for your expression of respect for me. I am glad to see this moving back the direction I always expected it to take from the beginning. I can easily accept (without agreeing, obviously!) that you literally believe “a theory of werewolfism could contain less implausible and less ad-hoc hypotheses than Christianity does.” That’s your position, and I accept it for what it is.

As to “party tricks,” though, which you continue press as a literal Christian belief, I still take exception to that. My son is an amateur magician who does party tricks. They’re not intrinsically dishonorable, but they’re not what Jesus did at Cana, either. The term generally connotes something trivial, an illusion rather than something real, done to draw attention to the trickster. I don’t think what he did was trivial (though I know some might argue that it was). It was certainly not an illusion nor was it done to draw attention to himself.

I also have some problems with equating “magical” with “supernatural,” and with viewing God as a wish-granter, but if that is how you viewed it, that is how you viewed it, and who am I to dispute that? Here’s the distinction: if you represent something as your belief about Christianity (that it’s less plausible than werewolf-ism, for example) then it is what it is, you believe what you believe, and that’s not subject to dispute. If you have a strongly negative view of Christianity overall, again, that’s not open to debate, I merely ask that you continue to bear in mind that this website is for “disciplined debate,” which includes respect for persons who hold those beliefs. If on the other hand you represent something as being an actual belief of Christianity, and if, as in the case of “party tricks,” it’s not an accurate representation, then you can expect to be corrected on that.

Now, on to the matter of explanations. I think we can meld your answers and mine and come up with a good overall definition. In my last post I focused more on “what is an explanation?” than on “what makes it a good explanation?” I can attribute this to a discussion going on simultaneously at Thinking Christian, where one commenter has claimed that under no circumstances could God ever be an explanation for anything (“there is no such thing as a supernatural explanation,” and other similar statements following). He has a rather idiosyncratic take on what an explanation is, and that was on my mind when I wrote Letter 11.

One small thing: on further reflection I would revise my statement there to begin,

x is explanatory of y if…
or
x contributes to the explanation of y if …

rather than

x is a good explanation of y if …

The reason for that change is that it makes the qualification I added in my first footnote much less necessary. I still mean to imply that x‘s contribution to the explanation of y is a good contribution provided that (A), (B), and (C) are met.

You say,

Saying that theory x is a good explanation for y is precisely what gives us reason to believe that x is true (as I understand explanation). So to put (B) in the requirements for a good explanation just sets us back to the starting point.

I agree and disagree. We need some reason to believe x is true. Its fit with y is certainly one potential reason (see below), but just one among many.

I did acknowledge and I still agree with your list of explanatory virtues. By them we can assess whether x fulfills (A), (B), and (C). So my suggested statement as a result would be,

x contributes to the explanation of y if:

(A) by knowing x we understand y better than by not knowing x; and if

(B) there is adequate reason to believe that x is true, reliable, trustworthy, etc.; and if

(C) there is no competing z, such that z is contrary or contradictory to x and fulfills the conditions of (A) and (B) more successfully than does x.

And we assess whether x fulfills (A), (B), and (C) by reference to a set of explanatory virtues including:

1. Testability – a good explanatory hypothesis should be testable.
2. Consistency with background knowledge – a good explanatory hypothesis should not contradict our background knowledge.
3. Past explanatory success – a good explanatory hypothesis should fit in a tradition with much past explanatory success.
4. Simplicity – a good explanatory hypothesis should be simple, not making a lot of ad-hoc assumptions.
5. Ontological simplicity – a good explanatory hypothesis should not add more unknown things to our ontology than necessary.
6. Informativeness – a good explanatory hypothesis should allow us to deduce precise details of its effects.

Note that when the argument is of the form, inference to the best explanation, part of the test for x is its fit with y, compared with other candidate explanations’ fit with y. My argument in the last several letters has been of this form, and the test I have been proffering has been a test of fit.

So my guess is that we’re in agreement on what we’re going to call a good explanation, or how we’re going to assess the quality of explanations. Am I right?

Regards,

Tom

Tom,

I think that when I talk about werewolf-theory and about Jesus performing “party tricks” you may think I am disrespecting you. If you could hear me say these things face to face I think you would not feel disrespected, by my tone of voice is not communicated clearly via a text medium.

So allow me to reinforce what I said about werewolf theory and Jesus performing party tricks, as examples. I do not mean these to be flippant, dismissive remarks. I mean them very specifically and literally. I literally think a theory of werewolfism could contain less implausible and less ad-hoc hypotheses than Christianity does. And Jesus literally performed party tricks: turning water into wine at a wedding. So I mean that all quite literally, not flippantly.

I’m doing the same thing when I tell people my deconversion story and I recall the moment when I realized I literally believed I had an invisible magical friend who granted me wishes. I believed Jesus was my friend, I believed he was invisible, I believed he had magic powers (aka “supernatural” powers), and I believed he sometimes fulfilled telepathic wishes I made (prayers). When telling this story I do not mean that statement flippantly or dismissively, either. I mean it very specifically and literally.

And of course it’s possible that I really did have an invisible magical friend who granted me wishes. When I realized this I did not give up my faith. I had many reasons to believe it was true. It was only after these reasons were undermined by argument and evidence that I had to admit it was not true that I had an invisible magical friend who granted me wishes.

With that in mind, let me reaffirm the respect I have for you. Tom, you’re a smart guy. A loving guy. A genuine seeker of truth. I don’t think you’re being willfully dishonest (like some apologists are). And I respect that. I think it’s literally true that a common conception of Jesus is that of an invisible friend who sometimes grants wishes, and I think it’s literally true werewolfism could be less implausible and less ad-hoc than Christianity, but none of that means I disrespect you. It just means I think you’re wrong, for very specific reasons.

Okay, back to the main topic.

We’re at the point again where we need to find some common ground before we can continue. Before I say anything more about whether Christianity or naturalism offers a superior explanation for the human condition, we should get clear on what we think a good explanation is.

You gave three criteria for a good explanation:

(A) by knowing x we understand y better than by not knowing x; and if

(B) there is adequate reason to believe that x is true, reliable, trustworthy, etc.; and if

(C) there is no competing z, such that z is contrary or contradictory to x and fulfills the conditions of (A) and (B) more successfully than does x.

But I’m not sure how this can work. Saying that theory x is a good explanation for y is precisely what gives us reason to believe that x is true (as I understand explanation). So to put (B) in the requirements for a good explanation just sets us back to the starting point. We are left asking, “Well, is that x a good explanation of y, so that we have good reason to believe x is true?”

When you say that e is evidence for x, what you’re saying is that x provides the best explanation for e. Otherwise, e could just as well be evidence for the Zeus theory of lightning as for the electric theory of lightning. The reason e is evidence for the electric theory of lightning instead of being evidence for the Zeus theory of lightning is that the electric theory is a better explanation of e. (And it is a “better” explanation of e in that it has better explanatory scope, explanatory power, and so on – compared to the Zeus theory. See below.)

At least, that’s what I mean when I use these terms. But they are used in varying ways, and that’s probably why we need this cleared up.

So how do we tell which hypothesis (or theory, a collection of hypotheses) is the best explanation of some phenomenon (say, “the human condition”)? Some prefer the approach of Bayesian confirmation theory, such that the best explanation is one who final probability is greater than 0.5. One non-fatal difficulty with this approach is that we cannot assign accurate probability numbers to the terms of the equation. Another is that it seems to set the bar too high. Scientific theories are often accepted before we have reason to believe their probability is greater than 0.5.

One alternative view to confirmation theory is “explanationism.” The explanationist may accept a theory whose probability does not exceed 0.5. Instead, the explanationist may merely require that a theory possess certain explanatory virtues to a greater degree than competing theories.

Scientists and philosophers have developed an indefinite but useful set of explanatory virtues. The list looks something like this:

1. Testability – a good explanatory hypothesis should be testable.
2. Consistency with background knowledge – a good explanatory hypothesis should not contradict our background knowledge.
3. Past explanatory success – a good explanatory hypothesis should fit in a tradition with much past explanatory success.
4. Simplicity – a good explanatory hypothesis should be simple, not making a lot of ad-hoc assumptions.
5. Ontological simplicity – a good explanatory hypothesis should not add more unknown things to our ontology than necessary.
6. Informativeness – a good explanatory hypothesis should allow us to deduce precise details of its effects.

Of course, there are other virtues we might consider. I could also spend time arguing why these explanatory virtues, if fulfilled by explanation x, argue in favor of x being the correct explanation. But I suspect we agree on several or all of these explanatory virtues already, and I have said enough for one post.

So, Tom, are we any closer to a common understanding of what a “good explanation” is?

Your turn!

Luke

Greetings, Luke,

I could also rattle off a long list of things I think are wrong with atheism, if that were our purpose here. I could “give a hint at” why I think atheism is “a terrible explanation for the world,” for indeed I think it is. I haven’t done that (“not even close!”) because that’s not what we’re about here. So I’ll take it as duly noted that you believe Christianity is a “tangled mess,” etc., and move on to the discussion we’ve agreed to have.

I’m also not going to get into your misrepresentations of Christian theology, because that would just throw us off track. That discussion is proceeding just fine without me in the comments.

Now for the meat of it. You said,

I think the natural explanations for these phenomena are plausible, while the Christian explanation is absurd. You think natural explanations for these phenomena are weak, while the Christian explanation for them is strong. Therefore I suspect we have very different ideas about what makes an explanation strong or weak, good or bad…. Personally, I tend to think of a good explanation as one that possesses many explanatory virtues: testability, consistency with background knowledge, simplicity, informativeness, and so on.

I would center my conception of explanation around this:

x is a good explanation for y if

(A) by knowing x we understand y better than by not knowing x; and if

(B) there is adequate reason to believe that x is true, reliable, trustworthy, etc.; and if

(C) there is no competing z, such that z is contrary or contradictory to x and fulfills the conditions of (A) and (B) more successfully than does x.

That’s a fairly simplified and generalized picture of what I think explanation is about. It leaves open the door for variations like tentative explanation (the best explanation we have so far, where (B) could possibly be true, but is not well established), multiple coinciding explanations (where z is complementary rather than contrary or contradictory to x), and so on.¹ It also provides for some comfortable flexibility in how we determine whether (A), (B), and (C) are each true. Certainly your list of explanatory virtues is relevant.

Now I think that Christianity certainly meets condition (A) for the phenomena I focused on in my last two posts. The Judeo-Christian doctrines of the imago Dei and the fall certainly help us to understand the confusing, paradoxical human condition. We were created in dignity, love, and moral innocence for an intimate relationship with a holy God and with others of his creation. We rebelled against God and fell away from that state of complete dignity, love, and relationship; and we feel the sting of it in the form of alienation, moral failure, and multiple indignities including physical illness and death. Our longing for something better or something more is a sense that is built into us by virtue of being in God’s image. Our moral awareness, which is largely (not perfectly) in tune with truth, is a reflection of God’s moral character still stamped in us. Our successes in love, giving, sacrifice, etc. are real successes in living in tune with that character as it is known to us through conscience and natural law (and also, for those who know and accept it, by revelation from God). Our awareness of failure, on the other hand, is a result of really not matching up to a really existent standard. It’s an awareness that is not born of our own imaginings, it’s not a mere social construction, it’s not the result of evolution, but it is based on a standard that has stood forever.

Knowing our origin and state—created in the image of God but fallen—certainly gives us a better understanding of our paradoxical condition than not knowing it does. Thus it meets part (A) of my definition for an explanation. Does it meet the test of (B)? I think so. I know you disagree. I cannot present the whole case for that here; it would require having our whole discussion about all of Christianity at once! I am content to show that it meets (A), but not just that; that it meets (A) better than the competing naturalistic explanation. If it succeeds in that, then that by itself contributes to the case (B) for its being true.

The naturalistic explanation for the human condition seems problematical to me. Yes, I know there’s more than one naturalistic explanation, but they all share the view that every cause that produces every effect, including every human feeling, thought, or behavior, is purely natural: matter and energy interacting according to necessity (natural law) and chance (quantum effects).² The naturalistic worldview entails the causal closure of the physical; that is, that all causes are physical. Matter and energy interacting by law and chance constitute the entire ultimate causal picture for everything.

More proximately, the cause of all human feeling, thought, behavior, etc. is to be found in evolutionary processes that are built in randomness and seek and reward³ only reproductive fitness. If there is any other naturalistic explanation available for any organism’s characteristics, well, it hasn’t been suggested yet.⁴ Random variation and reproductive success constitute the entire causal schema for all organisms’ attributes. This, too, is causal closure. On this level of analysis, there are no other causes for human behavior.

The naturalistic explanation requires one to suppose that the painful, paradoxical experiences that go with being human led to reproductive success sometime in our species’ natural history. Now, I have trouble seeing how a purely natural process could have thrown up any reliable sense of right or wrong whatever: it’s just not in the nature of physical things to be right or wrong. I can imagine natural processes producing a sense of right or wrong, but not a reliable sense; for there is nothing about natural processes that could be expected to produce such a thing, in a way it could be relied upon as true. So on naturalism, I have to take it that our sense of being somewhat right, somewhat wrong; somewhat good, somewhat bad; somewhat dignified, somewhat dirt; somewhat eternal in value, somewhat eternally flawed—all of this is just a sense, and there is no reason to consider it the least bit reliable. It is only a sense, one that was developed and preserved in humans just because it served our ability to make more babies that made more babies.

I have spoken of two levels of explanation for human behavior. On the most ultimate level, all causation is accounted for in terms of physics, interactions of particles by lawlike or chance processes. On a more proximate level, all causation relevant to human behavior is accounted for by random variation and natural selection (reproductive success). This does not obviate or contradict the most ultimate level of causation. Evolution expresses physics, at a more complex level of organization, it does not deny it or override it.

Of course there are still other levels of explanation, including our mental processes of belief-forming and decision-making; but on naturalism, given the causal closure of the physical and the causal closure (on another level) of random variation and reproductive success, it is impossible for this level of explanation to escape the causal necessities of the more ultimate levels of explanation. Therefore whatever mental processes we think are influencing our behavior, those mental processes are ineluctably and inescapably caused by physical interactions of particles (on one level) and by randomness and reproductive success on another level. There is no other causal stream contributing to human attributes, and thus no other causal stream contributing to human mental processes.

To state it baldly: all of my and your most human mental attributes—our aspirations, hopes, dreams, confusions, moral victories, moral shames—do not exist because we really do have dignity, worth, moral value, etc. They exist in us because attributes like this helped our ancestors reproduce more successfully and for no other reason.⁵ This is the conclusion Dawkins draws in The Selfish Gene, and if naturalism is a true picture of reality, then I don’t see how he could be wrong on this.

Further evidence for my position is found in the immense difference between humans and animals in this respect. Animals do not agonize over unfulfilled aspirations, and they don’t mourn their moral failures. They don’t reach for something higher. What could evolution have done to some population of primates to produce this in homo sapiens? It would have had to introduce in us a sense that some things were right, and other things were wrong. But is it plausible to think of this happening in that population of primates, when what evolution cares about⁶ is getting the male and the female together in some environment where they’ll make babies that will live to make babies? It works awfully well in animal populations without all this mental anguish of hope and failure; why would evolution have bothered to do more than that for humans? Where is the reproductive advantage in it?

So naturalism could plausibly explain the sense of human aspiration and failure, but it cannot explain it as a reality. It can tell how we feel worth and dignity, but it certainly cannot say how we really have worth and dignity beyond those feelings. If that sense or feeling is actually tapping into something real, then naturalism’s explanation for it is inadequate. Or as I put it last time, what we know to be true about the human condition fits within the Christian worldview. It takes some serious Procrustean hammering to make it fit with naturalism.

1. I’m also saying we can call x a good explanation without requiring that it be the whole explanation. The pot is boiling on the stove because I want to make spaghetti is a good explanation, even if it doesn’t make reference to how the electricity was supplied, how the resistive stovetop coil produces heat, how heat makes water boil, etc. The greater the explanatory scope of x, the better an explanation it is, of course. [↩]
2. I have seen attempts by naturalists to avoid the such reductionism, but I’ve never seen any that succeeded in avoiding self-contradiction. [↩]
3. Please pardon the anthropomorphism [↩]
4. Genetic drift is just randomness writ large. [↩]
5. There actually is another possible reason: that these things came up as a co-adaptation along with some other human feature that served purpose of reproductive success. If so, and if there is any actual truth to be found in this co-adaptive state, that’s just marvelously lucky indeed. The odds are not in favor of such a thing happening. [↩]
6. I know, I’m anthropomorphizing again [↩]