A demarcation principle for Miracles

Randal Rauser has proposed that we may be able to use a design filter as an epistemic tool for identifying miracles. What he means by a design filter here is really just an adaptation of William Dembski’s notion of ‘specified complexity’. The inspiration for this seems to be in part because Rauser realizes that the term ‘miracle’ did not, to the pre-scientific mind of the Biblical authors and their audience(s), entail something as strong as a violation of, or an exception to, the laws of physics, or anything like that. A miracle was an act of God (whether mediate or immediate) signifying something with theological or revelatory import. Therefore, it’s not out of the question that God can be said to perform a miracle even where one can give a satisfactory physical account of the happenstance. For instance, when a pastor prays to God for money needed to keep a church or ministry going, and the money comes in just in time; perhaps one can tell a causal story about how the money happened to be on the way before the pastor ever prayed for it, but that may not take away from the pastor’s justified belief that God had answered her prayer.

I think this is an extremely interesting proposal. Randal Rauser is in effect offering us a tool for justifying the inference of a miracle from the principle by which we can presumably make a legitimate design inference. One merely needs to identify three elements: 1) contingency, 2) complexity, 3) specificity. For those not already familiar with Dembski’s work on specified complexity, I will direct readers to his book. Now, Rauser’s proposal has received some criticism, for instance by Jonathan M.S. Pearce. However, I think Pearce says nothing directly opposed in principle to Rauser’s recommendation. Indeed, as I’ve followed the relevant blogs, posts and the overly rhetorical comments I simply found Pearce and others attacking the suggestion that any anecdotal examples Rauser put forward were really improbable after all. However, that just is to say that they don’t pass Rauser’s own standard. Maybe nothing passes Rauser’s design filter – that wouldn’t be any sort of defeater for the filter itself (even if it may be a weak defeater for miracles).

Rauser provides the following thought experiment on his blog (to which I’ve already linked):

Let’s say that Dave lives in a pre-scientific culture in the year AD 500. Dave is a missionary to a barbarous people and is about to have a power encounter with their head witch doctor in which he aims to show the superiority of the Christian god over their tribal deity of blood and soil… [Witch doctor does some magic] Dave is undeterred. “Yahweh, the one true God, is superior!” Dave says. “Watch him darken the very source of all fire, our sun!” And at that moment an eclipse which would have been predicted by able Ptolemaic or Newtonian astronomers, had any been around to predict it, commences. The crowd gasps at the awesome display.

Would it be rational for the audience to conclude that they had witnessed a miracle at the hands of Dave’s deity? Of course it would.

And would it in fact be the case that they had witnessed a miracle at the hands of Dave’s deity?

That depends. Presumably if Dave’s deity exists then he set up the scientific laws in the first place, and he did so knowing that at precisely the right moment Dave would predict an eclipse which would in fact occur due to those scientific laws.

Or, if we’re libertarians, God set the world up so that Dave would have the libertarian-free choice to trust God’s direction imparted by the Holy Spirit when God invites Dave to say such a thing. God doesn’t determine, nor is it determined by anyone other than Dave, that Dave will proclaim thus to the indigenous audience.

In the example above, all Randal Rauser is attempting to establish is that a person can be epistemically justified in inferring that a miracle has occurred, even where the laws of physics have not been (apparently) ‘broken’ or ‘suspended’, and that there may in fact be some instances of miracles which would be correctly inferred.

There may be some significant advantages to using Rauser’s design-filter to identify miracles. For instance, one who is so Humean as to think that nobody can ever be justified in believing that the laws of nature have been suspended (i.e., that the universe is ALWAYS a closed system), may find herself having to confront the resurrection of Jesus as a miracle to which our best physics assigns the quality of physical possibility. Whether there are any scientific laws which hold with nomic necessity (whose application yields deductive closure) and which allow room for a resurrection (making it a physically possible event), or not, nothing of significant import follow from this; regardless of whether such laws can be cited in one’s explanatory account of the resurrection, the inference form Jesus’ resurrection to the hypothesis that in the person of Jesus of Nazareth God revealed himself definitively to the world (and ultimately vindicated his claims by raising Jesus from the dead), remains entirely legitimate. I am, of course, very tempted to say that there are no actual laws of physics (that is, no ‘correct’ laws of physics) stipulated with nomic necessity which make it possible that Jesus rose again naturally from the dead, but the point is that even if there were such laws one could legitimize calling the resurrection a miracle precisely because it passes this demarcation principle.

If one can use this demarcation criteria for miracles then one may also, in principle, be able to distinguish paranormal activities from miracles, (thus there is no fear of taking any miraculous paranormal activity as the explanandum of scientific investigation, or even not being sure whether to take such events as scientific explananda). I think that this principle has a lot of mileage; it is both a brilliant demarcation principle for miracles, and also one for which I cannot think of any available defeater.

So, not only is a miracle not an exception to any scientific law which is stipulated to hold with nomic necessity (as I have argued and explained previously), but even where some event can be explanatorily captured by such scientific laws it may nevertheless remain genuinely miraculous. The demarcation principle is independent of whether one can give a satisfactory scientific explanation of the physical events in question, so that the question of whether what occurs can be captured by the language of science is just superfluous to the question of whether what occurs is miraculous.

Advertisements

About tylerjourneaux

I am an aspiring Catholic theologian and philosopher, and I have a keen interest in apologetics. I am creating this blog both in order to practice and improve my writing and memory retention as I publish my thoughts, and in order to give evidence of my ability to understand and communicate thoughts on topics pertinent to Theology, Philosophy, philosophical theology, Catholic (Christian) Apologetics, philosophy of religion and textual criticism.
This entry was posted in Apologetics, Epistemology, Philosophy, Philosophy of Religion, Theology. Bookmark the permalink.

10 Responses to A demarcation principle for Miracles

  1. I’ll look this over and comment when my workload calms down!

    Cheers

    JP

  2. I can think of a better criteria for inferring design: background knowledge of intentions. If “specified” means intentions, then we’re on the same page–though I think Dembski’s meaning is different. Can you give me an example of a design inference that isn’t based on background knowledge of intentions? I’m pretty sure there are examples of specified complexity that aren’t regarded as designed.

    So, when you infer God, I assume you’re inferring omniscience, omnipotence, and omnibenevolence, and not just intelligence simpliciter. It doesn’t seem possible that any observation can justify those 3 properties. If there appears to be a more optimal way to achieve the posited divine goals, then it seems we can’t infer the tri-omni God; because an tri-omni god would know the best way to achieve the goal, has the power to do it, and should be assumed to be instrumentally rational. I’m not saying that an observation rules out a tri-omni God; I’m asking if the inference to tri-omniness is reasonable.

    There’s another reason to deny theistic explanations: they have a bad track record as a research programme–they don’t have a history of predictive success. In the Dave pre-scientific example, he may be justified, but we’re no longer in that position.

  3. An interesting piece. Ideas of warranted true beliefs aside, this is essentially about arguing what constitutes specified complexity, much akin to arguments about ID. Now, especially if one invokes skeptical theism such that it is hard to decipher the intentions of God (theists love to invoke this to answer the problem of evil, but are less than happy when it is reversed on them, such as with the work of Weilenberg), we have a problem of subjectivity qua arbitrariness concerning where to draw the line of what is considered specified enough to qualify.

    In principle, this actually becomes a Bayesian style probability argument (unless God comes down and announced to the whole world that it was, indeed, a miracle, though this would still be data thrown into a BT calculation).

    It literally becomes a case of assessing the claims on available data. Thus Dave’s contemporaries would be justified, on their knowledge, in believing it to be a miracle (damn, returned to epistemic justification!) but later people might derive a different calculation given background knowledge etc.

    I don’t think anyone like myself would a priori rule out miracles. It’s just that the standard of evidence must match the standard of claim. BT is merely a codification of extraordinary claims require extraordinary evidence. miracles, by the improbable nature of their claims, require exceptionally high standards of evidence.

    Resurrection claims of Jesus are not good quality evidence. In fact, the Gospels represent pro-Jesus ex post facto believers evangelising an agenda, written by unknown people in unknown places and times. Yes, we can intelligent guess, but that is about it.

    The resurrection of Jesus may have happened, but given the standard of evidence available to us, and the prior probability being exceptionally low given the torrents of similar false miracle claims of other religions, then we are not justified in believing in its truth.

    That takes cognitive biases…

    • “BT is merely a codification of extraordinary claims require extraordinary evidence.”

      This is just mistaken methinks. What one needs in order to justify an extraordinary claim is for the evidence (however ordinary) to make it the case that the conditional probability of the extraordinary claim, given the evidence, is more likely than not. More pedantically and precisely, considering an extraordinary event X, and the evidence Y, what we need is for P(Y|X)>>P(Y|~X) regardless of the probability of X simpliciter (meaning on only the background knowledge, which I’ve omitted above). In other words, the probability that we would have the evidence we do have (however ordinary) on the supposition that X did not occur, is much much less probable than the probability that we would have the evidence we do have on the supposition that X did occur. However unlikely prima facie X may be, one need not have an equally extraordinary body of evidence Y in order to justify it. Think of the lottery paradox – supposing somebody reads the number off their ticket and comes to believe it is the winning ticket: one may think that it is always more likely that they are misreading the ticket wishfully, or that they are being punked, or some such more plausible explanation is the case. However, that would mean that lottery winners are always acting epistemically irresponsibly in claiming the prize based on the belief that they have won. The fact, though, is that they don’t need extraordinary evidence in order to believe they have the winning ticket (such as the evidence of reaping the benefits), but rather they just need it to be the case that the evidence they have in hand is less likely on the supposition that they haven’t won, than it is on the supposition that they have won. That all seems to me introductory philosophy of probability stuff. This may or may not be the case with the resurrection (as you no doubt know, the trouble with the probability calculation is always with the ‘background knowledge’). However, the evidence for the resurrection doesn’t come from taking the evidence of testimony in the Gospel accounts – that’s also just a misrepresentation of the argument for the resurrection.

      Also, as a student of Theology, I can’t pass up the opportunity to point out that the historical-critical method of textual criticism has given us a very good idea about who wrote what when and where, and the Gospels end up looking very early, and not unlikely written by the autographs they traditionally carry, or else written in close connection to communities established by those figures. The scientific approach to scripture scholarship makes much less of a mystery of these questions, and we can have reasonable confidence in the method.

      Finally, I just wanted to register a note about epistemic vantage points. You write “we have a problem of subjectivity qua arbitrariness concerning where to draw the line of what is considered specified enough to qualify.” On the one hand, I am in deep agreement with you here (I note again in passing though that this doesn’t disqualify the principle in principle, but only poses a problem on which work is wanting), however, I would like to suggest that there may be an easier response to this than you imagine. In the problem of evil we can have only at best speculative views about what morally sufficient reason God has for allowing some evil (of course if God does exist it follows with certainty that he does have a morally sufficient reason, and thus that gratuitous evil cannot exist in any logically possible world). However, for something to be a miracle seems here to mean that it is intended as a sign to somebody. What makes a miracle a miracle is not that it defies scientific expression (if it even does), but rather that it is an act of God whereby he signifies, communicates, something to somebody or some body (definite or indefinite) of persons. However, that seems to require subjective probability assessments. Suppose that the laws of physics have it that that in three years I will cry out for a sign of God’s presence in my life, only then to be met with some, to me very significant, event, an event which impresses it’s significance on me. I cannot help but read the event through the lens of my personal history, and cannot help but take it as a sign. Now, the laws of physics which in the first place had it that I would cry out for such a sign are the same, let us suppose, that had it that the significant event would happen as it did and when it did. Wouldn’t I still be justified in believing that this event was a miracle, even if I was later made aware that the ‘event’ was as ‘inevitable’ as my desire for it? It seems to me it would. The point is this, in the case of evil, we just aren’t at an epistemic vantage point which allows us to say with any certainty whether there is a morally sufficient reason in God’s mind for it. However, I am in a privileged epistemic position for discerning what, to me, would be a sign from God. It seems to me that nobody in all the world is better situated for the task of assessing the plausibility of something’s being a sign to me than me myself. The fear of subjectivism may be felt here and combined with the too pervasive theophobia of the western modern intelligentsia, I can appreciate the reasonable aversion you and others may feel. However, it seems to me that there’s nothing for it, if there are miracles this is the only way they can be discovered. Indeed, if no miracle is ever thus discovered, then no miracle exists!

      • Hi there

        This is just mistaken methinks. What one needs in order to justify an extraordinary claim is for the evidence (however ordinary) to make it the case that the conditional probability of the extraordinary claim, given the evidence, is more likely than not.

        Yes, and in one sense no. It has to be the most likely hypothesis of all hypotheses, as well as being more likely than not. The idea of a miracles claim is the extraordinary low prior probability. Thus the level of evidence must be high to overcome that.

        So do sets of specified complex situations, for want of better words, qualify as good quality evidence? And does such a hypothesis remain more probable than all competing hypotheses?

        Any miracles claim is going to fit in with that kind of specified complexity, irrespective of the religion. We can think of any number of rival miracles claims which, given the claims about the particular background knowledge, may seem more likely to be true. It comes down to the robustness of the claims about the background knowledge.

        What I am saying is that given the failure of so many other previous global miracle claims to, say, one in the bible, we are justified in thinking that the likelihood of the biblical one was low. We would expect such a claim, as all others would have been, to have the criteria of coherence. But I suppose the notions themselves with which the miracle claim must be coherent must themselves be called into question.

        Think of the lottery paradox – supposing somebody reads the number off their ticket and comes to believe it is the winning ticket: one may think that it is always more likely that they are misreading the ticket wishfully, or that they are being punked, or some such more plausible explanation is the case.

        OK, the lottery paradox is a big can of worms, and we don’t want to start getting on to formal logical structures of the logic and probability of aggregates…

        Suffice to say that they are justified in believing someone will win, but that it will not likely be them. However, the likeliness is made more probable on account of the evidence at hand (thus warranted knowledge isn’t just mere probability). The evidence of having the ticket, and perhaps asking someone else that they are not seeing things, is a pretty good standard of evidence.

        they don’t need extraordinary evidence in order to believe they have the winning ticket

        This is where I disagree – they do. But that first hand witness evidence of something which does not defy known natural laws, and which is easily verifiable form an independent witness, qualifies as a very high standard of evidence, at least for that person themselves.

        The problem with evidence is that when you multiply entities, that standard does not hold across those people, from the point of view of the far removed people themselves. Think of me reading the Gospels compared to the evidence of someone seeing firsthand a miracles of Jesus. The fact is, we have very poor access to such people.

        This may or may not be the case with the resurrection (as you no doubt know, the trouble with the probability calculation is always with the ‘background knowledge’). However, the evidence for the resurrection doesn’t come from taking the evidence of testimony in the Gospel accounts – that’s also just a misrepresentation of the argument for the resurrection.

        Well, if one rejects background knowledge assumptions (ie that God exists, that that God is the one of the Old Testament and New, that omni characteristics make sense, that free will makes sense and is necessitated by a non-Calvinist view of that God etc etc), then it really does come down to that evidence. And it’s not very strong at all. For example, Sathya Sai Baba has committed more, and better attested miracles in contemporary society with many, many more eyewitnesses than Jesus ever afforded, and yet we reject those claims.

        I think these sorts of problems are compounded by the very circular nature of faith and appraisal of evidence of the New Testament as set out here:
        http://www.skepticink.com/tippling/2013/04/26/the-circularity-of-believing-the-new-testament/

        If I believed all the things about God, the universe, and the New Testament as you do, then I would find the Resurrection claims to be thoroughly persuasive. I just think that is a circular enterprise, and I thoroughly reject all the ‘background knowledge’ which gets thrown into the BT analysis.

        Also, as a student of Theology, I can’t pass up the opportunity to point out that the historical-critical method of textual criticism has given us a very good idea about who wrote what when and where, and the Gospels end up looking very early, and not unlikely written by the autographs they traditionally carry, or else written in close connection to communities established by those figures. The scientific approach to scripture scholarship makes much less of a mystery of these questions, and we can have reasonable confidence in the method.

        We can have intelligent guesses, but these are thoroughly biased by our cognitive faculties. Do we really know who wrote them? More importantly, do we know their sources? Do they use critical historiography? Even Suetonius used better historiography than all of them, Luke included, and half of what he said was bullshit – and he was a historian! The Gospel writers were not. They were ex post facto evangelists. Would I get my historical information on David Koresh from a follower convert of his who had never met him, and who wrote a ‘biography’ of him at least 40 years after his death AFTER coming to believe him to be the Messiah? Not on your nelly.

        Because that is what is going on here.

        Finally, I just wanted to register a note about epistemic vantage points. You write “we have a problem of subjectivity qua arbitrariness concerning where to draw the line of what is considered specified enough to qualify.” On the one hand, I am in deep agreement with you here (I note again in passing though that this doesn’t disqualify the principle in principle, but only poses a problem on which work is wanting),

        Sure, I don’t think I have disagreed with the principle per se.

        however, I would like to suggest that there may be an easier response to this than you imagine. In the problem of evil we can have only at best speculative views about what morally sufficient reason God has for allowing some evil (of course if God does exist it follows with certainty that he does have a morally sufficient reason,

        Well, if YOUR particular version of God exists. But Christians cannot even agree on that. An open theist would surely have different ideas on the allowance of evil than a classical theist, no?

        The point is this, in the case of evil, we just aren’t at an epistemic vantage point which allows us to say with any certainty whether there is a morally sufficient reason in God’s mind for it.

        Absolute certainty aside, the evidential problem of evil sets out to state that God probably doesn’t exist, which means that it is a rationally warranted judgement.

        However, I am in a privileged epistemic position for discerning what, to me, would be a sign from God. It seems to me that nobody in all the world is better situated for the task of assessing the plausibility of something’s being a sign to me than me myself.

        What was the quote, “The easiest person in the world to fool is yourself”?

  4. It seems you and Pearce are disagreeing on whether to use Bayesian analysis or likelihoods P(X|Y) vs P(Y|X). I’m hardly an expert on this and there are people on both sides.

    “…considering an extraordinary event X, and the evidence Y, what we need is for P(Y|X)>>P(Y|~X) regardless of the probability of X simpliciter.”

    To take a page from Sober, let’s take X to be “gremlins are bowling in my attic”; let’s take Y to be “hearing noises in the attic.” It seems P(Y|X) = 1, and P(Y|~X) seems inscrutable. Seeing as how we don’t have any independent evidence for X, we would probably not believe that gremlins are bowling in my attic even though P(Y|X) = 1.

    “In the problem of evil we can have only at best speculative views about what morally sufficient reason God has for allowing some evil.”

    On your view, shouldn’t every natural disaster be a case for celebration because of the greater good achieved? If a tree falls and crushes a baby, isn’t a greater good achieved?

    • Morally sufficient reason does not a best of all possible worlds make. Granted there is no such thing strictly speaking as the best of all possible worlds, generally we want to say that some worlds are better than others. A world in which a tree falls and crushes a baby may be a maximally good feasible world for reasons we can’t understand if the Molinist is right, but the occasion of evil is not a cause of celebration. Obviously some greater good is achieved, but it could have been achieved more ideally (the catch here is that it is achieved less-than ideally because of the fault not of God but of some set of libertarian-free agencies). Note that the libertarian-free agents may be situated in the future, or in the past, or perhaps may not be human (spiritual beings like angels and demons count as libertarian-free if they exist).

      I am also no expert on probability theory.

  5. Lots to talk about here. Off to work now, will try to reply tonight.

Leave a Reply

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