THEISM AND EXPLANATION: Degree of testability, coherence with the rest of our knowledge, previous success of the research tradition to which it belongs, simplicity, ontological economy, informativeness.

CHAPTER 1:

Could an explanation that invokes a divine agent be a good explanation?

Could the existence and action of a divine agent be the primary causal factor in a satisfactory explanation?

TWO KINDS OF NATURALISM:

1-Epistemic Naturalism: (Quine) “the recognition that it is within science itself, and not in some prior philosophy, that reality is to be identified and described” “the most we can reasonably seek in support of an inventory and description of reality is testability of its observable consequences.” This is not inconsistent with theism; of positing the existence of God were to offer some ‘indirect explanatory benefit’ we could embrace this without giving up naturalism. This kind of naturalism is a view of how we gain epistemic access to reality.

Many agree with Quine that ‘the only means we have of figuring out what the world is like, is our experience of the world and our explanatory theorizing about it.”

2-Ontological Naturalism: there is nothing besides nature, nothing in addition to nature, nothing outside or beyond nature.

3-Metholodological Naturalism: We must proceed in our endeavor to understand the universe AS IF there were no supernatural agents. Or more strongly yet, we should proceed AS IF ontological naturalism were true. However, this does not specify the content of our knowledge.

Could you be a methodological naturalist without being an ontological naturalist? Some say yes, such as Michael Ruse, while Phillip Johnson says no. Dawes thinks that if the naturalistic research programme of the sciences has been overwhelmingly successful, then one could argue that the best explanation of its success is the truth of ontological naturalism. Indeed, he says that methodological naturalism is practically indistinguishable from ontological naturalism. But, since it is a procedure rather than a proposition it does not actually entail ontological naturalism.

RESPONSE:

1-Methodological naturalism can and does support theism. Both the kalam cosmological argument and the fine-tuning argument garner justification for their premises according to epistemological or methodological naturalism, but entail the contradictory of ontological naturalism.

2-http://www.apologeticsinthechurch.com/27/post/2011/11/naturalistic-induction-replacing-supernatural-explanations-with-natural-explanations.html

3-This argument has a couple of assumptions that seem dubious to me. One, I don’t think we have any way to know the antecedent probability of how many supernatural explanations there would be in the universe on theism compared to naturalism. Second, a miracle, by its very nature, wouldn’t ever replace an established natural explanation because in that case it would cease to be a miracle and become a new law of nature. But let’s try to assume we can know something of the antecedent probability of how many miracles there would be if theism were true. I think there are a number of reasons we shouldn’t be surprised that there are laws of nature that operate ‘uninterrupted’ regularly. It would be irrational for God to set up the laws to constantly be suspending them. Second, it would betray a lack of perfection to constantly have to tinker with them. Third, the laws of nature must operate regularly in order for us to be able to make rational decisions. Fourth, just because God COULD do something doesn’t entail that he SHOULD do it. It seems God shouldn’t perform a miracle without having a sufficient reason such as confirming a message that cannot be read off of nature (e.g. the resurrection of Jesus after his ministry), and the occasions where such a miracle should be done may be as low as one. I think this shows that the antecedent probability on theism for there to be a HIGH FREQUENCY of miracles is (self-defeating for one thing, but also) different than the antecedent probability on theism to expect some PARTICULAR INSTANCE (however infrequent) of a miracle that we do or don’t have evidence for, and it is the latter that needs to be shown for this type of argument to work; especially in light of the considerations being raised here. Fifth, there can be and I think there is good evidence for God’s existence that doesn’t require a supernatural explanation to REPLACE a natural one, but is instead complementary or constituted by both a proximate natural explanation and an ultimate supernatural explanation. Sixth, this objection is basically saying that it is very unlikely on theism that there would be laws of nature! What? Quite the opposite according to some metaphysical arguments. Seventh, it seems that this argument begs the question since there are supernatural explanations of the origin of the universe, the fine-tuning of the universe, and the resurrection of Jesus. But, as I say, the main problem seems to be that both the NUMBER and NATURE of supernatural explanations for events in the world is supposed to be x-number of law-like events, but since we don’t observe this, this confirms metaphysical naturalism over theism. I don’t know how you could know this and in fact, there are plausible considerations to think it is wrong.

AN A PRIORI COMMITMENT? Natural explanations may exhibit certain explanatory virtues that religious explanations lack such as testability, simplicity, economy, and precision which would be a prior considerations; not a posteriori. However, given the success of the sciences and the failure to find any evidence in support of alternative views this a prior commitment is not dogmatic or merely presupposed.

Richard Swinburne thinks that personal explanations (Dawes calls them intentional explanations) are not, and cannot be reduced to scientific explanations. Dawes thinks that such explanations yield testable predictions and so can be regarded as scientific. The real question though is whether they are adequate explanations and Dawes and Swinburne agree that they can be.

DE FACTO OBJECTION TO RELIGIOUS EXPLANATIONS: Religious explanations are acceptable IN PRINCIPLE but IN FACT we have insufficient reason to accept them. Religious explanations represent a failed

research tradition, and need not be taken any more seriously than Ptolemaic astronomy. However, such an objection is provisional at best, rather than permanent.

IN PRINCIPLE OBJECTION TO RELIGIOUS EXPLANATIONS: An objection that seeks to show that a theistic explanation does not meet some condition that must be met by any successful explanation.

1-Darwin and Huxley argued that all explanations must depend on laws and that a Creator by the nature of the case can exemplify no law.

2-Lawson and McCauley argue that propositions about God are ‘semantically anomalous; insofar as they are open to an indefinite number of interpretations which allows them to explain any possible state of affairs. This is a vice rather than a virtue. Basically, they cannot be falsified since they explain everything in that they are consistent with any possible state of affairs.

3-Pennock argues that since God is capable of miracles we don’t know how to connect an effect with the will of a supernatural agent so there is no way to disconfirm religious explanations.

4-Dawkins argues that any explanation that leaves unexplained the existence of their explanans are not good explanations. Every successful explanation should also explain its own explanans.

RESPONSE to 1: Appendix (3.3)

Unless one holds that causation is NOTHING OTHER than a regular associateion of events, there is no reason to deny the possibility of singular causation; an event ‘could be unique in the history of the universe, and yet be, and be known to be, a case of causation (e.g. the origin of the universe, the fine-tuning of the universe, the singularity).

Epistemological Issue: Intentional explanations need not cite causal laws (even if the facts to which they refer could also be explained using the language of the natural sciences)—Appendix 3.3.1

It is true that one could use an intentional explanation to construct a general law: If there existed antoher rational agent of precisely this type (having the same beliefs and desires) then when placed in exactly the same situation he would act in the same way (this entails a potential law-like regularity). However, this is pretty loose. The more important point is that intentional explanations do not rely on generalizations linking intentions and actions. What they rely on is a calculation of how a rational agent in this situation would be expected to act, given certain beliefs and desires from which it follows that our expectation regarding an agent’s behavior is not based on any general intention-action rule.

If a theist is offering a singular explanation, we cannot test it by replicating condition C and seeing if event E is observed, but this doesn’t mean it is untestable (Appendix 3.3.2). There are two descriptions of an event, one which is ‘physical’ and one which is intentional. After all, causation is a feature of the world, not a feature of our descriptions of it so that it would embody the representational fallacy which happens when we wrongly assume that we can deduce the structure of the world from the language we

use to describe it. It is on the level of the event that there exists a cause; we need to abandon premise 1 then, (causation involves law-like relations).

RESPONSE TO 2 / 3 senses of Unfalsifiable: 1-There are no possible observations that would demonstrate its falsity; 2-The opponents of some theory adopt all manner of AD HOC strategies to avoid admitting that their theory has been falsified. 3-An explanation that lacks empirical content is unfalsifiable; but even then sometimes a theory can still be made falsifiable by a more careful formualtion.

Assuming a rational agent is at work with certain beliefs and desires, we can have certain predictable consequences that will extend beyond the fact to be explained; an intentional explanation will be independently testable.

RESPONSE to 3 (4.4): If the theist is to have any trust in our knowledge of the world, he is bound to assume that his God will not work TOO MANY miracles. This objection presents us with a false dichotomy. It assumes that the world is either governed by natural laws or it is the product of mere whim on the part of a supernatural agent. The world could also be the work of a supernatural agent acting rationally in order to achieve his goals so that we can predict behavior is we cite a goal. What would be the most rational way for such an agent to act in order to achieve this goal? The answer may or may not include miracles.

RESPONSE to 4: Many of our most successful explanations raise new puzzles and present us with new questions to be answered. Lipton remarks, “a drought may explain a poor crop, even if we don’t understand why there was a drought;…the big bang explains the background radiation, even if the big bang is itself inexplicable.

Dawes isn’t going to survey all the in principle objections ever offered, but will instead start from scratch. He wants to know the form and content of religious explanations; what kind of agent do they invoke, under what conditions, if any, would invoking such an agent be explanatory? We cannot rule them out in principle since a religious explanation would be yet another theoretical explanation, positing the existence of yet another unobservable entity. Such explanations may have no explanatory force, of is they have some they are not the best explanation. Dawes wants to spell out the conditions that a successful religious explanation would have to meet to show that it is unlikely that any proposed religious explanation would be successful which implies that religious explanation should be, at best, a last resort.

CHAPTER 2: On Explanations in General

-What Dawes will argue is that there are certain features that all adequate explanations will share.

RELIGIOUS & THEISTIC: Dawes is interested in theistic explanations which posit ‘a person without a body who is necessarily eternal, perfectly free, omnipotent, omniscient, perfectly good, and the creator of all things. He is the ‘greatest possible being.’

PROPOSED, POTENTIAL, & ACTUAL EXPLANATIONS: ‘Explanation’ can be thought of as a success term to be used of theories that have been shown to be true or worthy of provisional acceptance. On this account, theistic explanations get ruled out once they have been superceded. The problem is that this doesn’t allow for false explanations and it undermines the distinction between a de facto and in principle objection to theistic explanations. The de fact objection you recall holds that some theistic explanations have some force but that we no longer have sufficient reason to regard them as true. Moreover, talking about explanation as a success term doesn’t allow for false explanations or better and worse explanations. So, Dawes prefers the term potential explanation, or that the best way to defend a theistic explanation would be by way of ABDUCTIVE REASONING:

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

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

3-Hence, there is reason to suspect that H is true.

POTENTIAL EXPLANATION: Any proposition which satisfies H is a potential explanation of E; if it were true, it would make the fact to be explained intelligible.

ACTUAL EXPLANATION: It is true, or at least, we have sufficient reason to regard it as true. This will be so when two conditions are fulfilled: 1-It must be a potential explanation, and 2-We must have sufficient reason to accept it. Lipton rejects this because it entails that all true potential explanations are actual explanations but in the case of a causal explanation this is false, since a potential cause may exist yet not be an actual cause because some other cause pre-empted it (LIPTON, 60). Dawes responds by appealing to ceteris paribus clause and that it is unlikely that divine action could be pre-empted. He draws a nested diagram with ‘actual explanations’ in the center, potential explanations next, and the proposed explanations third. He uses a thought experiment to try to show causal pre-emption allows for potential explanations that are not true potential explanations.

PROPOSED EXPLANATION: Covers hypotheses and theories. They lie on a continuum, a theory being a more complex explanatory posit (it may include a number of distinct hypotheses). They can be used interchangeably though some draw a distinction:

Hypothesis: a tentative statement about the world leading to deductions that can be tested.

Theory: Well-substantiated explanation of the natural world that can incorporate facts, laws, inferences, and tested hypotheses.

Dawes is not making this distinction.

ACCEPTING AN EXPLANATION: In what sense would we ‘expect’ to observe E if H were true? What are sufficient grounds to consider that H is an actual explanation of E and thus consider H as true? The second question has been answered in two ways. Collins and Swinburne employ confirmation theory which uses the rule of inference that the considerations that lead us to accept a theory must be such that they show they theory to be at least probably true. What does this mean? It means that we are

justified in accepting a potential explanation only if it is confirmed in the absolute sense, if the evidence in its favor makes it more probable than not, and if it has no probable competitor:

Where E is evidence, H is a potential explanation of E, and K our background knowledge, then we can regard H as a successful explanation if: Pr(H/E & K)>0.5 and there exists no competing hypothesis whose probability given E and K is higher.

The probability of H given E and K (its posterior probability) can be discovered by estimating the prior probability of H and the likelihood of E given H and then applying Bayes theorem. This is how Swinburne argues.

CRITCISM: There are difficulties assigning precise numerical values to probability estimates. There are also difficulties in spelling out the required conception of probability: ‘physical’ ‘statistical’ and ‘inductive.’ Physical probability has to do with the propensities that are actually at work in the universe. In a fully deterministic universe every actual event would have a probability of 1. Statistical probability is a measure of the proportion of the members of a class who possess a certain property. Swinburne doesn’t use these, he uses inductive probability which is a measure of the extent to which one proposition r makes another one q likely to be true. Does it make sense to speak of the probability of an explanatory hypothesis? Swinburne assumes that since both hypotheses and bodies of evidence are expressed in propositions, inductive probability can also be applied to explanations. Swinburne advocates an objective, normative view of inductive probability, rejecting what he calls subjective probability: the force that evidence has for someone with a certain way of assessing that evidence, and a certain ability to do so. Swinburne is interested in probability estimates according to logical probability: that measure of inductive support that would be reached by a logically omniscient being (one who knows what are all the relevant logical possibilities and knows what they entail, and has correct inductive criteria.) One might argue that this is not very useful since we are not omniscient beings and the only probability measure we could possibly make use of is what Wesley Salmon terms ‘ascertainable’ or what Swinburne calls epistemic probability. This measures the degree of support that the evidence lends to the hypothesis, when that hypothesis is assessed by subjects with correct rules of reasoning but limited knowledge (Subjective probability does not assume that subjects are using the correct rules of reasoning). If epistemic probability is the best we can achieve, Swinburne implies that it is logical probability at which we should aim. What are the criteria of logical probability? Swinburne argues that we can assess the logical probability of a hypothesis by reference to its explanatory power, its narrowness of scope, its fit with background knowledge, and its simplicity.

Dawes says that some would take issue with each of these criteria. Popper would deny that we can speak of the probability of a hypothesis; Sobel rejects Swinburne’s idea logical probability in favor of measures of the degree to which actual individuals are confident about their beliefs; McGrew argues that logical probability sets our standards for justified belief impossibly high. Dawes believes that the use of confirmation theory has serious difficulties and instead will propose his own criteria if theory choice in Chapter 7: degree of testability, simplicity, economy, and fit with background knowledge. Moreover, Dawes preferred notion of simplicity differs from Swinburne’s and he will not be using these criteria to estimate probabilities; he will make no use of confirmation theory.

CONCESSION: In discussing the work of those theists who do appeal to probability arguments, I shall simply assume that they have some answer to these objections. After all, these are no objections to religious explanations in particular; they are objections to a particular way of defending explanations in general. But as I hope to show, it is not the only way in which explanations could be defended.

EXPLANATIONISM: This is the view that we are justified in accepting a potential explanation when it displays, to a greater degree than any competitor, certain explanatory virtues, EVEN IF WE CANNOT DEMONSTRATE IT TO BE PROBABLY TRUE. Dawes thinks that confirmation theory sets the bar too high; a bar that is not actually applied within the sciences. Dawes gives the example of Darwinian evolution when its posterior probability was not greater than .5, but when it was ‘the least objectionable theory’ on offer. It was capable of explaining hitherto puzzling phenomena, it posited a mechanism for which there existed a familiar analogy, and it was potentially fruitful. What features justify this kind of epistemic position, why we should value them, and how they relate to the confirmation theorist’s probability calculations is a question for (6.2).

POTENTIAL EXPLANATIONS: If H were true, E would be a matter of course. What does this mean? Two answers: 1-One can argue that H ENTAILS E; the explanation in question taking the form of a deductive argument. 2-H renders E more likely than it would be otherwise; argue that the truth of H makes E more likely than it would be if H were not true. Swinburne takes the second approach arguing that intentional explanations make E more likely than it would be otherwise. Dawes will take the first deductive approach. This is not because he believes that all explanations can be reduced to deductive arguments; maybe they can maybe they can’t, but because theistic explanations are a species of intentional explanations which Dawes will argue (in Appendix 2.1) are best reconstructed as deductive arguments. If one can deduce a description of the explanandum from the hypothesis, then one’s hypothesis corresponds to the second premise of Pierce’s schema. A deductivist approach does not exclude probability calculations. However, it is true that the explanation ITSELF would not be probabilistic but if the prior probability of H is less than 1.0, the posterior probability could still be less than 1. So, just because our explanans would not assign some degree of likelihood to the explanandum this does not mean that we cannot assign some degree of probability to our explanans.

CONCESSION: A theistic explanation could have some degree of explanatory force, and in the absence of a natural explanation we could, conceivably have reason to accept it, but this should always be an explanation of last resort.

CHAPTER 3: What are theistic explanations?

Positing A Divine Agent / The Explanans:

Intentional explanation is itself a particular kind of theoretical explanation since it posits unobservable mental states (intentions, beliefs, desires) in order to explain some observable behavior of the agent. Theism does not merely posit an unobservable intention, but also an unobservable AGENT. In ordinary life we already know that the agent exists and we are trying to understand WHY this agent acted as she did but with theism the existence of the proposed agent is itself contested. So, the theist faces a double burden; she must both posit the existence of an unobservable agent, and show that the

explanandum is best understood as the work of that agent. This means that there must be something about the event to be explained that would be less puzzling if it were described as an intentional action. Merely positing the existence of God is a necessary but not a sufficient condition of a proposed theistic explanation, at least if that explanation is to have any degree of content. But as Swinburne argues, we can at least conceive of explanations that are parallel to that offered by the theist, in which we posit both the existence and the activity of an agent (e.g. hypothetical poltergeist). However, an unembodied agent may simply be a surprising fact about the world akin to the strangeness of quantum mechanics.

Both the kalam and the fine-tuning argument posit both the existence and the activity of an agent without necessarily needing to say WHY this agent needed to act so it seems Dawes is incorrect if he is saying that we need to posit the existence, activity, and INTENTION to move from a PTE to an ATE. Such an explanation would be adequate even if not fully complete but Dawes already defended the notion that we don’t need to have a complete explanation for the explanans. Indeed, SETI is a great example of this. Here, we are searching for evidence that would confirm both the existence and activity of agents which we do not know exist (even if we are unclear on there intentions). Murder cases are also similar to this. Other arguments for the existence of God are metaphysical in nature posit God as an ultimate and/constitutive cause rather than a proximate and immediate cause.

Spurious Unification: Successful explanations unify our knowledge, bringing disparate facts under the same pattern of explanation but this can be spurious when a pattern of explanation is offered that could embrace any state of affairs:

1-God wants it to be the case that a.

2-What God wants to be the case is the case.

3-Therefore, a.

Since a has no restrictions, and the empirical content of a theory is measured by the possible state of affairs it excludes, then 1 has no empirical content. Dawes thinks premise 1 is false since he thinks and OMNI-God would exclude states of gratuitous suffering. But still, the empirical content of ‘God wills a’ may be very low and unhelpful. This is especially so when we recognize that God is the cause of every event that occurs and of every actual state of affairs. If a proposed theistic explanation is to have any significant degree of empirical content, it must answer the question: ‘Why did God do that?’We need to know not just ‘that’ God did it, but ‘why’ God did it; without a particular divine goal otherwise there won’t be any significant degree of empirical content.

It isn’t true that God is the cause of every event that occurs; that is occasionalism or theological determinism. Moreover, while is can be helpful to specify why God did something to generate testale predictions, it isn’t always necessary to know WHY and agent did something before you can know THAT an agent did something:

Both the kalam and the fine-tuning argument posit both the existence and the activity of an agent without necessarily needing to say WHY this agent needed to act so it seems Dawes is incorrect if he is

saying that we need to posit the existence, activity, and INTENTION to move from a PTE to an ATE. Such an explanation would be adequate even if not fully complete but Dawes already defended the notion that we don’t need to have a complete explanation for the explanans. Indeed, SETI is a great example of this. Here, we are searching for evidence that would confirm both the existence and activity of agents which we do not know exist (even if we are unclear on there intentions). Murder cases are also similar to this. Other arguments for the existence of God are metaphysical in nature posit God as an ultimate and/constitutive cause rather than a proximate and immediate cause.

THE COHERENCE OF THEISM: Do intentional explanations make sense when applied to God? Perhaps God has analogous pro attitudes that humans do, but this is made more difficult by the doctrine of simplicity. There is also the difficulty of deciding which if any of our language about God is univocal, or analogical. Does analogical language undermine the force of proposed theistic explanations. No, because we already have successful explanations which use analogical terms such as the beliefs and desires of animals.

The doctrine of simplicity is incoherent.

The Concept of God: Are God’s attributes internally consistent with one another?

Anselmian theism is the way to go! Or even a finite theism.

Direct and Indirect Proofs: Even if theistic explanations seemed apparently contradictory we could still legitimately adopt them as a provision heuristic if they possess other explanatory virtues.

Intentional and Causal Explanations: Causal explanations do not necessarily involve causal laws. But perhaps the problem is that theistic explanations say nothing about the mechanism of divine action and that is why they fail to be causal explanations. Grunbaum writes, “the hypothesis of divine creation does not even envision, let alone specify an appropriate intermediate causal process that would link the presence of the supposed divine causal agency to the effects that are attributed to it. “

RESPONSE: It is the essence of the theistic hypothesis that God can bring about events and states of affairs directly, without any intermediate causal process merely by willing it as a basic action. One might propose that this would be like ‘magic’ and not warranted by the evidence. One might point out that we know of no other causal process like this. (We do know of other causal processes like this: the expansion of space, quantum non-locality, quantum teleportation, potential cosmological models of the universe, etc.)

But this is no objection to the IDEA of divine action. In fact, Grunbaum’s argument would lead to an infinite regress of causal attributions;

if A causes B, we must be able to specify an ‘intermediate causal process’ that links A and B (e.g. A1), but what is the intermediate causal mechanism linking A1 and B, and so on ad infinitum. The way to avoid this is to point out that A1 is not a distinct cause, but nothing more than a specification of how cause A operates which is to affirm a direct causal link between A and B; and there is nothing incoherent about the idea that the divine will could be a direct cause of this type.

God and Causation: It is difficult to find a concept of causation that is applicable to the posited divine agent. Quentin Smith’s De Facto (in fact) argument:

1-No existing sense of the word cause that is applicable to God;

Consider premise (1) If the claim that God caused the Big Bang cannot be analyzed in terms of extant definitions of causality, then God cannot have caused the Big Bang. I see no reason to think that this premise is true. In general, arguments to the effect that some intuitively intelligible notion can't be analyzed in terms of certain philosophical theories should make us suspect the adequacy of those theories rather than reject the common sense notion. The idea that God caused the universe is intuitively intelligible. A cause is, loosely speaking, something which produces something else and in terms of which the thing that is produced can be explained. This notion certainly applies to God's causing the universe. If God's causing the universe cannot be analyzed in terms of current philosophical definitions of causality, then so much the worse for those theories! This only shows that the definitions need to be revised. Indeed, the standard procedure in terms of which proposed definitions of causality are assessed is typically to propose some counterexamples in terms of intuitively plausible cases of causation and then show how the definition fails to accommodate these new cases. In the same way, if God's causing the universe cannot be accommodated by current philosophical definitions of causality, then that plausibly constitutes a counterexample to the definition, which shows that it's inadequate as a general metaphysical analysis of the causal relation, however adequate it might be for scientific purposes. Moreover, there's no reason to believe that we have arrived at the final and correct analysis of causation. In fact, as I'll point out in a minute, there's good reason to believe quite the opposite. The point that I'm making , I think, is especially plausible when you recall that the philosophers who drafted the definitions quoted by Quentin were exclusively concerned with natural causes, even physical causes. They weren’t even considering such recondite examples as divine causation of the origin of the universe. It's hardly surprising, therefore, that their theories should fail to capture this notion. So I see no reason to think premise (1) is true and good reasons for thinking that it is false.

Now what about premise (2), The claim that God caused the Big Bang cannot be analyzed in terms of extant definitions of causality? As it stands, premise (2) is clearly false, and Quentin himself, in effect, admits that it is false. For in his discussion of the analysis of causation in terms of necessary and sufficient conditions, he does not deny that God's willing the universe or the Big Bang is a case that satisfies the definitions proposed. On the contrary, he explicitly states that it does satisfy the proposed definitions. Rather he attacks the adequacy of the definitions themselves. Now that puts an entirely different light on the matter! What Smith is really defending is

(2) The claim that God caused the Big Bang cannot be analyzed in terms of any adequate extant definitions of causality.

But premise (2) is extremely problematic, for it is, I think, generally acknowledged that there is no adequate extant definition of causality to date. The very proliferation of different definitions which were only partially surveyed by Quentin in his talk testifies to the uncertainty and the dissatisfaction which exists in the philosophical community today with proposed analyses of the

causal relation. Thus the expression in (2) "adequate definitions of causality" may well be a nonreferring expression, so that (2) cannot be true.

But let's assume that the definitions surveyed by Smith are adequate. The fact is that God's causing the universe does satisfy at least some of these definitions, so that (2) is false. Take David Lewis' analysis of causation, for example: According to Lewis, c causes e if and only if c and e are both events, and both occur, and if c had not occurred e would not have occurred. Now God's willing the Big Bang clearly satisfies this definition. God's willing and a Big Bang are both events which occur, and if God's willing had not occurred, the Big Bang would not have occurred. No problem! But, Quentin rejoins, if the Big Bang had not occurred, God's willing would not have occurred. So is the Big Bang the cause of God's willing? Well, obviously not; but what this calls into question is the adequacy of Lewis' analysis, not whether divine causation satisfies it. Lewis remedies the problem by stipulating that if e had not occurred, c would still have occurred, a remedy that won't work for divine causation. Actually, Lewis' remedy won't work for many natural causes either, since in some cases the counterfactual, "if e had not occurred, c would not have occurred" is true. So what Lewis' definition provides is not a definition of "c causes e," but rather it is a definition of "c and e are causally related," and it fails to specify the direction of causation. But here the theist, you see, faces no problem because it is metaphysically impossible for God's willing to have an external cause. There is no possible world in which the Big Bang causes God's willing or God's volition. Therefore, given Lewis' definition of "c and e are causally related" and the impossibility of the Big Bang causing God's willing, it follows that God's willing causes the Big Bang, and thus divine causation satisfies Lewis' definition of causality.

Or, again, if we hold with Isaac Newton, Richard Swinburne, Nicholas Wolterstorff, and others that God exists prior to the Big Bang in a metaphysical time, then there's no objection to adopting an analysis of causality which involves the component of temporal priority of cause to effect. God would be temporarily prior to the Big Bang. So it seems to me that the premise (2) fares no better than premise (1). Both premise (1) and premise (2) are false, and therefore the argument against divine causation of the universe is unsound.

Read more: http://www.reasonablefaith.org/does-god-exist-the-craig-smith-debate-1996#ixzz2VvaQmovM

2-3 conditions of causality—temporal priority, spatial continguity, and law-like conjunction.

Causes can be simultaneous with their effects, God could exist in metaphysical time, quantum non-locality is simultaneous and there is no continguity; law-like conjunction presupposes intentional explanations are metaphysically impossible and that there cannot be singular causal explanations. In biology, there are no laws of nature and hence no law like conjunction. The singularity violates all of these conditions as well and surely it is absurd to say that it didn’t cause the rest of the universe.

Smith’s In Principle argument: If an event or state, x, is a logically sufficient condition of another event or state, y, then x cannot also be the cause of y.

For example, ‘the body is in motion’ is a logically sufficient condition of the truth of the proposition ‘the body occupies space.’ But the former doesn’t cause the latter. Thus, there is a distinction between logically sufficient conditions and causes. But then, since God’s willing y is a logically sufficient condition of y, it follows that God cannot be the cause of y, and since all explanations are causal it seems that God cannot be the cause of the universe.

Smith’s principle is just false. For example, increasing the Pressure and holding the volume constant for a gas in a container is logically sufficient for increasing the Temperature while also being the cause of it which doesn’t show that a theistic explanation cannot be causal.

Are all explanations causal? It would seem that the theist wants them to be.

A Case of Fallacious Reasoning: If Smith’s in principle argument was valid, then every truth would be a necessary truth which is fallacious.

A Category Mistake: Smith confuses logical and metaphysical issues. There is a sense in which God’s willing E ;necessitates the occurrence of E, but this necessity is not logical, it is metaphysical.

THE UBIQUITY OF DIVINE CAUSATION: If God exists, then any complete explanation of any fact must at some point make reference to divine action. We do not need to have a complete explanation in order to have an explanation. There are two forms of secondary causation: a miracle and natural causes. How might the theist reconcile these two forms of explanation?

Chapter 4: What Would They Explain? / The Explananda or Explanandum

3 features of proposed theistic explanations: 1-they cover both events and states of affaris; 2-they are generally singular explanations; 3-there is a certain relationship between theistic and natural explanations…on some occasions TE cover facts that admit of a natural explanation. On other occasions divine action invokes an agent capable of miracles.

EVENTS AND STATES OF AFFAIRS: Facts are events, and ‘states’ could be subsumed under a common definition of events. They are both exemplification by substances of properties at a time. This is not without its relevance to causation/

STATES OF AFFAIRS: Theism doesn’t always explain an event in space and time that is bounded but sometimes it is a state of affairs that exists over time and may be distributed in space; it may even be co-extensive with time (as in the case of the existence of the universe, the laws of nature, why the universe exists).

EVENTS: These explanations are narrow in scope (floods, earthquakes, resurrections)

In general, theistic philosophers favor theistic explanations of states of affairs over explanations of particular events. The broader the better (e.g. the laws nature); what would it even mean to offer a natural explanation of the laws of nature?

SINGULAR AND GENERAL EXPLANATIONS: Theistic explanations are typically of the form ‘C caused E’ rather than ‘C-events cause E-events.’ This a description of theistic explanations not an explanation itself. But does it make sense to speak of a singular causal explanation? The first question is metaphysical regarding causation, and the other is epistemological regarding explanation. Something is a cause only if it is regularly associated with an effect, the relata here being event-types (necessary condition). Causation is a matter of law-like association. Events of type A are regularly followed by events of type B. Any causal explanation must reflect this law-like association. However, Donald Davidson argues that intentional explanations are law-less (or anomalous) while still holding to a HUmean view; Davidson reconciles these two kinds of explanations by distinguishing between two levels of explanation: Intentional explanations do not cite laws, but facts to which they refer could also be explained using the language of the natural sciences, and such language would cite causal laws. Singular causation is not merely a claim about the structure of our explanations it involves a metaphysical claim about causation; Dawes thinks that unless we hold that causation is nothing other than a regular association of events, then there is no reason to deny the possibility of singular causation; an event could be unique in the history of the universe and yet known to be a case of causation.’ Even if this is wrong, this still wouldn’t eliminate ALL proposed theistic explanations such as prayer patterns that do purport to related event-types or law-like regularities.

TESTING A PROPOSED EXPLANATION: Singular explanations cannot be tested by replicating condition C and seeing if E is observed, but this doesn’t mean it is untestable because we can test it in other ways (Appendix 3.3.2).

THEISTIC AND SECULAR EXPLANATIONS: Occasionalism and Causation-not created secondary causes; there are no natural causal explanations though there could be regularities in occasions.

Facts lacking a natural explanation: In principle and in fact.

Naturally Unexplained Facts: We may not have a potential natural explanation but we might some day such as a person who mysteriously recovers from an illness.

Naturally Inexplicable facts: We could never have a natural explanation of THIS. All the atheist could say is that it is just a brute fact such as the existence of the universe or the laws of nature.

Facts Having Natural Explanations: Theists sometimes offer complementary explanations of natural facts and other times competing explanations of natural facts.

Fact With Accepted Natural Explanations: Walls of Jericho.

Are Theistic Explanations Redundant? If there is already a natural explanation then why offer a theistic one at all. This is not a case of causal overdetermination where two factors are sufficient to bring about an effect but neither is necessary since even if one were not present the other would ensure the same

outcome. No, the natural cause is thought to be dependent on the divine cause; God’s willing of the event is a necessary condition of its occurrence; for an event that was a secondary cause. But God’s willing of the event is also a sufficient condition of its occurrence since the divine will is unfailingly efficacious. Natural causes are thought to be the means by which God achieves his goals. Dawes argues that any potential theistic explanation must tell us why God would choose any created cause as a means of achieving his goals instead of a direct basic action. But perhaps the charge of redundancy is not very serious. For in the case of intentional explanations there does seem to be a sense in which one can ofer two explanations for the same fact; an intentional explanation tells us what the agent was trying to achieve in performing this action while natural-scientific explanation describes the mechanism that brought about the action. Neither explanation would be complete in the sense that each leaves some explanation-seeking questions unanswered but as incomplete explanation is still an explanation ‘the motion of my hand may be explained by neuroscience but it may also be explained by me bringing it about through the intention and power to do so.

MULTIPLE PROBLEMS HERE: Only if you think physical reality has to be sustained in existence by God’s power does it make sense to suggest that God is a necessary condition of its occurrence. Even here however, one would say that God either positively wills what happens, or He merely permits it while having morally sufficient reasons to do so even though he doesn’t positively will it (e.g. some evil in the world). It is not true that God’s willing something is sufficient for it occurrence because God must also act on that willing; God could have higher-order willings, not everything God wills may be metaphysically possible or logically possible to achieve, etc. Moreover, it makes sense to seek ultimate/proximate causes in science (biology) and so God can use natural means as proximate causes but still be the ultimate cause, and indeed I have already given reasons why God would use natural means and why he would use miraculous means.

Explaining under a description: But to show that these two levels of explanation are complementary does not yet escape the charge of explanatory redundancy; for if someone accept the natural explanation why offer an intentional explanation as well? Any fact is explained under a particular description; The physiological explanation explains why my hand moved while the intentional explanation explains why I moved my hand. To say that a fact already has a successful natural explanation is to say that there exists some description of that fact under which it can be explained in non-intentional terms. So if the theist wishes to argue that the same event requires a complementary, intentional explanation, she must show that there exists another true description of the event that requires this new kind of explanation.

Again, we just need to make the ultimate/proximate distinction and give an argument for some fact likely has its ultimate explanation in something supernatural even though the proximate fact has a natural explanation.

The surprising implication of this is that a theist doesn’t have to begin with the naturally inexplicable, or naturally unexplained, or natural facts they reject but with accepted natural facts and show that there exists another true description of the same fact that demands an intentional theistic explanation. But

since both the existence of the agent and its intentions are doubted by the atheist it is understandable that the theist beings with facts that can’t be explained naturally.

Facts With Contested Natural Explanations: Fine-tuning

The Problem of Miracles: Miracles are no explanation at all since that actions of God are incalculable or arbitrary. If we could not say what would follow if a theistic hypothesis were true; such a hypothesis would not count as even a potential explanation, but is this the case for a miracle working agent? A related objection is that without the constraint of lawful regularity, inductive inference cannot get off the ground; so if the theist is to have any trust in our knowledge of the world, he is bound to assume that his God will not work too many miracles! And of course there is something odd about the idea of a God who establishes the laws of nature only to violate them. But this objection suggest that even positing an agent who can work miracles is fatal to the task of explanation.

Rational Agent: The above objection assumes that intentional explanation is nomological; that we must cite laws connecting the agent’s observed behavior with some particular set of beliefs and desires but whether or not such laws exist an explanation of this form does not rely on them. It assumes a rationality principle which says that we can predict an agents behavior by assuming that she will act in a way that is consistent with her beliefs and desires in order to achieve her goal. In other words, the objection above assumes a false dichotomy; either the world is governed by laws or it is the product of whimsical supernatural agent. But the world could the work of a rational supernatural agent so that if we posit a particular goal we might be able to predict its behavior. The question is: What would be the most rational way for such an agent to act in order to achieve this goal? The answer may or may not include miracles.

HUMES’ ARGUMENT: On what basis do scientists search for previously unobserved events? Presumably because they have a promising theory which predicts that this event will occur. This suggests that the prior probability of a miracle is dependent on the prior probability of the theistic hypothesis.

CHAPTER 5: Potential Theistic Explanations

We must ask if this particular story of divine action constitutes a potential explanation of this particular state of affairs.

1-Theistic explanations must posit not the mere existence of a creator, but explain WHY God willed what is being explained.

2-The coherence of theism-Anselmian theism

3-The theist needs to show that there is some sense of the word cause that is applicable to God.

Not necessarily, consider that there is no accepted account of knowledge yet surely we know things. Or consider that there are no necessary and sufficient conditions for what counts as science but surely we can still practice it. Same for God. But, all we have to do is say that God is an personal efficient cause without material cause and we have examples of efficient causation without material causation.

4-The optimality condition

5-

THEOLOGICAL SKEPTICISM: Can we predict what God would do? Perhaps God is so different from anything we know that we cannot make predictions about what He would do. Or perhaps we know what God is like but we don’t know what options are available to him. If either of these are correct, then we cannot IN PRINCIPLE make judgments about what would count as even a potential theistic explanation.

We can reason from our available background knowledge (e.g. the known laws of nature) plus the Optimality Condition rather than a pure a prior Optimality Condition. This way, we control our ignorance without letting it land us in complete skepticism.

DESIGN ARGUMENTS: Sober argues that ID doesn’t win by default or just because there is no good natural explanation but an ID proponent also needs to show that the likelihood we should observe the same facts is higher on design and to do this we need to be able to predict what would follow if the hypothesis were true. If one cannot say what would follow if a hypothesis were true, then that hypothesis is untestable, but worse still, we cannot even know that it is a potential explanation. Sober argues that we cannot know the goals and abilities of the designer in a predictive manner so we cannot predict anything about ID to test.

http://www.lydiamcgrew.com/PhiloTestability.pdf

Sober does consider that we could discover the designers intention by inspecting the products of his design, what we see around us but this begs the question since we have to assume that what we see around us is designed in the first place.

I don’t think this begs the question; Antikythera Mechanism, or ruins on some extraterrestrial planet, or some piece of technology that isn’t human but comes into Earth’s orbit. In any case, with fine-tuning, we can make a testable prediction using theism and an auxillary hypothesis if need.

IS SOBER CORRECT? The criterion of independent specification: if we are to judge a theistic explanation to be a potential explanation, we must be able to specify what would follow if it were true. Since God can do anything logically possible, this makes the problem even worse since if an engineer had limited options he would be more predictable. But do theistic explanation place no constraints no how God might be expected to work?

An Explanatory Constraint: We are perfectly at liberty to make assumptions and then see what follows from them; this is precisely what it means to offer a theoretical explanation supported by abductive reasoning. So we can posit the existence of God having the requisite goals and abilities and then ask what we would expect to observe as a result. If the explanandum is among the things we would expect

to observe given the hypothesis, then we would have satisfied the second premise of Peirce’s abductive schema, and with the principle of rationality it seems we can make such a prediction.

The Optimality Condition: God doesn’t have any false beliefs, or unrecognized desires, or any constraints that aren’t logical, no weakness of the will, cannot be under any compulsion to adopt a means to an end, he could bring about what he desires merely by willing it, which all goes to mean that God would choose the best possible means of achieving what he wills. The explanandum must meet this requirement. So, an atheistic could discredit a proposed theistic explanation by arguing that there exists a better way in which God could have achieved his goal. Suboptimality arguments often miss this point when and instead try to argue:

1-If God is responsible for this apparent design, we would expect it to be perfect.

2-It is not perfect.

3-Therefore, God is not responsible.

This implies that a theistic explanation merely attributes some fact to divine action, but Dawes argues that an explanation of this kind would lack empirical content, and makes suboptimality arguments of this kind too easy to defeat. For efficiency must be measured against some goal, and the notion of a ‘perfect x’ is hard to specify: ‘given the conventional concept of the creator, there seem to be no limits on what is possible nor any reason why one hypothetically possible panda should be preferred, as a counterfactual ideal, to another. If ‘perfection is limited only by the extent of one’s imagination, then specifying an ideal phenotype, for the panda or any other organism, quickly becomes a fanciful exercise. Why wouldn’t the creator have given pandas the ability to fly?’

The optimality condition doesn’t specify some kind of overall optimality (a perfect panda thumb) but only optimality in relation to a specified divine purpose. It is up to the theist to nominate just what that purpose is. Once we do this we can see that Gould’s argument works since there would be a more efficient design plan for stripping bamboo on theism.

Objections To Optimality Condition:

1-God is not obligated to act optimally: God does not have limited time and limited resources.

2-There is no optimal divine action: Just as there is no best possible world there is no optimal realization of a divine intention. The optimality condition entails that if this world is created by God, then it is the best possible realization of whatever intentions God had in creating.

No best possible world: 1-There may be no ‘single scale of value’ against which different possible worlds could be ranked. There many different kinds of value and how are they to be compared? 2-Such values may have no upper limit.

RESPONSE: 1-What this shows is that there can be no greatest possible God either. To not act optimally would be to act irrationally. Moreover, this is not a constraint on God, it is a constraint on those offering

theistic explanations. 2-The optimality condition is the ‘best relative to the ends and purposes one has in mind.’ While this may entail that a world created by God is the best possible world, they are not identical ideas. 3-Let’s grant the argument that there is no optimal realization, what matters is that we can conceive of a state of affairs which, from the point of view of a rational creator, would be better than that which actually obtains. What we can say then is that if worlds cannot be compared in terms of overall value, then not only is there no best possible world, there is no BETTER possible world. But Dawes doesn’t have to face this objection because even if there is no best possible realization of a divine plan, there would still be better or worse realizations of a divine plan (assuming we can compare different possible realizations of a divine plan).

3-We cannot make such judgments: How do we know which options are available to God?

Response: We are required to make modal judgments if we are to OFFER theistic explanations. If we can’t know this in any corrigible manner, then we are in no position to know what kind of a world an OMNI-GOD would created. We would no longer be able to give content to omnipotence, or logically possible. To test an explanation is to ask what else would follow, if it were true.

4-ID does not entail that the design is optimal: Evidential arguments are comparative. This exposes a huge weakness in Dawe’s sense of orediction.

Concluding remarks: We must be able to specify how we would expect a divine agent to act in order to achieve his goals. We are warranted in regarding a theistic hypothesis as a potential explanation of some state of affairs ONLY IF we cannot conceive of any better way in wchi the postied divine goal could have been attained. If we can’t make these judgments, then there cannot be any theistic explanations. Even a moderate degree of skepticism about modal judgments will undermine our confidence in theistic explanations.

CHAPTER 6: Inference to the best explanation

What would constitute a sufficient reason for accepting a theistic hypothesis, given that it has been show to be a potential explanation of the fact in question?

Induction & Abduction: Dawes will use the STRUCTURE of Swinburne’s arguments as a foil.

Inductive Reasoning: Swinburne claims to offer C-inductive arguments which he describes as arguments that make the theistic hypothesis more probable than it would be otherwise.

The Likelihood of the Evidence: Swinburne argues that if H renders E more probable than it would be otherwise, then H is confirmed by E. This is a weak sense of ‘confirmation’ that is incremental rather than absolute—for E may lend some support to H even when H is clearly false. For example when you hear noises in your attic, this confers a high likelihood on the hypothesis that there are gremlins up there bowling, but few of us would conclude that this hypothesis probably true because of a low prior probability. The second step in Swinburne’s reasoning is to turn various C-inductive arguments into a P-inductive argument. A P-inductive argument is one that makes the conclusion more probable than not. Swinburne argues that the prior (or intrinsic) probability of theism is high, at least relative to other

hypotheses. For prior probability will be directly proportional to the simplicity of a hypothesis, its narrowness of scope, and its conformity to background knowledge. Swinburne argues that background knowledge need not be taken into account when assessing the theistic hypothesis since theism purports to explain everything logically contingent. Swinburne concludes that theism is as probable as not until we include religious experience and history which makes it more probable than not.

Abductive Reasoning: Dawes thinks Swinburne is actually engaged in abductive reasoning. Although it is true that abductive reasoning can be supported by inductive arguments.

6.2 The Best Explanation: IBE is a form of abductive reasoning that is particularly well-suited for theism. First, it can be used to support the positing of unobservable (or unobserved entities) even when the cause we posit is not something we have observed before. Second, IBE can be used to support singular causal claims. Even if we only know of one event of the type in question (such as ‘the big bang’) we can still posit a cause and argue for its existence in the manner Pierce indicates. All we need is a non-inductive way of arguing that ‘if H were true, E is what we would expect,’ one that does not need to appeal to past instances of this correlation especially in the case of intentional explanations.

The Problem of Abduction: The mere fact that some hypothesis would explain the data, if it were true, does not entail that it is true. Moreover, even if an explanation is the best amongst competitors, it may still be false. So, here is defensible argument about the ‘reach’ of IBE:

1-The surprising fact, E, is observed.

2-H would be a satisfactory explanation of E.

3-No available competing hypothesis would explain E as well as H does.

4-It is reasonable to accept the best available potential explanation of any fact, provided that explanation is a satisfactory one.

5-Therefore, it is reasonable to accept H.

It is rational to accept a falsehood provided that one does not know it is false (e.g. Ptolemy before Copernicus).

The Best Explanation 6.2.2: What criteria do we judge H to be the best explanation of E; even if it doesn’t have any rivals? What if a proposed theistic explanation were the only possible explanation of some fact? Would a proposed theistic explanation win by default even if it did not appear to be a very good explanation? Some event may turn out to be a brute fact, one lacking an explanation. Moreover, we rarely know that a proposed theistic explanation is the only alternative. Should we go with a ‘justificationist’ answer like Swinburne in his appeal to confirmation theory where we are justified in accepting a potential explanation H if: Pr (H/E&K)>0.5 and there exists no competing hypothesis

whose probability given E and K is higher. We can then use Bayes’s theorem to take us from the likelihood of E given H to the probability of H given E and K. Dawes thinks this sets the bar too high since people accept scientific theories even though the posterior probability has not yet been shown to be greater than .5.

What we should go with is an explanationist method where we accept a theory so long as it displays certain explanatory virtues to a greater degree than any competitor. It is true that is may be impossible to quantify such judgements (hence the attraction of Bayesian formulas) and there may exist no algorithm, no decision procedure that will ensure that our judgements are correct; ‘two scientists may make their choices in different ways; and yet in doing so both of them might be acting rationally.

Explanatory Virtues:

1- What features of theories count as explanatory virtues?The problem here is that EV’s are difficult to define, there exists no exhaustive list, and it is far from clear just how they are supposed to function. William Lycan lists three functions EV’s may be thought to play: A-they may be thought of as purely descriptive (how scientists actually choose theories) B-They may be thought of as normative; as accounts of how we would choose theories if we were acting rationally, or C-They may be regarded as practical suggestions for a doxastic decision procedure. Dawes will take them as primarily normative.

2- Why do we regard these features as virtues, or desiderata? Dawes says he doesn’t have a satisfactory answer to this question, but that nobody else does either.

3-What would give us sufficient reason to accept a potential explanation as the actual one?Dawes says an appeal to explanatory virtues is fundamental and cannot be justified by an appeal to anything deeper since these are the principles we use to justify and which cannot be justified in turn, ‘that which is used to prove everything else…cannot itself be proved.’ When we look at other instances of theory-choice that we judge to be rational we discover that these are the kind of principle they employ.We can however offer an epistemic justification of these principles in a broadly evolutionary sense; ‘these methods of theory choice are the ones that a wise and benevolent Mother Nature would have given us, …because our having these methods rather than others has survival and welfare advantage.’

CHAPTER 7: Successful theistic explanations

What must a theistic explanation accomplish?

1-It must justify the double burden of positing a divine agent and a divine intention.

2-It must show that the fact in question was the best way in which that intention could have been realized ‘optimality condition.’

3-Once 1 & 2 are accomplished, then we should assess the proposed explanation against a list of explanatory virtues. These virtues should be seen as desiderata (something desired) rather than (jointly) necessary conditions: we need not demand that an explanation displays all of these virtues.

And any theory choice will have to balance these various desiderata against each other, for they can come into conflict. ‘Our preference for any one of the ‘explanatory’ virtues always comes qualified by ‘other things being equal,’ and the ‘other things’ are the respective degrees of the other virtues. Unfortunately, there doesn’t seem to be any algorithm that would guarantee a correct balance in any particular case (one can see the appeal of a Bayesian approach). But the lack of a clear decision procedure does not count against the insight that this is how we actually choose theories. Nor does it show that we are acting irrationally in doing so. Since there exists no definite or exhaustive list of explanatory virtues we are forced to choose which we shall adopt. Here is my choice:

1-High degree of testability; it will have survived independent tests; it will posit mechanisms that are at least analogous to those which we are already familiar, and will form part of a previously successful research tradition. Many regard this as a necessary condition of calling something a scientific explanation. All branches of the science derive implications that can be checked intersubjectively and perform experimental and observational tests on those implications. Falsifiability is testability. Once a theory survives independent tests, it can be said to have been corroborated.

A hypothesis is independently testable if we can use it to make predictions about facts OTHER THAN those it purports to explain, that is to say, about different kinds of phenomena. For example, if ‘cathose rays’ are really negatively charged particles, then when they enter an enclosure they ought to carry into it a charge of negative electricity. And if we are dealing with competing potential explanations, we should be able to make predictions that would turn out to be true given our favored hypothesis, but not true given the alternative. In other words, there needs to be some chance that the prediction it makes will turn out to be false; it needs empirical content; there needs to be at least one state of affairs that it excludes.

1970’s: Lakatosian Idea: Only ‘novel’ predictions make a program progressive. But what counts as ‘novel?’ Temporal novelty (in which prediction must be made before evidence is gathered) has been argued against by appealing to intuition and the history of science.

In its place has come heuristic or use-novelty which says that evidence counts as a use-novel prediction of a theory if the scientist did not use knowledge of the evidence in constructing the theory.

1980’s: Attention shifted from Lakatosian accounts of theory change and assessment of research programmes to formal theories of confirmation (Bayesian) until the mid 1990’s.

The central question in all these cases is: whether and to what extent successful novel prediction conferred stronger epistemic weight?

Theism is independently testable because it can be used to make prediction about different kinds of phenomena: scientific, philosophical, mathematical, and historical.

Are proposed theistic explanations testable?: It is not sufficient that a PTE attributes some event or state of affairs to a divine agent because if God exists then any actual event or state of affairs is attributable to him because whatever God wills must happen.

This would only apply to Facts With Accepted Natural Explanations, not naturally inexplicable facts, or facts lacking a natural explanation (as with the kalam, the fine-tuning, the resurrection, etc.)

So, a PTE must say WHY God is supposed to have willed a fact. Then we can use the rationality principle. What a theist should not do is treat theism as a single hypothesis like: There exists a God, who created the world’ since this has little empirical content insofar as it fails to single out the particular facts to be explained;

No, the hypothesis that God created the universe has significant empirical content especially since on metaphysical naturalism the universe will not have an explanation or supernatural cause the produces it and this can be empirically tested by cosmology.

what theist should seek to corroborate is ‘there is a God who wills G’ where G is a posited divine goal.

Specifying a divine goal certainly adds empirical content, but it isn’t essential or necessary to the theistic hypothesis in all cases, only in those facts that have an accepted natural explanation already.

A theist must do more than show that explanations appealing to a divine agent can cover a wide range of events she must show that this particular proposed explanation can cover a wide range of events. They must have intentional specificity, or tell us WHY God would do what He does:

1-There exists a divine agent with goal G.

2-E is the best means of achieving G.

3-A divine agent will always choose the best means of achieving G.

4-Therefore, the divine agent will do E.

The theist needs to specify a particular goal, and since there aren’t any laws of divine behavior, we should employ the rationality principle.

This can be done:In contrast to the naturalistic explanation of ultimate reality, theism would lead us to expect an all-good God to create a reality with at least a positive, if not optimal, balance of good over evil. Thus, arguably, theism should lead us to expect a universe structured toward the realization or moral and aesthetic value. Given that we glimpse some special value that highly vulnerable embodied conscious agents like us can realize that could plausibly thought to outweigh the evils—such as suffering—resulting from such embodiment, it follows that it is not surprising that God would create an ECA-

structured universe. I propose that one such value is the ability to engage in particular kinds of virtuous actions—such as self-sacrificial love, courage, and the like—that the vulnerability that comes from embodiment allows…these actions allow for eternal connections of appreciation, contribution, and intimacy between conscious agents. For example, if someone significantly helps me in times of suffering, it can create a connection of appreciation in me that has the potential of lasting all eternity, and hence growing in value.

PREDICTION & RETRODICTION: Prediction is ambiguous, it could refer to the deduction of hitherto unobserved facts or it could be used in a looser sense of a retrodiction of known facts. The former severe Popperian kind of prediction is too demanding, given actual scientific practice (e.g. the perihelion of Mercury prior to GTR). What seems to matter when we are testing a theory is that its consequences are true which is compatible with the looser kind of prediction. But when can a known fact support a retrodiction?

1-Everyone agrees that accommodated evidence does provide some epistemic support so that novel prediction is not the only way to gain epistemic support as it was for the Lakatosians, but predicitivts argue that novel prediction provides greater epistemic support than accommodation. The view that only novel predictions provide epistemic support is no longer on the table.

2-Novel prediction is now taken to mean heuristic or use-novelty but the question remains whether scientists considered the already know evidence when constructing the theory: If they did, then they performed an accommodation, if not, then their theory predicted the evidence (Even though the evidence was already available).

3-Many contemporary authors maintain that prediction is better than accommodation but that it is not irreducible, or that it is reducible to another epistemic virtue, namely, explanatory power (Harker). Thus, some have distinguished between strong predicitivism (the view that there is some intrinsic value to prediction) and weak predictivism (the view that prediction is a surrogate for something else).

4-Pluralist Instrumental Predicitivism: The view that is orthogonal to the strong/weak opposition which supposes that prediction is not an intrinsic epistemic virtue, but rather, it serves as an instrumental indicator yet that it can be a proxy for disparate sources of epistemic assurance making it the case that it does not reduce to one particular virtue (as in weak predictivism) nor does it have some kind of intrinsic epistemic virtue (as in strong predictivism).

THESIS: Novel prediction provides some instrumental assurance in a range of cases. We will first structure the levels of scientific inference for which novel prediction can be relevant. We consider inferences from data to phenomena, inferences from phenomena to theory , and inferences from theory to framework. We distinguish these levels and then proceed to examine reasons for why novel prediction might be useful at each level.

A heuristic account: The facts that are considered to corroborate a theory should not be those that were used to construct it. We must also add that a retrodiction must be able to predict some fact that cannot be explained by another competing theory, or that falsifies a competing theory. What happens when there is no predecessor? Alan Musgrave argues that a solitary hypothesis can be corroborated by reference to known facts just as the ‘first testable theory in any field will be confirmed by all the phenomena which it explains.’ So long as there is a consilience of explanatory successes (even in the absence of strong predictions). This is why explanatory power is so important.

The Solitary Potential Explanation: Sober rejects the notion that a solitary hypothesis can be tested. Sober thinks this because there isn’t any probabilitsitc equivalent of modus tollens.

RESPONSE: Alan Musgrave argues, ‘the first testable theory in any field will be confirmed by all phenomena which it explains so that if a theistic explanation doesn’t have a competitor it could still be corroborated by a conislience of explanatory successes, even in the absence of successful predictions and competitors. We don’t have to use probabilistic arguments like the law of likelihood. We could use deductive arguments (of which intentional explanations are a form). Moreover, even if a probabilistic theistic account IS the only proposed explanation on offer, it may be possible to test one or more of its elements comparatively. One can do so by choosing some fact about the world and inventing a rival naturalistic hypothesis, to see which of these hypotheses would render the fact more likely. Paul Draper does this with the problem of evil: (HI) Neither nature nor the condition of sentient being on earth is the result of benevolent of malevolent actions performed by non-human persons. One could also use the null hypothesis where we predict that there would be no statistically significant difference in outcome between a test group and a control group. Chance would be the naturalistic alternative. Sober is right that if a proposed explanation is not corroborated by independent tests, it does not necessarily follow that it is false but it does follow that we have less reason to regard it as a good explanation, one we are entitled to accept.

2-It will also be simple:

SIMPLICITY: What makes a theory simple? Swinburne adopts a metaphysical conception of simplicity: simplicity is a matter of its postulating few (logically independent) entities, few properties of entities, few kinds of entities, few kinds of properties, properties more readily observable, few separate laws with few terms relating few variables, the simplest formulation of each law being mathematically simple.

AUXILLARY HYPOTHESES: Dawes prefers a different kind of simplicity; one where simplicity is equivalent to its empirical content or degree of falsifiability which means that simplicity has to do with the number of auxillary hypotheses that theory requires in order to explain the fact in questions. The fewer the auxillary hypotheses, the simpler the theory. Thus, we shouldn’t prefer theory 1over theory 2 simply because it postulates the existence of fewer entities since we may prefer a theory that postulates more entities than its competitors if it has greater explanatory power (or what Thagard calls consilience). Thus, a theory that is preserved from refutation only with the aid of auxillary hypotheses loses empirical content making it less simple and more complex.

How Simple is Theism (Van Inwagen’s Defence): Atomist, long spikes holding particles of air up, and no evidence of the spikes equals lack of simplicity; or ad hoc and should be rejected unless it exhibits other important virtues to a high degree.

The hypothesis…must include fewer new suppositions which are not already implied to some extent by existing beliefs

3-Ontologically Economical: Similar to simplicity above but it is an ‘ontological type-economy’ which suggest we should not posit NEW KINDS of entities without SUFFICIENT REASON. Just as Dawes argued that we would have more reason to accept a hypothesis positing unobservable entities if we could point to familiar mechanism that bore some analogy to it, this virtue suggests that we should not posit a hitherto unknown type of cause without sufficient reason. In other words, we should avoid positing new kinds of entities when familiar ones will suffice. If what we can observe CAN be explained without reference to God, then it OUGHT to be so explained.

Ontological economy is a ceteris paribus condition that suggests we shouldn’t posit new kinds of entities unless we are required and it is at least conceivable that positing a divine agent might be required to explain some phenomenon. RIGHT, we offer arguments for the existence of God.

4-Informative: ‘Loveliness’ of a proposed explanation. A ‘lovely’ explanation specifies some articulated causal mechanism…whose description allow us to deduce the precise details of the effect. Two aspect of this definition:

1-Description of the mechanism; Dawes has already argued that a PTE will often lack an INTERMEDIATE causal mechanism (one that would mediate the divine will and its effect) and this is not fatal.

2-Ability to deduce what would follow if the proposed explanation were true; a hypothesis should have empirical content in that it will exclude certain states of affairs but it will specify in some detail just what it does and does not predict.

OBJECTION:

But perhaps we can object that theism doesn’t make QUANTIFIABLE PREDICTIONS written in the language of mathematics and that many of our everyday intentional explanation also fail to make quantitative predictions and instead use vague terms to explain and predict peoples behavior even when we lack any clear idea of what mental states they denote or the mechanisms by which they are expressed amking our everyday intentional explanations less than lovely (e.g. belief, desire, hope, fear). They can still be useful, still have explanatory force, and still enable us to make rough and ready predictions, and unless you are prepared to reject all forms of intentional explanation, the fact that a proposed theistic explantion precision does not seem fatal.

RESPONSE: The problem with the above response is that it fails to take into account the mysteriousness of the posited divine agent. We have very little idea just what would be denoted by the predicates that we use of God in terms of two types of predicates: 1-mental (discussed in 3.3.3) and 2-action; What would it mean to say that God creates or speaks to us? If appeal to analogical language then all that

really amounts to is saying we are unable to grasp just what these terms would mean when applied to God or we could do what Alston does and use such predicates in a literal sense. But this won’t work either because there is a disanaology still between how embodied creatures create and how an unembodied thing could create thus leaving us with the difficulty of not being able to specify what we might observe, if he did it.

Classical theism, Davies explains, is the view of the majority of the medieval philosophers like St. Thomas Aquinas, St. Anselm, Augustine, and Avicenna to name a few. God’s nature—according to the classical theists—is immutable (unchanging), impassible (cannot be affected by creation), and can only be spoken of in negative terms. God is not a person like you and I are, that is, like human beings, but God is simple and entirely above any of these classifications (10-11). However, theistic personalism, says Davies, views God through the lens of Descartes. Descartes viewed person as essentially immaterial beings. Each of us are essentially incorporeal beings according to Descartes. God then, being an immaterial being, is described as a person in this sense. In highlighting differences between the two conceptions, theistic personalists pit up God to be a sort of person like us. With respect to creation, God is simply a spectator “who is able to step in and modify how things are” (11). Davies goes into much more detail about the differences, however his point, as he states, is simply to show that there are differences within the theistic camp on how they conceive of God.

An OBJECTION TO THE ABOVE RESPONSE: Perhaps a theist will say that we should expect God’s creative activity to be basic action and so we can specify how God would act in order to create so that whatever God wills will come into existence ‘by magic.’ God need not employ means which means we should expect a miracle. However, this reponse still is not specify the divinely-willed goals, or the purpose God has for acting in some basic manner and so it will lack empirical content; and so, even if basic, we still need to specify a goal in order to explain an action; otherwise we won’t know what to expect from a divine basic action and without that, we cannot explain. For example, we must be able to say that an event such as a tsunami is how we would expect God to punish sinners. As Matthew Ratcliffe has recently argues, much of our everyday social interaction is guided by our shared understanding of the social norms governing our behavior which allows us to understand behavior without without attributing certain beliefs and desires to the person in question in order to understand his behavior (e.g. meeting the eyes of a coffee barista, I need not assume that ‘he believes that I want attention’ because in the context he takes me as a customer and I take him as a barista; what we want and what we do flows from this mutual understanding). But this make theistic explanations even more unlikely because we lack ANY-THING analogous to the shared social norms that can help us to understand human behavior (what about the problem of evil, divine hiddenness, etc.). An appeal to past actions in Scripture begs the question that the believer already has a successful theistic explanation. Even if told that God was loving we wouldn’t know how God would act. Moreover, an omnipotent being would presumably have an infinite range of choices and only a few which we could begin to comprehend and

since we do not know the full range of divine options, we cannot know with any degree of confidence how God would or would not act, in order to achieve his goals and even the rationality principle cannot give us confident judgments about how God would act.

THE DANGER OF ACCOMODATION: So, if we cannot deduce the precise details of the effect it creates the danger of seeing a theistic where one doesn’t exist.

CAUSAL MECHANISM: It is the essence of the theistic hypothesis that God can bring about events and states of affairs directly, without any intermediate causal process merely by willing it as a basic action. One might propose that this would be like ‘magic’ and not warranted by the evidence. One might point out that we know of no other causal process like this. (We do know of other causal processes like this: the expansion of space, quantum non-locality, quantum teleportation, potential cosmological models of the universe, etc.)

But this is no objection to the IDEA of divine action. In fact, Grunbaum’s argument would lead to an infinite regress of causal attributions;

if A causes B, we must be able to specify an ‘intermediate causal process’ that links A and B (e.g. A1), but what is the intermediate causal mechanism linking A1 and B, and so on ad infinitum. The way to avoid this is to point out that A1 is not a distinct cause, but nothing more than a specification of how cause A operates which is to affirm a direct causal link between A and B; and there is nothing incoherent about the idea that the divine will could be a direct cause of this type.

EMPIRICAL CONTENT: This would only apply to facts with an accepted natural explanation.

Beginning of the universe

This can be done:In contrast to the naturalistic explanation of ultimate reality, theism would lead us to expect an all-good God to create a reality with at least a positive, if not optimal, balance of good over evil. Thus, arguably, theism should lead us to expect a universe structured toward the realization or moral and aesthetic value. Given that we glimpse some special value that highly vulnerable embodied conscious agents like us can realize that could plausibly thought to outweigh the evils—such as suffering—resulting from such embodiment, it follows that it is not surprising that God would create an ECA-structured universe. I propose that one such value is the ability to engage in particular kinds of virtuous actions—such as self-sacrificial love, courage, and the like—that the vulnerability that comes from embodiment allows…these actions allow for eternal connections of appreciation, contribution, and intimacy between conscious agents. For example, if someone significantly helps me in times of suffering, it can create a connection of appreciation in me that has the potential of lasting all eternity, and hence growing in value.

5-Consistency With Background Knowledge: Theism scores poorly here since the agent they posit is so dissimilar to any other agent with which we are familiar. When the mechanisms posited by potential explanation are consistent with what we already know about the world this should contribute to our willingness to accept it, and when there is an inconsistency we should treat it with suspicion. They should be consistent with observable facts and with our best existing theories. Darwin did this with natural selection by pointing to artificial selection. Better explanations (other things being equal) introduce mechanisms, entities, or concepts that are used in established explanations. ‘the key power…that of fulfilling intentions directly without any physical or causal mediation without materials of instruments is not made comprehensible by anything in our background knowledge, let alone likely that anything should have such a power. All of our knowledge of intention-fulfillment is of embodied intentions being fulfilled indirectly by way of bodily changes’ Dawes tries to show a disanalogy between substance dualism and God but I fail to see what it is.

http://www.apologeticsinthechurch.com/27/post/2011/06/response-to-god-doesnt-have-a-brain-so-probably-he-doesnt-exist-argument.html

The fact that a theory may have a narrow scope—in the sense that it purports to explain only a small range of phenomena—should not count against it. Narrow scope is not necessarily an explanatory vice. What makes us favor a theory is its capacity to explain facts that its competitor SHOULD be able to explain, but cannot. In this sense alone is explanatory power a suitable criterion for theory choice.

--Quantum teleportation, or entanglement-assisted teleportation, is a process by which a qubit (the basic unit of quantum information) can be transmitted exactly (in principle) from one location to another, without the qubit being transmitted through the intervening space.

--In physics, nonlocality or action at a distance is the direct interaction of two objects that are separated in space with no intermediate agency or mechanism.

Is Divine Action inconsistent with our background knowledge?

1-Efficient causation without material causation/material mechanisms is inconsistent with our background knowledge.

No, it is not. The expansion of space, quantum non-locality, certain cosmological models that have been offered seem to have no problem with the notion of efficient causation without material causation (e.g. Steady State Theory, and that other model I mentioned: (From Craig) Many scientists say that perhaps the positive energy in the universe and the negative energy in the universe exactly cancel out, so that on balance there is no energy in the universe. In that case you wouldn’t need to have a material cause to bring the positive and negative energy into being. You could just have an efficient cause.

Craig’s suggestion that God may have thought the universe into being: Perhaps it would be

helpful here to think of cases where we could have efficient causation without material causation.  I’ve been working heavily on the topic of abstract objects like numbers, sets, propositions, and so on.  Many philosophers believe that these immaterial objects exist necessarily and eternally.  But there are many abstract objects which seem to exist contingently and non-eternally, for example, the equator, the center of mass of the solar system, Beethoven’s Fifth Symphony, Leo Tolstoy’s Anna Karenina, and so forth.  None of these is a physical object.  Tolstoy’s novel, for example, is not identical to any of its printed exemplars, for these could all be destroyed and replaced by new books.  Nor can Beethoven’s Fifth be identified with any particular series of ink marks or any performance of the symphony.  Now these things all began to exist:  the equator, for example, didn’t exist before the earth did.  But if they began to exist, did they have a cause or did they come into being out of just nothing?  (Notice that it makes sense to ask this question even though these entities are immaterial and so have no material cause.)  Many philosophers would say that they did indeed have a cause: it was Tolstoy, for example,  who created Anna Karenina.  So in cases such as these (and they are legion), we do, indeed, have instances of efficient causation without material causation.  You may not agree that such abstract objects really exist;  but I think we have to say that the view defended by our philosophical colleagues is a coherent one. 

The examples of literary and musical creation are suggestive.  Could God have analogously thought the universe into being, just as Tolstoy created Anna Karenina?  It’s a provocative idea.

Craig again in a debate: He then asserts that it is logically impossible to have a creation of a material world by an immaterial being. Well, I simply invite him to give me the proof. That is an enormous claim that he is making—to say that it is logically impossible. I can certainly conceive of this happening, so how can he show me it is logically impossible?  In fact, I suggested that we have an analogy to this in the mind-body relation. In the mind-body relation we do see an immaterial entity acting upon a physical reality.

2-It violate the law of conservation of mass/energy:

No it doesn’t: http://home.messiah.edu/~rcollins/Mind-Body%20Problm/Modern%20Physics%20and%20the%20Energy%20Conservation%20Objection%20to%20Mind-body%20Dualism.doc

3-All the minds we know of have brains, so God probably doesn’t exist since he has a brain:

This would only show that everything IN the universe that is sentient will also have a brain, not that something BEYOND the universe without a brain probably doesn’t exist. A parallel argument that nobody defends can be made against the existence of abstract objects: Everything we see is physical, therefore nothing non-physical SIMPLICITER probably exists. What this

argument shows is that nothing IN the universe is probably non-physical, but it doesn’t show that there are no ABSTRACT objects which exist BEYOND the universe. This is the very same point that Paul Draper tries to make against the first premise of the kalam: everything that begins to exist WITHIN the universe has a cause is not the same as things that begin to exist WITH the universe. Craig says both are true because the first premise is a metaphysical principle, and I think he is right; whereas people have yet to show us that every sentient being having a brain is nothing more than a PHYSICAL principle like the law of gravity rather than a METAPHYSICAL PRINCIPLE about being and consciousness

However, consistency with background knowledge is a contingent matter: if we already had a tradition of successful theistic explanations this could change

There is nothing particularly authoritative about this list. The features it names are closely related and it may be that with a bit of reworking the number could be reduced to three or four. And perhaps more desiderata could be added. For example, the FECUNDITY of a theory-its ability to suggest new lines of research-would be one obvious addition. But one has to stop somewhere, I have chose these for reasons of convenience…If a proposed theistic explanation were to rate poorly when measured against these criteria this would be a sufficient reason to seek a naturalistic alternative.

Swinburne denies the relevance of background knowledge for theistic explanations because as the scope of our theories increases the significance of background knowledge diminishes. We should not demand that the entities whose existence it posits resemble those that the theory is meant to explain. For example, it is no objection to some theory of physics postulating fundamental particles and purporting thereby to explain the physical and thereby chemical behavior of medium sized objects that it postulates particles quite unlike those medium-sized objects. But theism seeks to explain the widest scope imaginable, the existence of the universe and so there is no background knowledge other than purely logical knowledge, which it has to fit. Dawes thinks we should divide background knowledge and evidence differently: (E)-The facts that fall within the scope of the explanation; (K)-Those that do not constitute background knowledge. While positing the existence of an unembodied agent might be warranted, if that hypothesis possessed other explanatory virtues, the fact that it posits an otherwise unknown kind of mechanism counts against it.

PAST EXPLANATORY SUCCESS: This can be classed under background knowledge. From a Bayesian point of view, you might argue that the past failure of the tradition of theistic explanation lowers the prior probability of any proposed theistic hypothesis. I am not speaking of what we might call the ‘track record’ of the same hypothesis. For given that a hypothesis can be corroborated by known facts, considerations of past explanatory success would come under the heading of corroboration. Any particular hypothesis can be seen as part of a research programme or a research tradition, which unites a series of proposed explanations sharing certain common assumption. But of course, we may yet require a theistic hypothesis.

RELATIONSHIP OF NATURAL FACTS AND THEISTIC EXPLANATIONS:

Facts lacking a natural explanation / Fine-tuning, beginning of the universe: In principle and in fact.

Naturally Unexplained Facts / God of the Gaps: We may not have a potential natural explanation but we might some day such as a person who mysteriously recovers from an illness.

Naturally Inexplicable facts: We could never have a natural explanation of THIS. All the atheist could say is that it is just a brute fact such as the existence of the universe or the laws of nature.

Facts Having Natural Explanations: Theists sometimes offer complementary explanations of natural facts and other times competing explanations of natural facts.

Fact With Accepted Natural Explanations: Walls of Jericho.

This criteria seems correct with respect to naturally UNEXPLAINED facts but I don’t see how it could make the prior probability of a proposed theistic explanation that uses accepted facts that either have or lack a natural explanations.

Theistic explanations are typically of the form ‘C caused E’ rather than ‘C-events cause E-events’ which makes them singular by their very nature. This a description of theistic explanations not an explanation itself. But does it make sense to speak of a singular causal explanation? The first question is metaphysical regarding causation, and the other is epistemological regarding explanation. Something is a cause only if it is regularly associated with an effect, the relata here being event-types (necessary condition). Causation is a matter of law-like association. Events of type A are regularly followed by events of type B. Any causal explanation must reflect this law-like association. However, Donald Davidson argues that intentional explanations are law-less (or anomalous) while still holding to a HUmean view; Davidson reconciles these two kinds of explanations by distinguishing between two levels of explanation: Intentional explanations do not cite laws, but facts to which they refer could also be explained using the language of the natural sciences, and such language would cite causal laws. Singular causation is not merely a claim about the structure of our explanations it involves a metaphysical claim about causation; Dawes thinks that unless we hold that causation is nothing other than a regular association of events, then there is no reason to deny the possibility of singular causation; an event could be unique in the history of the universe and yet known to be a case of causation.’

3 features of proposed theistic explanations: 1-they cover both events and states of affaris; 2-they are generally singular explanations; 3-there is a certain relationship between theistic and natural explanations…on some occasions TE cover facts that admit of a natural explanation. On other occasions divine action invokes an agent capable of miracles.

EVENTS AND STATES OF AFFAIRS: Facts are events, and ‘states’ could be subsumed under a common definition of events. They are both exemplification by substances of properties at a time. This is not without its relevance to causation/

STATES OF AFFAIRS: Theism doesn’t always explain an event in space and time that is bounded but sometimes it is a state of affairs that exists over time and may be distributed in space; it may even be co-extensive with time (as in the case of the existence of the universe, the laws of nature, why the universe exists).

EVENTS: These explanations are narrow in scope (floods, earthquakes, resurrections)

In general, theistic philosophers favor theistic explanations of states of affairs over explanations of particular events. The broader the better (e.g. the laws nature); what would it even mean to offer a natural explanation of the laws of nature?

RESPONSE:

1-Methodological naturalism can and does support theism. Both the kalam cosmological argument and the fine-tuning argument garner justification for their premises according to epistemological or methodological naturalism, but entail the contradictory of ontological naturalism.

2-http://www.apologeticsinthechurch.com/27/post/2011/11/naturalistic-induction-replacing-supernatural-explanations-with-natural-explanations.html

3-This argument has a couple of assumptions that seem dubious to me. One, I don’t think we have any way to know the antecedent probability of how many supernatural explanations there would be in the universe on theism compared to naturalism. Second, a miracle, by its very nature, wouldn’t ever replace an established natural explanation because in that case it would cease to be a miracle and become a new law of nature. But let’s try to assume we can know something of the antecedent probability of how many miracles there would be if theism were true. I think there are a number of reasons we shouldn’t be surprised that there are laws of nature that operate ‘uninterrupted’ regularly. It would be irrational for God to set up the laws to constantly be suspending them. Second, it would betray a lack of perfection to constantly have to tinker with them. Third, the laws of nature must operate regularly in order for us to be able to make rational decisions. Fourth, just because God COULD do something doesn’t entail that he SHOULD do it. It seems God shouldn’t perform a miracle without having a sufficient reason such as confirming a message that cannot be read off of nature (e.g. the resurrection of Jesus after his ministry), and the occasions where such a miracle should be done may be as low as one. I think this shows that the antecedent probability on theism for there to be a HIGH FREQUENCY of miracles is (self-defeating for one thing, but also) different than the antecedent probability on theism to expect some PARTICULAR INSTANCE (however infrequent) of a miracle that we do or don’t have evidence for, and it is the latter that needs to be shown for this type of argument to work; especially in light of the considerations being raised here. Fifth, there can be and I think there is good evidence for God’s existence that doesn’t require a supernatural explanation to REPLACE a natural one, but is instead complementary or constituted by both a proximate natural explanation and an ultimate supernatural explanation. Sixth, this objection is basically saying that it is very unlikely on theism that there would be laws of nature! What? Quite the opposite according to some metaphysical arguments. Seventh, it seems that this argument begs the question since there are supernatural explanations of the origin of the universe, the fine-tuning of the universe, and the resurrection of Jesus. But, as I say, the main

problem seems to be that both the NUMBER and NATURE of supernatural explanations for events in the world is supposed to be x-number of law-like events, but since we don’t observe this, this confirms metaphysical naturalism over theism. I don’t know how you could know this and in fact, there are plausible considerations to think it is wrong.

If the theist is to have any trust in our knowledge of the world, he is bound to assume that his God will not work TOO MANY miracles. This objection presents us with a false dichotomy. It assumes that the world is either governed by natural laws or it is the product of mere whim on the part of a supernatural agent. The world could also be the work of a supernatural agent acting rationally in order to achieve his goals so that we can predict behavior is we cite a goal. What would be the most rational way for such an agent to act in order to achieve this goal? The answer may or may not include miracles.

Unless one holds that causation is NOTHING OTHER than a regular associateion of events, there is no reason to deny the possibility of singular causation; an event ‘could be unique in the history of the universe, and yet be, and be known to be, a case of causation (e.g. the origin of the universe, the fine-tuning of the universe, the singularity).

Epistemological Issue: Intentional explanations need not cite causal laws (even if the facts to which they refer could also be explained using the language of the natural sciences)—Appendix 3.3.1

It is true that one could use an intentional explanation to construct a general law: If there existed antoher rational agent of precisely this type (having the same beliefs and desires) then when placed in exactly the same situation he would act in the same way (this entails a potential law-like regularity). However, this is pretty loose. The more important point is that intentional explanations do not rely on generalizations linking intentions and actions. What they rely on is a calculation of how a rational agent in this situation would be expected to act, given certain beliefs and desires from which it follows that our expectation regarding an agent’s behavior is not based on any general intention-action rule.

If a theist is offering a singular explanation, we cannot test it by replicating condition C and seeing if event E is observed, but this doesn’t mean it is untestable (Appendix 3.3.2). There are two descriptions of an event, one which is ‘physical’ and one which is intentional. After all, causation is a feature of the world, not a feature of our descriptions of it so that it would embody the representational fallacy which happens when we wrongly assume that we can deduce the structure of the world from the language we use to describe it. It is on the level of the event that there exists a cause; we need to abandon premise 1 then, (causation involves law-like relations).

CHAPTER 8: Conclusion

Chpts 1-4: Objections to theistic explanations considered

1-No possible state of affairs is excluded by theistic explanations (TE’s)

2-Actions of an agent capable of miracles would be unpredictable

3-The Concept of God is incoherent

4-The will of God cannot be a cause

Chpt. 5: Set out the circumstances in which invoking a divine agent would constitute a potential explanation of some state of affairs (e.g. the optimality condition). This will be difficult given the existence of pointless evils.

Chpt. 7: TE’s would also have to meet a list of accepted explanatory virtues and that theism is simply incapable of achieving a high score. TE’s are not consistent with our background knowledge since it has failed in the past and naturalistic explanation have not; it does not allow us to predict the precise details of the effect. So, it seems remote that there will even be even a POTENTIAL THEISTIC explanation since they fail to meet the optimality condition. We will always have good reasons to seek natural alternative.

However, consistency with background knowledge is a contingent matter: if we already had a tradition of successful theistic explanations this could change; ontological economy is a certeris paribus condition that suggests we shouldn’t posit new kinds of entities unless we are required and it is at least conceivalb ethat positing a divine agent might be required to explain some phenomenon; while TE’s might not be lovely and allow us to deduce the precise details of an effect, this does not seem (by itself) fatal if TE’s were to score high on other criteria. So, even a presumption of naturalism is defeasible.

APPENDIX: Intentional Explanations

Ratcliffe argues that we can explain a man behavior without attributing any mental states such as beliefs, desires, and intentions by reference to social norms; this might not be a causal explanation but we often understand behavior this way; unfortunately we don’t even have anything analogous to social norms in the case of God. All we can do is post a divine agent, attribute certain beliefs and desires and see what explanatory force this offers.

ACTIONS, REASONS, AND CAUSES:

Can a reason also be a cause? Davidson’s two theses:

1-BD (belief-desire thesis)-R is a primary reason why an agent performed the action A under the description d only if R consists of a pro attitude of the agent towards actions with a certain property, and a belief of that agent that A, under description d, has that property.

A pro attitude can be an evaluative belief and not every desire constitutes a reason to act. This thesis needs amending. Dawes does this by distinguishing between an intention and the formation of an intention; an intention is a kind of pro attitude while the latter is a judgment that something is to be done. An intentional explanation is causal because it posits an intention and shows that the action to be explained is a rational one, given that intention. Just as the act of accepting a proposition can be a voluntary act, even though forming a belief is not, so that act of forming an intention can be a voluntary act, even though having a desire is not. We may want to go further and say even if an intention is a sui generis mental state, and even if the formation of an intention is an act on the part of the agent, we can

still ask why an agent formed this intention rather than another. It is at this point that we will need to invoke the agent’s beliefs and desires. How would invoking his beliefs and desires explain his intention.

THE RATIONALITY PRINCIPLE: We explain an agent’s action by positing a particular intention and by offering a practical syllogism that has a description of that action as its conclusion. But to do this is to presume that the agent is acting rationally. What does it mean to treat something as a rational agent?

1-A system’s beliefs are those it ought to have, given its perceptual capacities, its epistemic needs, and its biography…

2-A system’s needs are those it OUGHT TO HAVE, given its biological needs, and the most practicable means of satisfying them…

3-A systems’ behavior will consist of those acts that it would be rational for an agent with those beliefs and desires to perform.

The first two are uncontroversial so long as we interpret ‘ought’ in expectations not moral. The third requirement is about rational agents and Davidson has a similar principle ‘the principle of charity’ in which he demands that we find ‘a large degree of rationality and consistency in the behavior of those whose actions we are explaining.’ This principle is helpful because it works when treat people as if they were rational agents it allows us to understand and even predict their behavior on the basis of three criteria: 1-consistency, 2-perceived efficacy, 3-efficiency

2-CT (causal thesis)-A primary reason for an action is its cause.

If we do not recognize a reason as a cause, then it is hard to make sense of the idea that I am acting for a reason; this leaves open the question what kind of cause a reason is; it may be one that does not operate according to strict and it might not eliminate the agent herself from the process.

EXPLANATION AND ARGUMENTS: There may be a time when we cannot offer a plausible intention that would an agents behavior rational, even by her own lights. In this situation we should seek some other kind of explanation. We can do this by means of a practical syllogism which posits an intention as a premise and the action as its conclusion. The extent to which this syllogism actually reflects the agents reasoning process will differ from case to case:

Major Premise: the motivational premise-the agent wants x

Minor premise: the cognitive premise: her A-ing would contribute to realizing x.

Conclusion: the practical judgment-the agent should A.

This form of the argument has two weaknesses: 1-having a desire is not the same as having a motivation and it is not clear how the conclusion follows from the premises. Here is how Dawes would do it:

1-There exists a rational agent A with intended goal G.

2-A has beliefs B1, B2,…, Bn relating to the attainment of G.

3-If B1, B2,…Bn were true, E would be the best way of achieving G.

4-Rational agents always choose the best way of achieving their goals (relative to the intended goal, the available options, etc.)

5-Therefore, A will do E.

EXPLAINING THE EXPLANANS: Assuming that an explanation can be complete even if it does not explain its explanans, the above syllogism represent an intentional explanation. But, then we can ask why an agent acted on one set of potential reasons rather than another by forming a new syllogism with the agent’s desires and beliefs for its premises, and a description of her intention as its conclusion. This will explain how the agent weighed up the various potential reasons for action to which her bleifs and desires could have given rise.

(1) Moral agents are a good thing.(2) Existing alone an OMNI-PERFECT moral agent the best possible world quantitatively.(3) At the same time, there could be more qualitative goods in a world where an OMNI-PERFERT

MORAL AGENT exists with other moral agents. (4) It would be better for me to create a world with moral agents in it.(5) Therefore, I should create a world with moral agents in it.

THE PROBLEM OF AKRASIA: First-order intentional explanation takes us from posited intention to action, and second order explanation, which explains why the agent formed this particular intention. Akrasia is still an intentional action even though it may not be fully rational.

TESTING INTENTIONAL EXPLANATIONS: Are they testable?

Asking the Agent-We could predict what an agent will do by asking them. This is not always an option of course (e.g. the agent may be dead as in historical explanations).

Weakness of Will: Akrasia also highlights the need to appeal to the character of the agent since it will be the case that an agents beliefs and desires will lead to an action depending on her character as well .

Intentions and Laws: They are anomalous in the sense that they don’t rely on laws, and so it is unclear how we can make predictions that are testable. And it is hard to see how the causal thesis can be defended as well without appealing to causal laws.

Causation and Laws: Metaphysical question: Is there a law-like relation between events of type A and events of type B, and an Epistemic Question: in order for an explanation to be adequate do I have to explain event B by citing a law that connects events of type with events of type B? Dawes will answer no. It is true that one could use an intentional explanation to construct a general law where we could lsay: if there existed another rational agent of precisely this type (having the same beliefs and desires) then when placed in exactly the same situation he would act in the same way. In this sense one could

argue that any singular explanation entails the existence of a regularity, in the sense of a potential regularity. But this isn’t very helpful. It would be better to say that agents with similar character, in similar situations, with similar beliefs and desires, will act in similar ways which will allows us to construct some generalizations about human behavior. We could also connect reports of agent’s intentions with observed behavior inductively thereby constructing a pattern of behavior. This is not necessary however since there exists another constraint on intentional explanations, namely, the rationality principle which presumes that rational agents will act rationally. But the main point is that intentional explanations do not rely on generalisations linking intentions and actions. What they rely on is a calculation of how a rational agent in this situation would be expected to act, given certain beliefs and desires.

Causation without laws:

1-Causation involves law-like regularities.

2-Intentional explanations do not cite laws.

3-Intentional explanations are causal explanations.

This triad, if consistent, will only work if 2 is epistemic and 3 is ontological. So, this would reduce reasons to epiphenomena. Thus we must deny 1. Dawes does this by arguing that it is wrong to assume that we can deduce the structure of the world from the language we use to describe it.

The Prediction of Behavior: Robbing a bank prediction allows for corroboration even though they don’t depend on laws. We can test an intentional explanation because there will be certain predictable consequences that extend beyond the fact to be explained. While it is true that hypotheses are not tested one-by-one, but in bundles this doesn’t prevent scientific theories from being independently testable and we often test the auxillary assumptions independently. Eventually, we will decide whether the theory in question represent a degenerating research programs and should be abandoned.

FALSIFICATION AND FAILURE: