An Argument against the Argument against Miracles

[Don Mc]

According to Hume's famous "general maxim" against the confirmation of miracles in the Enquiry Concerning Human Understanding, "no testimony is sufficient to establish a miracle, unless the testimony be of such a kind, that its falsehood would be more miraculous, than the fact, which it endeavors to establish." The basic idea is that the laws of nature being what they are, and human nature being what it is, the probability of a miracle is always lower than the probability that the testimony given for it is simply false. In this Hume seems to have anticipated the logic of Carl Sagan, who popularized the idea that "Extraordinary claims require extraordinary evidence."

While this principle appears rational enough at first blush, there are reasons to think it's not sound. First, it was Hume himself who spelled out the problem of induction – that there is no logical basis for inferring future outcomes from past experiences. Assuming there exists a set of well-defined "laws of nature," those regularities would seem to be descriptive rather than prescriptive. But if the laws of nature are descriptive, there is no reason to think miracles cannot or should not occur. Second, the argument against miracles is essentially circular. Hume asserts that there is "uniform experience" against the resurrection, for example, adding that a man risen from the dead "has never been observed, in any age or country." The question of the resurrection, however, is precisely whether or not Jesus was observed by his disciples to have risen from the dead. To say that a resurrection event was never observed because there is "uniform experience" against it is to beg that question (and we should bear in mind that there is equally uniform experience that life does not arise from nonliving elements – yet here we are). Finally, while it's true that human nature has the potential to corrupt the testimony of eyewitnesses and the writings of biographers and historians, it also has the potential to corrupt the field reports, lab results, journal articles, textbooks, etc., that lead us to accept the same scientific theories thought to render miracle reports implausible or even impossible. The problem of "confirmation bias" among humans, and scientists in particular, is well documented.

Evidently underlying popular skepticism of miracles is a belief that miracles are inherently, extremely improbable. But that seems to hold only if a miracle is defined in naturalistic terms. After all, the proposition "A man rose from the dead by natural processes" appears considerably less probable on its face than the proposition "Jesus Christ rose from the dead by the power of God." As Paul put it, "Why should any of you consider it incredible that God raises the dead?" (Acts 26:8)

Questions for debate:
Are miracles improbable? If so, how improbable are they and why?
Could historical evidence for a miracle give us good evidence for theism?

[Mithrae]

Backing up a bit might be helpful; sometimes I feel a bit bogged down in the back-and-forth of so many post segments. So just trying to get my bearings here... to begin with, I'm not sure if or where you've shown yet that a divine intervention hypothesis is substantially different from any other hypothesis of comparable specificity?

Divine intervention can literally explain any event. It's the least specific (i.e. restricted) explanation you could possibly have for anything. There are no "comparables".

If you want a type of explanation supposed to cover any event, natural laws fit the bill just as well as divine action and arguably moreso than the divine intervention implied by miracle claims. The concept of natural laws is as if not more nonspecific and subject to revisions/adjustments as that of divine intervention. Obviously if this were a basis for proclaiming the broad concept of divine intervention unfalsifiable and/or inadmissible, then the same must apply to the broad concept of natural laws.

In post #48 you said that divine intervention hypotheses can and often have been falsified 'colloquially'

All I really meant was that some people would say "Oh this falsifies the miracle claim", and that such a statement is meaningful. Strictly speaking, they would be wrong, or at least what they mean by the 'miracle' is just in reference to the outcome (I have been using "effect" until now, but I actually think "outcome" is better).

When you falsify the outcome of a supposed event, for the mechanism ("cause") you are making any explanation unnecessary in the first place. This is not the same as showing the supposed mechanism to be false, but some people will not recognize/appreciate this distinction and will conflate. That's all there is to it.

This seems to be quite a sweeping assumption: I would say on the contrary that many known falsifications of miracle claims don't involve disputing the outcome at all. Weeping statues shown to be caused by condensation or hidden pipes are falsifications entirely focusing on the mechanism, not the outcome of water down the statue's face; 'miraculous' healing shown to be caused by natural processes unknown to the patient's carers again involves the mechanism, not the outcome. It's not clear whether your comments imply denying that those cases exist, or denying that they are falsifications involving mechanism rather than outcome, or denying that they are falsifications at all - but any of the three would seem to be incorrect assumptions.

(such as being proven as not having occurred, hoaxes/fraud, more plausibly explained etc.) just not 'epistemically,'

Proving hoax/fraud is just a variation of showing that the supposed outcome did not actually occur. "More plausibly explained", on the other hand, is entirely different and certainly NOT what I meant; it is erroneous/circular because it implies there was any plausibility to the miracle explanation in the first place.

Supposing that the miracle explanation has zero plausibility is either an unjustified/erroneous philosophical assumption, or a circular argument based on its supposed unfalsifiability (which is what you are trying to demonstrate). As shown above, in some cases proving hoax/fraud doesn't involve disputing the outcome at all and are indeed simply showing a better explanation.

in terms of general knowing the purpose of falsifiability is to recognize and avoid the scenario in which any circumstances are taken as confirming a conclusion.

Yes I agree. It's just important to remember that the outcome being falsifiable does not necessarily imply that the proposed mechanism is.

You say that you agree, but you've snipped and failed to address the obvious consequence of that purpose for falsifiability, namely that:
Demonstrated examples of 'colloquial' falsification more than accomplish that; so we can indeed say that some circumstances are far less compatible with a miracle conclusion, meaning that we can validly conclude other circumstances to be far more indicative of a miracle conclusion.

In post #60 you suggested that divine intervention hypotheses can be 'deductively' falsified - just not 'inductively' - but (setting aside the more dubious latter part for now) I don't think the core of my response was addressed: That inasmuch as deductive falsification is even in view, the distinction seems to be an artefact of your focus on and my reply regarding the functionality of a miracle theory (its logical coherency or possibility), in contrast to a more practical focus on the relative probabilities of that and competing theories.

Your argument went something like this:

1) "Miracle" is a good explanation unless idolatry occurs
2) Idolatry occurs
C) "Miracle" is not a good explanation

Therefore "Miracle" is "falsifiable".

The problem is that if no idolatry occurs, then you could observe the whole universe in both space and time and not be able to falsify the "miracle" explanation, even if it is, in fact, false. For example, you could replace "idolatry occurs" with any event, such as "Donald Trump does a backflip" or "Bertrand Russell's teapot exists" and claim that you proved the miracle to be falsifiable. Surely you see how arbitrary and silly such an argument is? You're just having idolatry occur in this hypothetical universe to meet the "observation" criteria I had set, when in fact there is no necessity for it to occur at all. At best you could call it "maybe falsifiable", as weird as that sounds. Scientific theories do not work this way; if a theory is false you would necessarily be able to verify its falsehood by observing the whole universe.

I called it "deductive" falsification because all it amounts to word games on your part. You can "falsify" literally anything in some hypothetical world if you insert a falsification criteria into its definition and then have that criteria met in that hypothetical world; and that's exactly what you did.

But this was not my best objection to your example. The best one was to simply point out that your (1) is itself what's unfalsifiable. For this reason, if you are confused, you can just forget I said anything about "deductive falsifiability".

Scientific reasoning and good empirical reasoning outside of science generally goes something like this:
A) Observation
B) Explanation/hypothesis, preferably (and esp. in the sciences) with predictive capacity
C) Further observation or experimentation
D) Falsification, correction or provisional confirmation of hypothesis

That is the model followed by the example I used to respond to your 'functionality' argument:
- remarkable observation in religious context,
- explanation as a miracle by an Abrahamic God, with implied prediction against idolatry
- further observations inconsistent with that explanation
- falsification of that explanation

Now firstly, you claim that without that specific observation of idolatry "you could observe the whole universe in both space and time and not be able to falsify the 'miracle' explanation"; but your claim is simply and obviously false. As we've already discussed ad nauseum, further observations showing the remarkable event to be caused by human trickery or some other more normal process would obviously falsify the miracle explanation. Simply denying such possibilities or stubbornly insisting that these falsifications do not involve the mechanism at work is not a rational position.

Secondly, because the specific type (from many possible forms) of falsification used in my example was tailored to address your 'functionality' argument, you say that it is "arbitrary and silly" having the occurrence of idolatry as that falsifying observation and that "in fact there is no necessity for it to occur at all." But what exactly do you think you're arguing there? Are you honestly trying to claim that for every other explanation which is falsifiable in principle, and in fact false, there exists some "necessity" that a particular falsifying observation will be made? That's obviously ridiculous, as your appeal to universal observation throughout all of time and space tacitly acknowledges. We don't have access to omniscient observation - and if we did it would falsify all false explanations alike, whether divine or otherwise - and within the scope of observation we do possess, miraculous and mundane explanations can both be falsified in practice at times, and at other times escape falsification even if incorrect and falsifiable in principle. All you're trying to do is invent some imaginary middle ground where all non-divine explanations are falsified if false and divine explanations never are (despite being often falsified even with our limited observations), apparently by implicitly including a theology of divine causation being totally distinct from "space and time."

And thirdly you're claiming that the... let's say the theory behind the explanation itself (as opposed to its application to particular observations) is unfalsifiable. But as I noted above, and in discussion with Bluegreenearth, that is just as true of alternative broad theories such as natural laws as it is of theism or its particular variants. Such broad theories can be subject to philosophical debate, refinement - the view of God in the Abrahamic traditions alone has changed considerably over time! - and varying degrees of plausibility relative to their competing or complementary theories, but they can't really be proven false; there's nothing unique about theism in that regard. But more to the point, an abstract unfalsifiable theory with no rational connection to the real world of observation obviously has little or no merit or purpose; it is essentially ruled out or falsified in pragmatic terms even if not on its own theoretical level: That's why I have consistently tried to steer the discussion back to the relative probabilities of competing explanations for observed phenomena rather than their abstract 'functionality.'

[FarWanderer3]

Backing up a bit might be helpful; sometimes I feel a bit bogged down in the back-and-forth of so many post segments. So just trying to get my bearings here... to begin with, I'm not sure if or where you've shown yet that a divine intervention hypothesis is substantially different from any other hypothesis of comparable specificity?

Divine intervention can literally explain any event. It's the least specific (i.e. restricted) explanation you could possibly have for anything. There are no "comparables".

If you want a type of explanation supposed to cover any event, natural laws fit the bill just as well as divine action and arguably moreso than the divine intervention implied by miracle claims. The concept of natural laws is as if not more nonspecific and subject to revisions/adjustments as that of divine intervention. Obviously if this were a basis for proclaiming the broad concept of divine intervention unfalsifiable and/or inadmissible, then the same must apply to the broad concept of natural laws.

In post #48 you said that divine intervention hypotheses can and often have been falsified 'colloquially'

All I really meant was that some people would say "Oh this falsifies the miracle claim", and that such a statement is meaningful. Strictly speaking, they would be wrong, or at least what they mean by the 'miracle' is just in reference to the outcome (I have been using "effect" until now, but I actually think "outcome" is better).

When you falsify the outcome of a supposed event, for the mechanism ("cause") you are making any explanation unnecessary in the first place. This is not the same as showing the supposed mechanism to be false, but some people will not recognize/appreciate this distinction and will conflate. That's all there is to it.

This seems to be quite a sweeping assumption: I would say on the contrary that many known falsifications of miracle claims don't involve disputing the outcome at all. Weeping statues shown to be caused by condensation or hidden pipes are falsifications entirely focusing on the mechanism, not the outcome of water down the statue's face; 'miraculous' healing shown to be caused by natural processes unknown to the patient's carers again involves the mechanism, not the outcome. It's not clear whether your comments imply denying that those cases exist, or denying that they are falsifications involving mechanism rather than outcome, or denying that they are falsifications at all - but any of the three would seem to be incorrect assumptions.

You provide excellent examples of mechanisms. We're narrowing things down and that's good.

So then, what makes these particular alternative natural explanations better than a divine intervention explanation?

(such as being proven as not having occurred, hoaxes/fraud, more plausibly explained etc.) just not 'epistemically,'

Proving hoax/fraud is just a variation of showing that the supposed outcome did not actually occur. "More plausibly explained", on the other hand, is entirely different and certainly NOT what I meant; it is erroneous/circular because it implies there was any plausibility to the miracle explanation in the first place.

Supposing that the miracle explanation has zero plausibility is either an unjustified/erroneous philosophical assumption, or a circular argument based on its supposed unfalsifiability (which is what you are trying to demonstrate). As shown above, in some cases proving hoax/fraud doesn't involve disputing the outcome at all and are indeed simply showing a better explanation.

Again, here you specifically call them "better". Why? What makes them so?

in terms of general knowing the purpose of falsifiability is to recognize and avoid the scenario in which any circumstances are taken as confirming a conclusion.

Yes I agree. It's just important to remember that the outcome being falsifiable does not necessarily imply that the proposed mechanism is.

You say that you agree, but you've snipped and failed to address the obvious consequence of that purpose for falsifiability, namely that:
Demonstrated examples of 'colloquial' falsification more than accomplish that; so we can indeed say that some circumstances are far less compatible with a miracle conclusion, meaning that we can validly conclude other circumstances to be far more indicative of a miracle conclusion.

Please do not presume that I also agree with your claim that "Demonstrated examples of 'colloquial' falsification accomplish that".

Also, I made the mistake of not editing your statement to "in terms of general knowing the purpose of falsifiability is to recognize and avoid the scenario in which any circumstances are taken as being consistent with a conclusion".

In post #60 you suggested that divine intervention hypotheses can be 'deductively' falsified - just not 'inductively' - but (setting aside the more dubious latter part for now) I don't think the core of my response was addressed: That inasmuch as deductive falsification is even in view, the distinction seems to be an artefact of your focus on and my reply regarding the functionality of a miracle theory (its logical coherency or possibility), in contrast to a more practical focus on the relative probabilities of that and competing theories.

Your argument went something like this:

1) "Miracle" is a good explanation unless idolatry occurs
2) Idolatry occurs
C) "Miracle" is not a good explanation

Therefore "Miracle" is "falsifiable".

The problem is that if no idolatry occurs, then you could observe the whole universe in both space and time and not be able to falsify the "miracle" explanation, even if it is, in fact, false. For example, you could replace "idolatry occurs" with any event, such as "Donald Trump does a backflip" or "Bertrand Russell's teapot exists" and claim that you proved the miracle to be falsifiable. Surely you see how arbitrary and silly such an argument is? You're just having idolatry occur in this hypothetical universe to meet the "observation" criteria I had set, when in fact there is no necessity for it to occur at all. At best you could call it "maybe falsifiable", as weird as that sounds. Scientific theories do not work this way; if a theory is false you would necessarily be able to verify its falsehood by observing the whole universe.

I called it "deductive" falsification because all it amounts to word games on your part. You can "falsify" literally anything in some hypothetical world if you insert a falsification criteria into its definition and then have that criteria met in that hypothetical world; and that's exactly what you did.

But this was not my best objection to your example. The best one was to simply point out that your (1) is itself what's unfalsifiable. For this reason, if you are confused, you can just forget I said anything about "deductive falsifiability".

Scientific reasoning and good empirical reasoning outside of science generally goes something like this:
A) Observation
B) Explanation/hypothesis, preferably (and esp. in the sciences) with predictive capacity
C) Further observation or experimentation
D) Falsification, correction or provisional confirmation of hypothesis

That is the model followed by the example I used to respond to your 'functionality' argument:
- remarkable observation in religious context,
- explanation as a miracle by an Abrahamic God, with implied prediction against idolatry
- further observations inconsistent with that explanation
- falsification of that explanation

Now firstly, you claim that without that specific observation of idolatry "you could observe the whole universe in both space and time and not be able to falsify the 'miracle' explanation"; but your claim is simply and obviously false. As we've already discussed ad nauseum, further observations showing the remarkable event to be caused by human trickery or some other more normal process would obviously falsify the miracle explanation. Simply denying such possibilities or stubbornly insisting that these falsifications do not involve the mechanism at work is not a rational position.

Secondly, because the specific type (from many possible forms) of falsification used in my example was tailored to address your 'functionality' argument, you say that it is "arbitrary and silly" having the occurrence of idolatry as that falsifying observation and that "in fact there is no necessity for it to occur at all." But what exactly do you think you're arguing there? Are you honestly trying to claim that for every other explanation which is falsifiable in principle, and in fact false, there exists some "necessity" that a particular falsifying observation will be made? That's obviously ridiculous, as your appeal to universal observation throughout all of time and space tacitly acknowledges. We don't have access to omniscient observation - and if we did it would falsify all false explanations alike, whether divine or otherwise - and within the scope of observation we do possess, miraculous and mundane explanations can both be falsified in practice at times, and at other times escape falsification even if incorrect and falsifiable in principle. All you're trying to do is invent some imaginary middle ground where all non-divine explanations are falsified if false and divine explanations never are (despite being often falsified even with our limited observations), apparently by implicitly including a theology of divine causation being totally distinct from "space and time."

And thirdly you're claiming that the... let's say the theory behind the explanation itself (as opposed to its application to particular observations) is unfalsifiable. But as I noted above, and in discussion with Bluegreenearth, that is just as true of alternative broad theories such as natural laws as it is of theism or its particular variants. Such broad theories can be subject to philosophical debate, refinement - the view of God in the Abrahamic traditions alone has changed considerably over time! - and varying degrees of plausibility relative to their competing or complementary theories, but they can't really be proven false; there's nothing unique about theism in that regard. But more to the point, an abstract unfalsifiable theory with no rational connection to the real world of observation obviously has little or no merit or purpose; it is essentially ruled out or falsified in pragmatic terms even if not on its own theoretical level: That's why I have consistently tried to steer the discussion back to the relative probabilities of competing explanations for observed phenomena rather than their abstract 'functionality.'

I could argue these points, but frankly it's a waste of time. I already said it wasn't my best objection, and it seems to be all you want to talk about. Twice now I have made the point that your "It's a miracle unless idolatry occurs" scenario is unfalsifiable (even if idolatry occurs), and twice you have elected not to address it.

[Mithrae]

This seems to be quite a sweeping assumption: I would say on the contrary that many known falsifications of miracle claims don't involve disputing the outcome at all. Weeping statues shown to be caused by condensation or hidden pipes are falsifications entirely focusing on the mechanism, not the outcome of water down the statue's face; 'miraculous' healing shown to be caused by natural processes unknown to the patient's carers again involves the mechanism, not the outcome. It's not clear whether your comments imply denying that those cases exist, or denying that they are falsifications involving mechanism rather than outcome, or denying that they are falsifications at all - but any of the three would seem to be incorrect assumptions.

You provide excellent examples of mechanisms. We're narrowing things down and that's good.

So then, what makes these particular alternative natural explanations better than a divine intervention explanation?

Supposing that the miracle explanation has zero plausibility is either an unjustified/erroneous philosophical assumption, or a circular argument based on its supposed unfalsifiability (which is what you are trying to demonstrate). As shown above, in some cases proving hoax/fraud doesn't involve disputing the outcome at all and are indeed simply showing a better explanation.

Again, here you specifically call them "better". Why? What makes them so?

If we observe hidden pipes leading to a statue's eyes, and water being drawn up those pipes before running down the statue's face, we would conclude somewhere in the order of 100% probability of that being the explanation for the 'weeping' statue, leaving very little scope for plausibility in the competing miracle explanation. As I outlined in post #46,
"The sum probabilities of all mutually exclusive, exhaustive possibilities must equal 100%: If we have 90% confidence in one explanation, all other explanations (and the unknown) can only add up to 10%. Hence if 'miracle' had been the strong leading explanation for an event (at say 60% confidence) but then aliens explicitly taking credit and demonstrating how they'd done it led to a 90% confidence in that explanation, it would radically reduce the viability of the 'miracle' explanation for that event. Same goes for finding the hidden wires of a hoax, obviously."

If you're asking why those observations would lead to a near-certainty in that conclusion, beyond simply 'common sense,' I'm thinking that the relevant metrics would be the explanation's parsimony/elegance/simplicity and breadth/scope/comprehensiveness (as suggested in posts #21, 42 etc) and depth/specificity/predictive capacity (#58, 71 etc). By introducing fewer unknowns, a demonstrable explanation in terms of hidden pipes, condensation or the like would have better parsimony than a miracle explanation, and perhaps more importantly would have much greater depth/specificity.

You say that you agree, but you've snipped and failed to address the obvious consequence of that purpose for falsifiability, namely that:
Demonstrated examples of 'colloquial' falsification more than accomplish that; so we can indeed say that some circumstances are far less compatible with a miracle conclusion, meaning that we can validly conclude other circumstances to be far more indicative of a miracle conclusion.

Please do not presume that I also agree with your claim that "Demonstrated examples of 'colloquial' falsification accomplish that".

Also, I made the mistake of not editing your statement to "in terms of general knowing the purpose of falsifiability is to recognize and avoid the scenario in which any circumstances are taken as being consistent with a conclusion".

I'm not presuming you agree - that's why I said you didn't address it. It does seem obvious that in many if not most cases we can indeed avoid the scenario in which any circumstances are taken as supporting a miracle conclusion - which is the epistemic purpose of falsifiability - and hence that we can validly conclude some circumstances to be far more indicative of a miracle conclusion than others. Just as obviously, that's contrary to the position you've been arguing... but now for a third time you haven't answered it.

Scientific reasoning and good empirical reasoning outside of science generally goes something like this:
A) Observation
B) Explanation/hypothesis, preferably (and esp. in the sciences) with predictive capacity
C) Further observation or experimentation
D) Falsification, correction or provisional confirmation of hypothesis

That is the model followed by the example I used to respond to your 'functionality' argument:
- remarkable observation in religious context,
- explanation as a miracle by an Abrahamic God, with implied prediction against idolatry
- further observations inconsistent with that explanation
- falsification of that explanation

Now firstly, you claim that without that specific observation of idolatry "you could observe the whole universe in both space and time and not be able to falsify the 'miracle' explanation"; but your claim is simply and obviously false. As we've already discussed ad nauseum, further observations showing the remarkable event to be caused by human trickery or some other more normal process would obviously falsify the miracle explanation. Simply denying such possibilities or stubbornly insisting that these falsifications do not involve the mechanism at work is not a rational position.

Secondly, because the specific type (from many possible forms) of falsification used in my example was tailored to address your 'functionality' argument, you say that it is "arbitrary and silly" having the occurrence of idolatry as that falsifying observation and that "in fact there is no necessity for it to occur at all." But what exactly do you think you're arguing there? Are you honestly trying to claim that for every other explanation which is falsifiable in principle, and in fact false, there exists some "necessity" that a particular falsifying observation will be made? That's obviously ridiculous, as your appeal to universal observation throughout all of time and space tacitly acknowledges. We don't have access to omniscient observation - and if we did it would falsify all false explanations alike, whether divine or otherwise - and within the scope of observation we do possess, miraculous and mundane explanations can both be falsified in practice at times, and at other times escape falsification even if incorrect and falsifiable in principle. All you're trying to do is invent some imaginary middle ground where all non-divine explanations are falsified if false and divine explanations never are (despite being often falsified even with our limited observations), apparently by implicitly including a theology of divine causation being totally distinct from "space and time."

And thirdly you're claiming that the... let's say the theory behind the explanation itself (as opposed to its application to particular observations) is unfalsifiable. But as I noted above, and in discussion with Bluegreenearth, that is just as true of alternative broad theories such as natural laws as it is of theism or its particular variants. Such broad theories can be subject to philosophical debate, refinement - the view of God in the Abrahamic traditions alone has changed considerably over time! - and varying degrees of plausibility relative to their competing or complementary theories, but they can't really be proven false; there's nothing unique about theism in that regard. But more to the point, an abstract unfalsifiable theory with no rational connection to the real world of observation obviously has little or no merit or purpose; it is essentially ruled out or falsified in pragmatic terms even if not on its own theoretical level: That's why I have consistently tried to steer the discussion back to the relative probabilities of competing explanations for observed phenomena rather than their abstract 'functionality.'

I could argue these points, but frankly it's a waste of time. I already said it wasn't my best objection, and it seems to be all you want to talk about. Twice now I have made the point that your "It's a miracle unless idolatry occurs" scenario is unfalsifiable (even if idolatry occurs), and twice you have elected not to address it.

More than three times now I've made the point that what matters for our knowledge and understanding of the real world is the relative probability of competing explanations for observed phenomena rather than their abstract 'functionality,' and that this scenario showing the potential for negation of a miracle explanation's functionality is little more than a distraction as far as I'm concerned. I made that point in the four and a half lines mentioning your acknowledgement of deductive falsification in post #64... to which you devoted over thirteen lines in reply. Suggesting that it's all I want to talk about is hardly an accurate characterization.

As for your best objection to that tangential scenario - the unfalsifiability of its broader theory that "If there's idolatry it's not a miracle," distinct from specific application of that theory to explain remarkable events which turn out to be idolatrous - that's exactly what is addressed in some detail by my third point above.

[Mithrae]

Trying to summarize what I think I've learned from the discussion so far, at the top of the epistemic pyramid we have broad abstract theories such as materialism, deism, theism, naturalism/natural laws and so on which we might legitimately describe as 'unfalsifiable.' More precisely I would say that none of them seem likely to be completely ruled out in the foreseeable future, although they may still be subject to correction or revision in light of advances in philosophical and empirical understanding, and may be debated as to their respective merits compared against one another: Crucially, if and to the extent that some of them may remain merely abstract with little or no empirical basis, that would obviously imply little or no purpose or merit in the real world and they could be viewed as provisionally 'falsified' in pragmatic terms even if not in terms of their logical functionality.

The foundation of our epistemic pyramid consists of observations (or perhaps that's the ground floor, with logical axioms being the foundation ). Disparate observations - waves, clouds, falling apples, healing wounds, moving planets etc. etc. - are linked by explanations, for which I think my comments in post #23 still serve as an adequate starting point: "Near as I can tell, explaining an observed phenomenon means situating it within a broader, coherent theory with better scope and parsimony than competing alternatives," to which I would now add specificity/depth/predictive capacity as a third important metric for explanatory merit. Obviously not all theories are of the extremely broad, abstract type found at the top of the pyramid, but have a range of different scopes and dependence on empirical observation. Parsimony or simplicity is an important metric of explanatory merit because introducing many new or unknown concepts to an explanation or theory leaves it more in the realm of unsubstantiated speculation; breadth or scope is an important metric because the whole purpose of explanations and theories is to link disparate observations, which an explanation of very narrow scope scarcely accomplishes; and depth, specificity or (especially in the sciences) predictive capacity is an important metric because they increase the potential correspondence between observation and explanation/theory, or conversely, increase the potential options for falsifiability.

Falsifiability itself is not some kind of logical necessity, philosophical abstraction or absolutely-defined criterion; it is simply one of the means by which to differentiate between stronger and weaker explanations and theories, by recognizing and avoiding the scenario in which any circumstances are taken as supporting a given conclusion. Contrary to your post #72 I would not say that absolute "inconsistency" between circumstances and explanation is a required criterion of falsification; that is indeed a desirable criterion, particularly in the hard sciences, but the role/purpose of differentiation between stronger and weaker conclusions doesn't require the potential for them to be completely ruled out and indeed for most of the things which we as individuals accept as true and useful knowledge that potential simply doesn't exist! For one fairly well-known illustration of this point, even the theory of geocentrism is 'consistent' with all observations particularly in light of the relativistic rejection of an absolute frame of reference... but we would hardly dismiss heliocentrism as 'unfalsifiable' because of that! The falsification of geocentrism (and hence potential falsification criterion of heliocentrism) simply consists of which model given all known observations is better overall, and in those terms heliocentrism easily wins out in the metrics of both parsimony/simplicity and the scope of explanation for the motion of other planets.

Hence given all of the above, the obvious, common sense or 'colloquial' fact that many miracle explanations have been falsified is not so difficult to understand: Even in the hard sciences falsification doesn't require absolute proof of impossibility or negation of logical functionality. Somewhat stricter standards may indeed apply in the hard sciences, but in terms of general knowing all that is fundamentally required is the means of differentiating to a lesser or preferably greater extent between better and worse explanations - means which obviously are available in principle and in practice for many if not most alleged miraculous incidents. (And potentially, to a much lesser extent, all miracle theories generally, though I suppose our discussions of materialism and determinism were a bit of a sidebar on the abstract theoretical level.)

[FarWanderer3]

So then, what makes these particular alternative natural explanations better than a divine intervention explanation?

If we observe hidden pipes leading to a statue's eyes, and water being drawn up those pipes before running down the statue's face, we would conclude somewhere in the order of 100% probability of that being the explanation for the 'weeping' statue, leaving very little scope for plausibility in the competing miracle explanation. As I outlined in post #46,
"The sum probabilities of all mutually exclusive, exhaustive possibilities must equal 100%: If we have 90% confidence in one explanation, all other explanations (and the unknown) can only add up to 10%. Hence if 'miracle' had been the strong leading explanation for an event (at say 60% confidence) but then aliens explicitly taking credit and demonstrating how they'd done it led to a 90% confidence in that explanation, it would radically reduce the viability of the 'miracle' explanation for that event. Same goes for finding the hidden wires of a hoax, obviously."

If you're asking why those observations would lead to a near-certainty in that conclusion, beyond simply 'common sense,' I'm thinking that the relevant metrics would be the explanation's parsimony/elegance/simplicity and breadth/scope/comprehensiveness (as suggested in posts #21, 42 etc) and depth/specificity/predictive capacity (#58, 71 etc). By introducing fewer unknowns, a demonstrable explanation in terms of hidden pipes, condensation or the like would have better parsimony than a miracle explanation, and perhaps more importantly would have much greater depth/specificity.

So then you are saying that the more unknowns the more likely the miracle explanation is to be accurate? Isn't that just God of the gaps?

You say that you agree, but you've snipped and failed to address the obvious consequence of that purpose for falsifiability, namely that:
Demonstrated examples of 'colloquial' falsification more than accomplish that; so we can indeed say that some circumstances are far less compatible with a miracle conclusion, meaning that we can validly conclude other circumstances to be far more indicative of a miracle conclusion.

Please do not presume that I also agree with your claim that "Demonstrated examples of 'colloquial' falsification accomplish that".

Also, I made the mistake of not editing your statement to "in terms of general knowing the purpose of falsifiability is to recognize and avoid the scenario in which any circumstances are taken as being consistent with a conclusion".

I'm not presuming you agree - that's why I said you didn't address it. It does seem obvious that in many if not most cases we can indeed avoid the scenario in which any circumstances are taken as supporting a miracle conclusion - which is the epistemic purpose of falsifiability - and hence that we can validly conclude some circumstances to be far more indicative of a miracle conclusion than others. Just as obviously, that's contrary to the position you've been arguing... but now for a third time you haven't answered it.

There's never been anything to answer, because your question has been based on what I consider false premises.

The epistemic purpose of falsifiability is to avoid the scenario in which any circumstances are taken as being consistent with a conclusion.

You don't "falsify" an explanation by failing to support it. There is nothing inconsistent with the existence of pipes and the notion that God put the "tears" in place.

Having read your following post I understand that you disagree. My response is below.

Trying to summarize what I think I've learned from the discussion so far, at the top of the epistemic pyramid we have broad abstract theories such as materialism, deism, theism, naturalism/natural laws and so on which we might legitimately describe as 'unfalsifiable.'

I agree that none of these worldviews are falsifiable (I would not call them theories).

More precisely I would say that none of them seem likely to be completely ruled out in the foreseeable future, although they may still be subject to correction or revision in light of advances in philosophical and empirical understanding, and may be debated as to their respective merits compared against one another: Crucially, if and to the extent that some of them may remain merely abstract with little or no empirical basis, that would obviously imply little or no purpose or merit in the real world and they could be viewed as provisionally 'falsified' in pragmatic terms even if not in terms of their logical functionality.

I would also agree (?) that none of these worldviews even necessarily have any epistemological relevance in as much as they, depending on their respective variants, might not make any predictions that pertain to our lives.

I presume induction is a valid form of constructing knowledge and have been operating under the assumption that everyone else does too. Induction requires the uniformity of nature*. The uniformity of nature is itself simply assumed (which is a variation of being "unfalsifiable") or, I suppose, could be validated by another source (like revelation from a trustworthy God, for example).

*or whatever this empirical reality is that we occupy- don't get too hung up on the word "nature"

The foundation of our epistemic pyramid consists of observations (or perhaps that's the ground floor, with logical axioms being the foundation ).

It sounds to me like you are taking the uniformity of nature (empirical reality) to be a first axiom (assumption), too.

Falsifiability itself is not some kind of logical necessity, philosophical abstraction or absolutely-defined criterion; it is simply one of the means by which to differentiate between stronger and weaker explanations and theories, by recognizing and avoiding the scenario in which any circumstances are taken as supporting a given conclusion. Contrary to your post #72 I would not say that absolute "inconsistency" between circumstances and explanation is a required criterion of falsification; that is indeed a desirable criterion, particularly in the hard sciences, but the role/purpose of differentiation between stronger and weaker conclusions doesn't require the potential for them to be completely ruled out and indeed for most of the things which we as individuals accept as true and useful knowledge that potential simply doesn't exist! For one fairly well-known illustration of this point, even the theory of geocentrism is 'consistent' with all observations particularly in light of the relativistic rejection of an absolute frame of reference... but we would hardly dismiss heliocentrism as 'unfalsifiable' because of that! The falsification of geocentrism (and hence potential falsification criterion of heliocentrism) simply consists of which model given all known observations is better overall, and in those terms heliocentrism easily wins out in the metrics of both parsimony/simplicity and the scope of explanation for the motion of other planets.

Geocentrism is "falsified" because it would require a ridiculous amount of ad hoc reasoning to make it consistent, especially in comparison to the Theory of Gravity. "Miracles" never require ad hoc reasoning to make them consistent.

[Don Mc]

We've got some information - not always especially reliable or detailed - about John the Baptist, about Jesus, his brother James, his followers Cephas, John and James, his executors Annas, Caiaphas and Pilate... and that's about it, as far as I know. I suppose you could throw in Paul and his alleged mentor Gamaliel too, despite no direct connection to the Jesus story. But the other ten of the thirteen apostles are little more than names, a couple of whom are useful props in one or two gospel stories and one of which became wrongly associated with our extant first gospel.

Right, but that seems like quite a bit of background information besides what's already provided in the Gospels, especially given that the Gospels are not meant to be general histories but intentionally condensed (readily reproducible and transmittable) histories of the ministry of Jesus.

The all-important 'Joseph of Arimathea' who supposedly seized custody of Jesus' body and provided the much vaunted tomb appears out of nowhere and disappears just as mysteriously.

You portray Joseph as if he's some shadowy character arbitrarily inserted into the narrative who pilfers the body while no one is looking, but Mark, the shortest and presumably the first Gospel, doesn't really read that way: "Joseph of Arimathea, a respected member of the Council, who was also himself looking for the kingdom of God, took courage and went to Pilate and asked for the body of Jesus. Pilate was surprised to hear that he should have already died. And summoning the centurion, he asked him whether he was already dead. And when he learned from the centurion that he was dead, he granted the corpse to Joseph."

Here Mark includes references to Joseph's hometown, his vocation, his favorable reception of Jesus' ministry, and his deferential interaction with Pontius Pilate to request official custody of the body. Like any historical figure, he appears where he is relevant to the narrative and disappears otherwise. And I don't have to tell you that the Gospel writers were not concerned with writing general history for its own sake, but sacred history or "salvation history." Naturally that means more attention to Jesus and less attention to figures like Joseph.

There's no reason to imagine that 'Matthew's' story about Herod the Great trying to kill baby Jesus is true, nor Luke's story about Herod Antipas' involvement in his trial, and the story of the census looks like total nonsense (though it's hard to imagine why Luke would come up with such a feeble tale).

There's no reason to imagine…other than the fact that the Gospel writers go out of their way to include the accounts. This brings me to what I call the "presumption of error," the hypercritical notion that the biblical narratives must be considered false until proven true. So King David was presumed to be an imaginary figurehead for the tribes of Israel until discovery of the Tel Dan stele; Pilate was considered legendary until someone stumbled upon a piece of limestone in Caesarea with his name and title inscribed upon it; victims of Roman crucifixion were thought too dishonorable for burial until an ossified skeleton was discovered in a tomb near Jerusalem with a nail driven through one of the heels; etc. For critics, extrabiblical information sources (documents, inscriptions, etc.) are valid in themselves, but the biblical narratives have to be supplemented by extrabiblical information to be valid. Thus in any conflict between a biblical source and a secular source, critical scholars too often take it as proof that the Scripture is in error rather than that the question is still open. One begins to wonder how many times archaeological and documentary evidence will have to overturn the presumption of error before critical scholars begin to consider the NT documents valid sources in their own right.

And Luke's account of the census is hardly "total nonsense," a notion that evidently springs from the presumption outlined above. For one thing, archaeologists tell us that Luke tends to get things right, even in cases where he was previously thought to be wrong.

The only direct reference to Jesus himself outside of Christian literature is an obviously tampered with passage in Josephus; his brother James, implied by both Paul and Acts to be the leader of the movement after Jesus' death, seems to have been downplayed by Christians themselves in favour of Cephas, John and Paul, though at least a passing reference to James' death in Josephus provides some solid evidence as to his fate. Even in the case of Cephas and John we're talking about merest scraps of information, often on the level of second- or third-hand rumour: Cephas might have gone to Caesarea, Antioch and eventually Rome, possibly with Mark as an interpreter and eventually being killed there; John perhaps ended up in Ephesus, opposing Cerinthus and writing some or all of the NT's Johannine literature?

This paucity of information isn't a problem unique to Christian origins (though the two wars in Judea may well have resulted in some lost information) it's just the reality of much of history and ancient history in particular.

The key phrase there is "outside of Christian literature." Within Christian literature, Jesus is the central figure by design; everyone else receives far less attention. Even then, Paul's adventures in Acts are recounted in considerable detail. I personally think we have plenty of information, roughly what could be expected of a man highly revered by his followers but thought to be an insignificant religious troublemaker by most everyone else. Whatever fame Jesus may have had among the Romans basically evaporated once he was crucified along with a couple of local shoplifters.

Nonetheless, there are some early non-Christian references to Jesus. Josephus mentions Christ in not one but two passages, the Testimonium and the comparatively uncontroversial reference to James as "the brother of Jesus, who was called Christ." I've never quite understood why the Testimonium has to be an obvious interpolation. Josephus' commitment to the cause against Rome lasted no longer than the revolt itself, and if he could so readily switch allegiances politically it seems plausible that he could entertain the popular religious (that is, Christian) understanding of who Jesus was. Tacitus recounts the death of Christ ("Christus) at the hands of Pilate. Suetonius likewise appears to mention Jesus, as "Chrestus," thought to be responsible for civil unrest in Rome. Jewish rabbis also mention Jesus, albeit infrequently, and one reading in the Babylonian Talmud said to derive from the Tannaitic period has it that "On the day before the Passover they hanged Jesus…" His alleged crime was practicing "sorcery," which suggests a Jewish polemical reaction to Christ's miracles.

As for textual criticism, its conclusions tend to weigh very heavily against trusting anything the gospels say on face value; that's the reason why we can pretty sure that 'Matthew' and Luke were shameless propagandists changing and editing information at will to suit their agendas, and have some reason to suspect that gJohn may be a composite work with several layers of redaction and editing.

Far as I can tell, what you describe are the expectations rather than the conclusions of textual criticism. The remarkable consistency in the overall text resulting from the last several decades of work in textual criticism has surprised even relatively conservative scholars like Bruce Metzger – the seemingly overblown protests of his apprentice Bart Ehrman notwithstanding.

I think you're right about John, though, in that many responsible scholars of various ideological stripes share the view that John was essentially an editor, more so than the other Gospel writers.

Some scholars even point to the chiastic structure of Mark and frequent parallels to stories from the Tanakh to conclude that it too is as historically unreliable as the other three. I'm not entirely persuaded by all those claims myself, but it seems very strange to appeal to textual criticism, form criticism and redaction criticism as a defense for the reliability of the gospel stories!

Well, by reference to those disciplines I don't mean the more radical metaphysical presuppositions (hence conclusions) that so many in the German school in particular have imported into their work. I mean the basic operational assumptions underlying the disciplines. Memorization of short pericopes as part of an oral tradition was common in the first century, for example, in keeping with form criticism. And clearly the Gospel writers were editors (redactors) – as are historians generally! – drawing from various sources both oral and written in creating their narrative.

It all works well enough, I think, if we abandon the simplistic scenario according to which the disciples witness the ministry of Jesus, keep that knowledge mostly to themselves for the next 25-40 years of so, and only then decide to go ahead and publish their accounts. Much evidence suggests that their material was compiled and circulated for many years before the Gospels themselves were "officially" published. To me it seems reasonable to adopt a theory of inspiration in which the Holy Spirit superintended the recording of miracles in the context of actual ancient history using actual ancient historical methods and conventions. On this view ancient historiography is the ordinary, mundane medium through which the miracle accounts are recorded and transmitted, but not so much a miracle in itself.

Do you honestly think that the manner of speech and dialogue presented in the fourth gospel reflects that of a wandering Galilean preacher? Really? The gospel pretty much screams its theological rather than historical intent right from its opening chapter.

I've never been to Galilee myself, let alone met any preachers wandering around there, so I can't say for certain. But I am fairly certain that having a theological intent does not contradict having a historical intent – especially when the theology intended, right from the opening chapter, is that God immersed himself in history ("the Word became flesh") in Jesus Christ. To assert otherwise would appear to be a false dichotomy, not to mention something much like the genetic fallacy: i.e., the history must be faulty because the greater interest was theology. May as well say that published modern histories are suspect because they are written by professors tainted by the profit motive.

The sayings of Jesus in Mark and Q look plausible, and quite consistent with each other, whereas John is not only radically different in style from those but utterly implausible on its own with its long, convoluted monologues and heavily-directed, occasionally nonsensical dialogues. It may be baffling why John would do so, but that he did invent words to put in his mentor's mouth seems quite obvious (except of course on the possibility that it wasn't written by him at all). Given the incongruence and weirdness of supposing that Jesus imparted the Holy Spirit (v22) without noticing that at least one apostle was absent, the 'doubting Thomas' pericope may be a later addition, perhaps (in an extremely speculative vein) intended to downplay/sideline the set of traditions represented in or which eventually produced the Gospel of Thomas.

You're right that John differs greatly from the Synoptics in both style and content. That might seem baffling, but it does not follow from our ignorance of John's experiences or motives that the words he attributed to Jesus were sheer fabrications. Having access to the three extant Synoptic Gospels, he may well have set out to compose a Gospel of his own along different lines altogether, and recalled more obscure messages of Jesus that were not already popularized in the Synoptics and their sources. Maybe he was a prophet, and was trying to preempt twenty-first century intellectuals complaining about a "paucity of information" regarding Jesus.

It should be quite obvious to the readership that Luke removed all reference to the eschatological "abomination of desolation" from that prediction and instead inserted an indefinite 'times of the gentiles.' Conversely it should be obvious to the readership that 'Matthew' added a point of emphasis to the abomination, explicitly directing the reader to Daniel, invented a brand new prophecy for Jesus' imminent return (10:23) and changed the wording of Mark 9:1 to be unambiguously eschatological in Matt. 16:28. But how many Christians even allow themselves to be consciously aware of (let alone actually accepting) those facts? Those two authors obviously were deliberately misrepresenting Jesus; I think you're overestimating how critically people would be reading those documents, as the video I posted says

A video on YouTube says it. I believe it. That settles it!

Seriously, your points are noted, if somewhat irrelevant and exaggerated. I see no reason to think that Luke's "times of the Gentiles" is not eschatological simply because it diverges from Matthew, for instance, any more than I think Matthew's "abomination of desolation" somehow directly contradicts Luke. It may be that Matt and Luke reworded the same original discourse of Jesus, or that Jesus uttered the same basic discourse more than once. Either possibility (and I'm guessing other possibilities exist) would make it considerably less than obvious that they were deliberately distorting the message of Christ. On what grounds is the least charitable reading of a text the most obvious anyway?

Now contrast those seemingly – or at least plausibly – minor and innocuous discrepancies with the prospect of Matt and Luke (along with the other two Evangelists) setting out to deceive readers into thinking that Jesus had "prophesied" the destruction of the Temple when He had not; and how utterly, willingly naive their audience must have been not to notice that a "prediction" of an inescapable and widely known fact only surfaced after that fact.

The appendix to the fourth gospel even goes so far as to suggest that Jesus never really predicted his imminent return at all, that it was all just an off-the-cuff expression of whimsy misunderstood by those dim-witted disciples!

If we're talking about the same text (21:20-25), I think you're taking more liberties than even artistic license would justify. All Jesus suggested there is that John's fate was none of Peter's business. The key word there appears to be the conditional "if" – even if it were the case that Jesus willed it for "that other disciple" to enjoy a full and happy life until the second coming, Peter's obligation to take up his own cross and obey the will of Christ would not be affected. But you wouldn't be the first to read more into Jesus' meaning than what he actually said, per John's brief aside in 21:23.

If Jesus really believed that nationalism, a material focus on God's supposed throne in Jerusalem, and over-zealous messianic aspirations were going to result in the destruction of his people and their culture... do you think he might be willing to get himself deliberately martyred in order to fulfill the prophecies in a different way beforehand, in the hopes of saving at least some of his countrymen from their folly? Something to think about - I'll leave it at that for now

Well, you've left it at quite a bit. While you've raised quite a few objections, it seems to me most of them are stated as facts in passing rather than arguments – that John's portrayal of Jesus is "utterly implausible," for example, or that his account of "Doubting Thomas" means that Jesus simply failed to notice Thomas' absence at the impartation of the Spirit, or that any discrepancy between Matthew's and Luke's versions of the Olivet Discourse entails that both are "shameless propagandists," etc. I will be the first to admit that such a "shotgun" approach is effective, though, as it's difficult to answer such a litany of charges all at a go.

[Don Mc]

Forgive me for the delay in responding to you, FW (assuming you're still following the thread). For a time I was content to let this discussion fade, but then I took another look at your latest rebuttal to me and thought it really deserved a reply.

Feeling a bit disadvantaged, I took the liberty of reading the text in question. Fortunately it was (much!) shorter than I expected. This is really the only part that matters:

there must be a uniform experience against every miraculous event, because otherwise the event wouldn't count as a ‘miracle’. And as a uniform experience amounts to a proof, we have here a direct and full proof against the existence of any miracle.

He's saying any event worthy of being considered as a miracle will necessarily (by definition, basically) have uniform experience against it, and anything with uniform experience against it has proof against it.

Where I think you are getting things wrong is that you seem to think that what qualifies as "uniform experience" actually matters. He provides numerous examples of what he considers "uniform experience" for illustration purposes, which you contest; but it's rather irrelevant to contest them as you do, because the logic of the core argument works by any consistent definition of "uniform experience" (which is actually the whole point). If you want to tear his argument down you'll have to show he's equivocating between two different meanings of "uniform experience" or you'll have to attack elsewhere.

Here's a formal version of the argument. Have at it.

1) Any supposed miracle will have uniform experience against it.
2) Any supposed event with uniform experience against it will have full proof against it.
C) Any supposed miracle will have full proof against it.

Well if that's Hume's argument it's even weaker than I thought. That is, if he's saying in premise 1 that miracles have uniform experience against them because no one has ever witnessed or will ever witness a miracle, then he's simply arguing in a circle.

The logic of the overall argument is valid, though, and if that's what you mean by "the core logic works," then I agree. But even on a charitable reading the truth of the first premise is far from evident,

The first premise is supposed to be true by definition, or at least by a logically equivalent conclusion derived from the definition. It is not a question of whether it is "evident" or not, because its criteria for being true has nothing to do with evidence, just logic and definitions. Miracles are supposed to be unique events, hence the premise.

That may be so, but in principle at least an event is something that actually happens; and there is no reason to think there cannot be experience of something that actually happens – even if it only happens once. So on the definition of a miracle as a unique event, it's still far from clear that the truth of the first premise can be logically derived from that definition.

If Hume is saying only that miracles have uniform experience against them prior to the present miracle claim, on the other hand, then there could be experience of it rather than against it – namely experience of the miracle that is presently claimed.

Prior and after (again, miracles are supposed to be unique events). As for the experience itself, it is what's in question, so to claim it as evidence for itself is circular.

Miracles are unique events, yes, but so are all other events. In historical terms, an event is a singular configuration of various persons, objects and entities in time and place.

I do agree that it would be circular to argue that an experience is evidence of that experience. Yet the testimony of witnesses is not the experience itself but evidence of it (along with documentary evidence, archaeological evidence, etc.). It's the difference between someone simply asserting "X happened," and someone testifying under oath, "I know that X happened because I saw it and I can tell you where it was, who else was there, what they said, what day of the week it was, the time of day…" etc. Now that clearly doesn't mean that the testimony is true. Like any other form of evidence, testimonial evidence only makes a claim more probable than it would have been otherwise. To what degree the evidence makes the claim more probable depends upon things like prior probability of the evidence, explanatory power, and in the case of eyewitness testimony, the perceived credibility of the witnesses.

Consider how the same logic would apply to a scientific discovery (the discovery of Neptune in 1846, for example, or the implication of relativity that time can slow down). By definition the thing discovered had not, like a miracle on Hume's definition, been directly apprehended by anyone prior to the discovery; in other words just prior to the discovery there is uniform experience against it. For reference, let's call your argument above H1, and the argument directly below H2:

i) Any scientific discovery will have uniform experience against it.

Opposite the P1 I provided above, this one is false by definition. No one would call it a scientific discovery in the first place if it weren't something that were continually verifiable.

I see what you're saying, but think of it like this: if there were uniform experience of the phenomenon in question prior to its discovery, Hume's procedure would never have allowed it to be verified at all, because it filters out anything that has uniform experience against it from the outset. And for something to be continually verifiable, it must first be verified. Thus on Hume's procedure nothing new could ever be discovered. To put it another way, any genuine scientific discovery means that Hume was simply wrong about the implications of previous experience for new evidence.