Dianelos Georgoudis wrote:

[me:]
>> X is evidence for Y when P(Y|X) > P(Y). (Of course that could use
>> some refinement.)

[Dianelos:]
> Well, the problem I see with this definition is that one can’t define
> the concept of probability without making use of the concept of
> evidence somewhere down the line, so there is a hidden circularity in
> there.

I don't see any more reason to think that than to think that
one can't define the concept of "feeling confident to use something
as a premise in an argument" without making use of the concept
of evidence. (I don't see much reason to believe either of them;
it probably depends on what you're prepared to consider to be
a workable definition.)

>        It seems to me that a good definition should only use terms
> that are simpler (more likely not to be misunderstood) than the term
> it purports to define.

I almost agree, but it seems to me that "simpler" and "more likely
not to be misunderstood" are quite different notions and that
different reasons for wanting a definition might lead to different
preferences between the two. (Or to some other notion of
prior-ness to take their place.)

It doesn't seem to me that "feel" and "confident" are simpler
notions than "probability". I'm not sure about "premise" and
"argument"; they might be, but it probably depends on how
narrowly one construes "argument".

> [1] A good definition in that sense would be St. Anselm's definition
> of God "God is that than no greatest can be conceived", because it
> defines a concept that is easy to misunderstand (God) using simpler
> concepts, such as "greatest" and "conceive".

Perhaps it's a good definition in *that* sense (at least once we
add the word "which" and correct "greatest" to "greater" :-) ) but
it's a terrible definition in other respects: that vital word
"greater" is left entirely undefined, and indeed Anselm's infamous
ontological argument depends on using it in silly ways. I think
"conceived" is quite problematic, too: for instance, since this
alleged definition doesn't say anything about who's doing the
conceiving we're left to guess whether it means "conceived by
ordinary human beings", "conceived by any human being, no matter
how extraordinarily clever and imaginative", "conceived by any
actually existing being", or "conceived by any possible being,
including God-whatever-that-may-mean". (It should be clear that
that last meaning leads to all kinds of trouble.)

> There may also be a problem with your definition. Consider this
> counterexample: X="concept C refers to a natural number" and
> Y="concept C refers to a prime". P(Y|X) > P(Y), correct? But I don't
> think one can reasonably say that X is evidence for Y.

I would be happy to say that it is. (At least, given a concrete
situation in which, say, you know that someone is thinking of
something but you initially don't know anything about what it is.)
I'd also say that it's very weak evidence, barring special
circumstances that make primes much more likely than other
natural numbers.

> Here's another counterexample: X="we possess libertarian free will"
> and Y="God exists". An atheist may well agree that P(Y|X) > P(Y) but
> disagree with the claim that X is evidence in the first place, never
> mind that X is evidence for Y.

I'd have thought that most atheists would do one of the following:

1. Agree that P(God|LFW) > P(God) and that LFW is likely, but
   not think that this raises P(God) enough to make atheism wrong.

2. Agree that P(God|LFW), but think P(LFW) is small.

3. Deny that LFW is a coherent concept, and therefore not think
   that P(God|LFW) is well enough defined to ask whether it's
   bigger or smaller than P(LFW).

4. Deny that P(God|LFW) > P(God).

And I don't see any reason why those in categories 1 and 2 shouldn't
say that LFW is, or would be if it were real, evidence for God's
existence.

(If your point is just that one doesn't say "X is evidence for Y"
when one doesn't believe in X, then I suppose I agree; so the
kinda-definition I gave would be of "X is, or would be, evidence
for Y", and "X is evidence for Y" becomes something like
"P(Y|X) > P(Y), and P(X) is close to 1".)

> Finally I think the definition of "evidence" should be independent of
> how that evidence is used in an argument.
...
> I think the concept of evidence E entails a) that E us true b) that E
> is used as a premise to argue about the truth of other propositions.

Noted. I think that's an odd combination, for what it's worth.
Anyway: the first of those opinions offers no grounds for choosing
between your definition and mine; the second certainly offers grounds
for preferring yours, but I don't know why you believe it.

> Perhaps a simpler definition is this: Evidence is any piece of
> information that one judges to be evident enough to be assumed as true
> without further justification.

Perhaps it's simpler. It's certainly wrong. Example: One piece of
evidence for the evolutionary account of life on earth is the fact
that when trying to work out when two lineages diverged you tend
to get very similar results (1) by examining fossils and (2) by
looking at "silent" changes in DNA to see how much change there's
been. But this isn't "evident enough to be assumed as true without
further justification", nor anything like it: it's the result of
doing a whole lot of science of multiple different kinds.

Perhaps we're not in any sense trying to define the same term;
perhaps you mean by "evidence" something more like what I'd express
by "axiom". But that seems to me a very strange use of the word
"evidence".

>>>> In our other discussions you have said, contrariwise, that naturalists
>>>> *must* reject anything that is not physical, which you regard as
>>>> absurd on its face, therefore naturalism is wrong, QED.
>> 
>>> I think there is a misunderstanding here. Naturalists do tend to
>>> reject any evidence that is not physical/objective because their
>>> ontology is based on the premise that at bottom all is physical or
>>> explained on physical principles, and hence all evidence that is not
>>> imaginary must be of this sort too. I don’t personally believe that
>>> they must do so; in fact I agree with Chalmers that whether objective
>>> or subjective data is data.
...
> Indeed – once somebody is committed to materialism then concepts such
> as consciousness become meaningless.[2] That's the point I was making
> before: To be consistent a materialist must reject any non-physical
> evidence;

Either you are contradicting yourself or I am misunderstanding you.
You appear to be saying that (1) it isn't true that materialists
must reject anything that is not physical and (2a) anyone committed
to materialism must regard "concepts such as consciousness" as
meaningless and (2b) consistent materialists must "reject any
non-physical evidence".

But here, once again, we're getting into the territory of our
previous discussion which you have chosen to abandon, and for
the reasons already stated multiple times I'm not going to take
this any further.

> [2] At the very least no materialist philosopher has found a way to
> describe consciousness (and the various concepts related to it) within
> materialism. Compare this to how easy and natural it is to describe
> matter (and the various concepts related to it) within idealistic
> theism.

Already thoroughly answered back in 2006.

>> I have no idea how I can interpret those statements, other than
>> as saying that adherents of materialism can't meaningfully talk
>> about (let alone coherently believe or disbelieve) propositions
>> like "I am conscious" or "Bach's music is beautiful" or "Eating
>> people is wrong".
> 
> That's what I meant.

Once again, either you are contradicting yourself or I am
misunderstanding you.

>> The fact that planets appear to move in elliptical orbits is
>> evidence that gravity obeys an inverse-square law. The fact
>> that the world contains as much suffering and evil as it does,
>> distributed as they are, is evidence that Christianity is wrong.
> 
> Actually you must build some arguments in order to justify either
> claim, and these arguments may contain errors or else be weak.

Well, of course. So what? (It doesn't appear to me that the
rest of your paragraph answers that question.)

>> The fact that some people claim to have been miraculously healed
>> at Lourdes is evidence that Christianity is right. All those
>> statements about evidence would remain true even if somehow
>> my brain were coaxed into a state in which none of them seemed
>> true to me. The fact that I can (sincerely, though of course
>> you'll have to take my word for that) affirm the foregoing
>> sentence demonstrates that what I mean by "X is evidence for Y"
>> is not some proposition about my own thoughts or feelings.
> 
> Yes, but surely when you say "X is evidence for Y" I am justified to
> infer that you feel very confident about X being true. That you feel
> very confident about X being true is an important part of why you call
> X "evidence" in the first place.

(I think it's best to consider "X is evidence for Y" as the combination
of two propositions: "X is true" and "X, if true, is evidence for Y".
I've generally been referring to the latter in our discussions above,
because I think that's where the interesting questions lie. My apologies
for any resulting confusion.)

Anyway: my confidence is an important *cause* of my calling X evidence;
it isn't (usually, at least) an important *reason* for me to call X
evidence.

I think this point would be clarified by the separation I mentioned
above; we're really talking here about the "X is true" part. And
I say that when I say "X is true", my own feelings about X form no
part of my *meaning* even though they contribute to *causing* my
statement. The statement "snow is white" is a statement about snow
and light and vision and so forth, not a statement about what's
inside my head. (Unless my head happens to be full of snow, which
to the best of my knowledge it isn't.)

>> I didn't say that the fine-tuning argument is any good, only
>> that the best way to respond to it is unlikely to be "I'm
>> entitled to believe whatever I want to about the laws of
>> physics". Let me give an extreme example: suppose we found
>> a way of measuring the fine structure constant very accurately,
>> and discovered that if you write it in binary then bits
>> 32 onwards spell out "Jesus is Lord" in ASCII. I think
>> that would count as something that atheistic scientific
>> naturalists would do well to have an explanation for.
>> (Fraud on the part of whoever did the experiment would
>> have to be high on the list of candidate explanations,
>> of course.)
> 
> Good example. (Have you read Sagan's "Contact"? there a message is
> found within the expansion of pi in base 11).

Yes, and I found that really stupid. (I think the only way
gods or aliens or whatever could put a long message early
in the digits of pi would be to mess with my brains so that
we see something that isn't really there.)

> I understand naturalist thinkers have tried many ways out of the
> perceived problem of the fine-tuning of the fundamental physical
> constants, such as the possible existence of some kind of ur-
> explanation that binds all constants together, and various kinds of
> models of reality in which universes are continuously brought forth
> having different sets of constants, and so on, but without really
> succeeding to remove the problem - even though the suggested solutions
> themselves are wildly speculative.

It's far from clear to me whether there really is a problem.
Much of the alleged fine-tuning seems to me to be greatly
exaggerated.

If there is a problem, it's certainly a smaller problem than
the Jesus-is-Lord one I proposed above would be :-).

>>        In so far as anyone can be said to be "right" or "wrong"
>> about a point of definition, that is;
> 
> Well I suppose one can reasonably say that a definition is wrong when
> its acceptance would contradict the way that a concept is normally
> used.

Yes (with the proviso that sometimes you might make a technical
definition and say "this isn't quite what everyone's meant by this
before, but it's the nearest thing that makes sense to what they
meant").

>       For example the definition of atheism as the lack of belief in
> God implies that all newborn babies are atheists, or that a brain-dead
> Pope is an atheist - and these statements contradict the way the
> concept is normally used.

That isn't my preferred definition of atheism, but I think someone
who prefers that definition can fairly reasonably say either
(1) "the term 'atheist' can only properly be applied to thinking
beings; newborn babies and brain-dead popes don't count, any more
than rocks do; so those counterexamples are no use" or (2) "the
term 'atheist' has generally only been applied to thinking beings,
and I'm really primarily interested in how it applies to them;
I think my definition works better for them than its rivals because
blah blah blah, and if it gives counterintuitive results for strange
corner cases then I'm happy to say that our intuition is wrong for
those strange corner cases, as intuition often is for corner cases".

(I prefer other definitions of atheism because I think that one
*doesn't* match general usage in the ordinary cases; it's not only
in strange corner cases that it gives counterintuitive results.)

>                            A simple definition which fits well with
> the way a concept is normally used can reasonably be said to be right.

Depends on how well, and whether there's a simpler better-fitting
definition.

>> I suppose what I mean is
>> that (1) "explain" as generally used means something more than
>> "spot patterns in", (2) this stronger usage is, or at least can
>> be tweaked so as to be, both coherent and useful, and (3) it is
>> in fact more useful than the pure-pattern-spotting meaning that
>> you apparently prefer.
> 
> My point is this: All cases of explaining I know about can be
> described as the discovery of a pattern in what is being explained.

So you're making an inference from "X can be described as Y"
to "X *is* Y, and nothing more than Y"? You've explicitly
condemned such inferences elsewhere in this thread.

Anyway. Let's take a concrete example: the explanation of
gravity in terms of spacetime curvature. It seems to me that
any formulation of this explanation that does it justice is
going to involve more than just describing patterns in
our observations of things moving under the influence of
gravity. For instance, it leads immediately to conclusions
that go way beyond such observations: for instance, if you
take your digital watch up a mountain then it will run
faster there.

    (See http://www.leapsecond.com/great2005/ .)

It seems to me that for your account of "explanation" to be
right, you'll need to use so broad a notion of "discovering
a pattern" that just about any act of cognition comes under
that heading; in which case, your definition fails to define.

>> And, actually, it isn't just a point of definition, because
>> the question here is one about what counts as evidence for what,
>> and *that* isn't a merely definitional matter. It seems to me
>> that when A explains X better than B does, that's evidence for A
>> (at least if X is known to be true). I don't know whether that
>> can even be *stated* with a pattern-spotting definition of
>> "explain" (what does it mean for A to explain X better than B
>> does with such a definition?), but it seems to me that something
>> along those lines is plainly correct; if you don't want to use
>> the word "explain" then some other might have to do instead.
> 
> Actually I think you present a good example here. As a case in point
> let’s use X="gravitational phenomena", A="general relativity" and
> B="Newton’s mechanics". The claim that A explains X better than B
> means that A describes more accurate patterns in X than B.

That works fine when X is a broad set of phenomena. But I had
in mind the case where X is a single proposition. "How come
the sky is blue?" "Rayleigh scattering goes like the (-4)th
power of wavelength." That's an explanation, or rather a
very brief sketch of one. The fact that it works is evidence
for the theory that gives rise to it, even without there
being a whole lot of other phenomena for it to "describe
patterns" in.

    (Of course, the fact that the same theory applies
    to a wider range of phenomena is even better evidence,
    which e.g. enables us to distinguish "Rayleigh scattering"
    from "lots of little pixies wearing blue coats" and
    prefer the former. None the less, a theory can explain
    single propositions as well as whole sets of them, and
    I don't see how your account deals with this.)

>                                                            For a
> scientific naturalist, who is committed to the belief that scientific
> models describe objective reality, this evidences that A's respective
> model of reality is more probably true than B's.

And the fact that this sort of reasoning is so effective is
one reason for liking scientific naturalism. (In so far as
such inferences are really specific to scientific naturalism,
which frankly isn't very far; but you do insist on writing
as if they are.)

> One way or the other what's hugely more relevant is how well the
> implications of these postulates work. Also, that reality is similar
> to how we are (namely personal) has a prima facie plausibility that is
> higher than any other model which postulates that reality is
> fundamentally very different from us.

Clearly reality has two arms, two legs, and one head.

>                                       Not to mention the gargantuan
> prima facie implausibility of the various naturalistic models of
> quantum mechanics.

Given the gargantual prima facie implausibility of various
facts about the universe, clearly predicted by QM and verified
about as conclusively as anything can be, I don't think we're
entitled to expect that the underlying facts (whatever they
may be) are prima facie plausible.

>>> incidentally by far the more implausible ontological postulates I have
>>> come across come from naturalists trying to interpret quantum
>>> mechanics.)
>> 
>> Plausibility is, unfortunately, notoriously not something people
>> all agree about :-)
> 
> One more reason not to use this concept to attack the postulates of
> the worldviews one is comparing.

Or (better, I think) to see what common ground we can find
among our various notions of plausibility.

>                                  Perhaps we should then ignore the
> prima facie plausibility of the postulates and proceed to check how
> well the ontological theory based on them works.

Very well. Then my theory is simply this: Everything is exactly
as it happens to be. Every observation we make is an axiom of my
theory; so is every true statement about those observations.
(I'm taking "observations" as broadly as possible here; e.g.,
if as you claim we can Just See that we experience a "phenomenal
consciousness" that is necessarily non-material, then that's
an axiom too.)

"There is no God" is another, extra, axiom.

This theory (by construction) works *perfectly* as an explainer
of experience, at least in your sense: any fact about our experience
is fully described by the theory. I am sure you will immediately
convert to atheism, since the nonexistence of God is also an
implication of the theory. (At which point I'd better convert
back to some variety of theism, in the face of another theory
which is identical to mine except that it has "There is a God"
as its extra axiom instead.)

This is, of course, perfectly preposterous. The reason why it's
preposterous is that the "theory" I've described, if you take
my description as defining it, is almost entirely unknown; and
that any specific candidate for an instantiation of my theory
is vastly complicated and therefore vastly improbable: to be
confident that it was correct we'd need far more evidence than
we could ever really collect.

>>>             The important point though is that any hypothesis, whether
>>> ontological or not, has implications on which that hypothesis can be
>>> judged. If the hypothesis works well then it is justified no matter
>>> how implausible it may appear at first.
>> 
>> Because a sufficient quantity of working-well can generally
>> far outweigh the (bounded and usually not very large) difference
>> in plausibility at the outset.
>> 
>> This depends on there being a good supply of test cases. In
>> empirical science, for instance, that's often true. In questions
>> of philosophy or religion, it's often not, which is one reason
>> why arguments about such topics so often come down to a clash
>> of one person's intuitions against another's.
> 
> Well, reality as we subjectively experience it (which is the greatest
> and most relevant part of the human condition) is impossible to
> measure scientifically but this does not mean that that reality is
> therefore not quite real or not quite important.

You appear to be imagining that I've claimed that things that
can't be measured scientifically aren't real or aren't important.
To the best of my knowledge I have never made any such claim.

>                                                  As for test cases,
> these exist in this realm: According to all great religious traditions
> there are spiritual exercises, or more generally ways of life, which
> transform to a lesser or greater degree the subjective part of the
> human condition.

These do not constitute useful test cases because they lack
discriminating power: some degree of "transformation of the
subjective part of the human condition" might result from
these exercises even if, say, scientific naturalism is correct.

Using such things as test cases is like using "this stone will
fall downwards when I drop it" as a test case for general
relativity: yes, GR predicts that that will happen, but so do
lots of other theories, including Newtonian mechanics.

>                  The empirical sciences fit as well with theism as
> with naturalism (actually they fit better), but theism also explains
> and thus gives us control over the qualitative part of our existence.
> Now one may argue that these test cases for theism are rather vague
> and subjective, but in any case, in comparison, scientific naturalism
> does not enjoy of any test cases whatsoever. At this juncture an
> atheist may suggest that in the future brain science might explain why
> these spiritual exercises are so beneficial, which is indeed probable
> but also beside the point.



>> Please don't imagine that my reluctance to discuss these
>> questions with you at present has anything to do with a lack
>> of interest or engagement in the issues themselves.
> 
> Swell. We have time to discuss these issues in the future then.

Just as soon as you give me some evidence that doing so won't
be as colossal a waste of time as it has been in the past.

>> I adopt something-like-scientific-naturalism as a working
>> hypothesis. It seems to work rather well (but then, I find
>> myself without many of the extraordinary metaphysical intuitions
>> that you seem to think every 10-year-old has; perhaps I'm
>> mentally defective).
> 
> I never said or implied that a 10-year-old has extraordinary
> metaphysical intuitions; I only claimed that a normal 10-year-old uses
> the concept of "consciousness" in a coherent and intelligent matter
> which evidences that a normal 10-year-old understands the meaning of
> that concept.

... which concept you also claimed to carry all sorts of
metaphysical freight. But, again, no more on this.

>>>> Yes, those words appear in "River out of Eden", and so far as
>>>> I can tell Dawkins really means them. But I don't think he means
>>>> by them what you are trying to claim he means.
...
>>> Well it seems to me that the clear implication of what he writes
>>> is that as the universe we observe has the same properties we would
>>> expect if there is at bottom no evil and no good one has therefore
>>> no reason to believe that evil and good do at bottom exist. Why else
>>> would have he written the above?

(I gave several possible reasons he might have had. Apparently
you don't like them. Fair enough, but unless you think the reasons
I give are all so transparently crazy that Dawkins couldn't possibly
have written what he did for any combination of those reasons, it
doesn't matter in the least for present purposes if you think they
have flaws. I'll discuss only one of your specific criticisms.)

>> Perhaps to suggest that attempting to derive our beliefs about
>> what is right from a careful observation of what is actual and
>> some idea that "whatever is, is right" is silly and dangerous.
> 
> Well, at least for a methodological naturalist, there is no way to
> derive any knowledge except by objective observation.

That doesn't mean that methodological naturalists have to believe
"whatever is, is right", and *that* is what I was suggesting Dawkins
might have been arguing against.

>> There seem to me to be any number of possible explanations that
>> have nothing to do with moral nihilism.
>> 
>> And Dawkins does not, on the whole, appear to be a moral nihilist;
>> it seems pretty clear that he finds some things Wrong and others
>> Right. (I don't know whether he's a moral *realist*, which is
>> not the same question.)
> 
> I think it is. A moral nihilist is somebody who believes that nothing
> is objectively wrong; on the other hand a moral nihilist cannot deny
> that moral values exist, as people obviously do hold them, but will
> consider them all inventions.

I suggest that there are two possibly useful distinctions here:
firstly, between those who think moral judgements are statements
of objective fact (or anti-fact) and those who don't; secondly,
between those who make moral judgements at all and those who don't.

It seems to me that "moral nihilist" suggests the second distinction
much more readily than the first; that "moral nonrealist" will do
nicely for the first, but is clearly inappropriate for the second;
and that using the term "moral nihilist" for the first distinction
is liable to give an entirely misleading impression. (I would regard
that as a bug; perhaps for some it might be a feature.)

>>> Well the frequency of "existence of god" in a site that discusses
>>> the existence or non-existence of God is to be expected. The large
>>> frequency of "intelligent design" speaks in favor of my argument,
>>> namely that atheists often focus on red herrings.
>> 
>> The fact that *you* regard that as a red herring doesn't mean
>> that it *is* a red herring.
> 
> Well, of course when one speaks one is expressing one's own thoughts.
> Of course one can always be wrong. So I don't really understand the
> point you are making here.

Gosh, if only I'd done something like, I dunno, not ending the
paragraph there but instead going on to explain what my point
was. Oh look, I did.

>> Lots of people, some of them quite
>> clever, think that "intelligent design" gives reason for believing
>> in God. Many of them would say things like "Every specific claim
>> of Christianity could be false, but intelligent design would
>> still show that there is a Designer".
> 
> I said that in my experience atheists often focus on red herrings.
> Clearly most atheists believe (as I do) that intelligent design (i.e.
> the thesis that the best scientific explanation for the origin of the
> species is not evolution but design by some exterior intelligence) is
> nonsense. Further, intelligent design is not entailed in theism, nor
> in Christianity, so for an atheist to point out the foolishness of
> intelligent design is a red herring. I don’t see how the probable fact
> that some intelligent people think as you describe changes anything in
> that.

That might all make sense if richarddawkins.net had some sort of
charter as a place where all discussion was focused on the single
question: "Is there, or is there not, a god?". Curiously enough,
that isn't the case.

1. Lots of people, some of them quite clever, think that
   "intelligent design" gives reason for believing in God.
   Therefore, if someone -- on richarddawkins.net or elsewhere --
   wants to engage with actual theists rather than with some
   idealized theoretical theist who isn't at all swayed by
   bad arguments, then they might want to say something about
   "intelligent design".

2. Creationists, in some cases funded by people with deep pockets,
   have used the ideas of "intelligent design" as weapons in their
   campaign to destroy science education in US public schools.
   Therefore, if someone -- on richarddawkins.net or elsewhere --
   is concerned that science education not be destroyed, then
   they might want to say something about "intelligent design".

3. Some atheists might think (and in fact I expect some do) that
   "intelligent design", or at least some of the ideas associated
   with that label, provides the nearest thing there is to a good
   reason for believing in God. Such a person, even if trying to
   engage an idealized theoretical theist, might therefore want
   to say somethign about "intelligent design".

For all these reasons, the fact that "intelligent design" gets
a lot of mention on richarddawkins.net need not indicate that
the people there are using red herrings.

>                                                      As for Dawkins's
> site as well as the success of God Delusion I think it evidences how
> little critical thought there is around including in atheistic
> circles.

Is it supposed to be news that lots of people aren't good thinkers?
Or that atheism isn't a guarantee of being a good thinker? Gosh.

>>> Well, atheism is not even a positive ontological position.
>> 
>> Nor is theism.
> 
> Surely theism is. Theism says what is out there.

Theism says very little about what is out there -- there are
vastly many very different things that could be called "God".
(Some of them might not really be rightly called "things",
but let's ignore that technicality for now.)

>                                                  Indeed the common
> ground of all theism (what I call plain vanilla theism) makes a very
> concrete and radical thesis: that reality is based on the existence of
> one personal being who is perfect in all respects.

There are plenty of people who should plainly be categorized
as theists but who do not believe that.

>> Both leave a great many important questions
>> unanswered. I mean, clearly neither theism nor atheism gives
>> you more than about one bit of information, right?
> 
> No, not right, even plain vanilla theism gives you a huge amount of
> information as the implications of its thesis are very far-reaching.

What you call "plain vanilla theism" is not the same thing
as theism; it says a lot more than theism-as-such does. And
it *still* leaves a great deal open.

>> See above: the same complaint can be made against theism;
>> knowing merely that someone is some kind of theist tells you
>> very little about what she believes.
> 
> In almost all cases it tells you quite a bit: that one God exists,
> that God is a person who is perfect in all respects,

I would guess that among theological sophisticates (let's remember
that you consider it a red herring to address the views of anyone
other than theological sophisticates...) probably not more than
half believe that.

>                                                      that God has
> created the environment in which we find ourselves,

I would guess that among theological sophisticates rather a lot
(though I wouldn't care to put a figure to it) would say that
that, if true at all, requires an awful lot of qualifications.

In any case, it certainly isn't entailed by theism as such,
nor even by your so-called "plain vanilla theism" (though
those words "based on" do smuggle in something related).

>                                                     that our life has
> a deeper meaning namely salvation,

*Obviously* not part of theism as such, nor even of "plain
vanilla theism".

>                                    that there are objective ethical
> values,

Obviously not part of theism as such, nor even of "plain vanilla
theism". (It's not even presumed by the definition of the latter;
God might be perfect according to every objective measure, but
goodness might not be among the things that have objective measures.)

>         that we possess libertarian freedom,

Obviously not part of theism as such, nor even of "plain vanilla
theism".

>                                              that life is eternal,

Obviously not part of theism as such, nor even of "plain vanilla
theism".

> that the choices we make carry relevance for the afterlife,

Obviously not part of theism as such, nor even of "plain vanilla
theism".

>                                                             and so on.

Of course I can't comment on whatever items you didn't get round
to listing :-).

> Indeed virtually all theists will agree with the above propositions.

Maybe so, though I think most of those propositions are commonly
disputed by sophisticated theologians. (Whereas unsophisticated
theologians commonly believe things like "the Bible is inerrant"
and "intelligent design is good evidence for the existence of God",
and therefore discussing *their* views would be a Red Herring.)

> Theists disagree about a lot, but not about the fundamentals. The same
> goes for scientific naturalists; virtually all of them will agree with
> the proposition "The whole of reality is based on the existence of an
> impersonal world as described by science." I think the basic
> differences are quite clear between the two worldviews are quite
> clear; what remains is to see which works better.

If you are taking "theism" to include all the propositions you've
listed there, and (I guess) others that enjoy similar levels of
consensus among theists, then I agree that what-you-call-theism
has some actual content to it and might be suitably pitted against
scientific naturalism (or some other such view) rather than mere
atheism.

(But I also think it's perfectly reasonable to put it up against
mere atheism; if it's reasonable to ask whether X is true, it's
never unreasonable to ask whether X is more plausible than not-X.)

-- 
Gareth McCaughan
.sig under construc

Dianelos Georgoudis wrote:

[me, about "that than which no greater can be conceived":]
>> Perhaps it's a good definition in *that* sense (at least once we
>> add the word "which" and correct "greatest" to "greater" :-) ) but
>> it's a terrible definition in other respects: that vital word
>> "greater" is left entirely undefined, and indeed Anselm's infamous
>> ontological argument depends on using it in silly ways. I think
>> "conceived" is quite problematic, too: for instance, since this
>> alleged definition doesn't say anything about who's doing the
>> conceiving we're left to guess whether it means "conceived by
>> ordinary human beings", "conceived by any human being, no matter
>> how extraordinarily clever and imaginative", "conceived by any
>> actually existing being", or "conceived by any possible being,
>> including God-whatever-that-may-mean". (It should be clear that
>> that last meaning leads to all kinds of trouble.)

[Dianelos:]
> Well, I think from the context it's clear that by "greater" St. Anselm
> means "closer to perfection/more valuable/more admirable" and I think
> that as a practical matter people do agree in their judgment of
> personal greatness.

Nope, still hopeless. Socrates, Newton, Gauss, Bach, Shakespeare
(substitute Homer if you prefer and think he was a single person),
Rembrandt, St Francis, the Buddha: I bet you'd get rather little
agreement on how those people should be ranked in "personal greatness",
and that's just within the domain of (1) ordinary human beings and
(2) people whose accomplishments are in some (very broad) sense
of the mind.

For Anselm's alleged definition to work, we need there to be some
universal notion of "greatness" that applies universally -- gods
as well as human beings, and (at least for his use of it in the
infamous ontological argument) imaginary beings as well as real
ones.

>                     As for your second point about who is actually
> doing the conceiving, it is a point well taken and actually shows how
> St. Anselm's definition works especially well: God is not completely
> conceivable by any finite mind, but to any degree a mind is capable in
> this sense St. Anselm's definition is equally valid.

I'm afraid that sentence is not completely conceivable by my
finite mind, and I find myself wondering whether it is in fact
valid. In other words, could you please try to explain it more
clearly? (Unless you're saying *absolutely nothing more* than
in the following sentence of yours.)

>                                                      So it would be
> wrong for a child to understand under "God" a person who is less great
> than what that child's mind can conceive; and it would be equally
> wrong for a great mystic to understand under "God" a person who is
> less great than what that mystic's mind can conceive.

OK, fine; I have no problem (apart from the aforementioned vagueness
of the concepts of "greatness" and "conception") with the idea that
God is by definition at least as great as what any of us can conceive.

>                                                       In other words
> St Anselm's definition says: to the degree of understanding your mind
> is capable of God is a person such you cannot conceive any greater.

But that manifestly isn't a definition.

>>> There may also be a problem with your definition. Consider this
>>> counterexample: X="concept C refers to a natural number" and
>>> Y="concept C refers to a prime". P(Y|X) > P(Y), correct? But I don't
>>> think one can reasonably say that X is evidence for Y.
..
> Well, let me then give another counterexample which may be clearer.
> According to your definition "X is an animal" is evidence for "X is
> male" because indeed the truth of the former increases the probability
> of the truth of the latter. But by the same measure "X is an animal"
> would be evidence that "X is female". That's not the way people use
> the concept of evidence.

If it's not previously known that X is either male or female
then indeed "X is an animal" is evidence for both "X is male"
and "X is female", and I don't think there's any real problem
with this other than the illusory one that comes from the fact
that in most typical situations when one considers the propositions
"X is male" and "X is female", it's already known that one of them
holds and the question is which one. In such a situation, my
definition has the (expected) property that evidence for one
of them is also evidence against the other.

> I understand the idea behind your definition, but it would need more
> work.

Or, as I said myself immediately after stating it: "(Of course that
could use some refinement.)".

>       But, again, the most serious problem is that "probability" is a
> more obscure concept than "evidence".

Says who? (And on what evidence, and with what probability?)

>                                       Indeed, what exactly you mean
> under "probability" here? You are not using "probability" in the sense
> of mathematical statistics? I suppose you mean it in the sense of
> warrant; so that given the truth of X one has more warrant to believe
> in Y.

Actually, I *am* using it in the mathematical sense, though
of course in practice quantifying the probabilities and
degrees-of-evidence we encounter in everyday life is neither
feasible nor a good use of resources. Just as if you asked
me what "red" means I might give an answer in terms of
spectra and wavelengths, even though most of the time we
don't and shouldn't quantify those things when looking at
coloured objects.

>>> Here's another counterexample: X="we possess libertarian free will"
>>> and Y="God exists". An atheist may well agree that P(Y|X) > P(Y) but
>>> disagree with the claim that X is evidence in the first place, never
>>> mind that X is evidence for Y.
>> 
>> I'd have thought that most atheists would do one of the following:
>> 
>> 1. Agree that P(God|LFW) > P(God) and that LFW is likely, but
>>    not think that this raises P(God) enough to make atheism wrong.
[SNIP: three other categories]

> I concern myself with the category 1 atheists. Would you say that they
> would agree that there is evidence for the existence of God, albeit
> evidence which is no stronger than the evidence for atheism?

More or less. Better: no stronger than the evidence against the same
gods as there's evidence for. (There are notions of "God" for which
there isn't, and perhaps couldn't be, any appreciable amount of evidence
either way.)

>                                                              Because
> in this case I wonder what the evidence for atheism is, for I know of
> none. Incidentally, in this context instead of free will one could
> also use moral realism.

Since your definition of "evidence" is more or less "what I feel
like using in an argument", and since so far as I can tell you
think theism is basically just obvious, I'm not surprised that
you "know of none".

I say that something is evidence for atheism if learning it makes
atheism more probable relative to theism. One handy property of
this definition is that if X is evidence for Y, then not-X is
evidence for not-Y. So anything that could have been evidence
for theism but that isn't actually the case is evidence against
theism (though possibly very weak evidence; consider Hempel's
ravens).

In practice, things usually end up being evidence for atheism
by being evidence against particular varieties of theism. So,
e.g., the amount and distribution of evil in the world is strong
evidence against one particularly widely-believed sort of god:
one who cares about our wellbeing (in something like the same
sense as that in which we think of other people doing so) and
has vast power and goodness. And the repeated success of the
reductionist approach in science is (not so strong) evidence
against forms of theism, like yours AIUI, that claim that mind
is more fundamental than matter rather than vice versa.

Some versions of theism are almost perfectly immune to disconfirmation.
For instance, if you say that there is a vastly powerful god whose
principal goal is to prevent our getting really good evidence of his
existence, then it's hard to see how anything could be really good
evidence for or against its existence. I think it's perfectly reasonable
for someone to call himself or herself an atheist while acknowledging
that they don't have strong evidence against such beings. (Just as
someone could call himself or herself a Christian while admitting
that it's possible in principle that Christians are the victims of
systematic deception by an evil god of immense power.)

>> (If your point is just that one doesn't say "X is evidence for Y"
>> when one doesn't believe in X, then I suppose I agree; so the
>> kinda-definition I gave would be of "X is, or would be, evidence
>> for Y", and "X is evidence for Y" becomes something like
>> "P(Y|X) > P(Y), and P(X) is close to 1".)
> 
> Right. But I think a good definition is one that in the simplest terms
> captures what people mean when they use a concept.

It often happens that there's no reasonably simple definition
that does that, because people use the concept inconsistently
or vaguely, or because it just isn't a concept that corresponds
neatly to any chunk of idea-space that's marked out by a simple
definition. (The standard example, since Wittgenstein, is "game".)

And it often happens that the way people use some term is
vague or inconsistent or hard to reason about, but that one
can come up with a definition that comes close to matching
everyday use but that's much cleaner. In such cases, I think
it's reasonable to adopt that definition in contexts that
require hard thinking. (But then one must take care not to
confuse oneself or others by getting confused between the
cleaned-up definition and the subtly different everyday
usage.)

>                                                    So I stand by my
> original definition: by "evidence" people mean a piece of information
> they are very confident is true and which they use to argue for the
> truth of some other belief. Famously, for example, Descartes used the
> fact that he thinks as evidence for his existence.

If you wish to use the word that way, then fair enough. But
then the question of whether there's "evidence for God" and
what various doctrines about God imply on that front (which
is where this discussion began) becomes (1) a question that
only makes sense when one specifies *for whom* -- there's
then no such thing as evidence per se, only evidence for
you, or for me, or for a majority of the Academie Francaise --
and (2) a much less interesting question, since surely it's
more important whether the available information gives reason
for believing in God than whether some particular person
is inclined to use that information as a premise in an
argument.

> Yes, good point. The necessary property of some evidence is that is
> judged to be reliable, but not necessarily self-evident. Scratch then
> my latter definition.

OK. :-)

> Well, my point was that to claim "E is evidence for P" entails that
> both that one believes that E is true and that one believes that the
> argument for P based on E is a good one. But the argument itself may
> be a bad one so that E is not really evidence *for* P even though E is
> still good and valid evidence by itself. Therefore I think it is best
> to define the concept of "evidence" in a way that is independent from
> what one believes E is evidence for. For example many people, theists
> and atheists alike, may well agree that LFW is evidence, but disagree
> about whether it evidences the existence of God.

It seems as if you want to have some notion of being-evidence
with the property that "E is evidence" could be true while "E is
evidence for P" is false for all choices of P. I think that's
crazy.

If not: I think the questions "is E evidence for P?" and "is E
evidence for something?" are separate ones, and I don't see why
anything you've said gives reason to think that the latter is a
good question and the former a bad one. In fact, I think the
former question is much more often interesting or important.

>> Anyway: my confidence is an important *cause* of my calling X evidence;
>> it isn't (usually, at least) an important *reason* for me to call X
>> evidence.
>> 
>> I think this point would be clarified by the separation I mentioned
>> above; we're really talking here about the "X is true" part. And
>> I say that when I say "X is true", my own feelings about X form no
>> part of my *meaning* even though they contribute to *causing* my
>> statement. The statement "snow is white" is a statement about snow
>> and light and vision and so forth, not a statement about what's
>> inside my head. (Unless my head happens to be full of snow, which
>> to the best of my knowledge it isn't.)
> 
> I don’t understand what you mean. How you feel about something causes
> you to say something about it, but has nothing to do with what you
> mean by saying it? Can you elaborate?

I'll try, but I'm not sure I can explain better than I already have.
Please forgive me if I succeed only in being equally unclear but
wordier and more annoying.

Consider the sentence S: "George, you are really stupid".
It is a sentence I'm extremely unlikely ever to utter, because
I'm usually quite polite. I'm probably much more likely to say
such a thing if I'm under stress: let's say I have a pressing
deadline at work, I haven't had enough sleep recently, and I
have a headache.

So, my (hypothetical) headache would be a major contribution to
*causing* my (hypothetical) utterance of S. But "Gareth has a
headache" is no part of the *meaning* of S; what S means is
that some person called George is stupid.

This would still be true even if you could somehow establish
(say, by some kind of colossal brain simulation) that the *only*
circumstance in which I'd utter sentence S would be if I had
a headache.

In other words: we have to distinguish sharply between the
states of affairs that *cause* some utterance, and the ones
that constitute its *meaning*, its criterion for truth.

OK. So now consider the statement W: "Snow is white". What that
sentence *means* is that snow is white. What *causes* it to
be uttered is that I think snow is white. (I'm ignoring all
sorts of corner cases, subtleties, etc. I don't think they're
relevant here.) The two are not the same.

There is no direct causal link from the whiteness of snow
to my utterance of W. Conversely, there is no direct semantic
link from the state of my brain to my utterance of W. (I don't
like that sentence, and worry that it's misleading. I hope I've
explained myself well enough by now for it not to mislead.)
There are *indirect* links of both kinds, and they go via
that correlation between the facts about snow and my beliefs
about snow.

(I cannot resist remarking that perhaps "W" would have been
quite a good name for a sentence about someone called George
being stupid, and indeed "S" would have been fine for one
about snow.)

> Also if the statement "snow is white" is a statement about vision
> surely it's also a statement about what’s inside your head, isn't it?
> At least from the point of view of a naturalist who believes that
> one’s brain produces one’s experience of vision?

Not specifically *my* head, no; not unless I say "snow seems
white to me" rather than "snow is white". I could say "this
apple is green and that one is red", meaningfully and truthfully,
even if I suffer from a form of colour blindness that makes me
not see a difference in colour between the two apples. When I
say "snow is white" I am saying something about vision *generally*,
which involves facts about what happens inside people's heads
*generally*. I am not saying anything specifically about my
head, even though the causal path from the whiteness of snow
to my saying "snow is white" involves the insides of my head
in ways it doesn't involve the insides of other people's.

> Well I think Sagan considered the value of pi to be contingent on the
> structure of our physical universe, and if someone has actually
> created that structure then presumably they would also have created
> the value of pi. Now I think I understand you sentiment in the sense
> that math not contingent and therefore stays the same in all possible
> worlds, but I wonder if that's a warranted belief. Why can't there be
> worlds with different rules of logic (and hence with radically
> different math) than our own?  It's not a particularly relevant
> question, but I wonder what you think about this.

I think it's entirely possible that there are worlds in which
any thinkers would end up with mathematics of a very different
sort from ours. I'm not sure what it would mean for there to
be worlds with different logic from ours, but let's suppose
for the sake of argument that there could be such worlds.

Even so: I can tell you what pi is quite concisely, in ways
that seem to me to have very little "flexibility" in them of
the sort that would be needed for even a super-omnipotent god
who can change the laws of logic to be able to make there be
a world in which pi had some specified lengthy string of digits
in it. For instance: all solutions of the differential equation
y'' = -y have a certain period; pi is half that period. For
instance: choose a very large positive integer N and consider
all pairs (a,b) of positive integers from 1 to N; then the
fraction of them that have no common factor is approximately
6/pi^2, and the approximation gets better as N increases.
For instance, pi = 4 - 4/3  4/5 - 4/7  4/9 - ... .

Now, surely an omnipotent god, even without the ability to
mess with logic, could mess with the geometry of spacetime
so that an attempt to measure pi by drawing large circles and
measuring their area would give some particular value of
that god's choice. But that isn't how one actually calculates
pi; I don't actually remember how the calculation in "Contact"
was supposed to be done, but I bet it didn't involve having
the machine actually draw lots of circles and measure them
really accurately.

>> If there is a problem, it's certainly a smaller problem than
>> the Jesus-is-Lord one I proposed above would be :-).
> 
> If God exists then I don't see any particular reason why God would not
> choose to incarnate as a human, and if God did incarnate as a human
> then such a human would not live unlike Jesus of Nazareth lived. So
> once one establishes the truth of theism why do you think Jesus-is-
> Lord belief is especially problematic for theism?

You're confused, sorry. I was talking about the bizarre hypothetical
situation in which we find "Jesus is Lord" encoded in the binary
expansion of the fine structure constant. That would (in that
bizarre hypothetical situation) be a problem for materialism and
atheism. I wasn't talking about any problem (1) for theism, or
(2) in the real world.

>>> My point is this: All cases of explaining I know about can be
>>> described as the discovery of a pattern in what is being explained.
..
>> Anyway. Let's take a concrete example: the explanation of
>> gravity in terms of spacetime curvature. It seems to me that
>> any formulation of this explanation that does it justice is
>> going to involve more than just describing patterns in
>> our observations of things moving under the influence of
>> gravity. For instance, it leads immediately to conclusions
>> that go way beyond such observations: for instance, if you
>> take your digital watch up a mountain then it will run
>> faster there.
> 
> The discovery of a pattern allows us to make predictions and these
> sometimes are surprising.
> 
> General relativity's equations describe a deep pattern present in the
> gravitational phenomena we observe. We find it practical to work with
> these equations by visualizing a curved spacetime, but the equations
> do not really imply that there is such a thing as spacetime, never
> mind that it is curved. For example should we exist within a computer
> simulation then general relativity's explanatory power of the
> gravitational phenomena we observe will remain exactly as true/
> relevant/useful as it is now, even at the absence of curved spacetime.

If we exist within a computer simulation that has the properties
predicted by general relativity, then there *is* curved spacetime,
inside the simulation; and those properties predicted by general
relativity are not merely facts about what observers inside the
simulation see, they are structural facts about what the simulation
does.

(Of course it is possible e.g. that the simulation includes special-case
code to deceive intelligent observers within it about how the world
works; perhaps when things in the simulation aren't observed they
obey Newtonian mechanics, but when observed they are somehow forced
to match up with general relativity. I think it's reasonable to
ignore such silly possibilities; not because they're impossible but
because there is no end to all the silly things that *could* be
true, because we couldn't possibly have good evidence for or
against them, and because empirically it's never been much use
to take such possibilities seriously.)

> I understand that some cosmological theories use complex numbers to
> represent time, but surely nobody thinks that therefore time has a
> real and an imaginary component.

I don't think the machinery of "imaginary time" ever involves
giving time both real and imaginary components simultaneously.
And I think the question "is the time coordinate real or
imaginary?" is a non-question.

>> It seems to me that for your account of "explanation" to be
>> right, you'll need to use so broad a notion of "discovering
>> a pattern" that just about any act of cognition comes under
>> that heading; in which case, your definition fails to define.
> 
> Well, I do use the concept of "explaining" in a particularly broad
> sense. For example when one sees a tree outside of one's window then,
> in my use of the concept, one is explaining one's raw visual
> information by discovering a tree-pattern in it.

So far you have not managed to communicate to me what acts of
cognition you class as "discovering a pattern" or as "explaining".

>>>> And, actually, it isn't just a point of definition, because
>>>> the question here is one about what counts as evidence for what,
>>>> and *that* isn't a merely definitional matter. It seems to me
>>>> that when A explains X better than B does, that's evidence for A
>>>> (at least if X is known to be true). I don't know whether that
>>>> can even be *stated* with a pattern-spotting definition of
>>>> "explain" (what does it mean for A to explain X better than B
>>>> does with such a definition?), but it seems to me that something
>>>> along those lines is plainly correct; if you don't want to use
>>>> the word "explain" then some other might have to do instead.
..
>> That works fine when X is a broad set of phenomena. But I had
>> in mind the case where X is a single proposition. "How come
>> the sky is blue?" "Rayleigh scattering goes like the (-4)th
>> power of wavelength." That's an explanation, or rather a
>> very brief sketch of one. The fact that it works is evidence
>> for the theory that gives rise to it, even without there
>> being a whole lot of other phenomena for it to "describe
>> patterns" in.
> 
> We use semantic shortcuts when explaining things, but at bottom they
> all refer to patterns. The explanation of a particular phenomenon may
> be (and often is) only a sub-pattern within a broader one, but it is
> still a pattern. "Rayleigh scattering goes like the (-4)th power of
> wavelength" does describe a pattern after all.

Sure, it does, but that isn't the point. The point is that
the datum I'm claiming as evidence for the models of physics
that include Rayleigh scattering needn't itself have any
"patterns" worth mentioning at all in it, again unless you
define "pattern" so broadly that your statements become
content-free.

I think that in fact you *do* define "pattern" so broadly;
the result is that it seems to me that you've not really given
any account of what you mean by explanation. To explain something
is to describe a pattern it fits with? But everything is a pattern
and every act of cognition is pattern-spotting; therefore everything
we do is an explanation of everything. I'm sure you don't really
think *that*, but I haven't yet been able to find out how what
you really think differs from it.

>>>                                                            For a
>>> scientific naturalist, who is committed to the belief that scientific
>>> models describe objective reality, this evidences that A's respective
>>> model of reality is more probably true than B's.
>> 
>> And the fact that this sort of reasoning is so effective is
>> one reason for liking scientific naturalism.
> 
> No, I think you are making a conceptual jump here. Once one is
> committed to scientific naturalism then it is of course reasonable to
> believe that a more powerful scientific model describes reality better
> than a less powerful model. But the fact that there exist scientific
> models that more or less well describe the physical phenomena we
> observe does not in any way evidence that scientific naturalism is
> true. This is a basic point. Do you agree with it?

Nope.

(Of course it's not *conclusive* evidence. I don't think it's
even very strong evidence. But evidence, I think, it certainly
is.)

>                                                    Again, should we
> live within a computer simulation, or should there be an evil
> Cartesian demon deceiving us, scientific naturalism would be false but
> nothing would change with scientific knowledge or with the way we
> conduct science.

Except that at any point the Lords of the Matrix, or the Cartesian
Demon, might decide to mess with the input they're feeding us and
all our science would suddenly go terribly awry. If they never
choose to do that, then in fact our world does have structural
features that are more or less isomorphic to the ones science
describes. They might in fact be implemented in computer hardware
or in the mind of the Cartesian Demon, but they would be no less
real for all that.

>>> One way or the other what's hugely more relevant is how well the
>>> implications of these postulates work. Also, that reality is similar
>>> to how we are (namely personal) has a prima facie plausibility that is
>>> higher than any other model which postulates that reality is
>>> fundamentally very different from us.
>> 
>> Clearly reality has two arms, two legs, and one head.
> 
> Not really, for these are neither necessary nor relevant properties of
> what it is to be a person.

"Reality is not fundamentally very different from us" (which
might perhaps have some prima facie plausibility, though I
rather doubt it) has now metamorphosed into "Reality is not
fundamentally very different from those aspects of us which
Dianelos chooses to associate with the word 'person'" (which
I think has much less prima facie plausibility, e.g. on account
of being a much more complicated statement).

Furthermore, the only persons we are definitely acquainted with
have all sorts of features in common that it is not at all
reasonable to assume Reality has too. (For instance, they are
living things made of meat, based on DNA and proteins; they
do their thinking using lumps of tissue of size comparable
to a human head; they have emotions that are influenced by
chemical changes in the fluids flowing through those lumps
of tissue; they experience anger and lust; they typically
have a lifetime of between one and one hundred years.)

I think it's perfectly ridiculous to expect that Reality
will not be "fundamentally very different from us". (That's
assuming arguendo that the statement even makes sense,
which I think doubtful.)

>>>                                       Not to mention the gargantuan
>>> prima facie implausibility of the various naturalistic models of
>>> quantum mechanics.
>> 
>> Given the gargantual prima facie implausibility of various
>> facts about the universe, clearly predicted by QM and verified
>> about as conclusively as anything can be, I don't think we're
>> entitled to expect that the underlying facts (whatever they
>> may be) are prima facie plausible.
> 
> QM does not predict any fact whatsoever about the universe; it only
> makes predictions about the quantum mechanical phenomena we observe.

True only in the following vacuous sense: everything we know
about the universe we know from our observations, and all our
theories can if we choose be stated in terms of what we will
observe in given circumstances.

> Whether these phenomena are produced by the will of God or perhaps by
> some kind of mechanical substratum or perhaps by XYZ ontological view
> is an entirely different question.

Don't forget Cartesian demons. And pixies.

>>>                                  Perhaps we should then ignore the
>>> prima facie plausibility of the postulates and proceed to check how
>>> well the ontological theory based on them works.
..
>> This is, of course, perfectly preposterous. The reason why it's
>> preposterous is that the "theory" I've described, if you take
>> my description as defining it, is almost entirely unknown; and
>> that any specific candidate for an instantiation of my theory
>> is vastly complicated and therefore vastly improbable: to be
>> confident that it was correct we'd need far more evidence than
>> we could ever really collect.
> 
> Right. Any candidate for your hypothesis is vastly complicated,
> because your theory does not claim any patterns, and therefore does
> not explain anything.

On the contrary: for every pattern that in fact exists in our
observations, "my hypothesis" includes a statement of that pattern
as an axiom. (As I said: "so [sc. an axiom] is every true statement
about those observations".)

So my (absurd) theory does, in fact, claim every pattern that
exists; and no others.

> Anyway I do wish you'd explain what your ontological beliefs are,
> instead of explaining what they aren't.

Why should I? You've responded to all my attempts to get you
to be more specific about *your* beliefs with smoke-blowing
and bluster. I think the ball's in your court at this point.

But, briefly: I think many ontological questions -- are there
*really* such things as numbers, sets, minds, opinions, propositions,
preferences, emotions, etc.? -- are pseudo-questions; radically
different ontologies can be equivalent in practice. So questions
like "Does God exist?" can be obfuscated ad lib by saying things
like "Existence is the wrong thing to predicate, or not predicate,
of God; rather, he is the essential precondition of all existence".
Or indeed "Of course God exists; 'God' is simply the name we give
to our highest thoughts and aspirations.". If we are somehow able
to forget about this sort of philosophical quibbling and focus on
the non-pseudo questions that are commonly meant by things like
"Does God exist?", I am cautiously in favour of the material things
that we (seem to) observe in everyday life and study scientifically,
and against gods and demons and angels and spirits and other such
irreducibly mental things. 

I don't have a thoroughly filled-in philosophical position that
takes definite positions on all the quibbly questions. I think
that a better way to characterize what you've described as a
crisis in materialist thinking is as follows: Many people have
hoped to find a simple, clear-cut philosophical position that
makes all the difficulties of subjectivity go away. Unfortunately,
no one has succeeded; the fact that all our thinking is *our*
thinking and is done with the imperfect equipment of our brains
and our language is too big an obstacle. So I don't *expect* to
find, or for anyone else to find, any way of looking at the
issues that just makes them disappear. The only reason why this
might look like a problem *for materialism* is that materialism
has been pretty much the only position with the slightest prospect
of being clearly enough defined for it to work. It's not that
other positions fare any better; it's that their holders had
less ambition.

>> You appear to be imagining that I've claimed that things that
>> can't be measured scientifically aren't real or aren't important.
>> To the best of my knowledge I have never made any such claim.
> 
> Well, you do not really say what you believe about reality Gareth.
> Perhaps you are basically an agnostic.

I find the widespread bickering about the terms "atheist"
and "agnostic" very boring and largely ill-conceived. I'm
atheist in the sense that I see no reason to believe in
the existence of any god, and much reason to disbelieve
in the existence of the most widely-believed-in ones.
I'm agnostic in the sense that I don't think that even
such specific things as evangelical Christianity can be
refuted outright, and in the sense that I think there
are (kinda-sorta) theistic claims for which no one is
ever likely to have good evidence either for or against.
I'm atheist in the sense that if you pick any particular
notion of god and ask me to assign it a probability,
then the probabilities I'd assign would all be very
low. I'm agnostic in the sense that none of them would
be zero.

And, stepping aside from the specific question(s) of the
existence of gods, I am agnostic about plenty of things
about the universe. I don't know whether there is intelligent
life elsewhere in it. I don't know what a unification of
general relativity and quantum mechanics will end up looking
like. I don't know how life on earth began. I don't know
whether there will be superhuman-level AI within a century.
And so on.

>>>                                                  As for test cases,
>>> these exist in this realm: According to all great religious traditions
>>> there are spiritual exercises, or more generally ways of life, which
>>> transform to a lesser or greater degree the subjective part of the
>>> human condition.
>> 
>> These do not constitute useful test cases because they lack
>> discriminating power: some degree of "transformation of the
>> subjective part of the human condition" might result from
>> these exercises even if, say, scientific naturalism is correct.
..
> I see your point. But theism makes specific claims about our
> subjective experience,

I do not believe that Dianelosism makes claims about our
subjective experience that are specific enough to be testable.

>                        whereas right now scientific naturalism can't
> even describe subjective experience in naturalistic terms, let alone
> make coherent claims about it.

As I've said before, I think you are using a double standard
here; if those difficulties appear you to be real ones for
scientific naturalism and not for theistic idealism, I think
it's only because you demand more of the former than of the
latter. And you have yet to make it clear to me what you
mean by "subjective experience" and "consciousness" and all
those other grand terms you use.

    (I shouldn't have to say this, but: Kindly refrain from
    saying things like "Gareth pretends not to know what
    consciousness is", because I have said nothing even
    slightly resembling that.)

>>>> And Dawkins does not, on the whole, appear to be a moral nihilist;
>>>> it seems pretty clear that he finds some things Wrong and others
>>>> Right. (I don't know whether he's a moral *realist*, which is
>>>> not the same question.)
>> 
>>> I think it is. A moral nihilist is somebody who believes that nothing
>>> is objectively wrong; on the other hand a moral nihilist cannot deny
>>> that moral values exist, as people obviously do hold them, but will
>>> consider them all inventions.
>> 
>> I suggest that there are two possibly useful distinctions here:
>> firstly, between those who think moral judgements are statements
>> of objective fact (or anti-fact) and those who don't; secondly,
>> between those who make moral judgements at all and those who don't.
>> 
>> It seems to me that "moral nihilist" suggests the second distinction
>> much more readily than the first; that "moral nonrealist" will do
>> nicely for the first, but is clearly inappropriate for the second;
>> and that using the term "moral nihilist" for the first distinction
>> is liable to give an entirely misleading impression. (I would regard
>> that as a bug; perhaps for some it might be a feature.)
> 
> Yes, well, it seems to me that as a matter of psychological fact a
> normal human being cannot possibly avoid making moral judgments.

I suggest that a normal human being who adopts (any position
that deserves the name of) moral nihilism would, at any rate,
not advertise their moral judgements and admit openly that
they make them.

>> Theism says very little about what is out there -- there are
>> vastly many very different things that could be called "God".
>> (Some of them might not really be rightly called "things",
>> but let's ignore that technicality for now.)
> 
> That there is a lot of disagreement about the details does not imply
> that theism says very little about what is out there. There is much
> disagreement between scientific naturalists also (see the various
> interpretations of QM), but scientific naturalism is a positive
> ontological position too.

In so far as there is much disagreement (about ontology) between
scientific naturalists, scientific naturalism is not a positive
ontological position but a family of positive ontological positions.
The same goes for theism.

But the question isn't really "should we call this one position
or a family of positions?", which is only a question about language.
It's "how much would knowing that X is correct tell you about the
world?".

Theism-as-such leaves an enormous amount open.

I suspect that we are merely arguing about what counts as "very
little". Let's continue this if, and only if, you think there's
some actual point at issue that matters; if so, please say what.

>>>                                                  Indeed the common
>>> ground of all theism (what I call plain vanilla theism) makes a very
>>> concrete and radical thesis: that reality is based on the existence of
>>> one personal being who is perfect in all respects.
>> 
>> There are plenty of people who should plainly be categorized
>> as theists but who do not believe that.
> 
> Who for example?

Well, for instance, Mormons. (Who claim, AIUI, that God is part
of the natural order rather than "outside" or "beneath" it; and
that he "was once as we are now".)

>>> In almost all cases it tells you quite a bit: that one God exists,
>>> that God is a person who is perfect in all respects,
>> 
>> I would guess that among theological sophisticates (let's remember
>> that you consider it a red herring to address the views of anyone
>> other than theological sophisticates...) probably not more than
>> half believe that.
> 
> There must be some misunderstanding here. I don't think you will find
> one grown-up theist, no matter how unsophisticated, who will agree
> that there is more than one God, or that God is not a personal being,
> or that God is less than perfect in some respects.

You have my point exactly backwards. I suggest that the *more*
theologically sophisticated a theist is, the *less* likely they
are to agree with your definition. For instance: Christians
who have given a lot of thought to the Trinity will be very
much disinclined to say that God is "a person". Those following
in the footsteps of Tillich, or the earlier apophatic theologians,
will want to quibble with "one God exists", on the grounds that
"exists" gives the too-simple impression that God is just one
more thing alongside all the other things that exist. Process
theologians will be inclined to say that God becomes more perfect
with the passage of time; they may perhaps say that this gradual
increase in perfection is itself perfection on a higher level,
but I don't think they all do and in any case I think it's a
fudge.

[SNIP: a whole lot of things Dianelos says are implied by saying
that God exists, but which seem to me not to be implied either by
that or by the much more specific claim that Dianelos has called
"plain vanilla theism".]

> Well I disagree and would happily describe what I think the existence
> of a perfect personal being implies – but before I do that, what would
> *you* say does it imply? Imagine for a moment a world in which the
> only thing that exists is a personal being who is perfect in all
> respects. What do you think would such a person do? I am curious about
> your answer, because I think the implications of the core theistic
> hypothesis are independent of one’s ontological assumptions.

I don't know what such a being would do, not least because I don't
think the phrase "perfect personal being" has much meaning -- for
all the reasons we've discussed already about the ambiguity of
"perfect", etc. I rather doubt that anything much is strictly
*implied* by the existence of such a being.

And since even in my most arrogant moods I wouldn't claim to
have anything like infinite intelligence or wisdom, how could
I possibly make confident predictions of what such a being would
do?

I can make some guesses, with various degrees of confidence.
It seems pretty plausible that a perfect being might create
things. It seems likely that it would create *good* things.
If it created persons, I'd expect it to treat them with at
least as much decency as they might reasonably be expected
to treat one another. It would probably have an infinitely
rich "inner life"; just looking at its interactions with its
creation(s), if any, might be uninformative.

It's much easier to come up with things it probably wouldn't
do. (If you claim to have a machine that plays perfect chess,
then I can tell you all sorts of things I wouldn't expect it
to do, but I wouldn't have a hope of answering 'Suppose it
plays a game against itself; what do you expect the moves
to be?'.)

> I agree. But it may be the case that it is not *practical* to think
> about whether X is more reasonable than not-X. Let me try to find an
> analogy.
[SNIP: To investigate whether the planets are pushed around
by invisible angels, look for a specific alternative explanation
of their motion.]

Yeah, sure. But now imagine that we're in, say, Kepler's time.
We know that the planets appear to move in ellipses; we don't
yet have Newtonian physics; we don't have any specific candidate
explanation. None the less, I think we should say "The planets
are probably not moving the way they do on account of being
pushed around by invisible angels". Do you agree or not?

> Similarly then in order to show that core theism is more reasonable
> than atheism one has to instantiate a positive atheistic ontology and
> compare core theism with it.

You have given no actual reason to believe that; just one
not particularly close parallel where one might argue that
the corresponding proposition is true.

Consider "core theism" plus the proposition that God will
make every conscious being other than himself spend eternity
in torment in hell. I claim that this doctrine is less
reasonable than its negation, because it has such serious
internal problems. (Namely, at least, an allegedly perfect
being behaving in an apparently odious way.)

So on what grounds are you so sure that "core theism" itself
needs to be compared with a specific variety of atheism to
evaluate it?

-- 
Gareth McCaughan
sig under construc