Jason Pratt on Lewis

This is by frequent commenter Jason Pratt, on what I am missing in Lewis's argument from reason.

....... C. S. Lewis' (most?) Dangerous Idea.......

When Professor Victor Reppert wrote his detailed analysis of variants of
the Argument from Reason (especially in connection to the work of C. S.
Lewis), a few years ago, I opined more-or-less steadily through various
drafts and into the finished copy (and beyond in occasional comments since
then), that despite his book's title (_C. S. Lewis' Dangerous Idea_, wryly
riffing on the title of Daniel Dennett's _Darwin's Dangerous Idea_) Victor
was downplaying the most "dangerous" application of Lewis' argument: the
application Lewis himself culminates his chapter with.

I recently mentioned this again in a private critique of some potential
encyclopedia entries; and while collating discussion on the theistic
Argument from Reason, Victor has publicly opened the debate here on his
second DangIdea site. (To which initial comments Victor will probably link
this letter for prior reference, below.)


The discussion between us, is on the value of the portion of Lewis' 3rd
chapter from the 2nd edition of _Miracles: A Preliminary Study_, which (for
quick reference sake) begins "'But,' it will be said, 'it is incontestable
that we do in fact reach truths by inferences.'", and which (for our
purposes) ends at roughly the place where Lewis writes, "On these terms the
Theist's position must be a chimera nearly as outrageous as the
Naturalist's." Victor has posted a bit more than half of this in his
initial remarks (opening the public debate). Of course, none of us should
be reading and analyzing this apart from the progression and development of
Lewis' argument up to the beginning of that climactic section; so for
reference sake I encourage readers unfamiliar with the work to please refer
(if a copy isn't handy) to the page Victor linked several days earlier,
containing a text of MaPS Chp 3 (2nd edition). This page can be found at

http://www.philosophy.uncc.edu/mleldrid/Intro/csl3.html (That's CSL 3, not
CS thirteen, for readers with some kinds of fonting...)

As usual, I will caution that there is a prevalent tendency to conflate
between the category of atheism/theism and the category of
naturalism/supernaturalism. While I find Lewis was actually better about
avoiding category error switches on this topic than the average (especially
in MaPS 2nd ed), his presentation is still worded with this conflation in
view, and there are times when (in acceding to the popular use of the term
Naturalism) he verges hard on category error. Between this, and the fact
that most subsequent commentators are in the habit of conflating the
categories (not infrequently in convenience to whatever argument they're
trying to make), figuring out what Lewis was primarily attempting in this
chapter can become (even?) more difficult.


In opening his recent public comments on the topic, Victor contended, or
seemed to contend, that Lewis was disagreeing with Darwinians that
evolutionary development could lead to effective mental behaviors in
relation to true facts. Lewis does have something to say about the
difficulty of evolutionarily developed expectations being proposed as
leading to a power to recognize truths; and this can be found beginning at
the very top of page 30 (paginated according to the hyperlinked text
referenced above) down to roughly the first sentence completion at the top
of page 31 (with some commentary and explication afterward leading into the
portion I have identified as the climactic summary; this can be found
beginning at the bottom of page 32.)

However: if we go back to the beginning of the first full paragraph on page
28, down to the bottom of page 29, we will find Lewis nevertheless
_agreeing_ with the notion that evolutionary processes could plausibly
result in the development of effective mental reactions (the phrase
"responses to stimuli" is how Lewis puts it) which, in the consequent
accuracy of connecting behaviors to true facts of reality, Lewis admits,
"might serve us as well as reason or in some circumstances better."

I contend that in making this concession, Lewis is actually agreeing with
the sort of notion represented by Victor's quotation of Antony Flew; not
disagreeing with it. (Not so far as that quote itself goes anyway, aside
from whatever Flew does with the notion elsewhere.)

Victor finds the portion of Chp 3 which I call the climactic summary, to be
dissatisfactory as a way of ending-and-illustrating Lewis' disagreement
with (specifically) Darwinian claims of effective process development. I
agree that this portion would in fact be dissatisfactory as such an
ending-and-illustration; _if_ Lewis intended for it to be used that way.

I believe he did not intend for the climactic summary to be used that way.
He is not primarily aiming at critiquing this or that particular Darwinian
explanation for the rise of effective epistemic behavior. He isn't aiming
at various fish; but for Leviathan itself.

(In passing, let me add that I may have misread Victor originally to be
trying to say that one part of this climactic portion was Lewis presenting
"the Darwinian objection to his argument from reason", instead of pointing
to a different subsequent portion where this objection would be better
located instead, which Victor did go on to give after some interspersed
commentary. In retrospect, I expect that when Victor wrote "Here Lewis
presents the Darwinian objection..." he was thinking in terms of what he
would quote from Lewis _next_, after a commentary digression, instead of
what he _had just quoted_. If so, I readily retract an incidental criticism
I made in a preliminary comment, with apologies.)


My contention is that in MaPS chapter 3, Lewis is carefully building up to
a tactical deduction of atheism out of the pool of philosophical options.
Whether or not he succeeds or fails at this, _this_ is his "dangerous
idea". He is going for the heart; not against various fingers and toes or
even arms and legs. He intends this argument to be a deadly (if polite)
threat to atheism across the board, at the very root of any person's ground
for choosing to believe atheism or not.


In order to trace and demonstrate this most fully, I would (of course) be
required to analyze the progression of Lewis' argument point for point;
something that would require far more length from me than chapter 3 itself!
As a quick-n-dirty illustration, though, I can combine (a) Lewis' treatment
of the word "Naturalism" in chp 3, in topical connection with (b) his
treatment of the words 'rational' and 'non/not/sub-rational' (which
not-incidentally begin on page 28), in conjunction with (c) his original
title for the chapter ("The Self-Refutation of the Naturalist", or
"Self-Contradiction", my sources differ--either of which in itself ought to
be sufficient evidence for his original intention at least!); along with
(d) Lewis' removal of the term 'irrational' from his original 1st edition
of the chapter, substituting for it in the 2nd edition the _stronger_
negative terms mentioned at (b). (Terms insisted upon by Anscombe and,
ironically, Flew.)

Lewis did not gear down chapter 3, after his loss to Anscombe, his more
cautious title notwithstanding. He took Anscombe's critiques, incorporated
them, revised his argument, and _ramped up_ at least the apparent (I would
say actual) strength of his argument thereby.


Leaving aside even a fast summary of _how_ Lewis is getting there, his
final aim may (I believe) be accurately described thus: Lewis intends to
demonstrate that the claim of atheism sooner or later _requires_ the
atheist to have to do one or the other of two logically illegitimate
moves--justify that our justification abilities can be possibly accurate,
or justify that our mental behaviors can be possibly accurate even if we
don't (in fact) have justification abilities. Ultimately there is no other
option (once the topic turns to epistemology) under rigorous atheism. We
can see that non-rational causation frequently produces non-rational
effects, and atheism requires that our own apparently rational behaviors
must be ultimately produced and maintained by non-rational behaviors.

In order to fully appreciate the problem, rational behavior must be
discussed in context of intentional actions while avoiding various pitfalls
(e.g. the externalistic fallacy), in comparison to and distinction from
non-intentive purely automatic reactive behaviors.

At which point, two basic paths emerge: either true action capability (and
not merely the illusion thereof) is produced somehow by-and-only-by
reactions and counterreactions; or else only non-intentive reactions and
counterreactions exist. The latter directly calls into question any claims
we ourselves may make to personal cogency, but in either case it is readily
apparent from experience that usually non-rational causation only causes
non-rational effects. _Why then_ should we insist that our own real
reasoning is _not_ only knee-jerk reaction to stimulus? Or, alternately, if
our own ostensible reasoning is in fact only knee-jerk reaction to stimulus
_why then_ should we insist on being taken seriously?

The potential answers for 'why then' in either case, however, are
irrelevant, for in _any_ case they will require necessarily presuming what
they are hoping to defend.

Lewis doesn't present his argument in quite these terms and fashions, but I
find that this is the dichotomy he is trying to set up. If the proposition
of atheism necessarily requires that we end up having to justify that our
real-or-apparent justification abilities can in fact possibly be effective,
then dichomatic fallacies necessarily follow from proposing atheism.
Consequently, atheism should be removed from the list and not-atheism
(naturalistic or otherwise) should be believed instead. Formality aside,
not-atheism, of course, effectively equals theism.

Among other things, the argument recognizes that our effective reasoning
capability be a (if not _the_) necessary presumption for _any_ argument.
This means we cannot prove thereby that our reason exists; we can only
demonstrate and follow out what happens when we presume otherwise in a
self-reflexive fashion, observing along the way, not-incidentally, that
even to try to presume otherwise and operate on that presumption, we
essentially still must presume _not_-otherwise!

This is why Lewis writes (near the top of page 33), "If the value of our
reasoning is in doubt, you cannot try to establish it by reasoning. [...]
Reason is our starting point. There can be no question either of attacking
or defending it." Lewis' aim is to demonstrate that atheism necessarily
leads sooner or later to having to explain _why_ our justificational
ability (more specifically our own, _my own_--not some hypostatized
"humanity's") should be accepted as being possibly accurate. One way or
another, the atheist cannot get around this without begging the question.

(In passing, I will here note that the argument cannot be considered
finished at this point, for theism would have to be tested too for a
similar failure!--something Lewis, so far as I have ever been able to tell,
completely misses. Nor does Victor discuss it, so far as I recall.)


It will (and should) be noted that this is a soft deduction, even if it is
successful. In effect, by using Lewis' most dangerous argument (or a
refined variant of it), I am discovering that so long as I insist on taking
my own intellectual claims seriously, I ought to believe God exists, and I
ought not to believe that God doesn't exist. What I find I must necessarily
believe to be true about myself, for purposes of making _any_ argument, I
find I must doubt without any way to legitimately escape the doubt, so long
as I also propose atheism to be true.

The extent to which this may be considered to be a hard proof of God's
existence could be easily critiqued (and I expect rebutted). That it leads
deductively to a conclusion that I _should believe_ God exists, with the
action of that choice (to believe or not to believe) still to follow, is I
think less debatable--if the argument actually holds, of course.

Let me also say that this is only a barest rendering of the argument, and
that I do not even remotely expect it to be fleshed out sufficiently for
acceptance by an opponent yet. I can think of many topical considerations
still needing discussion in regard to the argument, myself. My main purpose
here is not to argue exhaustively for the position, but to contend that
this is ultimately what Lewis is trying to accomplish in his argument: the
deductive removal of a position by demonstrating that a necessarily fatal
dilemma necessarily follows from treating the position as being true. This
is also reflected, though more crudely, in practically every reference by
Lewis to what we now would call the theistic AfR (keeping in mind Lewis
never referred to it by this title), outside of MaPS, either edition. It is
also reflected in Lewis' move to become an absolute idealist upon
acceptance of the argument's conclusion: he precisely rejected _atheism_
and accepted _not-atheism_ instead, making the minimal move necessarily
implied by the conclusion.


Steve Lovell seems to think that Lewis only sees a _possibility_ of the
atheist making such a question-begging response--which is surprising given
Steve's own thorough analysis of the argument. I contend that Lewis
(rightly or wrongly) sees the eventual _necessity_ of the atheist making a
question-begging response; and _not_ to an argument by Lewis from reason to
not-atheism (an argument which, it should be observed, _Lewis hasn't
already given_ at the point Steve is commenting on!--though Lewis is in the
process of giving the concluding elements there), but to the challenge
posed by atheism's own first implications in regard to human epistemology:
_from_ nothing, comes _nothing_. From no-reason comes... reason? Or just
more no-reason? Either claim requires a subsequent follow-up that can only
beg the question in regard to the epistemic claims we ourselves are making
right that moment _in order to_ follow-up on either claim. This is
completely reflected in the quotes, from outside MaPS, with which Steve
begins his own analysis of the argument's history and implications, btw. It
will also be noticed that this is completely reflected in the title (and
contents?) of John Lucas' new debate with Elizabeth Anscombe, held several
years after Lewis' death, with Lucas taking the role of defending Lewis'
argument: "Is Mechanism Self-Refuting?"

(It would be a mistake to infer from this title that Lewis ended up only
aiming at a particular kind of determinism, by the way; as most of his 2nd
edition chp 3 is explicitly presented in discussion of a 'naturalism' that
need not even be purely materialistic! This explicit announcement occurs
right after the quote from Haldane, on page 20, which tends to lead some
critics, who couldn't be bothered to read a sentence or two further, to
think Lewis only has mechanistic determinism in view.)


Steve also claims (in relation to the quote from Flew presented by Victor
in his opening statement) that Flew "is not trying to remove doubts about
our cognitive faculties, he is attempting to stop those doubts from arising
in the first place."

It looks to me, from the quote given, like the doubt has already been
established, and Flew is trying to defend against it. (i.e. Ernest Gellner
writes, "If [naturalism] is true, then it is always _a mere coincidence_
that what we believe is also true." [my emphasis] To which Flew is, by
Steve's report, responding.)

But supposing Steve is correct instead, this still raises a portentous
question: why should Flew have to attempt to stop those doubts from arising
in the first place? Is he _having to_ attempt to stop those doubts from
arising? If so, and if he is appealing to the effectiveness of a process in
order to prevent those doubts, then he is still falling afoul of what Lewis
is actually attempting. For Lewis' ultimate point is that Flew, from the
presumption of atheism, _will have to_ make a defense (whether that
involves stopping doubts from arising in the first place, or whatever.)
Flew's doubt prevention strategy cannot possibly be successful at avoiding
begging what Lewis says is being questioned--which is why Lewis (in his 2nd
edition) answers Flew (in effect) the way that he does.

Put another way, Flew wants to succeed at keeping the possible reliability
of his cognitive faculties from being called into question, and proposes to
do so by using his presumably possibly reliable cognitive faculties to
explain why ancestors of his who had possibly reliable cognitive faculties
were more likely to survive to pass any improvements (accidental ones per
atheism, of course) in those possibly reliable cognitive faculties along,
etc. It doesn't matter what his defense attempt is, though. He has already
shot himself in the foot. (Which, I expect, is why Lewis can seem ambiguous
about whether he accepts or rejects portions of Flew's defense. It's a
secondary matter, and he isn't primarily concerned about it.)


BDK, in commenting on perhaps the key statement of the climactic summary
(where Lewis writes, "If the value of our reasoning is in doubt, you cannot
try to establish it by reasoning."), notes that if we accept this, we kill
epistemology. To the extent that epistemology currently involves justifying
our justification capabilities, that would indeed be true!--but that
doesn't make the statement less correct. It would just mean epistemology
has gone badly off the tracks somewhere (or numerous somewheres.)

In any case, Lewis can be said to proceed by assuming BDK's (1) ("That
human reasoning is epistemically valuable"); but Lewis is _not_ actually
proceeding by going on to BDK's (2) ("Argue that this skill couldn't have
evolved via natural processes.") That would be a variant of the argument
from reason, certainly, and a popular variant, too--Victor, for example,
frequently appeals to it in various ways (as do I on occasion. So does
Steve in the article Victor links to.) But it is not what Lewis is doing.

Also, for Lewis to make the claim quoted above ("If the value of our
reasoning is in doubt..."), manifestly does _NOT_ undercut the first of the
steps listed by BDK. To observe that we cannot reasonably reestablish the
value of our reasoning once that value is doubted, is not the same as
actually _doubting_ the value of our reasoning. (BDK's further comments are
not inappropriate, but Victor is handling those elsewhere.)


I have an expectation that Victor's denigration of Lewis' efforts in what I
am calling the climactic summary portion of MaPS chp 3, hinges on Victor's
preference for Best Explanation variants of the AfR instead of Sceptical
Threat variants; and be he right or wrong, Lewis in _that_ portion is
certainly making a type of Sceptical Threat AfR. But I don't believe Lewis
is making quite the same kind of ST-AfR that Victor nominally rejects.

What Victor objects to (as Steve in his own analysis rightly summarizes
from Victor's work), is the kind of argument that proceeds by raising
sceptical doubts about the validity of reasoning and then goes on to argue
that such doubts can be resolved only if 'naturalism' (more specifically
atheism) is denied. Victor correctly observes that no absolute security
against such doubts is available from any quarter. Of course not!--if we
appeal to theism (or not-atheism) in order to reestablish our security on
this, then we are only doing what Lewis is excoriating the atheists for
doing. But that is not the point of the Lewisian AfR; or anyway that is not
the point of its 'gist', so to speak, for as Lewis presents it the argument
does need better phrasing in order to clarify its proper implications. (I
am especially but not exclusively thinking here of the chp 3 material
subsequent to the climactic summary, which if read out of context can
easily seem to be appealing to theism in order to justifiably restore our
confidence in the possible reliability of our justification ability. I
don't believe Lewis is actually doing this even here, any more than he does
on a similar application later in chapter 13 "On Probability", but it could
be misunderstood that way.)

Properly presented, though, the most dangerous version of the Lewisian AfR
needs no such re-establishment. It proceeds by beginning from Reason, just
as Lewis says, and by looking to see whether one of a dichomatic group
itself necessarily raises questions about reasoning which necessarily need
answering but can only be answered by begging the question. At a more basic
level, the Lewisian AfR isn't even really about deducting away a threat to
the _validity of reasoning_ per se; a phrase that is probably a non
sequitur anyway (though used by Lewis in his presentation.) Lewis is not
assuming validity (or whatever may be considered most important about our
reasoning for purposes of presenting our own arguments) is a fact and then
asking whether in an atheistic reality one can account for the assumed
fact. He is demonstrating that under the proposition of atheism, we have to
try to account for something otherwise necessarily presumed to be factual.
This is subtly but crucially different as a tactic.

We may, if we wish, reduce Lewis' point down to this: _from_ the necessary
presumption of our reasoning ability, for sake of argument, we can consider
atheism or not-atheism. Does either of these require us to try to
subsequently account for the existence of something we are already
necessarily presuming to be factual? Since, under atheism, our reasoning
behaviors must depend upon non-rational behaviors, this necessarily
requires us to try to subsequently account for the existence of a type of
behavior which we are already presuming to factually exist, even in order
to try making that account. The distinction between theism and atheism is
precisely a distinction about the existence of that type of behavior.
Atheism denies that behavior's existence at a fundamental level; requiring
it to be produced or substituted in derivation. Theism at least doesn't
involve the denial of that behavior's existence, and so doesn't introduce
an immediate disparity in regard to our own derivative reasoning claims.


Does atheism involve introducing conflicts with the character and quality
we ourselves claim for our own reasoning when making any argument? Lewis
demonstrates (or at least aims at demonstrating) the answer is yes; and
then effectively assumes without further evaluation that the answer would
necessarily be no for not-atheism (i.e. for theism.)

If Lewis' demonstration is properly accurate, though; _and_ if an
evaluation of theism's implications does not reveal necessary introduction
of conflicts with the character and quality we ourselves claim for our own
reasoning when making any argument; then we will have deduced, not strictly
that atheism is false, but that so long as we take our own reasoning
seriously we ought to reject atheism as false--and accept not-atheism as
true.

Which, in a cruder but equivalent fashion, is how and why Lewis the atheist
eventually decided to believe theism to be true instead. (While, at first,
_keeping_ his philosophical _naturalism_.)


Jason Pratt


Addendum: it may be asked if this means no Best Explanation AfR can be
derived from Lewis' work. I believe Victor and others have amply
demonstrated that various kinds of BE-AfR may be derived from, or
illustrated by, Lewis' work as well. I also believe, however, that Lewis
was not primarily aiming at this, which explains why in his own climactic
summary he is clearly not making that kind of argument. Efforts to try to
demonstrate that the primary importance of Lewis' AfR is in providing a
BE-AfR instead, are always going to be unable to account for that climactic
portion, and must at best simply wave it aside as being a peculiar
mis-step.