CHAPTER 6
Strong Higher-Level Theistic
Evidentialism:
Reflective Rationality and the
Project of Natural Theology
The kinds of evidentialist requirements
I have been developing in the last two chapters contribute to what I have called modest evidentialism. This version of
evidentialism has emphasized the causal and epistemic significance of propositional evidence
for individual believers in their specific epistemic situation. I have developed the
partial sustaining and overdetermining function of reasons, especially under defeating
conditions for theistic belief. For some people in certain circumstances, their holding
(firm) theistic belief and its having a positive epistemic status will require that their
theistic belief be at least partly sustained by reasons. In chapter 5 I extended this to include instances in which
theistic belief is based wholly on propositional evidence alone. The next two chapters
will have as their objective a strengthening of modest evidentialism by the introduction
of another kind of evidentialist requirement, what I will be calling the strong higher-level evidentialist requirement.
In the present chapter I begin by
utilizing William Alston’s multi-level foundationalism to present a distinction
between some person S being justified in believing that p and S being justified in
believing that <S is justified in believing that p>.
Alston suggests that even if a person S is immediately justified in some theistic
belief that Pt, S’s being justified in believing that <S is immediately justified
in believing that Pt> requires mediate justification. Based on this I develop the idea
that reasons are necessary for a cognitive desideratum I will be calling reflective rationality - our coming justifiably to
believe that some belief that p is justified, rational, or possesses some epistemically
meritorious status. After laying out the Alstonian position in Part I, I extend
Alston’s account in Part II and develop more fully the notion of reflective
rationality, distinguishing between externalist and internalist theories of reflective
rationality. In Part III I consider the nature of natural theology in the light of these
developments and argue that we may plausibly view the project of natural theology as a
pursuit of reflective rationality towards our theistic beliefs.
I.
Higher-Level Theistic Evidentialism
Aristotle (Posterior Analytics Bk. II,
72b19-24) seems to have thought that the immediate truths which end the epistemic regress
are and must be indemonstrable. This has contributed to the common view that foundational
beliefs must in some way be unique, wear their justification of their face, be
self-justified, or possess epistemic immunities of some sort. Otherwise foundational
beliefs would appear to be wholly arbitrary. This worry becomes acute if we reject
classical foundationalism and adopt a version of modest foundationalism. One thereby opens
oneself to the charge of arbitrariness with respect to what counts as a basic belief. I
explained in chapters 1 and 2 that one of the objections to theistic foundationalism is
based on a more general objection to modest foundationalist theories of justification (and
knowledge). Terminating the chain of justification in beliefs that are (wholly) justified
by something other than their relation to other justified beliefs appears to lead to
arbitrary dogmatism. If all justificatory paths terminate in immediately justified
foundations, the entire structure of beliefs appears to rest on beliefs for which there
are no reasons or for which no reasons can be given.
A. Alston and Multi-Level
Foundationalism
Alston has devoted several papers to articulating and defending a version of modest
foundationalism that, among other things, provides an answer to this objection to
foundationalism.[1] The answer directly
challenges the Aristotelian “indemonstrable foundational beliefs” thesis.
Alston’s version of foundationalism distinguishes between various levels built up
from the introduction and iteration of pistic, justificatory, and epistemic operators,
thereby creating multiple levels of belief, justification, and knowledge.
From any proposition p, we can build levels in the following manner:
1. Pistic Levels
p
S believes that p.
S believes that <S believes that p>.
2. Justificatory Levels
p
S is justified in believing that p.
S is justified in believing that <S is
justified in believing that p>.
3. Epistemic Levels
p
S knows that p.
S knows that <S knows that p>.[2]
One of the important consequences of these level distinctions for foundationalism
is that it suggests another way in which beliefs may be assessed in terms of reasons, even
where beliefs are immediately justified. I have explained in previous chapters that the
foundationalist commitment to immediate justification need not preclude mediate
justification of the same belief (even at the same time). Alston’s epistemic level
distinctions highlight a specific way immediately justified beliefs are open to evaluation
in terms of reasons, and so by implication open to a kind of mediate justification. Even
where some person S’s belief that p is
immediately justified, it remains possible (at least in principle) to find reasons for the
higher-level proposition <S’s belief that p is immediately justified>. So, for
example, where S’s belief that <It is raining outside> is immediately
justified, we may say that there exists a correlative higher-level proposition
<S’s belief that It is raining outside
is immediately justified>. Alston maintains that, even if a putative belief that p is immediately justified, this does not preclude
seeking reasons for correlated propositions at the higher-level. Since every nonepistemic belief that p may be
thought of as having a correlative epistemic belief that <S’s belief that p is
justified>, even if one is restricted to immediate justification on the lower-level for
the belief that p, one may adduce reasons at the higher-level for the belief that
<S’s belief that p is immediately justified>. So, though S may be immediately justified in believing that <it
is raining outside>, S may have reasons for regarding his belief that <it is raining
outside> as immediately justified. In other
terms, S may have a mediately justified belief that <S’s belief that It is raining outside is immediately justified>,
even if the belief that p is immediately justified.[3]
Nor does this perpetuate the regress as Aristotle might have thought, for
higher-level beliefs are not needed to stop the regress but only to show that we have
reason for regarding the regress as stopped.
Recalling Alston’s concept of justification, a belief that p is justified if and only if it is based on an internally accessible, objectively adequate ground. What the justification of any belief b depends upon is the existence of a valid epistemic principle that applies to b. S need not (justifiably) believe or know anything about the justificatory conditions that obtain (though the ground - not its efficacy - will be cognitively accessible). To be justified in some belief that p requires (a) the existence of a valid epistemic principle and (b) S’s belief that p satisfying the conditions the principle lays down. But it is precisely the subject's coming to have (justified) beliefs the content of which is given in (a) and (b) that will come to bear on the higher level. For (a) and (b) are the types of reasons we will have for regarding some belief that p as justified. Call these meta-reasons. They are reasons for regarding some belief as justified (in the appropriate sense). So whereas an immediately justified belief that p requires the existence of a valid epistemic principle that relates to the belief in question, a mediately justified belief that <S’s belief that p is immediately justified> will be based upon (justified) beliefs that the epistemic principle for the belief that p is valid and that the belief that p may be subsumed under the principle.[4]
Hence, we may add the following principles as constituents of modest
foundationalism:
The general principle of epistemic levels:
[L1]
Given any proposition p, there is a correlated higher-level proposition P*, where P*
is built up from the iteration of pistic, justificatory, or
epistemic operators.
For every putative immediately justified belief that p, there is a higher-level evidential option:
[L2] Given any person S, if S’s belief that p is immediately justified, then S’s belief that <S’s belief that p is immediately justified> can (in principle) be mediately justified.[5]
B. The Strong Higher-Level
Evidentialist Requirement
But the position envisaged here is stronger than it might first appear. According
to Alston (1989c, pp. 24-25, 36-38; cf. 170), the justification of any higher-level belief would actually require mediate justification. This gives rise to what I call the strong higher-level evidentialist requirement. Any belief that <S’s belief that p is
immediately justified> is, if justified, mediately justified. This is not to say that one can’t be
immediately justified in the lower-level belief that p without being mediately justified
in a higher-level belief (a clear case of level confusion). It is only to say that if one is justified in any higher-level belief that
<the belief that p is immediately justified> then that belief must be based upon adequate reasons.
The higher-level evidentialist requirement I am speaking of is not a requirement for being
justified in some belief that p, only for being justified in some belief that P*.
In “Two Types of Foundationalism,” Alston explains how this strong
requirement for justification at the higher-level answers the charge of arbitrariness and
dogmatism.
Though the simple foundationalist
requires some immediately justified beliefs in
order to terminate the regress of justification, his position permits him to recognize
that all epistemic beliefs require mediate justification. Therefore, for any belief that
one is immediately justified in believing, one may
find adequate reasons for accepting the proposition that one is so justified. The curse (of dogmatism) is taken off immediate
justification at the lower level, just by virtue of the fact that propositions at the
higher level are acceptable only on the basis of reasons. A foundational belief, b, is immediately justified just because some valid
epistemic principle lays down conditions for its being justified which do not include the
believer having certain other justified beliefs. But the believer will be justified in
believing that he is immediately justified in
holding b only if he has reasons for regarding that principle as valid and
for regarding b as falling under that principle.
And if he does have such reasons, he certainly cannot be accused of arbitrariness or
dogmatism in accepting b. The absence of reasons
for b is “compensated” for by the
reasons for the correlated higher level belief. (Alston
1989c, p. 37)
Alston’s position here is by no means uncontroversial, and I shall address its plausibility in the next chapter over against some basic objections. For the moment I simply note that according to Alston the justification of this strong evidentialist requirement follows from what he considers to be the implausibility of being immediately justified (in a “truth-conducive” sense of justification) in beliefs about the epistemic status of some putative belief that p.[6] Many beliefs that are immediately justified are so justified by virtue of the direct experiential presentation of the (content of the) target proposition to the subject’s awareness. But Alston is doubtful about this mode of justification generating immediately justified beliefs about epistemic facts or statuses; for it does not seem plausible that we are “presented” with epistemic statuses. Other beliefs are immediately justified because they are self-evident. But even if a person takes it without question that a belief of theirs is justified, this hardly constitutes a case for the self-evidence of epistemic beliefs. Doubtless, there are other ways of spelling out the range of immediately justified beliefs that might yield a case for immediately justified epistemic beliefs. I shall come to these in the next chapter. Alston’s position, though, rests rather firmly on the conviction that for each plausible model of immediate justification we will not find a decent case for subsuming epistemic beliefs under any such models.[7]
Let us, therefore, lay down (at least provisionally) the principle of the strong
higher-level evidentialist requirement:
[E*] Given any person S, S is justified in believing that <S’s belief that p is justified> if and only if S’s belief that <S’s belief that p is justified> is based upon adequate reasons.[8]
The distinction between epistemic levels entails that there are actually two distinct levels on which the evidentialist
requirement for reasons or mediate justification may operate. There is a lower-level
evidentialism which requires a mediate justification of nonepistemic beliefs, the
theistic evidentialist correlate of which is:
[E]
Given any
person S, S’s belief that Pt is justified only if S’s belief that Pt is based
upon adequate reasons,
and there is a higher-level evidentialism ([E*]) which involves the imposition of an
evidentialist requirement on the correlative epistemic beliefs of nonepistemic beliefs,
which for theistic belief entails:
[ET*]
Given any person S, S is justified in believing that <S’s belief that Pt is
justified> only if S’s belief that <S’s belief that Pt is justified> is
based upon adequate reasons.
The modest evidentialism of chapters 4 and 5 may now be thought of as modest lower-level
evidentialism, which we are now interested in extending by adopting the principle of
strong higher-level strong evidentialism. I emphasize that strong evidentialism in no way
compromises the stance of foundationalism which terminates justification in immediately
justified beliefs, for the mediate justification of any higher-order belief may be thought
of as resting on reasons which themselves terminate in immediately justified beliefs at
the lower-level. Foundationalism is a first-order theory of the justified structure of our
beliefs. There is no reason to be committed
to meta-foundationalism, holding that the structure of epistemic beliefs divides into
foundational and superstructure beliefs. The path of any mediately justified belief bm
ultimately leads to some immediately justified belief bi. The justification of
bm is spelled out in terms of other beliefs, and the justification of bi
in terms of some justification-conferring property X. Bi is justified if in fact it possesses
the appropriate justification-conferring property X.
To push the justification question beyond this is to raise the question about whether Bi
does in fact possess X and whether property X does confer justification. If S is to be
justified in supposing that <S’s belief that p is immediately justified>, then
S must have (good) reasons for believing that Bi possesses X and the possession of X renders Bi (immediately) justified. At
this point though we are talking about beliefs as to what is the case (at the lower level)
and what sorts of principles of evaluation are valid, not beliefs about epistemic
statuses. In this way, a path from a mediately justified belief that bm will
lead to an immediately justified belief Bi, and a path from a mediately
justified higher-level belief BJm*
will lead to some immediately justified lower-level belief (even if the path passes
through one or more mediately justified lower-level beliefs). I take it that these considerations show that
within the foundationalist structure of justified belief we should distinguish between the
structure of lower-level beliefs and the structure of beliefs which involve attributions
of justification.
II. Externalist and Internalist
Reflective Rationality
Alston’s multi-level foundationalism and the higher-level evidentialist
requirement have some far-reaching epistemological implications. These implications I will
subsume under the notion of reflective rationality
(or justification). By reflective
rationality I will understand a particular evaluative stance we may take towards our
beliefs, evaluating them in terms of their justificatory or epistemic status. When I come
to hold that my belief that God exists is immediately justified, I am evaluating that
belief in terms of the property of justification. As about to be explained, what I take
justification to be will determine the specific nature of my evaluative stance, as well as
the sorts of particular reasons that will justify me in holding the higher-level belief.
But the general point is that we have a distinct cognitive desideratum (distinct from
being justified in believing that p) which is satisfied when we come justifiably to
believe that our beliefs have some positive epistemic status. The satisfaction of the
higher-level evidentialist requirement vis-à-vis the belief that P* entails being reflectively rational with
respect to the belief that p.
A. Externalism and Reflective
Rationality
Let’s look at Alston’s account a little more closely. A belief that p is
justified if and only if it is based upon an internally accessible, objectively adequate
ground. As we have seen, objective adequacy is taken in an externalist sense of actual
truth-conducivity. S’s belief that p on ground G will be justified only if beliefs of
the sort that p is tend to be true when based on grounds of the sort that G is. So for
instance, my belief that I see a tree when appeared to a certain way will be justified
only if sensory experience (plus other features of the situation which functioned as input
in the formation of the belief, such as my possessing reliability-relevant
characteristics) tends to yield mostly true beliefs, so that the objective probability
(see chapter 1, section II. D) of such beliefs being true is fairly high given those
grounds. For theistic beliefs based on the putative direct experiential awareness of God,
a belief that <God is present> will be justified on such grounds only if there is
something like a nonsensory perceptual awareness of God which usually produces true
theistic beliefs (of the appropriate sort). What is needed for S to be justified in
believing that <S’s belief that God is present is immediately justified> are
reasons for regarding the experiential grounds of his theistic belief as adequate. Since
adequacy entails reliability, one will need reasons for supposing that the grounds of
one’s theistic belief (or the processes leading to its formation and/or sustenance)
mostly yield true beliefs.
In the case of sensory perceptual beliefs, determining whether they have been
reliably produced will depend on what we can learn about our cognitive faculties and
physical environment from psychology, physics, and physiology. Alston has argued at length (“Epistemic
Circularity” in 1989c; 1991c; 1993b,c) that there can be no epistemically noncircular
case for the reliability of any basic source of belief. To determine whether our sensory
perceptual mechanisms are reliably produced we must rely on our perceptual beliefs, and
thereby assume (in practice) that those beliefs are
reliably produced. The same will be true for our other basic sources of belief:
introspection, rational intuition, and induction. This is no bar to our using them to
actually show that some source is reliable, for if they are reliably produced, then they
can be used to show that the sensory perceptual experience is reliable. It is only to say
that all such arguments for the reliability of a basic source must assume the conclusion
(i.e., X is a reliable source of belief) if we are to be justified in accepting the
premises. Moreover, Alston has claimed that just as the reliability of our sensory
perceptual beliefs is decidedly an empirical matter, it is primarily a theological matter
as to whether some source is a reliable source for theistic beliefs. We must have recourse
to what it is that we (allegedly) know about God, his nature, purposes, and actions if we
are to determine whether our beliefs about him are likely to be true when based on some
particular source, such as religious experience.
Generally speaking, to determine the reliability of a particular belief-forming
mechanism we have to rely on what we know (or reasonably believe) about the reality with
which this mechanism deals and our cognitive relations thereto. For empirical
belief-forming mechanisms, this means carrying out empirical investigations, including the
sophisticated versions thereof known as science. As for religious belief-forming
mechanisms, this means carrying out investigations into the nature and doings of God, and
this means doing theology. (1992, pp. 42-43)
On Alston’s account, S’s being justified in believing that <S is
immediately justified in believing that Pt> requires a consideration of the conditions
causally operative in the formation or sustenance of S’s belief that p and reasons
for regarding these conditions as conferring justification (i.e., as being
reliable-indications of the truth of the belief). There are two aspects to reflective
rationality here. First, a source-relevant view of justification: what confers
justification on a belief is causally operative as a ground in the process of the
belief’s formation or sustenance. Secondly: externalism with respect to the adequacy
of the relevant grounds or cognitive processes. Reflective rationality here involves
self-referential epistemic beliefs. This corresponds to the notion of personal or doxastic
justification we looked at in previous chapters. Externalism entails a source-relevant
view of justification, and so vis-à-vis higher-level considerations it entails reflective
rationality towards one’s actual belief on the basis of some actual ground. It is a
thoroughly doxastic conception of reflective rationality.
So let’s say that:
[RRE]
S exhibits externalist reflective
rationality with respect to the belief that p just if S has adequate reasons for supposing
that the causal source of S’s belief that p (e.g., the grounds or relevant cognitive
mechanisms involved in the production or maintenance of the belief) mostly yields true
beliefs.
And where [RRE] entails the conditions of
doxastic reflective rationality:
[RRD]
S exhibits doxastic reflective
rationality with respect to the belief that p just if S has adequate reasons for supposing
that the grounds for S’s belief that p provides adequate support for S’s belief
that p.
[RRD] suggests a different
construal of reflective rationality - an
impersonal or propositional version of reflective rationality according to which
individuals arrive at beliefs about the justificatory status of a certain proposition p by virtue of the reasons there are in
support of p, even if those reasons were not causally operative in the formation or
sustenance of S’s belief that p. Impersonal or propositional justification, as
explained earlier, concerns there being a justification available for a certain
proposition, typically by virtue of the existence of evidence which provides some relevant
degree of evidential support for the proposition. We might think of such evidence as
potential grounds, for if S were to believe p on the basis of the evidence, S would have
an adequate ground for his belief. It is open
to a person, then to be reflectively rational about his beliefs in two ways. First, by
considering his actual belief that p and the
ground of his belief that p. Secondly, by
reflecting on the proposition p (which he may or may not actually believe) and the
evidence (or potential grounds) available in support of it. This becomes particularly
relevant when a person’s belief that Pt is immediately justified, say by virtue of
being based on some experience which adequately supports the belief that Pt. In this
situation, one may exhibit doxastic reflective
rationality by considering the extent to which the experience in question does provide the
appropriate justificatory grounding for the belief. Alternatively, one could exhibit propositional reflective rationality by considering
the adequacy of the reasons for the proposition Pt, though such reasons were not causally
operative in one's coming to believe or presently holding the belief that Pt. Here adequacy will be something like the extent
to which Pt is rendered probable by evidence e given some set of inductive standards.
So in contrast to [RRD] we
have something like:
[RRP] S
exhibits propositional reflective rationality
with respect to the proposition p just if S has adequate reasons for supposing that there
is evidence e that provides adequate evidential support for the proposition p.
B. Internalist Reflective Rationality
Moreover, we may contrast [RRE]
with an internalist conception of reflective rationality that is closely related to
propositional reflective rationality:
[RRI]
S exhibits internalist reflective
rationality with respect to the belief that p just if S has adequate reasons for supposing
that there is adequate evidence for the belief that p, the adequacy of which is
cognitively accessible upon reflection.
[RRI] certainly seems to
entail [RRP], but it is compatible
with [RRD]. The internalist might
require that the evidence mentioned in [RRP]
and [RRI] be the basis of one’s
belief that p if that belief is to be justified. Internalism is compatible with a
source-relevant view of justification. There would simply be a kind of straightforward
correspondence between what makes the evidence for p adequate and what makes the ground
for one’s belief adequate, and so what provides justification for the proposition p
provides justification for one’s belief that p. The point should be briefly
elucidated. What determines the rationality of a proposition (or belief) is - generally
speaking - the epistemic adequacy of the supporting evidence or grounds. Given the truth
oriented nature of epistemic adequacy, some sort of probability relation is typically
viewed as the determinant of adequacy. Evidence (or a belief’s ground) is adequate
just if it renders the proposition or belief h probable. On the externalist view, the
appropriate probability relation will be objective features of the subject’s
empirical situation measured in terms of something like the statistical likelihood of a
belief’s being true when it is based on the grounds (or produced by cognitive
mechanisms) of a certain sort. On the internalist view, the appropriate probability
relation will be a logical relation between evidence e and a proposition h such that given
some set of inductive criteria e makes h probable. Recall the earlier distinction between
epistemic and subjective evidential probability that I used in chapters 1 and 2 to suggest
ways of thinking about justification. There is a minimal kind of propositional (and
internalist) justification for a proposition h which results from S’s evidence e
rendering h probable given S’s own inductive standards, what we can call
justification by subjective evidential probability. There is also the stronger notion that
S’s evidence renders h probable given correct inductive standards (and where S is
less than a logically omniscient being), what we can call justification by epistemic
evidential probability. If one is inclined to adopt some source-relevant position, one can
add the based-on requirement and think of the evidential conditions as a ground of belief
and as setting out the necessary and sufficient conditions for doxastic internalist
justification. One can think of the first case as involving belief formation or sustenance
satisfying S’s own inductive standards, and the latter case as satisfying correct
inductive standards. In each case, a person responds to the evidence before him. It is
merely a question of whether the response satisfies correct standards.
On this internalist construal, when it comes to the higher-level there are three
sorts of considerations which are involved in reflective rationality: (a) a person’s
inductive standards, (b) a person’s evidence-base e, and (c) the probability relation
between h and e given (a). Take the matter of
inductive standards first. It is one thing to respond to evidence given one’s own
inductive standards, it is another thing to subject those standards to critical scrutiny.
Now S might subject his standards to criticism which he takes to be adequate, which is
adequate given his inductive standards, or which is adequate given correct inductive
standards. Similarly, a person’s evidence may result from investigation that S takes
to be adequate, which is adequate given S’s own inductive standards, or which is
adequate given correct standards. Likewise, whether h is rendered probable by e is
something that S may check to an extent adequate in his own view, given his own standards,
or given correct standards. In each condition, we have two kinds of perspectival
constraints on the adequacy of investigation and criticism. In each instance, the last
option represents a critical reflection that satisfies objective standards.
There is then a minimal subjective internalist reflective rationality. In
reflecting on whether e renders h probable, S critically assesses his inductive standards
to a degree that he takes to be adequate. His evidence is also the sort of thing that has
resulted from investigation that S takes to be adequate. S also checks - adequately in his
view - to see whether h is probable on e. A
person fails to have reflective rationality in this sense when there is a failure on the
person’s part to collect the appropriate sort of evidence (or when the person
intentionally collects evidence biased in favour of the desired proposition) or when the
evidence is dubious. A person will not be reflectively rational in this sense if he does
not critically consider whether his inductive standards are the correct ones. Moreover, if
S does not check to see whether the appropriate probability relation obtains between e and
h, given his standards, he fails to be subjectively reflectively rational. Now clearly,
one may exercise critical reflection on some of these conditions or all of them. There are
degrees of subjective reflective rationality (say between partial and complete reflective
rationality). A stronger sense of subjective reflective rationality arises when the
perspectival adequacy of the critical reflection, etc. is supported by S’s own
inductive standards. A person fails to have
this sort of reflective rationality when, for instance, he has standards as to what is the
adequate amount of investigation into evidence, but they are not satisfied by his actual
investigation (and there is no overriding reason for him to deviate from his normal
standards). In the last case, there is a kind of reflective rationality in which S’s
evidence base results from investigation which was in fact adequate, S’s critical
assessment of his standards was adequate, and S has adequately checked to see if e does
render h probable. Call this last version objective internalist reflective rationality.
Another important distinction is between what we might call synchronic and diachronic reflective rationality. I began the discussion of reflective rationality by identifying it with a synchronic state - a person’s being in possession of the relevant evidence that renders him justified in some higher-level belief. Call this synchronic reflective rationality. The last few paragraphs, though, have introduced notions like “evidence resulting from investigation” and “critically assessing one’s inductive standards.” S goes through a process of gathering and evaluating evidence, investigates whether his inductive standards are the correct ones, and checks to see whether p is made probable given his evidence and inductive standards. These matters are temporally prior to one’s coming into a state of reflective rationality. Reflective rationality, then, is broader than merely possessing a justified higher-level belief. It includes the activities of gathering evidence and evaluating the matters included in (a)-(c) above. This is important since one may go through such a process and not have a justified higher-level belief, but nevertheless one has engaged in the kinds of voluntary activities for evaluating one’s belief from the epistemic point of view. Call this diachronic reflective rationality. It is this sense of reflective rationality that is exemplified in procedural epistemic rationality (delineated in chapter 2, section III.A). Being a responsible truth seeker involves voluntarily engaging in the activities that will help one see whether one’s belief that p is likely to be true. This is diachronic reflective rationality. As I understand it, synchronic reflective rationality entails (some degree of) diachronic reflective rationality.[9]
We should recognize a final point here. Suppose S is immediately justified in some
belief that h (in an externalist sense of justification). In addition to S's exhibiting
externalist reflective rationality toward h by considering reasons for supposing that his
belief that h was reliably produced, he may consider the extent to which there is evidence
e for the proposition h. That is, he may
consider the extent to which e makes h probable given S’s own inductive principles,
and the extent to which the kind of inductive standards S holds are correct ones. Much in
the same way doxastic and propositional justification may be combined at the lower level,
one may combine an externalist lower-level justification with an internalist higher-level
propositional justification. The implications of this for S’s remaining immediately
justified (in a source-relevant sense) as opposed to becoming mediately justified (in a
source-relevant sense) after one has become externalistically or internalistically
reflectively rational, I leave for the next chapter. I want now to consider how the
preceding distinctions bear on how we understand the project of natural theology.
III. Natural Theology: An Exercise in
Reflective Rationality
The distinction between unreflective and reflective rationality permits, by way of
application, a more sophisticated perspective on the nature, relevance, and function of
natural theology. Although earlier chapters presented several arguments for the importance
and restricted necessity of natural theology for being justified in (lower-level) theistic
beliefs, my claim here will be that it also provides the resources and framework for our
coming to be justified in believing that theistic belief is justified.
A. Natural Theology and First-Order
Justification
Since the nature of higher-level evaluation depends on what justification amounts
to at the lower level, I begin by unpacking externalist and internalist views of natural
theology as a basis for first-order justificational status.
Here are some putative reasons for belief in God typically associated with natural
theology.
(1) There exists a
complex physical system that has existed for finite time.
(2) There exists a
complex physical system (of either finite or infinite temporal
duration).
(3) The Universe
conforms to simple, intelligible scientific laws.
(4) The Universe
exhibits various kinds and degrees of beauty.
(5) There exist, within the
Universe, conscious human persons with knowledge,
freedom, and power.
(6) Human persons have
a sense of right and wrong.
(7) There are binding moral
truths.
(8) Individual human
persons have had requests made in prayer come to pass.
(9) Individual human
persons have witnessed sudden healings or phenomena otherwise characterized as
“miraculous.”
(10)
People of many different cultures and periods in history have believed that God
exists.
(11) People of many different cultures and periods in history have claimed to have an experience of God.[10]
On an externalist account of justification (or knowledge), such as Alston’s,
the extent to which any one of (1)-(11) contribute to a person’s being mediately
justified in believing in God will depend on the actual truth-conducivity of these
reasons. Suppose a person believes in God on the grounds of certain beliefs he has about
general features of the Universe which are evident to him (e.g., the existence of the
Universe as a complex physical system and its conformity to relatively few, simple,
intelligible scientific laws). For Alston, theistic belief is justified in such cases just
if this propositional evidence is an adequate ground. And it will be adequate only if
propositional grounds of that sort (as what was
taken into account during the formation of the belief) typically yield true theistic
beliefs of the appropriate sort. I say
“only if” because presumably other features, such as reliability relevant
characteristics of the person, would be a further determinant of adequacy (recall the
Swainian account in chapter 1). So, for instance, an especially gullible or overly
credulous person would lack justification in cases of evidence like (8) or (9) (or
potentially all of the reasons listed above), whereas the same evidence would suffice for
a justified belief in God for a more discriminating person. Moreover, in cases of explicit inference, there
is the additional question as to whether the principles of inference involved mostly yield
true conclusions given true premises. What this shows is that for the externalist adequacy
supervenes on contingent features of a person’s empirical situation, not merely the
evidence in question.
In chapter 1 I explained how the notion of “adequate” grounds or evidence
is complicated by the association of adequacy with probability, as there are several
senses of probability which yield different senses of adequacy. When spelled out in terms
of conditional evidential probability, the Alstonian account does not conditionalize
solely on the evidence for h (and which S is aware of). But it conditionalizes also on
factors external to the person’s mental life, factors which the externalist takes to
be essential for determining the likelihood of S possessing a true belief that h. The
person’s possessing reliability-relevant characteristics is essential (as is the
absence of relevant unreliability characteristics). More generally, if the actual causal
source of a person’s belief is a factor in determining the justification of their
belief, and reliability is necessary for justification, then a person may have good
evidence for the proposition h, but not be justified in believing h (or, if belief origin
is essential to knowledge, S may be justified but not have knowledge). Consequently,
natural theology may supply one with propositions that constitute good evidence for the
proposition that God exists (perhaps one even believes it on the basis of such evidence).
But if the evidence causally operative in the acquisition or sustenance of one’s
theistic belief is not this evidence (or is some other evidence or grounds which are
inadequate, say wishful thinking, paranoia, or a dream one had), one will not be
justified.
On an internalist view, matters are different. Here the idea of S’s evidence is central, and the empirical matters of belief formation and reliability of belief-forming processes drop out of the picture. The evidential force of the evidence for h given certain inductive standards determines "adequacy." Typically the internalist spells this out in terms of evidential probability. But owing to the earlier distinction between subjective and epistemic evidential probability, we have at least one important distinction. S may have as his evidence for the proposition <God exists> a set of beliefs he has about the Universe, for instance the presence of spatial and temporal regularities within the vast complex physical system we call the Universe. This evidence may render S’s belief that God exists probable given S’s own inductive standards. S’s criteria might include the following. "The simpler hypothesis is more likely to be true. The probability of a hypothesis h is raised if observed phenomena are entailed by or likely to occur if h is true. The hypothesis fits with background knowledge, and so contributes to h’s prior probability. H’s explanatory power is a function of the prior probability of the evidence and the predictive power of h, etc." Then again S may be mistaken in his inductive standards or he may not be justified in holding to what is evidence for the theistic hypothesis. Perhaps he thinks that more complex theories are more likely to be true or that fit with background knowledge is the sole determinant of prior probability. Or suppose the evidence is an alleged healing S observes at a revival meeting - evidence which S might not be justified in taking as evidence (perhaps it wasn’t actually a healing which S observed). In other words, we can distinguish between instances in which S’s belief that God exists is made probable by evidence given his own inductive standards and cases in which the evidence is justifiedly held and S’s inductive standards are the correct standards (such as the Bayesian criteria above). So we can distinguish between two kinds of rationality. First, there is the sort of rationality that goes with a proposition S believes given evidence e makes h probable given S’s own standards (subjective evidential rationality). Then there is the rationality that goes with e making h probable given correct inductive standards (and when S is justified in holding the evidence he does), a kind of objective evidential rationality. Notice here though that what makes a ground adequate is not that beliefs formed on the basis of it are statistically likely to be true (though that may be the case). Rather there is a certain body of evidence that renders a proposition likely to be true given certain inductive standards.[11]
B. The Higher-Level Relevance of
Natural Theology
On Alston’s account, the role of natural theology at the higher-level will naturally be restricted to whatever support the content of natural theology can provide for the reliability of cognitive mechanisms producing theistic belief. Since Alston believes that the reliability of theistic sources of belief is in large part a theological question, it would seem that natural theology would have something to offer here. And since natural theology is not an immediate source of belief, it would be a way of validating the outputs of practices of theistic belief formation on the basis of mystical experience without assuming the reliability of the premises of those practices. But as Alston notes (1991c, p. 144) natural theology operates at too high a level of abstraction to grant much input to the sorts of concrete considerations which are involved in determinations about reliability. In much the same way an independent argument for the existence of the external world would not provide adequate support for the claim that the sensory perceptual practice is a reliable way of knowing the world. Even if we can support the notion that God would display himself in our experience, we would still need to stretch this to include many particular occasions when different people sense that God is manifesting himself to their consciousness then and there. If Alston’s position is correct, it may be that the Bible, tradition, and religious experience itself each play a larger role than natural theology in determining whether certain putative sources of immediate belief in God are reliable.[12],[13]
Propositional reflective rationality, though, and the internalist understanding of
evidential probability provides interesting prospects. A person may be immediately
justified in holding to some theistic belief that Pt, and yet S may consider the extent to
which reasons that provide putative support for the proposition <God exists> are
adequate. As we have seen, the adequacy of evidence on the internalist view is most
plausibly construed as a function of whether the evidence is justified, the extent to
which the evidence makes h probable, and what kinds of inductive standards are being
employed. Cognitive reflection on adequacy from the internalist perspective will be a
reflection on these three general considerations. And the adequacy of one’s
investigation and critical assessment of these factors may be judged in terms of
subjective or objective adequacy, thereby yielding multiple levels of (diachronic and
synchronic) reflective rationality.
Clearly some of the propositions (1)-(11) which constitute potential evidential
bases for theistic belief are uncontroversial and there can be little doubt about a
person’s being justified in accepting them, even with a very high degree of
confidence. That there exists a complex physical system which conforms to relatively few,
simple scientific laws is believed and accepted by nearly everyone, even if there is some
debate regarding whether the Universe has existed for finite or infinite time. Other bits
of evidence are philosophically controversial and the question of whether a person is
justified in holding them is more pressing. Here I am thinking of beliefs about humans
possessing indeterministic free will and beliefs that entail moral realism. Still more
controversial are empirical phenomena some people take to be miracles. Even if we suppose
that such evidences are believed by the person and contribute to the probability of there
being a God, investigating the extent to which one is justified these beliefs (which
constitute one's alleged evidence for the existence of God) usually will require the
process of diachronic reflective rationality, especially when a person realizes the
controversial nature of such evidence.
There are of course several matters that are relevant to the adequacy of such
investigation. For instance, the level of importance of the thesis that the evidence e is
supposed to support, how much the probability of h depends on e being true, how widely
accepted e is, whether S has good reasons to doubt the truth of e, whether e is
representative evidence, etc. A person who, for instance, attends a faith healing service
and sees various phenomena he immediately characterizes as “healings” could and
perhaps should ask himself whether he is justified in believing this. This will be
especially so if he knows that there are many charlatans who conduct "healing
services" in order to bilk simple folk of their earnings, or that the excitement of
religious meetings easily leads people to make incorrect judgements about their physical
condition, etc. Similarly, a person who takes answered prayer as evidence of the existence
of God might have forgotten many instances in which nothing came to pass in response to
her prayers and thus her evidence may not be representative. On the best case scenario, a
person could investigate the justification-of-evidence matter adequately given objective
standards. More frequently, though, a person will judge the adequacy of his investigation
by standards that are not all correct. Nevertheless, we must appreciate that it is one
thing to have evidence for the proposition that <there is a God> and it is another
matter to investigate, to varying degrees, whether one is actually justified in holding
such evidence. To do the latter is to embark upon an important aspect of internalist
reflective rationality. And it is certainly a part of the project of natural theology to
consider the extent to which some of its more controversial premises are justified.
In addition to investigating the justificational status of our evidence-base, we
can engage in critical reflection on the “makes probable” relation. Since S can
inquire into the extent to which evidence e renders h probable (or increases h’s
probability) given S’s own inductive standards or correct ones, the critical
assessment of the probability relation between e and h is really a two-pronged matter.
There is the issue of whether h is probable on e given the sorts of inductive criteria
which guides S’s normal inferential practices, and then considering whether these
standards are correct. Some degree of caution
is required here, though, as a person’s normal inferential practices frequently
involve inductive standards which are restricted in application to mundane matters and so
do not admit of application to the question of whether there exists an essentially
incorporeal, omnipotent, omniscient being. A person who is reflecting on whether the
existence of the Universe and its temporal regularities
make it probable (or increases the probability) that God exists by his own
inductive standards must sufficiently distinguish the particular standards he employs (say
in empirical matters) from more general standards which have cross-area application, and
from which more particular criteria may be derived and which would be applicable in the
religious sphere. But having reached some understanding of what his general inductive
principles are, he can turn to the question of whether given these standards e renders h
probable (or increases h’s probability). Where his inductive standards involve, say,
explanatory relations of some sort, he will check to see that theism has the appropriate
sort of explanatory power relative to the evidence in question and perhaps whether any
other proposition has this sort of explanatory power. This may involve a careful analysis
of just what it is <there is a God> means, as well as what features of the evidence
require explanation in the ways in which S is thinking of explanation.
The particulars here are better spelled out if we have a definite set of inductive standards with which to work. For this purpose I will follow Swinburne’s treatment in The Existence of God (1979) and adopt the principles of confirmation theory expressed in terms of Bayes’s theorem. Let ht = <there is a God>, e = evidence, and k = background knowledge. The probability of ht on [e and k] is a function of the prior probability of ht <there is a God> (how likely ht is independent of the consideration of the evidence e) and the explanatory power of ht. The latter depends on the prior probability of e (how likely is whether or not ht is true) and the predictive power of ht (how likely e is given ht). The explanatory power of ht increases as its predictive power increases, and it decreases with the greater the prior probability of e. Reflective rationality within such a framework requires an assessment of the relevant prior probabilities (of e and h) and the extent to which we would expect the phenomena in e if ht is true.[14]
Checking one’s assignment of prior probabilities will include an analysis of
what one takes to be the determinants of prior probability (such as scope, fit with
background knowledge, and simplicity), as well as which of these determinants is most
appropriate for assessing the prior probability of theism. Swinburne, for instance,
emphasizes the purely a priori category of
simplicity. He holds that where the scope of empirical data to be explained is exhaustive,
background knowledge is mere tautological evidence and intrinsic probability is the more
relevant notion, and simplicity is the main determinant of intrinsic probability. A person
for whom the probability of theism depends on simplicity will have to evaluate what
simplicity amounts to (e.g., the postulation of very few mathematically simple laws
holding between few entities of a very intelligible kind), as well as how simple the
theistic hypothesis is on that view. The latter requires a critical grasp of the divine
attributes and their relations to each other (e.g., in what ways some of the divine
properties may be deduced from more simple properties), especially how the attributes
constitute a unity. The simplicity of theism
will also be contrasted with the simplicity of other hypotheses of what exists. In evaluating how probable e makes ht, one will
have to decide whether the predictive power of theism relative to that evidence is high
because the phenomena in the evidence base would be very likely if God existed (as seems
to be the case where e = temporal regularities and beauty) or because the evidence base
would be very unlikely to occur if God did not exist (what is sometimes claimed for the
existence of the Universe). Considerations of the divine nature (such as God’s
goodness) may suggest that, if God creates a world, He has overriding reason to create one
of the sort that actually exists (one which exhibits order and beauty). And the
probability of there being a Universe given the existence of God may be high not because
God has overriding reason to create it but because its existence is inexplicable except we
suppose that it is brought about by the action of God (i.e., the prior probability of
<there is a universe> is very low).
Moreover, given that often the evidence base for theism will consist of several
bits of evidence, say the conjunction of some of (1)-(11), there is the matter of S’s
sorting out these relevant bits of evidence and considering the contribution each makes to
the probability of theism. It may be that individual bits of one’s evidence increase
the probability of theism, and that the conjunction makes theism more probable than not.
Determining whether <there is a God> is probable on e given one’s inductive
standards will often call for assessing how one’s inductive standards are applicable
to each bit of evidence e1, e2, e3, . . .and then to the
conjunction of all the bits {e1, . . . ,en}. Clearly, this task
admits of degrees, and one will judge the adequacy of his investigation to differing
degrees. But if S’s being reflectively rational about the belief that God exists
requires checking to see whether his evidence makes the proposition probable, it is going
to require analysis of the kinds of inductive principles and their application to bodies
of (possible) diverse evidence as I have outlined.
What the preceding account suggests is that natural theology is a multi-level
cognitive enterprise with both diachronic and synchronic dimensions. It does more than
state the kinds of evidence people may have for theistic belief, and which - given the
possession of certain inductive standards - confirms or makes probable God’s
existence. It is equally concerned with subjecting to critical analysis the relevance,
credibility, and force of that evidence, and the latter will obviously depend on the
nature and application of our inductive criteria (and their correctness). It is only by
considering whether we are justified in holding our evidence (with the degree of
confidence we do), critically examining how are standards are applicable to theistic
belief, and ultimately examining our reasons for thinking that our inductive standards are
the correct ones, that allows us to form justified beliefs about the adequacy of our
evidence for the proposition <there is a God>. In other terms, reflective
rationality involves reflecting on first-order evidence, and that involves raising
second-order questions. The distinction between first- and second-order evidence is just
the distinction between what adequately supports p and what adequately supports <the
evidence adequately supports p>.[15] As
explained earlier, on the internalist view the rationality of S’s belief that p will
depend on p’s being made evidentially probable by e given S’s inductive
standards, where e consists of other propositions S believes. S will be justified in
believing that p if his reasons for p do make p probable. But S will be justified in
believing that <his belief that p is (propositionally) justified> only if S has
(good) reasons for regarding the evidence as in fact rendering p probable given the
evidence and S’s inductive standards. This, as I have tried to show, means
investigating and critically evaluating the evidential probability making features by
which p is made probable and so may be regarded as propositionally justified.
IV.
Conclusion
If my analysis and arguments in this chapter have been approximately correct, then
though S may be immediately justified in some theistic belief that Pt (even where
“justification” is understood in the externalist sense), S may engage in the
project of natural theology as an exercise in propositional reflective rationality -
examining to various degrees the extent to which the arguments of natural theology provide
adequate evidential support for the proposition <there is a God>. In this way S may
become justified in the belief that <there is a
God is propositionally justified>. Whether, and how much, this affects the doxastic
justificational status of S’s belief that Pt will be discussed in the next chapter.
What should be clear is that there is a distinction between putative beliefs that Pt and
beliefs about the justificational status of beliefs that Pt. Furthermore, there is a distinction between being
in the latter state and regulating one’s cognitive life through epistemically
oriented activities of critical investigation and reflection which place one in a good
position to be justified in the kinds of higher-level theistic propositions I have
introduced in this chapter. The project of natural theology, then, provides a framework
for both unreflective and reflective rationality.