PROBLEMS

WITH THE BOOTSTRAPPING

OBJECTION TO THEISTIC ACTIVISM
Christopher Menzel

Introduction
Many traditional theists are deeply committed to the doctrine of divine aseity the
doctrine that God alone exists a se; all other things exists ab alio (Craig 2014, p.
113). More specifically, the doctrine has two components: (i) Gods existence is
independent of all other things, and (ii) all other things depend on God God both
creates all other things and sustains them in existence.1 On the face of it, divine
aseity, so understood, is inconsistent with classical Platonism, i.e., the view that
there are objectively existing, abstract objects numbers, pure sets, general
properties like wisdom, redness, etc. For according to the classical Platonist, at least
some abstract entities are wholly uncreated, necessary beings and, hence, as such,
they also exist a se; like God, they depend on nothing for their existence.2
The thesis of theistic activism purports to reconcile aseity with a robust
Platonism (see Morris and Menzel 1986, Menzel 1987). Specifically, the activist
holds that God creates the necessary, abstract objects of traditional Platonism no
less than the contingent concrete objects of the physical universe3 and hence that,
like all created things, they exist ab alio after all, their necessity notwithstanding.4
But many theists believe a severe roadblock for activism remains a problem
known as the bootstrapping objection that (so it is purported) renders activism
hopelessly incoherent (Gould 2011, p. 259). The locus classicus of the objection
indeed, it appears to be the only detailed exposition of it is a 2006 paper by
Michael Bergmann and Jeffrey Brower (henceforth, B&B). Despite widespread faith
in the deliverances of this argument, I will show that it is open to significant
objections on several fronts.

1. The Bootstrapping Objection

William Craig (ms, p. 4) gives us the bootstrapping objection in a nutshell: [I]n
order to create properties, God must already possess properties, which involves a
vicious circularity. In our 1986 paper, Morris and I (henceforth M&M) explored a
simple version of this objection and found it unconvincing. More specifically, as we
saw it, the bootstrapping objection depends on an unwarranted conflation of
existential and causal dependence to deliver the alleged circularity the
metaphysical impossibility that God is causally dependent upon God. Say that is
existentially dependent on if and only if s existence entails s, i.e., if and only if,
necessarily, exists only if does,5 and that is causally dependent on if and only
if (crudely) has caused it to be the case that exists. Now, let be Gods nature,
divinity. According to activism, since God creates all the properties and is a
property, God creates . Hence, is causally dependent upon God. At the same time,
as is a necessary being, it is necessarily the case that God exists only if does.
Hence, God is existentially dependent upon Gods own nature . But the conclusion
that God is causally dependent on God follows only if one adopts a bridge principle
asserting that transitivity holds across these two very different types of dependency
that is, a principle asserting that, from the fact that is causally dependent on
and is logically dependent on , we can infer that is causally dependent on . But
there is simply no reason whatever to think that such a principle is true. Thus, M&M
conclude (ibid., p. 360):
It just seems to us that there is nothing logically or metaphysically
objectionable about Gods creating his own nature in precisely the way
indicated.
In fact, B&B find nothing awry in the reasoning that leads M&M to this
conclusion; the conclusion itself however, B&B claim (2006, p. 364), is a non
sequitur:

[w]e suggest that the reason it just seems to [Morris and Menzel] that there is
no objectionable circularity is that [they arent] clear enough about precisely
what the objectionable circularity is.
That is, B&B argue that M&M missed the fact that there is a further fundamental
dependency relation beyond existential and causal dependency that they call logical
posteriority. The real bootstrapping objection involves this variety of dependence
(and its converse logical priority). They sketch the argument thus (ibid., p. 366):
If...theistic activism is true, then every property...will be a product of Gods
creative activity. But this implies the general principle that, for any property
, Gods creating is a prerequisite for, and hence logically prior to, .
Notice, however, that in order to create , God must have the property of
being able to create a property. Here is where the trouble begins. For on the
one hand, it would seem that this property (i.e., being able to create a
property) must be logically prior to Gods creating it, since Gods having it is a
prerequisite for the creation of any property. On the other hand, however, it
would also seem that this property must be logically posterior to Gods
creating it, since insofar as it is a property..., it must...be a product of Gods
creative activity. Evidently, therefore, in order for it to be true that God is the
creator of all properties, there must be a propertynamely, being able to
create a propertythat is both logically prior and logically posterior to Gods
creating properties. Assuming that logical priority is an asymmetrical
relation, however, this conclusion is obviously absurd.
B&Bs bootstrapping objection henceforth, BBBO is subtle and interesting
but, as noted, is vulnerable to a number of important objections. My criticism is two-
pronged. First, BBBO is very informal, and this hinders comprehension. I will
provide a logical framework that I think makes the logic of the argument much
clearer. This in turn will uncover a problem that, while perhaps not insurmountable,
at least reveals that BBBO rests upon shaky and incomplete foundations. I will then
turn to two more philosophically substantive objections. Specifically, I will argue (a)
that there are no cogent grounds for taking the relation of logical priority that is so

central to BBBO to be a genuine, fundamental relation and (b) even if we admit such
a relation exists, there are strong reasons for thinking it is not asymmetrical.

2. A Logical Framework for BBBO

Logical priority rather than its converse logical posteriority takes center stage in
B&Bs detailed version of BBBO. I will represent it with the less-than predicate <
a natural enough symbol for one thing coming before another in order of logical
priority. Now, two significant questions are: what sorts of syntactic entities are the
fundamental arguments to this predicate and, relatedly, what are the relata of the
corresponding relation? In B&B , one finds both sentential gerunds (S-gerunds) like
Gods creating and verb phrase gerunds (VP-gerunds) like being able to do
flanking is logically prior to. But one also finds ordinary singular terms like God,
the thinker, the variable , and even plurals like its parts. However, I think it is
reasonably clear that S-gerunds are the fundamental arguments of is logically prior
to and that we can reduce all the other cases to those. Specifically, when is
anything other than an S-gerund, we can take is logically prior to and is
logically prior to to mean s existing is logically prior to and is logically
prior to s existing. The answer to our second question, then, depends on what we
take S-gerunds to denote. I will follow B&B here and take them to denote states of
affairs. B&B do not address the metaphysics of states of affairs and, on the face of it,
their exact nature is not central to the argument. For the sake of clarity, I will make
use of square brackets around English gerunds in my natural language rendering of,
and discussion of, BBBO. This is especially useful when a VP-gerund is embedded
within an S-gerund, as in Gods exemplifying being able to create, which I will
write as [Gods exemplifying [being able to create]].
To represent BBBO formally, I will use a flexible second-order language (call it
) in which predicates also count as singular terms that can occur as arguments in
atomic formulas, as the argument seems to require and certain predicates play both
syntactic roles.6 For purposes here, the only second-order variables that seem to be
required by the argument are 1-place variables , G, H, ; intuitively, of course,

these range over properties. It is also convenient to introduce special variables ,

q, r, ranging over states of affairs (as opposed to a dedicated predicate
expressing the property of being a state of affairs); the exact syntactic classification
of these variables is an issue that will arise below. To represent S-gerunds formally,
I will use square brackets as term-forming operators so that, when a formula
represents a sentence of natural language, [] represents the corresponding S-
gerund. Thus, s being [or exemplifying] will be represented as []; and s
existing will be represented as [! ]. To express complex properties, I will use
familiar -notation. Thus, in particular, being able to do , where indicates some
action property, will be represented as [] read literally: being an such
that is able to do . (It is important to note that nothing deep is built into the use
of the standard possibility operator to capture the modal element of being able to
perform some action. I am in particular not meaning to imply that the power of to
do is to be analyzed simply as the possibility of doing . For purposes of the
argument, all that is needed is that the contribution of is able to in is able to do
, however construed exactly, is represented so that being able to do and
doing can be taken to indicate distinct properties.) This plays out in my
representation of BBBO as follows. Note first that, although BBBO as B&B present it
involves a binary depends on the creative activity of predicate as well as creates,
it seems to me they only introduce the former for expository reasons and that
creates is all one needs. I will represent it with the binary predicate .7 In the
case of the property [being able to create a property], i.e., [being able to create some
property or other], then, I will let , when used as a 1-place predicate, stand for
the complex property [creating a property], i.e., =!" [].8 Accordingly, I
will represent [being able to create a property] as []. (Henceforth I plan to
be more careless about quotes and will usually leave it to context to establish
whether formal expressions are being used or mentioned.)

3. The Argument
We begin our formalization of BBBO with five general principles that B&B invoke as
justification for several premises and inferences in the argument. In fact, I dont
think all of these assumptions are needed; in a couple of cases, I think the general
principles are more dubious than the instances of those principles that are used in
the argument. Nonetheless, I present all five in order to remain as faithful as
possible to the form of BBBO as presented in B&B .
A1 For any property , if God creates , then [Gods creating a property] is logically
prior to (i.e., to [s existing]).
( [] < )
A2 For any and any (action) property , [s being able to do ] is logically prior
to [s doing ].
([] < [])
A3 For any , if [s exemplifying ] is logically prior to , then is logically prior
to .
([] < < )
A4 For any and any (action) property , [s being able to do ] = [s
exemplifying [being able to do ]].
([] = [[]])
A5 Logical priority is asymmetric.
( < )
Now for the argument.9
1.

God creates all the properties.

2.

For any property , [Gods creating a property] is logically prior to .

[] <

3.

(Assumption for reductio)

(From 1 and A1)

[Gods creating a property] is logically prior to the property [being able to

create a property].
[] < []

(From 2)

4.

[Gods being able to create a property] is logically prior to [Gods creating a

property].
[] < []

5.

(From A210)

[Gods exemplifying [being able to create a property]] is logically prior to [Gods

creating a property].
[[]] < []

6.

[Being able to create a property] is logically prior to [Gods creating a property].

[] < []

7.

(From 5 and A3)

Contradiction.
<

8.

(From 4 and A411)

(From 3, 6, and A5)

God does not create all the properties.

(1-7, by reductio}

4. Problems with the Argument

The above argument is valid and, I believe, it is faithful to BBBO. However, the
argument suffers from three problems that call its soundness into question. I discuss
these in detail in the following three subsections.

4.1. Problems for Principle A4

Fundamental to B&Bs justification for BBBO and, in particular, for principle A4, is
the claim that theistic activisms commitment to Platonism is motivated by the
desire for a unified account of predication in terms of exemplifiables, as expressed
in the following principle (362):
P

The truth of all true predications, or at least of all true predications of the form
is , is to be explained in terms of a subject and an exemplifiable (however
exemplifiables are themselves to be conceived).12

As I understand it, P is a sort of comprehension principle. Comprehension principles,

recall, in their property theoretic guises, essentially tell us which of the conditions
formulable in a given language express properties thus the connection to

Platonism. Such principles typically take the following form, for some class of
formulas of a given (second-order) language :
( ), for any formula in which does not occur free.
Now, in fact, we have in effect built the most liberal comprehension principle
possible into our logical framework (or, at least, we have not ruled it out) insofar as
weve put no explicit restrictions on the creation of -predicates. For all we have
said, given any arbitrary formula of our language , there is a -predicate [].
Given the standard -conversion principle governing the logic of complex
predicates, i.e.,
LC

[] !! ,13

the full second-order comprehension principle for is immediate:

( ), for any formula of in which does not occur free.
However, as B&B realize,14 and as the history of logic in the 20th century has taught
us, the injudicious introduction of comprehension principles can quickly lead to
paradox. And that is indeed the case here: the flexible language that seems to be
required for BBBO contains the formula ( = ) and hence, the -
predicate [( = )] denoting, intuitively, the property of non-self-
exemplification. Given LC, a property-theoretic version of Russells Paradox follows
straightaway.15
Clearly, then, if the activist is committed to P and hence, formally, to some sort of
comprehension principle, the formation of -predicates must be restricted to a
proper subset of the formulas of . Assuming that a predication is is
represented as , then the simplest rendering of principle P takes simply to
be the class of primitive monadic atomic formulas of , i.e., those atomic formulas
formed from s primitive monadic predicates. This, then, yields all -predicates of
the form [], for variables and primitive monadic predicates .16
Of course, though, this wont begin to get us what we need for A4. In particular,
the instance of A4 used in step 5, viz.,

(1)

[] = [[]],

involves two properties, and [], whose logical forms are non-atomic and
hence whose existence, if it is to be justified by a comprehension principle, requires
to encompass a much broader class of formulas of than the primitive atomics.
And, indeed, there are many sentences of the form is where the is itself
logically complex, e.g., John is wealthy but unhappy and God is able to create a
property. Ive argued above that the logical form of the latter sentence is in fact that
of a predication qualified by an intensional operator , God could create a
property. However, by means of the (somewhat stilted and artificial) such that
construction, most any singular assertion ...... can (with proper care re
consistency) be transformed into an equivalent predication is such that
...he/she/it...: from God created a property we have God is such that he created a
property and from God could create a property we have God is such that he could
create a property, which, given P, entail the existence of the properties (=
[]) and [] that appear in (1).
But note that the preceding observation is reversible: any predication involving
a logically complex predicate is such that ...he/she/it... can be transformed into
an equivalent, logically complex sentence ...... in which the complex predicate
...he/she/it... is eliminated. (These two observations together, of course, constitute
the two directions of (the ordinary language reflection of) the -conversion
principle LC.) Given this, the activist seems in no wise compelled to accept the
existence of logically complex properties. For, as far as I can see, the activist is
committed only to Gods having created some set of fundamental properties and
(more generally) relations, whatever they may be. Good candidates might be
fundamental physical properties, universals like wisdom, loving, and the creating
relation, mathematical properties like being a set, perhaps also mental properties
and relations like thinking, being conscious, etc. logically simple (let us suppose)
properties and relations of a sort that we can assume are expressed by the primitive
predicates of our language . Since, as just noted, predications involving logically
complex predicates can always be unpacked in terms of the logically simple

predicates from which they are built up, it is far from obvious that the sorts of
logically complex properties allegedly expressed by those complex predicates are an
essential part of the activist picture. Thus, with respect to the argument proper, the
activist holds that God created the fundamental properties and, hence, that God
created some property or other, i.e., that . It does not obviously follow from
this at all that there must also be a corresponding property [creating a property],
[] (a.k.a., ) and thence a modal property [being able to create a
property], [] that God exemplifies. No general comprehension principle
that entails the existence of logically complex properties in particular, the
properties and [] involved in the relevant instance (1) of principle A4
seems to follow from the basic activist picture of the creation of the fundamental
Platonic properties and relations that the activist is committed to.
Contrary to B&B, then, Principle P, understood so as to entail commitment to
logically complex properties, does not seem to be a part of the basic activist picture.
Granted, it could be that an activist might be motivated by the desire for a unified
semantics of predication. But there are several other motivations than P for wanting
abstract entities in ones ontology. For instance, ones motivation might simply be
the classic Platonic explanation of the one over the many in terms of shared
universals. If that is ones motivation, there is no obvious need to reify the unique
and essentially non-shareable properties of the Godhead like [].17 Indeed,
when it comes to such alleged properties, there seems no reason why an activist
with this motivation might not opt for B&Bs (2006, Part II) own truthmaker
account of the semantics of divine predication. Commitment to principle P,
therefore, appears to be entirely avoidable.18 For activists that reject it, the move
from step 4 to step 5 in the argument is undercut, as the existence of the requisite
properties needed for the instance (1) of A4 can be denied.
That noted, a number of robust varieties of Platonism embrace logically complex
properties (e.g., Bealer 1982, Zalta 1983 and 1988, Menzel 1993). Hence, some
activists might consider the preceding response a Pyrrhic victory if the cost is that
logically complex properties are in principle unavailable to them. So suppose an
activist has identified a paradox-free class of formulas that permits the -

predicates expressing the properties and [] in (1).19 Even granting the

existence of those properties, nothing follows by way of justification for A4
generally or for (1) in particular. For B&B specify no criteria of identity for states of
affairs and hence provide no warrant for the claim that modal states of affairs of the
form [] are identical to predicative states of affairs of the form [[]].20
They are of course logically equivalent21 indeed, B&B (ibid., p. 369) themselves
refer to A4 as an equivalence, its actual form notwithstanding but this will not do:
if A4 is weakened to logical equivalence, step 5 will no longer follow directly from
step 4 simply by the logic of identity; and the obvious logical principle that would
warrant step 5 under A4 so weakened viz., the principle that states of affairs are
identical if logically equivalent is one the activist would almost surely reject as
too coarse-grained. (It would follow, for example, that [Gods either existing or not
existing] is identical to [Somethings being wise if Socrates is].) But even if the
identity in question is somehow justifiable, the larger point here is that such
justification is absent from BBBOs logical foundations. Hence, A4 as it stands is ad
hoc and, pending justification, is reasonably rejected by the activist. However, even
if ample justification for A4 is available, serious objections to BBBO remain. To those
I now turn.

4.2. There is no Fundamental Relation of Logical Priority

Essential to BBBO is the claim that logical priority is a fundamental relation distinct
from existential and causal dependence. In this section I will argue that there simply
is no such relation at least, no metaphysically fundamental relation of the sort
that B&B need.
The best way to clarify the notion [of logical priority], B&B tell us, is by way of
example. They give us three (ibid., pp. 367, 370):

The priority of the parts of a whole to the whole.

The priority of a thinker to her thoughts.

The priority of the constituents of a state of affairs to the state of affairs.

Presumably, then, we are to intuit in these examples a single fundamental relation of

logical priority <. Critically, as B&B note (p. 368), because < is asymmetrical, it
cannot be identified with the relation [being a necessary condition of] or with
existential priority (as defined above) as these both have counterexamples to
asymmetry (notably, when their relata are both necessary beings22). Moreover, it
cannot be the relation of temporal priority, as an eternal whole would be co-eternal
with its parts but its parts would still be logically prior to it. As for positive
characteristics, B&B (p. 369) tell us that perhaps the most illuminating is that,
when < , depends on in such a way that [i.e., I take it, s existing] at least
partially explains [i.e., s existing] whereas is not even a partial explanation of
.
I certainly see the priority of parts to wholes, thinkers to thoughts, and
constituents to states of affairs. And I see, in each case, that the priority in question
seems to have the characteristics of logical priority that B&B suggest. Intuitively, it
is in the nature of parts and wholes that, if is a (proper) part of , s existence
depends on and, hence, is partially explained by, s existence in a way that s
existence is not explained by s. And it seems that a similar tale can be told about
states of affairs and their constituents, as states of affairs likewise seem to consist of
their constituent parts plus structure. And thinkers clearly explain the existence of
their thoughts in a way that thoughts do not explain the existence of their thinkers
for thinkers create or generate their thoughts and not vice versa.
I do not, however, perceive the distinct, common, fundamental relation of logical
priority that B&B allege. At the very least, there is no compelling reason for
introducing one. Rather, by my lights, the examples simply illustrate three distinct,
asymmetrical priority relations: <! , <! , and <! . There is of course a common, non-
fundamental priority relation here as well, viz., the disjunctive relation
[ <! <! <! ]. But that, of course, will not do for the purposes of
BBBO, as all three differ considerably from the notion of logical priority found in
B&Bs argument, specifically in principles A1-A3. As I see it, then, B&Bs examples
are at best evocative. They provide us with three different priority relations that

only bear a certain family resemblance to one another but which are not species of a
more fundamental relation.
It seems to me, then, that the best B&B have to offer is a hypothesis, viz., that
there is further priority relation, <, whose relata can be (states of affairs involving)
agents (notably, God), actions (notably, creation), and properties, and, moreover,
that this relation bears the same family resemblance to the relations noted, in
particular, in being asymmetrical. Finally, according to the hypothesis, we grasp this
relation with sufficient clarity as to be able, at least, to find the truth of principles
A1-A3 reasonably compelling. So it seems to me that we simply need to evaluate
these principles on their own merits. That is, primed by the family of relations
discussed in the examples, we need to evaluate the hypothesis in question and see
whether those principles are plausible for some appropriate notion of priority.
And here is where the trouble begins. For my part, I would agree that there are
notions of priority under which all three principles A1, A2, and A3 are true.
However, it seems quite clear that, like the ones identified in B&Bs examples, they
are different and, moreover, already well-understood notions of priority. A1 is the
priority of, specifically, an act of creation to the thing created; the act certainly
seems to explain the thing created in a way that the thing created does not explain
the act. But this seems like nothing other than the asymmetry of causation the act
is the cause of the thing, but not vice versa. A2, by contrast, involves a sort of
priority relating abilities and actions: an agents ability to perform some action at
least partially explains the existence of the action the action exists because some
agent with the ability to perform it exists. But the agents ability is not explained by
the action, especially if the action is free. So again we do seem to have a reasonable
notion of asymmetrical priority here, but it is in no obvious way at all the same sort
of priority as the causal priority of acts of creation to things created in A1. And
neither notion of priority warrants principle A3. For there is nothing about the
priority of creative acts to their products or the priority of abilities to actions that
seems in any clear way to bear on the truth of the claim that, for arbitrary properties
and states of affairs , the property constituent of a predication [] is prior to a

state of affairs if the predication itself is prior to . If anything, this is a further

notion of priority related to constituency.
No two of the principles A1-A3, then, appear to involve the same notion of
priority and, as in the examples B&B provide, there is nothing whatever to suggest
there is any more fundamental priority relation they all exhibit. If so, then BBBO is
rife with ambiguity and the only way to preserve its validity would be to invoke a
series of ad hoc principles relating the three very different notions of priority in the
argument. This would leave the argument explicitly open to the charge M&M
initially leveled against the bootstrapping objection, viz., that it runs together
different notions of priority/dependence that can only be connected by means of
implausible bridge principles.

4.3. Counterexamples to the Asymmetry of Logical Priority

But let us suppose that the problem of ambiguity can be met and that B&B give us
strong reasons to think that there is a fundamental priority relation underlying the
examples they provide and that that relation is plausibly the one at work in BBBO. In
this final section I point out that there are examples of the very sort that B&B draw
upon to pump our intuitions about dependency where the dependency in question
is in fact mutual and, hence, that the (incorrectly and tendentiously named) logical
priority relation is not, in general, asymmetric. I will provide three counter-
examples.
Non-well-founded Sets. I doubt that B&B would balk at the idea that set
membership provides a good example of logical priority. Pure sets constitute a
particularly cogent case for distinguishing the dependence associated with logical
priority from mere existential dependence. For, as classic necessary beings, every
pure set is trivially existentially dependent on every other pure set and, in
particular, the members of a pure set are as existentially dependent on as is on
them. But the axiom of Extensionality for sets appears to provide a clear sense in
which a set is asymmetrically dependent on its members (and hence its members
asymmetrically prior to ): by Extensionality, s identity is completely determined
(collectively) by its members, but the identity of its members is not determined by

. Hence, each of s members at least partially explains it is one of the objects

that determines its identity but not vice versa. Hence, sets appear to provide a
fine example of asymmetrical priority.
But this is too quick, for the preceding argument for asymmetry depends on the
assumption that set membership is well-founded, and it is a well-known fact that the
axiom of Foundation in, say, ZFC set theory can, without any loss of consistency, be
replaced by an anti-Foundation axiom notably, the axiom AFA of Aczel 1983
that asserts the existence of non-well-founded sets.23 In the resulting set theory, in
addition to all of the well-founded sets of ZFC, there is a vast variety of non-well-
founded sets. In particular, there are sets , such that and , e.g., let
= , and =

, .24

As Extensionality holds in both well-founded and non-well-founded set theory,

non-well-founded sets are still determined by their members and, hence, are at least
partially explained by them. However, consider the sets and above. As
= , , depends on, and is partially explained by . But, as =

, ,

depends on, and is partially explained by . Hence, notwithstanding the intuitions

that arise when we restrict our gaze to well-founded sets, by broadening our
perspective beyond the well-founded universe, the priority of the members of a
set to the set itself is not, in general, asymmetric.25
Constituency, Common Knowledge, and Self-awareness. As noted above, B&B
suggest that states of affairs/propositions have constituents. I like this idea, too.
And, as with sets, if we focus on common examples, constituency might appear to
indicate a type of asymmetrical dependence. However, some states of affairs seem to
require themselves as constituents. Among these are states of affairs involving
common knowledge where two or more agents are aware of a set of facts, among
which is each agents awareness of those very facts. Consider for example a game of
5-card stud involving two players and in which each player can see four of the
other players cards. Let be s visible hand and be s and let be the having
relation between players and hands. Within the overall game situation, then, there is
a state of affairs involving not only the facts [] and [] but also, critically,

the facts of each players awareness of all of those facts, i.e., of each players
awareness of , and, hence, each players awareness of the others awareness of .
This reflexivity is most naturally captured by taking to include itself among its
constituents; specifically, where is the awareness relation:
(2) = [ ].
This simple, intuitive representation of the common knowledge in the game
situation yields the sort of inferences we expect. For example, on the assumption
that awareness of a conjunctive state of affairs entails awareness of each conjunct, it
follows from s awareness of , , that, e.g., is aware that is aware that (i) s
hand is and that (ii) is aware that is aware that s hand is , [[
[[]]]].
Cartesian self-awareness seems to fit the same pattern in an even simpler and
more direct way. Descartes breakthrough involves his awareness of the fact of his
own thinking. But that he is simply thinking something, [], wont capture the
reflexive self-awareness underlying the cogito. Rather, Descartess breakthrough
comes in his awareness that that very fact, i.e., that he is thinking, is the object of his
thinking. And, as in the preceding example, the most natural way of characterizing
this propositional object of his thought is for it to have itself as a constituent.
Specifically, where is the thinking relation, is Descartes, and is the object of
Descartess thinking, we have:
(3) = []
Similar to extensionality for sets, then, it is intuitive that states of affairs, qua
structured wholes, depend on their constituents and, hence, that the constituents of
a state of affairs partially explain it. Hence, since is among its own constituents, it
depends upon itself. As with non-wellfounded sets, consideration of a broader class
of states of affairs reveals that the priority of constituents to states of affairs is not,
in general, asymmetrical.26
Numbers and the Natural Number Structure. Finally, consider the system of
natural numbers, that is, the collection of natural numbers in their natural ordering

0, 1, 2,; call it ! . On the one hand, like other complex systems, ! is dependent
on its parts, i.e., the individual natural numbers intuitively, ! is a complex
whole constituted entirely by its parts + structure. Hence, like sets and states of
affairs, it is at least partially explained by its parts ! is nothing more than the
natural numbers structured in a certain way.
However, on what is arguably the most compelling realist account of the natural
numbers, viz., the ante rem structuralism of, notably, Shapiro 2006, the natural
numbers of the system ! are themselves nothing but the places or positions in
! . Thus, as we might put it, the sum parts + structure in the case of the system of
natural number is redundant the parts, the individual natural numbers, being
positions in the overarching structure, are nothing over and above the structure.
The number 2, for example, just is the third position in the natural number structure
(counting from 0). It has no nature independent of that structure; the structure is
what makes it what it is. As Shapiro (p. 120) puts it: the essential properties of a
natural number consist of its relations to the other natural numbers. Hence, the
parts of the system ! depend just as much upon the system as the system
depends upon them. Once again, an example in the same ballpark as those B&B
draw upon to illustrate logical priority indicate that, insofar as there is any such
relation at all, it is not in general asymmetric.
Upshot and Reflection. Of course, the existence of these counterexamples to the
asymmetry of any purported logical priority relation does not entail that there
isnt a species of priority that is asymmetrical. However, I believe they do undermine
B&Bs inference from their examples of their alleged logical priority relation to their
claim that that relation is asymmetrical, as the counterexamples here are very much
in the same conceptual neighborhood. For them to assert, in light of these
counterexamples, that the species of logical priority in BBBO the existence of
which weve already found solid independent reasons to reject is asymmetric
appears to question-begging. I conclude that A5 in BBBO has been undermined.
I close with a quick reflection. Well-foundedness has a strong intuitive pull, so
much so that there seems to be a natural philosophical resistance to most any threat

of circularity. I think a useful heuristic function of the above examples is to

undermine the intuition that circularity is inevitably vicious. Were inclined to think
that wholes (of any sort) are dependent upon their parts but not vice versa. More
generally, we are inclined to think that if is in any way dependent upon , then
cannot be in any way dependent upon . And so it often is but not always. At the
ontological margins we find counterexamples. Overgeneralizing from familiar well-
founded examples can make us blind to them, at the risk of missing important
theoretical insights. There is no time here to explore the possibility that new
theological insights lurk in the circularity of dependencies between God and the
properties that activism seems to embrace. But it seems to me that the latter-day
theoretical embrace of non-well-founded phenomena raises the prospect of hitherto
unrecognized avenues for the activist to explore.
Texas A&M University

Notes
This paper developed out of comments I delivered in response to a talk by William
Lane Craig at the Philosophy of Religion group meeting on 27 February 2014 at the
APA Central Division meetings in Chicago. My sincere thanks to an anonymous
referee for cogent, detailed comments on the original version of this paper. The
current version is much improved because of them.
1. Some philosophers reserve aseity for the first component here only. Thus,
Bergmann and Brower (2006, p. 358) refer to (i) and (ii) together as the aseity-
dependence thesis.
2. The qualification at least some is due to the fact that there might be impure
abstract entities like, say, the singleton set Socrates . If Socrates hadnt existed,
neither would Socrates have existed (as it would have been the empty set and,
hence, by the axiom of Extensionality, distinct from Socrates ). Hence, Socrates is
both contingent and, insofar as it depends for its existence on the existence of
something that exists ab alio, it also exists ab alio.


3. Craig (ms) helpfully distinguishes between what he calls conceptualist and non-
conceptualist forms of the thesis. On the conceptualist version, abstract entities exist
in the divine mind they are literally the products of divine intellective activity; as
Morris and I (1986, p. 355) put it, properties are Gods concepts, propositions are
Gods thoughts. On the non-conceptualist variety, abstract entities are merely
products of Gods creative activity but do not exist in the divine mind, at least, not as
divine concepts and thoughts. The advantage of the conceptualist variant lies in its
specificity in identifying abstract entities with ideas in the divine intellect, it
broaches the question of the nature of their nature that the non-conceptualist
variety leaves wide open. The advantage of non-conceptualist activism is that it is
not subject to a variety of objections that arise for the conceptualist variant. Morris
and I come down pretty solidly on the side of the conceptualist variant but, for
purposes here, I dont believe the distinction is important, as it appears to me that
the bootstrapping objection applies with equal force to both forms.
4. The idea of necessary existence ab alio has historical precedence in traditional
theism. Both Ibn Sina and al-Farabi, for example, distinguished between things that
are necessary per se God being the only such thing and those that are
necessary ab alio those things that are eternal and immaterial but which
necessarily require a cause outside of themselves in order to exist (Fackenheim
1947, p. 40). Granted, these and other early defenders of necessary ab alio existence
probably did not have platonic universals in mind but, rather, the denizens of the
superlunary realm. But on the face of it, at least, there is no obvious conceptual
problem with extending the idea to abstract entities. Even granting that, however,
traditional theists might argue that there remain theological problems, notably, that
the existence of beings that are necessary ab alio impinges on Gods freedom a
necessary being cannot have failed to exist and, hence, a created necessary being
would be one that God could not have failed to create and so would not have been
free not to create. Morris and Menzel (1986, p. 357) did not find this argument
compelling. For, the necessity of abstract objects notwithstanding, their creation


does not obviously impinge on divine freedom in any theologically objectionable
way, insofar as doing so is conscious, intentional, and neither constrained nor
compelled by anything independent of God and [Gods] causally efficacious power.
For more detailed discussions of divine freedom that bear on this matter, see Flint
1983 and Morris 1984.
5. Existential dependence is called logical dependence in both B&B 2006 and M&M
1986. Ive opted to abandon this terminology as, informally, logical dependence
sounds like the converse of logical priority, whereas it is not at all the case that
existential dependence as defined above is the converse of B&Bs notion of logical
priority, which will be the focus of our discussion here.
6. Higher-order syntactic features notwithstanding, the natural semantics for
languages of this kind is first-order. See Menzel 1993 for rather more complex
examples.
7. Perhaps this is because they want to allow for cases of dependence on Gods
creative activity in which God is not the sole creator. But there is no reason why
creates itself cannot be understood in such a way. If one needs the stronger sense
one can simply define solely creates in the obvious way.
8. As far as the argument goes, nothing hangs on this particular logical form for
[creating a property]. The idea is simply to have a property representing Gods act of
property creation and, concomitantly, a property representing his power so to act.
See B&B 2006, p. 369: The act of creating seems to be logically prior to the creature
(and not vice versa); and, the having of an ability seems to explain (at least
partially), and hence to be logically prior to, the exercise of that ability (and not vice
versa).
9. My rendition of BBBO is somewhat less general than B&Bs. B&B version applies
generally to any view realist view of properties (and relations) that analyzes
predication generally in terms of exemplification. By contrast, my rendition is


directed at the specific thesis of theistic activism, in particular, the thesis that God
creates all the properties. Nothing hangs on this for purposes here.
10. And, of course, the premise that [creating a property] is an action property.
11. Instantiating God () and [creating a property] () for and , respectively, in
A4. And, of course, given this identity, the substitutivity of identicals is what actually
gives us step 5 from step 4.
12. I have swapped out the schematic variable instead of B&Bs because Im
using as a second-order variable.
13. LC tells us that an individual has the property expressed by just in case it
satisfies the condition . Thus, for example, Annie has the property being a political
scientist born in Seattle just in case Annie is, in fact, a political scientist who was
born in Seattle.
14. See 2006, fns 2 and 10. I am indebted to a reviewer for pointing out that Id
overlooked B&Bs sensitivity to the threat of paradox seen in these footnotes and for
comments that led me to characterize P as a comprehension principle.
15. Let = [( = )]. Either or . If the former, then by LC,
( = ). Let be such a . Then = and . Hence, by basic
properties of identity, . Suppose then the latter, i.e., that . Then, by LC
again, ( = ), i.e., ( = ). Hence, by universal instantiation,
= . So . Contradiction.
16. Assuming the property theoretic version of the principle known as (monadic) -
conversion in the -calculus = [], for predicates and individual
variables and a classical quantification theory in the underlying logic for , the
addition of this class of -predicates is logically trivial and all instances of the
corresponding comprehension principle ( ), for primitive monadic
atomic formulas , are easily derived logical truths. For example,
follows directly from ( ), -conversion (to derive (
[]), and existential generalization in any standard classical second-order


deductive system. We could of course consider moving to a free quantification
theory (for second-order quantifiers, anyway), which would render existential
generalization invalid and, hence, the comprehension principle non-trivial but, for
purposes here, this would be a needless exercise.
17. I owe this insightful point to my friend and colleague Robert Garcia.
18. I have recently discovered that arguments against P bearing similarities to mine
in this section are also given by Boyce in section II of his 2012 ms.
19. For example, one might disallow formulas containing quantifiers that bind
predicate position occurrences of predicate variables from serving in -predicates.
See Menzel 1993 for a language containing such a restriction and section 3.1 of the
same paper for a metaphysical justification of it.
20. Boyce (ibid., p. 10) also recognizes that there is nothing compelling the activist
to acknowledge that the states of affairs in question are identical.
21. Where, presumably, states of affairs [] and [] are logically equivalent if and
only if and are.
22. That is, more exactly, when their relata are both of the form [! ], where is a
necessary being.
23. For a somewhat more accessible overview of non-well-founded set theory than
Aczels book that includes numerous applications to issues in philosophy, linguistics,
and computer science, see Barwise and Moss 1996.
24. Ive defined =

, instead of = , as, under AFA, it would have

followed from the latter that = .

25. A response here might be that non-well-founded sets are just a sort of
mathematical curiosity and dont reflect any intuitive conception of set. Barwise
(1988) argues cogently for a conception of sets as attenuated situations that
encompasses both well-founded and non-well-founded sets alike.


26. Both of the examples here are found in Barwise 1988, although he does not spell
out the Cartesian case. A similar solution to the problem of common knowledge is
foreshadowed in Lewis 1969.

References
Aczel, P. 1983. Non-Well-Founded Sets (Stanford, CA: CSLI Publications).
Barwise, J. 1988. Situations, Sets, and the Axiom of Foundation, in The Situation in
Logic (Stanford, CA: CSLI Publications), pp. 177200.
Barwise, J. and L. Moss. 1996. Vicious Circles: On the Mathematics of Non-
Wellfounded Phenomena (Stanford, CA: CSLI Publications).
Bealer, G. 1982. Quality and Concept (New York: Oxford University Press).
Bergman, M. and J. Brower. 2006. A Theistic Argument against Platonism (and in
Support of Truthmakers and Divine Simplicity), in Oxford Studies in Metaphysics,
Volume 2, ed. D. Zimmerman (New York: Oxford University Press), pp. 357-386.
Boyce, K. ms. In Defense of Theistic Activism: A Response to Bergmann and
Brower

(version

dated

11

April

2012),

http://kennethboyce.files.wordpress.com/2012/04/in-defense-of-theistic-
activism.doc.
Craig, W. L. ms. God and Abstract Objects: A Survey of the Landscape.
------ 2014. Anti-Platonism, in Beyond the Control of God? Six Views on the Problem
of God and Abstract Objects, ed. P. M. Gould (New York: Bloomsbury Publishing
Group), pp. 113126.
Flint, T. 1983. The Problem of Divine Freedom, American Philosophical Quarterly,
vol. 20, pp. 255264.
Gould, P. M. 2011. The Problem of God and Abstract Objects, Philosophia Christi,
vol. 13, no. 2, pp. 255-274.
Lewis, D. 1969. Convention: A Philosophical Study (Cambridge, MA: Harvard
University Press).


Menzel, C. 1987. Theism, Platonism, and the Metaphysics of Mathematics, Faith
and Philosophy, vol. 4, no. 4, pp. 365382.
------ 1993. The Proper Treatment of Predication in Fine-grained Intensional Logic,
in Philosophical Perspectives Volume 7: Language and Logic, ed. J. E. Tomberlin
(Atascadero, CA: Ridgeview Publishing Co.), pp. 61 87.
Morris, T. V. 1984. Duty and Divine Goodness, American Philosophical Quarterly,
vol. 21, pp. 261268.
Morris, T. V. and C. Menzel. 1986. Absolute Creation, American Philosophical
Quarterly, vol. 23, no. 4, pp. 353362.
Shapiro, S. 2006. Structure and Identity, in Identity and Modality, ed. F. MacBride
(Oxford: Clarendon Press), pp. 109145.
Zalta, E. 1983. Abstract Objects: An Introduction to Axiomatic Metaphysics (Dor-
drecht: D. Reidel).
------ 1988. Intensional Logic and Metaphysics of Intentionality (Cambridge, MA: MIT
Press).