Phronesis 52 (2007) 188-209

www.brill.nl/phro

The Ontological Argument of Diogenes of Babylon

Michael Papazian
Department of Religion and Philosophy, Berry College,
Mount Berry, GA 30149-0550, USA
[email protected]

Abstract
An argument for the existence of gods given by the Stoic Diogenes of Babylon and
reported by Sextus Empiricus appears to be an ancient version of the ontological argument. In this paper I present a new reconstruction of Diogenes argument that diers
in certain important respects from the reconstruction presented by Jacques Brunschwig.
I argue that my reconstruction makes better sense of how Diogenes argument emerged
as a response to an attack on an earlier Stoic argument presented by Zeno of Citium.
Diogenes argument as reconstructed here is an example of a modal ontological argument that makes use of the concept of being of such a nature as to exist. I argue that
this concept is a modal concept that is based on the Philonian denition of possibility,
and thus that Diogenes argument is a source of important evidence about the use of
non-Stoic modalities in the post-Chrysippean Stoa. I conclude by arguing that the
objections made against considering Diogenes argument as ontological are unfounded
and that Diogenes argument clearly resembles modern versions of modal ontological
arguments.
Keywords
Stoics, ontological argument, ancient logic

Even the most casual student of philosophy will associate the ontological argument for the existence of God with St. Anselm. Anselm is widely
celebrated as the inventor of this argument and, accordingly, a philosopher who has made an original contribution that has no Greek precedents. tienne Gilson has made the even stronger claim that no ancient
Greek philosopher could have invented the argument: Thinkers like
Plato and Aristotle, who do not identify God and being, could never
dream of deducing Gods existence from His idea.1 For Gilson, it is an
Koninklijke Brill NV, Leiden, 2007

PHRO 52,2_f4_188-209.indd 188

DOI: 10.1163/156852807X180072

4/26/07 11:39:32 AM

M. Papazian / Phronesis 52 (2007) 188-209

189

undisputed fact that no trace of [the ontological argument] exists in

Greek thought.
Several historians of philosophy have called Gilsons claim into question. Some have found adumbrations of Anselms argument in Plato
and Aristotle.2 Perhaps the most promising candidate for an ancient
ontological argument is the argument for the existence of gods attributed by Sextus Empiricus to the Stoic Diogenes of Babylon. In this
paper I focus on three distinct but related matters concerning Diogenes argument. First, I present a reconstruction of Diogenes argument, a reconstruction that I argue is preferable to an alternative
reconstruction presented by Brunschwig (Part I). I then discuss the
conception of possibility that plays a central role in Diogenes argument and examine its relation to the other modal denitions current
in Hellenistic logic (Part II). Finally, I consider Brunschwigs claim that
Diogenes argument is not ontological. I argue that Brunschwigs conclusion that Diogenes saw the possibility of reasoning along the lines of
the ontological argument but deliberately refrained from using it3 is
unwarranted. Instead I hold that Diogenes argument is indeed an ontological argument, specically a modal ontological argument (Part III).
I. Diogenes Argument
The argument in question is a reformulation by Diogenes of Babylon of
an argument given by Zeno of Citium. The arguments of Zeno and
Diogenes are reported by Sextus Empiricus (M. 9.133-6). Sextus claims
that Zeno presented the following syllogism:
A1) One may reasonably honour the gods. [
]
A2) One may not reasonably honour those who do not exist. [
]
A3) Therefore, the gods exist. [ ]

1)
2)
3)

Gilson (1936), 59.

See Brunschwig (1994), 170 for references.
Brunschwig (1994), 172.

PHRO 52,2_f4_188-209.indd 189

4/26/07 11:39:34 AM

190

M. Papazian / Phronesis 52 (2007) 188-209

In response, certain philosophers constructed a parallel or parody argument ():

B1) One may reasonably honour the wise. [
]
A2) One may not reasonably honour those who do not exist.
B3) Therefore, wise men exist. [ ]
The parallel argument cannot be accepted by the Stoics, since according
to Stoic doctrine the wise man has remained undiscoverable up to now
(M. 9.133-4).
Zenos syllogism is one of a series of arguments he presented in
shorter and more compressed (brevius angustiusque) form.4 Schoeld
has convincingly argued that the parodies or parallel arguments were all
the work of Alexinus, a Dialectician and contemporary of Zenos.5
Alexinus standard method against the Zenonian syllogisms was to keep
one premise and to substitute the other premise with one of identical
form. The conclusion of the parody would always be repugnant to the
Stoics. For example, Philo of Alexandria reports one of Zenos syllogisms concerning whether the wise get drunk:6
If one would not act reasonably in entrusting a secret to a drunk,
but if one would act reasonably in entrusting a secret to a good man,
it follows that the good man does not get drunk.

In this case, the parallel argument alters the major premise to read,
One would act unreasonably in entrusting a secret to someone who is
melancholy, asleep, or dying (Philo 177). It follows that the good man
is not melancholy but also that he needs no sleep and is immortal.
Diogenes attempted to save Zenos syllogism for the existence of god
from the parody by claiming that what Zeno really meant to say (or
should have said) in A2 was:
4)

See Cicero, De natura deorum, 2.20.

Schoeld (1983), 36. DL 2.109 states that Alexinus was one of the successors of
Eubulides.
6)
Philo, De plantatione, 176 (Colson and Whitaker).
5)

PHRO 52,2_f4_188-209.indd 190

4/26/07 11:39:34 AM

M. Papazian / Phronesis 52 (2007) 188-209

191

C2) One may not reasonably honour those who are not of such a
nature as to exist [
].
From C2 and A1, it may be concluded that
C3) The gods are of such a nature as to exist. [
] (M 134-5)
Now, however, there is the problem of getting from C3 to the original conclusion of Zenos argument. Diogenes tries to show that if the
gods are of such a nature as to exist, then the gods do exist. The reasoning is as follows:
D1) If the gods did exist at one time, then they also exist now. [
, ]
The support for D1 is that the gods are indestructible and ungenerated
according to their conception (
) just as atoms are indestructible and ungenerated
according to their conception (M 9.135). Thus, not being liable to
destruction, the gods of the past must exist even today.
Diogenes must assume that
D2) If the gods are of such a nature as to exist, then they must have
existed some time in the past.
From C3, D1, and D2, the desired conclusion A3 can be derived.
Diogenes then shows that a parallel argument leading to the conclusion that the wise exist is not sound because it is not the case that
E1) Since () the wise are of such a nature as to exist, then they
must also exist now. (M 9.135-6)
Diogenes argument raises a number of questions. First, what does it
mean to say that something is of such a nature as to exist? Second, why
must it follow from the proposition that the gods are of such a nature

PHRO 52,2_f4_188-209.indd 191

4/26/07 11:39:34 AM

192

M. Papazian / Phronesis 52 (2007) 188-209

as to exist that they must have existed in the past? And third, why does
the past existence of the wise men not follow from their being of such
a nature as to exist? Sextus leaves us in the dark about these questions.
Any attempt at reconstructing Diogenes argument will require an
interpretation of existing by nature or natural existence. Brunschwig
presents three possible interpretations:7
1. Necessarily existing
2. Possibly existing
3. Normally existing
1 can be excluded because it would make Diogeness attempt to argue
from the gods existence by nature to their actual existence superuous.
If Diogenes understood natural existence to mean necessary existence,
he would not have bothered to devise the supplementary argument that
attempts to show that existence by nature entails actual existence. 2 is
too weak according to Brunschwig because merely possible existence
does not ensure existence at some time in the past, thus making D2 a
false conditional. Instead, Brunschwig proposes that existence by nature
must be some third mode of existence weaker than 1 but stronger than
2. He calls this mode of existence normally existing. For Brunschwig,
the properties which a given type of being possesses by nature are normally possessed by any token of this type; a given token of the type
cannot be deprived of any of these properties, unless by accident, or . . .
unless through the action of prohibiting external causes.8 Brunschwig
understands the claim that the gods exist by nature to mean that tokens
of the divine type must exist for the most part or most of the time. It
may be that there are no such tokens now or at certain times in the past
or future, but it cannot be the case that at no time in the past did any
token of the gods exist. On this interpretation of natural existence, D2
is a true conditional.
There is, however, one major problem with Brunschwigs reconstruction that seems to be a decisive objection against it. It is dicult to see

7)
8)

Brunschwig, 185.
Brunschwig, 186.

PHRO 52,2_f4_188-209.indd 192

4/26/07 11:39:35 AM

M. Papazian / Phronesis 52 (2007) 188-209

193

how Diogenes could deny that the sages exist by nature, which is something that Diogenes does in Brunschwigs interpretation.9 For Diogenes
could not have held that one may not reasonably honour the sages. If
he did, then he could have very easily dismissed the parody argument
by refusing to give his assent to B1 rather than reformulating premise
A2 of Zenos argument. But if Diogenes wanted to maintain that one
may reasonably honour the wise, he would have to believe that the wise
exist by nature, since Diogenes reformulated premise, C2, commits
him to the belief that things that do not exist by nature are not worthy
of honour.
Now it is true that some Stoics attempted to discharge the parody
argument by appealing to the ambiguity of the word or to honour.10 To honour may mean either to worship or to hold in esteem.
These Stoics took A1 to mean that one may reasonably worship the
gods. B1 would be false if to honour means to worship, but would be
true in the second sense of honour. So either B1 is false or the parody
argument commits a fallacy of equivocation by using honour in the
sense of holding in esteem in B1 and in the sense of worship in B2. The
reasoning behind this may be that one can reasonably hold in esteem or
think highly of ctional characters though it would be absurd to worship them.
Diogenes did not invoke the ambiguity of honour but instead seems
to have held that while the gods exist because they exist by nature the
natural existence of the wise does not entail their actual existence. The
question then is what the relevant distinction between the gods and the
wise is. If Diogenes believed that the wise exist by nature, then existence by nature cannot mean normal existence or existence most of the
time. There does seem, however, to be good reason to believe that
something has a property F by nature if it has or will come to have that

See Brunschwig, 187-9. E1 is a . According to DL 7.71, a (which is distinguished by the use of rather than the of simple
9)

conditionals) asserts that (i) if p, then q, and (ii) p is true. Brunschwig argues that in
denying E1, Diogenes is denying both (i) and (ii). In my reconstruction, Diogenes only
denies (i), not the truth of the antecedent. That is, Diogenes denies that the sages present existence follows from their natural existence but accepts their natural existence.
10)
See M 9.136.

PHRO 52,2_f4_188-209.indd 193

4/26/07 11:39:35 AM

194

M. Papazian / Phronesis 52 (2007) 188-209

property whenever external circumstances do not prevent it from being

F. Support for this understanding of having a property by nature is
found in Sextus treatment of the Stoic theory of signs. Sextus says that
the Stoics divided the non-evident () into the absolutely nonevident, the naturally non-evident, and the temporarily ( )
non-evident (M. 8.145). The temporarily non-evident are those things
that are naturally evident ( ) but are nonevident to us at certain times due to external circumstances (
). The example given is the city of Athens, which is not
evident to us because of our distance from it. It is important to note
that Athens could remain non-evident to us most, if not all of the time,
if we never go near it. In this case, of course, the property is relational,
so that Athens is evident to those in or near it, but non-evident to those
of us who are far away from it. But the same point would hold in the
case of non-relational properties such as existence or wisdom. Things
exist by nature or are wise by nature if they either do come to exist or
become wise except when external circumstances prevent them from
existing or being wise. This is the sense of natural existence that I wish
to defend as the one used by Diogenes. This understanding of natural
existence is preferable to Brunschwigs because it allows us to arm
that sages are worthy of honor and to deny that sages exist most of
the time.
My understanding of natural existence faces its own diculty,
however, because if it is correct, the Stoics would have had to believe
that the development of a sage is natural and that only external circumstances could prevent a human from becoming a sage. Can such a view
be attributed to the Stoics?
There is support in a number of Stoic sources for attributing to the
Stoics the doctrine that wisdom is the natural result of normal human
development and that the ability to become a sage is naturally implanted
in all humans.11 According to these sources, human development to
wisdom is natural in the same way that it is natural for seeds to develop
into mature plants. But the claim that wisdom is a natural feature of
11)

This discussion of the naturalness of the development of wisdom is heavily indebted

to Jackson-McCabe (2004). I am grateful to a referee of this journal for drawing my
attention to Jackson-McCabes paper.

PHRO 52,2_f4_188-209.indd 194

4/26/07 11:39:35 AM

M. Papazian / Phronesis 52 (2007) 188-209

195

human development is nevertheless consistent with the rarity of wise

people, just as it is natural for a seed to develop into a full-grown organism even if very few seeds actually do so because their natural development is often frustrated or prevented by external conditions.
In particular, Plutarch notes that Chryippus said that his teaching
concerning good and bad things connects especially with innate or
implanted preconceptions ( ).12 These implanted
preconceptions, Jackson-McCabe has argued, are mentioned by
Chrysippus in reference to preconceptions of the ethical sphere whose
formation is guaranteed by oikeisis, and which in this sense represent a
type of inborn knowledge (327). It is oikeisis that is responsible for the
ability of animals to distinguish between what is benecial and harmful
to their constitutions. In humans, that tendency is conjoined with reason and the ability to form concepts, so that humans are endowed with
the natural tendency to form concepts of what is good and bad. As
Cicero points out, once man possesses understanding or the ability of
conception (intellegentiam vel notionem potius), he will ultimately by
exercise of intelligence and reason come to grasp the concept of goodness in itself, that is, what is praiseworthy and desirable for itself (per
se laudandum et expetendum).13 Thus, the formation of the moral concepts and their discernment is a natural process that results from the
combination of oikeisis and rationality that is characteristic of human
nature.
Seneca makes explicit the seed analogy by noting that nature has
given us the seeds of knowledge (semina nobis scientiae dedit),14 seeds
which by their nature result in knowledge though they often will not
reach their natural end. Since it is the possession of knowledge that
characterizes the sage, it follows that the Stoics believed that sages exist
by nature, as understood according to my proposed sense of natural
existence.
My interpretation of natural existence faces a further problem because
although Brunschwigs interpretation of existence by nature clearly

12)
13)
14)

Plutarch, De Stoicorum repugnantiis, 1041E.

Cicero, De nibus, 3.21 (Rackham)
Seneca, Ep. 120.4.

PHRO 52,2_f4_188-209.indd 195

4/26/07 11:39:36 AM

196

M. Papazian / Phronesis 52 (2007) 188-209

makes the conditional D2 true, my interpretation does not. It does not

follow from the fact that something exists by nature that it existed some
time in the past, since external circumstances may always have prevented it from existing. If my interpretation is correct then it must
make D2 true without making the general inference from existence by
nature to existence once upon a time licit. Fortunately, there is a way to
do this. The argument assumes that part of the concept of god is being
ungenerated and indestructible. It may also be part of the concept of
the divine that it is impassible it cannot be aected by external causes
or at least cannot be prevented by external causes from having the qualities to which it has an appropriate relation. This is, of course, true of
the Stoic god. The Stoic god is identied with the active principle of the
universe, the designing re that creates and sustains the world.15 Since
the only other principle recognized by the Stoics is the passive and inert
matter, there are no causes or inuences external to god. Therefore, if
the Stoic god exists by nature, it will exist, because there are no external
circumstances that can prevent it from existing. In contrast, sages also
exist by nature but because there are external circumstances that can
prevent humans from becoming sages, the natural existence of sages
does not entail their actual existence.
Another objection to my interpretation is that it does not make sense
of the way in which Diogenes tries to show that the gods natural existence implies the gods actual existence. If the gods exist by nature and
there are no external circumstances that could prevent them from existing, then it should be obvious that the gods must exist now. Why, then,
does Diogenes only infer from the gods natural existence that the gods
existed at some time in the past, and then bring in D1 to draw the
desired conclusion? Perhaps Diogenes was anticipating a potential
objection along the following lines. One could object to Diogenes that
the mere fact that something exists by nature and is not prevented by
external circumstances from coming into existence does not entail that
that thing will immediately come into existence. For example, the Stoic
cosmos by nature results in a state of conagration. Further, no external
circumstances prevent the conagration from coming into existence.
But the nature of the cosmos is such that it gradually develops into this
15)

See, for example, DL 7.134.

PHRO 52,2_f4_188-209.indd 196

4/26/07 11:39:36 AM

M. Papazian / Phronesis 52 (2007) 188-209

197

state. It is understandable that the word for nature, , which derives

from , to grow, would be associated with processes of development and unfolding. Thus, the natural existence of the gods (and the
absence of external constraints) would not necessarily imply the gods
present or actual existence but at best only a process of development
that would result at some time in the existence of gods. Accordingly,
the move from the gods natural existence to the gods actual existence
needs further support.
That support comes from the conceptual understanding of the gods
as ungenerated and indestructible beings. If we add these two characteristics to the divine nature, then the desired result follows. Diogenes
complete argument may be reconstructed as follows.16 There are two
auxiliary arguments whose conclusions are used in the main argument
to get to the nal conclusion that the gods exist now. The rst auxiliary
argument is
I1: If the gods are impassible, then nothing can prevent the gods from
existing.
I2: The gods are impassible according to their conception. [a basic tenet
of Stoic physics, in which god or re is the active principle and
matter is the passive principle]
I3: Therefore, nothing can prevent the gods from existing. (I1 and I2 by
the Stoics rst indemonstrable or modus ponens)17

16)
The reconstruction is much more complex than the argument as reported in Sextus, but Sextus account is clearly truncated. The point of my reconstruction is to make
explicit all the assumptions and logical inferences Diogenes would need to reach his
conclusion from the given premises.
17)
In reconstructing the argument it is important to make only those inferences
known to be acceptable to the Stoics. It has been convincingly argued that Stoic logic
is a form of relevance logic in which many sequents that are valid in classical logic do
not hold (see Bobzien (1996), 185f.). I use Bobziens account of Stoic inference
to show that the Stoics would have accepted this reconstructed argument as valid.
Accordingly, arguments must either be one of the ve indemonstrables or be capable
of reduction to one of the indemonstrables using only the four Stoic themata. See
Bobzien (1996), 134-141 for a discussion of the indemonstrables, and 143-163 for the
themata.

PHRO 52,2_f4_188-209.indd 197

4/26/07 11:39:36 AM

198

M. Papazian / Phronesis 52 (2007) 188-209

The second auxiliary argument is

H1: If the gods are not of such a nature to exist, then one may not reasonably honour them.18
H2: One may reasonably honour the gods. [A1 of Zenos argument]
H3: Therefore, the gods are of such a nature as to exist. (H1 and H2 by
the second indemonstrable or modus tollens)
The main argument begins by using the rule of addition19 to form a
conjunction from I3 and H3:
M1: The gods are of such a nature as to exist and nothing can prevent
the gods from existing.
Given our understanding of natural existence, the following conditional must be true:
M2: If the gods are of such a nature as to exist and nothing can prevent
the gods from existing, then the gods must exist either in the past
but not now or in the future but not now or the gods must exist
now.20
M1 and M2 entail
M3: The gods must exist either in the past but not now or in the future
but not now or the gods must exist now.
18)

This is an instantiation in conditional form of the general principle C2 of the

original argument. The reformulation as a conditional is necessary in order for this
argument to be in the form of the second indemonstrable.
19)
The sequent A, B A and B is valid in Stoic logic. Application of the rst thema
(for which, see Apuleius, Int. 209.12-14) reduces this sequent to the third indemonstrable (for which, see M 8.266 and DL 7.80):
not(A and B), A not B
20)
The Stoic or is exclusive (see Gellius, Noctes Atticae, 16.8.12-14). The disjunction
X exists in the past or X exists in the future can be false in Stoic logic since both disjuncts can be true. In order for M2 to be a sound conditional, the consequent must be
formulated as above.

PHRO 52,2_f4_188-209.indd 198

4/26/07 11:39:37 AM

M. Papazian / Phronesis 52 (2007) 188-209

199

By denition of destruction and generation we can assume

M4: If the gods exist in the past but not now, then they have been
destroyed.
and
M5: If the gods exist in the future but not now, then they will be
generated.
But according to their conceptions,
M6: It is not the case that the gods are destroyed.
and
M7: It is not the case that the gods are generated.
M4 and M6 entail by the second indemonstrable
M8: It is not the case that the gods exist in the past but not now.
and M5 and M7 likewise entail
M9: It is not the case that the gods exist in the future but not now.21
We can combine M8 and M9 to get
M10: It is not the case that either the gods exist in the past but not now
or that the gods exist in the future but not now.22

21)
The account in Sextus must be abridged here, giving premise M8 but not M9 and
thus not explaining the relevance of gods ungenerated nature in the argument.
22)
A valid move in Stoic logic since not A, not B not (A or B) can be reduced to the
fth indemonstrable A or B, not A B by one application of the rst thema. For the
fth indemonstrable see DL 7.81.

PHRO 52,2_f4_188-209.indd 199

4/26/07 11:39:37 AM

200

M. Papazian / Phronesis 52 (2007) 188-209

Finally M10 and M3 take us to the desired conclusion again by

means of the fth indemonstrable:
M11: The gods exist now.23
As it stands the reconstruction makes use only of principles that are
attributed either to Diogenes or the Stoics or which they in general
would accept, and all the inferences are valid in Stoic logic. The argument reconstructed in this way has the advantage over Brunschwigs
reconstruction because it allows us to explain why Diogenes reinterpreted Zenos argument in light of the parody in his own way rather
than simply discharging the parody by pointing out the ambiguity of
honour. Further, my reconstruction, unlike Brunschwigs, does not
attribute to Diogenes the belief that the sages are not worthy of
honour.
II. Natural Possibility and Hellenistic Modal Logic
Having presented the reconstruction, we turn now to the conception of
possibility that Diogenes uses in the argument and its relation to the
other modal concepts in Hellenistic logic. One of the reasons that Diogenes argument is so interesting is that it may help to shed light on the
development of modal logic in the post-Chrysippean Stoa, and in particular, to provide evidence of the continued use of older, pre-Chrysippean modal concepts among the Stoics. The concept of being of such a
nature as to exist, which plays a central role in the reconstructed main
argument, seems to be a special form of a modal concept, namely a
form of possibility according to which something is possibly F if it will
come to have that property F whenever external circumstances do not
prevent it from having F. Let us refer to Diogenes conception of possibility as natural possibility.
Natural possibility closely resembles the denition of possibility
attributed to the dialectician Philo:24
That is, (A or B) or C, not (A or B) C, which is a substitution instance of the fth
indemonstrable.
24)
Alexander, An. Pr. 184, 9-10. See also Boeth., Int. 234, 10-21 and Simplicius, Cat.
23)

PHRO 52,2_f4_188-209.indd 200

4/26/07 11:39:37 AM

M. Papazian / Phronesis 52 (2007) 188-209

201

[the possible] is this: what is said of the subject according to its bare tness (
), even if it has been prevented from coming to be by some external
necessity ( ). Accordingly Philo said that it is possible that cha
lying in the uncut [wheat] ()25 or in the depth of the sea be burned where
it is, even though by necessity it is prevented by its surrounding.

This Philonian possibility must be contrasted with the Stoic denition

of possibility usually associated with Chrysippus. For Chrysippus a
possible proposition is that which admits of being true and external
factors ( ) not preventing it from being true.26 According to
Chrysippus, it is not possible that the cha at the bottom of the sea
burn, since it is prevented by external circumstances from burning.
Therefore the conception of possibility (i.e. being of such a nature as to
be F) that Diogenes uses in his argument is closer to the Philonian
conception of possibility, since for Diogenes the wise man is of such a
nature as to exist even though external circumstances may always prevent him from existing. Let us refer to Diogenes conception of possibility as natural possibility.
Is natural possibility the same as Philonian possibility? Their identity
would make sense on the very plausible assumption that Diogenes is
arguing with Alexinus, who belonged to the same school as Philo.27
7, 195f. It should be noted that this modal denition, like the Stoic ones, is de dicto.
That is, a proposition is possible just when the very nature of the statement is supportive of truth, as Boethius puts it, even if factors extrinsic to the proposition prevent
the proposition from being true. But presumably the reason why propositions like I
will read Theocritus Bucolica today and Cha burns are possible is because of a de re
possibility or suitability possessed by the referent of the subject. Thus, Simplicius notes
that on the oor of the ocean a piece of wood is combustible as far as it is in itself and
according to its nature ( ). Note that both Boethius and Simplicius characterize Philonian possibility in terms of the nature of either the statement
or object in question.
25)
Following the translation in Alexander (1999), 94 rather than Long and Sedleys
atomic dissolution (LS 38B2).
26)
DL 7.75. See Frede (1974), 107-117 and Bobzien (1993), 76-84 for discussion of
this denition.
27)
DL 2.109. Philo was a student of Diodorus (DL 7.16), who was a student of Apollonius, himself a student of Eubulides (DL 2.110). According to DL 2.109, Alexinus
was among the successors of Eubulides (
).

PHRO 52,2_f4_188-209.indd 201

4/26/07 11:39:38 AM

202

M. Papazian / Phronesis 52 (2007) 188-209

Since Diogenes is arguing dialectically, it is understandable that he

would make use of a modal denition propounded by a colleague of
Alexinus, and therefore likely to be the one accepted by Alexinus.28
Although the sources are too meager to warrant certainty, the combination of three considerations (the use of in the accounts of both
Philonian possibility and natural possibility, the fact that Diogenes
argument emerges from dialectical engagement with a member of the
same school as Philo, and the fact that the interpretation of natural
existence as a Philonian modality allows for a more plausible reconstruction of Diogenes argument than does Brunschwigs interpretation) makes it highly likely that natural possibility is essentially Philonian
possibility.29

28)

There is reason to suppose that some of the pre-Chrysippean Stoics also accepted
the Philonian modalities. The early Stoa generally inherited the logical doctrines of the
Dialecticians: Zeno studied logic under Diodorus and Philo (DL 7.16), and according
to Cicero, Zeno contributed much less to the development of logic than previous
philosophers (Gell. 7.2.6-13). Frede and Ebert both provide support for the dependence of pre-Chrysippean Stoic logic on the Dialecticians by noting the several parallels between the titles of logical books written by Cleanthes and his disciple Sphaerus
and the titles of books by Diodorus and Philo (Frede (1974), 22; Ebert (1987), 107-8).
It is likely, therefore, that the Stoic logicians prior to Chrysippus did not make any
signicant advances beyond the logic of the Dialecticians. The two Dialectical modal
denitions are those of Diodorus and Philo. Since Cleanthes rejected Diodorus Master
Argument, which was used by Diodorus to support his denition of the modal concepts, and rejected the necessity of the past (see Epictetus 2.16.5), he presumably
accepted the Philonian denitions because true propositions about the past are always
necessary under the Diodorean denition of necessity but not under the Philonian
denition. For the Philonian modalities are concerned solely with the intrinsic features
of the proposition, and the passage of time is an extrinsic factor that by itself would not
alter the Philonian modal status of a proposition. If Zeno also accepted Philos
denition, there is another reason why Diogenes would invoke Philonian possibility.
Since Diogenes is attempting to reconstruct Zenos thinking to defend the argument
against the parody, he is obliged to use the denition of possibility available to Zeno.
29)
Brunschwig (p. 186) seems to hint at the relation of natural existence to Philonian
possibility though he does not make an explicit connection: cups are of such a nature
as to be broken; they normally break, unless one or even lots of them have sunk in
the depths of sea.

PHRO 52,2_f4_188-209.indd 202

4/26/07 11:39:38 AM

M. Papazian / Phronesis 52 (2007) 188-209

203

III. Is Diogenes Argument an Ontological Argument?

We can now turn to the question of whether Diogenes argument as
reconstructed is an ontological argument. One of the diculties in
deciding whether an argument is ontological is the obscurity of the
criteria usually used to dene such arguments. The term, of course, is
due to Kant, who dened ontological arguments as those that abstract
from all experience, and argue completely a priori, from mere concepts,
to the existence of a supreme cause.30 It is often dicult, however, to
determine when an argument argues completely a priori or from mere
concepts. This requires that we be able to distinguish between properties that are part of the concept of God (or the gods) and those that are
not. Thus Zeno, whose argument for the existence of the gods Diogenes tried to salvage from parody, used the premise One may reasonably honour the gods. Is the property reasonably honoured part of the
concept of the gods, something that may be inferred a priori without
any reference to empirical considerations? One may plausibly argue
that claims about reasonableness are matters of prudential consideration dependent upon empirical facts about human nature and not
simply upon some conceptual truths about the gods. Alternatively, one
could maintain that being worthy of honour pertains to the gods independent of whether honouring them has good consequences for
humans. If one follows the Kantian denition alone, the question of
whether an argument containing such a premise is ontological cannot
be conclusively answered.
Perhaps a better way to proceed is to look at some later versions of
ontological arguments and determine if there are any such arguments
that resemble the argument forms of the Stoic arguments. Useful in this
regard is Graham Oppys taxonomy of ontological arguments. Oppy
divides ontological arguments into seven categories: denitional, conceptual, modal, Meinongian, experiential, mereological, and Hegelian.31 In general Oppy prefers to characterize ontological arguments as
those that proceed from considerations that are entirely internal to the
theistic worldview. That is, the premises of ontological arguments
30)
31)

Kant (1929), 500.

Oppy (1996), 1.

PHRO 52,2_f4_188-209.indd 203

4/26/07 11:39:38 AM

204

M. Papazian / Phronesis 52 (2007) 188-209

involve terms that non-theists would object to if used in contexts that

bestow existential import upon those terms, examples of which include
greatest possible being, beings than which nothing greater can be conceived, and most perfect being. Ontological arguments use such
terms, but in order to avoid circularity the terms that appear in the
premises must fall outside of ontologically committing contexts.
Diogenes argument appears to resemble most closely the modal
arguments. This is not surprising given the rich tradition of modal logic
that preceded Diogenes both within the Stoa and in the work of the
Megarians and Dialecticians such as Diodorus Cronus.32 Modal ontological arguments are arguments with premises that make modal claims
about God, claims involving the possibility or necessity of Gods existence.33 An example would be
It is possible that God exists.
Either it is impossible that God exists or it is necessary that God
exists.
It is necessary that God exists.
Therefore, God exists.
Applying ancient denitions of the modal operators, it is easy to construct an ancient modal ontological argument. For example, Diodorus
held that the possible is what either is or will be true.34 Given this
denition of the possible, one can easily imagine Diodorus arguing in
this way:
1. It is possible that the gods exist.
2. It is true or will be true that the gods exist. (by 1 and Diodorus
denition of the possible)

32)

There is disagreement about the existence of a Dialectical school and its relation to
the Megarians. I follow Sedley (1977) in arming the existence of a distinct Dialectical school though the main theses argued for here do not depend on acceptance of
Sedleys views.
33)
Examples of modern versions of such modal ontological arguments can be found
in Malcolm (1960) and Plantinga (1974), 196-221.
34)
Epictetus, Diss. 2.19.1-5 (Schenkl); Boethius, Int. 234, 22-6 (Meiser).

PHRO 52,2_f4_188-209.indd 204

4/26/07 11:39:38 AM

M. Papazian / Phronesis 52 (2007) 188-209

205

3. If it will be true that gods will exist, then gods exist at all times
(because the gods are not of such a nature as to come into existence or go out of existence).35
Therefore the gods exist now.
We would not have any trouble identifying this as a modal ontological
argument since the argument has premises that make modal claims
about the gods in a similar way that modern modal arguments do about
God. Diodorus has his own peculiar temporal analysis of modal concepts but his conceptions of these concepts are still recognizable as corresponding to modern conceptions of possibility and necessity.
Of course there is no evidence that Diodorus actually formulated
such an argument. Pace Gilson, however, there is no reason to think
that Diodorus could not have dreamt of producing this argument. We
do have, though, the actual argument of Diogenes as reported by Sextus Empiricus, and I think that Diogenes argument is best understood
as a modal ontological argument, since, as I have argued, Diogenes
makes use of a modal concept resembling Philonian possibility.
Before making the case that Diogenes argument is a modal ontological argument, however, it is necessary to dispose of one possible
objection to considering an ancient argument like Diogenes as an
ontological argument. The most obvious dierence between the ancient
arguments and modern ontological arguments is that the Stoic arguments are arguments for the existence of gods, not of God. But this is
only an apparent dierence once one sees that the Stoics were monotheists who veiled themselves as polytheists. For although the Stoics
used the names of the traditional gods, it is clear that these names
describe dierent features of, and were intended to refer to the one god
that the Stoics identied with the cosmos:
[God] is called by many names according to his powers. For they call him Zeus
() through whom ( ) all things come to be. They call him Life () in
so far as he is the cause of life or extends all through life, and Athena with respect
to the extension of his commanding-faculty into the ether ( ), and Hera
with respect to its extension into the air ( ). (DL 7.147)
35)

A principle that is appealed to in Diogenes argument. See Sextus Empiricus,

M 9.135.

PHRO 52,2_f4_188-209.indd 205

4/26/07 11:39:38 AM

206

M. Papazian / Phronesis 52 (2007) 188-209

There is also evidence concerning Diogenes own inclination to allegorize things commonly taken to be divine. Cicero states that Diogenes
of Babylon . . . in his book Minerva rationalizes the myth of the birth of
the virgin goddess from Jove by explaining it as an allegory of the processes of nature.36 Minucius Felix wrote that Diogenes of Babylon
again adopts a similar school of teaching in expounding the birth of
Jupiter, the production of Minerva and the like, as terms denoting
things, not gods.37 Although Diogenes argument concludes that the
gods exist, we should read this to mean that the one Stoic god that can
be described in numerous ways exists. I will use the term god to refer
to the one Stoic god the establishment of whose existence is the real
purpose of Diogenes argument. God is reserved for the supreme
being of classical theism, the God of Anselm. In the reconstruction of
the propositions forming the ancient arguments I retain the plural form
but add the proviso that it is a simply a manner of speaking.
Returning to the question of whether Diogenes argument meets the
criteria Oppy proposes for ontological arguments, recall that for Oppy
ontological arguments are arguments that proceed from considerations
that are entirely internal to the theistic worldview. It seems that Diogenes argument is ontological in this sense. Diogenes makes use of
concepts that are internal to Stoic theology, specically the impassible,
ungenerable, and indestructible nature of the Stoic god and then proceeds from these concepts to derive the existence of that god now.
Accordingly, Diogenes argument would be a form of the modal
ontological argument that uses the natural possibility of the gods existence to infer their current existence. As in other modal ontological
arguments, the special attributes of god ensure that god must exist given
natural existence while the inference from the natural existence of other
things (such as sages) to their actual existence is blocked.
There remains, however, another reason to doubt that Diogenes
argument is ontological. The natural existence of the Stoic god is not
part of the idea of god but follows from a normative claim about the
propriety or rationality of humans honouring gods. The auxiliary argu-

36)
37)

Cicero, De natura deorum, 1.15 (Rackham).

Minucius Felix, Octavian, 19.12 (Rendall).

PHRO 52,2_f4_188-209.indd 206

4/26/07 11:39:39 AM

M. Papazian / Phronesis 52 (2007) 188-209

207

ment I1-3 is not based on a conceptual truth about god internal to

theology but on what it is reasonable for humans to do.
One should rst note that the premise that one may reasonably honour the gods can be read in two ways. An anthropocentric reading
interprets the premise to say that it is reasonable to honour god on
prudential or pragmatic grounds. It is not that god is intrinsically honourable but, reasoning along the lines of Pascals wager, the safer policy
for humans is to behave as if there is a god.38 A theocentric reading
understands the premise to be making a claim about the intrinsic nature
of the god. God is by nature worthy of honour.39 Barnes and Bett support a theocentric reading while Brunschwig and Schoeld advocate an
anthropocentric reading. In the case of an anthropocentric reading, the
premise is not internal to Stoic theology but one that is more widely
acceptable on strictly prudential grounds. Likewise, if the premise is
read anthropocentrically, then it would be a normative claim about
rational human behaviour involving empirical considerations rather
than a conceptual truth about god. In either case such a reading would
prevent the argument from being purely ontological.
As Bett points out,40 an anthropocentric reading of A1 would make
Zenos premises obviously inconsistent and thereby make his argument
unsound. For if the reason that one may reasonably honour god is that
ones failure to honour god may result in dire consequences should god
exist, then honouring non-existing things may be reasonable as long as
one does not know for certain that they do not exist. Bett notes that
there are no good prudential reasons for honouring beings that are
known not to exist. But even that can be questioned. If it is reasonable
to honour gods because it is safer to follow convention than to out it,
it may be reasonable even to honour things that one knows do not
exist. But A2 denies the reasonableness of honouring non-existing
things. So the reasonableness of honouring some non-existing things is
entailed by A1 on the anthropocentric reading and denied by A2. The
theocentric reading does not face this problem since the claim that the
gods are by their nature worthy of honour does not entail that it is
38)
39)
40)

See Brunschwig, 175 and Schoeld, 38-39.

Barnes (1972), 17.
Bett (1996), 85-86.

PHRO 52,2_f4_188-209.indd 207

4/26/07 11:39:39 AM

208

M. Papazian / Phronesis 52 (2007) 188-209

reasonable to honour non-existing things. The theocentric reading of

A1 is, therefore, more plausible. Further, since Sextus report does not
state that Diogenes altered Zenos rst premise, one may assume that
Diogenes reading of A1 is the same as Zenos.41
However, regardless of whether one accepts the theocentric or anthropocentric reading of this premise, the status of Diogenes argument as a
modal ontological argument should not be aected. The reasonableness
of honouring god is invoked simply to arrive at the possibility of gods
existence. Once the possibility is established, the main argument is used
to infer the actual existence of god. It should not matter that the auxiliary argument makes use of empirical or prudential considerations.
Suppose that one presented a modal ontological argument in which the
possibility of Gods existence is used to infer the necessity of Gods existence. In order to support the possibility of God, the arguer uses some
empirical or psychological fact (that humans do conceive of God) and
then adds the premise that whatever one conceives of existing can exist.
The use of such an argument in support of the possibility of Gods existence would not call into question the ontological character of the main
argument. Neither then should the auxiliary argument of Diogenes, if
indeed it contains an empirical premise, undermine the ontological
character of his main argument.42
If this is correct, then Diogenes main argument is a modal ontological argument because its premises are purely conceptual claims
about god internal to Stoic theology and a modal concept plays a central role in the argument. There is, however, an important distinction
between Diogenes argument and Anselms argument. Anselms argument is much simpler, making use only of one feature of God (that He
is that than which nothing greater can be conceived) and deriving all
other attributes (e.g. eternity and impassibility) as corollaries.43 Thus,
41)

Cf. Bett, 87.

Whether Zenos argument is ontological or not does depend on the interpretation
of A1, since this premise is an essential part of his argument. Even if Zenos premise
should be taken theocentrically, Zenos argument is at best a crude rst attempt at an
ontological argument, and credit is due to Diogenes for constructing a more sophisticated and rigorous proof.
43)
Anselm demonstrates the impassibility of God in chapter 8 of the Proslogion and
eternity in chapter 13.
42)

PHRO 52,2_f4_188-209.indd 208

4/26/07 11:39:39 AM

M. Papazian / Phronesis 52 (2007) 188-209

209

while Diogenes argument is ontological, it lacks the simplicity, elegance, and power of Anselms argument. Gilsons claim that no trace of
the ontological argument exists in Greek thought cannot be sustained
in the light of Diogenes argument, though it remains true that Anselm
went beyond ancient thought by arguing from the mere idea of God as
that than which nothing greater can be conceived to the conclusion of
Gods necessary existence.44
References
Alexander of Aphrodisias (1999) On Aristotle Prior Analytics 1.14-22, translation by
I. Mueller (London).
Barnes, J. (1972) The Ontological Argument (London).
Bett, R. (1996) Hellenistic Essays Translated, Apeiron 29, 75-97.
Bobzien, Susanne (1986) Die stoische Modallogik (Wrzburg).
(1993) Chrysippus Modal Logic and Its Relation to Philo and Diodorus, in
Dring and Ebert (eds.), Dialektiker und Stoiker (Stuttgart), 63-84.
(1996) Stoic Syllogistic, Oxford Studies in Ancient Philosophy 14, 133-192.
Brunschwig, J. (1994) Papers in Hellenistic Philosophy, translation by J. Lloyd
(Cambridge).
Dring, K. and Ebert, T. (eds.) (1993) Dialektiker und Stoiker (Stuttgart).
Ebert, T. (1987) The Origins of the Stoic Theory of Signs in Sextus Empiricus, Oxford
Studies in Ancient Philosophy 5, 83-126.
Frede, M. (1974) Die stoische Logik (Gttingen).
Gilson, . (1936) The Spirit of Medieval Philosophy (Notre Dame).
Jackson-McCabe, M. (2004) The Stoic Theory of Implanted Preconception, Phronesis 49, 323-347.
Kant, I. (1929) Critique of Pure Reason, translation by N.K. Smith (New York).
Long, A.A. and Sedley, D.N. (1987) The Hellenistic Philosophers (Cambridge).
Malcolm, N. (1960) Anselms Ontological Arguments, Philosophical Review 69,
41-62.
Oppy, G. (1996) Ontological Arguments and Belief in God (Cambridge).
Plantinga, A. (1974) The Nature of Necessity (Oxford).
Schoeld, M. (1983) The Syllogisms of Zeno of Citium, Phronesis 28, 31-58.

44)

I am grateful to Richard Bett, Henry Dyson, Steven Strange and an anonymous

referee of this journal for their comments on earlier drafts of this paper.

PHRO 52,2_f4_188-209.indd 209

4/26/07 11:39:39 AM