Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

History and Philosophy of Logic

ISSN: 0144-5340 (Print) 1464-5149 (Online) Journal homepage: http://www.tandfonline.com/loi/thpl20

Stoic Conditionals, Necessity and Explanation

SCOTT LABARGE

To cite this article: SCOTT LABARGE (2002) Stoic Conditionals, Necessity and Explanation,
History and Philosophy of Logic, 23:4, 241-252, DOI: 10.1080/0144534021000041236

To link to this article: http://dx.doi.org/10.1080/0144534021000041236

Published online: 10 Nov 2010.

Submit your article to this journal

Article views: 51

View related articles

Full Terms & Conditions of access and use can be found at


http://www.tandfonline.com/action/journalInformation?journalCode=thpl20

Download by: [Temple University Libraries] Date: 10 June 2016, At: 22:57
HISTORY AND PHILOSOPHY OF LOGIC, 23 (2002), 241±252

Stoic Conditionals, Necessity and Explanation

SCOTT LABARGE
Department of Philosophy, Santa Clara University, 500 El Camino Real, Santa Clara,
CA 95053-0310, USA

Received 15 February 2002 Revised 3 September 2002

An examination of a particular passage in Cicero's De fatoÐFat. 13-17Ðis crucial to our understanding of


the Stoic theory of the truth-conditions of conditional propositions, for it has been uniquely important in the
Downloaded by [Temple University Libraries] at 22:57 10 June 2016

debate concerning the kind of connection the antecedent and consequent of a Stoic conditional should have
to one another. Frede has argued that the passage proves that the connection is one of logical necessity, while
Sorabji has argued that positive Stoic attitudes toward empirical inferences elsewhere suggest that that
cannot be the right interpretation of the passage. I argue that both parties to the debate have missed a
position somewhere between them which both renders a connection between antecedent and consequent
that is not merely empirical and makes sense of the actual uses to which the Stoics put the conditional.
This will be an account which grounds the connection between antecedent and consequent in a proleÃpsis,
a special kind of concept which plays a special epistemological role for the Stoics, especially in grounding
scienti®c explanations. My contention will be that Stoic conditionals are true when there is a conceptually
necessary connection between antecedent and consequent such that the former explains the latter via a
proleÃpsis.

1. Stoic conditionals, necessity and explanation1

`Look, even the ravens on the rooftops are cawing, ``What follows from what?'' 'Ð
Callimachus

Anyone who studies the history of philosophy harbours fears once in a while that we
have simply lost too much information about a philosopher's cultural and intellectual
milieu to reconstruct his view and its motivation. The rediscovery of the Stoics' logical
system gives us some reason to think these fears are justi®ed; until very recently, Stoic
logic was dismissed by modern scholars as a variant of Aristotle's syllogistic logicÐat
best a useless variant, at worst a confused one. The development of the propositional
calculus in the ®rst half of this century, however, led to a reassessment of the Stoics'
logic. Once philosophers had a new logical paradigm to work with, they could
recognize that the Stoic theory had not been Aristotelian at all, but was actually an
early precursor to modern propositional logic. This recovery of what now appears
a fresh and innovative logical system that anticipates many important modern
insights and problems has generated a lot of interest in Stoic logic. In particular,
ever since the publication of Michael Frede's seminal Die Stoische Logik (DSL) a
steady stream of papers and books has attempted to frame the Stoics' logical
system in a modern context.

1. After this piece had already been accepted for publication, it came to my attention that Chapter 4 of
Suzanne Bobzien's recent book, Determinism and Freedom in Stoic Philosophy (1998), contains a
discussion of Stoic conditionals which shares many of the conclusions set forth here. As I arrived at
these conclusions independently, I can only hope that, where my arguments resemble hers, this might
be taken as some con®rmation of the attractiveness of the shared view.

History and Philosophy of Logic ISSN 0144-5340 print/ISSN 1464-5149 online # 2002 Taylor & Francis Ltd
http://www.tandf.co.uk/journals
DOI: 10.1080/0144534021000041236
242 Scott LaBarge

One of the most controversial parts of this broad and multi-faceted project has
surrounded the Stoic theory of implication. Just as modern logicians today, the
ancients struggled to understand what we mean when we use `if±then'
constructions, but if the Hellenistic poet Callimachus is to be believed, the ancients
had a better public relations system for their work; the ancient controversy was
heated enough to generate widespread interest even in society at large. More
interesting to us, however, is the fact that individual philosophers took up positions
about the truth conditions of conditional propositions that are apparently very
similar to positions that yet ®nd defenders. Philo of Megara (late 4th to early 3rd
centuries BCE), for instance, a member of the Dialectical school, endorsed what is
essentially a truth-functional de®nition of the truth conditions of conditionals,
while his teacher Diodorus Cronus (died c. 284 BCE) defended an omni-temporally
true Philonian conditional.2 It is clear that the Stoics, and in particular Chrysippus
Downloaded by [Temple University Libraries] at 22:57 10 June 2016

(c. 280±207 BCE), the third head of the Stoa and its greatest logician, were a party
to this debate, and we have substantial treatments of their logical systems in the
works of, for instance, Diogenes Laertius and Sextus Empiricus. Nevertheless, the
Stoic position on implication has resisted any precise description because of some
apparent inconsistencies in the Stoic account. On the one hand, the Stoics seem to
have de®ned the truth-conditions of implications much along the lines of modern
relevance conditionals; for a conditional to be true there must be some kind of
logically necessary connection between the contents of the antecedent and the
consequent. On the other hand, we know from the works of Cicero, the great
Roman orator and sometime philosopher, that the Stoics also accepted
conditionals that would seem to be unable to meet such strong conditions; in
particular, they endorse some medical inferences that seem clearly to be at best
connected by some kind of empirical necessity. Presently, modern interpreters of
Stoic logic have found themselves splitting on the genuine truth-conditions of Stoic
implications, some endorsing the view that true implications require a logically
necessary connection, others that it requires only an empirically necessary
connection. This essay is an attempt to break this impasse by taking a closer look
at the crucial evidence of Cicero's De fato in the light of another work by CiceroÐ
his often neglected De diviniatione.

2. Stoic implication in Cicero's De fato


An examination of a particular passage in Cicero's De fatoÐFat. 13-17Ðis
absolutely crucial to our understanding of the Stoic theory3 of the truth-conditions
of conditional propositions.4 This passage has been uniquely important in the
debate concerning the kind of connection the antecedent and consequent of a Stoic

2. That is, `If p then q' is true iff there is no time at which `p'= is true and `q' is false. Thus while Philo
accepts the conditional `If it is day then Dion is walking about' so long as it is true that Dion is
walking about, Diodorus rejects it on the grounds that there is some point in time at which it is day
but Dion is not walking about.
3. It is, of course, dangerous to talk of the Stoic view, since it is well-known that there were differences of
opinion in all areas of philosophy among the Stoics, and the possibility of a disagreement on the nature
of conditionals seems suggested by Sextus Empiricus in Outlines of Scepticism (PH) II.110-112, where
Stoics may be Sextus' source for two of four different views of the conditional. It does seem,
however, that Chrysippus' view has a special claim to represent Stoic orthodoxy in dialectic, so
though I only claim to be interpreting Chrysippus' theory of the conditional, I yet feel it is justi®ed
to speak of his view as the Stoic view in a loose sense.
Stoic Conditionals, Necessity and Explanation 243

conditional should have to one another. Frede and Sambursky (1959), for instance,
have argued that the passage proves that the connection is one of logical necessity,
while Sorabji (1980), on the other hand, has argued that positive Stoic attitudes
toward empirical inferences elsewhere suggest that that cannot be the right
interpretation of the passage. I think, however, that both parties to the debate have
missed a position somewhere between them which both renders a connection
between antecedent and consequent that is not merely empirical and makes sense of
the actual uses to which the Stoics put the conditional. This will be an account
which grounds the connection between antecedent and consequent in a proleÃpsis, a
special kind of concept which Chrysippus included in his list of epistemological
criteria of truth and falsity.5
First, however, we should consider some of the other evidence we have about the
nature of Stoic conditionals. The theory that is commonly taken to be the orthodox
Downloaded by [Temple University Libraries] at 22:57 10 June 2016

(Chrysippean) theory of the conditional, the sunarteÃsis or `connection' view, can be


found, for instance, in the text of Diogenes Laertius: `Thus a true conditional is
one the contradictory of whose consequent con¯icts with its antecedent. For
example, ``If it is day, it is light'' ' (DL VII.73).6 We likewise ®nd this theory in
Sextus Empiricus as the unascribed third view of the conditional: `Those who
introduce ``connectedness'' say that a conditional is sound when the opposite of its
consequent con¯icts with its antecedent' (PH II.111).7 What is striking about this
view is that, in contrast to the views of the Dialecticians Philo and Diodorus
Cronus8, Chrysippus requires a direct connection of some kind between the
antecedent and consequent such that the content of the conditional's parts is
relevant (in a vague but fairly demanding way) to the conditional's truth, apart
from any question of those parts' truth or falsity.
Of course, the question that follows is what the connection between antecedent
and consequent must be for the conditional formed from them to be counted
sound, and it is precisely here that the evidence of Cicero's De fato is so crucial.
Our central passage involves Chrysippus' dispute with Diodorus Cronus about the
possibility of unrealized events; Diodorus argues (through the famed Master
Argument9) that since all future events are caused by events that have become
necessary once they have entered the past, the future events are necessary also.
Cicero puts Diodorus' point thus:

For if this is a true conditional, `If you were born with the Dogstar rising, you will
not die at sea,' and the ®rst clause in the conditional, `You were born with the
Dogstar rising', is necessaryÐfor all true statements about past things are
necessary, in the view of Chrysippus . . . if the ®rst clause in the conditional is
necessary, what follows becomes necessary as well. True, Chrysippus does not

4. I share Frede's (1974) and Bobzien's (1986, 1998) concern that `propositions' is a somewhat misleading
translation of axiomata, since `proposition' brings with it a lot of modern logical baggage that is
inappropriate in a Stoic setting (e.g. axiomata are tensed and change modalities, while modern
propositions do not). Nevertheless, so long as we remember that the two are not wholly
interchangeable, the familiarity of `proposition' seems to me suf®cient justi®cation for its use.
5. Diogenes Laertius, Lives of Eminent Philosphers (DL) VII.54.
6. Quotations from Diogenes Laertius are from Long and Sedley's (1987) The Hellenistic Philosophers.
7. Quotations from PH are from Annas' and Barnes' (1994) translation.
8. See PH II.110±112
9. The Master Argument is preserved in Epictetus' Discourses II.19.1±5.
244 Scott LaBarge

think this applies in every case;10 but nevertheless, if there is a cause in nature for
Fabius not dying at sea, Fabius cannot die at sea. (Fat. 14)11

But Cicero immediately gives Chrysippus' response:

At this point Chrysippus, becoming agitated, hopes that the astrologers and the
other diviners can be foiled, and that they will not make use of conditionals but
rather of conjunctions, so that they will not declare their observations as
follows, `If someone was born with the Dogstar rising, that man will not die at
sea', but rather will speak as follows, `It is not the case both that someone was
born with the Dogstar rising and that that man will die at sea'. (Fat. 15)

Cicero then berates Chrysippus for this move, arguing that there is no inference that
Downloaded by [Temple University Libraries] at 22:57 10 June 2016

could not be thus altered into a negated conjunction, whether the inference was
medical, geometrical, or whatever:

. . . I ask, if the astrologers are going to speak in such a way as to assert negations
of inde®nite conjunctions, rather than inde®nite conditionals, why should the
doctors, the geometricians and all the rest not be able to do this? The doctor,
®rst of all, will not put forward what has been observed by him as follows, `If
someone's pulse is like this, that person has a fever', but rather, in the manner
described, `it is not the case both that someone's pulse is like this and that the
person does not have a fever'. Similarly the geometrician will not say `Greatest
circles on a sphere bisect one another', but rather, in the manner described, `it is
not the case both that there are greatest circles on a sphere and that they do not
bisect one another'. What is there that could not in this way be transformed
from a conditional to a negated conjunction? . . . there are many types of
expression, but none more tortuous than this one which Chrysippus hopes the
astrologers will accept for the Stoics' sake. However, none of them does talk
like that; for it is a bigger undertaking to learn these contorted expressions than
to learn the rising and settings of the constellations' (Fat. 16±17).

Cicero's rebuttal of Chrysippus certainly leaves something to be desired; the


comparative ease with which Stoic dialectic may be learned and the astrologers'
actual practice are not arguments against the correctness of Chrysippus' account of
conditionals, and the fact that all inferences can be rendered as negated
conjunctions does not mean they all must be so rendered, as Chrysippus claims the
astrologers' predictions must be. Indeed, Cicero seems to have missed Chrysippus'
point entirely; he is ignoring some crucial difference the Stoics recognize between
the astrologers' predictions and the inferences of a doctor or geometer. What
Cicero does not tell us in this work, and what we should very much like to know, is
what this difference is.

10. The exceptions mentioned here are probably like those recounted by Alexander of Aphrodisias at In
Analytica Priora 177.25±178.1. There Chrysippus claims that `If Dion is dead, then this one [referring
to Dion] is dead' can be a true conditional so long as Dion is alive, but as soon as Dion has died and
`Dion is dead' has thus become necessary, `this one' no longer has a reference, and so the consequent
`this one is dead' is no longer true or false but, in Chrysippus' terms `destroyed'. Questions of the
attractiveness of Chrysippus' view aside, if this is the sort of thing Cicero is referring to in the
passage above, it is not particularly relevant to the case at hand.
11. Quotations from De fato are from Sharples' (1991) translation.
Stoic Conditionals, Necessity and Explanation 245

Of course, a number of answers to this question have been offered, from which I
will discuss a small but representative sample. Perhaps the most in¯uential has been
defended by Frede in DSL, where he argued that the passage was evidence that the
connection manifested in a sound conditional has to be more than empirical:

This spot in Cicero's De fato can thus not be introduced to show that for
Chrysippus incompatibility included empirical incompatibility. One could rather
suppose the exact opposite on the basis of this discussion. If one supposes the
opposite, immediately one has an explanation at hand why Chrysippus forbade
the formulation of astrological theorems as implications. Even in the best case,
if one has faith in divination, they rely on empirical, but not logical,
incompatibility. Therefore perhaps we come closer to an answer to our question
if we seek for the grounds which Chrysippus could have marshalled to hold
Downloaded by [Temple University Libraries] at 22:57 10 June 2016

astrological precepts in the form of implications as basically false. (1974: 86)12

Frede goes on to argue that, since the Stoics were committed to the usefulness and
truthfulness of divination, and furthermore to the claim that diviners could and
had developed general principles for predicting the future, it must be the case that
conditionals grounded on such empirically generated principles still could not meet
the Stoic criterion for soundness in a conditional. Thus the con¯ict between
antecedent and consequent in a sound conditional must require something more
than empirical con¯ict, however law-like that con¯ict is, and more akin to a
logically necessary connection.
Frede's is not, however, the last word on the subject. Sorabji (1980), for instance,
has challenged Frede's interpretation. He recounts a number of problems with Frede's
account, the ®rst two of which Frede himself recognizes as problems. These problems
are:

(1) Chrysippus seems to have allowed that some empirical connections can be cashed
out as conditionals, because Cicero implies that Chrysippus thought some
obviously empirical medical inferences, e.g. from a fast pulse to a fever, could
be appropriately rendered as a conditional (p. 267).
(2) For the Stoics, a sign is a true antecedent in a sound conditional, and Sextus
Empiricus hands down many examples of valid Stoic signs which seem to be
empirically related to the things they signify. `[M]ilk in the breasts is a sign of
having conceived, blushing of shame, sweat of pores in the skin, a viscid
bronchial discharge of a wound in the lungs, bodily movements of the presence
of soul, smoke of ®re, a scar of a wound, and a puncture in the heart of
imminent death' (p. 267).
(3) If, as our evidence strongly suggests, Chrysippus is committed to the view that all
events have a natural cause, and that it is impossible for a cause not to be
followed by its effect, then even if the move from conditional to negated
conjunction is justi®ed in the case of the conditional in which the antecedent
refers only to astrological evidence, once we ®ll in the antecedent with the
genuine natural cause of the consequent (in this case, the thing that genuinely
prevents Fabius from dying at sea), the move to negated conjunction should

12. My translation from the German.


246 Scott LaBarge

no longer be available. If this is so, then the only way Frede can be right about the
truth-conditions for Stoic conditionals is if the relationship between natural
causes and their effects is one of logical necessity, and Sorabji doubts that that
is plausible (p. 268).

Having marshalled these objections, Sorabji concludes that the evidence weighs
against Frede's interpretation and suggests that `Chrysippus had not worked out
when to treat something as a material implication [equivalent to a negated
conjunction] rather than a genuine conditional' (p. 270).
Sorabji's objections are genuinely troubling. They leave us in a nasty muddle, for it
is not immediately clear what Chrysippus' resistance to astrological conditionals
could stem from if not from some kind of mistrust of conditionals grounded only
empirically, but Cicero is right to suggest that it is hard to see why any objection
Downloaded by [Temple University Libraries] at 22:57 10 June 2016

Chrysippus had to astrology would not apply as well to medicine, another


avowedly empirical science. On the other hand, the puzzle also offers an approach
to a solution; if we can ®nd some other difference between the inferences of the
astrologers and those of the doctors, this may provide us with the wedge we need
to explain Chrysippus' different treatment of the two.

3. Stoic implication in the De diviniatione


Cicero's De diviniatione, which he wrote just before De fato, may offer some
purchase on this issue. At a minimum, the work tells us something more about why
astrological inferences are problematic for Chrysippus, for we frequently hear in
the work's ®rst book that the Stoics are committed to the effectiveness of
divination, but we also read repeatedly that they maintained this view despite the
obscurity surrounding the source of this effectiveness. The following passage,
rebutting an attack of the sceptical Academy, is representative of the Stoic position:

The truth is that no other argument of any sort is advanced to show the futility of
the various kinds of divination which I have mentioned except the fact that it is
dif®cult to give the cause or reason of every kind of explanation . . .
You ask why everything happens. You have a perfect right to ask, but that is not
the point at issue now. The question is, Does it happen, or does it not? For
example, if I were to say that the magnet attracted iron and drew it to itself,
and I could not tell you why, then I suppose you would utterly deny that the
magnet had any such power. At least that is the course you pursue in regard to
the existence of the power of divination, although it is established by our own
experience and that of others, by our reading and by the traditions of our
forefathers. (Div. I.85±86)13

This passage clearly indicates that the Stoic commitment to divination rests not on
any genuine understanding of why divination works, but rather on the empirical
fact of its frequent accuracy. When confronted with his fundamental ignorance of
the underlying sources of divination's effectiveness, the Stoic has no response but
to say that the established fact of that effectiveness makes a further response

13. Quotations from De diviniatione are from Falconer's (1923) translation.


Stoic Conditionals, Necessity and Explanation 247

unnecessary. Its empirically demonstrated success removes the need (though not the
hunger!) for an explanation of that success.
This fact about the Stoic defense of divination may give us a better understanding
of Chrysippus' refusal to accept astrologers' inferences as conditionals. Perhaps the
de®ciency of divinatory conditionals is not simply that their foundation is
empirical, but rather that we have no explanation of why the antecedent (e.g.
`Fabius was born at the rising of the Dogstar') and the consequent (e.g. `Fabius
will not die at sea') should have the connection that we have empirical reason to
believe exists. If this is accurate, then the ¯aw of the astrologers' inferences is an
explanatory short-coming.
For this account of Chrysippus' position to be a genuine advance on Frede and
Sorabji, it has to produce some difference between medical and astrological
conditionals that would explain why Chrysippus accepts the former but not the
Downloaded by [Temple University Libraries] at 22:57 10 June 2016

latter. At ®rst glance, this may be problematic, for in De diviniatione Cicero gives
examples of medical inferences we make without understanding why they work as
an argument for an acceptance of divination despite our ignorance of its inner
workings: `We may wonder at the variety of herbs that have been observed by
physicians, of roots that are good for the bites of wild beasts, for eye affections,
and for wounds, and though reason has never explained their force and nature, yet
through their usefulness they have won approval for the medical art and for their
discoverer' (Div. I.13). On the basis of this passage we would have to admit that
medical inferences seem no better explained than divinatory inferences.
That admission, however, would be hasty. After all, the passage above indicates
only that some medical inferences are not well understood, not that all of them are.
The fact that doctors had no explanation for the effectiveness of certain herbs says
nothing about the availability of explanations for the effectiveness of the rest of the
doctor's tools, and there is no reason to think that the obscurity of the powers of
herbs carries over to the rest of medical knowledge. Indeed, if we look at the list of
medical signs which we know the Stoics to have endorsed as valid conditionals, the
connections between the signs and the things they signify intuitively seem less
obscure than the connections between herbs and their effects. Above I gave
Sorabji's brief list of such signs, most of which are medical in some sense: ` . . .
[M]ilk in the breasts is a sign of having conceived, blushing of shame, sweat of
pores in the skin, a viscid bronchial discharge of a wound in the lungs, bodily
movements of the presence of soul . . ., a scar of a wound, and a puncture in the
heart of imminent death' (Sorabji 1980: 267), and we could include Cicero's
example of the quickened pulse indicating a fever (Fat. 15). None of these signs,
however, seems to resist an explanation as the connection between herbs and their
effects does.14 If the Stoics did have an explanation for the effectiveness of these
signs, then there would be a genuine difference between the inferences of the
astrologers and (at least some) medical inferences that could account for
Chrysippus' different treatment of them.
Unfortunately, there is, to my knowledge, no way to verify the claim that the
Stoics recognized such a difference between different kinds of medical inference.
Without a text which records a Stoic explanation of the effectiveness of medical
practices, the postulation of such a difference must remain speculative.

14. I say this, of course, from the perspective of ancient, not modern science. The ancients did not have our
tools to analyze the effectiveness of herbs.
248 Scott LaBarge

Nevertheless, it does not seem unreasonable to think that the Stoics could offer an
explanation of the connection between milk and conception or between a wound to
the heart and death more readily than an explanation of the bene®cial effects of
certain herbs.
This is not, however, all that can be said for drawing the difference between sound
medical conditionals and de®cient astrological conditionals in terms of the availability
of an explanation, for Stoic epistemology provides a tool that seems ready-made for
this distinctionÐproleÃpseis, or preconceptions. ProleÃpseis are concepts formed
naturally in the course of human experience which play a number of roles in Stoic
dialectic, roles which are uniquely suited to give them pride of place in a Stoic
account of explanation.
Downloaded by [Temple University Libraries] at 22:57 10 June 2016

4. ProleÃpsis and explanation


It is perhaps advisable to say a little about the philosophical history of
preconceptions and the uses to which they were put before we discuss their role in
explanation. The term proleÃpsis was ®rst introduced as a philosophical term of art
by Epicurus and seems originally to have represented a solution to Plato's Meno
paradox: the problem of how, starting from ignorance, we could ever know that we
had learned what we set out to learn, since we would not recognize that we had
learned unless we already knew what it was we were supposed to be learning.
Plato's answer to his own problem, namely that what we call learning is really
recollection of knowledge we carried over from past lives, was not available to a
materialist like Epicurus who denied the immortality of the soul. Instead he
whipped up preconceptions to ®ll the gap left by the denial of recollection.
Diogenes Laertius records how this solution is supposed to work:

Preconception, they [the Epicureans] say, is as it were a perception, or correct


opinion, or conception, or universal `stored notion' (i.e. memory), of that which
has frequently become evident externally; e.g. `Such and such a kind of thing is
a man'. For as soon as the word `man' is uttered, immediately its delineation
also comes to mind by means of preconception, since the senses give the lead.
Thus what primarily underlies each name is something self-evident. And what
we inquire about we would not have inquired about if we had not had prior
knowledge of it. For example: `Is what's standing over there a horse or a cow?'
For one must at some time have come to know the form of a horse and that of
a cow by means of preconception. (DL X.33)

Where Plato introduced recollection to provide an epistemological starting point for


what we call `learning', Epicurus introduced preconceptions, which played the same
epistemological role as Plato's remembered Forms without requiring the existence
of an eternal soul. Furthermore, we know from other authors that the Stoics
followed Epicurus in this solution to the Meno paradox.15
Interestingly, neither Epicurus nor the Stoics stopped the employment of
preconceptions at that point. Both capitalized on the supposed universality and
natural origins of these preconceptions in order to use them not only as a starting
point for further investigationÐthough that is all that is needed to free us from the

15. See for example Cicero's Academica I.42 and II.21


Stoic Conditionals, Necessity and Explanation 249

Meno problemÐbut rather took them to be the standard by which we must test
whatever theories we have concerning the object of the preconception. Thus we ®nd
Chrysippus, for example, approving his ethical theory because it `connects best with
the innate preconceptions' (Plutarch, On Stoic self-contradictions 1041E), and
Cicero's Stoic spokesman Balbus in the De natura deorum measures his theory of
the divine nature against our preconception of it (ND II.45). Given this much
broader epistemological role for preconceptions, it is unsurprising that Chrysippus
counted preconception as one of our criteria of the true and false (DL VII.54).
To return to the main issue, how could preconceptions be of use in providing
explanations of unclear things? Given what we have already seen about
preconceptions, they are ideally suited to play an explanatory role for some of the
inferences the Stoics wish to make. After all, preconceptions are general concepts,
developed without the intervention of education, which tell us broadly what these
Downloaded by [Temple University Libraries] at 22:57 10 June 2016

things are like, and when re®ned are even the source of de®nitions.16 In this role of
pre-theoretical evidence, preconceptions both serve as correctives for theories,
allowing us to dismiss theories which con¯ict with them (as, for instance, the Stoics
reject Epicurean theology), and provide evidence on the basis of which we can
make further inferences and extend our knowledge of the object preconceived (as
the Stoics make inferences about the nature of the gods on the basis of our
preconception of them.) Preconceptions are therefore tremendously useful as we
seek to draw or deny connections between properties of one or more things,
connections which can then undergird sound Stoic conditionals.
Furthermore, the Stoics seem to have employed preconceptions just where we
would expect to see something akin to a scienti®c or conceptual explanation. That
is to say, the Stoics do not refer to preconceptions as an epistemological criterion
where, for instance, the most basic sensory information is concerned. When what
we seek to know is the colour of some object, it is the apprehensive appearance
(kataleÃptikeà phantasia) of that object to which we turn. But when the Stoics seek to
test their views on some subject where sensory information is unavailable or
exceedingly limitedÐsubjects like theology (as in much of De natura deorum), ethics
(as in De ®nibus, e.g. III.33±34), or the nature of the true and false (as in Acad.
II.33)Ðthey turn to preconceptions. Likewise, it is dif®cult to imagine the Stoics
explaining the validity of the inferences about abstract geometrical ®gures which
Cicero mentions at De facto 15±16 without relying on the preconceptions of those
®gures we have naturally developed from experience of individual shapes.
Furthermore, we also ®nd the Stoics employing preconceptions in areas we
normally take to be empirical, such as physics. For instance, Alexander of
Aphrodisias records Chrysippus' reference to our various preconceptions of kinds
of mixture as that philosopher's chief argument for the existence of total
blending.17 According to Chrysippus, by witnessing different kinds of blending we
form preconceptions of these different kinds, and by means of these
preconceptions we know not only that the different kinds of blending exist, but
also what their natures are. It is but a small step to imagine preconceptions
playing a similar role in, for instance, the medical inferences which Sorabji lists; if,
for instance, our preconceptions of hearts and of life include something like the
information that the heart is by de®nition an organ essential to the proper

16. See Cicero's Tusculan disputations IV.53 and Augustine's City of God VIII.7
17. See On mixture 216,14±218,6.
250 Scott LaBarge

function of human bodies, then these preconceptions would not only justify the
inference from a heart-wound to imminent death, they would explain why that
connection exists as well. And indeed, in the case of at least one of the medical
inferences the Stoics made, we know the inference was justi®ed by a preconception,
since Sextus reports, ` . . .[T]hat sweat ¯ows is revelatory of the fact that there are
pores, because it is a preconception that liquid cannot be carried through a solid
body' (PH II.142). If this is correct, sound conditionals are sound not merely
because they have empirical veri®cation, nor because they display something as
strong as a logically necessary connection, but rather because they possess
something like a conceptually necessary connection.
It yet remains to be shown why such preconceptions would not exist in the case of
divinatory inferences. For if such a preconception could exist in these cases, providing
an explanatory link between the diviner's sign and the thing he predicts, there would
Downloaded by [Temple University Libraries] at 22:57 10 June 2016

seem to be no reason why Chrysippus should single these inferences out as de®cient
conditionals. Ironically, however, Chrysippus' opponents seem to have believed
that if an explanation was impossible anywhere, that place was divination. Cicero,
attacking the practice of reading entrails, puts the argument bluntly:

But what relationship have [divinatory signs] with the laws of nature? Assuming
that all the works of nature are ®rmly bound together in a harmonious whole
. . ., what connection can there be between the universe and the ®nding of a
treasure? For instance, if the entrails foretell an increase in my fortune and they
do so in accordance with some law of nature, then, in the ®rst place, there is
some relationship between them and the universe, and in the second place, my
®nancial gain is regulated by the laws of nature. Are not the natural
philosophers ashamed to utter such nonsense? And yet a certain contact
between the different parts of nature may be admitted and I concede it. The
Stoics have collected much evidence to prove it. [A list of empirically
established connections follows.]
Concede that it does exist; it does not contravene the point I make, that no sort
of a cleft in a liver is prophetic of ®nancial gain. What natural tie, or what
`symphony', so to speak, or association, or what `sympathy', as the Greeks term
it, can there be between a cleft in a liver and a petty addition to my purse? Or
what relationship between my miserable money-getting, on the one hand, and
heaven, earth, and the laws of nature on the other? (Div. II.33-34).

Surely this is a forceful criticism of divination; it urges that the human mind simply
cannot imagine how the state of an animal's liver and the state of my bankbook
could be connected. Perhaps Chrysippus' rejection of divinatory conditionals was
his way of acknowledging the force of this objection, in which case Fat. 13±17
would record Chrysippus' recognition that the inferences of the astrologers, though
accurate, do not have the same kind of justi®cation as (most) medical inferences.
That is, divinatory inferences cannot point to a preconception which conceptually
links the prophetic sign with the object of prophecy. And this despite the fact that
all inferences, be they astrological, medical, or geometrical, rely fundamentally on
empirical data, whether in the form of individual simple impressions or
preconceptions compounded from these.18

18. See Bobzien (1998: 170), with which I am in full agreement.


Stoic Conditionals, Necessity and Explanation 251

So, once we have made the switch from Frede's insistence on a logically necessary
connection between antecedent and consequent to a conceptually necessary
connection in which the antecedent explains the consequent, we get clear answers to
Sorabji's objections. We can reply to his ®rst two objections that some empirically-
based sign-inferences will be properly rendered as conditionals, and some will not.
If our warrant for making the inference is simple empirically established correlation
and nothing more, as in the case of birth under the Dogstar and not dying at sea, a
conditional statement is not appropriate. But when we have the additional warrant
of a grasp of the relevant concepts by which we see that truth of the antecedent of
the conditional explains the truth of the consequent, the conditional form is ®ne.
That leaves Sorabji's third objection, that an inference involving the true cause of
the diviner's predictive success would be appropriately rendered as a conditional even
if the inference from purely divinatory evidence is not. Nothing I have said cuts
Downloaded by [Temple University Libraries] at 22:57 10 June 2016

against the fact that, even if we do not know why astrological inferences work, the
Stoics may be committed to the existence of natural causes that necessitate the
astrologers' predictions despite our ignorance of the causes' operation. There are
then two possibilities: either the Stoics believe that there are natural causes that
explain why, for instance, Fabius' birth under the Dogstar will prevent him from
dying at sea, or they do not. Let us take them in turn.
First, imagine that the Stoics do believe that some cause in nature does exist that
brings it about that Fabius does not die at sea. If that is so, and if we can come to
understand what that cause is and how it brings its effect about, then it seems clear
that a conditional would be appropriate to report our grasp of that connection. But
if we had that kind of information, then the connection between antecedent and
consequent would no longer be established by empirical correlation alone, but by
the causally explanatory connection between the two as well. In other words, this
new conditional that referred to the true cause of Fabius' survival at sea would
now differ from the astrological inference in just the way that the other sound Stoic
conditionals do. No problem there.
It is more likely, however, that the Stoics would not say that Fabius' survival at sea
had a unique natural cause that explains it. Consider this Stoic passage from Cicero:

If we succeed in [defending divine providence] . . . then surely it must follow that


the gods give to men signs of coming events. But it seems necessary to settle the
principle on which these signs depend. For, according to the Stoic doctrine, the
gods are not directly responsible for every ®ssure in the liver or for every song
of a bird; since manifestly, that would not be seemly or proper in a god and
furthermore is impossible. But, in the beginning, the universe was so created
that certain results would be preceded by certain signs, which are given
sometimes by entrails and by birds, sometimes by lightnings, by portents, and
by stars, sometimes by dreams, and sometimes by utterances of persons in a
frenzy. And these signs do not often deceive the persons who observe them
properly. (Div. I.117-18)

This passage suggests that the Stoics thought that the effectiveness of divination
ultimately springs not from the proper natures of the things that we use as
divinatory signs, but from the providential will of the gods, re¯ected here in the
basic organization of the universe. What, then explains why Fabius' birth under the
Dogstar entails he will not die at sea? The gods wanted it that way. There is
252 Scott LaBarge

nothing about Fabius or the Dogstar or the sea that explains why this connection
exists. It just does, as a sort of brute factÐeven, perhaps, a miracle. If an
explanation of the sort that justi®es a conditional requires something more than
this, as it surely does, then the success of divination is not just unexplained, but is
fundamentally inexplicable, at least by anything more epistemically satisfying than
the claim that `the gods want it that way'.19
We end up, then, with a much more satisfactory account of Stoic conditionals than
anyone has offered hitherto. Callimachus' ravens have been cawing on the rooftops
about what follows from what for a very long time, but they all fell off one side of
the roof or the other. Frede slipped off one side when he set the bar for Stoic
conditionals too high; demanding a logically necessary connection between
antecedent and conditional rules out too many conditionals that the Stoics clearly
accepted. But Cicero and Sorabji set the bar too low; because they did not
Downloaded by [Temple University Libraries] at 22:57 10 June 2016

understand the crucial role that explanation mediated by proleÃpsis plays for the
Stoics, they could not understand why the Stoics exclude the conditionals they
clearly exclude. My hope is that my account will allow us to perch comfortably on
the rooftop and caw to our hearts' content.

Acknowledgements
I owe thanks to Frank Romer, Julia Annas, Michael White, Shaughan Lavine,
and Robert Audi for their helpful responses to early drafts of this essay. Any
remaining shortcomings are, of course, my own.

References
Alexander of Aphrodisias. 1883. In Aristotelis Analyticorum priorum librum I commentarium, M. Wallies.
ed. Berlin.
Alexander of Aphrodisias. 1892. Quaestiones, De fato, De mixtione, I. Bruns. ed. Berlin.
Augustine. 1981. City of God (De civitate Dei), ed. B. Dombart and A. Kalb. Leipzig.
Bobzien, S. 1986. Die stoische Modallogik, WuÈrzburg: KoÈnigshausen and Neumann.
Bobzien, S. 1998. Determinism and Freedom in Stoic Philosophy, Oxford: Clarendon Press.
Cicero. 1914. De Finibus, ed. H Rackham. Cambridge, MA: Harvard UP (Loeb Classical Library).
Cicero. 1923. De Senectute, De Amicitia, De Diviniatione, ed. W.A. Falconer. Cambridge, MA: Harvard
UP (Loeb Classical Library).
Cicero. 1927. Tusculan Disputations, ed. J.E. King. Cambridge, MA: Harvard UP (Loeb Classical Library).
Cicero. 1951. De Natura Deorum, Academica, ed. H. Rackham. Cambridge, MA: Harvard UP (Loeb Clas-
sical Library).
Cicero. 1991. Cicero: On Fate (De Fato), ed. R.W. Sharples. Warminster: Aris and Phillips.
Diogenes Laertius. 1964. Lives of Eminent Philosophers (Vitae Philosophorum), ed. H.S. Long. Oxford:
Oxford UP.
Frede, M. 1974. Die stoische Logik, GoÈttingen: Vandenhoeck and Ruprecht.
Long, A.A. and D. Sedley. 1987. The Hellenistic Philosophers, Cambridge: Cambridge UP.
Plutarch. 1976. `On Stoic Self-Contradictions', in Moralia, v. XIII, part II, ed H. Cherniss. Cambridge,
MA: Harvard UP (Loeb Classical Library).
Sambursky, S. 1959. Physics of the Stoics, London: Routledge and Kegan Paul Ltd.
Sextus Empiricus. 2000. Outlines of Scepticism, ed. J. Annas and J. Barnes. Cambridge: Cambridge UP.
Sorabji, R. 1980. `Causation, Laws, and Necessity', in M. Scho®eld, M. Burnyeat, and J. Barnes Doubt and
Dogmatism, Oxford: Oxford UP.

19. On a side note, this also bears on the connection between herbs and their effects that we mentioned
earlier. Are the effects of herbs unexplained, or inexplicable? I know of no text to decide the issue for
the Stoics, but I suspect they would attribute their lack of an explanation not to the non-existence of
an explanation but to their contingent ignorance of the herbs' natureÐi.e. to the fact that they
lacked proleÃpseis of the herbs' causal properties.

You might also like