SOLMSEN, FRIEDRICH, Abdera's Arguments for the Atomic Theory , Greek, Roman and 
Byzantine Studies, 29:1 (1988:Spring) p.59 

Abdera's Arguments for 
the Atomic Theory 

Friedrich Solmsen 

H ISTORICAL LINKS between one school of philosophy and its 
predecessor must have a certain intrinsic logic if they are to 
carry conviction. Yet there seem to be instances where a logic 

overconfident of its powers and oblivious of our very limited knowl­
edge constructs links and influences which, while theoretically pos­
sible, are yet on closer examination found to be arbitrary. Long-held 
opinions about the origin of the atomic theory in the minds of Leucip­
pus and Democritus may be a case in point. 

Although Guthrie does not endorse these opinions in their tradi­
tional form, he offers a convenient report about them.l Dealing with 
Melissus, and having discussed his relations with Empedocles and 
Anaxagoras, Guthrie continues: "Finally there are the atomists. In 
this connexion much has been made of the conclusion expressed by 
Melissus in fr.8: 'If there were many things, they would be such as the 
One is.' It is commonly held that the atomism of Leucippus and 
Democritus was a response to this challenge: there were many things 
with the properties of the Eleatic One-indestructibility, homogene­
ity, indivisibility, lack of internal change-though they were micro­
scopic in size and innumerable." Guthrie next points out that these 
entities move (for the concept of empty space has been re-estab­
lished), and by combining build up the variety of physical objects. He 
concludes this paragraph: "Except for the reference to Melissus, this 

1 W. K. C. GUTHRIE, A History of Greek Philosophy II (Cambridge 1965 [hereafter 
'Guthrie']) 117, referring to J. Burnet, Early Greek Philosophy4 (London 1930) 335, 
and to G. S. Kirk and J. E. Raven, The Pre-Socratic Philosophers (Cambridge 1957) 
306, 406. We may add C. Bailey, The Greek Atomists and Epicurus (Oxford 1928) 
44f. The second edition of Kirk and Raven (with M. Schofield [Cambridge 1983]), 
offers a more complicated account (408f) but arrives at an unqualified endorsement 
of Aristotle's construction (409 n.4: "excellent historical sense"; no question of "over­
schematization"). In conformity with such opinions Melissus has in the new edition 
been removed from the neighborhood of Parmenides and Zeno and given a place 
after Anaxagoras and Archelaus and immediately preceding the atomists. The 
wisdom of this decision is open to serious doubt. 

59 



SOLMSEN, FRIEDRICH, Abdera's Arguments for the Atomic Theory , Greek, Roman and 
Byzantine Studies, 29:1 (1988:Spring) p.59 

60 ABDERA'S ARGUMENTS FOR THE ATOMIC THEORY 

is, roughly, Aristotle's account of the atomic theory, and there is no 
reason to doubt it."2 

Before we look at Aristotle's authoritative account, we must, in 
fairness to Guthrie, take notice of his departure from what he presents 
as widely-held opinion. His own reconstruction runs thus (39): "Leu­
cippus looked rather [sci!. than to Melissus] to Parmenides, whose 
theory he modified only in the minimal way3 just mentioned. What 
exists must still be ungenerated and imperishable, unchangeable, in­
capable of being added to or subtracted from, homogeneous, finite 
and a plenum, continuous and indivisible." And there must be void, 
and "the sensible world [can] be accounted for on the supposition that 
there are millions of such solid imperishable entities." 

The Aristotelian passage that has encouraged these reconstructions 
is De gen. et corr. 1.8 (325a23-32): 

A ' ~\," 'Ll ~, " \ \ "Ll r ~ , 
EVIC&'2T'2TO~ uE EXnv cp1Jll1J I\.0'Y0V~' O&T&VE~ '2TpO~ T1JV a&CTIl1JCT&V OP.0I\.0'Y0V-

~ , ,,' ," " .... Ll \ " , 
p.Eva I\.E'Y0VTE~ OVIC aVa&p1JCTOVCT&V OVTE 'YEVECT&V OVTE ."Ilopav OVTE IC&V1JCT&V 

\ \ ~ ~Ll ~" • ~ , ~ \ ~ \ ~.... ' 
ICa& TO '2T1\.1JIlO~ TWV OVTWV, OP.OI\.0'Y1JCTa~ uE TaVTa P.EV TO&~ ."a&Vop.EVO&~, 

~ ~\ \" ,/",.,,,, 9" ~, 

TO&~ uE TO EV ICaTaCTICEVa~OVCT&V W~ OVIC av IC&V1JCT&V OVCTav aVEV ICEVOV, TO TE 
\ \" \ ,.. " ,~\ \ " '" ... ,\ ,,, 

ICEVOV p.1J OV ICa& TOV OVTO~ OVuEV p.1J OV .,,1JCT&V Uva&' TO 'Yap ICVP&W~ OV 
~ ~ ", ~ ~'9 \ ~ ,," ~ ~, " \ ~ ~Ll \ 

'2Tap.'2TI\.1JPE~ OV· al\.I\. Uva& TO TO&OVTOV OVX EV, al\.I\. a'2Tnpa TO '2T1\.1JIlO~ ICa& 
, , ~ \ , ~,' ~~" ~ ~ .... ' Ll ( \ 
aopaTa u&a CTP.&lCpOT1JTa TWV O'YICWV. TaVTa u EV Tcp ICEVCP ."EpECTlla& ICEVOV 

\ 9 ) \ , \, ~ ~ ~, ~ \ .... Ll ' 
'Yap Uva& lCa& CTVV&CTTap.Eva P.EV 'YEVECT&V '2TO&E&V, u&al\.vop.Eva uE ."Ilopav. 

It need not be pointed out how much Guthrie and others owe es­
pecially to the last sentence of this section. In listing a considerable 
number of qualities of Parmenides' (6v that survive in the atom, they 
have gone beyond Aristotle but they have stuck close to his method. 

What is this method? It may be defined as the construction of a 
historical development by separating what in the transition from one 
system to another remains unchanged from what is altered. The re­
sults are undeniably of interest provided they are used with caution 
and not for the wrong purpose. The purpose for which they should not 
be used is the inference of motives, intentions, and deliberate strate­
gies from the final product. Even if the doctrines of the Abderites were 
known in their original wording and in their initial form, they would 
not necessarily disclose what questions, considerations, or desires 
prompted their enterprise. That the Eleatic, and in particular Par­
menides', definition of the (OV furnished criteria of reality must be 
admitted. And given their paramount importance, the Eleatics may 
well have contributed more. 

2 Guthrie 117; cf 392. 
3 Sc. by replacing one entity fulfilling the condition by an infinite number of them. 



SOLMSEN, FRIEDRICH, Abdera's Arguments for the Atomic Theory , Greek, Roman and 
Byzantine Studies, 29:1 (1988:Spring) p.59 

FRIEDRICH SOLMSEN 61 

Jonathan Barnes, besides clearing away misconceptions about an­
cient atomism,4 has the merit of pitting the Abderites against Zeno 
rather than against Parmenides or Melissus. With good reasons, al­
though with less than complete success, he looks for Democritus' 
reply to Zeno's dichotomy. At the same time Barnes, far from dis­
missing the other Eleatics, gives the Parmenidean legacy the full 
measure of its due: "The atomists asked themselves what were the 
properties of onta qua onta. . . . Every substance, they argued, was 
unitary ... and solid. What is solid is by physical necessity indivisible 
or atomic ... eternal ... and also immutable or impassive" (50). In 
content this description is correct; regarding its form I object less to 
the mildly anachronistic use of Aristotelian terms than to the assump­
tion that we know what question started the atomists on their search. 

Is it at all possible to find in Aristotle, his ancient commentators, or 
in any other author trustworthy information about the motives that 
produced the atomic theory? Aristotle De gen. et carr. 1.2 (316a14-
317a2) is the passage that has attracted most attention and has been 
analyzed with the most optimistic expectations. But the assessments 
differ greatly. It is called by Barnes, where he summarizes, "an in­
volved argument of Zenonian flavor," described four pages later as 
"Aristotle's in form even ifit is not so in substance," and again shortly 
afterwards as "Democritus' reply to the Zenon ian argument by posit­
ing physically indivisible atoms."5 

Each of these three parties-Zeno, the Abderites, and Aristotle­
has its sponsors, and some scholars are confident that they may re­
cover the outlook of more than one party.6 Evidently we cannot move 
ahead before we know with whose deliberations we are here dealing. 
As the section is too long to be copied, I shall try to convey its gist by 
as unbiased a summary as proves possible. The section is preceded by 
an emphatic-in fact brutally emphatic-declaration of Democritus' 

4 The Presocratic Philosophers II (London/Boston 1979 [hereafter 'Barnes']) 40-75. 
I appreciate particularly the virtual removal from consideration of theoretical and 
mathematical divisibility. My observations suggest that the Abderites' concern is 
strictly physical. 'Parts of the atom' are an Epicurean innovation; see 68f infra. 

5 Barnes 52, 56f. Note also his comment after reporting the conclusion of 
Aristotle's disquisition: "Zenonian anxiety causes the spots of atomism" (52). 

6 For Zeno see, besides Barnes' inclination towards him, Bailey (supra n.1) 72. The 
Abderites are favored by H. H. Joachim, Aristotle on Coming to Be and Passing Away 
(Oxford 1922) 76 ad 316a13ff; D. J. Furley, Two Studies in the Greek Atomists 
(Princeton 1967) 83; and Guthrie 503f ("Democritus' point-by-point reply to the Ele­
atics"; Guthrie provides further references). R. Sorabji, Time, Creation, and the 
Continuum (Ithaca 1983) 336ff, finds here Zeno's challenge and Democritus' reaction. 
In favor of Aristotle, note Harold Cherniss' admirable caution, Aristotle's Criticism of 
Presocratic Philosophy (Baltimore 1935) 113; and see also J. Mau, Zum Problem des 
Infinitesimalen bei den antiken Atomisten (Berlin 1954) 25ff. 



SOLMSEN, FRIEDRICH, Abdera's Arguments for the Atomic Theory , Greek, Roman and 
Byzantine Studies, 29:1 (1988:Spring) p.59 

62 ABDERA'S ARGUMENTS FOR THE ATOMIC THEORY 

superority to Plato (scil., in devising a theory of indivisibles) because 
as a man thoroughly conversant with physical subjects he chose as 
first principles some of wider scope (or applicability).7 

The subject of the disquisition in 316a 14ff is a body or magnitude 
regarded as divisible through and through (7TavT!1 aLaLpETov). Such 
breaking up in every place-and in every place simultaneously-is 
pronounced to be perfectly possible even if it is unlikely to happen 
(a22f). What would be left? Certainly not a body or any kind of mag­
nitude or extension. Rather what results from such complete division 
is either nothing at all or points (a25-29). On either assumption, it 
proves impossible to put the magnitude again together (a third possi­
bility, scil., that something in the nature of sawdust develops in the 
dividing, and evaporates, is briefly envisaged, and dismissed as con­
tributing nothing to a solution, a34-b2).8 Since the experiment has led 
to such unsatisfactory results (for lJ.T07TOV EIC fJ.~ fJ.E'YE8ii>V fJ.E'YE8os EtvaL, 
b4, b 14ff; and points are not P.E'YE87]), the theory of indivisible magni­
tudes appears to be inescapable (b 14-16; the same conclusion, b29-
34). Yet this theory too is open to objections, for which we are 
referred to other treatises.9 A way out of the dilemma must be found 
and can be found. On renewed consideration, the complete breaking­
up of a body everywhere-the premise of the whole argument-is 
seen to be not only actually but even potentially impossible. Poten­
tially because the breaking up of the whole body into points (CTTL'Yp.aL) 
would be conceivable only if points were contiguous and through 
their contiguity could compose a body. The truth is that points do not 
touch (317a2-12, 12-14). 

A close study of this section with attention given not only to its 
thought but also to its style, and most of all to the temperament dis­
played in it, removes all doubt that we are confronted with Aristotle's 
own intellectual efforts. The vividness and the zeal with which he 
explores every aspect of the question and pursues it into every comer 

7 316a5-14. Words like ~"'P.OI<P'TOS ~' av q,avfl." olulo,s I<a~ q,VU'l<o'is AOyO'S 7rnrf'iuOa, 
(al3f) prompt the unwary to think that a report about Democritus' arguments will 
follow. Actually by MiAOV ~' lUTa, a AEyOP.fV 7rpo'·ovuw lXf' yap a7roplav ... (14), 
Aristotle himself takes over. The reference to Democritus' arguments is motivated by 
the contrast with Plato. E. Asmis, Epicurus' Scientific Method (Ithaca 1984) 254ff, 
understands a 14 differently and proceeds to an interpretation of the argument that 
leaves me unconvinced. Lucretius 1.540-50, 588-83, which she compares, is not 
similar. 

8 Still another possibility-that owing to a special 7relOOS or ft~OS' points could build 
up a body (316bI2-15)-may here be ignored. 

9 Serious objections are developed in De caelo 3.4 (303a3-b8). An objective that 
Aristotle pursues throughout De gen. et. corr. 1.3 is to show that even if atomism 
were a tenable doctrine it could not account satisfactorily for genesis; for true and 
"simple" genesis does not come about through uVYI<P'u,s (see 317a12ff, esp. 17-22). 



SOLMSEN, FRIEDRICH, Abdera's Arguments for the Atomic Theory , Greek, Roman and 
Byzantine Studies, 29:1 (1988:Spring) p.59 

FRIEDRICH SOLMSEN 63 

do not leave us with the impression of an author who serenely reports 
about other thinkers' problems and predicaments. There is nothing 
here of the continuous and progressive decrease of the cut-off parts: 
not only do we fail to find the characteristic words a7T€XOV and 7TpO­
€XOV;lO the breaking-up Aristotle describes is of a quite different type. 
He imagines a magnitude breaking up everywhere simultaneously, 11 

and, in order to bring out the awkward conditions he sees arising, he 
operates repeatedly-at first implicitly, yet at the end quite frankly 
and openly-with his characteristic antithesis of dynamis and entele­
cheia. 12 Moreover, behind the thought experiment here performed are 
Aristotle's own conceptions of every continuum (<TVVEX€s) as divisible, 
and, more importantly, of points (<TTLY/J-al) as having no extension, 
being unable to touch one another, adding nothing to a magnitude 
and therefore incapable of building up a continuum. Points as the 
final result of a divison are not known to have figured in Zeno's 
paradoxes or in the Abderites' arguments for the atom; yet this is the 
hypothesis on which Aristotle's argumentation, throughout a good 
part of the section, is predicated. Whether Aristotle and Zeno agree in 
holding that a 'nothing' cannot produce an increase is of no conse­
quence; nor would it make a difference if the aLalpE<TLS KaTtt TO P.€<TOV at 
al9 is identical with, or refers to, Zeno's characteristic aLXoTop.la. 

As this section in De generatione et corruptione refuses to shed light 
on the motives behind the Abderite conception of atomic magni­
tudes,13 we have to fall back on the very meager-but, on account of 
this very meagerness, less suspect-statement at Physica 1.3 (187a 
Iff): 

,/ 10" '10 ~ \ / ',f,. / ~'" / "" '" " fVWL u fVfuouav TotS' I\.O;'OLS' ap.."oTfpOLS', T~ P.fV on 7TaVTa fV Et TO OV fV 
, tl " \, " ,.., 5::'" ,.., ~ , " I 

ullp.aLVEt, on fun TO P.1I OV, T~ uf fK TlIS' utXOTop.LaS' aTop.a 7TOLlIuaVTES' 

fJ. f ;,fOll. 

That the fVLOL who made such concessions to the Eleatie challenges are 
the Abderites has been generally recognized since Burnet and W. D. 

10 Zeno 29Bl D.-K. The meaning of these words has not been cleared up beyond 
doubt. The best explanation may be G. Vlastos' in Gnomon 32 (1959) 196ff (=R. E. 
Allen and D. Furiey, edd., Studies in Presocratic Philosophy II [London 1975] 177ff). 
The term a:rr/xovTa in 316b29 suggests a condition of separation and can hardly have 
the same meaning as in the Zeno passages. 

11 316a 15ff, 17ff (lfJJ,a, 19ff). Sorabji (supra n.6: 338) notices the incongruity 
between division everywhere and infinite division. 

12 316a15ff, bI9-24. Cf Mau (supra n.6). 
13 I see no need to discuss De gen. et corr. 325a23-b5 (included in 67 A D.-K.), a 

section that sets forth essential theories of Leucippus, stressing the divergence from, 
and opposition to, Eleatic doctrines. There is no statement referring to the origin of 
the Abderite position. 



SOLMSEN, FRIEDRICH, Abdera's Arguments for the Atomic Theory , Greek, Roman and 
Byzantine Studies, 29:1 (1988:Spring) p.59 

64 ABDERA'S ARGUMENTS FOR THE ATOMIC THEORY 

ROSSl4 explained the passage in this sense. They 'yielded' to the 
Eleatic monism by asserting reality for the fJ-~ lJv and they posited 
indivisible magnitudes in response to Zeno's 'cutting'. Aristotle has 
been engaged in a critical scrutiny of Eleatic ontology ever since 
184a15, but Zeno's dichotomies have not yet been mentioned. The 
sudden and completely unprepared appearance in 187alff of the AO­
yos EK dis a'XOTOfJ-,as increases the probability that it (cautiously put) 
played some part in calling the atomic theory into existence. Do we 
find support for this impression? 

One piece of support is immediately available. The very word aTo­
fJ-0V, often translated 'indivisible', means literally 'uncuttable'. Should 
we not read this term as a protest and barrier erected against Zeno's 
TOfJ-~ ds a7rupov? Probably yes. But having expressed distrust of inten­
tions or motivations constructed by modem scholars we need to ask 
two questions: do we find any information on why the Abderites 
wished to bring Zeno's TOfJ-~ to a halt? and what justification did they 
claim for deciding one stage in this TOfJ-~ to be the final one? 

Lucretius 1.551-64 may suggest the answers: 

denique si nullam finem natura parasset 
frangendis rebus, iam corpora materiai 
usque redacta forent aevo frangente priore 
ut nil ex illis a certo tempore posset 

555 conceptum summum aetatis pervaderefinem. 
nam quidvis citius dissolvi posse videmus 
quam rursus refici; quapropter longa diei 
infinita aetas anteacti temporis omnis 
quod fregisset adhuc disturbans dissolvensque 

560 nunquam relicuo reparari tempore posset. 
at nunc nimirum frangendi reddita finis 
certa manet quoniam refici rem quamque videmus 
et fin ita simul generatim tempora rebus 
stare quibus possint aevi contingere florem. 

Clearly what these lines proclaim is the need to erect a safeguard 
against Zeno's TOfJ-~ Els a7rupov, to establish a finis certa as the neces­
sary alternative to nulla finis frangendis rebus. It is a physicist's reply 
and perhaps simpler and less sophisticated than what some of our 
contemporaries have constructed as the intellectual foundation for 
atomism. It is also less sophisticated than the case that Aristotle 
builds up-and immediately proceeds to invalidate-for indivisible 
magnitudes. ls 

14 Burnet (supra n.l) 335fT; W. D. Ross, Aristotle's Physics (Oxford 1936) 480fT. 
IS While formulating the case for Abderite origin, I am far from denying that in the 



SOLMSEN, FRIEDRICH, Abdera's Arguments for the Atomic Theory , Greek, Roman and 
Byzantine Studies, 29:1 (1988:Spring) p.59 

FRIEDRICH SOLMSEN 65 

The point here made is that nature cannot allow the breaking-up to 
continue ad infinitum, and that no theory of nature can be devised on 
such a basis. To be sure, for the actual concept of the atom, as distinct 
from the relatively vague idea of a finis jrangendi, we need more 
specific arguments. But before looking for them we may as well com­
plete our diagnosis of this section in Lucretius. The physicists who 
found it necessary thus to protect their enterprise against theories apt 
to render it futile are, to judge by the tenor of their protests, the 
Abderites, and we may even think of the founder of the school, 
Leucippus, as the author of this protest l6 (though on the whole I have 
found it unprofitable in this paper to distinguish between his thought 
and that of Democritus). Historically and chronologically it makes far 
better sense to pit the Abderites against Zeno than to assign this role 
to Epicurus, who had quite different adversaries to contend with.17 

The p.tYE8os discussed by Zeno in 2981 does not give the impression 
of a physical object, nor are the parts that continue to emerge de­
scribed as results of a physical act of breaking, cutting, or destroy­
ing. ls Very understandably, however, the physical philosophers per­
ceived a bearing of this Grundlagenkrise upon their province of 
thought. Anaxagoras, to judge by 5983, saw his way to absorbing the 
results. The Abderites did not. 

For a physical body of three dimensions 'breaking' (jrangere) seems 
a more appropriate word than 'cutting' (secare, which in a context like 
that of 5.559 would look rather odd if it were to occupy the place of 
jrangere). Still the choice of llTop.ov instead of, say, llpp1JKTov proves 
that Zeno's TOP.~ created the issue. There are noteworthy instances of 
amicable association between 'cutting' and 'breaking' in Lucretius. In 
the listing of all threats to which the atom is immune (1.528-35), we 
meet necfragi necfindi in bina secando (=O,xoTofLla, 533), i.e., the two 
concepts in question placed side by side. And towards the end of Book 
I, in the polemic against Empedocles and other pluralists, we find one 
of their principal errors specified (746-48): 

various stages of transmission from, say, Democritus to Lucretius erosion has been at 
work, leaving us with the skeleton of an argument that had been more subtle and 
elaborate. 

16 For Leucippus' opposition to Zeno cf Bailey (supra n.1) 73. 
17 Asmis (supra n.7) 236 says quite correctly: "By the time of Epicurus the con­

troversy provoked by the Eleatic arguments on being had subsided. The early 
atomists ... were clearly at the center of the controversy." To identify Epicurus' 
adversaries is not easy, pace Furley, BICS 13 (1966) 13ft'. I still include the Stoics 
among them, though I agree about the superior importance of Plato and Aristotle, 
especially the exoteric Aristotle. About the relations between Epicurus and the 
resuscitated Megarians, I look forward to further enlightenment. 

18 Cf G. Vlastos in W. Kaufman, ed., Philosophical Classics (New York 1962) 31. 



SOLMSEN, FRIEDRICH, Abdera's Arguments for the Atomic Theory , Greek, Roman and 
Byzantine Studies, 29:1 (1988:Spring) p.59 

66 ABDERA'SARGUMENTSFOR THE ATOMIC THEORY 

deinde quod om nino finem non esse seeandis 
eorporibus /aciunt neque pausam stare /ragori 
nee prorsum in rebus minimum eonsistere quiequam. 

Anaxagoras is more simply criticized for not allowing cor pori bus fi­
nem esse secandi (884). These polemical sections are Epicurean in 
origin and so is the list of fatal dangers in 528-35. 19 Given Epicurus' 
adherence to most of Democritus' physical doctrines2o it is not as­
tonishing if an association of concepts that had been especially mean­
ingful for the Abderites survives in Epicurus' texts. Some readers may 
prefer to conclude that it was Epicurus who substituted 'breaking' for 
'cutting' and changed the concept of a finis seeandi into a finis fran­
gendi. I have given much thought to this alternative interpretation of 
the evidence but cannot help regarding /rangere as the obvious word 
appropriate for a three-dimensional object. 

The suggestion that Epicurus adopted-along with the concept of 
the atom-an original argument for its necessity may meet with some 
resistance. Actually the situation is more complex. At a later point in 
this paper when studying Epicurus' own derivation of the atom, we 
shall find him reaching his goal along quite different lines and observe 
that his defense against infinite divisibility has been removed to the 
discussion of another problem. Our new suggestion about the origins 
of the atomic theory may well be in need of further clarification and 
refinement, but it seems entitled to a place beside the explanations so 
far in vogue, based as it is on an authentic atomist text and having less 
recourse to speculation. 

Because it was imperative to give the terminal point of the division 
a maximum of stability, the atomon was declared unchanging and 
eternal. In Lucretius soliditas and aeternitas are inseparably tied to 
one another; in Epicurus himself CTTEp£iJ. and ciauIAvTa are concepts 
similarly connected.21 

With aeternitas most of the other predicates ofParmenidean reality 
listed above (59) establish their hold on the atom. Simplicitas and 
immutabilitas, which stand out in the text,22 may be assumed to carry 

19 As W. RosIer has shown in his masterly article, "Lucrez und die Vorsokratiker," 
Hermes 1 0 1 (1971) 48ff, the polemical sections (1.635-910) go back to an Epicurean 
version (or adaptation) of a Hellenistic doxography. 

20 On 1. 526ff see 7lf infra. 
21 See 1.500, 518ff, 538f, 577-83, etc.; Epicurus ad Herod. 54. Epicurus inherited 

these qualities from the Abderites. For the Eleatie origin cf the theories of Guthrie 
and others recorded above (59f). None of these theories makes allowance for thejinis 
frangendi motif (or motive). It seems correct to say that Parmenides was used to 
counteract Zeno. 

22 Simplicitas, 1.542, 574, 612; immutabilitas, 591-98; cf 670ff. Note, however, 



SOLMSEN, FRIEDRICH, Abdera's Arguments for the Atomic Theory , Greek, Roman and 
Byzantine Studies, 29:1 (1988:Spring) p.59 

FRIEDRICH SOLMSEN 67 

many or most of the others with them. In the transition, however, 
from a sublime reality far removed from all secular concerns to a 
minimal entity needed as safeguard, these predicates change their 
status. While losing dignity they acquire concreteness. The distance is 
large between the €oV, which Parmenides called l1.vapxov l1.7TavCTTov and 
by sublimely simple logic saved from genesis and destruction,23 and 
the Democritean or Epicurean atoms which cannot be broken either 
from the outside or from the inside, and which remain unharmed by 
moisture, fire, and "spreading cold. "24 They are eternal because they 
are impenetrable, and they owe their impenetrable quality to their 
hardness. Surely the substitution of many or countless oVTa for one 
was not the only change Parmenides' reality underwent when passing 
from Elea to Abdera. Simplicitas, which conveniently sums up vari­
ous original Eleatic predications,25 might seem more appropriate to 
the partless atom of Democritus than to the Epicurean which has 
acquired parts; still Lucretius does not hesitate to emphasize simplici­
tas in the very same paragraph in which he describes the partes 
(1. 599-612; note especially 606-12). Inasmuch as the partes are not 
separable and the atom remains intrinsically homogeneous, the con­
tinued use of simplicitas for it seems, after all, justified. 

Having discovered what appears to be an Abderite argument, we 
wonder whether in the neighborhood of it there may be lurking one or 
several others that may likewise be restored to the 'inventors'.26 
Zeno's relentless divisions must somehow be reflected in the reason­
ing of 1.615-27: 

615 praeterea nisi erit minimum, parvissima quaeque 
corpora constabunt ex partibus in/mitis, 
quippe ubi dimidiae partis pars semper habebit 
dimidiam partem nec res praefiniet ulla. 
ergo rerum inter summam minimamque quid escit? 

that the paragraph in which the finis is established (551-64) includes no reference to 
soliditas, aeternitas, or simplicitas. This difference from all preceding arguments 
(503-50) and from the program in 499f alerts us to the argument's peculiarity, which, 
I believe, is explained by my hypothesis of its original purpose. 

23 B8.27. Note also 8.2-4, 8, 19-21, and in 8.29 Ka6' EavTo n KElTaL. 
24 See Lucr. 1.528-35. 
25 In B8.22-25, OV3f 3LaLpETov, 1Tav op.o'iov, and (TVVEX£S are particularly significant. 
26 The argument of 1.577-83 that is also directed against nulla frangendis reddita 

corporibus finis must be closely related to the thought of 551-64. One may wonder 
whether someone decided to give the reductio ad absurdum a slightly different form. 
The reasoning is reminiscent of 232-37. In both passages I seem to find aVTLp.apTVpT/­
(TLS, a method illuminatingly discussed by Asmis (supra n.7) 143 and passim; cf her 
index s. v. "counter-witnessing." 



SOLMSEN, FRIEDRICH, Abdera's Arguments for the Atomic Theory , Greek, Roman and 
Byzantine Studies, 29:1 (1988:Spring) p.59 

68 ABDERA'S ARGUMENTS FOR THE ATOMIC THEORY 

620 nil erit ut distet; nam quamvisfunditus omnis 
summa sit infinita, tamen parvissima quae sunt 
ex infinitis constabunt partibus aeque. 
quod quoniam ratio reclamat vera negatque 
credere posse animum, victus fateare necessest 

625 esse ea quae nullis iam praedita partibus extent 
et minima constent natura. quae quoniam sunt, 
ilia quoque esse tibi solida atque aeterna fatendum. 

Here again the threat that must be faced and eliminated is division 
going on ad infinitum, although the protest is this time not uttered in 
the interest of nature and a theory of nature but on behalf of ratio 
vera. The unwelcome and absurd condition here pointed out, scil., no 
difference between the largest and the smallest,27 is the direct outcome 
of Zeno's cutting. Intended originally, we may suppose, as a paradox, 
a provocation and a cause of embarrassment, it has now become 
serviceable to a theory that escapes this state of things. Anaxagoras' 
"Of the small there is no smallest" inevitably comes to mind; for even 
if, as Furley maintains, it has a different orientation28 it yet reflects (as 
does the reference to TOP.~ in the same context) the perspective created 
by Zeno's intellectual adventures. For Lucretius the miminae partes 
at which he insists the division must stop are not the atoms them­
selves, but the atom's parts with which he has acquainted us in the 
paragraph immediately preceding (1.599-614).29 Whatever Epicurus' 
reason for introducing something smaller than the atom,30 it is hard to 
imagine that the minimae partes were in their original conception 
meant to serve as terminus for the Zenonic TOP.~. The atom itself as 
the minimum that rescues physics from the Zenonic predicament 
makes excellent sense and involves no historical or chronological im­
probability. To the atom and its soliditas we are actually brought back 

27 Bailey (supra n.1: 72 n.2) speaks of Lucretius as "employing" here a "half" of 
Zeno's argument against infinite division. But in his commentary, Titi Lucreti Cari de 
rerum natura libr; sex (Oxford 1947) II 703, he describes the proof of 1.615-27 as "a 
commonplace paradox in antiquity." The fullest collection of evidence known to me 
is in Sorabji (supra n.6) 3431£, esp. 344 n.18. 

28 Supra n.6: 37, with reference to S. Sambursky, The Physical World of the Greeks 
(London 1956) 22. 

29 By focusing on Lucretius' procedure I do not mean at all to exclude his depen­
dence on a Greek source. 

30 Furley (supra n.6: chapters 1, 2, and 8) has, in my opinion, made some advance 
without solving the problem completely. Lucretius 1.631-34, if in part obscure, does 
provide some rough indication why atoms must have parts. Most of the conclusions 
reached by Vlastos in his comprehensive discusson of the minimae partes (Isis 54 
[1965] 1211£) seem to me convincing, but I am at a loss how to reconcile Lucretius' 
indications with the line of thought Vlastos pursues. 



SOLMSEN, FRIEDRICH, Abdera's Arguments for the Atomic Theory , Greek, Roman and 
Byzantine Studies, 29:1 (1988:Spring) p.59 

FRIEDRICH SOLMSEN 69 

at the end of this argument, by a tum of thought that comes most 
unexpectedly and lacks logical cogency as well as subtlety (1.626f). 
The very awkwardness suggests that the move was necessary. The 
atom was the end and essence of the argument. 

Although we may not reach the same degree of confidence, it seems 
worth pondering whether the argument of 1.565-76 also goes back to 
the original proponents of atomism. The atoms, it is here argued, 
while themselves hard, offer no obstacle to a perfectly satisfactory ac­
count of soft (mollia) objects. All that is needed is the presence of void 
between the solida. By contrast, physicists who posit mollia primordia 
rerum (570) are unable to account for the existence of hard objects 
since no composition of mollia may conceivably produce them. It is 
not easy to believe that this benefit of their invention should have 
escaped Leucippus or Democritus, whose apxal could thus be con­
sidered superior to those of their contemporaries and 'rivals'-Em­
pedocles, Anaxagoras, and perhaps still others. Even so, our attempt 
to claim the contents of these lines for the Abderites remains in the 
realm of conjecture. 31 

Our study of Lucretius 1.551-64 prompted the question what kind 
of place and what degree of importance the insistence on a finis certa 
might have in the system of Epicurus. His own Letter to Herodotus 
shows by what road he arrived at the atom. Lucretius' Book 1 moves 
by parallel steps, and as it is fuller and offers for our purposes more 
helpful information, we shall use it as a basis for our reconstruction, 
and, after gaining from it as much light as it may yield, tum to the 
Letter to see how far it bears out our conclusions. 

We need not analyze the entire development of doctrine between 
1.149 and 1.550. A few highlights will suffice. In 151-214 Lucretius 
establishes the axiom nihil e nihilo gigni, and in 215-64 the comple­
mentary one, nihil ad nihilum interire. These are traditional Preso­
cratic topics.32 In the latter-nihil interire-the Epicureans (to put it 

31 I am aware of two objections. In the polemical sections at the end of Book 1, the 
disadvantages inherent in mollia principia are played off against Empedocles (1. 743 
and 753ff) and Anaxagoras (847ft) If these sections are a product of the Epicurean 
school composed in reliance on a doxography (cf Rosier [supra n.19] 448ff), I see no 
reason why the author(s) should refrain from using Abderite thoughts (or Abderite 
criticisms of rival systems). More serious may be a point raised by my friend Peter 
M. Smith: Empedocles 873, where the bones are hardened by the agency of fire, and 
perhaps some other passages in Empedocles show what can be done with mollia prin­
cipia. May one suppose that Democritus ignored such instances? 

32 See, among others, Parmenides 88.6-11; Empedocles 812; Anaxagoras 817. For 
an exhaustive account see Robin's note in A. Emout and L. Robin, edd., Lucrece. De 
rerum natura (Bude: Paris 1925) ad 1.150. For the Abderites the meaning of fJ.~ lJv 
excludes genesis from it and passing into it. Cf Guthrie 29f. 



SOLMSEN, FRIEDRICH, Abdera's Arguments for the Atomic Theory , Greek, Roman and 
Byzantine Studies, 29:1 (1988:Spring) p.59 

70 ABDERA'S ARGUMENTS FOR THE ATOMIC THEORY 

cautiously, although most probably it was the master himself) intro­
duced some modifications, two of which should here be recorded. 
Everything that comes to be and passes away has its specific, certa, 
materies; and this materies does not perish, but is eternal (aeterna).33 
Like the Stoics, the Epicureans have borrowed the extremely useful 
hyle concept from Aristotle,34 yet the notion of materies aeterna 
presents a remarkable new departure, which so far as I am aware has 
failed to receive due attention.35 We shall find this concept and the 
reasoning behind it taken up again and given a precise meaning after 
the existence and peculiarity of the primordia have been settled. 

Presocratics after Parmenides were agreed on replacing genesis by 
combination, and destruction by separation.36 For Epicurus, what­
ever his views of genesis,37 combination (ul)"YKPLCrLs) remains of car­
dinal importance. Lucretius 1.483f introduce the basic distinction: 
corpora sunt partim porro primordia rerum partim concilio quae con­
stant principiorum.38 While the Abderites posited atoms and the void 
as the basic reality and archai of their system, Epicurus begins instead 
with corpora and the void, and from the corpora, by the distinction 
just quoted, works his way to the atoms (primordia).39 For the ap­
proach to the latter this makes no small difference.4o When Lucretius 

33 For certa materies and certa semina see 1.167-71; also 173, 176, 189, 203f. For 
aeterna semina, 221; materies certa, 239; aeterna, 239, 245, 518f, 540. 

34 H. Usener's Glossarium Epicureum, edited by M. Gigante and W. Schmid (Rome 
1977), lists for iiA71 ad Pyth. 93 and 112, as well as passages from the Volumina 
Herculanensia, one of which (X 95) is still attributed to Epicurus by the editors. In G. 
Arrighetti, Epicuri opera 1 (Turin 1973) I find iiA71 confidently deciphered or restored 
three times in an unidentified book of Epicurus (32.8.6; 11.17; 12.18). In associating 
iiA71 with the name of Aristotle, I do not intend to dissent from the attractive opinion 
that the concept was used also by other members of the Academy in the last decades 
of Plato's life. It may have been appreciated as a very serviceable alternative to Pla­
to's xwpa. Cf now especially F. H. Sandbach, Aristotle and the Stoics (=PCPS Suppl. 
to [1985]) 34-37 with nn.74-81. 

3S Bailey's laconic comment on materies (supra n.27: ad 1.58) shows little awareness 
of the historical background. It may be awkward to avoid the use of the word 'matter' 
in dealing with the physics of, say, Anaxagoras or Democritus, but indiscriminate re­
sort to this term is bound to create confusion. Sorabji (supra n.6: 357) mentions the 
Epicurean materies aeterna without comment. 

36 See Empedocles 88f, Anaxagoras 817, Democritus A37; there is a good deal of 
less explicit evidence. 

37 For the rehabilitation of genesis in Plato and Aristotle see F. Solmsen, Aristotle's 
System of the Physical World (Ithaca 1960) 44ff, 74, 321 ff. 

38 The return to corpora at 483 is in order, since at 329 Lucretius has broken off his 
discourse on corpora to deal with the void and some other aspirants to reality whose 
claims he rejects (449-82). His first and more general discussion of corpora, which 
ends at 329, while its beginning is less easy to determine, poses problems that will be 
discussed elsewhere. 

39 See Democritus 88 and 125; Epicurus ad Herod. 40. 
40 1.485 sets the tone for what follows: note 499-502. 



SOLMSEN, FRIEDRICH, Abdera's Arguments for the Atomic Theory , Greek, Roman and 
Byzantine Studies, 29:1 (1988:Spring) p.59 

FRIEDRICH SOLMSEN 71 

after 1.483ffocuses on the primordia, their role asfinisfrangendi is by 
no means his first concern. What has to be established before anything 
else is that the primordia are firm and unbreakable, solido corpore. 
Along with soliditas they acquire simplicitas and aeternitas. Three 
arguments (1.505-27) may be intended to prove the soliditas,41 
though it is just as possible, and probably even truer to Lucretius' 
intention, to read them as proving the existence and necessity of the 
prima corpora. All three arguments operate in one way or other with 
the antithesis between body and void. The first (503-10)42 postulates 
that each of these two must exist purum and per se (Ka8' Eavni, 506), 
the second (511-19) that the void needs to be surrounded and, as it 
were, protected by a solidum (which, as it turns out, is a concilium of 
atoms, but the purpose is served no less well). The third argument 
(520-27) declares it impossible that either body or void can fill the 
whole (omne); the body that figures in the reasoning is again the 
corpus primum and solidum; solidum implies that there is no void at 
all within it. 

In 1.528-37 Lucretius, evidently satisfied with his proofs for the 
existence of the corpora prima, dwells on their complete immunity to 
dissolution.43 None of the common dangers associated with the de­
structive powers of the elements-water, fire, etc.-can possibly affect 
them. At 540 the argument changes its direction. Recalling what 
earlier in the Book has been laid down about the aeterna materies as 
being necessary rebus reparandis and to prevent destruction ad ni­
hilum, Lucretius shows that this requirement of aeterna materies is 
implemented by the immortali primordia corpore (545). Their eter­
nity (immortality) and solidity, their function as materies, and their 
capacity-in fact exclusive capacity-ex infinito iam tempore res 

41 Bailey (supra n.27: 688) is able to make a case for soliditas as the demonstran­
dum of all three arguments. In the second it is noteworthy that solidum is also 
associated with the concilium materiai, especially at 1.516f. There may well have 
been a concern to slow the breaking-up process even before the stage of the atom is 
reached. Indicative of such a tendency is 238-47. We can hardly decide whether it 
goes back to the Abderites. 

42 I share Diskin Clay's uneasy feeling about 1.503-10 (Lucretius and Epicurus 
[Ithaca 1983] 129) but hesitate to conclude that Lucretius devised these arguments 
himself. 

43 Bailey (supra n.27: 688) misses the significance of 1.528-39, which for him are 
"an appendix rather than a digression." It is correct that aeternitas is here secured for 
the atom, but what matters more is that the atom itself is now regarded as estab­
lished. This is the reason why in 540ff it takes over the role previously assigned to 
materies (as Bailey quite rightly observes ad 543). The analysis of 1.503-598 as a 
sequence of eight proofs (Bailey 684, 690, 694) plus, it would appear, an appendix is 
misleading since it fails to bring out the special weight of the first three 'proofs' that 
carry the burden of the argument. 



SOLMSEN, FRIEDRICH, Abdera's Arguments for the Atomic Theory , Greek, Roman and 
Byzantine Studies, 29:1 (1988:Spring) p.59 

72 ABDERA'SARGUMENTSFOR THE ATOMIC THEORY 

reparare (550) have thus been established before thefinisfrangendi is 
brought into play at 551. Originally, if our reconstruction is correct, 
the finis frangendi had been the argument and quite probably the 
fundamental argument for the existence and necessity of the atom, 
while indivisibility, eternity, homogeneity, and other Parmenidean 
qualities had served to fortify it. Now the Parmenidean qualities 
create the atom and the necessity to stop the breaking has been 
relegated to a subsidiary place and function. 

In the Letter to Herodotus, too, the first principle (7rPWTOV) put down 
is that nothing comes to be from not-being and that if there were 
destruction into not-being this would mean the end of all things, for 
nothing would be left for their restoration. There follows a statement 
about uwp.aTa and KEVOV, 44 the mainstay of the entire system, and this 
in tum is followed by the division of uwp.aTa into compounds (uvy­
KpluEL~) and "what the compounds are made of": TaVTa a' EUTLV l1.Top.a 

" , Q.... >I ".... .... ' ", '" "'() , () , ........ , KaL ap.ETafJI\TJTa EL7rEp p.TJ P.EI\I\EL 7raVTa EL~ TO p.TJ OV .." apTJUEU aL, al\l\ 
lUXVOVTa iJ7rEPP.EVE'iv EV Ta'i~ aLaAvUEUL TWV uVYKpluEWV (41). The phrase 
El~ TO p.~ 8v «p()ap~uEu()aL corresponds to Lucretius' ad nihilum redire. 

Epicurus adds a further essential requirement for the atom, 7rA~PTJ 
\ I" d h .." r '" b ' tl c.I '" TTJV «PVULV OVTa, an e vanes LUXVOVTa V7rOP.EVELV y OVK EXOVTa 07r'!1 TJ 

Cf7rw~ aLaAv()~uETaL.45 At the end of this cardinal section he repeats: 
<I ,,' " , ~... '''' ' E 'f WUTE Ta~ apxa~ aTop.ov~ avaYKaLov ELVaL uwp.aTWV .."VUEL~. ven 1 no 
indi vidual word in this paragraph (41) corresponds to solidum, ap.ETa­
~ATJTa, 7rA~PTJ, luxvovTa V 7rOP.EVE'iV , and the persistent exclusion of 
aUIAvuL~ do the same service. And l1.TOP.O~ has the full weight of its 
meaning 'uncuttable'. (Later on in 54 the atom, while lacking most of 
the usual qualities, must yet itself remain UTEPEOV KaL aaLaAvTov; the 
former of these adjectives corresponds to solidum.)46 

Having settled these fundamental requirements, Epicurus feels jus­
tified in embarking on the infinite number of atoms and on various 
aspects of their behavior. In the course of this account we learn that 
the process has no beginning, for the atoms (and the void) are ataLa 
(44), but there is no reference to the finis frangendi. A reference to it 
does occur slightly later (56), where Epicurus takes his stand against 
the presence of infinitely numerous atoms in a finite object, and in 
following up this decision declares, T~V El~ l1.7rELPOV TOP.~V E7rL To15AaTTov 
, ,."" () ~ ~ , ~ .... ',I, ~, () , avaLpETEOV ,va p.TJ 7raVTa au EVTJ 7rOLWP.EV Kav TaL~ 7rEPLI\TJ."EUL TWV a pOWV 

44 I accept the text as printed in the edition ofP. Von der Mtihll and G. Arrighetti. 
4S The differences Clay finds (supra n.42: 125f) between ad Herod. 40f and Lu­

cretius 1.483-86 do not affect the substance. I appreciate-here and elsewhere­
Clay's observations about the forcefulness of Lucretius' language. 

46 Note also uTEPEOT71~ in ad Herod. 44 and, for other occurrences of UTEPEOV in 
Epicurean texts, cf Usener (supra n.34). 



SOLMSEN, FRIEDRICH, Abdera's Arguments for the Atomic Theory , Greek, Roman and 
Byzantine Studies, 29:1 (1988:Spring) p.59 

FRIEDRICH SOLMSEN 73 

~ls TO fL~ Sv ava"YKa'WfL~8a Ttt ()VTa 8A{{30VTES KaTavaA{uKEtv.47 These 
clauses and those surrounding them bristle with problems of meaning 
into which we fortunately need not digress since the feature of special 
interest to us is not affected by them. It is a surprise to find the JJ.~ lfv as 
terminus of infinite division. Pedantic as it may seem, we must insist 
on the difference between the frequently occurring motif of disappear­
ance or dissolution into nothing and the progressive <breaking' or 
'cutting'. Both, to be sure, are anathema to physicists, but while the 
one leaves nothing for the 'repair' or rebuilding of things, the other 
only leaves less and less and does not allow the time needed for 
rebuilding. Moreover, while nihil ad nihilum is common ground for 
post-Parmenidean physicists, the protest against continuous division 
points specifically to Abdera. 48 We may note that Lucretius presents 
the argument against ad nilum and that against nulla finis frangendis 
rebus in separate, if successive, sections (1.540-50, 551-64). It is 
futile to speculate whether Epicurus in his authoritative larger work 
avoided the "contamination" he allowed himself in an epitome, or 
whether Lucretius here depends on later members of the school who 
reformed the founder's arguments.49 What matters for us is something 
different. The stand against the ~ls a7rEtpov TOJJ.~ is now made to refute 
the notion of infinitely many atoms present in a finite body. The 
reality and necessity of the atom have been established on other 
grounds. 50 

CHAPEL HILL 

December, 1987 

47 Furley (supra n.6: 13) translates: "we must ... do away with division into smaller 
and smaller parts to infinity, so that we may not make everything weak and in our 
conception of the totals be compelled to grind away things that exist ... into the non­
existent." In my view a8poa are compounds, and Ta C;vTa are not limited to the atoms. 
There follows in Epicurus' text another clause excluding a IA-fTa.{3aau ... fIr CJ.7TfLPOV 
E7Tt Tov>"aTTov. The meaning is very uncertain. Furley (14) may be right in understand­
ing IA-fTa.{3aCTLr as a 'traversing' in one's mind. What interests us is that the finis 
frangendi has no place in Epicurus' derivation of the atom. 

48 Dependence, direct or indirect, conscious or not, on Zeno's notorious passing 
from the smallest to nothing (29Bl) is a definite possibility. Vlastos (supra n.30: 122, 
141) presents a critical analysis of what may be Zenonian thought in the Letter (56f). 

49 For A. A. Long, Hellenistic Philosophy2 (Berkeley/Los Angeles 1986) 32ft', 
"Infinite divisibility implies nothing about reduction to sheer non-existence. Lucre­
tius avoids this fallacy." I do not accept all his reports about divisibility but I cer­
tainly endorse the observations here quoted. 

so On the intellectual level as much as in technical detail, the manuscript owes far 
more than a few words can express to the wise and patient help of Alexander 
Mourelatos. In response to his criticism I have modified some of my positions and 
stated others more precisely. I am also indebted to my colleague Peter M. Smith and, 
for technical assistance, to Darrel Rutkin.