by Allan Marquand

        When we think of the Epicureans we picture a friendly brotherhood in a garden, soothing each other's fears, and seeking to realize a life of undisturbed peace and happiness. It was easy, and to their opponents it became natural, to suppose that the Epicureans did not concern themselves with logic; and if we expect to find in their writings a highly developed formal logic, as that of the Stoics, our search will be in vain. But if we examine the letters of Epicurus, the poem of Lucretius, and the treatise of Philodemus
* with a view to discovering the Epicurean mode of thought, we find a logic which outweighs in value that of their Stoic rivals. This logic is interesting to us, not only because it is the key to that school of Greek Philosophy which outlasted every other, but because a similar logic controls a powerful school of English thought.
* Gomperz: Herkulanische Studien I, Leipzig, 1865. Bahnsch: Des Epicureers Philodemus Schrift Peri sêmeiôn kai sêmeiôseôn. Eine Darlegung ihres Gedankengehalts. Lyck, 1879.

        The logic of Epicurus, like that of J. S. Mill, in opposition to conceptualism, attempts to place philosophy upon an empirical basis. Words with Epicurus are signs of things, and not, as with the Stoics, of our ideas of
/2/ things,* There are, therefore, two methods of inquiry: One seeks for the meanings of words; the other, for a knowledge of things. The former is regarded as a preliminary process; the latter, the only true and necessary way of reaching a philosophy of the universe.
*The hypothesis of lekta, or of immaterial notions, was a conceptualistic inconsistency on the part of the Stoics. The Epicureans and the more consistent empiricists among the Stoics rejected them. Sextus Empiricus, Math. viii, 258.

        All our knowledge is to be brought to the test of sensation, pre-notion, and feeling.
(Diogenes Laertius, x. 31) By these we do not understand three ultimate sources of knowledge. Democritus held to only one source, viz., Feeling; and Epicurus, who inherited his system, implicitly does the same. (Sextus, Math., vii, 140) But each of these modes of feeling has its distinguishing characteristic, and may be used to test the validity of our knowledge. It is the peculiarity of sensation to reveal to us the external world. Sensation reasons not, remembers not; it adds nothing, it subtracts nothing. (D.L., x, 31) What it gives is a simple, self-evident, and true account of the external world. Its testimony is beyond criticism. Error arises after the data of sensation become involved in the operations of intellect. If we should compare this first test of truth with Hume's "impressions," the second test, pre-notion, would correspond with Hume's "ideas." Pre-notions were copies of sensations in a generalized or typical form, arising from a repetition of similar sensations. (D.L., x, 33) Thus the belief in the gods was referred to the clear pre-notions of them. (D.L., x. 123, 124) Single effluxes from such refined beings could have no effect upon the senses, but repeated effluxes from deities sufficiently similar produce in our minds the general notion of a god. (Cicero: De Nat. Deor., I, 49; D.L.x, 139) In the same /3/ manner, but through the senses, the continued observation of horses or oxen produce in us general notions, to which we may refer a doubt concerning the nature of the animal that moves before us.

        The third criterion, Feeling (in the limited sense), was the ultimate test for ethical maxims. The elementary forms are the feeling of pleasure and the feeling of pain. A fourth criterion was added, viz., The Imaginative representations of the intellect. Its use is by no means clear.

        Upon this foundation rises the structure of Epicurean logic. When we leave the clear evidence of sense we pass into the region of opinion, away from the stronghold of truth to the region where error is ever struggling for the mastery of our minds. A true opinion is characterized as one for which there is evidence in favor or none against; a false opinion, one for which there is no evidence in favor or some against.
(D.L., x, 34, 51) The processes by which we pass to the more general and complex forms of knowledge are four: Observation, Analogy, Resemblance, Synthesis.* By Observation, we come into contact with the data of the senses; by Analogy, we may not only enlarge and diminish our perceptions, as we do in conceiving a Cyclops or a Pygmy, but also extend to the unperceived the attributes of our perceptions, as we do in assigning properties to atoms, the soul, and the gods; by Resemblance, we know the appearance of Socrates from having seen his statue; by Synthesis, we combine sensations, as when we conceive of a Centaur.
* D.L., x, 32. The Stoics held a similar view; see D.L., vii, 52.

        As a matter of fact, Epicurus regards only two processes,—Observation and Analogy. Our knowledge, then,
/4/ consists of two parts: (Philodemus: Rhet. lib. iv., i, col. xix) (1) The observed, or phenomena clear and distinct to consciousness; and (2) The unobserved, consisting of phenomena which are yet to be observed, and of hidden causes which lie forever beyond our observation.* The function of logic consists in inference from the observed to the unobserved.** This was called a sign-inference. According to Epicurus there are two methods of making such an inference; one resulting in a single explanation, the other in many explanations. (Ibid, x, 86, 87) The former may be illustrated by the argument Motion is a sign of a void. Here the void is regarded as the only explanation to be given of motion. In other cases many explanations are found equally in harmony with our experience. All celestial phenomena belong to this class. That explanation which alone represents the true cause of such a phenomenon being unknown, we must be content to admit many explanations as equally probable. Thus thunder is explained by supposing either that winds are whirling in the cavities of the clouds, or that some great fire is crackling as it is fanned by the winds, or that the clouds are being torn asunder or are rubbing against each other as they become crystallized. (Ibid., x, 100. Cf. Lucretius, lib. vi. 95-158) In thus connecting celestial and terrestrial phenomena, Epicurus aimed only to exclude supernaturalistic explanations. This done, he was satisfied.
* That is, to prosmenon kai to adêlon, D.L., x. 38
** D.L. x,32. hoden kai peri tôn adêlon apo tôn phainomenon chrê sêmeiousthai

        In the garden at Athens this logic took root and grew; and by the time that Cicero visited Greece and sat at the feet of Zeno—
See Zeller's Stoics, Epicureans, and Sceptics (London, 1880, p. 412, n. 3)—he may have listened to that great /5/ representative of the Epicurean School discussing such questions (Philodemus, Peri sêmeiôn, col. xix-xx) as:

    How may we pass from the known to the unknown?

    Must we examine every instance before we make an induction?

    Must the phenomenon taken as a sign be identical with the thing signified ?

    Or, if differences be admitted, upon what grounds may an inductive inference be made? And,

    Are we not always liable to be thwarted by the existence of exceptional cases?
But such questions had no interest for Cicero. He was too much an orator and rhetorician to recognize the force of the Epicurean opposition to dialectic. The Epicurean logic to him was barren and empty. (Cicero: De Fin., I, 7, 22) It made little of definition; it said nothing of division; it erected no syllogistic forms; it did not direct us how to solve fallacies and detect ambiguities. And how many have been the historians of philosophy who have assigned almost a blank-page to Epicurean logic!

        With a supreme confidence in the truth of sensation and the validity of induction the Epicureans stood in conflict with the other schools of Greek philosophy. The Stoics, treating all affirmation from the standpoint of the hypothetical proposition, acknowledged the validity of such inductions only as could be submitted to the modus tollens. The Sceptics denied the validity of induction altogether. Induction was treated as a sign-inference, and a controversy appears to have arisen concerning the nature of signs, as well as concerning the mode and validity of the inference. The Stoics divided signs into suggestive and indicative.
(See Prandtl's Gen. d. Log., i. 458) By means of a suggestive sign we recall some previously associated fact: as from smoke we infer fire. By indicative signs we infer something otherwise unknown: thus motions of /6/ the body are signs of the soul. Objectively a sign was viewed as the antecedent of a valid conditional proposition, implying a consequent. Subjectively, it was a thought, mediating in some way between things on the one hand, and names and propositions on the other. The Epicureans looked upon a sign as a phenomenon, from whose characters we might infer the characters of other phenomena under conditions of existence sufficiently similar. The sign was to them an object of sense. In considering the variety of signs, the Epicureans appear to have admitted three kinds; but only two are defined in the treatise of Philodemus. (Philod. loc. cit., col. xiv) A general sign is described as a phenomenon which can exist whether the thing signified exists or not, or has a particular character or not. A particular sign is a phenomenon which can exist only on the condition that the thing signified actually exists. The relation between sign and thing signified in the former case is resemblance; in the latter, it is invariable sequence or causality. The Stoics, in developing the sign-inference, inquired, How may we pass from the antecedent to the consequent of a conditional proposition? They replied, A true sign exists only when both antecedent and consequent are true. (Sextus: Math., viii, 258) As a test, we should be able to contrapose the proposition, and see that from the negative of the consequent the negative of the antecedent followed. Only those propositions which admitted of contraposition were allowed to be treated as hypothetical. (Cicero, De Fato, 6,12; 8.15)

        On this propositional ground, therefore, the Epicurean must meet his opponent. This he does by observing that general propositions are obtained neither by contraposition nor by syllogism, nor in any other way than /7/ by induction. (Philod., loc. cit., col. xvii) The contraposed forms, being general propositions, rest also on induction. Hence, if the inductive mode of reasoning be uncertain, the same degree of uncertainty attaches to propositions in the contraposed form.
(Ibid. col. ix) The Stoics, therefore, in neglecting induction, were accused of surrendering the vouchers by which alone their generalizations could be established. (Ibid. col. xxix) In like manner they were accused of hasty generalization, of inaccurate reasoning, of adopting myths, of being rhetoricians rather than investigators of Nature. Into the truth of these accusations we need not inquire. It is enough that they cleared the way for the Epicureans to set up a theory of induction.

        The first question which Zeno sought to answer was, "Is it necessary that we should examine every case of a phenomenon, or only a certain number of cases?"
(Ibid. col. xix, 13-15) Stoics and Sceptics answered, The former is impossible, and the latter leaves induction insecure. But Zeno replied: "It is neither necessary to take into consideration every phenomenon in our experience, nor a few cases at random; but taking many and various phenomena of the same general kind, and having obtained, both from our observation and that of others, the properties that are common to each individual, from these cases may we pass to the rest". (Ibid., col. xx, 30—col. xxi, 3) Instances taken from a class and exhibiting some invariable properties are made the basis of the inductive inference. A certain amount of variation in the properties is not excluded. Thus from the fact that the men in our region of country are short-lived, we may not infer that the inhabitants of Mt. Athos are shortlived also; for "men in our experience are seen to vary considerably in respect to length or brevity of life." (Ibid., col. xvii, 18-22) /8/ Within limits, then, we may allow for variation due to the influence of climate, food, and other physical conditions; but our inference should not greatly exceed the limits of our experience. But, in spite of variations, there are properties which in our experience are universal. Men are found to be liable to disease and old age and death; they die when their heads are cut off, or their hearts extracted; they cannot pass through solid bodies. By induction we infer that these characteristics belong to men wherever they may be found, and it is absurd to speak of men under similar conditions as not susceptible to disease or death, or as having the ability to pass through iron as we pass through the air. (Philod. loc. cit., col. xxi)

The Epicurean looks out upon Nature as already divided and subdivided into classes, each class being closely related to other classes. The inductive inference proceeds from class to class, not in a haphazard way, but from one class to that which resembles it most closely. (Ibid., col. xviii, 20; col. xxviii, 25-29) In case the classes are identical, there is no distinction of known and unknown; and hence, properly speaking, no inductive inference.(Ibid., col. vi, 8-10) In case the classes are widely different, the inference is insecure. But within a certain range of resemblance we may rely as confidently upon an inductive inference as we do upon the evidence of sense. (Ibid., col. xxxiii, 33—col. xxxiv, 34)

        In speaking of the common or essential characters, the basis of induction, it was usual to connect them with the subject of discourse by the words "hêi", "katha", or "para". These words may be taken in four senses: (Ibid., col. xxxiii, 33—col. xxxiv, 34) (1) The properties may be regarded as necessary consequences; so we may say of a man that he is necessarily corporeal and liable to disease and death. (2) Or as essential to the conception or definition of the subject. This is what is /9/ con
veyed in the expression, "Body as body has weight and resistance; man as man is a rational animal." (3) That certain properties are always concomitant. (4) The fourth sense, lost in the lacunae, appears from the following examples to involve degree or proportion: " The sword cuts as it has been sharpened; atoms are imperishable in so far as they are perfect; bodies gravitate in proportion to their weight."

        Zeno's theory of induction may be formulated in the following Canons: —

        Canon I.—If we examine many and various instances of a phenomenon, and find some character common to them all, and no instance appears to the contrary, this character may be transferred to other unexamined individuals of the same class, and even to other closely related classes.

        Canon II. — If in our experience a given character is found to vary, a corresponding amount of variation may be inferred to exist beyond our experience.

        The most important objection made to this theory was, that phenomena exist in our experience exhibiting peculiar and exceptional characters, and that other exceptions might exist beyond our experience to vitiate any induction we may make. The following examples are given:
(Philod., loc. sit., col. i., ii) The loadstone has the peculiar property of attracting iron; amber, of attracting bran; the square number 4 x 4, of having its perimeter equal to its area. Exceptional characters are found in the Alexandrian anvil-headed dwarf, the Epidaurian hermaphrodite, the Cretan giant, the pygmies in Achoris. The sun and moon also are unique; so are time and the soul. Admitting such exceptional phenomena, the Epicurean replies that the belief that a similar state of things exists beyond our experience can /10/ be justified only inductively. (Philod. loc. cit., col. xxv) And exceptional phenomena must be viewed not as closely resembling, but as being widely different from, other phenomena. Inductions concerning loadstones must be confined to loadstones, and not extended to other kinds o£ stones. Each class of exceptional phenomena offered a new field for induction, and hence could be said to strengthen and not to weaken the inductive argument. (Ibid. col. xxiv 10—col. xxv, 2)

        The correctness of all inductions could be tested by the rule of Epicurus for the truth of opinion in general. An induction is true, when all known instances are in its favor, or none against; it is false, when no instances are in its favor, or some against. When the instances are partly one way and partly another, we cannot reach universal conclusions, but only such as are probable. (Ibid. col. xxv, 31-34)

        This theory of induction was completed by a consideration of fallacies, summarized in a work called the "Demetriac." (Ibid., col. xxviii, 13—col. xxix, 24) These consisted in —

     (1) Failing to see in what cases contraposition is applicable.

     (2) Failing to see that we should make inductions not in a haphazard way, but from properties which resemble each other very closely.

     (3) Failing to see that exceptional phenomena are in no way at variance with the inductive inference, but on the other hand add to its force.

     (4) Failing to observe that we infer from the known to the unknown, only when all the evidence is in favor and no shadow of evidence appears to the contrary.

     (5) The failure to perceive that general propositions are derived not by contraposition, but by induction.

When we compare the work of Zeno with that of /11/ Epicurus, an important logical difference is brought to view. Both are occupied with the sign-inference, and look upon inference as proceeding from the known to the unknown. Epicurus, however, sought only by means of hypothesis to explain special phenomena of Nature. Zeno investigated generalizations from experience, with a view to discovering the validity of extending them beyond our experience. This resulted in a theory of induction, which, so far as we know, Epicurus did not possess. In the system of Aristotle, induction was viewed through the forms of syllogism, and its empirical foundation was not held in view. The Epicureans, therefore, were as much opposed to the Aristotelian induction, as they were to the Aristotelian syllogism. It was Zeno the Epicurean who made the first attempt to justify the validity of induction. The record of this attempt will give the treatise of Philodemus a permanent value in the history of inductive

        It is refreshing to see the formalistic and rhetorical atmosphere which had surrounded the subject of logic breaking away, and an honest attempt being made to justify the premises of syllogism. As yet, this had not been done by all the moods of the philosophers.

        It is also interesting to find in the ancient world a theory of induction which rests upon observation, suggests experiment, assumes the uniformity of Nature, and allows for the variation of characters.