§19. Sensibus: it is important to observe that the word sensus like αισθησις means two things, (1) one of the five senses, (2) an individual act of sensation. Deus: for the supposed god cf. T.D. II. 67. Non videam: this strong statement is ridiculed in 80. De remo inflexo et de collo columbae: cf. 79, 82. The κωπη εναλος κεκλασμενη and περιστερας τραχηλος are frequently mentioned, along with numerous other instances of the deceptiveness of sense, by Sext. Emp., e.g. Pyrrhon. Hypot. I. 119-121, Adv. Math. VII. 244, 414. Cicero, in his speech of the day before, had probably added other examples, cf. Aug. Cont. Ac. III. 27. Epicurus hoc viderit: see 79, 80. Epic. held all sensation, per se, to be infallible. The chief authorities for this are given in R. and P. 343, 344, Zeller 403, footnote. Lumen mutari: cf. Brut. 261. Intervalla ... diducimus: for this cf. Sext. Pyrrh. I. 118 πεμπτος εστι λογος (i.e. the 5th sceptic τροπος for showing sense to be untrustworthy) ‛ο παρα τας θεσεις (situs) και τα διαστηματα (intervalla) και τους τοπους. Multaque facimus usque eo: Sext. Adv. Math. VII. 258 παντα ποιει μεχρις αν τρανην και πληκτικην σπαση φαντασιαν. Sui iudicii: see for the gen. M.D.F. II. 27; there is an extraordinary instance in Plaut. Persa V. 2, 8, quoted by Goer. Sui cuiusque: for this use of suus quisque as a single word see M.D.F. V. 46.

§20. Ut oculi ... cantibus: Halm after Dav. treats this as a gloss: on the other hand I think it appropriate and almost necessary. Quis est quin cernat: read Madvig's strong remarks on Goerenz's note here (D.F. II. 27). Umbris ... eminentia: Pliny (see Forc.) often uses umbra and lumen, to denote background and foreground, so in Gk. σκια and σκιασμα are opposed to λαμπρα; cf. also σκιαγραφειν, adumbrare, and Aesch. Agam. 1328. Cic. often applies metaphorically to oratory the two words here used, e.g. De Or. III. 101, and after him Quintilian, e.g. II. 17, 21. Inflatu: cf. 86 (where an answer is given) and αναβολη. Antiopam: of Pacuvius. Andromacham: of Ennius, often quoted by Cic., as De Div. I. 23. Interiorem: see R. and P. 165 and Zeller's Socrates and the Socratic Schools, 296. Quia sentiatur: αισθησις being their only κριτηριον. Madv. (without necessity, as a study of the passages referred to in R. and P. and Zeller will show) conj. cui adsentiatur, comparing 39, 58; cf. also 76. Inter eum ... et inter: for the repetition of inter cf. T.D. IV. 32 and Madv. Gram. 470. Nihil interesse: if the doctrine of the Academics were true, a man might really be in pain when he fancied himself in pleasure, and vice versa; thus the distinction between pleasure and pain would be obscured. Sentiet ... insaniat: For the sequence cf. D.F. I. 62 and Wesenberg's fine note on T.D. V. 102.

§21. Illud est album: these are αξιωματα, judgments of the mind, in which alone truth and falsehood reside; see Zeller 107 sq. There is a passage in Sext. Adv. Math. VII. 344, 345 which closely resembles ours; it is too long to quote entire: αισθησεσι μεν ουν μοναις λαβειν ταληθες (which resides only in the αξιωμα) ου δυναται ανθρωπος. ... φυσει γαρ εισιν αλογοι ... δει δε εις φαντασιαν αχθηναι του τοιουτου πραγματος "τουτο λευκον εστι και τουτο γλυκυ εστιν." τωι δε τοιουτωι πραγματι ουκετι της αισθησεως εργον εστιν επιβαλλειν ... συνεσεως τε δει και μνημης. Ille deinceps: deinceps is really out of place; cf. 24 quomodo primum for pr. quom. Ille equus est: Cic. seems to consider that the αξιωμα, which affirms the existence of an abstract quality, is prior to that which affirms the existence of a concrete individual. I can quote no parallel to this from the Greek texts. Expletam comprehensionem: full knowledge. Here we rise to a definition. This one often appears in Sextus: e.g. Adv. Math. VII. ανθρωπος εστι ζωον λογικον θνητον, νου και επιστημης δεκτικον. The Stoic ‛οροι, and this among them, are amusingly ridiculed, Pyrrh. Hyp. II. 208—211. Notitiae: this Cic. uses as a translation both of προληψις and εννοια, for which see Zeller 79, 89. In I. 40 notiones rerum is given. Sine quibus: δια γαρ των εννοιων τα πραγματα λαμβανεται Diog. VII. 42.

§22. Igitur: for the anacoluthia cf. Madv. Gram. 480. Consentaneum: so Sextus constantly uses ακολουθον. Repugnaret: cf. I. 19 and n. Memoriae certe: n. on 106. Continet: cf. contineant in 40. Quae potest esse: Cic. nearly always writes putat esse, potest esse and the like, not esse putat etc., which form is especially rare at the end of a clause. Memoria falsorum: this difficulty is discussed in Plato Sophist. 238—239. Ex multis animi perceptionibus: the same definition of an art occurs in N.D. II. 148, D.F. III. 18 (see Madv.), Quint, II. 17, 41, Sext. Pyrrh. Hyp. III. 188 τεχνην ειναι συστημα εκ καταληψεον συγγεγυμνασμενων ib. III. 250. Quam: for the change from plural to singular (perceptio in universum) cf. n. on I. 38, Madv. D.F. II. 61, Em. 139. Qui distingues: Sext. Adv. Math. VIII. 280 ου διοισει της ατεχνιας ‛η τεχνη. Sextus often comments on similar complaints of the Stoics. Aliud eiusmodi genus sit: this distinction is as old as Plato and Arist., and is of constant occurrence in the late philosophy. Cf. Sext. Adv. Math. XI. 197 who adds a third class of τεχναι called αποτελεσματικαι to the usual θεωρητικαι and πρακτικαι, also Quint. II. 18, 1 and 2, where ποιητικη corresponds to the αποτ. of Sext. Continget: "will be the natural consequence." The notion that the verb contingit denotes necessarily good fortune is quite unfounded; see Tischer on T.D. III. 4. Tractabit: μελλει μεταχειριζεσθαι.

§23. Cognitio: like Germ. lehre, the branch of learning which concerns the virtues. Goer. is quite wrong in taking it to be a trans. of καταληψις here. In quibus: the antecedent is not virtutum, as Petrus Valentia (p. 292 ed. Orelli) supposes and gets into difficulty thereby, but multa. This is shown by etiam; not merely the virtues but also all επιστημη depends on καταληψεις; cf. I. 40, 41, with notes, Zeller 88, R. and P. 367. Stabilem: βεβαιον και αμεταπτωτου. Artem vivendi: "tralaticium hoc apud omnes philosophos" M.D.F. I. 42. Sextus constantly talks about ‛η ονειροπολουμενη περι τον βιον τεχνη (Pyrrh. Hyp. III. 250) the existence of which he disproves to his own satisfaction (Adv. Math. XI. 168 sq). Ille vir bonus: in all ancient systems, even the Epicurean, the happiness of the sapiens must be proof against the rack; cf. esp. D.F. III. 29, 75, T.D. V. 73, Zeller 450, and the similar description of the σοφος in Plato's Gorgias. Potius quam aut: Lamb. ut; but I think C.F. Hermann is right in asserting after Wopkens that Cic. never inserts ut after potius quam with the subj. Tischer on T.D. II. 52 affirms that ut is frequently found, but gives no exx. For the meaning cf. De Off. I. 86, Aug. Cont. Ac. II. 12 who says the sapiens of the Academy must be desertor officiorum omnium. Comprehensi ... constituti: cf. the famous abiit, evasit, excessit, crupit. Iis rebus: note the assumption that the sensation corresponds to the thing which causes it. Adsensus sit ... possint: nearly all edd. before Halm read possunt, but the subj. expresses the possibility as present to the mind of the supposed vir bonus. Cf. Madv. Gram. 368.

§24. Primum: out of place, see on 21. Agere: the dogmatist always held that the sceptic must, if consistent, be ανενεργητος εν βιωι (Sext. Pyrrh. Hyp. I. 23). Extremum: similar attempts to translate τελος are made in D.F. I. 11, 29, V. 17. Cum quid agere: cf. I. 23 for the phrase Naturae accommodatum. a purely Stoic expression, ‛ωμοιωμενον τη φυσει; cf. 38 and D.F. V. 17, also III. 16, Zeller 227, footnote, R. and P. 390. Impellimur: κινουμεθα, Sext. Adv. Math. VII. 391, as often.

§25. Oportet videri: "ought to be seen." For this use cf. 39, 81 and 122 of this book. Videri at the end of this section has the weak sense, "to seem." Lucretius often passes rapidly from the one use to the other; cf. I. 262 with I. 270, and Munro's n., also M.D.F. II. 52, Em. Liv. p. 42. Non poterit: as the Academics allege. Naturae ... alienum: Cic. uses this adjective with the dat, and also with the ablative preceded by ab; I doubt whether the phrase maiestate alienum (without the preposition) can be right in De Div. II. 102, where the best texts still keep it. Non occurrit ... aget: occurrit is probably the perfect. Cf. n. on 127.

§26. Quid quod si: Goer., outrageously reads quid quod si, si. Tollitur: the verb tollere occurs as frequently in this sense as αναιρειν does in Sextus. Lux lumenque: Bentl. dux The expression dux vitae is of course frequent (cf. N.D. I. 40, T.D. V. 5 and Lucretius), but there is no need to alter. Lux is properly natural light, lumen artificial, cf. Ad Att. XVI. 13, 1. lumina dimiseramus, nec satis lucebat, D.F. III. 45 solis luce ... lumen lucernae. There is the same difference between φως and φεγγος, the latter is used for the former (φεγγος ‛ηλιου) just as lumen is for lux (si te secundo lumine his offendere—Ad Att. VII. 26, 1) but not often vice versa. Trans. "the luminary and the lamp of life," and cf. Sext. Adv. Math. VII. 269 where the φαντασια is called φεγγος. Finis: so in the beginning of the Nicom. Eth. Aristot. assumes that the actual existence of human exertion is a sufficient proof that there is a τελος. Aperta: a reminiscence of the frequently recurring Greek terms εκκαλυπτειν, εκκαλυπτικος etc., cf. Sextus passim, and D.F. I. 30. Initium ... exitus = αρχη ... τελος. Tenetur: MSS. tenet, the nom. to which Guietus thought to be ratio above. Αποδειξις: cf. the definition very often given by Sext. e.g. Pyrrh. Hyp. II. 143 λογος δι' ‛ομολογουμενων λημματων (premisses) κατα συναγωγην επιφοραν (conclusion) εκκαλυπτων αδηλον, also Diog. VII. 45, λογον δια των μαλλον καταλαμβανομενων το ‛ηττον καταλαμβανομενον περαινοντα (if the reading be right).

§27. Notio: another trans. of εννοια. Conclusisse: although the Greeks used συμπερασμα instead of επιφορα sometimes for the conclusion of the syllogism, they did not use the verb συμπεραινειν which has been supposed to correspond to concludere. It is more likely to be a trans. of συναγειν, and conclusum argumentum of συνακτικος λογος, which terms are of frequent occurrence. Rationibus progredi: to a similar question Sextus answers, ουκ εστιν αναγκαιον τας εκεινον (the dogmatists) δογματολογιας προβαινειν, πλασματωδεις ‛υπαρχουσας (Adv. Math. VIII. 367). Sapientiae ... futurum est: for the dat. with facio and fio see Madv. Gram. 241, obs. 5, Opusc. I. 370, D.F. II. 79, and cf. 96 of this book. Lex veri rectique: cf. 29; the constitutio veri and the determination of what is rectum in morals are the two main tasks of philosophy. Sapientique satis non sit: so Manut. for the sapientisque sit of the MSS. Halm after Dav. reads sapientis, neque satis sit, which I think is wrong, for if the ellipse be supplied the construction will run neque dubitari potest quin satis sit, which gives the exact opposite of the sense required. Ratum: cf. 141.

§28. Perceptum: thoroughly known and grasped. Similar arguments are very frequent in Sextus, e.g. Adv. Math. VIII. 281, where the dogmatist argues that if proof be impossible, as the sceptic says, there must be a proof to show it impossible; the sceptic doctrine must be provable. Cf. 109 of this book. Postulanti: making it a necessity for the discussion; cf. De Leg. I. 21. Consentaneum esse: ακολουθον ειναι. Ut alia: although others. Tantum abest ut—ut: cf. Madv. Gram. 440 a.

§29. Pressius: cf. De Fato 31, 33, N.D. II. 20, T.D. IV. 14, Hortensius fragm. 46 ed. Nobbe. The word is mocked in 109. Decretum: of course the Academics would say they did not hold this δογμα as stabile fixum ratum but only as probabile. Sextus however Pyrrh. Hyp. I. 226 (and elsewhere) accuses them of making it in reality what in words they professed it not to be, a fixed dogma. Sentitis enim: cf. sentis in D.F. III. 26. Fluctuare: "to be at sea," Halm fluctuari, but the deponent verb is not elsewhere found in Cic. Summa: cf. summa philosophiae D.F. II. 86. Veri falsi: cf. n. on 92. Quae visa: so Halm for MSS. quaevis, which edd. had changed to quae a quovis. Repudiari: the selection depended on the probabile of course, with the Academics. Veri falsique: these words were used in different senses by the dogmatist and the sceptic, the former meant by them "the undestructibly true and false." This being so, the statements in the text are in no sense arguments, they are mere assertions, as Sext. says, ψιλη φασει ισον φερεται ψιλη φασις (A.M. VII. 315), φασει μεν φασις επισχεθησεται (ib. 337). Cognoscendi initium: cf. 26, "This I have," the Academic would reply, "in my probabile." Extremum expetendi: a rather unusual phrase for the ethical finis. Ut moveri non possint: so κινεισθαι is perpetually used in Sext. Est ut opinor: so Halm after Ernesti for sit of the MSS. I think it very likely that the MSS. reading is right, and that the whole expression is an imitation of the Greek ‛ικανος ειοησθω and the like. The subj. is supported by D.F. III. 20, De Off. I. 8, Ad Att. XIII. 14, 3, where ut opinor is thrown in as here, and by Ac. II. 17, D.F. III. 21, 24, N.D. I. 109, where si placet is appended in a similar way.

§§30—36. Summary. With respect to physical science, we might urge that nature has constructed man with great art. His mind is naturally formed for the attainment of knowledge (30). For this purpose the mind uses the senses, and so gradually arrives at virtue, which is the perfection of the reason. Those then who deny that any certainty can be attained through the senses, throw the whole of life into confusion (31). Some sceptics say "we cannot help it." Others distinguish between the absolute absence of certainty, and the denial of its absolute presence. Let us deal with these rather than with the former (32). Now they on the one hand profess to distinguish between true and false, and on the other hold that no absolutely certain method for distinguishing between true and false is possible (33). This is absurd, a thing cannot be known at all unless by such marks as can appertain to no other thing. How can a thing be said to be "evidently white," if the possibility remains that it may be really black? Again, how can a thing be "evident" at all if it may be after all a mere phantom (34)? There is no definite mark, say the sceptics, by which a thing may be known. Their "probability" then is mere random guess work (35). Even if they only profess to decide after careful pondering of the circumstances, we reply that a decision which is still possibly false is useless (36).

§30. Physicis: neuter not masc.; cf. I. 6. Libertatem et licentiam: et = "and even." Libertas = παρρησια as often in Tacitus. Abditis rebus et obscuris: cf. n. on I. 15, and the word συνεσκιασμενος Sext. Adv. Math. VII. 26. Lucem eripere: like tollere (n. on 26), cf. 38, 103 and N.D. I. 6. For the sense see n. on 16, also 61. Artificio: this word is used in Cic. as equivalent to ars in all its senses, cf. 114 and De Or. II. 83. Fabricata esset: the expression is sneered at in 87. Quem ad modum primum: so Halm rightly for MSS. prima or primo, which latter is not often followed by deinde in Cicero. Primum is out of position, as in 24. Appetitio pulsa: = mota, set in motion. For ‛ορμη see 24. Intenderemus: as in the exx. given in 20. Fons: "reservoir," rather than "source" here. It will be noted that συγκαταθεσις must take place before the ‛ορμη is roused. Ipse sensus est: an approach to this theory is made in Plat. Theaet. 185, 191. Cf. especially Sext. Adv. Math. VII. 350 και ‛οι μεν διαφερειν αυτην των αισθησεων, ‛ως ‛οι πλειους, ‛οι δε αυτην ειναι τας αισθησεις ... ‛ης στασεως ηρξε Στρατον. All powers of sensation with the Stoics, who are perhaps imitated here, were included in the ‛ηγεμονικον, cf. n. on I. 38. Alia quasi: so Faber for aliqua. "In vera et aperta partitione nec Cicero nec alius quisquam aliquis—alius dixit, multo minus alius—aliquis," M.D.F. III. 63. Goer. on the other hand says he can produce 50 exx. of the usage, he forbears however, to produce them. Recondit: so the εννοιαι are called αποκειμεναι νοησεις (Plut. De Sto. Repug. p. 1057 a). In Sext. Adv. Math. VII. 373 μνημη is called θησαυρισμος φαντασιων. Similitudinibus: καθ' ‛ομοιωσιν Sext. Pyrr. Hyp. II. 75. Cic. uses this word as including all processes by which the mind gets to know things not immediately perceived by sense. In D.F. III. 33 it receives its proper meaning, for which see Madv. there, and the passages he quotes, "analogies" will here best translate the word, which, is used in the same wide sense in N.D. II. 22 38. Construit: so MSS. Orelli gave constituit. Notitiae: cf. 22. Cic. fails to distinguish between the φυσικαι εννοιαι or κοιναι which are the προληψεις, and those εννοιαι which are the conscious product of the reason, in the Stoic system. Cf. M.D.F. III. 21, V. 60, for this and other inaccuracies of Cic. in treating of the same subject, also Zeller 79. Rerumque: "facts". Perfecta: sapientia, virtus, perfecta ratio, are almost convertible terms in the expositions of Antiocheanism found in Cic. Cf. I. 20.

§31. Vitaeque constantiam: which philosophy brings, see 23. Cognitionem: επιστημην. Cognitio is used to translate καταληψις in D.F. II. 16, III. 17, cf. n. on I. 41. Ut dixi ... dicemus: For the repetition cf. 135, 146, and M.D.F. I. 41. The future tense is odd and unlike Cic. Lamb. wrote dicimus, I would rather read dicamus; cf. n. on 29. Per se: καθ' αυτην, there is no need to read propter, as Lamb. Ut virtutem efficiat: note that virtue is throughout this exposition treated as the result of the exercise of the reason. Evertunt: cf. eversio in 99. Animal ... animo: Cic. allows animus to all animals, not merely anima; see Madv. D.F. V. 38. The rule given by Forc. s.v. animans is therefore wrong. Temeritate: προπετεια, which occurs passim in Sext. The word, which is constantly hurled at the dogmatists by the sceptics, is here put by way of retort. So in Sext. Adv. Math. VII. 260, the sceptic is called εμβροντητος for rejecting the καταληπτικη φαντασια.

§32. Incerta: αδηλα. Democritus: cf. I. 44. Quae ... abstruserit: "because she has hidden." Alii autem: note the ellipse of the verb, and cf. I. 2. Etiam queruntur: "actually complain;" "go so far as to complain." Inter incertum: cf. Numenius in Euseb. Pr. Ev. XIV. 7, 12, διαφοραν ειναι αδηλου και ακαταληπτου, και παντα μεν ειναι ακαταληπτα ου παντα δε αδηλα (quoted as from Carneades), also 54 of this book. Docere: "to prove," cf. n. on 121. Qui haec distinguunt: the followers of Carneades rather than those of Arcesilas; cf. n. on I. 45. Stellarum numerus: this typical uncertainty is constantly referred to in Sext. e.g. P.H. II. 90, 98, A.M. VII. 243, VIII. 147, 317; where it is reckoned among things αιωνιον εχοντα αγνωσιαν. So in the Psalms, God only "telleth the number of the stars;" cf. 110. Aliquos: contemptuous; απονενοημενους τινας. Cf. Parad. 33 agrestis aliquos. Moveri: this probably refers to the speech of Catulus; see Introd. p. 51. Aug. Cont. Ac. III. 15 refers to this passage, which must have been preserved in the second edition.

§33. Veri et falsi: these words Lamb. considered spurious in the first clause, and Halm brackets; but surely their repetition is pointed and appropriate. "You talk about a rule for distinguishing between the true and the false while you do away with the notion of true and false altogether." The discussion here really turns on the use of terms. If it is fair to use the term "true" to denote the probably true, the Academics are not open to the criticism here attempted; cf. 111 tam vera quam falsa cernimus. Ut inter rectum et pravum: the sceptic would no more allow the absolute certainty of this distinction than of the other. Communis: the απαραλλακτος of Sextus; "in whose vision true and false are confused." Cf. κοινη φαντασια αληθους και ψευδους Sext. A.M. VII. 164 (R. and P. 410), also 175. Notam: the σημειον of Sextus; cf. esp. P.H. II. 97 sq. Eodem modo falsum: Sext. A.M. VII. 164 (R. and P. 410) ουδεμια εστιν αληθης φαντασια ‛οια ουκ αν γενοιτο ψευδης. Ut si quis: Madv. in an important n. on D.F. IV. 30 explains this thus; ista ratione si quis ... privaverit, possit dicere. I do not think our passage at all analogous to those he quotes, and still prefer to construe quem as a strong relative, making a pause between quis and quem. Visionem: Simply another trans. of φαντασια. Ut Carneades: see Sext. A.M. VII. 166 την τε πιθανην φαντασιαν και την πιθανην ‛αμα και απερισπαστον και διεξωδευμενην (R. and P. 411). As the trans. of the latter phrase in Zeller 524 "probable undisputed and tested" is imperfect, I will give Sextus' own explanation. The merely πιθανη is that sensation which at first sight, without any further inquiry, seems probably true (Sext. A.M. VII. 167—175). Now no sensation is perceived alone; the percipient subject has always other synchronous sensations which are able to turn him aside (περισπαν, περιελκειν) from the one which is the immediate object of his attention. This last is only called απερισπαστος when examination has shown all the concomitant sensations to be in harmony with it. (Sext. as above 175—181.) The word "undisputed," therefore, is a misleading trans. of the term. The διεξωδευμενη ("thoroughly explored") requires more than a mere apparent agreement of the concomitant sensations with the principal one. Circumstances quite external to the sensations themselves must be examined; the time at which they occur, or during which they continue; the condition of the space within which they occur, and the apparent intervals between the person and the objects; the state of the air; the disposition of the person's mind, and the soundness or unsoundness of his eyes (Sext. 181—189).

§34. Communitas: απαραλλαξια or επιμιξια των φαντασιων; Sext. A.M. VII. 403, P.H. I. 127. Proprium: so Sext. often uses ιδιομα, e.g. A. M. IX. 410. Signo notari: signo for nota, merely from love of variety. The in before communi, though bracketed by Halm after Manut., Lamb. is perfectly sound; it means "within the limits of," and is so used after notare in De Or., III. 186. Convicio: so Madv. Em. 143 corrected the corrupt MSS. readings, comparing Orator 160, Ad Fam. XV. 18. A.W. Zumpt on Pro Murena 13 rightly defines the Ciceronian use of the word, "Non unum maledictum appellatur convicium sed multorum verborum quasi vociferatio." He is wrong however in thinking that Cic. only uses the word once in the plural (Ad Att. II. 18, 1), for it occurs N.D. II. 20, and elsewhere. Perspicua: εναργη, a term used with varying signification by all the later Greek schools. Verum illud quidem: "which is indeed what they call 'true'." Impressum: n. on 18. Percipi atque comprehendi: Halm retains the barbarous ac of the MSS. before the guttural. It is quite impossible that Cic. could have written it. The two verbs are both trans. of καταλαμβανεσθαι; Cic. proceeds as usual on the principle thus described in D.F. III. 14 erit notius quale sit, pluribus notatum vocabulis idem declarantibus. Subtiliter: Cic.'s constant trans. of ακριβως or κατ' ακριβειαν (passim in Sext. e.g. P.H. II. 123). Inaniterne moveatur: MSS. agree in ve for ne, on which see M.D.F. IV. 76. Inaniter = κενως = ψευδως. Cf. n. on I. 35, also II. 47, D.F. V. 3 (inaniter moveri), T.D. IV. 13, De Div. II. 120, 126, 140 (per se moveri), Greek κενοπαθειν (Sext. P.H. II. 49), κενοπαθεια (= inanis motus, Sext. A.M. VIII. 184), κενοπαθηματα και αναπλασματα της διανοιας (ib. VIII. 354), διακενος ‛ελκυσμος (ib. VII. 241), διακενος φαντασια (ib. VIII. 67), and the frequent phrase κινημα της διανοιας. For the meaning see n. on 47. Relinquitur: so in Sext. απολειπειν is constantly used as the opposite of αναιρειν (tollere).

§35. Neminem etc.: they are content to make strong statements without any mark of certainty. Primo quasi adspectu: the merely πιθανη φαντασια is here meant; see 33.

§36. Ex circumspectione, etc.: the διεξωδευμενη; see n. on 33. Primum quia ... deinde: for the slight anacoluthia, cf. M.D.F ed. II. p. 796. Iis visis, etc.: i.e. if you have a number of things, emitting a number of appearances, and you cannot be sure of uniting each appearance to the thing from which it proceeds, then you can have no faith in any appearance even if you have gone through the process required by Carneades' rules. Ad verum ipsum: cf. 40. Quam proxime: cf. 47, and also 7. Insigne: σημειον, the same as nota and signum above. Quo obscurato: so Lamb. for MSS. obscuro which Halm keeps. Cf. quam obscurari volunt in 42 and quo sublato in 33. Argumentum: Cic. seems to be thinking of the word τεκμηριον, which, however, the Stoics hardly use. Id quod significatur: το σημειωντον in Sext.

§§37—40. Summary The distinction of an animal is to act. You must either therefore deprive it of sensation, or allow it to assent to phenomena (37). Mind, memory, the arts and virtue itself, require a firm assent to be given to some phenomena, he therefore who does away with assent does away with all action in life (38, 39).

§37. Explicabamus: 19—21 and 30 (quae vis esset in sensibus). Inanimum: not inanimatum, cf. M.D.F. IV. 36. Agit aliquid: I. 23. Quae est in nostra: Walker's insertion of non before est is needless, cf. n. on I. 40. It is the impact of the sensation from without, not the assent given to it, that is involuntary (Sext. A.M. VIII. 397 το μεν γαρ φαντασιωθηναι αβουλητον ην). For in potestate cf. De Fato 9, N.D. I. 69

§38. Eripitur: cf. 30. Neque sentire: Christ om. neque; but the sceptics throughout are supposed to rob people of their senses. Cedere: cf. εικειν, ειξις in Sext. P.H. I. 193, 230, Diog. VII. 51, των δε αισθητικων μετα ειξεως και συγκαταθεσεως γινονται [‛αι φαντασια]; also 66 of this book. Οικειον: cf. 34. Adsentitur statim: this really contradicts a good deal that has gone before, esp. 20. Memoriam: cf. 22. In nostra potestate: this may throw light on fragm. 15 of the Ac. Post., which see.

§39. Virtus: even the Stoics, who were fatalists as a rule, made moral action depend on the freedom of the will; see n. on I. 40. Ante videri aliquid for the doctrine cf. 25, for the passive use of videri, n. on 25. Adsentiatur: the passive use is illustrated by Madv. Em. 131, the change of construction from infin. to subj. after necesse est on D.F. V. 25. Tollit e vita: so De Fato 29.

§§40—42. Summary. The Academics have a regular method. They first give a general definition of sensation, and then lay down the different classes of sensations. Then they put forward their two strong arguments, (1) things which produce sensations such as might have been produced in the same form by other things, cannot be partly capable of being perceived, partly not capable, (2) sensations must be assumed to be of the same form if our faculties do not enable us to distinguish between them. Then they proceed. Sensations are partly true, partly false, the false cannot of course be real perceptions, while the true are always of a form which the false may assume. Now sensations which are indistinguishable from false cannot be partly perceptions, partly not. There is therefore no sensation which is also a perception (40). Two admissions, they say, are universally made, (1) false sensations cannot be perceptions, (2) sensations which are indistinguishable from false, cannot be partly perceptions, partly not. The following two assertions they strive to prove, (1) sensations are partly true, partly false, (2) every sensation which proceeds from a reality, has a form which it might have if it proceeded from an unreality (41). To prove these propositions, they divide perceptions into those which are sensations, and those which are deduced from sensations; after which they show that credit cannot be given to either class (42). [The word "perception" is used to mean "a certainly known sensation."]

§40. Quasi fundamenta: a trans. probably of θεμελιος or the like; cf. ‛ωσπερ θεμελιος in Sext. A.M. V. 50. Artem: method, like τεχνη, cf. M.D.F. III. 4, Mayor on Iuv. VII. 177. Vim: the general character which attaches to all φαντασιαι; genera the different classes of φαντασιαι. Totidem verbis: of course with a view to showing that nothing really corresponded to the definition. Carneades largely used the reductio ad absurdum method. Contineant ... quaestionem: cf. 22 and T.D. IV. 65 una res videtur causam continere. Quae ita: it is essential throughout this passage to distinguish clearly the sensation (visum) from the thing which causes it. Here the things are meant; two things are supposed to cause two sensations so similar that the person who has one of the sensations cannot tell from which of the two things it comes. Under these circumstances the sceptics urge that it is absurd to divide things into those which can be perceived (known with certainty) and those which cannot. Nihil interesse autem: the sceptic is not concerned to prove the absolute similarity of the two sensations which come from the two dissimilar things, it is enough if he can show that human faculties are not perfect enough to discern whatever difference may exist, cf. 85. Alia vera sunt: Numenius in Euseb. Pr. Ev. XIV. 8, 4 says Carneades allowed that truth and falsehood (or reality and unreality) could be affirmed of things, though not of sensations. If we could only pierce through a sensation and arrive at its source, we should be able to tell whether to believe the sensation or not. As we cannot do this, it is wrong to assume that sensation and thing correspond. Cf. Sext. P.H. I. 22 περι μεν του φαισθαι τοιον η τοιον το ‛υποκειμενον (i.e. the thing from which the appearance proceeds) ουδεις ισως αμφισβητει, περι δε του ει τοιουτον εστιν ‛οποιον φαινεται ζητειται. Neither Carneades nor Arcesilas ever denied, as some modern sceptics have done, the actual existence of things which cause sensations, they simply maintained that, granting the existence of the things, our sensations do not give us correct information about them. Eiusdem modi: cf. 33 eodem modo. Non posse accidere: this is a very remarkable, and, as Madv. (D.F. I. 30) thinks, impossible, change from recta oratio to obliqua. Halm with Manut. reads potest. Cf. 101.

§41. Neque enim: a remark of Lucullus' merely. Quod sit a vero: cf. Munio on Lucr. II. 51 fulgor ab auro. Possit: for the om. of esse cf. n. on I. 29.

§42. Proposita: cf. προτασεις passim in Sext. In sensus: = in ea, quae ad sensus pertinent cf. I. 20. Omni consuetudine: "general experience" εμπειρια, cf. N.D. I. 83. Quam obscurari volunt: cf. I. 33. quod explanari volebant; the em. of Dav. obscurare is against Cic.'s usage, that of Christ quam observari nolunt is wanton without being ingenious. De reliquis: i.e. iis quae a sensibus ducuntur. In singulisque rebus: the word rebus must mean subjects, not things, to which the words in minima dispertiunt would hardly apply. Adiuncta: Sext. A.M. VII. 164 (R. and P. 410) πασηι τη δοκουσηι αληθει καθεσταναι ευρισκεται τις απαραλλακτος ψευδης, also VII. 438, etc.

§§43—45. Summary. The sceptics ought not to define, for (1) a definition cannot be a definition of two things, (2) if the definition is applicable only to one thing, that thing must be capable of being thoroughly known and distinguished from others (43). For the purposes of reasoning their probabile is not enough. Reasoning can only proceed upon certain premisses. Again to say that there are false sensations is to say that there are true ones; you acknowledge therefore a difference, then you contradict yourselves and say there is none (44). Let us discuss the matter farther. The innate clearness of visa, aided by reason, can lead to knowledge (45).

§43. Horum: Lamb. harum; the text however is quite right, cf. Madv. Gram. 214 b. Luminibus: cf. 101. Nihilo magis: = ουδεν μαλλον, which was constantly in the mouths of sceptics, see e.g. Sext. P.H. I. 14. Num illa definitio ... transferri: I need hardly point out that the ‛ορος of the Academics was merely founded on probability, just as their "truth" was (cf. n. on 29). An Academic would say in reply to the question, "probably it cannot, but I will not affirm it." Vel illa vera: these words seem to me genuine, though nearly all editors attack them. Vel = "even" i.e. if even the definition is firmly known, the thing, which is more important, must also be known. In illa vera we have a pointed mocking repetition like that of veri et falsi in 33. In falsum: note that falsum = aliam rem above. For the sense cf. Sext. P.H. II. 209 μοχθηρους ‛ορους ειναι τους περιεχοντας τι των μη προσοντων τοις ‛οριστοις, and the schoolmen's maxim definitio non debet latior esse definito suo. Minime volunt: cf. 18. Partibus: Orelli after Goer. ejected this, but omnibus hardly ever stands for omn. rebus, therefore C.F. Hermann reads pariter rebus for partibus. A little closer attention to the subject matter would have shown emendation to be unnecessary, cf. 42 dividunt in partis, T.D. III. 24, where genus = division, pars = subdivision.

§44. Impediri ... fatebuntur: essentially the same argument as in 33 at the end. Occurretur: not an imitation of εναντιουσθαι as Goer. says, but of απανταν, which occurs very frequently in Sext. Sumpta: the two premisses are in Gk. called together λημματα, separately λημμα and προσληψις (sumptio et adsumptio De Div II. 108). Orationis: as Faber points out, Cic. does sometimes use this word like ratio (συλλογισμος), cf. De Leg. I. 48 conclusa oratio. Fab. refers to Gell. XV. 26. Profiteatur: so ‛υπισχνεισθαι is often used by Sext. e.g. A.M. VIII. 283. Patefacturum: n. on 26, εκκαλυπτειν, εκκαλυπτικος, δηλωτικος (the last in Sext. A.M. VIII. 277) often recur in Greek. Primum esse ... nihil interesse: there is no inconsistency. Carneades allowed that visa, in themselves, might be true or false, but affirmed that human faculties were incapable of distinguishing those visa which proceed from real things and give a correct representation of the things, from those which either are mere phantoms or, having a real source, do not correctly represent it. Lucullus confuses essential with apparent difference. Non iungitur: a supposed case of διαρτησις, which is opposed to συναρτησις and explained in Sext. A.M. VIII. 430.

§45. Assentati: here simply = assensi. Praeteritis: here used in the strong participial sense, "in the class of things passed over," cf. in remissis Orat. 59. Primum igitur ... sed tamen: for the slight anacoluthia cf. Madv. Gram. 480. Iis qui videntur: Goer. is qui videtur, which is severely criticised by Madv. Em. 150. For Epicurus' view of sensation see n. on 79, 80.

§§46—48. Summary. The refusal of people to assent to the innate clearness of some phenomena (εναργεια) is due to two causes, (1) they do not make a serious endeavour to see the light by which these phenomena are surrounded, (2) their faith is shaken by sceptic paradoxes (46). The sceptics argue thus: you allow that mere phantom sensations are often seen in dreams, why then do you not allow what is easier, that two sensations caused by two really existing things may be mistaken the one for the other? (47). Further, they urge that a phantom sensation produces very often the same effect as a real one. The dogmatists say they admit that mere phantom sensations do command assent. Why should they not admit that they command assent when they so closely resemble real ones as to be indistinguishable from them? (48)

§46. Circumfusa sint: Goer. retains the MSS. sunt on the ground that the clause quanta sint is inserted παρενθετικως! Orelli actually follows him. For the phrase cf. 122 circumfusa tenebris. Interrogationibus: cf. I. 5 where I showed that the words interrogatio and conclusio are convertible. I may add that in Sextus pure syllogisms are very frequently called ερωτησεις, and that he often introduces a new argument by ερωταται και τουτο, when there is nothing interrogatory about the argument at all. Dissolvere: απολυεσθαι in Sext. Occurrere: cf. 44.

§47. Confuse loqui: the mark of a bad dialectician, affirmed of Epicurus in D.F. II. 27. Nulla sunt: on the use of nullus for non in Cic. cf. Madv. Gram. 455 obs. 5. The usage is mostly colloquial and is very common in Plaut. and Terence, while in Cic. it occurs mostly in the Letters. Inaniter: cf. 34. There are two ways in which a sensation may be false, (1) it may come from one really existent thing, but be supposed by the person who feels it to be caused by a totally different thing, (2) it may be a mere φαντασμα or αναπλασμα της διανοιας, a phantom behind which there is no reality at all. Quae in somnis videantur: for the support given by Stoics to all forms of divination see Zeller 166, De Div. I. 7, etc. Quaerunt: a slight anacoluthon from dicatis above. Quonam modo ... nihil sit omnino: this difficult passage can only be properly explained in connection with 50 and with the general plan of the Academics expounded in 41. After long consideration I elucidate it as follows. The whole is an attempt to prove the proposition announced in 41 and 42 viz. omnibus veris visis adiuncta esse falsa. The criticism in 50 shows that the argument is meant to be based on the assumption known to be Stoic, omnia deum posse. If the god can manufacture (efficere) sensations which are false, but probable (as the Stoics say he does in dreams), why can he not manufacture false sensations which are so probable as to closely resemble true ones, or to be only with difficulty distinguishable from the true, or finally to be utterly indistinguishable from the true (this meaning of inter quae nihil sit omnino is fixed by 40, where see n.)? Probabilia, then, denotes false sensations such as have only a slight degree of resemblance to the true, by the three succeeding stages the resemblance is made complete. The word probabilia is a sort of tertiary predicate after efficere ("to manufacture so as to be probable"). It must not be repeated after the second efficere, or the whole sense will be inverted and this section placed out of harmony with 50. Plane proxime: = quam proxime of 36.