James Adam, The Republic of Plato

The reading φυλάξασθαι καὶ λαθεῖν, οὗτος δεινότατος καὶ ἐμποιῆσαι, which has slight MS authority, is defended by Boeckh (Kl. Schr. IV pp. 326 ff.), with whom Zahlfleisch (Zeitschr. f. öst. Gymn. Vol. XXVIII 1877, pp. 603 ff.) and others agree. Boeckh points out that καὶ λαθεῖν (sc. νόσον, according to his view) suggests (from its notion of clandestine cunning) the idea of stealing. This may be admitted, but the idea of stealing is much more forcibly suggested (as Stallbaum points out), if καὶ λαθεῖν is construed with οὗτος δεινότατος κτλ., and this involves the necessity of changing (with Schneider) ἐμποιῆσαι of the MSS to ἐμποιήσας, for the construction λαθεῖν ἐμποιῆσαι, though retained by Campbell, is destitute of authority.

Even if Schneider's emendation be adopted, the argument is (as stated in the notes) fantastical and inconclusive. In order that the conclusion ὅτου τις ἄρα δεινὸς φύλαξ, τούτου καὶ φὼρ δεινός should be valid, φυλάξασθαι should be φυλάξαι, and the objects of the two verbs in proposition (1) should be identical, as well as those in propositions (2) and (3). As it is, if we express φυλάξασθαι in terms of φυλάξαι, they are not identical: for in (1) it is the enemy whom you smite, but yourself whom you guard: in (2) it is yourself (or your patient) whom you guard, but the disease which you secretly implant: in (3) you guard your own army, but steal the enemy's plans, etc. Nevertheless Schneider's emendation is preferable to the traditional reading, which not only contains all the same fallacies as the other, but leaves the three stages of the argument in comparative isolation, attaches the first hint of ‘stealing’ (λαθεῖν) to the wrong member of the clause, and involves the use of the somewhat strained expression λαθεῖν νόσον. It should be added that the change from ἐμποιῇσαι to ἐμποιήσας is not greater than the insertion of καί before ἐμποιῆσαι, and that ἐμποιήσας was very likely to be corrupted under the influence of δεινὸς φυλάξασθαι just before. The emphatic position of καὶ λαθεῖν is necessary to call attention to the first suggestion of the idea contained in κλέψαι; nor can I agree with J. and C. that in Schneider's emendation “the emphasis falls on the wrong word.” In λαθεῖν ἐμποιήσας, which is virtually a single expression, λαθεῖν is more important, in view of the conclusion καὶ κλέπτειν δεινός, than ἐμποιήσας.

Hartman condemns the words καὶ λαθεῖν, and thinks ὅστις and οὗτος have changed places: “cum enim ubique τὸ φυλάξασθαι urgeatur (ὁ πατάξαι δεινότατος, οὗτος καὶ φυλάξασθαι—ὅσπερ κλέψαι..., ὁ αὐτὸς φύλαξ ἀγαθός), requiritur οὗτος δεινὸς φυλάξασθαι, ὅστις δεινότατος κτλ.; quibus tribus exemplis praemissis inversa ratione concludit ὅτου τις ἄρα δεινὸς φύλαξ, τούτου καὶ φὼρ δεινός.” Tucker revives the old conjecture καὶ ἀλθεῖν (‘heal’) instead of καὶ λαθεῖν, and suggests (as an alternative) that λαθεῖν should be μαθεῖν (i.e. καὶ μαθεῖν οὗτος δεινότατος ἐμποιῆσαι ‘clever at learning how to implant’). None of these conjectures appears to me so probable as that of Schneider.