In XXIX, from 76-79, he introduces the distinction:
cum videat = videns, quia videt
cum… navigaret = dum navigat
From the chapter: Cum Arion … ex Italia in Graeciam navigaret…, nautae pauperes … eum necare constituerunt.
(I don’t know enough Latin to say whether the “cum” in reference to the main action of the verb should be distinguished from “cum” in reference to temporis, only that Orberg doesn’t distinguish.)
Oh my, Lewis and Short (a lie!) on this:
In adverbial clauses of coincident time dependent on preterites ( = eo tempore quo), the clause with cum designating the time at which or during which the main action took place, when, as, while. [The theory of the use of tenses and moods in these clauses is not fully settled. The older grammarians require the indicative if cum denotes pure time, but the subjunctive if denoting cause or relations similar to cause. Zumpt and others acknowledge that the rule is frequently not observed, attributing this to the predilection of the Latin language for the subjunctive. Recently Hoffmann (Zeitpartikeln der Lateinischen Sprache, 1st ed. 1860; 2d ed. 1873) and Lübbert (Syntax von Quom, 1870) have advanced the theory that cum requires the indicative if denoting absolute time, but the subjunctive if denoting relative time. They define absolute time as time co-ordinate or parallel with, or logically independent of, the time of the principal action, which performs the function of a chronological date for the principal action, and they consider it as a criterion that the clause might have constituted an independent sentence; while relative time is logically subordinate to the principal action. Hoffmann condenses his theory in the following words: cum with indicative names and describes the time at which the action of the principal sentence took place; cum with the subjunctive, on the contrary, designates the point of time at which, or the space of time during which, the action expressed in the principal sentence commenced or ended. The chief objections to this theory are: (1) Its vagueness.—(2) The facts that in many instances cum with the subjunctive clearly dates the main action (C. 3. a. β, 2, and 4.; C. 3. a. 5.; C. 3. b. β, 3. and 5.; C. 3. b. γ infra); that many of the subjunctive clauses with cum may be transformed into independent sentences (C. 3. b. β, 2. and 3. infra); that many indicative clauses with cum are logically subordinate to the main action (C. 3. a. α, 2. infra), and that when both moods are used in two co-ordinated clauses with cum belonging to the same main sentence, Hoffmann must account for the difference of the moods by explanations not drawn from his theory (Cic. Agr. 2, 64, 64; id. Clu. 30, 83; id. Div. 1, 43, 97; id. Fin. 2, 19, 61; id. de Or. 67, 272; Caes. B. C. 2, 17; Liv. 6, 40, 17; 30, 44, 10).—(3) The impossibility of clearly drawing the line between logical co-ordination and subordination; and the fact that, wherever it is drawn, there will be many passages not accounted for (cf. 1. init. and many passages under C. 3. a. α, 3.; C. 3. a. δ; C. 3. b. γ, etc.).—(4) That the supposed use of cum with the imperfect indicative is inconsistent with the received doctrine that the imperfect always designates a time relative to another time—a difficulty not satisfactorily met by Hoffman’s assumption of an aoristic imperfect.]GENERAL RULE.—The predicate after cum is in the perfect indicative (or historical present) if the action is conceived as a point of time coincident with the time of the main action. It is either in the imperfect indicative or in the imperfect subjunctive if the action is conceived as occupying a period of time within which the main action took place (e. g.: quid enim meus frater ab arte adjuvari potuit, cum … furem se videre respondit? Quid in omni oratione Crassus … cum pro Cn. Plancio diceret? Cic. de Or. 2, 54, 220; where dicebat might stand for diceret, but not responderet for respondit: cum ad tribum Polliam ventum est, et praeco cunctaretur, etc., Liv. 29, 37, 8; cf.: cum tecum Ephesi collocutus sum, Cic. Fam. 13, 55, 1; and: cum te Puteolis prosequerer, id. ib. 3, 10, 8: cum primum lex coepta ferri est, Liv. 3, 14, 4; and: cum ferretur lex, id. 5, 30, 4; also, Cic. Fam. 4, 3, 1, and Liv. 3, 58, 7).
Attributively with nouns denoting time (tempus, dies, etc.), in ordinary sentences.
With pres. or fut. indic.: incidunt saepe tempora cum ea commutantur, Cic. Off. 1, 10, 31: longum illud tempus cum non ero, etc., id. Att. 12, 8, 1; id. Verr. 2, 5, 69, § 177; id. Quint. 2, 8; id. Sen. 23, 84.—With potential subj., Cic. Att. 3, 3.—
With past tenses, indic., Plaut. Am. prol. 91; id. Rud. 2, 6, 12; Ter. And. 883: atque ille eo tempore paruit cum parere senatui necesse erat, Cic. Lig. 7, 20: memini noctis illius cum … pollicebar, id. Planc. 42, 101; id. Phil. 2, 18, 45; 2, 35, 88; id. Imp. Pomp. 15, 44; id. Sest. 7, 15; 29, 62; id. Sull. 18, 52; id. Fam. 11, 8, 1; 11, 27, 3; id. de Or. 1, 11, 45; Sall. J. 31, 20; Ov. Tr. 4, 10, 6; Prop. 1, 10, 5; 1, 22, 5; Gell. 1, 23, 2 et saep.—So with nouns implying time: illa pugna quom, etc. ( = in quā), Plaut. Poen. 2, 26; Marcellino Consule, cum ego … putabam ( = anno Marcellini, quo, etc.), Cic. Att. 9, 9, 4: patrum nostrorum memoriā cum exercitus videbatur ( = tempore quo), Caes. B. G. 1, 40; Cic. Fam. 13, 1, 2; Liv. 6, 40, 17.—
With preterites in subj., Ter. Hec. 652: > accepit enim agrum iis temporibus cum jacerent pretia praediorum> , Cic. Rosc. Com. 12, 33; so id. Off. 2, 19, 65: > numerandus est ille annus cum obmutuisset senatus> ? id. Pis. 12, 26; so id. Verr. 2, 4, 35, § 77; id. Rep. 2, 37, 62; id. Font. 3, 6; Liv. 3, 65, 8: > haec scripsi postridie ejus diei cum castra haberem Mopsuhestiae > (cf. habebam, as epistolary tense), Cic. Fam. 3, 8, 10.—If the clause does not define the noun, but is a co-ordinate designation of time, it follows the rule of adverbial clauses: > eodem anno, cum omnia infida Romanis essent, Capuae quoque conjurationes factae> , while, Liv. 9, 26, 5; Cic. Rep. 2, 36, 61; id. de Or. 2, 3, 12; Liv. 8, 15, 1; 1, 41, 6.—