The first example from the Symposium occurs in the passage set for transformation into direct speech on p. 182 of Dickey (and further examples occur in this exercise), but she doesn't mention this point in her discussion of indirect discourse. I've seen this frequently in reading, and Paul will have seen this in Herodotus, but when I encountered it in doing the exercise, I had to go hunting in Smyth to confirm the analysis. She does recommend reading just about all of Smyth in the course of her book, and the above-quoted section is part of the recommended reading for Chapter XVIII, on oratio obliqua. So you're expected to be aware of this when you do the exercises. (I have to confess, though, that personally, I didn't read Smyth as diligently as she would have liked as I went through her book.)2631. In subordinate temporal and relative clauses the infinitive is often used for the indicative or optative by attraction to an infinitive standing in the principal clause after a verb of saying. In some cases ἔφη may be mentally inserted.
ἔφη . . . ἐπειδὴ δὲ γενέσθαι ἐπὶ τῇ οἰκίᾳ τῇ Ἀγάθωνος, ἀνεῳγμένην καταλαμβάνειν τὴν θύραν he said that, when he arrived at the house of Agathon, he found the door open P. S. 174d ( = ἐπειδὴ ἐγενόμην, καταλαμβάνω). See also the sentence quoted in 1228 b, end. So οὗτοι δὲ ἔλεγον ὅτι πολλοὺς φαίη Ἀριαῖος εἶναι Πέρσας ἑαυτοῦ βελτίους, οὓς οὐκ ἂν ἀνασχέσθαι αὐτοῦ βασιλεύοντος and they said that Ariaeus said that there were many Persians better than himself, who would not endure his being king X. A. 2.2.1 ( = πολλοί εἰσι ἐμαυτοῦ βελτίους, οἳ οὐκ. ἂν ἀνάσχοιντο ἐμοῦ β.). Here the relative is equivalent, in sense, to καὶ τούτους. The infinitive occurs even in clauses with εἰ (T. 4.98, and often in Hdt.), and with διότι (Hdt. 3.55).
a. The infinitive is rare in such relative clauses as ““διορίζουσι σαφῶς ἐν οἷς ἐξεῖναι ἀποκτιννύναι” they make a clear distinction in cases where it is permitted to kill” D. 23.74.
See also Goodwin, Greek Grammar, sec. 1524.