Went hunting for the images of the books CYRILLUS Alexandrinus was digitized from and found them[1]. Apparently google doesn’t index the contents of all their books. I wasn’t finding it by searching on a text string. It turns out that one of the citations was indeed the same kind of digitization error where δ’ was read as δι’ but the other one wasn’t an error. The first citation from De adoratione et cultu in spiritu et veritate was read correctly, at least the line in question was correct:
ἔχοι τὸ ἄμωμον. Εἶεν δι’ ἂν εἰς τύπον ἡμῖν σαφῆ τὰ
The photo of the physical copy of the second citation shows δ’ followed by ἂν.
στὰς, προσθείην δ’ ἂν ὅτι καὶ αὐτοὺς τοὺς προφή-
CYRILLUS Alexandrinus De adoratione et cultu in spiritu et veritate
Volume 68, page 953, line 24
{ΠΑΛΛ.} Πῶς γὰρ οὔ;
{ΚΥΡ.} Οὐκοῦν ἀπόβλητον μὲν τὸ ἄναλκι, καὶ τὸ
οἱονεὶ παρειμένον, δεκτὸν δὲ τὸ εὐσθενὲς, ὡς ἐν
ἀρσένων φύσει ζωγραφούμενον, εἰ παρεζευγμένον
ἔχοι τὸ ἄμωμον. Εἶεν > δι’ ἂν > εἰς τύπον ἡμῖν σαφῆ τὰ
τοιάδε Χριστοῦ. Ἄρσην γὰρ ὁμοῦ καὶ ἄμωμος ὁ
Ἐμμανουὴλ, εἴπερ ἐστὶν ἀληθῶς Θεοῦ δύναμις, καὶ
οὐκ εἰδὼς ἁμαρτίαν. Ἄρσην δὲ καὶ ἑτέρως· προὖ-
χον γὰρ ἀεὶ, καὶ ἐν ἀμείνοσι θατέρου τὸ ἄρσεν· καὶ
πρός γε δὴ τούτῳ καὶ ἡγεμονικώτατον. Προέστηκε
CYRILLUS Alexandrinus Commentarius in Isaiam prophetam
Volume 70, page 344, line 10
συνέχειν αὐτοὺς ἰσχύοντος πρὸς εὐσέβειαν. «Οὐ γὰρ
ἐπ’ ἄρτῳ μόνῳ ζήσεται,» φησὶν, «ὁ ἄνθρωπος, ἀλλ’
ἐπὶ παντὶ ῥήματι ἐκπορευομένῳ διὰ στόματος Θεοῦ.»
Ὕδωρ μὲν οὖν ἀποκαλεῖ τὸν ζωοποιὸν τοῦ Θεοῦ, λόγον·
πηγὰς δὲ τοὺς ἁγίους ἀποστόλους τε καὶ εὐαγγελι-
στὰς, προσθείην > δι’ ἂν > ὅτι καὶ αὐτοὺς τοὺς προφή-
τας· Σωτήριον δὲ τὸν Χριστόν.
If anyone wonders why this matters, people rely on these digital texts to launch arguments of all sorts. Knowing that there are errors in standard sources like TLG is important, particularly for people publishing their ideas. Anyone who has worked for a while with digital resources knows they are not infallible. I recall decades back a NT Greek professor getting very heated about statistical “errors” in Accordance. I found this ironic. Running statistics on grammatical features in a tiny sample is a dubious undertaking to begin with. Accordance was treating punctuation as words which had an impact on the sample. I assume this was fixed eons ago.
[1] http://www.roger-pearse.com/weblog/patrologia-graeca-pg-pdfs/