I have some quibbles with the Aramaic, though. For instance: Casey discusses A. Meyer's idea that there is an underlying Aramaic wordplay at Matt. 3:9 || Luke 3:8: "God is able of these stones (Gk. lithon = Aram. abnayya ) to raise up children (Gk. tekna = Aram. benayya) unto Abraham." This suggestion has made its way into many a commentary.And Ed goes on to make some further useful points. The passage in question also raises other concerns with retroversion (back translation) projects. One difficulty is that any given word might have a variety of words from which it could be alleged to have derived, and this can provide the invitation to the scholar looking to discover puns, alliteration, hidden meanings and the like. When you add this to the fact that there are many passages in the Gospels that do not lend themselves to this treatment even given the variety of possibilities for retroversion, one cannot help wondering whether the choice passages are themselves simply happy coincidences.
But Casey says:That is not unreasonable, but it does involve the selection of benayya, which might well have been translated huious, rather than ynqyn, which was bound to be rendered tekna. (p. 13)Casey's point, and it is well taken, is that you can't just translate the Greek backwards into Aramaic to find wordplays and such; you have to imagine how a translator would most likely have rendered any putative Aramaic original. He thinks that benayya would most likely have been rendered "sons," not "children"; yanqin is the Aramaic word most likely to have been rendered "children."
One bit of data I like to add in to this kind of discussion is the importance of paying careful attention to ways in which the Greek of the passage may in fact show evidence of the redactional tendencies of the Gospel in which it appears. In the discussion of Matt. 3.9 above, it is worth noting that Matthew does in fact have a rather odd use of the term tekna elsewhere in a context where he is clearly discussing sons, 21.28-32, the Parable of the Two "Sons" (always called that, and never "Children"). I'm not sure how this helps the Aramaic retroversion discussions, but I am inclined to think that if Matthew is capable of idiosyncracies in his terminology, e.g. using tekna in contexts where we might have expected huioi, we should be very careful of talking about how a word "would have" been translated, or even how a word is "most likely" to have been translated.
One more point: the case for Matthean composition of the passage Matt. 3.7-10 is I think pretty strong; see, e.g., Michael Goulder's Midrash and Lection in Matthew and Luke: A New Paradigm. What Matthew has done is to compose a short speech for John the Baptist in his own style (even using phrases he will repeat in Jesus' mouth later on, like "Brood of vipers!" plus rhetorical question), around the theme suggested by the idea inherited from Mark that John's baptism was one of repentance.
PS: it is worth noting that Maurice Wile's Aramaic Sources of Mark's Gospel is on Google Print, and also An Aramaic Approach to Q.