In his Harmony of the Gospels, Augustine spends a few paragraphs why Matthew would have referred a prophecy to Jeremiah that more properly ought to be attributed to Zechariah (cf. Matt 27:9). Like a good historical-critical NT scholar, he first considers a possible textual corruption:
If any one finds a difficulty in the circumstance that this passage is not found in the writings of the prophet Jeremiah, and thinks that damage is thus done to the veracity of the evangelist, let him first take notice of the fact that this ascription of the passage to Jeremiah is not contained in all the codices of the Gospels, and that some of them state simply that it was spoken by the prophet. It is possible, therefore, to affirm that those codices deserve rather to be followed which do not contain the name of Jeremiah. For these words were certainly spoken by a prophet, only that prophet was Zechariah. In this way the supposition is, that those codices are faulty which contain the name of Jeremiah, because they ought either to have given the name of Zechariah or to have mentioned no name at all, as is the case with a certain copy, merely stating that it was spoken by the prophet, saying, which prophet would assuredly be understood to be Zechariah (3.7.29).
Nonetheless, he quickly rejects this solution:
However, let others adopt this method of defence, if they are so minded. For my part, I am not satisfied with it; and the reason is, that a majority of codices contain the name of Jeremiah, and that those critics who have studied the Gospel with more than usual care in the Greek copies, report that they have found it stand so in the more ancient Greek exemplars. I look also to this further consideration, namely, that there was no reason why this name should have been added [subsequently to the true text], and a corruption thus created; whereas there was certainly an intelligible reason for erasing the name from so many of the codices. For venturesome inexperience might readily have done that, when perplexed with the problem presented by the fact that this passage could not be found in Jeremiah (Ibid.).
This is astounding. (To me, at least.) Augustine dismisses the text-critical argument for replacing “Jeremiah” by appeal to what we’d now call the canon lectio difficilior potior, which essentially recommends favoring the “more difficult reading” among two competitors. It’s more plausible, Augustine argues, that Matthew wrote the apparently erroneous “Jeremiah,” and that some scribes later corrected it, than to think that many, many scribes introduced an odd corruption into a perfectly intelligible text.
If that sounds commonsensical, stop and consider that this canon wasn’t formulated explicitly until the 18th century, probably by Johan Albrecht Bengel in 1725. It’s an essential tool of modern historical-critical research — but Augustine was using it in the 5th century! Maybe first century Romans had mechanical clocks after all…