I have always said that Matthew does not read to me like translation literature, but original composition. Somebody recently asked me for specific examples. This was my response:
Wow, that’s a really annoying question because it forces me to think about it. I’ve never really sat down and quantified it looking for the specific elements. It’s more a matter of feel based on familiarity with the language. It’s like when you get instructions on assembling a product that was made in Japan. It’s better these days, but often you would read through the instructions and have even less of a clue than before reading them, because although using English words it clearly wasn’t English… Let me take some time to come up with a specific example that illustrates the principles. A few elements to keep in mind, however:
1) The tendency for a translator is to render specific phrases from the original language always into the same phraseology in the target language consistently, whereas a native speaker tends to vary it. The same is often true of lexical stock.
2) The translator will tend to render certain constructions which reflect the way it’s said in the original language, rather than the best way to say it in the target language. That doesn’t mean that it’s wrong or unintelligible in the target language, only that it is unlikely that it’s precisely the way a native speaker would render it.
Now, this also tends to work with people who are not native speakers of the language they are communicating in, especially if they are less than truly fluent. I have edited papers for English style for Koreans and other speakers. Often their intent is clear, but oh man — that’s not the way a native speaker of English would say it! I recently read the entire Shepherd of Hermas in Greek for an article I was writing. Very easy Greek, but very tedious. Why? Because you can depend on the author to use the same vocabulary and constructions over and over again to communicate his points. It’s either translation literature or someone with limited competence in the language.
When I read Matthew, I don’t see much of this at all. I see variety of expression and lexical stock. I see a speaker familiar enough with his language of composition that he is comfortable saying things in “unexpected ways.” One test of this is to think of different ways that the communication could be rendered or paraphrased in the language of composition. Synoptic comparisons help here as well. Matthew doesn’t say things in the same way that Mark or Luke do…
A lot more could be said, particularly concerning the ancient views of translating Scripture as a sacred text vs. translating other types of literature (cf. Jerome’s letter on translating a Greek document into Latin vs. what he does with the biblical texts, http://www.newadvent.org/fathers/3001057.htm). But these are a few of the general reasons why Matthew does not read as translation Greek to me.