Of course, the same passage can contain at the same time several Semitisms mixed together, the conclusive force of which becomes more pronounced. Let us take a single example:
Matthew 23:25 . . . and so you will draw down on yourselves the blood of every just man (= justs) that has been shed on earth from the blood of Abel the just to the blood of Zachariah, son of Barachiah, whom you murdered between the Temple and the Altar.
Luke 11:50-51 … in order that there be demanded the blood of all the prophets poured out since the creation of the world to this generation, since the blood of Abel up to the blood of Zachariah, who had been murdered between the Altar and the House.
This passage contains two Semitisms of vocabulary (both of them in Luke): to demand the blood of someone means to hold him responsible for a murder and House refers to Temple. This passage contains (also in Luke) a Semitism of transmission: NQY: innocent, just (or, in the plural, NQYYM, as in Jeremiah 19:4) has been confused in Luke with NBY ’ (prophet), which can be written NBY in the spelling of Qumran. This necessitated the altering of the sentence to obtain: the blood of all the prophets. But why compare, in this fashion, the murder of this Zachariah, committed around the year 790 before Jesus Christ, with the murder of Abel committed at the very beginning of the world? One is actually at the beginning of a series, the other is far from being at the end of a series! It is because the murder of this Zachariah in the precincts of the Temple is reported toward the end of the second book of Chronicles, which is the final book of the Hebrew Bible but which is not the last book in the Greek Bible of the Septuagint. In Hebrew this means from the first page of the Bible to the last, but in Greek this no longer signifies anything (and actually for a long time commentators had no understanding of it). Behold, moreover, a Semitism of composition, at one and the same time in Matthew and in Luke! Who would dare to say that this passage had been composed in Greek by Matthew, or Luke, or anyone?
Carmignac, Jean. 1987. The Birth of the Synoptic Gospels. Translated by Michael J. Wrenn. Chicago, Ill: Franciscan Pr. p. 39
Conclusion
For Jean Carmignac, the evidence that the original language of the Gospels of Mark and Matthew and documents used by Luke was Hebrew makes his hypothesis secure.
In order to contradict this conclusion, it would be necessary to provide satisfactory explanations, valid in Greek, for all the cases that have been mentioned. (p. 40)
Arguing a fortiori Carmignac presented only the evidence relating to three of the nine types of Semitisms.
Only the Semitisms of the final three categories (composition, transmission, translation) have been retained in order that no objection could be made regarding either the mother tongue of the authors or their desire to imitate the Septuagint. But even in the first five categories (borrowing, imitation, thought, vocabulary, syntax) and especially in the sixth (style), the abundance of evidence presented goes far beyond any possibility that the author was influenced by his mother tongue or by the prestige of a venerable text. For the Greek of our Gospels testifies to a good knowledge of the language: nouns are correctly declined, the verbs accurately conjugated, and the vocabulary is relatively rich. Our Greek Gospels were not written by semiilliterates; they were written by people who possessed a good solid Greek culture, but who did not express themselves with the independence of a redactor, and who believed themselves obliged to render these precious documents in the most slavish fashion possible. Our Synoptic Gospels are not compositions which were realized in Greek; they are translations made upon the Hebrew (except for the Prologue and the introductions of Luke). And therefore the real authors of Mark and Matthew are their Hebrew redactors. For Luke, the situation is less clear, for we do not know if he himself was the translator or if he relied on the competence of some bilingual collaborator; we cannot be specific about which revisions he inserted into the documents which he found before him. But, in general, these reworkings must have been superficial, as the numerous Semitisms which still exist bear witness. (p. 40)
Carmignac acknowledges that his arguments as set out so briefly for a wide audience will not be enough to persuade specialist scholars. They will want more in-depth technical discussions. No doubt. But till then surely his work makes it difficult to ignore a real possibility of a Hebrew background to the gospels.