The Carmignac Challenge, Part II of III: Semitisms in the Gospel of Mark and the Gospel of Matthew

​Summary

In the Abbé Jean Carmignac’s short book, The Birth of the Synoptic Gospels (Chicago: Franciscan Herald Press, 1983), he identifies and classifies Semitisms in the synoptic gospels. He evaluates the value of the listed Semitisms as supporting evidence for his theory that the synoptic gospels were originally written in a Semitic language (GMark and GMatthew) or were translated from Semitic documents (most of GLuke). He acknowledges that their value as evidence will vary. He is also interested in the Semitisms for their own sake, as they “contribute to a better understanding of the word of God.” (p. 21) Here, I review only those Semitisms in the Gospel of Mark and the Gospel of Matthew. This list is exhaustive as a reader service, because the book is rare and expensive.

Initially, I intended to only assess whether the listed Semitisms supported Carmignac’s theory of a Semitic-language GMark. I found that the Semitisms that Carmignac identifies in the Gospel of Mark are all explainable by one of the following: (1) Mark incorporated into his narrative (and his play) literal Greek translations of Hebrew-language sayings, parables, etc., from documents in Mark’s library (Sayings of Jesus or similar). This is not evidence of a Semitic-language GMark. (2) Mark wrote his narrative in a literary style that emulated the Septuagint (LXX). That raised the question, did Mark’s emulation include a Hebrew version?

In Part I, I explained why I think that the answer is yes. Item (2), Mark’s emulation of the LXX, was a multi-step process that included an intermediate step: an idiomatic translation of Mark’s first Greek draft into Hebrew. The resulting Hebrew document was discarded after it was re-translated, literally, into Greek.

My review of the list of Semitisms yielded another observation of value to scholarship. Matthew used texts that were written with Qumran Hebrew letter forms. That fact must be incorporated into the history of second-century Christianity. I discuss this in Part III.

​The nine types of Semitisms

Carmignac identifies nine types of Semitisms in the synoptic gospels. I provide most of the detailed Semitisms that involve the Gospel of Mark, and my comments regarding their value as support for Carmignac’s theory. It turns out that only a few instances (mainly #7, Semitisms of Composition) support Carmignac’s theory. However, these instances are enough to convince me that GMark did have a Hebrew version.

​1 Semitisms of Borrowing

These are words borrowed from Semitic languages, like amen and pasche. Their presence does not indicate that the text in which they are embedded was composed in a Semitic language. (Comment: I agree.)

​2 Semitisms of Imitation

These include constructions improper in Greek that reproduce a turn of phrase from the LXX. Their presence does not indicate that the text in which they are embedded was composed in a Semitic language. (Comment: I agree.)

​3 Semitisms of Thought

These include the absence of subordinate phrases which are necessary for the periodic sentences that characterize the Greek of the time. Instead, the synoptics use the Semitic amplitude, e.g., “he took the floor and he spoke.” (Comment: I think it would take a great effort to consistently write these phrases in a Greek text. They are not accidental byproducts of growing up in an Aramaic-speaking environment. I think these Semitisms are evidence of translation from a Semitic language.)

​4 Semitisms of Vocabulary

Carmignac gives an example of a Semitism of vocabulary: Hebrew/Aramaic uses “son” idiomatically. “Instead of saying citizen of the kingdom, invited to the banquet, condemned to Hell… one will say son of the kingdom (Mt 8:12; 13:38); son of the banquet (Mt 9:15; Mk 2:19; Lk 5:34); son of Gehenna (Mt 23:15)…sons of thunder (Mk 3:17).” Carmignac says that such Semitisms “are vestiges of the language which is most familiar to the author,” not proof he used it. (p. 23) (Comment: I don’t agree. As with #3 above, I can’t see a Greek speaker using such a construction naively, no matter how weak his Greek. I think these Semitisms are evidence of translation from a Semitic language.)

​5 Semitisms of Syntax

An example of a Semitism of syntax is the omission of a definite article in a Greek text. “These syntactical peculiarities generally do not belong in our translations…but they are, in fact, extremely numerous.” (p. 24) (Comment: It would be fair to say that many of these derive from Mark’s emulation of the LXX. The question is whether Mark could have emulated the LXX by thinking and writing in Greek only, or whether a Hebrew intermediary step was required. I addressed that in Part I.)

​6 Semitisms of Style

Carmignac provides examples of Hebrew poetic style in GMatthew and GLuke; unsurprisingly, these are mainly from poems and prayers. These Semitisms tell us nothing about the language used by the gospel authors. GMark contains two awkward phrases that are evidently translated from a Hebrew construction that begins wayyehȋ. (Comment: Yes, these are evidence of an underlying Hebrew text.)

​7 Semitisms of Composition

Semitisms of Composition are “cases in which the text would not exist in its present form if had not been composed in a Semitic language. Each of these cases tends to be a strict proof (unless one always chooses to invoke chance) that the original had been composed in a Semitic language and translated subsequently into Greek” (pp. 26-27).

1. Carmignac gives almost a full page to Mark 1:1-4. “The words bemidbâr (in the wilderness) and qôré’ crying or preaching are taken from Isaiah and applied to John the Baptist according to the process which is known as pesher, such as it was practiced at the time at Qumran (and elsewhere)….The citation from Isaiah only agrees with the account of Mark in Hebrew, but not in Greek in which its meaning disappears. If the author had written in Greek, he would not have employed it . This constitutes a Semitism of composition.” p. 27. (Comment: I concur. But the pesher could have originated in a Hebrew-language text, source (1) in the Summary above.)
2. “Another triple play on words exists in Mark 3:14-15. (Jesus) sends (the apostles) to preach and to have power to cast out devils: to send translates the word shâlah, to have power the verb shâlat, and to cast out the verb shâlak….These verbs are even placed in alphabetical order (het, tet…kaf). pp. 28-29. (Comment: I believe that 3:14-15 was written by an editor of GMark (see my book, The Two Gospels of Mark: Performance and Text . This editor conformed to the ‘emulation of the LXX’ style of GMark, that is, source (2) in the Summary above. So this Semitism tells us that Mark emulated the LXX; it does not tell us how he did so.)
3. Now to some plays on words. Example: “Why does Mark 1:13…specify that Jesus was with the wild beasts? Wouldn’t this be because of the excellent play on words wehâyâh im hahayyâh?” p. 29. (Comment: I believe that Mk 1:13 was written by an editor of GMark (see my book, The Two Gospels of Mark: Performance and Text.) This editor conformed to the ‘emulation of the LXX’ style of GMark, that is, source (2) in the Summary above. So this Semitism only tells us that Mark emulated the LXX; it does not tell us how he did so.)
4. “In Mark 2:6 the scribes are sitting and reflecting: yôshebîm whôshebîm.” p. 29. (Comment: I believe this phrase was written by an editor. Again an editor uses a Semitism because he knows that Mark emulated the LXX. We are again referred to how Mark did so.)
5. “In Mark 2:21 and Matthew 9:16, the tear (qéra’) does not become larger, as would be expected, and as the Vulgate actually translates it, but worse (ra’).” p. 29. Mk 2:21 in full is: “No one sews a piece of unshrunk cloth on an old garment; if he does, the patch tears away from it, the new from the old, and a worse tear is made.” (Comment: Mk 2:21 is a literal translation from a Hebrew-langauge source document, source (1) in the Summary above.)
6. “In Mark 3:10 those who seek to touch (nâga’) Jesus are, not the sick but those who had all types of afflictions (nèga’).” p. 29. (Comment: I think 3:10 is by an editor, conforming to Mark’s ‘emulation of the LXX’ literary style.)
7. “In Mark 4:6 and Matthew 13:6, mention of the sun is not that important (actually Luke 8:6 suppresses it) but sun (shèmèsh) is phonetically linked to the word root (shôrèsh).” p. 29. (Comment: This Semitism occurs in a parable that is translated from a Hebrew-language source document, source (1) in the Summary above.)
8. “In Mark 6:38 Jesus says to them (lâhèm) how much bread (lèḥèm) do you have (lâkèm)? Go…(lekou).” p. 29. (Comment: This line occurs during the First Feeding Miracle, a scene I believe was written by an editor. The editor conforms his writing to Mark’s ‘emulation of the LXX’ style.)
9. “In Mark 9:18 the possessed foams at the mouth (weyâraq) and grinds his teeth (wehâraq).” p. 29. (Comment: This phrase is a stage direction. Therefore, I think Mark wrote it. Unless the similarity is a coincidence, I can only explain its presence with Mark’s emulation process described in Part I: Mark writes in Greek, an assistant translates the Greek sense into Hebrew, then another assistant translates the Hebrew literally into Greek.)
10. “In Mark 10:34 Jesus predicts that he will be mocked (wesâḥaqû bo) and will be spat upon (weyâraqû bo).” p. 29. (Comment: This verse points to Isaiah 50:6, “I did not hide my face from mocking and spitting.” Probably Mark used the Greek version of that verse. No support here for Carmignac’s theory.)
11. “In Mark 11:15 and Matthew 21:12 the money changers tables are mentioned, because tables is signified by the word shûleḥânôt, and changers, sḥûlehânim.” p. 29. (Comment: Tables and money changers are necessary to the theatrical action of Mark’s play. I think the words’ resemblance is a coincidence. No support for Carmignac’s theory.)
12. “In Mark 13:8, Matthew 24:7 and Luke 21:11 Jesus predicts earthquakes (ra’âshîm) and famines (ra’âbîm).” pp. 29-30. Comment: This prediction is part of the Olivet Discourse, which I believe was heavily edited (see my book). An earthquake may be original, but “famines” are not. They cannot be dramatized—they refer to the world outside the world of the play. “Famines” (and possibly “earthquakes”) was added by an editor. Again we have an editor conforming to the ‘emulation of the LXX’ style of GMark.)
13. “In Mark 13:21: See here…see there corresponds to hinnéh hénnâh…hinnéh hénnâh.” p. 30 . (Comment: This is in another edited verse of the Olivet Discourse. Again, editorial addition.)
14. “In Mark 14:11, the high priests listened (wayyishme’û) and were delighted (wayyismeḥû).” p. 30. (Comment: I think this is an original stage direction that, like (I) above, supports Carmignac’s theory of a Hebrew GMark.)
15. “In Mark 14:16, the two disciples leave (wayyese’yû) … and find (wayyimṣe’yû) .” p. 30. (Comment: I think these words aren’t similar enough to be a Semitism.)
16. “In Mark 14:41 and Matthew 26:45, sleep and take your rest, according to the Hebrew translation of the New Testament done by Delitzsch, supposes nûmû wenûḥû.” p. 30. (Comment: This [apparent] Semitism is not a quotation from a source text. It can only be explained by a Semitic-language GMark. It is evidence for Carmignac’s theory.)
17. “In Mark 14:65 to blindfold (the face) and to hit correspond to the verbs sâtar and sâṭar, and perhaps to the very same causative form.” p. 30. (Comment: I am not sure this is original. An editor may have added the “blindfold (the face).” If it is original, it supports Carmignac’s theory.)

Summary of this section: there are several sources for Semitisms of Composition in GMark. Three can be explained only by a Semitic-language GMark (9, 14, 16, maybe 17) and are therefore evidence for Carmignac’s theory.

​8 Semitisms of Transmission

Now we switch gears away from Mark’s writing process. Now we talk about Carmignac’s second contribution to scholarship (in my opinion): the discovery of a Qumran-Hebrew-trained copyist used by Matthew.

Here, Carmignac is concerned with transmission, that is, differences between synoptic texts, mainly between the texts of GMark and GMatthew. He lists nine examples (I provide only six here). They can all be explained by the following process (which I think is reasonable given that Matthew is widely thought to have produced an Aramaic version): a) Matthew obtained a copy of (Greek) GMark; b) Matthew (or an assistant) translated GMark into Hebrew or Aramaic; c) the translation (let us say it was Aramaic) was written down by a secretary/copyist who used distinctive Qumran Hebrew letter forms (which Carmignac recognized); d) Matthew used this Aramaic text of GMark as the basis for Aramaic GMatthew. In the process, Matthew or a (non-Qumran-Hebrew-trained) assistant occasionally misread the first copyist’s letter forms. (Later, e) Matthew and his team translated Aramaic GMatthew into Greek.)

I note that when several people are involved, human error increases. Carmignac adds another explanation for these errors: the ancients didn’t have glasses or artificial light.

Note that here I do not address Carmignac’s claim that these Semitisms support the existence of Aramaic GMatthew. I accept it.

1. “In Mark 1:7 and Luke 3:16, John the Baptist says: I am not worthy to unfasten (lâshèlèt) the strap of his sandals, but according to Matthew 3:11 he says: I am not worthy to carry (lâs’ét) his sandals…” (p. 31).

2. “Matthew 13:32 and Luke 13:19 say that the grain of mustard becomes a tree, which is a definite exaggeration, for this plant hardly exceeds a meter and a half or two meters. Mark 4:32, on the contrary, says that it has branches that are so great that birds nest there. All would be explained if Mark, who seems to be inspired by Ezekiel 17:23, had rendered ‘NP (pronounce ‘anâph) branch, as in Ezekiel, and if a copyist prior to Matthew and to Luke had read ‘S (pronounce ‘ès) tree, since in the style of calligraphy in Qumran, the letters N and P, if they come together and touch one another, resemble the letter Ṣ” (p. 31). (Comment: At some later point, an editor of GLuke conforms it to GMatthew rather than to GMark.)

3. “In Mark 8:31 Jesus begins to teach (LHWRWT = lehôrôt) and in Matthew 16:21 he begins to show (LHR’WT = lehar’ôt). The two words are very easily confused with each other since, according to the style of calligraphy, in Qumran, the first could lose a W and the second its ’ so that each one would end up looking like LHRWT.” (p. 31)

4. “Mark 8:27 says that Jesus passed through the villages of Caesarea Philippi and Matthew 16:13 the regions of Caesarea Philippi. The general meaning is the same, but the passage from one word to the other could have been brought about by the resemblance between QRYWT = qiryôt: villages and QSWWT = qesawwôt: regions, since Y and W are written almost in the same way.” (p. 32)

5. “Mark 9:43 and Matthew 18:8 conclude a similar recommendation, one with the verb to go and the other with the verb to be thrown, that is to say by WHLKTH OR WHSLKTH, with only one letter difference.” (p. 32)

6. “Mark 11:14 speaks of eating of the fruit = YWKL (according to the spelling of Qumran) and Matthew 21:19 to produce fruit YWBL: as the letters B and K resemble each other so greatly, the possibility for confusion is very likely.” (p. 32)

​9 Semitisms of Translation

Carmignac writes, “In all languages the same word has several meanings…in English lime signifies chalk or a fruit or a tree. In these cases the Synoptic Gospels can present different translations, which are explained by a same original document.” p. 34. He gives examples from the three synoptic gospels. I here address only those that include GMark.

The Semitisms of Translation have a variety of sources and motivations. They provide little windows into the complex textual histories of the synoptic gospels. I cite the only one that I think is evidence for a Hebrew GMark. None reveal the Qumran-Hebrew-trained copyist described in #8 above.

1. “In Mark 5:13 the herd of pigs is estimated to number about two thousand. This is absolutely improbable, given the less than sociable nature of pigs and the scant vegetation of this area in Transjordan (today’s Golan Heights). Moreover, Matthew 8:32 and Luke 8:33 had great concern to suppress this detail. But, in Hebrew K’LPYM, which signifies about 2,000 if it is vocalized ke’alpayim, can also be vocalized ka’alapim and can mean by herds: the pigs went over the cliff by herds into the lake, but their number is no longer specified.” pp. 36-37. (Comment: I think that Mark wrote 5:13; it’s stage action, and not a quotation from a sayings document. But I don’t think Mark originally wrote “by herds.” Possibly the translator of Mark’s narrative into Hebrew took some literary license and translated the meaning of a word like “groups,” as ka’alapim. Then another translator mistranslated that word into Greek as “about 2,000.” Alternatively, Mark is referencing a now-lost text. In either case, this example does support Carmignac’s belief that GMark had a Hebrew version.)

​Other sources of Semitisms in the synoptic gospels

There are additional explanations of Semitisms in the synoptic gospels:

• Editors and copyists of the gospels may have referred to Hebrew or Aramaic texts, as more ‘historically accurate’ about Judean history than the LXX, or closer to the Word of God, and ‘corrected’ the gospels.
• Such corrections might also be artifacts of the ethnicity, education and native language of the copyists.
• Such corrections might also be the result of rabbinic-style exegesis of the gospels.

​Conclusions

Semitisms occur in the synoptic gospels for a number of reasons. A number of Semitisms are explained by translations (by Mark or before Mark, by his congregation) of Hebrew source documents (such as Sayings) into Greek. Other Semitisms are explained by Mark’s emulation of the LXX, or by editors of GMark who copied Mark’s literary style. The question is, whether Mark’s emulation involved a Hebrew intermediary step. Here, I have cited several Semitisms of Composition and a Semitism of Translation that I think can only be explained by the existence of a Hebrew GMark. The logical location for that Hebrew GMark was during the emulation process, after Mark had written his first draft of his narrative. Separately, the Semitisms of Transmission are evidence that a Qumran-Hebrew-trained copyist was involved in the creation of GMatthew. That will be addressed in Part III.