In the section of his book entitled "Shem-Tob, Mark, and Luke" (p. 199,
1995 ed.), George Howard does comment upon the unusual Lukan textual
character of the Hebrew Gospel of Matthew. He notes that there are many
unusual agreements between HMt and Luke (as well as Mark) in those
passages where Hebrew Matthew diverges from the standard Greek Matthew.
And then Howard says,
should be noted
that Shem-Tob's harmonistic readings with Mark and Luke are not evenly
distributed. He agrees with Luke much more often than with Mark in the
triple tradition. Second, and perhaps even more significant, he agrees
with the Q sections of Luke more than two times as often as he does
with the non-Q sections." (G. Howard, THE HEBREW GOSPEL OF MATTHEW,
1995, p. 199)
In regard to Howard's first observation, i.e. that HMt "agrees with
Luke much more often than with Mark in the triple tradition", there's
no question that he is correct there. But it is his second claim, i.e.
that HMt "agrees with the Q sections of Luke more than two times as
often as he does with the non-Q sections", that I find highly
Because I have been making such comparisons myself recently, and the
truth of the matter seems to be quite the opposite...
In actual fact, it is in the non-Q sections,
i.e. in the triple
tradition passages (where Mt, Mk, and Lk all have text in common), that
the agreements between HMt and Lk are especially concentrated. There
are of course also quite a few agreements in the Q sections, i.e. in
the double tradition (the common passages between Mt and Lk only), but
these are far less frequent.
Based on my own comparisons, the Q sections agreements generally seem
to be running at the rate of perhaps 1 per 10 verses of HMt.
In the triple tradition, on the other hand, we may find as many as 5
agreements with Lk per 10 verses of HMt -- at least... And in many
pericopes, such as the one I've already cited (Healing of a Blind
Beggar in HMt 20:29-34), there will be even more agreements with Lk! In
fact, in the triple tradition passages, it's not unusual to find such
agreements at the rate of one per every verse. In other words, this
would be about 10 times more often than in the sayings material!
I have no idea how Howard arrived at his estimate, as cited above --
i.e. that, in the sayings materials, HMt agrees with Lk twice as often
as in its narrative parts. But it seems like his main focus had been on
the hypothetical Q source, and how it may be reflected in the HMt...
Perhaps Howard was trying to appeal to the mainstream Q theorists with
all this, so that they might start paying some attention to the HMt. So
he formulated some quite interesting theories in this area but,
unfortunately, following the publication of his studies, the Q
theorists basically ignored all that. They were not interested...
Well, myself, I don't really believe in Q, so my own focus is very
UNIFORMITY OF THE SAYINGS MATERIAL
One basic consideration needs to be kept in mind in all this. The fact
of the matter is that, in general, there's a lot more agreement between
Mt and Lk in the double tradition than in the triple tradition. Most
Synoptic scholars are well aware of the fact that the sayings material
(i.e. the double tradition) is fairly homogeneous, relatively speaking
-- as compared to the narrative material of the gospels.
Indeed, there are long passages where the sayings material in the Greek
Mt is running almost word-for-word identical to Lk. For example, in
sections like these,
Jesus' Thanksgiving to the Father
Lk 10:21-22/Mt 11:25-27
God's Answering of Prayer
Lk 11:9-13/Mt 7:7-11
Return of the Evil Spirit
Lk 11:24-26/Mt 12:43-45
Lk 12:22-32/Mt 6:25-34.
So, quite obviously, since Mt and Lk are so close in these sections, it
would be quite impossible for HMt to have any sort of a special
affinity to Lk in these sections...
In contrast, it's in those passages where Mt and Lk do differ a lot
that the Hebrew Mt is likely to demonstrate its Lukan character most of
In general, the Q theorists tend to see those sections in the double
tradition where Mt agrees very closely with Lk as the 'bedrock Q', i.e.
very early material. But I think it is a lot more realistic to see such
sections as a rather late material; in other words, these close
agreements seem to be the result of late editing and polishing.
So, even if considered from this perspective alone, Howard's estimate
that in the sayings materials HMt agrees with Lk twice as often as in
its narrative parts should already be seen as rather suspect.
The evidence for what I'm saying is all over the place in the Hebrew
Matthew. One can open up just about any triple tradition section and
verify these things easily.
For example, we can consider these two consecutive sections, "Stilling
the Storm" (HMt 8:23-27), and "The Gadarene Demoniacs" (HMt 8:28-34).
The rate of agreement between HMt and Lk in these sections is running
at about one per every verse. 12 verses -- 12 agreements (and there
seem to be even some more in the Gadarene pericope).
OF THIS TEXTUAL PHENOMENON IN HEBREW MATTHEW
This apparent error in Howard's analysis that I've found is certainly
quite significant. Indeed, it has now been 18 years since Howard
published this new text, and yet, essentially, it still remains a black
sheep of New Testament scholarship. Other than some pat dismissals from
certain quarters, nobody is doing any serious work on this new text.
Nobody quite knows what to do with it, and how to approach it...
Nobody, including Howard himself, has offered any coherent theory as to
how this Hebrew text came about, and where its true place within the
history of NT gospel tradition lies.
The only thing that George Howard, himself, has ever tried to prove is
that this text contains much early material, and goes back to ancient
times... And, while I believe he for the most part succeeded in this
task, this of course still leaves all sorts of important questions
unaddressed and unanswered.
It always appeared to me that, indeed, Howard's Hebrew Matthew
contained all sorts of original material... I constantly found very
unusual early readings there all over the place -- the readings that
were very often supported in the ancient Aramaic manuscripts of Matthew
-- but, still, any coherent theory as to how this text originated had
always been rather elusive.
Well, perhaps we can answer this question now; Howard's Hebrew Gospel
of Matthew is in fact nothing other than the crucial missing link that
solves the Synoptic Problem handily, all by itself... In effect, this
Hebrew text is the best proof that Luke was the earliest gospel, from
which our canonical Matthew eventually derived. Because this Hebrew
text is demonstrably an all-important middle term along that trajectory between Luke and Matthew!
There's simply no other logical and
economical way to explain all these Lukan features that are all there
in the Hebrew Matthew.