Matthew is the most obviously Jewish of the four Gospels, so a Hebrew or Aramaic original can sound quite plausible. But the evidence does not really go there. The claim depends heavily on one ambiguous sentence from Papias. The Gospel we actually have reads as Greek, uses Greek Mark, draws heavily on Greek Scripture, and is supported by an early Greek manuscript tradition. Matthew is profoundly Jewish. That does not mean our Gospel of Matthew was translated from Hebrew.

Matthew feels Jewish from the opening line. It begins with Abraham and David. It is saturated with Scripture. Jesus speaks about the law and the prophets, righteousness, the kingdom, judgement and Israel’s hope. So the idea that Matthew first wrote in Hebrew or Aramaic has an obvious appeal. Surely this most Jewish of Gospels began in Hebrew or Aramaic rather than Greek?

But sounding plausible is not the same as being supported by the evidence. Once we ask what the surviving Gospel actually looks like – what language it reads like, what sources it uses and what manuscript trail it has – the Hebrew-original theory creates more problems than it solves.

Let’s be clear about what is and isn’t being argued. Jesus and his earliest disciples lived in a Semitic-speaking Jewish world. Some sayings of Jesus may well have circulated in Aramaic or Hebrew before they were written in Greek. Matthew is deeply Jewish in thought, argument and scriptural background. None of that is controversial. The much larger claim is that the canonical Greek Matthew in our Bibles is a translation of an earlier Hebrew Gospel. That claim is much harder to sustain.

The ancient peg: Papias

The whole theory really hangs off one sentence from Papias, bishop of Hierapolis, writing in the early second century. His work is lost. We know the sentence because Eusebius quoted it in the fourth century – about two centuries later:

“Matthew wrote the oracles in the Hebrew language, and every one interpreted them as he was able.”[1]

This is the one indispensable text for the Hebrew Matthew theory.  Without it, the theory largely collapses.  And it is worth noting that while Papias is broadly agreed to be very early we are relying on second hand testimony centuries later. 

First, Papias does not use an unambiguous word for a complete Gospel. The Greek term logia can mean “oracles”, “sayings”, sacred utterances, or perhaps a collection of traditions. It may refer to a Gospel, but it does not have to mean “the Gospel of Matthew exactly as we now have it”. Hagner’s judgement is fair: “nearly every element” in the sentence can be understood in more than one way.[2]

 Second, the phrase usually translated “in the Hebrew language” may refer to Hebrew, Aramaic, or, on one reading, to a Jewish or Semitic manner of arrangement. The verb translated “interpreted” can mean translated, explained or expounded. So the key witness is early, but it is not clear.

Third, even if Papias meant that Matthew wrote something in Hebrew or Aramaic, it does not follow that he meant our full canonical Gospel. He may have had in mind a collection of sayings, fulfilment texts, or some other Semitic material later associated with Matthew. Hagner thinks a sayings collection is more probable than Papias describing our canonical Gospel.[3]  

The International Standard Bible Encyclopedia draws attention to the two problems which keep returning in this discussion: our Greek Matthew does not look like a translation, and its use of Greek Mark is very difficult to fit with the traditional reading of Papias.[4]

So Papias may preserve a real memory that Matthew, or a group associated with him, collected Jesus traditions in a Semitic language. That is quite possible. But Papias does not give us a clear publication history proving that the Gospel in our Bibles was translated from Hebrew.

The missing manuscript problem – where is the Hebrew Matthew?

Now for the very basic problem: we do not possess an early Hebrew Matthew. The early manuscript tradition is Greek. The papyri are Greek. The great uncials are Greek. The later minuscules, patristic quotations, lectionaries and early translations all point back to a Greek Gospel. Turner and Bock list second- and third-century Greek fragments of Matthew, followed by abundant Greek manuscript evidence thereafter.[5]

Absence of evidence is not always evidence of absence. Fair enough. But here the absence has real weight. If an apostle had written the authoritative original of Matthew in Hebrew, we would expect some trace of it: a fragment, a quotation clearly identified as coming from it, a known textual tradition, somebody comparing it with the Greek, or at least an early Christian able to say they had seen it. Instead the supposed original vanishes at exactly the point where it should have mattered most. Lenski presses the point well. If Matthew, one of the Twelve, had written a full Gospel in Hebrew, the church would have treasured it. Producing a Greek translation would not suddenly make the apostolic original worthless. It would still matter for authority, memory and checking the translation. Yet the early fathers cannot name anyone who had actually seen or handled this alleged Hebrew Matthew.[6]

This does not rule out every possible Semitic source behind Matthew. There may have been sayings, notes, testimonia or oral traditions in Hebrew or Aramaic. But it makes the larger claim (that our canonical Matthew is a Greek translation of an apostolic Hebrew Gospel) historically very strained.

Greek Matthew reads like a Greek composition

Then there is the Greek itself. The objection is not merely, “We have not found the Hebrew manuscript.” Scholars also point out that the text we do have does not behave like a translation from Hebrew or Aramaic.

Hagner states the basic problem plainly: Greek Matthew shows no signs of being translated from Aramaic and shows abundant evidence of dependence on the Greek of Mark.[7]  Guenther is even blunter: that Matthew’s language is not translation Greek is “too widely documented to require further discussion.”[8]  Luz likewise argues that the evangelist’s good Greek and familiarity with the Septuagint suggest that Greek was probably his mother tongue.[9]

Now that needs a qualification. Matthew certainly has a Semitic flavour. It is full of Jewish themes, Semitic turns of phrase and some Aramaic expressions. But that does not prove translation. Jewish Greek could sound Jewish. The Septuagint had already given Greek-speaking Jews a language for Scripture, worship and theological argument. A Jewish-Christian author could write directly in Greek while thinking entirely within Israel’s scriptural world.

This is where some arguments for Hebrew Matthew push the evidence too far. Words such as raca, Beelzebul, corban and Bar-Jonah show contact with the Semitic world of Jesus. They do not prove the whole Gospel was translated from Hebrew. Mark also preserves Aramaic expressions. John preserves Semitic names and then explains some of them. As the Dictionary of Jesus and the Gospels notes:

As a generalization, the Gospels are written in a simpler style and contain some expressions that stretch the limits of what might be considered “proper” Greek. Some of these features of style could be ascribed to the Semitic background of the NT. In this line of thought, the native Aramaic language of NT authors colored their Greek style, so that many of the grammatical constructions and word usages that seem unusual as Greek are simply Aramaic (or Hebrew) idioms coming literally into Greek. ^[10]

That is the point. Semitic colouring is what we would expect from writings produced in the Jewish world of the first century. No one therefore concludes that all four Gospels must be translations of Hebrew originals.

Some details also fit naturally with Matthew editing Greek Mark. Mark 7:11 preserves the Semitic term corban and then explains it as “a gift”. Matthew 15:5 dispenses with Mark’s Semitic term and simply gives the Greek equivalent, “gift”. That does not prove Matthew composed in Greek, but it is entirely natural if he was rewriting Greek Mark for his own readers. It provides no positive evidence that Matthew was translating a Hebrew Gospel.

Matthew’s use of Mark creates a major logical problem

For me, this is the biggest logical problem for a Hebrew original. Most scholars regard Mark as the earliest Gospel and conclude that Matthew used Mark as a source. That is not just because Mark is shorter. Matthew and Luke reproduce large amounts of Mark, often with close agreement in the Greek, while also smoothing Mark’s grammar, abbreviating or expanding stories, and rearranging material in ways which make sense as deliberate editing.

The NIV Cultural Backgrounds Study Bible summarises the point: about ninety per cent of Mark appears in some form in Matthew, and Mark’s rougher grammar makes it much more likely that Matthew improved Mark than that Mark made Matthew rougher.[11]

The relationship between the gospels can be represented as follows (source: Two-gospel hypothesis – Wikipedia)

Goodacre’s argument from “editorial fatigue” makes the relationship even more interesting. In some stories Matthew changes Mark’s setting at the start but later slips back into wording or assumptions which only make sense in Mark. Matthew’s healing of the leper, for example, has crowds present, yet Jesus still commands secrecy; the command makes better narrative sense in Mark’s more private scene. In another case Matthew has Jesus’ mother and brothers standing “outside”, although Matthew has not put Jesus inside a house. Mark had, which explains the detail.[12]

This does not prove every part of Markan priority beyond argument. But now try to fit the evidence into a Hebrew-original theory. Matthew first writes a Hebrew Gospel. An unknown person later translates it into Greek. While translating, that person also has Greek Mark open, repeatedly adopts Mark’s wording and narrative details, and edits Mark as part of the process. The finished work still circulates as Matthew’s Gospel, while the apostolic Hebrew original disappears.

Could that happen? Yes. Almost anything can be made possible if enough extra steps are added. But it is not the natural reading of the evidence. At that point those claiming a Hebrew original are not following the evidence; they are stacking assumptions to protect their original claim.

Matthew’s Bible is Greek Scripture

Matthew’s use of the Old Testament pushes in the same direction. This needs some care. Matthew does not quote one simple, uniform Septuagint text. His fulfilment quotations are complex. Some are closer to the Hebrew, some to Greek, and some may reflect revised Greek texts or Matthew’s own rendering.

But that complexity does not help the Hebrew-original theory.  Menken’s study of Matthew’s Old Testament text argues that Matthew’s fulfilment quotations are best explained as a revised Septuagintal text, and that “Matthew’s Bible” was, on the whole, a revised Greek Scripture text.[13]  Keener and Walton likewise observe that where Matthew overlaps with Mark, his quotations follow the standard Greek translation as Mark’s do.[14]

That does not by itself prove Greek composition. But if our Gospel were a fairly direct translation of a Hebrew original, we might reasonably expect a more consistent relationship with the Hebrew text than we actually find. Instead, Matthew repeatedly works with Greek textual forms of Scripture. To preserve the Hebrew-original theory, the proposed translator must therefore do more than translate: he must also substantially recast the scriptural quotations.

The later Hebrew Matthews do not rescue the theory

Some modern arguments move from Papias to an actual surviving Hebrew text: Shem-Tob’s Hebrew Matthew, embedded in the fourteenth-century Jewish polemical work Even Bohan. George Howard argued that it preserved an old Hebrew substratum which might somehow be related to a primitive Hebrew Matthew.[15]

This claim has very little scholarly support.  McKnight notes that Howard’s conclusions have received little acceptance even among some who are sympathetic to the possibility of a Hebrew Gospel tradition.[16]  Petersen’s study is particularly damaging. He argues that Shem-Tob belongs to the western harmonised Gospel tradition, is related to the tradition behind the Middle Dutch Liège Harmony, and was translated from a medieval Latin Vorlage.[17]

The date and setting matter. A text is not early simply because it is written in Hebrew. Shem-Tob is more than a thousand years removed from the apostles and appears inside a medieval Jewish anti-Christian polemic. It has to demonstrate an ancient textual ancestry; we cannot simply assume one. On Petersen’s analysis it looks medieval, harmonised and secondary, not like the recovery of an apostolic original. Howard himself notes that Shem-Tob’s Matthew differs theologically from Greek Matthew in striking ways. It changes titles for Jesus, presents John the Baptist differently, and lacks the Gentile-oriented force of Matthew 28:19–20.[18]  Some readers are attracted to those very differences, especially when a reading appears to weaken a later doctrinal formula. But that is precisely where we need to be careful. A late medieval variant cannot be placed above the overwhelmingly attested Greek text because we happen to prefer its theology.

This is important for Christadelphians. We do not need a speculative Hebrew Matthew to challenge later Trinitarian theology. Our doctrine should rest on a careful reading of the biblical text we actually have, not on an unstable reconstruction of a lost Gospel.

What about the patristic chorus?

Then comes the claim that the early fathers are almost unanimous. It sounds impressive. But it may be far less independent evidence than it first appears. Much of the later testimony may simply repeat or expand Papias.

There is another complication. Later Christian writers knew of Jewish-Christian works called the Gospel according to the Hebrews, the Gospel of the Nazarenes and the Gospel of the Ebionites. Those works are confusing to us, and the distinctions were not always clear to ancient writers either. It would have been easy to connect one of them with Papias’ statement about Matthew’s logia. Hagner notes that the existence of these Hebrew or Aramaic Jewish-Christian gospels may have encouraged later fathers to identify Papias’ statement with canonical Matthew.[19]  The older Dictionary of the Bible makes the same point: Jerome at first identified a Hebrew gospel he examined in Syria with Matthew’s work, but later withdrew that conclusion.[20]

This does not make the patristic testimony useless. It simply means we should not flatten several distinct pieces of evidence into one supposedly unanimous tradition. The fathers may preserve memories of Semitic gospel materials and of Jewish-Christian gospels. That is not the same thing as independent evidence that canonical Matthew was translated from Hebrew.

The strongest pro-Hebrew argument

To be fair, the strongest scholarly case is not the internet version of “Papias said it, so the matter is settled.” James R. Edwards argues that a Hebrew Gospel tradition influenced the development of the Synoptic Gospels, particularly material unique to Luke. He draws on patristic references, Semitisms in Luke, and the possibility that a Hebrew Gospel functioned as a genuine source in early Gospel formation.[21]

This is a serious argument and deserves a fair hearing. It is also not really the simple claim now being made in some circles. Edwards does not establish that canonical Matthew is a Greek translation of a Hebrew Matthew, and he cannot point to an extant copy of the Hebrew Gospel he proposes. Reviewers have naturally highlighted that problem.[22]  Goodacre goes further, arguing that Edwards overstates the Lukan links, underplays Matthean and Markan material, and does not sufficiently allow for the possibility that some “Hebrew Gospel” fragments come from later Gospel harmonisation rather than a primitive source.[23]

At most, the strongest version of the case may show that Semitic traditions and Jewish-Christian gospels were part of the complicated early history of Jesus tradition. I have no problem with that. But it is a much smaller claim than saying the canonical Gospel of Matthew in our Bibles was originally written in Hebrew.

A cleaner explanation

For me, the cleaner explanation is:

• Jesus taught in a Semitic-speaking Jewish environment.
• Some sayings, scriptural testimonia and oral traditions about Jesus may have circulated in Hebrew or Aramaic.
• Matthew is a Greek composition shaped throughout by Jewish Scripture, Jewish argument and Semitic tradition.
• The author used Greek Mark and Greek forms of the Old Testament.
• Papias probably preserves a compressed, confused or partial memory of Semitic material associated with Matthew, not a reliable publication history of our canonical Gospel.
• Later Hebrew Matthews, especially Shem-Tob, are medieval and secondary rather than surviving apostolic originals.

This accounts for why Matthew feels so Jewish. It accounts for Papias without pretending his sentence is clearer than it is. It accounts for the Greek manuscripts, Matthew’s close use of Mark, the Septuagintal texture of the Gospel, and the later confusion with Jewish-Christian gospels.

The Hebrew-original theory has to keep adding qualifications: perhaps Matthew wrote in Hebrew; perhaps an unknown person translated it; perhaps the translator used Greek Mark throughout; perhaps the translator also revised the Old Testament quotations towards Greek forms; perhaps the apostolic original then disappeared completely; and perhaps medieval Hebrew texts still preserve pieces of it. At some point the growing stack of qualifications becomes part of the evidence against the claim.

Conclusion

Matthew is deeply Jewish. That is not in dispute. There may even have been Semitic material associated with Matthew, sayings, testimonia, notes or oral traditions. But the canonical Gospel does not look like a translation from Hebrew or Aramaic. It looks like a Greek Gospel written by someone immersed in Israel’s Scriptures, using Greek sources, especially Mark, and writing for people able to hear Israel’s story and arguments in Greek.

So I think the claim needs to be stated much more carefully. Not: “Matthew was originally written in Hebrew.” Rather: “Matthew is a Greek Gospel with a profoundly Jewish and Semitic background.”

That retains what is true in the tradition. It just stops the tradition being pushed further than the evidence will take it.

by Daniel Edgecombe

---

[1] Eusebius, Ecclesiastical History 3.39.16, trans. Arthur Cushman McGiffert, in Nicene and Post-Nicene Fathers, Second Series, vol. 1, ed. Philip Schaff and Henry Wace. The online New Advent text preserves the Papias statement and its immediate context.

[2] Donald A. Hagner, Matthew 1–13, Word Biblical Commentary 33A (Dallas: Word, 1993), xliv.

[3] Hagner, Matthew 1–13, xlv–xlvi

[4] D. A. Hagner, “Matthew, Gospel according to,” in The International Standard Bible Encyclopedia, rev. ed., ed. Geoffrey W. Bromiley (Grand Rapids: Eerdmans, 1979–1988), 281–282

[5] David Turner and Darrell L. Bock, Cornerstone Biblical Commentary, Vol. 11: Matthew and Mark (Carol Stream, IL: Tyndale House, 2005), 6–7

[6] R. C. H. Lenski, The Interpretation of St. Matthew’s Gospel (Minneapolis: Augsburg, 1961), 12–18

[7] Hagner, Matthew 1–13, xlv

[8] Heinz O. Guenther, “The Sayings Gospel Q and the Quest for Aramaic Sources: Rethinking Christian Origins,” Semeia 55 (1991): 41–76, quote at 49

[9] Ulrich Luz, Matthew 1–7: A Commentary, rev. ed., Hermeneia (Minneapolis: Fortress, 2007), 47

^[10] M. Graves, “Languages of Palestine,” in Dictionary of Jesus and the Gospels, Second Edition, ed. Joel B. Green, Jeannine K. Brown, and Nicholas Perrin (Downers Grove, IL: IVP Academic, 2013), 490.

[11] Craig S. Keener and John H. Walton, eds., NIV Cultural Backgrounds Study Bible (Grand Rapids: Zondervan, 2016), 1682

[12] Mark Goodacre, “Fatigue in the Synoptics,” New Testament Studies 44 (1998): 45–58

[13] M. J. J. Menken, Matthew’s Bible: The Old Testament Text of the Evangelist (Leuven: Leuven University Press / Peeters, 2004)

[14] Keener and Walton, NIV Cultural Backgrounds Study Bible, 1682

[15] George Howard, “The Textual Nature of Shem-Tob’s Hebrew Matthew,” Journal of Biblical Literature 108 (1989): 239–240

[16] Chris McKnight, “Matthew, Gospel Of, Hebrew Version of,” in The Lexham Bible Dictionary, ed. John D. Barry et al. (Bellingham, WA: Lexham, 2016)

[17] William L. Petersen, “The Vorlage of Shem-Tob’s ‘Hebrew Matthew’,” New Testament Studies 44, no. 4 (1998): 490–512

[18] George Howard, “Matthew, Hebrew Version of,” in The Anchor Yale Bible Dictionary, ed. David Noel Freedman (New York: Doubleday, 1992), 642–643

[19] Hagner, Matthew 1–13, xlv

[20] James Hastings et al., Dictionary of the Bible (New York: Charles Scribner’s Sons, 1909), 592

[21] James R. Edwards, The Hebrew Gospel and the Development of the Synoptic Tradition (Grand Rapids: Eerdmans, 2009)

[22] Michael J. Thate, review of James R. Edwards, The Hebrew Gospel and the Development of the Synoptic Tradition, Themelios 36, no. 2 (2011)

[23] Mark Goodacre, review of James R. Edwards, The Hebrew Gospel and the Development of the Synoptic Tradition, published in abbreviated form in Catholic Biblical Quarterly 73, no. 4 (2011): 862–863