[Note: Readers are encouraged to explore the embedded links in this four-part essay, many of which lead to digital images of Mark 1:2 in manuscripts.]
Part 1: Reviewing the Greek Manuscript Evidence
Did Mark 1:2 originally say “in the prophets” (ἐν τοῖς προφήταις) or “in Isaiah the prophet” (ἐν τῷ Ἠσαίᾳ τῷ προφήτῃ)? As we embark on a multi-part exploration of this question, let’s thoroughly describe the external evidence, beginning with the manuscript-evidence for each rival variant:
● ἐν τοῖς προφήταις, according to the
apparatus, is supported by Codex
Alexandrinus (A, 02), Codex
Washingtoniensis (W, 032), f13, 28, 180, 579, 597,
1006, 1010, 1292, 1342, 1424,
1505, and Byz.
“Byz” represents not only hundreds of Greek Gospels-manuscripts that are less than 900 years old, but also the following manuscripts:
Codex Basiliensis (E, 07), Codex Boreelianus (F, 09), Codex Seidelianus II (H, 013), Codex Cyprius (K, 017), Codex Campianus (M, 021, which has “Isaiah” in the margin), Codex Guelferbytanus A (Pe, 024), Codex Vaticanus 354 (S, 028), Codex Nanianus (U, 030), Codex Mosquensis II (V, 031), Codex Monacensis (X, 033), Codex Macedonianus (Y, 034), Codex Petropolitanus (Π, 041), Codex Rossanensis (Σ, 042, from the 500’s), Codex Beratinus (Φ, 043), Codex Athous Dionysiou (Ω, 045), 047, 0133, and minuscules 24, 27, 29, 34, 67, 100, 106, 123, 134, 135, 144, 150, 161, 175, 259, 262, 274, 299, 300, 338, 344, 348, 364, 371, 376, 399, 405, 411, 420, 422, 478, 481, 564, 566 (paired with Λ, 039), 568, 652, 669, 771, 773, 785, 875, 942, 1055, 1073, 1076, 1077, 1078, 1079, 1080, 1110, 1120, 1172, 1187, 1203, 1223, 1225, 1266, 1281, 1346, 1347, 1357, 1379, 1392, 1422, 1426, 1444, 1458, 1507, 1662, 1663, 1701, 1816, 2142, 2172, 2193, 2290, 2324, 2368, 2369, 2370, 2373, 2414, 2474, 2509, 2545, 2722, 2790, 2800, 2811, 2812, 2854, 2907, and 2929.
● ἐν τῷ Ἠσαίᾳ τῷ προφήτῃ is supported by Oxyrhynchus Papyri LXXVI 5073, Codex Sinaiticus (01, Aleph (À), Codex Vaticanus (B, 03), Codex Regius (L, 019, “Isaiah” is spelled Ϊσαϊα), Codex Sangellensis (Δ, 037, Greek-Latin), 33, 151, 892, 1241, about 10 other minuscules, and the D’Hendecourt Scroll (from the 1300’s). (Minuscule 151’s retention of this reading may have something to do with the inclusion of Eusebius’ apologetical composition Ad Marinum in the same volume.)
● ἐν Ἠσαίᾳ τῷ προφήτῃ is supported by Codex Bezae (D, 05, Greek-Latin), Codex Koridethi (Θ, 038), the core members of f1, 565, and 205 (from the mid-1400’s), plus 700, 1243, and 1071. (Only these last three lack close affiliation with either the Western or Caesarean Text.) A few other minuscules support ἐν Ἠσαίᾳ τῷ προφήτῃ but were not listed in the
UBS apparatus; these include 22. Minuscule 22 shares some readings with f1
and 205, and also shares a note about the ending of Mark; in 22 the note
is shorter (failing to claim that the Eusebian Canons omit Mk. 16:9-20, very
probably because where and where 22 was made, the Canons had been adjusted to
include those verses) but it is recognizably the same note. (Minuscules 15, 22, 1110, 1192, and 1210 all have the note about Mark
16:9-20.) Also included: 61 (Codex
Montfortianus, on 55r; this manuscript is famous for its inclusion of the Comma Johanneum), 372 (assigned to the
1500’s, with some Latin notes in the margins), and 391 (produced in 1055).
● ἐν βίβλω λόγων Ἠσαΐου τοῦ προφήτου is read by 1273 (the George Grey Gospels) and 544 and a similar text is in the book in a full-page picture of Mark in Lectionary 1635.
For a convenient summary of versional and patristic evidence, see the STEPBible Textual Apparatus and Wieland Willker’s Textual Commentary on the Greek Gospels – Volume 2, Mark, 2015 Edition. Readers should be aware that 2427 (cited throughout Mark in the 27th edition of the Nestle-Aland Novum Testamentum Graece) has been proven to be a forgery, based on a printed text from the 1800’s, and that although f1 is cited for “in Isaiah the prophet,” this represents only a consensus of its core members.)
The Armenian version was listed in
UBS2 as support
for ἐν τοῖς προφήταις, but in UBS4 the Armenian
version was listed as support for ἐν Ἠσαίᾳ τῷ προφήτῃ. The
older Armenian manuscripts tend to not have “Son of God” in Mark 1:1, and to read “in Isaiah the prophet” in 1:2 – following the Caesarean form. However, a competition of influences upon the
Armenian tradition began very early in its history, in addition to later
influence from the Vulgate.
Most of the Greek lectionaries, such as Lect 123 (an illustrated lectionary from the 900’s), Lect 379 (from the 800’s), Lect 1599 (from the 900’s), Lect 71 (from 1066), Lect 183 (from the 800’s or 900’s), and the illustrated Lect 120 (from the 1100’s) support ἐν τοῖς προφήταις, but there are some exceptions, such as Lect 562 (from A.D. 991), which supports ἐν Ἠσαίᾳ τῷ προφήτῃ.
● One more variant seems to be attested by the Old Latin Codex Usserianus Primus (VL 14); the Byzantine and Alexandrian readings are combined, so as to read, “in Isaiah and in the prophets.” However the text is difficult to read due to damage to the parchment (See fol. 150r at the page-views at the Trinity College Dublin website .)
In this essay, the reading “in the prophets” will be defended as original, and I will argue that its Alexandrian and Western rivals originated in the following way:
In the 100’s, some copyists were mildly averse to non-specific references to Old Testament books, and added specific names in place of the original non-specific references. Mark 1:2 is one of the passages affected by this tendency toward specificity. Some copyists, understanding the paraphrastic opening phrase – which could be understood as a reference to Exodus 23:20 (in the Law, rather than the Prophets) – as merely an introduction to Isaiah’s words, adjusted the text so as to identify the prophet being cited.
This happened independently in Alexandrian and Western transmission-streams, which is why the Alexandrian witnesses consistently have τῷ before Ἠσαίᾳ, while the major Western and Caesarean witnesses do not. When (and where) copyists and commentators were confident that Mark was using Malachi rather than Exodus, Christian scholars whose manuscripts read “in Isaiah the prophet” developed inventive explanations about how Mark could appear to identify Malachi’s words as if they had been written by Isaiah. These explanations were sufficiently convincing to allow the reading to remain in the Alexandrian and Western transmission-lines.
The insertion of specific names, where the original text has no specific name, is a recurring scribal practice, and one which is observable in some of our very earliest New Testament manuscripts. For example, in the early Alexandrian transmission-stream, in Luke 16:19, in the story about the rich man and Lazarus, the rich man was given a name: he was named
. This reading is found in Luke 16:19 in Sahidic
copies, and in the manuscript known as Codex
Sinaiticus Arabicus, or Nineveh CSA, one of the
documents discovered at St. Catherine’s Monastery in 1975. (A collation of CSA
by Hikmat Kachouh was released in 2008 in the journal Novum Testamentum.) But there
is much earlier evidence for that reading.
75 reads named Neuhs in the same
passage (ονοματι νευης), and this is the same name, Nineveh,
disfigured by a parableptic error in which the copyist skipped the first
syllable. (Two Greek manuscripts,
minuscules 36 and 37, have margin-notes which also identify the rich man as Nineveh.)
When a character in the Gospels plays a prominent role, but has no name, frequently a name is provided. Bruce Metzger documented this phenomenon in his essay, Names for the Nameless in the New Testament, which serves as chapter 2 of New Testament Studies: Philological, Versional, and Patristic.
The scribal tendency to provide names for unnamed individuals comes into play repeatedly in passages where the text refers to the fulfillment of prophecies. The non-specific attribution “through the prophet” is often turned into a specific attribution. Usually the attribution is correct, but sometimes it is incorrect.
The Old Latin Codex Colbertinus (VL 6) displays this tendency. Its text of Mark 15:27 provides names for the two robbers who were crucified with Jesus – Zoathan and Chammatha. VL 6, like practically all Latin manuscripts of Mark, also reads “in Isaiah the prophet” in Mark 1:2. In Matthew – where Matthew quoted Isaiah without naming him (simply saying that “what was spoken by the Lord through the prophet” was fulfilled) – Codex Colbertinus states specifically that the prophecy was given by Isaiah.
This phenomenon is not limited to one medieval Old Latin copy. Codex Bezae, which D. C. Parker has assigned to c. 400, also includes Isaiah’s name in the text of Matthew 1:22, both in its Latin text and in its corresponding Greek text. Old Latin Codex Veronensis (VL 4, from the 400’s) also has Isaiah’s name in Matthew 1:22. So do the Old Latin codices Brixianus (VL 10, from the 500’s) and Sangermanensis (VL 7, c. 810) and Vercellensis (VL 3, probably from the 370’s). (Metzger expressed some uncertainty about Codex Vercellensis’ reading in his Textual Commentary, but “ESEIAM PROPHETAM” is shown clearly in Irici’s 1748 presentation of Codex Vercellensis.) “Isaiah” is practically the normal Old Latin reading in Matthew 1:22.
The earliest evidence for the reading “in Isaiah the prophet” in Matthew 1:22, however, may be even earlier than the earliest Old Latin manuscript: in the Latin text of Irenaeus’ Against Heresies 3:21:4 (composed in Greek c. 184), Irenaeus quotes Matthew 1:22: “et quoniam Angelus in somnis dixit ad Joseph: Hoc autem factum est, ut adimpleretur quod dictum est ab Esaia Propheta: Ecce virgo it utero concipiet.” It is possible that the form of this quotation was altered by the Latin translator of Irenaeus’ work, but that too would be early testimony.
In the Syriac tradition, the same scribal tendency is on display. In Matthew 1:22, the inclusion of the name “Isaiah” is attested by the Sinaitic Syriac, the Curetonian Syriac, the Harklean Syriac, and the Palestinian Aramaic.
Another Western witness that displays the tendency to fill the vacuum when a prophet’s statements are cited without specifying his name is the Middle Egyptian Glazier Codex of Acts (G67, from the 400’s). Instead of “in the prophets” in Acts , G67 reads, “in Habakkuk the prophet.”
The scribal tendency toward specificity is also displayed by the core members of family-1. Although these manuscripts are medieval, they are generally thought to represent a text of the Gospels similar to a text used by Origen at Caesarea in the 200’s; this is indicated by their support for the reading “Jesus Barabbas” in Matthew 27:17; according to a Latin translation of Origen’s Commentary on Matthew, Origen stated that some of his copies had this reading.
The text of f1 indicates that copyists of the manuscripts used by Origen were not exempt from the tendency toward specificity, and that occasionally the scribal attempt to make the text more specific was poorly executed. In Matthew , where most manuscripts simply read “by the prophet,” without naming the prophet being quoted, the text in f1 includes a specific name: Isaiah.
That is not a correct reference; Matthew’s quotation clearly comes from Psalm 78:2, which was composed by Asaph, not by Isaiah. Yet an early copyist’s need for specificity was greater than his grasp of the contents of the Old Testament, and the name “Isaiah” was perpetuated in various manuscripts, including minuscules 1, 543, 788, 230, 983, and 1582 (and some others), and Codex Θ.
Eusebius of Caesarea, in his Commentary on the Psalms, mentioned that some copies read “in Isaiah the prophet” in Matthew , but not the accurate copies.
Jerome, in his Homily 11 on Psalm 77 (our Psalm 78), cited Matthew 13:35 and claimed that the reading “through the prophet Asaph” is supported by “all the ancient copies” – “in omnibus ueteribus codicibus” – but it was changed by ignorant individuals (see Amy Donaldson’s Explicit References to New Testament Variant Readings Among Greek and Latin Church Fathers, Vol. 2, pages 369-370). In addition, Jerome wrote that Porphyry (an anti-Christian author who wrote c. 270) made an accusation against Matthew that can only be accounted for by Porphyry’s use of a copy of Matthew with the reading “in Isaiah the prophet” in Matthew :
“Porphyry, that unbeliever . . . says, ‘Your evangelist, Matthew, was so ignorant that he said, “What is written in Isaiah the prophet: I will open my mouth in parables, I will utter mysteries from of old.”’ . . . Now, just as this was the scribes’ error, it was, likewise, their error to write ‘Isaiah’ instead of ‘Asaph.’”
Jerome proceeded to offer a theory that some early copyist, reading “Asaph the prophet” in his exemplar, did not recognize the name “Asaph,” and replaced it with “Isaiah.” He offered the same line of reasoning in his Commentary on Matthew. On the premise that Jerome was not being altogether deceptive, it would appear that the text of Matthew in copies that he considered ancient had been expanded to include Asaph’s name. (We shall take a closer look at Jerome’s testimony later.)
The tendency to make non-specific quotations of Old Testament prophets more specific – via the insertion of a prophet’s proper name rather than “through the prophet” or “by the prophet” – was so strong that copyists in the Western and Caesarean transmission-streams inserted prophets’ names in various passages – and, in the case of Matthew 13:35 in the Caesarean transmission-stream, perpetuated a specific name even when it was the wrong name.
The scribal tendency toward specificity was so strong in the Old Latin transmission-line that in Old Latin Codex Vercellensis (VL 3) a copyist felt that it was necessary to identify the prophet being quoted in Matthew 2:5. Four copies of the Harklean Syriac display the same tendency, but their copyists exercised restraint by only putting Micah’s name in the margin of this passage. In VL 3 (probably produced in the 370’s), the copyist (or his exemplar’s copyist) embedded the prophet’s name directly into the text – and, making matters worse – the identification is incorrect: VL3 reads there, “per Eseiam propheta,” that is, “through Isaiah the prophet.”
What about Alexandrian witnesses? Yes; although not as heavily as elsewhere, the tendency toward specificity impacted Alexandrian manuscripts too: “Isaiah the prophet” is the reading of Codex ﬡ at Matthew .
In the margin of Matthew 2:15 in Codex Sinaiticus, we see how precarious it would be to assume that copyists knew the Old Testament too well to attribute to Isaiah a passage from a different Old Testament book. Matthew contains a quotation of Hosea 11:1 – “that it might be fulfilled which was spoken by the Lord through the prophet, saying, ‘Out of
I called My Son.’” Someone did not recognize that the passage
being quoted was Hosea 11:1 (because in the Septuagint, Hosea 11:1 reads
differently, as “When Israel was a child
then I loved him, and called his sons out of Egypt”) and thought instead
that Matthew was referring to a passage in Numbers – maybe 15:41 or 20:16 – and
for that reason, he wrote in the margin of Codex Sinaiticus, in small
vertically stacked lettering, “In
However reasonable it might seem to assume that copyists knew the Old Testament so well that they would not have risked giving the impression that they attributed a passage to Isaiah that did not originate with Isaiah, there is evidence against such an assumption. Not only does the text of VL3 attribute Micah 5:2 to Isaiah in Matthew 2:5, but in Matthew 21:4 (according to Metzger in Textual Commentary, page 54), a few Vulgate copies, Bohairic copies, and Ethiopic copies add Isaiah’s name, although the quotation is from Zechariah.
Not all copyists were familiar with the Old Testament text, and for most of those who did know the Old Testament well, the text they knew was the Septuagint. Consequently there was a risk that copyists would imagine that their exemplars contained an error when a Gospels-manuscript contained a form of an Old Testament passage that did not match up with the form in which it was found in the Septuagint.
Mark’s use of Malachi 3:1 is one such case. His utilization of Malachi 3:1 closes with the phrase, “who shall prepare your way” (ὃς κατασκευάσει τὴν ὁδόν σου) in the Alexandrian text, or, in the Byzantine Text, “who shall prepare your way before you” (ὃς κατασκευάσει τὴν ὁδόν σου ἔμπροσθέν σου). Neither is an exact match with the text of Malachi 3:1 in the Septuagint, which ends with the phrase “καὶ ἐπιβλέψεται ὁδόν πρὸ προσώπου μου” – “and he shall carefully look for a way before me.” (See Maurice Robinson’s article Two Passages in Mark in Faith & Mission, 13/2 (Spring 1996), pp. 66-111.) An additional factor to consider is that the Septuagint’s text of Exodus 23:20a reads Καὶ ἰδοὺ ἐγὼ ἀποστέλλω τὸν ἄγγελόν μου πρὸ προσώπου σου – “And behold, I send my messenger before your face” – which might have caused some copyists to wonder if the readers of their copies would suppose that Mark was using a passage in the Law, rather than in the Prophets.
So if anyone wonders, “If the reading ‘in Isaiah the prophet’ in Mark 1:2 is not original, where did it originate?”, let the copyists of ﬡ, D, Θ, the Old Latin copies, and the main manuscripts of f1 reply: from the same place that their readings “in Isaiah the prophet” in Matthew 1:22, Matthew 2:5, and Matthew 13:35 originated: from the propensity of some early scribes to make non-specific references more specific.
A faint echo of the kind of scribal confusion that led to the insertion of Isaiah’s name in Mark 1:2 (or an independent repetition of it) may be seen in two medieval Bohairic manuscripts. Boh-E1 (a Coptic-Arabic manuscript produced in 1208), has the Bohairic words for “Exodus” and for “Malachi” in the margin near Mark 1:2. An Arabic note says, “A copy has, ‘the prophets.’” Boh-O1 (a Coptic manuscript produced in the 1300’s) has an Arabic note that says, “Isaiah prophesied with the voice of one crying, and Moses and Malachi prophesied with the sending of the messenger.” The notes in both copies show that to some copyists, Mark 1:2 was understood to refer not just to the Prophets, but to a passage in Exodus.
To be continued in Part Two: Mark 1:2, Origen, and Jerome