There is something in Luke 9:23 that is not found in Matthew 16:24 or Mark 8:34. In all three of the Synoptic Gospels, Jesus delivers His saying about taking up the cross shortly after Peter made his famous confession, recognizing that Jesus is the Christ. And in all three, Jesus says that whoever wishes to follow Him ought to deny himself. But when it comes to taking up the cross, Luke 9:23 includes a little phrase that Matthew and Mark did not mention: καθ’ ἡμέραν – “daily.”
|Luke 9:23 in Codex Alexandrinus.|
Or does it? In most Greek manuscripts in which Luke 9:23 is extant, καθ’ ἡμέραν / “daily” is not in the text. Here we have one of those unusual occurrences where the Textus Receptus (the KJV’s base-text), instead of agreeing with the Byzantine reading (which is supported by most manuscripts), agrees with the Alexandrian Text.
Let me say that again: in Luke 9:23, the longer reading is opposed by most Greek manuscripts, but it was adopted in the Textus Receptus, and it is also supported by Papyrus 75, Codex Sinaiticus, and Codex Vaticanus. The longer reading is also supported by the Peshitta, and the Gothic version, and Codices A K L M W Π, and by Cyril of Alexandria (Sermons 50 and 51). Gregory the Great used the longer reading in his Homily 32 on the Gospels.
This situation poses a problem for the widely-circulated model of the text’s history in which there are two distinct transmission-lines – one from
perpetuating the Textus Receptus, and
one from ,
producing the base-text of modern versions such as the ESV). Some KJV-Onlyists use that model to give the
impression that the Textus Receptus
is supported by the majority of manuscripts; however, while that is generally the case (at least, in
Matthew-Jude), it is not always the
case. Things are not really that simple. In the case at hand, instead of seeing the
“Antiochan” transmission line support the reading in the Textus Receptus, and the Egyptian text support some other reading,
the situation is reversed. The inclusion
of καθ’ ἡμέραν in Luke 9:23 is just one of many readings in the Textus Receptus that are only supported
by a minority of Greek manuscripts, and which cannot plausibly be considered to
have an advantage in terms of either their popularity or their scope of influence
in the Greek-speaking churches. Alexandria
To emphasize: in Luke 9:23, the Textus Receptus agrees with the Alexandrian Text and disagrees with the Byzantine (Majority) Text.
Is there any way that the Byzantine reading here could be original? Consider the following:
● The Byzantine reading in Luke 9:23 is shorter than its rival. There are still some textual critics who routinely teach their students that the shorter reading is to be preferred (although if one examines how editors have applied that flawed and obsolete principle, they might as well have said, “prefer the shorter reading unless it is Byzantine”).
● The Byzantine reading maintains agreement among the Synoptic Gospels. To some scholars, this is a point against the reading; the argument being that a scribal tendency to harmonize the accounts led to the deletion of καθ’ ἡμέραν so that all three parallel-passages would be identical. But this raises two questions: first, if (a) καθ’ ἡμέραν is original, and (b) copyist typically expanded the text, as detractors of the Byzantine Text typically assume, then why didn’t scribes seeking to harmonize the parallel-passages simply insert καθ’ ἡμέραν into Matthew 16:24 and Mark 8:34? We do not see such an insertion (at least, not on any scale) in these passages, nor in Mark 10:21 where the invitation to take up the cross appears in most manuscripts (but not in À B C D). And, second, reckoning that Luke was dependent upon Marcan material, where did Luke get the idea to add καθ’ ἡμέραν?
● The longer reading could have been created as a means of interpreting the phrase “take up your cross” through the lens of Paul’s words in First Corinthians 15:31, where καθ’ ἡμέραν appears at the beginning of the verse. John Burgon advocated this explanation, describing the Textus Receptus’ reading here as a “spurious accretion.” Burgon also pointed out (in Causes of Corruption, pages 176-178) that Chrysostom connected the two passages; however, a consultation of Chrysostom’s Homily 5 On the Statues, part 14 shows that the passage that Chrysostom cites there is not Luke 9:23, nor one of its parallel-passages; it is Matthew 10:38, where, as far as I know, καθ’ ἡμέραν never appears in any Greek manuscript.
Jerome, in Epistle 127, To Principia, On Marcella, part 6, connects First Corinthians 15:31 and Luke 14:27 – but as he quotes Luke 14:27, it is with the Latin equivalent of καθ’ ἡμέραν included, and he states that this is how the verse appeared in the ancient copies; a quotation will be illustrative:
“Marcella . . . often quoted with approval Plato’s saying that philosophy consists in meditating on death – a truth which our own apostle [i.e., Paul] endorses when he says: for your salvation I die daily. Indeed, according to the old copies our Lord himself says, Whoever does not bear His cross daily and come after me cannot be my disciple.”
|Luke 9:23 in a Byzantine minuscule.|
Amy Donaldson, in her 2009 dissertation Explicit References to New Testament Variant Readings Among Greek and Latin Church Fathers, proposed that Jerome implied via his reference to “the old copies” that other copies existed in his day which omitted Luke 14:27. Although there are nowadays copies extant that do indeed omit Luke 14:27 (Codex M, Γ, 2*, 69) due to parableptic error (skipping from μαθητής at the end of verse 26 to μαθητής at the end of verse 27), it seems to me more likely that Jerome was thinking of a contest between the presence and absence of “daily.” Affirming this, Jerome says in his Commentary on Matthew, in a comment on Matthew 10:38, “In alio evangelio scribitur, Qui non accipit crucem suam quotidie,” that is, “In another Gospel, it is written, he who does not take up his cross daily.” (The focus of these comments is Luke 14:27, but they might be considered relevant to the question about the text of Luke 9:23, for they seem to indicate that a scribe expanded Luke 14:27 by adding the word “daily,” and what would motivate a scribe to do this, if not the presence of “daily” in Luke 9:23?)
● Although the ancient codices that represent the text used in
favor the inclusion of καθ’
ἡμέραν, there are witnesses of almost the same age for the shorter
reading. Not only is the Sinaitic Syriac
among the witnesses which weigh in for the shorter reading, but so are several Old Latin copies, including (according
to Burgon): Egypt
b, VL 4, Codex Veronensis, late 400s
c, VL 6, Codex Colbertinus
e, VL 2, Codex Palatinus, 400s
ff2, VL 8, Codex Corbeiensis secundus, 400s
q, VL 13, Codex Monacensis (a.k.a. Codex Valerianus), 500s or 600s.
The thing to see, regarding these witnesses, is not only their relatively early production-dates, but that they echo a text (or texts) translated into Latin sometime before the Vulgate. That is, readings upon which they widely agree are echoes from the 100s-300s. To this testimony is added the voices of Codices C X Δ E F G H S U V Γ Λ and, as already mentioned, most minuscules. The margin of the Harklean Syriac version, from 616, chimes in too; “daily” is in its text but a note affirms that this is not found in all manuscripts.
Codex Bezae, meanwhile, along with the Old Latin codices Vercellensis (late 300s) and Rehdigeranus, is sidelined from this particular context: due to parableptic error (involving a skip from one “and” to another), the entire phrase between “deny himself” and “and follow Me” has dropped out of their text.
And, as Willker observed, Codex Vaticanus has a ¨ symbol alongside the line where καθ’ ἡμέραν is in the text; this symbol signifies a scribe’s awareness of a textual variant in the line which it accompanies; the date at which such symbols were added to Codex B, however, is unknown.
● Hort noticed that the Western Text has a penchant for expansion; this consideration elicited Hort’s theory of Non-Western Interpolations, which specially affected Hort’s compilation in Luke 24. Hort’s preference for short Western readings had a heavy impact and it influenced the Revised Standard Version. That is why, if you take in hand a copy of the Revised Standard Version, you will not find the following:
► The phrase “of the Lord Jesus” in Luke 24:3,
► The phrase “of the Lord Jesus” in Luke 24:3,
► The words “He is not here, but has risen” in Luke 24:6,
► Any of Luke 24:12,
► Jesus’ words of peace in Luke 24:36,
► Any of Luke 24:40,
► Luke’s statement that Jesus was carried up into heaven in 24:51, or
► The word “And they worshiped Him” in 24:52.
The discovery of Papyrus 75 elicited a reversal on the part of the experts who made the RSV, and most of these readings were put back into the text in the New Revised Standard Version (but not the one in 24:3). Subsequently, the producers of some younger versions have done an excellent job of letting their readers forget how different their version’s textual grandfather was in Luke 24; the Christian Standard Version, for example, has plenty of footnotes in Luke 24 but none of them mention these textual variants. Similarly in the Tyndale House edition of the Greek New Testament, only one of these seven variants – the Western Text’s failure to mention Jesus’ ascension – is covered in the textual notes.
But I digress; the thing to see is that the Western Text, or at least an early form of it, represented by the Sinaitic Syriac and several Old Latin manuscripts, resisted its tendency toward embellishment, and maintained the shorter reading in Luke 9:23. Thus an argument in favor of the shorter reading in Luke 9:23 could include the same points that Hort made for the Western Non-Interpolations (a case which persuaded most of the editors of the RSV for a while),
However, Burgon’s intriguing idea about καθ’ ἡμέραν being an insertion by an early scribe cannot withstand the force of the variety of the external evidence in which καθ’ ἡμέραν is confirmed. We face a textual tree which has most of its fruit on a single branch (the medieval minuscules); another branch has the same fruit. But on both of those branches, and on several other branches, there is a different fruit. Grafters have been at work. In this case I think we are obligated to favor the reading which is supported, not by the most fruit, but by most branches of the tree. This means that harmonization-by-deletion occurred in Luke 9:23 in the Old Latin and Byzantine text-lines (and in the Sinaitic Syriac text), and that the longer reading in Luke 9:23 (supported by the earliest stratum of the Byzantine Text, by the earliest manuscripts which attest to the Alexandrian Text, by the Vulgate, by the Gothic version, by the Peshitta, by some of the Old Latin, by and the Armenian version, etc.) should be adopted.
Readers are invited to double-check the data in this post.