John 3:13: The Son of Man Who is in Heaven

            Today we shall look into the question of whether or not the phrase “who is in heaven” was in the original text of John 3:13.

            In two recently published translations of the New Testament – the Evangelical Heritage Version and the Modern English Version – John 3:13 ends with the phrase, “who is in heaven.”  This is also the reading of the King James Version.  It is supported by the vast majority (over 95%) of Greek manuscripts, as well as the Old Latin, the Vulgate, the Peshitta, the Ethiopic version, and a wide variety of early patristic writers. 

Manuscript 0141, in which the text of John is
interspersed with extracts from patristic commentaries, has
an unusual difficulty-relieving variant in John 3:13.

            The NIV and ESV, however, do not include this phrase, following instead the shorter Alexandrian text that is displayed in Papyrus 66, Papyrus 75, Vaticanus, Sinaiticus, some Egyptian versions, and some patristic writers. 

            The late Bruce Metzger, in his Textual Commentary on the Greek New Testament, defended the decision of the compilers of the United Bible Societies’ printed Greek New Testament to reject this phrase:  “The majority of the Committee, impressed by the quality of the external attestation supporting the shorter reading, regarded the words ὁ ὢν ἐν τῷ οὐρανῷ as an interpretive gloss, reflecting later Christological development.” 

            Against this theory of “later” – but earlier than Hippolytus, who cited John 3:13 with the phrase “who is in heaven” in Against Noetus – expansion, Wieland Willker responded effectively:  “Internally the longer reading is clearly the harder reading and there is no reason why the words should have been added.  Metzger says it could be an “interpretive gloss, reflecting later Christological development”, but is this probable?  It seems more probable that scribes omitted the difficult words or changed them as 0141, Sy-S [the Sinaitic Syriac] and e [Old Latin codex Palatinus, from the mid-400s], Sy-C [the Curetonian Syriac] did.” 

            Willker was referring to alterations in the text of Old Latin codex Palatinus (from the mid-400s) and the Curetonian Syriac that yield the meaning of “was in heaven” and in the Sinaitic Syriac manuscript that yields the meaning of “is from heaven.”

            This textual variant overlaps with an interpretive question:  writing in ancient Greek with no (or only minimal) punctuation, did John intend to report that Jesus told Nicodemus, at the time of their conversation, that the Son of Man was in heaven?  Or, if the phrase is assumed to be original, was it intended to be understood, not as part of Jesus’ words, but as a parenthetical phrase made by John? 

            As Willker noticed, it is not hard to see why early copyists would consider the phrase puzzling:  if the phrase is not understood as a parenthetical comment by John, then Jesus seems to say that the Son of Man is in heaven, while He is right there on the scene talking to Nicodemus.  Internal considerations thus weigh in heavily against the shorter reading:  to remove this phrase would be to reduce the risk of misinterpretation, whereas a copyist who added this phrase would be adding an interpretive difficulty where there previously was none.    

The phrase “ὁ ὢν ἐν τῷ οὐρανῷ” is far more likely to be part of the original text, excised in the Alexandrian text-stream by a copyist prone to relieve perceived difficulties, that it is to have originated as a scribal expansion.  The absence of this phrase in the fifth edition of the UBS/Biblica Greek New Testament, and in the 28th edition of the Nestle-Aland Novum Testatementum Graece, is an echo of previous compilers’ reliance upon poorly represented data combined with a preference for manuscripts that happened to be stored in a dry climate.  In the Sinaitic Syriac and the Curetonian Syriac and the Old Latin Codex Palatinus (and in the uncial 0141, in which the closing phrase states that the Son of Man is from heaven) we see copyists surrendering to the temptation to alter the text in order to resolve a perceived difficulty; the manuscripts that lack the phrase echo the work of an early scribe who took things a little further.  

(Those who would object, “But we should follow the oldest manuscripts” are advised to notice that Papyrus 75 reads πιστεύετε (not πιστεύσετε) at the end of 3:12, and in nearby 3:31, the scribe of papyrus 66 initially omitted the word ἐρχόμενος, and also in 3:31, the scribes of Papyrus 75 and Sinaiticus both did not include the final phrase ἐπάνω πάντων ἐστίν; the compilers of the UBS/NA texts obviously felt no obligation to follow the oldest manuscripts unthinkingly.  Nor should we.)

The recently published Tyndale House edition of the Greek New Testament likewise omits the phrase; the pitiable brevity of its apparatus prevents the reader from seeing the early patristic evidence.  Had the editors accepted the judgment of Samuel Tregelles, the scholar from the 1800s whose work laid the foundation for the THEGNT, this phrase would have been retained.  (But at least they mention the variant; the New American Standard does not even do that.)

Those who wish to read more about this variant-unit may wish to consult Dr. David Alan Black’s article (from Grace Theological Journal, 1985) on the subject.

P.S.  A comparison of the treatment of John 3:13 in different editions of the UBS Greek New Testament does not build confidence in the reliability of the resources upon which the Committee-members depended, or in the consistency of their presentations of the evidence.  In the first edition (1966), the Ethiopic version was listed as a witness for the non-inclusion of “who is in heaven.”  The Arabic Diatessaron was listed as a witness for the inclusion of the phrase.  The Georgian version was listed in support of the longer reading.  Didymus was listed as a witness for both readings.  In the fourth edition (1993), part of the Georgian evidence was listed as support for the shorter reading, the Diatessaron was only listed for the shorter reading, Didymus was listed only as support for the shorter reading, and the Ethiopic version switched sides, favoring the inclusion of the phrase.  And the certainty-rate varied from A (“the text is certain”) to C (“the Committee had difficulty”) to B (“the text is almost certain”).   

[Readers are invited to double-check the data in this post.]

  

Hand to Hand Combat: Codex D versus 1324

            It’s time once again for hand-to-hand combat!  Today’s combatants are Codex Bezae and minuscule 1324.  The arena is Luke 8:19-25.

            Codex Bezae, also known as Codex 05 and as Codex D, is named after Theodore Beza, an important Protestant theologian and textual researcher of the 1500’s who owned the manuscript, and who donated it to Cambridge University in 1581, where it resides to this day.  (Because it is at Cambridge, it is also known as Codex Bezae Cantabrigiensis.) 

            Readings from Codex Bezae were cited, albeit inaccurately, in the footnotes of Stephanus’ 1550 edition of the Greek New Testament.  This poses a problem for those who claim that the Textus Receptus was compiled using only late manuscripts, because Codex D is not late.  It has been assigned to the end of the 300’s or early 400’s by David C. Parker (though earlier researchers considered it a century younger).  It is the primary Greek support for what has been called the “Western” text of the Gospels. 

            Codex Bezae is a Greek-Latin manuscript; it contains not only the Greek text of (most of) the Gospels and Acts, but also the Latin text, on alternating pages, so if one opens the codex to any undamaged portion, the Greek text of a passage will be on the page to the reader’s left, and the Latin text of approximately the same passage will be on the page to the reader’s right.  (Also, a Latin page containing a snippet of text from the end of Third John survives, testifying that the codex originally had more books than it does now.)

            Minuscule 1324 is a Greek Gospels-manuscript from the 1000’s, and is part of the collection held by the Jerusalem Patriarchate; it is catalogued as Panagios Taphos 60.  Before the text of the Gospels begins, 1324 has the Eusebian Canon-tables, elaborately decorated, and before the canon-tables is Ad Carpianus (a letter from Eusebius to Carpian, which serves as a manual on how to use the canon-tables to find parallel-passages in the Gospels).

            Notably, the text of Ad Carpianus is framed within a quatrefoil, similar to the same feature in minuscule 157, and also like the empty framework found in minuscule 1191 (a Gospels-manuscript at Saint Catherine’s monastery), and like the format of the Armenian text of Ad Carpianus in Walters MS 538 – but especially reminiscent of the cut-out framework of Ad Carpianus in minuscule 2812 (the Zelada Gospels).

            Miniscule 1324 is a beautiful manuscript, neatly written and finely illustrated.  Each Gospel in Codex Bezae begins with red text (see Matthew, John, Luke, and Mark), and its lettering is sufficiently neat, but the artistry in 1324 is far more impressive; each Gospel in 1324 has a full-page illustration of its author, and an intricately detailed headpiece (see Matthew, Mark, Luke, and John).  (In these pages, by the way, notice the use of a metobelus-mark (÷) to signify the beginning of a section of text, rather than a textual variant.)  In addition, the parchment of the pages on which the chapter-lists for the Gospels are written has been dyed purple – a rare and sumptuous feature.

            But how will the text of 1324 perform in the ring against an opponent that is 600 years older?  Let’s find out by comparing the text of Luke 8:19-25 in each manuscript to the text of the 27th edition of the Nestle-Aland Novum Testamentum Graece.     

            First, the rules of the ring:  each non-original letter in a manuscript’s text is counted as a point, and each original letter that is absent in a manuscript’s text is counted as a point; the winner will be the manuscript with fewer points.  Transpositions are mentioned but not counted if no text is lost.  Contractions are not considered variants.  Itacisms, iota adscript/subscript, and movable-nu variants will be listed and counted, but a separate calculation will be made which does not take them into consideration. (These inconsequential variants are indicated by underlined verse-numbers.)  

            Here is a comparison of 1324’s text to the text of NA27: 

19 – 1324 reads Παρεγένοντο instead of Παρεγένετο (+2, -1)

19 – 1324 reads αυτωι instead of αὐτῷ (+1)

20 – 1324 reads Καὶ before ἀπηγγέλη instead of δὲ after it (+3, -2)

20 – 1324 reads λέγοντων after αὐτῷ (+8)

20 – 1324 reads σε θέλοντες instead of θέλοντες σε (transposition)

21 – 1324 reads αὐτόν at the end of the verse (+5) [This Byzantine reading is not noted in NA27.]

22 – no variants.

23 – 1324 reads ἀφύπωσε instead of ἀφύπωσεν (-1)

23 – no variants.

24 – 1324 reads Καὶ before προσεθόντες instead of δὲ after it (+3, -2)

24 – 1324 reads ἐγερθεὶς instead  of διεγερθεὶς (-2) [This Byzantine reading is not noted in NA27.]

24 – 1324 reads ἐπετίμησε instead of ἐπετίμησεν (-1)

24 – 1324 reads ἀνέμωι instead of ἀνέμῳ (+1)

24 – 1324 reads τωι instead of τῳ (+1)

24 – 1324 reads μεγάλη after γαλήνη (+6)

25 – 1324 reads ἐστιν after ποῦ (+5)

25 – 1324 reads τωι instead of τῳ (+1)

25 – 1324 reads αυτωι instead of αὐτῷ (+1)

            Thus, in these six verses, 1324 has 37 non-original letters, and is missing 9 original letters, for a total of 46 letters’ worth of corruption.  When inconsequential orthographic variants are removed from the equation, 1324 has 32 non-original letters, and is missing 7 original letters, for a total of 39 letters’ worth of corruption.  

            Now let’s compare Codex Bezae’s text of Luke 8:19-25 to NA27, and see how well “one of our oldest witnesses” (as described by Bart Ehrman) performs:

19 – D reads αυτου after μήτηρ (+5)

19 – D reads οτι after αυτω (+3)

20 – D reads εξω εστήκασιν instead of εστήκασιν εξω (transposition)

20 – D reads ζητουντες instead of ιδειν θέλοντες (+5, -9)

21 – D does not have προς after ειπεν (-4) 

21 – D reads αυτοις instead of αυτους (+1, -1)

21 – D reads οι before αδελφοι (+2)

22 – D does not have και after ημερων (-3)

22 – D reads αναβηναι αυτον instead of αυτος ενεβη (+6, -3, assuming a transposition)

23 – D reads λελαψ instead of λαιλαψ (+1, -2)

23 – D reads πολλη after ανεμου (+5)

24 – D reads κε κε (i.e., κυριε κυριε) instead of επιστατα επιστατα (+4, -16) [This could fairly be counted as +10, -16.]

24 – D reads ἐγερθεὶς instead  of διεγερθεὶς (-2)

24 – D reads επετειμησεν instead of επετιμησεν (+1)

24 – D does not have του υδατος (-9) 

25 – D reads εστιν after που (+5)

            Thus Codex D has 38 non-original letters in Luke 8:19-25, and is missing 49 original letters, for a total of 87 letters’ worth of corruption.  When inconsequential orthographic variants are removed from the equation, Codex D has 36 non-original letters, and is missing 47 original letters, for a total of 83 letters’ worth of corruption. 

            Let’s go to the scorecards.  When accretions are compared, 1324 wins:  it has only 32, whereas D has 36.  (Even if itacisms and such were considered, 1324 still wins by one.)  And when omissions are compared, 1324 virtually knocks Codex Bezae out of the ring:  1324 omitted 7 original letters but Codex D omitted 47!  The clear winner:  minuscule 1324! 

Some Post-Fight Analysis

            This comparison, though anecdotal, suggests that a few common axioms should be challenged or significantly adjusted:

            ● The oldest manuscripts should be preferred . . . right? 

            It seems perfectly reasonable to think that the older a manuscript is, the better its text is likely to be.  Every time the text was copied, there was a risk of the introduction of new corruptions.  The older a manuscript is, the fewer generations of copies are likely to be between it and the autograph. 

            However, the force of this mere likelihood shrinks when one observes the liberties that were taken by the copyists in the Western textual tradition in the second and third centuries.  Situated in locales where Greek and Latin competed to be the lingua franca, Western copyists prioritized the meaning of the text, and were not averse to replacing original expressions with different expressions that seemed to them to be more precise, more reverent, and less vulnerable to misunderstanding. 

            An example:  in Mark 7:19, Jesus says that after a man has eaten food, what remains – that is, dung – “goeth out into the draught,” as the KJV puts it.  The “draught” (Greek ἀφεδρῶνα) is a latrine or toilet.  Some English translators, softening Jesus’ earthy reference, translate this as “sewer” (see for example the MEV, NET, NLT, and NRSV).  Some others are yet more evasive, simply saying that the food is eliminated or expelled from the body (see for example the NKJV, ESV, CSB, NIV, and NASB).  Just as our modern English translators have tended to avoid Jesus’ reference to a latrine or toilet in Mark 7:19, so did the person or persons responsible for the Western Text:  Codex D replaces the latrine with a sewer, reading οχετον (i.e., “sewer”). 

            One might say that if the text of the Gospels were translated into English using a “dynamic equivalence” technique, occasionally resorting to paraphrase, and then translate the resultant English text back into Greek, the result would be similar to the Western Text.  Very many alterations to the form of the text resulted as copyists attempted to maximize what they perceived to be the meaning of the text – and this was happening in the 100’s and 200’s (not, as far as the Gospels-text is concerned, as a one-time revision, but as an ongoing process).  There is thus no reason to expect that manuscripts fished out of a transmission-stream heavily contaminated by Western corruptions will have fewer corruptions than  manuscripts from some other transmission-stream, regardless of their ages.

            If the editors of the Textus Receptus had trusted their earliest available manuscript in the 1500’s, then instead of introducing 33 letters’ worth of corruption (working from the premise that NA27’s compilation of Luke 8:19-25 is completely correct), they would have introduced 83 letters’ worth of corruption.

            ● Scribes tended to add rather than omit . . . right? 

            In Luke 8:24, 1324 displays a harmonization:  “great” (μεγάλη) was added to “calm” (γαλήνη), which brings the passage into closer agreement with the parallel-passage in Matthew 8:26 and Mark 4:39.  Yet, on balance, when compared to Codex D – a manuscript 500 years older – 1324 has slightly fewer accretions in Luke 9:19-25.  If (as Daniel Wallace has claimed) copyists were applying “If in doubt, don’t throw it out” as a basic principle, and were thus expanding the text for 600 years (from 400 to 1000), how is it that the Byzantine manuscript from the 1000’s has a text of Luke 8:19-25 with fewer accretions than a text from the early 400’s?       

            The more the text-critical canon, “prefer the shorter reading” is tested, the more its wrongness is demonstrated.  Yet the critical text of Nestle-Aland/UBS still agrees quite closely with the Westcott-Hort 1881 compilation, for which this canon was constantly in play.            

            ● The Western Text of the Gospels is characterized by expansion and elaboration . . . right?

            The Western Text, according to the late Bruce Metzger (see p. 213, The Text of the New Testament) “is usually considered to be the result of an undisciplined and ‘wild’ growth of manuscript tradition and translational activity.”  I draw your attention to the word “growth.”  A better term might be “change,” because it is not unusual at all to find readings in the Western Text that are shorter than their rivals.  In the six verses studied here, inasmuch as Codex D’s text has 36 non-original letters, and is missing 47 original letters, its text is 11 letters shorter than the original text (using NA27 as the basis of comparison). 

            In one variation-unit in this passage, the text of D is shorter because of the excision of perceived superfluity:  in verse 24, the phrase “of the water” has disappeared, and one can picture a Western copyist thinking, “There’s no need to say that the waves were made of water.”  The same phenomenon is observable in modern paraphrases; look in the CEV, the Easy-To-Read Version, the God’s Word translation, the hyper-paraphrase known as “The Message,” and the New American Bible for examples.    

            Therefore, when encountering a Byzantine reading that is longer than its Western rival, we should consider the intrinsic character of the Byzantine reading, and ask, “Could a translator consider the content of this variant superfluous?” and if the answer is “Yes” then the Western reading should be considered suspect.  Perhaps one could go further and say that the same approach should be in play when comparing Byzantine and Alexandrian readings – and it does not seem absurd to suggest that when Alexandrian and Western witnesses support the non-inclusion of a superfluous-seeming word or phrase, we may be seeing the effects of the same scribal tendency in both transmission-streams.

           

            ● The Textus Receptus is essentially a late medieval text . . . right? 

            Setting aside itacisms and other such variations, the Textus Receptus reads just like 1324 in Luke 8:19-25, with a few exceptions:
            22 – TR reads Καὶ εγένετο instead of Ἐγένετο δε (+3, -2)

            24 – TR reads προσεθόντες δὲ instead of Καὶ προσεθόντες (agreeing with NA27)

            24 – TR does not read μεγάλη after γαλήνη (agreeing with NA27)

Which means that the Textus Receptus is slightly more accurate in Luke 8:19-25 than 1324’s text.  If one were to select any medieval Byzantine manuscript at random, its text of Luke 8:19-25 would very probably trounce the text of Codex D in a direct comparison.  

_______________
Readers are invited to double-check the data and calculations in this post.