Matt 7:26 την οικιαν αυτου

Influenced by the word order of the preceding μου τοὺς λόγους in 7:24, 26, the common ancestor of certain manuscripts (ℵ B W Z Θ Σ f1 892 pc) both here and in 7:24 reflects αὐτοῦ τὴν οἰκίαν, but most scribes resisted the transposition and retain the more common order of τὴν οἰκίαν αὐτοῦ (including C E K L M S U V X Γ Δ Π Ω 047 0211 f13.35 Byz 565. 1424. 1500. 2224). Yet the minority reading is followed by Bover, Greeven, Lachmann, Merk, Soden, Tischendorf (7th, 8th), and Vogels. Cf. the note on Matt 7:24 την οικιαν αυτου for further explanation on this variant, and also Matt 5:20 η δικαιοσυνη υμων for discussion on a similar transpositional variant.

Matt 7:24 την οικιαν αυτου

A small but diverse contingent of manuscripts (ℵ B C W Z Θ Σ f1 33. 892 al), along with Bover, Greeven, Lachmann, Merk, Soden, Tischendorf (7th, 8th), and Vogels, supports the word order αὐτοῦ τὴν οἰκίαν instead of the more common τὴν οἰκίαν αὐτοῦ, as appears in most manuscripts (including E G K L M S U V X Δ Π Φ Ω 047. 0211 f13.35 Byz 565. 1424. 1500. 2224).

Alford (1:75) plausibly reasons that the majority reading reflects "a transposition to more usual order," while Bloomfield (Annotations, 7), after conceding the possibility of Alford's explanation, counters, "but so may the other have been a transposition to a more classical order, such as may be found in very many other passages, . . . and in all of them, with the present, the transp[ositio]n was more likely to attach to a few than to many copies."

Admittedly, the present textual variation involves merely a word order trifle that does no harm to the sense. Nevertheless, it is notable that no reasoned eclectic editor (to my knowledge) supports or even indicates the critical observation that the order with the preposed personal pronoun (i.e. αὐτοῦ τὴν οἰκίαν) in both 7:24 and 7:26 might reflect assimilation to the identically less common and unasailable word order of the expression μου τοὺς λόγους that occurs earlier in both verses.

A brief investigation indicates that of the roughly 420 times where the genitive of ἐγώ (70x), σύ (120x), or αὐτός (230x) modifies a noun or noun phrase in Matthew, only 20 times (5%) does the pronoun precede the modified word(s) without much doubt:

2:2 αὐτοῦ τὸν ἀστέρα
5:39 σου σιαγόνα
6:4 σου ἡ ἐλεημοσύνη
6:17 σου τὴν κεφαλήν
7:24 μου τοὺς λόγους
7:26 μου τοὺς λόγους
8:3 αὐτοῦ ἡ λέπρα
8:8 μου ὑπὸ τὴν στέγην
9:5 σου αἱ ἁμαρτίαι
9:6 σου τὴν κλίνην
9:30 αὐτῶν οἱ ὀφθαλμοί
13:25 αὐτοῦ ὁ ἐχθρός
15:28 σου ἡ πίστις
16:18 μου τὴν ἐκκλησίαν
17:15 μου τὸν υἱόν
22:13 αὐτοῦ πόδας καὶ χεῖρας
23:8 ὑμῶν ὁ καθηγητής/διδάσκαλος
26:43 αὐτῶν οἱ ὀφθαλμοί
26:51 αὐτοῦ τὸ ὠτίον
28:9 αὐτοῦ τοὺς πόδας

Only very rarely is the order in the above passages altered to the "more usual order" in any significant witnesses:

5:39 σιαγόνα σου - B D
6:4 ἡ ἐλεημοσύνη σου - D
9.30 οι οφθαλμοι αυτων - D it vg
13.25 ο εχθρος αυτου - pc it vg etc.
16:18 τὴν ἐκκλησίαν μου - D it vg
28.9 τοὺς πόδας αὐτοῦ - D it vg

Moreover, not including the present case of 7:24, 26, and also 9:2 (which involves a possible omission and assimilation to 9:5), only 6 passages are in dispute. Four involve a change in one direction:

12:13 σου τὴν χεῖρα - ℵ* B L f1 33 pc
23:9 ὑμῶν ὁ πατήρ - ℵ B 0102 33. 892 pc
23.30 αὐτῶν κοινωνοί - B D f1.13 700 pc
27:49 αὐτοῦ τὴν πλευράν - ℵ B C L Γ pc (but Origen: τὴν πλευρὰν αὐτοῦ)

But two involve a change in the opposite direction:

20:33 οἱ ὀφθαλμοί ἡμῶν - ℵ B D L Z 0281vid 33. 892 pc; Or
26:52 τὴν μάχαιράν σου - ℵ B D L 0281vid f1.13 892 1424 pc

In the two places just mentioned, Alford inconsistently rejects the less common word order predominating in the textual tradition (i.e. ἡμῶν οἱ ὀφθαλμοί in 20:33 and σου τὴν μάχαιραν in 26:52), demonstrating the lack of or at least inconsistent value that may be attached to his explanation for rejecting the consensus reading in 7:24, 26.

Furthermore, in other places in Matthew the pronoun is found to be preposed in certain witnesses, especially Alexandrian ones, in opposition to the very rule that forms the foundation of internal evidence for rejecting the less common word order in 7:24, 26. In fact, from the following alterations a case might be made that scribes were just as likely to alter the word order in one direction as they were the other:

4.24 αὐτοῦ ἡ ἀκοή - D
18.31 αὐτοῦ οἱ σύνδουλοι - B
20.13 αὐτῶν ἑνί - B
20.34 αὐτῶν τῶν ὀμμάτων - B
22.6 αὐτοῦ τοὺς δούλους - Origen
24.47 αὐτοῦ τοῖς ὑπάρχουσιν - K Π
27.29 αὐτοῦ τῇ κεφαλῇ - 33

For all of the above observations the likeliest explanation for the preposed personal pronoun in a minority of witnesses in both 7:24 and 7:26 remains assimilation to the same word order that occurs just words before both occurrences, i.e. μου τοὺς λόγους. If so, the consensus reading in both passages actually reflects the one less harmonized to the immediate context and therefore the most likely to be original. For why would so many witnesses reflect conscious alteration of the "uncommon" word order of the second expression in both places while leaving the same "uncommon" word order just words before in both places completely untouched? For a similar transpositional variant, see Matt 5:20 η δικαιοσυνη υμων.

Matt 7:5 την δοκον εκ του οφθαλμου σου

Apparently only four largely related Greek manuscripts (ℵ B C 0281), influenced by the order of the words in 7:3, transpose την δοκον from before the prepositional phrase εκ του οφθαλμου σου to the more emphatic position following it. However, most witnesses, followed by Greeven, preserve the more common word order (E G K L M N S U V W X Δ Θ Π Σ Φ Ω 047 0211 0233 [Byz ca. 1520 mss] f1.13.35 33. 565. 892. 1424. 1500. 2224 latt sy-h; Ir-lat Chr).
     Many editors (Bover, Lachmann, Merk, Soden, Tischendorf [7th, 8th], Vogels) follow the minority  of manuscripts on the ground that all the others assimilated to the order in 7:5b or the parallel passage in Luke 6:42 (so, e.g., Alford, 1:70; Soden, 2:17). For this reason alone Soden (1:1448) remarks that the minority reading might even be the reading of the Jerusalem-Egyptian-Byzantine archetype!
     Alternatively, it seems much more likely that four related manuscripts should share a common ancestor that had the transposition, the origin of which was caused by (a) assimilation to the order of words in 7:3 (εν τω σω οφθαλμω δοκον), (b) a desire to emphasize the object δοκον, or (c) an initial accidental omission of την δοκον by homoeoteleuton error (ον...ον) which the scribe repaired by adding the skipped words after copying the prepositional phrase without doing harm either to the sense or to the beauty of the exemplar. It is hardly necessary to point out that the order εν τω οφθαλμω σου δοκον also appears in Luke 6:42, although accidental or intentional assimilation by the few related manuscripts to the nearer preceding order in 7:3 seems more likely than to the same in Luke.
     In short, due to the conflicting claims of internal evidence, the question is whether just four manuscripts of one type harmonized the text or whether all other manuscripts of all types did so. If the former scenario is not more likely, judgment on this variation may be held in suspense until the character of the few manuscripts in question might be ascertained from the examination of all other variations where a judgment is more certain.

Matt 6:22 ο οφθαλμος σου απλους η

Some witnesses (ℵ B W Φ pc lat; Bas) transpose the verb η to the first position in the clause, while most witnesses have it, as in the following apodosis and both clauses of 6:23, in the final position (including E G K L M S U V Γ Δ Θ Σ Π Ω [Byz ca. 1500 mss] f1.13 33. 565. 892. 1424 it; GregThaum? [Fragmentum in evangelium Matthaei 6.22–23 {PG 10:1189}] Chr Aug). Some lectionaries (l 844. l 2211) naturally omit the verb altogether as it is unnecessary.
     While it is more likely, all things being equal, that an editor would have wished to place the verb last in order to harmonize it with the same position of the verbs in the following clauses (the preposing of the same verb in 6:23 in ℵ* W is a case in point!), it is at least possible that the harmonization is by Matthew's own design (cf., e.g., 5:46–47; 10:13; 16:19; 18:18; 20:26–27; 23:18). In this case, however, two transcriptional considerations may account for the divergence of those few witnesses that place the verb first.

1. The accidental dittography of the final letter of ΟΥΝ (i.e. ΟΥΝΝ) could have caused the introduction of the verb (Η) in that position, as the confusion of the letters Η and Ν was one of the most common in the uncial period, especially in old, damaged, or hard-to-read exemplars.
2. Similarly, the confusion of the final letter of the particle ΟΥΝ for the verb Η would have resulted in the ungrammatical and nonsensical εαν ου η, which also might account for the omission of ουν in ℵ (the others in that textual tradition repairing the "ου" by adding a ν where it apparently had been lost), although the omission of the 3-letter ουν in ℵ also may be accounted for by its similarity with the preceding 3-letter word (εαν) or by harmonization toward an early version (cf. lat sy-c mae bo-ms).

     As internal evidence is rather inconclusive, one may preliminarily accept the order of words found in the vast consensus of manuscripts on the ground of their predominant accuracy in such trivial matters throughout the long sequence of variations investigated up to this point.

Matt 5:36 μιαν τριχα λευκην η μελαιναν ποιησαι

The word order found in most manuscripts (including E K M S U V Δ Π Σ 1006 1506 sy-p.h) contributed to the many existing variations of this phrase, either through (1) scribal desire to bring the complementary infinitive closer to the verb, (2) editorial attempt to polish the expression, (3) imperfect correction of an initial homoeoteleuton error from δυνασαι to ποιησαι, which resulted in a text without μιαν τριχα λευκην η μελαιναν ποιησαι, (4) attempted repair of a conjecture, or (5) a combination of these factors. As this pericope is unique to Matthew, there was no standard to which scribes could turn for help, and consequently the following variations appear:

(1) μιαν τριχα λευκην ποιησαι η μελαιναν (ℵ B W Θ 33 892 1342 pc lat) [ℵ*: τριχαν]
(2) μιαν τριχα λευκην ποιησαι μελαιναν (L pc; Chr) [L: τριχαν]
(3) μιαν τριχα ποιησαι λευκην η μελαιναν (0250 f13 700 h)
(4) ποιησαι τριχα μιαν λευκην η μελαιναν (D f1 205 k; Cl Cyp GrN Aug) [D*: ποιειν; f1 205: μιαν τριχαν]

Fritzsche (240–41) thinks reading #2 arose from someone trying to remove a perceived absurdity, namely, that no one would ever even want to make or produce a white hair, which is a sign of age, and so the expression "you cannot make a white hair black" arose. Incidentally, Kühnöl (159) mentions an expositor (Wakefield) who actually conjectures this reading, but with a different word order (μιαν τριχα λευκην μελαιναν ποιησαι). Still the reading (or conjecture) is old as confirmed by Codex L/019 and Chrysostom, but, not surprisingly, it also created another difficulty, namely, that people can make a white hair black through dye. Thus Bloomfield (GNT, 1:36) considers reading #1 "a vain alteration proceeding from the polishing school of Critics . . . for the sake of removing a difficulty," and indeed the reading is a more classical construction, making it, all things being equal, less likely to be original.

Matt 5:30 βληθη εις γεενναν

Some manuscripts (ℵ B f1 33 205 892 pc [lat] sy-c [mae1] bo) alter the word order and the verb to read εις γεενναν απελθη against the reading of most manuscripts (including E G K [L] M S U V W Γ Δ Θ Π Σ 0233 f13 f vg-ms sy-p.h sa [L: βληθησει εις την γεεναν]). Matthäi (72–3), commenting on 5:29, blames the the widespread influence of Origen for the alteration of βληθη into απελθη, which is represented there in D/05, the Old Latin (a b c d g1 h), the Old Syriac (sy-s.c), and the Bohairic (bo). These same versions also in 5:30 form the major support for απελθη, which is in only two Greek manuscripts from before the ninth century (ℵ B). From this situation it is at least possible, as many early Western and Alexandrian manuscripts were bilingual, that the alteration toward απελθη in Greek in both 5:29 and in 5:30 occurred due to versional cross-contamination. Also in this connection is Codex Bobiensis (k/1), an early and influential Old Latin manuscript with strong and sometimes unique affinities with ℵ B, which has no verb at all for rendering βληθη/απελθη in either verse, a circumstance that could have prompted the introduction of a verb other than the original in any descending copies and then in those that depended on them. Alternatively, Wettstein (1:302), Griesbach (1:52–3), Fritzsche (235), Kühnöl (152), and Bloomfield (GNT, 35) all consider απελθη to be an alteration derived from Mark 9:43, where that word fits well to the context but is "less suitable" (so Griesbach) and "thoroughly inept" (so Fritzsche) at this place in Matthew, perhaps due to the subject (το σωμα). Fritzsche also suggests a solution involving error of the eye (or one should say more likely of the ear while a scribe copied by dictation), namely, that the last syllable of γεενναν before βληθη caused confusion and resulted in αβληθη, which was immediately altered by conjecture to απελθη. This solution requires the earlier transposition of words, perhaps from 3:10 or 7:19. Another passage that could have suggested απελθη is 25:46. Yet Whitney (1:66) reasons that απελθη was "originally an endeavor to avoid the harsher word βληθη," and rightly cautions that internally the dissimilarity of βληθη in 5:29 and απελθη in 5:30 should carry no weight in the decision, since Jesus' "discourses, parables, and conversations generally are distinguished for the sameness of the phraseology in which identical thoughts are expressed," noting elsewhere (1:150) Jesus' fondness for lack of variety of expression, one example being ηκουσατε οτι ερρεθη (5:21, 27, 33, 38, 43). In the end, versional influence or harmonization to Mark 9:43 seem the most likely explanations for prompting the editorial alteration.

Matt 5:25 εν τη οδω μετ αυτου

Although the phrase is transposed to read μετ αυτου εν τη οδω in a number of witnesses (ℵ B D L W 0275 f1.13 33 205 892 pc it sy-s.c.p sa-mss bo-pt; Did), the word order of most witnesses (including E K M S U V Γ Δ Θ Π Σ 0233 f ff1 g2 k vg sy-h sa-mss mae1) is corroborated by  Clement of Alexandria in the third century (Stromata 4.95). Bloomfield (Annotations, 5) calls the transposition an apparent "correction of position such as was likely enough to occur to Alexandrian critics." The transposition in the versions may be due to idiomatic usage, and once entrenched in enough Old Latin, Old Syriac, and even Coptic manuscripts, the regard for those better-understood languages in bilingual manuscripts of the earliest centuries would naturally have led to alterations in some of their Greek counterparts and then in their descendents in turn. Hence the endurance of this transposition, which easily could have occurred independently in Greek copies regardless, would have been strengthened. Still the preponderance of the consensus reading, now extant in nearly 1600 Greek manuscripts and supported by the diversity of major representatives from every text-type, could hardly be diminished. Moreover, were the transposition as reflected in the few Greek manuscripts original, there is no good reason why nearly all the Greek copies would have militated against it.

For Clement of Alexandria, Stromata 4.95, cf. Wilhelm Dindorf, Clementis Alexandrini opera (4 vols.; Oxonii: Clarendoniano, 1869) 2:371.

Note: Carpocrates of the second century (as cited by Epiphanius, Panarion 27.5.3 [Karl Holl, Epiphanius {2 vols.; Die griechische christliche Schriftsteller der ersten drei Jahrhunderte 25, 31; Leipzig: J.C. Hinrichs, 1915, 1922} 1:306]) indeed has "ισθι ευνοων τω αντιδικω σου εν ω ει εν τη οδω μετ αυτου," yet after this he has elements that are more Lukan, such as "και δος εργασιαν απηλλαχθαι απ αυτου" and "και ο υπηρετης βαλη σε εις φυλακην." Nevertheless, the precision of the former definitely aligns with the corresponding Byzantine form of Matt 5:25.

Matt 5:20 η δικαιοσυνη υμων

Almost all earlier and many later manuscripts prepose the pronoun to read υμων η δικαιοσυνη (p64[67]vid ℵ B E K L M S V W Γ Δ Θ Π Ω 047 0211 892 1241 pm; Ju Cl-pt), but many manuscripts (perhaps most according to Tischendorf and Soden) place the pronoun after the subject (including U Σ f1.35 33. 565. 1424 pm latt; Cl-pt Or). Lachmann's departure (1:20) from the earlier manuscripts is remarkable if not accidental. While the reading with the preposed pronoun could have originated due to the order of the words in 5:16 (υμων τα καλα εργα), it is equally likely (if not more so) that this reading could have prompted scribes to place the pronoun after the subject in conformity with its more usual position. The preposing of the pronoun may indicate an emphasis, either by the author himself or secondarily by scribes, on the nature of the listeners' righteousness in contradistinction to that of the scribes and the Pharisees (cf. Bengel, Gnomon, 171–2). Griesbach (1:46) notes that "not a few manuscripts, having changed the order, read υμων η δικαιοσυνη," and further cautions that "rarely can a decisive judgment be rendered due to the great license of scribes in transposing words and the negligence of the majority of learned men who have examined manuscripts in noting discrepancies of this kind." Because (a) the reading with the preposed pronoun has impressive external support as well as perhaps a slight advantage on internal grounds, and (b) the picture of most Byzantine minuscules is not sufficiently clear (cf. Hodges-Farstad's "Mpt" [indicating a Byzantine split] for both readings), it seems best to acknowledge the intrinsic value of both readings until the state of the early minuscules (9th–11th centuries) is better known.

Matt 4:16 ειδεν φως μεγα

Indicative of the atticistic alterations found elsewhere in the same major witnesses (cf., e.g., Matt 3:4 αυτου ην and Matt 4:9 ταυτα παντα σοι), a few witnesses alter the word order to φως ειδεν μεγα in an attempt to improve the style (ℵ B C W Σ f1.13 33 892 1241 it vg-pt; Or-pt Eus Chr Cyr). The normal Koine word order is intrinsically characteristic of Matthew's style, and the many more noticeable verbal dissimilarities of the passage with the LXX make harmonization to the word order there highly unlikely. The alteration was no doubt encouraged by a critical and poetic attempt to align the position of φως here with the position of φως later in the verse (φως ειδεν μεγα . . . φως ανετειλεν αυτοις). Furthermore, the word order found in most witnesses is early and diverse (D E K L M P S U V Γ Δ k vg-pt; Hipp Or-pt).

Matt 4:9 ταυτα παντα σοι

Most manuscripts read ταυτα παντα σοι (including C2 D E K L M P S U Δ Θ Ω f13 lat sy co; Ir-lat), yet an early scribe or editor, motivated by a revival of the Attic dialect, sought to enhance the perfectly good Koine by changing it into the classical ταυτα σοι παντα as reflected in some manuscripts (ℵ B C* W Z Σ f1 33 al; Or). Marks of Atticistic revision in the text of the New Testament are always highly suspect, especially when evidenced by only a few manuscripts primarily of Alexandria, a place esteemed for its critical revision of texts. Cf. also the notes on Matt 2:22 and Matt 3:4. Another possible cause for the minority reading is that one or another scribe accidentally skipped παντα due to homoeoteleuton error (τα...τα), then after writing σοι he repaired his text without loss of meaning by adding the skipped word. Even Lachmann (1:15) supports the reading of most manuscripts, which is corroborated by the oldest witness of any kind, namely, Irenaeus' Old Latin translation, which is significant because it differs from the later Vulgate, which follows the Alexandrian word order.

Matt 4:3 αυτω ο πειραζων ειπεν

     Some manuscripts have αυτω ο πειραζων και ειπεν αυτω (D a b c g h sy-s.c) and others ο πειραζων ειπεν αυτω (p101vid ℵ B W f1.13 33 205 700 892vid al aur ff1 l vg sy-p mae1 bo). Griesbach (1:33–4) defends the consensus reading and describes the genealogy of the variations in this way: first the reading of Bezae (D) and some Old Latin manuscripts arose from scribes adding αυτω after ειπεν, "as happens 600 times elsewhere," but the double αυτω recurring on the same line struck some poorly, and so, second, some naturally omitted the first αυτω and retained only the second, giving rise to the reading in Sinaiticus (ℵ), Vaticanus (B), and others; "and so out of the received reading the origin of all the others is most easily explained" (1:34). Griesbach also cautions that the extra αυτω in similar places in Matthew was added either just after the προσελθων or just after the ειπεν. Bloomfield (GNT, 1:22) argues that "internal evidence is rather in favour of the position after προσ[ελθων] as being the less likely to come from the polishing school of the Alexandrian critics," and that the received reading is "most agreeable to the evangelist's simple phraseology" (Annotations, 3).
     Matthew's usual pattern is to use προσελθων/προσελθοντες absolutely, i.e., without the goal of motion specified (cf. 8:19, 25; 9:20; 13:10, 27; 14:12; 15:12, 23; 16:1; 17:7 [Byz]; 25:20, 22, 24; 26:50, 60, 73; 28:2, 9, 18). Also in 8:2 and 9:18, where the participle varies in the MSS between ελθων and προσελθων, it is in either case absolute. Verses where the goal of motion is indicated after προσελθων/προσελθοντες (by a dative pronoun or articular noun) include 17:19, 21:28, 30; 26:49; 27:58. In 4:3 and 18:21 the Byz MSS indicate the goal but Alexandrian MSS do not; in 19:16 the reverse occurs. As the absolute use of προσελθων was much more common, the likelihood for scribes to have preferred προσελθων without αυτω is increased. Of the various constructions of clauses that contain προσελθων/προσελθοντες in Matthew, the most common is participle—noun phrase/subject—verb—dative pronoun (5x: 8:19; 13:10, 27; 15:12; 28:18). The pattern participle—dative pronoun—noun phrase/subject—verb only occurs twice (4:3 and 18:21), and in both cases it is more likely that the Alexandrian text alters it to the more common pattern than the reverse. In 19:16, however, as in 8:2 and 9:18, the pattern is different in that the subject comes before the participle along with the emphatic ιδου. This appears to require the main verb to follow the participle according to Matthew's usage. (The entire substance of this paragraph is Tony Pope's.)
     Furthermore, as in, e.g., 4:4, Matthew need not always include an indirect object after verbs of speaking, yet it is not unreasonable to suggest that the frequency of occurrence of αυτω after verbs of speaking (as, e.g., in 4:6, 7, 9, 10) could have prompted the transposition in a minority of manuscripts. Also important is the theoretical observation laid out by Griesbach that the consensus reflects the original, the Western reading came second, and the Alexandrian reading came third and is dependent on the earlier Western corruption.

Matt 3:16 ανεβη ευθυς

     The few witnesses (ℵ B Ds W f1 33vid 205 700 pc lat sy-p co) that transpose the words to read ευθυς ανεβη reflect either the critical idea to give ευθυς its more usual place before the verb (so Bloomfield, Annotations, 3; Meyer, 72) or harmonization to the word order of Mark 1:10 (cf. von Soden, 2:7). The word order of most manuscripts (including C E K L M P S U V Γ Δ [0233] f13 d h sy-h), as it is not inconsistent with Matthew (cf. 25:15, but note the alternative punctuation in NA27), is less common, and is not in harmony with Mark, is thus more likely to have undergone alteration than the minority reading, were it original. It is also possible that the versions, which frequently alter word order based on their own conventions, could have affected the word order of a minority of manuscripts here or there, especially since it is known that bilingual manuscripts were valued early on as a way of validating the versions, and vice versa.

Matt 3:11 βαπτιζω υμας

     After εγω μεν at the beginning of the verse, a few manuscripts (p101 ℵ* B W 33 700 990 f1.13) transpose the words to read υμας βαπτιζω "for the purpose of better adapting the words to those at the end of the verse" (Bloomfield, GNT, 19). There was no need for Matthew to emphasize the first υμας, since doing so would have lessened the significance of the emphasis of the second υμας. By far the more memorable clause is the second of the two that contain υμας, and the word order of the second (υμας βαπτισει εν) may well have influenced an early scribe to conform the word order of the less familiar first occurrence to the more familiar second. Additionally, the critic could have thought to improve the construction by choosing a simple way to avoid the hiatus of βαπτιζω υμας, or to attach the prepositional phrase (εν υδατι) to the verb βαπτιζω just as the more famous prepositional phrase at the end of the sentence (εν πνευματι αγιω) is attached to βαπτισει. It is far less likely that all other manuscripts should reflect harmonization either to Mark 1:8 or Luke 3:16 or alteration of a perfectly fine construction than that a minority of manuscripts would reflect conformation to the word order that appears in the very same verse.

Matt 3:4 αυτου ην

     Reflecting an attempt to improve the text, a few manuscripts (ℵ B C D W f1) have ην αυτου, an atticism indicative of the critical editing found in those manuscripts and their relatives. Bloomfield (Annotations, 2) judges that ην αυτου is "the very position that a classical Greek writer would have used," while αυτου ην is the one a Hellenistic Greek writer would have used, thus making it intrinsically more probable. A critical desire to give variance to the phraseology repeated already twice in the same verse (το ενδυμα αυτου . . . την οσφυν αυτου . . . η δε τροφη αυτου) may have motivated the change, but such is precisely the style of the kind of Koine Greek that Matthew writes (cf., e.g., 3:12: τη χειρι αυτου . . . την αλωνα αυτου . . . τον σιτον αυτου, etc.). The slim support for ην αυτου cannot overthrow the aggregate weight of all other manuscripts, and since scribes tended to make the text not less elegant but rather, indeed, to polish and refine unexceptional expressions, it is precisely this class of readings that arouses suspicion, especially when found in only a few manuscripts. Cf. Matt 2:22 επι for a similarly-motivated improvement of the text.

Matt 2:22 Ηρωδου του πατρος αυτου

Reflecting either individual preference or an emphasis on Archelaus' relationship to Herod the Great, a handful of manuscripts (ℵ B C* W) transpose the words to read του πατρος αυτου Ηρωδου, a stylistic adjustment similar to that made by two of the same manuscripts (ℵ B) in Matt 2:3. Such a word order usually reflects emphasis of some kind, and is not uncommon (Mark 11:10; Luke 1:59; John 4:12; cf. also Mark 1:20; Luke 1:59; John 4:12; 8:39, 53; Acts 2:16; 8:28, 30, etc.). Since the word order of the few copies is extremely isolated and, were it original, should hardly have caused all other copies and versions to alter it given its familiarity to scribes, the consensus reading is to be preferred.

Matt 2:19 κατ οναρ φαινεται

Some witnesses (ℵ B D Z Σ 0250 f1.13 157 1574 pc lat) reflect scribal assimilation to the exact expression six verses earlier (2:13) and thus have φαινεται κατ οναρ, but the less exact and slightly more remote passage of 1:20, where the manner of the appearance is given emphasis, demonstrates that κατ οναρ φαινεται is also Matthean, even though at that place some witnesses (Θ 543 788 826 1689 it) alter the word order to εφανη κατ οναρ (removing the emphasis). Fritzsche (98) thinks that Matthew "deliberately preposed κατ οναρ here so that the words more carefully fixed for making up expressions with φαινεται would not perpetrate an annoyance on the reader, had they been reckoned a broken series," namely, "he appears -- (indeed even) in his sleep -- to Joseph -- in Egypt -- with this mandate," etc. There is no good reason to abandon the intrinsically good and externally well-supported reading of the consensus. Note that in 2:13 Codex Vaticanus (B) alone has κατ οναρ εφανη, an obvious assimilation to 1:20, and εις την χωραν αυτων after αυτων, an accretion from 2:12.

Matt 2:8 ακριβως εξετασατε

     A few witnesses (ℵ B C* D Σ f1.13 33vid 157 205 pc lat) transpose the words to read πορευθεντες εξετασατε ακριβως, perhaps (so Wettstein, 1:247) an attempted improvement of the text by removing any ambiguity concerning which word ακριβως was to modify (was it "go directly and search" or "go and search accurately"?). Bloomfield (Annotations, 1) argues that the reading of the few manuscripts "certainly makes better composition" and thus is indicative of "critical emendation," whereas Griesbach (1:19) says the transposition was accidental rather than intentional. But if the error was accidental, it is much more likely that scribes, after writing πορευθεντες, would have skipped ακριβως in anticipation of the imperative since such a construction is relatively common (cf. Matt 9:13; 11:4; 28:19; Luke 7:22; 13:32; 17:14; 22:8) than, were the reading of the few manuscripts original, that they would have skipped the imperative and first written ακριβως. Matthew frequently places the adverb before the word it modifies (cf. 1:18, 19; 2:7; 4:20, 22 [versus αφεντες ευθεως in Mark 1:20]; 5:24, 28; 8:8, 21 [versus απελθοντι πρωτον in Luke 9:59]; 21:28; 26:34; 28:7, etc.). See Matt 14:15, where Matthew's ηδη παρηλθεν is transposed by a few manuscripts (ℵ Z f1 pc), a transposition that may or may not have been influenced by the possible earlier omission of ηδη as reflected in some versions (sy-c.p sa bo); cf. also παλιν in 26:44, αρτι in 26:53, εξω in 26:69, etc., for similar examples of transposition with omission. Griesbach notes that two manuscripts of not much authority omit ακριβως, an error that he believes arose due to the earlier transposition, but which was more likely caused by accident, as mentioned above, due to anticipation of the imperative following πορευθεντες. But scribes were more likely to remove possible ambiguity than vice versa. Accordingly, in Matt 26:44 the same principle militates against the reading of some manuscripts (ℵ B C D L 067 pc lat sa bo), which alter the slightly ambiguous απελθων παλιν προσηυξατο ("again he went away and prayed" or "he went away and prayed again") into the less ambiguous παλιν απελθων προσηυξατο ("again he went away and prayed"; note the early omission of παλιν [p37 Θ f1.13 pc a sy-s], presumably to omit what was thought superfluous, especially if παλιν was taken with προσηυξατο which is already modified by εκ τριτου that follows). For all of these reasons the authenticity of the word order reflected in most Greek manuscripts, namely, ακριβως εξετασατε, is to be maintained.

Matt 2:3 Ηρωδης ο βασιλευς

A few manuscripts (ℵ B D Z f1.13 157 273; b c k bo) transpose the expression to read ο βασιλευς Ηρωδης. However, in a phrase that contains a proper name and a title or other classifying word the normal order in the NT is the title followed by the name. For example, one finds Ιερεμιου του προφητου (2:17; 27:9), Ησαιου του προφητου (3:3; 4:14; 8:17; 12:17), Ιωνα του προφητου (12:39; 16:4), Δανιηλ του προφητου (24:15), Ανδρεαν τον αδελφον (4:18), Ιωαννην τον αδελφον (4:21), Ανδρεας ο αδελφος (10:2), Ηρωδιαδα την γυναικα Φιλιππου του αδελφου αυτου (14:3), Ιωαννην τον αδελφον (17:1), Ιησου χριστου (1:1, 18), Ιησους ο λεγομενος χριστος (1:16), Ιησου υιε του θεου (8:29), Ιησους ο προφητης (21:11, but a few manuscripts have ο προφητης ιησους), Ιησου του Γαλιλαιου (26:69), Ιησου του Ναζωραιου (26:71), Ιησουν τον λεγομενον χριστον (27:17, 22), Ιησους ο βασιλευς των Ιουδαιων (27:37). On the other hand, Matthew also writes της μητρος αυτου Μαριας (1:18) in contrast to Μαριας της μητρος αυτου (2:11), perhaps to draw attention to the relationship. Thus it is possible that the variant reading preposing ο βασιλευς is a critical alteration, either to give variation to the same expression that occurs in 2:1 (so Bloomfield, Annotations, 1) or to put focus on the title to highlight Herod's role as king in the narrative. Matthäi (37) mentions the omission of ο βασιλευς in six manuscripts, which, if not due to simple scribal error, probably resulted because scribes either considered the expression superfluous or sought to conform it to the places in the rest of the chapter where the title is missing (2:7, 12, 13, 15, 16, 19, 22). Especially as many early papyri were originally bilingual Greek/Latin manuscripts, it is also feasible that the title, after being omitted accidentally in some Latin manuscripts where it consists of only three letters (rex), could have contributed to the transposition upon its reintroduction to the text (e.g., it is written above the line in Codex H; 7th cent.). As for other possible causes, it is at least doubtful that the transposition could be due to assimilation to so remote a passage as Mark 6:14, where ο βασιλευς Ηρωδης is secure with only minor variation. In the end, there is little reason to reject the order Ηρωδης ο βασιλευς, which, in addition to most Greek manuscripts, has the support of four Greek manuscripts from the 6th century and earlier (C N W Σ), the oldest Old Latin manuscript (a/3 Vercellensis; 4th cent.) and the important Brixianus (f/10; 6th cent.), and also the Vulgate and Sahidic versions.