² tc The verb λατρεύω (latreuō; here the participial form, λατρεύοντες [latreuontes]) either takes a dative direct object or no object at all, bearing virtually a technical nuance of “worshiping God” (see BDAG 587 s.v.). In this text, πνεύματι (pneumati) takes an instrumental force (“by the Spirit”) rather than functioning as object of λατρεύοντες. However, the word after πνεύματι is in question, no doubt because of the collocation with λατρεύοντες. Most witnesses, including some of the earliest and best representatives of the Alexandrian, Western, and Byzantine texts (א* A B C D² F G 0278^vid 33 1241 1505 1739 1881 2464 M co Ambr), read θεοῦ (theou; thus, “worship by the Spirit of God”). But several other significant witnesses (א² D* P Ψ 075 365 1175 lat sy Chr) have the dative θεῷ (theō) here (“worship God by the Spirit”). P⁴⁶ is virtually alone in its omission of the divine name, probably due to an unintentional oversight. The dative θεῷ was most likely a scribal emendation intended to give the participle its proper object, and thus avoid confusion about the force of πνεύματι. Although the Church came to embrace the full deity of the Spirit, the NT does not seem to speak of worshiping the Spirit explicitly. The reading θεῷ thus appears to be a clarifying reading. On external and internal grounds, then, θεοῦ is the preferred reading.

² tc ‡ The best witnesses read the imperative here (so P⁴⁶ א A B P 0278 33 81 1739 pc). The participle is found primarily in the Western and Byzantine families (D² Ψ 1175 1505 1881 2464 M latt [D* F G are slightly different, but support the participial reading]). But the participle is superior on internal grounds: The structure of v. 17 almost requires an imperative after ἀλλά (alla), for this gives balance to the clause: “Do not become foolish, but understand…” If the participle is authentic, it may be imperatival (and thus should be translated just like an imperative), but such is quite rare in the NT. More likely, there is an implied imperative as follows: “Do not become foolish, but become wise, understanding what the will of the Lord is.” Either way, the participle is the harder reading and ought to be considered original. It is significant that seeing an implied imperative in this verse affords a certain symmetry to the author’s thought in vv. 15-21: There are three main sections (vv. 15-16, v. 17, vv. 18-21), each of which provides a negative injunction, followed by a positive injunction, followed by a present adverbial participle. If συνίετε (suniete) is original, this symmetry is lost. Thus, even though the external evidence for συνιέντες (sunientes) is not nearly as weighty as for the imperative, both the transcriptional and intrinsic evidence support it.

¹ tc Several mss (P⁴⁶ D² Ψ 1175* 1505 M) have πνεύματος (pneumatos, “Spirit”) instead of φωτός (phōtos, “light”). Although most today regard φωτός as obviously the reading of the initial text (UBS⁵ gives it an “A” rating), a case could be made that πνεύματος is what the author wrote. First, although this is largely a Byzantine reading (D² often, if not normally, assimilates to the Byzantine text), P⁴⁶ gives the reading much greater credibility. Internally, the φωτός at the end of v. 8 could have lined up above the πνεύματος in v. 9 in a scribe’s exemplar, thus occasioning dittography. (It is interesting to note that in both P⁴⁹ and א the two instances of φωτός line up.) However, written in a contracted form, as a nomen sacrum (pMnMs)—a practice found even in the earliest mss—πνεύματος would not have been easily confused with fwtos (there being only the last letter to occasion homoioteleuton rather than the last three). Further, the external evidence for φωτός is quite compelling (P⁴⁹ א A B D* F G P 33 81 1175^c 1739 1881 2464 latt co). It is thus doubtful that the early and widespread witnesses all mistook πνεύματος for φωτός. In addition, πνεύματος can be readily explained as harking back to Gal 5:22 (“the fruit of the Spirit”). Thus, on balance, φωτός appears to be autographic, giving rise to the reading πνεύματος.

² tc The unusual expression διὰ θεοῦ (dia theou, “through God”) certainly prompted scribes to alter it to more customary or theologically acceptable ones such as διὰ θεόν (dia theon, “because of God”; F G 1881), διὰ Χριστοῦ (dia Christou, “through Christ”; 81 630 sa), διὰ ᾿Ιησοῦ Χριστοῦ (dia Iēsou Christou, “through Jesus Christ”; 1739^c), θεοῦ διὰ Χριστοῦ (“[an heir] of God through Christ”; א² C³ D (P) 0278 (6 326) 1175 1241 (1505) 2464 M ar sy), or κληρονόμος μὲν θεοῦ, συγκληρονόμος δὲ Χριστοῦ (klēronomos men theou, sugklēronomos de Christou, “an heir of God, and fellow-heir with Christ”; Ψ [cf. Rom 8:17]). Although it is unusual for Paul to speak of God as an intermediate agent, it is not unprecedented (cf. Gal 1:1; 1 Cor 1:9). Nevertheless, Gal 4:7 is the most direct statement to this effect. Further testimony on behalf of διὰ θεοῦ is to be found in external evidence: The witnesses with this phrase are among the most significant in the NT (P⁴⁶ א* A B C* 33 1739*^vid lat bo Cl).

² tc For προσετέθη (prosetethē) several Western witnesses have ἐτέθη (etethē, “it was established”; so D* F G it Ir^lat Ambst Spec). The net effect of this reading, in conjunction with the largely Western reading of πράξεων (praxeōn) for παραβάσεων (parabaseōn), seems to be a very positive assessment of the law. But there are compelling reasons for rejecting this reading: (1) externally, it is provincial and relatively late; (2) internally: (a) transcriptionally, there seems to be a much higher transcriptional probability that a scribe would try to smooth over Paul’s harsh saying here about the law than vice versa; (b) intrinsically: [1] Paul has already argued that the law came after the promise (vv. 15-18), indicating, more than likely, its temporary nature; [2] the verb “was added” in v. 19 (προσετέθη) is different from the verb in v. 15 (ἐπιδιατάσσεται, epidiatassetai); virtually all exegetes recognize this as an intentional linguistic shift on Paul’s part in order not to contradict his statement in v. 15; [3] the temper of 3:1-4:7 is decidedly against a positive statement about the Torah’s role in Heilsgeschichte.

¹ tc Most witnesses, including some significant ones (P⁴⁶ א² C D² F G Ψ 0209 M lat sy Tert), read ὑμεῖς…ἐστε (humeis…este, “you are”) instead of ἡμεῖς…ἐσμεν (hēmeis…esmen, “we are”) here, but several other early and significant mss (א* B D* L P 0243 6 33 81 326 365 1175 1739 1881 2464 co Cl Or) have ἡμεῖς…ἐσμεν. The external evidence is somewhat in favor of the first person pronoun and verb; the internal evidence weighs in even stronger. In light of the parallel in 1 Cor 3:16, where Paul uses ἐστε (“you are the temple of God”), as well as the surrounding context here in which the second person verb or pronoun is used in vv. 14, 17, and 18, the second person reading seems obviously motivated. The first person reading can explain the rise of the other reading, but the reverse is not as easily done. Consequently, the first person reading of ἡμεῖς…ἐσμεν has all the credentials of authenticity.

² tc There are three viable variant readings at this point in the text. (1) The reading ἡ γυνὴ ἡ ἄγαμος καὶ ἡ παρθένος (hē gunē hē agamos kai hē parthenos, “the unmarried woman and the virgin”) is represented by ancient and important mss, as well as some significant versions (P¹⁵ B 104 365 1505 vg co). (2) The reading ἡ γυνὴ ἡ ἄγαμος καὶ ἡ παρθένος ἡ ἄγαμος (“the unmarried woman and the unmarried virgin”) is also found in ancient and important mss (P⁴⁶ א A 33 1739 1881). (3) The reading ἡ γυνὴ καὶ ἡ παρθένος ἡ ἄγαμος (“the woman and the unmarried virgin”) is found in Western mss (D F G) and the majority of Byzantine cursives. Based upon external evidence, the first and second readings are the strongest; the readings both reach deep into the second century with strong testimony from mss of the Alexandrian group of witnesses. Internal evidence seems equally balanced: Scribes may have wanted to add ἡ ἄγαμος to παρθένος for stylistic reasons, but they might also have wanted to remove it because it sounded redundant. Because Paul’s meaning is not quite clear, a decision on the proper textual reading is difficult. On the whole scribes tended to add to the text, not take from it. Thus the first reading should be favored as earlier, but this decision should be regarded as less than certain.

¹ tc Some significant Alexandrian and Western mss (א B D*^,c 1506 bo) read νῦν (nun, “now”) here. A few other mss (33 365 sa) have ὕστερον (husteron, “finally”). mss that lack the word are P⁴⁶ A D¹ F G Ψ 81 1175 1241 1505 1739 1881 M latt. External evidence slightly favors omission with good representatives from the major text-forms, and because of the alliance of Alexandrian and Byzantine mss (with the Byzantine going against its normal tendency to embrace the longer reading). Internally, scribes could have added νῦν here to give balance to the preceding clause (οὗτοι νῦν ἠπείθησαν…αὐτοὶ νῦν ἐλεηθῶσιν [|outoi nun ēpeithēsan…autoi nun eleēthōsin; “they have now been disobedient…they may now receive mercy”]). However, it seems much more likely that they would have deleted it because of its seeming inappropriateness in this context. That some witnesses have ὕστερον presupposes the presence of νῦν in their ancestors. A decision is difficult, but νῦν is slightly preferred, since it is the more difficult reading and is adequately represented in the mss.

² tc ‡ Several variants occur at this juncture. Some of the earliest and best witnesses (P⁷⁴ א B* C Ψ 33^vid Did^pt) read “Jesus Christ” (᾿Ιησοῦς Χριστός, Iēsous Christos); others ([A] 36 1175 it) have “the Lord Jesus Christ” (ὁ κύριος ᾿Ιησοῦς Χριστός, ho kurios Iēsous Christos); a few read simply ὁ Χριστός (614 1241 1505); the majority of mss (B² E 1739 M Did^pt) have “Jesus the Christ” ( ᾿Ιησοῦς ὁ Χριστός). Although the pedigree of this last reading is relatively weak, it draws strength from the fact that (a) the other readings are much more natural and thus more predictable, and (b) there are several variants for this text. It seems hardly likely that scribes would intentionally change a more common expression into a title that is used nowhere else in the NT (although 1 John 2:22; 5:1 come close with “Jesus is the Christ”), nor would they unintentionally change a frequently used designation into an unusual one. Thus, in spite of the external evidence (which is nevertheless sufficient to argue for authenticity), ᾿Ιησοῦς ὁ Χριστός is the reading that best explains the rise of the others.

⁶ tc Some mss read θεῷ (theō, “God”) here, a variant much easier to understand in the context. The reading “God” is supported by א² A C E Ψ 33 1739 M lat sy co. The more difficult οἴκῳ (oikō, “house”) is supported by P⁷⁴ א* B D H 049. Thus the second reading is preferred both externally because of better ms evidence and internally because it is hard to see how a copyist finding the reading “God” would change it to “house,” while it is easy to see how (given the LXX of Ps 132:5) a copyist might assimilate the reading and change “house” to “God.” However, some scholars think the reading “house” is so difficult as to be unacceptable. Others (like Lachmann and Hort) resorted to conjectural emendation at this point. Others (Ropes) sought an answer in an underlying Aramaic expression. Not everyone thinks the reading “house” is too difficult to be accepted as original (see Lake and Cadbury). A. F. J. Klijn, “Stephen’s Speech—Acts vii.2-53, ” NTS 4 (1957): 25-31, compared the idea of a “house within the house of Israel” with the Manual of Discipline from Qumran, a possible parallel that seems to support the reading “house” as authentic. (For the more detailed discussion from which this note was derived, see TCGNT 308-9.)

¹ tc It is difficult to decide between ἐζήτουν οὖν (ezētoun oun, “then they were seeking”; P⁶⁶ א A L W Ψ ƒ^1,13 33 pm lat), ἐζήτουν δέ (ezētoun de, “now they were seeking”; P⁴⁵ and a few versional witnesses), καὶ ἐζήτουν (kai ezētoun, “and they were seeking”; D), and ἐζήτουν (P^75vid B Γ Θ 700 pm). Externally, the most viable readings are ἐζήτουν οὖν and ἐζήτουν. Transcriptionally, the οὖν could have dropped out via haplography since the verb ends in the same three letters. On the other hand, it is difficult to explain the readings with δέ or καί if ἐζήτουν οὖν is autographic; such readings would more likely have arisen from the simple ἐζήτουν. Intrinsically, John is fond of οὖν, using it some 200 times. Further, this Gospel begins relatively few sentences without some conjunction. The minimal support for the δέ and καί readings suggests that they arose either from the lone verb reading (which would thus be prior to their respective Vorlagen but not necessarily the earliest reading) or through carelessness on the part of the scribes. Indeed, the ancestors of P⁴⁵ and D may have committed haplography, leaving later scribes in the chain to guess at the conjunction needed. In sum, the best reading appears to be ἐζήτουν οὖν.

⁴ tc A few significant witnesses lack ὑμῶν (humōn, “your”) here (P^66,75 B L W 070), while the majority have the pronoun (א C D Θ Ψ 0250 ƒ^1,13 33 565 892 M al lat sy). However, these mss do not agree on the placement of the pronoun: τοῦ πατρὸς ὑμῶν ποιεῖτε (tou patros humōn poieite), τῷ πατρὶ ὑμῶν (tō patri humōn), and τῷ πατρὶ ὑμῶν ταῦτα (tō patri humōn tauta) all occur. If the pronoun is read, then the devil is in view and the text should be translated as “you are practicing the things you have heard from your father.” If it is not read, then the same Father mentioned in the first part of the verse is in view. In this case, ποιεῖτε should be taken as an imperative: “you [must] practice the things you have heard from the Father.” The omission is decidedly the harder reading, both because the contrast between God and the devil is now delayed until v. 41, and because ποιεῖτε could be read as an indicative, especially since the two clauses are joined by καί (kai, “and”). Thus, the pronoun looks to be a motivated reading. In light of the better external and internal evidence the omission is preferred.

⁵ tc There is a seeming myriad of variants for this text. Many mss read εἰς τόπον ἔρημον (eis topon erēmon, “to a deserted place”; א*^,2 [1241]) or εἰς τόπον ἔρημον πόλεως καλουμένης Βηθσαϊδά (eis topon erēmon poleōs kaloumenēs Bēthsaida, “to a deserted place of a town called Bethsaida”; [A] C W Ξ^mg [ƒ^1,13] [565] M) here, while others have εἰς κώμην λεγομένην Βηδσαϊδά (eis kōmēn legomenēn Bēdsaida, “to a village called Bedsaida”; D), εἰς κώμην καλουμένην Βηθσαϊδά εἰς τόπον ἔρημον (eis kōmēn kaloumenēn Bēthsaida eis topon erēmon, “to a village called Bethsaida to a deserted place”; Θ), or εἰς τόπον καλουμένον Βηθσαϊδά (eis topon kaloumenon Bēthsaida, “to a place called Bethsaida”; Ψ). The Greek behind the translation (εἰς πόλιν καλουμένην Βηθσαϊδά, eis polin kaloumenēn Bēthsaida) is supported by (P⁷⁵) א¹ B L Ξ* 33 2542 co. The variants can be grouped generally into those that speak of a “deserted place” and those that speak of a place/city/town called Bethsaida. The Byzantine reading is evidently a conflation of the earlier texts, and should be dismissed as secondary. The variants that speak of a deserted place are an assimilation to Mark 6:32, as well a harmonization with v. 12, and should also be regarded as secondary. The reading that best explains the rise of the others—both internally and externally—is the one that stands behind the translation and is found in the text of NA²⁸.

¹ tc ‡ A difficult textual problem is found here, involving three different variants: καὶ λέγουσιν (kai legousin) is found in א; οἱ δὲ εἶπον (hoi de eipon) is the reading of P⁴⁵ A D W Θ ƒ^1,13 33 M it; and καὶ λέγουσιν αὐτῷ (kai legousin autō) is supported by B C L (Δ 579 892). The first two variants would not be translated differently; the third reading, however, would add “to him” after “they replied.” What complicates the issue is that the external evidence is fairly evenly split between the second and third readings, though the first reading is in agreement with the second reading in lacking the dative pronoun. Indeed, another layout of the problem here could treat this as two distinct problems: καὶ λέγουσιν vs. οἱ δὲ εἶπον and αὐτῷ vs. omission of the word. In this second arrangement of the problem, the reading without the pronoun has slightly stronger support (P⁴⁵ א A D W Θ ƒ^1,13 33 M it). Internally, Mark never elsewhere uses the form εἶπον for the third person plural indicative form of this verb (it is always εἶπαν [eipan]). And although only one other time in Mark is the object lacking after λέγουσιν (6:38), it is a similar context (viz., the disciples’ response before Jesus feeds the 5000). Very tentatively, the reading that is followed here is καὶ λέγουσιν. NA²⁸ puts αὐτῷ in brackets, indicating some doubt as to its authenticity.

¹ tc The translation here follows the reading στήσητε (stēsēte, “set up”) found in D W Θ ƒ¹ 28 565 2542 it sy^s,p Cyp. The majority of mss here read τηρήσητε (tērēsete; א A L ƒ¹³ 33 M co) or τηρῆτε (tērēte; B), both translated “keep.” It is hard to know which reading is best: On the one hand, τηρήσητε/τηρῆτε has much stronger external support, but στήσητε is a more difficult reading. What makes “keep” suspect is that it appears in two different forms, suggesting independent alterations of a difficult reading. Further, scribes may have been influenced by the preceding “commandment of God” to change the text toward “keep” (TCGNT 81), a common enough expression (cf. Matt 19:17; John 14:15; 1 Tim 6:1; 1 John 5:3; Rev 14:12). Thus, the more difficult reading is “set up.” Also, the more natural opposite of “reject” (ἀθεῖτε [atheite], literally “you set aside”) is “set up.” However, the Western reading may have been influenced by Exod 6:4 or Heb 10:9, but this likelihood seems remote. Thus, “set up” is more likely to be the earlier reading here.

¹ tc ‡ The witnesses here support several different readings: αὐτῇ πολλά (autē polla, “to her insistently”) is found in D Θ 565 700 it; πολλά is the reading of P^45vid 28; both words are lacking in L; and א A B C^2vid ƒ¹³ 33 M lat have just αὐτῇ. The best candidates for authenticity, on external grounds, are αὐτῇ πολλά and αὐτῇ. So the issue revolves around whether πολλά is part of the text. On the one hand, πολλά used adverbially is a distinctive Markanism (10 of the 16 NT instances are found in Mark; of the other Gospels, Matthew alone adds a single example [Matt 9:14]). It could be argued that such an unremarkable term would go unnoticed by the scribes, and consequently would not have been inserted in imitation of Mark’s style observed elsewhere. On the other hand, the largest cluster of instances of an adverbial πολλά are in Mark 5-6, with the most recent example coming just three verses earlier (Mark 5:23, 38, 43; 6:20). Scribes may well have imitated the usage so recently and so frequently seen. Further, the best Alexandrian witnesses, as well as good representatives of the Western and Byzantines texts, lack πολλά. On the whole, though a decision is difficult, it is probably best to read the text without πολλά. NA²⁸ places the word in brackets, indicating some doubt as to its authenticity.

² tc The phrase “whom he named apostles” is lacking in the majority of mss (A C² D L ƒ¹ 33 565 579 1241 1424 M latt sy; SBL). Several primary Alexandrian and other key witnesses (א B C* W Δ Θ ƒ¹³ 28 co) include the phrase, rendering the external evidence strongly in favor of this reading. It is possible that the Alexandrian witnesses have inserted these words to bring the text in line with Luke 6:13 (TCGNT 69), but against this is the internal evidence of Mark’s style: Mark tends toward gratuitous redundancy. However, significant Western and Byzantine mss along with other authorities lack the clause, which is against the Byzantine tendencies. The key issue, however, is that both the wording of the clause and its location in the verse varies significantly among the witnesses, which suggests that it was indeed borrowed from the Lukan parallel. The NA²⁸ puts the words in brackets indicating doubts about their authenticity.

¹ tc The phrase “he appointed 12” at the beginning of v. 16 is lacking in the majority of mss (A C² D L Θ ƒ¹ 33 700 1241 1424 M lat sy bo), including several key witnesses. Some significant authorities include the phrase (א B C* Δ 565 579 pc). The omission may have been caused by haplography in combination with homoioarcton: The first word of the clause in question is καί (kai), and the first word after the clause in question is also καί. And the first two letters of the second word, in each instance, are επ (ep). Early scribes may have jumped accidentally from the first καί to the second, omitting the intervening material. Metzger suggests that “the clause seems to be needed in order to pick up the thread of ver. 14 after the parenthesis ἵνα…δαιμόνια” (TCGNT 69). This seems to be a stretch. Further, the external evidence in favor of the words is not as compelling as it could be (the addition of “whom he named apostles” in 3:14 actually has stronger evidence, yet we considered it spurious). A decision is difficult but the shorter reading is preferred. NA²⁸ puts the words in brackets.

² tc Most witnesses, including some significant ones (A D W Γ Δ 28^c 700 1241 1424 M lat sy), have τῆς βασιλείας (tēs basileias) between τὸ εὐαγγέλιον (to euangelion) and τοῦ θεοῦ (tou theou): “the gospel of the kingdom of God.” On the one hand, it is perhaps possible that τῆς βασιλείας was omitted to conform the expression to that which is found in the epistles (cf. Rom 1:1; 15:16; 2 Cor 11:7; 1 Thess 2:2, 8, 9; 1 Pet 4:17). On the other hand, this expression, “the gospel of God,” occurs nowhere else in the Gospels, while “the gospel of the kingdom” is a Matthean expression (Matt 4:23; 9:35; 24:14), and “kingdom of God” is pervasive in the synoptic Gospels (occurring over 50 times). Scribes would thus be more prone to add τῆς βασιλείας than to omit it. Further, the external support for the shorter reading (א B L Θ ƒ^1,13 28* 33 565 579 892 sa Or) is significantly stronger than that for the longer reading. There is little doubt, therefore, that the shorter reading is authentic.

¹ tc Although the external evidence for the inclusion of “Jesus” before “Barabbas” (in vv. 16 and 17) is rather sparse, being restricted virtually to mss of what was formally labeled the “Caesarean” text (Θ ƒ¹ 700* sy^s arm geo²; Or^mss), the omission of the Lord’s name in apposition to “Barabbas” is such a strongly motivated reading that it can hardly be original. There is no good explanation for a scribe unintentionally adding ᾿Ιησοῦν (Iēsoun) before Βαραββᾶν (Barabban), especially since Barabbas is mentioned first in each verse (thus dittography is ruled out). Further, the addition of τὸν λεγόμενον Χριστόν (ton legomenon Christon, “who is called Christ”) to ᾿Ιησοῦν in v. 17 makes better sense if Barabbas is also called “Jesus” (otherwise, a mere “Jesus” would have been a sufficient appellation to distinguish the two). Metzger notes that codex S, a tenth-century majuscule, along with a score of minuscules, have a marginal comment on this verse as follows: “In many ancient copies which I have met with I found Barabbas himself likewise called ‘Jesus.’” The attribution of this scholium is variously given as Anastasius, Chrysostom, or even Origen (TCGNT 56).