The textual problem in 1 Th 2:7 of nhpioi or hpioi("babes" or "gentle") is one of the more intriguing textual problems in the Thessalonian letters. Most English translations go with hpioi, including the NASB, ESV, NRSV, NIV, and KJV. Of the English versions to which I have access, only the NET, TNIV, and, interestingly, the Message (based on my interpretation of it) choose nhpioi. The NA²⁷/UBS text reads nhpioi, while the Majority Text and the TR both have hpioi.

The External Evidence[1]

Reading #1: nhpioi P⁶⁵ 01* B C* D* F* G I Y* 104* 326^c it vg^cl.ww sa^ms bo

Reading #2: hpioi 01^c A C² D² K L P Y^c 0278 33 81 365 630 1505 1739 1881 Byz vg^st (sy) sa^mss Cl

The external evidence seems firmly in favor of reading 1, nhpioi. It enjoys the support of the early Greek MSS; of the 7 extant MSS from the first eight centuries which contain this verse, only A, from the fifth century, reads hpioi. The other six read nhpioi, including three MSS that are much earlier (P⁶⁵ is third century, while 01* and B are fourth century) and exhibit a higher textual quality than A. Admittedly, P⁶⁵ is fragmentary in 1 Thessalonians, but it has nonetheless been judged to be extremely reliable (Aland tentatively concludes that it has a "strict text" and ranks it in his category 1 [Aland-Aland, Text, 100]), while 01* and B are primary Alexandrian and are two of the most respected MSS in the entire tradition; codex A on the other hand is categorized as secondary Alexandrian, which is somewhat less reliable. Perhaps not coincidentally, hpioi appears in a number of other secondary Alexandrian witnesses and one later primary Alexandrian witness (33, 81, 1506, 1881, and 1739). Based on this evidence, hpioi must have entered the Alexandrian text-type somewhere between the primary and secondary Alexandrian MSS (perhaps through Clement of Alexandria or through a simple scribal error). It is found in the Byzantine MSS, but it is uncertain whether it entered the Byzantine tradition from the secondary Alexandrian text, or if it entered the secondary Alexandrian text due to a Byzantine influence; however, it is commonly known that the secondary Alexandrian texts are marked by a Byzantine influence, which seems to make the latter more likely than the former, and would explain the existence of hpioi in the correctors of 01, C and D (which typically follow a Byzantine Vorlage). In any case, nhpioi certainly appears in earlier and more reliable MSS (third-fourth century, early primary Alexandrian texts).

However, due to genealogical solidarity within both the Alexandrian and Western text-types, the date of reading 1 can be pushed even further back than the third century. The consensus of the early Alexandrian MSS (P⁶⁵ 01* B) implies that the reading can be dated back to the archetype of the Alexandrian tradition, which probably originated in the second century. This is because the reading's presence in the archetype is the likely reason for its presence in the different streams of primary Alexandrian text. Furthermore, the reading's second-century date is doubly reinforced by the consensus of the Western text-type; if the reading is present in the Western Greek uncials (D* F G) as well as our early Latin versions (it), then it probably existed in the archetype of the Western text-type which also has been traced back to the second century. As for reading 2, hpioi appears in the mass of Byzantine MSS, so its date can be pushed back only as early as the fourth century with the Byzantine archetype. Therefore, based on genealogical solidarity, nhpioi can be dated as far back as the second century, while hpioi can only be dated to the fourth century.

Finally, reading 1 is not only demonstrably earlier-it is also more geographically widespread. The agreement between the Alexandrian and Western text-types shows that at an early date the reading existed both in North Africa and in the Western areas of the Roman Empire. The reading's penetration of the various geographical areas of very early Christianity is yet another argument for its primacy.

Due to the superior date, character, genealogical solidarity, and geographical diversity of the MSS in support of reading 1, the external evidence strongly favors nhpioi as the original reading.

The Internal Evidence

Reading #1: nhpioi EGENHQHMENNHPIOI

Reading #2: hpioi EGENHQHMENHPIOI

The pronunciation of these two readings is virtually indistinguishable, since the N at the end of the first word might be pronounced together with the N of the second word. But this does not necessarily imply that our erring scribe copied by hearing the words of his exemplar read aloud by another scribe, for he may have merely read his exemplar and pronounced the words internally to himself. In any case, when read at a normal pace, the two readings sound very similar. In addition, the visual similarity of the uncial letters N and H could cause an error of faulty eyesight. Therefore, if the variants are the result of purely accidental transcriptional error (perhaps even dating all the way back to Paul's amanuensis), it is extremely difficult to determine whether the accidental error was haplography or dittography; a scribe reading and/or pronouncing the words to himself while copying could just as easily have removed the N on accident as he could have added it.

But if this is not purely a case of accidental, transcriptional error, then it is likely that our scribe was influenced by other stylistic and contextual considerations. For example, in Paul, nhpio" occurs far more frequently than hpio";Paul uses nhpio" 10 other times (NA²⁷: Rom 2:20, 1 Cor 3:1, 1 Cor 13:11 [5 times], Gal 4:1, Gal 4:3, and Eph 4:14), and he uses hpio" only one other time (NA²⁷: 2 Tim 2:24). Clearly, nhpio" is the word that Paul uses more regularly. But does this fact support a Pauline or a scribal origin of the word here? On one hand, if Paul uses nhpio" 10 times more frequently than he uses hpio", one could simplistically argue that Paul was therefore more likely to have used nhpio" here than hpio". But on the other hand it has also been argued that, because scribes were more familiar with Paul's more frequent use of nhpio", our erring scribe could have changed the less familiar hpio" to the more familiar nhpio".

However, Timothy Sailors has shown that in Greek literature of the first and second century, the use of hpio" is not so uncommon that a scribe would have considered it an unfamiliar word (Sailors, "Wedding Textual and Rhetorical Criticism," 86-7). In fact, the ratio of nhpio" to hpio" falls in the second century to as low as 5:1. In light of this evidence, it seems unlikely that a scribe would change hpio" to nhpio" on account of his lack of familiarity with the word.

It is also argued that Paul would not use nhpio" here because (1) he only uses it to refer negatively to babies, new converts, or the spiritually immature, (Delobel, "One Letter Too Many," 128) and (2) he never uses nhpio" to refer to himself (Metzger, Text, 232). These facts seem to imply that Paul was more likely to have used hpio" rather than nhpio" to refer to himself in 1 Thess 2:7; to some (like Metzger), the likelihood is greater that he chose to use hpio" rather than vary his usage of the nhpio".

There are several problems with this line of reasoning. First, even if these two statements were true, it would not necessarily follow that nhpio" is therefore the inferior reading. For example, if Paul never uses nhpio" to describe himself, a scribe is highly unlikely to have consciously created a situation in which he does. Scribes were sometimes sleepy and sometimes stupid, but it seems rather unlikely that a scribe would assign a term to Paul that Paul never once uses to describe himself-especially a pejorative term that might disrespect Paul's nearly unsurpassed position of authority. Unless this proposed error was committed by a scribe unusually uneducated in Pauline style, it probably did not have a scribal origin. Thus, to the question, "Who varied Paul's style in the usage of nhpio"?" we must answer, "it was unlikely that it was Paul, but more unlikely a scribe." It would be much more likely that Paul would vary his style here than a scribe vary it for him.

Second, these two assertions are actually inaccurate. Paul's use of nhpio" in 1 Cor 13:11 (which accounts for half of the instances in Paul) is not necessarily pejorative; it might be argued that Paul is using nhpio" in a neutral, descriptive sense within a simple before-and-after comparison. Furthermore, recent studies have shown that writers in Paul's time would not necessarily have assigned a negative connotation to nhpio"; of all the uses of nhpio" during the first century B.C. and the first century A.D., 75% are neutral and only about 20% are pejorative (Sailors, "Wedding Textual and Rhetorical Criticism," 91-2). Thus, one cannot cite an assumed negative connotation of nhpio" as strong internal evidence for hpio" as the preferred reading.

Next, the assumption that Paul never uses nhpio" to refer to himself is exaggerated, because Paul himself is the referent of all five uses in 1 Cor 13:11. Admittedly, Paul is describing himself as he was in his youth, but to say that he would not use nhpio" in 1 Thess 2:7 since he never uses nhpio" to describe himself is at least somewhat inaccurate. In fact, it might even be more accurate to conclude that, since Paul does refer to himself in 50% of his uses of nhpio", it is entirely appropriate that he might use it to refer to himself here in 1 Thess 2:7 as well!

Finally, in addition to the analysis of Paul's normal use of words, there are a number of important contextual and grammatical considerations. These issues are complicated, and are too numerous to evaluate with depth in this paper; however, the primary considerations involved are the nature of the shift in metaphor and the grammatical structure of the passage.

Metzger asserts that nhpioi is so difficult that it violates common sense due to the resulting violent shift in metaphor (Paul is a babe, yet he also is a mother-nurse). But Paul's sudden shift in metaphor is not without precedent; Gal 4:19 also contains a similar shift in which Paul is "in labor" with the recipients until Christ is formed within the wombs of the recipients themselves!

However, it is not even necessary to assume that these two images in 1 Thess 2:7 are part of the same metaphor. It is not insignificant that the entire phrase alla egenhqhmen [n]hpioi en mesw umwn can be linked with the previous clauses, which results in a separation of the two metaphors into two distinctly different thoughts: (1) Paul and friends had the power to use their authority as apostles of Christ, but instead they behaved as spiritual babes or new converts (vv. 6-7b), and (2) like a mother-nurse Paul and friends cared for them and were happy to share the gospel and their lives with them (vv. 7c-8). With this re-punctuation, the full stop would be moved forward, linking alla egenhqhmen [n]hpioi en mesw umwn of 1 Th 2:7b to the preceding and not the following clauses. This separation certainly eases the violent shift in metaphor, and is often overlooked (see Metzger's Textual Commentary, 562, in which the change in punctuation appears only as an afterthought). The term nhpioi is then placed in contrast with apostoloi, to emphasize that Paul's behavior was far different than that which his authority afforded him. Apostles were the leaders of the church, and as such they could have wielded their authority or sought glory like other traveling orators; on the other hand, new converts (or spiritual "babes") did not have authority and did not seek their own glory, but instead were humble and innocent in their behavior. Likewise, Paul did not use his authority as an apostle, but was instead like a spiritual babe in his innocent behavior. This use of the term nhpioi likely spurred his use of the next, now-separated metaphor of the mother-nurse. Thus, 1 Thess 2:7b-c may not contain a single mixed metaphor, but the end of one metaphor (and sentence!) and the beginning of another.

Jeffrey Weima points out three grammatical arguments in favor of this re-punctuation: (1) Paul's use of alla following a negative (as in 1 Thess 2:6-7b) typically introduces the second portion of an ou_alla contrast, (2) Paul includes five such ou_alla contrasts throughout 1 Thess 2:1-8, and (3) Paul's use of the w"_ outw"combination (as in 1 Thess 2:7c-8) typically indicates a new correlative clause (Weima, "The Case for Nhpioi in 1 Thess 2.7," 555-556). These facts are strongly in favor of the re-punctuation of this particular text, but they do not inherently support either reading over the other, since both readings might actually fit well with the re-punctuation; rather, they show that nhpioi is not so difficult as to be lectio impossibilis, especially when taking the proper punctuation into account.

After establishing that reading 1 is far less unreasonable than Metzger asserts, it becomes obvious that nhpioi is the preferred choice because it is the harder reading: (1) it creates a different variation in Paul's use of nhpio", (2) it creates a violent shift in metaphors, and (3) the violent shift is not impossibly violent, if accompanied by the proper re-punctuation. Reading 2, on the other hand, is the easier reading and is relatively free of difficulty, except for the very surmountable fact that hpio" is only used once elsewhere in Paul.

Based on the transcriptional and intrinsic probabilities (which are neutral and in favor of reading 1, respectively), the internal evidence falls more in favor of nhpioi as the original reading.

Conclusion

Most scholars recognize that the external evidence is strongly in favor of nhpioi as the original reading; most scholars also recognize nhpioi as the more difficult reading. However, many have concluded that nhpioi is too difficult, and therefore also concluded that hpioi is the original reading on the basis of over-simplistic arguments. But analysis of the stylistic, lexical, contextual and grammatical evidence shows that nhpioi is indeed not too difficult, and is in fact the preferred reading. The internal evidence, although less overwhelming, agrees with the external evidence in its support of nhpioi as the original.

Chart of Manuscript Evidence[2]

Reading #1: nhpioi P⁶⁵ 01* B C* D* F* G I Y* 0150 5 38 61 69 102 103 104* {122*}[3] 131 142 206 209^c 263 309 326^c 393 421 425 429 451 459 460 {460^c}[4] {491}[5] 582 620 623 886 941 1101* 1102 1115 1311 1398 1409 1524^c 1646 1718 (1729) 1798^c 1830 1838 1890* 1904 {1906}[6] {1912}[7] 1914 1918 1921 1922 1943 1951 1952 1962 1991 2002 2003 2086 2104 2482 2495 2502 2516 2576 2659^c {it vg^cl.ww sa^ms bo}[8]

Reading #2:hpioi01^c A C² D² K L P Y^c 0278 33 81 365 630 1505 1739 1881 Byz {vg^st (sy) sa^mss Cl}[10]

Bibliography

Aland, Kurt, et al. Kurzgefasste Liste der griechischen Handschriften des Neuen Testaments. 2d ed., revised and enlarged. ANTF 1. Berlin: de Gruyter, 1994.

Aland, Kurt and Barbara. The Text of the New Testament: An Introduction to the Critical Editions and to the Theory of Practice of Modern Textual Criticism. Trans. by Erroll F. Rhodes. 2d ed. Grand Rapids, Michigan: Eerdmans, 1989.

Aland, Kurt, et al., eds. Text und Textwert der griechischen Handschriften des Neuen Testaments. II: Die Paulinischen Briefe. Vol. 4. ANTF 19. Berlin and New York: de Gruyter, 1991.

Delobel, Joel. "One Letter Too Many in Paul's First Letter? A Study of (n)hpioi in 1 Thess 2.7," Louvain Studies 20 (1995): 127-128.

Holmes, Michael W. "New Testament Textual Criticism." In Introducing New Testament Interpretation, ed. Scot McKnight, 53-74. Grand Rapids: Baker, 1989.

Metzger, Bruce M. The Text of the New Testament: Its Transmission, Corruption, and Restoration. 3d ed. New York: Oxford University Press, 1992.

Metzger, Bruce M. A Textual Commentary on the Greek New Testament. 2d ed. Stuttgart: Deutsche Bibelgesellschaft/German Bible Society, 1994.

Sailors, Timothy B. "Wedding Textual and Rhetorical Criticism to Understand the Text of 1 Thessalonians 2:7," JSNT 80 (2000): 81-98.

Weima, Jeffrey A. D. "'But We Became Infants Among You': The Case for NHPIOI in 1 Thess 2.7," NTS 46 (2000): 547-564.

[1] For a more complete list of MS evidence, see the Chart of Manuscript Evidence below.

[2] MS support is summarized from Aland, Text und Textwert, 301-303, unless otherwise indicated by brackets {} and explanatory footnotes; MS dates and text-types are found in Aland, The Text of the New Testament, 107-142, with a few updates from Holmes, "New Testament Textual Criticism," 59.

[3] Found in Tischendorf; interestingly, this MS is listed in support of hpioi in Aland, Text und Textwert, 301.

[4] Both Tischendorf and von Soden list the correction, 460^c.

[5] Found in von Soden.

[6] Found in Tischendorf.

[7] Found in von Soden.

[8] Found in NA²⁷.

[9] The following MSS are not classified as any particular text-type in Aland-Aland, The Text of the New Testament: 38 102 131 421 460 941 1102 1115 1311 1646 1729 1798 1830 1890 1904 1943 1991 2002 2086 2104 2482 2576 2659.

[10] Found in NA²⁷.