Bruce M Metzger, the canon of scripture and the denial of 2 Peter 3.16:

A focus on Clement of Rome's epistle.

   Home                                    Scanned Sections p41-44 from Prof Metzger’s book                Scanned Text of Clement’s Epistle in Greek                   Hoole’s translation of Clement                                                                                                         (Without permission under 'fair use')                                           Taken from this site with thanks                                                          Taken from this site with thanks
                                                 Scanned Sections p271-274 from Prof Metzger’s book           

“Even as our beloved brother Paul also according to the wisdom given unto him hath written unto you; 16  As also in all his epistles, speaking in them of these things; in which are some things hard to be understood, which they that are unlearned and unstable wrest, as they do also the other scriptures, unto their own destruction.” (2 Peter 3.15b,16)

2 Peter is the Apostle Peter’s last warning Epistle of the dangers of apostasy, in particular of false teachers and of mockers. In chapter 3 and verse 16 he describes the character of Paul’s Epistles and of the danger of those who wrest their meaning and significance. To those who acknowledge the Divine character of the New Testament writings here is an impregnable assertion that New Testament scripture was recognised and acknowledged as scripture from the time of its receipt. The NT uses the noun grafh or writing 51 times, as here, its unique reference is to Holy Scripture. 

An image of Professor Metzger from his autobiography.

Bruce Metzger was professor emeritus at Princeton Theological Seminary, and a board member of the American Bible Society. He was chief editor of the Reader's Digest Condensed Bible, and according to its preface approved each section. He was the general editor of the New Revised Standard, and one of the editors of the United Bible Society's standard Greek New Testament.

 Without doubt he was a scholar of considerable breadth of learning and erudition. His 1987 work ‘The Canon of the New Testament’ is often cited by evangelicals as a valuable reference work. Metzger’s view of the New Testament is plain from his first sentence. ‘The recognition of the canonical status of the several books of the New Testament was the result of a long and gradual process, in the course of which certain writings, regarded as authoritative were separated from much larger body of early Christian literature’[1].

Clement’s Epistle is the first piece of historical data Metzger’s book examines. It is of singular importance to his thesis, being the first early church writing after the apostles.

Metzger claims that the Pauline epistles cited in Clement are regarded as, ‘filled with good counsel’, however, ‘he does not present them as invested with divine authority’. On the basis of a rapid transition from a quote in the Septuagint to a paraphrase of Romans 1.29-32, Metzger concludes, ‘for Clement the Pauline Epistles were not Scripture, though he obviously regards them as possessing a certain kind of authority.’ He concludes his section, that though the New Testament epistles hold considerable significance, Clement ‘never refers to them as authoritative “Scripture”.’[2]

Clement’ s References to the OT.

His first line of evidence is the difference in Clement’s citations of the Old and New Testaments. Quotes from the Septuagint (LXX) are ‘for the most part made with great exactness’ he claims[3], although this is softened slightly later, ‘more or less exactness’2.

Distinguishing between allusions and citations and counting allusions that may be incidental is prone to some subjective judgement. The criteria used here are set out below in the endnote. In brief, a citation must include a reference indicating it as a quote, or be lengthy and close to verbatim, an allusion does not.[4] There are 65 citations from LXX[5], 47 are directly ascribed to Divine authority, 5 to spokesmen who in context would be regarded as inspired (Job, David, Adam and Moses), and the remaining 13 are introduced simply by narrative verbs (he says, he witnessed). There are at least 41 other allusions to LXX, which do not classify as citations. 

One basic question is the standard of reference by which textual accuracy is determined. The degree of textual variation in the Septuagint is not clear, and it seems likely that slightly differing texts were available throughout the Roman Empire – but it also seems unlikely that these variations were substantial – given the similarity of the surviving texts of LXX, or of Origen’s subsequent comparison of LXX with the 2^nd century retranslations of Aquila, Symmachus and Theodotion in the Hexapla, named after its six parallel columns of text. For the purpose of this study, I have assumed that Clement had available a version of the LXX identical to the Codex Vaticanus version, derived from the Hexapla, although the exercise might profitably be repeated with any other surviving LXX text.

There are many modestly precise quotes of LXX, in 27.7 for example Clement cites Psalm 19.1-3. Apart from the addition of the definite article to the first half of v.2, kai to second and the reversal of ‘language’ and ‘speech’ (both plural in LXX) in v.3, it is exact. Or in 35.7-12 with confidence he cites Ps. 50.16-23[6], introducing it with ‘therefore scripture says’ (legei gar h grafh). Even here his precision is moderate. He substitutes epi for dia in v.16, the noun anomia apparently in an adverbial sense instead of the vocative case of its adjective in v.21, omits se in v.21, and ws lewn in v.22.

However most of Clement’s citations of LXX contain extensive paraphrase and some can only be identified with difficulty. So for example in Clement’s first citation in 3.1, with the words ‘that which is written’(to gegrammenon) a clause of 8 words in LXX at Deut.32.15, by the omission of two words and the addition of 5 becomes 11 words. Similarly in his citation of Jer.9.23-24 in 13.1, introduced with the strongest ‘that which is written, for the Holy Spirit says’ (to gegrammenon, legei gar to pneuma to agion) there is a high degree of imprecision. The 42 words of the quote from the point where he starts, in Clement’s citation become 28, the word order of the last clause of the quote changed substantially, and the last clause of his sentence is a complete deviation from LXX. Interestingly his quote of Jeremiah at the crucial focus ‘let him that boasts boast in the Lord’ is much closer to the New Testaments citations of Jeremiah than it is to LXX. These are found in 1 Cor. 1.31, where the five word phrase is identical and 2 Cor.10.17, to which again Clement’s word order is closer than to LXX. This is a theme to which we’ll return.

8.3 demonstrates another remarkable piece of collation, even though Clement attributes the quote to ‘an excellent saying’. It appears to tightly spin many different threads in one small cloth, starting with a paraphrased citation of Ezek. 33.11, but distinct allusions to Isa. 46.8, Isa. 1.18 , Ps 103.11, Isa 50.3, 2 Chron. 6.38, Isa.64.8 and 62.12 can all be drawn.

Would Metzger argue this is an example of ‘more’ or of ‘less exactness’? Clement’s aim is not verbatim precision, but to represent the distilled essence of argument. 

The passages Clement cites in 29.3 seem a rough amalgam of Num. 18.27 and 2 Chron. 31.14. His citation by ‘for it is written’ (gegraptai gar) in 50.4 of Isa.26.20 is also loose paraphrase, and other examples could be multiplied. Just how removed these are from the original can be judged quite well by the reader in an English translation of LXX and Hoole’s of Clement, the comparison of the texts below (see panel 1). On five occasions Clement tightly weaves together a tapestry of citations (18.1, 26.2, 39.3-9 as well as 8.3 and 29.3 above). In 39.3 the natural sequence of the texts is interrupted by an insertion, Job 4.16-18 is parted from Job 5.1-5 by Job 15.15.

 
Panel 1 Some of Clement’s less precise OT citations  
Of course this is not in the least to denigrate from Clement’s knowledge of Scripture, quite the contrary. His practice of paraphrasing is common amongst early church writers, and is sanctioned by inspiration in the NT. He was well aware that LXX was a Greek translation of Hebrew text, and would have felt as much at liberty to paraphrase as any modern preacher or writer. Nevertheless this examination of how precise the quotations are and the many unobtrusive LXX allusions sets the context for an examination of Metzger’s further claims about the NT. 

Clement’s References to the Messiah’s words. Metzger cites the two passages where Clement recalls the Lord Jesus’ words, 13.2 and 46.8. The citation is not preceded by reference to a written record, and contains words interwoven from different Gospel records, and ideas inserted not directly reported in our Gospels. From these and subsequent passages of a similar nature in two other subsequent writers, Metzger claims, ‘it is not certain…if he [Clement] had written copies, that he felt impelled to quote exactly’[9]. Again he concludes, ‘it seems he has a written record in mind, but does not call it a gospel’ and that Clement, ‘does not appear to enquire how their [the record of the Saviour’s words] authenticity is ensured’². 

Does the evidence warrant these far reaching assertions, however tentatively framed?

First there is the question of dating of Clement, if as a few writers including J. Robinson claim[10], the epistle was written at or before the destruction of the temple – there may indeed have been few if any available Gospel manuscripts in Rome, indeed John’s Gospel was still unborn. Some support for this view derives from Clement’s description of Peter and Paul’s martyrdom in 5.4-7as ‘noble examples of our own generation’, his putative identification as Clement in Phil. 4.3, and the description of an apparently still functioning temple 41.2-3. This is currently not the majority view and such an early date seems doubtful. 

The first citation of the Saviour in13.2 contains words and ideas derived from Mt 5.7, 6.14, 7.2a, 7.12, Lk. 6.36, Mt 7.2b in order. The word ‘be kindly’ (Crhsteuesqe) is not from our Gospels, nor is the extended idea of reciprocation for other acts than forgiving or judging. Nevertheless as we have seen such paraphrases are common practice for Clement, and so are his extrapolations, which may simply be drawn from Mt.6.14 and 7.2. When it is plain he did not ‘feel impelled’ to quote exactly from a text which we know he must have had in his hands (LXX) even when he is quoting it, why must we therefore assume that he had no Gospel account? It is a wholly unwarranted inference. So is the suggestion that Clement is unconcerned about the preservation of the authenticity of the Lord’s words, especially given the intense contemporary efforts to forge documents, mislead and disturb believers (Gal. 6.11, 2 Thess. 2.2, 2 Pet. 2.1, 2 Cor. 4.2 etc.).

The second recollection of Messiah’s words in 46.8, recalls Lk. 17.1 in sequence and content (Metzger also lists two other parallel passages in Matthew but they add no content except the word katapontisqh ‘he should be plunged’ from Mt. 18.6). Metzger points out there is no parallel to Clement’s ‘he should cause one of my elect to offend’, but in the Alexandrine codex Clement here has the closer parallel to Luke, ‘offend one of my little ones’. Again the idea of a doom so tragic that birth is undesirable has an obvious parallel in Mt. 14.21 or 26.24. 

So these data do not justify Metzger’s doubtful conclusions. Again as we have seen with LXX, Clement frequently does not make explicit, complete or precise citations, only 38% of his all his references to the OT contain explicit indication of the Lord’s authorship, even of these few are close to being verbatim quotes from LXX.

Indeed given Clement’s own use of the preservation of the temple ordinances and the severe penalties for deviation in 41.2-4 as argument a fortiori for guarding NT ordinances as carefully as those of the OT – it would be a striking inconsistency for him to unconcerned about the more glorious revelation of Christ the Heavenly Temple. If Moses in 43.1 was meticulous to faithfully record his commands in the holy books (en taiv ieraiv bibloiv), would not Clement also consider it negligent to an extraordinary degree, if the documents of the greater mission were not also kept strictly and precisely? Or despite His promise of the Spirit to His apostles, were the Redeemer’s words to be left at the mercy of failing memories and fallible transmission? 

Clement’s References to the NT Epistles

Clement has at least 34 allusions to New Testament Epistles. Since some may be contested, I have listed them in the below (panel 2). This includes a broader sample than Professor Metzger reports, registering only Romans, 1 Corinthians, Galatians, Philippians, Ephesians, Hebrews, and perhaps Acts, James and 1 Peter.

The three allusions to 2 Peter are of particular interest, 7.6, 9.4, and 20.12, not least because of the intense distaste shown towards this epistle by unbelieving opponents of the Bible, which is hardly surprising given its content. It is an opposition Professor Metzger himself briefly entertains by mentioning proposals to omit 2 Peter and Jude from the canon. It is found on the same page as his sympathetic description of Aland’s suggested mutilation of the canon for church unity. ‘On the surface’ he claims, ‘such a proposal may appear have much to commend it’, though ultimately he rejects this cynical proposal on political grounds[11]. 

7.6 and 9.4 contain reference to Noah’s preaching ministry and to its purpose and two aspects of its content which is disclosed by no other parts of scripture than 2 Peter 2.5. Clement also here refers to the saving of his hearers, his own house, uniquely revealed by 2 Pet.2.5 and 1 Pet.3.20-1.

20.12 contains a doxology which, like 2 Peter 3.18 and Heb 13.21 uniquely among the epistles, attribute eternal glory particularly to Christ. Jude’s parallel passage, v. 25, also contains Clement’s word ‘majesty’ (megalwsunh), though Jude’s ascription is to the only wise God our Saviour. 2 Peter shares 10 words with Clement - all but one in the same case, part of the title ascribed to Christ, and a more direct attribution of glory to Christ, indicated in Clement by the five word phrase before the relative pronoun. Hebrews with 11 common words is marginally closer, and shares the same terminating phrase, though the ascription of glory is not so clearly directed to Christ primarily. 

Clement’s OT quotes often derived from NT not LXX.

Clement’s use of OT quotes are much closer to the New Testament text than to the Vaticanus text of LXX, some as allusions others as citations. This demonstrates his regard for the NT epistles as scripture per se. For example, as mentioned above, the relatively loose citation of Jer. 9.23-4 in 13.1 contains the Greek five words ‘he who boasts let him boast in the Lord’ phrase verbatim from 1 Cor. 1.31, though here numbered among OT citations. 

Some serial OT quotes closely follow the theme and content of NT passages, even though the citations have been numbered here to the OT, for example from Hebrews 1.2-13 in 36.1-5. Here his citation of Ps. 104.4 is closer in the case of two words and the number of one to Heb. 1.4 than LXX, although his quote of Psalm 2 in 36.4 is extended by a verse from that found in Heb. 1.5.

Likewise in 23.5 the cited paraphrase of Mal. 3.1 is closer to Heb. 10.37. In 17.5 the reference to Num. 12.7 is slightly closer to Heb. 3.2 (personal pronoun and word sequence). The tense for ‘delay’ in his allusion in 23.5 is closer to the NT rendering in Heb 10.37 and to the Hebrew OT than to the LXX for Hab. 2.3. The allusion to Ps. 24.1 in 54.3 matches 1 Cor. 10.26 slightly better than LXX. 

34.3 contains a citation which alludes to both Isa. 40.10 and 62.11 and distantly perhaps to Mal. 3.1, but the structure and wording of the text is much closer to Rev. 22.12, with obvious implications for dating Clement if true.

NT citations.

By the given criteria, there are 9 citations of the NT in Clement. Again since some are likely to be contested they are given below (panel 3). Only citations of shared OT passages which show the greatest proximity to the NT are included. For example, at 34.8 Clement cites a 19 word verse from scripture ‘h grafh’ (from the context of 34.6). The reference alludes to the last 18 words of Isa 64.4, of which the only one common to Clement is the simple conjunction kai, but used in quite a different location, all the rest is mere paraphrase. However when compared with 1 Cor. 2.9, it is immediately apparent from which source Clement drew. Of the 20 words in Paul’s text, 15 are absolutely identical, 13 consecutively and 15 in perfect order, the whole verse forming a much closer paraphrase. This begs an important question, Clement had the text of LXX and quotes liberally from Isaiah for example, from verses never cited by the NT, why does he then now rely so heavily on Paul’s epistle to cite scripture? If Metzger were correct in claiming a radical dichotomy in Clement’s thinking between the Divine oracles behind the LXX on one hand and the humbler, less exalted epistles of the apostles on the other, this would show remarkable inconsistency.

Clement’s citation at 23.3 particularly deserves closer examination. “Let not that scripture (h grafh) be applicable unto us which saith, Wretched are the double-minded, even they that doubt in their heart and say, We have heard these things in the time of our fathers; and lo, we have grown old, and none of them hath happened unto us.” Here is a chimera of two NT verses cited as scripture - ‘h grafh’ a description the NT and Clement reserve elsewhere only for divinely inspired texts. No verse in LXX or the apocrypha come close to James 1.8 and 2 Peter 3.4.
 

Since Metzger does not recognize Petrine authorship of 2 Peter[12], we scrutinise the strength of the citation. First both contain a warning against unbelief, which in Clement is continued in the next verse. Second, both contain reference to the ‘fathers’, third both suggest the passage of considerable period of time, fourth both warn of the skepticism toward the reality of God’s warnings expressed in scoffing terms, which fifthly is based on the absence of change in history. In James 1.8 the unique NT occurrence of diqucov and the contextual challenge to doubt is shared with Clement here. The shock to an infidel theologian of finding so plain a reference to an epistle he regards as both later than and unwritten by Peter should be no excuse for overriding hard data to the contrary, any more than the complete incongruity of discovering a supposedly 68 million year dinosaur fossil with intact red blood cells and soft tissues![13] Indeed the parallels between the situations run deeply.

Metzger’s suggests that a transition at 35.5-6 from a paraphrased allusion to Rom. 1.29-32 to a solid citation of Ps. 50.16-23, which we examined above, is proof of a marked dichotomy in Clement’s view toward OT scripture and Paul’s Epistles². What then of 18.1,2 where Clement cites God’s word about David, found uniquely in Acts 13.22, to lead on to David’ words about God, not specifically described as inspired but quoted from Ps. 51.1-17?  Should we then speculate Clement held a lower view of the inspiration of the Psalms than of Acts?  His argument is also manifestly both weak and tendentious, given Clement’s parallel citation of scriptures (grafh) in 23.3 and 23.5. Here the chimeric quote of 2 Peter 3.4 and James 1.8 as scripture is echoed shortly afterwards by a quote from Mal. 3.1, ‘the scripture also witnessing’.  His citation of 1 Cor. 1.10-13, is introduced ‘ep alhqeiav pneumatikwv’ in truth spiritually, again indicating Clement’s view of Paul’s epistle as inspired. He makes an obvious allusion to Paul’s description of Divine inspiration in 1 Cor. 2.13-5 where the verbal root pneumatik- found here in Clement is used by Paul 4 times either as a noun or adverbially. 

In other parts of this work on the canon, Metzger suggests the Bible is not complete or sufficient, and by inference that it is a human product and not directly superintended by the Holy Spirit, in accordance with the Lord’s promise. “From a theoretical point of view the way is open for the possible addition of another book or epistle to the NT canon”[14] He briefly toys with the suggestion that the Ignatian epistles should be added to the canon¹¹. Thus he betrays his own gradualist presuppositions, a curious fideism in its own right. However Clement’s colourful if rather absurd diversion into the fable of the phoenix in 25.1-26.1, however seriously treated and precisely applied, carries a distinctly banal ring to a modern reader.  It is noteworthy that true scripture, for all its speaking asses and other distinctive miracles, is quite free of accounts of such a character. Clement’s epistle itself should be a firm antidote to ideas of opening up the canon. Though it is lofty, saturated with scriptural allusions and intensely challenging – it is unoriginal, dependent on and wholly inferior to the simplest inspired epistles.

Conclusion

Shortly before his martyrdom, the Apostle Peter solemnly warned of the profound danger of tampering with Paul’s divinely inspired epistles, as with other scriptures. It is a warning Clement both heeded and firmly repeated. The very first body of evidence Metzger examines radically exposes the error of central assumptions of his thesis of a gradually evolving canon. On closer examination, the sharp dichotomy that he claims for Clement’s treatment of OT scripture on one hand and the NT documents as significant but uninspired writings on the other melts away. On the contrary, Clement witnesses with the NT itself that the early post apostolic church regarded the writings of the Messianic Apostles and Prophets as Holy Scripture.