1. Context: The OT Citations in Hebrews and Current Issues in Septuagintal Studies
‘Replete with quotations, allusions, general references and echoes, Hebrews packs more of the Old Testament into its complex discourse than any other New Testament writing… Simply stated, the uses to which Hebrews has put the Old Testament are the book's bone and marrow’.Footnote 1 George Guthrie's account of the extent of the influence of OT texts and themes on the structure and theology of the Letter to the Hebrews is hardly controversial, and recognition of this fact has resulted in numerous publications on the use of the OT in the epistle. Some of these have attempted detailed analyses of the specific OT citations, investigating either their text form or the exegetical techniques employed by the author to interpret them. Among the most influential of these studies of the OT citations in Hebrews are those by Caird,Footnote 2 Thomas,Footnote 3 Howard,Footnote 4 McCullough,Footnote 5 and Cadwallader.Footnote 6
The existence of such an extensive and valuable body of literature on this subject might be taken as an indication that opportunities for further fruitful research in this area are likely to be limited, but, on the contrary, I suggest that in fact the time is opportune for a new investigation into the form(s) of Greek biblical text underlying the citations in Hebrews, one which deliberately seeks to make connections between NT research and the study of the LXX. Two factors in particular have prompted this conclusion. First, a study of Syriac forms of the book of Psalms by Robert HiebertFootnote 7 brought to my attention the existence of the reading in Papyrus Bodmer XXIV of the prepositional phrase ἐν δοκιμασίᾳ at the beginning of line two of Ps 94(95).9. This form, of course, agrees with the citation of this text in Heb 3.9 against other major LXX witnesses, which use the verbal form ἐδοκίμασαν. Although it is now almost twenty years since the publication of Hiebert's book, this important textual witness does not appear to have been noted by any commentators on Hebrews. This suggests that detailed study of other individual OT citations in the letter may also uncover previously neglected textual variants, or help to extend our understanding of the forms in which scripture circulated in the first century ce.
Secondly, important advances have been made in the field of Septuagintal Studies over the last two decades, some of which have not yet been fully appreciated by NT scholars. The significant manuscript finds at Qumran and Naḥal Ḥever, together with the work of a new generation of LXX specialists, are changing perceptions about the multiplicity of forms in which the scriptures circulated in both Greek and Hebrew in the early post-biblical period,Footnote 8 and about the value of specific manuscript traditions, particularly the Lucianic or Antiochian recension.Footnote 9 This means that the serious exegete can no longer be content, as some of the earlier commentators on Hebrews were, with comparing the textual form of OT citations in the NT with only the major LXX witnesses like Codices Alexandrinus and Vaticanus. It also suggests that there is a very strong case for a reappraisal of some of the earlier critical editions of individual books of the LXX, which were published at a time when the editors either did not have access to all the manuscripts and other evidence now available, or were not sufficiently aware of the need to take seriously alternative readings from the Lucianic and other traditions. Discussing the Göttingen edition of the Psalms, for example, Albert Pietersma has concluded that: ‘More so-called L (i.e. Lucianic) readings have ancient roots than is evident from Rahlfs’ edition’.Footnote 10
The need for NT commentators to be more alert to this recent research and more sophisticated in their approach to the use of the Greek bible in the early Christian texts has already been recognised by some, for instance Timothy McLay.Footnote 11 Detailed research into the Vorlage of the OT citations in certain NT books, drawing on all available textual evidence, is also currently being undertaken by a small number of scholars, notably Maarten MenkenFootnote 12 and Gert Steyn.Footnote 13 Considerable scope remains, however, for the expansion of these efforts. Section two of this paper will therefore examine in detail the seven direct OT citations found in Hebrews chapter 1, and attempt to establish very precisely the nature of the text form underlying them. The final part of this study will then draw out some conclusions and consider the implications of these findings for current issues in the study of both the LXX and the use of the OT in the NT. Hebrews chapter 1 would seem to make a useful representative starting point for an investigation of this kind, despite its short length, because of the number of definite scriptural citations which it includes and the evident importance of the OT to its author.Footnote 14 In addition, almost all of the quoted texts are taken from the Psalms, a point of interest for this study, as it is the view of a number of leading contemporary LXX scholars that the Psalms volume in the Göttingen series, published by Rahlfs in 1931, is in particularly urgent need of re-editing and updating. Flint, for example, has estimated that Rahlfs elected to use only about one hundred of the over one thousand manuscripts available to him, largely discounting the evidence of the versions, patristic sources, and Lucianic readings,Footnote 15 and Pietersma has critiqued Rahlfs’ method for determining relationships between manuscripts.Footnote 16 Since other important sources, such as the Bodmer papyri and Qumran manuscripts, have been discovered subsequently, it is more than possible that a detailed examination of the Psalms cited in Hebrews will uncover significant new readings, which should influence the debate about the extent to which the author of the epistle cited his scriptural sources accurately.
2. The LXX Sources of Specific OT Citations in Hebrews
2.1. Citations Showing No Divergence from Standard LXX Readings
Three of the seven OT citations in Hebrews chapter 1 are given in a form identical to the reading found in all the major witnesses to the LXX, namely Ps 2.7 in Heb 1.5a and Ps 109(110).1 in Heb 1.13. The citation in Heb 1.5b, whether it is assumed to be taken from 2 Sam 7.14 or 1 Chron 17.13, likewise reflects the LXX exactly. In all three cases, there is no evidence for the existence of any textual variants.
2.2. Citations with Minor Variations from Standard LXX Readings
2.2.1. Psalm 103(104).4 in Hebrews 1.7
There is just one difference to be noted between the citation as given here in Hebrews and almost all witnesses to the text of the LXX. At the end of the second line, Hebrews has the form πυρὸς φλόγα (flame of fire) rather than the standard LXX reading πῦρ φλέγον (fire and flame/flaming fire). The text as given in Hebrews is attested in one correction to LXXA, some manuscripts assigned to the Lucianic group and in the Bohairic and Sahidic versions, as noted in the apparatus in the Göttingen edition.Footnote 17 However, these witnesses are few and not particularly valuable, and were judged by Rahlfs to be secondary and influenced by the reading in Hebrews. Most other commentators (for example AttridgeFootnote 18 and KatzFootnote 19) agree. This view is now open to question, however, in the light of the discovery of support for the form in Hebrews in Papyrus Bodmer XXIV, a Christian but possibly pre-Hexaplaric manuscript. Even Cadwallader accepts this reading as a possible genuine variant,Footnote 20 despite the general thrust of his study, which stresses the tendency of scribes to correct the text of the OT citations in Hebrews towards the LXX. The fact that alternative readings for this phrase are found also in Symmachus (πύρινην φλόγα) and Aquila (πῦρ λάθρον)Footnote 21 makes it more likely that the source text used by the author of Hebrews contained a variant reading.
2.2.2. Psalm 44(45).7–8 in Hebrews 1.8–9
This citation is largely given in Hebrews in exactly the same form in which it is found in most witnesses to the LXX. The only differences occur in the form of the second line of v. 7 of Psalm 44(45) as reproduced in the last line of Heb 1.8. Here, Hebrews has an extra καί at the beginning of the line and places the definite article ἡ before the first occurrence of ῥάβδος, rather than the second, where it occurs in the LXX, thus making ἡ ῥάβδος τῆς ɛὐθύτητος (the sceptre of righteousness) the subject of the sentence, not the predicate. No witnesses to support the reading in Hebrews, nor any other variants, are given in the textual apparatus of Rahlfs’ edition.Footnote 22 Most commentators conclude that the author of Hebrews has made minor changes to his source here,Footnote 23 probably to emphasise the parallelism between God's throne in the first line and the royal staff of God or the son in the second clause. The addition of καί in particular is taken as an example of his characteristic tendency to divide an OT citation into two to make two different points.Footnote 24 However, Ellingworth notes the textual and exegetical problems in the Hebrew original underlying the first line of this citation,Footnote 25 which could have resulted in minor differences in translation and thus led to variant readings.
2.3. Citations Differing More Significantly from Standard LXX Readings
2.3.1. Deuteronomy 32.43(LXX) in Hebrews 1.6
Determining the scriptural source of this citation is not a straightforward task, and this verse and its underlying OT text have consequently been the subject of several detailed studies, such as those by CockerillFootnote 26 and McLay.Footnote 27 One possibility is that it is taken from Ps 96(97).7, which reads προσκυνήσατɛ αὐτῷ πάντɛς οἱ ἄγγɛλοι αὐτοῦ (worship him, all his angels). However, most commentators take the view that it is in fact a citation of the LXX form of Deut 32.43,Footnote 28 καὶ προσκυνησάντων αὐτῷ πάντɛς υἱοὶ θɛοῦ (and let all sons of God worship him). This conclusion presents some difficulties, because the verse differs from the equivalent section in the MT. However, the Qumran discoveries have yielded a Hebrew text (4QDeutq) which does contain the clauses found in the LXX but absent from the MT, thus making it more probable that the Greek form of the passage is based on a genuine Hebrew tradition.Footnote 29
Deuteronomy 32.43 has the third person plural aorist imperative form of the verb ‘to worship’, as in Heb 1.6 but not Ps 96(97).7, which uses the second person plural, but there is a difference between Hebrews and the LXX at the end of the line, where Hebrews has πάντɛς ἄγγɛλοι θɛοῦ (all God's angels) and the LXX πάντɛς υἱοὶ θɛοῦ (all God's sons). Some commentators think it likely that the author of Hebrews made this change to his source himself, as the reading of the majority of LXX manuscripts would not suit his whole argument in ch. 1 about the uniqueness of Christ's sonship. However, the apparatus in the Göttingen edition reveals considerable textual uncertainty throughout this verse,Footnote 30 some of which was perhaps occasioned by the potentially theologically difficult phrase ‘God's sons’, which may have been changed in some textual traditions to ‘angels’, as is the case in some other verses in the LXX where ‘sons of God’ or ‘gods’ occurs in the MT (e.g. Ps 8.6[5]; 97[96].7; 138[137].1; Job 1.6; 2.1; 38.7). The variation may also have arisen from simple transposition, either by the author of Hebrews or already present in his source, of the parallel clause in the fourth line of the verse. A version of this text almost exactly the same as that given in Heb 1.6, including the phrase καὶ προσκυνησάντων αὐτῷ πάντɛς οἱ ἄγγɛλοι θɛοῦ (note the additional οἱ, not present in Heb 1.6), is attested in the Ethiopic versions, and also the Odes (2.43), a group of hymns attached to the Psalter of Codex Alexandrinus and other major LXX manuscripts from about the fifth century. Ellingworth therefore considers this evidence irrelevant, as it is of a relatively late date and may have been influenced by the text in Hebrews.Footnote 31 McLay, on the other hand, concludes that Odes 2.43 and Heb 1.6 are related to a longer reading of this text which was preserved in the Old Greek but not the MT,Footnote 32 and this view is shared by Steyn.Footnote 33 I agree that it would be difficult to rule out completely the possibility that the Odes was drawing on earlier textual traditions. The evidence of the Odes and of the Qumran text (4QDeutq) would, therefore, seem to indicate that the author of Hebrews was citing his source faithfully, but that it differed from the text printed as standard in modern critical editions of the LXX.
2.3.2. Psalm 101(102).26–29 in Hebrews 1.10–12
There are a number of differences between Hebrews and the majority reading of the LXX in this case. First, the pronoun σύ, addressing the Lord, is placed at the beginning of the citation in Hebrews, whereas it is the third word of the line in the LXX. Most commentators regard this as a deliberate change on the part of the author for emphasis, or to connect this scriptural text clearly with the previous one, which ends with the word σου. This is the view of Thomas,Footnote 34 for example, and although he may be correct, his study as a whole is weakened by the fact that he does not consider much evidence beyond that contained in the two major textual witnesses to the LXX, Codices A and B. In fact, there is some evidence, particularly in the versions such as the Old Latin, for considerable variation in the word order in this line.Footnote 35
The next differences occur in the first line of Heb 1.12, where the last line of Ps 101(102).27 is being quoted. Hebrews has ἑλίξɛις (you will roll up) against the LXX reading ἀλλάξɛις (you will change), the verb which occurs also at the end of the line, and repeats ὡς ἱμάτιον (like a garment) after this verb. The textual apparatus in the Göttingen edition lists some supporting evidence for ἑλίξɛις rather than ἀλλάξɛις in witnesses from the Alexandrian, Vaticanus, and Lucianic traditions, as noted by Attridge,Footnote 36 but not for the repetition of the noun.Footnote 37 There are, then, sufficient textual witnesses to convince many commentators that ἑλίξɛις may have been present in the text before the author of Hebrews,Footnote 38 but most conclude that the repetition of ὡς ἱμάτιον was a deliberate addition made to his source for stylistic reasons, in order to make the imagery more vivid, or to emphasise the contrast between the changeable nature of the angels and the eternal son.Footnote 39 However, McCulloughFootnote 40 has argued for accepting the entire reading in Hebrews as a genuine LXX variant, and now Flint's more recent study of the Qumran Psalms Scroll (11QPsa) has revealed the existence of a Hebrew text containing the repetition of ‘like a garment’Footnote 41 which would seem to support his position.
3. Conclusions and the Implications for NT and LXX Study
This analysis of Greek versions of the biblical texts cited in Hebrews chapter 1 has, then, highlighted information relevant to the study of the Letter to the Hebrews, of the use of the OT in the NT generally, and of the LXX. First, evidence has emerged to support some readings in Hebrews which differ from the form of the text in standard critical editions of the LXX. Manuscripts uncovered in recent decades, such as Papyrus Bodmer XXIV, 4QDeutq, and 11QPsa, for example, indicate that the author of Hebrews may well have been citing a genuine textual variant rather than deliberately altering his source where his version of Deut 32.43, Ps 101(102).27, and Ps 103(104).4 (cf. also Ps 94[95].9 in Heb 3.9) differs slightly from the standard LXX form. This serves to strengthen the growing consensus of contemporary Septuagintal scholarship about the fluidity of the scriptural text in the first century ce. Indeed, the Letter to the Hebrews, and the writings of the NT as a whole, should now be seen as themselves witnesses to the textual pluriformity which characterised this period.Footnote 42
Secondly, this study has sought to take account of changing scholarly perceptions about the antiquity of the Lucianic recension and its value as an authentic witness to the Old Greek. If this view is correct, then greater weight will need to be given to those cases where readings found in manuscripts from the Lucianic tradition support the form of the text given in Hebrews chapter 1 (for example, πυρὸς φλόγα in the citation from Ps 103[104].4 in Heb 1.7, or ἑλίξɛις rather than ἀλλάξɛις in the last line of Ps 101[102].27 in Heb 1.12). Wider consideration of other citations of the OT in the NT which might also be related to Lucianic readings could, therefore, be useful in adding to our understanding of both the history of this recension and of the use of scripture by the NT authors.
Thirdly, this investigation adds support to the calls of scholars such as Pietersma and Flint for a thorough revision and updating of the Göttingen Psalms volume. The evidence of the Qumran biblical manuscripts and other recent discoveries such as the Bodmer papyri all needs to be incorporated in a new critical edition of the LXX Psalms. As noted above, the NT itself may also witness to some genuine alternative LXX readings, and scholars are aware of other potentially valuable sources for the task of better reconstructing the original text which are largely untapped, for instance biblical quotations in Philo and Josephus, the vast body of patristic writings, and early versions of the bible based on the LXX such as the Old Latin and the Coptic. Dines, for example, regards the use of the LXX in the works of Philo as ‘a resource that awaits further study…’,Footnote 43 and Fernández Marcos states that ‘the question of biblical quotations is the weakest point of the Cambridge and Göttingen editions’.Footnote 44 There are certainly great difficulties in making use of the patristic biblical commentaries and other writings for the critical study of the LXX, since they frequently include citations from memory, or citations which have been conflated, or influenced by parallel passages, or by the form of a text in the NT. A pre-requisite for further progress in this area is, therefore, the publication of good new critical editions of patristic works.Footnote 45
Fourthly, this study suggests that the author of Hebrews generally cited his scriptural source texts accurately. It concludes that differences between the form of the text in Hebrews chapter 1 and the LXX are only likely to result from authorial intent in the case of Ps 44(45).7–8, cited in Heb 1.8–9, and that the changes made to the text in that instance are very minor. These findings have a bearing on judgments about the attitude of the author to scripture, a subject which has been a matter of debate among Hebrews commentators, with Thomas and McCullough perhaps representative of opposite ends of the spectrum. For Thomas finds in the citation of the OT in Hebrews ‘a pattern of significant changes [from the LXX] which must be more than accidental…’,Footnote 46 while McCullough,Footnote 47 on the other hand, argues that almost all of these differences reflect genuine variant readings, or are only minor stylistic changes. This examination of the OT citations in Hebrews chapter 1 supports McCullough's emphasis on the complexity of the textual state of the LXX in the first century ce rather than Thomas' view of the author's creative approach to his scriptural sources. Indeed, both the results of this study and the consensus among current LXX scholars about the extent of textual fluidity in the first century would indicate that it is now safer to operate with the presumption that the form of all the citations in Hebrews can be supported by textual evidence rather than rush to the judgment that the author freely altered scriptural quotations to suit his theological purposes. That the writer of Hebrews should have cited biblical texts faithfully is not very surprising in view of his context within early post-biblical Judaism, given that respect for the accurate reproduction of the words of scripture is evident among his exegetical contemporaries, the rabbis,Footnote 48 the targumists,Footnote 49 and the commentators who produced the Qumran pesharim.Footnote 50 For all of these ancient Jewish exegetes, as it would seem also for the author of Hebrews, it is through the interpretation of scriptural texts and their arrangement in different contexts that they can receive new meanings, not through the alteration of their wording.
Fifthly, this study has consequences for the study of the NT beyond Hebrews. A detailed examination of each OT citation individually of this kind does yield results which are far more precise than the general and rather vague statements about ‘affinities with Codices A and B’ which are still to be found in some commentaries on NT writings. It has become clear that establishing as accurately as possible the form of the source text underlying an OT citation in the NT is a necessary first step to its thorough exegesis, and to a fuller consideration of the ways in which each individual NT author approaches the scriptural text. This study has emphasised the situation of textual pluriformity which existed in the first century ce, which has been acknowledged, at least in theory, by some commentators for many years—McCullough, writing in 1980, states this clearly at the outset of his paper, for example, and a recent article by McLay, likewise claims that ‘an essential characteristic of the Scriptures of the early church was pluriformity’.Footnote 51 However, this recognition has not yet been acted upon sufficiently by the majority of NT interpreters.
It must, however, also be acknowledged that an investigation such as this faces substantial difficulties in the task of definitely determining genuine variant readings, in the absence of any hard and fast method for doing so, and in view of the likelihood that copies of the LXX were altered by Christian scribes over time to conform to the readings in the NT and vice versa.Footnote 52 Several criteria for weighing up the worth of different manuscripts are obviously well established in text criticism and will continue to be valuable, namely assessing the date, provenance, and general nature of the manuscripts witnessing to a particular reading, and the extent of independent support for it. It would also seem sensible to consider the use of the OT by an NT author as a whole, so that if he is found to be generally faithful to his source in his citations, it would seem wise to be wary of attributing to him an intentional change in some cases, without fully considering the evidence for a textual variant. It may be necessary also to undertake a detailed examination of individual LXX manuscripts on their own merits when seeking to reach conclusions about the genuineness of alternative readings; if, for example, an LXX manuscript which offers support for a variant reading contained in Hebrews can be shown to agree consistently with the NT in disputed cases, it may legitimately be suspected of having been altered by Christian scribes to conform to it. In view of the tendency for copies of texts regarded as sacred to converge over time, there may be a case for establishing readings on the basis of an approach of ‘maximum textual dissimilarity’, accepting as the best readings those in the Hebrew tradition which differ most from the LXX, and in the LXX those which differ most from the MT and the NT. These very real problems notwithstanding, I contend that efforts to reach judgments about text form in citations should continue to be made. It cannot be sufficient in every case for commentators to keep an open mind about whether or not an NT author was deliberately altering his OT source when quoting it, as a decision about this impacts crucially on conclusions about his exegetical methods and attitude to scripture. In short, I am suggesting that the current state of thinking within the field of Septuagintal Studies now reverses the burden of proof, placing it on those who would argue against a variant reading and for a definite theologically motivated alteration of a biblical source-text, rather than the other way round.
