Recherche – Detailansicht






Altes Testament


Pietersma, Albert


A New English Translation of the Septuagint and a Commentary Series to Follow


A New English Translation of the Septuagint and the Other Greek Translations Traditionally Included Under That Title (NETS) can best be viewed as the first stage in a two-stage interpretive endeavour. That is to say, the first stage, effectively begun in 1996 with the publication of the NETS Translation Manual,1 is a translation into English with a minimum of notation, while the second stage commenced de facto with the publication of "A Prospectus for a Commentary on the Septuagint" in 19992 aims to be a full-fledged, book-by-book, verse-by-verse commentary series on the literary collection dubbed the Septuagint by tradition. In both cases, however, the focus is on the production or constitutive character of the text in distinction from its reception or history of transmission and interpretation. Both projects are sponsored by the International Organization for Septuagint and Cognate Studies (IOSCS), and NETS is being published by Oxford University Press, New York-Oxford.3

It may be called a two-stage interpretive endeavour, since no translation, ancient or modern, from one language into another, can be anything but an interpretation. Thus both NETS and the Commentary have to do with interpretation.

What kind of interpretation their object of enquiry (i. e. the LXX) is perceived to be (of its source) and the methodology deemed appropriate to the textual linguistic make-up of that translation will, I hope, become clear in what follows.


That the time has come for such an ambitious undertaking in the English language is clear in light of a number of considerations, not least of which is that the best English has at present to offer students of biblical literature is the Thomson translation of 1808,4 and the Brenton translation of 1844,5 the latter of which has been available since 1851 in a Greek-English diglot. Since the publication of these two translations, now more than a hundred and fifty years ago, significant advances have been made in modern linguistics and Greek lexicography. Moreover, numerous ancient manuscripts have come to light, and great progress has been made in recovering the original text of individual books of the Greek corpus. By way of comparison it may be noted that whereas both Thomson and Brenton were based on (essentially) diplomatic editions of a single manuscript (MS Vaticanus [B] of IV CE), the critical edition of the Göttingen Septuagint for the book of Genesis rests on a foundation of some one hundred and forty manuscripts (nine pre-dating the fourth century CE), ten daughter-versions, plus biblical citations in Greek and Latin literature.

No less importantly, the past several decades have seen a renewed interest in the Septuagint. La Bible d' Alexandrie, announced by Marguerite Harl in 1981, has been followed by at least eight translation projects of the Septuagint into modern languages, including Septuaginta-deutsch elucidated 989 ff in this issue of ThLZ by Wolfgang Kraus. A Greek-Dutch interlinear has recently been launched by Chr. Fahner and J. Poeder.6

Title and Scope

As the long-title reflects, NETS has had some difficulty delimiting the anthology of literature subsumed under the title Septuagint. Though its practical extent is determined in large measure by the biblical canon, the interest of NETS per se is of a historical-critical rather than of a canonical or scriptural nature. So, for example, Odes has been excluded, since it has dubious integrity as a literary work, and, in any case, almost all of the individual Septuagint odes are included in their native setting in other books. The sole exception is Ode 12 in Rahlfs' edition, the Prayer of Manasses, which for that reason has been separately appended to the Psalter. Moreover, since the NETS endeavour is concerned primarily with that body of translation literature that was inherited by the Christian Church from Greek-speaking Jewry and accepted as scripture, Odes qua book does not in any case belong.

It must be admitted, however, that in the final analysis NETS has bowed to the weight of tradition rather than insisting on its theoretical stance. As a result, the Göttingen Septuagint, editio maior, has been largely followed. Consequently, the NETS reader will find parallel texts of such books as Daniel (plus Addenda), Judges, Tobit and Esther without there being any suggestion that both texts can lay equal claim to being Old Greek (OG). Furthermore, though in Job the Septuagint (OG) has been presented as the main text of NETS, the asterisked materials (attributed to Theodotion), sanctioned solely by ecclesiastical usage, have been included, albeit in footnotes. Similar procedures have been followed in a few other books.

Since NETS is firmly committed to the principle of critical text, the best critical editions have been used, the Göttingen editio maior where available and the editio minor (Rahlfs) elsewhere. While NETS translators have been discouraged from deviating from their respective Greek texts, IOSCS commentators are encouraged to participate in the ongoing quest for the original text.7 Insistence on original text, however, and adherence to constitutive character have been understood to be two sides of one and the same coin.

Target audience

Ancient texts, including biblical texts, have been translated from time immemorial, and the need for such work clearly continues. What is often less clear is the precise reading-public a translation should target. This is perhaps especially true for biblical literature with its typically diverse readership. Writing specifically on the topic of Bible translations, Nida and Taber envisaged no fewer than three such audiences.

It is usually necessary to have three types of Scriptures: 1) a translation which will reflect the traditional usage and be used in the churches, largely for liturgical purposes (this may be called an "ecclesiastical translation"), 2) a translation in the present-day literary language, so as to communicate to the well-educated constituency, and 3) a translation in the "common" or "popular" language, which is known to and used by the common people, and which is at the same time acceptable as a standard for published materials.8

NETS has been aimed primarily at the reading public identified in Nida and Taber's second grouping, namely, a biblically well-educated audience, on the assumption that it is most probably this audience that has a more than passing interest in biblical traditions other than their own, even if that entails what Wilfred Cantwell Smith9 might call a pre-scriptural reading of texts. In light of this, NETS seeks to be faithful not only to the meaning of the Greek text but also tries to reflect something of the style in which the Greek translation was produced. Its general modus interpretandi is to render idiomatic Greek by idiomatic English but to resort to less than felicitous English when the Greek is perceived to be translationese rather than standard usage. In light of the linguistic register typically used by Septuagint translators, the reader of NETS can expect to encounter less than felicitous English.

Principles and Perceptions

Axiomatic for both NETS and the Commentary is the distinction between the production and the reception of the text of the Septuagint. Typical phrases used to express this distinction are constitutive character, on the one hand, and reception history, on the other. This distinction per se is by no means new to Septuagint Studies, even though its significance and implications, not infrequently, are either ignored or given short shrift. No one has drawn the line of demarcation more clearly than James Barr, first in his book The Semantics of Biblical Language,10 and more pointedly in his response to David Hill's critique11 of his book, when he wrote:

He [Hill] does not make the obvious and necessary distinction between two sets of mental processes, those of the translators themselves, whose decisions about meaning were reached from the Hebrew text, and those of later readers, most of whom did not know the original ...12

This same kind of distinction, though within the larger context of Descriptive Translation Studies (DTS), has been introduced by Gideon Toury. For Toury all translations, as facts of their recipient cultures, are brought into being by a felt need and, as such, intrinsically have three interdependent aspects designed to meet that need, namely, function, product and process. By function (or position) he has in mind not so much the use for which a translation has been designed, but rather what systemic literary slot it aims to fill within the recipient culture. Product comprises a translation's textual-linguistic make-up, that is to say, the network of relationships introduced by the translator. By process he means the strategies by which a translation is derived from its source text, which therefore includes the relationships that hold the target text and the source text together. Though the aspects are distinct they are nevertheless interdependent, according to Toury. Diagrammatically he portrays them as: position/function determining textual-linguistic make-up (product), which in turn governs the relationship that holds translation and original together (process). Since in essence the three aspects or foci form a complex whole, the central object of research into a translation is the exposing of the interdependencies of the three, with the aim of uncovering the underlying concept of translation, and the model used to shape what is produced. It is thus clear that at the production stage a translation is equipped with its intrinsic design aimed at meeting the cultural need.

During its reception history, however, the text in question may end up playing a role contrary to its original design. Thus, a translation not originally designed as e. g. a literary work of high prestige may in time be assigned that position. Be it noted, however, that such re-articulation has no bearing on its textual-linguistic make-up.13 Differently put, though the text representation of the reader/hearer has changed, that of the writer/ speaker has not.

The distinction between production and reception noted above has obvious applications for both translation of and commentary on the Septuagint. Most obviously it applies to the form of the text, since it informs the standard distinction between original text versus received text.14 Though a translation (and commentary) of the Septuagint might legitimately be based on a text-form received at a particular time and place during its reception history, e. g. the text-form of I CE Palestine, NETS, as stated above, is based on the critical text, that is to say, on the most original and pristine text hitherto re-constructed, on the basis of all extant evidence, by means of commonly agreed upon text-critical principles. Secondly, the distinction in question is directly relevant for the meaning of the text, as per Barr's critique of Hill. That the translators themselves read their target text in the same manner as readers of that text in I CE Palestine is inherently improbable, and in any case must be demonstrated rather than assumed. In point of fact, the difference between the produced text and the received text might be so great as to necessitate speaking of different Septuagints, lest there be a tacit assumption in scholarly discussion that the Septuagint is the Septuagint, while in reality quite different entities and distinct methodologies are at issue. Thus, though the Septuagint may not exist, there is every reason for flagging basic distinctions that identify one's object of enquiry.

Although at the semantic level the distinction between constitutive character and reception history is of more deep-seated importance for the Commentary than it is for NETS, both stages of the present endeavour are nonetheless based on the same explanatory model or paradigm for the original text.

That most books or translation units of the Septuagint have been written in Koine Greek with considerable linguistic interference from the source languages (Hebrew and Aramaic) is a broadly based, if not in fact universal, consensus in the discipline.15 NETS fully subscribes to this consensus. Within DTS it may further be noted that, in accordance with Gideon Toury's law of interference, "phenomena pertaining to the make-up of the source text tend to be transferred to the target text."16 Even so, as seems clear from secondary literature, Septuagint Greek remains to be viewed as a semiticized translational variety of Koine Greek the employment of which needs to be accounted for. Significantly, modern scholarship and ancient ecclesiastical judgment coincide on this point.17 What remains controversial, however, is how that linguistic character is to be accounted for and what bearing it is deemed to have on its exegetical dimension. Should it be treated as a linguistic phenomenon indicative of prospective design and aim, or should it be regarded as a theological phenomenon communicating divine mystery?18

In NETS' view, however, the only reasonable and possible arbiter in a dispute about appropriate interpretive strategies for such a text is its textual-linguistic make-up. In other words, interpretive strategies and rules of exposition must arise from clues the text itself provides, rather than being wholly superimposed from the outside. To put another way, though it is a truism to state that translation means interpretation, whether translation is tantamount to redaction is quite another matter. The clues for that any given text itself must give.

Accordingly, for NETS the Letter of Aristeas, though of great interest as a witness to the reception history of the Septuagint - especially its re-articulated function and scripturalization - has little if anything to offer regarding its production. Nor for that matter do any other external pronouncements from antiquity. Furthermore, Aristeas's legend of origins has direct relevance only for the Greek Pentateuch, but even if one were to extend it, mutatis mutandis, to other books of the corpus,19 it remains perforce an apologia for the Septuagint a century and a half after its inception, with the manifest aim of defending it as a text in its own right, genealogically a translation but genetically a work of great literature and philosophy, never to be amended or emended. From that perspective it is scarcely a surprise if Aristeas's translational terminology, as Holger Gzella has recently maintained, bespeaks a work intended to be both translation (Übersetzung) and interpretation (Interpretation).20 In fact, had it been otherwise, one might well have raised serious questions about the coherence of the Aristean discourse. But to regard Aristeas's depiction as reflective of the constitutive character of the text itself and thus to elevate it to the status of explanatory model for its linguistic make-up, and hence its exegetical dimension, cannot be accepted. Rather than being rooted in the text, a paradigm built on Aristeas is nothing more than a superimposition upon the text as produced. It comes as no surprise, however, that on the basis of such an explanatory model conclusions are being drawn about the distinctive ideological profile of the translated text. As I have argued earlier, there is of course nothing improper about making the-Septuagint-of-reception-history the object of one's investigation - as long as reception and production are not confused. Original text and constitutive character go together in the same way that textus receptus and re-articulated reader's text representation go together.

NETS, on the other hand, views the Septuagint as a text (originally) designed to play an ancillary role to the Hebrew source text, hence a translation intended to bring the reader to its source, rather than a translation that brings its source to the reader. Judging from its textual-linguistic profile, when the Greek translation of the LXX is mapped onto its Semitic source, the Greek was not designed to take the place of the Hebrew original nor was it a text of equal status to its impeccable parent. This conclusion is further underscored when the translated text is subjected to discourse analysis. Not only does it typically have few cohesive links that do not merely represent items already in the source text, but also an isomorphic and atomistic translational modus operandi has not infrequently produced items that are disruptive of textual coherence, and might thus be labeled anti-links. Consequently, in NETS's view, the dominant characteristic of the majority of books within a heterogeneous Septuagintal anthology is that of isomorphic representation of the source text. If that is its dominant characteristic it must also be labeled its defining characteristic and, therefore, presents itself as an interpretive presupposition for the reader. Again, in NETS's view what we typically have in the Septuagint is a translation rather than a redaction. More particularly, the translation is in a linguistic mode rather than in a textual one, and decidedly not of a literary variety.21

As stated, NETS is committed to the historical-critical approach, which is understood to include not only a focus on the most original text form but also on its constitutive character. Moreover, though NETS fully recognizes that at some stage in its reception history the Septuagint not only was assigned a freestanding status and became regarded as sacred scripture, that re-articulation (or even re-contextualization) of the original text does not fall within its purview.

A second major distinction for NETS is that between translational and non-translational literature. As is well known, the corpus of the Septuagint is comprised mostly of translations from Hebrew and Aramaic. Included, however, is a number of texts originally composed in Greek: 2-4 Maccabees, Wisdom of Solomon and Additions to Esther. Not only are these two categories of texts distinct in origin and production, it is for that very reason that they are subject to different interpretive methodologies.22 It is the former of the two categories that constitutes the primary focus of NETS.

Interlinearity23 and its Explanatory Power

NETS has been based on the interlinear paradigm for essentially three reasons. First, this paradigm best explains the "translationese" aspect of typical Septuagintal Greek, with its strict, isomorphic, quantitative equivalence to the Hebrew. This aspect cannot be fully accounted for on the basis of Toury's law of interference governing translations generally. For beyond instances of what Toury has termed positive transfer from the source text, i. e. the overuse of grammatical features that exist in the target language,24 one finds negative transfer as well, i. e. deviations from normal codified practices of the target system.

The latter, perhaps, needs brief elucidation. What NETS has in mind is items such as en emoi for Hebrew by, ego eimi + finite vb. for 'nky) + finite vb., dynamai + tu = + infin. for ykl + l + infin., anthropos anthropos for 'ys 'ys, pronoun + finite vb. for pronoun + participle, Greek article to represent the so-called nota accusativi or inseparable prepositions, transcriptions of Hebrew, (certain) prepositions as verbal completion, pleonastic pronouns and adverbs, and so the list could go on. One finds negative transfer on both the level of grammar and on the level of discourse, not to speak of semantic obscurity and even unintelligibility. This is not to say that one encounters such items on every line of Greek or even in every paragraph, but since an explanatory model must be able to accommodate both unintelligibility and intelligibility, translationese as well as high literary usage, NETS has opted for the interlinear paradigm. Secondly, the interlinear paradigm provides the translator with a principled way of drawing on the Hebrew parent text as an arbiter of meaning when the Greek is ambiguous or nonsensical. Differently put, the interlinear paradigm recognizes that unintelligibility is an inherent characteristic of the Greek text qua text. Thirdly, the interlinear paradigm safeguards the Greek-ness of the Septuagint by not forcing the model reader to posit meanings (syntactic as well as lexical) not attested in compositional Greek or in translational contexts where compositional use can be demonstrated. Lexicographically, for instance, one is not hamstrung by the rule that context determines meaning, when it is clear that the use of a given lexeme is determined more by the occurrence of its counterpart in the source text than by contextual considerations of the target text.

It can finally be argued that "interlinear" signals that the LXX qua text is two-dimensional, that is to say, it has a vertical as well as a horizontal dimension. On the horizontal plane morphemes are knit together into syntactic units to convey information, on the vertical plane the source text tends to form the de facto context for units of meaning and, as a result of excessive isomorphic dependence on the source text, the target text may be rendered disjointed or worse. That is to say, in an interlinear text one expects that, on occasion, the vertical dimension interferes with the horizontal to such an extent that the text lacks semantic coherence.

I have elsewhere25 likened the two dimensions of an interlinear text to the literary genre "sentimental romance" as characterized by the well-known literary critic Northrop Frye.26 Frye compares the romance with the modern novel. In the novel as a realistic narrative - Frye argues - the writer attempts to keep the action horizontal, using the technique of causality to keep the narrative moving from within and to knit the plot firmly together. Romance, on the other hand, tends to be more sensational, "that is," says Frye, "it moves from one discontinuous episode to another, describing things that happen to characters, for the most part, externally." He then goes on to speak of the "hence" narrative (the novel) and the "and then" narrative (romance). I may perhaps remind my readers at this point of my earlier note on discourse analysis, cohesive links and anti-links.

It is in light of the two-dimensionality of the translated text that the Prospectus for the IOSCS Commentary series states:

When the text is a translation rather than an original composition, one should take an essentially two-pronged approach: First, because it is a translation, the contextual sense of Greek words or expressions may have suffered interference from the Greek's close relationship to the parent text. Consequently, one may be forced to treat the Greek text as being disjointed [cf. vertical dimension]. Second, because, in spite of its precise relationship to its parent text, the Greek text is nevertheless a new entity, one should treat it, as much as is warranted, as a unitary whole [cf. horizontal dimension].27

NRSV as base text for NETS

While it is obvious that the Septuagint at some point in its reception history, in Egypt, became a text read independently from its Semitic parent,28 it is equally true that it was, in its inception, a Greek translation of an original in another language. Furthermore, judging from the textual-linguistic make-up of the vast majority of its books, the dependence it was designed to have was not simply a matter of genealogical descent, in the way that all translations are descended from their sources. Rather, it was a dependence that signaled an intrinsically ancillary role. To reflect, at least in part, that role of subservience and service to its source text, the NETS Committee decided to establish a similar, even if not identical, relationship between NETS and an existing English translation of the Masoretic Text. While it might be objected that the text of MT (vocalized or unvocalized) was by no means always the parent text of the Greek, it remains true nevertheless that the percentage of agreement between the Vorlagen of both is so great as to warrant the methodological dictum: "the Greek translates a text identical to MT until proven otherwise." Furthermore, if the Greek by design played an ancillary role, it seems appropriate not to read it as though it were an entirely new, free-standing entity, or as a kind of redaction of its source. Thus NETS's dependence on the NRSV accords well with the interlinear paradigm which functions as the explanatory model for both NETS and the IOSCS Commentary series on the Septuagint. As the general introduction to NETS29 makes clear, the specific choice of the NRSV as base text was guided by purely practical considerations.

An added benefit of the virtual diglot of NSRV and NETS is that English translations of MT and LXX can, mutatis mutandis, be read synoptically.


To the extent that Septuagint Studies has as its central object of enquiry the transfer of meaning from a Semitic source text to a Greek target text, to that extent both NETS and the IOSCS Commentary series can be said to be of direct relevance to the discipline as a whole.


1) Pietersma, Albert, Translation Manual for "A New English Translation of the Septuagint" (NETS). Uncial Books, Ada, Michigan 49301: 1996. Accessible at:


2) BIOSCS 31 (1998), 43-48. Accessible at:

3) The first volume of NETS appeared in 2000 (A New English Translation of the Septuagint, The Psalms. Oxford University Press, New York-Oxford).

4) Thomson, Charles, The Holy Bible, containing the Old and the New Covenant, commonly called the Old and the New Testament. 4 vols. Philadelphia, 1808.

5) Brenton, Sir Lancelot Charles Lee, The Septuagint Version of the Old Testament, According to the Vatican Text, Translated into English: with the Principal Various Readings of the Alexandrine Copy, and A Table of Comparative Chronology. 2 vols. London, 1844.

6) De Griekse text van het Oude Testament. Genesis. De Banier, Utrecht 1999.

7) Prospectus, principle (1), reads: "the principle of original text, which is understood to mean that though for any given book the best available critical edition will form the basis of interpretation, commentators shall improve upon that text where deemed necessary, and thus assist in the ongoing quest for the pristine Greek text" (44).

8) Nida, E. A. and C. R. Taber, The Theory and Practice of Translation. Leiden, 1982, 31.

9) "Scripture as Form and Concept: Their Emergence for the Western World", in: Rethinking Scripture: Essays from a Comparative Perspective (Ed. M. Levering, Albany: State University of New York Press 1989), 29-57.

10) Oxford 1961.

11) Hill, D., Greek Words and Greek Meanings: Studies in the Semantics of Soteriological Terms. Cambridge, 1967.

12) "Common Sense and Biblical Language", Biblica 49 (1968), 377- 387, 379.

13) See Descriptive Translation Studies and Beyond, Benjamins, Amsterdam/Philadelphia, 14 (1995), 272-273.

14) I use this phrase loosely as indicating any secondary text-form, including recensions.

15) Wolfgang Kraus in his article in this issue of ThLZ in fact uses the term "barbarisch".

16) Op. cit., 275.

17) See e. g. Natalio Fernandez Marcos, The Septuagint in Context: Introduction to the Greek Version of the Bible, Brill, Leiden/Boston/ Köln 2000.

18) The most ardent ancient proponent of this view was Jerome, but in recent literature it has been suggested by Adrian Schenker in his "Gewollt dunkle Wiedergaben in LXX? Am Beispiel von Ps 28 (29),6", Biblica 75 (1994), 546-555.

19) See recently Holger Gzella, Lebenszeit und Ewigkeit. Studien zur Anthropologie und Eschatologie des Septuaginta-Psalters. Philo, Berlin 2001.

20) Ibid., 39.

21) See Toury, op. cit., 171.

22) See Toury, op. cit., 28.216; cf. also Prospectus 47.

23) For a recent application of the concept to the Septuagint see Hermann-Josef Stipp, "Bemerkungen zum griechischen Michabuch aus Anlass des deutschen LXX-Übersetzungsprojekts", JNSL 29 (2003), 103-132 [115].

24) H. St. John Thackeray (Grammar, 29) citing J. H. Moulton refers to this feature as over-working.

25) "A New Paradigm for Addressing Old Questions: The Relevance of the Interlinear Model for the Study of the Septuagint" in Bible and Computer. The Stellenbosch AIBI-6 Conference. Proceedings of the Association Internationale Bible et Informatique "From Alpha to Byte". University of Stellenbosch 17-21 July, 2000. Johann Cook (ed.). Brill, Leiden/Boston: 2002, 337-364 [351]. Accessible at:

26) The Secular Scripture. A Study of the Structure of Romance. Harvard University Press, 1976.

27) Op. cit., 47.

28) See the "Letter of Aristeas".

29) A New English Translation of the Septuagint, The Psalms, viii-x.