The discovery at wadi Qumran in 1947 by a shepherd was arguably the most significant as far as biblical texts are concerned.¹ After all, texts such as 1QIsa^a are more than a millennium older than the Textus Receptus. There are conspicuous correspondences between Massoretic Text (MT) and some Dead Scrolls. At first glance, 1QIsa^a seems identical to MT Isaiah; both have 66 chapters in the published volumes. However, when scrutinised, prominent differences in addition to correspondences become apparent.

This contribution will take a closer look at a number of Dead Sea manuscripts and/or fragments in order to determine their linguistic and exegetical value. The article will, firstly, address textual material that is largely in agreement with the MT – 1QIsa^a. Secondly, fragments that are on the face of it less relevant will be discussed. The ‘insignificant’ fragments from the Biblical books Proverbs and Job are cases in point. Finally, highly significant textual differences, such as a fragment from Genesis 1 and one from the books of Jeremiah, will be evaluated.

This contribution will depart from the reality of textual plurality in the pre-common era (Tov 1985). According to this point of departure, the MT is one of the textual witnesses available (Septuagint [LXX], Tgg, Pesh, etc.) but not the most important one. Even so, MT is used as a basis of comparison. The Biblia Hebraica Stuttgartensia (BHS) edition is used for the Hebrew text (MT), DJD 32 for 1QIsa^a and the Rahlfs pocket edition for LXX.

As is well known, the books of Isaiah and the Psalms are well represented in the Dead Sea Scrolls. Of the first, there are two larger mss, 1QIsa^a and 1QIsa^b, as well as numerous fragments (Ulrich & Flint 2010). 1QIsa^a at first sight seems to be identical to MT Isaiah, for one thing, both have 66 chapters in the modern publications.² Closer inspection reveals prominent differences. Firstly, the scribes and/or copyists were less meticulous than MT Isaiah’s.³ The number of later handwritten marginal notes testifies to this (Ulrich & Flint 2010:61⁴). Secondly, there is evidence that the scroll was divided into two parts. In this regard, there is a clear gap of 3 lines at the end of chapter 33, which was used by Giese (1988:61) as main evidence for proposing a bisection (Part 1 = chapters 1–33 and Part 2 = 34–66) in 1QIsa^a. There is additional evidence of a dichotomy in this scroll.

This bisection can be observed from some orthographical characteristics of 1QIsa^a (Cook 1989, 1992). Data reworked in an electronic Qumran database (Cook 1988) as well as Logos 6, provide the necessary information. The first category is the aleph as mater lectionis used in various positions. The first example is the particle יכ. There are 323 examples in this scroll and 328 occurrences in the Textus Receptus. Defective and plene forms are used simultaneously; however, the plene form (איכ), which does not appear at all in BHS, has been used in the majority of cases, namely 194 times according to Logos. In 129 instances the defective form is used. There is a definite pattern as far as these forms are concerned. In the first part of the scroll (1–33), יכ occurs 123 times and the plene form איכ 34 times. In the second part, יכ appears 4 times and איכ 159⁵ times.

In chapters 34 through to 66, there are only 4 cases of the defective form out of 129 occurrences. There is thus a concentration of plene forms in the second part of 1QIsa^a.

The interrogative particle ימ occurs 63 times in 1QIsa^a and in 23 cases the spelling is plene. These are all concentrated in chapters 33–66.

The suffix ה/occurs 134 times in 1QIsa^a and in 5 cases (30:8; 34:10, 11; 62:4 and 66:10) has the aleph added to the he-אה. What is striking, firstly, is that all the plene forms are added after chapter 33. Secondly, not all the suffixes actually had the aleph added. The following pattern prevailed: without the aleph Isaiah 5:14; 6:13; 8:21; 24:6; 27:4; 30:32, 33; 33:24; 34:17; 37:29; 42:5; 51:3; 56:2; 59:8 and 65:19.

There is thus a pattern of sorts as far as 1QIsa^a is concerned, with plene forms being concentrated in the second half of the scroll. On the contrary, 1QIsa^b consistently uses defective forms. This phenomenon is probably the result of different scribes.

According to some scholars, there can be no doubt that the dating of textual material is of crucial importance. Hurvitz (2006) phrases it as follows:

Not everybody agrees with this point of view.⁶ Concerning the value of orthography for introductory issues, scholars also have deviating views. Girdlestone (1892:176) held the opinion that issues of spelling could be useful to determine the age and authorship of sources. Whereas formerly scholars argued that 1QIsa^a seems to have been copied by more than one person (Ulrich & Flint 2010:61), there seems to be a consensus that one scribe was responsible for copying the whole book and that later revisers made some changes and added expansions (Ulrich & Flint 2010:63). As stated above, Ulrich and Flint (2010:63) found evidence of scribal activity that can be dated to the Herodian period.

Cross and Freedman (1952:60) also thought that the study of orthography could be valuable for linguistic analyses. The problem with their work concerns the fragmented nature of their subject matter. Andersen and Forbes (1986:63) executed important research in this regard, but concentrated on the Hebrew Bible. This applies to James Barr (1989:2f.) also. He was highly sceptical of endeavours to find solutions to introductory questions via spelling characteristics. He followed a strictly descriptive approach without presenting historical solutions.

There are also more positive views regarding this issue. Martin (1958) took the Dead Sea Scrolls into account. He held the opinion that social factors in fact influenced scribal material. The environment of the scribe and/or copyist therefore also is a determinative factor. The standard research on 1QIsa^a and MT remains the book by Kutscher (1974). He also stressed the plurality of textual material at Qumran and suggested that the secluded geographical situation of Qumran should be accounted for by the researcher. Tov (1986) followed a historical approach in this regard. He is of the opinion that there existed two groups of scrolls at Qumran, which can be distinguished on account of their orthography. According to him, one group had a ‘Qumran orthography’ in which plene spelling systems predominated. The second group lacks such forms. It is immediately clear that it is difficult to classify 1QIsa^a according to this criterion. Even though plene forms dominate in this scroll, defective ones appear abundantly. On the other hand, 1QIsa^b could fit this picture.

Hopefully, it has become clear that the large Isaiah scroll differs in various respects from MT. In my opinion, orthographic characteristics can be useful for addressing introductory issues, as demonstrated by Tov, Ulrich and Flint et al.

The paucity of textual material of the Dead Sea Scrolls for the books of Proverbs is a problem. Whereas for some books of the Hebrew Bible (e.g. Isaiah and the Psalms) in the Dead Sea Scrolls there are many texts and fragments available, for Proverbs only a few fragments are extant. However, there is a multitude of mss that attest to MT's.

These fragments attest to some Hebrew variants. In Proverbs 1:32 ת ב וש מ there is a variant, תבשומ, testified to by 4QProv^a. According to De Waard (2008:31), it could be the result of metathesis. In 11:23, there is a variant, ה ד ב ע for ה ר ב ע that appears in De Rossi (De Waard 2008:41). In 14:35, MT readsותרבעו, which in De Waard’s (2008:45) opinion seems to be read in 4QProv^b as ותבעו. According to De Waard (2008:45), תבע in fact was the Vorlage for the LXX rendering. Chapter 8 contains a crucial variant, ןומא (artificer) for ןוּמא in verse 30 (De Waard 2008:14). In 11:7, there is a possible variant, for תלחותו, the reading תלהותו, which could entail a scribal error (see also 20:21).

4QProv^b is fragmentary, but according to De Waard (2008:6) in at least five instances supports MT. These are 13:6; 14:32; 15:2, 24 and 28. He also thinks that 3 other cases probably are related to MT, namely 15:22, 26 and 31. In connection with ותבעו in 14:35, there seems to be some relationship between 4QProv^b, LXX and Pesh.

The discovery of fragments of the Hebrew of Job in the Dead Sea Scrolls has unfortunately not thrown light on the issue at stake. There are only a few smaller fragments of Job available. See 4QpaleoJob^c in P.W. Skehan et al. (1992:155–157), Qumran Cave 4 IV Palaeo-Hebrew and Greek Biblical Manuscripts; Cf. also M. Abegg et al. (1999:590–593), The Dead Sea Scrolls Bible. The Oldest Known Bible Translated for the First Time into English.

It is a pity that these two books do not have more extant textual material available. The reason why there is a paucity of material of these two books is probably because they were not deemed significant books by the Essene community, at least not as much as Isaiah and the Psalms. In respect of LXX Proverbs, it would have been helpful as far as the difference in the order of chapters towards the end of the book is concerned, had there been additional Hebrew evidence. The only ‘evidence’ of a different text is the Septuagint, a text that has been freely rendered. As it stands, the hypothetical reconstruction of a different parent text remains just that.

The first chapter of Genesis has a complex history of origin and transmission (Schmidt 1964). The difference between MT and LXX testifies to that situation. There are numerous differences in the so-called Hexaemeron (vv. 1–31).

The following general structure can be reconstructed from the two versions (Cook 2001:317):

Focusing upon the Hebrew, one becomes aware that there is a discrepancy between the number of days (six) and the number of works (eight) completed during these days, which undermines this ideal structure. Moreover, this structure is not followed in all the days and/or works. The first day of creation does not have a Tatbericht. Likewise, the creation of man (vv. 26–30) has no ending formula following, neither Wort- nor Tatberichten. Finally, the name giving formulae are not found in connection with the third and fourth days. However, one can offer explanations for these apparent discrepancies. The non-existence of a Tatbericht in connection with the first day is natural because light was seen as the fundamental substance that came about solely by fiat (Westermann 1974:155). It is also possible that the author(s) of Gen 1 decided that the final work, man, should be included in the final expression of satisfaction in verse 31, דֹאְמ בוֹ֖ט. This could act as an explanation why verses 26–30 have no ending formula.

Turning to the Septuagint, there are prominent differences in comparison with the Hebrew. For one thing, the Septuagint has a much more closely knit structure than MT. Firstly, the equivalent of the Hebrew formula ןֵֽכ־יִהְיַֽו does not occur in verse 6 following the Wortbericht, but in verse 7 after the Tatbericht. This applies also to the sixth work (vs 20), where it is not used at all in the MT. The main issue in this regard is whether these differences should be attributed to a deviating Vorlage or to the translator. In general, Rösel (1998:64) argues that some of the conspicuous differences are the result of harmonising by the Greek translator. However, there are differences of opinion on this issue. Hendel (1998, 1999), Brown (1993, 1999) and Cook (2001) argue that the differences between the MT and LXX in Genesis 1 are primarily the result of a deviating Hebrew Vorlage. It must be said that some scholars have nuanced views in this regard. Hendel (1998:121), for one, does interpret some of the additions as the result of harmonisations. However, he is clear on the issue at stake: ‘In sum, it is more plausible and cogent methodologically to describe Gen-LXX, as, in general terms, a careful conservation of its Hebrew Vorlage than to explain each deviation from MT as the free composition of the Greek translator’ (Hendel 1999:34).

This argument is primarily based upon a translation technical assumption. According to Hiebert (2007:1), ‘The overall assessment of Genesis is that, lexically and syntactically, it is a strict, quantitative representation of its source text’. The Greek translation, in other words, is a relatively faithful rendering of the creation stories, which are related by means of the particle ἔτι in Genesis 2:9 and 19. These chapters are also harmonised in that Genesis 1:26 and 2:18 both read the plural ᴨοιήσωμεν, whereas the Hebrew has a singular form in Genesis 2.

These discrepancies were problematic for Jewish exegetes. Ber R iv:6, for example, contains a discussion of this issue: ‘He made – how remarkable! Surely it came into existence at God’s word’. This Rabbinic passage also includes attempts to explain why the ending formula for the Tatbericht (בוֹֽט־יִּכ םיִ֖הלֱֹא אְרַּ֥יַו) in verse 8 is omitted in MT. This applies also to verse 9 where TG Hagigah 12 (a) offers an explanation for the fact that the Tatbericht is missing. Clearly, the discrepancies were difficult for Jewish exegetes to understand, who then formulated some explanations. It should be remembered that for these exegetes, scripture was not discrepant at all.

What is remarkable is that many of the differences between LXX and MT are connected to water. I have argued that the Vorlage of MT differs from LXX’s parent text because water played too prominent a role in the creation process (Cook 2001:319). This was unacceptable to a conservative Hebrew redactor who simply adapted his Hebrew text. Seen as such, the Tatbericht in verse 9 was not added in LXX in order to harmonise apparent discrepancies. The change was rather brought about in the parent text of MT. As a matter of fact, the LXX in general represents the Urtext of Genesis 1. This interpretation has theological ramifications. The MT places more focus on the sovereignty of God by bringing about ideologically inspired adaptations. This is another example of the fact that there should be an interactive relationship between textual and literary criticism (Cook 2009).

This explanation is primarily based on internal considerations – the Septuagintal evidence be deemed as primary evidence. But to return to the topic of this study, there fortunately exists fragmentary Qumran material, recently published by Davila (1990:8f, 1994:61) that underscores the Septuagint reading of verse 9.

To be sure, the material is fragmentary and includes only three words, as well as the consonant שׁ in separate lines. However, the third line contains a crucial Hebrew word, הוקמ. This word corresponds with the Greek word συναγωγὴ in the LXX, which in turn corresponds with the Old Latin against MT, SP, Pesh, V, TO, Neof, the FT and 4QGen^b. These textual witnesses all read the equivalent of םוֹ֣קָמ.

The mentioned Greek word appears twice in verse 9, once in the Wortbericht and once in the Tatbericht as can be observed in the text below.

⁹Καὶ εἶᴨεν ὁ θεός Συναχθήτω τὸ ὕδωρ τὸ ὑᴨοκάτω τοῦ οὐρανοῦ εἰς συναγωγὴν μίαν, καὶ ὀφθήτω ἡ ξηρά. καὶ ἐγένετο οὕτως. καὶ συνήχθη τὸ ὕδωρ τὸ ὑᴨοκάτω τοῦ οὐρανοῦ εἰς τὰς συναγωγὰς αὐτῶν, καὶ ὤφθη ἡ ξηρά (Rahlfs & Hanhart 2006).

Hence, there existed a Hebrew Vorlage of the Tatbericht that underscores the interpretation I offered above. However, to be sure the Hebrew text is fragmentary, one of the golden rules of textual criticism is that textual evidence should be weighed and not counted.

This reconstruction has theological implications. Some Hebrew scribe, probably in Egypt where water plays a crucial role in creation stories – one example is the water god who inseminates the earth from above – deliberately removed the Tatbericht from Genesis 1:9, because it undermined the sovereignty of Elohim. In the formulation of a theology of the Septuagint version of Genesis, this could be a prominent topos (Cook 2010).

The books of Jeremiah have a complex transmission history, which is evidenced in the large number of extant texts. However, the differences between LXX and MT are significant; in this regard, it should also be remembered that the LXX of Jeremiah is circa 15% shorter than MT. A prominent issue is the differences in the order of chapters vis-à-vis MT and other textual witnesses. This phenomenon occurs elsewhere in the LXX too. LXX Proverbs is a case in point. However, there is a significant difference between these translated units. As will be demonstrated below, LXX Jer has additional Hebrew evidence which is lacking in Proverbs.

As far as the differences in the order of chapters are concerned, the most striking example concerns the so-called Oracles against the Nations. Whereas in the MT, these prophecies are found at the end of the book (chs 46–51), in the Septuagint they occur approximately in the middle (chs 26–32). There are also other larger order differences, such as chapter 50 in MT that is chapter 43 in the LXX, the passage that I will discuss.

These textual differences are of determinative importance for an understanding of the Jeremiah collections. Depending on the approach followed by the interpreter, the crucial question in this regard would probably be which of these arrangements should be taken as the original. If one concentrated on the reception of these different texts in diverse contexts, then the original (or earlier) version of the description would probably be of less importance. However, as I demonstrated above, I am of the opinion that, in the process of reconstructing texts, one gathers significant exegetical perspectives (Cook 2009) that can be helpful in arriving at an understanding of the different texts.

Various reasons have been suggested for these remarkable differences. Diverging Hebrew Vorlagen are taken as the reason for these deviations by most scholars (Janzen 1973; Tov 1992a). Rudolph (1968), on the contrary, ascribes to the translator practically all deviations from the MT. Few scholars will today agree with this point of view.

Janzen (1973), who was probably the first scholar to address systematically the issue of Qumran Jeremiah, concluded that the Hebrew Vorlage of the LXX was very conservative and that it had been created not very far removed in time from the Urtext of Jeremiah (Janzen 1973:68). He also deems the LXX to be a better witness to the original text of Jeremiah than the MT, which in his view is expansionistic, and the end result of scribal activity of many manuscript generations (Janzen 1973:68). According to him, the Oracles against the Nations initially circulated separately and were only added to the finished book later (Janzen 1973:115).

As far as the position of a different parent text is concerned, Tov (1985) presents a nuanced position. He distinguishes between literary and textual criticism, arguing that biblical books grew stage by stage throughout a period of several generations (Tov 1992a:316). According to him, even when a book seemed to have attained a completed state, it was often re-edited in a revised edition (Tov 1992a:316). Contextual factors accordingly led to the existence of more than one edition of specific books because, after they were completed, they were accepted as final and remained in use. However, later ‘revised editions’ of the books, which were intended to replace the earlier ones, were prepared and circulated. Because the process of substitution was not complete, the early editions did not simply disappear. Consequently, in Palestine the new edition that was later to become the MT.

In line with this explanation, Tov therefore maintains that Jeremiah LXX is the result of a different Hebrew Vorlage, in the sense of a different edition to the Massoretic parent text. Edition I contains the LXX source text, and Edition II that which later became the MT and which was almost identical to the edition contained in some of the ancient versions (Tg, Pesh and Vul), only partly replaced the earlier texts. They were still deemed holy texts by certain communities. Thus, the earlier editions remained in use in places that were not central from a geographical and sociological point of view, such as the Qumran repository of texts and the various manuscripts from which the Greek translation was prepared in Egypt (Tov 1992a:316).

The problem with this suggestion is that it is speculative. However, the situation changed dramatically in the wake of discoveries in the Judaean desert. This applies especially to 4QJer^d, which represents an older text and/or edition than the Vorlage of the MT.

The following fragment has been reconstructed by Tov (1992b:538):