We use cookies to enhance your experience on our website. By continuing to use our website, you are agreeing to our use of cookies. You can change your cookie settings at any time. Find out more
Select Bible Use this Lookup to open a specific Bible and passage. Start here to select a Bible.
Make selected Bible the default for Lookup tool.
Book: Ch.V. Select book from A-Z list, enter chapter and verse number, and click "Go."
  • Previous Result
  • Results
  • Look It Up Highlight any word or phrase, then click the button to begin a new search.
  • Highlight On / Off
  • Next Result

The Oxford Handbook of Biblical Studies Provides a comprehensive survey of Biblical scholarship in a variety of disciplines.


We now turn to the issue of methodology: how to realize the goal of textual criticism. We will deal with this matter both on a theoretical level and on a practical one by providing a few examples which may be illustrative.

It can not be denied that in practice the above-stated purpose will not be achieved in every instance in a convincing way, as there are many obstacles, such as the fact that the number of ancient witnesses in Hebrew is very limited indeed. Surely, we now have an impressive number of biblical texts from the Dead Sea area, but for most books this material is of a fragmentary nature. Besides, there are cases which are complicated, and cases which in the end are difficult to decide as to which reading or wording might have been the (more) original one. All this means that textual criticism can be carried out only by trial and error, including, as is usual in textual and historical research, an element of subjectivity.

To reach one's goal as a textual critic, there are roughly speaking two steps to be taken:

  • 1. First, one should collect the relevant data in a case, or in a group of cases, and analyse these data, each in its own context.

  • 2. Second, the evidence has to be evaluated in view of the question of which reading may be the preferred one, in the sense of the (more) original one.

As has been indicated above (sect. 1), the reason behind the first step may be either the availability of one or more variant readings in a case, or the idea that a reading or passage in the MT may be ‘corrupt’, or both. In the case of a difficulty, or supposed corruption, it may be that variant readings are not available (i.e. all ancient witnesses support MT, at least concerning its ketib). In such a case, one may consider a conjectural emendation.

The collection of the relevant data (first step) concerns the issue of variant readings (variant, plus, or minus) in a case. These readings are to be noted, be it in Hebrew in the case of a Hebrew witness, or in the language of one or more ancient translations. It may be useful to compare the witnesses involved with the help of a synoptic overview of a given passage (verse, or pericope). Having done so, the relevant data should be analysed, each variant reading in the context (verse, or pericope) of the witness involved against the background of the characteristics of a given witness. The purpose of this analysis is to find out whether one has to do with a variant reading which is not due to a translator, or scribe, but may reflect an ancient reading. However, in cases in which it turns out to be difficult to reach such a decision, the question should be left open, and the evidence should be treated as a virtual variant in the second step, when the data are to be evaluated in view of the question as to which reading may be the (more) original one.

In the next, and final section, of this paragraph I will address some aspects of method, illustrated by a few examples of cases in which variant readings are attested. (For a discussion of ‘conjectural emendation’, including examples, see Tov 2001: 351–69.)

As to the weighing of the evidence, several factors are involved, which may differ from case to case, such as linguistic, philological, exegetical, literary, and cultural considerations. As a rule, one factor will not be sufficient for an evaluation of a case. This also applies to the well-known rule lectio difficilior potior.

This rule may do in some cases, such as the following one: Gen. 17: 16. The MT reads: ‘I will bless her, and she shall be to nations; kings of peoples shall come from her’; cf. Targ. However, the other witnesses—LXX, Pesh., Vulgalte, and SamPent. (partly)—read ‘him’ (i.e. Isaac), ‘he’, ‘him’. The MT is likely to offer the more difficult reading in the sense of the preferable one.

The rule of the more difficult reading can thus be helpful, but in many instances one needs additional arguments or considerations. For example, Gen. 2: 2:

MT: God stopped his work ‘on the seventh day’; cf. Targ., Vulg.;

LXX: ‘on the sixth day’; cf. SamPent.; Pesh. (see also Jubilees 2. 16).

This is a well-known case of what might be considered a good example of lectio difficilior potior: the reading ‘seventh’ is clearly the more difficult one since, from the narrative context, one would expect ‘sixth’. Thus, the reading ‘seventh’ should be the preferred one. However, the reading ‘sixth’ is the reading which seems to make more sense, while the reading ‘seventh’ could be explained as due to a scribal error, triggered by the next clause in the verse (Hendel 1998: 33). Yet, as an additional factor, or consideration, one may point to the fact that, as the section about the sixth day (Gen. 1: 24–31) ended at v. 31, one would not expect a reference to that day afterwards, the more so since Gen. 2: 2–3 is about the seventh day. This is in favour of the reading ‘seventh’ as the original one. The idea of v. 2 may well be that, as soon as the seventh day began, God stopped working, and rested from his work during that day.

Another additional argument may be the rule that the non- harmonized reading is to be preferred to the harmonized one: for example, Deut. 34: 1–3. In this case, the MT contains a fairly detailed description of the promised land, which Moses was allowed to see; cf. LXX, Targ., Pesh., Vulg.; SamPent., on the other hand, offers a text which is much shorter than the MT: ‘from the land of Egypt to the great river, the river Euphrates, and to the Western Sea’. Since it represents a case of harmonization (see Gen. 15: 18), it can not be regarded as reflecting the more original text.

Or an additional argument may be of a linguistic nature: for example, Isa. 45: 2 as far as the reading הרורים in the MT is concerned. The other witnesses offer the following readings and renderings: 1Qa הררים, 1Qb הרורים, LXX ὄρη, Targ. ‘the walls’, Pesh ‘rocky place’ (cf. 40: 4 for MT העקב), Vulg. gloriosos (‘notables’). Hebrew הדורים presents a difficulty, since its meaning is uncertain and disputed (‘spiral roads’, ‘uneven places’?). Scholars have therefore argued that the variant reading attested by 1Qa and LXX (and 1Qb?), ‘mountains’, is to be preferred, as this makes good sense in the context. However, the difficulty with this solution is that the reduplicated plural (הררים) does not occur in biblical Hebrew in the absolute state (Koole 1997: 435). As has been argued by other scholars, MT Ketib may well represent the original reading if taken in the sense of ‘the walls’, in line with Akkadian dūru (cf. Targ.) (Southwood 1975: 802). Contextually, this would fit even better because of the ‘doors of bronze’ and ‘bars of iron’ in the rest of the verse.

An important matter concerns the characteristics of the witnesses involved, including the MT itself. For instance, for the weighing of variant readings, it is important to know that a text like 1QIsa-a represents a witness which, from a linguistic point of view, displays many readings of a secondary nature in comparison to the MT. However, although an overall view of a witness may help us in weighing the evidence, this is not to say that it is decisive in every case. On the contrary, although a text such as 1QIsa-a contains many readings which, linguistically speaking, are of a secondary nature, one should not exclude the possibility that this scroll also offers readings which are very significant and which may be preferred to MT. So, in principle, each case should be treated in its own right.

A good example of such a reading is found in Isa. 19: 18, where 1QIsa-a offers a variant (החרס, ‘[city of] the sun’; cf. 4QIsa-b) which is likely to be regarded the original reading, whereas the reading of the MT (ההרס, ‘[city of] destruction’) is of a secondary nature. (The variant reading of the LXX—ἀσεδεκ—is due to influence from Isa. 1: 26.) Another interesting case is the plus of אור ‘light’ in Isa. 53: 8, attested in 1QIsa-a and in 1QIsa-b, 4QIsa-d, and the LXX as well, over against the MT and the other, later, witnesses (Targ., Pesh., Vulg.). The fact that this variant is found in all the available early witnesses favours the idea that this reading represents an older text tradition than the MT. This is the more probable since 1QIsa-b, a conservative type of text, joins the other texts.

Another interesting case concerns the book of Jeremiah. The text of this book is attested in two versions, a longer one (MT, 4QJer-a, Theod., Aq., Sym., Targ., Pesh., Vulg.), and a shorter one (LXX, 4QJer-b). Many scholars are of the opinion that the shorter version represents the older one, whilst others hold the opposite view. Whatever theory one adheres to, also in this case each case should be treated, first of all, as an individual. For example, in Jer. 31: 39–40 a remarkable difference between the MT and the LXX concerns the beginning of v. 40 (MT): ‘the whole valley of the dead bodies and the ashes’, which is not found in the LXX (38: 39–40). It may well be that this phrase was not present in the parent text of LXX, but even then the question remains as to whether it represents a secondary addition, or whether it was left out for one reason or another. In the text attested by MT (cf. Targ., Pesh., Vulg.) the phrase about the dead bodies and the ashes is part of a context in which the city (Jerusalem) is described as a territory ‘sacred to the Lord’. Since a place with dead bodies and ashes (graves?) in Jerusalem as holy city creates serious difficulties for reasons of purity, it is easier to imagine that this phrase was left out than that it was added at a later date. As we know from Jewish sources of the Hellenistic era (4QMMT, Temple Scroll), the issue of the purity of Jerusalem as holy city was a matter of serious concern, particularly after the dramatic events in the seventies and sixties of the second century BCE (profanation of the temple).

Interestingly, LXX-Jeremia presents a version according to which the sacred area has been interpreted as ‘sanctuary’ (ἁγίασμα). Hebrew קרש has been taken here in the sense of ‘sanctuary’, and not as ‘sacred (to the Lord)’ as e.g. in LXX-Zech. 14: 20, 21 (ἅγιον). The ending of v. 39 in Greek which, unlike MT, refers to an encircling wall of precious stones (καὶ περικυκλωθήσεται κύκλω ἐξ ἐκλεκτω̑ν λίθων) is in line with this interpretation. It seems that a particular wall is meant here, viz. the περίβολος, the enclosing wall of the temple, just as in LXX-Isa. 54: 12: καὶ τὸν περίβολόν σου λίθους ἐκλεκτούς. Whether one would consider the translator the one who left out the phrase in v. 40, or not, it is clear that this phrase does not fit the interpretation of the text as attested by the LXX.

The rule that each case should be treated in its own right is an important one, but at the same time it is a rule which does not apply to every case. There are also text-critical cases which do not stand on their own, but are related to other cases in a book, or collection of books. An interesting example is Deut. 32: 8, which seems to be part of a network consisting of Deut. 32: 43, Gen. 46: 20, 27, and Exod. 1: 5 (Barthélemy 1978; Van der Kooij 1994).

Finally, I will deal with another aspect of method: viz. that of the relationship between text-critical and literary-critical methods. This has become a major issue in modern text-critical research, mainly in view of the issue of different ‘editions’ as far as attested by textual evidence (MT, Qumran, LXX). Methodologically speaking, one might argue that textual criticism should come first, and literary (or redactional) criticism next. In theory this makes good sense, but in practice it does not work. Considering the complexities involved in both types of research, it seems better to apply both approaches, each in its own right, that is to say, in interaction with each other, as will be illustrated in the following two examples.

(1) Joshua 20. The Old Greek of Joshua 20, the chapter about the cities of refuge, is much shorter than the MT (= Targ., Pesh., Vulg.): vv. 4–5, most of v. 6, and one expression in v. 3 (‘unintentionally’) are not attested in this version. The latter (LXX) reflects a text of Joshua 20 which is fully in line with the related passage in Num. 35: 11–12. This might evoke the idea of a harmonization, but since both texts belong to the same stratum, viz. the Priestly one (P), this is not plausible. As has been pointed out by scholars, from a literary-critical point of view, the pluses in MT are of a secondary nature (Cortese 1990: 79 f.). If so, the text attested by LXX corresponds to the results of a literary-critical analysis of MT.

As to the pluses in MT Joshua 20, the question arises as to whether they are part of a stratum, or layer, in the book. Rofé (1985: 145) and Tov (1992: 330) regard these pluses as ‘Deuteronomistic’. According to the latter, ‘[t]he layer of additions … in Joshua contains words and sections from Deuteronomy 19 which are meant to adapt the earlier layer to Deuteronomy—an assumption which is not surprising regarding the book of Joshua, whose present shape displays a deuteronomistic revision elsewhere in the book’ (Tov 1992: 330). However, the pluses actually contain elements from D (Deut. 19) and P (in v. 6, compare Num. 35: 25, 28) as well. Consequently, the reworking of the text is of a late date when P and D elements were easily combined (as e.g. is the case in the book of Chronicles, or in the Temple Scroll). It is therefore not plausible to assign the pluses in Joshua 20 to a stratum such as a ‘Deuteronomistic’ redaction of the book. Moreover, if this were the case, one would expect that LXX would attest at other ‘Deuteronomistic’ places in Joshua a pre- or proto-Deuteronomistic text, which actually is not the case, at least not in terms of sizeable differences. The pluses of Joshua 20 seem to have more of an occasional character than part of a redaction of the book as a whole.

(2) 1 Samuel 17. The text-critical issue concerning 1 Samuel 17, the story of David and Goliath, represents another most interesting example in terms of the relationship between textual criticism and literary criticism. The text of this story is attested in two versions, a longer one, MT (58 verses) and a shorter one, LXX (32 verses; the larger minuses are 17: 12–31, 41, 48b, 50–8). It is a well-known theory that the latter reflects the earlier version of the text, whereas the longer version (MT) is the result of a later expansion (Lust, and Tov, both in Barthélemy et al. 1986; for the opposite view, see Barthélemy, and Gooding, also both in Barthélemy et al. 1986). The crucial question is whether a literary-cricital analysis of the longer version would confirm, as in the case of Joshua 20, the idea that the shorter version may well be considered as the (more) original form of the text. There is reason to believe that such an analysis does not support the idea that the version reflected by LXX would converge with a source component of the MT version (Van der Kooij 1992; Aurelius 2002). This would imply that the LXX attests a version of the story which belongs to the history of reception and not to that of the redaction of the book.

  • Previous Result
  • Results
  • Look It Up Highlight any word or phrase, then click the button to begin a new search.
  • Highlight On / Off
  • Next Result
Oxford University Press

© 2020. All Rights Reserved. Cookie Policy | Privacy Policy | Legal Notice