Monday, March 7, 2011

A Text-Critical Analysis of Deuteronomy 32:35-37

Why is it that in Deuteronomy 32:35, the ESV reads "Vengeance is mine, and recompense (i.e., the noun form for repayment of vengeance)...," while the NIV reads "It is mine to avenge; I will repay..."? The difference in the second part is not due to translation style, but to a complicated set of textual variants in the beginning of this verse. The Massoretic Text (MT) for this passage reads לי נקם ושלם "Vengeance is mine, and recompense." The Samaritan Pentateuch (SP) reads instead ליום נקם ושלם "On the day of vengeance and recompense." The Septuagint (G) reads εν ημερα εκδικησεως ανταποδωσω for ליום נקם אשלם "On the day of vengeance, I will repay." Romans 12:19 and Hebrews 10:30, on the other hand, read εμοι εκδικησις εγω ανταποδωσω for לי נקם אשלם "Vengeance is mine, I will repay." There is one Qumran fragment 4QPaleoDeut(r) frg. 41, which contains on line 2 the top tips of the letters לי, and even though the text is not extant at the point in question, the editor (Skehan) claims that the traces below line 1 best fit the word ליום "On the day."

4QPaleoDeut(r) frg. 41 from DJD

Thus, all four possible combinations of readings for these two phrases are attested in ancient witnesses, a perfect example of how there were numerous textual differences even in early times. How do we resolve the differences, since there is ancient support for each combination of readings? The key is to recognize that these phrases together form a parallelism with the following לעת טמות רגלם "In due time their foot will slip" and that the variants must be evaluated together.

SP's ליום נקם ושלם "On the day of vengeance and recompense" is grammatically strained, as it lacks a main verb, and also has no verbal parallel with the following clause, so it is intrinsically unlikely. The NT's εμοι εκδικησις εγω ανταποδωσω for לי נקם אשלם "Vengeance is mine, I will repay"  is parallel within itself, and thus cannot function as a proper parallel in the context of Deuteronomy, where it would have to function as an awkward parallel within a parallel.

This leaves MT and G as possible candidates. לי נקם ושלם "Vengeance is mine, and recompense" is a complete sentence and grammatically tenable in combination with the following clause, but the parallelism is formally poor, even though one could argue for some sort of synthetic parallelism. Fortunately, one need not resort to this explanation, as G proves itself to be intrinsically the superior reading with its εν ημερα εκδικησεως ανταποδωσω for ליום נקם אשלם "On the day of vengeance, I will repay," which makes for tight formal, synonymous parallelism with the temporal adjunct and main verb of the following clause.

The reading ליום נקם אשלם "On the day of vengeance, I will repay" easily explains all the other variants with simple text-critical explanations, so there is no reason to doubt the originality of the G reading. The ום of  ליום נקם was accidently dropped off the end of the word when the scribe confused it with the final ם of נקם. The א of אשלם was replaced by ו by simple scribal error and/or as a secondary correction to better correspond to the variant לי נקם.

Thus, Deuteronomy 32:35 probably originally read with the Septuagint ליום נקם אשלם "On the day of vengeance, I will repay...," rather than the לי נקם ושלם "Vengeance is mine, and recompense" of the ESV or the לי נקם אשלם "It is mine to avenge; I will repay..." of the NIV. This is a fine examlpe of how formal parallelism can be used to solve text-critical problems and restore the text to a more pristine, original form.

For a more detailed analysis of the evidence and critique of contrary possibilities, as well as further discussion on the remaining simpler problems in the text of Deuteronomy 32:35-37, see my paper "A Text-Critical Analysis of Deuteronomy 32:35-37."


  1. I like the flow of logic, Drew.
    One thing came to my mind, though:
    Does this mean that the New Testament originally quoted a mistake? How do we deal with NT quotes that don't match the original Hebrew? Does it create a theological problem or is it acceptable as still being God-breathed? Basically, I'm asking if it creates a dilemma or if it isn't really a problem.

  2. James,

    The preceding two textual problems do indeed raise significant issues for the textual critic to work through with regard to the New Testament use of the Old Testament. For those who view the New Testament as inspired and authoritative, an understanding of the authority of New Testament citation text types is critical for a proper understanding of textual criticism. A couple of points and suggestions for future consideration are in order for determining how best to apply these quotations.

    First, the textual critic should be aware that G is the most common text type for New Testament citations. This should not be viewed as an automatic rule, however, as Deuterenomy 32:35-37 proved. Account should be made for each individual style and usage. In our discussion, we found that in Romans Paul actually generally prefers MT over G. But in Hebrews, the author almost always uses G. Even then, the text type is also complicated by the fact that it is often actually more accurately described as a hybrid of multiple text types. In the complicated textual environment, it must be recognized that there were many differing text types floating around in any given context that could influence a reading. This is compounded by the fact that often the writers of the New Testament would likely cite Scripture from memory, and not as a direct quote from a written biblical text.

    Whether these cited text types should be granted much authority or not in textual criticism might depend in part on one’s opinion concerning the theory of a sensus plenior. For those open to the idea that the New Testament can legitimately give a sense of a passage not present in the Old Testament text, a differing New Testament text type will likely cause fewer problems. Those who hold to a stricter grammatico-historical hermeneutic, however, must consider whether they can accept differences between the preferred Old Testament texts and their New Testament citations. What if the New Testament disagrees with the Old Testament in precise form, but still agrees in general tenor? Should that be seen as an authoritative affirmation of the precise form of a contrary text type? Or what if, by all appearances, the New Testament significantly alters the Old Testament text in both form and meaning? Can both be retained and held in tension? Must one be conformed to the other?

  3. Given the complex hermeneutical issues involved, perhaps a great degree of caution is appropriate in attributing authoritative status to New Testament citations in textual criticism. If the general tenor and meaning of the New Testament reference is in accord with that of the Old Testament and is not dependent on either reading against the other, it would be wise not to force the New Testament text type onto the Old Testament. The fact that the New Testament cites a text type other than MT does not necessarily mean the authors accepted that alternative reading as text-critically more accurate, especially if it is a side issue to the author’s main point. Not every passage should be pressed into use for textual criticism with apostolic authority, particularly when no such decision was intended by the author.

    When the New Testament author’s main thesis is dependent on one reading against another, the picture is more complicated. This dilemma indicates the author’s clear choice for (or at least access to) one variant over another. This gets into many of the hermeneutical and theological questions discussed above. But given the complexity of these questions and the exegesis of the passages in view, it may be preferable not to jump to speedy conclusions for the Old Testament text based on New Testament usages. Overall, caution should be practiced in claiming authority for New Testament citations of the Old Testament. The usages of the Old Testament by the apostolic church is still not well understood, and we run the risk of misreading the Old Testament when we read the New Testament into it prematurely. The Old Testament text-critic should generally, then, treat the New Testament citations of his passage as one among many pieces of evidence to be considered in toto.

    Thus, I would say that in general it would be dangerous to assume that NT citations indicate an authoritative, apostolic preference for a given text. The NT textual evidence for the OT is complex and often contradictory and reflective of the pluriformity of the text in the first century. There appears to have been a certain permissibility of textual diversity in the first century that did not bother the apostles. For the overwhelming majority of variants this really makes no big difference in the NT author's argument, but those few cases where the NT author relies heavily on apparently non-original readings may still require further explanation. Any other ideas about those? :)

  4. Roger Nicole has a good discussion under the title: "3. ACCURACY OF OLD TESTAMENT REFERENCES"

  5. Thanks, Drew. Your response has helped clarify my thinking.
    As to your parting question, the only instance that comes to mind are the references by Steven and Paul about the length of the sojourn of Israel in Egypt, a difference of 215 years in chronology. Paul seems to suggest the 430 years began with Abraham, where as the MT seems to say it started when Jacob went down to Egypt.
    Currently, I don't really have a lot of clarifying thoughts on that issue.