The New Testament Use of the Old Testament: Text Form and Authority Moisés Silva

The New Testament Use of the Old Testament: Text Form and Authority

Moisés Silva


Most Christians, I suspect, have at some point or another in their spiritual pilgrimage been seriously disturbed by the statement “Such-and-such a word is not found in the better biblical manuscripts.” To hear that kind of a remark for the first time with reference to the Holy Scriptures is most surely a rude introduction to textual criticism. It would be altogether different, of course, if we grew up learning, along with our multiplication tables, that the great plays of Aeschylus, for instance, have come down to us in a hundred or so corrupt manuscripts—the earliest of which was copied a millennium and a half after the original was composed—and that not infrequently we are at the mercy of imaginative scholars who must make educated guesses (“conjectural emendations”) as to what the tragedian might really have said, every available manuscript having failed us.

Even a superficial acquaintance with the processes involved in the transmission of any literature would lead us to expect, rather than be surprised at, the kinds of scribal errors present in all biblical manuscripts. And only a moment’s reflection would then persuade us of the implausibility, indeed the absurdity, that it might be otherwise. If an enemy of Christianity took it in hand to copy a biblical book, must God prevent him from distorting the material? More to the point: did our Lord ever promise that any fool who wished to copy some portion of Scripture would automatically be kept from error? Dare a modern printer, for that matter, forego proofreading when producing copies of the Bible?

Both skillful and unskillful scribes, but none infallible, have produced biblical manuscripts. It is merely in recognition of that fact that affirmations regarding the infallibility of Scripture are normally qualified by some such phrase as “in the original manuscripts.” Unfortunately, what is intended only as an acknowledgment of the realities of textual transmission has become for many a basis for ridicule. What is the use, we are told, of “inerrant autographs” if we do not have them? Why fuss about infallible documents if only errant ones are available to us? The simple answer is that, with regard to the bulk of Scripture, we know what the autographs said. To be more specific: the possibility of textual variation hardly ever affects those passages that are claimed by some to teach error or falsehood. If we made a list of the really controversial portions of Scripture (Adam’s creation, Paul’s teaching on women, etc.), we would be hard pressed to find any in which textual variation becomes a factor; that is, no one really doubts what the original writers said concerning these matters—though we may disagree on what they meant! Occasionally, let us grant, scholars might debate whether a particular error should be attributed to the original or to its transmission; but when it comes to virtually every issue of substance, we are as certain as it is possible to be regarding what the autographs said.1

It should be stressed, incidentally, that this last statement applies, with minor qualifications, to most of the ancient literature that has survived. For example, we have only about a half dozen truly valuable manuscripts of Aristotle’s Metaphysics, two of which date back to the tenth century (the others to the twelfth and fourteenth centuries). At hundreds of points the text is uncertain, but the vast majority of the variations do not materially affect the meaning of those passages; certainly no significant Aristotelian doctrine hangs on a textual variant!2 If this principle holds true even for the average piece of ancient literature, how shall we react to the richness of textual attestation characteristic of Scripture? Greek manuscripts of the New Testament are so early and numerous—to say nothing of supporting ancient translations and patristic quotations—that no other writing of antiquity begins to compare with it.3 Textual evidence for the Hebrew Old Testament is of a peculiar character, with a few books (notably Samuel-Kings) presenting special difficulties; still, the extraordinary faithfulness of medieval scribes (the Massoretes), combined with the independent evidence of the Dead Sea Scrolls on the one hand and of the Greek versions on the other, results in a truly formidable set of materials. Yet long before most of these materials were available, the Westminster divines had already sufficient evidence to speak of “the singular care and providence” of God in preserving His Word.4

The editors of this volume, however, have asked me to discuss how the concept of biblical authority is affected, not by textual transmission in general,5 but by those variations found in a very special subset of texts: Old Testament quotations in the New Testament. When studying these quotations, we are faced with the further complications of determining the transmission of one text within the transmission of another; and this factor happens to complicate matters in the manner of geometric progression—much in the same way as the addition of, say, a fourth child enlivens the household by approximately sixteenfold.

But this is not all. It would be more accurate to say that we are dealing with transmission-within-transmission-within-transmission. For the New Testament authors, writing in Greek, could not quote the Old Testament in Hebrew; they had to translate the Old Testament passages right then or, more often, use some existing Greek version. And, as it turns out, this “bridging” element is the most perplexing of all. Indeed, the textual transmission of the Greek Old Testament, when set against any other piece of ancient literature, almost certainly ranks first in complexity.6

In addition to all these textual problems, our difficulties are further intensified by the issue of apostolic hermeneutics. Even a superficial reading of the quotations found in the New Testament reveals that its authors felt no need to quote the Old Testament verbatim; their concern for interpretation and application led, through paraphrase, to formal changes (whether or not it led to substantive changes is a question that will soon occupy us). Not surprisingly, the mere presence of inexact quotations has been used to argue that the apostles had no view of “verbal inspiration”; but this argument can hardly be taken seriously. It is not even quite adequate to infer that the New Testament writers were more interested in ideas than in words.7 Loose quotations and paraphrases may readily be found also in the writings of modern authors who fight tooth-and-nail for the doctrines of inerrancy and verbal inspiration.

It has often been pointed out that the ancients had no clear-cut system of identifying direct quotations, yet it continues to be difficult for us to appreciate how deep and pervasive are the changes in writing convention that have taken place over the centuries. The availability of quotation marks, ellipses, question marks, exclamation points, dashes and parentheses, a variety of type faces, footnotes and precise bibliographic information, chapter-and-verse references, standard critical texts with apparatus—all of these create a frame of mind and a set of expectations far removed from those of ancient writers and their readers. To read a document with all the words run together, and with almost none of the punctuation helps listed above,




It may not be altogether possible, or even beneficial, for us to shed our modern perspective; but surely every effort should be made not to expect from an ancient writer what we are accustomed to read.

In any event, the fact that the New Testament authors often do not quote verbatim (for theological reasons),8 in addition to the textual problems previously mentioned, presents a need for detailed, technical research, much of which is yet to be done. Clearly, then, there can be no question in this one article of attempting a definitive solution to our problem. We know enough, however, to lay out the choices available to us, to reject certain facile solutions, and to suggest promising avenues of approach.


We can lay out some of these choices quite clearly by dealing with a relatively simple problem that has come up in discussions of inerrancy. In Hebrews 11:21 we read that Jacob, when he was at the point of death, blessed Joseph’s sons “and worshiped upon the top of his staff.” These words are an exact quotation from the Septuagint (LXX) of Genesis 47:31 (kai prosekynēsen Israēl epi to akron tēs rhabdou autou; Hebrews 11:21 omits Israēl because Jacob has already been identified in the first part of the verse). The difficulty arises when we check our English versions of Genesis, which say nothing about a staff; rather, they speak of Israel (Jacob) as bowing down on the head of his bed. Now the Hebrew word in question, which consists of the three consonants mṭh, can be vocalized in two different ways. At the time the LXX was produced, no system for marking vowels (or doubling consonants) was available; so the translator had to supply them in his mind (not a difficult task for a fluent speaker of Hebrew). The LXX translators assumed that the word was maṭṭeh, meaning “staff.” When the Masoretes (medieval Hebrew scribes) designed a system of vowel notation, they marked this word in Genesis 47:31 differently: miṭṭāh, “bed.” Most scholars seem agreed that the Masoretes were right (thus the rendering in the standard English versions) and the Greek translators wrong. It might appear, then, that the author of Hebrews, having been misled by the LXX, misquoted the Old Testament. If this interpretation of the facts is accurate, we would have to agree that the author has made a mistake (i.e., he is affirming something that happens to be false) and that we should in all honesty renounce any doctrine of inerrancy or verbal inspiration.

But this is not the only reasonable interpretation of the facts. Problems of this type, in fact, can be handled in many ways. Most of the possible solutions may not apply to this particular instance, but it will be useful to list them for illustration.

1. The first possibility to be considered is that of corruption in the transmission of the Hebrew text. In other words, when we encounter a textual difference between our Hebrew Bibles (= the Masoretic Text) and the LXX, we cannot assume that the former must be right. This consideration, however, is not really applicable to the problem in Genesis 47:31, where the difference is due to interpretive rather than textual reasons.9

2. Second, we may consider whether there is a textual corruption in the transmission, not of the Hebrew text, but of the LXX. Is it possible that the Greek translators rendered mṭh with the Greek word for “bed” but that later copyists changed the word to “staff”? The gifted and erudite Puritan theologian, John Owen, took this position, not only with reference to Genesis 47:31 but more generally whenever the LXX differed significantly with the Masoretic Text (MT) while agreeing with a New Testament quotation. His theory was that late copyists, familiar with the New Testament, were likely to change LXX manuscripts at those points where they differed from apostolic citations of the Old Testament.10 In principle, this kind of textual development is quite possible; indeed, we know that many variants in LXX manuscripts can be accounted for in some such fashion. Owen, however, had no access to a great deal of information available today and his proposed solution cannot be sustained. In particular, none of the many surviving LXX manuscripts of Genesis has the reading “bed”; we cannot doubt, then, that “staff” was the original Greek translation.

3. We are therefore led to consider the possibility of textual corruption, not in the Hebrew text nor in the LXX, but in the Epistle to the Hebrews itself. It is theoretically possible that the author used the Greek word for “bed” in conformity with the Masoretic understanding of Genesis, but that a later copyist, familiar with the LXX reading, suspected that his master copy was mistaken and therefore changed the word to “staff.” Once again we may say that, in principle, there is nothing far-fetched about such a process. This is precisely what has happened, for instance, in Acts 7:32, where the original reading, “the God of Abraham and of Isaac and of Jacob,” has been changed, in conformity with the LXX of Exodus 3:6, to “the God of Abraham and the God of Isaac and the God of Jacob” (this corrected reading appears in the great majority of New Testament manuscripts). In Hebrews 11:21, however, the evidence clearly points in a different direction, since no manuscript, to our knowledge, gives the variant “bed.”11 Furthermore, the author of Hebrews is known to cite the LXX even when it differs from the Hebrew. We are therefore as certain as it is possible to be that “staff” was the reading in the autograph of the epistle.

4. So far we have considered the possibility of textual corruption in each of the three documents that concern us and we have concluded that, though the attempt to isolate such corruption is legitimate and may work in some places, the problem in Hebrews 11:21 cannot be solved by an appeal to variation in the manuscripts. Assuming then that our texts are well established, we may turn to other considerations. For example, we may reconsider the difference between the MT and the LXX, this time focusing on the interpretive, rather than the textual, issue. Since both the translators’ text and that of the Masoretes read the same consonants, the question is not, which text is more accurate, but whose interpretation is more probable?

The Masoretes, we should keep in mind, were masters of the Hebrew language; so their opinions on lexical and grammatical problems should be set aside only on the basis of very weighty evidence. Still, they were not inspired, infallible scribes; furthermore, since they lived many centuries after the Scriptures were composed, some of their information was faulty. Almost all modern commentators on Genesis agree with the Masoretic interpretation, “bed.” It would be gratuitous, however, to suggest that nothing at all can be said in support of the LXX rendering. To begin with, it may be necessary to point out that the decision to translate “staff” can certainly not be attributed to ignorance or carelessness, for two verses later they translate Hebrew miṭṭāh with Greek klinē, “bed.” Second, we cannot forget that the Hebrew construction is not free of difficulty.12 It has been suggested, though one cannot easily verify it, that the Alexandrian (LXX) translators may have been influenced by Egyptian customs or by the fact that Hebrew beds had no head.13 However that may be, the Greek rendering was thoughtfully considered. Quite probably, the significance of Jacob’s staff, as brought out in Genesis 32:10, influenced their decision; it is also possible that they were reflecting some Jewish tradition that emphasized the staff (see below).

We must admit that the weight of probability lies on the side of the Masoretic interpretation. In other words, we can, but probably should not, try to solve the problem in Hebrews 11:21 by claiming that the LXX is correct after all. In particular, we should resist the strong temptation to assume that quotations in the New Testament, having been written by inspiration, must determine the form of the Old Testament passages; as we have already noted, the New Testament writers often introduce intentional changes. Still, the view that the LXX translators were right remains a possible, even if unlikely, solution.

5. Another proposed solution is to suggest that both the MT and the New Testament are correct. We can support this proposal in one of two ways: (1) by arguing that the author of Hebrews is referring to a different event, or (2) by suggesting, with John Owen, that Jacob bowed toward the head of the bed while supporting himself on his staff. We could also justify this approach by pointing out that the author does not explicitly say he is quoting (e.g., by the use of some introductory formula, such as in Hebrews 3:7, “the Holy Spirit says”). There are, however, two serious problems with this approach. In the first place, it gives the impression of grasping at a straw for the sake of harmonization. Although the picture of an aged patriarch leaning on his staff while bowing toward the head of his bed is convincing enough, and though in theory almost anything is possible,14 we should normally aim for hermeneutical sobriety! Second, it stretches one’s credulity to assume that the author received by revelation this bit of trivial information, especially in view of the fact that elsewhere he clearly depends on the LXX. Easy solutions of this sort must be resisted.

6. Assuming, then, that the MT is correct and that the author of Hebrews is intentionally quoting the LXX, we could argue that, precisely because he is quoting, a trivial error in the quotation should not be attributed to him. Calvin, Leupold, and, at least implicitly, many others apparently take this position.15 There are two distinct, though related, issues involved here. One is that of triviality; the other has to do with whether the author is affirming everything that is part of the quotation. In this section we will consider the second issue, which hinges on the distinction between conventional (or culturally conditioned) statements and positive teachings. It is agreed on all sides that a phrase such as “the four corners of the earth,” though used by the authors of Scripture, does not constitute an affirmation regarding the shape of the earth; rather, it belongs to a class of acceptable expressions found in all languages.16 A special subclass consists of quotations. The statement “There is no God” is made by an author of Scripture (Ps. 14:1), but no one attributes error or falsehood to the psalmist, since the comment is introduced by the words “The fool says in his heart.” Unfortunately, the problem in Hebrews 11:21 is not analogous, for the author of the epistle clearly regards his quotations as authoritative. Perhaps it is possible to argue that the authoritative element is only the act of worship on Jacob’s part and that the reference to the staff, while naturally included to remind the readers (themselves familiar with the LXX) of the event in view, is not part of the author’s affirmation. This approach is hardly distinguishable from 7:3 below.

7. What about the appeal to the trivial nature of the problem? Unhappily, such an appeal can suggest different things to different people. (1) It may reflect a relaxed (or, depending on one’s point of view, careless) attitude toward details. One must admit that there is a certain wholesomeness to the insistence that we should refuse to “major on the minors.” Yet one must ask whether an aversion to details really serves to protect “the majors.” The most insignificant sin in the life of Christ would destroy the central truth of His sinlessness; as far as that doctrine is concerned, one need not discover some gross, immoral act. Can we consistently hold the central truth that all Scripture is God’s very breath if we allow for a series of minor errors? When we add the further consideration that one man’s mountain is another man’s mole hill, it appears that this solution is no solution at all.17

(2) But to speak of trivialities may suggest a slightly different approach, namely, the supposed need to distinguish between revelational and nonrevelational matters in Scripture.18 In this approach, the distinction does not lie between important fact and insignificant detail, but between matters that affect salvation and those that do not. We are faced here, however, with a disturbing dichotomy between the religious (the sphere of faith) and the nonreligious (the sphere of scientific research), and it is very doubtful whether this general approach can be made to work.19

(3) More satisfactory is the position that takes “trivial” to mean something like nondidactic, so that we are back to the matter discussed above (section 6). In this section, however, my comments are not tied to the fact that Hebrews 11:21 happens to be a quotation. All responsible formulations of inerrancy have recognized that the Scriptures may, for example, use imprecise (i.e., round) numbers and that this fact does not disturb any meaningful understanding of infallibility.20 If a writer intends—so far as intention can be exegetically determined—to give precise numbers, yet fails to do so, his failure would indeed entail error or falsehood. If, on the other hand, he states (or implies), “I am giving round numbers for convenience,” we all accept his figures as “true.” Again, Jesus’ well-known comment that the mustard seed is the smallest of all seeds was certainly not intended as an absolute scientific statement, but as an appeal to such knowledge as was available to his Palestinian audience—the statement was true within that defined context.21

Such appears to be the consideration that led Calvin to comment that the writer of Hebrews accomodated himself (on a point that did not materially affect his teachings) to the Greek translations available to his Hellenistic Jewish readers.22 Of the possibilities we have thus far surveyed, this one may be the least objectionable; but it is not entirely convincing, since the problem in Hebrews 11:21 is not strictly analogous to the previous examples. If Jacob did not in fact “worship on the top of his staff,” then to say that he did should probably be interpreted as positive error rather than as imprecise statement. Furthermore, it is not at all clear that the reference to the staff was as insignificant as is often suggested. The possibility that the author saw some special significance in this detail leads us to our next consideration.

8. Otto Michel comments on this passage that both the LXX and the Epistle to the Hebrews “presuppose a haggada concerning Jacob’s staff” as the symbol of “wandering.”23 The haggada to which Michel refers is found in various forms in several rabbinic documents. We may summarize the legend as follows: Toward the end of the sixth day of creation, God created ten special objects, one of which was the rod, hewn from the sapphire of God’s throne. God gave this rod to Adam and it was transmitted to Enoch, Noah, Shem, Abraham, Isaac, and Jacob. (According to one tradition, Jacob took the rod away from his brother Esau. As he was fleeing from Esau, he used this rod, his only possession [cf. Gen. 32:11], to divide the waters of the Jordan.) Jacob brought it to Egypt and gave it to Joseph. After Joseph’s death, the Egyptians took it. Jethro, Moses’ father-in-law, stole it from the Egyptians and eventually it came into Moses’ possession. Aaron used it to perform miracles in Egypt. After miraculously blossoming (Num. 17), it was placed in the Ark of the Covenant. David used it in his encounter with Goliath, and it remained in the possession of his descendants until the destruction of Jerusalem, when it mysteriously disappeared. In the future, however, Elijah will give it to the Messiah.24

Now it is impossible to determine how much, if any, of this tradition existed at the time the LXX was composed; but that it was in substance known to the author of Hebrews and to his readers seems likely. Moreover, we may be sure that the significance of the staff did not escape the author: he makes reference to “Aaron’s rod that budded” (9:4) and, more important, to the royal messianic scepter (1:8, same Greek word, rhabdos, translating Ps. 45:6). Of immediate significance to our passage, however, is the emphasis that chapter 11 places on the wandering patriarchs who looked for a city without foundations (v. 10). They were believers who recognized their status as exiles and who desired a better, heavenly country (vv. 13–16).25 In the light of these facts, it appears doubtful that the reference to Jacob’s staff was insignificant or even incidental. Rather, it would seem that the author deliberately used the LXX rendering to make a theological point. In other words, he makes a conscious hermeneutical decision in order to lay stress on Jacob’s faith as a wanderer who longed for the messianic hope!

But does this approach solve our problem? Does it not rather intensify the difficulty? Do the New Testament writers feel free to change the original for theological reasons? Some conservative scholars would prefer almost any view to the one I am proposing. For example, S. Lewis Johnson, Jr., argues that the doctrine of inerrancy, without demanding verbatim quotations, “requires that the meaning the New Testament author finds in the Old Testament and uses in the New is really in the Old Testament.”26 Many evangelicals would agree with this statement, and perhaps it is accurate. But how do we determine whether or not it is accurate? Is it a criterion we have obtained from scriptural teaching or only an assumption that informs our way of thinking? At this point I merely wish to leave open the possibility that the author of Hebrews deliberately and for theological reasons used the LXX reference to Jacob’s staff and that such a use of the Old Testament—though it incorporates nonbiblical Jewish traditions—is not necessarily incompatible with the doctrine of inerrancy.

The rest of this chapter consists largely of an elaboration of that possibility. First, however, we need to assess where we have come in the argument so far. The reader, remembering my earlier reference to Hebrews 11:21 as a “relatively simple problem,” may by now have lost all confidence in my credibility. Let me stress, therefore, that if we were interested only in that particular quotation, the discussion would have been considerably briefer, since several of the proposals were either inapplicable or unworthy of serious consideration, while some others, though possible, seemed very unlikely; of the suggested solutions, only three (4, 7.3, and 8) should be seriously considered. But our interests are not limited to that verse. The solutions that I have rejected for it are often applicable to other passages. Indeed, there are some additional considerations (particularly “subtypes” of the first three proposals) that provide persuasive answers for a number of Old Testament quotations. For example, the presence of Targumic text forms in first-century Palestine is known to have influenced the New Testament writers.27 Again, the possibility of typological interpretation helps to solve still other problems.28 In short, we may have as many as fifteen or even twenty distinct approaches to the difficulties raised by the New Testament writers’ use of the Old Testament, though only a couple of them may be applicable at any one time. It is clear, therefore, that we should avoid quick solutions and simplistic answers. But it is no less clear that the tendency, in some quarters, to point out biblical difficulties without acknowledging the great complexity of the issues and without giving a sympathetic hearing to the wide range of possible solutions available to us, far from promoting intellectually honest inquiry, obstructs it.


During the past decade or two, biblical scholarship has shown a growing obsession with the issue of hermeneutics, a harmless enough word, but one occasionally used as a euphemism for “the skill of all but totally ignoring the Bible while appearing to accept it.” Although one may be excused for feeling irritated at the way the word is thrown about as the ultimate panacea, it would be a grave mistake to dismiss the issue altogether. It is so easy for us to read the evening paper and understand it—that is, interpret it accurately—that we tend to think of interpretation as an eminently simple process. In reality, we depend on a massive framework of assumptions slowly formed by innumerable experiences.29 As a result, those aspects of interpretation that appear to us to be the most obvious are often the ones that cause us the greatest difficulty. In particular, when we confront a text written by someone whose “framework of assumptions” differs significantly from ours, how can we possibly bridge the two? The attempt to answer that question is what hermeneutics is all about.

The contemporary discussion has relevance to this chapter for the simple reason that textual differences in Old Testament quotations cannot be accounted for by merely appealing to the mechanical aspects of text-critical work. Indeed, some of the most perplexing quotations are those that, without exhibiting textual changes, are used by the New Testament writers in ways foreign to us; for example, “Out of Egypt I have called my son” (Hosea 11:1) is applied in Matthew 2:14–15 to what appears to be a different and unrelated event. The literature on this general subject is immense,30 and we cannot hope to do full justice to it here. Nevertheless, the following suggestions may prove helpful for our concerns.

Perhaps the most basic, yet often overlooked, consideration is quite simply that the New Testament writers used the Old Testament at different times in different ways for different purposes. Although in principle everyone might agree with that statement, it seldom seems to make any practical difference. A scholar, for example, might offer a general judgment on whether Paul’s use of the Scripture is compatible with modern exegetical principles yet fail to specify just where in Paul’s letters that judgment happens to be applicable. For truly even the most negative critic will have to admit that at least some of Paul’s Old Testament quotations do not at all conflict with our grammatico-historical interpretation of the corresponding Old Testament passages.

In particular, we need to rid ourselves of the almost universal assumption that whenever a New Testament writer uses the Old Testament, he intends to prove a theological point. It may turn out, of course, that the assumption is correct, but this has to be demonstrated; does the doctrine of inspiration demand that the New Testament writers may refer to the Old Testament for dogmatic reasons only? Most of us balk at the possibility we are considering because it seems to open the door to abuse, as in the case of modern works that appear to make careless references to the Scriptures. For example, Edith Hamilton’s popular book on Greek culture abounds with allusions to biblical passages that, she believes, illustrate some point or another of ancient literature. In a discussion of Pindar’s poems to athletic heroes she quotes Hebrews 12:1; she describes Sophoclean tragedy in the words, “Lo, I come to do thy will”; Paul’s comment that “the invisible must be understood by the visible” is regarded as the basis of all great art; Ephesians 6:12 becomes illustrative of the view that the most divisive human conflicts are those waged “for one side of the truth to the suppression of the other side.”31 One’s immediate reaction is to smile condescendingly or express indignation that the true meaning of Scripture is being perverted. This judgment may well be valid, but is it so in every case?

Some time ago I asked a prominent evangelical scholar to evaluate an article I had written; with undue modesty, he prefaced his criticisms with the qualification that, in regard to the subject matter of the article, he was like ho anaplērōn ton topon tou idiōtou, “one who occupies the place of the unskilled,” a quotation from 1 Corinthians 14:16. Now it would not occur to anyone to accuse this scholar of violating grammatico-historical exegesis or of adopting a low view of Scripture. Is it then completely out of the question that the New Testament writers, intimately acquainted as they were with the Scriptures, might make relatively casual references to the Old Testament? If they did, these casual references would reveal nothing about their exegetical method, much less about their doctrine of inspiration.

It must be admitted, of course, that in the very nature of the case casual allusions are unlikely. The New Testament documents are not casual letters, but eminently serious and often urgent. It is indeed doubtful whether one could come up with a really persuasive example, though we might consider 2 Corinthians 13:1, where Paul, having announced his third visit to Corinth, quotes Deuteronomy 19:15, “Everything shall be established by the mouth of two or three witnesses.”32 That Paul is thereby threatening his opponents with a formal trial seems improbable; this view fails, among other things, to appreciate that the quotation (“three witnesses”) must be related to the remark about three visits. On the other hand, one must seriously doubt the suggestion (in spite of its impressive pedigree—it goes back to Calvin) that three visits by one person constitute fulfillment of the judicial requirement. Barrett comes closer to a satisfactory interpretation when he comments that “Paul does not use his quotation as a proof (there is no ‘as it stands written’—contrast 4:13; 8:15; 9:9), but says, in effect: You have had due warning, as prescribed; I am now about to take action.”33 Some have objected that this approach makes Paul appear whimsical in a very serious context. Yet the very seriousness of the matter leads Paul, more than once, to use irony and ridicule (cf. “super-apostles” in 11:5; note also Gal. 5:12). We should therefore not be too quick to dismiss the possibility that the reference to a third visit has reminded Paul of a well-known Jewish principle and that he proceeds to use it, not in order to demonstrate a theological or ethical point, but purely for its emotional impact. We should immediately remind ourselves, however, that even if this approach is correct, one could hardly accuse Paul of a loose view of Scripture or conclude that this passage betrays his principles of interpretation.

Yet again, we may consider the possibility that some quotations are neither purely illustrative (as I have just suggested) nor used for proof, but somewhere in between. Perhaps an expression such as “shifts in application”34 accurately describes what I have in mind here. Notice, as a possible example, Romans 15:21, where Paul quotes the words of Isaiah 52:15 (“what they were not told, they will see, and what they have not heard, they will understand”) in connection with his missionary method of working where Christ has not been preached. Now it is clear that the prophet himself did not have Paul in mind. Furthermore, we cannot here be satisfied with an appeal to divine authorship (“Maybe the prophet didn’t know, but God did”). While the principle may be valid in certain situations, it is often abused, and here it is simply inadequate. Granted that God in His omniscience and sovereignty foresaw Paul’s missionary labors among the Gentiles, that factor alone does not account for the peculiar application of the passage.

Paul is in effect explaining why he has not yet visited Rome: since a Christian community has already been established there, he must give priority to other, unevangelized fields. Do we need to believe that Paul considered the words in Isaiah a set of instructions concerning the arrangement of his missionary travels? Does Paul suggest that if he had gone to Rome the prophecy would not have been fulfilled? It would seem more reasonable to see in this passage something analogous to what modern Christians do when they apply to a present situation a biblical statement that clearly does not speak directly to that situation. Interestingly, Charles Hodge, while favoring the view that Paul was acting on the basis of a prediction, allows for my suggestion: “There is, however, no objection to considering this passage as merely an expression, in borrowed language, of the apostle’s own ideas.”35 In short, we should be open to the possibility that the New Testament writers may in some instances quote the Old Testament, not because they understand their use of it to correspond with the original writer’s intention (or even God’s original intention), but because their minds would naturally turn to passages that might have some kind of association, even a purely formal one, with the subject at hand.

Third, we may consider whether some passages that give every appearance of being dogmatic proofs, may turn out to reflect an opponent’s position. Herman Ridderbos, we should note, has persuasively argued that Paul’s hostile tone when speaking of the law should be understood in the light of the synagogue’s handling of the law.36 That factor alone alerts us to the possibility that a particular quotation may in fact be a short-hand pointer to the Judaizers’ interpretation of that passage. Consider Galatians 3:11–12, where Paul apparently opposes Habakkuk 2:4 to Leviticus 18:5, as though the Old Testament taught two mutually exclusive approaches to salvation. One of many attempts to solve the problem is to suggest that Leviticus 18:5 was something like the Judaizers’ motto, so that Paul’s use of that passage would have been understood by his readers as a reference to the Judaizing point of view.37 Even if we disagree with this particular interpretation of Galatians 3, is there a principial reason to set aside such an approach? From a slightly different standpoint, and with particular reference to Galatians 4:21–31, C. K. Barrett argues that Paul often refers to “passages that have been used by his opponents, correcting their exegesis, and showing that their Old Testament prooftexts were on his side rather than theirs.”38

We have seen, then, that the New Testament authors may well have used the Old Testament in at least three ways, none of which requires an understanding of the quotations as proof texts. Let us now turn our attention to those—no doubt a majority—that are indeed used for doctrinal demonstration. What do we learn from them regarding apostolic exegesis? Or, to ask a more specific question, What are the similarities and differences between apostolic and rabbinic interpretation?39

First of all, the differences. If we compare the bulk of quotations in the New Testament with the bulk of quotations in rabbinic literature, we cannot but be struck by the greater sensitivity of New Testament writers to the original context. As we will see shortly, the rabbinic approach may not be so faulty as it appears at first sight; nonetheless, a sympathetic study of the relevant New Testament passages reveals a notably sane, unfanciful method. Thus, C. H. Dodd, as is well known, argued that the early Christians normally quoted Old Testament passages

rather as pointers to the whole context than as constituting testimonies in and of themselves. At the same time, detached sentences from other parts of the Old Testament could be adduced to illustrate or elucidate the meaning of the main section under consideration. But in the fundamental passages it is the total context that is in view, and is the basis of the argument.40

Dodd himself believed that the New Testament contained “a fringe of questionable, arbitrary or even fanciful exegesis,” but that

the main line of interpretation of the Old Testament exemplified in the New is not only consistent and intelligent in itself, but also founded upon a genuinely historical understanding of the process of the religious … history of Israel as a whole.41

In other words, Dodd was not at all interested in the possibility of inerrancy—he believed that the New Testament authors did engage in some faulty use of Scripture. But instances of such use, being rare and atypical, were disregarded by him. In his judgment, the distinguishing feature was the New Testament authors’ reasonableness. With some qualifications, Dodd’s viewpoint has been widely accepted.42

But, we may ask, if the New Testament writers were characterized by sane interpretation, should not that very factor shed light on the apparent exceptions? Significantly, one of the reasons Barrett credits the Judaizers with the initial use of the Sarah-Hagar “allegory” is that Paul’s treatment in Galatians 4 is quite unusual for the apostle.43 We need not follow Dodd in speaking of fanciful exegesis unless we have first considered whether the few passages at issue may be classed in one of the three categories previously discussed or whether they may be understood along lines yet to be set forth in this article.

So much for the differences. While it is certainly true that the New Testament compares favorably with rabbinic literature, we must also recognize the deep affinities between the two. In particular, we are interested in the possibility that, as is the case with rabbinic hermeneutics, certain citations may simply reflect accepted interpretations. One need not find it difficult to see that, in the absence of chapter-and-verse references, a writer may quote a verse when he in fact means to draw our attention to a rather long passage. Would it be very different to quote a verse and mean not just the verse of the passage but an interpretive framework associated with this verse? One of the great difficulties in reading the Mishnah—particularly in the Hebrew rather than in a helpful translation—is the extreme compression of the argument. Brief citations from the Scriptures or from “the sages” are clearly intended to evoke sizeable theological structures apart from which the material is unintelligible.44

Some appreciation of that factor can temper our attitude toward rabbinic interpretation. It is very easy to reduce an audience to hysterics by reading some examples of Jewish exegesis.45 I suspect that, in some cases at least, the joke might be on us. For consider: How does one account for the fact that intelligent people, indeed scholars, should come up with interpretations that have nothing to do, so far as we can tell, with the original passage? The only reasonable answer is that the connection between text and interpretation does exist, only that time and tradition have obscured it.

The process of biblical interpretation is very ancient—it goes back to the Old Testament itself. For example, the author of the books of Chronicles certainly made use of the books of Samuel—Kings, and it is obvious that in using the earlier material he has reworked and interpreted it.46 Any interpretation or application of a passage, even the most legitimate, necessitates some degree of contextual separation. Thus, for instance, it may be valid to appeal to Ephesians 4:30 (“Do not grieve the Holy Spirit”) as proof for the personality of the Spirit, but in its context the verse is rather making a point about the Christian life; moreover, even this last statement is now found in the new context of “The New Testament Use of the Old Testament”!

Now the process of interpretation that began in Old Testament times grew through oral (and sometimes written) tradition over many centuries, with the New Testament writers right in the middle of the process. Even the wildest talmudic application may be the result of layer upon layer of deduction, so that we cannot see the connection between the beginning and the end. It may be that the New Testament writers, without rejecting such a hermeneutical process in principle, avoid its extreme application. It may be, indeed, that the difference between biblical and rabbinic interpretation, in this respect, is quantitative rather than qualitative.47 If so, we need not be surprised at the presence of difficult passages in the New Testament that reveal similarities to rabbinic methods.48


We may now return to our test case, Hebrews 11:21, and to the suggestion that the author, perhaps influenced by Jewish traditions about Jacob’s staff, deliberately used the LXX rendering to make a theological point. In light of the complex hermeneutical issues we have briefly surveyed, it would seem facile to dismiss this proposed solution on doctrinal grounds. Serious consideration should be given to the possibility that it does not conflict with the doctrine of inerrancy.

Several obstacles stand in our way of accepting this suggestion. For example, we find it difficult to dissociate the results of inspiration (an infallible Scripture) from the process of inspiration (involving that writer’s subjective experience). While everyone would agree in principle that inspiration is not deification,49 we often assume, unwittingly, that it does entail omniscience. With regard to Hebrews 11:21, do we need to assume, if we wish to preserve the doctrine of inerrancy, that the author must have been acquainted with the Hebrew and that he was therefore aware of the apparent discrepancy between it (or rather its probable interpretation) and the LXX? Do we need to assume, as John Owen did, that the author was conscious of revealing (or confirming) a historical fact not recorded in the Hebrew text, namely, Jacob’s use of the staff when he worshiped? These and similar questions remind us that many important issues are well beyond our power to explain, and our ignorance in turn suggests that accepting a particular interpretation must not depend on unproved assumptions regarding those issues.

Another obstacle consists of a different set of assumptions, this time with regard to historiography. The rise of “scientific” history-writing in modern times has conditioned us to expect certain standards of any author who purports to deal with facts. Consciously or not, we assume that the failure of an ancient author to conform to such standards necessarily implies falsehood or error. The point we are making is, of course, easily abused. The differences between ancient and modern methods have often been exaggerated, sometimes with the insinuation that ancient authors did not care at all about facts. It is perfectly evident, however, that the ancients recognized, and indeed emphasized, the difference between fact and fiction.50

What we need to appreciate is that in the reporting of historical events ancient authors were seldom concerned about verbatim quotations and precise information. Furthermore, it is clear that in Judaism haggadic interpretations were closely linked with those historical events—to refer to the one was almost necessarily to refer to the other. We ourselves may be reluctant to admit it, but even our own appeals to Scripture are in effect appeals to some interpretation of Scripture, even if it is a fairly obvious one (but obvious to whom?).51 May not the New Testament writers appeal to (a particular understanding of) a particular Scripture by citing, in effect, a generally received tradition? In short, Hebrews 11:21 may be interpreted merely as an appeal to a particular interpretation of Jacob’s act and not necessarily as an affirmation regarding the historicity of Jacob’s bowing on his staff.

One final objection to our proposed interpretation must be considered, and that is the question whether Hebrews 11:21 could then be regarded as a model for our exegesis. The significance of this question can hardly be overestimated. Indeed, the word “authority” in the title of this chapter faces its most serious challenge right here. For what is the use of affirming infallibility if the apostles’ very handling of Scripture proves invalid for us?

Articles and books that touch on this general subject often conclude with a remark to the effect that, though the New Testament writers’ use of the Old Testament can be appreciated and in some respects defended, we have no business copying it.52 Others, recognizing the fundamental inadequacy of such a conclusion, affirm that we must pattern our interpretation after that of the apostles, though one is not always clear how it is possible to do so.53 Perhaps more realistic is Longenecker’s attempt to deal with the seriousness of this issue without playing down the difficulties accompanying it. His answer54 is that, first, we cannot follow the apostles when their exegesis “is based on a revelatory stance”55 or when it is culturally conditioned (midrash, allegory, ad hominem); and second, we can reproduce their exegesis when it follows that line of grammatico-historical exegesis.

Longenecker’s approach advances the discussion in several respects; for example, in his recognition that the various Old Testament quotations in the New Testament do not follow a uniform pattern and in his reminder that we normally agree to distinguish between the descriptive and the normative in theological discussions. His formulation, however, appears to me to be less than persuasive. Indeed, we may question both the negative and the positive features of his solution—the former because it is too negative, the latter because it is “too positive.” (Perhaps another way to phrase my concern is to say that his formulation assumes too sharp a contrast between what biblical writers used to do and what we do today.)

With regard to the culturally conditioned statements, Longenecker’s formulations imply that the apostles’ handling of Scripture was indeed faulty at times, and this faultiness in turn suggests that their argumentation was sometimes invalid. Possibly this is not what he means (my own position, as we will see, may not be very far removed from his), but one may question the appropriateness of his description.

But why should we further suggest that his positive evaluation is “too positive”? In a sense, Longenecker gives the apostles more credit than is wise or necessary. Modern exegesis involves a series of skills that in the very nature of the case, were unavailable before modern times. For example, we cannot very well do responsible exegesis without relying on precise, scientific philological information. Further, we are all agreed that the Bible is no more a textbook of philology (or of textual criticism, form criticism, source analysis, etc.) than it is a textbook of zoology. The fact that the biblical writers (though authoritative) had a “prescientific” understanding of the animal world is no reason for us to remain prescientific and burn our modern zoology textbooks. Similarly, it is hardly a troubling thought that the apostles had no scientific literary hermeneutics—unless we decide that inspiration does entail omniscience after all. The authority and validity of apostolic interpretation, therefore, do not depend on its conformity to modern exegetical method.56

Perhaps I can best illustrate my point with a contemporary, though still “prescientific,” example. On Saturday evening, May 17, 1980, riots broke out in a Black community in Miami, Florida. The following morning Mount St. Helens erupted, producing one of the most sensational natural phenomena of our generation. Almost immediately, political cartoons all over the country took advantage of this fortuitous combination and editorials appeared with such titles as “A Black Cloud Over Miami.” One might, I suppose, protest that the two events had no scientific connection—how could journalists be so ignorant? But of course, we all understand, accept, and even appreciate such associations. They may have no scientific validity, yet they have validity (even in our scientific age) at a different level, functioning usefully within established social and literary expectations.57

Well, then, if God wished to reveal something of the significance of the Old Testament through His inspired apostles, would He do so through “scientific” methods that were to take twenty centuries to develop and would therefore have been totally incomprehensible to first-century readers? Might He not rather use those very associations and interpretive clues that would awaken the intended human response? Just as the use of imperfect human languages like Hebrew and Greek can prove an adequate channel for conveying divine truth unmixed with error,58 so does prescientific apostolic exegesis serve to communicate, infallibly, the teaching of the Old Testament.

It also follows, however, that just as we are not required to write in Greek (as the author of Hebrews did) in order to produce valid exegesis, so are we not required to use Jewish haggadic tradition to communicate our understanding of Scripture. But to say that much is not for a moment to suggest that we can dispense with apostolic interpretation. Quite the contrary. If we refuse to pattern our exegesis after that of the apostles, we are in practice denying the authoritative character of their scriptural interpretation—and to do so is to strike at the very heart of the Christian faith.

What I wish to point out is that adherence to this basic principle does not entail indiscriminate imitation, any more than faithfulness to the pattern of apostolic evangelism requires us to board ships rather than airplanes or to make Antioch the headquarters for modern missions.59 While we are committed to discover and pursue the interpretive framework that characterized apostolic interpretation, we need not suppose that such a commitment compels us to reproduce it in all its features. With particular reference to Hebrews 11:21, our decision to accept the solution proposed in this chapter should not depend on whether we are required to reproduce the precise method of exegesis used by the author of the epistle.60

In conclusion, I may anticipate a general objection to my suggestions. Are we not playing with words when we insist on “inerrancy” or some equivalent? Are we not stretching these words beyond their reasonable limits—or, conversely, narrowing the sense of “error” so that practically nothing will fit its meaning? This kind of objection, though often heard, and though it may evoke considerable sympathy, reveals a failure to understand the concerns that gave rise to the modern formulations of inerrancy in the first place. For clearly, our better theologians assumed from the beginning that because of the human form in which Scripture was given the claim that the Scriptures teach only truth would be measured according to the usual canons of veracity, not according to some artificial criterion of “absolute truth” (i.e., absolute precision, exhaustive and verbatim reports, etc.). It makes no more sense to accuse the New Testament writers of error or falsehood in their use of the Old Testament than to hurl the same charge at the journalists who linked the Miami riots with Mount St. Helens’ eruption.


1 To be sure, a number of related and interesting questions can be asked, such as, Why would God be concerned about inerrant autographs if they were not to survive? That very common question has probably been discussed more intensively than it is worth. Particularly interesting, and difficult, is the problem of identifying what we mean by “autograph” when the book in question has had a long history of development. For example, is the autograph of the Pentateuch the shape it had during Moses’ lifetime (assuming Mosaic authorship), or its shape after his death when some “updating” took place, or its “canonical shape” (to use a fashionable expression)? This question is extremely complicated and calls for serious discussion.

2 The reason for this phenomenon lies in the redundancy inherent in language. The meaning of most sentences can be captured even if we miss individual words; also, the main points of a discourse come through though we may occasionally daydream. I have treated this matter in Biblical Words and Their Meaning (Grand Rapids: Zondervan, 1983), chap. 6.

3 Note some relevant statistics in Bruce M. Metzger, The Text of the New Testament, 2nd ed. (Oxford: The University Press, 1968), pp. 31–35.

4 The Westminster Confession of Faith, I.viii.

5 This subject has been treated by many writers. Note John H. Skilton, “The Transmission of the Scriptures,” in The Infallible Word, ed. N. B. Stonehouse and Paul Woolley, 3rd ed. (Philadelphia: Presbyterian and Reformed, 1967), pp. 141–95. More recently, and with special emphasis on Old Testament problems, Douglas Stuart, “Inerrancy and Textual Criticism,” in Inerrancy and Common Sense, ed. Roger R. Nicole and J. Ramsey Michaels (Grand Rapids: Baker, 1980), pp. 97–117.

6 The primary reason for this complexity is that several Greek translations and revisions of the same Old Testament books were produced from the fourth century b.c. to the fourth century a.d., and that surviving manuscripts contain “mixed texts,” making it almost impossible to identify the original translation. Cf. Robert A. Kraft’s discussion of the “Earliest Greek Versions” in The Interpreter’s Dictionary of the Bible, Supplementary Volume (Nashville: Abingdon, 1976), pp. 811–15.

7 This inference does contain a measure of truth. Cf. Donald A. Hagner, “The Old Testament in the New Testament,” in Interpreting the Word of God: Festschrift in Honor of Steven Barabas, ed. Samuel J. Schultz and Morris A. Inch (Chicago: Moody, 1976), pp. 78–104, esp. p. 90.

8 It should be noted that we are not interested here in exact quotations that happen to be used in unusual ways—though this subject will necessarily demand our attention later on. See O. Palmer Robertson, “Genesis 15:6: New Covenant Expositions of an Old Covenant Text,” Westminster Theological Journal 42 (1979–80): 259–89, esp. p. 279.

9 That is, the Hebrew text used by the LXX translators had the same three consonants found in the Masoretic text (mṭh). The translators and the Masoretes disagreed on which of two possible meanings was conveyed by that combination of consonants. A good example of textual variation is Jeremiah 23:9, “I have become like a drunken man.” The work “drunken” translates the Hebrew consonants škwr (vocalized šikkôr). The LXX, however, must have used a Hebrew text that read šbwr (vocalized šābûr), for they translated it with a Greek word that means “broken.”

MT Massoretic Text

10 John Owen, An Exposition of Hebrews, 7 vols. in 4 (1855; reprint of Goold ed., Marshallton, Del.: National Foundation for Christian Education, 1969), 1:114–17.

11 More often than not, New Testament copyists resisted assimilation with the LXX. See Kenneth J. Thomas, “The Old Testament Citations in Hebrews,” New Testament Studies 11 (1964–65): 303–25, esp. pp. 303–4, n. 7.

12 E. A. Speiser says that the usual translation and the LXX rendering are “equally implausible” (The Anchor Bible: Genesis [N.Y.: Doubleday, 1964], p. 327); cf. also G. von Rad, Genesis: A Commentary, rev. ed. (London: SCM, 1972), p. 414, “The meaning of the gesture … is not quite clear.”

13 See especially Moses Stuart, A Commentary on the Epistle to the Hebrews (London: Fischer, Fischer and Jackson, 1834), pp. 492–93. It is true that the expression roʾš hammiṭṭāh does not occur elsewhere in the Bible. The view that the LXX has translated correctly is supported by a few other conservative scholars, such as Philip E. Hughes, A Commentary on the Epistle to the Hebrews (Grand Rapids: Eerdmans, 1977), p. 489.

14 For an interesting, real-life example of apparently contradictory reports that were surprisingly harmonized when additional information became available, see the Summer 1980 issue of Update (published by the International Council on Biblical Inerrancy), p. 3. Harmonization is a standard historical tool—I am concerned here about its possible abuse. See also my article, “Ned B. Stonehouse and Redaction Criticism,” Westminster Theological Journal 40 (1977–78): 77–88, 281–303, esp. p. 79.

15 John Calvin, Commentaries on the Epistle of Paul the Apostle to the Hebrews, tr. John Owen (Grand Rapids: Eerdmans, 1948), pp. 290–91; this particular edition includes an introduction by Ned B. Stonehouse, who commends Calvin precisely because of his interpretation of 11:21. H. C. Leupold, Exposition of Genesis, 2 vols. (1942; reprint, Grand Rapids: Baker, 1950), 2:1141–42. Note also the standard conservative commentaries on Hebrews (e.g., Delitzsch, Westcott).

16 For example, “The sun rose this morning at six o’clock.” Some writers argue (e.g., H. Ridderbos, Studies in Scripture and Its Authority [Grand Rapids: Eerdmans, 1978], p. 30) that the parallel is not real, since the biblical writers did believe some of these things; but to raise the objection that these writers were in fact limited is to confuse inerrancy with omniscience. (See also below, n. 21.) I might add in this connection that Daniel P. Fuller’s comments on “Benjamin B. Warfield’s View of Faith and History,” Bulletin of the Evangelical Theological Society 11 (1968): 75–83, esp. pp. 80–81, do not take into account Warfield’s own clear qualification: “No one is likely to assert infallibility for the apostles in aught else than in their official teaching. And whatever they may be shown to have held apart from their official teaching, may be readily looked upon with only that respect which we certainly must accord to the opinions of men of such exceptional intellectual and spiritual insight” (emphasis mine). See his collection of articles, The Inspiration and Authority of the Bible (Philadelphia: 1893; reprint ed., Presbyterian and Reformed, 1964), pp. 196–97.

17 One cannot help but recall the scorn with which the serious christological controversies of the fourth century have been described as “the whole world convulsed over a diphthong” (homoiousios). Interestingly, Stephen T. Davis, The Debate About the Bible: Inerrancy versus Infallibility (Philadelphia: Westminster, 1977), first defines infallibility by saying that “in matters of faith and practice [the Bible] does not mislead us” (p. 16), but later redefines it to include only “matters that are crucially relevant to Christian faith and practice” (p. 118, emphasis mine). I suspect one could honestly present good arguments for an estimate that only 5 percent of the Old Testament material is crucially relevant for Christian faith and practice (for example, in Genesis 1 only verses 1 and 26 might qualify). Even more interestingly, Davis himself considers an error the biblical statement that God commanded the Israelites to kill the Canaanites (pp. 96–98). But is not the ethics of killing crucially relevant for Christian practice? It would seem that, according to Davis, the Bible does mislead us in gravely significant matters of practice. So much for the complaint that inerrantists worry too much about trivial details.

18 This formulation is characteristic of Daniel P. Fuller; cf. his articles on Warfield (see above, n. 16) and on “The Nature of Biblical Inerrancy,” Journal of the American Scientific Affiliation 24 (1972): 47–51. Unfortunately, it is not at all clear what he means. His attempts at clarification in JETS 16 (1973): 67–69 would suggest that his view belongs, not here, but in the next item, 7.3. If so, his terminology is unfortunate. Cf. also Davis’s evaluation, The Debate About the Bible, pp. 37–48.

19 Though overstated, Montgomery’s criticisms on this matter are generally on target; see God’s Inerrant Word: An International Symposium on the Trustworthiness of Scripture, ed. J. W. Montgomery (Minneapolis: Bethany Fellowship, 1974), pp. 29–30.

20 Cf. the quotations gathered by Ned. B. Stonehouse in Origins of the Synoptic Gospels: Some Basic Questions (Grand Rapids: Eerdmans, 1963), p. 110n. In what follows, my use of the word “intention” should be distinguished from that of Daniel P. Fuller, who emphasizes the general salvific purpose for which the Scriptures were given. My only concern is that the meaning of specific passages must be exegetically determined before we proceed to state what those passages infallibly teach.

21 Even scholars hostile to any notion of infallibility have long recognized this elementary hermeneutical factor. One must therefore consider it a disturbing retrogression in the contemporary debate when Davis insists that Jesus’ comment about the mustard seed contains “at least some sort of error” (Debate, p. 101). This statement is not only bad exegesis; it also erects a straw man by implying that inerrancy entails some artificial (= unnatural to human communication) standard of absolute truth.

22 See above, n. 15.

23 Der Brief an die Hebräer, 11. Aufl. (Göttingen: Vandenhoeck und Ruprecht, 1960), p. 270. Cf. also C. Spicq, L’Épître aux Hébreux, 2 vols. (Paris: J. Gabalda, 1952–53), 2:355, who states that, according to the LXX, the staff was a sign that Jacob professed faith in the future city—“or, at least, that his people would leave Egypt to possess the Promised Land.”

24 This material is most accessible in Louis Ginzberg, The Legends of the Jews, 7 vols. (Philadelphia: The Jewish Publication Society, 1912–28), 1:83, 347; 2:291ff.; 3:306–7, with corresponding notes; see also 2:34 for the view that the staff was a symbol of messiahship. Cf. Friedrich Schröger, Der Verfasser der Hebräerbriefes als Schriftausleger (Biblische Untersuchungen 4; Regensburg: Friedrich Pustet, 1968), p. 221.

25 Cf. Ernst Käsemann, Das wandernde Gottesvolk. Eine Untersuchung zum Hebräerbrief, 2. Aufl. (Göttingen: Vandenhoeck und Ruprecht, 1957), esp. chap. 2, though the author does not comment on 11:21.

26 S. Lewis Johnson, Jr., The Old Testament in the New: An Argument for Biblical Inspiration (Grand Rapids: Zondervan, 1980), p. 66. He makes this statement concerning the use of “body” (instead of “ears”) in Hebrews 10:5. In a note, p. 104, he mentions the view of Simon Kistemaker, an Evangelical, that the author of Hebrews used the LXX precisely because it lent itself to his interpretation. Johnson responds that this suggestion “raises questions concerning the biblical doctrine of inspiration, to which [Kistemaker] has not addressed himself.”

27 One common example is the quotation of Isaiah 6:10 in Mark 4:12; among other details, the verb “be healed” is changed to “be forgiven,” in correspondence to the Targum (Aramaic translation) of Isaiah. With regard to the importance of the Syriac tradition, see Earl Richard, “The Old Testament in Acts: Wilcox’s Semitisms in Retrospect,” Catholic Biblical Quarterly 42 (1980): 330–41.

28 Cf. Johnson, The Old Testament in the New, chap. 4, and the literature cited by him. O. Palmer Robertson, “Genesis 15:6: New Covenant Expositions of an Old Covenant Text,” Westminster Theological Journal 42 (1979–80): 259–89, says concerning the quotation of Hosea 11:1 in Matthew 2:15 that “the ‘fulfillment’ envisioned involves the ‘bringing of fruition’ of a principle of redemptive history that had an earlier manifestation” (p. 285).

29 For a comprehensive review, see Anthony C. Thiselton, The Two Horizons: New Testament Hermeneutics and Philosophical Description (Grand Rapids: Eerdmans, 1980), esp. chap. 4.

30 Note the survey by Anthony J. Saldarini, “Judaism and the New Testament,” in The Bible and Its Modern Interpreters, III (forthcoming). The basic issues are clearly laid out by E. Earle Ellis, “How the New Testament Uses the Old,” in New Testament Interpretation: Essays on Principles and Methods, ed. I. Howard Marshall (Grand Rapids: Eerdmans, 1977), pp. 199–219. Also useful is C. K. Barrett, “The Interpretation of the Old Testament in the New,” in The Cambridge History of the Bible, vol. I, ed. P. R. Ackroyd and C. F. Evans (Cambridge: The University Press, 1970), pp. 377–411.

31 The Greek Way (1930; reprint, New York: Avon, 1973), pp. 69, 187, 208, 247.

32 This example, among others, has been used by Anthony T. Hanson, Studies in Paul’s Technique and Theology (Grand Rapids: Eerdmans, 1974), pp. 173–74, though his approach is in some respects different from mine.

33 C. K. Barrett, A Commentary on the Second Epistle to the Corinthians (New York: Harper and Row, 1973), p. 333.

34 I mean by this phrase something quite different from C. H. Dodd’s similar term, “transposition,” in According to the Scriptures: The Sub-Structure of New Testament Theology (Digswell Place: James Nisbet, 1952), p. 130.

35 Commentary on the Epistle to the Romans (1886; reprint, Grand Rapids: Eerdmans, 1968), p. 441. Concerning the quotation in Romans 10:18, Hodge similarly remarks that Paul “is not to be understood as quoting the Psalmist as though the ancient prophet was speaking of the gospel. He simply uses scriptural language to express his own ideas, as is done involuntarily almost by every preacher in every sermon” (p. 349).

36 Paul: An Outline of His Theology (Grand Rapids: Eerdmans, 1975), pp. 154ff.

37 See Daniel P. Fuller, Gospel and Law: Contrast or Continuum? (Grand Rapids: Eerdmans, 1980), p. 98; see also Douglas J. Moo’s critical review in Trinity Journal 3nS (1982): 99–103.

38 “The Allegory of Abraham, Sarah, and Hagar in the Argument of Galatians,” in Rechtfertigung. Festschrift für Ernst Käsemann zum 70. Geburtstag, ed. J. Friedrich et al. (Tübingen: Mohr, 1976), pp. 1–16, esp. p. 6.

39 Strictly speaking, “rabbinic interpretation” describes the approach evident in works that date no earlier than the second century of our era. We are not interested, however, in whether a particular method (say, one of the supposed rules of Hillel) or line of interpretation existed in the first century; rather, we are concerned with a basic approach to Scripture. No one imagines that “the rabbinic mind” appeared ex nihilo after a.d. 70; indeed, the general features that have a bearing on our topic go back at least to the second century b.c. Note J. Weingreen, From Bible to Mishna: The Continuity of Tradition (Manchester: The University Press, 1976), passim, and my review article, “The Pharisees in Modern Jewish Scholarship,” WTJ 42 (1979–80): 395–405, esp. pp. 402–3. Note also J. Weingreen et al., “Interpretation, History of,” in The Interpreter’s Dictionary of the Bible, Supplementary Volume, esp. pp. 436–48; and G. Vermes, “Bible and Midrash: Early Old Testament Exegesis,” in The Cambridge History of the Bible, 1:199–231.

40 Dodd, According to the Scriptures, p. 126.

41 Ibid., p. 33.                                            

42 Note, however, James Barr, Old and New in Interpretation: A Study of the Two Testaments (New York: Harper and Row, 1966), p. 142, n. 2. Ernst Käsemann, Commentary on Romans, trans. Geoffrey W. Bromiley (Grand Rapids: Eerdmans, 1980), p. 86, accuses rabbinic exegesis and Paul of taking passages out of context.

43 Barrett, “The Allegory,” p. 10.

44 Herbert Danby, The Mishnah (Oxford: The University Press, 1933), p. xxvi, n. 2, tells us that the ten opening words of Arakhin 4 required, for translation, seventy-seven English words (even so, many English readers would no doubt find the translation quite perplexing). For a taste of the complexities involved, see Jacob Neusner, Invitation to the Talmud: A Teaching Book (New York: Harper and Row, 1973).

45 Franklin Johnson argued that the resemblance between Jewish and biblical interpretation “is chiefly in appearance; when the reader pierces below the surface, he finds but little of it; and it vanishes wholly when he searches in the New Testament for the obscurities, the superstitions, the cabalisms, the puerilities, the absurdities, the insanities, which stare at him from every page of the rabbinic interpretations of the sacred writings”; see The Quotations of the New Testament from the Old, Considered in the light of General Literature (Philadelphia: American Baptist Publication Society, 1896), p. 379. More sober scholars, while avoiding such extremes, still tend to overemphasize the differences; cf. the highly regarded study of Robert H. Gundry, The Use of the Old Testament in St. Matthew’s Gospel, With Special Reference to the Messianic Hope (Supplements to Novum Testamentum 18; Leiden: Brill, 1967), p. 205, which speaks of Qumran and rabbinic interpreters as “supremely oblivious to contextual exegesis whenever they wish.” For a different viewpoint, see J. W. Doeve, Jewish Hermeneutics in the Synoptic Gospels and Acts (Van Gorcum’s Theologische Bibliothek 24; Assen: Van Gorcum, 1954), p. 89.

46 Notice how the phrase, “the king went to Gibeon” (1 Kings 3:4, right after a negative comment on Solomon’s worship) becomes a five-verse apologetic in 2 Chronicles 1:2–6, where emphasis is put on Solomon’s obedience to the instructions of Leviticus 17:1–7. (I owe this observation to Dr. Raymond B. Dillard, who is preparing a commentary on 2 Chronicles; cf. his article, “The Chronicler’s Solomon,” Westminster Theological Journal 43 [1980–81]: 289–300.) Important works on the chronicler’s theology include T. Willi, Die Chronik als Auslegung (Göttingen: Vandenhoeck und Ruprecht, 1972), and H. G. M. Williamson, Israel in the Books of Chronicles (Cambridge: The University Press, 1977). More controversial is the question whether the Evangelists felt free to modify Jesus’ deeds and teachings with a view to bringing out their significance; note my article “Ned B. Stonehouse,” Westminster Theological Journal 40 (1977–78): 281–303, esp. pp. 289ff., and the massive development of this idea by Robert H. Gundry, Matthew: A Commentary on His Literary and Theological Art (Grand Rapids: Eerdmans, 1982).

47 Needless to emphasize, there are qualitative differences, particularly in the christological perspective of New Testament writers. My statement does not have reference to all aspects of interpretation, but only to the question of “distance” between text and interpretation. Moreover, I do not wish to deny the possibility that some rabbinic interpretations were developed “to justify long-established usage not expressly ordained or permitted by Scripture” (Danby, The Mishnah, p. xv).

48 Hanson, Studies, pp. 129 and 146–47, has stressed his view that what appears to us, at first blush, arbitrary Pauline interpretation often turns out to be established contemporary exegesis. Hanson takes things too far, however, as when he approvingly (p. 203) quotes Doeve’s opinion that “there is no essential difference between rabbinic and New Testament use of Scripture.”

49 As Ridderbos, Studies (see above, n. 19), p. 25, rightly argues, although his own distinction between Scripture and the Word of God runs afoul of 2 Timothy 3:16.

50 Cf. my comments in Westminster Theological Journal 40 (1977–78): 295n.

51 Weingreen, From Bible to Mishna, repeatedly alludes to the Talmudic saying (B. Kiddushin 49a): “What is Tora? It is the exposition of Tora.” In her famous article “Midrash” for the Dictionnaire de la Bible, Renée Bloch emphasizes the positive aspect: midrash is not to be understood as fable, but as an amplification intended “to show the full import” of Scripture, including its present adaptation. “So long as there is a people of God who regard the Bible as the living Word of God, there will be midrash: only the name might change. Nothing is more characteristic in this regard than the use of the Old Testament in the New Testament: it always involves midrashic actualization.” See Approaches to Judaism: Theory and Practice, ed. William Scott Green (Brown Judaic Studies 1; Missoula: Scholars, 1978), pp. 29, 32. Note also Brian M. Nolan, The Royal Son of God: The Christology of Matthew 1–2, in the Setting of the Gospel (Orbis biblicus et orientalis, 23; Göttingen: Vandenhoeck und Ruprecht, 1979), pp. 52ff. For a valuable survey and bibliography, see M. P. Miller, “Midrash,” The Interpreter’s Dictionary of the Bible, suppl. vol., pp. 593–97.

52 Cf. C. Dietzfelbinger, Paulus und das Alte Testament (München: Chr. Kaiser, 1961), p. 41, quoted by Marcus Barth, who responds: “I am not yet convinced that the hermeneutical methods developed since the enlightenment have yielded results so superior to those employed by the authors of the New Testament that we are entitled to put their hermeneutics on a Schandpfahl [pillory] or into a museum for good.” See “The Old Testament in Hebrews: An Essay in Biblical Hermeneutics,” in Current Issues in New Testament Interpretation: Essays in Honor of Otto A. Piper, ed. William Klassen and Gordon F. Snyder (New York: Harper and Row, 1962), pp. 53–78, esp. p. 78.

53 Johnson, The Old Testament in the New, p. 67, is undoubtedly correct in remarking that “if the apostles are reliable guides in biblical teaching, then they surely are reliable guides in the doctrine of interpretation, and we must follow them.” One must wonder, however, whether Johnson himself uses Scripture in a manner comparable to that in which, say, Matthew 2:15 uses Hosea—or whether he would approve of preachers who do!

54 Richard N. Longenecker, Biblical Exegesis in the Apostolic Period (Grand Rapids: Eerdmans, 1975), p. 218.

55 Here Longenecker has in mind particularly pesher-type interpretation: “This is the fulfillment of that.” One wonders whether Clement of Rome was overstepping his bounds when he used pesher in addressing the Corinthians as follows: “All glory and enlargement was given to you, and that which has been written was fulfilled, ‘My beloved ate and drank, was enlarged and made fat, and he kicked’ ” (1 Clement 3:1, quoting Deut. 32:15). We may here recall G. Vos’s insistence that Christians, though not inspired, stand with Paul in their redemptive-historical perspective, so that the interpretive tasks are the same; see the discussion by Richard B. Gaffin, Jr., The Centrality of the Resurrection: A Study in Paul’s Soteriology (Grand Rapids: Baker, 1978), pp. 23ff.

56 Longenecker says as much; my own concern is to clarify what a phrase such as “conformity to modern exegesis” does and does not imply.

57 Cf. C. S. Lewis’s discussion of “second meanings” in Reflections on the Psalms (New York: Harcourt, Brace and World, 1958), pp. 99ff. The question of sensus plenior should probably be related to the present discussion. Unfortunately, I have nothing to contribute to this very complicated issue. For a firm denial of any kind of allegorizing, see Walter C. Kaiser, Jr., “The Current Crisis in Exegesis and the Apostolic Use of Deuteronomy 25:4 in 1 Corinthians 9:8–10,” JETS 21 (1978): 3–18.

58 Cf. John Frame, “God and Biblical Language: Transcendence and Immanence,” in God’s Inerrant Word (see above, n. 19), pp. 159–77.

59 Cf. Longenecker’s reference to the descriptive over against the normative, p. 219.

60 It may be worthwhile to point out again that I do not regard this proposal as the only possible (or even necessarily the best) solution, but as an approach that should not be summarily dismissed by evangelicals.