by B. B. Warfield

IT would be difficult to invent methods of showing profound reverence for the text of Scripture as the very Word of God, which will not be found to be characteristic of the writers of the New Testament in dealing with the Old. Among the rich variety of the indications of their estimate of the written words of the Old Testament as direct utterances of Jehovah, there are in particular two classes of passages, each of which, when taken separately, throws into the clearest light their habitual appeal to the Old Testament text as to God Himself speaking, while, together, they make an irresistible impression of the absolute identification by their writers of the Scriptures in their hands with the living voice of God. In one of these classes of passages the Scriptures are spoken of as if they were God; in the other, God is spoken of as if He were the Scriptures: in the two together, God and the Scriptures are brought into such conjunction as to show that in point of directness of authority no distinction was made between them. 

Examples of the first class of passages are such as these: Gal. 3:8, "The Scripture, foreseeing that God would justify the heathen through faith, preached before the gospel unto Abraham, saying, In thee shall all the nations be blessed" (Gen. 12:1–3); Rom. 9:17, "The Scripture saith unto Pharaoh, Even for this same purpose have I raised thee up" (Ex. 9:16). It was not, however, the Scripture (which did not exist at the time) that, foreseeing God's purposes of grace in the future, spoke these precious words to Abraham, but God Himself in His own person: it was not the not yet existent Scripture that made this announcement to Pharaoh, but God Himself through the mouth of His prophet Moses. These acts could be attributed to "Scripture" only as the result of such a habitual identification, in the mind of the writer, of the text of Scripture with God as speaking, that it became natural to use the term "Scripture says," when what was really intended was "God, as recorded in Scripture, said." 

Examples of the other class of passages are such as these: Matt. 19:4, 5, "And he answered and said, Have ye not read that he which made them from the beginning made them male and female, and said, For this cause shall a man leave his father and mother, and shall cleave to his wife, and the twain shall become one flesh?" (Gen. 2:24); Heb. 3:7, "Wherefore, even as the Holy Ghost saith, To-day if ye shall hear his voice," etc. (Ps. 95:7); Acts 4:24, 25, "Thou art God, who by the mouth of thy servant David hast said, Why do the heathen rage and the people imagine vain things" (Ps. 2:1); Acts 13:34, 35, "He that raised him up from the dead, now no more to return to corruption, … hath spoken in this wise, I will give you the holy and sure blessings of David" (Isa. 55:3); "because he saith also in another [Psalm], Thou wilt not give thy holy one to see corruption" (Ps. 16:10); Heb. 1:6, "And when he again bringeth in the first born into the world, he saith, And let all the angels of God worship him" (Deut. 32:43); "and of the angels he saith, Who maketh his angels wings, and his ministers a flame of fire" (Ps. 104:4); "but of the Son, He saith, Thy throne, O God, is for ever and ever," etc., (Ps. 45:7) and, "Thou, Lord, in the beginning," etc. (Ps. 102:26). It is not God, however, in whose mouth these sayings are placed in the text of the Old Testament: they are the words of others, recorded in the text of Scripture as spoken to or of God. They could be attributed to God only through such habitual identification, in the minds of the writers, of the text of Scripture with the utterances of God that it had become natural to use the term "God says" when what was really intended was "Scripture, the Word of God, says." 

The two sets of passages, together, thus show an absolute identification, in the minds of these writers, of "Scripture" with the speaking God. 

In the same line with these passages are commonly ranged certain others, in which Scripture seems to be adduced with a subjectless λέγει or φησί, the authoritative subject—whether the divinely given Word or God Himself—being taken for granted. Among these have been counted such passages, for example, as the following: Rom. 9:15, "For he saith to Moses, I will have mercy on whom I have mercy, and I will have compassion on whom I have compassion" (Ex. 33:19); Rom. 15:10, "And again he saith, Rejoice, ye Gentiles, with his people" (Deut. 32:43); and again, "Praise the Lord, all ye Gentiles; and let all the people praise him" (Ps. 107:1); Gal. 3:16, "He saith not, And to seeds, as of many; but as of one, And to thy seed (Gen. 13:15), which is Christ"; Eph. 4:8, "Wherefore he saith, When he ascended on high, he led captivity captive, and gave gifts unto men" (Ps. 68:18); Eph. 5:14, "Wherefore he saith, Awake thou that sleepest and arise from the dead and Christ shall shine upon thee" (Isa. 60:1); 1 Cor. 6:16, "For the twain, saith he, shall become one flesh" (Gen. 2:24); 1 Cor. 15:27, "But when he saith, All things are put in subjection" (Ps. 8:7); 2 Cor. 6:2, "For he saith, At an acceptable time, I hearkened unto thee, and in a day of salvation did I succor thee" (Isa. 49:8); Heb. 8:5, "For see, saith he, that thou make all things according to the pattern that was showed thee in the mount" (Ex. 25:40); James 4:6, "Wherefore he saith, God resisteth the proud but giveth grace to the humble" (Prov. 3:34). 

There is room for difference of opinion, of course, whether all these passages are cases in point. And there has certainly always existed some difference of opinion among commentators as to the proper subauditum in such instances as are allowed. The state of the case would seem to be fairly indicated by Alexander Buttmann, when he says:

"The predicates λέγει or φησίν are often found in the New Testament in quotations, ὁ θεός or even merely ἡ γραφή being always to be supplied as subject; as 1 Cor. 6:16, 2 Cor. 6:2, Gal. 3:16, Eph. 4:8, 5:14, Heb. 8:5, 4:3 (εἴρηκεν). These subjects are also expressed, as in Gal. 4:30, 1 Tim. 5:18, or to be supplied from the preceding context, as in Heb. 1:5 seq."

Of the alternatives thus offered, Jelf apparently prefers the one:

"In the New Testament we must supply προφητής, ἡ γραφή, πνεῦμα, etc., before φησί, λέγει, μαρτυρεῖ."

Winer and Blass take the other:

"The formulas of citation—λέγει, 2 Cor. 6:2, Gal. 3:16, Eph. 4:8 al., φησί, 1 Cor. 6:16, Heb. 8:5; εἴρηκε, Heb. 4:4 (cf. the Rabbinical ואומר); μαρτυρεῖ, Heb. 7:17 (εἶπε, 1 Cor. 15:27)—are probably in no instance impersonal in the minds of the New Testament writers. The subject (ὁ θεός) is usually contained in the context, either directly or indirectly; in 1 Cor. 6:16 and Matt. 19:5, φησί, there is an apostolic ellipsis (of ὁ θεός); in Heb. 7:17, the best authorities have μαρτυρεῖται."

"In the formulas of citation such as λέγει, 2 Cor. 6:2, Gal. 3:16, etc.; φησίν, 1 Cor. 6:16, Heb. 8:5; εἴρηκε, Heb. 4:4—ὁ θεός is to be understood ('He says'); in 2 Cor. 10:10, φησίν (א DE, etc. [?], 'one says'), appears to be a wrong reading for φασίν (B), unless perhaps a τις has dropped out (but cp. Clem. Hom., xi. 9 ad init.)."

The commentators commonly range themselves with Winer and Blass. Thus, on Rom. 9:15, Sanday and Head-lam comment: "λέγει without a nominative for θεὸς λέγει is a common idiom in quotations," referring to Rom. 15:10 as a parallel case. On Gal. 3:16, Meyer says: "sc. Θεός, which is derived from the historical reference of the previous ἐῤῥέθησαν, so well known to the reader"; and Alford: "viz., He who gave the promises—God"; and Sieffert: "οὐ λέγει sc. θεός which flows out of the historical relation (known to the reader) of the preceding ἐῤῥέθησαν (cf. Eph. 4:8, 5:14)." On Eph. 4:8, Meyer's comment runs: "Who says it (comp. 5:14) is obvious of itself, namely, God, whose word the Scripture is. See on 1 Cor. 6:16; Gal. 3:16; the supplying ἡ γραφή or τὸ πνεῦμα must have been suggested by the context (Rom. 15:10). The manner of citation with the simple λέγει, obviously meant of God, has as its necessary presupposition, in the mind of the writer and readers, the Theopneustia of the Old Testament." Haupt, similarly: "The introduction of a citation with the simple λέγει, with which, of course, 'God' is to be supplied as subject, not 'the Scripture,' is found in Paul again 5:14, 2 Cor. 6:2, Rom. 15:10; similarly φησί, 1 Cor. 6:16 (εἶπεν with the addition ὁ θεός, 2 Cor. 6:16)." A similar comment is given by Ellicott, who adds at Eph. 5:14: "scil. ὁ θεός, according to the usual form of St. Paul's quotations; see notes on chap. 4:8 and on Gal. 3:16": though on 1 Cor. 6:16 he speaks with less decision: "It may be doubted what nominative is to be supplied to this practically impersonal verb, whether ἡ γραφή (comp. John 7:38, Rom. 4:3, 9:17, al.) or ὁ θεός (comp. Matt. 19:5, 2 Cor. 6:2, where this nominative is distinctly suggested by the context): the latter is perhaps the more natural: comp. Winer, Gr., § 58, 9, and notes on Eph. 4:8." On 1 Cor. 6:16, Edwards comments: "sc. ὁ θεός, as in Rom. 9:15. Cf. Matt. 19:4, 5, where ὁ ποίησας supplies a nom. to εἶπεν. Similarly in Philo and Barnabas φησί introduces citations from Scripture." On 2 Cor. 6:2, Waite says: "A statement of God Himself is adduced"; and De Wette: "sc. θεός, who Himself speaks." On Heb. 8:5, Bleek comments: "That there is to be understood as the subject of φησί, not, as Böhme thinks, ἡ γραφή, but ὁ θεός, can least of all be doubtful here, where actual words of God are adduced"; and Weiss: "This statement is now established (γάρ) by appeal to Ex. 25:40, which passage is characterized only by the interpolated φησίν (cf. Acts 25:22) as a divine oracle.… The subject of φησίν is, of course, God, neither ὁ χρηματισμός (Lün.) nor ἡ γραφή (Bhm.)." On James 4:6, Mayor comments: "The subject understood is probably God, as above, 1:12, ἐπηγγείλατο, and Eph. 4:8, 5:14, where the same phrase occurs; others take it as ἡ γραφή. Cf. above, 5:5." 

Most of these passages have, on the other hand, been explained by some commentators on the supposition that it is ἡ γραφή that is to be supplied, as has sufficiently appeared indeed from the controversial remarks in the notes quoted above. This circumstance may be taken as precluding the necessity of adducing examples here. Suffice it to say that those so filling in the subauditum are entirely at one with the commentators already quoted in looking upon the citations as treated by the New Testament writers as of divine authority, it being, in their apprehension, all one in this regard whether the subauditum is conceived as ἡ γραφή or as ὁ θεός. 

In the meantime, however, there has occasionally showed itself a tendency to treat these subjectless verbs more or less as true impersonals. Thus we read in Delitzsch's note on Heb. 8:5: "For 'see,' saith He, i. e., ὁ θεός, or taking φησί impersonally (that is, without a definite subject), 'it is said' (i. e., in Scripture), (Bernhardy, 'Synt.,' 419)." So Kern on James 4:6 comments: "λέγει here impersonaliter, instead of the foregoing λέγει ἡ γραφή"; and accordingly Beyschlag, in his recent commentary says: "to λέγει, ἡ γραφή is to be supplied, or it is to be taken with Kern impersonally." Similarly Godet on 1 Cor. 6:16 says: "The subject of the verb φησίν, says he, may be either Adam or Moses, or Scripture, or God Himself, or finally, as is shown by Heinrici, the verb may be a simple formula of quotation like our 'It is said.' This form is frequently found in Philo." Some such usage as is here supposed may seem actually to occur in the common text of Wisdom 15:12 and 2 Cor. 10:10. But in both passages the true reading is probably φασίν; in neither instance is it clear that, if φησίν be read, it has no subject implied in the context; if φησίν be read and taken as equivalent to φασίν it still is not purely indefinite; and in any case the instances are not parallel, inasmuch as in neither of these passages is it Scripture, or indeed any document, that is adduced. 

The fact that a few very able commentators have taken this unlikely line of exposition would call for nothing more than this incidental remark, were not our attention attracted somewhat violently to it by the dogmatic tone and extremity of contention of a recent commentator who has adopted this opinion. We refer to Dr. T. K. Abbott's comment on Eph. 4:8, in his contribution to "The International Critical Commentary." It runs to a considerable length, but as on this very account it opens out somewhat more fully than usual this rather unwonted view of the construction, we shall venture to quote it in extenso. Dr. Abbott says:

"Διὸ λέγει. 'Wherefore it saith' = 'it is said.' If any substantive is to be supplied, it is ἡ γραφή; but the verb may well be taken impersonally, just as in colloquial English one may often hear: 'it says' or the like. Many expositors supply, however, ὁ Θεός. Meyer even says, 'Who says it is obvious of itself, namely, God, whose word the Scripture is. Similarly Alford12 and Ellicott. If it were St. Paul's habit to introduce quotations from the Old Testament, by whomsoever spoken in the original text, with the formula ὁ Θεός λέγει, then this supplement here might be defended. But it is not. In quoting he sometimes says λέγει, frequently ἡ γραφὴ λέγει, at other times Δαβίδ λέγει, Ἡσαΐας λέγει. There is not a single instance in which ὁ Θεός is either expressed or implied as the subject, except where in the original context God is the speaker, as in Rom. 9:15. Even when that is the case he does not hesitate to use a different subject, as in Rom. 10:19, 20: 'Moses saith,' 'Isaiah is very bold, and saith'; Rom. 9:17, 'The Scripture saith to Pharaoh.' 

"This being the case, we are certainly not justified in forcing upon the apostle here and in chap. 5:14 a form of expression consistent only with the extreme view of verbal inspiration. When Meyer (followed by Alford and Ellicott) says that ἡ γραφή must not be supplied unless it is given by the context, the reply is obvious, namely, that, as above stated, ἡ γραφή λέγει does, in fact, often occur, and therefore the apostle might have used it here, whereas ὁ θεὸς λέγει does not occur (except in cases unlike this), and we have reason to believe could not be used by St. Paul here. It is some additional confirmation of this that both here and in chap. 5:14 (if that is a Biblical quotation) he does not hesitate to make important alterations. This is the view taken by Braune, Macpherson, Moule; the latter, however, adding that for St. Paul 'the word of the Scripture and the word of its Author are convertible terms.' 

"It is objected that although φησί is used impersonally, λέγει is not. The present passage and chap. 5:14 are enough to prove the usage for St. Paul, and there are other passages in his Epistles where this sense is at least applicable; cf. Rom. 15:10, where λέγει is parallel to γέγραπται in ver. 9; Gal. 3:16, where it corresponds to ἐῤῥήθησαν. But, in fact, the impersonal use of φησί in Greek authors is quite different, namely = φασί, 'they say' (so 2 Cor. 10:10). Classical authors had no opportunity of using λέγει as it is used here, as they did not possess any collection of writings which could be referred to as ἡ γραφή, or by any like word. They could say: ὁ νόμος λέγει and τὸ λεγόμενον."

It is not, it will be observed, the fact that Dr. Abbott decides against the subauditum, ὁ θεός, in these passages, which calls for remark. As he himself points out, many others have been before him in this. It is the extremity of his opinion that first of all attracts attention. For it is to be noticed that, though he sometimes speaks as if he understood an implied ἡ γραφή, or some like term, as the subject of λέγει, that is not his real contention. What he proposes is to take the verb wholly indefinitely—as equivalent to "it is said," as if the source of the quotation were unimportant and its authority insignificant. This interpretation of his proposal is placed beyond doubt by his remarks on chap. 5:14. There we read:

"Διὸ λὲγει. 'Wherefore it is said.' It is generally held that this formula introduces a quotation from canonical Scripture.… The difficulties disappear when we recognize that λὲγει need not be taken to mean ὁ θεὸς λὲγει—an assertion which has been shown in 4:8 to be untenable. It means, 'it says,' or 'it is said,' and the quotation may probably be from some liturgical formula or hymn—a supposition with which its rhythmical character agrees very well.… Theodoret mentions this opinion.… Stier adopts a similar view, but endeavors to save the supposed limitation of the use of λὲγει by saying that in the Church the Spirit speaks. As there are in the Church prophets and prophetic speakers and poets, so there are liturgical expressions and hymns which are holy words. Comparing vv. 18, 19, Col. 3:16, it may be said that the apostle is here giving us an example of this self-admonition by new spiritual songs."

So extreme an opinion, as we have already hinted, naturally finds, however, little support in the commentators, even in those quoted to buttress it,—of course, in its fundamental point. Braune says: "We must naturally supply ἡ γραφή, the Scripture, with λὲγει, 'saith,' (James 4:6, Rom. 15:10, Gal. 3:16, 1 Cor. 6:16: φησίν), and not ὁ θεός (Meyer, Schenkel), or ὁ λέγων (Bleek: the writer)": to which Dr. M. T. Riddle, his translator, however, adds: "The fact that Paul frequently supplies ἡ γραφή (Rom. 4:3, 9:17, 10:11, Gal. 4:30, 1 Tim. 5:18) is against Braune's view; for in some of these passages there is a reason for its insertion (see "Romans," p. 314), and as the Scriptures are God's Word (Meyer), the natural aim and obvious subject is ὁ θεός. So Alford, Ellicott and most." Moule's comment runs: "Wherefore he saith] Or it, i.e., the Scripture, saith. St. Paul's usage in quotation leaves the subject of the verb undetermined here and in similar cases (see, e.g., chap. 5:14). For him the word of the Scripture and the word of its author are convertible terms." Macpherson alone, of those appealed to by Dr. Abbott, supports, in a somewhat carelessly written note, the indefinite interpretation put forward by Dr. Abbott,—being misled apparently by remarks of Lightfoot's and Westcott's. His comment runs:

"A very simple quotation formula is here employed, the single word λέγει. It is also similarly used (chap. 5:14; 2 Cor. 6:2; Gal. 3:16; Rom. 15:10). This word is frequently employed in the fuller formula, The Scripture saith, λέγει ἡ γραφή (Rom. 4:3, 10:11, 11:2; Jas. 2:23, etc.); or the name of the writer of the particular scripture, Esaias, David, the Holy Spirit, the law (Rom. 15:12; Acts 13:35; Heb. 3:7; 1 Cor. 14:34, etc.). Of λέγει, φησί, εἴρηκε, and similar words thus used, Winer ("Grammar," p. 656, 1882) says that probably in no instance are they impersonal in the minds of the New Testament writers, but that the subject, ὁ θεός, is somewhere in the context, and is to be supplied. On the contrary, Lightfoot, in his note on Gal. 3:16, remarks that λέγει, like the Attic φησί, seems to be used impersonally, the nominative being lost sight of. In our passage we have no nominative in the context which we can supply, and it seems better to render the phrase impersonally, It is said. The same word is used very frequently in the Epistle to the Hebrews, but always with God or Christ understood from the immediate context. Westcott very correctly remarks (p. 457) that the use of the formula in Eph. 4:8, 5:14, seems to be of a different kind."

Outside of these commentators quoted by himself, however, Prof. Abbott's extreme view has (as has, indeed, already incidentally appeared) the powerful support of Light-foot and Heinrici. The former expresses his opinion not only in his note on Gal. 3:16, to which Macpherson refers, but more fully and argumentatively in his note on 1 Cor. 6:16 printed in his posthumous "Notes on the Epistles of St. Paul." In the former of these places he says:

"οὐ λέγει seems to be used impersonally, like the Attic φησί in quoting legal documents, the nominative being lost sight of. If so, we need not inquire whether ὁ θεός or ἡ γραφή is to be understood. Comp. λέγει, Rom. 15:10, Eph. 4:8, 5:14; and φησίν, 1 Cor. 6:16, 2 Cor. 10:10 (v. l)."

In the latter, speaking more at large "as to the authority assigned to the passage" quoted by St. Paul, he says:

"What are we to understand by φησίν Is ὁ θεός to be supplied or ἡ γραφή? To this question it is safest to reply that we cannot decide. The fact is that, like λέγει, φησίν when introducing a quotation seems to be used impersonally. This usage is common in Biblical Greek (λέγει, Rom. 15:10, Gal. 3:16, Eph. 4:8, 5:14; φησίν, Heb. 8:5, 2 Cor. 10:10 (v. l.), more common in classical Greek. Alford, after Meyer, objects to rendering φησίν impersonally here, as contrary to St. Paul's usage. But the only other occurrence of the phrase in St. Paul is 2 Cor. 10:10, where he is not introducing Scripture, but the objections of human critics and of more than one critic. If then φησίν be read there at all, it must be impersonal. The apostle's analogous use of λέγει points to the same conclusion. In Eph. 5:14 it introduces a quotation which is certainly not in Scripture, and apparently belonged to an early Christian hymn. We gather therefore that St. Paul's usage does not suggest any restriction here to ὁ θεός or ἡ γραφή. But we cannot doubt from the context that the quotation is meant to be authoritative."

In his own commentary on I Corinthians (1880), Heinrici writes as follows:

"To φησί, just as to λέγει (2 Cor. 6:2, Gal. 3:16) nothing at all is to be supplied, but like inquit it stands, sometimes as the introduction to an objection (2 Cor. 10:10, where Holsten refers to Bentley on Horat., Serm., i, 4, 78), sometimes as a general formula of citation. It is especially often used in the latter sense by Philo, in the quotation of Scripture passages, and by Arrian-Epictetus, who supplies many most interesting parallels to the Pauline forms of speech. Schweighäuser, in his Index, under φησί, remarks of it: nec enim semper in proferenda objectione locum habet illa formula, verum etiam in citando exemplo ad id quod agitur pertinente. J. G. Müller (Philo the Jew's Book on the Creation, Berlin, 1841, p. 44) says that φησί, after the example of Plato (?), became gradually among the Hellenistic Jews the standing formula of citation."

In his edition of Meyer's "Commentary on I Corinthians" (eighth edition, 1896), this note reappears in this form:

"φησίν). Who? According to the usual view, God, whose words the sayings of the Scripture are, even when they, like Gen. 2:24 through Adam, are spoken through another. Winer, 7 § 58, 9, 486: Buttmann, 117. But the impersonal sense 'es heisst,' 'inquit,' lies nearer the Pauline usage; he coincides in this with Arrian-Epictetus and Philo, with whom φησί sometimes introduces an objection, sometimes is the customary formula of citation. Cf. 2 Cor. 10:10, 6:2, 1 Cor. 15:27, Eph. 4:8; Winer, as above; Müller, in Philo, De op. mund., 44; Heinrici, i. 181. In accordance with this, are the other supplements of subject—ἡ γραφή or τὸ πνεῦμα (Rückert)—to be estimated."

Even in the extremity of his contention, therefore, Dr. Abbott, it seems, is not without support—on the philological side, at least—in previous commentators of the highest rank. 

He himself does not seem, however, quite clear in his own mind: and his confusion of both considerations and commentators which make for the fundamentally diverse positions that there is to be supplied with λέγει some such subject as ἡ γραφή, and that there is nothing at all to be supplied but the word is to be taken with entire indefiniteness, is indicatory of the main thing that calls for remark in Dr. Abbott's note. For, why should this confusion take place? It is quite evident that in interpreting the phrase the fundamental distinction lies between the view which supposes that a subject to λέγει is so implied as to be suggested either by the context or by the mind of the reader from the nature of the case, and that which takes λέγει as a case of true impersonal usage, of entirely indefinite subject. It is a minor difference among the advocates of the first of these views, which separates them into two parties—those which would supply as subject ὁ θεός, and those which would supply ἡ γραφή. That one of these subdivisions of the first class of views should be violently torn from its true comradeship and confused with the second view, betrays a preoccupation on Dr. Abbott's part, when dealing with this passage, with considerations not of purely exegetical origin. He is for the moment less concerned with ascertaining the meaning of the apostle than with refuting a special interpretation of his words: and therefore everything which stands opposed in any measure to the obnoxious interpretation appears to him to be "on his side." Put somewhat brusquely, this is as much as to say that Dr. Abbott is in this note dominated by dogmatic prejudice. 

There do not lack other indications of this fact. The most obtrusive of them is naturally the language—scarcely to be called perfectly calm—with which the second paragraph of the note opens: "We are certainly not justified in forcing upon the apostle here and in chap. 5:14 a form of expression consistent only with the extreme view of verbal inspiration." Certainly not. But because we chance not to like "the extreme view of verbal inspiration," are we justified in forbidding the apostle to use a form of expression consistent only with it, and forcing upon him some other form of expression which we may consider consistent with a view of inspiration which we like better? Would it not be better to permit the apostle to choose his own form of expression and confine ourselves, as expositors, to ascertaining from his form of expression what view of inspiration lay in his mind, rather than seek to force his hand into consistency with our preconceived ideas? The whole structure of the note evinces, however, that it was not written in this purely expository spirit. Thus only can be explained a certain exaggerated dogmatism in its language, as if doubt were to be silenced by decision of manner if not by decisiveness of evidence. So also probably is to be explained a certain narrowness in the appeal to usage—that rock on which much factitious exegesis splits. Only, it is intimated, in case "it were St. Paul's habit to introduce quotations from the Old Testament, by whomsoever spoken in the original text, with the formula ὁ θεὸς λέγει," "could this supplement here be defended." One asks in astonishment whether St. Paul really could make known his estimate of Scripture as the very voice of God which might naturally be quoted with the formula "God says," and so render the occurrence of that formula occasionally in his writings no matter of surprise, only by a habitual use of this exact formula in quoting Scripture. And one notes without surprise that the narrowness of Dr. Abbott's rule for the adduction of usage supplies no bar to his practice when he is arguing "on the other side." At the opening of the very next paragraph we read, "It is objected that although φησί is used impersonally, λέγει is not": and to this the answer is returned, "The present passage and chap. 5:14 are sufficient to prove the usage for St. Paul"; with the supplement, "And there are other passages in his epistles where this sense is at least applicable"; and further, "But in fact, the impersonal use of φησί in Greek authors is quite different." One fancies Dr. Abbott must have had a grim controversial smile upon his features when he wrote that last clause, which pleads that the meaning assigned to λέγει here is absolutely unexampled in Greek literature, not only for λέγει but even for φησί, as a reason for accepting it for λέγει here! But apart from this remarkable instance of skill in marshaling adverse facts—a skill not unexampled elsewhere in the course of this note, as any one who will take the trouble to examine the proof-texts adduced in it will quickly learn—might not the advocates of the supplement, ὁ θεός, say equally that "the present passage and chap. 5:14 are sufficient to prove the usage for St. Paul, and there are other passages in his epistles where this sense is at least applicable." And might they not support this statement with better proof-texts than those adduced by Dr. Abbott, or indeed with the same with better right; as well as with a more applicable supplementary remark than the one with which he really subverts his whole reasoning—such as this, for example, that elsewhere, in the New Testament, as for instance in the Epistle to the Hebrews, the usage contended for undoubtedly occurs, and a satisfactory basis is laid for it in the whole attitude of the entire body of New Testament writers, inclusive of Paul, toward the Old Testament? Certainly, reasoning so one-sided and dominated by preconceived opinions so blinding is thoroughly inconclusive. The note is, indeed, an eminent example of that form of argumentation which, to invert a phrase of Omar Khayyam's, "goes out at the same door at which it came in": and even though its contention should prove sound, can itself add nothing to the grounds on which we embrace it. At best it may serve as the starting-point of a fresh investigation into the proper interpretation of the phrase with which it deals. 

For such a fresh investigation we should need to give our attention particularly to two questions. The first would inquire into the light thrown by Paul's method of introducing quotations from the Old Testament, upon his estimate of the text of the Old Testament,—with a view to determining whether it need cause surprise to find him adducing it with such a formula as "God says." Subsidiary to this it might be inquired whether it is accurate to say that "there is not a single instance in which ὁ θεός is either expressed or implied as the subject, except where in the original context God is the speaker," and further, if Paul's usage elsewhere can be accurately so described, whether that fact will warrant us in denying such an instance to exist in Eph. 4:8. The second question would inquire into the general usage of the subjectless λέγει or φησί in and out of the New Testament, with a view to discovering what light may be thrown by it upon the interpretation of the passages in question. It might be incidentally asked in this connection whether it is a complete account to give of φησί in profane Greek to say that the "impersonal use of φησί in Greek authors is quite different from that of the New Testament, inasmuch as with them φησί = φασί, 'they say.' "

It is really somewhat discouraging at this late date to find it treated as still an open question, how Paul esteemed the written words of the Old Testament. And it brings us, as the French say, something akin to stupefaction, when Dr. Abbott goes further and uses language concerning Paul's attitude toward the Old Testament text which implies that Paul habitually distinguished, in point of authority, between those passages "where in the original context God is the speaker" and the rest of the volume, so that "we have reason to believe" that the formula ὁ θεὸς λέγει "could not be used by Paul" in introducing Scriptural language not recorded as spoken by God in the original context. He even suggests, indeed, that Paul shows an underlying doubt as to the Divine source of even the words attributed to God in the Old Testament text—"not hesitating to use a different subject" when quoting them, "as in Rom. 10:19, 20, 'Moses saith,' 'Isaiah is very bold and saith'; Rom. 9:17, 'The Scripture saith to Pharaoh' "—and deals with the text of other portions with a freedom which exhibits his little respect for them—"not hesitating to make important alterations" in them. It would seem to require a dogmatic prejudice of the very first order to blind one to a fact so obvious as that with Paul "Scripture," as such, is conceived everywhere as the authoritative declaration of the truth and will of God—of which fact, indeed, no better evidence can be needed than the very texts quoted by Dr. Abbott in a contrary sense. 

For, when Paul, in Rom. 9:15, supports his abhorrent rejection of the supposition that there may be unrighteousness with God, with the divine declaration taken from Ex. 33:19, introduced with the formula, "For he"—that is, as Dr. Abbott recognizes, God—"saith to Moses," and then immediately, in Rom. 9:17, supports the teaching of this declaration with the further word of God taken from Ex. 9:16, introduced with the formula, "For the Scripture saith unto Pharaoh"—the one thing which is thrown into a relief above all others is that, with Paul, "God saith" and "Scripture saith" are synonymous terms, so synonymous in his habitual thought that he could not only range the two together in consecutive clauses, but use the second in a manner in which, taken literally, it is meaningless and can convey an appropriate sense only when translated back into its equivalent of "God saith." The present tense in both formulas, moreover, advises us that, despite the fact that in both instances they are words spoken by God which are cited, it is rather as part of that Scripture which to Paul's thinking is the ever-present and ever-speaking word of God that they are adduced. It is not as words which God once spoke (εἶπεν, LXX.) to Moses that the former passage is here adduced, but as living words still speaking to us—it is not as words Moses was once commanded to speak to Pharaoh that the second is here adduced, but as words recorded in the ever-living Scripture for our admonition upon whom the ends of the world have come. They are thus not assigned to Scripture in order to lower their authority: but rather as a mark of their abiding authority. And similarly when in that catena of quotations in Rom. 10:16–21, we read at ver. 19, "first Moses saith," and then at ver. 20, "and Isaiah is very bold and saith," both adducing words of God—the implication is not that Paul looks upon them as something less than the words of God and so cites them by the names of these human authors; but that it is all one to him to say, "God says," and "Moses says," or "Isaiah says": and therefore in this catena of quotations—in which are included four, not two, quotations—all the citations are treated as alike authoritative, though some are in the original context words of God and others (ver. 16) words of the prophet—and though some are adduced by the name of the prophet and some without assignment to any definitely named human source. The same implication, again, underlies the fact that in the catena of quotations on Rom. 15:9 seq., the first is introduced by καθὼς γέγραπται, the next two by καὶ πάλιν λέγει and καὶ πάλιν, and the last by καὶ πάλιν Ἡσαΐας λέγει—the first being from Ps. 78:50, the second from Deut. 32:43, the third from Ps. 117:1, and only the last from Isaiah—Isa. 11:10: clearly it is all one to the mind of Paul how Scripture is adduced—it is the fact that it is Scripture that is important. So also it is no more true that in Gal. 3:16, the λέγει "corresponds to ἐῤῥήθησαν" of the immediately preceding context, than that it stands in line with the "and the Scripture foreseeing that God would justify the Gentiles by faith, preached the Gospel beforehand unto Abraham" of 3:8—a thing which the Scripture as such certainly did not do; and with the "for it is written" of 3:10 and 3:13, and the unheralded quotations of the Scriptures as unquestioned authority of 3:11 and 3:12; and with the general appeal in 3:22 to the teaching of Scripture as a whole as the sole testimony needed: the effect of the whole being to evince in the clearest manner that to Paul the whole text of Scripture, inclusive of Gen. 12:3, Deut. 27:26, Hab. 2:4, Lev. 18:5, and Gen. 22:18, was as such the living word of the living God profitable to all ages alike for divine instruction. 

We need not go, indeed, beyond the first sentence of this Epistle to the Romans from which all but one of Dr. Abbott's citations are drawn, to learn Paul's conception of Scripture as the crystallized voice of God. There he declares himself to have been "separated unto the gospel of God which he promised afore by his prophets in the Holy Scriptures" (Rom. 1:2). Dr. George T. Purves, in a singularly well-considered and impressive paper on "St. Paul and Inspiration," printed in The Presbyterian and Reformed Review for January, 1893, justly draws out the meaning of this compressed statement thus:

"Not only did Moses and the prophets speak from God, but the sacred Scriptures themselves were in some way composed under divine control. He not only affirms with Peter that 'moved by the Holy Ghost, men spake from God,' but that 'the Scriptures themselves are inspired by God.' Paul plainly recognizes the human authorship of the books, and quotes Moses and David and Isaiah as speaking therein. But not only through them, but in these books of theirs did God also speak. Many readers notice the first part of Paul's statement, but not the second. God spake 'through the prophets in the Holy Scriptures.' "

This emphasis on the written Scriptures as themselves the product of a divine activity, making them as such the divine voice to us, is characteristic of the whole treatment of Scripture by Paul (1 Cor. 10:11, Rom. 15:4, 4:23, 1 Cor. 9:10, 4:6): and it is thoroughly accordant with the point of view so exhibited, that he explicitly declares, not of the writers of Scripture, but of the sacred writings themselves, that they are theopneustic—breathed out, or breathed into by God (2 Tim. 3:16). For he applies this epithet not to "every prophet," but to "every Scripture"—that is, says Dr. Purves, to "the whole collection to which he had just referred as the 'sacred writings,' and all their parts": these writings are theopneustic. "By their inspiration, he evidently meant," continues Dr. Purves justly, "that, as writings, they were so composed under God's particular direction that both in substance and in form they were the special utterances of His mind and will." 

It could be nothing more than an accident if Paul, under the dominance of such a conception of Scripture, has nowhere happened to adduce from it a passage, taken out of a context in which God is not expressly made in the Old Testament narrative itself the speaker, with the formula, ὁ θεὸς λέγει, expressed or implied. If no instance of such an adduction occurs, it is worth while to note that fact, to be sure, as one of the curious accidents of literary usage; but as there is no reason to doubt that such a formula would be entirely natural on the lips of Paul, so there is no propriety in calling it impossible in Paul, or even in erecting a distinction between him and other New Testament writers on the ground that they do and he does not quote Scripture by such a formula. As a matter of fact, the distinction suggested between passages in Scripture "where in the original context God is the speaker" and passages where He is not the speaker—as if the one could be cited with a "God says," and the other not,—is foreign to Paul's conception and usage, as has abundantly appeared already: so that whatever passages of the former kind occur—"as in Rom. 9:15," says Dr. Abbott—are really passages in which Scripture is quoted with a "God says." It cannot be held to be certain, moreover, that passages do not occur in which the "God says" introduces words not ascribed to God in the original context—so long, at least, as it is not obvious that "God" is not the subauditum in passages like Acts 13:35, Rom. 15:10, Gal. 3:16. It is no doubt, however, also worth observing that it is equally matter of fact, that it is rather to the Epistle to the Hebrews than to those that bear the name of Paul that we shall need to go to find a body of explicit instances of the usage in question. This is, as we have said, an interesting fact of literary usage, but it is not to be pressed into an indication of a divergent point of view toward "Scripture" between the Epistle to the Hebrews and the epistles that bear Paul's name. 

Even Dr. Westcott seems, to be sure, so to press it. In the interesting dissertation "On the Use of the Old Testament in the Epistle," which he has appended to his "Commentary on the Epistle to the Hebrews," he sets out in some detail the facts that bear on the mode in which that epistle cites the Old Testament:

"The quotations," he tells us, "are without exception made anonymously. There is no mention anywhere of the name of the writer (4:7 is no exception to the rule). God is presented as the speaker through the person of the prophet, except in the one place where He is directly addressed (2:6).… In two places the words are attributed to Christ.… In two other places the Holy Spirit specially is named as the speaker.… But it is worthy of notice that in each of these two cases the words are also quoted as the words of God (4:7, 8:8). This assignment of the written word to God, as the Inspirer of the message, is most remarkable when the words spoken by the prophet in his own person are treated as divine words-as words spoken by Moses: 1:6 (Deut. 32:43); 4:4, comp. vv. 5, 7, 8 (Gen. 2:2); 10:30 (Deut. 32:36); and by Isaiah: 2:13 (Isa. 8:17 f), comp. also 13:5 (Deut. 31:6). Generally it must be observed that no difference is made between the word spoken and the word written. For us and for all ages the record is the voice of God. The record is the voice of God, and as a necessary consequence the record is itself living.… The constant use of the present tense in quotations emphasizes this truth: 2:11, 3:7, 12:5. Comp. 12:6."

Every careful student will recognize this at once as a very clear and very true statement of the attitude of the author of the Epistle to the Hebrews toward the Old Testament. But we cannot help thinking that Dr. Westcott overshoots the mark when he throws it into strong contrast with the attitude of the rest of the New Testament writers to the Old Testament. When he says, for example: "There is nothing really parallel to this general mode of quotation in the other books of the New Testament"—meaning apparently to suggest, as the subsequent context indicates, that the author of this Epistle exhibits an identification in his mind of the written text of the Scriptures with the voice of God which is foreign to the other writers of the New Testament—he would seem to have attached far too great significance to what is, after all, so far as it is real, nothing more than one of those surface differences of individual usage which are always observable among writers who share the same fundamental view-point, or even in different treatises from the same hand. Entirely at one in looking upon the Scriptures as nothing less than τὰ λόγια τοῦ θεοῦ (Rom. 3:2, Heb. 5:12)—in all their parts and phrases the utterance of God—the epistles that bear the name of Paul and this epistle yet chance to differ in the prevalent mode in which these "oracles" are adduced: the one in its formulas of citation emphasizing the sole fact that they are "oracles" it is quoting, the others, that these "oracles" lie before them in written form. Let the fact of this difference, of course, be noted: but let it not be overstrained and, as if it were the sole relevant fact in the field of view, made to bear the whole weight of a theory of the relations of the two in their attitude toward Scripture. 

Impossible as such a procedure should be in any case, it becomes doubly so when we note the extremely narrow and insecure basis for the conclusion drawn, which is offered by the differences in usage adduced between Hebrews and the rest of the New Testament—which means for us primarily the epistles that bear the name of Paul. Says Dr. Westcott in immediate sequence to what we have quoted from him:

"There is nothing really parallel to this general mode of quotation in the other books of the New Testament. Where the word λέγει occurs elsewhere, it is for the most part combined either with the name of the prophet or with 'Scripture': e.g., Rom. 10:16, Ἡσαΐας λέγει; 10:19, Μωυσῆς λέγει; 11:9, Δαυεὶδ λέγει; 4:3, ἡ γραφὴ λέγει; 9:17, λέγει ἡ γραφή, etc. Where God is the subject, as is rarely the case, the reference is to words directly spoken by God: 2 Cor. 6:2, λέγει γὰρ (ὁ θεός); Rom. 9:15, τῷ Μωυσεῖ λέγει; 9:25, ἐν τῷ Ὡσηὲ λέγει. Comp. Rom. 15:9–12 (γέγραπται … λέγει.… Ἡσαΐας λέγει). The two passages in the Epistle to the Ephesians (4:8, 5:14, διὸ λέγει) appear to be different in kind."

The last remark is apparently intended to exclude Eph. 4:8 and 5:14 from consideration. The immediately preceding one seems intended to suggest that the subject to be supplied to λέγει in Rom. 15:10, which carries with it also Rom. 15:11, is ἡ γραφή; if we rather supply with Sanday-Headlam θεός, this citation would afford an instance to the contrary. Other cases similar to this, e. g., Acts 13:35. and (with the parallel φησί) 1 Cor. 6:16, are simply passed by in silence. If such cases were considered, perhaps the induction would be different. 

It is possible, on the other hand, that the usage of the Epistle to the Hebrews also is conceived by Dr. Westcott a shade too narrowly. It scarcely seems sufficient to say of 2:6, for example, that this passage is not an exception to the more general usage of the Epistle inasmuch as it is "the one place where God is directly addressed"—and is therefore not ascribed to Him, but to "some one somewhere." According to Dr. Westcott's own exposition, we have in 1:10 also words addressed to God and yet cited as spoken by God, and in a number of passages words spoken of God nevertheless cited as spoken by Him; and, in a word, the fundamental principle of the mode of quotation used by this Epistle is that the words of Scripture as such are the living words of God and are cited as such indifferently—whether in the original context spoken by Him or by another of Him, to Him, or apart from Him. In any event, therefore, the citation in the present passage by the formula "someone hath somewhere borne witness" is an exception to the general usage of the Epistle, and evidences that the author of it, though conceiving Scripture as such as a body of divine oracles, did not really lose sight of the fact that these oracles were delivered through men, and might therefore be cited on occasion as the deliverances of these men. In other words, here is a mode of citation of the order affirmed to be characteristic of the letters bearing the name of Paul. It is at least not beyond the limits of possibility that another such instance occurs in 4:7: "saying in David." No doubt, "in David," may be taken here, as Dr. Westcott takes it, as meaning "in the person of David," i. e., through his prophetic utterances; but it seems, on the whole, much more natural to take it as parallel to ἐν τῇ βίβλῳ Μωυσέως (Mark 12:26), ἐν τῷ Ὡσηέ (Rom. 9:25), and as meaning "in the book of David"—exhibiting the consciousness of the author that he is quoting not merely "God," but God in the written Scripture—written by the hand of men. This is the more worth insisting on that it is really not absolutely certain that the subject of the λέγων here is immediately "God" at all. There is no subject expressed either for it or the ὁρίζει on which it depends; and when we go back in the context for an express subject it eludes us, and we shall not find it until we arrive at the "even as the Holy Ghost saith" of 3:7. From that point on, we have a series of quotations, introduced, quite in the manner of Philo, with formulæ which puzzle us as to their reference—whether to God, who is the general subject of the whole context, or to Scripture, conceived as the voice of God (e. g., 3:15, ἐν τῷ λέγεσθαι,—by whom? God? or "the Scripture" already quoted? 4:4, εἴρηκεν—who? God? or Scripture? 4:5, καὶ ἐν τούτῳ πάλιν). Something of the same kind meets us in the eighth chapter, where quite in the manner of Philo, we begin at ver. 5: "Even as Moses was oracularly warned when about to make the tabernacle, for 'see,' φησίν, etc." and proceed at ver. 8, with a subjectless λέγει, to close with ver. 13 with an equally subjectless ἐν τῷ λέγειν. It certainly is not obvious that the subject to be supplied to these three verbs is "God" rather than "oracular Scripture." 

One can but feel that with a due regard to these two classes of neglected facts, a somewhat broader comparison of the usage of the Epistle to the Hebrews and that of those letters that bear the name of Paul would not leave an impression of such sharp and indubitable divergence in point of view as Dr. Westcott's statement is apt to suggest. In the Epistle to the Hebrews, the verb λέγω is used to introduce citations, (1) with expressed subject: 2:6, "But someone somewhere hath borne witness, saying.…"; 3:7, "Even as the Holy Ghost saith.…"; 6:14, "God.… sware by himself, saying.…": (2) with subject to be supplied from the preceding context: 1:6, "And when he (God) again bringeth in the firstborn into the world, he saith.…"; 1:7, "And of the angels he (God) saith.…"; 2:12, "He (Christ) is not ashamed to call them brethren, saying.…"; 5:6, "As he (God) saith also in another place.…": (3) with subject to be supplied from the general knowledge of the reader: 10:5, "Wherefore when he (Christ) cometh into the world, he saith.…"; 10:8, "Saying (Christ) above.…"; 12:26, "But now hath he (God) promised, saying.…": (4) without obvious subject: 3:15, "While it is said, To day, etc." (by whom? God? or the Scripture quoted, 3:7 seq.?); 4:7, "He [or it?] again defineth a certain time, saying in David.…"; 8:8, "For finding fault with them, he [or it?] saith.…" (cf. 8:13, "in that he [or it?] saith.…"). On the other hand, in the epistles that bear the name of Paul we may distinguish some four cases of the adduction of Scripture by the formula λέγει. (1) Sometimes, quoting Scripture as a divine whole, the formula runs ἡ γραφὴ λέγει or λέγει ἡ γραφή: Rom. 4:3, 9:17 (λέγει ἡ γραφή τῷ Φαραῳ̈́), 11:2 (ἡ γραφὴ ἐν Ἡλεία), Gal. 4:30, 1 Tim. 5:18. (2) Sometimes it is adduced by the name of the author: Δαυεὶδ λέγει, Rom. 4:6, 11:9; Ἡσαίας λέγει, Rom. 10:16, 20, 15:12. (3) Sometimes it is quoted by its contents: ὁ νόμος λέγει, Rom. 3:19, 7:7, 1 Cor. 9:8, 10, 14:34; the righteousness that is of faith λέγει, Rom. 10:6 (cf. ver. 10); ὁ χρηματισμός λέγει, Rom. 11:4. (4) Sometimes it is adduced by the verb λέγει without expressed subject. (A) In some of these cases the subject is plainly indicated in the preceding context: Rom. 9:25 = "God," from ver. 22; 10:10 = "the righteousness of faith," (?) from ver. 6; 10:21 = "Isaiah," from ver. 20. (B) In others it is less clearly indicated and is not altogether obvious: [Acts 13:34 = "God," from εἴρηκεν?]; Rom. 9:15 = "God," from ver. 14?; Rom. 15:10 = "Scripture," from γέγραπται?; 2 Cor. 6:2 = "God," from preceding context; Gal. 3:16 = "God," from the promises?; Eph. 4:8 and 5:12. It should be added that parallel to the use of the subjectless φησί in Heb. 8:5 we have the similar use of it in 1 Cor. 6:16. 

When we glance over these two lists of phenomena we shall certainly recognize a difference between them: but the difference is not suggestive of such an extreme distinction as Dr. Westcott appears to indicate. The fact is that for its proper estimation we must rise to a higher viewpoint and look upon the two lists in the light of a much larger fact. For we cannot safely study this difference of usage as an isolated phenomenon: and we shall get the key to its interpretation into our hands only when we correlate it with a more general view of the estimate of Scripture and mode of adducing Scripture prevalent at the time and in the circles which are represented by these epistles. Dr. Westcott already points the way to this wider outlook, when at the end of his discussion he adds these words:

"The method of citation on which we have dwelt is peculiar to the Epistle [to the Hebrews] among the writings of the New Testament; but it is interesting to notice that there is in the Epistle of Clement a partial correspondence with it. Clement generally quotes the LXX. anonymously. He attributes the prophetic words to God (15, 21, 46), to Christ (16, 22), to the Holy Word (13, 56), to the Holy Spirit (13, 16). But he also, though rarely, refers to the writers (26, Job; 52, David), and to Books (57, Proverbs, 'the all virtuous Wisdom'), and not unfrequently uses the familiar form γέγραπται (14, 39, etc.). The quotations in the Epistle of Barnabas are also commonly anonymous, but Barnabas mentions several names of the sacred writers, and gives passages from the Law, the Prophets and the Psalms with the formula, 'the Prophet saith' (vi. 8; 2; 4, 6)."

And, he should have added, Barnabas also repeatedly adduces what he held to be the Word of God with the formulas γέγραπται, (iv. 3, 14, v. 2, xi. 1, xiv. 6, xv. 1, xvi. 6) and λέγει ἡ γραφή (iv. 7, 11, v. 4, vi. 12, xiii. 2, xv. 5): and indeed passes from the one mode of citation to the other without the least jar, as, for example, in chap. 5.: "For it is written concerning him, some things indeed with respect to Israel, and some with respect to us. For it saith this (Isa. 53:5, 7).…. And the Scripture saith (Prov. 1:17).… And still also this (Gen. 1:26).…. For God saith (Zech. 13:7).… For the prophesier saith (Ps. 22:21, etc.).…. And again it saith (Isa. 50:6)." Though adverting thus to these facts, however, Dr. Westcott quite misses their significance. What they mean is shortly this: that the two modes of citing Scripture thought to distinguish Hebrews and the letters that bear the name of Paul, do not imply well-marked distinctive modes of conceiving Scripture; but coëxist readily within the limits of one brief letter, like the letter of Clement or that of Barnabas. No wonder, when laid side by side, we found the usages of the two to present no sharply marked division line, but to crumble into one another along the edges. And when we look beyond Clement and Barnabas and take a general glance over the literature of the time, it is easily seen that we are looking in the two cases only at two fragments of one fact, and are seeing in each only one of the everywhere current methods of citing Scripture as the very Word of God. It seems inconceivable that one could rise from reading, say, twenty pages of Philo, for example, without being fully convinced of this. 

Philo's fundamental conception of Scripture is that it is a book of oracles; each passage of it is a χρησμός or λόγιον, and the whole is therefore οἱ χρησμοί or τὰ λόγια: he currently quotes it, accordingly, as "the living voice" of God, and whole treatises of his may be read without meeting with a single citation introduced by γέγραπται or with the Scriptures once called ἡ γραφή. Nevertheless, when occasion serves, he adduces Scripture readily enough as ἡ γραφή, and cites it with γέγραπται, and calls it τὰ γράμματα. We have no more reason for assuming that such modes of citing Scripture would have been foreign to the author of the Epistle to the Hebrews (whose mode of citing Scripture is markedly Philonic) than we have for assuming that the author of the tract de Mutatione Nominum, in which they do not occur, but where Scripture is almost exclusively οἱ χρησμοί, or the author of the tracts de Somniis, where again they do not occur, but where Scripture is almost exclusively ὁ ἱερὸς (or ὁ θεῖος) λόγος (i. 14, 22, 33, 35, 37, 39, 42, ii. 4, 9, 37, etc.; i. 33, ii. 37)—which designations are rare again in de Mutatione Nominum (ὁ θ. λ., 20; ὁ ἱ. λ., 38)—held a different conception of Scripture from the author of the tract de Legatione ad Caium (§ 29) or the tract de Abrahamo (§ 1), in which the Scriptures are spoken of as τὰ γράμματα or αἱ γραφαί. There is no reason, in a word, why, if the Epistle to the Hebrews had contained even a single other verse, it might not have presented the "exotic," ἡ γραφή or γέγραπται. Because Philo or the author of this Epistle was especially accustomed to look on Scripture as a body of oracles and to cite it accordingly, is no reason why he should forget that it is a body of written oracles and be incapable on occasion of citing it from that point of view. Similarly because Paul ordinarily cites Scripture as written is no reason why he should not be firmly convinced that what is written in it is oracles, or should not occasionally cite it from that point of view. In a word, the two modes of citing Scripture brought into contrast by Bishop Westcott are not two mutually exclusive ways of citing Scripture, but two mutually complementary methods. The use of the one by any writer does not argue that the other is foreign to him; if we have enough written material from his hand, we are sure rather to find in him traces of the other usage also. This is the meaning of the presence in the Epistle to the Hebrews of suggestive instances of an approach to the citation of Scripture as a document: and of the presence in the epistles bearing the name of Paul of instances of modes of citation which hint of his conception of Scripture as an oracular book. Where and when the sense of the oracular character of the source of the quotation is predominatingly in mind it tends to be quoted with the simple φησί or λέγει, with the implication that it is God that says it: this is most richly exhibited in Philo, and, within the limits of the New Testament, most prevailingly in the Epistle to the Hebrews. Where and when, on the other hand, the consciousness that it is from a written source that the authoritative words are drawn is predominant in the mind, it tends to be quoted with the simple γέγραπται or the more formal ἡ γραφὴ λέγει: this is the mode in which it is most commonly cited in the epistles that bear the name of Paul. Both modes of citation rest on the common consciousness of the Divine authority of the matter cited, and have no tendency to exclude one another: they appear side by side in the same writer, and must be held to predominate variously in different writers only according to their prevailing habits of speaking of Scripture, and at different times in the same writer according as the circumstances under which he was writing threw the emphasis in his mind temporarily upon the Scriptures as written oracles or as written oracles. 

From this point of view we may estimate Dr. Westcott's remark: "Nor can it be maintained that the difference of usage is to be explained by the difference of readers, as being [In Hebrews] Jews, for in the Gospels γέγραπται is the common formula (nine times in St. Matthew)." This remark, like his whole treatment of the subject, seems conceived in a spirit which is too hard and narrow, too drily statistical. No one, doubtless, would contend that the difference of readers directly produced the difference of usage, as if the Scriptures must be quoted to Jews as "oracles of God," and to Gentiles as "written documents." But it is far from obvious that the difference of readers may not, after all, have had very much to do with the prevalence of the one mode of citation in the Epistle to the Hebrews and of the other in the epistles that bear the name of Paul. The Jews were certainly accustomed to the current citation of the Scriptures as the living voice of God in oracular deliverances—as the usage of Philo sufficiently indicates: and it may be that this was subtly felt the most impressive method of adducing the words of the Holy Book when addressing Jews. On the other hand, the heathen were accustomed to authoritative documents, cited currently, with an implication of their authority, by the formula γέγραπται: and it may well be that this subtly suggested itself as the most telling way of adducing Scripture as authoritative law to the Gentiles. We need not ride such a notion too hard: but it at least seems far from inconceivable that the selfsame writer, addressing, on the one hand, a body of devout Jews, and, on the other, a body of law-loving Romans, might find himself using almost unconsciously modes of adducing Scripture suggestive, in the one case, of loving awe in its presence and, in the other, of its binding authority over the conscience. Be this as it may, however, it is quite clear that the fact that Paul ordinarily adduces Scripture with "the forms (καθὼς) γέγραπται (sixteen times in the Epistle to the Romans), ἡ γραφὴ λέγει, and the like, which never occur in the Epistle to the Hebrews," implies no far-reaching difference of conception on his part from that exhibited by that Epistle, as to the fundamental character of the Scriptures as an oracular book—which, on the contrary, is just what he calls them (Rom. 3:2)—and certainly raises no presumption against his occasionally quoting them as an oracular book with the formula so characteristic of the Epistle to the Hebrews, ὁ θεὸς λέγει, or its equivalents. And the fact that "Paul not unfrequently quotes the words of God as 'Scripture' simply (e. g., Rom. 9:17)" so far from raising a presumption that he would not quote "Scripture" as "words of God," actually demonstrates the contrary, as it only in another way indicates the identification on his part of the written word with the voice of the speaking God. 

If we approach the study of such texts as Eph. 4:8, 5:14, therefore, from the point of view of the Pauline conception of Scripture, there is no reason why they should not be understood as adducing Scripture with a high "God says." To say that "we have reason to believe" that such a formula "could not be used by Paul," is as wide of the mark as could well be. To say that it is a formula more in accordance with the point of view of the Epistle to the Hebrews, is to confound mere occasional differences in usage with fundamental differences in conception. To Paul, too, the Scriptures are a book of oracles, and though he cites them ordinarily as written oracles there is no reason why he should not occasionally cite them merely as oracles. And in any case, whether we take the subauditum in such passages as "God," or "Scripture," or prefer to render simply by "it," from Paul's point of view the meaning is all one: in any case, Scripture is to him the authoritative dictum of God and what it says is adduced as the authoritative word that ends all strife.

In seeking to estimate the likelihoods as to the meaning of such a locution as the διὸ λέγει of Eph. 4:8, 5:14, we should not lose from sight, on the other hand, the fact that the Greek language was not partial to true "impersonals," that is, absolutely indefinite uses of its verbs. Says Jelf:

"Of impersonal verbs (in English, verbs with the indefinite it) the Greek language has but few."

Says Kühner:

"Impersonal verbs, by which we understand a verb agreeing with the indefinite pronoun it, are not known to the Greek language: for expressions like δεῖ, χρή … λέγεται, etc.… the Greek always conceived as personal, in that the infinitive or subjoined sentence was considered the subject of these verbs."

No doubt, the subject often suffers ellipsis—especially when it may be counted upon readily to suggest itself, either out of the predicate itself, or out of the context, or out of the knowledge of the reader: and no doubt this implied subject is sometimes the indefinite τις. But it remains true that as yet there has turned up no single instance in all Greek literature of λέγει in the purely indefinite sense of "someone says," equivalent to "it is said" in the meaning of general rumor, or of a common proverb, or a current saying; and though there have been pointed out instances of something like this in the case of the kindred word φησί, it still remains somewhat doubtful precisely how they are to be interpreted. The forms commonly used to express this idea are either the expressed τις, or the third person plural, as λέγουσι, φασί, ὀνομάζουσιν, or the third person singular passive, as λέγεται, or the second person singular optative or indicative of the historical tenses, as φαίης ἄν, = dicas, or the like. 

We find it, indeed, occasionally asserted that φησί is used sometimes or frequently as a pure impersonal, in the sense of "it is said." The passage from Bernhardy, to be sure, to which reference has been made in support of this assertion, by more than one of the commentators adduced above, has its primary interest not in this point, but in the different one of the use of the singular φησί for the plural—like the Latin inquit, and the English "says" in that vulgar colloquial locution in which it is made to do duty not only in the form "he says," but also in such forms as "I says" and "you says," and even "they says" and "we says." What Bernhardy remarks is:

"The rhetorical employment of the singular for the plural rests on the Greek peculiarity (K. 3, 5; 6, 13c.) of clearly conceiving and representing the multitude by means of the individual. A ready instance of this is supplied by the formula φησί, like the Latin inquit an expression for all persons and numbers for designating an indefinite speaker (den beliebigen Redner)—'heisst es'; and by the more classic εἰπέ μοι in appeal to the multitude in Attic life, Arist. (as Pac., 385, εἰπέ μοι τί πάσχετʼ ὦνδρες; coll. Eccl., 741), Plat. (clearly in a turn like εἰπέ μοι, ὦ Σώκρατές τε καὶ ὑμεῖς οἱ ἄλλοι), Demosth., Phil. 1, p. 45; Chers., p. 108; Timocr., p. 718."

The usage of φησί here more particularly adverted to—for all numbers and persons—seems a not uncommon one. Instances may possibly be found in the "Discourses" of Epictetus i. 29, 34 (Schenkl, p. 95). "Even athletes are dissatisfied with slight young men: 'He cannot lift me,' φησί," where φησί might perhaps be rendered by our vernacular, "says they," referring to "the athletes." Again, iv. 9, 15 (Schenkl, p. 383): "But learn from what the trainers of boys do. The boy has fallen: 'Rise,' φησί, 'wrestle again, till you become strong!' " where we may possibly have another 'says they,' viz., the trainers. Possibly again ii. 10, 20 (Schenkl, p. 133), "But consider, if you refer everything to a small coin, not even he who loses his nose is in your opinion damaged. 'Yes,' φησί, 'for he is mutilated in his body,' " where possibly φησί is "says you," referring to the collocutor, addressed in the preceding context in the second person—though, no doubt, another explanation is here possible. Indeed, in no one of the instances cited is it impossible to conceive a singular subject derived from the contextual plural as specially in mind. If φησί were genuine in Wisdom 15:12, 2 Cor. 10:10, these might well supply other instances—the "says they" in each case continuing the contextual or implicated plural. But in none of these instances, it is to be observed, would the subject be conceived as in the strict sense "indefinite." It is a perfectly definite subject that is present to the mind of the writer, given either in the immediate context or in the thorough understanding that exists between the writer and reader. There is in them nothing whatever of the vagueness that attaches to the French "on dit," or the German "man sagt," or the English "it is said." The Greeks had other locutions for expressing this idea, and if it was ever expressed by the simple φησί, only the slightest traces of it remain in their extant literature. 

In the seventh edition of the Greek Lexicon of Liddell & Scott, nevertheless, this usage is expressly assigned to φησί. We read:

"φασί parenthetically, they say, it is said, Il. 5, 638, Od. 6, 42 and Att.; but in prose also φησί, like French on dit, Dem. 650, 13, Plut. ii 112 C., etc. (so Lat. inquit, ait, Gronov, Liv. 34, 3, Bent. Hor. 1 Sat. 4, 79;—especially in urging an objection or counter-argument, 5. Interpp. Pers. Sat. 1, 40);—so also ἔφη, c. acc. et inf., Xen. An. i, 6, 6."

It is far from obvious, however, that the passages here adduced will justify precisely the usage which they are cited to illustrate. In the passage from Demosthenes—ἔστω, φησὶν, ὑπὲρ αὐτοῦ ἡ αὐτὴ τιμωρία, etc.—it seems to be quite clear, as the previous sentence suggests and the editors recognize, that the subject of the φησί is ἕκαστος τῶν γεγραφότων, and is far from a purely indefinite τις. The passage from Plutarch ("Consolatio ad Apollonium," xxi) is more specious. It runs: ἀλλʼ οὐ γὰρ ἤλπιζον, φησί, ταῦτα πείσεσθαι, οὐδὲ προσεδόκων; and is translated in the Latin version, "At, inquiunt, præter spem mihi hic casus et expectationem evenit"; and in Holland's old English version, "But haply you will say, I never thought that this would have befallen unto me, neither did I so much as doubt any such thing." A glance at the context, however, is enough to show that there is no purely indefinite φησί here, though it may be that we have here another instance of its usage without regard to number and person. In any case, the subject is the quite definitely conceived interlocutor of the passage. That the ἔφη adduced at the end of the note as in some degree of the same sort is not an indefinite ἔφη, but has the Clearchus of the immediately preceding context as its subject, is too obvious for remark. Clearchus was present by the request of Cyrus at the trial of Orontes, and when he came out he reported to his friends the manner in which the trial was conducted: "He said (ἔφη) that Cyrus began to speak as follows." It is not by such instances as these that the occurrence of a purely indefinite φησί can be established. 

The subjectless φησί, to be sure, does occur very thickly scattered over the face of Greek literature, introducing or emphasizing quotations, or adducing objections, or the like: but the "it" that is to be supplied to it is, ordinarily at least, a quite definite one with its own definite reference perfectly clear. A characteristic instance, often referred to, is that in Demosth., "Leptin," § 56: καὶ γάρ τοι μόνῳ τῶν πάντων αὐτῷ τοῦτʼ ἐν τῇ στήλῃ γέγραπται, ἐπειδὴ Κόνων, φησίν, ἠλευθέρωσε τοὺς Ἀθηναίων συμμάχους.—Ἔστι δὲ τοῦτο τὸ γράμμα.…." Here F. A. Wolf comments: "Absolute ibi interjectum est φησίν, aut, si mavis, subaudi ὁ γράψας"; and Schaefer adds: "Subaudi ἡ στήλη." It does not appear why we should not render simply "it says": but this "it" is so far from an " 'indefinite' it" that it has its clear reference to the inscription just mentioned. Perhaps even more instructive is a passage in the third Philippic42 of Demosthenes, which runs as follows:

"That such is our present state, you yourselves are witnesses, and need not any testimony from me. That our state in former times was quite opposite to this, I shall now convince you, not by any arguments of mine, but by a decree of your ancestors (γράμματα τῶν προγόνων), which they inscribed upon a brazen column (στήλην) erected in the citadel.… What, then, says the decree (τί οὖν λέγει τὰ γράμματα)? 'Let Arithmius,' it says (φησίν), 'of Zelia, the son of Pythonax, be accounted infamous and an enemy to the Athenians and their allies, both he and all his race.'… The sentence imported somewhat more, for, in the laws importing capital cases, it is enacted (γέγραπται) that 'when the legal punishment of a man's crime cannot be inflicted he may be put to death', and it was accounted meritorious to kill him. 'Let not the infamous man,' saith the law, 'be permitted to live' (καὶ ἄτιμος, φησί, τεθνάτω), intimating that he is free from guilt who executes this sentence (τοῦτο δὴ λέγει, καθαρὸν τὸν τούτων τινὰ ἀποκτείναντα εἶναι)."

In both cases it is doubtless enough to render φησί, "it says," its function being in each case to call pointed attention to the words quoted: but the "it" is by no means "indefinite" in the sense that its reference was not very definitely conceived. On the second instance of its occurrence Wolf comments: "s. ὁ φονικὸς νόμος," while Schaefer says:44

"Pleonastice positum cum γέγραπται praecesserit. Verumtamen h. I. sensum paulo magis juvat quam ubi post εἶπον, εἶπε, continuo sequitur ἔφην, ἔφη. Ad φησί subaudi ὁ νομοθέτης."

These instances will supply us with typical examples of the "absolute" φησί; and, in this sense, "subjectless φησί" is of very common occurrence indeed in Greek literature. 

But really "subjectless φησί," i. e., φησί without any implied subject in context or common knowledge, which therefore we must take quite indefinitely, is very rare indeed, if not non-existent. Perhaps one of the most likely instances of such a usage is offered us by a passage in Plutarch's "Consolatio ad Apollonium," 34. Holland's old version of it runs thus:46

"And verily in regard of him who is now in a blessed estate, it has not been naturall for him to remaine in this life longer than the terme prefixed and limited unto him; but after he had honestly performed the course of his time, it was needfull and requisit for him to take the way for to returne unto his destinie that called for him to come unto her."

From this we may at least learn that φησίν here presented some difficulty, as Holland passes it by unrendered. The common Latin version restores it, reading the last clause thus: "Sed ita postulabit natura ut hoc expleto fatale quod aiunt iter conficeret, revocante eum jam ad se natura"; the Greek running thus: "ἀλλʼ εὐτάκτως τοῦτον ἐκπλήσαντι πρὸς τὴν εἱμαρμένην ἐπανάγειν πορείαν, καλούσης αὐτῆς, φησίν, ἤδη πρὸς ἑαιτήν." The theory of the Latin version obviously is that φησίν here is to be taken indefinitely, that is as an index hand pointing to a current designation of death as an entering upon the "fated journey"—ἡ εἱμαρμένη πορεία. This is explained to us by Wyttenbach's note:

"φησίν] non debebat offendere viros doctos. Est ut ait poeta ille unde hoc sumptum est. Videt hoc et Reiskius. Correxi versionem. De Tragici dicto in Animadversibus dicetur."

Accordingly, in the Animadversions, he addresses himself first to showing that the expression here signalized was a current poetical saying—appealing to Plato,49 Julian, Philo; and then adds:

"Cæterum φησίν ita elliptice usitatum est: v. c. Plutarcho, p. 135 B., 817 D., Dion. Chrys., p. 493 D., 532 A., 562 B. Notavit et Uptonus ad Epict. in Indice. In annotatoribus ad Lambertum Bosium de Ellipsibus unus Schoettgenius, idque ex uno Paulo Apostolo hunc usum annotavit, p. 74. Et. Latine ita dicitur inquit, quod monuerunt J. F. Gronovius et A. Drakenborch. ad Livium xxiv. 3, J. A. Ernestus in Clav. Cic. voce Inquit."

It does not seem, however, that Wyttenbach would have us read the φησί here quite indefinitely, as adducing for example a current saying: judging from his own paraphrase this might appear to him as a certain exaggeration of its implication. Its office would seem rather to be to call attention to the words, to which it is adjoined, as quoted, and thus, in the good understanding implied to exist between the writer and his readers, to point definitely to its source: so that it might be a proper note to it to say, "subaudi ὁ τραγικός, vel ὁ ποιητής"—and this might be done with a considerable emphasis on the ὁ; nay, the actual name of the poet, well known to both writer and reader, though now lost to us, might equally well be the subauditum, and such, indeed, may be the implication of the subauditum suggested by Wyttenbach: ut ait poeta ille unde hoc scriptum est. Surely, an instance like this is far from a clear case of the absolutely indefinite or even generally undefining use of φησί. 

Among the references with which Wyttenbach supports his note, the most promising sends us to Epictetus, whose "Discourses" abound in the most varied use of φησί, and offer us at the same time one of our most valuable sources of knowledge of the Greek in common use near the times of the apostles. We meet with many instances here which it has been customary to explain as cases of φησί in a wholly indefinite reference. But the matter is somewhat complicated by the facts that we are not reading here Epictetus' "Discourses" pure and simple, but Arrian's report of them; and that Arrian may exercise his undoubted right to slip in a φησί of his own whenever he specially wishes to keep his readers' attention fixed upon the fact that they are his master's words he is setting down, or perhaps even merely out of the abiding sense, on his own part, that he is reporting Epictetus and not writing out of his own mind. When such a φησί occurs at the beginning of a section it gives no trouble: every reader recognizes it at once as Arrian's. But when it occurs unexpectedly in the midst of a vivacious discussion, the reader who is not carrying with him the sense of Arrian's personality, standing behind the Epictetus he is attending to, is very apt to be stumbled by it, and to resort to some explanation of it on the theory that it is Epictetus' own and is to find its interpretation in the context. An attempt has been made by Schenkl in the index to his edition of Epictetus to distinguish between the instances in which φησί occurs "inter Epicteti verba ab Arriano servata," and those in which it occurs "inter Arriani verba." It will be found that most of the instances where it has been thought markedly indefinite in its reference are classed by him in the second group and are thus made very definite indeed—the standing subauditum being "Epictetus." Opinions will, no doubt, differ as to the proper classification of a number of these: and in any case many instances remain which cannot naturally be so explained—occurring as they do in the midst of vividly conceived dramatic passages. In this very vividness of dramatic action, however, is doubtless to be found the explanation of these instances. So far are the verbs here from being impersonal, that the speakers in these little dialogues stood out before Epictetus' mind's eye as actual persons; and it is therefore that he so freely refers to them with his vivid φησί. 

The following are some of the most striking examples of his usage of the word. "But now we admit that virtue produces one thing, and we declare that approaching near to it is another thing, namely progress or improvement. Such a person, φησίν, is already able to read Chrysippus by himself. Indeed, sir, you are making great progress" (i. 4, 9). Here Schenkl suggests that the φησίν is Arrian's, and this would seem to be a good suggestion, as it illuminates the passage in more ways than one. If not, the subauditum would seem to be the collocutor of the paragraph: a "some one," no doubt, but rather the "some one" most prominent in the mind of writer and reader in this discussion. "But a man may say, Whence shall I get bread to eat, when I have nothing (καὶ πόθεν φάγω, φησί, μηδὲν ἔχων;)?" (i. 9, 8). Here again the φησί seems best explained as Arrian's (Schenkl): if not, the subauditum is again the collocutor prominent through the context, and only, in that sense, indefinite. "Who made these things and devised them? 'No one,' you say (φνσίν). O amazing shamelessness and stupidity" (i. 16, 8). The reference is to the collocutor. "They are thieves and robbers you may say (κλέπται, φησίν, εἰσι.…)" (i. 18, 3). Either Arrian's (Schenkl), or with the collocutor as the subauditum. "How can you conquer the opinion of another man? By applying terror to it, he replies (φησίν), I will conquer it" (i. 29, 12). Subaudi the collocutor. "For why, a man says (φησί), do I not know the beautiful and the ugly?" (ii. 11, ?). Either Arrian's (Schenkl), or subaudi the collocutor. "How, he replies (φησίν), am I not good?" (ii. 13, 17). Either Arrian's (Schenkl), or subaudi the collocutor. So also similarly in ii. 22, 4; iii. 2, 5; iii. 5, 1, etc. Cf. also ii. 23, 16; iii. 3, 12; ix. 15; xx. 12; xxvi. 19. Similarly, in the "Fragments" we have this: "They are amusing fellows, said he (ἔφη = Epictetus), who are proud of the things which are not in our power. A man says, I (ἐγώ, φησί) am better than you, for I possess much land and you are wasting with hunger. Another says (ἄλλος λέγει).…." ("Frag.," xviii. [Schw., 16]). Here the φησί is brought in as the initial member of a series and in contrast with ἄλλος λέγει: it would seem to be Epictetus' own, therefore, and to mean "says one," as distinguished from another; and thus it appears to be the most likely instance of the "indefinite φησί" in the whole mass. But even it seems an essentially different locution from the really indefinite "it is said," "on dit," "man sagt." 

A glance over the whole usage of φησί in Arrian-Epictetus leaves on the mind a keen sense of the lively way in which the word must have been interjected into Greek conversation, but does not greatly alter the impression of its essential implication which we derive from the general use of the word. Take a single instance of its current use in the "Discourses" in its relation to kindred words:

"So also Diogenes somewhere says (που λέγει) that there exists but one means of obtaining freedom—to die contentedly, and he writes (γράφει) to the king of the Persians, 'You cannot enslave the city of the Athenians, any more,' says he (φησίν), 'than fishes.' 'How? Can I not catch them?' 'If you catch them,' says he (φησίν), 'they will immediately leave you and be gone, just like fishes: for whatever one of them you catch dies, and if these men die when they are caught, what good will your preparations do you?' " (iv. 1, 30).

The lively effect given by such unexpected interpositions of φησίν is lost in our decorous translation of the New Testament examples: but it exists in them too. Thus: "But she, being urged on by her mother, 'Give me,' says she, 'here upon a charger, the head of John the Baptist' " (Matt. 14:8); "But he, 'Master, speak,' says he" (Luke 7:40); "But Peter to them, 'Repent,' says he, 'and be baptized each one of you' " (Acts 2:38); " 'Let those among you,' says he, 'that are able, go down with me' " (Acts 25:5); " 'To-morrow,' says he, 'thou shalt hear him' " (Acts 25:22); "But Paul, 'I am not mad,' says he, 'most noble Festus' " (Acts 26:25). The main function of φησί then would appear to be to keep the consciousness of the speaker reported clearly before the mind of the reader. It is therefore often used to mark the transition from indirect to direct quotation: and it lent itself readily, therefore, to mark the adduction both of objections and of literary citations. But, one would imagine, it did not very readily lend itself to vague and indefinite references. 

If we desire to find cases of "subjectless λέγει" in any way similar to those of φησί, we must apparently turn our back on profane Greek altogether. We have fortunately in Philo, however, an author, the circumstances of whose writing made literary quotation as frequent with him as oral is in the lively pages of Epictetus' "Discourses." And in Philo's treatises λέγει takes its place by the side of its more common kinsman φησί, and is used in much the same way, though naturally somewhat less frequently. In harmony with his fundamental viewpoint—which looked on the Scriptures as a body of oracular sayings—Philo adduces Scripture commonly with verbs of "saying"—φησί, λέγεται, λέγει, εἶπεν (γέγραπται falling into the background). Passages so adduced are often woven into the fabric of his discussion of the contents of Scripture; and where the words adduced are words of a speaker in the Biblical narrative, the subject of the φησί or λέγει which introduces them naturally is often this speaker—whether God or some other person. Equally often, however, the subject given immediately or indirectly in the context is something outside of the narrative that is dealt with: in this case it is sometimes Moses, or "the prophet," or "the lawgiver"—at other times, "the Holy Word," or "the sacred Word," or "the Oracle," or "the Oracles" (ὁ θεῖος λόγος, ὁ ἱερὸς λόγος, ὁ χρησμός, τὸ λόγιον, οἱ χρησμοί, τὰ λόγια)—at other times still it is "God," under various designations. Often, however, the verb—φησί or λέγει—stands not only without expressed subject, but equally without indicated subject. The rendering of these cases has given students of Philo some trouble, arising out of the apparent confusion, when the subject is expressed, of the reference of the verb,—now to a speaker in the text of Scripture and now to the author of the particular Scripture, to God as the author of all Scripture, or to Scripture itself conceived as a living Word. This apparent confusion is due solely to Philo's fundamental conception of Scripture as an oracular book, which leads him to deal with its text as itself the Word of God: he has himself fully explained the matter, and we should be able to steer clear of serious difficulties with his explanation in our hands. 

Nevertheless, a somewhat mechanical mode of dealing with his citations has produced, on more than one occasion, certain odd results. Prof. Ryle says:

"The commonest forms of quotation employed by Philo are φησί, εἶπεν, λέγει, λέγεται, γέγραπται γὰρ. Whether the subject of φησί be Moses or Scripture personified cannot in many cases be determined."

In no case is the subject strictly indeterminate, however, and the failure to determine it aright may introduce confusion. Thus, for example, in "De Confus. Ling.," § 26 (Mangey, i. 424), Philo mentions the Book of Judges, and cites it with the subjectless φησί. Prof. Ryle comments thus:

"He does not mention any opinion as to authorship, and introduces his quotation with his usual formula φησίν. We are hardly justified in assuming that Philo intended Moses as the subject of φησίν, and regarded him as the author of Judges (so Dr. Pick, Journal of Biblical Literature, 1884). Moses is doubtless often spoken of by Philo as if he were the personification of the Inspired Word; but we cannot safely extend this idea beyond the range of the Pentateuch. All that we can say is that φησίν, used in this quotation from Judges, refers either to the unknown writer of this book or to the personification of Holy Scripture."

Or else, we may add, to God, the real author, in Philo's conception, of every word of Scripture. Prof. Ryle, however, has not caught precisely Dr. Pick's meaning: Dr. Pick does not commit himself to the extravagant view that wherever subjectless φησί occurs in Philo the subauditum "Moses" is implied: he only says, in direct words, that here—in this special passage—"Moses is introduced as speaking." It would seem obvious that he had a text before him which read "Moses says," and not simply "says," at this place. This text was doubtless nothing other than Yonge's English translation, which reads Moses here, as often elsewhere with as little warrant: " 'For,' says Moses, 'Gideon swore, etc.' " The incident illustrates the evil of mechanically supplying a supplement to these subjectless verbs—which cannot indeed be understood except on the basis of Philo's primary principle, that it is all one to say "Moses says," "the Scripture says," or "God says." The simple fact here is that Philo quotes Judges, as he does the rest of Scripture, with the subjectless "says," and with the same implication, viz., that Judges is to him a part of the Word of God. 

As has been already hinted, by all means the commonest verb used by Philo thus,—without expressed or obviously indicated subject,—to introduce a Scripture passage, is φησί. Perhaps, however, the one instance to which we have incidentally adverted will suffice to illustrate the usage—other instances of which may be seen on nearly every page of Philo's treatises. It is of more interest for us to note that λέγει seems also to be used in the same subjectless way—examples of which may be seen, for instance, in the following places, "Legg. Allegor.," i, 15; ii, 4; iii, 8; "Quod Det. Pot. Insid.," 48; "De Posterit. Caini," 9; 22; 52; "De Gigant.," 11; 12; "De Confus. Ling.," 32; "De Migrat. Abrah.," 11; "Fragment, ex Joh. Monast." (ii, 668). In "Legg. Allegor.," i, 15, for instance, we have a string of quotations without obvious subject, introduced, the first by the subjectless φησίν, the next by the equally subjectless ἐπιφέρει πάλιν, and the third (from Exod. 20:23) by λέγει δὲ καὶ ἐν ἑτέροις. In "Legg. Allegor.," ii, 4, we have Gen. 2:19 introduced by λέγει γὰρ without any obvious subject. Yonge translates this too by "For Moses says": but to obtain warrant for this we should have to go back two pages and a half (of Richter's text), quite to the beginning of the treatise, where we find an apostrophe to the "prophet." In "De Posterit. Caini," 22, λέγει ἐπὶ μὲν Ἀβραὰμ οὕτως (Gen. 11:29), though Yonge supplies "Moses" again, that would seem to be demonstrably absurd, as the passage proceeds to place "Moses," in parallelism with Abraham, in the object. Similarly the passages adduced from "De Gigant.," 11 and 12 (Num. 14:44 and Deut. 34:6) are about Moses, and it would scarcely do to fill out the ellipsis of subject with his name. Examples need not, however, be multiplied. 

It would seem quite clear that both the subjectless φησί frequently, and the subjectless λέγει less often, occur in Philo after a fashion quite similar to the instances adduced from the New Testament. And it would seem to be equally clear that the lack of a subject in their case is not indicative of indefiniteness, but rather of definiteness in their reference. Philo does not adduce passages of Scripture with the bare φησί or λέγει, because he knows or cares very little whence they come or with what authority; but because he and his readers alike both know so well the source whence they are derived, and yield so unquestionably to its authority, that it is unnecessary to pause to indicate either. The use of the bare φησί or λέγει in citations from Scripture is in his case, obviously, the outgrowth and the culminating sign of his absolute confidence in Scripture as the living voice of God, fully recognized as such both by himself and his readers. In the same sense in which to the dying Sir Walter Scott there was but one "Book," to him and his readers there was but one authoritative divine Word, and all that was necessary in adducing it was to indicate the fact of adduction. The φησί or λέγει serves thus primarily the function of "quotation marks" in modern usage: but under such circumstances and with such implications that bare quotation marks carry with them the assurance that the words adduced are divine words. 

It would seem to be very easy, in these circumstances, to give ourselves more uneasiness than is at all necessary as to the precise subauditum which we are to assume with these verbs. It may serve very well to render them simply, "It says," with the implication that Philo is using the codex of Scripture as the living voice of God speaking to him and his readers. The case, in a word, would seem to be very similar to that of the common New Testament formula of quotation γέγραπται—meaning not that what is adduced is somewhere written, but that it is the authoritative law that is being adduced. Just so, "It says," in such a case would mean not that somebody or something says what is adduced, but that the Word of God says it. As the one usage is the natural outgrowth of the conception of the Scriptures as a written authoritative law, the other is the equally natural outgrowth of the conception of Scripture as the living voice of God. How very natural a development this usage is, may be illustrated by the fact that something very similar to it may be met with in colloquial English. In the same circles where we may hear God spoken of as simply "He," as if it were dangerous to name His name too freely, we may also occasionally hear the Bible quoted with a simple "It says," or even with an elision of the "it," as "'Tsays": and yet the "it," though treated thus cavalierly, is in reality a very emphatic "It" indeed—the phrase being the product of awe in the presence of "the Book," and importing that there is but one "It" that could be thought of in the case. Somewhat similarly, in the case of Philo, the Scriptures are cited with the bare φησί, λέγει, because, in his mind and in the circles which he addressed, there stood out so far above all other voices this one Voice of God embodied in His Scriptures, that none other would be thought of in the case. The phrase is the outgrowth of reverence for the Word and of unquestioning submission to it: and the fundamental fact is that no special subject is expressed simply because none was needed and it would be all one whether we understood as subject, Moses, the prophet and lawgiver—the holy or sacred Word or the oracle—or finally, God Himself. In any case, and with any subauditum, the real subject conceived as speaking is GOD.

If now, in the light of the facts we have thus brought to our recollection, we turn back to the New Testament passages in which the Old Testament is cited with a simple φησί or λέγει, it may not be impossible for us to perceive their real character and meaning. There would seem to be absolutely no warrant in Greek usage for taking λέγει, and but very little, if any, for taking φησί really indefinitely: and even if there were, it would be inconceivable that the New Testament writers, from their high conception of "Scripture," should have adduced Scripture with a simple "it is said"—somewhere, by some one—without implication of reverence toward the quoted words or recognition of the authority inherent in them. It is rather in the usage of Philo that we find the true analogue of these examples. Like Philo, the author of the Epistle to the Hebrews looks upon Scripture as an oracular book, and all that it says, God says to him: and accordingly, like Philo, he adduces its words with a simple "it says," with the full implication that this "it says" is a "God says" also. Whenever the same locution occurs elsewhere in the New Testament, it bears naturally the same implication. There is no reason why we should recognize the Philonic φησί in Heb. 8:5, and deny it in 1 Cor. 6:16: or why we should recognize the Philonic λέγει in Heb. 8:8 and deny it in Acts 13:35, Rom. 9:15, 15:10, 2 Cor. 6:2, Gal. 3:16, or in Eph. 4:8, 5:14. Only in case it were very clear that Paul did not share the high conception of Scripture as the living voice of God which underlies this usage in Philo and the Epistle to the Hebrews, could we hesitate to understand this phrase in him as we understand it in them. But we have seen that such is not the case: and his use in adducing Scripture of the subjectless φησί and λέγει quite in their manner is, rightly viewed, only another indication, among many, that his conception of Scripture was fundamentally the same with theirs, and it cannot be explained away on the assumption that it was fundamentally different. 

It does not indeed follow that on every occasion when a Scripture passage is introduced by a φησί or a λέγει it is to be explained as an instance of this subjectless usage—even though a subject for it is given or plainly implied in the immediate context. That is not possible even in Philo, where the introductory formula often finds its appropriate subject expressed in the preceding context. But it does follow that we need not and ought not resort to unnatural expedients to find a subject for such a φησί or λέγει in the context, or that acquiescing, whenever that seems more natural, in its subjectlessness, we should seek to explain away its high implications. Men may differ as to the number of clear instances of such a usage, that may be counted in the New Testament. But most will doubtless agree that some may be counted: and will doubtless place among them Eph. 4:8 and 5:14. Some will contend, no doubt, that in the latter of these texts, the passage adduced is not derived from the Old Testament at all. That, however, is "another story," on which we cannot enter now, but on which we must be content to differ. We pause only to say that we reckon among the reasons why we should think the citation here is derived from the Old Testament, just its adduction by διὸ λέγει—which would seem to advise us that Paul intended to quote the oracular Word. 

There may be room for difference of opinion again as to the precise subauditum which it will be most natural to assume with these subjectless verbs: whether ὁ θεός or ἡ γραφή. In our view it makes no real difference in their implication: for, in our view, the very essence of the case is, that, under the force of their conception of the Scriptures as an oracular book, it was all one to the New Testament writers whether they said "God says" or "Scripture says." This is made very clear, as their real standpoint, by their double identification of Scripture with God and God with Scripture, to which we adverted at the beginning of this paper, and by which Paul, for example, could say alike "the Scripture saith to Pharaoh" (Rom. 9:17) and "God.… saith, Thou wilt not give thy Holy One to see corruption" (Acts 13:34). We may well be content in the New Testament as in Philo to translate the phrase wherever it occurs, "It says"—with the implication that this "It says" is the same as "Scripture says," and that this "Scripture says" is the same as "God says." It is this implication that is really the fundamental fact in the case. 


Source: Revelation and Inspiration, by B. B. Warfield

By Topic


By Scripture

Old Testament









1 Samuel

2 Samuel

1 Kings

2 Kings

1 Chronicles

2 Chronicles








Song of Solomon


















New Testament







1 Corinthians

2 Corinthians





1 Thessalonians

2 Thessalonians

1 Timothy

2 Timothy





1 Peter

2 Peter

1 John

2 John

3 John



By Author

Latest Links