In verses 1-5 of this chapter of Romans, Paul focuses on the tension between God’s promises to Israel and Israel’s plight. The people to whom God has promised so much have, apparently, rejected God’s most profound revelation – His Son. In this verse, Paul presents the last of several privileges enjoyed by the Jews: Descent from the ‘fathers’ – the patriarchs – to whom God made promises which were still valid for those descended from them; and the Messiah who is ‘from’ them.
As Paul defined his own relationship to the Jews as “according to the flesh,” so he defines the Messiah’s relationship to them as well. Paul is not the spiritual kin of the Jews, but rather is their kin in terms of “this world.” So, to, the Messiah is ‘from’ the Jews in the strictly human sense. There is an implicit contrast between shared physical kinship and spiritual disunity.
Does Paul complete this thought by explicitly denoting a further aspect of the Messiah – that He is Deity? Or does he leave the contrast as implied, and offer a doxology to God the Father, praising Him for the fulfilled Messianic promise in Jesus?
These questions have been much debated, with scholars, grammarians, and translators failing to reach complete consensus. The questions turn on punctuation. Does the phrase “who is over all, God blessed forever” go with the previous clause – in which case Paul calls Jesus “God;” or does it stand alone as its own sentence – a doxology to the Father? Since early manuscripts of the NT lack all but rudimentary punctuation marks, these questions can only be decided by secondary evidence and interpretation.
Despite the varied opinions of scholars, there is substantial evidence that Paul is attributing “God” to the Messiah in this verse. This evidence is cumulative in nature. That is, I do not regard any one piece as being decisive, but when put together, the pieces strengthen one another and provide a strong inductive case for our conclusion. We may summarize this evidence as follows:
1. The phrase “the one who is over all” is most naturally taken as a relative clause modifying “the Messiah.” The Greek phrase ho ôn (“the one who is”) almost always modifies the preceding head noun, not one that follows1.
2. As Douglas Moo points out, Paul’s doxologies are never independent, but always are closely linked with the preceding context (Moo, Romans, p. 567). This context stresses the tragedy of the Jews rejecting their Messiah by enumerating the blessings God has promised the Jews, and which they could claim, if they would but believe. Paul laments that the Jews have not received the fulfillment of God’s promises, the most profound of which is the coming of the Messiah. The true irony of the Jews’ rejection of Jesus is that not only is He “from” them according to the flesh, He is – in fact – God over all. Their rejection is thus the greatest of all tragedies – a rejection of God Himself. If Paul is here breaking off his lament to praise God for sending the Messiah, this would tend to undercut the person of Christ: “I grieve that you have rejected Christ, who is from you according to natural descent, but praise be God who is over all for sending Jesus anyway!” Indeed, Paul’s continued grief is presupposed at the start of the next verse:
But it is not as though (ouch hoion de hoti). Supply estin after ouch: “But it is not such as that,” an old idiom, here alone in N.T. (RWP).
The transition between verses 5 and 6 is smoother if the doxology refers to Christ – as a statement of just how profound is the Jews’ rejection of Jesus. It is not impossible that Paul praises God in this moment – as the One who keeps His promises, even when His people reject Him. But if so, his doxology breaks not only Paul’s thought but his mood as well.
3. Doxologies to God in the Bible which contain the word “blessed” (Greek: eulogêtos; Hebrew: bârak) always place this word in the first position2. As Bruce Metzger notes, it is “altogether incredible that Paul, whose ear must have been perfectly familiar with this constantly recurring formula of praise, should in the solitary instance have departed from established usage” (Metzger, Punctuation, p. 107).
4. The qualifying phase “according to the flesh” implies a contrast, and Paul usually supplies this contrast in the immediate context. It is true that this is not always the case; Paul implies but does not delineate the contrast in verse 3. Nevertheless, in most cases he does (e.g., Romans 1:3-4), and when we find a phrase that provides this contrast as we do here, it would seem probable that this is Paul’s intention. As Metzger notes:
“If Christ did not have some other relation, or stand in some other position besides the one connected with the Jews, and different from it, there would seem to be no occasion for mentioning any such limitation. In other words, Paul’s language here, having called attention to the human ancestry of Christ as a Jew (‘according to the flesh’), naturally implies that he was more than a Jew” (Metzger, Punctuation, pp. 103-104).
5. While a slight majority of later Greek manuscripts favor a doxology to the Father, these are not conclusive. No Greek manuscript prior to the 5th Century has been found with a full stop after “flesh.” Other ancient translations, however, almost all take “God” as attributed to Christ, even those prior to the 5th Century..
6. The majority of early Church Fathers understood Paul to be calling Christ “God” in this verse. Only two Greek fathers held the opposite view. Some have suggested that the later fathers of the church argued in favor of Christ being called “God” in response to Arius and his followers. While this is certainly true, as Moo rightly points out, the evidence is too early and too widespread to ignore (Moo, Romans, p. 566 n. 64). Further, at least one of the dissenting fathers apparently taught that the incarnation was nothing more than a supreme instance of inspiration and grace; if some of the fathers allowed their theology to guide their interpretation contra Arius, we must allow that this one may have done so for his own theological ends. The other wrote in the 9th century, far too late to be of much use in this discussion.
The primary objection to seeing Paul as calling Christ “God” in this verse is based on Paul’s usage of “God” elsewhere. It is argued by some that because Paul does not use “God” of Jesus elsewhere, that he cannot be doing so here – particularly in what appears to be such a casual way. However, Paul almost certainly calls Jesus “God” in one other verse (Titus 2:13), attributes to Christ all the fullness Deity (Col 2:9), quotes OT passages referring to YHWH and directs them at Jesus (e.g., Isaiah 45:23; Philippians 2:10), and speaks of Christ in the highest possible terms (Col 2:3). For one as devoted to Christ as Paul was, it is not surprising that he attributes full divine status to the Messiah. And the casual manner in which he does so merely demonstrates that for Paul, such an attribution was not such a rare occurrence – either in his writing or his preaching.
Thus, it seems on balance, the evidence favors the view that Paul is here attributing to Christ a title he normally reserves for the Father alone. The One rejected by the Jews is supreme over all as God blessed forever!
To get rid of the bright testimony here borne to the supreme divinity of Christ, various expedients have been adopted:
(1) To place a period, either after the words “concerning the flesh Christ came,” rendering the next clause as a doxology to the Father–“God who is over all be blessed for ever”; or after the word “all”–thus, “Christ came, who is over all: God be blessed.”, &c. [ERASMUS, LOCKE, FRITZSCHE, MEYER, JOWETT, &c.]. But it is fatal to this view, as even Socinus admits, that in other Scripture doxologies the word “Blessed” precedes the name of God on whom the blessing is invoked (thus: “Blessed be God,” Psa 78:35; “Blessed be the Lord God, the God of Israel,” Psa 72:18). Besides, any such doxology here would be “unmeaning and frigid in the extreme”; the sad subject on which he was entering suggesting anything but a doxology, even in connection with Christ’s Incarnation [ALFORD].
(2) To transpose the words rendered “who is”; in which case the rendering would be, “whose (that is, the fathers’) is Christ according to the flesh” [CRELLIUS, WHISTON, TAYLOR, WHITBY]. But this is a desperate expedient, in the face of all manuscript authority; as is also the conjecture of GROTIUS and others, that the word “God” should be omitted from the text. It remains then, that we have here no doxology at all, but a naked statement of fact, that while Christ is “of” the Israelitish nation “as concerning the flesh,” He is, in another respect, “God over all, blessed for ever.” (In 2 Cor 11:31 the very Greek phrase which is here rendered “who is,” is used in the same sense; and compare Rom 1:25 Greek). In this view of the passage, as a testimony to the supreme divinity of Christ, besides all the orthodox fathers, some of the ablest modern critics concur [BENGEL, THOLUCK, STUART, OLSHAUSEN, PHILIPPI, ALFORD, &c.] (JFB).
ων οι πατερες και εξ ων ο χριστος το κατα σαρκα ο ων επι παντων θεος ευλογητος εις τους αιωνας αμην
ÔN hOI PATERES KAI EX ÔN hO CHRISTOS TO KATA SARKA hO ÔN EPI PANTÔN THEOS EULOGÊTOS EIS TOUS AIÔNAS AMÊN
Whose [are] the fathers and from whom [is] the Christ according to the flesh, who is over all God blessed to the ages. Amen.
- Insofar as the physical is concerned (BDF).
- As far as physical descent is concerned (Moule, Idiom; c.f., BAGD 744a).
- Used of natural or physical origin, generation or relationship, born of natural generation (Thayer).
- As concerning the flesh (to kata sarka). Accusative of general reference, “as to the according to the flesh.” Paul limits the descent of Jesus from the Jews to his human side as he did in Rom 1:3. (RWP).
OTHER VIEWS CONSIDERED
The New Word Translation renders the latter half of this verse: “from whom Christ sprang according to the flesh: God who is over all be blest forever. Amen.” (NWT, 1950). The Watchtower provides a brief defense of this translation in the Appendix of several editions of their Bible (e.g., 1950, 1984). Greg Stafford has written a more extensive defense (Stafford, pp. 143 – 152). Both will be examined, below.
objection: The Watchtower cites several scholars who state that grammar alone cannot decide the most accurate rendering of this verse, It quotes AT Robertson at length:
As is well known, the difficulty here is a matter of exegesis and the punctuation of the editor will be made according to his theology. But it may be said in brief that the natural way to take ò wn and qeos is in apposition to ò CristoV. – Grammar, page 1108. (NWT, 1950, Appendix – Romans 9:5).
The Watchtower immediately follows this quote by saying: “We take this passage as a reference to God.” It states that the grammar “admits” this rendering, and cites several translations that agree with theirs, including Moffatt, the RSV, and the Riverside New Testament.
Response: It must be frankly said that this is not so much a defense as it is an admission that the NWTTC rendered this verse on the basis of theology. While Robertson supports doing so to some degree, he also points out that the “natural” way to understand this verse is to link “who is God” with “the Christ.” Of the four other translations listed as agreeing with the NWT translation, one – the RSV – was revised in 1989 so that, in the main text, the doxology is now attributed to Christ (a footnote reflects the older rendering). Similar revisions occur in UBS3 (1975) and NA26 (1979). While exegesis must decide the proper punctuation of this verse, the Watchtower offers no exegetical reasons for its translation whatsoever.
objection: Greg Stafford’s detailed defense of the NWT punctuation of Romans 9:5 is divided into five sections: “Evidence from early translations;” “Punctuation in early Greek manuscripts;” “The view of early church fathers;” “Grammatical analysis;” and “Contextual considerations.” These generally parallel corresponding sections in Bruce Metzger’s “The Punctuation of Rom. 9:5.”
Mr. Stafford concludes the first section, “Evidence from early translations,” as follows:
The above [summary of Metzger’s review of early translations] constitutes evidence in favor of the rendering found in the NIV and other, similar translations. But this early evidence is countered by other early evidence relating directly to the transmission of the Greek text itself” (Stafford, p. 144).
Response: Mr. Stafford concedes that the evidence from early translations of the Greek text favor a rendering which ascribes the doxology to Christ. However, he says that this evidence is “countered” by other evidence from early Greek manuscripts. Thus, if the manuscript evidence can be shown to be questionable, then – according to Mr. Stafford’s own argument – it would no longer “counter” the evidence from early translations. Even if this is not the case, most of the early translations Metzger reviews are dependent upon Greek exemplars, and hence provide indirect evidence of early Greek manuscripts that support a doxology to Christ. Also in this regard, we should consider the patristic evidence from Greek-reading fathers; if they support Christ being called “God,” this would strengthen Metzger’s contention and weaken Mr. Stafford’s.
objection: In the second section of his defense, “Punctuation in early Greek manuscripts,” Mr. Stafford offers evidence ‘countering’ that provided by early translations. Mr. Stafford notes the specific Greek punctuation marks that concern us: “A middle point is usually taken to indicate a pause such as we might indicate by use of a colon or comma, while a high point is generally used to indicate a full stop” (Stafford, p. 144 n39). He summarizes Metzger, indicating that some manuscripts – notably Codex A – have a middle point after “flesh,” while others “such as B, L, 0142, and 0151 have a high point after ‘flesh,’ also indicating a pause or break of some kind” (Ibid.). Mr. Stafford notes that Metzger identifies Codex B as a middle point, but argues that “it is quite possible” that B is actually a high point (Ibid, n39). Mr. Stafford agrees with Metzger that the use of punctuation in these manuscripts is “oddly placed,” but argues that Codex A is an exception, “and yet uses a mid- or highpoint and what appears to be a small space between sarka and the article ho” (Ibid, p. 145). Mr. Stafford concludes this section with another agreement with Metzger:
Metzger is probably right in saying that “the most that can be inferred from the presence of a point in the middle position after sarka [sarka, ‘flesh’] in the majority of the uncial manuscripts is that scribes felt some kind of pause was appropriate at this juncture of the sentence” (Metzger, p. 99 in Stafford, p. 145).
Response: Mr. Stafford’s burden, as he has set it forth, is to provide sufficient evidence from the punctuation of early Greek manuscripts to “counter” the significant evidence from early translations. We may first note that none of the Greek manuscripts is earlier than 4th Century; in fact some of the translations Metzger discusses are actually earlier than the Greek manuscripts Mr. Stafford cites:
|Codex Vaticanus (B)||4th Century||Mid-point (Stafford: High-point)|
|Codex Alexandrinus (A)||5th Century||Mid-point|
|Peshitta (Syriac)||5th – 6th||Comma|
|Coptic (Beatty MS)||6th||Comma|
|Codex Regius (L)||8th Century||High-point|
Figure 1 (Greek MSS in brown)
Any information we may glean about the punctuation of Romans 9:5 from these manuscripts is secondary evidence, at best. With that caveat in mind, the evidence itself does not appear to be particularly decisive in Mr. Stafford’s favor. While Mr. Stafford notes that the high-point indicates a “full-stop,” he defines the mid-point as being equivalent to “a colon or a comma.” I take Mr. Stafford to mean that either a mid-point or a high-point supports the kind of “pause” indicated by the NWT’s colon after “flesh.” However, I don’t believe the mid-point can be so construed. It is true that Metzger refers to a “mid-point colon,” but the sources I’ve found that discuss the mid-point indicate that it is equivalent to our comma:
Two kinds of stop may be seen in texts of the late ii. B.C. and of i. B.C.: one is placed high in the line [Greek Ano Stigme], the other in a middle position [Greek Mese Stigme]… Normally the high stop marks period end. The stop in the middle position serves as a subdivision inside the period, with the effect of a modern comma (Turner & Parsons, p. 9).
The point at the top of the line (·) (stigmh teleia, ‘high point’) was a full stop; that on the line (.) (upostigmh) was equal to our semicolon, while a middle point (stigmh mesh) was equivalent to our comma. But gradually changes came over these stops till the top point was equal to our colon, the bottom point became a full stop, and the middle point vanished, and about the ninth century A.D. the comma (,) took its place (Robertson, Grammar, p. 242).
Thus, the manuscripts that contain a mid-point cannot be considered evidence in favor of the NWT punctuation. It will be observed (see figure 1) that the manuscripts with the high-point date from the 8th Century or later, which can hardly be decisive in determining how Paul or his amanuensis would have punctuated this verse. With regard to Codex B and Mr. Stafford’s disagreement with Metzger, it is not at all clear to me that the placement of the point after sarka differs markedly from other mid-points in the surrounding context. Though I have studied textual criticism at the graduate level, I am not an expert in Biblical texts. My opinion, therefore, is not to be valued above Mr. Stafford’s. But the same is not true of Dr. Metzger’s opinion. He is one of the most well-known and widely-respected scholars in the field of NT textual criticism. He has worked with primary texts throughout his long career, taught NT textual criticism at Princeton, served on the Editorial Committee of the United Bible Societies Greek New Testament (UBS GNT), and on the Translation Committee for the NRSV. On balance, it would seem Dr. Metzger is in a better position to correctly identify the point in question than is Mr. Stafford.
Mr. Stafford’s agreement with Metzger’s conclusion that the presence of the mid-point indicates “some kind of pause” is not an argument in favor of taking “God” in reference to the Father. Metzger says the evidence is inconclusive. Therefore, it does not appear that Mr. Stafford has met his burden and demonstrated evidence that counters that of the early translations which attribute “God” to Christ.
objection: The next ‘leg’ of Mr. Stafford’s argument is “The view of the early church Fathers.” Mr. Stafford first seeks to cast doubt on the earliest father who quotes Romans 9:5. Metzger cites Irenaeus’s 2nd Century Against Heresies (3.16.3) as very early evidence that “God” was taken as a reference to Christ. In response, Mr. Stafford quotes Abbot, who notes that Irenaeus “does not quote it to prove Christ is qeoV [theos, G-god]” (Abbot, quoted in Stafford, p. 145). Stafford also cites Abbot’s observation that Irenaeus’ text is preserved only in Old Latin, “which, of course, cannot determine the construction which Irenaeus put upon the Greek” (Ibid.). Finally, Mr. Stafford repeats Abbot’s argument that the title “the God over all” is elsewhere always used by Irenaeus of the Father, and if Irenaeus intended to call Christ by that title, “the question naturally arises, how the Father can be ‘the God over all,’ unless the term ‘God’ as applied to Christ is used in a lower sense” (Ibid., p. 146, n47).
Mr. Stafford goes on to attempt to undermine the evidence of Hippolytus by pointing out that for Hippolytus, while “God over all” refers to Christ in this verse, elsewhere, Hippolytus makes it clear that the Father is the “Lord” of Christ, and thus the title “God over all” must be seen in a qualified sense. Mr. Stafford argues that even though Hippolytus refers to Christ in this verse “in a somewhat Trinitarian sense,” he does so because he interprets the relationship of God to Christ as “light from light, or water from fountain, or as a ray from the sun” (Against Noetus, 11). Such analogies, says Mr. Stafford, are not used in the Bible.
Mr. Stafford notes that while almost all of the fathers cited by Metzger attribute “God” to Christ, there are two Greek fathers who do not: “Tarsus [sic] and Photius” (Stafford, p. 146). Mr. Stafford concludes this section with a long quote from Metzger, which culminates as follows:
The prevailing patristic interpretation of the passage [which supports the attribution of “God” to Christ] is altogether counterbalanced by what we have seen came to be the prevailing scribal tradition of punctuation in the later manuscripts … each tradition neutralizing, so to speak, the force of the other (Metzger in Stafford, p. 147).
Response: In his examination of Irenaeus, Mr. Stafford stands upon the broad shoulders of Ezra Abbot, the noted 19th Century Unitarian scholar. In most cases, this would be a sound strategy, but I do not believe that it is, this time. Abbot attempts to demonstrate that it is “doubtful” that Irenaeus attributed “God” to Christ in Romans 9:5 (Abbot, Romans 9:5, p. 136), but his arguments are uncharacteristically strained and unconvincing. He points out that Irenaeus’ text is preserved only in Old Latin, and thus cannot prove how Irenaeus understood the Greek. But this is an ad hoc argument. Abbot has not demonstrated that the Old Latin is inaccurate at this point in the text, and Abbot himself accepts its accuracy in his subsequent arguments. Abbot says that Irenaeus is not using this verse to prove Jesus is God, but to demonstrate the unity of the Christ with the man, Jesus. This is beside the point. Irenaeus quotes the entire verse and attributes the latter half to Christ. Finally, Abbot argues that the title, “the God over all” is used throughout Against Heresies and very often elsewhere, as an exclusive designation of the Father. But Abbot is being arbitrary with the evidence. Abbot says that the “absolutely decisive” evidence that Paul did not call Jesus “God” is that he does not do so elsewhere (he rejects Titus 2:13). Thus, by Abbot’s own methodology, if a writer frequently calls Jesus “God,” there would be no “absolutely decisive” evidence against him doing so in Romans 9:5. Irenaeus, of course, regularly calls Jesus “God,” and even speaks of Him in terms equivalent to being such “over all” (e.g., Against Heresies, 2.13.8; 3.6.1, 3.8.3). There is every reason to accept the testimony of Irenaeus. This father of the early church, long before the Arian controversy, understood Romans 9:5 to call Christ “God over all.”
Mr. Stafford’s comments about the third Century father, Hippolytus, are also largely derived from Ezra Abbot. Mr. Stafford and Abbot both note that Hippolytus’ first reference to Romans 9:5 in his work Against Noetus is in the context of answering the Noetians’ modalistic interpretation of this verse – that is, that “God over all” was attributed to Christ and hence made Him the Father. Both Abbot and Mr. Stafford, however, miss two important points:
1. The Noetians not only understood the latter half of Romans 9:5 as referring to Christ, they apparently were publicly promoting this interpretation in support of their theology. It would seem far easier for such a misinterpretation to grow if it were planted in the soil of widespread understanding that Romans 9:5 called Christ “God.” If the early church understood that this verse actually concluded with a doxology to the Father, Noetian eisegesis would certainly have been countered with arguments making this very point. But this is not what the record shows.
2. Hippolytus answers the Noetians by agreeing that this verse attributes “God over all” to Christ, but explains that this fact does not mean that Jesus is the Father. Again, if the general understanding of the church was that Romans 9:5 contained a doxology to the Father, this argument would have suited Hippolytus’ apologetic much better than the one he actually offers – and, indeed, such a view would have been more in accord with the theology of Christ’s subordination that Mr. Stafford claims Hippolytus believed and taught.
On this last point, Mr. Stafford says that Hippolytus understood Paul to be calling Christ “God over all” in a “somewhat Trinitarian sense,” but Hippolytus understood “over all” to be qualified “in such a way that allowed the Father to be Lord over Christ” (Stafford, p. 146). One wonders which Trinitarian creed Mr. Stafford has in mind that denies the Father’s headship over Christ (1 Corinthians 11:3)? Hippolytus teaches that Christ is “God over all,” but is not the Father, and in fact is actually subordinate to the Father. This teaching is not Trinitarian “in a sense,” but Trinitarian in every sense.
Mr. Stafford also says that the Bible does not use the same language Hippolytus does to describe the relationship of the Father to Christ. This is a red herring. Hippolytus’ explanation of how Christ can be “God over all” does not obviate the fact that he understands Romans 9:5 to attribute this phrase to Christ, not the Father. Hippolytus, writing in the third Century – well before the Arian controversy – answers a modalistic interpretation of Romans 9:5 in part by agreeing that this phrase describes the Son. This is very strong evidence that in the earliest records available to us, Romans 9:5 was consistently viewed as calling the Christ “God over all,” regardless of how individual writers may have understood that title.
Mr. Stafford mentions that Metzger lists Tertullian and “several other early writers” who support the view that “God” in this verse refers to Christ. Abbot is more forceful in admitting that the Latin fathers almost to a man attribute “God” to Christ:
But Mr. Stafford notes that Metzger also lists two Greek fathers who refer to “God over all” as a doxology to the Father: “”Tarsus [sic] and Photius” (Stafford, p. 146). The first, Diodore of Tarsus (d. 390 AD), “emphasized the humanity of Christ tending to make the incarnation nothing more than a supreme instance of inspiration and grace” (The Ecole Glossary). Abbot argues that we should disregard the testimony of the great majority of the fathers (who, of course, support the opposite view), because all it proves is that they interpreted an ambiguous grammatical construction to suit their theology (Abbot, Romans 9:5, p. 133). But sauce for the goose is sauce for the gander; the same may also be said of Diodore. In any event, Diodore’s view was certainly not pervasive in the fourth Century, as the evidence presented above demonstrates. Indeed, not even his own star students, John Chrysostom and Theodoret, followed their teacher in his view of Romans 9:5.4The second Greek father mentioned by Metzger, Photius (d. 897 AD), really is far too late a witness to have much, if any, bearing on the correct punctuation of this verse.
Mr. Stafford, by way of his concluding quotation of Metzger, suggests that the patristic evidence is completely balanced by the textual evidence of later Greek manuscripts. However, while I have great respect for Dr. Metzger, I think he is giving ground far too easily, here. First, as Metzger notes, there is no evidence of any punctuation (mid- or high-point) after sarka in Greek manuscripts prior to the fourth Century. The fourth Century Vaticanus and fifth Century Alexandrinus contain mid-points, which are not conclusive evidence of a full stop. They may, in fact, indicate that a comma was intended – as reflected in early translations of the same period. The testimonies of Irenaeus, Hippolytus, Tertullian, Cyprian, and Novatian all date from the same period, and cannot be ignored simply on the basis of anti-Arian bias (given that all predate Arius and the controversy that bears his name). While some scholars have overstated the importance of the patristic evidence, others such as Moo (quoted above), Sanday and Headlam (Critical and Exegetical Commentary on the Epistle of the Romans, p. 234), Cranfield (Romans, 469-70), and Faccio (De divinitate Christi justa S. Paulum, Rom 9, 5) present a balanced view in which the patristic evidence is placed in its proper perspective. Dwight presents the case cogently:
The value of patristic interpretation may be questioned, indeed, and in the case of some of the fathers it is possible that reasons may be suggested which influenced their minds, apart from the mere language which is used by the Apostle. But whatever may be said in this way, and however we may estimate these writers, their substantial or complete unanimity is a circumstance which should not be disregarded (Dwight, p. 42).
On balance, it is fair to say that as early as the 2nd Century, Christians were quoting Romans 9:5 in such a way that Christ was called “God,” and placing those quotes in settings rich with other acclamations of Christ’s Deity. This fact is hardly matched by ambiguous evidence of haphazard punctuation marks from later Greek manuscripts5.
objection: In his “Grammatical Analysis” section, Mr. Stafford argues that taking “God who is over all” in reference to Christ must be understood as “an appositive for ‘Christ according to the flesh,’ which would then create a conflict with Trinitarian thinking in terms of a deification of Christ’s human nature” (Stafford, p. 147). He notes that Murray J. Harris and others “attempt to find an antithesis in this verse between Christ’s human and divine natures” (Ibid., p. 148). Stafford argues that there is no antithesis to “according to the flesh,” but that Paul uses it in the same way he does in verse 3. Mr. Stafford accuses Harris of redefining theos as “a category of being” which is not articulated in the Bible, and thus importing a post-Biblical theology into the text.
Mr. Stafford repeats a common objection to “blessed forever” being attributed to Christ on the basis that eulogêtos is never used of Christ in the Greek New Testament. He points to the overwhelming number of times Paul uses theos of the Father.
Mr. Stafford interacts with the arguments raised by Harris and others regarding the placement of eulogêtos, and notes what he sees as a double-standard in their methodologies, when they argue on the basis of regular usage of eulogêtos but disregard Paul’s regular usage of theos. Mr. Stafford accuses Harris of not “fully appreciating” Abbot’s point about the position of eulogêtos in Romans 9:5: “Paul wishes to stress … the overruling providence of God as ‘the Ruler over All'” (Abbot in Stafford, p. 150).
Mr. Stafford concludes this section of his defense of the NWT rendering as follows: “The grammatical arguments given in support of the translation which makes theos predicate for Christ are relevant, but they are certainly not incontrovertible” (Ibid., p. 151).
Response: Mr. Stafford’s asserts that if “God who is over all” refers to Christ, it is appositional to “Christ according to the flesh,” which results in deifying Christ’s humanity. This assertion does not rest on any solid grammatical ground. Apposition merely requires that two substantives in close proximity refer to the same person or thing (Wallace, p. 48). An appositive need not modify intervening relative clauses (e.g., 2 Corinthians 11:31). As for Mr. Stafford’s contention that “according to the flesh” need not imply an antithesis, most scholars – even those advocating Romans 9:5b as a doxology to the Father – disagree. Abbot, for example, says, “the phrase kata sarka undoubtedly implies an antithesis” (Abbot, Romans 9:5, p. 101). The question turns not on whether an antithesis is implied, but whether it must be explicitly stated. Most scholars agree that it need not be explicitly stated, as verse 3 indicates.6 Nevertheless, in many cases, the antithesis expressly follows (e.g., Romans 1:3-4) and since “God who is over all, etc.” provides such an antithesis, the burden of proof lies with Mr. Stafford and those who agree with him that such is not the case here.
In response to Mr. Stafford’s assertion that Harris redefines theos as an unbiblical “category of being,” I would point out that the Bible does, on several occasions, use theos in this very sense and uses other words that mean the same thing. First, many experts in Greek grammar have noted that anarthrous nouns in general often signify the qualities, essence, or nature of the noun.7 If this principle is true of other nouns, we may wonder why it cannot be true of theos?8Second, in Galatians 4:8, Paul speaks in negative terms of those who are “not gods by nature” (mê phusis ousin theois).9 Paul’s statement presupposes that there is at least One who is “God by nature,” and thus the concept of Deity (“that which makes God, God”) is a Biblical concept. This concept is echoed in Acts 17:29 (where theios means “divine nature”), Col 2:9 (where theotes signifies “Deity”), and 2 Peter 1:4 (theios, again, signifying “the divine nature”). We may debate what each specific reference to “divine nature / Deity” may mean in its context, but it cannot be denied that the idea that God has a unique nature which sets Him apart from all creation is a Biblical teaching. The question is, then, is theos ever used to signify the essence, nature, or qualities of “God?” The Watchtower itself argues that theos in John 1:1c is used in this manner: “Careful translators recognize that the articular construction of the noun points to an identity, a personality, whereas an anarthrous construction point to a quality about someone” (NWT 1950, p. 774). Mr. Stafford agrees that this semantic sense is present in theos in this same verse:
The inspired apostle shows that the Word has the same kind of nature and qualities that “the God” (not simply the “person”) he existed with has (Stafford, p. 349).
The Watchtower and Mr. Stafford, of course, do not regard the qualitative aspect as the only semantic force present in theos in John 1:1c,10 but they acknowledge its existence, and therefore concede that theos is used in the Bible to signify the nature of God.Other verses that use theos in a qualitative sense include (LXX): Deuteronomy 4:35, Joshua 24:17, 3 Kings 18:24, 27; (GNT): John 1:1; Romans 9:5; 1 Corinthians 8:4; 2 Corinthians 1:3; 2 Thessalonians 2:4.
Mr. Stafford’s objection that eulogêtos (“blessed”) is nowhere used of Christ in the NT is certainly true, but we must place it in perspective. Apart from this verse, there are only seven other instances of its use (four others by Paul). This would seem an insufficient sample from which to draw firm conclusions. It is used both of men and God in the LXX, as is it’s close cousin, eulogeô, which is used of Christ six times in the NT. There is thus scarce evidence that Paul would have refrained from using eulogêtos of Christ on this occasion.
The argument Mr. Stafford raises about Paul’s “regular” use of theos is, I believe, the most reasonable objection to “God” being attributed to Christ in this verse. It is an argument raised by virtually every proponent of the ‘doxology to the Father’ view, though some treat it as proving their view, which it cannot do.11 In response, most scholars who advocate the view argued here have answered in two general ways:
1. Paul calls Jesus theos in Titus 2:13.
2. Paul refers to Jesus in the highest possible terms elsewhere, effectively calling Him “God” by using other terms. Therefore, it is not surprising to find Him called “God” here.
While I agree with both points, and have utilized them in the Commentary (above), I believe there is another pertinent point to raise. Paul uses theos about 490 times in his writings. If he has called Jesus “God” here and in Titus 2:13, that represents about .4%. The NT as a whole contains “God” about 1315 times, and most Trinitarians would – at most – accept seven verses as calling Jesus “God.” This is a ratio of .5%. If Paul is here referring “God” to Jesus, he is not doing so outside the ‘norm’ of the NT. John, who attributes theos to Jesus more than any other writer, only does so three times out of just over 200 uses – about 1%. We are dealing with a sample of data in which there is very rare use of theos in reference to Jesus (though, of course, the data also show that Jesus is exalted to the highest degree using other terms). It therefore cannot be special pleading to say that Paul is doing so, here.
If it is not special pleading to claim that Paul could have called Jesus “God” in this verse, the question then turns on whether one believes that Paul knew and approved of Thomas’ confession, as recorded in John 20:28 (that is, that other Apostles were comfortable with this affirmation, albeit on rare occasions); On whether Paul ascribes Deity to Jesus in Col 2:9; On whether he exalts Jesus with the name of YHWH and says of Him that every knee will bow in worship (the clear sense of in the OT setting) in Phillipians 2:10 (c.f., Isaiah 45:23). If Paul can say of Christ that He is the Lord of Glory; the Lord from Heaven; the Lord of the living and dead; that in Him are hid all the treasures of wisdom and knowledge; and that He is raised above all principalities and powers and might and dominion, and every name that is named, not only in this world and the next; If Paul can say all this, he can surely call Christ “God above all” and not exceed the bounds of reason or usage.
Mr. Stafford’s objection that Metzger and Harris are applying a double-standard with regard to the placement of eulogêtos (in comparison to Paul’s usage of theos) is ultimately a tu quoque (“you too”) argument. Even if Paul is using theos counter to his normal usage, this does not relieve Mr. Stafford (and other scholars who argue as he does) from accounting for the unprecedented placement of eulogêtos in this verse. Mr. Stafford undertakes this burden by, once again, relying heavily on Ezra Abbot. Mr. Stafford cites Psalm 67:19 as an example of eulogêtos occurring after theos, but most scholars – including Abbot – recognize that because of its chiastic structure, this verse is not a valid counterexample (see Note 2, below). Abbot’s argument about the placement of eulogêtos, which Mr. Stafford says that Harris is not fully appreciating, is essentially that Greek syntax is flexible enough to allow Paul to place eulogêtos after theos, in order to emphasize God’s role in sending the Messiah. It is difficult to understand how Harris could fail to appreciate Abbot’s argument when he spends almost a page dealing with it (Harris, pp. 162-163). Harris’ response – which I believe Mr. Stafford ‘fails to appreciate’ (or at least interact with) – is two-pronged:
1. A doxology is a fixed formulaic phrase, not merely single word like theos. When a phrase has been fixed, particularly one of such devotional meaning, it is proper to take the established norm into account when determining the meaning of an isolated variation.12
2. “It is hard to imagine that nowhere else in the Greek Bible does the subject in a doxology bear an emphasis comparable to that in Romans 9:5 so that the customary word order is reversed” (Harris, p. 163).13
This final point seems to me to be decisive. If Abbot and Mr. Stafford contend that the word order of a fixed doxology can be varied to place emphasis in certain contexts, they must do so in the absence of any evidence supporting them.
Mr. Stafford concludes that while the grammatical arguments in favor of “God” being ascribed to Christ are “relevant,” they are not “incontrovertible.” The same can certainly be said of the arguments attributing the doxology to the Father. I would suggest, based on the evidence presented here, the probabilities strongly favor the former view.
objection: In his concluding section, “Contextual considerations,” Mr. Stafford notes that Metzger and Harris have both argued that the preceding context of Romans 9:5 supports the view that “God over all” is attributed to Christ, while there appears to be no real support for a doxology to the Father. Mr. Stafford replies that both Metzger and Harris have “failed to appreciate” how Paul’s preceding sadness turns to joy in Romans 9:5 and “is expressed in praise to God for sending Christ ‘according to the flesh'” (Stafford, p. 152). Mr. Stafford complains that both Metzger and Harris fail to interact with Abbot’s detailed argument in support of this view. Finally, Mr. Stafford concludes with a quote from Dwight who, while arguing in favor of “God over all” being ascribed to Christ, nevertheless acknowledges that a doxology to the Father would not be wholly out of place in this context.
Response: In arguing for Paul’s sorrow turning to joy, Mr. Stafford says that Paul breaks out in praise to the Father for sending Christ “according to the flesh.” But we have noted earlier that Mr. Stafford has “failed to appreciate” the implied contrast in this phrase that every scholar Mr. Stafford mentions (and many that he has not) – including his exemplar, Ezra Abbot – recognize as being there. Paul is not here speaking about Christ “coming as a man,” but rather of his descent from the Patriarchs. In the catalog of blessings the Father has bestowed upon His people, the Messiah is the greatest, and last. Paul feels such kinship for his ‘brother’ Jews, and so longs for their salvation, that he wishes himself “accursed” and “separated from Christ” if by such a profound sacrifice the Jews could be saved. This is no garden-variety sorrow; indeed, such a willingness to consign oneself to eternal separation from Christ is found nowhere else in Scripture. It speaks both of Paul’s great love for his people, and his even greater love for Christ, that he would so example Christ’s sacrificial love towards those who – in many cases – beat and stoned him, and wished him dead.
While I don’t regard it impossible that Paul breaks out of his grief to praise God (not for sending Messiah “in the flesh,” but for being a faithful God who keeps His promises, despite the unfaithfulness of a stubborn people), nevertheless such a mood-swing seems most unlikely. The next verse begins with “But it is not as though the word of God has failed.” The presupposition here is that the reader may be thinking that the word of God had failed, because of the Jews rejection of Jesus. But this does not fit if Paul has just broken his mood and praised God for sending the Messiah. Had Paul just concluded a doxology to the Father, one would expect Paul to begin the next verse with “Because” or “For” (Greek gar). We would expect him to continue with the thought that God keeps his promises (by sending Messiah) and because of this, His word has not failed, for others have come to saving faith outside of Israel. On the other hand, if Paul has just proven how very grave the Jews rejection of Jesus was, because not only have they rejected the promised Messiah, but also God Himself, then it makes perfect sense that Paul would begin his next sentence as he does: “But not that…” (Greek: ouch hoion de…).
Whether Metzger or Harris should have interacted more with Abbot’s arguments depends largely on how compelling one finds Abbot’s arguments. Metzger apparently gave them little weight as Mr. Stafford is correct – he does not specifically interact with his arguments. But the same cannot be said of Harris. Metzger and Harris actually spend a great deal of time developing their contextual arguments (Metzger, Punctuation, pp. 103 – 112; Harris, pp. 154 – 165). Both offer detailed exegesis, substantial support for their views from relevant literature, and respond to the major objections. In the case of Harris, these include Abbot’s (e.g., Harris, p. 158, 162, 163, 165).
Mr. Stafford’s concluding quote from Dwight establishes the point that he did not regard a doxology to the Father being impossible in this context. Such is my view as well. However, immediately after Mr. Stafford’s quote, Dwight goes on to say:
But, while we admit this, we must observe that the progress of the author’s thought is towards the sixth verse and what follows it, and that the balance of probability cannot be determined without considering the five verses in connection with the sixth and the rest of the chapter. As we look at the matter from this point of view, we find that the thought moves on in an easy and natural way, if we make the reference of these words, which are under discussion, to be Christ (Dwight, p. 41).
And this, too, echoes my thoughts. We have here a verse with an ambiguous construction in the Greek. Neither view is impossible from the standpoint of grammar alone – but one is more likely when all other considerations are taken into account. I believe that view is clearly the one I have advocated. However, ultimately, it is God who reveals the truth of who His Son is, not carefully crafted arguments and endless scholarly quotations. If you are inclined to Mr. Stafford’s view, but find the arguments here presented troubling, perhaps God is working in your heart, even now, as you read these words. I invite you to pray that God will show you who His Son truly is, and to seek Him in the pages of God’s Holy Word.
In conclusion, I will follow Mr. Stafford and quote the words of Timothy Dwight:
It is not vital to the doctrine of the Divinity of Christ to find the declaration that he is God in this verse. The Apostle Paul may have believed that his Lord and Savior was Divine, and may teach this in his Epistles; and yet may have chosen to limit himself in the use of the name God, so far as to apply it to the Father only…. If, however, this verse does contain the apostolic testimony that Christ is God, it is a direct affirmation of what the opposite doctrine would deny, and excludes that doctrine altogether (Dwight, pp. 53-54).
Soli Deo Gloria
Woodland Hills, 2003
1. Of the 13 examples of ho ôn in the GNT and LXX, only two (John 3:31; 8:47) begin a new phrase. In each case, John has constructed his sentences in such a way that it is impossible to construe ho ôn asmodifying a preceding head noun. It has been argued that when ho ôn is used to introduce a relative clause, the noun it modifies immediately precedes. In Romans 9:5, of course, the phrase to kata sarka is between the head noun Christos and ho ôn. However, in two cases (John 6:46; 2 Corinthians 11:31), this ‘rule’ does not pertain. While this is an admittedly small sample of data, and it is impossible to draw absolute conclusions, we may nevertheless say that the Biblical authors were aware that ho ôn could be construed as a qualifying phrase, or introducing a phrase in apposition to a preceding head noun, and so appear to have taken care when using it to start a new sentence that it could not be so understood.
2. The lone exception is Psalm 67:19. But as Dwight argues at length, this verse is really not a proper parallel to Romans 9:5 in that it differs from ordinary doxologies by doubling eulogêtos (Dwight, pp 32-33). Ezra Abbot, one of the most articulate proponents of “God over all” being a doxology to the Father, agrees: “I do not urge it as a parallel to Rom. ix. 5” (Abbot, Romans 9:5, p. 107).
3. Ambrosias actually seems quite clear that he understands “God” to be attributed to Christ: “As there is no mention of the Father’s name in this verse and Paul is talking about Christ, it cannot be disputed that he is called God here…If someone does not think that it is said about Christ that he is God, then let him name the person about whom he thinks it is said, for there is no mention of God the Father in this verse (Commentary on Paul’s Epistles, in ACC: Romans, p. 247).
4. Abbot says Chrysostom and Theodoret are to be distinguished from the other fathers, “for sobriety and good sense in interpretation” (Abbot, Romans 9:5, p. 140). Nevertheless, he notes that they both, “adopted that excessively unnatural if not impossible construction of 2 Cor iv. 4” (Ibid.). But Abbot has made a hasty generalization. Simply because they may have adopted an allegedly “impossible” construction of one verse does not prove that they have done so in Romans 9:5.
5. Metzger himself characterizes the punctuation in early Greek manuscripts as “quite erratic” and provides a number of examples from the very manuscripts under consideration (Metzger, Punctuation, p. 99). Abbot puts it succinctly: “The truth is, that this whole matter of punctuation in the ancient MSS. is of exceedingly small importance” (Abbot, Romans 9:5, p. 152).
6. Mr. Stafford’s assertion that kata sarka does not imply an antithesis in Romans 9:3 overlooks the fact that out of almost 130 uses of “brother” (adelphos) in Paul’s writings, in every case except Romans 9:3, it means either a spiritual brother (i.e., a fellow Christian) or a literal brother (“James, the brother of the Lord”). But Paul does not consider the Jews his spiritual brothers; rather he qualifies the term to mean: “kinsmen according to the flesh.” Thus, the implied antithesis is between Paul’s “brothers” in the Lord and Paul’s “brothers” as Jews. This contrast is so apparent that it is a virtual commonplace among commentators and other scholars writing about this verse (e.g., Gill (Commentary), Robertson (Word Pictures), Barnes (Notes), Moo (Romans), Godet (Romans), Moule (Romans), Hodge (Romans), Stauffer (TDNT 3:105), Phillipi (Romans), The Expositor’s Bible Commentary, and Harris (Jesus, p. 156). In verse 5, Paul is not speaking about physical versus spiritual brotherhood, but physical versus spiritual descent (ex ôn).
8. One answer to this question often offered by Jehovah’s Witness apologists, including Mr. Stafford himself (Stafford, p. 339), is that theos is a count noun and count nouns (because “countable”) can only be definite or indefinite – not qualitative. The definition of a “count noun” preferred by these apologists is a contextual one – that is, if a noun is “countable” in a given context, it is a count noun. If it is not countable, it is a mass noun. Based on this definition, it is begging the question to suggest that theos in Romans 9:5 cannot be a qualitative noun because it is a count noun. The Witnesses must first establish that theos is a count noun in this context. If theos is here an appositional predication of Deity to Christ, it is not countable.Thus, any arguments based on a contextual definition of mass/count terms are of little value in determining the semantic force of theos in a specific context.
9. See also Deuteronmy 32: 17 – 21, in which YHWH calls the “demons” (LXX: daimoniois) “not-God” (JPS). The meaning here is not merely that the demons were not YHWH (the person), but that they were not theos – not God by nature.
10. It is not my intention to here engage Mr. Stafford’s argument on this point in detail, but I will say that the idea that a word may contain more than one semantic force (i.e., “meaning”) in a given context (unless the author intends ambiguity) is a lexical principle that requires proof beyond assertion. It seems counter to the way lexical semantics actually works (that is, that we use words to mean only one of their possible denotations in any given context), and is actually an example of what D.A. Carson has called the exegetical fallacy of “unwarranted adoption of an expanded semantic field” (Fallacies, p. 60-61). After noting that a word outside of a context actually “does not have a meaning” but rather various potential meanings, Louw says:
“When used in a context, the situation and the syntactic environment contribute to the choice between the several possibilities of meaning. The word has a specific meaning in that context” (Louw, Semantics, p. 40).
Cotterell and Turner concur: “The context of the utterance usually singles out (and perhaps modulates) the one sense, which is intended, from amongst the various senses which the word is potentially capable” (Cotterell, p. 175, emphasis in original). Silva quotes Vendryes: “Among the divers meanings a word possesses, the only one that will emerge into consciousness is the one determined by context (Vendryes, in Silva, p. 139) and says this principle is “one of the few universally accepted hermeneutical guidelines” (Ibid., p. 138).
11. Arguments based on statistics can only prove probabilities, not actualities. They do not allow for exceptional cases. For example, the fact that no other human beings have raised themselves from the dead does not disprove that Jesus did so.
12. Harris quotes Phillipi as follows: “In the interpretation of a formula that has become fixed, empiricism is altogether in its right place, and still more where, for the established usage, a sufficient ratio can be alleged” (Phillipi in Harris, p. 162).
13. Harris’ footnote to this point is as follows: “Dwight (Romans, p. 36-37) cites several LXX passages where an inversion might be expected on this principle, but is not found (e.g., eulogêtos in 1 Sam. 25:33 and 2 Macc. 15:34)” (Ibid., p. 163 n61).