Grace Theological Journal 6.1 (1985) 91-112.

            [Copyright © 1985 Grace Theological Seminary; cited with permission;

                            digitally prepared for use at Gordon College]



                THE SEMANTICS AND



                    CONSTRUCTION IN

                THE NEW TESTAMENT



                                    DANIEL B. WALLACE



            A survey of the grammatical terminology, identification, and

semantics of the object-complement construction in the Greek NT

demonstrates that the treatment of this construction in the ma:jor

grammars is inadequate. A rather extensive listing of NT examples

of this construction supports the thesis that the object complement

construction is semantically equivalent to the subject-predicate nom-

inative construction. Thus, any principles which apply to subject-

predicate nominative constructions (e.g., "Colwell's Rule') are equally

applicable to object-complement constructions.


                                                *     *     *




Although some would insist that grammar is one of those elemen-

tary things which is better left behind as we press on to maturity,

there are still a few die-hards who feel that not all has been said on

the topic. Lars Rydbeck, for example, recently asked the question,

"What happened to New Testament Greek grammar after Albert

Debrunner?"1 H is answer is that it "has come almost to a standstill,

one of the reasons being that "There is a prevalent but false assump-

tion that everything in NT Greek scholarship has been done already.”2

Rydbeck goes on to suggest that one major area in NT grammar

which has yet to be resolved is the nature of NT Greek.3 This, indeed,


            l The title of a paper presented to the Fifth International Congress on Biblical

Studies (Oxford: September. 1973), published in NTS  21 (1974-75) 424-27.

            2 Ibid., 424.

            3 Ibid., 425.


92                    GRACE THEOLOGICAL JOURNAL

is a critical issue; but there are others. Among them is the relation of

structure to semantics. This is a problem area because most grammars

are satisfied with presenting the structural phenomena of the NT in a

descriptive manner (i.e., a mere tagging of structures as belonging to

certain syntactical categories), while hardly raising the question of the

differences in the fields of meaning that 'synonymous' structures can

possess.4 One construction which can be profitably put through the

structure-semantics grid is that of the object-complement double



                                    DEFINITION OF TERMS

            Not all are agreed on which terms to use when describing this

grammatical phenomenon. Thus it is appropriate to begin by defining



Double Accusative

            The nomenclature "double accusative" is customarily used in

grammars to refer to two different kinds of constructions:5 (1) a

person-thing double accusative (in which a verb takes two direct

objects in the accusative, one being the person affected, the other

being the thing effected);6 and (2) an object-complement double


            4 Some specific areas of inquiry with reference to this problem are: the genitive of

possession vs. the dative of possession; the simple infinitive vs. the genitive articular

infinitive (or ei]j/pro<j plus the accusative articular infinitive) to express purpose; the

overlap in the use of simple cases and prepositions plus cases (e.g., simple dative vs. e]n

plus the dative); the anarthrous generic noun vs. the articular generic noun; the various

structures used to express result, causality, etc. To be sure, some of these topics are

discussed in the grammars, but as of yet, grammars by and large make no attempt to

be systematic in dealing with the differences in the fields of meaning that 'synonymous'

structures can have.

            5 There are other double accusative constructions as well, but which occur so

infrequently as to call for little attention in the grammars. Besides the person-thing and

object-complement constructions, BDF list the "accusative of object and cognate

accusative" and "accusative of object and of result" (86-87).

            6 Cf. BDF, 85; A. T. Robertson, A Grammar of the Greek New Testament in the

Light of Historical Research (4th ed.; Nashville: Broadman, 1934) 482-84; G. B. Winer,

A Treatise on the Grammar of New Testament Greek, translated and revised by W. F.

Moulton (3d ed., revised; Edinburgh: T. &T. Clark, 1882) 284-85; H. W. Smyth, Greek

Grammar, revised by G, M. Messing (Cambridge, Mass.: Harvard University, 1956)

363-64. Others call this construction "an Accusative of the remoter object as well as of

the immediate object" (C. F. D. Moule, An Idiom-Book of New Testament Greek

[2d ed.; Cambridge: Cambridge University, 1959] 33), or a double accusative of "per-

sonal and impersonal object" (H. E. Dana and J. R. Mantey, A Manual Grammar of

the Greek New Testament [Toronto: Macmillan, 1927] 94).




accusative. This paper is concerned only with this second type of




            An object-complement double accusative is a construction in

which one accusative is the direct object of the verb and the other

accusative (either noun, adjective or participle) complements the object

in that it predicates something about it.7 This construction is called a

double accusative of object and predicate accusative by Robertson,

Blass-Debrunner, Turner, Smyth, Mayser, Kuhner, Jannaris, and

others.8 It is described as "an accusative of the object affected and an

object complement" by Funk,9 "accusative of subject [!] and predi-

cate" by Winer,10 and "A direct and predicate object" by Dana and

Mantey.11 Others describe the construction in still different terms,12

even as I have done. I use the name "object-complement" because it is

brief and to the point.13


            7 Another way of defining this construction which perhaps is technically more

correct is that given by Goodwin and Gulick: "A verb and an accusative depending on

it may together be treated as a single word having another accusative as its object"

(W. W. Goodwin, Greek Grammar, revised by C. B. Gulick [Boston: Ginn & Co.,

1930] 227).

            8 Robertson, Grammar, 480; BDF, 86; J. H. Moulton, A Grammar of New Testa-

ment Greek, vol. 3: Syntax, by N. Turner (Edinburgh: T. & T. Clark, 1963) 246; Smyth,

Greek Grammar, 362; E. Mayser, Grammatik der griechischen Papyri aus der Ptole-

maerzeit, vol. 2, part 2: Satzlehre (Berlin: Walter de Gruyter, 1934) 320 ("Ein Akkusativ

des Objekts und ein Pradikatsakkusativ"); R. Kuhner, Grammar of the Greek Language

(Boston: B. B. Mussey, 1849) 398; A. N. Jannaris, An Historical Greek Grammar

(London: Macmillan, 1897) 332; H. P. V. Nunn, A Short Syntax of New Testament

Greek (Cambridge: Cambridge University, 1943) 41; Goodwin and Gulick, Greek

Grammar, 228; C. Vaughan and V. E. Gideon, A Greek Grammar of the New Testa-

ment (Nashville: Broadman, 1979) 66.

            9 R. W. Funk, A Beginning-Intermediate Grammar of Hellenistic Greek (2 vols.;

2d, corrected ed; Missoula: Scholars, 1973) 2:725.

            10 Winer, Treatise, 285.

            11 Dana and Mantey, Manual Grammar, 94.

            12 J. A. Brooks and C. L. Winbery (Syntax of New Testament Greek [Washington,

D. C.: University Press of America, 1979] 47) describe the construction as involving "a

direct or primary object and a predicate or secondary object"; William Webster (The

Syntax and Synonyms of the Greek Testament [London: Rivingtons, 1864] 64) states

that "The second accusative often appears as a tertiary predicate or an apposition";

Moule (Idiom-Book, 35) comes close to the 'normal' description when he speaks of

"The Accusative used Predicatively, i.e. to "predicate" something of a noun already in

the Accusative."

13 It should be observed that those grammars which do speak of the "object com-

plement" mean by this the second accusative only, i.e., the predicate accusative. By the

use of the hyphen in "object-complement," I am indicating both accusatives (hence, the

whole construction)--the object and its complement.


94                    GRACE THEOLOGICAL JOURNAL



            There are three issues I wish to discuss, namely, (1) the identifi-

cation of the construction (i.e., how does one know when he has an

object-complement construction?), (2) the identification of the com-

ponents (i.e., how can one tell which is object and which is comple-

ment?), and (3) the semantics of the construction (i.e., in addition to

the obvious fact that predication is involved, what else can the

construction indicate?).


Identification of the Construction

The problem in identifying the construction is due primarily to

the fact that every verb which can take an object-complement con-

struction is not required to do so.14 Consequently, not all would make

a positive identification of the construction in a given instance.15 For

example, Phil 3:18 reads, ou{j polla<kij e@legon u[mi?n, nu?n de> kai>

klai<wn le<gw [,] tou>j e]xqrou>j tou? staurou?.  It is possible to take tou>j

e]xqrou>j as an appositive to ou{j (thus, "whom often I used to mention

to you, and now weeping I say, [they are] the enemies of the

cross. ..").16 But a second possibility is to consider le<gw as having

the meaning 'I call' here and to treat tou>j e]xqrou>j as the complement

to an implied pronominal object (thus, ". . . but now, weeping, I call

[them] the enemies of the cross. . .").17 There are not many question-

able constructions such as this, but there are a few that are exegetically



Identification of the Components

The problem in identifying the components is that occasionally

the natural order of object, then complement, is reversed. In most of


14 E. V. N. Goetchius, The Language of the New Testament (New York: Charles

Scribner's Sons, 1965) 141. It is to be noted, however, that some verbs regularly or

almost exclusively take object-complements (e.g., h[ge<omai, o]noma<zw, and fa<skw).

15 No grammar gives an exhaustive list of object-complements in the NT. Conse-

quently such lists cannot be compared to discover the questionable instances. But by

comparing translations and by attempting to reconstruct the semantic range of every

possible object-complement construction (i.e., to see whether the construction in ques-

tion must be or might be an object-complement), the definite and the questionable

instances can be determined.

16 In support of this view, cf. Winer, Treatise, 665; Robertson, Grammar, 413;

M. R. Vincent, A Critical and Exegetical Commentary on the Epistle to the Philippians

and to Philemon (ICC; Edinburgh: T. & T. Clark, 1897) 117.

17 In support of this view, cf. H. A. A. Kennedy, "The Epistle to the Philippians" in

vol. 3 of The Expositor's Greek Testament, ed. W. R. Nicoll (New York: Dodd,

Mead & Co., 1897) 461.




the instances it is obvious which is object and which is complement.

For example, Phil 3:1 7 reads e@xete tu<pon h[ma?j. A very literal trans-

lation would not render this, "you have a pattern/example in us," for

that would require e@xete tu<pon h[mi?n. Rather, it should be rendered,

"you have us as [a] pattern/example." It is obvious, then, that this is

an object-complement construction and that the order has been

reversed. Such a clear instance demonstrates the reversal phenomenon

and, at the same time, raises two questions: (1) What are the criteria

for determining which is which since word order is not an infallible

guide? and (2) Why is the order sometimes reversed?


The Semantics of the Construction

The third issue involves the semantics of the construction. As

mentioned earlier, by definition an object-complement construction is

a construction in which a predication is made. But beyond this given,

what else can the construction indicate? Specifically, what is the dif-

ference semantically between the order of object, then complement

and complement, then object? For example, is it possible that when

Paul wrote e@xete tu<pon h[ma?j in Phil 3:17 he did not mean exactly the

same thing as e@xete h[ma?j tu<pon?



Concerning the identification of the construction, the standard

grammars make almost no advances beyond defining the construc-

tion18 and giving an abridged list of the kinds of verbs which take

object-complements. Some of the grammars do point out that the

complement is often preceded by ei]j or w[j.19 Unfortunately, not only

is there a very high percentage of cases where ei]j and w[j are absent,

but even when either one is present, there is not, ipso facto, an object-

complement construction.20 With reference to the identification of the

components, only one of the more than thirty grammars examined

explicitly addressed the question of order in an object-complement


18 Goetchius (Language, 141) is a lone exception to the silence of the grammarians:

"Object complements occur only with certain verbs (all of which also occur with

'ordinary' direct objects, i.e., without object complements), e.g., call, make, find, think,

deem, choose, elect. Some of these verbs also occur with indirect objects (e.g., call,

make, find), so that it may not always be immediately apparent whether sentences

containing them are structurally similar to (3) ['The child gave the dog a bone'] or to

(4) ['The general called the captain a fool']; usually, however, the meanings of the

nouns N2 and N3 [in the construction N1-V-N2-N3] are compatible with only one

interpretation (and, hence, with only one structural analysis) of a sentence."

19 Robertson, Grammar, 480-81; BDF, 86-87; Turner, Syntax, 246-47.

20 Cf., for example, Matt 9:38; 22:13; Mark 1:12; Rom 6:22; 2 John 10.


96                    GRACE THEOLOGICAL JOURNAL


construction.21  Some grammars did, however, deal with the issue of

order implicitly, giving some guidelines which will be discussed below.

Concerning the semantics of the construction, apart from the fact

that the complement is making an assertion about the object, again

only one grammar gave any explicit guidelines.22 But not one ad-

dressed the question of the difference in force between the normal

order and the reversed order.




Method of Research Used

In order to come to any sound conclusions, it was necessary to

be as exhaustive in the inductive process as possible. By means of the

grammars, thesaurus, concordance, and lexicon, I discovered more

than fifty verbs which take object-complements23 and more than three

hundred object-complement constructions in the NT. The raw data

gathered is at least enough to provide guidelines which may help to

inform and possibly resolve the three issues.


Solution Proper

In dealing with each of the three issues, some guidelines or prin-

ciples that have been derived from the study are first set forth, and

then some of the exegetically significant passages affected by this

study are briefly discussed.


21 Goetchius (Language, 142) again was the lone exception, stating, "The constitu-

ents of these Greek sentences may, as we might expect, occur in any order; both the

direct object and the object complement are in the accusative case, but the direct object

is always more 'definite' than the object complement."

22 Goetchius, Language, 142.

23 Included in the list of verbs are the following: a[gia<zw, a@gw, ai]te<w, a]natre<fw,

a]podei<knumi, a]polu<w, a]poste<llw, geu<omai, ginw<skw, de<xomai, di<dwmi, doke<w, e]gei<rw,

ei#don, ei#pon, e]kba<llw, e]kle<gw, e]ndeiknumi, e]pikale<w, eu[ri<skw, e@xw, h[ge<omai, qe<lw,

qewre<w, i[kano<w, i!sthmi, kaqi<sthmi, kale<w, khru<ssw, kri<nw, lamba<nw, le<gw, logi<-

zomai, nomi<zw (in spite of the protests by BDF [86] and Robertson [Grammar, 480]

that nomi<zw does not take an object-complement in the NT, there are two unmistakable

instances [cf. I Cor 7:26nomi<zw ou#n tou?to kalo>n u[pa<rxein and I Tim 6:5—nomizo<n-

twn porismo>n ei#nai th>n eu]se<beian]), oi#da, o[mologe<w, o]noma<zw, paralamba<nw, par-

e<xw, pari<sthmi, pei<qw, peria<gw, pisteu<w, poie<w, proori<zw, prosfe<rw, proti<qhmi,

proxeiri<zw, suni<hmi, suni<sthmi (sunista<nw), ti<qhmi, u[pokri<nomai, u[ponoe<w,

u[yo<w, fa<skw, xrhmati<zw. In addition to these are three questionable verbs—katakli<nw,

katanoe<w, and o]ri<zw. As well, the NT uses e]pile<w, e]ponoma<zw, and prosagoreu<w in

the passive which, in the active, would take object-complements (in the passive, the

object is converted to the subject and the complement to the predicate nominative/




Identification of the Construction

General Principles. With reference to the identification of the

construction, I have counted about thirty questionable instances. The

most common of these involved an infinitive as the complement.24

The question here is whether the infinitive is functioning substan-

tivally as the complement to the direct object or in some other

capacity.25 But however the infinitive is tagged, the meaning of the

total construction is not altered. A second group of instances was

debatable because the alleged complement could possibly be a simple

appositive to the direct object.26 Other constructions were questionable

because of the relation of the adjective to the direct object,27 the

function of ei]j before a second accusative,28 the ambiguity of the case

of the second noun,29 etc.30

Since there was a positive identification of more than 90% of all

possible object-complements examined,31 and since the questionable

instances fell into very specific structural categories, certain principles

for determining the identification of the construction become evident.

First, what must be established is that the verb related to the con-

struction in question can, indeed, take an object-complement. In the

case of hapax legomena and other rare verbs, appeal can certainly be

made to extra-NT Greek literature for verification.32 Second, the


24 Cf. Rom 1:13; 11:25; I Cor 10:1; 12:1; 2 Cor 9:5; Phil 2:25; 3:13; I Thess 4:13;

I Tim 2:4.

25 In particular, as a complementary infinitive to the verb.

26 Cf. Matt 27:32 (here a@nqrwpon Kurhnai?on might be a Semitic periphrastic

construction [cf. Matt 11:19] in which wyx is left untranslated when followed by

an appositional substantive. The idiom, however, is also found in Greek. Cf. W. E.

Jelf, A Grammar of the Greek Language Chiefly from the German of Raphael Kuhner

[2d ed.; 2 vols.; Oxford: James Wright, 1851] 1:102; and Demosthenes I. I, 2, 4, 6,

8,9, 10; 9. 19,23,25,36, etc.); Acts 13:6,23; Rom 10:9; Phil 3:18; Col 2:6, I Pet 3:15;

Rev 13:17.

27 Cf. Acts 6:13; 24:20 (interrogative pronoun); Titus 2:10.

28 Cf. Eph 1:5.

29 Cf. Heb 4:7 and Rev 9:11 (here, of course,  ]Apollu<wn is nominative in form, but

the author may possibly be treating it as an indeclinable noun functioning as an


30 Other constructions were debatable because the adjective could be substantival

and the pronominal adjective related to it could be modifying it (John 2:11; 4:54), or

the verb was not found with any clear object-complements (I Pet 3:15), or a]na< was

wedged between w[sei< and the second accusative (Luke 9:14).

31 When the instances involving infinitives are discounted, the positive identifica-

tion is closer to 95%.

32 E.g., a[gia<zw seems to take an object-complement construction in I Pet 3:15

(though there are some dissenters among the translations), but no other clear NT

examples can be found (though I Thess 5:23 comes close). However, in the LXX there


98                    GRACE THEOLOGICAL JOURNAL

specifics of the structure in question must have parallels in positively

identified object-complements. Thus, for example, if the possibility

that e]sthsa<n te ma<rturaj yeudei?j is an object-complement in

Acts 6:1333 is even to be entertained, clear instances of an anarthrous

object with a predicate adjective must be found. Finally, once these

first two points are established in a given text, I believe that, barring

contradictory contextual factors,34 the antecedent probability is that

the construction in question is indeed an object-complement.35

Exegetically Significant Texts. From my count, there are at

least eight exegetically significant passages which are affected by the

issue of the identification of the construction.36 Four of these pas-

sages are affected by the other two issues as well,37 and consequently

will not be discussed here. Of the remaining four, two passages,

Phil 3:18 and Titus 2:10, warrant a brief treatment at the present



are two examples (Exod 29:1 and 30:30) in which an infinitive probably functions as a

complement as well as one example (Isa 8:13, the text which lies behind I Pet 3:15) in

which a[gia<zw clearly takes an object-complement.

33 The difference exegetically between taking yeudei?j predicatively and attributively

is that a predicative yeudei?j makes more explicit the intention of Stephen's enemies to

produce false witnesses (thus, "and they brought forth witnesses [to be] false").

34 An illustration of possibly contradictory contextual factors is found in Acts

13:23—o[ qeo>j . . . h@gagen t&?  ]Israh>l swth?ra  ]Ihsou?n. If the construction is taken as

an object-complement ("God has brought to Israel Jesus [as] Savior") rather than

simple apposition ("God has brought to Israel a savior, [namely] Jesus"), one is faced

with the difficulty that Jesus is introduced in the message as though the residents of

Pisidian Antioch were already familiar with his name.

35 This antecedent probability varies in certainty directly in proportion to how well

the first two principles are established in a given instance. If they are established at all,

tagging the construction as object-complement must at least be given serious con-


36 John 2:11; 4:54; Acts 13:23; Rom 10:9; Phil 3: 18; Col 2:6; Titus 2: 10; I Pet 3:15.

37 Acts 13:23 (for a brief discussion, see n. 34 above); Rom 10:9; Col 2:6; I Pet 3:15.

38 The two remaining constructions are found in John 2:11 and 4:54. John 2: 11

reads, Tau<thn e]poi<hsen a]rxh>n tw?n shmei<wn o[  ]Ihsou?j. The ASV, RSV, NASB and

NIV all take e]poi<hsen here in the sense of 'he did,' with the RSV and NIV treating

a]rxh>n as an appositive to Tau<thn and the ASV and NASB regarding Tau<thn as

modifying a]rxh>n. However, if e]poi<hsen has the sense of 'he made' here, then the

construction is an object-complement (thus, "Jesus made this [to be] [the] first of his

signs"). The object-complement construction makes more explicit the idea of design on

the part of Jesus while the other reconstruction of the text only speaks of his power.

John 4:54 reads, tou?to de> pa<lin deu<teron shmei?on e]poi<hsen o[  ]Ihsou?j. Here again the

translations all treat e]poi<hsen as 'he did.' Although they all seem to recognize the

construction to be an object-complement, they weaken its force by treating e]poi<hsen

as though it belonged in a relative clause (almost as though they were translating tou?to

de> pa<lin h#n deu<teron shmei?on o{ e]poi<hsen). But if e]poi<hsen has the force of 'he made'

(thus, "Now again, Jesus made this [to be] [the] second sign"), then not only is there

design in the selection of miracles recorded (cf. John 20:30-31), but also in the sequence

and performance of them as well.


In Phil 3:18 Paul says, polloi> ga>r peripatou?soin ou{j polla<kij

e@legon u[mi?n, nu?n de> kai> klai<wn le<gw [,] tou>j e]xqrou>j tou? staurou?

tou? Xristou?. If tou>j e]xqrou>j is in apposition to ou{j, then there

appears to be a change in description, but not a change in status, of

the object. One of the problems with this view, however, is the func-

tion of ga<r. Unless it is equivalent to de<, the polloi> of v 18 apparently

belong to the same camp as "those who are thus walking" (tou>j ou!tw

peripatou?ntaj) in the previous verse. However, if le<gw has the sense

of 'I call,' and if nu?n de> has a contrastive force rather than a con-

tinuative force, then there is an object-complement construction here.

If so, it becomes apparent that there is a shift in status from the ou{j

to the tou>j e]xqrou>j (thus, "For many are walking, about whom often

I used to speak to you, but now, even weeping, I call [them] the

enemies of the cross of Christ.")39 Obviously the interpretation of this

text cannot be solved on the basis of grammar alone, but the fact that

an object-complement construction is at least possible here gives

some breathing room to the exegete in this thorny passage.

In Titus 2:9-10 Paul commands Titus to exhort Christian slaves

to be obedient to their earthly masters. In v 10 he describes both a

negative and a positive aspect of what their conduct is to be. The

positive aspect is described in the participial clause pa?san pi<stin

e]ndeiknume<nouj a]gaqh<n. Although most would understand a]gaqh<n as

an attributive adjective modifying pi<stin (thus, "showing forth all

good faith"), it is possible that a]gaqh<n is a predicate adjective, func-

tioning as the complement to pi<stin (thus, "showing forth all faith [to

be] good"). Grammatically and exegetically this may be valid, though

the grammarians and exegetes do not mention the possibility.

            Although there are other grammatical arguments in favor of a

predicate a]gaqh<n,40 the concern here is only with those which are


39 It should be mentioned that there are several clear examples of the omission of a

pronominal object in an object-complement construction (thus paralleling the con-

struction here). Cf. Matt 23:9; John 6:15; Rom 1:22; 2 Cor 11:2; Phil 3:8; I Thess 2:13;

3:15; Heb 11:11; 2 Pet 1:8; 2 John 4.

40 ln particular, the relation of adjective to noun in anarthrous constructions could

be cited in favor of a predicate a]gaqh<n here. In cursory form, the evidence derived

from such a consideration is as follows. In non-equative clauses and phrases I have

discovered over forty completely attributive relations in adjective-noun-adjective con-

structions in the NT (e.g., Matt 7:17; 23:35; Eph 1:3; Rev 18:2). However, none of the

constructions involving pa?j and only one other attributive construction had an inter-

vening word between the noun and second adjective (cf. Rom 1:11). Also, seven of the

pa?j constructions were in prepositional phrases, a situation which does not parallel

Titus 2:10 (e.g., Col 1:10; 2 Tim 3:17; Titus 3:1).

I also discovered thirteen instances in which one adjective was attributive and one

was predicate in non-equative clauses/phrases (e.g., Matt 5:36; John 7:23; Col l:28

[here with pa?j and, interestingly enough, an object-complement construction]). In four

instances the second adjective was separated from the noun by an intervening word or



directly relevant to object-complements. By applying the three maxims

related to the identification of an object-complement construction, at

least the possibility of an object-complement construction here can be


First, e]ndei<knumi does indeed take an object-complement else-

where in the NT.41 Second, there are other instances of object-comple-

ments which involve an anarthrous object and a predicate adjective,42

as well as scores of passages which exhibit the more general parallel

of a predicate relation in an anarthrous noun-adjective construction.43

Third, other exegetical considerations do allow for this possibility,44

and there are apparently not any contextual factors which exclude it


phrase (cf. Mark 7:2; 8:19; Acts 4:16; Rev 15:1). John 10:32 also has an intervening

verb between the noun and adjective (polla> e@rga e@deica u[mi?n kala<), but there is

ambiguity as to the function of the second adjective.

Therefore, although the attributive constructions outnumbered the constructions

in which the second adjective was predicate three to one, the second type of construc-

tion commonly had an intervening word between noun and second adjective. Further-

more, none of the definitely attributive relations with pa?j; in the first attributive

position had an intervening word between the noun and second adjective. Thus,

although the construction in Titus 2:10 is similar to wholly attributive constructions in

that it has pa?j before the noun (but cf. Col 1:28 for an example in which the pa?j

preceding the noun is attributive and the adjective following is predicate), it is similar

to part attributive/ part predicate constructions in that there is an intervening word

between the noun and second adjective. There is, then, a good possibility (might one

even say, an antecedent probability?) grammatically that a]gaqh<n is a predicate adjec-

tive in Titus 2:10.

For more information on the whole area of the relation of adjective to noun in

anarthrous constructions, see D. B. Wallace, "The Relation of Adjective to Noun in

Anarthrous Constructions in the New Testament" (unpublished Th.M. thesis; Dallas

Theological Seminary: May, 1979) and the article by the same title (which is derived

from the thesis) in NovT 26 (]984) 128-67.

41Cf. Rom 2:15. As well, at least one of the cognate verbs also takes an object-

complement (a]podei<knumi, 1 Cor 4:9). Furthermore, 2 Macc 9:8 has a precise parallel

to Titus 2:10 (fanera>n tou?  qeou? pa?sin th>n du<namin e]ndeiknu<menoj). This is obviously an

object-complement construction because the adjective fanera>n is outside of the article-

noun group th>n du<namin.

42 Cf. Luke 3:8; John 9:]; Acts 10:28; Col l:28.

43 See Wallace, "The Relation of Adjective to Noun"(thesis), Appendix II:73-102

in which almost 400 such constructions are charted. For the more precise parallel, cf.

n. 40 above.

44 The main question exegetically has to do with the meaning of pi<stij. This noun

seems to be used in the pastoral epistles frequently as a technical term for the Christian

religion (cf. I Tim 1:2; 3:9; 4: 1, 6; 2 Tim 2:18; 3:8; Titus 1:13; 3:15). In two of the three

occasions in which pi<stij is modified by an adjective (in Titus 1:4 koinh<n modifies

pi<stin, suggesting more about the scope of this faith than about its character), the

adjective used is a]nupokri<toj (cf. I Tim 1:5; 2 Tim 1:5). The author seems concerned

that one's faith be a sincere faith. An insincere faith is apparently not genuine (cf.

I Tim 1:19; 4:1; 5:8; 6:21; 2 Tim 3:8), but a sincere faith is closely associated with holy


from consideration. Consequently, the antecent probability is that

Titus 2:10 does contain an object-complement construction. If it does

then the sense of Titus 2:9-10 could be expressed in the following

loose translation: "Slaves should be wholly subject to their masters

. . . demonstrating that all [genuine ]45 faith is productive, with the

result46 that they will completely adorn the doctrine of God.”47

Again, grammar does not solve all of the exegetical problems by

any means, but if the principles for identifying object-complement

constructions have any validity at all, then one must at least deal

seriously with the possibility of such a construction in Titus 2:10,

even though such a possibility apparently has hitherto gone unnoticed.


Identification of the Components

General Principles. With reference to the identification of the

components of an object-complement construction, it has already

been pointed out that word order is not an infallible guide. Therefore,

some other criteria must be used to supplement if not supplant the

principle of word order.

On the basis of several strands of evidence, I believe the follow-

ing overall thesis for solving the problem of the identification of the

components can be stated: the object-complement construction is

semantically equivalent to the subject-predicate nominative construc-

tion. This thesis is the major point of this article. Therefore, any

principles which help to resolve the identification of the components

in a subject-predicate nominative construction are equally applicable

to the object-complement construction. Two points must be estab-

lished in order to validate this thesis. First, it needs to be established

that there is analogy between the two types of constructions. And


behavior (cf. 2 Tim. 3:15-17; Titus 1:13-16--the author links faith with holy behavior

outside the pastorals as well (cf. Eph 2:8-10; Coll:4, 6, 10]).

Thus if a more technical sense for pi<stij is understood in Titus 2:10 (J. W. Roberts

["Every Scripture Inspired by God," Restoration Quarterly 5 (1961) 35] apparently

leans toward a more technical sense for pa?j here, for he writes, ". . . the context shows

that the word pas means 'perfect' or 'complete' faith"), the author maybe instructing

Titus to exhort slaves to demonstrate that their faith is sincere and that it results in

holy behavior.

45 ‘Genuine' may either be implied from the flow of argument or may be considered

as part of the field of meaning for pa?j when it is used with abstract nouns (cf. BAGD

on pa?j I. a. 0.). .

46   !Ina here is taken as having an ecbatic force.

47 A further argument to help validate this sense is the possibility of a synthetic

parallel between the two halves of v 10 which is evident only when a]gaqh<n is taken as a

predicate adjective. Thus, to demonstrate that genuine faith is productive is to adorn

the doctrine of God.



second, the thesis needs to be tested on specific object-complement


The following lines of evidence establish, I believe, that the

object-complement construction is semantically equivalent to ;the

subject-predicate nominative construction. (1) By definition, both the

complement and the predicate nominative make an assertion about

another noun in the same case. (2) The terms used to describe the

object-complement construction in most grammars strongly suggest

such semantic equivalence. As the reader will recall, it was mentioned

earlier that many of the major grammars call this construction an

object and predicate accusative construction.48 And Winer goes so far

as to call the construction an "accusative of subject and predicate

[italics mine]."49 (3) The infinitive of the copula occasionally occurs

in an object-complement construction, linking this construction to the

subject-predicate nominative construction semantically.50 (4) Many

of the verbs which take an object-complement also take a declara-

tive/recitative o!ti clause (and even, occasionally, some other use of o!ti

which involves its own subject-predicate nominative clause) in which

there is a subject-predicate nominative construction.51 (5) Occasion-

ally, the manuscripts even vacillate between an object-complement

construction and a subject-predicate nominative construction in a o!ti

clause,52 illustrating that the scribes probably considered the two con-

structions to be semantically equivalent. (6) As several grammars

point out, when a verb which takes an object-complement construc-

tion in the active is transformed into a passive, the object becomes

the nominative subject and the complement becomes the predicate


48 See the definition of terms above and n. 8.

49 Winer, Treatise, 285.

50 Cf. Matt 16:13; Mark 8:27, 29; Luke 9:20; 20:41; 23:2; Acts 5:36; 8:9; 16:5;

17:7; 19:35 (in D), 20:6, 28:6, Rom 1:22, 14:14, 15:8, 16:19, I Cor 7:7, 26, 32, 10:20, 2 Cor

11:16; Phil 3:8, etc.

51 Cf. John 4: 19; 10:34-36 (though a slightly different situation here); 20:31; Matt

21:26-Mark 11:32; Acts 16:3; Rom 8:18; Phil 2:11; etc.

52 Cf. Rom 10:9 (o[mologh<s^j . . . ku<rion  ]Ihsou?n in most manuscripts; o[molo-

gh<s^j . . . o!ti ku<rioj  ]Ihsou?j in B). We might add here that the biblical authors

occasionally vacillate between the two constructions. For example, Mark 11:32 has a

mixed construction (object-o!ti-predlcate nominative: ei#xon to>n  ]Iwa<nnhn o@ntwj o!ti

profh<thj h#n) which parallels the object-complement in Matt 21:26 (w[j profh<thn

e@xousin to>n   ]Iwa<nnhn). In John 10:34-35 there are parallel thoughts in which one is an

object-complement and the other is direct discourse (though not directly introduced by

a recitative  o!ti:  o!ti e]gw> ei#pa: qeoi< e]ste . . . ei] e]kei<nouj ei#pen qeou>j. Notice also v 36 in

which the thought is carried on: ui[o>j tou? qeou? ei]mi). Cf. also Rom 9:25 and I Pet 2:10

for a similar parallel.



nominative.53 (7) Occasionally, such a passive transform is in a paral-

lel text to an object-complement.54 (8) "The predicate nom. and the

predicate acc. are somet. replaced by ei]j w. acc.",55 suggesting that

both constructions were treated as semantically identical by the bibli-

cal and Koine writers. (9) Finally, the few principles which the

grammars do mention for distinguishing object from complement are

identical with the ones they suggest for distinguishing subject from

predicate nominative.56

Now all of this may seem like a case of linguistic overkill. How-

ever, by firmly establishing that the object-complement construction

is semantically equivalent to the subject-predicate nominative con-

struction, it is possible to make logical deductions both with regard to

the identification of the components and with regard to the semantics

of the construction.57

Having established that the object-complement construction is

semantically equivalent to the subject-predicate nominative construc-

tion, principles used in identifying the components in this latter con-

struction can now be applied to the former. Unfortunately, as

McGaughy laments, "Although the problem of subject identification

. . . appears to be elementary, traditional grammars provide little or

no help in solving it.”58 The introductory grammar by Goetchius is a

rare exception.59 Therefore, I will begin with his principles, making


53 Cf. Robertson, Grammar, 485; Radermacher, Grammatik, 120; Goodwin and

Gulick, Greek Grammar, 228; Kuhner, Grammar, 398. For examples of texts, cf.

Matt 21:13; Luke 1:76; 15:21; Acts 1:23; 4:36; 10:5, 18, 32; 11:13; 1 Cor 4:2; 2 Cor 5:3;

Gal 2:11; Rev 5:4; etc.

54 For similar texts (though not strictly parallel), cf. Luke 1:13 (kale<seij to> o@noma

au]tou?  ]Iwa<nnhn), v 59 (e]ka<loun au]to> . . . Zaxari<an), and 2:21 (e]klh<qh to> o@noma au]tou? 

 ]Ihsou?j). These may be considered parallel in the sense that the verbage is similar

though expressed by two different constructions.

55 BAGD, s.v., "ei]j," 230. sec. 8.

56 Normally the only principle mentioned for either construction is that the article

will be with subject/object, but not with predicate nominative/complement. Goetchius

is a lone exception, giving five principles by which to identify the subject and predicate

nominative. Furthermore, he does, via analogy, apply these principles to the object-

complement construction (cf. Language, 45-46, 142).

57 Although the exegetical implications are far greater in relation to the semantics

of the construction, it is necessary first to establish this semantic equivalence argument

in consideration of the identification of the components.

58 L. C. McGaughy, Toward a Descriptive Analysis of Ei#nai as a Linking Verb in

New Testament Greek (Missoula: Society of Biblical Literature, 1972) 25.

59 Of the more traditional grammars, S. G. Green (Handbook to the Grammar of

the Greek Testament [revised ed.; New York: Fleming H. Revell, n.d.] 179) is the only

one examined to mention that, besides the fact that the subject will have the article, the

subject will often be a pronoun.


refinements along the way.60 Goetchius states:


We may lay it down as a general principle that, if two nouns in the

nominative case are connected by an equative verb in Greek, the more

definite of the two is the subject. Thus:

(a) If one of the two nouns is a proper name, it is the subject. . . .

(b) If only one of the nouns has the article, it is the subject. . . .

(c) If both nouns are equally definite (or indefinite), the one which

has the narrower reference is the subject. . . .

(d) If one of the two nouns has been referred to in the immedi-

ately preceding context, it is the subject. . . .

(e) If an equative verb joins a noun to a pronoun, the pronoun is

the subject. . . 61


From a pragmatic point of view, only two refinements need to be made

of Goetchius' principles. (1) The grid of definiteness vs. indefiniteness

is overly simplistic. One should at least bear in mind that this seman-

tic range is not cut and dried. Rather, there is a continuum from

indefiniteness to qualitativeness to definiteness.62 (2) Goetchius appar-

ently does not believe that the subject-predicate nominative construc-

tion can sometimes be a convertible proposition.63 If so, he virtually

stands alone among grammarians.64


60 From a linguistic standpoint, McGaughy's critique of Goetchius' principles is

well taken (Analysis of Ei#nai, 29-33; cf. 36-54 for McGaughy's solution). However,

from a practical standpoint, Goetchius' treatment does solve the problems in most


61 Goetchius, Language, 46.

62 P. B. Harner has ably pointed out the importance of seeing this continuum

("Qualitative Anarthrous Predicate Nouns: Mark 15:39 and John 1:1," JBL 92 [1973]

75-87). Perhaps the grid of general to specific might be better nomenclature (so

M. Zerwick, Biblical Greek Illustrated by Examples [Rome: Scripta Pontificii Instituti

Biblici, 1963] 55).

63 Goetchius (Language, 46) uses I John 3:4 as an example of his principle '(c)' with

the suggestion that "there are other kinds of lawlessness besides sin." McGaughy

(Analysis of Ei#nai, 32-33) rightly questions Goetchius' use of I John 3:4 in this way:

"rule (c) must be questioned since the meaning of 'definite-indefinite' has been shifted

from a grammatical to a semantic one. In the example under this rule Goetchius

explains that he has chosen a[marti<a as the subject of the sentence because '. . . there

are other kinds of lawlessness besides sin.' In other words, sin is the subject, according

to Goetchius, because it is the more definite of the two concepts. If one were to

interpret this verse theologically, however, he could argue for just the opposite inter-

pretation on the basis of Goetchius' rule: h[ a]nomi<a is the subject because there are

other kinds of sin besides lawlessness. In either case, the point to be noted is that the

determination of the subject on the basis of rule (c) is arbitrary and inadmissable,

therefore, as a grammatical rule."

64 Cf. Robertson, Grammar, 768; Zerwick, Biblical Greek, 56; Harner, "Qualitative

Anarthrous Predicate Nouns," 75, 77; et al. Robertson (Grammar, 769), in fact, uses

Goetchius' same proof text (I John 3:4) as an illustration of a convertible proposition!



As far as the application of these principles to the object-com-

plement construction is concerned, all that needs to be said here is

that they are, indeed, valid. Of the more than sixty constructions

examined in which the order had been reversed between object and

substantival complement, the identification of the object could be

positively made in every instance by using these principles.65 The

verification of this is that in only one passage was there even a slight

possibility of confusion between the object and complement.66 There-

fore, our examination of the reversed order in object-complement

constructions has overwhelmingly confirmed the thesis that the

object-complement construction is semantically equivalent to the

subject-predicate nominative construction.

Exegetically Significant Texts. The only exegetically significant

text which is affected by the issue of the identification of the com-

ponents is also the only one which was slightly ambiguous. But both

the context and the fact that one accusative had the article rendered

the components in John 5:18 as clearly identifiable. The text reads

pate<ra i@dion e@legen to>n qeo<n. It must, of course, be rendered, "he

called God his own father," rather than, "he called his own father



65 For examples involving a proper noun as the object, cf. Matt 3:9; 21:26; 27:22;

Luke 3:8; Acts 8:37 (v.1.); 17:7; Rom 10:9; Rev 9:11. For examples involving a pro-

noun as the object, cf. Matt 14:5; 21:46; Mark 10:6; John 16:2; 19:7, 12; Acts 2:36;

17:22; Rom 4:17; 2 Cor 11:16; Gal 4:14; Phil 3:17; 1 Pet 3:6; I John 1:10. For examples

involving the definite article with the object, cf. Matt 16:13; John 8:41; Phil 2:6; Heb

7:24; 11:26; Jas 5:10; I Pet 1:17; 2:16.

I would also suggest that this analogy between the object-complement and subject-

predicate nominative constructions is valid in distinguishing the subject of an infinitive

from a predicate accusative. Thus, whereas H. R. Moeller and A. Kramer (“An Over-

looked Structural Pattern in New Testament Greek," NovT 5 [1962] 27) argue for word

order as the normal guide when one is faced with "two consecutive case substantives

constructed with an infinitive," when such a construction also involves an object-

complement, there is a better semantic approach than mere word order. Perhaps the

principles for distinguishing subject from predicate nominative are even valid for all

seventy-seven infinitival constructions examined by Moeller and Kramer (and would

thus supplant their word order principle which, at bottom, strikes me more as a

phenomenological approach than a semantic one).

66 I.e., in all but one text (John 5:18) the considerations of sense determined what

was object and what was complement. In all of these the 'rules' coincided with the

obvious sense of the passage. John 5:18 was the lone exception for, apart from these

'rules,' one could conceivably see pate<ra as object and to>n qeo<n as complement.

However, in light of the overall context, such a meaning would be absurd. And even if

the context had been ambiguous, since the validity of the 'rules' has been established in

all other reversed order constructions, such grammatical evidence would be wholly on

the side of taking pate<ra as complement and to>n qeo<n as object.



The Semantics of the Construction

General Principles. With reference to the semantics of the con-

struction, the main question has to do with the difference in force

between the order object followed by complement and the order

complement followed by object. In order to resolve this issue, one can

start with the established thesis that an object-complement construc-

tion is semantically equivalent to a subject-predicate nominative


Specifically, a "rule" developed by E. C. Colwell comes into

consideration here. In an article in JBL in 1933, Colwell stated the

following rule: "Definite predicate nouns which precede the verb

usually lack the article.”67 He went on to point out that "a predicate

nominative which precedes the verb cannot be translated as an

indefinite or a "qualitative" noun solely because of the absence of the

article; if the context suggests that the predicate is definite, it should

be translated as a definite noun. . . “68 The implication from this

study is that to the extent that Colwell's rule is applicable to predicate

nominatives it is equally applicable to predicate accusatives. But

before making the transfer from nominative to accusative, a warning

is in order. Colwell's rule has been abused almost from the time it was

penned. Most grammarians and exegetes have assumed the converse

of Colwell's rule to be equally true, namely, that anarthrous predicate

nominatives which precede the copula will usually be definite. But

such is not the case, as Harner69 and Dixon70 pointed out. Suffice it


67 E. C. Colwell, "A Definite Rule for the Use of the Article in the Greek New

Testament," JBL 52 (1933) 20.

68 Ibid.

69 Harner, "Qualitative Anarthrous Predicate Nouns," 75-87.

70 P. S. Dixon, "The Significance of the Anarthrous Predicate Nominative in John,"

(unpublished Th.M. thesis; Dallas Theological Seminary, 1975). Dixon illustrates the

illegitimate application of the converse of Colwell's rule: "The rule does not say: an

anarthrous predicate nominative which precedes the verb is definite. This is the con-

verse of Colwell's rule and as such is not a valid inference. (From the statement 'A

implies B,' it is not valid to infer 'B implies A.' From the statement 'Articular nouns are

definite,' it is not valid to infer 'Definite predicate nominatives are articular.' Likewise,

from the statement 'Definite predicate nominatives preceding the verb are anarthrous,'

it is not valid to infer 'Anarthrous predicate nominatives preceding the verb are

definite.')," (pp. 11-12).

The problem, methodologically speaking, is that Colwell began his study with a

semantic category (definite predicate nominatives which precede the verb) rather than a

structural category (anarthrous predicate nominatives which precede the verb). This

problem was compounded by the fact that Colwell assumed definiteness in certain

passages (e.g., John 1:1) which were highly debatable. Both Harner and Dixon began

with structural categories and determined the semantic range of such. Their conclusions

were virtually identical: anarthrous predicate nominatives which precede the verb are

usually qualitative (cf. Harner, "Qualitative Anarthrous Predicate Nouns," 87; Dixon,

"Anarthrous Predicate Nominatives," 54-55).



to say here that anarthrous pre-copulative predicate nominatives will

fall within the semantic range of qualitative-definite71 and anarthrous

post-copulative predicate nominatives will usually fall within the

semantic range of qualitative-indefinite.72

Unfortunately, the application of Colwell's rule to the object-

complement construction is severely hampered by the fact that (1) the

infinitive of the copula does not usually occur and (2) when it is

present, the complement usually follows the verb.73

However, there is a further implication derived from Colwell's

study which may prove beneficial to the issue at hand. I have dis-

covered that, as a general rule, in verbless sentences, when the predi-

cate nominative precedes the subject it has the same semantic range

as though it had preceded a verb.74 Thus, by analogy, when an


71 Cf. Harner, "Qualitative Anarthrous Predicate Nouns," 75-87 and Dixon, "Anar-

throus Predicate Nominatives," 31-53, 54-55. As well, from my cursory observation of

this phenomenon in the entire NT, I would agree substantially with their conclusions

(allowing for a somewhat higher percentage of definite predicate nominatives), noting

that I have not discovered one clear example of an indefinite pre-copulative anarthrous

predicate nominative. (The implication of this for John 1:1, then, is still that, on gram-

matical grounds, the translation of qeo>j h#n o[ lo<goj as "the Word was a god" is

inadmissable.) The passages in the NT which contain an anarthrous pre-copulative

predicate nominative that I have discovered thus far are: Matt 4:3, 6; 5:34, 35 (twice);

12:8, 50; 13:39 (twice); 14:26, 33; 23:8, 10; 27:40, 42, 54; Mark 2:28; 3:35; 6:49; 11:17,

32; 12:35; 14:70; 15:39; Luke 4:3, 9, 22; 5:8; 6:5; 11:48; 22:59; 23:6; John 1:1, 12, 14, 49;

2:9; 3:4, 6 (twice), 29; 4:9, 19; 5:27; 6:63 (twice), 70; 7:12; 8:31, 33, 34, 37, 39, 42, 44

(twice), 48, 54; 9:5, 8,17, 24, 25, 27, 28, 31; 10:1, 2, 8, 13, 33, 34, 36; 11:49, 51; 12:6,

36, 50; 13:35; 15:14; 17:17; 18:26, 35, 37 (twice); 19:21; Acts 3:15; 7:26, 33, 52; 9:15;

10:27, 36; 13:33; 16:3, 17 (v. 1.), 21, 37; 22:27, 29; 23:6, 27; 28:4; Rom 1:9; 13:4 (twice),

6; 14:23; I Cor 1:18 (twice); 2:14; 3:16,19; 4:4,16; 6:15, 16, 19; 11:3 (twice); 2 Cor 1:24;

2:15; 6:16; 11:22 (thrice), 23; Gal 3:29; 4:1, 25, 28; 5:4; Phil 2:13; I Thess 5:5; I Tim 6:2,

10; Heb 1:5,10; 3:6; 5:5, 13; 9:15; 11:16; Jas 1:27; 2:23; 4:4; 5:17; I John 1:5; 2:2, 4; 4:8;

and Rev 17:4; 21:22.

72 Cf. Harner, "Qualitative Anarthrous Predicate Nouns," 76.

73 It should be noted here that the verb which takes the object-complement con-

struction only introduces the construction but does not playa part in the semantic

equivalence of this construction with the subject-predicate nominative construction.

Therefore, its position is inconsequential with regard to the semantic range of the

substantival complement (cf., e.g., Mark 11:17 and Luke 19:46; I Cor 9:5).

74 When an anarthrous predicate nominative stands before the subject, it will either

be qualitative or definite. This is apparently due to the fact that (1) had the verb been

present, it more than likely would have come after the predicate nominative (thus

approximating the semantic range of the anarthrous pre-copulative predicate nomina-

tive), and (2) by placing the predicate nominative before the subject, an author is mak-

ing the predicate nominative emphatic (cf. BDF, 248) and if emphatic, then by the

nature of the case, it is moving toward the semantic range of qualitative-definite and

away from the semantic range of indefinite-qualitative (since it is difficult to conceive of

an indefinite predicate nominative being emphasized, though not entirely impossible).

A few illustrations ought to suffice. In John 4:24 Jesus says to the woman at the

well, pneu?ma o[ qeo<j. The anarthrous predicate nominative comes before the subject and




anarthrous complement precedes the object, it will fall within the

semantic range of qualitative-definite. And when an anarthrous com-

plement follows the object, it will tend to fall within the semantic

range of qualitative-indefinite.

For example, when Jesus is called ui[o>j qeou? / a]nqrw<pou in an

object-complement construction, either ui[o<n is anarthrous and pre-

cedes the object (as in John 19:7), or it is articular and follows the

object (as in Matt 16:13). When this is compared with the subject-

predicate nominative constructions, the same pattern emerges. Thus,

in John 10:36 ui[o<j is anarthrous and it precedes the verb, while in

John 20:31 and 1 John 5:5 it is articular and it follows the verb.75

Exegetically Significant Texts. There are literally scores of exe-

getically significant passages which are affected by the issue of the

semantics of the object-complement construction.76 However, one

passage in particular holds some interest for me. In Rom 10:9 there is,

apparently, a soteriological-christological confession: e]a<n o[mologh<-

s^j e]n t&? sto<mati sou ku<rion  ]Ihsou?n . . . swqh<s^. Not only is

this passage exegetically significant, but it serves as an ideal model

text to illustrate the validity of all three issues related to the object-

complement construction. Therefore, this passage will be approached

one issue at a time.


there is no verb. Here, despite the KJV's rendering, pneu?ma is most certainly qualita-

tive, stressing the nature or essence of God. In Phil 2:11 Paul proclaims that ku<rioj

 ]Ihsou?j Xristo<j ("Jesus Christ is Lord"). Here, as in John 4:24, there is no copula

and the anarthrous predicate nominative precedes the subject. In light of the allusion to

Isa 45:23, it is most probable that ku<rioj should be taken as definite ("the Lord"). In

the least, it should be taken as qualitative, not indefinite. By the use of parallel passages

it is possible to confirm the semantic equivalence a bit further. Phil 1:8 reads ma<rtuj

ga<r mou o[ qeo<j. Rom 1:9 reads ma<rtuj ga<r mou e]stin o[ qeo<j. The force of the two

constructions appears to be identical, though only in one is the verb present. However,

in both constructions the predicate nominative precedes the subject. Rom 10:4 reads

te<loj ga>r no<mou Xristo>j in which the sense is most probably, "Christ is the end of the

law." Cf. also Mark 13:8 and 1 Thess 4:6 for other examples.

75 For other texts which seem to demonstrate this analogy, cf. Matt 21:26 with

Mark 11:32 (in which the construction in Matt 21:26 is a reversed order object-

complement and the construction in Mark 11:32 approximates an anarthrous pre-

copulative predicate nominative [see discussion in n.52]); 1 Pet 1:17; John 19:7 with

Matt 26:63. For examples of the semantic range of qualitative-indefinite for a comple-

ment which follows the object, cf. Mark 12:23 (note that e@sxon au]th>n gunai?ka ["they

had her as a wife"] is parallel to the subject-predicate nominative construction in the

first part of the verse: ti<noj au]tw?n e@stai gunh< ["for which of them shall she be a

wife?"]); John 10:33 (in which both a@nqrwpoj and qeo<n are apparently qualitative,

stressing the nature or essence of Jesus); Luke 20:43 (Acts 2:35 and Heb 1:13);

Acts 26:28; Rev 3: 12.

76 Cf., e.g., Matt 10:25; 22:43, 45; Mark 12:37; Luke 20:6, 41; 23:2; John 4:46; 5:18;

10:33, 35, 36; 19:7; Acts 2:36; 13:23; 14:5; 17:7; 28:6; Rom 2:19; 2 Cor 4:5; Phil 2:6, 11;

2 Thess 2:4; 1 Pet 1:17; 2:3 (v. 1.); 3:6, 15; 1 John 1:10 and 5:10 (cf. John 8:44).



            (1) The first question that needs to be asked here is, Is this an

object-complement construction? In answer to that, note that it meets

all three of the principles used in identifying an object-complement

construction: (a) o[mologe<w is used elsewhere with the object-

complement construction;77 (b) there are several clear instances of an

object-complement construction involving two anarthrous nouns, thus

affording a parallel to this text;78 and (c) not only are there no con-

textual factors barring the object-complement from consideration

here, but there are in fact compelling factors to argue in its favor.79

Consequently, the antecedent probability is extremely high that this

construction is, indeed, an object-complement.

(2) The next question involves the identification of the com-

ponents. The analogy of the subject-predicate nominative construc-

tion indicates that the proper noun,  ]Ihsou?n, must be the object and

ku<rion its complement.

(3) Finally, the semantics of the construction needs to be exam-

ined. Specifically, what is the meaning of ku<rion here? Because it

precedes the object, it has already been established that it falls within

the qualitative-definite range. If qualitative, then the meaning is

probably "master." If definite, then the meaning is more likely "Yah-

weh" (i.e., "the Lord”).80 I believe that the meaning "Yahweh" is

probably what is meant here. In support of this are the following lines

of evidence.

(a) From my count, there are five other passages in which the

assertion is made that Jesus Christ is Lord (i.e., ku<rioj is not in

simple apposition with  ]Ihsou?j / Xristo<j, but the two are in a predi-

cate relation). In Col 2:6, the most dubious example, the text reads

parela<bete to>n Xristo>n  ]Ihsou?n to>n ku<rion. This may be read,

"you received Christ Jesus the Lord" (a statement in which no predi-

cation is made), or "you received Christ Jesus [as] the Lord" (an

object-complement construction). If the construction is an object-

complement, it is not insignificant that, although the complement


77 Cf. John 9:22; I John 4:1; 2 John 7. Curiously, Robertson only admits these,

ignoring Rom 10:9 (480), contra BDF (86).

78 Cf. Luke 23:2; 2 Cor 5:4; Jas 1:2; Rev 9:11.

79 Although the force of o[mologe,w is most compelling on the side of an object-

complement, I found the Douay and KJV to deny the construction here; and of the

modern texts examined, I found the same error curiously enough 'preserved' only in

the New KJV.

80 The qualitative idea, of course, would stress more what he does rather than

specifying who he is (cf. I Pet 3:6). A definite ku<rion would probably have a par

excellence force to it. Thus, by implication, since Yahweh is the one who deserves the

name "Lord" above all others, Yahweh could well be implied by a definite ku<rion.



(ku<rion) follows the object (Xristo>n  ]Ihsou?n), it too has the article.81

2 Cor 4:5 records the apostle's proclamation: khru<ssomen . . . Xristo>n

 ]Ihsou?n ku<rion. Since Paul has placed the complement (ku<rion) after

the object (Xristo>n  ]Ihsou?n), and has not added the article, this

could be an exception to the suggestion made here about Rom 10:9

(i.e., it seems, by the grammatical principles laid down, that Paul is

only declaring Christ to be master here, not Yahweh). But the context

makes it clear that the author's emphasis is indeed that Christ is

master, without reference to his deity, for the apostle goes on with the

mildly antithetic parallel:  khru<ssomen. . . e[autou>j de> dou<louj. There-

fore, this text in no way nullifies the proposal for Rom 10:9. In

I Cor 12:3 the apostle puts up the challenge: ou]dei>j du<natai ei]pei?n:

Ku<rioj  ]Ihsou?j ei] mh> e]n pneu<mati a[gi<&. There is dissension among

the Greek witnesses, with several of the key Western and Byzantine

texts converting this into the accusative (and hence, an object-

complement construction). But even in these manuscripts, the order is

the same.82 These three texts, in the least, do not argue against the

view of Rom 10:9 suggested here. In the first text (Col 2:6), the

complement followed the object and was articular; in the second

(2 Cor 4:5), though the complement was anarthrous, it was argued

that Paul's emphasis was on Christ as master, not as Yahweh; and in

the third (1 Cor 12:3), the statement and word order were parallel to

Rom 10:9.

There are two other texts, however, which make a substantial

contribution to this discussion. In one, Phil 2:11, a subject-predicate

nominative construction is in a o!ti clause (e]colmologh<shtai o!ti ku<rioj

  ]Ihsou?j Xristo<j); in the other, 1 Pet 3:15, there is a probable

object-complement construction introduced by a[gia<zw (ku<rion de> to>n

Xristo>n a[gia<sate). In both of these texts, there is an allusion to the

OT and specifically to Yahweh himself (Isa 45:23 and 8:13 respec-

tively).83 Thus, in the two parallel passages where the ku<rioj clearly


81 This, of course, is in keeping with Colwell's rule which asserts that a definite

predicate nominative will either lack the article and precede the verb or have the article

and follow the verb (or, in this case, the object).

82 This text is in reality parallel to Rom 10:9 for it too makes a particular con-

fession the test of faith. Rom 10:9 should be the basis for interpreting I Cor 12:3,

rather than vice versa, because the evidence for I Cor 12:3 is far more scanty than in

the Romans text.

83 Isa 45:23 reads, Nvwl-lc fbwt jrb-lk frkt yl-yk (cf. vv21-22 for the

identification of the speaker as God [v 22—lx-ynx yk], i.e., Yahweh [v 21—hvhy ynx]),

and the LXX translates, o!ti e]moi> ka<myei pa>n go<nu kai> e]comology<setai pa?sa

glw?ssa t&? qe&?. Paul quotes this text in Rom 14:11 with reference to God and alludes

to it in Phil 2:11 with reference to Jesus. Isa 8:13 reads, vwydqt vtx tvxbc hvhy-tx

(LXX: ku<rion au]to>n a[gia<sate). (Note that the direct object marker tx makes possible

an object-complement



refers to Yahweh, even though this predicate noun is anarthrous, the

biblical author places it before the object/subject to indicate that it is

definite. Apparently, not only was the article unnecessary, but the

reversed order seems to be the 'normal' way to express the idea that

ku<rioj is definite.84

(b) Codex Vaticanus strays from the pack in Rom 10:9, changing

the object-complement to a subject-predicate nominative construction

following o!ti. If the preceding argument has any validity at all, then

the variant only strengthens the view that ku<rion is equivalent to

Yahweh here.

(c) Finally, Paul continues his message in v 13 by adding a quote

from Joel 3:5, "Everyone who calls upon the name of the Lord shall

be saved." The Hebrew text of Joel 3:5 has hvhy for 'Lord' here. In

vv 11 and 12 of Rom 10, Christ is still clearly in view; thus, to suggest

that kuri<ou refers to the Father ignores the obvious connection Paul

is making here: to confess that Jesus is Lord is to confess that he is

the Lord of v 13. If so, then the confession is of Jesus as Yahweh.



The object-complement construction can be profitably put

through the structure-semantics grid. Three issues with respect to this

construction were raised in this study: (1) the identification of the

construction, (2) the identification of the components, and (3) the

semantics of the construction. With reference to the identification of

the construction, three principles were suggested: (a) the verb related

to the construction must be able to take an object-complement,

(b) the specifics of the structure in question must have parallels in

positively identified object-complements, and (c) there must be strong

contextual overrides to prevent one from so tagging such a construc-


Under the heading of the identification of the components the

major thesis of the paper was stated, namely, the object-complement

construction is semantically equivalent to the subject-predicate nomin-

ative construction. Hence, the guidelines for one are guidelines for the

other--both with reference to the identification of the components

and with reference to the force of the construction semantically.


construction in the Hebrew; the Greek is very clear. Elsewhere in the LXX, a[gia<zw

takes an object-complement [cf. Exod 29:1 and 30:30 and the discussion of these texts

in n. 32].)

84 It is possible that the article was not added to ku<rioj in order to distinguish the

subject/object from the predicate noun.



            Concerning the semantics of the construction, it was noted that

when the order was complement then object, the complement would

fall within the semantic range of qualitative-definite. When the com-

plement followed the object it would tend to fall within the range of

qualitative- indefinite.

With application to exegesis, just a few of the scores of passages

affected by this study were noted. Among them, Titus 2:10 and

Rom 10:9 received lengthy treatments and I suggested that the per-

severance of the saints and the deity of Christ were implicit in these

texts, respectively.

In conclusion, although the reader may find some of the exe-

getical suggestions stated herein to be debatable, he should remember

that the purpose of this paper is not primarily to come to exegetical

conclusions, but to raise exegetical questions on the basis of a better

understanding of the semantics of a particular grammatical construc-

tion. Therefore, if the grammatical arguments set forth in this paper

help the exegete to see new possibilities (e.g., in Titus 2: 10; John 2:11;

4:54), or to strengthen old views (e.g., in John 5:18; Rom 10:9), this

purpose has been accomplished.





This material is cited with gracious permission from:

            Grace Theological Seminary

            200 Seminary Dr.

            Winona Lake,  IN   46590

Please report any errors to Ted Hildebrandt at: