Where the world comes to study the Bible

Chapter 2: Discussion of Research Relevant to the Target Cluster

Chronological Review of Previous Research

Henry Moeller and Arnold Kramer first address the broader problem of distinguishing the semantic roles between two consecutive substantives.1 They write, “In Koine Greek the function of a substantive is not always unanimously signaled by its inflectional form. Thus it occurs that within a given phrase-structure, two words may be similar in their inflectional terminations, yet be quite different in function. In such a case, some criterion other than form must be invoked to help determine the functions of the two otherwise formally similar items.”2 The conclusion of their study provides a good launch point for future work, but not without some concerns. It is stated as follows:

Of two consecutive accusative case substantives constructed with an infinitive, the first in order functions as the subject term, the second as the predicate term, except that the predicate terms stands first in the following circumstances:

    1. When one of the two accusative terms

      a. is a form of τίς , τί in direct interrogative use (Pattern III, 7 exx.); or

      b. is τινα (indef. Pron.), occurring with a reflexive pronoun after εἶναι (Pattern IV, 2 exx.); or

      c. is κόπον with the inf. παρέχειν (cf. discussion of Pattern I, above); or

    2. When the infinitive (except εἶναι, γενέσθαι, and ὑπάρχειν) follows the accusatives (unless the accusatives are preceded by δεῖ' or ἔξεστιν (Pattern I.2., 3 exx.; IIa, 1 exx.)). 3

The limitations of the study are fourfold: (1) varying semantic situations, (2) loose commitment to structural priority, (3) overemphasis on the verbal proximity theory, and (4) focus on a peripheral case form. First, the study fails to preserve the semantic situation by not focusing on one verb and by not focusing on the same types of substantive comparisons. In other words, it lacked a tightly defined cluster of syntactical units. Second, though the study makes semantic assertions for what it calls patterns and reflects a desire to give priority to structure, it can only do so much with multiple verbs (without regard for whether they are true copulas) and multiple types of substantives. Third, Moeller and Kramer’s objectives are somewhat contradictory in that they try to show that the subject is the first of the two substantives and that it is the closest substantive to the verb. This idea of verbal proximity confuses their observations. Perhaps, this is why the idea is less pronounced toward the end of the article.

Finally, the article looks at an oblique case. It is not faulted for this because such was its objective. However, methodologically speaking, it may be slightly off-center because the challenge is to distinguish subject from predicate in constructions of double substantives. Since “rules” derived from the naming case also provide solid footing for deciding between double accusatives, it makes more sense to study this problem in the nominative case. The most common use of this case is subject of a finite verb. Technically speaking, participles and infinitives take no subject.4 This suggests that the nominative case is the “core case form” for a study on substantives competing for subject of a clause.

Eugene Van Ness Goetchius may be the first to provide a clear statement of the core problem: “the absence of a fixed word order in Greek makes it difficult for us to distinguish which of two nouns in the nominative case is the subject of an equative verb and which is the predicate nominative.”5 He solves the puzzle by identifying subset propositions, noting that one substantive is more definite than the other. This becomes the key for Goetchius, asserting that the notions of definiteness and indefiniteness provide the way to identify the subject of an equative verb. He then provides a “pecking order” based on this principle:

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:

Goetchius’ work contributes much to the discussion. It provides one of the best descriptions on linguistic structure and grammatical analysis and proceeds to demonstrate a strong commitment to structural priority, as evidenced by the five-part statement of conclusions.6 As a result, it provides the first semblance of a pecking order, focuses on the naming case forms, and fixes the discussion on the equative verb, εἰμί. It is also the first to state clearly the principle upon which it bases its conclusion. Yet the work possesses a few gaps.7 First, the choice of definiteness as the discriminating principle, limits the study from the beginning. Too many of the double nominatives stump the exegete precisely on this point. They appear to have the same degree of definiteness, whether arthrous or anarthrous. Second, it provides very little empirical evidence to support each of the five parts in the pecking order. Finally, it does not address the precise problem of this investigation, SPN constructions involving the target cluster.

Moeller and Kramer’s article introduces the broader challenge of choosing a subject from consecutive double substantives with equative verbs. Goetchius focuses the study on the nominative case (the core case form of the problem). Gordon Fee starts the discussion of the exegetically significant passages of this thesis.8 Though heavy on text-critical issues and hardly focused on the issues of a pecking order, Fee’s article makes a couple of important observations. First, in defining the semantic situation within which an older observation regarding anarthrous uses of ᾿Ιησοῦς holds true, he indirectly expresses the view that in John 20:31 ᾿Ιησοῦς functions as subject of the clause.9 In effect, this statement covers the other four Johannine passages as well because three match the semantic situation exactly and one differs only by a particle of negation.10 He then adds a two-fold expansion to the comment by saying that the practice of anarthrous forms of ᾿Ιησοῦς is not unique to John and that it holds true for all personal names in the nominative case when the name precedes the verb. The significance of these statements resides in their underlying presupposition, that proper nouns take precedence over articular nouns. It appears that Fee applied Goetchius’ principle (a) rather indiscriminately by labeling the proper name as subject in John 20:31 without explaining why he did not apply principle (b).11 Fee’s study concerned the use of the article with personal names in John, so the lack of explanation is understandable.

Next, Lane C. McGaughy tackles the specific problem with much better footing. Using nomenclature from Funk’s Beginner’s Grammar, he lists six kernel sentence types and then focuses on Type S-II because it reflects the copulative use of the εἰμί verb.12 He concludes: (1) Rule 1: The subject is that word or word cluster which agrees in person and number with the personal ending (bound morpheme) of the equative verb; (2) Rule 2: The word or word cluster with head term in the nominative case is the subject; (3) Rule 3: The subject is determined by its antecedent; (4) Rule 3a: Demonstrative and relative pronouns are subjects; (5) Rule 3b: The subject is indicated by zero anaphora; (6) Rule 3c: The word or word cluster determined by an article is the subject; and (7) Rule 3d: If both words or clusters are determined by an article, the first one is the subject.13

His work adds to previous studies in several ways. First, it provides an excellent review of older solutions to the problem.14 Second, he presents a very fair critique of Goetchius’ work.15 Third, it pushes the discussion forward by setting out to order the rules so as to minimize conflicts. He shows how the ordering of the rules requires an understanding of the interplay between the discourse structure (sentence transcending) and the syntactic one. Fourth, McGaughy’s Rule 1 and Rule 2 hold true for all cases and are truly syntactical. Fifth, he begins a well-pronounced shift in the discussion from syntactical considerations to phenomenological ones in pursuing a solution to situations where the subject, as he contends, can not be determined purely from the sentence.16

Finally, his methodology provides a sound model for research on problems where syntax is silent. Rule 3 sprouts from the new ground of discourse analysis, a study which relies heavily on an elaborate reference system that is highly dependent on semantic configurations.17 He uses the notions of anaphora, contrast, and coordination to create the rules before coming back to the explicitly labeled “words and word clusters” which fill the functional spots of subject and predicate. In other words, he uses discourse analysis (participant identification) to determine the subject in his sample passages and then organizes what he has termed as subject to show the statistical representation of his theory. Unfortunately, this fourth advance also leads to some weaknesses.

McGaughy’s work raises a few concerns. First, he appeals much to discourse analysis, and to “participant identification” in particular, which results in the correct shift to use sentence transcending examinations in order to better understand what may be the subject in the target cluster. He explains that, in Greek, an elaborate reference system signals identification of the participants.18 He formulates the third rule largely from this system. It assumes the solution must remain outside of the sentence itself. This may be true, but McGaughy’s study does not prove this. It simply succeeds in showing that certain cases require this kind of analysis. Such conclusions fail to exhaust the potential of his approach.

Second, he defines the construction under investigation in terms of functional rather than formal nomenclature.19 Although he maintains “structural” integrity by keeping the discussion focused on S-II sentences, his work does not technically adhere to structural priority of forms (morphological and syntactical).

Finally, not only is the study not based on syntactical structure types, but McGaughy’s work does not help the present study because it does not provide a solution for five of the six target cluster passages in the New Testament. He simply labels these exceptions. Regarding their subject, he explains the variance as due to their formulaic character as early christological confessions following certain verbs. Regarding the predicate of these exceptions, he points out that the earliest occurrence of this alleged confessional statement assigns the articular word cluster a predicate function.20

Although coming from a philosophical angle, focusing on Attic Greek, and not examining the specific target cluster, Charles H. Kahn still adds to the discussion in three ways.21 First, he continues to call attention to the challenge of determining subject from among double nominatives joined by a copula. Second, he highlights the fact that the greater challenge is with convertible propositions.22 But his more pronounced input has to do with methodology rather than with offering a solution to the problem.23 He argues strongly against any hypothesis based on word order.24 And he challenges the significance of statistical preponderance, cautioning against explaining variations thereof as special stylistic or rhetorical intentions.25

Goetchius returns to the discussion with a book review of McGaughy’s dissertation.26 He strongly affirms Rule 1 and only mildly refines Rule 2.27 However, by urging that Rule 3 be “applied rigorously” to the so-called exceptions in Johannine literature, he launches the discussion of these passages into a new level. He suggests that ᾿Ιησοῦς is the predicate nominative in all five. To support this claim he appeals to the analogous situation involving infinitives and the accusative case (Acts 5:42, 18:5, and 18:28). However, he misses the fact that his sample passages do not match the word order of the five passages in question (a proper name, first, followed by the equative verb, followed by an articular noun). In his examples the double substantives are consecutive and the articular noun precedes the anarthrous proper name.28 Goetchius also questions Rule 3d on grounds that it does not qualify as a Type S-II sentence. However, he bases this on issues of “definiteness,” the very thing for which he was critiqued by McGaughy.

Shortly after Goetchius’ book review, D. A. Carson writes an article which interacts heavily with Goetchius’ appraisal of McGaughy’s work and which references the work of Moeller and Kramer to support his translation of John 20:31.29 He contributes to the discussion primarily by highlighting the exegetical significance of this thesis. His article’s stated purpose is “to examine an overlooked syntactical unit in John 20:31 and, on this basis, to suggest an alternative understanding of the first ἵνα-clause.”30 Unfortunately, Carson does not identify any passages which previous works have not already discussed.31 Instead, he attempts to synthesize the observations from McGaughy and Goetchius in order to create a syntactical polemic for his view.

Carson begins the argument by observing that all the Johannine passages which McGaughy calls exceptions to Rule 3c are christological in nature. Next, he points out that the three passages from Acts also deal with Christology. By this he means to collect eight pieces of evidence which can be shown to adhere to McGaughy’s Rule 3c. Then he appeals to Goetchius’ review of McGaughy and, by way of corroboration, to Moeller and Kramer’s study in order to show that the subject in the Acts passages can and should be the articular noun, τὸν Χριστὸν. He then returns to the four passages in 1 John to argue from historical background that the problem John addressed was that of a protognosticism which refused to connect “the Christ” and “the Son of God” to the historical person, Jesus. Consequently, he concludes that “it is entirely understandable if the crucial christological confession, from his [John’s] point of view, is that the Messiah is Jesus (2:22b; 5:1) or that the Son of God is Jesus (4:15; 5:5c).”32

Finally, he claims that “there is every syntactical reason for thinking that the crucial clause should be rendered, ‘that the Christ, the Son of God, is Jesus.’”33 What he means by this is elusive. The majority of this article deals with issues of biblical theology, historical context, literary context, and textual criticism. Carson devotes two pages to the discussion on syntax and eight to implications and reflections on the purpose of John’s gospel. Very little of these two pages has anything to do with a fresh syntactical analysis of the target cluster. Of the twelve implications and reflections listed, only one truly addresses an issue of grammar.34

The investigations on syntax to which he appeals have hardly established his case. As was shown above, Moeller and Kramer’s work approached the analogous problem rather imprecisely. And even so, their stronger points pertain to consecutive accusatives and not those split by the verb. As for Goetchius’ review, the problems with his conclusion are stated above. Both Goetchius and Carson make much of McGaughy’s Rule 3c but virtually ignore Rule 3d. And yet the protasis of Rule 3d better reflects the problem which the target cluster faces. Carson and Goetchius fail to see that McGaughy’s approach requires clear or strong participant identification in order to perceive the patterns which are then translated into rules. The procedure requires a way to distinguish between varying degrees of determinacy. Rule 3d confirms this.35 In summary, I think Carson’s article offers little towards resolving the problem of distinguishing S from PN in the target cluster. It simply highlights the exegetical significance of the present study.

Jeffrey T. Reed provides a great example of how to proceed with the problem at hand as it relates to the accusative case.36 In addition to a lucid and fair critique of Moeller and Kramer’s seminal work, Reed describes and executes a new research approach that effectively deals with the many exceptions (Moeller and Kramer) which the original study identified.37 Reed’s new rule states, “Of two accusative case nominal, pronominal, or adjectival substantives with an infinitive, the first in order functions as the subject, the second as the object/predicate.”38 The major contribution of this study is that it provides a great template for stating and then testing a falsifiable hypothesis, as it involves the semantic realm.39 However, it is not without limitations, and is best explained by the word imprecision.

The first limitation is a loose use of functional nomenclature. Discussions of unmarked and marked word or phrase order typically employ S for subject and O for object.40 These broad terms serve the purposes of linguistic studies quite well. However, empirically oriented grammatical studies benefit more from syntactical and functional labels which possess greater specificity. It would have been helpful to see the semantic configurations broken down into the twelve permutations of the five semantic terms (S, O, P, Ic, Inc).41 Reed’s use of O for both predicate and object does help to display a rule which “handles the most examples and has the fewest ‘exceptions.’”42 But it also creates the appearance of stronger empirical substantiation than actually exists.43 The lack of clarity as to how he determines the subject in each of his examples also limits the study. Some of the discussions imply that he is using linguistic analysis as well as a grammatically based pecking order. Finally, there also appears to be a lack of structural priority. The problem that all studies face, the paucity of the data, may explain why Reed looked at such a broad pool of substantives. Unfortunately, such an approach limits his study’s ability to make semantic statements about specific syntactical structures.44 For example, it does not contribute to our understanding of the pecking order among substantives. As a matter of fact, it seems he uses a system of grammatical tags to determine which of the two substantives is subject. Then he notes its position in relationship to the other.

In his commentary, Carson’s discussion of John 20:31 maintains the position of his first article.45 He simply states that there is high probability, on syntactical grounds, that his translation best represents the intent of the author.46 As with his journal article, it adds little to resolving the problem of distinguishing S from PN in the target cluster.

Gordon Fee re-enters the discussion at the text-critical level with an essay entitled “On the Text and Meaning of John 20:30-31,” but not without making some relevant comments on Carson.47 In the last footnote, he notes that his conclusion clashes with Carson’s and proceeds to argue against Carson’s over-reliance on McGaughy’s Rule 3c. Fee points to his own findings regarding Johannine use of the anarthrous ᾿Ιησοῦς as explanation for McGaughy’s exceptions. He writes, “Johannine usage on the whole suggest that ᾿Ιησοῦς functions as the subject in this clause.”48 Because Fee addresses the topic of my thesis minimally, by way of one footnote, it is difficult to critique his comments. One concern does surface. In attempting to show the significance of his observations of proper names in Johannine ὅτι-clauses, he devalues the significance of observable tendencies within the target cluster. He avers that the tendencies of a proper noun in a ὅτι-clause are “of more significance than ‘syntactical links to ἐστίν.’”49

Matthew A. Cripe approaches with greater clarity the question of whether word order helps to distinguish subject from object in infinitive clauses containing two accusatives.50 His review of the Moeller and Kramer article concludes that though it is somewhat helpful, “it is still inadequate in many respects.”51 His assessment of Reed’s article is equally forthright.52 His inductive study shows a clear difference between copulative and non-copulative constructions. It also shows that the rules governing the naming case apply to copulative infinitive clauses involving double accusatives.53 It borrows from what is known in the nominative case to make assertions about this oblique case. In this sense, it highlights the value of any work which expands our understanding of the pecking order for the nominative case. One criticism I have of this study is that it mixes functional nomenclature with formal nomenclature resulting in a confusion of what should be meant by structural type.54

Daniel B. Wallace’s work provides the most explicit and fair treatment to date. His method for distinguishing S from PN in the target cluster addresses both subset and convertible propositions. The rule for subset propositions is as follows: (1) the subject will be a pronoun, whether stated or implied in the verb; (2) the subject will be articular; and (3) the subject will be a proper name. The pecking order for convertible propositions is as follows: (1) the pronoun has greatest priority and (2) articular nouns and proper names seem to have equal priority.55 Wallace’s work advances the discussion in four ways. First, it states a clear core guiding principle, that the subject is the known entity. Second, it delineates the two semantic categories of SPN constructions—subset and convertible. Third, it provides the clearest statement of the problem which this study examines, pecking order #2. Finally, Wallace tentatively suggests that word order signals the subject and invites empirical substantiation. I have but one reservation. The claim that ὄνομα is an exception to #3 may be overstated. My study captured every use of this noun with the article and provides some additional insights.56

Andrew K. M. Adam may be the latest grammarian to supply a system for distinguishing S from PN in the target cluster. He asserts, “A demonstrative pronoun or a relative pronoun is more likely to be the subject than any other nominative. A personal pronoun or proper noun is more likely to be the subject than any other of the remaining sorts of nominatives. An articular noun is more likely to be the subject than the anarthrous noun. If either nominative is distinctly more definite than the other, the linking verb may simply be equating the two, and the word that comes first is probably the subject.”57 Adam’s system is fair in that it consistently speaks of likelihoods rather than certainties. It is insightful in that it suggests what to do in the case of convertible propositions. However, it seems archaic at the foundational level. Like Goetchius, it is based on the meaning-based principle of definiteness which McGaughy has shown to be limited. Another critique of Adam’s pecking order is that it lacks substantiation. There are no references to other works or empirical studies to support it.

Carson enters the discussion yet another time.58 In an article awaiting publication, he claims that while the original work aspired to show the bearing of McGaughy’s research to the interpretation of John 20:31, this latest manuscript aims to respond to Fee’s essay in The Four Gospels 1992: Festschrift für Frans Neirynck.59 Carson does just this and with an exceptionally helpful evaluation of Fee’s article on arthrous and anarthrous proper names in the Johannine corpus.60 Unfortunately, he does not provide any new examples of the target cluster or any new insight regarding it.

Despite this lack of additional data, several of the comments added to the original piece are worth mentioning. First, Carson claims that his translation of the verse, and Goetchius’, is in line with a rising number of commentators.61 But this hardly constitutes a solidifying of the alleged grammatical grounds for his translation. It simply means that more people are agreeing with his view on the historical setting of the gospel.

Second, he adds an enigmatic comment to the original argument which garnished support from Moeller and Kramer. He writes, “In each of these contexts, the ‘given’ for the hearers is ‘the Christ.’”62 He is referring to the contexts of Acts 5:42, 18:5, 18:28, and John 20:31. But what he means by ‘given’ is unclear. If he is alluding to Goetchius’ underlying principle (definiteness) for determining subject, then he appeals to a deficient method. Wallace’s core guiding principle, that the subject is the known entity, produces much better results. If he is referring to McGaughy’s umbrella Rule 3, that the subject is determined by its antecedent, then he rightly identifies the subject for the Acts passages but dismisses McGaughy’s assessment that the more sensible antecedent for John 20:31 (and for the four passages from 1 John) is “Jesus.”

Third, Carson asserts that the recent work of Reed supports his translation.63 But this is plain wrong. He misses the point. Reed determines the subject based on word order, and by way of improvement to the research of Moeller and Kramer, his study includes accusatives split by the infinitive. If anything, Reed’s work supports the traditional translation of John 20:31 since ᾿Ιησοῦς is the first nominative of two split by the equative verb.

Fourth, Carson acknowledges Fee’s parting comments in the 1992 essay and more importantly, offers significant interaction, albeit as rebuttal. He begins very graciously by pointing out that he does not over-rely on McGaughy’s rule, but more precisely he over-relies on an extension of it. He clarifies, “In all fairness to McGaughy, however, in his original study he saw John 20:31, and four passages in 1 John, as exceptions to his own rule. I followed Goetchius in arguing that McGaughy’s rule is better than McGaughy himself thought it was. If Fee is right on 20:31, then I am wrong, but not McGaughy. The ‘major flaw’ in my argument, if there is one, is not in my ‘reliance on L. C. McGaughy,” but in my extension beyond him.”64

Carson then proceeds to pick apart Fee’s parting footnote, rightly showing that the absence of the article on ᾿Ιησοῦς should have no bearing on the rule, whatever it may be, governing the target cluster in question. He does so by evaluating Fee’s original study which yielded nine ὅτι-clauses in which ᾿Ιησοῦς precedes the verb.65 However, before dismantling Fee’s case, he insightfully points out that the primary cause for disagreement in grammatical approaches to subject determination is the fact that they are based on differing constructions. He writes, “Fee is interested in all instances where ᾿Ιησοῦς precedes the verb within a ὅτι-clause; McGaughy is interested in all instances of ἐστίν as a linking verb that joins a subject and complement.”66 Once noting this he quickly eliminates the four passages which do not use ἐστίν and highlights that Fee allows ᾿Ιησοῦς to function as predicate in John 5:15, 20:14, and 21:4. Having shown this he questions Fee’s reluctance to allow it elsewhere. Consequently, Carson sees no reason why it should not function as such in John 20:31 as well.67

In summary, this chronological review captures some of the major discussions and advancements that relate to subject determination in equative clauses involving anarthrous proper nouns and articular nouns. It accentuates the need for the present research and informs it on how to proceed so that it will yield sound conclusions and observations regarding the behavior of the target cluster in Koine Greek.

Methodological Direction for Present Research

The fundamental question is whether an author tended toward a certain practice in expressing the grammatical S in a SPN construction involving the target cluster, the most difficult form of this question being when the target cluster is part of a convertible proposition. The falsifiable hypothesis asserts that the first nominative in target clusters functioning as SPN constructions is subject.

There are at least three approaches to proving this true—indirectly, directly, or a combination of the two. “Indirectly” refers to the use of analogy. It has been shown that corresponding challenges exist for the oblique cases. A study by analogy would gather and organize all such analogous examples and observe patterns of more broadly defined structures. Reed’s work with the accusative case provides a fitting example and excellent starting point. He expanded the limits and was able to get more samples. Rather than focusing on nouns, or pronouns, or adjectives, Reed looked for substantives of all three. The same principle can be expanded even further to include all matching case double substantives with a copulative verb. The value of this approach is that it increases the number of observable data within a given time period. It increases the chances of finding plenty of New Testament text examples of the observed practice (rule) at work. The challenge of this approach, however, is to show that the situations are in fact analogous. The other approach is to define a specific structural type, to look for it in the extant Greek literature, to record its propensity to function one way over another, and to test it against the falsifiable hypothesis. The value of this approach is the ability to make semantic assertions about specific structural types. The difficulty with this approach is the paucity of the data when such a specific syntactic structure is in view.

After careful review of the various directions taken by previous studies, it seems prudent to proceed with the latter approach. It is admittedly more arduous but promises greater clarity for the specific problem at hand. In examining all of the other studies related to the accusative case I did not see many examples, if any, of this particular target cluster in another case form. I cannot remember seeing a single example in an oblique case of the specific problem with which this research deals. This leads me to believe that it is even less common in the other cases. This is one reason why I chose to focus on the nominative case.

Another reason is that recent studies of SPN constructions in the accusative case appeal much to the rules which govern the nominative case. They look to the solid work which has already been advanced in this area for the naming case.68 It is treated as the core case form for answering the same or similar question regarding S in an SPN construction. Focusing on the target cluster will provide additional insight. This provides some reasoning for choosing the nominative case but two more questions remain—why such restriction on the verb and why use only nouns?

Simple pragmatism stands behind the reasoning for restricting the verb so much. The preliminary research did allow for all moods of the verb. However, it yielded no examples other than the indicative. But the primary reason the verb was restricted to three forms is because of search limitations on Thesaurus Linguae Graecae #E (TLG #E).69 Terms are not grammatically tagged in its data bank. Therefore, the case, gender, or number of a noun can not be specified for the search strings. Neither can the syntactical labels of verb, noun, or proper noun. For example, the only way to find substantives was to use each of the three nominative articles for the query. Since this provided too much data without further specification, the verb was limited to the three forms. This reduced the number of matches significantly and yet not enough. In any case, this is the primary reason why the verb is so tightly defined. By way of reiteration, the original study did utilize Accordance Bible Software 5.1 (Accordance) for searches in the Septuagint, New Testament, and Apostolic Fathers, which allowed for queries of the unrestricted verb.70 Yet it only produced examples of the three verb forms chosen for the TLG #E.

Articular non-noun substantives were not included for a couple of reasons. The first reason is that they slightly complicate the semantic situation. With articular substantives consisting of syntactical units other than nouns, the article functions primarily to make substantive an inherently non-substantive. Therefore, it essentially adds another variable to the problem at hand. Is the article only nominalizing the non-substantive or is it also marking the syntactical unit for a specific function? The articular noun does not present this extra question. The article simply marks the syntactical unit for use, like subject or apposition. To state it differently, it affects it functionally but not lexically, as with non-substantives.

These are a few reasons why the search string was so tightly defined. One other merits mentioning. Simplicity brings clarity. In research, reduction of variables helps to achieve this. The tightly defined target cluster reflects an attempt at clarity.


1 Henry R. Moeller and Arnold Kramer, “An Overlooked Structural Pattern in New Testament Greek,” Novum Testamentum 5 (1961): 25-35.

2 Ibid., 25.

3 Ibid., 32.

4 Daniel B. Wallace, Greek Grammar beyond the Basics: An Exegetical Syntax of the New Testament (Grand Rapids: Zondervan, 1996), 38.

5 Eugene Van Ness Goetchius, The Language of the New Testament (New York: Scribner, 1965), 45.

6 Goetchius, Language, 13-27.

7 In appendix two of this thesis, the review of McGaughy’s work adds to the list of three presented here.

8 Gordon D. Fee, “The Use of the Definite Article with Personal Names in the Gospel of John,” New Testament Studies 17 (1970-71): 168-83.

9 Ibid., 179. Fn. 3 explains that Wies (1913) had also noted that ᾿Ιησοῦς is anarthrous whenever it preceded the conjunction in Johannine literature but failed to state that this is the pattern only when the subject precedes the verb. And fn. 6 cites John 20:31 as one of nine examples where the subject, ᾿Ιησοῦς‚ precedes the verb in a ὅτι-clause.

10 1 John 2:22 ( ὅτι ᾿Ιησοῦς οὐκ ἔστιν); 1 John 4:15 ( ὅτι ᾿Ιησοῦς ἐστιν) ; 1 John 5:1 ( ὅτι ᾿Ιησοῦς ἐστιν); and 1 John 5:5 ( ὅτι ᾿Ιησοῦς ἐστιν).

11 In all fairness to Fee, I could not find a single reference to Goetchius in the article so it is possible that Fee may have operated on older notions for determining subject in double nominative constructions. The point, however, remains the same. It appears that Fee presupposes that the proper name is subject regardless of other factors.

12 Lane C. McGaughy, Toward a Descriptive Analysis of Einai as a Linking Verb in New Testament Greek (Nashville: Society of Biblical Literature for the Linguistics Seminar, 1972), 21-23. McGaughy credits the six kernel sentence types to Robert Funk in fn. 1 which states, “These are the six basic sentence types of New Testament Greek as listed by Funk, Beginner’s Grammar of the Greek New Testament, §§554-571. As is clear from the above summary, the six kernel sentence types are, in fact, six predicate types. Consequently, the sub-types of S-II discussed in this study will be the different predicates occurring with S : be. N. B. S-I, etc., = Sentence Type I, etc.” (McGaughy, Descriptive Analysis, 21). Without the accompanying illustrative passages, the types are as follows: (1) S-I = S-V (S=Subject; V=Intransitive Verb), (2) S-II = S-E-SC (E=Equative Verb; SC=Subjective Complement), (3) S-III = S-V-O (V=Transitive Verb; O=Object), (4) S-IV = S-V-IO-O (IO=Indirect Object), (5) S-V = S-V-OC (OC=Objective Complement), and (6) S-VI = S-V-O-O.

13 McGaughy, Descriptive Analysis, 36-54.

14 Ibid., 23-26.

15 Ibid., 29-33. McGaughy states the critique in two broad statements, that Goetchius’ solution rests on meaning based-distinctions (definiteness) and that the analysis contains major difficulties. He explicates four from the latter: (1) grammatical (morphological and syntactical) categories are mixed with contextual and meaning based (semantic) ones; (2) the rules are not ordered (i.e., it does not account for conflicts like John 8:39); (3) Goetchius fails to show that the rules do derive from Jespersen’s basic principle; and (4) rule c does not actually hold true, even for the example given (therefore, rule c is “arbitrary and inadmissible”).

16 Ibid., 40-41.

17 McGaughy, Descriptive Analysis, 45. He asserts, “the following sub-rules specify the major surface signals of S-II subject identification which derive from the underlying discourse structure of Greek.”

18 According to McGaughy, this system includes nominal anaphora, pronominal anaphora, zero anaphora, contrastive pronouns, and coordinating conjunctions. He adds that an elaborate system of verb désinences, that is endings, or inflected suffixes also signals participant identification. Ibid., 42.

19 See Funk’s sentence types above.

20 McGaughy, Descriptive Analysis, 51-52.

21 Charles H. Kahn, The Verb “Be” in Ancient Greek, The Verb ‘Be’ and its Synonyms: Philosophical and Grammatical Studies, ed. John W.M. Verhaar, vol. 16, 17 vols. (Dordrecht: D. Reidel Publishing Co., 1973), 40-46, 51.

22 Ibid., 39.

23 Kahn does not really provide an answer although he suggests that context and/or the article may provide the key.

24 Kahn, Verb “Be”, 426.

25 Ibid., 428.

26 Eugene Van Ness Goetchius, “Review of Lane C. McGaughy's Toward a Descriptive Analysis of Einai as a Linking Verb in New Testament Greek,” Journal of Biblical Literature 95 (1976): 147-49.

27 Goetchius points out that with “embedded sentences” consisting of the infinitive, the subject tends to be in the accusative case.

28 Acts 18:5 διαμαρτυρόμενος τοῖς ᾿Ιουδαίοις εἶναι τὸν Χριστὸν ᾿Ιησοῦν (“testifying to the Jews that the Christ is/was Jesus”); Acts 18:28 ἐπιδεικνὺς διὰ τῶν γραφῶν εἶναι τὸν Χριστὸν ᾿Ιησοῦν (“showing by the scriptures that the Christ is/was Jesus”); and the verbal ellipsis Acts 5:42 εὐαγγελιζόμενοι τὸν Χριστὸν ᾿Ιησοῦν (“telling the good news that the Messiah [was] Jesus”). The falsifiable hypothesis of the present work yields the same translations Goetchius suggests, but for a different reason.

29 D. A. Carson, “The Purpose of the Fourth Gospel: John 20:31 Reconsidered,” Journal of Biblical Literature 106 (1987): 639-51.

30 Carson, “Purpose,” 642.

31 John 20:31; 1 John 2:22b; 4:15; 5:1, 5c; Acts 5:42; 18:5 and 28.

32 Carson, “Purpose,” 643.

33 Ibid., 643.

34 Carson, “Purpose,” 648-49. See number (9).

35 This study aims to show that the target cluster, when found in convertible propositions, more aptly falls under McGaughy’s Rule 3d than Rule 3c.

36 Jeffrey T. Reed, “The Infinitive with Two Substantival Accusatives: An Ambiguous Construction?” Novum Testamentum 33 (1991): 1-27.

37 Moeller and Kramer made much of the consecutive accusatives characteristic and virtually dismissed non-consecutive examples as too ambiguous. The decision to look at all examples, including those split by the infinitive, advanced the work immensely.

38 Reed, “Ambiguous Construction,” 26.

39 In other words, Reed examined the data in general (combining objects and predicates under one category, objects) semantic terms (subject, copulative, non-copulative, object). The generalization not withstanding, he rightly identified the six functional types (SOI, OSI, SIO, OIS, ISO, and IOS). The present study refers to these as semantic configurations. Reed explained that his study aimed to show that ordering SO outnumbered the ordering OS. In fact, it showed SO:OS ratio = 41:2 for non-copulatives and SO:OS ratio = 46:6 for copulatives.

40 For Reed, “unmarked” refers to the normal word order between subject and object and “marked” refers to the transposition of the two due to contextual factors. The theory is that sentence transcending items will often explain the reason for changing the normal word order. Essentially, three camps exist on this discussion—those who hold to SO being the normal word order, those who hold OS being the normal word order, and those who believe that there is no “normal” word order. For a helpful investigation on this subject see Brad Lee Van Eerden, “An Examination of Some Issues Relating to Greek Word Order and Emphasis” (Th.M. thesis, Dallas Theological Seminary, 1994).

41 On the surface it might seem that five terms would result in 120 combinations (5x4x3x2x1). However, this is misleading because the copulative infinitive only takes a predicate, or stated differently, a non-copulative infinitive only takes an object. Noting that two categories are in view, we see that there are really only 12 permutations (3x2x1 + 3x2x1). Reed’s Copulative table (pg. 26) may very well only represent one category S and P ordering and the Non-copulative table represents S and O ordering.

42 Eerden, “Examination,” 25.

43 The combined total of examples is less than 100. Only 44 of the 52 copulative constructions use the equative verb, εἶναι. And only three involve a proper noun and an articluar noun. Such few examples prove little more than that the construction is rare in the New Testament. Fortunately, Reed is careful to speak more about ratios than about percentages.

44 He spread the net wide (nouns, pronouns, and adjectives) in order to increase the number of samples but, in so doing, he limited the ability to make semantic statements about each of the various combinations of those substantives. Observation of syntactical structures helps to surface the limitations of syntax and show if, indeed, sentence-transcending features ultimately help to identify the subject.

45 D. A. Carson, The Gospel according to John (Grand Rapids: Eerdmans, 1991), 662.

46 Carson still holds to rendering ὅτι ᾿Ιησου) ς ἐστιν ὁ Χριστὸς ὁ υἱὸς τοῦ θεοῦ as, “that the Christ, the Son of God, is Jesus.”

47 Frans Neirynck and Frans van Segbroeck, eds., The Four Gospels 1992: Festschrift für Frans Neirynck, 3 vols., vol. 3 (Leuven: University Press, 1992), 2193-205.

48 Ibid., 2205 fn. 29.

49 Ibid.

50 Matthew Allen Cripe, “An Analysis of Infinitive Clauses Containing both Subject and Object in the Accusative Case in the Greek New Testament” (Th.M. thesis, Dallas Theological Seminary, 1992). This essay contributes immensely to the discussion of the infinitive clause but only those elements which pertain to copulative constructions are highlighted.

51 Ibid., 20-24. Cripe identifies five problems with Moeller and Kramer (MK): (1) it neglects passages where the two accusatives are split by the infinitive, (2) the choice of patterns seems arbitrary and confusing, (3) the grammatical method employed included unclear or ambiguous examples and based conclusions on insufficient data, (4) it presents its results inaccurately, and (5) the statement of the rule is too complex.

52 Cripe, “Analysis,” 30-34. He gives five critiques of Reed: (1) it does not really provide a new rule because it only removes the exceptions disclaimer on MK’s rule, (2) it understands the MK rule exceptions in relationship to the proximity rule rather than in relationship to the word order rule, (3) the grammatical method includes ambiguous examples and slightly prejudices the results, (4) it fails to explain the contradictions to the rule, and (5) the analysis does not separate the copulative infinitive from non-copulative infinitive.

53 This relies heavily on an article by Daniel B. Wallace, “The Semantic and Exegetical Significance of the Object-Complement Construction in the New Testament,” Grace Theological Journal 6 (1985): 91-112.

54 The meaning of this is clarified below.

55 Wallace, Greek Grammar, 43-44.

56 These will be discussed in full in the next chapter.

57 Andrew Keith Malcolm Adam, A Grammar for New Testament Greek (Nashville: Abingdon Press, 1999), 63-64.

58 D. A. Carson, “Syntactical and Text-Critical Observations on John 20:30-31: One More Round on the Purpose of the Fourth Gospel,” (Draft of journal article mailed to Mario Cerda on 21 January 2005, from a paper which was to be presented at the 56th Annual Meeting of the Evangelical Theological Society, San Antonio, TX, 17-19 November 2004), January 2005 TMs (photocopy), Author's personal holding, Trinity Evangelical Divinity School, Deerfield. Carson’s new piece concludes the chronological review of past research. The reader is reminded that my survey focused on discussions relevant to the target cluster. For this reason, some of the reviews may seem minimalistic. Grammars and monographs dealing more squarely with the target cluster or with analogous constructions received more interaction than articles that simply use a syntactical postulation about it in order to advance an argument. Carson provides a fitting example. His articles attempt to support a particular position regarding the historical setting of John’s gospel. Comparatively speaking, his discussions spend little time on original or supporting evidence from new studies dealing with the target cluster.

59 Neirynck and Segbroeck, eds., Four Gospels, 2193-205.

60 Fee, “Definite Article,” 168-83.

61 He claims this on the onset and later reiterates by naming a few scholars that “follow [his] arguments” (Carson, “One More Round,” 4, 13). It seems that by this he means that they agree with him. However, the titles of their books suggest that they contain minimal, if any, new focused study of the target cluster. Cf. Robert Tomson Fortna, The Fourth Gospel and its Predecessor (Philadelphia: Fortress, 1988), 201 fn. 474, 323; Robert Gordon Maccini, Her Witness Is True: Women as Witnesses according to John (Sheffield: Sheffield Academic Press, 1996), 32; Derek Tovey, Narrative Art and Act in the Fourth Gospel (Sheffield: Sheffield Academic Press, 1997), 88.

62 Carson, “One More Round,” 4.

63 Carson, “One More Round,” 4.

64 Ibid., 22.

65 John 4:1, 47; 5:15; 6:24; 7:39; 11:20; 20:14; 20:31; and 21:4.

66 Carson, “One More Round,” 24-25. Carson uses functional terminology to describe McGaughy’s goal but McGaughy is very careful to use truly syntactical considerations (Rules 1 and 2) to determine subject. And when form fails to signal the subject, then sentence transcending considerations (Rule 3) are employed.

67 Ibid., 22-26.

68 Matthew Cripe’s thesis is one of the more recent studies which clearly rely heavily on Wallace’s “pecking order” for the nominative case.

69 Thesaurus Linguae Graecae Ver. #E (Los Angeles: University of California, 1999). See first mention in chapter one for the description of Old Testament (OT) and New Testament (NT) texts it uses.

70 Roy Brown, Accordance Bible Software Ver. 5.1 (Vancouver: OakTree Software, Inc., 2004). This program also uses Alfred Rahlfs’ 9th edition text of the Septuagint but, unlike TLG #E, it utilizes the newer Nestle-Aland 27th edition of the NT text (Alfred Rahlfs, Septuaginta, 9th ed. [Stuttgart: Württemberg Bible Society, 1935 (repr. 1971)], 1-941; Eberhard Nestle, The Greek-English New Testament, ed. Erwin Nestle, rev. and ed. by Kurt Aland and Barbara Aland, 27th ed. [Stuttgart: Deutsche Bibelgesellschaft, 1998], 1-886). Accordance labels the manuscripts Codex Vaticanus and Codex Alexandrinus (BA) as LXX1 and Codex Sinaiticus (S) as LXX2.


Related Topics: Grammar