The Analysis of Determiners (Definite Articles)

Determiners Followed by Noncongruent Vocabulary 8.1

Determiners as Pronouns 8.2

Determiners as Relative Pronouns (Articular Participles) 8.3

In its regular analysis a determiner, or definite article, is not complicated. It may be any of five cases, three genders, and two numbers. We consider ú a particle (qs), not a vocative article. For us the vocative article is the corresponding nominative article when used vocatively. In this case the article is simply dv rather than dn^dv. Determiners serve infrequently as demonstratives (which was their original function), as does é (dnms^apdnm-s) in 1 Corinthians 7.7. In the nominative case, followed by d™ or m™n and when there is no overt substantive, it points to a previously mentioned referent.

Occasionally an article is followed by a noncongruent word, usually of different gender, number, or case. This occurs in three distinct situations or constructions, each deserving comment. The first is exemplified in Mark 12.17: t (danp+) Ka°sarov (n-gm-s). Clearly "things" or some equivalent might be supplied to give the necessary sense, "the things of Caesar" or "Caesar’s things." We chose not to indicate this in the tags, neither in the determiner tag as danp^danp&n-an-p nor in the noun tag as n-an-p&n-gm-s. The plus in the determiner tag indicates the missing substantive.

In the second construction the article is followed by a phrase. This is usually a prepositional phrase, as in Mark 13.16: é e¸v tèn ‡grèn. Here the article é is simply marked dnms+; we do not represent "man" or "one" in the determiner tag or anywhere else.

In the third construction the article is followed by a single word, usually an adverb. For example, tè (dans+) žswqen (ab) in Luke 11.40. Again the plus indicates that there is no overt substantive. With respect to this third construction type, we note that adverbs receive functional analysis as substantival adjectives in our system only when they are anarthrous, that is, when there is no determiner tag on which to place the plus. The two main instances of this are preposition-followed-by-adverb constructions (e.g. œwv ˆrti, for which see section 7.11 above) and anarthrous pljs°on (e.g. Luke 10.29), where "neighbor" and not "nearby" is contexually required.

Determiners can also be used like pronouns. This function is limited to nominative-case determiners and must be followed by m™n or d™. It is a narrative device to reintroduce a participant into the role of actor (hence the nominative case). These determiners, when functioning like pronouns (e.g. é, dnms+) and followed by participles, look very much like articular participles, introduced in 8.3 below. A determiner functioning like a pronoun serves to reintroduce someone who has already been identified; an articular participle, by means of the participle, serves to point out someone. hO dš ‡koÀsav eºpen (Matthew 9.12) is ambiguous apart from context. It can mean "But when he heard (this), he said…" or "The one who heard (this) said…." In the first case the speaker is a definite individual identified earlier in the context. In the second, the speaker is being introduced, for the first time, at this point. Both are similarly tagged: the former is marked with a determiner tag and following plus, the latter with similar conventions introduced next.

Articular participles, like relative clauses, are a grammatical device for relating two clauses through a noun. Take, for example, this sentence: é ‡gapòn tèn ‡delfèn aÇto n tþ fwt± m™nei (1 John 2.10). It has two verbs and therefore two clauses that need to be related. The main verb is m™nei. It makes a statement so that the main clause reads, "(someone) remains in the light." The articular participle serves to identify that someone: "he who loves his brother." The tags we give to the words in this sentence are all predictable except for the tag for the first article, in which we relate the clauses: dnms^npnm3s&aprnm-s. This working complex-tag analysis is to be read: the article functions like a noun substitute (the antecedent, if we may say so) and a relative pronoun. The chief difference between this derived relative pronoun and a real one is that the former takes a participle as its verb form, the latter a finite verb. The simplified tag for the determiner é is dnms+, with the plus pointing out the missing substantive.

Approximately sixty percent of the articular participles in the Greek New Testament are of the kind just presented, with the semantic antecedent supplied in the tag. Though the overwhelming majority of them are nominative case, they can be any of the five cases. For example: é qewròn mš qewre² tèn pšmyant€ me (John 12.45). Though this sentence has two articular participles, we are interested here only in the second, which is accusative. "The one seeing me sees" someone. Who is that "someone"? "It is the one who sent me." The working analysis of tèn is dams^npam3s&aprnm-s, which means that the article functions like a noun substitute (the object of qewre²) and a relative pronoun (the subject of p™myant€). It is very interesting that derived relative pronouns always act like the subject of the following participle, for which reason they receive a nominative-case tag, here aprnm-s. Whereas a real relative pronoun may stand in any relationship to the verb of the subordinated clause, an article followed by a participle may only function like the participle’s subject. If the participle is passive, then the article used as a relative is still that participle’s grammatical subject. The simplified tag for tèn is again dams+.

Another thirty-five percent of the articular participles have their own antecedents preceding them in the Greek text. For example, in this sentence, … tÒv c€ritov to qeo tÒv (dgfs^aprnf-s) doqe°sjv moi … (Ephesians 3.7), "grace" is the antecedent. Because the antecedent is overt, the repeated genitive feminine article receives the working analysis of an article used as a relative. (The simplified tag is merely that of the determiner, without any plus, for the antecedent is overt in the surface structure.) Notice again that the case of the functional relative is nominative, the subject of the passive participle. This example is normal in that the case of the repeated article is the same as that of its antecedent. The case need not be the same, however, as numerous instances in Revelation confirm. An instance of a working analysis from Colossians might be more convincing: ‡pè HEpafr‚ (n-gm-s) … é (dnms^aprnm-s) ka± djlðsav (1.7-8). (The intervening relative clause might have conditioned the case of the article. Notice, incidentally, the two comments about Epaphras, one in a real relative clause, the other in a functional one.) Articular-participle derived relative clauses may also have pronouns as their antecedents: aÇtÞ tÞ kaloum™nÛ ste°r‹ (Luke 1.36).

Another similarity with relative clauses is that articular participles may have their antecedents following (rather than preceding) them. Constructions of this type constitute the remaining five percent. Most readers will quickly recognize this as a case of the participle being used as an attributive adjective, that is, article–participle-as-adjective–noun. At the very least this is a relative construction when viewed semantically. And there are also grammatical indications that it is. For example, several words that can fill the position of the noun would not qualify if the participle were replaced with an adjective. Among these are mo° (tþ q™lonti mo±, Romans 7.21) and toÂto (tè gegramm™non toÂto, Luke 20.17). In all these cases of a following substantive, we have given the article a working functional tag with a plus sign. Thus tè in Luke 20.17 has the working-analysis tag of dnns^aprnn-s+. One recurring instance of a following noun is the correlative-like construction in which the identity of a person or thing is expressed in an articular participle, which in a following reflex of oÆtov or a similar demonstrative is made to join a main clause. For example, é (dnms^aprsm-s+) pisteÀwn e¸v mš … k‡ke²nov poiÒsei (John 14.12). The simplified tag for this construction is merely the tag of the determiner with a plus, dnms+. This construction parallels that of 7.6.2, subclass f^3.

In analyzing articular participles the way we do, we are making no claims about how they should be translated. Our only claim is that semantically these constructions parallel real relative constructions. We have so analyzed all articular participles, no matter how reduced they are; for example, to qeo to (dgms^aprnm-s) zòntov (Matthew 26.63). (The simplified tag is clearly dgms.) Let us look at three nearly identical constructions and the implications they raise. Our working analysis of Hebrews 10.34 reads: tòn (dgnp^npgn3p&aprnn-p) Ãparcçntwn Ãmòn. "Your possessions" is a translation that would probably be widely accepted, and yet our analysis seems to force the translation, "the things that exist of yours." hUmòn, rather than Ãm²n, follows the participle, and this seems to tip the scales toward taking the participle as a substantive and forgetting any relative construction. (See Luke 12.1, however, where the antecedent is possessed by a phrase, tòn Farisa°wn, that is interrupted by a real relative clause.) Second, Luke 19.8 is similar, but with the possessor preceding the participial construction: mou tòn (dgnp^npgn3p&aprnn-p) Ãparcçntwn. Finally, Luke 8.3 gives a more convincing functional relative construction, with a dative pronoun replacing the genitive: tòn (dgnp^npgn3p&aprnn-p) Ãparcçntwn aÇta²v. Here the possessor is within the participial construction, as seen in clauses with finite Ãp€rcw. The examples we have just inspected show the range and variability of these constructions, being possessed within or without, and relating to the participle as verb or to the antecedent implicit within the article. One must be alert to these articular participles, remembering that our working analysis is based on semantic function, not grammatical form.

Observe the three following constructions: (1) p‚v é ˆnqrwpov poie², (2) p‚v ëv poie², and (3) p‚v é poiòn. In the first, p‚v is tagged a--nm-s without controversy; in the second, p‚v is ap-nm-s as substantival to following relative clause. How should it be tagged in the third? It might be tagged either a--nm-s, because this construction is parallel with the first (quantifier plus determiner), or ap-nm-s, because it is parallel with the second (real and functional relatives, respectively). This latter analysis is possible, and in keeping with it our working relative analysis of adjoining articular participles would then be dnms^aprnm-s, understood as representing the underlying structure "everyone (p‚v) who." We have chosen, however, to analyze it as a--nm-s. Here the working analysis suggests the determiner tag as dnms^npnm3s&aprnm-s with the combined quantifier-determiner "translation" as "every [supplied antecedent] who"). The actual simplified tag for the determiner in the third construction is dnms+, pointing out the missing antecedent.

In Luke 1.35 and Matthew 2.2 we had to decide whether the articular participle contains in the article the antecedent to the construction and „gion and basileÁv, respectively, are complements to the participles; or whether these last named are the (following) semantic antecedents such constructions require. Our usual rule of thumb is to take kal™w and l™gw, and especially passive instances, as requiring a complement and so, where an antecedent is lacking, to supply it in the tag. In both Luke 1.35 and Matthew 2.2, we decided in favor of the first possibility. Other cases are analyzed individually.

Articular participles, like real relative clauses, can be left hanging. See, for example, Hebrews 1.7, where é (dnms^npnm3s&aprnm-s, as working analysis, and dnms+, the simplified tag) poiòn has no main clause to which to relate. In the original context for this phrase (Psalm 104), nothing is left hanging.

In several places our relative analysis of articular participles runs into apparent trouble: 1Timothy 4.3 and Titus 1.15. In these passages a single article governs a set of one adjective and one participle joined by ka±. The problem is that for articular participles (and prepositional phrases) we indicate an unexpressed substantive by a plus on the article, whereas for adjectives used substantivally, the designation is carried by ap. What shall we do, for example, with 1Timothy 4.3? Shall it be ddmp and ap-dm-p to satisfy the adjective construction or ddmp+ and vpradm-p for the articular participle construction? One thing is clear and that is that semantically only one participant set is in view. To this end we have labeled the determiner with a plus, ddmp+, in both references, indicating a unified substantive of two characteristics, while the adjective is analyzed as a-. (Romans 2.8, another conjoined articular construction, is not problematic, for the items joined (prepostional phrase and participle) each individually take the ddmp+ analysis tag.)

These examples raise the general question: Do not adjectives work the same way articular participles work? And if they do, should they not receive similar treatment? At the very deepest, most abstract level of language, adjectives are viewed as parts of relative clauses. "The happy child" is viewed as "the child who is happy." From this same viewpoint the relative and the verb "to be" are lost and the adjective is transposed into attributive position. How this might work in practice is not our concern. It is enough to note that copula verbs are often missing in Greek; other verbs are missing much less often. This accords with what we find concerning adjectives and articular participles. Adjectives in attributive position can be viewed as abstract relative clauses with e¸m° or even as articular-participle derived relatives with øn. The verb of being is lost and an adjective results. When the copula is not deleted, we have either a true relative clause with e¸m° (1 John 2.8) or an articular participle with øn (2 Corinthians 11.31). (Note, incidentally, that these immediately foregoing examples have some adjunct information. For example, "… true in him." A lone adjective, it seems, must lose its relative-clause trappings. They may be retained with adjunct material or with an indication of time other than present. For example, see John 9.24.) Real and derived relative clauses with verbs other than e¸m° cannot have their verbs deleted without losing some element of their meaning. Thus their verbs are retained. Therefore, we hold that there is a difference between attributive adjectives and articular participles that warrants different treatment.

Articular participles can be first- or second-person constructions in the same way as real relatives can. When a first- or second-person personal pronoun is on one side of an equivalence statement (e¸m°) and an articular participle is on the other, we have extended the first or second person of the pronoun across the equivalence to the derived relative construction. Thus in John 6.51 Jesus claims to be the living bread that came down from heaven. This analysis is given in the case of either claimed or denied identity, but not of questioned identity ("Are you…?"). The reason why such an articular participle can be marked 1 or 2 (on the participle) when complement (predicate) to a personal pronoun and a form of e¸m° is seen in John 8.18. The reflexive pronoun mautoÂ, rather than moÂ, gives strong evidence that é marturòn should be considered first person. In many of these constructions, it is as if the first- or second-person reflex of e¸m° should be read "It is I/you" and the articular participle is as if a simple functioning relative with the overt personal pronoun as antecedent.

As with providing antecedents for true relatives that involve e¸m°, so with the so-called functional relatives: one must ask whether the writer is predicating equivalence or existence. In Galatians 1.7 Paul predicates only existence. He is not saying that "some are the ones who…" or that "the ones who … are some." Rather he is saying that some ones exist; the articular participle identifies the "some ones." Because the antecedent is overt, the article is tagged in the working analysis as dnmp^aprnm-p, and dnmp in simplified form. In Mark 4.16 Jesus asserts equivalence rather than existence: "These are equivalent to the ones who…." Here the working analysis of the construction is dnmp^npnm3p&aprnm-p, or dnmp+ in simplified form, because no antecedent is available. In those cases where either existence or equivalence is possible, we have picked one based on our judgment of the discourse requirements.

Our analysis of derived relative pronouns stops with participles that have the definite article. Many participles have no governing article, and these too must bear some relation to finite verbs. We have not analyzed these. Some, even though they lack an article, may be related as semantic relatives to the main verb. Many of these are not related to the main verb through the noun, but bear to the verb instead an adverbial relationship. These remain untouched except for the analysis of the form itself.