Postpositives

Most GNT students think they know about postpositives like δέ and γάρ and τέ because they have learned about them in their postmodern grammars and can give some examples. Most current GNT grammars are postmodern. Postmodern means that concepts are discussed without giving operational definitions. The big advantage of omitting operational definitions is that any author can make up claims about terms without research and cannot be pinned down or proven wrong. However, a real definition must clearly provide criteria that include all members and exclude all nonmembers.

Postmodern grammars tell us that “A postpositive is positioned (posited) after (post) other words in the clause” (https://tinyurl.com/ybrzbp34). This is not even a useful discriminator, let alone a definition. It is a preposterous statement since every word in the clause except the first one is posited after other words, but most of them are not postpositives. The majority of words in the dictionary do not come first in any sentence or clause in the GNT but are not postpositives.

The other defect in this statement of concept is that it fails to tell the learner that the postpositive must come second or near second in some expression. Hopefully, the reader’s introduction to postpositives involved some idea of second. However, second has no meaning if we do not know second in what. The advantage of omitting this crucial information is that the author does not need to prove how early in the unit. For example, in Acts 20:21 the postpositive τέ comes 62nd in its clause. So we know that the non-testing academic author of this “definition” has chosen the wrong unit (a clause), even in his inadequate descriptions of postpositives.

Here is another postpositive internet attempt at explaining postpositives without defining them.

Postpositive: A word that cannot be placed in the first position within the sentence, but is translated first: γάρ, γέ, δέ. (https://tinyurl.com/y7kkdbcg)

As is usual in postmodern thinking, an inadequate list of examples is given instead of a definition. If there are ten or twenty of them, how hard is it to give the complete list? The “definition” does provide more of a test than the previous one, but we suspect that Greeks understand postpositives without reference to translation. Order in target languages is too flexible to be a valid test. The target language order test is valid for and and for but not for therefore and somehow. A definition that is sometimes true and sometimes false is postmodern. Why would people even produce such sloppiness, let alone be satisfied with it? The sentence unit is even worse that the clause unit because some postpositives come near the end of sentences rather than the beginning.

Stanley Porter gives a better postmodern description of postpositives.

1.3.2 Postpositive words. The following words tend not to begin a clause (or phrase) in NT Greek, but usually appear in the second structural position: ἄν, γάρ, γέ, μέν, οὗν, and enclitics such as ποτἐ, πῶς, ποῦ, and a number of pronouns με, μου, μοι, τέ. These are often called postpositive words. Postpositive means that the word occurs anywhere from right after the first word in a syntactical unit. (https://tinyurl.com/yc5kvus5)

Clearly, the list is vague and open-ended and uses weasel words like tend, which destroy the definitional criteria and obscure meaning. The unit is wrong, as shown above. The word ἄν acts like a postpositive when used to mean ever or even, as in Luke 9:47, but not when it means if, as in John 16:23. So ἄν is not a postpositive lexeme per se. It fails the definitional test.

There is a natural confusion between postpositives and enclitics, which both follow other words in some sense. However postpositive is a syntactical concept defining phrases while enclitic is a diacritical concept defining accents. μου is enclitic, but it is doubtful that μου is a postpositive, as Porter and Robertson claim, since it often precedes its head or follows it. One of the telltale signs of postmodern grammarians is that their example lists are open-ended and do not match each other and their “definitions” contain insufficient criteria to prove which list is right.

Enclitics are inflections, not always whole lexemes, since some inflections of ἐγὠ like με, μου, μοι are enclitic but ἐγὠ itself is not. A valid definition of postpositives would need to specify whether it refers to lexemes or inflections. However, we see no proof that any lexemes have both postpositive and non-postpositive inflections, as Porter suggests.

Now that we have seen that definitions of postpositive in dictionaries and grammars are generally circular, vague, non-existent, non-committal or lacking either operationally inclusive or exclusive criteria, we can turn to crafting a reasonable definition of postpositive that accords with tests we can devise on our NT syntax-coded sample. The difference between vagueness and precision is the availability of syntax-coded statistical tools.

Considering a postpositive as generally a second (not first) word in a phrase, we immediately encounter the need to define the kind of phrase. As we have seen above, clause or sentence are completely inadequate for this purpose, since postpositives can occur very late in these units. The phrase must be both smaller and well defined. It cannot be some vague expression; otherwise there is no operational test.

Also, clearly the phrase cannot be the structure consisting of the postpositive itself and its subordinates, since that is not the phrase a postpositive is regularly second in. The phrase that the postpositive structure is regularly second in is its matrix. The matrix is the structure of the head of the postpositive. The matrix includes the postpositive, its head, siblings and all their subordinates, i.e. its head and all subordinates. For example, the matrix of the direct object is the verb phrase (VP) of the verb whose direct object it is.

In the syntax diagram, the eye goes to the head and includes all its subordinates below. This set of words is always unambiguously defined in GNTCG. The correctness of the choice of the matrix as the phrase of the postpositive is confirmed by the fact that most of the most popular examples of postpositives never come first in the matrix but always come early. The simple concept of the matrix, basic to a hierarchical grammar like GNTCG, is unknown to many conventional grammarians.

The next definitional decision concerns the word first. Should words like ἄν that usually come second but sometimes come first be called postpositives? This is a point that is very hard for postmoderns to decide on. Most postmoderns will squirm at the idea of discrimination (even though all definitions involve discrimination) and will prefer the familiar Orwellian environment of definitional slipperiness. However, GNTCG is not postmodern. It is for people who prefer clearly defined terms that provide operational criteria. Accordingly, the GNTGC provides the definition: a postpositive is a conjunction or particle lexeme that does not occur in the initial position of its matrix but usually in the second position or sometimes in the third or fourth position or rarely in the fifth or sixth position.

Good examples of the cutoff are the three enclitic particles ποτέ, πώς and πού that are often considered postpositives. πού and πώς never occur first in their matrix, but ποτέ occasionally does (for example, Luke 22:32—καὶ σύ ποτε ἐπιστρέψας στήρισον τοὺς ἀδελφούς σου and you, once you turn back, strengthen your brothers). Here, ποτε initiates the circumstantial participial phrase ποτε ἐπιστρέψας (likewise for Col. 1:21).

The fact is that the majority of particles (other than negatives and positives) never initiate their matrix, but we do not call them postpositives because they to not cluster in the second position. It is possible to form a general rule that is true most of the time, but that would be arbitrary. The simple absolute rule is better. It is better to say that ποτέ is virtually a postpositive (likewise for ἄν). If definitions have any meaning, the line needs to be drawn. 

Postpositives usually precede their head but may follow. The inflection μου is disqualified from being a postpositive because it often comes first in its matrix and does not gravitate toward second place, even if it never comes first in its clause or sentence. Clauses and sentences are invalid units/phrases for postpositives because they are not the units in which the postpositives cluster in second position. The natural matrix concept of hierarchical GNTCG finally provides an opportunity for grammarians to define what they have long been struggling (but unable) to define accurately.

Matrices are a natural part of hierarchical syntax in a word grammar like GNTGC but not traditional units of GNT grammar. The fact that they provide a framework for both expressing the definition of postpositives and the definitional criteria for validating the adherence of the members to the definition proves that such hierarchy, dependent largely on conjugation, class and declension, is a natural and inherent property of the original language and not an arbitrary construct imposed on it by the grammarian.

Here is a table of the positions and frequency of the ten postpositives in GNTGC in the coded sample.

Next, we turn to the question of how many and what sort of words may precede the postpositive in the matrix, i.e. be forerunners. There are 1336 postpositives in our sample, and all but 84 are in position two of the matrix. The 84 have the postpositive between positions 3 through 6. Nineteen of the 84 have the head as a forerunner and so have the whole set of forerunners as one structure. The other 65 with no head forerunner all have two or three forerunners.

The majority of forerunners are markers like articles, conjunctions, relative pronouns, prepositions, interrogatives and other postpositives, that all strive to occur near an initial position in structures. Other than markers, each of the 84 postpositives has at most one forerunner sibling that may occasionally have one dependent preceding the postpositive. So, the structure of compound forerunners is still simple, consisting of one head or one sibling structure of the postpositive and/or one or two markers. Markers have no dependents, but whether the dependent of the head or sibling forerunner is placed after or occasionally before the postpositive appears to be an optional stylistic variation.

The precise discoveries of the nature of the forerunners allows us to restate the definition of the postpositive: A postpositive is a word that occurs in the second through sixth position of its matrix and is only preceded by markers and/or the head or one sibling and occasionally by one or two of their dependents.

Matthew 22:20 has the postpositive τε in the fifth place, after the head: πάντας οὔς εὖρν, πονηρούς τε και ἀγαθούς all that they found, both good and evil. Matthew 2:6 has the postpositive in the third position after one sibling with a dependent: ἐκ σοῦ γὰρ for from you. Matthew 12:33 has the same sibling preposition forerunner, but the dependent follows the postpositive, which is in second position: ἐκ γὰρ τοῦ καρποῦ for from the fruit.

In addition to being a lexical term for particles and conjunctions, postpositive may occasionally also be a syntactical term for prepositions. Normally in Koine, proper prepositions precede the object, although in Attic they could sometimes follow. However ἐναντίος before may be postpositive to the object as in 1 Corinthians 2:15 or “prepositive” as in Luke 1:6.

—Dennis Kenaga, April 2018