Predicative/attributive - position and function

Iver Larsen alice-iver_larsen at
Mon Jan 22 06:22:11 EST 2001

Dear Moon,

> Dear Iver,
> thanks for the great insight. But I have some problems with your
> explanation as well.
> > Some examples:
> > KALON TO hALAS (Mk 9:50)    KALOS hO NOMOS (1 Tim 1:8)
> > Salt (is) (a) good (thing). Law (is) (a) good (thing).
> > Here we have a NP KALON/KALOS consisting of an adjective and an implied N (a
> > good thing). This NP functions as a predicate in the sentence and is in
> > agreement with the subject. The other NPs (salt, law) function as
> subjects. Each
> > sentence consists of two NPs with the verb ESTIN implied.
> >
> [Moon]
> I have no problem with saying KALON in the above sentence is in a
> predicate position. Supplying the implied noun or verb complicates
> the analysis. To me "I am happy" and "I am a happy person" mean
> different things either in English or Greek. So, KALON in KALON TO hALAS
> IS a predicate to TO hALAS rather than a modifier of the implied noun
> thing.

This analysis would be acceptable to me, too, and it may well be better. I don't
see any substantial disagreement here, just a matter of terminology. I am used
to see a sentence (or proposition) in terms of a VP (Verb Phrase) and zero, one
or more nominal phrases (subject NPs, object or indirect object NPS and various
prepositional NPs. Many languages have descriptive verbs corresponding to
English adjectives. E.g. they would say "she beautifuls" as a verb rather than
"she is beautiful." Whether you take "(be) beautiful" as a VP or "beautiful" as
a predicate, I can live with both.
I have studied a non-European language (Sabaot) extensively. It has VSO word
order, but does not have regular adjectives at all. Sabaot would have to say
either: "Beautiful she" without any copula and the adjective "beautiful" in the
verb (or predicate) position in the sentence. "She" is a personal pronoun in the
subject NP position. If I want to say "a beautiful woman" in that language I
would have to say "woman who beautiful (is)" with a rankshifted verbless
relative clause, i.e. a clause that modifies a head noun in an NP.

Concerning "I am happy" and "I am a happy person" there is often a slight
difference in meaning as to whether a concept is implied and not mentioned, or
whether it is made explicit. "I am happy" I understand as meaning "I am in the
state of being happy", or "I am being happy". "I am a happy person" assumes that
there are different classes of people, happy ones and sad ones. I belong to the
class of happy ones. I did not intend to say that modifying an implied noun is
the same as modifying an explicit noun.

What I am trying to address is the so-called substantive use of adjectives,
particles etc. We could say that when such words which normally modify a head
noun are used substantively, then they are predicates, and that may be the best
option. Or we could focus on the fact that these words are normally used as
modifiers and so could posit an implied N which would fit with the word
"substantive". In English translations from Greek we very often have to supply
these implied nouns from the context or use general nouns that are almost
without semantic content (like "thing" or "one").
> [Iver]
> > Since KALOS is a regular, descriptive adjective, if it had been
> part of an NP
> > with an explicit N it would have been positioned differently:
> > TO KALON hALAS  the good salt (as opposed to bad salt)
> > TO DENDRON TO KALON the good tree (as opposed to another good thing)
> [Moon]
> KALON in TO KALON hALAS is attributive, because it specifies an attribute
> of
> hALAS to determine the referent of hALAS.

I don't have much problem with the word "attributive" although I prefer to call
it "descriptive."

> TO DENDRON TO KALON is really two
> phrases combined into one, as Carl said: "The tree, i.e. the good one".
> Hence
> TO KALON is an apposition to TO DENDRON. KALON is attributive but so with
> respect to the implied noun one/thing, rather than DENDRON.
> So, KALON is attributive in the same way in both examples. It is confusing
> to say that KALON in TO DENDRON TO KALON comes after the modified noun
> DENDRON. In fact, KALON comes before the implied modified noun one/thing.
> attributive with respect to the implied noun thing.

Here I would offer a different analysis. I don't think the repetition of the
article justifies the analysis of this as two NPs, the second in apposition to
the first. It is a possible analysis, but not the only possible one.

Some languages, e.g. Norwegian and some dialects of Danish, have double articles
without implying that this splits up the NP into two NPS. Norwegian members can
correct me if I am wrong, but as far as I know "The nice house" in Norwegian
would be "Det fine huset" (The nice house-the). This developed from the fact
that the article follows the noun when it stands without a modifier: Danish:
"huset" (house-the), but "et hus" (a house) and "det fine hus" (the nice house).

The article is related to specificity. Just like a proper name is more specific
than a common name or title which is more specific than a pronoun (Peter, a man,
he), so the addition of an article moves the word up the so-called "specificity
cline". That is why in most languages that have articles, proper names normally
do not have them. They are already as specific as can be. (Confer the use of the
article, when the reference is in doubt: the Peter I am talking about is not the
Peter you are talking about.) When proper names in Greek have an article, it is
most likely a discourse function related to old or new information or
participant tracking. When a general noun has a number of modifiers added it
becomes more specific. "A nice house" is more specific than just "a house". "The
nice house" is more specific than "a nice house". "The nice, beautiful house on
the top of the mountain" is more specific again.

The double use of the article in Greek could as well be necessitated by the free
word order. Let me explain.
KALON DENDRON means "a GOOD tree" with emphasis on good (contrast to trees that
are not good.)
DENDRON KALON means "a good tree", the default order with no particular
TO KALON DENDRON means "the GOOD tree" (there is no doubt that KALON belongs to
the NP since it is sandwiched between the article and the head noun.)
TO DENDRON KALON does not mean "the good tree", but rather "the tree (be) good"
as in POIHSATE TO DENDRON KALON (Matt 12:33) "Make  a/the tree (be) good".
How, then, could Greek possibly talk about "the good tree" if it is not
sufficient to add the article in front of the head noun? How could the default
order DENDRON KALON be made specific by use of the article? The obvious thing to
do is to repeat the article before the adjective to indicate that in this case
the adjective is part of the NP, so we get
TO DENDRON TO KALON for "the good tree".

In this analysis there is no need to talk about two NPs in apposition. In fact,
that kind of analysis bears the risk of wrong implications. If I say "the tree,
the good one" in English there is some emphasis on the good tree in contrast to
a bad tree. But this kind of emphasis is the opposite of what happens in Greek.
The GOOD tree in Greek would be TO KALON DENDRON, and not TO DENDRON TO KALON.
Furthermore, a construction like "the tree, the good one" is highly unnatural in
English. And to posit so many NPs in apposition in Greek strikes me as being
highly unnatural and therefore an inadequate analysis. (I am not denying the
existence of appositional NPs in Greek, but I am uncomfortable with the
hypothesis that the constructions we are talking about here involve apposition.)

All for now,

Iver Larsen
Kolding, Denmark
alice-iver_larsen at

More information about the B-Greek mailing list