THEORY: branchedness [was Re: Word order]
|From:||Thomas R. Wier <trwier@...>|
|Date:||Tuesday, August 6, 2002, 23:13|
Quoting Barry Garcia <Barry_Garcia@...>:
> CONLANG@LISTSERV.BROWN.EDU writes:
> >So, why is this the case? I've heard people talking about
> >left-branching and right-branching languages, but I don't really
> >understand the theoretical interpretation of this.
> Yeah, neither do i. Can someone explain this as simply as possible,
> perhaps giving examples?
The basic idea derives from a lot of early generative ideas
about phrase structure. All sentences, it is claimed, are composed
of groups of phrases generated by "heads", words which define
phrases' internal structure. So, for example, we could break
down a sentence like (1a) into its constituent phrases in (1b):
(1) a. I like that picture of Sally.
b. [[I]-NP [like [that picture [of Sally]-PP ]-NP ]-VP ]-S
You'll note that the nesting involved here seems to be directional,
that is, when a head word such as "like" takes a complement like
"that picture of Sally", or an NP like "picture" takes "of Sally",
the complement tends always to be the right of (spoken after) the
head which generates it. This apparent directionality is what we
mean by "branchedness", since, like a tree branch, the phrase
structure appears to move uniformly in one direction. Branchedness
is a language specific phenomenon. Georgian, for example, is
a highly left-branching language, since all modifiers appear to
come before the verb, which is usually final:
(2) P'avle jmas c'erils gaugzavnis
Paul.NOM brother.DAT letter.DAT he will send
"Paul will send his brother a letter"
(<j> here is [dz], and <c'> is [ts'])
(3) chemi axali k'ost'umi unda naxot.
my new suit must see-Pl.IMP
"You all must see my new suit."
(4) "In dem von der US-Regierung finanzierten Vortrag
hieß es weiter..." lit. "In the by the US-government
financed speech is it further claimed..." (from _Der
Spiegel_, "Notfalls Saudi-Arabische Ölfelder
(5) That incredibly irritating neighbor of ours...
(Not: neighbor that irritating incredibly of ours)
As you can see in (2) and (3), expressed subjects and objects
come *before* the verb, not after. In the generative tradition,
this has been explained by reference to the branchedness of
complements: in Georgian, as in many other languages, like, say,
Japanese, complements are placed to the left of their head. Even
prepositional phrases may in such left-branching languages be
placed before the noun they modify, as we can see from the
German example in (4). These generalizations will broadly hold
true across languages, although individual languages will obide
by them to varying extents. English, although usually highly
right-branching, places most adjectives before, not after, the
modified noun in (5), and in matrix clauses, German places objects and
prepositional complements after, not before, the head verb.
Branchedness in general explains the fact that most SOV languages
will have postpositions and adjectives and genitives will come
before the modified noun, while most SVO languages will have
prepositions, and adjectives and genitives will follow the
The general idea to take away from all this is that in constructing
your language, you want to think about what the head word is for
any clause you're constructing, and then consistently branch
complements off to the same side in that clause.
Dept. of Linguistics "Nihil magis praestandum est quam ne pecorum ritu
University of Chicago sequamur antecedentium gregem, pergentes non qua
1010 E. 59th Street eundum est, sed qua itur." -- Seneca
Chicago, IL 60637