Re: Right-Branching vs. Left-Branching

On Wed, 17 Sep 2003 13:55:25 -0700, Heather Fleming <hfleming@PLANET-
SAVE.COM> wrote:

>Aha. Now it comes back to me. If you go by number of individual languages,
>SVO comes out ahead. But if you go by language GROUPS then SOV wins by a
>long shot.
What source(s) do you have for that?  Also, how many non-IE SVO languages
were originally SOV or something else, and then became SVO through IE
influence?

>Did a search and found it near the bottom of the article. Yep, that's all
>he says. "Right-branching languages are inherently easier to parse for
>both computers and humans."
>
>Lacking any kind of proof, I would think that this is a purely personal
>viewpoint. It may be accurate for computers, but probably because computer
>programming languages are mostly invented by English speakers or at least
>speakers of primarily right-branching languages (I'm guessing), and read
>left to right.
I think that syntax may be the hardest thing for one to change when
learning another language with different syntax, probably because it's a
pattern.  Look at Latin and German -- they're considered "difficult" to
learn by native English-speakers because of their canonical SOV word-order.

There's a good quote by Winfred P. Lehmann: "Languages, in spite of their
outward differences, are formed by identical principles" (Syntactic
Typology, 1978).  I think this rings true.  When looking at the larger
picture, people generally have the same mental structure (if they didn't,
language would never have arisen).  Different syntax patterns, grammatical
structures, etc. are just different ways of essentially communicating the
same thing.

Some people (mainly auxlangers) believe that purely isolating languages,
especially pidgins and creoles, are the "best."  However, the evidence
shows that such languages generally arise only through contact between
speakers of disparate languages, who are not acquainted with each other's
syntax, inflections, etc.  So a pidgin or creole is the solution: an
extremely basic language that is initially used for communication
between/among speakers of two or more different languages.  Languages that
have generally been isolated from extensive contact with others tend to
show many more inflections and other traits of synthetic grammar than ones
that have had such contact.

Anyway, this whole thread (and the one I started about adpositional heads)
was started because I am attempting to create a "logical language" -- i.e.,
a language that is based more on formal logic than any natural language
(that I know of).  I started out thinking that when dealing with numbers,
the most significant digit is the one furthest to the left (i.e., the digit
in the largest place-holder).  So I concluded that in this language that
I'm creating, the most important information will be furthest to the left.
I then thought that, generally speaking, the verb would usually be the most
important piece of information -- more important than its subject(s) or
object(s), because it (according to predicate logic anyways) describes a
relationship between the two.  In this way, I concluded that canonical VSO
syntax was appropriate for my "logical language."

However, things got stickier when dealing with adpositions.  In my opinion,
there are two ways of looking at them: either they describe relationships
with one or more nouns, or they are simply modifiers of single nouns.
Modern theories of grammar assume that adpositions are "heads"; that is,
they delimit a phrase.  There seems to be some evidence of this in natural
languages, but I think there is also evidence to the contrary.  In some SOV
languages, postpositions are more or less clearly discernible as coming
from nouns, where the "object of the postposition" is in the genitive
case.  The best example I can think of is Finnish: talon edessä "in front
of (the) house" literally means "(the) house's [front]-in."  Here we can
discern that there was originally a word *eti (or something similar),
having a meaning (analogous to) "front," which then took an inessive case
ending -ssA.  I also have a (probably) unique theory about changes in
syntax during the history of Proto-Indo-European, specifically about the
change from postpositions to prepositions.  Anyways, it appears that
different languages treat what we normally call "adpositions" in different
ways, and that there doesn't always seem to be a clear-cut location of
the "head" in adpositional phrases.

>About the term "right-branching" anyway. Just because WE write (and draw
>tree diagrams) from left to right doesn't mean "right-branching" is a
>universally intuitive term. "Head-initial" or "head-final" would be more
>accurate.
I agree.  Thus, by "right-branching" I meant "head-initial."  But as I said
before, it's not always easy to point out the heads!

>One of my favourite profs was schooled in Chomskyan linguistics and then
>left because it was so far removed from the real world and so full of
>itself.
>
>Heather
I would wager that that's an accurate description. :)

- Rob

Reply