From: Mike Sangrey (firstname.lastname@example.org)
Date: Tue Jun 20 2000 - 09:03:18 EDT
> There are artificial languages (e.g. machine languages) which can be
> "understood" by reading one token at at time in one pass reading left
> to right. But I think this approach will lead you to a dead end when
> applied to a natural language. You can work left to right but you will
> need to make multiple passes. This is a topic I haven't thought much
> about in the last decade. I gave it a lot of thought in the mid 80's
> but that was a long time ago.
I've written simple artificial language compilers, so I know what you mean.
To further your comment, a lot of careful thought goes into the design of
an artificial language to make sure it is "one token look ahead". As you
would know, it makes the parser SO much simpler. Obviously, natural
language is not that way.
So, thanks for pointing this out; being true it gives me the chance to
clarify something which may not have been clear. I did not mean to imply
that one could determine the meaning of a larger semantic unit AT EACH
WORD UNIT ALONG THE WAY. As one reads (or hears) one makes choices
along the way; or one holds a very small number of choices in one's
head waiting for a clue that will clarify which choice would have been
the correct one; or one determines which choices were the wrong ones.
This sets up a rather dynamic mechanism for determining the meaning.
Therefore, in a sense, from my perspective, it does take multiple
passes; yet, this too is not machine like. It's more like a reader
frequently reaching back into earlier parts of the larger constituents
and adjusting his/her thinking. (Somewhat like single pass compliers
resolving forward references, but much more complex, more fluid.)
I also tend to think there are certain phrases and types of constructions
which are caught by the mind's ear as a unit. For example, "in my not
so humble opinion" approaches so closely to an atomic constituent that
in email the author abbreviates it as 'IMNSHO.'
Overall it is very difficult to get at this since the mind works so
quickly in building the meaning. It is very hard to tell what part
of the meaning was obtained from the word and what part was obtained
from the complex of relations. However, when working with a language
one is NOT fluent in, the mistake is to approach it with such rigor,
even machine like, that one misses the naturalness the author used in
constructing his/her idea. Take a 20 word Greek sentence with the verb at
the end. IMCO (In My Current Opinion), it is the START of a mistake to
take that whole sentence, put it on the dining room table and construct
the relations from the whole. The mistake is made then when one builds
the meaning of the sentence from the diagram working from the verb out.
The author did not do it that way; the reader shouldn't either.
Actually, one perhaps should do diagramming--I certainly have--in order
to get the relations in one's head. But then, I think it best to force
oneself to move left to right. Having the relations ready at hand one
can more fluently move through the sentence allowing the author to have
his/her way with the meaning.
I whole heartily agree with you that one absolutely must expose the
relations between every level of constituent. Fluency does this for us
automatically; not having fluency one must rely on rigorous application
of specific techniques. My caution is that the application of these
artificial techniques causes their own problems; and the remedy is to
force ourselves to be as natural as we possibly can (whatever natural
> However when a biblical language neophyte picks of a "word study
> dictionary" and launches off to do some "exegesis" I don't think that
> their primary focus is going to be the analysis of the relationships
> between constituents at all levels of the discourse. They are just
> going to dig in and study some words. Meanwhile the meaning of the
> TEXT will continue to remain and enigma to them since it will never be
> discovered in this manner.
I think this is absolutely right.
Thanks Clay, and others.
Every Christian library should have a plaque which states:
"There is one book which explains all of these."
B-Greek home page: http://metalab.unc.edu/bgreek
You are currently subscribed to b-greek as: [email@example.com]
To unsubscribe, forward this message to leave-b-greek-327Q@franklin.oit.unc.edu
To subscribe, send a message to firstname.lastname@example.org
This archive was generated by hypermail 2.1.4 : Sat Apr 20 2002 - 15:36:29 EDT