[MUD-Dev] Re: Prepositions and parsing

coder at ibm.net coder at ibm.net
Sat May 17 14:36:40 New Zealand Standard Time 1997


On 16/05/97 at 11:21 PM, cg at ami-cg.graysage.edmonton.ab.ca (Chris Gray)
said: > [ChrisL:]
>: [ChrisGray:]

>:>In English we disambuate such things by a slight rephrasing:
>:>    shoot the paper which is under the book on the table
>:>    shoot the paper under the book which is on the table

>:Both of these are still ambiguous.  The first could mean to shoot the
>:paper which is located under the book, so that it moves onto the
>:table, or to shoot the paper which is located under the book which is
>:located on the table.	Simiarly for the second one.
>:
>:  shoot the paper which is under the book onto the table
>:
>:Works a bit better, as does
>:
>:  shoot the paper under the book which is on top of the table
>:
>:or
>:
>:  shoot the paper under the book which is atop the table
>:
>:for the other side.
>:
>:>Perhaps handling syntax something like that is a way to go?
>:
>:<kof>
>:
>:Given the average state of western education and general language
>:articulacy, I'm not keen on requiring a gramatically unambiguous
>:parse.

>Grin!

Quite.

I had anough difficulty when I used to teach school (math, Comp Sci)
trying to get kids to realise that the actually had no clue what most of
the words they used everyday in speech actually meant,  Given that, trying
to get them to understand a page of directions or explanstions of a topic
was a lost cause.  They had no chance -- they didn't understand what three
quarters of the mono- and bi-syllabic words on the page meant.

Now you want to add simple grammar rules?

>However, if the intent is to be able to have a textual form of what the
>user does in a graphical world with a mouse or something, any unambiguous
>form will do. 

True.  Its is comparitively easy to map from a graphical entry to an
unambiguous textual command.  However I don't want to drop the texzt
interface, and given that, _and_ the graphical presentation, I am going to
have users trying to enter:

  shoot the paper which is under the book which is on the table

in broken half-understood pidgen english.

>That leaves what to do with ambiguous forms, and maybe just
>not allowing them is enough?

Which is why I can't do this.

>    shoot the paper which is under the book which is on the table
>    shoot the paper under the book which is on top of the table
>    shoot the paper which is under the book onto the top of the table

>There *are* ways! English likely does have truly ambiguous statements,
>but I suspect there are always unambigous forms.

I would argue that there are no ambiguous statements expressable in
english which can't be re-stated in an unambiguous form in the same number
or less sentences, in english.  It does of course take a certain command
and facility with the language however,

Think of it this way:

  A good definition for a word is subsitutable for that word in a sample
sentence with no loss of meaning, context, or stretching of the rules of
grammar.

Now pick something easy like:

  The dog sat on the mat.

and define "sat".  Its certainly do-able, but to watch many try; they're
confronting Einstein.

--
J C Lawrence                               Internet: claw at null.net
----------(*)                              Internet: coder at ibm.net
...Honourary Member of Clan McFud -- Teamer's Avenging Monolith...




More information about the MUD-Dev mailing list