Re: A Normalization Question

"Ralph Becket" <rafe_at_cs.mu.oz.au> wrote in message news:3638acfd.0407221716.387e1eb_at_posting.google.com...
> neo55592_at_hotmail.com (Neo) wrote in message
news:<4b45d3ad.0407211939.47e9bebd_at_posting.google.com>...
> > In the above (doctored), the schema (which is also data) has a
> > redunancy. Both tables are representing Subject/Verb/Object
> > relationships. They should be in one table.
>
> You're right:
>
> event name subject verb object
> ------------------------------------------
> thingA john takes final
> thingB john receives 95
> thingC thingA causes thingB
>
> > In the above (doctored), the schema (which is also data) has a
> > redunancy.
>
> Are you aware of the term "referential transparency"? Do you
> really believe that
>
> 2 + 2
>
> and
>
> let x = 2 in x + x
>
> *denote* different things? (Note, I am *not* asking about
> syntactic structure.)
>
> > Where/how would one store the following:
> > ((john take final) cause (john receive 95)) VerbX ObjectY.
>
> (I have no idea what VerbX and ObjectY mean here.)
>
> If you're asking about how we describe relationships with
> more than two arguments, all we have to do is have one table
> with n + 1 columns for each relationship of arity n, then for
> each relationship instance put the relationship name
> in the first column and its corresponding
> arguments in the subsequent columns. Then we need to agree on
> the meanings of names. As a consequence we need a means of
> disambiguating semantically different things that coincidentally
> happen to have the same name in English, such as the name "Brown"
> and the colour "Brown". Two ways of doing this are (1) context: we
> know from the type of a given argument position of a given
> relation whether "Brown" must denote a name or a colour or a
> particular spelling and (2) by simply giving them different
> symbolic names that just happen to elsewhere be realated to the
> same spelling.
>
> -- Ralph

Hi,

Out of curiosity, how would one deail with different types of semantics across concepts or entities, such as role slots, in these relationships. In some cases they may be bidirectional. In others, directed. I mean, we have relata such as hypononyms, hyperonyms, troponyms, mereonyms, metaphors, subsumptive relationships, etc. They all have certain characteristics that might not be sufficiently general.

Regards,

Dan Received on Fri Jul 23 2004 - 07:01:48 CEST