Re: what is a loglang?

Hey, that's quite interesting. I'm trying to do
something similar in some aspects, but I was trying to
get to a higher level in abstraction.

My idea was trying to express the whole knowledge
using a very simple meta-syntax, essentially:
<x R y>
meaning: there is a relation R between x and y

The first time you mention a concept x, you have to
use a class definition, like:
<lead is-kind-of metal>, where:
"lead" is the new concept you're introducing
"is-kind-of" is a relational concept you already
defined before (in fact, it is the privileged
relational concept)
"metal" is a "nominal" concept you already defined
before (I say "nominal" because I have no better word
at hand, but of course it doesn't refer to any
syntactical natlang or conlang category, it is purely
conceptual)

To this basic definition you add specific features,
like:
<lead has-density "11.3"(dens-quant1)>, where:
'lead' is the same as before
'has-density' is a relational concept you already
defined before
'"11.3"(dens-quant1)' is a compound concept of
measurement, required by the definition of
'has-density' . It consists of the ordered association
of a constant value "11.3" and the concept
"dens-quant1", defined before (essentially dens-scale1
consists of the association of a specific scale, for
ex: "mini=0, maxi=22, continuous" and a measurement
unit, like "kg/dm3". The interest of such a scale
could be that all possible elements on Earth would be
included between the mini and the maxi (22 for
iridium, I believe), so the value "11.3" would have
some meaning of its own. In other cases (the inside of
a star, for ex), you would use of course a different
scale and/or measurement unit.

Joining both assertions would be done by an operator,
like "and", "such-as", "constraint" (list under
construction). So the definition would become:
<lead is-kind-of metal such-as lead has-density
"11.3"(dens-quant1)>
which follows the syntax:
<x R y op x R y>
and you could of course add more features in the same
way, even if basically only one <x is-kind-of y> is
required (because you don't have necessarily all
information at hand). For ex the atomic structure of
lead, of course, or its usages, or other physical
attributes...

If we looked at the definition of 'has-density' :
<has-density is-kind-of
physical-quantifiable-attribute such-as has-density
is-quantifiable-by dens-quant>
(the former mentioned dens-quant1 being a kind-of
dens-quant, etc)

Then:
<physical-quantifiable-attribute is-kind-of
physical-attribute and physical-quantifiable-attribute
is-subject-to quantification>

Then:
<quantification is-kind-of logical-relation and
quantification is-ordered-association
(value,scale_unit)>
(here we see that the last y-term consists of a list
of 2 ordered terms, of course to be defined earlier;
scale_unit being defined as an ordered association of
scale + unit)

<physical-attribute is-kind-of attribute constraint @y
is-kind-of physical-concrete-concept>
(here the @y restriction means that, when using the
'physical-attribute' relation, you must use a y-term
earlier defined (transitively or directly) as a
physical-concrete-concept, like a substance (lead,
stone, milk...) or a physical entity (a stone, a
chair, a human body...)
(if you noticed that I mentioned 'stone' both as a
substance and as an entity: there would be two entries
in the database, namely stone#1 and stone#2, the first
one defined as a substance, the secund one as an
entity: the fact that English uses the same word for
both should not confuse us. I'm trying to avoid the
'or' in the definitions: I hate ambiguity).

Also:
<physical-entity is-kind-of physical-concept and
physical-entity is-kind-of entity>
(here we have the intersection of two concepts of
higher level, producing a new compound concept)

At the highest possible level, you would find the
universal root:
<concept is-prime concept>, meaning that you won't go
any further.

All this of course under construct, and probably
forever.

I have some remarks about you "Metaclass Descriptions"
(3.2.3), I'll try to send in another message.


--- "Mark P. Line" <mark@...> wrote:
>
> There's a recent draft of the upper ontology I have
> in mind at:
>
>     http://www.polymathix.com/papers/socs-upper.html
>
=====
Philippe Caquant

"High thoughts must have high language." (Aristophanes, Frogs)


	
		
__________________________________
Do you Yahoo!?
Win a $20,000 Career Makeover at Yahoo! HotJobs
http://hotjobs.sweepstakes.yahoo.com/careermakeover