One of the problems is that both semantics and ontology are generally under the umbrella of “semantic technologies”, which leads people to think they are the
same.
However, semantics is about ways of referring (principally language) to the objects and categories of the world, whereas ontology is about those referents.
To some extent, the semiotic triangle (triangle of significance) tries to show that there are 3 nodes: 1) word, 2) sense/concept, and 3) world reference, with
some folks attempting to eliminate (2). I usually consider (2) to be a kind of mental placeholder for (3), so that when you try to obtain the denotation of a word, you go through (2) to find (3). Linguists (semanticists) are usually interested in (1-2), ontologists
are usually interested in (2-3), with, however, often different philosophical “stances” on how (3) (and sometimes 2) is interpreted, i.e., idealism, nominalism, conceptualism, and realism (and other shades). And for (1-2) there is lexical semantics and there
is compositional semantics, with different semantical “stances” on how to interpret these, e.g., model-theoretic semantics and cognitive semantics.
Using our language we can talk about impossible objects like “square circles” or anomalies like “green ideas”; these do not have real world referents.
Thanks,
Leo
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx]
On Behalf Of Amanda Vizedom
Sent: Sunday, February 19, 2012 12:08 PM
To: [ontolog-forum]
Subject: Re: [ontolog-forum] What goes into a Lexicon?
David,
On Sun, Feb 19, 2012 at 10:41, David Eddy <deddy@xxxxxxxxxxxxx> wrote:
At one point I entertained delusions that ontologies would help with this issue (one conceptual label = many physical labels). Obviously I no longer hope in that direction.
That's a shame, and a failure of the ontologies you've worked with. Many ontologies and applications exist that *do* help with this. From federated information integration to multilingual search and retrieval to
cross-domain (or even just multi-user-type) systems of all sorts, there are numerous existing ontology usages in which the ability to help with this issue is a key benefit of ontologies and a critical part of the application's success.
There are also plenty of insufficiently trained ontologists who do not understand the concept/_expression_ distinction sufficiently. I have been known to have long technical discussions in which I ban the use of
"term" "vocabulary" and several other expressions that are, in practice, ambiguous between the conceptual thing and the lexical thing; that was an extreme action necessitated by a lack of consistent understanding across the development team, such that some
would revert to thinking of ontological things lexically. That meant both applying unsuitable design criteria to the ontology *and* failing to capture the lexical expressions and relationship appropriately. Not to mention massive bog-downs as people debate
what "location" "really means", when there is no ontological conflict; there are simply multiple, distinct, and significant concepts which should be captured in the ontology, defined sufficiently (in formalisms, not just text annotations) to capture the differences
between them, and mapped lexically (i.e., given the label, etc.) to "location."
This is often a training and personnel issue. It is also, IMNSHO, one of the areas in which ontology as a field can improve by developing explicit dimensions of ontology requirement specification and explicit, experience-
and research-backed guidelines regarding ontology features and characteristics that are suitable or necessary for an ontology that must meet those various requirements.