W3C home > Mailing lists > Public > www-tag@w3.org > June 2007

Terminology (was Re: article on URIs, is this material that can be used by the)

From: Henry S. Thompson <ht@inf.ed.ac.uk>
Date: Tue, 12 Jun 2007 17:01:36 +0100
To: Pat Hayes <phayes@ihmc.us>
Cc: Tim Berners-Lee <timbl@w3.org>, Dan Brickley <danbri@danbri.org>, www-tag@w3.org
Message-ID: <f5br6ohyxen.fsf_-_@hildegard.inf.ed.ac.uk>

Hash: SHA1

Pat Hayes writes:

> HST writes:
>>OK, so further to this and your previous response, care to define
>>'represent', 'depict' and 'describe' as you would like to see them
>>used in this kind of discourse?
> 'Describe' relates a textual or symbolic document to what it, well,
> describes. Its the basic relation between a symbolic text and some
> part of the (or a possible) world, aka a situation, aka an
> interpretation structure. It presumes that the describer has a
> symbolic, parsable, structure to which meaning can be attached.

OK, that seems pretty straight-forward.  For example, in the WebArch
introductory example http://www.w3.org/TR/webarch/#intro, we can say
that the Oaxaca weather report 'describes' the weather in Oaxaca.

> So in a game, for example, the Eiffel tower jpeg might be a symbol
> for a goal state the players are racing to attain.

What???  This seems to be an example for 'represents', not for
'describes'. . .

> 'Depict' relates an image to what it is an image of. "Image" here can
> be construed broadly, to include audio "images" (sound tracks) for
> example. Cameras and recorders create depictions, but they can also be
> made by artists of course. Typically, depictions are not constructed
> from symbols, they have no syntax, and they represent by being in some
> sense similar to (a projection of) the thing they depict. In the
> literature often called a "direct" representation.

I'm good with this too.

> 'Represent' is an overarching term meaning any relationship between
> some information-bearing object and the thing or things it bears
> information about. It encompasses both the above and probably other
> things as well.

Hmm -- I'm less happy here.  Leaving aside the political science
meanings, seems to me there are two primary ordinary-language meanings
of 'represent':

 1) "stands in for"/"takes the place of", as in "the yellow disk
    represents the attacking force, the blue bar the defenders".

    Here we have some kind of arbitrary, externally stipulated,
    connection between an object or a symbol with a referent.

    I think _this_ is how I would label your game/Eiffel tower jpeg

 2) "corresponds in a regular way"/"re-presents", as in "The position
    of the arrow on this dial represents the amount of fuel in the
    tank" or "these WFFs represent propositions about the structure of
    my family" or "this (metadata, bits) represents today's Oaxaca
    weather report" or "the value of the variable *total* represents
    the number of registrants for the conference to date".

> I think tag-represent is a very narrow, special case of depiction, in
> this scheme of things.

Interesting.  I think it follows from what you said above that I
agreed with, and what I said above, that depiction is a sub-case of
restriction(2), where, as you say, there's little-or-no 'syntax', but
there is some kind of isomorphism.  In the digital world, it's almost
always some kind of projection (e.g. from 3-D to 2-D, or from stereo
to mono) and quantisation.  webarch:represents doesn't feel like that
at all to me -- it crucially involves syntax (that's what the
'metadata' part of the domain determines, after all), it usually
doesn't involve sampling or quantisation, nor is there necessarily any
kind of isomorphism.

Having said that, I don't think we've gotten to the bottom of the
ontology of the Web (in the old-fashioned sense of 'ontology' -- what
is the nature of the things involved).  In particular, I think
that trying to tell essentially the same story about




and what they identify (an image of Bill Clinton and a biographical
sketch of Bill Clinton, respectively) is to oversimplify, and to blur
a crucial distinction, namely that a (image/jpg, bits) representation
once rendered conformantly requires no further interpretation to be
what it is, whereas a (text/html, bits) representation _does_ require

But that's a story for another day. . .

- -- 
 Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
                     Half-time member of W3C Team
    2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
            Fax: (44) 131 650-4587, e-mail: ht@inf.ed.ac.uk
                   URL: http://www.ltg.ed.ac.uk/~ht/
[mail really from me _always_ has this .sig -- mail without it is forged spam]
Version: GnuPG v1.2.6 (GNU/Linux)

Received on Tuesday, 12 June 2007 16:02:00 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:32:52 UTC