copyright declared on today's material [RE: [humanmarkup-comment] OurNext Meeting]

[cognite@zianet.com]

Better copyright this immediately.

Copyright S. Candelaria de Ram, 17 7/2002.

Requested response to:
At 04:17 PM 16-07-2002 -0500, Len wrote:
>...  I was adjusting the content model:
>
><!ELEMENT sign (sign*, signifier, signified+, referent) >
><!ATTRIBUTE sign
>  id ID #REQUIRED
>  type (symbol | icon | index) #REQUIRED >
>
>A signifier may be associated with multiple signifieds. 
>That may be a place to put context modifiers. 
 ...[where as ATTRIBUTE?] 
>type indicates a relationship between signifier and 
>signified, does it apply to all of the possible 
>signifieds?  If that is so, then we should go at 
>this one again.  ...
>
>Also, a sign and a signifier are essentially the same 
>thing, yes, so we could drop that or use it as a 
>container for representations (eg, refer to pictures 
>of images) for cases where we know a sign can have 
>multiple forms.
>
>Comments?

Here's one initial reaction:

Markup assists a communication process between agents.
Let us term any "signing agent", whether sender or receiver,
a SEMIOTE.  Then explicitly, HUML markup is to aid accurate 
transfer of meaning in this particular communicative semiosis process.

        SEMIOTEi  -- markedup signal --> SEMIOTEj

Adopting the Kleene star again to allow for possible multiplicity of
participants would give something on the order of:

        SEMIOTEi*  -- markedup signal-complex --> SEMIOTEj*

There has ever been the possibility of multiple listeners and speakers.
The arrow has implications of temporality as well as channel.

Each SEMIOTE acts within a [geo]temporal context, which is
to admit that this process takes place in the real world, a matter
that we have to deal with rather more specifically perhaps than
semiotics has to date.  (If the SEMIOTE does not act, there is
no communication.)

The signal transmission is also contextually bound.  It is not
necessarily lossless nor adsorptive in any given case.

Adding context to each component of the description explicitly
yields something like this:

        (SEMIOTEi - in context)* 
                -- (marked up signals transmitted - in context)* -->
        (SEMIOTEj - in context)*

In more detail, there is genesis of the signals:
        (contextual stimuli)* X (SEMIOTEi) --> reactive interpretation
        with emission of signal-complexes 
                at [geolocations and] time(s) A
                in media M
                complexes include signs: 
                        patterns, which may be traditional in context
                        inadvertent and advertent [intentionalities]

(The indented features might be rendered as "attributes".) We add to this
unadorned process:
                markup
                        described (parsed?) portions of the signal-complex
                                attribute markings (right?)
                        additions, like ID for convenience

Those [marked up] emitted signal-complexes get conveyed between SEMIOTES:

        signal-complex with optional markup

        travels between SEMIOTES thru channel(s) - in context(s)
                at geolocations and times B >= A
                undergoing possible transformation (loss, aggregation, mutation)

        becomes part of the context for 
                receiver(s) with capacity to do semiotic processing, i.e.,
SEMIOTESj

All told, the semiosis we would seem to need to describe includes our own
enhancement, a signal transmission, and context at every step.  

Of a SEMIOTE we may or may not need to explicitly include, for HUML purposes,
detail of stimuli being both external and internal, of what I've called
above "reactive interpretation" and Len has included as ineluctable effect.
Some of this effect we
might consider primary in that it yields the signal-complex requisite for the
communicating we model.  When the effect of a SEMIOTE's signaling is
re-absorbed we have the idempotent case of "reflexion", where SEMIOTEi
communicates with SEMIOTEi itself.  But it is rather much for HUML to
try to model all psychology, as Len says!

There seem to be these two sequenced components in our semiosis model, having
this nature:

        o       SEMIOTEi X context
        o       transmittable signal-complex  X context

To handle context at every stage, Len's idea of bundling CONTAINERS along
with components seems quite well-reasoned.  
                
Our criterion for design, in this framework, is that good HUML markup lets
the receiver accurately interpret (comprehend) the reaction to stimuli by
the sending SEMIOTE.   Our additions should not distort the signals.  The
critical gap we may
need to fill involves mismatch of communicant SEMIOTEs' current contexts.
Presumably the "referents" reside in those contexts.

Put in practice, we may also face issues related to transformation of
signal-complexes as they pass thru channels that convert them to different
media.  (Will the attributes
characterizing the stream need to be transformed also?)  Also, I'm not convinced
at this point that a signal-complex can be neatly characterized as signs,
nor helpfully distinguished as icon vs. symbol vs. index, perhaps because
I've been thinking more in terms of token.  

Signifier seems like a real good term, implying process.  And, as noted,
Len's idea of
treating things as CONTAINERS might be just the ticket for including contexts.

SC

Other reactions?


----------------------------------------------------------------
To subscribe or unsubscribe from this elist use the subscription
manager: <http://lists.oasis-open.org/ob/adm.pl>