W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > January to March 2003

Re: 4.1 latest version

From: Maurizio Boscarol <maurizio@usabile.it>
Date: Thu, 30 Jan 2003 18:24:37 +0100
Message-ID: <008901c2c884$7f289700$6dce2a97@k7l2g9>
To: "Lisa Seeman" <seeman@netvision.net.il>, "W3c-Wai-Gl@W3.Org \(E-mail\)" <w3c-wai-gl@w3.org>

Yes, I think Lisa's right. This is a good way to merge my suggestion
into the checkpoint...


Lisa Seeman:
> I am going to try to suggest a way to merge this into our checkpoint,
>
> 1, Cohesiveness

... Ops: I made an error in my definitions! I named the two with the
same 'coherence' label... Sorry, it was a my mistake. What you name
'cohesiveness', it should be named 'coherence', according to what I
mean.
The meaning is right, anyway:

> Where we  state that one idea per paragraph is preferred,  we could
> recommend that at least all ideas/concepts in a paragraph are strongly
> related.
> Where we recommend  'Logic in the order and flow of information,"  we
could
> recommend that the logic and flow of ideas is reveled to the user (so
that
> when you switch from one topic to another , it is made clear - like a
new
> sub heading.. (this is related to checkpoint 1.3 and 1.4 on
structure. -I do
> not see the requirement of actually separating ideas into a structure)
>
> these recommendations could be in accommodation explanatory text

Good, you hit the point.
To correct my mistake, the one below should be named cohesiveness, not
coherence:

(> 2, Coherence)

2, Cohesiveness:
>
> I am not sure what you meant hear- is it the conjunctions like
promotes are
> easy to resolve? For example when you say "He" - which he? The main
subject
> of the paragraph or the last male to have been refereed to.
> for example:
> "this is a story about David...... David has a dog called Tag. He has
big
> brown eyes."
> compare with
>  "this is a story about David..... David has a dog called Tag. Tag has
big
> brown eyes."

Yes, perfectly, *cohesiveness* involves this example. Every ambiguity in
the reference caused by lexical choices should be avoided. But this
concept involve any appropriate lexical word that convey relationship
between two sentences or concept. I mean: consequentiality, causality,
simple co-occurrency, etc. The use of And, Then, But, Furthermore, and
so on, is to explicitate the relationship between phrases. They
'connect' the phrases, giving at the same time a meaning to their
relations.

Perhaps this is a topic that may vary in different languages. In
Italian, a tipical example is the use of 'but'. It' is to avversate the
concept it preceed, *but* (au contraire, instead...) sometimes is used
inappropriately to simply
separate or distinguish some concept - not to avversate it! It's a very
common mistake. (in
English 'but' is used even in the sense of "except": Everyone but me; in
Italian we use 'exept', not 'but', so I don't know if I can explain very
well).
Or a generic And is used instead of a better, more specific 'connector'.
I can try to make some better example, if you think it's useful.


> This is called Semantic ambiguity. Each word is not ambiguous, but
other
> knowledge sources are required to determine the meaning of a sentence.
For
> example, if we already know that David has blue eyes, then we know how
to
> resolve the sentence. However for a slow reader who may  forget minor
> details of  what he read yesterday, the ambiguity remains. ( from a
machine
> perspective it is also hard to resolve .)

Yes. The coherence and cohesiveness involves the capacity of reader to
make the right inference between two lexical informations. The right
inference leads to the correct mental rapresentation of the text. A good
reader can make the right inference even from a bad text, with lack of
cohesiveness or coherence. A bad reader (a reader with reduced cognitive
ability,
like some disable are) cannot. A un-cohesive or un-coherent text is
always more
difficult than well-cohesive/coherent one: the problem is also in the
term of
cognitive load.

The lack of coherence leads to 'creative', 'non-authorized' inferences:
we tend to re-create a coherent mental model even if the phrases are
unrelated, as many research has demonstrated. It can be a creative
writing strategy, too... If the author use it in a conscious way!
You can eliminate lack of coherence only introducing 'bridge-concept'
that
explicitare the reference, the relations between the two non-coherent
sentences.
But you can eliminate lack of cohesiveness just by choosing the right
word instead of the wrong one!...It's a very 'superficial' question: so
it's called lexical. Even if it is in some way conceptual, obviously.
But it can be addressed working on the 'surface of the text'. The
coherence cannot.


>
> I think that this is a classic case of reviewing "Consequences of
ambiguity
> or abstraction" .
> but it is a good discussion and certainly explanatory text  could make
this
> clearer.

Yes, I will think to some better example, trying to facilitate Avi's
work!

Best things.

Maurizio Boscarol
http://www.usabile.it
Received on Thursday, 30 January 2003 12:12:24 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:47:21 GMT