Re: [port-peer-review] reviews
Phillipe, (01)
> > Given that PORT is about finding the right mix of technologies (my
> > interpretation) with which to complete collaborative tasks, I must ask
> > what's not to like about concept maps?
>
> Personnaly, I would not call them a "knowledge representation" language
> since, correct me if I am wrong,
> - you can invent whatever label you want for concepts/relations nodes
> (you can but do not have to declare and organize the labels into
> a set of labels the use of which is constrained by subsumbtion
> relationships, exclusion relationships, relation signature, or more
> elaborate constraints);
> - there is no quantifier (exist, forall, numerical quantifiers,
> set quantifiers and set interpration modes, quantifier scopes, ...)
> and therefore not only "no denotation" but no meaning at all; (02)
These are all features of "formal" representations. Associative
representations such as concept maps are different. Associative
structures are not used for inferences. Instead they are used for
analogical reasoning. They certainly have meaning but not in the
formal sense. They are very important in human communication.
Associative arguments are acceptable to a person if they resonate
with what the person already believes. Politicians, for example,
use more associative than formal reasoning: their arguments have
to appeal to the public, they need not be logically correct. (03)
> - I do not remember if there are contexts or not (04)
Contexts for associative and formal structures are different.
Formal structures use abstract, global contexts (such as formal
contexts in FCA). Associative structures use local contexts.
They often require very concrete, cultural, spatial, social
contexts. Without context associative structures cannot be understood.
(But that does not mean that concept maps use explicit contexts.
They should but they are not perfect.) (05)
> In short, if the above is correct, you can write graphs quickly
> with Topic Maps but you cannot represent anything, and the result seems
> very hard to check and exploit for inferencing. (06)
Associative structures should not be used for inferencing. They
are used for communication. For example, think of the diagrams
often used in the business environment to communicate ideas
to bosses or investors. (07)
There is no doubt for me that associative structures, such as
concept maps are essential. But the main question is: how can they
be integrated with formal structures? More precisely, could
ontologies and formal techniques be used to extract formal
knowledge from associative representations? (08)
Uta (09)