I finally managed to grab some time to read Mika’s article and found it very interesting – not in a mathematical sense, but maybe in more of a ‘social science’ sense.
I liked the notion of a tripartite model of actors-concepts-instances. This move from the bipartite model (concept-social) to include the ternary element of ‘actor’ or ‘agency’ is interesting. It kind of mirrors something I’ve come across in my own research into cultural semiotics where, in moving from notions of binary opposition in meaning generation, the idea of a ternary system is introduced, the third element being the process of interaction… whether social or otherwise. It seems to me that in a triangulation of elements, a certain notion of balance emerges – perhaps that’s because the ‘thirdspace’ situates or anchors the propositional knowledge? So, in this way… the introduction of the ‘actor’ by Mika lends itself to his later elaboration of ‘semantic identity’ being something that emerges from a community network or collaborative group, for example in the production of ‘tags’.
The functionality of a ‘group’, as it were, automatically ‘frames’ their thinking (tagging) and collective memory (Mika’s ‘agreed upon terms’) and this, of itself, helps to categorise the apparently ‘loosely’ defined tags generated by the members of the community. So that, if you like, even though each individual makes an individual choice as to which tags to use, the tags are used within a ‘visualised set’ belonging to the group (Mika’s ‘concensus’), which makes them less problematical as identifiers because the element of randomness is thus lessened.
I was intrigued by Mika’s raising of temporality as a problem in concept definition… (his ‘ontology drift’) but, for myself, saw this as a natural part of the evolving semantic map and not as problematic as he suggests as, surely, there are causal links which tie new concepts back to old ones, even as communities change and evolve? Knowledge may be ‘codified’ as Mika suggests, but must codification always signify something which is fixed?
I liked Mika’s vision of a dynamic ontology as a ‘community of self-organising, autonomous, networked and localized agents… establishing connections and negotiating meaning only when it becomes necessary for co-operation’. He suggests that there is a lack of an abstract model for such a system but, actually, it seems to me that this ties into Yuri Lotman’s concept of semiosphere (forgive me for plugging my own research interests here). That said, Lotman’s semiosphere is demarcated as an abstract model for meaning making/meaning generation, whereas Mika’s interests seem to lie in connectivity of concepts for meaning making. Both, however, envisage meaning made on multiple levels and as multilayered.
Mika points out the lack of ‘one-to-one correspondence’ in keywords generated by individuals when tagging… however, although not explicitly stated… the notion of ‘many-to-many correspondence’ seems somehow appealing in the context of tagging and folksonomies.
I really liked his explanation of the relationship between user, object and concept and the notion of ternary associations. Forgetting the maths and the fancy language – lovely word ‘hypergraph’ but don’t they block the way sometimes? and just fixing on the concepts, I liked the idea that the combination of these three produces a ‘higher level’ concept that is accessible in different ways, so that, even where tags are non-identical, some overlap in conceptual framing at this ‘higher level’ may bring them together into one category – so, where two might not meet, three have the potential to do so… (thanks for the nice graphs, Yish). I did chuckle a little at the understatement:
Tripartite graphs and hyperedges are rather cumbersome to understand and work with. (Uhuh!)
How I summarised the rest… take two people who share an object, expressing its value with different tags… the object’s defining potential is multiplied by two… and so on. For me, I ignored the hyperedges and affiliations and instead pictured an ever-growing spirograph, growing rapidly more dense at different nodes around a central point (actor or object) – equivalent to the weighting/filtering properties of the ‘affiliation network’ – well, it kind of works for me and makes for interesting reflection. The idea of ‘overlap’ in conceptual framing generating an emerging semantic network seems eminently feasilbe (and somehow sensible) to me.
Skipping along, some other things I picked up on and which interested me:
- the notion of observing tagging behaviour to determine the kinds of meanings they generate – whether collectively or individually
- reflecting on the idea of a ‘collective mind set’
- the ‘well-known law of community formation’
I really like this: “interaction creates similarity, while similarity creates interaction” – I think it’s a nice phrase to link to the whole idea of social networking, although I’m curious as to the scope of the notion of ‘similarity’ and what this might engender in terms of collaborative or communal interactions. It also kind of brings up the question of ‘the chicken or the egg’ and which comes first. Something worth thinking on, just the same.
Some quick notes without too much reflection (from later sections):
social bookmarking: link, describe, tag (keyword) and share
bookmarks: searchable, affiliated, associated, temporally situated
issues: synonyms, ambiguity, generalisation
community: core interests, clustering, evolving, correlations
folksonomy: tagging, keyword associations, community ontologies
Finally, the comparison of ‘object’ (say, search engine) and ‘actor’ in terms of effectiveness as meaning negotiators had me thinking (and I’m still thinking)… when I read this:
Overlapping communities turn out to be a stronger link than overlapping sets of web pages. A possible explanation is that even after including the disambiguating term in the query, the search engine still suffers from knowing too much, blurring away community-specific interpretation.
The idea of a search engine ‘knowing too much’ made me smile… sure had that ‘glut’ problem myself. Quite apart from the fact that a community has more consciousness (by virtue of its members) – the idea of the ‘overlap’ between communities generation meaning in and of itself is an interesting one. I’m wondering about the point of intersection and the ‘zone’ where meaning is made in such a case… which kind of ties in with Lotman’s notion of ‘explosion’ and the idea that where two apparently correlating actors and concepts interact… meaning is self-generating at some level but not necessarily immediately perceived… I’m wondering whether Mika’s interpretation of the interpretative impact of overlapping communities is similar to Lotman.