Thursday, 15 December 2011

Triangles, trapeziums and three Cs. And a missing T.

A few weeks back SIRG hosted a very interesting talk by Amanda Spink, Professor of Information Science at Loughborough University.  When asked about a definition of information, she commented that for her and for the Information Science community, the definition of information wasn't really an issue (ie they aren't agonizing over it the way I am!), but, if pushed for a definition, it would be something like 'data that is meaningful for someone'.*

Compare some the definitions used by various authors in Ramage and Chapman:
  • Information is comprised of data in context (page 26, Bissell quoting Cordes)
  • INTELLIGENCE provided, a PERSON is informed by a SIGN about some THING in a certain CONTEXT (page 27, Bissell quoting Borgmann)
  • data plus meaning (interpretation) in a particular context at a particular time (page 72, Howell)

Ignoring the reference to time in Howell's definition for the moment - I'll come back to that - I think we can encapsulate all these definitions in a semiotic triangle.

But what is 'data'? What is an 'interpreter? And what is 'meaning'?

With reference to the definitions of information above, the interpreter is the 'someone', 'person' or 'context'. Though several of the formulations above assume this to be human, my contention is that it need not be.

Likewise, my contention is that 'meaning' is not only something associated with people. At DTMD2011, Rainer Zimmermann proposed adopting "Wittgenstein's perspective that the meaning of a piece of information is to be found in the action this information provokes" (see Zimmermann's presentation on this page). I've not read Wittgenstein so I suppose that I could be doing damage to what he was talking about, but I'm using a definition of meaning that allows a mechanical response to data to constitute 'meaning'.

Data is a difference of some sort.  So - maybe you spotted it coming - by equating 'provoking action' with 'making a difference' this conveniently aligns with Bateson's 'difference that makes a difference'.


Rather than the triangle, though, I prefer to use a trapezium, drawing from conventions in the engineering of communication systems.

There are two reasons for preferring the trapezium. (Actually it could equally well be a rectangle. You see both used in engineering, but I favour the trapezium because it distinguished the top and bottom, and looks nicer.)

1) It emphasises that information is always about communication. (I mentioned this here. But see also below, comments on tripleC.)

So the trapezium that I've drawn above, which is a 'receiver', must always somewhere be matched with a transmitter that is the source of the meaning. Or rather, you should always think about what the source of the meaning is, which you may recall was very revealing when exploring school reports.

2) It makes it easier to concatenate them in layers - so that the meaning from one becomes the data for the layer above. 'data' and 'meaning' are not absolutes: meaning is only ever relative to data, mediated by the interpretation.

*** Added later. I should add that I am fully aware that whereas the three vertices of the semiotic triangle have significance, the four vertices of the trapezium do not. The trapezium still has three significant features - top, bottom and content - so maybe the trapezium is not so clever. ***

TripleC: cognition, communication and cooperation

Wolfgang Hofkirchner and co-workers have developed the concept of 'triple C': cognition, communication and cooperation. Hofkirchner presented an interpretation of information exploiting the triple C framework in his presentation at DTMD2011.

Maybe the triangle is for cognition, and the trapezium for communication. Then when we bring in the dimension of time, we allow for cooperation.

Information doesn't emerge from one-off events. Meaning only ever emerges from conversation. A single message can convey information, but only because there have been messages before.  This is the beginning of cooperation.

So we need a diagramming convention that will display cooperation. I've no idea what that might be, at the moment, but it'd be really handy I can describe it by a word that begins with 'T' so we have 3Cs and 3Ts: cognition illustrated by a triangle; communication illustrated by a trapezium; and cooperation illustrated by a t...?

OK, so I know this doesn't all hold together, but this is a blog, not an academic paper, and it's work progress.

Any comments would be appreciated, though.


* I may not have got that exactly right. Since I was organising the talk I was distracted by the mechanics of the event, and not fully focussed on listening to what she said and not taking notes!

1 comment:

Magnus Ramage said...

Petri nets? They do focus on co-operation (at least between two entities) and they certainly include the time element.