Rui,
> but basically Biology has moved from
> being a hypothesis-driven to being a data-driven science. As Wilkins says:
>
> "While the great majority of biologists would agree with the dictum that
> 'nothing in biology makes sense except in the light of evolution', most
> can conduct their work quite happily without particular reference to
> evolutionary ideas."
If you want a good read on this issue you should get Ian Hacking's
'Representing and Intervening' (CUP, 1983). Briefly, as an ex-physicist
turned philosopher of science he rebalances Popper's strong
hypothesis-driven model of the scientific process by pointing out the
importance of 'data'. In fact he uses the somewhat more useful concept of
the 'creation of phenomena'. He does this to explain the importance of
experimentalists in physics:
"One role of experiments is so neglected that we lack a name for it. I call
it the creation of phenomena. Traditionally scientists are said to explain
the phenomena that they discover in nature. I say that often they create the
phenomena which then become the centrepiece of theory." (p.220)
This is relevant to the space syntax field precisely because of the
epistemic divide between modelling and analytic approaches to the study of
urban processes. Space syntax is analytic and so based on data and the
creation of 'phenomena' mainly through developing new ways of representing
data on urban morphology, human behaviours/social structures and the
relations between the two. The justified graph, the inequality genotype and
the axial integration core are 'created phenomena' in the sense that they
allow us to see regularities that are present in morphology but not obvious
to the unaided eye. Often this approach seems to bring us into conflict
(well - perhaps amicable discussion :-) with followers of the long and
honourable modelling tradition who start with a 'theory' - (for example:
that people are attracted to people according to a distance decay model in a
manner analogous to gravity), and see the role of data as one of setting the
parameters in that model.
The point that Hacking makes, and with which I agree, is that in any science
you need both hypotheses and the creation of phenomena about which to
hypothesise. You also need at certain points in the process to construct
models and calibrate these against data.
Alan
|