[Siguse-l] 'Linear' and 'non-linear' models.

[Tom Wilson]

Since nothing much seems to be happening on the SIG-USE list, I
thought I'd throw something in for debate :-)

I've been bothered for some time about the notion of 'non-linear'
models of ISB - a notion introduced, I think, by Allen Foster
(http://informationr.net/ir/10-2/paper222.html) and this concern has
been sharpened as a result of editing ISIC papers for publication.

Partly, my problem is with the word 'non-linear': I think that the
word 'linear' is usually associated with the other terms to do with
dimensionality - 'two-dimensional', 'three-dimensional' and
'n-dimensional'.  What Foster describes is not a 'non-linear' model
(except in that it is two-dimensional!), but a 'non-sequential' model.
 A number of authors have proposed 'sequential' models - sometimes
called 'staged' models, involving a series of stages through which an
individual is assumed to proceed in the process of seeking information
- Kuhlthau's is the obvious example, but the notion of a stages in a
process is also implicit in Dervin's 'sense-making' model.  However,
everyone that I know of notes that these are idealised models and that
feedback loops will inevitably occur in the process - indicating that
the process is not problem-free.  The feedback loops are explicit
(but, even there, simplified) in my 'problem-solving' model of ISB.

A further problem, if we accept that 'non-linear' actually means
'non-sequential', is that we have to ask in what sense information
behaviour can be other than sequential - we do one thing after
another. Certainly, we may break off from, say, an online search to
visit the library and browse through copies of print journals not yet
online, but behaviour implies action and action is inevitably
sequential. We can model the setting, or the context, within which ISB
takes place, which is partly what Foster does in locating his 'core
processes' within 'cognitive approach', within 'internal context',
within 'external context', but that is not a model of behaviour, since
it does not reveal what the person actually *does* in the process of
information seeking - and a model of behaviour must show what that
behaviour is, or at least, what it might be.

Thoughts, anyone?


Professor Tom Wilson, PhD, Hon.Ph.D.,
Publisher and Editor-in-Chief
Information Research: an international electronic journal
Website: http://InformationR.net/
E-mail: wilsontd at gmail.com
______________________________________