Monday, September 10, 2007

Kuhn, paradigms, rules, and concepts of water

Kuhn makes the point that a scientific discipline does not need a set of articulated rules, even though it works within a paradigm. Moreover, uncovering any rules may be a frustrating and fruitless process for the academic.

In many ways, characterizing the concept of water within a paradigmatic framework is analogous to having rules bounding the definition of the concept of water within a discipline. For example, in your discipline:
  • Which water do you account for?
  • How do you measure water?
  • What are the more important characteristics of water that one might try and quantify?
  • How can your area of interest (with regard to water) be modeled?
  • What is a preferred result in a model scenario (assuming that a modeled scenario is preferred)?
Kuhn makes the observation that as paradigms come under question, the importance of rules becomes greater. In an analogous manner, this occurred with regard to the Augusta Ck project on which I worked. Explicit rules of the quantification of different, discipline-important aspects of water emerged (although I don't think we ever "defined" a unified theory of water). Kuhn doesn't make mention of cross-disciplinary work (at least as of chapter VII), but it makes sense to me that one way in which paradigms come under question is in such a setting, and when one works with various scientists, engineers, and other professionals, any (or all) of whom may not have a strong understanding of the origins and implications of their own discipline's central tenets, arriving at a unified goal (let alone an agreed-upon set of rules of procedure) appears (to the insider - me) all the more daunting.

2 comments:

Anonymous said...

Ah….not to rain on the Kuhn-parade, but although these are important things to learn and internalize as a scientist, ultimately there is nothing new under the sun.

Kuhn is making the (extremely) banal observation that different disciplines have different priorities and biases, based on their different approaches and theoretical foundations.

This is only a 'problem' if you insist on having a single Unified Field Theory of all science. Since human cognition actually *is* capable of mastering approaches and integrating multiple perspectives on the fly...well, you can see where I’m going. Furthermore, note that many scientific theories rely on excising as many correlates and outside factors as possible from the thing being studied, and thus NECESSARILY REQUIRE input and interaction with other theories and methods in order to actually describe Real World systems.

As a result, the current SOP in the social sciences (at least) is that a multidisciplinary, multi-methodological approach is of great merit. IIRC, this is similarly the case in biology and neurosciences. Indeed, there is *no* scientific discipline that is not in some way reflected in or by another scientific discipline, no matter how distant the connection. Think of it as the Kevin Bacon game of sciences – every science has insights that can speak to another; we thus have a Confederation of Science Theories, if not a Unified Science.

To respond to the Kuhnian "problem" of a definition of water, how do you 'bound the definition'?

Answer: ultimately, you don't, nor should you want to. As some wiseacres are fond of saying, "This has been another episode of simple answers for simple questions."

==========

I think Kuhn sometimes gets carried away with the philosophy of science and overlooks (or forgets) the practicing realities of it.

In multidisciplinary work, different (sub)disciplines will NOT have identical results or starting points or definitions or concepts of even basic items. But to complain about this is to miss the forest for the trees, because...

NO DISCIPLINE has a single set of core concepts or definitions. None. Nada. Zip. To pretend otherwise is to give scientific knowledge an absolute truth that it inherently does not claim to have.

If you doubt this, simply go to your nearest graduate library and find a scientific article in your discipline from 50 years ago. As the State of the Field changes, even the bounds and definitions very basic concepts can change slightly (or even greatly).

This is the reason why it is good scientific practice for EVERY paper, every journal article, every science book, to clearly identify what the Heck it is talking about and how it is defining it, what the methods were, what the results were, etc. This not only allow for replication, it also allows for a reasoned critical analysis of the findings, and the relevance and comparability (or not) to other findings you might have tracked down.

This ability to readily cross-compare work and build incrementally on existing research is something made *easier* for you when you work solely within a single sub-discipline that has a large publication base, and frequently such ‘insider’ papers will cheat on space by _assuming_ you share the same basic concepts/principles. But this is largely a result of the modern scientific industry, and NOT a necessary consequence of doing science. Historically, scientists could NOT count on existing work to tie neatly together for them, until well into the Age of Enlightenment.

Which I know Kuhn *does* recognize, but his philosophical musings often seem to disguise this fact.

The problems you raise about shifting paradigms in different disciplines will be inevitable when you are not working purely within one small sub-discipline of science. And the method and practice and tradition of science has work-arounds that do really work.

===========

Finally, this issue of hyper-specification with increasingly detailed definitions – this is an inevitability of hyper-specification. Human beings do NOT define words like dictionaries or scientists. Human beings “define” words by a very messy, regularly (potentially) ambiguous process of concept association and metaphorical extension. To define a word concisely is very useful in logical systems (which form the basis of much scientific practice), but such definitions – as well as the general system of logic itself – are inherently a very INhuman thing to do.

As a result, you end up with a delightful little paradox that was recognized by the ancient Greeks (and most certainly by philosophers even centuries before that) – that you are attempting to do something very, very weird by human standards (logic and precise definitions of words) using a system that is inherently human (language) and thus a VERY bad fit, and very bad tool, for the task at hand.

So in order to do it precisely and well, you have to over-specify.

Note that you can see a similar problem arise in legal systems.

That said, this is potentially a very useful thing for non-linguists (and philosophers not aware of the history of their field) to remember. Language is NOT a precise tool, but despite that is regularly asked by logical systems (like science) do very precise things. Sort of like trying to repair a Stradivarius with a chainsaw. That’s why you gotta be so careful and detailed and anal about the whole ‘definitions’ thing – and why you can end up with so many seemingly ‘different’ definitions when you’re all talking about ‘The Same Thing’.

Umlud said...

I suppose the answer to how to effectively "bound" a scientific a specific-use definition of "water" is to do it on a case-to-case basis. While this may seem IMHO to be unwieldy, I believe that this is a good method of doing it, especially in the course of transdisciplinary research.

This study-by-study definition of water will allow two things: a single, agreed-upon definition of the what the concept (e.g., "water") is and is not, and a means of shaping the course of studying the tangibles (or "operationalized" intangibles) in question.

Of course, an explicit definition may need to be written out when conducting final reporting, but I think the process will be illuminating for all involved, since doing so will likely force people to question why their own disciplines think certain ways about the concept, and with that understanding [hopefully] put together a methodology that makes intrinsic (rather than numerical, post hoc) sense.