Saturday, August 18, 2007

Determinism and uncertainty

There is a tendency (usually unspoken and probably un-, or semi-conscious) in the 'software world' to write about approaches and technologies in a deterministic manner – adopt this approach and the benefits will be realized, the millennium will come. There is something of an imperative to research, develop, and write from this point of view – what you are trying to do is establish absolute cause-and-effect. What would be the point of developing a piece of software, or a method of using it, unless doing so in some prescribed way is going to bring about some definite, repeatable result? If and when you don't come from this point of view, it feels wrong, and is looked at as deficient.

For a long while I've felt that my general approach doesn't somehow fit in the software world, or any of the realms where there is an unquestioned kind of determinism underlying any effort -- that is, what you are trying to do is come up with a method, process, tool, theory, etc. that, if applied and followed correctly, will have a certain cause and effect, a predictable result. Doing X will produce Y, if you do it well and correctly enough. Claims are made for approaches on this basis. If only everyone blogged with tags, or used GOMS to analyze usability, or followed rhetorical principles in discussions, or (for that matter) used Compendium (with or without IBIS), you'd attain some kind of utopia, a better world, or at least better software or products.

I don't think that way. There are no silver bullets in human affairs and there never will be. There are, sometimes, incremental improvements (that often have unintended consequences), and there are tools and approaches that can sometimes help in some situations. Of course I believe (passionately, maybe foolishly) that Compendium can be helpful, that making it better, and helping to help people to apply and use it, can have beneficial effects, especially in understanding each other. But I don't think that even if we "perfected" it (not possible), and everyone used it, that the millenium would arrive, paradise regained, etc. That's not even an underlying assumption or fond hope for me.

I sometimes feel that this puts me, and my work, and my writing, in the wrong light, and puts me at odds even with some of my close friends and collaborators. If I don't have an underlying millenial assumption, an underlying conviction that what I'm after is the development of a deterministic, infallible path, then I am not writing, working, developing, researching the right way. Perhaps that is what is missing for people in my writing -- if I at least acted like I was convinced of the infallibility and rectitude, that getting to that state was my goal, people would get what they wanted (and what they are now missing) from the work.

Mark Aakhus recently recommended reading the first chapter of Colin Grant's "Uncertainty and Communication: New Theoretical Investigations." I was arrested in the first few paragraphs:
"the indeterminacy of the range of communication options we have at our disposal in different cultures, different languages, different media and different social roles is not actually determinable at all. The porous form of communication actually means that even when we think we choose a clear, stable form, the penumbra of unselected
information remains." (p. 2)
What I like about the way Grant puts this is his emphasis on what is unknown over what is known. I don't know where he is going yet, but I'm hopeful there is some useful framework here for research that, even though it tries as hard as it can to contribute to effective, useful, and generative forms of communication, recognizes that there are no absolutes, there will always be gaps. It is up to the people using those forms in their unique situations to try to fill the gaps as best they can, in temporary and provisional ways. It is not a failure of the research, or the forms, to recognize this. Maybe Grant's work can help to frame the research in a way that better characterizes its value than holding it up to an implicit deterministic yardstick.

No comments: