September 20, 2006 (2): (Anti-?) Social Software

It seems as if I always have the urge to respond to one of JP Rangaswami's optimistic items about social software in Confused Of Calcutta with my own measure of pessimism. This is not (just?) that I am naturally contentious so much as that, for better or worse, I still believe in the kind of synthesis that can arise from dialectical argumentation. However, if there is a "dark side" to social software, I would attribute it to the fact that most of the products I have seen that claim to be "social" may exhibit some very clever technology but do not necessarily exhibit much appreciation for why the product deserves to be called "social" in the first place.

I realized that this was the source of my skepticism this morning while reading Anthony Giddens' introduction to the essays he collected in Studies in Social and Political Theory. The subtitle of the introduction is "Some Issues in the Social Sciences Today;" and the fact that "today" happened to be in 1977 does not make this essay any the less relevant (which probably says something about our progress). The first issue he addresses is the one that his the social software nail squarely on the head: "Problems of method and epistemology." My argument is that any "social software" product that deserves to be called "social" needs to have firm foundations in both method and epistemology; and I am just not sure that this is a realistic requirement given our current understanding.

Now an optimist might react by saying that we should not worry if the foundations are weak. We can only learn about them through our attempts to build software products. While I am not against the scientific method, I am also haunted by some Randy Newman lyrics from my youth: "Let's drop the big one./See what happens." (I assume most readers know what Newman meant by "the big one!") Exploratory experimentation is all very well and good, as long as you remember the mantra that actions have consequences! Having said that, let me return to the theme of my argument.

The epistemological problem is one that I have been exploring for several years, most recently on this blog. To reiterate the key point, the epistemology of technology products is an epistemology of nouns (or, as software developers prefer to call them, "objects"). Whatever a piece of technology may do, it is not really covered by an epistemology of verbs. It is sufficient to deal with what the technology does in terms of transitions from one state to another. In other words there is always "before" and "after;" but there is never "doing." Much of my current interest in Kenneth Burke stems from his interest in the verb-based and the fact that he actually launched a "trial balloon" for a verb-based epistemology; but, as far as I can tell, Burke is pretty thoroughly terra incognita (if not "here be dragons") in the world of software development.

That brings us to the problem of method. One of the points that Giddens tries to make in his essay is that the social sciences may not be served by "scientific method;" and, since I have tried to make that point myself, I was glad to see it! In my own remarks I dared to suggest that we still do not have methods to evaluate social software and that we may not be able to do better than collect usage anecdotes. This is fine for a first step but only if we are not deluded into believing it is also the last one.

Am I taking the fun out of social software? I hope not. For better or worse I remember the "rape in cyberspace" case history in the LambaMOO user community; so I am naturally cautious about consequences greater than what we expect. Actually, I believe that social software may eventually be beneficial because it can be fun; but it will be fun the way swimming is fun. You have to go into the water with enough sensibility to avoid things like drowning and shark attacks!