Death of Theory?
I have been wrestling much of this week with a WIRED article by Chris Anderson entitled “The End of Theory: The Data Deluge Makes the Scientific Method Obsolete”. [Link] The wrestling is probably more a matter that I do a lot of analysis and modeling (as in the construction of informational models, not the other kind that has to do with clothing and other goods) which makes me more of a theoretician than an experimentalist. Otherwise I would probably have shrugged the piece off as the typical superficial puff piece that characterizes traditional media.
What the article has at its roots is the shift from information poverty to information abundance that the internet represents and embodies. Then, based on the success of Gooey using correlative analysis, or some such – they are not really forthcoming about this, the obvious conclusion is that theory is antiquated, archaic, and useless. And from that the Scientific Method is extinct.
Sorry. Not obvious at all. The argument is so seriously flawed that it is obvious why people whose universe is the ‘net and information stuff can’t see it, because this world isn’t about science, it’s about information.
Science is fundamentally about understanding how the universe works. The internet is only secondarily connected to that reality. The internet is a technologically enabled social construct. As such it has to do with reality only through the operations of the technology that enables it and through the humans whose efforts populate it with information. And what Gooey is doing is not being successful at understanding things but being successful at a particular way of making money. They are, or at least have been, very good at that but even if they actually do understand what that reality is, it is not the reality of the universe.
Theory is the codification of our understanding of reality. As such it is inseperable from experiment. Experiment affirms or destroys theory; theory provides direction for experiment. That is direction as in azimuth, not mandate.
The problem has always been what experiments are important, and how can predictions of theory be tested. The internet can enhance this only as a medium of transportation of the information. It does not analyze the data, it does not understand the data, it only manipulated the data at best.
Understanding is human, not mechanical.
