… are so crazy. Their editor-in-chief Chris Anderson, for instance, defended in a recent piece that we were now living in a time when science did not need theories anymore. ** !!!!! ** And why that? Because they could be advantageously replaced by data mining. In other words, he writes that a priori constructs can be dropped, when new computational technologies allow for powerful inferences to be made and validated.
It is clearly a provocative exaggeration, as a good editorial can be. Still, I wonder. Could we say in any meaningful sense that, the same way cosmologies were big theories that lost their central place in science around the turn of 1900 (I am thinking primarily of Herbert Spencer’s), theories as we know them could now be displaced by a mix of data + algorithms + computing power?
This idea kind of seduces me, even if I expect the reply “but data will always be theory embedded!” to be shot back immediately!
[The Erasmus Studio in Rotterdam will discuss this article in its research seminar on the 7 November – just contact me if you wish to attend.]