I think Sensei clearly points out the limits of traditional scientific methodology here, but I feel he limits the term science to include only clinical studies utilizing minimal variables. I think that science is beginning to voyage into the real world more and more...
I think of Sudhir Venkatesh'swork in the sociology of urban environments and how that spawned economical studies of crack dealing and prostitution by Levitt and Dubner, or Doug Tallamy's work assaying wild life on different trees, and how there was less biomass on non-natives than natives. These are not sterile laboratory computer modelled abstracts; these are measures of natural worlds and natural systems. As we finally catalogue the world around us the new science is showing Sensei's intuitions to be correct; that we cannot understand the full complexity of the system, but we can glimpse enough to know trends.
In much the same way that Venkatesh could not have gotten good data without a real and true immersion into the culture of urban life, with ALL it's attendant dangers and issues, we cannot get natural gardens without risking some issues, accepting some losses. There is a certain amount of fatalism in either adoption, but the rewards are worth it in the long run... but I am not sure the abandoning of science is entirely necessary as long as we use that definition of science as making the most accurate measurement you can. We need yardsticks to measure success; they simply must be accurately marked...
HG