Page 380 - Deep Learning
P. 380
11
Elements of a unified Theory
… the supreme goal of all theory is to make the irreducible basic elements as
simple and as few as possible …
Albert Einstein 1
… condensing of a multitude of laws into a small number of principles affords
enormous relief to the human mind …
Pierre Duhem 2
The founders of cognitive science found it useful to remind everybody that
no benevolent deity has issued psychologists an insurance against comple x
ity. They implied that if the mind is complex, then we should expect theo
3
ries of mind to reflect that complexity. H. A. simon and A. newell wrote that
their “… theory posits internal mechanisms of great extent and complexity. …
That is all there is to it.” The three microtheories of creativity, adaptation and
4
conversion proposed in this book conform to this expectation. in each case,
the behavioral phenomenon that the theory was designed to explain can be
described in one or two paragraphs, but its explanation requires a chapter
length exposition. if the ultimate theory of cognition is a conjunction of many
such microtheories, then that theory is complex indeed.
But the founders might have taken their realism a shade too naively.
scientific theories are human constructions and they are shaped as much by
our ambitions, desires and needs as by the reality they explain. scientists strive
for concise principles because they cannot think effectively with complex ones,
due to the very limits on cognitive processing that the founders themselves
5
did so much to reveal. The undeniable complexity of reality reappears in the
process of articulating abstract principles visàvis particular phenomena, the
6
way the air in a balloon bulges out at the bottom when you squeeze its top.
newton’s equations of mechanical motion can be written on a single page, but
363