What is Intelligence? : A proposed framework of four different concepts of intelligence

Abstract

Intelligenceis an ambiguous and generally poorly understood concept, and this becomes especially apparent in disagreements on the abilities and capacities of animals and artificial intelligence. In this thesis, I propose a framework of four different concepts of intelligence, that have in common that they describe capacity for complex problem solving, but otherwise have very different meanings and implications. These are algorithmic intelligence, which describes the intelligence of algorithms, system intelligence, which describes the intelligence of systems (artificial or biological), heterophenomenal intelligence, which describes conscious problem solving, and homophenomenal intelligence, which describes problem solving based on conceptual understanding. The thesis is an examination of the arguments that justify treating each of these as distinct and separate concepts, and an attempt to show what can be gained in philosophical discourse from such a distinction.