Neoclassicism, artificial intelligence, and the marginalization of ethics

Purpose – The paper examines the dependence of the positivist and welfarist preference satisfaction paradigm of neoclassical economics upon an implicit functionalist philosophy of mind. Functionalism is the doctrine that mental states are strictly materialistic and understandable in cause-effect terms. An important aspect of functionalism is the multiple realizability thesis, namely, that mental states can be realized in any type of hardware, whether human brain or computer. Design/methodology/approach – The approach used involves investigating the fact-value distinction after Robbins in terms of the positivist meta-ethical view known as emotivism, and then explaining emotivism as inherently functionalist. Functionalist thinking itself is explained in terms of contemporary philosophy of mind. Findings – An important finding is that the preference satisfaction paradigm can be shown to be as suitable to artificial intelligence systems as to human beings. A consequence of this is that normative concerns are increasingly difficult to address in connection with the neoclassical thinking about economic agents. Research limitations/implications – The paper does not investigate more recent research programs in economics (such as behavioural economics) that depart from basic neoclassical assumptions. Practical implications – A practical implication of the paper is that it shifts attention to previously un-emphasized aspects of neoclassical thinking. Originality/value – The paper's value to explain the relation of economics to ethics in neoclassical economics in connection with functionalist philosophy of mind.