The Problem with Phi: A Critique of Integrated Information Theory

摘要

IIT is a novel new theory of consciousness proposed by Guilio Tononi [1–8]. IIT appears to be gaining popularity, and Tononi has recently teamed up with neuroscientist Chistof Koch to promote the theory [4,6]. IIT defines a property of a system called integrated information (Φ) and postulates that this is an exact measure of the quantity of consciousness of any system. Tononi defines integrated information as “the amount of information generated by a complex of elements, above and beyond the information generated by its parts” (Tononi [3], p. 216) and states, “The integrated information theory (IIT) of consciousness claims that, at the fundamental level, consciousness is integrated information” (Tononi [3], p. 217, italics in original). While the details of the mathematical calculation of integrated information have varied with each new version of IIT, this abstract definition of integrated information has remained unchanged [1–8]. The claim that consciousness is integrated information has also remained consistent through the various versions of IIT [1–8]. Tononi hopes to someday use measures of integrated information to determine the level of consciousness in patients who appear comatose [2–4]. Tononi and Koch have also claimed that, according to IIT, a neuromorphic electronic device with similar connectivity and dynamics to the brain would be conscious, while a human-equivalent emulation of a brain run on a Von Neumann digital computer would not be conscious [2–6,8].