Abstract

Steels & Belpaeme (S&B) are clearly interested in the possible test their models may provide for human language theories. However, they only superficially address the assumptions underlying their own agent architecture, while these are of crucial relevance to the topic of human language. These assumptions fit an Augustinian picture of language, which Wittgenstein challenges in his Philosophical Investigations. It is too early to draw conclusions regarding human language evolution from such models.