Connectionism is rooted in long-standing views about the human mind. As a philosophy, it goes back to Hume's view that mental activity mainly consists in associations among ideas. As a scientific theory, it originates in the works of McCulloch and Pitts (1943), Hebb (1949), and Rosenblatt (1958), who developed theories of neural circuitry based on mutual activations and inhibitions among neurons. And as an approach in AI and cognitive science, it was instigated by a number of independent researchers during the 1970s and 1980s – for example, Dell (1986), Grossberg (1976), and McClelland and Rumelhart (1986). Most notably, the publication of Parallel Distributed Processing by the PDP Research Group in 1986 marks a resurgence of interest in connectionist approaches, which has persisted until now.

One of the active areas of connectionist modeling is language processing, which is an intriguing topic of research in AI and cognitive science. Probably no other topic has as immensely fascinated researchers in these fields, and none has been as challenging, elusive, and thwarting as natural language.