PL
This article – based on a larger study (Pawelec 2009) – has two aims. The more limited one is to present network models proposed by Ronald Langacker and George Lakoff. I try to show that both ventures rest on manifestly different assumptions, contrary to the widespread view that they are convergent or complementary. Langacker’s declared aim is “descriptive adequacy”: his model serves as a global representation of linguistic intuitions, rooted in convention. Lakoff, on the other hand, offers a developmental model: a fairly general abstract schema is “imagistically” specified and transformed, while the more specific schemas serve as the basis for metaphorical transfers. My wider aim is to offer a preliminary assessment of theoretical justifications and practical potential of network models in lexical semantics.