ABSTRACT
Morphologically complex words are not only concatenations of lexical elements. They also denote combinations of meanings: a “classroom” is something related to both classes and rooms. From a cognitive perspective, this has led to the question as to whether, in order to understand the meaning of a complex word, we need to access the meanings of its morphemic elements. A central topic of research is hence to what extent the processing of a complex word is influenced by its semantic properties and in particular its transparency, that is, the degree of contribution provided by the morphemic meanings to the meaning of the whole word. Results in this respect have been partially inconsistent, leading to a fierce debate about the role of semantics in the processing of complex words. More recent proposals suggest that a redefinition of the role of morpheme meanings may help solve this conundrum. When complex-word processing is conceptualized as the basis for an active and productive meaning combination process, empirical results are consistently observed and dissociations can be explained. This perspective, computationally implemented in models like FRACSS and CAOSS, reframes the role of semantics in complex-word processing in terms of meaning combination, and allows us to investigate not only how we understand familiar words, but also how we assign meanings to novel forms.
