Keywords: Lexical frequency information, generative linguistics

1. Introduction
1.1 Studies on Word Structure in Formal Linguistic Theory
Looking back over the history of the theory of generative linguistics, we notice that the structure of the lexicon or the lexical component of the grammar, which was once considered mere storage filled with unpredictable information, has drawn a number of researchers' attention since early 1970s. From the late 1970s through the mid-1980s, a theory called Lexical Morphology (Siegel, 1974; Kiparsky 1982; Mohanan 1986, among others) took shape and played an undeniably important role in explicating a number of intricate phenomena related to the interface between phonology and morphology. The theory seemed successful in explaining the following observations. First, there are two types of suffixes in English with respect to their phonological properties: Class I suffixes and Class II suffixes. Class I suffixes may affect word stress assignment while Class II suffixes show no such effect. Second, suffixation is subject to a certain ordering relation. Put simply, a Class I suffix is always attached to the stem before any Class II suffixation takes place. To handle these observations, the theory was so constructed as to have a multi-layered structure which allows Class I suffixation to take place at the first layer (Level 1) and Class II suffixation at the second layer (Level 2), thereby ruling out such forms as *neighborlity. This ill-formed word offers a contrast to its well-formed counterpart neighborliness in that -ly, a Class II suffix, can precede another Class II suffix -ness for the attested form neighborliness, while for the unattested form *neighborlity, -ity is no longer available at Level 2 due to its Class I status. The internal structures of neighborliness and *neighborlity are represented in (1) and their licit and illicit derivational processes are illustrated in (2).

To view full article please see PDF or MS word file.