This paper examines and analyzes so-called measure adjectives in Korean, English, Japanese and Romani in a comparative fashion. Apart from the cases in Romani dialects, only a subset of measure adjectives which are pairs of antonym members allow morphological rule applications. That is, among the measure adjective pairs such as English high-low and deep-shallow, only the first members of the pair allow -t(h) nominalization, while the second members do not. This kind of morphological behavior is by no means random, but rather systematic and rule-governed. Unlike the general tendency of idiosyncratic morphological behaviors, the measure adjectives whose exact definition should be given later behave in quite an expected, predictable way cross-linguistically. As an extension of Shin (1992), Jang (2002), and Jang & Shi (2003, 2006), we will further show that only the unmarked, basic, positive measure adjectives allow a set of particular morphological rule applications, while marked, derived, negative measure adjectives do not. Furthermore, we show that the morphological patterns found in many Romani dialects are only apparent and not real. That is, the uniform pattern of morphological rule applications in many Romani dialects are not lexical, basic, or primary. Rather the morphological pattern in these languages is rather secondary, syntactic, and somewhat leveled.
(0)
(0)