Similarities among different languages that are met with by means of empirical, confrontative analysis should not be mistaken for universals in the logical or philosophical sense. What Verma has described as the "propositional constituents" of a sentence (participants, process, and a temporal relation) may be considered logically universal. On the other hand, most "surface-structure" generalizations about language similarities do not refer to universalities. These non-universals include, among other categories, phonological changes that are explicable in terms of anatomical characteristics of human speech production, "taboo" lexical items, the notion of estimate from a distance that underlies the comparative degree, the construction of progressive verbal action by means of common lexemes, and the general use of nouns from the semantic field of diminutiveness in the function of negative intensifiers. (Author/JB)