Systemic Lexicography as a Basis of Dictionary-making

COMMENTARIES Systemic Lexicography as a Basis of Dictionary-making Jurij D. Apresjan J. find both Wierzbicka's and Atkins's papers insightful and stimulating, and the list ofwhat I agree with is far greater than the list of my disagreements. Yet to further the purpose of the present enterprise , I think it best neither to comment on what is acceptable to me nor on the controversial points, but rather to set forward my own position on the issues raised in both articles. In what follows I proceed from the following three assumptions . First, the supertask of any explanatory dictionary consists in presenting the vocabulary of a language as a system, that is, as a set of lexemes (= word senses) falling into well-defined classes on the basis of shared properties. Second, lexemes display a number ofquite non-trivial recurrent features that extend over the senses of a polysemous word and over the vocabulary of a language at large. Those recurrent features pertain to the semantic, pragmatic, communicative (theme-rheme, or topic-comment), syntactic, selectional, prosodie, and other properties of lexemes. Third, theoretical, or systemic, lexicography sets out to capture the maximum number ofsuch recurrent features and to record them in a stricdy unified way in the dictionary entries of the respective lexemes and in all the other parts of linguistic description where the need to mention them may arise, including the grammar. From the above it immediately follows that there are two things to systematize: lexicographic types (roughly speaking—sets of lexemes with recurrent properties) and types of lexicographic information to be entered in the dictionary. There is a meeting point between theoretical lexicography thus understood and practical dictionary-making: at least one category of dictionaries, namely dictionaries for speakers (geared to the needs of 80Jurij D. Apresjan text production rather than text interpretation), will serve their purpose best if they take into account the lexicographic types for a given language and the types of relevant lexicographic information to be included in the dictionary entries. I will take up those two notions in more detail in the following two sections. One more important problem of systemic lexicography is that of its defining language, or metalanguage . That will be dealt with in the final section. Lexicographic types As has been stated above, the lexicon of every language falls into a number of lexicographic types. By lexicographic type I mean a group of lexemes having a number of properties in common that are sensitive to the same or similar sets of linguistic rules—morphological , syntactic, prosodie, semantic, etc. This notion is thus different in two respects from the more familiar notion of a semantic class. First, the underlying properties of a lexicographic type are not necessarily semantic. Second, lexicographic types are defined against the background of various linguistic rules (broadly speaking, "grammar "). In other words, they make sense only within the framework of an integrated theory of language, that is, an ideally coordinated description of the lexicon and the grammar. Every lexicographic type should be treated in the dictionary in a unified way. The set of lexicographic types for the given language is determined above all by the way its conceptual material is patterned. This pattern of conceptualizations inherent in lexical and grammatical meanings ofthe language and obligatory for all its speakers has come to be called the "naive picture of the world," or the "world-view." The naive picture ofthe world can be partitioned into naive geometry, naive physics of time and space, naive psychology, and so on. In many important respects these differ from the corresponding fragments of the scientific picture of the world and display a number ofsufficiently general features cross-linguistically (the predominant trend) or across some specific part of the lexicon within a single language. Reconstructing the naive picture on the basis oflanguage evidence is a prerequisite for systemic, lexicographic description. The first use to which such a reconstruction may be put lexicographically is the extracting of a general scheme for a uniform description of a certain class of words. I will illustrate this with a fragment of naive psychology concerned with emotions, where, apart Systemic Lexicography as a Basis of Dictionary-making81 from my own findings...