This paper presents a rationale and a conceptual framework for a wider research project dealing with mathematical communication, and in particular with actions performed by interlocutors whenever they wish to clarify their use of a symbol, a word or an expression. The aim of such actions is often to repair a communicational breach resulting from differences in the interlocutors' uses of words. As was found in our study, only some of the defining actions would result in texts known as mathematical definitions. The point of departure of our research project is that the effectiveness of the defining actions is as much a function of the action itself as of its contexts. Our focus in this research is thus broader than in the past studies on definitions, and includes the when and why of defining along with their how. (Contains 7 footnotes and 4 figures.) [For complete proceedings, see ED500858.]

Definitions are not simply descriptions of meaning. They are acts that have different purposes and conditions. They can be dialogical tools for altering and sometimes manipulating the hearers’ commitments. They can be rhetorical instruments that can lead the interlocutor to a specific decision. The concept of persuasive definition captures the rhetorical dimension of the definitions of specific words, called “emotive”. By modifying their meaning or the hierarchy of values that they are associated with, the speaker can redirect the interlocutor’s attitudes towards a situation. From a pragmatic perspective, the meaning of a word can be described in different fashions, and be the content of different types of speech acts. Not only can the speaker remind the audience of a shared meaning, or stipulate or advance a new one; he can also perform definitional acts by omitting definitions, or taking them for granted. These silent acts are potentially mischievous, as they can be used to manipulate what the interlocutors are dialogically bound to, altering the burden of proof. The implicit redefinition represents the most powerful tactic for committing the interlocutor to a meaning that he has not agreed upon, nor that can he accept.

Whereas natural language concepts are based on prototypes, classical terminological definitions (CTDs) are based on necessary and sufficient conditions. The sociocognitive approach to terminology rejects both the possibility and the desirability of CTDs. In this paper, I argue that CTDs are needed for certain types of term, but not for others. In a first step, I distinguish two overlapping classes of expressions, based on two different criteria for termhood. Specialized vocabulary is the set of items whose use is restricted to specialized communication. Specialization is a gradual property, so that the boundaries of this class are vague. TERMS in the narrow sense have a concept with a clearcut boundary. In scientific and legal contexts, clearcut boundaries are necessary because classification is important. Only for TERMS in this narrow sense do we need CTDs. Some of the problems raised for CTDs can be solved straightforwardly by restricting their domain to TERMS. Discussions about t...