Chapter 2 — Exploring Semantic Relations

2.1 Semantic (conceptual) relations

2.1.1 Semantic relations and the field of terminology

According to Robison (1970, 273), "Human language can be viewed as a vehicle for describing relationships". In the field of terminology specifically, determining concepts, conceptual relations, and conceptual networks is vital. In the words of Felber, "any terminology work should be based on concepts and not on terms" (1984, 116). He defines concepts as "the mental representations of individual objects....not only of beings or things...but...also of qualities...actions...and even of locations, situations or relations." (1984, 115). Lyons' definition is the following: "By concept is to be understood an idea, thought or mental construct by means of which the mind apprehends or comes to know things....concepts mediate between words and objects" (1977, 110).

Concepts do not exist in a vacuum, though; they have meaning only by being set in relation to other concepts. In fact, "the sense of a lexical item may be defined to be, not only dependent upon, but identical with, the set of relations which hold between the item in question and other items in the same lexical system" (Lyons: 1968, 443). Ahmad and Fulford reflect this as well: "The terms of a domain do not exist in isolation, they are related to one another in a variety of ways" (1992, 1). In any subject field, then, there are semantic links among the concepts. The web of such links constitutes what is known as the conceptual network of the subject field. In French, this network is called the réseau notionnel. In the words of Rondeau (1981, 85), "une notion se délimite par le biais de ses rapports avec d'autres notions et par la place qu'elle occupe dans un réseau notionnel."

Terminologists engage in concept analysis to determine the relations between the concepts that the terms designate, for the purpose of understanding the subject field and adequately representing that knowledge on term records. According to Meyer (1994, 7):

High-quality concept analysis is a sine qua non for high-quality terminology work: without some understanding of the conceptual structures underlying the domain, the terminologist cannot properly carry out many of the practical tasks related to the production of a vocabulary.

During a terminology project, terminologists aim for as full an understanding of the domain as possible. Sager (1994b, 43) explains the difference between passive and full understanding: "We understand 'passively', when we have only a vague idea of the place of a concept in the knowledge space. We understand 'fully' when we know the precise place of a concept in relation to others". We can consider a given subject field as a special knowledge space, and "pour un domaine donné, chaque notion occupe une place définie à l'intérieur d'un système organisé de relations" (Larivière: 1996, 410).

Moreover, conceptual analysis of the subject field is not an isolated activity in a terminology project; rather, it permeates all aspects of the project. In the words of Meyer (1993, 140), "Concept management is crucial to all stages of terminology work as it is practised today and will become even more so for the next generation of term banks." Being able to manage concepts adequately and establish conceptual networks presupposes a thorough understanding, on the part of the terminologist, of the various semantic relations and the ways in which relations between concepts are expressed in real language.

2.1.2 Linguistic patterns that express semantic relations

In the documentation of any subject field, definitions and explanatory material(3) are used to make explicit the semantic relations that exist between concepts in that subject field. Each semantic relation can be expressed by a variety of linguistic structures consisting of one or more words that indicate to readers what relation is being dealt with at that moment. The pattern consisting of, used in the previous sentence, expresses meronymy, i.e. the "parts of" relation. This structural pattern is, therefore, a powerful semantic device that makes it clear in the reader's mind that the relation being expressed between concept X and concept(s) Y is meronymy.

2.1.3 Definitions as expressions of semantic relations

Austin considers definitions to be a type of performative utterance (1962, 65). He explains that the term performative utterance "indicates that the issuing of the utterance is the performing of an action--it is not normally thought of as just saying something" (1962, 6-7). For example, a person uttering the sentence "I promise to call you tomorrow" is making a promise by saying so. Said another way, by uttering the above sentence, a person is doing two things: saying something and making a promise. The word promise in this sentence is called a performative verb.

Pearson (1996, 817-818) explains that performative utterances are not always prefaced by the performative verbs that act as markers, such as "I promise..." or "I swear..." Indeed, definitions generally do not appear with the explicitly stated performative verb phrase "I hereby define..." signalling that a definition or some explanatory material is about to follow. Readers are nevertheless able to recognize a definition when they are reading a text. How? Austin explains that illocutionary acts are conventional acts, i.e. "an act done as conforming to convention" (1962, 105). In the case of definitions, those conventions take the form of a) an equational structure and/or b) formulaic linguistic patterns or structures that mark the utterance as a definition.

As Picht and Draskau (1985, 51) explain, "the definition has a considerable similarity with the mathematical equation." Indeed, following Aristotle, the structure of a formal definition can be written as a simple equation:

Definiendum = Definiens
(that which is to be defined) = (that which is defining)
= Genus + Differentia specificae

In natural language, the "=" sign is usually expressed by a verb such as is or consists of. For example: a car is a means of transportation that has four wheels and travels on roads, which distinguishes car from other means of transportation such as motorcycle or boat. The diagram above shows the classical definition structure (called an intensional definition, in the terminology literature). However, as Picht and Draskau (1985, 51) point out, "with certain modifications to the right hand side of the equation, this structure may also be applied to other types of definitions." One type of definition described by Picht and Draskau (1985, 52) that has relevance here is the extensional definition.

The extensional definition involves naming the units that make up a whole or naming the list of specific instances of a generic concept. Consider the following examples:

a) North America is a continent made up of three countries: Canada, the United States, and Mexico.

b) Types of operating systems include DOS, Unix, and Windows.

Regarding the order of elements in a definition, Pearson (1996, 818-819) points out (with reference to the classical definition structure) that "contrary to what one might normally expect...the term which is being defined can appear before or after the defining statement", i.e. on either side of the definitional equation. For example, "The black substance made from the remains is called humus" (Pearson: 1996, 823).

At this point, it should be pointed out that the definiens side of a definitional equation can contain actual terms from the same subject field in which the definiendum occurs. Ahmad and Fulford (1992, 18) note that "definitions largely comprise descriptions of semantic relations holding between terms [concepts]." For example, consider the following sentence which defines the concept compost process, from the field of composting: "The compost process is a partial breakdown of organics by microorganisms such as bacteria and fungi." The sentence fragment following the is a structure contains the words organics, microorganisms, bacteria, and fungi, which are, themselves, terms from the field of composting. This is useful for terminologists during the stage of term identification and the fleshing out of their concept tree of the domain.

Loffler-Laurian (1990, 14-18) deals with definitions as expressed in scientific discourse. She describes five catégories définitoires corresponding to five ways of answering the question "Qu'est-ce que c'est?". This question can be answered by a) naming the item (Denomination); b) providing a lexical equivalent having a larger extension (Equivalence); c) indicating the characteristics of the item, usually through adjectivisation (Characterisation); d) describing the parts/components of the item (Analysis); and e) revealing the item's function (Function). Loffler-Laurian (1990, 19-20) also points out which types of scientific discourse the various types of definitions can be found in. She explains that, in highly specialised scientific discourse, definitions are very rare, but when provided, they tend to be from all definitional categories except Function. In semi-popularised discourse, Analysis and Function definitions are most common. Popularised discourse (the goal of which is to inform the reader, yet lending the appearance of scientificité) seems to provide all types of definitions except Characterisation. In pedagogical discourse, the reader can expect to find Denomination, Characterisation, and Analysis, with Equivalence and Function being uncommon.

The information outlined in the preceding paragraph is evidence of the importance, in terminology, of including in the corpora texts having different levels of technicality. Terminologists, seeking as much information about the given subject field, will need access to all types of definitions, and not just to those provided in, say, text written by and for specialists.

Section 2.2 describes the relations of hyponymy, meronymy and functionality and includes a discussion on previous research related to linguistic structures, appearing within definitions and explanatory material, that express these relations.


Footnote #3

By explanatory material, I mean any definition-like information that cannot be classified as a formal definition.

back to main text