One of the long-established misconceptions about the lexicon is that it is neatly and rigidly divided into semantically related sets of words. In contrast, we claim that word meanings do not have clear boundaries.1 In this paper we will give proof of the fuzziness of meaning through an analysis of the semantic field of MOVEMENT in the English language. We will show that many MOVEMENT verbs belong not only to several subdomains within the field of MOVEMENT, but also to various semantic domains through metaphorical extension.
Before dealing with the double or even triple membership of MOVEMENT verbs, let us first present the model on which our description of the lexicon is based, the Functional-Lexematic Model (Martín Mingorance, 1984, 1985a,b; 1987a,b,c; 1990a,b).
1. The Functional-Lexematic Model
The FLM integrates Coseriu’s Lexematics (1977), Dik’s Functional Grammar (1997a) and some fundamental principles of cognitive linguistics. Following Faber and Mairal (1998: 4-5), the two main objectives of this model are, on the one hand, the construction of the linguistic architecture of the lexicon of a language, and on the other hand, the representation of knowledge based on the linguistic coding of dictionary entries.
The FLM establishes three axes of analysis: the paradigmatic, syntagmatic and cognitive axes. The elaboration of the paradigmatic axis entails the structuring of the lexicon in semantic domains —each corresponding to a basic area of meaning,2 and the organization of lexical domains into hierarchically constructed subdomains elaborated on the basis of shared meaning components A subdomain is “a subdivision of semantic space derived from the factorisation of the meaning definition of its members”3 (Faber and Mairal 1998: 6). Word definitions are built according to Dik’s method of Stepwise Lexical Decomposition. This means that the definition structure of each lexeme consists of the nuclear word —the archilexeme— and a series of semantic features which mark its distance from the preceding members of the subdomain.
Following Faber and Mairal (1999), the domain of MOVEMENT is organised into four subdomains. The first subdomain describes generic movement, while the other subdomains subsume lexemes which denote movement in a number of contexts: liquid, atmosphere and land. Cutting across this major configuration of the domain, the parameters of manner and direction introduce further divisions within each subdomain.4 For instance, these parameters traverse the following subdomains within the subdomain lexicalizing generic movement:
To move towards a place/person/thing
To move back
To move up
To move down
To move quickly
To move slowly
To move smoothly
To move in a circular manner
As an example of a subdomain structured paradigmatically, we have selected the subdomain To move down:
fall: to move down from a high position/the sky/a tree.
plunge: to fall suddenly a long way from a high position.
plummet: to fall very quickly from a high position.
come down: to fall (rain/snow) heavily.
descend: to move down a slope/stairs (fml).
The verbs indented to the right (plunge, plummet, come down) are defined in terms of the verb immediate above them (fall), which thus becomes their definiens. They are basically differentiated from one another in terms of manner. The other archilexeme of this subdomain is descend.
The construction of the syntagmatic axis implies the analysis of the complementation patterns of each lexeme using predicate frames as integrated formulae.
The following types of information are captured in predicate frames:
(i) the form of the predicate
(ii) the syntactic category to which it belongs
(iii) its quantitative valency, i.e. the number of arguments that the predicate requires
(iv) its qualitative valency, i.e. the semantic functions of the arguments and the pertinent selection restrictions
(v) the meaning definition
Predicate frames describe a state of affairs and specify the relationship between the predicate arguments (represented by the variable x). Each argument is characterized by a selection restriction —described in terms of binary semantic features— and fulfills a semantic function (Agent, Experiencer, Goal, Recipient, etc.).
Consider the predicate frame of the verb bow:
[ (x1: prototyp. human)Ag (x2: prototyp. part of the body)Go ]Action
DEF = to bend your head and upper body as a greeting or as a sign of respect.
This frame describes an Action and specifies the relationship between a human argument, performing the function of Agent, and an argument fulfilling the function of Goal and semantically marked as part of the body (head).