Polysemy and metaphor in perception verbs: a cross-linguistic study

1.1.2. Methodological Principles 
Human categorisation is one of the major issues in Linguistics. The ability to 
categorise, i.e., to judge that a particular thing is or is not an instance of a particular 
category, is an essential part of cognition. Categorisation is often automatic and 
unconscious, except in problematic cases. This can cause us to make mistakes and make 
us think that our categories are categories of things, when in fact they are categories of 
abstract entities. When experience is used to guide the interpretation of a new 
experience, the ability to categorise becomes indispensable. How human beings 
establish different categories of elements has been discussed ever since Aristotle. 
The classical view on categorisation, that of Aristotle
3
, claims that categories are 
defined in terms of a conjunction of necessary and sufficient binary features, that is to 
say that linguistic analytical categories impose a set of necessary and sufficient 
conditions for the membership in the category. This requirement not only implies that 
categories have clear boundaries and that all members of a category have equal status 
(Taylor 1995: 25) but also that there is an abstract, general definition with which all the 
members of that category must comply. For instance, the different senses of the word 
hand in John’s hands are very big and in The hour hand of the clock would be 
considered as related to one general, core abstract sense of hand.
However, this abstract definition of ‘core meaning’ is problematic; as Sweetser 
(1986) points out, in cases when the extension of meaning has been carried out by means 
of metaphor or metonymy, it is very difficult to identify this abstract meaning. In the 
examples above, it could be argued that this core sense is a ‘pointing function’, but this 
core meaning cannot account for other instances of hand as in My life is in your hands, 
They are taking new hands or This matter is out of my hands. These other examples lead 
us to another problem: no matter how complex this core abstract meaning could be, it 
will leave some likely candidates outside of the domain. 
These above examples would be analysed quite differently under the Cognitive 
Linguistics methodology. Instead of relating these different senses to an abstract default 
3
Aristotle distinguished between the essence of a thing (what makes a thing be what it is, 
indicates its individuality, its destruction is the destruction of all) and the accidents of a thing (incidental 
properties, not determining part). 

---

B. Iraide Ibarretxe-Antuñano 
Polysemy and metaphor in perception verbs 
20
sense that includes all of them, the cognitive approach adopts a prototype categorisation 
model (cf. Rosch 1973, 1977, 1978, 1983; Rosch and Mervis 1975; Mervis and Rosch 
1981). In this model human categories have two types of members: the ‘prototype’ and 
several less-central members related to the former in a motivated way. The prototype is 
the best, the most prominent and the most typical member of a category. It is the 
example that first comes to mind when one thinks of that category. 
Prototype categorisation
4
stems from Ludwig Wittgenstein’s (1953) thesis that 
necessary and sufficient conditions are not appropriate for defining the meanings of 
many words. His example of the concept of game showed how there are very few 
properties, if any, that are shared by all games; instead, one game shares some properties 
with another game, and this other game may share some properties with a third and so 
on. This concept of game is based on what he called ‘family resemblance’: members of a 
family resemble one another in various ways. But, everyone in the family does not need 
to share the collection of properties that define that family, gradience (how much a 
member belongs to the family) and centrality (central/good and non-central/bad 
members). In this same line of research, we find other linguists and philosophers such as 
Austin (1961) (relationships among meanings of words; analogy; primary nuclear 
sense
5
); Zadeh (1965) (study of categories with fuzzy boundaries); Lounsbury (1964) 
(kinship categories); Berlin and Kay (1969) (colour categories as an empirical 
establishment of Wittgenstein’s ideas of gradience and centrality), and the primary study 
of basic-categories of Brown (1958, 1965). However Eleanor Rosch (see references 
above) was the first to provide a general perspective on these problems. 
Following Rosch’s approach to categorisation, a cognitive methodology 
identifies the prototypical use of hand as that referring to a part of the body, and would 
treat the other uses of this lexical item as motivated, non-prototypical senses, related to 
the prototypical sense in a systematic way. In The hour hand of the clock, My life is in