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

4.1.1. DISCUSSION 
Based on the cognitive semantic principle that language is based on human 
understanding and experience of the world, Sweetser claims that the paths of semantic 
change are one-way and lead from the external (socio-physical) domain to our internal 
(emotional, psychological) domain. In the case of English perception verbs, the source 
domain is the vocabulary of physical perception, whereas the target domain is the 
vocabulary of internal self and sensations. These domains are structured by means of 
metaphor. Finally, Sweetser suggests that these metaphorical mappings are not particular 
to one language, but constitute a cross-linguistic phenomenon. 
In Chapter 2, I analysed the semantic field of sense perception in three different 
languages: English, Basque and Spanish. Based on these data, it can be said that – as 
Sweetser predicts – many of the semantic extensions in this field are shared by the three 
languages analysed, despite the fact that in some cases, each language had even more 
extended meanings that are not present in the other two. 
In the previous section, I presented Sweetser’s metaphorical mappings for 
English perception verbs. These metaphorical mappings were: 
VISION
→→→
KNOWLEDGE 
HEARING
→→→
HEED
→→→
OBEY 
TASTE
→→→
LIKES / DISLIKES 
TOUCH
→→→
FEELINGS 
SMELL
→→→
DISLIKEABLE FEELINGS 
As shown in Chapter 2, these are not the only possible metaphorical mappings in 
sense perception verbs. What Kövecses (1995, in press) calls the ‘metaphorical scope’
105
of sense perception verbs is much larger than that proposed by Sweetser; even in cases, 
such as smell, where Sweetser claims that it “has fewer and less deep metaphorical 
connections with the mental domain than the other senses” (1990: 43). Smell, for 
105
“The range of the application of particular source domains to particular target domains” 
(Kövecses 1995: 316). 

---

B. Iraide Ibarretxe-Antuñano 
Polysemy and metaphor in perception verbs 
115
instance, is not only mapped into dislikeable feelings, but also into meanings such as ‘to 
investigate’, ‘to suspect’ and ‘to guess’ (Ibarretxe-Antuñano 1997). 
Another factor that is not discussed in Sweetser’s analysis is that in some cases 
the extensions of meaning do not shift towards more abstract or metaphorical domains, 
but they remain physical. For instance, in the case of the meaning ‘to affect’ in the sense 
of touch illustrated in (1). 
(1)
Just don’t touch anything in my room 
As discussed in Section 2.3.3, the meaning of (1) does not only imply physical 
contact but also a change of state. In (1), the person who utters this sentence does not 
want the other person to change the state / position of anything in his / her room. 
Another example of this meaning can be seen in (2). 
(2)
Blackfly touched the flowers (
COL
) 
This example states that a type of insect physically touched the flowers and since 
this insect is harmful to them, the flowers were physically affected. 
As seen in Chapter 2, semantic extensions from physical perception onto a 
physical domain are not only found in the sense of touch. In the sense of smell, we have 
the meaning ‘to trail something’; in vision, ‘to visit’ and ‘to receive’ among others. 
These physical extensions of meaning are not taken into account in Sweetser’s analysis. 
Metaphor is the cognitive device that Sweetser proposes as the structuring means 
to link the physical prototypical domain to more abstract domains. However, metaphor 
cannot explain or structure these physical extended meanings, because the mappings in 
these cases are not between a physical and an abstract domain, but between a physical 
and another physical domain. In Chapter 6, I will propose the process called Property 
Selection as a possible solution for this problem. 
Another point that remains unanswered in her analysis is why certain source 
domains get mapped onto certain target domains; the reasons why it is possible to say 
Mary smelt the joke but not Mary touched the joke. In the Cognitive Linguistics model, 
the bodily basis of these verbs should be taken as the motivation for such semantic 
extensions, but this approach does not discuss what it is exactly the bodily basis in the 
field of perception. In Chapter 5, I will characterise the five senses in terms of properties 

---

B. Iraide Ibarretxe-Antuñano 
Polysemy and metaphor in perception verbs 
116
and state that these properties can be considered the bodily basis of such semantic 
extensions. 
A related issue is also the question of how much of the source domain is mapped 
onto the target domain in these extensions of meaning. In other words, what and how 
much information from the source domain is selected and transferred onto the target 
domain.
It is commonly agreed among metaphor researchers that not everything from the 
source domain gets mapped onto the target domain. Lakoff and Johnson (1980: 52) call 
this partial map of the structure of the source domain the ‘used’ part of metaphor. 
However, in order to constrain metaphorical mappings it is not enough to say that there 
is a selection of the source domain, it is necessary to show exactly what it is that is