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

CHAPTER 2: THE SEMANTIC FIELD OF SENSE 
PERCEPTION 
Perception verbs have supplied a rich field of research in linguistics: 
grammaticalisation (Heine et al., 1991), complementation (Horie 1993) and semantic 
change (Sweetser 1990). In this chapter, I analyse the different meanings that perception 
verbs can convey in the three different languages under investigation: English, Basque 
and Spanish. 
The organisation of this chapter is as follows: Section 2.1 describes the sources 
of the language material used in this thesis. Section 2.2 focuses on the prototypical 
physical meanings in perception verbs; their classification according to the semantic 
roles of the arguments that these verbs take and the hierarchies that have been 
established in this semantic field. Section 2.3 offers an account of the different non-
prototypical extended meanings that these verbs can convey from a cross-linguistic point 
of view. Section 2.4 summarises the results from previous sections. 
2.1. sources of the language material 
The linguistic material used in support and illustration of the discussions in the 
various parts of the study belongs to three different sources. 
(i) Monolingual and bilingual dictionaries. The dictionaries that I have made use 
of when writing this thesis are listed as a particular subgroup in the Bibliography 
section. These examples are followed by an abbreviated reference within brackets. 
(ii) Corpora of written Basque and Spanish
34
. Present-day Basque Reference 
Corpus (EEBS)
35
is the corpus used for Basque. EEBS is made available by the Basque 
Centre for Terminology and Lexicography (UZEI) and the Royal Academy of the 
34
English corpora, although consulted, have not been used as sources of language material. I 
decided not to use these corpora because I considered that the dictionaries available provided enough data 
to illustrate the main ideas put forward in this thesis. This was not the case in Basque and Spanish either 
because there were not enough dictionaries at my disposal or because these dictionaries did not offer 
enough examples. Although I am aware that the lack of English corpora may be an important imbalance in 
the case of a more quantitative kind of research (for instance, a study of the frequencies in the use of 
polysemous senses), it does not affect the results obtained in this case. 

---

B. Iraide Ibarretxe-Antuñano 
Polysemy and metaphor in perception verbs 
42
Basque Language (Euskaltzaindia). This corpus contains everything published in Basque 
from 1900 to 1995, and it is updated anually. It makes up to a total of 3,553,000 forms. 
Reference Corpus for Present-day Spanish (CREA
36
) is the corpus used for Spanish. 
CREA is provided by the Institute for Lexicography at the Royal Academy of the 
Spanish Language. This corpus contains literary, journalistic, scientific and technical 
texts, transcriptions from spoken language and from media broadcasts. A total of 200 
million words. These examples are also followed by an abbreviated reference within 
brackets. 
(iii) Examples that occur without any bracketed indication of the source have for 
the most part been constructed by me, occasionally on the basis of an utterance that I 
have seen or heard used. In addition, some of them have been taken from other linguistic 
studies, and whenever I can straightforwardly pinpoint the origin of such specially 
invented examples, I shall do so either in a note or in the accompanying text. I have 
always consulted native speakers concerning the naturalness of these examples. 
Being the result of a dictionary/corpus-based study, the analyses offered cannot 
represent one individual’s linguistic system. They can only reflect language use in a 
given community. Consequently, individual native speaker’s intuitions may deviate from 
the proposed analyses sometimes. These cases are indicated in the discussion. 
I would also like to point out that the main aim of this study is not to show how 
frequent or salient the meanings presented are in each language, but just the fact that it is 
possible to infer them. Therefore, I have not included any data on frequencies.