The Interplay of Synonymy and Polysemy

lanzar
versus 
arrojar
in the 
SOUNDS
schema) or a higher level of productivity (
echar
compared to 
lanzar 
and 
arrojar
in the 
LOOKS 
schema). There is then a delicate balance between sharing semantic space and each verb 
claiming its own areas.
A third result of this study is a view of polysemy as a network of related meanings. In 
the conception of language that seems to me to be the most compatible with the results 
obtained in this thesis, a word may have one (or more) central or prototypical meanings. Then 
other meanings and senses are derived through metaphor, metonymy, generalization and 
other semantic processes. The type and degree of motivation is extremely varied: for 
example, there seems to be very little motivation for the phrase 
echar
de menos 
‘to miss’, 
while a phrase such as 
tirar golpes 
‘throwing hits’ shows many semantic similarities to the 
central 
THROWING 
schema. But motivation helps a speaker make sense of their language. 
Language need not be simple a list of all the meanings of a word, as the traditional view 
contends. Speakers acquire new uses and extensions from the central or primary meaning of 
words (Bosque 2004:CXII). The semantic analysis of the various constructions in this study 
sought to provide plausible motivations for the (unpredictable but not arbitrary) meaning 
extensions.

---

152 | 
I posit that the 
THROWING 
schema is the central schema for all four verbs (though 
there are certainly reservations about that choice; see the introduction to Chapter 4). The 
characteristics of this central schema are often maintained in the different meaning 
extensions. For example, the majority of the data contain animate 
INITIATORS
, across a variety 
of meaning extensions. Even when the 
INITIATOR
is not animate, the position retains some 
traits of animates. In the majority of cases, the 
INITIATOR
,
be it an animate being or a 
document, is conceptualized as a source and as a causer. Likewise, the central notion of 
caused motion, is maintained in many of the meaning extensions (
HITS
,
BLAME
,
SOUNDS
,
throwing oneself, etc.) even when the object that is moved is not a physical object. 
Characteristics of the 
THROWING 
schema, including the participant roles and part of its 
meaning are maintained in the various meaning extensions. 
As a final point, this study has also shown that various linguistic phenomena can be 
described and understood using constructions, links and correspondences, all derived by a 
schematization process that traces the commonalities between different instances of language 
in use. Some schemas are mostly lexically filled, such as the 
SHED LIGHT 
schema that had a 
fixed verb, 
MOVANT
and preposition in the 
DIRECTIONAL
, while other constructions are 
completely schematic, such as the semantically-schematic idiom schema which only requires 
a verb and a 
DIRECTIONAL
or 
MOVANT
of some sort. Sometimes a schema is mostly 
compositional (all the collocational schemas) while other schemas are almost completely 
non-compositional (the schema for 
echar de menos 
and 
echar en falta 
‘to miss’). Some 
schemas can be highly productive (LVC constructional schema) while other schemas are 
limited in their application (the 
SLAM 
(
DOOR
) schema). All these phenomena are captured as 
constructions, but they still are capable of displaying their own unique traits and 
characteristics. All are constructions with their own specific flavors and details. 
This project also points to future areas of research. Little focus has been placed on the 
pragmatic or connotational details of the data under study. Still, I argue that we can notice 
some important differences between