The Interplay of Synonymy and Polysemy

Figure 5
. Schema illustrating the functional synonymy of 
tirar 
and 
lanzar
when expressing 
the throwing of hits.
The two lower-level schemas are partially specified. They include a lexically filled 
verb, 
tirar 
in the left schema and 
lanzar 
in right schema. The second element of each phrase 
is not a specific word, but can be any noun that means ‘a hit’. I use capital letters to indicate 
word classes. 
The semantic relation between the two lower level schemas is captured by the higher 
level schema (the topmost box). The two lower-level schemas not only share a semantic 
similarity, but there is a structural similarity too. They are built up in similar ways and 
include similar parts. The first element of both is a verb that roughly means ‘to throw’, while 
the second element is a noun that means ‘a hit’.
tirar 
HIT 
THROW 
HIT 
lanzar 
HIT 

---

28 | 
The higher schema is connected to the lower schemas by elaboration links (solid 
lines). The higher schema only captures information shared by both lower schemas. The 
elaboration links mean that the lower schemas are instances of the higher schema. The 
schema for 
lanzar golpes ‘
to throw hits’ contains all the information in the higher level 
schema, plus potentially some additional more specific information.
For Pustejovsky (1995), Jackendoff (1990), Murphy (2003) and even Goldberg 
(1995), synonymy is not a part of the mental lexicon. If it is captured at all, it would likely be 
part of extra-linguistic knowledge as Murphy (2003:42) states. Though, I cannot give a 
detailed argument in favor of why synonymy should be part of the mental lexicon, some 
general benefits of my analysis can be given.
First of all, part of the knowledge of a speaker is knowing how and when to use a 
word (Kilgarriff 1997:95). A speaker must already independently know that 
tirar
can 
combine with nouns that refer to hits, and that 
lanzar
combines with nouns that refer to hits. 
To use each verb properly a speaker will have the lower-level schemas in figure 5. It is then a 
very simple process to arrive at the higher-level schema: if a speaker at some point in time 
notices the semantic and syntactic commonalities of the lower-level schemas he will derive 
the higher-level schema, thereby noting the functional synonymy. In this analysis synonymy 
can also be productive: a speaker can derive or generate a higher-level schema in a specific 
situation when necessary.
A further benefit is that synonymy is captured using the same tools used throughout a 
language. Figure 5 does not differ drastically from schemas capturing polysemy or other 
types of relations. Schemas and links between schemas are the main tools to capture all kinds 
of aspects of language,
including synonymy. This means too that synonymy can be captured 
implicitly. A speaker can have a representation of figure 5 in the mental lexicon, without 
necessarily labeling it as synonymy.