The Interplay of Synonymy and Polysemy

collostructional analysis. 
Their analysis begins with a construction. Then they investigate which lexical items are 
attracted or repelled by a slot or position in that construction (2003:214). The constructions 

---

40 | 
range in complexity and schematicity, from a single verb (where they explored its object) to 
the ditransitive construction (where they explored the verb slot). My second study employs 
the Stefanowitsch & Gries (2003) methodology, along with their statistical measures. 
For each 
throw
-verb, I found the ten most commonly occurring nouns within a seven-
word window. There is no standard window size (Stubbs 1995:11), though a 5-word window 
is often used (Gilquin 2010:172). A larger window was chosen for two reasons. First, since 
Spanish tends to have free word order, the elements in a sentence can vary in location, 
sometimes across a fairly large distance. Secondly, the frequency of each noun was quite low, 
so increasing the window size could increase the count. The result of this search was a list of 
32 words representing the top ten nouns that collocate with each verb (some of the nouns 
appeared in the top ten list for more than one verb).
The next step was to extract all co-occurrences of a noun from the list with each of the 
four verbs. For example, 
carcajada 
‘laugh’ appeared in the top ten list for 
lanzar.
This 
second step involved searching for all occurrences of 
carcajada 
‘laugh’ with each verb: 
arrojar, echar, lanzar
and 
tirar.
This was done for all 32 nouns. The resulting data had to be 
manually sorted, in order to ensure that the noun occurred in a specific semantic relationship 
with the verb. The study focuses on two relationships: one where the noun was the 
MOVANT
of the verb and another where the noun represented the 
DIRECTIONAL
.
Stefanowitsch & Gries (2003) use a