The Interplay of Synonymy and Polysemy

3.3 Experimental Study 
Several authors have recommended the use of multiple data sources when exploring 
linguistic phenomena (Labov 1972, Gilquin & Gries 2009, Gillham 2007). Arppe & Järvikivi 
(2007) is a recent example of a corpus study using experimental data to test their corpora 
results. Having this second data source allows one to corroborate previous corpus findings 
(Gilquin & Gries 2009:17, Arppe & Järvikivi 2007:27). It is also possible to use experimental 
data to explore phenomena that were not found in the corpus, since corpora will necessarily 
have a limited data set (Gilquin & Gries 2009:13). Using two types of data allows a more 
nuanced analysis of the phenomena. 
The primary goal in this project was to complement both corpus studies (the 400-
sentence corpus study and the collostructional analysis): since the results on 
DIRECTIONALS
were limited, an experiment allows me to expand on the available data and confirm specific 
hypotheses (Tummers et al. 2005:233). The limited data set in the main corpus study results 
in little data for 
DIRECTIONALS
. There are so many types of 
DIRECTIONALS
and so few 
repeated words or semantic classes that little can be gleamed from the 400-sentence corpus 
data. The collostructional analysis did show tendencies in the types of 
DIRECTIONAL
, but 
these cannot be confirmed in the 400-sentence corpus.
The experimental data explores two results, one from the collostructional analysis and 
another from the 400-sentence corpus. In the first case, the verbs seems to differ when it 
came to expressing direction, involving the nouns 
aire 
‘air’, 
piso 
‘floor’ and 
suelo 
‘ground’. 
The first can be seen as contrasting with the latter two in directionality. 
Lanzar 
is attracted to 
upward motion, while it is not attracted to either phrase representing downward motion. 
These results, though, cannot be confirmed in the 400-sentence corpus, due to the low 
occurrence of these specific nouns in the sample.
Another trend seems to arise concerning the optionality of 
DIRECTIONALS
. 
Echar
has 
fewer sentences in the 400-corpus that lacked a 
DIRECTIONAL
, compared to the other verbs. 
This result is statistically significant. This seems to indicate that a 
DIRECTIONAL
is a profiled 
aspect of this verb’s meaning and that 
echar 
has a stronger requirement on expressing this 
argument.

---

43 | 
To give a fuller picture of the role that the 
DIRECTIONAL
plays for these verbs, a test 
was set up. The purpose was to explore two specific issues: 1) directionality of motion and 2) 
optionality of a 
DIRECTIONAL
. The questionnaire was limited to these two issues in order to 
keep the test short, that way avoiding fatigue effects (Dörnyei 2003:14).