The Interplay of Synonymy and Polysemy

desde 35 metros
. 
was he who at the 28 minute threw violent from 35 meters 
‘It was 
he
, who at 28 minutes 
SHOT
with force 
from 35 meters
.’

---

60 | 
Sentences with no overt 
MOVANT
were rare in the data. There were 11 sentences each 
with 
echar 
and 
tirar
without a 
MOVANT
expressed. We saw in §4.1, that Spanish allowed the 
subject to be unexpressed in a sentence. The same is not true with objects. They tend to be 
expressed. Therefore, the lack of an overt 
MOVANT
is relevant. The distribution tables do not 
include the sentences with no overt 
MOVANTS 
because there are too few examples to be tested 
statistically, but they are discussed in §4.2.6.
Having discussed the elements that are considered 
MOVANTS
, I will now detail how 
these were classified. All 
MOVANTS
were divided into three categories: nonphysical 
(inanimate) movants, physical animate movants and physical inanimate movants. First of all, 
many uses of the verbs involve nonphysical elements: things that cannot be touched or 
handled, and therefore cannot be physically thrown. It is not the same to throw a candy (25) 
as it is to throw a laugh (26).
(25) 
…
todos
me
LANZABAN
caramelos…. 
(CdE:19-F, El nombre prestado) 
everyone CL.1
st
throw.3
rd
.pl candies 
‘
Everyone
THREW
me
candy’ 
(26) 
…
LANCÉ
una carcajada… 
(CdE:19-F, Fecundación fraudulenta) 
threw.1st a guffaw 
‘I burst out laughing’ 
Distinguishing between physical and nonphysical elements is especially helpful for 
studying light verb constructions (Vaamonde et al. 2010:1906). Sentences with physical 
objects as 
MOVANTS
tend to express literal motion, while the sentences with nonphysical 
elements are usually figurative or metaphorical in some way.
A majority of 
MOVANTS
fall into the physical category. In order to make meaningful 
distinctions within such a large category, I divided the physical category into animate and 
inanimate elements. Separating physical entities into animate and inanimate allowed me to 
separate the data into more categories, making the semantic analysis simpler. It also helped to 
isolate the pronominal uses from non-pronominal ones. 
This method exemplifies a bottom-up analysis (Gilquin 2010): I used the details of the 
corpora to guide my annotation process. The annotation choices I made are justified in at 
least three respects. Previous studies have used animacy as a parameter, as mentioned in §4.1. 
The distinction between physical and nonphysical elements is basically one of concreteness 
and abstractness, which has also been used in several other corpus studies (Gries & Otani 
2010, Liu 2010, Glynn 2009, Vaamonde et al. 2010). The innovation in this case was to use 
the parameters concurrently.

---

61 | 
Secondly, both Divjak & Gries (2008) and Liu (2010) mentioned adapting their 
annotation process based on the data they had available; this type of bottom-up analysis has 
been adopted before. Finally, the results of the statistical measures show that there is a 
statistically significant difference in the behavior of the verbs across these three variables. 
This indicates that the three variables are good predictors of the behavior of the 
throw-
verbs.
The results from the data are shown in table 3. The p-value for the entire matrix of 
data is 6.697E-14 (with a X-squared of 73.83 and a df of 6). The Cramer’s V is 0.3133, 
showing a medium effect size (King & Minium 2008:327-329). Each of the verbs shows a 
different preference in the type of 
MOVANT
. The table also shows how each verb varies from 
the expected values. The goodness-of-fit p-value indicates that the observed data differs from 
the expected. The arrows in the table show in which direction each verb differs (See §3.1). 
Table 3
. Distribution of 
MOVANT
types across all four 
throw-
verbs
14
Physical inanimate 
Physical animate 
Nonphysical 
GOF p-value
GOF p-value
GOF p-value 
arrojar 
50 ↑ 
6.94E-03 
23 ↓ 
1.72E-02 
27
echar 
24
37
27
lanzar 
11 ↓ 
7.47E-08 
41
48 ↑ 
2.04E-05 
tirar 
54 ↑ 
2.13E-06 
28
6 ↓ 
5.59E-06 
Beginning with physical inanimate 
MOVANTS
,
the number of examples for 
echar 
is not 
significantly different from the expected value (approximately 1/3 physical animate 
MOVANTS
).
Both 
arrojar
and 
tirar
have significantly more physical 
MOVANTS 
than expected. 
In fact, (over) half of their total sentences include physical inanimate 
MOVANTS
. 
Lanzar
in 
contrast has significantly fewer 
MOVANTS
in this category.
With physical animate 
MOVANTS
, only 
arrojar
has a significant result, having fewer 
physical 
MOVANTS
than expected. The remaining three verbs have approximately 1/3 
MOVANTS
. In the final category, nonphysical 
MOVANTS
, the number of examples for both 
arrojar 
and 
echar
fall within the expected values, with approximately 1/3 nonphysical 
MOVANTS
. 
Tirar
has significantly fewer nonphysical 
MOVANTS
, while 
lanzar 
has significantly 
more of this 
MOVANT
type. 
Table 3 shows that the verbs can be distinguished based on the 
MOVANT
type. The 
goal now is to understand why the verbs show this distribution. In the following, I will 
discuss each 
MOVANT 
type, describing the most common uses and meanings that the verbs 
14
The numbers for 
echar
and 
tirar
only add up to 88. There are 11 sentences with each verb that do not have 
an overt 
MOVANT 
(see