The Interplay of Synonymy and Polysemy

de un lado
de mi pequeña barca. 
‘The sea 
THREW
me from 
one side
of my little boat’ 
(CdE:19-F, Ventana Abierta al Mar) 
(130) …la 
ARROJÓ
al mar
. 
‘S/he 
THREW
her 
to the sea
.’ 
(131) …
ECHA
a andar
… 
(CdE:19-F, Arbusto y Piedra. ) 
‘S/he 
STARTS
to go
.’ 
As we can see, (129) includes the source and the goal, (130) describes the goal and 
the path and in (131) there is an inchoative verbal periphrasis and the intended goal. The 
possibility of expressing two 
DIRECTIONALS
in a single sentence has led some to propose 
separate types of 
DIRECTIONALS
. For example, Morante et al. (1998) propose a schema for 
(autonomous) motion verbs which includes both a source (marked by 
de 
‘from’) and a goal 
(marked by 
a
‘at/to’). Spanish FrameNet includes three frame elements for our single 
DIRECTIONAL
: the source, the path and the goal. Only the goal is a core element.
But it is also possible to conflate the different aspects of a trajectory into a single 
category, as I have done. Morimoto (2001) speaks of an 
argumento espacial 
‘spatial 
argument’
 
which includes any element of a trajectory. I follow Morimoto in conflating all 
elements that refer to a trajectory into a single participant. These double 
DIRECTIONALS 
only 
appear in 16 sentences in the 400-sentence corpus. Though they are possible, they are 
infrequent. Most of the time only one 
DIRECTIONAL
is expressed even though multiple 
DIRECTIONALS 
are allowed.

---

100 | 
Delving into the types of 
DIRECTIONALS 
in the
400-sentence corpus provides few 
statistically significant results. The specific prepositions and adverbs that appear vary greatly, 
but most are so infrequent that they can not be tested statistically. It is also possible to group 
prepositions into types, based on their usual function. Tests (see §4.3.2) show no significant 
variation among the verbs on the choice of preposition type.
A prepositional phrase also includes a nominal that describes a location or an object 
(Morimoto 2001:27). The nominals found in the corpus refer to a wide variety of objects, 
persons, places and directions, so little can be gleamed from the 400-sentence corpus. The 
collostructional analysis, though, does provide some specific information on common 
collocates in the 
DIRECTIONAL 
position, specifically the nouns as complements of the 
preposition. The results (§4.3.3) sometimes signal meaning differences between the verbs, 
but often they seem to signal collocational preferences
20
. That is, a verb attracts a specific 
DIRECTIONAL
because speakers have conventionalized the combination, but the meaning is 
entirely compositional.
The final data source on 
DIRECTIONALS
is from a questionnaire (§4.3.4). The 
questionnaire explores two issues that arose from the 400-sentence corpus and the 
collostructional analysis. The results of the questionnaire are compared to those from the two 
corpus studies in order to give a well-rounded picture of the data.
The discussion then turns to some additional issues regarding 
DIRECTIONALS
.
First I 
discuss
DIRECTIONALS
that contain nouns (or verbs) referring to time and actions. In some 
cases, these aspectual 
DIRECTIONALS 
form part of inchoative constructions, signaling the 
beginning of an action. The semantics and structure associated with these types of 
DIRECTIONALS
are discussed in §4.3.5.
Then I provide (§4.3.6) a semantic proposal connecting the central 
THROWING 
schema 
for 
tirar
with the meaning ‘to pull’. I argue that the type of structures that appear with this 
meaning (‘to pull’) can be understood if the phrase introducing the object pulled is both a 
MOVANT 
and a 
DIRECTIONAL
. This analysis also helps explain the meaning extension from 
throwing to pulling. I end this section with a summary of the findings (§4.3.7).
20 
Researchers often have different working definitions of collocations (See for example Alonso Ramos 1995 for 
a discussion). I am using the term 
collocation
as defined by Cruse (1986:40) “The term collocation will be used 
to refer to sequences of lexical items which habitually co-occur, but which are nonetheless fully transparent in 
the sense that each lexical constituent is also a semantic constituent.”

---