|
3.3.1 Materials and Procedure
The questionnaire consisted of 30 sentences. In each case, a verb was missing from
the sentence. The subjects were instructed to fill in the blank with the verb or verbs that
would best complete the sentence. The only options were the four
throw
-verbs. Respondents
were told that they could write several verbs if they felt that they were (almost) equally as
good in that context. Even so, more than half the respondents (19 out of the 30) only provided
one verb per test sentence.
The test was printed as a booklet with 8 pages. This format was chosen since it feels
and looks short and is therefore more welcomed by respondents (Dörnyei 2003:19). The
initial page was a cover page which included the instructions and three example sentences.
The 30 test sentences spanned the next 5 pages. There were four different randomized
versions of the test in order to avoid any effects from order of presentation. The final page
asked for some basic demographic information: age, education level and the place where the
respondent was raised. The task took 15-30 minutes to complete.
The 30 sentences in the test were divided into 4 types. Two sentences were included
as controls. These were always placed as the first and last sentence of the test. The control
sentences included relatively fixed phrases, where there should only be one choice for an
answer. The sentences from the tests are shown in (4) and (5); the correct verb is in
parentheses.
(4)
Tiene que __________ de la palanca.
(tirar)
‘You have to ___________ on the lever’
(5)
Ellos le ________ la culpa al jefe.
(echar)
‘They _________ the blame on the boss.’
The remaining 28 sentences tested the variation under consideration. Half of the
sentences explored the issue of directionality (to the air versus to the ground/floor) and the
other half explored the optionality of a
DIRECTIONAL
.
In the directionality set, 7 sentences expressed upward motion, 4 with the phrase
al
aire
‘to the air’ and 3 with other phrases:
al techo de la casa
‘to the roof of the house’,
encima del closet
‘on top of the closet’, and
al segundo piso
‘to the second floor’. The other 7
44 |
sentences of this set expressed downward motion; 2 with
al
suelo
‘to the ground’, 2 with
al
piso
‘to the floor’, and 3 with other phrases:
debajo del ropero
‘under the dresser’,
al pozo
‘to
the well’, and
abajo
‘down’.
The other set testing the optionality of the
DIRECTIONAL
was set up in a similar
fashion. Seven sentences did not have an overt
DIRECTIONAL
. These were paired with 7 other
sentences that had a
DIRECTIONAL
, always a goal. Examples include,
al lago
‘to the lake’,
en
la fuente
‘in the fountain’ and
en el cesto
‘in the basket’.
The seven sentence sets form a multi-item scale (Dörnyei 2003:33-34). The seven
sentences from the upward directionality set focus on the same target (upward motion), but
do so with different types of sentences, different objects thrown and different overall context.
The benefit of multi-item scales is that they “maximize the stable component that the items
share and reduce the extraneous influences unique to the individual items” (Dörnyei
2003:34). This means that it is more likely that the results we find are true of all sentences of
that type and are not due to the various specificities of a specific sentence.
All sentences had human subjects. The
MOVANTS
used in each subset varied. This was
done in order to try to ensure that the preference applied across different kinds of objects
being thrown. In the directionality set for example, each upward sentence was paired with a
downward sentence. This means that they were similar in word order, length and the type of
object that was thrown. The object was not always identical, in order to avoid too much
repetition in the test. For example, one sentence contained the word
sombrero
‘hat’ and its
pair had
gorro
‘cap’.
In three cases the semantic group of the
MOVANT
was repeated. Three pairs of
sentences included various clothing items. Two pairs of sentences included only a clitic, so
that the identity of the object was unknown (the equivalent of using
it
). Finally one pair used
the word
piedra
‘stone’ while another used
piedras
‘stones’, differing only in plurality.
According to the collostructional analysis (§4.3.3), the singular and plural forms patterned
differently. The expectation was that they would prefer different verbs in the test.
Some of the sentences were based on the 400-sentence corpus, though they were often
shortened or edited to make them clearer. Other sentences, though, had to be created. They
were still based on the corpus to an extent, by using the same types of
MOVANTS
and
DIRECTIONALS
as were found in the corpus. Priority was not given to repeating the sentences
from the corpus, because the focus was to derive data that was not available in the corpus.
The questionnaire’s cover page and the 30 sentences used in the test are given in appendix A.
Do'stlaringiz bilan baham: |
