![3D"Interfaz](3D"1.Buyse_archivos/image001.jpg")
![3D"Cuadro](3D"1.Buyse_archivos/image002.gif")
The impact of motivational
didactic interventions on the writing process and its output: the case of
descriptive texts
El impacto de
intervenciones didácticas motivacionales en el proceso y el producto=
de
redacción: el caso de los textos descriptivos
Kris Buyse
KU Leuven & Nebrija University
kris.buyse@kuleuven.be
ABSTRACT
In this article we describe =
the
impact of motivational didactic interventions on the process and the output=
of
writing descriptive texts in Spanish as a Foreign Language by Dutch speakin=
g students
of the 3rd Bachelor year of the Applied Linguistics curriculum at
the Flemish KU Leuven university.
The texts were stored and analysed in th=
e Aprescrilov learner corpus application. The didactic
intervention was tested in a pretest – test – posttest setting,
both with consecutive and with simultaneous groups: the results of the year
2006-2007 (without intervention) were compared with the ones of 2007-2008 (=
with
intervention), as well as the results of two groups in 2011-2012 (experimen=
tal
versus control group). The results of quantitative and qualitative tests sh=
ow
an important impact on both process (motivation, enjoym=
ent
and perceived improvement) and output (lexical richness and dynamic style) =
of
the writing of the descriptive texts.
Keywords: error analysis,
interlanguage, learner corpora, task-based language learning, motivation
RESUMEN
Palabras clave: análisis de errores, interleng=
ua,
corpus de aprendices, aprendizaje de lenguas basado en tareas,
motivación
1.
Introduction
The 3rd
Bachelor curriculum at the department of Applied Linguistics of the Flemish
University of KU Leuven aims at teaching students to produce in the foreign
language –in this case Spanish- qualitative texts in the major text
genres at a general language level of B2 of the Common European Framework of
Reference for Languages. Since our learner corpus Apre=
scrilov
(see section 3 for more details) stores all texts of students of Spanish as=
a
Foreign Language (SFL) and tags them a.o. by ge=
nre
and error category, it enables us to discover certain tendencies in the out=
put
of the learners. In this way the error analysis of the descriptive texts
produced between 2003 and 2006 learnt us that one of the most frequent prob=
lems
in those texts is what we call ‘lexical poverty’: (over)use of =
the
verbs ser, est=
ar,
hay (all hyper frequent verbs meaning ‘to be’) and tener (&#=
8216;to
have’) – instead of a variety of semantically ‘richerR=
17;
variants-, while this error category never enters the top 10 of the most
frequent problems in other text genres. A second problem that emerges from =
the
analysis of these texts is their monotone, static character, which is obvio=
usly
a consequence of the static “to be” and “to have”
verbs, but also of the general point of view chosen by the authors, i.e. st=
atic
instead of dynamic. Thirdly, in the writing portfolios that learners are as=
ked
to hand in after each writing assignment, students showed low motivation wh=
en
writing the descriptive texts, with an average score of 2.3 on a scale of 1=
to
5.
Therefore =
we
adopted for the 2006-2007 edition of the course a TBLT approach, with a str=
ong
emphasis on real-life, meaning-making tasks, as well as focus on creativity,
learner autonomy, motivation and team work, with=
pre
and post tasks. In the pretasks, focus was set =
on
lexical richness and dynamic style. As students with a =
task
based motivation are expected to consider tasks as challenging and
valuable learning experiences, and therefore perform them in a more structu=
red
and thorough way (Vandekerckhove, Vandergraesen
and Cruysweegs, 2009), we expected this approac=
h to
entail a more motivating, meaningful and empowering writing process as well=
as
a better output.
In the
following sections we will go into the details of the study. In Section 2 we
will provide a brief overview of the general theoretical framework and the =
key
notions used in this study. Section 3 will outline the architecture of the =
Aprescrilov corpus, Section 4 the methodology and Sec=
tion 5
the results of the present study.
2.
Theoretical framework
The Aprescrilov corpus and Aprescril=
ov-based
case-studies performed to date can be located at the interface of five
frameworks, viz. Error Analysis, Interlanguage studies, (Learner) Corpus
Research, Task Based Language Learning and Teaching, and Studies on Motivat=
ion
and Learning. In what follows, we will present succinctly these frameworks.=
2.1 Error Analysis
Error Anal=
ysis
(EA) has become relevant thanks to the work of Corder<=
/span>
(Corder, 1981). It is, essentially, a scientific
procedure whose objective is to determine the nature, cause and consequence=
s of
errors made by non-native language learners when learning/acquiring a forei=
gn
language.
Additional=
ly,
the final objective of EA is to draw conclusions from the identification, <=
span
class=3DGramE>description and explanation of errors, with the aim of
proposing didactic procedures designed to help avoid those errors in the
interlanguage (IL) of non-native speakers (see Section 2.2). IL and EA are =
thus
closely related.
Interestin=
gly,
in view of severe criticism of EA for paying too much and exclusive attenti=
on
to learner’s errors and deficiencies instead of also analyzing
learners’ strengths, EA has shifted focus “from mere error anal=
ysis
to the analysis of performance in its entirety” (Callies
et al., 2015: 166). Although errors are obviously more likely to draw
analysts’ attention, we also keep track of learners’ successful
uses of a particular linguistic item in our Aprescrilo=
v-based
case-studies.
2.2 Interlanguage
The
Interlanguage framework (IL) was proposed by Selinker<=
/span>
(Selinker, 1972). He hypothesized that in addit=
ion to
the mother tongue and the learner’s foreign language, there is a sepa=
rate
(idiosyncratic) linguistic system in learning processes, viz. the
interlanguage. This system is based on observable output which results from=
a
learner’s attempted production of a Target Language (TL) norm. Selinker called this linguistic system
“interlanguage”.
IL has also
come to be associated with another scholar, viz. Granger, in
particular since her seminal paper on Contrastive Interlanguage Anal=
ysis
(Granger, 1996). Granger proposes contrastive IL as the best methodological
approach to Learner Corpora. This method compares the target language or L2
with the learner’s native language or L1. Preferably, it also include=
s a
comparison between distinct types of learners, with reg=
ard to
their proficiency level and mother tongue.
2.3 Learner Corpus Research
Research in
Learner Corpus Research (LC(R), see a.o. Grange=
r,
2009) has only recently been recognized as a worthy field or subdomain with=
in
Corpus Linguistics2. Learner Corpora are generally defined as
“systematic collections of authentic, continuous and contextualized
language use (spoken or written) by L2 learners, stored in electronic
format” (Callies and Paquot,
2015: 1). Although the framework initially mainly focused on English, LC of
many languages are currently freely available3.
As is the =
case
of corpora of native speaker productions, corpora diverge largely (among ot=
hers
as to size, text genres, year, etc.), usually in function of the corpus
designer’s research interest.
The
introduction of corpus linguistics into language teaching makes it possible=
for
the results of EA and IL studies not to be merely intuitive, descriptive
and structuralist, but objectively based on solid data, in this case data f=
rom
a LC. However, Hasko (2013: 4-5), among others,
criticizes the lack of progress in establishing strong, bidirectional links
between LC and Second Language Acquisition (SLA) and Foreign Language Teach=
ing
(FLT), due to (i) a shortage of longitudinal st=
udies
that would allow scholars to establish causality in interpreting corpus dat=
a analyses,
and even better, test the efficiency of pedagogical adjustments and (ii) the
fact that it is more typical of LC analysis to describe learner language ra=
ther
than attempt to explain it. The Aprescrilov pro=
ject
intends to move a step forward in applying LCR to FLT (Cruz Piñol,
2012). Although it has been primarily designed to improve the didactic mate=
rial
and habits in SFL-teaching in Belgium, it allows to analyze the possible
interference of more than one language (L1, but also other L2) in foreign l=
anguage
learning (Buyse, Delbecque=
and Speelman, 2009).
While both=
EA
and IL originated in the 1960-1970s, LCR emerged at the turn of the 1990s (=
Callies and Paquot, 2015:=
1). EA
and IL initially included useful studies on several languages, but from the=
90s
onwards they were used predominantly in the field of English as a
Second/Foreign Language (ESL/EFL). Their use for SFL remained limited. LCR
researchers were also primarily concerned with EFL at the outset, but rapid=
ly
infected SLA researchers of many languages with their enthusiasm. In the li=
ght
of promising benefits of LC for language learning and teaching, the field is
increasingly gaining in interest, but is still considered to be ‘on t=
he
move’ or ‘under construction’ (Calli=
es
and Paquot, 2015; Callies<=
/span>
et al., 2015). Hence, with respect to English L2 learner corpus research,
Spanish L2 learner corpus research is gradually bridging the gap.
2.4 Task Based Language Learning
Task Based
Language Learning and Teaching (TBLT) is an approach in which learning revo=
lves
around the completion of meaningful tasks. In the TBL approach (Ellis, 2003=
),
the main focus is the authentic use of language =
for
genuine communication. Tasks can be real-life situations or have a pedagogi=
cal
purpose. They should provide opportunities for students to exchange informa=
tion
with a focus on meaning, and have a clear purpos=
e:
learners should know the outcome they are expected to produce when they fin=
ish
performing the task. The outcome may vary, and usually results in an outcom=
e that
can be shared with more people.
2.5 Affectivity in Language Learning and Teaching
Affect in
learning and teaching is nowadays seen as more effective than a purely
cognitive teaching approach (Arnold, 2011; Dewaele,
2005, 2015). As Stevick (1980) states: “S=
uccess
[in language learning] depends less on materials, techniques and linguistic
analyses and more on what goes on inside and between the people in the
classroom” (p. 4). The “inside” refers to individual lear=
ner
factors such as self-concept, anxiety, learner styles, but also to
teachers’ own personal development.&=
nbsp;
The “between” is about the relational aspects which deve=
lop
between the participants in the classroom – between students or betwe=
en
teacher and students - or possibly between learners and the target language=
and
culture (Dewaele and MacIn=
tyre,
2014). Positive affect can provide invaluable support for learning just as
negative affect can close down the mind and prev=
ent
learning from occurring altogether (MacIntyre, =
Gregersen and Mercer, 2016).
This expla=
ins
the growing interest of the intertwining of emotion and cognition both with=
in
an institutional context and in extra-institutional contexts where multiple
languages and cultures meet (Berdal-Masuy and <=
span
class=3DSpellE>Pairon, 2015; Dewaele, 20=
18).
In this st=
udy
we investigated the possible influences of didactic interventions both on t=
he
skills and on the emotions of students when writing (descriptive) texts in =
SFL,
more particularly on perceived L2 improvement and enjoyment (Nakamura, 2018=
).
3. The
corpus
The
architecture of Aprescrilov – which stand=
s for Aprender a Escribir en Lovaina ‘Lea=
rning to
write in Leuven’ (see Buyse, 2011) –=
; is
based on the models proposed by Díaz-Negrillo and Fernández
Domínguez (2006), Granger (1996) and others. It is an online corpus =
with
restricted access for researchers4. It allows (i)
to perform quasi-longitudinal studies of writings by SFL-students, (ii) to
objectively determine the interference of more than one language, (iii) to =
take into account both task and learner variability (c=
f.
Granger 2015; Tracy-Ventura & Myles, 2015), and, eventually, (iv) to
successfully implement conclusions from studies based on LC in the developm=
ent
of new didactic material (cf. Fernández Pereda, Buyse
and Verveckken, 2014).
Aprescrilov consists of two =
subcorpora. The first, Aprescril=
ov
I, is composed of 2700 texts written in the academic years 2004-2010 by
students of Spanish Linguistics and Literature at the Faculty of Arts of th=
e KU
Leuven and of Applied Linguistics at the Lessius Hogeschool (now “KU Leuven @ Antwerp”). T=
he
compositions were written by 1st, 2nd and 3rd ye=
ar
Bachelor students of these two institutions and have been digitally marked =
with
the same customized version of the Markin progr=
am
(see a.o. Buyse and
González, 2013). This “button set” allows systematic mar=
king
of problems or ‘errors’ in the texts, as well as of positive
aspects. The annotations cove=
r all
components of writing – from spelling to discourse structure,
punctuation, morphology, morphosyntax, pragmatics, lexicon, etc.5
![](3D"1.Buyse_archivos/image004.jpg")
Figure 1. Search interface of the Aprescrilov
corpus.
The operat=
ing
extension of the corpus results in an ongoing creation of the second
(sub)corpus, A=
prescrilov II, composed of Spanish texts from
Dutch-speaking students of the same institutions, Spanish texts from Dutch
students of the Radboud Universiteit (Nijmegen, the Netherlands) and from
French-speaking Walloon students of the UCL (Louvain-la-Neuve,
Belgium) written in the same period (viz. the academic year 2013-2014) and =
on
comparable subjects. These texts are gathered in an online corpus with the =
same
interface as Aprescrilov I. The corpus is curre=
ntly
being expanded both with new versions of the same (writing) tasks in the
following academic years with new students, and with writing tasks of stude=
nts
in higher years.
Texts were
annotated with our customized version of Markin=
(see
Section 3) by annotators who had been trained in order =
to
obtain a systematic treatment of the same errors (inter-rater reliability t=
est
for Kappa =3D 0.85 with p < 0.001). The Aprescrilov=
corpus allows to perform queries per level, per year and per assignment, and
the integrated information on words per text enables us to convert the abso=
lute
figures into relative ones.
With regard to the aforemention=
ed
shortcomings of LC studies (see 2.3), it is worth mentioning that Aprescrilov varies as to task, learner and other aspe=
cts:
(1)&=
nbsp;
Task: genre,
assignments vs test.
(2)&=
nbsp;
Learner: it inclu=
des
texts written both by beginners and by advanced learners (viz. successful
learners in higher years, allowing pseudo-longitudinal research). Since 201=
3,
metadata on the authors have been available, more precisely on their native
language(s), the (amount of) contact with the Spanish language and their
proficiency level of other foreign languages. Overall, the corpus contains
texts of three main groups of students: (1) Dutch-speaking students in Flan=
ders
(Dutch as L1, French as L2); (2) Dutch-speaking students in the Netherlands
(Dutch as L1, English as L2); (3) French-speaking Walloon students (French =
as
L1, Dutch as L2).6
(3)&=
nbsp;
Other variables:
course, academic year, institution.
4.
research design and Methodology
As already
mentioned in the introduction, the analysis of the descriptive texts of
(1)&=
nbsp;
“lexical poverty”, viz.=
the
(over)use of the verbs ser, estar, hay and tener, and a lack of semantically ‘richer’ <=
span
class=3DGramE>variants
(2)&=
nbsp;
its monotone, static character, as =
result
of the static “to be” and “to have” verbs, but also=
of
the general point of view chosen by the author, i.e. static instead of dyna=
mic.
(3)&=
nbsp;
low student motivation when writing=
the
descriptive texts, with an average score of 2.3 on a scale of 1 to 5.
Problems t=
wo
and three are all the more striking, since one y=
ear
earlier the authors of these texts, during the writing course of the 2nd
bachelor year, had received a two hours class on “lexical povertyR=
21;
and been trained to avoid it with exercises on sentence and paragraph level
(although not on text level), such as the following two (underlining is our=
s):
(1)&=
nbsp;
Rewrite:
Me
gustó mucho el pueblo; pero lo que más me
gustó del pueblo fue la plaza porticada. Sin embargo, a mis padr=
es les
gustó más la torre de la iglesia del pueblo.
(“I like=
d
the village a lot; but what I liked most about the village=
was its arcaded square. Nevertheless, my parents liked more the towe=
r of
the church of the village”).
(2)&=
nbsp;
Find an alternative for the followi=
ng
verbs:
¿Por
qué están los coches en las aceras?
(“Why ar=
e the
cars on the sidewalks?”)
En esa tumba
Esa ley tien=
e
40 artículos muy extensos. (“That law has 40 very extensive
articles”)
(…)
Despite of
this training, one year later, the same students (now in the 3rd Bachelor),
when writing descriptive texts, seem to have forgotten the knowledge and sk=
ills
they had been training on sentence and paragraph level, as can be deduced f=
rom
the following example, with 18 instances of the “poor” verbs ser, estar=
, hay
and tener<=
/span>
on a total of 310 words (italics and underlining are ours):
La
casa de mis sueños
Aunque
todavía vivo con mis padres, de vez en cuando ya pienso en mi propia
casa. Me gusta leer los folletos publicitarios y los catálogos con
muebles modernos, pinturas nuevas, aparatos electrodomésticos... En
resumidas cuentas, me hace gracia imaginarme cómo será=
mi
casa futura.
En
realidad, mi sueño es vivir en la casa –o mejor dicho la
hacienda– de mis padres que tiene un patio, un establo, una ca=
sa
espaciosa y un jardín grande. Después de 18 años
aún no está renovada totalmente. Además, quiero
modificar y sobre todo modernizar mucho. Prefiero tener un interior
moderno y acogedor que incluso sea fácil para limpiar.
Creo
que la cocina y la sala de estar son los cuartos más importan=
tes
puesto que se está mucho tiempo allí. Me gusta cocinar=
, de
modo que una cocina grande y sobre todo práctica es
imprescindible. Voy a pintar la sala de estar de colores cáli=
dos,
pero no demasiados oscuros. Además, una chimenea y un parquet oscuro en el rincón para sentarse t=
ienen
que aportar al ambiente acogedor.
Los
otros cuartos de la planta baja son un estudio donde quiero colocar
estantes llenos de libros, un cuarto de los niños, un cuarto con la
lavadora y la secadora, y claro, un servicio. También voy a comprar =
una
sauna, ya que es bastante sano y no me gusta el frío.
En
el piso de arriba hay 4 habitaciones y un pequeño cuarto de
baño. Me gustaría tener un cuarto de baño mayor=
con
baño, dos lavabos y una ducha grande. Voy a tener 2 dormitori=
os
para los niños y uno para los huéspedes. En mi propio dormito=
rio
quiero otro cuarto de baño y un guardarropa muy grande.
No
voy a tener tiempo para cuidar del jardín ni de un huerto, sin
embargo, quiero tener un jardín con un césped y muchas
flores que florecen en diferentes estaciones.
In conclus=
ion,
students do not link their knowledge on “lexical poverty” to the
genre of descriptive texts.
Therefore =
we
redesigned the introductory class on this genre, including guidelines on ho=
w to
write this type of descriptive texts. The introduction “new styleR=
21;
contains a comparison between a descriptive text of a Spanish writer and an
(anonymized) one by a former student.
A first pr=
e-task
consists in listing and counting the instances of the four aforementioned
‘poor’ verbs in both texts, as well as the lexical and grammati=
cal
alternatives used by each author in order to enh=
ance
the variation in the students’ productions. Together with the teacher
they draw the following conclusions:
· =
Text of professional writer: Ser 6, Estar 5, Hay 4, Tener 3 (Total =
=3D
18/500, or 3.4%, versus 6% in student text)
· =
Lexical alternatives:
o
for Estar
> pasar muchas horas (“to spend a =
lot of
hours”), disfrutar de (“enjo=
y”),
vivir en (̶=
0;to
live in”), quedarse (“to
stay”), cocinar [vs estar
en la cocina] (=
“to
cook” [vs “to be in the kitchen”]), jugar
(“to play”)…
o
for estar/hay
> nos encontramos con (“we
find”), irradiar (“irradiate=
”),
caer (“to fall”), saltar a la vista (“to hit in the eye”=
;),
llegar a (“to arrive at”), <=
span
class=3DSpellE>entrar en (“to =
enter”)…
o
…
· =
Syntactic alternatives:
o
Postponed adjective / participle: rodeado
de (“surrounded byR=
21;)
o
Relative clause: [est&aacu=
te;
claro >] lo que salta a la vista ([“it is obvious that”] &g=
t;
“what hits in the eyes is”)
o
Pronominal verbs: [est&aac=
ute;
dividido en] > s=
e divide
en ([“it
is divided into”] > *”it divides itself in”)
o
(“one can…”=
/
“we can…”)
A second pre-task requires the student to rewrite a (poor) descripti=
ve
text, aiming at more variation and dynamics, using techniques such as
“travelling”, where the author moves the angle like a moving ca=
mera
does in movies. In that exercise he discovers formulas such as the followin=
g:
o&nb=
sp;
Dynamics: bienvenida (“welcome”), como puedes ver
(“as you can see”), empecemos con (“let
us start with”) ...
In order to measure the effe=
ct
of that didactic intervention, we designed the following experiments.
1. 2006-2007 (N =
=3D
21): pretest – test – post test wit=
hout
control groups, texts of ±300 words; this experiment was repeated in
2007-2008 (N =3D 20)
a.
“pretest” (assignment): descriptive text without new
instructions
b.
didactic intervention: new introduction (cf. supra)
c.
“Test” (assignment)
d.
“Post test” (test, 2 months later)
2. 2011-20=
12:
pretest – test – post test with con=
trol
group, students being randomly assigned to experimental vs control group (N=
=3D
10 vs 11)
a.
“pretest” (assignment): descriptive text without new
instructions
b.
didactic intervention: new introduction, only for experimental group=
c.
“Test” (assignment)
d.
“Posttest” (test=
,
2 months later)
All activi=
ties
were assessed by 2 independent raters.
Despite of=
the
very homogeneous character of the population, the students were asked to
deliver the following metadata: gender, age, nationality, mother tongue,
±bilingual, education, earlier Spanish courses, earlier Spanish
immersion period(s), other languages.
Participan=
ts
were also asked to answer a short list of questions in a portfolio. Besides=
procedural
items such as “did you look at the model of the text genre before
starting to write?”, “did you brainstorm on the topic
beforehand?”, “did you work out a structure for the text before
starting to write?”, students were asked to score their motivation, t=
heir
enjoyment and their perceived improvement when writing the descriptive text=
on
a scale from 1 (not motivating / enjoying / improving at all) to 5 (extreme=
ly
motivating / enjoying / improving).
Our resear=
ch
hypotheses were the following.
(RH1) The didactic intervention will have a positive and significant
effect on lexico-grammatical richness and on the
dynamic and creative character of the descriptive text. Hence, the variables
“lexico-grammatical richness” and
“dynamic expression” will increase, as well as the general score
for the test. The variables “lexical poverty” and “static
expression” will decrease.
(RH2) The level of lexico-grammatical ri=
chness
and textual dynamics will drop significantly in the post test (due to time
lapse and different conditions of the activity).
(RH3) The student will be more motivated when writing texts of this =
text
genre.
5.
Results and discussion
Table I sh=
ows
the results of the first experiment (with consecutive groups). They seem to
confirm RH1: the general score=
s for
the years 06-07 and 07-08, respectively, improve with 14,5 to 20% between
pretest and test, and accordingly for lexico-gr=
ammatical
richness and dynamic expression (increase of occurrences with 94 till 115%)=
, on
the one hand, and lexical poverty and static expression (decline of 38 till
125%), on the other. A series=
of
paired samples t-tests revealed that the differences between pretest and te=
st
are significant: for 2006-2007 there was a significant difference in the sc=
ores
for pretest (M=3D11.3, SD=3D1.62) and test (M=3D14.2, SD=3D1.17) conditions;
t(20)=3D-16.02, p =3D 0.000; for 2007-2008 there was a significant differen=
ce in
the scores for pretest (M=3D11.0, SD=3D1.21) and test (M=3D15.0, SD=3D1.39)=
conditions;
t(19)=3D-17.69, p =3D 0.000.
RH2, on its turn, does not seem to be confirmed, as t=
here
is only a minor decrease of the general score when comparing test and postt=
est:
for 2006-2007 there was no significant difference in the scores for test
(M=3D14.2, SD=3D1.17) and posttest (M=3D14.1, SD=3D1.18) conditions; t(20)=
=3D1.45, p =3D
0.162; for 2007-2008, there was no
significant difference neither in the scores for test (M=3D15.0, SD=
=3D1.39)
and posttest (M=3D14.5, SD=3D1.76) conditions; t(19)=3D-16.02, p =3D 0.016.=
The same h=
olds
for the more specific parameters (lexical poverty, lex=
ico-grammatical
richness, static expression and dynamic expressi=
on):
the differences between pretest and test are significant (p<0.01), the o=
nes
between test ant posttest are not (p>0.01).
A Spearman
rank correlation analysis revealed no highly significant influences of the
variables listed in the metadata document: criteria such as gender, age,
nationality, mother tongue, other languages, type of education showed no
correlation with the results (r=3D< .10); even criteria with stronger po=
ssible
influences such as ±bilingual upbringing, earlier Spanish courses and
earlier Spanish immersion period(s) showed only small possible influences
(r=3D.10 - .29).
|
|
General score /20 points |
Lexical poverty /300 words |
/300 words |
Static expression /300 words |
Dynamic expression /300 words |
|||||
|
|
0607 |
0708 |
0607 |
0708 |
0607 |
0708 |
0607 |
0708 |
0607 |
0708 |
|
Pre t |
11.3 |
11.0 |
10.3 |
12.7 |
2.8 |
5.4 |
35.3 |
33.8 |
2.6 |
4.3 |
|
Test |
14.2 |
15.0 |
2.6 |
3.2 |
23.1 |
24.3 |
10.7 |
8.7 |
25.3 |
27.3 |
|
Post t |
14.1 |
14.5 |
3.6 |
3.4 |
20.3 |
24.1 |
12.2 |
8.9 |
23.7 |
26.8 |
Table I. Resu=
lts
of quantitative analysis. Average scores of 2006-2007 and 2007-2008<=
/a>
The analys=
is
of the portfolios confirms RH3=
: the
writing of the descriptive texts is now scored as highly motivating (averag=
es
of 4 and 4.2 on a scale of 5 vs the average of 2.3 in the portfolios of ear=
lier
years before the didactic intervention), highly enjoyable (averages of 4.1 =
and
4.0, vs 2,4 before) and with a highly positive effect on the perceived
improvement (averages of 4.1 and 4.3, vs 2,6 before).
Table II s=
hows
the results of the second experiment (with experimental and control group).
Also in this experiment the general scores improve between pretest and test=
, with
30% for the EG and 11.5% for the CG, and accordingly for lexico-grammatical
richness and dynamic expression (increase of occurrences with 115 till 116%=
for
EG vs 6 to 21% for CG), on the one hand, and lexical poverty and static
expression (decline of 45 till 127% for EG vs 3.5 to 8.5 for CG), on the ot=
her.
A series of paired samples t-tests revealed that the differences between
pretest and test are significant in both groups: for EG there was a signifi=
cant
difference in the scores for pretest (M=3D10.5, SD=3D0.97) and test (M=3D16=
.5, SD=3D1.50)
conditions; t(9)=3D-23.24, p =3D 0.000; for CG there was also a significant
difference in the scores for pretest (M=3D11.2, SD=3D1.33) and test (M=3D13=
.5,
SD=3D0.93) conditions; t(10)=3D-11.66, p =3D 0.000.
In this ca=
se, RH1 is confirmed by a series of
independent samples t-tests, which revealed that, on the one hand, the resu=
lts
at the pretests of both groups are comparable: there was no significant
difference in the scores of the EG (M=3D10.5, SD=3D0.97) and CG (M=3D11.2, =
SD=3D1.33) conditions;
t(19)=3D-1.33, p =3D 0.199; and, on the other hand, the differences between=
the results
of both groups at the tests and post tests were significant: for the test, =
EG
(M=3D16.5, SD=3D1.50) and CG (M=3D13.5, SD=3D0.93) conditions; t(19)=3D5.62=
, p =3D 0.000; for
the post test, EG (M=3D16.2, SD=3D1.16) and CG (M=3D13.2, SD=3D0.60) condit=
ions;
t(19)=3D7.71, p =3D 0.000. The same holds for the differences between the v=
alues of
EG and CG of the more specific parameters (p<0.01).
RH2, on its turn, does not seem to be confirmed here
neither: we only note a minor decrease of the general score when comparing =
test
and posttest (-1.5%, both for EG and CG). For EG there was no significant
difference in the scores for test (M=3D16.5, SD=3D1.51) and posttest (M=3D1=
6.2,
SD=3D1.14) conditions; t(9)=3D-1.96, p =3D 0.081=
; for CG,
there was no significant diff=
erence
neither in the scores for test (M=3D13.5, SD=3D0.93) and posttest (M=3D13.2=
, SD=3D0.60)
conditions; t(10)=3D-1.40, p =3D 0.192.
The same h=
olds
for the more specific parameters: the differences between pretest and test =
are
significant (p<0.01), the ones between test ant posttest are not
(p>0.01).
Again, a
Spearman rank correlation analysis revealed no highly significant influence=
s of
the variables listed in the metadata document, neither for criteria such as
gender, age, nationality, mother tongue, other languages, type of education
(r=3D< .10), nor for criteria such as ±bilingual upbringing, earl=
ier
Spanish courses and earlier Spanish immersion
period(s) (r=3D.10 - .29).
|
|
General score /20 |
Lexical poverty /300 |
/300 |
Static expression /300 |
Dynamic expression /300 |
|
|
|
|
|
|
|
|
EG |
10.5 |
11.2 |
2.3 |
35.6 |
2.6 |
|
CG |
11.2 |
12.3 |
2.5 |
32.3 |
3.2 |
|
Test |
|
|
|
|
|
|
EG |
16.5 |
2.2 |
25.6 |
10.2 |
25.6 |
|
CG |
13.5 |
11.6 |
6.7 |
30.6 |
4.4 |
|
Post test |
|
|
|
|
|
|
EG |
16.2 |
3.3 |
23.2 |
12.6 |
23.4 |
|
CG |
13.2 |
12.6 |
5.6 |
32.3 |
3.4 |
Table II. Results of quantitative analysis: experimental group (EG) =
vs
control group (CG). Average scores of 2011-2012.
The analys=
is
of the portfolios confirms also RH3: the writing of the descriptive texts is
scored as highly motivating by the students of the EG (average of 4.4 on a
scale of 5, vs an average of 2.4 in the portfolios of the control group),
enjoyable (EG: average of 4, vs 2,3 for CG) and with a highly positive effe=
ct
on the perceived improvement (EG: average of 4.1, vs 2,8 for CG).
By way of
illustration we copy here an example of a text (posttest) produced by a stu=
dent
of the EG (3 “poor verbs” vs 27 “rich expressions” =
on a
total of 283 words):
Se
vende burbuja financiera: la Residencia de los Reyes Magos
Entrando por
la puerta inmediatamente notamos qu=
e
los Reyes Magos han incorporado=
su
origen en la decoración de la casa. Pasamos
por estatuas griegas igual que budistas y vudúes. Acudiendo a la cocina ya olemos las especias que provienen de tres continentes
diferentes. En la cocina vemos
grandes hornos que los Reyes utilizaban para cocer artículos de
confitería. Colgados por=
todas
partes hay crucifijos que nos recuerdan el origen de este día=
tan
especial para los españoles. Las otras habitaciones a nivel del suel=
o se utilizaban para producir los re=
galos.
Máquinas de empaquetar apare=
cen
al lado de mesas de diseño.
Subiendo a
la primera planta por una escalera ancha encontramos
las habitaciones de los reyes y sus pajes. En un rincón oscuro se esconde una escalera muy estrec=
ha que
nos permite el acceso al
desván. Allí arriba, =
nos
damos cuenta de que había un museo. Colgadas en la pared ve=
mos
múltiples imágenes re=
latándonos
el origen de los Reyes. Para bajar utilizamos el medio más rá=
pido
siendo un tobogán espiral que da al jardín.
Situado en
un monte, el jardín nos pres=
ta una
vista espectacular sobre Andalucía. Descendiendo del monte buscamos la entrada al túmulo don=
de apilaban todas las cartas recibida=
s de
los niños. A los nuevos propietarios les costará vaciar el si=
tio.
Seguimos descendiendo hasta el pie del monte donde se encuentran los establos de los camellos. Abriendo la verja p=
ara pasar a los establos, en un
santiamén, estamos rodeados por ovejas, cabras y gallinas. Nos
estorban el paso manifestando claramente que están hambrientas.
Empujando los animales a un lado nos
dirigimos hacia la zona segura, fuera de la verja. Un telesquí nos lleva a la cumbre del monte do=
nde
termina nuestra visita.
Since this
didactic intervention is now systematically applied when the text genre of
descriptive texts is introduced, the category of ‘lexical povertyR=
17;
has left the top 5 of most frequent problems in general, and in descriptive=
texts in particular.
On the oth=
er
hand, the Aprescrilov corpus has certain limita=
tions
and could benefit from a number of extensions that are currently missing du=
e to
lack of funding: lemmatization and POS-tagging would open a whole range of =
new
research perspectives; tagging of all contexts without a certain error, i.e.
which may have a learning potential for a specific problem (in Aprescrilov, for legibility reasons this is only done=
for a
few studies, as Markin does not allow multi-lay=
er
annotation and the annotated versions are first sent to the students).
6.
Conclusions
This study
offers insights into the essential role of learner corpora such as Aprescrilov in the evolution of errors by students and
student groups, and shows how powerful some
methodological and didactic changes can prove to be, both for the writing
process (motivation, enjoyment, perception of improvement) and for the outp=
ut
(richness of the writing product).
However, t=
he
study is limited to the writing of one text genre by students of one type of
department in a Flemish university, with a very homogeneous population
regarding age, education, mother tongue, etc. More studies into other langu=
ages
at other departments in other countries should be carried out in order to be able to extrapolate to other population=
s.
NOTES
1 Some results of a pilot of this study were
previously published in Buyse, Fernández
Pereda and Verveckken (2016).
2 The emergence is situated at “the tu=
rn
of the 1990s” (Callies and Paquot
2015: 1) but the field has developed rapidly. It now has a proper internati=
onal
academic association (the Learner Corpus Association) holding an internatio=
nal
conference every two years. A proper handbook has been published by Granger=
and
colleagues. Since April 2015, it also has its own international scientific
journal (International Journal of Learner Corpus Research, John Benjamins) =
(cf.
Callies and Paquot,=
2015:
1-3; Callies et
alii, 2015: 160-161).
3 Cf. the LC listed in the online overview by
Granger et al.:
http://www.uclouvain.be/en-c=
ecl-lcworld.html.
4 Aprescrilov is
available at http://ilt.kuleuven.be/aprescrilov
under the acceptance of the terms and conditions stated in the introduction.
The interface and search buttons are both in Dutch (the language of the
institution, KU Leuven, Belgium) and in Spanish.
5 Markin, elabor=
ated
by Creative Technology, allows noting down compositions digitally. “I=
t is
a Windows program which runs on the teacher's computer. It can import a
student's text for marking by pasting from the clipboard, or directly from =
an
RTF or text file. Once the text has been imported, Mar=
kin
provides all the tools a teacher needs to mark and annotate the text. When
marking is complete, the teacher can export the marked text as an RTF file =
for
loading into a word-processor, or as a web page so that students can view t=
he
marked text in a web browser. Marked work can even be emailed directly back=
to
the student, all from within the Markin
program.” (https://www.cict.co.uk/markin/ind=
ex.php)
6 The corpus takes advantage of the
institutional differences between Belgium and the Netherlands in foreign
language teaching at primary/secondary education level. The Aprescrilov
corpus includes three groups of SFL-learners, characterized by two distinct
mother tongues (Dutch vs. French) and distinct L2s (French – English
– Dutch), and in doing so, allows to determine not only the inference=
of
L1 on the IL of SFL-learners but also the interference of L2. The underlying
hypothesis is that differences in performances may be found between the thr=
ee
learner groups and that some differences may be due to the a distinct degre=
e of
interference from French, according to the following cline: French-speaking
students will probably make more errors reflecting the influence of French =
than
Dutch students, while Flemish students will be situated at the center of the
scale, between French-speaking and Dutch students, due to the different sta=
tus
of French as a FL in Flanders and the Netherlands.
References
Arnold, J. (2011). Attention to Affect in Language
Learning. Angl=
istik. International Journal of English Stu=
dies,
22, pp. 11-22.
Berdal-Masuy, F. and Pairon, J. (2015). Special issue “Affects et acquisition des langues”. Le Langage et l'Homme, 50 (2).
Buyse, K., Delbecque, N. and Speelman,
D. (2009). Trampas y pistas para la
expresión escrita de los neerlandófonos<=
/span>. Edelsa-Averbode: Madrid-Averbode.
Buyse, K. (2011). Effective Writing Tasks and Feedback for the Internet
Generation. Language Learning in Hi=
gher
Education 1 (2), pp. 1–22.
Buyse K., Fernández Pereda L., Verveckken K.
(2016). The Aprescrilov Cor=
pus, or
Broadening the Horizon of Spanish Language Learning in Flanders. In: M. Alo=
nso
Ramos, ed., Studies in Corpus
Linguistics, vol: 78, Spanish Learner Corpus Research Current trends and fu=
ture
perspectives. Amsterdam: Benjamins, pp.
143-168.
Buyse K. and González Meló=
;n
E. (2013). El corpus de aprendices Aprescrilov =
y su
utilidad para la didáctica de ELE en la Bélgica multiling&uum=
l;e.
In: B. Blecua et alii, eds., Multil=
ingüismo
y enseñanza de ELE en contextos multiculturales. Gerona:
Asociación para la Enseñanza del Español como Lengua
Extranjera, pp. 247-261.
Callies, M. and Paquot, M. (2015). Learner Corp=
us
Research. An interdisciplinary filed on the move. International Journal of Learner Corpus Research 1(1), pp. 1-6.=
Callies, M., Paquot, M. and Tono,
Y. (2015). An interview with Yukio Tono. In mem=
oriam
Geoffrey Leech. International Journ=
al of
Learner Corpus Research 1(1), pp. 160-171.
Corder, S. (1981). Error analysis =
and
Interlanguage. Oxford University Press: Oxford.
Cruz Piñol, M.
(2012). Lingüística de corpus =
y enseñanza
del español como 2/L. Madrid: Arco/Libros.
Dewaele, J.-M. (2005). Investigating the psychological and the emotional
dimensions in instructed language learning: Obstacles and possibilities. Modern Language Journal, 89, pp.
367-380.
Dewaele, J.-M. (2015). On emotions in foreign language learning and use. The Language Teacher, 39, pp. 13-1=
5.
Dewaele, J.-M. (2018). (ed.) Special issue “Emotions in SLA”. Studies in Second Language Learning and
Teaching, 8 (1).
Dewaele, J.-M. and MacIntyre, P. (2014). The two
faces of Janus? Anxiety and enjoyment in the foreign language classroom. Studies in Second Language Learning and
Teaching, 4, pp. 237-274.
Díaz-Negrillo, A. & Fernández
Domínguez, J. (2006). Error Tagging Systems for Learner Corpora. =
Revista española<=
/span>
de lingüística aplicada,
19, pp. 83-102.
Ellis, R. (2003). Task-based
Language Learning and Teaching. Oxford, New York: Oxford Applied
Linguistics.
Fernández Pereda, L., =
Buyse,
K. and Verveckken, K. (2014). Error analysis, contrastive linguistics and learner corpora, or how to use current
linguistic tools to improve the level in SFL class: the case of change-of-
state verbs. Studies van de BKL/Travaux du CBL/ papers of the LSB (Belgische Kring
voor Linguïstiek), 8, pp. 59-68.
Fernández, S. (2007). Interlengua y análisis de errores en el aprendizaje del
español como lengua extranjera. Madrid: Edelsa=
.
Granger, S. (1996). From CA to CIA and back: An
integrated approach to computerized bilingual and learner corpora. In: K. <=
span
class=3DSpellE>Aijmer, B. Altenberg, M.
Johansson, eds., Languages in Contr=
ast.
Text-based cross-linguistic studies. Lund: Lund University Press, pp.
37-51.
Granger, S. (2009). The contribution of learner
corpora to second language acquisition and foreign language teaching: A
critical evaluation. In: K.=
Aijmer, ed., Corpora and language teaching. Philadelphia/Amsterdam: John Benjamins,=
pp.
13-32.
Granger, S. (2015). Contrastive interlanguage
analysis. A reapprai=
sal.
International Journal of Learner Co=
rpus
Research 1(1), pp. 7-24.
Hasko, V. (2013). Capturing the Dynamics of Second Language Development v=
ia
Learner Corpus Research: A Very Long Engagement. The Modern Language Journal, 97 (1), pp. 2-10.
MacIntyre<=
/span>, P., Gregersen, T., and Mercer, S. (Eds=
.).
(2016). Positive psychology in SLA<=
/i>.
Bristol: Multilingual Matters.
Nakamura, S. (2018). How I see it: An exploratory
study on attributions and emotions in L2 learning. Studies in Second Language Learning and Teaching, 8 (3), pp.
553-574.
Santos Gargallo, I. (2=
008). El análi=
sis de
errores en la interlengua del hablante no nativo. In: J. Sánchez Lob=
ato
and I. Santos Gargallo, eds., Vadem=
écum
para la formación de profesores: enseñar español como
segunda lengua (L 2) / lengua extranjera (LE). Madrid: SGEL, pp. 391-410.
Selinker, L. (1972). Interlanguage=
. International Review of Applied Lingui=
stics
in Language Teaching, 10:3, pp. 209-232.
Stevick, Earl W. (1980). Teaching
Languages: A Way and Ways. Rowley, MA: Newbury House.
Suby, J. and Asención-Delaney, Y. (20=
09). El uso del
español del profesor en las clases de principiantes. Hispania, vol. 92, n. 3, pp. 593-607.
Vandekerckhove, J., Vandergraesen, F., Cruysweegs,=
B.
(2009). Competent: een algemene did=
actiek
in 101 lemma's. Wommelgem: Van In.