Professional Documents
Culture Documents
Errors of Commission Omission and The in
Errors of Commission Omission and The in
Errors of Commission Omission and The in
---
---
In Partial Fulfillment
of the Requirements for the Course
Second Language Acquisition and Bilingualism
---
by
Veronika Richtarcikova
February 22, 2013
Abstract
Learning a foreign language is notoriously difficult for adults and few ever
achieve native like competence. It has been claimed that fossilization – cessation
of learning in some aspect of interlanguage - is inevitable, permanent and
partially caused by L1 transfer. By building on previous research by Franceschina
(2005) and Zyzik (2006), this study is designed to investigate the fossilization of a
morphosyntactic construct of the Spanish clitic se present in the causative-
inchoative alternation of certain verbs. Through a series of grammaticality
judgment and oral production tasks in Spanish, performance of L1 English
speakers, immersed in L2 Spanish environment is compared to a control group of
speakers of an L1 where a comparable clitic is present. It is expected that the
findings would prove the hypothesis that a native language with no clitic feature
is the main predictor of fossilization. Secondary hypotheses are that errors of
omission are the most enduring, and that lack of formal instruction is a significant
contributing factor to fossilization. The study is relevant to the issue of L1 transfer
and ultimate attainment in L2 morphosyntax competence that adult learners in
immersion context can reach.
Table of contents
0. Introduction..........................................................................................................1
0.1. Structure of the paper.................................................................................1
1. Literature Review: Fossilization..........................................................................2
1.1. Conceptualizations, characteristics and causal variables of fossilization.....2
1.2. Notable empirical findings..........................................................…...........2
2. Present study........................................................................................................4
2.1. Spanish clitic se.........................................................................................4
3. Methodology........................................................................................................6
3.1. Experimental design..................................................................................6
3.2. Research questions and hypotheses...........................................................7
3.3. Subjects......................................................................................................8
3.4. Data collection, instruments and analysis..................................................9
3.5. Expected results and discussion.................................................................9
4. Conclusion.........................................................................................................10
5. References..........................................................................................................11
6. Appendices.........................................................................................................12
0. Introduction
Vast majority of adult second language (L2) learners never achieve native-
like proficiency in one or more linguistic aspects of the target language (TL)
despite strong motivation, effort and otherwise favorable conditions. Why does
everyone, under normal circumstances, reach full competence in their native
language but the same seems impossible even for the most talented L2 learners?
Understanding of this non-trivial limitation to second language acquisition (SLA)
is relevant both to educators and researchers alike. Identifying what is not
learnable/not teachable has implications on pedagogical strategies and curriculum
development. From the theoretical point of view, on the other hand, study of the
failure to reach the target L2 performance directly contributes to a resolution of
one of the controversial questions in SLA - whether fossilization is a valid
explanation of the phenomenon. Selinker first introduced the term fossilization in
his seminal paper in 1972 and later explicitly defined it as “the process whereby
the learner creates a cessation of interlanguage (IL) learning, thus stopping the IL
from developing, it is hypothesized, in a permanent way...The argument is that no
adult can hope to ever speak a second language in such a way that s/he is
indistinguishable from native speakers of that language.” (Selinker, 1996).
Important strides that have been made in the research of fossilization in the past
four decades will be briefly discussed in the ensuing subsections.
1
1. Literature Review: Fossilization
1.1. Conceptualizations, characteristics and causal variables of fossilization
The definitions and alleged scope of fossilization vary greatly across SLA
literature (see Han, 2004, for an in-depth analysis). Multiple angles on the
construct include “backsliding” (R. Ellis, 1985), temporary “plateauing” (Flynn
and O'Neil, 1988), permanent “cessation of learning” (Odlin, 1993), equating
fossilization with “ultimate attainment” (Towell and Hawkins, 1994), etc.
Fossilization is believed to be “local”, affecting only certain subsets of IL, or
“global”, affecting all aspects of the learner's IL.
2
environment and are of great interest for the light they throw on the non-target
nature of the end-state grammatical L2 knowledge in immersion conditions.
Many researchers, however, employ a working definition of IL
“stabilization” and instead of following a population of L2 learners longitudinally,
they use a cross-sectional design. Such studies also provide useful insights, but
arguably generalize to a certain language community (e.g. through an analysis of
typical persistent errors of advanced speakers of a uniform L1) and cannot capture
what is happening to individual learners. One example of this type of study is
Mukkatash (1986) whose methodological approach included a “defossilization”
treatment. In his study, corrective-feedback provided to 80 Jordanian learners with
an average of 11 years of prior English instruction failed to improve their
grammatical performance. Fossilization was proposed as an explanation.
A puzzling question in the fossilization research is why some grammatical
structures are affected and seem to be unlearnable for adult L2 learners but not
others. Long (2005) suggests that this can be explained by the interaction of a
feature's perceptual salience characteristics, such as frequency, communicative
value and semantic weight with the input sensitivity of some individual learners.
If such a proposal is correct, it would predict which classes of linguistic elements
are more likely to fossilize (or stabilize) than others. Long cites Todeva (1992) in
the discussion of three high-risk categories (HRCs) of features that are especially
prone to fossilization: (i) categories lacking a straightforward form-function
relationship (e.g articles); (ii) semi-productive rules with a high number of
exceptions (e.g. English stress shift in verb-to-adjective formations); and (iii) units
of a highly arbitrary nature (e.g. gender assignment). Franceschina (2005) looked
specifically at one such HRC feature, grammatical gender, and hypothesized that
the absence of this grammar feature in L1 plays a large role in the subjects' non-
target performance. The methodology involved oral interviews combined with a
series of elicitation tasks with the aim of tapping the relevant grammar feature
through different measures. 68 subjects were divided into two groups: -gen group
(speakers of English where the feature is not observed), and +gen group (speakers
of a number of languages, French, Arabic, etc. where this feature is observed). By
3
comparing the results and observing that -gen group (81.33% accuracy) was
systematically less target-like than +gen (92.4%), Franceschina provided evidence
that the absence of the gender feature is the determining factor that accounts for
the differences in performance between these two groups.
In light of the research findings to date and lack of definite evidence of
whether IL grammars fossilize, further research is warranted. One such design
proposal will be delineated in the ensuing subsections.
2. Present Study
In line with Franceschina's (2005) interest in HRC grammatical features that
differ in L1 and L2 and present a learnability problem because of their
exceedingly complex distribution and form-function mapping, the present
research study will focus on a Spanish morphosyntactic clitic se. To my
knowledge, there has not been a previous study on English native speakers,
linguistically and culturally immersed in the country of the TL for a substantial
period of time, and their acquisition of the clitic se in the inchoative-causative
verb alternation in Spanish.
4
For the purposes of the present study only the decausative function will be
considered. As illustrated by the following examples, the same state of events can
be expressed from two perspectives, depending on whether they occur with an
influence of an external agent or cause. In syntax this is referred to as a causative-
inchoative alternation and is morphologically coded in Spanish but not in English.
Zyzik (2006) tested English native speaking subjects' performance on this type of
alternating verbs. Her hypothesis that English-speaking learners would produce
zero morphology on intransitive verbs that depict inchoative events, i.e. commit
the error of omission, was confirmed by the results (see Table 2).
The advanced learners' grammar failed to converge with that of NSs, in fact,
errors of omission were more enduring and only 50% of correct answers in
transitive→intransitive alternation is seen by Zyzik as evidence for the lasting
influence of one's L1 (zero-derivation in this case) on the SLA process. The
advanced subjects in Zyzik's study were college students enrolled in an advanced
Spanish course at an American university, with less then one academic year spent
abroad and English reported as being the only language spoken in their household.
The current proposal, as delineated in the following section, builds on this
research by looking at subjects in an immersion context where, as opposed to
Zyzik's advanced subject group, fossilization might be reasonably expected.
5
3. Methodology
3.1 Experimental Design
The study will compare grammatical judgment and oral production data collected
from high-proficiency speakers of L2 Spanish, both male and female, living in
Spain (TL environment) for a substantial period of time (approximately 10 years).
The participants in the study are screened for, among other factors, their level of
acculturation and linguistic immersion through a self-assessment questionnaire
and tested on their proficiency through a standardized test. Thus, the individual
differences are controlled to the highest ability of the researcher. The subjects are
divided into two groups of distinct linguistic backgrounds: one group of English
native speakers (-clitic experimental group) and one group of speakers of a
language where a comparable clitic is present in the L1 grammar (+clitic control
group). A group of native Spanish speakers will provide the baseline data. L1
being the independent variable, with secondary independent variable being the
amount of formal instruction, the focus of the study is to measure the dependent
variable of amount of errors of commission (overgeneralization) and omission
(avoidance). The data is collected one time, through administering of
grammaticality judgment and oral production tasks. If the -clitic group performs
significantly worse than the +clitic group, this will serve as supporting evidence
that L1 transfer is a crucial factor in the fossilization of certain grammatical
features in adult SLA.
Figure 1 below graphically represents the experimental design.
baseline
6
3.2. Research Questions and Hypotheses
Adopting a working definition of fossilization as a stabilized non-target
like IL state despite a number of years in the TL environment and plentiful
linguistic input, and assuming enduring L1 transfer even for advanced learners,
the following research questions guided the design of the current study proposal:
1. Does L1-L2 distinction in morphological coding of causative-inchoative
alternation in the two languages lead to a general failure by English NSs to
fully acquire the Spanish clitic se in an immersion context?
2. Is fossilization of avoidance of a structure different from fossilization of
overgeneralization, as measured by a quantitative difference in the amount
of errors of omission vs. errors of commission?
3. Considering the various suggestions for causes of fossilization in the
literature (as mentioned in Section 1) and given the interesting expatriate
situation of my subjects, what other factors reflected in their lives will
affect the tendency to fossilize in this aspect of Spanish morphosyntax?
3.3. Subjects
Following the design of Franceschina's (2005) study, in which she compared
two groups, +gen and -gen, this study also compares two groups of participants,
one where the grammatical feature of a clitic is observed and one where it is not.
The experimental, -clitic group consists of 30 native speakers of English of mixed
gender. These participants would be selected from among adult expatriates who
7
have lived in Spain (Spanish speaking regions, e.g. Madrid) for a substantial
amount of time (>10 years) and have the intention to stay indefinitely. Even
learners who reside in a TL environment for an equal period of time can vary in
their ability, motivation and opportunity to learn during that time (it is not
uncommon for such individuals to spend considerable proportions of their lives in
L1 “linguistic ghettos”). Therefore, such factors as the quality and quantity of
language input, assessed by a questionnaire (see Appendix 1), will be controlled.
The ideal subjects are highly-immersed (e.g. use Spanish to communicate with
their spouse or in their place of employment), of similar age, socioeconomic and
linguistic (speakers of only L1 and Spanish) background. The experimental group
will be divided further into two subgroups: those subjects that have received some
formal instruction, and those who have received none or a negligible amount.
The control +clitic group consists of 30 expatriates with an L1 that has a clitic
construction similar to Spanish, for example Slovak, similarly screened and
comparable to the experimental group in their high immersion scores. Both
+clitic and -clitic will be homogeneously highly proficient in Spanish, measured
by using a standardized test without regard to the particular grammatical construct
of clitic se. Because not even NSs perform at a perfect ceiling-level, 10 Spanish
NSs will set the target-performance norm to which the other results will be
compared, The subjects will be compensated for their participation.
8
The error data collected will be quantifiable in respect to the error type (omitting
se in the transitive-intransitive alternation - type 1, overgeneralizing in the
opposite direction - type 2).
The data analysis step of the experiment naturally begins with scoring the
participants' responses. In order to compare the experimental group to the control
groups, the statistical tool SPSS will be employed. The independent variables in
the calculations are the participant's L1 and the amount of formal instruction they
have received. The dependent variables are the amount and type of errors.
Because of the complexity of the data, a mixed model analysis of variance
(MANOVA), using the Tukey-Kramer post hoc test, is appropriate to reveal
potential patterns and to show whether the differences between the other two
groups are significant. The interpretation of results and comparison to the
previous empirical data in the literature is the final step in this process.
9
level. In addition, if the t-test statistical tool reveals significant correlation
between the amount of previous formal instruction and the percentage of errors,
the stated hypotheses 3 would be confirmed. Such hypothetical results would
build upon the existing literature and further contribute to the current
understanding of which “local” morphosyntactic areas of IL do or do not fossilize
and the causes of why that is. The opposite results would possibly serve as
counter-evidence to fossilization and L1 transfer being its qualifying factor.
4. Conclusion
SLA research over the past forty years has ushered in a myriad of
explanations for fossilization. The current study proposal does not claim to be able
to provide definite and decisive conclusion on the permanence of IL fossilization
for every individual English speaker, but aims to draw generalizations over L1
English population. Furthermore, ”permanence”, if interpreted literally, is
impossibly difficult to document and prove empirically. To do so, it would be
necessary to demonstrate that the fossilized item in question – correct use of the
Spanish clitic se - has completely ceased developing and a learner would not
make progress regardless of any additional time spent in Spain or personalized
corrective feedback. However, this would require that a researcher reanalyzes the
learners performance repeatedly, ideally until the learner’s death. This
methodology is admittedly unrealistic. In SLA research, we must remain
cognizant of inter-speaker and intra-speaker variability as well as the ever elusive
character of the issue under question – fossilization.
10
5. References
Zyzik, Eve. 2006. Transitivity Alternations and Sequence Learning: Insights from
L2 Spanish Production Data. In SSLA, Vol. 28 (pp. 449-485). Cambridge
Unviersity Press.
11
6. Appendices
Appendix 1 – Screening questionnaire
(Thank you for your participation in this research study.
The information you provide will be kept strictly confidential.)
1. Personal Data
Name: ___________________ Age: _______ Sex: M / F
Country of birth: ___________ Marital status: _____________
Telephone number or e-mail:_______________________________________
1. What is your highest level of education completed? (tick one):
□ some high school □ high school □ some college □ college graduate
2. Language Background
2. Country of origin: ____________ 3. Age of arrival to Spain: ______
4. What language did your caregivers speak before you were 5? _______
5. At what age did you first begin to learn Spanish? ______ tick one:
Mainly through: □ classroom instruction □ interacting with people □ mixture of both
6. How would you characterize your level of Spanish at the time you moved to Spain? (circle one)
□ none □ very basic □ conversational □ very good □ advanced
7. How would you characterize your level of Spanish at the moment? (circle one)
□ none □ very basic □ conversational □ very good □ advanced
8. Have you attended a Spanish language course? If yes, answer following:
Start year: ______ End year: _______ Average number of hours per week: __________
7. What languages other than Spanish and your native language do you speak? _____________
8. If you are married, what language do you speak with your spouse? _____________
13. In general, what language do you currently prefer to use? □ my native language □ Spanish
14. In general, what language do you currently use at home? □ my native language □ Spanish
15. In general, what language do you currently use at work? □ my native language □ Spanish
16. In general, what language do you currently use with the friends you see most often?
□ my native language □ Spanish
17. In general, what language do you currently use at work? □ my native language □ Spanish
18. Would you like to improve your Spanish language skills? □ Yes □ No
Why? ______________________________________________________________________
What specifically would you like to improve?_______________________________________
What, if anything do you do to improve your Spanish language skills? ___________________
______________________________________________________________________
12
Appendix 2 – Sample of the tasks
1. acceptability judgment test – sample questions
Test item with alternating verb: non-alternating verb:
Researcher: Este hombre se pesa. ¿Y qué sucede en este dibujo? (The man is weighing himself and
in this picture?)
Participant: e.g. El hombre pesa unas manzanas. (He’s weighing some apples.)
By providing the source input, the researcher primes the target lexical item and the participant
does not have to guess which verb to manipulate. This is done in order to avoid a wide range of
appropriate but off-target responses. The following pictures represent these verbs: 1. pesar-se, 2.
secar-se, 3. mirar-se, 4. acostar-se.
1. 2.
3. 4.
13