FormalProposal2 41KB Nov 23 2010 04:45:44 PM

advertisement
Ohio University
Proposal for Research in Partial Fulfillment
of the Requirements for the Degree of Master of Arts
Department of Linguistics
Title: Cross-Language Perceptual Assimilation of French and German front vowels by native
speakers of American English.
Submitted by: Gabriella Ruiz
359 Gordy Hall
Department of Linguistics
Athens, OH 45701
Date of Submission: November 23, 2010
Program: Applied Linguistics
First Reader:
Name: Dr. Scott Jarvis
Signature: _________________________
Date: _______________
Second Reader
Name: Dr. Michelle O’ Malley:
Signature: _________________________
Date: _______________
2
Cross-Language Perceptual Assimilation
INTRODUCTION: RESEARCH AREA
The current study will investigate the role of native language (L1) phonological influence
of American English (AE) over the categorization of L2 and L3 high- front rounded and
unrounded vowels of French and German. Previous studies have provided evidence for the
highly influential nature that listeners’ perception has over the realization of speech sounds in an
additional language (cf. Flege 1987, 1992, 1995; Strange 2007; Trofimovich, Gatbonton &
Segalowitz, 2007; strange, Levy & Law II, 2009). Flege (1987) maintains that L2 language
learners’ (LLs) phonological perception differs with respects to phonological categorization of
their L1. Referred to as ‘equivalence classification’, this is a process in which L2 learners realize
phones of the target language (TL) as sounding similar to their L1 due in part to segmental
correlations (e.g. physiological speech constraints). Flege hypothesizes that the phonological and
phonetic effects of listeners’ native languages affect their realization of L2 phones. L2 learners
“identify L2 phones in terms of native language (L1) categories” (Flege, 1987). As a result,
perception of L2 phones is an important part of L2 learning with respects to the role that L1
categorization plays for L2 auditory and articulatory accuracy (Ingram & Park, 1997). In the past
few years, researchers have empirically examined the role that gained learning experience
(gradual increase in study of a non-native language) in an L2 has in regards to L2 phonological
perception (Trofimovich et. al, 2007; York, 2008). The studies have built on earlier work that
offered theoretical models of the effects of gained experience in an additional language, such as
3
Cross-Language Perceptual Assimilation
the “gradual diffusion model” (Gatbonton 1975, 1978) and the “dynamic paradigm” (Bailey,
1973).
The scope of this study focuses on the implications of gained L2 experience of nativespeakers of AE by investigating novice and advanced language listener’s perceptions of the
following high front rounded and unrounded vowels of French: /i, y/ and high-lowered, short/
long lax rounded vowels of German: /y: Y/. Attention will be paid to potential perceptual
differences of these vowel sets between advanced language learners of either French or German,
and by naïve listeners (e.g. monolingual native speakers of AE). As few studies have attempted
to investigate the psycholinguistic role of L1 to L2 and L3 phonology in language acquisition,
the present study intends to find how language learners of two languages with similar sounding
vowel systems, can hear differences between the vowels, or if they continue to draw similarities
of the L2 vowels, to their L1. Provided that the data reveal significant differences of perception
between aforementioned language listeners, further evidence for theories of equivalent
classification can be supported.
AIM/JUSTIFICATION
The theoretical focus of the current study is modeled after Flege’s 1987 study titled, “The
Production of “new” and “similar” phones in a foreign language: evidence for the effect of
equivalence classification.” In this study, Flege hypothesizes that L2 speakers’ tendencies of
matching (“normalizing”) L1 phones with similar L2 phones will inhibit them from sounding
“native-like” for that attempted TL phone (or word). ‘Normalizing’ (e.g. assimilating) any given
4
Cross-Language Perceptual Assimilation
L2 phones with phones in the L1, provides evidence that the L2 speaker cannot distinguish1 a
difference between sounds; in an attempt to compensate for this, the speaker searches the
phonetic categories of their L1 for an equivalent. Perceivably “new” sounds, however, that L2
speakers attempt to produce, will sound closer to the target sound if the sound is not found in
the phonological categorization of the speaker’s L1. Flege and other similar studies (Best 1995;
Ingram & Park 1997) have looked at only the perceptual phonological relationship between a
speaker’s L1 and L2, however. This may account for some of the reasons why additional
considerations for each study suggest a further developed model for cross-language perception.
As a result, this study intends to investigate if advanced English-speaking learners of L2 French
and German perceptually assimilate phones in the same way, or if the L1 continues to influence
their perception of their L2/to L3 correspondence.
LITERATURE REVIEW
Over the past two decades, several studies have investigated the role of perceptual ( i.e.
observed) vowel assimilation between the speech sound correspondences of native speaking (L1)
American English (AE) language learners, of a second language (L2) of French or German (e.g.
Flege 1987, Levy & strange, 2007; Trofimovich, Gatbonton & Segalowitz, 2007; Strange, Levy
& Law II, 2009) Additionally these studies have investigated the implicational role of
phonological context and gained experience in target language (TL) learning over a language
learner’s (LL) perception towards the phonology of a non-native language. It has been a constant
theoretical tenet among these studies that perceived phonemic similarity between native and nonnative speech segments can posit challenges for second language (L2) learning. By conducting
perceptual assimilation production tasks with L2 language learners, these studies have found that
1
Or maybe they can but are unable to produce the sound.
5
Cross-Language Perceptual Assimilation
increased exposure and experience in a TL affects the LL’s perception towards the sounds of an
additional language.
While previous research has provided implications for the pathways in which LLs
process new linguistic information from the L1 to the L2, few studies have attempted to explore
the influential phonological dynamic of perceptual vowel assimilation between an L2 and L3. As
it stands currently, no study has yet attempted to demonstrate if cross-language speech sound
assimilation can occur between two non-native languages. The current study intends to provide
an initial attempt at researching this unexplored territory with specific attention paid to listeners’
cross-language similarity between front rounded and unrounded vowels of French and German.
Although many studies have researched the dynamic between L1 to L2 phonetic assimilation
(Trofimovich et. al 2007) less research has focused on the perceptual effects that gained
language learning L2 experience may have in perceiving the speech sounds between French and
German fronted vowels.
The Role of L1 L2 cross-language perceptual assimilation
Although the current study bases most of its inquiry on previous research done on L1 to
L2 perceptual speech similarity, it departs from the traditional theories in so much that there is
yet to be any evidence of whether perceptual speech sound assimilations can occur between two
non-native languages, and not just between the L1 to L2 or L1 to L3. Evidence for a
phonological categorization framework, was first offered by Trubetzskoy in his 1939 theory on
“filtering” perceptually acoustic distinctions. Trubetzskoy hypothesized that a language listener
will sort out perceptually different sounds that are not represented in their L1 (Flege, 1987)..
Flege’s (1987) study took aspects of this hypothesis further by investigating the formation of
6
Cross-Language Perceptual Assimilation
vowels and voice onset measurements of French speakers (English as L2) and native English
speakers (French as L2--varying at three levels of French proficiency) producing French and
English words. An acoustic analysis was done to measurements of both groups of speakers
speaking in the L2 (English French, FrenchEnglish). Flege hypothesized that L2 speakers
normalization of similar phones inhibits them from sounding “native-like” for those given
“similar realizations” between the L1 and the L2, but that “new” sounds will sound native like if
the sound is not found in the phonological categorization of the speaker’s L1. The results of this
study indicated that adult L2 learners are able to learn how to modify their articulation of similar
phones due to the voice onset time (VOT) results produced by experienced L1 English -L2
French speakers for the French realization of /t/ and /t ʰ/ in English.
The role of phonetic context
Trubetzkoy’s filter theory has provided important grounds for further empirical research
investigating the role of speech perception in phonology; namely, the role of phonological
context impacting the categorization of speech sounds by language learners of a non-native
language.
Flege’s (1987) raised a significant point about perceptual sound assimilation when nonnative sounds are positioned in specific contextual phonological environments either similar or
dissimilar to the L1 of the listener. Unlike Trubetzskoy’s “filtering” process, Flege maintained
that not only does the L1 listener isolate distinctions of L2 speech segments, but that the context
in which the segments occur, is an important part of how s/he perceives, and produces the TL.
7
Cross-Language Perceptual Assimilation
Similarly, recent studies have also investigated the significance of context and the role it plays
for speech sound assimilation.
Levy & Strange (2007) investigated the effects of language experience by evaluating
American English listener’s perceptions of the following Parisian French vowel sets /y, oe, u, i/ .
The study involved experienced and inexperienced AE LLs of French. Studies found that AE
listeners faced higher tendencies to categorize similarities between high and back rounded and
unrounded vowels in alveolar contexts than in bilabial environments. Participants in this study
were asked to discriminate between vowels by listening to disyllabic nonsense words in carrier
phrases and then rating vowel closeness by clicking on a number “1” or “3”. These numbers
were given values” based upon their correctness in a given environment. Discrimination of
vowels differed with respects to the differences in carrier phrases. The results of this study
concluded that the naïve listeners (L1 speakers either without or no current background
knowledge in the TL) perceived the vowel sets differently depending on the context in which the
vowels were set. This provides further implications for further research between contextual
significance and cross-language perceptual similarity.
Two years after this study, Winifred and Strange (2009) tested further significance of the
context of perceptual vowel assimilation in correspondence with gained language experience.
This study examined perceptual assimilation of North German (NG) and Parisian French (PF)
vowels in disyllabic (a word with two syllables) and multisyllabic contexts. Participants rated the
acoustic similarity of these vowels with American English (AE) on a nine point Likert scale.
This scale measured the participants opinions on how “close” given vowels were to vowels
represented in their native language of AE. The results of participant’s categorization found that
participants faced difficulties observing distinctions between fronted vowels of NG and PF.
8
Cross-Language Perceptual Assimilation
Participants demonstrated that accuracy for fronted vowel assimilation varied in respects to
context and a prosodic (e.g. rhythm, stress, intonation) dependent environment; that is
participants were able to respond with less difficulty to speech stimuli that closely represented
phonological categories of their L1. This is, however, contrary to Flege’s assertion in that
similar categories between the L1 and non-native language present more difficulty for the L1
listener, as they are unable to make any distinction in the sound. An explanation for this is that
although Flege’s study focused on context, it did so within a set contextual environment, and not
by randomization (generating a non-sequenced pattern), as Winifred and Strange’s study
demonstrated. This is important as this provides implication for further investigation surrounding
the importance that context can play on the perception of sound in the TL.
Effects of gained experience in the TL
While drawing from previous studies implicational models for investigating L1 to L2 phonetic
correspondences, the current research also investigates the role that the effects of gained
experience in a TL may have towards the perception of speech sounds in that given language.
Trofimovich, Gatbonton, and Segalowitz (2007) investigated the effects of gradual gained
experience for L2 phonological learning. Results showed that as the learning stage progresses,
cross-linguistic similarity determines the effects of perception towards the speech sounds in the
L2. Trofimovich et. al supports Flege’s equivalence classification theory in asserting that a
speaker’s L1 plays an important role in L2 learning. Trofimovich et. al. describes this process as
a gradual diffusion framework, stating that as increased experience in an L2 takes place, nonnative ‘segments’ are then replaced by ‘native’ ones of the TL. Evidence for this theory was
9
Cross-Language Perceptual Assimilation
made by investigating native Québécois French speaker’s realizations of /ð/ in 80 English tokens
given in eight phonetic contexts. Contextual basis was found to be quite significant in so much as
it plays an important role of similarity processing in the L2 speaker. Trofimovich et. al offers
another explanation for perceptual assimilation of speech sounds between and L1 and L2, in
saying that frequency of segments in the L1 also pays a role in the success and facilitation of
learning and producing TL segments.
As previous research has demonstrated, a few factors contribute greatly to an L2
language learner’s overall perception of target language sounds in the L2. The role that
phonological realizations between the L1 and L2 plays, provides evidence for whether LL
distinguish between speech sounds that closely represent speech sounds in their native
language(s). While some studies have demonstrated that this provides challenges for the L2 LL,
other studies have implied that categorizing these sounds and identifying distinctions for them
can be relatively easy for the LL provided the context the speech sounds are placed in. As a
result, current research emphasizes that context of speech sounds does matter for the success of
L2 comprehension. Furthermore, gained experience in a TL can also provide differences in how
the speech sounds of the L2 are perceived. Although there is significant research on the study of
speech sound perception, there still remains little evidence for perceptual speech correspondence
between an L2 and an L3. The closest evidence for this implication comes from DeAngelis’
(2001) hypothesis involving “Foreign talk mode”; thusly, she explains that at times, the L2
speaker may not know a word in the TL but for time’s sake will resort to another
one of the languages s/he may feel is relatively similar to the TL, and will create an
utterance based upon their own idea of what it should sound like .Considered, “talk
foreign”, or “foreign language mode,” this idea speculates that a speaker may not
10
Cross-Language Perceptual Assimilation
perceive their native language to sound “foreign” enough to use in their
interlanguage.
Provided that DeAngelis’ hypothesis refers more to “lexical” usage, she
raises an interesting point stating that “foreign” talk mode involves the “sounds” of
another known language that the LL may feel is more representative of the TL
structure. That is the LL may not feel the speech sounds of their native language do
not correspond well with the TL; therefore, they will draw from another “learned”
language, as they perceive that the other language has a closer representation in
phonological structure than their L1. Given this hypothesis, perceptual speech
sound assimilation may not only be a matter of L1 realizations in an L2 or L1  L3
however, but may also imply that L2 learners when presented with an additional
language, in this case an L3, may draw from their L2 for perceptual accuracy,
especially if they decide that the L2 speech sounds are a more accurate
approximation of the TL speech sound(s).
RESEARCH QUESTIONS
1.1 Do native speakers of American English (AE) perceive /categorize similar phones in
two non-native languages as belonging to the same phoneme?
1.2 Does gained experience in the language affect the participant’s realization between
NG [i, y: Y] and PF [i, y,]?
1.3 Do advanced learners of German and French categorize NG and PF front vowels the
same way or differently?
11
Cross-Language Perceptual Assimilation
METHODOLOGY
SUBJECTS
The current study will consist of three groups of subjects ranging between 30-60 L1 AE
speaking participants. Participants will consist of naïve/ novice language experience in either
French or German, or advanced levels of proficiency in either of the target languages (TL). If
participants from the less experienced L1 English population have taken either French or
German in junior-high to high-school (as this is not un-common), they are still qualified
candidates since it is assumed that they are not currently exposed to either language on a daily
basis. One group of advanced participants will be L1 AE advanced LLs of French. The subjects
will be students (Ss) from an advanced French class from the Department of Modern languages
at Ohio University. The number of Ss may range from15 in one class to 30 from two advanced
sections combined. The other group of advanced language learning subjects will also come from
the Department of Modern Languages at Gordy Hall, and will be L1 AE advanced LLs of
German. Expected number of Ss is around the same as the French LLs.
MATERIALS/INSTRUMENTS
Testing will be done in a quiet computer lab. Each participant will have access to a
computer and a set of headphones. The assimilation tasks will be elicited by Qualtrics survey
system; this system will de divided into three main tasks with a fourth section allowing for
12
Cross-Language Perceptual Assimilation
open-ended feedback on challenges the participants encounter. Additionally, this system will
calculate participant responses in a spreadsheet, which will then be submitted automatically to
the researcher. Target vowel sounds will be recorded by native speakers of both German and
French along with a recording of American vowel sounds by a native a speaker of AE used for
the second to last task of the study. Sounds will be randomized throughout each task. Each task
will consist of 15 task items in which the participants must respond.
PROCEDURE
Testing will take place in a computer lab at Ohio University. Computers will be set-up
prior to participant arrival. It is assumed that not all participants will be able to meet at the same
time; therefore, testing will take place over a one-two day span. The researcher will serve as the
proctor of the testing. Testing Groups will be able to accomplish the required tasks within a 1520 minute time span. The researcher will guide the participants in a 3-4 minute training/priming
session prior to the actual stage of testing. This session will get participants accustomed to the
sounds they will be listening to, and will guide them on how to respond to the three stages of the
listening tasks.
Once the participants are seated, the researcher will pass out paper(s) with directions on
how to submit their test results once they have completed the tasks. Once all participants have
been seated, the researcher will guide them through a “training” screen. This screen will look
like an actual slide used in the study. The researcher will have use of a computer that will be
connected to a projector screen. Participants will be asked to watch the screen, as the researcher
explains the directions, and performs a trial-run for everyone to view. Shortly thereafter, the
researcher will begin priming the participants on the target sounds for which they will be making
13
Cross-Language Perceptual Assimilation
judgments. Participants are instructed during the “training” that although they are allowed as
much time as they need to complete a single item, once they have responded to that item, they
are not allowed to go back. The system will make the question disappear to only be randomized
within the rest of the task items for that phase of testing.
Participants are then asked to put on their headphones and begin. For each task, the
screens will show instructions at the top of the screen followed by a symbol for the participants
to ‘click’ on and listen. The first task will ask participants to identify if the sounds they are
listening to exist in English or not. Below the question they will be asked to choose an answer
from a drop down box of “yes” or “no”. Once participants have answered all 15 items from the
first task, a transition slide will let them know that they have completed the first stage and are
about to begin the second. At this point, there will be a “next” button indicating them to move on
to the next phase.
The testing instructions in task two will be the same as in task one, with the exception
that the participants will listen for the target sounds in this task within monosyllabic ‘non’ words
/h_b ə/ and /t_ bə/. They will be asked to ‘click’ on the sound within the word that is marked by
a symbol; this symbol will be arbitrary to eliminate grapheme phoneme correspondence. The
distributional criteria of the non words will follow the phonology of the French and German
languages. Participants are asked again as in the first task; to make judgments based on whether
they think the vowels they are listening to exist in English. These vowels will also be
randomized within the monosyllabic contexts. Again participants will complete a total of 15 task
items, and will not be aloud to go back to change their responses.
14
Cross-Language Perceptual Assimilation
In task three, participants will be prompted by the screen to listen to the following vowel
pairs. This time the pairs will be randomized as single vowel pairs, and non-word pairs. English
high and mid front unrounded and back rounded/unrounded vowels will also be incorporated in
the item sets. After listening to each pair of items, participants will be asked to choose if the
target sounds they are listening for sound the ‘same’ or ‘different’. If they respond “different”,
they will then be asked to provide their alternative response, by typing in a text box according to,
which vowel(s) they believe they are actually hearing. Specific instructions for this task will ask
that participants with alternative perceptions indicate which vowel/non-word they believe sounds
different, along with providing which vowel they believe they are hearing.
Once task three has been completed, the participants will be led to a new screen with an
open text box. This box will ask participants to take 15-20 seconds to provide feedback on some
of the more challenging portions of the testing.
The researcher will thank everyone for their participation and offer to send those
interested, the results of the study once it has been completed. Once all participants have left the
room, the researcher saves all “saved” files on each computer to an external hard drive where
data will be stored for analysis.
TYPE OF DATA
Responses from the first two tasks will be scored as the number of times respondents
from each group choose “yes” or “no” if the vowel exists in AE ((60 participants)(15 trials)=900
responses/vowel * 3 tasks= 2700 responses/item). The consistency of item choice for each
category will be represented as a percentage of the total trial per group. A median response item
15
Cross-Language Perceptual Assimilation
will be determined for those groups that chose an alternative item(s). To account for individual
responses, each participant’s American vowel selection for the German or French vowels will be
tallied.
ANALYSIS
As this study consists of three independent groups with different variables, there will be
one quantitative analysis for this study performed with SPSS. A one-way ANOVA with post-hoc
tests will be used to test for differences in the means of the three groups in relation to their ability
to distinguish between similar vowels of the two target languages and of English.
ANTICIPATED PROBLEMS/LIMITATIONS OF THE STUDY
Although the number of advanced language learning participants is not difficult to find,
the permission to use them in the study is given by the department head of the Modern languages
program. The study may be allowed access to only one section of the advanced level French and
German classes; if this happens, alterations will have to be made in how many naïve/novice
participants will participate.
Leaving open-ended time for participants to respond may skew some of the results, as subjects
may second-guess their initial response(s). An initial pilot will be conducted to test if the
allowance of no time marker will effect participant responses.
RESULTS
16
Cross-Language Perceptual Assimilation
As this study is modeled after Strange & Levy (2009), I expect to receive similar results.
I am, however testing for second language (L2) to additional language (L3) perceptual
assimilation, a result of which Strange et. al , discovered incidentally, and also producing
minimal results, as the focus of their study did not explicitly test for such occurrences.
For this study I expect that participants in the advanced German and French groups will
assimilate the French and German vowels in the same way. Additionally, I expect to find that
many respondents from both groups will make opposite responses for /y, Y/ vowels; that is
where the German/Y/ vowel occurs, the AE advanced German language learners, will respond
the same AE vowel as they do for the French /y/. For the AE advanced French language
learners, I expect these participants to respond the same for both /y: / and /Y/ vowels. As for the
naïve/novice L1 AE participants, I expect their assimilation responses to be different from the
advanced language learners of the target languages. Each naïve/novice participant will respond
similarly for their independent group, but differently from the advance language learner groups.
CONCLUSION
Several studies (cf. Flege 1987, 1992, 1995; Strange 2007; Trofimovich, Gatbonton &
Segalowitz, 2007; strange, Levy & Law II, 2009) have demonstrated the nature of speech sound
categorization between phones in a speaker’s L1 and L2. Speech sound correspondence is
significant for L2 performance as previous research has shown that perceivable similarities
between as speaker’s L1 and L2 may present some challenges for L2 learners’ speech
production (Strange et. al, 2009). As considerable research has been conducted demonstrating
the relationship between L1 to L2 speech correspondence, very little research has been done
investigating the relationship of speech correspondence between two non-native languages. For
17
Cross-Language Perceptual Assimilation
this reason, the current study will explore aspects of speech categorization between two nonnative languages while testing for speech categorical tendencies made by native speakers of
American English. Significant results demonstrating perceptual correspondences made between
two non-native languages will provide further discussion for the role of an L2 over the
perception of speech sounds in an L3, that is, if the LL perceives speech similarities between
both languages. Alternatively, results that do not show significance between the categorization
of vowels in an L2 and L3 will provide additional support for previous hypotheses surrounding
the idea of L1 influence over non-native language learning.
18
Cross-Language Perceptual Assimilation
REFERENCES
Cenoz, Jasone, Hufeisen, Britta, & Jessner, Ulrike. (2001). Cross-linguistic
Influence in third language acquisition . Multilingual Matters Ltd
Flege, J. E. (1987). The production of "new" and "similar" phones in a foreign
language:evidence for the effect of equivalence classification. Journal of phonetics , 4765.
Flege. J.E (1992). Speech Learning in a Second Language. In C.A. Ferguson, L.
Memm, & C. Stoel-Gammon (Eds.), Phonological development: Models,
research, implications (pp.565-604). Timonium, MD: York Press.
Flege. J. E. (1995). Second language Speech Learning: Theory, findings, and
problems. In W. Strange (Ed.), Speech Perception and linguistic experience:
issues in cross-language research (pp 233-277). Batimore: York Press
Gatbonton, E. (1975). Systematic variations in second language speech: A
sociolinguistic study. Unpublished doctoral dissertation, McGill University,
Montreal, Canada.
Gatbonton, E. (1978). Patterned phonetic variability in second-language speech: A
gradual diffusion model.Canadian Modern Language Review, 34, 335-347.
Ingram,J. C.L.,& Park, S. ( 1997). Cross-Language vowel perception and production
19
Cross-Language Perceptual Assimilation
by Japanese and Korean learners of English. Journal of Phonetic, 25, 343-370.
Levy, E.S., & Strange, W. (2008). Perception of French Vowels by American English
adults with and without French language experience. Journal of Phonetics, 36,
141-157.
Trofimovich, P., Gatbonton, E., & Segalowitz, N. (2007) A Dynamic Look AtL2
Phonological Learning: Seeking Processing Explanations for Implicational
Phenomena. Studies in Second language Acquisition, 407-448.
Strange, W., Levy, E.S., Law II, F.F. (2009). Cross-language categorization of French
and German vowels by naïve American listeners. Acoustical Society of America,
126, (1461-1467).
Download