pROPOSAL kbk rUSWAN

advertisement
THE PROCESSING OF AMBIGUOUS SENTENCES
BY NATIVE SPEAKERS AND FOREIGN LANGUAGE LEARNERS
OF ENGLISH
A. INTRODUCTION
1. Background
Ambiguity is a common phenomenon in human language. It exists in both spoken and
written language. In other words, it is pervasive in human linguistic expressions (Vuong
2010). Interestingly, however, people almost never notice ambiguity in spoken language.
When someone says “I went to the bank,” people tend to interpret it as “I visited a financial
institution where people can invest or borrow money.” The word “bank” in this case is not
interpreted as “the side of a river.” People appear to ignore the fact that a bank does not
only mean a financial institution. Perhaps, this is because in everyday conversation
contexts help people guess the meaning of a bank.
If the sentence is examined “in isolation,” the sentence is lexically ambiguous.
Ambiguity occurs because the word “bank” has more than one meaning. In addition to
lexical ambiguity, sentences can have more than one interpretation due to structural
ambiguity, which means ambiguity that occurs when a sentence has more than one
syntactic structure.
A famous example of structure ambiguity is “I saw the boy with the binoculars.” The
ambiguous part of the sentence is the prepositional phrase “with the binoculars.” The
prepositional phrase may be interpreted as “the tools that I used for seeing the boy” or “the
tools that the boy was holding.” If the sentence is taken out of context, it is confusing
because it may represent two different events.
Observing the above examples, it is interesting to investigate how the human mind
handles the interpretation of ambiguous sentences when contexts are not present. In other
words, what strategies are used by people to determine the correct meaning of ambiguous
sentences? Is there such a thing that applies universally when it comes to interpreting
ambiguous sentences?
Thus, this proposal will seek to examine the universality of the human sentence
processing mechanism. The proposal will also explore whether there are cross-linguistic
variations in parsing ambiguous sentences.
1|Page
2. Statement of Problems
The present study tries to answer the following questions: (1) What types of strategies
are employed in interpreting ambiguous sentences by native speakers of English?, (2)
What types of strategies are employed in interpreting ambiguous sentences by foreign
language learners of English?, and (3) How do sentence processing theories explain the use
of those strategies?
Some concepts used in the present study can be understood as followas:
(1)
Ambiguity is undesrtood as a linguistic form that has more than one meaning such as
words or syntactic structures; (2) Late Closure
can be defined as the situation that:
“requires that new incoming elements be attached to the phrase currently being processed.”
This strategy is also called “low attachment” (3) Minimal Attachment is a condition in a
sentence that “requires that new incoming material be attached in a way that the fewest
necessary phrase structure nodes are used in accordance with the well-formedness of
language rules. This strategy is also called “high attachment.”
3. Research Objectives
The present study tries to describe: (1) the types of strategies employed in interpreting
ambiguous sentences by native speakers of English?, (2) the types of strategies employed
in interpreting ambiguous sentences by foreign language learners of English?, and (3)
theoretical explanation that can be derived from the data found. This study will also look at
the differences of the processing ambiguous sentences between native speakers of English
and the respondents speaking English as a foreign language. It will reveal the whether the
cognitive language processing between the different respondents are different.
2|Page
B. RESEARCH ROAD MAP
Penelitian yang sudah
dilakukan
Didi Sukyadi
1) Coreferences in
English and Indonesian
detached participle
2) The prototype of
“anger” and “hate”
3) The metaphorical use
of English address
terms
4) A Semiotic analysis of
cyber emoticons
Penelitian yang akan dilakukan
TARGET KEPAKARAN
Semiotics
Morphology
Pyscholinguistics
DIDI SUKYADI
DADANG SUDANA
SUKYADI
RUSWAN DALLYONO
1) The non-iconic clause ordering and
discourse function of sentenceinitial adverbial clauses in students
theses and dissertations
2) The positioning of temporal
adverbial clauses in Indonesian
3) The visual iconicity of UN test
item illustration
4) Phonasthemes in Sundanese
5) The metaphors used in Indonesian
Cartoons
Penelitian saat ini
The iconic construction of
gestures used by science
teachers using English as the
language of instruction
Dadang Sudana
1) Relating the Prefix
{meng-} and {ber-} to
verb roots (2007)
2) Semantic Mapping of
Affixation: an attempt
to explain affixation in
Bahasa Indonesia
(2007)
3) Derivative – Based
Materials Development
to Improve Students,
Vocabulary
Acquisition (2006)
Ruswan Dallyono
1) The Role of Literacy
Education to National
Development.
Unpublished.
2) The Use of Multimedia
in The Teaching of
Speaking.
Unpublished. (2008)
3) The Use of heading in
Academic Discourse.
Educationalist Journal.
1) The study of the relation
between botanical vocabulary
and thought and attitude
towards the preservation of
environment.
2) The semantic study
concerning the the affixation
process of single and double
afix.
1) The acquisition of TLUS 1) Syntax
subject in Indonesian by 2) Morphology
foreign language learners
of Indonesian.
2) The use of modality in
British English in written
academic text
C. REVIEW OF RELATED LITERATURE
3|Page
1) The study of how fonem is stored
and accessed in our brain.
2) The study of how lexical and
grammatical
morphemes
are
accessed and retrieved from the
brain.
3) The study of how formulaic
expressions are stored and retrieved
from the brain.
4) The study of how sentences are
formulated to store messages
5) Studying the phenomenon of
language
lost
and
language
deficiency
6) Studying the relation betwwn
brain and language activities
The mental sentence processing theories have been revolutionized by Fodor (1983)
who wrote a book entitled “The modularity of the Mind” which says that we should not
analogize the mind as a massive maximally interconnected network, such that, one's
language processing can probably be influenced by what was eaten in the morning or the
colour of one's eyes (Jackendoff 2000). According to this modular theory, the mind has
distinct parts that have their own functions. They still constitute one system but they are
not directly connected to one another.
Sentence processing is not an exception. There must be one or several functions in the
brain that take cake of the mechanisms involved. There are a number of theories that have
developed within the modular framework, expounding how the mind works when an
individual processes a sentence. One of these functions is the part that is called the parser.
This function has an essential role in sentence processing. The parser is defined as
something that determines the way a sentence is syntactically analyzed (Field 2004).
Interestingly, the parser seems to work in a rule-based manner so that it does not randomly
choose or prefer one specifically rather than another. Several theories will be described as
to how the human parsing mechanisms work in the mind.
According to the universal models of sentence processing, the human parsing
mechanism has the same patterns across different languages and the parsing actions are
dictated by grammatical constraints (Abney 1989; Crocker 1996; Gorrell 1995; Philips
1996; Pritchett 1988, 1992 in Papadopoulou 2006). In this perspective, people are
predicted to interpret ambiguous sentences in relatively the same ways.
Most universal parsing theories argue that parsing preferences are controlled by a
locality principle which requires new information to be attached to the phrase being
processed and which is thought to come from grammatical constraints (Philips 1996;
Weinberg 2001 in Papadopoulou 2006). The locality principle maintains that the parser
will attach a clause to the closest DP. This principle is also called “low attachment.”
Philips (1996) says that by drawing upon the principle-based model, the grammar and the
parser actually embody the same system. Thus, they are like two sides of a coin. They only
differ in the sense that the parser is the type of grammar which operates within limited
resources such as memory constraints and expectations. These are psycholinguistic
variables that might interfere in interpreting ambiguous sentences (Papadopoulou 2006).
Philips argues that all syntactic considerations involved in interpreting ambiguous
sentences depend on a single universal principle called Branch Right. This syntactic
4|Page
tendency is determined by the mental grammar of the individual and is justified by an
economy principle of grammar. The Branch Right principle requires that grammatical
derivations move from left to right. In Frazier's term, this phenomenon is called Late
Closure (Frazier 1978 in Papadopoulou 2006) or in Gibson's term it is called Recency
(Gibson et al. 1996A, 1999 in Papadopoulou 2006 ). Nonetheless, the locality principle has
failed to explain cross-linguistic experimental findings. Spanish subjects, for example,
were found to prefer the minimal attachment strategy in interpreting ambiguous sentences
(Fernandez and Cairns 2011).
Another perspective to the processing of ambiguity is the Garden Path model. This
model presupposes that there is a serial, modular and phrase-structure driven parsing
mechanism in interpreting ambiguous sentences. According to this model, in responding to
an ambiguous sentence, the parser chooses one analysis rather than using multiple
syntactic analyses or postponing analyses of input (Clifton et al. 1991; Frazier 1987;
Frazier and Rayner 1988 and Fodor1998 in Papadopoulou 2006).
In cases, the first analysis is found to be inappropriate, the parser goes to the garden
path and reanalyze the sentence. There are two phases to this model, namely the parsing
and the interpretation phases. In the parsing phase the parser determines a structural
analysis to the input received based on phrase structure rules. In this phase, the parser
draws on syntactic information to get to the initially preferred analysis. Thus, the parsing
phase represents a purely syntactic analysis, ignoring non-syntactic information such as
semantics and pragmatics (Frazier 1987; Ferreira and Clifton 1986 in Papadopoulou 2006).
The interpretation phase, however, makes use of all the non-syntactic resources such
as semantics, pragmatics and other general knowledge. In this phase, the parser comes to a
conclusion whether the analysis is correct or incorrect (Papadopoulou 2006). The parser
has two principles of parsing, namely Late Closure and Minimal Attachment. The Late
Closure principle states that new constituents should be attached to the phrase being
processed. This principle is also known as the lower attachment strategy. The Minimal
Attachment, however, states that new constituents should be attached in such a way that
the fewest obligatory phrase nodes are employed for the sake of linguistic well-formedness
(Papadopoulou 2006).
These structural principles function to maximize the speed and efficiency for
integrating new information into the occurring analysis. These two principles merge into
what is called the First Analysis Constraint. The parser, therefore, is thought to be
5|Page
universal and all parsing differences occurring cross-linguistically are caused by specific
grammatical variations among differing languages (Papadopoulou 2006).
Fodor (1998: 291–293 in Papadopoulou 2006) assumes that the human sentence
processor is designed to make a ‘least effort’ and its goal is to parse as fast as possible,
therefore in his view the First Analysis Constraint tends to construct the simplest analysis,
which is actually due to the ‘laziness’ property of the parser. The Closure principle is
normally used in (1) and it dictates that the constituent to Sue should be attached to the last
DP constituent (the letter) instead of the higher DP(the memo, the note) or to the VP
(read):
(1) Alex read the note, the memo, and the letter to Sue.
This prediction has reaffirmed by several studies (Ferreira and Henderson 1991a;
Frazier and Rayner 1982; Kennedy and Murray1984; Mitchell 1987a in Papadopoulou
2006). Nevertheless, the universal applicability of Late Closure has been questioned by
Cuetos and Mitchell (1988) in (Papadopoulou 2006), who investigated RC attachment
choices in English and Spanish. The Garden Path model presupposes that parsing
principles are universal and valid across different languages. In sentences such as (2), Late
Closure predicts that people will attach the RC low to the second DP, the actress:
(2) Someone shot the servant of the actress who was on the balcony.
This prediction has been validated by empirical findings in such languages as English
(Cuetos and Mitchell, 1988; Frazier and Clifton, 1996; Gilboy et al., 1995; off-line studies
undertaken by Traxler et al. 1998), Swedish (Ehrlich et al. 1999), Norwegian (Ehrlich et
al., 1999), Romanian (Ehrlich et al., 1999), Brazilian Portuguese (Miyamoto, 1998), and
Arabic (Abdelghany and Fodor, 1999), where the RC is suitably attached low, that is, to
the second DP. Nonetheless, the traditional Garden Path model has failed to expound the
tendency to choose the first DP attachment in Spanish by Cuetos and Mitchell (1988) and
also by subsequent studies in Spanish (Carreiras and Clifton, 1993, 1999; Gilboy et al.,
1995 in Papadopoulou 2006).
Construal theory
6|Page
A more refined parsing theory is the Construal theory (Frazier and Clifton, 1996, 1997 in
Papadopoulou 2006). The Construal theory is presupposes different parsing mechanisms
for different structures. Syntactic constructions are categorized into two parts: primary and
non-primary phrases. Primary relations or primary phrases cover the subject and main
predicate of any finite clause in addition to the complements and necessary constituents of
primary phrases (Frazier and Clifton, 1996: 41 in Papadopoulou 2006). Other types of
relations are classified as non-primary relations. The Construal theory asserts that only
primary relations are parsed in a definite manner, as explained by the Garden Path model
(Papadopoulou 2006).
Non-primary relations are parsed in an non-definite fashion and they are associated to
one domain in terms of the Construal principle including thematic and interpretation
processes. Thus, an underspecified analysis is given only to non-primary relations; in that
case, the Construal theory is about syntactic underspecification and might be viewed as a
development of parsing options that stressed the difference between argument and adjunct
attachment (Igoa et al., 1998 in Papadopoulou 2006). Crocker (1996: 220–221 in
Papadopoulou 2006) considers modifier attachment to be an example of local reanalysis,
instead of simple attachment because modifiers are not predicted.
Crocker further contends that ‘statistical or interpretative knowledge might be
employed in such cases to conduct reanalysis’ (Crocker, 1996: 221 in Papadopoulou 2006),
although he does not explain how to specifically describe modifier attachment ambiguities
in terms of quantitative analysis. In the Construal framework, the distinction between
primary and non-primary phrases is triggered by the fact that non-primary relations are not
compulsory (Frazier and Clifton, 1996: 46–47, 1997: 285–286 in Papadopoulou 2006). In
order to rapidly construct an analysis, the parser can, first of all, neglect the non-primary
relations. However, in reality underspecification of primary relations may lead to
inconsistent analyses of the sentence in disambiguating the input material. As Traxler et al.
(1998: 586) say, mis-attachment of a modifier has no serious consequences for the overall
grammatical meaning of the rest of the sentence (Papadopoulou 2006).
The Construal theory posits that relative clauses belong to non-primary phrases
because they are nominal modifiers. This argument means that, if a RC can be attached to
two heads, it will not be directly attached to the second DP, as Late Closure tends to
predict, but it will be associated to the extended maximal projection of the last theta-role
assigner. If more than one potential head is available in the present thematic domain, then
interpretative principles establish the attachment choices for the RC. (Frazier and Clifton,
7|Page
1997 in Papadopoulou 2006). A good example is given in a sentence such as (10), the last
thematic domain is determined by the lexical preposition with:
(10) Someone was looking at the servant with the actress who was on the balcony.
In examples such as (10) only the second DP is within the last theta-domain and thus,
it is the only available node for the RC and, therefore, a low-attachment choice can be
projected. This prediction has been affirmed by a number of studies across different
languages (Gilboy et al., 1995; Traxler et al., 1998, 2000 for English, Gilboy et al., 1995
for Spanish, Hemforth et al., 1996, 1998 for German and Frenck-Mestre and Pynte, 2000
for French). The consideration of thematic clues for the construction of initial input
analyses has also been used as a component in previous parsing models (cf. Pritchett, 1988,
1992) in Papadopoulou (2006).
As for a sentence such as (2) rewritten here as (11), its last thematic domain, however,
covers both DPs because the preposition “of” does not instantiate a new thematic domain:
(11) Someone shot the servant of the actress who was on the balcony.
In (11) the RC tends to be associated to the extended maximal projection of the final
theta-assigner (the first DP) and, hence both DPs should be evenly available nodes for the
RC. The Garden Path model and its refinements also use of thematic information to
explicate processing patterns discovered in various languages (Papadopoulou 2006).
D. RESEARCH METHOD
This research proposal will primarily employ a qualitative method. However, it will
also use simple statistics to compute percentages to show the distribution of strategies used
by the participants. This is a case study of how the participants interpret ambiguous
sentences in English.
Participants
The participants of this study will be eight English native speakers (four Americans and
four English people) and eight Indonesian students from the English department (all of
whom are seniors). The native speaker participants will be selected based on their
8|Page
education. Preferably, they should hold a master's degree, but at least they should have a
bachelor's degree.
Research Procedure
1. A research instrument will be prepared consisting of 15 syntactically ambiguous
sentences.
2. These ambiguous sentences will be read by a native speaker of English and recorded.
3. Each participant will be asked to listen to each ambiguous sentence once and then they
will be asked to figure out what the sentence means by choosing an answer provided by
the researcher.
4. The data will be classified and then computed in the form of percentages.
5. The quantified data will presented in tabular form and interpreted based on the existing
theories.
The processing of ambiguous sentences
by native speakers and foreign language learners of English
Data analysis
Data
Collection
Library
Research
Drawing
Conclusion
Constructing
Instrument
Field
observation
Implication and
recommendation
Possible
solution
Problem
formulation
Dissemination
Effectiveness of
Language use in
CLIL Classrooom
E. RESEARCH SCHEDULE
No. Activities
Months
August
September
October November
Decembe
r
1.
Instrumet validation
9|Page
V
V
2.
Data collection
V
V
3.
Data display
V
V
4.
Data analysis
V
V
5.
Drawing conclusion
V
V
6.
Writing research report
V
7.
Writing journal article
V
8.
Dissemination
V
F. REFERENCES
Fernandez & Cairns, 2011. Fundamentals of Psycholinguistics. West Sussex: John Wiley
& Sons Limited.
Field, J. 2004. Psycholinguistics: The Key Concepts. London & New York: Routledge,
Taylor & Francis Group.
Jackendoff, R. in Grodzinsky, Y., Shapiro, L., & Swinney, D. (Eds.) Language and the
Brain: Representation and Processing. New York, London, & Tokyo: Academic
Press.
Papadopoulou, D. 2006. Cross-linguistic Variation in Sentence Processing; Evidence from
RC Attachment Preferences in Greek. Dordrecht: Springer.
Vuong, L. C. 2010. The Role of Executive Control in Garden Path Reinterpretation.
Houston: UMI Dissertation Publishing
10 | P a g e
11 | P a g e
APPENDIX
Research Instrument
Answer the following questions:
1. Someone shot the maid of the actress who was on the balcony.
(Fernandez 2011)
Who was on the balcony in the event?
a. the actress
b. the maid
2. The student told the professor that everyone hated a lie?
What did the student tell the professor?
a. A lie
b. that everyone hated a lie
3. I met the manager of the actress who had a scandal.
Who had a scandal?
a. the manager
b. the actress
4. I saw the boy with the binoculars.
Whom did I see?
a. the boy (the binoculars were the tools that I used for seeing the boy)
b. the boy carrying the binoculars
5. Sue met the woman with her husband.
Whom did Sue meet?
a. the woman
b. her husband
6. Paul loves the woman along with her children.
Whom does Paul love?
a. the woman
b. her children
7. They were not those of the girls who went to the mall.
Who went to the mall?
a. Those of the girls
b. the girls
8. James met the father of the boys who went to London last year.
Who went to London?
12 | P a g e
a. the father
b. the boys
9. I saw the man with the telescope.
What happened?
a. I used the telescope to see the man
b. I saw the man carrying the telescope
10. I saw the mother of the girl who found the wallet yesterday.
Who found the wallet?
a. the mother
b. the girl
11. Jack spoke to the man from the village that is very nice.
Who/which is very nice?
a. the man
b. the village
12. Lucia did not like the shape of the fruit which is thorny.
Which is thorny?
a. the shape
b. the fruit
13. Daniel sat on the chair of the king that is very old.
What/who is very old?
a. the chair
b. the king
14. Jamie kissed the the mother of three children who came to the event.
Whom did Jamie kissed?
a. the mother
b. three children
15. I visited the grandmother of the girl who was sick.
Who was sick?
a. the grandmother
b. the girl
13 | P a g e
Download