lOMoARcPSD|11705684

UNIT 3 - contrastive analysis

Dẫn nhập âm vị và ngữ âm (Trường Đại học Ngoại ngữ, Đại học Đà Nẵng)

Scan to open on Studocu

Studocu is not sponsored or endorsed by any college or university

Downloaded by Hoa Nguy?n ([email protected])
 lOMoARcPSD|11705684

UNIT 3 TERTIUM COMPARATIONIS AND PROCEDURES OF CA

Identifying a common ground for comparison

All comparisons require that there be a common ground against which variation may be
noted, a constant that underlies and makes possible the variables that are identified; this is
known as the tertium comparationis (TC). In CA and translation, this tertium comparationis
is not readily identifiable.

I. Comparability criterion and tertium comparationis

All comparisons require that there be a common ground against which variation may
be noted, a constant that underlies and makes possible the variables that are identified; this is
known as the tertium comparationis (TC). In CA and translation, this tertium comparationis
is not readily identifiable.
Whenever we decide to contrast languages, we should first determine the criteria for
our analysis, because obviously two objects may appear similar or different, depending on the
category we choose to compare. Accordingly, two sentences may turn out to be similar or
different. These two sentences are very similar regarding the syntactic structure (S-V-A), but
different if we compare the use of the subject and verb in the framework of case grammar
where the subject in 1) is viewed as the affected participant while the subject in 2) is regarded
as the agent participant, as far as their participant or theta role is concerned.
e.g. M t con trâu bu c b i tre.
A buffalo stood by the bamboo hedge.
The criterion of our comparison may also be called tertium comparationis and is
determined by the purpose of contrasting a phenomenon in two languages. The comparability
criterion, one of the major theoretical concepts of contrastive studies, has to be established
prior to any analysis itself. Effectively, the analyst is supposed to answer the question what
can be compared in the observed languages. Traditionally, there are three main ways of
dealing with the problem of comparability. Originally, it used to be established either at the
semantic or formal/grammatical levels. The third way of establishing comparability criterion
assumes defining the relations of equivalence, similarity and difference in the observed
languages.
II. What exactly do they refer to?
The notion of equivalence was originally taken from theory of translation and it
involved the concept of translation equivalence. More specifically, equivalence in contrastive
studies assumes that there is a universal feature, a common platform of reference, tertium
comparationis, representing the starting point of any comparison. The actual realization of
that universal feature in the two languages is what the contrastivist is interested in. In other
words, equivalence is one of the key issues of contrastive analysis, and the basic working law
of the discipline can be presented graphically as a triangle, interrelating the contrasted
features in the observed languages by means of tertium comparationis (cf. Fig. 1 below;
Djordjevi 1987: 58).

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

A B
Fig. 1 Equivalence and tertium comparationis
Similarities and differences, on the other hand, are to be observed in the form,
meaning and distribution of the relevant language segments.
Tertium comparationis can be understood as common platform of reference enabling
the process of contrastive analysis. It refers to the third part of comparison. TC does not only
determine the possibility of comparison but also governs the result of comparison. Objects
can be compared via different features to yield the result that these objects are not only similar
in some respects but also different in others. For example, a square and a rectangle: they have
the same number of angles but different side lengths.
TC is different from the similarity in that TC is the basis of comparison without which
a comparison is impossible whereas similarities are just the result from the comparison. In
CA, besides similarity there may also be differences and these are two sides of CA with TC as
the common platform of reference against which differences can be stated. An emphasis on
the former or the latter depends on the purpose and objectives of the study whereas TC is
always the center of the comparison.
TC and equivalence are not equal either. The latter is a notion established on the
foundation of relations concerning with values and ability of substitution. To say A is
equivalent to B means A has the same value as B and can be used to substitute B. By
Krzeszowski “Equivalence and tertium comparationis are two sides of the same coin”. (Cited
from Bui Manh Hung 2008: 99).
In the classical period of contrastive analysis, comparability criterion involved two
basic relations, namely similarities and differences, and they were observed at three separate
levels: in form, meaning and distribution. This standpoint was originally proposed by Lado
(1957).
Following that standpoint, contrasted elements can be similar in form, but different in
meaning and distribution, etc. The introduction of the notion of contrast refined the
contrastive analytical process further, defining differences among the observed language in
more precise terms. Namely, the relation of contrast is to be seen in the so-called convergent
and divergent relations between the analyzed linguistic segments, while the relation of
difference was now observed in the so-called ‘zero relations’. Let us briefly have a closer look
at each of these notions.
Convergent relations between the observed language segments can be established in
the situation when two or more symbols in language A are confronted with only one symbol in
language B representing the same segment of reality. These relations can be observed at both
grammatical and lexical levels. Consequently, divergent relations are to be established in the
situation when one particular symbol in language A is confronted with two or more symbols
in language B representing the same segment of reality. Again, these relations can be

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

observed at both grammatical and lexical levels. Finally, the notion of difference in
contrastive studies is represented by ‘zero relations’ (cf. Carroll 1963). These relations can be
spotted in a situation when there is a symbol in language A labeling a certain segment of
reality and the corresponding symbol in language B cannot be found. Again, zero relations
can be observed at both grammatical and lexical levels. Fig. 2 below summarizes these basic
contrastive analytical relations fundamental in establishing the comparability criteria (cf. also
Whitman 1970; Djordjevi 1987; Kurteš 1991).

Language A Language B Type of relation

features features
Equivalence +/- +/- 1:1

Similarity +++ ++- all

Contrast ++++ ++-- all

Difference +/- -/+ 1:0/0:1

Fig. 2 Basic contrastive relations

An Encyclopaedia of The Arts Vol. 4 (9):830 - 839 (2006)
http://arts-lasu.org/publications/contents/vol4/9/kurtes.pdf
III. Types of TC
Language is a complicated system of structures consisting of various levels and
aspects. It is a sophistication of units and relations.. Thus TC in CA is said to cover equivalent
types corresponding to the comparative levels, aspects under investigation.
By Krzeszowski (1990) “2-text” in any form of text, spoken or written, in 2 languages
can be used as corpus or data for a contrastive analysis. Texts in 2 languages usually form
equivalent pairs which can be translational version to each other (2-texts [+trans] which
provides data for the qualitative contrastive studies. Besides, equivalent pairs of texts may not
be translational version to each other (2-texts [- trans]) but they must be of the same style or
genre, to mention the same topic, say the least. Texts of this type can be used for quantitative
contrastive studies.
This classification bases itself on the distinction of language and speech, which in turn
is the basis for the distinction between contrastive studies of system and those of within texts.
Typical studies within texts are qualitative studies and quantitative studies basing on the
corpus. Typical studies in systems are contrastive studies of systems, constructions, rules …
Thus, there are 7 possible types of TC corresponding to different types of CA in the scope of
texts and system.
1) Statistical equivalence:

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

The structures in two languages are considered equivalent in statistics if they occur as
translational version of each other with a highest frequency and/or, if compared with other
structures with the same meaning, their frequency of occurrence in the texts in question is
maximally the same or almost exactly the same. This kind of equivalence can be recognized
as 2-texts [+ trans] or 2-texts [- trans]. The structures that are equivalent in terms of statistics
are not necessarily equivalent in syntactico-semantics. Equivalence of this type functions as a
base for quantitative contrastive studies.
2) Translational equivalence:
Equivalence of this type is established depending on 2-texts [+ trans]. Such 2-texts
usually provide data for qualitative contrastive studies basing on corpus, and contrastive
studies of systems, which constitute the main bulk of contrastive studies. 2-texts which are not
translations, marked [-trans], can be used as data for quantitative contrastive studies.
3) System equivalence:
Equivalence of this type functions as basis for contrastive studies of systems. A CA of
systems must be executed on the paradigmatic axis. However, this must be done along with an
analysis on the syntagmatic axis because an analysis of a systems, say, system of pronouns,
will be impossible without dealing with the structures or constructions with the combination
of the units or members of the system in question. For the establishment of the equivalents in
Language A and Langue B, it is necessary to examine the equivalent structures in those
languages. In other words, equivalents in systems and equivalents in structures are not
separable. In reality, linguistic means called pronoun in English and those labeled i t in
Vietnamese are comparable for an assumption that given such a ‘common’ label, they can be
regarded as system equivalence. Under such an initial ‘shared umbrella’ as TC for the
establishment of the equivalents, further contrastive analysis will yield various specific details
of each system.
4) Semanto-syntactic equivalence:
Equivalence of this type lays the foundation for the contrastive studies of
constructions. By Krezeszowski, the equivalent can be determined on the similar basis of deep
structure known as semantic structure, structure as input for the grammatical derivation.
5) Rule equivalence:
Equivalence of this type functions as basis for contrastive studies of rules. Like
contrastive studies of systems, any comparison of rules cannot be divorced from an implicit
comparison of constructions on which these rules operate. By Krezeszowski, rules are
interpreted in the view of Transformation-Generative Grammar. They are rules to form Phrase
Structure Rules, Transformational Rules. When sentences undergo or experience change
according to the similar formal rules in the transformation process, they are said to have
similar rules. Most rules have a construction as the input and a construction as the output.
Therefore, semanto-syntactic equivalence also underlies rule equivalence.
6) Equivalents in objects:

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

These equivalences concern with objects or entities outside language, expressed by

units of phonology, vocabulary of the language in use. Equivalents of this type functions as
basis for contrastive studies of phonology and vocabulary which are done on the paradigmatic
axis at the expense of the syntagmatic axis. Krzeszowski stresses that these entities are not
simply physical entities but mental psychological image in the language user.
7) Pragmatic equivalents:
Equivalence of this type functions as basis for contrastive studies of pragmatics,
stylistics or socio-linguistics. These are relations between texts of two different languages
which illicit from the language user the maximally similar cognitive effects.
Among those TC mentioned above, formal equivalences are the least important, for an
assumption that these equivalences of means cannot be regarded as TC. By Wilems et al.
(2004) the comparison of the linguistic devices of languages is significant only if these
comparative devices or means have a function that is comparable to each other.
We can now see that tertium comparationis is in fact the reason why any two texts are
brought together as a 2-text and/or why any two items in two languages are juxtaposed for
comparison. Each type of contrastive studies has its own types of tertium comparationis.
Within each type of comparationis, it is possible to distinguish more specific subtypes,
subsubtypes, etc., unique within each type. Each type of tertium comparationis is connected
with a specific type of equivalence.
Questions:
1. In what way can similarity/difference and TC be distinguished?
2. In what way can equivalence and TC be distinguished?
3. State types of TC.
4. What can be TC in a CA of sound systems of two languages?
5. Explain this statement: “Any comparison presupposes similarity as tertium comparisonis”
(Krzesowski, 1990)
6. What is the most common type of equivalence?
7. Can we conduct a contrastive study with such as title as “A Comparison/Contrastive
Analysis of Articles in English and Vietnamese?”
If yes, please provide the reason. If no, please rephrase the title so that it can conform to the
knowledge and principle of TC.
8. What is TC in the contrastive study with the title “A contrastive analysis of Pronouns in
English vs. Vietnamese”?

Glossary:
An Encyclopaedia of The Arts Vol. 4 (9):830 - 839 (2006)
Comparability criterion: the starting point in the contrastive analytical process which
subsumes establishing what is comparable in the two languages.

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

Contrast: a contrastive relation referring to a relative low degree of likeness between the
analysed grammatical segments of the two languages. This is observed in the so-called
divergent and convergent relations.
Contrastive analysis: a branch of theoretical linguistics and a principle of applied linguistics
whose aim is to ascertain in which aspect the observed languages are alike and which they
differ, based on a systematic comparison of their grammatical structures.
Difference: a contrastive relation referring to the situation in which there is no corresponding
category in language B for the category found in language A. This is also known as a zero
relation.
Equivalence: a contrastive relation referring to the relative sameness in meaning.
Similarity: a contrastive relation referring to a relative high degree of likeness between the
analysed grammatical segments of the two languages.
Tertium comparationis: a common platform of reference enabling the process of contrastive
analysis. Literally, it is the “third term of a comparison”; that which remains invariant in
translation or in contrastive analysis, which forms the basis for the comparison
textual equivalence
Definition: relation that is observed to exist between an element of a source text and a
corresponding element in its translation, as accepted by a competent bilingual
Source: Catford 1965
translation equivalent
Definition: expression in a target language which can translate a source-language expression
in certain contexts

IV. Procedures of CA
Traditional contrastive methodology subsumed two basic processes – description and
comparison. Krzeszowski (1990), however, speaks about three main steps in classical
contrastive studies – description, juxtaposition and comparison.
1. Steps in Contrasting Two Language Systems
These four steps are relevant to all levels of languages structure, namely, syntax,
lexicon, phonology, pragmatics and discourse.
1) Description
The first step in executing a contrastive analysis is to provide description of the
aspects of the languages to be compared.
Description includes the selection and preliminary characterization of the items under
comparison in the framework of language- independent theoretical model.
No comparison is possible without a prior description of the elements to be compared.
Therefore, all contrastive studies must be founded on independent descriptions in that they

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

should be made within the same theoretical framework. It will not do to describe one
language in terms of transformational grammar and another language in terms of, say
relational grammar and then to attempt to compare them. The results of such descriptions will
be incompatible and incomparable.
Not all linguistic models are equally well suited as foundations of cross-language
comparisons. It seems that those models which make explicit references to universal
categories are more suitable than those which are connected with language isolationism,
inherent in many variants of structuralism.
The minimum requirement of ‘parallel description’ is that the two languages be
described through the same model of description. Why, we may ask, must the two
descriptions be framed in the same model? There are several reasons: First, different models
can describe certain features of language more successfully than other models/ We saw
instances of this in T-G Grammar which can effectively account for native speaker’s
intuitions that certain construction-type are somehow related (Active and Passive sentences,
for example) and that certain others are ambiguous (e.g. She’s a beautiful dancer); Case
Grammars, on the other hand, provide apparatus for explaining the semantic affinity between
a pair of sentences like
This key opens that door
and
That door opens with this key.
Now, it follows that if the ‘same data from L1 and L2 are described by two different
models, the descriptions are likely to highlight different facets of the data. When this happens,
the subsequent comparison will be unnecessarily difficult, and, what is more serious still, the
analyst will be uncertain of the status of the contrasts he identifies: are they linguistic
contrasts, in representing differences between the L1 and L2 data? Or are they reflections of
the use of two different models, i.e. description-induced rather than data-induced contrast? It
was for this reason that Harris (1963:3) insisted that comparable descriptions of two
languages will only be guaranteed if identical ‘methods’ of description are used for
description of the two: “since any differences between these descriptions will both be due to
differences in method used by the linguists, but to differences in how the language data
responded to identical methods of arrangement’.
Linguistic typology tells us that human languages fall into several types according to
which grammatical, phonological or lexical features they show preferences for. If some
models are better at describing certain features, it must follow that some models will describe
certain languages better than others. It is possible that T-GG, a product of American
Linguistics, describes English better than it describes other languages. It seems that
Applicative Generative grammar, a model devised by the Soviet linguist Shaumjan (1965) is
eminently better suited to describe Russian, a language with a complex morphology, than it is
to describe English. Obviously, distortion would result if we did a CA of Russian and English
using a model which favours one of these languages at the expense of the other: the
descriptions, while being ‘parallel’, would be unequal.

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

We seem to be faced by a dilemma, then. On the one hand, there, there are good
theoretical reasons for using the same model for yielding the descriptions of L1 and L2; on
the other hand there are equally cogent practical reasons why this undesirable. There would
seem to be two ways out of the dilemma: bilateral CA and unilateral CA.
i) Describe L1 and L2 data independently, using the models which yield the fulliest
descriptions of either language, and then translate these two descriptions into a form which is
model-neutral. There is a precedent for this in Translation Theory, where use is made of an
artificial ‘etalon language’ (Melchuk, 1963:62) which is a neutral intermediary between L1
and L2; in fact it is a composite of the two, or ‘supralingual, in containing the features both of
the L1 construction and of the L2 construction. Catford (1965: 39) illustrates this convention
(see page 65) in comparing an English and a Russian sentence which hare transformationally
equivalent.
Note that the English construction selects from the etalon features 1contrastive studies
3, 5 and 6, while the Russian selects 1, 2, 3, 4, 5 and 7. These sets of features are those which
a good grammar of either language would generate all seven of.
Features in
Sentence (E) the Etalon Sentence (R)
I__________________________1 speaker _____________________________ ja
2 female
have arrived 3 arrival prishla
4 on foot
5 anterior
6 current relevance
7 completed
ii) A second solution would be to abandon the requirement that the two description
need to be equally exhaustive, or, to use Halliday’s term (1961:272) ‘delicate’. A number of
contrastivists have suggested that a CA should indeed show a descriptive imbalance, in favour
of the L2. Sciarone (1970:126) points out that “If both languages are described beforehand,
too much, ie. superfluous work is done for the sake of CA”. He suggests that less attention
needs to be paid to the L1 than to the L2, since it is the latter which must be learnt. Slama
Caxaru (1971) suggests a “procedural adjustment: of CA which she terms ‘contact analysis’:
we should be more concerned with what the learner does with the L2 than with what linguistic
knowledge (the L1) he enters the learning situation. Filipovic (1975:15) openly assert that his
CA of Serbo-Croatian (L1) and English (L2) has been descriptively biased toward the latter.
This unilateral CA is carried out with the contrasted languages that are not equal for
the execution of the contrastive techniques. The unilateral CA can be done with 2 phases:
- The first phase: Establish the subsystem for CA in Language 1.

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

- The second phase: List out the language means in Language 2. These language
means are used to mark or express the meanings of the factors, categories of the
subsystems in Language 1.
For example, the comparison of ‘possessive case’ in Vietnamese and English should
start with the establishment of the possessive meaning of Noun in Vietnamese in specific
instances (e.g. tình yêu c a Lan) then the CA will go on with the examination of the number
of linguistic devices to express this possessive meaning in English (e.g. by preposition of and
the suffix ‘s bearing the possessive meaning.
Thus, CA of this type is often imbalanced in its nature. In a language, the object of CA
is one linguistic form whereas in the other contrasted language, various forms have to be
listed to form a system of means to express the meaning conveyed by the only form in the
source language.
As Nguyen Van Chien (1992) has put, the unilateral CA can be beneficial to the
foreign language beginners who are typically said to express or represent his/her ideas in
mother tongue in his/her mind, and then automatically translate these ideas into the target
language.
Unilateral CA can become the most essential technique in the case where the target
language there is no equivalent subsystem to the subsystem in the source language or mother
tongue. For instance, for the CA of English and Vietnamese, the morphological category of
number can be found in the former but this category is absent in the latter.
English Vietnamese
one book∅ m t quy n sách∅
two books hai quy n sách∅
n books n quy n sách∅
a box∅ m t cái h p∅
2 boxes hai cái h p∅
Accordingly, in this case, the first step of a typical unilateral CA is:
- to describe the system of system of number of English (analyse the number meaning
of nouns in English and the suffixes denoting these plural meanings of nouns);
- to point out the linguistic means to denote the plural meaning of nouns in English in
Vietnamese (Vietnamse has at its disposal such lexical means (determiners) as
nh ng/các/m i/nhi u …). These means are counted as qualitative information).
The result from a unilateral CA based on the detailed description will help the English
learners master different ways of expressing the plural meanings of nouns in Vietnamese.
As a complementary technique to a unilateral CA, a bilateral CA can help to solve the
problems of the imbalance in the analysis of the contrasted languages. Basing on this
approach, the contrastive analysis can point out the linguistic means to denote the meanings of
category of the units in L1 by means of L2, and “simultaneously” list out the linguistic means

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

to express this category in L2 by means of units in L1. In fact, these are two successive phases
of a parallel CA. For in stance, a CA of the possessive case in Vietnamese and English, it is
necessary to point out the linguistic means to express the possessive meanings of nouns of
Vietnamese in English, then we have to look for the means to denote the possessive meanings
of nouns of English in Vietnamese.
With the aim of second language teaching, the description of the linguistic means to
express the categories of the mother tongue (L1) in the target language (L2) is at most
important. It is because the result of CA of this type will help the learners prevent the negative
transfer or interference in expressing the meanings of a certain category of a foreign language
with linguistic means of that language.
A bilateral CA is impossible without the balance in means or ways of expressing
categories of the linguistic units in L1 and L2: In some cases, a certain category of the
linguistic units in one of the contrasted language is expressed by the open-system of units
(lexical units) whereas the same category can be expressed by a close-system of units
(grammatical units). In other case, a certain category of the linguistic units in one of the
contrasted language is expressed by the synthetic means (affixes) whereas the same category
can be expressed by the analytic means (determiners). For example, the category of
possessive meaning in Vietnamese can be expressed by the analytic means (Tình yêu c a
Lan) whereas this meaning can be represented in English by either synthetic or analytic means
(Lan’s love; the love of Lan; the love for Lan).
This technique approaches the CA of language phenomena, facts by searching the
linguistic means to express the categories of notions in the contrasted languages.
2) Juxtaposition
Juxtaposition is a step where one decides what is to be compared with what. "The first
thing we do is make sure that we are comparing like with like.”
Juxtaposition involves a search for, and identification of cross/inter-linguistic/cultural
equivalent while the comparison proper evaluates the degree and types of correspondence
between items under the comparison.
This step is crucial in deciding what is to be compared with what. In classical
contrastive studies, this step was based on intuitive judgements of bilingual competence, i.e.
the knowledge of two languages, enables one to make decisions about whether or not element
X in one language is equivalent with element Y in another language. If the two given
elements are equivalent, they are said to be comparable. For example, anyone competent in
English and in Vietnamese intuitively knows on the basis of his “bilingual competence”, that
such ng i mà and cái mà in Vietnamese and which/who in English are equivalent, given
appropriate contexts. Likewise, “bilingual competence” manifests itself in judging the
following pair of sentences as equivalent:
(1) I want John to come.
(2) Tôi mu n John n.

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

Such judgments are taken for granted in classical contrastive studies, so that elements
recognized as equivalent are intuitively deemed to be comparable. One of the obvious
weakness of this approach consisted in the lack of clearly stated principles underlying
decisions about what to compare and why. Formal resemblance and semantic resemblance
were resorted to, but both, as we have seen, led to circularities: similarity was presupposed
before comparisons yielded results allowing to ascertain it. In many instances, formal
resemblance (at least at the level of surface structures) so drastically contrasts with the
disparity of meaning that comparisons based on formal criteria alone are reduced ad
absurdum. Consider one of the early examples quoted by Stockwell et al. (1965: 40):
(3) English: I said to be sure.
(4) Spanish: Dije estar seguro. ‘I said I was sure.”
Although formally very similar, (3) and (4) are semantically very different. Therefore, they
are incomparable since they do not share a semanto-syntactic tertium compartionis. This
example shows again that formal considerations alone do not suffice in establishing
comparability. Therefore, juxtapositions based on formal criteria alone, though naturally
possible, are ill-conceived and must be discarded in contrastive studies.
In classical contrastive studies, the investigator himself often acts as the bilingual
informants and decide what to compare on the basis of his own knowledge of the two
languages. Unless more criteria constraining the data are applied, such a procedure often leads
to arbitrary decisions, which seriously undermine the rigor required in scientific
investigations. Contrastive Generative Grammar attempts to make explicit “bilingual
competence”, underlying intuitive judgements of bilingual informants.
3) Comparison
In the comparison stage, the actual comparison and contrast of the two systems or sub-
systems are performed. Not always are the two steps of juxtaposition and comparison are kept
discrete.
Here again we encounter a number of theoretical problems, mainly surrounding the
issue of criteria for comparison, or the tertium comparationis. We concentrate here on how to
compare rather than on what basis to compare. Admittedly, this is a somewhat arbitrary
approach, since the ‘how’ and ‘why’ are inextricable.
We compare ‘types’ rather than ‘tokens: that is, to refer again to Catford’s example
above, we do not compare these sentences as strings of sounds or graphic substance, but their
structures. Their structures are:
Pronoun + 1st Person + Sing – Auxiliary – Past, Participle
v v
I have arrived.
Pronoun + 1st Person + Sing – Prefix+ Verb+ Perfective + Past + Feminine
v v
Ya prishla

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

Any structure, being idealization, represents an infinite number of possible

realizations: if the structure is a sentence, it is the basis of many utterances, as Lyons (1968)
points out. He explains the difference by reference to de Saussure’s famous distinction
between parole and langue: “Utterances stretches of the parole produced by native speakers
out of sentences generated by the system of elements and rules which constitute the langue.”
From the premise that CA compares abstract elements rather than their concrete
realizations it follows that each of its statements has very broad coverage of potential
utterances. We shall now illustrate how CA utilizes parallel description and comparison of
types in L1 and L2.
We distinguish three basic areas of comparisons:
1. Comparisons of various equivalent systems across languages, such as pronouns, articles,
verbs, and in phonology consonants, vowels, as well as subsystems, such as nasals, laterals,
etc. depending on the degree of “delicacy” of the grammar.
2. Comparisons of equivalent constructions, for example, interrogative, relative, negative,
nominal phrase, etc. and in phonology clusters, syllables, diphthongs, and various
distributions of sounds.
3. Comparisons of equivalent rules (in those models where the concept of rule appears), for
example, subject raising from the embedded sentence, adjective placement, interrogative
inversion, passivization, etc., and in phonology assimilation, dissimilation, metathesis, etc.
(cf. Sussex 1976:7 – 11).
In each area of comparison one of three possible situations may arise:
(1) XLi = XLj
when item X in Li may be identical in some respects with an equivalent item in Lj.
(2) XLi XLj
when item X in Li may be different in some respects with an equivalent item in Lj.
(3) ( XLi + - ∅Lj
when item X may be present in Li but absent in Lj.
In fact, (1) and (2) are just two facets of one possibility, for the assumption that two
linguistic phenomena supposed to be identical in some aspects can possibly be different in
some others, and vice versa. This is necessary because as in (1), the contrastivist may consider
the similarities of the two phenomena compared more important. As for (2) the differences are
said to be more important. It is implied that in (1), (2) and (3) there is no case for the systems,
constructions or rules to be completely similar in the two languages compared. In Russian and
French it is noted that gender is a grammatical category to mark the subtypes of nouns.
However, a close examination of the nouns in the two languages still shows the detailed
differences: the category of gender in Russian is established with a three-way distinction
masculine-feminine-neuter whereas the category of gender in French can be characterized
with a two-way contrast masculine-feminine.

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

In respect of the specification and the ‘inclusiveness of reference’ (Hawkins), the

article ‘các’ of Vietnamese is said to be similar to ‘the’ of English, as in
a) Please remove the books from the table.
b) Please remove some books from the table. (Dik 1989)
in a) ‘the books’ can be interpreted as all the books on the table whereas in b) ‘some books’
refers to not all the books on the table. Similarly, in Vietnamese, a noun phrase with ‘các’ is
presupposed to refer to all the individuals mentioned by the noun without any exclusion of
any other individual as compared with a noun phrase with ‘m t s /nh ng’, as in
a) Các sinh viên ã có m t.
b) Nh ng sinh viên ã có m t.
c) M t s sinh viên ã có m t.
However, there is a difference in the function of marking and collocation between
‘các’ and ‘the’: the former is restricted to the marking of plurality of a noun and it cannot
combine freely with any noun whereas the latter can be neuter in number and is free in
combining with nouns. (Bùi M nh Hùng, 2000).
In reality, there may be a possible situation as (3) where the equivalent of X in L1
cannot be found in L2. However, the so-called absence of the equivalent of X can be
interpreted in different ways; accordingly the next steps of contrastive analysis can be
executed or not. Two possible situations can be mentioned here:
3a) When X in Li is a phonological means or morphological means and does not indicate any
meaning in Lj, and
3b) When X in Li is a linguistic means bearing meaning which is said to be universal or at
least present in Lj.
In the situation 3a) the so-called no equivalent can be interpreted as transparent as it
may mean. For example, the presence of tone in Vietnamese and the absence of this prosodic
means in English.
On the other hand, the contrastive analysis can be conducted where X is a unit bearing
meaning in Li, as mentioned in 3b) though it is absent in Lj. For example, in contrasting
English and Vietnamese in the respect of tense which is present in the former but absent in the
latter, the researcher has to point out that while English makes use different forms of tense to
mark the time, Vietnamese may make use of other means to fulfill the same function.

Further Reading (Cited from Bui Manh Hung (2008)

Basic approaches in contrastive analysis in languages
We can execute a bi-lateral or unilateral contrastive analysis depending on the purpose
and task of our research study.
A bilateral or multi-lateral contrastive study examines the comparable phenomenon in
two or more languages in an inter-relation on a common platform for contrast, depending on a

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

specific TC. First, we select a TC and determine the linguistic means which belong to the
category in the languages compared. For this procedure, we have to answer which means are
available in Li and Lj to indicate or belong to the category selected as TC.
Due to the lack of explicitly stated comparationis expressed in some universal terms
(such as a semantic representation of the compared items), typical classical contrastive studies
were directional: depending on the aims of a particular contrastive analysis, one could start
with a description of linguistic forms in L1 and match them for comparison with equivalent
items in L2 and look for their equivalents in L1. Presumably an exhaustive contrastive study
concerning the entire grammars of both languages has to be bi-directional. Typical directional
contrastive studies would bear such titles as {system X/Construction Y} in Li and its/their
equivalents in Lj, for example, “English modal auxiliaries and their equivalent constructions
in Polish” …
Establishing semantic tertium comparationis creates the possibility of adopting an
alternative approach, which consists in selecting a concept and examining the ways in which
it is realized through various grammatical means in comparative languages. Typically, such
contrastive studies have the titles of the form Ways of expressing a category X in L1 and L2.
For example, “The expression of future in English and Serbo-Croatian” (Kalogjera 1971),
“Ways of expressing cause in English and Polish” (Danilewicz 1982) or “Directives in
English and Finnish” Markanen 1985).
In bilateral CA, in the scope of tertium comparationis we should deal with one or
some specific meanings. The other meanings that do not fall within the scope of TC in
question will not be mentioned. For example, a study of the linguistic devices to denote the
specification of noun with articles should be restricted to this function of the articles,
accordingly the other functions, if there are any, such as the distinction of gender, number and
substantivisation will not be dealt with.
The so-called bilateral CA is not due the fact that the execution of the contrastive
analysis is done starting from language A then with Language B, and vice versa. In fact, the
bi-lateral CA examines the linguistic devices in 2 comparative languages, then analyses the
similarities and differences in these two languages on the presupposition that neither of them
is considered the source or target language. This is illustrated in the diagram below:
Language A TC Language B

X1
X1
X2
X2
X3
X3
X4

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

(Adapted from Sternemann et al. 1989, cited from Bui Manh Hung, )
According to the diagram, TC has 4 linguistic devices or means in Language A and 3
in Language B. The diagram shows the relations of convergence or divergence of the
linguistic means in the two languages with the crossing lines linking X1, X2, X3, X4 of
Language A and X1, X2, X3 of Language B. The vertical lines represent the contrast of the
linguistic means within each language.
On the other hand, unilateral CA just examines the meanings of a certain linguistic
means or form in Li and identifies or determines the means that represent the equivalent
meanings in Lj. The CA can be executed with a description of the linguistic forms in Li first,
and then contrasts these with the equivalents in Lj, or vice versa.
This approach is considered unilateral because the contrastivist has to select one
language as the source language and the other as target language. The selection depends on
the aim and purpose of the study. This is illustrated in the diagram below:
X
(in Language A)

meanings
(of X)

Y1 Y2 Y3 Y4 Y5
(in Language B)
(Adapted from Bui Manh Hung, 2008:162)
The diagram shows that the form X in Language A denotes various meanings, and Language
B makes use of 5 different means to denote these meanings.
In CA with this approach, when Language A is selected as the source language, the
result of CA should be presented with the reference of the similarities and differences in the
priority of Language A, i.e. we should say Language B is similar or different from Language
A in terms of a certain aspect, but not Language A is similar or different from Language B.
On the other hand, in bi-lateral CA, we can present the result of CA by referring to the
similarities and differences in terms of a certain aspect in expressing a TC, i.e. we can say
Language A and Language B are similar or different in terms of a certain aspect.
Typically, unilateral contrastive studies have the titles of the form System X/
Construction Y in Language A and the equivalent system/constructions in Language B. For
example, Tag-Questions in English and the Equivalent structures in Vietnamese, the Passive
sentences in English and the equivalent structures in Vietnamese …
As Bui Manh Hung (2008) states, in bi-lateral CA, we cannot carry out such a study as
‘A Study of the consonant systems in English and the equivalents in Vietnamese’ because we
cannot describe the English system of consonants and then base on the equivalents of these
English consonants in terms of functions to establish and contrast the equivalents in

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

Vietnamese with English consonants. A TC such as Consonantal Systems should be

established as first place for a bi-lateral CA.
The execution of CA in the bi-lateral or unilateral approach is supposed to yield
different results, accordingly we can select a CA with one of these approaches depending on
the aim and problems under investigation. A bi-lateral CA of adverbs in 2 languages such as
English and Vietnamese (given that Vietnamese has adverbs) will be formulated with the title
like “Adverbs in English and Vietnamese”. Then TC is established within the scope of
adverbs. A typical unilateral CA of English and Vietnamese will have the title “Vietnamese
adverbs and the equivalent construction/patterns in English”.
If a bi-lateral CA is executed, as in the study of adverbs as mentioned above, all the
patterns identified as adverbs in the two languages will fall within the scope of description
and comparison. On the other hand, if a unilateral CA is executed with the study of adverbs,
the scope of description and comparison will cover all the constructions or patterns that are
identified as adverbs in Vietnamese and the equivalents in English.
Modern contrastive analysis introduces some methodological innovations into its
analytical framework (Chesterman 1998), essentially drawing from Popper’s view expressed
in his philosophy of science (Popper 1972). According to this view, objective knowledge is
gained through an endless process of problem solving, basically consisting of suggesting,
testing and refuting initial hypotheses, which are revised and tested again, etc. Following this
line of argument, a new methodological framework is proposed, its main stages being the
following:
1) Collecting primary data against which hypotheses are to be tested. Primary data
involve all instances of language use, utterances that speakers of the languages in
question produce.
2) Establishing comparability criterion based on a perceived similarity of any kind.
3) Defining the nature of similarity and formulating the initial hypothesis.
4) Hypothesis testing: determining the conditions under which the initial hypothesis
can be accepted or rejected. This process will normally include selection of a
theoretical framework, selection of primary and additional data and use of corpora,
appeal to one’s own intuition or other bilingual informants, even the results of error
analysis of non-native usage.
5) Formulating the revised hypothesis.
6) Testing of the revised hypothesis, and so on.
These contrastive formulations can be successfully tested by finding them in a corpus
or checking the behaviour of speakers. The real task for the contrastivist is to specify the
conditions under which the formulations are valid, which is essentially in traditional
contrastive studies known as the contrastive rule. Depending on the comparability criterion,
these conditions can be syntactic, semantic, pragmatic, stylistic, contextual, etc. (Chesterman
1998).

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

4) Prediction
Under the influence of the mother tongue the differences are transferred into the
learner's language – i.e., interlanguage – hence, interference is created in certain deviant
structures that are expected to be generated.
This expectation is called prediction. But how do these deviant forms present
themselves? The general assumption is that deviant structures reflect the structure of the
mother tongue.
Lado (1057) states that “The plant of this book rests on the assumption that we can
predict and describe the patterns [of L2] that will cause difficulty in learning and those that
will not cause difficulty”. Odler (1971: 79) again speaks of CA as “… a device for predicting
points of difficulty and some of the errors that learners will make”. By Carl James (145) there
seems then to be three things that a CA can predict – in the sense of ‘pre-identify’ – what
aspects will cause problems; or it can predict difficulty; or it can predict error, and in his view
there is a suggested fourth possibility: of CA predicting the tenacity of certain errors, that is,
their strong resistance to extinction through time and teaching.
However, the phrase “Prediction of errors” may be ambiguous because it may mean
that there will be error or prediction of the form of that error. Obviously, to claim that CAs
have predictive capacity of the second kind would, given the present ‘state of the art’, be quite
presumptuous. According to James (1980:146), rather than risk making wrong predictions
about the form of errors, contrastivists have more cautiously made predictions of an either/or
type: learners with a certain L1 leaning this L2 will produce either x or y or y types of errors.
There are, of course, purely quantitative limitations on the numbers of learner errors
that CAs can predict, limitations stemming from the fact that not all errors of L1 interference,
i.e. interlingual errors. Other major sources of errors have recognized (Selinker, 1972;
Richards, 1974) which are of a ‘non-contrastive’ origin. These include:
- the effects of target-language asymmetries (intralingual errors);
- transfer of training;
- strategies of L2 learning; and
- communication strategies
Given that a CA predicts “behaviour that is likely to occur with greater than random
frequency” (Lado, 1968:125) about 60% of the third to half of all errors, it will not try or
claim to predict the other 70 to 80%. One must be careful not to exaggerate the claims made
on behalf of CA. (James: 146).
There is a further aspect of their predictive capacity that is of great pedagogical
relevance: this is their alleged capacity to predict a scale of incremental difficulty. If this scale
can be validated, it will have powerful implication for pedagogic Grading and for Evaluation
(Testing).
Scale of Difficulty

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

The most well-known hierarchy of FL learning difficulty is that proposed by

Stockwell & Bowen (1965) for phonology, and again, with certain elaborations by Stockwell,
Bowen & Martin (1965). Attempts to design scales for the level of vocabulary are those of
Higa (1965) and Rodgers (1969). The Stockwell et al. Scale based on the conditions the
notions of positive and negative transfer potential, and conditions for such transfers are
assumed to be statable in terms of the relations holding between matched rules of L1 and L2
(cf. p.172). There are three possible interlingual rule relationships:
a) L1 has a rule and L2 an equivalent one.
b) L1 has a rule but L2 has no equivalent.
c) L2 has a rule but L1 has no equivalent.
The second step is to identify the types of choices that either language makes
available, and relating these choices. There are three types of choice: optional, obligatory and
zero (∅).
Hierarchy of Difficulty:
Optional choice: possible selection among phonemes, e.g. English can have /p/ or /b/ word
initially
Obligatory choice: the selection of conditioned allophones and the limitations in
distribution of phonemes:
- English word initial /p/ must be aspirated
- /s/ but not /z/ and before /m/ at the beginning of a word /z/
- distribution of / / or / /
Zero choice: existence of a certain sound in one language that has no counterpart in
another language, e.g. Vietnamese has no counterpart for / / in English.
An optional phonological choice “refers to the possible selection among phonemes”:
one is free, in English and German, to choose either / / or / /, etc. in word-initial position, to
say (English) show/so, (German) Schau/Sau. Russian allows the free choice of either on …..
or …..to express future reference. An obligatory phonological choice involves little freedom,
since phonetic context determines which of a set of allophones is required to represent freely
selected phonemes: thus /l/ and / / are optional choices in Russian while [ ] and [ ] as
realizations of / / are each obligatory choices in English.
These different availabilities of choice in L1 and L2 allow eight kinds of relationship
between the two languages: the result is an eight-point hierarchy of difficulty, which is
simplified to a scale of three orders of difficulty by coalescing 123(I), 456(II) and 78(III):
Order of Difficulty Comparison of Choice Type
Most L1 L2
1 ………. ∅ Ob
I 2 ………. ∅ Op

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

3 ………. Op Ob
4 ………. Ob Op
II 5 ………. Ob ∅
6 ………. Op ∅
7 ………. Op Op
III
8 ………. Ob Ob
Least
Note: Op = Optional; Ob = Obligatory
About the Hierarchy:
- Does it make predictions?
- Are these predictions testable?
- How can they be tested? Error counts? Production tests, perception tests?
- Originally developed as a guide to curriculum development
How are counterparts determined? Same phonetic symbol? Same grammatical
category? Same translated meaning. An important ingredient of the teacher’s role as monitor
and assessor for the learner’s performance is to know why certain errors are committed. It is
on the vais of such diagnostic knowledge that the teacher organizes feedback to the learner
and remedial work. Even the learner should know why he has committed errors if he is to
self-monitor and avoid these errors in the future.
Wardhaugh (1970) suggested that the CA hypothesis is only tenable in its ‘weak’ or
diagnostic function, and not tenable as a predictor of error: “The weak version requires of the
linguist only that he use the best linguistic knowledge available to him in order to account for
observed difficulties in second language learning” (Wardhaugh, op. cit.:126) and “reference is
made to the two systems (L1 and L2) only in order to explain actually observed interference
phenomena” (ibid.: 127). Since there are very few published CAs of such on-the-spot ad-hoc
mini CAs anyway. The purpose of doing them is to see if a particular attested error is
explicable in terms of L1 interference. If no L1 structure can be found that the structure of the
errors seems to be a reflection of, then we have to start the long job of finding some cause -
other than L1 transfer. One is certainly given an illuminated short-cut when the L1 suggests
the obvious source of the error.
5) Testing
One of the requirements of a good language test is that it should have validity: it
should be a true measure of the student’s command of the language he has been taught. The
most valid test therefore would be one that was comprehensible, i.e. it would test everything
that has been taught. For obvious reason such a test would be impracticable to administer to
students after their first week or two of instruction. Therefore we must attempt to achieve test
validity by testing a representative sample of the student’s repertoire. This is where CA has a
part to play, and Lado (1961) based his theory of testing to a considerable extent of CA.

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

Testing experts since Lado have endorsed his approach: “If a test is constructed for a single
group of students with identical language background and identical exposure to the target
language then contrastive analysis is essential” (Davies, 1968: 12).
CA will have two roles to play in testing. First, since sampling is required, it will carry
suggestions about what to test, and to what degree to test different L2 items. If items
isomorphic in L1 and L2 are assumed to be easy for the learner, they can be bypassed in the
test. It will be more informative for the tester to test only the learning problems predicted by
the CA. As for the degree to which to test, it depends on the level of the learner, but a test for
the intermediate student that is CA-based should contain more items of, say, difficulty levels
4,5 and 6 on the Stockwell , Bowen Scale (q.v.) than items of difficulty levels 1 and 2.
Turning to the matter of how to test, if a multiple-choice type of objectives test is
being constructed, a CA of L1 and L2 will suggest the types of distracters to use: as Harris
says: “The most effective distracters in a test item will be those which evoke first-language
responses from those subjects who have not fully mastered the very different patterns of the
target language” (Harris, 1968: 39). For example, since Vietnamese has no participle form
corresponding to English participle form of adjective ‘boring’, expressing inherent
characteristic of an entity but uses the same form ‘bored’ (for person) in such cases, the
Vietnamese will tend to say the erroneous:
* The film was very bored
for the intended
The film was very boring.
corresponding to Vietnamese
Cu n phim này th t chán.
Therefore a discrete-point test of the English participle forms of adjectives for Vietnamese
learners ought to contain at least one distracter evoking * Ving, Ved forms. It is less obvious
how CA predictions might inform the writer of the ‘integrative’ test that are in vogue today:
cloze tests and noise tests for example; but it is not inconceivable that a cloze test could be
designed in which only those elements of the L2 test are deleted which are predictably
difficult for learners of a given L1 to operate: for instance, deleting the articles in an English
test for learners whose L2 is Vietnamese or Chinese.

Further Reading
It is reported that different things are not always the most difficult ones. Students'
perception of difficulty does not always correlate with CA predictions.
That is why the framework of CA we have been explaining and that we will be using
is called the Strong Version of Contrastive Analysis Hypothesis. This is a version in which
practically most, contrastive analysis activities are performed.
Two other versions, namely weak and moderate, are named in the literature which are
not well cultivated yet.

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

The strong version of CA holds that the degree of difficulty correlates with the
intensity of differences between the two structures in L1 and L2. However, the moderate
version claims that minimally distinct structures are more problematic for learners.
Procedures of CAH
Whitman (1970: 191) breaks the contrastive analysis down to a set of component
procedures. The four steps are (1) taking the two languages, LI and L2, and writing formal
descriptions of them (or choosing descriptions of them), (2) picking forms from the
descriptions for contrast, (3) making a contrast of the forms chosen, and (4) making a
prediction of difficulty through the contrast. Here, the term "form" refers to any linguistic unit
of any size. To describe the prediction stage, Stockwell et al. (1965) propose a "hierarchy of
difficulty" based on the notions of transfer (positive, negative, and zero) and of optional and
obligatory choices of certain phonemes in the two languages in contrast. When the structures
of the given two languages are similar, positive transfer will occur while with those that are
different, a negative transfer will take place. Where there is no relation between those
structures of the two languages, zero transfer will occur. When an English speaker selects a
word among phonemes /p/ or /b/, an optional choice occurs. On the other hand, when he has
/p/ at the beginning of a word, he should choose the aspirated allophone [ph] in that
environment, which is called an obligatory choice. Stockwell et al. used the following criteria
to establish the "preferred pedagogical sequence":
(1) Hierarchy of difficulty
(2) Functional load
(3) potential mishearing
(4) pattern congruity.
Hammerly (1982 : 26) described as "adequate" the a priori hierarchy of difficulty by
Bowen et al. in representing initial difficulty with a second language sound system. He
proposed his own hierarchy that represents the hierarchy of difficulty in terms of the
persistence of pronunciation errors after considerable instruction. His hierarchy is classified
into forty-five items by mean error.
Hierarchy of difficulty (after Prator)
Fries noted that L1 and L2 learning are very different tasks, new set of habits against a
background of old habits, as opposed to no habits at all. Lado held that similar structures will
transfer easily and different structures will cause interference.
Assumptions of CA Language is a habit Major source of L2 error is L1 The greater the
number of differences, the greater interference and learning difficulty L2 involves learning the
differences
The following hierarchy of difficulty is not necessarily predictive of difficulty, which
calls into question the entire theory of contrastive analysis. The problem with the theory is
that similar structures may often prove more difficult to acquire, since the differences are
difficult to perceive. A common Spanish-English contrast is aplicación, which means

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

'diligence,' and not application. These kinds of close correspondences cause many problems
for students.
Despite the problems with CA theory, the categories identified are found useful by
teachers and students alike. The teacher can use these categories to organize the material to be
taught into logically arranged groupings.

Level 0 Transfer No contrast between the languages: mortal > mortal

(sometimes called ‘correspondence’)
Level 1 Coalescence Two items in L1 become one item in L2:
his/her/your/their > su
Level 2 Underdifferentiation Item in L1 is absent in L2: do > 0 in negatives and
interrogatives) (sometimes called ‘absence’)
Level 3 Reinterpretation Item in L1 used differently in L2: He's a teacher > Es
maestro.
Level 4 Overdifferentiation Item not in L1 must be learned for L2: in Spanish, gender
of nouns (sometimes called ‘new category’)
Level 5 Split The opposite of coalescence, item in L1 becomes two in
L2: know > saber/conocer; to be > ser/estar; to ask for >
preguntar por/pedir (also called ‘differentiation’)

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

Summary

Unit 3 TERTIUM COMPARATIONIS & PROCEDURES OF CA

I. Comparability criterion and tertium comparationis (TC)

Tertium comparationis:
- a common platform of reference enabling the process of contrastive analysis
- “third term of a comparison”
C

A B
- remains invariant in translation or in CA which forms the basis for the comparison
II. Equivalence:
A contrastive relation referring to the relative sameness in meaning
E.g. Objects can be compared via different features -> similar in some respects but
different in others
– A square & a rectangle:
• Same number of angles;
• Different side lengths
– Box A & Box B: Volume (A > B); Weight (B < A)
Joseph Vendryes:
- under the variety, languages share common attributes -> Foundation for general linguistics
James (1980):
- Translation equivalence is the best TC for CA
- Translation equivalence = semantic equivalence + pragmatic equivalence (contextual
equivalence)
- Formal equivalence is incomplete for CA
TC at different levels of linguistics:
• Phonetics & phonology:
TC: The issues of Position/Manner articulation; Suprasegmental units; distinctive features can
be discussed in both English and Vietnamese
E.g. /p/ & /b/ in English vs. Vietnamese in terms of the aspects mentioned above.

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

• Lexis:
TC: The issues of mental images in the surrounding world can be discussed in both English
and Vietnamese
E.g. words naming colours in English vs. Vietnamese
• Grammar:
TC: Both English and Vietnamese have corresponding structures & meaning in some aspects
E.g. Existential sentence in English vs. Vietnamese
• Pragmatics:
TC: Both English and Vietnamese share some corresponding language functions
E.g. act of greeting in English vs. Vietnamese
III. Types of TC
2-texts [+/-trans]: data collected as corpus for CA
- 2-texts [+trans]: texts that are translatable
- 2-texts [-trans]: texts that are untranslatable
1) Statistical equivalence (for quantitative Contrastive Studies (CSs))
- Translational version of structures in L1 & L2 with a highest frequency
- Semantic/pragmatic equivalent with almost the same frequency
2) Translational equivalence:
- 2-texts [+trans]: data for qualitative
- 2-texts [-trans]: data for qualitative CS (Contrastive Studies)
3) System equivalence (for CS of systems):
- Equivalent established on paradigmatic + syntagmatic axis
- Examine members of system + their collocation
4) Semanto-syntactic equivalence (for CS of construction):
- On the similar basis of deep structure as semantic structure, as input for the
grammatical derivation
5) Rule equivalence (for CS of rules):
- Based on comparison of constructions on which these rules operate
- Interpreted in the view of Transformation-Generative Grammar: Phrase Structure
Rules, Transformational Rules, e.g. input & output of Wh-question vs. Vietnamese
equivalents
6) Equivalents in objects:

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

- Objects or entities outside language expressed by vocabulary in L1 & L2, e.g. foods,
festivals in English culture vs. Vietnamese
7) Pragmatic equivalents (for CS of pragmatics, stylistics or socio-linguistics):
- Relations between texts of two different languages which illicit from the language
user the maximally similar cognitive effects:
+ Functions of a unit, construction, structure
+ How these linguistic devices behave in speech acts in each speech community
- Formal equivalences are the least important
- Comparative devices of languages: significant only if they have a function that is
comparable to each other
IV. Procedures of CA
4 Steps in Contrasting Two Language Systems

Description Juxtaposition Comparison Prediction

1. Description:
• Selection & preliminary characterization of items under comparison
• Conducted within the same framework of language- independent theoretical model
• 2 approach for description of CA: bilateral/unilateral CA
+ Bilateral CA:
Describe L1 and L2 data independently
Use etalon language form which is model-neutral

Features of
completed
relevance
anterior
On foot
speaker

Current

Etalon lang.
female

arrival

Language CA

Sentence (E) + - + - + + -
I have arrived

Sentence (R) + + + + - + +
Ja prishla

Unfavorable points of bilateral CA:

- No need for the description of L1 & L2 to be equally exhaustive
- Too much work is done for comparison

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

- impossible without the balance in means or ways of expressing categories of the

linguistic units in L1 and L2
E.g. Intonation: [+] in English but [-] in Vietnamese
- A descriptive imbalance, in favour of the L2
- More concerned with what the learner does with the L2
The unilateral CA can be done with 2 phases:
- The first phase:
Establish the subsystem for CA in Language 1
E.g. possessive category in Vietnamese “tình yêu c a Lan”
- The second phase:
List out the language means in Language 2
E.g. Vietnamese English
The love of Lan
tình yêu c a Lan
Lan’s love

2. Juxtaposition
- decides what is to be compared with what, like with like
- identification of cross/inter-linguistic/cultural equivalent
- bilingual competence, enables one to make decisions about the equivalence of
element X & element Y in L1 & L2 respectively X &Y: comparable
E.g.

Language for CA Vietnamese English

Sentences to be Vì th mà bát cháo hành c a th N This onion soup offered by Thi No

juxtaposed làm h n suy ngh nhi u. made him think much.

- formal considerations alone do not suffice in establishing comparability or TC

E.g.

Language for CA Vietnamese English

Sentences to be Tôi thích th t ngu i. I like the meat cold.

juxtaposed

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

3. Comparison
- compare ‘types’ rather than ‘tokens, i.e. not strings of sounds/graphic substance but their
structures
E.g.
Pronoun + 1st Person + Sing – Auxiliary – Past, Participle

I have arrived.
Pronoun + 1st Person + Sing – Prefix+ Verb+ Perfective + Past + Feminine

Ya prishla

CA compares abstract elements rather than their concrete realizations

Three basic areas of comparisons:

CA of various equivalent CA of equivalent CA of equivalent rules

systems across languages constructions

( pronouns, articles, verbs, (interrogative, negative, (subject raising, adjective

and in phonology consonants, nominal phrase…); in placement, interrogative
vowels); subsystems (nasals, phonology (sound clusters, inversion, passivization), in
laterals) syllables, diphthongs, & phonology (assimilation,
distributions of sounds dissimilation, metathesis

Possible situations in each area of comparison:

(1) XLi = XLj
when item X in Li may be identical in some respects with an equivalent item in Lj.
• Similarities of the two phenomena compared more important
(2) XLi XLj
when item X in Li may be different in some respects with an equivalent item in Lj.
• Differences are said to be more important
(3) XLi + - ∅Lj
when item X may be present in Li but absent in Lj.
Xli has no equivalent in Ylj, e.g. Tone in Vietnamese
4. Prediction

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

From assumptions of differences of L1 & L2, hypotheses/predictions are made about learner’s
transfer of habit of mother tongue into the use of target language:
• Interference is created in certain deviant structures
• CA power: prediction of errors
- from influence of mother tongue
- the effects of target-language asymmetries;
- transfer of training;
- strategies of L2 learning; and
- communication strategies

Pedagogical relevance of predictive capacity: to predict a scale of incremental difficulty

Three possible interlingual rule relationships based on positive and negative transfer potential:
• L1 has a rule and L2 an equivalent one.
• L1 has a rule but L2 has no equivalent.
• L2 has a rule but L1 has no equivalent.
Three types of choice in the Hierarchy of Difficulty:
The contrastivists identify the types of choices that either language makes available, and
relating these choices

Hierarchy of Difficulty:
1. Optional choice:
Possible selection among phonemes, e.g. English can have /p/ or /b/ word initially
2. Obligatory choice:
The selection of conditioned allophones and the limitations in distribution of
phonemes:
- English word initial /p/ must be aspirated, e.g. pin [ ] pin
- distribution of /n/ or / /, e.g. / / is restricted to the final position of the syllable in
English, as compared with both initial and final position in Vietnamese, e.g. / /sing
(English); / nga ngang (Vietnamese)
3. Zero choice:
Existence of a certain sound in one language that has no counterpart in another
language, e.g. Vietnamese has no counterpart for / / in English.

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

Words & Expressions

1. Comparability criterion (n) Tiêu chí i sánh

i m xu t phát trong quá trình phân tích i chi u bao g m c vi c xác l p nh ng y u
t có th so sánh c trong các ngôn ng , vd: có th so sánh ph âm /p/ c a ti ng
Vi t v i ph âm /p/ c a ti ng Anh, c n ph i xác l p tiêu chí so sánh trên bình di n c u
t o âm (place of articulation); cách th c phát âm/ thoát h i (manner of articulation) và
thanh tính (voicing)
2. Contrast (n) i l p/t ng ph n
M t quan h ch m c gi ng nhau t ng i th p gi a các n v ng pháp c
phân tích c a 2 ngôn ng . M c này c quan sát theo các quan h h i nh p
(convergence) và phân ly (divergence)
3. Difference (n) D bi t
M t quan h ch tình hu ng không có m t ph m trù t ng ng ngôn ng B i v i
ph m trù c tìm th y ngôn ng A. Quan h này c g i là quan h zero, vd:
Language for CA Vietnamese English
Tone + Ø

4. Equivalence quan h t ng ng
M t quan h i chi u ch s gi ng nhau t ng i v ng ngh a
5. Similarity (n) T ng ng
M t quan h ch m t m c gi ng nhau t ng i cao gi a các n v ng pháp c
phân tích c a 2 ngôn ng .
6. Tertium comparationis (n) C s so sánh
M t n n t ng chung c a s qui chi u cho phép phân tích i chi u. Theo ngh a en,
ây là “y u t th 3 c a m t s so sánh”, và y u t này không thay !i trong khi d ch hay
trong phân tích i chi u, làm c s cho s so sánh.
Theo Wikipedia, ây là thu c tính/ph"m ch t c a 2 s v t c so sánh có i m
chung. ây là i m so sánh g i ý cho tác gi c a s so sánh khi so sánh m t ng #i hay
v t v i m t ng #i hay m t v t khác. Hai s v t c so sánh không nh t thi t ph i ng
nh t/gi ng y nhau. Tuy nhiên hai s v t này ph i có it nh t m t thu c tính hay ph"m ch t

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

chung. Ph"m ch t chung này c g i là c s so sánh hay thu c tính c so sánh

(tertium comparationis)
Theo phép "n d , tertium comparationis là c s hay i m chung cho phép so sánh,
vd:
• Necessity is the mother of invention. (English proverb)
• Tính c n thi t là m$ c a phát minh.
• it ng so sánh: quan h gi a m$ và con, quan h gi a tính c n thi t và phát minh
• C s so sánh (Tertium comparationis): ngu n, n i m t s v t nào ó phái sinh
• Woman is the nigger of the world. (John Lennon)
• Ph n là ng #i da màu c a th gi i.
• i t ng so sánh: s i x% c a v&n hóa M' v i ng #i da en, s i x% c a v&n hóa
toàn c u i v i ph n
• C s so sánh (Tertium comparationis):s i x phi nhân, s áp b c
7. Textual equivalence (n) T ng ng v n b n
Quan h t n t i gi a m t y u t c a v&n b n g c và m t y u t t ng ng b n d ch,
c ch p nh n b i m t ng #i có kh n&ng song ng
22. Translation equivalent (n) T ng ng i d ch
Di(n t ngôn ng ích (Target languge) dùng d ch m t di(n t ngôn ng
ngu n (Source language) trong m t s ng c nh nh t nh. Thu t ng này còn ch m c
mà các n v ngôn ng (vd: t), c u trúc cú pháp) có th c d ch sang m t ngôn ng
khác mà không th t thoát ý ngh a. Hai n v ngôn ng có cùng ng ngh a trong 2 ngôn
ng c cho là các i d ch hay t ng ng i d ch.
23. Corpus (n) s nhi u corpora Kh i ng li u
Kh i ng li u ( c thu th p nghiên c u phân tích, c th là phân tích i chi u)
24. 2-texts [+/-trans] kh i ng li u c u thành t 2 v n b n (có th /không th d ch)
Kh i ng li u c xây d ng t) 2 v&n b n dùng phân tích i chi u, có th d ch
ho c không th d ch
25. Quantitative Contrastive Studies (CSs)) Các nghiên c u i chi u nh l ng
Theo ngh a h$p, là b t k* nghiên c u có s% d ng qui trình thao tác thu th p d li u
d i d ng s . R ng h n, thu t ng này còn ch ph ng pháp v i m c ích gi i thích quan
h nhân qu c a hi n t ng c quan sát qua vi c xác nh các bi n s d c s% d ng làm
c s cho vi c i u tra th c nghi m
26. Qualitative Contrastive Studies (CS) Các nghiên c u i chi u nh tính
Theo ngh a h$p, là b t k* nghiên c u có s% d ng qui trình thao tác thu th p d li u
không ph i d i d ng s , nh trong các nghiên c u ph+ng v n, quan sát di(n ti n phát
tri n c a cùng m t i t ng, tham gia quan sát

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

27. paradigm1 (n) paradigmatic adj h hình

M t danh sách/t p h p hay bi u th c ch các d ng c a m t t) trong m t h th ng ng
pháp. Ví d , trong ti ng Anh:
singular plural
boy - boys
boy’s - boys’
(of the boy) (of the boys)
H hình c,ng có th c dùng ch các d ng khác nhau c a m t t). Ví d , trong ti ng
Pháp:
singular plural
je parle “I speak” nous parlons “we speak”
tu parles “you speak” vous parlez “you speak”
il parle “he speaks” ils parlent “they speak”
elle parle “she speaks” elles parlent “they speak”
M c dù h hình th #ng ch các bi n d ng c a m t t), ôi khi thu t ng này c,ng dùng
ch các t) phái sinh c t o ra t) m t t) g c (nh ví d d i ây m c syntagmatic
relations & paradignmatic relations)

28. Syntagmatic relations (n) Quan h cú o n (k t h p) paradigmatic relations (n)

quan h h hình/liên t ng/t v
Quan h cú o n là quan h theo ó các n v ngôn ng (vd, t), m nh (cú)) quan h
v i các n v khác b i chúng có th cùng xu t hi n trong m t chu-i. Ví d : m t t) có th
c cho là có quan h cú o n v i các t) khác xu t hi n trong cùng m t câu, nh ng
chúng l i có quan h h hình v i các t) có th thay th chúng trong cùng m t câu.
Ví d :

I gave Tracy the book

passed = syntagmatic relations

handed
= paradigmatic relations

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

Threw
paradigmatic axis Tr c h hình
syntagmatic axis Tr c cú o n

passed

I gave Tracy the book

syntagmatic axis
Paradigmatic axis

handed

threw

29. Generative grammar (n) Ng pháp T o sinh

H th ng ng pháp nh.m xác nh và mô t v i m t h th ng các qui t/c hay nguyên lí t t
c các câu ÚNG NG0 PHÁP c a m t ngôn ng và không mô t các câu sai ng pháp.
Ki u ng pháp t o sinh hay s n sinh các câu úng ng pháp. (Xem lý thuy t T o sinh)
30. Generative theory (n) Lý thuy t T o sinh
Thu t ng bao g m các li thuy t ngôn ng khác nhau v i m t m t m c ích chung (a)
cung c p s lí gi i các c i m hình th c c a ngôn ng , xác nh các qui t/c thuy t gi i
cách th c t o thành các các câu úng ng pháp c a m t ngôn ng và không mô t các câu
sai ng pháp (theo nguyên t/c phù h p miêu t Descriptive Adequacy), và (b) lí gi i t i
sao các ng pháp có các c i m v n có và cách tr1 con th /c chúng trong m t th#i k*
ng/n (theo nguyên t/c gi i thích phù h p Explanatory Adequacy)
31. Collocation (n) collocate (v) k t h p t /ng k t h p
Ph ng th c theo ó các t) c s% d ng k t h p th #ng xuyên. Các k t h p ch các ch
nh v cách th c các t) c s% d ng k t h p, ví d , các gi i t) nào c s% d ng v i
các ng t) c th , hay các ng t) và danh t) c s% d ng k t h p
Vd: trong ti ng Anh ng t) perform c s% d ng v i t) operation (cu c gi i ph"u),
nh ng không th c s% dung v i t) discussion (cu c th o lu n)
The doctor performed the operation.
* The committee performed a discussion.
Thay vào ó ta ph i nói:
The committee held/had a discussion.

Downloaded by Hoa Nguy?n ([email protected])

 lOMoARcPSD|11705684

perform c dùng v i (k t h p v i) operation, và hold / have k t h p v i discussion.

high k t h p v i probability, nh ng không k t h p v i chance:
a high probability but a good chance
do k t h p v i damage, duty, và wrong, nh ng không k t h p v i trouble, noise,
và excuse:
do a lot of damage do one’s duty do wrong
make trouble make a lot of noise make an excuse
Questions:
1. State 4 major steps of the procedures of CA.
2. Briefly describe the two phases of a unilateral CA.
3. Briefly describe the two phases of a bilateral CA.
4. Make a description of the lexical item “Gi ” and its equivalents in English, using a
unilateral CA.
5. How can we generate a hypothesis or make prediction about the learner’s difficulties in
learning and mastering a language construction or lexical item? For example, make prediction
about the Vietnamese learner’s difficulty in using “until” in English.
6. Design a diagnostic test to support a hypothesis about the Vietnamese learner’s difficulty in
using “until” in English.

Downloaded by Hoa Nguy?n ([email protected])