Filetype PDF | Posted on 21 Sep 2022 | 3 years ago
Authentication
digital resources
133x Filetype PDF File size 0.31 MB Source: celta.paris-sorbonne.fr
Wlodarczyk André (1998) “The Proper Treatment of the Japanese ‘wa’ and ‘ga’ Particles” , in
Proceedings of the International Workshop on Human Interface Technology 1998 (IWHIT '98) - Aizu-
Wakamatsu, Japan, pp. 63-70.
THE PROPER TREATMENT
OF THE JAPANESE wa AND ga PARTICLES
André WLODARCZYK
Centre for Analysis and Social Mathematics - LaLIC Group
CNRS - EHESS -University Paris-Sorbonne (Paris IV)
96, Boulevard Raspail, 75006 PARIS
wlod@ext.jussieu.fr
0. Abstract 1. Introduction
For modern Linguists, wa and ga particles The Modern Japanese particles wa and ga
constitute one of the most interesting and arduous cannot be explained properly without taking
problems of Japanese Grammar. There are two kinds into account the other particles that belong to
of logical relations that wa particle can mark in the same classes of morphemes that ga and wa
different sentence positions. These relations are of represent; i.e., case particles (kaku-joshi) and
set-theoretical and predicative types. However, we do «concordance» particles (kakari-joshi). The
not take for granted that logical relations have their
equivalents or one-to-one mappings in languages. We interpretation according to which wa is a marker
only intend to suggest that, in order to explain both of the topic and ga is a marker of the subject is
series of particles, we need to recognise their not satisfactory because it makes use of a
functioning as markers of some logical relations. deletion rule concerning ga particle when the
On the other hand, our approach accepts that wa subject is to be topicalised by wa. Furthermore,
and ga particles present historically motivated such interpretation is not proper for explaining
ambiguities and that these ambiguities can be multiple meanings of the Japanese wa and ga
explained as the result of a boomerang relation particles.
between the mentioned classes of morphemes. This In order to make explicit the differences
kind of opposition is unknown in Structural
Linguistics, nevertheless there is much evidence on between wa and ga particles, we should keep in
the material of Japanese grammar that such mind the following facts:
opposition (defined possibly as a double privative 1. European grammars are based upon a
relation) should be added to the realm of oppositions
that proved to be so useful in language studies. predicative sentence structure (with
obligatory subject).
The ambiguity problem of particles wa and
ga leads to another problem that is also related 2. The Japanese sentence is, in this respect,
to the logical concept pair of Universal and somehow different from that of European
Existential quantifiers. In order to explain the languages because its Subject constituent is
intuition (that many Japanese linguists have optional and the Predicate part (Verb or
had) of the relation between the logical Adjective) alone is sufficient to make a
quantification and the particles wa and ga, we sentence.
need to reinterpret this relation as New/Old
information. Viewed as such, the determination Indeed, since in Japanese there is no
part of Logical Quantification appears as a morphological agreement between Subject and
discrete simplification of the continuum of Predicate, the subject is not obligatory and the
values contained between two poles (Generic verb (or adjective) is the only obligatory
and Specific) with a common (null) point of constituent of the sentence. On the other hand,
both these antinomies. the particle ga can refer to more than one
syntactic functions (such as the Object or the Natural Language expressions are produced
Location) and the particle wa, when attached to by mechanisms using two kinds of fundamental
a subject constituent, is not always a operations: selection and combination thus
topicalisation marker. As a matter of fact, we endowing the language with two-dimensional
can observe the same opposition between wa character3 (Roman Jakobson). The other
and ga when they follow subject or object evidence of the same fact is related to the
phrases and when they affect other kinds of psychological research where two distinct
phrases. processes are found to lay at the base of human
understanding4 of NL expressions (in this case
2. Boomerang opposition between «wa» novel noun phrases): comparison and scenario
and «ga» particles creation (Edward J. Wisniewski, 1997).
As an example, let us consider two kinds of
In Functional Linguistics, three following identity :
types of systemic relations (called oppositions)
are distinguished in the synchrony (a given state 1) a is b. p(a), (predicative identity of a with
of a language): respect to b), where p = "is b",
- privative oppositions: +/- or marked/unmarked 2) a belongs to A. (a ∈ A) ; i.e.: set-theoretical
- equipollent oppositions: C1 versus C2 identity of a.
If we want to formulate both at the same
- gradual oppositions: (C1 ( C2 ( ....( Cn)))) time, we must consider that there are two
However, during the transitory (evolutive, different orders in sentences : actual (explicit)
dynamic1) phases of linguistic systems, we can and virtual (implicit). These orders are called, in
encounter important changes that are possible classical structural linguistics (F. de Saussure),
only because categories enter into new syntagmatic and paradigmatic axis.
oppositive relations between them. An attempt to formulate the two orders at
Two traditional points of view: once could look as follows :
1- equipollent opposition between both p(a ∈ A; b ∈ B) where p = the copula "to be"
particles wa/ga is a predicative relator (is_a).
2 - concatenation ga and wa with deletion of The meaning of such a formula would be
ga something like this: «a taken as an element of
the set A is to be related to b taken as an
Our point of view: element of the set B». However, such parallel
3 - bilateral (boomerang) relation between treatment of relations should be distinguished
wa and ga particles from what is known as Restricted Quantification
since, in this approach, we are not concerned
As we consider that both wa and ga particles with quantification problems.
are now undergoing mutational processes, we
introduced the boomerang relation defined as a On a somehow higher level of abstraction, we
special kind of bilateral or double privative must also recognize that the result of a
opposition. This opposition uses the characteristic function on a universe is
markedness/ unmarkedness2 of linguistic forms equivalent to that of the abstraction operator
with regard to categorial meanings and is, in our (or lambda operator) when applied to the same
view, characteristic of the transitory situations. universe.
Thus, we claim that introducing the concepts of {char. Function} ≡ {abstraction Operator}
opposition and markedness we can capture more a → F(x) ≡ {x: F(x)} (a)
properly than before the ambiguous (since
complex) character of the Japanese wa and ga. Subject and Topic are special cases of each of
the above identities: predicative identity of an
3. Logical foundations of the Japanese argument for Subject and set-theoretical identity
of an element for Topic. Let us mention also
kernel sentences
1) Cf. Jakobson R., 1963 3 ) Cf. Jakobson R. (1959)
2 ) Givon T., 1995. 4 ) Edward J. Wisniewski (1997)
that for semantic and pragmatic reasons Subject that the meaning of a sentence contains a
and Topic are often associated in speech semantic agent. Therefore, using AVL-like
processes. logic expressions, it is possible to establish the
following correspondences:
4. Word order (1) can be
interpreted as
Word order in Japanese can be described by
three different positions of the Subject and/or (2) can be
Topic : distant, intermediate or close. From the interpreted as
dynamic point of view, the opposition between Roughly (because Natural Languages are in
morphemes wa and ga appeared in the no way ideal objects), the first correspondence
intermediate position : 0 <1 wa< ... (1) is characteristic of the Japanese language
...>1 ga > 0 (see figure 1) (Taro: ga aruite kita) and the second one (2) is
Distant Intermediary Close Referen proper to English (Taro came on foot). As the
position position position ce point matter of fact, in different languages, we can
find both morpho- or ordino-syntactic elements.
Topic Topic wa Subject Verb In the structure of the Japanese language
< wa < Subject ga > ga > some grammatical morphemes refer to
Fig. 1: Ascending and descending orders contiguity (such is the function of case
particles) whereas others refer to similarity
The intermediate position is therefore the (morphemes called traditionally in Japanese
common distribution of both Topic and Subject kakari-joshi, i.e.; concordance particles).
and it constitutes the reason why the Therefore, the grammatical theory of the
«boomerang opposition» between the particles Japanese language should take into account not
wa and ga could take place. only syntagmatic (actual, in praesentia)
relations but also paradigmatic (virtual, in
absentia) relations. In this paper, we argue that,
5. The Japanese case particles and in Japanese, at least in a simple predicative
«concordance» particles revisited sentences, Topic seems to be the result of the
Although we do not pretend to present here mapping of a paradigmatic relation into the
an alternative to deep case structure5, in order to syntagmatic organisation.
understand better what is the nature of case Thus, in our approach, we accept that wa and
relations, let us compare some characteristic ga particles present historically motivated
features of Attribute Value Logic (AVL) and ambiguities and that these ambiguities can be
First Order predicate Logic (FOL). The explained as the result of a boomerang relation
advantage of AVL with respect to FOL consists of the mentioned classes of morphemes.
in making symbol manipulation more flexible.
Namely, owing to AVL not only the problem of 6. Element Particles (wa, mo, koso, sae)
order of arguments and that of the predicate
arity disappeared but also the representation of Generally speaking, the element particles are
structured information became less restricting. markers of absolute and relative identity in the
Let us recall for instance that the following set-theoretical sense. The figure 2 shows how
predicate p(x,y,z) has 3 arguments which must some of these particles can be classified
appear in that order, i.e.: x, y and z. If we assign according to the criterion of belonging of a
ordinal attributes to these arguments such as chosen element to a virtual set.
pos1: x, pos2: y, pos3: z, then obviously we free
the order of arguments because the latter is absolute identity relative identity
explicitly stated.Thus, we can extrapolate that a wa b mo
A A
ideally the kernel syntactic relations can be simple a
belonging b a
linearized either in a morpho-syntactic or in an indication comparison
ordino-syntactic manner. For example, suppose a koso b sae
B B A
complex A
belonging a b a
insistence concession
5 ) On this particular point, cf. WARD Nigel
(1992).
Fig. 2: Particles marking identity of an between Japanese argument particles and
element belonging to a set or to a subset element particles.
1) wa is the marker of belonging of the Table 1: Case and ‘concordance’ particles in
element a to the set A; i.e.: aRa, (a ∈ A) Modern Standard Japanese (cp = case particle)
2) mo is the marker of belonging of the Case name cp cp+wa cp+mo
element b to the set A, this belonging being nominative ga *) *)
established with respect to a which is another accusative o o-ba**) o mo
element belonging to the same set A; i.e.: bRa,
(b ∈ A) & (a ∈ A) dative ni ni wa ni mo
allative e e wa e mo
3) koso is the marker of belonging of the instrumental de de wa de mo
element a to the set A, the latter set A being a comitative to to wa to mo
subset of B; i.e.: aRa, (a ∈ A) & (A ⊆ B) ablative kara kara wa kara mo
4) sae is the marker of belonging of the terminative mad made wa made mo
element b to the set B, the set B being a subset e
of A and the identity of b is established by
contrast with a; i.e.: bRa, (b ∈ B) & (a ∈ A) & *) Neither ga+wa nor ga+mo are possible.
**) o-ba is no more used.
(B ⊆ A)
It is very important for our purpose to notice
NB: In addition, wa and koso particles are that neither *ga+wa nor *ga+mo are
markers of reflexive relations. grammatically correct6). Although as such wa
We have used here what has been called and ga particles belong to different classes (wa
«attributive relations» (cf. [Desclés J.-P., is representative for set-theoretical relations -
1987]); i.e.: (a) as a relation of belonging and Element’s identity or image - on the one hand
(b) as the subset relationship. As the matter of and ga is representative for predicative relations
fact, the attributive relations can be seen as the - Argument - on the other), but their usage is
result of a kind of predicative projection such that they interact in a way that the system
between elements of a virtual (paradigmatic) which combines argument particles with
axis such as a Subject and those of the actual element particles seems to be changing in
(syntagmatic) axis such as the attribute part of contemporary Japanese. If we represent
the Predicate. Consider the sentence where predicative relations by the horizontal axis, ga1
element particles may occur alternatively: as a nominative (though sometimes ambiguous)
case marker goes not only together with o as an
Tori wa naku. (wa has no equivalent in accusative case marker, but also with the
English) /Birds sing./- «Birds belong to the set particles wa and mo . Analogously, if we
of beings that can execute the action of singing» 2 2
represent set-theoretical relations by the vertical
Tori mo naku. (mo corresponds to «also», axis, wa1 as an exocentrically oriented
«even» etc.) /Birds also sing./ - «Birds belong to (reflexive) identity marker goes not only
the set of beings that can execute the action of together with ga2 as an endocentrically oriented
singing, this belonging is established with (reflexive) identity marker, but also with the
respect to other beings that have the same particles o2 and mo1. Thus, we observe that the
property» «boomerang opposition» described above is the
Tori koso naku. (koso corresponds to result of interactions between «cases» and
«exactly,» «just» etc.) /It is birds that sing./ «elements» in the dynamic synchrony of the
«Birds belong to the set of beings that can standard Japanese syntax.
execute the action of singing, this set is seen as
a subset of another set.»
Tori sae naku. (sae corresponds to «even»,
«also» etc.) /Even birds sing./ «Birds belong to
the set of beings that can execute the action of
singing, this set is a superset of another set and 6) The dialects of Kyushu where the
the identity of birds is established by contrast combination of ga and wa is possible cannot be
with that of beings belonging to the subset. taken as a proof of the contrary because dialects
Table 1 shows the syntagmatic relations develop their own structures
no reviews yet
Please Login to review.
