| 421.1.9.1: LANGUAGE UNIVERSALS AND TRANSLATION
STUDIES |
|
It must
be a common experience for most of us that though we may
know one or two more languages besides our mother tongue,
we do not have equal command over all the languages. Quite
often then we get stuck at expressing ourselves in one
language our knowledge of the language in which we have
better proficiency comes handy. We tend to extend the
words, idioms or structural patterns of such a language
to the other, often by means of translation. Once I happened
to buy a booklet in English consisting of quotations from
Indian classics. The title of the booklet in English was
'Sources of Indian wisdom'. The backside of the booklet
also had a title in Devanagari script  which
can be Romanised as srote bhaartiya kii buddhii. Now you
know that from any stretch of imagination the above title
cannot pass as normal Hindi. The translator has obviously
indulged in a word translation of English ignoring the
genius of Hindi grammar. In the Hindi phrase the head
noun sort or sroten should come after the genitive phrase
bhaartiya kii buddhii. But even if this order is rectified
to give the title as bhaartiiya kii buddhii sorten the
resulting expression is not a very good translation. We
will not go into the details of what could have been a
better way of saying 'sources of Indian wisdom'. The reason
this example was brought before you was to show that the
very fact that transferences like this take place when
we are dealing with more than one language, indicates
that human languages share a number of common linguistic
properties which we tend to exploit unconsciously as users
of one or more languages. As long as such transference
of linguistic features falls within the scope of the shared
elements things work. But languages, including closely
related ones, also differ from each other in many specific
details and if the transferred elements happen to fall
in such an area we get unwanted results like sorten bhaartiya
kii buddhii. |
|
| In translating
from a source language (SL) into the target language (TL),
a translator looks for equivalents between linguistic
elements of the two languages in question at the levels
of lexemes, idioms, phrases and clause structures. A translator
basically has to be concerned with both the semantic and
formal aspects of the SL and TL, and what formal options
or alternatives are available for carrying out this task
satisfactorily. As translators, thus, we are immediately
concerned with the vocabularies of the language in question
on the one hand and with their structural details on the
other. |
|
| In what
follows we will discuss the relevance of universals in
translation first at the level of vocabulary and then
at the grammatical level. |
|
| 421.1.9.2: VOCABULARY |
|
| While considering
the vocabularies of SL and TL, the first question that
would come up is are there certain universals principles
along with the vocabularies of natural languages are structured.
That is, are there intertranslatable items of vocabulary
shared by languages so that these can be compiled in the
form of dictionaries and lexicons to serve as helping
tools in translating. There are extreme views regarding
how the semantic universe is mapped in human languages
and how far, therefore, can languages be comparable in
the way their lexicon is structured. On the one hand,
we have the view that semantic mapping of the universe
is arbitrary and therefore languages can differ widely
from each other in the organization of their vocabulary.
But if this were really so translating back and forth
between languages in bilingual and multilingual speech
communities would have been a task next to impossible.
Not only, this, the whole enterprise of machine translation
and automatic translation would not have emerged as viable.
We should therefore, agree with Weinreich (1963:143) that
"despite the basically arbitrary quality of semantic "mapping"
displayed by languages, there are nevertheless remarkable
parallelisms between both related and unrelated languages".
If we take this stand, then the question would be to find
out what kind of generalization can be made about the
vocabulary structures of languages in common and how to
formulate them. |
|
| While considering
the structure of vocabulary Ullmann (1963) points out
certain basic issues with universal implications. The
first of these relates to the area of what he calls 'lexical
constant'. If we find that there are objects, events and
other notions of such fundamental importance that they
must find expression in all languages they would then
constitute the 'lexical constants'. Form the practical
point of view of translation, we would be interested in
finding out whether such a list of lexical constants could
be drawn. The items of this list need not necessarily
be expressed by a comparable set of formal elements in
the SL and TL. What would be expressed in one language
by a word may be shown by a bound form, compound or phrase
in the other. Dictionaries or lexicons complied with the
purpose of serving as a tool for translation between two
or more languages will then have to specify how the lexical
constant is manifested in the languages in question. |
|
| It should
be borne in the mind here that even if such constants
can be set up one should allow for differences between
languages. Ullmann points out, for instance, the idea
of 'father-hood' can be taken as a lexical constant in
the sense that every language ought to express such a
concept. But in Latin there are two words for father 'genitor'
for the physiological relationship and 'pater' which carries
the social connotations. What it means is that various
aspects of a lexical constant may be expressed by separate
words in some languages. A good dictionary should take
note of such subtle details in order to serve as a good
helping tool in translation. |
|
| Another
area of interest with universal implication pointed out
by Ullmann is 'lexical fields'. The systems of colours,
kinship relations, ethical and aesthetic values or religious
and mystical experiences etc. can be taken as examples
of 'lexical fields'. Within each one of these fields one
may want to see what are unique or idiosyncratic to a
language and what elements are common so that they ought
to be expressed in all languages. For instance, in the
field of kinship relation, the general notion of siblings
i.e. 'children of the same parent or parents' is expressible
in all languages. But the manner in which this part of
the semantic field is structured in individual languages
is different. For instance, in Malay there is only one
term saudara which means not only brother and sister both
but also younger and elder brother as well as sister.
The term can also be used for cousins. English on the
other hand has two terms to express the concept of male
and female siblings, namely 'brother' and 'sister'. Tamil
has a different system. It has four terms, aNNaa 'elder
brother', akkaa 'elder sister', and tampii 'younger brother',
tankai 'younger sister'. Hindi is more like English in
providing separate words for 'brother' and 'sister'. Now
such a trivial looking difference may in fact pose a problem
in translation. For instance, take the Hindi examples.
In this language it is customary in public speech to address
the audience as bhaaiiyo aur bahino 'brothers and sisters'.
How should this expression be translated into Tamil? The
native Tamil words will not suffice as is obvious from
the lack of a common term for the noun 'brother' and a
common term for the notion 'sister'. The best alternative
will be to take help form Sanskrit equivalents for the
two terms which are nativized in Tamil as sahodara sahodarikaLe
'Brothers and Sisters" (vocative). |
|
| The search
for a universal set of vocabularies for scientific communications
had been a concern among western scholarship even much
before machine translation and computer assisted translation
emerged as a possibility. As Hutchins (1986) points out
as early as 1669 Johann Becker came out with a book consisting
of 10,000 Latin words for which equivalent entries in
othe rlanguages like German, French and Arabic were to
be found. However Becker had no idea as to how to deal
with the problems of syntactic differences that may be
involved in the equivalent words. According to Hutchins
a different approach to arrive at universal set of words
attempted by John Wilkins in 1668. Wilkins did not start
with the lexical items of a language for which equivalents
need to be found out in other languages. Instead he tried
to systematically enumerate and describe those objects
and concepts which ought to be expressed in all languages,
Whether by means of words or other markings. In other
words, he was attempting at, a universal classification
of concepts and objects, termed in later years as 'lexical
constants' by Ullmann. These attempts thus can be taken
as a precursor to the real search for a universal base
with a view to establishing equivalents in different languages
for machine translation purposes. |
|
|
