| 421.1.3.1: TWO MAJOR APPROACHES TO LANGUAGE
UNIVERSALS |
|
| As was mentioned
earlier, most modern studies on language universals, especially
pertaining to syntax, fall broadly within the frameworks
adopted by Joseph Greenberg and Noam Chomsky. It is necessary
for us, therefore, to understand the basic points of these
two approaches to language universals. Let us call the
Greenbergian approach the 'typological approach' and Chomsky's
approach the 'generative approach'. We will take them
up in that order. |
|
| 421.1.3.1.1: TYPOLOGICAL APPROACH |
|
| The reason
Greenberg's approach is called typological is that linguists
working in the Greenbergian framework see a close connection
between 'typological' study of language and setting up
of universals of linguistic features. It may be worth
elaborating here what is meant by typological study. Languages
are generally classified either genetically or typologically.
In genetic classification of languages two or more languages
are grouped together if it can be proved that they have
family affiliation. That is, if they can be viewed as
either two daughters deriving from the same common parent
language or if it could be shown that they represent different
stages of one and the same language. It is in this sense
that we talk of Sanskrit as the mother of modern IndoAryan
languages like Hindi, Bengali, Punjabi, Oriya etc. in
typological classification, on the other hand, languages
are grouped or classified not by historical consideration,
but on the basis of whether or not they share a given
linguistic trait. For example, if we take the relative
order of occurrence of subject, verb and object in a simple
sentence as a criterion for classification, we will group
together Tamil, Hindi, Telugu and Bengali as belonging
to the type SOV. As different from this, Kashmiri will
be assigned with English to the SVO type of languages.
It should be obvious from this illustration that typological
classification offers a different grouping of languages
from genetic classification. Genetically Hindi, Bengali
and Kashmiri belong to the Indo-Aryan family of languages
as distinct from Tamil and Telugu, which are languages
of the Dravidian family. But we have noticed that in typological
classification of languages it is legitimate to group
languages together even though they may not be genetically
related. Further languages that can be grouped together
on typological basis need not necessarily be in geographical
proximity to each other. |
|
| Coming to
the connection between typological study of language and
linguistic universals, the very assumption that we can
compare language in terms of their structural properties
means that there exist some universal features of language
on the basis of which we can compare languages and classify
them into certain types. As Comrie (1982:) points out
it is for this reason that "the study of linguistic typology
has in general gone hand in hand with the study of language
universals". |
|
| Greenberg's
basic ideas on language universals were put forth by him
in his paper "some universals of Grammar with particular
reference to the order of Meaningful element" (1963).
As the title of his paper indicates, Greenberg approached
the question of language universals by considering the
order of words, more appropriately, order of constituents
in the sentence i.e., the relative order of subject, object
and verb, besides the order of constituents in the noun
phrase and verb phrase for the formulation of his universals.
As a result works on universals carried out in his framework
are also called 'word order typologies' of 'word order
universals'. |
|
| Greenberg
examined a sample of 30 languages belonging to different
families with a wide geographical coverage. As has been
pointed out one feature that he used to compare the languages
of his sample is the relative order of the constituents
subject, object and verb in a sentence as the most dominant
pattern for the language in question. He noticed that
though majority of languages in his sample showed alternative
orders of these constituents, they had one order as the
dominant one. A significant observation that he made was
that although the relative order of the three constituents
S, O and V should yield logically a total of six possibilities,
all the possible types were not attested in the language
examined by him. Only three of these possible orders occurred
more commonly and they were SOV, SVO and VSO. The three
order types which did not occur or occurred only rarely
were VOS, OSV and OVS. Note that in all of these non-occurring
types we have the Direct Object coming before the subject.
This suggested to Greenberg that languages generally do
not tolerate sentences in which the object precedes the
subject unless such an order is exploited by the language
for specific purposes such as for emphasizing or focusing
an element. Consideration of this type led Greenberg to
propose his first universal as follows, "In declarative
sentence with nominal subject and object, the dominant
order is almost always one in which subject precedes the
object". |
|
| As was mentioned
earlier, Greenberg considered other order types of grammatical
elements besides the order of constituents in a sentence.
For instance languages may differ in the relative order
of adjective (A) and the head Noun (A) that it modifies,
the two possible orders are thus NA or AN. Indian languages
like Tamil, Telugu, Hindi etc. generally use the order
AN. But there are languages like Hebrew, Arabic, Zapotec
etc., which systematically place the modifying adjective
after the head Noun thus following the NA pattern. Similarly
languages may differ in the order in which they express
the possessive of genetive element (G) and the head Noun
in the Noun phrase, of the two possible orders GN and
NG, Dravidian languages like Tamil, Telugy and Malayalam
etc. use consistently the GN order, for instance, Te.
nii peeru 'your name' , mana illu 'our house'. As different
from this, Persian puts the genitive element consistently
after the head Noun. Hindi basically goes with other Indian
languages in this respect in having the order GN as in
meraa ghar 'my house', sudhiir kii patnii 'Sudhir's wife'.
But Hindi also allows the reverse order e.g. kissaa kursii
kaa 'story of the chair', or bhaaii uskaa bahut biimaar
hai 'His brother is very ill' literally 'brother his is
very ill'. |
|
| Greenberg
considered many more possible orders of grammatical elements
in proposing his 45 universals. But besides considering
word order as a parameter he also considered other factors
one such parameter that he used was whether a language
has prepositions or postpositions. We have already noted
elsewhere that most Indian languages employ postpositions
where English used prepositions. Now let us consider the
four parameters mentioned so far, namely (i) the relative
position of S, O and V in clause (ii) the relative order
of A and N, (iii) the relative order of G and N and (iv)
whether a language is postpositional (PO) or prepositional
(Pre). A priori there is no way of knowing what kind of
relationship holds between these different features Therefore,
given these four parameters there is a possibility of
a wide range of variation among languages since individual
languages may choose different permutations and combinations
of these four features in question. But the interesting
and insightful observation made by Greenberg is that there
seems to be a high degree of correlation or connection
among these parameters in the sense that selection of
one of these puts restrictions on the way a language can
choose from alternatives available on other parameters. |
|
|
