<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="Author" content="Giovanni Mandorino">
<meta name="GENERATOR" content="Mozilla/4.6 [en] (WinNT; I) [Netscape]">
<title>S.Giannini - S.Scaglione - Abstract</title>
</head>
<body text="#000000" bgcolor="#FFFFFF" link="#660000" vlink="#CC66CC" alink="#FF0000" background="Artgalry.exe">
<center><b><font size=+1>On defining the notion of areal coefficient: the
diagnostic value of quantitative criteria</font></b>
<br><b><font size=+1></font></b> <b><font size=+1></font></b>
<p>Stefania Giannini & Stefania Scaglione University for Foreigners,
Perugia</center>
<p>Abstract
<p>Research on classifying and synchronically comparing languages geographically
delimitable has focused on two basic diagnostic criteria, crucially for
improving our general knowledge of areal linguistic phenomena.
<br>Areal linguistics (AL) has established such crieeria as the followings:
<br>a. a quantitative criterion: several linguistic traits shared by various
languages genetically unrelated in a geographically contiguous area do
represent a first indicator of a probable linguistic area (LA) (Sherzer
1973, Campbell, Kaufmann & Smith-Stark 1986);
<br>b. a contrastive criterion: traits which are widely distributed in
a linguistic area and rarely attested (or absolutely absent) elsewhere
do have much more diagnostic value for delimiting a LA; in other words,
the higher the markdness of a trait, the greater its areal diagnostic value.
<br>In recent years developments in the study of genetic and/or geographical
relatedness across linguistic boundaries reached a good level of theoretical
definiteness (see Nichols 1993, devoted to identify stable structural features
in a sample of families and in a sample of areas; and at last van der Auwera
1997, who defined isophlets the abstract lines delimiting LAs with the
same number of traits).
<br>Yet two basic questions are at issue:
<br>1. Evaluate and measure to what extent established areal linguistic
traits can or must expected to be typologically consistent. On this field,
LA is related to typological linguistics (the more promising directions
for such areas of study, both separately and in their mutual interchange,
are clearly illustrated by Ramat 1998);
<br>2. Estimate and measure the different status of each linguistic trait,
in terms of the diagnostic value it has in delimiting LA where it spread.
<br>Concerning question 2., this study is designed to identify the diagnostic
values of each areal linguistic trait, in a qualitative perspective by
means of estimating its areal coefficient (AC).
<br>We will use the term areal coefficient referring to a numerical value
of lower/equal height to 2. This kind of count will be employed for all
convergent or areal diffused features, in order to assess their specific
diagnostic weight: marked and uniformly shared traits will get the top
score; on the other side, the down score will be basically assigned to
unmarked and unevenly diffused traits.
<br>To determine AC we used three distinct sets of relevant parameters:
<p>a. absolute structural parameters:
<br>a1. Markedness
<br>a2. Grammatical integration
<br>a3. Kind of borrowing
<p>b. relative structural parameters:
<br>b1. Typological consistency
<br>b2. Hierarchic implication
<p>c. areal diffusion parameters:
<br>c1. Density function of areal distribution (number of points of discontinuity
for each trait within LA)
<br>c2. Patterns of areal distribution (ways and shape of diffusion for
each trait within LA)
<br>c3. Presence/absence of each trai t in neighbouring languages outside
the LA
<p>Count type a. (a1.-a3.) takes into consideration the general relevance
of areal traits, in terms of their higher or lower markedness and availability
to be borrowed. Linguistic features will be of different weights, depending
on how marked they are, how integrated into the respective grammars and
how easily they can be adopted in the component of a language.
<br>Elaboration criteria type b. (b1.-b2.) can give an estimate of the
degree of typological conformity with a given language in the area: the
more consistent a trait, the less likely to have been transferred by means
of language contact within LA.
<br>Thirdly (criteria c1.-c3.) we will describe and evaluate the geographical
distribution of each trait within LA. Since all of the indexes calculated
in point c. are supposed to weaken the degree of areal convergency for
each trait, the final score of c. will be subtracted from the general amount
of two previous sets of parameters (a.-b.).
<br>In this paper the results of our study will be fruitfully applied to
some of the linguistic phenomena identified as characterizing and delimiting
features of the Mediterranean linguistic area.
<br>The goal of this study is to find linguistic traits to be rightly used
as indicators of areal convergence and diffusion. Areal coefficient may
represent a block of a theory of areal diffusion. It was created by a classification
and measures of structural and geographical parameters, which correspond
respectively to linguistic and historical circumstances that gave rise
and contributed to the development of LAs.
<p>References
<p> Campbell, L., Kaufmann T. & T.C. Smith-Stark. 1986. Meso-America
as a Linguistic Area, Language 62/3: 530-570.
<p>Nichols, J. 1993.Diachronically stable structural features, in Historical
Linguistics 1993, Selected Papers from the 11th International Conference
on Historical Linguistics, Los Angeles, 16-20 August 1993, edited by H.Andersen,
Amsterdam/Philadelphia, Benjamins, pp. 337-355.
<p>Ramat, P. 1998. Typological comparison and linguistic areas: some introductory
remarks, Linguistic Sciences 20/3: 227-240.
<p>Sherzer, J. 1973. Areal linguistics in North America, in Current Trends
in Linguistics, ed. by T. A. Sebeok, vol 10, Linguistics in North America,
Mouton de Gruyter, pp.749-795.
<br>
</body>
</html>