Terminological Representation of Specialized Areas in Conceptual Structures: The case of Gender Stud

1
Terminological Representation of Specialized
Areas in Conceptual Structures The case of
Gender Studies

• María J. López-Huertas. Universidad de Granada
  (Spain)
• Mario Barité. Universidad de la República
  (Uruguay)
• Isabel de Torres. Universidad de Granada (Spain)

2
INTRODUCTION

• Knowledge is in an unprecedented stage of
  diversification, specialization and disciplinary
  intersection.
• This rise of interdisciplinarity or hybridation
  (Klein, 1996 Dogan, 1996) presents serious
  problems for the representation and retrieval of
  information.

3
INTRODUCTION

• Ordinary documentary languages follow the classic
  paradigm of the sciences, dividing knowledge into
  disciplines (Beghtol, 1998 Landridge, 1995
  Iyer, 1995).
• As a result, the structures and the terminology
  of interdisciplinary areas are not visible
  enough, or situated properly in these languages
  (Williamson, 1998 Vaver, 2002).

4
INTRODUCTION

• Gender Studies are an emerging interdisciplinary
  domain (Caidi, 2001)
• A) their literature is growing and diversifying
  at great speed (Marcella, 2002)
• B) its subject area includes contributions from
  different disciplines and subject areas, whose
  impact and relative participation have not been
  studied in depth and
• C) it is information with important social and
  multicultural repercussions (López-Huertas
  Barité, 2002).

5
INTRODUCTION

• Recent research has detected two potential
  problems with respect to KO in interdisciplinary
  areas
• 1) its thematic nuclei and its epistemological
  borders are diffuse and not agreed upon by all
  specialists and
• 2) many disciplines participate in the
  construction of these spaces and generate an
  overflow of terminology, and a unbalanced
  conceptual structure.

6
OBJECTIVES

• The analysis of the domain and of the
  terminological behaviour within the area of
  GENDER STUDIES
• to provide guidelines based on objective and
  consistent criteria, the proper selection,
  validation and normalization of terminology
• and striving for consensus about the conceptual
  representation of the domain in thesauri.

7
MATERIAL AND METHODS

• Four specialized thesauri in Gender Studies from
  different geographic areas
• The European Women Thesaurus (IIVA, 1998),
• Tesauro Mujer from Spain (Instituto de la Mujer,
  2002),
• Thesaurus dHistòria Social de la Dona from
  Catalonia (Sebastiá i Salat, 1988),
• Tesauro para Estudos de Género e sobre Mulheres
  from Brazil (Bruschini, Ardaillon Unbehaum,
  1998).

8
MATERIAL AND METHODS

• A twofold methodology was applied
• a study of the terminology used
• a structural analysis of the domain.

9
METHODOLOGY. Study of terms

• 1) Thorough compilation of terms (T), descriptors
  (D), and non-descriptors (ND) of the thesauri
  (with the exception of the Brazilian thesaurus,
  as the digital version could not obtained)
• 2) A quantitative comparison and evaluation of
  the number of T, D and ND given to the domain
  and

10
METHODOLOGY. Study of terms

• 3) A determination of the degree of overlapping
  among the three thesauri altogether, considering
  the limiting factor
• Limiting factor or LF (overlapping terms divided
  by maximum number of possible overlaps).
• In this case, the LF coincides with the number of
  T, D and ND of Catalonian thesaurus (714, 624 and
  90, respectively) as it is the one with the least
  terminology.

11
METHODOLOGY. Domain Analysis

• The structural analysis of the domain involved
• 1) identification of the classes or main topic
  areas of the four thesauri, understood as only
  those established in the first level of the
  subdivision
• 2) comparison of these classes or areas, to
  determine their overlapping. 1 co-occurrence (in
  two thesauri) 1 hit
• 3) Identification of the subject areas with a
  greater number of hits.

12
RESULTS. Terminological behaviour
Table 1. Contributions of terms among thesauri
coefficient of equivalence
13
RESULTS. Terminological behaviour

• Table 2. Overlap among the three thesauri

14
RESULTS. Terminological behaviour
Table 3 expresses two values simple
co-occurrence (overlapping terms multiplied by
three, divided by total co-occurrences) and
co-occurrence with a limiting factor or LF
(overlapping terms divided by maximum number of
possible overlaps).

• Table 3. Percentages of co-occurrence in the
  thesauri considering the LF

15
RESULTS. Structural analysis of the domain

• A) Identification of the classes or main subject
  areas of each thesaurus
• 34 different subject areas used by the 4 thesauri
  in the first level of subdivision, revealing
  extremely high diversification.
• Similarly extreme is the diversity of the domain
  models. In general, the subject areas correspond
  to disciplines, but also to broad areas (e.g.
  Health).

16
RESULTS. Structural analysis of the domain

• B) Identification of subject areas with the
  greatest impact
• None of the 34 subject areas obtained 3 hits
  (none of them are included in the four thesauri).
• Eight areas (23.5) co-occurred in this first
  level in three thesauri (2 hits).
• Eight areas (23.5) co-occurred in this first
  level in two thesauri (1 hits).
• Eighteen areas (53) are only in one thesaurus (0
  hits).

17
TABLA 3. Áreas temáticas según los hits de
co-ocurrencia en los tesauros
18
RESULTS. Structural analysis of the domain

• The previous table shows the following
  interesting data
• Areas that are not associated with Gender Studies
  in a nuclear sense (Phylosophy and
  Communication/Media) are included (2hits)
• Other areas that are indeed nuclear in the
  literature and expert opinion are almost
  invisible (Womens Studies, Sexuality, Family)
  because they are located under other main areas.
  Maybe reflecting the cultural patron of
  invisibility of women.

19
CONCLUSIONS. Terminological behaviour

• There is a lack of uniformity in the terminology
  provided by the thesauri. There is scarce
  consensus regarding the terminology belonging to
  the domain.
• Common epistemological criteria seems to be
  lacking.

20
CONCLUSIONS. Terminological behaviour

• There is a need for objective methods of
  terminological extraction from the literature.
• The percentage of nuclear terms (D and ND) in the
  set of overlapping terms is remarkably low. If
  this is confirmed in future studies for the
  vocabulary set as a whole, we could conclude that
  a significant percentage of the terms proposed by
  the thesauri is not relevant for indexing
  specifically about Gender Studies.

21
CONCLUSIONS. Domain structure

• There is a severe conceptual dispersion of the
  models built to represent the domain.
• There is a lack of consensus as to the scopes of
  the domain and their structural representation in
  the consulted sources.
• There is a need for a KO model for such
  interdisciplinary domains.

22
SUGGESTIONS

• Organizing these conceptual structures and their
  terminology applying methodologies founded upon
  objective criteria
• Taking into account the revision of predecessors
  in the subject matter, the rational application
  of the principle of literary warrant
• Improving the process of identification,
  selection and formalization of terminology and
  its structural expression.