Nuopponen, Anita (1994). Begreppssystem
för terminologisk analysis (Concept systems for terminological
analysis). Acta Wasaensia No 38, 266 p.
CONCEPT SYSTEMS FOR TERMINOLOGICAL ANALYSIS -
Summary
Introduction
The main purpose of this study is to devise a
framework for the classification of concept systems and concept
relations and in this way to develop the current basic conceptual
apparatus for terminological analysis.
Concept systems constitute in many respects a
particularly interesting object of research. They are of interest
not only to terminology science but also to philosophy, the general
theory of science, psychology, artificial intelligence, lexicology,
semantics, pedagogy, information science, classification theory,
database development, etc. All these disciplines represent
different views of concept systems.
Concept systems are fundamental to human
existence. They are part of man's mental activity, of his ability
to structure experience and knowledge. The organization of concepts
and the linguistic coding of experiences, observations, etc. are
especially important to the specialist, since more exacting demands
are made on concept formation and classification in special subject
fields than in everyday life. The concepts required in everyday
communication are vague and more or less unconscious in comparison
with those in special fields. The specialist needs a sophisticated
conceptual apparatus and a system of terms to cope with his
professional duties. He is also frequently faced with the task of
developing, revising, defining, standardizing, structuring,
classifying and naming concepts.
Researchers in the field of terminology science
are concerned with the conscious structuring of concept systems as
carried out in different special fields, while, for instance,
psychologists also investigate unconscious structures. The
importance of concept systems to special languages was stressed by
Eugen Wüster (18981977), who laid the foundations of the
theory the General Theory of Terminology which serves as the
theoretical basis for the present study. Wüster's successors
have also stressed the importance of concept systems to
terminological analysis and terminology work, but no comprehensive
investigation has been produced in this domain. In many quarters,
not least among those concerned with the development of term banks,
further development of terminological theory in this respect has
been wished for. Thanks to the computer, the graphical
representation of concept systems is no longer equally
circumscribed as earlier. It is now possible to illustrate quite
complex systems comprising different types of concept relations and
to retrieve these when needed. However, this makes certain demands
on the accuracy of theory and methods, which means that a more
sophisticated conceptual apparatus is required for the work.
Besides, attention has been drawn to the need for a theoretical
framework which could also be used in other special fields than
those of science and technology. Thus both the inadequate current
theory and practical demands have given an impetus to undertaking
the present investigation.
About the
investigation
Wüster provides the point of departure for
the development of my theory, but I endeavour to present a
synthesis of several different approaches, modifying them and
supplementing them with concepts and divisions of my own. I rely
for the most part on terminological literature, but also on
linguistic, semantic, philosophical and other literature in which
concept systems and concept relations or similar matters are dealt
with. The terminological literature consists mainly of terminology
standards, handbooks, textbooks, articles, conference papers, etc.
Some unpublished material is also included, for instance notes from
the Wüster Library at Infoterm in Vienna.
The study first deals with general aspects of
concept systems qua systems and their properties. Next the theory
of terminology is contrasted with different methods and theories
and concept structures used in studying and structuring the data
and concepts of several other disciplines. In the third and fourth
chapters I consider the components connected with concept systems,
i.e. concepts and concept relations, and work out a preliminary
classification of concept relations. This classification serves as
a point of departure for the classification and description of
concept systems in the fifth chapter.
Concept systems
The concept of concept system, which is one of
the most central theoretical notions in the theory of terminology,
is usually defined in terminological literature as a system of
related concepts which form a coherent whole. Starting from the
idea of system, concept systems could be regarded as systems
consisting of several components (concepts) and their relations
(concept relations). They are mental, i.e. abstract, artificial,
theoretical, man- made systems. They are static because they
represent the conceptual apparatus reflecting the knowledge which
exists at a particular time. New data result in new concepts, and
the emergence of new concepts changes existing concept systems as
has repeatedly happened for instance in the history of
biology.
It is important for terminology research to
distinguish between ontical systems, concept systems and term
systems. These three levels have their analogues in the classical
semantic triangle, which relates object, concept and designation to
one another. The ontical level represents the world and material
and immaterial things; at the conceptual level we find concepts and
at the level of expression there are the linguistic and other
symbols, which are used to refer to concepts. This study is mainly
concerned with the first two levels.
I distinguish between macro and micro concept
systems. The concept system at macro level encomprises all the
concepts within a special subject field or a clearly-defined
section regardless of the relations between them. A macro system
consists of several micro concept systems in which the concept
relations are more homogeneous. These micro systems are what
terminology science has normally been concerned with, and they will
not be unimportant in the future either. All the possibilities of
organizing concepts at micro level have not yet been chartered, nor
have all types of relations been defined. I focus for the most part
on the micro level, but at the end of the fifth chapter I propose
models for mixed concept systems which can cover several different
types of micro concept systems.
Concept relations
Concept relations and concept systems are
inseparable, since without relationships there would be no system,
and since relationships depend on the systemic context. Concept
relations may be strictly logical connections or freer associations
between one concept and another. They are mental entities which
link concepts to one another. Concept relations are thus one type
of concept, concepts of relationship, and, like other concepts,
they are the result of abstraction. Their referents are the
relations between individual entities, whether it is a question of
similarity or other relations.
In terminological literature concept relations
are classified in many different ways, which at least partly seems
to depend on the point of view adopted. Nor is there any agreement
concerning the terms used for concept relations. In connection with
concept relations it is also important to distinguish the ontical
from the conceptual level because it is sometimes difficult to keep
these two levels apart, especially as far as ontological concept
relations are concerned. Ontological concept relations are namely
based on ontical relations and they are simplifications of
relationships which can be observed among individual real-world
phenomena.
The difference between the various types of
concept systems depends on what kinds of relationships exist
between the concepts which they encomprise. I distinguish between
three different points of view from which concept relations can be
viewed and classified: a) concept relations which determine the
type of concept system (qualitative relations), b) relationships
concerned with the intension or extension of the concepts
(quantitative relations: intensional and extensional) and c) formal
relations which obtain between the concepts in a concept system
(systemic relations) (see fig. 136). It is difficult to draw a line
between these as all three groups of relations are intertwined in
diverse ways. The purpose of this classification is on the one hand
to bring together the different ways of viewing concept relations
which are represented in terminological literature and on the other
hand to create new concepts of relations and modify the
old.
Figure 136. Concept
relations.
The classification above is based on the
hypothesis that a concept analysis is carried out in three stages:
qualitative analysis, quantitative analysis and system-oriented
analysis. The first stage establishes which one of the three basic
types of relation is involved; in the second stage this relation is
further scrutinized, and in the third the position of the concepts
in the concept system is defined. This is, of course, only a model;
in reality it is difficult to keep these three phases
apart.
Qualitative aspects of concept relations and
concept systems
The qualitative classification of concept
relations is fundamental and forms the basis for the primary
division of the concept system as presented here. The division is
not based on a single criterion but is among other things
determined by what type the concepts represent, e.g. whether they
are object or process concepts. In accordance with Wüster I
divide both concept relations and concept systems first into
logical and ontological categories and then further into their
subcategories. It is stated that the essential difference between
logical (or generic) and ontological concept relations is that
logical concept relations are immediate relations between concepts
while ontological concept relations between concepts arise
indirectly (see Wüster 1985: 9, 12).
Figure 137. Qualitative concept
relations.
In agreement with Wüster I further divide
the ontological concept relations into concept relations of
contiguity and concept relations of influence. Among concept
relations of contiguity I include partitive concept relations
(house roof, door); accessorial concept relations (camera tripod,
film); local concept relations (lake fish, algae); material concept
relations (beer alcohol); attributive concept relations (wool
warmth); temporal concept relations (pre-wash washing); and concept
relations of rank (president vice- president).
Concept relations of influence I divide into
causal concept relations, developmental concept relations,
functional concept relations and interactional concept relations.
They relate to processes of influencing and causing and have a more
or less marked causal component. In causal concept relations the
causal component is the most prominent. They can all form concept
systems.
Figure 138. Concept relations of
influence.
Developmental concept relations comprise
phylogenetic concept relations (the development of a species),
genealogical concept relations (e.g. father son), ontogenetic
concept relations (chrysalis butterfly), material developmental
concept relations (e.g. crude oil petrol), and concept relations of
role change (e.g. accused guilty). They can all form more or less
comprehensive concept systems.
Functional concept relations I divide into
concept relations of activity and concept relations of origin
which, especially in terms of concept systems, resemble each other.
The primary concept in a functional concept system constitutes the
point from which the other concepts are viewed (Haarala 1981: 25).
Concept relations of activity obtain between a concept of
activity and one or more concepts linked with it, e.g. concepts
referring to object, instrument, material, etc. The element which
relations of origin have in common is the concept of a
concrete or abstract entity while the other concepts refer to the
origin of the object (e.g. original material, producer, instrument,
manufacturing method, manufacturing process, place of origin,
etc.). The term 'function' is here used in a very broad
sense.
Interactional conceptual relations are based on
an interplay of referent phenomena. I divide them into concept
relations of transmission, dependency and representation. The first
one is based on the relation between agents in a process of
transmission in which A gives / sends / transmits something to B
(e.g. transmitter receiver, coder decoder) while the other is based
on various types of economic, legal and other similar relations
which may obtain between different parties (e.g. employer
employee), and the third refers to the relation between an entity
and its representative (e.g. concept term, people
parliament).
Table 37. Concept systems based on the
qualitative concept relation
Concept
systems
- logical
concept systems
-
ontological concept systems
- concept
systems of contiguity
-
partitive concept systems
-
accessorial concept systems
- local
concept systems
- material
concept systems
-
attributive concept systems
- temporal
concept systems: with concept systems of succession or process
- concept
systems of rank
- concept
systems of influence
- causal
concept systems (stress on cause effect)
-
developmental concept systems (stress on change)
-
functional concept systems (stress on activity/origin of an
object)
-
interactional concept systems (stress on interactivity of
participators)
Quantitative aspects of concept relations and
concept systems
Formerly only the logical concept relations were
described from a quantitative point of view, but I assume that this
kind of approach could be extended and the same criteria applied
when analysing other types of relations to the extent that this is
feasible and expedient. I make use of quantitative classification
in this context, however, mainly when dealing with logical and
partitive concept relations. As regards logical concept systems,
quantitative analysis is fruitful and, in conjunction with systemic
analysis, leads to a classification of these types of system (see
below).
With reference to the classification of logical
concept relations it is expedient to regard the intension of a
concept as comprising its characteristics and the extension of a
concept as comprising a number of subordinate concepts or referents
and to deal with them quantitatively. Therefore I divide
quantitative relations into intensional and extensional and
distinguish both intensional and extensional identity, inclusion,
disjunction, intersection and conjunction.
When comparing the concepts in a partitive
concept system, it is not possible to use extension or intension as
criteria in the same way as in connection with logical concept
systems, but the comparison is made in terms of the sum of
components. It would thus also be possible to analyse the partitive
concept systems at least to some extent in terms of quantitative
relations (identity, inclusion, overlapping, disjunction) and not
only restrict these to logical concept systems. The study of
further possibilities of applying quantitative criteria in other
concept system is left for future research.
Systemic aspects of concept
systems
The last category of concept relations is the
system-oriented one, which at the same time is an alternative way
of classifying all concept systems. The starting-point is the
formal properties of concept systems, for instance the direction of
the relation and the position of the concepts in relation to each
other. On the basis of such criteria it is usual to distinguish
between hierarchic and non-hierarchic concept
relations.
The types of concept system that have been
classified on the basis of qualitative criteria seem to fall into
three different structural categories. Two of them are hierarchic
and sequential concept systems. The characteristic property of
hierarchic concept systems is that they have a superordinate
concept to which the other concepts are
subordinate.
The logical concept systems are hierarchic and
they can be divided into different categories according to their
structure: mono/polyhierarchic systems, mono/polydimensional
systems and combinatory and non- combinatory systems (see table).
The concept systems with contiguity relations belong more or less
distinctly to hierarchic concept systems.
The sequential concept systems make reference to
temporal connections between phenomena and consequently contain
simultaneous and consecutive concept relations. They thus for
instance contain concepts that refer to the beginning and the end
(if applicable) and to intermediate stages. Sequential concept
systems are usually presented as a type of non-hierarchic concept
systems without distinguishing any other
subcategories.
In addition, in my classification there are
systems in which no particular hierarchy nor any definite temporal
order between concepts can be perceived. To this group belong
functional and interactional concept systems as described above.
Causal concept systems constitute a borderline case as they contain
sequential relations but other relations as well (e.g. relations
based on co-ordination of cause). Some of the relations in
functional concept systems could also be considered sequential. The
individual functional concept systems at micro level often have an
unambiguous primary concept, but if a greater number of concepts
are included, a system with several potential primary concepts will
easily develop. Which concept assumes the role of primary concept
in a system depends on the angle of approach. In a scheme on a
larger scale, for instance a computer-aided presentation, the
question of primary concept may even be quite arbitrary. These
concept systems are difficult to classify as either hierarchic or
sequential. I suggest the term 'heterarchic concept system' for
this type of concept system. The term 'heterarchy' is defined by
Collins Dictionary (1989: 719) as "a formal organization of
connected nodes, without any single permanent uppermost node".
Heterarchic concept systems could thus be defined as concept
systems whose structure instead of displaying hierarchy or sequence
resembles a network with several possible points of departure or
primary concepts.
Table 38. Concept systems according to the
formal structure.
Concept
systems
hierarchic concept systems (the concepts
in a hierarchy, usually with a superordinate
concept):
- logical
concept systemsmonohierarchic logical concept system (one
superordinate for each concept)
-
polyhierarchic logical concept system (several superordinates for a
concept)
-
monodimensional logical concept system (one criterion of division
on each node)
-
polydimensional logical concept system (several criterion on one
node)
-
combinatory logical concept system (concepts can be combined with
each other)
-
non-combinatory logical concept system (concepts cannot be
combined)
- concept
systems based on contiguity (except temporal concept system)
sequential (concepts in a sequence or in
parallel alternative or coordinate chains)
- temporal
concept system
- causal
concept system (mostly)
-
developmental concept system
heterarchic (several concepts can
function as superordinates depending on viewpoint)
-
- (causal
concept systems, partly)
-
functional concept systems
- concept
systems with interactional relations (purely heterarchic
systems)
- satellite
systems
-
In addition to the
concept systems mentioned above I also include mixed concept
systems in my classification. They can be mixed hierarchic systems
and mixed systems comprising relations from different systemic
categories. Heterarchic concept systems are very often mixed by
nature. Both types of functional concept systems can thus be
divided into homogenous and heterogeneous systems; the former
including only one type of functional concept relations while the
latter include several types of relations. As a model for a macro
concept system I introduce a concept system that I call 'satellite
system' (see fig. 130133). The idea of a satellite system is that
the most important concept or a title is placed at a node and its
different aspects or most closely related concepts are arranged
around it. These "satellites" in their turn take their own
satellites etc. until, finally, the whole subject field is covered.
The satellite systems can be used in different phases of
terminological analysis: first they serve as an ordering device
when collecting concepts and terms, in the second phase they
connect different micro concept systems to each other and finally
they provide the macro concept system of the subject field under
scrutiny.
-
Conclusion
-
There
are many possibilities and approaches with regard to the
structuring of concept systems. In the present study I have summed
up several of them and endeavoured to devise a consistent
conceptual apparatus and terminology. The classifications of
concept relations and concept systems presented here are far from
exhaustive and final, for the more closely concept systems and
concept relations are examined, the more types of relations emerge
and the more types of concept systems could be included in the
classification. Similarly, with regard to the development of their
designations more work is needed. At this stage I employ terms
which are long but give some clue to the
concept.
-
The
classification can be developed and expanded through further
research in the field. I have aimed at universal applicability but
some of the concept relations and concept systems described above
are of limited use. If the classification is pursued further, one
will soon arrive at categories which are more or less dependent on
the subject field concerned. These problems are for other research
projects to deal with.
-
Further
efforts are also needed to devise graphic methods of presentation
and notations suited to each type of relation and each type of
system. Wüster devoted a great deal of time to creating a
notation system for concept relations, which, however, became so
complicated that it has not been adopted. He used notations which
are rather difficult to remember and to reproduce on a screen. In
the present study I have instead laid emphasis on describing and
exemplifying different types of systems without developing a
consistent system for graphic representation or
notation.
-
The aim
of this study has been theoretical, and therefore I have not dealt
with possible applications of concept systems in greater detail.
The needs of terminological analysis have however, been, a guiding
factor. It is my hope that part of what I have proposed here can be
used for the development of tools for terminological work,
especially in connection with computer applications, but also for
the analysis of terms and concepts for various other
purposes.
-
