RANGANATHAN'S THEORY OF FACET ANALYSIS

AND KNOWLEDGE REPRESENTATION

Documentation Research and Training Centre. Indian StatisticalInstitute

Bangalore-560059

Abstract
Presents the basis of than's theory ofFaeetAnalysis and
identifies its application to the analysis of knowledge. Discusses
its value in knowledge representation in different contexts.

1. Preamble of Library Classification, normative prin-

ciples for cataloguing, scientific prin-
l992 is the birth centenary year
ciples for library management, efficient
of Dr SR Ranganathan, National Re-
approaches to handling library measure-
search Professor in Library Science
m e n t s , d e v e l o p m e n t of n a t i o n a l
(1965-1972). Government of India. He documentation systems and library net-
was a savant of library and information w o r k s . He developed s t a n d a r d s ,
services; who promoted & organised specifications and guidelines through
university library system. public library the National Standards Bureau of India.
network on state-wide basis, specialised In short, he built many infrastructures
R&D, industrial, and business informa- to promote organised library services in
tion centres. He also promoted and cul- India . This article highlights the value
tivated specialised personnel with of facet analysis in knowledge repre-
expertise a n d skills for h a n d l i n g sentation in the contemporary scene.
knowledge in different ways. He created
and contributed to the literature of 2. introduction
library science and propounded the Five Ranganathan strived to develop an
laws of Library Science, dynamic theory a n a l y t i c o - s y n t h e t i c s t r u c t u r e of

16 DESZDOC Bull InfTech 12(5)September 1992. pp. 16-20

Dr SR Rangamthan Birth Centenary Issue

knowledge. He identified his studies on Subject'; with each basic subject, several
the way knowledge gets embodied in entities. their properties, and actions
books, periodical articles. and other can be studied in relation to space and
rendered material; and based it on the time. Based on this analysis he identified
way in which the seekers approach a scheme for classification. he postu-
knowledge. Further, his analysis of con- lated categories that gave generalised
ceptual a s p e c t s of a n information abstractions. In each instance, a subject
storage and retrieval system would pro- may have the presence of one or more of
vide for achieving the following aspects : these categories. These categories are
stated a s follows :
(a) Providing better access to contents
of documents; % T h e : The pervasive temporal
aspects of the world such as day.
(b) Narrowing down and broadening the night. year, century. etc.
search:
4: Space: Geo a n d o t h e r s p a t i a l
(c) P r o v i d i n g a l t e r n a t i v e s e a r c h concepts in the relational system
patterns; such as India, USA. USSR etc.
(d) Distinguishing between no matches Energy: Actions of various varieties
due to search error, and no matches s u c h as r u n n i n g , m a n a g i n g ,
between t h e i t e m s n o t i n t h e communicating, modelling, etc.
database.
1 Matter: The properties and other
However, this analysis of information attributes of our object, such a s
retrieval should be backed u p by the name, valency. mobility, viscosity.
ability to analyse and synthese texts in etc.
a variety of ways. His postulates on
1 Personality: T h e o b j e c t s , t h e i r
analysis lead to facet analytic approach.
instantial variations, s u c h a s
Ranganathan said on the role of clas-
automobiles, passenger cars, sports
sification thus : 'A traveller will be helped
car, etc.
muck more, saved many wrong turns
and attracted to the home more easily, if * Basic subject: Chunk of a field of
the light put up by the host is a search knowledge, convenient a s a field of
light, which not only acts as a beacon but activity; such as 'Vehicle Engineering,
also lights u p the region through which Land Vehicle Engineering and Motor
he has to walk.Then there will not mere- Vehicle Engineering.'
ly the goal be glimmering but every inch This is a core model from which one
of the way will be illuminated. The for- can derive any number of representations
mula of facets and phases of a well- of knowledge embodied in documents and
other kinds of materials.
designed analytico-synthetic scheme of
classification will serve the purpose of
such a light'(see reference 1).
4. E x a F ,,, D i
Database For Public Finance
3. Facet-analytic Approach Example 1 Economic and politicd
concept modelling t h r o u g h foundations of tax structure. (Published
analytico-synthetic approach, we can i n American Economic Review. 1988,
adopt Ranganathan's approch (see refer- 78(4), 701-12).
ence 2). In his model, he proposed a Analysis:
universe of discourse, called 'Basic Time : 1988

DESIDOC Bull Inf Tech '1215) September 1992. pp. 16-20 17

 Dr SR Ranganathan Birth Centenaty Issue

Space : USA 5. Knowledge Representation Sys-

Matter: Tax Structure tems
Personality 1 : Political System
Knowledge bases have to be or-
in relation to
ganised and structured a s the rules em-
Personality 2 : Economic System
bodied in them go on increasing. Thus,
Basic Subject : Public Finance
knowledge representation language
Here we see no instance of 'Energy' comes into play. Many issues in this
category. context handle both conceptual factor-
E x a m p l e 2 Another c o n c e p t : ing and computational efficiency. In
Federal Government Budget deficits techniques such a s semanllc nets, con-
a n d i n t e r e s t r a t e s : a n empirical ceptual taxonomies, frame systems, etc;
analysis for the US. 11955-84. (Publish- knowledge is conceptually factored ac-
ed in Public Finance. 1988, 43[3),337- cording to a hierarchy or generalisation
48)
structure. I t is a kind sf taxonomic or
classificatory structure. KL-ONE, is a
Analysis:
knowledge r e p r e s e n t a t i o n s y s t e m
Time : 1988 oriented around a taxonomy of struc-
Space : USA, tured concepts. The taxonomy for or-
Energy : Analysis Empirical analysis ganising a knowledge base of rules is
Matter : Rate of interest described in the following paragraphs.
Personality : Public Expenditure
Basic Subject : Public Finance A concept mode in KL-ONE is as-
in correlation with sociated with a set of rules (a generdisa-
Personality : Economic System tion) of the notions of attribute, part,
constituent feature, etc] and a set of
S u b Field : Deficit Budget
s t r u c t u r e d conditions expressing
correlated with
relationships among them. These con-
Co-entity : Political System
cepts are linked to more general con-
Sub Field: Federal Government
cepts by explicit links. The more general
Basic Subject : Public Finance concept in such a relationship is called
Thus, we see that the expressions of the super concept and is said to sub-
summarisation of information contents sume the more specific subconcept.
of objects can be analysed into Basic Some of the concepts' roles and struc-
subject, Personality, (=EnUty)and other tural conditions are attached to it direct-
categories. Such an analysis leads to a n ly, while others are inherited indirectly
intermediate lexicon of ideas which can from more general concepts.
be used as modelling media for input The concepts and roles of KL-ONEare
information in any specialised field a s similar in structure to classical data
well a s a media for searching informa- structure notions of record and field or
tlon from the database. Lists of various the 'frarnel/'scheme'/'unit'and 'slot' of
f e a t u r e s of Political S y s t e m a n d much A1 terminology. However, they dif-
Economic System for Public Finance, fer subtly in function and interpretation
useful for analyac-system can be easily since they are motivated more by an
built into a design database. We shall attempt to model the semantics and con-
discuss this process in knowledge-based ceptual structure of an abstract space of
Management systems also. concepts, than by issues of data and

X)C Bull I n j Tech. 1215) September 1992. pp. 16-20

Dr SR Ranganahart Birth Centenary Issue

referral operations. This difference s e n t a t i o n r e s e a r c h , with t h o s e of

manifests itself in a variety of details, database organisation and retrieval in
such as the way subsumption is defined order to develop the techniques neces-
and used, the pressure of associated sary to handle large knowledge bases of
structural conditions attached to a con- rule-like formation.
cept, and the representation of explicit
relationships between roles a t different
6. Development Of Decision Sup-
levels. port System For information
The ability to a s s i d a t e new descrip- Retrieval
tions into an &sting Wonomy a t any Knowledge representation languages
level permits a n evolutionary system can provide a base for an intermediate
design that achieves the same standards lexicon. The creation of a n intermediate
of rigour a s a "top-down design" without lexicon as a base for inpubs and outputs
requiring concepts to be defined in.a pre- in information referral helps in the
determined order. For most applications, development of intermediary system.
The intermediary system will have the
even if one could get the initial design
following features :
carefully laid out in top-down mode,
subsequent changes will require a n (a) It would accept statements made in
ability to modify a system i n more a natural language a s expression,of
flexible ways, the query by the information seeker.
A system's taxonomy of recognisable (bj It would structure this statement
situations should be viewed as a n evolv- i n t o d e s c r i p t i o n s o r key
ing knowledge structure that continues words/phrases for matching the
to be defmed and developed throughout same with system's thesaurus.
the Uetime of a system, just as it is for (e) I t a r r a n g e s d e s c r i p t i o n s i n a
human beings. In particular, eoncep- sequence wherein the search for
t u a l l y - f a c t o r e d , taxonomic repre- information is conducted efficiently
sentation systems s u c h as KL-ONE and productive retrieval results are
appear to be well suited to such applica- provided.
tions (see reference 3). These knowledge (d) It would interactively negotiate witti
structures can be used to perform a kind the information retrieval system in
of parsing of the situation, using the modulating the response to a query.
patterns and schemata of the knowledge It would trigger interaction between
base a s its 'grammar'. The way in which user and the system.
elements of knowledge base are accessed (el It would ask the user for terminology
in this process, differs substantially modulation in understanding the
from the way in which elements of a user's request and system's term
traditional databases are accessed. (synonym, homonym, a n d other
Such access will require techniques hierarchical structuring). These
that are analogous to and generalisa- actions would of course be well
tions of tradional parsing algorithms to structured by :
achieve reasonable response times. A (i) Establishing the field of knowledge
substantial challenge lies ahead of u s to to be searched in the database (boun-
combine the insights of natural language daries) and the comprehension of an
understanding and knowledge repre- information retrieval thesaurus.

DESIDOC Bull Inf Tech 1.2151 September 1992.pp. 16-20 19

 1Dr SR Rangmathan Birth Centenary h e

IiiJ Understanding of user's levels of concept codesjmodels for knowledge

expertise in the subject area concerned. representation is one of the primary
(lii) Protocols In t h e system to ways we can develop software for utilis-
negotiate with the user on and off the ing the large knowledge base systems for
search time and process. intelligent front-end usage. This has
Ranganathan speaks of classiflca- b e e n e x p r e s s e d by R a n g a n a t h a n
Uon features in this vein 'CPasstflcaffon aptly,Telassification can illumine the
is an uncovering of the thought-content field of knowledge, it can be prophetic.'
of a written or expressed unit of thought. 'P"s make it fit for this new role, it must
ln dl its minutest detaiis it is what it be placed on a new type of foundation
should be only when the thought-con- and built in a new way. This calk for
tent as a whole is uncovered including considerable research. Communication
all the phases. facets, and foci by those of thought has to depend on helpful ar-
faculties which apprehend all the details rangement and selection. tever be
e&austively, and yet in terms of entirety
the psychology of the arranging and
rather than in terms of merely parts. It
selecting process, t h i s process is
is only an analytico-synthetic scheme
facilitated by the u s e of artificial
that can uncover thought-content in
such a holistic way' (ref. 1, p. 1 15).Thus, counters or symbols . . . Lmpiy a s much
classification provides a kind of relation- purposive abstraction a s mathematical
al framework for knowledge-base sys- symbols or symbols of f o r d logic (ref.
tems. 1. p. 260).

7. Conclusion References
The development of conceptual 1. Ranganathan, S.R. Classificatiofi
structure. cIassification structure, and a n d communication. Sarada
the knowledge structure representation R a n g a n a t h a n Endowment for
techniques indicate the varieties of ways Library Science, Bangalore. 1989.
in which concept modelling can proceed 2. ganathan, S.R. Prolegomena to
towards developing the basic taxonomy
library classification, Ed 3, Asia
for information processing and modell-
Publication House, New Deihi, 196'9.
ing for use in information retrieval sys-
tems. With the hardware development 3. Brachman, R.J. & Schmolze, J. An
for compact storage, faster retrieval and overview of the KL-ONE hawledge
search processes are to be made efficienr representation system. Cognitive
and effective. The software in the form of Science. 1985, 9, 171-216,

"Man's perqonaiity is dynamic. It is ever in flux. it seeks

unfoldment at its own speed, in its own way, and towards its
--ownfullness. Such unfoldment is true education.@
1 -SR Ranganathan in Library Book Selection
I

DESIWUC Bull f n J T e d ~1215) Sepremkr 1932, pp. 16-23