A SCIENTIFIC THEORY OF CLASSIFICATION AND INDEXING AND ITS PRACTICAL APPLICATIONS
| DOI | https://doi.org/10.1108/eb026155 |
| Pages | 83-99 |
| Published date | 01 February 1950 |
| Date | 01 February 1950 |
| Author | J.E.L. FARRADANE |
| Subject Matter | Information & knowledge management,Library & information science |
A
SCIENTIFIC THEORY OF CLASSIFICATION
AND
INDEXING AND ITS PRACTICAL
APPLICATIONS
by
J. E. L.
FARRADANE,
B.SC.,
A.R.C.S.
Scientific Information
Officer
Tate & Lyle
Research
Laboratories
SUMMARY
A classification is a theory of the structure of knowledge. From a discussion of
the nature of truth, it
is
held that scientific knowledge is the only knowledge which
can be regarded as true. The method of induction from empirical data is therefore
applied to the construction of a classification. Items of knowledge are divided into
uniquely definable terms, called
isolates,
and the relations between them, called
operators.
It is shown that only four
basic
operators exist, expressing appurtenance,
equivalence, reaction, and causation; using symbols for these operators, all subjects
can be analysed in a linear form called an
analet.
With the addition of the permis-
sible permutations of such analets, formed according to simple rules, alphabetical
arrangement of the first terms provides a complete, logical subject index. Examples
are given, and possible difficulties are considered. A classification can then be con-
structed by selection of deductive relations, arranged in hierarchical form. The
nature of possible classifications is discussed. It is claimed that such an inductively
constructed classification is the only true representation of the structure of know-
ledge, and that these principles provide a simple technique for accurately and fully
indexing and classifying any given set of data, with complete flexibility.
I. THEORY
THIS
paper concerns the classification of knowledge as a whole, and not only
the relatively simple classification of limited groups of objects.
Classification is a theory of the structure of knowledge, i.e. of the relations
between different parts of knowledge. No arbitrary method of grouping,
however carefully applied, is true classification. The problem is primarily
epistemological. What is true knowledge, and what are true relations
between the parts of knowledge? It is essential to define these if the classifica-
tion is to be true and logically sound.
This theory is being propounded from a scientific viewpoint, since it is
held that the only true knowledge is scientific knowledge. This is of course
a much debated subject and cannot be discussed fully here, but a brief
expo-
sition is essential, since the various systems of classification of knowledge as
a whole have hitherto not been based oh scientific principles.
Types of knowledge
Knowledge may be postulated as being of three possible kinds: (1)
a
priori,
(2) empirical, i.e. directly experienced, and (3) logically derived by a process
84 THE JOURNAL OF DOCUMENTATION
VOL.
6,
No.
2.
of induction or deduction (or related forms of reasoning). We can immedi-
ately eliminate (1), since any truth in a
priori
knowledge rests only upon
assumptions concerning the nature of mind, or upon belief in intuitions;
such knowledge is moreover uncommunicable in any exact form, and is
unverifiable by a second person. To a certain extent we are obliged to accept
empirical knowledge, since any kind of knowledge is obviously related to
primary forms of experience, learning of language, and awareness of our
surroundings through the human
senses.
It is recognized, of course, that such
empirical knowledge is in some ways as much lacking in proofs as
a priori
knowledge; its claim to truth, without further use of logical processes,
depends upon the repetition of experience both in one individual and in
many individuals, and is ultimately a matter of wide consensus of opinion
concerning such communicable experiences.
The only methods known to us to enhance the likelihood of truth in such
empirical experiences are the well-known logical processes. It should be
noted that logic is in itself partly an
a priori
process, in the assumptions of the
truth of certain linguistic processes, and partly rests upon empirical know-
ledge, especially of the primary visual, tactile, and auditory experiences.
Attempts to place logic upon a mathematical basis do not enhance its intrinsic
reliability, since mathematics is probably based initially upon empirical
experience (the five fingers of the hand, &c.), and, if it is not so, then it can
only be considered (as it is, in fact, often claimed) to be initially of an
a priori
nature. However, as we have to take human beings, and human knowledge,
as we find them, we must assume that absolute truth will always be beyond
us,
or at least beyond recognition as such, and that such relative truth as we
can achieve must be related to the human senses. We must accept the best
forms of logical processes as providing proofs of truth within the relative
framework. What forms of logic shall we use? Results from purely deduc-
tive logic, as generally employed by philosophers to the exclusion of induc-
tive processes (despite the fact that, in the fullest sense, philosophy should
embrace all methods and all knowledge), are inescapably dependent upon
the truth of the initial generalizations from which the deductions are made.
It must be a generalization, and not a single empirical fact, which
is
employed
in the beginning of the argument, since no deduction can be made from
a single fact, other than statements deriving from linguistic usages (e.g. 'This
is a black object, therefore it is not a white object', which usually leads to
the 'A or not A' fallacy, whereas many objects may be 'partly A'). But
a generalization can be obtained, in the first instance, only by
a priori
or by
inductive processes. Deductive logic from an
a priori
generalization bears no
more truth than that of the initial
a priori
statement.
Induction
We therefore come to consider inductive
processes.
It
is
agreed by modern
logicians (e.g. Keynes, Bertrand Russell) that the truth of an induction is
a matter of probability, and attempts have been made to express the prob-
ability in mathematical form. Attacks by philosophers upon the scientific
To continue reading
Request your trial