ON THE PROCESSING OF PRINTED SUBJECT INDEX ENTRIES DURING SEARCHING
| Date | 01 April 1977 |
| DOI | https://doi.org/10.1108/eb026645 |
| Published date | 01 April 1977 |
| Pages | 266-276 |
| Author | E. MICHAEL KEEN |
| Subject Matter | Information & knowledge management,Library & information science |
ON THE PROCESSING OF PRINTED SUBJECT INDEX
ENTRIES DURING SEARCHING
E. MICHAEL KEEN
College of Librarianship
Wales,
Aberystwyth
Reports a laboratory experiment in which verbalized tape-recorded
searches on five printed subject indexes reveal something of the linguistic
processing that took place. Some 20% of the entries examined were
changed grammatically, by word order change or function word supply,
according to the linguistic form of the index concerned. Extracts from the
search transcriptions are given, and the search processing modes of seeking,
scanning, and screening are discussed.
INTRODUCTION
THE OPERATION of searching an index is still a black box: it is an activity
frequently undertaken, a skill that may be improved on by experience, and even
an operation whose outcome may provide the basis for index performance
measurement in evaluation
tests.
But so far no one has even slightly raised the lid
of
this
box, and there seems to be very little encouragement to do so from the
current state of fields such
as
cognitive studies, psychology, linguistics, and
so
on.
Indeed a complete understanding of the operation would have to include human
mental processing,
as
well
as
the hard-to-capture eye-movements and judgements
of aboutness and relevance.
However, the view has already been advanced that in spite of the present
impossibility of flow charting or modelling the process of searching in detail it
is already possible to obtain some first answers about the process of searching.1
The questions for which answers are being sought are similar to some of those
that have been asked in the field known as psycholinguistics: since retrieval
systems of all kinds provide an output that
is
to be read or visually scanned by the
searcher, the linguistic content of this output clearly needs to be studied in the
context of
its
use in searching.2 These linguistic questions are most clearly seen
in the realm of different kinds of printed subject index entries: for example, are
entries that follow active rather than passive grammatical form 'better'? Are
entry schemes that are incomplete grammatically (consisting only of content-
bearing terms without connectives or function words) better than schemes that
are like complete sentences? What about schemes of contextual dependency that
fragment the order of terms in entries, such as the PRECIS scheme: do they aid
the objectives of index searching?
This paper presents the results of
a
small experiment in which five different
kinds of printed subject indexes were searched, and from transcripts of the
verbalized tape-recordings made some evidence of the way the entries were
processed can be found. These results also throw further light on the relative
merit of the index entry types under test in the project of which this experiment
was a part, namely EPSILON.1,3 Before the experiment and its results are pre-
Journal
of
Documentation,
Vol.
33,
No. 4, December
1977,
pp. 266-276
266
To continue reading
Request your trial