The indexing of museum objects
Leonard Will
The principles of indexing museum objects are the same as for printed publications, but several
new problems arise because background information has to be indexed as well as information
about the object itself, and museum indexes are increasingly part of computerized collections man
agement systems rather than just catalogues. Comprehensive data standards are being developed
by several projects. Thesaurus-controlled indexing, classification schemes and free-text indexing
all have their parts to play.
Introduction
Most museums use computers to record the objects in
their collections, so this article is written with that
approach in mind, but it is primarily concerned with the
information science aspects and the data which is stored
and indexed rather than with the information technology;
the principles generally apply to manual systems as well as
to small and large computer applications. There are some
strongly-held differences of opinion in the museum world,
and the views in this article are my own assessment of the
problems and my judgements of what is desirable; they arc
not necessarily shared by any organization with which I am
associated. Another view of the problems of indexing
museum objects was given in The Indexer three years ago
by Roger Smither of the Imperial War Museum1.
There are several categories of index term by which this
object might be sought: what it is, the disciplines or areas
of study to which it relates, and where, when and by whom
it was built and used. Appropriate indexing terms might be:
Indexing objects rather than text
The indexing of museum objects has many similarities
to the indexing of written material, but there are many dif
ferences too, and these present some interesting and chal
lenging problems. The main one is that while most
indexing jobs start with a block of text, museum catalogu
ing and indexing starts with a box of objects. These may
have words on them, but usually the words have to be sup
plied. The cataloguer has to decide whether a vessel is a
cup, a beaker, a mug or just a drinking vessel.
In theory, the process should be straightforward: a cura
tor should write a description of the object in whatever
terms seem natural and appropriate, and an indexcr can
then work on this textual description as on any other text,
providing standardized access points under controlled
terms at a level of specificity appropriate to the collection.
Systematic retrieval is thus provided for, and when an item
has been identified as being of interest the curator's
description of it is available.
To take an example from the Science Museum, we have
an object with the following semi-free-text description:
1986-1502: Experimental bed for continuous weight
measurement during psychic experiments, used by Prof.
J. B. Hasted in his research, English, c. 1980
The Indexer Vol. 18 No. 3 April 1993
In practice, the distinction between description and
access points is not so easy to maintain. Curators common
ly do write free-text descriptions of objects, particularly
when recording accessions or when preparing scholarly or
evaluative catalogues, but it is often desirable for them to
provide structured descriptions too, with information
broken up into fields and groups. This: (1) allows data to
be selected, sorted and presented in different formats; (2)
acts as a reminder to the curator of data that ought to be
included; and (3) makes it easier to scan a list of records
looking for or comparing specific items of information.
Recording of data in a structured form blurs the distinc
tion between description and access points, and there is a
temptation to make a single set of data fields perform both
functions. They can often do so, and then there may be
economies in the amount of data to be stored. The lack of
distinction has, however, led to enormous confusion, with
curators complaining that indexers and other information
staff were trying to force them to use terms that are
anachronistic or that do not reflect scholarly discrimina
tions, while indcxers complain that curators are not pre
pared to accept the standardization of terminology which is
needed to make an indexing system reasonably efficient.
For most museums, the curator, the cataloguer and the
157
INDEXING OF MUSEUM OBJECTS
indexcr are the same person, who has to reconcile the con
brief mention of people, organizations and places.
flicting points of view.
Object names and subjects
Data structures and standards
Much work has been done nationally and international
ly over the past fifteen years or so to develop agreed struc
tures for museum data and terminology. In the UK the lead
has been taken by the Museum Documentation Association
(MDA), while international work has been done by work
ing parties of CIDOC, the International Committee for
Documentation of ICOM, the International Council of
Museums. Reviews of work in many different institutions
and organizations have been published in the proceedings
of two meetings.2-1 Over the past year there has also been a
major project undertaken by consultants, on behalf of a
consortium of several museums. This project, called
LASSI (LArge Scale Systems Initiative), is investigating
the data and procedures used in managing collections,
including their cataloguing and indexing, to determine
whether it is feasible for several museums to use a common
computer system.
The MDA's work over the last twenty years has made
museum staff much more aware of the potential of their
collections as a source of information and of the need to
organize that information so that it can be retrieved effec
tively. The latest edition of the MDA data standard4 pro
vides for a great deal of analysis of data into a complex
hierarchical structure, but I fear that it has reinforced the
confusion between description and access points referred to
above; besides, much of the detail is too subtle for practical
application. The MDA has therefore set up a UK Museums
Data Standards Initiative, under which a wide range of rep
resentatives of museums of all kinds are reviewing data
and procedures, in the hope that they can achieve the
demanding task of producing a standard which is practical,
understandable, acceptable, implemcnlable, compatible
with the work of LASSI, and hospitable to data which has
already been recorded.
These projects arc wide-ranging, as they arc trying to
model the overall data needs of museums, and therefore
their time-scales are measured in years. Many museums
have had to take more pragmatic decisions to set up their
own systems in the meantime; some of these systems arc
very effective, but sometimes the basic principles have had
to be learned the hard way, by trial and error.
Another MDA working group is concerned with termi
nology, and has recently been concerned with the drafting
of guidelines for the construction of a thesaurus of museum
object names.s The hope is that specialist subject thesauri
developed in accordance with these guidelines will be com
patible with one another, so that they can be combined
without major structural changes.
In this paper 1 shall not attempt to cover the work on
collections management, although that introduces many
interesting questions of indexing and terminology control
(e.g., for methods of acquisition or storage location codes).
Instead, I shall concentrate on subject indexing, with a
158
Fields and structure
The MDA standard includes fields* called name, simple
NAME, FULL NAME, OTHER NAME, CLASSIFIED NAME, CATEGORY,
CLASSIFIED
CATEGORY,
FULL
CATEGORY,
SIMPLE
CATEGORY,
descriptor and concept, among others. Many of these
fields can have further sub-structure. Although the standard
gives definitions, the distinctions are difficult to convey to
non-specialists.
It is intended that a computer system
should act as an interface so that users do not need to be
aware of the underlying detail, but the only such system
which exists, modes, uses a very reduced and simplified
version of the standard.
Although the MDA data standard allows an object to
have several occurrences of simpi£ name, many museums
have imposed a limitation of one simple name per object.
This allows lists to be produced in which each item appears
only once, and a single name can even be used for the
physical arrangement of objects in storage, shelving all the
teapots together, for example. Under this system, the object
described above would probably be given the simple name
bed, although it could well be argued that it should be
weighing machine or even psychic apparatus. These addi
tional terms would be recorded in the other name or full
name fields, while terms such as levitation might be record
ed in the concept or full category fields. The full name
field is sometimes used to contain terms that qualify or
amplify the simpi£ name.
I think such distinctions arc more bother than they are
worth. In particular, dividing index terms into many differ
ent fields means either that each field has to be specified
separately when constructing a search statement, or else a
combined index has to be prepared, which contradicts the
argument for separating them in the first place. Most
thesauri used for indexing bibliographic material have no
problem in accommodating different types of term, includ
ing names of objects, disciplines and activities; these arc
not hierarchically related, but it is convenient to be able to
make associative relationships ('related term' or 'sec also'
references) between them.
Pre- or post-coordinate indexing
The assumption in most museum cataloguing systems is
that searching will be done by computer, rather than from
printed output. A browsing display of pre-coordinated
strings can be helpful, but a post-coordinate system gives
greater flexibility and economy by indexing each item with
individual terms rather than strings. There is some lack of
precision in this, but my experience is that most searchers
* I have used the convention that names of fields are shown in
italic capitals, such as simple name, while examples of data which
may occur in these fields are shown in italic lower case, such as
beds:
The lndexer Vol. 18 No. 3 April 1993
INDEXING OF .MUSEUM OBJECTS
prefer to search on individual words anyway, not being
confident that they can construct a search string which will
match any stored string. The necessary complexity of syn
tax rules for pre-coordination makes them unlikely to be
used consistently (cf. udc and precis) although a scheme is
described in the introduction to the Art and architecture
thesaurus.*
Vocabulary control
The names of objects and other subject indexing terms
used for retrieval need to be standardized by using a con
trolled vocabulary list of some kind, preferably a thesaurus
that gives a full structure of relationships between preferred
and non-preferred terms, broader and narrower terms. Beds
could then be retrieved in a search for all kinds of
furniture, and weighing equipment in a search for all types
of measuring equipment. Links could also be made
between related terms for disciplines, phenomena, objects
and so on, such as parapsychology, levitation, and weigh
ing equipment.
Though the principles of constructing a thesaurus set
out in the British Standard apply to thesauri for museum
objects as well as for any other type of material, no single
thesaurus is generally accepted and used. Various term lists
have,
however,
been
used
or developed
by
different
museums. Published lists include Nomenclature7 and the
Hertfordshire simple name list,* neither of which has a
terms timers and cooker components, and a handle broken
from a vase might be indexed as handles and vase frag
ments. There needs to be local agreement on how this
approach is to be applied to a particular collection, because
if the components term is introduced at too low a level the
size of the thesaurus could almost be doubled, with every
object name having a corresponding components term.
The inherent problem is that we need role indicators,
but introducing these brings complexity into the simple
post-coordinate system we would like to use. Anything
more complex would not be catered for by any of the avail
able software packages, and would be difficult to explain to
the many people without specialist indexing knowledge
who will have to input and retrieve information. Any fur
ther ideas on this point would be welcome.
Classification schemes
So far we have just discussed alphabetical subject
indexing. A complementary technique is the assignment of
classification numbers or codes to object records. These
will be used primarily to generate printouts of specific col
lections in a subject order, and correspond to the guide
cards used in a file of inventory cards for the collection.
Classification codes have the advantages that they allow
subject lists to be produced under complex subject head
ings, which may be specific to the context of the collection;
they allow physical arrangement of objects in store in sub
BT/NT relationship, and the Art and architecture
thesaurus, which does not allow a term to have more than
ject order, like shelfmarks in a library; and they allow
one broader term and which is very large because of the
other hand, they have the disadvantages that an alphabeti
inclusion of very specific terms in some areas. In the
Science Museum we have made some use of the British
Standards Institution's ROOT thesaurus,'' which is quite
well structured although it docs not cover all the areas in
which we are interested.
Since many objects have been catalogued without
agreed guidelines on terminology, there have to be some
compromises. In particular, we may have to accept that an
cal index to the classification has to be provided, and ideal
ly a searcher needs to have a copy of the classification
selection of a prc-dctcrmined subset of a collection. On the
scheme, to identify the range of classification headings
which are relevant.
hang-over from the view that it is a description of an
object, rather than the plural, beds, as would be more logi
A classification scheme shares many of the advantages
and disadvantages of a thesaurus, and, indeed, the hierar
chical relationships of a good thesaurus lead to an implied
classification in any case. One classification scheme quite
widely used in the UK is the Social History and Industrial
Classification (SHIC)," which uses as its primary principle
of division the historical context in which objects were
cal for an indexing term describing the category to which
used. This gives a different perspective from the thesaurus
the object belongs, and as is recommended in the British
schemes which focus on what the objects actually arc.
object name will be recorded in the singular, bed, as a
Standard for thesaurus construction.1"
Parts and wholes
The British Standard recommends that for indexing
Free text
As well as controlled indexing terms and classification
schemes, a third necessary approach is the indexing of the
parts or components, separate terms should be assigned for
individual words in object descriptions. This generally uses
the component and for the object of which it forms part, so
some form of inverted file in conjunction with a stop list,
that aircraft engines would be indexed by the two terms
and enables fast retrieval of records containing any word or
aircraft and engines. This causes problems in a museum
combination of words, without the computer having to read
through all entries.
collection, because items indexed in this way would be
retrieved in a search for aircraft, when only complete air
craft were being sought. It therefore seems preferable to
use a term such as aircraft components. A particular engine
may well be an aircraft component, but it is not an aircraft.
Similarly a timer from a cooker can be indexed by the
The Indexer Vol. 18 No. 3 April 1993
Free text has the great advantage that it is quick to
implement, with no specific maintenance effort required
from curators, apart from writing informative and consis
tent descriptions of objects. Searching on the actual words
used in the descriptions is helpful in identifying half-
159
INDEXING OF MUSEUM OBJECTS
remembered items, and will retrieve by specific terms
peculiar to individual objects, trade names, jargon, etc.
On the other hand inconsistencies in description mean
that the searcher has to think of all terms which might have
been used, including synonyms, grammatical variants,
generic and specific terms. Many existing inventory
descriptions are inadequate, and editing them will take a lot
of work, generally requiring examination of the objects.
When searching free text there is no guarantee that
items retrieved represent all items of that type in the
museum, because other items may have been differently
described.
People and places
As well as naming objects and retrieving them by sub
ject, museum documentation staff have also to set up
indexes for many other entities, such as people and organi
zations, modem and historical places, events and materials.
Detailed rules for the forms of names to be used for people
and organizations are given in the Anglo-American
Cataloguing Rules12 widely used in libraries, and there is
general acceptance that these can be used as guidelines for
museum applications too, although many existing users
apply simpler rules of their own. The rules for the choice of
name applicable to books are not so easy to apply to
objects, because names arc not generally printed on the
items themselves but have to be obtained from accompany
ing documents or supplied from the cataloguer's own
knowledge or from reference sources. There arc several
projects in progress to compile lists of names of artists and
scientific instrument makers, for example. These files can
be expanded to form biographical data files, and name
indexing is then done by creating links between biographi
cal and object records.
Biographical records can take advantage of a thesaurus
structure to hold links between related people and organi
zations, as well as references from non-prefcrred to pre
ferred forms. In indexing places, it is convenient to create a
full hierarchical structure, so that, for example, an item
which has been indexed with the term Inverness will be
retrieved in a search for any Scottish items, assuming that
the computer system can accept an enquiry of the form
* Scotland and all its narrower terms'. Such systems are not
widespread at present, so that museums often have to use
the more cumbersome approach of including every level of
the hierarchy in every record catalogued. The above item
might then have to be given the terms Europe, United
Kingdom, Scotland, Highland region and Inverness rather
than just the one most specific term.
Conclusion
Indexing in museums is still in its infancy.
Documentation has never held the central place which it
does in libraries, but it is growing in importance and recog
nition, and the number of information specialists in
museums continues to increase. It is still very unusual to
find a museum which makes its full catalogue available for
160
searching by visitors in the way that libraries do, but the
increasing development of visitor-oriented information ser
vices drawing on integrated curatorial and library resources
encourages us to look forward to a period of rapid change
in the next few years.
References
1.
Smithcr, Roger. 'Is Britannia a personality?': some questions
arising while indexing the Imperial War Museum's collections.
The Indexer 17 (I) April 1990. 7-11.
2.
3.
Terminology for museums: proceedings of an international
conference held in Cambridge, England, 21-24 September
1988. Edited by D. A. Roberts. Cambridge: Museum
Documentation Associalion. 1990. 623p. isbn 0-905963-62-8.
Thesauri for museum documentation: proceedings of a work
shop held at the Science Museum, London. Cambridge:
Museum Documentation Association. 1992. [In press], isbn 0905963-84-9.
4.
The MDA data standard. Revised edition. Cambridge:
Museum Documentation Associalion. 1991. I vol. (loose-leaf).
isbn 0-905963-74-1. Previous versions of this were distributed
5.
informally, and extracts from il have been incorporated as (he
data structure distributed wilh the MDA's MODES (Museum
Object Data Entry System) software package.
Guidelines for constructing a museum object name thesaurus.
Compiled by Stuart A. Holm. Cambridge:
Museum
Documentation Associalion, 1993. (In press.]
6.
Getty Art Hislory Information Program. Art and architecture
thesaurus. New York; Oxford: Oxford University Press, 1990.
3 vols. isbn 0-19-506403-8.
7.
The revised nomenclature for museum cataloging: a revised
and expanded version of Robert G. Chenhall's system for clas
sifying man-made objects, (edited by] James R. Blackaby,
Patricia Grceno. Nashville, Tcnn.: American Association for
Stale and Local Hislory. 1988. isbn 0-910050-93-7.
8. Hertfordshire simple name list: a thesaurus of terms for use in
cataloguing general social history collections: designed for
use with Museum Documentation Association museum object
cards; compiled by a working party of the Hertfordshire
Curators' Group. [Hertford?]: Standing Committee for
Museums in Hertfordshire. cl984. 177 leaves, isbn 0-90135430-9.
9.
British Standards Institulion. ROOT thesaurus. 3rd ed.
London: BSI. 1988. 2 vols.
10. British Standard guide to establishment and development of
monolingual thesauri. (BS 5723: 1987: ISO 2788-1986).
London: British Standards Institute. 1987. 32p.
11. Social hislory and industrial classification (SHIC): a subject
classification for museum collections. [Shefficldl: Published
for Ihe SHIC Working Party by the Centre for English Cultural
Tradition and Language. 1983. 2v.
12. Anglo-American cataloguing rules. 2nd ed., prepared under the
direction of ihc Joint Steering Committee for revision of
AACR. London: Library Association Publishing. 1988. 677p.
ISBN 0-85365-598-7.
Leonard Will is Head of Library and Information Serx'ices
at the Science Museum, London, and is Chairman of the
MDA Working Group on Terminology.
The Indexer Vol. 18 No. 3 April 1993