The Yearbook of International Organizations (Volume 1) in 1996 describes or lists over 28,036 bodies which can in some way be considered international organizations. Whether governmental or nongovernmental, their activities interweave in a myriad ways in the processes of the international community. Although organizations are listed in alphabetical order of titles and abbreviations in Volume 1 of the Yearbook and a multilingual index is provided in this volume, this nevertheless fails to provide an ordered, comprehensible overview of how such activities interrelate. In the absence of any such ordering, tendencies to fragmentation are reinforced and subtler approaches to integration are hindered.

This paper clarifies this challenge and describes the factors entering into the design of the process from which the activity classification in this volume emerged as a first product. It is important to note, as is explained below, that it is unnecessary to read or agree with the contents of this paper in order to derive practical benefit from the classification in its present form. The concern of this paper is to point out other ways of making use of the classification and the possibilities for its further development.

1. Review of classification of organizations in the Yearbook of International Organizations

When the predecessor of the Yearbook of International Organizations was first produced in 1910, the organizations were classified in it according to the Universal Decimal Classification (UDC) system. This was to be expected given that the person co-responsible for both the UDC and for the organizations publishing the Annuaire de la Vie Internationale (as it was first known) was Paul Otlet, often referred to as the 'father of international documentation'. In 1910, on the occasion of the 1st World Congress of International Associations, he produced a 'Tableau de l'Organisation Internationale' grouping organizations (and conferences) by subject area. An improved version of this was produced in 1924 by him, on the occasion of the 4th Conference of International Associations, covering some 400 international bodies with comments on their activities (2). The practice of using the UDC for classifying international organizations in its archives was in fact continued up until 1960 by the Union of International Associations.

The use of the UDC proved however to be too cumbersome for the organization of the Yearbook of International Organizations after its resuscitation in 1949. Between 1951 and 1965 (10th edition), organizations were grouped into some 20 subject chapters and allocated a simple filing number for indexing purposes. The number changed from edition to edition as a result of additions. Intergovernmental bodies were grouped in a separate non-subject chapter. This system proved progressively less satisfactory due to the emergence of organizations which could be usefully allocated to more than one subject chapter.

In the 11th and 12th editions the organizations were ordered alphabetically in an encyclopedia format. A systematic permanent numbering system was maintained in parallel as a development of the earlier subject division. The approach created filing problems so that, in anticipation of the conversion to computer processing, organizations were given a permanent filing number from the 13th edition (1970-71). The subject-based numbering was abandoned from the 14th edition. The original subject 'chapter' division was however maintained, with some additions, until 1980, in order to ensure statistical continuity. But from the 15th edition (1974) such statistics proved increasingly suspect due to the problem of overlap between categories and despite the introduction of 'secondary' classifications. It was recognized that a totally different approach would have to be used.

2. Review of other approaches to international organization classification

The Union of International Associations is obviously not the only body faced with the problem of classifying international organization activities. In searching for better approaches it is therefore important to take into account other initiatives, even if their focus is not solely concerned with international organizations.

Of greatest potential value is the Macrothesaurus; a basic list of economic and social development terms (3). This was first published in 1972 by the Organisation for Economic Cooperation and Development (OECD) in collaboration with other bodies, including the FAO and the ILO. Unfortunately its value is limited by the range of subjects indicated by its sub-title. It is however adapted to computer processing and exists in a multi-lingual version. It uses a 6-digit subject code.

Also of great potential value is the initiative of the International Federation for Documentation (FID), under contract to UNESCO, to design a Broad System of Ordering (4). This is known as BSO and reflects FID's experience as the agency responsible for the UDC. The BSO was intended as the basis for interconnecting information systems within the framework of the Intergovernmental Programme for Co-operation in the Field of Scientific and Technological Information (UNISIST). The most recent draft was published in 1978. It has met with severe criticism and is not particularly well-designed for computer processing. In addition, as might be expected from the priorities of UNISIST, the range of subjects does not respond to the detail or variety encountered in the Yearbook of International Organizations.

Simpler in many respects, and therefore of greater practical value, is the inter-organizational exercise within the United Nations system carried out by the Inter-Organization Board for Information Systems (IOB) with the approval of the Administrative Committee on Coordination (ACC). This resulted in the production of a list of Broad Terms for United Nations Programmes and Activities in 1979 (5). The 2,500 terms are grouped in 16 activity divisions defined at this stage by a 3-digit code permitting further development. The difficulty here is that the system does not appear to have been further developed and does not yet respond to the variety encountered in the Yearbook, especially as reflected in the concerns of nongovernmental organizations.

Also of great interest as a practical approach is the technique used by the publishers of commercial subject directories for multi-lingual users. An example is the 'yellow page' directory produced for Belgian telephone subscribers. Subjects are given a 4-digit numeric code which does not however have any classificatory significance except to provide a numeric sequence. Separate indexes in English, Flemish, French and German enable users to locate each subject.

It is significant that none of the above initiatives is especially concerned with the pattern of relationships between activities or subjects. The allocation of numbers to activities is basically arbitrary. The project of Ingetraut Dahlberg, Editor of the journal International Classification, resulting in the production in 1981 of an Information Coding Classification (ICC) system (6), therefore merits special attention in a following section. One of its advantages is the use of a 4-digit code. But one difficulty in relation to this project is that the schedule of terms has so far only been published for 3-digits, raising problems in handling other topics with which international bodies are concerned.

3. Possibilities of an alternative approach

Serious attempts were made to use several of the above schemes in the period 1979-81, either singly or in combination. For a variety of reasons they proved impractical. The decision was therefore made to design a new scheme adapted to the specific problem of handling international organizations and their activities.

Once this decision was taken it created the opportunity of responding to many of the less apparent constraints encountered when attempting to use the above general schemes. These have been discussed in a separate paper on anti-developmental biases in thesaurus design (17), on the occasion of a conference initiated by the Committee on Conceptual and Terminological Analysis (COCTA).

4. Preliminary design considerations

The point of departure was the system, mentioned above, developed by Ingetraut Dahlberg, following proposals first made by her in 1971. The general outline of her ICC scheme may be seen from Figure 1.

The following features of the scheme are of special interest:

• a) It is based on a concern for 'man's ability to perceive the world, and to construct a system of knowledge units to facilitate his understanding of the world and communication about its nature' (8, p.6).
  
• b) It recognizes that the 'structuring of man's knowledge about the world may be seen as being related to the optical levels of general, world-immanent objects by an evolutionary sequence which, however, is of a spiralling rather than of a linear nature' (8, p.7).
  
• c) It is ordered vertically in terms of 9 optical levels associated with a progressive complexication of perceived reality:
• Pure forms and structures (magnitudes, proportions)
• Pure matter and energy (atoms, forces, etc)
• Aggregated matter in motion (cosmic bodies)
• Animated, non-intelligent beings (micro-organisms, plants, animals)
• Animated, intelligent beings (individual human beings)
• Aggregated, intelligent beings (human societies)
• Material products (goods and services)
• Intellectual products (documents, information)
• Spiritual products (language, works of art and other meta physical works (8, p.35). These are distinguished by the first digit of the ICC code.
  
• d) It is ordered horizontally from the non-fundamental disciplines at each level (on the left) to those concerned with application of that knowledge (on the right). These are distinguished by the second digit of the ICC code.
  
• e) Within any area of the resulting matrix, a structured sequence for the system positions was applied for the repeatable arrangement of the elements of each group. These are defined as follows:
• 1. General and theoretical statements (axioms, etc)
• 2. Object-related statements (elements of objects, parts, kinds of object, etc)
• 3. Activity-related statements (states and processes in objects, operations applied to them, etc)
• 4,5,6. Statements related to specialities of the objects and/or activities concerned in 2 and 3
• 7. Statements on influences onto 2 and 3 from outside ('instru mental', technical relationship)
• 8. Statements on the use of 2 and 3 in other fields ('potential', resource orientation, application relationship)
• 9. Statements of the knowledge about 2 and 3 in distributing it by human beings, societies, documents, etc ('actualization', synthesizing, environmental relationship). These are distinguished by the third digit of the ICC code.

Dahlberg has elaborated, published and applied the scheme (6) using three digits (some 700 classes) and hopes to publish a more extended four-digit version (some 7000 classes) in the future (9).

If the four-digit version had been available it is probable that it would have been used to design the coding system for international organizations. In experimenting with the various possibilities however it became apparent that there was a basic awkwardness and bias in making all the preoccupations of such bodies subservient to 'knowledge' of 'objects'. This problem is particularly striking when a social reality like 'homelessness' is classified under an intellectual discipline, namely 'sociology', as in the case of the UNESCO Thesaurus (10). Similarly a value and condition of fundamental importance like 'peace' is classified under an intellectual discipline such as 'political science', or, again, 'friendship', 'love' and 'hatred' are classified under 'psychology'. Positioning values, conditions and forms of praxis in this way can be seen as reinforcing the dominance of the knowledge function during a period when the international community recognizes a need to enhance action, the 'will to change', as well as the emergence of new values. Many organizations perceive themselves as concerned with praxis and do not relate directly to the intellectual disciplines by which their actions are supposedly governed according to university faculties.

In the light of the ICC scheme the question then became one of de-emphasizing this bias in favour of knowledge, whilst at the sametime respecting the concerns reflected in the ordering of the matrix. One criterion of an interesting matrix, for example, would be the possibility of mapping onto it at different locations the various agencies and institutions required for the 'operation' of a country or the world (eg various government ministries, hospitals, factories, farms, airports, military bases, etc). In this way the matrix would become a tool reflecting operational reality to a greater degree, rather than responding primarily to the difficulties of designing information retrieval systems to facilitate research and the generation of further knowledge.

Another valuable feature of such a matrix would result from ensuring that it told a developmental 'story'. This feature is to some extent present in the ICC matrix in that the 'lower' optical levels reflect the earlier phases in an evolutionary process, whilst the 'higher' levels reflect the relatively recent phases of civilization. But it is possible that a more interesting developmental story (or stories) could be embedded in the structure of the matrix. This would be especially valuable if it highlighted the stages at which different functions emerged in society (eg social organization, mutual care, shelter, artifact construction, etc). As argued in an earlier paper (7), this implies a dynamic emphasis on processes in contrast to the conventional static emphasis in classification schemes on states and objects. A number of authors are now arguing against the insidious effects of static (Euclidean, Newtonian, Cartesian) descriptions of reality as favoured by the 'Western' mode of thought (11, 12, 13). It can certainly be argued that this emphasis undermines a dynamic approach to development (7).

Clearly the above features would emphasize the 'interweaving' of the cells of the matrix. This approach is to be contrasted with the practice adopted in the design of many thesauri. So little attention is devoted to the relationship between major classes that it is easy to get the impression that any such relationship is totally arbitrary - isolated subject clusters ('science', 'religion', 'art', 'commerce', etc) denoted by digits from 1 to 9, etc. The 'lumping' of major classes together in this way does not appear to have changed significantly throughout the history of classification schemes from 1200 BC to the recent initiatives of the intergovernmental community (14). It is not difficult to argue that it is this arbitrariness which deprives the pattern of classes of any significance as a whole. As such it reinforces the fragmentation of society which many authors have deplored, as well as undermining any efforts towards an 'integrated', 'interdisciplinary' or 'holistic' pattern of action (11).

5. Insights from periodic classification

As a guide to further insights for the design of a more interesting solution, what appeared to be required was some matrix-type model incorporating developmental features reflecting the emergence of a series of qualities organized into corresponding 'groups' at a succession of 'levels'. The richest conceptual scheme of this kind appears to be the periodic table of (chemical) elements. The possibility of generalizing this periodic system seems first to have been explored by Edward Haskell (15). Inherent in such a scheme are many interdependency relationships. Furthermore, in comparing J W van Spronsen's history of the development of the periodic classification system (16) with that of Samurin's history of the development of the classification of knowledge in general (17), it is possible to conclude that a scheme such as that of Dahlberg corresponds in structure to the penultimate development phase prior to the emergence of the fully fledged periodic system. Many conventional classification schemes correspond however to much earlier phases in this development with only rudimentary relationships between major classes.In considering the possibility of such a fully-fledged periodic system, it is useful to bear in mind the following remark by A J Ihde in the foreword to van Spronsen's survey:

'Facts soon reach a point where they become less and less manageable unless an attractive and meaningful system of classification is brought into being... Equally important is the role of tools in science...

It is frequently not recognized that tools may be conceptual as well as physical... The Periodic System has fulfilled both of these roles. It has served as a classificatory device but it has contributed much more than mere classification. It has been a conceptual tool which has predicted new elements, predicted unrecognized relationships, served as a corrective device, and fulfilled a unique role as a memory and organization device. The periodic table has contained an innate flexibility which has prevented it from becoming frozen into a rigid structure. It lends itself to a large variety of forms.Although many of these are unique only as schemes representative of the author's originality, certain forms have unique value in bringing out particular relationships.' (6, p.ix).

On this last point it is striking to compare the range of experiments with spirals, tables, circles, cones, cylinders and other figures (see Figure 2) in portraying the classification of elements (16) against the seemingly universal preoccupation with simply structured lists in the case of the classification of knowledge (14). In this sense the Dahlberg scheme is indeed an exception. To clarify the discussion it is useful to note how one frequent form of the periodic table (Figure 3) can also be presented in another way (Figure 4) which resembles more closely Dahlberg's ICC scheme. The 'groups' of chemical elements then tend to appear in columns, analogous to those denoted by the ICC second digit. The transformation from Figure 3 to Figure 4 clarifies the distinction between two 'sub-groups'. This is even clearer in a circular form of the table (Figure 5)