1. Matrix (1983-85)

As stressed above the classification is designed to be modified. For the 1983-85 editions the design can best be described in terms of the classification matrix and specifically in terms of the definition of its "semantic cells" on the basis of the levels (rows) and columns in Figure 7. As pointed out above, the choice of levels, columns and cells resulted from an iterative process aimed at ensuring an interesting and functionally meaningful balance within the pattern as a whole. As pointed out in an earlier paper (26), this process could perhaps be best described as analogous to "tuning" a "semantic piano". Clearly different tuning systems are possible, none of which balances qualities in a totally satisfactory manner. This problem of balance and tuning has been highlighted elsewhere (27, 28) in an attempt to reconcile the qualitative results, from different cultures, of exercises in classifying the same range of attributes into 1,2,3,... or N categories (see also ref 68). The situation here is of course complicated by the heterogeneous nature of this approach.

It is for this reason that the level and column headings should be considered as tentative indications of dimensions that it seemed valuable to distinguish. Similarly the terms attached to individual "semantic cells" of the matrix have been selected as recognizable common terms indicative of some percentage of the significance to be associated with each such cell. At this stage no attempt has been made to modify the discipline-oriented terms conventionally used for many common subject areas. The consequence is that the lower half of the table, at least, has familiarly named major classes.

1.1. General structure

Before discussing the structure in more detail it is appropriate to note a general structuring device which has been used. As pointed out above in discussing the work of Jantsch, there would appear to be value in attempting to "capture" some aspects of the alternation on which the organization of self-organizing systems is based. It provides a means of acknowledging the functional reality of the operational hostility frequently experienced by those concerned with "mutually irrelevant" functional domains (eg science, religion, commerce), whether as expressed in relations between international agencies, between university faculties or between their departments (28). This phenomenon is also reflected in the manner in which categories may be conceived or perceived. For example, current investigations are demonstrating the complementary roles of modes of perception associated with the right and left hemispheres of the brain (29). These can be linked to such dichotomies asqualitative/quantitative, art/science, "soft"/"hard", image/text, context/structure and process/strasis.

Levels 0-7 and columns 1-8 have therefore been organized in an alternating manner to reflect the extremes of these dichotomies. In levels 8-9 and column 9, areas are reserved for the "transcendence" of these dichotomies. Where the row and column dichotomies do however "interact", cells of three distinct types are defined: quantitative/quantitative (crossed cells), qualitative/qualitative (circled cells), and quantitative/qualitative (unmarked cells). The resulting pattern is an interesting first approximation, especially when the crossed cells are seen as primarily associated with well-defined categories and text, whilst the circled cells are seen as primarily associated with fluidly defined categories and imagery. The unmarked cells are then associated with a blend of art and science (eg design, artefact production by industry, or technology as a useful art).

1.2. Level (row) structure

The levels may first be considered in pairs as: nature (0-1), praxis (2-3), theory (4-5), developmental principles (6-7), and existential experience (8-9). Here "nature" is split into the physical sciences and the biosphere. "Theory" and "developmental principles" may also be grouped as the "noosphere". This approach has the merit of preventing innovative change and development from being obscured and denatured by including them under descriptive sciences and theory. It also provides space for the values and experiential conditions in the name of which change is proposed and implemented, rather that disguising them as the subject matter of psychology or philosophy.

Levels 0-3 constitute the "material world", its description, and the more concrete forms of action in society. Levels 4-7 provide space for the reflections and interpretations of those acting in the "material world", whereas levels 8-9 provide space for experience in its own right. As such it is the least tangible but the most intimate, figuring in much of the current debate on human values and non-material human needs (30).

Levels 0-1 are also associated with the natural sciences and as such figure prominently in university departments. Levels 2-3 may be directly related to government agencies, public services and institutions. Levels 4-5 correspond to the "soft sciences" whose subject matter tends to be defined rather than given. It is at levels 6-7 that new directions of social change are defined.

The reservation of cells in levels 8-9 for values and conditions of awareness must necessarily be considered extremely tentative given the lack of attention to the problems of classifying experiences in their own right.

1.3. Column structure

The columns may also be first considered in pairs in terms of a possible set of (social) patterning implications: establishment and consolidation (1-2), maintenance and appreciation (3-4), adaptation and propagation (5-6), and innovation and exploitation (7-8). Column 9 is concerned with the resulting symmetry or imbalance. Column 0 is used for formal concepts calling for qualifiers.

Taken singly the columns may be tentatively described as follows: domain definition (1), organized relations (2), differentiated order (3), contextual renewal (4), controlled movement (5), communication reinforcement (6), redistribution of resources (7), and environmental manipulation (8). Needless to say, such descriptions are indicative rather than exhaustive.

1.4. Individual cells

In the matrix some cells call for special comment:

• Fundamental sciences (00): The special situation of the "fundamental sciences", including mathematics, physics and chemistry, can be usefully modelled by the peculiar situation at the beginning of the periodic table of chemical elements (see Figure 3 and Figure 4). Without stressing the resemblance, special status has been given here to "00" treating it as a kind of formal precondition or "pre-level". Within it are to be found the fundamentals of relationship (as partly reflected in mathematics) and of matter (as partly reflected in physics and chemistry).
• Society (21): This is distinguished from sociology (41) to separate the function of reflection about society from the entities acting within society. Note that such a separation is not called for with respect to levels 0 and 1, in which the subjects of attention are to a much greater extent taken as given.
• Health care (32): Treatment in general, and its necessary infrastructure, is distinguished from the analysis of disease under medicine (18).
• Societal problems (29): This is used to group problems of imbalance in the functioning of society, including crimes and disasters. It also includes preventive measures such as safety and hygiene.
• Science (51): This includes science in the broadest sense (eg humanistic sciences) as well as science as a phenomenon in its own right (science of science) and as such is appropriately distinguished from the natural sciences (levels 0-1).

1.5. "Harmonic" relationships

Given the alternation of levels, although semantic cells in the same column have qualities in common, the relationships between those in odd-numbered levels or in even-numbered levels is stronger. This is most evident from the second level. For example: religious practice (36), theology (56), morals and ethics (76), transcendence (96). This series clearly goes from tangible manifestations of religion, through associated beliefs, to transcendent experience. In this sense the cells in the higher levels bear a "harmonic" relationship to those in the lower ones.

1.6. Complementary relationships

The set of cells at any given level can be seen as representing functional complementaries. The expression of one in a society calls for the expression of the others to complement or counter-balance that function. This is most clearly seem at levels 2 and 3. Each function would seem to be necessary for the society to be viable at that level, whatever the views of those who identify with the categories of a particular function. For example, despite the words chosen to label cells 31 ("research, standards") and 36 ("religious practice"), a society will engender an investigative, "normalizing" function, as well as some measure of ceremonial, even if it is only to celebrate secular values.

1.7. "Empty" cells

It is important to emphasize that although most cells have words associated with them, those words may only signify a very small percentage of the meaning that could come to be associated with the cell. In this sense many of the cells are effectively "empty", especially those at the less tangible levels (from 4-9), as well as those in column 9.

Given the importance attached to guidance from the historical development of the periodic table, empty cells can be welcomed as a provocative challenge. Cell 52 may be expected to correspond (in part) to the theory of health care and health improvement (as opposed to the treatment of disease), given that cell 72 corresponds (in part) to sensitivity training of an individual in groups. Cell 62 may be expected to correspond to enlightened management techniques in which the worker-employer relationships is redefined in a new form of partnership. Cell 78 may be expected to correspond (in part) to enlightened forms of agriculture, such as organic farming and similar experiments.

2. Initial word coding (1983)

Using the computer-extracted word list (at that time containing 11,000 items) a procedure was adopted for the 1983 edition whereby significant words were first given a 2-digit code according to Figure 7 [Integratrive Matrix]. After resorting, these words could then be regrouped within each semantic cell taking into account their frequency of occurrence, namely the number of organizations with a particular preoccupation. Sub-classes were thus created using a 3rd and 4th digit, resorting whenever a new overview of the result was required. In the final run some 7,500 words were used, the remainder being proper names or insignificant, or else having too many distinct connotations (polysemes).

Clearly allocating a single code to a word assumes that a given word cannot be associated with several semantic cells. This is certainly not true in the case of homonyms. But given the practical orientation of this project, the question is whether it produces a useful result in a sufficiently high percentage of cases. If obvious mismatches do result but the user can easily eliminate them by visual inspection, then the approach continues to have merit. Such mismatches are after all inherent in current word-oriented information retrieval systems in which the user is obliged to filter what is supplied. The allocation of codes to single words (eg "economic" or "development") rather than to multi-word terms (eg "economic development") has advantages and disadvantages. It increases the problem noted in the previous paragraph, since qualifiers reduce the incidence of mismatches. It does however create the possibility of highlighting links between distinct semantic cells, namely interdisciplinary or inter-functional links. This is discussed further below.

Recalling the periodic table once again, a basic difficulty in elaborating it was that for centuries the distinction between a chemical element and a compound of several chemical elements could not always be clearly established. As a result attempts were made to classify compounds on the basis of their properties which were thus confused with those of chemical elements. It was consequently difficult to produce a meaningful table. It would seem that a very similar problem exists in elaborating a classification scheme for the societal functions with which international organizations are associated. There is a need to distinguish between those which can be considered as "elements" and those which should be treated as "compounds" of such functional elements. Although only in the first stages of its elaboration, this is the intended distinction between Section W and Section X.

The question is obviously not just one of considering single words as denoting such functional elements. Many such single words, even when they are not compounds in their own right (eg psychosocial), should appear only in Section X because of their multi-functional nature. Further attention will continue to be given to this problem in future editions.

At this stage there is clearly a problem in determining whether it is more appropriate to associate the word "art", for example, with its manifestations in performances and works (eg level 3), the intellectual study of it (eg level 5), its transformative function (eg level 7), or the experience of artistic creativity (eg level 9). These of course bear a harmonic relationship to one another as discussed above, but the word "art" may be used indiscriminately to describe all of them. Indeed it could be considered a compound of them which could be more appropriately located within the framework of Section X. Such distinctions have been made wherever possible in order to highlight the functional significance of levels 6 to 9. A balance was however sought between emphasizing this harmonic spread and locating the word where the user might expect to find it in the light of past classification schemes.

In contrast to Dahlberg's approach discussed earlier, the organization of the word coding within the semantic cells is not systematic at this stage. Grouping within the cells has often been done on the basis of word frequency. In the case of levels 8 and 9, no grouping has been attempted within the cells. It is hoped that for a future edition the organization within the 2-digit cells can be made analogous to that within the matrix as a whole. This would, for example, make it possible to distinguish at levels 8 or 9 between values and experience which are more concrete and those which are more transcendental.

3. Initial section generation (1983-85)

Once the word coding was complete for what then constituted a computer-based thesaurus, this was used to determine with which codes each organization should be associated. Here a distinction must be made between the three main sections of this volume (W, X, and Y).

Where only one code was associated with an organization, the latter was allocated to Section W, X or Y, depending upon the letter associated with that code in the thesaurus. Where several codes were associated with an organization, that organization was allocated to all the corresponding categories in the volume. In addition the codes were combined to select "interfunctional" preoccupations with which the organizations could be associated in the categories of Section X. The "combination" has been done on the following basis. If the codes for the organization all corresponded to the same 2-digit semantic cell, they were not combined. Where the first two digits of codes were different, they were combined. For example: W3310 ("schools") and W3900 ("law") were combined to create the additional codes X3339 ("schools/law") and X3933 ("law/schools"). If in addition the code Y5000 ("Europe") was present, then additional codes Y5039 ("Europe/law") and Y5033 ("Europe/education") were generated. On the basis of these, the organization name/address entry was allocated to the relevant categories in the different sections of this volume.

Where the kinds of keywords in the name of an organization made the above procedure inadequate, was supplemented or by-passed by the traditional form of "manual" allocation of codes directly to the individual organization rather than to words in the thesaurus (eg in the case of "trade unions", or "international relations"). The finalsorting of the individual entries into sections was then completed by computer up to and including the generation of photocomposition pages ready for print.

4. Procedural revision (1984-1985)

The results of the above exercise were reproduced in the first edition of Global Action Networks in 1983. As planned, those results were reviewed as part of the production of subsequent editions.

In 1985 the thesaurus was automatically extended from 11,000 to 13,604 words by incorporating words from names of new international organizations and world problems. Nearly 8,000 words received subject allocation codes placing them in the active part of the thesaurus. The major modification made was to remove limitations in the computer programme used to identify words on the basis of which subject codes were associated with the organization.

Three techniques were used to reduce the percentage of mis-allocations:

• a) Words generating an excessive number of mis-allocations were eliminated from the active part of thesaurus. In 1984 85 words, and in 1985 60 words were treated in this way (respectively 1.2 and 0.7 percent of the active thesaurus).
• b) The mis-allocation of specific subject codes to some selected organizations was inhibited. In 1984 390, and in 1985 504 such allocations were treated in this way (respectively 2 and 6.4 per cent of the active thesaurus).
• c) Allocation of some subject codes to specific organizations was forced where it would not normally have occurred. In 1984 145, and in 1985 545 allocations were forced in this way (respectively 0.7 and 2.6 percent of the organization allocations). In addition some subject groups, not well defined by particular words, were deliberately created in this way (eg "religious orders", "trade unions", "regional studies", "intergovernmental organizations").

The pattern of codes in the matrix was reviewed, and in 1985 37 changes were made.

Aside from the changes noted above, of the 8,000 words in the active thesaurus, code changes were made to 50 of them (0.6 percent) in 1984, and to 201 of them (2.5 percent) in 1985. Some modifications to levels 8 and 9, the most difficult area of the matrix, were made in the light of work on "human values" and "states of consciousness" for the 1986 edition of the Encyclopedia of World Problems and Human Potential (70). Work on other sections of the Encyclopedia was also incorporated, updating the information included in the 1983 edition drawn from the 1976 Yearbook of World Problems and Human Potential.

5. Procedural revision (1986-1987)

Following the 1985 revision two developments resulted in further modifications to the procedures used. The first of these was the transfer of the database onto an in-house local area network with many more software possibilities, in addition to those which were specially designed to facilitate production of this particular publication. In 1986 these procedures were further rationalized and modified to remove system errors. The second development was the completion of the 1986 edition of the Encyclopedia of World Problems and Human Potential, from which information on world problems, strategies and values was drawn in order to clarify their relationships to the fields of international organization activity.

The main modifications were therefore as follows:

• a) Whereas the previous editions have been based on the allocation of subject codes based on single words, the software has since been developed to permit subject coding based on word pairs in an organization title. In 1986, approximately 1,000 such word pairs were added to the thesaurus to avoid the earlier need for manual intervention to resolve such coding problems as "flying saucers" or "brain drain"; in 1987 500 word pairs were added.
• b) Whereas in the previous editions the subject codes allocated for a single organization were based on the title only, the software was developed in 1986 to permit the allocation of codes to specially italicized words text in the "Aims" paragraph of the description in Volume 1. (The resulting entries are flagged in Section W to distinguish them from those arising from the title of the organization.) In 1987 Sections X and Y also included entries generated through allocation of codes via keywords in the "Aims" paragraph.
• c) Because of the cruder procedures employed in the past, many words had been maintained in the inactive portion of the thesau rus to prevent them from generating ambiguous or misleading subject codes. With the above-mentioned procedures these inactive words were reviewed and some 1,000 were activated for the 1986 edition, and several hundred more in 1987.
• d) The new software permitted a much more flexible approach to any necessary fine-tuning of the pattern of allocated codes. A more sophisticated approach was possible to forcing the allocation of codes (whether forcing into a subject area or out of it). With the introduction of word pairs into the thesaurus, fewer such interventions were necessary.
• e) Relatively few changes were made to the central portion of the code matrix (Rows 0 through 6); some 200 words were transferred to new codes. Major changes however were made to Rows 7, 8 and 9, as part of the continuing exploration of ways of incorporating human values and strategies in the light of the work on the 1986 Encyclopedia of World Problems and Human Potential. Given the inherently ambiguous nature of the words associated with those rows of the matrix, a much higher proportion of words was forceably associated with particular codes, partly in an effort to take advantage, at least provisionally, of the groupings which emerged in the Encyclopedia. In particular the 1986 edition explored the use of polarities as categories whereby values or strategies can be grouped. This approach offers some advantages in handling words which are usually an embarrassment to any scheme of subjects, despite the importance attached by society to the concepts to which they may refer. The value and strategy polarities of the Encyclopedia (Sections VP and SP) were therefore used as categories in Rows 7 and 8 in the 1986 edition. In the case of the 2,000 value words from Sections VC and VD of the Encyclopedia, which are grouped under one (but usually more) of these polarities, these were forced into the relevant categories on the basis of the information in the Encyclopedia. This procedure was also adopted for the strategies arising from Sections SS and ST of the Encyclopedia. In this case however the allocation to a polar category was based only on the gerund term used in the strategy to emphasize the activity inherent in it. Any other words resulted in the allocation of the strategy to other subject codes as in the case of organization titles.
• f) In 1987 work continued on rationalizing rows 7, 8 and 9. The value and strategy polarities (Sections VP and SP of the Encyclopedia of World Problems and Human Potential) continued to be used as categories in rows 7 and 8, with value words from Sections VC and VD being grouped under the row 8 polarities and strategies form sections SS and ST under row 7 polarities. To improve transparency, all other row 7 and 8 headings were grouped as W70 or W80 terms; organizations, problems or strategies indexed under such headings therefore appear together, rather than interspersed with strategy and value polarities. In line with the structural philosophy of the matrix, the third digit of the heading code reflects the overall headings of the matrix columns. A number of minor changes were made to the central portion of the code matrix.
• g) The basic organization of the major sections (W,X and Y) of the 1986 volume remained the same. In 1986 Section Z was added to group types of organizations. In the few cases where such types were already allocated to particular codes, it was appropri ate to move them to this new section.
• h) The Citation index, previously published as an Appendix, was omitted in 1987. This is because a revision of Volume 1 has led to cited organizations being listed, together with their Yearbook reference numbers, in the descriptions of organizations citing them.

In 1986 the thesaurus contained 20,000 words, including 1,000 word pairs. Of these some 9,000 words were inactive in that they did not result in the allocation of codes leading to the incorporation of an entry at some place in this volume. The thesaurus was therefore increased by 6396 words since the 1985 edition. In 1987 the thesaurus contained over 23,000 words, including 1,500 word pairs. Of these some 11,000 words were inactive. Overall, there was an increase of 3,000 index terms since the 1986 edition.

6. Procedural revision (1988)

The main innovation this year was the inclusion of a proportion of items from the Human Development section (Section H) of the Encyclopedia of World Problems and Human Potential (1986, second edition), as a result of work for the forthcoming third edition of the Encyclopedia. This section is in two parts - human development concepts (sub-section HH) and modes of awareness (sub-section HM). The parts have been treated differently. The 635 items from the human development concepts part have beenincluded in Section W, the subject section, being treated in an exactly similar manner as the problems. They have not, however, been included in the subject combination section, Section X; and, not being geographical in nature, they have not been included in Section Y. Of the nearly 1,000 modes of awareness, over 300 have been coded to appear in the top row of the subject matrix, that concerned with awareness. The codes have been assigned on the basis of the philosophy behind the choice of matrix columns as described above. Row 9 of the matrix has been rationalized along the lines adopted for rows 7 and 8 in 1987. That is to say, the modes of awareness are grouped under row 9 polarities 1 to 9, while organizations, problems, strategies and human development concepts otherwise appearing in row 9 are grouped together in column 0, the third digit of the heading code reflecting the overall headings of the matrix columns.

The modes so far included are those which are most readily considered individually. Work continues on grouping the remaining modes, which already form part of well recognized series. The task is to arrange the series in a meaningful manner within the constraints of the matrix structure.

Other modifications in 1988 have focused on further refining the thesaurus. Among the 2,500 "active" terms included since the last edition are a number of abbreviations, proper names and non-English words. These are not included in the subject index at the beginning of the volume, as they are not considered helpful in this respect; however, they do increase the accuracy of coding and should improve the quality of information offered. An additional 1,500 completely new terms have been added to the thesaurus in 1988, and 225 paired terms. The thesaurus now includes nearly 24,500 terms, of which 14,500 are active. Further minor changes have also been made to the matrix; and Section Z now lists international conference series and international agreements and treaties.

7. Procedural revision (1989)

A massive expansion of the thesaurus has been undertaken, with the inclusion of terms from the text of organization and problems descriptions. There are now over 51,500 terms in the thesaurus, of which some 22,500 are active. A particular innovation is the additional inclusion of major subject terms in other languages - notably French, Spanish, German and Russian - with a view to enhancing this volume as an index to Volume 1. The intention is to make Volume 3 a more specific means of access to organizations, replacing some of the functions of the index to Volume 1 which has, in consequence, been streamlined.

As an additional cross-referencing tool and to assist in identifying organizations whose exact title is unknown, the subject headings and index words in Volume 3 are now also indicated in the index to Volume 1, the former in English, French, Spanish, German and Russian.

As an aid in tracing terms in Russian, these terms are included under two different systems of transliteration. Another aid to tracing organizations is the entry Z6300 in Z section, which lists organizations whose titles contain proper names.

Since work is still in progress on the new edition of the Encyclopedia of World Problems and Human Potential, it is not considered useful this year to include the entries from the Human Development Section, nor the Strategies or Values. This has allowed sufficient space for items from Section G of Volume 1 (internationally oriented national bodies) to be included in Section X of Volume 3 for the first time in several years.

The world problems included this time are those for which it is expected a full description will be published in the new edition. The section code for the world problems (second letter of the reference number) is the result of preliminary work on classifying the problems according to their nature and to the breadth of their coverage.The order in which entries are printed under the subject headings has been altered. Previously, entries were quoted alphabetically within the sections in which they appeared in Volume 1 or in the Encyclopedia.

The order has now been simplified so that entries appear within alphabetical order of title independent of the section in which they appear. However, organizations still appear before entries from the Encyclopedia. Their reference numbers are preceded by a figure 1, while Encyclopedia entries are preceded by a figure 4.

8. Procedural revision (1990-1996)

Procedural revision (1990)

As well as a further large increase in the active terms in the thesaurus, and matching increases in paired terms, a major innovation isthe further expansion of the index. This now includes not only thesaurus terms relating to Volume 3 but a keyword index (in English and French) to organizations in Volume 1. The index of Volume 3 (this volume) is therefore a means of direct entry to the organizations in Volume 1. In order to allow space for this longer index, Section X has been amended. In previous editions, the largest and smallest "mirror-image" entries were omitted. In this edition, all "mirror-image" entries are omitted, although headings are included with an indication of where to find the information (see the introduction to Section X). Again, for space reasons, the strategies, values and human development sections of the Encyclopedia of World Problems and Human Potential have not been included. The thesaurus now contains 53,434 terms, of which 24,405 are active, and includes 2,325 paired terms.

Procedural revision (1991)

The main effort this year has been the manual forcing in/out of entries under the various subject codes where ambiguities of meaning (often arising from non-English titles) have caused mis-coding. In addition, the expanded index (again including keywords from rotated English and French titles) has been made more easy to use. As was the case for 1990, "mirror image" entries in Section X have been omitted, but headings are included, together with an indication of where to find the information. The thesaurus has increased by 2,327 terms to a total of 55,761. Of these, 25,131 are active, an increase of 726. There are 2,627 paired terms.

Procedural revision (1992)

As indicated in Appendix 1, "New Features 1992/1993", all three volumes of the Yearbook have been rearranged to simplify access. They now list all entries in alphabetical order. The previous sections W, Y and Z of Volume 3 have been combined into one continuous alphabetic sequence by major subject/regional categories, subdivided by suject categories also in alphabetical order, with interleaved subject indicators directing the user to the entry required. Volume 1, now also being in alphabetical order, is considered an index in itself, but detailed index entries (alternative and former titles/initials, executive offcers' names) now appear in the index to this Volume. This index now becomes the main index for the whole Yearbook. To allow space for this expanded index, Section X: subject combinations (already much reduced in previous years) has been omitted entirely from this edition. The thesaurus has increased by 1,252 terms to a total of 57,013. Of these 26,915 are active, an increase of 1,784. There are 2,884 paired terms, an increase of 457 over last year.

Note that statistical tables (Appendix 3) still refer to the old W, Y and Z sections and may be compared with statistics from previous years. Statistics for section X are not published this year.

Procedural revision (1993)

Building upon the changes of last year, several items have been rationalized to avoid repetition. New items have been added, including a list of UNESCO bodies (under Culture/UNESCO Bodies). Considerable effort has gone into improving the multilingual character of the thesaurus used for coding and indexing. This has resulted in an increase of 14,851 words over last year (71,864 words altogether). Work is continuing on coding these new words. Currently there are 26,515 non-English or English/other languague words and word-pairs in the thesaurus, of which 11,740 have subject codes. In total there are 31,539 active words (an increase of 4,624 over last year). There are 3,222 word-pairs (an increase of 338). Another innovation is the inclusion of founding personalities in the index. This work is also continuing. Eliminated from the index are the thesaurus words not occurring as title or keywords in organizations or world problems.

Procedural revision (1994)

The main emphasis this year has again been the building up of the thesaurus with non-English words occuring in organization titles and descriptions. A further 3,859 such words have been added, 14.6% more overall than the already expanded thesaurus of 1993. There are now 30,374 non-English or multiple-language words, of which 13,578 have meaningful codes (a 15.7% increase over 1993). In all there are 82,682 words in the thesaurus, including 3,370 word pairs, 34,030 of the words having meaningful codes. A major change in 1994 is the removal of items from the Encyclopedia of World Problems and Human Potential. These will now appear in their ownsubject index as part of Volume 3 of the new edition of theEncyclopedia, this volume being entitled Actions - Strategies -Solutions and planned to be published in 1995. The removal of these items has provided the opportunity, through increased space, to expand the index (now 38,554 entries), which for the first time in several years includes reference to the countries in Volume 2 of the Yearbook as well as items from currently inactive and from newly founded organizations. It has also allowed inclusion in the main body of the volume of, for example, listings of individual membership bodies.

Procedural revision (1995)

The thesaurus has been augmented considerably this year, largely from the new strategies volume of the Encyclopedia of World Problems and Human Potential. There are currently 87,075 words, an increase of 5.3% on last year, and this is expected to increase further in the near future, both from the strategies volume and from the new French language edition of the Yearbook to appear shortly on CD-Rom. Work is going ahead on coding these words. There are currently 34,646 words with codes leading to the subject index, 13,798 of which are in a language other than English. Amendments to the subject classification itself include a new heading "Law/Prohibition" and the splitting of "Societal Problems/Safety" to include additional items "Societal Problems/Prevention" and "Societal Problems/Accidents".

Procedural revision (1996)

A significant further increase in the thesaurus took place this year in conjunction with the preparation of a French-language version of the Yearbook on CD-Rom. There are currently 98,166 words in the thesaurus, an increase of 12.7% on last year. Of these, 39,575 are non-English and/or multiple language words, an increase of 26.1% in one year. Nearly 2,000 words were subject coded this year, with 36,630 now leading to the subject index - an increase of 5.7%. Much of the increase was in the French and other non-English words, 15,279 of these now leading to the index (1,581 coded during the year). Amendments to the classification include a new heading "Exchanges" under "Communication" (largely transferring from "Social Activity/Friendship", a new heading "Victims" under "Societal Problems", the transfer from "Geography/Resources" to the main heading "Resources" of much that relates to natural resources; and the rationalization of several organization "types" to appear within the relevant subject areas (notably "Exile Bodies", which are now listed under "Society/Exiles", together with other organizations dealing with the subject which were previously listed under "Society/Migration").