The Evolution of Chain Indexing: Ranganathan’s Innovation

Picture a library in the 1930s. A new book arrives, and a cataloguer has to decide which subject words to attach to it so that future readers can find it. There was no consistent rulebook for this. Two cataloguers could look at the same book and assign completely different subject headings, sending readers down different paths. S.R. Ranganathan, the mathematician-turned-librarian known as the father of library science, saw this problem clearly and built a logical solution around it. That solution was chain indexing, and it reshaped how subject access to documents was approached for decades.

The origins of chain indexing
Where the idea came from
How it changed subject cataloguing
Sought, unsought, and false links
An economic and efficient approach
Not as automatic as it looks
Adoption by national bibliographies
Comparison with other indexing methods

The origins of chain indexing

Chain indexing, also called chain procedure, is a method for deriving subject index entries or subject headings directly from the class number of a document. Ranganathan first formally described it in his 1938 work Theory of Library Catalogue. The core idea was elegant: instead of treating classification and subject cataloguing as two separate jobs, treat them as two sides of the same coin.

Here is the logic. When a classifier assigns a class number to a book, they are translating the book’s subject into an artificial language of ordered digits. Each digit or group of digits stands for an idea, and these ideas are linked together in sequence, forming a “chain.” Ranganathan reasoned that this same chain could be read in reverse and translated back into ordinary words to produce subject headings. Because the chain is used to derive subject entries through a set of rules, he named the system chain procedure.

The first practical use came even before the formal description. Chain indexing was first applied at the Madras University Library in 1936, where Ranganathan was working. Although it was originally designed to work alongside his own Colon Classification, the method can be applied to any classification scheme whose notation follows a hierarchical pattern.

Where the idea came from

Ranganathan did not invent the concept entirely from nothing. He drew the seed of the idea from Dewey’s Relative Index. Melvil Dewey had used a similar technique to connect index terms to class numbers in the Dewey Decimal Classification, but he never gave the method a name or developed it into a structured theory. Ranganathan took that scattered practice, studied it deeply, and turned it into a systematic, rule-based procedure that could be taught and applied consistently.

How it changed subject cataloguing

Before chain indexing, libraries leaned heavily on ready-made lists of standardised subject headings. The pioneers among these were the ALA List of Subject Headings, the voluminous Library of Congress Subject Headings (LCSH) for large research libraries, and the Sears List of Subject Headings (SLSH), which appeared in 1923 for smaller collections. These lists brought some uniformity, but Ranganathan considered them illogical and ineffective in important ways.

His criticisms were specific. The lists mostly offered a choice among synonymous terms, current concepts were often missing, and there were no uniform rules for selecting the most specific heading. Identical material could end up under different headings, leading cataloguers and users down inconsistent paths. The lists were also heavily biased towards the country and culture in which they were created, which forced indexers elsewhere to make constant adaptations.

Chain indexing replaced this guesswork with a logical sequence. Because subject headings are derived mechanically from a class number that has already been worked out through classification, the choice of heading is no longer an arbitrary decision. The class number does the analytical heavy lifting, and the cataloguer simply reads it backwards to produce the verbal entries. This shift, from random and subjective heading assignment to systematic derivation, is the heart of what chain indexing changed.

Sought, unsought, and false links

Not every link in a chain becomes a subject heading. The procedure recognises different kinds of links, and only some are used. A sought link is a substantive heading that users are likely to look for, so it earns an index entry. An unsought link is a heading that exists in the chain but is unlikely to be searched, often because little material is published under it. A false link is represented by a non-semantic digit, such as a punctuation mark or a connecting symbol, which signals the type of facet to follow rather than a subject in its own right. Sorting links into these categories is what keeps the catalogue free of useless entries while still capturing every meaningful access point.

An economic and efficient approach

One of the strongest arguments for chain indexing has always been cost. Subject heading lists like LCSH and SLSH are expensive publications that are revised frequently to incorporate new subjects, which means libraries must keep buying updated editions. Chain indexing removes this dependency entirely. Because headings are generated from the class number that the library already assigns, no separate purchased list is required at all.

The savings extend to staff time. In chain procedure, the cataloguer starts where the classifier left off and works from the same class number. This avoids duplication of effort, since the document does not need to be analysed twice. Chain indexing therefore saves the time of the indexer by reusing the analytical work that has already gone into the class number. The procedure mechanises a large part of the choice and rendering of subject headings, which makes the work simpler and faster while reducing arbitrary decisions.

This combination of low cost and high efficiency made the method especially attractive to libraries and bibliographic agencies working with limited budgets and large volumes of incoming material.

Not as automatic as it looks

It would be misleading to suggest the process runs entirely on its own. The method depends completely on the quality of the underlying classification scheme, and the class number is its foundation. If a document’s class number is built on a structurally weak scheme, chain procedure cannot index it properly. The method also suffers from what is called the problem of the disappearing chain, where a needed concept does not surface as a separate link. Human judgement and experience remain essential for formulating and assigning the headings, so the system is better described as semi-mechanical than fully automatic.

Adoption by national bibliographies

The real proof of any indexing method is whether large, serious organisations rely on it. Chain indexing passed that test twice over.

The British National Bibliography (BNB), established around 1950 to record everything published in the United Kingdom, adopted chain indexing as its method for deriving subject entries. It used the procedure through the 1950s and 1960s, which gave the method a high-profile, real-world showcase. The BNB eventually moved on to PRECIS, the Preserved Context Index System developed by Derek Austin and applied to the bibliography in the early 1970s, but for two decades chain indexing was the engine behind subject access in one of the world’s major national bibliographies.

Closer to home, the Indian National Bibliography (INB) has used chain procedure for its subject headings since its launch. The INB, compiled at the Central Reference Library in Kolkata, is built as an authoritative record of current publications in 14 major languages including English. Its entries are arranged in a classified sequence following the Dewey Decimal Classification, with the corresponding Colon Classification number printed alongside, and chain procedure supplies the subject headings throughout. The first issue appeared in 1958, and the method has remained central to its subject arrangement.

Comparison with other indexing methods

How does chain indexing stand against the alternatives? The clearest contrast is with the enumerated subject heading lists such as LCSH and SLSH. Those lists are pre-built vocabularies that the cataloguer consults and selects from. Chain indexing flips this around by generating the vocabulary from the document’s own class number, which means it adapts automatically to whatever subject arrives, including new and highly specific topics that a fixed list might not yet contain.

The trade-off is dependence. Enumerated lists work with any catalogue regardless of its classification, while chain procedure needs a logical, hierarchical classification scheme to function at all. Chain procedure scores heavily on economy and provides both general and specific access while avoiding arbitrary decisions, but its reliance on the class number is its main weakness.

Chain indexing also influenced the methods that came after it. PRECIS, which replaced it at the BNB, and POPSI (Postulate-based Permuted Subject Indexing), developed at the Documentation Research and Training Centre in India, both grew out of attempts to keep the strengths of classificatory subject indexing while shedding the dependence on a class number. POPSI in particular was designed to remain rooted in Ranganathan’s postulates and principles without being tied to the class number itself. In this sense, chain procedure was not just a method but a foundation that later, more sophisticated systems built upon.

What do you think? If chain indexing reduces arbitrary decisions by deriving headings from a class number, do you think a method can ever be truly objective when human judgement is still needed to decide which links are sought and which are not? And given how heavily it depends on the classification scheme behind it, would you trust an automated catalogue built on chain procedure more than one built on a fixed subject heading list?

How useful was this post?

Click on a star to rate it!

Average rating 0 / 5. Vote count: 0

No votes so far! Be the first to rate this post.

References

PDF 📄

Table of Contents

The origins of chain indexing 🔗

Where the idea came from 🔗

How it changed subject cataloguing 🔗

Sought, unsought, and false links 🔗

An economic and efficient approach 🔗

Not as automatic as it looks 🔗

Adoption by national bibliographies 🔗

Comparison with other indexing methods 🔗