Chain Indexing: Ranganathan’s Method for Subject Indexing

Every library faces the same basic problem: a reader knows the topic they want, but the books are arranged by class number. How do you connect a person searching for “dogs” to the shelf marked 636.7? Chain indexing is the elegant answer that Dr. S.R. Ranganathan worked out in the 1930s. Instead of inventing subject headings from scratch, it pulls them straight out of the class number a document already has. This single idea reshaped how classified catalogues were built and influenced subject indexing systems for decades. Let’s unpack what chain indexing is, how the process actually works step by step, and where it shines and stumbles.

What chain indexing is and why it exists 🔗

Chain indexing, also called chain procedure, is a systematic method of deriving subject headings or class index entries from the class number of a document. Ranganathan first described it in his 1938 work Theory of Library Catalogue, having designed the technique a few years earlier for his Colon Classification. The central claim is that classification and subject indexing are two sides of the same coin. When a classifier assigns a class number, they have already analysed the subject into its component ideas and arranged them in a logical order. Chain indexing simply translates that number back into words.

The purpose is practical. A classified catalogue has two parts: a classified file arranged by class number, and an alphabetical subject index that points readers to those numbers. Most users do not think in notation. They search using ordinary words. The alphabetical index is therefore the doorway through which they enter the classified arrangement. Chain indexing is the mechanism that generates the entries for that doorway in a consistent, repeatable way.

The meaning of the “chain” 🔗

A class number is not a random string of digits. Each digit represents a step in a hierarchy, moving from a broad subject to a progressively narrower one. This descending sequence of subordinate ideas is the chain, and each individual step is a link. For the class number 636.7 in the Dewey Decimal Classification, the chain runs from Technology (6) to Agriculture (63) to Animal husbandry (636) and finally to Dogs (636.7). Reading the chain from top to bottom shows the subject getting more specific at every stage. This built-in hierarchy is what makes the whole method work.

Why Ranganathan designed it 🔗

Before this, librarians often relied on enumerated lists of subject headings such as the Sears List or the Library of Congress Subject Headings, where every heading had to be pre-determined and looked up. Ranganathan saw two problems with that approach: it duplicated the intellectual work the classifier had already done, and it did not guarantee that headings matched the structure of the classification scheme. His solution was that the indexer should start from where the classifier left off rather than repeat the analysis. The result was a method described as semi-mechanical, because much of it follows fixed rules once the class number is in hand.

How the chain indexing process works 🔗

The process moves from a finished class number to a set of ready-to-file subject entries. While the exact rules vary slightly between the Colon Classification and Dewey, the underlying sequence stays the same.

Step one: construct the class number 🔗

Everything begins with classification. The indexer either takes the class number already assigned to the document or builds it by analysing the document’s specific subject and synthesising the notation using a chosen scheme. Chain indexing was created for the Colon Classification, but it was later found to work with any hierarchical scheme whose notation shows clear subordination, which is why it has commonly been demonstrated using Dewey as well.

Step two: unwind the chain link by link 🔗

Next, the indexer expands the class number into its full chain. The notation is read one digit (or one meaningful step) at a time, and each step is written out with its verbal equivalent taken from the classification schedule. This produces a vertical list running from the most general concept at the top to the most specific at the bottom. This unwinding is the heart of the procedure, because it exposes every concept hidden inside the notation.

Step three: identify the kinds of links 🔗

Not every link in the chain deserves its own index entry. Ranganathan classified the links into distinct types, and learning to tell them apart is the skill that separates a good indexer from a clumsy one. According to the standard description used in subject cataloguing texts, the main types are:

Sought links are the concepts a reader is genuinely likely to use as a starting point for their search. Each sought link will eventually generate one entry. Unsought links represent concepts that exist in the hierarchy but that no realistic user would search for. False links do not stand for any valid subject at all; they are usually connecting symbols, indicator digits, or, in Ranganathan’s rules, the time facet, and they are simply ignored. Missing links are concepts that the subject genuinely contains but that the classification scheme failed to provide a digit for, which the indexer must supply through verbal extension.

An important caution is that the boundary between sought and unsought is not absolute. As noted in the IGNOU teaching material, whether a link is sought depends on the readers a particular library serves, so a link that one library treats as unsought may be very much sought in another. When in doubt, the convention is to give the benefit of the doubt and treat it as a sought link.

Step four: derive headings through reverse rendering 🔗

Now comes the move that defines chain indexing. The specific subject heading is derived from the last sought link and built upwards through the necessary higher links in a backward, or reverse rendering, process. The most specific term leads, followed by the broader terms that give it context, with components usually separated by commas. This is described in detail in the guidance on subject indexing techniques, which explains that headings should use the minimum number of upper-link terms necessary to make the meaning clear. After the specific entry is made, additional reference entries are prepared from each higher sought link, pointing the reader from broader headings towards the specific one.

A worked example 🔗

Take a document on dogs with the Dewey number 636.7. The chain unwinds like this: Technology, then Agriculture, then Animal husbandry, then Dogs. Reading from the bottom upward, the sought links produce a series of entries. “Dogs” becomes the specific subject heading. From the higher sought links, the indexer prepares reference entries such as “Animal husbandry” and “Agriculture,” each guiding the reader towards the more specific topic. Notice how a single class number quietly produced several access points. The most specific entry carries the full subject, while dropping terms one at a time creates broader entries that serve readers searching at a more general level. This gives both specific and generic search routes from one act of classification.

Advantages of chain indexing 🔗

The strongest selling point is economy of effort. Because the indexer works from the class number, there is no duplication of the subject analysis the classifier already performed, which saves real time. The method is largely mechanical, so once the rules are mastered it operates quickly and yields consistent results rather than depending on each cataloguer’s personal taste. It also offers multiple approaches to a single document, since one class number generates a specific heading plus several broader reference entries, supporting both narrow and wide searches.

Another advantage is versatility. The same procedure can be applied to any classification scheme whose notation shows clear subordination, and it can be used to build indexes to classification schedules and to books, as well as headings for shelf guides and stockroom guides. Its influence has been considerable. Chain indexing was adopted by the British National Bibliography in the 1950s and 1960s and has long been used by the Indian National Bibliography, demonstrating its value at a national scale before later systems such as PRECIS took over in some places.

Limitations of chain indexing 🔗

The method’s great strength is also its great weakness: it is entirely dependent on the classification scheme. If the scheme is weak, the index inherits that weakness. The commonly cited critique is that a broad scheme like Dewey, which groups documents only loosely, does not individualise subjects finely enough, so chain indexing performs far better with a rigorous, faceted system like the Colon Classification.

It is also less mechanical than it first appears. Human judgement is needed to decide which links are sought, and that dividing line is thin and hazy, requiring experience and knowledge of both the subject and the language. The scheme offers no help in discovering synonyms, so a reader using a different word for the same idea may miss the entry. The problem of missing links forces the indexer to step in and supply concepts the notation skipped. And any error in the original class number is carried straight into the subject headings.

Why computerization is challenging 🔗

Chain indexing is often called amenable to computerization, and programs have indeed been written to generate headings from class numbers by reverse rendering. Yet full automation runs into the same human-judgement problem described above. A computer can unwind a chain and reverse-render terms reliably, but deciding which links are sought, supplying missing links, and editing headings to remove awkward context all need interpretation that resists a fixed algorithm. Reverse rendering can also read oddly to users, since the term order does not match natural speech. These factors explain why later post-coordinate and computer-friendly systems gained ground even though chain procedure laid the conceptual groundwork.

What do you think? If a classification scheme decides what your subject index can express, how much does the choice of scheme shape what readers are able to find? And in an age of full-text search and keyword retrieval, does the disciplined logic of chain indexing still offer something that algorithms miss?

References

1. https://www.librarianshipstudies.com/2017/04/chain-indexing.html
2. https://egyankosh.ac.in/bitstream/123456789/38416/5/Unit-14.pdf
3. https://ebooks.inflibnet.ac.in/lisp3/chapter/subject-cataloguing-chain-procedure-popsi-and-precis/
4. https://egyankosh.ac.in/bitstream/123456789/33116/1/Unit-17.pdf
5. https://www.studocu.com/in/document/university-of-delhi/library-science/unit-14-pdf/89625416