Library Twin · Reference

Reference

Work, edition, copy

The word “book” means three different things at once. Most catalogues only record one of them, and that is why they quietly stop being useful.

The three levels

Say you own Pride and Prejudice. That sentence is already ambiguous, because it could mean any of three things.

Every physical book you own sits at all three levels simultaneously. A catalogue that records only the title has collapsed all three into one, and cannot answer questions about the other two.

Why this breaks a home library

The failures are not theoretical. They are the ordinary frustrations people describe about their own shelves:

• Accidental duplicates. You buy a book you already own, because your list says “I have this title” and cannot tell you that the copy you have is a battered reading paperback rather than the gift edition you meant to buy.
• Inherited books lose their history. A copy inherited from a parent is not interchangeable with a copy bought last year. The inscription, the provenance and the marginalia belong to that object. Merge them by title and the history is gone.
• Lending becomes untraceable. “Who has my copy?” is a question about a copy. If three copies of a title exist in the house, a title-level record cannot say which one left.
• Location stops working. A title does not sit anywhere. A copy sits somewhere specific, and can be moved without the title changing at all.
• Condition and value become meaningless. Condition is a property of an object. A first printing in a dust jacket and a water-damaged book club edition of the same work are not the same asset.

Each of these failures has the same root cause: a level of identity was discarded at the moment the book was recorded, and no amount of later tidying brings it back.

Where this model comes from

This is not a new idea, and it is not ours. Librarians formalised it decades ago. The current standard is the IFLA Library Reference Model, published in August 2017 by the International Federation of Library Associations, which unified three earlier models: FRBR, FRAD and FRSAD.

It defines four levels, known together as WEMI:

• Work — the intellectual creation.
• Expression — a specific realisation of it, such as a particular translation or a revised text.
• Manifestation — the published form: this printing, by this publisher, in this year.
• Item — the single physical object.

An honest simplification

Library Twin uses three levels rather than four. Our work is LRM’s Work, our edition corresponds to LRM’s Manifestation, and our copy corresponds to LRM’s Item. We do not model Expression as a separate level.

That is a deliberate trade, and it has a cost worth stating plainly. Expression is what cleanly separates two translations of the same novel, or an author’s revised text from the original. In our model, those distinctions live at the edition level instead, where they are recorded but not structurally separated. For a research library that would be a real loss. For a home collection, where the practical questions are which copy do I own, where is it, and where did it come from, the fourth level adds complexity without answering the questions people actually ask. Anyone cataloguing to a professional standard should use the full model.

How Library Twin uses it

Library Twin keeps work, edition and copy as separate records that stay linked. Two copies of the same edition remain two records, each with its own location, condition, acquisition history, inscriptions, loans and reading history. Identity operations are built not to silently merge records that might be different things: where the evidence is ambiguous, the ambiguity is shown for a person to resolve rather than guessed at.

This is the reason Library Twin can answer “which of my three copies is the one Dad gave me, and where is it” — a question that a title-level catalogue cannot represent at all.