What the Book Gives Off

Ask a book what it smells like and you get chemistry. Ask a person and you might get chocolate.

In an experiment at the Birmingham Museum and Art Gallery, visitors encountered an unidentified smell made from historic paper. The most frequently offered word was not library or dust. It was chocolate: 21 mentions among 79 respondents, who supplied 120 descriptors in all. Add cocoa, Cadbury and chocolatey, and the chocolate family reaches 27 mentions. Coffee came next, with ten. These are counts of descriptions, not a majority vote declaring that books are chocolate. [1]

I like the result because it scrambles a familiar category. A book is something to read; chocolate is something to eat. A nose has no obligation to respect the shelving.

But the experiment is more interesting than the headline version. The visitors were not sniffing an intact old volume. The researchers had heated a small paper sample, extracted it with methanol, and presented the extract on gauze after allowing solvent to evaporate. The canister did not identify its contents. Even so, the activity included other smells—including chocolate and coffee. The authors explicitly discuss the possibility that those neighbors influenced the answers.

So this is not a clean demonstration that old books inherently smell like dessert. It is a meeting of a prepared chemical mixture, a human vocabulary and a particular exhibition's context. All three belong in the explanation.

Paper has a recipe

“Old paper” sounds like one material at a later stage of life. It isn't.

In a 2009 study of paper degradation, Matija Strlič and colleagues examined 72 historic papers from the nineteenth and twentieth centuries. The collection covered different fibers and manufacturing choices: rag papers, bleached pulp, groundwood, gelatin sizing, rosin sizing, coatings. Cellulose gives paper its structural framework, but it does not travel through history alone. Lignin, sizing and other constituents alter the chemical possibilities. A whole book also brings bindings, applied media and its environmental history into the mixture. [2]

As materials degrade, some products escape as volatile compounds. Among the paper-related compounds discussed in these studies is furfural, formed through degradation pathways involving cellulose and hemicellulose. The 2017 study's table associates it with the odor descriptions bread, almond and sweet. Benzaldehyde is associated with almond and burnt sugar. Acetic acid gets the less romantic sour. [1, Table 1]

Those words are useful handles, not a list of ingredients that guarantees a particular experience. The table's aroma descriptions came from established odor databases. They do not establish how much each compound contributed to what these visitors smelled. Detecting a molecule is one job; establishing its perceptual importance in a mixture is another.

This is where the simple “books smell like vanilla because lignin” explanation loses me. It may point toward one part of a real chemistry, but it makes the object sound as though it has one precursor, one product and one inevitable aroma. The studies describe a network, not a flavoring bottle.

A chemical footprint is not a birthday

The 2009 researchers asked whether emitted compounds could reveal something about paper's composition and condition. They measured volatile patterns and related them to properties such as lignin content, acidity and the degree of polymerization of cellulose—a measure of its chain length.

That is a more demanding question than whether something smells old. It also needs the experimental conditions attached.

Their small paper samples were predegraded at 80°C for 24 hours before the released compounds were collected and analyzed. This was an accelerated laboratory preparation, not an instrument quietly reading an untouched library shelf. The authors distinguish the destructive development work from the possibility of future nondestructive monitoring.

The statistical models were informative but not magical. The authors describe their calibrations as relatively low quality, while finding value in what the patterns suggested about paper components and degradation. For acidity, the reported cross-validation error was 1.2 pH units. This was not a demonstration that a person—or a machine with no appropriate calibration—could sniff an arbitrary book and announce its age or remaining life. [2, Table 2]

That limit changes what I think the smell is telling us. It is not time itself made fragrant. It is the consequence of particular materials living through particular conditions. Two equally old books need not have followed the same chemical road.

The room is another sample

Cecilia Bembibre and Strlič's 2017 heritage-smell study did not stop at the museum extract. They also sampled the headspace of a 1928 copy of Panait Istrati's Les Chardons du Baragan, held in a bag at room temperature, and the air of the historic library at St Paul's Cathedral. Those were separate sampling arrangements from the heated-paper extract shown to visitors. [1, Experimental methodology]

And they brought seven untrained assessors into the library itself. Given a list of 21 odor descriptors and space to add others, every assessor selected woody.

That is not a contradiction of the chocolate result. It is a different object of attention, a different setting and a different task. The library panel could see wooden furnishings and old books. Its members were choosing from supplied words, not naming a concealed smell freely. The authors flag both visual and verbal context as questions needing further exploration.

The two papers share Strlič as an author; they are complementary investigations, not independent replications of one finding. But together they make a useful distinction visible: a record of chemicals is not yet a record of experience.

The 2017 researchers tried to connect those records in a Historic Book Odour Wheel: broad aroma categories, more specific human descriptors and likely chemical counterparts. They present it as a preliminary documentation tool. It is not an exact recipe for reproducing every library, or a validated universal diagnosis of paper condition.

Its ambition is still worth taking seriously. If you preserve only the chemical list, you lose the words people used. If you preserve only woody or chocolate, you lose the material evidence. If you preserve both, you still have to ask why this particular smell mattered.

Worth keeping is a different judgment

The study quotes visitors who valued the smell at St Paul's as part of their encounter with the place. That is evidence of an attachment, not a reason to encourage deterioration. A material can be changing in ways a conservator wants to slow while giving off an aroma a visitor wants to remember.

I started with the idea that chemistry explains the smell and nostalgia explains the affection. I now think that division is too tidy. The preparation of the sample, the room around it and the language offered to a nose all enter the account before we even reach the question of value.

A book gives off chemicals. A person supplies a description. A place can make that description mean something.

If we could reproduce the aroma perfectly, then, what would we have preserved: the book, the library—or one way of arriving there?

Sources

[1] Cecilia Bembibre and Matija Strlič, “Smell of heritage: a framework for the identification, analysis and archival of historic odours,” Heritage Science 5, 2 (2017). Open full text. Museum descriptors, library sensory assessment, Table 1, odour wheel and experimental methodology. Full main text, methods, table and captions consulted; supplementary questionnaires not reanalyzed.

[2] Matija Strlič et al., “Material Degradomics: On the Smell of Old Books,” Analytical Chemistry 81 (2009), 8617–8622. DOI; published paper, academic-hosted PDF. Materials and Methods, Table 2, Results and Discussion, and conclusions. This essay is a reading of the research, not a new smelling experiment, library visit or conservation assessment.