Growth vs Obsolescence in Literature: How They Influence Each Other

Every year, more research is published than ever before. Scientific output keeps multiplying, yet a strange thing happens alongside this explosion: older papers stop getting cited and quietly fade from active use. These two forces-the rapid growth of literature and its steady obsolescence-are not separate phenomena. They are deeply connected, and understanding how one shapes the other reveals a great deal about how knowledge ages, circulates, and eventually retires. This relationship sits at the heart of scientometric research, and the answers it produces are more surprising than they first appear.

The interplay between growth and obsolescence 🔗

Growth of literature refers to the increase in the volume of published documents over time. Obsolescence refers to the decline in the use or citation of documents as they age. At first glance, these look like opposite movements-one going up, one going down. But the speed at which a field grows directly conditions how quickly its older material appears outdated.

The logic is intuitive once stated plainly. When a discipline produces a large amount of new work each year, the newest publications dominate the citation pool. Researchers naturally cite the most recent contributions, so older items are crowded out and seem to “age” faster. A field that grows slowly behaves differently: with fewer new papers competing for attention, older works retain their relevance for longer. This is why fast-moving fields like medicine and engineering show sharp citation peaks followed by rapid decline, while slower fields keep older references alive much longer.

The crucial point is that growth influences obsolescence, not the other way around. The rate of publication is a driver, and the apparent rate of aging responds to it. This is precisely why the growth of a literature and the growth of its contributing authors were long suspected of distorting any honest measurement of how literature ages.

How growth of literature actually behaves 🔗

Before connecting growth to obsolescence, it helps to understand how literature grows. The pioneer of this study was Derek de Solla Price, who argued that science follows a law of exponential growth: at any moment, the rate of growth is proportional to the size already achieved. In simple terms, the bigger a body of literature already is, the faster it adds to itself.

Price observed that scientific literature tended to double in volume roughly every ten to eleven years, a pattern confirmed across many disciplines. Modern estimates place the overall doubling time somewhat differently-one large analysis found an average growth rate of around 4 percent with a doubling time near 17 years-but the underlying exponential character holds.

From exponential to logistic growth 🔗

Pure exponential growth cannot continue forever. Price himself noted that the initial exponential phase eventually slows and flattens into a logistic, or S-shaped, curve once a field matures and approaches a ceiling. A review of growth models across more than fifty subjects found that close to half of the studied literatures followed this exponential-then-logistic pattern, while the rest fitted linear, power, or other models. The growth rate, in other words, is not uniform-and because growth conditions obsolescence, the aging of a literature changes as the field passes through these phases.

Theoretical investigations: Egghe’s findings 🔗

For a long time the relationship between growth and obsolescence was discussed informally. Researchers sensed that faster growth meant faster apparent aging, but no formal mathematical model tied the two together. This gap was filled most decisively by Leo Egghe, whose 1993 paper directly tackled the question.

Egghe built a model assuming an increasing exponential function for production (growth) and a decreasing exponential function for aging (obsolescence). His central conclusion was elegant and unexpected. He proved that in the synchronous case, an increase in growth implies an increase in obsolescence, while in the diachronous case the exact opposite occurs. The same underlying force-faster growth-pushes the measured obsolescence in opposite directions depending on how you choose to study it.

This is one of the most important results in the field, because it shows that the answer to “does faster growth cause faster aging?” depends entirely on the measurement method. There is no single answer; there are two correct answers that contradict each other, and both are mathematically valid.

Why the obsolescence factor is not constant 🔗

Egghe’s work, often alongside I. K. Ravichandra Rao, also corrected an earlier assumption. Bertram Brookes had treated the obsolescence (aging) factor as a fixed constant. Egghe and Rao demonstrated that this factor is not a constant but a function of time, because citation data is not distributed in the simple exponential way Brookes assumed. Instead, citation patterns typically show an initial rise followed by an exponential decline-a shape that the lognormal distribution describes well. This refinement matters because it means obsolescence cannot be captured by one number; it shifts as literature accumulates and as the field grows.

Synchronous vs diachronous studies 🔗

To understand Egghe’s split result, you have to understand the two ways of measuring how literature ages. These are not minor technical variations-they are fundamentally different vantage points, and they produce different results from the same underlying reality.

The synchronous approach 🔗

A synchronous study looks backward from a single point in time. It takes the references in a set of current articles and asks how old those references are. As one survey puts it, synchronous studies are concerned with plotting the age distribution of the material used at one moment. This is a retrospective method based on references made. You stand in the present and survey the ages of everything being cited right now.

Because a fast-growing field publishes far more recent papers, the references in today’s articles will lean heavily toward recent years. The age distribution looks compressed, suggesting rapid obsolescence. So in synchronous terms, more growth produces the appearance of faster aging.

The diachronous approach 🔗

A diachronous study works in the opposite direction. It selects a set of documents published in a particular year and then follows them forward through time, tracking the citations they receive in subsequent years. This is a prospective method based on citations received. The concept of half-life-the time taken for a body of literature to receive half of its eventual citations-is naturally discussed in this diachronous context.

Here the effect of growth reverses. Because the field keeps expanding, there are more and more potential citing papers in later years, which can keep older documents in use longer than expected. Faster growth, measured diachronously, can therefore make literature appear to age more slowly. This is exactly the opposition Egghe formalised.

Do the two methods ever agree? 🔗

This contradiction worried researchers, because if two valid methods disagree, which should be trusted? Stinson and Lancaster examined this directly using the literature of human and medical genetics. They compared a synchronous study over a 19-year period with a diachronous one and reached a reassuring conclusion. They found that once the first two years of synchronous data are excluded, the synchronous and diachronous rates of obsolescence become statistically equivalent.

Even more striking, they tested whether synchronous studies needed to be corrected for the growth of the literature, as Brookes had argued. Their data supported Brookes’s idea that the growth of the literature and the growth of the number of contributors have opposite effects-and these two corrections effectively cancel each other out. A synchronous study with no corrections at all gave results equivalent to one carefully corrected for both factors.

Experimental and empirical evidence 🔗

Theory is only half the story. Most studies on growth and obsolescence are empirical, fitting models to observed citation data through regression analysis. These experimental investigations consistently confirm the link between how fast a literature grows and how it ages.

A clear example comes from physics. Using the synchronous method on major physics journals, Gupta studied citations to Physical Review and found an exponential decrease in citation density with age and a half-life of about 4.9 years. Fast-moving “hard” sciences like physics and engineering show short half-lives, reflecting rapid growth and quick turnover of cited material. Computing literature, similarly, was found to have a half-life of around four years.

Empirical work on specific fields reinforces the pattern that growth and obsolescence must be studied together. Research modelling obsolescence in disciplinary journals confirms that citation distribution always combines aging with literature growth, because the early years of any discipline simply contain fewer documents than its later years. You cannot observe pure aging in isolation; the growth signal is always mixed in. This is why empirical curves typically show that initial growth phase followed by exponential decline rather than a clean downward slope.

The role of field-specific growth 🔗

Experimental studies also reveal why obsolescence rates differ so much between disciplines. The deciding factor is the speed of knowledge diffusion, which is tightly linked to growth. Fields with rapid diffusion-and rapid publication growth-show a sharp citation peak followed by a steep drop, as seen in medicine and engineering. Fields with slower diffusion show a gentler decline, keeping older work relevant for longer. The correlation between the pace of growth and the pace of obsolescence is not an abstraction; it appears repeatedly across subject after subject when real citation data is analysed.

What do you think? If a fast-growing field makes its own literature appear to age more quickly, does that mean rapid scientific progress comes at the cost of forgetting valuable older work? And given that synchronous and diachronous methods can point in opposite directions, which measure should libraries and researchers trust when deciding what literature is truly obsolete?

References

1. https://journals.sagepub.com/doi/abs/10.1177/016555158701300201
2. https://garfield.library.upenn.edu/price/pricequantitativemeasures1951.pdf
3. https://www.nature.com/articles/s41599-021-00903-w
4. https://www.academia.edu/43490926/Revisiting_De_Solla_Price_growth_dynamics_studies_of_various_subjects_over_last_one_hundred_years
5. https://link.springer.com/article/10.1007/BF02016550
6. https://ebooks.inflibnet.ac.in/liscp10/chapter/different-models-to-explain-the-phenomena-of-growth-and-obsolescence-of-literature/
7. https://ebooks.inflibnet.ac.in/liscp10/chapter/obsolescence-factor-definition-and-calculation/
8. https://journals.sagepub.com/doi/10.1177/09610006231196891
9. https://link.springer.com/article/10.1007/s11192-022-04359-w