Beyond words

GUEST POST: How numbers should and shouldn’t shape our understanding of history

Recently, The Economist published an article warning that popular history is becoming oversaturated with listicles – a formula whereby the past is told through ‘an intriguingly specific number of unexpected things’.¹ You know the kind. In the last few years, readers have been inundated with books promising to tell the history of the world in eight plagues, ten dinners, 21 women, 50 plants, or 100 objects housed at the British Museum. Such histories-by-numbers are ‘appealingly accessible’, the author writes. However, they risk giving the impression, to quote Arnold Tonybee, that history is merely ‘one damned thing after another’.

The article got me thinking about a question which has been at the back of my mind ever since I completed my master's degree with the Biography of an Uncharted People Project in 2020. How should and shouldn’t numbers shape our understanding of the past? I should say upfront that I am not a fan of the listicle. I am a grey-area-thinking sort of person and dislike the idea of history being neatly packaged in ways which ignore larger questions around contingency and the often murky interactions between people and things. That said, I recognise the need to make history more accessible and think numbers can play an important role in this process. 

In A History of the World in Numbers, Emma Marriot argues that numbers can be used to tell histories both big and small.² Large numbers can illuminate the broad sweep of history, from the mass movement of populations to the expansion of empires, while small numbers often measure the ‘living details of history’ such as a person’s age or the cost of bread. Marriot’s book offers a fascinating compendium of some of the most important and surprising numbers in history. For example, in 1637 a Dutch florist sold a single tulip bulb for 3000 guilders, twenty times the annual salary of a skilled craftsman. In the 1850s, the life expectancy of a slave in Brazil was 23, while during the Second World War, US troops in the Pacific were supplied with 20 cigarettes as part of their daily rations. Marriot’s book also mentions the 1913 Natives Land Act, which prevented black Africans from buying or renting land across 90% of South Africa.

Moving beyond precise quantities, numbers can also be helpful in identifying trends. Like most historians, I believe that explaining why something happened in the past is very different from predicting how and when it will happen again. Consequently, I am wary of Peter Turchin’s suggestion that, like physics, history can be used to produce laws. However, quantitative history comes in many shapes and sizes. Social historian Beverly Bunch-Lyons has written an excellent article on how quantitative methods helped her research on African American women’s involvement in underground economies.³ Johan Fourie, myself and several colleagues have also published a recent book using quantitative methods to tell the stories of indigenous peoples, women, enslaved peoples, and other marginalised groups in South African history.⁴  

For historians and social science scholars just starting in the field, I would suggest consulting Pat Hudson and Mina Ishizu’s History by Numbers.⁵ Gary J. Kornblith’s online guide Making Sense of Quantitative Evidence is another useful primer.⁶ In it, Kornblith argues that the challenge for budding quantitative historians is twofold: (1) they must learn how to pose good questions of quantitative sources, and (2) they must learn how to ‘read’ the data to best answer these questions. Below I include four of my own rules for working with quantitative data.  

1. Context matters

To ask the right questions, scholars must contextualise the quantitative data they use. For example, the 1913 Land Act, which prevented blacks from owning 90% of land, formed part of a broader pattern of racial dispossession whereby black South Africans, who constituted 67% of the population at the time, were systematically denied citizenship rights. Reading widely in and beyond one’s discipline is the best way to gain contextual knowledge. Qualitative research, biographies, and even fiction can be useful. I am writing a chapter on demographic trends in late-apartheid South Africa, and have found Sindiwe Magona’s Living, Loving and Lying Awake at Night helpful in imagining the reproductive experiences of black women living in the 1970s and 1980s.⁷

2. Relative vs. absolute quantities

A paucity of data has meant that scholars are sometimes liable to report absolute rather than relative figures, but such figures can be misleading. Researchers must adjust for population and make other simple efforts to compare numbers accurately. The transatlantic slave trade was undoubtedly a horrific and profoundly traumatising event in human history, the social cost of which cannot be overlooked. Nevertheless, it would make little sense to compare the life expectancy of a 19th-century Brazilian slave with that of Westerners today when, in 1850, the life expectancy for the average Britain was a mere 42 years (with much lower figures for Welsh coal miners and other rural poor).

3. Beware spurious but plausible correlations

We all know the truism that ‘correlation doesn’t imply causation,’ but when we observe two lines sloping together, it can be difficult to avoid seeing a relationship. Amusing examples of spurious correlations abound. See Harvard Law School student Tyler Vigen’s website, which allows anyone to chart absurd correlations using large, mostly US-based datasets.⁸ The site shows correlations between visitors to Universal Orlando’s ‘Islands of Adventure’ ride and annual car sales, as well as per capita margarine consumption and rising divorce rates. The problem with these examples is that they are obviously implausible. Nobody expects there to be a relationship between margarine and divorce (autocorrect just tried to change margarine to marriage). It is much harder to be sceptical, though, when we, as serious authors, have chosen the variables involved, usually because of a possible interaction. Yet, the same statistical rules apply. Just because it seems likely that there is a causal relationship between income and voting patterns, does not mean that such a relationship exists. Further tests are needed, and even then, causation is tricky to prove without the ability to perform experiments.

4. Like words, numbers can wield power

There is something solid about numbers which can make them appear neutral bearers of truth, but this is rarely the case. Quantitative researchers need simply consult history to recognise that numbers, as well as words, can be used to wield power. During World War II, concentration camp victims were deliberately branded with numerical identifiers, and Foucault argues in his theory of ‘biopolitics’ that more insidious examples exist of statistics being used to control and dehumanise.⁹ I am currently working on Stanford biologist Paul Ehrlich’s 1968 book The Population Bomb, which presented ‘irrefutable’ data on global population growth and relative food consumption to advocate for extreme methods of population control (including forced sterilisation and the succession of food aid to India).¹⁰ Ehrlich’s predictions have proven false.

In summary, I put little store by the aphorism attributed to Mark Twain amongst others that, ‘There are three kinds of lies: lies, damned lies, and statistics.’ Instead, I believe that numbers and words are similar: both can obscure as well as reveal the truth. To achieve the latter outcome, scholars need to approach each medium with a combination of seriousness and scepticism, becoming trained in the various skills necessary to accomplish their research goals.

‘Beyond words was first published on Our Long Walk. The image was created with Midjourney v6.

1. The Economist. ‘Can a dozen shipwrecks tell the history of the world?’, 27 February 2024, https://www.economist.com/culture/2024/02/27/explaining-histo (date accessed 15/03/2024).

2. Emma Marriott. A History of the World in Numbers (Michael O'Mara, 2014)

3. Beverly Bunch-Lyons. ‘A Social Historian Retools and Reframes’, Perspectives on History, 1 April 2015, https://www.historians.org/research-and-publications/perspectives-on-history/april-2015/a-social-historian-retools-and-reframes (date accessed 15/03/2024).

4. Johan Fourie ed. Quantitative History and Uncharted People: Case Studies from the South African Past (Bloomsbury, 2023).

5. Pat Hudson and Mina Ishizu. History by Numbers: An Introduction to Quantitative Approaches (Bloomsbury, 2016).

6. Gary J. Kornblith. ‘Making Sense of Quantitative Evidence’, https://historymatters.gmu.edu/mse/numbers/numbers.pdf (date accessed 15/03/2024).

7. Sindiwe Magona. Living, Loving and Lying Awake at Night (David Philip, 2004).

8. Tyler Virgen. ‘Spurious Correlations’, https://www.tylervigen.com/spurious-correlations (date accessed 15/03/2024).

9. Leonard Lawlor and Johns Nale eds. ‘Biopolitics’  in The Cambridge Foucault Lexicon (Cambridge University Press, 2014), 37-43.

10. Paul Ehrlich. The Population Bomb (Ballntine Books, 1968).