Abstract
We study the effects of scientific changes on broader cultural discourse, two phenomena that the economics literature identifies as key drivers of long-term growth, focusing on a unique episode in the history of science: the elaboration of the theory of evolution by Charles Darwin. We measure cultural discourse through the digitized text analysis of a corpus of hundreds of thousands of books as well as of Congressional and Parliamentary records for the US and the UK. We find that some concepts in Darwin’s theory, such as Evolution, Survival, Natural Selection and Competition, significantly increased their presence in the public discourse immediately after the publication of On the Origin of Species. Moreover, several words that embedded the key concepts of the theory of evolution experienced semantic and sentiment changes—further channels through which Darwin’s theory influenced the broader discourse. Our findings represent the first large-sample, systematic quantitative evidence of the relation between two key determinants of long-term economic growth, and suggest that natural language processing offers promising tools to explore this relation.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Mokyr (2013), moreover, distinguishes between “macro” and “micro” scientific discoveries. He argues that only the former creates a discontinuous change. The latter are important to guarantee continuous improvements, but do not cause any scientific or cultural breakthrough.
Similarly, Schiller (2017) advances the idea that certain individuals may introduce “narratives” that spread in a society and affect broader beliefs and consequent actions.
See for example Guiso et al. (2004).
Giuliano and Nunn (2021) have recently advanced an agenda to measure cultural persistence and change.
We limit the semantic and sentiment analysis to the Google Book corpus because of the demanding sample size requirements of the underlying methodologies.
See in particular Desmond and Moore (1994) for details on the personal and intellectual biography of Darwin.
Similarly, studies of the literary production prior to the publication of On the Origin of Species point out how some of Darwin’s ideas connected to images already developed by these writers. A frequently cited example is the work of Alfred Tennyson, and in particular his poem In Memoriam, published in 1850. Scholars also investigated the connections between broader worldviews, such as Enlightenment and Romanticism, on Darwin’s ideas (Cartwright and Baker 2005; Chapple 1986; Gianquitto and Fisher 2014; Lansley 2016; Otis 2009; Richards 2013; Scholnick 2015).
The year 1859 saw also the publication of other important works, John Stuart Mill’s On Liberty, Tennyson’s Idylls of the King, Eliot’s Adam Bede and Dickens’ A Tale of Two Cities. These publications make it harder to identify a connection between the publication of The Origins of Species and changes in the public discourse. However, in our study, we focus on specific concepts that are central in Darwin’s work but not in the other works mentioned above; we also consider the presence of those concepts in the public discourse before 1859. Below, we also assess the presence of any abnormal trend in the number of words in our dataset, and in the introduction of new words, around 1859.
For various applications in political economy, the study of media, innovation, marketing and finance, see also Balsmeier et al. (2018), Bandiera et al. (2017), Catalini et al. (2015), D’Amico and Tabellini (2017), Enke (2020), Iaria et al. (2018), Jelveh et al. (2014), Kearney and Liu (2014), Kozlowski et al. (2019), Yin et al. (2021). Gentzkow et al. (2018) provide a survey of the use of text as data in economics.
Available at: http://books.google.com/ngrams.
Over this period, there are about 1.26 million unique words on average per year.
Parliamentary Debates (Series 1 to 4 – 1903 to 1908).
These include Congressional Record (1873–1997), the Congressional Globe (1833–1873), the Register of Debates in Congress (1824–1837), and the Annals of Congress (1789–1824). It is worth noticing that before 1873 each House was only required to keep an internal journal of its proceedings. Only from 1874 onwards were external reporters allowed to witness debates and granted full permission to report them (McPherson 1942).
The softmax function maps scores into probability distributions as follows: \(\mathrm{p}\left(c|w;\theta \right)=\frac{{e}^{{v}_{c}{ v}_{w}}}{\sum_{{c}^{^{\prime}}\in C}{e}^{{{v}^{^{\prime}}}_{c}{ v}_{w}}}\), where \({v}_{c}\) and \({v}_{w}\) are vector representations of context word c and focal word w respectively, and C is the set of all possible contexts. The estimation procedure thus consists of maximizing the probability that a given context word occurs within a given window around each focal word of interest.
Kozlowski et al. (2019), for instance, project vectors related to different musical genres (e.g., jazz, rap, etc.) onto a plan that measures “affluence”, to determine whether, say, jazz is more associated with wealthier strata of a population than rap, and how these associations vary over time.
The projection of a word vector on a vector-dimension is equivalent to the cosine of the angle between the two vectors if the vectors are normalized.
Interestingly, the increase in the use of Evolution occurred especially since the 1870s, coinciding with the use of the term in the 1872 edition of Darwin’s book.
In Figure A1 of the Appendix, we show additional frequency analyses where we “stem” the main Darwinian (Evolution, Selection, Survival, Competition and Adaptation) and consider also other words with the same roots. Specifically, we plot the frequency of one of five key words, and the average frequency of a set of other words with the same root. There are two main patterns. In the cases of Evolution, Competition and, to a lesser extent, Selection, the average frequency of the words with the same root appears to follow a similar pattern as the corresponding word of interest. However, these other words have substantially lower frequency and therefore presence in the written language; in other words, in these three cases our main word of interest within a given etymological root is dominant. In the other two instances, Survival and Adaptation, the words with those same roots have similar diffusion on average, but their pattern over time is erratic. This evidence corroborates our focus only on the key Darwinian terms rather than the stemmed words overall.
The statistical significance of the estimates is robust to multiple hypothesis corrections; we applied the Romano-Wolf procedure to the six regressions whose estimates are in Table 1, to account for the use of multiple left-hand-side variables. See Romano and Wolf (2005a, b, 2016) as well as Clarke et al. (2019) for the Stata procedure. Standard error estimates are very similar if we use the Newey-West in lieu of the Huber-White correction (see Tables A8 through A10).
In the Appendix, Table A1 reports the list of these words. The list includes both general terms like number, animals and nature, and more specific ones such as eggs and insects. In Appendix Figure A3, we display the relative occurrence of some of the ninety-nine control terms as an example: Nature, Number, Life, Animals, Flowers, Plants. For none of these words is there any discernible change in diffusion in the decades immediately preceding and following the publication of On the Origin of Species. Appendix Table A3 reports estimates from spline regressions (with one knot at year 1859) on these six words. Appendix Figure A4 plots the estimated frequency slope in 1820–59 and 1860–99 for each of the ninety-nine words; the estimates gravitate around the 45-degree line, thus indicating limited changes in the rate of diffusion after the publication of On the Origin of Species.
The list of these alternative control words is in Appendix Table A4.
With the exception of Chinese, the other languages that we considered are linked to a predominantly Christian culture, where creationism was well accepted. The N-gram database does not include many languages that refer to non-Christian environment. In addition to Chinese, the only available one is Hebrew. We report our analyses on this language in Appendix Figure A8. Because a Hebrew version of On the Origin of Species was only available starting from 1960, we also extended the period of observation to the end of the XX century to 2000 in order to assess the evolution of Darwinian words after the book’s publication. We observe minimal, if any, use of most of the Darwinian terms in the XIX century, with an increase in diffusion in the second half of the 1900s.
We add the following combinations of names, middle names and last names: Alfred Russel Wallace, Alfred Wallace, Charles Darwin, Charles Robert Darwin, Robert Chambers, Jean-Baptiste Lamarck, Jean-Baptiste de Lamarck, Jean Baptiste Lamarck, Jean Baptiste de Lamarck.
Spline regression analyses confirm that the words more related to Darwin’s theory were significantly more likely to enter the political debate, after 1859, especially in the US Congress, with some lag with respect to the publication of On the Origins of Species. The less significant estimates in the UK Parliament corpus may be due to the fact that some of the Darwinian terms were perhaps more common in the UK than in other English-speaking countries such as the US. We do not find any specific pattern for the most frequent words in Darwin’s book (estimates are in Appendix Table A7).
In Figure A9 of the Appendix, we report additional analysis of semantic similarities between pairs of words. The graphs show, on the one hand, that some of the patterns in Fig. 9 are not specific to a narrowly defined pair of words. For example, the similarity pattern over time between Human and Animal is similar to the one for Human and Ape, Man and Animal, and Man and Ape. On the other hand, broader semantic trends that indicate the role of science in society, and not specific aspects of the theory of evolution, do not experience any change around 1860 (see for example Science-Knowledge, Religion-Knowledge, Science-Religion, Science-Nature, Evolution-Nature).
Additional word similarity rankings (available from the authors) show, consistently with the evidence in Fig. 9, that Evolution became one of the most semantically similar words to terms traditionally used to describe the origin of the world by the religious doctrine, such as Creation and Genesis. Moreover, words like Progenitor, Anthropoid and Descended raised among the words with the most similar meaning to Ape.
References
Aiden, E. and Michel, J.B. (2014). Uncharted: Big data as a lens on human culture. Penguin.
Alesina, A., & Giuliano, P. (2015). Culture and institutions. Journal of Economic Literature, 53(4), 898–944.
Armstrong, N. (1987). Desire and domestic fiction: A political history of the novel. Oxford University Press.
Balsmeier, B., Li, G. C., Assaf, M., Chesebro, T., Zang, G., Fierro, G., Johnson, K., Lück, S., O’Reagan, D., Yeh, B., & Fleming, L. (2018). Machine learning and natural language processing on the patent corpus: Data, tools, and new measures. Journal of Economics & Management Strategy, 27(3), 535–553.
Bandiera, O., Hansen, S., Prat, A., and Sadun, R. (2017). CEO Behavior and Firm Performance (No. w23248). National Bureau of Economic Research.
Bellego, C., and Pape, L. D. (2019). Dealing with the log of zero in regression models (No. 2019–13). Center for Research in Economics and Statistics.
Bolukbasi, T., Chang, K.-W., Zou, J. Y., Saligrama, V., & Kalai, A. T. (2016). Man is to computer programmer as woman is to homemaker? Debiasing word embeddings. Advances in Neural Information Processing Systems, 29, 4349–4357.
Bush, V. (1945). Science, the endless frontier: A report to the President. US Govt. print.
Caliskan, A., Bryson, J. J., & Narayanan, A. (2017). Semantics derived automatically from language corpora contain human-like biases. Science, 356(6334), 183–186.
Cartwright, J. H. and Baker, B. (2005). Literature and science: Social impact and interaction. Abc-Clio.
Catalini, C., Lacetera, N., & Oettl, A. (2015). The incidence and role of negative citations in science. Proceedings of the National Academy of Sciences, 112(45), 13823–13826.
Chapple, J. (1986). Science and Literature in the 19th Century. Macmillan.
Ciccarelli, C., & Nuvolari, A. (2015). Technical change, non-tariff barriers, and the development of Italian locomotive industry, 1850–1913. The Journal of Economic History, 75(1), 860–888.
Cohen, M. (1999). The sentimental education of the novel. Princeton University Press.
D’Amico, L., & Tabellini, M. (2017). Measuring attitudes towards immigration using local newspapers’ data and congressional speeches. Working Paper.
Davies, M. (2010). The corpus of contemporary American English as the first reliable monitor corpus of English. Literary and Linguistic Computing, 25(4), 447–464.
Desmond, A. J., and Moore, J. (1994). Darwin. WW Norton & Company.
Dubossarsky, H., Tsvetkov, Y., Dyer, C., and Grossman, E. (2015). A bottom up approach to category mapping and meaning change. NetWordS, pp. 66–70.
Enke, B. (2018). Moral values and voting. Journal of Political Economy, 128(10), 3679–3729.
Fetter, F. W. (1975). The influence of economists in Parliament on British legislation from Ricardo to John Stuart Mill. Journal of Political Economy, 83(5), 1051–1064.
Fuller, R. (2017). The book that changed America: How Darwin’s theory of evolution ignited a nation. Penguin.
Galor, O. (2011). Unified growth theory. Princeton University Press.
Garg, N., Schiebinger, L., Jurafsky, D., and Zou, J. (2017). Word embeddings quantify 100 years of gender and ethnic stereotypes. arXiv preprint arXiv:1711.08412.
Gentzkow, M., Kelly, B. T., and Taddy, M. (2018). Text as data. Journal of Economic Literature, forthcoming.
Gentzkow, M., & Shapiro, J. (2010). What drives media slant? Evidence from US daily newspapers. Econometrica, 78(1), 35–71.
Gerow, A., Hu, Y., Boyd-Graber, J., Blei, D. M., & Evans, J. A. (2018). Measuring discursive influence across scholarship. Proceedings of the National Academy of Sciences, 115(13), 3308–3313.
Gianquitto, T., & Fisher, L. (Eds.). (2014). America’s Darwin: Darwinian theory and US literary culture. University of Georgia Press.
Giuliano, P., and Nunn N. (2021). Understanding cultural persistence and change. Review of Economic Studies, forthcoming
Gramsci, A. (1948). Selections from the prison notebooks. London: The civil society reader. University Press of New England.
Gray, A. (1860). Darwin on the origin of species. The Atlantic, July issue.
Guiso, L., Sapienza, P., & Zingales, L. (2004). The role of social capital in financial development. American Economic Review, 94(3), 526–556.
Guiso, L., Sapienza, P., & Zingales, L. (2006). Does culture affect economic outcomes? Journal of Economic Perspectives, 20(2), 23–48.
Gulordava, K. and Baroni, M. (2011). A distributional similarity approach to the detection of semantic change in the google books Ngram corpus. In Proceedings of the GEMS 2011 Workshop on Geometrical Models of Natural Language Semantics, pp 67–71. Association for Computational Linguistics.
Hamilton, W. L., Leskovec, J., and Jurafsky, D. (2016a). Diachronic word embeddings reveal statistical laws of semantic change. arXiv preprint arXiv:1605.09096.
Hamilton, W. L., Leskovec, J., & Jurafsky, D. (2016b). Cultural shift or linguistic drift? Comparing two computational measures of semantic change. In proceedings of the conference on empirical methods in natural language processing. Conference on empirical methods in natural language processing, 2016b, pp. 2116–2121.
Heuser, R., & Le-Khac, L. (2011). Learning to read data: Bringing out the humanistic in the digital humanities. Victorian Studies, 54(1), 79–86.
Heuser, R. (2016). Word vectors in the eighteenth century. IPAM workshop: Cultural Analytics.
Huxley, T. (1859). Darwin on the Origin of species. The Times, 26 December: 8–9.
Jatowt, A. and Duh, K. (2014). A framework for analyzing semantic change of words across time. In Digital Libraries (JCDL), 2014 IEEE/ACM Joint Conference, pp. 229–238. IEEE.
Jelveh, Z., Kogut, B., and Naidu, S. (2014). Detecting latent ideology in expert text: Evidence from academic papers in economics. In proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP), pp. 1804–1809.
Jenkins, J., Russell, W., & Suci, G. (1958). An Atlas of semantic profiles for 360 words. American Journal of Psychology, 71(4), 688–699.
Jones, C. I. (2002). Sources of U.S. economic growth in a world of ideas. American Economic Review, 92(1), 220–239.
Kearney, C., & Liu, S. (2014). Textual sentiment in finance: A survey of methods and models. International Review of Financial Analysis, 33, 171–185.
Kelly, B., Papanikolaou, D., Seru, A., & Taddy, M. (2021). Measuring technological innovation over the long run. American Economic Review: Insights, 3(3), 303–20.
Kim, Y., Chiu, Y.-I., Hanaki, K., Hegde, D., and Petrov, S. (2014). Temporal analysis of language through neural language models. arXiv preprint arXiv:1405.3515.
Kirby, S., Dowman, M., & Griffiths, T. L. (2007). Innateness and culture in the evolution of language. Proceedings of the National Academy of Sciences, 104(12), 5241–5245.
Kozlowski, A. C., Taddy, M., & Evans, J. A. (2019). The geometry of culture: Analyzing the meanings of class through word embeddings. American Sociological Review, 84(5), 905–949.
Kulkarni, V., Al-Rfou, R., Perozzi, B., and Skiena, S. (2015). Statistically significant detection of linguistic change. In Proceedings of the 24th International Conference on World Wide Web, pp 625–635. International World Wide Web Conferences Steering Committee.
Landes, D. S. (1999). The wealth and poverty of nations, why some are so rich and some are so poor. WW Norton & Company.
Lansley, C. M. (2016). Charles Darwin-s debt to the Romantics. PhD thesis, University of Winchester.
Lévi-Strauss, C. (1963). Structural anthropology. Basic Books.
Levy, O., Goldberg, Y. and Dagan, I. (2015). Improving distributional similarity with lessons learned from word embeddings. Trans. ACL, 3.
Lin, Y., Michel, J.-B., Aiden, E. L., Orwant, J., Brockman, W., and Petrov, S. (2012). Syntactic annotations for the Google books Ngram corpus. In Proceedings of the ACL 2012 system demonstrations, pp. 169–174. Association for Computational Linguistics.
Lyons, M. (2003). New readers of the nineteenth century: Women, children, workers. In G. Cavallo & R. Chartier (Eds.), A history of reading in the West. University of Massachusetts Press.
MacKinnon, J., & Webb, M. (2018). The wild bootstrap for few (treated) clusters. The Journal of Econometrics, 21(2), 114–135.
Manovich, L. (2009). Cultural analytics: visualising cultural patterns in the era of more media. Domus March.
Markevich, A., & Zhuravskaya, E. (2018). The economic effects of the abolition of Serfdom: Evidence from the Russian empire. American Economic Review, 108(4), 1075–1117.
Masci, D. (2019). Darwin in America. Pew Research Center.
Mayr, E. (1982). The growth of biological thought: Diversity, evolution, and inheritance. Harvard University Press.
Mayr, E. (1995). Darwin’s impact on modern thought. Proceedings of the American Philosophical Society, 139(4), 317–325.
Mayr, E. (2001). The philosophical foundations of Darwinism. Proceedings of the American Philosophical Society, 145(4), 488–495.
McCloskey, D. N. (2016). Bourgeois equality: How ideas, not capital or institutions, enriched the world (Vol. 3). University of Chicago Press.
McPherson, E. G. (1942). Reporting the debates of congress.
Michalopoulos, S., & Xue, M. M. (2021). Folklore. Quarterly Journal of Economics, 136(4), 1993–2046.
Michel, J.-B., Shen, Y. K., Aiden, A. P., Veres, A., Gray, M. K., Pickett, J. P., Hoiberg, D., Clancy, D., Norvig, P., Orwant, J., et al. (2011). Quantitative analysis of culture using millions of digitized books. Science, 331(6014), 176–182.
Mikolov, T., Sutskever, I., Chen, K., Corrado, G. S., and Dean, J. (2013). Distributed representations of words and phrases and their compositionality. Advances in neural information processing systems 3111–3119.
Mokyr, J. (2010). The enlightened economy an economic history of Britain 1700–1850. Yale University Press.
Mokyr, J. (2013). Cultural entrepreneurs and the origins of modern economic growth. Scandinavian Economic History Review, 61(1), 1–33.
Mokyr, J. (2016). A culture of growth: The origins of the modern economy. Princeton University Press.
Moretti, F. (2013). Distant reading. Verso Books.
Nguyen, D., Liakata, M., DeDeo, S., Eisenstein, J., Mimno, D., Tromble, R., & Winters, J. (2020). How we do things with words: Analyzing text as social and cultural data. Frontiers in Artificial Intelligence, 3, 62.
Otis, L. (2009). Literature and science in the nineteenth century: An anthology. Oxford University Press.
Pakes, A., & Sokoloff, K. L. (1996). Science, technology, and economic growth. Proceedings of the National Academy of Sciences, 93(23), 12655–12657.
Pechenick, E. A., Danforth, C. M., & Dodds, P. S. (2015). Characterizing the google books corpus: Strong limits to inferences of socio-cultural and linguistic evolution. PloS One, 10(10), e137041.
Richards, R. J. (2013). The impact of German romanticism on biology in the nineteenth century. The impact of idealism: The legacy in philosophy and science. Cambridge University Press.
Romer, P. M. (1990). Endogenous technological change. Journal of political Economy, 98(5), S71–S102.
Roodman, D., Nielsen, M. Ø., MacKinnon, J. G., & Webb, M. D. (2019). Fast and wild: Bootstrap inference in stata using boottest. The Stata Journal, 19(1), 4–60.
Roth, S. (2014). Fashionable functions: A Google Ngram view of trends in functional differentiation (1800–2000). International Journal of Technology and Human Interaction, 10(2), 35–58.
Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986). Learning representations by back-propagating errors. Nature, 323(6088), 533–536.
Schiller, R. J. (2017). Narrative economics. American Economic Review, 107(4), 967–1004.
Scholnick, R. (2015). American literature and science. University Press of Kentucky.
Sen, A. (2004). How does culture matter? In V. Rao (Ed.), Culture and public action. Orient Blackswan.
Spolaore, E. (2020). Commanding nature by obeying her: A review essay on Joel Mokyr’s a culture of growth. Journal of Economic Literature, 58(3), 777–792.
Spolaore, E. (2014). Culture and economic growth. Edward Elgar Publishing.
Stephan, P. E. (2012). How economics shapes science (Vol. 1). Harvard University Press.
Thompson, B., Roberts, S. G., & Lupyan, G. (2020). Cultural influences on word meanings revealed through large-scale semantic alignment. Nature Human Behaviour, 4(10), 1029–1038.
Turner, F. (2010). From counterculture to cyberculture: Stewart Brand, the whole earth network, and the rise of digital utopianism. University of Chicago Press.
Whorf, B. L. (1956–2012). Language, thought, and reality: Selected writings of Benjamin Lee Whorf. MIT press.
Wilkens, M. (2015). Digital humanities and its application in the study of literature and culture. Comparative Literature, 67(1), 11–20.
Williamson, O. (2000). The new institutional economics: Taking Stock, looking ahead. Journal of Economic Literature, 38(3), 595–613.
Winans, R. B. (1975). The growth of a novel-reading public in late-eighteenth-century America. Early American Literature, 9(3), 267–275.
Yin, Y., Dong, Y., Wang, K., Wang, D., & Jones, B. (2021). Science as a public good: Public use and funding of science (No. w28748). National Bureau of Economic Research.
Acknowledgements
We thank Dora Costa, Ryan Heuser, Graeme Hirst, Xander Manshel and Yang Xu for their suggestions; and participants to presentations at Brown University, the University of Toronto, the University of Munich, the NBER Productivity Lunch, the 2018 REER Conference at Georgia Tech, the Workshop in Memory of Luigi Orsenigo at Bocconi University, the 2019 NBER Summer Institute, the 2018 Academy of Management Annual Meetings, the Virtual Economic History Seminar Series, and Monash University for their helpful feedback.
Funding
We gratefully acknowledge the financial support of the National Bureau of Economic Research through the Innovation Policy Grants Program.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
About this article
Cite this article
Giorcelli, M., Lacetera, N. & Marinoni, A. How does scientific progress affect cultural changes? A digital text analysis. J Econ Growth 27, 415–452 (2022). https://doi.org/10.1007/s10887-022-09204-6
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10887-022-09204-6