In the new expanded edition of our book, The Privileged Planet: How Our Place in the Cosmos Is Designed for Discovery, Jay Richards and I argue that total solar eclipses are part of a wider pattern wherein the conditions for planetary habitability correlate with the conditions for doing science. A recent study, if confirmed, adds a layer by showing that total solar eclipses are a kind of “multiplying factor” in spurring economic growth and furthering technological advancements.

With the intriguing title “Solar Eclipses and the Origins of Critical Thinking and Complexity,” the scholarly paper was published in the May 2024 issue of The Economic Journal. Appropriately, this was just a month after the great American solar eclipse of 2024. The authors, Anastasia Litina and Èric Roca Fernández, are from universities in Greece and France, respectively. 

A Trigger for Curiosity

They explore the relationship between curiosity, triggered by natural phenomena such as solar eclipses, and economic development in pre-modern societies. They hypothesize that exposure to inexplicable events prompted curiosity and critical thinking as people attempted to understand these mysteries, thereby enhancing human capital, technological progress, and ultimately economic growth.

The study focuses on solar eclipses as a trigger for curiosity due to their unique characteristics. Solar eclipses are impressive yet harmless natural events that occur infrequently enough to remain mysterious but regularly enough to allow for intergenerational knowledge transmission. The authors argue that pre-modern societies more exposed to such phenomena gained a comparative advantage in critical thinking as people attempted to understand these mysteries.

The research combines multiple datasets to investigate this hypothesis:

1. The Ethnographic Atlas, which provides information on over 1,000 pre-modern societies.
2. The Seshat database, which contains details on historical polities.
3. Country-level historical data from various sources.
4. A novel dataset on historical scientific occupations constructed by the authors using Wikidata.

The main empirical strategy employs cross-sectional regressions with the number of visible solar eclipses as the key independent variable and various outcomes as dependent variables. The authors control for a wide range of geographic and climatic factors that may influence development, including elevation, ruggedness, malaria prevalence, potential caloric yield, latitude, temperature, precipitation, and major habitat types. For panel data (also called longitudinal or cross-sectional time series data), they use fixed effects models to account for time-invariant characteristics of polities and general trends over time.

Key findings of the study include:

1. Economic Development: Societies exposed to more solar eclipses show higher levels of economic development, as measured by population density, settlement patterns, jurisdictional hierarchy, class stratification, and political integration.
2. Human Capital: There is evidence of greater human capital in eclipse-exposed societies. This includes more prevalent strategy games, more developed writing systems, and more sophisticated understanding of eclipses as reflected in folklore.
3. Technological Sophistication: The number and complexity of tasks performed in a society, as well as various technological indices, are positively associated with eclipse exposure.
4. Curiosity in Folklore: Societies with more eclipse exposure have folklore that contains more references to eclipses, calendars, and curiosity-related concepts.
5. Individual-Level Effects: Using the novel dataset constructed from Wikidata, the authors find that individuals who observed a solar eclipse during childhood were more likely to pursue scientific occupations later in life.

Potential Confounding Factors 

The authors conduct several robustness checks to address potential confounding factors:

1. Lunar Eclipses: They repeat analyses using lunar eclipses, finding similar but weaker effects. This supports the idea that the visibility and impressiveness of the phenomenon matter.
2. Area Controls: They control for homeland area in various ways to address the mechanical relationship between area and number of eclipses visible.
3. Religion: They examine the relationship between eclipses and religious concepts/occupations, finding no significant associations. This suggests the effects operate through curiosity rather than religious channels.
4. Placebo Tests: Using concepts unrelated to human capital or thinking (e.g., colors, inanimate objects), they find no relationship with eclipse exposure.

The authors acknowledge some limitations of their approach. The cross-sectional nature of much of the data limits causal interpretation. The Seshat database uses interpolated data, requiring cautious interpretation. The individual-level occupation data from Wikidata may have selection biases.

The paper includes several appendices with additional details:

Appendix A provides a more detailed conceptual framework, including a theoretical model that incorporates destructive and non-destructive shocks. This model helps to rationalize why harmless but inexplicable events like solar eclipses might be particularly effective in spurring curiosity and development.

Appendix B explains methods of eclipse prediction in pre-modern times, distinguishing between predicting the timing of eclipses (which was possible with some accuracy) and their local visibility (which was very difficult). This information supports the authors’ argument that eclipses remained mysterious enough to spur curiosity even as societies developed some ability to predict them.

Appendix C presents additional robustness checks and extensions. These include analyses with lunar eclipses, alternative specifications to address the relationship between area and eclipses, and examination of the cumulative effects of eclipses over time. The authors also investigate the relationship between solar eclipses and scientific discoveries in Greece and India over time.

Appendix D provides detailed descriptions of all variables and data sources used in the analysis. This includes information on how the eclipse data was compiled, how various measures of economic development and human capital were constructed from the Ethnographic Atlas and other sources, and how the individual-level data on historical occupations was collected and coded.

The Role of Curiosity

This paper presents a novel hypothesis about the role of curiosity in long-run economic development and provides a range of empirical evidence consistent with this hypothesis. While the nature of the data and analysis precludes definitive causal claims, the consistency of results across multiple datasets and specifications suggests an intriguing relationship between exposure to mysterious natural phenomena and economic, technological, and intellectual progress in pre-modern societies.

The authors argue that their findings shed light on the early origins of human cognition and contribute to our understanding of growth among societies that had not yet mastered science. By highlighting the importance of curiosity as a natural predecessor to human capital accumulation, they provide a new perspective on the roots of comparative development across societies.

Do I Believe These Findings? 

In the area that I know something about, namely ancient and modern astronomical observations, I can say that the authors are careful and accurate. For example, it was smart for them also to analyze lunar eclipses, which are more frequently observed and are more predictable than total solar eclipses, removing some of the “wow” factor. Societies would have already developed explanations for a disappearing moon during the monthly new moon phase, and the moon appears red when it is rising and setting. For these reasons, one would expect the correlation between lunar eclipses and economic development to be weaker, which is what the authors found to be the case.

I’m not a social scientist or an economist, and I haven’t tried to follow the authors’ analyses in detail. But I think there is something here.