The past is never dead. It’s not even past.
Why are leaders chosen in free and fair elections in some countries while in others they inherit power, seize power, or engineer victory in rigged elections? Why is political speech unregulated in some countries and tightly controlled in others? Why are opposition leaders free to express their views in some countries while in others they languish in jails, are subject to harassment or assassination, or go into exile?
There are many possible answers to the question of regimes and a great deal of work has accumulated on the subject.Footnote 1 Yet, for all its richness and sophistication this body of work is narrowly circumscribed in time. Most studies focus on the postwar era. A few peer into the nineteenth century. Prior to that, work is thin and tends to be focused on particular historical and regional contexts.
As a point of departure, we assert that patterns of democracy and dictatorship observable across the world today are not simply the product of recent history. They have deeper roots.
In this book, we concern ourselves with those distal causes. This means that we will be dealing with factors that are structural rather than eventful. Causes and effects are separated by centuries, in some cases by millennia. And the pathways from X to Y are long and tortuous. Some time will be required before we can hope to unravel them.
Without further ado let us lay out our arguments, which center on geography and long-term patterns of diffusion. Following that, we shall review alternate explanations, introduce our data and methodology, and define key terms. By way of conclusion, we provide a roadmap of the book.
Harbors
Our geographic argument centers on access to the sea. This is afforded by natural harbors, which may be located on the coast or on navigable rivers that flow, unimpeded, to the coast. Harbors enhance mobility – of people, goods, capital, and ideas.
The extraordinary connectivity of harbor regions affects economic development (nurturing trade, urbanization, human capital, and economic growth), the structure of the military (away from standing armies and toward naval power), statebuilding (toward smaller states, confederations, or overseas empires with semi-autonomous colonies), and openness to the world (through trade, migration, tourism, religious pilgrimages, and conquest).
Each of these developments shifts the balance of power between rulers and citizens. As a result, areas situated close to harbors are more likely to evolve in a democratic direction than areas surrounded by large landmasses or inaccessible coasts.
As is the case with most geographic arguments, there is no identifiable point of onset. We surmise that the importance of harbors increased as shipping technology improved and ships displaced overland travel as the dominant mode of communication, travel, and trade. This was a long process, occurring at different speeds in different parts of the world. In Asia, it was well underway in the premodern era; in many parts of the New World it is a comparatively recent phenomenon. Accordingly, we expect that the influence of harbors on political institutions unfolded over a very long period of time in regions of the world where shipping has a long history, and over a shorter period of time in regions where shipping is a more recent technological development. Everywhere, we expect some attenuation as other forms of transport, travel, and communication supplant the primacy of shipping at the end of the twentieth century.
European Ancestry
Although there was considerable variety in political institutions throughout the premodern world, only one area developed systems of representation through parliaments. Defined in this fashion, democracy (i.e., representative democracy) was invented in Europe.
Beginning about 1500, with the advent of sailing vessels capable of circumnavigating the globe, Europeans began to populate the distant abroad, often in the shadow of colonial conquest. By 1900, they could be found virtually everywhere, in varying proportions. We argue that the resulting ratio of Europeans to non-Europeans – which we call European ancestry – structured the fate of regimes around the world.
Larger numbers of Europeans meant greater exposure to the idea of democracy for non-Europeans, transmitted through schools, churches, newspapers, radio, direct contact with settlers, settler societies abroad (such as the new hegemon, the United States), and the metropole. Larger numbers of Europeans brought additional features that we refer to as the infrastructure of democracy including education, advanced transport and communications, urbanization, a nation-state form of political organization, property rights, capitalism, and wealth. Larger numbers of Europeans, finally, changed incentives. For Europeans, it meant that they were likely to control political outcomes under democratic rules. Consequently, they had an incentive to invest in democratic institutions and were more likely to respond positively to claims by slaves and indigenes for inclusion.
It is important to appreciate that democracy in the first instance meant democracy for Europeans. Where they were numerous they established some form of representative democracy, usually with restrictions (de jure or de facto) limiting suffrage and officeholding to those of European descent. Where they were in the minority they were more reticent about popular rule and often actively resisted democratization. And where Europeans were entirely absent, the concept of representative democracy was unfamiliar and the practice undeveloped. In this fashion, Europeans served as agents of diffusion as well as agents of constraint.
We expect that the impact of European ancestry on democracy applied with increasing force as European influence over the world grew, reaching its apogee during the twentieth century – at which point most of the world was directly controlled by, or under the regional hegemony of, a European power or transplanted European settlers. By the turn of the twenty-first century, the influence of Europe and various “neo-Europes” across the world was declining and the concept of representative democracy no longer a European preserve. Thus, we conjecture that the impact of European demography on regimes increased monotonically across the early modern and modern eras with some attenuation in recent decades (and expected further attenuation in the future).
Synthesis
These two arguments are interconnected. That representative democracy developed on the European continent is, in part, a product of its aqueous geography. And that Europe discovered the world, rather than the reverse, was also a product of an aqueous geography. Harbors foster an outward orientation, as trade and seafaring come naturally to civilizations bordering the sea. From this perspective, it is not surprising that the age of globalization was led by Europeans rather than Asians or Africans. Finally, we must make allowances for how Europe traversed the world. It was of course through ships. And this means that areas (outside Europe) well endowed with natural harbors were the first to be discovered, the most likely destinations for European migrants, and the areas most closely linked to Europe through trade and cultural exchange. Harbors connect.
To summarize, we argue that natural harbors serve as a prime mover of political institutions in the modern era, having a direct effect on the development of representative democracy as well as an indirect effect through the European diaspora. This historical model is illustrated schematically in Figure 1.1.
Although the schema is admirably compact, it is important to signal that the mechanisms at work in our framework – represented by the solid arrows in Figure 1.1 – are quite complex. We are hard put to identify a parsimonious theory, one that reduces to a few fundamental precepts that operate consistently across millennia on a global scale. So far as we can tell, natural harbors and European ancestry mattered for many reasons.Footnote 2
Extant Work
Big arguments about big subjects do not materialize out of thin air. They arise from an ongoing conversation with the scholarly literature. In this case, the conversation extends beyond disciplinary boundaries, encompassing work by historians, archeologists, classicists, as well as historically inclined social scientists. The Bibliography includes nearly two thousand articles and books, with additional sources on related subjects listed in Appendix E (online at Dataverse). Here, one will find precedents for nearly every claim made in this long book.
Yet, the schema represented in Figure 1.1 is not simply a restatement of accepted wisdom. Indeed, most of the secondary studies cited in this book are not focused on democracy per se but rather on the histories of specific regions or adjacent subjects such as maritime history or colonial history.
To our knowledge, no study of democracy has identified a causal role for natural harbors, though a few historical studies emphasize the liberating role of the sea and of maritime culture more generally (e.g., Abulafia Reference Abulafia2019; Lambert Reference Lambert2018). Likewise, no study has systematically explored the role of European ancestry as a conditioning factor in modern regimes, though the idea is presented briefly in Hariri (Reference Hariri2012) and the impact of European settlement on economic development outside Europe is widely acknowledged.Footnote 3 In these respects, our arguments may claim novelty.
Alternate Explanations
When one paints on a large canvas, as we do in this book, there may be an unspoken expectation that everything about the topic will be revealed. That is not the case here. The schema presented in Figure 1.1 is far-reaching but by no means comprehensive. Democracy is affected by myriad factors, many of which have nothing to do with natural harbors or European ancestry. Lots of things are going on.
How do these other causal factors fit with our own explanation? This is important as a matter of clarity and also as an issue in causal inference. Note that when dealing with observational data (where the treatment of theoretical interest is not randomized) every rival hypothesis poses a potential problem of confounding.
Even if rival explanations are orthogonal (non-confounding), we still want to know something about their relative impact and robustness. Some factors are trivial and others substantial. Some are robust to a wide variety of measures, specifications, and estimators; others are more delicate, depending upon stronger assumptions. Unless we test alternate causal factors together with our own, and unless these tests are directly comparable, we cannot assess these issues. Cumulation requires standardization.
For the narrow reason of reaching causal inference, and the broader rationale of building general knowledge, we take an inclusive approach to alternate explanations. These are introduced briefly and tested against available evidence from the modern era, wherever possible. Of course, we do not have the space to discuss each theory at great length or to conduct extensive tests of every possible hypothesis, and many theories have never been operationalized on a global scale. Accordingly, these tests – summarized in the next section – should be regarded as probative.
A Brief Survey
From extant work on geography we can identify numerous factors that might influence institutional development. These include climate, irrigation, agricultural potential, mountains, and islands.Footnote 4 Each can be operationalized in a variety of ways, leading to a large number of testable hypotheses. Our analyses show that distance from the equator is a strong predictor of democracy in the modern era, though it should be noted that this geographic factor appears to operate almost entirely through European ancestry. Other geographic factors are not as robust. These theoretical and empirical questions are vetted in Chapter 12.
Extant work on European diffusion has focused on several pathways from Europe to the world. Principal among these are colonialism and Protestantism.Footnote 5 These factors can be measured in a variety of ways, offering a variety of testable hypotheses. Our analyses confirm that Protestantism is a strong predictor of democracy in the modern era, though questions may be raised about its causal status. These issues are discussed in Chapter 13.
In addition, there is a large residual category of potential causal factors that might be described as economic, institutional, or cultural. This includes the timing of transitions to sedentary agriculture, modernization, inequality, labor scarcity, patterns of marriage and family, feudalism, parliaments, Christianity, Catholicism, state size, state–society relations, and ideas. Our global analyses (which, of necessity, leave aside Europe-specific causes) show some evidence in favor of agricultural transitions, kinship intensity, and state history, though the relationship of these factors to democracy is sensitive to choices in specification. In any case, most of these factors are downstream from our framework, illustrated in Figure 1.1. Indeed, some of these factors are identified as causal mechanisms in our framework. These issues are discussed in Chapter 14.
Of course, this does not exhaust the panoply of potential causes of democracy. A short tour of recent work would include additional factors such as civil society (e.g., popular protest, civil organizations, trust), constitutional rules (e.g., presidentialism, electoral systems, federalism, decentralization), elites (e.g., divisions, alliances, pacts, leadership succession), macroeconomics (e.g., inflation, unemployment, per capita GDP growth, trade), human capital (e.g., education, health, the age distribution of the population), social diversity (ethnic, linguistic, or religious), and international influences (e.g., international organizations, foreign aid, foreign intervention, short-term diffusion).Footnote 6 We regard these factors as proximal and thus downstream from our argument.
To conclude, our empirical tests – presented in Part IV of the book – suggest that one geographic factor (equator distance) has an effect on modern democracy comparable to the impact of natural harbors and that one alternate pathway from Europe (Protestantism) is as strongly correlated with modern democracy as European ancestry. Other factors do not seem to be robust predictors of democracy in the modern era or are downstream from the factors of theoretical interest in this study. Accordingly, our two-part theoretical framework (if true) adds substantially to what is currently known about the long-term causes of democracy.
In Search of Distal Causes
We turn now to methodological considerations. How can one test non-manipulable causes that operate slowly, almost imperceptibly, over the course of centuries?
Let us begin with the most common species of explanation for the rise of democracy. This sort of explanation focuses on features specific to Europe, e.g., classical Greece, the Roman Empire and its dissolution, feudalism, the Catholic Church, family and marriage patterns, a culture of individualism, a spirit of liberty, and so forth.Footnote 7 These sorts of explanations generally offer limited variability within Europe, or none that could be regarded as the basis for a strong empirical test given the interdependent relationship of European states and cultures. Beyond Europe, one is left with a binary contrast between Europe and everywhere else, which is not very revealing given that Europe is different from other regions in many ways, each of which constitutes a potential confounder. Explanations that feature something distinctive about Europe are therefore not very tractable.
Our theory is more falsifiable. All three nodes of theoretical interest – natural harbors, European ancestry, and democracy – are measurable on a global scale across the modern era and offer considerable variability across regions and within regions. This facilitates a wide variety of empirical tests.
Of course, these tests are not as conclusive as experimental tests, where the treatment of theoretical interest is manipulated by the researcher. By way of entrée, it may be helpful to identify the hard and soft points of the empirical analyses to come.
As a feature of geography, natural harbors are causally exogenous. There is no risk that the outcome could impact the cause, or that other causes could affect both. However, there is a risk that any association found between harbors and democracy might be spurious if the former serves as a proxy for some (unmeasured) factor; hence the importance of specification tests that include all geographic factors that could plausibly impact the long-term development of political institutions. Doubts would be further eased if one could trace the impact of harbors on democracy through a set of observable mechanisms. Unfortunately, a long and tortuous route connects geography to regimes in the modern era. Some of these putative pathways can be measured and tested. But since natural harbors are static, and affect social relations slowly over the course of millennia, we cannot process-trace their relationship to democracy.
European ancestry changes over time and bears a more proximal relationship to the outcome; as such, the mechanisms are less mysterious. However, there are questions about the exogeneity of this causal factor that resist easy answers, despite our attempts to model the relationship with instrumental variables.
These are the problems of causal inference that we wrestle with in subsequent chapters. In this section, we review our general approach. This is complicated ground and we cannot hope to cover it in any detail. Nonetheless, it is important to foreshadow the evidence that lies ahead.
Data
Data for this project is gathered from a variety of sources. Port locations in the ancient world are drawn from The Catalogue of Ancient Ports and Harbours compiled by Arthur de Graauw (Reference de Graauw2017), who culls from the Barrington Atlas, the Pleiades project, the Digital Atlas of the Roman Empire, and many other primary and secondary sources. In the modern era, we rely on Lloyd’s List (Ducruet et al. Reference Ducruet, Lee and Ng2018) and several editions of the World Port Index (US Navy Reference Putterman and Weil1953; NGIA 2017). Helpfully, all of these sources provide the precise GIS location of ports, allowing for fine-grained analyses. We note parenthetically that this project makes extensive use of maps, all of which employ a standard WGS 84 geographic coordinate system (EPSG 4326).
Maritime history is tracked through inland transport rates, Soundtoll traffic, trade, ship departures from Europe to Asia, ship tonnage and speed, number of mariners, and migration rates – for which we draw on a wide variety of historical sources. We explore the relationship of ports to early cities, city foundings, contemporary megalopolises, urbanization, colonial capitals, state size, ethnic diversity, and democratic values (the latter with data drawn from the latest round of the World Values Survey, which provides the GIS location of each interview). We also construct a ratio measure of the degree to which states relied on naval or land forces. And we collect original data on the share of Europeans in colonies and countries around the world from 1600 to the present, the largest database of its kind.
We devote special care to the outcome of interest. To measure democracy in the early modern era we enlist data on legislatures in British North America (Graham Reference Graham2018), the British Caribbean (from various sources), and across all European colonies (Paine Reference Paine2019). Within Europe, we enlist data on parliaments extending back to the Middle Ages (Abramson, Boix Reference Abramson and Boix2019). And across ethnic groups we utilize the Ethnographic Atlas (Murdock Reference Murdock1967), from which a simple measure of democracy is derived from procedures used to select headmen in bands, tribes, and small states.
Even so, data sources on political institutions in the premodern era are quite limited. Consequently, our analysis of democracy during this period rests primarily upon traditional historical materials of a qualitative nature – the vast secondary literature produced by historians, archeologists, anthropologists, classicists, and a few historically minded social scientists. Here, we adopt a case study approach focused on within-region comparisons (e.g., across Europe or across South Asia) or within-colonizer comparisons (e.g., across British colonies or across Spanish colonies).
In the modern era, it is possible to measure democracy on a global scale with greater precision and nuance. Our benchmark measure is the Polyarchy index from the Varieties of Democracy (V-Dem) project (Coppedge, Gerring, Glynn, et al. Reference Coppedge, Gerring, Glynn, Knutsen, Lindberg, Pemstein, Seim, Skaaning and Teorell2020; Teorell et al. Reference Teorell, Coppedge, Lindberg and Skaaning2018). Although V-Dem stretches back to 1789 and extends to all sizeable countries and some colonies, it omits micro-states and many colonies. To generate more comprehensive coverage we interpolate missing data with other sources.
As secondary measures of democracy in the modern era we enlist the Polity2 index from the Polity IV project (Marshall, Monty, Jaggers Reference Meier2016) and the Lexical index of electoral democracy (Skaaning et al. Reference Skaaning, Gerring and Bartusevičius2015) Occasionally, we home in on specific aspects of democracy like male suffrage (measured by V-Dem for the twentieth century and Bilinski Reference Bilinski2015 for the nineteenth century).
Statistical Tests
Methods of analysis follow from what is plausible with the data that is available. In many cases, it is possible to conduct statistical tests of hypotheses using colonies, countries, cities, grid-cells, or individuals as units of analysis.
A cross-sectional research design is preferred, as the predictors of theoretical interest are static (in the case of natural harbors) or sluggish (in the case of European ancestry). In the latter setting, we supplement cross-sectional models with fixed-effect models.
To control for time-effects across the observed period we include time dummies and cluster standard errors according to the unit of observation. As an additional test we focus on a single point in time close to the present (2000 ce), eliminating the problem of temporal autocorrelation but with considerable information loss (since all other years are dropped). Potential problems of causal identification are probed with instrumental variables that attempt to model assignment to treatment. Myriad specification tests are enlisted in order to probe different assumptions about the data-generating process. Descriptions of each variable and its source are placed in Appendix A and descriptive statistics in Appendix B (online at Dataverse).
Our approach to causal identification is eclectic. We recognize that there is rarely a single “correct” approach to the analysis of observational data. At the same time, there are a limited number of plausible approaches, each of which rests on a somewhat different set of assumptions about the data-generating process. If the results of these tests are relatively stable, one can justify greater confidence that the relationship might be causal. But we should not kid ourselves. Macro-level causal relationships occurring across centuries can never be estimated precisely and always require supporting theory and assumptions.
Assuming that a factor plays a causal role, its impact on democracy may be understood along several dimensions. First, we present graphs of estimated values for the outcome as the variable of theoretical interest varies. Second, we examine changes in impact over time using rolling regressions. Where possible, we examine the impact of a factor when grid-cells are aggregated differently, which helps to address potential problems posed by interference across units (stable unit treatment assumption, aka SUTVA) as well as the modifiable area unit problem.
In comparing the strength of relationships across rival predictors we employ several techniques. First, we take note of what is called (loosely) “statistical significance,” as registered in p values and t statistics. While prone to abuse, these statistics are informative if we bear in mind their assumptions and their limitations. Second, we compare the impact of a set of variables on total model fit, understood in linear models with the R2 statistic. None of these approaches is sufficient on its own. However, taken together, they give us a fairly nuanced sense of how valuable different predictors might be in explaining regime types in the modern era.
Qualitative and Quantitative Evidence
Not all our arguments can be subjected to formal statistical tests. This is perhaps true for any large project but it is especially true for the present project. Because of the nature of the available evidence, chapters focused on the premodern era incline towards qualitative data while chapters focused on the modern era incline towards quantitative data. Let us briefly review the strengths and weaknesses of these rather different styles of evidence as they pertain to the present study.Footnote 8
Large-N cross-case analyses in this book rely on variables that are either objective in nature or coded by others, reducing the scope of potential authorial bias. Large-N models are also capable of handling stochastic error. Finally, findings are apt to be generalizable since our global samples contain nearly all the sizeable polities in the world, functioning more like a census than a sample.
Small-N comparisons of a most-similar variety eliminate many background factors that might serve as confounders in a large-N analysis. They also allow us to look closely at the development of those cases through time, which may reveal processes (i.e., causal mechanisms) that are not visible in statistical analyses. However, they are subject to biases in the secondary literature, upon which we are reliant. It should be noted that the sources consulted for this project are mostly English-language, which limits our reach. Material of this nature also requires greater authorial judgment and is subject to questions about generalizability.
Ideally, these methodologies complement each other, shedding light on different aspects of our research question and invoking different assumptions. In this instance, neither is very solid without the other.
Key Terms
As a final act of theoretical exposition, we must define several key terms that recur in this chapter and the chapters that follow. This includes democracy, the state, and various historical periods (premodern, modern, etc.).
Our proposed definitions are not unusual. However, because these concepts are understood in many ways it is important to define them formally so that readers know what we have in mind. Doing so will also help to clarify the meaning and scope-conditions of our theory.
Democracy
Generally understood, democracy means rule by the people. In ascertaining the degree to which people rule we distinguish two dimensions: membership (the number of people living within a polity who are granted citizenship rights) and accountability (the control they have over policymaking). Membership varies, in principle, from one (only the ruler is empowered) to all (full citizenship rights are granted to all permanent residents). Accountability varies from none (there is no accountability whatsoever) to full (whatever that might mean). Democracy is thus a matter of degrees.
When we characterize a polity as democratic we mean that the polity in question was more democratic (autocratic) than other contemporaneous polities. We realize that no sizeable polity in antiquity bestowed full membership on all permanent adult residents, and thus none would be considered a democracy today. Nonetheless, to say that ancient Athens was a democracy is meaningful if this statement is understood relative to polities existing at that time. Similar caveats apply to our use of the term democracy prior to the twentieth century, a period when suffrage was generally restricted along gender and/or racial lines.
We recognize, finally, that two quite different forms of popular rule qualify as democratic.
Direct democracy means that citizens make policy decisions themselves, usually through popular assemblies or consultative bodies. From what we can tell, these mechanisms were widespread in the premodern era. Although classical Athens is the best-known example there is no reason to suppose that it was unique. Citizens probably played important deliberative and decision-making roles in other city-states, as well as in bands and tribes without statelike forms of political organization. We show in Chapter 2 that this sort of democracy was widespread throughout the premodern world, and by no means limited to Europe.
Representative democracy means that citizens rule through representatives who are accountable to them. This entails a representative assembly (a parliament), regular elections to that body, and a constitution prescribing formal rules of operation and perhaps specific rights reserved for citizens. Here, Europe played a leading role. During the Middle Ages, the concept and practice of representative democracy developed in various places across the European subcontinent. In the modern era, it spread throughout the world and is now considered to be essential for the achievement of democracy in statelike political organizations. Since our theoretical objective is primarily focused on the modern era and on statelike entities, in the chapters that follow we usually refer to democracy as a set of representative institutions.
Further discussion of the conceptualization and measurement of democracy, and its applicability to regimes past and present, can be found in Chapter 2.
States
The framework illustrated in Figure 1.1 applies to statelike political organizations. These organizations have a single locus of power (a capital), sovereignty or semisovereignty (some degree of self-rule, as enjoyed by most overseas European colonies), a recognized territory that the government controls, and a governmental apparatus that persists from one ruler to the next.
This definition may be regarded as an ideal-type, which existing polities reflect to varying degrees. Inevitably, there are problems of operationalization. When does a chiefdom take on the attributes of a state? When does a state lose that designation, due to deficits in one or more of the foregoing attributes? There are plenty of borderline cases that would be difficult to code even with full information. As one moves back in time, information becomes scarcer and details are even harder to resolve.
Nonetheless, the distinction between states and non-state or pre-state entities such as bands, tribes, and small chiefdoms is crucial. Our framework does not purport to explain modes of organization or degrees of democracy within the latter.Footnote 9
Periods
Because we are examining vast stretches of history it is important to make some distinctions among periods. Although these distinctions are arbitrary, and much blurrier around the edges than the following dates imply, they are nonetheless indispensable. Accordingly, we adopt the following conventions:
Premodern: from the first civilizations to 1789 or first European contact
◦ Ancient: from the first civilizations to 800
◦ Medieval: from 800 to 1500
◦ Early modern: from 1500 to 1789 or first European contact
Modern: from 1789 (or first European contact) to the present
Contemporary: from 1960 to the present
Note that the dividing line between modern and premodern is sometimes defined by the French Revolution (1789) and sometimes by the cataclysmic encounter between Europeans and other peoples (which occurred in different places at different times and with varying severity). In the latter instance, “premodern” refers to societies as they existed, or were thought to exist, prior to European contact. We trust that these varying usages will be clear from context.
A Roadmap
Chapter 2 explores the meaning and measurement of democracy through the ages and offers a brief survey of its manifestations. As such, it sets the stage for arguments to follow.
Part II of the book focuses on maritime geography. In Chapter 3, we explain how the presence of natural harbors might affect the long-run prospects of democracy. In Chapter 4, we discuss maritime history, the history of ports, and our strategy for identifying natural harbors, culminating in our key empirical variable – natural harbor distance. The final section discusses an obstinate units-of-analysis problem. Chapter 5 explores the role of ocean exposure on the long-run development of political institutions by looking at each of the major regions of the world with statelike polities in the premodern era. Chapter 6 probes the relationship of natural harbors to democracy in a series of crossnational tests, focused mostly on the modern era. Chapter 7 offers an empirical exploration of possible mechanisms at work in the relationship between harbors and democracy.
Part III focuses on the diffusion of democracy from Europe. Chapter 8 lays out our argument for how European ancestry influenced regime types around the world. Chapter 9 discusses issues of conceptualization and measurement – who qualifies as European and how their numbers can be estimated in colonies and countries around the world and through time. Chapter 10 explores the colonial and post-colonial era, focusing on particular colonizers and regions. Chapter 11 analyzes the relationship between European demography and democracy in a global context.
Note that Parts II and III are organized in a parallel fashion. The first chapters (3 and 8) introduce each theoretical framework and are relatively short. Readers wishing to know about our arguments can head straight to these chapters. Later chapters deal with conceptualization and measurement (Chapters 4 and 9), case studies (Chapters 5 and 10), and crossnational empirics (Chapters 6 and 11).
Part IV explores alternate explanations for the deep history of modern democracy. Chapter 12 focuses on geographic factors including climate, irrigation, agriculture, mountains, and islands. Chapter 13 considers alternate pathways of European influence including colonialism (especially English), religion (especially Protestant), and language (understood as including any European language). Chapter 14 explores a wide variety of causal factors centered on economics, institutions, and culture.
Part V is the concluding section of the book. Chapter 15 summarizes the arguments and evidence presented in preceding chapters. Following that, we come to terms with an issue looming over these arguments, which may be viewed by some readers as “deterministic.” Chapter 16 moves beyond the arguments and evidence of the book, sketching a broader vision. There, we argue that democracy is advanced when human connectedness is enhanced. This process includes many forms of transport and communications (not just shipping) and involves migration and travel on everyone’s part (not just Europeans).
Appendices offer further details. Appendix A defines all variables used in the book and their sources. Appendix B provides descriptive statistics. Appendix C provides extensive lists of early cities and their port status, European colonies and their relationship to ports, and coastal and inland cities and their assumed levels of diversity. Appendix D explores additional outcomes connected to the quality of governance that might be affected by natural harbors and European ancestry. Appendix E shows data sources and historical coding of European ancestry for each country, colony, or dependency from 1600 to 2019. Appendix F lists sources cited in the previous appendices and also additional sources – not cited in the Bibliography – that provide background information pertaining to various sections of the book. Because of their ungainly size Appendices B–F are posted online in Dataverse, along with replication materials.
