I now turn from describing and analyzing events to the more challenging task of interpreting and explaining them.

Part 3 addresses expansion. The central question is why, from the fifteenth century onward, a few west European states governed so many lands and peoples in so many parts of the world.

The subject’s vast scope and complexity and the variety of forms expansion took in specific situations pose formidable obstacles to efforts to devise a theory of European imperialism. Such a theory cannot be simultaneously comprehensive in scope, accurate in accounting for all the story’s details, and parsimonious in identify­ing causal factors. My task is not to account for every event described in the preced­ing chapters but rather to make sense of broad historical patterns. It is not to search for one factor explaining everything but to identify several factors whose interaction and cumulative impact go a long way toward making broad patterns comprehensi­ble. For reasons given earlier, the search is for conditions conducive to overseas expansion, since sufficient conditions cannot be found.

A theory of European imperialism is strengthened by being neither too re­stricted nor too broad in temporal and spatial coverage. Is its time span sufficiently broad to account for several centuries of overseas rule? Answering this question entails a search for factors prominent in both expansionist phases, not just one. Is spatial coverage sufficiently comprehensive to account for parallel empire building by several European states? Answering this question entails a search for features widely shared throughout western Europe, not features prominent in a few metro­poles but absent in others.

Does the theory avoid covering too long a span? Specifically, does it help explain not only why Europeans formed overseas empires from phase 1 onward but also why they did not do so in earlier centuries? The best way to answer this question is to focus on early phase t to see if changes in the fifteenth and sixteenth centuries significantly increased European capacity and will for effective aggression overseas.

Does the theory avoid overshooting the mark spatially? That is, does it explain not only why Europeans founded overseas empires but also why other societies with the maritime capacity to do so did not? The best way to answer this question is to compare western Europe with Arab-speaking peoples and China around the start of phase i. A feature present in Europe but less prominent or absent in the other two leading candidates for global dominance helps account for the direction world his­tory actually took.

So many factors vie for inclusion as important contributing causes that the search for explanation threatens to become unmanageably complex.¹ If some factors can be discounted for failure to pass the “both phases” temporal comprehensiveness test, then the task becomes more tractable. Expansion occurred in the preindustrial era, during the first Industrial Revolution of iron, textiles, and steam power, and during the second industrial revolution of steel, chemicals, and electricity. Expansion took place when capitalist institutions were in nascent form and when they reached the advanced “finance capital” stage described by Lenin; when Europe produced little for export and when it faced periodic overproduction crises. Colonies were founded during periods of intense intra-European warfare and decades of unusually peaceful interaction; during periods of rising and falling prosperity; before national­ist ideas gained currency and after they became popular; when settlers were integral to the process and when they were absent. It follows that a theory of European imperialism cannot plausibly identify industrialization, advanced capitalism, over­production, war, a particular phase of the business cycle, populist nationalism, or settler activity as essential causal agents. There are simply too many instances when expansion occurred in the absence of these factors for any of them to be conducive to a process lasting five centuries.

By the same logic, variations among metropoles can be used to rule out other­wise plausible causal factors for failing to pass the spatial comprehensiveness test. Empires were constructed by states that were large and small (by European stan­dards), rich and relatively poor, industrialized and agrarian, monarchical and re­publican, democratic and authoritarian, feudal and postfeudal in social structure. The size, wealth, level of industrial development, type of political system, degree of democratization, or social structure of metropoles cannot be considered necessary, sufficient, or even conducive conditions for European global dominance. There is simply too much variation in each category among states sharing the capacity and will to establish overseas possessions. To say this is not to deny that each metropole differed in important ways from the others. Chapter 9 discusses these distinctive features, showing how they help account for variations in the extent, duration, and character of various empires.

The search for dissimilarities between phases and among metropoles performs the useful function of decreasing the number of plausible causal factors. The search for widely shared features of west European geography, society, and history takes enquiry in a more positive direction. What merits inclusion as a conducive cause? The criterion is how well a given factor fares against the comprehensiveness tests just described. It would be unreasonable to expect any factor to pass all four tests. For example, a feature of European geography shared by all metropoles but not present in other world regions is too comprehensive temporally: it cannot distinguish be­tween Europe pre- and postphase 1 because it remains constant across time. A more reasonable standard is that a causal factor pass at least one and preferably two or three tests, and that each of the four tests be passed by at least one factor. What sets modern western Europe apart, both from its medieval predecessor and from other societies that might have attained world dominance, are the conjuncture and inter­action of factors noted in chapters 8-10?