Chapter 2 Patterns of Inadvertent Expansion, 1816–2014

Inadvertent Expansion Data

To get a better understanding of the conditions under which inadvertent expansion occurs, I compiled and collected data on territorial expansion by the great powers from 1816 to the present. I made use of a wide variety of sources in collecting the data, including existing data on the topic, encyclopedias, historical dictionaries and chronologies, historical surveys, in-depth histories, and even primary source documents when necessary.1

The most detailed and comprehensive existing data on territorial expansion is the Correlates of War’s (COW) “Territorial Change” data, collected by Jaroslav Tir, Philip Schafer, Paul F. Diehl, and Gary Goertz.2 Despite the immense value and important contributions of the territorial change data, I did not simply take it “off the shelf” and rely on it for two reasons. First, it was sometimes a challenge to identify each and every observation of expansion in the data.3 This poses problems in studying inadvertent expansion, since each individual observation must be carefully examined to uncover whether the initiative originated in the capital or on the periphery. And second, in the course of researching, I came across a number of observations that, as best as I could tell, are not included in the territorial change data.4 The data presented here, therefore, provide a more complete picture of the history of great power territorial expansion.

The unit of analysis for the data is the successful great power expansion observation. So, for example, despite the fact that it unfolded over several months, the US invasion of Florida in 1818 is still considered a single observation. In chapter 1, I defined expansion as the coercive acquisition of foreign territory that is intended to be long term or permanent for the expanding state. In accordance with this definition, expansion observations had to meet five criteria to be included in the data. First, the expansion observations had to be coercive in nature. Thus, voluntary purchases, transfers, and trades of territory are not included in the data.5 Second, the territory acquired had to be foreign at the time of acquisition. Various forms of domestic expansion—such as expropriation, counterinsurgency campaigns, and territorial pacification—are, therefore, not included in the data.6 Third, the territory had to be inhabited or claimed by another political entity. Genuine terra nullius claims are not included in the data.7 Fourth, the territorial acquisition had to be non-temporary, at least in its intent. Cases of temporary military occupation are not included in the data.8 Finally, the expansion observation had to be successfully carried out to its full fruition. Cases where expansion was attempted but not successfully completed are not included in the data.9

As this fifth and final criterion makes clear, the quantitative data do not include “negative,” or failed, cases of great power territorial expansion. The primary reason negative cases are excluded is that they are not easily observable in the historical record.10 When it comes to intentional expansion, failed or negative cases are incredibly difficult to distinguish from interstate war, brief military occupations, and other temporary armed incursions on foreign territory. To make these distinctions, one would need to have a fairly solid understanding of state leaders’ plans and intentions before each and every armed incursion on foreign territory by each of the great powers over the course of two centuries—a tall order, to be sure. When it comes to inadvertent expansion, failed or negative cases are difficult to distinguish from militarized territorial disputes and often do not show up in the historical record in a meaningful way at all. What typically occurs in these cases is a peripheral agent claims or conquers a piece of foreign territory and then is either ignored or quickly (and often quietly) rebuffed by leaders in the capital. Events such as these are often simply not notable enough to show up in most histories or even in many published collections of government documents. Compared to cases that ultimately lead to new territory being gained, these negative cases are far more difficult to observe and far less likely to be a prominent part of recorded history. Therefore, generating a representative sample of such negative cases, let alone the entire universe of them, is likely to be prohibitively difficult for a single book by a single scholar.

On the one hand, the exclusion of failed or negative cases of territorial expansion in the quantitative data obviously introduces the risk of “selection bias,” since the cases in which expansion succeeds may differ in important ways from those where it fails and, therefore, may not be representative of the phenomenon as a whole.11 On the other hand, the data collection effort for this book was already substantial, involving traversing the histories of nine great powers over the course of, in some cases, two hundred years. And, given the difficulties of observing “attempted” or failed expansion, as discussed earlier, I think this is simply a problem to be aware of, and in light of which the results should be interpreted, not one that can be readily solved.12 Five failed or negative cases of inadvertent expansion—where faits accomplis were presented by peripheral actors but rejected by leaders in the capital—are given detailed qualitative examination in chapters 3–7.

I have restricted the empirical scope of the data to the great powers from 1816 to the present. I use the COW’s State System Membership data on “Major Powers” to indicate the identity and the tenure of the great powers, with a few modifications.13 I have limited my focus to the great powers for three central reasons. First, the great powers do most of the territorial expansion. According to the territorial change data, great powers account for 76 percent of all cases of successful conquest and annexation globally from 1816 to 2014. Second, territorial expansion by the great powers tends to be more consequential. As just one possible indicator of “consequence,” the territorial change data suggest that great power expansion accounts for 79 percent of the territorial area gained and 62 percent of the population gained through conquest and annexation globally since 1816.14 And third, the empirical record is most complete for the great powers, and reliably coding expansion as inadvertent or intentional often requires a great deal of in-depth historical research.

Altogether, I compiled and collected data on 258 cases of territorial expansion by the great powers between 1816 and 2014. For each observation, I include a set of variables with basic information about the event in question. For the purpose of this book, the most important of these is inadvertent—a dichotomous variable indicating whether the expansion observation is inadvertent or not (and, therefore, implicitly “intentional”). As outlined in chapter 1, I consider an expansion observation to be inadvertent when it is planned and executed by actors on the periphery, without the authorization of leaders in the capital. My basic coding procedure for each observation was to seek out information on who specifically ordered the territorial acquisition in question. When I found evidence in the historical record that the acquisition was ordered or authorized in advance by leaders in the capital—the chief executive or key members of the security apparatus, including the foreign minister, defense minister, or colonial minister—I considered the observation intentional.15 When I found a primary source or at least two secondary sources indicating that the territory was acquired without orders or prior authorization of any of these leaders in the capital, I coded the observation as inadvertent.16

For three reasons, who ordered a given territorial acquisition is often a surprisingly difficult piece of information to unearth. First, in writing on territorial expansion and empire in both history and political science, causes and consequences tend to be emphasized at the expense of process.17 That is, scholars usually focus on why a piece of territory was acquired, or how it was governed after acquisition, but the actual process of acquisition gets far less attention. A second issue is the widespread personification of the state or empire, which is also common in both history and political science.18 Frequent references to “Russia,” “Japan,” or “the British Empire” acquiring various territories obscure the nature of the specific actors involved as well as the authorities under which they were operating. Relatedly, many scholars operate with implicit (or explicit) rational and unitary assumptions of the state, which can encourage post hoc reasoning and reliance on “revealed preferences”—they acquired the territory, so they must have wanted to. Third, the historical record on a good number of these cases is incredibly thin. Cases of inadvertent expansion, in particular, are often less dramatic than their intentional counterparts and, therefore, tend to receive considerably less attention in historical research. Furthermore, as noted in chapter 1, the leaders who experience inadvertent expansion often have incentives to cover up these episodes, so as to not appear incompetent or out of control, and may suppress, distort, or even destroy records of events. It seems likely, therefore, that inadvertent expansion is even more common than I have been able to uncover here. In short, coding expansion observations as either intentional or inadvertent is both time consuming and challenging.19

Besides whether the expansion was inadvertent, each observation includes the year and month in which it took place, the gaining great power (gainer), the territorial entity acquired, and the region in which the observation occurred.20 I define regions broadly, dividing the world into just six major regions: the Western Hemisphere, Europe, sub-Saharan Africa, the Middle East and North Africa, South and Central Asia, and the Asia Pacific. I also include variables that indicate whether the territorial entity was acquired via armed conquest or political annexation. Following the territorial change data, and in line with international legal definitions, I consider expansion to be annexation when territory is acquired primarily through diplomacy, and I consider expansion to be conquest when territory is acquired primarily through the use of military force.21

The theory presented in chapter 1 argues that inadvertent expansion should be more likely when leaders in the capital have limited control over their peripheral agents and in situations that involve relatively low levels of geopolitical risk. To capture the concept of control over the periphery, I include the variable telegraph, which is a dichotomous variable indicating whether the territorial entity acquired was connected to the global telegraph network at the time of acquisition.22 The idea here is that, if there was a telegraph station in the territory at the time of acquisition, this would allow leaders in the great power’s capital to rapidly communicate with any of their agents in and around the territory and thereby better monitor and potentially control their behavior.23 Another variable included to get at the concept of control is extra_regional, a dichotomous variable indicating whether the territorial entity acquired is outside of the great power’s own region. The idea here is that, all else being equal, agents within a great power’s own region are easier to control than those in other regions. A third, more granular, measure that I include along these lines is the distance of the entity acquired from the great power’s capital.24 This, again, is on the assumption that, all else being equal, agents in more distant territories should be more difficult to control than those in more proximate territories.

I also include two variables meant to capture the second factor: geopolitical risk. While, according to the theory of inadvertent expansion, this factor mainly influences the second step of subsequent central authorization, it is worth seeing if there are broad differences between intentional and inadvertent expansion with respect to geopolitical risk. Building off of arguments made by Colin Elman and Randall Schweller, I construct the dichotomous variable risk.25 I consider territorial expansion to be risky when it is: (1) onto the territory of another great power, (2) onto territory adjacent to another great power’s national borders, (3) onto the territory of a political entity allied with another great power, (4) in violation of a prior agreement with another great power on the territorial integrity of the entity in question, and/or (5) onto the territory of a regional power—a relatively powerful state that does not meet the typical threshold of great power status. I use the Alliance Treaty Obligation and Provision data, as well as the Correlates of War’s Formal Alliances data, as measures of formal alliances.26 Examples of the kind of states that are considered regional powers are Mexico, Austria (post-1918), Spain, Romania, the Ottoman Empire, Egypt, China (pre-1949), and Taiwan (post-1949). A related variable that I include is the dichotomous variable conflict, which indicates whether the expansion observation is part of a broader conflict. I consider expansion to be part of a broader conflict when it is undertaken during, and as part of, a broader war, or when it takes place in the immediate aftermath, and as a direct result of, a broader war. I consider wars as being those conflicts included in the Correlates of War’s “Inter-state,” “Intrastate,” and “Extra-state” war data.27

I also include a number of variables that can be thought of as controls for other plausible explanations for inadvertent expansion. First, and as noted earlier, I include the dichotomous variable annexation. Given that political annexations generally require fewer resources and are less difficult to organize than armed conquest, it is possible that inadvertent expansion is more likely to occur via annexation. Second, I include a measure of the capacity of the state’s institutions. It seems plausible that states with weak institutions may be more likely to experience inadvertent expansion, since they have smaller central state apparatuses and less capacity to monitor their agents at any distance from the capital. For this, I include the great power’s “information capacity” (info_capacity), an annual index that measures the state’s ability to collect and process reliable information about its population and territory.28 Third, I include a number of regime type variables. The dichotomous variable democracy indicates whether the expanding great power has a Polity score of six or greater at the time of acquisition.29 There is an enormous literature in international relations pointing to the unique foreign policy behavior of democratic states. And it is at least plausible that, allowing more individual liberty for their citizens and being more subject to popular pressures, democracies may be more likely to experience inadvertent expansion. I also include an autocracy indicator for the expanding great power, its simple polity score, and its “Varieties of Democracy” polyarchy score (vdem_polyarchy), a common alternative regime type measure.30 Fourth, I include the great power’s Composite Index of National Capabilities score (gainer_cinc), a standard measure of relative power.31 Again, many theories in international relations point to relative power as an important explanatory factor for a variety of foreign policy outcomes, so it is accounted for here.

Descriptive Statistics

The data collected show inadvertent expansion to be a surprisingly common phenomenon.32 Of the 251 cases of great power territorial expansion that I have coded as either intentional or inadvertent, a total of 55—or 22 percent—are inadvertent. This point, I think, deserves emphasis: in the modern history of great power politics, nearly one-in-four cases of territorial expansion are the result of decisions made by actors without the authority to make them. Thus, in almost a quarter of cases of territorial expansion by the great powers, there is a sharp disconnect between the intentions of the leaders in charge and the state apparatuses that they are purportedly in charge of. And the kinds of territories acquired in this manner run the gamut, from tiny islands in the South Pacific—such as France’s annexation of Tahiti in 1842—to large swaths of territory containing tens of millions of people—such as Japan’s conquest of Manchuria in 1932. All cases of inadvertent expansion are listed in table 2.1.

Besides being a common occurrence, the data on inadvertent expansion show some other interesting patterns. First, there are cases of inadvertent expansion as early as 1818 and as late as 1932, though, notably, it tends to occur in earlier years than intentional expansion. As shown in figure 2.1, inadvertent expansion is a regular occurrence throughout the nineteenth century and into the early twentieth century, with a small cluster of observations during World War I. Intentional expansion, in contrast, is more common throughout and sees increases during the late nineteenth-century “scramble” for empire and during the two world wars. Second, a majority of the great powers experience inadvertent expansion, though they vary considerably in the extent to which they do. As indicated in table 2.1, France, the United Kingdom, and Russia experienced a considerable amount of inadvertent expansion. Japan, Germany, and the United States, by contrast, experienced moderate-to-low amounts of inadvertent expansion. And great powers such as Austria, Italy, and China have seen no successful cases of inadvertent expansion at all.33 Third, inadvertent expansion occurs in all world regions, though some regions experience a lot more than others. As suggested by the cases listed in table 2.1, South and Central Asia, sub-Saharan Africa, and the Asia Pacific see significant amounts of inadvertent expansion, whereas the Western Hemisphere, the Middle East and North Africa, and Europe see little to none.

In sum, inadvertent expansion occurs in nearly one-in-four cases of great power territorial expansion, a wide variety of territories have been acquired in this manner, it covers over a century of history, it is experienced by a majority of the great powers, and it occurs in all world regions. It is, therefore, a general phenomenon—a regular occurrence in the history of great power politics.

However, the questions that primarily concern this book are when and why it occurs. Throughout this book, I have argued that two factors are central to the process: the degree of control leaders in the capital have over their peripheral agents, and the geopolitical risk associated with acquisition. The key variables here are telegraph and risk. If the theory of inadvertent expansion is correct, with respect to the telegraph, expansion should be less likely to be inadvertent in cases where there is a telegraph station in the territory at the time of acquisition and more likely to be inadvertent in cases where there is not a telegraph station in the territory at the time of acquisition. With respect to risk, expansion should be less likely to be inadvertent in cases that are risky and more likely to be inadvertent in cases that are not risky. Figure 2.2 presents two difference-in-proportions comparisons.

As the figure clearly shows, these theoretical expectations appear to be borne out. Starting with telegraph, on the left-hand side, 8 percent of expansion observations are inadvertent in cases where there is a telegraph station in the territory at the time of acquisition. In contrast, 38 percent of observations are inadvertent in cases where there is not a telegraph station in the territory at the time of acquisition—a large and highly statistically significant difference. A similar pattern is evident with respect to risk, on the figure’s right-hand side. Nine percent of expansion observations are inadvertent in cases that do involve considerable risk, whereas 32 percent of observations are inadvertent in cases that do not involve much risk—again, a large and statistically significant difference.

The average differences in value between intentional and inadvertent expansion for all the remaining variables are presented in table 2.2. As is clear, some variables see much greater differences than others across the two types of expansion. Variables such as year, extra_regional, distance, and conflict show considerable differences between intentional and inadvertent expansion. Based on these figures, it appears that inadvertent expansion is more likely: in earlier years, when the territorial entity is outside the great power’s home region, and at a greater distance from the great power’s capital. It appears that inadvertent expansion is less likely, in contrast, when the expansion is part of a broader conflict. For the remaining variables, there appears to be little difference between the two forms of expansion: annexation, info_capacity, the regime type variables (polity, democracy, autocracy, and vdempolyarchy), and gainercinc scores show little to no meaningful difference between cases of inadvertent expansion and intentional expansion. Inadvertent expansion, it seems, is no more likely to take the form of annexation than conquest. Great powers with weak and strong institutions are equally likely to engage in inadvertent expansion. Democracies and autocracies alike seem to expand both intentionally and inadvertently. And relatively powerful as well as relatively weak great powers have a similar record with respect to both types of expansion.

Analysis

These comparisons are certainly interesting, and even enlightening. But with the evidence presented so far, it is difficult to tell which factors most distinguish inadvertent and intentional expansion. Because some of these measures may be correlated with each other, and many are likely correlated with time, it is necessary to observe the independent relationship of each while controlling for the others. This is the task I turn to now.

As noted earlier, the unit of analysis is the great power territorial expansion observation and the outcome of interest is inadvertent, a dichotomous variable indicating whether the expansion was inadvertent or not. Thus, the variation being analyzed here is between successful forms of expansion: inadvertent and intentional. In an ideal world, we would have data that include all cases of successful intentional and inadvertent expansion as well as all cases of failed intentional and inadvertent expansion. This would provide a more complete picture of the broad conditions under which inadvertent expansion opportunities arise as well as when leaders either accept and or reject the fait accompli. However, as discussed in chapter 1, cases of failed inadvertent expansion are not easily observable in the historical record, making comprehensive collection difficult. As a result, this quantitative analysis examines variation between forms of expansion. The qualitative case studies in the chapters that follow take up the question of variation between acceptance and rejection of the fait accompli.

The primary explanatory variables I include are telegraph and risk. One concern with the telegraph measure may be that it is simply a proxy for the passage of time. Because the telegraph emerged in Europe and North America the mid-nineteenth century and gradually spread across the globe in the decades that followed, any apparent relationship between the telegraph and inadvertent expansion may simply be due to some other variable that is also correlated with time. To try to deal with this concern, I include the year in which the expansion observation occurred as a control variable. Another concern with the telegraph measure may be that it is simply a proxy for distance, since territories that are closer are likely to be telegraphically connected to the capital sooner than those that are more distant. To try to deal with this concern, I include distance as a control variable in the analysis. As for geopolitical risk, one concern may be that it is simply picking up the difference between peacetime and wartime territorial expansion. To try to address this concern, I include conflict as a control variable. As more general controls for other plausible explanations, I include annexation, info_capacity, democracy, and gainer_cinc. The analysis is conducted using a linear probability model. In all cases, robust standard errors are reported to account for error heteroskedasticity. Table 2.3 presents the results.

As the results make clear, both the telegraph and geopolitical risk appear to be correlated with expansion being inadvertent rather than intentional, even when accounting for other factors, such as time, distance, the form of expansion, state institutional capacity, regime type, and relative power. Even with these controls, observations of territorial expansion are 23 percent less likely to be inadvertent when there is a telegraph in the territory acquired at the time of acquisition. Similarly, when the expansion observation involves considerable risk, it appears to be 14 percent less likely to be inadvertent, even when controlling for these other factors. Both of these measures are statistically significant at conventional levels, and none of the other variables reaches statistical significance.

As a whole, the analysis suggests that two factors are most importantly associated with expansion being inadvertent rather than intentional. The first is the absence of a connection to the global telegraph network in the territorial entity acquired at the time of acquisition. The second is the geopolitical risk associated with the expansion observation in question. The great power’s state capacity, form of expansion, regime type, and relative power, notably, do not seem to be correlated with expansion being intentional or inadvertent.

I ran a number of tests to see how robust these results are to alternative model specifications and variable inclusions. The full results are available in the online appendix (https://doi.org/10.7910/DVN/JBGNNH), though I briefly describe the procedures and findings here. First, I reran the analysis using logistic regression rather than a linear probability model. The results are unchanged. Second, I reran the analysis including extra_regional rather than distance as the distance control. The results are, again, unchanged, and extra_regional is not statistically significant. Third, instead of using democracy as the great power’s regime-type measure, I reran the analysis including an autocracy indicator, a simple polity score, and a Varieties of Democracy Polyarchy score (vdem_polyarchy) for electoral democracy. The results are unchanged, and none of the alternative regime-type measures reach conventional levels of statistical significance. Fourth, I reran the analysis with dichotomous variables controlling for each of the great powers, one by one, to see if the results are driven by any particular great power actor. The results are unchanged, and the statistical significance of risk even increases when controlling for the United Kingdom, Italy, or Japan. Fifth, and finally, I reran the analysis with dichotomous variables controlling for each region in which expansion takes place, one by one, to see if the results are driven by any particular region. The results are mostly unchanged, with an important exception: risk loses its statistical significance when controlling for sub-Saharan African expansion. This suggests that the cases in sub-Saharan Africa play a statistically influential role in low-risk conditions being associated with inadvertent territorial expansion. However, given that sub-Saharan African cases account for nearly half of all cases of inadvertent expansion (24/55 or 44 percent), this may not be entirely surprising.34 Overall, the results are quite robust to alternative models and variables. The presence or absence of rapid modern communications in the form of a telegraph station, as well as relatively low geopolitical risk, are strongly associated with whether territorial expansion is inadvertent or intentional.

Discussion

These results provide strong, though only partial, support for the theory of inadvertent expansion presented in chapter 1. First, the argument that inadvertent expansion is more likely when leaders in the capital lack adequate control over their agents on the periphery is supported by the analysis. The strongest predictor of when expansion is inadvertent rather than intentional is the absence of a globally connected telegraph station in the territorial entity acquired at the time of acquisition. The presence of such a station is associated with a 30 percent reduction in the probability of expansion being inadvertent (see figure 2.2), and this relationship holds even when controlling for a variety of other plausible explanatory factors. As the evidence from this chapter suggests and as is illustrated in the historical case studies in the chapters ahead, the inability of leaders in the capital to rapidly and reliably communicate with, and thereby monitor, their agents on the periphery plays a crucial role in the ability for peripheral agents to spring territorial faits accomplis upon them—what I refer to as unauthorized peripheral expansion.

Second, the argument that inadvertent expansion is more likely when there is little significant geopolitical risk associated with territorial acquisition is also supported by the analysis. The only other significant predictor of when expansion is inadvertent rather than intentional is the absence of geopolitical risk. The presence of such risk is associated with a 23 percent reduction in the probability of expansion being inadvertent (see figure 2.2), and this result holds even when controlling for other explanatory variables. As these results indicate, and as is further supported in the qualitative chapters that follow, the presence of geopolitical risk will discourage leaders in the capital from accepting the territorial faits accomplis presented by their peripheral agents—what I refer to as subsequent central authorization.

Other possible explanatory factors—including whether expansion takes the form of annexation, the distance of the territory from the capital, whether it occurs in peacetime, the expanding great power’s state institutional capacity, its regime type, or its relative power—were not supported in the analysis.

This chapter has provided a macroscopic perspective of the phenomenon of inadvertent expansion in world politics. In it, I introduced and analyzed data on great power territorial expansion from 1816 to 2014. The chapter presented two central findings. First, inadvertent expansion is a surprisingly general phenomenon. Far from being an occasional and obscure “accident of history,” it has been a regular feature of the modern history of great power politics. It occurs in nearly one-in-four cases of great power territorial expansion, spanning over one hundred years, from the early nineteenth century to well into the twentieth century. It is successfully carried out by six different great powers, from the sprawling British Empire to the late-modernizing Imperial Japan. And a wide variety of territories have been acquired in this manner, from small remote islands in the South Pacific, to topographically forbidding central African kingdoms, to densely populated regions of East and Southeast Asia.

Second, the most important conditions under which inadvertent expansion is likely to occur are when the capital lacks sufficient control over its agents on the periphery and when the acquisition in question involves little geopolitical risk. Expansion into territorial entities that were not connected to the global telegraph network at the time of acquisition is significantly more likely to be inadvertent than expansion into territories that were connected. Similarly, expansion in conditions involving little geopolitical risk is significantly more likely to be inadvertent than expansion in conditions involving considerable risk. Other plausible factors do not appear to be related to whether expansion is intentional or inadvertent. What this chapter has primarily accomplished is to uncover what is distinctive about inadvertent expansion within the broader universe of cases of territorial expansion, and this distinctiveness has been shown to be consistent with the theory of inadvertent expansion presented in chapter 1.

However, two caveats are necessary in closing. First, a lack of rapid communications technology is neither a necessary nor a sufficient condition for inadvertent expansion. There are ten cases in the data of inadvertent expansion into territories that are connected to the global telegraph network, showing that it can occur under conditions of modern communications technology.35 And there are seventy-five cases in the data of intentional expansion in the absence of a proximate telegraph station, indicating that great powers can still readily direct territorial expansion from the capital, even without the aid of rapid modern communications. This also applies to geopolitical risk. There are eleven cases of inadvertent expansion in conditions that are objectively risky, showing that the existence of geopolitical risk is not a foolproof guarantee of expansion being intentional.36 And there are seven cases of inadvertent expansion that are both into territories with a global telegraph network connection and that involve considerable geopolitical risk.37 However, what the evidence presented in this chapter does suggest is that relatively low peripheral control by the capital and low geopolitical risk seem to make inadvertent expansion significantly more likely.

The second caveat is that, as mentioned earlier, the structure of the data and the variation examined in the analysis are not ideal for testing the theory of inadvertent expansion as it is presented in chapter 1. The ideal data would include not only all cases of successful intentional and inadvertent expansion but also all cases of unsuccessful intentional and inadvertent expansion. This would give us a much better sense of the universe of cases of inadvertent expansion opportunities—helping us deal with problems of selection bias—and allow us to quantitatively examine the variation between acceptances and rejections of the fait accompli by leaders in the capital. Difficulties of observation and data collection rule this possibility out. However, it is possible to examine variation between successful and failed inadvertent expansion on a smaller sample of cases and in a more focused and detailed way. Conducting such an examination is the task of the remainder of this book.