Levels of analysis

Some fifteen years ago, Nancy Green concluded that migration history (and social history in general) was in need of ‘poststructural structuralism’.Footnote ¹ Thus, the fruits of structural and broad comparisons could still be enjoyed, provided that they paid attention to and were embedded in specific cultural and historical contexts. When we compare developments or processes, either in space or time, we therefore have to be more aware not only of the levels and units of analysis but also of the way in which we structure our comparisons. She and others concluded that in most cases the meso level is to be preferred over highly aggregated (macro) approaches.Footnote ² In the parlance of migration history: networks and agency, rather than push and pull. Although in principle we fully subscribe to her plea, in our original 2009 article, ‘The mobility transition revisited’, we nevertheless opted for the macro dimension, both for data on the choice of migration categories and for the ensuing typology. Before going into the extremely useful commentaries on our article we would first like to explain why, for comparative reasons, we ostensibly disregarded Green’s advice.

The main reason for our approach was to enable global comparisons by constructing uniform migration rates and a new standardized migration typology, or at least to develop systematic ideas in that direction. This was also why we cast our approach in a formula that allows other scholars to apply our method in other periods or continents. Only then are systematic comparisons possible. The comments by Adam McKeown are a good illustration of such a standardized scholarly exchange. By using our model, his reconstruction of Chinese migration patterns between 1600 and 1900 shows similarities but also important differences in both level and types of migration. The juxtaposition of China and Europe, embedded in the political and socioeconomic history of these two large parts of the world, automatically leads to new insights and questions.

One of the things they do is to open our eyes to the different timing and different effects of military migrations and colonization, as Figures 1 and 2, based on our and McKeown’s own calculations, show. Competition between states in Europe led to almost constant warfare, as Tilly stressed.Footnote ³ This produced not only massive military migrations, but also important technical innovations. In China, the effects in the long transition period from Ming to Qing may have been comparable. From the Taiping Rebellion (1850–1864) onwards, however, the effects were more destructive (politically and economically). Another apparent difference between Europe and China in the nineteenth century is the importance of rural to urban migrations, which we will discuss later.

Figure 1.

Share of migration modes in Europe, 1601–1900

Figure 2.

Share of migration modes in China, 1601–1900

Units of analysis

Such very broad comparisons, however, as Adam McKeown and Leslie Moch rightly argue, also have their downside, as they tend to hide as much as they reveal. We fully agree that it might make more sense to compare China as an empire with the Russian, Ottoman, or Austro-Hungarian empire. With respect to human mobility, the logic of empires is different from that of nation-states. They tend to move their borders over people and thus rulers are confronted with much more ethnic, religious, and linguistic diversity than in most nation-states, especially following the ‘religious cleansing’ of large parts of Europe from the fifteenth century onwards.Footnote ⁴ The imperial logic, linked to greater (partly feudal) restrictions on peasant mobility, explains why male and female labour migration to commercialized areas was less widespread. But empires were also much more active in sending peasants as colonizers to frontier regions, a category that, in the Chinese case, was partly conflated with the migration of soldiers.Footnote ⁵

Which unit of analysis is meaningful, however, depends on the research question. In the ‘Great Divergence’ debate on the economic performance of China and Europe, economic historians did not use empires as their point of departure but compared England with the Lower Yangtse delta. The reason was simple: these two represented the engines of economic growth in their respective parts of the world.Footnote ⁶ However, contrasting comparisons (such as our comparison of migration rates in Russia and the Dutch Republic) need not be misleading or useless per se, as Leslie Moch seems to assume. Notwithstanding structural differences in scale and nature, such sub-comparisons on the level of states open our eyes to different migration dynamics, and help to put mass migrations in their historical context.

The migration categories

McKeown, Moch, and Ehmer also discuss problems with the categories that we used to systematize the various forms of cross-community migrations in Europe. We agree that our typology is somewhat ambiguous. The initial aim was to come up with categories that would enable us to measure total migration rates for Europe as a whole, with as little overlap as possible. So far, our critics have not really suggested an alternative, apart from adding nomadism and transhumance (Ehmer), which did not concern very large numbers, whether in Europe or beyond.Footnote ⁷ We do think, however, that these kinds of moves do fit in our seasonal category.Footnote ⁸

Our migration typology indeed harbours two different organizing principles. One is to differentiate different forms of migration: to cities, seasonal, soldiers and sailors, and colonization. The other consists of two logical categories, namely emigration from the region chosen – in this case Europe – and immigration into it. That the latter two categories are of a different nature may be demonstrated easily: if the region chosen is the entire world, the categories emigration and immigration would vanish automatically. The next question is what the cross-community effects of emigration are on the region of origin. In the period under investigation they could be threefold: through return migration, communication with the home region (by letters, etc.), and books about non-European parts of the world. The degree to which such experiences actually reached the sending area depended in the first place on the life expectancies of emigrants, and also on their communication skills, in particular literacy. Both these varied greatly according to place and time. Mortality of Europeans was much higher in the tropics, for example in the Caribbean and Asia, than in North America.Footnote ⁹ Literacy was much higher in north-western Europe than in southern and eastern parts of the continent.Footnote ¹⁰ Mortality figures dropped from the second half of the nineteenth century, whereas literacy figures started to rise at the same time. If we were to apply such considerations to coolies, for example, we might be able to determine effects of emigration from outside Europe. This would also include a comparison of destinations. Did it matter that most European emigrants to the United State landed up in cities, whereas many Chinese and Indian coolies were restricted to plantations and mines?

Finally, we point at the possibility that within the period being researched the geographical unit may change, which has consequences for both the emigration and the immigration categories. In our case, one could argue that, from the early nineteenth century onwards, North America became part of the European orbit, as the consequence of the first round of globalization, which constituted a coherent Atlantic economic and cultural migration field.Footnote ¹¹ As most European emigrants went to US cities, the erstwhile ‘emigrants’ in such a new spatial unit then shifted to the urban category. The relative weight of rural to urban migrations in ‘the Atlantic’ would then contrast even more strongly with China, even if we restrict ourselves to the Lower Yangtze delta or the urbanized Chinese east coast, where levels of urbanization were still considerably lower than in Europe for the period 1850–1900.Footnote ¹²

Migration rates and the mobility transition

Several critics have questioned our definition of migration. Moch doubts the utility of Manning’s cross-cultural concept, pointing to the often low intensity of contacts between migrants and ‘natives’. Many migrants were isolated from others (such as soldiers or seasonal workers), kept to themselves, and returned to their home regions. We do not disagree with her but do not regard this as a major problem. The application of our model at lower levels of aggregation will indeed enable scholars to research and qualify the intensity of the contact, and to analyse the conditions under which cross-cultural connections à la Manning are likely to lead to innovations. This will undoubtedly differ between categories, but also from case to case. As we explained in our original article, we expect that migration to cities had a greater innovative capacity than colonization or seasonal migration. However, in the latter case we should pay more attention to the effect of changes wrought by return migrants, a phenomenon that we have not dealt with so far. We would also like to add that migrants – and in particular immigrants in cities, sailors, soldiers, and seasonal migrants – mostly had to earn their income as wage labourers, and consequently entered into labour relations, with other possibilities for cross-cultural contacts.

Josef Ehmer also looks critically at our definition, but from a different perspective. Based on his own and others’ extensive and excellent work on travelling journeymen and domestics in the Austro-Hungarian empire, he suggests a more encompassing definition of migration, which encapsulates micro and repetitive moves by domestics, journeymen, and the like. We agree fully that leaving them out, unless they settled in cities, has the big disadvantage of underestimating high levels of mobility in early modern Europe. Besides, it might even completely level off the jump in the figures during the second half of the nineteenth century. As a consequence, in a future version of our data collection, we are planning to include figures for tramping artisans, and if possible also migratory servants, even if it is very difficult to collect data at an equal level for different periods and parts of Europe. We will subsume these tramping artisans and servants (not settling down in cities) under the heading of temporal migrants, which presently exclusively consists of soldiers and sailors.

Since the publication of our 2009 article we have thoroughly elaborated our data in our 2010 IISH Research Paper, in which we extensively document our sources and explain our methods.Footnote ¹³ Combined with new data, this has led us to lower the migration rates in the period 1850–1900 by almost five percentage points (Table 1), thus moving in the direction that Ehmer suggests.

Table 1.

Total migration rates in Europe, 1501–1900

Source: Lucassen and Lucassen, ‘Mobility transition revisited’.

Modernization and human capital

Perhaps more interesting are Ehmer’s remarks about the broader implications of the high mobility of young European men and women in the early modern period for the nature of Europe’s society and economy. His ideas fit perfectly with Jelle van Lottum’s critique of our lack of sophistication with respect to the quality of the migrants, or human capital.Footnote ¹⁴ Van Lottum uses a new institutionalist approach, pointing at the ‘modern’ features of north-western European societies long before the nineteenth century, such as free wage labour for men and women, relatively high levels of social mobility, high levels of urbanization, the dominance of the nuclear family and neo-local marriages, effectively functioning markets, and the protection of property rights.Footnote ¹⁵ That geographical mobility for both men and women was so normal in this highly commercialized, urbanized, and proletarianized part of Europe, as Ehmer argues, was crucial for the functioning of labour markets and the effective allocation of labour and human capital in the economy. This perspective also offers comparative possibilities, making an interesting contrast with the information provided by Adam McKeown that wage and seasonal labour in China did not necessarily occur in the commercialized parts of the empire, but often in frontier regions. This difference would support the position of those in the Great Divergence debate who stress the in-built advantages of north-western, or more broadly western, Europe.

Van Lottum, furthermore, rightly warns against the temptation to assume too easily a causal relation between cross-cultural migrations and innovation or economic performance. We agree that the quality of the human capital of migrants matters, and that studies therefore have to take the specific historical context into consideration. He exemplifies this by using the concepts of ‘migration fields’ and ‘intervening obstacles’ in his comparison of Amsterdam and London in the seventeenth and eighteenth centuries. Thus he shows that their respective migration fields differed in character and changed over time. Amsterdam, in particular, was confronted with emerging competitors in a migration field that stretched to Scandinavia and northern Germany – for example, Copenhagen, Hamburg, and Bremen. This was much less the case in the British Isles, which constituted London’s migration field. As a consequence, in contrast to London, the human capital of Amsterdam’s (maritime) migrants dropped, thus accelerating the decline of the Dutch economy.

Van Lottum urges us to pay more attention to the highly ambiguous relation between migration and economic performance. First of all, it depends what a certain economy needs. The cotton factories of Manchester and Liverpool in the mid nineteenth century asked for cheap unskilled or semiskilled labour, whereas the London service economy depended largely on migrants with higher rates of human capital. In the latter case, brains were more important than muscles, and the innovative effect of migration was probably higher. We think that Van Lottum’s insights from new growth theory and neo-institutionalism are important and may help migration historians to rethink the relation between cross-community migration and innovation, as postulated by Patrick Manning.Footnote ¹⁶ To test his hypothesis and formulate the conditions under which such an effect is likely, detailed comparative case studies are needed, in which the interplay between the human capital of migrants and the institutional opportunities offered by receiving communities is crucial. We do not, however, share his ideas on ‘quality’ versus ‘size’. The number of migrants does matter, because this determines the quantity of cross-cultural contacts, and thus the extent of innovative potential.

State-centredness

A last important critique, formulated by Leslie Moch, concerns the extent to which our calculations might be a statistical artefact, caused by the obsessions of state officials. We do not deny that states, at all levels, are important actors in stimulating or slowing down migration, both through legislation and by hiring migrants to work for the state, as is most clearly the case with soldiers. Furthermore, like all migration historians, we are indeed highly dependent on state sources that measure migrations. We doubt, however, that this is a big problem. Apart from the fact that the state played a key role in hiring soldiers, and in attracting or sending out colonizers, we did not entirely rely on state sources. Our estimates on people moving to cities, for example, are not so much based on state registrations as on what we know about urbanization processes and urban mortality. In the end, migrations in Europe were triggered by a combination of three interlinked political and economic processes that worked out at the micro, meso, and macro levels, and that can be summarized as shown in Table 2.

Table 2.

The relationship between three major historical processes and types of migration

The mobility transition

Finally, we return to the original aim of our quantitative exercise, which was to test the idea of a mobility transition. Although everybody agrees that mobility rates in early modern Europe (and China, for that matter) were much higher than the mobility transition assumes, this does not preclude the possibility that the jump from 20% to 30% in the nineteenth century had different causes from early modern cross-community migrations. According to Moch, the jump in migration rates is not only explained by the transport revolution but also by the general demographic transition, which led to unprecedented population growth in Europe, for which emigration offered a solution. Moch suggests that this coincided with a shift in mentality, ‘a belief that one could have a better life’. Or, in the words of Clark and Slack, there was a significant shift from ‘subsistence’ to ‘betterment’ migration.Footnote ¹⁷ Through what Dirk Hoerder called ‘the secularization of hope’,Footnote ¹⁸ migrants fostered greater aspirations and a belief in a more meritocratic society on the other side of the ocean, with ample chances for upward social mobility and freedom of religion and speech in a vibrant civil society, as Alexis de Tocqueville explained in his famous book Democracy in America (1835/1840). Moreover, closer to home, the French Revolution led to a more democratic and meritocratic society, and nurtured similar hopes and aspirations in the old continent.

This is an interesting and tempting idea, and for future research it could serve as a starting point to look more closely into the relationship between nineteenth-century emigration from Europe and cross-community effects. These might be discerned both across the ocean and back home, through return migrations, which were considerable.Footnote ¹⁹ Greater expectations, and the greater openness of societies, may have meant that migrants were selected differently, and were more open to new influences. This may have contributed to the nature of newly formed states beyond Europe.

We have started this rejoinder by situating our approach in the wider field of historical migration studies. Our macro approach has the great advantage of a formalized comparison of developments worldwide over long periods of time, through which not only can broad migration trends be detected but also their implications for innovation and economic performance. This is certainly not the nec plus ultra in our field because to explain such trends and their implications it is indispensable to look at meso and micro levels, as ‘poststructural structuralism’ suggests. For a truly encompassing global migration history, all three levels should therefore be studied in constant interaction.