Empires - Comparative History
Content
Original
Empires:
Article
aResearch
problem
ofLtd
comparative
history
Blackwell
Oxford,
Historical
HISR
©
0950-3471
XXX
Institute
UK
Publishing
of
Historical
Research 2006
Empires: a problem of comparative history*
Susan Reynolds
Institute of Historical Research
Abstract
Most historians of empires probably start by assuming that what they are interested
in are relatively large polities that consist of a ruling part (the metropolis) and
other parts (colonies or peripheries) that it dominates as a result of military
conquest or political or economic bullying, and that are retained and governed
separately from the metropolis rather than being directly absorbed in it. Not all
the polities called empires over the centuries, however, have had all these
characteristics. Surveying some of the variations in characteristics may help in
deciding what it may be most profitable to compare with what.
The first problem to confront in comparing empires is what kind of polity,
and which individual polities, we are to count as empires. One cannot begin
to make serious comparisons without deciding what comes into the category
of cases one wants to compare. It is not so much a matter of defining the
word ‘empire’ as of, first, deciding which phenomena – which actual
polities, past and present – that those who use the word refer to, and
deciding what characteristics these phenomena share that make them a
category; and second, considering whether there are other phenomena
that share significant characteristics with the first group and therefore
ought to be brought into the discussion even if they are not usually called
empires. ‘Empire’ is a European word, derived from Latin, while some
of the polities that we need to consider were outside Europe and came
to be called empires by Europeans for various reasons, some of which, as
I shall argue, were not based on any clear or consistent concept or notion
of what an empire is. The relationship between word and phenomenon
is at best indirect, mediated by the notions or concepts that users or
hearers of the word have of either the word or the phenomenon. 1
* This article is a revised version of a plenary lecture given at the Anglo-American Conference,
Beveridge Hall, University of London, July 2005. I am grateful to participants, and particularly
to Pamela Crossley, for their suggestions.
1
Words, concepts and phenomena are discussed (under different names) in C. K. Ogden and
I. A. Richards, The Meaning of Meaning: a Study of the Influence of Language upon Thought and the
Science of Symbolism (1923), pp. 13–15. There are further discussions in, e.g., J. Lyons, Semantics (2 vols.,
Cambridge, 1977), i. 95–119, 175; and R. Tallis, Not Saussure: a Critique of Post-Saussurean Literary
Theory (Basingstoke, 1988), pp. 114-16. As applied to medieval history, see R. Schmidt-Wiegand,
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Published by Blackwell Publishing Ltd., 9600 Garsington Road, Oxford, OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.
152
Empires: a problem of comparative history
It may be useful to make a preliminary distinction of empires, as I use
the word, from states. For present purposes I ignore all the arguments
about the monopoly (or control) of the legitimate use of force, about
boundaries, about when ‘true states’ appeared and whether the word is
only to be used about modern polities, and so on.2 States I take to be more
or less independent or autonomous polities, enforcing their own laws (or
failing to do so), which, whether unitary or federal in constitution, have
a different and less divided character from the polities that I am going to
talk about and that for convenience I lump together as empires. Prima
facie, and for what it is worth at this stage, I take the difference between
federal states and empires to be that the component parts of federal states are
supposed to be equal, while the polities that I am considering as empires
are normally thought of as consisting of a ruling polity (sometimes called
the metropolis) and subordinate polities (peripheries or colonies). They
came together, moreover, not by a voluntary (or supposedly voluntary)
federation but on the initiative of what thus became the metropolitan
polity, often by some kind of armed conquest. This working definition
is not intended to exclude any other, but merely to serve for this
discussion. It raises problems about the distinction between metropolis
and colonies, to which I shall return, and about whether to call each of
the colonies a state (as did, among others, Edmund Burke and Moses
Finley)3 or to reserve that title for the metropolis, or whether the whole
thing constitutes a single state. I shall not return to that as it seems to me
an unhelpful argument merely about words.
Words, nevertheless, are my starting point, since they are what we have
in the sources and what we must start from when considering notions and
the phenomena that they represented. That means starting with Rome
and going on through medieval Europe. I realize that most historians of
empires are primarily interested in modern ones. Most surveys of empires
tend to jump over the middle ages, but both Rome and the medieval
west matter: our notions of empire, like the word itself, started from
Rome, were adapted by medieval use, and filtered through the rather
narrow view of Greek and Roman history imparted by centuries of
classical education. For Europeans the Roman empire long remained the
‘Historische Onomasiologie und Mittelalterforschung’, Frühmittelalterliche Studien, ix (1975),
49 – 78; and S. Reynolds, Fiefs and Vassals: the Medieval Evidence Reinterpreted (Oxford, 1994),
pp. 12–14.
2
S. Reynolds, ‘The historiography of the medieval state’, in Companion to Historiography, ed.
M. Bentley (1997), pp. 117 – 38; and S. Reynolds, ‘There were states in medieval Europe: a
response to Rees Davies’, Jour. Historical Sociology, xvi (2003), 550 –5.
3
Edmund Burke, Writings and Speeches, iii: Party, Parliament and the American War, 1774 –80,
ed. P. Langford (Oxford, 1996), p. 132; M. I. Finley, ‘The 5th-century Athenian empire: a
balance sheet’, in Imperialism in the Ancient World, ed. P. D. Garnsey and C. R. Whittaker
(Cambridge, 1978), pp. 103–16.
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Empires: a problem of comparative history
153
prototype empire, the great result of great conquests, with supreme, even
virtually universal, authority. Much as I have learned from sketchy reading
about non-European and modern empires, I still think that looking at the
way earlier Europeans referred to empires and took their words and ideas
about them to the rest of the world explains some of the anomalies of
modern usage and the varying phenomena included in the whole category.
That, of course, is what a medievalist who has rashly undertaken to talk
about empires in general would say.
Romans used the word imperium for all kinds of authority and power
from that of heads of households up. When they used it for what we call
the Roman empire they apparently only occasionally gave it a territorial
connotation, as when they said that Palmyra lay between the summa
imperia of Romans and Parthians. The word imperator also started with
wider uses before it came to be used as the title of the emperor.4 After
the collapse of the western half of the empire, while a Roman emperor
still, of course, continued to rule in Byzantium, rulers in the west could
also be called emperors if they had or claimed authority over other kings
and kingdoms, that is, over polities that would have normally been
thought of as independent.5 Kings who claimed some kind of authority,
for instance over other kings within Britain in the early middle ages and
within Spain for longer, could be said to have imperium or to be called
imperatores. But much more formally and importantly the title belonged to
Charlemagne and his successors whose coronation in Rome symbolized
their claim to revive its ancient glories and, after them, to German kings
who were similarly crowned there as emperors. It is sometimes said that
the title was an empty one, but both Carolingian and, until at least the
end of the twelfth century, German emperors were the most powerful
rulers of Europe. Since they ruled multiple kingdoms, with their power
base in the kingdom of Germany, the rest of their dominions could rank
in our terms as peripheries or colonies, with Germany as the metropolis.
Their empire could therefore, for what it is worth, come within the
modern category, allowing perhaps for a shifting kind of inner metropolis
in whatever part of the kingdom of Germany the current emperor had
his particular base. In the twelfth century a foreigner could refer to the
emperor as emperor of Germany, but it was not until imperial authority
had become restricted to Germany, and within Germany, in the later
4
Oxford Latin Dictionary, ed. P. G. W. Glare (Oxford, 1982), pp. 842– 4; P. A. Brunt, ‘Laus
imperii’, in Garnsey and Whittaker, pp. 159 – 91.
5
D. Bullough, ‘Emperors and emperordom from late antiquity to 799’, Early Medieval Europe,
xii (2003), 377 – 87; R. Folz, L’idée d’empire en occident du ve au xive siècle (Paris, 1953), pp. 66 –
81, 114–23; J. L. Nelson, ‘Kingship and empire’, in The Cambridge History of Medieval Political
Thought c.350 –c.1450, ed. J. H. Burns (Cambridge, 1988), pp. 211–51; and J. L. Nelson,
‘Kingship and empire in the Carolingian world’, in Carolingian Culture: Emulation and Innovation,
ed. R. McKitterick (Cambridge, 1994), pp. 52–87.
Historical Research, vol. 79, no. 204 (May 2006)
© Institute of Historical Research 2006.
154
Empires: a problem of comparative history
middle ages that the empire came to be officially called the Holy Roman
Empire of the German Nation.6
It is not only wrong to suppose that the medieval western empire, far
from being an empire in our sense, was just an insubstantial dream; it is
also wrong to suppose that it was generally thought to have authority over
all the kingdoms of Catholic Europe. It is true that universal authority
was sometimes claimed on its behalf, but these claims, like those about the
ancient Roman empire that they copied, were rhetorical or polemical. 7
Most often they were made by lawyers engaged in disputes between
emperor and pope that solipsistically ignored not only Byzantium and
non-Christian lands known to western Christendom, but all the kingdoms
within western Christendom that were not involved in the particular
controversy. An anecdote about twelfth-century diplomacy may warn
those who meet excessive claims about empires in other periods or
societies with less good records neither to accept them nor to discount
them too easily.
In 1157 King Henry II of England sent a letter to the emperor
Frederick I in which he said that his kingdom was under Frederick’s
authority and that all within it would be ordered at the emperor’s nod
and command. He also said that he was sending Frederick rich presents,
which one could perhaps interpret as tribute to a superior. But, as Karl
Leyser pointed out nearly thirty years ago, historians who have used the
letter to show that other kings accepted the empire’s claims to authority
over them have ignored its ending. The letter ended with a brief remark
that Henry’s ambassadors would give by word of mouth his reply to what
Frederick had said (in a now lost letter) about the hand of St. James. This
needs explanation. Henry’s mother had been the daughter of Henry I of
England, and before marrying Henry II’s father, had married an earlier
emperor and king of Germany, also confusingly called Henry. They had
no children, and when he died in 1125 she came home to her father,
bringing a good deal of loot from the imperial treasure, including at least
two crowns and the hand of St. James. The hand had been a precious
relic in the imperial chapel and, by the time Frederick was asking Henry
II to return it, enjoyed the same status in Reading abbey. Henry II gave
his ambassadors the unpleasant task of saying ‘No’, Frederick could not
have his hand back. It remained at Reading.8 Flattering the emperor with
6
William fitz Stephen, ‘Vita . . . sancti Thomasii’, in Materials for the History of Thomas Becket,
ed. J. C. Robertson (7 vols., 1875–85), iii. 99–101; cf. John of Salisbury, Letters, ed. W. J. Millor
and others (2 vols., Edinburgh, 1955– 79) (hereafter John of Salisbury), i. 202– 6; L. Scales, ‘Late
medieval Germany: an under-Stated nation?’, in Power and the Nation in European History, ed.
L. Scales and O. Zimmer (Cambridge, 2005), pp. 166 – 91.
7
L. E. Scales, ‘France and the empire: the view of Alexander of Roes’, French History, ix
(1995), 394 – 416 (and references in n. 80).
8
K. Leyser ‘Frederick Barbarossa, Henry II and the hand of St. James’, Eng. Hist. Rev., xc (1975),
481–506 (repr. in K. Leyser, Medieval Germany and its Neighbours, 900 –1250 (1982), pp. 215–40).
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Empires: a problem of comparative history
155
professions of humble obedience was good cover for denying his request.
The episode does not demonstrate either the overweening pretensions of
the medieval empire or their emptiness. Frederick I was a powerful ruler,
even if his military campaigns, like those of other powerful rulers (even
today), did not always go according to plan. He presumably thought the
recovery of St. James’s hand worth the exertion of diplomatic pressure on
the king whose mother had taken it from Germany. What the story
illustrates is the importance of applying a modicum of scepticism to the
language of international relations.
That is not to say that diplomatic language or rhetoric is not historically
significant. It clearly is. Henry’s letter was kept and used by Frederick’s
spin-doctors as evidence of his greatness and that of his empire. The claim
to universal or quasi-universal rule mattered, but on a rhetorical level,
like, I suggest, similar claims about other empires. Rulers who spent as
much time and trouble on defending or advancing their borders as did,
for instance, those of ancient Rome, Byzantium or China, were clearly
aware on a practical level that they did not really rule the world in the
way that some of their flatterers suggested.9 Calling one’s country the
Middle Kingdom, for example, shows that while, like most people, one
thinks that one is the centre of the world one also knows that there are
other kingdoms outside. How different is the rhetoric of universalism
from that of talking about an empire on which the sun never sets?
In the later middle ages, although other western kings did not, I think,
call themselves emperors, their kingdoms were sometimes called empires.
Henry VIII of England is sometimes said to have made new and
revolutionary claims by calling England an empire, but the idea, with its
implication of absolute independence, was not a new one in England and
had long been a commonplace in other kingdoms where Roman law was
more studied.10 The word imperium and its vernacular equivalents were
used by western Europeans, to judge from a casual and superficial survey,
9
F. Millar, ‘Emperors, frontiers and foreign relations 31 B.C. to A.D. 378’, Britannia, xiii
(1982), 1–23; C. R. Whittaker, Frontiers of the Roman Empire: a Social and Economic Study
(Baltimore, Md., 1997); D. Obolensky, The Byzantine Commonwealth: Eastern Europe 500 –1453
(1971), pp. 108–21, 159 – 62, 200 –21; J. Gernet, ‘Comment se présente en Chine le concept
d’empire’, in Le concept d’empire, ed. M. Duverger (Paris, 1980), pp. 397 –414; N. Standen,
‘Frontiers of 10th-century north China’, in Frontiers in Question: Eurasian Borderlands, 700 –1700,
ed. D. Power and N. Standen (Basingstoke, 1999), pp. 55– 79; M. Rossabi, ‘Introduction’, in
China among Equals: the Middle Kingdom and its Neighbours, 10th–14th Centuries, ed. M. Rossabi
(Berkeley, Calif., 1983), pp. 1–13.
10
John of Salisbury, ii. 580; Dictionary of Medieval Latin from British Sources (Oxford, 1975–),
i. 1249; A. A. M. Duncan, The Kingship of the Scots, 842–1292: Succession and Independence
(Edinburgh, 2002), pp. 115, 260, 270 –1; F. Calasso, I glossatori e la teoria della sovranità (3rd edn.,
Milan, 1957), pp. 22 – 40; Folz, pp. 114–23; P. Munz, Frederick Barbarossa: a Study in Medieval
Politics (1969), p. 233; W. Ullmann, ‘The development of the medieval idea of sovereignty’,
Eng. Hist. Rev., lxiv (1949), 1–33; J. R. Strayer, Les gens de justice du Languedoc sous Philippe le
Bel (Toulouse, 1970), p. 44 n.; M. J. Rodriguez-Salgado, ‘Christians, civilized and Spanish:
multiple identities in 16th-century Spain’, Trans. Royal Hist. Soc., 6th ser., viii (1998), 233–53.
Historical Research, vol. 79, no. 204 (May 2006)
© Institute of Historical Research 2006.
156
Empires: a problem of comparative history
primarily for the Roman empire and the German empire, but sometimes
for any large or important and independent polity. The sense of territories
conquered overseas occurs occasionally in Spanish and English from the
seventeenth century, but did not apparently become common until latish
in the eighteenth.11 Meanwhile the title of emperor remained above all a
mark of status. An emperor did not necessarily have authority over kings
but he ranked above them.
As Europeans came into contact with more peoples elsewhere they
used either what they could make of the words, like khan or sultan, that
those people used for their rulers or such of their own words as seemed
suitable. While Marco Polo seems to have thought generally in terms of
kings and kingdoms and called the great khan the king of kings rather
than emperor, the thirteenth-century missionaries called him emperor.
Other rulers whom they met they called kings or used some version of
those rulers’ own titles.12 Later on emperor and empire came to be used
generally for any Asian rulers who seemed so rich and powerful and
maybe, above all, so dangerous that they deserved the grander title. 13
How the ruler of Japan came to be called an emperor is puzzling, at least
to me. In the seventeenth century European traders there gave the title
to the shogun and saw the other dignitary, whom the Japanese called the
tenno, as a sort of pope. In the nineteenth century they began to call him
the mikado. When the shogun was ousted at the Meiji restoration of
1868, they transferred the title of emperor to the mikado.14 Yet Japan
surely fitted none of the usual notions of an empire until the very late
nineteenth century.
In the eighteenth century, although European diplomats and politicians,
to judge from what one of them wrote, thought that ‘a pompous style
11
Oxford English Dictionary (2nd edn., 20 vols., Oxford, 1989), v. 187– 8; A. O. Meyer, ‘Der
britische Kaisertitel zur Zeit der Stuarts’, Quellen und Forschungen aus italienischen Archiven und
Bibliotheken, x (1907), 231– 7; Dictionnaire historique de la langue française, ed. A. Rey (2 vols.,
Paris, 1993), i. 682; Le Grand Robert de la langue française (6 vols., Paris, 2001), ii. 2054–5;
J. Fisch, ‘Imperialismus I–II’, in Geschichtliche Grundbegriffe, ed. O. Brunner and others (8 vols.,
Stuttgart, 1972–92), iii. 173–5; Diccionario de la lengua castellana (9 vols., Madrid, 1726 – 39), iv.
224 (facsimile repr. as Diccionario de autoridades (3 vols., Madrid, 1984)); J. H. Elliott, Spain and
its World 1500 –1700 (New Haven, Conn., 1989), pp. 7 – 91; J. Lynch, Spain 1516 – 98: from Nation
State to World Empire (Oxford, 1991), pp. 95– 6.
12
Marco Polo, Voyages (Paris, Société de Géographie, Recueils de Voyages, i, 1824), e.g.,
pp. 301, 304, 312, 347, 466; Sinica Franciscana: Itinera et Relationes Fratrum Minorum saeculi XIII
et XIV, ed. A. van den Wyngaert (7 vols., Florence, 1929), i. 29 – 30, 137 142, 188, 201;
Matthaei Parisiensis, Chronica Majora, ed. H. R. Luard (7 vols., Rolls ser., 1872 –84), iii. 488– 9.
13
K.-H. Ruffmann, ‘England und der russische Zaren- und Kaisertitel’, Jahrbücher für
Geschichte Osteuropas, new ser., iii (1955), 217 –24.
14
History of Japan compiled from the Records of the East India Company, comp. P. Pratt (2 vols.,
New York, 1972), i. 24, 32, 38, 350, 85; J. W. Hall, Japan from Prehistory to Modern Times (New
York, 1970), pp. 247, 302– 7; H. Bolitho, ‘Japanese kingship’, in Patterns of Kingship and
Authority in Traditional Asia, ed. I. Mabbett (1985), pp. 24–43, although M. Satow’s work
(referred to there, n. 45) does not seem to support Bolitho’s exact dates.
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Empires: a problem of comparative history
157
and . . . embellished letters’ pleased more barbarous nations,15 it was not
only barbarians who cared about titles, rank and status. After the Holy
Roman Emperor had lost virtually all his authority over the princes who
ruled different parts of Germany, the kings of England and France
resented more than before any suggestion that they were inferior to him. 16
Russia, too, was beginning to be a problem in international relations. When
the rulers of Muscovy began to call themselves tsars, a title obviously
derived through Byzantium from Rome, translating this as emperor
seemed harmless enough to western rulers and diplomats – so long as
Russia was seen as a more or less barbarous country outside the circle of
European diplomacy. When Russia entered that circle both the English
and the Austrian governments, for their different reasons, refused for
some time to allow the tsar to count as a proper emperor.17
The words emperor and empire seem to have become even more
detached from each other than they had been when medieval kingdoms
claimed to be empires because they were fully independent even though
their kings were not emperors. Rulers of the kind of overseas empires of
conquest to which the word empire was beginning to be applied in the
eighteenth century were not normally emperors. Emperor was a title that
no longer depended on size, independence or conquest, whether overseas
or not. If held by an established European ruler it was a mark of status
that was virtually immune to forfeiture. After Napoleon had complicated
things by becoming emperor over real, though short-lived, conquests and
had abolished the Holy Roman Empire, its emperor nevertheless
remained an emperor, as emperor of Austria. Both the Second Empire in
France and the German empire of 1870 got their titles, I imagine, not
because of conquests but because Napoleon III was the heir to his uncle’s
glory while the Kaiser, with evocations of the medieval empire, had to
be put above the other kings within Germany and on a level with the
emperor of Austria-Hungary.
All this probably impinges on the work of historians of modern empires
only in so far as it can be difficult to free oneself from assumptions
embedded in the established use of key words. Many historians are
reluctant to get bogged down in definitions, often because they are more
interested in one individual polity than in fitting it into a category: that
is entirely reasonable, but it can make their work hard to use
comparatively. Those who translate as empire and emperor the words
used in non-European sources for rulers and polities without further
explanation evidently assume that everyone knows what the words empire
15
Ruffmann, pp. 222–3.
H. Duchhardt, ‘Imperium und regna im Zeitalter Ludwigs XIV’, Historische Zeitschrift,
ccxxxii (1981), 555–81; P. Burke, The Fabrication of Louis XIV (1992), p. 72.
17
G. Hosking, Russia: People and Empire, 1552–1917 (1997), pp. 5–9, 39 –47; Duchhardt,
pp. 577 – 81.
16
Historical Research, vol. 79, no. 204 (May 2006)
© Institute of Historical Research 2006.
158
Empires: a problem of comparative history
and emperor mean.18 But it is hard to be sure that we do when both have
long been used in ways that make the boundaries of the categories they
imply so difficult to draw. Some use the Latin word imperium as if it
represented a more exact concept than ‘empire’, but it is not clear that it
does or why it should.19 As I hope I have shown, it has had many
different meanings. Others avoid the word empire altogether. Quite apart
from the conquest of the U.S.A. itself from sea to shining sea, which
would fit some criteria of empire, the American acquisition and rule of
the Philippines, for example, and of Hawaii and Alaska until they were
made states of the union, suggest that the U.S.A. had something that
should rank as an empire long before it became fashionable to talk about
one.
My argument is not that historians of empires need a tighter definition
of the word, let alone one that would exclude some of the polities
traditionally called empires that seem not to fit the current models. What
I suggest – if it is not impertinent for someone who does not herself do
any kind of imperial history to make any suggestion – is that looking
directly at the variety, not to say confusion, of words and concepts that
we have inherited may make it easier to lay both words and concepts
aside for the moment and concentrate on deciding what seem to be the
most important characteristics of the phenomena with which I think that
historians of empires, particularly more modern empires, are concerned.
If some empires seem to lack certain of those characteristics that is not a
reason for saying they are not empires, but it may help in deciding what
it is most profitable to compare with what. Irrespective of the definition
that anyone uses or implies it is surely necessary when one is making
comparisons to identify the characteristics that one wants to compare.
If one starts by thinking of what we are interested in as, by and large,
relatively large polities that consist of a ruling part (the metropolis) and
other parts (colonies or peripheries) that it dominates as a result of military
conquest or some kind of political or economic bullying, and that are
retained and governed separately from the metropolis rather than being
absorbed in it, then one has to note that not all the polities that look
likely cases seem to have all these characteristics. Size, to start with, has
to be relative, depending on the technology of communications and the
nature of economies: polities comprising areas that support only thinly
scattered populations look more impressive empires on a map than those
that cover smaller but more densely settled and intensively exploited
18
E.g., J. F. Fletcher, Studies on Chinese and Islamic Inner Asia (1995), ch. 7 (‘Turco-Mongolian
monarchic tradition in the Ottoman empire’); G. Rachewilz, ‘Qan, Qa’an and the seal of
Güyüg’, in Documenta Barbarorum: Festschrift für Walther Heissig, ed. K. Sagaster and M. Weiers
(Wiesbaden, 1983), pp. 272 – 81; I do not find clear the distinction between state and empire
in T. J. Barfield, Perilous Frontier: Nomadic Empires and China (Cambridge, Mass., 1989), pp. 5–27.
19
E.g., Barfield, p. 229; B. Porter, The Lion’s Share: a Short History of British Imperialism, 1850 –
2004 (4th edn., 2004), p. 8.
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Empires: a problem of comparative history
159
areas. The Athenian empire is often included in surveys of empires but it
was pretty small as well as short-lived. Some like to talk of Athenian
hegemony rather than empire but, following Finley, I do not think that
is very useful. Hegemony is ambiguous in modern use and the Athenians
really ruled the city-states that they took over, and ruled some of them
brutally.20
Conquest and division into metropolis and colonies are both as difficult
to assess as size. Some conquests were more violent than others, some
acquisitions, such as purchases, may have been quite peaceful, while the
longer an empire lasts the more the initial conquest may fade from memory.
But there seems to be a significant distinction to be drawn between
empires defined as the result of unilateral annexation of other polities and
either voluntary federations or the kind of ‘aristocratic empires’ that
Kautsky discussed, which seem more like what I would consider states –
aristocratic states maybe, although I would want to qualify some of his
ideas about government by traditional aristocracies. These seem to be
significantly different from empires of the kind that most historians are
concerned with, and not only because Kautsky’s aristocratic empires were
uncommercialized or pre-capitalist.21 The differences need to be thought
about.
The ancient Indian Mauryan empire fits the criteria of conquest, size
and division into metropolis and periphery, with some further distinctions
within the peripheral areas.22 The Roman empire fits the model of
conquest and probably also that of size, although it was less exceptionally
enormous than Eurocentric history tends to assume. As for metropolis
and colonies, Rome started with a fairly clear division between first the
city and the rest of Italy and then between Italy and the provinces, but
both divisions became less clear with time.23 China seems to qualify better
on the ground of size than of anything else: after the unification of what
became known as China in 221 B.C.E., when one ruler took over what
had been separate kingdoms, it became at first more like a single, if
enormous, state, although some subordinate rulers were allowed an initial
measure of independence.24 This first unification of China thus looks,
except in size, rather like that of the little medieval kingdom of England,
which was formed out of the conquest and absorption of a number of
mini-kingdoms. The larger and later kingdom of Spain was formed in
20
Finley, ‘The 5th-century Athenian empire’; ‘Introduction’, in Garnsey and Whittaker,
p. 1. D. Lal, In Praise of Empires: Globalization and Order (Basingstoke, 2004), p. 4, citing M. W.
Doyle, Empires (Ithaca, N.Y., 1986), thinks Thucydides distinguished between Athens’s empire
and Sparta’s hegemony but Doyle, pp. 12, 40, 81 makes it clear that the distinction is modern.
21
J. H. Kautsky, The Politics of Aristocratic Empires (Chapel Hill, N.C., 1982), pp. 1–31, 62 – 7.
22
R. Thapar, Early India: from the Origins to A.D. 1300 (2002), pp. 174– 9, 194– 7.
23
P. Garnsey and R. Saller, The Roman Empire: Economy, Society and Culture (1987), p. 9.
24
Cambridge History of China, ed. D. Twitchett and others (Cambridge, 1978–), i. 20, 53–4,
123, 429, 470 –2.
Historical Research, vol. 79, no. 204 (May 2006)
© Institute of Historical Research 2006.
160
Empires: a problem of comparative history
rather the same way, but there some of the constituent kingdoms kept
their separate identities and some of their institutions for centuries, while
being effectively governed from Castile. Peninsular Spain might thus itself
count as an empire quite apart from its overseas dominions.25 But China
did not apparently develop like that. Different dynasties during the long
centuries of what is called the Chinese empire had different centres of
power, so that, although historians refer to ‘metropolitan areas’ around
successive capitals that were differently governed from other provinces, it
is difficult (as in what became known as the Holy Roman Empire) to
identify a single and permanent metropolis in the sense that the word is
used by theorists of empires. China certainly had an outer periphery in
central Asia, where its rulers intermittently controlled, or tried to control,
areas beyond what they ruled more regularly, but a consistent distinction
between metropolis and peripheries within China is hard to find.26 Russia’s
conquests qualified it by size even if the date when it began to count as
an empire is a matter of opinion, but again the metropolis/colony division
is unclear. On the other hand, different as Russia’s nineteenth-century
land-based empire may look from the overseas colonies of other European
powers of the time, it is less exceptional if one takes a longer view. 27
Leaving aside cases of the absorption of neighbouring areas into the
kingdoms of medieval England or France, or into the U.S.A., the
distinction between metropolis and colonies also does not seem to fit, for
instance, all stages of the early Islamic empire or the Ottoman empire. I
do not, incidentally, take seriously the claim that in the last years of
French rule Algeria was part of metropolitan France.28
Variations in size, in methods of acquisition and in methods of
government clearly offer scope for comparisons that raise questions about
individual cases and the categories into which we put them. I propose to
look briefly at methods of government, which, I suggest, need to be seen
in the context of the norms and assumptions of those who made
conquests or acquisitions. Empires have been acquired not only by
monarchies but by what are called republics or even democracies –
although it is surely slightly misleading to go on calling Athens a
democracy, given any of the senses that the word has in current English.
Leaving that aside, there is the fairly obvious point that, however
metropolitan states rule their colonies, they do not themselves need to be
25
Rodriguez-Salgado; Lynch, pp. 1–45, 96 –100, 211–49.
Cambridge History of China, i. 429, 505; iii. 160, 293; vii. 10 –11. Barfield, pp. 2–27; J. A.
Millward and P. C. Perdue, ‘Political and cultural history’, in Xinjiang: China’s Muslim
Borderland, ed. F. S. Starr (New York, 2004), pp. 27 – 62.
27
Hosking, pp. 39 –40; D. Lieven, ‘The Russian empire and the Soviet Union as imperial
polities’, Jour. Contemp. Hist., xxx (1995), 607 –36; and D. Lieven, The Russian Empire and its
Rivals (2000).
28
I. S. Lustick, Unsettled States, Disputed Lands: Britain and Ireland, France and Algeria, Israel
and the West Bank-Gaza (Ithaca, N.Y., 1993), pp. 89 –112.
26
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Empires: a problem of comparative history
161
ruled autocratically, let alone by someone called an emperor, to count
usefully as empires.
One point that I also want to make may be less obvious to some
modernists. What I know about the European middle ages, and what I
have picked up through patchy reading on traditional societies outside
Europe, suggests that many, however arbitrary, unjust and oppressive
their governance seems to us, appear to have been able to rely on a good
deal of voluntary submission. Rulers ruled, and were supposed to rule,
by and with the advice and consent of their great men, while the
customary law that was practised in many polities, even if declared and
enforced by the ruler, was supposed to be made and applied by some kind
of collective judgements. Kingdoms and other units of government
were often regarded as natural communities of common descent, law and
culture.29 When neighbouring territories that they conquered had similar
cultures and customs that had not yet been fossilized by written records,
they may thus have been relatively easy to absorb and integrate into the
cultures and solidarities of the conquerors. The distinction between
metropolis and periphery could therefore be lost – as it was in some
medieval European kingdoms, and apparently in early China. How far
absorption was helped or hampered by allowing acquired territories to
keep separate institutions and bits of their own law seems to have varied.
Both in Spain and in central and eastern Europe, for instance, the special
privileges of particular acquired areas tended to become eroded with
time, but separate names and collective identities could still be preserved
in these quasi-imperial situations.
It might be reasonable to guess that what preserved conquests as empires
rather than turning them into new provinces within the conquering state
may have been the practicalities of distance and communication. Or was
it that they started with such fundamental differences of culture, customs
of government and physical appearance of the inhabitants as to make
absorption impossible or difficult? Or that, as in the case of some of the
nomadic empires of Asia, all that the new ruler wanted to do anyway was
to take tribute from conquered territories? Or that violent conquest and
resistance left a legacy of hostility and violence? Or a combination of any
of these?
Distance and difference of culture and solidarities may go together but
need not: they did not, for instance, in the case of English conquests in
Wales and Ireland. In the example of Wales, in particular, distances were
minimal even in medieval terms. Although they fairly obviously went
together in European conquests overseas, cultural differences may have
been exaggerated by conquerors who could not or would not look
carefully at societies that they wanted to conquer and profit from anyway.
29
S. Reynolds, ‘The idea of the nation as a political community’, in Scales and Zimmer,
pp. 54– 66, and works cited there.
Historical Research, vol. 79, no. 204 (May 2006)
© Institute of Historical Research 2006.
162
Empires: a problem of comparative history
In Asia it was easier to ignore existing laws and rights of property if one
classed a society as an oriental despotism. Societies in north America, on
the other hand, could be seen as in a state of nature, with no property
rights at all. In practice a good many conquests (how many?) started with
mere tribute-taking and sometimes used existing systems and officials,
even those of rulers condemned as despots. However strange other aspects
of a newly conquered society may be, conquerors seem to find out about
their taxes quite easily. But, easy as it may be to take tribute once or
twice, one tribute does not make an empire. Getting payments and dues
regularly, and in amounts to satisfy growing demands, may then lead to
rebellion, fiercer enforcement and more formal and oppressive government.
How does this affect the metropolitan government itself and its norms?
Although oppression can be just as fierce in states without empires, I
wonder how much and in which empires repression abroad has affected
ideas about the rights of subjects or citizens at home. Whether they are
called subjects or citizens is, incidentally, I suggest, more a matter of
words than of substantive rights, although the different words may suggest
new ideas to those concerned. How much does empire lead either rulers
or their metropolitan subjects to think new thoughts about rights and law
in general and in their own country?
Comparison suggests that variations in the ways empires have been
ruled are endless and raises a lot of questions about possible reasons for the
variations. It may seem obvious that empires should have been nurseries
of bureaucracy, needing tax-lists and systems of written communication
between centre and periphery, if no more. Leaving aside the more fleeting
nomadic empires, the tendency seems to have been fairly general. China,
of course, although in some other ways non-imperial, was famously
bureaucratic, while the way that the Roman empire, although it started
with a very limited bureaucracy and delegation of a good deal of everyday
government to local elites, became more bureaucratic with time, may
illustrate the connection well enough.30 On the other hand, while empires
may tend to need bureaucracy, bureaucracy does not need empires. In
post-Roman western Europe it developed most notably in Italian citystates, the small kingdom of England and the papacy, which ruled neither
a state nor an empire.
Even the most centralized and bureaucratic governments, moreover,
whether of empires or any but the very smallest states, need to delegate
authority. They needed to do so especially when lines of communication
were long and travel was slow. Indirect rule through trading companies
was one way of managing far-flung empires; giving local governors some
freedom in making decisions was another; while co-opting existing local
elites and leaving local government to them might produce a system of
30
Garnsey and Saller, pp. 20 – 40; C. M. Kelly, ‘Later Roman bureaucracy’, in Literacy and
Power in the Ancient World, ed. A. K. Bowman and G. Woolf (Cambridge, 1994), pp. 161– 76.
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Empires: a problem of comparative history
163
government rather like that of the metropolis. Leaving local elites in
charge presumably worked more easily insofar as it maintained traditional
units of government, as in the Roman empire, so that the bulk of the
population might be relatively unaffected by the conquest until demands of
government became so heavy as to provoke revolt, or, in modern empires,
until ideas of new kinds of popular government or democracy began to
percolate down. I refer here to ideas of democracy rather than nationalism
because the ideas of solidarity and political community that I have argued
were attached to units of government long before the eighteenth century had,
I maintain, a good deal in common with what is now called nationalism. 31
While they served to support and validate existing states, they could also
console the elites of former states that preserved some of their traditional
institutions within larger states or empires. It was the combination of
these old solidarities with new ideas of the rights of man that created the
kind of modern nationalism that has frustrated modern empires.
Settler colonies seem, at least at first, to have been easier to keep on a
light rein. Some, indeed, started as completely independent, like Iceland
and Greenland – nice exceptions to the trend from empire to
independence – or like the medieval German colonies in kingdoms to the
east, which had no formal attachment to the kingdom of Germany but
received privileges from their local kings.32 Some early modern emigrant
colonies may have been kept under metropolitan government as much to
keep out rivals as because they were initially seen as particularly valuable,
except as places to send poor and landless troublemakers. Assuming that
emigrants from the home country could be left a fair degree of autonomy
because they shared the values and laws of the metropolis could work
well – as the white British commonwealth in the First World War might
suggest if it were not that Britain got even greater service in the war from
India. It could, however, turn out to be a mistake, either because settlers
used their old values to question the taxes and controls imposed on them
or because their own demands on native populations or imported slaves
provoked more resistance than they could handle on their own. Settler
colonies were, after all, never populated only by free settlers: slaves,
indentured servants and convicts in penal colonies all tended to complicate
relations between metropolis and periphery. Nor were any colonies
without settlers of some kind, even if only temporary ones, who would
or could go home or go elsewhere. There were generally governors,
garrisons and administrators, some of whom became too corrupt or
tyrannical, while others, in the British phrase, ‘went native’. Did which
of these happened depend on distance, geographical or cultural, on the
amount of central control, or on changes in metropolitan policies? Where
31
Reynolds, ‘The idea of the nation’.
These would not be colonies at all by Finley’s classification (M. I. Finley, ‘Colonies – an
attempt at a typology’, Trans. Royal Hist. Soc., 5th ser., xxvi (1976), 167 – 88).
32
Historical Research, vol. 79, no. 204 (May 2006)
© Institute of Historical Research 2006.
164
Empires: a problem of comparative history
indigenous populations seemed to the empire-builders entirely what we
now call ‘Other’ one would expect them to be ruled more unsympathetically
and oppressively, but that may be too simple. Otherness has been measured
so differently at different periods, by cultural or physical differences or by
class: native rulers and elites might seem less ‘Other’ than peasants, let
alone slaves, but some who were less ‘Other’ to their British rulers, like
middle-class Indians who received a western education, could arouse
more fear and hostility from their rulers than peasants who could more
easily be seen as the salt of the earth and gratefully devoted to Us.
How consistently the reasons for acquiring any empire affected the way
it was then governed seems doubtful. Modern historians tend to stress
economic motives, but in the long ages before governments had economic
policies a ruler’s conquests, although perhaps starting with raids for slaves
and loot (which presumably count as economic), were surely often
motivated more by a drive to defend and extend his power and prestige.
Economic purposes become most obvious with the European search for
spices in the east and gold and silver in the west, but in these adventures
more or less independent entrepreneurs could be as important as the ruler
who ended up with an empire. Then there were missions, some undertaken
with complete sincerity, others not, whether they were intended to spread
a religion, law to lesser breeds without it, or the benefits of commerce
and taxation. Whatever their purpose and sincerity, missionaries had
varying and changing relationships with imperial governments and
varying impacts on the way that they governed.
Since the tendency to self-justification seems common in human beings,
and rulers like to be loved and to be considered legitimate, it is not
surprising that many empires produced ideologies of benevolent imperialism,
but it might be rash to assume that these ideologies were very different
from those of benevolent governance in non-imperial states. The degree
to which a metropolitan society’s own ideology was changed by having
an empire, or whether the two were kept separate, and if so how, seems
to me worth investigating – for all periods and empires. All these
questions are particularly difficult for periods in which one has to deduce
reasons from results, and possible motives from what look like possible
reasons. Even with better sources, the numbers of people with varying
motives and interests involved in conquering, governing, settling and
trading in any empire, combined with the force of muddle in human
affairs, make the problem of disentangling causes and effects almost
insoluble – but still worth thinking about? How far did any empire (apart
from the much discussed Spanish and British empires) increase the power
or wealth of its metropolis, and of what sections of its population? The
correlation of relative oppression with official policy, with degrees of
discontent or with the fall of the empire is liable to be complex.
I cannot, however, go into the reasons for the fall of empires,
except to say that ‘the natural and inevitable effect of immoderate
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Empires: a problem of comparative history
165
greatness’,33 like a good many other less lapidary judgments on the Roman
empire, looks even less convincing when one compares it with nonEuropean empires. Nor can I go into the meaning of imperialism or the grey
area of quasi-empires – whether labelled as hegemonic, invisible, informal,
soft power, or simply put into quotation marks as ‘empires’. Some of the
labels are helpful in reminding us that degrees of effective control correlate
poorly with formal connections, but some of them have been chosen for
particular cases in a way that impedes comparison. Seeing British commercial
activity in south America in terms of informal empire was illuminating about
nineteenth-century British history and the growth of the British empire
but perhaps looks less easy to interpret in terms of empire in general
when United States involvement alongside is brought into the picture. 34
Anyone who works on any of the empires that I have mentioned in this
rush through time and space will realize how little I know about their
own field. I hope that my mistakes and misunderstandings will not enable
them to discount entirely my plea for more comparisons. The subject
cries out for more of them, including, I submit, comparisons between
periods, which some historians do not approve of – although disapproval
does not always inhibit carefree allusions to the Roman or Chinese
empires. Even for those who are only interested in one case or one
period, comparison with others may surely be illuminating in raising
questions about assumptions and chains of cause and effect. The real
argument against comparison is that it is such hard work, especially when
it goes over different periods and demands a range of different skills, and
especially if it involves more than looking outside one’s own field to pick
something out of its context elsewhere because it looks like something in
one’s own. Even at the most superficial level there is another
impediment. Historians working in different periods and in different
national traditions of historiography, even if they do not have conscious
ideological differences, tend to work from different presuppositions and
within different paradigms. They therefore focus on different subjects and
use different terminologies (or the same terminology in different senses),
so that it can be hard, especially if one cannot read their sources, to be
sure that one is comparing like with like. I have found this when
comparing just a few aspects of a few kingdoms in medieval Europe.
Empires, however defined, pose much greater problems, even if one
ducks (as I have) the question of whether they were or are a Good Thing
or a Bad Thing. But grappling with these problems is a matter for the
historians of empires, not for a single lecture by a non-imperial historian.
33
E. Gibbon, Decline and Fall of the Roman Empire, ed. D. Womersley (5 vols., 1994), ii. 509
(ch. 38, ‘General observations’).
34
J. Gallagher and R. Robinson, ‘The imperialism of free trade’, Econ. Hist. Rev., 2nd ser.,
vi (1953), 1–15; C. M. Dobbs, ‘The Monroe Doctrine’, in The Dictionary of American History,
ed. S. I. Kutler (3rd edn., 10 vols., New York, 2003), v. 446– 7.
Historical Research, vol. 79, no. 204 (May 2006)
© Institute of Historical Research 2006.
Empires:
Article
aResearch
problem
ofLtd
comparative
history
Blackwell
Oxford,
Historical
HISR
©
0950-3471
XXX
Institute
UK
Publishing
of
Historical
Research 2006
Empires: a problem of comparative history*
Susan Reynolds
Institute of Historical Research
Abstract
Most historians of empires probably start by assuming that what they are interested
in are relatively large polities that consist of a ruling part (the metropolis) and
other parts (colonies or peripheries) that it dominates as a result of military
conquest or political or economic bullying, and that are retained and governed
separately from the metropolis rather than being directly absorbed in it. Not all
the polities called empires over the centuries, however, have had all these
characteristics. Surveying some of the variations in characteristics may help in
deciding what it may be most profitable to compare with what.
The first problem to confront in comparing empires is what kind of polity,
and which individual polities, we are to count as empires. One cannot begin
to make serious comparisons without deciding what comes into the category
of cases one wants to compare. It is not so much a matter of defining the
word ‘empire’ as of, first, deciding which phenomena – which actual
polities, past and present – that those who use the word refer to, and
deciding what characteristics these phenomena share that make them a
category; and second, considering whether there are other phenomena
that share significant characteristics with the first group and therefore
ought to be brought into the discussion even if they are not usually called
empires. ‘Empire’ is a European word, derived from Latin, while some
of the polities that we need to consider were outside Europe and came
to be called empires by Europeans for various reasons, some of which, as
I shall argue, were not based on any clear or consistent concept or notion
of what an empire is. The relationship between word and phenomenon
is at best indirect, mediated by the notions or concepts that users or
hearers of the word have of either the word or the phenomenon. 1
* This article is a revised version of a plenary lecture given at the Anglo-American Conference,
Beveridge Hall, University of London, July 2005. I am grateful to participants, and particularly
to Pamela Crossley, for their suggestions.
1
Words, concepts and phenomena are discussed (under different names) in C. K. Ogden and
I. A. Richards, The Meaning of Meaning: a Study of the Influence of Language upon Thought and the
Science of Symbolism (1923), pp. 13–15. There are further discussions in, e.g., J. Lyons, Semantics (2 vols.,
Cambridge, 1977), i. 95–119, 175; and R. Tallis, Not Saussure: a Critique of Post-Saussurean Literary
Theory (Basingstoke, 1988), pp. 114-16. As applied to medieval history, see R. Schmidt-Wiegand,
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Published by Blackwell Publishing Ltd., 9600 Garsington Road, Oxford, OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.
152
Empires: a problem of comparative history
It may be useful to make a preliminary distinction of empires, as I use
the word, from states. For present purposes I ignore all the arguments
about the monopoly (or control) of the legitimate use of force, about
boundaries, about when ‘true states’ appeared and whether the word is
only to be used about modern polities, and so on.2 States I take to be more
or less independent or autonomous polities, enforcing their own laws (or
failing to do so), which, whether unitary or federal in constitution, have
a different and less divided character from the polities that I am going to
talk about and that for convenience I lump together as empires. Prima
facie, and for what it is worth at this stage, I take the difference between
federal states and empires to be that the component parts of federal states are
supposed to be equal, while the polities that I am considering as empires
are normally thought of as consisting of a ruling polity (sometimes called
the metropolis) and subordinate polities (peripheries or colonies). They
came together, moreover, not by a voluntary (or supposedly voluntary)
federation but on the initiative of what thus became the metropolitan
polity, often by some kind of armed conquest. This working definition
is not intended to exclude any other, but merely to serve for this
discussion. It raises problems about the distinction between metropolis
and colonies, to which I shall return, and about whether to call each of
the colonies a state (as did, among others, Edmund Burke and Moses
Finley)3 or to reserve that title for the metropolis, or whether the whole
thing constitutes a single state. I shall not return to that as it seems to me
an unhelpful argument merely about words.
Words, nevertheless, are my starting point, since they are what we have
in the sources and what we must start from when considering notions and
the phenomena that they represented. That means starting with Rome
and going on through medieval Europe. I realize that most historians of
empires are primarily interested in modern ones. Most surveys of empires
tend to jump over the middle ages, but both Rome and the medieval
west matter: our notions of empire, like the word itself, started from
Rome, were adapted by medieval use, and filtered through the rather
narrow view of Greek and Roman history imparted by centuries of
classical education. For Europeans the Roman empire long remained the
‘Historische Onomasiologie und Mittelalterforschung’, Frühmittelalterliche Studien, ix (1975),
49 – 78; and S. Reynolds, Fiefs and Vassals: the Medieval Evidence Reinterpreted (Oxford, 1994),
pp. 12–14.
2
S. Reynolds, ‘The historiography of the medieval state’, in Companion to Historiography, ed.
M. Bentley (1997), pp. 117 – 38; and S. Reynolds, ‘There were states in medieval Europe: a
response to Rees Davies’, Jour. Historical Sociology, xvi (2003), 550 –5.
3
Edmund Burke, Writings and Speeches, iii: Party, Parliament and the American War, 1774 –80,
ed. P. Langford (Oxford, 1996), p. 132; M. I. Finley, ‘The 5th-century Athenian empire: a
balance sheet’, in Imperialism in the Ancient World, ed. P. D. Garnsey and C. R. Whittaker
(Cambridge, 1978), pp. 103–16.
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Empires: a problem of comparative history
153
prototype empire, the great result of great conquests, with supreme, even
virtually universal, authority. Much as I have learned from sketchy reading
about non-European and modern empires, I still think that looking at the
way earlier Europeans referred to empires and took their words and ideas
about them to the rest of the world explains some of the anomalies of
modern usage and the varying phenomena included in the whole category.
That, of course, is what a medievalist who has rashly undertaken to talk
about empires in general would say.
Romans used the word imperium for all kinds of authority and power
from that of heads of households up. When they used it for what we call
the Roman empire they apparently only occasionally gave it a territorial
connotation, as when they said that Palmyra lay between the summa
imperia of Romans and Parthians. The word imperator also started with
wider uses before it came to be used as the title of the emperor.4 After
the collapse of the western half of the empire, while a Roman emperor
still, of course, continued to rule in Byzantium, rulers in the west could
also be called emperors if they had or claimed authority over other kings
and kingdoms, that is, over polities that would have normally been
thought of as independent.5 Kings who claimed some kind of authority,
for instance over other kings within Britain in the early middle ages and
within Spain for longer, could be said to have imperium or to be called
imperatores. But much more formally and importantly the title belonged to
Charlemagne and his successors whose coronation in Rome symbolized
their claim to revive its ancient glories and, after them, to German kings
who were similarly crowned there as emperors. It is sometimes said that
the title was an empty one, but both Carolingian and, until at least the
end of the twelfth century, German emperors were the most powerful
rulers of Europe. Since they ruled multiple kingdoms, with their power
base in the kingdom of Germany, the rest of their dominions could rank
in our terms as peripheries or colonies, with Germany as the metropolis.
Their empire could therefore, for what it is worth, come within the
modern category, allowing perhaps for a shifting kind of inner metropolis
in whatever part of the kingdom of Germany the current emperor had
his particular base. In the twelfth century a foreigner could refer to the
emperor as emperor of Germany, but it was not until imperial authority
had become restricted to Germany, and within Germany, in the later
4
Oxford Latin Dictionary, ed. P. G. W. Glare (Oxford, 1982), pp. 842– 4; P. A. Brunt, ‘Laus
imperii’, in Garnsey and Whittaker, pp. 159 – 91.
5
D. Bullough, ‘Emperors and emperordom from late antiquity to 799’, Early Medieval Europe,
xii (2003), 377 – 87; R. Folz, L’idée d’empire en occident du ve au xive siècle (Paris, 1953), pp. 66 –
81, 114–23; J. L. Nelson, ‘Kingship and empire’, in The Cambridge History of Medieval Political
Thought c.350 –c.1450, ed. J. H. Burns (Cambridge, 1988), pp. 211–51; and J. L. Nelson,
‘Kingship and empire in the Carolingian world’, in Carolingian Culture: Emulation and Innovation,
ed. R. McKitterick (Cambridge, 1994), pp. 52–87.
Historical Research, vol. 79, no. 204 (May 2006)
© Institute of Historical Research 2006.
154
Empires: a problem of comparative history
middle ages that the empire came to be officially called the Holy Roman
Empire of the German Nation.6
It is not only wrong to suppose that the medieval western empire, far
from being an empire in our sense, was just an insubstantial dream; it is
also wrong to suppose that it was generally thought to have authority over
all the kingdoms of Catholic Europe. It is true that universal authority
was sometimes claimed on its behalf, but these claims, like those about the
ancient Roman empire that they copied, were rhetorical or polemical. 7
Most often they were made by lawyers engaged in disputes between
emperor and pope that solipsistically ignored not only Byzantium and
non-Christian lands known to western Christendom, but all the kingdoms
within western Christendom that were not involved in the particular
controversy. An anecdote about twelfth-century diplomacy may warn
those who meet excessive claims about empires in other periods or
societies with less good records neither to accept them nor to discount
them too easily.
In 1157 King Henry II of England sent a letter to the emperor
Frederick I in which he said that his kingdom was under Frederick’s
authority and that all within it would be ordered at the emperor’s nod
and command. He also said that he was sending Frederick rich presents,
which one could perhaps interpret as tribute to a superior. But, as Karl
Leyser pointed out nearly thirty years ago, historians who have used the
letter to show that other kings accepted the empire’s claims to authority
over them have ignored its ending. The letter ended with a brief remark
that Henry’s ambassadors would give by word of mouth his reply to what
Frederick had said (in a now lost letter) about the hand of St. James. This
needs explanation. Henry’s mother had been the daughter of Henry I of
England, and before marrying Henry II’s father, had married an earlier
emperor and king of Germany, also confusingly called Henry. They had
no children, and when he died in 1125 she came home to her father,
bringing a good deal of loot from the imperial treasure, including at least
two crowns and the hand of St. James. The hand had been a precious
relic in the imperial chapel and, by the time Frederick was asking Henry
II to return it, enjoyed the same status in Reading abbey. Henry II gave
his ambassadors the unpleasant task of saying ‘No’, Frederick could not
have his hand back. It remained at Reading.8 Flattering the emperor with
6
William fitz Stephen, ‘Vita . . . sancti Thomasii’, in Materials for the History of Thomas Becket,
ed. J. C. Robertson (7 vols., 1875–85), iii. 99–101; cf. John of Salisbury, Letters, ed. W. J. Millor
and others (2 vols., Edinburgh, 1955– 79) (hereafter John of Salisbury), i. 202– 6; L. Scales, ‘Late
medieval Germany: an under-Stated nation?’, in Power and the Nation in European History, ed.
L. Scales and O. Zimmer (Cambridge, 2005), pp. 166 – 91.
7
L. E. Scales, ‘France and the empire: the view of Alexander of Roes’, French History, ix
(1995), 394 – 416 (and references in n. 80).
8
K. Leyser ‘Frederick Barbarossa, Henry II and the hand of St. James’, Eng. Hist. Rev., xc (1975),
481–506 (repr. in K. Leyser, Medieval Germany and its Neighbours, 900 –1250 (1982), pp. 215–40).
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Empires: a problem of comparative history
155
professions of humble obedience was good cover for denying his request.
The episode does not demonstrate either the overweening pretensions of
the medieval empire or their emptiness. Frederick I was a powerful ruler,
even if his military campaigns, like those of other powerful rulers (even
today), did not always go according to plan. He presumably thought the
recovery of St. James’s hand worth the exertion of diplomatic pressure on
the king whose mother had taken it from Germany. What the story
illustrates is the importance of applying a modicum of scepticism to the
language of international relations.
That is not to say that diplomatic language or rhetoric is not historically
significant. It clearly is. Henry’s letter was kept and used by Frederick’s
spin-doctors as evidence of his greatness and that of his empire. The claim
to universal or quasi-universal rule mattered, but on a rhetorical level,
like, I suggest, similar claims about other empires. Rulers who spent as
much time and trouble on defending or advancing their borders as did,
for instance, those of ancient Rome, Byzantium or China, were clearly
aware on a practical level that they did not really rule the world in the
way that some of their flatterers suggested.9 Calling one’s country the
Middle Kingdom, for example, shows that while, like most people, one
thinks that one is the centre of the world one also knows that there are
other kingdoms outside. How different is the rhetoric of universalism
from that of talking about an empire on which the sun never sets?
In the later middle ages, although other western kings did not, I think,
call themselves emperors, their kingdoms were sometimes called empires.
Henry VIII of England is sometimes said to have made new and
revolutionary claims by calling England an empire, but the idea, with its
implication of absolute independence, was not a new one in England and
had long been a commonplace in other kingdoms where Roman law was
more studied.10 The word imperium and its vernacular equivalents were
used by western Europeans, to judge from a casual and superficial survey,
9
F. Millar, ‘Emperors, frontiers and foreign relations 31 B.C. to A.D. 378’, Britannia, xiii
(1982), 1–23; C. R. Whittaker, Frontiers of the Roman Empire: a Social and Economic Study
(Baltimore, Md., 1997); D. Obolensky, The Byzantine Commonwealth: Eastern Europe 500 –1453
(1971), pp. 108–21, 159 – 62, 200 –21; J. Gernet, ‘Comment se présente en Chine le concept
d’empire’, in Le concept d’empire, ed. M. Duverger (Paris, 1980), pp. 397 –414; N. Standen,
‘Frontiers of 10th-century north China’, in Frontiers in Question: Eurasian Borderlands, 700 –1700,
ed. D. Power and N. Standen (Basingstoke, 1999), pp. 55– 79; M. Rossabi, ‘Introduction’, in
China among Equals: the Middle Kingdom and its Neighbours, 10th–14th Centuries, ed. M. Rossabi
(Berkeley, Calif., 1983), pp. 1–13.
10
John of Salisbury, ii. 580; Dictionary of Medieval Latin from British Sources (Oxford, 1975–),
i. 1249; A. A. M. Duncan, The Kingship of the Scots, 842–1292: Succession and Independence
(Edinburgh, 2002), pp. 115, 260, 270 –1; F. Calasso, I glossatori e la teoria della sovranità (3rd edn.,
Milan, 1957), pp. 22 – 40; Folz, pp. 114–23; P. Munz, Frederick Barbarossa: a Study in Medieval
Politics (1969), p. 233; W. Ullmann, ‘The development of the medieval idea of sovereignty’,
Eng. Hist. Rev., lxiv (1949), 1–33; J. R. Strayer, Les gens de justice du Languedoc sous Philippe le
Bel (Toulouse, 1970), p. 44 n.; M. J. Rodriguez-Salgado, ‘Christians, civilized and Spanish:
multiple identities in 16th-century Spain’, Trans. Royal Hist. Soc., 6th ser., viii (1998), 233–53.
Historical Research, vol. 79, no. 204 (May 2006)
© Institute of Historical Research 2006.
156
Empires: a problem of comparative history
primarily for the Roman empire and the German empire, but sometimes
for any large or important and independent polity. The sense of territories
conquered overseas occurs occasionally in Spanish and English from the
seventeenth century, but did not apparently become common until latish
in the eighteenth.11 Meanwhile the title of emperor remained above all a
mark of status. An emperor did not necessarily have authority over kings
but he ranked above them.
As Europeans came into contact with more peoples elsewhere they
used either what they could make of the words, like khan or sultan, that
those people used for their rulers or such of their own words as seemed
suitable. While Marco Polo seems to have thought generally in terms of
kings and kingdoms and called the great khan the king of kings rather
than emperor, the thirteenth-century missionaries called him emperor.
Other rulers whom they met they called kings or used some version of
those rulers’ own titles.12 Later on emperor and empire came to be used
generally for any Asian rulers who seemed so rich and powerful and
maybe, above all, so dangerous that they deserved the grander title. 13
How the ruler of Japan came to be called an emperor is puzzling, at least
to me. In the seventeenth century European traders there gave the title
to the shogun and saw the other dignitary, whom the Japanese called the
tenno, as a sort of pope. In the nineteenth century they began to call him
the mikado. When the shogun was ousted at the Meiji restoration of
1868, they transferred the title of emperor to the mikado.14 Yet Japan
surely fitted none of the usual notions of an empire until the very late
nineteenth century.
In the eighteenth century, although European diplomats and politicians,
to judge from what one of them wrote, thought that ‘a pompous style
11
Oxford English Dictionary (2nd edn., 20 vols., Oxford, 1989), v. 187– 8; A. O. Meyer, ‘Der
britische Kaisertitel zur Zeit der Stuarts’, Quellen und Forschungen aus italienischen Archiven und
Bibliotheken, x (1907), 231– 7; Dictionnaire historique de la langue française, ed. A. Rey (2 vols.,
Paris, 1993), i. 682; Le Grand Robert de la langue française (6 vols., Paris, 2001), ii. 2054–5;
J. Fisch, ‘Imperialismus I–II’, in Geschichtliche Grundbegriffe, ed. O. Brunner and others (8 vols.,
Stuttgart, 1972–92), iii. 173–5; Diccionario de la lengua castellana (9 vols., Madrid, 1726 – 39), iv.
224 (facsimile repr. as Diccionario de autoridades (3 vols., Madrid, 1984)); J. H. Elliott, Spain and
its World 1500 –1700 (New Haven, Conn., 1989), pp. 7 – 91; J. Lynch, Spain 1516 – 98: from Nation
State to World Empire (Oxford, 1991), pp. 95– 6.
12
Marco Polo, Voyages (Paris, Société de Géographie, Recueils de Voyages, i, 1824), e.g.,
pp. 301, 304, 312, 347, 466; Sinica Franciscana: Itinera et Relationes Fratrum Minorum saeculi XIII
et XIV, ed. A. van den Wyngaert (7 vols., Florence, 1929), i. 29 – 30, 137 142, 188, 201;
Matthaei Parisiensis, Chronica Majora, ed. H. R. Luard (7 vols., Rolls ser., 1872 –84), iii. 488– 9.
13
K.-H. Ruffmann, ‘England und der russische Zaren- und Kaisertitel’, Jahrbücher für
Geschichte Osteuropas, new ser., iii (1955), 217 –24.
14
History of Japan compiled from the Records of the East India Company, comp. P. Pratt (2 vols.,
New York, 1972), i. 24, 32, 38, 350, 85; J. W. Hall, Japan from Prehistory to Modern Times (New
York, 1970), pp. 247, 302– 7; H. Bolitho, ‘Japanese kingship’, in Patterns of Kingship and
Authority in Traditional Asia, ed. I. Mabbett (1985), pp. 24–43, although M. Satow’s work
(referred to there, n. 45) does not seem to support Bolitho’s exact dates.
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Empires: a problem of comparative history
157
and . . . embellished letters’ pleased more barbarous nations,15 it was not
only barbarians who cared about titles, rank and status. After the Holy
Roman Emperor had lost virtually all his authority over the princes who
ruled different parts of Germany, the kings of England and France
resented more than before any suggestion that they were inferior to him. 16
Russia, too, was beginning to be a problem in international relations. When
the rulers of Muscovy began to call themselves tsars, a title obviously
derived through Byzantium from Rome, translating this as emperor
seemed harmless enough to western rulers and diplomats – so long as
Russia was seen as a more or less barbarous country outside the circle of
European diplomacy. When Russia entered that circle both the English
and the Austrian governments, for their different reasons, refused for
some time to allow the tsar to count as a proper emperor.17
The words emperor and empire seem to have become even more
detached from each other than they had been when medieval kingdoms
claimed to be empires because they were fully independent even though
their kings were not emperors. Rulers of the kind of overseas empires of
conquest to which the word empire was beginning to be applied in the
eighteenth century were not normally emperors. Emperor was a title that
no longer depended on size, independence or conquest, whether overseas
or not. If held by an established European ruler it was a mark of status
that was virtually immune to forfeiture. After Napoleon had complicated
things by becoming emperor over real, though short-lived, conquests and
had abolished the Holy Roman Empire, its emperor nevertheless
remained an emperor, as emperor of Austria. Both the Second Empire in
France and the German empire of 1870 got their titles, I imagine, not
because of conquests but because Napoleon III was the heir to his uncle’s
glory while the Kaiser, with evocations of the medieval empire, had to
be put above the other kings within Germany and on a level with the
emperor of Austria-Hungary.
All this probably impinges on the work of historians of modern empires
only in so far as it can be difficult to free oneself from assumptions
embedded in the established use of key words. Many historians are
reluctant to get bogged down in definitions, often because they are more
interested in one individual polity than in fitting it into a category: that
is entirely reasonable, but it can make their work hard to use
comparatively. Those who translate as empire and emperor the words
used in non-European sources for rulers and polities without further
explanation evidently assume that everyone knows what the words empire
15
Ruffmann, pp. 222–3.
H. Duchhardt, ‘Imperium und regna im Zeitalter Ludwigs XIV’, Historische Zeitschrift,
ccxxxii (1981), 555–81; P. Burke, The Fabrication of Louis XIV (1992), p. 72.
17
G. Hosking, Russia: People and Empire, 1552–1917 (1997), pp. 5–9, 39 –47; Duchhardt,
pp. 577 – 81.
16
Historical Research, vol. 79, no. 204 (May 2006)
© Institute of Historical Research 2006.
158
Empires: a problem of comparative history
and emperor mean.18 But it is hard to be sure that we do when both have
long been used in ways that make the boundaries of the categories they
imply so difficult to draw. Some use the Latin word imperium as if it
represented a more exact concept than ‘empire’, but it is not clear that it
does or why it should.19 As I hope I have shown, it has had many
different meanings. Others avoid the word empire altogether. Quite apart
from the conquest of the U.S.A. itself from sea to shining sea, which
would fit some criteria of empire, the American acquisition and rule of
the Philippines, for example, and of Hawaii and Alaska until they were
made states of the union, suggest that the U.S.A. had something that
should rank as an empire long before it became fashionable to talk about
one.
My argument is not that historians of empires need a tighter definition
of the word, let alone one that would exclude some of the polities
traditionally called empires that seem not to fit the current models. What
I suggest – if it is not impertinent for someone who does not herself do
any kind of imperial history to make any suggestion – is that looking
directly at the variety, not to say confusion, of words and concepts that
we have inherited may make it easier to lay both words and concepts
aside for the moment and concentrate on deciding what seem to be the
most important characteristics of the phenomena with which I think that
historians of empires, particularly more modern empires, are concerned.
If some empires seem to lack certain of those characteristics that is not a
reason for saying they are not empires, but it may help in deciding what
it is most profitable to compare with what. Irrespective of the definition
that anyone uses or implies it is surely necessary when one is making
comparisons to identify the characteristics that one wants to compare.
If one starts by thinking of what we are interested in as, by and large,
relatively large polities that consist of a ruling part (the metropolis) and
other parts (colonies or peripheries) that it dominates as a result of military
conquest or some kind of political or economic bullying, and that are
retained and governed separately from the metropolis rather than being
absorbed in it, then one has to note that not all the polities that look
likely cases seem to have all these characteristics. Size, to start with, has
to be relative, depending on the technology of communications and the
nature of economies: polities comprising areas that support only thinly
scattered populations look more impressive empires on a map than those
that cover smaller but more densely settled and intensively exploited
18
E.g., J. F. Fletcher, Studies on Chinese and Islamic Inner Asia (1995), ch. 7 (‘Turco-Mongolian
monarchic tradition in the Ottoman empire’); G. Rachewilz, ‘Qan, Qa’an and the seal of
Güyüg’, in Documenta Barbarorum: Festschrift für Walther Heissig, ed. K. Sagaster and M. Weiers
(Wiesbaden, 1983), pp. 272 – 81; I do not find clear the distinction between state and empire
in T. J. Barfield, Perilous Frontier: Nomadic Empires and China (Cambridge, Mass., 1989), pp. 5–27.
19
E.g., Barfield, p. 229; B. Porter, The Lion’s Share: a Short History of British Imperialism, 1850 –
2004 (4th edn., 2004), p. 8.
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Empires: a problem of comparative history
159
areas. The Athenian empire is often included in surveys of empires but it
was pretty small as well as short-lived. Some like to talk of Athenian
hegemony rather than empire but, following Finley, I do not think that
is very useful. Hegemony is ambiguous in modern use and the Athenians
really ruled the city-states that they took over, and ruled some of them
brutally.20
Conquest and division into metropolis and colonies are both as difficult
to assess as size. Some conquests were more violent than others, some
acquisitions, such as purchases, may have been quite peaceful, while the
longer an empire lasts the more the initial conquest may fade from memory.
But there seems to be a significant distinction to be drawn between
empires defined as the result of unilateral annexation of other polities and
either voluntary federations or the kind of ‘aristocratic empires’ that
Kautsky discussed, which seem more like what I would consider states –
aristocratic states maybe, although I would want to qualify some of his
ideas about government by traditional aristocracies. These seem to be
significantly different from empires of the kind that most historians are
concerned with, and not only because Kautsky’s aristocratic empires were
uncommercialized or pre-capitalist.21 The differences need to be thought
about.
The ancient Indian Mauryan empire fits the criteria of conquest, size
and division into metropolis and periphery, with some further distinctions
within the peripheral areas.22 The Roman empire fits the model of
conquest and probably also that of size, although it was less exceptionally
enormous than Eurocentric history tends to assume. As for metropolis
and colonies, Rome started with a fairly clear division between first the
city and the rest of Italy and then between Italy and the provinces, but
both divisions became less clear with time.23 China seems to qualify better
on the ground of size than of anything else: after the unification of what
became known as China in 221 B.C.E., when one ruler took over what
had been separate kingdoms, it became at first more like a single, if
enormous, state, although some subordinate rulers were allowed an initial
measure of independence.24 This first unification of China thus looks,
except in size, rather like that of the little medieval kingdom of England,
which was formed out of the conquest and absorption of a number of
mini-kingdoms. The larger and later kingdom of Spain was formed in
20
Finley, ‘The 5th-century Athenian empire’; ‘Introduction’, in Garnsey and Whittaker,
p. 1. D. Lal, In Praise of Empires: Globalization and Order (Basingstoke, 2004), p. 4, citing M. W.
Doyle, Empires (Ithaca, N.Y., 1986), thinks Thucydides distinguished between Athens’s empire
and Sparta’s hegemony but Doyle, pp. 12, 40, 81 makes it clear that the distinction is modern.
21
J. H. Kautsky, The Politics of Aristocratic Empires (Chapel Hill, N.C., 1982), pp. 1–31, 62 – 7.
22
R. Thapar, Early India: from the Origins to A.D. 1300 (2002), pp. 174– 9, 194– 7.
23
P. Garnsey and R. Saller, The Roman Empire: Economy, Society and Culture (1987), p. 9.
24
Cambridge History of China, ed. D. Twitchett and others (Cambridge, 1978–), i. 20, 53–4,
123, 429, 470 –2.
Historical Research, vol. 79, no. 204 (May 2006)
© Institute of Historical Research 2006.
160
Empires: a problem of comparative history
rather the same way, but there some of the constituent kingdoms kept
their separate identities and some of their institutions for centuries, while
being effectively governed from Castile. Peninsular Spain might thus itself
count as an empire quite apart from its overseas dominions.25 But China
did not apparently develop like that. Different dynasties during the long
centuries of what is called the Chinese empire had different centres of
power, so that, although historians refer to ‘metropolitan areas’ around
successive capitals that were differently governed from other provinces, it
is difficult (as in what became known as the Holy Roman Empire) to
identify a single and permanent metropolis in the sense that the word is
used by theorists of empires. China certainly had an outer periphery in
central Asia, where its rulers intermittently controlled, or tried to control,
areas beyond what they ruled more regularly, but a consistent distinction
between metropolis and peripheries within China is hard to find.26 Russia’s
conquests qualified it by size even if the date when it began to count as
an empire is a matter of opinion, but again the metropolis/colony division
is unclear. On the other hand, different as Russia’s nineteenth-century
land-based empire may look from the overseas colonies of other European
powers of the time, it is less exceptional if one takes a longer view. 27
Leaving aside cases of the absorption of neighbouring areas into the
kingdoms of medieval England or France, or into the U.S.A., the
distinction between metropolis and colonies also does not seem to fit, for
instance, all stages of the early Islamic empire or the Ottoman empire. I
do not, incidentally, take seriously the claim that in the last years of
French rule Algeria was part of metropolitan France.28
Variations in size, in methods of acquisition and in methods of
government clearly offer scope for comparisons that raise questions about
individual cases and the categories into which we put them. I propose to
look briefly at methods of government, which, I suggest, need to be seen
in the context of the norms and assumptions of those who made
conquests or acquisitions. Empires have been acquired not only by
monarchies but by what are called republics or even democracies –
although it is surely slightly misleading to go on calling Athens a
democracy, given any of the senses that the word has in current English.
Leaving that aside, there is the fairly obvious point that, however
metropolitan states rule their colonies, they do not themselves need to be
25
Rodriguez-Salgado; Lynch, pp. 1–45, 96 –100, 211–49.
Cambridge History of China, i. 429, 505; iii. 160, 293; vii. 10 –11. Barfield, pp. 2–27; J. A.
Millward and P. C. Perdue, ‘Political and cultural history’, in Xinjiang: China’s Muslim
Borderland, ed. F. S. Starr (New York, 2004), pp. 27 – 62.
27
Hosking, pp. 39 –40; D. Lieven, ‘The Russian empire and the Soviet Union as imperial
polities’, Jour. Contemp. Hist., xxx (1995), 607 –36; and D. Lieven, The Russian Empire and its
Rivals (2000).
28
I. S. Lustick, Unsettled States, Disputed Lands: Britain and Ireland, France and Algeria, Israel
and the West Bank-Gaza (Ithaca, N.Y., 1993), pp. 89 –112.
26
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Empires: a problem of comparative history
161
ruled autocratically, let alone by someone called an emperor, to count
usefully as empires.
One point that I also want to make may be less obvious to some
modernists. What I know about the European middle ages, and what I
have picked up through patchy reading on traditional societies outside
Europe, suggests that many, however arbitrary, unjust and oppressive
their governance seems to us, appear to have been able to rely on a good
deal of voluntary submission. Rulers ruled, and were supposed to rule,
by and with the advice and consent of their great men, while the
customary law that was practised in many polities, even if declared and
enforced by the ruler, was supposed to be made and applied by some kind
of collective judgements. Kingdoms and other units of government
were often regarded as natural communities of common descent, law and
culture.29 When neighbouring territories that they conquered had similar
cultures and customs that had not yet been fossilized by written records,
they may thus have been relatively easy to absorb and integrate into the
cultures and solidarities of the conquerors. The distinction between
metropolis and periphery could therefore be lost – as it was in some
medieval European kingdoms, and apparently in early China. How far
absorption was helped or hampered by allowing acquired territories to
keep separate institutions and bits of their own law seems to have varied.
Both in Spain and in central and eastern Europe, for instance, the special
privileges of particular acquired areas tended to become eroded with
time, but separate names and collective identities could still be preserved
in these quasi-imperial situations.
It might be reasonable to guess that what preserved conquests as empires
rather than turning them into new provinces within the conquering state
may have been the practicalities of distance and communication. Or was
it that they started with such fundamental differences of culture, customs
of government and physical appearance of the inhabitants as to make
absorption impossible or difficult? Or that, as in the case of some of the
nomadic empires of Asia, all that the new ruler wanted to do anyway was
to take tribute from conquered territories? Or that violent conquest and
resistance left a legacy of hostility and violence? Or a combination of any
of these?
Distance and difference of culture and solidarities may go together but
need not: they did not, for instance, in the case of English conquests in
Wales and Ireland. In the example of Wales, in particular, distances were
minimal even in medieval terms. Although they fairly obviously went
together in European conquests overseas, cultural differences may have
been exaggerated by conquerors who could not or would not look
carefully at societies that they wanted to conquer and profit from anyway.
29
S. Reynolds, ‘The idea of the nation as a political community’, in Scales and Zimmer,
pp. 54– 66, and works cited there.
Historical Research, vol. 79, no. 204 (May 2006)
© Institute of Historical Research 2006.
162
Empires: a problem of comparative history
In Asia it was easier to ignore existing laws and rights of property if one
classed a society as an oriental despotism. Societies in north America, on
the other hand, could be seen as in a state of nature, with no property
rights at all. In practice a good many conquests (how many?) started with
mere tribute-taking and sometimes used existing systems and officials,
even those of rulers condemned as despots. However strange other aspects
of a newly conquered society may be, conquerors seem to find out about
their taxes quite easily. But, easy as it may be to take tribute once or
twice, one tribute does not make an empire. Getting payments and dues
regularly, and in amounts to satisfy growing demands, may then lead to
rebellion, fiercer enforcement and more formal and oppressive government.
How does this affect the metropolitan government itself and its norms?
Although oppression can be just as fierce in states without empires, I
wonder how much and in which empires repression abroad has affected
ideas about the rights of subjects or citizens at home. Whether they are
called subjects or citizens is, incidentally, I suggest, more a matter of
words than of substantive rights, although the different words may suggest
new ideas to those concerned. How much does empire lead either rulers
or their metropolitan subjects to think new thoughts about rights and law
in general and in their own country?
Comparison suggests that variations in the ways empires have been
ruled are endless and raises a lot of questions about possible reasons for the
variations. It may seem obvious that empires should have been nurseries
of bureaucracy, needing tax-lists and systems of written communication
between centre and periphery, if no more. Leaving aside the more fleeting
nomadic empires, the tendency seems to have been fairly general. China,
of course, although in some other ways non-imperial, was famously
bureaucratic, while the way that the Roman empire, although it started
with a very limited bureaucracy and delegation of a good deal of everyday
government to local elites, became more bureaucratic with time, may
illustrate the connection well enough.30 On the other hand, while empires
may tend to need bureaucracy, bureaucracy does not need empires. In
post-Roman western Europe it developed most notably in Italian citystates, the small kingdom of England and the papacy, which ruled neither
a state nor an empire.
Even the most centralized and bureaucratic governments, moreover,
whether of empires or any but the very smallest states, need to delegate
authority. They needed to do so especially when lines of communication
were long and travel was slow. Indirect rule through trading companies
was one way of managing far-flung empires; giving local governors some
freedom in making decisions was another; while co-opting existing local
elites and leaving local government to them might produce a system of
30
Garnsey and Saller, pp. 20 – 40; C. M. Kelly, ‘Later Roman bureaucracy’, in Literacy and
Power in the Ancient World, ed. A. K. Bowman and G. Woolf (Cambridge, 1994), pp. 161– 76.
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Empires: a problem of comparative history
163
government rather like that of the metropolis. Leaving local elites in
charge presumably worked more easily insofar as it maintained traditional
units of government, as in the Roman empire, so that the bulk of the
population might be relatively unaffected by the conquest until demands of
government became so heavy as to provoke revolt, or, in modern empires,
until ideas of new kinds of popular government or democracy began to
percolate down. I refer here to ideas of democracy rather than nationalism
because the ideas of solidarity and political community that I have argued
were attached to units of government long before the eighteenth century had,
I maintain, a good deal in common with what is now called nationalism. 31
While they served to support and validate existing states, they could also
console the elites of former states that preserved some of their traditional
institutions within larger states or empires. It was the combination of
these old solidarities with new ideas of the rights of man that created the
kind of modern nationalism that has frustrated modern empires.
Settler colonies seem, at least at first, to have been easier to keep on a
light rein. Some, indeed, started as completely independent, like Iceland
and Greenland – nice exceptions to the trend from empire to
independence – or like the medieval German colonies in kingdoms to the
east, which had no formal attachment to the kingdom of Germany but
received privileges from their local kings.32 Some early modern emigrant
colonies may have been kept under metropolitan government as much to
keep out rivals as because they were initially seen as particularly valuable,
except as places to send poor and landless troublemakers. Assuming that
emigrants from the home country could be left a fair degree of autonomy
because they shared the values and laws of the metropolis could work
well – as the white British commonwealth in the First World War might
suggest if it were not that Britain got even greater service in the war from
India. It could, however, turn out to be a mistake, either because settlers
used their old values to question the taxes and controls imposed on them
or because their own demands on native populations or imported slaves
provoked more resistance than they could handle on their own. Settler
colonies were, after all, never populated only by free settlers: slaves,
indentured servants and convicts in penal colonies all tended to complicate
relations between metropolis and periphery. Nor were any colonies
without settlers of some kind, even if only temporary ones, who would
or could go home or go elsewhere. There were generally governors,
garrisons and administrators, some of whom became too corrupt or
tyrannical, while others, in the British phrase, ‘went native’. Did which
of these happened depend on distance, geographical or cultural, on the
amount of central control, or on changes in metropolitan policies? Where
31
Reynolds, ‘The idea of the nation’.
These would not be colonies at all by Finley’s classification (M. I. Finley, ‘Colonies – an
attempt at a typology’, Trans. Royal Hist. Soc., 5th ser., xxvi (1976), 167 – 88).
32
Historical Research, vol. 79, no. 204 (May 2006)
© Institute of Historical Research 2006.
164
Empires: a problem of comparative history
indigenous populations seemed to the empire-builders entirely what we
now call ‘Other’ one would expect them to be ruled more unsympathetically
and oppressively, but that may be too simple. Otherness has been measured
so differently at different periods, by cultural or physical differences or by
class: native rulers and elites might seem less ‘Other’ than peasants, let
alone slaves, but some who were less ‘Other’ to their British rulers, like
middle-class Indians who received a western education, could arouse
more fear and hostility from their rulers than peasants who could more
easily be seen as the salt of the earth and gratefully devoted to Us.
How consistently the reasons for acquiring any empire affected the way
it was then governed seems doubtful. Modern historians tend to stress
economic motives, but in the long ages before governments had economic
policies a ruler’s conquests, although perhaps starting with raids for slaves
and loot (which presumably count as economic), were surely often
motivated more by a drive to defend and extend his power and prestige.
Economic purposes become most obvious with the European search for
spices in the east and gold and silver in the west, but in these adventures
more or less independent entrepreneurs could be as important as the ruler
who ended up with an empire. Then there were missions, some undertaken
with complete sincerity, others not, whether they were intended to spread
a religion, law to lesser breeds without it, or the benefits of commerce
and taxation. Whatever their purpose and sincerity, missionaries had
varying and changing relationships with imperial governments and
varying impacts on the way that they governed.
Since the tendency to self-justification seems common in human beings,
and rulers like to be loved and to be considered legitimate, it is not
surprising that many empires produced ideologies of benevolent imperialism,
but it might be rash to assume that these ideologies were very different
from those of benevolent governance in non-imperial states. The degree
to which a metropolitan society’s own ideology was changed by having
an empire, or whether the two were kept separate, and if so how, seems
to me worth investigating – for all periods and empires. All these
questions are particularly difficult for periods in which one has to deduce
reasons from results, and possible motives from what look like possible
reasons. Even with better sources, the numbers of people with varying
motives and interests involved in conquering, governing, settling and
trading in any empire, combined with the force of muddle in human
affairs, make the problem of disentangling causes and effects almost
insoluble – but still worth thinking about? How far did any empire (apart
from the much discussed Spanish and British empires) increase the power
or wealth of its metropolis, and of what sections of its population? The
correlation of relative oppression with official policy, with degrees of
discontent or with the fall of the empire is liable to be complex.
I cannot, however, go into the reasons for the fall of empires,
except to say that ‘the natural and inevitable effect of immoderate
© Institute of Historical Research 2006.
Historical Research, vol. 79, no. 204 (May 2006)
Empires: a problem of comparative history
165
greatness’,33 like a good many other less lapidary judgments on the Roman
empire, looks even less convincing when one compares it with nonEuropean empires. Nor can I go into the meaning of imperialism or the grey
area of quasi-empires – whether labelled as hegemonic, invisible, informal,
soft power, or simply put into quotation marks as ‘empires’. Some of the
labels are helpful in reminding us that degrees of effective control correlate
poorly with formal connections, but some of them have been chosen for
particular cases in a way that impedes comparison. Seeing British commercial
activity in south America in terms of informal empire was illuminating about
nineteenth-century British history and the growth of the British empire
but perhaps looks less easy to interpret in terms of empire in general
when United States involvement alongside is brought into the picture. 34
Anyone who works on any of the empires that I have mentioned in this
rush through time and space will realize how little I know about their
own field. I hope that my mistakes and misunderstandings will not enable
them to discount entirely my plea for more comparisons. The subject
cries out for more of them, including, I submit, comparisons between
periods, which some historians do not approve of – although disapproval
does not always inhibit carefree allusions to the Roman or Chinese
empires. Even for those who are only interested in one case or one
period, comparison with others may surely be illuminating in raising
questions about assumptions and chains of cause and effect. The real
argument against comparison is that it is such hard work, especially when
it goes over different periods and demands a range of different skills, and
especially if it involves more than looking outside one’s own field to pick
something out of its context elsewhere because it looks like something in
one’s own. Even at the most superficial level there is another
impediment. Historians working in different periods and in different
national traditions of historiography, even if they do not have conscious
ideological differences, tend to work from different presuppositions and
within different paradigms. They therefore focus on different subjects and
use different terminologies (or the same terminology in different senses),
so that it can be hard, especially if one cannot read their sources, to be
sure that one is comparing like with like. I have found this when
comparing just a few aspects of a few kingdoms in medieval Europe.
Empires, however defined, pose much greater problems, even if one
ducks (as I have) the question of whether they were or are a Good Thing
or a Bad Thing. But grappling with these problems is a matter for the
historians of empires, not for a single lecture by a non-imperial historian.
33
E. Gibbon, Decline and Fall of the Roman Empire, ed. D. Womersley (5 vols., 1994), ii. 509
(ch. 38, ‘General observations’).
34
J. Gallagher and R. Robinson, ‘The imperialism of free trade’, Econ. Hist. Rev., 2nd ser.,
vi (1953), 1–15; C. M. Dobbs, ‘The Monroe Doctrine’, in The Dictionary of American History,
ed. S. I. Kutler (3rd edn., 10 vols., New York, 2003), v. 446– 7.
Historical Research, vol. 79, no. 204 (May 2006)
© Institute of Historical Research 2006.
