Wood EM - Landlords and Peasants, Masters and Slaves - Class Relations in Greek and Roman A

Ellen Meiksins Wood
Landlords and Peasants, Masters and Slaves:

Class Relations in Greek and Roman Antiquity 1
Few historians would hesitate to single out slavery

as an essential and distinctive feature of the social
order in ancient Greece and Rome. Many might
even be prepared to accept that slavery is in some
sense the distinctive and essential characteristic,
so that Greece and Rome can be meaningfully and
informatively described as ‘slave societies’, ‘slave
economies’, or instances of the ‘slave mode of pro-
duction’. There is little agreement, however, about
what precisely it means so to characterise them and
how – or even whether – such characterisations are
intended to shed light on historical processes.
At least two separate questions must be raised in
assessing the usefulness of describing Greece and
Rome as slave societies. The rst concerns the extent
and location of slavery in the Greek and Roman
1
This article, written in 1983, has been sitting in my drawer ever since, no doubt
in the hope that I would get back to it. Some of the arguments concerning ancient
Greece were later developed in Wood 1988. But, since I am now unlikely to return to
systematic work on the ancient world, I have agreed to let this piece be published as
is, even though I would do some things differently if I were writing it now. The text
here remains as it was when I left it in 1983, with the exception of footnotes which
refer to publications that have appeared since then, a few places where a misleading
word has been replaced, and the deletion of one especially confusing passage.
Historical Materialism, volume 10:3 (17–69)

© Koninklijke Brill NV, Leiden, 2002
Also available online – www.brill.nl
18 Ellen Meiksins Wood
economies, and, in particular, the degree to which production rested on

slave labour. The second concerns the sense in which slavery, whatever its
location and extent, can be said to account for historical movement, economic
development, social change, political and cultural processes. One way of
formulating the latter question might be to ask whether slavery should be
looked upon as a kind of prime mover, the ‘distinctive characteristic’ in
reference to which other features of Greco-Roman antiquity can be explained;
or whether it is more fruitful to treat Greek and Roman slavery as effect rather
than cause, the ‘distinctive characteristic’ which itself most demands expla-
nation, by reference to other, prior, historical factors.
In what follows, it will be argued that the focus on slavery as the primary
or dominant characteristic of Greco-Roman antiquity may obscure more
than it explains, and that historical processes in ancient Greece and Rome
would be more clearly illuminated by the proposition that their ‘distinctive
feature’ is a particular and unique relationship between free producers and
appropriators and, more specically, between landlords and peasants. This
is true only partly because the extent of slave production, especially in Greece,
is itself open to question. Even in Rome, where the evidence for widespread
slavery in basic production is far less ambiguous (though, even here, there
has been some exaggeration), the relations between landlord and peasant are
the framework within which historical processes – including the rise and
decline of slavery – can most fruitfully be understood.
Master and slave vs. landlord and peasant

The proposition that Greece (meaning, for the most part, specically Athens)
and Rome were ‘slave societies’, ‘slave economies’ or instances of the ‘slave
mode of production’ is problematic, rst, because there is little agreement
not only about the explanatory value of these formulae but even about
what facts they are meant to represent. These designations were relatively
unproblematic for those who believed – as many Marxists, but not only
Marxists, have traditionally done – that the bulk of production in Greece and
Rome was performed by slaves and that class divisions were reasonably clear
and transparent, opposing a legally dened community of free men – and
especially citizens – to a subjected producing class of slaves. Since it is now
more commonly accepted that production throughout Greek and Roman
Landlords and Peasants, Masters and Slaves 19
history rested at least as much on free labour as on slavery, and since this
inevitably raises questions about class divisions within the free population,
the role of slavery as the key to ancient history has become a rather more
thorny question. Many commentators still nd it useful to speak of ancient
Greece and Rome as slave societies; but the formula alone no longer tells
us what claims its users want to make about Greek and Roman economic
and social organisation, about the prevailing forms of production and appro-
priation, or about the dominant social relations and conicts.
In the case of Athens, there are at least two senses in which the adequacy
of the formula ‘slave society’ or ‘slave mode of production’ must be ques-
tioned. First, there is considerable ambiguity in the evidence concerning the
degree to which slaves were engaged in production (particularly as distinct
from domestic service), especially in agricultural production, which was
still the basis of the Greek economy, and hence also the degree to which they
produced the surplus that created and sustained the propertied classes. Second,
it is far from clear how the formula helps to account for historical movement,
social change, political development, or the production of ideology – processes
in which the relations between propertied classes and free producers were
at all times central.
The role of slavery in Athenian agriculture has been a matter of hot
dispute. Historians have generally tended toward the view that slavery in
classical Athens was far less important in agriculture than in industry and
trade, which is clearly a fact of critical signicance in this predominantly
agrarian economy. Recently, this view has been challenged by some historians
who claim that agricultural slavery was far more widespread and important
than is generally believed; but their arguments have been far from convincing.2
Given what is known about patterns of landholding, the relatively limited
concentrations of property and disparities of wealth, the austerity of the
material culture, and the simplicity of the state apparatus, it seems unlikely
that there was a great deal of scope for labour on the land beyond the peas-
ant family unit. Landholdings were generally modest, and even wealthy
2
There is no room in this article to do more than assert this boldly and without
supporting argument. A detailed elaboration of this assertion can be found in Wood
1983 where I canvass the evidence and critically examine the two most important
recent arguments in favour of widespread agricultural slavery in classical Athens:
Jameson 1977 and Ste. Croix 1981, especially in Appendix II.
landlords tended to own several scattered smaller properties rather than large
estates. There were no plantations like the Roman latifundia. The bulk of the
proprietors were smallholders who worked their own land, while wealthier
landlords probably ‘farmed out’ their several smallholdings to tenants
and sharecroppers. On large estates, there was undoubtedly a permanent
but probably not very large stock of farm slaves. It is known that casual
wage-labour was widely used at the harvest, and it seems likely that it was
available at all times in the form of propertyless citizens and smallholders
whose own lands (or tenancies) were insufcient to support their families.
Nevertheless, if the contribution of slaves to production has in the past
been exaggerated, and if the degree to which free men not only produced –
both in agriculture and in craft-production – but were subject to exploitation
has been underestimated, it remains true that chattel slavery was much
more widespread and signicant in Greece and Rome than anywhere else
in the ancient world, sufciently so that it can reasonably be considered
a ‘distinctive characteristic’ of these societies. It also remains true that free
producers, especially in democratic Greece, were free from exploitation to an
unusual degree, so that this, too, must be regarded as a distinctive feature of
that society. The question then might be: what is the relationship between
the two primary sets of social relations – slaveowners/slaves, and propertied
classes/free producers or, more specically, landlords/peasants? And, if either
of these should be accorded explanatory priority, which is it to be?
The peasants of Attica differed from their counterparts elsewhere (with the
partial exception of Rome) above all in one respect: their citizenship. In no
other known ancient society – and in no society since to the same extent –
were peasants full members of the political community. The peasant-citizen
was certainly a ‘distinctive characteristic’ of Athenian society (and Roman,
up to a point). Among the many implications of this unique social formation
is that Attic peasants, as well as small urban producers and tradesmen,
enjoyed an unusual degree of freedom from various forms of exploitation
by landlord and state. The civic status of the small producer limited the
two principal forms of surplus extraction to which peasants, in particular,
have historically been subject: rent and tax. The democracy – or rather,
the conguration of social and political power that it represented – restricted
the wealth and power of landlords by limiting the opportunities for con-
centrating property and by protecting small producers from various forms
of personal dependence – slavery, serfdom, debt-bondage. Thus, both the

form and the extent of rent-extraction were severely restricted. Furthermore,
as M.I. Finley has suggested, exemption from regular taxation was a hallmark
of ‘that novel and rarely repeated phenomenon of classical antiquity, the
incorporation of the peasant as a full member of the political community.’3
Taken together with the relatively simple state apparatus that characterised
the polis, this meant a freedom from state exploitation which distinguishes
Athens (and early republican Rome) radically from other societies where
tax-burdened peasants have borne the weight of kingdoms and empires.
It is not, however, only the free producers themselves who were affected
by their unique civic status. The whole system of production and surplus
extraction, including the emergence and development of slavery, must be
understood against the background of the peasant-citizen. It has often been
said that slavery made Athenian democracy possible. This proposition has
often been associated with the profoundly mistaken view that Athenian
citizens were generally free from the necessity to labour and that production
rested essentially on slave labour. Since it is now a generally accepted fact
that the majority of citizens worked for a livelihood – in agriculture, crafts,
or trade – and since the contribution of slaves to production, especially in
agriculture, is, to say the least, open to question, the association of democ-
racy with slavery must be formulated somewhat differently.
One possibility is to say (as Michael Jameson does, for example 4) that slaves
supplemented the labour of the ordinary farmer in a way that made possible
his performance of civic and military functions. Even this formulation, while
it assumes that agricultural slavery was essential and widespread, implies
that the growth of slavery presupposed the emergence of the farmer-citizen.
The argument is not simply that because slaves somehow became available
in large numbers, Attic farmers found themselves able to become citizens.
On the contrary, the object of an argument like Jameson’s is precisely to
explain what was unique about ancient Athens which compelled farmers
to seek an unusual means of responding to increasing demands on their
productive capacities. In other words, according to this view, it is because
3
Finley 1973, p. 96.
4
This is, in fact, the essence of his article cited in note 2 above.
the Athenian farmer had attained the apparently unprecedented status of full
citizenship that he was obliged to seek the assistance of slaves.
This formulation, though it places slavery and the farmer-citizen in the
correct order, will not quite do. Since, as we have seen, the civic status of
the farmer actually restricted the need for surplus production by limiting the
pressures of surplus extraction in the form of rent and taxes, the citizenship
of the peasant could just as easily be regarded as a limitation on his need for
slave assistance.5 What can, however, be said with some assurance is that
the status of the peasant citizen and his freedom from dependence created
an incentive for wealthier landlords to seek alternative sources of labour;
and it is almost certainly true that slavery grew as Athenian smallholders
themselves became unavailable as dependent labourers.
Even here, caution is needed. Not only must we keep in mind that
smallholders remained available to their wealthy compatriots as tenants,
sharecroppers, and casual wage-labourers, and also that there were property-
less citizens who required employment; but we must also consider the extent
to which the relations between landlords and peasants restricted the form
and extent of slave-utilisation itself. In particular, the conguration of class
power within the citizen body, to the extent that it curtailed concentration
of property, also limited the possibilities of slave exploitation. As long as
properties remained small and peasant tenures relatively secure – and even
wealthier landlords tended to own several scattered smaller holdings – the
scope for the utilisation of labour in production beyond the peasant family
was limited. In the forms of exploitation more appropriate to smallholdings
and a free peasantry – tenancy, sharecropping, casual wage-labour – family
labour would still have been the predominant productive force. The growth
of the urban economy, craft-production, and trade expanded the scope of
slave exploitation (though the extent of these developments should not be
exaggerated, since production for the market remained undeveloped); and it
is worth noting that the very few known large slave-enterprises in Athens,
in addition to the mines, were ‘industrial’ rather than agricultural. In sharp
contrast to Rome, the intensive exploitation characteristic of latifundia worked
5
In Wood 1983, I elaborate this argument in reply to Jameson’s contention that the
utilisation of slave labour was probably the norm on ordinary small farms, because
it was the best way to intensify labour without sacricing the farmer’s civic status.
by slave-gangs, made possible by enormous accumulations of property, never

existed in classical Greece. These signicant differences between Greek and
Roman slavery reect different congurations of power and property which
characterised relations between peasants and landlords in the two cases.
The signicant point, then, is that however many complexities and
qualications are introduced into the relation between slavery and that ‘rare
social phenomenon’, the peasant-citizen, we are always brought back to the
latter as historically prior. While there can be no simple equation between
chronological and causal precedence, it would be difcult to sustain an
argument that explained the emergence in Greece of a special kind of
small producer – and a unique set of relations between property, class, and
state – by referring to an increase (for whatever reason) in the availability
of slaves. The slave factor is neither sufcient nor necessary to explain this
development. In contrast, while the relative unavailability of free producers
for certain kinds of exploitation may not be sufcient to explain the parti-
cular development of slavery in Greece, it is certainly a necessary factor in
accounting for the nature and extent of that development.
It is not only slavery itself that must be explained within the more
inclusive framework of class relations within the citizen body. The latter are
also more inclusive in their ability to explain other essential characteristics
of Athenian society, notably its distinctive political and ideological features.
In particular, Athenian democracy, that most remarkable phenomenon of
classical antiquity, cannot be explained (as Marx, for example, appears to
suggest) as the means by which a community of proprietors organised itself
militarily and politically against the outside world and especially against
an alien subjected producing class of slaves; instead, it is more fruitfully
understood as a means of organising the class struggle between lords and
peasants. The democracy, whose foundations were laid before large-scale
slavery became a signicant factor in Athenian social life, grew out of class
conict between small proprietors and the landlords who had subjected many
peasants to dependence, especially in the form of debt-bondage.6 The nature
6
These arguments are developed in Wood and Wood 1978, Chapter 2. While
there is much that I should like to improve in this account, the basic outline of the
class conict between producers and appropriators and how it contributed to the
development and shape of the democracy still seems to me essentially correct. It
is interesting to note that a similar argument is made in Part II of Ste. Croix 1981.
of that class conict within the citizen-body changed as slavery grew and
as the number of citizens engaged in ‘urban’ production as craftsmen and
artisans increased; but the class conict among citizens never ceased to play
an essential role in Athenian politics. Although political divisions between
democratic and oligarchic factions never coincided neatly with class
divisions, these political conicts cannot be understood without reference
to the antagonisms between those for whom the democracy represented a
protection from various forms of surplus extraction and those for whom it
represented a limitation on their powers of appropriation.
These class divisions within the citizen body are equally visible in the great
cultural products of Athenian society. The political philosophy of Plato, for
example – indeed much of his non-political philosophy, even his conceptions
of knowledge and virtue – is unintelligible if abstracted from the social
division between those who laboured for a livelihood and those who lived
on the labour of others. His contempt for labour and those who are bound
to the world of material necessity by the need to work for a living, the
principle that such material bondage is morally and politically debilitating
and renders true knowledge (which is virtue) impossible – a principle that
lies at the very heart of his whole philosophical project – has often been treated
(especially by Marxists) as a reection of the ‘slave mode of production’;
but it is demonstrably an aristocratic reaction to the realities of Athenian
democracy and the unprecedented power, protection, and freedom it afforded
to the free producing classes, the ordinary peasants, craftsmen, and trades-
men who constituted the bulk of the Athenian citizenry.7 For Plato, as for
many aristocrats in non-slaveowning societies as well as ‘slave economies’,
all labour is in essence servile; and Athenian democracy violated the essen-
tial principle that political life should be the exclusive preserve of people
freed from the realm of material necessity by the labour of others.
Clearly, any attempt to explain the dynamics of Athenian history must, at
the very least, take into account the complex interactions between the two
primary sets of social relations, those between slaveowners and slaves and
Despite his emphasis on slavery in the conceptual arguments of Part I, his historical
account of Athenian society focuses almost exclusively on the class struggle within
the citizen body, with virtually no reference to slavery. More on this later. [A later
elaboration of my argument occurs in Wood 1988.]
7
See Wood and Wood 1978, Chapter 4.
those between propertied classes and free producers. If primacy must be

given to one set of relations over the other, however, the case for the latter
seems more convincing, because the development of slavery itself took
place within the framework of the relation between propertied classes and
free producers, and because the latter accounts for much that is distinctive
in Athenian social, political, and ideological formations.
A similar case can be made for Rome. The fact that here – or at least in
Roman Italy – slaves were far more prominently employed in agricultural
production than in Greece does not alter the primacy of the relations between
landlords and peasants in Roman history. It is true that slavery in agricul-
ture was extensive, and that slave plantations – which never existed in Athens –
gured prominently in the Roman economy in certain periods and places.
Nevertheless, even in Rome, the large-scale use of slavery, and especially
slave latifundia, developed late and remained localised, concentrated largely
in the West and in Italy and Sicily in particular. Peasant production survived
even here and always predominated in other parts of the Empire, where other
forms of surplus extraction prevailed, such as various forms of tenancy and
sharecropping. Slavery in agriculture was never important, for example, in
North Africa or in the East. It has been estimated that, even in Roman Italy,
when slavery was at its height, the majority of the population outside the
city of Rome were probably still peasants.8
Even granting the extraordinary prominence of slavery in the Roman
economy, that very prominence demands explanation. The historical rarity
of ‘slave societies’ alone should turn our attention to the peculiar circum-
stances that made them possible and/or necessary;9 but the dramatic growth
of slavery in Roman Italy is especially remarkable for reasons that go
beyond the numerical scarcity of ‘slave societies’ throughout history. As Keith
Hopkins has pointed out, there is a signicant difference between Roman
(and Greek) slavery and the slavery of the other, more recent slave societies
in the Americas. In the latter, slaves did not displace local labour but existed
8
Hopkins 1978, p. 7.
9
‘Slave societies’ – societies with ‘an institutionalised system of large-scale
employment of slave labour in both the countryside and the city’ – have been rare
throughout history, as pointed out, for example, in Finley 1980, p. 67. Keith Hopkins
also emphasises the rarity of slave societies, suggesting that there have been only a
handful of cases: classical Athens, Roman Italy, the West Indian islands, Brazil, and
the southern states of the USA. Hopkins 1978, pp. 99–100.
alongside a local labour-force inadequate to the demands of the ‘incipient

industrial revolution’ (as Hopkins puts it) or, more precisely, to the growing
needs of capitalism. By contrast, Hopkins explains, ‘In Roman Italy (and to
a much smaller extent in classical Athens), slaves were recruited to cultivate
land already being cultivated by citizen peasants.’10 In the early Republic,
before imperial expansion, the relations between landlord and peasant
had probably been roughly similar to our description of classical Attica
with respect to the size of landholdings, the disparities of wealth, and
the prevailing forms of surplus extraction. Peasants produced surplus for
landlords as tenants, sharecroppers, casual labourers; and there was a limited
scope for labour beyond the peasant family unit. By the end of the Republican
era, Rome was characterised by huge concentrations and disparities of wealth
and a growing imperial apparatus. The period in between was also marked
by the tremendous growth of slavery and the partial displacement of peas-
ants by slaves in part of the empire.
The question of slavery, then, is rst and above all the question of the
peasantry. How, why, and under what circumstances were relations between
landlords and peasants transformed in such a way as to displace peasants
on a signicant scale as agricultural producers and to permit or necessitate
a massive intrusion of slaves into the Roman economy? The question becomes
even more pressing if (as Hopkins convincingly argues) slavery was a difcult
and costly mode of exploitation and ‘by no means an obvious solution to the
élite’s needs for agricultural labour’.11
If the history of slavery in Rome is the history of peasants, it is also the
history of peasant expropriation and struggles over land. Some have argued
(e.g. Pierre Dockès)12 that the massive inux of slaves provided the Roman
‘élite’ with an incentive, or even a need, to squeeze out small proprietors in
order to concentrate land and take full advantage of the opportunities afforded
by slavery. No doubt it is true that, once set in train, the intrusion of slaves
created pressures for expropriation; but this argument evades the essential
questions. How and why did slavery on such a scale become an option in
the rst place? If slavery was the result of conquest and empire, how and
10
Hopkins 1978, p. 7.
11
Hopkins 1978, pp. 9, 108 ff.
12
See below, pp. 54–5.
why did Roman imperial expansion become possible and necessary, and why
did the Romans transform captives into slaves when this disposition of
conquered peoples was far from an obvious or universal practice? It can
hardly be maintained that the inux of slaves provided the original impetus
for expropriation, since the very motivation for large-scale slavery presupposes
a signicant degree of land concentration; therefore, where did this impetus
come from and how did it become possible to expropriate peasants on such
a scale that vast numbers of slaves could be employed? One might even
ask why it became necessary for landlords to adopt a form of exploitation
that required concentration of property to make it economically feasible. At
any rate, what made the advantages of this burdensome form of exploitation
outweigh its disadvantages?
Whatever the answers to these questions, the important thing is that they
are there to be asked. In other words, slavery cannot be taken for granted.
And whatever the answers, the questions will inevitably lead us back to the
relation between landlords and peasants and to the peculiar circumstances
surrounding the displacement of peasants by slaves. This process has two
equally problematic aspects: the process by which many peasants were pushed
off their properties while larger proprietors gained possession of vast
concentrations of land; and the process by which slaves were set to work on
this land instead of the dispossessed peasantry. Furthermore, it is not simply
the peculiar outcome of struggles over land that needs to be explained. The
very fact of these struggles requires explanation.
Hopkins writes that ‘The central place of land in Roman politics sprang
from the overwhelming importance of land in the Roman economy.’13 At rst
glance, this may seem a trivial observation about the self-evident importance
of land in an overwhelmingly agrarian economy. Nevertheless, although
Hopkins does not make this clear, the point is that land and conicts over it
held a special place in Roman history beyond anything that can be explained
by its agrarian economy alone. The Roman ‘élite’ was arguably distinctive
in the degree to which it depended for its wealth on the acquisition of
land. Hopkins contrasts the Roman imperialist élite, for example, to the
Manchu conquerors of China in the seventeenth century ‘who latched on to
13
Hopkins 1978, p. 6.
the existing bureaucracy and became pensioners or sinecurists of the tax

system’.14 He might have added that, in the ancient world, ruling classes whose
wealth and status rested on possession of the state (of which possession of
landed property might be an adjunct) and on surplus extraction in the form
of tax or compulsory services was more the rule than the exception. Class
power rmly and primarily grounded in private property and on individual,
‘private’ surplus extraction may be precisely the special characteristic of
Greco-Roman antiquity (more on this later). Even as imperial administrators,
the Romans had as one of their principal concerns the looting (ofcial or
unofcial) of the local population largely for the purpose of investing the
prots of ofce in land, the only secure and steady source of wealth for
the Roman aristocracy. The particular salience of land, then, lent a special
character to relations and conicts between landlords and peasants. It may
also go a long way toward explaining other remarkable characteristics of the
Romans, including their extraordinary militarism and imperialism. It is only
against the background of these factors that the growth of Roman slavery
can be understood.
Perhaps nothing better expresses the peculiar and paradoxical nature of
the relations between landlords, peasants, and the state in Rome – as well as
their effects in determining the growth of slavery – than the military position
of the Roman peasantry. The Roman peasant was the military backbone of
Roman imperial expansion.15 For many peasants, it was by means of their
exploitation as soldiers rather than through direct extraction of surplus that
they produced the wealth of their aristocratic compatriots. This had several
often contradictory effects. The military role of the peasantry, for example,
was the basis of slavery in more ways than one – not only in the obvious
sense that the slave supply was directly dependent on military conquest,
or that absent peasants were often replaced by slaves in agricultural pro-
duction, but also in the sense that the frequent and prolonged absence
of peasants on military campaigns was a critical factor in making their
14
Hopkins 1978, p. 14.
15
It is worth noting that, while Athenian peasants, too, had military obligations,
they were not as onerous as those of their Roman counterparts. Athenians were never
compelled to bear either the military or the scal burden of creating and sustaining
a vast imperial apparatus. These factors no doubt helped to limit the necessity and
the possibilities of agricultural slavery in Athens.
expropriation possible. ‘Roman peasant soldiers,’ as Keith Hopkins puts

it, ‘were ghting for their own displacement’.16 The military activities
of peasants made them less available for direct exploitation by virtue of
physical absence, but no doubt also by virtue of the powers of resistance
afforded by possession of arms. Imperial expansion also, at least for a time,
provided peasants with an alternative source of income which made them
less dependent on land they had lost and gave them a mobility that often
allowed them to replace ancestral farms which they had been compelled to
give up. In many respects, then, the military role of Roman peasants was a
two-edged sword. In any case, it made the growth of slavery both possible
and, in a sense, necessary. The paradoxes are nicely summed up in Keith
Hopkins’ argument that ‘one of the main functions of slavery was that it
allowed the élite to increase the discrepancy between rich and poor without
alienating the free citizen peasantry from their willingness to ght in
wars for the further expansion of the empire’.17 However this contradictory
situation is viewed, it is clear that the growth of Roman slavery can be
explained only within the framework of the very particular relations between
Roman landlords and peasants.
In both Greece and Rome, then, the distinctive relations that had evolved
between landlord and peasant and the development of the peasant-citizen
account for much else that is distinctive in their social formations when
compared to other ancient civilisations, including the unprecedented growth
of slavery. At the same time, there are signicant differences between Athens
and Rome; and these can be traced to differences in the outcome of the
struggle between landlords and peasants. In both cases, the early relations
between the two agrarian classes issued in a reduced availability of poor
citizens as dependent labour, which may have acted as an incentive to
subject aliens. In Athens, however, the dominant landed class never succeeded
in expropriating peasants or concentrating land to the extent that the Romans
later did, nor did they succeed in maintaining an aristocratic state. In Rome,
the victory of large landed proprietors was more thorough, though never
complete; and the ‘senatorial’ economy together with the aristocratic state,
16
Hopkins, p. 30.
17
Hopkins, p. 14.
unlike the democratic polis, made possible imperial expansion and property-
concentration, together with forms of slave-utilisation that never existed in
Greece. In Rome, too, the groundwork was laid for a reversal of the process
which had created the peasant-citizen and limited the exploitation of
peasants, and a foundation was established for a new subjection in medieval
serfdom.
It is, nally, the relation between landlord and peasant that provides the
element of continuity in Greco-Roman history between the formation of the
early Greek states to the decline of the Roman Empire and the rise of
feudalism. The rise and decline of slavery appears in this historical process
as a ‘dependent variable’. We shall return to a brief consideration of these
processes and to the question of the role of slavery in the ‘decline of the
Roman Empire’ – an historical process that perhaps more than any other has
been explained by reference to the institution of slavery, the consequences of
declining slave supplies, the stultifying effects of slavery on technological
progress, and so on. A consideration of this question should put the explana-
tory value of the ‘slave society’ to the ultimate test. First, however, a look at
some of the most important recent works that have in various ways treated
Greece and Rome as ‘slave societies’, ‘slave economies’ or instances of the
‘slave mode of production’.
‘Slave society’, ‘slave economy’, ‘slave mode of production’

Several important books have appeared in recent years which, in one way
or another, treat slavery as the distinctive feature of the Greco-Roman
world. Each has its own conception of what it means to designate Greece
and Rome as slave societies, though none rests this characterisation on
the predominance of slavery in production. In fact, their arguments often
proceed, explicitly or implicitly, as explanations of the sense in which it
is still possible and important to identify Greco-Roman civilisation as a
slave system despite the fact that the bulk of production was not generally
performed by slaves.
M.I. Finley – most recently in Ancient Slavery and Modern Ideology – suggests
that Greece and Rome were slave societies in the sense that, while slavery
occurred elsewhere, there existed in the Greco-Roman world ‘something rare
throughout history’ (there have been only ve ‘slave societies’ throughout
human history): ‘. . . an institutionalised system of large-scale employment of

slave-labour in both the countryside and the cities’.18 He does not mean by
this that slavery in general displaced or even predominated over free labour
but rather that slaves constituted the permanent workforce ‘in all Greek or
Roman establishments larger than the family unit’,19 and that slavery tended
to displace other forms of ‘involuntary’ labour. The centrality of slavery does
not here rest on its weight or preponderance in production generally – for
example, the question is not how much of Greek and Roman production
actually took the form of large-scale units worked by slaves rather than
small-scale farming and petty commodity production dominated by free
producers and family labour. Nor is it Finley’s primary object to explain
the general social consequences of a system of production in which larger
enterprises are dominated by slave labour and where slave labour predom-
inates over other forms of involuntary labour. The essential point is that the
very existence of such a rare formation signies something important that
demands an explanation. (We shall return to the explanation offered by Finley.)
A similar argument is made by Keith Hopkins in Conquerors and Slaves, which
also emphasises the rarity of slave societies.
Perry Anderson in Passages from Antiquity to Feudalism adopts the term
‘slave mode of production’ rather than the more exible ‘slave society’ to
describe Greece and Rome from early Attica to late imperial Rome.20 His
intention, however, is not to make any untenable claims about the preponder-
ance of slavery in production. He acknowledges the continuing importance
of free labour, especially in agricultural production, in these overwhelmingly
agrarian societies. Nevertheless, he argues, slavery represented the dominant
mode of production, in the sense that it cast its shadow over all other coex-
isting forms of production – and he has in mind especially its ideological
shadow. So, for example, the social degradation of labour by its association
with slavery stied technical inventiveness and obstructed the development
of productive forces. Anderson also suggests that, though class divisions
existed within the citizen communities of both Greece and Rome, these
18
Finley 1980, p. 67.
19
Finley 1980, p. 81. The emphasis on permanent is Finley’s.
20
Anderson 1974.
divisions paled before the more essential opposition between citizens, or

free men generally, and slaves. Furthermore, the most important cultural
expressions of Greek and Roman society – such as the works of Plato, Aristotle,
or Xenophon in Athens – must also be understood as reections of the slave
mode of production.
Pierre Dockès, whose Medieval Slavery and Liberation is largely concerned
with the decline of slavery and the consequent emergence of feudalism, puts
slavery at the heart of things for a more fundamental reason. For him, ‘Slavery
is not one type of domination and exploitation among others . . . [but] the
primary and primordial relation of exploitation, that form of exploitation
out of which serfdom and wage labour arise, and that form toward which
the master always strives . . .’.21 Thus, presumably, any exploitative society
could be fruitfully explained in terms of its natural tendency toward slavery
and the factors which promote or obstruct that tendency. Since Greco-Roman
antiquity represents a case in which chattel slavery actually, and not just
potentially, was the ‘chief form of exploitation and oppression’,22 a consider-
ation of its internal contradictions, the conditions that sustained it and the
forces that led to its decline, provides a key to the nature of exploitation in
general and to the possibility of its supersession. In any case, the particular
dynamic of Greco-Roman history derived from the exploitative relations of
slavery and more specically from the class struggles in which slaves resisted
oppression. This resistance ultimately destroyed the imperial state upon which
the whole system eventually depended and brought about the ‘medieval
liberation’, the decline of slavery and its replacement by serfdom.
The most systematic argument has been presented by G.E.M. de Ste. Croix
in The Class Struggle in the Ancient Greek World. He denies quite explicitly
that in Greek and Roman society ‘. . . the bulk of production was done by slaves,
or even (at least until the later Roman Empire) by slaves, serfs and all
other unfree works put together – I am sure it was not: in my opinion, the
combined production of free peasants and artisans must have exceeded that
of unfree agricultural and industrial producers in most places at all times, at
any rate until the fourth century of the Christian era, when forms of serfdom
21
Dockès 1982, p. 2.
22
Dockès 1982, p. 4.
became general in the Roman Empire’.23 Consequently, he continues, ‘. . . it

would not be technically correct to call the Greek (and Roman) world “a slave
economy”’.24 And yet, he nonetheless regards this designation as reasonable
and accepts the opinion of Marx that ‘Direct forced labour is the foundation
of the ancient world’.25 Again citing Marx, he argues that the most essentially
distinctive characteristic of any society is not the specic form in which the
labour of production itself is conducted, but the form in which surplus is
extracted, that is, the dominant mode of exploitation. Unfree labour was the
dominant form of surplus extraction, and slavery ‘played at some periods a
dominant role and was always a highly signicant factor’26 in unfree labour –
indeed, was (with debt bondage) ‘the archetypal form of unfree labour’.27 The
formula ‘slave economy’, therefore, can be accepted as a characterisation
of Greco-Roman society. This emphasis on the dominant mode of surplus
extraction is at least implicit in other accounts (notably Anderson’s and
Dockès’s, but perhaps even Finley’s); but only Ste. Croix states it so precisely.
The argument, then, is that even if the bulk of production is performed by
producers who remain outside the system of exploitation – acting neither as
exploiters nor exploited – a society can still be characterised by the specic
form in which surplus is extracted from the minority of exploited producers.
In the Greek world, of course, slaves (and other unfree labourers) played
a substantial if not predominant role in production; and, though free pro-
ducers were in practice often exploited, Ste. Croix suggests,28 the ‘typical’
Greek (like the ‘typical’ American Southerner in the era of slavery) was an
independent small producer not involved in the system of exploitation.29 In
this sense, ‘direct forced labour’ and slavery in particular, the predominant
form in which labour was exploited, remains the distinctive feature and
foundation of the ancient Greco-Roman world.
Although Ste. Croix never makes it entirely clear why the mode of surplus
extraction should be accorded such centrality, he provides some important
23
Ste. Croix 1981, p. 133.
24
Ibid.
25
Ibid.
26
Ibid.
27
Ste. Croix 1981, p. 173.
28
Ste. Croix 1981, p. 33.
29
Ste. Croix 1981, p. 55. On p. 53, Ste. Croix seems to suggest that the same is true
of the Romans.
hints – and Marx himself does more than that. In particular, the mode of
surplus extraction is the form in which surplus is supplied to the dominant
classes, that is, it is the specic form in which those classes and their
dominance are created, sustained, and reproduced. And, as Marx tells us in
Volume III of Capital,
The specic economic form, in which unpaid surplus labour is pumped out
of the direct producers, determines the relationship of rulers and ruled, as
it grows directly out of production itself and, in turn, reacts upon it as a
determining element. . . . [and] reveals the innermost secret, the hidden basis
of the entire social structure, and with it the political form of the relation
of sovereignty and dependence, in short, the corresponding form of state.30
Thus, to extrapolate from Ste. Croix’s argument, if slavery was the dominant
form of exploitation in the Greek world, then it provides the key to the whole
system of domination and the nature of the state. Indeed, one might argue,
assuming that slavery was the very condition of existence of a dominant
propertied class, it was also the condition which underlay the relations between
that class and the class of free small producers and any conicts that may
have existed between these classes. Free producers may have been essential
to production in general; and the conicts between them and the propertied
classes may have been crucial in determining social change, political devel-
opment, and historical movement in the Greco-Roman world. Nevertheless,
it could still be argued that slavery remains the essential force in Greco-Roman
history.
Slavery as the key to Greco-Roman history?

While the ‘slave society’ means many different things to different people,
it must be assumed that all those who apply it to the Greco-Roman world
share the conviction that slavery in one way or another provides the essen-
tial key to social life and historical process in ancient Greece and Rome. Again,
however, there is little agreement about precisely what, and in what sense,
it explains.
30
Marx 1971, p. 791.
In Finley’s case, slavery is not the ‘primordial fact’ in relation to which

other phenomena can be explained. Instead, slave society is a rare occurrence,
the fact which itself most demands explanation. He certainly agrees that ‘Once
established, a slave society had its own dynamic’,31 its own conditions of
maintenance, expansion, and decline, which helped to shape the course of
social relations and historical movement; but the essential explanatory value
of the formula ‘slave society’ lies in the nature of slavery as effect even more
than as cause.
Anderson, Dockès, and Ste. Croix, in contrast, are more inclined to take
slavery for granted and thus to treat it as prime mover. Dockès especially
proceeds from slavery as ‘primordial fact’. Ste. Croix very nearly does – at
least he assumes that slavery was the naturally preferred mode of exploita-
tion whose absence or decline is more in need of explanation than its
presence or emergence. It is true that he places conditions on the intrinsic
preferability of slavery as a mode of exploitation but, as we shall see, his
account of these conditions does little to dissuade us from taking slavery for
granted.
We shall return in a moment to the question of slavery as given, ‘primor-
dial fact’. The issue now is precisely how slavery, once established and by
whatever means, can be regarded as the key to Greek and Roman social life.
The most important questions have to do with the role of slavery in social
conict and class struggle, its relation to the form and evolution of the state,
and its ideological effects. Only Anderson and Ste. Croix have attempted
a comprehensive history of Greco-Roman antiquity which explores social,
political, and ideological developments together as a more or less coherent
whole, in such a way that we can judge whether the notion of ‘slave society’
or ‘slave mode of production’ has been vindicated.
If production in the Greek world rested at least equally on free labour as
on slavery, the most convincing grounds for singling out slavery as the
‘distinctive feature’ is the argument offered by Ste. Croix about the centrality
of surplus extraction.32 As a general principle, it seems to me quite correct –
31
Finley 1980, p. 92.
32
I have discussed Anderson’s argument in more detail elsewhere: Wood 1981a.
(Please note that an error in the text of this article is corrected in History Workshop
Journal Autumn 1982, pp. 178–9 in a letter to the editors in no. 14. I do not know why
publication of this correction was so long delayed.)
both as an interpretation of Marx and as a principle of historical explana-

tion – to say that the whole structure of social domination, the political power
that sustains it, the social conicts and ideological struggles that surround
it, and the historical dynamic generated by these conicts and struggles, are
all grounded in the process of surplus extraction and shaped by the specic
form in which this process takes place. It also seems to me demonstrably
true that Greek and Roman history can best be explained in these terms. The
most important social, political, and cultural developments in Greece and
Rome are fundamentally unintelligible without tracing their connections
to the prevailing modes of surplus extraction and the conditions of their
maintenance and reproduction, the struggles arising from the efforts of some
classes to maintain and enhance their exploitative powers and of other classes
to resist them.
Whether or not these principles justify the special role assigned to slavery
in the formula ‘slave society’ or ‘slave mode of production’, however, is
another matter. The proposition is especially problematic in the case of Athens,
if the role of slavery in agricultural production was negligible and if peas-
ants produced wealth for their rich fellow citizens as tenants, sharecroppers,
and casual wage-labourers. There is, then, considerable doubt about the
relative degrees to which slaves and free men were exploited, and it is
questionable whether the evidence actually permits the exclusion of free
producers from the system of exploitation to quite the extent that these
formulae suggest. But, even if we accept both the centrality of surplus
extraction and the proposition that the ‘typical’ free man stood outside the
system of exploitation, questions arise.
If the mode of surplus extraction and exploitation is the essential char-
acteristic of any society, then the fact of exploitative surplus extraction is
more fundamental still. In other words, the rst essential characteristic of
any society is whether surplus is ‘pumped out of the direct producers’ by
someone other than themselves. One cannot simply set aside free producers
on the grounds that they neither exploit others nor suffer exploitation
themselves. Especially if they exist in substantial or even preponderant num-
bers and perhaps even account for a greater proportion of basic production
than do other forms of labour, it must in itself be a highly signicant fact
that they are not exploited and that much or most of production takes place
outside the system of exploitation and without surplus being pumped out of
the direct producers. The existence or non-existence of exploitation, the

organisation of production in exploitative rather than non-exploitative ways,
is, after all, the difference between a class system and a classless society.
The freedom of so many direct producers from exploitation must be treated
as an essential factor in describing and explaining the nature of social arrange-
ments, class conict, ideology, and the form of the state. One must ask how
such a strikingly unusual arrangement came about, and precisely what in
the system of social organisation, class relations, and political power made
it possible. And, at the very least, in explaining the nature and function
of slavery itself, one must ask whether the unusual development of slave
production was cause or effect of the equally unusual system of class rela-
tions in which so many direct producers managed to free themselves, or to
remain free, from exploitation.
At the same time, it must be said that the form in which production is
carried out cannot be treated as peripheral, even if we place surplus extrac-
tion at the centre of our analysis. If it turned out that the bulk of production
was performed by free men rather than by slaves (as seems to be the case in
Athenian agriculture), this alone might raise questions about slavery as the
primary form of surplus extraction. At the very least, it would be necessary
to ascertain whether the proportion of social production carried out by
slaves was sufcient to account for the amount of surplus appropriated by
the propertied classes. After all, surplus must rst be produced in order to be
extracted.
As it happens, both Anderson and Ste. Croix, however much they stress
the centrality of slavery, at least implicitly place free producers and the
class struggle between free producers and their appropriating compatriots at
the heart of their historical explanations. In both cases, there is, in fact, a
noticeable disjunction between, on the one hand, the conceptual analysis
which denes the slave economy or the slave mode of production, and on
the other, the actual explanation of historical processes and the particular
conguration of social, political, and cultural forms in Greece and Rome. Ste.
Croix, for example, provides both the most convincing theoretical case for
regarding the Greek world as a slave society and the most detailed historical
account of social, political, and ideological developments which cannot be
explained without according a determining role to class struggles within the
citizen population.
Ste. Croix’s book is divided into two parts. The rst is devoted to a very
useful conceptual explanation and a detailed catalogue of property forms
and modes of exploitation in general and in the Greek world in particular
(though one is never quite sure when he is speaking of Greece, when of Rome,
or in what periods). The second part purports to apply the Marxist categories
explicated in Part I, notably the concept of class struggle, to an explanation
of historical developments, the processes of social and political change in
the Greek world. It is worth noting that Part II begins with a section entitled
‘Class Struggle in Greek History on the Political Plane’. Here, the relations
between ‘propertied’ and ‘unpropertied’ classes within the free population
take centre stage and slavery recedes. Although Ste. Croix points out that
there was seldom a simple coincidence between political divisions and class
divisions, he clearly believes that class divisions within the citizen body were
essential in determining the direction of social and political change.
In other words, it is precisely when the object is to explain specic historical
processes that the centrality of this class struggle, as distinct from the relations
between masters and slaves, becomes evident. Although Ste. Croix breaks
little new ground here, it becomes clear, for example, that the nature of Greek
democracy, its rise and fall, were the product of a class struggle between free
producers and their appropriating compatriots; that the democratic state
was a means of protecting the one class from exploitation by the other;
that the tensions between these classes never ceased to play a central role
in determining the direction of social change and political action; that (and
this is a point on which Ste. Croix is especially good) the imperial govern-
ment in both the Hellenistic state and the Roman Empire took much of its
shape from the survival of democratic polis and the conscious efforts of the
state to strengthen itself and the propertied classes by gradually absorbing
and destroying democratic forms.
Ste. Croix’s explanation of ‘class struggle on the ideological plane’ also
centres on the opposition between free producers and their propertied class
enemies. For example, he treats Plato’s philosophy, with its contempt for
labour, as anti-democratic propaganda directed against the role of ordinary
peasants and artisans in Athenian political life.33 And Aristotle’s perceptive
33
Ste. Croix 1981, p. 412. [A similar argument was made earlier in Wood and Wood
1978.]
analysis of social conict, which Ste. Croix deeply admires – and which he
cites as conrmation that the notion of class struggle as applied to ancient
Greece is no latter-day imposition but a reality perceived by the Greeks
themselves34 – is concerned precisely with relations between classes within
the citizen body, not relations between masters and slaves. The role of
slavery in all this is far from clear. There is, signicantly, no chapter on
‘class struggle on the economic plane’ in the historical section of Ste. Croix’s
work, no discussion of how the ‘plane’ on which slaves presumably played
their most prominent role gures in the historical process, or even how the
political and ideological ‘planes’ are related to it.35
In Part I, Ste. Croix demonstrates why it is wrong to speak of class only
where there is class consciousness and active political conict (as Ralf Dahren-
dorf, incorrectly, claims that Marx does). He also makes a case for referring
to the relations between masters and slaves as class struggles even when
no overt struggle occurred. His point here is ‘to bring back exploitation
as the hallmark of class’. This is no doubt important, though broadening
the denition of class struggle (to represent any class relations, oppositions, or
antagonisms) is not necessary to achieve that object. It is certainly import-
ant to recognise that the objective fact of exploitation and the inherently
antagonistic relations it entails have profound effects on historical processes,
even in the absence of overt conict and political struggle – something
that many historians and sociologists forget. Exploitation always requires
34
Ste. Croix 1981, pp. 77–80.
35
To speak of economic and political ‘planes’ as if they were separate compart-
ments is in any case rather misleading, especially in the case of precapitalist societies
where ‘economic’ powers of exploitation generally rest to a great extent on access to
‘extra-economic’ powers. (For a detailed discussion of this point, see Wood 1981b. [A
somewhat revised version of the article appears in Wood 1995.]) Ste. Croix’s failure
to acknowledge this difference between capitalist and precapitalist societies allows
him to make some misleading statements about the relationship between political and
economic divisions, class conicts and political contests, in ancient Greece, and, for
example, some misleading analogies between liturgies and property conscations in
ancient Athens and taxation in modern democratic states, Ste. Croix 1981, p. 97. He
also seems to misunderstand the signicance of certain Marxist analyses that stress
the degree to which exploitation in precapitalist societies is based on ‘direct relations
of domination and subjection’. Ste. Croix regards such analyses as ‘clearly contrary
to the views of Marx’ (p. 98), when, in fact, the distinction between ‘economic’ and
‘extra-economic’ modes of surplus extraction, the latter involving various forms of
direct domination and subjection, is precisely the basis of Marx’s distinction between
capitalist and precapitalist modes of exploitation.
means and institutions to sustain and enforce it and to overcome, suppress,

contain, or forestall resistance; and these requirements go a long way toward
explaining the nature of social organisation, political arrangements, and
ideological forms. Ste. Croix’s conceptual analysis of class in general and
slavery in particular is especially welcome for reminding us that the failure
of slaves to play a signicant political role does not by itself entitle us to
deny their existence as a class (though other problems may remain).36 But, to
demonstrate by denition that slaves are a class is not the same as showing
how this ‘classness’ actually affected historical process.
The point, then, is that it needs to be made clear how the largely silent
class struggle between masters and slaves determined the course of Greek
and Roman history, if we are to accept the explanatory value of the propo-
sition that slavery was the distinctive feature of Greco-Roman civilisation.
Ste. Croix has surprisingly little to say about this, much less than he says
about the other class struggle. He now and then hints at ways in which
slavery may have affected this other class conict: for example, he draws
36
Although much of what Ste. Croix says about the existence of class in the absence
of political struggle is undoubtedly correct and important, his argument is somewhat
vitiated, again, by his failure to acknowledge the special role of the ‘political’ in pre-
capitalist societies in this respect. His argument is, however, important in countering
certain objections to the application of class analysis to classical antiquity. It is often
argued, for example, that the concept of class is inapplicable here because people –
and the masses of the peasantry in particular – were never conscious of themselves
as a class nor able to act together in conscious pursuit of their class interests. Although
Ste. Croix may obscure the issue by using the concept of class struggle too loosely,
his argument is useful in pointing out that the relations of exploitation have profound
consequences for the organisation of society and social processes even when they
are not expressed in conscious political struggles between classes. For example, the
necessity of enforcing property relations and ensuring and enhancing the appropria-
tion of surplus against the resistances of direct producers deeply affects the nature of
juridical and political forms, even when producers are not organised consciously as
a class, and certainly nds expression in ideological forms – in attitudes toward labour,
in philosophical justications of social hierarchies, and so on.
As E.P. Thompson has argued, if class is something more than a theoretical con-
struct imposed upon the evidence, it is because it can be seen as a pattern in social
relationships, ideas, and institutions when they are observed over time and through
periods of social change: Thompson 1968, pp. 10–11. To say this, incidentally, is
not to say (as Ste. Croix accuses Thompson of doing) that class exists only in the
presence of class consciousness; but it does imply that if ‘class’ is to have any mean-
ing, we must do more than simply declare its presence by denition. We must show
its dynamic at work in historical processes and in the shaping of social and political
relationships, whether or not this dynamic is expressed in overt displays of class
consciousness.
an analogy between slaves in Greece and immigrant workers in modern

capitalist societies or blacks in South Africa, suggesting that citizen labourers
in Greece, like indigenous workers in capitalist Europe or white workers in
South Africa, may constitute a kind of ‘labour aristocracy’ united with their
exploiters against the ‘alien’ element. Even here, however, Greek slavery
appears less as a primary determinant than as a modication of another,
more central class struggle. Slavery re-emerges from the wings only in the
nal scene, the ‘decline’ of the Roman Empire. Now, according to Ste. Croix,
a fall in the slave supply and a resultant decline in the protability of
slave-exploitation compelled the propertied classes to tighten the screws
on free producers and to create new forms of exploitation which were
eventually to become the foundation of feudal serfdom. The intensied
exploitation of free producers in turn undermined the foundations of the
Empire. The decline of the Empire is, in fact, the only aspect of the historical
process in Greece and Rome which Ste. Croix actually explains with specic
reference to slavery (though not really slaves in class struggle) as the central
determining factor. Even in this case, as we shall see in the concluding
section, the process is intelligible only when slavery, both its rise and its
decline, is placed in the context of another, overarching class relation between
landlords and peasants.
As it is, if one were to read only Part II of Ste. Croix’s massive work, the
section in which by his own testimony he begins to apply his conceptual
categories to the explanation of actual historical processes, one might con-
clude that the truly ‘distinctive’ characteristic of the Greek world was not
slavery but, rather, a particular kind of free producing class and a particular
kind of class struggle between that class and its propertied compatriots. The
centrality of free producers and the importance of slavery are not, of course,
mutually exclusive. On the contrary, the two are inextricably related, and
their historical trajectories are closely intertwined. The question is whether
to accord explanatory priority to slavery is the most fruitful way of looking
at these historical trajectories, or whether it obscures as much as it reveals.
Slavery: ‘primordial fact’ or historical oddity?

To begin at the beginning: the question may come down to the origin of ‘slave
society’ and whether it is ‘primordial fact’ or historical rarity. To argue the
latter is already to look for a prior historical determinant. This may not require
us to abandon slavery, once established, as ‘distinctive feature’ or ‘dominant
mode of production’; but it does compel us to ask different questions than
are raised by formulae which take slavery as given. The proposition – as put
forward by Pierre Dockès (following in the tradition of his distinguished
compatriot Fustel de Coulanges)37 – that slavery is not simply one form of
exploitation but the ‘primordial’ form from which all others grow and toward
which they constantly tend has little explanatory value. The proposition is
rather empty, given the empirical rarity of large-scale chattel slavery as a
form of production, the widely diverse forms of surplus extraction that have
existed historically, and the great differences in their respective conditions
of emergence, maintenance, expansion, and decline. As an historical pro-
position, it is useless, even if, in some very broad sense, we could accept that
in the best (or worst) of all possible worlds, and without the constraints of
historical circumstance, all exploiters would prefer to have slaves. It would
be better to say, at the very most, that exploiters can, on the whole, be expected
to strive for the most complete and effective means of enhancing their
surplus – at least, the means most effective under the prevailing material,
social, and historical circumstances. It remains an open question what means
are most effective under what circumstances, and to what degree and in what
conditions exploiters can full their wishes in a context of class struggle, which
is never one-sided. And, of course, it still needs to be said that all systems
of exploitation are not equally in need of maximum surplus extraction – in
particular, the necessity for maximising surplus-value is uniquely typical
of the capitalist system with its pressures for accumulation and the self-
expansion of capital. Dockès acknowledges this problem, at least indirectly,
in his effective criticism of those who apply anachronistic notions of ‘prot-
ability’ to the judgement of slavery; but nonetheless takes far too much for
granted the inherent and universal preferability of slavery as a mode of
exploitation.
Ste. Croix takes slavery for granted in an apparently more conditional form.
He argues that slavery was the best means of extracting ‘the largest possible
surplus from the primary producers’ under the prevailing conditions.38
37
Finley 1980, p. 67.
38
Ste. Croix 1981, p. 40.
Underlying this proposition is a more fundamental and universal rule that

unfree labour and hired labour are the only ways of consistently achieving
a large surplus.39 Under the conditions prevailing in ancient Greece and Rome,
slavery was the best available means of extraction. His attempt to prove that
agricultural slavery was prevalent in classical Athens, for example, rests to a
great extent on his abstract assumption concerning the inherent superiority
of unfree labour, and the presumption that what is in principle superior must
be historically predominant.40
There were, according to Ste. Croix, three basic conditions that made
slavery the preferred form of exploitation: the prevailing level of technology
and productivity; ‘the poor supply of free, hired labour’; and the easy avail-
ability and cheapness of slaves.41 All this may be true,42 and his account at
least has the advantage of acknowledging in principle that the preferability
of slavery is in some way conditional on historical circumstances – though
the superiority of unfree labour in general is apparently absolute. Nevertheless,
his argument leaves the critical questions unanswered, and unasked, and
effectively still takes slavery too much for granted. The abstract principle that
unfree labour and hired labour are intrinsically more effective than other
modes of surplus appropriation – such as the extraction of rents, nes, taxes,
and labour services from ‘free’ producers – must be argued rather than merely
asserted; but, even if we were to accept it in the abstract, Ste. Croix provides
no adequate reason for concluding that these inherently superior forms have
been predominant in reality, and that we can take their existence for granted.
The tremendous and universal importance of rents, taxes, and labour
services extracted from peasant proprietors throughout history makes
it difcult to take seriously the dismissal of these forms as major sources
of wealth (as Ste. Croix so easily dismisses them in the case of Greece), 43
simply on the a priori grounds that they are in principle inferior types of
exploitation. And to use the unavailability of hired labour as an argument
to support and explain the prevalence of slavery is questionable in the
39
Ste. Croix 1981, p. 53.
40
In Wood 1983, I discuss this argument at greater length and examine in detail
Ste. Croix’s use of the textual evidence concerning both slavery and hired labour.
41
Ste. Croix 1981, p. 40.
42
For a different evaluation of slavery and its costs, see Hopkins 1978, p. 10 and
pp. 108–11.
43
See Wood 1983.
extreme. Quite apart from the fact that hired labour may not have been as
unimportant as Ste. Croix claims, even by his own testimony, (since it seems
to have been the typical form of harvest labour in Athens, and since the
harvest has always accounted for a substantial proportion of labour in
agrarian societies), the argument is profoundly ahistorical. Until the advent
of capitalism, wage-labour has never and nowhere been a predominant
form of exploitation. (It is worth noting here that Ste. Croix may have a
tendency to generalise from capitalism in several ways, perhaps treating
the capitalist drive for accumulation as his criterion for determining the
advantages of forms of surplus extraction and identifying their preferability
with protability – according to anachronistic principles of cost-accounting.)
Neither has large-scale slavery been common. What can it possibly mean, as
an historical proposition, to say that wage-labour is the chief alternative whose
absence proves or explains the presence of slavery? How do we account for
the massive surpluses produced and appropriated by landlords and states
in other civilisations with material cultures and state apparatuses more
lavish than those of Greece, in similar conditions of low productivity and yet
without either widespread slavery or hired labour?
Here, we encounter another difculty in Ste. Croix’s argument: the ten-
dency to lump chattel slavery together with other forms of unfree labour –
which, in his catalogue, includes various forms of debt-bondage and serfdom.
Surely, these forms differ from one another sufciently in critical respects
(for example, in some cases, producers remain in possession of the means of
production, while, in others, they themselves become chattel property) so
that their conditions of existence, as well as their relative advantages and
disadvantages, vary considerably in nature and degree? In any case, how
should we measure advantages and disadvantages? Are we entitled to give
as little weight as does Ste. Croix to the disadvantages of slavery – such
as the difculties and costs of supervision or the problems arising out of
the fact that the master’s investment is embodied not simply in means of
production or labour-time but in the very person of the slave?
Furthermore, to conate the various forms of unfree labour is to obscure
a truly fundamental question: why slavery and not other forms, and why
in Greece and Rome and nowhere else in the ancient world, or, for that
matter, in very few places at any time? Perhaps the answer is that the civic
status of Greek and Roman peasants and artisans made them unavailable as
dependent labourers and made necessary other forms of surplus extraction.

This, by itself, cannot prove that slavery predominated in production, if
we do not accept without question Ste. Croix’s rst premise about the
intrinsic superiority of unfree labour over free or his assumption that abstract
superiority means historical predominance. What such an argument would
suggest, however, is precisely that the importance of slavery depended upon
the condition of free producers, that the growth of slavery cannot be taken
for granted as predetermined by its inherently greater effectiveness as
a mode of exploitation and cannot be explained without rst accounting
for the special character of Greek and Roman peasants and their relations to
landlords.
There are, in fact, hints of doubt in Ste. Croix’s own argument. In at
least one place, for example, he suggests that, in the late Roman Empire,
when there was ‘a considerable increase in the exploitation of small free
producers, the use of slave labour in the strict sense was in principle less
necessary’.44 This statement would seem to run counter to the assumption
that slave labour as a mode of surplus extraction is in principle and funda-
mentally preferable to the exploitation of small free producers (rent-paying
peasants and craftsmen, not hired labourers). The suggestion now seems to
be that slavery is preferable only when free producers are not available for
exploitation.
Here is, at least, a hint of how the crucial question might be posed: why,
and under what conditions, did free producers cease to be readily available
for exploitation, how, and under what conditions, was their availability restored
and increased, and why did their exploitation take the particular forms
that it did? Put this way, as we shall see, the question also sheds a different
light on the ‘decline’ of the Roman Empire; for, then, the explanation no
longer hinges – as Ste. Croix’s account suggests – on the slave supply. The
suggestion that the availability of free producers for exploitation determines
the importance of slavery would seem to undermine the argument that the
decline of slavery and the rise of feudal forms were essentially determined
44
Ste. Croix 1981, p. 113. He makes a similar suggestion about the relationship
between the unavailability of free producers for exploitation and the rise of slavery
in Athens on p. 141.
by a decrease in the slave supply, an accompanying decline in the prot-

ability of slavery, and the consequent need to tighten the screws on free
producers. The whole sequence of causality changes when one ceases to take
slavery for granted and if one assumes that the rise and the very existence
of slavery need to be explained, and not simply its absence or decline.
Slavery was undoubtedly a lucrative means of surplus extraction for Greek
and Roman propertied classes. ‘Yet,’ writes M.I. Finley,
for all the advantages (or apparent advantages), slavery was a late and
relatively infrequent form of involuntary labour, in world history generally
and in ancient history in particular. Advantages and disadvantages are not
essences but historical attributes that come and go under changing social
and economic conditions.45
Even if all exploiters could be shown in principle to be striving for slavery,

it would still remain to explain under what conditions they can achieve it
and why Greece and Rome alone among ancient civilisations succeeded in
establishing it as a widespread form.
Even this, however, is not the most fruitful way to ask the question. If
the desirability of slavery is to be treated as an ‘historical attribute’, slavery
must also be conceived as a totality, a social form with specic conditions
of maintenance, requiring other social institutions as well as political and
ideological forms to sustain and reproduce it. One might conclude by ask-
ing not simply how Greek and Roman propertied classes succeeded in
establishing and maintaining such a rare though ‘advantageous’ form of
exploitation, but why they were compelled to adopt such a complicated
form, one that required, for example, not only periodic coercive acts of
appropriation but a continuously coercive supervision of production, as
well as an elaborate and expensive state apparatus to support it. It is not
surprising that an ancient agricultural authority like Columella in his De Re
Rustica (I.vii.4–7) should question the practicability of allowing slaves to work
agricultural land when the landlord is unable to supervise them directly (as
must have been frequently the case in ancient Greece and Rome where large
proprietors often owned several widely scattered holdings); in such cases, he
suggests, tenants are preferable. The increasing subjection of free producers
45
Finley 1980, p. 77.
in the later years of the Roman Empire, and its accompaniment by the
partial liberation of slaves who were placed in possession of land to be
exploited in serf-like dependence, do not testify only to the decreasing prot-
ability of slavery in conditions of declining supply, as Ste. Croix suggests,
nor simply to a decline (as Dockès argues) in the political and military powers
of the state. These developments may also signal a decreasing need on the
part of the propertied classes for an unwieldy form of exploitation and the
burdensome apparatus required to maintain it. The growing weakness, for
a variety of reasons, of the Roman imperial state was certainly a factor in
rendering slavery unworkable; but it is also arguable that this weakness was
itself conditioned by the state’s decreasing usefulness, indeed its positive
obtrusiveness, to a propertied class already in the process of establishing new
relations of exploitation and increasingly able to subject free producers to
conditions of personal dependence.
Slavery and the ‘decline’ of the Roman Empire

This brings us to the question of slavery in the ‘decline’ of the Roman Empire.
The ‘decline and fall’ of Rome are often associated, if not equated, with the
decline of slavery; and the dissolution of the Roman Empire is the one specic
historical process which historians have most systematically sought to explain
by reference to slavery. So it is perhaps here that we ought to look for
the ultimate vindication of the ‘slave society’ and the notion that slavery
provides the key to Greco-Roman history.
The central question seems to be: what factors led to the decline of slavery
and its replacement by other forms of labour and exploitation, notably serf-
dom? Various explanations have been offered. Max Weber emphasised
the decline of the slave trade as the Empire reached its limits. Marc Bloch
argued that a general decline in trade and the money economy was inimical
to slavery. Marxists have stressed the limitations on productivity inherent in
slavery and its effects on the development of productive forces.46 Ste. Croix’s
recent work has rened the Weberian analysis from a Marxist perspective
46
I have criticised the Marxist argument concerning the effects of slavery on the
development of productive forces in Wood 1981a, pp. 14–15, 19–20. [A better discus-
sion of this point appears in Wood 1988.]
by explaining why the decline in the slave supply was signicant, arguing
that various factors contributing to a decreasing slave supply reduced the
inherent protability of slavery and compelled landowners to tighten the
screws on free producers, thereby undermining the social, political, and
military foundations of the Empire. Each of these arguments can be faulted
on its own terms as an inadequate explanation of slavery’s decline. 47 The
more fundamental issue, however, is whether the question about Rome’s
decline has been correctly posed at all. Is it fruitful to focus our primary
attention on factors leading to the decline of slavery, or should this question
be subordinate to, or subsumed under, another question or set of questions?
Rodney Hilton, in Bond Men Made Free, has indirectly suggested why and
how our attention should be redirected:
. . . there seems little doubt that peasantries were the basis of the ancient
civilisations out of which most European feudal societies grew; and that the
class of slaves, though economically and culturally of great signicance at
certain times and in certain sectors of the ancient world, was numerically
inferior and of less permanent importance than the peasant producers.
In fact, viewed from the standpoint of this most numerous class of rural
society, the difference between late Roman and early medieval civilisation
may not have been all that easy to discern. 48
Viewed in this perspective, the rise and decline of Greco-Roman civilisation

– including the rise and decline of slavery ‘at certain times and in certain
sectors’ – might be seen as the history of peasantries, their changing relations
to landlords and to the state. The question of the decline of Rome and the
rise of feudalism could then be reformulated, again more inclusively, as
follows: what factors led to changes in the form and extent of peasant exploita-
tion, including – temporarily – its relative decline and partial replacement
by slave-exploitation, and its re-establishment in the new forms of medieval
serfdom? Related to and co-extensive with this question – and again more
inclusive than the issue of slavery – are other questions concerning changes
in the nature of appropriating classes and the state. This mode of analysis
among other things has the advantage of focussing on continuities rather
47
For criticisms of various arguments on the decline of slavery see, for example,
Dockès 1982, pp. 117–49.
48
Hilton 1973, p. 10.
than upon a radical break between the Roman Empire and what came after.
The myth of Rome’s decline and fall, whether viewed as a sudden
cataclysmic collapse or a gradual dissolution, has tended to make feudalism
appear out of nowhere (or, at best, out of the alien barbaric North). First,
Rome declines and falls in accordance with its inner logic and inadequacies,
then feudalism lls the void. What is needed instead is a vantage point from
which the continuities are clearly visible and which permits us to discern the
emergence of specically ‘feudal’ relations and institutions within the social
and institutional framework of Imperial Rome.
European feudalism was characterised by three essential features: the ‘par-
cellisation’ of the state and its replacement by a patchwork of jurisdictions
in which state functions were both vertically and horizontally fragmented;
a ‘parcellisation’ of the economy, contraction toward a ‘natural’ economy;
and, above all, a signicant growth of personal dependence, a condition of
serfdom which to an unprecedented degree bound formerly free producers
both to the land and to individual appropriators in a relationship of depen-
dence that was at once and inextricably economic and political. These three
features can be viewed as three aspects of a single phenomenon: a new mode
of surplus extraction constituted by a decentralised fusion of political and
economic power.49 This new form of exploitation took the shape of a juridical-
political relationship between producer and appropriator, so that each unit
of production and appropriation was at one and the same time a fragment
of the state, and the lord was both private exploiter and ruler at once.
The question of the transition from Imperial Rome to Western feudalism
should therefore focus on whatever there was in the logic of Roman social
relations that tended toward fragmentation of the state, the growth of a parcel-
lised unity of political and economic power, and the increasing personal
dependence of formerly free producers. The specic institutional forms assumed
by these new relationships – the forms of infeudation and subinfeudation,
vassalage, etc. – may owe a great deal to alien intrusions; but, even these,
together with the manorial system, have Roman antecedents and could not,
in any case, have been implanted if an appropriate matrix of social relations
had not already developed to receive them in Rome.50
49
See Anderson 1974, pp. 147–8, and Wood 1981b, pp. 86–9.
50
For a brief summary of the ambiguities in the evidence concerning the prove-
nance of specic feudal institutions, see Anderson 1974, pp. 130–1.
No explanation that gives primacy to slavery can bear the weight of all
these developments. Ste. Croix’s explanation of Rome’s decline, for example,
hardly even attempts to account for the processes of economic and political
parcellisation which constituted new forms of personal dependence and
exploitation, nor does his analysis of class struggle encompass class relations
that might have impelled such processes. If his focus on the slave supply and
slave protability is meant to explain these larger economic and political
tendencies, it is only in the very general sense that the declining protability
of slavery, according to his argument, led to an increasing burden on free
producers which made them increasingly unable to support the military
and administrative apparatus that had always rested on their backs. There
is too much ‘decline and fall’ in this, and not enough on-going historical
process to explain what came after. Apart from the fact that Ste. Croix may
be reversing cause and effect, his explanation is inadequate because it makes
the emergence of feudal dependence and parcellisation contingent upon the
(relatively late) decline in the availability of slaves with no explanation of the
long-term tendencies in these directions without which no decline in slave
supplies could have produced the specic relations of feudalism. So, among
other things, he cannot even consider the possibility that the supply of slaves
may have been affected by the demand for them, and that the reasons for the
change in demand should be sought in larger economic and political trends.51
Ste. Croix explicitly contrasts his own approach to M.I. Finley’s account
of slavery’s decline. Finley, Ste. Croix tells us, notes the decline of slavery
and the fact that it ‘requires explanation’; and he ‘comes very near to saying
something valuable, when he declares that “the key lies not with slaves but
with the free poor” . . .’.52 Unfortunately, Ste. Croix continues, Finley conceives
of the process ‘from a supercial point of view’ as simply a general trend
(quoting Finley), ‘a cumulative depression in the status of the lower classes
among the free citizens’.53 This formula conceals ‘the mainspring and essen-
tial character’ of the trend and its foundation in class exploitation.54 Finley fails,
therefore, to explain ‘the changeover . . . from slave production to what I would
call mainly serf production. . . . The “explanation” should be precisely the
51
Finley 1980, p. 86, and Finley 1973, p. 93.
52
Ste. Croix 1981, p. 462, quoting Finley 1973, p. 86.
53
Ste. Croix 1981, pp. 462–3, quoting Finley 1973, p. 87.
54
Ste. Croix 1981, p. 463.
other way round: it was because slavery was not now producing as great a
surplus as it did in Rome’s palmiest days that the propertied classes needed
to put more pressure on the free poor.’55
It can be argued, however, that Finley comes closer here than does Ste.
Croix to offering an explanation, or at least pointing us in the direction of one,
and that this explanation – for all Finley’s doubts about status and class –
has implicitly more to do with class struggle than does Ste. Croix’s account.
Ste. Croix reads Finley’s argument as if it merely describes a trend – a depres-
sion in the status of lower class free men – which is symptomatic but which
cannot explain the ‘changeover from slave production to serf-production’.
The argument can, however, be read quite differently. When Finley writes
that the decline of slavery ‘requires explanation’, he means that it must be
explained in terms of a prior and more inclusive development, ‘a structural
transformation within the society as a whole’.56 His remark that ‘the key lies
with the free poor’ does in fact help to explain the ‘changeover’, by suggesting
that the relations between the free poor and their wealthier compatriots
is the key to the ‘structural transformation’ of which the decline of slavery
is simply one aspect or consequence. Where Ste. Croix argues that because
slavery was not producing an adequate surplus, free producers had to be
more exploited, Finley seems to be arguing essentially that, because free
producers could now be more exploited, slavery was no longer so necessary.57
The difference between these two positions rests on a different evaluation
of slavery and its place in history: as we have seen, Finley makes no universal
assumptions about the inherent preferability of slavery as a mode of exploita-
tion or about its ‘primordial’ character. Instead, he treats the advantages of
slavery as historical attributes within a context of historical conditions – social,
economic, and political. He acknowledges the rarity of large-scale slavery
and its distinctiveness as an unusual social formation whose very existence
needs to be explained. Since he poses the question of slave society in terms
of the conditions that made it possible and/or necessary in the rst place,
demanding an explanation of its emergence and rise, he must also pose the
question of its decline in a different way. Finley consequently looks for more
55
Ibid.
56
Finley 1973, p. 86.
57
Finley 1973, p. 93.
inclusive factors which can explain the structural developments within

which slavery rose and declined. Although he fails perhaps to pursue his
own argument to its logical conclusion by referring unambiguously and
unapologetically to class exploitation and class struggle, there can be little
doubt that his explanation of these structural developments rests heavily on
the class relations between landlords and peasants. Explaining the emergence
of slave society, Finley writes
The peasantry had won their personal freedom and their tenure on the land
through struggle, in which they also won citizenship, membership in the
community, the polis. This in itself was something radically new in the world,
and it led in turn to the second remarkable innovation, a slave society.58
Finley acknowledges that the slave society, once established, ‘had its own
dynamic [and that] the conditions that led to its creation were not identical
with the conditions that led to its maintenance, expansion or decline’. 59
Nevertheless, he clearly sees the decline as in some sense a reversal of the
original process, a restoration of an older social formation in which landlords
and peasants confronted one another as exploiting and exploited classes.60
And, just as earlier struggles had consolidated the personal freedom and land
tenure of the peasantry by making them members of the civic community,
the process which increasingly deprived them of their freedom and made
them more available for exploitation and involuntary labour took the form
of their gradual extrusion from the civic community. Finley traces this process
to the very beginning of monarchical government in Rome, a gradual decline
in the political and military power of the citizen poor, and the increasingly
insupportable burden imposed upon them by the state.61
Finley’s account has several advantages over Ste. Croix’s. It need make
no ahistorical assumptions about the inherent exploitative superiority of
slavery. It need not assume that landed proprietors in classical antiquity
operated according to anachronistic principles of protability and compara-
tive cost-accounting (which in any case can be used by historians either for
58
Finley 1973, pp. 89–90. Although Finley here speaks of the polis, he seems to have
in mind the Roman state as well, see p. 89.
59
Finley 1973, p. 92.
60
For example, Finley 1973, p. 93.
61
Finley 1973, pp. 86–7.
or against the efciency of slavery), especially principles that probably

make sense only in conditions where the uniquely capitalist compulsions of
accumulation prevail and where the preferability of a mode of exploitation
is synonymous with its productivity.62 It encourages us to look for larger
and more inclusive structural transformations. And it can take account of
both the similarities between Greece and Rome when compared to other
ancient civilisations, and the differences between Greece and Rome, which
can be explained by divergences in the outcomes of class struggle between
Above all, Finley’s argument gives us a sense of historical process which
Ste. Croix’s lacks. Finley’s argument at least suggests an ongoing process, a
constant tension, not a sudden and one-sided tightening of screws by the
propertied classes, but a continuing and changing relationship marked by
struggles in which free producers resisted subjection with varying degrees
of success. In both Greece and Rome, the success of free producers was
sufcient substantially to restrict the options of the dominant classes; but
it did not, even temporarily, exempt all free producers altogether from
exploitation or completely remove from within the free population the ten-
sions and conicts that characterise exploitation and resistance to it. Such a
view of things forces us to consider from the outset the various factors that
determined the relative successes of these classes, and to assess the role of
slavery in that context.
Finley’s approach has a further – and perhaps unintended – advantage.
Paradoxically, although it is Ste. Croix who emphasises class struggle and
justly remarks on Finley’s tendency to avoid the concepts of class and exploita-
tion, it is Finley’s argument that, at least by implication, places its stress on
class struggle. Like Ste. Croix, Finley asks ‘what motivated the upper classes . . .
to change over from slave gangs to tied peasants?’,63 as if the explanation
were to be sought only in the intentions of the upper classes. For Ste. Croix,
however, these intentions and the factors affecting the need to change from
62
It is worth noting that Ste. Croix’s reply to Finley’s criticism of such ‘cost-account-
ing’ arguments addresses itself simply to whether slave exploitation was inefcient
and unprotable. He apparently fails to see that the issue is whether the question of
‘protability’ is relevant at all in the way he suggests and whether it is legitimate to
attribute these principles of comparative cost-accounting to the ancients themselves.
See Ste. Croix 1981, p. 462, n. 18.
63
Finley 1973, p. 85.
one mode of exploitation to another have little to do with the action and
resistance of exploited classes. Finley, in contrast, offers an explanation
that leaves more room for the struggles of producers. It is an explanation
much more compatible, for example, with the insight so fruitfully applied by
Robert Brenner to another great structural transformation, from feudalism to
capitalism.
The historical evolution or emergence of any given class structure is

not comprehensible as the mere product of a ruling-class choice and
imposition, but . . . represents the outcome of class conicts through which
the direct producers have, to a greater or lesser extent, succeeded in
restricting the form and extent of ruling-class access to surplus labour. . . .
Finally, even if an established pre-capitalist ruling (or propertied) class does
not maximise its surplus, there is generally no economic necessity for it
to be surpassed by more effective ‘surplus maximisers’ (more ‘suitable’
methods of labour control) [i.e. methods that maximise productivity and
protability, the need for which results from the pressures of accumulation
specic to capitalism – pressures that are the result not, the cause of
capitalist relations of production. – EMW].64
The decline of slavery, then, cannot be explained solely by reference to its

inherent advantages nor to considerations of protability that owe their logic
to a different set of property relations. The explanation must be sought
in particular structural and historical conditions which made slavery the
‘preferred’ form in specic times and places. Among these conditions, the
factor of class conict, which affected and restricted the ‘form and extent of
ruling-class access to surplus labour’, is critical. What complicates the issue
in the case of Greece and Rome is that the conicts in question encompass
two separate but interacting sets of class relations, not only those between
masters and slaves but also, and in the rst instance, those between rentier-
lords and free producers.
Despite Finley’s focus on relations between classes, it can be argued –
as Pierre Dockès maintains – that, in the end, class struggle in his account
‘gives way to a process in which slaves are totally passive and the peasants
unresponsive, and in which even the masters, the great landowners, seem to
64
Brenner 1977, pp. 59–60.
be bystanders rather than agents of the transition to dependence from

which they were to prot’.65 The process to which all classes succumbed in
Finley’s explanation, argues Dockès, is (in Finley’s own words) ‘the iron
law of absolutist bureaucracy that it grows both in numbers and in the
expensiveness of its life-style’.66 Dockès, instead, places the accent on class
struggle in the most literal sense, indeed class violence, and more particularly
on the struggles of slaves. In Finley’s perspective, he suggests, ‘slaves are
in some sense strangers to their own history, to the history of their living
conditions; their struggles are “tempests in a teapot”, said to be without
consequence. The cause of the end of slavery lies “elsewhere”, in relations
among the “other classes”’.67 Although Dockès stresses how important it
is to recognise, as Finley does, ‘that relations between masters and slaves
cannot be studied independently of the relations between large landowners
and free peasants’,68 for him, the primary and critical factor is slavery – indeed,
so much so that he places an inordinate burden of explanation on the overt
and sporadic outbreaks of violent slave revolt (which he nds everywhere,
often on the imsiest evidence). It was such struggles that, according to
Dockès, nally undermined the foundations of the Roman imperial state.
The primacy accorded by Dockès to slave revolts (which often leads him
to treat free producers as passive pawns in much the same way that he accuses
Finley of treating slaves) is partly conditioned by his original assumption
about the primordial nature of slavery as the primary form of exploitation
out of which all others grew and toward which they tend. His analysis
is further coloured by his political objective, which is to demonstrate the
parallels between, on the one hand, ancient slavery and Roman imperialism,
and on the other, capitalism and modern imperialism. Thus, the decline of
the Roman Empire and the rise of feudalism are seen not so much as a process
in which free producers were enserfed as a process in which slaves were
liberated. The struggles of slaves – the primordial and paradigmatic struggle
against exploitation – becomes an object lesson for all exploited classes.
The weaknesses of an argument based on the primordial and universal
character of slavery have already been noted. There will no doubt also be
65
Dockès 1982, pp. 201–2.
66
Dockès 1982, p. 202.
67
Dockès 1982, p. 203.
68
Ibid.
critics for whom Dockès, in his arguments against the impersonal deter-
minisms of ‘vulgar’ Marxism or Finley’s ‘iron law of absolutist bureaucracy’,
goes too far in the opposite direction toward an excessively ‘voluntaristic’
conception of class struggle. This is not the place to engage in theoretical
debates about the appropriate balance in historical explanation between
objective necessity and subjective agency. There is, in any case, more to
Dockès’s argument than class violence. Dockès does not, in fact, conceive of
the historical process of Rome’s decline and the rise of feudalism as simply
a contingent series of violent and ‘voluntaristic’ encounters between classes.
Rather, he attempts to identify an ‘internal dynamic’ inherent in Roman
relations of production, what might be called a ‘developmental logic’, that
explains not only the decline of slavery and its replacement by other forms
of exploitation but also the disintegration of the state. He rejects the simplistic
determinism of those versions of Marxism which conceptualise away the
problem of historical change by postulating a unilinear, progressive, and
virtually mechanical succession of modes of production. At the same time,
he rejects explanations that rely too heavily on contingent or external factors
(like the closing of long-distance trade routes) or on demographic patterns
which are notoriously unreliable since population factors can so often be
shown to have opposite effects in different contexts.69 Instead, without denying
the effects of external factors, he tries to decipher the logic of process within
the prevailing social relations. In this respect, his argument has something in
common with Robert Brenner ’s account of the origins of capitalism, which
has proved such a fruitful advance over other approaches. It is, then, in
Dockès’s attempt to identify this ‘internal dynamic’ or developmental logic
that the importance of his argument lies, whether or not we are satised with
his answer.
The developmental logic that Dockès is seeking is the logic of a slave system.
To summarise his argument very briey: slavery produces a tendency, which
it shares with capitalism, toward expropriation of small proprietors and a
concentration of property and labour-forces into ‘centrally managed large-
scale productive units’70 (i.e. slavery produces latifundia just as capitalist
relations of production produce factories). This, in turn, creates a need for a
69
For an example of such contradictory patterns, see Brenner 1976 and Brenner
1982. [These articles appeared later in Aston and Philpin (eds.) 1985.]
70
Dockès 1982, p. 226.
strong centralised state, especially to deal with the dangers created by large
numbers of exploited workers joined together. These processes, however,
contain certain contradictions. There is, of course, an inherent antagonism
between masters and slaves, which tends to express itself in resistance and
revolt; and, in Rome, the oppression of the free producers tended eventually
to lead them into an alliance with slaves. The defection of the peasantry in
Rome helped to undermine the foundations of the state. There is also an
inherent contradiction between the landlords’ collective need for a strong
centralised state and their individual antagonism to it as an obstacle to and
a drain upon their exploitative powers. The inherent antagonism between
landlord and state further undermines the state’s foundations. The collapse
of the imperial state in Rome was thus brought about by slave revolts,
eventually aided by peasant allies, together with the weak foundations of
the state in the propertied class; and this collapse, in turn, brought about
the destruction of slavery. In other words, it was not the decline of slavery
(for whatever reasons) that determined the disintegration of the imperial
state but rather the disintegration of the state, encouraged especially by slave
resistance, that determined the decline of slavery.
This is a strong argument. Even if we set aside Dockès’s probably excessive
emphasis on slave revolts, there still remains a ‘logic’ in his slave society
which might account for structural weaknesses, contradictions, and an impulse
toward disintegration. His emphasis on the state and the nature of its
relation to slavery, and his suggestion of a causal sequence in which the
collapse of the state precedes the destruction of slavery, are particularly
important. A few questions, however, immediately arise: if slavery produces
a natural tendency toward the concentration of property, what were the
factors that encouraged this tendency in Rome but not so much in Greece?
And why did the tendency appear to operate even in some parts of the Empire
where slavery was not an important factor? If the squeezing out of small
producers and the concentration of property in the hands of large land-
owners occurred in order to make possible the concentration of a labour-force
in ‘centrally managed large-scale production units’, how do we explain the
fact that large-scale landownership so often took the form of scattered smaller
holdings?71
71
See, for example, Duncan-Jones 1974, pp. 323–6, on the sizes of estates in Roman
The whole analogy between slavery and capitalism is especially problematic,

not least because slavery did not and could not have an ability – or a need –
to squeeze out other modes of production comparable to capitalism’s unprece-
dented tendency to do so. If capitalist relations of exploitation have promoted
the development of a particular system of production – factory production,
with a concentrated, integrated, and centrally controlled labour-force – it
is not at all clear that slavery carried with it a similar impulse toward lati-
fundial production. Capitalism encourages certain forms of production because
it creates unprecedented pressures for accumulation and competition, and
because the capitalist owns only the labour-power of the worker and only
for a xed period of time. Capitalism therefore typically responds to the pres-
sures of competition and accumulation by increasing labour productivity,
and this requires transformations in the organisation of the labour-process.
The resulting high productivity, within the competitive system, tends to drive
out less productive forms of labour. The same pressures of accumulation and
competition and the need to increase labour productivity are not present
in slavery – especially since the master owns the worker’s person and not
simply his labour-time.
Slavery may create pressures to concentrate property; but this is, arguably,
less because of its inherent superiority as a mode of production and exploitation
than because of its disadvantages, the need to compensate for its costs and
its shortcomings in competing with other forms of exploitation. So, for example,
a need may arise to consolidate holdings in order to employ slaves full-time,
by mixing crops etc., because the investment in their persons might otherwise
exceed the returns, a problem that does not arise with wage-labourers or
tenants. In this case, the question is not how a ‘primordial’ system of exploita-
tion, by its own inherent natural logic and superior strength, squeezes out
other forms of production, but, rather, how and under what specic historical
restrictions such a difcult mode of exploitation becomes the most eligible
one. It is also worth stressing that the ‘logic’ of slavery depended very much
on surrounding circumstances. For example, it might in certain conditions
adapt itself to prevailing forms of production rather than establish forms
Italy and the ‘two contrasting tendencies in large-scale Italian landownership’:

the aggregation of enormous single units, and the ownership of scattered smaller
holdings.
of its own which tended to drive out others. Especially in classical Athens,
under the constraints of the democracy, traditional forms of individual craft
production were never superseded by gang production, and slaves commonly
produced as individual craftsmen. Even the few known large enterprises
that employed many slaves under one roof never achieved anything like
the integrated labour-force and division of labour characteristic of the
modern factory but essentially brought together in juxtaposition a number
of individual craft producers.
The question raised by Dockès’s analysis of slavery and its internal dynamic
may point to another, more fundamental question: if there was a tendency
toward concentration of property and a consequent – and contradictory –
need for a strong centralised state, should that tendency be traced to the
logic of slavery at all or to some other, prior source? In other words, is the
developmental logic for which we are looking the distinctive logic of a slave
society at all?
It is true that the possibility of large-scale slave-utilisation, especially with
‘centrally managed large-scale production units’, presupposes landowner-
ship on a scale sufcient to require and make possible a large labour-force.
This is almost tautological, but it does not necessarily mean that the ten-
dency toward concentration is produced by slavery. Apart from the fact
that latifundial slavery was only one form of slave-utilisation, temporally
and geographically limited, it could just as easily be argued that the very
motivation to acquire a large force of slaves was preceded, and in some sense
caused, by a tendency toward concentration of property. Dockès does little
but assert – usually by analogy with capitalism – that the logic of slavery
was to drive out small producers and concentrate property, except to say that
the massive inux of slaves helped to ruin the independent small peasant in
Italy by making him unable to compete in the market not only with imported
goods but with produce from large Italian estates. Questions must, however,
be raised about the degree to which peasants depended on the market for
survival, especially markets in which they would be forced to compete with
large estates. Peasant production was largely subsistence farming, and the
markets in which they operated were essentially local ‘peasant’ markets in
which petty producers exchanged necessities with one another.72 Dockès may
72
Finley 1973, p. 107.
here again be imposing the logic of capitalism and capitalist competition

on Roman society. But, even granting that the logic of slavery encouraged
concentration of property, this logic presupposes the existence of large estates
and a prior process of concentration which slavery only aggravated.
At the same time, and paradoxically, the growth of slavery seems also to
presuppose a failure on the part of large landowners to reduce the peasantry
to dependence – that is, a failure to deny peasants access to the means of
subsistence and reproduction without being compelled to perform involun-
tary labour for others. Again, slavery rose and declined in inverse ratio to the
availability of free producers for exploitation. This would seem to suggest
that slavery might be most necessary precisely where peasants remained in
possession of the means of production. Is this true, and does it contradict our
earlier assumption that slave-utilisation and peasant expropriation go hand
in hand?
Large-scale landownership existed in various parts of the Roman Empire,
but large-scale slave-utilisation was not equally widespread. Slavery on a
large scale was primarily conned to the Western Empire, and especially to
Italy and Sicily. Large landownership was not so conned. For example,
Roman senators owned vast properties in the province of Africa (Pliny the
Elder claims that six men at one time owned half the province, see Natural
History, 18. 35.); yet slavery here seems to have been relatively unimportant.
Although landholdings in the East were never as great as in the West, there
were, nevertheless, substantial concentrations of land; and, here, forms of
exploitation other than slavery predominated.73 Especially in these areas,
surviving peasantries were not generally independent proprietors largely
free of exploitation by the rich, but often subject populations available for
various forms of exploitation by great landowners supported by the Roman
imperial apparatus.
The crucial variable in determining the spread of slavery seems to be not
only the amount of land available to large proprietors but the degree to which
free populations were subjected. Even Dockès writes that
. . . in the remainder of the empire [i.e. outside Sicily, and the western
provinces] . . . large estates were not dependent on slave-labour groups of
73
On the probable distribution of slavery and other forms of agricultural labour in
the Empire, see Jones 1973, pp. 792–4, and White 1970, pp. 411–12.
the classical type; instead, virtually the entire local population was sub-
jugated, which accounts for the fact that these regions (1) did not import
slaves, and/or (2) were hunting grounds for pirates and slave traders
supplying the West. 74
In other words, it is the fate of free producers that accounts for the growth
of slavery and not the reverse.
The speci c dynamics of class

The growth of slavery in Greece and Rome, then, seems to have depended
on a delicate balance between two contradictory tendencies: a concentration
of property in the hands of large landowners, and a resistance to expropriation
and/or subjugation – successful in varying degrees – on the part of small free
producers. Are we, then, looking at a ‘developmental logic’ in the relations
between landlords and peasants? And, if so, how might it account not only
for the development of slavery but also (directly, and not just through the
medium of slavery) for the contradictory tendencies toward state centralisa-
tion and political fragmentation, which Dockès rightly identies as central
to the logic of Roman history?
Greece and Rome are distinguishable from other ancient civilisations
not only by their utilisation of slaves in unprecedented ways and degrees,
but also by their distinctive relations between landlords and peasants and
between both these classes and the state. The typical ancient state was the
‘bureaucratic’ kingdom in which the state exercised substantial control over
the economy, property in land tended to be closely bound up with state
service, and peasant producers were subject to surplus extraction less in the
form of personal subjection to individual private proprietors than in the form
of collective subjugation to the appropriating, redistributive state and its
ruling aristocracy, especially through taxation and compulsory services.75
Classes confronted each other not simply as individual appropriators and
producers, or large and small proprietors, but collectively as appropriating
states and subject peasant villages. Although states of this kind, at least on
74
Dockès 1982, p. 54.
75
For a more detailed discussion of the contrast between these different forms of
state, see Wood 1981b, pp. 82–6.
a small scale, seem to have existed in Bronze Age Greece (as the archaeo-
logical remains of Mycenaean civilisation and the decipherment of Linear B
reveal), they completely disappeared and were replaced by new forms of
social and political organisation. Unfortunately, the process by which this
replacement occurred remains obscure.
What is important from our point of view is that, in Greece and Rome,
in the absence of this form of state and its characteristic relations between
ruling and subject groups, appropriators and producers confronted one
another more directly as individuals and as classes, as landlords and peasants,
not primarily as rulers and subjects. Private property developed more
autonomously and completely, separating itself more thoroughly from the
state. In other words, a new and distinctive dynamic of property and class
relations was differentiated out from the traditional relations of (appropriat-
ing) state and (producing) subjects. We have seen this specic dynamic at
work in the struggles over land, which were so central to Greco-Roman
history. Indeed, one might say that Greece and Rome were distinctive precisely
in the degree to which a differentiated dynamic of class conict was at work,
with a logic of its own.
New forms of state emerged out of these relations. The ancient ‘bureaucratic’
state had constituted a ruling body superimposed upon and appropriating
from subject communities of direct producers. Although such a form had
existed in Greece, both there and in Rome a new form of political organisation
emerged which combined landlords and peasants in one civic and military
community.76 The very notions of a civic community and citizenship, as distinct
from a superimposed state apparatus and rulership, were distinctively Greek
and Roman. The unity of appropriators and producers, rich and poor, embod-
ied in this new form of state was, as it were, a ‘harmony of opposites’ (to
adopt a concept beloved by the Greeks), imbued throughout with the ten-
sions and contradictions, the internal dynamic, of the conicts between and
within these two classes.
The special characteristics of these states are reected in the classics of
ancient political thought. When Plato, for example, attacked the democratic
76
This description does not, of course, apply equally to all parts of Greece. Sparta
and Crete are the most notable examples of Greek states in which the citizen com-
munity ruled over a subject population of producers.
polis of Athens, he did so by opposing to it a state-form which departed

radically from precisely those principles most unique and specic to the Greek
polis and which bore a striking resemblance in principle to certain non-Greek
states. In the Republic, Plato proposes a community of rulers superimposed
upon a ruled community of producers, primarily peasants, a state in which
producers are individually ‘free’ and in possession of property, not depen-
dent on wealthier private proprietors, but collectively subject to the ruling
community and compelled to transfer surplus labour to their non-producing
rulers. Political and military functions belong exclusively to the ruling class,
according to the traditional separation of military and farming classes which
Plato and Aristotle both admired.77 Plato, no doubt, drew inspiration from
the Greek states that most closely adhered to these principles – notably Sparta
and Crete; but it is likely that the model he had more specically in mind
was Egypt – or, at least, Egypt as the Greeks understood it.78
Other classical writers defended the supremacy of the dominant classes
in less radical and more specically Greco-Roman ways. In particular, the
doctrine of the ‘mixed constitution’ – which appears in Thucydides and in
Plato’s Laws and gures prominently in the writings of Aristotle, Polybius,
and Cicero – can be said to reect a uniquely Greek and Roman reality and
the special problems faced by a dominant class of private proprietors in a
state that incorporates rich and poor, appropriators and producers, into a
single civic and military community. In the idea of the ‘mixed constitution’,
rich and poor are respectively represented by ‘oligarchic’ and ‘democratic’
elements; and the predominance of the rich is achieved not by drawing a
clear and rigid division between a ruling apparatus and subject producers,
or between military and farming classes, but by tilting the constitutional bal-
ance toward oligarchic elements.79 The Roman Republic, with its aristocratic
senatorial government presiding over a free community of Roman citizens
consisting in large part of peasants – SPQR80 – can be said to represent the
77
Aristotle praises the separation of farming and ghting classes, which he attrib-
utes to Egypt and Crete: Politics, 1329 a–b.
78
See Wood and Wood 1978, pp. 168–9, for an argument suggesting that Egypt may
have served as a model for Plato’s Republic.
79
I owe this suggestion on the mixed constitution to Neal Wood. Also, see his dis-
cussion of Aristotle’s mixed ‘polity’ in Wood and Wood, pp. 243–4.
80
Senatus Populusque Romanus, the Senate and the Roman People, the emblem
of the Roman Republic.
‘mixed’ constitution in its most successful practical form. In both theory and
practice, then, the specic and differentiated dynamic of property and class
relations was woven directly into the fabric of Greek and Roman states in
unprecedented ways and degrees.
The developmental logic which Dockès seeks in slave society – the contra-
dictory logic which demands state centralisation but is, at the same time,
inimical to it – can be said to inhere not just in slavery but in the very nature
of private property and class. Private property and class exploitation require
coercive power to sustain them; and the appropriating powers of the indi-
vidual lord always depend in various ways and degrees on a collective
class power. Direct producers, even when exploited individually, never
confront their exploiters solely as individuals. Even peasant proprietors
who are relatively isolated in production tend to be organised in communal
groups, especially in village communities.81 Appropriators must nd ways of
counteracting the divisions within their own class, the intraclass conict which
results from private property and competition over land and limited sources
of surplus labour. It can also be argued that the balance of power between
appropriators and producers may be less one-sided in favour of the former
when petty producers are confronted by private appropriators, divided and
competing among themselves, rather than by a centralised ‘public’ appro-
priator. There is, therefore, always a tendency toward centralisation which
will permit individual exploiters to withstand resistance by producers and
to maintain their hold on property.
That tendency, however, is accompanied by countervailing forces. The
resistance of producers may itself act as a force against centralisation, as may
the intraclass conict within the ruling class. More particularly, to the extent
that the dominant class is not directly organised as an appropriating state –
in other words, to the extent that class and state are not co-extensive – they
will represent two separate and often competing powers. Until the advent of
capitalism, in which appropriators can rely on ‘economic’ modes of surplus
extraction which depend not on the coercive extraction of surplus but on
increasing the productivity of labour, the dominant class and the state must
81
See Brenner 1976, pp. 56–60, for an example of how village organisation can
function as a kind of peasant class organisation and affect the relationship between
confront each other, in varying degrees, as competing ‘extra-economic’ powers

of appropriation. Both landlords and state must rely on the application of
direct force to extract surplus from the same limited source, the same peasant
producers, one in the form of rent, the other in the form of tax (in this
context, a kind of centralised rent).
On the one hand, then, the antagonisms inherent in any exploitative
relationship may be quite enough to produce a compulsion toward political
centralisation. Where the state itself is the major direct appropriator, it is,
by denition, highly centralised. Where appropriation is based on private
property, the conditions for the maintenance of private property are likely
to include political centralisation. A need for political centralisation is not
exclusive to cases where there is a tendency toward large-scale production
units. On the other hand, the antagonisms and contradictions inherent in
private property, inter- and intraclass conict, and the antagonisms between
ruling class and state, are enough to produce a countervailing tendency toward
political fragmentation.
There is, then, a constant tension between political centralisation and
fragmentation, centripetal and centrifugal forces, inherent in the very nature
of private property and class relations; and the specic ‘logic’ of private
property and class operated, as it were, more purely and completely in Greece
and Rome than anywhere else in the ancient world. The rise and decline of
Greek and Roman states, and the rise and decline of slavery, can best be
understood within the framework of that logic (it is not a trivial point to
note, in the rst place, that chattel slavery, by denition, presupposes private
property), even while it is acknowledged that slavery had a logic of its own
which helped to shape the way that more inclusive logic worked itself out.
The rise and decline of the Roman Empire cannot be understood without
identifying the conditions that alternately promoted or obstructed one or
the other of the contradictory tendencies toward political centralisation and
fragmentation inherent in the relations between landlords and peasants.
Factors that have traditionally been cited in explaining the decline – a fall
in the slave supply, ‘barbarian’ invasions, etc. – were either effects as much
as causes, or had their particular effects because of these larger, ‘structural’
tendencies. For example, while barbarian raids were a recurring theme of
Roman imperial history, they became disastrous only when the Empire’s
inherent tendency toward disintegration was well advanced.
Even at the height of its centralisation, the Roman imperial state contained
the seeds of its own fragmentation. The burden of centralisation fell on
the peasantry whose ability to support it was limited. At the same time,
imperial ‘absolutism’ had developed in large part to counteract erce com-
petition and a self-defeating political ‘individualism’ within the aristocracy,
so that the state’s foundation in the ruling class was always fragile and
contradictory.82 More particularly, one might say that parcellisation was,
paradoxically, the very essence of Roman imperial administration. The
Romans (like the Hellenistic monarchies before them) undoubtedly borrowed
methods of administration from the bureaucratic traditions of the Eastern
kingdoms and to some extent allowed the old modes of political organisa-
tion and state-economy to survive in Egypt and the East. Nevertheless,
there had evolved a new pattern of imperial rule unique to the Greco-Roman
world and better suited to its class relations and to the level of development
of private property. The Romans (like the Hellenistic rulers) dominated
their empire largely by means of a ‘municipal’ system in which vast ter-
ritories and heterogeneous populations were administered through local
community organisations, with varying degrees and kinds of civic institu-
tions, local magistrates, councils, and assemblies. In a sense, the old Greek
polis-organisation – with all its municipal particularism – had been adopted
as a mode of imperial administration, imposed not only by making use of
already established cities but often by the establishment of new cities and
more particularly by the creation of rural ‘municipalities’. These imperial
municipalities were not, however, a means of establishing democratic self-
government, but on the contrary, a means of strengthening or even creating
Romanised local aristocracies (often supplemented by Roman senators) through
which the Roman state ruled. In fact, the Roman Empire was, to a great extent,
a confederation of local landed aristocracies.
The old ‘redistributive’ empires had typically ruled by means of a bureau-
cratic hierarchy descending from the monarch – in principle, owner of
land – to administrative districts governed by royal functionaries and s-
cal ofcials who extracted surplus labour from subject villages of peasant
82
The notion that the ‘absolute monarchy’ in Rome was the product of a kind of
intraclass conict is suggested, for example, by Matthias Gelzer in his classic studies
of the Roman nobility. See Seager 1975, p. 139.
producers.83 A new type of imperial hierarchy was created in the Hellenistic

state, more in keeping with the distinctive relations between property, class,
and state which had evolved in archaic and classical times. This hierarchy
descended from the monarch to the city, often the possessor of free land (the
monarch frequently transferred royal property to the polis) and dominated
by a local aristocracy of private landlords, who often had land grants from
the monarch. The latter mode of imperial rule was perfected by the Romans.
It was a form of empire especially well-suited to reect and enhance the
dominance of private property, both in Rome and in the provincial periphery.
In this, the Roman Empire was quite unlike the bureaucratic, redistributive
states whose very existence obstructed the full and autonomous development
of private property or a propertied class independent of the imperial bureau-
cracy. The Roman Empire had the effect of strengthening private property.
Thus, there was, in Rome, a unique synthesis of a strong imperial state
and strong private property, two distinct foci of power, based on a delicate
balance between a centralised state and a parcellised municipal rule.
The tendency toward fragmentation was thus built into the very structure
of the Roman state, not only in the particularism and divisiveness of private
property on which the state was based, but also in its parcellised mode of
administration. In the end, the tendencies toward fragmentation prevailed.
The disintegrating state had long since become an intolerable burden to
peasants and a dispensable nuisance to landlords. It left behind a network
of personal dependence binding peasants to landlord and land. The state
itself had helped to create this network when, in a period of crisis, measures
were enacted, most probably for scal purposes, tying many tenants to the
land and in the process creating a captive agricultural labour-force for the
landlords. No doubt it is also true that peasants often accepted dependence
on landlords in order to obtain protection – including protection from oppres-
sion by the state. At any rate, the tied colonate was a mode of exploitation
with clear advantages for the landlord; and it gradually absorbed slavery
itself, as many slaves were placed in possession of land to be exploited as
83
For a discussion of the new form of imperial administration which resulted
from ‘the meeting of the imperial idea with the form of the polis’ in the Hellenistic
state, and the inuence of this ‘meeting’ on the later form of the Roman Empire, see
Ehrenberg 1969, Part II, Chapters 2–3.
coloni or serfs. To explain the transition from antiquity to feudalism is to

explain the unique conditions in which, after several attempts by dynastic
monarchies to recentralise this fundamentally fragmented polity, landlords
were able for some time to maintain this parcellised power and to subject
peasants to personal dependence without the support of a centralised state.
References
Anderson, Perry 1974, Passages from Antiquity to Feudalism, London: New Left Books.
Aston, T.H. and C.H.E. Philpin (eds.) 1985, The Brenner Debate: Agrarian Class Structure
and Economic Development in Pre-Industrial Europe, Cambridge: Cambridge University
Press.
Brenner, Robert 1976, ‘Agrarian Class Structure and Economic Development in

Pre-Industrial Europe’, Past and Present 70: 30–75.
Brenner, Robert 1977, ‘The Origins of Capitalist Development’, New Left Review, I, 104:
25–92.
Brenner, Robert 1982, ‘The Agrarian Roots of European Capitalism’, Past and Present,
97: 16–113.
Dockès, Pierre 1982, Medieval Slavery and Liberation, London: Methuen.
Duncan-Jones, Richard 1974, The Economy of the Roman Empire, Cambridge: Cambridge
University Press.
Ehrenberg, Victor 1969, The Greek State, Second Edition, Oxford: Blackwell.
Finley, Moses I. 1973, The Ancient Economy, Berkeley: University of California Press.
Finley, Moses I. 1980, Ancient Slavery and Modern Ideology, London: Chatto and Winders.
Hilton, Rodney 1973, Bond Men Made Free, London: Temple Smith.
Hopkins, Keith 1978, Conquerors and Slaves, Cambridge: Cambridge University Press.
Jameson, Michael 1977, ‘Agriculture and Slavery in Classical Athens’, Classical Journal,
73: 122–41.
Jones, A.H.M. 1973, The Late Roman Empire, Oxford: Blackwell.
Marx, Karl 1971, Capital, Volume III, Moscow: Progress Publishers.
Seager, Robin 1975, The Roman Nobility, Oxford: Blackwell.

Ste Croix, Geoffrey de 1981, The Class Struggle in the Ancient Greek World, Ithaca: Cornell
University Press.
Thompson, Edward P. 1968, The Making of the English Working Class, Harmondsworth:
Penguin.
White, K.D. 1970, Roman Farming, London: Thames and Hudson.
Wood, Ellen Meiksins 1981a, ‘Marxism and Ancient Greece’, History Workshop Journal
11: 16–20.
Wood, Ellen Meiksins 1981b, ‘The Separation of the Economic and the Political in
Capitalism’, New Left Review, I, 127: 66–95.
Wood, Ellen Meiksins 1983, ‘Agricultural Slavery in Classical Athens’, American Journal
of Ancient History, 8, 1: 1–47.
Wood, Ellen Meiksins 1988, Peasant-Citizen and Slave: The Foundations of Athenian
Democracy, London: Verso.
Wood, Ellen Meiksins 1995, Democracy Against Capitalism: Renewing Historical Materialism,
Cambridge: Cambridge University Press.
Wood, Ellen Meiksins and Neal Wood 1978, Class Ideology and Ancient Political Theory,
Oxford: Blackwell.

Wood EM - Landlords and Peasants, Masters and Slaves - Class Relations in Greek and Roman A

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Wood EM - Landlords and Peasants, Masters and Slaves - Class Relations in Greek and Roman A

Uploaded by

Copyright:

Available Formats

Ellen Meiksins Wood

Landlords and Peasants, Masters and Slaves:

Few historians would hesitate to single out slavery

Historical Materialism, volume 10:3 (17–69)

economies, and, in particular, the degree to which production rested on

Master and slave vs. landlord and peasant

of personal dependence – slavery, serfdom, debt-bondage. Thus, both the

by slave-gangs, made possible by enormous accumulations of property, never

those between propertied classes and free producers. If primacy must be

alongside a local labour-force inadequate to the demands of the ‘incipient

the existing bureaucracy and became pensioners or sinecurists of the tax

expropriation possible. ‘Roman peasant soldiers,’ as Keith Hopkins puts

‘Slave society’, ‘slave economy’, ‘slave mode of production’

human history): ‘. . . an institutionalised system of large-scale employment of

divisions paled before the more essential opposition between citizens, or

became general in the Roman Empire’.23 Consequently, he continues, ‘. . . it

Slavery as the key to Greco-Roman history?

In Finley’s case, slavery is not the ‘primordial fact’ in relation to which

both as an interpretation of Marx and as a principle of historical explana-

the direct producers. The existence or non-existence of exploitation, the

means and institutions to sustain and enforce it and to overcome, suppress,

an analogy between slaves in Greece and immigrant workers in modern

Slavery: ‘primordial fact’ or historical oddity?

Underlying this proposition is a more fundamental and universal rule that

dependent labourers and made necessary other forms of surplus extraction.

by a decrease in the slave supply, an accompanying decline in the prot-

Even if all exploiters could be shown in principle to be striving for slavery,

Slavery and the ‘decline’ of the Roman Empire

Viewed in this perspective, the rise and decline of Greco-Roman civilisation

inclusive factors which can explain the structural developments within

or against the efciency of slavery), especially principles that probably

The historical evolution or emergence of any given class structure is

The decline of slavery, then, cannot be explained solely by reference to its

be bystanders rather than agents of the transition to dependence from

The whole analogy between slavery and capitalism is especially problematic,

Italy and the ‘two contrasting tendencies in large-scale Italian landownership’:

here again be imposing the logic of capitalism and capitalist competition

The speci c dynamics of class

polis of Athens, he did so by opposing to it a state-form which departed

confront each other, in varying degrees, as competing ‘extra-economic’ powers

producers.83 A new type of imperial hierarchy was created in the Hellenistic

coloni or serfs. To explain the transition from antiquity to feudalism is to

Brenner, Robert 1976, ‘Agrarian Class Structure and Economic Development in

Dockès, Pierre 1982, Medieval Slavery and Liberation, London: Methuen.

Jones, A.H.M. 1973, The Late Roman Empire, Oxford: Blackwell.

Marx, Karl 1971, Capital, Volume III, Moscow: Progress Publishers.

Seager, Robin 1975, The Roman Nobility, Oxford: Blackwell.

White, K.D. 1970, Roman Farming, London: Thames and Hudson.

You might also like

by a decrease in the slave supply, an accompanying decline in the prot-

or against the efciency of slavery), especially principles that probably

The speci c dynamics of class