Edo's Importance in the Changing Tokugawa Society

Author(s): Gilbert Rozman

Source: The Journal of Japanese Studies , Autumn, 1974, Vol. 1, No. 1 (Autumn, 1974),
pp. 91-112
Published by: The Society for Japanese Studies

Stable URL: http://www.jstor.com/stable/133438

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

is collaborating with JSTOR to digitize, preserve and extend access to The Journal of Japanese
Studies

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
GILBERT ROZMAN

Edo's Importance in the

Changing Tokugawa Society

By 1800 Edo ranked as one of roughly 70 cities in the world (five

were in Japan) with more than 100,000 residents, as one of about
20 cities (three in Japan) in excess of 300,000 population, and as
probably the only city easily to surpass 1,000,000 inhabitants. By
1970 Japan was one of a small number of countries that started the
modernization process as latecomers yet rushed through a complex
transition to become highly modernized societies. What is the con-
nection between a country's premodern urban heritage and its pace
of modernization? By raising some general questions about Edo in
comparative perspective, I will examine in this paper aspects of this
city's impact on the development of strategic factors favorable to
subsequent modernization.
A single, large city represents many things to social scientists: a
separate spatial entity with distinctive features, a component con-
tributing various functions to a greater societal network of cities,
a case designated for comparison with cities chosen from other
societies and, in potential accord with each of these viewpoints, either
a mirror on which broad lines of change are vividly reflected or a
breeding ground for the dissemination of change to some wider area.
The city of Edo commands attention from each of these perspectives.
First, as the most populous city in Japan for slightly longer than the
final two centuries of Tokugawa rule, Edo glitters in the spotlight
thrown by Japanese historians on many features of their compact
country's extraordinary urban past. Second only to Kyoto, which was
unsurpassed for a more than 1,000 year tenure as one of the world's
great cities, Edo intrigues Japanese scholars, who in recent years have
continued to add to the unusually well-documented history of the
city's various social strata and its successive phases of growth.'

1. For information on Edo's chonin see Nishiyama Matsunosuke, ed.,

91

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
92 Journal of Japanese Studies

Second, as the crown atop Japan's highly efficient pyramidal

structure of 1,700 or more central places (administrative and market-
ing centers) during the second half of the Tokugawa period, Edo
occupies a crucial position in analyses of networks of cities.2 Japanese
historians have partially documented the flow of goods between hun-
dreds of small marketing centers, many of the more than 200 castle
cities (Q5kamachi), some tens of ports and post stations with regional
significance, and the three principal cities. Inferior, perhaps, only to
the detailed local histories compiled for English urban history, these
materials assembled on large numbers of Japanese cities and their
interrelationships represent one of our richest resources in the analysis
of premodern urban networks.
Third, as the premier city in a country which later succeeded in
rapid modernization, Edo would seemingly attract comparisons with
the main cities in other countries alleged to have possessed unusually
favorable preconditions for modernization. Unfortunately, little is
yet known about the roots of modernization, especially among the
countries which were latecomers to the process. Recognition remains
long overdue of the utility of comparisons between cities, including
such major centers as London and Paris in the early modernizing
countries and St. Petersburg in Russia, another rapidly modernizing
latecomer.
In the absence of direct comparison, articles by Thomas C. Smith
and E. A. Wrigley in Past and Present indirectly pose the most com-
pelling challenge to consider in new comparative ways Edo's signifi-
cance for changes conducive to subsequent modernization in Japan.
On the one hand, Smith's stimulating work, "Pre-modern Economic
Growth: Japan and the West," cannot but leave doubts as to whether,
in comparison to London or Paris, Edo made any major contribution
at all to the premodern phase of economic growth, since Smith boldly
contrasts Japanese rural-centered to European urban-centered devel-

Edo chonin no kenkyui (Studies of the Edo ch6nin), vols. 1-3 (Yoshikawa
Kobunkan, 1972-74); and Minami Kazuo, Edo no shakai kozo (The social
structure of Edo) (Hanawa Shob6, 1969). On Edo's hatamoto see Kozo
Yamamura, A Study of Samurai Income and Entrepreneurship (Cambridge:
Harvard University Press, 1974); and on other categories of bushi see the
bibliography in Yamamura. For histories of Edo's growth see Naito Akira,
Edo to Edojo (Edo and Edo castle) (Kajima Kenkyfij5, 1966); Nomura
Kentaro, Edo (Shibundo, 1966); Ikeda Yasaburo, Hiroshige no Edo (Kodan-
sha, 1968); and Kawasaki Fusagoro, Edo happyaku hachi ch5 (The 808 wards
of Edo) (Togensha, 1967).
2. Gilbert Rozman, Urban Networks in Ch'ing China and Tokugawa
Japan (Princeton: Princeton University Press, 1973).

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
 Rozman: Edo's Importance 93

opment.3 On the other hand, Wrigley's arti

London's Importance in Changing English Society and Economy
1650-1750," proposes a model for assessing London's considerable
importance in English society with obvious (though unstated) appli-
cations for studying Edo's importance in Japanese society.4 Despite
initial impressions, interpretations of the two views need not be en-
tirely inconsistent. If carefully interpreted, Smith's contrast implies
two stages of Japanese development, with urban-centered growth
during the early Tokugawa period giving rise to later rural-centered
growth. And Wrigley's interest really centers on London's long-run
impact on rural England, not excluding a comparable two-stage pro-
cess. Neither author draws comparisons between Edo and London;
nonetheless, this is the obvious comparison which emerges from a
juxtaposition of their conclusions.
Wrigley suggests that London's relationship with the rest of
England can be conceived as that of "a potent engine working toward
change."5 He begins with population data, noting that England's
capital grew rapidly from roughly 200,000 in 1600 to 575,000 in
1700 (by which time it had become the largest city in Europe) to
as many as 675,000 inhabitants in 1750. Students of Japan have long
been aware that in absolute growth Edo easily outdistanced London
during this period. In 1590, when London already ranked as a great
city of Europe with 200,000 inhabitants, Edo was just being founded
on the site of a small settlement distinguished only by a branch castle
and an insignificant market.6 Within 130 years, as London's popula-
tion increased by 400,000 or 450,000, Edo added more than one
million residents.7 Indeed, for almost 100 years until London's grow
gained momentum in the early nineteenth century, Edo probably

3. Thomas C. Smith, "Pre-modern Economic Growth: Japan and the

West," Past and Present 43 (1973):127-60.
4. E. A. Wrigley, "A Simple Model of London's Importance in Changing
English Society and Economy 1650-1750," Past and Present 37 (1967):44-70.
5. Ibid., p. 70.
6. A recent account of Edo's early history can be found in Mizue Renko,
"Shoki Edo ch6nin" (The chonin of early Edo), in Nishiyama, ed., Edo
chonin no kenkyft, 1:43-124.
7. Estimates of Edo's total population can be found in many sources,
including Nait6 Akira, Edo to Edoj p pp. 124-42; and Nomura Kentar6, Edo,
pp. 103-12. These estimates indicate growth from about 100,000 in 1610- to
roughly 400,000 in the 1640's to as many as 800,000 in the 1680's and finally
to over 1,000,000 persons by the 1720's. More recently Nait6 Akira has esti-
mated the city's peak population at 1.3 to 1.4 million, including 650,000
persons on bushi estates. See his book, Edo no toshi to kenchika (The city and
architecture of Edo) (Mainichi Shinbunsha, 1972), pp. 23-25.

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
94 Journal of Japanese Studies

ranked as the largest city in the world, rivaled only by one or two
cities in far more populous China. Under these circumstances, are
we not justified in asking whether Edo did not likewise operate as
44a potent engine working toward change?" Should we not both ob-
serve the reflection in Edo of nationwide currents of change and
examine the central role of this city in initiating changes which spread
across Japan?
Wrigley's model links demographic conditions, marketing, and
new forms of social mobility and consumption to the growth of Lon-
don and the establishment of a foundation for the industrial revolu-
tion. It credits London with exerting a pervasive impact on the rest
of England, largely through new patterns in the movement of people
and goods centering on this city. Before examining whether new pat-
terns of migration and marketing into Edo could have exerted a
similar impact, we shall first want to establish that changes similar to
those listed by Wrigley for English society prior to 1750 were also
occurring in Tokugawa society. Then we shall briefly consider the
prevailing image of Edo, showing how it differs from that of London.
After looking in some detail at the evidence for recasting Edo's
image, we should emerge in a position to draw preliminary conclu-
sions about Edo's impact on Japan in comparison to the impact of
other great cities on their premodern societies.
Included among Wrigley's checklist of ten changes which the
growth of London may have promoted and which by their occurrence
may have succeeded in engendering the magic ."take-off" are: 1) the
fostering of changes in agricultural methods which increase the pro-
ductivity of those engaged in agriculture so that the cost of foodstuffs
will fall and real wages rise; 2) the interplay between fertility, mortal-
ity, and nuptiality such that population does not expand too rapidly
for some time after real income per head has begun to trend upwards;
3) the steady spread of environments in which the socialization pro-
cess produces individuals with different orientations in their patterns
of action; 4) the establishment of conditions in which upward social
mobility need not necessarily lead to the recirculation of ability within
traditional society; and 5) the spread of the practice of aping one's
betters.8 While emphasizing that "it is not so much that London's

8. Also on Wrigley's checklist are: 6) the creation of a single national

market; 7) the development of new sources of raw materials; 8) the provi-
sion of a wider range of commercial and credit facilities; 9) the creation of
a better transportation network; and lO)the securing of a steady rise in real
incomes. Of these changes, all stated here in a simplified form, Japan may only
have lacked the pronounced development of new sources of raw materials such

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
 Rozman: Edo's Importance 95

growth was independently more important than the other major

changes which modified English economy and society during the
century, as that it is a most convenient point of entry into the study
of the whole range of changes which took place ," Wrigley
presents a model for exploring the impact of a great premodern city
on its wider society.9
Reference to many of the same social changes that Wrigley
identifies has surfaced during the past two decades in reappraisals
of Tokugawa history, most recently in the sweeping challenge posed
to the lore of past studies by Susan Hanley and Kozo Yamamura.
Together these two social scientists have offered a string of fresh
insights in applying the relatively precise tools of demographic and
economic history to the study of the Tokugawa period, pointing
specifically to similarities between Japan and England in slow popu-
lation growth and rising per capita income.'0 The impressive docu-
mentation by Japanese scholars and by Ronald Dore and Thomas C.
Smith of dramatic rises in literacy, non-agrcultural by-employments
and labor mobility within rural areas also leaves little doubt that
Wrigley's checklist could be reproduced virtually intact for Tokugawa
Japan.'" The similar general pattern of social transformation is no
longer problematic; only the basic explanations for change vary from
one country to the other. Curiously, unlike the overwhelming empha-
sis placed on London in explanations of the dynamism of English
society, Edo is still virtually ignored in interpretations of major social
changes within Japan.
The image of Edo conveyed in writings on the Tokugawa period

as coal, increasingly mined in England before 1750. See Wrigley, "A Simple
Model of London's Importance," pp. 65-67.
9. Ibid., p. 65.
10. Among the numerous articles authored individually or jointly by
Susan Hanley and Kozo Yamamura are: "Population Trends and Economic
Growth in Pre-industrial Japan," in D. V. Glass and Roger Revelle, eds.,
Population and Social Change (London: Edward Arnold, 1972), pp. 451-99;
"Toward an Analysis of Demographic and Economic Change in Tokugawa
Japan: A Village Study," The Journal of Asian Studies 31 (May 1972):
515-37; and "Toward a Reexamination of the Economic History of Tokugawa
Japan, 1600-1867," The Journal of Economic History 33 (September 1973):
509-41. The authors point specifically to similarities between England and
Japan in the first of these articles on pp. 451 and 485 and in the second
article on p. 536.
11. R. P. Dore, Education in Tokugawa Japan (Berkeley: University of
California Press, 1965); and Thomas C. Smith, "Farm Family By-employ-
ments in Preindustrial Japan," The Journal of Economic History 29 (December
1969): 687-715.

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
96 Journal of Japanese Studies

is generally more that of a caboose than of a locomotive. Unlike

London's towering primacy in the English context, Edo appears as a
late addition coupled on at the rear of a progression of Japanese
cities. During the Tokugawa period, innovation in the sphere of pro-
duction and distribution is credited primarily to the Kinai area within
the Kinki region, not to the Kanto region in which Edo was located.
Osaka is commonly identified as the center of a prosperous national
market, while Edo's economic importance is typically relegated to
the sphere of consumption, a contribution often ignored or even
regarded as parasitic.'2 In the mistaken impression that cities can be
neatly classified according to a single primary function, Osaka is
labeled an economic city and Edo a political city. From this per-
spective, Edo appears to have coasted along, fueled by the momentum
gained through resources mobilized in Osaka and the various castle
cities.
Mounting evidence that in critical respects Tokugawa society re-
sembled English society some 100 years earlier, of course, does not
establish that Edo's primacy should be equated with that of London.
In at least one obvious way Edo's position in Japan differed consid-
erably from London's position in England. After all, London con-
tained seven per cent of England's population in 1650 and about
10-11 per cent from the late seventeenth century through the first
half of the eighteenth century, while Edo maintained a comparativel
meager three per cent of Japan's population during the second half
of the Tokugawa period. Yet, this difference should be attributed,
above all, to the small scale of England and Wales with a total of
just 5.5 million persons in the late seventeenth century. London re-
mained the only city in England until after 1700 with as many-as
30,000 inhabitants, while Edo was one of 20 to 25 Japanese cities
in excess of this minimum, which together encompassed some 8-9
per cent of the national population. While we might conclude from
these figures that it would be useful to compare a number of Japanese
cities together to London (and for some purposes it undoubtedly
would be), there are many respects in which Edo merits comparison
as a direct counterpart to London. As the national administrative
center, as one of two fairly equal centers of a national market, and
as the point into which flowed Japan's numerous circulating elite,

12. See, for example, George Sansom, A History of Japan 1615-1867

(Stanford: Stanford University Press, 1963). Among the many writings in
Japanese which convey a negative impression of Edo's economic importance
versus that of Osaka is Matajima Masamoto, Edo jidai (Iwanami Shoten,
1966), pp. 106-18.

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
 Rozman: Edo's Importance 97

Edo's influence was carried down through the various levels of the
urban hierarchy. At the very least, we should examine the possibility
that Edo also provides a "convenient point of entry into the study
of the whole range of changes which took place."''3
In the absence of studies focusing on Edo's primacy in generating
social change, we are confronted with incomplete interpretations of
the mechanisms of social change during the first half of the Tokugawa
period and, more seriously, with one-sided impressions of the nature
of social change during the second half of this period. By now there
is general agreement that the seventeenth century was characterized
by unprecedented rapid urbanization. What is less clear is how this
urban growth transpired. What made possible a sudden four- or
five-fold jump in the urban population of Japan and a more than
doubling of the urban percentage? Comparisons with other premodern
societies suggest that three basic factors were involved: 1) the prior
existence by the late sixteenth century of a solid foundation of com-
mercial exchange; 2) the deliberate restructuring of administrative
practices and settlement patterns to maximize urban concentrations;
and 3) the promotion of new growth mechanisms which stimulated
the continued mobilization of increasing amounts of resources into
cities. By design, Edo became the center in the restructuring process
and at the same time an heir to the commercial legacy of the past.
A case will be made below that, in a less deliberate manner, Edo
also became the main source for the new mechanisms of resource
accumulation which pervaded all of Japan.
During the seventeenth century Edo's special significance for
overall urban growth in Japan stemmed above all from new patterns
of elite migration associated with the sankin k3tai system of alternate
residence and from the continually rising demand for goods and reve-
nues to meet the responsibilities commensurate with each elite position
within the city's finely stratified population. Unlike the sixteenth cen-
tury efforts at urban consolidation promoted by local lords,'4 impulses
emanating from the top of the urban hierarchy now produced waves
of urban growth in successively lower level cities. Daimyo transformed
their castle cities to support new mobilization of local resources to
meet expenses in Edo. Centralization in Edo spurred increased accu-
mulation and production also in Osaka and Kyoto and in smaller
cities, reaching eventually to the local commercial nexes subordinate

13. Wrigley, "A Simple Model of London's Importance," p. 65.

14. On sixteenth century cities see Nakabe Yoshiko, Kinsei toshi no seiritsu
to koz5 (Establishment and plan of Tokugawa cities) (Shinseisha, 1967),
part 1.

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
98 Journal of Japanese Studies

to castle cities. Lending support to this emphasis in t

during the first half of the Tokugawa period are the data assembled
in my book on Ch'ing and Tokugawa cities showing that as much as
one-fourth of all urban growth between 1590 and 1720 was accounted
for by the growth of Edo.'5
While rural Japan shared in the dynamism with respect to popula-
tion growth, to commercial specialization, and to social differentiation
in the seventeenth century, it was overshadowed by the unparalleled
transformation within the urban sector. With Edo, Osaka, and Kyoto
leading the way, new patterns of consumption and new modes of social
organization developed in Japanese cities. If any period of premodern
history anywhere can properly be labeled urban-centered, it is this
period from about 1600 to the 1720's in Japan.
The justification for Edo's centrality becomes less obvious after
the first quarter of the eighteenth century, when Japan's overall urban
growth stopped, while, as various studies have shown, the rural sector
increasingly exhibited unusual dynamism in the proliferation of com-
mercial orientations. Indeed, Japanese cities have long taken a back-
seat to villages in studies of the second half of the Tokugawa period;
and this trend has culminated recently in Thomas Smith's hypothesis
that this phase of Japan's premodern economic growth was rural-
centered. Yet, given the continuing high percentage of Japanese in
cities (about 16 or 17 per cent if the population in all marketing
centers with more than 3,000 residents and one-half the population
in smaller centers with intermediate markets are included, not so far
below the 20 or 21 per cent recorded in England for the 1680's and,
perhaps, slightly higher than the percentage in France before 1789) ,16
the striking rural changes ought not to be seen in isolation from closely
connected urban phenomena. Increasing village real incomes, literacy
rates, and commercial orientation corresponded to similar changes in
cities. Where the proximity to large cities was greatest and the urban
presence was most pervasive, as near Edo, rural dynamism reached
its peak, evidenced by dramatic declines in population, the sudden
proliferation of small-scale industries, and the gradual disappearance
and replacement of periodic markets by village stores and by daily
commerce in major cities. These changes penetrated most thoroughly
in areas in close contact with the consequences of the earlier urban-

15. Gilbert Rozman, Urban Networks in Ch'ing China and Tokugawa

Japan, pp. 285-88.
16. See chapter 5, "Urban Networks of Stage G Societies," in Gilbert
Rozman, Urban Networks in Russia, 1750-1800, and Premodern Periodization
(Princeton: Princeton University Press, forthcoming).

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
 Rozman: Edo's Importance 99

centered phase of societal development and with the continued re-

organization of life in the cities.
Acknowledgment of a perceptible shift in the balance of social
dynamism to the rural sector is not inconsistent with emphasis on
Edo's continued importance in eliciting change. English and French
urban data on the eve of early modernization likewise reveal little
increase in the percentage of the population in cities. Moreover,
London between 1670 and 1750 and Paris during the century before
the French Revolution seem to have maintained a fairly constant
percentage of their national populations, as did Edo in the second
half of the Tokugawa period. Even with a population probably rising
quite slowly after 1720, Edo could easily have continued to be in a
position in no way inferior to that of these other cities to act as the
principal source of changes which spread across rural areas.
The significance of the turnabout in the fortunes of Osaka and
Kyoto by the middle of the eighteenth century was to make more
direct Edo's impact over much of Japan. As the combined population
of these two neighboring cities within the Kinki area declined from
roughly 900,000 to the range of 600,000 to 700,000, Edo's predomi-
nance over over Japanese cities strengthened and its sources of supply
widened to encompass expanded rural production of non-agricultural
goods. The continued economic implications of a city that was ex-
tremely populous and sheltered expensive tastes now reached rural
Japan in full force.
If Edo's influence paralleled that of London, then the mechanisms
by which it penetrated into rural areas might presumably have been
the same as those identified for London. Did Edo exert a powerful
impact over much of Japan through the mechanisms of migration and
marketing? How did the character of these mechanisms vary from the
first half to the second half of the Tokugawa period? Answers in
greater detail to these questions are essential for reassessing Edo's
image.
I offer some preliminary approximations on migration into Edo
in the hope of stimulating more careful and complete studies. First,
since the population of Edo rose by about one million in 130 years
from 1590 to 1720, it will on the average have been increasing
annually by nearly 8,000. It seems likely that this growth was spaced
quite evenly during the 130 years, which means that the percentage
increase continually declined. Second, similar to London and other
large cities, the crude death rate in Edo was probably substantially
higher than the crude birth rate over the period as a whole. In the
absence of direct information, we may assume for the moment that

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
100 Journal of Japanese Studies

the gap between the two rates was roughly as in London and held
throughout the period. Accordingly, if the difference between the two
rates had fluctuated around 10 per 1,000 per annum, then at the time
when the population of Edo was one million, the shortfall of births
each year is assumed to have been 10,000. To make good this short-
fall and to permit an annual increase of the total population of 8,000,
the net immigration into Edo must have fluctuated around 18,000 per
annum. As Edo's population rose between 1590 and 1720, the net
immigration figure must also have been generally rising, with the city
at first requiring fewer persons to replace its losses through the excess
of deaths over births. Based on these assumptions, we should envision
a net immigration of about 10,000 persons during the first half of the
seventeenth century, 15,000 persons in the second half of that century,
and about 18,000 persons by the early eighteenth century. After the
relative stabilization of Edo's population at 1.0 to 1.2 million, a net
immigration figure of about 10,000! would again have been sufficient.
Two types of migration into Edo can be distinguished. Most widely
noted has been the elite inter-city mobility of daimyo, rejoining mem-
bers of their families left in Edo and accompanied by large entourages
of their samurai retainers often travelling with their families.17 This
closely regulated annual migration under the sankin k5tai system
involved increasing numbers of persons during the seventeenth cen-
tury, but its growth must have virtually stopped by the early eighteenth
century. Excluding the stationary hatamoto and gokenin, who together
with their families comprised perhaps ten per cent of Edo's popula-
tion, this assembled elite continuously replenished from f5kamachi
throughout Japan constituted roughly 25 to 30 per cent of the city's
population.
It would be difficult to exaggerate the importance of this regular
circulation of Japan's elite population between local f5kamachi and
Edo. In my opinion, this system of alternate residence ranks as the
single greatest accomplishment of Japanese leaders precisely because
it built on the already considerable scale of urban and, commercial
development to accelerate the mobilization of resources at both
national and local levels. Its multiplying effects, beginning with the
demand for supporting personnel in Edo, had ramifications that
reached from city to village throughout all parts of the country carried
by the migrating messengers of change.
The estimates of Edo's population classified as bushi, which are

17. See Toshio G. Tsukahira, Feudal Control in Tokugawa Japan: The

Sankin KOtai System (Cambridge: Harvard East Asian Monographs, 1966).

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
 Rozman: Edo's Importance 101

in excess of 500,000, include a category of persons whose number

is difficult to determine. These were the servants or hired help of the
bushi. An immediate consequence of both strict prohibitions on slav-
ery and the rapidly rising bushi population, some permanently and
some temporarily in Edo, had been the replacement of hereditary
servants with hired wage labor during the seventeenth century.18 These
were hired help serving in the most menial positions and meeting
quotas set by the bakufu for a minimal complement of personnel
appropriate to each samurai office depending on rank, holdings, and
official responsibilities. Normally called hokonin, these employees
generally lived in the bushi estate complexes and served fixed terms
of service. Despite the fact that the hokonin of the bushi were not
counted as part of the ch5nin population, they were virtually indis-
tinguishable by origin from other migrants until they registered at
one of Edo's employment bureaus for new arrivals.'9 At a minimum,
some five to ten per cent of Edo's population should be classified as
hired subordinates of the bushi.
The second type of migration, which in all essential respects
should also include the hdk5nin entering bushi service, differed from
the first in bringing to the city persons mostly rural in origin, younger
on the average, more predominantly male (although the migration
of samurai also involved an unspecified surplus of males over fe-
males), and with rare exception poorer. Unlike samurai migrants,
these would-be ch5nin arrived in Edo without guaranteed incomes,
jobs, or places of residence, although not necessarily without contacts
from their native areas which could ease the transition. Renting lodg-
ings and finding work as servants, peddlars, or other hired laborers,
these migrants probably experienced high rates of mortality despite
their youth.
After the period of rapid growth had come to an end by the
1720's, Edo's enumerated chonin population (including persons liv-
ing in the separately administered jurisdiction over the city's temple
and shrine areas) hovered between 500,000 and 600,000. An addi-
tional 20,000 or so seasonal migrants regularly spent the winter in
the city, returning to home areas for the agricultural season. Available
data suggest that after an early peak in 1721 the total number of
chonin remained near the bottom of this population range through
the 1740's and, after a period of gradual growth, a new peak was

18. Sori Yoshio, "Edo no dekaseginin" (The migrant laborers of Edo),

in Nishiyama, ed., Edo chonin no kenkyfi, 3:263-308.
19. For information on the life of migrants to Edo see Minami Kazuo,
Edo no shakai kozo.

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
102 Journal of Japanese Studies

reached at the top of this range by the 1850's.20 Conceivably, un-

registered persons could have pushed the total chonin population
figures well over 600,000 and the city total over the widely accepted
1.0-1.1 million range to as high as 1.2 million.
While the size of the chlnin population appears to have fluctuated
within a narrow range during the second half of the Tokugawa period,
its changing composition provides pertinent information on migration
patterns. Enumerations of residents not born in the city show a sharp
drop from roughly one-third of the total chonin population to barely
one-quarter during the 25 years after 1843.21 Prior to 1843, only a
slight decrease in this percentage had occurred since the 1720's, al-
though given the rapid growth of the seventeenth century one might
expect that over the entire Tokugawa span figures of one-half or
more of all residents born outside the city would represent the start-
ing point for this transition. No doubt, figures for the percentage of
Edo's able-bodied inhabitants born outside the city always exceeded
these averages for all age groups.
As the proportion of chonin residents born inside the city in-
creased, the ratio of males to females approached unity. During the
1720's there were almost twice as many males as females in the
chonin population; however, by 1844 this figure had fallen sharply
until males numbered roughly 52 per cent of the total. Indeed, the
relatively complete data of 1844, 1849, and 1853 indicate that the
total male chonin population was falling and then rising slightly,
while the female population showed consistent increases of more than
6,500 over each five year interval.22 Edo's pull on rural residents
attracted an increasing proportion of females during the second half
of the Tokugawa period. By 1853 males outnumbered females among
the enumerated chOnin by barely 15,000. Despite this reduction in
the sex ratio, which ordinarily would be expected to have been accom-
panied by an increase in household size, the average size of the more
than 140,000 cho5nin households declined somewhat.23 Small families
and relatively low birth rates for a premodern population predomi-

20. Sori, "Edo no dekaseginin," pp. 296-98.

21. Takeuchi Makoto, "Kansei-Kaseiki Edo ni okeru shokaikyfi no d6ko"
(Tendencies of the various classes in Edo during the Kansei to Bunka and
Bunsei periods), in Nishiyama, ed., Edo chonin no kenkyui, 1:387-90.
22. Yoshihara Ken'ichiro, "Bakumatsuki Edo chonin no sonzai keitai"
(The state of existence of Edo ch6nin in the late Tokugawa period), in
Nishiyama, ed., Edo ch6nin no kenkyui, 1:533-34.
23. Yoshihara, "Bakumatsuki Edo ch6nin no sonzai keitai," 533-34.

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
 Rozman: Edo's Importance 103

nated throughout most of Japan,24 and must have been especially

noteworthy in Edo, where in the mid-nineteenth century the average
chO household numbered just 3.9 members.25
Just as it is possible to distinguish between categories of bushi
more or less stationary in Edo, a rough distinction can be made be-
tween houseowners and lodgers among the chinin. Data for 1828
indicate that almost all cho areas ranged between 55 and 85 per cent
of the population renting their accomodations.26 Scattered throughout
the densely settled chi sections of the city, the poor occupied back
rooms and tiny quarters. While families that owned their own resi-
dences were likely to have long been present in the city, the roughly
two-thirds of all chonin living as lodgers moved most frequently
within the city and included a disproportionate number of recent
arrivals from the countryside.
Migrants originating from certain localities had access to special
channels of entry into Edo's competitive job market. For many born
in the vicinity of Kyoto or in nearby Ise and Omi provinces, entry
into Edo came as employees of branch stores whose main offices were
located in their home territories. The flow of apprentices consisted of
young men or boys paid only in daily necessities, who after a fixed
period of service may either have returned to their home areas or have
been permitted to advance in the Edo shop as wage laborers.27 The
main stores regulated the flow of personnel as well as the supply of
retail goods to Edo, in some cases setting up several branch outlets,
each with as many as ten or more employees.
Taking into consideration this all-male labor force in branch
stores, the large numbers of males employed in construction during
the seventeenth century, the predominance of males on bushi estates,
and the general ambience favoring men in the city, Nishiyama
Matsunosuke labels Edo a male city.28 Entertainment largely catered
to male tastes, including a system of reservations at houses of prosti-
tution which provided the counterpart of modern corporation-owned

24. Susan B. Hanley, "Fertility, Mortality and Life Expectancy in Pre-

modern Japan," Population Studies 28:1 (1974).
25. Yoshihara,"Bakumatsuki Edo chonin no sonzai keitai," pp. 533-34.
26. Takeuchi, "Kansei-Kaseiki Edo ni okeru shokaikyft no dok6," pp. 391-
92. See also Ikegami Akihiko, "Koki Edo kas6 chonin no seikatsu" (The life of
lower class chonin in Edo during the late Tokugawa period), in Nishiyama,
ed., Edo chonin no kenkyui, 1:167-71.
27. Hayashi Reiko, "Edo dana no seikatsu" (Life in the shops of Edo), in
Nishiyama, ed., Edo chonin no kenkyui, 2:95-138.
28. See the introduction of Nishiyama, ed., Edo chonin no kenkyii, 1:5-13.

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:1 Jan 1976 12:34:56 UTC
All use subject to https://about.jstor.org/terms
104 Journal of Japanese Studies

season tickets parcelled among employees who take turns visiting the
facilities. It remains to be explored whether this male-oriented climate
was affected by the gradual shift in the balance of population main-
tenance away from in-migration toward natural increase.
At a time of declining village populations in the hinterlands of
big cities, Edo's continued slight population growth suggests that the
city maintained, and relative to other big cities increased, its appeal
to peasant migrants. The dynamics of migration changed markedly
from the first half of the Tokugawa period, reflected in a growing
mobile labor force finding employment outside of Edo and other
large cities. Policies to restrict entry into Edo and at times to return
recent arrivals may have contributed to, but were not the principal
cause of, this gradual realignment of labor mobility. Population move-
ment in eastern Japan was once oriented almost exclusively toward
Edo, but by the mid-nineteenth century intra-regional mobility in-
creased; as the volume of migration to the largest cities declined and
the number of migrants from village to village and to small cities in-
creased. In this respect, repeated in the histories of tens of other
Tokugawa cities, urban-centered patterns of movement bred new
rural-centered patterns as well. The details of this transformation in
migration patterns remain to be discovered. Furthermore, the problem
of explaining migration to Edo in terms of the characteristics of the
places of origin, as has been done for the major nineteenth century
Russian cities by Barbara Anderson, looms as an important task for
future study.29
While a much smaller proportion of Japanese had direct experi-
ence with life in Edo than did Englishmen with life in London, Edo's
demonstration effect took on added significance because of the elite
nature of its migrants. The daimyo and their retainers represented a
circulating elite with impressive control over resource allocation in
every local area of Japan except for areas directly administered from
Edo in which resources were mobilized without need for such an elite
through the more common procedure of dispatching officials to repre-
sent central interests. While chonin and other non-samurai throughout
Japan did not themselves exhibit a regular pattern of migration to
Edo, their efforts to control resources became closely intertwined with
this recurrent circulation of the elite. To a large extent, the popula-
tions of the castle cities engaged in accumulating goods and revenues

29. Barbara A. Anderson, "Internal Migration in a Modernizing Society:

The Case of Late Nineteenth Century European Russia," (a doctoral disserta-
tion submitted to the Department of Sociology of Princeton University, 1973).

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
 Rozman: Edo's Importance 105

for use in Edo. In this respect, the regular movement to and from
Edo of individuals with exceptional control over local resources
swelled the city's impact throughout Japan. Given the inevitable
limitations on centralization in premodern societies as large as Japan,
a circulating elite proved to be a strikingly effective device. Elsewhere
I have argued that this was a similarity between Japan and Russia,
which helps explain their subsequent speed of modernizations
While the number of individuals moving in and out of Edo each
year totaled some tens of thousands, a much larger number, certainly
reaching into the millions, participated in the production, transporta-
tion, and exchange of goods bound for Edo. Indirectly nearly every-
one in Japan contributed in some way to the sizable han revenues
and the lively national commerce which supported new habits of
consumption in Edo. In turn, behavior and attitudes reflected the
decreasing self-sufficiency and growing outside orientation. Changing
family patterns resulting from popular aspirations for a higher living
standard most likely closely corresponded to the commercialization of
rural life. Marketing, probably more than migration, broadened the
horizons of ordinary villagers.
With respect to marketing, Edo's preeminence in the transforma-
tion of Tokugawa society is not nearly as indisputable as was Lon-
don's domination over English society. Unlike London, Edo was not
the first city in Japan associated with the development of a national
marketing system, nor was it the city during most of the Tokugawa
period best known for extensive commercial and credit facilities. In-
stead, Osaka and, at least during the seventeenth century, Kyoto have
long captured attention for economic supremacy, most notably through
a huge grain market, diverse specialty products from practically every
locale, and numerous handicraft industries. The heartland of com-
mercialized agriculture within the densely settled Kinki area pumped
goods directly into these neighboring cities. Does this mean that Edo's
contribution to social change centered on migration, while Osaka and
Kyoto dominated in marketing? Separate examination of the stages of
marketing in the Tokugawa period is needed to show Edo's position
relative to these cities.
During at least the first half of the seventeenth century Edo's
impact on the production and accumulation of goods throughout all
but a few areas of Japan was mediated largely through Osaka, nearby

30. Cyril E. Black, Marius B. Jansen, Herbert S. Levine, Marion J. Levy,

Jr., Henry Rosovsky, Gilbert Rozman, Henry D. Smith, II., S. Frederick
Starr, The Modernization of Japan and Russia (New York: The Free Press,
forthcoming).

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
106 Journal of Japanese Studies

Sakai, and Kyoto.3' As late as 1630, Edo's chonin

about one-half the number in either Kyoto or Osaka. Because of the
undeveloped nature of the surrounding Kanto region, Edo relied
heavily on goods from afar and on requisitions in kind for daily neces-
sities supplied from nearby. Meanwhile, much of the rapidly growing
national exchange to support costs incurred in Edo occurred through
the direct intervention of daimyo with the close support of small
numbers of privileged merchants.
During the next century Edo established itself as a center of na-
tional marketing equal to or surpassing the Kinai complex of cities,
but its growth did not come at the expense of other cities. Rather,
Osaka, Kyoto, and many large castle cities flourished at their peak
in the first half of the eighteenth century. New shipping routes made
all areas of the country more accessible to the main cities, particularly
reinforcing ties between Edo and coastal areas along the Japan Sea
in the Tohoku and Chubu (Hokuriku) regions. Productivity rose
sharply in the areas closest to Edo. Private commerce increased
rapidly, accompanied by an expansion of marketing centers withou
administrative functions and by a growing competition among mer
chants successfully encroaching on the old privileges once monopolized
by a few.
Even in this stage of general urban prosperity, Edo's dependence
on Osaka was far less than once was thought. As William B. Hauser
observes, with the exception of certain goods which required a sophis-
ticated processing technology, Edo received little from Osaka.32 The
majority of goods imported into Osaka were consumed locally and
not reexported. Edo's market for grain and other agricultural and
forestry products from eastern Japan rivaled the separate market for
these goods in the West.
During the final century of Tokugawa rule marketing increasingly
bypassed the biggest cities. This is the stage of jimawari commerce,

31. On Edo's commerce see the writings of Hayashi Reiko, including

Edo tonya nakama no kenkyii: bakuhan taiseika no toshi shbgy5 shihon
(Studies of the Edo tonya nakama: urban commercial capital under the
Tokugawa shogunate) (Ochanomizu Shobd, 1967).
32. William B. Hauser, Economic Institutional Change in Tokugawa Japan:
Osaka and the Kinai Cotton Trade (London: Cambridge University Press,
1974), pp. 14, 30. For a more detailed treatment of the relations between Edo
and Osaka see the writings of Oishi Shinzaburb, including "KybhM kaikakuki
ni okeru Edo keizai ni taisuru Osaka no chii: Ky6h5 kaikakuki ni okeru shij6
koz6 ni tsuite" (The position of Osaka in the Edo economy during the
Ky6h6 reform period: concerning the market structure during the Kyoho
reform period), Nihon rekishi 191 (April 1964):2-3 1.

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
 Rozman: Edo's Importance 107

signifying the specialization of production within eastern Japan di-

rectly for the Edo market. Villages within the Kanto region became
actively engaged in commercial production, including to a rapidly
increasing extent textile processing. Various goods such as cotton and
soy sauce previously imported from the Kinai cities now came pri-
marily from local producers. Unlike the decline of Osaka and Kyoto
in the face of this rural dispersion of processing activities, Edo con-
tinued to prosper into the 1850's.3 Rural competition, the extension
of han monopolies, and the breakdown in exclusive merchant organi-
zations all disrupted the commercial centrality of other cities, while
improving the supply of goods and revenues to Edo. This is not to
suggest that Edo was left unaffected internally by these various
changes. A redistribution of real income from bushi to merchant, a
move from forced migration within the city to freer urban relocation
and sprawl, and a continued disruption of old commercial organiza-
tions in favor of freer associations all accompanied the emergence of
Edo's new marketing patterns.
By the mid-nineteenth century Edo's chonin population may have
outnumbered that of Osaka and Kyoto combined. Edo's dominance
had reached its peak, yet even much earlier there is reason to think
that Edo's impact on marketing loomed large. To clarify this point,
it is necessary to reconsider in a general way the nature of marketing
in premodern societies.
Just as employment opportunities shape migration patterns, so
too does consumption shape marketing patterns. In premodern soci-
eties the value of urban craft production, serving primarily local
customers, comprises a tiny fraction of the national product. Much
more value is contributed by commerce, particularly involving rural
products. Thus the impact of a city is less a function of its notable
handicraft industries or its exports of processed goods to other cities
than a function of its consumption demands. Already in the seven-
teenth century Edo established itself as the single dominant center
of consumption in Japan. This dominance was never relinquished,
although its character changed with the declining relative prosperity
of bushi as opposed to chonin. As peasant real income rose, patterns
of consumption spread from Edo to rural areas.
Edo's enormous consumption needs directly reshaped production
patterns in the Kanto region and in parts of the Tohoku and other

33. Evidence on this point is presented in Hauser, Economic Institutional

Change in Tokugawa Japan pp. 33-58. See also Toyoda Takeshi and Kodama
K6ta, eds., Ryitsishi (History of marketing), vol. 13 in Taikei Nihonshi
sosho (Yamakawa Shuppansha, 1969).

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
108 Journal of Japanese Studies

regions and indirectly filtered through other cities

country. To the extent that more revenues and pro
able for expenses in Edo, more pressure was exert
to supply the city's needs. Through the conversion
methods and the reorganization of commercial ope
increased income and reduced self-sufficiency. The
Japan to make urban needs felt in rural areas.
With three per cent of Japan's total population
stimulate changes among the rest of the populatio
become the cornerstone of later modernization? H
tribute to the diffusion of literacy, technical skills, an
toward family and work which count as possible p
modernization? The two principal mechanisms of m
keting, I have argued above, responded to changes
specifically to new patterns of employment and con
emphasis should be given to the latter, since employm
mainly met consumption needs. Migration tapped the h
of the country for the city's needs and imprinted the
of living on individuals returning to their home a
tapped the material resources of the country for th
reoriented the city's suppliers within their home areas
isms, activated in Edo, operated with exceptional f
Mutual emulation of higher standards of living insi
contributed to new patterns of employment and co
sparked a chain reaction throughout the nation.
Edo's basic capacities for activating these mecha
erating change had much in common with those of
throughout the world. In general, the existence of l
change through a limited number of ways: throug
of securing resources, through individual pursuit of co
through making possible costly modes of living, an
market for hired labor. Most essential is the conce
viduals within the city with both the means and
mobilize resources from the rest of the country.
A large premodern city thus draws resources f
area and in the process transforms itself and the
cities around it. Urban- and rural-centered develo
initially the urban sector displays greater dynamis
the much larger rural population becomes the major
For these reasons, in premodern societies changes i
city are difficult to divorce from changes over wider
city is heavily dependent on foreign trade (which p

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
 Rozman: Edo's Importance 109

happen in small-scale societies with populations well below that of

England), its capacity to generate jobs and to increase consumption
rests on an intricate support system nestled in a network of marketing
and administrative centers.
Through an urban networks approach, we can think of Edo's
impact reaching out in a number of steps: most directly to the 2-3
per cent of Japan's population in Osaka and Kyoto, who shared in
the widest access to resources, then to another 2-3 per cent of the
population in cities with at least 30,000 inhabitants scattered across
the country, and step by step down to the smallest marketing centers.
Simultaneously the process of diffusion reached out directly from the
most populous cities to the surrounding countryside. In particular,
Edo functioned as an administrative center for Osaka, Kyoto, and
some other major cities, as a marketing center for the entire urban
network, but especially its eastern half, as a center of elite migration
from most cities with 10,000 or more residents as well as from many
cities in lower population ranges, and as a center for peasant migrants
from the Kanto region and other areas in eastern Japan. In these ways,
Edo's impact spread across Japan.
There is no need here to repeat the details of a seven-level hier-
archy of central places which I have discussed elsewhere.35 It will
suffice to note developments during the second half of the Tokugawa
period at both ends of the hierarchy. At the bottom, level 7 settle-
ments, defined as standard marketing centers with fewer than 3,000
residents, begin to lose their central place functions. The number of
periodic markets in advanced regions no longer increases and, indeed,
begins to decline. This phenomenon also appeared in England and
then in France during the first decades of the eighteenth century and
in Russia by the early nineteenth century, interestingly in each country
at a time of little or no increase in the percentage of the national
urban population. Some of the commercial functions typical of this
type of settlement become dispersed in ordinary villages and at the
same time other activities are increasingly concentrated in settlements
at the higher levels of 6 and 5, that is, intermediate marketing centers
with fewer than 3,000 residents and cities with 3,000 to 10,000 resi-
dents respectively. The continued growth of zaikata central places in
the second half of the Tokugawa period supports this impression of

34. My view on the unimportance of foreign trade contrasts with the view
expressed by Smith, "Pre-modern Economic Growth," pp. 147-49.
35. Gilbert Rozman, Urban Networks in Russia, 1750-1800, and Pre-
modern Periodization, chapter 1.

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
110 Journal of Japanese Studies

redistribution of functions among central place

local areas.
At the top of the urban hierarchy in these countries, level 1 cities,
national administrative centers with more than 300,000 residents,
generally increased their central functions at the expense of cities at
levels 2, 3, and 4 (with populations in excess of 300,000; 30,000 and
10,000 respectively). In this regard, the decline of Osaka, Kyoto, and
some fr-kamachi, as well as the falling percentage of the total urban
population within some of these countries in cities other than the
national center which exceeded 10,000 in population, indicates a new
centripetal force. Obviously, regional variations must be examined
carefully before firm conclusions should be drawn. It does, however,
appear likely that Edo's rising preeminence reflects a common pattern
of concentration of resources in a single city where the scale of the
society permitted.
Through comparisons of five countries I have tentatively con-
cluded that the two stages of Tokugawa society closely parallel similar
divisions in the late premodern histories of England, France, and
Russia, but not China. After a century or more of rapid growth in
the percentage of the national population in cities, the urban per-
centage leveled off in each case. In comparison to the other countries,
Japan supported the highest average population per central place
(17,000 for each of its more than 1,700 ft-kamachi and marketing
settlements), the highest average urban population per central place
(3,000 based on a total of 5.3 million urban residents), and the largest
city. Its extremely efficient urban network signifies that a small number
of lower level central places supported relatively large numbers of
populous cities. The absence of much further urban growth in Japan
as a whole and in Edo in particular during the second half of the
Tokugawa period reflects the fact that Japan, as England, already
boasted an extremely efficient premodern urban network. The dyna-
mism of local areas at this time in both countries reveals a restruc-
turing, perhaps, essential for subsequent urbanization on a new
foundation.
The kind of changes taking place in Edo during the second half
of the Tokugawa period reveals that the city was expanding its sources
of support. Whereas during the first half of the period Japan achieved
extraordinary centralization of population at the various high levels
in the urban pyramid, now the direct channeling of resources through
low levels and villages advanced. In the process the number and
wealth of ch6nin increased in the urban total and urban sprawl
reflected more diverse contacts with the outside. Changes in the
sources of supplies, in the distribution of wealth, in urban land use,

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
 Rozman: Edo's Importance 111

and in other urban characteristics should be further examined to

determine Edo's new-found position in Tokugawa society.
Among the largest cities in these five countries, only London
topped Edo's percentage of the national population. No city rivaled
Edo's massive elite migration and, with the possible exceptions of
England and France, no country could have approached Japan's
high percentage of production redistributed to cities. Japan's efficient
urban network speedily conveyed to village residents changes in
consumption, in styles of living, and in aspirations. Japan not only
resembled England in the sorts of changes identified by Wrigley,
but Edo also resembled London in its capacity for generating such
changes. Similar to London, the city "must have acted as a powerful
solvent of the customs, prejudices and modes of actions.... There
were many more lodgers than in the countryside, as well as servants
[and] apprentices.... Outside the household, moreover, a far higher
proportion of day-to-day contacts was inevitably casual.... The shop,
a most important new influence upon consumer behavior, was a nor-
mal feature . . ." In these and other ways Edo generated improved
standards of performance and new tastes for living better.36
While not every characteristic of London was repeated in Edo
(for instance, the merchants of Edo were excluded from the ruling
elite), Wrigley's conclusion for London would likely hold about as
well for Edo during the second half of the Tokugawa period. It "was
so constituted sociologically, demographically and economically that
it could well reinforce and accelerate incipient change." Of the ten
cities in the world with populations in excess of, 500,000 in 1800,
London and Edo were probably uniquely in a position to influence
their countries.
Looking backward from the experiences of modernization in
numerous societies, explanations of success in this process must take
into consideration the premodern base from which change took place.
No less important than the goal of throwing light on the global origins
of this momentous period of rapid change is the need to uncover the
preconditions of successful latecomers. Anything which distinguished
Japan from other latecomers during the century or more preceding
its initial modernization may help to throw light on the conditions
which made possible rapid development once the initial outside forces
of change were introduced.
Premodern societies varied in their urban development as they
varied in the extent to which the mechanisms of migration and mar-
keting interrupted the preexisting routines of peasant existence. Large

36. Wrigley, "A Simple Model of London's Importance," pp. 50-51.

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms
112 Journal of Japanese Studies

cities with concentrated wealth and plentiful employment opportuni-

ties had the potential to dislodge these old routines. In countries al-
ready endowed with complete premodern urban networks the special
impetus of a single, great city could be realized through two phases
of successive changes. The first phase, operating primarily in cities,
was characterized by a substantial increase in the percentage of the
population in cities, by a restructuring of the urban network increas-
ing its efficiency in various ways, and by a growing circulation of those
with the greatest control over resources between local centers and a
national administrative center. Not many societies experienced this
first phase of advanced premodern growth; probably none entered it
before the late sixteenth or early seventeenth centuries. Edo along
with London and St. Petersburg emerged as one of the world's great
cities as this- phase of development progressed.
The second phase was one of a diffusion of skills and orientations
from cities to villages. Striking changes in rural areas and in settle-
ments at the bottom levels of the central place hierarchy went along
with continued, though less dramatic, changes in cities. The rural
phase may not have occurred before the late seventeenth century.
It seems to have occurred only in societies relatively urbanized for
the world at that time and in which one great city bore special impor-
tance as an incubator of change. In each of these phases Edo trailed
London by roughly half a century. However, by the late eighteenth
century London's premodern growth had given way to the first exam-
ple of modernization, in which the pace of change accelerated drasti-
cally. This is not the place to consider why early nineteenth century
Japan did not follow England's precedent with self-initiated modern-
ization. What should be emphasized is that Japan did continue its
second phase, most likely securing a firmer foundation for moderniza-
tion as a latecomer.
In conclusion, I would note that despite Edo's position as probably
the world's largest eighteenth century city, it has scarcely been noticed
by urban specialists outside of Japan. Moreover, among specialists in
Japan, Edo has not been credited with the significance in generating
economic growth which it rightfully deserves. Comparisons with cities
in other societies show that indeed Edo possessed the basic charac-
teristics that elsewhere are identified as the sources of major social
change. New forms of mobility and consumption originating in Edo
did spread through Japan. Further comparisons are essential to show
more precisely how Edo's impact on Japan facilitated the emergence
of preconditions for modernization.
PRINCETON UNIVERSITY

This content downloaded from

200.3.152.96 on Fri, 09 Apr 2021 14:46:56 UTC
All use subject to https://about.jstor.org/terms