York, Abigail, Michael E. Smith, Benjamin Stanley, Barbara L. Stark, Juliana Novic, Sharon L ...

17 Pages·2011·0.33 MB·English

48(11) 2399–2415, August 2011
Ethnic and Class Clustering through the
Ages: A Transdisciplinary Approach to
Urban Neighbourhood Social Patterns
Abigail M. York, Michael E. Smith, Benjamin W. Stanley, Barbara L. Stark, Juliana Novic,
Sharon L. Harlan, George L. Cowgill and Christopher G. Boone
[Paper first received, March 2010; in final form, April 2010]
Abstract
This paper presents initial findings from longer-term transdisciplinary research
concerning the social dynamics of urban neighbourhoods. It examines the spatial
clustering of ethnicity and class in neighbourhoods over urban history, from Bronze
Age Mesopotamia to contemporary cities. Fourteen distinct drivers of social clustering
are identified, grouped under the headers of macro-structural forces, the state, local
regimes and institutions, and bottom–up processes. The operation of these processes
is examined through three historical and three archaeological case studies of clustering.
It is concluded that: clustering is a common, but not universal, attribute of cities; there
is much variation in clustering patterns, both within and between cities and urban
traditions; and, consideration of a wide variety of drivers is required to understand
historical and modern residential dynamics.
For [cities] are each one of them many cities, neighbourhoods with concentrated ethnic,
not a city, as it goes in the game. There are two class, occupational or religious groups has
at the least at enmity with one another, the been repeatedly documented from Plato
city of the rich and the city of the poor (Plato,
to the analyst of modern census tracts. In
The Republic, Book IV 1937 edn).
some eras, writers praise such social cluster-
Observers of human society have written ing; the Han dynasty Chinese philosophical
about social clustering in cities since at aphorisms recorded in the Guanzi annals
least the time of Plato. The existence of include this statement
Abigail M. York, Michael E. Smith, Barbara L. Stark, Juliana Novic, Sharon L. Harlan, George L. Cowgill
and Christopher G. Boone are in the School of Human Evolution and Social Change, Arizona
State University, PO 872402, Tempe, Arizona, 85287-2402, USA. E-mail: [email protected],
[email protected], [email protected], [email protected], [email protected], [email protected] and
[email protected].
Benjamin W. Stanley is in the School of Sustainability, Arizona State University, Tempe, Arizona,
USA. E-mail: [email protected].
0042-0980 Print/1360-063X Online
© 2010 Urban Studies Journal Limited
DOI: 10.1177/0042098010384517
Downloaded from usj.sagepub.com at ARIZONA STATE UNIV on August 1, 2011
2400 ABIGAIL M. YORK ET AL.
The scholar-official, the peasant, the draftsman in the midst of a long-term investigation of
and the merchant ... should not mix with one urban life and the dynamics of change from
another, for it would inevitably lead to conflict the earliest cities to the present. Although
and divergence of opinions and thus complicate
our empirical research is far from complete,
things unnecessarily ... Let the scholar-official
we have identified 14 distinct drivers that
reside near school areas, the peasants near
promote or discourage social clustering in
fields, the craftsmen in the constructions
pre-modern2 and modern cities, and have
workshops near the officials’ palace, and the
reached preliminary conclusions about the
merchants in the shia [commercial wards]
(quoted in Kostof, 1992, p. 102). empirical variation in urban neighbourhoods
across time and space. These initial findings
In the modern era, writers tend to disapprove will assist in development of a major system-
of residential clustering typically discussed in atic comparative study to answer some of the
terms of discriminatory segregation.1 There larger questions posed earlier.
is a general consensus that heterogeneous
neighbourhoods have social advantages
Background
(Fainstein, 2005; Talen, 2006) despite the
fact that homogeneity appears to be more
Intellectual Context
frequently observed in modern cities (for
example, Knox and Pinch, 2006, pp. 168–187). The historical and archaeological records are
Urban planners distinguish “socially accept- replete with examples of cities with socially
able clustering” from “clustering that is homogeneous neighbourhoods and cities
undesirable” (Marcuse, 2005, p. 15). Others with heterogeneous residential zones. Many of
investigate socially heterogeneous urban the earliest cities with good data—in southern
neighbourhoods in order to identify the fac- Mesopotamia—had mixed neighbourhoods,
tors that encourage diversity (Nyden et al., whereas some cities in the same region exhib-
1997; Talen, 2010). ited marked clustering by wealth (see later).
What can the historical record tell us about Processes of segregation and mixing are far
the occurrence of social clustering in urban from uniform within individual historical
neighbourhoods? Some scholars assert that and regional contexts. Both medieval Europe
socially homogeneous neighbourhoods have and Aztec Mexico were settings in which
been the norm throughout history (Rapoport, some cities exhibited marked social cluster-
1980/81). New urbanists argue that the ing while others did not. What accounts for
homogeneity of modern neighbourhoods this variation? Some writers claim that social
represents an extreme that is neither socially clustering is always imposed by the state
healthy nor rooted in history (Talen, 2006). (Marcuse, 2002), while others endeavour to
Were the earliest cities socially segregated, or show that extensive clustering can result from
did ethnic and class groups live interspersed the unco-ordinated and unintentional actions
with one another? How similar are the of individuals (Schelling, 1960).
dynamics in pre-modern and modern cities? Our point of departure is the notion that
Can information about these patterns con- all cities share a set of basic social dynamics
tribute to a better understanding of general that permit comparative analysis of urban
processes of urbanisation? life. We agree with recent calls for greater
In this paper, we report initial findings from comparison in urban studies (Nijman, 2007;
a transdisciplinary research project entitled Sellers, 2005), but argue that most compara-
‘Urban organization through the ages: neigh- tive works do not go deeply enough into the
borhoods, open spaces, and urban life’. We are past. In order for comparative urbanism to
Downloaded from usj.sagepub.com at ARIZONA STATE UNIV on August 1, 2011
NEIGHBOURHOOD SOCIAL DYNAMICS 2401
address underlying processes such as social broken new ground in applying systematic
clustering, comparison should not be limited comparisons to pre-modern urban societies,
to the modern period. A broad understanding although their focus is governance and col-
of ‘urban’ must incorporate pre-industrial, lective action rather than urbanism.
non-Western and interdisciplinary perspectives Intensive comparisons employ fewer cases
(Smith, 2009). and greater social and historical contextu-
Some may question our effort to compare alisation (Tilly, 1984; Trigger, 2003). In an
modern urbanism with ancient and non- approach that might be called ‘exemplary
Western urban experiences because capitalism comparison’, Xavier de Souza Briggs (2004)
has fundamentally changed land markets, or describes a project close to our own in
because transport and other technologies its goals. He compares ethnic diversity in
have altered human interaction, or because imperial Rome, medieval Córdoba and con-
democratic institutions have changed social temporary Los Angeles in order to generate
relationships within cities. We do not propose insights on how such diversity has been
a single experience or trajectory of historical handled by governments. Another kind of
development; rather, we suggest that a set of intensive approach can be called ‘typological
drivers influence clustering patterns observed comparison’; cities (or other phenomena)
throughout time. One of the goals of our proj- are divided into types and the problem of
ect is to identify and analyse such processes interest is analysed separately for each type.
and conditions through comparative analysis. A good example is Grillo’s (2000) comparison
We are not the first to make empirical com- of four types of ‘plural cities’: pre-industrial
parisons among diverse kinds of cities—we patrimonial cities, colonial cities, modern
build on the insight of a number of scholars, industrial cities, and neo-liberal post-modern
including Besim Hakim (2007), Jill Grant cities. He identifies characteristic patterns in
(2001), Xavier de Souza Briggs (2004) and the political-economic role and significance
Ralph Grillo (2000). of ethnicity in each of his types.
While the research of Briggs and Grillo is
Approach to Comparison
valuable, we argue that deeper empirical anal-
Our project is comparative, but we find it yses of specific cities are required to model
difficult to situate our approach within exist- processes of social clustering. Our approach
ing categories of comparative social science combines elements of the systematic and
methods (for example, Ember and Ember, intensive strategies of comparison. Another
2001; Mahoney and Rueschemeyer, 2003; distinctive feature of our research is its
Ragin and Becker, 1992; Tilly, 1984; Nijman, transdisciplinary nature. We started with the
2007; Trigger, 2003). Approaches to compari- notion that an adequate comparative under-
son are often discussed in terms of a contrast standing of urban neighbourhoods and social
or continuum between systematic and inten- clustering, like many other phenomena in the
sive comparative methods (Caramani, 2009). human sciences, requires research that goes
Systematic comparisons typically involve beyond the confines of individual disciplines
large-number random sampling strategies (Polimeni, 2006; Wallerstein, 2003). Our team
and the statistical analysis of many variables includes scholars from archaeology, geogra-
(Ember and Ember, 2001; Ragin, 1987). The phy, sociology, political science and sustain-
application of these methods to pre-modern ability studies. By combining the insights of
cities is limited by the scarcity of available these diverse disciplines, we hope to generate
data. A recent project by Richard Blanton new understandings of neighbourhoods and
and Lane Fargher (2008), however, has social clustering in pre-modern and modern
Downloaded from usj.sagepub.com at ARIZONA STATE UNIV on August 1, 2011
2402 ABIGAIL M. YORK ET AL.
cities. In the following section, we review in terms of both free choice (Dahya, 1974)
drivers that generate urban social clustering as well as social constraint (Rex and Moore,
under different conditions. 1967). As the debate between Dahya (1974)
and Rex and Moore (1967) regarding drivers
of Pakistani clustering in Birmingham illus-
Drivers of Ethnic and Class
trates, constraint and choice should often be
Clustering
understood as mutually reinforcing.
There is a large and varied academic literature Other authors tend to focus on ‘structure’—
on the drivers of ethnic and class cluster- the larger, often global, drivers of change
ing in contemporary cities, spanning many generating class differences or bringing
disciplines and theoretical approaches, but ethnicities into close contact in urban areas.
most writers focus entirely on modern cities. Some believe that the literature on Western
The historical and archaeological literature clustering is
on social clustering in pre-modern cities is
much thinner and typically case-oriented dominated by those adhering to the structural
approach ... Structural global economic
with little concern for comparison, gener-
processes and, recently, also structural
alisation or theory. Some of the dynamics of
differences between welfare states are brought
clustering appear to be quite similar in ancient
to the fore as the main forces behind the
and modern cities, while others are distinct,
social and spatial processes [of segregation]
and one of the tasks of our larger project is
(Musterd et al., 1999, p. 578).
to disentangle and examine these processes.
The Chicago School of Sociology first In the following section, we briefly lay out
highlighted social clustering (or segrega- the major types of top–down, bottom–up
tion) by presenting a human ecology model, and structural theories often presented to
which hypothesised that American ethnic explain social clustering. Although multiple
groups and neighbourhoods proceed through processes are usually at work in any given
a series of cultural stages (Park, 1926). situation, the interesting issues concern the
Subsequent researchers, both supportive and relative importance of each driver in specific
critical of this approach, focused on the exis- cases and the varied ways in which they may
tence of clustering and measurement issues interact. Except in the smallest societies, the
(for example, Duncan and Duncan, 1955; labels ‘top–down’ and ‘bottom–up’ are over-
Goldsmith and Stockwell, 1969), the impact simplifications because complex societies are
of clustering on populations (Marschall multilayered hierarchies. Nevertheless, these
and Stolle, 2004), the causes of segregation provide a convenient way to present drivers
(Bruch and Mare, 2006) and policy prescrip- of social clustering.
tions (Nelson et al., 2004).
Macro-structural Processes
Arguments about top–down versus bottom–
up drivers are common within this literature. A variety of macro-level processes—such
Ceri Peach (2003) has described several as major shifts in socioeconomic systems—
archetypes of clustering present within the clearly influence residential patterns. Our
American and Canadian urban experience, focus is primarily on relatively proximate fac-
which reflect both top–down and bottom– tors affecting neighbourhoods and clustering.
up pressures. Peach (1998) has framed this We recognise that a variety of prominent pro-
dichotomy in terms of ‘constraint’ and ‘choice’. cesses operate at a broad level and affect ethnic
Increasing patterns of ethnic-based cluster- relations, occupations and class, which in turn
ing observed in Britain have been explained can affect urban clustering—processes such
Downloaded from usj.sagepub.com at ARIZONA STATE UNIV on August 1, 2011
NEIGHBOURHOOD SOCIAL DYNAMICS 2403
as European colonial expansion, pre-modern clustering (Sassen, 1991, Levitt and Jaworsky,
state expansion (Stark and Chance 2008, 2007). Other scholars question the impor-
pp. 24–32), urbanisation and world-system tance of globalisation processes in generat-
interactions (Chase-Dunn and Manning, ing social clustering, pointing instead to
2002; Smith, 1996; Hall 1998). These processes a variety of state and civic mechanisms
often are a precursor to specific neighbour- (van Kempen, 2007). The world-systems
hood patterns, initially operating to bring literature covers a broader temporal range
together a diversity of ethnic, class and family than globalisation studies and the effects of
groups in urban areas. world-systems processes on city size have
become an important topic of research
(1) Industrialisation. Industrialisation, which (Chase-Dunn and Manning, 2002). David
leads to rural-to-urban labour migration, Smith (1996) has addressed the impact of
brings highly diverse groups into close con- differential world-system position on devel-
tact. This driver of clustering was important opment, poverty and inequality, but so far
in the US and Europe in the 19th century and specific linkages to clustering processes have
it remains a significant factor in many parts not been studied.
of the developing world today. The industri-
alisation of African cities triggered extensive (4) Pre-modern commercialisation. The
migration from rural villages to urban centres. role of commercialisation did not originate
Since rural Africa contains highly diverse tribal with capitalism. Some pre-capitalist econo-
groups, this directly led to ethnically diverse mies had well developed commercial insti-
African cities characterised by various degrees tutions, while others had state-controlled
of conflict and co-operation between groups non-commercial economies; the former are
(Hanna and Hanna, 1981). distinguished from capitalism by the impor-
tance of wage labour and land markets in
(2) Capitalism. David Harvey (1989) and capitalist economies (Smith, 2004). In at least
Richard Sennett (1990) argue that capital- one documented case, growing pre-modern
ism changed the relationships between indi- commercialisation led to increased social
viduals and their environments—particularly clustering at the neighbourhood level (see the
their relationships to real property—leading Chang’an case study later).
to class clustering. Post-Fordist economic
restructuring following the crisis of the (5) Religious rules. Islamic law contains
mid 1970s has been implicated as a cause of numerous provisions concerning the urban
increased segregation. The retraction of wel- built environment and social relations
fare benefits like public housing provision and between groups of people (Akbar, 1989;
expansion of free market reforms accelerated Hakim, 1986, 2007) and these have affected
growth of wealth-based disparities and class- processes of social clustering in some
based residential enclaves in developed cities Islamic cities. Wirth (1956) argues that
such as Hamburg (Dangschat, 1994). Jewish neighbourhoods in European cities
began as voluntary clusters enabling the
(3) Globalisation and world systems. Some practice of the Jewish faith and the main-
scholars argue that the effects of globalisation— tenance of the community, but were for-
such as the loss of industries leading to malised as ghettos by the Christian Church
unemployment and immobility, or the and cities, as Christian intolerance grew
growth of industrial centres in develop- towards Jews during the Crusades (see also
ing countries—contributed to class-based Haverkamp, 1995).
Downloaded from usj.sagepub.com at ARIZONA STATE UNIV on August 1, 2011
2404 ABIGAIL M. YORK ET AL.
The State The policies and decisions of pre-modern
Formal state policies addressing the built states also affected social clustering. We
environment directly generated residential include the impact of imperialism and colo-
clustering. Whether expressly intended to nialism (both recent and ancient) on clus-
bring about clustering—such as segrega- tering in this category. In the ancient world,
tionist or colonial planning policy in the Inca emperors in Peru often moved whole
20th century—or aimed at a broader range villages to new areas and Aztec rulers in cen-
of social objectives, policies established by tral Mexico granted urban land to immigrant
nation-states often lead to clear-cut examples ethnic groups for political gain (for example,
of clustering. Hicks, 1982), leading to ethnic clustering.
Augustus famously divided Rome into 14
(1) Federal law and policies. In modern residential districts or areas (Lott, 2004), but
Europe, state policies regarding assimilation there is little evidence of how this may have
(for example, France) and multicultural affected social clustering.
or pluralist approaches (for example, the
Netherlands) influence ethnic settlement pat- (2) Planning and public works. Modern
terns amid complex political forces operating state planning projects and public works
at more local scales (Musterd, 2005). French often displace people, frequently the poor
federal policy aimed at the socio-spatial or disenfranchised, because of concerns
assimilation of highly clustered immigrant about social unrest, public health or the
groups has been largely ineffective, distorted aesthetic of crowded slums. The size of
by the ways in which municipal levels of gov- such works in urban areas was limited prior
ernment interpret and implement such poli- to Haussmann’s major reconstruction of
cies. Laws preventing the legal recognition of Paris in the 19th century (Jordan, 1995;
minorities have been cited to explain increas- Rabinow, 1989). In the mid 20th-century
ing patterns of ethnic clustering in suburban US, urban renewal efforts supported by fed-
public housing estates (Simon, 1998; Rhein, eral governments fuelled clustering in cities
1998). In the US, Supreme Court cases across the country (Anderson, 1964). In
opened the housing market to non-Whites, many cities, such as Algiers and New Delhi,
yet local institutions and norms—such as colonial city planning drastically altered
mortgage and insurance discrimination by the spatial layout forming new enclaves
the public and private sectors—reduced the or ghettos outside the central city (Çelik,
ability of minorities to purchase homes in 1997; King, 1976; see discussion of Algiers
White neighbourhoods (Gotham, 2002). later). In modern Chinese cities, concerted
More explicit state policies about citizenship efforts by the state to redevelop inner-city
and rights also affect class and ethnic cluster- areas displaced large numbers of residents,
ing. In southern Africa, colonial administra- breaking up older economically diverse
tions and local political élites co-created tribal neighbourhoods (Abramson, 2007; He
ethnic and racial categories, which were codi- and Wu, 2007). Fully planned cities often
fied (Vail, 1989) and designated specific lands show premeditated forms of clustering-
for settlement (Marks, 1989). South Africa’s such as Chandigarh, the administrative
apartheid policy led to neighbourhoods and capital of Punjab, where construction of
townships segregated by race (Christopher, various sizes of housing mitigated Indian
2001) that continue to cluster after the tendencies to cluster by ethnicity and caste,
legal repeal of apartheid in 1991 (Kotze and but promoted class-based areas in the city
Donaldson, 1998). (D’Souza, 1968).
Downloaded from usj.sagepub.com at ARIZONA STATE UNIV on August 1, 2011
NEIGHBOURHOOD SOCIAL DYNAMICS 2405
(3) Housing policy. Residential housing the US, single use zoning (Young, 1990) and
projects built by federal and local governments growth politics (Peterson, 1981) or machines
generated specific clustering patterns in cities (Logan and Molotch, 1987) fuelled clustering
around the world. In the former Soviet Union, at a neighbourhood level.
state allocation and residential construction
policies only partially mitigated class-based (2) Real estate practices. Informally within
clustering; many neighbourhoods continued the modern system of building codes, plan-
to show concentrations of wealth, while oth- ning and zoning, realtors and developers
ers exhibited clustering based by occupation imposed their ideas about separation of uses
and age (Smith, 1989; Gentile, 2004). In and people. In the US, until the Fair Housing
20th-century western Europe, Kesteloot and Act of 1968, it was legal to steer buyers to
Cortie (1998) show that different public hous- different residential areas based on race.
ing policies in Belgium and the Netherlands Prior to the US Community Reinvestment
led to divergent residential patterns among Act, and even afterwards, local banks refused
Turkish and Moroccan immigrants; Dutch to grant mortgages for particular areas of
policies scattered public housing leading to town or for Blacks moving into White areas.
a dispersal of immigrants, whereas a lack of Thus, even without government interference,
social housing in Belgium pushed immigrants informal city-level institutions contribute to
into specific districts of older housing stock. clustering.
Local Regimes and Institutions Bottom–Up Processes
Urban regimes, coalitions of municipal gov- Bottom–up processes—a catch-all term
ernments, commercial interests and local describing the initiatives of individuals, small
élites drive social clustering, although this groups and grassroots movements—affect
may be a recent phenomenon. Van Kempen social clustering in both modern and pre-
notes that in the modern period modern settings. The degree to which such
processes can be conceptually separated
the state gradually becomes only one of
from top–down constraints continues to be
the actors [generating segregation and
a subject of academic debate (Peach, 1998).
clustering], especially when all kinds of
coalitions, partnerships, and governance
(1) Individual and household preferences.
emerge (van Kempen, 2002, p. 50).
Scholars of contemporary cities have shown
(1) Local land use policy. Contemporary city how individual and household actions can gen-
administrators commonly restrict land uses erate macro social processes at the neighbour-
through zoning and related mechanisms. In hood level. An influential work in this area was
earlier periods, laws regulating landownership Schelling’s (1960) computer simulation, which
served a similar purpose. Dan Smail’s (2000) showed how simple threshold rules and a pref-
study of medieval Marseille illustrates the erence for a majority of similar neighbours can
growing encroachment of municipal authori- generate patterns of hyper-segregation. More
ties as city officials took over regulative activities recently, Bruch and Mare (2006) demonstrate
formerly managed by informal neighbourhood that smoothing the preference functions of
organisations. Decisions about building, trans- individual agents leads to lowered segregation
port, demolition and renewal often occur at the in simulations, yet segregation still persists.
city level. In the Western context, notorious Economists often use arguments based on
institutions such as racial zoning and restrictive individual choice to explain the ‘White flight’
covenants led to neighbourhood clustering. In during the 1950s and 1960s (Gotham, 2002).
Downloaded from usj.sagepub.com at ARIZONA STATE UNIV on August 1, 2011
2406 ABIGAIL M. YORK ET AL.
(2) Mutual support. Ethnically clustered stabilisation can occur through both formal
neighbourhoods can provide mutual sup- and informal governance arrangements.
port for vulnerable and marginal groups in neighbourhoods may self-govern by provid-
modern cities. Gerald Suttles (1968) argues ing public goods, leading to Tiebout-like sort-
that residential clustering in Chicago and ing in which individuals ‘vote with their feet’
other US cities persisted for three reasons: to and move to neighbourhoods with preferred
minimise conflict between different groups; amenities (Aronson, 2001). Although it is dif-
to maximise political voice through such ficult to apply simulation models to ancient
functions as block voting; and, to establish cities, the concept of neighbourhood self-
the greater self-control and self-policing made regulation is consistent with Grillo’s (2000)
possible in homogeneous groups (see also model, which suggests that ethnic clustering
Knox and Pinch, 2006, pp. 175–177). Similar in pre-industrial cities arose from bottom–up
processes have been identified in squatter and forces without the intervention of state or city
slum settlements in developing countries. authorities.
Nijman (2010) shows that aspects of spatial
organisation in the slums in Mumbai can be
Dynamics of Change in Ethnic
seen as the result of individual choice relating
and Class Clustering: Three Case
to issues of defence and support.
Studies from History
(3) Chain migration. For pre-industrial cit- To show the operation of many of our cluster-
ies and some modern cities in the developing ing drivers, we describe three case studies—
world, rural-to-urban migration is the most Chang’an, Algiers and Prague—selected from
important bottom–up driver. Ethnographers our growing database. These cases illustrate
have identified numerous cases in which clustering pattern dynamics in response to
immigrants from particular rural areas settled changes in major drivers.
in distinct urban neighbourhoods that were
perpetuated by continued in-migration Chang’an, China, 9th–10th Century:
From State Control to Commercialised
(Mangin, 1970; UNCHS/Habitat, 1982).
Economy
Greenshields (1980) shows how this works
in historical and recent Near Eastern cities. During its height in the T’ang period (581–
Established migrants help to find homes for 907 A.D.), the Chinese imperial capital
immigrant relatives and friends near their Chang’an was the largest city in the world with
own dwellings and ethnic solidarity becomes around 1 million inhabitants (Xiong, 2000).
a form of migrant adaptation to urban life. There is considerable historical documenta-
Another kind of rural-to-urban migration tion of this walled city and modern research
that can generate clustering is the relocation reveals a causal relationship between com-
of particular social groups to cities in reaction mercial expansion and increased social clus-
to the intrusive actions of states in rural areas tering by neighbourhood. During the T’ang
(Chance and Stark, 2007, pp. 216–218; Scott, period, the state exerted strong control over
1998, pp.185, 395). neighbourhoods and patterns of city life. In
what is called the ‘ward system’, large walled
(4) Neighbourhood self-regulation. Empirical residential wards or districts called fang had
work has demonstrated that contemporary gates that were guarded and closed at night
neighbourhoods ‘self-regulate’, leading to to traffic. Marketplaces were restricted to two
temporal stability in indicators such as crime walled zones inside the city and commercial
rates (Galster et al., 2007). This process of activity was not permitted in residential
Downloaded from usj.sagepub.com at ARIZONA STATE UNIV on August 1, 2011
NEIGHBOURHOOD SOCIAL DYNAMICS 2407
areas. The social composition of Chang’an and a number of diverse ethnic groups
was reflected in broad spatial patterns. Most clustered in specific residential spaces,
foreigners lived in the western part of the city including Andalusians, Moors, Kabyles, Jews,
and the houses of aristocrats were roughly Saharans and Europeans (Çelik, 1997; Shuval,
clustered. Nevertheless, social groups were not 1998). On the eve of French colonialism in the
extensively clustered within the city (Xiong, 1820s, this residential clustering was generated
2000, ch. 8; Tatsuhiko, 1986). and maintained by bottom–up drivers such as
The transition to the following Song period chain migration and mutual support and by
(960–1127 A.D.) was marked by population Islamic law, a macro-structural driver. Most
growth, a florescence of commercial activity groups mixed in public streets and markets. As
and a reduction in state control over residence in many Islamic cities, religious law encour-
and movement. Buying and selling escaped aged wealth-based mixing within clan-based
the confines of the T’ang marketplaces and compounds (Miege, 1985).
streets with shops became busy centres of The imposition of French rule drastically
commerce. The ward system broke down affected the social makeup of Algiers. The
as gates were opened and officials no longer French government rased neighbourhoods in
enforced curfews or regulated movement the Marine quarter and built a new modern
(Heng, 1999, ch. 4). These developments led urban district (Rabinow, 1989) that soon
to major changes in land use and residential became home to French, Italian, Spanish and
patterns, a process Tatsuhiko (1986) calls Maltese immigrants. The historical Casbah
‘urban specialisation’. One component of this neighbourhood became the cultural heart of
was the formation of neighbourhoods (“social the stigmatised Muslim population.
communities”; Tatsuhiko, 1986, p. 176) that
Racial, cultural, and historical otherness
appear to be clusters of people with similar
constituted the main paradigm that dominated
social characteristics. Thus, in Chang’an, the
all building activity in Algiers during the
relaxation of central political control accom-
French occupation, and spatial separation
panied by commercialisation produced urban
in the most concrete sense reinforced the
restructuring, including an increased social difference (Çelik, 1997, p. 5).
clustering at the neighbourhood level. In
our terminology, state drivers were replaced Although some groups continued to clus-
by structural and bottom–up drivers as the ter residentially in traditional patterns, the
major forces affecting social clustering. breakdown of Islamic rule led to new neigh-
bourhoods segregated according to income.
Algiers, 19th Century: Ottoman to Official segregation policies instituted in the
French Colonial City
1920s helped to reinforce clustering, although
Prior to French colonial rule, Algiers, Algeria, certain exempt neighbourhoods, like the
resembled many Near Eastern cities in Marine quarter, showed signs of heterogene-
having strong neighbourhood clustering ous mixing.
based on ethnicity and rural place of origin By the 20th century, an acceleration of
(Greenshields, 1980). Although neighbour- rural-to-urban migration transformed
hoods were separated by walls in these Algiers into “essentially a Berber-European
cities, the walls were built and maintained city” (Miege, 1985, p. 176). Public housing
by residents, in contrast to the centrally camps built by colonial authorities far from
built walled neighbourhoods of Chang’an the centre of the city housed some of this
(Abu-Lughod, 1987). In the 16th century, Algiers influx, but many immigrants were shunted
contained over 50 separate neighbourhoods towards growing squatter settlements and
Downloaded from usj.sagepub.com at ARIZONA STATE UNIV on August 1, 2011
2408 ABIGAIL M. YORK ET AL.
the now-overcrowded Casbah. On the eve Suburban luxury villas are increasingly
of independence, both state policies (in the common, as are luxury infill developments
form of colonial planning practices) and more proximate to lower classes. Socialist-era
bottom–up processes (such as additional eth- housing estates, now operated as rent-
nic clustering) worked to reinforce historical controlled public housing, have been increas-
trends towards ethnic segmentation in Algiers ingly deregulated and opened to market
(Çelik, 1997). forces, spurring new forms of class-based
stratification. Free-market business and tour-
Prague, Czech Republic, Late 20th
ism have fuelled gentrification of central-city
Century: Post-communist Social
neighbourhoods (Sýkora, 1999). Although
Stratification
the processes of socioeconomic stratification
At the start of World War II, élites and work- have been relatively slow in the post-socialist
ing classes clustered in different neighbour- years, shrinking state control over housing has
hoods in Prague, similar to many European been accompanied by increasing class-based
cities. The centre of the city, however, was clustering, a trend evident in other Soviet,
“socially and culturally highly mixed”, with eastern European and Chinese cities (Sýkora,
proximate populations of elderly, young, 1999; He and Wu, 2007). Prague presents a
Jewish, Gypsy, working-class and established case where the imposition of stronger state
families of the wealthy élite (Musil, 1987, controls (under socialism) worked to reduce
pp. 30–31). The Communist takeover of ethnic and class clustering and then the reduc-
1945 led to central state attempts to increase tion in state power after 1989, coupled with
class-based heterogeneity through aggressive increasing capitalist penetration, permitted
allocation and relocation policies and con- greater expression of macro-structural, local
struction of large peripheral housing estates. and bottom–up forces, resulting in increased
By 1970, socioeconomic segregation had been social clustering.
considerably reduced, as many working-class
families were relocated to the historical core,
Ethnic and Class Clustering in
but clustering based on occupation and age
Ancient Cities
increased due to housing estate allocation
policies favouring young families and specific The three case studies illustrate the roles of
professions. This sort of favouritism increased specific clustering drivers in well documented
into the 1980s, creating a distinctive class- and episodes of historical transformation. It is
occupation-based clustering. Pre-war class- likely that many of the factors identified also
based clustering endured in a small number operated in ancient cities, but the limitations
of neighbourhoods as well, since the socialist of data prevent dynamic analyses of the sort
government left most élite neighbourhoods already described. In this section, we briefly
undisturbed (Musil, 1987). review three examples that illustrate vari-
Since the fall of the socialist government in ability in social clustering in the earliest cities.
1989, Prague, like other post-socialist coun- The earliest cities—in Bronze Age
tries, has exhibited increasing levels of class- Mesopotamia—already exhibit variability in
based spatial clustering. Sýkora observes that neighbourhood organisation. The first cit-
ies arose before 3000 B.C. and, by the Old
the major factors which influence growing
Babylonian period (2000–1600 B.C.), urbanism
socio-spatial disparities in post-communist
Prague are increasing income inequalities and was firmly established in southern Mesopotamia
newly introduced market-based mechanisms (van de Mieroop, 1999). Neighbourhoods can
of housing allocation (Sýkora, 1999, p. 680). be identified in both archaeological plans and
Downloaded from usj.sagepub.com at ARIZONA STATE UNIV on August 1, 2011