Introduction

This paper examines literature on immigrants in Canadian cities from the perspective of Urban Studies.¹ The most striking feature about this literature is its paucity, which in turn indicates uneven coverage of both specific cities and specific immigrant groups. Indeed, it is still common for contemporary analyses of urban issues to ignore completely the issues raised by immigration.

There is, however, a large and rapidly growing literature on urban areas and spatial aspects of urbanisation, especially in relation to the themes of the changing world economy and urban restructuring, the status of Canadian cities in the new world economy, the socio-cultural complexity of Canada's urban populations, social relations, urban landscapes and urban culture and the nature of urban politics, both in relation to city governments and urban social movements. Although we have not reviewed this literature systematically (largely because other reviewers are addressing literature in the fields of economics, sociology and social psychology), we have tried to read the material on immigration within the theoretical context of urban systems engaged in processes of transformation in order to identify the range of themes that have influenced inquiry into immigration and Canadian cities. We therefore address the broader literature quite frequently.

Literature on urban studies is much more extensive in both the United States and Britain (cf. The Canadian Journal of Urban Research which was only established in 1992). Both countries have longstanding traditions of urban studies, and have profoundly influenced theoretical approaches in Canada. There is also a more recent but burgeoning literature from Australia which has also been drawn upon. Empirical literature from these and other countries is used, therefore, although only where it sheds light upon the Canadian experience.

It should also be noted that there is a close overlap in the interests of urban sociologists, geographers and regional economists, especially in concerns about ethnic segregation, ethnic identity, demography, ecology and the spatialisation of labour markets (see Balakrishnan and Kralt 1987; Darroch and Marston 1987; Driedger 1991; Ley 1983; McGahan 1982). Our approach is interdisciplinary, with an emphasis upon questions that affect city lives and city forms, rather than upon the disciplinary background of the writers.

There is also an overlap between issues of immigration per se and broader issues of ethnicity (see Beaulieu 1983, for example, on this conflation in relation to urban research in Quebec). It is important to make a distinction between the two since 1) we wish to avoid the common, prejudiced notion that all members of minority ethnic groups are 'immigrants' even when they are born in Canada; and 2) it follows that all Canadians are members of some ethnic group. Nonetheless, it is often difficult to distinguish ethnicity from immigrant status when they are conflated within the literature. Furthermore, because ethnic distribution is the result of immigration of further generations, it is necessary to make some links within a wider community context. Therefore we often move from the more limited analysis of immigrants and immigration to broader questions such as ethnic demography, ethnic communities, and policy issues surrounding multiculturalism and racism.

Finally, this review presents a selective discussion of the most significant works that address immigration and emerging characteristics of Canadian metropolitan areas. A more comprehensive review of the literature is found in the attached annotated bibliography. With the exception of a short section on immigrant women in the labour market in which we have tried to take a more comprehensive approach, our focus in on work that examines conditions in particular cities, rather than work undertaken at a national scale.

Historical Background

Modern Urban Studies grew out of attempts by members of the Chicago School in the 1920s and 1930s, particularly Robert Park and Ernest Burgess, to develop spatial theories of the city based on Chicago as a giant urban laboratory that came to represent "the norm in North American urbanization" (Ley 1983, p.22). It was a time of turmoil in the large, industrial cities of the north-eastern United States which were experiencing large scale immigration both from the southern United States and Europe. Robert Park began a remarkable series of writings on the city by noting the tendency for coincidence of spatial and social distance (Park 1924), which he then characterised as a spatial and moral order (Park 1926). Explicitly identifying the city as a social laboratory (Park 1929), he applied established models from the life sciences to view the city as an expression of "human ecology", governed by a Darwinian process similar to that of biological organisms, whereby "primitive forces led to conflict, both economic and on occasions physical, and so to a hierarchy of dominants and subordinates among social groups in the city" (Ley 1983, p.23; Park 1967, Chap. 2). Therefore, whereas, "In a plant community this dominance is ordinarily the result of a struggle among the different species for light...the principle of dominance operates in the human as well as in the plant and animal communities...the area of dominance in any (urban) community is usually the area of highest land values" (Park 1936, quoted in Ley 1983, p.27).

The immigrant becomes a catalyst for change within this ecological system, as a foreign organism that enters the city at its weakest point, the degraded and devalued land adjacent to the city's central business district, then adapts his or her way to a position of greater dominance, moving through the property zones to the wealthier suburban areas in the process. In this way, immigrants also shed the unsophisticated ways of "folk culture" in response to the civilizing influence of the city (Park 1928). Indeed, civilization depends upon the process of assimilation. As groups move across urban space from centre to periphery, they gain North American traits, and foster national solidarity based on 'like-mindedness' as "the immigrant readily takes over the language, manners, the social ritual, and outward forms of his adopted country" (Park 1967, p.115). 'Race', however, presents a barrier to assimilation since skin colour always establishes a symbolic distance between groups (Park 1967, p.117).

Members of the Chicago School thus established an enduring scholarly preoccupation with assimilation as both the dominant and the ideal social process associated with immigration. But, as Jackson notes, "The advocates of minority group 'assimilation' rarely pause to consider precisely whose interests such a process would serve, casually assuming it to be a universally desirable goal of social policy" (1987 p. 5; see also Ray 1992, Chap. 2). Thomas and Znaniecki (1984), whose work is remarkable for its insight into the experiences of Polish immigrants, describe a society in which immigrants become ideally and progressively, but never fully, assimilated, so that cultural traits of the 'old' world become simplified and fall away in favour of standardised American culture. As Brian Ray describes it, "The inescapable argument of the book, however, is the breakdown of traditional immigrant social orders, alienation and creeping assimilation" (Ray 1992, p.15) as immigrants enter a world where "...structure and prevalent attitudes is neither Polish nor American but constitutes a specific new product whose raw materials have been drawn partly from Polish traditions, partly from the new conditions in which the immigrants live, and partly from American social values as the immigrant sees and interprets them" (Thomas and Znaniecki 1984, p.240). This normative and reified notion of 'American' culture, which denies the emergence of a plural society, set the tone for much of the research on immigration.

The work of Louis Wirth (1938) on urbanism as a way of life also helped set the scene for the study of immigrants in the city. Wirth implied that there was some kind of connection between type of settlement and the psychic life of individuals (Savage and Warde 1993). He claimed that large settlement size, along with increasing density and heterogeneity, resulted in an urban environment in which specialisation was necessary to survive. Locality-based or community life would break down and be replaced by secondary ties based on workplace and interest groups. He aligned folk, or traditional, family-oriented, modes of social organisation based on co-operation with rural life, while he associated with urban areas a condition of anomie (of social normlessness), anonymity, social disorganisation, and impersonal behaviour, based more on self-interest than co-operation and the breakdown of primary social ties and communities. Hence there emerged a coupling at conceptual and empirical levels between patterns of social change and fragmentation and of a genre of work that considered urban culture in relation to loneliness, isolation and deviance in the city.

The theme of degradation of 'peasant' life in urban North America is perhaps best illustrated in the work of Oscar Handlin (1951), for whom the physical environment of the large city led to contagious poverty, ill health and moral turpitude, as "their lives were dominated by disorganisation rather than the norms, values and mores of the communities they left behind" (Ray 1992, p. 20). Ray argues further that Handlin's argument that assimilation is a necessary, "unilinear process revealed in ethnic identity loss" has more recently been challenged by the notion that, "Assimilation into the culture of a larger society and retention of some forms of ethnic identity can and often do take place concomitantly" (Isajiw 1990, p. 35 quoted in Ray 1992, p.20).

In the early 1960s, there began a reaction to the pessimism and unreality of these assumptions, with the counterargument that ethnic neighbourhoods and communities continued to survive in cities. Glazer and Moynihan (1963; 1970) challenged the notion of the American 'melting pot', and described New York City as a specific physical environment in which ethnicity influences politics, the character of residential neighbourhoods and a range of personal and social relationships. A number of empirical studies subsequently arose to illustrate the positive aspects of the ways of life of various ethnic or class groups in the city, organised around the theory of urban subcultures (Fischer 1975; 1976; 1982). In the United States the emphasis was on ethnicity, in work such as Suttles' (1968) and Gans' (1962). Conzen's (1976) study of the German immigrant community in 19th Century Milwaukee demonstrates that Germans were "an important and internally stratified subculture in terms of culture, class and occupation....'precisely because the Germans were so numerous and various, "community", to be effective, required definition in terms sufficiently broad and vague as to encompass and even draw strength from a diversity which permitted it to perform for Germans of all lifestyles its critical cushioning functions'" (Conzen 1976, p.156, quoted in Ray 1992, p.26-27). Recently, Lal (1986) has called for a return to the work of Robert Park and others of the Chicago School to recognise its importance in establishing urban processes as the basis for understanding human relations in the city, suggesting that Park's work could indeed be developed as a basis for understanding pluralism.

In British studies, the emphasis has been on social class rather than ethnicity as a dominant factor. Jones's The Social Geography of Belfast (1960) presents a classic study of the class effects of urban segregation along ethnic lines. The work of Ceri Peach and his students at Oxford from the 1960s to the 1980s produced an array of work on issues of pluralism and urban form (Clarke, Ley and Peach 1984). For Rex and Moore (1967), class hierarchy is inevitably a result of racial pluralism due to immigration. This observation provided an important background to more recent work which ties the emergence of racism in cities in developed contexts directly to the results of colonialism. John Rex's (1973) Race, Colonialism and the City is a pioneering work in this field which clarifies the role of ideology in the emergence of racism, and which identifies social construction as the process by which racial hierarchies are created.

In Canada there was a later, small scattering of studies in the tradition of urban ethnography. Jane Jacobs (1962) The Life and Death of Great American cities, makes a somewhat romantic plea for pluralism based on a physical environment enriched by immigration and other processes that create social diversity. Barry Wellman's 1970s study of social networks in East York, Toronto (Wellman and Leighton 1979) paints a similarly positive picture of urban immigrant life. More recent work on immigrant neighbourhoods has contributed a rich ethnographic tradition for a relatively small number of ethnocultural groups (Harney 1985; Hiebert 1987).

A second strand in the historical development of immigration research is associated with the development of quantitative analyses in geography. During the 1960s, with the introduction of computers and the computerisation of large data bases such as censuses, came the development of sophisticated multivariate techniques such as Social Area Analysis, Factorial Ecologies and Discriminant Analyses which were used to 'test' the assertions of the Chicago School. Spatial variations at the intra-urban level in the location of social groups and the character of different parts of the city came to be an overwhelming concern. These studies, largely descriptive in tone, lay the groundwork for the development of principles of social change in various areas of the city. These, in turn, led to the development of models of urban-social change.

Of particular interest is the work of Shevky and Bell (1955) on Social Area Analysis. They highlight three specific components of socio-spatial variation: social rank, urbanisation (family status) and segregation (ethnicity). Each of these constructs is underpinned by three variables, each of which can be measured from census data to the smallest spatial scale of census tracts. Their work has since received a great deal of criticism, but Social Area Analysis became the standard method for measuring social differentiation in cities. In the 1960s its use was largely superseded by the more technically sophisticated technique of Factorial Ecology, which can utilise a larger number of variables and,

In addition, it used precise statistical measures of similarity, such as correlation coefficients, to summarize the degree of association between each variable measured in every...census tracts. The heart of the method, the factorial phase, can be considered as a type of statistical induction. Each factor is a new mathematical vector that entails the rewriting of the data set in more concise form; summarizing a common source of variation in the similarity matrix. Each factor, therefore, can be considered as a quantitative measurement of a dimension or component of variation, or constructs lying behind the empirical reality of the individual variables. (Davies and Herbert 1993, p.42)

Factorial ecologies do not predetermine the number or description of factors to emerge from the rotation of the variables. Each variable has a different loading, or value, in relation to each factor. Thus, those variables that load highly on a factor can be used to define it. Also of importance in this technique is the fact that factor scores can be derived for each factor in relation to each spatial unit at which data are collected.

In Canada, factorial ecologies were still being conducted in the early 1990s (Davies and Murdie 1992) but they are largely out of fashion. One unresolved issue is that of the number of factors needed to capture adequately the level of social variation in urban areas, although it is generally agreed that there are more than three axes of social differentiation and also that they will vary from place to place and over time. A major Canadian study (Davies and Murdie 1992) attempted to resolve the problem of non-comparability of studies of different cities by conducting a factorial ecology of all 24 Census Metropolitan Areas using 1986 data. The study resulted in a nine factor solution, with one factor specifically related to immigrants' ethnicity. Ethnicity is a dimension typically found in factorial ecologies but its usage is fraught with difficulties. For example, because ethnic groups are spatially concentrated in particular areas Davies and Murdie measured ethnicity by the percentage of immigrants. They also only measured those of French ethnic origin and those of non-British and French ethnic heritage. Hence their results can only measure the extent of ethnic deviation from those of British origin. In order to measure specific ethnicities, other techniques, such as the Dissimilarity Index, must be used.

Ultimately, there are two major problems with quantitative approaches: first, while they are useful descriptive devices which can classify areas according to a range of dimensions they cannot deal with the dynamics or processes underlying these, including processes involving political power; secondly, they cannot cover elements such as attitudes or behaviour. This is not to suggest that quantification is not an important part of the analytical process, but the shortcomings of studies based entirely upon quantification arise largely because their theoretical focus is not on human conditions per se, but on independent patterns which have a putative regulatory effect upon human behaviour.

A third strand in the development of research on immigrant groups has resulted from the politicisation of the social sciences. This has seen two stages. In the 1970s, in particular, concern for social relevance and social justice (Harvey 1973), led to a radical interpretation of the urban space economy, although this interest initially was focussed on African Americans (Bunge 1971) and subsequently on the effects of central city decay and gentrification on neighbourhoods defined by class and race (Smith 1979). Although there is little or no reference in this literature to immigration, it is usually cited as the basis upon which later studies of the political economy of immigration rest.

More recently, since the mid-1980s, the critical turn in the social sciences with the development of postmodern modes of theorising and the central concern with questions of difference and marginality has led to a recognition of the need to untangle modes of oppression (and at the same time recognise that they are inherently intertwined and constitutive of peoples identities). Such work follows theoretically both from the politicisation of culture in the work of writers such as Sharon Zukin (1995) and from the attempt to theorise the city as an expression of the restructuring of capitalism (Castells 1996; Fainstein and Campbell 1996; Harvey 1996; Soja 1989). Recent developments have been depicted by many as a "crisis" (see Smith and Feagin 1995, esp. Waldinger 1995) epitomised by violence in Los Angeles and other North American cities, including Toronto, following the beating of Rodney King during the summer of 1992. Even more recently, there has been a return to recognizing the connection between immigrants and their countries of origin, with the development of theories of 'transnationalism' as a condition of globalisation (Basch et al. 1994; Glick-Schiller et al B. 1992) wherein destination cities are transformed both to accommodate established tradition and to negotiate a new presence in the face of practices that discriminate against and marginalise immigrants. Before exploring these developments in more depth, however, we turn to a discussion of some of the assumptions that underlie this review.

Some Conceptual and Methodological Considerations

Terminology:

1. Immigrant. There exists a technical definition of the term 'immigrant' which refers to a 'landed immigrant', or one who has attained permanent residency in Canada but who is not yet a citizen. We wish to emphasise, however, that 'immigrant' is also a socially constructed category which is used as a means of stratification. The popular conception of immigrants refers to people of colour who come from 'Third World' countries, who do not speak fluent English and who occupy lower positions in the occupational hierarchy (Ng and Estable 1987). White, middle-class professionals from Britain or the United States are not commonly perceived as 'immigrants'. Hence the notion of the term as a social construction; labelling citizens as immigrants divides them from other Canadians and places them - culturally, economically, socially and politically - in an inferior position.
2. 'Ethnic' is often used an a synonym for immigrant yet ethnicity is not a characteristic that is limited to immigrant or minority people. The bias in its definition is illustrated by frequent references in the academic literature to social relationships with members of the 'host' society, or in the press to 'ethnic communities'. In the province of Quebec, there is even an official Ministry of 'Cultural Communities', which refers euphemistically to those of neither French nor English ethnocultural background. This popular use of the term reveals complex relations that are dynamic and spatially variable and involve the attributes of ancestry, culture, religion and language; but the terminology also involves exclusion and inclusion and relations of dominance and subordination related to racism (Giles and Preston 1991). It is not therefore merely an ascribed trait: "...ethnic affiliations can be best understood, not as tendencies inherited from a common past, but as social responses to specific social and economic circumstances, opportunities and limitations" (Darroch and Marston 1987), however, the distinction between 'ascribed' and 'achieved' characteristics is less clear than it used to be. Not only does its definition and importance change over time and among places, but it is interpreted through a highly politicised lens -- tinted in Canada by the official policy of multiculturalism -- which allows the concept to be transformed according to particular interests.
3. 'Race'. Despite the long association of variables 'measuring' race and immigration, the association of the study of racism and that of immigrants is a relatively recent one. The term 'race' was used largely as a natural category until the 1960s when the use of the term 'ethnic' came to replace it. Since the mid-1980s, there has emerged a more comprehensive understanding of the term, with the recognition of its social construction as an expression of racist ideology, variable in both time and place to reflect specific conditions of subordination of people of colour (Kobayashi 1993; Satzewich 1990a,1990b). Racism is most effective when institutionalised (Anderson and Frideres 1981), and has specific effects upon the social geography of cities (Anderson 1991; Jackson 1987).
4. Spatial Scale. One recurrent paradox of many of the studies conducted by geographers is that their concern with space, place and location has resulted in studies which are conducted at a variety of spatial scales, including metropolitan areas, municipalities, neighbourhoods or census tracts. As a result, it is difficult to make direct comparisons of studies of the same immigrant group in different cities or even of different studies of the same group in the same city if they have been conducted at different spatial scales. It is here that we run into the problem of what is referred to as the ecological fallacy i.e., of assuming that findings that are found to be true at one spatial scale hold true at all spatial scales.