In our everyday lives, many of us use the language of class to refer to a social hierarchy and 'knowing your place' within it. Class is a matter of 'breeding' and of social background. It is reflected in our attitudes and our lifestyles, our accents, and our ways of dressing. For this reason, class has also come to be seen as something that is outmoded and old-fashioned. Class distinctions are tied to a world of tradition and subordination that no longer exists, and the language of class is incompatible with contemporary attitudes and values. Many journalists and social commentators hold this view. These writers see the persistence of class-ridden attitudes and the continued use of the word 'class' itself, as evidence that Britain has failed to adopt the more modern - or perhaps 'post-modern' - approach to life found in such apparently classless societies as the United States.

Sociologists, on the other hand, have more often used the word class to describe economic divisions and inequalities, especially those that are rooted in property and employment relations: in other words a particular kind of social division. This has led them to question whether Britain actually is any more, or any less, of a class society than the United States or the countries of continental Europe and other parts of the world. All these societies are unequal societies - they show vast and continuing differences in income, wealth, and property ownership - and so all can be described, in these terms, as class societies. Other differences of culture, attitudes, and life style are not disregarded, but they are seen as pointing to quite different kinds of social division. In sociological analyses, therefore, the economic relations of class are often contrasted with cultural matters, in particular with 'status', those more visible 'styles of life' led by people, and which affect their standing in the community. What non-sociologists describe in the everyday language of class, sociologists describe in the language of status.

Not all sociologists take this point of view, of course. There are few things on which sociologists are unanimous. For some, especially in the United States, class should, in fact, be seen as a matter of culture and identity, and not a matter of underlying material inequalities. According to these sociologists, cultural change has indeed undermined the power and relevance of the language of class. 'Class is dead', it is a thing of the past, even though economic and other inequalities may persist.

Even if it were true that class is dead, and that we wished to ague for an alternative framework of analysis, we would still need to know class was supposed to have been, before we could confine it to the grave of history. Unfortunately, as we have already noted, class does not mean the same thing to everybody. There is a considerable task in clarifying the idea of class, before we can make an informed judgement on its importance or lack of importance as a social division.

Indeed, in any field, linguistic confusion is a recipe for misunderstanding. If progress in sociological understanding is to be made, and if sociologists are to contribute to public debates, there needs to be some common ground in the language and concepts that they use. We need concepts that will help us to understand the many inequalities in resources, health, and education that exist in contemporary societies. This will help us to see how they relate to other social divisions, such as those discussed elsewhere in this book: in particular, gender, sexuality, ethnicity, and age. We must also agree about which aspects of these inequalities can legitimately be described and analysed as aspects of class relations. Only then would it be possible to investigate whether these class inequalities and class differences are increasing or declining in importance in particular societies. This kind of investigation raises very directly the relationship between the facts of class and their interpretation by commentators and by those who experience them in their everyday lives.

This chapter therefore begins with a conceptual exploration of the terms required to make some sense of this kind of social stratification. This exploration returns to the works of the classic sociologists. It does this not for reasons of 'ancestor worship', but in order to recapture the important distinctions that they recognised and that have been lost in many contemporary debates over class and stratification.

Social inequalities are central to any understanding of social stratification; but social stratification itself consists of more than simply inequalities in life chances. The concept of social stratification as a particular form of social division emphasises the idea that individuals are distributed among the levels or layers of a social hierarchy because of their economic relations. These layers or 'social strata' are real social groupings, forged together through both their economic relations and their associated social relations and interactions; groupings that are able to reproduce themselves over time. Work in similar occupations, marriage, kinship, and informal interaction connect individuals together and build up boundaries that close one stratum and divide it off from another.

They are not simply statistical categories defined in an unequal distribution of resources. They are not just the 'top 10 per cent' or 'bottom 30 per cent' of a distribution of income and wealth. Social strata are real social groupings. They are a particular form of what elsewhere (Chapter 1) has been called the 'categories' that comprise a social division. In this sense social stratification is a typical social division, but differing from others in that it is solidly based in economic relations.

The interconnections between social stratification and other divisions like gender, sexuality, ethnicity, and age are very complex. Social inequalities are invariably structured by all of these social divisions: inequalities are gendered, sexualised, racialised, and aged. On the one hand, each division has an independent importance in sociological analysis, as do relations of social stratification. Social exclusion and discrimination on the basis of gender and ethnicity, for example, generate inequalities in life chances.

On the other hand, this often operates through a process that combines exclusion in one division with exclusion in another, so that these other social divisions can become crucial conditions for the formation of social strata. This comes out clearly both in the other individual chapters in this book, and particularly in the chapters on health and community that deal with how social divisions interact in ordinary life. However, this does not mean that the social inequalities arising from other divisions should simply be equated with stratification. 'Social strata' as referred to in this chapter exist only when the economic and social relations of people tie them into larger, more complex and more permanent structures that take the form of a system of 'social stratification'.

To develop these arguments, I want to return to some distinctions made by Max Weber (1914; 1920) in his comparative investigations into economy and society. In two places in his work - in an early draft and in an incomplete re-draft of this - Weber set out an argument that has been massively influential in contemporary sociology. He identified three distinct aspects or dimensions to the distribution of power within societies. These are class, status, and authority or command. Weber's discussions concentrate on class and status and their relationship to 'party', which has led to the common misunderstanding that his three dimensions were class, status and party. His third dimension is better understood as 'authority', as I will shortly show. This misunderstanding, due also in part to problems of translation, is not however my main concern here (a more extensive discussion of this particular point can be found in Fulcher and Scott (1999: Ch. 15) while the rest of this section draws on Scott (1996: Ch. 2)). My main concern is to distinguish the core elements and how they relate to social stratification, as set out in the next two sections.

For Weber, class, status, and authority are aspects of power each of which has a separate effect on the production of life chances. He defined class relations as those that result from the distribution of property and other resources in capital, product, and labour markets. It is the possession or non-possession of economic resources that gives people a specific capacity or power to acquire income and assets and, therefore, to enhance their life chances in all sorts of ways. Company shares yield an investment income and can be sold at a profit on the stock market, while educational credentials or technical expertise may give people the opportunity to earn a higher income in the labour market. The distribution of economic resources, then, defines the various class situations that individuals can occupy in a society. A class situation exists wherever the distribution of resources is such that a particular category of individuals have similar abilities to secure advantages and disadvantages for themselves through the use of their marketable resources. A person's class situation is a causal component in determining both their life chances and the interests that they have in protecting and enhancing these life chances.

In setting out his view of class, Weber quite deliberately sought to build on Marx's work, much of which he took as valid. Marx saw property ownership, and especially property in the means of production, as the basis of class relations. For him, it was the ownership or non-ownership of factories, machines, and land that determined people's life chances and shaped their actions. Marx recognised just two principal class situations in contemporary societies, those of capitalists and proletarians. Capitalist class situations rest on the ownership and control of capital. Their occupants include industrial entrepreneurs, bankers, and landowners, as well as those who simply live on an income from company shares. Proletarian class situations, on the other hand, involve a lack of ownership and are based on the exercise of labour power as an employee. They include those of people involved in skilled work, manual labour, or office work; in other words, people involved in sets of similar types of occupations.

Debates largely within Marxism have modified this basic picture of two classes, leading to the recognition of lines of differentiation among capitalists and among proletarians, and to the identification of a whole array of 'intermediate' class situations. The proletarian class, being larger, has more obviously been sub-divided, and arguments about the relative merits of how this sub-division should be carried out, and which resulting classification scheme of the class situations is best, have been extensive. The most important of these debates in recent work are covered in the debates surrounding the work of Wright (1985; 1989; 1997). These debates are very much in line with Weber's suggestion that it is necessary to recognise a great variety of class situations in any society. In later sections, I will show how it is possible to build on these ideas to investigate contemporary class divisions.

One of the things that Weber objected to in Marxism was its economic determinism. He pointed out that non-economic factors were important, alongside the economic, in determining life chances and, therefore, in shaping patterns of social stratification. The first of these non-economic factors that he discussed was status. He saw divisions of status as originating in the distribution of prestige or social honour within a community. People judge one another as superior or inferior in relation to the values that they hold in common, and so a person's status is their standing or reputation in the eyes of others. When people act in conformity with these values, they build up a good reputation and, therefore, a high status in their community. Those who deviate from these values are accorded a lower status and may be excluded from the benefits and advantages given to others.

Most typically, a person's status follows from what Weber called the 'style of life'. The ways that people carry out the tasks that are associated with their occupations and their sex-gender roles, and the customs and practices that they follow as members of ethnic and other social groups define their particular and distinct styles of life (Chapter 4). Thus, manual workers, women or Asians may have a low status because of the way that their style of life is perceived and evaluated. 'Dirty' or routine work may be seen as undesirable or demeaning, a feminine style of life may be valued less highly than a masculine one, or the Asian way of life may be devalued.

What Weber called 'status' corresponds very closely to what in everyday life is often referred to by non-sociologists as 'class'. A style of life involves specific types of dress and bodily adornment, types and sizes of house, areas of residence, clothing, accent, methods of cooking and eating, and so on. These markers of social identity - symbols of status - are often of great importance. Much public discussion of 'class', however, tends to highlight status and to concentrate on its relatively minor and superficial aspects, so disregarding and disguising economic class.

However, for a more sophisticated and sociological analysis, we need to recognise that inequalities in life chances must be seen as reflecting the effects of both class and status situations. Status situation, like class situation, is a causal component in life chances. Class situations comprise the property and employment relations through which control over marketable resources is organised, while status situations are the communal relations through which prestige is given to a particular life style. Both operate to determine the life chances of individuals. An occupation, for example, involves both specific employment relations in the labour market and a particular level of occupational prestige (Parkin 1971). These two aspects of occupational position operate interdependently in determining life chances, and their separate effects can be difficult to disentangle. To claim that class and status should be seen as two separate and distinct dimensions of stratification is not to prejudge their relative importance. This is always an empirical matter. The distinction does, however, allow questions about the salience of class to be formulated with much greater clarity than is the case in much discussion of the subject.

Weber's third dimension of stratification is to be found in his discussions of authority and bureaucracy. Authority relations in states and business enterprises, he argued, involve relations of command in which one person is empowered to give orders to another. What Weber had to say on this can best be understood by introducing the concept of a 'command situation' to parallel his concepts of class situation and status situation. A command situation is a causal component in life chances that results from differentials of authority in formal organisations.

Weber's analysis of domination and authority gave him an acute understanding of the formation of ruling minorities in the top command situations of political, economic and other hierarchies, but, for all his insights, the conceptualisation of ruling minorities was one of the least developed parts of his sociology. His insights were independently developed, so far as social stratification is concerned, by Mosca (1896; 1923) and Pareto (1916: 1430). Their particular view was that the crucial social division was between a dominant minority and a subordinate majority. In other words, there were essentially two groups, a small one at the top of society, and a large one below it. They believed that this was an inevitable consequence of any social organisation of authority and that authority relations involve the formation of ruling minorities or what they called 'elites'. While we do not have to adopt their model of a society based on elites and masses, it illustrates how the delineation of class situation, status situation, and command situation is the crucial basis for any analysis of patterns of the social inequalities that comprise social stratification.

In the previous section, in distinguishing class, status and command in contemporary societies, I have already noted that occupations are central to the analysis of class and status situations. Occupations also lie at the core of structures of command. Occupations are positions in economic and social organisations in which each occupation has a different degree of authority, or lack of authority, over others. In all companies, government agencies, or charities, people in some occupations direct and manage the work of those in other occupations. Any particular occupation also rests on specific marketable skills or resources, and has a particular level of occupational prestige. A sociological analysis of the class, status, and command situations of contemporary societies is, for many purposes, a mapping of the occupational structures. Investigating social inequalities involves assessing the relative importance of class, status, and command in the life chances that are associated with particular occupations and groups of occupations.

Sociological research has used a variety of 'class schemes' to represent this, taking occupation as an indicator of class situation, in order to study classes as 'employment aggregates' (Crompton 1998: Chapter 4). This work is extremely important - and I make use of it below. It is sometimes held that the usefulness of particular class schemes and their advantages over others are to be judged on purely pragmatic grounds: whichever has the best predictive capacity for dealing with whatever is being studied) is to be preferred (Erikson and Goldthorpe 1993. See the criticism in Morris and Scott 1996). However, the implication of my argument is that class schemes must by judged on theoretical as well as empirical grounds, and that they must be seen as more or less adequate attempts to offer a coherent and operational concept of class. It is Weber's framework that offers the best basis for this latter requirement.

A system of social stratification can be thought of as consisting of a number of separate social strata, layered one above another. Each social stratum contains people in similar distinctive class, status and command situations. Each stratum is a social group with boundaries marking off those inside from those outside. The group contains a cluster of social positions, with individuals moving both between these social positions and also interacting (or 'associating') with each other as they do so. They may modify their own and their children's social position by marriage or changing jobs ('circulating' within the boundary of their group). Through kinship and close intimate interactions such as leisure-time socialising and club membership, they associate with other people like themselves who make up their social stratum. For these reasons, both occupational mobility and patterns of association must be seen as central to strata formation.

These movements, on the one hand of individuals from one occupation to another, and on the other hand of interactions among those engaged in particular occupations, define the key boundaries within a system of social stratification. Occupational categories are a part of the same social stratum if there is easy and frequent movement between them. This might involve the lifetime mobility of individuals between types of occupations (called intra-generational mobility), or it might involve occupational movement between generations, e.g. from father to son (inter-generational mobility: see Payne 1989). Equally, it might involve frequent intermarriage between those in similar occupations, or their interaction with one another in leisure-time activities. Whenever these relations of circulation and association reinforce one another in such a way as to create regular and established patterns of connection among the people working in a set of occupations, the connected occupations form part of a single stratum.

Occupations fall into the same social stratum when they are connected through chains of frequent and relatively easy circulation and association. They fall into different social strata when they are connected - if at all - only through infrequent circulation and association. Boundaries between social strata exist wherever patterns of circulation and association produce divisions between categories of occupations. In sociological research, individuals are allocated to the occupations that combine together various class, status, and command situations; rates of mobility and social interaction must be examined in order to uncover how those occupations are clustered together into social strata.

Marx held, for example, that the occupants of capitalist situations form a distinct social stratum because they circulate freely around the different forms of capital, because they are involved in extensive networks of intermarriage, and because their children are able to inherit accumulated capital and to enjoy the advantaged life chances that it generates. In the same way, he saw occupants of proletarian situations as forming a separate social stratum because they move from one type of work to another similar type, and from work to unemployment, they marry other workers, and their children have no choice but to try to enter paid employment.

One perennial topic in studies of stratification has been the question of whether the individual or the family household should be taken as the unit of analysis (Abbott and Sapsford 1987; Goldthorpe 1983; Scott 1994a). This debate has been fired in recent years by feminist criticisms of those conventional approaches to stratification that have allocated women to social classes according to the occupational position of their husbands (Stanworth 1984). Proponents of this approach claim that women's class attitudes and behaviour can be adequately 'approximated' by this method (Erikson and Goldthorpe 1993). This pragmatic argument offers little theoretical purchase on class, and its critics have rightly argued that women's work must be considered independently of men's work and must play an equal part in determining positions in a system of social stratification.

When considering the class, status, and command situations of individuals, it is of the utmost importance that each individual is classified on the basis of his or her own occupation. However, this is not the end of the matter. When considering the formation of social strata (rather than just the occupancy of class, status, and command situations), the social relations among members of family households cannot be ignored. Marriage connects or divides occupations from one another, and the question of which social stratum a person (man or woman) belongs to cannot be decided without reference to their various family relations. Households are formed through marriage or cohabitation, and it is within households that the educational and other opportunities that shape occupational mobility are generated. Households are also bases for the organisation of much free-time interaction. In all these ways, household membership is central to the allocation of individuals to social strata.

Class, status, and command operate alongside one another, but their relative importance can vary a great deal. In contemporary capitalist societies, as we saw above, these all operate in and through the occupational division of labour and the associated system of property relations. One of the central features of the development of capitalist societies has been the way in which class situations have become, through the occupational system, the fundamental determinants of all other aspects of stratification. Status relations and command relations are, to a considerable extent, consequences of or reflections of class relations. It was for this reason that Weber thought it appropriate to describe contemporary capitalist societies as 'class societies'. A class society is one in which class situation is the most important determinant of the life chances of individuals. Weber contrasted class societies with 'status societies', where the most important determinants of life chances and the overall pattern of social stratification were status situations. Similarly, we can recognise 'command societies' where command situations are of paramount importance. These distinctions are important for comparative investigations, but they need not concern us here.

In a class society, social strata take the form of what Weber called 'social classes'. A social class is a cluster of households whose members owe their life chances principally to their property ownership or employment relations. It has been argued that the very language of 'class' developed in the eighteenth and nineteenth centuries as an attempt to grasp the new social relations that emerged with the increasing salience of class situations in capitalist industrial economies (1960). The language of class was central to the recognition of 'social classes'. By the early nineteenth century, a recognition of capitalist societies as containing three social classes - an upper class, a middle class, and a working class - had become widely recognised.

The final conceptual point that needs to be made concerns class consciousness and class identity. All of us attempt to interpret and to understand the social world in which we live, and a central part of this social understanding is the attempt to interpret the social inequalities that we experience. Where people have life experiences in common, they are likely to develop a shared awareness and outlook on those experiences. Those who occupy the same class, status, or command situation and the members of a social stratum, then, are likely to develop a shared understanding of patterns of social stratification and the boundaries of social strata. This common awareness is most likely to result to the extent that people work together, live in the same neighbourhood, and engage in the same leisure activities. It is also influenced by the imagery conveyed in the mass media, though this is often mediated by personal, face-to-face contacts. This consciousness of social stratification (whether strong or weak) forms a central part of overall images of society, those cognitive maps that guide us in our relations with others.

What Marx and Weber called 'class consciousness' is that form of class awareness that develops in social classes and that, at its fullest, defines the interests that their members have in maintaining or enhancing their life chances. This will shape their political outlook and may lead them to form what Weber called 'parties' - conflict or interest groups of various kinds that give voice to their interests in political struggles. It is also likely to shape individual political action, and even in the absence of strong class-based parties, a class consciousness may, for example, bring about a close association between social class position and voting.

The formation of a class consciousness is by no means automatic, as the awareness of social stratification may be only weakly developed or may be outweighed by other elements in social experience. One important aspect of investigations into social stratification, therefore, is the extent to which people's social imagery actually does centre on a sense of class identity. This means that we can add to the concepts that we have set out so far - principally those of social stratification based on economic relations, the differences between class, status and command situations, and the way occupations have been used to define strata - the ideas of identity and actions.

Mapping classes

We observed earlier that to understand class fully, we needed both clear concepts and a coherent way to operationalise them in research. The ideas just discussed have provided a not always explicit framework for the extensive discussion about how the class situations and social classes of contemporary Britain can be mapped in clear and reliable categories. Many different official and unofficial schemes have been proposed, by sociologists and non-sociologists alike. In these schemes, both the number of classes and their composition vary quite considerably. In most cases, the theoretical basis of the scheme is not spelled out, and few have been constructed along explicitly Weberian lines. It sometimes seems as if there is a huge gap between stratification theory and empirical research on social stratification, with the latter showing a bewildering variety of class maps. This has led some sociologists to suggest a radical revision of how stratification is conceptualised and measured. The Cambridge approach, for example, sees stratification as involving a continuous scale rather than discrete categories (Prandy 1991; Blackburn and Prandy 1997).

The range of disagreement can be seen quite clearly in the differences between two of the leading class schemes, those of Wright (1985) and Goldthorpe (1980). Wright's scheme - the most developed of a number that he has produced - contains 12 distinct class categories. These categories are defined by ownership and control of property, by organisational assets of authority, and by skills and education. These 12 categories are seen by Wright as a half-way house between the detailed listing of class situations and a briefer listing of real social classes, and they were intended to grasp broad differences in life chances in a large comparative study. Marshall et al. (1988) applied Wright's scheme to British data and showed that the classes ranged from a 'bourgeoisie' of property owners, accounting for 2.0 per cent of the population, to a 'proletariat' containing 42.9 per cent of the population. Between this bourgeoisie and proletariat was a whole range of propertied and non-propertied classes covering those working in various kinds of managerial, supervisory, and technical work.

Goldthorpe's scheme was an attempt to move closer to a true social class map, and he allocated class situations to seven separate categories. The extremes of this scheme were a 'service class' containing 25 per cent of the population and a class of non-skilled manual workers containing 22 per cent of the population (Erikson and Goldthorpe 1993). Unlike Wright, Goldthorpe did not incorporate substantial property ownership into his scheme, on the grounds that there were too few large property holders to figure in any national sample survey. Nevertheless, he recognised that, for some purposes, it was sensible to try to distinguish an 'elite' of property owners that would amount to a fraction of one per cent of the population. Goldthorpe has also modified his basic scheme in order to use it in the wide range of very different societies that he has investigated in comparative research. This enlarged scheme is less useful for investigations of any one society.

Neither Wright nor Goldthorpe can be taken as having produced a completely accurate map of the class structure of contemporary Britain. Nevertheless, a wealth of secondary evidence suggests that Goldthorpe's basic scheme - modified by the inclusion of his eighth 'elite' category - does provide an approximation to an accurate mapping of contemporary social class boundaries.

The older and rival 'official' (i.e. government-produced) schemes, however, are still widely-used for a whole variety of purposes related to the investigation of life chances and life styles, and they are likely to continue in use for some time. In much of this chapter, these categories will therefore be the ones used. In many studies and especially in official statistics, two alternative sets of 'official' class categories have been used. These are the so-called 'Registrar General's scheme' and the SEG (Socio-Economic Groups) scheme. These are far from having the plausibility of the Goldthorpe scheme (not least because their origins and subsequent changes of detail were not grounded in a systematic sociological analysis), but they are nevertheless useful bases for mapping contemporary differences in life chances and life styles. Their particular attraction is that for many years they were the only schemes available to access data collected by the governments large-scale social surveys.

A significant advance in techniques of sociological research has recently been made with the introduction of a new official social class scheme that extends the Goldthorpe classification and integrates it with the Registrar General's and the SEG schemes. This scheme is the SEC, produced by David Lockwood and David Rose (see Rose and O' Reilly 1998), and it will soon allow a far more effective integration of official and sociological research into social class. Until it begins to be used in research, however, we must make do with the existing schemes.

The basic Registrar General's scheme contains five social classes, one of which (class III) is divided into manual and non-manual sections. In its most developed form, therefore, the Registrar General's classification contains six social classes. This scheme attempts to see class differences in terms of the skill and employment relations of occupations, and it combines this with an assessment of the social standing of the various occupations. The 17 category SEG scheme is more systematically economic in its basis, and in its 'collapsed' version also contains six social classes. These boundaries, however, are drawn at different points and they do not correspond to the Registrar General's categories in a one-to-one way. The SEG scheme has sometimes been converted into a seven category scheme by the subdivision of white collar workers by their grade. With numerous qualifications, it is possible to summarise the relationship between the various schemes as in Table 2.1.

[Table 2.1 about here]

We can use these class schema to map the shape and size of classes, and how these evolve. Class relations have not remained unchanged in British society. The structure of class relations itself is subject to change, and these patterns of change cannot be reflected in a fixed scheme of social classes. Nevertheless, the use of these categories in historical research does give some indication of how class relations have altered. A key element in this is the declining significance of the subordinate classes of manual workers. The subordinate classes amounted to over 80 per cent of the population in 1911, and in 1951 they still accounted for almost three quarters of the population. By 1981, however, they represented only just over a half of the population. Ten years later, they made up just 45 per cent of the population.

The decline in the number of manual workers, both skilled and unskilled, has been matched by a growth in the number of non-manual workers employed in the new service industries that have grown massively over the course of the century. Between 1911 and 1951, non-manual workers increased from just over 12 per cent to around a quarter of the population. By 1991 they made up 55 per cent of the population.

These figures, like all official statistics, must be treated with caution. However, they do give an general impression of the principal change in the system of social stratification. The key to understanding this change is what Payne (1987a; see also Clark 1940) has described as the 'occupational transition'. This term refers to the shift in occupational structure that occurs with developments in technology and the division of labour. Industrialism involved a shift from a structure of primary sector occupations (those in such industries as agriculture and mining) to one in which secondary sector occupations (those involved in the manufacturing of machinery and consumer goods from the products of the primary sector) played the leading part. Developments in manufacturing technology, however, led to the growth of technical and specialist occupations that require high levels of education and training. These occupations grew at the expense of purely manual work and, at the same time, there was an expansion in the tertiary sector of public service and commercial occupations concerned with distribution, banking, and insurance. The occupational transition, then, is one in which, successively, the primary, the secondary, and the tertiary sectors have been the most important bases of occupational differentiation.

The wider implications of these changes in the occupational division of labour have been widely discussed. For many observers, the changes point to a long-term development from an 'industrial' to a 'post-industrial' society, while others have described it as one aspect of a transition from a 'modern' to a 'post-modern' society. Whatever might be concluded about this, it is clear that the occupational transition has been associated with the development of forms of social structure in the second half of the twentieth century that are radically different from those that existed in the first half. The old pattern of social classes, their boundaries, and their forms of consciousness have all been transformed. The large, cohesive social classes that existed in the early twentieth century have developed into the more fragmented and divided classes that are entering the new millennium.

These changes in class relations will be discussed later in this chapter, but I would first like to look in some detail at the distribution of life chances across the social classes as they exist today. This will establish a benchmark from which the social changes of the twentieth century can be assessed. The data presented here come from the recent and valuable compilation produced by Reid (1998), where the evidence is discussed more fully.

Tables 2.2 and 2.3 show the distribution of money, work, and property, as measured in relation to class situation. In each case, a sharp class gradient is apparent. Disposable income, for example, ranges from the £519 per week earned by professionals to the £213 per week earned by unskilled manual workers. The more detailed categorisation used for the distribution of income shows that the middle levels of the scheme do not form a neat and clear-cut hierarchy for all measures. Skilled manual workers, for example, earn slightly more per week than routine non-manual workers.

[Table 2.2 about here]

A similar gradient is apparent for the ownership of shares and for membership in a pension scheme. Level and continuity of income determines the chances that a person has to purchase property. In addition to inequalities in income-generating shares, there are marked inequalities in home ownership. Table 2.2 shows that the rates of home ownership are much lower for semi-skilled and unskilled workers than they are for all others. The rate for unskilled workers is half that found among professionals. Gender differences have rarely been explored in relation to class, but Table 2.2 brings out its importance. The proportion of women in pension schemes is lower, for each class, than it is for men.

A national survey in 1996 found that most people thought that a weekly income of £211 was necessary to avoid poverty, and it can be seen that the average for unskilled manual workers is only just above this level. Unemployment is a life experience that can push people into poverty, and Table 2.3 shows that unskilled workers are especially likely to become unemployed.

[Table 2.3 about here}

Rates of unemployment are almost five times as high for these workers as they are for professionals and managers, and unskilled workers are also far more likely to remain unemployed for a year or more.

These class inequalities are reflected in the inequalities of health, life, and death documented in Tables 2.4 and 2.5.

[Table 2.4 about here]

Again, a clear - though far from perfect - class gradient is apparent. The proportion of babies with low birth weight, rates of infant mortality, rates of child mortality and life expectancy all show the same class gradient as found in income and property ownership, and these differentials in life chances persist into adulthood. Even less crucial aspects of health like long-term illness, eye, ear, and tooth problems (see also Chapter 10, below) have this gradient, as Table 2.5 shows.

[Table 2.5 about here]

In addition to these demographic life chances, class situation appears as a clear determinant of both housing and education. The proportion of those living in overcrowded housing and the proportion without central heating both show the familiar gradient (Table 2.6).

[Table 2.6 about here]

We also find the same pattern in the measures of educational achievement. Rates of attendance at private schools, where attendance depends on income, are predictably unequal, but so also are rates of higher education and the proportion of those leaving school with no formal educational qualifications of any kind.

[Table 2.7 about here]

Inequalities of life chances in education of the kind shown in Table 2.7 are especially important, as these are central to the reproduction of class relations. Chances of entering highly-paid occupations are closely related to educational achievement, which is, in turn, related to class situation as measured by occupation. This is the most striking result of research into social mobility (the movement between classes), the consequences of which are shown in Table 2.8 These data, from the 1972 Nuffield study, show high levels of self-recruitment in each class. Over a half of men born with semi-skilled or unskilled fathers remained in manual work, with only 16 per cent entering professional or managerial work. Conversely, almost two-thirds of those born into professions and management families remained there, only 12 per cent taking up manual work. Despite apparently high levels of social mobility, class inequalities in chances of mobility remain strong, and they closely follow the other documented inequalities in material life chances.

[Table 2.8 about here]

These figures on inequalities in life chances are illustrative only -such inequalities are documented more comprehensively in Reid's (1988) own summary. My purpose in using them is to demonstrate that class situation, however this is measured, remains an important determinant of life chances in contemporary Britain. There is, for a number of key measures, a clear class gradient, and this has to be recognised in any discussion of the relevance of class analysis. Class differences persist and have, in many respects, become sharper. However, these are far less directly reflected in distinct differences of social status, and so they are less directly reflected in sharp differences of attitude and outlook. Class consciousness - at least as conventionally understood - is no longer a central feature of contemporary class relations. Traditional status values, that for so long defined the character of British class relations, have decayed. This has undermined the everyday use of the language of 'class', which I showed to rest, in fact, on 'status' differences in accent, dress, and social background.

Cannadine (1998) has convincingly shown that the language of stratification in Britain has drawn on three distinct, and often contradictory images. These he terms the hierarchical, the triadic, and the dichotomous. People draw on these to varying degrees and according to their social position (Lockwood 1966). They are rhetorical and interpretative devices that allow people to construct an understanding of the complex social structures in which they live. When thinking about the social stratification of their society, people are 'silently and easily shifting from one social vision to another' (Cannadine 1998: 165).

Hierarchical imagery has been dominant in social thought, especially in the form of a status hierarchy focused around traditional distinctions of education, dress, accent, ancestry, and other aspects of style of life. This imagery began its long-term decline in the inter-war years, when a tripartite imagery of upper class, middle class, and working class consolidated its position in the popular mind. For many in the working class of subordinate manual workers, this shaded into a dichotomous opposition between 'us' and 'them'.

In the second half of the twentieth century, social imagery became more complex and less clear-cut. The traditional imagery continued to decline as traditional and deferential values generally decayed with the decline of Empire and the sense of 'Greatness' that went with it. The growth of consumerism and the encouragement of a greater diversity in life styles fragmented the tripartite and dichotomous images and encouraged the formulation of new hierarchical images in which spending power and consumption were the central elements. Such 'money models' of society established new and highly flexible forms of status division that mask, rather than solidify, underlying differences in economic resources. Class, as defined by Marx and Weber, remains the crucial determinant of life chances, and it shapes the opportunities that people have for pursuing particular life styles, but it is less visible to people, who tend to define their positions in status terms.

Having sketched the persistence of class differences, it is now possible to look in more detail at each of the classes in turn. This will involve taking a broadly historical perspective, examining the most important social changes that have taken place at the various levels of the system of social classes, as a means of understanding contemporary patterns and also recent debates about those patterns. It is not possible fully to understand class today without grasping how the classes have evolved, a social process that is arguably the key change in society. We will start with the subordinate, manual classes, followed by the intermediate 'middle classes, before discussing the 'advantaged' or upper classes.

The employment relations that define the social position of the subordinate manual worker are those of the wage labour contract. Manual workers are engaged to work in exchange for a wage that is supposed to compensate them for their labour time. Most typically, this kind of work has been paid on a weekly basis, though large numbers of manual workers are employed on a part-time or casual basis for shorter periods. The limited autonomy allowed or expected for manual workers is apparent in the fact that their wages are generally calculated on an hourly basis, with employers exercising controls over both the time and the pace of their work.

The market situation of the manual worker, then, depends on the skills that they can bring to the labour market. Those skills that are in the greatest demand are likely to command a higher rate of pay than skills that are less in demand, and a significantly higher rate than unskilled labour. The major source of internal division within this class, then, has tended to be along the lines of skill (see Scott 1996 for an elaboration of this argument).

The employment relations of manual workers also involve them in distinct work situations. Within the occupational division of labour of the enterprise that employs them, they are in a subordinate position. They are subject to managerial authority over all aspects of their work. Manual workers in large organisations, such as car plants, engineering works, chemical plants, local authorities, banks, and so on (and formerly shipyards and coal mines) are subject to an extensive managerial hierarchy, the top levels of which may be far removed from them both socially and physically. Manual workers in smaller organisations, such as much light industrial and building work, farm work, and domestic service, are employed in smaller work groups and may have closer and more immediate contact with the managers who supervise their work. Cross-cutting the lines of skill, then, is a secondary line of internal division between those involved in large-scale and those in small-scale work situations.

Similarities in their conditions of employment - their involvement in a wage labour contract and their subordination to managerial authority - form these workers into a distinct class at the economic level. Unless lines of division by skill and work situation differences become particularly sharp, workers will form what Marx referred to as a 'class in-itself'. They are united by the shared economic conditions of propertylessness and subordinate employment, and by the life chances with which these are associated. Where internal divisions become marked and are reflected in differences in life chances - and I have documented some evidence for this in the previous section - it may make more sense to identify a number of different subordinate classes, or sub-classes, at the economic level.

The identification of economic class boundaries, however, is merely the first step towards the identification of social class boundaries. Those in various economically-defined class situations form a single social class, it will be recalled, whenever mobility and association are easy and frequent among them. When a class in-itself is forged into a social class through the relations of mobility and association of its members, it is on the way to becoming what Marx described as a 'class for-itself'. In its fully developed form, a class for-itself has not only a real existence as a clearly-bounded social class, but also has a consciousness of its own position and of its interests relative to other social classes. The class for-itself is most likely to act politically in pursuit of its collective interests.

For much of the nineteenth century, manual workers formed neither a class in-itself nor a class for-itself. They were, for the most part, employed in relatively small workshops where there were few opportunities for them to develop a consciousness and solidarity of themselves as members of a single class. At the same time, they were greatly divided by differences of skill (Thompson 1968). From around the middle of the century, however, the scale of production began to increase and workers were brought together in larger and larger productive settings. The discipline of the factory became the common experience for manual workers (Foster 1974). Within large-scale factory systems of production, the occupational division of labour brought workers with different skills into closer contact with each other, and many forms of craft work underwent a process of de-skilling. Manual workers, while still differentiated by their skills, became more homogeneous in terms of both their market situations and their work situations.

This gradual forging of manual workers into a class in-itself was matched by high levels of mobility and association across the lines of skill within the working class during the nineteenth and early twentieth centuries. Manual workers were sharply separated from non-manual workers, and they were internally unified as a distinct social class. Very few manual workers or children of manual workers entered non-manual work, even at its lowest levels, and there were very few marriages across the manual/non-manual line (Savage and Miles 1994). On the other hand, skilled workers whose jobs disappeared with technological change had little option but to take up unskilled work. Consequently, unskilled workers and their children could sometimes enter semi-skilled work through on-the-job training, especially in times of economic growth and relative prosperity.

These relations of mobility and intermarriage closely linked workers in a particular locality to one another, though they were far less closely linked to similar workers in other localities. As a result, the great industrial towns and inner cities each became distinct, but structurally similar, communities of families dependent on manual work. It was in these conditions that the idea of a 'working class' took root. The proletarian communities were fertile seed beds for sentiments of class solidarity and for the forging of a shared consciousness of class.

By the inter-war years of the twentieth century, this process was well-developed. Contemporaries could refer to 'the working class', and everyone would know what this meant. Manual workers themselves embraced the label whole-heartedly and took it as their basic marker of social identity. The social institutions and agencies of the working class community sustained this identity. The dense networks of kinship and friendship that tied together neighbourhood and place of work (Young and Willmott 1957) had a more formal expression in the shared leisure-time involvement in pubs and clubs, cinemas, and dance halls with which they were associated (Dennis et al. 1956; Jackson 1968), and in participation in trades unions, co-operatives, and the Labour Party. Through the first half of the century, the electoral strength of the Labour Party grew in parallel with the crystallisation of these working class communities. Support for the Labour Party was seen as a natural reflection of the stronger and deeper-rooted local solidarity of a working class community.

The degree of unity found in the working class until the middle of the twentieth century should not be exaggerated. Skill differences persisted, these were associated with differences in life chances, and they could often undermine class solidarity. Highly skilled male craft workers - those who had served an apprenticeship or some similar period of training - could generally command higher wages and had greater job security than their semi-skilled and unskilled counterparts. Many were also able to exercise a degree of authority at work as the foremen and supervisors of other manual workers. Their relatively high wages and their involvement in authority meant that there was a tendency for them to see their interests as distinct from those of unskilled workers (Gray 1981). As an 'aristocracy of labour', they were especially likely to be committed to ideals of self-help, getting-on, and respectability. One aspect of this respectability was that their wives withdrew from the labour market and took on full-time responsibility for domestic work and family obligations.

In a similar way, the persistence of small-scale production, often in small towns, meant that large numbers of manual workers were isolated from the mainstream of the labour movement. Their work situation encouraged a greater sense of identification with - or, at least, of accommodation to - their employers and managers. Isolated from the supportive institutions of the oppositional sub-culture, they were more likely to defer to the authority of their employers (Stacey 1960; Newby 1975).

At the other end of the scale, there were large numbers of manual workers who lived in poverty and experienced long periods of unemployment. Casual work, a lack of skill, old age, and illness all made people susceptible to economic downturns that could push them below the poverty line (Booth 1901-2; Rowntree 1901). Poverty was the fate that could all-too-easily befall any manual working family. Once forced into poverty, their lives were degraded and demoralised, and they became the targets of moralising attack by respectable society - manual and non-manual alike - which saw them as the 'rough' and 'undeserving' members of society. The poor were not, however, a distinct social class, they were the lower level of the working class. The rise and fall of families across the poverty line ensured that there was no sharp separation between the poor and the rest of the working class.

The mainstream of the proletarian working class, therefore, coexisted with the small-town working class of deferential workers and workers who aspired to improve themselves within the system (Lockwood 1966; Parkin 1971; Bulmer, 1974). The lower levels of this working class formed a poor working class that amounted to 30 per cent of the whole population at its peak. The cultural division between the 'rough' and the 'respectable' working class, rooted in real economic differences, became a persistent theme in national politics.

Although there have been great changes, not everything has changed significantly since the middle of the twentieth century,. Many of these changes had their roots in the early years of the century, while others are only just beginning to make themselves felt. It is undoubtedly the case, however, that the mid-century marked a watershed in the history of the working class.

Manual workers, as I have shown, have come to account for a much smaller proportion of the population than was the case a century ago, and the economic boundary between manual and non-manual work has become less sharp and less salient. Changing technology has transformed much manual work. It has led to the collapse of whole industries and, through the resulting migration of workers in search of work, it has undermined and caused the disappearance of the stable working class communities with which they were associated. The sociologists who first commented on these changes tended to see them as a process of 'embourgeoisement': a process in which manual workers, and society as a whole, were becoming more 'middle class' in character and outlook and, as a result, were abandoning old patterns of class solidarity and class politics (Zweig 1961; Klein 1965). These claims involved a serious over-statement of their case, and many of the changes that were occurring were misinterpreted. They did, nevertheless, highlight the quite critical changes in class structure that were taking place.

Manual workers were not becoming middle class, but many of them were becoming far better-off than earlier generations. The 1950s and 1960s saw a real growth in manual worker incomes, and the general improvement in economic conditions allowed many more people to buy domestic consumer goods that were, in the past, available only to the privileged or were simply not available at all. 'Affluent' workers were able to buy houses, washing machines, cars, and televisions, and they were able to furnish their homes in much more comfortable ways than before.

This partial homogenisation of standards of consumption between manual and non-manual workers, however, was not matched by changes in their property and employment relations or in their life chances more generally. The conditions of work for manual workers remained inferior even to those of routine non-manual workers. Their work was dreary and monotonous, it was not a source of intrinsic satisfaction, they had less security of employment, inferior pension and holiday provision, and few opportunities for promotion. They had often to move long distances to obtain this work. It was tolerated only because of the size of the pay packet - the higher their pay, the greater was their willingness to tolerate their inferior working conditions (Goldthorpe et al. 1969. See also Goldthorpe and Lockwood 1963; Goldthorpe 1964; Lockwood 1960; Devine 1992).

Most importantly, consumer goods were acquired because of the opportunities that they opened-up, and not because they were the status symbols of a middle class life style. It is still worth making the obvious point that people buy washing machines to wash clothes, televisions to provide entertainment, and cars to provide them with mobility. These things are not accumulated simply in order to make a claim to higher status. Indeed, status divisions, like class divisions, remain sharp. Manual workers do not regularly and frequently entertain non-manual workers at home, nor do they engaged in many common forms of leisure-time associations with them. They have, in fact, been forced to become more 'home-centred', more concerned with the life chances and life style of their own family, as they no longer have extensive working class communities to support them in times of need.

'Affluence' was a cyclical phenomenon, not a permanent state, though levels of consumption have remained much higher than in the past. For a significant minority, however, poverty conditions have persisted (Townsend 1979). For some commentators, the poverty of the post-war period has been seen as indicating the emergence of an 'underclass' separate from the social class of subordinate manual workers (Murray 1984). The evidence does not support this claim.

In the first place, poverty is associated with other social divisions, particularly gender, age and ill health, as we shall see in Chapters 3, 6 and 10. Equally, as in the past, poverty remains an ever-present possibility for all manual workers, and its incidence is generally beyond their individual control. The collapse of a firm or an industry in an area of high unemployment can throw workers at all levels of skill into protracted periods of poverty. On the other hand, those living in poverty are not necessarily condemned to a life of poverty. Many are able to obtain employment or better-paid work after a period of poverty, and there is, overall, a substantial turnover among those in poverty (Morris 1995). Those who do persist in poverty are, in particular, the aged and the infirm, whose families and friends often do not themselves live far enough above the poverty line to help them out. Poverty is a real and important problem in contemporary society, an inevitable consequence of the way that it operates, but the poor do not form an underclass that exists separately from the rest of the class of subordinate manual workers (Scott 1994b). Poverty is not a class situation per se, although aspects of it can be thought of as arising from many very poor people's lack of opportunities to sell their labour in the market (Payne et al. 1996).

Changes in consumption and the spending of income have largely concerned the extent to which a class of subordinate manual workers can be regarded as forming a class for-itself. Their market and work situations, and their associated life chances, remain distinct from those of other workers, and patterns of mobility and association still form them into a distinct social class, but these ties are neither so strong nor so intense as they were in the past. There is not the institutional support for the degree of class solidarity and cohesion that characterised the period before the Second World War and that persisted into the 1950s.

Instead of a class consciousness rooted in common cultural and political traditions and in participation in the labour movement, manual workers are today more likely to see their society, and their own position within it, in terms of a more open and flexible 'money model' of society that stresses relative spending power rather than political power. A focus on consumption and spending has made traditional class and status distinctions less relevant and has replaced them by new distinctions rooted in differences in consumption. To the extent that they see the language of class as relevant -and most will use it if given some encouragement by a researcher - it is not regarded as the sole or even the most relevant way of describing themselves.

People are likely to report themselves as being members of a large class that contains virtually the whole of society, only the very rich and the very poor falling outside it; i. e. a tripartite model. Within the main class, people make distinctions on the basis of the income and standards of living that they believe people to have achieved. Their knowledge of the true extent of income differences, however, is generally very poor, and the correspondence between real conditions and perceived conditions is slight. Most manual workers feel that those who are better-off than themselves are not that much better-off, and so they look for only relatively modest improvements in their economic situation. They see the less-well off as being not much worse off than themselves, often tending to be unsympathetic towards welfare payments. For some, this central class is described as a 'middle class', while for others it is a 'working class'. For yet others - perhaps for a growing number - it is a sign of 'classlessness': if almost everybody is in a single class, then 'class' distinctions are no longer relevant.

Subordinate manual workers, therefore, no longer live class in the same way as their counterparts did in the working class communities of the past. This makes it more difficult to use the term 'working class' to describe them. They are more fragmented social class than before, the differences between 'rough' and 'respectable', skilled and unskilled, proletarian and 'deferential' having become more visible and, at the same time, submerged in a larger pool of social identities. Despite the inferior life chances that were documented in the previous section, manual workers have neither the solidarity nor the consciousness that is needed to sustain a strong commitment to collective class action. This declining sense of class identity is one of the most important factors underlying the decline in the traditional Labour vote from the 1940s to the 1990s. If a greater percentage of manual workers supported Labour in 1997 than at any other time in the post-war period, this is evidence not of a resurgence of class consciousness but of the transformation of the Labour Party itself. New Labour's commitment to the 'classless' society commits the party to mobilising the support of those who see themselves in the large central class through pursuing the politics of the centre.

Just as the subordinate classes can be thought of as either a single block or as sub-divided, the middle classes cover a range of positions. The three main segments of the intermediate classes comprise those who are employed in professional, managerial, administrative, and technical occupations; those who are self-employed; and those with small-scale property. The employment relations of the latter are distinguished from those of subordinate workers by the much greater autonomy and involvement in authority that their work situations involve and, for those at the more senior levels, by the possession of a longer-term 'service' contract of employment rather than a time-based labour contract. They are contracted to employ their skills in the service of an employer - generally a large-scale organisation - and they have a great deal of autonomy and discretion as to when and how they provide these services (Goldthorpe 1980). They are paid on a monthly, salaried basis, and they have significant additional benefits over and above their basic salary.

Self-employed workers, such as lawyers, accountants, doctors, and many business consultants, whether working alone or in partnership, exchange their professional services for a fee paid by their clients. By virtue of their self-employment, they, too, have considerable discretion and autonomy over the conditions under which they carry out their work, and they have much leeway in determining the level of their own fees. Small-scale property owners involved in the running of manufacturing or service businesses may, legally, be employees of their own businesses or they may be self-employed. In practice, however, it is their property that defines the petty bourgeoisie's market situation and allows them to earn an income.

These differences in market situations have meant that the intermediate classes have never had quite the degree of economic homogeneity achieved by subordinate manual workers. Nevertheless, there was a sharp economic gulf between the two classes. The intermediate classes were more closely linked to each other than they were to those who stood below them in the economic structure. This was matched by the sharpness of the social imagery that, for most of the nineteenth century, separated the 'middle class' - often called the 'middle classes' - from the mass of society.

The middle classes were the leading elements in the towns and cities, occupying positions of power in local councils and running the whole array of church, civic, charitable, and other voluntary associations (Morris 1990). Around the formal associations were formed extensive networks of interlocking committee memberships, informal meetings, and leisure-time activities. Their economic circumstances and life style sustained a cultural and political individualism that precluded any collective action except that in pursuit of sectional professional privileges and business interests. Individuals were to help themselves, and not expect the state to act on their behalf, though this self-help often involved the mobilisation of the informal social networks that abounded within the middle classes. In small towns across the country, this pattern persisted into the 1950s (Stacey 1960).

Internal differences within the intermediate classes did, however, sharpen towards the end of the nineteenth century. The numbers of doctors, lawyers, and teachers grew only very slowly in the years leading up to the First World War, as did the numbers of small property-owners. However, there was rapid and substantial growth in the number of managerial and clerical workers: the number of clerks was almost three times as high in 1911 as it was in 1880. Clerks accounted for 0.8 per cent of all male employees in 1851, but by 1911 they accounted for 5.7 per cent. By 1951, they formed 10.5 per cent.

While these clerks did not have the long-term service contract and the autonomy enjoyed by the old middle class, they did enjoy conditions of employment that were far superior to those of subordinate manual workers. They were generally recruited through the interpersonal connections that they or their families could mobilise, and their promotion depended on their maintaining good personal relations and loyalty to their employers (Lockwood 1958). These employment conditions secured them clear advantages in income and status over manual workers. Like the similarly expanding numbers of shop assistants, commercial travellers, and technicians, however, they remained at some distance from the established middle classes and tended to be defined as 'lower middle class' (Crossick 1977).

The continued growth in the numbers of clerical and managerial workers in the twentieth century further increased the diversity of the 'middle classes'. A growth in the scale of management in large business enterprises produced a growing 'service class' of managers who became increasingly sharply separated from the equally large number of clerical workers whose working conditions were being transformed. Instead of being dependent on the mobilisation of personal contacts, clerks were more dependent on the educational qualifications that they could secure. The expansion of management, however, was also associated with a mechanisation of office work that progressively devalued the skills that these qualifications signified. Office work became more routine and, in many respects, more 'manual' than 'mental' in character (Braverman 1974). Promotion opportunities for clerks declined, while the work itself became more feminised (Holcombe 1973; Anderson 1976). By 1951, over a half of all clerks were female. The women who entered clerical work were often the wives or daughters of manual workers, further reducing the significance of the boundary between clerical work and manual work. Clerks were no longer so unambiguously 'middle class' as they had been in the nineteenth century, while the middle classes had become more exclusively professional and managerial in character.

The transformation of the middle classes, therefore, began much earlier than the changes experienced by subordinate manual workers. Nevertheless, the middle of the twentieth century does mark an important point of transition for the intermediate classes also. Economic competition has meant that small businesses have come under greater pressure from the large enterprises (in which the vast array of managers are employed), and the position of the petty bourgeoisie has consequently been much weakened. Small business owners have also had high levels of self-recruitment, and correspondingly much lower rates of mobility in and out of other intermediate class situations (Scase and Goffee 1982). The growth of large-scale enterprise has also transformed the position of many managerial and professional workers, forcing many of them to convert from self-employed to employed positions (Johnson 1972). Those in service locations have become more diverse. Particularly significant has been the growth of the public sector and of the 'semi-professionals' (teachers, social workers, nurses, etc.) employed within it. The vast growth of financial services in the 1980s has led to an expansion in the numbers of non-traditional service workers in insurance, pensions, and estate agency. At all levels, then, the intermediate classes have been tied to bureaucratised conditions of employment.

The class cohesion of the middle classes was always rather low: they formed a clearly-bounded social class in relation to the working class, but they were internally divided by the relative difficulty of moving from propertied to employed and self-employed positions. Savage et al. (1992) have argued that this separation has become more marked in the second half of the twentieth century, and that the managerial, professional, and propertied sections of the intermediate classes are now sharply distinct from each other. The intermediate layers of the class structure have, like its subordinate levels, become more fragmented. This fragmentation at the intermediate levels has been along the lines of organisational assets, educational assets, and property assets.

As a result, traditional middle class individualism and civic power have weakened (Stacey et al. 1975). Those in the vast managerial bureaucracies - and especially those in the public sector bureaucracies - are more likely to consider collective action in pursuit of their interests. In place of the individualism of the old middle classes there has emerged the collectivism of the new middle classes.

The term 'middle class', however, may be less useful as a designation for those in these intermediate employment and property relations. The term 'middle class' as a self-description, as I have already shown, is as likely to be used by manual workers as it is by non-manual workers. Large bureaucracies, with their extensive professional and managerial hierarchies based on entry by educational qualification, technical competence, measured performance and competitive promotion, form the bases of intermediate class situations. This encourages their occupants towards the social imagery of an ostensibly open and 'classless' society. Intermediate class situations still exist, and they still enjoy life chances that divide them from subordinate manual and routine non-manual workers, but their self-perception - even when it employs the phrase 'middle class' - is not one that emphasises either their traditional status or a solidaristic class identity.

The advantaged classes are those who are involved in property and employment relations that give them ownership and/or control over large amounts of capital. Whether in the form of land, buildings, machinery, or financial securities, property is the key to their power. In some cases this involves the direct personal ownership of substantial property. This might be as the exclusive or part-ownership of a large integrated block of capital (such as a landed estate or a large business enterprise) or it might be the holding of diversified portfolio investments in a large number of different types of capital. In other cases, the class situation might involve an employment relation that defines a position of control over these kinds of assets. The chief executives of large enterprises, their part-time directors, and the investment managers of large financial enterprises are all able to use their control over capital to secure, directly or indirectly, significant advantages for themselves.

For much of the nineteenth century, there was a sharp separation between land and manufacturing as sources of class advantage. Landowners -closely linked to the financial sector - saw themselves and were seen by others as forming an 'upper class', rooted in the aristocracy and, more narrowly, in the peerage. Landed wealth was the source of the biggest fortunes (Rubinstein 1981), and it was the land-owning upper class that dominated the political machinery of the state (Guttsman 1963). Manufacturing capitalists, often provincial in their origins, became an increasingly significant category of wealth, but they were barely distinguished by their contemporaries from the vast mass of the middle classes. The social imagery of upper class, middle class, and working class made manufacturing capitalists virtually invisible as a class. Towards the end of the nineteenth century, land and industry came closer together. Peers and other landowners, who were experiencing a fall in the returns on their estates, became more willing to invest in industrial enterprises and to sit on the boards of the large business enterprises that were formed from the 1880s in banking, insurance, and railways, as well as in manufacturing. A 'lord on the board' was an important marker of stability and probity in a more complex business environment. Through these connections, and through mobility and intermarriage, manufacturing capitalists came grudgingly to be accepted as part of the upper class (Scott 1991; Thompson 1963; Stone and Stone 1984).

The advantaged class at the turn of the twentieth century numbered just under 40,000 people, approximately 0.1 per cent of the population. Its position in society was confirmed more by its status than by any widespread recognition of its economic resources. The popularity of the Royal Family was central to this, as the monarchs were, as major landowners, core members of the class. Victoria, Edward, and the two Georges were well-integrated into the advantaged class, and through the ceremony and glamour that was attached to royalty they ensured that the social class as a whole received a high level of public approval and at least overt deference. The cement that held together the increasingly diverse upper class as a distinct status group was the round of 'Society' activities: the dances, dinners, parties, and salons held by the aristocracy and other prominent families in their London and country homes (McKibbin 1998; Weiner 1981; Rubinstein 1993).

As it became less exclusively aristocratic, Society became more ostentatious and more focused around glamour and frippery. The relaxation of traditional standards helped to weld the diverse elements of the upper class together, but in the long-run Society was incapable of preventing the disintegration of the class in the face of this diversity. The institutions of Society became more open to sheer wealth and less able to impose a single style of life on all its members. This is clear in the informal social networks that were rooted in kinship and friendship and sustained through attendance at public schools and at prestigious Oxford and Cambridge colleges. These networks were important bonds of solidarity for the male members of this upper class, who could rely on the mobilisation of the social capital of their 'old boy networks'. As the century progressed, these networks were becoming weaker. They could no longer guarantee success for land-owning families, though they were still highly successful in the social placement of the sons of the aristocracy, and they could not meet the needs of those who were closer to the margins of the class. advantaged class situations were becoming ever more business-based in character, and less exclusively landed. In parliament and government, in the public schools, in the old universities, and in the top levels of the military and the civil service, the upper class was beginning to lose its distinctiveness and its cohesion.

By the early 1950s, 'Society' had virtually ceased to operate in its traditional form, its demise having been hastened by war-time conditions. The advantaged classes could no longer really be defined as an 'upper class'. They formed, instead, a diversity of overlapping cliques and sets, an 'upper circle' rather than an 'upper class'. These cliques were very diverse. Some followed traditional country pursuits, some were involved in the media and entertainment, some were members of a 'jet set', and others simply kept to themselves and made their money.

This decay in the traditional status of the upper class - the death of the upper class - was the counterpart of the growing economic complexity of property and its management. The middle decades of the twentieth century saw a transition from the direct personal ownership of land and business that had prevailed in the nineteenth century to more indirect and impersonal forms of corporate ownership. The number of large personal shareholdings in major companies declined and there was a corresponding growth in the proportion of shares that were owned by financial enterprises such as insurance companies, banks, and pension funds (Scott 1997). In this situation, it was the directors and top executives who controlled the capital embodied in these enterprises.

These directors and executives were recruited disproportionately from those who were, individually, property owners, although new channels of mobility to advantaged class situations had been opened-up. Wealthy families had diversified their assets by taking stakes in a large number of enterprises and by entrusting the management of these assets to the growing financial enterprises. At the same time, they were able to ensure their continuing control over the assets that generated their wealth by taking top board and management positions in them. Whether as rentiers, who depended passively on the management of their capital by others, or as active finance capitalists and finance executives, the propertied were able to secure a continuation of their advantaged life chances. Indeed, the demise of Society made their privileges far less visible and, therefore, less open to attack, than in the past (Scott 1997: Chapter 8).

This trend has been described as a 'managerial reorganisation' of the capitalist class (Mills 1956). Business enterprises have become ever more closely tied together through the interweaving shareholdings and interlocking directorships of the big financial investors. The directors and executives of these enterprises are overwhelmingly drawn from a propertied background, though with a growing influx from management and the professions and whose children are, thereby, able to become the propertied inheritors of the next generation. The capitalist class of today combines personal wealth with top-level participation in corporate management.

I have tried to show that a Weberian approach to social stratification - and especially his distinction between class and status - allows us to understand both the contemporary contours of class division and the long-term processes that have brought them about. There has been a move away from a stable tripartite system of social classes (upper, middle, and working) to a more fragmented class structure in which the advantaged, intermediate, and subordinate levels are no longer so directly defined in status terms. It was the close association of class and status from the 1880s to the 1950s, underpinned by the economic conditions of the period, that allowed the formation of a stable system of social classes. In these circumstances, the members of the various classes readily accepted this in their own social imagery, class consciousness, and political action.

Economic change and the demographic changes that have gone along with it have destroyed this structure. These changes have long roots, but they have been especially marked since the 1950s. Class divisions still exist, but they are no longer so easily mapped into a tripartite imagery of social classes. There is still a sharp gradient of life chances, as shown in the evidence on health, mortality, housing, and education that I have presented in this Chapter, but it is no longer so obvious exactly where the real social boundaries are to be drawn on this gradient. Boundaries are more fuzzy, less distinct. Indeed, it is not clear whether it is more useful to focus on the broadly-defined advantaged, intermediate, and subordinate classes or on their internal divisions along the lines of their property and employment relations.

In this sense, class as a social division is a complex one that does not divide society neatly or conspicuously into two, three, or even the half a dozen categories of the Registrar-General's scheme. It is also true that, certainly compared to the past, many more individuals can move between classes during their lifetimes (Payne 1992). Other social divisions, not least gender and ethnicity, have become increasingly important to our sense of identity, as will be seen in Chapters 3 and 4. However 'fuzziness' of boundaries, or complexity in the structure of sub-divisions, do not necessarily invalidate the idea of social divisions.

The new complexity of multiple class situations should not disguise this fact. It would be wrong to mistake contemporary society's increased individual capacities to manipulate the cultural trappings of identity as meaning that class actually no longer matters. No amount of personal choice to 'mix and match' consumption behaviour or symbolic life goals will remove the underlying constraints of class situations. The fact that we encounter so much political rhetoric about 'classlessness', and the importance of individual choice or the latest fashion in 'achievement', does not mean that class has disappeared. Indeed, we should be asking in whose interests is it to propagate such myths?

Everybody does not have to believe in the existence of class, nor constantly think of themselves in terms of class identity, for class to be a social division. The system of class situations is not dependent on people's awareness of, or faith in, it. Social divisions do not always lead systematically to a sense of distinctive group identities, or to action with others on the basis of such shared personal social identities.

Despite lacking in visibility, and however imperfectly measured in the existing social classifications, class divisions remain central to social life in contemporary Britain. Of course, we recognise that in contemporary society, people are less likely spontaneously to describe their own experiences in the language of class. They search for more direct and specific determinants of their life chances to put alongside their recognition of class, and they recognise the independent part played by age, gender, and ethnicity. We do not, then, live in a 'classless' society, though we do live in a society whose members no longer spontaneously and unambiguously use the language of class as the obvious, taken-for-granted way of describing social inequalities. Class is not dead, but perhaps the monolithic social imagery of class has, indeed, had its day.

The most comprehensive and up-to-date overview of data on class differences is Ivan Reid's Class in Britain, London: Polity Press, 1998. A summary of the debates over poverty and inequality, which brings together historical and contemporary studies, can be found in Scott, J. Poverty and Wealth, Harlow: Longman. Geoff Payne's Mobility and Change in Modern Society, London: Macmillan, 1987 is an excellent summary of the debates and trends in social mobility, while Class and Cultures: England 1918-1951 is a social history of earlier class differences that adds some life to the statistical bones. Finally, a very useful, balance summary of the theoretical debates can be found in Crompton, R. Class and Stratification, Cambridge: Polity press 1998 (2^nd. Edn.).