Learner differences in theory and practice†

ABSTRACT This paper explores the complex and shifting dimensions of the social, cultural and bodily differences that impact on learners and their learning. Our theoretical argument proceeds in five stages. First, we build a typology of terms used to classify demographic differences for the purposes of designing, implementing and evaluating the effectiveness of educational institutions and programs: material conditions (social class, locale and family); corporeal attributes (age, race, sex and sexuality, physical and mental abilities); and symbolic representations (language, ethnos, communities of commitment and gendre). Second, we address the paradigms of civic association that modern nation-states have used to negotiate these differences: exclusion, assimilation and an aspirational regime that we call ‘civic pluralism’. Third, we explore complications that render the demographic categorizations problematic. Fourth, we propose an alternative and supplementary frame for social and learner differences based on ‘lifeworld differences’. Finally, we explore the ways in which civic pluralism might be translated into educational practice. We interleave these theoretical explorations with an analysis developed for an evaluation of an inclusive education program in Roma communities in Northern Greece. The Roma serve as a case study of the complex ways in which categories of difference play out in social and educational reality.


5.
Resembling what is customary among or characteristic of gipsies; often applied to open-air meals or pic-nics. Oxford English Dictionary, Ed.1, Vol. IV, 1901. When first encountering a class of students, a teacher may note what each student 'is'-'Gypsy' for instance. Every such naming, every such categorization, is inevitably burdened with a weight of complex historical meaning. Every such naming (or refusal to name) brings with it pedagogical implications.
The Oxford Dictionary's definition of 'Gypsy' serves a symptomatic starting point for our analysis of learner differences. Words and their meanings not only reflect social frames of reference. They also have social effects. The alternative framing of learner differences that we develop through this paper represents a waypoint in an interdisciplinary endeavor that we have been undertaking for some decades now, traversing the domains of history, linguistics, philosophy and education (Kalantzis, 2000;, 2012b. This paper is an attempt to revisit and expand upon these ideas. As well analysis, we also want to outline an agenda for social and educational action we call 'civic pluralism'. We use the 'Gypsy' or 'Roma' example because that was our focus as evaluators in the project, 'Education of Roma Children in the Regions of Central Macedonia, West Macedonia and East Macedonia and Thrace'. The evaluation itself is a separate piece (Kalantzis, Cope, & Arvanitis, 2015), although we will discuss it briefly at the end of this paper. As one of the most marginalized cultural and linguistic groups in Europe, there strong parallels as well as striking contrasts with earlier research we have undertaken with Indigenous and immigrant communities in Australia (Cope, 1998;Cope et al., 1994;Kalantzis & Cope, 1999).

In theory: categorical differences
Naming is an act of theory. Names classify, categorize, delineate. In the case of the OED definition of Gypsy, notions of 'race', 'language', 'life', 'custom', and 'habit' are put to theoretical work. These are the kinds of metacategories that generate the specific 'is' of every person, every learner. The metacategories spawn empirical categories, specific classifications such as 'Roma' person, or 'Romany-speaking', or 'Roma culture'. As we will argue in theory and exemplify in the case of the Roma, these metacategories do a lot of work, and because they do so much work, they beg special scrutiny.
Following are some common classifiers in a working theory of human differences that we have used to map the range of possible key factors that can be used to define identity for the self, or have identity ascribed by others: (1) material conditions (social class, locale and family); (2) corporeal attributes (age, race, sex and sexuality, and physical and mental abilities); and (3) symbolic differences (language, ethnos, communities of commitment and gendre).
We will explore the meanings of these metacategories, refining them for the purposes of conceptual clarity (Kalantzis & Cope, 2012b, Chapter 5). However, as we will go on to argue, each one of these categories is fraught with ambiguities and difficulties in general social application, as well as the specifics of educational practice. They evoke gross demographics that stares social actors in the face with a certain kind of obviousness, only to become not-so-obvious on closer examination. But first, we want to revise the naming and refine the definitions that attach to these classifiers: 1. Material differences 1a. Social class -'they make a living by basket-making, horse-dealing and fortunetelling, etc.?' (OED) 'Class' describes the segmentation of society as a consequence of systematic differential relationships of groups to material resources and social power. Slave-owner and slave, serf and feudal lord, worker and capitalistthese are some of the descriptors that been used to capture class differences in various historical epochs (Hindness & Hirst, 1975). Social class is an economic measure of access and control to material resources in unequally ordered social structures. It is a relationship of power and control over others in hierarchical societiesthe slave master owns the slave; the capitalist commands the workers during their working hours. Marx's classical analysis of class relations sets out to explain the processes by means of which workers extend the wealth of the capitalist (Marx & Engels, 1848/1973. Weber analyzes the dynamics of class as social status (1922/1968). In contemporary times, material inequality as a consequence of class differences has been exacerbated (Piketty, 2014), although the gradations and permutations of class have become more complicated (Wright, 1985). Relative success and failure in school correlates with social class (Bernstein, 1971;Bowles & Gintis, 1976).
1b. Locale -'their nomadic life and habits' (OED) Opportunities for education, employment and access to social resources vary by location: different neighborhoods within a city; urban versus suburban; urban versus rural or remote; different regions; developed versus developing countries; and the degree and forms of connectedness between the local and the global. To significant degree, geolocation may correlate with class (Harvey, 1996), though increasingly, class differences appear in sharply juxtaposed spaces (Cope & Kalantzis, 2015b).
1c. Family Different relationships of domesticity and cohabitation create different conditions of socialization. Families are economic units to the extent that they involved the unpaid labor of caregiving, which may or may not be adequate to the needs of working or dependent members. They are also critical sites of socialization, that may or may not align with the culture of schooling (Bourdieu, 1973). In both these respects, the form and practice of family is a critical factor in for educationboth informal learning, and articulation with formal institutions of learning.
2. Corporeal differences 2a. Age From infant and child development, to adulthood and aging, age is a determinant of bodily and mental capacities, as well as relevant and appropriate modes of learning. Developmental psychology tracks cognitive and emotional phases of learning and development in childhood (Piaget, 1923(Piaget, /2002Vygotsky, 1934Vygotsky, /1986, and adult education speaks to the specifics of adult learning (Knowles, Holton, & Swanson, 1973. Today, the uniform and universal 'normal-ness' ascribed to different life phases is becoming less clear. 2b. Race -'they have a dark tawney skin and black hair' (OED) 'Race' refers to the phenotypical differences between one human population and another: skin color, facial features, hair color and texture, height and physique. Such differences, however, have little biological significanceacross a range of biological variables there is greater inherited variation within populations (such as blood types and range of heights and body shapes) than the average variation between populations. Nor does cultural or linguistic variation correlate with peoples typically classified by 'race' (Cavalli-Sforza, 2000). And, despite provocative claims to the contrary (Eysenck, 1971;Herrnstein & Murray, 1995), the consensus among learning scientists today is that differences and intelligence or ability to learn are not heritable in patterns that correspond with phenotypical difference (Fraser, 1995). However illusory race may be as a biological phenomenon or a factor affecting capacities to learn, it remains a critical category in the social imaginary. It underpins 'racism' as an enduring cultural-linguistic reality where spurious biological claims are used to rationalize inequality (Miles, 1989;Roberts, 2011).
2c. Sex and sexuality Sex is the biologically inherited difference between females and males. Sexuality consists of biologically grounded proclivities to form bonds of sexual intimacy. Once only recognized to be an attraction between females and males, today a wider spectrum of sexualities is recognizedincluding among others intersex, homosexuality, bisexuality, transsexuality or asexuality (Herdt, 1994;Mayo 2013).
2d. Physical and mental abilities Bodily form and cognitive capability are spread across a spectrum. These may be framed as 'disabilities', which special schools and other social institutions may be created to address. The range of disabilities may affect hearing, sight, speech, mobility, child/adult development, learning, psycho-emotional challenges, and chronic illness. Disabilities, however, can equally be conceived as institutional problems, or the failure of social and physical contexts to make adequate accommodations to the range of physical and mental abilities (Bowe, 1978).
3. Symbolic differences 3a. Language -'their language (called Romany) is a greatly corrupted dialect of Hindi' (OED) Languages consist of oral and written words and texts. Perhaps there are seven thousand languages in the world today -'perhaps' because the boundaries of languages are blurred by the complexities of shared ancestral languages, dialect divergence, and registers which are mutually incomprehensible (Crystal, 1997;Kalantzis & Cope, 2012a, Chapter 1). In education, learners are differentially impacted, depending on whether their first language is the language of instruction, whether it is a written language or a language of intellectual prestige, and whether the dialects or registers of their everyday usage align with 'academic literacies' (Kalantzis & Cope, 2012a, Chapter 12).
3b. Ethnos -'resembling what is customary among or characteristic of gipsies; often applied to open-air meals or pic-nics … [and] A cunning rogue' (OED) Ethnos refers to the cultural identity of a 'people'. In the modern world, attempts are to create 'nations' that align with geographically defined nation-states (Anderson, 1991;Gellner, 1983). Notwithstanding these geopolitical efforts and the ideology of nationalism used as their justification (in which schools play a key socializing role), ethnic diversity persists in many forms within nations and spanning international borders, including ongoing counter-claims to sovereignty on the part of colonized indigenous communities (Nakata, 2001), and the phenomenon of diasporic community as a consequence of migration (Glazer & Moynihan, 1975).
3c. Communities of commitment Religion and political orientation are communities of commitment that cross-cut language and ethnos (Bourdieu, 1993). Islam, Christianly or atheism; conservatism, liberalism or radicalismthese represent specific commitments to meanings-in-the-world accompanied at times by more or less formal processes of affiliation and institutional structures. Equally, to be areligious or apolitical involves a certain kind of commitment, defined by rejection of others' arguably spurious commitments.
3d. Gendre -'a contemptuous term for a woman, as being cunning, deceitful, fickle, or the like' (OED) We have proposed the word 'gendre' to describe symbolic or cultural attributes ascribed to and associated with sex and sexuality (Kalantzis & Cope, 2012b, Chapter 5).
In an earlier modernity, gender was regarded to be the primary cultural overlay of sex and sexuality. The male/female was the fundamental dichotomy of sex and heterosexual affinity was positioned as 'normal'. Feminist theorists began a critique of these gender relations to the extent that they institutionalized relationships of inequality (de Beauvoir, 1952(de Beauvoir, /1993Mitchell, 1971). Later, theorists of sexuality and social movements advocating for LGBT rights problematized the classical male/female symbolic typology (Mayo, 2013). Our intention in using the word 'gendre' (a word from Middle English meaning 'kind' or 'type'), is to extend the notion of gender, in order to capture the complex range of differences that now manifest themselves in the close interplay of sex, sexuality and their symbolic manifestation as identities.
The metacategories are important because they help to capture aspects of history and experience. The categories that they generate shape the ways we act and interact. They turn our pasts into our futures. For these reasons, they need to be thought through carefully and analytically. However, when speaking of social groups and educational practices, the metacategories often appear as an ad hoc list of differences-not-to-forget (ethnicity and gender and disability and … ). We list them in policy documents in ways that at times feel like an imposed list of caveats, as if to apologize for the otherwise neat heuristics of our social designs. They appearoften jarringlyas overly simple check-in lists for admission to hospitals and schools. The categories are part of our (begrudging it seems?) need to 'manage diversity'.
The metacategories, and the specific classifications of empirical people and practices that they evoke (for instance, 'Roma person', 'Romany language') are necessary in a wide variety contexts. They help us know the people with whom we are interacting, and interact with them effectivelyas clients, or employees, or patients, or students. For the very reason of their necessity, we need to think carefully and analytically. We need to define the terms carefully. We need a comprehensive schema and rigorous social theoryhence our classification of material, corporeal and symbolic differences. Material differences can be addressed by redistribution of resources. Corporeal differences require institutional accommodations. Symbolic differences can be addressed through processes of recognition in sites of civil society such as schools. These are qualitatively different kinds of social process, demanding different kinds of social action.
However, as important as it is to theorize these metacategories, as this paper proceeds we will complicate the categorization, to the point where the categories fall awayalmost. But then, of practical and historical necessity, we bring them back, albeit in qualified form.
In practice: the Roma Now we'll explore the practicalities of classification of social and learner differences by analyzing the application of the metacategories in the case of the category 'Roma'.
The word 'Roma' (or Gypsy, or Traveller) is a term associated with a distinct group of peoplethat part would seem clear enough. But it's not so clear, starting with the numbering of people who may fall under this classification. By geography, Roma are a European people and have been since time immemorial. However some European censuses do not count Romaincluding Greece's census. The classification 'Roma' is made in order that Roma are not counted. Naming in this case removes the name from the official record. Whatever the intent and effect, it's still naming.
The result of this naming in order not-to-be-named is that estimates of the number of Roma in Europe vary by as much as 50% (Fraser, 1992, pp. 298-299). Perhaps ten million Roma live in the area encompassed by the Council of Europe, says the Council's Commissioner for Human Rights (Commissioner for Human Rights, 2012, p. 31). In Greece, there are perhaps 180,000 Roma; perhaps 365,000. Or perhaps there are 50,000, perhaps 500,000, say other sources. Many have no official papers or registration of identity. Resistant to classification as minority status, Roma often simply regard themselves as Greek citizens by birth (Mitakidou, Tressou, & Karagianni, 2015, p. 234;Ziomas, Bouzas, & Spyropoulou, 2011, pp. 3-4). The history of classification, of naming, has brought with it such terrible racism, violence and genocide, that Roma have good reason to resist naming. At which point, Roma seem to disappear, for the Greek authorities, and the Roma themselves who want to avoid the kinds of official attention they have received in the past. But even not-to-be-named still requires naming.

Categorized by name
This paper uses the word 'Roma', transliterating Ρομά from Greek. 'Rom' is the word for 'man' or 'husband' in the Romani language. 'Roma' is the pluralmen, husbands. This term is preferred in European Union documentation because it is a term of self-appellation. 'Romani' is an alternative contemporary usage (Matras, 2015), and 'Rroma' an alternative spelling. The other of this classification is 'gadje'non-Rom or stranger.
However, as a namable European people, the oldest word for this 'people' is Γύφτος -Yftes or 'Gypsy' and its correlates. The root is a medieval Greek word, Aegypti, or 'Egyptians', in use from as early as the fourteenth century. By the high Middle Ages, when it was considered a duty to offer charity on penitents and pilgrims, these self-appointed 'Egyptians' would carry documents assuring passage, supported by a narrative (told here by a sixteenth century chronicler), of: how their ancestors in Lesser Egypt had formerly abandoned for some years the Christian religion … and that, after their repentance, a penance had been imposed on them that … some members of their families should wander about the world and expiate the guilt of their sin. (Fraser, 1992, p. 48, 65) Another commonly used word in Greek is Τσιγγάνοι -Tsiganes, a word that may derive from the medieval Greek 'Athingani', the name for a social group of 'untouchables' (Matras, 2015, p. 21).

Categorized by language
'Approximately two thirds of the Rroma speak Rromanes, the mother tongue of the Rroma,' claims a Rroma Foundation document. 'This means that more people in Europe speak Rromanes than, for example, Norwegian, Dutch or Finnish.' The Romani language offers evidence, the document continues, that 'the Roma originated in India … From their language, Rromanes, and its vocabulary one can infer that the Rroma already left India before the 8th century' (Laederich, 2009, p. 8, 6).
However, there is no continuous tradition among Roma to self-attribute origins in India. There is no historical or orally transmitted memory of Indian origins, though this is now the conventional wisdom. Such origins were 'discovered' by European linguists. The first systematic 'proof' of 'Hindustani' origins was offered in 1783 by a young scholar at Göttingen University in Germany, Henrich Moritz Gottlieb Grelleman, finding that the Romani language was strongly related to north-Indian Sanskrit languages (Willems, 1997, pp. 22-83).

Categorized by class
In Greece today, 'Yftes' (not 'Roma' in everyday discourse) are often encountered as purveyors of Chinese plastic chairs or watermelons from the back of trucks, announcing their arrival with loudspeakers strapped to the roof. They are encountered as Roma women with a child, begging on a street corner. They are, it is said, thieves. They are said to be welfaredependent (Karagianni, Mitakidou, & Tressou, 2013, p. 84). When the European Union frames its focus in class terms, it uses the word 'Roma' to express its official concern: Many of the estimated 10-12 million Roma in Europe face prejudice, intolerance, discrimination and social exclusion in their daily lives. They are marginalized and live in very poor socio-economic conditions. The EU's Europe 2020 strategy for a new growth pathsmart, sustainable and inclusive growthleaves no room for the persistent economic and social marginalization of what constitutes Europe's largest minority. (European Commission, 2011, p. 2) Categorized by race 'At the beginning of the 15th century, a small number of darkskinned people, whose culture, language and way of life were different appeared in Western Europe,' says the Rroma Foundation document (Laederich, 2009, p. 4). In the 1930s, Eugène Pittard, physical anthropologist and founder of the Museum of Ethnography at the University of Geneva, wrote of the 'Tsiganes': Their swarthy complexion, jet-black hair, straight well-formed nose, white teeth, dark-brown wide-open eyes, whether lively or languid in expression, the general suppleness of their deportment, and the harmony of their movements, place them high above many European peoples as regards physical beauty. (Quoted in Fraser, 1992, p. 23)

Categorized by ethnos
Here is the essential 'Gipsy', as portrayed by Kenneth Grahame in his 1908 children's book, The Wind in the Willows: [T]here, drawn out of the coach house into the open, they saw a gipsy caravan, shining with newness, painted a canary-yellow picked out with green, and red wheels. 'There you are!' cried the Toad, straddling and expanding himself. 'There's real life for you, embodied in that little cart. The open road, the dusty highway, the heath, the common, the hedgerows, the rolling downs! Camps, villages, towns, cities! Here to-day, up and off to somewhere else tomorrow! Travel, change, interest, excitement! The whole world before you, and a horizon that's always changing! (Grahame 1908(Grahame /2009 (Figure 1) Later in the same century, in his 'Gypsy Woman', Jimmy Hendrix sings of being hypnotized by a woman's 'gypsy eyes'. Searching for a woman's love, he found himself walking along her 'rebel roadside', a roadside that 'rambles for a million miles'. He wonders where she might be now, this elusive gypsy woman, 'still roamin' in the country side'. Hendrix chooses to walk down the same road, searching for this woman, this love, and his own soul too.
And here's another version of 'gypsy', this time in the words of Madalin Voicu, a prominent Romani politician in Romania: Our gypsies are stupid. They could at least be crafty but they aren't. They are just primitives and they manage to irritate the entire society which is already watching them closely [ … ] They run through the country and Europe barefoot, slimy and dirty, wearing clothes which are more likely to disgust you than make you feel sorry for them [ … ] Begging, soliciting and being disorganized will never bring them any advantages. (Quoted in Nicolae & Slavik, 2013) Categorized by locale Conditions of location frame social and educational experience. In our evaluation of the educational inclusion program for Roma in Northern Greece, we find ourselves on the outskirts of Περαία -Peraia, in July 2012. This community collects scrap for recycling, has no running water or sewerage, and uses electricity stolen from a pole nearby. The children do not attend school, though the aspiration for school is expressed in a small structure with a sign above its door, 'ΣΧΟΛΕΙΟ'. In theory: negotiating categorical differences Since the rise of the modern nation-state, social differences have been negotiated in a variety of ways, based on alternative paradigms of civic association. Historically, there have been three paradigmatic ways of negotiating these differences, in societies and in schools: exclusion, assimilation, and processes of inclusion that we callstrategically and aspirationally -'civic pluralism'. In each of these social models, the categories of difference that we have defined thus far, are put to different kinds of use.

Exclusion
Exclusion is one kind of historic reaction to diverse populations in contact. It is a process of negotiating differences by drawing strict boundaries and insisting on conditions of similarity within a social space. Some types of people are excluded by categorical definition. They are prevented from entry or physically removed. Categorization in this case is a necessary precondition of separation. Forms of exclusion range from the violence of war and genocide to a de facto segregation that is practical though barely articulated. Exclusionary social processes may involve practices that push differences over a normative divide of 'deviance'. Whether brutal or subtle, the underlying premise is that socially uniform groups are to be preferred; difference does not work and it cannot be made to work. Not even assimilation of different others is practicable or achievable.
Instances of exclusion cover as broad a range as the metacategories of difference. Here are a few examples using the categories of social and learner differences. Class: 'affordability' excludes poorer people from certain places, from expensive private schools for instance. Locale: the pragmatics of spatial location open or limit opportunities, in available education and other social resources. Family: creates closures as well as openings in life opportunities, for instance in the disjunction or alignment of family conditions with institutions of education or work. Age: determines specific social possibilities, for instance in strict age and peer segregations. Race: modes of exclusion range from the explicit civic racism of apartheid, to the violent racism of genocide, to the subtleties of attitudinal racism and largely unvoiced de-facto exclusions of institutional racism. Sex and sexuality: exclusion by sex-role, or exclusion of sexualities by classification of deviance against the canons of normality. Physical and mental abilities: from the 'eugenics' or breeding out 'dis'abilities, to quasi-incarceration, to institutional separation in 'special' schools. Language: linguistic imperialism (Crystal, 2000;Phillipson, 1992) and official 'national' languages and monolingual school education that bans or ignores other languages. Ethnos: from ethnonationalist wars to exclusionary nationalist narratives in textbooks (Cope, 1987). Communities of commitment: from religious violence to exclusive religious schools. Gendre: from sexism and imposed heteronormality to the subtle glass ceilings of 'merit' and 'cultural fit' (Kalantzis, Cope, & Issaris, 1988).

Assimilation
Assimilation is quite another orientation to human diversity. It is a process of negotiating differences by means of which a dominant group sets conditions of similarity for entry into a social space. Just as concerned as exclusion and separatism to make virtue of homogenous community, its terms of engagement with difference contrast sharply: 'We will accept people who are different, so long as they are willing and able to become like us.' Failing this, the orientation of assimilation might merely be to attempt to pretend that differences do not exist.
Here are some examples of assimilation at work. Class: scholarships to expensive schools for a select few, those who have been 'passed' according to measures of exceptional excellence styled in the image of the accepting institution, followed by an expectation that those who have been accepted rise to the standards and values of the self-styled elite institution. Locale: internal or international migration to places of 'opportunity'. Family: aligning norms and practices with work or school. Age: norming to grade levels. Race: 'integration', ostensible 'color-blindness'. Sex and sexuality: voluntarily conforming to norms. Physical and mental abilities: 'mainstreaming' where the mainstream puts no effort into accommodation. Language: mainstream 'remedial' or second language learning programs that do not recognize or extend home languages. Ethnos: socialization into national culture. Communities of commitment: acquiring the belief system of the social milieu or institution. Gendre: assuming conventional gender identity and heteronormativity.

Civic pluralism
Given that exclusion and assimilation are bothalbeit in their different waysdiscriminatory orientations to human differences, we want to speak now to an inclusive alternative. We also want to highlight some complex and seemingly contradictory moves towards a more humane sociality that at first glance appear to be reversions to practices exclusionary and assimilationist practices. But first to articulate the general shape of a regime of civic pluralism, and the ways in which the demographic metacategories of difference are put to analytical and programmatic work.
Civic pluralism develops institutions and nurtures human dispositions that work to redress the historic inequalities and injustices that accompany differences between persons as a consequence of variable access to material resources, corporeal attributes and the ascription of symbolic meanings to personhood (Kalantzis, 2000).
To take up the demographic categories again: Class: we know from historical experience that pushing material inequalities to extremes is unsustainable and that they lead inexorably to uncivil states (Polanyi, 1944(Polanyi, /1975. While education cannot in itself redress material inequalities, it is almost the only path that offering intergenerational mobility in access to resources for individuals and categories of person. Locale: the forces of globalization, increased geographical mobility and new communications media offer greater opportunities that could at least in part overcome traditional rigidity of locale-based inequalities. Family: so that forms of domesticity might vary without prejudice to social and educational access. Age: traditional gradations of age and age-related learning institutions may be opened out, and peer separations may become less rigid. Sex and sexuality: are recognized to be varied and malleable. Race: may be discredited as a classificatory of biological classification, while continuing to address racism, or the historical and cultural consequences of having judged phenotypical differences to be markers of superiority and inferiority. Physical and mental abilities: accommodations are created to create conditions of ability for everybody. Language: multilingualism is the norm, in which home languages are deepened and leveraged as an extension of communicative capacities into other, strategically useful languages. Ethnos: such that you don't have to act and feel the same as others in order to be their equal. Communities of commitment: so varieties of belief system can co-exist, to the extent that they are without prejudice to each other. Gendre: action to ensure that gender does not portend inequality, and that divergent cultures of sexuality can flourish (Kalantzis & Cope, 2012b, Chapter 5).
Civic pluralism is both a reality, albeit incomplete, and to the extent that it is incomplete, an aspiration. The reality is the seemingly insistent demographic tendencies and pressures of our late modernity. Inexorably, it seems, the differences underpinning civic pluralism are becoming living and normative realities, buttressed by an expanded conception of human rights. However, severe exclusion and pressures to assimilation persist. The trajectory of change is to this extent also contingent, an outcome of ongoing struggle. History has its arc, however frequently this appears more as a normative or aspirational arc than an actual one. There have been also intermediate steps, some of which are less than satisfactory. A tokenistic 'multiculturalism' (Castles, Cope, Kalantzis, & Morrissey, 1992;Glazer, 1997;Taylor, 1994) may celebrate ethnic differences (the cosmopolitan city of ethnic restaurants, or the school projects on 'other cultures'), while not enhancing social access. This is step beyond exclusion nor assimilation, but on the more ambitious measure of civic pluralism, the outcomes disappoint.
At this point we want to put to theoretical work our threefold categorization of social and learner differences into material conditions, corporeal attributes and symbolic representations. Here, we follow Nancy Fraser who, in debate with Axel Honneth (Honneth, 2003), proposes 'a set of analytical distinctionsfor example, cultural injustices versus economic injustices, recognition versus redistribution' (Fraser, 2008, p. 13). In our terminology, the project of recognition addresses symbolic differences, and the project of redistribution addresses material differences. If the predominant politics in the era of two global systems in the twentieth century, capitalism and communism, was oriented to redistribution, in the era of a single, global capitalist system, the focus of claims to social justice have become more oriented toward a politics of recognition.
And after all, recognition seems more readily achievable (for instance gender equity, non-racism, gay rights) than to overturn material inequalities. This is how (and why) tokenistic multiculturalism only addresses symbolic differences. Its politics is limited by design. By recognizing differences, it represents a 'live and let live' approach that doesn't necessarily deal with inequalities accompanying differences in the realm or redistribution. It is not an holistic program. In a similar line of reasoning, Fraser notes limitations in 'second wave feminism. … It is often said that the movement's relative success in transforming culture stands in sharp contrast with its relative failure to transform institutions' (Fraser, 2009, p. 98).
In fact, says Fraser, 'far from comprehending the totality of moral life, recognition for me is one crucial but limited dimension of social justice' (Fraser, 2003, p. 199). Moreover, these two sets of goals, do not always align comfortably, 'recognition claims … calling attention to [and] … affirming the value of specificity … [and thus to] promote group differentiation; … redistribution claims … often call for abolishing economic arrangements that underpin group specificity … [and thus] tend to promote group de-differentiation' (Fraser, 2008, p. 18).
Notwithstanding the potential for cross-purposes, Fraser also argues that it is possible to bring the two agendas together into a consistent program: I assume that [capitalist society] encompasses two analytically distinct orders of subordination: class stratification, rooted primarily in economic system mechanisms, and status hierarchy, based largely in institutionalized patterns of cultural value. … Whereas class stratification corresponds to maldistribution, status hierarchy corresponds to misrecognition. … both orders of subordination violate a single overarching principle of justice, the principle of participatory parity. (Fraser, 2003, p. 218) Civic pluralism, to connect Fraser's case with our terminology, is an holistic agenda that aims to address both recognition (symbolic differences) and redistribution (material differences). Ours is both a categorical holism in which symbolic differences do not make sense except in a context where material differences are taken into account, and a policy holism, in which any agendas of redress inequality must work on both material and symbolic fronts, in the realms of redistribution as well as recognition. We also want to add corporeal differences into the agenda of a civic pluralism as a metacategory in its own right, neither material in the sense of economic or social resources, nor purely symbolic. To the extent that corporeal differences are differences-in-nature and to so often immutable or hard to change, social justice demands institutional accommodations in order to facilitate equitable participation.
Civic pluralism, in our proposal, abounds in complexities, including some apparent strategic and pragmatic returns. Some of its actions might seem like assimilation. Social access often looks like passing over into the world of the powerful, when women seem to act like men, or linguistic minorities speak dominant languages, or members of faith communities become more 'liberal' in their outlook, or when social mobility looks like 'selling out'. Such moves may be consistent with the agenda of civic pluralism, so long as they are without prejudice to differences. Other actions oriented to civic pluralism may seem like exclusion. Take, for instance, schools designed for specific ethnic or minority communities that help learners because they support their languages, cultures and values but that also offer students an affirming environment that helps them succeed in terms of the 'mainstream' (Cope, 1993). Or schools for girls designed to give them a space to succeed without having to compete with boys. These moves may serve the cause of civic pluralism when their effect is to expand the range of viable, comfortable and authentic personae. The measure of civic pluralism is whether the effect is to extend participation, to equalize access to resources, and to support a balanced, two-way flow of symbolic interactions.

In practice: Roma exclusion, assimilation and inclusion
How, then, have the lives of Roma have been touched by these varied approaches to human differences?

Roma exclusion
Exclusion has been the most persistent and pervasive experience of Roma since their arrival in Europe, right through to the present day.
'Athinganoi' were to be found in the Byzantine east of Europe, perhaps as early as 1068 when they are mentioned in a religious text composed in the monastery of Mount Athos (now Greece), and later in a text by the canonist Theordore Balsamon (d.1204). The Athinganoi, Balsamon says, are fortune-tellers, 'false prophets' who 'would tell one person that he was born under an evil star, and the other under a lucky star; and they would also prophesy about forthcoming good and evil fortunes'. A fifteenth century Byzantine canon prescribes five years excommunication for 'those who consult the Egyptian women for fortune-telling, or the those who bring a soothsayer into their homes to practice sorcery upon them, when they are ill or suffer some other cause' (Quoted in Fraser, 1992, pp. 46-47). In late medieval Europe, these 'Egyptians' are reported to have been traveling tinkers, knife-grinders, fortune-tellers, healers, hawkers, minstrelsand also thieves. Their presence stretched right across Western Europe by the fifteenth century. By virtue of their nomadism, they were marginal in a feudal order that was largely sedentary.
The early modern state began to take exception to 'masterless men' and 'vagabonds', albeit with little authority or capacity to solve the 'problem'. Diderot's Encyclopédie (1751-1772) defined Gypsies as 'vagabonds who profess to tell fortunes by examining hands. Their talent is to sing, dance and steal'. An English statute of 1713 prescribed whipping and hard labor for 'all Persons pretending to by Gipsies, or Wandering in the Habit or form of Counterfeit Egyptians, or pretending to have skill in Physiognomy, Palmistry or like Crafty Science, or pretending to tell Fortunes or like Phantastical Imaginations.' (Fraser, 1992, p. 147, 137) Across Europe at various moments in early modern times, the penalty for such infractions was torture and sometimes death (Figures 2 and 3).
With the arrival of the modern nation-state Gypsies' rights to move, to use common lands, and to practice their various trades without regulation were increasingly restricted. The sedentary poor came to be supported by local authorities in 'poor houses', with stricter records kept of who was local and who was a potentially dangerous, vagrant outsider. 'Vagabonds' and wanderers were viewed with heightened suspicion, and as always-potential criminals. With the rise of the bureaucratic state, the controls became stricter. In order to 'combat the Gypsy nuisance', in 1906 Prussia started to issue licenses for those practicing itinerant trades. In Munich, Alfred Dillman produced a 'Gypsy Book' in which he listed the names of Gypsies, their families and origins. By 1926, a Bavarian law made settlement compulsory and authorized sending Gypsies and other 'work shy' people to workhouses for up to two years, in the interest of 'public security' (Fraser, 1992, pp. 252-253).
These more systematic, methods of exclusion laid the ground for the Gypsy Holocaust or 'Porajmos' -'destruction' in Romani. A key actor in this process was German 'race scientist', Dr Robert Ritter. A psychologist and medical doctor by training, in 1936 Ritter was appointed Director Eugenic and Population Biological Research Station of the Reich Health and Sanitation Office of the Nazi government. Not only did he classify Gypsies as non-Ayrian (arbitrarily, because in the conventional schema of racial classification that uses this term, they should have been Ayrian). He also associated their 'race' with a congenital 'criminal biology'. He and his office proceeded to classify and register individuals as wholly Gypsy and fractions of Gypsy blood, in order to determine the level of social danger that they presented. The purpose in the first instance was to prevent miscegenation, under the principles of eugenics or breeding out of racial weaknesses. On the same principles, there were also forced sterilizations (Willems, 1997, pp. 196-274) Taking this social policy towards its own 'final solution', 219,700 Gypsies are documented to have been killed in German concentration camps, though the number could be as high as 500,000 (Taylor, 2014, p. 174) (Figure 4).
In the twenty-first century, ideologies and practices of Roma exclusion persist. As part of a 'tough new approach on immigration and crime' in 2010, President Sarkozy of France began a process of expulsion of undocumented Roma from France. Five hundred camps were targeted and thousands deported based, it was later disclosed, on a police database of 'travelling people (Gypsies) … and travelling criminals from Eastern Europe (Roma … )' (Davies, 2010;Taylor, 2014, p. 231).
In the context of the Greek crisis of the 2010s there has been a documented rise of xenophobic violence (Angouri & Wodak, 2014;Human Rights Watch, 2012). The first half of that decade also saw the rise of the neo-Nazi party Golden Dawn (Ellinas, 2013). Golden Dawn won 6.3% of the vote in the 2015 elections, and 17 of 300 seats in the Greek Parliament. In the European elections of 2014, they won 9.4% of the vote, and 3 of Greece's 21 seats in the European Parliament. There are Golden Dawn mayors in 14 municipalities. A band called 'Pogrom' sings at Golden Dawn rallies. Its hit song 'Rock for the Fatherland' includes a line, 'We do not want any foreigners and parasites in our land' (Bistis, 2013, p. 47).
69.7% of Roma in Greece have never gone to school; 14.9% have only attended the first four grades of primary school; 10% have graduated from primary school; 0.9% have graduated from high school (Kek, 2009, p. 5). In a case study involving interviews and discussion groups with school personal, community officials and parents in Greece, Triandafyllou and colleagues encountered several main lines of argument for the segregation of Roma and migrant children. One is a 'school quality' case 'that disguises a strong ethnic prejudice' by saying that Roma or migrant children 'are worse than "our" children and immigrant children lower the standards at a school'. Another argument is that 'they create trouble and interrupt the teaching'. Exclusion in these circumstance can take the form of nonattendance because the school is an inhospitable place. Or it may take the form of segregation into separate school annexes allegedly so the school 'can better cater to the needs of the Roma children but with the implicit scope of keeping these children physically away from the local "normal" school'. In the words of a mayor of a neighborhood with a high Roma population, this is also because 'Roma children are not vaccinated and suffer from various skin or other contagious diseases, thus representing a health hazard for other children' (Triandafyllidou, 2011, p. 4, 17;Triandafyllidou & Kouki, 2012).
In 2008, 11 'Greek Nationals of Roma Origin' brought a case to the European Court of Human Rights arguing that the local schools in Aspropyrgos had failed to provide schooling for their children in the 2004-2005 school year. Attempts to enroll them in the regular school had been blockaded by protests staged by non-Roma parents. Special classes were subsequently created for the Roma children in a separate building. However, the court held that 'the conditions of school enrolment for those children and their placement in special preparatory classes resulted in discrimination against them' (European Court of Human Rights -Chamber Judgment, 2008).
In September 2012 in the village of Anthili, near Lamia, a meeting was convened by locals about the Roma camp that the mayor had ordered closed. At the meeting complaints were also voiced about having Roma in the local school because they were 'dirty and sick'. According to a local reporter, 'a dozen local members of Golden Dawn were in attendance and also itching for a fight'. A brawl in fact ensued, and the police had to be called in to break up the meeting. 'The Roma are saying that hundreds of members of Golden Dawn want to attack them and burn them alive. But there are also rumors that the Roma will attack villagers. To sum up, everyone is afraid of everyone' (Parmenopoulos, 2012).

Roma assimilation
The first explicit attempts on the part of the state to assimilate Gypsies were put in place in the eighteenth century. A decree of 1758 by Hapsburg Empress Maria Teresa ordered Gypsies in the Austro-Hungarian Empire to settle, carry out services to landowners and churches, and pay taxes. Wearing of traditional Gypsy dress and speaking the Romani language were banned. In 1782, her son Emperor Joseph II issued an order that Gypsy children must go to school. Hithertofore, they were to be called 'New Citizens', rather than Gypsies (Matras, 2015, pp. 186-187). In Spain, Gypsies were forced into settled communities during the reign of Phillip V, and the practice of traditional trades was prohibited. The word 'Gitano' was banned, and later even the euphemism 'New Castilian' (Fraser, 1992, pp. 164-168).
The intended effect of assimilation policies is that minorities change to become indistinguishable from and thus invisible within the host community. Today, the term 'integration' is often used to the same effect. Schools play a key role in this process. In Greece today, conclude Georgiadis and Zisimos, 'the formal policy from the Greek Ministry of Education' is one of 'assimilating the Roma children in the Greek culture'. As a consequence: there are still no provisions in place for encouraging effectively the promotion of diversity in Roma education: the language, history and culture of various minorities is still not taught in any school; [and] very limited language support is offered to students whose mother tongue is not Greek. (Georgiadis & Zisimos, 2012, pp. 48-49)

Roma inclusion
Inclusion is the normative presupposition that underlies the paradigm of civic pluralism. This reflects both a recognition of difference and a program to address unequal access to material resources. Inclusion underpins any case for a just society.
In 1913, a Georgian Bolshevik, Joseph Stalin, articulated an agenda for national autonomy that would for some decades become a key element in the program of the Communist Party of the Soviet Union.
[A] nation is not a racial or tribal, but an historically constituted community of people … formed … as a result of lengthy and systematic intercourse, as a result of the fact that people live together from generation to generation. The right to self-determination means that a nation can arrange its life according to its own will. It has the right to arrange its own life on the basis of autonomy. … Nations are sovereign and all nations are equal. … [T]he policy of repression … [n]ot infrequently … passes from a 'system' of oppression to a 'system' of inciting nations against each other, to a system of massacres and pogroms. … 'Divide and rule'such is the purpose of the policy of inciting nations against each other. … There can be no possibility of a full development of the Tatar or Jewish worker if he is not allowed to use his native language at meetings and lectures, and if his schools are closed down. (Stalin, 1913(Stalin, /1942 In this context, the state-sanctioned All-Russian-Gypsy Union was founded in 1925. The first Romani language school was opened in Moscow in 1926. A standardized written form of the language was created, based on a North Russian Romani dialect. By the 1930s, hundreds of translations, textbooks, political pamphlets and magazines were published in Romani. Romani industrial collectives were established in chemical and food industries, as well as more than 50 collective farms. A Romen theater was opened in Moscow, which is still open today, as well as a traveling Gypsy theater (Matras, 2015, p. 210;Taylor, 2014, pp. 160-162). In 1981, Yugoslavia granted 'Rom' national status, choosing this word in preference to 'Cigan' when was by then considered pejorative. Romani was taught in a number of primary schools beginning in 1983. TV and radio stations began regular Romani programming. Numerous local Rom social and cultural associations were formed (Fraser, 1992, p. 282).
In the twenty-first century, the EU also speaks to Roma inclusion. The European Commission against Racism and Intolerance was established in 1993 as a monitoring body the Council of Europe including, as per Law 3304/2005, the 'Implementation of the principle of equal treatment regardless of racial or ethnic origin, religious or other beliefs, disability, age or sexual orientation'. Reporting specifically on the educational conditions of the Roma: ECRI urges the Greek authorities to strengthen measures taken to address problems faced by Roma children in education including exclusion, discrimination and under-performance. … However, the absence of disaggregated data on the situation of Roma pupils makes any indepth assessment of their situation and the ability to devise specific programmes targeting this group difficult. (European Commission against Racism and Intolerance, 2009, p. 7, 23, 21) In 2011, the European Union adopted the 'Framework for National Roma Integration' by means of which 'Member States began to coordinate their efforts to close the gap between Roma and non-Roma in access to education, employment, healthcare and housing'. As a part of this agenda, member states were: to ensure, as a minimum, primary school completion, to widen access to quality early childhood education and care, ensure that Roma children are not subject to discrimination or segregation and to reduce the number of early school leavers. … The persistence of segregation of Roma children in special schools or classes remains a key challenge. This was backed by bureaucratic insistence: 'As part of the ex-ante conditionality mechanism for Cohesion Policy 2014-2020, a national Roma inclusion framework has to be in place where funds are programmed for Roma integration' (Directorate-General for Justice, 2014, p. 1, 4, 10). The project which we were evaluating was funded by the EU under these terms, this 'ex-ante conditionality'.
Also in 2011, the Greek Ministry of Labour and Social Security issued a National Strategic Framework for Roma focusing on housing, employment, education, and health. In the words of its opening statement: Dealing with social exclusion and poverty is a fundamental policy priority for all Member States of the European Union. Social integration and equal treatment of groups with distinctive cultural features are particularly important challenges, given that securing harmonious coexistence within the broader social fabric, on fair and democratic terms, and respect for the personality are both essential conditions for the achievement of social cohesion and prosperity. (Hellenic Republic Ministry of Labour and Social Security, 2011) However, apart from several EU-funded projects, very little has been done to address Roma educational exclusion in Greece (Fox & Vidra, 2012). This was the policy context in which our evaluation of the project, 'Education of Roma Children in the Regions of Central Macedonia, West Macedonia and East Macedonia and Thrace', was undertaken.
In theory: it's complicated: factors that confound categorization The metacategories of difference ('ethnicity', 'class', and the like) and the classifiers consequently derived in order to name specific groups ('the Roma', 'the poor'), are used to conceive and enact policies and practices that address social and learner differences. Now we're going to start to complicate these classifications, almost to undo them. We do this because all the categories are too-simple in practice, and at times bring with them problems as serious as those they attempt to resolve. The problems start with the fact that the categories can be put to such dramatically variant kinds of use, to operationalize any of exclusion, assimilation or civic pluralism.
Of course, the demographic categories that describe and classify material, corporeal and symbolic differences capture powerful historical and social realities. A century and more of research tells us that they are rough if not always reliable predictors of social and educational outcomes. However, as soon as we begin to negotiate differences in good faith, we find ourselves confounded by these very categories. They don't tell enough to provide a sufficiently subtle heuristic or guide for our everyday interactions.
The categories also are lined up in lists which, in an era now more sensitive to difference, all-too-often come to sound like a glib litany. What then, do we do to rise above the glibness and the sometimes justified accusations of platitudinous 'political correctness' (Cope & Kalantzis, 1997b)? For history's sake, we need to address the gross demographics captured by such categorization, and today, a lot more, with greater seriousness, and with a keen eye to consequences. Indeed, the more insistent the dynamics of difference become in our schools and in the wider society, and the more critically essential the project of negotiating diversity, then the less satisfactory the demographic metacategories. As soon as the metacategories and categories are applied, human reality resists.
Here are some of the main problems:

Unmanageable lists of differences
Now we have these metacategoriesreferring to sexualities, for instance, or mental and physical abilities, or language. In good faith, we consider these metacategories to be of significance in our social relations and institutions. But the more assiduously we apply ourselves to the task of classification, the more complex the world becomes. Sexuality is not two genders and sexualities but a panoply of forms and shades of sexuality and identity metrosexuals or gay body builders, asexuality or bisexuality, gay marriage or polyamory. Special educators do not just focus on autism or Asperger's but an ever-more finely differentiated socio-behavioral spectrum, to give just one example across the wide range of behavioral, physical, hearing, sight, speech, language and learning abilities. Language is not just an easily namable thing (perhaps there are six thousand in the world today), as mutual intelligibility is confounded varieties of dialect and register. Nor is this a level playing field where all languages are valued equally but a terrain of contextual variation where some languages and modes of literacy have greater local or global efficacy than others. Language learning is not just a matter of first or second language acquisition but variable capacities in any number of languages, dialects and registers ). The metacategories demand empirical substance. In order to name specific groups, this creates a secondary level of classification and subclassification that is endless so long as you keep trying. We face real-world specificities such that, to align people with ostensible categorical norms, is to over-simplify their actually experienced conditions of life. And the smaller differences are significant. Fine distinctions need to be made that are at times just as important or even more important than the gross distinctions. In the Rwandan massacre of 1994, Tutsis could only be definitively distinguished from Hutus when they presented an ID card with an ethnic identification. Freud had a formulation to describe this phenomenon: 'the narcissism of minor differences' (Freud, 1930(Freud, /1975. When one needs to categorize something that is in empirical reality infinitely complexthe forms of the human condition, for instanceit soon becomes evident that the categories must undergo continuous refinement so they progressively become more operationally valid. Indeed, this may be an occupational hazard in any process of categorization and classification, as Bowker and Starr point out in the case of the history of racial categorization in apartheid South Africa (Bowker & Star, 2000), and as Burbules points out in a reading of difference that has a Wittgensteinian sensitivity to contextual blurriness in the of naming things (Burbules, 1997). This is why the project of improving the categories that describe human differences becomes a series of receding horizons. In classrooms, teachers today may come to feel that they need to know an impossible-to-manage encyclopedia of empirically identifiable learner differencesthe shades of disability, or the range of second language learning needs, or the variable effects of different forms of poverty, or the endless complexity of cultural differences.

Internal group variations
The groups named in each of the demographic classifications are massively internally differentiateddifferences among Roma, or women, or Arabic language speakers, for instance. In fact, a rough general theorem would be that the spread of internal differences within any demographically defined group is greater than the average difference between groups. This means that the demographic classifications, while helpful to our understanding of the historical and experiential basis for certain moral agendas and social claims, are oversimplified and often counterproductively so.
Worse, the categories of gross demographics can easily lead to stereotypical generalizationsabout 'Asian learning styles', boys' personae vis-a-vis academic learning, or the educational consequences of socio-economic disadvantage, for instance. Indeed, exclusionary and assimilationist agendasracism, or sexism, or homophobiacan and do use the self-same classifications for their own purposes.
As a consequence, even well intentioned diversity programs based on one or several of these categories can at times prove ill-judged or irrelevant to individual or subgroup circumstances, and even at times counterproductive. They may oversimplify critical successand failure-determining differences within groups and between individuals. For instance, some students in disadvantaged groups do succeed; background is not all-determining. Indeed, in some conjunctions of circumstance, a student's 'disadvantaged' background may become the basis for their particular resilience (McGinty, 1997), their peculiar success. Sometimes also, the demographic categories become invidious labels, implying a deficit on the part of the student, when in fact they may be an opportunity upon which to build constructive learning experiences.

Intergroup relationality
Differences are invariably relational. Group membership cannot be neatly categorized and described as though the boundaries of the group are its beginning and end points. The very act of categorization tends to imply that in-group cohesion, cultural commonalities and personal identifications are more important than intergroup relations and processes of co-construction of difference. Rather, groups are the products relationships in which one group is constituted in relation to another. Groups exist in dynamic, and neverstable, tensionclass to class, gender to gender, ethnos to ethnos, and disability in relation to the affordances of physical and social structures designed for certain kinds of ability. The one group would not even exist, or at least not its current manifestation, but for its counter-position in relation to another group in-its-difference. The social phenomena that the demographic categories purport to describe are in fact almost invariably relational. They are as much a creature of the dynamics of the social whole as they are products of isolatable group commonalities. They are defined in and through social relationshipsof comparative power, privilege and access to resources. Each group is created and evolves over time through a series of historical and ongoing intergroup relationships. Without the one, there would not be the 'other', or at least the other would be something quite different. These relationships (racism, sexism, comparative socio-economic privilege and the like) often play themselves through in schools and classrooms via deeply relational dynamics, in which the forces of mutual intergroup definition are stronger than the purely endogenous attributes of groups and their members.

Intersectionality
The differences intersect. More than the dozen or so key metacategories of difference that we identified earlier in this paper, and more than the endless potentials for empirical subcategorization, for any individual, the chance of any one particular combination (class, gendre, race, body form … ) is so low that, in their mixed-up peculiarity, they can only ever belong to the tiniest of minorities. Indeed, the specification of difference does not have to be too finely grained before every minority is a minority of one. This means that to classify a person into a singular demographic category may do disservice to their more precisely defined needs and interests. No matter how long and finely grained the list of categories to be applied, and no matter how many of them are applied to a person, the categorical list of this plus this plus that is always all-too-neat. In every person, the categories interact to form new and different permutations of experience. The whole of identity is more than the sum of its parts. In combination, the parts transform each other. Every person is always a new formation. This phenomenon of multilayered, multifaceted identity is frequently named 'intersectionality' (Crenshaw, 2015).
Moreover, in a person's life experience, the groups with which they have been affiliated are not separate; they are overlapping, simultaneous, multilayered. The consequent differences are never things-in-themselves. Rather, every aspect of material, corporeal and symbolic difference manifests is deeply overlaid, forming an integrated whole. Every person represents a peculiar conjunction of dimensions of difference, a unique mix of group or community experiences. The constitution of that individual can only be understood through the nuanced intricacies of their narratives of life experience. By the time the layers are combined, what might be predicted by one demographic variable is transformed to mean something else by its interrelation with other variables.

Change
Differences never stay still. They are not states simply to be found, classified and dealt with. One of the dangers of group categorization is to assume stasis, to over-emphasize continuity for the purposes of heuristic clarity. On the contrary, groups are always moving, always fluid, always in a state of dynamic and unstable change. And we social actors do not necessarily remain content to leave differences the way they are. We may want to move them along. This can either be from the perspective of an insidera woman who wants to change the role of women, or an indigenous activist struggling to improve the conditions of life of their people, or an individual striving for social mobility, for instance. Or it can be from an outsider's perspective, for instance, the ways in which educators support learners in their self-transformation or growth, to achieve dreams and aspirations that may have seemed beyond the scope of possibility within the frame of their lifeworlds.
The operational validity of the demographic categories has nowalmostfallen away. But we can't let them go. They have deep historical and pragmatic meanings. They are the tokens and sites of significant struggles for justice, both for redistribution and recognition, to adopt Fraser's terminology once again, as well as the institutional accommodation of corporeal differences. In these senses the categories of are of strategic value; they are a theoretical and social heuristic.
In natural language, the meanings of words are rife with ambiguities and semantic untidiness. Dictionaries capture the historical nuances of natural language in everyday use. The metacategories of difference (class, language, physical and mental abilities, and the like), and also the empirical subcategories (Asperger's, Arabic, asexuality, and the like), constitute what we call a 'strategically un-natural language' (Cope, Kalantzis, & Magee, 2011, Chapter 13). They abstract criterial features of the world, thus simplifying for the sake of observation and argument. They classify the world for the purposes of analysis and action. We can't discard such categorical languages because they are drenched in historical experience as well as today's future-oriented calls to action. They are born in the common ground of shared struggle. We still need to operate in these categorical modesin our classrooms and in the wider world of human differences. However, we also live and work in a world where we know that the differences must be nuanced in ways that are more realistic than these metacategories and their consequent demographic categories allow. We call this more nuanced perspective, after Husserl and his successors, 'lifeworld attributes'. But before reaching this discussion, we return to the Roma.

In practice: Roma experiences
To categorize by name 'Roma'/'Rroma'/'Romani'/'Romany'; 'Gypsies'/'Yftes'; 'Tsiganes'/'Ciganos'these are some contemporary names, and variants of 'Gypsy' and 'Tsigane' are at times regarded to be pejorative. There are other names, for some subgroups and sometimes several or all: 'Travellers', 'Sinti', 'Kalé', 'Manoche', 'Ashkali', 'Gens du Voyage'. 'It is not always clear who should be included and who should be excluded from these categories,' says the Commissioner for Human Rights of the Council of Europe. 'The minorities labeled "Roma", "Gypsies" and "Travellers" in fact comprise a multitude of ethnicities and distinct linguistic communities, heterogeneous groups that are viewed as a unit primarily by outsiders' (Commissioner for Human Rights, 2012, p. 32).
The situation is also complicated by refusals to name, on the part of authorities and persons themselves who are the objects of the naming. The Greek census does not collect data on ethnic affiliation or language, and there is no other systematic form of registration or enumeration of the Roma population. Nor does France, for the ostensible reason that to do so would contradict principles of equal treatment of citizens and their right to privacy. There is also a concern in France that this information might fuel the racism of the National Front, as well as the historical memory Vichy collaboration with Nazis in the Jewish holocaust, and Gypsy internment and deportation to concentration camps in this period.
However, not classifying and not counting also amounts to a certain kind of refusal to address discrimination and inequality. In the words of an Open Society Institute report: there is a fundamental lack of accurate information on Roma in Central, Eastern, and South Eastern Europe. … This failure has far-reaching consequences: the lack of information renders policy and planning ineffective and makes it impossible to monitor changes. Inadequate data also perpetuates vast disparities in access to quality education. Inadequate information will continue to enable governments to evade responsibility for failing to create, fund, and implement effective programs. (Open Society Institute, 2006, p. 2) On the other hand, Roma often have good reason to prefer that they are not named. 'We are Greeks; we are citizens'this is a legitimate counter-claim to a label that describes marginalization, another kind of principled refusal to name. As the Commissioner for Human Rights admits: not surprisingly, many Roma continue to see the authorities as a threat. When required to register or to be fingerprinted they fear the worst. This is all the more understandable when they explain how they see similarities between much of today's anti-Roma rhetoric with the language used in the past in Europe by Nazis and fascists and other extremists. (Commissioner for Human Rights, 2012, p. 8) So, when it comes to naming Roma as a group, it's complicated.

To categorize by language
Romani consists of a number of variants that are often not mutually intelligible. Here is the Ethnologue classification, where the each major variant is sufficiently divergent to be classified as a language in its own right: Balkan Romani (dialects: Arlija, Dzambazi, Tinners Romani), Baltic Romani (dialects: Estonian Romani, Latvian Romani (Lettish Romani), North Russian Romani, Polish Romani, White Russian Romani), Carpathian Romani (dialects: East Slovakian Romani, Moravian Romani, West Slovakian Romani), Kalo Finnish Romani, Sinte Romani (dialects: Abbruzzesi, Serbian Romani, Slovenian-Croatian Romanim German Romani), Vlax Romani (dialects: Churari, Bisa, Ghagar, Grekurja, Kalderash, Lovari, Machvano, North Albanian, Sedentary Bulgaria, Sedentary Romania, Serbo-Bosnian, South Albanian, Ukraine-Moldavia, Zagundzi, Kalderash, Ursari, Churari, Argintari, Lingurari), and Welsh Romani. Any number of classifications and subclassifications are possible; other namings and classifications are strikingly different (Laederich, 2009, p. 9). Apart from several 'isolates' that may be remnant indigenous languages (Basque, Finnish, Hungarian), like other major European language. Romani is a language with origins in the easthence its classification as an Indo-European language. Like other European languages it includes historical accretions, in the case of Romani from Persian, Armenian and Byzantine Greek. In contemporary usage, it contains many loan words from local languages. As a principally oral language, Romani is a particularly dynamic language, and locational divergence continues at a faster pace than written and standardized languages. Some variants such as the Catalonian, Spanish, Portuguese, and Brazilian Caló have become dialects of the host language. In Laederich's estimation, perhaps one third of Roma do not speak Rromanes (her spelling) (Laederich, 2009, p. 8). Roma who speak Romani are also necessarily bi-or multilingual in local as well as home languages. So, when it comes to classify Roma by language, it's complicated.

To categorize by class
For the (perhaps) millennium-long life in Europe, manybut by no means all -Roma have been part of a class outside of the main class system, neither serfs nor lords in European feudalism, neither capitalists nor workers in modernity. Roma are have long been traveling salespersons, scrap collectors, entertainers and beggars. But some have also been in almost as many other classes as there can be, and are today. They were slaves where slave states existed, such as Wallachia and Moldova until slavery was abolished there in 1856 (Taylor, 2014, pp. 126-128). They were rounded up by order of Ferdinand IV of Spain in 1794 and forced to work in the state's naval arsenals (Fraser, 1992, p. 166). They joined the proletariat in factories and collective farms in the Soviet Union (Taylor, 2014, pp. 160-161). They have migrated to the New World, finding themselves in located in conventional class formations as workers or business people. So when it comes to locating Roma by class, it's complicated ( Figure 5).

To categorize by race
Like all European populations, Roma genetic ancestry is mixed. In Fraser's hypothetical calculation over a millennium of European settlement, if only four Roma marriages in every hundred were with non-Roma, then the Roma of today would have 70% non-Roma ancestry. This is why the eugenicists and lineage list-makers of Bavaria then the Third Reich found themselves confounded by the complexities of blood percentages (Fraser, 1992, p. 6). Recent genetic research confirms Roma origins in Northwest India and suggests emigration of a single group about 1500 years ago. However, the researchers also report 'extensive genetic variation' based both on recent and older 'genetic admixture' (Mendizabal et al., 2012). In these respects, the Roma share a generalized pattern of movement and inter-relationship across Europe populations over recent millenniaa migration from the east (beginning several thousand years ago with the first farmers), genetic admixture with indigenous Europeans, and the supplanting indigenous languages an Indo-European language (Cavalli-Sforza & Cavalli-Sforza, 1995;Renfrew, 1998, pp. 184-186;Sykes, 2001, p. 145). As a consequence of this history and variety of experience, some Roma have dark skin; others have white skin. Nor is 'race' any more a fixed categoryas narrow and problematic term as it may begiven the contemporary fashion among young Roma women in Greece to use chemical skin whiteners (Karagianni, 2015). So, when it comes to identifying Roma by race, it's complicated.

To categorize by ethnos
Roma are not recognized as an ethnic minority by the Greek state. Either they register as Greek citizens, or they remain stateless (Triandafyllidou & Kouki, 2012, pp. 24-25). When it comes to identify ethnic continuity of the Roma in Greece, Mitakidou, Tressou, and Karagianni say: it would be inaccurate to consider [Roma] as a homogeneous group in any aspect, when, in fact, they represent a very wide spectrum of in-group, often in-family, variability. In terms of  (Fraser, 1992, p. 225).
their variations, Greek Roma are mainly Christian or Muslim (in Thrace) and they speak Romani (an oral language), Greek, and Turkish (in Thrace), as well as combinations of these three and other languages. They are divided in subgroups, such as Tsigganoi, Fitsiria, Gyftoi, Tourkogyftoi, Katsiveloi, and Athiganoi, according to their geographical descent, dialect, occupation, and settlement. (Mitakidou et al., 2015, pp. 234-235) The variations of Roma identity across Europe, and the Roma New World diaspora, are endless (Matras, 2015, pp. 283-293). 'There is no sense of our having been a single, unified people,' says Roma activist and scholar Ian Hancock (Hancock, 2010, p. 17). Concludes another Roma scholar, Brian Belton: what people see themselves to be is, in the main, not a one-dimensional social or ethnic type, but what I have come to than of as 'resultant beings': they point out a range of influences and lineages that are expanded over time and across circumstances. (Belton, 2010, p. 39) There are, however, cultural motifs that we have come to see as distinctively 'gypsy': a style of clothing; motifs in the music of Brahms, Liszt and Bartók; flamenco as a dance form. But these are have also become other peoples' ethnos. How also 'Spanish' is flamenco, and how also 'Hungarian' the musical motifs? And when a non-Roma wear gypsy clothes, or hum a riff from what has become the canon of classical music, or dance flamenco, these have become a deeply corporeal parts of their identities, too. We can all be gypsies now. Such elements of identity defy attempts neatly to align of persons with groups defined by ethnos.
Moreover, there may also be some aspects of 'being Roma' that some Roma may wish to abandon, or that should be abandoned, or even must be forcibly abolisheda 'cleanliness' regime that restricts the lives women, and arranged child marriage, for instance (Commissioner for Human Rights, 2012, p. 19, 104-105).
Nevertheless, the construction of ethnos becomes a strategic rallying point. The first World Romany Congress was held in London in 1971. Delegates from 14 countries agreed to adopt the term 'Rom', and took on those accouterments of nationa flag, a 'national anthem' and a national day (Council of Europe, 2011;Fraser, 1992, p. 316;Matras, 2015, p. 32). The Rroma foundation was established in Switzerland in 1993, with financial support from the Soros Foundation (Laederich, 2009, p. 19). So, when it comes to Roma ethnicity, it's complicated.

To categorize by locale
Earlier in this paper, we saw the Roma camp at Peraia in Northern Greece. Now we'll visit two more, strikingly different Roma locales. We do this to show the sheer range of geospatial experience, even within the same region.
In Ottoman times the residents of the village of Φλάμπουρο -Flamapouro most of whom are Roma origin, worked as share-croppers on the lands two wealthy Turkish landowners, or as fisherpeople on the local lake. After the departure of the Turks, many gained ownership of small pieces of farming land, supplementing their livelihoods with day labor. During the Second World War and the subsequent civil war, many joined the resistance led by the Communist Party of Greece. In the post war period became trade unionists and active members of the Pan-Hellenic Socialist Movement. Although educational participation by the Roma of Flamapouro has been below the norm in Greece, it has been much higher than in other Roma locales (Zachos, 2010(Zachos, , 2011. Nor is this a unique phenomenon in Northern Greece (Themelis, 2013).
҅ Aγία Σοφία -Agia Sophia, our third locale, is different again, both from Flamapouro and Peraia. Unlike Peraia, it is a planned settlement of prefab homes, laid out in a rectangular road pattern, and with electricity (which the community complains is expensive). The children are enrolled in local schools, but attendance is poor. In evidence are the traditional, highly visible hawking trades, such as selling watermelons. The 50€ notes strewn across the dashboard of a pickup truck also point to another market that is considerably more lucrative than watermelons. So do the large new houses under construction on the sites of the former prefabs. So, when it comes to examining Roma locales, it's complicated.
The categories, in short, fail us. We do need some of their aggregations for strategic purposes. However, the classical categorizations of modernity no longer work for us as easily as once it seemed they might. So, the strategic use of these categories in policy and practice must necessarily be balanced with a disaggregation of 'lifeworld differences'. Such is the extra categorical work we must do if we are to identify and act upon social and learner differences.

In theory: lifeworld differences
Because 'it's complicated', we propose an alternative and supplementary framing of social and learner differences focused on the notion of 'lifeworld'.
The 'lifeworld' consists of the things we end up knowing without having to think how we came to know them (Bourdieu, 1993;Cope & Kalantzis, 2000;Habermas, 1981Habermas, /1987Husserl, 1954Husserl, /1970Kraus, 2014). It is the way we end up being without ever having consciously decided to be that way. The lifeworld is not particularly explicit. It is made up of things that seem so obvious to insiders that they don't need saying. It is a set of habits, behaviors, values and interests that go without saying in a particular context. These are things that go without saying because they have come without saying. Knowledge of the lifeworld does not have to be taught in a formal way. We learn how to be in the lifeworld just by living in it, and this learning is mostly so unconscious that it is rarely consciously experienced as learning. The lifeworld is the ground of our existence, the already learned and continuously being-learnt experience of everyday life. It is also the locus of our subjectivity and identity, the source of our motivation, the basis of our agency. It is intuitive, instinctive and deeply felt. We are both determined by it in its pervasive 'surroundingness' and we determine it by our choices, our everyday actions and practices. It is dynamic, shifting, and sometimes unpredictable in its pace and scope of change.
In a formal educational context, the lifeworld is the everyday lived experience that learners bring to a place of learning. It is the person they have so far become through the influence of their family, their local community, their friends, their peers and the particular slices of popular or domestic culture with which they identify. The lifeworld is what has shaped them. It is has made them who they are. It is what they like and unreflectively dislike. It is who they are. The underlying attributes of lifeworld experience form the basis of identity and subjectivity.
To take note of lifeworld attributes may give concrete substance to the differences identified in the demographic categories. However, just as often in the lifeworld, one finds oneself in the presence of differences that can only be grasped at a level which defies tidy categorization. Looking through the frame of the lifeworld differences also may also reveal the points at which the demographic categories overgeneralize, or miss the mark when these categories prove to be unhelpful or plain wrong, and even at times create counterproductive stereotypes. Here are some vectors in the constitution of lifeworld attributes:

Affinities
We are who and what we associate ourselves with, and what that association stands for. 'Affinity' captures senses of community, from formal membership connections with groups, to relations in networks, to informal senses of connection in common causes. Affinity may connect a person to sports or teams, through shared enthusiasm and common interests, in workplace 'cultures' or loyalty to products, in the collective sympathies of social or environmental movements, in professions or shared communities of practice.
Personae 'Persona' captures the kind of person we envision ourselves to be, and style and present ourselves as. It is how we see others seeing us. It may be affected or unarticulated. It may be conscious, semi-conscious or unconscious. Persona may be manifest in gesture, demeanor, social intersubjectivity, and modes of 'presentation of self' (Goffman, 1959) including fashion and varieties of person-to-person interaction. Different personae may represent stances, values, worldviews, dispositions and sensibilities. They may also reflect cultivated or un-self-reflective quirks of 'personality'.

Agencies
In the era of a politics of recognition, there is greater scope for self-making, in symbolic areas even if the scope for material self-making remains as restricted as ever. The consequence is a plethora of identitiesaesthetic, epistemological, discursive, interpersonal. The essence of the change is epitomized in transition from mass production of uniform products for massified, putatively homogenous market (Henry Ford: 'Any color you like, as long as it is black'), to customization for a myriad of niche markets, each representing an identity position (now, you are what you drive) (Cope & Kalantzis, 1997a). Inequalities in access to material resources may stubbornly persist and even worsen, but within the realm of symbolic differences, there is greater space for agency.

Narratives
Affinities, personae and agencies come together into narratives of person. These are lifehistorical, ongoing, and aspirational. They are simultaneously retrospective and constructivelife trajectories tracing where a person has been, in relation to where they are now, in relation to where they might imagine their life course to be are heading. Such identity narratives are the givens that are constitutive of who a person is, and how they enact their being. They tell of how the social and historical is instantiated in the personal and contemporary.
Thus lifeworld experience constitutes 'identity', but not just in the sense implied the etymology of this term -'things alike' in a group of people. To a greater degree today than in our modern pasts, identity is constituted by things that are noticeably unlike. In the era of a politics of recognition, and also in the strategic vision of civic pluralism, difference sits deep in our consciousness, our epistemologies, our subjectivities and our means of production of meaning. No longer can we assume there to be a universal personality (normal, or deviant but remediable, or excluded because irredeemable), because the universal today is a humanity of personalities emphatically in the plural (the range of our differences), and also in the multiple (the layered complexity of the differences within usfor every individual the unique intersection of attributes, the nature and sources of which may often be ascribed to groups and socialization).
We live today in an era that affords greater scope for agency in the realms of symbolic and corporeal difference. We are more able now than we were in an earlier modernity to make ourselves more different. And because we can, we do. Take for instance the rainbow of gender identifications and expressions of sexuality in the newly plastic body; or the shades of ethnic identity and the juxtapositions of identity which challenge our inherited conceptions of neighborhood; or the locale that highlights its peculiarities to tourists; or the panoply of identities supported by the new, participatory media; or the bewildering range of products anticipating any number of consumer identities and product reconfigurations by consumers themselves. This becomes the stuff of our personalities in the plural and the multiple. Together, these manifest themselves as the multiplicity of our dispositions, our sensibilities, our identities.
We also live in a world where the poles of comparative access to material resources are pulling further apart. Here too, there is scope both for greater agency, as well as the active suppression of that agency. Material inequalities create fissures and tensions that pervade society, from small acts of anger and resistance, to larger frames of democratic and undemocratic political action. In the case of learners, one line of agency may be to use education as a path to access, through formal educational qualifications that secure intergenerational upward mobility. Another might be to learn to be political, offering vicarious or practical support to reform agendas that aim to reduce or eliminate material inequalities. They dynamics of engagement with material inequality are complex and highly variable from context to context and person to person.
Pursuing these lines of exploration into learner lifeworldsbe that self-understanding on the part of learners or the understanding that teachers have of their learnersbrings us closer to the realities of persons and their learning needs than was ever possible with gross demographic categorization alone. In fact, it both adds substance to the categories and qualifies their overgeneralizing tendencies. It takes us past the unmanageable list of group categories because it focuses on narrative specificities of particular persons and their inter-relations. It leads to the discovery of internal group variations. It finds intergroup relationality because that's what the life narratives speak to. It accounts for the unique intersectionality of every person's group-and history-related position. It takes notice of change, and anticipates further change. Importantly also, it asks these questions of every learner, thus getting away from a tendency to start diversity agendas with a perfunctory list of equity groups.
And why must we go to these lifeworld attributes when we address learner differences? Learning succeeds or fails to the extent that it engages the varied subjectivities of learners. Behind the demographics are real people, who have always-already learned. The range of their learning possibilities are both boundless and circumscribed by what they have learned already and who they have come to be through their learning. Education, then, needs to engage with difference at a far deeper level and in a much more nuanced way that the gross demographic categorization allows. The challenge is, how do we engage all learners in classrooms of deep difference? In other words, how do we do civic pluralism?
In theory: towards the inclusive school The modern schooland the nation-state, and the 'Fordist' factory, and the newspapers and television of the old 'mass media' ofaddresses differences by imposing a epistemological and moral architecture of sameness. Differences inexorably exist, but the institutions of an older modernity attempt to ignore, or elide, or write over these differences. This is how the modern school establishes its program of sameness: the school takes students out of life and puts them into classroom whose discourse is exophoric and singular. Exophoric: school perennially refers to things outside of the classroom. The referring is a purely discursive and abstract act of 'bringing in' potentially anything or everything of the outside worldplanets, or nineteenth century novels, or mathematical theorems. Singular: the voices of the knowing teacher or the synoptic textbook refer unequivocally to facts and norms.
The syllabus lays out the officially sanctioned program of learning. The textbook brings the world into the classroom by summarizing the worlda singularly authoritative synopsis of facts and theories deemed significant. The students are arrayed behind rows of desks, eyes directed to teacher at the front of the classroom. The teacher lectures the class: 'Listen, O my son, to the precepts of thy master, [for it] belongeth to the master to speak and to teach; it becometh the disciple to be silent and to listen' (St Benedict c.530 [1949]). Mostly, the students have to listen because there are many students to one teacher. The teacher may ask a question: teacher initiation ('what?' 'how?'); followed by student response (one student puts up their hand, attempts to give the answer they anticipate the teacher expects, so acting as a proxy for all others on the assumption that they could or should provide the same answer); followed by teacher evaluation ('yes' that's right', or 'no, try again') (Cazden, 2001). The students undertake activities, rehearsing the presented curriculum narrative in notes, written answers, or essays. The cycle of classroom discourse comes to a temporary close with the test that measures the degree to which learners and their teacher have remembered the content prescribed in the syllabus. This is an architecture, discourse, epistemology of sameness. Knowledge is a hierarchically structured system of transmission, syllabus => textbook => teacher delivery => learner test. The curriculum is univocalsingular narratives of mathematics, or the triumphal history of the nation-state, or the rules of official languages. Classes are carefully constituted to align with this regime of sameness. Other languages are forbidden, or just not offered. Classes are finely differentiated by age (grade n) and 'ability' (classes nA, nB … ). 'Disabilities' are removed into 'special schools'. The rich are sorted into exclusive educational enclaves by affordability. Local schools sort by unequal locale. In single-sex schools, students learn what seems appropriate to traditional gender roles.
Everything about this modern school could change, and should change. We name this aspiration, a learning architecture of productive diversity and inclusion. In the era of civic pluralism, such an architecture may become a child of necessity. For educators, it must also be an agenda of invention. This is something we need to work to design, a pedagogical order that we educators need to parent.
Inclusive education is a way of working with learner differences such that differences are without prejudice to social access and symbolic recognition. Inclusion means that you don't have to be the same to be equalto have similar opportunities, not identical opportunities, but the same kinds of opportunities measured in terms of comparable access to material resources in the form of employment, civic participation and senses of belonging to a broader as well as a localized community (Kalantzis & Cope, 2012b, Chapter 5;Slee, 2011). In this context, learning is not a matter of unilinear 'development' in which you leave your old self behind, jettisoning lifeworlds that would in earlier times have been framed by education as less inadequate to the task of modern life. Rather, it is an open-ended process of extending one's cultural repertoire, starting with a recognition of lifeworld experience and using that experience as a basis for extending what one knows and what one can do. An inclusive process of transformation, then, is not a matter of vertical development; rather, it is a process of expanding horizons. These new horizons have a reciprocal impact on the lifeworld: learners engage in and with their lifeworlds in new ways, changing those lifeworld conditions through engagement, but not necessarily in order to leave those lifeworlds behind in a kind of one-way trip.
Following are some items in an agenda for inclusive education: To know learner profiles Don't assume learners are the same, even though in an educational architecture of sameness this usually the institutional and pedagogical default. When beginning to get beyond the architecture of sameness, the demographic classifiers are an important starting point and they do mean something. However they are no more than a starting point and a few simple demographic descriptors are nowhere near enough. Elsewhere, we have worked on policy development in which personal profiles are developed at a level of granularity that is appropriate to a service, with information requirements determined on a need-to-know basis (Department of Immigration and Multicultural Affairs, 1998). If surgery, you need to know quite a lot of specific information (about religion, language spoken, dietary culture), more than if you are a public transport provider, for instance. If you are in education, you need to know a lot about the student starting from the point of enrollmenthome languages spoken? socio-economic conditions? In fact, you probably need to cover nearly every aspect of the demographic metacategories that we outlined earlier in this paper.
However, don't assume that initial or seemingly obvious demographic classifications will work. The only way to find out more is to build curriculum processes and community relations that are open, dialogical and continuous. To frame this in terms of our Learning by Design or Multiliteracies pedagogy, essential pedagogical moves include 'experiencing the known', where students bring their own lifeworld experiences into the curriculum, and 'applying creatively' which takes them back to apply what they have learned to real-world settings (Cope & Kalantzis, 2015d). In these ways, every moment of learning also a moment of mutual knowing-each-other between teachers and learners. The lifeworld comes to school and the school comes to the lifeworld. So the business of building learner profiles is continuous and embedded into the logic of the curriculum itself. The curriculum is an opening for students to voice their lifeworld experiences and life aspirations. This continuous voicing is an essential part of engaging effectively with learners' lifeworld experiences. Then, beyond recognition alone, the school must extend these experiences? It must expand learners' horizons.

To create differentiated curriculum
The old, one-size-fits-all, on-the-same-page curriculum is no longer appropriate in the era of civic pluralism. Nor is it pragmatically necessary in the era of digital media and computer-mediated learning . These new media make differentiated instruction more practicable. Learners can be doing the same thing at their own pace, or they can be doing different things according to their needs or interests. Such is the objective of adaptive, personalized or differentiated instruction which calibrates learning to individuals (Walkington, 2013;Wolf, 2010). However, we need to remain alert to ensure the comparability or equivalence (albeit not sameness) of educational and social outcomes.
To provide recursive feedback Learner differences demand customized feedback based on differential interests, chosen topical focus, and learning needs. Recursive feedback or formative assessment (Black & Wiliam, 1998;Wiliam, 2011) provides responses to learners that are always calibrated to the specifics of who they are, and the knowledge they are representing in their learning. In the era of inexpensive and accessible social knowledge technologies, no learning environment should be without always-available feedback mechanismsmachine feedback and machine-mediated social feedback (Cope & Kalantzis, 2013;Cope, Kalantzis, McCarthey, Vojak, & Kline, 2011).

To recruit learner agency
A more inclusive approach will recruit learner agency, subjectivity and identity as an energy that drives learning. Inclusive education conceives schools as knowledge-producing communities. It fosters in learners a sense that they themselves are knowledge producers rather than consumers of knowledge that has been pre-processed and transmitted to them, facts and definitions and theorems to be committed to memory and correctly applied. Instead of having textbooks, inclusive curriculum has students summarize the world, navigating and critically evaluating the plethora of sources available online and in community experience. Such has been the long-cherished the aim project-based learning (Kilpatrick, 1918;Rugg & Shumaker, 1928;Waks, 1997). Instead of ingesting a singular narrative that is textbook content, learners negotiate, synthesize a diverse range of sources and perspectives. Then, they write the textbook, figuratively speaking. In so doing they are reinventing the world, no more or less than an expert does (Cope & Kalantzis, 2015a).
Inclusive curriculum must also open out alternative starting points for learningwhat the learner perceives to be worth learning, what engages the particularities of their identity. It must allow for alternative forms of engagementthe varied experiences that need to be brought to bear on the learning, the different conceptual bents of learners, the different analytical perspectives the learner may have on the nature of cause, effect and human interest, and the different settings in which they may apply or enact their knowledge (Kalantzis & Cope, 2010). It must cater to different learning orientationspreferences, for instance, for particular emphases in knowledge making and patterns of engagement. It must offer different modalities in meaning-making, embracing alternative expressive potentials for different learners. And it must provide for alternative pathways and destination points in learning.
If we could allow this much scope for learner agency, we will allow a thousand differences to bloom. We will also create a more powerful sense of belonging in the processes of learning and the institutions of education. The paradox of belonging in the era of civic pluralism is that this belonging must be in all-your-difference. The more we take agency into account, the more multifarious its manifestations become and the more complex its intersections and matrices. Moreover, not simply to leave differences the way they are along with their underlying inequalities, inclusive education also means that learners have more opportunities to jump out of the rut of narrow lifeworld destiny, opening their horizons of possibility and their potentials for self-transformation.

To nurture productive diversity
Centering educational energies on learner agency in all its variety will also nurture a new ethics of collaborative learning. When lifeworlds are so varied, diversity of perspective becomes a resource for learning. Knowledge construction and learning become all the more potent for their productive engagement with diversity. This is the basis for learning and knowledge ecologies of 'productive diversity' where navigating differences of experience, perspective, paradigm, expertise is integral to the work of knowledge making and learning. We call the generative synergies of difference 'productive diversity' (Cope & Kalantzis, 1997a;Page, 2007). The educational objective is to support the self-development of kinds of persons who have the capacity to learn and act in particular ways. They can navigate change, negotiate deep diversity and make and lead change rather than be helpless to its forces. They can engage in sometimes difficult dialogues. They can compromise and create shared understandings. And they can comfortably extend their cultural and knowledge repertoires into new areas. They are tolerant, responsible and resilient in their differences.

To measure comparabilities
Learner transformation is a central mission of education. It occurs through the extension of the learner's repertoire of knowledge and capacities. It involves boundary crossing and expanding their horizons in a world of differences. This does not mean having to leave one's old self behind as was the case in the days of assimilation. Inclusive education is about learner self-transformation, learners acting as an agents to transform their worlds. How, then, do we create forms of assessment and evaluation that enable learners to meet high standards and can tell us in meaningful ways how learners have grown through their learning experiences? How do you measure progress in achieving education's most basic promises, for individuals and the groups to which they belong? The answer must mean using innovative assessment and evaluation practices that provide meaningful feedback such as portfolio evaluation, peer-or self-review, and 'big data' analyses of learner progress (Cope & Kalantzis, 2015c). The measure for all of these innovative assessment modes should not be identical 'standards' but comparabilitiesevidence in the form of learning activity and knowledge representations which can never be the same, but may be comparable in terms of epistemic effort and intellectual outcomes.

In practice: towards inclusive Roma education
Returning now to the starting point of this analysis, the project we were evaluating. Between 2010 and 2014, a project team led by Litsa Tressou, Soula Mitakidou and Yiota Karagianni conducted an intervention, 'Education of Roma Children in the Regions of Central Macedonia, West Macedonia and East Macedonia and Thrace'. The project worked in 39 Roma communities, involving 78 schools, and engaging 96 social workers. It aimed to achieve improvements in Roma education that might be termed 'inclusive'. Actions included: preschool education in order to ease the transition to school; the use of community-based psychologists to support children and families with a view to improving the rate of attendance of school-aged children; efforts to support access to second-chance schools and adult education, particularly for women; attempts to connect families with the local community and school; the use of information and communication technologies to network communities; two international conferences on Roma education; and a cultural program that included two festivals of Roma culture and music at the foot of White Tower, the icon of Thessaloniki. These actions are described in detail elsewhere Karagianni et al., 2013;Mitakidou et al., 2015).
The project finished in 2014 when EU funding ended. With this, the initiatives that had been developed during these years also came to an end. So where, after this, is Roma education in Northern Greece measured in terms of the ambitions for an inclusive education? Without sustained attention and resources, inclusive education for Roma is going almost nowhere. And what has been achieved in the project? As much as possible under circumstances constituted by the weight of history and the present crises of Greek and European society. Some Roma, however, have at the very least experienced an inkling of educational and social possibility.

From theory to practice: towards social and educational transformation
Over the course of modernity, we have experienced virulent moments of exclusion of based on human differences. We have also experienced attempts to assimilate differences on the working assumption that modern society and its institutions function more effectively when their members conform to the canons of sameness. The subsequent experiences of exclusion and assimilation are ongoing. Modern schooling was founded on these anthropological presuppositions, and we live with that legacy today.
Processes of exclusion and homogenization have, however, always encountered resistances. As they unavoidably become manifest, the contours of the differences have been named by resistors and proponents alike, using the metacategories that we have defined and analyzed in this paper. After centuries of resistance, this is perhaps a moment of paradigm shift, demanding not only a renewed recognition of differences, but also redress of inequalities in the distribution of material resources that are a key dimension of difference. If so, we need to extend our categories in order to capture the complexities of identity and inequality. And we need to reform our schools so they no longer attempt to frame differences through their legacy architecture of sameness. The task of the moment is to redesign schooling around an architecture of productive diversity and inclusion.

Disclosure statement
No potential conflict of interest was reported by the authors.

Notes on contributors
Mary Kalantzis is Dean of the College of Education at the University of Illinois, Urbana-Champaign. She was formerly Dean of the Faculty of Education, Language and Community Services at RMIT University in Melbourne, Australia, and President of the Australian Council of Deans of Education. With