IDEOLOGIES OF DISADVANTAGE: LANGUAGE ATTRIBUTIONS IN A DUTCH MULTICULTURAL PRIMARY SCHOOL CLASSROOM

Current ethnographic research shows that Dutch educational policy is caught between two positions. First, it constructs pupils from immigrant minority groups as educationally disadvantaged and, as a consequence, fosters mainstream (language) education as the means for their social integration and emancipation (Bezemer 2003). Second, it leaves Dutch primary school teachers with the challenge of dealing with the cultural and linguistic diversity brought about by their pupils (Bezemer & Kroon 2008; Spotti 2007). Against this background, the present article, stemming from a larger comparative ethnographic enquiry in the Netherlands and Flanders, focuses on the analysis of the discourse of a Dutch native primary school teacher in a multicultural classroom in the Netherlands. By means of socio-culturally informed discourse analysis (Gee 2005), it is shown that the identities of immigrant minority pupils are constructed, in the class teacher’s discourse, on the basis of language attributions that find their pivotal point in ideologies of language disadvantage provoked by the lack of Dutch language skills on the part of these pupils’ parents. The analysis, however, indicates that at the level of the discourse that populates the classroom, the ideologies that lay beneath the language attributions through which these pupils’ identities are constructed are eroding. Such erosion might also hold consequences for the way in which immigrant minority pupils’ identities are constructed in the discourse of Dutch governmental institutions.


Introduction
The separation of people from their native culture through physical dislocation as refugees, immigrant guest workers or expatriates as well as the dissolution of colonisation processes have been formative experiences of the last century for many Western European nation-states.
In 2005, the year in which this case study was carried out, it was estimated that in the Netherlands out of a total population of slightly more than 16 million inhabitants, 3.1 million had at least one parent born outside the country (CBS 2006). The last century's immigration phenomena are not only tangible through numbers but also through current political and public discourse. On the one hand, immigrant minority group members addressed as westerse allochtonen (western non-indigenous people), are thought to share a common European history and a 'European' identity (cf. Extra & Spotti 2008). On the other hand, immigrant minority group members addressed as niet-westerse allochtonen (non-western nonindigenous people) -mostly Turks, Moroccans and more recently Somali -are presented as people in need of societal and linguistic integration. From these two examples, it appears that Dutch public discourse is armoured with a 'jargon of minorities' (Extra & Gorter 2001: 5) through which immigrant minority group members, their descendants, their cultural backgrounds and their languages hit the headlines. As a consequence, the Dutch public discourse constructs immigrants and their descendants as other than the majority group, and their languages as other than the majority language (Kroon 2003: 40). These attributions of otherness are also present in (primary) education and go beyond mere jargon alone. The attribution of linguistic resources -or lack thereof -to one group of pupils rather than another is an endemic feature present in the Dutch educational discourse. Although these attributions often remain unarticulated, they are still informative. By functioning as index of language abilities, they tacitly inform the way in which immigrant minority pupils' identities are constructed as 'good' or 'bad' users of the dominant language. In contexts characterised by immigration and globalisation, the attribution of linguistic resources may appear difficult in that it may not be totally exhaustive of the heterogeneous language repertoires, styles and genres that pupils bring about in the socio-cultural spaces they inhabit (cf. Bezemer 2007;Gogolin & Kroon 2000;Jaspers 2005;Kroon & Sturm 1996;Spotti 2006a).
Against this background, we focus on how a Dutch-native primary school teacher constructs the identities of her immigrant minority pupils in a regular multicultural classroom. More precisely, this paper explores the ideological complex nested in the attributions of linguistic resources (or lack thereof) as proposed by the class teacher. The attributions of these linguistic resources indicate that the class teacher tries to make sense of the pupils' multilingual realities through a monolingual lens (cf. Gogolin 1994;Spotti 2006b). Our analysis reveals that the ideologies behind the attribution of language disadvantage, provoked by these pupils' parents lack of skills in Dutch, are under erosion. Two considerations are made in discussing these outcomes. First, we consider whether it is feasible or indeed necessary for (primary school) teachers to be aware of how language ideologies work. Second, we ponder on the consequences that this erosion holds for the way in which immigrant minority pupils' identities are constructed in the macro-discourses of Dutch educational institutions.

Conceptual framework
Central to identity construction is categorisation. Categorisation is a process that involves 'identifying oneself (or someone else) as someone who fits a certain description or belongs to a certain category' (Brubaker & Cooper 2000: 17). Once made operational, categorisation leads people to construct their own and/or someone else's identity as a member or as an outsider of a given community. This membership happens on the basis of the fulfilment of certain characteristics which consist of thinking, acting, valuing and interacting, in the 'right' places, at the 'right' times through the use of the 'right' objects, including language, in ways that are considered appropriate for community members. In other words, the characteristics that someone ought to fulfil constitute the conditio sine qua non for someone to subscribe him -or herself and/or ascribe someone else as a community member (cf. Carbaugh 1996;Holland & Quinn 1987;Wieder & Pratt 1990). However, people's judgment of what is appropriate in order to be considered a community member does not happen just because. Rather, this judgment relies on the basis of what Gee (1999: 43) calls 'discourse models'. Discourse models are explanatory theories of mind, either idiosyncratic or culturally transmitted, that people hold to make sense of the world around them. They are formed on the basis of those associative networks that people have been part of throughout their lives. Discourse models are channelled through discourse where discourse is understood as the whole of possible forms of expression, e.g., oral, written, pictorial and multimodal, produced by the action of an institution and/or of an individual within a particular socio-cultural space (cf. Blommaert 2005).
The array of discourse models that people may hold is wide. Discourse models may range from the rituals that someone should follow for having a cup of coffee in a certain socio-cultural space to why certain gestures are applicable and others are not when engaged in a PhD viva with an opponent. In relation to language, the discourse models people may hold about the language or languages someone speaks, and the linguistic resources someone may or may not own, supply a means through which identities are constructed and negotiated, along with membership of certain communities. Ideologies of language and identity guide the ways in which individuals use linguistic resources to index and/or conceal their identities as well as to attribute the use of linguistic practices to others. The discourse models that guide the analysis presented in this paper have a metonymic function, i.e., they are the pars pro toto of larger language ideologies that are nested beneath the attribution of linguistic resources.

The study
The present study has adopted a sociolinguistic-ethnographic perspective (Creese 2008;Erickson 1986). Such perspective is best described as wanting to investigate " [...] what people are, how they behave, how they interact together. It aims to uncover people's beliefs, values, perspectives, motivations, and how all these things develop or change over time or from situation to situation. It tries to do all this from within the group, and from within the perspectives of the group's members" (Woods 1986: 4). Within this ethnographic perspective, the study has aimed at understanding the construction of immigrant minority pupils' identities in the discourses of a Dutch-medium primary school teacher, in terms of these pupils' cultural, ethnic, religious and linguistic belongings. Further, it has sought to shed light on how the identities constructed through the teachers' discourses about her pupils' language attributions can be understood in relation to the meso-discourses held at school level and to the Dutch macro-discourses of cultural, ethnic and linguistic otherness. The study was designed so to produce a 'cultural ecology' of the classroom (Rampton et al. 2004: 2) and at the same time, to adopt a critical perspective, that is a perspective that questioned the normative nature of the macrodiscourses in which the investigated classroom was inserted. From the outset of the study, care was taken to comply with research ethics. Pseudonyms were used for the school, the class teacher and the pupils, to preserve confidentiality, informed consent was sought from the parents of all participating pupils, and all interview transcripts were authorised by the interviewees.
The fieldwork started on February 15 th , 2005, when the first author visited St. Joseph Catholic Primary for the first time and explained the purpose of the study to the school Head and to Miss Sanne, the class teacher of Form 8a. After gaining both their approvals, one month was spent in Form 8a as a (non-participant) observer. In order to establish a working relationship with the teacher and allow the pupils to get used to the presence of a stranger in the classroom, this month of fieldwork was gradually built up from two days a week up to a complete school week. In that month, classroom events were audio recorded for a total of 54 hours and 46 minutes. Following the writing up of the field notes in a synoptic format, supplemented by the transcription of all the audio-recorded events, interviews were carried out with Miss Sanne. The main interview was based on the model of the long open-ended interview (McCracken 1988: 9). This was done to explore the class teacher's biography and her primary schooling experience and professional career. In this way, a body of knowledge was gathered that would permit us to identify the associative networks that had populated the class teacher's life. Altogether, four interviews were carried out with Miss Sanne. These were all audio recorded and, soon afterwards, they were transcribed and made available to the teacher for confirmation of content and accuracy of transcription. Once authorised, the transcripts were analysed using Gee's (2005) socio-culturally informed discourse analysis. The aim was to identify in the teacher's discourse, those informal theories of mind, i.e., discourse models (Gee 1999: 43) that contributed to the construction of her pupils' identities. This analytic work was done by means of a continuous sifting process. This involved reading the interview transcripts several times, and then identifying and coding those sections where the teacher's discourse models and language ideologies were most clearly manifested. In this paper, when the excerpts are taken from the interviews with the teacher, they are presented primarily in English with a Dutch translation below.
In 2005, Form 8a counted for eighteen pupils in total, eight boys and ten girls. The age of the pupils ranged from eleven to thirteen years due to some pupils repeating the school year. Thirteen pupils had attended this school since Form 1. Following the class teacher, all Form 8a pupils have an educational weight of 1.90. This means that, educationally speaking, because of their socio-ethnic backgrounds these pupils are as 'heavy' as almost two pupils with an educational weight of 1.0 which generally are pupils from indigenous Dutch educated parents.

Identities based on the lack of linguistic resources
In illustrating the background of St. Joseph's pupils, Miss Sanne starts talking about the district where the school is located asserting it to be a district "with many foreign families in particular also because here there are still very many rented houses" (S02: 256), where a rented house denotes lower incomes and therefore the presence of foreign families. Further, the discussion about the background of the pupils at her school develops as follows: Sanne: You also just notice it, right, if the parents have not followed absolutely any education at all. And some some families, they want it very much but they have let's put it simply (…) the children have gone a bit off track. And but you also have families there who are really well educated and those set the good example. Merk je ook gewoon hè, als de ouders totaal geen opleiding hebben gevolgd. En sommige (…) sommige gezinnen, die willen heel erg graag maar die hebben zeg maar gewoon (…) de kinderen zijn een beetje uit de band gesprongen. En maar je hebt er ook gezinnen bij die echt prima opgeleid zijn en (uh) die geven dan het goed voorbeeld. (S02: 273) In the utterances above, the educational level of the pupils' families becomes the central theme of Sanne's discourse. Parents are grouped in three categories. Those who have not followed any study, those who -even though eager to participatehave children who have gone 'a bit off track' and those who 'set the good example' because they are well educated. Soon, a link between the parents' educational level and their children's attainment is drawn:

Sanne:
And then you also really realize the difference in the child, right. First, Miss Sanne uses the presence of Dutch native families to help explain why the majority of the pupils at this school manage to attend, at the end of their primary schooling career, a prestigious type of secondary school. Second, Miss Sanne's statement 'I have not a single Dutch child in my class' is used as an explanation for why her pupils perform worse than those pupils at the other school. The lack of parental qualifications and these parents being non-native Dutch are at the basis of the informal theory of mind that Miss Sanne uses to explain St. Joseph's extra investment in language with a particular focus on vocabulary.
We now move further in the reconstruction of Miss Sanne's discourse models. First, we present the cases of two pupils, i.e., Mohammed and Lejla, whose language attributions marked the opposite ends of the category 'immigrant minority pupil with a language disadvantage'. Second, we present two pupils, i.e., Walid and Micheline, whose language attributions are in contrast with the discourse models so far reconstructed by the class teacher.

Mohammed
To give an example of the language disadvantage at St. Joseph, Miss Sanne starts off with Mohammed, a thirteen-year-old Somali child who attended Miss Sanne's Form 8a in the previous school year. At that time, Mohammed had been in the Netherlands since he was eight years old and "he was fluent in the Somali language" (S02: 314). However, proficiency in the Somali language turned out to be detrimental to Mohammed's Dutch language development because: At the age of eight, Mohammed was already fluent in his mother tongue, i.e., Somali, and he had to learn a second language, i.e., Dutch. As Miss Sanne reports in the coordinate phrase that follows, the Somali language has a different sentence structure to Dutch, which led Mohammed to use Somali's syntax in Dutch and to always speak 'in twisted sentences'. Mohammed's difficulties in speaking Dutch properly are found in the syntactical interference hypothesis where the second language learner inappropriately transfers structures of his first language to the second (cf. Van de Craats 2000: 335). As Miss Sanne adds:

Sanne:
And if you get it also at home, because that mother, she, of course, was also having problems with that herself En als je dat ook van thuis uit, want die moeder, die was, natuurlijk, daar ook mee aan het stoeien Max: (hmm) Sanne: and that father too, he also spoke hardly any Dutch. Mohammed not only uses 'strange sentences' in Dutch because his speech is based on the structure of Somali, a language that has SOV-order in its main clause in comparison with Dutch SVO-order (cf. Saeed 1999). Also, as introduced by the causative conjunction 'so', both Mohammed's parents are responsible for the syntactical interference among Somali and Dutch. The father, in fact, spoke no Dutch and the mother also 'suffered' from Somali sentence structure in her use of Dutch, both language situations that feed the discourse model of immigrant minority pupils' language disadvantage because of their parents' lack of Dutch proficiency.

Lejla
Miss Sanne's discourse dealt also with Lejla, an eleven-year old girl born in Bosnia-Herzegovina to Bosnian parents who came to the Netherlands when she was three years old:

Sanne:
Lejla she is also (…) let's see she has lived here ever since she was three or so, therefore also still really very young when she already a new language (…) look and small children can pick up a (…) another language really easily that is simply, yeah, scientifically proven. Lejla die is ook (…) even kijken die woont hier al sinds dat ze drie is of zo dus ook nog heel erg jong dat ze al een nieuwe taal (...) kijk en kleine kinderen kunnen heel makkelijk een andere taal oppikken dat is gewoon, ja, wetenschappelijk bewezen Max: (hmm) Sanne: and indeed she is also better in Dutch than other children and that is also be-cause her parents also just spoke Dutch at home from the beginning.

En zij is ook inderdaad beter in het Nederlands dan andere kinderen en dat komt ook omdat haar ouders ook gewoon vanaf het begin af aan hier gewoon ook thuis Nederlands praten. (S02: 443-445)
In the utterances above, Lejla appears to be in an advantaged position in picking up a second language because she came to the Netherlands at a very young age. Further, Miss Sanne tries to obtain objectiveness for her claim. In the utterance 'look and small children can pick up a (…) another language really easily', she uses the imperative 'look' to substantiate the evidence of her claim. Further, she calls upon the critical age hypothesis (cf. McWhinney 1992) implying that young children learn a second language more easily than those who approach a second language at an older age. Not only is the age at which Lejla came into contact with Dutch relevant; also her parents' language behaviour is now regarded as key element to Lejla's 'good' language development. Interesting to notice that Lejla's parental language behaviour is accompanied by the adverb 'simply'. The use of this adverb may indicate that the practice of speaking Dutch at home is regarded by Miss Sanne as nothing more than what parents should do by default with their children at a young age. However, at home, Lejla and her parents have a language repertoire that includes Croatian, English, Dutch and Bosnian. Bosnian is the language Lejla denotes as her language, and she claims to have both passive and active literacy skills. Further, she reports to use it for verbal exchanges with her younger siblings and with her parents.

Walid
It is also worth focusing on the case of Walid who, although assigned an educational weight of 1.9, is considered by Miss Sanne as a pupil with an educational weight of 1.0. It is during the second long interview with Miss Sanne that, while talking about Form 8a pupils, she expands on Walid's case and states:

Sanne:
Right, so there is one pupil with one point zero. But when they [his parents; MS/SK] came to an advisory meeting in preparation to secondary education that mother asked like how is that possible. I say and also he himself is therefore completely just one hundred percent Moroccan so in principle he is also simply one point nine. en hij is zelf ook gewoon dus helemaal honderd procent Marokkaan dus in principe is hij ook gewoon een punt negen. (S02: 287-289) Miss Sanne constructs Walid's identity as an exception to the mechanistic educational weight system. She singles him out from the rest of Form 8a as 'so there is one pupil with one point zero'. In the adversative clause that follows 'but it is (…) Walid is also simply a child of two Moroccan parents', Walid is ascribed, on the basis of ius sanguinis, to the Moroccan immigrant minority group. This ascription is confirmed through the coordinate clause 'and also he himself is therefore completely just one hundred percent Moroccan' that, as introduced by the causative connective 'so', stands for the fact that being Moroccan would automatically qualify him as a 1.90 pupil. Further on, through an adversative clause, Miss Sanne states:

Sanne:
But his parents are indeed very highly educated so (…) I do not know whether it has been a little mistake or that at the time they have asked for a one point zero but she was rather puzzled in my view so I think that it is a little mistake because all the children here in the class are offspring of foreign parents.

Max:
(hmm) okay so it could very much be that he is a one comma nine then. (uh) he is a really clever boy and his parents are also both (uh) highly educated and you really experience that.
(uh) het is een hele slimme jongen en zijn ouders zijn ook allebei (uh) goed opgeleid en dat merk je toch wel Max: (hmm) Sanne: they also speak correct Dutch at home so he also gets it taught well so yes. die spreken thuis ook correct Nederlands dus hij krijgt het ook goed aangeleerd dus ja. Max: (hmm) and what do you mean by correct Dutch? (hmm) en wat is volgens jou correct Nederlands? Sanne: Good sentence structure (...) (uh) yes good vocabulary.

Goeie zinsopbouw (…) (uh) ja goede woordenschat. (S02: 291-301)
Like the rest of Form 8a pupils, Walid should have an educational weight of 1.90. However, there are elements that trigger Miss Sanne's doubt as to whether an educational weight other than 1.90 could technically apply to him. It is true, in fact, that 'all the children here in the classroom are offspring of foreign parents' which is an indicator of these pupils holding an educational weight of 1.90. However, as shown by Walid's educational weight being referred to as possibly a 'little mistake', the teacher seems to reconsider Walid's identity ascription. Miss Sanne claims that Walid's parents are both highly educated and that he is 'a really clever boy'. Further, she adds that his parents 'also speak correct Dutch' that implies 'good sentence structure' and 'good vocabulary' and because of his parents' language practice, Walid 'gets it [Dutch; MS/SK] taught well'. It would seem that while technically a 1.90 pupil, practically Walid's case erodes the discourse model of parental language behaviour being a precondition for a pupil's language disadvantage. Matching therefore his mother's demand, Walid is 'in practice' an immigrant minority pupil who has an educational weight of 1.0.

Micheline
Further, the teacher addresses the case of Micheline, an eleven-years-old Antillean girl born in Curacao to Antillean parents who is among the brightest pupils of Form 8a.
Micheline is the only pupil of Form 8a who scored just above the school forecast for her primary schooling final examination results. She is also the only pupil of Form 8a who has been advised to attend havo (Senior General Secondary Education). In dealing with Micheline's case, Miss Sanne tries to unravel the reasons why she does not suffer from the language problems of Form 8a and she states:

Sanne:
She [Micheline] is she is also Antillean so she also has, right, in Curacao they also speak Dutch. Ze is ze is ook Antilliaans dus ze heeft ook, hè, in Curaçao spreken ze ook Nederlands Max: (hmm) Sanne: so her parents can also both speak Dutch. dus haar ouders kunnen ook allebei Nederlands. Max: (hmm) (S02: 425-428) At first, Miss Sanne categorises Micheline on the same level as other Antillean pupils on the basis of her origin as 'she is also Antillean'. The teacher then connects her being Antillean with the fact that 'in Curacao they also speak Dutch' and, as reported in the following causative sub-clause 'so her parents can also both speak Dutch'. If we were to follow the discourse model that Miss Sanne has so far proposed, having parents who speak Dutch implies the presence of Dutch at home and Dutch at home, in turn, explains Micheline's good Dutch language skills. In a latter adversative clause, Miss Sanne expands her reasoning to Antillean children and more generally about their Dutch language skills in general, when she states: Sanne: But I also notice that simply very often (…) with Antillean children that that language [Dutch] they simply posses it a bit better already because that, because Dutch is not really a second language Maar dat merk ik gewoon ook heel veel (…) bij Antilliaanse kinderen dat die taal er ook al gewoon wat beter in zit omdat dat omdat Nederlands niet echt een tweede taal is Max: (hmm) Sanne: Dutch is also their mother tongue along with Papiamentu. Nederlands is ook hun moedertaal met het Papiaments. (S02: 431-433) Through the opening statement that Antillean children posses Dutch 'a bit better already', Miss Sanne implies a term of comparison. Supposing that Miss Sanne is drawing a comparison with Mohammed, the pupil previously used in her discourse model as an example for the current limitations that Form 8a pupils face in Dutch, this could imply that Antillean children are better at Dutch than other immigrants. This is so because, as Miss Sanne states, for Antilleans "Dutch is not really a second language", meaning that Dutch for them is closer to a first language since "Dutch is also their mother tongue along with Papiamentu". The sociolinguistic position of Dutch on Curacao sketched by Miss Sanne is not that of a foreign language, as it is for most of the other immigrant minority pupils who have parents who do not speak Dutch at home. Rather, in Curacao, education has to cater for learning Dutch and learning in Dutch. Further, Dutch is used as a vreemde voertaal in society, i.e., a foreign language that is used in education and in other official situations but that in practice holds little sociolinguistic relevance in people's everyday life (cf. Narain 1998: 7). The generalisation drawn by Miss Sanne holds two consequences. First, given that it is common to all Antillean pupils to be 'a bit better at Dutch', it corroborates that Micheline is good at Dutch because Dutch is almost a first language for her. Second, it confirms the discourse model that sees parental Dutch language skills as supportive to pupils' good Dutch. Micheline, however, is not the only Antillean pupil in the class. It is at this point that erosion comes into play. Miss Sanne's discourse model about the Dutch of Antillean pupils and Micheline's good results was based on her parents' use of Dutch. Joshwa and Rhonda, the two other Form 8a Antillean pupils, were also born in the Netherlands and, like Micheline, they both speak Dutch at home with their parents. However, they both attend Leerweg Ondersteunend Onderwijs (Learning Supportive Education) and are both ascribed by Miss Sanne as "weak" pupils (S03: 127). This leads Miss Sanne to add a second explanation for Micheline's good results: "And yes where else does it come from, I think that it is also simply part of her nature" (S02: 429). However, this second explanation is not related to Micheline's ethno-linguistic background and parental language skills. Rather, it boils down to Micheline's intrinsic nature.

Conclusions and discussion
Mohammed, to whom the rest of the pupils of Form 8a were compared, appeared as the prototype of the immigrant minority pupil with a language disadvantage due to parental language practices. He had come to the Netherlands when he already mastered Somali and, following Miss Sanne's discourse, it was because of his mastering of Somali that he encountered syntactical and vocabulary limitations in Dutch. Further, he was not sufficiently exposed to 'good Dutch' because his father spoke Dutch with funny sentences and his mother's spoken Dutch suffered of syntactical problems too. From Mohammed's case, Miss Sanne moved to Lejla, a Bosnian pupil of her current Form 8a who came with her parents to the Netherlands at the age of three. Within the discourse model of immigrant minority pupils with a language disadvantage, Lejla appeared to be the opposite of Mohammed. Lejla's Dutch was good and, following Miss Sanne's discourse model, her Dutch was good because her parents 'simply' spoke Dutch at home and also because she has learnt Dutch at a young age. However, Lejla's own linguistic resources differed from those formulated by Miss Sanne. Lejla addressed Bosnian as her own language and she reported to use it extensively with her immediate siblings and parents. After having discussed Mohammed and Lejla, Miss Sanne's discourse turned to Walid and Micheline. Following the discourse model so far reconstructed that sees abundant contact with immigrant minority languages and parental lack of Dutch proficiency as deterrent for the pupils' Dutch, Walid would be a 1.90 pupil. 'Technically speaking', in fact, Walid has Moroccan parents and therefore is a hundred percent Moroccan himself. Yet again, Miss Sanne's own experience differs from the 'technical' aspects that would construct Walid's identity as a 1.90 pupil.
Walid 'in practice' is a smart boy with highly educated parents who, in the teacher's view, speak good Dutch. This last reality comes to erode the discourse model that Miss Sanne had so far drawn about parental language practices, the language attribution of immigrant minority pupils and the construction of immigrant minority pupils' identities. Walid's educational weight should be a 1.0. For the teacher, though, it still remained difficult to grasp how an educational weight of 1.0 could be possible for a pupil who like all the other pupils in Form 8a is a descendant of foreign parents. Finally we have Micheline, an Antillean pupil born in Curaçao to Antillean parents and grown up in the Netherlands. Following Miss Sanne's discourse, Micheline was among the brightest pupils of Form 8a and she did not have a language disadvantage. This was so for two reasons. First, in agreement with the model of pupils' language disadvantage because of parental lack of skills in Dutch, Micheline's parents spoke Dutch and therefore their language behaviour catered for her good results. Second, following Miss Sanne's own 'ethnic hierarchy' (cf. Verkuyten, Hagendoorn & Masson 1996) combined with the attribution of linguistic resources, Antillean pupils possess Dutch 'a bit better' than the rest of the immigrant minority pupils of Form 8a. This was so because Dutch is not really a second language to them, rather it is 'also their mother tongue with Papiamentu'. The other two Antillean pupils of Form 8a, although they also had Dutch at home, were categorised as 'weak' pupils. Their connotation of 'weak' pupils then goes against Miss Sanne's attribution of linguistic resources that saw Antillean pupils possessing Dutch a bit better. As a way out, the teacher did not make the link anymore between Micheline's good results, the Dutch spoken by her parents at home and the fact that Dutch is almost a first language for Antilleans. Rather, she backed it up by referring to Micheline's own 'bright' nature.
The reconstruction we propose shows how in a regular multicultural primary school classroom, the teacher's attributions of linguistic resources (or lack thereof) construct students' multilingual realities through a monolingual lens. The lack of Dutch language skills at home is linked to the limitations that these pupils experience in their Dutch in the classroom. Alternatively, the use of Dutch at home is proposed as the 'good'/'normal' way. The language attributions and the indexical order of identities that emerged from Miss Sanne's discourse -although general human principles that people use for ordering the world -are reminiscent of the last three decades of work carried out in sociolinguistics and education (cf. Keddie 1971;McDermott & Gospodinoff 1979) that saw the sociolinguistic background of pupils matched to their ascribed ethnic identities and, together with the latter, became fertile ground for preconceived barriers to school success. However, Miss Sanne's attributions, the discourse models on which they are based, the language ideologies nested within them and the erosion detected once she engages in reflecting upon her classroom experience make us wonder whether, in the training of teachers, more attention should be paid to ideologies and their workings.
Furthermore, the erosion reconstructed in the classroom might hold consequences for the way in which immigrant minority pupils' identities are constructed in the macro-discourse of Dutch educational institutions. If (primary) educational discourse is contingent on giving accurate attributions of pupils' language repertoires and resources, which it often is, mismatching in language attributions can cause the construction of a priori disadvantaged identities in these pupils' schooling trajectories (cf. Kroon & Vallen 2006). The sooner these processes of attribution of linguistic resources and categorisation of identities are made aware of the working of ideologies, the greater will be the chance that (primary) education understands objectively its students' communicative world, their organisation of multilingual repertoires and the particular sociolinguistic economies that characterize pupils' background in an era of migration and globalisation.