New speakers of new and old languages: an investigation into the gap between language practices and language policy

This paper explores the intersection of new speakers in conditions of globalisation led mobility and it investigates the implications the phenomena may have for language policy making. It first describes two historical phases in language policy development that are closely related to a sociolinguistics of stability. In this, it criticises how present-day language policy is attached to specific time and space constraints whose focus is a by now outdated concept of language and of speaker as its prescriptive objects—thus leading institutional language policies to not being ‘in sync’ with contemporary new speakers’ socio- and geo-political movements and developments. This proposition is illustrated in two case studies, both located in the Netherlands and dealing with the language practices and connected policies of two types of new speakers. The first case deals with the experiences of asylum seekers being engaged with ‘techno-literacies’. That is asylum seekers being part of ICT assisted classes for civic integration through the learning of Dutch (new speakers of a new language, learning through new means of language learning). The second case deals with Chinese students who are fully proficient in Dutch, attending language heritage classes for learning Mandarin through book based lessons (new speakers of an old language, learning through old means of language learning). In both cases, the observed language practices and meta-pragmatic judgements of the individual language users elect them as initiators of bottom-up sociolinguistic change that, while offering grassroots solutions for local challenges, also plays a role as local evidence for informing future top-down language policy development.


Introduction
Urgesh: Look at this Sir, look at this Wassif: These are cool bruv, these are cool Urgesh: I have seen them on a gig Wassif: Yeah yeah, look at that, power, broder Max, pure power (Asylum 2.0, field notes 14102013) The above quote is taken from the fieldwork log notes of a project aimed at investigating what it means to be an asylum seeker in an age of globalization. Its main characters are Urgesh, a young Bengali asylum seeker engaged in watching a YouTube video, and Wassif, an Afghani asylum seeker, both of them at an asylum seeking center in Flanders, the Dutch speaking part of Belgium, the place where these two young men wish to stay once their refugee status will be approved. In the excerpt, these men are convivially commenting the video using their own varieties of English-as the Bengali band on the screen, called Sultana Bibiana, plays a cover from the American band Metallica while performing at a concert in Indonesia. In the above quote, several are the issues at play. First, as exemplified by the total absence of Dutch in the exchange, we do not see any trace of State driven language policies being taken on board by the two language users. Second, we see that the interaction at hand implies that the interlocutors are rather proficient language users of an international language, i.e. English. Last, we can also observe that they are proficient techno-literates in that they use the internet as a means for accessing popculture content (Kurvers and Spotti 2015).
The situation pictured above confronts us with questions that are fundamental for the field of language policy. First, is there any purpose or room left for institutional top down language policies and language policing (Blommaert et al. 2009) in contexts that are characterized by globalization led mobility and its inherent complexity? And, if so, what role can these policies have in the life of those people, who do not necessarily belong to the sociolinguistic mainstream, in that they hold either an indigenous minority background or a globalized migratory background? Last, what does this say about the newly coined term 'new speaker' (O'Rourke et al. 2014(O'Rourke et al. , 2015(O'Rourke et al. , 2017O'Rourke and Ramallo 2011)? Is it yet another incremental term for describing sociolinguistic phenomena caused by globalization led mobility or does it have a transformative potential to itself? These questions, we feel, can be of help toward the formulation of a refreshing answer to what is the aim and object of language policy, i.e., what concept of language it should actually embrace, what it should wish to achieve, and what kind of speakers it does have in mind. For finding an answer to these questions, however, a reappraisal of what the object of language policy actually is, i.e., a reappraisal of what language actually has come to be in the sociolinguistic conditions of globalization (see Pennycook 2017: 125), is deemed necessary. Against this background, our contribution deals with new speakers of new and old languages in the Netherlands. In doing so, it also deals with the means through which these new and old languages are learned in the classroom.
After a brief introduction on language policy and its paradigms, we venture in exploring some of the terms within the gamut of contemporary sociolinguistics' terminology. From there, we plead for replacing the concept of 'language' as the traditional object of language policies with the concept of 'language register' that we see as more fit for describing contemporary language practices in times of globalization and developing relevant language policies. In order to substantiate our conceptual claims, we present data from two ethnographic case studies that deal with new speakers, language classrooms and language policing. The first case deals with the experiences of asylum seekers being engaged with techno-literacy. That is, with the experiences of asylum seekers being engaged with an ICT supported class for the development of their civic integration thanks to the learning of Dutch (new speakers of a new language through new means of language learning). The second case deals with Chinese students who are fully proficient in Dutch, attending Chinese language heritage classes and learning Mandarin through the means of a book, i.e., new speakers of an old language learning through old means of language learning (Li 2016). In our approach, we consider the observed educational language practices and meta-pragmatic judgements of the individual language users involved and we see them as initiators of bottom-up sociolinguistic change that while offering grassroots solutions for local challenges, also plays a role as local evidence for informing future top-down language policy development.

Language policy for stability and homogeneity
In the field of language policy development there can be distinguished two main consecutive historical phases, these being (1) language policy connected to decolonization processes after World War II and (2) language policy connected to immigration movements starting in the early 1970s.
The first phase of language policy making, is connected to worldwide decolonization related issues in newly emerging nation states. Here the approach to language had a dual perspective. First, to establish a national language as a binding element of the new nation state following the Herderian constellation of one language, one nation, one state, one territory (Gal 2006: 378) and secondly to give a role and a place to all of the (other) indigenous languages present in the newly born nation state. From this standpoint, language policy and planning try to cater for language related questions from a strongly Fishmanian understanding of the sociology of a given language as a finite entity that stands to symbolize national identity and ethnolinguistic identification (see e.g. Fishman 1989). Through this approach language policy and planning become tools for fostering sociolinguistic stability, finding its pivotal point in the homogeneity of a nation and its language. Connected to the establishment of national languages, at the same time we also see language policy and planning as instruments in an ecologically or otherwise motivated endeavor for the preservation, maintenance and (even) revitalization of minority languages in an effort to prevent language loss and language death (Ricento 2016). Although at first sight, language policy making for safeguarding minority languages might appear to be the opposite or even in conflict with establishing a national language, such endeavors basically originate from the same principle: the ambition to define also for minority languages an ensemble of language status, corpus and acquisition in a fixed time and space, fulfilling again the purpose of constructing societal stability. The use of the term planning in such contexts subsumes another set of thoughts around languages and nation states, i.e., the perspective of language diversity (i.e. people using a variety of majority and minority languages) as a problem and planning (i.e. defining fixed forms, functions and spaces of use for these languages) as a solution. This approach, irrespective of the societal changes that have taken place thanks to globalization, seems to be still in place across many nation states as well as across many of their institutional environments, e.g. education, that necessitate language planning.
The second phase in language policy making has emerged in response to migration to Europe, brought about by international decolonization waves, an intra-European demand for cheap manual labor and by unpredictable, yet ongoing, movements of asylum seekers brought by conflicts, poverty, climate jeopardy and political fear of prosecution. In this second phase, language policy has tried to present language as a fundamental human right and bilingualism, later on addressed as multilingualism, as a fact of life, as something to be nurtured and preserved (see Spotti and Blommaert 2017). This wave of language policing started with the 1977 Council of Europe Directive on the Education of the Children of Migrant Workers and made legal room for second language teaching as well as immigrant minority language teaching in the Community's member states (see Reid and Reich 1992). Irrespective of the fact that education is a strongly regimented institution language wise, what teachers and students do with language and how they use language in education, may be well far away from the normative principles spelled out by language planning. Although language policy in handbooks is (still) mostly dealt with in terms of Cooper's (1989) distinction of status, corpus and acquisition planning, which informed much of the work done in the two phases of language policy making reported so far, ethnographic research into language policy has, in the meantime, made abundantly clear that there is no such thing as a straightforward and unproblematic implementation of top-down language policies (cf. McCarty 2011). Even if Cooper's work has by now been engaged in a much more sophisticated and critical fashion (see Gogonas and Michail 2014;Hornberger 2006;Pennycook 2017) it remains that status, corpus and acquisition still mainly focus on a rather fixed understanding of what language is. An understanding that no longer fits the sociolinguistic practices of contemporary language users engaging in trans-and pluri-lingual languaging (Arnaut et al. 2016(Arnaut et al. , 2017Blommaert 2010;Blommaert and Rampton 2016;García et al. 2017a, b).

Language policy and globalization led mobility
In addition to the above considerations regarding the linkage of language policies with a specific chronotopicity of institutions' and people's own language needs (Blommaert 2017) that is potentially hampering the effectiveness of such policies in the long run, there is also another potentially disturbing aspect, i.e., the top-down-bottom-up divide in language policy making. If we ponder for a moment the example of a teacher working in a multilingual classroom, a bottom-up perspective would entail this teacher analyzing the problem she is facing, coming up with a local solution and then implementing it in her classroom. The day-to-day language teaching practices that result from this process are basically a reflection of local bottom-up language policies. Johnson (2013) characterizes such bottom-up policies as micro-level, covert, implicit and de facto policies whereas top-down policies are instead macro-level, overt, explicit and de jure policies. Bottom-up and topdown policies more often than not, do appear to be very different from one another, although trying to find a solution to the same problem. This mainly relates to a difference in perspective. The practitioner in everyday practice is faced with participants and their problems that cannot wait for policy makers first going through all the different stages of the policy cycle before reaching a possible solution (Howlett et al. 2009). The policy maker, on the other hand, is in a position that aims to come up with an approved policy plan. As Spolsky (2004) noted, language policy can indeed take the shape of a text with rules and regulations, it however has to be taken up in people's language practices and attitudes as well. People's language practices, in as much as their attitudes and their meta-pragmatic judgements, do not necessarily coincide with what has been decided upon in language policy documents dealing with the same issue. The ideas and practices of those involved in solving language related problems on the ground are basically a reflection of bottom-up, grassroots policies that provide local solutions for local challenges, but at the same time can play a role as local bottom-up evidence for a certain approach in the traditional process of top-down policy development.
Where 'language' is defined as a verb, i.e., 'to language', rather than a noun, the concept of a speech community comes to lose its value and language users are instead understood as being part of communities of practice and knowledge that share and use language repertoires and registers consisting of a continuously changing ensemble of linguistic features, stemming from their own and others' linguistic resources, that are then mapped onto the sociolinguistic functions one has to fulfil when engaged in a communicative exchange. Language policy making still seems to lag behind, engaged, as it is, in defining the function, form and teaching of specific, well-defined languages in a specific time and place. As a consequence, language policy making loses the connection to new speakers' everyday sociolinguistic practices and realities. As a remedy to this problem, we propose to bypass the a priori assignment of specific sociolinguistic statuses to specific linguistic-communicative resources. Such statuses-e.g. the assumption that certain resources are "English", "standard English" or "standard American English" (see Pennycook 2017) and the assumption that the identities of those who use these varieties fall within certain categories-should on the contrary become the object of ethnographic interpretive inquiry. This calls into question the presupposed stability of "-lingualism", as in "bi-" or "multi-", and gives priority to language being understood as a set of sociolinguistically performed empirically observable forms mapped on to functions in which new speakers become not only confronted with "languages" and "codes", "-lects" and "registers" but also with the ideological by-product and the means through which someone learns 'language' (see Spotti and Blommaert 2017). We illustrate the above by discussing two case studies in which top-down perspectives on language policy making turn out to be 'outdated'. We plead for an approach that focuses on language practices on the ground where the understanding of language is re-evaluated and new perspectives for language policy making are being shaped (Jaspaert 2015).

New speakers of a new language through new means for language learning
In both the public and political discourses on integration of newly arrived migrants in the Netherlands, we find a language policy effort that does imply integration through (inviting) new speakers to learn the (official) language of the host country. While the Dutch discourse regarding integration used to be a discourse combining Dutch as a Second Language as well as ethnolinguistic and ethno-cultural identity preservation in language classes for students with an immigrant minority background (Bezemer 2003;Spotti 2007), the Netherlands has meanwhile witnessed a discursive shift toward the learning of Dutch as the main way for integration into mainstream Dutch society. On top of this, recent language policy efforts both at a European and national level (EU Commission 2006) as well as the formulation of digital competences for all European citizens (Ferrari 2013), have brought in the use of ICT as the means through which faster linguistic and socio-cultural integration can be achieved. Although regulated by a ministerial decree in 2007 asking to make integration trajectories shorter, the present field of ICT and integration through Dutch as a second language learning comes across as a 'Far West' scenario where private and public providers take the lead on the basis of their low costs in hosting (online) Dutch as a second language classes for newly arrived migrants (see also Spotti 2011 for a critique on the emergence of this sector). In what follows, we examine a case of ICT supported learning of Dutch as a second language. In doing so, we deal with a dual learning ICT supported method (ITpreneurs 2007) called 'The Virtual Neighborhood' (De Virtuele Wijk) and we present, through the means of focused group discussions, migrant learners' own perception of language learning and their reaction to the language policy implementation that compels integration classes to use ICT.

'The Virtual Neighborhood' in the classroom
In the following, we re-appraise some results emerging from an ethnographic interpretive enquiry into migrant students' drives and barriers toward ICT. This research was conducted at a regional educational center, here called Vondenland, back in 2010 (see Driessen et al. 2011). The Dutch language courses provided at the center that used 'The Virtual Neighborhood' were split in two streams: Dutch Language (Nederlandse Taal) and Civic Integration (Inburgering) (cf. Vondenland 2010). The civic integration course is aimed at newly arrived migrants who are engaged in an integration trajectory. Courses in preparation to the integration exam, including language production and reading comprehension have a multimedia component, through which the learner 'learns the Dutch language with the help of exercises and films on the computer' (cf. Vondenland 2010: 3). ICT supported second language learning can be carried out in class, in the open access learning center or at home. If a student wishes to pursue a paid job she will be assigned to the stream 'Integration and Work'. If a student instead wishes to be engaged in child rearing and housekeeping tasks, she is assigned to the stream 'Integration and Pedagogical Support'. As stated by the Vondenland center's own policy: in order to be integrated students must learn the Dutch language and build knowledge of Dutch society. Integrating means being able to function in practical situations, as -for instance -a talk with your GP, with a teacher or with your employer. (Vondenland 2010: 2; our translation from Dutch)

New speakers' perspectives on their new language learnt through new means for language learning
The class under study at the educational centre has 17 students preparing for the integration exam. All of them follow the Integration and Pedagogical Support stream. They are all women aged between 21 and 56 who have either a low or middle educational background, except for two of them who have followed higher secondary education. The first group to be interviewed included Muhlise from Turkey, Tamusa from Russia and Arzoo from Afghanistan. Muhlise has been in the Netherlands for the past 9 years, she married a Turkish man who already lived and worked in the Netherlands. Tamusa, instead, came to the Netherlands in 1999 because of her husband's job. In Russia she was a librarian in a local primary school. Arzoo arrived as a refugee from Afghanistan in 2002 and her integration trajectory, following the regulations spelled out in the Netherlands for integrating newly arrived migrants had to be completed in November 2010. None of them were working, though the local municipality required them to carry out voluntary work because in such a way they could practice their Dutch. Muhlise's experience with ICTs is limited to the classroom environment. At home, in fact, she had no PC; Arzoo and Tamusa did have a PC at home that they use for keeping in touch with their home countries but also for learning Dutch. Both Mulhise and Tamusa agree that ICTs make it difficult to learn Dutch because of the feedback that one gets: Yeah, yeah sir, I don't understand, you get wrong, wrong, wrong, good but then why wrong? It is not the first language I learn but here I don't understand (Vondenland fieldnotes 2010: 97; our translation from Dutch) While 'The Virtual Neighborhood' helps them in encountering new situations that reflect real life settings, they are also confronted with specific bits of the Dutch language, such as the jargon of insurance companies that was often mentioned in the focused group discussion.
When confronted with such vocabularies, these students could not find their way in the ICT program from which they get only straightforward mechanical feedback of right or wrong. The second focused group discussion involved Amar and Sera, both refugees from Iraq who came to the Netherlands 11 and 13 years ago respectively and Rana, who is originally from Afghanistan and who had come to the Netherlands 12 years ago. All of them are low educated. Sera reported to look forward to her exam and that she is currently engaged in voluntary work because she wants to be in touch 'with the Dutch'. They all have a computer and a laptop at home, which they use regularly: Amar: I use it for all sorts of things, also for Dutch Rana: Yeah, yesterday, yesterday, you know, yesterday I used it for the exercises of the integration programme, but I love talking and with 'The Virtual Neighborhood' you cannot talk Amar: Yeah Rana: And also the accent counts, we cannot practice accent and then when we talk, we are foreigners, it is a matter of accent, they understand a lot but sometimes when I talk they look at you, like I don't understand.' In terms of the ICT application, the conversation unfolded as follows Sera: I search for a job, I am almost done with finding one and this helps me for a job for real life, I am almost done, it is important for me to know words, for my children, for my children at school Max: And how about the exercise that you have done yesterday and today, the exercise of going to the library? Rana: It was easy, easy, every day you use these words outside Amar: With the mouse I have not learned it here, I have learned it before, before I did not know anything, two three times and then it is ok Sera: I knew it already, you don't need school for using a mouse, though some people here did learn it here (Vondenland fieldnotes 2010: 97; our translation from Dutch) On the one hand, Rana is not satisfied because of the lack of speaking skills involved in this ICT application. Other learners, although they see the benefits of being exposed to real life situations, also call upon a lack of any possibility to practice their accent, a component of their Dutch language skills that still makes them sound foreign. Last, the ICT application confronts the learners with genres and vocabulary that can be an obstacle. This lack of experience with a certain register, though, is more a development stage that learners may encounter with their Dutch language than a limitation of the ICT application. Native Dutch speakers too are not acquainted with every register of Dutch, as there is no native speaker who knows his/her language as a whole and who knows his/her language perfectly well. This lack of register proficiency becomes even more problematic when there is a lack of background vocabulary knowledge, as the learner does not understand the dialogue assigned to her. Another obstacle learners report in working with 'The Virtual Neighborhood' is that there are no wrong answers.
Last, we have Filiz, and Akanan. Filiz is of Moroccan origin and a speaker of Moroccan Arabic as well as of German as she has lived in Germany since she was 9 years old. Akanan is from Turkey. She has left the country to marry a Turkish man who has lived in the Netherlands for 23 years. Filiz reports to understand pretty much everything that people say to her in Dutch through her knowledge of German, but her speaking and writing skills are, in her view, both still rather poor. Furthermore, Filiz, at times, does not know why certain things happen in Dutch and that is when the use of ICT, such as online dictionaries and grammar buddies from Dutch to German, come in handy. When addressing 'The Virtual Neighborhood', Filiz reports that the level of understanding that is required from this ICT application is fairly low. However:

Filiz:
This was easy, easy, the library is easy, the school is easy, I know it but the grammar is very difficult Akanan: Yes with the children.

Summary of case 1
The data collected during these focused group discussions report a paradox in that 'The Virtual Neighborhood', an application that matches the top-down policy aiming at the development of language skills in real-life situations, in fact, presents new speakers with a double set of obstacles. First, it turned out that it was not always clear to the new speakers what they should do with language in the ICT based 'real life' situations, a register bound language learning, and whether what they do with a given register is right. Further, an obstacle to obtain optimal results from this ICT application could be found in the lack of techno-literacy that these new speakers experience when dealing with ICT-lead language learning and more specifically with grammar, something that they are used to learn from books. These students in fact complained having to respond to the prompts given by the application. They felt that for learning a foreign language, one needs books, rules and a good dose of grammar while they missed out on the development of communication skills and in training their pronunciation with their teacher. The result they reported was that they ended up memorizing the right answers without knowing whether and/or why these answers were either well or ill-fitting the communicative exchange at hand.

New speakers of an old language: observations from Chinese heritage language classrooms in the Netherlands
The Chinese belong to the oldest established immigrant communities in the Netherlands. July 2011 witnessed the 100 years anniversary of Chinese presence in the Netherlands, which was celebrated with a Lion Dance featuring a crew of 100 Chinese lions performing at the Dam Square in Amsterdam (Li 2016: 19). Figures of the number of Chinese living in the Netherlands range from 84,310 in 2015 (according to the Central Bureau of Statistics' definition in terms of persons' or their parents' nationality and country of birth) and 150,000 (according to the more inclusive estimate by the Chinese embassy in the Netherlands). In the pre-1990s the majority of Chinese immigrants came from Hong Kong, Zheijang and Guandong and were speakers of Cantonese. After 2000, however the number of people coming from Zheijang and other provinces in mainland China has increased to over 50%, most of them being speakers of .
The Chinese community in Eindhoven provides an excellent reflection of the above sketched demographic development and its consequences for people's language and identity repertoires. This can be perfectly observed in the Chinese school in Eindhoven, which was established in 1978 by the Chinese community to provide Cantonese lessons to children of Cantonese origin. The school is a complementary or heritage school that is community-run and self-financed. In 2012 it had over 300 children from Kindergarten to grade 12 who on Saturdays from 9.15 to 11.45 mainly follow lessons in .
In what follows we will discuss two key incidents (Kroon and Sturm 2007) that illustrate the position of Dutch-Chinese new speakers in the polycentric Chinese classroom in Eindhoven run by a Chinese teacher whom we call here Mrs Sun.

Chinese classrooms
The first episode in Mrs. Sun's class that we deal with here is a teacher-led discussion of a curriculum text with eight (university and higher secondary education) students, aged 17-20. Mrs. Sun's class is very heterogeneous: three students are of Hong Kong Cantonese background; two are of Wenzhounese background and one each is of Fuzhounese and Mandarin background. The classroom language is exclusively Mandarin or Putonghua. All students have native-like proficiency in Dutch.
The text that was discussed, 'The song of a little brook', is a well-known Chinese folk story, written in 1959 during China's Big Leap Forward campaign, striving to transform China into a modern communist society trough industrialization and collectivization. The story aims at producing and instilling traditional values, collectivity and community in the students. The little brook, running through the landscape to the sea, turning into a stream and ultimately a big river is a metaphor for the socialist revolution and construction of China, praising hard work and collective achievement. And that exactly is what Mrs. Sun wants her class to get out of the text. Her opening question how they feel about the text, however does not lead to much interest. 'I don't have any feelings about this text' (我没有什么没感 受), Mandarin background student Tao says, and 'Nothing' (没意思) Whenzounese Hong says. After the teacher's explanation of the metaphorical meaning of the story, the students start giving their own interpretations of the text. Tao for example, reacts to Mrs. Sun's statement that 'the Netherlands absolutely makes people lazy, makes people making no efforts' (我觉得荷 兰太让人不努力了) by stating 'Dutch people are more efficient than Chinese' (荷兰人比中 国人efficiënt. 中国人是没办法) culminating in his claim that 'my way of thinking is Dutch' (我的想法是荷兰人的 想法). In that cultural framework there is clearly no room for the message of a traditional Chinese folk story. As Li (2016: 56) concludes: "While the teacher seemed to believe that teaching 'language' and 'culture' through folk stories was a means of reproducing 'Chinese' identity in the young people's minds, the imposition of such Chineseness was explicitly challenged and renegotiated in the classroom. The students assertively considered themselves Dutch citizens fully participating and entrenched in Dutch culture and society, and rejected the deeper metaphorical meaning and moral lesson embedded in the story." Another example of such competing perspectives between a Chinese teacher and Dutch-Chinese youngsters is nicely illustrated in a teacher-student classroom interaction on tone in Chinese in Mr. Zhou's class attended by four students (age 16-19), two of Hong Kong Cantonese, one of Wenzhounese and one of mixed Guandong and Hong Kong background. Mr. Zhou, an earlier migrant from Guandong province who is a speaker of Cantonese asks who can make a sentence or give the meaning in Dutch of 本质 (bĕn zhí, 'property', pronounced with rising tone). Guandong-Honkongese Tongtong gives the correct answer but also asks the teacher: 不是本质 吗 ('Shouldn't it be bĕn zhì', with falling tone). The teacher looks at the book again and agrees: 本质啊, 应该读第四声啊, 对不起 ('Bĕn zhì, ah, should be pronounced with the fourth tone ah, sorry'). The same story is repeated with the wrong pronunciation of 比较 (bǐ jiăo instead of bǐ jiào) again corrected by TongTong and other students and agreed upon by the teacher. Li (2016: 42) concludes: "This classroom episode presents a serious deviation from the traditional Chinese language class where the teacher has all the 'knowledge' (…) The language teacher's pronunciation is 'corrected' by his students." This documents the change in the Chinese school from one language regime (Cantonese) to another (Mandarin/Putonghua) and Dutch Chinese students as the agents of this change, reflecting in the diaspora a geopolitical repositioning of the Peoples Republic as a new economic world power.
In conclusion, these two examples centered around being a Dutch-Chinese student in dealing with Chinese make clear that the classroom is a polycentric environment in which the Chinese school provides the language (Mandarin/Putonghua) and the values from the PRC. The students however only take the language (as a commodity) and at the same time challenge the teacher's authority regarding Chinese language and values. As such the school becomes a site of conflict, contestation and negotiation over Chineseness from a Dutch perspective by students who are fully competent in Dutch and who are at the same time competent Dutch new speakers of Chinese.

Chinese online
In addition to the above offline classroom examples we will now deal with Chinese youth's metalinguistic interaction online, more specifically focusing on the Dutchmedium teenage Asian and Proud community on the Dutch social network site Hyves (http://asian -and-proud .hyves .nl) and the adolescent platform Jonc-Connecting Asians (http://www.jonc.nl). The Asian and proud community was established in 2007 by a Cantonese-speaking Chinese-Dutch girl who studied Chinese (Mandarin) in complementary education in Amsterdam. It was closed down in 2013 after Facebook had taken over Hyves. Here we will have a look at one of the discussion forums on the Asian and Proud Hyve, as it was introduced April 20, 2008 at 21:37 by Leon and asked: Welk Chinees dialect spreken jullie? (What Chinese dialect do you speak). Leon introduces himself as follows (originally in Dutch): The answers that were given to Leon's question can be read as an archive of self-articulations of Dutch-Chinese identity and language. The language varieties named by 89 persons in 95 posts included as the most frequently mentioned: Chinese (41x), Mandarin (63x), Cantonese (53x) and Wenzhounese (21x). In addition also Vietnamese (8x), Hakka (4x), and Fuzhounese, Hokkien, Teochew, Qingtianese, Shanghainese and Suzhounese (each 1x) were mentioned. This is not the place to elaborate on the participants' posts on their positive as well as negative experiences and attitudes regarding (being forced by their parents to learn) Chinese and the (limited) level of proficiency they generally reported to reach. The main point that we want to make here is that knowing Chinese in the Netherlands can mean a plurality of things. First, it means knowing Dutch, especially online youth variety of Dutch language, characterized by free use of punctuation marks, emoticons, word contractions and abbreviations, extensive borrowing from English, deliberate misspellings, use of vulgar or foul language, onomatopoeic exclamations, colloquialisms, duplication of letters to express intensity and occasional switches to Chinese (Li 2016: 97-98). This all, except for the switches to Chinese, makes their use of Dutch language undistinguishable from indigenous native speakers of the variety of Dutch associated with youth online interacting with one another. As Li concludes: "In multilingual Europe 2.0, ethnicity (being Chinese) or the mother's tongue (Chinese, Wenzhounese) is not a valid criterion for native-speakerness anymore (…) we are dealing with new natives and new European identities, with Chinese-Dutch and many other hyphenated, polycentric identities" (p. 105). The new speaker is no simple aggregate of one new language added to an old one, it is a complex and multilayered ensemble of linguistic resources that manifests itself in multifaceted languaging practices.
On jonc.nl, the website of one of the two Chinese youth and young adults organizations in the Netherlands, our focus is on a Forum discussion initiated by Faraway from Eindhoven, asking on August 6, 2007 (originally in Dutch): "How often have you been to China and do you speak the language well?" The topic ran for 48 h and 46 min and had 11 posts by 9 contributors. As far as travel is concerned it turns out that most contributors travelled (or want to travel) to China for touristic reasons (not unlike their Dutch peers). The languages mentioned include Mandarin, Wenzhounese, Cantonese, Weitouhua and Hakka and the contributors indicate that their proficiency in one of these Chineses is not perfect or what they would like it to be. "For the young adults on jonc.nl and the teenagers on the Asian and Proud community," Li (2016: 114) concludes, "being Chinese invariably means knowing Dutch and sometimes means knowing very little Chinese or no Chinese at all (…)."

Summary of case 2
A main conclusion of the Eindhoven Chinese school case study is that these Chinese youngsters, although from the perspective of new speakerism they would be considered new speakers of Dutch, are in fact (new) near native speakers of Dutch that embrace Dutch identity and values. At the same time, although (a variety of) Chinese is their heritage language, they are new speakers of Chinese, i.e. Putonghua but without following traditional Chinese values. They are in fact not really speakers of their heritage language(s), but are becoming more and more internationally oriented as is shown by them opting for Dutch and even mainly English on online forums. It can therefore be concluded that the languages and identities displayed by Chinese youngsters in the Netherlands are not categorical but situated, i.e. performed and negotiated from moment to moment, from space to space: in the Chinese classroom seeking voice in opposition with the teachers and online seeking voice in e-coming together and sharing experiences with other international Asians. As such they constitute a complex type of new speaker, i.e. combining being a (new) near native speaker of Dutch, a new speaker of Chinese (Mandarin/Putonghua) and a new speaker of English. The levels of proficiency in these languages may vary: as second or third generation immigrants their Dutch proficiency might be (near) native, their proficiency in Mandarin/Putonghua might not be perfect as they themselves stipulate-but it's notably better than that of their Cantonese background teachers-and their English proficiency certainly meets international online youth standards. What they however miss is "old", i.e. Chinese heritage language proficiency, which is becoming obsolete not only in the Netherlands and in the international online community they engage with but also in contemporary mainland China where they would wish to go for a holiday or family visit.
From a language policy perspective we see that the students' bottom-up language practices and metalinguistic perspectives in the Chinese school and on the internet challenge the top-down family language policies of their parents who after all wanted them to learn "Chinese" as well as the Chinese school's focus on combining learning Chinese with learning Chinese values. These speakers of Dutch and Chinese in the Netherlands seek and manage to acquire their own voice offline as well as online. In doing so they represent global youth in a superdiverse society that eclectically opts for the language and identity resources that come in handy for achieving their communicative goals irrespective of the policies set for them.

The new speaker and the language policy abyss
The two cases we have reported here show that ethnography has the potential of opening up the process of language policy making while tying it down to those actors that are initiators of sociolinguistic change. More concretely put, as already pointed out by Hornberger and McCarty (2012), ethnography shows that large policy structures and globalizing forces are reconfigured in local language ideologies and practices or-as they theorize-in "globalization from the bottom up". What these examples show is that the language users at hand do not seem to do things with language following the exact ways that official language policies would want or expect them to. It is exactly there that a starting point can be found for transforming language policy making from a top-down to a bottom-up endeavor. All over the world language academies like the Nederlandse Taalunie in the Netherlands, Flanders and Suriname, the Accademia della Crusca in Italy and the Académie Française in France-all members of the European Federation of National Institutions for Language (EFNIL)-deal with rules and regulations about what is allowed and what is not allowed in using, mainly writing, a specific language. At the same time, however, language users engage with language in polylingual ways (Jørgensen et al. 2016), thereby using, combining and re-combining the linguistic resources that they have available from their sociolinguistic repertoires so to reach the end goal of having their voice heard when engaged in an interaction. These repertoires are in turn used-in many cases successfully-to make meaning in communities of practice and knowledge that are no longer stuck to the concept of a speech community with a fixed and prescribed linguistic form and format. Such 'poly-languaging practices' are far away from the exclusive and normative perspective that traditionally seems to be embraced by language policies, as well as by language policy agents and political discourses regulating language acquisition, for example in integration contexts. As such they can be considered local, bottom-up language policies that can provide empirically grounded inspiration for institutional language policy development.
The two cases we have dealt with in this contribution lead us to the following conclusions. The always existing distance between top-down policies and bottomup practices in the field of language seems to have developed into an abyss where 'new speakers' of both new and old languages learning language through new and old means for language learning are in jeopardy of sinking into. The actual language practices and the underlying attitudes that can be observed in times of globalization led mobility are no longer a valid reflection of institutionally supported language policies, understood as "an attempt by someone to modify the linguistic behaviour of some community for some reason" as Kaplan and Baldauf (1997: 3) wrote somewhat more than 20 years ago. Time and space compression, together with the top-down-bottom-up divide as well as their potentially negative consequences for the efficiency and effectiveness of concrete policies may apply to all policy making irrespective of the policy object under consideration, whether this is language or the means through which a given language should be learnt for a given purpose. In the domain of language policy, however, there is another complicating and potentially damaging factor, i.e., the very concept of 'language' that is (still) in general use here as is reflected in for example many older and newer handbooks (cf . Cooper 1989;Johnson 2013). In order for a language policy to be successful in catering for the needs of new speakers, it has to reorient itself to a definition of language that includes people's linguistic resources, registers and repertoires and their uses in the everyday complex though "ordered" reality of the actual language practices it should serve.
The domain of actual language practices, in other words, can be considered the meeting place of formal policies and norms from above and informal practices and emergent norms from below. For exactly that reason, following Ricento and Hornberger (1996: 419), we plead for foregrounding the agency of language users in deciding on language policies. Furthermore, given the cut that we embrace here, we are compelled to reflect on the power relations and power results that have shaped the present ways of conceptualizing language. We ought, in fact, to pay attention to emergent ways of conceptualizing language that people use on the ground and, in doing so, we hope to have showed that those emergent more inclusive ways in which language tends to be used, bear in mind the purpose of 'doing togetherness' and conviviality rather than for evoking and defending even further the construction of difference and inequality. In so doing the concept of new speaker manages, in fact, to avoid falling into the incrementality that characterizes, as Makoni (2012) denounces, the 'barren verbiage' surrounding the contemporary sociolinguistic debate about language and society.