Introducing a Polynomic Approach in Ryukyuan Language Learning

Abstract

The polynomic model as a model of language codification treats regional and social variation within language as inherently good and does not hierarchise this variation. This research is an argument in advocating the application of a polynomic model in Ryukyuan language learning, focussing on Okinawa Ryukyuan. The Ryukyuan languages consist of a minimum of 5 abstand languages that are spoken in the Ryukyuan island chain in southern Japan. The Ryukyuan languages have no standard languages and show internal variation. A monolingual language ideology implemented since the annexation of the Ryukyus by Japan in the late 19th century has caused a language shift towards Japanese, rendering all Ryukyuan languages endangered. A small revitalisation of the Ryukyuan language has been taking place since the early 21st century, necessitating the development of an infrastructure for language learning. In this research, I give reasons why the application of a polynomic model would allow for an inclusive language revitalisation that respects local language practices and identities in the Ryukyus, without reproducing the colonialist attitudes towards language that led to language endangerment in the first place.

1. Introduction

The Ryukyuan languages are a complex of endangered languages autochthonously spoken by the inhabitants of the Ryukyuan Island chain between mainland Japan and Taiwan. There are several organisations and initiatives that strive to revitalize the Ryukyuan languages. The Ryukyuan languages display strong regional variation. A minimum of five different abstand languages (Kloss 1967) can be discerned (Pellard 2015), and no standard languages have developed.

The regional variation that exists within the Ryukyuan languages has given rise to prestige differences between varieties. Helping (new) speakers of negatively characterised varieties to gain a sense of empowerment and an appreciation for what is their heritage, and to redefine themselves on their own terms, has emerged as a goal of decolonization and the Ryukyuan language revitalisation effort, but achieving this is difficult.

In view of this situation, this study argues for the application of a polynomic model when teaching spoken Ryukyuan to adults. Regional variety and the absence of standard languages means that speakers who are willing to teach do not necessarily speak the variety that learners want to learn. An application of a polynomic model would facilitate the acquisition by the learner of a variety other than that of the instructor and enable the learner to understand other varieties than their own target variety. Moreover, the application of the polynomic model would contribute to the decolonisation of the discourse around Ryukyuan language revitalisation, because it does not project the Japanese standard language ideology upon the Ryukyuan languages.

I shall focus on Okinawa Ryukyuan, or Uchinaaguchi, the largest Ryukyuan language in terms of speakers and area. I provide a description of variation within the Okinawan language, and a description of the polynomic model of language. After that, I shall give an overview of the factors that necessitate the inclusion of regional variation in Okinawan language learning, along with a short section on how to touch upon variation in Okinawan.

2. Variation in the Okinawan Language

2.1. Sociolinguistics of Okinawan

Okinawan does not have a written standard language, but there is an archaic, supraregional poetic register that is used in folk songs and poetry. This register shows hardly any regional variation but is largely based on South-Central Okinawan. However, this poetic register is not used as a spoken language by traditional speakers (Nishioka 2017). Spoken Okinawan, as well as the other Ryukyuan languages, are regionally and socially diverse, and they have been so since well before the spread of the Japanese language after the annexation of the Ryukyus by Japan was formalised in 1879. Domain loss to Japanese, and the advent of modern modes of transport and communication, have given rise to intergenerational diversification in Ryukyuan as well. The Ryukyuan languages of younger speakers tends to be more influenced by Japanese, and it has in some cases become regionally koinenised. Regional koinenisation tends to be more pronounced in regions where the shift towards monolingualism is less advanced.

Due to the interruption in intergenerational transmission, Okinawan is classified as ‘definitely endangered’ by UNESCO (2009). Practically all speakers of Okinawan are bilingual in Japanese and Okinawan. Anderson (2009) categorises speakers into four generational subgroups based on their Okinawan language proficiency. Traditional speakers who learned Okinawan at home or in daily life during childhood can be divided into (1) ‘Full speakers’ (born prior to the mid-1930s), and (2) ‘Rusty speakers’ (born between the mid-1930s and the mid-1950s). The name ‘rusty speaker’ is infelicitous for the Okinawan case, for it includes fluent speakers who use Okinawan on a daily basis but have not acquired the full range of vocabulary because of domain loss. People with no or a very limited active command of Okinawan can be divided into (3) Semi-speakers (born between the mid-1950s and the mid-1980s), who have a passive command of Okinawan, and (4) Non-speakers (born between the mid-1980s and the present), who can neither understand nor speak Okinawan. Note that Anderson’s categorisation is based on speakers in the Shuri-Naha area, the political and economic centre of Okinawa prefecture. There are indications that Okinawan language proficiency is higher outside of the Shuri-Naha area, and the subgroups proposed by Anderson tend to be a minimum of ten years younger than in Andersons original report. For instance, on Kume Island, speakers born in the mid-1940s can still be regarded as full-speakers, and the youngest cohort of active Okinawan speakers were born in the late 1960s (Van der Lubbe et al. 2021).

Heinrich (2007) is one of the very few pieces of research that investigates the domains of use of Okinawan without being clear about which region of Okinawa his results represent. His results indicate that the family (‘addressing grandparents’) is the domain where most Okinawan use takes place, more than the domains ‘neighbours’ and ‘colleagues’.

As mentioned above, spoken Okinawan has significant regional and social variation. In the minds of many, there is a hierarchy between these varieties. Shuri Okinawan, or Sui-kutuba in Okinawan, carries a certain prestige due to Shuri’s former position as the seat of government of the Ryukyu Kingdom. The features that set Sui-kutuba apart from other Okinawan varieties are mostly found in the lexicon pertaining to Shuri’s court culture (Lawrence 2015), and the fact that there are class differences in pronunciation and vocabulary (Ishihara et al. 2019; Kokuritsu Kogugo Kenkyūjo 1963, 1982, 1985, 1987).

Certain historical developments in the phonology of Okinawan spoken in the Shuri-Naha area (/p/ > /h/: pana > hana ‘nose’. /k/ in front of /i/ and /j/ > /tʃ/: sabaki > sabatʃi ‘comb’) have spread throughout a large part of the South-Central area of Okinawa, and some of the adjacent islands. Although there still exist small local differences, Uemura (2003) points out that a form of Okinawan that ‘cannot be identified with any particular village’ has developed. I shall refer to this form of Okinawan as ‘Common Okinawan’. There are indicators that before the spread of Japanese, Common Okinawan had (limited) currency as an interregional lingua franca. Common Okinawan has been influencing, or even replacing, local Okinawan varieties (Ishihara et al. 2019). This has led to the rise of local koinenised Okinawan varieties. More traditional speakers, or even the younger speakers themselves tend to characterise these varieties in negative terms, as ‘not the language of the settlement’, or even as ‘incorrect’.

There are indicators that the general attitude towards Okinawan and other Ryukyuan languages has improved in the 21st century (Ishihara 2014); however, many speakers of Okinawan still characterise their native language as ‘uncouth’ or ‘rustic’. This is especially the case with speakers of varieties outside of the Shuri-Naha area. In some localities, this negative perception of the local variety vis-à-vis Sui-kutuba has even caused a faster language shift to Japanese (Osumi 2001).

The improved attitude towards Okinawan has led to a small revitalisation effort, and there are now language courses organised by universities, municipal institutions, and through private initiatives (Ishihara et al. 2019). New speakers have appeared but are still a marginal phenomenon (for a profile of new speakers, refer to Zlazli 2021).

2.2. Regional Variation within Okinawan

The Okinawan dialect cluster can be divided into a Northern and South-Central group based on Pellard (2015) and Lawrence (2006). Although there are certain phonological reflexes that are more widely attested in the northern dialects, it is difficult to draw a clear border dividing North and South-Central Okinawan. Note that UNESCO regards the varieties spoken in Northern Okinawa as part of a distinct language called ‘Kunigami’. I shall focus on the South-Central variety of the Okinawan language in this research.

The nine varieties in

Table 1. ‘I am going to the fields, but where are you going?’ in IPA in different Okinawan varieties. The horizontal line marks what can be considered the border between Northern and South-Central varieties.

North	Oku	waɴ=ja	pʰaru=ɴkai	iku-ʃiga	ura=ja	daː=ɴkai	ikuɴ-ga
	Jana	waɴ=ja	pʰaru=tʃi	itʃuː-ʃiga	ʔjaː=ja	daː=tʃi	itʃu-ga
	Sokei	wanu=ja	haru=gatʃi	itu-mu	jaː=ja	da=ttʃi	itu-ga
	Onna	waɴ=ja	paru=ɴkeː	iki-gwa	jaː=ja	maː=ɴkeː	iki-geː
South-Central	Sobe	waɴneː	haru=ɴkeː	iku-ʃiga	ʔjaː=ja	maː=ɴkeː	iku-ga
	Yakena	wanoː	haru=ɴkai	itsu-iga	jaːmeː	maː=ɴkai	itsu-kutu
	Shuri	waɴneː	haru=ɴkai	itʃu-ʃiga	ʔjaː=ja	maː=ɴkai	itʃu-ga
	Itoman	wanoː	haru=ɴkai	iku-ʃiga	jaː=ja	maː=ɴkai	iku-ga
	Janadō	wanoː	haru=katʃi	itsu-higa	jaruː=ja	maː=katʃi	itsu-ga
		1SG:TOP	field=ALL	go-but	2SG=TOP	where=ALL	go-WQ 1

¹ Glosses: 1SG = first-person singular pronoun, TOP = topic marker, ALL = allative case marker, 2SG = second-person singular pronoun, WQ = wh-question marker.

nicely illustrate the parameters of regional variation within Okinawan. The following is the sentence ‘I am going to the fields, but where are you going?’ rendered in nine local varieties.

The variation we see in Table 1 is mostly phonological. For instance, the word ‘field’ appears as pʰaru, and haru, and the verb ‘go’ with the suffix meaning ‘but/however’ appears as iku-ʃiga, itsu-iga, itʃu-ʃiga, iku-ʃiga, itsu-higa, iki-gwa, and itu-mu. We can observe phonological as well as lexical variation in the allative particle; most regions use a variant of ɴkai (with a fused vowel: ɴkeː), whereas Janadō uses katʃi, Sokei gatʃi or ttʃi, and Jana tʃi. A difference in morphology can be observed in the topicalized form of the first-person pronoun singular; it appears as either a fused form waɴneː, or wanoː, or a form unfused form waɴ=ja, or wanu=ja ‘as for me’.

As of 2022, no research exists on the mutual intelligibility of Okinawan varieties. There are some indications that speakers of Northern Okinawan varieties understand South-Central varieties better than the other way around (Heinrich and Fija 2007). This is also true in the experience of the author, with the addition that it depends largely on how much a speaker is exposed to other varieties.

3. The Polynomic Model

The concept of a ‘polynomic language’ (langue polynomique) stems from Corsican linguistics (Blanchet 2020). Corsican linguist Jean-Baptiste Marcellesi described it as follows:

A polynomic language is a language whose unity is abstract, and which is recognised by users as existing in several forms, each tolerated equally without hierarchical or functional distinctions. It is characterised by mutual acceptance of phonological and morphological variation by users of different varieties; likewise, lexical variety is seen as a source of richness (Marcellesi 1986; as cited in Sallabank 2010).

Succinctly, Sallabank (2010, p. 311) describes the polynomic model as ‘a pluralistic model of language without a single prestige variety or functional distinctions’.

According to Jaffe (2020), polynomie as a language ideology ‘crowd-sources’ perspectives on what is to be considered as one language. Polynomie calls upon speakers of a language to acknowledge existing regional and social varieties of that language. It requires speakers to commit to a critical engagement of the concept of language (Blackwood 2011). This translates into language practice by speakers of different varieties of Corsican accepting each other’s varieties as Corsican and using Corsican amongst each other regardless of regional differences.

We find places in the Ryukyus where a polynomic practice can already be observed today. Speakers of Okinoerabu Ryukyuan use their language regardless of dialect differences (fieldnotes August 2021), and the same seems to be the case with Miyako Ryukyuan notwithstanding considerable differences between the varieties used (Thomas Pellard, personal communication March 2013).

There are places in Okinawa where the use of the local Okinawan varieties for communication between inhabitants of different settlements is all but eradicated. This is the case in most of Northern Okinawa and the area around the Katsuren peninsula in Central Okinawa (fieldnotes for Oku, Kunigami Village 2012–2015, fieldnotes for Katsuren Heshikiya, Uruma City 2016–2017). The author is frequently confronted with claims that it is not possible to have a conversation with someone from another settlement using Uchinaaguchi because of dialect differences. Upon further inquiry, speakers always admit that feelings of shame or ridicule towards one’s own or the others variety, rather than mutual unintelligibility, are the actual reason why Japanese is favoured (field recordings Oku, Kunigami Village 2013). It is likely that in pre-colonial times such dialect differences where not seen as an impediment to successful communication.

The polynomic model is an attempt to raise the status of a minoritized language, while maintaining tolerance of its variation. As such, it is in opposition with a hierarchical model where one standard language is promoted as the ‘correct’ way to speak and write, degrading in so doing all the other regional and social varieties of that language (Milroy and Milroy 2012). Speakers of minoritized languages as well as advocates for language revitalisation tend to be only familiar with the hierarchical model, as it tends to be the language ideology associated with the language that is dominant (=the replacing language in case of language shift) in their societies (Murchadha 2015). This is especially the case in Japan, where the standard language ideology is strong. This ideology surfaces in the Okinawan context in two ways: (1) Okinawan is a group of local dialects of Japanese, or (2) Okinawan is a language separate from Japanese and is in need of standardisation for functional efficiency. Needless to say, the latter view tends to be more prevalent amongst Okinawan revitalisationists.

A hierarchical notion of language is also evident in public discourse on the revitalization of the Ryukyuan languages, where the fact that every settlement has their own distinct Ryukyuan language variety is often pointed out as a problem in the development of learning materials (Heinrich and Ishihara 2017). The polynomic model could be key in challenging these notions. Challenging these notions is a crucial part of decolonisation and can be viewed as an exercise in ‘applied disinvention’ (Makoni and Pennycook 2006) where the heteroglossic nature of Okinawan language practices in terms of social and regional variation is accommodated, and recognised as normal and even desirable, as well as represented in learning materials. The polynomic model could be key in this disinvention in that it does not uncritically reproduce majority ideologies, regimentations and uses of language which stand in the way of reclaiming local and social varieties.

Given the sociolinguistic reality of rich local variation, which constitutes the linguistic and cultural heritage of Okinawans, we come to understand that any hierarchical model of language will produce a range of problems that are remindful of colonial language settings. Let us therefore turn to current second language learning to get a more comprehensive idea about the nature and the scope of the problem.

4. Applying a Polynomic Model on Okinawan Language Learning: Why and How

Standardisation as a method of language codification is an inherently colonial concept in the Ryukyuan context, because the concept of one spoken standard language was only introduced in the Ryukyus through the annexation and the consequent assimilation to Japanese language and culture (Heinrich 2012). A polynomic model offers a way around the concept of a monolithic standard. As we will see below, there are many additional benefits of applying a polynomic model where regional variation is touched upon during the learning process in the case of Okinawan (and other Ryukyuan languages).

Okinawan is the only Ryukyuan language for which already several textbooks exist that offer a step-by-step roadmap for language learning; however, all of these focus on Shuri Okinawan and do not touch upon regional differences (for an overview, see Ishihara et al. 2019).

If language varieties are carriers of cultural knowledge and function as economic and aesthetic resources, then one cannot have enough variations. The polynomic model applied to language learning can help learners understand variation in Okinawan as a natural phenomenon as well as individually and socially enriching and facilitate the development of learning methods that allows learners to study together amongst like-minded people, even if their target varieties differ.

Descriptive linguists, like Tohyama (2019) and Sakihara (2015), call for the construction of learning materials and/or the establishment of language classes for every subvariety of Ryukyuan. It is questionable whether this is feasible logistically, for there are over 800 traditional settlements in the Ryukyus that can all be considered to have their own linguistic variety (Uemura 2002). Specific skills are needed for the construction of language learning materials. Few people possess these skills, and even fewer are willing to pour their time into a very local and endangered language variety.

Moreover, it is questionable whether there are enough language consultants for every one of the traditional settlements who can assist in the construction of language materials. It is already becoming increasingly difficult to find traditional speaker consultants for some varieties, and many traditional speakers are reluctant to teach their language.

The construction of learning materials based on a polynomic model, touching upon the common characteristics of Okinawan as a language as well as the parameters of variation would be a logistically feasible alternative. This would also facilitate the acquisition of local varieties, especially when combined with a master–apprentice programme (Hinton et al. 2018), as is currently being set up for Yaeyama Ryukyuan (Topping, forthcoming). In this way, learners can enjoy the benefits of contact with other learners who are going through the same experience, as well as having the benefit of being instructed by a teacher.

Traditional speakers of Okinawan tend to be monodialectal. If the goal of the Okinawan language learner is to acquire the ability to communicate in a meaningful and appropriate way with traditional speakers in Okinawan, it is imperative that they learn about regional differences in the language in order to communicate in Okinawan with these monodialectal speakers without having to resort to Japanese too much (Van der Lubbe et al. 2021; Van der Lubbe 2021a). The small Ryukyuan language revival effort has produced a few competent new speakers. However, these new speakers are still rare. Traditional speakers of Okinawan tend to regard new variation in the form of newspeaker variation with scepticism.

The legacy of linguistic colonization, and the language ideologies that accompanied it, manifests in two ways. Both amount to obstacles in language revitalization, but these can be addressed by applying a polynomic model. Firstly, Hammine (2020) identifies local language ability as being associated with a lack of formal education as one reason why many traditional speakers of Miyara Yaeyama Ryukyuan lack the confidence to teach their language. The author has also observed the same internalised oppression in Okinawan speakers, as well as well as a form of linguistic traditionalism similar to what is described by Sallabank (2017) for Guernsey. Secondly, the Okinawan of younger speakers is often characterised as ‘broken’ or ‘incorrect’ by both older and younger speakers alike. The Okinawan of older speakers, or even speakers of olden days is presented as the standard of what is ‘authentic’ and ‘uncorrupted’ Okinawan, leading to unattainable standards of how much Okinawan one should know to be able to teach the language. Younger speakers who do show a willingness to teach Okinawan often feel the need to justify this lack of reluctance by claiming that they spent a lot of time with old people when they were young, or that they were raised by their grandparents who did not know Japanese. New speakers’ Okinawan is often not accepted, because it is perceived as ‘funny’, ‘incorrect’, ‘awkward’, or ‘forced’. These traditionalist views further complicate revitalisation efforts. Applying the polynomic view of language not only on regional variation, but also on generational differences in Okinawan could help revitalisation for it would help ‘legitimise’ younger speakers and new speakers as language teachers.

The benefits of the polynomic model become also evident if we consider it in the light of a so-called manifesto that was developed by the Ryukyuan Heritage Language Society (RHLS). It consists of twelve points that clarify how and why a Ryukyuan–Japanese bilingual society will constitute an improvement to present-day society in the Ryukyus (Heinrich 2014). Note that the points are meant to be applicable to the entire Ryukyus, and Okinawa in particular.

(1)

Transmit and promote a deeper reflection of the Ryukyus in Ryukyuan;

(2)

Restore Ryukyuan self-esteem and confidence;

(3)

Promote Ryukyuan perspectives on language, history, and culture in education;

(4)

Restore cohesion between older and younger generations;

(5)

Familiarise the younger generations with Ryukyuan heritage culture;

(6)

Maintain, strengthen, and apply Ryukyuan cultural heritage;

(7)

Contemporise Ryukyuan language and make it relevant for the future;

(8)

Regain control over Ryukyuan self-image and education;

(9)

Maintain choices for language, identity, and culture;

(10)

Stop conformism in Ryukyuan identities and behaviours to models from the Japanese mainland;

(11)

Contribute to communal happiness and wellbeing;

(12)

Recognise Japan’s cultural diversity, and promote intercultural tolerance.

It is the author’s opinion that the goals of the RHLS as stated in the manifesto can be fully supported through the application of a polynomic model where learners are encouraged to learn about regional differences. The following subsection shall provide a further clarification of several of these points. Refer to Van der Lubbe (2022) for a clarification of all 12 points from the perspective of the polynomic model.

4.1. Why to Apply a Polynomic Model in Okinawan Language Learning

(2)

Restore Ryukyuan self-esteem and confidence.

The suppression, and subsequent minoritisation of Ryukyuan languages has caused many Ryukyuans, including Okinawans, to internalise this oppression. In this colonised setting, use of Ryukyuan languages was stigmatised as bad behaviour and being uneducated. This has caused many Ryukyuans to end up with what Hammine (2020) calls ‘linguistic self-orientalism’. In order to counter this, respecting variation within Okinawan is imperative. Ignoring regional variation for the benefit of promoting and teaching only Shuri Okinawan, or Common Okinawan, disenfranchises speakers of other Okinawan varieties even further (Zlazli 2021). What is more, it reproduces the colonial linguistic hierarchisation that was earlier used to promote Japanese at the cost of Ryukyuan languages. The polynomic view of language treating variation as ‘good’ allows language revitalisists to get around counterproductive hierarchisation.

(3)

Promote Ryukyuan perspectives on language, history, and culture in education.

It has been more than 140 years since the Ryukyus were annexed by Japan. Education has been the main tool by which Ryukyuans were assimilated to become Japanese. Even in the present day, the centralised nature of the Japanese education system does not leave much room for Ryukyuan and Okinawan perspectives on language, history, and culture. Note again that Okinawa is culturally and linguistically diverse, and that there are different perspectives on language, history, and culture depending on region, social class, etc. Even within regions, there is linguistic and cultural diversity that a polynomic approach could help to become recognised and respected.

(4)

Restore cohesion between older and younger generations.

Traditional speakers of Okinawan (and any other Ryukyuan language) tend to be bilingual in Japanese. However, in Okinawan, traditional speakers only speak the variety of their community (Ishihara et al. 2019). Introducing a minority language as if it was one monolithic entity does not help younger, new speakers relate to their elders, as is testified by the undesirable situation that has occurred in Breton; the younger generation is taught a (standardised) variety that traditional speakers fail to recognise as their own language (Hornsby 2005). Research by Hammine also points at some sort of resistance to include new speakers with their new varieties into the traditional speech community. Fostering awareness of, and a positive attitude towards linguistic diversity would help to avoid such situations. The pluralistic nature of the polynomic model makes it ideally suited to addressing this problem.

(7)

Contemporise Ryukyuan language and make it relevant for the future.

Heinrich (2014) calls for corpus planning to develop a ‘socially neutral’ form of Okinawan and maintains that it may be recommendable to draw from the language varieties spoken by the former gentry. It could be argued that Common Okinawan is as close to ‘socially neutral’ as any variety of Okinawan could get. Promoting Common Okinawan as a unitary standard for Okinawan would certainly be easy to formalise, since it (1) already exists, and (2) would allow for learning materials that do not have to touch upon regional differences. However, this would mean imposing a hierarchy on the Okinawan varieties, and this would risk perpetuating the negative characterisation, and internalised oppression of the speakers of ‘peripheral’ Okinawan varieties like Itoman, Yakena, and Sobe. Any form of language revitalization that discourages the use of some varieties favouring what is perceived to be a ‘higher variety’ contributes to this sense of inferiority, and it perpetuates the same attitude that led to language endangerment in the first place. If the aim of language revitalisation is to do away with inferiority complexes, to empower endangered language speakers, and to provide opportunities to appreciate what is one’s own, any form of confirmation of (self-)repressive attitudes is obviously counterproductive.

There are examples that show that maintaining variety and language adaptation to contemporary communicative requirements do not need to be in opposition. Corsican as a polynomic language is one such example (Marcellesi 1986), but there are other languages where diversity is the norm and does not seem to be regarded as an impediment for their continued use. For instance, Norwegian and Swiss German regional varieties have never lost their contemporary relevance as means of communication and expression (Leon 2014; Christen 2008).

Having been involved in teaching Okinawan language to young Okinawans, the author has observed a strong desire amongst new speakers to use Okinawan words where possible without resorting to Japanese. Okinawan is not as functionally developed as Japanese, and there are regional differences in lexicon. In order to meet new speakers’ demands, different Okinawan varieties could borrow from each other and adapt items to their individual phonologies. For instance, the Shuri variety of Okinawan possesses a lexicon for words pertaining to court culture as well as Sino-Japanese borrowings that most other varieties lack (Lawrence 2015). These words can be simply borrowed into other varieties. By the same token, Itoman Okinawan has a rich vocabulary concerning fishery and sea life that can easily be appropriated by speakers of other Okinawan varieties.

(8)

Regain control over Ryukyuan self-image and education.

Assimilation education has promoted Japanese culture and language and has attempted to actively erase Ryukyuan languages until the 1970s. Education has ever since not been engaged in actively addressing the damage that it has done. Assimilation education has made Ryukyuan self-image as something that is only defined in opposition with mainland Japanese culture (Barclay 2006). As such, the general Ryukyuan self-image concerning language and culture is one of low self-esteem, as is testified by Arakaki and Oyakawa (2014). Language revitalisation could help Ryukyuans appreciate their heritage and foster a more positive self-image that is not necessarily defined in contrast with mainland Japan, or with any place other than one’s own. A Ryukyuan self-image based solely on the economic and political centre of Okinawa prefecture, the Shuri-Naha area, would force Okinawans from economically and politically more peripheral areas to define themselves in opposition to the Shuri-Naha area. Okinawan language education by means of one single standard language would risk stigmatising other varieties and their speakers’ self-image as ‘peripheral’. Recognising Okinawan in education as a polynomic language would allow for a more pluralist and inclusive imagining of the self.

(9)

Maintain choices for language, identity, and culture.

Japan’s adoption of an ideology that portrays Japan as a monoethnic, monocultural, and monolingual for the purpose of nation building since the Meiji period (1868–1912), has not left much room for diversity in terms of language, identity, and culture other than just Japanese (Hammine 2019). This point is related to the previous one in that it concerns self-image. Appreciation of one’s own linguistic and cultural heritage would be empowering for all Ryukyuans. A polynomic model of language would allow for a maximum number of choices for language, identity, and culture, for it recognises that there exist several ways of speaking Okinawan, or any other Ryukyuan language.

(10)

Stop conformism in Ryukyuan identities and behaviours to models from the Japanese mainland.

In the assimilation effort, Ryukyuan customs and language were labelled as primitive, and undesirable (Barclay 2006). Ryukyuan culture became associated with, amongst others, tardiness and superstition. Mainland Japan on the other hand became associated with modernity and progress. Becoming Japanese, and leaving all things Ryukyuan behind, became perceived as a tool to achieve ‘success’ in life (see Hammine 2020). Revitalisation of the Ryukyuan languages would be extremely hard to achieve if this conformist attitude towards the perceived sophistication of a place outside of one’s own is not critically assessed, that is, without efforts of decolonization.

In addition to the conformist attitude towards mainland Japan, a linguistic inferiority complex towards the Shuri-Naha region and gentry sociolects can be observed throughout Okinawa. This has proven detrimental to local language varieties like Sokei Okinawan (Van der Lubbe 2020a), and in Ōgimi in the northern part of Okinawa, it has even sped up the language shift towards Japanese (Osumi 2001). The author has observed that often speakers and learners think that their local variety is not ‘real Okinawan’, but rather a ‘corrupted’ form of it. As a consequence, otherwise competent speakers think that they themselves are not qualified to teach Okinawan, as long as they are monodialectal in their own variety, and that speakers from Shuri or with a gentry background are the only ones who are qualified to teach. As for learners, this means that they often feel reservation to ask local speakers to teach them, and thus, lose an important source in their learning process. Regarding Okinawan as a polynomic language could help speakers and learners alike in getting rid of conformist attitudes by coming to realise that all varieties of Okinawan are indeed ‘real’ Okinawan. This could become crucial in creating a wider and more representative base for Okinawan language revitalisation.

(12)

Recognise Japan’s cultural diversity and promote intercultural tolerance.

Intercultural tolerance has to start with the recognition of diversity and with the abandonment of attempts to hierarchise that diversity. The raising of the status of Okinawan as a language in its own rights, with a place in education, could help in doing away with the negative perception of one’s own language and culture that is still so widely diffused in Okinawa, and the spread of intercultural tolerance. The pluralist nature of the polynomic model of language could help promote interregional cultural tolerance within Okinawa.

4.2. How to Apply the Polynomic Model in Okinawan Language Learning

The author, in cooperation with Okinawan language revivalist Misato Matsuda, has been involved in constructing learning materials for Okinawan based on the polynomic model. These materials were originally intended for use in the online Uchinaaguchi Shūtoku Binchōkwai study group, and they were discussed in Matsuda and Van der Lubbe (2020) and Van der Lubbe (2021a, 2022). The materials were also adopted by the Mabu-E language course at the Sakurazaka Theatre in Naha, as well as for use in Okinawan language classes at Okinawa Christian Junior College, where they serve to teach Okinawan to Okinawans from different regions and non-Okinawans alike, but have not been published for the market yet.

The goal was to enable learners of Okinawan to become aware of the parameters of regional variation in the Okinawan language, enabling them to learn their variety of choice through a unified system of instruction. We also believe it possible to make instructors and potential instructors who already have a good command of spoken Okinawan aware of these parameters of variation. This would allow for a system of language transmission where language classes, or self-study, can be supplemented with a master–apprentice system. Obviously, instilling an awareness of the parameters of variation within the Okinawan language is to be regarded as only one aspect of an Okinawan language course, the main focus being learning to speak and understand Okinawan.

The parameters of regional variation introduced in the constructed learning materials were based on the author’s field work and participant observation using Okinawan as a spoken language over the course of 10 years. Fieldwork was done in the following localities: Kunigami Village (once every two months from 2012–2015), Kumejima (monthly in the period 2015–2019, sporadically since the start of the COVID-19 pandemic in Okinawa in 2020), Katsuren Peninsula, including Yakena (7 times in 2016–2018), Ginoza (weekly in 2018–2019), and Itoman (2014), Ishikawa (weekly since 2022) and Naha (monthly 2016–present), amongst others1. Fieldwork was carried out focusing on the acquisition of data for descriptive linguistic purposes, and it has resulted in descriptive linguistics publications on Kumejima Okinawan (Van der Lubbe 2018), Sokei Okinawan (Van der Lubbe 2019, 2020a, 2020b), Sobe Okinawan (Van der Lubbe 2021b), and Naha Okinawan (Van der Lubbe 2017, 2021c).

We choose to focus on the regional variation within the South-Central varieties of Okinawan. I believe that the scope of the regional variation covered could be widened to include Northern Okinawan varieties as well.

The key to instilling awareness of Okinawan regional variation is to convey to teachers and learners with no or little background in linguistics where exactly variation resides in the language. Namely, in three levels of the language: (1) lexicon, (2) phonology, and to a limited extent, in (3) morpho-syntax. Examples are given from the five South-Central varieties listed in Table 1. We chose to not evade linguistics altogether. However, we took great care to ensure that the information we presented was understandable for a non-linguist audience. An important point is that variation in (2) phonology is based on regular, easy-to-learn sound correspondences.

The following is a sound correspondence that is widely attested in South Central Okinawan varieties. In some varieties, *s has changed into /h/ (or /ɸ/ in front of /u/, represented by f in

Table 2. Reflexes of /s/ in different Okinawan varieties.

	Shuri	Janadō	Itoman	Yakena	Sobe
fast	feesan	heesan	heesan	heehan	heehan
high	takasan	takahan	takasan	takahan	takahan
bring	sooin	hooyun	sooin	sooin	sooin
do	sun	sun	sun	fun	sun

.

The fact that /s/ turns into /h/ in some Okinawan varieties is discussed in our learning materials as part of a lesson on adjectives. Okinawan adjectives tend to end in -san (e.g., muchikasan ‘difficult’), but in varieties where /s/ has turned into /h/ -san, this tends to become -han (muchikahan ‘difficult’). The learner is encouraged to make their own targeted enquiry about the variety they wish to acquire.

Differences in lexicon are mostly found concentrated in certain areas of the lexicon. Kinship terms are an area where regional as well as social differences in lexicon can be observed (Lawrence 2015). Observe

Table 3. Variation in kinship terms in different Okinawan varieties.

	Shuri		Janadō	Itoman	Yakena	Sobe
	gentry	commoner
mother	ayaa	anmaa	anmaa	anmaa	anmaa	anmaa
father	taarii	suu	suu	suutaa	suusuu	chaachaa

:

The introduction of kinship terms provides for an opportunity to make learners aware of the fact that variation in Okinawan depends not only on geography but also on commoner–gentry differences.

There exist relatively few differences in morpho-syntax between Okinawan varieties. Morphosyntactic differences can be introduced when a form or construction that displays variation is introduced. For instance, when discussing adjectives, learners need to be made aware that there are several ways in which adjectives can be negated depending on the regional variety. One is the result of fusion of the adverbial form of adjectives ending in -ku with the topic marker ja followed by neen/neeran ‘to be absent’. The other one is the result of fusion of the abstract noun suffix -sa (or -ha depending on the region) with the topic marker ja followed by neen/neeran ‘to be absent’. Observe

Table 4. The negation of adjectives in different Okinawan varieties.

	Shuri	Janadō	Itoman	Yakena	Sobe
cold	fiisan	hiisan	hiisan	hiisan	hiihan
not cold	fiikoo neen	hiikoo neen	hiikoo neen	hiisaa neen	hiihaa neen

.

Showing the different negation of adjectives allows learners to comprehend the difference and allows them to make their own targeted inquiry about the variety they intend to master.

Introducing variety in the language in the way described above shows learners in a very concrete way that there are many ways in which Okinawan can be spoken, and by extension, that there are many ways in which one can be Okinawan. Moreover, it encourages learners to actively engage with Okinawan speakers from their own region by making their own inquiries.

I have introduced one example for each level of variety where variation resides in the Okinawan language, for a more detailed discussion is beyond the scope of this article. Refer to Van der Lubbe (2022), or Van der Lubbe (2021a, in Japanese) for a more detailed discussion, as well as other examples.

5. Conclusions

The dominant Japanese standard language ideology presents the availability of a unified standard as a prerequisite for the raising of the status of a language and the construction of learning materials. The absence of standard varieties in Ryukyuan languages is sometimes regarded as an impediment to language revitalisation (Heinrich and Ishihara 2017). I have argued that the establishment of one standard per Ryukyuan abstand language would amount to simply applying the dominant ideology on smaller regional units, and that it would fail to address the colonialist attitudes that led to the endangerment of Okinawan and the other Ryukyuan languages in the first place: hierarchisation and discrimination of variation.

In my view, the application of a polynomic model of language in Okinawan language learning could help instil an appreciation of regional and social variation in learners as well as teachers. It would have the additional benefits of allowing learners to learn Okinawan from a speaker of a variety other than their own target variety, and equipping learners and speakers alike with the ability to communicate in Okinawan with speakers of different Okinawan varieties.

Based on extensive fieldwork, I have made an attempt to identify the parameters of variation within South-Central Okinawan. Knowledge of these parameters should help teachers to convey the extent of regional variation in South-Central Okinawan to learners and should enable learners to make inquiries about their target variety.

Understanding parameters of variation can help speakers and learners comprehend that the differences within one abstand language tend to be regular and do not necessarily hamper mutual intelligibility. However, whether such parameters can be easily discovered for other Ryukyuan languages depends on many factors. Application of computational tools on wordlists of several Ryukyuan varieties seem to indicate that the phonological differences that exist in some of the Ryukyuan abstand languages indicated by Pellard (2015) are much greater than in South-Central Okinawan (Karimata 2020). For instance, Pellard subsumes the varieties of the islands of Yoron, Okinoerabu, and Tokunoshima under the roof of the Amami abstand language. However, Yoron, Okinoerabu, Tokunoshima, and Amami proper appear to be individual dialect clusters and are reportedly not mutually intelligible with each other.

In other cases, for instance, in Northern Okinawan varieties, a dialect continuum exists, but with vast differences between the ends of the continuum. This may exceed the limit of how much variation learning materials based on a polynomic model of language can reasonably cover. Extralinguistic factors may come into play here, and research with the help of traditional speaker consultants could help establish where to draw borders between varieties that could be treated as polynomic languages in their own right.

There is also anecdotal evidence that the five abstand languages indicated by Pellard (2015) are not necessarily recognised by speakers in their language choices. In Section 2, we mentioned places where colonialism has almost eradicated the use of the local Okinawan varieties in communication between inhabitants of different settlements. There are indications that this is also the case in many other places in the Ryukyus. For instance, Ishigaki Yaeyaman and Miyara Yaeyaman are mutually intelligible, but speakers of these two Yaeyaman varieties use Japanese to each other (Hammine Madoka, personal communication, May 2021). A similar situation allegedly exists in northern Amami Island. Speakers of different varieties prefer to speak Japanese, citing ‘shame’ and ‘awkwardness’ as reasons not to use the local language, even when mutual intelligibility is high (Shigeno Hiromi, personal communication, June 2020). A polynomic model of language in these varieties might help to promote interregional communication in the local languages also among traditional speakers.