Endangered Language Research Articles

Enhanced unsupervised neural machine translation by cross lingual sense embedding and filtered back-translation for morphological and endangered Indic languages

ABSTRACT The fast growth of communication technology has brought nations and their cultures closer together, and the demand for cross-language communication has risen tremendously. There is a different learning method to connect the source language to the target language in which unsupervised learning is a blessing for low-resource languages. The unsupervised machine translation is always problematic to those languages which are morphologically rich and low resources languages. Morphologically rich and low-resource language does not provide good results in machine translation if the translation is from morphologically less complex language to morphologically more complex languages. In this paper, we have improved the unsupervised neural machine translation by tackling the ambiguity problem and the quality of pseudo-parallel sentence pairs generated through back-translation for morphologically rich languages. The ambiguity problem is solved by taking the cross-lingual sense embedding at the source side instead of cross-lingual word embedding. By giving more weight to better pseudo-parallel sentence pairs in the back-translation step, the quality of pseudo-parallel sentences is increased. Different evaluation metrics have been used to check the robustness of the model and compared with different baseline models. The experiment is performed on different morphologically rich languages English-Hindi, English-Tamil, English-Telegu, and one low-resource endangered kangri language.

The acquisition of demonstratives in a complex noun class system

We present an exploratory study of 2- to 3-year-old children’s acquisition of the demonstrative system of Eegimaa (ISO 369–3 bqj), an endangered language belonging to the Jóola cluster of the Atlantic family of the Niger-Congo phylum, spoken by about 13,000 speakers in southwestern Senegal. Eegimaa demonstratives express distance from speaker (proximal, medial and distal) and the agreement categories of number and gender, as well as having four morphological types that create an additional dimension of complexity for children to learn. These demonstrative types are each associated with a range of syntactic functions with partial overlaps.From nearly seven hours of recordings, including children at three age points (2;0, 2;6 and 3;0), we extracted 218 demonstrative tokens from the children’s speech, matched with 205 tokens from a sub-sample of caregiver speech. The youngest children can be described as restricting their use of demonstratives to a small set of learned items, with evidence of generalisation and productivity arising over the course of development, alongside an increase in frequency and development in distribution patterns of the various demonstrative forms to more target-like usage in the 3;0 sample. At age three we observe more variation by syntactic function. As has been found in other languages, children acquiring Eegimaa seem to make use of the diverse forms of demonstratives early, but they do not yet make use of the full range of forms even at age three, when they are beginning to produce more systematic forms of the demonstratives across syntactic contexts and with a variety of genders.

A Modern Western Aramaic Account of the Syrian Civil War

Modern Western Aramaic is one of the most critically endangered Aramaic languages, and the only extant member of the Western Aramaic subfamily. Its speakers are among the few who have not migrated away from their original territory, but the Syrian Civil War (2011–present) has accelerated its endangerment and resulted in the abandonment of one of the three villages in which it was previously spoken. Modern Western Aramaic thus provides us with an opportunity to document a language precisely as it succumbs to some of the primary causes of language endangerment, i.e., conflict and migration. The following text, collected in 2020, discusses this community’s experience during this conflict from the perspective of one of the few remaining competent speakers. We have contextualized this text with a discussion of the language and its grammar in comparison with the most recent descriptions, all of which were made decades prior to the war, when the language was much less endangered.

Directions for Indonesian Language Policy in Education: Towards A Translingual Perspective

This study overviews literature regarding the development of Indonesian, language endangerment, language attitudes, and ideologies, as well as emerging teaching methodologies to provide the context of the current challenges facing Indonesia’s sociolinguistic situation. The establishment of Bahasa Indonesia as the national language helped unite the country but, at the same time, created new challenges for addressing linguistic diversity. The dissemination of English education has complicated the situation further and has been associated with exacerbating inequalities. These inequalities have not only had consequences for linguistic diversity but also for access to employment and education. Language policies continue to treat language as static rather than dynamic processes which involve one’s communicative repertoire. We argue that while some language scholars studying Indonesia have begun to acknowledge the advantages of translingual/multilingual education, there still remains under-addressed obstacles to its successful implementation and the risk of co-option for ulterior purposes. We conclude that advocacy for language policy in Indonesia should be a part of a broader effort to address the needs of marginalized communities and harness students’ diverse and dynamic linguistic repertoires to improve the quality of learning and inclusivity in the classroom

Phonological features of Caijia that are notable from a diachronic perspective

Abstract This study sets out several phonological features in Caijia that are notable from a diachronic point of view. The Caijia language is an endangered language spoken in northwestern Guìzhōu, China. It was first formally documented in the early 1980s and is generally viewed as a Sinitic language. Some aspects of Caijia phonology are noteworthy from the perspective of historical phonology. There exist features which cannot be accounted for in terms of Middle Chinese (MC), such as the retention of the contrast between Old Chinese (OC) T-type and L-type onsets in words with d- or dr- in Middle Chinese. Moreover, Caijia also demonstrates features which are observed or preserved in Middle Chinese, but absent in mainstream modern Sinitic varieties, including the retention of bilabial stops in words with initials Fēi/Fū/Fèng. This study will also explore the implications certain phonological features have for the classification of Caijia in the Sinitic clade and examine the relationship between Caijia and Bai.

Globalization and the African Praise-Poet: The Example of ??b?b? K?n? ?z?n Ibe (The Chief Oral Poet of the ?z?n Nation)

Scholars exploring the role of praise-poets in contemporary Africa have not yet looked at praise-poets' response to globalization-related issues, such as immigration, gender, environment, HIV/AIDS, and language endangerment and death. This article focuses on the relations between African praisepoets and globalization. It examines how Chief Adolphus Munamuna, an Ịzọn praise-poet, has used his art to address two issues of globalization affecting Africa in general and the Ịzọn in particular: HIV/AIDS and language endangerment and death. It points out that even globalization has not wiped out the social function that praise-poets have been serving since precolonial times; rather, it has opened up new topics and contexts for them to explore.

Språkdokumentasjon innen fennistikken og kvensk

The study of the Finnish language – called Fennistics – was focused on collecting Finnish dialect material from very early on. During the 19th century interest in studying dialects was governed by the idea that dialects could be used to develop modern written Finnish. However, gradually the study of dialects also became an area of study in its own right. Collecting material on Kven dialects belonged to the larger project of Fennistic materials collection from the very beginning. Therefore, many Kven dialect materials can be found in Finnish dialect archives. The documentation which resulted from dialect collections gathered via the field of Fennistics have been used in the process of revitalizing Kven. In this article, language documentation is defined to be an activity which also includes traditional dialectology. By contrast, documentary linguistics is a field of linguistics established in the 1990s which focuses on revitalizing endangered languages. The difference between language documentation in the Fennistic tradition and documentation of endangered languages in the field of documentary linguistics today is also discussed.

Low hanging fruit and the Boasian trilogy in digital lexicography of morphologically rich languages

Online lexicographical resources for the morphologically rich Indigenous languages in Canada use a wide range of strategies for conveying their language’s morphological system, i.e. how words are inflected and derived, which this paper illustrates in a survey of seventeen bilingual online resources. The strategies these resources employ boil down to two basic approaches to the underlying structure of the resource: 1) a lexical database, or 2) a computational model. Most resources we surveyed are constructed around lexical databases. These assume the word(form) as the basic unit, an assumption that makes it difficult to incorporate the language’s sub-word, morphological structure in full detail. However, one resource uses a computational morphological model to bring the language’s morphology into the core of the lexicon – this proved to be a “low-hanging fruit” in the application of language technology that had been accomplished within a reasonable time-frame, as has been advocated by Trond Trosterud. We discuss the value created and questions raised by this approach and argue that it successfully overcomes the traditional Boasian three-way partition of dictionary, grammar, and text, creating integrated language resources that meet the modern needs of low-resource endangered languages and their communities.

Creating a corpus for Kven, a minority language in Norway

Language documentation, including the development and use of corpora, is frequently linked to revitalisation. This is also the case for the Kven language, a Finnic minoritised language, traditionally spoken in the two northernmost counties of Norway. Kven is a recognised minority language in Norway, protected by the European Charter for Regional or Minority Languages. This status led to increased efforts to document Kven, including the development of the Ruija Corpus, consisting of recordings of interviews in Kven. The corpus was an important tool for the standardisation of Kven. In this article we describe how the corpus was developed and account for search functions, including a discussion of the limitations of the corpus. We also discuss the role of corpora and other online tools for language revitalisation, with a particular focus on the standardisation of Kven and conclude by reflecting on how expertise also resides with the speakers of an endangered language and that they have a right to be involved in efforts of language documentation and revitalisation.

Unsupervised SMT: an analysis of Indic languages and a low resource language

ABSTRACT The rapid globalisation in language technology and the Internet’s fast expansion have brought nations and their cultures close together, and the demand for inter-language interactions has risen enormously. However, in many low-resource languages (LRL) pairings and areas, Machine Translation (MT) is still not viable because of a lack of parallel data. The challenge of MT is still unsolved. Recent studies employing monolingual datasets have shown excellent outcomes in Phrase-based Statistical MT (PBSMT) and Neural MT (NMT) systems. However, earlier researchers have demonstrated that unsupervised Statistical MT surpasses unsupervised NMT, especially for different language pairings. The study unveils the compendium of ten unsupervised SMT systems translation tasks utilizing a monolingual dataset from the Dravidian and Indo-Aryan language families; and a low-resource endangered language. The machine-translated experimental outcomes examined the system using different tokenizers and investigated them for various language pairs using different evaluation metrics for various iterations. The statistical significance of test results has been computed for each evaluation metric to check the true system quality of the translation tasks.

Language Endangerment: A Critical Destiny of Indigenous Mother Tongues Decline in Nepal

Language is the vehicle for people to communicate socio-cultural reality which transmits culture and knowledge from one generation to another. Language and culture are intimately related and dependent on each other. Endangerment is one of the stages of language loss which is a very critical issue in the 21st century because the extinction of each language results in the irrecoverable loss of unique expression of the human experience and the culture of the world. A language is dying every time in the world which will create difficulty to understand the pattern, structure, and function of human languages, prehistory, and the maintenance of the world’s diverse ecosystems in the future. Language is thus essential for the ability to express cultural knowledge and the preservation of one is paramount to the preservation and further development of the others. It is believed that half of the world's languages will disappear within this century. In Nepal, there are 92 indigenous or minority languages out of 123 are in endangered situation. This paper describes the context and stages of language endangerment in the global and Nepalese context and indigenous language situations of Nepal.

Language variation in a shifting community: Different patterns of noun incorporation in Modern Chukchi

Research questions: This study asks whether an interface phenomenon such as noun incorporation (NI) displays meaningful socially conditioned variation in the endangered polysynthetic language, Chukchi, by investigating whether speakers of all levels of experience or proficiency make use of NI in a consistent, rule-governed way. Design and methodology: This study compares production data from small groups of speakers of a moribund language. Study tasks include a controlled production task in which speakers are asked to construct sentences using provided lexical items. The lexical items were conditioned so as to trigger NI in certain stimuli (on the basis of verbal valency and argument animacy). Data and analysis: The production data was transcribed and coded for the occurrence and structural type of NI (compounding vs. syntactic incorporation). The results were compared across three groups of speakers: conservative older speakers, younger attriting speakers, and new speakers. Findings/conclusions: NI frequency and productivity clearly differ among the three groups. CSs use incorporation frequently and productively in the expected contexts, while ASs use productive incorporation only in familiar contexts, followed by NSs who make little to no use of incorporation. All speaker groups display knowledge of the appropriate circumstances in which to use incorporation. Originality: This study makes use of a novel experimental methodology in studying several under-researched areas: variation in traditional Chukchi, shift-induced variation in a polysynthetic language, and NI as a locus of variation. Significance/implications: This study contributes to our understanding of the behavior of non-normative speakers of endangered languages and demonstrates that they play a role in language preservation. The study shows that the diffuse nature of the Chukchi speech community is different from comparatively well-studied shift settings (especially in the North American and European contexts) in its lack of a community of use or practice, which presents unique challenges in language maintenance.

A variationist analysis of the subjunctive mood in a sociolinguistic corpus of spoken Francoprovençal

Variationist studies on the subjunctive in Romance remain focused on dominant nation-state languages. Conversely, very little research focuses on the minoritised varieties (often labelled ‘regional’ languages) with which these dominant languages are in contact. The framework deployed for elucidating on change in the subjunctive has not yet been adequately tested in language endangerment. Using a novel sociolinguistic corpus of spoken Francoprovençal, the variationist analysis presented in this paper offers a preliminary investigation of the synchronic shape of the subjunctive in one variety, and assesses the extent to which social and linguistic factors influence variant selection. The results indicate a progressive weakening of a subjunctive mood that is increasingly lexically motivated. The available evidence suggests that a combination of tense-aspect-mood variation and morphophonological regularisation are contributing to the decline in subjunctive selection, with no social factors playing a role in the variation.

Literary Translation Techniques in Professor Pamela Smith’s Translation of Akinwumi Is̩ola’s Ogun Omode to Treasury of Childhood Memories

Literary translation is a veritable tool to mitigate the endangerment and imminent extinction of African indigenous languages and literature. Professor Pamela Smith has taken up the challenge to translate Ogún O̩mo̩dé written by Professor Akinwumi Is̩o̩la into Treasury of Childhood Memories among many others. Yorùbá literary critics, translation experts, and linguists are yet to scrutinize the literary translation techniques in the translated text. This study, therefore, examined the literary techniques adopted by the translator in the Target Language (TL). The study employed a qualitative research design with a close reading and content analysis of both the Second Language (SL) and Target Language (TL) texts using the Hutardo’ (2002, 498) model of literary techniques for data analysis. The findings of the study showed that: the translator adopted many literary techniques that make the TL fascinating and pleasurable to readers, but the following techniques were more predominant, these are modulation; compression; elision/omission; linguistic amplification; borrowing, calque; compensation; adaptation; and particularization. The essay concluded that the translator's high level of bilingual and bicultural competence and the literary translation techniques adopted to make the contents of the source text easily transposed and rendered in impeccable English language in the TL.

The Metaphorical Use of Body Parts in Forming Counting expressions: Evidence from Tāti Language Group

AbstractMathematics is a particularly important challenge for embodied approaches to cognition, as it is probably the most abstract domain of human knowledge. Humans use metaphors in all aspects of life. This paper studies the effects of human body parts on numerals, numeral systems, and mensural and sortal classifiers. The evidence for this paper comes from the Tāti language group, an endangered Iranian language of the Indo-European language family. The Tāti data shows these languages make use of base-10, base-20, and base-50 numeral systems, some of which are among the most common and earliest counting systems worldwide, while the last is unique and peculiar to the area. Body parts may also play an important role in forming mensural and sortal classifiers, as is the case in the Tāti language group.

Cross-dialectal synchronic variation of a diachronic conditioned merger in Tlingit

Abstract Crosslinguistically rare sounds may be uncommon as a result of being phonologically marked (Trubetzkoy 1939) or due to articulatory or perceptual biases (Maddieson 1998). Certain types of sound changes are often argued to have roots in articulatory and perceptual biases (Blevins 2004). But in cases where there is limited data available, such as with understudied languages, it may be difficult to find evidence for the roots of sound changes. Synchronic variation can be used to provide evidence for diachronic sound changes (Blevins 2004; Lindblom 1990; Ohala 1993), which is particularly useful when historical data is limited. In this investigation we discuss phonetic biases, including acoustic and perceptual factors, that contribute to a set of sound changes in Tlingit, a critically endangered Indigenous language of Alaska, British Columbia, and the Yukon, that resulted in a primary split-merger (Blust 2012). This investigation provides further support for including explicit discussion of synchronic variation as part of the description of understudied languages. We propose that there should be a stronger emphasis on documenting and analyzing variation within understudied languages because excluding variation potentially masks significant intralinguistic and crosslinguistic phenomena.

Edutaining with indigeneity: Mediatizing Ainu bilingualism in the Japanese anime, Golden Kamuy

This article examines how the Ainu language, a critically endangered indigenous language of northern Japan, is deployed in the Japanese animation Golden Kamuy (Noda, 2018-). The Ainu language is mediatized as loanwords and the language of the community primarily in Ainu–Japanese interactions. The Ainu co-protagonist’s skillful translation of loanwords makes Ainu culture accessible to the viewers. Such considerate representations reflect the creator's extensive research and commitment to overcoming discrimination against Ainu. Unfortunately, however, careful analysis of overall Ainu language allocational patterns reveals that the show cannot escape from reproducing race-related discourses in Japan, reflecting Japan's colonial legacy.

Language simplification in endangered languages?

Abstract The present paper examines a hypothetical correlation between language endangerment and the simplification of nominal and verbal inflections. After contrasting the complexities exhibited by two endangered languages (Eastern Huasteca Nahuatl and Wymysorys) with the complexities of their non-endangered predecessors (Older Nahuatl and Middle High German, respectively), the authors conclude that the endangerment-simplification entanglement cannot be demonstrated. First, although Wymysorys (a more endangered code) is slightly more simplified than Nahuatl (a less endangered code) as far as the nominal domain is concerned, this relationship is reversed in the verbal domain. Second, simplifying tendencies are not radical, with a number of innovative complexifying processes being also present. Third, when attested, simplification constitutes part of a “natural” language evolution rather than a process resulting from the endangerment.

Acoustic Analysis of the Baram Phonemes

Baram is a critically endangered Tibeto-Burman language spoken in the Gorkha District of central Nepal. This work carries out a preliminary acoustic study of the language. Baram plosives are acoustically characterized by VOT, the fricatives by the peak value and the affricates by the rise time and friction duration. Similarly, the nasals, lateral, trill and glides are characterized by formant structure. Similarly, the distribution of the F1 and F2 reveals the height of the tongue and frontness/backness of the vowels. The height of the vowels is also depicted by their intrinsic F0 as it increases as the height of the vowel increases. Baram has five diphthongs which are close diphthongs and the second target is always [i].

Enhanced Electronic Grammars Online

Enhanced Electronic Grammars Online (EEG) is a full-text linguistics database available on the De Gruyter platform. EEG offers the ability to conduct searches across dozens of comprehensive grammars describing rare and endangered world languages. Intended for use by researchers into general linguistics and linguistic typology, EEG would also be a good resource for undergraduate teaching. The review notes some issues with the search interface and metadata.