DOCSLIB.ORG

SIGMORPHON 2020 Shared Task 0: Typologically Diverse Morphological Inflection Ekaterina Vylomova@ Jennifer Whiteq Elizabeth Saleskyz Sabrina J

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
SIGMORPHON 2020 Shared Task 0: Typologically Diverse Morphological Inflection Ekaterina Vylomova@ Jennifer Whiteq Elizabeth Saleskyz Sabrina J SIGMORPHON 2020 Shared Task 0: Typologically Diverse Morphological Inflection Ekaterina Vylomova@ Jennifer WhiteQ Elizabeth SaleskyZ Sabrina J. MielkeZ Shijie WuZ Edoardo PontiQ Rowan Hall MaudslayQ Ran ZmigrodQ Josef ValvodaQ Svetlana ToldovaE Francis TyersI;E Elena KlyachkoE Ilya YegorovM Natalia KrizhanovskyK Paula CzarnowskaQ Irene NikkarinenQ Andrew KrizhanovskyK Tiago PimentelQ Lucas Torroba HennigenQ Christo Kirov5 Garrett Nicolaiá Adina WilliamsF Antonios Anastasopoulosì Hilaria CruzL Eleanor Chodroff7 Ryan CotterellQ;D Miikka Silfverbergá Mans HuldenX @University of Melbourne QUniversity of Cambridge ZJohns Hopkins University EHigher School of Economics MMoscow State University KKarelian Research Centre 5Google AI áUniversity of British Columbia FFacebook AI Research ìCarnegie Mellon University IIndiana University LUniversity of Louisville 7University of York DETH Zürich XUniversity of Colorado Boulder [email protected] [email protected] Abstract 1950 and more recently, List et al., 2016), gram- matical features, and even abstract implications A broad goal in natural language processing (NLP) is to develop a system that has the capac- (proposed in Greenberg, 1963), each language nev- ity to process any natural language. Most sys- ertheless has a unique evolutionary trajectory that tems, however, are developed using data from is affected by geographic, social, cultural, and just one language such as English. The SIG- other factors. As a result, the surface form of MORPHON 2020 shared task on morpholog- languages varies substantially. The morphology ical reinflection aims to investigate systems’ of languages can differ in many ways: Some ability to generalize across typologically dis- exhibit rich grammatical case systems (e.g., 12 tinct languages, many of which are low re- in Erzya and 24 in Veps) and mark possessive- source. Systems were developed using data from 45 languages and just 5 language fam- ness, others might have complex verbal morphol- ilies, fine-tuned with data from an additional ogy (e.g., Oto-Manguean languages; Palancar and 45 languages and 10 language families (13 in Léonard, 2016) or even “decline” nouns for tense total), and evaluated on all 90 languages. A (e.g., Tupi–Guarani languages). Linguistic typol- total of 22 systems (19 neural) from 10 teams ogy is the discipline that studies these variations were submitted to the task. All four winning by means of a systematic comparison of languages systems were neural (two monolingual trans- (Croft, 2002; Comrie, 1989). Typologists have de- formers and two massively multilingual RNN- based models with gated attention). Most fined several dimensions of morphological varia- teams demonstrate utility of data hallucination tion to classify and quantify the degree of cross- and augmentation, ensembles, and multilin- linguistic variation. This comparison can be chal- gual training for low-resource languages. Non- lenging as the categories are based on studies of neural learners and manually designed gram- known languages and are progressively refined mars showed competitive and even superior with documentation of new languages (Haspel- performance on some languages (such as In- math, 2007). Nevertheless, to understand the po- grian, Tajik, Tagalog, Zarma, Lingala), espe- tential range of morphological variation, we take a cially with very limited data. Some language families (Afro-Asiatic, Niger-Congo, Turkic) closer look at three dimensions here: fusion, inflec- were relatively easy for most systems and tional synthesis, and position of case affixes (Dryer achieved over 90% mean accuracy while oth- and Haspelmath, 2013). ers were more challenging. Fusion, our first dimension of variation, refers to the degree to which morphemes bind to one an- 1 Introduction other in a phonological word (Bickel and Nichols, Human language is marked by considerable diver- 2013b). Languages range from strictly isolat- sity around the world. Though the world’s lan- ing (i.e., each morpheme is its own phonolog- guages share many basic attributes (e.g., Swadesh, ical word) to concatenative (i.e., morphemes 1 Proceedings of the Seventeenth SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology, pages 1–39 Online, July 10, 2020. c 2020 Association for Computational Linguistics https://doi.org/10.18653/v1/P17 bind together within a phonological word); non- In this year’s iteration of the SIGMORPHON linearities such as ablaut or tonal morphology shared task on morphological reinflection, we can also be present. From a geographic perspec- specifically focus on typological diversity and aim tive, isolating languages are found in the Sahel to investigate systems’ ability to generalize across Belt in West Africa, Southeast Asia and the Pa- typologically distinct languages many of which cific. Ablaut–concatenative morphology and tonal are low-resource. For example, if a neural net- morphology can be found in African languages. work architecture works well for a sample of Indo- Tonal–concatenative morphology can be found in European languages, should the same architecture Mesoamerican languages (e.g., Oto-Manguean). also work well for Tupi–Guarani languages (where Concatenative morphology is the most common nouns are “declined” for tense) or Austronesian system and can be found around the world. Inflec- languages (where verbal morphology is frequently tional synthesis, the second dimension considered, prefixing)? refers to whether grammatical categories like tense, voice or agreement are expressed as affixes (syn- 2 Task Description thetic) or individual words (analytic) (Bickel and Nichols, 2013c). Analytic expressions are com- The 2020 iteration of our task is similar to mon in Eurasia (except the Pacific Rim, and the Hi- CoNLL-SIGMORPHON 2017 (Cotterell et al., malaya and Caucasus mountain ranges), whereas 2017) and 2018 (Cotterell et al., 2018) in that synthetic expressions are used to a high degree in participants are required to design a model that the Americas. Finally, affixes can variably sur- learns to generate inflected forms from a lemma face as prefixes, suffixes, infixes, or circumfixes and a set of morphosyntactic features that derive (Dryer, 2013). Most Eurasian and Australian lan- the desired target form. For each language we guages strongly favor suffixation, and the same provide a separate training, development, and holds true, but to a lesser extent, for South Ameri- test set. More historically, all of these tasks can and New Guinean languages (Dryer, 2013). In resemble the classic “wug”-test that Berko (1958) Mesoamerican languages and African languages developed to test child and human knowledge of spoken below the Sahara, prefixation is dominant English nominal morphology. instead. Unlike the task from earlier years, this year’s These are just three dimensions of variation in task proceeds in three phases: a Development morphology, and the cross-linguistic variation is Phase, a Generalization Phase, and an Evaluation already considerable. Such cross-lingual variation Phase, in which each phase introduces previously makes the development of natural language pro- unseen data. The task starts with the Develop- cessing (NLP) applications challenging. As Ben- ment Phase, which was an elongated period of der (2009, 2016) notes, many current architectures time (about two months), during which partici- and training and tuning algorithms still present pants develop a model of morphological inflection. language-specific biases. The most commonly In this phase, we provide training and develop- used language for developing NLP applications is ment splits for 45 languages representing the Aus- English. Along the above dimensions, English is tronesian, Niger-Congo, Oto-Manguean, Uralic productively concatenative, a mixture of analytic and Indo-European language families. Table 1 pro- and synthetic, and largely suffixing in its inflec- vides details on the languages. The Generaliza- tional morphology. With respect to languages that tion Phase is a short period of time (it started exhibit inflectional morphology, English is rela- about a week before the Evaluation Phase) during tively impoverished.1 Importantly, English is just which participants fine-tune their models on new one morphological system among many. A larger data. At the start of the phase, we provide train- goal of natural language processing is that the sys- ing and development splits for 45 new languages tem work for any presented language. If an NLP where approximately half are genetically related system is trained on just one language, it could (belong to the same family) and half are geneti- be missing important flexibility in its ability to ac- cally unrelated (are isolates or belong to a different count for cross-linguistic morphological variation. family) to the languages presented in the Develop- ment Phase. More specifically, we introduce (sur- 1Note that many languages exhibit no inflectional morphol- ogy e.g., Mandarin Chinese, Yoruba, etc.: Bickel and prise) languages from Afro-Asiatic, Algic, Dravid- Nichols (2013a). ian, Indo-European, Niger-Congo, Sino-Tibetan, 2 Siouan, Songhay, Southern Daly, Tungusic, Tur- sidered an important characteristic of the family. kic, Uralic, and Uto-Aztecan families. See Table 2 In addition, some of the families in the phylum use for more details. tone to encode tense, modality and number among Finally, test splits for all 90 languages are re- others. However, all branches use objective and leased in the Evaluation
Load more

Recommended publications
  • The Emergence of Hausa As a National Lingua Franca in Niger
    Ahmed Draia University – Adrar Université Ahmed Draia Adrar-Algérie Faculty of Letters and Languages Department of English Letters and Language A Research Paper Submitted in Partial Fulfilment of the Requirements for a Master’s Degree in Linguistics and Didactics The Emergence of Hausa as a National Lingua Franca in Niger Presented by: Supervised by: Moussa Yacouba Abdoul Aziz Pr. Bachir Bouhania Academic Year: 2015-2016 Abstract The present research investigates the causes behind the emergence of Hausa as a national lingua franca in Niger. Precisely, the research seeks to answer the question as to why Hausa has become a lingua franca in Niger. To answer this question, a sociolinguistic approach of language spread or expansion has been adopted to see whether it applies to the Hausa language. It has been found that the emergence of Hausa as a lingua franca is mainly attributed to geo-historical reasons such as the rise of Hausa states in the fifteenth century, the continuous processes of migration in the seventeenth century which resulted in cultural and linguistic assimilation, territorial expansion brought about by the spread of Islam in the nineteenth century, and the establishment of long-distance trade by the Hausa diaspora. Moreover, the status of Hausa as a lingua franca has recently been maintained by socio- cultural factors represented by the growing use of the language for commercial and cultural purposes as well as its significance in education and media. These findings arguably support the sociolinguistic view regarding the impact of society on language expansion, that the widespread use of language is highly determined by social factors.
    [Show full text]
  • Marketing Fragment 6 X 10.T65
    Cambridge University Press 978-0-521-87611-7 - A Linguistic Geography of Africa Edited by Bernd Heine and Derek Nurse Index More information Index Aari 277 consonant types 40–1 ablative, grammaticalization 235–6, 237–8 distinctively African characteristics 26 accusative case marking 253 distribution of case languages 251 accusative languages 253, 254 distribution of typological properties 31 Adamawa languages 31 documentation of languages 86, 103, 153, Adamawa-Ubangi languages 18 224 ATR vowel harmony 158 genetic relationships among languages 12, features of the Macro-Sudan belt 167 34 labial flaps 165 history of 1 labial-velar stops 157 phonological features 10 in the Macro-Sudan belt 178 prosodic features 68–80 minimal-augmented pronoun systems 169 relative frequency of typological properties question intonation 79 29 word order 161, 162, 164, 167 segmental features 39–68 adjectives 125–6, 279 specific properties 10, 28–34 paucity of 21 typological features 11 position of 279 typological relationships among 1 possibilities of expansion 125 variation space 36 reduplication of 21 African linguistics, history of 170–4 adpositions 124–5, 145, 280 genealogical super-group 173–4 adverbs 126, 237, 304 Greenberg’s nuclear African area 172–3 Afar 140 morpholosyntactic features 11, 148 Afitti 285 Niger-Kordofanian þ Nilo-Saharan Africa 15–33 super-group 173–4 ‘‘Africanisms’’ 19–27 Westermann’s Sudansprachen 170–2 as an areal-typological unit 10 ‘‘Africanisms’’ 18, 19–27, 39 comparative linguistics 17–18 gender or noun class systems 19 evidence for
    [Show full text]
  • SIGMORPHON 2020 Shared Task 0: Typologically Diverse Morphological Inflection Ekaterina Vylomova@ Jennifer Whiteq Elizabeth Saleskyz Sabrina J
    SIGMORPHON 2020 Shared Task 0: Typologically Diverse Morphological Inflection Ekaterina Vylomova@ Jennifer WhiteQ Elizabeth SaleskyZ Sabrina J. MielkeZ Shijie WuZ Edoardo PontiQ Rowan Hall MaudslayQ Ran ZmigrodQ Josef ValvodaQ Svetlana ToldovaE Francis TyersI;E Elena KlyachkoE Ilya YegorovM Natalia KrizhanovskyK Paula CzarnowskaQ Irene NikkarinenQ Andrew KrizhanovskyK Tiago PimentelQ Lucas Torroba HennigenQ Christo Kirov5 Garrett Nicolaiá Adina WilliamsF Antonios Anastasopoulosì Hilaria CruzL Eleanor Chodroff7 Ryan CotterellQ;D Miikka Silfverbergá Mans HuldenX @University of Melbourne QUniversity of Cambridge ZJohns Hopkins University EHigher School of Economics MMoscow State University KKarelian Research Centre 5Google AI áUniversity of British Columbia FFacebook AI Research ìCarnegie Mellon University IIndiana University LUniversity of Louisville 7University of York DETH Zürich XUniversity of Colorado Boulder [email protected] [email protected] Abstract 1950 and more recently, List et al., 2016), gram- matical features, and even abstract implications A broad goal in natural language processing (NLP) is to develop a system that has the capac- (proposed in Greenberg, 1963), each language nev- ity to process any natural language. Most sys- ertheless has a unique evolutionary trajectory that tems, however, are developed using data from is affected by geographic, social, cultural, and just one language such as English. The SIG- other factors. As a result, the surface form of MORPHON 2020 shared task on morpholog- languages varies substantially. The morphology ical reinflection aims to investigate systems’ of languages can differ in many ways: Some ability to generalize across typologically dis- exhibit rich grammatical case systems (e.g., 12 tinct languages, many of which are low re- in Erzya and 24 in Veps) and mark possessive- source.
    [Show full text]
  • Eleventh Creolistics Workshop
    Eleventh Creolistics Workshop Assessing old assumptions New insights on the dynamics of contact languages University of Gießen, Germany 23 to 25 March 2017 We gratefully acknowledge the support of John Benjamins Publishing Company Organising Team: Caroline Hoppe Magnus Huber Dmytro Iakovenko Melvy Imami Dieter Laufer Damaris Neuhof Annette Obert-Sochor Adrijana Repac Jasmin Ruckelshaußen Christine Stuka Viveka Velupillai Lena Wilhelm Contents 1. Programme outline _________________________ 1 2. List of Participants __________________________ 3 3. Map _______________________________________ 6 4. General Information ________________________ 8 5. Abstracts _________________________________ 13 6. Timetables for Gießen city buses and trains to Frankfurt Airport __________________________ 90 Eleventh Creolistics Workshop 1. Programme outline Time Wednesday 22 March 2017 19:30 Workshop warm-up at Restaurant Pizza Pie Time Thursday 23 March 2017 08:30- Registration 09:00 09:00- Welcome 09:30 09:30- Session 1a Session 1b 11:00 11:00 – 11:30 Coffee break 11:30- Session 2a Session 2b 13:00 13:00 – 14:30 Lunch break 14:30- Session 3a Session 3b 16:00 16:00 – 16:30 Coffee break 16:30 Language Disco 1 University of Giessen, Germany Time Friday 24 March 2017 09:30- Session 1a Session 1b 10:30 10:30 – 11:00 Coffee break 11:00- Session 2a Session 2b 12:30 12:30 – 14:00 Lunch break 14:00- Session 3a Session 3b 15:30 15:30 – 16:00 Coffee break 16:00- Session 4a Session 4b 17:30 19:30 Conference dinner at Restaurant Alt Giessen Time Saturday 25 March 2017 09:30- Session 1 11:00 11:00 – 11:30 Coffee break 11:30- Session 2 12:30 13:00 Workshop wrap-up at Restaurant Aspendos 2 2.
    [Show full text]
  • Language Contact, Areality, and History: the Songhay Question Revisited 1
    Language Contact, Areality, and History: the 1 Songhay Question Revisited Robert Nicolaï Institut Universitaire de France - Université de Nice 1. Historical background The proposal that Songhay be included in Nilo-Saharan (Greenberg 1963, see appendix 1 below) was questioned by Lacroix (1969) and debated by Nicolaï (1990). Large-scale studies of the internal subclassification of Nilo-Saharan (NS) conducted by Bender (1995) and Ehret (2001) nevertheless supported it. Both the latter scholars base their work on the genealogical tree model and share (in particular Ehret) the same assumption of the existence of a NS language family. Like Greenberg before them, they include the Songhay group of languages in the NS family, although each places it differently on his own genealogical tree. Their working methods are also different: Bender uses mass comparison and isoglosses while Ehret uses correspondence-based reconstruction. One might therefore assume that the question has been definitively settled, since major work by two independent scholars confirms the inclusion of Songhay in the NS family/phylum, even though subclassification varies. This is not however the case. The latest study (Nicolaï 2003) provides a detailed evaluation of the empirical proof adduced in support of the classification of Songhay as NS, and finds both Ehret's and Bender's results wanting. The analysis proceeds in three stages. • Impugning the NS hypothesis : it can be shown that the data and reasoning on which the classification of Songhay as NS is based do not withstand critical examination and must be rejected in view of their unconvincing empirical validity and their theoretical insufficiency. • Unexpected developments in lexical research involving the Afroasiatic languages : a Berber-Semitic origin can be postulated for a large number of terms invoked by both Bender and Ehret as proof of NS filiation and for many others as well.
    [Show full text]
  • Auxiliary Schedule 3 : Language
    AUXILIARY SCHEDULE 3 : LANGUAGE (1) This separate schedule is an auxiliary to the enumerated classes 2/9, A/Z. The concepts in it are available only to qualify those classes, and their classmarks are not to be used on their own. (2) These classmarks may be added to any class where the instruction appears to add letters from Schedule 3; e.g. in Class J Education, JV is special categories of educands, and at JVG Language groups, Non-English (or favoured language) speaking appears the instruction: Add to JVG letters C/Z from Schedule 3; so Education of Bantus would be JVG HN. Or, in Schedule 2, BQ is Linguistic group areas, and BQH N would be Bantu-speaking areas. (3) A prominent use of the Schedule will be to specify translations, version, etc., of particular works in classes like Literature, Religion and Philosophy; e.g. PM6 Versions of the New Testament, has the instruction Add to PM6 letters C/Z from Schedule 3; so PM6 RF is Greek versions. (4) Where no such special provision is made, the language of a document can be specified by adding Schedule 3 notation to 2X from Schedule 1, e.g. Encyclopaedia of economics, in French T3A 2XV. (5) If a library wishes to give a preferred position with a short classmark to a particular language (e.g. that of the place where the library is situated) the letter F or Z may be used. (6) Any language may be qualified by the preceding arrays C/E by direct retroactive synthesis; e.g.
    [Show full text]
  • Souag-Jpcl-2015.Pdf
    John Benjamins Publishing Company This is a contribution from Journal of Pidgin and Creole Languages 30:2 © 2015. John Benjamins Publishing Company This electronic file may not be altered in any way. The author(s) of this article is/are permitted to use this PDF file to generate printed copies to be used by way of offprints, for their personal use only. Permission is granted by the publishers to post this file on a closed server which is accessible only to members (students and faculty) of the author’s/s’ institute. It is not permitted to post this PDF on the internet, or to share it on sites such as Mendeley, ResearchGate, Academia.edu. Please see our rights policy on https://benjamins.com/#authors/rightspolicy For any other use of this material prior written permission should be obtained from the publishers or through the Copyright Clearance Center (for USA: www.copyright.com). Please contact [email protected] or consult our website: www.benjamins.com Explaining Korandjé Language contact, plantations, and the trans-Saharan trade Lameen Souag LACITO (CNRS — Paris III — INALCO) The intense Berber-Songhay language contact that produced Northern Songhay cannot be understood adequately without taking into account the existence of a Northern Songhay language outside the Azawagh valley — Korandjé, in Algeria — showing few, if any, signs of Tuareg contact. This article proposes a new explanation based on linguistic, epigraphic, and historical data: Western Berber- speaking Masūfa, present throughout northern Mali around 1200, founded Tabelbala to facilitate a new trade route; they chose Northern Songhay speakers, already a distinct group, for their experience in oasis farming and possibly cop- per mining.
    [Show full text]
  • Typological Studies in Language, Volume 64"KEYWORDS ""SIZE HEIGHT "240"WIDTH "160"VOFFSET "4">
    <DOCINFO AUTHOR ""TITLE "Studies in African Linguistic Typology"SUBJECT "Typological Studies in Language, Volume 64"KEYWORDS ""SIZE HEIGHT "240"WIDTH "160"VOFFSET "4"> Studies in African Linguistic Typology Typological Studies in Language (TSL) A companion series to the journal Studies in Language General Editor Michael Noonan Assistant Editors Spike Gildea, Suzanne Kemmer Editorial Board Wallace Chafe (Santa Barbara) Charles Li (Santa Barbara) Bernard Comrie (Leipzig) Edith Moravcsik (Milwaukee) R.M.W. Dixon (Melbourne) Andrew Pawley (Canberra) Matthew Dryer (Buffalo) Doris Payne (Eugene, OR) John Haiman (St Paul) Frans Plank (Konstanz) Bernd Heine (Köln) Jerrold Sadock (Chicago) Paul Hopper (Pittsburgh) Dan Slobin (Berkeley) Andrej Kibrik (Moscow) Sandra Thompson (Santa Barbara) Ronald Langacker (San Diego) Volumes in this series will be functionally and typologically oriented, covering specific topics in language by collecting together data from a wide variety of languages and language typologies. The orientation of the volumes will be substantive rather than formal, with the aim of investigating universals of human language via as broadly defined a data base as possible, leaning toward cross-linguistic, diachronic, developmental and live-discourse data. Volume 64 Studies in African Linguistic Typology Edited by F. K. Erhard Voeltz Studies in African Linguistic Typology Edited by F. K. Erhard Voeltz University of Cologne John Benjamins Publishing Company Amsterdam/Philadelphia TM The paper used in this publication meets the minimum requirements 8 of American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984. Library of Congress Cataloging-in-Publication Data Studies in African Linguistic Typology / edited by F. K. Erhard Voeltz.
    " class="panel-rg color-a">[Show full text]
  • 2. Historical Linguistics and Genealogical Language Classification in Africa1 Tom Güldemann
    2. Historical linguistics and genealogical language classification in Africa1 Tom Güldemann 2.1. African language classification and Greenberg (1963a) 2.1.1. Introduction For quite some time, the genealogical classification of African languages has been in a peculiar situation, one which is linked intricably to Greenberg’s (1963a) study. His work is without doubt the single most important contribution in the classifi- cation history of African languages up to now, and it is unlikely to be equaled in impact by any future study. This justifies framing major parts of this survey with respect to his work. The peculiar situation referred to above concerns the somewhat strained rela- tionship between most historical linguistic research pursued by Africanists in the 1 This chapter would not have been possible without the help and collaboration of various people and institutions. First of all, I would like to thank Harald Hammarström, whose comprehensive collection of linguistic literature enormously helped my research, with whom I could fruitfully discuss numerous relevant topics, and who commented in detail on a first draft of this study. My special thanks also go to Christfried Naumann, who has drawn the maps with the initial assistence of Mike Berger. The Department of Linguistics at the Max Planck Institute for Evolutionary Anthropology Leipzig under Bernhard Comrie supported the first stage of this research by financing two student assistents, Holger Kraft and Carsten Hesse; their work and the funding provided are gratefully acknowledged. The Humboldt University of Berlin provided the funds for organizing the relevant International Workshop “Genealogical language classification in Africa beyond Greenberg” held in Berlin in 2010 (see https://www.iaaw.hu-berlin.
    [Show full text]
  • The Subclassification of Songhay and Its Historical Implications
    DOI 10.1515/jall-2012-0008 JALL 2012; 33(2): 181 – 213 Lameen Souag The subclassification of Songhay and its historical implications Abstract: This paper seeks to establish the first cladistic subgrouping of Songhay explicitly based on shared arbitrary innovations, a prerequisite both for distin- guishing recent loans from valid extra-Songhay comparanda and for determining how Songhay spread. The results indicate that the Northern Songhay languages of the Sahara form a valid subfamily, even though no known historical records link Tabelbala to the others, and that Northern Songhay and Western Songhay (spoken around Timbuktu and Djenné) together form a valid subfamily, North- western Songhay. The speakers of Proto-Northern Songhay practised cultiva- tion and permanent architecture, but were unfamiliar with date palms. Proto- Northwestern Songhay was already in contact with Berber and probably (perhaps indirectly) with Arabic, and was spoken along the Niger River. Proto-Songhay itself appears likely to have been in contact with Gur languages, confirming its relatively southerly location. This result is compatible with two scenarios for the northerly spread of Songhay. On Hypothesis A, Northern Songhay spread out from an oasis north-east of Gao, probably Tadmakkat or Takedda, and North- western Songhay had been spoken in areas west of Gao which now speak Eastern Songhay. On Hypothesis B, Northern Songhay spread out from the Timbuktu region, and Western Songhay derives from heavy “de-creolising” influence by Eastern Songhay on an originally Northern Songhay language. To choose be- tween these hypotheses, further fieldwork will be required. Keywords: Songhay, historical linguistics, cladistics, language contact, Nilo- Saharan Lameen Souag: Laboratoire des langues et civilisations à tradition orale, Centre Nationale de la Recherche Scientifique, E-mail: [email protected] 1 Introduction The origins of Songhay have been a matter of debate for decades (notably, Greenberg 1963; Nicolaï 1990; Dimmendaal 1992; Nicolaï 2003; Kossmann 2005a).
    [Show full text]
  • Languages of the World
    CONCISE ENCYCLOPEDIA OF LANGUAGES OF THE WORLD COORDINATING EDITOR KEITH BROWN University of Cambridge Cambridge, UK CO-EDITOR SARAH OGILVIE University of Oxford Oxford, UK Amsterdam Boston Heidelberg London New York Oxford Paris San Diego San Francisco Singapore Sydney Tokyo SUBJECT CLASSIFICATION Note that italicized titles are included for classification purposes only and do not cross-refer to articles. Introduction Eastern List of Abbreviations Akkadian Classification of Languages Southern Ethiopian Semitic Languages Areal Linguistics Amharic Africa as a Linguistic Area Ga'sz Balkans as a Linguistic Area Tigrinya Ethiopia as a Linguistic Area Europe as a Linguistic Area Altaic Languages South Asia as a Linguistic Area Southeast Asia as a Linguistic Area Mongolic Languages Tungusic Languages Evenki Afroasiatic Languages Turkic Languages Ancient Egyptian and Coptic Azerbaijanian Berber Languages Bashkir Chadic Languages Chuvash Hausa Kazakh Cushitic Languages Kirghiz Highland East Cushitic Languages Tatar Oromo Turkish Somali Turkmen Omotic Languages Uyghur Wolaitta Uzbek Semitic Languages Yakut Eblaite Central Arabic Australian Languages Arabic Languages, Varation in Australia: Language Situation Aramaic and Syriac Mirndi Hebrew, Biblical and Jewish Wambaya Hebrew, Israeli Pama-Nyungan Jewish languages Arrernte Maltese Gamilaraay Phoenician Guugu Yimithirr Syriac Jiwarli Ugaritic Kalkutungu xii Subject Classification Kaytetye Caucasian Languages Morrobalama Abkhaz Pitjantjatjara / Yankunytjatjara Georgian Warlpiri Lak Southern Daly
    [Show full text]
  • A Grammar of Kusaal Agolle Dialect
    A Grammar of Kusaal Agolle Dialect David Eddyshaw 2019 i Contents Preface...................................................................................................................... vi Abbreviations.......................................................................................................... viii Interlinear glossing...................................................................................................ix Transcription conventions..........................................................................................x Sources...................................................................................................................... xi Other studies of Kusaal...........................................................................................xiii References............................................................................................................... xiv 1 Introduction............................................................................................................... 1 1.1 The Kusaasi people.............................................................................................1 1.2 The Kusaal language...........................................................................................4 1.2.1 Status...........................................................................................................4 1.2.2 Dialects........................................................................................................4 1.2.3 Related languages........................................................................................5
    [Show full text]