Week 6 Slides LIN200 2012 - LIN200 S Introduc/on to...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: LIN200 S Introduc/on to Language Winter 2012 1 Today – The topic for today’s lecture is language diversity and linguis/c universals. –  Diversity of speaking popula/ons –  Diversity of language structures –  Linguis/c universals – what might these be? 2 Big Picture •  Diversity vs. Uniformity: star/ng point is diversity; understanding diversity and its limits = understanding what it means to be (or not be) a possible human language; this in turn leads to an understanding of the core proper/es of human language, and the ability to speak meaningfully of possible universals. 3 Diversity of speaker popula/ons: spotlight on endangered languages •  How many languages in the world •  •  •  •  •  •  •  How many of these are endangered What is language death? Which languages are dying? Why do languages die? And why now? Why is the death of a language important? Can anything be done? Should anything be done? 4 UNESCO General Assembly, 1993: “Although its exact scope is not yet known, it is certain that the ex/nc/on of languages is progressing rapidly in many parts of the world, and it is of the highest importance that the linguis/c profession realize that it has to step up its descrip/ve efforts.” (Crystal:vii) 5 Endangered Language Fund, 1995: “Languages have died off throughout history, but never have we faced the massive ex/nc/on that is threatening the world right now. As language professionals, we are faced with a stark reality: Much of what we study will not be available to future genera/ons. The cultural heritage of many people is crumbling while we look on.” (Crystal: vii) 6 Founda/on of Endangered Languages, 1995 “There is agreement among linguists who have considered the situa/on that over half of the world’s languages are moribund, i.e. not effec/vely being passed on to the next genera/on. We and our children, then, are living at the point in human history where within perhaps two genera/ons, most languages in the world will die out.” (Crystal:viii) 7 “Thousands of the world's languages are vanishing at an alarming rate, with 90% of them being expected to disappear with the current genera/on.” (Abrams & Strogatz 2003. in Nature) 8 •  •  •  •  •  •  What is language death? Which languages are dying? Why do languages die? And why now? Why is the death of a language important? Can anything be done? Should anything be done? 9 What is language death? •  A language dies when it fails to be transmiged to new speakers. •  Concretely, a language dies when the last speaker dies, (though it usually has become moribund well before then). 10 Bruce Connell(in Crystal:1): “During fieldwork in the Mambila region of Cameroon’s Adamawa province in 1994 ­95, I came across a number of moribund languages… For one of these languages, Kasabe….only one remaining speaker, Bogon, was found….Im November 1996 I returned to the Mambila…to collect further data on Kasabe. Bogon, however, died on 5th Nov. 1995, taking Kasabe with him. He is survived a sister, who reportedly could understand Kasabe but not speak it, and several children and grandchildren, none of whom knew the language.” 11 Ole S/g Andersen (in Crystal:2) “The West Caucasian language Ubuh…died at daybreak, October 8th 1992, when the Last Speaker, Tevfik Esenc, passed away. I happened to arrive in his village that very same day, without appointment, to interview this famous Last Speaker, only to learn that he had died just a couple of hours earlier. He was buried later the same day.” 12 •  Language death is some/mes news: hgp://news.bbc.co.uk/go/pr/fr/ ­/2/hi south_asia/8498534.stm • But more open than not it happens quietly, without no/ce. 13 Which languages are dying? •  How many languages are at the point of death? •  How many are endangered? 14 •  •  •  •  First of all, need to establish how many languages there are alive in the world today. Most es/mates: 6000 ­7000 But es/mates have varied between 3000 and 10,000 in recent decades. 15 •  Why is this number so hard to come by? –  Un/l lager half of 20th century, there had been few surveys of any breadth. –  Surveys that have been done in recent decades are known to be incomplete. 16 •  Side note: not all of the languages spoken today have been documented! hgp://www.bbc.co.uk/news/science ­environment ­11479563 17 hgp://www.ethnologue.com/ •  The first agempt at a world ­wide survey was undertaken by Ethnologue in 1974 (the result: 5687) •  The 13th edi/on of Ethnologue (1996) cites 6703 languages. The current online version cites 6902 languages. 18 (Crystal p. 15) 19 •  A very small number of languages accounts for a vast propor/on of the world’s popula/on. •  The 8 languages with over 100 million speakers have nearly 2.4 billion speakers between them. •  The top 20 languages are spoken by over half of the world’s popula/on. •  4% of the world’s languages are spoken by 96% of the popula/on •  96% of the world’s languages are spoken by just 4% of the popula/on. 20 The Canadian context •  Some census facts –  popula/on: 29,639,035 –  english mother tongue: 17, 352, 315 –  French mother tongue: 6,703,325 –  Italian: 469,485 –  Cantonese: 322,315 –  Arabic: 199,940 –  Cree: 72,885 –  Inuk/tut: 29,010 21 The Canadian context •  Some census facts –  Popula/on repor/ng an Aboriginal iden/ty by mother tongue: •  Ojibwa: 20,890 •  Montagnais ­Nakaspi 9,655 •  Chipewyan 575 •  Some stats about speakers that don't appear in any census table: •  Oneia 200 (1991) •  Mohawk 350 (1998) •  Haisla 25 (1991) 22 •  So when considering stats on language death, this is the context we have to keep in mind. •  More stats: –  A quarter of the world’s languages are spoken by less than 1000 speakers –  Over half are spoken by fewer than 10,000 –  2/3rds are spoken by 20,000 or less –  Nearly 500 languages have less than 100 speakers 23 •  How many speakers does it take to guarantee the life of a language? •  It depends on context: –  In an isolated rural context, 500 speakers might permit robust transmission. e.g. In many Pacific island territories, 500 is considered a large group. –  but a minority community of 500 scagered around a large city don’t stand much of a chance. 24 •  Speaker numbers can never be seen in isola/on. –  E.g. Survey by Akira Yamamoto (Crystal: 12): –  Languages with 300 ­500 speakers •  Keresan (USA): community of 600 •  Ulwa (Nicaragua): community of 2000 •  Sahap/n (USA): community of 12,000 25 •  Linguists es/mate degree of endangerment depending on region/context –  e.g. Savanna zone in Africa, a language is considered to be endangered if it has less than 20,000 speakers –  In parts of West Africa, where English andFrench creole agract high numbers of speakers, many local languages are considered to be endangered even though they are currently spoken by several hundred thousand speakers –  Even Yoruba, with 20 million speakers, has been called ‘deprived’ because it has come to be dominated by English in higher educa/on. 26 •  Why consider a language with 100,000s of speakers to be endangered? •  Because in past century we have seen languages with high numbers fall to endangered status. e.g. in 1905 Breton had 1.4 million speakers today it has 250,000 27 •  UNESCO Experts Mee/ng on Safeguarding Endangered Languages (March 2003) a framework was proposed (Brenzinger et al. 2003) which uses nine factors of vitality and endangerment for the assessment of endangerment. 28 UNESCO 9 factors for assessing language vitality 1.  2.  3.  4.  5.  6.  7.  Intergenera/onal language transmission Absolute numbers of speakers Propor/on of speakers within the total popula/on Loss of exis/ng language domains Response to new domains and media Materials for language educa/on and literacy Governmental and ins/tu/onal language avtudes and policies 8.  Community members' avtudes towards their own language 9.  Amount and quality of documenta/on 29 Why do languages die? •  Any circumstance which threatens the safety of a community, threatens language. •  Many languages have become ex/nct because of drama/c loss of life in the community of speakers (natural disasters; invasions, disease, etc). 30 •  Languages also die because of cultural changes: – migra/ons – economic pressures – language contact – cultural assimila/on – loss of power/pres/ge – erosion of ins/tu/ons 31 Why should we care? •  Some people think we shouldn’t •  Babel myth: that prolifera/on of languages is a disadvantage. Reducing the number of languages would unify people. 32 •  But there are benefits to diversity: –  Same arguments that support biological diversity apply to language. –  Diversity enriches us –  having mul/ple perspec/ves on the world is beger than having only one 33 •  Language expresses iden/ty –  Diversity goes hand in hand with iden/ty. –  Welsh proverb: “A na/on without a language is a na/on without a heart.” •  Languages are respositories of history and collec/ve knowledge –  a language encapsulates its speakers’ history in its grammar and lexicon, and is the medium for the transmission of wrigen and oral culture and histories 34 •  Having said this, one needs to guard against reduc/onist whorfian thinking. It’s NOT the case that a culture, or a strong sense of iden/ty, cannot persist if the language spoken by that popula/on ships (e.g. Irish). Nor does language loss necessarily entail loss of world knowledge. hgp://languagesopheworld.info/russia ­ukraine ­and ­the ­caucasus/more ­on ­ endangered ­languages ­and ­saving ­ideas.html 35 •  Languages contribute to the sum of human knowledge/heritage –  each language reflects a unique collec/ve experience of human existence –  “The world is a mosaic of visions. With each language that disappears, a piece of that mosaic is lost.” ­Rodrigues, in Crystal: 45 36 •  From a linguis/c perspec/ve: the loss of diversity = devasta/ng loss of data. •  For more on language loss (and language revitaliza/on) see Chapter 14 of the textbook. 37 Language diversity, linguis'c typology and linguis/c universals •  Despite the many differences among them, there must be certain proper/es whereby the languages of the world are all recognized as falling into the category of human language. •  Linguis'c Typology: the sub ­discipline of linguis/cs concerned with discovering this unity by studying the rich structural varia/on found in languages of the world. 38 •  Typology: the study of structural varia/on across languages 39 •  Typology: the study of structural varia/on across languages •  What kind of varia/on do we find across languages? So much! –  Word order (sentence structure) We’ll see examples of –  Word structure these today –  Inventory of categories –  Displacement/movement pagerns –  Embedding strategies –  THESE ARE JUST THE TIP OF THE ICEBERG! 40 The core objects of typological study: •  a. typological features and types: typologists seek to establish viable points of comparison between languages (features), e.g. word order –  Within these, they classify languages according to types. e.g. word order; types: SVO, SOV, VSO, VOS, OSV, OVS •  b. correla/ons between types: typologists look for rela/ons between types –  e.g. Greenberg's (1963) Universal 3: –  Languages with VSO order are always preposi/onal 41 The core objects of typological study: •  c. Distribu/on of types I: typologists study the distribu/on of types (e.g. which are high frequency, which are low frequency, which (if any) are universal, and why) •  c'. Distribu/on of types II: typologists study which types are found where. there has been much work on sampling methodologies with the aim of teasing apart various factors contribu/ng to uniformity/ diversity, especially: gene/c rela/onships between languages; areal rela/onships between languages; geographic factors. 42 •  In modern typology, when we talk about comparing languages, this is really short ­hand for comparing 'structural proper/es of languages'. That is, languages, per se, are not typologized. It is their proper/es (i.e. par/cular structures, construc/ons) that are typologized. This is known as par'al typology. Even when clusters of features are studied together in pursuit of significant connec/ons between them, this is s/ll par/al typology. 43 •  Par/al typology has not always been the dominant approach. Prior to the 20th century, holis'c typology was the widely prac/ced. Languages themselves were thought to be classifiable into types. The dominant classical approach to typology divided languages according to their morphology: analy'c vs. synthe'c. (more on this shortly). 44 Universals •  How do we get from structural varia/on to universals? •  Universals emerge as characteriza/ons on the LIMITS OF VARIATION. •  Not all logically conceivable combina/ons of typological features are agested. •  NB: importance of correla'ons in the formula/on of universals (to be illustrated shortly) 45 Greenbergian Universals •  This approach – analyzing the limits on varia/on by looking at possible and impossible combina/ons of features – originates with the groundbreaking work of Joseph Greenberg (1915 ­2001): •  He saw the proper work of typology not as comparing languages, per se, but comparing structures: languages do not fall into discrete types, they are 'mixed'. 46 More on Greenberg •  Exploited the view that certain structural proper/es are correlated. •  Made use of probabilis/c statements •  Incorporated syntax into typological study •  Drew agen/on to importance of a proper database; used what was at the /me a large sample, with languages from many families. 47 Basic word order typology: some classical Greenbergian observa/ons •  Basic word order pagerns across languages: 6 logically possible alignments of S(ubjects), O(bjects) and V(erbs) –  SOV, SVO, VSO, VOS, OVS, OSV 48 (Song 2004) 49 50 51 •  The diversity evident in basic word order increases when the order of other cons/tents at other levels of structure is also taken into considera/on, e.g. order within NPs, PPs, etc. •  Typologists will compare mul/ple structural proper/es to ascertain whethere there exist correla/ons between them. 52 •  Consider the following basic word order pagerns at non ­clausal levels: PP—> P NP NP —> N (AdjP) NP —> N (GenP) NP—> N (RelCl) vs. vs. vs. vs. PP—> NP P NP—> (AdjP) N NP —> (GenP) N NP —> (RelCl) N 53 •  Consider the following basic word order pagerns at non ­clausal levels: PP—> P NP NP —> N (AdjP) NP —> N (GenP) NP—> N (RelCl) vs. vs. vs. vs. PP—> NP P NP—> (AdjP) N NP —> (GenP) N NP —> (RelCl) N Notes: 1. when P precedes NP it is referred to as a PREPOSITION (as you already know). When P follows NP it is referred to as a POSTPOSITION. 2. GenP refers to a GENITIVE PHRASE as in ‘Mary’s book’ 3. RelCl refers to a RELATIVE CLAUSE as in [the book [that Mary read]] 54 Song 2004 55 56 57 58 •  The different word order pairs illustrated in the previous slides, as well as the six permuta/ons on S, O, and V orders, are all logically and formally independent of one another. •  Yet, not all conceivable cross ­classifica/ons of these features are agested. 59 •  Greenberg (1963) proposed 45 puta/vely universal statements, based on a sample of 30 languages. •  Some examples follow on the next three slides: 60 •  Universal 1: In a declara/ve sentences with nominal subject and object, the dominant order is almost always one in which the subject precedes the object (i.e. SVO, SOV, VSO) •  Universal 2: In languages with preposi/ons, the geni/ve almost always follows the head noun, while in languages with postposi/ons it almost always precedes it. 61 •  Universal 3: Languages with dominant VSO order are always preposi/onal •  Universal 4: With overwhelmingly greater than chance frequency, languages with normal SOV order are postposi/onal 62 •  Universal 25: If a pronominal object follows the verb, so does the nominal object. •  Universal 27: If a language is exclusively suffixing, it is postposi/onal; if it is exclusively prefixing, it is preposi/onal. 63 •  No/ce that some of Greenberg’s universals are wrigen as ABSOLUTES (e.g Universals 25, 27) whereas others are wrigen as TENDENCIES (e.g. Universal 4). •  All of the universals I have shown you involve a CORRELATION between two features. This kind of universal is known as an IMPLICATIONAL UNIVERSAL –  If a language has property P (e.g. VSO basic word order) it will also have property Q (e.g. preposi/ons). –  NB: This implica/on is ONE WAY only! If you restate the generaliza/on the other way it will be false. It is NOT the case that all languages with preposi/ons also have V ­ini/al basic word order. 64 Summary •  Greenberg’s findings indicate that there exist constraint on possible varia/on within human lagnuage. •  There is no reason why the two independent proper/es should correlate, logically speaking. •  The ques/on then, is why should such correla/ons exist at all? What is this telling us about human language? 65 Some classical observa/ons about morphology •  Classical typologists believed it was possible to characterize an en/re language on the basis of word structure (much like biologists could classify the skeleton of an animal on the basis of a fossil jaw); syntax, seman/cs and phonology did not enter into the picture. 66 •  The word structure typology that dominated classical thinking involved the following set of morphological ‘types’: –  Analy'c: words tend to consist of free morphemes (or compounds). There are very few bound affixes. (Low morpheme to word ra/o, open 1:1) –  Synthe'c: words tend to be complex, involving affixa/on. (High morpheme to word ra/o) 67 •  The synthe/c category incudes the following sub ­types: •  Agglu'na'ng: can have several morphemes that agach to a base. Each morpheme has a dis/nct meaning and is usually clearly iden/fiable. •  Fusional: the bound morphemes that agach to a base can FUSE so that they contain more than one meaning (aka PORTMANTEAU) •  Polysynthe'c: highly agglu/na/ng languages (may also be fusional) in which a single word can include all/much of the meaning that would be expressed as a clause in a non ­polysnthe/c language. 68 Analy/c •  English modals •  Vietnamese 69 Synthe/c  ­ Agglu/na/ng •  Lushootseed •  Swahili 70 Synthe/c  ­ Fusional •  Spanish modals •   ­o (1st person, singular, present) •   ­aré (1st person, singular, future) •  German 71 Synthe/c  ­ Polysynthe/c •  Yup’ik (Eskimo ­Aleut) 72 •  As noted above, it was eventually realized that languages in their en/rety are not well suited to holis/c classifica/on. For one, most are mixed/hybrid. For another, it is not clear that it is a valid way to think about language. •  Nonetheless, the spectrum of isola/ng/ agglu/na/ng/synthe/c/polysynthe/c languages remains a point of interest for those who study morphology and syntax. 73 •  Greenberg’s work changed the way morphological typology was approached (in much the same ways already described for syntax above). •  About half of the universals collected in the Universals Archive (see final slide) involve morphology (There are over 2000 universals listed in this archive!) 74 •  There are a number of different types of puta/ve morphological universals, e.g. –  Generaliza/ons about affix order –  Generaliza/ons about morphological inventories 75 From Bobaljik 2008:1 Notes: 1. Case refers to affixation on nouns that reflects their syntactic function (e.g. whether they are functioning as a subject or an object. See textbook pp 164-166 on case. 2. Aspect refers to the internal temporal structure of an event, e.g. whether it is completed or not. Compare ‘I had eaten the cake’ vs. ‘I was eating the cake’ Both occurred in the PAST (i.e. prior to the moment of speech) but in the former it is implied that the cake-eating event is complete, and in the latter there is no such implication. 3. Ignore the example of external case/internal case 76 From Bobaljik 2008:1 •  If a language has gender dis/nc/ons in the first person [i.e., pronouns], it always has gender dis/nc/ons in the second or third person, or in both. (Greenberg 1963, Universal 44; Corbeg 1991:131). •  If a language has a dual number, it also has a plural (Greenberg 1963, Universal 34; see also Corbeg 2000). •  A language never has more gender categories in non ­ singular numbers than in the singular (Greenberg 1963, Universal 37) 77 Lexical categories across languages •  Another noteworthy point of structural varia/on between languages involves their inventory of syntac/c categories. •  Lexical categories tend to be introduced in linguis/cs courses as a given. However, it is not obvious that all of the major lexical categories are to be found across all languages. 78 •  Some points of varia/on about which there is ongoing debate: –  Universality of noun ­verb dis/nc/on (we’ll see some examples) –  Universality of adjec/ves (we won’t see example of this) •  There are also some important cross ­linguis/c differences in the proper/es of certain categories. –  E.g languages with restricted verb classes (verbs are a closed class; some of these languages have only a handful of verbs) •  e.g. some Australian and Papuan languages (Dixon 1980: 280 ­281, Foley 1986:113 ­128) 79 The Noun ­Verb dis/nc/on •  Some ways for Ns and Vs to look nondis'nct •  Purest iden/ty would be morphological and syntac/c nondis/nctness •  E.g sort of like predicate calculus (formal logical nota/on) ! 80 •  Straits Salish (Jelinek 1995) •  All major class lexical items said to func/on as predicates, but their appearance/func/on varies depending on role they are ‘sloged into’ in the clause. When used as predicates: ! 81 •  When used as arguments: 82 •  Nootka 83 •  Mundari ! 84 Pronominal categories •  Some of the most interes/ng work on universals has involved pronominal categories. •  The following slides show some findings reported in Bobaljik 2008: 85 Seven logically possible pronominal person categories 86 Note: ignore column C 87 Summary •  Linguis/c universals emerge when we study constraints on possible combina/ons of typological features across languages. •  Iden/fying the (relevant) typological features of languages, and the extent of cross ­linguis/c varia/on, is the domain of the field of Typology. •  A full understanding of language diversity is fundamental to the enterprise of iden/fying universals and coming to a characteriza/on of what cons/tutes a possible or impossible human language 88 Universals Archive Online hgp://typo.uni ­konstanz.de/archive/intro/ index.php 89 ...
View Full Document

Ask a homework question - tutors are online