Typology

Universals  

(Absolute) Linguistic universals are statements true of every language:

  • Every language has consonants.
  • Every language has vowels.(Has been challenged: see below)
  • Every language has nouns and verbs (Has been challenged: Nootka)
      Mamu:k-ma qu:?as-?i.
      Working-PRES man-DEF.
      'The man is working.'
      qu:?as-?i-ma Mamu:k-?i.
      man-PRES Working-DEF.
      'The working one is a man.'
  • Pronouns: Every known language has pronouns for at least the speaker and the addressee: the first person I,me and the second person you.
  • Every language has at least three vowels. (Has been challenged: Kabardian, language of the Northwest Caucasus claims from 0 to 2 made; Winfried Lehman, Proto-Indo-European; in both cases, surface vowels EXIST, but are phonetically predictable!)
  • Every langauge has stops.
  • Every language has more oral vowels than it has nasal vowels.
Universal
tendency 		 

Universal Tendency
A true statement of the form: Most languages have property X

  • In the basic word orders of languages of the world, there is an overwhelming tendency for subjects to precede objects:
      Note: Mallinson/Blake (100 languages) turned up 1 OSV language in their sample, Hixkaryana.

      To remember: OVS and OSV BOTH pretty rare.

      The following examples are from Dryer, Matthew, Order of Subject-Object, and Verb.

      Ruhlen's word order sample: 427 languages
      SOV 51.5 % Japanese
      Turkish       		97.6 %
      SVO 35.6 % English
      Mandarin
      VSO 10.5 % Arabic
      Irish
      VOS 2.1 % Fijian
      Malagasy       		2.3 %
      OSV 0.2 % Xavante
      Nadeb
      OVS 0 % Hixkaryana
      SOV

      Japanese. Kuno (1973:10)

        John ga tegami o yon-da
        S O V
        John NOM letter ACC read-PST
        John read the letter.
      SVO

      Mandarin. Li and Thompson (1981:217)

        Zhangsan shoudao-le yi-feng shin
        S V O
        John read-PERF one-CLF letter
        John read the letter.
      VSO

      Irish. Dillon and O'Croinin (1861:166)

        Leann na sagairt na leabhair
        V S O
        read-PRES the.PL priests.PL the.PL book.PL
        The priests read the books.
      VOS

      Nias. Austronesian; Sumatra Indonesia; Brown (2001:538)

        I-rino vakhe ina-gu
        V O S
        3SG.REALIS.cook ABS.rice mother.1SG.POSS
        My mother cooked the rice.
      OVS

      Hixkaryana. Carib; Brazil; Derbyshire(1979:87)

        toto y-ahosi-ye kamara
        O V S
        man 3:3-grab-DISTANT.PST jaguar
        The jaguar grabbed the man
      OSV

      Nadeb. Vaupes-Japura; Brazil; Weir(1994:309)

        awad kalapee hapiih
        O S V
        jaguar child see-INDIC
        The child sees the jaguar.

      Note: There are also languages that have truly flexible word order and lack a dominant order. Latin may have been such a language.

    1. High front vowels tend to be unrounded (countering the trend, French /y/)
    2. High back vowels tend to be rounded (countering the trend, Turkish /w/)
    3. The most common stop phonemes are /p/, /t/, /k/.
    4. The most common fricative phoneme is /s/.
    5. The most common nasal phoneme is /n/.
    6. Almost every known language has at least one nasal phoneme.
    7. The majority of languages have at least one liquid phoneme. [Exceptions: Blackfoot, Dakota, Efik (Nigeria), Siona (Ecuador)
    8. Sonorants are generally voiced, although there are voiceless sonorants, voiceless nasals, for example:
        Burmese nasals
        + Voiced - Voiced
        m, n, ng, ny hm hn hng hny
      We have voiceless liquids as allophones in English, but they are not distinctive. Can you give an example? Answer

      What kind of voiceless sonorant might English have phonemically, at least in some dialects? Answer.

Implicational
Universals 		 

Implicational Universal:
A true statement of the form: If a language has property X, it must have property Y. (Property X implies property Y)

  • If a language has contrasting long vowels, it has contrasting short vowels.

    How would you know?

      Finnish, Table 8.4
      Long vs. Long /vi:li/ 'junket' /va:li/ 'election'
      Short vs. Short /suka/ 'bristle' /suku/ 'family'
      Long vs. Short /tuli/ 'fire' /tu:li/ 'wind'

    What doesnt happen:

      Fictional Finnish: 2 long vowels, 1 short
      Long vs. Long /tu:li/ 'wind' /va:li/ 'election'
      Short vs. Short NO EXAMPLE   NO EXAMPLE  
      Long vs. Short /tu:li/ 'wind' /tuli/ 'fire'

    This means there would be LONG vowel phonemes with no corresponding short vowels.

  • If a language has contrasting nasal vowels, it has contrasting oral vowels.

      French
      Nasal vs. Nasal /l on/ (long) 'long' /l an/ (lent) 'slow'
      Oral vs Oral /klo/ (clos) 'shut' /klu/ (cloue) 'nail'
      Nasal vs. Oral /b on/ (bon) 'good' /bo/ (beau) 'handsome'

  • If a language has voiced obstruent phonemes, then it will also have voiceless obstruent phonemes. The reverse is not necessarily true. Ainu (Japan) has only voiceless obstruent phonemes: /p, t, k , t∫, s /.

      Voiceless obstruents Voiced obstruents possible (Ainu)
      yes no possible (Ainu)
      yes yes possible (English)
      no yes impossible

    Question: Which are the unmarked default case, voiced or voiceless obstruents?
  • If a language has voiceless sonorant consonants it will have voiced sonorant consonants.
  • If a language has fricatives, then it will also have stops. [Of course every language has stops]. Here are some languages with NO fricatives:
    1. Gilbertese (Gilbert Islands)
    2. Kitabal (Eastern Australia)
    3. Nuer (southeaster Sudan)
  • If a language has affricates, then it also has stops and fricatives.
Hierarchies  

Implicational universals can be summarized in hierarchies.

If a language has one of the soundclasses in any of the following series, then it also has all the sound classes to the left of it.

  1. voiced sonorants > voiceless sonorants
  2. stops > fricatives > affricates
Mixing  

Implicational TENDENCY

  • If a language has only one fricative, it is most likely to be /s/.
  • If a language has one nasal it is most likeley to be /n/.
  • If a language has two nasals they are most likeley to be /n/ and /m/.
Morphological
Types 		 

Languages can be classified according to their word building strategies:

Isolating or Analytic
Most words consist of a single morpheme. Mandarin (sometimes described as languages with NO morphology)
Synthetic I: Polysynthetic
Single words made out of long strings of roots and affixes. Single words that can express sentences.
  1. Lots of multi-root words
  2. Lots of morphophonology: putting morphemes together changes their shape.
Synthetic II: Agglutinating
Words may contain multiple morphemes, but the morphemese are easily divided up into their parts, normall roots and affixes. Each affix marks one grammatical category. Turkish.
  1. Stem + affixes
  2. Each affix expresses one grammatical categorry
  3. Not much or no morphophonology
Synthetic III: Fusion
Also called inflectional languages. Words of several morphemes. Unlike agglutnating in that a single affix may "fuse" together mutliple grammatical categories.
  1. Stem + affixes
  2. Affix fuse grammatical categories
  3. Morphophonology, yes!
Word-order
Universals 		 

When verb precedes objec, we call the language VO (SVO, VOS, and VSO). When object precedes verb, we call the language OV (SOV,OSV,OVS).

  • If a language is VO it has prepositions:
    • Irish (VSO)
        a.
        VSO
        Chonaic me mo mhathair
        saw I my mother
        I saw my mother
        b.
        Preposition
        sa teach
        in house

  • If a language has OV word order then it probably has postpositions rather than prepositions
    • Guugu Yimidhirr; Australia
        a.
        SOV
        Cudaa-ngun yarrga dyinda
        dog-ERG boy bit
        The dog bit the boy.
        b.
        Postposition
        yuwaal nganh
        beach from
        from the beach

  • PPs almost always precede the verb in OV languages, and usuallly follow the verb in VO languages. NOTE: PP now stands for either prepositional phrase or postpositional phrase.
    • Japanese
        a.
        SOV
        gakusei-ga hon-o yonda
        student-NOM book-ACC read-PST
        The student read the book.
        b.
        PP precedes verb
        Taroo-ga [PP nitiyobi ni] tsuita
        Taroo-NOM [Sunday on] arrive-PST
        Taroo arrived on Sunday

    • English
        a.
        SVO
        The student read the book
        the student read-PST the book
        The student read the book.
        b.
        PP follows verb
        George left [PP on Sunday]
        George leave-PST [Sunday on]
        Taroo left on Sunday

  • Manner adverbs overwhelmingly precede the verb in OV languages and generally follow the verb in VO languages.
      hayaku hasiru
      quickly run
  • In possessive structures, there is an OVERWHELMING preference for Genitive (possessive)+ N order in OV languages and weaker preference (which is still a preference) for N+Genitive in VO languages.
      Taroo no hon
      Taroo-GEN book
Headedness  

Some light can be shed on the directionality of the word order universals with the following observation (due to Joseph Greenberg, who originally compiled this data).

    Word order constraints generally conspire to make heads either uniformly phrase-initial or uniformly phrase-final. [Languages in which heads are SOMETIMES phrase-initial and sometimes phrase final are disfavored.]

Examples:

  1. SOV languages prefer postpositions
      • [VP [V' [NP O ] [V V]]]
      • [PP [P' [NP O ] [P P]]]
    both follow the pattern:
      [XP [X' [NP O ] [X X]]]
  2. SVO languages prefer prepositions
      • [VP [V' [V V][NP O ] ]]
      • [PP [P'[P P] [NP O ] ]]
    both follow the pattern:
      [XP [X' [X X] [NP O ] ]]
    3. IN OV languages the genitive precedes the head Noun:
      • [NP [NP possessor ] [N' [N N]]]
      This makes the possessive construction head-final
    • In VO languages genitive follows noun.
        [NP [N' [N N]][NP possessor ] ]
      This makes the possessive construction head-initial
Exc  

Classify each of the following consonant inventories as impossible, marked, or neither.

    A.
      bilab. interd. alveo. alveo-p. palat. vel. glot.
    stop p/b   t/d     k/g  
    fricative   th/ð s/z ∫/zh     h
    affricate       t∫/dzh      
    nasal m   n   nh ng  
    liquid     l
    r     		       
    glide w       j w  
    B.
      bilab. interd. alveo. alveo-p. palat. vel. glot.
    stop              
    fricative   th/ð s/z ∫/zh     h
    affricate       t∫/dzh      
    nasal m   n   nh ng  
    liquid     l
    r     		       
    glide w       j w  
    C.
      bilab. interd. alveo. alveo-p. palat. vel. glot.
    stop b   d     g  
    fricative   ð z zh      
    nasal m   n   nh ng  
    liquid     l
    r     		       
    glide w       j w  
    D.
      bilab. interd. alveo. alveo-p. palat. vel. glot.
    stop p/b   t/d     k/g  
    fricative   th/ð s/z ∫/zh     h
    affricate       t∫/dzh      
    nasal              
    liquid     l
    r     		       
    glide w       j w  
    E.
      bilab. interd. alveo. alveo-p. palat. vel. glot.
    stop p/b   t/d     k/g  
    fricative   th/ð s/z ∫/zh     h
    affricate       t∫/dzh      
    nasal           ng  
    liquid     l
    r     		       
    glide w       j w  
    F.
      bilab. interd. alveo. alveo-p. palat. vel. glot.
    stop p/b   t/d     k/g  
    fricative   th          
    affricate       t∫/dzh      
    nasal m   n   nh ng  
    liquid     l
    r     		       
    glide w       j w  
    G.
      bilab. interd. alveo. alveo-p. palat. vel. glot.
    stop p/b   t/d     k/g  
    fricative              
    affricate              
    nasal m   n   nh ng  
    liquid     l
    r     		       
    glide w       j w  
    H.
      bilab. interd. alveo. alveo-p. palat. vel. glot.
    stop p/b   t/d     k/g  
    fricative              
    affricate       t∫/dzh      
    nasal m   n   nh ng  
    liquid     l
    r     		       
    glide w       j w  
    I.
      bilab. interd. alveo. alveo-p. palat. vel. glot.
    stop p   t     k  
    fricative   th s ∫   x  
    affricate       t∫/dzh      
    nasal m   n   nh ng  
    liquid     l
    r     		       
    glide w       j w  

Answer key.x

Language
Families 		 

Language are groups of languages with a reliably reconstructable single ancestor.

The familial relationships of two languages in a family are established through the use of cognates, pairs of words in the two languages that can be traced back to a common source.

Resemblances between distant relatives can be faint, because they are family resemblances:

  1. A resembles B (in features 1, 2, 3, and 4)
  2. B resembles C (in features 2, 3, 4, 5)
  3. C resembles D (in features 4, 5 and 6)
All pairwise resemblances strong. But A resembles D only in feature 4.

Example: Two cognates in English and Latin

    Latin English
    unda wave
The case that these two are cognates cannot plausibly be made without looking at a number of other intervening language relatives.

Thus, constructing large language families takes a lot of data from a lot of languages.

We have such data in the form of a large volume of written records and surviving langauges in Europe.

Reconstructing the family relationships of African lanaguges has only recently been possible.

In many ways, reconstructing the language families of North America has been even more difficult.

Indo-
European 		 

Main branches of Indo-European (living languages)
-------------------------------------------
Germanic Hellenic Baltic
Celtic Albanian Slavic
Italic Armenian Indo-Iranian

Germanic  

[Germanic [EastGermanic [(Gothic)]][NorthGermanic [Icelandic] [Norwegian][Faroese][Swedish][Danish] ] [WestGermanic [English] [German] [Yiddish] [Dutch] [Frisian] [Afrikaans]]]

    Parentheses indicates dead langauge

Frisian is the language most closely related to English, still spoken by a small group of speakers in Holland.
Afrikaans is descended from 17th century Dutch (Boers of South Africa)
Not a complete list.

Indo-Iranian  

[Indo-Iranian [Iranian [Persian(Farsi)] [Pashto(Afghanistan)] [Kurdish]] [Indic [Hindi-Urdu] [Bengali] [Marathi] [Gujarati] [Romany(Gypsy)]]]

Hindi and Urdu count as one language Hindi-Urdu
Slavic  

[Slavic [EastSlavic [Russian][Ukrainian] [ByeloRussian]][WestSlavic [Czech] [Slovak] [Polish]][SouthSlavic [Bulgarian] [Macedonian] [Serbo-Croatian] [Slovene]]]

Slavs moved into Macedonia after the time of Alexander.
Serbian and Croatian count as one language Serbo-Croatian.
Romance  

[Romance [IberoRomance [Portuguese][Spanish] ][GalloRomance [French] [Catalan] [Romansch]][Italo-Romance [Italian] [Sardinian]][Balkano-Romance [Romanian]]

Other
language
families 		 

Other language families
-------------------------------------------
Caucasian   Caucasus: between Black and Caspian seas
Includes Georgian, 4 million speakers
Chechen-Ingush, language of Chechnya
Languages of Dagestan
Karbardian, other difficult consonant systems
Area includes: Armenian, Greek, Iranian, Altaic, Mongolian, and Russian speakers
Altaic   60 languages
Central and Eastern Asia (excluding China)
Korean, Turkic, Mongolic, Tungusic
Finno-Ugric (Uralic)   Northwestern Asia/Northern Europe + Hungary
Dravidian   Southern India
Austro-Asiatic   India, Southeast Asia
Mon-Khmer and Munda languages
Austronesian   1268 languages
Philippines, Taiwan, Madagascar (Malagasy),
South Pacific (Fijian, Samoan, Maori, Tongan, Marshall Islands, Solomon Islands Easter Island)
Sino-Tibetan   China, Tibet
Afro-Asiatic   Berber, Chadic, Egyptian, Semitic

Omitted here: Australian, Niger-Congo (Central and Southern Africa), Nilo-Sahran (North Central Africa), Khoisan (Southern Africa)

Small
families 		 

Many language families are quite small, especially languages of North America:

  1. Some small Amerindian famlies
    1. Wakashan (6 langauges)
    2. Caddoan (7 languages)
    3. Iroquoian (11 languages)
    4. Timucuan (2 languages)
    5. Chimakuan (2 languages): Quileute and Chemakum, two closely-related languages of Washington State. Could be considered an isolate if these two count as dialects of one language.
  2. Yukaghir: 2 languages. Tundra (Northern) and Kolyma (Southern) Yukaghir, spoken in several small communities in the Yakut (Saha) Republic in the North-East of Russia. As of 1987, there were approximately 150 native speakers of Tundra Yukaghir, and less than 50 speakers of Kolyma Yukaghir. --- Elena Maslova
  3. Khoisan (includes non-Bantu click langauges, Southern Africa) 8-12 languages depending on how counted. This is not a universally agreed upon language family, and may be better called a phylum, in which case it is an even BETTER example of small families, including as many as 7 language families in it.
Isolates  

  1. Basque
  2. Japanese: Origin has been controversial. Altaic, SinoTibetan, and Austronesian proposals. Isolate is best guess.
  3. Ainu (Japan)
  4. Burushaski: northwestern Kashmir
  5. Etruscan: Appenian Peninsula (Italy) Pre-Roman people. Wrtitin still undeciphered.
  6. Many Amerindian languages, including Haida, Kootenay, Zuni, Coahuilteco
Phyla
Stocks 		 
  1. Austric Phyla: Austronesian and various Southeast Asian families; a famous proposal is Austronesian + Austro-Asiatic
  2. Nostratic: Indo European, Uralic, Altaic [some others]. Defended by Bernard Comrie among others, sometimes includes Afro-Asiatic, Dravidian, Kartvelian (Caucasian)
  3. "Borean language" - the hypothetical ancestor of different language families of the northern hemisphere; possible etymological matches between what he considered the five major macrofamilies of the Old World, Eurasiatic, Afroasiatic, Sino-Caucasian and Austric with potential parallels from Amerind and several African language families --- Sergei Starostin. Note: Some suggestive vocabulary lists, but nothing close to a reconstruction of Proto Borean
  4. Proto-World: The ancestor of all languages. Ruhlen Merrit
  5. More modest proposals: New world
    1. Eskimo-Aleut (Sapirian "superstock", now widely accepted)
    2. Na-Dene (Haida,Athapaskan..) Sapirian "superstock"
    3. Dene-Caucasian: North Caucasian and Na-Dene (superfamily proposed by Sergei Sarostin)
    4. Penutian (many Californian language families) Sapirian "superstock"
    5. Algonkian-Wakashan Sapirian "superstock", widely rejected
    6. Uto-Aztecan Tanaoan Sapirian "superstock", widely rejected
    7. Hokan-Siouan Sapirian "superstock" as a whole widely rejected, but Hokan widely accepted
  6. Geographically convenient proposals
    1. Australian: Very hard to establish genetic relations solidly
    2. Khoisan: Southern Africa
    3. Papuan: Non-Austronesian langauges of New Guinea, the Solomons, and Eastern Indonesia. Could be as many as 46 language families, of the largest well-established family would be West Papuan
Presentation
Topics
Universals 		 
  • Comrie, Bernard. 1989. Language Universals and Linguistic Typology 2nd ed. Oxford: Blackwell.
  • Croft, William. 2003. Typology and Universals 2nd ed. New York: Cambridge University Press.
  • Greenberg, Joseph. (ed) 1966. Universals of Language 2nd ed. Cambridge, MA: MIT Press.
  • Mallinson, George and Blake, Barry J. 1981. Language Typology Amsterdam: North Holland)
  • Timoth Shopen. (ed) 1985. Language Typology and Syntactic Description. 3 volumes. Cambridge: Cambrdige UP.
  • Ruhlen, Merrit. 1991.Guide to the World's Languages Stanford University Press.
  • Ruhlen, Merrit. 1994.On the orgin of Languages Stanford University Press.

    But watch out! Ruhlen is extremely controversial. Wikipedia article on Ruhlen Makes some good points!

  • Brown, Cecil H. 1984. Language and Living Things: Uniformities in Folk Classification and Naming. New Brunswick, N.j: Rutgers UP.
  • Tomlin, Russell S. Basic Word Order:L Functional Principles. London: Croom Helm.
  • Butterworth, Brian, Bernard Comrie, and Osten Dahl. (eds) 1984. Explanations for Language Universals Berlin: Mouton.
  • Voegelin, C.F. and F.. M. Voegelin. 1977. Classification and Index of the World's languages New York: Elsevier.

Issues:

  1. Kinds of explanation for universals
    1. Cognitive: Pick a universal that has had a cognitive explanation proposed. Explain it. Examples
      1. color terms: Perceptual foci?
      2. Word order universals about headedness. Branching strcuture occurs in uniform way, facilating proecessing strategies
    2. Functional. Necesities of a system in which contrasts must be used for communicative purposes
        Constraints on consonant/ vowel inventories due to articulatory possibilities
    3. Historical: Why do languages with suffixes but no prefixes always have postpositions? [possible explanation: suffixes EVOLVED historically from postpositions]
  2. Folk taxonomy
Presentation
Topics 		 

  1. Phonemic analysis of a language
  2. Dialect comparison
  3. L2 analysis: Phonetic issues in comparing the sound inventories of two language. Especially systemic issues rather than articulatory issues (what is a phonemic or systematic distinction rather than just a "difficult phone" issue).
  4. Historical
    1. Controversial language groups: phyla, stocks, macrofamilies:
      1. Nostratic
      2. Panutian
      3. Cavalli-Sforza's genetic evidence and theories of macrofamilies
    2. Indo-European anything: homeland, vowels, laryngeals, Neo-Grammarian hypothesis
  5. Language diversity; endangered languages. A good starting point is:
      Vanishing Voices: The Extinction of the World's Languages. 2000. Nettle, Daniel and Romaine, Suzanne. Oxford: Oxford University Press.
  6. Language and Gender
  7. Sign language
  8. Language Maintenance: Dying languages and their survival through language maintenance programs
  9. A dialect study. Compare features of two English dialects. Identify regions. Do phonological and/or lexical differences.
  10. Phonological differences for speakers of some particular L1: A good place to start. International Dialects of English Archive. The speech accent archive.