7 typology
Language
A rvpolocv is simply a categorizationof some range of phenomenainto
various types. To "typologize" something is to group its parts into types. For
example,we often hearj (DP) V Finnish,Hindi . . . )
(Japanese,
OV Languages
" VP -+ V (DP) English. ' . )
(Mandarin,Indonesian,
VO Languages
E
E What is interestingabout this typology is that the order of object and verb in the
E verb phrasetends to correlatewith other aspectsof the syntax of the language.For
certainly have postpositions,
l pxample, if a Ianguagehas OV order, it will almost
it will almost
!$ather than prepositions. Conversely, if a language has VO order,
!$ertainly have prepositions. Also, in OV languages, inflectedauxiliariesalmost
f; $tways come after the verb, whereasin VO languages.auxiliaries usually precede
In short. there are, generally speaking,two major types of languagesin
$Ene v"rU.
j pre world: those in which syntactic heads normally precedetheir complements,
Eo?ndthose in which syntactic headsfbllow their complements:
E_&
RF+l Vo ianguages: oV languages:
;-E Head ComplementComplement Head
R; vP-+v DP DP v
ss IP -+ AUX VP VP AUx
,E.9 PP --> P DP DP P
E9 DP + D NP NP D
E ? B"ruuseeverylanguage alwaysin a stateof change,
is and the ordertlf head
(End complement a particular
in phrasal category onevariablethat may change
is
{ $vertime, these are
correlations not absolute. are
However,they highly signi{icant
t tiro*
pointof view.It is certainly merecoincidence languages
a statistical not that
correlate this way.The problemfbr linguistictheoryis why
in this shouldbe the
case.Many linguists haveapproached problemfrom differentdirections,and
this
typology
Syntactic 193
Table7 .1 Summaryof Greenberg's(Jniversals(from appendix2 of Greenberg1963)
Greenberg's
Universal Parameter coffelation
#l Main clauses v-o o-v
#34 Adpositions Prepositions Postpositions
#2 Genitive (possessor) and N-G G-N
head noun
#r7 Head noun & adjective N-Adj Adj-N
#24 Relative clausesand N-RelCL ReICL-N
head noun
#22 Comparatives Adj-Mkr-Std Std-Mkr-Adj
#16 Inflectedauxiliaries Aux-V V-Aux
#9 Questionparticles Sentence-initial Sentence-final
#12 Questionwords Sentence-initialor elsewhere Sentence-initial
#21 Affixes Prefixes Suffixes
we willnottry to summarizethese Rather,wewill
here. simplypresentthe lindings
of somevery importantfbundational research. and then give someexamples of
, languages represent
that eachof the majortypes.
.E The foundational work in syntactic typologywas doneby Joseph Greenberg
! in th. early 1960s. Greenberg compared syntacticcharacteristics thirty
the of
{ languages foundseveral
and interestingcorrelations.In particular, noticed
he that
in
languages his sample tendto havea basic,or unmerked, syntactic structure,
! the
; gnd that the order of certainelements this basic structure
in correlate with the
E Srdersof otherelements. Table?.1 summarizes correlations Greenberg
the that
53tSO:; observed VO and OV languages.
for Thesehavecome to be known as
$pGreenberg's Universals," to
sincetheywereassumed represent correlationsthat
$$rf a trueuniversally. for all languages.
i.e.,
i^E tt is importantto recognize that Greenberg simply observed certaincorrela-
'E$onr. He did not attcmptto providea reason (i.e.,to "motivate")thosecorre-
for
-Ffirtions,
or evento testthem for statistical In
significance. this sense, Greenberg
PF
g$ia not attempr predictconstituent
to ordersin asyet unstudied languages. Since
i fg6f , muchresearch revealed
has problems with Greenberg's originaltypology'
criticisms,anclextensions Greenberg's
of work are fbund
i {igniticant revisions,
s^'br Hawkins (1983). Payne
D. (1985),Mithun (1987), Dryer (1988,1992). and
E hawkins(1994).Inanimportantcorrection,Dryer(1988)showsthatGreenberg's
.17
E ilniversal number (theorderof adjective headnoun)doesnot holdwhena
and
ofr is
languages considered. Neverlheless,Greenberg's work stimu-
f$rger sample
-.Fted
A the field of typologicallinguistics to
andhascontinued be very influential.
- " In thefollowing sections will explainandillustrate
we someof thecorrelations
described tablel. I , usingexamples
in from two typologicallydistinctlanguages-
Japanese and Malagasy.
T94 LANGUAGE TYPOLOGY
Examples an OVand a VO language
of
Japanese andMalagasy aretwo languagesthatconform to Greenberg's
observationsvery closely. Most languagesare not this ideal, but thesewill suffice
to exemplify a rather remarkab.lerecuring pattern in the syntactic structuresof
the world's languages.
The following example illustrates OV constituent order in Japanese.In this
clause, the inflected VP is in brackets. Notice that the object, lntz, precedesthe
Yerb,nita:
(s) OV : Taro ga inu 0 mita 'Taro saw a clog.
Tar o NO M dog Acr(r saw
Here is a possibletree diagram of this clause:
(6) S
DP
-,+-. VP
N PD
I .-'--.
N DPV
€ ll,^
= I I NP D
_,,lll
EIIN
€lll
sl l
6 Taro sa inu o mita 0
..61
e99
3 Notl"" that under the IP node, the tree "branches" to the left. The syntactichead
fi
a branchesoff to the left.
5if eachphrasalcategory is on the right, and its complement
$$n tree diagrams of English sentences, you will notice that the branching tends to
bo..
HSixtena the right (see,e.g.,example7), in chaptertr).For this reason,languages
to
EQt . Japanese sometimescalled LBFT-BRANCHTNG languages, contrast
are in
E.b Englistrantl other VO languages, which can be termedRIGHT-ITRANCHING
E{anguages. Other terms sometimes used for these two language types are head-
pr$nal ancl heacl-initial languages,respectively,or complement*hea
cG ]- NOM man GEN dog ACC saw
'a
aaL
^
'Taro saw (0r lookeclat) a/the lriln's dog.'
Fc . l
Ffl Noti." that uncler IP nocleof examplelfi, thereis a VP, then a DP, then
+F
the
*.$n mp,then anotherDP. Among many other facts, this diagramcaptures the
- "rmportant fact that genitivenouns(thosefor which the determiner Japanese
in
is no) arenested within otherNPs.In otherwords,genitivenounsare a kind of
modifier,or optionalcomplement, othernouns.
of
r96 LANGUAGE TYPOLOGY
Exarnple i I illustratesa RELATIvE cLAUsE in Japanese. Relative clauses
are clauses that modify nouns and are embedded within noun phrases. We will
have much more to say about relative clausesin chapter N For now, just notice
{}.
that the Relative Clause (bracketedby {curly braces}; comes before the Noun it
modifies,ina:
6l) RC N: Taro ga [{niku o tabeta} inu ol mita
NOM meat ACC ate dog AC(l SaW
'Taro saw the dtlg that ate the meat.'
has
Japanese Tro.rrpositional phrases This is
phrases.
ratherthanprepositional
another commoncharacteristic verb-finallanguages. !l we seethe POST-
of In
fctllowingthe Noun it is relatedto:
position kctrct
(r2) N Postposition: Taro ga I rnado kara I inu o nlita
N0 M windorv fi orn clog A(r( silw
'Taro saw a doq l'ronr the rvindow.'
Auxiliarics in Japetnese follow the main Verb:
also
( 13) V ALIX: Taro-ga itru tl Imiru bekida ]
- N0M dog A C C : s e e shoulcl
'TAr0 shoulclsee a dog.'
g Here is a plausible tree diagram of a Japaneseclause with a postpositional
6 phrase and an inflectional element (something like an auxiliary) following the
€ verb:
o
x
a (14) s
s3 DP IP
=o
E+ ---
^- ^
l= NP D vP I
?s
booo
I I
c.r-\
csf-. N
r\ I
| PP DP V
F-o,
5 o :. , I
lll.^\/,,^\
azll
Sfi | DP P NP D
bo3llAll
e ' i l l - -l -r.
E$IINP?IT
g5 ll+
F g Taro ga mado
lll
0 kara inu o miru -bekida
dfi NOM window from dog ACC see should
,E"'f, 'Taro shouldseea dog from the window.'
E9
examples Malagasy,language in many
! $ Nowwe will lookat some fiom a that
[$spects exhibitssyntactic that is the "mirror image"of Japanese.
structure
F+ t" Malagasy verb normallycomesfirst in the clause.
thc Then comesthe O,
6,$nct (exarnples
trnallythe subject courtesy Andoveloniaina
of Rasolofb):
(1s) VO: I Nahita alika ] Rashu. 'Rasoitsaw a ciog.
saw dos Rasoa
Examples an OV and a VO language
of r97
In the noun phrase, GENitive(possessor;
the follows thepossessed
Noun:
(16) N GEN: Nahita Ini alika n'ilai rangahi Rashu
]
saw the dog the man Rasoa
'Rasoa saw the man's dog.'
Nouns:
followtheirhead
Also,ADJectives
(17) N ADJ: Nahita t;::, 'Rasoa a bigdog.'
saw
n, ff::;
in the
comebefore Verb:
Auxiliaries Malagasy
(18) AUX V: I afaka nlaita] alika be Rashu 'Rasoacall seea big dog.'
Again, Malagasy is exactly the opposite of Japanese placing relative clauses
in
afier their head Nouns. In this example, the head noun is alika,'dog,' and the
Relative Clause which modifies it follows:
(1e) N RC: Nahit a ilai I ali k a { n i h i n a n a i l a i hena } ] Rashu
silw the dog ats the lxeat Rastla
'Rasoa saw the clog that ate the nteat.'
Finally, Malagasy exhibits PREpositions rather than postpositions:
(.)
(20) PREP It{: Nahita alika I avi varavaranakely I Rashu
_a
saw dog through door slnall Rasoa
a.)
'Rasoa saw a dog tirrough the window.'
i
=
2 Thus we see that Malagasy exhibits exactly the mirror image of Japanesein
g $rms of basic syntactic structure.Here is a possible tree diagram of a Malagasy
! $tause that illustrates all of the phrasal structureswe have mentioned:
tr*
tI
s.Hz
(€ t--
50\
bo:. IP DP
=z
.im /,.A\
3C3 I VP DNP
:
E-S !,
|
I -
rr r \Y\ T\
I
E: I V DP PP N
85 | | D NP P DP
Rs | | | A -----------=-
--.^---
NA Dr D NP
sS | | | /\ =^-
EA
* "E
Eg
;a | | | | | l\
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| I
/
l\
t\
\r |
|
|
|
N
'n-}-',
r
|
A Dr
I
|
E5
E/
afaka maita 0 alika be avi eu am 0 varavarana kely 0 Rashu
i-e
OE
can see INDEF dog big through INDEF door small
'Rasoacan seea big dog througha window (lit. smalldoor).'
4=
FE
- " Notice that in this diagram the branchesunder the IP node extend to the rjght,
ratherthan the left.In this respect.Malagasycan be considereda"right-hranching"
Ianguage.
19 8 LANGUAGE TYPOLOGY
It is importantto emphasize factsaboutthe pioneeringwork of Greenberg.
two
First, most languages not as consistent
are with Greenberg'sgeneralfindings
as Japanese and Malagasyare. Second,Greenberg's samplewas quite inad-
equatein a number of respects. More recentwork (e.g., Dryer 1988, 1992)
has shownthat someof Greenberg's observationssimply cannotbe sustained
when a larger,more statistically sampleof languages takeninto
significant, is
account.
Pragmaticconstituent order languages
While the work of Greenberg (1963) and other early typologists was
instrumentalin establishingthe {ields of typological and functional linguistics,
there were several conceptual problems with this early work. In pafticular. one
of the assumptions was that all languagesemploy constituent order to express
grammatical relations. In other words, the very use of the terms SVO, SOV, and
others to describe languagetypes takes it for grantedthat "Subject" and "Object"
are the relevant terms that determine the order of words in languages.This does
hold true for English and many other languages, but is it necessarily true for all?
Could it be the case that a language may use the order of words in a clause to
€ express some other communicationally important information? What would such
6 a languagebe like?
-fr As a thought experiment, imagine a languagethat uses word order to express
! relative size rather than grammatical relations. {n this language, when a clause
X clescribestwo participants interacting, the larger participant is mentionetl first.
! Hnd the smaller one is mentioned later, regardlessof which one is subject and
!$t i"t't is object. Therefbre, a clause like the following would be ambiguous:
#;
HYPOTHETICAL DATA: Rearkill man.
$fizzl
EZ
gy putting beur carly in the clause,the speakeris assertingonly that the bear is
*E
E &tg"r than the man, not whether the bear is the subject or the object of the clause.
lr sentence could conceivablymean 'The bearkills the man' or 'The man kills
$St
g$he bear.'
i s fnis kind of ambiguity is not often toleratedin languages, becauseit is dysf'unc-
i {ionat - expressingwho acts and who or what is acted upon is such an important
E tg'ommunicativetask that it is not likely to be ignored by the morphosyntax of
E.$ny nurr..ul language.Therefore, if the languageis cleterminecl use worcl order
to
E|1l Oistinguish size, it would probably come up with some other solution to the
P$roblem of expressingwhich par-ticipant the actor, and which one is the aflected
is
e€
a Eanlclpant.
- " How about a morphological solution? This is the kind of solution illustrated in
chapter I with the Naga language:
Pragmatic order languages
constituent r99
HYPOTHtrTICAL DATA.
(23) a. Bear-akill mart. ' Bear k i l l s m a n . ' (A su{'fix-a nrarks the AGENT.)
b. Bear kill man-it. ' M an k i l l s b e a r . '
(2$ a. Bear k ill m an- p. ' Bear k i l l s l r A n . ' (A suffix -p marks the PATIENT.)
b. Bear-p kill nran. 'Man lrills bear.'
(25) a- B66r k ill r nhn. ' Bear k i l l s m a n . ' (Lligh tone marks AGENT; lou'
tone mitrks PATIENT.)
b. Bdirr kill mirn. 'Man kills beerr.'
The examples in )3, 1,4,and 25 illustrate three possible solutions to the prob-
lem of how to express the AGENT and PATIENT if word order is used for
some other purpose. Of course, any of the morphological, Iexical, or syntac-
tic expression types discussedin chapters I and 2 may be used. These are just
random possibilities, out of an infinite number. We will have a lot more to say
abouthow languagesactually do expressgrammatical relationsand semanticroles
in chaptersS and 9.
As you may suspectby now, there are in fact languagesthat use word order
for purposes other than to express grammatical relations. There are probably
none that use word order to express relative size, though such a language
is conceivable. What is much more common are languages that use the iin-
f ear order of words in clauses to express PRAGMATIc information, such as
I r onx r rn rAB rl rry , T o p rc A I-rry , R E FE R E N TIA LTTv emdothers.In thi s
I section we will give a few examples from such languages, and then present
j *o*" suggestions for how to analyze pragmatically based constituent order
1 Jpnguages.
(Hcath 1978:206 as cited in Mithun 1987), constituent orrler is
f;€ t" Ngandi
! 4overned by the following principle;
a6
New.indefinite, otherwise
or "newsworthy" informationis placedearlyin
fl&el
€'= the clause. alreadyintroduced
Given,delinite,or otherwise informationis
Placed later.
Eq
E_&
FE tn example from Ngandi,we seesubject-verb
I' refers
orderwhenthe subject
i,Eo a non-specific.
newlymentioned participant()?ll), andverb-subject
orderwhen
R $re sublect item (t?ti):
refersto a speciliciclentified
c.i I
F S27) a. Subjec t Verb
.
-q( )-a gu-jark-yul .
gu-ja-waik. . .
Naduwelefl-u I
E>
t hen- n n s C; LJ.WAIET.A I} S c u -now-gti:through
F c,) 'Then \ /ater passes thrclugh,'(lirst lnention of water)
^bo
0)E
SEb Verb Subject
a
\J l/
- '\
NaIuwelefr-un gu-ja-geyk-da-ni gu-jark-yutl
then-ans -now-throu'-AUG-llR G tJ-wiltel-n ns
G LJ
'Then fhe rvaterrushesthrough,' (subsequent
mention of water)
200 LANGUAGE TYPOLOGY
Coos (Frachtenberg l9l37) also follows this "indefinite early" principle. In
example ltr, the matting is the aff'ectedpafiicipant, and it comes before the verb
when it is mentioned for the first timc. In liib, the matting is again the affected
participant, but this time it comes after the verb, becausein this clausethe matting
has already been mentioned in the discourse,and thereforecan be taken as G IvEN
I N F O R M AT ION:
(28)a. Ob ject Subjec't Verb
TE tc ! i 'cil yiil- is y'tirqat. . .
that nratting we two split:it
'[-et's split this mat.' ({irst mention of tnat)
(they did so, and wellt dorn'n examine the earth. The earth was still not
to
s olid, ev en. . . )
b. Verb Object
i lau tci ux h i 't o u t s hE, tclicil.
when that there they:t\,'o:put:it:down the tnattil'lg
'af-Ler mention of mat)
thev had put down the lnat.' (subsequent
In both of theselanguages,thep
d t
Ctr I
ct ^to-)
E'=
(J
Pavneund McttthewI)ryer
F-r
t 0 t?t
E{ r
1 q-Sagua a LANG uAG E rsoLATE spokenby about4,000peoplein northeastern
q)
is
s\Eeru.
Ctr r-
AF
l+l !-
R
6\ r'\
-4. On the basistlf the exampleson the following page, identify whether
Yagua is basically an O-V or a V-O language., whether it is S-V
and
or V-S. Give the evidencef'or your clainrs.
Exercises 205
B. List the other constituentorder characteristics Yagua illustratedin
of
these data. For each chatacteristic,indicate whether it is expected,
unexpected, neither,given the basic order of V and O identifred
or
above.Cite examplesthat illustrateeach of your clainrs.
| . a. Sa-munaa-dee Alchico. 'Alchictl's piacenta.
3sc;-placenta-DM
I Alchic'tl
.Alcliico's placelltll.'
b. Alchico nlunaadec'.
n'Munaadee *Alchico satnutraadee.
Alchico,
c. Samunaadee His place'llta.'
?. Jirya rnunaadee. 'This placenta.'
uMunaadee
iirya.
3. Tinkii munaadee One placenta.
'*Munaadeetinkii.
4. Sarnunaadee kriritya. 'His placetrtarn''hispers.'
''K riittya santunaaclee .
5. Sakriftya Alchlco munaadee 'Alchico's plitcentawhispers.
6. J iryoonfr s$$y-anir sa-roori-myf Alch ico-nii
bus hr nas t er bit e- p A s 'l 3 s c ; - h o u s e - L 0 ( ' A l c h i c o - 3 s c ;
'A bushmaster (snake)bit hirn in Alchico's house.'
0,) '*Jiryoonfisasr,lpyanunii. bushnrastehe bit hirn.')
(.A r
'=
7. Sa-sltgy-anu jfryoonu Alchico roori-mvu-nii
-a
(.) 3sc -bite-pA S'l' bushmaster-3s Alchico house-t.tl
c C'-3s c;
'A bushmaster (snake)bit hirn in Alchico's house.'
a
.. c\ 8. srr-rr rqt?-kyu He rvill jurnP ! '
s-tX
3sc -rru'l' jump-r,t"l"1
E .f, '*r r'sar *sarqrirkyu
{gkyu silir, irirkyu sgle, $.
a u.)
booo 9. Sa-niy,' suv$-tyirajinu rnunftya sy-Llrxutggsfr
cS t--
= C} .
bo. .
3sc;- M A t-I-' f ear-tN't's this 3s
ancesttlr c;-behincl
=z
'\ rn
nrunirii jftiy sa-rpp-flii.
ooE
placenta Rt1t, 3sc;-jump-3sc;
o J9
;.v behinclthe placentathat tnakeshirn jr-rrnp.
'This ancestoris really at'raicl
XX
H-l tr
\i,q I 0. R?- a j
.j$p-charati iyu-day koodi-r,yiimri.
c'i I s c - F Llt ' t all- r n i g h th e r e - l ) A Y snake-irrsicle
2
c.i I "I nright ttrll here insiclea snake.'
(tr .t-
a
a4 c)
=.
q.)
Exercise7.2:Tshangla
F
^oo
>-o
cg( - Eric Anclvik
r\ r'\
vv
I . Ja-ga ata yigi ringrnu thur dri-ba.
I s c;-t) A'l' eld .brother letter long one rvlite-pi\ S'l
'My elder brother wrote ont: long letter.'
206 LANGUAGE TYPOLOGY
2. Ro-ka gari otha phai yanglu jap-kai tsuk-pa cha.
'l'
3sc;-r) A't car that house green behind-ABl. Put-t'ns is
'His car is parked behind that gree house.'
n
3. Ja-ga usin-git chhartl nan-gi ye-khan
ls c ; - DA' t ' y oung. s is - t r n 'l ' {r i e n d ? s c ; *l r R G s p c a k - n n t -
echha ngo-le re-be.
book buy - NPAs ' t ' c t t n - N P A S I '
'My vounger sister's frienclcan buy the book that you talked about.'
A. Where is Tshangla spoken I What is its genetic af{iliation? How many
speakers ale there'l
B. Describt: the head and complernent orders in the various syntactic
constituents illustrated. In what respects do tl-resedata conform to
Greenberg's (1963/1966) observations, and in what respects do they
not conforrn? Note any arnbiguous or problernatic clata.
7.5:
Exercise Shugnan
M. E. Alexeev, udapted bY Tom PaYne
Here are some noun phrasesin Shugnanand their translationsinto English:
1. kuzaahats 'jar of water'
'i
-Z 2. chalakzimaadi 'bucketof dirt'
€ 3. tambalbyuyun 'beardof a lazybones'
S 4. biig dyuyunaa 'pot ofcorn'
1 5. kuzaa gjev 'lid ofajar'
g^E 6. beechoraa zimaadj 'clirtof a beggar'
Er
#e. What languagefamily does Shugnanbelong to, and where is it
sPoken?
3"f
dE. into
Translate Shugnan:
ooP
'E oi 7 . ' b u cke t o f wa te r '
ogJ
EX 8. 'corn bcggar'
of'a
3F 9. 'jar of a lazybones'
o-
] P What determines the order of genitive and head noun in Shugnan?
cx
a
-g s )
ES Exercise translatingkees into bracketing
7.4:
a80
$E TomPayne
6$tt" structure
following are constituent treesfrom threetypologicallydistinct
languages.
A. "Translate"eachtree into a labeledand bracketedstring.