Transcript Document

Never too much of a good thing:
why so many languages have
more than one passive
Anna Siewierska & Dik Bakker
Lancaster University
The canonical construction
i) Agent defocusing & Patient promotion
ii) Semantic properties:


(a) Semantic valence: predicate (agent, patient)
(b) Subject is affected
iii) Syntactic properties


(c ) Encoding: agent  oblique or not expressed
patient  subject
(d) Valence of predicate: Active = P/n
Passive = P/n - 1
iv) Morphological properties
 Active = Predicate
 Passive = Predicate [+passive]
v) Pragmatically marked vis a vis the active
Multiple Passives
2
Skewed areal and genetic distribution
Multiple Passives
3
Language internal distribution

Enormous differences in use of
passive

Productivity of construction
Any transitive verb
 Only one verb: Skou (Donahue 2004)



Genre & register factors
Number and variety of passive
constructions
Multiple Passives
4
What counts as a distinct passive?

A construction-based approach


Any systematic difference in form, semantics
or use – too wide
 Agentive vs. agentless
 Different forms of agent marking
 Presence vs. absence of thematic subject
 Different thematic role of subject
This paper: distinctive verbal morphology



Different auxiliary or light verb
Different affixation; not evident allomorphs
Different form of lexical verb in periphrastic
Multiple Passives
5
English be vs. get passives

Same form of lexical verb different
aux


She got fired by her boss.
She was fired by her boss.
Multiple Passives
6
Kachiquel: suffixal & prefixal passive
a. Xta Maria x-tz’ub’-äx
r-oma’ a Juan
CL Maria COM-kiss-PASS 3-by
CL Juan
b. Xta Maria x- ki-tz’ub’-aj r-oma’ a Juan
CL Maria COM-PASS-kiss-TR 3-by
CL Juan
‘Maria was kissed by Juan.’
From
3pl to
passive
Multiple
Passives
7
Ukrainian ne/te vs. no/to passive

Same aux different form lexical verb
nemovlja bulo znajdene/znajdeno
u košyku
baby
AUX found:N:SG/found:N:SG in basket
`A baby was found in a basket.’
Multiple Passives
8
Russian periph. vs. synthetic passive
a. Takie stat’i
ne
such articles:NOM NEG
byli
opublikovany
were:PL published:PL
za
granicej
beyond border
`Such articles were not published abroad.’
b. V Rossii prodolzalo ne proizvodit’=sja
in Russia continued NEG produce=PASS
takix napitkov
such drinks:GEN
`There continued not to be any such drinks produced
in Russia.’
Multiple Passives
9
Current sample: 311 lgs with passives
90
80
70
Africa
60
Eurasia
50
Namer
40
Samer
30
SEA&Oc
20
AusNG
10
0
pass
Multiple Passives
10
more than 1 passive = 92/311 (30%)
pass
250
200
150
pass
100
50
0
1
2
3
Multiple Passives
4
4+
11
Areal distribution of 1 & 1+ passive
70
60
Africa
50
Eurasia
40
NA
30
SA
SEAOc
20
AusNG
10
0
1
pass
1+ pass
Multiple Passives
12
Areal distribution of 1 vs. 2 vs. 2+
70
60
Africa
50
Eurasia
40
Namer
30
Samer
SEA&Oc
20
AusNG
10
0
1 pass =219
2 pass = 52
2+ pass = 40
Verbal marking & 1 & 1+ passive
200
180
160
140
120
periph
100
80
both
synth
noVmarking
60
40
20
0
1+pass
1 pass
Multiple Passives
14
Why more than one passive?




No iconic or economic motivation for
structural synonymy within languages
(Croft 2003:105-106)
Some differences in function
What type of differences in function?
How are these differences distributed
globally and in relation to type of passive
verbal marking?
Multiple Passives
15
Type of differences between passives

Main constituents




Semantic factors




Subject
Agent
Verb
TAM
Nature of state of affairs
Assessment
Pragmatic factors


Information structure
Text type & genre
Multiple Passives
16
Subject




Presence
Semantic role
Animacy
Person
Multiple Passives
17
Subject: Presence: Kannada



(Sridhar 1980) padu `experience’ vs. agu
`become’
Krishnanu-indu Ramu-Ø
ko-pattu-nu
Krishna-INST Rama-NOM kill-PASS-PAST
`Rama was killed by Krishna.’
Rama-nannu kollal-ayi-yu
Ram-ACC
kill:INF-PASS-PAST
`Ram was killed.’
Multiple Passives
18
Subject: Semantic role: Dutch



Het boek wordt hem toegestuurd
the book become him sent
`The book is sent to him.’
*Hij wordt het boek toegestuurd.
Hij krijgt het boek toegestuurd.
he gets the book sent
`He is sent the book’
Multiple Passives
19
Subject: Animacy: Imbabura Quechua




Maria-ka juya-shka
ka-rka
Maria-TOP love-PAST.PART be-PAST:3
`Maria was seen.’
Maria-ka juya-y
tuku-rka
Maria-TOP love-INF
become-PAST:3
`Maria was seen.’
Aycha-ka miku-shka
ka-rka
meat-TOP eat-PAST.PART be-PAST:3
`The meat was eaten.’
*Aycha-ka miku-y tuku-rka
Multiple Passives
20
Subject: Person: Coeur d’Alene



–m (1SG or 3) vs. –t
Čε√lεq-n-t-Ø-m
LOC.burr—d-t-3SG-PASS
`He was buried.’
√cun√mεy-n-t-εli-t
point.know-d-t-1PAT-PASS
`We were taught.’
Multiple Passives
21
Agent




Possibility of agent expression
Obligatoriness of agent
Nature of covert agent; human vs.
any
Person of agent: non SAP vs. any
Multiple Passives
22
Agent: Presence: Buru (Grimes 1991)


Sira
dapa-k
eflali (*ringe)
they
get-ACP
beat (*him)
‘They got beaten up.’
Subu
di
ek-fuka-k
ringe
door
DIST PASS-open-ACP 3SG
‘The door was opened by him.’
Multiple Passives
23
Agent: Omissability: Mandarin


Jin-yu (bei) xiao-mao chi-diao le
goldfish PASS cat
eat-up PERF
`The gold fish has been eaten by the cat.’
Jin-yu (*rang) xiao-mao chi-diao le
goldfish PASS cat
eat-up PERF
`The gold fish has been eaten by the cat.’
Multiple Passives
24
Agent: any vs. human: Lunda

Kasumbi
Ø-na-sum-ew-j
domestic fowl S/A-TAM-bite-PASS-FV
`A domestic fowl has been bitten(e.g. by a
snake.)’
a-na-sum-j
kasumbi
PASS-TAM-bite-FV domestic fowl
` A domestic fowl has been bitten (by a
person not by a snake)

Multiple Passives
25
Agent: Person: Quiche




(Campbell 2000: 249)
š-in-čay-tax
aw-umal
ASP-1SG-hit-PASS 2SG.POSS-by
`I got hit by you.’
*š-in-ča:y
aw-umal
ASP-1SG-hit:PASS 2SG.POSS-by
`I got hit by you.’
*š-in-ča:y
r-umal
ri ačih
ASP-1SG-hit:PASS 3SG.POSS-by the man
`I got hit by the man.’
Multiple Passives
26
Main constituents: Verb

Lexical

morphological: derived vs. underived

valency
Multiple Passives
27
Verb: Lexical: Tzutujil


A few verbs take the archaic passive
marker vr, k’amo `carry’, tojooj `to
pay’, chapooj `to grab, scold’
All other verbs can take the simple
passive with the infix –j- or suffix –x or
the completive passive in -taj
Multiple Passives
28
Verb: Morphological: Older Egyptian





(Reintges 2008)
1. Internal passive ij
2. Suffixal passive –w
3. Reduplication applied to verbal root
Passive 1 & 2 can apply to any verb.
Passive 3 can only apply to transitive
verbs that are not derived
Multiple Passives
29
Valency: Krongo




(Reh 1985:228-9)
-atini (TR) vs. –aca (DTR)
N-apa-atiŋ
a?aŋ
½-PRF:hit-PASS I
`I have been hit.’
K-ada-aca
nan-kurusi a-kaaw
PL-PRF:give-PASS PL-money DAT-person
`The money has been given to the man.’
Multiple Passives
30
Semantics


TAM:
Nature of state of affairs




generic vs. episodic
stative vs. dynamic
accidental vs. volitional
Assessment


adversative vs. neutral
beneficial vs. adversative
Multiple Passives
31
Aspect: Polish być vs. zostać


Ciasto było pieczone / upieczone
cake was baked:IMPERF / baked:PERF
przez ciocię Helę
by
aunt Helen
`The cake was baked by aunty Helen.’
Ciasto zostało *pieczone / upieczone
cake remain baked:IMPERF / baked:PERF
przez ciocię Helę
by aunt Helen
Multiple Passives
32
Modality: Deontic

Gujarati


synthetic =a (possibility, necessity,
prohibition) vs. periphrastic =aa `come’
(non-modal)
Sardinian

Periphrastic-essere `be’ (non-modal) vs.
periphrastic kerrere `need’ & bollit ‘want’
(modal)
Multiple Passives
33
Gujarati -a vs. aa




Possibility (participant internal)
Chokr=aa=thii purU bol=a=y=U
nahi
boy=OBL=ABL fully say=PASS=P.PRT=N NEG
`The boy could not express himself.’
Non-modal
Prasav thay=aa=nii
bhaaii=o=ne jaaN
delivery being of brothers
information
kar=v=aa=mAA
aav=ii
do=INF=OBL=in
come=P.PL=FSG
`The brother’s were informed about the delivery.’
Multiple Passives
34
Sardinian éssere vs. kérrere vs. ?bole




Custa domo est istata fraicata dae un'Italianu
'This house was built by an Italian.'
Cussas fainas keren fattas de nos corcare
`Those chores need to be done before we go
to bed.’
Custa cicara de cafei bollit buffada
`This cup of coffee must be drunk.’
Multiple Passives
35
Modality: epistemic: Cayuvava

(Key1967:28)

Synthetic: ada-: certainty
ada-kæčæ `it is cut off.’
 mera-ada-boro `it will be given with
certainty’


Synthetic: bae-: probability
a-bæ-boroæ `it is given, probably’
 a-bæ-kočiro ‘if I am believed, I am
probably believed’

Multiple Passives
36
Generic vs. Episodic: Norwegian



–s vs. bli (Faarlund 1997:514)
Oppgavene levere-s
hver uke
the.exercises hand.in-pass every week
`The exercises are handed in every week.’
Oppgavene
ble
levert for seint
the.exercises become handed too late
`The exercises were handed in too late’
Multiple Passives
37
Stative vs. Dynamic: Estonian



olema `be’ + -tud part. vs. saama `get’ +
infinitive in -da (Torn-Leesik 2002:11-13)
Maja
projekt
oli
valmistatud
house:GEN plan:NOM be:PAST:3SG
make
Jürissoni poolt
Jürisson by
`The design plans of the house were drawn up
by Mr Jürisson.’
Patsiendid
said
arsti
käest noomida
patients:NOM get:3PL doctor:GEN from
reprimand:INF
`The patients got reprimanded by the doctor.’
Multiple Passives
38
Volitional vs. Accidental: Maanyan



(Gudai 1985:226)
Kawaweq yeruq na-jalak
daya-ni
deer
the
PASS-spear by he
`The deer was speared
by him.’
Bukuku ta-rakit
daya anak-ni
book-1SG PASS-burn by
son-3SG
`My book has been accidentally burnt by
his son.’
Multiple Passives
39
Adversative vs. Beneficial: Vietnamese



bi vs. duoc
Anh ay bi
nguoita danh
he
PASS someone hit
`He was hit by someone.’
Anh ay duoc Kim danh
he
PASS Kim hit
`He was hit by Kim (fortunately, he liked it)’
Multiple Passives
40
Adversative vs. Neutral: Indonesian



(Ing Djiang 1988:49) kena`be hit’ vs. diAli kena tipu
seorang dukun
Ali be hit deceive one-CLAS medicine man
`Ali was deceived by a medicine man.
Buku ini di-beli
oleh nya di Athenaeum
book this PASS-bought by him at Athanaeum
`The book was bought by him at the Athanaeum
bookshop.’
Multiple Passives
41
Pragmatics

Information structure

Register and genre
Multiple Passives
42
Information structure: Kakchiquel


(Broadwell & Duncan 2002)
In the ki-passive in contrast to the
standard äx–passive both Agent and
Patient convey given information
Multiple Passives
43
Register: Southern Min



(Matthews & Yip 2001:269)
Ngo seng4jat6 bei2 jan4
aak
I
always
PASS people cheat
`I keep being cheated.’
Jau5 saam bei6 hin2-faan2 loi6dei6
exist three PASS send-back mainland
`Three were sent back to the mainland.’
Multiple Passives
44
Distribution of differences

Areal

Type passive marking
Multiple Passives
45
Area & semantic differences
20
18
16
Sub
14
12
Agent
10
8
TAM
Verb
Event
6
4
Sem Aff
2
0
Africa Eurasia SEAOc AusNG NAmer SAmer
Multiple Passives
46
Area & semantic differences
20
18
16
Sub
14
12
Agent
10
8
TAM
Verb
Event
6
4
Sem Aff
2
0
Africa Eurasia SEAOc AusNG NAmer SAmer
Multiple Passives
47
Type verbal marking

Periphrastic

Synthetic

Both
Multiple Passives
48
V marking by area
25
20
periph
15
synth
10
both
5
0
Africa
Eurasia
SEAOc
AusNG
Multiple Passives
NAmer
SAmer
49
V marking and type of differences

In principle choices between
passives motivated by any of the
factors considered should involve
any type of passive: synthetic,
periphrastic or both
Multiple Passives
50
V marking & semantic differences
18
16
14
Sub
12
Agent
10
Verb
8
TAM
6
Event
4
Sem Aff
2
0
periph
synth
Multiple Passives
both
51
V marking & semantic differences
18
16
14
Sub
12
Agent
10
Verb
8
TAM
6
Event
4
Sem Aff
2
0
periph
synth
Multiple Passives
both
52
The major asymmetries




Adversative differences favour
distinctions between periphrastic
passives
TAM differences disfavour
distinctions between periphrastic
Verb distinctions favour synthetic
Agent differences disfavour
distinctions between periphrastic
Multiple Passives
53
Adversative semantics

Lgs with more than one passive





SEA&Oc = periph
Eurasia = periph
Africa: Coptic= periph
NAmerica: Central Yupik = synth
 Adverse in a general sense (inanimate subject)
 (?passive or resultative) -cir vs. -ma
Lgs with one passive

Japanese, Tungusic (Evenki, Even), Manchu =
synth
Multiple Passives
54
TAM

Synthetic vs. Synthetic


NAmerica: MesoAmer: Mayan, UtoAztecan
Periphrastic vs. Synthetic

Eurasia: Europe
Multiple Passives
55
Periph vs. synth in Europe

Periphrastic






perfective
specific event
subject not restricted
overt agent
higher register
Synthetic





imperfective
generic event
inanimate subject
no agent
lower register or neutral
Multiple Passives
56
Periph & synth outside Europe

Only 6 in the sample




TAM: not involved
Event: not involved
Subject: necessarily inanimate in the
more periphrastic in Cubeo
Agent: no agent in the periphrastic in
Coptic, Cubeo, Buru

Purepcha has calqued a periphrastic
passive on Spanish and introduced an
agent
Multiple Passives
57
TAM and periphrastic


No lgs where synthetic passive is
perfective and periphrastic
imperfective
Modal distinctions involving
capability, necessity and possibility
favour periphrastic
Multiple Passives
58
Person and synthetic

Restrictions involving person
relating to either the subject or the
agent favour synthetic passives


Subject: Interior Salish, Quileute,
Moseten, Ostyak
Agent: Mayan, Bantu, Uto-Aztecan
Multiple Passives
59
Conclusions



Contrary to what was stated in the abstract,
only about 1/3 of the languages with
passives have more than 1 (in terms of the
measure adopted here)
Lgs with more than one passive are mainly
in Eurasia, NAmerica (Meso); SEA&Oc
Lgs with periphrastic passives appear to be
more likely to have more than one passive
than languages with synthetic passives
Multiple Passives
60
Conclusions


The type of language-internal differences
displayed by passives are areally skewed
There are also asymmetries relating to
type of verb-marking of the passive



Adversative differences favour distinctions
between periphrastic passives
TAM differences disfavour distinctions between
periphrastic
Person differences favour distinctions between
synthetic
Multiple Passives
61
The wider context

Why some languages favour
multiple passives, while others do
not, and yet others have no
passives at all?
Multiple Passives
62