Language and modali y:
E ec s o he use o space in he
ag eemen sys em o lengua de signos
española (Spanish Sign Language)
LANGUAGE AND MODALITY:
EFFECTS OF THE USE OF SPACE IN THE AGREEMENT
SYSTEM OF LENGUA DE SIGNOS ESPAÑOLA
(SPANISH SIGN LANGUAGE)
ACADEMISCH PROEFSCHRIFT
e e k ijging an de g aad an doc o
aan de Uni e si ei an Ams e dam
op gezag an de Rec o Magni icus
p o . d . D.C. an den Boom
en o e s aan an een doo he College oo P omo ies inges elde
commissie, in he openbaa e e dedigen in de Agnie enkapel
op ijdag 29 janua i 2016, e 10:00 uu
doo
BRENDAN DENIS NICHOLAS COSTELLO
gebo en e Manches e , Ve enigd Konink ijk
(cc)2016)BRENDAN DENIS NICHOLAS (cc by-nc-sa 4.0)
P omo iecommissie:
P omo o : p o . d . A.E. Bake Uni e si ei an Ams e dam
Cop omo o es: d . R. P au Uni e si ei an Ams e dam
d . M. A. Landa A e alillo Uni e sidad del País Vasco/
Euskal He iko Unibe si a ea
O e ige leden: p o . d . P.C. Henge eld Uni e si ei an Ams e dam
p o . d . E.O. Aboh Uni e si ei an Ams e dam
p o . d . J. Que Villanue a Uni e si a Pompeu Fab a
p o . d . M. S einbach Geo g-Augus -Uni e si ä
Gö ingen
d . V. Kimmelman Uni e si ei an Ams e dam
Facul ei de Gees eswe enschappen
This doc o al hesis was ca ied ou in co u elle be ween he Uni e si y o
Ams e dam and he Uni e si y o he Basque Coun y.
The esea ch o his doc o al hesis ecei ed inancial assis ance om a
doc o al schola ship om he Basque Go e nmen ( e . BFI05.149).
Do m'a hai ,
do mo mhá hai ,
do mo dhei iú .
I ha e been o hell and back. And le me ell you, i was wonde ul.
Louise Bo geois
Table o con en s
ACKNOWLEDGEMENTS XI
ABBREVIATIONS OF SIGN LANGUAGE NAMES XVII
NOTATION CONVENTIONS XIX
1. INTRODUCTION 1
1.1. LANGUAGE AND MODALITY 2
1.1.1. Simul anei y in sign languages 4
1.1.2. Iconici y in sign languages 10
1.2. THE USE OF SPACE IN SIGN LANGUAGES 17
1.3. THE STUDY OF VERBAL AGREEMENT IN SIGN LANGUAGES 23
1.4. LENGUA DE SIGNOS ESPAÑOLA (LSE) 25
1.4.1. LSE: his o ical backg ound 27
1.4.2. LSE: sociolinguis ic se ing 29
1.4.3. P e ious esea ch on LSE 31
1.4.4. The LSE in his s udy 32
1.5. THE GOALS OF THIS THESIS 33
1.6. THE STRUCTURE OF THIS THESIS 34
2. THEORIES OF AGREEMENT 37
2.1. TWO APPROACHES TO AGREEMENT 38
2.2. TYPOLOGICAL APPROACH 42
2.2.1. Te minology 42
2.2.2. Con olle s 43
2.2.3. Ta ge s 45
2.2.4. Domains 56
2.2.5. Fea u es and alues 62
2.2.6. Condi ions 72
2.2.7. Canonici y 74
2.2.8. Summa y 77
2.3. THE MINIMALIST PROGRAM 78
2.3.1. Gene a i ism: issues and de elopmen s 79
2.3.2. The a chi ec u e o he language acul y 85
2.3.3. Ag eemen and Ag ee 88
2.4. SUMMARY 94
3. AGREEMENT IN SIGN LANGUAGES 97
3.1. PRONOMINAL REFERENCE 98
3.1.1. Loca ion assignmen 99
3.1.2. Role shi 101
3.1.3. R-locus and space 103
3.2. AGREEING VERBS 107
3.2.1. P o o ypical ag eeing e bs 109
3.2.2. Backwa ds ag eeing e bs 119
3.2.3. Single a gumen ag eemen 127
3.2.4. Summa y 129
3.3. AGREEMENT AUXILIARIES 130
3.3.1. AUX 131
3.3.2. Auxilia ies de i ed om lexical e bs 134
3.3.3. PAM 137
3.3.4. Issue a ising: wha ag eemen auxilia ies ell us abou ag eemen 140
3.4. NON-MANUAL AGREEMENT 142
3.4.1. Head il and eye gaze as ma ke s o subjec and objec ag eemen 143
3.4.2. Non-manual ag eemen in ole shi 147
3.4.3. Summa y 149
3.5. DP-INTERNAL AGREEMENT 150
3.6. SUMMARY 153
4. METHODOLOGY 157
4.1. METHODOLOGICAL CHALLENGES: THE ELUSIVE NATIVE SIGNER 157
4.2. INFORMANTS 160
4.3. DATA COLLECTION AND MATERIALS 163
4.4. TRANSCRIPTION 165
4.5. DATA ANALYSIS 167
4.6. SUMMARY 167
5. AGREEMENT PHENOMENA IN LSE 169
5.1. PRONOMINAL REFERENCE 170
5.1.1. Loca ion assignmen in LSE 170
5.1.2. Role shi in LSE 175
5.2. AGREEING VERBS 179
5.2.1. P o o ypical ag eeing e bs 180
5.2.2. Backwa d ag eeing e bs 183
5.2.3. Single a gumen ag eemen 184
5.3. AGREEMENT AUXILIARIES 189
5.3.1. AUX 190
5.3.2. Auxilia ies de i ed om lexical e bs: GIVE-AUX and BEAT-AUX 194
5.3.3. PERS 197
5.3.4. Summa y 200
5.4. CONSTRAINTS ON VERBAL AGREEMENT 200
5.4.1. Seman ic cons ain s on ag eeing e bs 201
5.4.2. Phonological cons ain s on ag eeing e bs 203
5.5. NON-MANUAL AGREEMENT 210
5.6. DP-INTERNAL AGREEMENT 213
5.7. DISCUSSION AND CONCLUSIONS 221
6. LSE AGREEMENT FROM A CROSS-MODAL TYPOLOGICAL PERSPECTIVE 227
6.1. CONTROLLERS 228
6.2. TARGETS 229
6.2.1. Ve bs and auxilia ies 229
6.2.2. O he a ge s o ag eemen 231
6.2.3. Means o exponence 231
6.2.4. Mul iple exponence 238
6.2.5. Summa y 238
6.3. DOMAINS 239
6.3.1. Clause-in e nal ag eemen 239
6.3.2. Ag eemen beyond he clause 241
6.4. FEATURES AND VALUES 242
6.4.1. Gende 243
6.4.2. Numbe 243
6.4.3. Pe son 247
6.4.4. O he ea u es: espec and case 253
6.4.5. Summa y 255
6.5. CONDITIONS 256
6.6. CANONICITY 259
6.6.1. Applying Co be ’s c i e ia o spa ial ag eemen in LSE 261
6.6.2. Applying Co be ’s gene al p inciples o spa ial ag eemen in LSE 267
6.6.3. O he e alua ions o he canonici y o sign language ag eemen 270
6.6.4. Summa y 271
6.7. DISCUSSION AND CONCLUSIONS 272
7. FORMAL ANALYSES OF AGREEMENT IN LSE 277
7.1. LOCATION, IDENTITY AND LOCATING IDENTITY 278
7.1.1. The loca ion o ϕ- ea u es 280
7.1.2. The loca ion o he iden i y ea u e 282
7.1.3. Op ionali y o he use o space 286
7.2. ACCOUNTING FOR SPATIAL AGREEMENT IN LSE 287
7.2.1. Loca ion assignmen 287
7.2.2. Ve bal ag eemen 289
7.2.3. P agma ic ag eemen 294
7.2.4. Summa y 298
7.3. “DEFECTIVE” AGREEING VERBS IN LSE: AN OT ACCOUNT 300
7.3.1. “De ec i e” ag eeing e bs in LSE 301
7.3.2. OT cons ain s 302
7.3.3. Applying he cons ain s 303
7.3.4. Ex ending he analysis o ISL da a 308
7.4. ISSUES ARISING 309
7.4.1. Op ionali y ( e isi ed) 309
7.4.2. Loca i e e sus locus 310
7.4.3. Linea i y 312
7.5. CONCLUSIONS 314
8. CLOSING REMARKS 317
8.1. WHAT THIS STUDY TELLS US… 318
8.1.1. …abou LSE 318
8.1.2. …abou sign languages in gene al 320
8.1.3. …abou language 325
8.2. WHAT THIS STUDY DOES NOT TELL US 327
8.3. FUTURE DIRECTIONS 329
REFERENCES 333
LANGUAGE AND MODALITY: EFFECTS OF THE USE OF SPACE IN THE
AGREEMENT SYSTEM OF LENGUA DE SIGNOS ESPAÑOLA (SPANISH SIGN
LANGUAGE) (SUMMARY) 371
TAAL EN MODALITEIT: EFFECTEN VAN HET GEBRUIK VAN RUIMTE BINNEN HET
CONGRUENTIESYSTEEM VAN LENGUA DE SIGNOS ESPAÑOLA (SPAANSE
GEBARENTAAL) (SAMENVATTING) 377
LENGUAJE Y MODALIDAD: EFECTOS DEL USO DEL ESPACIO EN EL SISTEMA DE
CONCORDANCIA DE LA LENGUA DE SIGNOS ESPAÑOLA (RESUMEN) 383
HIZKUNTZA ETA MODALITATEA: LENGUA DE SIGNOS ESPAÑOLA (ZEINU
HIZKUNTZA ESPAINIARRA)-REN KONMUZTADURA SISTEMAK ESPAZIOAREN
ERABILERAN DITUEN EFEKTUAK (LABURPENA) 389
x ii
Abb e ia ions o sign language names
ABSL Al-Sayyid Bedouin Sign Language (Nege Dese , Is ael)
ASL Ame ican Sign Language
Auslan Aus alian Sign Language
BSL B i ish Sign Language
CSL Chinese Sign Language
DGS Deu sche Gebä densp ache (Ge man Sign Language)
DSGS Deu schschweize ische Gebä densp ache (Swiss Ge man Sign
Language)
DTS Dansk Tegnsp ok (Danish Sign Language)
ESL Es onian Sign Language
GSL G eek Sign Language
ISL Is aeli Sign Language
IPSL Indo-Pakis ani Sign Language
KK Ka a Kolok (Bali, Indonesia)
KSL Ko ean Sign Language
LIU Lugha al-Ishā a al-U dunia (Jo danian Sign Language)
Lib as Lingua B asile a de Sinais (B azilian Sign Language)
LIS Lingua i aliana dei segni (I alian Sign Language)
LSA Lengua de señas a gen ina (A gen ine Sign Language)
LSC Llengua de signes ca alana (Ca alan Sign Language)
LSCol Lengua de señas colombiana (Colombian Sign Language)
LSF Langue de signes ançaise (F ench Sign Language)
LSFB Langue de signes ançaise de Belgique (Belgian F ench Sign
Language)
LSE Lengua de signos española (Spanish Sign Language)
LSM Lengua de señas mexicana (Mexican Sign Language)
LSQ Langue des signes québécoise (Quebec Sign Language)
NGT Nede landse Geba en aal (Sign Language o he Ne he lands)
NS Nihon Shuwa (Japanese Sign Language)
NZSL New Zealand Sign Language
RSL Russian Sign Language
SSL Swedish Sign Language
TİD Tü k İşa e Dili (Tu kish Sign Language)
TSL Taiwan Sign Language
VGT Vlaamse Geba en aal (Flemish Sign Language)
xix
No a ion con en ions
The examples in di e en signed and spoken languages ollow he (Feb ua y
2008 e sion o he) Leipzig Glossing Rules (LGR), de eloped join ly by he
Depa men o Linguis ics o he Max Planck Ins i u e o E olu iona y
An h opology and by he Depa men o Linguis ics o he Uni e si y o
Leipzig.
1
Whe e examples a e ci ed om o he wo ks, he ansc ip ion has
been adap ed o con o m o he LGR as closely as possible. Fo ease o
e e ence, he lis o common abb e ia ions speci ied in he LGR is
ep oduced a he end o his sec ion. Any abb e ia ions no included in he
common LGR lis a e explained below he example in which hey appea , and
ha e been added o he lis included he e.
The sign language examples include illus a i e s ills whene e possible
and a e ansc ibed using glosses in SMALL CAPS. While I ha e ied o main ain
he con en ions and abb e ia ions o he LGR, ce ain es ablished glossing
p ac ices in he sign language li e a u e ha e been main ained as s anda d,
and a e explained below:
Hyphens a e used when mo e han one wo d is equi ed o gloss a single
sign:
LOOK-AFTER
No e ha his di e s om he LGR usage, o which a hyphen sepa a es
dis inc mo phemes.
Spa ial modi ica ion o a sign is ma ked wi h a subsc ip . The subsc ip
may indica e a loca ion in he signing space (deno ed by x, y, z o neu o
he neu al loca ion a he unma ked cen e o he signing space) o on he
signe ’s body (deno ed by 1):
GROUPx
Poin ing o index signs (glossed as IX) in a iably include a subsc ip o
indica e he di ec ion o he poin ing.
Subsc ip s a e also used o e e en ial indices, ma king co e e en iali y,
bu a e dis inguished om spa ial modi ica ion by he index used: i, j and
k (as opposed o x, y o z o loca ions in he signing space).
Fo spa ial modi ica ion in ol ing mo emen be ween wo poin s, as is
he case o ag eeing e bs, a subsc ip a he beginning o he gloss
1
A ailable on-line: www.e a.mpg.de/lingua/ esou ces/glossing- ules.php
xx
deno es he ini ial loca ion and a subsc ip a he end o he gloss deno es
he inal loca ion:
1LOOK-AFTERx
No e ha his di e s om LGR usage, which would use he “>” symbol
o an a ix ha simul aneously exp esses wo a gumen s o a e b.
Finge spelling is shown by indi idual, lowe case le e s joined by
hyphens:
o-a-k
Classi ie cons uc ions a e indica ed by CL ollowed by a desc ip ion o
he o m/meaning in pa en heses:
CL(g oup)y
(No e ha classi ie cons uc ions a e in a iably loca ed in he signing
space, so a subsc ip is included o show his.)
Rele an non-manual ea u es a e shown abo e he glosses o he signs,
wi h ho izon al lines indica ing he scope o du a ion o he non-manual
ea u e in ques ion. The abb e ia ions used o ca ego ize he non-manual
ea u es a e in lowe case and may desc ibe unc ion (e.g. “q” o a
ques ion ma ke ) o o m (e.g. “eyeb ow aise”).
In mos ci cums ances only a single gloss is gi en o he sign s eam, bu
whe e he ac i i y on each hand is ele an , he ansc ip ion includes a
line o each hand, he uppe line glossing he dominan hand and he
lowe line he non-dominan hand. When a gi en hand pe o ms a hold
(main aining a gi en sign while he o he hand con inues o p oduce
signs), a dashed line shows he du a ion o he hold:
D hand
ESTI
BOYFRIEND
COME
IXmiddle- inge
ND hand
BUOYindex
BOYFRIEND
BUOYindex+middle----------
Lis buoys a e shown by means o he gloss BUOY and include a subsc ip
o indica e which inge s (o he non-dominan hand) a e ex ended. When
he dominan hand poin s a a buoy, he IX gloss is used wi h a subsc ip
showing which inge (on he non-dominan hand) is being poin ed a .
(See examples abo e.)
Fo examples wi h mul iple signs, he ele an i ems a e highligh ed by bold
ace o he glosses, and a shaded backg ound o he ele an images.
Examples aken om eco dings made o his s udy include he name o he
eco ding ollowed by he ime poin a which he example occu s.
xxi
Common abb e ia ions o he Leipzig Glossing Rules
This lis includes he common abb e ia ions published in he Feb ua y 2008
e sion o he Leipzig Glossing Rules ha a e used in his hesis, plus any
u he abb e ia ions ha we e equi ed o he examples included
(dis inguished in bold ace).
1
i s pe son
2
second pe son
3
hi d pe son
A
agen -like a gumen o
canonical ansi i e e b
ABL
abla i e
ABS
absolu i e
ACC
accusa i e
ADJ
adjec i e
AGR
ag eemen
AOR
ao is
ART
a icle
ASP
aspec
AUX
auxilia y
BEN
bene ac i e
CL
classi ie
COMP
complemen ize
COMPL
comple i e
D hand
dominan signing hand
DAT
da i e
DEF
de ini e
DEM
demons a i e
DEP
dependen
DIR
di ec ional
DISTR
dis ibu i e
DTS
di ec heme sign
DU
dual
ERG
e ga i e
F
eminine
FUT
u u e
G
gende class
GEN
geni i e
HON
hono i ic
IIND
independen indica i e
IMPF
impe ec
INAN
inanima e
INF
in ini i e
INS
ins umen al
INV
in e se
IX
index
LOC
loca i e
M
masculine
N
neu e
N-
non-
(e.g. NSG nonsingula , NPST
nonpas )
NEG
nega ion, nega i e
NEUT
neu al loca ion in signing
space
NOM
nomina i e
OBV
ob ia i e
P
pa ien -like a gumen o
canonical ansi i e e b
PL
plu al
PM
ph ase ma ke
POSS
possessi e
PRF
pe ec
PRS
p esen
PRO
p onoun/p onominal
PROG
p og essi e
PRV
p e e bal
PST
pas
PU
palms up
Q
ques ion pa icle/ma ke
REL
ela i e
RES
esul a i e
S
single a gumen o canonical
in ansi i e e b
SBJ
subjec
SBJV
subjunc i e
SG
singula
TOP
opic
TR
ansi i e
1
1. In oduc ion
This hesis examines he na u e o he ag eemen sys em in lengua de signos
española (LSE – Spanish Sign Language). Wi hin sign language linguis ics,
e bal ag eemen s a egies a e one o he mos s udied aspec s o hese
languages. The e a e se e al easons o such in ense in e es in he opic, and
hese ha e also mo i a ed his s udy in he con ex o a speci ic sign language.
Fi s ly, e bal ag eemen in sign language displays se e al unusual
cha ac e is ics, such as es ic ions on he numbe o e bs ha show
ag eemen , a ypologically uncommon s a e o a ai s. Secondly, e bal
ag eemen in sign language makes use o s a egies ha a e ancho ed o he
isual-ges u al na u e o sign languages, and hus una ailable o spoken
languages. Fu he mo e, he basic ag eemen mechanism is e y simila (i
no iden ical) ac oss many un ela ed sign languages, b inging in o ques ion
he in luence ha he language modali y may exe on a language’s s uc u e
and o ganiza ion.
The in e ac ion o modali y and language is he o e a ching heme o
his hesis, and he esea ch has been guided by a - eaching ques ions abou
he ole sign language da a can play in ede ining ou unde s anding o
human language in gene al. These guiding p inciples a e in oduced in
sec ion 1.1 o his chap e . Sec ion 1.2 concen a es on an aspec o sign
languages ha is a s ong candida e o u ning up modali y e ec s since i is
a mechanism ha is una ailable o spoken languages: he use o space. Sign
languages use space in di e en ways, and hese a e b ie ly desc ibed be o e
limi ing he discussion o one pa icula spa ial de ice in sec ion 1.3, namely
e bal ag eemen . Sec ion 1.4 gi es a basic in oduc ion o he speci ic
language unde in es iga ion, LSE, and i s mos ele an cha ac e is ics, such
as gene al sociolinguis ic in o ma ion and i s ela ion o o he sign languages.
This sec ion also includes an o e iew o p e ious esea ch in o he language.
Sec ion 1.5 a icula es he speci ic esea ch ques ions ha p o ided he
s a ing poin o his s udy, and sec ion 1.6 concludes his chap e by gi ing
an ou line o he emaining chap e s.
2 In oduc ion
1.1. Language and modali y
This s udy ocuses on a speci ic aspec o a speci ic sign language, bu is
couched in a much b oade pe spec i e. Fi s ly, he LSE da a will be
compa ed wi h da a om o he sign languages and also spoken languages o
p o ide a ypological con ex o ag eemen in LSE. Fu he mo e, he wide
consequences o he indings o linguis ic heo y in gene al will be examined
by aking a s ep back o see he bigge pic u e.
Sign languages o e he unique oppo uni y o look a he e ec o
modali y on language:
Why should we be in e es ed in whe he speci ic aspec s o linguis ic
s uc u e migh be a ibu able o he pa icula p ope ies o he
ansmission channel? Explo a ion o modali y di e ences holds ou he
hope ha we may achie e a kind o explana ion ha is a e in
linguis ics. Speci ically, we may be able o explo e hypo heses ha his
o ha p ope y o signed o spoken language is a ibu able o he
pa icula cons ain s ha a ec ha modali y. (Meie 2002: 5)
I linguis ic esea ch limi s i sel o spoken languages and he p oposals o
he undamen al na u e o language a e based solely on spoken language
da a, i will be impossible o know whe he ecu en p ope ies e lec
gene al design cha ac e is ics o human languages o a e me ely due o he
ocal/audi o y medium. By including sign languages in he linguis ic
p og am, he a iable o modali y is in oduced and we may hope o
dis inguish co e language p ope ies om modali y e ec s.
Can we hope o ind modali y e ec s by compa ing signed and spoken
languages, o a e hey essen ially he same? Al hough he ield has been
ma ked by a endency o highligh he simila i ies be ween spoken and signed
languages – due in la ge pa o a need o socially digni y sign languages and
o jus i y hei inclusion wi hin he discipline o linguis ics (Woll 2003) –
ecen esea ch has s a ed o look o possible di e ences be ween signed and
spoken languages (Meie , Co mie & Quin o-Pozos 2002; Ve mee be gen
2006). This “sign di e en ial” iew is an a emp o s udy sign languages in
hei own e ms wi hou applying inapp op ia e concep s o imposing models
de eloped in he con ex o spoken languages. This app oach o sign language
esea ch is closely linked o he idea ha sign languages a e quali a i ely
di e en o spoken languages and ha e di e en o ganiza ion and s uc u e.
These dissimila i ies a e due o he dis inc modali ies o signed and spoken
languages and he speci ic sociolinguis ic con ex o sign languages, especially
hei ela i e you h (A ono , Mei & Sandle 2005). The modali y di e ences
may be due o he con as ing na u e o he a icula o s used o language
Language and modali y 3
p oduc ion and he pe cep ual sys ems used o language comp ehension,
and he esul ing po en ial o iconici y ha a ises om he use o space
(Meie 2002, 2012). The a icula o s employed by sign languages gi e ise o
possibili ies o simul anei y (sec ion 1.1.1) and, oge he wi h he use o space,
o a g ea e exploi a ion o iconici y (sec ion 1.1.2); he use o space i sel , he
main ocus o his hesis, will be looked a in sec ion 1.2.
The no ion ha modali y shapes language coupled wi h he obse a ion
ha many di e en (and un ela ed) sign languages ha e simila s uc u es
and make use o analogous mechanisms (such as classi ie cons uc ions and
non-conca ena i e mo phology) leads o he sugges ion ha sign languages
a e no only di e en o spoken languages because o hei di e en modali y,
bu also ha hey a e simila o one ano he because o hei sha ed modali y.
Howe e , i is impo an o bea in mind ha sign languages show g ea e
a ia ion be ween hemsel es han was once ealised and g owing esea ch
on a wide ange o sign languages con i ms ha (supe icial) uni e sal
p ope ies a e ha d o come ac oss (Pe niss, P au & S einbach 2007).
Fu he mo e, wo k on non-wes e n sign languages and pa icula ly “sha ed
sign languages”, which exis in sociolinguis ic con ex s qui e unlike ha o
mos wes e n sign languages s udied o da e, ha e e ealed g ea e a iabili y
ac oss languages in he isual ges u al modali y. (Fo an o e iew see Nys
2012 and de Vos & P au 2015.) These di e ences be ween sign languages may
be accoun ed o in e ms o di e se ac o s such as modali y, ypology and
pa ame ic a ia ion wi hin he amewo k o Uni e sal G amma , in he
same way ha linguis ic di e si y is explained o spoken languages
(Hohenbe ge 2007).
This b ings us o he al e na i e o he “sign di e en ial” iew: he “sign
same” posi ion holds ha sign languages a e essen ially he same as spoken
languages, bo h being exp essions o he unde lying language acul y ha has
ixed co e p ope ies (Pinke & Jackendo 1995). O signi icance is he nuance
o unde lying simila i y: signed and spoken languages ha e undeniable
di e ences, bu he e a e enough simila i ies o claim ha a common
compu a ional componen se es bo h (Lillo-Ma in 2001, 2002, 2006). The
concep o a speci ic language componen is closely associa ed wi h
gene a i is linguis ics and he Chomskian adi ion which claims ha
language is an inna e human acul y ha exis s as a speci ic cogni i e module
in he b ain (Chomsky 1965; Fodo 1983; see sec ion 2.3 o he nex chap e o
a b ie o e iew o gene a i is linguis ics).
We now u n o speci ic aspec s o modali y di e ences be ween signed
and spoken languages (simul anei y and iconici y) be o e mo ing on o he
issue o space in sign languages.
4 In oduc ion
1.1.1. Simul anei y in sign languages
The a icula o s used o p oduce sign language a e adically di e en o hose
employed by spoken language. The mos salien di e ence is a ques ion o
scale: he hands, a ms, uppe body and head a e much la ge and occupy a
g ea e olume han he ocal appa a us. This gi es hem much g ea e
isibili y and allows hem o make use o space in a way ha he la ynx,
epiglo is, ongue and lips do no , as we shall see in sec ion 1.2. Fu he mo e,
oge he wi h his mac o-scale isibili y, he a icula o s a e ela i ely
independen o one ano he and can pe o m di e en mo emen s a he same
ime. This opens up he possibili y o simul anei y in sign languages.
Sign language p oduc ion may be di ided in o wo main channels: he
manual and he non-manual. The manual componen is a icula ed by he
hands, and he non-manual componen is exp essed by he head ( il s, nods,
shakes), he eyeb ows ( aised, owning), he eyes (gaze di ec ion, blinks,
ape u e), he nose (w inkling), he mou h (mou hing), he shoulde s ( aised)
and uppe body ( il s, u ns). I will look i s a simul anei y wi hin he
manual componen , and hen a he non-manual componen .
The ac ha sign language is a icula ed by he hands, o which he e
a e wo, makes i possible o be doing wo di e en (linguis ic) hings a he
same ime. Howe e , i is no ue o say ha signe s ha e wo independen
a icula o s equi alen o ha ing mul iple oices, like hose o Willie om he
ca oon The whale who wan ed o sing a he Me , who could sing duos wi h
himsel . On he whole, o mos signing p oduc ion, he hands wo k in
coo dina ion: one (non-dominan ) hand is subjuga ed o he o he (dominan )
hand, and he e a e es ic ions on wha he non-dominan hand can be doing
acco ding o he ac i i y o he dominan hand (Ba ison 1978). This holds ue
o mos lexical signs ( hose wi h a ixed o m whose meaning is no en i ely
subjec o con ex ), bu unde ce ain ci cums ances he hands may ac wi h a
g ea e deg ee o au onomy. A axonomy o di e en ypes o simul aneous
cons uc ion is p oposed by Mille (1994) and he e we shall look a wo b oad
g oups o simul aneous bimanual cons uc ions: co-occu ing lexical signs
and classi ie cons uc ions.
The i s ype o simul anei y occu s wi h one-handed signs. Some signs
do no make use o he non-dominan hand, and in he case o some wo-
handed signs he non-dominan hand may be supp essed (Ba ison 1974;
Padden & Pe lmu e 1987; B en a i 1998). Wi h his ype o sign, i is possible
o each hand o a icula e a di e en sign a he same ime, which may be
compa ed o u e ing wo wo ds simul aneously, as can been seen in he
B i ish Sign Language (BSL) examples in (1).
Language and modali y 11
sense) and i s e e en . An example is he sun symbol used on a wea he
o ecas map; bo h sign and e e en a e ound and b igh yellow.
Diag amma ic iconici y is a co espondence be ween pa s o a ep esen a ion
and pa s o he hing i ep esen s. An example is he numbe o lines on a
symbol o ep esen he s eng h o he wind; he ela ion be ween he wind
ba bs co esponds o he ela ion be ween di e en ypes o wind, such ha
he mo e ba s he e a e, he s onge he wind (as exempli ied in igu e 1.1). In
he linguis ic ealm, diag amma ic iconici y is p esen in he co espondence
be ween he empo al o de o o a ions and ha o he e en s desc ibed,
cap u ed by Jakobson’s (1965: 26) classic example ‘ eni, idi, ici’.
6
These
de ini ions o iconici y a e based on Pie ce’s (1932) seminal wo k on
semio ics.
7
ound
b igh yellow
mo e
ba s
s onge
wind
PERCEPTUAL
FEATURES OF
SYMBOL
RESEMBLE
PERCEPTUAL
FEATURES OF
REFERENT
RELATION
BETWEEN PARTS
OF SYMBOL
ANALOGOUS
RELATION
BETWEEN PARTS
OF REFERENT
Figu e 1.1. The ela ionship be ween o m and meaning in imagis ic (le ) and diag amma ic
( igh ) iconici y.
Iconici y is closely ela ed o he no ion o mo i a ion, and his is pe haps one
o he easons why sign languages we e excluded om linguis ic s udy o so
long.
8
The Saussu ean dogma o he a bi a iness o he linguis ic sign as a
de ining ea u e o language led o signed languages being dismissed as me e
pan omime. And indeed, he o ms o signs do end o show g ea e
mo i a ion han hose o wo ds. Conside ing ha much o wha language is
used o alk abou e e s o he isual wo ld, i is no su p ising ha a isual
language shows a g ea deal o mo i a ion in he o m o i s signs.
Con e sely, i we ocus on he domain o sound ela ed concep s, spoken
languages show a much g ea e deg ee o mo i a ion by means o
onoma opoeia and sound symbolism (Hin on, Nichols & Ohala 1994).
6
Fo he de elopmen o he concep o diag amma ic iconici y in spoken languages, see he
pape s in Haiman (1985).
7
Pei ce also included a hi d ype o iconici y: me apho s.
8
Fo an o e iew o he his o y o sign linguis ics, see McBu ney (2001).
12 In oduc ion
Equally, we seem o ha e a na u al bias o conside ing iconici y in isual
e ms and end no o pick up on co espondences be ween o m and meaning
in o he dimensions, such as empo al o ganiza ion (in his espec , o sign
languages see Wilbu 2008).
Much o he li e a u e on iconici y in sign languages deals wi h
imagis ic iconici y a he wo d/sign le el (e.g. Klima & Bellugi 1979; Pizzu o
& Vol e a 2000; Pie and ea 2002; Wilcox 2004; O mel, He mans, Knoo &
Ve hoe en 2009; Pe niss, Thompson & Vigliocco 2010; Baus, Ca ei as &
Emmo ey 2013).
9
The ac ha he sign o a gi en concep may bea a isual
connec ion o i s e e en is unques ionable, as can be seen om he LSE
examples in (8). The examples in (8a) and (b) a e anspa en enough ha hey
would appea in pan omimic ges u es o he same meanings; in (8c) BILBAO
depic s he adi ional musical ins umen s ( he xis u, a one-handed lu e, and
he danbolin, a d um) ypical o he egion whe e he ci y is loca ed
LSE (TZ)
(8) a.
b.
c.
CAR
COLD
BILBAO
Howe e , he e a e wo impo an obse a ions o be made. Fi s ly, al hough
signs may show a ce ain deg ee o isual mo i a ion, his does no exclude
some le el o abs ac ion. In (8a), he sign CAR uses me onymy o associa e
( he ac ion o handling) a pa o he e e en wi h he whole; (8b) COLD
associa es an ac ion ypically used o coun e ac he e ec s o a physical
sensa ion wi h he concep , in wha may be desc ibed as a ype o indi ec
synaes hesia; (8c) BILBAO depic s cul u al a e ac s associa ed wi h he e e en .
The p ocess o he c ea ion o an iconic sign may be b oken down in o
a ious s ages: concep ualizing, image selec ion, schema izing and encoding
(Taub 2001). This ela es o he second poin conce ning iconic signs: e en
hough a gi en sign may ha e a (high) deg ee o isual mo i a ion, his does
9
Wo k which looks a he ole o (imagis ic) iconici y a he discou se le el in sign languages
includes Salland e & Cuxac (2002) and Russo (2004).
Language and modali y 13
no make i s meaning anspa en and does no ob i o all a bi a iness. The
ac ha sign-naï e subjec s ail o guess he meaning o signs abo e chance
con i ms he ela i e opaci y o he o ms (Klima & Bellugi 1979), and (8c)
demons a es ha speci ic cul u al knowledge may be equi ed o deciphe
he isual mo i a ion behind he o m o a sign. Fu he mo e, he a ia ion in
lexical o m ac oss sign languages, as illus a ed in (9) by he di e en signs
used by h ee di e en sign languages – LSE, Aus alian Sign Language
(Auslan) an Colombian Sign Language (LSCol) – o he same meaning,
con i ms ha signs do ha e an elemen o he a bi a y.
Mo e ecen ly, sign esea che s ha e begun o examine he ole o
diag amma ic iconici y. Recall ha in he case o diag amma ic iconici y he e
is no need o he sign o esemble he e e en , bu a he he pa s o he sign
e lec a ela ionship be ween he pa s o he e e en (see igu e 1.1). Fo
spa ial desc ip ions, sign languages may make use o space o desc ibe
loca ion and mo ion. The placemen and mo emen o he signs ela i e o
each o he co esponds o he loca ion and mo ion o he e e en s o each
o he ; he signing space ac s as an map and is exploi ed opog aphically
(Emmo ey, Co ina & Bellugi 1995).
(9) The sign CAR in h ee di e en sign languages. (Image in (b) ep oduced wi h
kind pe mission om he au ho .)
a.
b.
c.
LSE
(TZ)
Auslan
(Vic o ia dialec )
(Johnson 1998: 285)
LSCol
(INSOR 2006: 310)
This opog aphic use o space no mally occu s in conjunc ion wi h classi ie
cons uc ions – see (4) o such a s uc u e in DGS – bu may also occu wi h
lexical signs, as demons a ed by he LSE desc ip ion o he wa e cycle in
(10), which mixes classi ie s wi h lexical signs such as SUN and RAIN wi hin a
diag amma ic spa ial amewo k. The di ec isomo phic mapping ha
opog aphic space es ablishes be ween he spa ial ela ions o he signs and
14 In oduc ion
hose o he e e en s c ea es a pe cep ual simila i y ha is eminiscen o
imagis ic iconici y. Howe e , no e ha he space is no he sign i sel bu
o ms pa o he ela ionship be ween he signs, and ha i is his (spa ial)
ela ionship ha is analogous o a (spa ial) ela ionship be ween he e e en s.
LSE (TZ)
(10)
(Use o space h oughou he discou se.)
D hand
CL(liquid)
SUN
CL(e apo a e)
CLOUD
RAIN
ND hand
CL(liquid)------hold----------
CL(e apo a e)
CLOUD
RAIN
D hand
CL(solid)
SUN
CL( low)
CL(liquid)
ND hand
CL(solid)------hold---------
CL(su ace)—hold-----
‘The wa e is e apo a ed by he sun and ises in o he ai as apou , whe e
i o ms clouds. These clouds hen mo e o e he landmass and become
dense , e en ually leading o p ecipi a ion. The wa e eezes on he
moun ains bu is hen mel ed by he sun and lows down he moun ain
back in o he sea.’
Apa om he opog aphic use o space, a u he ins ance o diag amma ic
iconici y in sign languages is iden i ied by Wilbu (2008), whose E en
Visibili y Hypo hesis s a es ha he pa h mo emen o a p edica e sign maps
on o he du a ion o he e en being desc ibed. In his case, he
co espondence is be ween spa ial ela ionships and empo al ela ionships.
Such ins ances o diag amma ic iconici y highligh he mo i a ed na u e o
Language and modali y 15
ce ain mappings employed by he language sys em and ecen wo k sugges s
ha his mo i a ion may be d i en by uni e sal pe cep ual mechanisms
(S ickland, Ge aci, Chemla, Schlenke , Kelepi & P au 2015). Ne e heless,
his is no he end o he s o y: al hough iconici y may be p esen in a
language, i s ole may be o eed possible o ms in o he sys em, which hen
con en ionalizes and g amma icalizes hese o ms so ha hey i in o he
linguis ic sys em (Wilbu 2008). Wo k on he g amma icaliza ion pa hs o sign
languages sugges s ha some linguis ic o ms may de i e om iconically
mo i a ed ges u es (P au & S einbach 2006a).
Tha sign language o ms a e abs ac , con en ionalized symbols is
con i med by e idence om acquisi ion s udies o sign languages: imagis ic
iconici y does no a ec he acquisi ion o signs in he ea ly s ages o language
de elopmen (O lansky & Bon illian 1984) and he acquisi ion o he e bal
ag eemen sys em is guided by he mo phology a he han he iconici y o
he o ms (Meie 1987). Fu he mo e, he classi ie sys em, wi h i s
diag amma ically iconic use o opog aphic space is ule-go e ned and
ope a es on sys ema ic linguis ic p inciples (Supalla 1982, 1986; Zwi se lood
2003). Iconici y is p esen bu his does no necessa ily al e he wo kings o
he linguis ic sys em.
When conside ing he ole o iconici y in spoken languages, we ha e
al eady seen ha phenomena such as onoma opoeia a e ins ances o imagis ic
iconici y; as a as diag amma ic iconici y is conce ned, he e is a g owing
body o wo k ha eases ou he diag amma ic ela ions in language s uc u e
a he le els o mo phology (Bybee 1985), syn ax (Haiman 1985) and discou se
s uc u e (Ka abæk 2003). Jus as he no ion o iconici y es s on he in ui ion
ha he s uc u e o language e lec s he s uc u e o expe ience (C o 2003:
102), i could be a gued ha he gene a i is s ance ha syn ac ic s uc u e
maps on o o mal seman ic s uc u e esembles an iconic ela ion in i s
isomo phism. Howe e , he s a us o iconici y in language s uc u e is
dispu ed, and i has been sugges ed ha p inciples based on iconic
conside a ions may be due o o he ac o s such as equency o use, o ha
he concep o iconici y is be e exp essed in e ms o no ions o economy and
dis inc i eness (Haspelma h 2008). This calls o mind Wilbu ’s sugges ion ha
iconici y may make mo i a ed o ms a ailable, bu he language sys em hen
g amma icalizes hese o ms.
Be o e closing his sec ion on iconici y, I wish o e u n b ie ly o he no ion o
a bi a iness. Aside om he deba e o e he ole o iconici y in language
s uc u e, he need o an a bi a y ela ion be ween o m and e e en has los
i s oo hold. Fi s ly, we ha e seen ha sign languages show a ela i ely high
deg ee o mo i a ion in he o ms o signs. Fu he mo e, s udies in
16 In oduc ion
synaes hesia show ha he naming o objec s is no a comple ely a bi a y
ma e : he so-called bouba/kiki e ec (see igu e 1.2) demons a es ha he e
is some so o unde lying co espondence be ween he isual appea ance o
he e e en and he o m o he linguis ic sign e en in spoken languages
(Ramachand an & Hubba d 2001). The au ho s claim ha his e ec may go
owa ds explaining he occu ence o sound symbolism, bu in he con ex o
sign languages he e ec may accoun o he high deg ee o mo i a ion.
Spoken languages ha e a long his o y and ha e unde gone housands o
yea s o e olu ion; he changes in he linguis ic sys em ha e led o an
a bi a y ela ionship be ween sign and e e en . Con e sely, (as we shall see
below in sec ion 1.4.2) sign languages a e ela i ely young languages and
hei e olu ion is s un ed by he pa icula sociolinguis ic ci cums ances in
which hey exis . As a esul , he naming p ocesses ha e occu ed much mo e
ecen ly
10
and so he mo i a ional link be ween e e en and o m is s ill
p esen . As such, a bi a iness may be a p ope y o old languages, bu no o
younge languages; he ac o common o bo h ypes o language is he ac
ha linguis ic sign is symbolic (Sandle & Lillo-Ma in 2006: 499).
Figu e 1.2. The bouba/kiki e ec . When asked o choose be ween wo names o hese wo
di e en shapes, 95% o subjec s choose “bouba” o he ounde shape and “kiki” o he
jagged shape. (Adap ed om Ramachand an & Hubba d 2001: 19.)
In his sec ion we ha e looked a he ole o iconici y in sign languages.
Al hough sign languages show a high deg ee o bo h imagis ic and
diag amma ic iconici y, bo h ypes o mo i a ed o m-meaning ela ionship
also exis in spoken languages. Fu he mo e, he non-a bi a y cha ac e o a
o m does no necessa ily de ac om i s linguis ic s a us, and he impo an
ques ion is whe he he o ms a e subjec o he ules o he language sys em.
10
Naming p ocesses may e en eoccu wi h each gene a ion. See sec ion 1.4.2 o his chap e
o he sugges ion ha sign languages unde go a cons an p ocess o c eoliza ion.
The use o space in sign languages 17
As languages in he isual-ges u al modali y, sign languages a e mo e
suscep ible o c ea ing and using isually mo i a ed o ms (and his may be
ein o ced by he ela i e you h o he languages) as well as mechanisms such
as opog aphic space and e en isibili y (Wilbu 2008). No ice ha bo h o
hese iconically d i en de ices in ol e he use o space and ha , once again,
he assessmen o modali y e ec s has led us o conside a ions o he use o
space in sign languages. The nex sec ion examines his opic in de ail.
1.2. The use o space in sign languages
Sign languages a e exp essed in he isual modali y and, as we ha e seen,
make use o a icula o s e y di e en o hose o spoken languages: he
hands and uppe body, including shoulde s, head and ace. Signs a e ei he
p oduced on he body o nea he body in he signing space, which, in mos
sign languages, is app oxima ely he qua e -sphe ical olume jus in on o
he signe (see igu e 1.3).
Figu e 1.3. The signing space occupies app oxima ely a qua e -sphe ical olume in on o
he signe .
This means ha sign languages can make use o space in a way ha spoken
languages canno , and his use o space is pe asi e h oughou he language:
In speech, he acous ic signal de i es om, bu is di e en om, he
mo ion o he a icula o s ( isible in o ma ion on he lips is ex emely
limi ed wi h espec o he whole phonological in en o y). In signing,
he isual signal – he hands mo ing – is he mo ion o he a icula o s,
ha is, wha is seen is he empo al dynamics and spa ial loca ion o
hand mo emen . The linguis ic sys em depends on he isual pe cep ual
sys em o p ocess he necessa y dis inc ions. (Wilbu 2008: 218)
A he phonological le el, space o ms one o he basic building blocks used o
make up an indi idual sign: cu en phonological models o sign language
18 In oduc ion
include some so o loca ion ea u e ha e e s o he poin o poin s in he
signing space whe e he sign is p oduced (Sandle 1989; B en a i 1998; an
de Kooij 2002). (11) shows an example o a minimal pai o signs in LSE ha
di e only in he loca ion o each sign while he handshape, mo emen and
o ien a ion o bo h signs a e iden ical: bo h BROWN and AMAZED in ol e
licking he middle inge o he humb wi h he palm acing away om he
body, bu he o me is a icula ed in neu al space while he la e is
p oduced in on o he ace (no mally in on o he o ehead).
LSE (TZ)
(11) a.
b.
BROWN
(Basque dialec )
AMAZED
Also a he phonological le el, many signs may include a (pa h) mo emen
ea u e in hei lexical speci ica ion: mo emen which, o cou se, occu s
h ough space.
11
Examples o signs wi h a pa h mo emen a e gi en in (12).
The sign THEN desc ibes a ci cle in he neu al space on he signe ’s ipsila e al
side; he sign BASQUE-POLICE in ol es a sho a c mo emen o he hand while
i is in con ac wi h he con ala e al side o he signe ’s ches .
A sign may also be mo ed h ough he signing space in o de o add
in o ma ion o he sign, making he mo emen componen mo phemic in
na u e. This s a egy is exempli ied by classi ie cons uc ions, which employ
a mo emen mo pheme o exp ess p edica es o exis ence o mo ion (Supalla
1982; Emmo ey 2003), as illus a ed in he examples in (13): (13a) desc ibes he
ela i e posi ions o h ee chai s, and (13b) gi es in o ma ion abou he
mo emen o he ca ela i e o he signe .
11
Phonological mo emen may also be in e nal – in Sandle ’s (1989: 92) e ms – o local – in
B en a i’s (1998: 130) e ms – in ol ing a change in he handshape o he o ien a ion o he
hand bu no a pa h mo emen h ough space. Such mo emen may be seen in he signs in
(11) abo e.
The use o space in sign languages 19
LSE (TZ)
(12) a.
b.
THEN
BASQUE-POLICE
LSE (TZ)
(13) a.
CHAIR
THREE
CL(chai )le
CL(chai )middle
CL(chai ) igh
‘The e a e h ee chai s: one on he le , one in he middle and one on he
igh .’
b.
YESTERDAY
CAT
MOUSE
CL(bi e)
CL(walk)
CLx(walk)1
‘Yes e day my ca b ough me a mouse i had hun ed.’
As we saw in sec ion 1.1.2 when looking a he issue o iconici y in sign
language, he signing space may be exploi ed opog aphically in o de o
p o ide spa ial desc ip ions. (Fo example, he spa ial posi ioning o he signs
in (3), (10) and (13) is analogous o he spa ial posi ioning o he e e en s.)
The exp ession o spa ial in o ma ion in sign languages and he use o space
ha his in ol es p o ide insigh in o he mechanisms and cons ain s a
wo k in he isual-ges u al modali y (see Pe niss 2007 o DGS). Howe e , he
20 In oduc ion
ocus o his hesis is he no ion o ag eemen and he use o space in he
exp ession o ag eemen ela ions; as such, spa ial desc ip ions (and he
associa ed use o opog aphic space) will be e e ed o only whe e ele an .
LSE (TZ)
(14)
(Use o space h oughou he discou se.)
SIGNx
SPANISHy
IXy
IXx
xLINKy
SAME
NOT
‘LSE and Spanish a e sepa a e languages ha a e no he same as each
o he .’
Space may also be used a he discu si e le el, wi h di e en discou se opics
o hemes being associa ed wi h di e en a eas o he signing space
(Emmo ey, Co ina & Bellugi 1995). In (14), he signe con as s wo languages
(Spanish and LSE) by associa ing each wi h ei he side o he signing space.
Since nei he o hese signs is body-ancho ed, hey a e a icula ed di ec ly a
he di e en loca ions in he signing space.
LSE
Spanish
Lengua de signos española (LSE) 27
This sec ion o e s a b ie in oduc ion o a ious aspec s o he language.
Sec ion 1.4.1 p o ides in o ma ion on he his o ical backg ound o LSE, which
is ela i ely well his o ically documen ed compa ed o o he sign languages.
The cu en sociolinguis ic si ua ion o LSE, essen ial o unde s anding many
o he ac o s ha condi ion he language, is desc ibed in sec ion 1.4.2. An
o e iew o p e ious linguis ic wo k on LSE is p o ided in sec ion 1.4.3, and
he las sec ion, 1.4.4, iden i ies he cha ac e is ics o he speci ic a ian o LSE
analysed in his hesis.
1.4.1. LSE: his o ical backg ound
LSE is used in he Dea communi y h oughou mos o Spain, including he
o sho e e i o ies o he Cana y Islands, Ceu a and Melilla, wi h he
excep ion o Ca alonia, whe e LSC is used (see abo e). The o igins o he
language a e no known, al hough some so o sign language has been in use
in Spain o a leas ou hund ed and i y yea s; Spanish Sign Language is in
he p i ileged posi ion o ha ing a ela i ely ich body o his o ical li e a u e
da ing om he mid-six een h cen u y onwa ds (c . Lasso 1550/1919). Many o
hese documen s a e eely a ailable as digi al acsimiles in he Biblio eca de
Signos, which o ms pa o he Biblio eca Vi ual Miguel de Ce an es hos ed by
he Uni e si y o Alican e.
18
Needless o say, he in o ma ion o e ed by hese
ex s is somewha e a ic and subjec o he au ho s’ p ejudices and in en .
Many o he w i ings, such as Bone ’s Reduc ion de las le as y a e pa a enseña a
abla los mudos (1620), deal wi h he educa ion o he dea , and speci ically
how o “imp o e” he sign language by making i mo e simila o he spoken
language. Fu he mo e, wha li le in o ma ion hese ex s o e is in he o m
o w i en desc ip ions and any g aphic ep esen a ions a e ine i ably o he
manual alphabe , which e eals nex o no hing abou he sign language i sel
(see igu e 1.5). E en so, hese his o ical ex s p o ide unques ionable
e idence ha he e has been a communi y o sign language use s in Spain o
se e al cen u ies. Ano he ma e is how simila he sign language o he
six een h o se en een h cen u y was o p esen day LSE. The sociolinguis ic
cha ac e is ics o sign languages gi e ise o a g ea deal o a ia ion bo h
his o ically and geog aphically, and his ma e is deal wi h in he nex
sec ion.
Along wi h he documen ed his o y o he sign language used by dea
people in Spain o e he las ew cen u ies, he e is a ich o al adi ion among
he Dea communi y wi h his o ical igu es such as Ped o Ponce de León o
Juan de Pablo Bone ea u ing in s o y- elling. Al hough his p ac ice is less
18
h p://www.ce an es i ual.com/seccion/signos/
28 In oduc ion
p e alen among he younge gene a ions o signe s, knowledge o he
his o y o he Dea Communi y is ega ded as a means o o ging social
iden i y. Con e sely, some au ho s ha e c i icised he appa en my hi ica ion
o ce ain cha ac e s and episodes (c . Gascón Ricao & S o ch de G acia y
Asensio 2004).
Figu e 1.5. The manual alphabe aken om a ex book o eaching dea s uden s (Balles e os
& Fe nández Villab ille 1845).
I has been claimed ha LSE belongs o he sign language amily o Sou h-
Wes Eu ope (Ande son 1979 ci ed in Woll, Su on-Spence & El on 2001: 26),
which includes F ench Sign Language (LSF) and Sign Language o he
Ne he lands (NGT). Howe e , as men ioned abo e, he o igins o LSE a e
unclea . Apa om a cu en lack o su icien knowledge on speci ic sign
languages o es ablish gene ic links be ween hem, he discipline o his o ical
linguis ics has no been de eloped o sign languages,
19
hus making i
19
Excep ions a e F ishbe g (1975), Woll (1987), Wi mann (1991), Janzen & Sha e (2002) and
Wilcox (2004). Fo an o e iew see McBu ney (2012).
Lengua de signos española (LSE) 29
di icul o go beyond specula ion abou he ela ions be ween he olde sign
languages o e en he exis ence o some “Eu opean P o o Sign Language”.
20
1.4.2. LSE: sociolinguis ic se ing
As is he case o mos wes e n sign languages, he sociolinguis ic se ing o
LSE is qui e complex: he language exis s in a pe manen s a e o bi- o
mul ilingualism, has no w i en o m, has been subjec o ins i u ionalized
opp ession and is isual-ges u al in o m. Lea ing aside he di e ence in
modali y, many o hese cha ac e is ics a e ypical o mino i y spoken
languages.
Wha eally se s sign languages apa is hei lack o gene a ional
con inui y: he as majo i y o dea child en a e bo n o hea ing amilies, and
as such hey will no no mally be able o acqui e sign language in a na u al
se ing. The igu e o en ci ed in he li e a u e is ha be ween 5-10% o dea
child en a e bo n in o dea amilies (Schein & Delk 1974), bu ecen wo k
sugges s ha his holds only o he Uni ed S a es, whe e a pa icula s ain o
gene ic dea ness and a ce ain deg ee o endogamy wi hin he dea
popula ion has led o a ou able condi ions o mul igene a ional dea
amilies (Nance, Liu & Pandya 2000). The igu e may be much lowe o o he
coun ies (Johns on 2006), Spain included (Cos ello, Fe nández & Landa 2008).
This si ua ion means ha he e a e ex emely ew na i e signe s in he signing
communi y. Hence, om he poin o iew o linguis ic esea ch, he
me hodology has o be adap ed o hese peculia condi ions. The esea ch
me hods and da a collec ion echniques used in his s udy a e desc ibed in
chap e 4.
The ac ha so ew dea child en a e bo n in o an en i onmen ha is
conduci e o hei acqui ing sign language means ha many lea n a sign
language a a la e age (ei he upon s a ing a o mal educa ion o e en la e ,
when hey ha e le school) and his has a huge impac on he language i sel .
As migh be expec ed unde such ci cums ances, each sign language shows a
ai ly high deg ee o a ia ion. Fu he mo e, i has been p oposed ha he
gene a ional discon inui y causes sign languages o unde go a con inual
p ocess o c eoliza ion (Fische 1978; Meie 1984; A ono , Padden, Mei &
Sandle 2003). Whe he o no he analysis o sign language as some so o
c eole is co ec (Lup on & Salmons 1996), he low le el o na i e-like
20
This clea ly is no he case o sign languages which ha e de eloped in he con ex o mo e
ecen his o ical p ocesses such as colonialism, as is he case o he a es ed p oximi y
be ween Auslan – Aus alian Sign Language – and BSL – B i ish Sign Language (c . Johns on
2003).
30 In oduc ion
acquisi ion amongs he use s ce ainly a ec s he language i sel (Cos ello,
Fe nández Landaluce, Villame iel & Mosella 2012).
A clea example o his is he ole ha educa ional policy can play in he
de elopmen o he language: since many indi iduals lea n sign language a
school a he han a home, he ype o language hey lea n is subjec o he
whims o educa ional policy (Bou e 1990). In mos wes e n coun ies, he
policy o Dea educa ion has been subjec o adical changes in he las 150
yea s and his is no less ue o Spain. Teaching p ac ice and philosophy ha e
gone om sign-based me hods o a long pe iod o o alism (du ing which he
use o sign language was discou aged o ac i ely punished), o he mo e
ecen ein oduc ion o sign language as pa o a bilingual/bicul u al
pedagogy accompanied by mains eaming (Plann 1997; Mingue So o 2001).
This endency o inse dea pupils in o dina y schools is signi ican because
he esiden ial dea schools we e adi ionally he sea o sign language
lea ning and o en ga e ise o a ian s ha o med he main egional dialec s
o a na ional sign language (Kyle & Woll 1986). The disappea ance o he
esiden ial schools has mean he loss o he oci o di e en dialec s.
The dialec al a ia ion o LSE has been s udied by means o lexical
compa ison using a Swadesh ype wo d lis in he glo och onology adi ion
(Gudshincksy 1956; Swadesh 1972) and mu ual in elligibili y es s (Pa khu s
& Pa khu s 2007). The s udy looked a sign language use in 18 di e en pa s
o Spain and he indings con i med ha ce ain lexical di e ences do exis
om one egion o he nex , bu he le el o mu ual in elligibili y be ween
signe s om di e en a eas is well abo e he 75% h eshold usually applied
o dis inguish di e en spoken languages (SIL 1991: 45, ci ed in Pa khu s &
Pa khu s 2007: 46). Fu he mo e, he indings dis inguish be ween LSE and
LSC, each wi h hei own in e nal dialec al a ia ion. The Pa khu s s udy
also p o ides a gene al o e iew o he si ua ion o sign languages in Spain,
including de ails on he ole o Dea Schools and Associa ions in he
de elopmen o language a ie ies. Fo u he in o ma ion on he signing
communi ies o Spain, see Vall e dú (2001) and Que , Mazzoni &
Sapoun zaki (2010); o mo e speci ic sociolinguis ic in o ma ion, see Mingue
So o (2001). Fo he sociolinguis ics o sign languages in gene al, see Lucas
(2001).
LSE was g an ed a ce ain deg ee o o icial ecogni ion by a law passed
in la e 2007 ha se ou o egula e he igh s and communica ion op ions o
dea indi iduals, hough i emains o be seen wha impac his law will ha e
on sign language use and he de elopmen o he language in he long e m.
21
21
The ull Spanish ex o he law is a ailable on-line:
Lengua de signos española (LSE) 31
1.4.3. P e ious esea ch on LSE
Taking a b oad iew, he desc ip ion o LSE appea s o con o m o many o
he indings o o he wes e n u ban sign languages ha ha e been s udied:
he language has a classi ie sys em and makes use o non-manual ea u es, o
ci e jus wo o he ea u es ha ypi y sign languages. The ela i e simila i y
be ween sign languages leads o he dange o o e -gene alizing indings
om one sign language o o he s, bu scien i ic me hod demands ha each
objec o s udy be examined in i s own igh . Al hough he la ge body o
esea ch on ASL ( he mos s udied sign language o da e) may in o m wo k
on lesse s udied languages, i is impo an o hose languages o be
in es iga ed in o de o ind ou wha hei own peculia i ies a e. So while
LSE may look ai ly simila o o he sign languages, he e is s ill a need o
ca y ou esea ch, i only o con i m i s imp essions. This sec ion p esen s
an o e iew o he esea ch wo k ha has al eady been done on LSE.
The e a e se e al dic iona ies o LSE, he i s published in he mid-
nine een h cen u y (Fe nández Villab ille 1851). Since 1957 a dic iona y has
been published unde he auspices o he Na ional Associa ion o Dea People,
CNSE, (Ma oquín Cabiedas 1957, ci ed in Vicen e Rod íguez e al. 2008, and
Pinedo Peyd ó 1981) and in ecen yea s specialized hema ic dic iona ies
ha e been c ea ed (Fundación CNSE 2002-2003). Mo i a ed by he 2007 law
ha p o ides ce ain legal ecogni ion o LSE and lays down p o ision o a
s anda diza ion p ocess o he language (see p e ious sec ion), CNSE has
published a no ma i e dic iona y wi h o e 4,000 en ies (Fundación CNSE
2008).
22
All hese dic iona ies a e limi ed o single sign en ies and p o ide no
g amma ical in o ma ion abou LSE.
The e has been e y li le mode n linguis ic analysis o LSE, and his
hesis o ms pa o a g owing body o wo k ha documen s he language.
P e ious wo k on he language includes i e doc o al heses: Rod íguez
González (1990) p esen s a gene al linguis ic analysis o LSE wi h espec o
he s uc u e o Spanish; Fe nández Sonei a (2004) on quan i ica ion in LSE;
Iglesias Lago (2006) on non-manual ea u es o exp ess modali y; G as Fe e
(2006) on he sociolinguis ic s a us o sign languages in Spain; Gu ié ez Sigu
(2008) on he ole o phonological ea u es in he p ocessing o LSE. Cu en ly
he e a e also se e al doc o al disse a ions on LSE in p og ess a a ious
Spanish uni e si ies.
Published wo k on LSE is limi ed o a lea ne s’ g amma (He e o
Blanco 2009) and a collec ion o a icles published by he Na ional Associa ion
h p://www.boe.es/boe/dias/2007/10/24/pd s/A43251-43259.pd
22
This dic iona y can be accessed on-line: h p://www. undacioncnse.o g/ eso olse
32 In oduc ion
o Dea People (CNSE 2000), a s udy o egis e in LSE (Chapa Baixauli 2001)
and a gene al ex book dealing wi h a ange o issues ha d aws on li e a u e
o o he sign languages (Mingue So o 2001). The e a e also se e al a icles
(almos exclusi ely in Spanish
23
) published by he a ious esea ch g oups
wo king on LSE: he main g oups a e o be ound a he Uni e si y o Vigo
and he Uni e si y o Alican e. The Alican e g oup has also published a ious
ma e ials on-line, including a bilingual (LSE-Spanish) basic g amma o LSE
(He e o e al. 2005).
24
1.4.4. The LSE in his s udy
This s udy looks speci ically a he a ie y o LSE used in he Basque Coun y,
a egion in he no h o Spain (shaded solid in igu e 1.4). As no ed abo e, LSE
shows a ce ain amoun o a ia ion bu his does no a ec in elligibili y
be ween use s om di e en a eas. In pa icula , signe s om he Basque
Coun y can easily unde s and and be unde s ood by signe s om o he pa s
o Spain. Fu he mo e, he membe s o he signing communi y o he Basque
Coun y conside hei language o be LSE. Bilbao, he la ges ci y in he
Basque Coun y, was home o a la ge esiden ial dea school (which in he las
20 yea s had been educed by mains eaming o a uni o pupils wi h mixed
special needs wi hin an o dina y school and ecen ly closed down al oge he ).
The LSE a ian o he Basque Coun y is gi en some uni o mi y by he ac
ha many olde signe s lea n o sign he e. Fu he de ails o he da a
collec ed o his s udy and he signe s who p o ided he da a will be
p o ided in chap e 4.
I should be emphasized ha he e is no Basque Sign Language, hough
he Basque Coun y does ha e wo main spoken languages: euske a, o
Basque, a language isola e o unknown o igin, and cas ellano, o Spanish, a
Romance language used h oughou all o Spain. This spoken language
bilingualism adds o he complex sociolinguis ic backg ound o he LSE
a ian used in he Basque Coun y, bu his issue will ba ely be ouched
upon in his wo k. Mos Dea people in he Basque Coun y ha e Spanish as
hei main spoken language (be i hei i s o second language), bu he
p omo ion o Basque in he compulso y educa ion sys em has mean ha he
dominan spoken language o some younge signe s is Basque (no mally o
hose whose amilies a e Basque speake s). Gi en he dispa i y be ween
23
A icles on LSE published in English include Cabeza Pe ei o & Fe nández Sonei a (2004)
and He e o Blanco & Salaza Ga cía (2005) in addi ion o he dialec s udy men ioned in
sec ion 1.4.2: Pa khu s & Pa khu s (2007).
24
Un o una ely he o ma o he on-line g amma is ou da ed and i has been di icul o
access he con en s since 2009.
The goals o his hesis 33
Spanish and Basque, his si ua ion p esen s a ascina ing oppo uni y o
examine language con ac phenomena be ween a sign language and wo e y
di e en spoken languages, bu I lea e his a ea o u u e esea ch.
1.5. The goals o his hesis
Ha ing ou lined he gene al opic and in oduced he indi idual language
ha his hesis deals wi h, I now u n o he speci ic objec i es o his esea ch
wo k.
Wha mechanisms does LSE use o ag eemen , and o wha ex en a e hey he
same as o di e en o ag eemen mechanisms employed by o he sign
languages?
The i s ask in hand is la gely desc ip i e as I se ou o desc ibe he e bal
ag eemen p ocess in LSE, looking a i s phonological mani es a ion and
ocusing on any peculia i ies i may ha e when compa ed wi h wha has
al eady been desc ibed o o he sign languages. Ag eemen o sign
languages, based on wo k on a a ie y o sign languages, is desc ibed in
chap e 3; he phenomenon in LSE is desc ibed in chap e 5. The ele an
compa isons a e made h oughou chap e 5 and in he concluding sec ion o
ha chap e .
E en hough he de ails need o be p o ided in he desc ip ion o he
LSE da a on ag eemen , we know ha his p ocess in ol es he use o space
and he associa ion be ween a locus and a e e en (see sec ions 1.2 and 1.3
abo e). In he sign language li e a u e, (ce ain cases o ) his spa ial
mechanism a e e e ed o as “ag eemen ” and conside ed o be analogous o
he same phenomenon in spoken language. One o he aims o his hesis is o
assess how alid his iden i ica ion is.
A e he spa ial mechanisms employed by LSE compa able o he ag eemen
mechanisms in spoken languages?
This will be done by using wo di e en linguis ic adi ions as ya ds icks o
he sign language da a. Fi s ly, a ypological app oach p o ides a b oad iew
ha de ines ag eemen in e ms o he di e en s uc u al elemen s ha play a
ole in he p ocess, and, u he mo e, o e s a ich o e iew o he di e en
op ions ha exis in ypologically di e se languages o he wo ld. Secondly,
minimalis syn ax, om he gene a i is adi ion, o e s a echnical and
highly speci ied no ion o ag eemen in e ms o s uc u al ela ions and
syn ac ic ope a ions. These di e en – bu complemen a y – app oaches
p o ide he means o hold up he LSE da a agains da a om o he spoken
languages, and heo ies de eloped based on spoken language da a.
34 In oduc ion
E alua ing he spa ial ag eemen p ocess in his way will also o e he
oppo uni y o de elop a o mal cha ac e iza ion o he phenomenon. In
desc ibing and analysing he da a, i will become e iden whe he o no
spa ial ag eemen in LSE i s in o exis ing models, and wha , i any,
adjus men s a e necessa y o accommoda e he model o he da a.
Can spa ial ag eemen in LSE be gi en a o mal cha ac e iza ion?
These, hen, a e he h ee esea ch ques ions ha d i e he esea ch laid ou in
his hesis. They can be summa ized as an a emp o compa e LSE ag eemen
o wha is known abou o he sign languages and abou spoken languages o
decide whe he he phenomenon can co ec ly be cha ac e ized as ag eemen
using (and, i necessa y, adap ing) cu en models.
1.6. The s uc u e o his hesis
This hesis is s uc u ed in he ollowing manne . I begin by p o iding he
heo e ical backg ound o ag eemen , om he poin o iew o gene al
(spoken language) linguis ics and sign languages, espec i ely. A e
desc ibing he me hodology employed, I ocus on ag eemen in LSE based on
he da a collec ed o his s udy and p o ide a de ailed accoun o how
ag eemen is mani es in his language. I hen apply wo di e en app oaches
( ypological and gene a i is ) o analyse he LSE da a wi h a iew o
e alua ing how well ag eemen in LSE i s in o he c oss-linguis ic landscape.
I also p o ide a o mal accoun o ag eemen in LSE based on he idea o a
basic spa ial ag eemen mechanism. I conclude by aking a s ep back o
conside wha his spa ial ag eemen mechanism in LSE can ell us abou
modali y e ec s and language in gene al.
Chap e 2 p esen s an o e iew o ag eemen as a linguis ic
phenomenon om wo di e en amewo ks: linguis ic ypology, which se s
ou o desc ibe da a om as b oad a ange o languages as possible; and
Gene a i e G amma , which o e s a se o concep s and mechanisms ha
p o ide a de ailed syn ac ic accoun o he wo kings o ag eemen . Al hough
hese wo amewo ks ep esen qui e di e en app oaches, I jus i y using
bo h as complemen a y me hods, each o which con ibu es i s own bene i s.
Chap e 3 p o ides an o e iew o he phenomenon o ag eemen as i
has been desc ibed in he sign language li e a u e. This in ol es looking i s
a how loca ions in space a e used o e e ence, including he p ocess o
loca ion assignmen . The o e iew akes in phenomena ha ha e p e iously
been analysed as ag eemen (namely, ag eeing e bs, backwa ds e bs,
ag eemen auxilia ies, and non-manual ag eemen ). Addi ionally, I also
desc ibe and conside simila uses o space as possible candida es o a spa ial
The s uc u e o his hesis 35
ag eemen p ocess, such as single a gumen ag eemen and DP-in e nal
ag eemen , and a gue ha hese p ocesses also cons i u e ins ances o a
gene al spa ial ag eemen mechanism.
Chap e 4 desc ibes he me hodology used in his s udy, which was
adap ed o he speci ic sociolinguis ic ci cums ances o LSE as desc ibed in
sec ion 1.4.2, and gi es de ails o he da a. This includes in o ma ion on he
in o man s, he da a collec ion echniques and how he da a we e ansc ibed
and analysed.
Chap e 5 o e s a desc ip ion o spa ial ag eemen in LSE. The s uc u e
o his chap e b oadly ollows ha o chap e 3 so as o p o ide a compa ison
be ween he ac s o LSE and wha has been desc ibed o o he sign
languages. Thus, desc ip ions a e gi en o a ange o phenomena ela ed o
spa ial ag eemen : ag eeing e bs, backwa ds e bs, ag eemen auxilia ies,
and non-manual ag eemen , as well as single a gumen ag eemen and DP-
in e nal ag eemen . I p o ide a gumen s o he e ec ha all hese
phenomena in ol e a basic p ocess o spa ial ag eemen . The chap e also
includes a desc ip ion o he cons ain s on pe son/numbe combina ions in
ag eeing e bs based on a simila s udy wi h o he sign languages. The
compa ison makes e iden ha ag eemen o ms in LSE do no show he
same pa e ns as in o he sign languages, bu a e subjec o he same ype o
cons ain s (i.e. phonological).
Chap e 6 examines he LSE da a om a c oss-modal ypological poin
o iew, d awing on he heo e ical amewo k p esen ed in sec ion 2.2.
Fi s ly, his assessmen o he spa ial ag eemen mechanism in LSE akes each
elemen o he ag eemen p ocess and looks a how he LSE ac s compa e o
he gene aliza ions d awn om ypological wo k on spoken languages. Thus,
he con olle s, a ge s, means o exponence, domains, ea u es and condi ions
ha appea in spa ial ag eemen in LSE a e held up agains wha has been
desc ibed o spoken languages. By and la ge, LSE i s wi hin he limi s o he
spoken language da a, wi h he impo an excep ion o he ea u es used in
spa ial ag eemen in LSE. I claim ha pe son is no a ele an ea u e o LSE
(in con as o i s uni e sal p esence in spoken languages) and p opose an
al e na i e ea u e, iden i y, based on e e en ial iden i y (de eloped in
chap e 7).
The second pa o he assessmen o he LSE da a in chap e 6 in ol es
exploi ing he no ion o canonici y de eloped by Co be (2003b, 2006).
Canonci y is de ined by a se o c i e ia ha p o ides a means o e alua ing
spa ial ag eemen in LSE o see how p o o ypical o canonical i is as an
ag eemen p ocess. Again, he esul s show ha LSE ag eemen beha es in a
ela i ely canonical manne . Mo e in e es ingly, his e alua ion highligh s
36 In oduc ion
hose aspec s in which LSE is unusual. Speci ically, he op ionali y o he
spa ial ag eemen p ocess is a ema kable cha ac e is ic (also aken up in
chap e 7).
Chap e 7 once mo e analyses he LSE ag eemen da a bu om he
poin o iew o gene a i is app oaches o language s uc u e: minimalis
syn ax and Op imali y Theo y. To p epa e he e ain, wo issues iden i ied in
he p e ious chap e a e ackled i s : he na u e o he iden i y ϕ- ea u e and
he op ionali y o he spa ial ag eemen sys em. Adop ing a dis ibu ed model
o ϕ- ea u es and a ea u e-sha ing heo y o ag eemen , I p o ide a syn ac ic
analysis o spa ial ag eemen in LSE ha can accoun o he p ocess o
loca ion assignmen , and e bal ag eemen o wo-place ag eeing e bs, o
single a gumen ag eemen and o ag eemen auxilia ies. Fu he mo e, his
syn ac ic accoun can p o ide an explana ion o he di e ence be ween
syn ac ic ag eemen and a o mally simila use o space ha leads o di e en
in e p e a ions, namely p agma ic ag eemen . This chap e also includes a
o mal analysis o a speci ic ype o ag eeing e b ha has unusual in lec ed
o ms due o a con lic be ween ag eemen ma ke s and lexically de ined
phonological ea u es. Using Op imali y Theo y, I p o ide an analysis o he
LSE ac s ha can also be ex ended o analogous da a om ano he sign
language.
Chap e 8 concludes his hesis by e isi ing he esea ch goals o
p o ide answe s o he ques ions se ou in he p e ious sec ion, and also o
assess he ex en o which i has been possible o answe hese ques ions. The
chap e also includes discussion o he issues ha a ise om his examina ion
o spa ial ag eemen in LSE, and poin s he way o u u e esea ch.
Typological app oach 43
Be o e mo ing on o look a each o hese ac o s, one mo e s ipula ion
wi h ega d o e minology mus be made. Th oughou his hesis I make use
o he e ms “subjec ” and “objec ” o ease o exposi ion. I am awa e ha
hese labels make assump ions abou he syn ac ic s a us o he e b’s
a gumen s, and ha he e is a need o look beyond such e ms o cap u e he
unde lying, mo e p imi i e no ions (McCloskey 2001). Com ie (1989: 70)
makes use o a se o labels ha ob ia e p e-emp ing he ques ion o
g amma ical ela ions: S is he a gumen o an in ansi i e e b; A is he
a gumen o a ansi i e e b ha co ela es mos closely wi h he agen ; and
P is he a gumen o a ansi i e e b ha co ela es mos closely wi h he
pa ien . While I ecognize he alue o using such e ms when wo king wi h a
ela i ely undesc ibed language, I do no adop hem he e (al hough hey a e
some imes used when discussing examples om he ypological li e a u e).
The eason is ha he dis inc ions hese labels make do no add any hing o
he analysis o he LSE da a and he challenge o spa ial ag eemen lies in
cha ac e izing he use o space a he han he a gumen ha is exp essed.
2.2.2. Con olle s
As Co be (2006: 35) s a es, “Con olle s a e ypically nominal in na u e.”
Nouns and noun ph ases o en con ol ag eemen wi h adjec i es and e bs,
espec i ely, as shown in (3). The adjec i e ag ees in numbe and gende wi h
he noun pe sonas, and he e b ag ees in pe son and numbe wi h he noun
ph ase las pe sonas icas:
Spanish
(3)
Las
pe sona-s
ic-as
abaj-an
poco.
DEF.F.PL
pe son(F)-PL
ich-F.PL
wo k-3PL
li le.
‘Rich people wo k li le.’
In he case o e bal ag eemen , he noun ph ase con olle s mos o en ul il a
p o o ypical seman ic ole, such as AGENT, THEME o GOAL, as exempli ied
by (1)-(3). Howe e , o he ypes o a gumen may be he sou ce o ag eemen
on he e b. In Chicheŵa [Nyanja] (Cen al Ban u, Nige -Congo), spoken in
Eas Cen al A ica, he e b ag ees in gende wi h he loca i e a gumen in
speci ic cons uc ions. In (4a) he ma ke ku- on he e b ‘come’ ma ks
ag eemen wi h he loca i e a gumen ‘ illage’. The ac ha he a gumen is
loca i e, and no ano he ole such as THEME o GOAL (as would be he case
in a sen ence like ‘The illage ecei ed hose isi o s’), is con i med by he
seman ics o he e b and he obse a ion ha unde di e en wo d o de
condi ions i ag ees wi h he o he a gumen , a-lendô-wo (‘ hose isi o s’), as in
(4b).
44 Theo ies o ag eemen
Chicheŵa (B esnan & Kane a 1989: 2)
(4) a.
Ku-mu-dzi
ku-na-bwé -á
a-lendô-wo.
G17-G3- illage
G17.SBJ-REC PST-come-IND
G2- isi o -G2. hose
‘To he illage came hose isi o s.’
b.
A-lendô-wo
a-na-bwé -á
ku-mu-dzi.
G2- isi o -G2. hose
G2.SBJ-REC PST-come-IND
G17-G3- illage
‘Those isi o s came o he illage.’
G=gende class
A con olle may also be a less ypical nominal elemen han a noun ph ase,
such as a clause (5a) o an in ini i al (5b):
Spanish
(5) a.
Que
las
modelo-s
gan-en
an o
pa ec-e
injus -o.
COMP
DEF.F.PL
model(F)-PL
ea n-3PL.SBJV
so_much
seem-3SG
un ai -M[SG]
‘I seems un ai ha models ea n so much.’
b.
T abaja
no
es
san-o.
wo k.INF
NEG
be.3SG
heal hy-M[SG]
‘Wo king (li . o wo k) is no heal hy.’
These con olle s a e de ec i e since hey do no ha e speci ic ag eemen
ea u es so he a ge shows de aul ag eemen , which in his case is hi d
pe son singula on he e bs. This is u he shown by he ac ha he
p edica i e adjec i es in bo h sen ences also ag ee wi h he de ec i e
con olle in he masculine singula de aul o m, despi e e en he p esence o
a eminine plu al con olle wi hin he embedded clause in (5a).
Con olle s may also be non-o e , as occu s in p o-d op languages
such as Pash o (Indo-Eu opean), an Indo-I anian language spoken in
A ghanis an, which does no equi e he subjec a gumen o be explici (6):
Pash o (Neeleman & Szend ői 2007: 672)
(6)
(Zə)
manna
xw -əm.
I.NOM
apple
ea -1SG
‘[I] ea he apple.’
I is also possible o he con olle o be comple ely absen such ha i canno
appea , as occu s wi h e bs ha desc ibe na u al phenomena, exempli ied by
he C oa ian example (7), in which no o e subjec is possible.
Typological app oach 45
C oa ian (Ma k Schmalz, pe sonal communica ion)
(7)
M ač-i
se.
(*
Ono
ač-i
se.)
ge _da k-3.SG
REFL
i .N
ge _da k-3.SG
REFL
‘I ’s ge ing da k.’
To summa ize, con olle s a e gene ally nominal in na u e – in he case o
e bal ag eemen hey a e NPs – and i is possible ha hey a e no o e ly
exp essed.
2.2.3. Ta ge s
The examples in he p e ious sec ion make i clea ha he mos
commonplace a ge s o ag eemen a e adjec i es (example 3) and e bs (all
examples). Howe e , he e is a se ies o o he elemen s ha ma k ag eemen ,
such as p onouns, nume als, quan i ie s, adposi ions and nouns. Be o e
conside ing hese a ge s, le us look mo e closely a ag eemen ma king on
e bs.
2.2.3.1. Ve bs and auxilia ies
In nea ly all he examples so a , e bal ag eemen has been ma ked di ec ly
on he e b, bu i is also possible o an auxilia y e bal elemen o bea
ag eemen in o ma ion. An auxilia y e b may be de ined as:
an elemen ha in combina ion wi h a lexical e b o ms a monoclausal
e b ph ase wi h some deg ee o (lexical) seman ic bleaching ha
pe o ms some mo e o less de inable g amma ical unc ion. (Ande son
2006: 5)
Auxilia y e bs ypically exp ess e bal ca ego ies o aspec and modali y,
and may also exp ess ense, nega i e pola i y o oice ca ego ies. These
ca ego ies encode in o ma ion abou he e b seman ics, and so do no
ep esen he displaced in o ma ion ha cha ac e izes ag eemen . Howe e ,
auxilia ies may also exp ess he ea u es associa ed wi h ag eemen , such as
pe son, numbe and gende . In (8a) om Bu ushaski, a language isola e o
No he n Pakis an, he auxilia y ma ks pe son o he subjec a gumen , and
pe son and numbe o he possesso o he objec a gumen .
Bu ushaski (Be ge 1998: 162, 161, ci ed in Ande son & Egge 2001: 240, 242)
(8) a.
kʰɑkʰɑ ɑy-umuc
pʰɑ
mée- -ɑɑ
walnu -PL
gobble_up
1PL-AUX-2
‘You gobbled up ou walnu s.’
46 Theo ies o ag eemen
b.
ɑ ɑ
ɑ-yúgusɑnc
moó-y-ɑ
bɑ -ɑ
I.GEN
1-daugh e .PL
2PL-gi e-1
AUX-1
‘I he ewi h am gi ing you my daugh e s.’
In such auxilia y e b cons uc ions, he e a e a numbe o ways in which he
in lec ional ma e ial is dis ibu ed be ween he lexical e b and he auxilia y
e b. In (8a) he lexical e b
pʰɑ
(‘gobble up’) is unin lec ed and he auxilia y
mée- -ɑɑ
bea s all he in lec ional ma ke s; in (8b), in con as , some in o ma ion
is ma ked on he lexical e b ‘gi e’ while he auxilia y s ill bea s an
in lec ional ma ke . Al hough Bu ushaski shows di e en ways o
dis ibu ing in lec ional ma e ial be ween he e b and auxilia y, many
languages consis en ly use one pa e n.
All in o ma ion may appea on he auxilia y; an example o such an
AUX-headed language is Ia mul (Sepik), as can be seen in example (9), in
which subjec ma king appea s on he auxilia y.
Ia mul (Foley 1986: 144, ci ed in Ande son 2006: 24)
(9)
klə-kə
lɨ-kə-wɨn
ge -DEP
AUX-PRES-1SG
‘I am ge ing i .’
DEP=dependen
Al e na i ely, i may be he lexical a he han he auxilia y e b ha bea s he
pe son in lec ion; his is he case o Mödö (Nilo-Saha an), shown in example
(10), which ma ks o subjec on he lexical e b ‘ escue’.
Mödö (Pe sson & Pe sson 1991: 19, ci ed in Ande son 2006: 24)
(10)
í
mók nyì
yí
FUT
1: escue
you
‘I will escue you.’
Ano he possibili y is ha bo h elemen s a e in lec ed: example (11) shows
how in Go um (Aus o-Asia ic) bo h he lexical e b ‘ea ’ and he auxilia y
bea ma king o he i s pe son subjec .
Go um (Aze 1973: 279, ci ed in Ande son 2006: 25)
(11)
miŋ
ne-gɑʔ- u
ne-lɑʔ- u
I
1-ea -PST
1-AUX-PST
‘I a e igo ously.’
The in lec ional ma e ial may be di ided be ween he lexical and auxilia y
e bs such ha each elemen bea s di e en in o ma ion; Jakal ek (Mayan) is
Typological app oach 47
a spli language o his ype and example (12) shows how he subjec is
ma ked on he lexical e b ila (‘see’) while he objec is ma ked on he
auxilia y.
Jakal ek (C aig 1977: 60, ci ed in Ande son 2006: 25)
(12)
-ɑch
w-ila
COMPL-ABS2
ERG1-see
‘I saw you.’
Ande son also iden i ies a i h ca ego y o auxilia y e b cons uc ions,
which he calls he spli /doubled ype. In languages o his ype, he
in o ma ion is ma ked on bo h he lexical and he auxilia y e bs, bu
incomple ely so on one o he wo elemen s. Bu ushaski has spli /doubled
auxilia y e b cons uc ions, as can be seen in (8b): he subjec is ma ked on
bo h elemen s, bu he objec is ma ked on he lexical e b alone. The di e en
ypes o auxilia y e b cons uc ions a e summa ized in able 1.1.
LEX
AUX
AUX-headed ype
–
+
LEX-headed ype
+
–
Doubled- ype
+
+
Spli ype
+i/‒ j
– i/+j
Spli /doubled ype
+i/+ j
– i/+j
– i/+j
+ i/+j
Table 1.1. A ypology o auxilia y e b cons uc ions showing he possibili ies o he
dis ibu ion o he in lec ional ma e ial be ween he lexical e b (LEX) and he
auxilia y e b (AUX). (Adap ed om Ande son 2006: 24-27.)
The di ision o labou be ween he lexical and he auxilia y e b in he spli
and spli /doubled ypes a ies om language o language, and he e a e
di e en g oupings ha con as nega ion, TAM ( ense, aspec and modali y)
and subjec /objec ma king. Lea ing aside nega ion, he in lec ion on he
lexical e b and auxilia y e b ends o di e en ia e be ween subjec on he
one hand and objec on he o he , o be ween TAM on he one hand and
subjec /objec on he o he . As we shall see in chap e s 5 and 6, his second
pa e n is compa able o how ag eemen auxilia ies in LSE ope a e. Fo he
momen we e u n o he di e en ypes o a ge s ha can ma k ag eemen .
2.2.3.2. O he a ge s o ag eemen
In addi ion o e bal elemen s, ag eemen ma king may appea on o he
ca ego ies o wo d. This sec ion looks a hose ha a e ele an o ag eemen
48 Theo ies o ag eemen
in LSE. As we shall see in chap e 3, he spa ial ma king ha appea s on sign
language e bs is closely ela ed o he use o space also p esen in p onouns;
addi ionally, spa ial ma king in LSE may also appea on nume als,
quan i ie s, adposi ions and nouns, so ag eemen on hese ca ego ies in
spoken languages is desc ibed he e.
P onouns o en ma k he pe son, numbe o gende o hei an eceden ,
o a combina ion o hese ea u es. In he case o Tamil (D a idian), he
p onoun ma ks pe son, numbe and hono i ic s a us. In example (13), he
p onoun a a u ma ches he hi d pe son singula hono i ic an eceden mu al
man i i (‘Chie minis e ’).
Tamil (Ashe 1985: 4-7, ci ed in B own e al. 2002)
(13)
A u a
aa am
mu al
man i i
namma
uu -ukku
a - a-a u.
nex
week
i s
minis e
ou
illage-DAT
come-PRS-3SG.HON
a a u
a u a
aa am
a- u a-a u
PRO.3SG.HON
nex
week
come-FUT-3SG.HON
‘The Chie Minis e is coming o ou illage nex week.
He will come nex week.’
PRO=p onoun
This would no be ea ed as ag eemen unde many models, since he
p onoun does no ha e o be wi hin a ce ain dis ance o he an eceden wi h
which i ag ees (see sec ion 2.2.4, on domains). Such a model is binding heo y
(Chomsky 1981), in which a p onoun is aken o be subjec o speci ic
es ic ions ega ding i s ela ion and ela i e posi ion wi h espec o he
an eceden . Condi ion B o binding heo y (“a p onominal is ee in i s
binding domain”) basically equi es ha he an eceden does no appea in a
syn ac ic posi ion in which i con ols he p onoun, and hus no ela ion
holds. Howe e , p onouns display he same ag eemen ea u es (such as
numbe , pe son and gende ) ha ypically show up on ag eemen ma ke s.
Ba low (1999: 200) claims ha ag eemen and an eceden -anapho a ela ions
make use o he same unde lying mechanisms, e en hough he e may be
mo e going on in he case o anapho ic e e ence (conce ning dis ibu ion and
con ol). Co be (2006: 228-30) suppo s his iew by poin ing ou ha he e is
no ob ious o logical poin a which ag eemen phenomena can be
quali a i ely ca ego ized in o di e en ypes (e.g. local e sus anapho ic
ag eemen ), and a gues o a uni ied model o ag eemen . The ques ion o
ela ion o p oximi y equi ed be ween a a ge and i s con olle will be
looked a in he sec ion on domains, and e u ned o in he o e iew o
ag eemen in he Minimalis amewo k in sec ion 2.3.3.
Typological app oach 49
Ano he elemen ha can ma k ag eemen is nume als. I is common o
he nume al ‘one’, which is o en closely ela ed o he inde ini e a icle, o
show gende ag eemen wi h he noun con olle , bu highe nume als may
also show a iance in some languages. In Ca alan, he nume al ‘ wo’ ma ks
he gende o he con olle noun, as can be seen by he dis inc ion be ween
dugues wi h he eminine noun ampolla (‘bo le’) and dos wi h he masculine
noun bidon (‘can’) in he examples in (14).
Ca alan
(14) a.
dugues
ampolle-s
d’
aiga
wo.F
bo le(F)-PL
o
wa e (F)
‘ wo bo les o wa e ’
b.
dos
bidon-s
d’
aiga
wo.M
can(M)-PL
o
wa e (F)
‘ wo cans o wa e ’
This o ms pa o a gene al c oss-linguis ic pa e n: he lowe he nume al,
he mo e likely i is o be a a ge o ag eemen . Con e sely, he highe he
nume al, he less likely i is o show ag eemen , and highe nume als show
mo e i egula ag eeing pa e ns: in Russian, he nume al ‘ wo’ dis inguishes
he eminine om he masculine and neu e gende s, while ‘ h ee’ and ‘ ou ’
do no dis inguish gende bu do ag ee acco ding o animacy (Co be 1991,
1993).
Quan i ie s and ques ion wo ds may also show ag eemen wi h a noun
con olle . This occu s in Tu kana (Nilo-Saha an), spoken in Kenya, shown in
he examples in (15) by he al e na ion o he o m o he wo d ‘which’
depending on he gende o he con olle noun.
Tu kana (Dimmendaal 1983a: 433-434, ci ed in B own e al. 2002)
(15) a.
e-kìle
a-li`
M.SG-man
M.SG.which
‘Which man is i ?’
b.
ɪ-k kʊ
a-ni`
N.SG-child
N.SG.which
‘Which child is i ?’
Adposi ions may also show ag eemen and ypically ag ee wi h he noun hey
go e n, as occu s in he case o many mode n Indic languages, some o he
Iwaidjan languages (o No he n Aus alia) and he Cel ic languages (Co be
2006: 46). Example (16) shows ag eemen in pe son and numbe be ween he
50 Theo ies o ag eemen
Welsh p eposi ion am (‘abou ’) and he noun i go e ns (only in he case
whe e he noun is p onominal).
Welsh (adap ed om Sadle 2003, example 7)
(16) a.
Roedd
Wyn
yn
sia ad
amdana
( i).
was.3SG
Wyn
PROG
speak
abou .2SG
2SG
‘Wyn was alking abou you.’
b.
Roedd
Wyn
yn
sia ad
amdanom
(ni).
was.3SG
Wyn
PROG
speak
abou .1PL
1PL
‘Wyn was alking abou us.’
The las ca ego y we shall look a ha can ma k ag eemen a e nouns
hemsel es. As we ha e seen, nouns ypically con ol ag eemen on ano he
elemen such as a e b o adjec i e, bu hey may also be ma ked o show
ag eemen . Noun a ge s usually ag ee wi h some o he noun, and o en in
he con ex o a possessi e cons uc ion, in which he possesso ag ees wi h
he possessum, o ice e sa. In Palauan (Aus onesian), spoken in a ious
islands o he Wes e n Paci ic, he possessum ag ees wi h he possesso (17).
Palauan (Josephs 1975: 66-68, glossed ollowing Po e 1992 and Tibe ius 2002)
(17) a.
a
bli-l
a
D o eo
PM
house-3SG.POSS
PM
D o eo[SG]
‘D o eo’s house’
b.
u e -i
a
e-dil
wo k-3PL.POSS
PM
PL-woman
‘women’s wo k’
PM=ph ase ma ke
I is common o nouns o ag ee in numbe when hey a e p edica i e, so as o
a oid a seman ic misma ch. Howe e , i should be kep in mind ha he e a e
languages ha do no equi e a noun p edica e o ag ee in numbe , especially
when he subjec is inanima e o non-human. This is he case o Hunga ian,
which shows numbe ag eemen in (18a), bu no in (18b) o he non-human
subjec s.
Hunga ian (Hall 1944, ci ed in B own e al. 2002)
(18) a.
Molná
és
Kos olányi
Magya
i ó-k
Molna
and
Kos olányi
Hunga ian
w i e -PL
‘Molna and Kos olanyi a e Hunga ian w i e s.’
Typological app oach 51
b.
a
ehén
és
a
ló
leg-hasznosabb
háziálla
ART
cow
and
ART
ho se
mos -use ul
domes ic_animal
‘The cow and he ho se a e he mos use ul domes ic animals.’
So a we ha e seen which ypes o elemen s may ma k ag eemen , namely
e bs (including auxilia y e bs), adjec i es, p onouns, nume als, quan i ie s,
adposi ions and nouns. Now le us u n o he ma e o how hese a ge s
ma k ag eemen .
2.2.3.3. Means o exponence
The mos common way o mani es ing ag eemen is by means o a ixes. This
is wha we ha e seen in he examples so a : mainly su ixes, bu also p e ixes
in he case o gende ma king in Chicheŵa (4) and in ixes in he case o
numbe ag eemen wi h he (absolu i e) objec in Basque (2). A pa icula
ype o a ixal ag eemen has ecei ed some a en ion in he li e a u e:
alli e a i e ag eemen , common in Ban u and o he Nige Congo languages.
Alli e a i e ag eemen in ol es he p esence o a ma ke on he con olle and
he same ma ke is used o ag eemen a ge s. In he Swahili (Cen al Ban u,
Nige -Congo) example in (19), he singula ma ke o gende 7/8, ki-, is
epea ed on he adjec i e, he nume al and he e b ha ag ee wi h he
subjec .
Swahili (Welme s 1973: 171, ci ed in Co be 2006: 87)
(19)
ki-kapu
ki-kubwa
ki-moja
ki-li-anguka
SG-baske (G7/8)
G7-la ge
G7-one
G7-PST- all
‘One la ge baske ell.’
The in e es ing cases a ise when he con olle lacks a gende ma ke , which
is a common si ua ion o loan wo ds. Wi h no ma ke on he con olle he e
is no hing a ailable o be e-used o ma k ag eemen on he a ge s. In his
case, wo op ions a e a ailable: in some cases a de aul ma ke is used, bu in
o he s he ma ke copies pa o he phonological o m o he con olle . An
example o his ype o adical o li e al alli e a i e ag eemen is shown in (20)
o Bainouk (A lan ic, Nige -Congo): (20a) shows ypical alli e a i e
ag eemen o a noun ha bea s a gende ma ke ; in (20b), on he o he hand,
he noun has no gende ma ke and he nume al uses he second s a egy by
copying he i s consonan and owel (CV) o ma k ag eemen . Such adical
alli e a i e ag eemen has been claimed o exis in o he languages such as
A apesh (To icelli), Wolo (Nige -Congo) and Landuma (Landoma, Nige -
Congo), bu e idence o a obus mechanism is scan (Co be 2006: 90).
52 Theo ies o ag eemen
Bainouk (Sau ageo 1967, ci ed in A ono , Mei & Sandle 2005: 320)
(20) a.
gu-s l
gu- ɛ
G7- unic
G7-whi e
‘whi e unic’
b.
ka a:m-ã
ka-nak-ã
i e -PL
CV- wo-PL
‘ wo i e s’
The impo ance o adical alli e a i e ag eemen lies in i s unde mining o he
p inciple ha syn ax is phonology- ee, since he (syn ac ic) ag eemen
p ocess needs o know abou he phonological o m o he con olle . I will
no en e in o his deba e, bu no e ha he phenomena is ele an o sign
language ag eemen as pa allels ha e been d awn be ween adical alli e a i e
ag eemen and ag eemen in sign languages (A ono , Mei & Sandle 2005).
Some e bal a ixes all in o a g ey a ea wi h espec o ag eemen , and
he e is a di ision o opinion as o whe he o no hey cons i u e a
mani es a ion o ag eemen p ope . These a e bound ma ke s o pe son,
numbe and gende ha a e ea ed as p onouns a ached o he e b a he
han ag eemen ma ke s. The dis inc ion is o en a di icul one o make, bu is
o consequence since om a syn ac ic poin o iew a p onoun is an a gumen
o a e b, whe eas a e b ma ked o ag eemen has independen a gumen s
( hough hey may no be o e , as in he case o p o-d op languages). Co be
(2006: 110) poin s ou ha he di e ence is also impo an i ag eemen is
es ic ed o he clausal le el, since p onouns may be indexed (on his iew
hey do no “ag ee”) wi h an eceden s ou side he immedia e clause, whe eas
ag eemen ma ke s mus s and in a local ela ion o hei con olle s. The
Aus alian language Bininj Gun-Wok [Gunwinggu] (Aus alian) makes use o
p onominal a ixes, shown by he p e ix gaban- on he e b (‘scold’) in
example (21).
Bininj Gun-Wok: Gun-djeihmi dialec (E ans 1999: 266, ci ed in Co be 2006: 104)
(21)
al-ege
daluk
gaban-du-ng
bedda
F-DEM
woman
3SG>3PL-scold-NPST
hem
‘Tha woman is scolding hem.’
These p onominal a ixes a e hyb id elemen s as hey a e mo phologically
bound o he e b, ye a he same ime hey a e a gumen s o he e b and
e e en ially hey unc ion much like independen p onouns (Mi hun 2003). In
his sense hey all be ween ypical ag eemen ma ke s and ee p onouns,
Typological app oach 59
b.
JON
IXx
MIREN
IXy
PROBLEM
xEXPLAINy
‘Jon explained he p oblem o Mi en.’
We ha e es ablished ha he ypical domain o ( e bal) ag eemen is
be ween a e b and i s a gumen s, bu he e a e a es ed cases o ag eemen
whe e he e b is con olled by an elemen ha is no one o i s a gumen s.
The i s case is possesso ag eemen , in which he e b ag ees wi h he
possesso o an a gumen a he han he a gumen i sel .
4
We ha e al eady
seen an example o his in Bu ushaski, epea ed he e as (28). The auxilia y
e b bea s he i s pe son ma ke mée- o indica e who he walnu s belong o,
a he han indica ing ag eemen wi h he walnu s hemsel es.
Bu ushaski (Be ge 1998: 162, 161, ci ed in Ande son & Egge 2001: 240, 242)
(28)
kʰɑkʰɑ ɑy-umuc
pʰɑ
mée- -ɑɑ
walnu -pl
gobble_up
1PL-AUX-2
‘You gobbled up ou walnu s.’
This ag eemen be ween he e b and he possesso o an a gumen occu s in
se e al languages, such as Mai hili (Indo-Eu opean), Banawá (Jamamadí,
A auan), Tabasa an (Daghes ian) and Fox [Meskwaki] (Algic) (Co be
2006: 61; Ande son 1997: 234). Fo his ype o s uc u e, i has been claimed
ha he possesso unde goes mo emen o a posi ion ypically occupied by
e bal a gumen s (“possesso aising”), hus p o iding he syn ac ic
jus i ica ion o he mani es a ion o he ag eemen ela ion (Landau 1999).
Howe e , such aising analyses a e uled ou o ce ain languages, such as
Mai hili, on he g ounds o wo d o de conside a ions: in spi e o ela i ely
ee wo d o de possibili ies in he language, he possesso canno be
ex ac ed om he majo cons i uen o which i o ms pa . This shows ha
syn ac ically he possesso mus be pa o he con aining possessum NP
a he han a sepa a e majo cons i uen in he clause and hus i is ha d o
know wha syn ac ic ela ion be ween he possesso and he e b cons i u es
he ag eemen domain (Com ie 2003: 335).
Con e sely, he ag eemen con olle does appea o occupy an
a gumen posi ion o he e b in he copying- o-objec o ma ions desc ibed
by Ande son (1997: 231-233). In hese cons uc ions he a gumen o an
embedded clause becomes he objec o he ma ix e b. In he example in
4
Possesso ag eemen be ween a e b and he possesso o one i s a gumen s is o be
dis inguished om ag eemen be ween possesso and possessum, as men ioned in sec ion
2.2.3.2. Recall ha his discussion o domain ocuses on e bal ag eemen , and as such
‘possesso ag eemen ’ he e e e s o ag eemen wi h a possesso on he e b (no on a
possessum).
60 Theo ies o ag eemen
(29a) om he No h Ame ican language Fox, he e b ‘wan ’ could ake a
clausal objec (which would be ea ed as an inanima e noun) bu in lec s o a
hi d pe son (anima e) objec . Ande son claims ha his is a copying
mechanism a he han some so o aising e ec since he e e en is
signalled wice and bo h clauses a e po en ially ee-s anding. F om he poin
o iew o ag eemen domains, he e is no hing oo unusual going on he e:
once he a gumen is copied o he objec posi ion in he ma ix clause, he
e b ma ks ag eemen wi h i . Howe e , Ande son epo s cases in which he
elemen ha igge s ag eemen on he ma ix e b has li le o do wi h he
embedded e b’s a gumen s uc u e. In (29b), he embedded i s pe son
opic is wha con ols ag eemen on he ma ix e b.
Fox [Meskwaki] (adap ed om Ande son 1997: 232, 233)
(29) a.
ne -ɑ ɑ:wɑ:n -ɑ:-wɑ=koh(i)
wi:h=ne:w-ɑ
1-wan -DTS(1 >)-3.IIND=you_know
FUT=see-1>3.AOR
Li . I wan him ( ha ) I will see him.
‘I do wan o see him.’
b.
ne-kehke:nem-ek(w)-wɑ
ni:nɑ
e:h=pwɑ:w -ke:ko:hi-ɑšeno-niki
l-know-INV-3(>1).IIND
1.TOP
AOR=no -any hing-disappea -INAN.OBV.AOR
Li . He knows me ha as o me no hing is missing.
‘He knows ( ha ) as o me no hing is missing.’
DTS=di ec heme sign, INV=in e se, IIND=independen indica i e,
AOR=ao is , INAN=inanima e, OBV=ob ia i e
(29b) could pe haps be ea ed as a combina ion o possesso aising in he
embedded clause and copying- o-objec in o he ma ix clause. Seman ically,
he possesso ag eemen cons uc ion is eminiscen o he da i e o in e es
common in Romance languages bu a he han occupying an oblique
a gumen posi ion as occu s in (30), he e e en in he possesso ag eemen
and copying- o-objec o ma ions is ma ked as an immedia e a gumen on he
e b.
Spanish
(30)
Es e
niño
no
me/ e/le
com-e
nada.
This
child
NEG
me/you/him/he .DAT
ea -3SG
no hing
‘This childi will ea no hing (and I’m/you’ e/(s)hej’s a ec ed).’
Toge he wi h possessi e ag eemen and o he da a, examples like (29b) a e
used by Ande son o a gue ha ag eemen in Fox is condi ioned by discou se-
d i en conside a ions o do wi h highligh ing anima e e e en s salien o he
discou se. Ini ially, we had desc ibed hese ag eemen -wi h-non-a gumen
Typological app oach 61
phenomena – possesso ag eemen and copying- o-objec o ma ions – as
being clause in e nal, bu Ande son’s line o easoning pushes he limi s back
beyond he clause (and he sen ence) o he le el o discou se.
5
Unde his
iew, ag eemen is (o , pu mo e cau iously, in some languages may be)
sensi i e o e ec s ha a e no con ained wi hin he mo phosyn ac ic domain
and canno be desc ibed in e ms o (syn ac ic) s uc u al ela ions. A mo e
adical iew o “ag eemen as a (pu ely) discou se phenomenon” is o e ed by
Ba low (1999), who sugges s ha he ela ionship be ween con olle s and
a ge s canno be educed o ea u e iden i y and is be e cap u ed as a
me ging o in e p e a ions associa ed wi h discou se e e en s. The
cha ac e iza ion o ag eemen in e ms o discou se conside a ions si s in s a k
con as wi h he minimalis ack, which conside s he domain o ag eemen
o be wi hin na ow syn ac ic s uc u e (see sec ion 2.3.3).
The las unusual ag eemen e ec ha we shall look a is a clea case o
ag eemen beyond he clause, ega dless o he ole gi en o discou se
conside a ions. Long dis ance ag eemen has been a es ed o a ious
languages such as Godobe i (No h Caucasian) (Haspelma h 1999), Basque
(Isola e) (E xepa e 2005), and Lokaa (C oss Ri e , Nige -Congo) (Bake 2008).
Example (31) comes om he Daghes anian language Tsez [Dido] (No h
Caucasian), showing how he ma ix e b ‘know’ ag ees in gende wi h he
objec o he embedded clause, ‘b ead’.
Tsez [Dido] (Polinsky & Po sdam 2001: 584)
(31)
eni
už
magalu
b c’ uł
b-iyxo
mo he
boy
b ead(G3).ABS
a e
G3-know
‘The mo he knows he boy a e he b ead.’
I can be shown ha mo emen in Tsez is s ic ly clause in e nal and he e is
no way o an a gumen o mo e ou side o i s clause. This means ha he
ag eemen be ween he e b and he embedded elemen mus be ac oss a
clause bounda y. Polinsky & Po sdam (2001: 641) conclude ha his makes
long dis ance ag eemen “p oblema ic o heo ies o ag eemen ha ei he
explici ly s ipula e o axioma ically de i e he claim ha all ag eemen
ela ionships a e clause bounded.”
To summa ize his sec ion, he domain o ag eemen is he ela ionship
ha holds be ween he wo elemen s in ol ed in he ag eemen con igu a ion,
5
Ano he phenomenon which ecommends he impo ance o aking in o accoun (anima e)
e e en s salien o he discou se is allocu i e ag eemen in Basque, in which he e b ag ees
wi h he add essee in pe son and gende e en hough he e e en is no an a gumen selec ed
by he e b (Oyha çabal 1993).
62 Theo ies o ag eemen
he con olle and he a ge , and ypically his is de ined in e ms o
g amma ical unc ion (“ he e b ag ees wi h i s subjec ”). We ha e looked a
a ious ag eemen phenomena ha esis any a emp o iden i y a
s aigh o wa d connec ion be ween he con olle and he a ge : in possesso
ag eemen , he e b ag ees wi h he possesso o i s logical objec ; copying- o-
objec o ma ions show he e b ag eeing wi h a discou se salien e e en
(which plays no pa in he e b’s a gumen s uc u e); and long dis ance
ag eemen allows he ma ix e b o ag ee wi h he a gumen o a di e en
e b in an embedded clause. One eac ion o hese unusual ag eemen e ec s
is o say ha ag eemen is sensi i e o discou se ac o s o , mo e adically,
ha i ope a es en i ely a he le el o discou se. An al e na i e s a egy, aken
up by he Minimalis P og am, is o cha ac e ize he domain o ag eemen in
pu ely s uc u al e ms, limi ing he desc ip ion o he ela ion be ween
con olle and a ge o syn ac ic con igu a ions.
2.2.5. Fea u es and alues
Fea u es a e he in o ma ion om he con olle ha is ma ked on he a ge
in he ag eemen p ocess. Mo e p ecisely, a ea u e is he ype o in o ma ion
ha is ma ked and he speci ic in o ma ion shown is he alue; o example, a
e b may ag ee in numbe ( he ea u e) wi h i s objec and in a gi en case ha
may be dual ( he alue). A ea u e has a se o possible alues ha a ies om
language o language: in he case o No he n Os yak [Khan y] (U alic),
numbe may be singula , dual o plu al, as can be seen om he di e en
ma ke s on he e b we:l (´kill´) in (32).
No he n Os yak [Khan y] (Nikolae a 1999: 334)
(32) a.
ma
am
kalaŋ
we:l-s-Ø-e:m
I
his
eindee
kill-TENSE-SG-1SG.SBJ
‘I killed his eindee .’
b.
ma
am
kalaŋ
we:l-sə-ŋil-am
I
hese
eindee
kill-TENSE-DU-1SG.SBJ
‘I killed hese ( wo) eindee .’
c.
ma
am
kalaŋ
we:l-sə-l-am
I
hese
eindee
kill-TENSE-PL-1SG.SBJ
‘I killed hese eindee .’
The main ea u es o ag eemen , he ϕ- ea u es, a e gende , numbe and
pe son. The e a e o he ea u es ha may be conside ed in he ag eemen
p ocess, such as case and espec . Be o e looking a each o hese ea u es and
Typological app oach 63
hei se s o possible alues, I wish o look a he gene al p ope ies o ea u es
hemsel es.
A ea u e may be ca ego ised in e ms o whe he o no i is equi ed by
he syn ac ic con ex , in much he same way ha in lec ion may be conside ed
inhe en o con ex ual (c . Booij 1996). Applying his dis inc ion o ea u es, a
con ex ual ea u e is one ha is equi ed by he syn ac ic con ex , while an
inhe en ea u e is no , al hough i may be ele an o he syn ac ic sys em.
This p ope y o a ea u e depends upon whe e he ea u e appea s: gende is
inhe en o nouns, bu con ex ual o adjec i es (Co be 2006: 123).
A u he dis inc ion (due o Zwicky 1992, ci ed in Co be 2006: 124)
may be d awn wi h espec o how a ea u e ela es o seman ics. Di ec
ea u es exp ess in insic con en and a e associa ed wi h p o o ypical
seman ics (numbe wi h nume osi y, gende wi h a classi ica ion o objec s,
e c.). Indi ec ea u es, on he o he hand, exp ess meanings indi ec ly, by
means o g amma ical ela ions (nomina i e case is associa ed wi h he
g amma ical ela ion o subjec , which in u n is associa ed wi h he seman ic
ole o AGENT). These wo p ope ies coincide closely bu a e no he same:
he i s ela es o syn ax (and as such depends upon whe e he ea u e
appea s), while he second is o do wi h seman ics. I is possible o a ea u e
o be inhe en and indi ec and his is o impo ance o he heo e ical
appa a us o Minimalism (c . unin e p e able ea u es in sec ion 2.3.3).
Ano he impo an p ope y o ea u es is ha hey ep esen a ce ain
le el o abs ac ion. A ea u e’s alues ac as ma ke s ha ca ego ize nominal
elemen s as belonging o a gi en ca ego y, such as plu al in numbe , o
masculine in gende . These alues ha e a seman ic basis and se e as a means
o ca e up he wo ld o linguis ic elemen s in o di e en ypes, which is
made use o by he g amma ical sys em. As such, many di e en i ems may
sha e he same ϕ- ea u es (and alues). This means ha ϕ- ea u es do no
uniquely indi idua e speci ic i ems, and hey a e o be con as ed wi h
indices, which a e speci ic labels o a single i em. The dis inc ion will be an
impo an one when we come o look a he ea u es a wo k in ag eemen in
LSE. We now u n o he di e en ea u es ha play a ole in ag eemen .
2.2.5.1. Gende
Gende is an inhe en ea u e o he noun and ca ego izes i acco ding o some
so o seman ically based axonomy. The be e known gende sys ems ha e
wo o h ee alues: masculine, eminine and neu e . Howe e , o he
languages, such as he Ban u languages, ha e mo e in ol ed gende sys ems
ha no mally dis inguish be ween se en and en gende s (o classes, as hey
a e known in he Ban uis adi ion). In he ex eme, Nige ian Fula (Nige -
64 Theo ies o ag eemen
Congo) has a ound wen y gende s, depending on he dialec (Co be 2008a).
Gende is an unusual ea u e in ha i may o may no igu e in a language’s
g amma : in a su ey o 257 languages o he Wo ld A las o Language
S uc u es (WALS – Haspelma h, D ye , Gil & Com ie 2008), Co be (2008a)
inds ha o e hal (145) ha e no gende sys em. Unsu p isingly, when a
gende sys em is p esen , he mo e dis inc ions i makes, he less common i
is.
Gende sys ems may be based on pu ely seman ic c i e ia o a
combina ion o seman ic and o mal c i e ia. In he i s case, he ca ego ies
es ablished by he sys em a e di ec ly ela ed o he meaning o he membe s
in each ca ego y, as is he case o Kannada (D a idian) o which all male
humans a e masculine gende , all emale humans a e eminine, and
e e y hing else is neu e (Co be 2008c). Al e na i ely, he gende sys em
may ha e a seman ic base supplemen ed by o he c i e ia ha esul in
ca ego ies wi h a mixed se o membe s ha do no seem o o m a na u al
class o any so . The addi ional assignmen ules o en ake in o accoun he
o m o each noun, and his may be done on he basis o phonology o
mo phology. The same WALS su ey o gende sys ems ound a oughly
equal spli be ween s ic ly seman ic and mixed seman ic/ o mal gende
sys ems (Co be 2008c).
A inal conside a ion o gende sys ems is he unde lying dis inc ion o
he seman ic c i e ia o assigning gende . The majo i y o languages ha
ha e a gende sys em in he WALS su ey (84 o 112) applies a sex-based
ca ego iza ion and he emaining languages make use o animacy as he basic
di e en ia ing ac o . In he mos limi ing case, animacy is es ic ed o
humans bu i may also be ex ended o animals and spi i s o ees. Many o
he languages ha make use o an animacy-based gende sys em a e om he
Nige -Congo and Algonquian amilies, bu his ype o language is also
ep esen ed all o e he wo ld (Co be 2008b).
In summa y, gende is an inhe en ea u e o nouns ha s ems om a
seman ic classi ica ion based on ei he sex o animacy, which may be
obscu ed by addi ional o mal c i e ia o gende assignmen . No all
languages ha e gende sys ems, and he ex en o which gende pa icipa es
in ag eemen p ocesses may a y om language o language: gende
ag eemen is gene ally limi ed o he nominal domain, bu may also play a
ole in e bal ag eemen . The ollowing sec ion looks a he ea u e o
numbe .
Typological app oach 65
2.2.5.2. Numbe
Numbe is an inhe en ea u e o nouns
6
and i s alue is no mally ma ked on
he noun (e.g. a plu al ma ke ), bu some nouns may be lexically speci ied o
numbe . Mos commonly, he numbe ea u e dis inguishes wo alues:
singula and plu al. The plu al ca ego y may be u he spli in o mo e
speci ic alues. Dual ma ks wo and only wo e e en s in Uppe So bian
(Indo-Eu opean) (Co be 2000: 20) and Hopi (U o-Az ecan) (Mo a csik 1978:
347), hus es ic ing he plu al alue o h ee o mo e. T ial ma ks h ee and
only h ee e e en s and appea s in languages ha ha e a dual o m, such as
La ike [La ike-Waksishu] (Cen al Malayo-Polinesian, Aus onesian) (Co be
2000: 21). The paucal is used o e e o a small numbe o dis inc e e en s: in
Bayso (A o-Asia ic) he paucal e e s o be ween wo and six indi iduals
(Co be 2000: 22). The paucal and he gene al plu al may also be di ided in o
lesse and g ea e ca ego ies, wi h he esul ha languages may ha e up o
i e di e en alues o he numbe ea u e: Mele-Fila (Cen al-Eas e n
Oceanic, Aus onesian) dis inguishes be ween singula , dual, paucal, plu al
and g ea e plu al (Co be 2000: 42).
In addi ion o hese numbe dis inc ions, some languages ha e a gene al
alue ha is ou side o beyond he numbe sys em. A language may ha e a
speci ic o m ha is neu al wi h espec o numbe . In he Fou a Jalon dialec
o Fula (Nige Congo), he wo d o i may e e o one o se e al oads, and
con as s wi h he o ms o he singula , o ii- u (‘a oad’), and plu al, o ii-ji
(‘ oads’). Many languages ha e a gene al meaning bu a he han use a
sepa a e o m, his is achie ed by means o one o he o ms o mo e
es ic ed numbe meaning. Thus, in Tu kish e can mean ‘house’ o ‘houses’,
while he plu al o m e le always means ‘houses’ (Co be 2000: 10-14). This
no ion o gene al numbe will u n ou o be ele an when in e p e ing he
LSE da a, especially gi en he appa en op ionali y o numbe ma king.
Numbe as a nominal ea u e needs o be dis inguished om e bal
numbe . We a e used o hinking o numbe as being a ea u e ha a e b
ag ees wi h ( ha is, is exp essed as a con ex ual ea u e on he e b as a
a ge ), bu a e b may ha e numbe as an inhe en ea u e. Ve bal numbe
e lec s he e en seman ics o he e b and indica es whe he an ac ion is
pe o med se e al imes o a se e al places (e en numbe ), o whe he i
a ec s o in ol es se e al pa icipan s (pa icipan numbe ). In his sense,
e bal numbe may o e lap wi h bo h aspec (i e a i e and dis ibu i e) and
6
The no ion ha numbe (and gende ) is inhe en o a noun ac ually depends on how his
in o ma ion is ep esen ed in he syn ac ic s uc u e. I e u n o his issue in sec ion 7.1.1
when e-examining he loca ion o ϕ- ea u es.
66 Theo ies o ag eemen
nominal numbe as e lec ed in ag eemen (Veselino a 2008; Co be 2000:
256). Howe e , i is o en possible o dis inguish e bal numbe as a
g amma ical ca ego y ha is ma ked on he e b. In Geo gian (Ka elian),
he o m o he e b s em ‘si ’ ma ks he plu ali y o he ac ion (one pe son
si s s. se e al people si ). The e b also bea s an ag eemen ma ke ha
indica es he numbe o he con olle a gumen . The con as be ween (33a)
and (b) highligh s he di e en e bal o ms associa ed wi h a singula subjec
a gumen /single e en (i.e. singula e bal numbe ) and wi h a plu al subjec
a gumen /mul iple e en (i.e. plu al e bal numbe ). Howe e , i is possible
o dis inguish be ween nominal (a gumen ) numbe and e bal numbe due
o he ac ha in Geo gian nume als equi e a singula noun and con ol
singula ag eemen on he e b. Thus, in (33c), wi h he nume ically
quan i ied subjec ‘my h ee iends’, he e b is ma ked o show ag eemen
wi h a singula subjec by he a ix -a, simila ly o (33a), bu also con ains he
a ix -sxd-, simila ly o (33b), o ma k plu al e bal numbe .
Geo gian (A onson 1982: 243, 406-7, ci ed in Co be 2000: 254)
(33) a.
i ane
še-mo- id-a
da
da-ǰd-a
John
PRV-PRV-en e -AOR.3.SG
and
PRV-si .SG-AOR.3.SG
‘John en e ed and sa down.’
b.
čem-i
mšobl-eb-i
še-mo- id-nen
da
da-sxd-nen
my-AGR
pa en -PL-NOM
PRV-PRV-en e -AOR.3.PL
and
PRV-si .PL-AOR.3.PL
‘My pa en s en e ed and sa down.’
c.
čem-i
sam-i
megoba -i
še-mo- id-nen
da
my-AGR
h ee-AGR
iend.SG-NOM
PRV-PRV-en e -AOR.3.PL
and
da-sxd-a
PRV-si .PL-AOR.3.SG
‘My h ee iends en e ed and sa down.’
PRV=p e e bal
Ve bal numbe is ma ked on he e b and is an inhe en ea u e o he e b.
As such, i does no ep esen a case o ag eemen since he e is no displaced
in o ma ion. Howe e , i is ele an o ag eemen because, as we ha e seen,
he e b may also ca y ag eemen ma ke s ha e lec he numbe ea u e
alue o one (o a ious) o i s a gumen s. This is ue o Geo gian, and i
was ela i ely s aigh o wa d o dis inguish he wo phenomena. Howe e , i
is no always so easy o di e en ia e be ween e bal numbe and ag eemen
ma ke s. Du ie (1986: 357-62) p o ides he ollowing diagnos ics:
Typological app oach 67
i) e bal numbe ope a es on an e ga i e basis, e lec ing he numbe
o he mos di ec ly a ec ed pa icipan , which is he subjec o
in ansi i e sen ences (S) o he objec o ansi i e sen ences (P), and
his may con as wi h o he ma king on he e b (e.g. subjec
ma king, which ag ees wi h S and A).
ii) e bal numbe may ma k di e en alues o hose ma ked by
ag eemen , especially when e bal ag eemen is es ic ed by some
condi ion (such as singula ag eemen o nume al ph ases in
Geo gian).
iii) e bal numbe may ha e a di e en se o alues o nominal
numbe ; al hough a e, i is possible o e bal numbe o include a
alue (such as dual) ha is no ma ked by nominal numbe in he
same language, o ice e sa.
i ) e bal numbe is e ained in con ex s whe e ag eemen is absen ,
namely non- ini e o ms ha lack ag eemen mo phology such as
con ol cons uc ions, impe a i es and a ibu i e usage.
) e bal numbe is p ese ed in de i a ional wo d o ma ion, bu
ag eemen in lec ion o (nominal) numbe is no .
These di e ences be ween e bal numbe and nominal numbe ma ked on
he e b se e o iden i y how much o he in lec ional ma e ial on a e b is
due o ag eemen (and con e sely, how much is ma king inhe en ea u es o
he e b). This will be use ul o delimi ing e bal ag eemen in LSE.
Numbe di e s om gende due o he ac ha i is held o be
uni e sally p esen in all languages: Uni e sal #42 p oposed by G eenbe g
(1963: 113) s a es “All languages ha e p onominal ca ego ies in ol ing a
leas h ee pe sons and wo numbe s.” This is no qui e ue. Fi s ly, he e a e
languages ha make no g amma ical dis inc ion in numbe . Co be (2000: 50-
51) men ions Kawi (Aus onesian), Classical Chinese (Sino-Tibe an) and
Pi ahã (Mu a) as examples o languages ha ha e no plu al nouns o
p onouns ( hough seman ic numbe may be exp essed by means o
conjunc ions and quan i ie s). Example (34) shows how he hi d pe son
p onoun is used indis inc ly in Pi ahã o singula o plu al e e ence.
Pi ahã (E e e 1986: 282, ci ed in Co be 2000: 51)
(34)
hiapióxio
soxóá
xo-ó-xio
3
al eady
jungle-LOC-DIR
‘He al eady wen o he jungle.’ o
‘They al eady wen o he jungle.’
DIR=di ec ional
68 Theo ies o ag eemen
A second p oblem wi h G eenbe g’s uni e sal is ha i is couched in e ms o
p onominal ca ego ies and he ques ion o wha coun s as a p onominal c oss-
linguis ically is a ho ny issue. This is closely connec ed o he ca ego y o
pe son so we shall look mo e closely a his ma e in he ollowing sec ion.
2.2.5.3. Pe son
Pe son is a ea u e inhe en o p onominal elemen s a he han nouns, which
a e aken o be hi d pe son by de aul .
7
Acco ding o G eenbe g’s Uni e sal
#42, pe son is a uni e sal ea u e o he p onominal ca ego y and always
dis inguishes be ween h ee di e en alues: i s ( he speake ), second ( he
add essee) and hi d (nei he speake no add essee). Since he g amma ical
ca ego y o pe son is closely linked o he ca ego y o pe sonal p onoun, he
de ini ion o wha coun s as a pe sonal p onoun is cen al o assessing he
uni e sali y o pe son. Cysouw (1997) claims ha some languages, such as
Wes -G eenlandic Inui (Eskimo-Aleu ), lack a hi d pe son p onoun, making
use ins ead o a demons a i e. Cysouw’s de ini ion o a hi d pe son
p onoun is as an in e subjec i e deic ic: “an i em ha can be used by all
speech-ac pa icipan s alike o e e o some hing” (1997: 9). Since
demons a i es a e no in e subjec i e (‘ his’ o me may be ‘ ha ’ o you,
while ‘he’ emains cons an o me and you) he concludes ha hey a e no
p onominal and hus ha hi d pe son p onouns a e no a uni e sal ca ego y.
In a simila ein, i p onouns a e de ined as a mo phosyn ac ic ca ego y, he e
a e languages such as Thai (Tai-Kadai) and Japanese (Japonic) whose pe son
ma ke s beha e mo e like nominals han p onominals, wi h he esul ha i
is no clea ha all languages ha e a p onominal ca ego y (Siewie ska 2004:
9).
The e a e wo esponses o hese claims. The i s is o use an al e na i e
de ini ion o p onouns in e ms o e e en ial ole and unc ions which ocuses
on hei e e en ial de iciency ( o dis inguish hem om nouns) and anapho ic
7
Howe e , some languages do allow non- hi d pe son ag eemen wi h a nominal ph ase.
Spanish is o en ci ed in his espec because o examples like he ollowing:
(i) a.
¿El
g upo
en e -o
-áis?
DEF.M.SG
g oup(M.SG)
whole-M.SG
go-2PL
‘Is he whole g oup (o you) going?’
b.
Los
ma ica-s
abund-amos
en
es -e
campo.
DEF.M.PL
poo (M)-PL
abound-1PL
in
his-M
ield(M)
‘(Us) poo s a e in abundance in his ield.’
Typological app oach 75
sho comings, wo o which I will add ess. Fi s ly, om a p ac ical poin o
iew, es ablishing he no m would be a huge unde aking as i would in ol e
aking s ock o he ag eemen sys ems o all he wo ld’s languages. Secondly,
a s a is ical app oach could also un in o p oblems due o he le elling na u e
o a e ages. To gi e a acile illus a ion, ake gende as a case in poin . Using
he igu es o languages wi h di e en numbe s o gende alues om he
WALS sample o 257 languages (Co be 2008a), a ough calcula ion gi es an
a e age o a ound 2 gende s. Ye he mos common ca ego y is o a language
no o ma k gende a all. Equally, a simple s a is ical a e age ails o cap u e
he in e ac ion be ween di e en ac o s: con inuing wi h he gende example,
he ac ha non-sex-based gende sys ems end o ha e many mo e alues
han sex-based ones would be los in an a e aging p ocess. Ob iously, hese
e o s could be o e come by imp o ed s a is ics (in hese cases using he
mode ins ead o he mean, and mo e complex a iance s a is ics), bu he e
emains an unde lying p oblem o ailing o cap u e he ull ex en o he
ag eemen phenomena. Es ablishing a s a is ical no m ails o delimi he
ex en o he phenomenon and only gi es us an image o he mos popula ed
pa o he ag eemen e ain a he han he pe iphe ies.
An al e na i e app oach, based on he no ion o canonici y, examines
he di e en ways in which ag eemen can a y and s ipula es c i e ia based
on hese a iables. The c i e ia lay ou he di e en op ions o ag eemen
sys ems and hus p o ide a mapping o he heo e ical space o possibili ies.
Fo each a iable a canonical alue is designa ed in acco dance wi h gene al
p inciples ha a e deemed o cha ac e ize (canonical) ag eemen . This means
ha he mos canonical sys em is he one ha bes con o ms o he gene al
p inciples and is no necessa ily he mos commonly occu ing sys em among
he wo ld’s languages. Each o he c i e ia p o ides a pa ame e o e alua e a
gi en ag eemen sys em agains he p o o ypical ag eemen sys em.
Co be (2003b, 2006: 10-27) de elops a canonical app oach ha I will
adop he e. I limi mysel o lis ing Co be ’s p inciples and c i e ia, p o iding
explana ion only whe e he e minology demands i . Reade s in e es ed in he
mo i a ion and jus i ica ion o Co be ’s choices should e e o his wo k. The
gene al p inciples o canonical ag eemen a e as ollows:
P inciple I: Canonical ag eemen is edundan a he han in o ma i e.
P inciple II: Canonical ag eemen is syn ac ically simple.
P inciple III: The close he exp ession o ag eemen is o canonical (i.e.
a ixal) in lec ional mo phology, he mo e canonical i is as
ag eemen .
76 Theo ies o ag eemen
These p inciples dic a e wha he mo e canonical alue is o he di e en
pa ame e s, which a e se ou in able 1.3.
C-1:
con olle is p esen
>
con olle is absen
C-2:
con olle has o e exp ession o
ag eemen ea u es
>
con olle has co e exp ession o
ag eemen ea u es
C-3:
consis en con olle (all a ge s
ake he same alue o a gi en
ea u e)
>
hyb id con olle ( a ge s ake
di e en alues o a gi en ea u e)
C-4:
con olle ’s pa o speech is
i ele an
>
con olle ’s pa o speech is
ele an
C-5:
ma king is bound
>
ma king is ee
C-6:
ma king is obliga o y
>
ma king is op ional
C-7:
ma king is egula (a ixal)
>
ma king is supple i e
C-8:
ma king is alli e a i e (ma ke on
all a ge s is he same and
iden ical o o man on con olle )
>
ma king is opaque (ma ke changes
om a ge o a ge and is no
iden ical o o man on con olle )
C-9:
ma king is p oduc i e (applies o
all membe s o a ca ego y)
>
ma king is spo adic (only appea s
on some membe s o a ca ego y)
C-10:
a ge always ag ees
>
a ge ag ees only when con olle
is absen
C-11:
a ge ag ees wi h single
con olle
>
a ge ag ees wi h mo e han one
con olle
C-12:
a ge has no choice o con olle
>
a ge has choice o con olle
C-13:
a ge ’s pa o speech is
i ele an
>
a ge ’s pa o speech is ele an
C-14:
domain is asymme ic
>
domain is symme ic
C-15:
domain is local
>
domain is non-local
C-16:
domain is one o a se
>
single domain
C-17:
ea u e is lexical
>
ea u e is non-lexical
C-18:
ea u es ha e ma ching alues
>
ea u e alues do no ma ch
C-19:
no choice o ea u e alue
>
choice o ea u e alue
C-20
no condi ions
>
condi ions
Table 1.3 C i e ia o canonical ag eemen . The symbol > means “is mo e canonical
han”. Adap ed om Co be (2006: 10-27).
These c i e ia p o ide a gauge o how canonical ag eemen in a gi en
language is, and may be applied o he ag eemen sys em o a language as a
whole, o o speci ic aspec s o ag eemen o ha language. Thus, o a gi en
language e b ag eemen may be s ongly canonical while DP-in e nal
ag eemen is less so. I shall apply his no ion o ag eemen in LSE o gi e us
an idea o whe he o no LSE has an ag eemen mechanism and, mo e
Typological app oach 77
gene ally, whe he wha has been ea ed as ag eemen in he sign language
li e a u e is jus i iably labelled as such.
2.2.8. Summa y
In his sec ion, we ha e looked a ag eemen om a ypological poin o iew,
su eying he phenomenon om he pe spec i e o he di e si y o i s
mani es a ions ac oss he wo ld's (spoken) languages. In o de o
accommoda e his a ia ion, no igid de ini ion o ag eemen is s ipula ed, bu
a he a e minological amewo k ha can desc ibe he di e en ypes o
ag eemen ha a e a es ed. This desc ip i e app oach iden i ies he di e en
elemen s ha en e in o play in he ag eemen ela ionship, and we ha e
examined each o hese in u n: con olle s, a ge s, ea u es, domains and
condi ions.
Con olle s, he elemen s ag eed wi h, a e gene ally nominal elemen s
(in he case o e bal ag eemen nominal ph ases), which may o may no be
o e ly p esen .
Ta ge s a e he elemen s ha ag ee wi h a con olle , and ca y some
so o ma king ha shows he ag eemen . The e is g ea e he e ogenei y
among a ge s and we ha e seen ha e bs (bo h lexical and auxilia y e bs),
adjec i es, p onouns, nume als, quan i ie s, adposi ions and nouns may be
a ge s. The e is also a a ie y o means by which ag eemen is ma ked on he
a ge including in lec ion, cli ics and mul iple exponence. Two ypes o
ag eemen ma king will be especially ele an o LSE. The i s is alli e a i e
ag eemen , ound in many Ban u languages, which in ol es he appa i ion on
he a ge o a o man (such as a gende -ma king p e ix) al eady p esen on
he con olle , and he mo e excep ional case o adical alli e a i e ag eemen ,
which in ol es copying phonological ma e ial om he con olle on o he
a ge (o en because no o man is a ailable). The second ype o ma king is
p onominal a ixes, which ep esen a g ey a ea be ween ag eemen ma ke s
and ee p onouns.
The domain is he con ex in which he ag eemen ela ion holds
be ween he con olle and he a ge , and is gene ally some so o local
g amma ical ela ion o syn ac ic con igu a ion. Delimi ing he ag eemen
domain is o many he de ining ac o o wha coun s as ag eemen and
wha does no . F om he ypological pe spec i e o his sec ion, we ha e
looked a he a ie y o he phenomenon and his has included ins ances o
“badly beha ed” ag eemen in which he ela ionship be ween he con olle
and he a ge is ex emely unusual: possesso ag eemen , copying- o-objec
cons uc ions and long dis ance ag eemen .
78 Theo ies o ag eemen
Ag eemen in ol es he ep esen a ion o displaced in o ma ion: some
aspec o he con olle is ma ked on he a ge . This in o ma ion may be
ca ego ised in o di e en ypes, o ea u es: he main ea u es o ag eemen
a e gende , numbe and pe son. Each ea u e has di e en alues and hese
alues de e mine he speci ic ma king ha appea s in an ag eemen ela ion
(e.g. i s pe son plu al inclusi e). Languages a y in bo h he ea u es ha
a e used and he se o alues a ailable o a gi en ea u e. The ea u es ha
en e in o a speci ic ag eemen ela ion may depend on speci ic p e equisi es
such as he wo d ca ego y, phonological o m o lexical in o ma ion.
Addi ionally, ag eemen may be subjec o syn ac ic o seman ic condi ions:
conside a ions such as animacy and opicali y may a ec he beha iou o
ag eemen .
This amewo k p o ides desc ip i e ools ha can accommoda e he
ange o ag eemen phenomena ac oss he wo ld’s languages, wi hou being
oo deeply en enched in any speci ic heo y conce ning he na u e o
ag eemen (o language s uc u e in gene al). As s a ed a he beginning o
his chap e , his o e s a way o desc ibing ag eemen in a ela i ely
undocumen ed language, in his case LSE, in such a way ha we can compa e
i o ag eemen in o he languages. A u he means o assessing ag eemen in
LSE is p o ided by he no ion o canonici y de eloped by Co be (2003b,
2006) on he basis o he desc ip i e amewo k al eady desc ibed. Table 2.3
con ains he c i e ia o canonici y, which se ou he p ope ies o ag eemen
in i s mos ag eemen -like mani es a ion.
As well as desc ibing ag eemen in LSE and placing i in he con ex o
ag eemen c oss-linguis ically, a u he objec o his s udy is o examine he
ole o ag eemen wi hin he language sys em, and speci ically he ex en o
which i o ms pa o he g amma ’s syn ac ic mechanisms. The Minimalis
P og am has de eloped a heo y o ag eemen ha dis ils he p ope ies o
ag eemen down o he ba es syn ac ic e ms, he eby con e ing i in o one
o he undamen al ope a ions ca ied ou by syn ax. The nex sec ion p esen s
he concep ion o ag eemen wi hin he Minimalis P og am.
2.3. The Minimalis P og am
As he la es inca na ion o he gene a i is school o linguis ics, he
Minimalis P og am is he p oduc o a esea ch adi ion ha ocuses on he
na u e o language as a unique cogni i e capaci y o humans. The guiding
p inciple behind he gene a i is app oach is he no ion ha language is a
sys em ha can be desc ibed in e ms o a se o ules: hese ules de e mine
wha is pe missible and should p oduce co ec ly- o med sen ences in he
The Minimalis P og am 79
language. Gene a i ism has been applied ac oss di e en a eas o linguis ic
esea ch, om phonology o language acquisi ion, bu much o he body o
wo k has cen ed on syn ax (and i s in e ac ion, o in e aces, wi h o he
linguis ic le els, such as seman ics and phonology). Ini ially, ules we e o he
o m N VP S (“pu a noun and a e b ph ase oge he and you ge a
sen ence”), bu hey ha e since e ol ed o a much g ea e le el o abs ac ion
exp essed in e ms o he s uc u al ela ions be ween elemen s (as will be
exempli ied in he explica ion o ag eemen om a minimalis poin o iew in
sec ion 2.3.3 below).
Recall ha o he s udy o ag eemen in LSE, he Minimalis P og am is
o ele ance on wo main coun s. Fi s ly, because ag eemen has aken cen e
s age wi hin he Minimalis P og am and is judged o be one o he basic
ope a ions used by syn ax o c ea e well- o med sen ences. As a esul ,
ag eemen is de ined in e y speci ic (syn ac ic) e ms. Secondly, he
Minimalis p og am s ems om a adi ion ha asks undamen al ques ions
abou he cha ac e is ics o language as a cogni i e capaci y. These ques ions
ie in well wi h hose ha unde lie his s udy o ag eemen in LSE conce ning
he way in which he use o space in a (signed) language is accommoda ed by
he language sys em.
The p esen a ion o he Minimalis P og am is o ganized as ollows: in
sec ion 2.3.1, I gi e a b ie backg ound o he gene a i is adi ion, p o iding
an o e iew o he impo an issues and he de elopmen s ha ha e shaped
cu en hinking and ha a e ele an o he Minimalis P og am. Sec ion 2.3.2
p esen s he a chi ec u e o he language acul y as en isaged by minimalism.
Finally, he heo y o ag eemen wi hin minimalism is se ou in sec ion 2.3.3.
2.3.1. Gene a i ism: issues and de elopmen s
A cen al concep ha has mo i a ed he Chomskian e olu ion in linguis ics
and guided he gene a i is en e p ise is he no ion o Uni e sal G amma
(UG), he idea ha a i s co e language is a cogni i e capaci y wi h a ixed se
o p ope ies. Indi idual languages may di e in de ails, bu all languages
sha e a common se o p ope ies ha a e shaped and limi ed by UG.
Fu he mo e, Chomsky (1965) makes s ong claims abou he na u e o UG
and main ains ha i is an inna e acul y ha is ha d-wi ed in o he human
b ain.
This concep ualiza ion o language as a limi ed se o ules leads o he
sea ch o hose ules. Ini ially, ules exp essed he means by which syn ac ic
elemen s could be combined and manipula ed o o m sen ences. This ga e
ise o he de elopmen o X-ba heo y, which p o ided he basic amewo k
80 Theo ies o ag eemen
o c ea ing syn ac ic s uc u es, mos commonly ep esen ed in he o m o
ee diag ams.
(39) a.
b.
Mo e speci ically, X-ba heo y p o ides a g ea e le el o abs ac ion han
me ely o mula ing ules, as i cha ac e izes he way in which ules a e
The Minimalis P og am 81
cons ained. The basic uni o X-ba s uc u e is he building block o a
ecu si e con igu a ion ha can accoun o he a angemen o elemen s in a
sen ence. (39a) shows a head X p ojec ing a maximum p ojec ion o he same
ype, XP. The complemen and speci ie posi ions may be occupied by o he
maximal p ojec ions, hus making i possible o one s uc u e o be nes ed
inside ano he . Applying his basic s uc u e o he syn ac ic analysis o an
English sen ence p oduces a ee diag am as shown in (39b).
The posi ions wi hin he syn ac ic s uc u e a e occupied by lexical
elemen s such as e bs, nouns, and adjec i es, which p ojec e b ph ases
(VP), noun ph ases (NP) o adjec i al ph ases (AdjP), bu he e a e also
posi ions ha a e unc ional in na u e and se e o accoun o he ole in
syn ax o elemen s such as in lec ion (IP), and complemen ize s (CP). IP
ini ially p o ided he s uc u e equi ed o a ini e e b o acqui e i s
in lec ed o m. Subsequen ly, he spli -In l hypo hesis o clausal s uc u e
subdi ided his pa o he s uc u e in o a ious p ojec ions such as TP ( o
ense), Ag SP ( o subjec ag eemen ) and Ag OP ( o objec ag eemen )
(Pollock 1989, Kayne 1989, Belle i 2001). O ele ance he e, he Ag SP and
Ag OP p ojec ions we e dedica ed explici ly o accoun ing o ag eemen
phenomena (bu , as shall be explained in sec ion 2.3.3, hey we e
subsequen ly done away wi h). Fu he mo e, he e b i sel has mo e
s uc u e han a simple p ojec ion: in addi ion o VP, a highe P p ojec ion
(o “shell”) domina es VP. The lexical e b occupies VP while a ligh e b,
and by ex ension e balizing a ixes, may occupy P (Hale & Keyse 1993;
Chomsky 1995). In he same manne , CP, also known as he le pe iphe y,
has been expanded in o a se ies o speci ic unc ional p ojec ions, bu de ails
will no be gi en he e as hey do no bea di ec ly upon he analysis o
ag eemen ( o de ails see Rizzi 1997, 2004). The basic clausal s uc u e wi h
he p ojec ions ha will be o ele ance in his s udy is shown in (40).
Ano he impo an de elopmen in he epe oi e o unc ional
p ojec ions ha popula e he syn ac ic s uc u e is he in oduc ion o he
de e mine ph ase (DP), which domina es he noun ph ase. Fu he mo e,
pa allels ha e been d awn be ween he in e nal s uc u e o he de e mine
domain and he clausal domain (Abney 1987; Ouhalla 1991; Aboh, Co e ,
Dyakono a & an Koppen 2010), wi h speci ic unc ional p ojec ions o
numbe (NumP) and a nominalizing equi alen o (nP). These p ojec ions
a e desc ibed in mo e de ail in sec ions 7.1.1 and 7.1.2 when looking a he
loca ion o he ϕ- ea u es wi hin he DP. The in e nal s uc u e o DP adop ed
in his hesis (showing only ele an p ojec ions) is displayed in (41).
82 Theo ies o ag eemen
(40)
(41)
The Minimalis P og am 83
The Minimalis P og am aims o make gene a i is heo y as pa simonious as
possible, and is d i en by conside a ions o economy. The unde lying no ion
is ha language achie es i s ends wi h he ewes possible esou ces. This has
b ough abou a econside a ion o he heo e ical appa a us equi ed o
accoun o syn ac ic phenomena, and a ejec ion o unnecessa y baggage. (As
we shall see when we look a ag eemen in Minimalism in sec ion 2.3.3, one o
he ic ims o his pu ge o economy is he se o Ag p ojec ions.) One o he
guiding p inciples o Minimalism is he Inclusi eness Condi ion, which
s a es ha no new ea u es a e in oduced by he compu a ional sys em
(Chomsky 2000: 113). This means ha syn ax mus make do wi h he se o
lexical i ems ha appea in he nume a ion: i may manipula e he i ems by
means o syn ac ic ope a ions, bu may no add any hing else in he p ocess o
he de i a ion.
The Inclusi eness Condi ion calls in o ques ion he alidi y o X-ba
heo y: he h ee di e en le els o each p ojec ion (X, X’ and XP) do no
igu e as pa o he lexical en y o a gi en i em and mus be added du ing
he de i a ion, hus con a ening he Inclusi eness Condi ion. Ra he han
de ine he di e ences be ween syn ac ic objec s in e ms o hei in insic
ea u es (as is done in X-ba heo y), a ela ional iew o p ojec ions ob ia es
he need o add unnecessa y labels. Unde his pe spec i e, a minimal
p ojec ion (X) is a lexical i em ha has been selec ed, a maximal p ojec ion
(XP) is a syn ac ic objec ha does no p ojec , and an in e media e p ojec ion
(X’) is a syn ac ic objec ha is nei he a minimal no maximal p ojec ion
(Chomsky 1995).
10
As a esul , X-ba heo y gi es way o ba e ph ase
s uc u e, a mo e s eamlined cha ac e iza ion o he way syn ac ic elemen s
a e con igu ed. (No e ha his is a change in he way o concep ualizing he
s uc u e and how i is ep esen ed by he compu a ional sys em. I is s ill
common p ac ice o e e o and o ep esen X-ba esque posi ions, e en
hough he unde lying concep s depend on ba e ph ase s uc u e. This is he
p ac ice I adop he e.)
Ba e ph ase s uc u e includes only lexical ea u es and he objec s
cons uc ed om hem. Syn ax cons uc s objec s om he basic elemen s
aken om he lexicon by means o syn ac ic ope a ions: Me ge and Mo e.
Me ge is mo e basic, and is a ecu si e, wo-place ope a ion ha combines
wo elemen s o o m a la ge one. The p ope ies o he esul ing objec
depend on hose o he elemen s ha a e me ged. Me ge is essen ially he
10
Addi ionally, he need o labels hemsel es has been b ough in o ques ion. As an
al e na i e, ca ego y labels may be eplaced by se s o g amma ical ea u es which
cha ac e ize he idiosync a ic p ope ies o indi idual wo ds. Also, labels may ep esen a
iola ion o he Inclusi eness Condi ion (U iage eka 2000).
84 Theo ies o ag eemen
simple mechanism by which wo ds (o mo e speci ically i ems om he
lexicon) a e pu oge he in a s uc u ed way o c ea e sen ences. In con as ,
Mo e allows an elemen o change loca ion du ing he de i a ion, as occu s
wi h wh-mo emen , o gi e a acile example. Mo e is no a basic ope a ion
and can be decomposed in o mo e undamen al s eps. Mo emen does no
simply place an elemen in a new posi ion he eby lea ing he o iginal
posi ion emp y, bu in ac i lea es behind some so o esidue (known as
aces in ea lie e sions o syn ac ic heo y). This is a es ed by he ac ha
he esidue o ace le by he mo emen c ea es e ec s such as cli iciza ion
blocking o may e en be pa ially p esen in non-s anda d language, as can be
hea d in auxilia y copying in child language and p eposi ion copying in
speech e o s. In (42a) he dele ed copy o should ( ep esen ed as ‘should’)
p e en s he auxilia y ha e om cli icizing on o he p onoun hey; in (42b), a
wo-yea -old epea s he auxilia y e b can in a ques ion in he posi ion ha i
would occupy in a decla a i e sen ence; and in (42c) a adio epo e epea s a
p eposi ion ha has al eady been mo ed o he beginning o he ela i e
clause (examples aken om Rad o d 2004: 157, 156, 192).
(42) a.
Should hey should ha e called he police?
(*Should hey’ e called he police?)
b.
Can i s wheels can spin?
c.
Ikea only ac ually has en s o es om which o sell om.
To accoun o his, mo emen is conside ed o be made up o wo ope a ions:
Copy and Me ge. The elemen o be mo ed is copied, and hen he copy is
me ged in o he new posi ion. To comple e he Mo e ope a ion, some so o
dele ion mechanism is equi ed, o he wise he e a e wo ins ances o he
mo ed elemen : he copy in he new posi ion and he o iginal in he ini ial
posi ion. The dele ion p ocess may be pos ula ed as a sepa a e ope a ion
(Chomsky 1995: 400) o as a ailu e o he o iginal o be gi en phonological
o m. The impo an poin is ha he dele ion p ocess canno be absolu e since
he o iginal ma e ial a ec s o he p ocesses – such as cli iciza ion, as in (42a) –
bu a he same ime he i em does no appea in i s ini ial posi ion in no mal
speech – in con as o (42b, c). An al e na i e explana ion is ha he dele ion
occu s a a speci ic poin in he de i a ion such ha i is dele ed a e i has
had an e ec , bu be o e i is assigned phonological o m.
In his sec ion we ha e seen ha he gene a i is adi ion o linguis ics
e ol es a ound he sea ch o ules o syn ax, o mo e gene ally, he way in
which hose ules a e cons ained. X-ba heo y p o ided a means o
cha ac e izing he ules o syn ax, bu conside a ions o pa simony in oduced
by he Minimalis P og am ha e led o he de elopmen o ba e ph ase
The Minimalis P og am 91
depends on he language: English e bal ag eemen in ol es pe son and
numbe , whe eas in A abic i also in ol es gende (Nasu 2001).
(48) a.
To p o ide a conc e e example o he Ag ee p ocess in ac ion, le us look a a
simple case o subjec - e b ag eemen in Spanish. In he sen ence Saioa habla
ancés (‘Saioa speaks F ench’), he e b ‘speak’ ag ees wi h he subjec ‘Saioa’
in pe son ( hi d) and numbe (singula ). In he de i a ion, he di e en
elemen s a e me ged in o he s uc u e om he nume a ion as shown in
(48a).
16
The e b habla me ges wi h i s di ec objec , he DP ancés, and
con inues o me ge wi h (emp y) s uc u e o o m a syn ac ic objec ( ’). The
16
The in e nal s uc u e o he DPs is no shown he e as he main aim is o illus a e how he
ag eemen mechanism wo ks in e ms o he unc ional heads in ol ed in e bal ag eemen .
The ques ion o he DP-in e nal s uc u e and loca ion o ϕ- ea u es is add essed in sec ion
7.1.1.
92 Theo ies o ag eemen
subjec DP, Saioa, me ges wi h his objec o o m a maximum p ojec ion, P.
This DP has in e p e able inhe en ea u es numbe and pe son, shown in
squa e b acke s. (No e ha he objec DP also has in e p e able inhe en
ea u es, bu hese a e no shown he e in o de o ocus on he subjec
ag eemen p ocess). The P me ges wi h he minimal p ojec ion, T, which
hos s unin e p e able, un alued ea u es ha equi e checking and a e
ma ked as u in (48a). Finally, he esul ing syn ac ic objec , T’, p ojec s a
maximal TP.
(48) b.
The ag eemen p ocess occu s in o de o he un alued ea u es on T o be
alued and ma ked o dele ion. Ac ing as a p obe, T sea ches wi hin i s
domain ( P) o a se o alued ea u es. The nea es possible candida e is he
subjec DP, Saioa, which has a ull se o in e p e able ea u es (pe son and
numbe ). (The objec DP also has a ull se o in e p e able ea u es, bu he
in e ening subjec DP blocks i om ac ing as he goal.) This p obe-goal
The Minimalis P og am 93
ela ion is shown by he do ed line in (48b). The unin e p e able ea u es on
he p obe a e alued and ma ked o dele ion, indica ed by he c ossed ou
ea u es in (48b).
Subsequen ly, he e b unde goes head mo emen o and om he e o T.
As a esul , he alued ea u es on T a e a ixed o he e b so ha he co ec
o m is p oduced a PF a e Spell-Ou . Finally, he subjec DP unde goes
mo emen o he Spec-TP posi ion ( o ul il an independen condi ion, he
Ex ended P ojec ion P inciple), yielding he co ec wo d o de : Saioa habla
ancés. This inal s age o he de i a ion is shown in (48c).
(48) c.
No e ha his de i a ion also wo ks unde mul iple Spell-Ou and cyclic
phases wi h mino adjus men s o he o de in which ope a ions a e applied
so as o espec each cycle. Since P ep esen s a phase, he head mo emen o
94 Theo ies o ag eemen
V o mus occu be o e he i s phase comple es and sends i s complemen ,
VP, o Spell-Ou . Once his happens, bo h he subjec DP and he e b a e
wi hin he phase edge (see (46) abo e), and hus a ailable o ope a ions in he
ollowing phase. This means ha he Ag ee ope a ion can ake place since he
goal is isible o he p obe, as can he emaining mo emen ope a ions.
This e ised cha ac e iza ion o ag eemen in e ms o he Ag ee
ope a ion does away wi h he need o mo emen o es ablish an ag eemen
ela ion.
17
Al hough he e was mo emen in he Spanish example in (48), he
Ag ee ope a ion did no depend on his and he mo emen ook place due o
independen conside a ions. Ag ee is essen ially an ope a ion ha dele es he
unin e p e able ea u es on he p obe, and is, Chomsky claims, speci ic o
language, unlike Me ge, which has analogues in o he cogni i e domains.
In his sec ion we ha e looked a he Ag ee ope a ion as o mula ed by
he Minimalis P og am. This ope a ion es ablishes a ela ionship be ween an
unin e p e able ea u e loca ed in a co e unc ional head ( he p obe) and an
inhe en ly alued ea u e ( he goal) wi hin a speci ic syn ac ic con igu a ion.
As a esul , he unin e p e able ϕ- ea u es on he p obe a e dele ed. Ag ee is a
basic syn ac ic ope a ion ha is unique o he language acul y.
2.4. Summa y
In his chap e , I ha e p esen ed wo e y di e en – bu no incompa ible –
iews o ag eemen ha ha e been de eloped on he basis o da a om
spoken languages. The i s , which I ha e called he ypological app oach,
aims o cap u e he di e si y o ag eemen phenomena in he wo ld’s
languages based on a e y open de ini ion: sys ema ic co a iance be ween he
p ope ies o wo linguis ic elemen s. The second is he cha ac e iza ion o
ag eemen as a undamen al syn ac ic ope a ion ha is a he co e o (and
pe haps unique o) he language acul y, as de eloped wi hin he amewo k
o he Minimalis P og am. Despi e he dispa i y be ween hese wo
app oaches, bo h o e amewo ks wi hin which o examine ag eemen in
LSE. In addi ion, hey p o ide a se ies o ools ha will se e o analyse he
phenomenon and o judge he deg ee o which ag eemen -like p ocesses in
LSE a e he same as wha is desc ibed as ag eemen in spoken languages.
17
Unde he cu en analysis, Ag ee becomes a p e equisi e o mo emen since bo h Me ge
and Ag ee a e each componen s o Mo e: Mo e es ablishes ag eemen be ween α and F and
me ges a ph ase de e mined by F o a p ojec ion headed by α (Chomsky 2000: 101). The
copying componen o Mo e men ioned a he end o sec ion 2.3.1 has been e o mula ed as
an Ag ee ela ionship plus some o he p ocess ha de e mines he ph ase ha en e s he
Me ge ope a ion.
Summa y 95
When we come o examine he LSE da a, hese aids will b ing us close
owa ds answe ing one o he basic esea ch ques ions mo i a ing his s udy:
a e we alking abou he same hing when we desc ibe ag eemen in signed
and spoken languages?
Be o e u ning o he LSE da a and e alua ing i using he con ibu ions
o heo ies de eloped by he s udy o spoken languages, he heo e ical
backg ound o his hesis would no be comple e wi hou looking a he wo k
ha has been ca ied ou on o he sign languages. Al hough esea ch in o sign
language is a much younge ield, he e is a conside able body o wo k
ela ed o ag eemen in se e al sign languages and a ious heo ies ha e been
de eloped conce ning e bal ag eemen . The nex chap e p esen s hese
heo ies.
97
3. Ag eemen in sign languages
The p e ious chap e e iewed he li e a u e on ag eemen as s udied o
spoken languages. This chap e o e iews he di e se mani es a ions and
analyses o ag eemen as ha e been p oposed in he li e a u e o di e en
sign languages, and a emp s o ake in as many ag eemen -like phenomena
om he epe oi es o sign languages. The objec i e, as se ou in sec ion 1.5,
is o look a all possible candida es o an ag eemen mechanism in signed
languages based on wha has been desc ibed in he li e a u e, be o e u ning
(in chap e 5) o he speci ic signed language ha is he ocus o his s udy,
Spanish Sign Language (LSE). This will pu us in a posi ion i s ly o si ua e
he LSE da a in p eceding wo k on o he sign languages and secondly o
assess o wha ex en ag eemen in LSE esembles ag eemen in o he sign
languages. This will also lay he g oundwo k o compa ing wha is called
ag eemen in LSE wi h ag eemen as unde s ood o spoken languages by
adop ing he amewo ks in oduced in chap e 2.
As men ioned in chap e 1, mos esea ch a en ion has been ocused on
ag eeing e bs and hey will also ake up much o his chap e . The ag eeing
mechanism unde lying ag eeing e bs is spa ial, and (as ou lined in sec ion
1.3) he e b changes ce ain aspec s o i s o m in o de o indica e one o
mo e o i s a gumen s. The ag eemen p ocess elies upon an associa ion
be ween a e e en and a poin o loca ion in he signing space. A mo e basic
use o his associa ion be ween e e en and loca ion is seen in he p onominal
sys em, and in many ways p onominal e e ence unde lies he e bal
ag eemen sys em. Fo his eason, in sec ion 3.1 p onominal e e ence and
he na u e o he spa ial e e ence sys em will be desc ibed be o e looking a
e bal ag eemen p ope in sec ion 3.2. The desc ip ion o ag eeing e bs
includes a de ailed look a p o o ypical ag eeing e bs, and he in e es ing
p ope ies hey show, as well as backwa d ag eeing e bs, which p o ide an
oppo uni y o e iew p e ious analyses o his ype o di ec ional e b.
1
1
This e iew is no exhaus i e as he e has been much wo k on e bal ag eemen in sign
languages, and I limi mysel o hose analyses ha a e ele an o his s udy. One no able
absence is Liddell’s (2000, 2003) wo k ha calls in o ques ion he linguis ic s a us o his
spa ial “ag eemen ” mechanism and has undeniably ca alysed much wo k in his ield.
Ne e heless, I e e o Liddell’s wo k whe e ele an and many o he issues aised by
98 Ag eemen in sign languages
Sec ion 3.2 also includes ano he ype o e b ha shows spa ial ag eemen
no no mally included in analyses o e bal ag eemen in sign languages. This
p ocess, which I e e o as single a gumen ag eemen , occu s when a e b is
a icula ed a a loca ion in signing space o ag ee wi h jus one a gumen .
Con inuing wi hin he e bal domain, sec ion 3.3 deals wi h ag eemen
auxilia ies and desc ibes he di e en ypes o auxilia y e bs ha ma k
ag eemen ha ha e been a es ed in he sign language li e a u e. The spa ial
ag eemen mechanism desc ibed in his chap e may also be exp essed by
non-manual ea u es ha can indica e loca ions in he signing space, such as
head il s o eye gaze. E idence o such non-manual ma ke s o ag eemen is
examined in sec ion 3.4. Jus as ag eemen is no es ic ed o he e bal
domain in spoken languages, his spa ial ag eemen mechanism in sign
languages may also be exploi ed o ag eemen in o he domains; in sec ion
3.5 we conside his possibili y o he nominal domain and look o e idence
o DP-in e nal ag eemen . The chap e closes wi h a sec ion ha summa izes
he main cha ac e is ics o spa ial ag eemen in sign languages.
3.1. P onominal e e ence
The p onominal sys em in mos sign languages is mos c udely desc ibed as
poin ing. In he case o physically p esen e e en s, he signe poin s owa ds
he e e en in o de o achie e deic ic e e ence. Fo non-p esen e e en s,
he signe ’s poin is o a loca ion associa ed wi h ha e e en (Co mie 2012).
To go beyond a me e c ude desc ip ion o p onominal e e ence in sign
languages, a ious e inemen s a e necessa y. Fi s ly, poin ing may be done
in a ious ways: manually, mos commonly wi h he ex ended index inge
(B), bu also wi h he ] handshape o he 2 handshape (P au 2011). In some
sign languages (e.g. Lib as, Be enz 2002), he handshape used o poin ing
may change acco ding o s yle o egis e .
2
In he non-manual domain,
poin ing may be ca ied ou by means o a head nod/ il , eye gaze o e en by
lip poin ing, as desc ibed o P o idence Island Sign Language (Washabaugh
1986, ci ed in Meie & Lillo-Ma in 2010: 348) and o an idiolec o ASL
(Bahan 1996: 86 n). Secondly, in he case o non-p esen e e en s, he e a e
a ious ways in which he e e en may be associa ed wi h a loca ion in space.
The gene al p ocess o making use o a poin in signing space has been called
Liddell ha e s imula ed (and a e, o some ex en , I hope, add essed by) he analysis o e ed in
his hesis.
2
The non-dominan hand may also be in oduced o “shield” he poin ing hand when using
he B handshape in o de o make he signalling mo e disc e e. This o m has also been
epo ed as a poli e p onoun in some sign languages, such as Lib as (Be enz 2002).
P onominal e e ence 99
“indexing” (F iedman 1975) o “nominal es ablishmen ” (Klima & Bellugi
1979). The speci ic p ocess by which a e e en is associa ed wi h a loca ion in
space, which I shall e e o as loca ion assignmen , is desc ibed in he nex
sec ion. Finally, he associa ion be ween a loca ion in space and a e e en is
discou se-dependen : he associa ion is c ea ed and alid o a gi en s e ch o
discou se, and may change om one discou se se ing o he nex .
Fu he mo e, he associa ion may change wi hin a gi en s e ch o discou se
in sys ema ic ways h ough he use o ole shi , desc ibed in sec ion 3.1.2.
3.1.1. Loca ion assignmen
F om he o mal poin o iew, he associa ion be ween a e e en and a
loca ion in space may be es ablished by means o a poin o index owa d ha
loca ion (which may be any a ie y o he manual o non-manual ypes
desc ibed abo e) in combina ion wi h a nominal sign. This can be seen in he
NGT example in (1), which in ol es he assignmen o he nominal INGEBORG
o a loca ion o he signe ’s igh by means o a poin immedia ely a e he
nominal. (The sen ence also con ains a second poin o he same loca ion,
which se es as an anapho ic e e ence o he same e e en .)
NGT ( an Gijn 2004: 18. Images ep oduced wi h kind pe mission om he au ho .)
(1)
INGEBORG
IXx
HAPPY
BECAUSE
IXx
GRADUATE
‘Ingebo gi is happy because shei will g adua e.’
An al e na i e s a egy, which ob ia es he need o any poin ing, is o
a icula e he nominal sign di ec ly a a loca ion (which I shall e e o as
localiza ion), hus associa ing he e e en wi h ha loca ion.
3
The a ailabili y
o localiza ion depends on he phonological p ope ies o he sign in ques ion:
body-ancho ed signs, which a e a icula ed in con ac wi h o nea o pa o
he body, canno be mo ed ou in o he signing space. To o e come he
immobili y o body-ancho ed signs, a classi ie may be used o localize he
3
Fo he sake o cla i y, i should be no ed ha localiza ion is no necessa ily loca ion
assignmen . A e e en may al eady be assigned a loca ion, and subsequen p oduc ions o
he localized sign me ely e e o ha e e en . I is he i s a icula ion o he nominal and
he loca ion ( ia poin ing o localiza ion) ha achie es he associa ion be ween he wo. The
con inued use o a localized sign may be due o conside a ions o e e en ial iden i y and
cohe ence. I is also possible ha p agma ic and discu si e unc ions play a ole, and he
signe may wish o easse a loca ion assignmen o ensu e ha he in e locu o s a e keeping
ack o he spa ial dis ibu ion o he e e en s.
100 Ag eemen in sign languages
e e en in he signing space. The use o classi ie s is pa icula ly sui ed o
spa ial desc ip ions (see sec ion 1.1.1) bu hey may also be used o e e en
di e en ia ion in gene al discou se.
4
Loca ion assignmen , hen, may in ol e
poin ing, localiza ion and classi ie s.
As a as he choice o loca ion is conce ned, a ious ac o s may play a
ole. I he e is a loca ion in he discou se se ing ha he in e locu o s
associa e wi h a e e en ( o example, he desk a colleague no mally si s a )
hen ha loca ion may be used (Co mie 2012). O en in such cases, he non-
p esen e e en is imagined as being p esen a he loca ion and occupies a
li e-size space (Liddell 1990). Fo Danish Sign Language (DTS), Engbe g-
Pede sen (1993: 71-74) iden i ies a ious con en ions ha may guide he
selec ion o a loca ion o a e e en : he iconic con en ion condi ions he
choice o loca ions so ha hey e lec he ac ual spa ial ela ionships be ween
he e e en s; he con en ion o seman ic a ini y places seman ically ela ed
i ems a he same loca ion in signing space; and he con en ion o compa ison
places e e en s being con as ed wi h each o he on he le - igh la e al axis
in on o he signe . Fu he mo e, loca ions may be chosen acco ding o
me apho ical schemes: o example, Nilsson (2008: 53) desc ibes how in a
Swedish Sign Language (SSL) discou se desc ibing an in e ac ion be ween a
doc o and a pa ien , he doc o may be associa ed wi h a loca ion highe han
ha associa ed wi h a pa ien o e lec he powe ela ion be ween he wo.
This so o con en ion has also been desc ibed o o he sign languages (e.g.
Bahan & Pe i o 1980 o Ame ican Sign Language – ASL). A a mo e o mal
le el, Ba be à (2014) desc ibes how in LSC di e en planes in he signing
space a e used o loca ion assignmen acco ding o he seman ic p ope ies o
he e e en s, such as speci ici y. Ne e heless, when such discou se and
p agma ic ac o s do no play a ole, i is gene ally assumed ha he choice o
loca ions is a bi a y.
In addi ion o loca ions in he signing space, e e en s may also be
associa ed wi h loca ions on he non-dominan hand, speci ically he ip o
each o he inge s. This s a egy is no mally used when dealing wi h wo o
mo e e e en s ha o m some so o na u al class o g oup, and is known as
a lis buoy (Liddell 2003: 223). The loca ion assignmen may o may no
in ol e poin ing, and may make use o he possibili ies o simul anei y
a o ded by ha ing he non-dominan hand do some hing di e en om he
dominan hand. Simila ly o loca ions in he signing space, once es ablished
4
In he use o classi ie s o pe sonal e e ence, o Tu kish Sign Language (TID) a dis inc ion
be ween a neu al and a hono i ic classi ie has been desc ibed (Zeshan 2003b: 64-67). The
neu al o m uses he B handshape, in con as o he 2 handshape o he hono i ic o m.
Ag eeing e bs 107
dis inc ions will be e u ned o in he discussion o he ϕ- ea u es p esen in
LSE in chap e s 5 and 6.
9
The loca ions employed o ack e e ences in sign languages a e no
pu e logical indices since a clean, cons an one- o-one co espondence is no
adhe ed o. Howe e , p onominal poin s a e dominan ly indexical in na u e,
as becomes clea when compa ing hem o p onominal e e ence in spoken
languages (Co mie , Schemb i & Woll 2013). T ea ing loca ions in space as
R-loci ha se e as an explici mani es a ion o an abs ac index (which is
mapped on o a discou se e e en ) p o ides a amewo k ha can explain he
la gely indexical na u e o p onominal e e ence in sign languages.
Fu he mo e, he model also accoun s o he unlis abili y, he discou se
de e minacy and he po en ial o shi o hese o ms. This is he model ha
will se e as he basis o he analysis de eloped in his hesis.
3.2. Ag eeing e bs
Ve bal ag eemen has a ac ed much a en ion in he sign language li e a u e,
and he exis ence o e bs ha show a spa ial means o ma king a gumen s
has been iden i ied and desc ibed o many sign languages (see Ma hu &
Ra hmann 2010, 2012 o o e iews o he phenomenon om desc ip i e and
heo e ical pe spec i es, espec i ely).
A s iking ea u e o e bal ag eemen in sign languages is ha no all
he e bs in a sign language show ag eemen . Al hough mos e bs can be
modi ied o aspec and many sign languages display a ich aspec ual sys em
(P au, S einbach & Woll 2012), no all e bs can in lec o ma k hei
a gumen s. Thus, o example, in Mexican Sign Language (LSM), e bs such
as KNOW, UNDERSTAND and LOVE do no mo e be ween poin s in space o
ma k hei a gumen s (C uz Ald e e 2009). These a e known as plain e bs.
O hose e bs ha a e di ec ional, an impo an dis inc ion is made be ween
spa ial e bs and ag eeing e bs (Padden 1983/1988
10
). Unde lying his
9
The glosses in he ansc ip ions dis inguish be ween loca ions on he signe ’s body (by
using a 1 subsc ip ) and loca ions in he signing space (by using a le e subsc ip , such as x o
y). (See he no a ion con en ions o mo e de ails.) Al hough his gi es he appea ance o a
i s /non- i s dis inc ion, his is done o cla i y gi en he phonemic salience o con ac wi h
he body. Using his no a ion con en ion does no commi me o a i s /non- i s pe son
dis inc ion, and, as shall become clea in he analysis de eloped o ag eemen in LSE, his is
no he pa h I ake.
10
Padden ini ially dis inguished be ween spa ial e bs and in lec ing e bs (1983/1988, 1986),
bu his was me ely a ques ion o nomencla u e, and in la e wo k (1990, 1998), she adop s he
e m “ag eemen e bs” ( ollowing Liddell & Johnson 1989). O he e ms ha a e used in he
li e a u e, such as “indica ing” (Liddell & Me zge 1998 and subsequen wo k by Liddell) o
108 Ag eemen in sign languages
dis inc ion is a di e ence in he use o space: spa ial e bs use he signing
space opog aphically o desc ibe spa ial ela ions, whe eas ag eeing e bs
use space ca ego ically o ma k pe son and numbe o he e bal a gumen s
(Padden 1990). This has se e al consequences o how each ype o e b
beha es. Spa ial e bs (such as MOVE, PLACE, THROW and STOP in LSM) use he
en i e signing space o show manne and pa h o mo emen such ha any
sligh change in he mo emen o he sign al e s i s meaning; ag eeing e bs
(such as HELP, TEACH, PAY and RESPECT in LSM) es ic mo emen o he
ho izon al plane and “ a y he posi ion o he beginning and end poin s o he
sign depending on he pe son ag eemen , an in lec ional ca ego y o which
spa ial e bs a e obli ious” (Padden 1983/1988: 47). Examples o each ype o
e b in LSM a e shown in (7).
LSM (adap ed om C uz Ald e e 2009: 733, 761, 747)
neg
(7) a.
PEPE
IXx
IX1
NOT-KNOW
‘Pepe doesn’ know me.’
eye gazex
b.
GLASS
IXx
TABLE
yMOVEx
‘Take he glass o he able.’
c.
PEPE
xRESPECTy
IXy
MARÍA
‘Pepe espec s Ma ía.’
The plain e b NOT-KNOW (7a) canno in lec o ma k i s a gumen s, e en
hough he e e en s in he sen ence a e associa ed wi h poin s in he signing
space (‘Pepe’ a locus x, and he i s pe son a he signe ’s ches ). The e b
MOVE (7b) is modi ied spa ially o al e he meaning o he e b: he s a poin
o he sign is he locus associa ed wi h he add essee, o mo e speci ically, he
place whe e he add essee is, and he end poin is he locus associa ed wi h
he able. Addi ionally, he end poin is also ma ked by eye gaze owa ds ha
loca ion, as ma ked in he glosses. The a gumen s associa ed wi h he e b a e
loca i e, gi ing a li e al meaning o ‘Take he glass om he place whe e you
a e o he able’. In con as , he a gumen s o he e b RESPECT (7c) a e no
loca i e bu pe son a gumen s: he e b s a s a he locus associa ed wi h he
subjec a gumen (PEPE) and mo es o he locus associa ed wi h he objec
a gumen (MARÍA).
“deic ic” (Mo ales-López e al. 2005) e bs, e lec he amewo k and analysis adop ed by he
au ho s, and will be men ioned whe e ele an in he desc ip ion o heo ies o e bal
ag eemen in sign language h oughou his sec ion.
Ag eeing e bs 109
These ag eeing e bs, as hey a e now mos commonly known, a e
desc ibed in sec ion 3.2.1, wi h an o e iew o he di icul ies in p o iding a
sys ema ic accoun o he phenomenon and o he di e en heo ies ha ha e
a emp ed o cha ac e ize his e bal p ocess. Addi ionally, he exis ence o
ce ain “backwa ds” e bs ha show he e e se pa e n o associa ion
be ween s a /end poin and subjec /objec makes inding a cohe en accoun
e en mo e challenging, and his is deal wi h in sec ion 3.2.2. Al hough wo k
on e bal ag eemen is no mally es ic ed o di ec ional e bs wi h wo
a gumen s, some e bs make use o a spa ial mechanism o ma k jus one o
hei a gumen s, and his mechanism is desc ibed in sec ion 3.2.3.
3.2.1. P o o ypical ag eeing e bs
The spa ial beha iou o mul i-di ec ional (F iedman 1976) o di ec ional
(Fische & Gough 1978) e bs, as hey we e i s desc ibed, has been he ocus
o mos o he wo k on e bal ag eemen in sign languages. The phenomenon
o in e es is shown by he di e en ealiza ions o he e b TEASE in New
Zealand Sign Language (NZSL) in example (8). A change in he di ec ion o
he e b changes he a gumen s ha a e iden i ied as he subjec and objec : in
(8a), he e b mo es om a poin associa ed wi h he i s pe son subjec , o a
poin associa ed wi h a non- i s pe son objec , yielding ‘I ease you’; while
he in e se mo emen (and, in his case, o ien a ion o he hand) gi es he
meaning ‘You ease me.’
NZSL (Online Dic iona y o New Zealand Sign Language)11
(8)
a. 1TEASEx
b. xTEASE1
‘I ease you.’
‘You ease me.’
The means hese e bs use o ma k ag eemen manually, by mo emen
and/o o ien a ion, a e desc ibed in sec ion 3.2.1.1., as well as ma king o
11
Images aken om ideos o sen ence examples in he Online Dic iona y o New Zealand
Sign Language, a ailable a h p://nzsl. uw.ac.nz/.
110 Ag eemen in sign languages
plu ali y. Sec ion 3.2.1.2 add esses he ac ha ag eemen does no always
appea , nei he on all he e bs o a gi en sign language no all he ime on a
gi en ag eeing e b, desc ibing ag eemen ma ke omission and de ec i e
ag eemen , wo phenomena ha con ibu e owa d he op ionali y o
ag eemen . The es ic ion o ag eemen o ce ain e bs is explo ed in sec ion
3.2.1.3 by looking a he possible p e equisi es and condi ions ha apply o
ag eemen and ha could o e an explana ion as o why ag eemen is
spo adic in sign languages. Any a emp o es ic ag eemen o ag eeing
e bs ine i ably leads o assessing wha se s hem apa om he o he g oup
o in lec ing e bs, namely spa ial e bs. Sec ion 3.2.1.4 e-examines he
dis inc ion be ween ag eemen and spa ial e bs, and e e s o analyses ha
ea all in lec ing e bs (bo h ag eeing and spa ial) as a single ca ego y. A
summa y is gi en in sec ion 3.2.1.5.
3.2.1.1. Ma king ag eemen
Ag eeing e bs in sign languages ma k wo a gumen s by mo ing be ween
loci associa ed wi h he e b’s a gumen s, speci ically om he subjec locus
o he objec locus. Howe e , his is no he whole s o y. The di e en o ms
o he NZSL ag eeing e b shown in (8) change in he di ec ion o he
mo emen bu also in he o ien a ion o he (dominan ) hand.
RSL (Schwage & Zeshan 2008: 536. Images ep oduced wi h kind pe mission om
John Benjamins Publishing.)
(9)
a) 1DISTURBx
b) xDISTURB1
‘I dis u b you.’
‘You dis u b me.’
Fo some e bs, o ien a ion alone may be used o ma k ag eemen , ypically
wi h he palm o he inge s o he hand acing owa ds he objec locus and
away om he subjec locus. This is he case o he Russian Sign Language
(RSL) e b DISTURB, which includes a downwa ds mo emen and so uses only
Ag eeing e bs 111
o ien a ion o ma k ag eemen , as shown in (9).
12
The dis inc ion be ween
mo emen and o ien a ion is cen al o some cha ac e iza ions o ag eeing
e bs ha will be discussed in sec ion 3.2.2.
Whe he o no he in lec ed e b o m mo es h ough he signing
space, he impo an idea is ha i s o m aligns wi h a ec o be ween he wo
loci associa ed wi h i s a gumen s (Padden 1990). As such, ag eeing e bs a e
always ansi i e o di ansi i e e bs, since wo a gumen s a e equi ed o
ag eemen o ake place. In he case o di ansi i e e bs, he second a gumen
ha is ma ked is in a iably he no ional indi ec objec a he han he di ec
objec (Co mie , Wechsle & Meie 1999), as shown in (10) o he LSM
di ansi i e e b GIVE. As men ioned in sec ion 2.2.4, his dis inc ion be ween
p ima y (di ec objec in a mono ansi i e clause o indi ec objec in a
di ansi i e clause) and seconda y (di ec objec in a di ansi i e clause)
objec s is no uncommon in spoken languages (D ye 1986).
LSM (adap ed om C uz Ald e e 2009: 749)
(10)
IXx
IXy
BOOK
xGIVEy
‘She ga e him a book.’
The claim o ag eeing e bs is ha hey ma k he pe son and numbe
ea u es o hei a gumen s. The issue o pe son ma king has been men ioned
al eady in he discussion o p onominal e e ence in sec ion 3.1.3; he
mani es a ion o numbe will be desc ibed he e. Numbe and plu ali y a e
ma ked in a a ie y o ways by sign languages, and in he e bal domain
he e a e a ious possibili ies o dis inguishing be ween singula and plu al
e e en s. Singula has ze o ma king, a s a egy ha is ex emely common in
spoken languages also (S einbach 2012). Fo non-singula e e en s, di e en
o ms ha e been desc ibed: mul iple o collec i e ma king by means o an a c
in he ho izon al plane; dis ibu i e o exhaus i e ma king by means o
mul iple eduplica ion oge he wi h a sidewa d mo emen ; dual ma king by
a single eduplica ion (wi h a change o loca ion be ween each ealiza ion), o
by simul aneous o sequen ial ealiza ion on each hand (Padden 1983/1988;
Ma hu & Ra hmann 2010: 181). The mul iple and dis ibu i e o ms o he
Aus alian Sign Language (Auslan) e b ASK a e con as ed wi h he singula
(ze o ma ked) o m in (11).
The dis ibu i e o m (11c) is e y simila in o m o ce ain ypes o
aspec ual ma king, such as he i e a i e (Wilbu 2008). Addi ionally, he
modi ica ion ela es o he empo al s uc u e and p ope ies o he e en
12
Fo mo e examples wi h pho os o di e en ag eeing e bs ha con as he use o
mo emen and o ien a ion, see (8) in sec ion 5.2.1.
112 Ag eemen in sign languages
being desc ibed, as p oposed by Wilbu ’s (2008) E en Visibili y Hypo hesis.
Taken oge he , hese obse a ions sugges ha he dis ibu i e ma king is as
much abou he e en exp essed by he e b as abou i s a gumen s. Recall
ha in sec ion 2.2.5.2 a dis inc ion was d awn be ween nominal numbe –
ela ing o he a gumen s con olling he ag eemen p ocess – and e bal
numbe , which is an inhe en ea u e o he e b. The dis ibu ed o m,
inso a as i in ol es eduplica ion o he e b, appea s o p o ide
in o ma ion abou e bal a he han nominal numbe . Fo he ime being, I
ocus on he mul iple o m (11b) as he ma ke on he e b o plu ali y o he
e b’s a gumen , bu will e u n o he issue o hese di e en plu al ma ke s
in sec ion 6.4.2 in he assessmen o he exp ession o numbe in e b
ag eemen in LSE.
Auslan (Johns on & Schemb i 2007: 148. Images ep oduced wi h kind pe mission
om Camb idge Uni e si y P ess.)
(11)
a) 1ASKx
b) 1ASKx.PL
c) 1ASKx.DISTR
‘I ask him.’
‘I ask hem.’
‘I ask each o hem.’
3.2.1.2. Lack o ag eemen ma king
Ve bal ag eemen in sign languages shows se e al in e es ing p ope ies
ela ed o he op ionali y o he appea ance o ag eemen . Fi s ly, no all e bs
in a gi en sign language show ag eemen . Secondly, hose e bs ha a e
ag eeing e bs o en do no show ag eemen o one o e en bo h o he
a gumen s. Finally, some ag eeing e bs show a de ec i e pa adigm due o
he ac ha hei phonological o m obs uc s he appea ance o he ma king
o he subjec a gumen . Each o hese phenomena will be desc ibed in u n.
As poin ed ou in he in oduc ion o his sec ion, ag eeing e bs
cons i u e a subse o he e bs, which means ha his mechanism is no
uni o m ac oss all e bs in a gi en sign language. C oss-linguis ically, his is
ex emely unusual since i a language has e bal ag eemen , i appea s on all
e bs (Co be 2003): al hough he e a e wo spoken languages ha show
Ag eeing e bs 113
ag eemen on only 30% o he e bs, he Nakh-Daghes anian languages
Chechen and Ingush (Bickel & Nichols 2007), his is a e y a e si ua ion.
Wi hin he class o ag eeing e bs, he e is a no able endency no o
show ag eemen . Fo ASL, Meie (1982) and Padden (1983/1988) no ed ha
he ma king o bo h a gumen s is no obliga o y and he e b may in lec o
jus one o i s a gumen s. I his happens, he a gumen ha is omi ed is
always he subjec , as shown in (12). The ag eeing e b GIVE s a s no a a
locus associa ed wi h he subjec a gumen WOMAN bu a a neu al loca ion in
he signing space. This phenomenon o ag eemen ma ke omission is also
a es ed in o he sign languages: example (13) om LSM shows how an
ag eeing e b may ma k only he objec a gumen (compa e wi h (7c) abo e,
in which he e b RESPECT ma ks bo h a gumen s). This will p o e impo an
o cha ac e izing ag eeing e bs when conside ing backwa ds e bs,
desc ibed in sec ion 3.2.2.
ASL (adap ed om Padden 1983/1988: 136)
(12)
WOMAN
neu GIVE1
NEWSPAPER
‘The woman ga e me a newspape .’
LSM (adap ed om C uz Ald e e 2009: 739)
(13)
JUAN
RESPECTx
‘Juan espec s you.’
Addi ionally, ag eemen may be comple ely absen and nei he a gumen is
ma ked on he e b. Co pus-based s udies in a ious sign languages ha e
e ealed ha ha ag eeing e bs mo e o en appea unin lec ed han
in lec ed, e en in con ex s in which ag eemen ma king could be expec ed (de
Beuze ille, Johns on & Schemb i 2009 o Auslan; Schui 2013 o Inui SL).
The op ionali y o ma king one o he a gumen s, desc ibed abo e as
ag eemen ma ke omission, should be dis inguished om he case o
ansi i e e bs ha can only show ma king o one a gumen , which has
been called single ag eemen (Meie 1982). In o de o a oid con usion wi h
he dis inc phenomenon o single a gumen ag eemen , o be in oduced in
sec ion 3.2.3, I use he e m de ec i e ag eemen o e e o his (phonological)
es ic ion o ag eemen o a single a gumen . In he case o ag eemen ma ke
omission, ma king o one o he a gumen s may no appea ; o de ec i e
ag eemen , ma king o one o he a gumen s canno appea . This occu s wi h
e bs such as SEE o TELL in ASL (Hahm 2006), which canno show ma king
since he o m o he e b includes a loca ion (nea he eyes and nea he
mou h, espec i ely, o hese e bs). As a esul , hese e bs show a de ec i e
ag eemen pa e n in which only he objec is ma ked. In his sense, he
114 Ag eemen in sign languages
phonological o m o he e b is a p e equisi e o ag eemen : he e b mus
ha e no speci ica ions o loca ion in o de o ull (i.e. wo-place) ma king o
ake place. This issue o de ec i e ag eemen will be e u ned o in he nex
sec ion when we look a he phonological cons ain s ha ope a e on ag eeing
e bs.
3.2.1.3. P e equisi es and condi ions
As he p e ious sec ion showed, ag eemen in sign languages is bo h
spo adic, in he sense ha only some o he e bs show ag eemen , and
op ional, since ag eemen may o may no appea . Lea ing aside he issue o
he op ionali y o ag eemen , which will be add essed in chap e 6, he
une en dis ibu ion o ag eemen ac oss he e bs o sign languages may be
due o a ious cons ain s on ag eemen in he o m o p e equisi es and
condi ions. In sec ion 2.2.6, we saw ha p e equisi es ope a e on ag eemen a
he phonological and mo phological le el, while condi ions a e conside a ions
o a syn ac ic o seman ic o de ha ha e an e ec on ag eemen .
Conside ing possible condi ions on ag eemen , a syn ac ic condi ion
could be ha ag eeing e bs mus be ansi i e o di ansi i e since wo
a gumen s mus be a ailable o ag eemen o ake place. This condi ion is
implici ly included in many au ho s’ de ini ions o ag eeing e bs, whe he
hey men ion ag eemen wi h subjec and objec (Padden 1983/1988: 47; Mei
2002: 421) o make explici men ion o wo a gumen s (Edge & He mann
1977: 147; Ma hu 2000: 212; Hong 2008: 170). As we saw in sec ion 3.2.1.2,
de ec i e ag eemen occu s when a gumen s canno be exp essed on he e b
( o phonological easons) bu he e b is s ill equi ed o ha e wo (o mo e)
a gumen s. Limi ing ag eemen o e bs wi h wo o mo e a gumen s is a
somewha a bi a y s a e o a ai s, and as I shall sugges in sec ion 3.2.3,
ag eemen can and does occu on e bs wi h a single a gumen . Fu he mo e,
he condi ion would be necessa y bu no su icien o ag eemen o occu :
many plain e bs a e ansi i e, such as KNOW in (7a) abo e.
A seman ic condi ion imposed on ag eeing e bs is he equi emen ha
bo h he e b’s a gumen s be [+human], o less es ic i ely [+anima e]. Some
au ho s include such a condi ion in hei de ini ion o ag eeing e bs (e.g.
Ma hu 2000: 212; Hong 2008: 170).
13
This seman ic condi ion may be subjec
o c oss-linguis ic a ia ion since he e a e clea di e ences in he li e a u e.
13
This es ic ion seems o be ela ed o he no ion o ans e men ioned below in sec ion
3.2.1.4: a gumen s which can be he sou ce o goal o ans e a e ei he loca i e (in which
case he e b is spa ial) o po en ial possesso s (and he e o e mos likely [+human], o a leas
[+anima e]). This idea is picked up in Mei ’s (1998b) analysis, which will be desc ibed in
sec ion 3.2.2.
Ag eeing e bs 115
Fo example, Yang & Fische (2002: 171) epo an ag eeing e b in Chinese
Sign Language (CSL) ha can in lec o he objec only i he objec is
anima e. In con as , o LSQ, we ind an example o an ag eeing e b
ma king o wo inanima e a gumen s: in (14) he e b FEND-OFF ma ks wo
non- i s pe son a gumen s, which a e ‘ he ozone’ (p esumably men ioned
ea lie in he discou se) and ‘ he sola sys em’.
LSQ (adap ed om Nadeau & Desou ey 1994: 153)14
(14)
xFEND-OFFy
SOLAR-SYSTEM
‘[Ozone] p o ec s us om he sola sys em.’
Al e na i ely, one way o explain he dis ibu ion o e bal ag eemen migh
be ha a phonological p e equisi e excludes many e bs om ag eeing. As
we saw abo e in he desc ip ion o de ec i e ag eemen (sec ion 3.2.1.2), e bs
wi h a speci ied loca ion canno show ull ag eemen . Thus, since ag eemen
may be mani es ed h ough mo emen and/o o ien a ion, i may be ha
e bs o which bo h mo emen and o ien a ion a e speci ied in hei
phonological o m canno show ag eemen . The ull speci ica ion o loca ion
(i.e. o all he loca ion slo s o he sign’s phonological ma ix) may also ba
ag eemen since his e ec i ely limi s mo emen and makes i una ailable o
he exp ession o ag eemen . This would go a long way o explaining why
body-ancho ed e bs a e plain e bs. Howe e , i would do li le o explain
why spa ial e bs do no show ag eemen . The dis inc ion be ween spa ial
and ag eeing e bs is ques ioned in he nex sec ion.
In ying o dis inguish ag eeing e bs om o he e b classes we ha e
looked a (syn ac ic/seman ic) condi ions and (phonological) p e equisi es
ha migh de e mine he appea ance o ag eemen . We now ocus on he class
o ag eeing e bs hemsel es o conside phonological cons ain s o which
hey a e subjec . These e bs all in o wo ca ego ies: he i s , al eady
men ioned abo e, is de ec i e ag eemen , which occu s when an ag eeing
e b canno in lec o bo h a gumen s; he second a ises due o gaps in he
ag eemen pa adigm caused by phonologically illici o ms o ce ain
pe son/numbe combina ions o he a gumen s.
In o de o ag eemen o ake place, a e b mus be able o mo e
h ough space (o a leas o o ien i sel in space). Howe e , many e bs ha e
14
In he o iginal example, Nadeau & Desou ey gi e he glosses in F ench, and he e b is
glossed as DEFENDRE. I gloss he e b as FEND-OFF in o de o show he a gumen s uc u e
mo e clea ly, since he subindices on he e b indica e ha i ag ees wi h wo hi d pe son
a gumen s and no wi h ‘us’, which he au ho s include in he English ansla ion o he
sen ence. The impo an poin is ha he e b is ag eeing wi h inanima e a gumen s.
116 Ag eemen in sign languages
a lexically speci ied loca ion ha cons ains he s a poin o he sign, hus
blocking he possibili y o ma king ag eemen wi h he subjec a gumen .
This gi es ise o a de ec i e ag eemen pa adigm. In he analysis o he
phonological clashes ha a ise be ween e b oo s and he mo emen needed
o ag eemen , Mei (1998b: 90) men ions ha se e al such e bs exis in
Is aeli Sign Language (ISL), such as ASK and ANSWER (bo h speci ied nea he
mou h), SEE (nea he eye) and TELEPHONE (nea he ea ). These e bs
gene ally ag ee wi h only he objec a gumen , and he subjec a gumen mus
be ma ked in some o he way (by he p esence o an explici nominal o
p onominal), as shown in (15) o he de ec i e ag eeing e b SEE in ASL.
ASL (adap ed om Co mie , Wechsle & Meie 1998: 220)
(15)
IXx
SEEy
‘She sees him.’
In ISL, hough, ce ain o ms o hese e bs may show ag eemen wi h bo h
a gumen s:
[ ]he only o ms o hese [de ec i e] e bs which ag ee wi h wo
a gumen s a e hose ha in lec o 1P objec . In such cases, he e b
o m has a complex pa h mo emen : i begins [a he locus associa ed
wi h] he subjec , mo es o he speci ied loca ion, and hen o he 1P
locus ( he signe ’s ches ). (Mei 1998b: 90-91)
When desc ibing de ec i e ag eeing e bs in ASL, Ma hu & Ra hmann (2010:
178) men ion ha a a ian o he e b TELL does allow ag eemen wi h bo h
a gumen s: in con as o ISL, which includes he speci ied loca ion as he
in e media e poin o he sign (subjec >mou h>objec ), he s a egy in ASL is
o begin he sign a he lexically speci ied loca ion and hen mo e o subjec
locus and om he e o objec locus (mou h>subjec >objec ).
The second ype o cons ain on he o m o ag eeing e bs a ises as a
esul o a icula o y incompa ibili ies. Fo example, he ASL sign GIVE is
lexically speci ied as ha ing he palm o he hand o ien ing upwa ds, which
equi es adio-ulna supina ion. A he same ime, i exp esses ag eemen
h ough mo emen and o ien a ion. A i s pe son objec o m would equi e
poin ing and mo ing he hand inwa ds ( owa ds he signe ) and a plu al o m
would in ol e adding an a c mo emen . I he a ge o m is i s pe son
plu al objec , his would equi e bo h o hese mo emen s. Howe e ,
combined wi h adio-ulna supina ion (in o de o keep he palm o he hand
acing upwa d), he esul pu s he a m in an awkwa d con igu a ion, and
he e o e he esul ing o m is no possible. In o de o s udy which e bs
allow which combina ions, Ma hu & Ra hmann (2001, 2006) collec ed da a
Ag eeing e bs 123
This sepa a e mechanism o acing allows a dis inc ion o be d awn
be ween seman ic oles ( e lec ed by he di ec ion o he mo emen ) and
syn ac ic oles ( e lec ed by he acing o he hands). Since Mei ’s p oposal
will be ele an o p ope ies o LSE e bal in lec ion as analysed in chap e 7,
some u he de ails o he p oposal will be gi en he e.
ISL (Mei 1998b: 84, 123. Images ep oduced wi h kind pe mission om he au ho .)
(18)
a) 1HATEx
b) xHATE1
‘I ha e you.’
‘You ha e me.’
c) xTAKE-ADVANTAGE-OF1
d) 1TAKE-ADVANTAGE-OFx
‘I ake ad an age o you.’
‘You ake ad an age o me.’
Based on he obse a ion ha ag eeing e bs ha e a speci ic lexical s uc u e
ha deno es ans e om a SOURCE a gumen o a GOAL a gumen , Mei
de eloped a hema ic s uc u e ag eemen analysis acco ding o which
ag eeing e bs me ge wi h a pa icula p edica e, DIR (called PATH in Mei
1998b), which deno es mo emen om one place o ano he (also p esen in
spa ial e bs, which also include he seman ic concep o mo ion). I is his DIR
p edica e ha shows ag eemen and no he e b i sel . To explain he
mechanisms o he ag eemen p ocess, Mei p oposes ha DIR is a bound
mo pheme ha uses wi h he oo o he e b, and desc ibes his in e ms o a
me ge o he lexical concep ual s uc u es o each elemen which esul s in a
complex e b. Essen ially, DIR ills he a gumen slo s o he e b oo wi h i s
own a gumen s (which ha e al eady been assigned hema ic oles o SOURCE
and GOAL); he complex e b hen checks he e e en ial ea u es o i s
a gumen s in he syn ax (Mei 2002: 438). The e e en ial ea u es o he
124 Ag eemen in sign languages
a gumen s a e exp essed as an explici index in he o m o spa ial loci and
hese loca ions a e copied on o he e b’s phonological ea u es a Spell-Ou .
F om a phonological poin o iew, he DIR mo pheme has wo emp y
loca ion slo s and may be ep esen ed in e ms o Sandle ’s (1989) Hand Tie
Model as in (19):
(19) DIR: (adap ed om Mei 1998b: 167)
μ
loca ion
[ ]
loca ion
[ ]
A e he me ge wi h he e b oo , hese emp y slo s se e as he landing si e
o he phonological loca ion ea u es copied om he a gumen s’ loci. Fo
his o happen, he e b oo mus also be unde speci ied in i s phonological
ma ix o loca ion and pa h mo emen ,
18
so ha he usion o he e bal oo
and DIR does no esul in a phonological clash. O he wise ag eemen canno
be exp essed, o is only pa ially exp essed. We ha e al eady men ioned cases
o his so : he examples in sec ion 3.2.1.1 showed how ag eemen by
mo emen may be comple ely blocked by lexically speci ied mo emen
(and/o loca ion/con ac ) in he e b’s phonological ma ix, esul ing in ei he
a plain e b (such as LOVE in LSM) o a e b ma king ag eemen h ough
o ien a ion alone (such as DISTURB in RSL). Al e na i ely, mo emen may be
pa ially blocked (as desc ibed in sec ion 3.2.1.2) in he case o de ec i e
ag eemen : a lexically speci ied loca ion p e en s he i s slo in he DIR
mo pheme om se ing as he landing si e o he loca ion o he SOURCE
a gumen .
Mei (1998b, 2002) cha ac e izes he independen mechanism o acing,
which ma ks he objec a gumen , as a case ma ke a he han an ag eemen
p ocess. The main mo i a ion o his is he ac ha in sign languages objec
ma king is mo e p ominen han subjec ma king, as demons a ed by
a gumen ma ke omission and de ec i e ag eemen (sec ion 3.2.1.2). While i
is ypologically unusual o ind ag eemen wi h only he objec , case ma king
o only he objec is common, so he ac ha sign languages ma k he objec
in his way i s in o common c oss-linguis ic pa e ns o case ma king o he
objec a gumen . Wha is sligh ly unusual is ha he case ma king appea s on
18
B en a i (1998: 4) dis inguishes be ween local and pa h mo emen . The dis inc ion is bo h
a icula o y and phonological: local mo emen s a e made by he w is , knuckles o inge
join s, and may cause a change in handshape o o ien a ion; pa h mo emen s a e made wi h
he elbow o shoulde and may be speci ied as a mo emen ea u e o a change in loca ion.
Ag eeing e bs 125
he e b a he han on he noun a gumen i sel , bu case ela ions may be
ma ked by e bal a ixes in head-ma king languages, ha is, languages ha
ma k ela ions on he head ( a he han he dependen s) o he ph ase
(Nichols 1986). The case ma king by means o acing appea s on hose e bs
whose lexical concep ual s uc u e includes he no ion o ans e , hus
making a dis inc ion be ween spa ial and ag eeing e bs. The o me deno e
mo ion bu no ans e and hus ha e no acing, whe eas he la e do deno e
ans e , acco ding o Mei ’s analysis. Again, i he phonological speci ica ion
o he e b oo con ains a speci ica ion o he palm and inge o ien a ion,
his blocks he possibili y o ma king he objec a gumen ia acing. Fo ISL,
Mei (1998b: 245-52) shows how di e en phonological clashes es ic he
way in which a e b exp esses ag eemen and case ma king.
One p oblem wi h a seman ically based accoun is ha i ails o
desc ibe he ag eemen phenomenon in con igu a ional e ms; ecall om
chap e 2 ha ag eemen ela ionships hold in he con ex o a speci ic syn ac ic
con igu a ion. In o de o ge ound his, Mei sugges s ha ag eemen is a
p ope y o he spa ial DIR p edica e (p esen in bo h spa ial and ag eeing
e bs) and ha i is he close ela ionship be ween he hema ic oles ha DIR
assigns o i s a gumen s and he ag eemen slo s hese a gumen s a e
associa ed wi h ha gi es ag eeing e bs hei hema ic la ou . As Mei
(2002: 440) pu s i , “ag eemen in ISL is ela ed o speci ic spa ial hema ic
oles because o he spa ial na u e o he ag eeing elemen ”. Essen ially,
ag eemen looks spa ial because wha ag ees in sign languages is a spa ial
elemen (DIR). This p oposal also deals wi h ano he anomaly: as men ioned in
sec ion 3.2.1.2, a ema kable p ope y o sign language e bal ag eemen is
ha i is no mani es ed by all e bs in a language, in con as o he case o
spoken languages ha whene e a language has e bal ag eemen , all he
e bs show ag eemen . A consequence o Mei ’s analysis is ha sign language
ag eemen beha es mo e consis en ly since i is no longe a p ope y o a
es ic ed class o e bs (o o any e bs o ha ma e ) bu a he i a ises
whene e a pa icula p edica e (DIR) is p esen .
Ano he ou come o his model o ag eemen is a e inemen o he
seman ic es ain s placed on he a gumen s o ag eeing e bs. As men ioned
in sec ion 3.2.1.3, in he li e a u e i is o en claimed ha ag eeing e bs can
only ake [+anima e] a gumen s (c . Janis 1995; Ma hu 2000; Ra hmann &
Ma hu 2005; Quad os & Que 2008); in con as , ollowing om he
cha ac e iza ion in e ms o ans e , Mei sugges s ha he ele an ea u e is
ha a gumen s be po en ial possesso s (1998b: 203 n).
126 Ag eemen in sign languages
3.2.2.4. P oblems wi h Mei ’s accoun
Al hough Mei ’s analysis dis inguishes he mo emen and acing as wo
sepa a e mechanisms, ea ing acing as a case ma ke e ec i ely excludes i
om he ag eemen p ocess, and he mo emen elemen is he only ag eemen
ma ke . This has wo consequences o he analysis. Fi s ly, he ag eemen
p ocess is common o bo h spa ial and ag eeing e bs, despi e he di e ences
in he use o space ha each ype o e b seems o in ol e (as desc ibed in
sec ion 3.2.1). Thus, ag eemen is cha ac e ized by he use o loci as he means
o exponence whe he he e b be spa ial o ag eeing, in he con ex o
mo emen be ween wo loci. As we saw in sec ion 3.2.1.4, he dis inc ion
be ween spa ial and ag eeing e bs is di icul o main ain, and many
p oposals ha e g ouped oge he bo h ypes o e bs as in lec ing e bs (c .
Quad os 1999), so his does no seem o be a d awback. The second
consequence is p oblema ic: since mo emen is closely ied o he (seman ic)
SOURCE-GOAL ela ionship (whe he his be in e ms o mo ion o ans e ),
his makes Mei ’s accoun o ag eemen essen ially seman ic in na u e. The
o mal de ice o es ic ing ag eemen o a p edica e (DIR) wi h a gumen s
ha happen o ha e speci ic seman ic oles (i.e. SOURCE and GOAL) explains
why a syn ac ic p ocess (ag eemen ) is seman ic in appea ance and p o ides a
syn ac ic amewo k o ag eemen , bu also means ha ag eemen is limi ed
o he seman ic con ex o SOURCE and GOAL. Fo spa ial e bs, his is no
issue since a e b ha desc ibes mo emen is necessa ily om SOURCE o
GOAL.
19
Howe e , Quad os & Que (2008) poin ou ha (in LSC and Lib as)
he second a gumen o an ag eeing e b is no always a GOAL, bu may be a
THEME, as in e bs like PRESS o INVITE. Fu he mo e, a ious ag eeing e bs,
such as CHOOSE o SUMMON, do no ha e a clea sense o ans e .
Mei ’s accoun does no add ess he ques ion o he ea u es in ol ed in
he ag eemen p ocess and limi s i sel o s a ing ha he e e en ial ea u es
o he language (R-loci) appea as he ag eemen ma ke s. This lea es he
issue o wha e e en ial ea u es sign languages use o a heo y o e e ence.
Howe e , since he ea u es in ol ed in ag eemen a e cen al o he p ocess,
and as we ha e seen (in sec ion 3.1), e e ence in sign language shows e y
pa icula cha ac e is ics, his issue will be examined ca e ully in chap e s 6
19
A possible excep ion is he use o mo emen o ace he shape o a e e en , such as a bend
in a oad, such ha he mo emen desc ibes a pa h and no mo ion be ween a SOURCE and a
GOAL. This is no mally achie ed wi h classi ie cons uc ions (see sec ion 1.2), which use a
s ongly isomo phic mapping o he signing space on o eal space. These s uc u es a e
usually conside ed dis inc om spa ial e bs and he issue is no deal wi h in his hesis, bu
he ques ion o how he use o space in classi ie s uc u es in e ac s wi h ha o lexical e bs
dese es u he a en ion.
Ag eeing e bs 127
and 7, when we assess hese ag eemen -like phenomena in e ms o he
heo e ical amewo ks de eloped o spoken language da a.
3.2.2.5. Summa y
Backwa ds e bs show he unusual p ope y o in e ing he associa ion
be ween he s a /end poin o he sign and he subjec /objec a gumen .
Simila lexical i ems show up as backwa ds e bs in many di e en sign
languages, hough he e is a ia ion om one language o ano he . These
e bs make clea ha he e a e wo mechanisms a play in ag eeing e bs: on
he one hand, he mo emen be ween he loci associa ed wi h he e b’s
a gumen s; and on he o he , he acing o he hands owa ds he locus
associa ed wi h he objec . Backwa ds e bs a e a subse o he ag eeing e bs
in ha hey ma k ag eemen o wo a gumen s. The nex sec ion looks a
e bs ha modi y spa ially o ag ee wi h jus one a gumen .
3.2.3. Single a gumen ag eemen
The e bal ag eemen mechanism we ha e examined so a has in ol ed
mo emen (o o ien a ion) be ween wo poin s in space, and he e b ag ees
wi h wo a gumen s. In sec ion 3.2.1.2 we saw cases o e bs ma king jus one
a gumen (in he con ex o ag eemen ma ke omission o de ec i e e bs)
bu hese we e si ua ions in which ma king o a possible second a gumen
was omi ed o blocked; he e bs a e di ec ional bu o some eason one o
he a gumen s is no ma ked. In con as , in single a gumen ag eemen he
e b is no di ec ional bu localizable: he e b is a icula ed a he locus
associa ed wi h he a gumen . As such, he spa ial mechanism employed by
he e b only e e allows o one a gumen o be ma ked, and only a single
ag eemen slo exis s. Example (20) shows how he NGT e b WAIT can be
a icula ed a a poin in he signing space associa ed wi h a e e en , hus
showing single a gumen ag eemen .
NGT (Zwi se lood & an Gijn 2006: 198. Images ep oduced wi h kind pe mission
om Ox o d Uni e si y P ess.)
(20)
a) WAIT
b) WAITx
ci a ion o m
‘He wai s.’
128 Ag eemen in sign languages
The obse a ion ha no all plain e bs a e body-ancho ed, and can be
a icula ed a di e en poin s in he signing space has appea ed equen ly in
he li e a u e: o example, Fische & Gough (1978: 22) men ion
“inco po a ion o loca ion” in e bs in ASL, and Be gman (1980) uses he e m
“localiza ion” o e bs in SSL. Be gman’s e m o his mechanism i s wi h
he de ini ion o localiza ion al eady adop ed (in sec ion 3.1.1): a sign is
a icula ed a a speci ic poin in he signing space. I use he e m o desc ibe
he modi ica ion o a loca ion o a sign; i does no e e o he unc ion o ha
modi ica ion (see n. 3 o u he cla i ica ion).
Some au ho s (Be gman 1980, 1990; Smi h 1990; Zwi se lood & an Gijn
2006) conside his phenomenon o be pa o he e bal ag eemen sys em, on
a pa wi h wo-place ag eeing e bs. Howe e , much o he li e a u e on
e bal ag eemen in sign languages a oids his phenomenon and es ic s he
discussion o ag eemen o p o o ypical and backwa ds ag eeing e bs as
desc ibed in he p e ious wo sec ions. Fo example, a ecen landma k pape
on ag eemen in ASL makes passing e e ence o he issue:
Al hough his class o [plain] e bs is conside ed non-ag eeing, some o
hem can ac ually be signed in a locus associa ed wi h a loca ion o an
e en (e.g. WANT, BUY, and LEAVE-ALONE). (Lillo-Ma in & Meie 2011:
106)
20
This has led o a ce ain amoun o con usion in he ield and dese es
a en ion o make he issues in ol ed explici . The exclusion o e bs
a icula ed a a single loca ion om he analysis o ag eemen s ems om he
obse a ion by Padden (1990) ha o ansi i e e bs such ma king is
ambiguous since i could be o he subjec o he objec . This is illus a ed by
(21), in which he ASL e b WANT is localized a di e en poin s in he signing
space. The loci may be associa ed wi h he e b’s subjec a gumen s o i s
objec a gumen s, and he in e p e a ion depends on he con ex . Since he
ma king o he a gumen is no sys ema ic, Padden claims ha his canno be
a case o ag eemen .
ASL (Padden 1990: 121)
(21)
WOMAN
WANTx
WANTy
WANTz
‘The womeni,j,k a e each wan ing.’
‘The woman wan s hisi, and hisj, and hisk.’
20
La e in he same a icle, Lillo-Ma in & Meie s a e ha “no e bs ma k ag eemen wi h
only he subjec (indeed, in ansi i e e bs a e no di ec ional)” (2011: 126) hus iden i ying
ag eemen wi h di ec ional e bs.
Ag eeing e bs 129
Howe e , ollowing up obse a ions made by Engbe g-Pede sen (1993), Mei
(1998b: 95) poin s ou ha he i s eading o (21), in which he subjec
a gumen is ma ked by ag eemen , occu s in a speci ic discou se con ex ,
namely, when se e al e e en s in he discou se a e being compa ed. Such
con ex s gi e ise o wha Engbe g-Pede sen calls p agma ic ag eemen , in
which o e iding p agma ic conside a ions indica e which a gumen is being
iden i ied by a icula ing he e b a a gi en locus. Ou side his speci ic
discou se con ex , Mei shows ha (in ISL) single a gumen ag eemen ma ks
he in e nal a gumen o he e b, i.e. he subjec o in ansi i e e bs and he
objec o ansi i es, as shown in he examples in (22).
ISL (Mei 1998b: 94)
(22) a.
STICK
IXx
CL(b eak)x
‘The s ick b oke.’
b.
BOY
IXx
GROW-UPx
‘The boy g ew up.’
c.
STICK
IXx
IX1
CL(b eak)x
‘I b oke he s ick.’
d.
POLICEMAN
IXx
THIEF
IXy
CATCHy
‘The policeman caugh he hie .’
This e inemen o he cha ac e iza ion o single a gumen ag eemen (by
excluding he ambiguous cases due o speci ic discou se conside a ions)
allows Mei o iden i y a mechanism ha uses space o sys ema ically ma k
a gumen s a he clausal a he han he discou se le el.
21
As such, hese e bs
ha can be localized o ma k an a gumen (and which clea ly do no in ol e
he DIR p edica e pos ula ed by Mei o ag eeing e bs) will be conside ed as
a possible mani es a ion o (spa ial) ag eemen when looking a he LSE da a
in chap e 5 and when assessing ag eemen in chap e s 6 and 7.
3.2.4. Summa y
This sec ion has looked a e bs in sign language ha ha e been desc ibed as
showing ag eemen by means o spa ial in lec ion. Mos a en ion in he
li e a u e has been gi en o di ec ional ag eeing e bs, which ma k o wo
a gumen s by mo ing om he locus associa ed wi h one a gumen o ha
associa ed wi h ano he . A small subse o hese e bs, backwa ds e bs,
21
Un o una ely, Mei does no analyse his cons uc ion in he wo k on ag eemen and
exp essly ocuses on di ec ional e bs.
130 Ag eemen in sign languages
shows he p ope y o mo ing om he objec o he subjec locus, con a y o
he p o o ypical mo emen om subjec o objec locus displayed by mos
ag eeing e bs. These ag eeing e bs ha e been he ocus o a g ea deal o
esea ch since hey display a se ies o unusual p ope ies, many ela ed o he
dis ibu ion o ag eemen in sign languages: no all e bs can show such
ag eemen , and hose ha can may omi ag eemen o ha e de ec i e
in lec ional pa adigms. Addi ionally, ag eemen has a e y s ong seman ic
la ou , since i commonly depic s ans e om a SOURCE o a GOAL. This
si s uneasily wi h a cha ac e iza ion o ag eemen as occu ing in a speci ic
con igu a ional o syn ac ic con ex . These ac s lead Mei (1998b, 2002) o an
analysis o ag eeing e bs in ISL in e ms o a speci ic spa ial p edica e, DIR,
which shows ag eemen and uses wi h seman ically app op ia e e bs ( hose
ha exp ess ans e ). As we ha e seen, his p o ides a syn ac ic amewo k
o a seman ically d i en p ocess, bu may commi ag eemen o a limi ed
seman ic con ex ha da a om o he sign languages sugges is oo es ic i e.
Finally, we ha e also looked a he phenomenon o single a gumen
ag eemen , in which a e b is localized o ma k jus one o i s a gumen s. This
mechanism has been gene ally o e looked in he li e a u e, bu appea s o
show a sys ema ic use o space o ma k a e b’s a gumen , in he same sense
ha (p o o ypical and backwa ds) ag eeing e bs do. As such, i will be
included in he possible lis o candida es o ag eemen o be assessed in LSE.
The nex sec ion con inues o look a e bal ag eemen , bu in he
con ex o e bal auxilia ies. The di e en auxilia ies ha ha e been iden i ied
o di e en sign languages unc ion p incipally o bea ma ke s o
ag eemen , and so a e highly ele an o he issue unde discussion.
Fu he mo e, he in e ac ion be ween lexical and auxilia y e bs p o ides
impo an insigh in o he na u e o he spa ial ag eemen p ocess in sign
languages.
3.3. Ag eemen auxilia ies
In he p e ious chap e , sec ion 2.2.3.1, we saw ha auxilia y e bs a e
common a ge s o ag eemen . In spoken languages, auxilia y e bs
gene ally se e o show in o ma ion ela ing o ense, aspec , modali y,
nega i e pola i y and oice, and he appea ance o ag eemen on hese
elemen s is mo e o a syn ac ic “acciden ” (S einbach & P au 2007). In sign
languages, howe e , a ious elemen s ha e been iden i ied ha se e as a
e bal auxilia y bu wi h he main unc ion o ma king subjec /objec
ag eemen when he lexical e b is no capable o doing so (i.e. when i is a
plain e b). Consequen ly, hese elemen s a e e e ed o as ag eemen
Ag eemen auxilia ies 131
auxilia ies o by he mo e speci ic e m “subjec objec ag eemen ” (SOA)
auxilia ies, coined by S einbach & P au (2007: 308). These elemen s occu
oge he wi h a lexical e b ( o o m a monoclausal e b ph ase) and pe o m
he g amma ical unc ion o ma king ag eemen , and as such hey all unde
he de ini ion o a e bal auxilia y p oposed by Ande son (2006: 5; see sec ion
2.2.3.1 o de ails).
Ag eemen auxilia ies ha e been iden i ied o a ious sign languages,
and hey can be ca ego ized in o h ee di e en ypes based on he
in e ela ed c i e ia o o m and o igin. The i s ype, deal wi h in sec ion
3.3.1, no mally glossed as AUX, is he mos equen c oss-linguis ically (based
on cu en da a) and consis s o an indexical elemen ha appea s o be
de i ed om conca ena ed p onominal o ms. The second g oup o
auxilia ies, desc ibed in sec ion 3.3.2, is de i ed om ull lexical e bs ha
ha e unde gone seman ic bleaching and aken on a mo e unc ional ole
wi hin he clause. The hi d ype o auxilia y (sec ion 3.3.3), PAM, is simila in
unc ion o AUX bu i s use ends o be mo e es ic ed and i appea s o ha e
i s o igins in he nominal PERSON. This sec ion p o ides an o e iew and
desc ip ion o each o hese ypes o ag eemen auxilia y in u n. (Fo a mo e
de ailed o e iew o auxilia ies in sign languages see Sapoun zaki (2012), and
o a discussion o he sou ces om which hey g amma icalize, see S einbach
& P au (2007)).
3.3.1. AUX
Many sign languages ha e an auxilia y elemen o ma k e bal ag eemen
ha consis s o signalling he loca ion associa ed wi h he subjec ollowed by
he loca ion associa ed wi h he objec . In o m, he hand adop s he B shape
ypically used o poin ing and he auxilia y looks like wo consecu i e poin s
joined by some so o a ced mo emen .
This AUX o m was i s desc ibed o TSL by Smi h (1990), shown in
igu e 3.1, and has subsequen ly been iden i ied o A gen ine Sign Language
(LSA) (Massone 1994; Massone & Cu iel 2004), Japanese Sign Language (NS)
(Fische 1996), Lib as (Quad os 1999; Quad os & Que 2008), Indo-Pakis ani
Sign Language (IPSL) (Zeshan 2000a, 2003a), LSC (Que e al. 2005) and G eek
Sign Language (GSL) (Sapoun zaki 2005).
132 Ag eemen in sign languages
1AUXx
Figu e 3.1 The indexical ag eemen auxilia y desc ibed o se e al sign languages, he e
showing mo emen om i s pe son o a non- i s pe son locus in TSL. (Image ep oduced
om Smi h 1990: 217, wi h kind pe mission om Uni e si y o Chicago P ess.)
The de ails o he beha iou o AUX a y om language o language, bu
basically he auxilia y appea s nex o he lexical e b and ma ks he
ag eemen o ha e b, as shown in examples (23a) and (24). AUX equen ly
occu s wi h plain e bs and se es o exp ess (spa ial) ag eemen ha he e b
i sel canno in lec o show due o i s phonological limi a ions. In some sign
languages, such as GSL and NS, AUX can accompany only plain e bs o
ag eeing e bs ha a e unin lec ed o ag eemen , and double ag eemen
(mani es ed on bo h he main e b and he auxilia y) is no possible, as
exempli ied by he ung amma ical NS sen ence (23b), in which AUX appea s
oge he wi h he in lec ed ag eeing e b HIT.
NS (adap ed om Fische 1996: 107)
(23) a.
CHILDx
TEACHERy
LIKE
xAUXy
‘The child likes he eache .’
b.
*MOTHER
FATHER
xHITy
xAUXy
IPSL (adap ed om Zeshan 2003a: 172)
q
(24)
UNDERSTAND
xAUX1
‘Do you unde s and me?’
Ta ge s 235
ph ases os ga os e melhos (‘ he ed o ks’) o as acas e melhas (‘ he ed
kni es’) is e y simila in o m o he de ini e a icles os/as bu his does no
mean ha he ma king in ol es inco po a ion o he a icle in o he adjec i e;
a he , bo h o ms make use o same unde lying ag eemen mechanism. This
boils down o a dis inc ion be ween he exponence and he a ge o
ag eemen , bu also equi es an analysis o space as he e e en ma king
sys em a wo k in sign languages.
Essen ially, he e a e wo issues a s ake he e. Fi s ly, whe he o no he
ma king is some so o p onoun, and secondly, wha he na u e o he
ma king is, an a ix o a cli ic. The o egoing discussion has shown ha ha
spa ial ma king beha es qui e di e en ly o a p onominal o m. A u he
obse a ion is ge mane o bo h issues: loca ion in i sel does no cons i u e
independen phonological ma e ial. As men ioned abo e, a loca ion is no a
p onoun, bu a he a pa o a p onoun. By he same oken, a loca ion canno
cli icize o o m an a ix wi h a e b as i does no ha e enough phonological
ma e ial o exis independen ly.
LSE (Ix_lion 1:11)
(4)
D hand
ansi ion(IX1)
IX1
1HELPX
ND hand
ansi ion(1HELPX)
1HELPx --------------------------------------
‘I’ll help you.’
Cli icized p onouns ha e ce ainly been obse ed in sign languages and wo
di e en mechanisms ha esul in cli icized p onouns ha e been desc ibed
(Sandle & Lillo-Ma in 2006). In he case o assimila ion, he p onoun adop s
he handshape o he neighbou ing e b; coalescence, on he o he hand,
in ol es he p oduc ion o he p onoun be o e o a e a wo-handed sign. The
LSE da a p esen bo h ypes o cli icized p onoun. Example (4) shows a
coalesced p onoun, isible on he dominan hand du ing he p io ansi ion
and he i s hold o he bimanual e b HELP as indica ed by he handshape
(wi h he index inge selec ed), demons a ing ha he p onoun is clea ly
236 Typological pe spec i e o LSE ag eemen
educed bu s ill p esen . In con as , when e bs a e spa ially ma ked, no
p onoun (o emains o a p onoun) is isible.
Mo e ecen wo k has con inued o a gue ha he spa ial ag eemen
ma king in sign languages beha es mo e like a p onominal cli ic (Kelle 2001),
d awing on e idence om i s syn agma ic p ope ies, dis ibu ion,
op ionali y, in en o y and ealiza ion (Ne ins 2011). On he o he hand, he
e idence p esen ed abo e – bo h Co be ’s diagnos ics and he a gumen ha
he ma king does no cons i u e enough independen phonological ma e ial –
sugges s ha he ma king is no de i ed om a p onominal o m. While
spoken languages commonly use p onouns as ag eemen a ixes, he ges u al
modali y makes a ailable he use o space o e e encing and his is ec ui ed
o he ag eemen . In bo h cases, a e e en ial mechanism becomes pa o he
e e ence-ma king sys em on e bs. Howe e , while spoken languages end
o do his wi h a ixes, he o m o ma king in sign languages is di e en .
The means o exponence o spa ial ma king in LSE, ha is o say, he
inco po a ion o a loca ion o loca ions in a sign, appea s no o in ol e he
addi ion o phonological ma e ial, as is he case wi h a ixes, bu a he a s em
al e na ion (Ma hu 2000). Phonological models p oposed o sign languages
include loca ion slo s ha can be assigned a gi en alue. The spa ial
ag eemen p ocess, hen, in ol es assigning a speci ic alue o one o he
loca ion slo s in he phonological ma ix (one in he case o single a gumen
ag eemen ; wo o ag eeing e bs). Howe e , i should be poin ed ou ha
he dis inc ions in loca ion wi hin he signing space (whe e loca ion
assignmen occu s) a e no no mally ele an as a as he phonology o
lexical i ems is conce ned. Mos phonologically con as i e loca ions occu on
(o nea o) he body. As such, he ag eemen sys em makes use o dis inc ions
in o m ha a e no ele an o he phonological sys em. On a specula i e
no e, i is possible ha loca ions ha a e no ancho ed o some poin on he
body a e no speci ic enough (and he e o e pe haps no s able enough) o
he phonological con as s needed by he lexicon. As such, sign language
phonologies make use o loca ion bu in a es ic ed domain. Howe e , he
spa ial medium o e s a a g ea e numbe o loca ions. Loca ions in he
signing space can be exploi ed isomo phically o spa ial desc ip ions, bu
also o e he possibili y o c ea ing a e e ence acking mechanism. One can
imagine ha mo i a ed spa ial loca ions ga e way o abs ac indexing, by
means o a g amma icaliza ion p ocess in ol ing he seman ic bleaching o
he loca i e meaning o poin s in signing space o some basic loca i e e b BE-
AT (Wilbu 1999). Such ideas ha k back o seminal wo k by Shepa d-Kegl
(1985) on space and loca i es in ASL, bu we ha e no his o ical da a om LSE
o suppo his specula ion. Howe e , e idence om an e ol ing language
Ta ge s 237
sys em, Nica aguan Sign Language, sugges s ha , a leas in he case o
manual poin ing, loca i e uses a e p esen a ea ly s ages o he language and
only la e do mo e abs ac , nominal uses o poin ing appea (Senghas &
Coppola 2011). Fu he mo e, his use o space o bo h loca i e and nominal
e e ence ies in well wi h he ac ha he dis inc ion be ween spa ial and
ag eeing e bs is o en blu ed, as men ioned in sec ion 3.2.1.4. I also helps o
explain why a e e en ial locus can egain i s spa ial meaning when ha
in o ma ion becomes ele an . This is demons a ed by Liddell’s (2000)
no o ious examples o he ype ‘I asked a ( all) man’, in which he sign ASK
mo es upwa ds in he signing space o indica e he ( ela i e) heigh o he
pe son e e ed o ( o a o mal seman ic cha ac e iza ion o his ole o spa ial
iconici y in sign languages see Schlenke 2011). I e u n o examples o his
ype and o he dis inc ion be ween spa ial and ag eeing e bs in he
discussion o condi ions in sec ion 6.5.
A unique p ope y o he means o exponence o spa ial ma king on
sign language e bs is ela ed o a p ope y men ioned abo e in he sec ion on
con olle s (6.1). The o m made mani es on he a ge is no some ea u e o
he con olle pe se, such as gende , o e en pa o he o m o he con olle ,
as occu s wi h adical alli e a i e ag eemen (desc ibed in sec ion 2.2.3.3).
Ra he , he a ge displays a s em al e na ion based on a o m ha has been
assigned o he con olle . In some senses, his is mo e simila o he ype o
alli e a i e ag eemen seen in Ban u languages, in which a gende ma ke on
he con olle appea s on he a ge o he ag eemen p ocess, as shown in
example (5).
Swahili (Welme s 1973: 171, ci ed in Co be 2006: 87)
(5)
ki-kapu
ki-kubwa
ki-moja
ki-li-anguka
SG-baske (G7/8)
G7-la ge
G7-one
G7-PST- all
‘One la ge baske ell.’
G=gende class
The e a e, o cou se, impo an di e ences. Fi s ly, he gende p e ix ki- in
Swahili is a e lex o pa o he lexical en y o he noun, and whene e he
noun appea s, i s gende is pa and pa cel o he syn ac ic elemen . In LSE, in
con as , a noun may appea wi hou being localized in space, and only in
ce ain si ua ions, which appea o be discou se dependen , will a loca ion be
assigned. Fu he mo e, in LSE a noun may be assigned one loca ion in a
ce ain s e ch o discou se and a di e en loca ion in ano he . Secondly, in
LSE, no all con olle s can admi he o m o he ma ke : body-ancho ed
signs canno be localized and he associa ion be ween he sign and he locus
238 Typological pe spec i e o LSE ag eemen
mus be achie ed by some o he means, o en by poin ing (as desc ibed in
sec ions 3.1.1 and 5.1.1). This makes i appa en ha he a ge does no copy
di ec ly om he o m o he con olle , as appea s o happen in adical
alli e a i e ag eemen (see sec ion 2.2.3.3), bu a he om some ea u e ha
has been associa ed wi h he con olle . Finally, he o m o he ma ke is
a ixal in Swahili and a s em al e na ion in LSE.
6.2.4. Mul iple exponence
Be o e closing his sec ion on a ge s and he way in which hey exp ess
ag eemen , I e u n o he issue o mul iple exponence. We ha e al eady
men ioned mul iple exponence ea lie in his sec ion as e idence o
conside ing spa ial ma king a mani es a ion o ag eemen a he han some
so o inco po a ed p onoun. In ha case, I e e ed o he appea ance o
spa ial ma king on di e en elemen s, namely, he lexical e b and he
auxilia y. Ano he ype o mul iple exponence occu s when a single a ge has
mo e han one ma ke o a gi en a gumen . Al hough he da a o LSE a e
scan in his espec , he possibili y ha spa ial loca ions a e ma ked no only
by he manual componen o a sign bu also by non-manual ea u es, such as
eye gaze o head il s (as desc ibed in sec ions 3.4 and 5.5), opens up he
possibili y o such mul iple exponence. Gi en ha he mul iple a icula o s
allow o simul aneous mul iple exponence, his may p o ide a modali y-
speci ic cha ac e is ic o ag eemen in sign languages, and dese es g ea e
in es iga ion.
1
6.2.5. Summa y
This sec ion has assessed he a ge s o ag eemen in LSE and how ag eemen
is mani es on hose a ge s. The LSE da a show ha he spa ial ma king we
a e conside ing as ag eemen , bo h on ag eeing e bs and o he elemen s
(unde he guise o single a gumen ag eemen ), occu s on a ange o
elemen s ha ha e also been a es ed o spoken languages. In he e bal
domain, he dis ibu ion o in o ma ion be ween he lexical and auxilia y e b
can be cha ac e ized as he spli /double ype, al hough, in con as o wha is
seen in mos spoken languages, he auxilia y seems o be specialized o
ca ying ag eemen in o ma ion since i does no in lec o aspec . The means
o exponence o his ag eemen mechanism in LSE is no a o m o cli icized
1
An al e na i e cha ac e iza ion o (manual and non-manual) mul iple ma king is as a
ci cum ix (Thompson, Emmo ey & Kluende 2006). While his e lec s he ac ha he same
in o ma ion is mani es as wo iden i iably di e en pa s, he no ion o a ci cum ix (a p e ix
and su ix) is di icul o econcile wi h he manual ma king as a s em al e na ion (and no an
a ix), and ails o cap u e he simul aneous na u e o he ma king.
Domains 239
p onoun, con a wha has been sugges ed o o he sign languages, bu a he
a s em al e na ion h ough he speci ica ion o a phonological ea u e
associa ed wi h he con olle . This ag eemen mechanism elies upon a
speci ic use o (loca ions in) he signing space and does no appea o ha e an
exac pa allel in any spoken language, al hough I ha e poin ed ou simila i ies
(and di e ences) wi h alli e a i e ag eemen . The use o spa ial loca ions is a
s ongly indexical e e encing mechanism ha pe ades he language well
beyond a es ic ed class o e bs.
6.3. Domains
A use ul di ision o make in di e en domains o e bal ag eemen is
be ween clause-in e nal ag eemen , in which he con olle and a ge a e
wi hin he same clause, and ag eemen beyond he clause, as se ou in sec ion
2.2.4. We will assess he LSE da a wi hin he con ex o each in u n. Wi hin
he clause, LSE c ea es ag eemen ela ionships be ween he e b and i s
a gumen s as has been a es ed o spoken languages, bu also shows
ag eemen ma king wi h a ious a ypical elemen s, no ably wi h loca i e
a gumen s, a phenomenon ha has been epo ed o some spoken languages
bu is none heless a a i y. Beyond he clause, LSE displays spa ial ma king
o “p agma ic ag eemen ”, as p e iously desc ibed o o he sign languages,
and o p onouns, al hough each appea s o in ol e a sligh ly di e en use o
he signing space.
6.3.1. Clause-in e nal ag eemen
In he con ex o clause-in e nal ag eemen , e bs gene ally ag ee wi h hei
a gumen s, wi h a g ea e endency o ag ee wi h mo e p o o ypical
a gumen s, such as subjec and objec , a he han indi ec objec o oblique
a gumen s. As we saw in sec ion 2.2.4, he Accessibili y Hie a chy
( ep oduced he e in (6)) es ablishes he ypes o ag eemen domain ha a e
p e equisi es o o he s o be p esen in a language.
(6)
subjec
>
di ec objec
>
indi ec objec
>
oblique
>
geni i e
>
objec o
compa ison
The Accessibili y Hie a chy (Keenan & Com ie 1977: 66)
LSE obeys his hie a chy since i exp esses ag eemen spa ially wi h subjec ,
di ec objec and indi ec objec . (The use o localiza ion wi h spa ial e bs
could be conside ed a case o oblique ag eemen .) In e es ingly, ag eemen o
bo h geni i es (possessi es) and objec s o compa ison occu s wi h he gene ic
ag eemen auxilia y AUX, as desc ibed in sec ion 5.3.1.
240 Typological pe spec i e o LSE ag eemen
Al hough he da a sugges ha LSE ollows he same pa e ns ha ha e
eme ged om c oss-linguis ic compa isons as a as ag eemen domains
a ailable o he language a e conce ned, he e a e some anomalies. Fi s ly,
di ansi i e e bs in LSE end o ag ee wi h indi ec objec s a he han di ec
objec s. Howe e , his endency has also been epo ed o many spoken
languages, which a e said o dis inguish be ween p ima y and seconda y
objec s a he han di ec and indi ec objec s (see sec ion 2.2.4 o de ails), and
so i s in wi h es ablished ypological pa e ns. Con e sely, a ai common
among sign languages, bu unusual in spoken languages, is he ac ha he
objec a gumen seems o be much mo e salien han he subjec a gumen .
This is e lec ed in he equency o ag eemen ma ke omission o he
subjec , while he objec a gumen ma king is main ained. This issue will be
examined u he in he con ex o op ionali y in sec ion 6.5.
Remaining wi hin he clause, ag eemen in spoken languages has also
been desc ibed be ween he e b and non- ypical elemen s, exempli ied in
sec ion 2.2.4 by possesso aising and copying- o-objec o ma ions, in which
an a gumen no cen al o he e b’s a gumen s uc u e is ma ked as i i
we e such an a gumen . The da a do no e eal ha LSE has any hing like
possesso aising, bu a pa allel may be ound in he ambigui y in ol ed in
using space o e e ence. The examples in (7) show how he same e b may
ag ee wi h an objec /loca i e, as in (a), o wi h a pe son, as shown in (b). (No e
ha he e b STEAL is a backwa ds e b, so he objec is ma ked by he i s
subindex.)
LSE
(7) a.
BIKEX
DISAPPEARx
xSTEALy
‘The bike’s been s olen ( om he e).’
b.
BIKEX
1STEALy
‘The bike’s been s olen ( om me).’
Fu he mo e, in he case o spa ial e bs in LSE, he e b ag ees wi h a
loca i e elemen . In sec ion 2.2.2, we saw an example om Chicheŵa, a Ban u
language ha admi s loca i e a gumen s as he con olle o e b ag eemen ,
and da a o his so led Thompson, Emmo ey & Kluende (2006) o p opose
ha ASL has loca i e ag eemen , and ha loca i es a e a gumen s ( a he
han adjunc s) o spa ial e bs in sign languages.
Con inuing wi h ag eemen wi h non- ypical elemen s, spa ial ma king
may also be used o connec seman ically o p agma ically ela ed concep s
o which i is di icul o cha ac e ize a speci ic syn ac ic ela ionship.
Engbe g-Pede sen (1994) iden i ied his ype o “p agma ic ag eemen ” o
Domains 241
DTS, as men ioned in sec ion 3.2.3, and his also occu s in he LSE da a. Fo
example, in a s e ch o discou se desc ibing he schooling, a signe uses he
space on he le o e e o he school, and he cen al signing space o e e o
high school. Thus, a icula ing a sign such as FOURTH-YEAR in he le -hand
space means ‘(in) ou h g ade a school’. In his sense, space is exploi ed o
associa e ela ed concep s o ideas, and gene ally in ol es he use o b oad
egions, a he han loca ions, in he signing space. An explana ion o
p agma ic ag eemen will be o e ed in sec ion 7.2.3 wi hin he o mal analysis
o spa ial ag eemen p oposed in he nex chap e . This so o use o space
may s e ch o e a leng h o discou se, and hus b ings us o domains beyond
he con ines o a single clause.
6.3.2. Ag eemen beyond he clause
A clea candida e o ag eemen beyond he clause, al eady men ioned o
spoken languages in sec ion 2.2.3, is p onouns, which display he same
ea u es (o he an eceden ) ha ypically show up on ag eemen ma ke s.
Tha (clause-in e nal) ag eemen and an eceden -anapho a employ common
unde lying mechanisms ce ainly appea s o be he case in LSE, which makes
use o he same spa ial ma king o bo h. In con as o he p agma ic
ag eemen men ioned in he p e ious sec ion, which uses b oade , less
de ined a eas o he signing space, p onominal e e ence elies on mo e
speci ic poin s in he space in he same way ha localiza ion o e bs does.
Thus, any hopes o cha ac e izing clause-in e nal ag eemen ma king as a
clea ly dema ca ed use o space and beyond he clause ela ions as a mo e
ague use o space (as e idenced by p agma ic ag eemen ) a e dashed by he
exis ence o he locus-based spa ial ma king on p onouns.
2
Fu he mo e, e en when he domain o ag eemen ma king is es ic ed
o a single clause, he p ocess has a s ong discou se la ou since ma ke s a e
assigned o a gi en s e ch o discou se, and no iden i ied in he lexicon.
This con as s wi h he si ua ion o ag eemen in spoken languages, such as
he gende ma king in Ban u languages seen abo e in example (5), o which
he ma ke s depend upon lexically de ined ea u es. In he discussion o
ag eemen domains in spoken languages in sec ion 2.2.4, i was men ioned
ha many o he ins ances o ag eemen wi h non-a gumen s ha e been used
o show ha ag eemen is condi ioned by discou se-le el conside a ions
ela ed o highligh ing salien anima e e e en s, o e en ha ag eemen is
essen ially a discou se phenomenon. In he case o LSE, spa ial ma king
2
A de e mining ac o in his espec , which is s a ing o be in es iga ed in a ious (Wes e n)
sign languages, is he ole o speci ici y in e e ence. Neidle e al. (2000) compa e de ini e and
inde ini e e e ence; Ba be à (2014) looks speci ically a speci ici y.
242 Typological pe spec i e o LSE ag eemen
appea s o ope a e a bo h le els, and se es o ma k clause-in e nal
a gumen s as well as discou se-le el ela ions. I is impo an o bea in mind
ha he same o mal mechanism may ha e di e en unc ions. Ag eemen
wi hin a speci ic syn ac ic con igu a ion may employ ma king ha is also
used o discou se le el e e encing, and he la e may no be subjec o he
same es ic ions.
A inal ype o ag eemen ela ion ha has caused much in e es in he
spoken language li e a u e is long dis ance ag eemen (desc ibed in sec ion
2.2.4), in which a e b ag ees wi h an a gumen no in he same clause. Wi h
he cu en LSE da a, I ha e ound no examples ha could con ibu e o he
deba e. Fu he mo e, since spa ial ma king can be bo h in aclausal and
in e clausal, inding a clea example o long dis ance ag eemen in LSE would
equi e showing ha a spa ially modi ied e b ag ees wi h a e e en om
ano he clause (which is no p esen in he e b’s clause ei he as a null opic
o a null a gumen , bo h o which a e possibili ies), and ha he spa ial
modi ica ion is no some discou se-le el mechanism. I lea e his a duous ask
o u u e esea ch.
In summa y, he spa ial ma king used in LSE ope a es bo h wi hin and
beyond he clausal le el, and is condi ioned by discou se in he sense ha he
ma ke s used a e c ea ed o a gi en s e ch o discou se. Al hough some uses
o space a he discou se le el (such as p agma ic ag eemen ) a e
cha ac e ized by a less ine-g ained di ision o he signing space, o he s,
no ably p onouns, use space in jus he same way as e bs ma k hei
a gumen s. Ne e heless, o m and unc ion should no be con la ed, and i
may be he case ha he same spa ial ma king is used o a ious unc ions, a
imes syn ac ically de e mined and a imes go e ned by b oade discou se
conside a ions. The nex sec ion looks a he ea u es ha a e in ol ed in LSE
ag eemen .
6.4. Fea u es and alues
The ype o in o ma ion, o ea u es, displayed in ag eemen p ocesses
gene ally alls in o h ee ca ego ies: gende , numbe and pe son. G amma ical
gende is no encoded in all languages, bu numbe and pe son a e aken o
be linguis ic uni e sals (al hough some excep ional cases do exis , as
explained in sec ion 2.2.5). Addi ionally, he e a e o he ea u es ha ha e
been conside ed o ake pa in he ag eemen p ocess, namely espec and
case. In his sec ion, we look a whe he each o hese ea u es is p esen in
LSE, and i so, how i is encoded.
Fea u es and alues 243
6.4.1. Gende
A i s sigh , LSE does no appea o code he no ion o gende in he e bal
ag eemen p ocess, a leas ce ainly no sex-based gende common o Indo-
Eu opean languages. The di e en hand con igu a ions used by classi ie
cons uc ions, which ca ego ize he e e en acco ding o i s physical
p ope ies, could be conside ed a ype o gende sys em and p oposals ha e
been made o classi ie s as an ag eemen mechanism in DGS (Glück & P au
1998) and speci ically as gende ag eemen in NGT (Zwi se lood 2003).
Such a mechanism would be mo e simila o he Ban u-s yle gende
ag eemen men ioned in sec ion 6.2.3 abo e, in he sense ha he gende
ma ke (i.e. he classi ie hand con igu a ion) depends upon he seman ics o
he e e en . In e ms o he c i e ia o he gende dis inc ion, hese a e pu ely
seman ic, since he phonological o m o a gi en noun does no a ec he class
i is assigned o. Howe e , in con as o he seman ic c i e ia o gende
sys ems in spoken languages, which end o be based on ei he sex-based
ca ego iza ion o on animacy, sign languages pay a en ion o he physical
(and mainly isual) p ope ies o he e e en s o nouns o classi y hem.
Al hough such an analysis could be applied o LSE (which also makes
use o classi ie cons uc ions, as desc ibed in sec ions 1.1 and 1.2), ou ocus
he e is on he use o spa ial ma king. This gende sys em, in con as , is
limi ed o modi ica ions o he hand con igu a ion, so I shall no pu sue an
analysis o gende ag eemen in o de o concen a e on spa ial ma king.
6.4.2. Numbe
Ano he ea u e o nominal elemen s ha may en e in o he ag eemen
p ocess is numbe . The mos common dis inc ion is be ween singula and
plu al, al hough some languages u he dis inguish be ween di e en
deg ees o plu ali y (dual, paucal, e c.). As we saw in sec ion 5.2.1, ag eeing
e bs in LSE may in lec o numbe o di e en ia e be ween dual, exhaus i e
and mul iple ma king, shown in igu e 6.1.
a
b
c
Figu e 6.1 In lec ional o ms o ma king o plu al objec s in LSE, as seen om abo e: a) dual
ma king; b) exhaus i e ma king; c) mul iple ma king.
244 Typological pe spec i e o LSE ag eemen
Addi ionally, single a gumen ag eemen ma ks numbe by means o
bimanual a icula ion o eduplica ion, somewha simila ly o he dual and
exhaus i e o ms o di ec ional ag eeing e bs.
An impo an dis inc ion mus be made he e be ween e bal and
nominal numbe . In sec ion 2.2.5, we saw ha numbe ma king on he e b
may e lec bo h he e en seman ics o he e b (e en numbe ) and he
pa icipan s (pa icipan numbe ), and as such i o e laps and in e ac s wi h
bo h aspec and nominal numbe as e lec ed in ag eemen . In con as o he
exp ession o (nominal) numbe in ag eemen , which displays he alue o he
numbe ea u e o he e b’s a gumen (s) on he e b, e bal numbe is
inhe en o he e b i sel and does no in ol e ag eemen wi h any hing else.
Thus, i is impo an o y o ease apa e bal and nominal numbe as
exp essed on he e b, as only he la e is an ins an ia ion o ag eemen .
O he ma ke s iden i ied o LSE, which ones e lec e bal numbe ,
and which nominal numbe ia ag eemen ? In ui i ely, he dual and he
exhaus i e o ms seem o a ec he e en s uc u e o he e b as hey imply
a leas one i e a ion o he e en , as al eady alluded o in he discussion o
numbe ma king in sec ion 3.2.1.1 based on Wilbu ’s E en Visibili y
Hypo hesis. Fo a lack o igo ous seman ic es s o con i m o e u e hese
in ui ions, le us u n o he diagnos ics sugges ed by Du ie (1986: 357-62),
desc ibed in sec ion 2.2.5.2, o dis inguishing e bal numbe om ag eemen
ma ke s:
i) e bal numbe ope a es on an e ga i e basis, e lec ing he numbe
o he mos di ec ly a ec ed pa icipan , which is he subjec o
in ansi i e sen ences (S) o he objec o ansi i e sen ences (P), and
his may con as wi h o he ma king on he e b (e.g. subjec
ma king, which ag ees wi h S and A).
Single a gumen ag eemen shows an e ga i e pa e ning, since he
e b ag ees wi h he in e nal a gumen (S o P) and he ac ha
numbe ma king does no occu o A is no e ealing. Di ec ional
ag eeing e bs, on he o he hand, a e no in ansi i e, so i is
di icul o make use o he S-P dis inc ion o hem. Howe e ,
conside ing he A a gumen (i.e. subjec o ansi i e e bs), he e
seems o be a clea dis inc ion be ween, on he one hand, he dual
and mul iple o ms, which may ma k A plu ali y, and exhaus i e
ma king, which canno . This sugges s ha he exhaus i e may be
e bal numbe since i appea s o be limi ed o S and P a gumen s.
ii) e bal numbe may ma k di e en alues o hose ma ked by
ag eemen , especially when e bal ag eemen is es ic ed by some
condi ion.
Fea u es and alues 251
dis inc ion in LSE. Wi hou his dis inc ion, he pe son ea u e does no play a
ole in ag eemen o in he e e en ial sys em o LSE. We now u n o he
p oposal ha loca ion is he ea u e encoded by ag eemen in sign languages.
Zwi se lood & an Gijn claim ha “ he only wo ele an ypes o
ϕ- ea u es in sign languages a e gende and loca ion” (2006: 195-6). (Thei
claim o gende is based on classi ie handshapes as desc ibed abo e in
sec ion 6.4.1 and I will no add ess ha issue u he he e.) Based on he
e e en ial use o space, he au ho s claim ha sign languages show loca ion
ag eemen . The e m “loca ion” may e e o a mo phophonological alue (a
locus in signing space), seman ic ea u e (loca i e) o o a speci ic e e en
(‘loca ion x’), bu i is no clea ha any o hese hings could be conside ed a
ϕ- ea u e. A ϕ- ea u e, as poin ed ou in sec ion 2.2.5, is abs ac and
ca ego ical in he sense ha he di e en alues o he ea u e p o ide a
means o classi ying di e en linguis ic i ems ( ypically nominal elemen s).
The e o e, loca ion mus be unde s ood as an abs ac alue ha can be
di e en ia ed om he ac ual subs an ia ion o a locus in he signing space;
his seems o be Wilbu ’s (2013) in e p e a ion when she dis inguishes
be ween a geome ic poin and an ac ual poin in space. I he e is some
abs ac loca ion ϕ- ea u e, wha alues does i ake? The ac ha loca ion
migh no ha e a lis able se o alues is dismissed by Wilbu (2013: 223) as
i ele an o i s linguis ic s a us; i is none heless he case ha o he ϕ- ea u es
ha e a small, closed se o alues.
5
Ano he issue o loca ion as a ϕ- ea u e wi h espec o he o he
ϕ- ea u es ha pa icipa e in ag eemen is he ea u e’s p ope ies. Since
ypical ag eemen ϕ- ea u es a e di ec , in he sense ha hey a e associa ed
wi h p o o ypical seman ics (gende ↔ seman ic class; numbe ↔ nume osi y;
pe son ↔ discou se ole), his should also be he case o abs ac loca ion.
Loca ion may be ins an ia ed as a poin in he signing space, and his may (o
may no ) ha e (loca i e) meaning, bu is his also he case o loca ion as a
ϕ- ea u e? Conside ing loca ion o be a ϕ- ea u e appea s o be a ma e o
hinking backwa ds: when aced wi h a use o loca ions in he o m o a
language, is i lici o pos ula e ha loca ion is pa o he abs ac g amma o
he language? This seems akin o sugges ing ha English has a sibilan
ϕ- ea u e because plu al nouns a e ma ked wi h a sibilan . O cou se, he
c i ical di e ence is ha an a gumen ’s being seman ically [±sibilan ] makes
li le sense, whe eas ha ing a gi en [loca ion] alue seems o make as much
5
I should poin ou ha Wilbu (2013) does no a gue o loca ion o be ea ed as a ϕ- ea u e,
bu a he de ends he linguis ic s a us o spa ial e e ence in sign languages. I ci e he wo k
he e as many o he a gumen s a e ge mane o he issue in hand.
252 Typological pe spec i e o LSE ag eemen
sense as being second pe son o masculine. This is a clea e ec o modali y,
since he isual channel makes i possible o employ o ms ha ha e a
ela i ely anspa en meaning ha can be p oduc i ely exploi ed. The
impo an ques ion is whe he he modali y e ec goes beyond me ely
in luencing how we analyse he language and ac ually c ea es a di ec
connec ion be ween unc ion (ϕ- ea u es) and o m (ag eemen ma ke s) ha
is unseen and unhea d o in spoken languages.
In sho , he p oposal ha sign languages make use o a loca ion
ϕ- ea u e is p oblema ic. Fi s ly, loca ion does no show he ypical p ope ies
o ϕ- ea u es o ha ing a small se o ca ego ical alues. Secondly, i in ol es
con la ing he su ace o m (loca ions in he signing space) wi h he
unde lying abs ac ea u e, which may be some hing else. While i could be
he case ha sign languages encode a g amma ical meaning ha is a ailable
o he languages’ medium, I p opose ano he possibili y. The ϕ- ea u e
esponsible o spa ial e e ence ma king, which Zwi se lood & an Gijn
(2006) call loca ion and which is indeed ealized by means o a loca ion in he
ag eemen p ocess, is an abs ac ea u e ha encodes iden i y bu is no
in insically ela ed o loca ion. This iden i y ϕ- ea u e may ake di e en
alues, which se e o dis inguish one discou se en i y om ano he . In his
sense, he ea u e is di ec since i is associa ed wi h he basic seman ics o
iden i y. Conce ning he lis abili y issue, in heo y he numbe o alues o he
ea u e is unlimi ed, bu in p ac ice i is limi ed o he easible numbe o
discou se en i ies. Admi edly, his does no educe he numbe o alues o a
closed se , bu is mo e limi ed han any possible loca ion.
A gi en alue o his iden i y ϕ- ea u e is ealized as a loca ion in he
phonological o m. The b eakdowns in indexicali y men ioned abo e, such as
s acking o a e e en appea ing in mo e han one locus, a e due o he
co espondence be ween he su ace o m and he unde lying ea u e alue.
This p oposal is based on he no ion o R-locus (in oduced in sec ion 3.1.3),
which also makes use o he co espondence wi h a (discou se-based)
e e en ial index in o de o esol e he “impe ec ” indexicali y displayed by
he spa ial ma king. The no ion o an iden i y ϕ- ea u e a emp s o in eg a e
he e e en ial mechanism in o he appa a us o ag eemen , and his p oposal
will be u he de eloped in chap e 7.
To summa ize, in spi e o he uni e sal p esence o he ca ego y o
pe son in spoken languages, he use o space o e e ence in LSE does no
employ his ea u e and as such i is no p esen in he e bal ag eemen
sys em. I ha e assessed he p oposal o an al e na i e ϕ- ea u e o sign
languages, namely loca ion, and ha e p oposed ha he abs ac ea u e is no
Fea u es and alues 253
ela ed speci ically o loca ion (e en hough i su aces in o m as a loca ion)
bu a he o iden i y.
6.4.4. O he ea u es: espec and case
In addi ion o he majo h ee ea u es o gende , numbe and pe son,
ag eemen sys ems may also ma k o he ea u es. As desc ibed in sec ion
2.2.5.4, espec and case a e ma ked in he ag eemen sys ems o some spoken
languages. This sec ion assesses whe he he da a in his s udy p o ide
e idence o pos ula e ha hese ea u es also exis in LSE, and concludes ha
he e is no e idence o ei he espec o case.
In spoken languages, espec is o en ma ked by means o ano he
ea u e, ypically pe son (in I alian a hi d pe son o m is used as a second
pe son espec o m, in bo h singula and plu al) o numbe (in F ench,
second pe son plu al is used as he espec o m o second pe son singula ).
O he languages ha e unique o ms o ma k espec ha a e no subsumed
unde ano he ea u e, as has been sugges ed o objec hono i ica ion in
Japanese (see sec ion 2.2.5.4 o de ails and examples). This dis inc ion occu s
in bo h he p onominal o ms and he e bal ag eemen ma king. Fo LSE, in
sec ion 5.1 i was no ed ha he hand con igu a ion o p onominal o ms
could change om he mo e common index- inge poin (B) o he la ]
handshape. Howe e , i seems unlikely ha his al e na ion is ma king
espec as a g amma ically encoded ca ego y o wo easons.
Fi s ly, he “ espec ul” o m is limi ed o use wi h physically p esen
e e en s, and is no used wi h non-p esen e e en s. This means ha he
o m is limi ed o deic ic con ex s and is no possible o anapho ic e e ence.
Such a s a e o a ai s could be expec ed o second pe son o ms, since he
add essee is in a iably p esen , bu he una ailabili y o he espec o m o
non-p esen hi d pe son e e ence sugges s ha he ] handshape is a s ylis ic
poli e o m o hose who a e p esen o pe cei e i . Secondly, and mo e
c ucially, he al e na ion is possible wi h i s pe son e e ence as well as
second and hi d pe son e e ence. Since espec ma ks he pe cei ed social
ela ionship be ween he speake and he e e en , i makes li le sense o ha e
a espec o m o e e o onesel . Again, his sugges s ha he use o he ]
handshape is pa o a poli e egis e in which ce ain o ms a e deemed mo e
app op ia e, a he han a g amma ically encoded ea u e.
6
Lea ing aside he p onominal sys em, is he e any e idence o a
espec ea u e in e bal ag eemen ma king in LSE? As was poin ed ou in
6
A pa allel can be ound in ce ain a ie ies o English, in which e lexi e p onouns a e
employed in mo e o mal se ings: ‘I mysel would like o ake his oppo uni y o hank
you sel es o you unde s anding in his ma e ’.
254 Typological pe spec i e o LSE ag eemen
sec ions 3.1.1 and 5.1.1, sign languages (LSE included) may assign e e en s o
di e en loca ions in he signing space acco ding o di e en con en ions
(Engbe g-Pede sen 1993). One such con en ion in ol es me apho ical
schemes in which he ela i e social s a us o e e en s is e lec ed by
di e ence in heigh in he signing space. Thus, o example, a e e en
pe cei ed o ha e highe social s anding han he signe , such as a doc o ,
would be assigned o a ela i ely high loca ion in he signing space, and any
e b using spa ial ma king wi h he doc o as an a gumen would inco po a e
ha highe loca ion (Nilsson 2008: 53; Ba be à 2012: 115). Could his use o
heigh be conside ed a espec ea u e in he ag eemen ma king? Again, as
was a gued o he p onominal o ms, his al e na ion does no in ol e a
g amma ical encoding o a espec ea u e.
Fi s ly, he ma king is no ca ego ical, bu in ol es he use o a
con inuum along he e ical axis o space. This con as s wi h he disc ee
alues displayed by he espec ea u e (and by ϕ- ea u es in gene al) in
spoken languages. The espec ea u e in LSE could be cha ac e ized as
showing ine -g ained dis inc ions in he same way ha some languages ha e
mul iple numbe alues a he han he simple singula -plu al dicho omy.
Howe e , his misses he poin ha he space is being exploi ed as a
con inuous scale in much he same way ha spa ial desc ip ions in sign
language in ol e an isomo phic mapping ha de ies a ca ego ical analysis.
Fu he mo e, his use o space is no obliga o y and is bes desc ibed as
a con en ion, e idenced by he ac ha i en e s in o play wi h o he
con en ions ha also in luence he choice o loca ion assignmen . Thus, o
example, in he case o physically p esen e e en s, conside a ions o iconici y
ha e g ea e weigh : a signe alking o a doc o will use a poin in space ha
coincides wi h he doc o ’s loca ion; i would no be possible o ma k espec
by using a loca ion abo e he doc o ’s eal loca ion. As such, his use o heigh
as an indica o o espec in he e bal ag eemen ma king does no o m a
consis en pa o he g amma . I is possible ha such a ea u e could be
g amma icalized o o m pa o he ag eemen sys em, bu he da a indica e
ha LSE cu en ly does no ha e a espec ea u e.
Case is o en conside ed a ele an ea u e in ag eemen p ocesses,
pa icula ly in he nominal domain. As poin ed ou in sec ion 2.2.5.4, al hough
case is no an inhe en ea u e o a con olle , i is closely ela ed o ag eemen
and can be ma ked oge he wi h mo e ypical ϕ- ea u es such as numbe and
gende . In he con ex o LSE, we saw in sec ion 5.3.3 when looking a e bal
auxilia ies ha he e is an elemen PERS (de i ed om he sign PERSON) ha
beha es e y simila ly o PRO[bC], a case-ma ked (da i e) p onoun desc ibed
o ISL (also de i ed om he sign PERSON). This leads us o ask whe he case
Fea u es and alues 255
is ma ked in he ag eemen sys em o LSE. E en hough he PERS elemen
desc ibed in sec ion 5.3.3 shows simila i ies o he ISL p onoun desc ibed by
Mei (2003; see sec ion 3.3.3), I also poin ed ou impo an di e ences. In
con as o PRO[bC], PERS does no appea o be p onominal, since i may appea
wi h co e e en ial p onouns. Addi ionally, he seman ic ole associa ed wi h
he ma ke is no consis en in LSE, since he e e en may be
AGENT/EXPERIENCER and no he jus THEME, as occu s in ISL. Thus, he
a gumen s o conside ing PRO[bC] o be a case-ma ked p onoun in ISL do no
hold o PERS in LSE. Mo e gene ally, a case analysis o sign languages has
been ques ioned due o he ac ha he e is no e idence o case mo phology
on nominal cons i uen s, he elemen s ha ypically ca y mo phological case
(Que 2011). In conclusion, he e is no e idence ha case plays a ole in he
ag eemen p ocesses o LSE.
6.4.5. Summa y
This sec ion has assessed which ea u es (and co esponding alues) a e
p esen in he spa ial ag eemen p ocess in LSE. O he ypical ϕ- ea u es
associa ed wi h ag eemen , namely gende , numbe and pe son, LSE makes
use o numbe alone. (Howe e , beyond he use o space, gende plays a ole
in he classi ie sys em in he shape-based classes dis inguished by di e en
hand con igu a ions.) Numbe in LSE can be bo h nominal and e bal: bo h
may be ma ked on he e b and I ha e made a poin o dis inguishing
be ween he wo. F om he poin o iew o ag eemen , nominal numbe (o
an a gumen ) ma ked on he e b esul s om ag eemen whe eas e bal
numbe exp esses a ea u e ha is inhe en o he e b i sel . Nominal numbe
is ma ked on he e b by means o loca ion(s) and o plu ali y his ansla es
in o mo emen ac oss mul iple loca ions; e bal numbe is exp essed wi h
mo emen by means o eduplica ion o he e b.
The lack o a pe son ea u e in LSE ag eemen is una guably a
ypological anomaly since pe son dis inc ions a e a es ed o all spoken
languages. This is a modali y e ec due o he e e en ial mechanisms
a ailable o a spa ial language. Al e na i e p oposals ha e sugges ed ha
loca ion is he ele an ea u e o sign languages, bu I ha e poin ed ou he
need o sepa a e he su ace ealiza ion (a loca ion in signing space) om he
unde lying unc ion. I ha e en a i ely p oposed ha a mo e basic ea u e o
iden i y, which se es o dis inguish one a gumen om ano he , could be a
wo k in LSE ag eemen , and his p oposal will be aken u he in sec ion 7.1.
E en wi h his adical di e ence in he se o ea u es a ailable o ag eemen
in LSE, he sys em s ill ul ils he gene aliza ion obse ed by Mo a csik (1978:
369) men ioned in sec ion 2.2.5.4: wha e e ea u es a e a ailable o he
256 Typological pe spec i e o LSE ag eemen
ag eemen sys em o a language will also be a ailable in he p onominal
sys em. The iden i y ea u e I p opose he e is also used by he p onominal
sys em since he indexical poin s also a ail o he loca ions ha a e he
phonological mani es a ion o he iden i y ea u e.
7
As such, LSE ollows he
es ablished pa e ns o he beha iou o he ea u es in ol ed in ag eemen .
6.5. Condi ions
Condi ions a e ac o s ha a e no ealized in he ag eemen p ocess, bu
which may de e mine how (o whe he ) ag eemen akes place. In sec ion
2.2.6, we saw how animacy and opicali y a ec ag eemen in a ious
languages, such ha ag eemen will only occu i an a gumen is [+anima e]
o a opic, o example. In he ligh o he p ope ies o ag eemen desc ibed in
he sign language li e a u e, especially he une en dis ibu ion o ag eemen
ac oss he e bs o a gi en language, we conside ed in sec ion 3.2.1.3 possible
condi ions and p e equisi es ha could explain he a es ed limi a ions on
ag eemen . This discussion was amed wi hin he cha ac e iza ion o
ag eemen on di ec ional ag eeing e bs, which display a speci ic se o
p ope ies: ag eemen occu s on (di) ansi i e e bs o ans e , and
a gumen s mus be [+human], [+anima e] o a possesso o some so .
Howe e , I ha e a gued ha ag eemen in LSE should no be es ic ed o his
ype o ag eeing e b and ha he e is a mo e gene al p ocess in ol ing an
associa ion be ween a e e en and a loca ion in signing space. This p ocess is
no limi ed o wo- o h ee-place p edica es and occu s wi h indi idual
a gumen s as single a gumen ag eemen (sec ions 3.2.3 and 5.2.3). In his
sec ion, I adop his b oade pe spec i e on ag eemen o examine he
condi ions ha ope a e on his spa ial ag eemen mechanism in LSE. This
discussion will lead o he issue o op ionali y in LSE ag eemen .
Animacy is a well a es ed condi ion on ag eemen in many spoken
languages, wi h less anima e a gumen s being less likely o igge ag eemen
(see sec ion 3.2.1.3 o examples). Va ious p oposals ha e sugges ed ha
ag eemen is es ic ed o anima e a gumen s in di e en sign languages ( o
ASL, Ma hu 2000: 212; o Ko ean Sign Language, Hong 2008: 170). Howe e ,
e en wi hin he domain o ( wo-place) ag eeing e bs, his condi ion does no
hold c oss-linguis ically: in sec ion 3.2.1.3 we saw example (14) om LSQ o
an ag eeing e b wi h inanima e a gumen s, and in sec ion 5.4.1, LSE example
7
Fu he mo e, he p oposal o classi ie handshapes as some so o class o gende
ag eemen (sec ion 6.4.1) also i s in o his gene aliza ion. I classi ie p o o ms a e conside ed
as ano he ype o p onominal e e ence, hen he gende ea u e is also p esen in he
p onominal sys em.
Condi ions 257
(23). Fu he mo e, expanding he no ion o ag eemen ma king o single
a gumen ag eemen makes his condi ion less enable: examples om ISL
(see (22) in sec ion 3.2.3) and LSE (see (10) in sec ion 5.2.3) demons a e ha
inanima e a gumen s such as ‘s ick’ o ‘exam’ may igge ag eemen .
Fu he mo e, he a gumen s o he so-called spa ial e bs a e inanima e.
Padden’s (1983/1988) ipa i e classi ica ion o e bs does no cha ac e ize he
spa ial modi ica ion o spa ial e bs as an ag eemen p ocess; I now eassess
his dis inc ion.
In sec ion 3.2.1.4, I ques ioned he dis inc ion be ween ag eeing and
spa ial e bs: a ious au ho s ha e al eady p oposed models ha do no
main ain such a dis inc ion ( o ASL, Janis 1992; o Lib as, Quad os 1999). I
will d aw upon wo sepa a e obse a ions o suppo he claim ha bo h
ag eeing and spa ial e bs mani es he same ag eemen p ocess. Quad os &
Que (2008) sugges ha ag eemen occu s in bo h ca ego ies o e b, which
may ag ee wi h ei he loca i e o pe sonal a gumen s.
8
This coincides wi h he
idea men ioned abo e in he sec ion on he con olle s o ag eemen (6.1) ha
sign languages ha e loca i e ag eemen (Thompson, Emmo ey & Kluende
2006), and loca i e a gumen s o m pa o ce ain e bs’ a gumen s uc u e.
The e o e, sign language e bs ag ee wi h hei a gumen s, some o which a e
loca i e, and his is achie ed by means o he same spa ial mechanism o
assigning he e e en o a loca ion in signing space.
This use o space o ag eemen should be dis inguished om he
exploi a ion o space o isomo phic mappings ypically employed in spa ial
desc ip ions. In he case o loca i e a gumen s, he o me ends o occu
wi hin he con ex o he la e , such ha he loca ions assigned o a gumen s
occu wi hin a spa ial map. By hei seman ic na u e, loca i e a gumen s a e
places, so hey occu in discou se con ex s in which loca ion is likely o be
ele an , and spa ial in o ma ion is he e o e ep esen ed (by means o
isomo phic mapping o he signing space). Howe e , he wo uses o space
a e independen , as is e iden om he ollowing wo cases ha highligh he
dis inc ion. Fi s ly, non-loca i e a gumen s may occu in isomo phic spa ial
mappings. This occu s in he case o p esen e e en s, in which he loca ions
assigned o e e en s co espond o each one’s eal-wo ld loca ion. Non-
p esen non-loca i e e e en s may also occu in a spa ial map, mo i a ing he
use o a highe loca ion in signing space in examples o he ype ‘I asked he
all man’ (c . Liddell 2000). Secondly, loca i e a gumen s may appea in
8
Quad os & Que main ain he p esence o a pe son ea u e, which I ha e ejec ed abo e in
sec ion 6.4.3. My analysis makes use o an al e na i e iden i y ea u e, bu he ele an poin
he e is ha a gumen s may be loca i e.
258 Typological pe spec i e o LSE ag eemen
con ex s in which ela i e loca ion is no ele an , and hus an isomo phic
mapping is no used. This dissocia ion is appa en in LSE due o he ac ha
isomo phic mappings o la ge physical dis ances end o occu in he e ical
plane: a signe desc ibing whe e she has s udied may place Bilbao, Ba celona
and London a di e en poin s on he e ical plane o show hei ela i e
posi ions, as i she had a map in on o he . Howe e , i he geog aphical
loca ion o each place is no ele an , each e e en may be associa ed wi h a
poin on he ho izon al plane. In his con ex , any e b ha ag ees wi h hose
loca ions (‘I mo ed om Ba celona o London’) is no suscep ible o he
(iconic and scala ) p ope ies o opog aphic space. Indeed, he p ope ies o
spa ial e bs ha Padden (1983/1988) desc ibed (and which seemed o se
hem apa om ag eeing e bs) a e in ac p ope ies o isomo phic,
opog aphic space in which spa ial e bs end o (bu do no necessa ily)
occu .
In summa y, since he spa ial ag eemen ma king in LSE ope a es in a
simila way o ag eeing e bs, single a gumen ag eemen and spa ial e bs,
he a gumen s a e no es ic ed o a speci ic seman ic p ope y, bu ake in
anima es, inanima es and loca i es. As such, he e does no appea o be a
condi ion ela ed o he seman ic ca ego y o he a gumen s o ag eemen .
Howe e , al hough a gumen s o di e en ypes may igge ag eemen , he
issue emains ha ag eemen in sign languages is highly op ional. As
desc ibed in sec ion 3.2.1.2, ag eemen ma king may be omi ed o one o he
a gumen s (ag eemen ma ke omission), o may be comple ely absen . In he
i s case, he subjec a gumen is no ma ked, while ma king o he objec
a gumen is p ese ed. As men ioned abo e in sec ion 6.3.1, his appea s o be
ela ed o a gene al endency o he objec a gumen o be mo e salien in
sign languages, and equi es u he in es iga ion o de e mine whe he his
could be a modali y d i en e ec . Ne e heless, al hough he subjec
a gumen is no ma ked by he e bal in lec ion, al e na i e mechanisms
could subs an ia e he ag eemen p ocess such as a ole shi ope a o , which
may be cha ac e ized as ano he ype o ag eemen (He mann & S einbach
2012). Al e na i ely, he lack o subjec ma king may be due o speci ic ypes
o e e ence: o example, null a gumen s (wi h null ag eemen ma king) may
be used o ma k impe sonal e e ence (Kimmelman 2015).
The comple e absence o ag eemen ma king on a e b ha migh
po en ially ag ee, on he o he hand, sugges s ha he ag eemen ela ion is
comple ely absen . Since ag eemen ma king is achie ed by he use o he
loca ions in signing space associa ed wi h he co esponding e e en s, his
issue is closely ela ed o whe he o no loca ion assignmen occu s. I
loca ions a e no assigned, hey a e no a ailable o he ag eemen
Canonici y 259
mechanism. Whe he o no a loca ion assignmen occu s seems o depend
upon discou se conside a ions o do wi h he p esence o o he e e en s and
whe he he p edica es gi e ise o he need o ack hose e e en s. Fo
example, a seman ically e e sible sen ence (‘Bea likes you’) may equi e
(spa ial) means o cla i y he seman ic/syn ac ic ole o each a gumen ,
whe eas his is no so o a non- e e sible p oposi ion (‘Bea likes ma hs’).
Equally, e en when a e b only ag ees wi h a single a gumen (as in he case
o single a gumen ag eemen ), he p esence o compe ing e e en s in he
discou se may o ce he use o space o dis inguish be ween hem. In his
sense, he use o spa ial e e ence (and hence o he ag eemen mechanism
ha depends upon i ) appea s “as equi ed”, pe haps along lines simila o
he use o explici p onouns in p o-d op languages. This “only i needed”
na u e o LSE ag eemen is an unusual p ope y, since ag eemen is gene ally
aken o be a basic, au oma ic mechanism in a language. The op ional na u e
o spa ial ag eemen will be e u ned o in sec ion 7.1.3 since i ep esen s an
impo an challenge o a syn ac ic accoun o ag eemen in LSE.
To conclude his sec ion on condi ions, ag eemen in LSE, unde s ood as
a basic mechanism ha uses loca ions in he signing space as ma ke s o
a gumen s, is no subjec o seman ic condi ions. In con as , he widesp ead
op ionali y o LSE ag eemen highligh s he ac ha discu si e and p agma ic
conside a ions de e mine he appea ance o spa ial ma king, and as such
ep esen a condi ion on ag eemen ha me i u he in es iga ion.
6.6. Canonici y
Ha ing desc ibed he wo kings o spa ial ag eemen in LSE in some de ail,
and ha ing assessed how he di e en aspec s (con olle s, a ge s, ea u es,
e c.) measu e up agains wha has been desc ibed o spoken languages, I now
u n o an e alua ion o how ag eemen -like his mechanism is. In o de o do
his, I use Co be ’s (2003b, 2006) no ion o canonici y. As desc ibed in sec ion
2.2.7, he no ion o canonici y is essen ially a means o de ining p o o ypical
ag eemen based on a se o gene al p inciples ha cha ac e ize he
phenomenon and on consensus (in he li e a u e and ac oss di e en
heo e ical amewo ks) abou wha coun s as ag eemen . Thus, he p ope ies
o p o o ypical ag eemen con o m o gene al cha ac e is ics o ag eemen
(e.g. “ag eemen is edundan a he han in o ma i e”) and a e hose ha a e
accep ed by mos linguis s as quali ying as a case o ag eemen (e.g. “ he
domain is local”). P o o ypical ag eemen is de ined in e ms o a se ies o
speci ic c i e ia (e.g. “ma king is bound”) and occupies he cen e o he space
o possible ag eemen -like phenomena. The emaining space is mapped by
260 Typological pe spec i e o LSE ag eemen
he less p o o ypical al e na i es o each c i e ion (e.g. “ he domain is non-
local”, “ma king is ee”).
C-1:
con olle is p esen
>
con olle is absen
C-2:
con olle has o e exp ession
o ag eemen ea u es
>
con olle has co e exp ession
o ag eemen ea u es
C-3:
consis en con olle (all a ge s
ake he same alue o a gi en
ea u e)
>
hyb id con olle ( a ge s ake
di e en alues o a gi en
ea u e)
C-4:
con olle ’s pa o speech is
i ele an
>
con olle ’s pa o speech is
ele an
C-5:
ma king is bound
>
ma king is ee
C-6:
ma king is obliga o y
>
ma king is op ional
C-7:
ma king is egula (a ixal)
>
ma king is supple i e
C-8:
ma king is alli e a i e (ma ke
on all a ge s is he same and
iden ical o o man on
con olle )
>
ma king is opaque (ma ke
changes om a ge o a ge and
is no iden ical o o man on
con olle )
C-9:
ma king is p oduc i e (applies
o all membe s o a ca ego y)
>
ma king is spo adic (only appea s
on some membe s o a ca ego y)
C-10:
a ge always ag ees
>
a ge ag ees only when
con olle is absen loca ion
assignmen occu s
C-11:
a ge ag ees wi h single
con olle
>
a ge ag ees wi h mo e han one
con olle
C-12:
a ge has no choice o con olle
>
a ge has choice o con olle
C-13:
a ge ’s pa o speech is
i ele an
>
a ge ’s pa o speech is ele an
C-14:
domain is asymme ic
>
domain is symme ic
C-15:
domain is local
>
domain is non-local
C-16:
domain is one o a se
>
single domain
C-17:
ea u e is lexical
>
ea u e is non-lexical
C-18:
ea u es ha e ma ching alues
>
ea u e alues do no ma ch
C-19:
no choice o ea u e alue
>
choice o ea u e alue
C-20
no condi ions
>
condi ions
Table 6.2 C i e ia o canonical ag eemen . The symbol > means “is mo e canonical
han”. Adap ed om Co be (2006: 10-27). Fo each c i e ion, he beha iou o
spa ial ag eemen ma king in LSE is indica ed in bold ace and shaded backg ound.
The o iginal o mula ion o C-10 is adap ed o he case o LSE and he change
indica ed by s ike h ough and i alics (see ex o de ails).
Canonici y 267
The inal c i e ion o canonici y ( able 6.2: C-20) e e s o condi ions:
canonical ag eemen has no condi ions. Once a a ge , con olle , domain and
ea u es ha e been de ined, his should be enough o ully speci y he
ag eemen ela ion. Howe e , o LSE, we ha e seen ha he p esence and
exp ession o ag eemen is subjec o a ious condi ions: no ably, discu si e
and p agma ic conside a ions de e mine whe he space will be used o
e e ence acking, and also he physical p esence o isibili y o he e e en s
in he communica i e con ex . As such, ag eemen in LSE is no canonical in
his espec .
In summa y, his assessmen o spa ial ag eemen in LSE shows ha on
he basis o mos o he c i e ia he p ocess is mo e canonical han no : 15 ou
o 20. I is impo an o bea in mind ha mos spoken languages also p esen
a ying numbe s o non-canonical p ope ies. As a compa a i e, a p ecu so y
e alua ion o e bal ag eemen in spoken Spanish sugges s ha he p ocess is
canonical acco ding o 16 o he c i e ia, bu has less canonical beha iou o
ou c i e ia (C-1, C-2, C-8 and C-18). Ne e heless, hese c i e ia se e as
guidelines, and do no u nish us wi h an index o canonici y: as Mea loa so
ap ly poin s ou , “ wo ou o h ee ain’ bad” as a as sco ekeeping goes, bu
mo e impo an is he na u e o he measu es being applied. To a oid ying
o make sense o a s a emen o he ype “LSE ag eemen is 75% canonical”, I
u n o he gene al p inciples ha unde lie he c i e ia o canonici y o
p o ide a b oade pe spec i e (Co be 2006).
6.6.2. Applying Co be ’s gene al p inciples o spa ial ag eemen in LSE
Th ee gene al p inciples unde lie he c i e ia o canonici y. These p inciples
conce n edundancy, syn ac ic simplici y and he mo phological exp ession o
ag eemen (Co be 2003b, 2006). Each c i e ion is mo i a ed by one (o mo e)
o hese p inciples, and hese dependencies a e shown in able 6.3. The able
also indica es whe he spa ial ag eemen in LSE is canonical o a gi en
c i e ion by means o a ick o c oss.
The i s gene al p inciple s a es ha canonical ag eemen is edundan
a he han in o ma i e. This p inciple g oups oge he c i e ia 1, 2, 10, 17, 18
and 19 (Co be 2006: 27). As can be seen in able 6.2, spa ial ag eemen in LSE
mee s jus hal o hese c i e ia, which indica es ha he sys em displays a
ela i ely low deg ee o edundancy. This lack o edundancy s ems om wo
aspec s o he e e en ial sys em ha he ag eemen mechanism exploi s:
i s ly, he ma king is s ongly indexical ( a he han lexical) in na u e (c . C-
17) and so in some sense he ma ke is adding a he han jus epea ing
some hing (which would be mo e edundan ).
268 Typological pe spec i e o LSE ag eemen
I
Canonical ag eemen is
edundan a he han
in o ma i e.
II
Canonical ag eemen is
syn ac ically simple.
III
The close he exp ession
o ag eemen is o
canonical in lec ional
mo phology, he mo e
canonical i is as
ag eemen .
C-1
C-2
C-3
C-4
C-5
C-6
C-7
C-8
C-9
C-10
C-11
C-12
C-13
C-14
C-15
C-16
C-17
C-18
C-19
C-20
Table 6.3. The gene al p inciples o canonical ag eemen , and he c i e ia associa ed
o each p inciple. The icks and c osses indica e whe he o no spa ial ag eemen in
LSE is canonical acco ding o a gi en c i e ion. The ligh e shading o C-6 o gene al
p inciple II indica es ha his c i e ion is only pa ially associa ed (see ex o
de ails).
The second p ope y o he e e en ial sys em ha educes edundancy is he
“only i needed” quali y men ioned in sec ion 6.5 (and e lec ed in C-10). I
spa ial e e ence we e obliga o y whene e a e e en is in oduced, his
would c ea e a much mo e edundan sys em. Howe e , LSE does no always
ancho e e ence o space (and hence use spa ial ag eemen ) bu only
whene e he need a ises.
The second gene al p inciple cha ac e izes canonical ag eemen as
syn ac ically simple, and akes in c i e ia 3, 4, 10, 11, 12, 13, 14, 15, 16, 18, 19
and 20 (and pa ially 6) (Co be 2006: 27). Acco ding o his p inciple,
Canonici y 269
ag eemen ha can be desc ibed by s aigh o wa d ules is mo e canonical.
In his espec , spa ial ag eemen in LSE is b oadly canonical (especially om
he poin o iew o he consis ency o con olle s, a ge s, hei ma king and
ea u es), bu ails on wo ela ed coun s. The op ionali y o he mechanism
means ha ag eemen does no consis en ly occu (c . C-6 and C-10) and his
complica ing ac o is (a leas pa ially) exp essed in e ms o condi ions on
he ag eemen p ocess (c . C-20).
The hi d gene al p inciple ela es o how ag eemen is ma ked: he
close he exp ession o ag eemen is o canonical (i.e. a ixal) in lec ional
mo phology, he mo e canonical he ag eemen mechanism. This mo i a es
c i e ia 5, 6, 7, 8 and 9 (Co be 2009: 27). As a as he ma king o spa ial
ag eemen in LSE is conce ned, i beha es like canonical in lec ional
mo phology wi h he peculia i y ha he ma king is a s em al e na ion. S ill,
his idiosync asy may e en be conside ed o be canonical in he con ex o he
simul aneous empla ic mo phology o signed languages (Sandle & Lillo-
Ma in 2006). Howe e , once mo e, he op ional na u e o he p ocess makes
spa ial ag eemen in LSE less canonical: he ac ha ma king may be omi ed
makes i less like (canonical) in lec ional mo phology.
In ligh o hese gene al p inciples, he anomalous o non-canonical
p ope ies o spa ial ag eemen in LSE can be seen mo e clea ly. Wha
becomes appa en is ha he ac o s ha pull he ag eemen sys em away
om canonici y de i e om wo gene al a eas: on he one hand, he ela ed
issues o edundancy and op ionali y, and, on he o he , he na u e o he
e e en ial sys em exploi ed by ag eemen in LSE.
The op ional cha ac e o ag eemen in LSE makes he sys em less
edundan since i o en will no appea i i is no equi ed. This aises he
ques ion o wha de e mines he op ionali y o he sys em: when can o mus
ag eemen appea , o no ? I ha e al eady indica ed ha p agma ic and
discou se conside a ions play a ole in his espec , bu u he s udy is
equi ed o es ablish a mo e p ecise cha ac e iza ion o he ac o s ha ha e
an e ec on he appea ance o ag eemen in LSE. To his end, a co pus-based
s udy could p o ide aluable e idence o iden i y wha causes ag eemen o
occu o no . A s udy o an Auslan co pus es ablished ha e bs modi y o
indica e hei a gumen s ela i ely in equen ly (less han hal o he okens)
(de Beuze ille, Johns on & Schemb i 2009). The au ho s sugges ha he
a ailabili y o o he s a egies (such as cons i uen o de , in o ma ion
s uc u e and cons uc ed ac ion) may in luence whe he o no spa ial
ag eemen occu s bu i emains o be seen i he choice o one s a egy o
ano he is sys ema ically condi ioned by discou se, p agma ic o syn ac ic
ac o s.
270 Typological pe spec i e o LSE ag eemen
Secondly, he ac ha he ma king is spa ial, and based on a spa ial
e e en ial sys em, comes o bea on he na u e o ag eemen in LSE. The
ag eemen p ocess in LSE in ol es adding a spa ial label o he con olle o
an ag eemen ela ion, a he han exploi ing an exis ing lexical ea u e (such
as gende ). This means ha bo h con olle and a ge a e ma ked wi h he
ea u e, possibly weakening he asymme ic di ec ionali y o he ela ion.
Ne e heless, he use o spa ial e e ence also has canonical aspec s. Co be
(2006: 24) poin s ou ha ea u es based on o mal assignmen a e mo e
canonical han hose o which assignmen is seman ically based. Al hough in
LSE he seman ic and o mal alues o a locus end o coincide, he impo an
obse a ion is ha he locus need no be assigned based on seman ics. This
was men ioned in he discussion o he dis inc ion be ween spa ial and
ag eeing e bs (in sec ion 6.5); he poin is ha spa ial loci may be used
exclusi ely o e e ence wi h no seman ic spa ial meaning. Thus, al hough
he use o loci can be (non-canonically) seman ic, i is possible o he loci o
se e a me ely e e en ial unc ion.
This exe cise in assessing spa ial ma king in LSE in e ms o canonici y
does no p o ide a de ini e classi ica ion o he phenomenon as “ag eemen ”
o “no ag eemen ”. Al hough i has been claimed ha he issue o
(non-)canonici y has been o e s a ed (Que 2011: 196) and i is ue ha he
ag eemen sys ems o many languages show non-canonical p ope ies ( ecall
ha all p o-d op languages ail C-1), I ag ee wi h Co be (2006: 27) ha “i is
mo e impo an o unde s and ag eemen and i s ela ed phenomena han o
d aw a p ecise line” be ween ag eemen and o he phenomena, and
conside a ions o canonici y p o ide a use ul pe spec i e o obse ing and
analysing he ma e .
6.6.3. O he e alua ions o he canonici y o sign language ag eemen
Be o e closing his sec ion on canonici y, I look a he wo o he a emp s o
apply his concep o ag eemen in a sign language, bo h o which ocus on
ASL (al hough he i s pape also p esen s da a om o he sign languages).
Ma hu & Ra hmann (2010) p o ide a somewha cu so y e alua ion o
sign language ag eemen based on Co be ’s c i e ia o canonici y. Thei
concep ion o ag eemen in sign language is limi ed o di ec ional ag eeing
e bs.
9
They e alua e all o Co be ’s c i e ia (wi h he excep ion o C-20,
which hey do no men ion), and mos o hei e dic s coincide wi h mine, so
9
Ma hu & Ra hmann (2010) dis inguish be ween double ag eemen (i.e. wi h subjec and
objec ) and single ag eemen (i.e. wi h he objec ). No e ha he la e is no wha I ha e called
single a gumen ag eemen bu is a di ec ional ag eeing e b wi h ag eemen ma ke
omission.
Canonici y 271
I will ocus on he poin s o di e gence. Conce ning a ge s, he e a e wo
di e ences, bo h o which conce n he issue o op ionali y: o C-6, Ma hu &
Ra hmann conside ha he ma king is obliga o y, whe eas I ha e claimed
ha i is op ional gi en he phenomenon o ag eemen ma ke omission; and
o C-10, hey main ain ha he a ge always ag ees, whe eas I poin ed ou
ha his only occu s when spa ial loci a e assigned. Un o una ely, hey do
no p o ide much explana ion o cla i ica ion o hei decisions, so i is no
possible o clea up his di e ence. The o he disc epancy comes om he
c i e ia o ea u es: o C-17, hey judge ea u es o ag eemen o be lexical;
o C-18, hey hold ha he ea u e alues do no ma ch. Thei unde lying
analysis o ag eemen (as in ol ing he ea u es o pe son and numbe ) is
wha leads hem o di e in hese c i e ia. Addi ionally, while hey conside
ole shi o be a eason o misma ches be ween ea u es (e.g. i s pe son
e bal o ms o non- i s pe son con olle s), I conside ed ha such cases
cons i u ed a di e en domain ha mo i a es a di e en exp ession o a
ea u e alue. In summa y, despi e mino di e ences due o di e en
cha ac e iza ions o he ag eemen p ocess (and i s scope), Ma hu &
Ra hmann’s assessmen o he canonici y o sign language ag eemen
coincides o a g ea ex en wi h ha o he LSE da a p esen ed he e.
In hei discussion o he linguis ic s a us o sign language ag eemen ,
Lillo-Ma in & Meie (2011) dedica e a sec ion o he non-canonical p ope ies
o di ec ionali y as ag eemen . Howe e , hey do no ame he discussion in
e ms o Co be ’s c i e ia, bu a he in e ms o he issues discussed in he
sign language li e a u e, namely, e b classi ica ion and he p ominence o
objec (o e subjec ) ma king. This leads o an in e es ing analysis o
ag eemen in sign language (and highligh s opics ha ha e c opped up in he
discussion o he LSE da a he e, no ably, op ionali y) bu does no p o ide an
e alua ion om wi hin an independen amewo k ha can p o ide a
sys ema ic means o gauging how and whe e a gi en ag eemen sys em
“misbeha es”. Addi ionally, Lillo-Ma in & Meie ’s discussion is limi ed o
di ec ional ag eeing e bs and does no con empla e spa ial ma king o o he
a ge s, such as single a gumen ag eemen .
6.6.4. Summa y
Spa ial ag eemen in LSE shows many p ope ies o canonical ag eemen as
de ined by Co be (2003b, 2006), and his b oadly coincides wi h he only
o he assessmen ha has sys ema ically e alua ed each o he c i e ia o
ag eemen in sign language based on da a om o he languages (Ma hu &
Ra hmann 2010). Fu he mo e, assessing he p ope ies o ag eemen in e ms
o wide eaching gene al p inciples has p o ided a way o analysing spa ial
272 Typological pe spec i e o LSE ag eemen
ag eemen in LSE o highligh i s peculia i ies as an ag eemen p ocess. Whe e
he phenomenon de ia es om canonical beha iou , his is in la ge pa due
o wo ac o s: he op ionali y o he p ocess and he spa ial e e ence
mechanism ha i makes use o .
6.7. Discussion and conclusions
This chap e has analysed he LSE da a on spa ial ag eemen om he poin o
iew o he ypological app oach se ou in sec ion 2.2. This examina ion has
shown ha ag eemen in LSE, unde s ood as he use o space o ma k
a gumen s on e bs and o he lexical ca ego ies (such as adjec i es,
de e mine s, and so on), alls wi hin he bounda ies o he phenomenon o
ag eemen as desc ibed o spoken languages. Impo an ly, he phenomenon I
ha e assessed is no limi ed o a subse o e bs ha show subjec and objec
ag eemen (i.e. ag eeing e bs) bu akes a wide iew o ag eemen as a basic
p ocess a use in he language.
The i s pa o his assessmen in ol ed e iewing he di e en
elemen s ha play a ole in ag eemen : con olle s, a ge s (including means
o exponence), domains, ea u es and condi ions (sec ions 6.1-6.5). In each
case, spa ial ag eemen in LSE is compa able o he indings o spoken
languages (al hough some unusual cha ac e is ics, o be discussed below, also
u ned up). This con o mi y was also con i med by he second phase o he
analysis, which applied Co be ’s (2003b, 2006) no ion o canonici y o he LSE
da a o see how ag eemen -like his mechanism is (sec ion 6.6), bo h om he
poin o iew o he indi idual c i e ia (sec ion 6.6.1) and he gene al
p inciples ha unde lie he no ion o canonical ag eemen (6.6.2). On bo h
coun s, he spa ial ag eemen mechanism in LSE a es well, and quali ies as
mo e ag eemen -like han no when iewed h ough he lens o canonici y.
As poin ed ou du ing his analysis, i is impo an o bea in mind ha
he end esul o his analysis is no he mos in e es ing inding (no , in he
inal eading, can i p o ide us wi h a conclusi e ca ego iza ion o spa ial
ag eemen as an ins ance o ag eemen ). Ra he , he p ocess o analysing he
da a in his way has led o a e inemen in how we concep ualize his spa ial
mechanism in LSE and has also b ough o ligh hose p ope ies ha a e
unusual o anomalous. In e ms o e inemen s, his chap e includes se e al
analyses o he LSE da a ha p o ide a clea e , mo e mo i a ed desc ip ion o
he gene al phenomenon.
When looking a he means o exponence o his ag eemen mechanism
(sec ion 6.2.3), I assessed p e ious claims ha he spa ial ma ke is some so
o p onominal a ix and applied a ious c i e ia ela ed o he p ope ies o
Discussion and conclusions 273
ag eemen ma ke s and p onominal a ixes in o de o ca ego ize spa ial
in lec ion in LSE as one o he o he . This e alua ion poin ed in he di ec ion
o ag eemen ma king, and also helped o s eng hen he case o ea ing
single a gumen ag eemen as a alid mani es a ion o his ag eemen
mechanism. I also highligh ed he di e ence be ween a p onoun and a poin
in space: when he wo a e con ounded, wha is essen ially a phonological
ea u e may e oneously be conside ed a cli ic. I p esen ed a cha ac e iza ion
o his spa ial ma king in LSE as a s em al e na ion, a mechanism also a es ed
o spoken languages.
The discussion o he numbe ea u e b ough up he dis inc ion
be ween e bal and nominal numbe (sec ion 6.4.2). Gi en ha LSE has
a ious mechanisms o ma king plu ali y on he e b, I applied a se ies o
diagnos ics o asce ain whe he his in lec ion was a e lex o he nume osi y
o he e b’s a gumen s, o o e bal numbe . The esul s con i med he ini ial
in ui ion ha he eduplica i e p ocess p esen in he dis ibu i e o m ma ks
e bal numbe , while a gumen numbe is ma ked h ough he use o space
(whe he ha be he a c mo emen o he mul iple ma ke , o he wo dis inc
loca ions in he dual ma ke ).
The inal e inemen came abou in he con ex o he condi ions ha
ope a e on his ag eemen p ocess (sec ion 6.5). When analysing he common
claim ha ag eemen in sign language is es ic ed o human o anima e
a gumen s, I poin ed ou ha conside ing spa ial ma king as he ag eemen
mechanism emo es his cons ain since single a gumen ag eemen shows a
much wide a ie y o a gumen s (and he da a show ha e en some ag eeing
e bs a e no subjec o his seman ic es ic ion). The p oposed seman ic
es ic ion may be ele an o (some) ag eeing e bs, bu no o he
ag eemen mechanism as a whole. Fu he mo e, he inspec ion o he
seman ic p ope ies o e bal a gumen s led o a eassessmen o he
di e ence be ween ag eeing and spa ial e bs: I p oposed ha spa ial e bs
ha e loca i e a gumen s bu he ag eemen mechanism is he same o bo h
ypes o e b (con a Padden 1983/1988). Wha can make a undamen al
di e ence in he use o space is isomo phic spa ial mapping, which in ol es
s ong iconic mo i a ion. The ac ha spa ial e bs (wi h hei loca i e
a gumen s) equen ly exploi isomo phic mappings con ounds his
dis inc ion. Howe e , I p o ided examples ha demons a e ha loca i e
a gumen s do no necessa ily imply an isomo phic use o space. Thus, space
may be used in sign languages in di e en ways, bu he mechanism o
in e es he e – he ma king o a gumen s ia loca ions – has many o he
hallma ks o ag eemen .
274 Typological pe spec i e o LSE ag eemen
Al hough his p ocess o spa ial ma king is a s ong candida e o
ag eemen based on his assessmen in e ms o c oss-linguis ic da a om
spoken languages, he analysis also poin ed ou a eas whe e LSE is doing
some hing ou o he o dina y as a as ag eemen is conce ned. The mos
s iking – and ypologically ex ao dina y – aspec o spa ial ag eemen is he
ac ha i does no make use o a pe son ea u e. Al hough a pe son
dis inc ion has been a gued o in o he sign languages, he da a do no
suppo upholding his ea u e o LSE. This makes LSE ex emely unusual as
a language. I a gue ha he lack o he pe son ea u e s ems om he
e e en ial sys em employed by he language. In u n, his e e en ial sys em
is based on he use o space and he lack o pe son ma king is a consequence
o he possibili ies a o ded by a spa ial e e ence mechanism. As an
al e na i e o he pe son ea u e, I ha e p oposed an iden i y ea u e, based
on he no ion o R-locus, and his p oposal will be de eloped in he nex
chap e .
Ano he anomaly o spa ial ag eemen in LSE ha canno be a oided is
i s pe asi e op ionali y. This issue has been men ioned a many poin s in his
hesis (and will con inue o ea i s head), and is one o he main causes o he
non-canonical ai s o ag eemen in LSE. This aspec o sign language
ag eemen also seems o be he cause o much scep icism as o i s s a us as
ag eemen , especially in he spoken language li e a u e, and I will men ion
wo di e en e e ences in his espec . Fi s ly, in his comp ehensi e s udy o
ag eemen , Co be (2006: 264), on whose wo k much o his hesis has so
hea ily lean , men ions sign languages in a oo no e and s a es ha he
p ocess desc ibed as ag eemen o hese languages does no seem o ha e he
“sys ema ic co a iance” o be conside ed ag eemen . E en hough I could y
o claim ha Co be ’s dismissal o sign language ag eemen is no ele an o
my analysis since I conside spa ial ag eemen pe se and no jus ag eeing
e bs, he unde lying issue ha his commen alludes o is no diminished:
op ionali y. The second e e ence is Cysouw’s (2011) eply o Lillo-Ma in &
Meie ’s (2011) landma k pape on he linguis ic s a us o sign language
ag eemen . Cysouw s a es ha Lillo-Ma in & Meie con incingly show ha
poin ing and di ec ionali y in sign languages is compa able o pe son
ma king in spoken languages, bu ha his does no allow hem o ake he
u he s ep o claiming ha his p ocess is ag eemen . Clea ly his depends
on he de ini ion o ag eemen , which Cysouw goes o g ea leng hs o ace
h ough i s his o ical e olu ion, and he unde lying p oblem appea s o be he
ac ha his pe son ma king mechanism does no ep esen a “sys ema ic
co a iance” o linguis ic exp essions. Again, I could a gue ha my concep ion
o ag eemen as a spa ial ma king mechanism is much mo e sys ema ic in ha
Discussion and conclusions 275
i applies o a wide a ie y o e bal ag eemen in addi ion o ag eemen in
he nominal domain. Howe e , I am s ill le holding he baby as a as
op ionali y is conce ned.
In he sec ion on condi ions (6.5) I wen some way o add essing he
issue o op ionali y by sugges ing ha i is due o discu si e and p agma ic
condi ions on ag eemen : only when he igh ci cums ances hold will
ag eemen ake place. This s ill lea es unanswe ed he ques ion as o how he
ag eemen sys em would handle his endemic op ionali y: he o mal accoun
de eloped in he nex chap e a emp s o explain his in e ms o de aul
alues.
Be o e closing his chap e on he commensu abili y o spa ial
ag eemen in LSE and ag eemen in spoken languages, we can now add ess
he second o he esea ch ques ions om chap e 1: A e he spa ial mechanisms
employed by LSE compa able o he ag eemen mechanisms in spoken languages?
This chap e has p o ided a de ailed analysis o he a ious spa ial
mechanisms ha I ea ma ked as being likely candida es o ag eemen in LSE
and desc ibed in chap e 5. This app aisal indica es ha spa ial ma king in
LSE shows s ong pa allels wi h ag eemen in spoken languages and ce ainly
appea s o all wi hin he limi s o he phenomenon as mani es ac oss spoken
languages. The assessmen o he canonici y o LSE spa ial ag eemen also
yielded a a ou able esul , sugges ing ha his mechanism is a ela i ely
canonical ins ance o ag eemen . As Ma hu & Ra hmann (2010: 196) poin ou
in hei e alua ion o canonici y o sign language ag eemen , “o he
app oaches o e bal ag eemen in signed languages may in e p e Co be ’s
c i e ia o canonical ag eemen di e en ly” and i is wo h bea ing in mind
ha his diagnos ic ool depends g ea ly on he p io concep ualiza ion o he
p ocess unde examina ion. Indeed, some disc epancies be ween Ma hu &
Ra hmann’s canonici y analysis and my own can be aced back o hei
es ic ion o ag eemen o ag eeing e bs and my inclusion o ela ed spa ial
phenomena.
In sum, spa ial ag eemen in LSE, unde s ood as he spa ial ma king o
one elemen o ma k co a iance wi h ano he , shows enough simili ude o he
ypological concep o ag eemen used o spoken languages, o bo h o be
ea ed as mani es a ions o he same linguis ic p ocess. Conside ing hese
phenomena om di e en modali ies as he same is bo h meaning ul and
use ul o a i ing a a be e unde s anding o how and why a language
c ea es ela ions be ween i s elemen s. (This is no o say ha he e a e no
di e ences, and his sec ion has also b ough hese o he eade ’s a en ion.)
The indings o his chap e also make i possible o build upon he
p elimina y answe o he i s esea ch ques ion o e ed a he end o he
276 Typological pe spec i e o LSE ag eemen
p e ious chap e . The ine uning o he cha ac e iza ion o spa ial ag eemen
in LSE om he analysis in his chap e pu s us in a be e posi ion o compa e
he LSE da a wi h wha has been desc ibed o o he sign languages. In many
ways, he cha ac e iza ion o LSE ag eemen I ha e de eloped has di e ged
om he si ua ion o o he sign languages as laid ou in chap e 3. Spa ial
ag eemen in LSE akes in a b oad ange o phenomena, well beyond he
domain o ag eeing e bs, and does no in ol e a pe son ea u e, which is no
wha we saw o o he sign languages. Some o hese di e ences may be due
o genuine in a-modal c oss-linguis ic a ia ion: he di e ences in he da a
o p onominal o ms be ween LSE and ASL, o ins ance, sugges ha some
sign languages may make use o a pe son ea u e while o he s do no .
Howe e , o he di e ences may ha e mo e o do wi h he e olu ion o he
heo y o ag eemen ha I ha e de eloped o LSE: only a eanalysis o he
da a o o he sign languages will e eal whe he his model can be applied o
explaining spa ial ag eemen in hose languages as well. In he in e im, he
da a om LSE and o he sign languages o e enough simila i ies o sugges
ha a simila mechanism is a wo k and enough di e ences o make
compa a i e s udies wo hwhile.
The nex chap e con inues o analyse he da a o his s udy, and
p o ides a o mal accoun o spa ial ag eemen in LSE.
Loca ion, iden i y and loca ing iden i y 283
dedica ed Num p ojec ion has been p oposed o numbe and an
independen ly mo i a ed nominalizing n p ojec ion o gende /wo d class. I
will deal wi h each in u n.
(2)
The Num p ojec ion is a midle el unc ional ca ego y analogous o T in he
e bal domain (Ri e 1991; Ca s ens 2001) and p o ides he s uc u al
posi ion o he numbe ea u e. Ri e (1993) u he claims ha o Romance
languages he gende ea u e is also loca ed a Num, and no on N, he noun
head, as is he case o Heb ew. This is based on obse a ion o he
in lec ional mo phemes in hese languages, and i is hus an empi ical
ques ion o a gi en language (o language amily) whe he gende appea s
on Num o no . I he iden i y ea u e occupies Num, his means ha i is in
he same unc ional head as he numbe ea u e (in much he same way ha
numbe and gende a e in Romance languages acco ding o Ri e ’s p oposal).
This could p o ide p edic ions abou he posi ion o numbe and iden i y
ela i e o o he elemen s ha could be es ed empi ically. Howe e , he ac
ha much mo phology in sign languages is simul aneous a he han
sequen ial makes i di icul o pinpoin he ela i e posi ion o hese ea u es
in he s uc u e based on su ace o m. Fu he mo e, bo h he numbe and he
iden i y ea u e may be mani es spa ially (and hus a he same ime),
obscu ing e en mo e how each ea u e con ibu es o he inal o m. The LSE
284 Fo mal analysis o LSE ag eemen
da a in his s udy do no p o ide any clea e idence ei he o o agains he
pe son and iden i y ea u es occupying he same s uc u al posi ion.
5
Al e na i ely, he hos o he iden i y ea u e could be n. This
p ojec ion immedia ely domina es he NP and is e ec i ely wha makes he
NP unc ion as a noun in he syn ac ic s uc u e. As we saw abo e, i has been
claimed ha he gende (o , mo e p ecisely, wo d class) ea u e is loca ed in
his p ojec ion (Kihm 2005). Gi en ha he iden i y ea u e is assigned o he
nominal, i is possible ha i occupies he head o a nea by (i.e. immedia ely
domina ing) p ojec ion in o de o he iden i y ea u e o a ix o he NP, hus
o ming a “syn ac ically comple e” nominal, ha is, an nP. Howe e , wo
conside a ions poin away om his possibili y. Fi s ly, wo d class/gende
p o ides a means o ca ego izing lexical i ems acco ding o some unde lying
seman ic classi ica ion, and his is no wha he iden i y ea u e appea s o do.
The second conside a ion conce ns he no ion o e e en ial index and is also
ele an when e alua ing N i sel as a po en ial hos o he iden i y ea u e.
So we will now examine his possibili y.
As men ioned abo e, i seems unlikely ha he iden i y ea u e is pa o
he lexical en y o a gi en noun since he same noun may appea wi h
di e en iden i y alues (i.e. e e ing o di e en en i ies). Ne e heless,
Bake (2003, 2008) de ines nouns as lexical ca ego ies ha ha e a e e en ial
index, and i is his e e en ial index ha he iden i y ea u e appea s o pick
up on. Indeed, Bake specula es ha bea ing a e e en ial index depends on
mo e undamen al unde lying c i e ia o iden i y (2008: 31-33). This link
be ween iden i y and he (modali y-independen ) e e en ial index is an
insigh ul connec ion and p o ides suppo o my p oposal o an iden i y
ea u e, bu I ques ion he associa ion be ween he noun i sel and he
e e en ial index. Bake men ions he p inciple ha an XP mus ha e a
e e en ial index in o de o ha e in insic ϕ- ea u es ( alues o pe son,
numbe and gende ). Fo Bake , his XP would be an NP; o a p oponen o
an nP accoun ske ched in he p e ious pa ag aph, i would be an nP.
Ne e heless, as we saw in sec ion 7.1.1, he ϕ- ea u es associa ed wi h a
nominal a e dis ibu ed h oughou s uc u e ha goes beyond he NP (and
he nP). As such, i is he en i e DP ha con ains he ϕ- ea u es and so, in
keeping wi h his p inciple, i mus be he DP ha has he e e en ial index.
Independen suppo o his idea comes om wo k on de e mine s ha
assumes ha he de e mine licenses he appea ance o a noun as an
a gumen (Longoba di 1994). E idence ha DPs can be a gumen s whe eas
5
Examining how classi ie cons uc ions in e ac wi h numbe and he use o loca ion may
p o ide a use ul means o p obing his issue and I lea e he ma e o u u e esea ch.
Loca ion, iden i y and loca ing iden i y 285
NPs canno comes om con as s such as (3) in Spanish. In (3a), a
de e mine less NP (he mano de Amaia) may occupy a p edica e posi ion,
whe eas in (3b) a ull DP (wi h he de ini e a icle el) is equi ed in an
a gumen posi ion and he sen ence is ung amma ical wi hou he a icle.
Fu he mo e, suppo om LSE i sel comes om he ac ha whole
sen ences can be localized and hus ope a e as a gumen s. (This link be ween
localiza ion and nominaliza ion will be de eloped in sec ion 7.2.3.)
Spanish
(3) a.
I zal
es
he mano
de
Amaia
‘I zal is Amaia’s b o he .’
b.
A
I zal
le
gus a
*(el)
he mano
de
Amaia
‘I zal likes Amaia’s b o he .’
Since he ull DP ca ies he e e en ial index (and he ϕ- ea u es), he e is no
need o he iden i y ϕ- ea u e o be es ic ed o he N (o n) posi ion.
Fu he mo e, he e y ac ha he ull DP is associa ed wi h he e e en ial
index makes D a much s onge candida e as he hos o he iden i y ea u e
han N (o n) is.
The connec ion be ween loca ion and he D head had been ouched
upon in p e ious wo k. In he wo k on LIS, Be one (2007) p oposes ha
space ea u es a e hos ed on D. Building on his analysis B unelli (2011) places
loca ion assigning poin s also in he head o he DP.
6
We saw abo e ha he
pe son ea u e is si ua ed on D o spoken languages: since he e I claim ha
LSE has an iden i y ea u e and no pe son ea u e, a pa simonious solu ion
would be o hese complemen a y ea u es o occupy he same posi ion.
Indeed, he e is a lo in common be ween he ea u es o pe son and iden i y,
since bo h ake alues ha depend upon he con ex o u e ance. The pe son
ea u e does his by di iding e e en s acco ding o hei ole as pa icipan s
in he discou se (speake , add essee, e c.) while he iden i y ea u e
dis inguishes e e en s ha appea in he discou se.
One possible p oblem wi h his p oposal is he op ionali y o he use o
loca ion. Abne (2012) poin s ou ha a non-obliga o y de e mine ha may o
may no appea lies in he ace o he idea ha languages canno exhibi ee
a ia ion be ween he p esence o absence o a de e mine o nominal
a gumen s. Thus, i a language has a lexical de e mine wi h a ce ain
meaning, i mus use ha de e mine o exp ess ha meaning (C isma 1997,
6
The idea ha loca ion ea u es a e hos ed on D is exploi ed by P au & S einbach (2013) in
hei analysis o he g amma icaliza ion o he sign PERSON in DGS.
286 Fo mal analysis o LSE ag eemen
ci ed in Longoba di 2001: 584). Howe e , he claim he e is no o a speci ic
lexical de e mine in D (along he lines o MacLaughlin’s (1997) analysis o
p enominal poin s in ASL, which is wha Abne was a guing agains ) bu
a he o he p esence o a ea u e ha en e s he syn ac ic s uc u e a a
speci ic place.
7
Fu he mo e, he op ionali y o a poin is a sepa a e issue om
he op ionali y o he iden i y ea u e: his is demons a ed by he possibili y
o localiza ion, in which an iden i y ea u e is p esen and exp essed, bu a
poin may (o may no ) be absen .
8
The appa en op ionali y o localiza ion
doub less complica es he issue bu i is no speci ically p oblema ic o he
iden i y ea u e being hos ed on D.
7.1.3. Op ionali y o he use o space
In he abo e discussion o he iden i y ϕ- ea u e, he issue o op ionali y
c opped up se e al imes. The ac ha loca ion is no always used aises an
impo an ques ion: is he unde lying iden i y ea u e op ional, o ,
al e na i ely, is he ea u e p esen bu (some imes) phonologically null?
Gi en ha he iden i y ea u e e lec s a undamen al unde lying concep , i
seems mo e likely ha i is p esen bu may gi e ise o a phonologically null
ealiza ion. O he ϕ- ea u es show simila beha iou : in sec ion 2.2.5.2, we
saw how numbe may ha e a neu al alue such ha he co esponding o m
is ambi alen wi h espec o nume osi y, and o en his neu al alue is
exp essed by means o a de aul o m. Thus, in Tu kish, he de aul “singula ”
o m e may mean ‘house’ o ‘houses’. By he same oken, he iden i y ea u e
may also ake a neu al alue ha is ou side he e e ence sys em. This does
no mean ha he seman ics o e e ence b eak down. Jus as a nume ically
agnos ic o m does no unde mine he exp ession o plu ali y, a neu al alue
o he iden i y ea u e means no mo e han ha he language does no ha e
o assign a speci ic alue in ce ain con ex s. Thus, much o he op ionali y o
ag eemen in LSE (and sign language in gene al) appea s o s em om he
ma e o whe he he iden i y ea u e has a speci ic (non-neu al) alue.
The ques ion ha now a ises is wha ac o s condi ion whe he he
iden i y ea u e has a speci ic alue o he neu al alue ha esul s in a
de aul o m. As men ioned in sec ion 6.6.2, his seems o be p incipally a
ma e o discou se a he han syn ac ic ac o s, bu he ma e equi es
7
Equally, I am making no claims abou de ini eness wi h ega d o ei he D o localiza ion.
Some au ho s ha e claimed ha loca ions in space always in ol e de ini e e e ence (e.g.
MacLaughlin 1997). Ba be à (2012) p o ides compelling e idence ha inde ini e e e en s
may be localized in space in LSC.
8
I owe he obse a ion ha poin op ionali y and localiza ion op ionali y a e sepa a e issues
o Na asha Abne .
Accoun ing o spa ial ag eemen in LSE 287
u he in es iga ion. A loose pa allel can be d awn be ween he use o space
in sign languages and he use o p osody in a spoken language like English:
a ious ac o s can play a ole in shaping p osody including p agma ic,
discu si e and syn ac ic ac o s as well as emo ional con en . The use o space
is also go e ned by mul iple ac o s, such as me apho ic schemes (men ioned
in sec ion 3.1.1) and seman ic conside a ions (such as he use o highe space
o unspeci ic e e en s). The de aul si ua ion is null ma king, bu i ce ain
condi ions hold, a speci ic alue spells ou as a mo e ma ked loca ion. One
such condi ion may be he need o con as be ween di e en discou se
e e en s (as we saw in sec ion 6.5), which licenses he use o di e en iden i y
alues and hus o dis inc i e loca ions. Ano he in luencing ac o is iconic
mo i a ion, which p o ides a mapping ha can accoun o he loca ion
assigned o p esen e e en s o opog aphical desc ip ions. The ole o
iconici y in he use o space is o mally accommoda ed in he wo k o
Schlenke (2011, 2014), and mo e wo k is needed in o de o o malize o he
ac o s ha igge and in luence he use o space.
In summa y, I claim ha he iden i y ϕ- ea u e in LSE occupies a
unc ional head ha domina es he NP. The e a e a ious candida es o his
posi ion (n, Num, D; I also conside ed he N head i sel ). The pa allels
be ween iden i y and he e e en ial index, on he one hand, and he
associa ion be ween he e e en ial index and DP, on he o he , lead me o
p opose ha he iden i y ea u e is hos ed on D, he head o DP. Once he
alued iden i y ea u e is in he nume a ion, i may en e in o ag eemen
ela ions. I now u n o how such a p oposal would accoun o spa ial
ag eemen in LSE.
7.2. Accoun ing o spa ial ag eemen in LSE
I ha e p oposed ha an iden i y ea u e pa icipa es in he ag eemen p ocess
in LSE. In his sec ion I spell ou how his p ocess ope a es in e ms o he
Ag ee p ocess s ipula ed wi hin he Minimalis P og am (desc ibed in sec ion
2.3.3), by looking i s a loca ion assignmen , and hen a e bal ag eemen . I
hen u n o p agma ic ag eemen , a use o space ha looks like an ag eemen
p ocess bu ha does no depend on a speci ic syn ac ic con igu a ion.
7.2.1. Loca ion assignmen
In LSE he e a e h ee mechanisms o loca ion assignmen : poin ing,
localiza ion and classi ie s uc u es (see sec ion 5.1.1). In he i s case, in
which a poin is used o es ablish he loca ion o he e e en , phonological
288 Fo mal analysis o LSE ag eemen
ma e ial is inse ed a D, as shown in (4) o he LSE DP IXx CAT (‘ he/a ca ’).
9
The p esence o a alued iden i y ea u e on he D head gi es ise o a s em
al e na ion a Spell-Ou ha gene a es a o m wi h speci ic spa ial p ope ies
(namely, he indexical poin is di ec ed owa ds a gi en locus in he signing
space). No e ha he phonological ma e ial inse ed a D may be manual –
gi ing ise o a manual poin – o non-manual, such as eye gaze o head il ,
which may also be di ec ed o a locus in he signing space.
(4)
When localiza ion occu s, he spa ial ma king associa ed wi h ( he speci ic
alue o ) he iden i y ϕ- ea u e is exp essed di ec ly on he phonological o m
o he noun i sel . This may be accoun ed o by head mo emen o he N o
he D posi ion so ha he iden i y ea u e a ixes o he lexical head N. This is
shown o he LSE DP HOTELx in (5). The N head HOTEL mo es up o he D
head and is associa ed wi h he alued iden i y ea u e. As a esul , a Spell-
Ou he s em al e na ion inco po a es a speci ic loca ion x in o he o m o he
noun.
9
This is somewha simila o he analysis p o ided by B unelli (2011) o loca ion assigning
indices in LIS and NGT, al hough he e I make use o he iden i y ϕ- ea u e.
Accoun ing o spa ial ag eemen in LSE 289
(5)
As poin ed ou in sec ion 5.6 in he con ex o DP-in e nal ag eemen in LSE,
loca ion assignmen is o en p edica i e in na u e (as e idenced by p osodic
in o ma ion). This is equen ly he case o pos -nominal poin s and classi ie
s uc u es and I ake hese s uc u es o in ol e some so o educed ela i e
clause. In such ins ances he loca ion assigning elemen ( he pos -nominal
poin o classi ie ) en e s in o an ag eemen ela ionship wi h he nominal
an eceden (o , mo e p ecisely, wi h he alued iden i y ea u e on D). This
mechanism o he ag eemen ela ionship o he iden i y ϕ- ea u e is
desc ibed in he nex sec ion.
7.2.2. Ve bal ag eemen
The ϕ- ea u e alues ha appea in loca ion assignmen as desc ibed in he
p e ious sec ion may also en e in o an ag eemen ela ionship. When his
happens, an un alued iden i y ea u e is alued by a alued iden i y ea u e
by means o a p obe checking a goal (as desc ibed in sec ion 2.3.3). This may
occu in he con ex o e bal ag eemen and DP-in e nal ag eemen , and I
ollow Ca s ens (2000, 2001) in assuming ha Ag ee can equally accoun o
he o me and he la e wi h no need o any addi ional specialized
mechanism. In his sec ion, I will ocus on he case o e b ag eemen .
Fu he mo e, as poin ed ou abo e in sec ion 7.1.1, I ollow Danon (2011) in
adop ing a ea u e-sha ing iew o Ag ee in o de o accoun o he ac ha
alued ea u es do no appea on a single head and may be loca ed a
posi ions dis inc om ha o he lexical head wi h which hey a e associa ed.
The DP a gumen o a e b bea s a alued iden i y ea u e. When he
e b en e s in o an ag eemen ela ionship wi h i s a gumen , his esul s in
290 Fo mal analysis o LSE ag eemen
he e b’s own un alued ϕ- ea u es becoming alued. Le us i s examine he
case o single a gumen ag eemen using he LSE sen ence IX1 EXAMx PASSx (‘I
passed he exam’).
10
In his sen ence (based on example (10) in chap e 5) he
e b PASS is a icula ed a he loca ion associa ed wi h he e e en EXAM,
ep esen ed he e as (6).
(6)
The e b’s in e nal a gumen , which occupies he complemen posi ion o V,
has alued ϕ- ea u es o iden i y ( he alue i) and numbe (singula ). The
head con ains un alued ea u es o numbe and iden i y and p obes wi hin
i s domain o ind a goal wi h in e p e able ea u es ha can alue hose on
he p obe. The DP EXAM is a sui able goal since i has alued ϕ- ea u es and
10
No e ha I con inue o use he subsc ip 1 o signs di ec ed owa ds he signe ’s body, such
as he indexical IX1 he e. As poin ed ou in oo no e 9 in chap e 3, his does no commi me o
a i s /non- i s pe son dis inc ion, bu me ely makes he o m o he sign clea e . The ma e
will be discussed in sec ion 8.1.2.
Accoun ing o spa ial ag eemen in LSE 291
he e is no in e ening goal be ween he p obe and his DP.
11
Thus, he alues
o he goal’s ea u es a e assigned o he p obe’s ea u es, as shown by he
do ed line in (6). Subsequen ly, he e b unde goes head mo emen o ,
shown by he dashed line in (6), and he ea u es a ix o he lexical head, hus
making i possible o he co ec phonological o m o be gene a ed by he
phonological s em al e na ion a Spell-Ou , namely, he e b is a icula ed a
he loca ion x associa ed wi h he iden i y alue i.
12
In he case o p o o ypical ag eeing e bs, which show bo h subjec and
objec ag eemen , wo ins ances o Ag ee occu , such as he LSE sen ence IXx
xTRICK1 (‘He’s icking me’), om example (7) in chap e 5. In he i s place,
he objec ag eemen happens along simila lines o wha we saw abo e o
single a gumen ag eemen , wi h he un alued ea u es in he head p obing
and ecei ing he alues o he ea u es on he e b-in e nal DP.
Subsequen ly, when he T head is me ged wi h he es o he s uc u e, he
un alued ea u es he e p obe wi hin he domain. The subjec DP (in Spec- P)
se es as goal, and he ea u es’ alues a e assigned o he ea u es on T. Once
again, head mo emen akes he e b o he T head, hus ensu ing ha he
ele an ea u es a e a ixed and a ailable a Spell-Ou so ha he app op ia e
phonological o m is gene a ed. The subjec ag eemen ma king ea u e spells
ou as loca ion x (associa ed wi h iden i y alue j), and he objec ma king
su aces as a loca ion on he signe ’s body (associa ed wi h iden i y alue i),
yielding he o m xTRICK1. The syn ac ic s uc u e o his p ocess is shown in
(7).
13
The mechanism ou lined he e o ag eemen in LSE can also accoun o
he appea ance and beha iou o auxilia y e bs, such as AUX, desc ibed in
sec ion 5.3.1. When AUX appea s, he Ag ee p ocesses ake place in he same
manne as shown in (7), bu he e b s ays in si u in he head o V and he
auxilia y is inse ed in o he syn ac ic s uc u e di ec ly a he head o . The
ac ha he e a e many pa allels be ween auxilia y e bs and ligh e bs,
11
I do no add ess he issue o whe he he goal is a maximal p ojec ion o a head. Al hough
in he ex I e e o he DP, no e ha bo h ea u es may be p esen on he D head. Likewise,
Danon’s analysis o ag eemen as ea u e sha ing allows o he ea u es o “collec ” a he
head o he DP and hus be a ailable o u he ag eemen ope a ions beyond he DP
domain.
12
In o de o accoun o he canonical su ace o m IX1 EXAM PASS, u he ope a ions a e
equi ed: i he e b con inues on o T, emnan mo emen could hen displace he emaining
s uc u e highe up ( o he le pe iphe y), hus c ea ing he o de . Since my aim he e is o
ocus on he Ag ee ope a ion, I will no dwell on his ma e .
13
Again, u he ope a ions a e equi ed o accoun o he su ace o m. This could be deal
wi h by assuming mo emen o he subjec DP in o Spec TP ( o ul il some so o EPP
cons ain ), o by means o emnan mo emen .
292 Fo mal analysis o LSE ag eemen
which no mally occupy , lends suppo o his posi ion o he auxilia y (see
P au & S einbach (2013) o he GSL auxilia y GIVE-AUX). Thus, om i s ini ial
posi ion AUX is associa ed wi h he ea u es alued by he objec DP, and
subsequen ly mo es up o T, which a ixes he ea u es alued by he subjec
DP. As a esul , AUX has he ea u es equi ed o he ele an phonological
s em al e na ion o be applied a Spell-Ou . Simila ly, in he case o he PERS
auxilia y, which in sec ion 5.3.3 I a gued should be conside ed an ag eemen
auxilia y based on pa allels wi h single a gumen ag eemen , inse ion o he
auxilia y a would p o ide an adequa e syn ac ic amewo k o he
ag eemen p ocess be ween PERS and he e bal a gumen .
(7)
Accoun ing o spa ial ag eemen in LSE 299
means o he syn ac ic ope a ion Ag ee, as s ipula ed by he Minimalis
P og am. In o de o p o ide as cohe en an accoun as possible, I p opose
ha his syn ac ic ag eemen always occu s bu may no show up in he
su ace o m o one o wo easons. Fi s ly, he e b may only allow o
ag eemen wi h one a gumen , as occu s wi h single a gumen ag eemen ,
since i s phonological o m does no allow he inco po a ion o mo e han one
loca ion. Secondly, he iden i y ea u e may ake a de aul alue ha gi es ise
o a neu al loca ion in he phonological o m. This goes some way o
explaining why spa ial ag eemen looks so op ional in sign languages, and
also why ag eemen ma ke omission is possible. The claim he e is ha he
syn ac ic p ocess is obliga o y, bu he a ailabili y o a de aul alue
equen ly ende s he p ocess in isible in he su ace o m.
Finally, I dis inguish be ween syn ac ic ag eemen p ope and
p agma ic ag eemen . In o m bo h exploi space in a e y simila manne , bu
syn ac ic ag eemen c ea es an unambiguous ela ionship be ween a e b and
i s a gumen s, whe eas p agma ic ag eemen gi es ise o an associa i e link
ha mus be esol ed p agma ically. I specula e ha his ela ionship in ol es
independen iden i y ea u es wi h di e en (i.e. unma ched) alues being
mapped on o he same loca ion. As such, he associa ion is o med a he
phonological le el and no he syn ac ic le el. In o de o explain how a
p agma ically ag eeing e b ob ains i s own alued iden i y ϕ- ea u e, I
sugges ha i unde goes nominaliza ion (by me ging wi h n), and me ges
wi h a D head wi h a alued iden i y ϕ- ea u e. This allows he e b o spell
ou a he same loca ion as ano he nominal o c ea e a ela ion o p agma ic
ag eemen be ween he wo.
This concludes he syn ac ic accoun o spa ial ag eemen in LSE. While
i does no claim o be exhaus i e in add essing all he issues ha such a
p oposal aces, he accoun does p o ide a wo king model ha can explain
he main spa ial ag eemen phenomena ound in he language, and
dis inguishes hem om simila uses o space ha do no me i a syn ac ic
accoun . The amewo k o minimalis syn ax has been applied o syn ac ic
ag eemen in o he sign languages (P au, Salzmann & S einbach 2011 o DGS
and NGT; Lou enço 2015 o Lib as) and he analysis p esen ed he e p o ides
an accoun o LSE and ex ends i o p agma ic ag eemen . The nex sec ion
looks a a speci ic aspec o spa ial ag eemen in LSE, namely a phonological
cons ain ha appea s on e bs wi h ce ain phonological cha ac e is ics, and
p o ides a o mal accoun o he di e en o ms ha appea in LSE.
300 Fo mal analysis o LSE ag eemen
7.3. “De ec i e” ag eeing e bs in LSE: an OT accoun
In sec ion 5.4.2.1, when looking a he phonological cons ain s on spa ial
ag eemen o ms in LSE, I desc ibed a g oup o ag eeing e bs wi h a
phonological o m ha complica es he exp ession o ag eemen o one o he
a gumen s. The o m o such e bs includes a lexically speci ied loca ion a
(o nea ) a pa o he body and an unde speci ied mo emen , which means
ha (objec ) ag eemen ma king a he end o he sign is possible, bu he
lexically speci ied loca ion causes p oblems o he exp ession o subjec
ag eemen a he beginning o he sign. In some sign languages his means
ha he ag eemen ma king is no exp essed, esul ing in a de ec i e
ag eemen pa adigm (see sec ion 3.2.1.3). Ne e heless, in LSE al e na e o ms
o he e bs appea ha in ol e modi ica ions o some so o accommoda e
he inclusion o bo h ag eemen ma ke s. Many such e bs in LSE belong o
he seman ic class o speech-ac e bs such as SAY, WARN and TEASE, and he
phenomenon has also been desc ibed o o he sign languages, such as ISL
(Mei 1998b). In his sec ion, I p o ide an accoun o he LSE ac s in e ms o
Op imali y Theo y ( ollowing P ince & Smolensky 1993; McCa hy & P ince
1993, 1999).
Op imali y Theo y (OT) se s ou o explain language o ms in e ms o
op imal ou pu esul ing om a hie a chical se o cons ain s. The basic
concei behind OT is ha an unde lying o m, o inpu , may gene a e many
di e en ou pu s; howe e , only one o hese ou pu s is he success ul
candida e and is he o m su acing in he language. The ques ion is how he
op imal ou pu comes o be selec ed: he se o possible ou comes is e alua ed
agains a se ies o cons ain s, some o which ha e mo e weigh ing han
o he s. The op imal ou pu is he op ion ha bes complies wi h he mos
impo an cons ain s (bu i may, ne e heless, iola e less impo an
cons ain s). This e alua ion p ocess is isualized in ableaux ha show he
di e en cons ain s and how each possible ou come a es in complying o (o
iola ing) each cons ain .
16
OT is mos commonly used o explain
phonological phenomena, bu has been ui ully applied o he
mo phosyn ac ic sphe e in sign languages, such as ecip ocal o ms in DGS
(P au & S einbach 2003).
I p o ide a desc ip ion o he o ms o hese “de ec i e” e bs in LSE
(sec ion 7.3.1), be o e speci ying he OT cons ain s ha a e equi ed o
accoun o hese ac s (7.3.2) and how hese a e applied (7.3.3). Finally, I look
a he da a om ISL o see how his analysis could be ex ended o accoun o
he phenomenon in ha language (7.3.4).
16
The con en ions used in hese ableaux a e explained below in he ele an pa o he ex .
“De ec i e” ag eeing e bs in LSE: an OT accoun 301
7.3.1. “De ec i e” ag eeing e bs in LSE
Commonly, de ec i e e bs can only ma k ag eemen o he objec a gumen ,
since he lexically speci ied (ini ial) loca ion blocks he appea ance o a subjec
ag eemen ma ke a he beginning o he sign. Al hough I ha e labelled hese
e bs as “de ec i e” in LSE, hey o en do show ma king o bo h a gumen s.
Howe e , in o de o achie e his ma king, he e b o ms a e o en mo e
complex han s anda d ag eeing e bs.
I will exempli y he beha iou o hese e bs aking he speci ic case o
WARN, a icula ed om he chin using a closed is wi h c ooked index inge ,
as shown in (11).
LSE (Ai_ag 4:18)
(11)
WARN
In he case o a subjec ha equi es a locus on he body (which would be
conside ed a i s pe son subjec in a pe son-based accoun ),
17
he lexically
speci ied loca ion on he body is su icien o ac as an ag eemen ma ke
locus and so no special modi ica ion is equi ed: he e b is a icula ed as
chin/1WARNx, as shown in igu e 7.1(a). The p oblem a ises when he subjec is a
a locus no on he body. In his case, he sign s a s a he lexically speci ied
loca ion, mo es o he loca ion associa ed wi h he subjec a gumen and hen
mo es on o he loca ion associa ed wi h he objec a gumen . Fo an objec
wi h a locus no on he body, his in ol es passing h ough wo di e en
poin s in he signing space, whe eas o an objec locus on he body, he sign
doubles back o end on he signe ’s body. These LSE o ms, chin-xWARNy and chin-
xWARN1/ches , a e illus a ed in igu es 7.1(b) and (c), espec i ely. Addi ionally,
17
Th oughou his analysis I e e o he ag eemen ma king ia loci ei he on he body o no .
These would co espond o i s pe son and non- i s pe son a gumen s on a pe son-based
accoun . This OT analysis does no depend on a ejec ion o he pe son ea u e, and could be
exp essed in such e ms, bu since I ha e p oposed ha he pe son ea u e does no exis in
LSE, I p esen his accoun in e ms ha a e cohe en wi h my model.
302 Fo mal analysis o LSE ag eemen
hese e bs ha e an al e na e o m o objec locus on he body: he e b s a s
a he loca ion associa ed wi h he subjec a gumen , mo es o he lexically
speci ied loca ion and hen mo es o he signe ’s ches . Thus, x-chinWARN1/ches ,
shown in igu e 7.1(d), has he same meaning as he o m in 7.1(c).
a) chin/1WARNx
b) chin-xWARNy
‘I wa n him.’
‘She wa ns you.’
c) chin-xWARN1/ches
d) x-chinWARN1/ches
‘You wa n me.’
‘You wa n me.’
Figu e 7.1 The exp ession o ag eemen o WARN in LSE. The squa e ep esen s he speci ied
loca ion o he sign ( he chin), he g ey ci cle shows he subjec locus, and he a ow-head he
objec locus.
LSE adds mo emen s and iming uni s o he o m o hese signs in o de o
accommoda e he lexically speci ied loca ion in he e b’s o m. The ollowing
sec ions p esen an OT analysis o accoun o hese o ms. Fi s we look a he
cons ain s ha a e equi ed o gene a e hese o ms.
7.3.2. OT cons ain s
To selec he co ec op imal ou pu o hese “de ec i e” ag eeing e bs in
LSE, h ee cons ain s a e equi ed. (A pu a i e ou h cons ain will be
in oduced in sec ion 7.3.4 o ex end he analysis o he ISL da a.) The
cons ain s I shall use a e gene al modali y-independen s uc u al
conside a ions which ha e been in oked o OT models o spoken language
phonology: IDENT(F), REALIZE(µ) and LINEARITY.
“De ec i e” ag eeing e bs in LSE: an OT accoun 303
The i s wo cons ain s a e ai h ulness cons ain s which ensu e ha
he ou pu espec s ce ain aspec s o he inpu . The i s , IDENT(F), s a es ha
any ea u es ha a e lexically speci ied in he inpu may no be changed in he
de i a ion: ea u es may be added, and unspeci ied ea u es may be speci ied,
bu hose ha a e al eady speci ied mus be espec ed (McCa hy & P ince
1995; 1999: 55-56). In he example unde conside a ion, his cons ain is
ele an o he place o a icula ion ea u e o he e b oo , he ini ial
loca ion, which is al eady speci ied, and ensu es ha he loca ion is
main ained in he ou pu .
IDENT(F): Fea u es speci ied in he inpu may no be changed.
The second ai h ulness cons ain , REALIZE(µ), s a es ha all mo phemes
p esen in he inpu mus be p esen in he ou pu (see Ku isu 2001: 39, and
e e ences he ein).
18
The ou pu may con ain ea u es ha do no appea in
he inpu , bu hose mo phemes ha a e he e in he inpu mus ha e a
phonological e lex in he ou pu . This cons ain ensu es ha in lec ion is
igge ed: ag eemen mo phemes included in he inpu mus be p esen in he
ou pu . Those e bs which ail o show ag eemen o bo h a gumen s iola e
his cons ain .
REALIZE(µ): All mo phemes in he inpu mus ha e a phonological
e lex in he ou pu .
The hi d cons ain equi ed, LINEARITY, speci ies ha he inpu and ou pu
ha e consis en p eceden s uc u es (McCa hy & P ince 1995; 1999: 55-56).
Fo spoken languages, his cons ain ules ou me a hesis, bu as will become
clea sho ly, he applica ion o sign language will be sligh ly di e en .
LINEARITY: The inpu is consis en wi h he p ecedence s uc u e o
he ou pu , and ice e sa.
The nex sec ion desc ibes how hese cons ain s a e applied and o de ed o
accoun o he LSE da a.
7.3.3. Applying he cons ain s
Re u ning o he example o he LSE ag eeing e b WARN, he inpu o he
inal mo phophonological o m is he usion o he e b oo and he
18
Ku isu’s o mula ion o REALIZE(µ) is mo e complica ed han he e sion used he e bu he
echnicali ies a e no equi ed o his analysis.
304 Fo mal analysis o LSE ag eemen
phonological s em al e na ions gene a ed as a esul o he alued iden i y
ea u e a he T and heads, esponsible o subjec and objec ag eemen ,
espec i ely (see sec ion 7.2.2 abo e o de ails). Each s em al e na ion is he
esul o a mo pheme ha a aches o he e b oo and he posi ion ha each
ag eemen mo pheme occupies on he lexical e b is de e mined (see he
discussion in sec ion 7.2.2 on he ela i e me i s o a syn ac ic o lexicalis
accoun o he o de ing o hese elemen s): o p o o ypical ag eeing e bs,
he subjec ag eemen mo pheme appea s a he beginning o he ou pu and
he objec ag eemen mo pheme a he end. (This s ipula ion will be e ined
du ing he analysis and discussed a he end o his sec ion.)
The e b oo is speci ied o handshape, mo emen and ini ial loca ion
[chin], and consis s o a syllable in he sense o a canonical loca ion-
mo emen -loca ion sequence (B en a i 1998; Sandle & Lillo-Ma in 2006). The
loca ion slo s o each ag eemen mo pheme a e illed wi h he loca ions [x]
and [y] esul ing om he alues on he espec i e iden i y ea u es. As such,
he inpu elemen s may be ep esen ed schema ically as (12).
(12)
WARN:
Ag eemen mo phemes:
σ
μ
μ
loca ion
[chin]
loca ion
[ ]
loca ion
[x]
loca ion
[y]
spell-ou si e o
[i iden i y]
spell-ou si e o
[j iden i y]
No mally o ag eeing e bs bo h loca ion slo s a e emp y in he phonological
ma ix. A e he me ge wi h he ag eemen mo phemes, he esul ing e b
complex has wo slo s which can se e as Spell-Ou si es o he locus
speci ica ions o he e b’s a gumen s and he inal ou come o he
mo phophonological p ocess is a monosyllabic sign wi h he o m [x]σ[y], such
as xTRICKy. Howe e , in he case o WARN- ype e bs, his is no possible: he
copying o [x] in o he i s slo o he complex e b is blocked by he p esence
o he speci ica ion o [chin], which was inhe i ed om he e b oo in he
me ge p ocess. The e a e se e al possible al e na i e op ions a ailable. The
sign may change in some way o accommoda e he blocking: he lexically
speci ied loca ion ea u e may be supp essed, o an ex a syllable may be
added o p o ide an addi ional loca ion slo . Al e na i ely, he ag eemen
may be only pa ially exp essed (ag eemen wi h one a gumen ins ead o
wi h bo h) o may no be exp essed a all. We now e alua e hese di e en
“De ec i e” ag eeing e bs in LSE: an OT accoun 305
ou pu op ions in e ms o he OT cons ain s desc ibed in he p e ious
sec ion.
Following s anda d con en ion, an OT ableau displays he inpu and
he compliance o he possible ou pu candida es wi h he cons ain s. The
inpu is shown in he op le cell. The di e en ou pu op ions a e in each ow
o he le mos column, and he cons ain s (in hie a chical o de ) in he
columns o he igh (wi h labels o each in he op ow). The iola ion o a
cons ain by a candida e is ma ked by * in he co esponding cell o he
ableau. The op imal ou pu (ma ked wi h he symbol ☞) is ha which ei he
has no iola ions o only has iola ions u he o he igh han all o he
op ions. When a iola ion ules ou a candida e, i is accompanied by an
exclama ion ma k o show i s a ali y, and he es o he ow is shaded in.
Cons ain s may be iola ed as a ma e o deg ee and mul iple iola ions a e
ma ked wi h he co esponding numbe o as e isks.
As a as he o de ing o hese cons ain s is conce ned, he p oposed
hie a chy is as ollows:
Ranking o cons ain s in LSE: IDENT(F) » LINEARITY » REALIZE(µ)
The cons ain s a e inse ed in o he ableau in his o de .
Tableau 1. Ag eemen o WARN- ype e bs in LSE.
[chin]σ, μx, μy
IDENT(F)
LINEARITY
REALIZE(µ)
[chin]σ[neu ]
**!
[x]σ[y]
*!
[chin]σ[y]
*
[chin]σ[x]
*!
*
☞ [chin]σ[x]σ[y]
[x]σ[chin]σ[y]
*!
Tableau 1 shows he e alua ion o a numbe o di e en ou pu op ions o
WARN- ype e bs in LSE. The ci a ion o m o he e b ( he i s candida e)
ails o show ag eemen and so iola es he REALIZE(µ) cons ain on wo
coun s since nei he o he ag eemen mo phemes is p esen in he ou pu .
The op ion o exp essing ag eemen as no mal ag eeing e bs do ( he second
candida e) iola es he IDENT(F) cons ain since he lexically speci ied ea u e
[chin] has been changed o [x]. Main aining he lexically speci ied ea u e and
ma king ag eemen wi h only he objec a gumen ( he hi d candida e),
iola es he REALIZE(µ) cons ain bu only once since one o he ag eemen
306 Fo mal analysis o LSE ag eemen
mo phemes is exp essed. Likewise, subjec ag eemen only ( he ou h
candida e) iola es REALIZE(µ) o ailing o include one o he ag eemen
mo phemes, and addi ionally iola es he LINEARITY cons ain since in he
inpu [x] canno be loca ed a he end o he e b, bu in he ou pu i appea s
in he coda posi ion o he e b syllable. The op imal ou pu (candida e i e)
ul ils all he cons ain s by main aining he lexically speci ied ea u e o he
e b oo and adding an ex a syllable, hus c ea ing loca ion slo s o he
ealiza ion o he ag eemen mo phemes. The ela ed s a egy o adding a
syllable bu changing he o de o he loca ions o [x]>[chin]>[y] (candida e
six) c ea es a iola ion o he LINEARITY cons ain . Al hough [x] is now in an
onse loca ion and [y] a he end o he sign, he ac ha he wo elemen s a e
no neighbou ing (i.e. [x] does no immedia ely p ecede [y] due o he
in e en ion o he lexically speci ied ea u e [chin]) does no comply wi h he
linea i y conside a ions.
The ableau shows ha his pa icula hie a chy o cons ain s co ec ly
p edic s he op imal ou pu o WARN- ype ag eeing e bs. Howe e , he
alue o a cons ain hie a chy lies in i s applicabili y o mo e han one
candida e se . Recall ha WARN- ype e bs in LSE allow wo o ms in he case
o ag eemen wi h an objec locus on he body: he gene al pa e n used o
any ype o objec (i.e. he op imal ou pu in ableau 1), and he [x]>[chin]>[y]
o m (co esponding o candida e six in ableau 1). How is i ha a candida e
ha was ejec ed in he gene al pa adigm becomes accep able in he case o an
objec locus on he body?
The answe lies in he ac ha he [chin] loca ion may coun as he locus
o he objec locus on he body, in which case he linea in eg i y o he
ag eemen mo phemes is p ese ed by he i s syllable ([x]σ[chin] is equi alen
o [x]σ[y]). No e ha he [chin] loca ion also se es as he subjec locus on he
body: in such o ms he mo emen o he sign is simply [chin]σ[y] (as shown in
igu e 7.1(a)), and no [chin]σ[ches ]σ[y].
19
This migh seem o ende he [ches ]
loca ion edundan since he objec ma king has al eady been achie ed.
Howe e , no e ha he inclusion o his second syllable pu s he lexically
speci ied ea u e [chin] in a syllable-ini ial posi ion, hus espec ing he
iden i y and linea i y condi ions s ipula ed in he inpu . The [ches ] loca ion
appea s o be some so o de aul o dummy loca ion o ill he emp y slo o
19
Fu he suppo o accep ing [chin] as a alid locus o ag eemen ma king on he body is
p o ided by he obse a ion ha many ag eeing signs a e a icula ed a a speci ic heigh , o
bo h he subjec and objec a gumen s (Liddell 1995). Fo example, LSE UNDERSTAND is
a icula ed a he heigh o he o ehead/ emple whe eas GIVE is a ches heigh . In he case o
WARN, he chin is he ele an heigh .
“De ec i e” ag eeing e bs in LSE: an OT accoun 307
he second syllable. Tableau 2 indeed shows ha LSE has wo op imal ou pu s
o ma king objec ag eemen on he body in WARN- ype ag eeing e bs.
Tableau 2. Ag eemen o WARN- ype e bs wi h objec locus on he body in LSE.
[chin]σ, μx, μy
IDENT(F)
LINEARITY
REALIZE(µ)
[x]σ[ches ]
*!
[chin]σ[ches ]
*!
☞ [chin]σ[x]σ[ches ]
☞ [x]σ[chin]σ[ches ]
Be o e mo ing on o ex ending his analysis o da a om ano he sign
language, I wish o make some commen s on he na u e o he LINEARITY
cons ain . Va ious candida es ha e been excluded on he g ounds o his
cons ain , which depends upon he p ecedence s uc u e o he inpu and
ou pu ma ching each o he . No ably, o he lexical oo , he lexically
speci ied loca ion does no need o be sign ini ial, bu canno be sign inal.
Equally, he same s ipula ion holds ue o he subjec ag eemen mo pheme:
i need no be sign ini ial, bu canno be sign inal. Addi ionally, he subjec
and he objec ag eemen mo phemes mus appea con iguously, ha is o
say, in he same syllable. This sugges s ha he ag eemen mo phemes a e no
jus a pai o independen loca ion slo s and ha he wo mus exis wi hin a
phonological uni . This lends suppo o Mei ’s (1998b, 2002) p oposal ha
ag eemen in ag eeing e bs is due o a DIR mo pheme wi h a syllabic
s uc u e (see sec ion 3.2.2.3 o de ails o his p oposal). The ac ha he wo
ag eemen ma ke s mus be connec ed in his speci ic way may se wo-place
ag eeing e bs apa om he gene al spa ial ag eemen mechanism a gued
o in his hesis (and may be why hey ha e a ac ed so much in e es in he
ield). Fu he mo e, p agma ic ag eemen only occu s when a single spa ial
ma ke is used and is no possible o e bs ha spa ially ma k wo
a gumen s. This lends u he suppo o he idea ha he mo emen be ween
he wo ag eemen ma ke s is ele an and con ibu es o he syn ac ic
in eg i y o wo-place ag eeing e bs.
This sec ion has laid ou an OT analysis o “de ec i e” ag eeing e bs in
LSE based on he a ious o ms ha exis in he language. We now u n o he
same phenomenon in ano he sign language o see whe he he cons ain s
p oposed can also explain he ac s.
308 Fo mal analysis o LSE ag eemen
7.3.4. Ex ending he analysis o ISL da a
Ha ing shown ha he OT analysis explains he LSE ac s, he nex ques ion is
whe he his cons ain hie a chy can be applied o he ISL da a. As desc ibed
in sec ion 3.2.1.3, WARN- ype e bs in ISL show ag eemen wi h he objec
alone (candida e h ee in ableau 1) o mos o ms and ag eemen wi h bo h
a gumen s when he objec locus is on he body (candida e ou in ableau 2).
Re iewing ableau 1 o LSE makes clea ha eo de ing he cons ain s
would no i he ISL ac s since candida e i e does no iola e any o he
cons ain s and will always come ou op. This sugges s ha ISL could ha e
ano he cons ain which excludes candida e i e. Le us call his cons ain C
and place i be o e REALIZE(µ) in he cons ain hie a chy. F om ableaux 3
and 4 i can be seen ha his modi ica ion would gi e he co ec esul s.
Tableau 3. Ag eemen o WARN- ype e bs in ISL.
[chin]σ, μx, μy
IDENT(F)
LINEARITY
C
REALIZE(µ)
[chin]σ[neu ]
**!
[x]σ[y]
*!
☞ [chin]σ[y]
*
[chin]σ[x]
*!
*
[chin]σ[x]σ[y]
*!
[x]σ[chin]σ[y]
*!
Tableau 4. Ag eemen o WARN- ype e bs wi h objec locus on he body in ISL.
[chin]σ, μx, μy
IDENT(F)
LINEARITY
C
REALIZE(µ)
[x]σ[ches ]
*!
[chin]σ[ches ]
*!
[chin]σ[x]σ[ches ]
*!
☞ [x]σ[chin]σ[ches ]
Al hough his solu ion i s wi h he da a, we need o be able o speci y wha
his new cons ain is in o de o he inding o be meaning ul.
Un o una ely, i is no ob ious wha he exac na u e o cons ain C could
be. One en icing op ion is a *SYLLABLE cons ain ha ules ou he inse ion o
an addi ional syllable. Such a cons ain has been pos ula ed independen ly in
he con ex o e lexi e o ms in DGS (P au & S einbach 2003). This would
ce ainly i well wi h he si ua ion shown in ableau 3, since i would exclude
he i h candida e on he g ounds o con aining an ex a syllable. Also, his
cons ain could be added o he LSE ableaux bu wi h a lowe anking wi h
Conclusions 315
desc ibed in chap e 5. This ou line accoun ed o loca ion assignmen , single
a gumen ag eemen , ag eeing e bs and ag eemen auxilia ies. Fu he mo e,
he syn ac ic model also p o ided a means o dis inguishing be ween hese
syn ac ic mani es a ions o ag eemen (in ol ing he Ag ee ope a ion) and
o he uses o space ha gi e ise o loose associa ions ha mus be esol ed
p agma ically (i.e. p agma ic ag eemen ). Wi h he modi ica ion o he
ϕ- ea u e, he LSE da a a e amenable o he ype o syn ac ic analysis ha has
been de eloped on he basis o spoken language da a. In he cookbook o
human languages, some ing edien s may be di e en , bu he basic ecipes a e
he same.
In addi ion o cha ac e izing spa ial ag eemen in LSE om a syn ac ic
poin o iew, his chap e has also examined a speci ic aspec o ag eeing
e bs. In sec ion 5.4.2, I examined he phonological cons ain s ha ope a e on
ag eeing e bs, and desc ibed a class o e bs ha ha e non-pa adigma ic
o ms due o a con lic be ween a lexically speci ied phonological ea u e and
he ag eemen ma king. Al hough hese ypes o e bs a e o en de ec i e in
o he sign languages, in LSE a ious s a egies allow he ag eemen ma ke s
o su ace on he e b. Sec ion 7.3 p o ided a o mal analysis, using he
amewo k o Op imali y Theo y, o accoun o hese o ms. The amewo k
and cons ain s i appeals o a e no language (o modali y) speci ic, ye hey
p o ide an accoun o he beha iou o hese e bs in LSE and go a long way
owa ds explaining he da a o simila e bs in ISL.
To conclude, we can now o e an answe o he hi d and inal ques ion
de ined in chap e 1: Can spa ial ag eemen in LSE be gi en a o mal
cha ac e iza ion? This chap e has o e ed a o mal cha ac e iza ion o he basic
mechanism o spa ial ag eemen in LSE. A gene al model based on he
heo e ical appa a us o e ed by minimalis syn ax p o ides an adequa e
accoun o he LSE da a and explains he ange o phenomena ha exhibi
spa ial ag eemen . Addi ionally, Op imali y Theo y was applied o a speci ic
spa ial ag eemen phenomenon in LSE and also p o ided an app op ia e
means o accoun o he phonological o m o his class o ag eeing e bs.
De eloping hese analyses b ough up a ious issues, no ably, op ionali y, he
loca i e/locus dis inc ion and linea i y conside a ions (sec ion 7.4). In
discussing and add essing hese ma e s, I examined he consequences o his
syn ac ic accoun , o unde line i s s eng hs and o acknowledge i s
limi a ions. In his sense, he o mal accoun o LSE ag eemen o e ed he e
no only con i ms ha his spa ial mechanism can be cha ac e ized
syn ac ically, bu also p o ides a e inemen o he second esea ch ques ion
by iden i ying how LSE di e s om spoken languages. Once mo e, he
in e ela ed issues o op ionali y and he basic e e en ial mechanism se sign
[Document text truncated for crawler view.]
