3RD PERSON PRONOMINAL CLITICS IN DIALECTS OF CATALAN
Euldia Bone
Uni e si a Au bnoma de Ba celona
This pape p o ides a desc ip ion and an analysis o combina ions o wo hi d
pe son cli ics, accusa i e and da i e, in ou a ie ies o Ca alan. While in one o
hese ou a ie ies, he ou pu o ms a e s aigh on a d, in he o he h ee he e is
an inc easing p esence o opaque ou pu s, ha is ou pu s ha di e om he o m
he cli ics ha e when hey appea in isola ion. Bo h he linea o de among he cli ics
and he p esence o opaque o ms is a gued o ake place in he Mo phology
Componen , be ween s-s uc u e and
PF
(a syn ac ic app oach o cli ic o de being
ejec ed). In his componen , he in e nal s uc u e o he cli ics, a hie a chical
s uc u e o ea u es,
can
be al e ed by se e al ypes o p ocesses (dele ion, inse ion):
inse ion is a gued o apply in one o he a ie ies unde discussion. Linea o de
among p onominal cli ics is de e mined by he mapping on o a empla e. This
mapping
can
be based on ei he whole cli ics o indi idual e minal ea u es. All
possibili ies exis in he ou a ie ies, and in wo cases hey coexis in a gi en
dialec . The es o he opaque o ms in he ou a ie ies a ise du ing he mapping
ope a ion, when wo ins ances o a gi en ea u e a e mapped on o one and he same
slo . An addi ional sou ce o a ia ion lies on be ea u e on which [ eminine]
depends.
When p onominal cli ics a e combined, hey o en adop an opaque o m; ha is, hey do no
appea wi h he shape hey ha e when hey appea in isola ion. The e a e se e al kinds o
modi ica ions cli ics can unde go, and he e is a e y conside able amoun o dialec al a ia ion
in his espec . Thi d pe son p onominal cli ics a e pa icula ly p one o su e modi ica ions
when hey a e combined wi h each o he . In his pape , ou di e en hi d pe son (accusa i e
and da i e) p onominal cli ic sys ems o Ca alan will be examined, which di e wi h espec o
Ca alan Wo quing Pape s in Linguis ics (CWPL)
3.1
(1993):
85-
11
1
Uni e si a Au bnoma de Ba celona
he ou pu o cli ic combina ions bu ha dly wi h espec o he o ms hose cli ics ha e in
isola ion, and a p oposa1 will be made o accoun o he a ia ion. One o hese sys ems
belongs o Cen al Ca alan, and i is sha ed by a la ge numbe o speake s. This is he dialec
wi h he mos opaque o ms. Ano he sys em ha will be aken unde conside a ion is
Majo can, in one o i s a ie ies, no spoken by younge people nowadays. The o he wo
sys ems o be examined co espond especially o s anda d a ie ies o Ca alan. One o hem is
S anda d Valencian, which is also spoken in pa s o Valbncia and Cas elló. The o he one is he
s anda d a ie y o Ca alan o he es o he Ca alan speaking a ea, also known as S anda d
Ca alan. Fo con enience, I will concen a e only on his a ie y as spoken in Cen al Ca alan.
The cli ic sys em o his a ie y is pa o he colloquial language spoken in ce ain, e y
limi ed, a eas. Th oughou his a icle
I
will use le e s o e e o each sys em in o de o a oid
any possible con usions. The e ms will be: a ie y A (S anda d Valencian), a ie y
B
(S anda d Caialan), a ie y
C
(Majo can), and a ie y
D
(Cen al Ca alan).
The pape is o ganized as ollows: in sec ion
1
I p esen he ele an da a conceming he ou
a ie ies o Ca alan men ioned in he p e ious pa ag aph, o ganized in inc easing deg ees o
opaci y. In sec ion 2 a p oposa1 is made o accoun o all he opaque o ms, which is based on
mo phology. Syn ac ic and phonological app oaches will be ejec ed in iew o he
cha ac e is ics o he da a being analyzed. The pape also con ains a las sec ion wi h
conclusions.
1.
The
da a
I .I.
Va ie y
A
This a ie y o Ca alan, spoken in ce ain a eas o Valbncia and Cas elló, has no opaque o ms
in combina ions o wo hi d pe son cli ics, accusa i e and da i e. As can be seen in mos cases,
he da i e cli ic p ecedes he accusa i e cli ic. These ac s a e illus a ed in (1) and (2). In (la-b),
and (2a-b), he cli ics appea in isola ion. The examples in (lc) and (2c) con ain combina ions
o he p e iously shown cli ics. In all he examples h oughou his pape
I
use he Cli ic-Le
Disloca ion cons uc ion in o de o clea ly show he in lec ional cha ac e is ics o he
an eceden . All cli ics and cli ic combina ions appea in bold ace.
(1)
a. Al xique ,
li
dona é la poma més a d.
o- he boy 3sgDa gi e-lsg-FUT he apple la e
b.
La
poma,
la
dona é al xique més a d.
he apple( em) 3sg emAcc gi e-lsg-FUT o- he boy la e
c. Al xique , la poma
li
la
dona é més a d.
o- he boy he apple 3sgDa 3sg emAcc gi e-lsg-FUT la e
'I
will gi e he apple o he boy la e .'
(2) a. Als xique s,
els
dona é les pomes més a d.
o- he boys 3plDa gi e-lsg-FüT he apples la e
b. Les pomes,
les
dona é als xique s més a d.
he apples( em) 3pl emAcc gi e-lsg-FUT o- he boys la e
c. Als xique s, les pomes,
els les
dona é més a d.
o- he boys he apples 3plDa 3pl emAcc gi e-lsg-FUT la e
'I
will gi e he apples o he boys la e .'
As can be seen in (3), below, he hi d pe son da i e plu al cli ic is phonologically iden ical o
he hi d pe son accusa i e plu al (masculine) cli ic. This is so in all he a ie ies o Ca alan ha
will be aken in o conside a ion, excep in one ( a ie y D). As will be shown la e , he iden i y
be ween he hi d pe son accusa i e plu al cli ic and he hi d pe son da i e plu al cli ic is only
phonological, no mo phological. The o n o he hi d pe son da i e cli ic in isola ion can be
seen in (2a).
(3) Als xique s, els p éssecs,
els els
dona é més a d.
o- he boys he peaches(masc) 3plDa 3plmascAcc gi e-lsg-FüT la e
'I
will gi e he peaches o he boys la e .'
1.2.
Va ie y
B
The hi d pe son p onominal cli ic sys em o his a ie y o Ca alan is iden ical o ha o
Valencian in all espec s excep when a hi d pe son accusa i e cli ic is combined wi h a hi d
pe son da i e
singula
cli ic. In his case opaque ou pu s a ise gi en ha he hi d pe son da i e
singula cli ic,
li
in isola ion, seems o adop he o m /i/ (spelled hi).
(4b) illus a es he combina ion o a hi d pe son accusa i e cli ic wi h a hi d pe son da i e
plu al cli ic. As can be seen by compa ison wi h (2c) he ou pu s in hese wo a ie ies a e
iden icai:
(4)
a. Als nois, els dona é les pomes més a d.
o- he boys 3plDa gi e-lsg-FUT he apples la e
b. Les pomes, als nois, els les dona é més a d.
he apples( em) o- he boys 3plDa 3pl emAcc gi e- lsg-FUT la e
'I
will gi e he apples o he boys la e .'
(5a) shows he o m o he hi d pe son da i e singula cli ic in isola ion. In (5b) his cli ic
appea s combined wi h a hi d pe son accusa i e cli ic, and adop s he o m li1 (hi). As shown,
he non-opaque ou pu s
les li
and
li les
a e non-exis en .
(5) a. Al noi, li dona é les pomes mCs a d.
o- he
boy
3sgDa gi e-lsg-FUT he apples la e
b. Al noi, les pomes, les hi dona 6 més a d.
o- he boy he apples( em) 3pl emAcc 3sgDa (/i/) gi e-lsg-FUT la e
(*les li
1
*li les)
'I
will gi e he apples o he boy la e .'
I is also impo an o no e ha , while in combina ions wi h a hi d pe son da i e plu al cli ic he
su ace o de is da i e
>
accusa i e, when he hi d pe son da i e cli ic is singula (when he
opaque o ms a ise), he o de is accusa i e
>
da i e.
1.3.
Va ie y
C
This a ie y o Ca alan, spoken in Majo ca, has no anspa en combina ions; all he ou pu s in
combina ions o wo hi d pe son cli ics a e opaque. Like in a ie y
B,
he hi d pe son da i e
singula cli ic, Ili1 in isola ion, adop s he o m li/ when i is combined wi h an accusa i e cli ic.
Mo eo e , he hi d pe son accusa i e cli ic ne e shows eminine ag eemen , e en hough a
eminine ma ke is p esen when his cli ic appea s in isola ion and he an eceden is eminine.
These wo aspec s a e illus a ed in (6a-c). I also gi e, in pa en heses, he non-exis en
anspa en o ms:
(6)
a. Ses pomes, les dona é a s'al.lo més a d.
he apples( em) 3pl emAcc gi e-lsg-FüT o he-boy la e
b. A s al.lo , li dona é ses pomes més a d.
o he-boy 3sgDa gi e-lsg-FLJT he apples la e
c. Ses pomes, a s al.lo ,
[alai
dona é més a d.
he apples
o
he-boy 3sgDa +3pl emAcc (Ilzil) gi e-lsg-FUT la e
(*les
li
I *li les)
'I
will gi e he apples o he boy la e .'
I one we e o say ha he o n /i/ co esponds o he da i e cli ic, he su ace o de be ween he
wo cli ics in (6c) would be accusa i e
>
da i e, like in a ie y
B
and unlike a ie y A.
I
gi e he opaque ihi d pe son da i e ou pu in i s phonological o n only because he e is no s anda d spelling
o ihis a ie y. O hen ise li1 would be spelled
hi.
I
will gi e only phonological o ns in such cases.

In combina ions o wo hi d pe son cli ics in which he da i e cli ic is plu al o he opaque
o ms a ise. In his case, he da i e plu al cli ic,
els
(/lz/) emains anspa en bu he accusa i e
cli ic becomes /u/ in all cases, ega dless o he gende and numbe o he an eceden . This
"change" o /o/ is also exis en in ce ain a ie ies o Valencian, as desc ibed in Todoli (1992).
Ca alan has an independen cli ic Iol (o /u/),
ho,
he so-called neu e cli ic, and la e i will be
claimed ha , h ough a speci ic mo phological p ocess, he hi d pe son accusa i e cli ic
becomes a neu e cli ic. This "change" is illus a ed in (7c). The cli ics ha show up in an
opaque o m in his combina ion appea in isola ion in (7a,b). As done ea lie ,
I
include, a e
(7c) and in pa en heses, he non-exis en anspa en ou pu o ms o he combina ion:
(7) a. Ses pomes, les dona é a e s al.lo s més a d.
he apples( em) 3pl emAcc gi e-lsg-FLJT o he boys la e
b. A e s al.lo s, els dona é ses pomes més a d.
o he boys 3plDa gi e-lsg-FUT he apples la e
c. Ses pomes, a e s al.lo s,
[alzo]
dona é mCs a d.
he apples o he boys 3plDa +3pl emAcc
[ah]
gi e-lsg-FUT la e
(*els les
/
*les els)
'I will gi e he apples o he boys la e .'
I we compa e (7a) wi h (7c) we will see ha , e en hough he hi d pe son accusa i e cli ic in
(7a) is emine and plu al, hose ma ke s do no show up in he combina ion in (7c). Ins ead, he
in a iable o m /o/ appea s.
While he su ace o de could be said o be accusa i e
>
da i e when he da i e cli ic was
singula (c . (6c)), he su ace o de would be da i e
>
accusa i e whene e he da i e cli ic is
plu al (c . (6c)).
1.4.
Va ie y
D
Va ie y D co esponds basically o Cen al Ca alan, and has no anspa en ou pu s in
combina ions o wo hi d pe son cli ics. A basic di e ence be ween a ie y D and he o he
a ie ies p e iously desc ibed is ha he o m o he hi d pe son da i e plu al cli ic is Ilzil,
while in he o he a ie ies his cli ic has he phonological o m Ilzl. As in Majo can, no
eminine ma king e e appea s in such combina ions. Mo eo e , whene e he wo cli ics a e
singula , he ou pu o m o he combina ion is Ilil, while whene e one o he wo cli ics o
bo h o hem a e plu al, he ou pu o m is IIziI. These aspec s a e illus a ed in (8) and (9),
below.
(8) a.
La
poma,
la
dona é als nois més a d.
he apple( em) 3sg emAcc gi e- lsg-FUT o- he boys la e
b. Als nois,
[alai
dona é la poma més a d.
o- he boys, 3plDa gi e-lsg-FüT he apple la e
c. Als nois, la poma,
[hi
dona més a d.
o- he boys he apple 3plDa +3sg emAcc (Ilzil) gi e-lsg-FUT la e
(*
[l*ll
*[Mal)
'I
will gi e he apple o he boys la e .'
In
(8b) we see he o m o he hi d pe son da i e plu al cli ic in isola ion, IIziI. As can be seen
in (8c), he eminine ma ke ha was p esen in
(84
does no appea in his las example.
F om examples like (8c), we migh conclude ha wha is eally happening in combina ions o
wo hi d pe son cli ics is ha only he da i e cli ic shows up in such combina ions. Recall ha
in his a ie y he phonological o m o he hi d pe son da i e plu al cli ic in isola ion is Ilzil,
like he o n ha appea s in
(SC).
The omission o he hi d pe son accusa i e cli ic is in ac
common in ce ain dialec s o F ench and also in e y es ic ed a eas o he Ca alan speaking
domain. Howe e , his canno
be
he case in a ie y D, as example (9c) shows:
(9)
a.
Les pomes,
les
dona é al noi m s a d.
he apples( em) 3pl emAcc gi e-lsg-FUT o- he boy la e
b. Al noi, li dona 6 les pomes m6s a d.
o- he boy 3sgDa gi e-lsg-FUT he apples la e
c. Al noi, les pomes, [ah] dona é més a d.
o- he boy he apples 3pl emAcc+3sgDa gi e-lsg-FUT la e
(*li les /*les li)
'I
will gi e he apples o he boy la e .'
As can
be
seen in (9c), he ou pu o m o he combina ion is Ilzil, like he hi d pe son da i e
plu al cli ic in isola ion, bu in his case he an eceden o he da i e is singula . The e o e we
canno conclude ha he oupu o he combina ion simply igno es he accusa i e cli ic. The
plu al ma ke in (9c) p ecisely has o come om he accusa i e cli ic. In (9c) i can also be seen
ha he eminine ma ke coming om he accusa i e cli ic does no show up in he combina ion.
(IOC), below, illus a es he ac ha he o m llzil is also he oupu o m in combina ions o wo
hi d pe son plu al cli ics. As usual, in (10a,b) I gi e he o m o he cli ics in ol ed in
(10c)
as
hey appea in isola ion. I also include he non-exis en anspa en possible o ms:
(10) a. Les pomes, les dona 6 als nois m6s a d.
he apples( em) 3pl emAcc gi e- lsg-FUT o- he boys la e
b. Als nois, [al3 dona 6 les pomes m6s a d.
o- he boys 3plDa gi e-lsg-FUT he apples la e
c. Als nois, les pomes, [alai dona 6 m6s a d.
o- he boys he apples 3plDa +3pl emAcc (Ilzil) gi e- lsg-FUT la e
(*
[alzilas] I
*
[laaahi)
'Dona 6 les pomes als nois m6s a d.'
In his a ie y o Ca alan i is impossible o ell wha he o de is be ween accusa i e and da i e.
1.5.
Summa y o
he
ac s
In (1 l), below, a cha is gi en wi h he ou pu phonological o ms o he combina ions o wo
hi d pe son da i e cli ics, accusa i e and da i e, in he ou a ie ies o Ca alan p e iously
desc ibed. In many cases, some epen he ic owels will be added (d=da i e; a=accusa i e;
s=singula ; p=plu al; m=masculine; = eminine):2
(11)
a .
A
a .
B
a .
C
a .
D
d.s.
+
a.s.m. Ai11 Ai1 Ai1 Ai1
ds.
+
a.s. . Aild Ilail Ai1 Ai1
d.s.
+
a.p.m. Aild Azil Azil ~zil
ds.
+
a.p. . Ailed Ilazil nzil nzil
d.p.
+
a.s.m. Azll Azll /lzo/ Azil
d.p.
+
a.s. . nzld Ilzlal Ilzol Azil
d.p.
+
a.p.m. nzld nzld
I~ZOI
nzi/
d.p.
+
a.p. . Azled Ilzlad /lzol llzil
In (12), below, I gi e he o ms o all he hi d pe son cli ics as hey appea in isola ion. All
a ie ies ha e basically he same o ms in all he cases, excep when he hi d pe son da i e
cli ic is plu al. In his case, a ie y
D
has a di e en o m. Ano he di e ence, which does no
ha e any bea ing on he opic o his pape , is he phonological o m o he eminine ma ke . In
mos a ie ies i has he o m la1 (o some /VI om which la1 can be de i ed by Vowel
Dele ion), while in a ie y
1
i has he o m
/a/
when eminine cooccu s wi h singula , and /e/
when i cooccu s wi h plu al:3
2
The o ms gi en in
(11)
a e he ones he cli ics ha e in p e e bai posi ion. In pos e bai posi ion he e a e
some modi ica ions no ela ed o he opic o his pape .
3
This pa e n
can
aiso be ound in S anda d Ca alan ( a ie y
B)
as spoken in he no h-eas enl egion o he
Ca alan
a ea
(18)
-
-
[ARG] [OBL]
In combina ions o wo hi d pe son cli ics, wo nodes [ARG] a e always p esen , one o hem
coming om he da i e cli ic and he o he one coming om he accusa i e cli ic. Bo h ins ances
o [ARG] will be mapped on o he empla e. (19) illus a es he mapping o wo hi d pe son
cli ics, accusa i e and da i e, bo h o hem singula and masculine ( ha is, wi h no ag eemen
ea u es). Recall ha he cli ics hemsel es a e no linea ized; he da i e cli ic could ha e also
been w i en be o e he accusa i e cli ic:
Resul ing ou pu o n: /li/
As indica ed, [ARG] will be spelled ou as /l/, h ough ule (15b). and [OBL] will be spelled
ou as li/, h ough ule (1%). When one o he wo cli ics (o bo h o hem) is plu al, he
ea u e [plu al] will be a dependen on [ARG] and will be mapped, wi h i , on o he slo . Since
all [ARG] ea u es will be mapped on o ha slo , [plu al] will also be always mapped,
ega dless o he cli ic i comes om. [plu al] will be spelled ou as
/z/
h ough ule (15e). Two
exa nples appea in
(20),
below:

a. CL b. CL
A
CL
A
n
CL
n
ARG
3RD
ARG
3RD
ARG
3RD
ARG
3m
--
[ARG] [OBL]
--
[ARG] [OBL]
Resul ing ou pu o m: Ilzil Resul ing ou pu o m: llzil
A plu al ma ke 121 will, hen, show up whene e a leas one o he wo cli ics is plu al
No ice ha he empla e in
(18)
has no slo o [3RD]. This means ha his ea u e will ne e
be
mapped. Because o he se ing (16a') he ea u e [ em], depending on [3RD], will hen no be
mapped; he e o e, i will ne e su ace.10 This is illus a ed in (21):
l0
1n Bone (1991). he absence o a eminine ma ke in combina ions o wo hi d pe son cli ics is amun ed o
h ough he applica ion o a mo phological de, ins ead o a pa ame e wi h wo se ings wi h espec o he
loca ion o he ea u e [ em]. The mo phological de esponsible o he absence o [ em] is shown in (i). [ em]
always depends on [ARG]:
(i)
CL
I
CL
I
ARG
I
3RD
+
[ em1
The de in (i) accoun s o he ac ha he eminine ma ke ne e shows up, no only wi h hi d pe son cli ics,
bu wi h all kinds o cli ics wi h he ea u e [3RD]. This also includes he geni i e o pa i i e cli ic and he
loca i e cli ic. Mo eo e , his ule also ensu es ha he ea u e [ em] is no los when he o he cli ic is i s
pe son, second pe son o he impe sonal o e lexi e; hese cl~ ics do no ha e he ea u e [3RD]. The p oposal
p esen ed in his pape is, in p inciple, be e han he al ema i e ske ched in his oo no e because i emiches he
ep esen a ions ins ead o pu ing he bu den on ules. Howe e i is no clea a his poin whe he he
ep esen a ional op ion
can
amun
as
well o he la e ac no ed he e: [ em] would always be a dependen on
[3RD], and would ne e ha e a spell-ou because he e would ne e be a slo o i in he empla e. The e o e
[ em] would also be w ongly los in combina ions wi h non-[3RD] cli ics, like i s pe son cli ics. This p oblem
migh
be
sol ed by ha ing a sligh ly di e en no ion o empla e. I lea e his issue o u he esea ch.
CL CL
n
ARG
3RD
n
y/>L
-
-
[ARGI [OBLI
Resul ing ou pu o m: Ili1
When he cli ics appea in isola ion, he empla e does no hold and all e minal ea u es a e
spelled ou . In his a ie y, [ em] is always -whene e licensed- spelled ou as
la1
h ough ule
(1%).
2.2.
Va ie yA
While a ie y
D
has he mos opaque hi d pe son p onominal cli ic combina ions, a ie y A has
only anspa en ou pu s. In o de o de i e hese anspa en o ms, he mapping o he cli ics
on o he empla e has o be based on whole cli ics (no on e minal ea u es), as shown in
(16b'). [ emenine], like [plu al], will
be
a dependen on [ARG] (c . (16a)). The empla e o
his a ie y appea s in (22):
(22)
[ARG OBL] [ARG 3RD]
(22) men ions only he ele an e minal ea u es o he cli ics in ol ed, unde each slo .
I
has
o be unde s ood ha he minimal cons i uen con aining he ea u es men ioned
(CL)
is
mapped. Thus, he i s slo will
be
occupied by he hi d pe son da i e cli ic (which has bo h
(uno de ed) [ARG] and [OBL]), while he accusa i e cli ic (wi h [ARG] and
[3RD])
will
occupy he second slo ( he da i e cli ic p ecedes he accusa i e cli ic in his a ie y).
(23), below, illus a es how he di e en hi d pe son cli ics a e mapped on o he empla e.
Recall, again, ha he wo cli ics could ha e been w i en in he opposi e o de ( he accusa i e
cli ic be o e he da i e cli ic) because he e is no linea o de a ha s age. The empla e
(unde nea h he cli ics) is esponsible o he linea o de .
CL CL
A
ARG 3RD
A
ARG 3RD
I
I
I
([PI]) OBL ([ em]) ([PI])
,
-
/
[ARG OBL] [ARG 3RDl
Resul ing ou pu o ns: Ili+ll, /li+la/, /li+lz/, Ili+lezl,
11z+11, Ilz+lal, /Iz+lzl, Ilz+lezi
The da i e cli ic, which is mapped on o he i s slo , will be spelled ou as lIl+li/, i his cli ic is
singula ( h ough he applica ion o ules (15b,c)), and as 111+12/ i he cli ic is plu al ( h ough
ules (15a,e)). In he la e case, no li1 will appea because in he con ex o [plu al], [ARG] and
[OBL] a e spelled ou oge he as 111. The accusa i e cli ic will be spelled ou as 111 i he cli ic is
masculine and singula ( ule (15b)),
as
/I/+/a/ i [ eminine] is p esen ( ules (15b, 1)), and as
/I/+/el+lz/ i bo h [ eminine] and [plu al] a e p esen ( ules (15b, ,e)). Rule (15 ) will apply
ins ead o (l5 ') because o he p esence o [plu al].ll
11 As shown in
(23).
he ou pu o n o he da i e singula cli ic in a ie y
A
is Ilil. 111 is he spell-ou o
[ARG]
and li/, he spell-ou o
[OBL]
(c .(lSb,c)) Howe e , no hing in he mechanisms p oposed p edic s he
o de be ween
[ARG]
and
[OBL]
in his a ie y, since he mapping on o he empla e picks up whole cli ics, and
he ea u es in e na1 o a cli ic s uc u e a e uno de ed. One could p obably assume ha in his a ie y
[ARG]
and
[OBL]
do no ha e independen spell-ou s (as shown in (15b,c)), bu ha he
{[ARG], [OBL])
cli ic has a
speci ic spell-ou Ilil. The ac ha his a ie y lacks an independen
[OBL]
cli ic ( he e is no loca i e cli ic)
makes his sugges ion ai ly plausible.
I
lea e his issue o u he esea ch.
2.3.
Va ie y
B
This a ie y is basically iden ical o a ie y A, bu only when he hi d pe son da i e cli ic is
plu al. Unde hese ci cums ances, he combina ion o wo hi d pe son cli ics is no opaque,
and he da i e cli ic p ecedes he accusa i e cli ic. The analysis o such combina ions will, hus,
be basically iden ical o he one p oposed o a ie y A in sec ion
2.1,
and, he e o e, he
co esponding empla e will be he one shown in (24), based on whole cli ics. When he da i e
cli ic is singula , a ie y B becomes e y simila o a ie y
D,
wi h only opaque ou pu s. The
empla e o a ie y B will, hen, be iden ical o he empla e p oposed o a ie y
D,
based on
e minal ea u es, no on whole cli ics. This empla e is epea ed in (24b), below:
(24) a.
-
b.
-
-
[pl OBL] [ARG 3RD] [ARG] [OBL]
(24a) will be in e p e ed in he ollowing way:
A
cli ic wi h he e minal ea u es [ARG] and
[3RD], ha is a hi d pe son accusa i e cli ic, will be p eceded by a cli ic wi h he e minal
ea u es
[pl]
and [OBL]. The only cli ic wi h hese e minal ea u es is he hi d pe son da i e
plu al cli ic. The Elsewhe e Condi ion, in oked ea lie , de e mines ha he applica ion o (24)
will always p ecede he applica ion o (24b), because (24a) is mo e speci ic (mo e complex)
han (24b). This condi ion is necessa y because o he wise we could gene a e ou pu o ms like
*lzi
ins ead o he exis en
els el
in combina ions o a hi d pe son accusa i e singula
masculine cli ic wi h a hi d pe son da i e plu al cli ic (which is wha happens in a ie y
D).l2
12
Since in a ie y
B
he da i e singula cli ic, lli1 in isola ion, adop s he o n li/, spelled
hi,
iden ical o he
o n o he loca i e cli ic, one wuld sugges ha he da i e cli ic is eally a loca i e cli ic o s a wi h, in he
syn ax. This would aiso
be
he case o o he a ie ies. This possibili y, plausible in p inciple, aces some
p oblems, hough. One o hem, concep ual, would
be
ha , in iguingly. he loca i e cli ic would only be
gene a ed in case he sou ce we e singula . no plu al. The e is also
an
empi ical p oblem: i is impossible o
ha e wo (independen ) loca i e ph ases in a gi en sen ence,
as
shown in (i):
(i)
*Ani a
a Reus a I'es ació.
go-3sg-FUT o Reus o he-s a ion
'Slhe will go o Reus o he s a ion.'
In a ie y
B,
he combina ions wi h a hi d pe son da i e singula cli ic a e iden ical o hose in
a ie y
D
excep in one espec : in a ie y
B,
[ eminine] is ne e los ; i su aces as
a
schwa
( h ough ule (15g)). In his a ie y, hen, [ eminine] will be a dependen on [ARG], like
[plu al] is (c . he pa ame e se ing in (16a)), ins ead o being
a
dependen on [3RD].
(25),
below, shows how he mapping wo ks o a combina ion o a hi d pe son accusa i e cli ic and
a hi d pe son da i e singula cli ic:
I he da i e cli ic is conside ed o be eally a loca i e cli ic, we p edic i o
be
impossible ha ing a hi d pe son
da i e cli ic and a loca i e cli ic in a gi en sen ence. This p edic ion is no bome ou , as (ii) shows:
(ii) En Pe e no
li
hi
(Ilil) ani 8, a Reus
he Pe e no 3 d-da (ben) loc go-3sg-FLIT o Reus
'Pe e will no go o Reus o himlhe .'
Unde he p esen p oposal he ac in (ii) is accoun ed o in exac ly he same ems as he cases wi h wo
ins ances o
[ARG],
seen so a : bo h he da i e cli ic and he loca i e ha e a ea u e
[OBL];
he wo ins ances o
his ea u e a e mapped on o one and he same slo , and, la e , ha ea u e is spelled ou as a single li1 h ough
de
(1%).
The in ui ion ha he da i e cli ic and he loca i e cli ic a e in some way he same is also cap u ed
unde he p esen p oposal: bo h cli ics ha e he ea u e
[OBL];
he only di e ence is he p esence e sus he
absence o he ea u e
[ARG].
The s uc u e o hese wo cli ics
can
be
compa ed in (iii) ((iiia) co esponds o he
hi d pe son da i e cli ic, while (iiib) co esponds o he loca i e cli ic):
(ui) a.
CL
A
b.
CL
ARG
3RD
I
3RD
I
OBL
I
OBL

Once he mo phological ule in (27a) has applied, he ou pu o m and he s uc u e
co esponding o he da i e plu al cli ic a e mapped on o a empla e based on whole cli ics (c .
(16b')). The da i e cli ic ([pl OBL]) will p ecede he neu e cli ic
([NEUT
3RD1). When he
hi d pe son da i e cli ic is singula , he ou pu o ms do no di e a all om a ie y
D:
he
o m Ili1 appea s whene e he wo cli ics in ol ed a e singula ; Ilzil appea s, ins ead, when he
accusa i e cli ic is plu al ( ecall ha in his case he da i e cli ic is singula ). No eminine ma ke
appea s. Unde he p esen p oposal, he lack o a eminine ma ke indica es ha he ea u e
[ eminine] depends on [3RD], like in a ie y
D.
Likewise, he mapping o
a
empla e will be
based on ea u es (c . (16b)). The wo empla es alid o a ie y
C
appea in
(28),
below:
(28) a.
-
b.
-
-
[pl OBL]
[NEUT
3RDl [ARG] [OBL]
Va ie y
C
cons i u es, hen, a mixed sys em wi h espec o mapping o a empla e, like a ie y
B: o ce ain combina ions he mapping is based on whole cli ics (in bo h a ie ies, his
looses
i s
ea u e
[NEUT]
when i woccu s wi h any
[3RD]
cli ic,
as
shown below:
I
CL
ARG
3RD
I
3RD
In a ie y
C,
de (i) has o p ecede de (27a). O he wise, he ea u e
[NEW]
in oduced by de (i) would be los
by he applica ion o ule (27a). The o de ing be ween hese wo ules could ollow om some s uc u e
p ese~a ion wndi ion.
Jim Ha is @.c.) sugges s ha Iol in his a ie y wuld be in e p e ed as a class ma ke (as he one ha appea s in
ca o), a he han he spell-ou o he ea u e
[NEUT].
This is ce ainly a possibili y wo h explo ing, e en
hough i would be mo e di icul o ule ou he o n */lzozl as he spell-ou o a wmbina ion wi h an
accusa i e plu al cli ic.
happens whene e a hi d pe son da i e plu al cli ic is in ol ed); o o he s, he mapping is
based on ea u es.14 The mapping o whole cli ics is shown in (29a), while he mapping o
ea u es appea s in (29b):
a. CL
2.
b. CL CL
A
A
A
ARG
3RD
ARG
3RD
ARG
3RD
ARG
3RD
I
I
I
,
[pll OBL,
,
NEüT
OBL
([pi]) ([ em])
[pl OBL]
PEüT
3RD]
[ARG] [OBL]
In (29a) he neu e cli ic ha appea s on he igh s~de (i could equally ha e been w ~ en on he
l4
We ha e seen ha when wo ea u es " igh " o a gi en slo in he empla e, no con iic anses and bo h o
hem a e mapped. La e on each such slo will ecei e a single spell-ou , which will be in e p e ed as ha ing wo
di e en sou ces. Among he combina ions aken in o conside a ion in his pape , he e a e no cases o wo
whole cli ics ha compe e o a single slo . These cases do exis , he po en ial combina ion o wo ins ances o
es
being an example o hem. Fi s and second pe son cli ics and he cli ic
es
(all non-[3RD] cli ics) a e mapped
as whole cli ics in all dialec s o Ca alan. Two ins ances o
es
could in p inciple cooccu i one o hem we e an
impe sonal cli ic and he o he one we e a e lexi e hi d pe son cli ic. Howe e , hese combina ions do no ha e
a g amma ical ou pu , as
can
be seen in (i):
(i) a. *En aques a casa
es es
en a.
in his house e l+imp washes
b.
*En
aques a casa
es
en a.
in his house e l+imp washes
'In his house, one washes onesel .'
As (ia) shows. wo ins ances o
es
canno cooccu ; in (ib) we see ha a single ins ance o
es
canno ep esen
bo h a e lexi e cli ic and an impe sonal cli ic. The ac s in (i), compa ed o he ac s aken in o accoun in he
ex wi h espec o he compe l lon o a gi en slo , lead us o conclude ha wo whole cli lcs
can
ne e be
mapped on o a single slo , whlle h~s
IS
no he case w~ h ln& ldual ea u es '%s cond~hon would amun o
he ln egn y (o eco e abill y, in ems aken om Bone (1991)) ha nou-[3RDl d l~s seem o shoW
(Masca ó
(1986 135-6) p esen s a mo e complica ed plc u e o he ac s ela
cl1 lcs)
le side) is he accusa i e cli ic a e he applica ion o he mo phological ule in (27a). The i s
slo in he co esponding empla e can only be occupied by hi d pe son da i e plu al cli ics,
which is he case in (29a). In (29b), he ea u e [ eminine] is no mapped on o he empla e
because he e is no slo o [3RD] and, he e o e, i canno be spelled ou .
3.
Conclusions
In his pape , ou di e en a ie ies o Ca alan ha e been analyzed wi h espec o hei
di e en ou pu s in combina ions o wo hi d pe son cli ics. I has been shown ha bo h linea
o de and he p esence o opaque o ms ollow, excep o one single aspec , om he way he
mapping on o a empla e wo ks. The excep ion is he change om an accusa i e cli ic o a
neu e cli ic in one o he a ie ies. This ype o opaque o m, a om being an excep ion, is
e y common in all Romance languages, and has been a gued o be de i ed om he applica ion
o a mo phological ule. These ules, by hei own na u e, canno gene a e andom ou pu s, bu
gi e as a esul a s uc u e co esponding o an independen ly exis ing cli ic. The esul s
p o ided by he p oposal made in his pape could ha dly be ob ained i a syn ac ic app oach
we e adop ed, which would, a mos , p o ide an accoun o linea o de -no o he p esence o
opaque o ms- by making obliga o y use, a leas o a ie ies
B
and C, o ce ain s ipula ions
(because o he exis ence o a di e en o de ing depending on he numbe o he da i e cli ic). I
has o
be
aken in o accoun , also, ha he amoun o dialec al and language a ia ion wi h
espec o p onominal cli ics and hei combina ion is ex ao dina y, while a ia ion in o he
aspec s is much smalle . Unde a minimalis iew o he g amma i does no make much sense
o y o accoun o all hese di e ences in syn ac ic ems. A pos -syn ac ic mo phological
app oach o his a ia ion seems mo e plausible.
I
would like o hank Jim Ha is, Joan Masca 6, an anonymous e iewe , and he audience a
he 2nd Wo kshop on he Syn ax o Cen al Romance Languages o insigh ul commen s and
help. All e o s a e my own.
110
Re e ences
Ande son, S.
R.
(1992)
A-Mo phous Mo phology,
Camb idge Uni e si y P ess (Camb idge
S udies in Linguis ics, 62), Camb idge.
Bone ,
E.
(1991)
Mo phology ape Syn ax: P onominal Cli ics in Romance,
doc o al
disse a ion, Massachuse s Ins i u e o Technology, dis ibu ed by MIT Wo king Pape s
in Linguis ics, Camb idge,
MA.
Chomsky,
N.
A. (1993) 'A Minimalis App oach o Linguis ic Theo y' in K. Hale and S.
J.
Keyse (eds)
The View om Building 20,
MIT P ess, Camb idge,
MA.
Halle, M. (1990) 'An App oach o Mo hology',
P oceedings o he No heas e n Linguis ics
Socie y,
20, 150- 184.
Halle, M. (1991) The La ian Declension' in G. Booij and
J.
an Ma le (eds)
Yea book o
Mo phology 1991,
Kluwe , Do d ech , 33-47.
Halle, M. and A. Ma an z (1992) 'Dis ibu ed Mo phology and he Pieces o In lec ion' in K.
Hale and S.
J.
Keyse (eds)
The View om Building
20,
MIT P ess, Camb idge,
MA.
Hanis,
J.
(1991) 'The Exponence o Gende in Spanish',
Linguis ic Inqui y,
22, 27-62.
Kipa sky,
P.
(1982) 'Lexical Mo phology and Phonology' in I.-S. Yang (ed)
Linguis ics in he
Mo ning Cal m,
Hanshin, Seoul, 3-91.
O e o, C. (1976) 'The Dic iona y in a Gene a i e G amma ', pape p esen ed a he 91s .
Annual Con en ion o he MLA, New Yo k.
P anka, P. (1983)
Syn ax and Wo d Fo ma ion,
doc o al disse a ion, Massachuse s Ins i u e
o Technology.
Todoli,
J.
(1992) 'Va ian s dels p onoms ebles de 3a pe sona al País Valencia: egles
onosin Ac iques i mo olbgiques subjacen s',
Zei sch i jii Ka alanis ik,
5,
137-160.
Depa amen
de
Filologia Ca alana
Uni e si a Au bnoma de Ba celona
E-O8193 Bella e a
e- nail: [email p o ec ed]