Unde s anding Pa iali y in p o-D op Languages: An
In o ma ion-S uc u e App oach
Ma a F asca elli and

Angel L. Jim
enez-Fe n
andez
1
Abs ac . This a icle explo es he accep abili y and in e p e a ion o e e en ial null subjec s in
se e al a ie ies o Spanish (bo h Peninsula and Ame ican a ie ies) in di e en syn ac ic
con ex s (ma ix and embedded) and dis inc clause ypes (unde b idge e bs, unde ac i e
e bs, and in ad e bial clauses). Based on he esul s o an o iginal online su ey, comple ed by
almos 300 esponden s all o e he wo ld, i is shown ha , in con as wi h a consis en -p o-
d op language like I alian, some Spanish a ie ies exhibi pa ial-p o-d op p ope ies; his
con ibu es o a ecen line o esea ch dealing wi h pa iali y in consis en -p o-d op languages.
We a gue ha such a ia ion can be accoun ed o wi hin an in o ma ion-s uc u e app oach, in
which he Topic C i e ion (F asca elli 2007), he o ma ion o opic chains, and he exis ence o
silen opics play a c ucial ole. Deg ees o pa iali y a e a ibu ed o an in e ace condi ion ha
combines in o ma ion-s uc u e equi emen s wi h PF isibili y o o e copies in opic chains.
Pa ial-p o-d op p ope ies can hen be explained by e e ence o independen syn ac ic
condi ions, such as he p e e ence o o e minimal links and he sensi i i y o islands.
1. In oduc ion
In i s o iginal o mula ion (Pe lmu e 1971), he p o-d op pa ame e was aimed a
cap u ing he empi ical obse a ion ha in some languages a defini e, e e en ial,
p onominal subjec in a fini e clause mus be exp essed o e ly.
1
This obse a ion was
aken up, ex ended, and ewo ked by Chomsky (1982) in he o m o he Ex ended-
P ojec ion P inciple (EPP), a basic ene in gene a i e g amma . Since hen, many
wo ks ha e been dedica ed o he p o-d op pa ame e , defining he o mal p ope ies
ha de e mine he se ing o i s di e en op ions, which yield consis en -, pa ial-,
adical-, semi-, and non-p o-d op languages ( o discussion and e e ences, see
Bibe aue e al. 2010, Cognola & Casalicchio 2018).
Since he seminal wo k ca ied ou by Jaeggli (1982) and Rizzi (1986), he null-
subjec pa ame e has been claimed o be dependen on he / ea u es ha a e specified
o encoded in he ele an licensing head (i.e., he Inflec ional head node). C i icism o
his app oach was soon elabo a ed (see Jaeggli & Safi 1989, Huang 1989), and an
al e na i e analysis has ecen ly been p oposed in he Minimalis amewo k by
Ea lie e sions o his a icle we e p esen ed a he ou h Camb idge Compa a i e Syn ax con e ence in
he UK, a he 26 h Colloquium on Gene a i e G amma in Spain, and a he se en h Encuen o de
G am
a ica Gene a i a in A gen ina. We hank all he audiences o hei encou aging commen s. We a e
especially g a e ul o he Syn ax e iewe s o hei challenges, which defini ely ha e con ibu ed o he
imp o emen o he a icle. The esea ch he e has been pa ially unded by esea ch p ojec FFI2013-41509-
P o Spain’s Minis y o Economy and En e p ise (MINECO), by esea ch p ojec PGC2018-093774-B-I00
o Spain’s Minis y o Science, Inno a ion, and Uni e si ies (MICINN), and by esea ch p ojec 2018/
00000491 o he se en h Plan P opio o he Uni e sidad de Se illa. The o de o au ho s is alphabe ical.
1
In his wo k we concen a e on null subjec s ha a e e e en ial; o he ypes, such as a bi a y and
exple i e null subjec s, a e le o one side.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
S Y N T 12184
Dispa ch: 6.7.19 CE: Wiley
Jou nal Code Manusc ip No.
No. o pages: 37 PE: Rajaseka an S.
Syn ax 2019 DOI: 10.1111/syn .12184
Holmbe g (2005), acco ding o which he licensing o a null subjec (he ea e , NS)
depends on he p esence o a D ea u e in T. (See also Camacho 2013, 2016, Saab 2016,
and Ticio 2018 on he occu ence o his D ea u e in T in di e en Spanish a ie ies.)
F om a di e en pe spec i e, F asca elli (2007) shi s he ocus o analysis om
licensing o he in e p e a ion o licensed NSs. Assuming wi h Holmbe g ha he
licensing mechanism equi es ma ching be ween in e p e able and unin e p e able
ea u es in Ag
0
, F asca elli concen a es on how a licensed p o can be co ec ly gi en
a e e en ial index and in e p e ed as in ended in he discou se. As a ma e o ac , /
ea u es only p o ide unc ional in o ma ion abou he pe son, numbe , and gende o
ap o; hey do no explain how i s discou se linking (D linking) is es ablished in
syn ac ic e ms in a consis en -p o-d op language.
Based on in e ace analysis o na u alis ic spoken co po a, F asca elli a gues o an
in o ma ion-s uc u e s a egy. Specifically, she p o ides e idence ha he in e p e a ion
o a e e en ial NS depends on a ma ching ela ion (“Ag ee”in Minimalis e ms)be ween i
and a specific ypeo opic. Based on I alian (spon aneous) da a, she iden ifies he la e as
he abou ness-shi opic (A-Topic) o F asca elli & Hin e h€
olzl 2007 (see sec ion 2.1 o
de ails, and see Jim
enez-Fe n
andez 2016 o a simila analysis based on Spanish da a). The
A-Topic combines he [abou ness] ea u e (which F asca elli p oposes as an “ex ended
EPP ea u e”) wi h he [shi ] ea u e, which is he p ope y ha A-Topics ha e o mo ing
he con e sa ion om one opic o ano he , upda ing he discou se con ex (see Bianchi &
F asca elli 2010). A-Topics a e p esen in e e y p edica i e sen ence o es ablish “wha he
sen ence is abou ”and, when o e ly ealized, o shi he opic ac oss sen ences. F asca elli
pu s o h a Topic C i e ion ha co ela es co e g amma wi h discou se equi emen s:
(1) Topic C i e ion
a. [+abou ness] is connec ed wi h an (ex ended) EPP ea u e in he high opic
field ha yields a specific discou se- ela ed p ope y, namely “abou ness.”
b. The [+abou ness] opic ma ches wi h an a gumen in he main clause
h ough Ag ee.
c. When con inuous, he A-Topic can be null (i.e., silen ).
The Topic C i e ion implies ha e e y sen ence con ains a posi ion in i s C domain
endowed wi h he [abou ness] ea u e. Howe e , since a shi is no ealized in e e y
sen ence, i is c ucial o assume ha , wi hin discou se, p edica ion can imply a
mul iclausal domain in which chains o clauses a e combined and e e o an
es ablished A-Topic, whose [abou ness] ea u e is kep con inuous.
Topic con inui y is enabled ac oss sen ences by he p esence o low- oned copies (G-
Topics, p onouns) and silen copies (null opics, NSs) o he es ablished A-Topic, heading
aso-called opic chain (see also F asca elli 2018). In his pic u e, he es ablished A-Topic
is kep silen un il a new A-Topic is p oposed (b eaking he p e ious opic chain).
2
2
A opic-o ien ed app oach o NSs can also be ound in p e ious wo ks (among o he s, Gi 
on 1983 and
mo e ecen ly Adli 2011, Camacho 2013, and Pe
sko 
a 2014). The c ucial di e ence in he p esen app oach
is he ac ha he ela ion be ween NS and opic is no “in e ed om he con ex ”bu is pa o discou se
g amma and, as such, implies specific o mal equi emen s.
2Ma a F asca elli and

Angel L. Jim
enez-Fe n
andez
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
Finally—gi en he (empi ically based) assump ions ha he in e p e a ion o an NS
depends on an Ag ee ela ion wi h he cu en A-Topic (c . (1b); his ela ion is local,
ollowing Minimalis ene s) and ha A-Topics a e only associa ed wi h clauses
endowed wi h illocu iona y o ce (c . he In e ace Roo Res ic ion in sec ion 2.3)—
he ollowing condi ion mus be added o he Topic C i e ion.
(2) Topic-Chain Condi ion (F asca elli 2018:(19))
a. An A- opic chain can only be s a ed om a oo (o oo -like) C domain.
b. The A-Topic heading he opic chain can be silen .
Indeed, he ele ance o silen ca ego ies is a majo issue in ecen analyses, and
subs an ial e idence has been p o ided o es ablish ha , echoing Kayne’s (2016:1)
wo ds, “many mo e elemen s in he sen en ial p ojec ion line (and elsewhe e) mus be
aken o be silen han is usually hough .”In his espec , see also Sigu ðsson 2004.
The cen al ole o opics in discou se is widely discussed and acknowledged in he
li e a u e (see, among many o he s, Ku oda 1965, Gi 
on 1983, Lamb ech 1994).
Topics ha e di e en unc ions, howe e . In his pic u e, i is impo an o emphasize
ha F asca elli’s (2007) p oposal (which we adop ) ea s opics as an en i y, one ha
is endowed wi h he discou se-seman ic p ope y o abou ness ( ollowing Reinha
1981), and his p ope y is cap u ed by a o mal ea u e as an ex ended EPP
equi emen . Fu he mo e, al hough F asca elli 2007 ocuses on consis en -NS
languages, F asca elli 2018 p o ides e idence ha he Topic C i e ion (1) can be
conside ed a mac opa ame e ha also accoun s o pa ial-NS languages (see
sec ion 3.2).
As a ma e o ac , he e is a gene al consensus in he ele an li e a u e ha
an NS language o he I alian, Spanish, and G eek ype may change g adually
in o a non-NS language like English and F ench (see Holmbe g 2010, Saab
2016). Based on a numbe o p ope ies, such as he p e e ence o p e e bal
subjec s o e pos e bal subjec s, he o e use o p onominal subjec s, and he
impac o c-command in he in e p e a ion o NSs, he e a e languages ha ha e
been claimed o be changing in o pa ial-NS languages, including Po uguese
and, o a lesse ex en , Spanish (see Ka o 2000, Modes o 2000, Rod igues 2004,
and Ba bosa, Dua e & Ka o 2005, which deal wi h B azilian Po uguese, and
Com
ınguez 2018 and Camacho 2016, which deal wi h Ca ibbean a ie ies o
Spanish). Ne e heless, he discussion on his opic is s ill li ely and open
(especially o Ca ibbean Spanish: see Camacho 2013, 2016, Com
ınguez 2018,
Ticio 2018; om a di e en pe spec i e, see Posio 2012, aking in o accoun
usage-based sociolinguis ics).
In he p esen wo k we con ibu e o his ho opic by discussing he syn ac ic and
in e p e a i e p ope ies o NSs (and, o a lesse ex en , o e p onominal subjec s) in
wo consis en -NS languages, I alian and Spanish. We p esen expe imen al esul s
om se e al Spanish a ie ies, and we examine whe he hey suppo he claim ha
(i) some Spanish a ie ies (no only Ca ibbean Spanish) a e becoming pa ial-NS
languages, like B azilian Po uguese, o ha (ii) Spanish a ie ies a e s ill consis en -
Unde s anding Pa iali y in p o-D op Languages 3
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
NS languages bu some o hem, e en so, show pa ial-NS p ope ies, suppo ing he
idea ha pa iali y in he cha ac e o languages is a ma e o deg ees.
3
F asca elli (2007) discusses his ques ion and p o ides mul iple a gumen s o he
ad an ages o an in o ma ion-s uc u e analysis. Among o he hings, she p o ides
e idence ha NSs canno be ea ed on a pa wi h o e p onouns in an NS language
like I alian, and she p o ides da a in ol ing quan ified DPs ha clea ly indica e he
necessi y o a o mal heo y ha includes some abs ac ion, such as null opics. In his
espec , conside o ins ance he I alian and Spanish scope ac s in (3)–(5).
(3) Ogni angolo della banca e a so eglia o da una gua dia. ∀,∃
Cada esquina del banco es aba igilado po un polic
ıa.
each co ne o . he bank be.PST.3SG gua ded by a police.o fice
‘E e y co ne o he bank was gua ded by a police o fice .’
As indica ed, he DP una gua dia/un polic
ıa‘a police o fice ’can ha e a na ow- o
wide-scope eading. Howe e , i we con inue (3) wi h a sen ence like (4), only an
exis en ial eading is a ailable.
(4) p o uma a come un ma o!*∀,∃
¡p o umaba como un loco!
smoke.PST.3SG like a mad
‘He smoked like a luna ic!’
We ake his o be e idence ha he NS in (4) does no ake as i s an eceden he DP
una gua dia in (3) bu a he a silen opic in i s own local C domain. This null opic is
specificbydefini ion and only allows o an exis en ial eading:
4
(5) ... so eglia o da una gua dia. [〈La/ques a gua dia
k
〉[p o
k
... gua ded by a police.o fice he/ his police.o fice
uma a come un ma o]].
smoke.PST.3SG like a mad
As we will see, silen opics will play an impo an ole in ou o maliza ion o he
in e p e a ion o NSs.
The a icle is o ganized as ollows. In sec ion 2 we p esen he ypology o opics
ha is pu o h in F asca elli & Hin e h€
olzl 2007, discuss he ole o he A-Topic in
con e sa ional dynamics, and illus a e i s unc ion and ele ance a he syn ax–
3
We belie e ha ou in o ma ion-s uc u e app oach o he encoding o D linking has ad an ages o e a
p agma ic app oach o anapho a esolu ion. A p agma ic heo y, claiming ha co e e en eadings a e
p e e ed unless a disjoin eading is implica ed (based on he G icean maxim o quan i y; see Le inson
1991), canno ully accoun o he complexi y o his phenomenon; an in eg a ed app oach is equi ed. Fo
discussion, see Van de Sand 1992.
4
Addi ional e idence o he specific in e p e a ion o he null opic in hese sen ence ypes comes om
he ac ha , whene e he opic is gi en an explici spellou , i mus be specific. This explains why only a
defini e DP like la/ques a gua dia ‘ he/ his police o fice ’can be o e ly ealized in he C domain o (3)–(5)
(see also Jim
enez-Fe n
andez 2016 o Spanish).
4Ma a F asca elli and

Angel L. Jim
enez-Fe n
andez
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
discou se in e ace as he elemen heading opic chains. In sec ion 3, we fi s p o ide
some backg ound on he di e ences be ween consis en - and pa ial-NS p ope ies
and hen in oduce ou expe imen and he ele an me hodology. In sec ion 4 he
expe imen al da a a e p esen ed and discussed in de ail. Based on he esul s ob ained,
sec ion 5 gi es an in o ma ion-s uc u e analysis o he da a and de ends he p esen
p oposal, aiming a a comp ehensi e explana ion o he c osslinguis ic a ia ion
a es ed.
2. Di e en Types o Topics, A-Topic Chains, and he A oid-P onoun P inciple
Based on an in e ace analysis o na u alis ic spoken co po a om I alian and
Ge man, F asca elli & Hin e h€
olzl (2007) p o ide subs an ial e idence o he
exis ence o a sys ema ic co ela ion be ween he o mal p ope ies o (di e en ypes
o ) opics and hei unc ions in discou se. Specifically, hey dis inguish be ween he
abou ness-shi opic (A-Topic), he con as i e opic (C-Topic), and he amilia /
gi en opic (G-Topic). The dis inc unc ion o each is encoded in a dedica ed posi ion
in he C domain, in line wi h a ca og aphic and Minimalis app oach o g amma (see
Rizzi 1997 and subsequen wo ks):
5
(6) [
Fo ceP
[
Shi P
[
Con P
[
FocP
[
FamP
[
FinP
[
IP
Fo he pu poses o he p esen analysis, we will specifically deal wi h A-Topics and
G-Topics, which a e b iefly desc ibed in he ollowing wo subsec ions.
2.1. A-Topics
As p e iewed in sec ion 1, he A-Topic connec s Reinha ’s (1981) abou ness wi h
he p ope y o being newly in oduced o ein oduced and changed o—hence, wi h
a shi ing unc ion in he discou se. In a ca og aphic app oach, in which discou se
p ope ies a e encoded in syn ax in he o m o o mal ( unc ional) ea u es, he Shi
0
head is hus assumed o be endowed wi h a specifica ion o in e p e able [abou ness]
and [shi ] ea u es. Since he [abou ness] ea u e is p oposed as an ex ended EPP
ea u e (see he Topic C i e ion in (1)), Spec,Shi P is a c i e ial posi ion (Rizzi 2006)
and mus be filled in e e y p edica ional sen ence.
Though his no ion migh seem eminiscen o Lamb ech ’s (1994) sen ence opic,
he cen al di e ence is ha , adop ing Reinha ’s app oach, he A-Topic is an en i y
wi h a specific ole in con e sa ional dynamics: an upda ing e ec leading o a new
con ex . Following his line o analysis, he discou se con ex includes a s uc u ed
eposi o y o sha ed in o ma ion (p oposi ions publicly accep ed by he pa icipan s),
and he discou se e e en deno ed by he A-Topic cons i u es he “add ess”(o file
5
The ca og aphic app oach, in which ea u es a e assumed o be me ged in dedica ed p ojec ions, is
ully compa ible wi h he Minimalis P og am, hough i may no seem ha way a fi s sigh . As Cinque &
Rizzi (2008) no e, we can ake he wo app oaches o ha e a di e en ocus. Minimalism ocuses on he
elemen a y mechanisms in ol ed in syn ac ic compu a ion, while ca og aphy deals wi h he de ails o he
syn ac ic s uc u e; hey a e no mu ually exclusi e (Mao & Meng 2016, Tanaka 2016).
Unde s anding Pa iali y in p o-D op Languages 5
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d

ca d, in he e ms o Heim 1982) unde which he p oposi ion con eyed by he
asse ion will be s o ed. This con e sa ional mo e upda es he in e locu o ’s common
g ound, in which p oposi ions a e s o ed unde defining en ies.
Syn ac ically, he A-Topic is me ged in he highes opic posi ion in he C domain
(Spec,Shi P), and i canno be i e a ed. As o in ona ion, he A-Topic has been
shown o be associa ed wi h a complex L*+H one in a numbe o ypologically
di e en languages (see F asca elli 2007, F asca elli & Hin e h€
olzl 2007, F asca elli
& Puglielli 2009). Howe e , since he p osodic in e ace is no pa o he p esen
in es iga ion, in ona ional p ope ies will no be u he men ioned.
As an illus a ion o a opic shi , conside he example in (7) ( om co pus da a in
F asca elli 2007), in which a s uden is alking abou compu e skills wi h a iend.
The opic cons i uen (s) unde examina ion a e bold aced.
(7) A: Il compu e
k[A-Topic 1]
u lo
k
a e i gi
a usa o qualche
he compu e you i ha e.PST.2SG al eady used any
ol a?
ime
‘Did you e e use a compu e be o e?’
B: S
ımi 
e capi a o pe 
op o
k
e a MacIn osh.
yes o.me be.3SG happened bu was Mac
‘Yes, I’ e expe ienced ha , bu i was a Mac.’
Ho semp e a u o un app occio di base ...
ha e.1SG always had a app oach o base
‘I only had basic skills ...’
Ecco, il CD-ROM
z[A-Topic 2]
non l
z
’a e o mai
well he CD-ROM no i ha e.PST.1SG ne e
usa o—p o
z

e u ilissimo!
used be.3SG e y.use ul
‘Well, I had ne e used a CD-ROM—i is e y use ul!’
As shown, he opic p oposed by speake A is il compu e (a disloca ed di ec objec
esumed by he cli ic lo), bu hen speake B shi s he opic o a specific ool, namely
il CD-ROM. No e ha he NS ollowing his opic shi is unambiguously ela ed o
he DP il CD-ROM and canno be linked o il compu e . This is e idence ha opic
shi s c ea e new opic chains.
2.2. G-Topics
Qui e di e en ly om A-Topics, he discou se unc ion o G-Topics is he e ie al o
gi en in o ma ion in he discou se. The G-Topic can he e o e be conside ed a D-
linked cons i uen , ei he in a “s ong”sense (Heim 1982) o in a weak/ amilia sense
(Robe s 2003).
Acco ding o F asca elli & Hin e h€
olzl’s (2007) heo y, he e ie al o gi en
in o ma ion can sa is y wo di e en discou se equi emen s. On he one hand, a G-
Topic can appea as pa o a opic chain, and in his case i is a “low copy”o an
6Ma a F asca elli and

Angel L. Jim
enez-Fe n
andez
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
es ablished A-Topic, se ing a con inui y unc ion. On he o he hand, he G-Topic
can e ie e gi en in o ma ion ha is pa o he backg ound bu is no linked o he
cu en A-Topic. The o me ype o G-Topics a e defined as abou ness G-Topics, he
la e as backg ound G-Topics.
Gi en his discou se cha ac e iza ion o G-Topics, F asca elli & Hin e h€
olzl
assume ha he head o FamP is endowed wi h he [abou ness] ea u e, which is
howe e unin e p e able in his posi ion. By Minimalis ene s, he p ope y o (un)
in e p e abili y is he d i ing o ce behind he es ablishmen o a syn ac ic dependency
(Chomsky 1995:277); acco dingly, we assume, as Adge & S enonius (2011:34) do,
ha “unin e p e able ea u es a e hose ha d i e he de i a ion, while he
in e p e able ones a e hose ha a e used, in he final ep esen a ion, o connec
wi h he seman ic sys ems o he phonological ones”(i.e., wi h in e ace le els). The
dis inc ion be ween he wo ypes o G-Topics hus depends on hei syn ac ic
de i a ion. When a G-Topic mo es o Spec,FamP, as illus a ed in (8), i en e s an
Ag ee ela ion wi h Shi
0
and Fam
0
, which a e endowed wi h, espec i ely, he
in e p e able ea u es [shi ; abou ness; e ] and he unin e p e able ea u es [abou -
ness; e ]. We assume ha he NS’s ea u es a e un alued bu in e p e able (a possible
dissocia ion o e ed by Pese sky & To ego 2007). As a consequence o his Ag ee
ela ion, he G-Topic’s in e p e able [abou ness] ea u e is alued in he syn ax and
hen in e p e ed acco dingly a he in e aces. This cons i uen is hus pa o a opic
chain and is in e p e ed as an abou ness G-Topic.
(8) [
Shi P
A-Topic
z
[
Shi
0
Shi
0
[shi ; abou ness; e ]
[
FamP
G-Topic/p o
z[abou ness; e ]
[
Fam
0
Fam
0
[abou ness; e ]
[
TP
z
T[
P
+V]]]]]]
No ice ha , besides [shi ] and [abou ness], he Ag ee ela ion unde examina ion also
allows o he in e p e a ion o he [ e ] o he NS. The Shi head, wi h in e p e able
[abou ness], [shi ], and [ e ] ea u es, hus plays a c ucial ole in in e ace
in e p e a ion by ac ing as a p obe.
In con as , when a G-Topic is no endowed wi h he un alued in e p e able
[abou ness] ea u e, mo emen o Spec,FamP is no igge ed and an Ag ee ela ion
wi h Shi
0
is no ealized. Since FamP is no needed, i does no p ojec . In his case
he G-Topic ge s a “backg ound” ealiza ion and is no pa o a opic chain. To
accoun o he backg ound eading o he G-Topic, we p opose—as an addi ion o
F asca elli & Hin e h€
olzl’s (2007) heo y— ha a backg ound G-Topic is in e p e ed
in Spec,TP (d awing on Jim
enez-Fe n
andez & Miyagawa 2014). This is illus a ed in
(9). The G-Topic does no e e back o he A-Topic.
(9) [
Shi P
A-Topic
z
[
Shi
0
Shi
0
[shi ; abou ness; e ]
[
TP
G-Topic T [
P
+V...]]]]
2.3. Discou se Ca ego ies, Con e sa ional Dynamics, and Topic Chains
Discou se ca ego ies ha e di e en unc ions in con e sa ional dynamics. In his
espec Bianchi & F asca elli (2010:51), adop ing K i ka’s (2007) amewo k o he
common g ound ha dis inguishes be ween common-g ound managemen and
Unde s anding Pa iali y in p o-D op Languages 7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
common-g ound con en , p o ide e idence ha a discou se ca ego y ha igge s an
upda e o he discou se con ex mus occu in a clause endowed wi h con ex -upda e
po en ial. They p opose he ollowing es ic ion.
(10) In e ace Roo Res ic ion (Bianchi & F asca elli 2010:(8))
In o ma ion-s uc u e phenomena ha a ec he con e sa ional dynamics
(common-g ound managemen ) mus occu in clauses endowed wi h
illocu iona y o ce ha implemen a con e sa ional mo e.
E alua ed in e ms o his es ic ion, A-Topics implemen a con e sa ional mo e,
and as such hey belong o he dimension o common-g ound managemen and a e
es ic ed o oo (o oo -like) clauses. By con as , since gi enness is calcula ed on
he basis o he common-g ound con en , G-Topics do no depend on illocu iona y
o ce and can be ound in any ype o clause (ad e bials included).
Finally, emembe ha opic con inui y implies he p esence o silen A-Topics,
whose exis ence is independen ly equi ed by he ex ended EPP (see (1)). Hence,
when no new opic is p oposed in a sequence o discou se- ela ed sen ences, he Spec,
Shi P posi ion is filled by a silen A-Topic, as is he case in he embedded clause in
(11) (angle b acke s indica e silen copies). In his case he embedded Shi
0
is
endowed only wi h [abou ness] and [ e ], and i en e s he Ag ee ela ion discussed in
sec ion 2.2 o cons i uen s mo ing o Spec,FamP.
6
(11) [
Shi P
Leo
k
[
FamP
p o
k
[
TP
k
pensa [
Fo ceP
che [
Shi P
〈Leo
k
〉
Leo hink.3SG ha Leo
[
FamP
p o
k
/lui
k
[
TP
k
comp e 
a una casa]]]]]]].
he buy.FUT.3SG a house
‘Leo
k
hinks ha he
k
will buy a house.’
No ice ha NSs and weak p onouns (o he abou ness kind) sha e a opic-con inui y
unc ion in he discou se and can eely al e na e in a opic chain (in pa ame ic
a ia ion ac oss languages and language a ie ies). In pa icula , he spellou o low-
oned copies can be conside ed a con e sa ional de ice in consis en -NS languages,
plausibly connec ed wi h s yle, p ocessing necessi ies (dis ance, memo y bu den), and
dialec s (mic o a ia ion). Since he ealiza ion o (c)o e subjec p onouns is
associa ed wi h he A oid P onoun condi ion, his heo y also allows o a
e o mula ion o ha condi ion based on in o ma ion-s uc u e conside a ions, as is
p oposed by F asca elli (2007:(32)):
(12) A oid P onoun
A oid using a s ong p onoun, whene e i would ag ee wi h he local
A-Topic.
6
The example in (11) is aken om he se o sen ences used in ou expe imen .
8Ma a F asca elli and

Angel L. Jim
enez-Fe n
andez
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
3. Compa ing Consis en - and Pa ial-NS Languages: Some Backg ound and
he Expe imen
3.1. Some Backg ound on he Di e ences be ween Consis ency and Pa iali y
I is e y di ficul o s a e wha a pa ial-NS language is, since he e is no gene al
consensus on exac ly wha ea u es such a language should show. Di e en pa ial-NS
p ope ies ha e been p oposed in he li e a u e (see e e ences in sec ion 1), wi h
di e en impo ance ac oss languages. In his une en scena io, Holmbe g, Nayudu &
Sheehan’s (2009) p oposal is o en aken as a landma k; on his basis, we can assume
ha a pa ial-NS language is cha ac e ized by (i) sensi i i y o locali y, (ii) sensi i i y
o syn ac ic con ol, (iii) conside able use o explici p onouns, and (i ) a p e e ence
o p e e bal subjec s.
In hei analysis o consis en - and pa ial-NS languages, Holmbe g, Nayudu
& Sheehan concen a e on he fi s wo o hese p ope ies, and hey a gue ha
in pa ial-NS languages a p o is only allowed i con olled (i.e., c-commanded)
by an an eceden in a highe clause.
7
They illus a e his condi ion wi h da a
such as (13), om B azilian Po uguese, a language ha is gene ally
acknowledged o ins an ia e pa ial p o d op ( hough, o a di e en iew, see
Ka o 2011).
(13) O Jo~
ao
i
disse [que os moleques
k
B azilian Po uguese
he Jo~
ao say.PST.3SG ha he child en
acham [que *(ele
i
)
e espe o]].
hink.3PL ha he be.3SG sma
‘Jo~
ao
i
said ha he kids hink he
i
is sma .’(Holmbe g, Nayudu &
Sheehan 2009:82)
They a gue ha pa ial-NS languages do no allow a con ol ela ion ac oss ano he
subjec , e en i he in e ening subjec has incompa ible ea u es wi h he NS. This is
why B azilian Po uguese p e e s he explici p onoun ele ‘he’in (13), so as o comply
wi h he s ic ly-local-con ol p ope y ha cha ac e izes pa ial p o d op (whe eas an
NS is used o ob ain co e e ence wi h an an eceden ha is in e nal o he sen ence
domain whe e he NS occu s).
A di e en ype o c-command-based es ic ion is also p oposed o Spanish
embedded NSs, specifically o dis inguish S anda d Spanish and Ca ibbean
a ie ies o Spanish (see Camacho 2013, 2016, Gup on & Lowman 2013), based
on Mon albe i’s (1986) e ec s on quan ifie binding ha depend on whe he he
p onominal subjec is o e o co e . Conside he example in (14) and he
possible in e p e a ions o null and o e subjec p onouns in he embedded
clause.
7
The au ho s deal wi h NSs in he complemen s o b idge and pe cep ion e bs, as well as in adjunc
clauses and indi ec ques ions. No ice ha he claim ha NSs in B azilian Po uguese a e subjec o a
c-command equi emen is al eady made by Modes o (2000).
Unde s anding Pa iali y in p o-D op Languages 9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
con ol bu on an in o ma ion-s uc u e s a egy. Hence, when no discou se-p osodic
cues a e a ailable o a clea indica ion o he A-Topic, ambigui y domina es (also in
PS a ie ies, con a y o he con as ha Ticio 2018 claims be ween PS and
Ca ibbean a ie ies; on he ambigui ies de i ing om “seman ic p edic abili y”in
opic con inui y, see also Gi 
on 1983).
Le us now conside he case in which an o e p onoun is ealized in he embedded
clause, o see whe he any di e ence eme ges be ween I alian and Spanish:
(22) (Same con ex as (21))
a. Leo ha de o che lui ha comp a o una casa.
Leo ha e.3SG said ha he ha e.3SG bough a house
‘Leo
k
said ha he
k
/he
z
bough a house.’
b. Leo dijo que 
el comp 
o una casa.
Leo say.PST.3SG ha he buy.PST.3SG a house
‘Leo
k
said ha he
k
/he
z
bough a house.’
Table 3 shows ha Spanish o e p onouns a e no discou se-con ex o ien ed (ei he
in PS o in AS) and ha hey p e e he ma ix subjec / opic as an an eceden o e he
embedded subjec (bo h possible A-Topics), whe eas no p e e ence is a es ed in
I alian da a. Indeed, he alues a es ed o ‘Leo’in able 3 a e highe han hose
epo ed in able 2, and he di e ence be ween hese alues is significan o bo h PS
(p=.0141) and AS (p=.0169) a ie ies. Pa icula ly high alues a e a es ed o
‘Leo’in Mex and Ca ib, whose alues wi h espec o able 2 a e significan ly
di e en (p=.0256 and p=.0441, espec i ely). Since he p e e ed op ion is s ill
‘bo h’ o PS and AS (in gene al), we can plausibly conclude ha , when an exclusi e
eading is equi ed, in he absence o in o ma ion-s uc u e cues he ma ix subjec is
in e p e ed as an A-Topic and he de aul discou se in e p e a ion associa ed wi h an
o e p onominal subjec is ha o a amilia en i y ha is pa o a opic chain— ha is
o say, an abou ness G-Topic.
Despi e his p e e ence, ambigui y is s ill he dominan esul . This suppo s he p esen
in o ma ion-s uc u e app oach, acco ding o which he in o ma ion-s uc u e iden ifica-
ion o he A-Topic is necessa y o he in e p e a ion o bo h NSs and o e p onouns (see
he con as be ween NSs and o e p onominal subjec s in able 14 below).
Table 3. O e p onoun embedded unde a ma ix b idge e b
I alian PS Anda Cas Ca Ex AS Ca ib Cen Mex Ri Pe
OK/N 128 173 111 23 35 4 91 29 14 15 18 15
Leo N 35 43 20 11 11 1 30 14 4 8 2 2
%27 25 18 48 31 25 33 48 29 53 11 13
His
b o he
N36 21 16 1 4 0 9 2 1 0 5 1
%28 12 14 4 11 0 10 7 7 0 28 7
Bo h N 57 109 75 11 20 3 52 13 9 7 11 12
%45 63 68 48 58 75 57 45 64 47 61 80
16 Ma a F asca elli and

Angel L. Jim
enez-Fe n
andez
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d

Le us now conside wha happens when an NS is double embedded unde wo
b idge e bs (condi ion ( ) in sec ion 3.2), hus ha ing wo possible con olling
an eceden s, one o which is close han he o he :
14
(23) (A he s adium)
A: Look!Juan and his b o he a e al eady on he ack!
B: ¡S
ı!Ped o dijo que Juan c ee que a a gana
yes Ped o say.PST.3SG ha Juan belie e.3SG ha go.3SG o win
la ca e a.
he ace
‘I see!Ped o
k
said ha Juan
z
hinks ha he
k
/he
z
is going o win he ace.’
Compa ing he esul s in able 4 wi h hose o one le el o embedding ( able 2),
whe e he second op ion was no p oposed in he syn ac ic con ex , i is
in e es ing o no ice ha alues a e almos iden ical o AS a ie ies (excep
Cen ), o which he bes op ion is again ‘bo h’. Ambigui y domina es in PS
a ie ies as well. Ne e heless, PS di e s om AS inso a as he embedded
subjec / opic (‘Juan’) is p e e ed o he ma ix subjec , and he di e ence
be ween hese alues and hose in able 2 is ex emely significan (p=.0008).
Hence PS a ie ies, especially Ex and Ca , seem o be mo e sensi i e o
syn ac ic equi emen s such as locali y o he selec ion o he an eceden , (a
poin ha will be aken up in sec ion 5). No ice, howe e , ha wi h he
excep ion o Ex and Ca he bes op ion is once mo e ambigui y, sco ing a ound
60% in bo h Spanish mac o a ie ies.
4.2. NSs Embedded unde Fac i e Ve bs
Le us now conside he second o he eigh condi ions lis ed in sec ion 3.2. Fo he
sake o space, we will no p esen he da a in ol ing one le el o embedding unde a
ac i e e b, since he esul s a e basically he same as wi h b idge e bs in
sec ion 4.1, and will ins ead pass immedia ely o NSs in complex sen ences (whose
ma ix e b is ac i e) wi h wo con olling an eceden s. This condi ion is pa icula ly
Table 4. NS double embedded unde wo b idge e bs
I alian PS Anda Cas Ca Ex AS Ca ib Cen Mex Ri Pe
OK/N 128 173 111 23 35 4 91 29 14 15 18 15
Ped o N 73 12 7 1 3 1 16 4 10 2 0 0
%57 7 6 4 9 25 18 14 74 9 0 0
Juan N 20 55 29 5 18 3 20 4 1 4 5 6
%17 32 26 22 50 75 22 14 5 27 28 38
Bo h N 30 106 75 17 14 0 55 21 3 9 13 9
%26 61 68 74 41 0 60 72 21 62 72 62
14
Fo he sake o space, we will no p o ide he I alian sen ence in his and he emaining examples,
excep whe e i is impo an o compa a i e pu poses.
Unde s anding Pa iali y in p o-D op Languages 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
in e es ing since he complemen o a ac i e e b is p esupposed in o ma ion and as
such does no ha e oo -like p ope ies. This means ha an A-Topic is no allowed in
i s C domain (see Bianchi & F asca elli 2010) and a p e e bal DP like ‘Ped o’in (24)
can only be conside ed a G-Topic.
(24) A: Juan is going wi h Ped o o he ace.
B: Ya sabes, Juan lamen a que Ped o c ea que a
al eady know.2SG Juan eg e .3SG ha Ped o hink.3SG ha go.3SG
a pe de la ca e a.
o lose.INF he ace
‘You know, Juan
k
is so y ha Ped o
z
hinks ha he
k
/he
z
will lose he ace.’
As we can see in able 5, he embedded p e e bal DP is selec ed wi h high equency
as a possible an eceden o he NS, as shown by he high combined sco es o he
‘Ped o’op ion and he ‘bo h’op ion. E en hough ‘Juan’is he only possible o e A-
Topic in (24), i is no p e e ed ei he in I alian o in Spanish a ie ies. This esul
suppo s, on he one hand, he hypo hesis ha he A-Topic heading he chain can be
silen and, on he o he , ha locali y plays a ole in he selec ion o he an eceden . In
his espec , i mus be no ed ha in Ex he ‘Ped o’op ion ac ually ies wi h he ‘bo h’
op ion (50% each), showing a deg ee o p e e ence o he closes DP ha is no
a es ed in he emaining Spanish a ie ies—especially no in Cen , whe e,
excep ionally, he ‘Ped o’op ion is he leas a o ed answe (7%). Bo h hese poin s
will be aken up in sec ion 5.
4.3. NSs in Ad e bial Clauses
I is gene ally ag eed in he li e a u e ha ad e bial clauses a e no endowed wi h
illocu iona y o ce. In his espec , howe e , Haegeman (2012) a gues o an
impo an dis inc ion be ween cen al and pe iphe al ad e bial clauses. Acco ding o
Haegeman, while he le pe iphe y o cen al ad e bials o ally lacks he unc ional
p ojec ions encoding speake - ela ed unc ions (speech ime, epis emic modali y,
illocu iona y o ce) and a e wi hin he scope o ope a o s, pe iphe al ad e bial clauses
seem o admi (some) oo phenomena.
Since we we e in e es ed in examining he in e p e a ion o NSs in non oo
ad e bial clauses (i.e., in s uc u al con ex s ha do no allow o an A-Topic in he
Table 5. NS double embedded unde a ac i e e b
I alian PS Anda Cas Ca Ex AS Ca ib Cen Mex Ri Pe
OK/N 128 173 111 23 35 4 91 29 14 15 18 15
Juan N 20 16 9 3 4 0 10 2 5 1 1 1
%16 9 8 13 11 0 11 7 36 7 6 7
Ped o N 45 50 36 3 9 2 21 6 1 6 6 2
%35 29 32 13 26 50 24 21 7 40 33 13
Bo h N 63 107 66 17 22 2 60 21 8 8 11 12
%49 62 60 74 63 50 65 72 57 53 61 80
18 Ma a F asca elli and

Angel L. Jim
enez-Fe n
andez
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
embedded C domain), we limi ed ou su ey o cen al ad e bials, conside ing in
pa icula empo al and condi ional clauses. Since he esul s ob ained o hese wo
ad e bial ypes a e e y consis en , o easons o space we will only p esen da a
conce ning condi ional clauses.
Condi ional clauses we e es ed in wo s uc u al condi ions: wi h he ‘i ’clause
ealized a e and be o e he ma ix clause. This was o check he in e p e i e e ec o
an os ensible mo emen ope a ion: he pos ma ix posi ion is usually aken o be
“basic”and he p ema ix posi ion is usually defined as “ on ed”(see Haegeman
2012). In (25) is a sample sen ence o he pos ma ix posi ion.
(25) (Ped o’s iends mee o a bee . They know ha Ped o is s ill a wo k wi h his
boss. They hope he can join hem la e . One o hem says:)
Ped o puede eni si e mina el abajo.
Ped o can.3SG come i finish.3SG he wo k
‘Ped o
k
can come i he
k
/he
z
finishes his wo k.’
The fi s in e es ing esul ob ained is ha NSs in ad e bial clauses a e no always
accep ed by Spanish in o man s, unlike wha we saw o NSs in he complemen s o
b idge and ac i e e bs. Fo pos ma ix condi ional clauses, as seen in able 6, 14%
o PS in o man s and 11% o AS in o man s ga e a nega i e judgmen . Fu he mo e,
i should be no iced ha nonaccep ance is sp ead ac oss all a ie ies. This esul can
be aken as a sign ha Spanish a ie ies a e no as “consis en ”as I alian is (since he
I alian in o man s p o ided a posi i e answe o all he condi ions examined). An
o e p onoun makes he sen ence accep able o all speake s in PS and AS; he
p onoun is in e p e ed as e e ing o Ped o (Com
ınguez 2011).
Le us conside in de ail he in e p e a ions p o ided by in o man s who answe ed
posi i ely. As shown, o he fi s ime ambigui y is no he bes op ion, and he
possibili y o es ablishing a “silen head” o he opic chain (i.e., an A-Topic ha is
men ioned in he con ex bu no o e ly ealized in he sen ence) sco es e y low,
excep in Mex and Pe , whe e again ambigui y wins (56% and 64%, espec i ely); in
all he o he AS and PS a ie ies, he ma ix subjec / opic ‘Ped o’qualifies as he bes
an eceden o he NS. Hence, he sco es o he ma ix p e e bal DP a e a he high;
compa ing i s sco es wi h NSs in ad e bial clauses o i s sco es wi h NSs embedded
Table 6. NS embedded in a pos ma ix condi ional clause
I alian PS Anda Cas Ca Ex AS Ca ib Cen Mex Ri Pe
OK/N 128/128 149/173 94/111 19/23 33/35 3/4 81/91 28/29 12/14 9/15 18/18 14/15
Ped o N 84 87 58 9 18 2 38 12 7 4 11 4
%66 58 62 47 55 67 47 43 58 44 61 29
His boss N 16 19 10 5 4 0 12 8 2 0 1 1
%12 13 11 26 12 0 15 28.5 17 0 6 7
Bo h N 28 43 26 5 11 1 31 8 3 5 6 9
%22 29 27 27 33 33 38 28.5 25 56 33 64
Unde s anding Pa iali y in p o-D op Languages 19
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
unde b idge and ac i e e bs, he di e ence is ex emely significan bo h in I alian
(p<.0001) and in bo h Spanish a ie ies (PS: p<.0001; AS: p=.0150).
Le us u n o he p ema ix condi ion, in which he embedded NS is no con olled
by he ma ix subjec in su ace s uc u e:
(26) (Same con ex as in (25))
Si e mina el abajo, Ped o puede eni .
i finish.3SG he wo k Ped o can.3SG come.INF
‘I he
k
/he
z
finishes his wo k, Ped o
k
can come.’
Once again, as seen in able 7, he e is less han ull accep ance in Spanish—less e en
han in he pos ma ix con ex : 23% in PS and 18% in AS. E iden ly on ing implies
some addi ional in e p e i e complica ions. These esul s place he Spanish a ie ies
unde examina ion close o pa ial languages wi h espec o his condi ion (see
F asca elli 2018).
As o in e p e i e esul s, he figu es in able 7 clea ly show ha ambigui y comes
back as he bes choice o AS in o man s; hey also selec ‘his boss’mo e equen ly
han he ma ix subjec / opic. PS a ie ies, on he o he hand, o e a mixed pic u e, in
which Cas seems o p e e an ex e nal an eceden while Anda and Ca show no
specific p e e ence be ween o e and co e an ecedence.
Apa om he ques ion a issue in he p esen a icle (i.e., he in o ma ion-
s uc u e app oach o an unde s anding o pa ial-p o-d op p ope ies), hese da a
a e also impo an because hey show ha on ed ad e bial clauses a e no
econs uc ed in hei basic posi ion o in e p e a ion: o he wise, he in e p e i e
esul s should be he same o bo h. This is in line wi h Haegeman’s (2007, 2012)
analysis, showing ha ad e bial-clause on ing in cli ic-le -disloca ion languages
p oduces no in e en ion e ec s since cli ic-doubled le -disloca ed opics “ o m a
sepa a e class om o he A0-dependencies”(Rizzi 2004:245). Tha is why
ad e bial-clause on ing does no in e e e wi h wh mo emen ou o he ma ix
clause in languages like I alian and Spanish, as shown in (27a) and (27b)
espec i ely.
Table 7. NS embedded in a p ema ix condi ional clause
I alian PS Anda Cas Ca Ex AS Ca ib Cen Mex Ri Pe
OK/N 128/128 133/173 86/111 16/23 28/35 3/4 75/91 26/29 13/14 6/15 17/18 13/15
Ped o N 59 34 23 3 6 2 17 10 2 0 3 2
%46 25 27 19 21 67 23 38 15 0 17 15
His
boss
N29 39 20 11 8 0 21 10 4 1 3 3
%23 29 23 69 29 0 28 38 31 17 17 20
Bo h N 40 60 43 2 14 1 37 6 7 5 11 8
%31 46 50 12 50 33 49 24 54 83 66 65
20 Ma a F asca elli and

Angel L. Jim
enez-Fe n
andez
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
(27) a. Se finisce il la o o, do e and 
a Ped o ?
b. Si e mina el abajo, ¿ad
onde i 
a Ped o ?
i finish.3SG he wo k whe e go.FUT.3SG Ped o
‘Whe e will Ped o
k
go i he
k
/he
z
finishes he wo k?’
This p ope y shows ha a on ed ad e bial clause does no in e e e wi h o he ypes
o mo emen and ha econs uc ion o he licensing and in e p e a ion o an NS
canno be in oked.
4.4. The In e p e a ion o NSs wi h an In e ening Nona gumen P onoun
Le us now u n o sen ences designed o check he in e p e a ion o an NS when a
nona gumen in e enes in he supe o dina e clause ha should, acco ding o
Holmbe g, Nayudu & Sheehan (2009; see discussion o (13) in sec ion 3.1), c ea e
in e e ence in pa ial-NS languages despi e being seman ically incompa ible. This
was es ed by means o sen ences like (28), which ob ained he esul s gi en in
able 8.
(28) Juan dijo que ue una sue e que hubie a
Juan say.PST.3SG ha be.PST.3SG a luck ha ha e.PST.3SG
ganado el segundo p emio en la lo e 
ıa.
won he second p ize in he lo e y
‘Juan
k
said ha i was lucky ha he
k
/he
z
won a he lo e y.’
This ype o sen ence was always accep ed. Ambigui y is once mo e he a o ed
op ion, in all a ie ies excep Cen and, o a less ex en , Ca ib. Locali y hus
seems o be a ele an equi emen in hese wo AS a ie ies. We will see in he
emainde o his in es iga ion whe he his sensi i i y o locali y in Cen and
Ca ib is confi med. Fo all o he a ie ies, we can simply conclude ha an NS can
be licensed and in e p e ed ac oss a nona gumen DP, ei he h ough a chain wi h
an o e nonlocal p e e bal subjec / opic (‘Juan’) o h ough a co e (silen ) A-
Topic. This p o ides addi ional suppo o he alidi y o he p esen in o ma ion-
s uc u e app oach.
Table 8. NS in e p e ed ac oss a nona gumen DP
I alian PS Anda Cas Ca Ex AS Ca ib Cen Mex Ri Pe
OK/N 128 173 111 23 35 4 91 29 14 15 18 15
Juan N 15 20 14 2 3 1 16 5 3 3 4 2
%12 12 13 9 9 25 19 17 21 20 22 13
Someone
else
N31 49 31 2 12 0 20 10 6 1 2 1
%24 28 28 6 34 0 22 35 43 7 11 7
Bo h N 82 104 66 15 20 3 54 14 5 11 12 12
%64 60 59 65 57 75 59 48 36 73 67 80
Unde s anding Pa iali y in p o-D op Languages 21
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d

4.5. The Licensing and In e p e a ion o NSs in Ma ix Clauses
The da a examined in he p e ious sec ions suppo s he exis ence o silen A-Topics,
in line wi h he Topic C i e ion (1), acco ding o which e e y p edica ional sen ence
con ains a posi ion in he C domain endowed wi h he [+abou ness] ea u e. As a
consequence, when an NS is ealized in a ma ix clause, a silen A-Topic mus be
assumed in he ele an C domain (as a gued in F asca elli 2018 and Jim
enez-
Fe n
andez 2016). This is illus a ed in (29) o Spanish; (30) is a ep esen a ion.
(29) Que 
ıa p esen a e a Julio. Es mi mejo amigo.
wan .PST.1SG in oduce.INF.you.CL o Julio be.3SG my bes iend
‘I’d like o in oduce Julio
k
o you. He
k
is my bes iend.’
Conside ing he p e e ence o ano e local an eceden and he pa ial-p o-d op p ope ies
a es ed in many Spanish a ie ies, he ques ion a ises whe he (and o wha ex en ) silen
A-Topics a e accep ed in he ma ix C domains o simple clauses in hese languages.
Fo his pu pose, he su ey con ained a numbe o s imuli in which a sen ence
con aining an NS was p eceded by a sen ence con aining wo possible o e
an eceden s, wi h di e en syn ac ic unc ions. Le us s a wi h he case in which
he choice was be ween wo a gumen DPs as an eceden o a ma ix NS:
(31) Juan habl
o con Ped o aye . Aho a comp ende lo que
Juan alk.PST.3SG wi h Ped o yes e day now unde s and.3SG he wha
pas
o.
happen.PST.3SG
‘Juan
k
alked o Ped o
z
yes e day. Now he
k
/he
z
unde s ands wha happened.’
As we can see in able 9, silen A-Topics a e no ully accep ed, especially in PS
a ie ies (22% nonaccep ance compa ed o only 5% in AS).
When i comes o in e p e a ion, in o man s o mos Spanish a ie ies mos ly
selec he ‘bo h’op ion; his ambigui y is pa icula ly s ong o PS speake s,
much less so in AS. In Ca ib he an eceden is less ambiguous han in he
Table 9. Ma ix NS
I alian PS Anda Cas Ca Ex AS Ca ib Cen Mex Ri Pe
OK/N 128/128 134/173 81/111 18/23 33/35 2/4 86/91 26/29 12/14 8/15 17/18 13/15
Juan N 68 40 23 6 11 0 31 12 3 2 8 6
%53 30 28 33 33 0 36 46 25 25 47 46
Ped o N 20 7 4120117 22 00
%16 5 5 6 6 0 13 27 17 25 0 0
Bo h N 40 87 54 11 20 2 34 7 7 4 9 7
%31 65 67 61 61 100 51 27 58 50 53 54
22 Ma a F asca elli and

Angel L. Jim
enez-Fe n
andez
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
emaining AS a ie ies. This migh be in e p e ed as e idence agains he
ele ance o a hie a chy o syn ac ic unc ions o he in e p e a ion o an NS. Ye
in Ri , Pe , and mos impo an ly in Ca ib, he subjec o he p e ious sen ence is
qui e p e alen ly selec ed.
Le us u n now o a case in which he fi s sen ence includes a p e e bal subjec DP
and a nona gumen DP and he second sen ence con ains an embedded NS and no
possible an eceden o ha NS:
(32) Juan se ue aye al cine con Ped o.
Juan REFL go.PST.3SG yes e day o. he cinema wi h Ped o
S
e que es aba muy con en o.
know.1SG ha be.PST.3SG e y happy
‘Juan
k
wen o he cinema wi h Ped o
z
yes e day. I know ha he
k
/he
z
was
e y happy.’
Compa ing able 10 wi h able 9, we can see ha accep abili y dec eases in bo h PS
( om 22% nonaccep ance o 35%) and AS ( om 5% nonaccep ance o 20%). This
dec ease can be a ibu ed o he ac ha he NS is embedded, wi h no explici DP in
he ma ix clause ha is a possible an eceden ; hence, he closes possible link is a
nonlocal silen A-Topic. This clea ly ep esen s a “double”in e p e i e di ficul y o a
pa ial-p o-d op speake . The ac ha he accep abili y dec ease is ela i ely d ama ic
o PS, while AS a ie ies a e less a ec ed (wi h Mex showing no dec ease a all and
Pe showing a small inc ease), can be aken as e idence ha PS a ie ies a e “mo e
pa ial-p o-d op-like” han AS a ie ies unde his pa icula condi ion. This acco ds
wi h wha eme ged om he analysis o ad e bial clauses (sec ion 4.3) and con as s
wi h gene al claims in he li e a u e. These poin s a e aken up in sec ion 5.
As o he in e p e i e da a, ambigui y is as usual he mos equen answe , e en
hough he p e e bal subjec DP is defini ely p e e ed o e he nona gumen DP,
especially in Ex , Ca ib, and Cen .
The final con ex whe e we es ed he in e p e a ion o NSs is one in which he
o e an eceden is a ‘by’ph ase, I has been a gued in ecen wo ks ha such an
an eceden is o ally excluded (see, o example, Samek-Lodo ici 1996). A sample
sen ence is in (33), and he esul s o his sen ence ype a e gi en in able 11 ( he
‘bo h’op ion is omi ed because i was ne e selec ed by in o man s).
Table 10. Embedded NS wi h no compa ible o e an eceden in he same sen ence
I alian PS Anda Cas Ca Ex AS Ca ib Cen Mex Ri Pe
OK/N 128/128 115/173 69/111 15/23 27/35 4/4 73/91 25/29 11/14 8/15 15/18 14/15
Juan N 62 42 26 4 10 2 30 15 6 1 6 2
%48 37 38 27 37 50 41 60 55 12.5 40 14
Ped o N 7 6 4200 64 11 00
%6 5 6 13 0 0 8 16 9 12.5 0 0
Bo h N 59 67 39 9 17 2 37 6 4 6 9 12
%46 58 56 60 63 50 51 24 36 75 60 86
Unde s anding Pa iali y in p o-D op Languages 23
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
(33) El cong eso ue inaugu ado po el Decano.
he con e ence be.PST.3SG inaugu a ed by he dean
Despu
es se ue a da sus clases.
hen REFL go.PST.3SG o gi e.INF his classes
‘The con e ence was p esen ed by he dean
k
. Then he
k
/he
z
/she
w
wen o
each his/he class.’
Accep abili y judgmen s exhibi a u he wo sening compa ed o hose in ables 9
and 10. This shows ha ha ing a ‘by’ph ase as a silen A-Topic ep esen s an
addi ional complica ion, and, as a consequence, i shows ha a hie a chy exis s
be ween syn ac ic unc ions in e ms o eligibili y as a silen A-Topic. Howe e , a
syn ac ic block canno be assumed (pace Samek-Lodo ici 1996) bu a he an
in e p e i e es ic ion ha is ope a i e a he in e ace be ween discou se and syn ax.
4.6. The Licensing and In e p e a ion o NSs ac oss Va ie ies
In ligh o he di e en a iables es ed, he da a illus a ed, and he esul s discussed
in sec ions 4.1–4.5, le us now y and conside he di e en a ie ies as a iables and
see i he alues ob ained p o ide any s a is ical associa ion wi h he dependen
a iable (i.e., NS accep abili y and an eceden selec ion).
Va ie ies ha e shown di e en deg ees o pa iali y depending on di e en
syn ac ic a iables. I is impo an o conside his aspec in mo e de ail, aking in o
accoun NS accep abili y fi s o all.
As shown in ables 2, 4, 5, and 8, he e a e some syn ac ic con ex s in which no
pa ial-NS p ope ies eme ge, ha is, whe e NS accep ance is always 100%: (i)
embedding unde a b idge e b (including double embedding); (ii) embedding unde a
ac i e e b (including double embedding), and (iii) in e en ion o a hi d-pe son
(non)a gumen subjec . Hence, he c ea ion o a opic chain wi h a nonlocal ma ix
p e e bal DP is no a dis u bing ac o o Spanish a ie ies; in his espec hey
pa e n wi h consis en -p o-d op languages.
Howe e , o he a iables ha e shown ha Spanish a ie ies canno be conside ed
en i ely consis en -NS languages. These a iables and hei (no malized) accep ance
a es in each a ie y a e gi en in able 12. The s a is ical significance o he de ia ions
a es ed is indica ed g aphically: ligh g ey o significan de ia ions, da k g ey
backg ound o e y significan de ia ions, and black backg ound o ex emely
significan de ia ions. Thus, on he one hand, Cen ’s and Pe ’s de ia ions wi h
Table 11. ‘By’ph ase as an eceden
I alian PS Anda Cas Ca Ex AS Ca ib Cen Mex Ri Pe
OK/N 128/128 119/173 81/111 15/23 19/35 4/4 59/91 19/29 10/14 9/15 8/18 13/15
‘By’
ph ase
N128 108 76 13 15 4 54 15 10 9 7 13
%100 91 94 87 79 100 92 79 100 100 88 100
Someone else N 0 11 524054 0010
%0 9 613210 821 0 012 0
24 Ma a F asca elli and

Angel L. Jim
enez-Fe n
andez
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
espec o he a iables conside ed a e no significan , and hus no pa iali y is a es ed
in hese a ie ies. On he o he hand, Mex seems o go in he di ec ion o quali ying as
a pa ial-NS language, since i s alues a e e y close o hose a es ed o Finnish in
F asca elli’s (2018) s udy;
15
and some deg ee o pa iali y can also be ound in he
o he a ie ies.
The de ia ion a es ed wi h ad e bial clauses is gene ally no significan ,
independen ly o hei posi ion. In e ms o he wo king hypo hesis assumed, his
esul suppo s he conclusion ha no in e p e i e p oblem is c ea ed as long as an
o e an eceden is p esen in he ma ix clause, heading he opic chain.
On he o he hand, pa iali y eme ges when a silen A-Topic mus be es ablished o
head he chain and se e as an an eceden . Hence, NSs in opic-less ma ix clauses
de e mine a e y significan de ia ion in Ex and Mex, while NSs in embedded
clauses wi h no o e p e e bal DP in he ma ix clause de e mine a significan
de ia ion in Anda and Cas and, once mo e, a e y significan de ia ion in Mex.
In e es ingly, he s onges de ia ion eme ges when he silen A-Topic o be
es ablished is co e e en ial wi h a ‘by’ph ase, possibly due o a discou se-seman ic
inconsis ency (since i in ol es ea ing as an A-Topic an elemen ha is gene ally
conside ed o be ocused, when o e ly ealized).
Le us now conside how hese a iables de e mine and/o a ec an eceden
selec ion. In pa icula , since pa ial-NS languages ha e been a gued o p e e subjec
an ecedence (Holmbe g, Nayudu & Sheehan 2009), we will only compa e subjec and
nonsubjec alues o each a iable, se ing aside he ‘bo h’op ion. As shown by
able 13, a significan p e e ence o subjec an ecedence is only a es ed when he NS
is loca ed in a pos ma ix ad e bial clause. In all he o he cases, a opic chain wi h a
nonsubjec elemen ha ins an ia es a silen A-Topic is always possible (i no
p e e ed, as o ins ance wi h p ema ix ad e bials in Cas , Ca , Cen , and Pe ),
which is consis en wi h ou hypo hesis. We will e u n o hese esul s o final
analysis in sec ion 5.
Table 12. Rele an a iables and significan de ia ions o each a ie y
%
Pos ma ix con-
di ional ad e bial
P ema ix condi-
ional ad e bial
Ma ix
clause
Embedded +no
o e an eceden
‘By’-ph ase
an eceden
Anda 84.5 77.5 75 62 73
Cas 82.5 69.5 82.5 65 65
Ca 94 80 97 77 54
Ex 75 75 50 100 100
Ca ib 96.5 89.5 93 86 65.5
Cen 85.5 93 85.5 78.5 71.5
Mex 94.5 60 53 53 60
Ri 100 94.5 94.5 83 44.5
Pe 93 86.5 86.5 93 86.5
15
Specifically, 100% o Finnish in o man s accep NSs in pos ma ix condi ional clauses %, 48% in
p ema ix condi ional clauses, 73% in ma ix clauses, and 27% in embedded clauses wi h no o e
an eceden . ‘By’ph ases could no be es ed since Finnish has no passi es.
Unde s anding Pa iali y in p o-D op Languages 25
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
connec ed wi h (ii)). This is exp essed in he Minimal-O e -Link Condi ion (40),
which ep esen s a p oposal o be used in u u e esea ch. In o he wo ds, hese
condi ions can be aken as alid diagnos ics o (deg ees o ) pa iali y in NS languages,
suppo ing Chomsky’s (1995) idea ha syn ax eeds in e p e a ion. Fu he mo e,
since hese condi ions we e designed o check he ele ance o ha ing an A-Topic
a ailable o se e as an an eceden , we can conclude ha he Topic C i e ion (1) and
he c ea ion o opic chains a e equi emen s o he in e p e a ion o p o in NS
languages in gene al; his suppo ing he heo y ha he accep abili y and
in e p e a ion o NSs c ucially depends on an in o ma ion-s uc u e s a egy.
Accep abili y judgmen s ha e also s ongly suppo ed he oo quali y o A-Topics,
showing ha in many Spanish a ie ies he associa ion wi h a subjec / opic in he
ma ix clause is pa icula ly impo an (e en hough A-Topics can also be me ged in
he complemen s o b idge e bs). This equi emen can also be easily de i ed om
he In e ace Roo Res ic ion (10), which is pa o he p esen app oach and aken as
an addi ional a iable o check o pa iali y ac oss NS languages.
Finally, judgmen s ha e shown ha accep ance dec eases in he case o silen A-
Topics, especially when no o e low copy is a ailable in he chain, in he o m o a
local abou ness G-Topic. The p e e ence o o e copies ( o which he Minimal-
O e -Link Condi ion (40) has been elabo a ed) can also be aken as a c i e ion in
e alua ing pa iali y.
To summa ize, he ele ance o opic chains, he an eceden ole o ma ix A-
Topics, and he impo ance o o e links a e consis en ing edien s wi h he
amewo k assumed. Consequen ly, a comp ehensi e accoun o he NS pa ame e
can be based on he ollowing e ision o he Topic C i e ion:
(45) Topic C i e ion ( e ised)
a. A hi d-pe son a gumen p o ma ches he [abou ness] [ e ] ea u es ha
a e encoded in he A-Topic p ojec ion (Shi P) in he C domain.
b. A-Topic chains can only be s a ed in oo domains ( he In e ace Roo
Res ic ion).
c. A-Topic chains ac oss sen ences equi e local links, which can be o e
o silen .
d. A-Topics can be silen .
The fi s wo pa s, (45a) and (45b), apply o all NS languages, whe eas (45c) and
(45d) a e subjec o pa ame iza ion: languages showing pa ial-NS p ope ies equi e
o e links (acco ding o he Minimal-O e -Link Condi ion) and/o an o e A-Topic
heading he opic chain.
The Topic C i e ion, as e ised, has ad an ages o e a p agma ic app oach, since i
p o ides a o mal amewo k and in o ma ion-s uc u e- ela ed diagnos ics o u u e
compa a i e in es iga ions ac oss languages. I is also supe io o a na ow syn ac ic
app oach, since syn ac ic equi emen s (o condi ions) define g amma icali y in clea -
cu e ms and can ha dly cope wi h ambigui y and g adien accep ance o “ uzzy”
phenomena (which may be due o ongoing change in pa ame e se ings).
32 Ma a F asca elli and

Angel L. Jim
enez-Fe n
andez
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d

I should be no iced ha he condi ions p oposed o define (deg ees o ) pa iali y a e
independen ly mo i a ed. As a ma e o ac , he ele ance o close, in e media e
links in long-dis ance ela ions is he basis o he Minimal-Link Condi ion, and island
sensi i i y is pa o he co e g amma o languages. As o he impo ance o spelled-
ou copies in opic chains, condi ions on PF me ge ha e been p oposed in he
li e a u e o explain a numbe o phenomena and pa ame e s; see, o example,
Bobaljik’s (2002) ea men o he VO–OV dis inc ion.
17
As o o e p onouns, he p esen expe imen shows ha o e p onouns a e no
(necessa ily) discou se o ien ed. On he con a y, hey end o be in o ma ion-
s uc u ally ela ed o he ma ix p e e bal DP (like NSs), which is aken as he mos
accessible o e A-Topic. Mo e specifically, he da a show ha when an exclusi e
eading is equi ed, in he absence o in o ma ion-s uc u e cues, he ma ix subjec is
in e p e ed as an A-Topic and he de aul discou se in e p e a ion associa ed wi h an
o e p onominal subjec is ha o a amilia en i y ha is pa o a opic chain: an
abou ness G-Topic.
I eme ged om he analysis ha di e en a ie ies can show di e en sensi i i y o
one o he o he o he wo pa ial-NS condi ions. When one o mo e o hese
equi emen s is no me , deg ada ion inc eases, in line wi h ecen app oaches (see
Haegeman, Jim
enez-Fe n
andez & Rad o d 2014, Jim
enez-Fe n
andez 2017, Villa a,
Rizzi & F anck 2016). In his connec ion i emains o us o ca y ou a mo e de ailed
s udy o pa ial-NS p ope ies in o he languages, in con as wi h non-NS languages,
in o de o each a ull unde s anding o he impac o pa ial-NS p ope ies ac oss
languages and on acquisi ion p ocesses.
Re e ences
Adge , D. & P. S enonius. 2011. Fea u es in minimalis syn ax. In The Ox o d handbook o
linguis ic minimalism, ed. C. Boeckx, 27–51. Ox o d: Ox o d Uni e si y P ess.
Adli, A. 2011. G adien accep abili y and equency e ec s in in o ma ion s uc u e: A
quan i a i e s udy on Spanish, Ca alan, and Pe sian. Doc o al hesis, Uni e si €
a F eibu g,
F eibu g, Ge many.
Alonso-O alle, L., S. Fe n
andez Sole a, L. F azie & C. Cli on. 2002. Null e sus o e
p onouns and he opic- ocus a icula ion in Spanish. Ri is a di Linguis ica 14(2):1–19.
Al a , M. 1996a. Manual de dialec olog
ıa hisp
anica: El espa~
nol de Am
e ica. Ba celona: A iel.
Al a , M. 1996b. Manual de dialec olog
ıa hisp
anica: El espa~
nol de Espa~
na. Ba celona: A iel.
Ba bosa, P., M. E. Dua e & M. A. Ka o. 2005. Null subjec s in Eu opean and B azilian
Po uguese. Jou nal o Po uguese Linguis ics 4:11–52.
Bianchi, V. & M. F asca elli. 2010. Is opic a oo phenomenon? Ibe ia 2:43–88.
Bibe aue , T., A. Holmbe g, I. Robe s & M. Sheehan. 2010. Pa ame ic a ia ion: Null
subjec s in Minimalis heo y. Camb idge: Camb idge Uni e si y P ess.
Bizza i, C. 2015. Russian as a p o-d op language: Da a and analysis om a new s udy. Annali
di Ca’Fosca i 49:335–362.
Bobaljik, J. D. 2002. A-chains a he PF-in e ace: Copies and “co e ”mo emen . Na u al
Language and Linguis ic Theo y 20.2:197–267.
17
Specifically, Bobaljik a gues ha Holmbe g’s Gene aliza ion is he esul o a mo phophonological
cons ain on e b inflec ion, equi ing me ge unde PF adjacency (suppo ed by VO–OV da a). In sho ,
PF can decide which copy o spell ou and de i e he e ec s.
Unde s anding Pa iali y in p o-D op Languages 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
Camacho, J. 2013. Null subjec s. Camb idge: Camb idge Uni e si y P ess.
Camacho, J. 2016. The null subjec pa ame e e isi ed: The e olu ion om null subjec
Spanish and Po uguese o Dominican Spanish and B azilian Po uguese. In The
mo phosyn ax o Po uguese and Spanish in La in Ame ica, ed. M. Ka o & F. O d
o~
nez,
27–48. Ox o d: Ox o d Uni e si y P ess.
Ca mina i, M. N. 2002. The p ocessing o I alian subjec p onouns. Doc o al hesis,
Massachuse s Ins i u e o Technology, Camb idge.
Chamo o, G. 2018. O fline in e p e a ion o subjec p onouns by na i e speake s o Spanish.
Glossa: A Jou nal o Gene al Linguis ics 3.1:1–16.
Chomsky, N. 1982. Some concep s and consequences o he heo y o Go e nmen and
Binding. Camb idge, MA: MIT P ess.
Chomsky, N. 1995. The Minimalis P og am. Camb idge, MA: MIT P ess.
Cinque, G. 1990. Types o A0-dependencies. Camb idge, MA: MIT P ess.
Cinque, G. & L. Rizzi. 2008. The ca og aphy o syn ac ic s uc u es. S udies in Linguis ics
2:42–58.
Cognola, F. & J. Casalicchio. 2018. Unde s anding null subjec s: A synch onic and diach onic
pe spec i e. Ox o d: Ox o d Uni e si y P ess.
Com
ınguez, J. P. 2011. Occu ence and in e p e a ion o subjec p onouns in empo al
embedded clauses in L2 English nea -na i e speake s o L2 Spanish. Ms., Ru ge s
Uni e si y, New B unswick, NJ.
Com
ınguez, J. P. 2018. The na u e and posi ion o subjec s in Pue o Rican wh-ques ions. In
Cu en esea ch in Pue o Rican linguis ics, ed. M. Gonz
alez-Ri e a, 67–89. New Yo k:
Rou ledge.
Emonds, J. 2004. Unspecified ca ego ies as in he key o oo cons uc ions. In Pe iphe ies:
Syn ac ic edges and hei e ec s, ed. D. Adge , C. De Ca & G. Tsoulas, 75–120. Do d ech ,
he Ne he lands: Kluwe .
Filiaci, F., A. So ace & M. Ca ei as. 2014. Anapho ic biases and null o e subjec s in I alian and
Spanish: A c oss-linguis ic compa ison. Language and Cogni i e P ocesses 29.7:825–843.
F asca elli, M. 2007. Subjec s, opics and he in e p e a ion o e e en ial p o: An in e ace app oach
o he linking o (null) p onouns. Na u al Language and Linguis ic Theo y 25:691–734.
F asca elli, M. 2017. Romance p o-d op languages a he in e aces: A compa a i e analysis.
Pape p esen ed a he 31s Going Romance symposium, Uni e si y o Bucha es , Bucha es ,
Romania, Decembe 8.
F asca elli, M. 2018. The in e p e a ion o p o in consis en and pa ial null-subjec languages:
A compa a i e in e ace analysis. In Null subjec s in gene a i e g amma : A synch onic and
diach onic pe spec i e, ed. F. Cognola & J. Casalicchio, 211–239. Ox o d: Ox o d
Uni e si y P ess.
F asca elli, M. & R. Hin e h€
olzl. 2007. Types o opics in Ge man and I alian. In On
In o ma ion s uc u e, meaning, and o m, ed. K. Schwabe & S. Winkle , 87–116.
Ams e dam: John Benjamins.
F asca elli, M. & A. Puglielli. 2009. Posi ion, unc ion, and in e p e a ion o opics in Somali.
In In o ma ion s uc u e and i s in e aces, ed. L. Me eu, 325–348. Be lin: G uy e .
Gi 
on, T. 1983. Topic con inui y in discou se: An in oduc ion. In Topic con inui y in
discou se: A quan i a i e c oss-language s udy, ed. T. Gi 
on, 5–41. Ams e dam: John
Benjamins.
Gup on, T. & S. Lowman. 2013. An F p ojec ion in Cibe~
no Dominican Spanish. In Selec ed
p oceedings o he 16 h Hispanic Linguis ics symposium, ed. J. Cab elli Ama o, J. Lo d, A.
de P ada P
e ez & J. Alana Aa on, 338–348. Some ille, MA: Cascadilla P ess.
Gu i
e ez-B a o, R. 2008. Topicaliza ion and p e e bal subjec s in Spanish wh-in e oga i es.
In Selec ed P oceedings o he 10 h Hispanic Linguis ics Symposium, ed. J. B uhn de
Ga a i o & E. Valenzuela, 225–236. Some ille, MA: Cascadilla P ess.
Haegeman, L. 2007. Ope a o mo emen and opicaliza ion in ad e bial clauses. Folia
Linguis ica 18:485–502.
34 Ma a F asca elli and

Angel L. Jim
enez-Fe n
andez
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
Haegeman, L. 2012. Ad e bial clauses, main clause phenomena, and he composi ion o he le
pe iphe y. Ox o d: Ox o d Uni e si y P ess.
Haegeman, L., 
A. L. Jim
enez-Fe n
andez & A. Rad o d. 2014. Decons uc ing he Subjec
Condi ion in e ms o cumula i e cons ain iola ion. Linguis ic Re iew 31:73–150.
Heim, I. 1982. The seman ics o defini e and indefini e noun ph ases. Doc o al hesis,
Uni e si y o Massachuse s, Amhe s .
Heycock, C. 2006. Embedded oo phenomena. In The Blackwell companion o syn ax, ed. M.
E e ae & H. an Riemsdijk, ol. 2, 174–209. Ox o d: Blackwell.
Holmbe g, A. 2005. Is he e a li le p o? E idence om Finnish. Linguis ic Inqui y 36:535–564.
Holmbe g, A. 2010. Null subjec pa ame e s. In Pa ame ic a ia ion: Null subjec s in
minimalis heo y, ed. T. Bibe aue , A. Holmbe g, I. Robe s & M. Sheehan, 88–124.
Camb idge: Camb idge Uni e si y P ess.
Holmbe g, A., A. Nayudu & M. Sheehan. 2009. Th ee pa ial null-subjec languages: A
compa ison o B azilian Po uguese, Finnish, and Ma a hi. S udia Linguis ica 63:59–97.
Huang, C.-T. J. 1989. P o-d op in Chinese: A gene alized con ol heo y. In The null subjec
pa ame e , ed. O. Jaeggli & K. Safi , 185–214. Do d ech , he Ne he lands: Kluwe .
Jaeggli, O. 1982. Topics in Romance syn ax. Do d ech , he Ne he lands: Fo is.
Jaeggli, O. & K. Safi . 1989. The null subjec pa ame e and pa ame ic heo y. In The null
subjec pa ame e , ed. O. Jaeggli & K. Safi , 1–44. Do d ech , he Ne he lands: Kluwe .
Jim
enez-Fe n
andez, 
A. L. 2016. When discou se me null subjec s. Bo ealis: An In e na ional
Jou nal o Hispanic Linguis ics 5.2:173–189.
Jim
enez-Fe n
andez, 
A. L. 2017. P eposi ions and islands: Ex ac ion om da i e and accusa i e
DPs in psych e bs. In Language use and linguis ic s uc u es: P oceedings o he Olomouc
Linguis ics Colloquium 2016, ed. J. Emonds & M. Jane ob
a, 155–172. Olomouc: Palack
y
Uni e si y Olomouc P ess.
Jim
enez-Fe n
andez, 
A. L. & S. Miyagawa. 2014. A ea u e-inhe i ance app oach o oo
phenomena and pa ame ic a ia ion. Lingua 145:275–302.
Ka o, M. 2000. The pa ial p o-d op na u e and he es ic ed VS o de in B azilian Po uguese.
In B azilian Po uguese and he null subjec pa ame e , ed. M. A. Ka o & E. Neg ~
ao, 223–
258. F ank u , Ge many: Ve ue .
Ka o, M. 2011. Acquisi ion in he con ex o language change: The case o B azilian
Po uguese. In The de elopmen o g amma : Language acquisi ion and diach onic change,
ed. E. Rinke & T. Kupisch, 309–330. Ams e dam: John Benjamins.
Kayne, R. S. 2016. The silence o heads. S udies in Chinese Linguis ics 37.1:1–37.
K i ka, M. 2007. Basic no ions o in o ma ion s uc u e. In e disciplina y S udies on
In o ma ion S uc u e 6:13–55.
Ku oda, S.-Y. 1965. Gene a i e g amma ical s udies in he Japanese language. Camb idge,
MA: MIT P ess.
Lamb ech , K. 1994. In o ma ion s uc u e and sen ence o m: Topic, ocus, and he men al
ep esen a ion o discou se e e en s. Camb idge: Camb idge Uni e si y P ess.
Le inson, S. C. 1991. P agma ic educ ion o he Binding Condi ions e isi ed. Jou nal o
Linguis ics 27:107–161.
Manzini, M. R. 1992. Locali y: A heo y and some o i s empi ical consequences. Camb idge,
MA: MIT P ess.
Mao, T. & F. Meng. 2016. The ca og aphic p ojec o he gene a i e en e p ise: An in e iew
wi h Guglielmo Cinque. Language and Linguis ics 17.6:917–936.
Meinunge , A. 2004. Ve b posi ion, e bal mood, and he ancho ing (po en ial) o sen ences. In
The syn ax and seman ics o he le pe iphe y, ed. H. Lohns ein & S. T issle , 313–341.
Be lin: Mou on de G uy e .
Modes o, M. 2000. Null subjec wi hou “ ich”ag eemen . In B azilian Po uguese and he null
subjec pa ame e , ed. M. A. Ka o & E. Neg ~
ao, 147–174. F ank u , Ge many: Ve eu .
Mon albe i, M. M. 1986. How p o is i ? In S udies in Romance linguis ics, ed. O. Jaeggli & C.
Sil a-Co al
an, 137–152. Do d ech , he Ne he lands: Fo is.
Unde s anding Pa iali y in p o-D op Languages 35
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d
Ojea, A. 2017. Co e in en ional ea u es in he syn ac ic compu a ion: De i ing he posi ion o
he subjec in Spanish. Lingua 195:72–91.
O d
o~
nez, F. & A. Ola ea. 2006. Mic o a ia ion in Ca ibbean/non Ca ibbean Spanish
in e oga i es. P obus 18.1:59–96.
Pe lmu e , D. 1971. Deep and su ace cons ain s in syn ax. New Yo k: Hol , Rineha and
Wins on.
Pese sky, D. & E. To ego. 2007. The syn ax o alua ion and he in e p e abili y o ea u es.
Ams e dam: John Benjamins.
Pe
sko 
a, A. 2014. In o ma ion s uc u e and he use o p onominal subjec s in Spanish. In
Me hodological issues in he s udy o in o ma ion s uc u e, ed. D. El Za ka & S. Heidinge ,
43–67. G az, Aus ia: G aze Linguis ische S udien.
Posio, P. 2012. P onominal subjec s in Peninsula Spanish and Eu opean Po uguese:
Seman ics, p agma ics, and o mulaic sequences. Helsinki: Uni e si y o Helsinki P ess.
Reinha , T. 1981. P agma ics and linguis ics: An analysis o sen ence opics. Philosophica
27:53–94.
Rizzi, L. 1986. Null objec s in I alian and he heo y o p o.Linguis ic Inqui y 17:501–558.
Rizzi, L. 1990. Rela i ized Minimali y. Camb idge, MA: MIT P ess.
Rizzi, L. 1997. The fine s uc u e o he le pe iphe y. In Elemen s o g amma : Handbook in
gene a i e syn ax, ed. L. Haegeman, 281–337. Do d ech , he Ne he lands: Kluwe .
Rizzi, L. 2004. Locali y and le pe iphe y. In The ca og aphy o syn ac ic s uc u es, ol. 3,
S uc u es and beyond, ed. A. Belle i, 223–251. Ox o d: Ox o d Uni e si y P ess.
Rizzi, L. 2006. On he o m o chains: C i e ial posi ions and ECP e ec s. In Wh-mo emen :
Mo ing on, ed. L. L.-S. Cheng & N. Co e , 97–134. Camb idge, MA: MIT P ess.
Robe s, C. 2003. Uniqueness in defini e noun ph ases. Linguis ics and Philosophy 26:287–350.
Rod igues, C. 2004. Impo e ished mo phology and A-mo emen ou o case domains. Doc o al
hesis, Uni e si y o Ma yland, College Pa k.
Saab, A. 2016. On he no ion o pa ial (non-)p o-d op in Romance. In The mo phosyn ax o
Po uguese and Spanish in La in Ame ica, ed. M. A. Ka o & F. O d
o~
nez, 49–57. Ox o d:
Ox o d Uni e si y P ess.
Samek-Lodo ici, V. 1996. Cons ain s on subjec s: An op imali y heo e ic analysis. Doc o al
hesis, Ru ge s Uni e si y, New B unswick, NJ.
Sigu ðsson, H. 2004. Meaning ul silence, meaningless sounds. Linguis ic Va ia ion Yea book
4:235–259.
Tanaka, H. 2016. A minimalis analysis o English opicaliza ion: A phase-based ca og aphic
complemen ize ph ase (CP) pe spec i e. Jou nal o UOEH 38.4:289–389.
Ticio, M. E. 2018. On Pue o Rican subjec s. In Cu en esea ch in Pue o Rican linguis ics,
ed. M. Gonz
alez-Ri e a, 90–108. New Yo k: Rou ledge.
To ibio, A. J. 2000. Se ing pa ame ic limi s on dialec al a ia ion in Spanish. Lingua
110.5:315–341.
an de Sand , R. 1992. P esupposi ion p ojec ion as anapho a esolu ion. Jou nal o Seman ics
9:333–377.
Villa a, S., L. Rizzi & J. F anck. 2016. In e en ion e ec s and Rela i ized Minimali y: New
expe imen al e idence om g aded judgmen s. Lingua 179:76–96.
Ma a F asca elli
Uni e si 
a degli S udi Roma T e
Dipa imen o di Lingue Le e a u e e Cul u e S anie e
Via del Valco di San Paolo, 19
Roma 00146
I aly
ma a. asca el[email p o ec ed]
36 Ma a F asca elli and

Angel L. Jim
enez-Fe n
andez
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d

Angel L. Jim
enez-Fe n
andez
Uni e sidad de Se illa
Depa amen o de Filolog
ıa Inglesa (Lengua Inglesa)
Calle Palos de la F on e a, s/n
Se illa 41004
Spain
[email p o ec ed]
Unde s anding Pa iali y in p o-D op Languages 37
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
©2019 John Wiley & Sons L d