Content uploaded by Ana Maria Martins
Author content
All content in this area was uploaded by Ana Maria Martins on Aug 17, 2018
Content may be subject to copyright.
Subject doubling in European Portuguese
dialects
e role of impersonal se*
Ana Maria Martins
University of Lisbon
is paper discusses a subject doubling construction found in European
Portuguese dialects where the impersonal clitic se is doubled by a strong
pronoun/DP. e availability of subject doubling is explained under the
hypothesis that dialectal se escapes the Case lter because it is ᶲ-incomplete
(plural, but person-less). Se and its doubler begin as a single constituent (a “big
DP”); later, the clitic se head-moves to T while the doubling pronoun/DP takes
one of the positions available to EP subjects (preverbal or post-verbal). is
analysis rightly predicts that: (i) there is no xed word order between se and its
doubler, (ii) the doubling pronoun/DP controls subject-verb agreement, (iii) the
interpretation of the double subject is compositional (the doubler establishes the
inclusive or exclusive reading of impersonal se). e analysis also enlightens why
the clitic se can behave as the universal plural anaphora or the expletive subject of
impersonal predicates in some dialects.
1. Introduction
European Portuguese (EP) dialects display a double subject construction where
the impersonal clitic se shares with a (plural) strong pronoun, or less commonly
a full DP, the role of sentential subject. In the illustrative sentences in (1) the sub-
ject clitic se is doubled by a gente, a former DP (“the people”) grammaticalised
as a rst person plural pronoun (see Menuzzi 2000; Lopes 2003). In the English
translations of the examples in (1), we is to be understood as non-specic, roughly,
“people here, including myself”.
*My thanks to Jairo Nunes, an anonymous reviewer and the audience of Going Romance 2007
(Amsterdam) for comments and helpful discussion of issues addressed herein.
1 Ana Maria Martins
European Portuguese dialects:
(1) a. A gente chama-se rãs a isto. (CORDIAL-SIN.1 FLF)
pron.1pl calls-SE rãs to that
“We call these ones rãs (frogs).”
b. Chama-se-lhe a gente espigas. (CORDIAL-SIN. AAL)
calls-SE-it.dat pron.1pl espigas
“We call it espigas (spikes of cereal).”
Standard European Portuguese does not allow the double subject construction
exemplied in (1). us while a gente and se can separately act as an indenite
subject, they are not allowed to co-occur in that role. See the paradigm in (2).
Standard European Portuguese:
(2) a. *Aˉgente chama-se rãs a isto. (ˉ≠ˉEPˉdialectsˉunderˉdiscussion)
pron.1pl calls-SE rãs to that
“We call these ones rãs (frogs).”
b. Aˉgente chama rãs a isto. (ˉ=ˉEPˉdialectsˉunderˉdiscussion)
pron.1pl callsˉ rãs to that
“We call these ones rãs (frogs).”
c. Chama-se rãs a isto. ( = EP dialects under discussion)
calls-SE rãs to that
“We call these ones rãs (frogs).”
e aim of this paper is to oer an analysis of the double subject construction
exemplied in (1),2 and so clarify what makes it available in the EP dialects under
observation, while standard EP (like standard Spanish or Italian) rejects it. e
facts are interesting from a theoretical perspective since within the generative
1. CORDIAL-SIN: Syntax-oriented Corpus of Portuguese Dialects. See http://www.clul.ul.pt.
2. e double subject SE construction is found in the archipelagos of Azores and Madeira
as well as in continental Portuguese. It is much more common in the Centre and South of
Portugal than in the North (where nonetheless it is also attested). It is fully ungrammatical
in standard EP and has gone totally unobserved by philologists and linguists who dealt
with dialect variation in European Portuguese. All the dialects where the double subject
SE construction is found also display the impersonal nominative SE construction of standard
European Portuguese, which is similar to the “non-agreeing” SE/SI construction of Spanish
and Italian (see Cinque 1988; Dobrovie-Sorin 1998; Rivero 2000; and D’Alessandro 2004,
among others). e data offered in this paper come from CORDIAL-SIN (Syntax-oriented
corpus of Portuguese dialects).
Subject doubling in European Portuguese dialects 11
framework, on which this work hinges, sentences with two subjects (neither of
which is le dislocated) should, in principle, be disallowed.3
e paper is organized in seven sections. Section 2 deals with word order.
e fact that there is no xed ordering between the two elements that express
the subject makes previous analyses of subject doubling unt to characterize the
dialectal EP construction. Section 3 describes agreement properties of the con-
struction showing that subject-verb agreement does not depend on impersonal
se but is set by the doubling nominative pronoun or DP. In Section 4, the inter-
pretative eects of subject doubling in the impersonal se construction are con-
sidered. It is proposed that the double subject has a compositional interpretation
with the strong pronoun or DP acting as a restrictor on the denotative set of
impersonal se and establishing its inclusive or exclusive reading. Section 5 draws
an integrative analysis of the syntactic and semantic facts introduced in the pre-
3. e type of se under discussion in this paper differs from the aspectual reflexive found in
Spanish (Strozer 1976; Zagona 1996; De Miguel 1999; Otero 1999) in that it is compatible with
intransitive (unergative) verbs and is not restricted to sentences with perfective predicates. See
(i) below, with aspectual se, in contrast to the examples above.
(i) a. Juan (*se) come normalmente en (Spanish.ˉFromˉDeˉMiguelˉ1999:2995)
JuanˉSE eats usually in
este bar.
this bar
“Juan usually eats in this bar.”
b. Juan (*se) come tortilla siempre que puede (idem)
Juan SE eats omelette always that can.3sg
“Juan eats omelette whenever he can.”
c. Juan (se) comió una tortilla él solo (idem)
Juan SE ate.3sg an omelette he alone
“Juan ate an omelette all by himself.”
It also differs from both the aspectual reflexive and the benefactive/malefactive dative (Strozer
1978, among others), exemplified in (ii), as it surfaces exclusively with the form se while the
aspectual and the benefactive/malefactive clitics display person distinctions (e.g. me.1SG,
te.2SG, etc.).
(ii) Pepe me le comió la manzana al niño
Pepe me.dat him.dat ate the apple to the.boy
“Pepe ate the boy’s apple.”
Moreover, the dative clitic is never phonologically realized as se in European Portuguese. In
this respect standard EP and dialectal EP are alike.
12 Ana Maria Martins
vious sections. e compositional interpretation is derived under the hypothesis
that the double subject results from the splitting of an earlier single constituent,
a complex DP, where se is the nuclear element. e particular property of se that
makes it irrelevant for subject-verb agreement and Case is taken to be its lack
of person. Hence nominative Case rests available for the doubling pronoun or
DP. In Section 6 independent evidence is brought into consideration in order to
support the approach developed in the paper. It is shown that the very same dia-
lects where the double subject se construction is attested display other peculiar
features pointing to the defectiveness of se with respect to ϕ-features, in sharp
contrast with what is observed in standard EP.
2. Word order as evidence for different types of subject doubling
In the double subject se construction, the two subject items do not appear in a xed
order. e subject clitic se may follow or precede the doubling strong pronoun/
DP, as illustrated by the root sentences in (1) – here repeated as (3) – and by the
embedded clauses in (4).4
(3) a. A gente chama-se rãs a isto. (CORDIAL-SIN. FLF)
pron.1pl calls-SE rãs to that
“We call these ones rãs (frogs).”
b. Chama-se-lhe a gente espigas. (CORDIAL-SIN. AAL)
calls-SE-it.dat pron.1pl espigas
“We call it espigas (spikes of cereal).”
(4) a. Este pode ser a coisa que a gente se
this can be the thing that pron.1pl SE
diz peixe-cavalo. (CORDIAL-SIN. CLC)
says peixe-cavalo
“is one could be what we call peixe-cavalo (a kind of sh).”
b. O mugeˉé o.que se chama aˉgente quando anda dentro
the muge isˉwhat SE calls pron.1pl when is inside
do rio. (CORDIAL-SIN. ALV)
of.the river
“Muge is the name we give to that kind of sh when it is found in the river.”
4. I take the preverbal non-clitic subject in sentences like (1) to sit in Spec,ΣP/subjP (Martins
1994; Cardinaletti 1997, 2004 respectively), a position of the high IP field. e post-verbal
non-clitic subject in sentences like (2) stays in Spec, vP.
Subject doubling in European Portuguese dialects 13
is variable word order sets the EP double subject se construction apart from
other cases of subject doubling reported in the literature and is le unaccounted
for by previous analyses. e construction falls outside the scope of the analyses
that handle subject doubling as a case of multiple realisation of copies of a move-
ment chain (Craenenbroeck & Van Koppen 2002; Van Koppen 2007, for Dutch;
Holmberg & Nikanne 2006, for Swedish). It is also not covered by the analyses that
derive subject doubling from multiple external merges of distinct lexical items in
dierent subject positions made available by clause structure (Holmberg & Nikanne
2006, for Finnish).
e sentences in (5), taken from Van Koppen (2007), illustrate subject doubling
in Wambeeck Dutch. Here the strong pronoun or DP, functioning as subject, neces-
sarily precedes the co-referring weak pronoun. e Wambeeck Dutch construction
is analysed by Craenenbroeck & Van Koppen (2002), and Van Koppen (2007) as an
instance of A’-movement coupled with multiple spell-out of chain links (see Nunes
2004; Corver & Nunes 2007). e subject moves from Spec,IP to a topic position in
the CP-domain. Concomitant spell-out of the higher and lower copies (in Spec,TopP
and Spec,IP respectively) is an available option because the head of chain diers
from the other copies in that the latter do not contain internal structure.5
(5) a. Zaai gui zaai nuir ojsh.
she goes she to home
“She is going home.” (Wambeek Dutch)
b. Dei vrou gui zaai nuir ojsh.
that woman goes she to home
“at woman is going home.” (Wambeek Dutch)
is type of subject doubling clearly diers from what is found in European
Portuguese dialects.6 at the doubling pronoun/DP in the EP impersonal se
construction is not a topic is clearly shown by the fact that it may surface in post-
verbal position (see (3b), (4b)), being in this case preceded by the impersonal
subject clitic.
Another type of subject doubling was reported in the literature where the weak
subject pronoun precedes the subject DP or strong pronoun. is is the Finnish
pattern (see Examples (6a–b)), which Holmberg & Nikanne (2006) analyse as
represented in (7).
5. According to Van Koppen (2007), lower copies are reduced in the sense that they only
contain the features of the maximal projection of the moved item (the reduced copy theory).
6. e Wambeek Dutch type of subject doubling is also found in Swedish (see Holmberg &
Nikanne 2006) and does not seem to significantly differ from Brazilian Portuguese or French
le dislocation of the subject with resumption.
14 Ana Maria Martins
(6) a. Se on Jari lopettanut tupakoinnin.
he has Jari quit smoking
“Jari has quit smoking.”
b. Ne sai kaikki lapset samat oireet.
they got all children same symptoms
“All the children got the same symptoms.”
(7) [FinP se [on+Fin [PredP Jari lopettanut tupakoinnin]
SE has Jari quit smoking
“Jari has quit smoking.”
According to Holmberg & Nikanne (2006), Finnish clause structure makes two
positions available for subjects: a position inside the information Focus domain,
that is PredP, and a position inside the Presupposition domain, that is FinP. In
the Finnish double subject construction, the subject is interpreted as information
focus because it remains inside PredP. e doubling weak pronoun directly merges
in Spec,FinP checking the EPP.7 According to Holmberg & Nikanne (2006), the
fact that Finnish has two pronouns with number but no person (i.e. se-SG and ne-
PL) is crucial to account for the availability of subject doubling:
Why is subject doubling of the Finnish kind not found in all languages?
Presumably either because the other languages lack the requisite categories,
i.e. decient pronouns, or because they lack the requisite ‘structural positions’
(Holmberg & Nikanne 2006)
As for the EP double subject impersonal se construction, given the non rigid order
of the two subject items, it is not reducible to the Finnish type of subject dou-
bling. Nevertheless, it will be proposed in Section 5 that the idea that person-less
pronouns lie behind subject doubling is to be maintained. e EP double subject
impersonal se construction thus lends further empirical support to the proposals
in Holmberg & Nikanne (2006).
7. e weak pronoun can only follow the subject DP if the latter is A’-moved to the CP
domain because it is a contrastive focus or a wh- phrase, as illustrated below.
(i) [CP Sinäkini [
FinP se [olet+Fin [PredPˉSinäkinˉi vihdoinkin
you-too SE have finally
lopettanut tupakoinnin]
quit smoking
“Soˉyou,ˉtoo,ˉhaveˉfinallyˉquitˉsmoking.”ˉ(contrastiveˉfocusˉonˉSinäkin)
(ii) [CP kukai [
FinP se [on+Fin [PredP who i vihdoinkin lopettanut tupakoinnin]
who SE has le the door open again?
Subject doubling in European Portuguese dialects 15
3. Agreement in the double subject impersonal se construction
In the double subject impersonal se construction, the verb displays agreement with
the strong doubling pronoun or DP, as shown in (8) below, where the verb agrees
with the nominative rst person plural pronoun nós “ we”. Sentences (9a) and (9b)
have the rst person plural pronoun a gente (literally, “the people”) in subject posi-
tion on a par with se. e pronoun a gente induces variable patterns of subject-verb
agreement (see Fn. 15), which allows the verb to surface in 3rd person singular (see
(1) above), in 1st person plural (see (9a) below), and in 3rd person plural (see (9b)
below). e two latter options make subject-verb agreement visible.
(8) Há várias qualidades que até ainda nós não
there.is several species that even already we not
se conhece-mos. (CORDIAL-SIN.ˉALV)
SE know-1pl
“ere are so many species among sh that even we (shermen) do not know
all of them yet.”
(9) a. Não sabem o.que a gente se
not know.3pl what pron.1pl SE
passá-mos aí. (CORDIAL-SIN. CLC)
passed-1pl there
“You don’t know all we’ve been through.”
b. A tesoura que a gente se tosquiava-m as ovelhas!
the scissors that pron.1pl SE shear-3pl the sheep!
A gente pegava-m-lhe
pron.1pl would.hold-3pl-them.dat
pela cabeça… (CORDIAL-SIN. MIG)
by.the head
“is is the scissors that we would shear the sheep with! We would hold
their head…”
Moreover, a null pronoun may be the doubler of se and control subject-verb agree-
ment.8 In the sentences in (10), a 1st person plural null subject pronoun is respon-
sible for the 1st person plural morphology on the verb.
(10) a. Nunca se vi-mos este peixe. (CORDIAL-SIN. CLC)
not SE see-1pl this sh
“We never saw that kind of sh.”
. European Portuguese is uniformly a null subject language; no dialectal variation is found
in this respect.
16 Ana Maria Martins
b. Com a licença de caça e tudo, não se
with the licence of hunting and all not SE
pode-mos caçar… (CORDIAL-SIN. EXB)
can-1pl hunt.INFIN
“Even having a hunting licence, we are not allowed to hunt…”
ese same dialects display non-doubling impersonal se sentences similar to those
found in standard EP, Spanish or Italian. In this case the verb invariably surfaces
in third person (singular or plural), like in (11) below.9 e fact that in the double
subject impersonal se construction the verbal inection always has its ᶲ-features
valued by the doubling strong pronoun or DP (see the ungrammaticality of (12))
shows that se is unable to enter the kind of Agree relation on which nominative
Case assignment is dependent (see Chomsky 2000, 2001). It will be proposed in
Section 5 that this is because the impersonal clitic se is person-less in the relevant
EP dialects.
(11) Há várias qualidades que ainda não se conhece(m).
there.is several species that already not SE k now. 3sg(/3pl)
“ere are many species among sh that are still unknown.”
(12) *Há várias qualidades que até ainda nós não se conhece(m).
there.is several species that even already we not SE k now. 3sg(/3pl)
“ere are so many species among sh that even we (shermen) do not know
all of them yet.”
4. Interpretative effects in the double subject impersonal se construction
e bipartite subject of the dialectal impersonal se construction usually has
a generic reading.10 e impersonal subject denotes an unspecied group of
9. On this matter, see Burzio (1986), Raposo and Uriagereka (1996), Rivero (2000) and
D’Alessandro (2004).
1. A definite reading is also attested in the corpus but appears to be infrequent. We find but
the two examples given below:
(i) Então, aˉgente deu-se o jantar e
then pron.1pl gave-3sg.SE the dinner and
cou melhorˉ (CORDIAL-SIN.ˉFLF)
stayed better
“en we offered that dinner to our neighbours and things started to go better.”
[‘we’ = ‘our family’]
Subject doubling in European Portuguese dialects 17
humans, which may include or exclude the speaker. In this respect, it does not dier
from the standard impersonal se construction of EP, Spanish and Italian (see the
references in Footnote (2) and Chierchia 1995). But the doubling pronoun/DP
acts as a restrictor on the denotative range of se, and establishes its inclusive or
exclusive reading.
4.1 Inclusive reading (the doubling pronoun/DP is 1st person plural)
e inclusive reading, which sentences (13) and (14) illustrate, arises when the
doubling pronoun/DP is 1st person plural, be it a null or an overt pronoun, as
shown in (13) and (14) respectively. ese and the other examples throughout the
paper come from the CORDIAL-SIN dialectal corpus (see Footnote (1)) and arose
in the context of spontaneous conversation.
In (13) we can see how the speaker reformulates the sentence to make clear
that he is mentioning what local people, including himself, used to do, not what
anyone would do. is is unambiguously conveyed by the double subject imper-
sonal se sentence, not by the standard impersonal se sentence that the speaker
uttered in rst place. Note that the 1st person plural agreement morphology on the
verb indicates the presence in the relevant sentence of a null pronoun with those
same ᶲ-features.
(13) Sabe às.vezes o.que se faz, o.que se
know sometimes what SE does, what SE
fazía-mos, antigamente? (CORDIAL-SIN. OUT)
would.do-1pl, in.older.times
“Do you know what one would do, what we would do, in older times?”
In (14) the contrast between the shermen of Câmara de Lobos (Madeira island)
and the inhabitants of Lisboa with respect to their eating habits can be set with
no need for a locative argument in the rst member of the coordinate structure
because the presence of the 1st person plural pronoun a gente is enough to restrict
(ii) Leis! Como a gente se falámos
laws as pron.1pl SE talked.1pl
ainda agora (CORDIAL-SIN. CLC)
just now
“at’s what laws are made for, as we have just talked about.”
[‘we’ = ‘the four of us who are here talking about this and that’]
1 Ana Maria Martins
the range of the group of humans denoted by se, and thus establish its inclusive
reading.11
(14) A gente não se come, mas os de Lisboa diz que comem
pron.1pl not SE eats but the.ones of Lisbon says that eat.3pl
daquele peixe. (CORDIAL-SIN. CLC)
of.that sh
“Here we don’t eat that sh but we heard that in Lisbon people eat it.”
4.2 Exclusive reading (the doubling pronoun/DP is 3rd person plural)
When an exclusive reading of se is intended, this is obtained by placing a 3rd per-
son plural pronoun (overt or null) as the doubler/restrictor of se,12 as illustrated by
(15). Although the speaker did not himself work with the farm tool whose name he
is being asked, he knows exactly what people used it for in former times. e sen-
tence conveys a generic reading that excludes the speaker. Sentences like (15) con-
rm that both members of the bipartite subject contribute to interpretation since
the generic reading would not be available if the pronoun eles “they” expressed
alone the sentential subject.
(15) Não sei lá a certeza também do nome, também não digo.
“I’m not sure about the name, so I won’t spell it out.”
Sei é de real certeza que isto era com
know.1sg is of real certainty that this was with
o.que se eles batia-m o centeio. (CORDIAL-SIN. FLF)
what SE they would.beat-3pl the rye
“But I am totally sure that this was the thing that people used when husking
the rye.”
11. e pronoun a gente is semantically 1st person plural even when it does not induce 1st
person plural agreement morphology on the verb.
12. e doubler may also be a full DP, like in (i) below, with the same effect of blocking the
inclusive reading. Here the speaker is describing the traditional cheese making process, but he
is not himself a cheese maker.
(i) A minha mãe e os outros todos tirava-m-se aquele punhadozinho
the my mother and the others all would.take-3pl-SE that handful
de coalhada e depois iam espremendo aquilo… (CORDIAL-SIN.ˉMTM)
of curd and then went.3pl pressing that
“My mother and everybody else would take a portion of curd in their hands and
would then go on pressing the curd (until it was ready to pack into the mould).”
Subject doubling in European Portuguese dialects 19
In (16), it is the null 3rd person plural pronoun that induces an unambiguous
exclusive reading (the 3rd person plural features are visible on the verbal inec-
tion). e sentential subject refers to the people in town who used to catch birds
(but would not use birdlime for that purpose); the speaker previously made clear
that he does not belong to the group of local bird-catchers.
(16) Cá nunca se usara-m isso. (CORDIAL-SIN. LVR)
here never SE used-3pl that (the birdlime)
“People would never use that here.”
5. Accounting for the availability of subject doubling in EP dialects
In order to account for the facts described in the previous sections, I will rely on
the hypothesis that the doubling pronoun/DP and the subject clitic se belon g to t he
same constituent, a complex DP where the doubler is an adjunct modier of the
clitic se (the latter, a minimal/maximal category in the terms of Chomsky 1995).
(17) [DP DP [ se ] ] (cf. Uriagereka 1995; Kayne 2000, among others)
In this way the interpretative eects described in Section 4 can be smoothly accom-
modated and the syntactic features of the construction can be thoroughly derived
as well, as will be claried in a moment.
Although I am here following the ‘complex DP’ analysis devised by Uriagereka
(1995) and Kayne (2000) to account for clitic doubling, it should be noted that the
double subject impersonal se construction does not reduce to regular clitic dou-
bling. e crucial dierence is that in the latter there must be a secondary source for
Case assignment (see how the lack of the Case-assigning preposition a “to” makes
(18b) ungrammatical), whereas in the former there is no such requirement.
(18) a. Não foi ele que a beijou a ela, foi ela que o
not was he that her kissed to she, was she that him
beijou a ele
kissed to he
b. *Não foi ele que a beijou ela, foi ela que
not was he that her kissed she, was she that
o beijou ele
him kissed he
“It wasn’t him who kissed her, it was she who kissed him.”
Going back to the derivation of subject doubling in dialectal EP, aer the complex
DP is assembled (see (17)), it is merged in the structure as the external argument
19 Ana Maria Martins
of transitive and unergative verbs, or (less oen) as the internal argument of unac-
cusative verbs, and is assigned a theta-role in its merging position.13 e two
elements of the complex DP later follow separate routes. e clitic head-moves to
T and undergoes cliticization according to the general pattern of clitic placement in
EP. e adjunct modier, as part of the remnant DP, either stays in its base-generated
position or takes the regular path of EP subjects. In this way, we account for the word
order facts. As the complex DP splits, each one of its constitutive parts just behaves
as a regular clitic and a regular DP-subject with regard to movement operations.
e clitic se has number (plural) but does not have a person feature or Case.
ough it has semantic content, it is syntactically expletive-like and that is what
makes the double subject construction available. Because se is inert to enter the type
of agreement relation that underlies Case assignment (cf. Chomsky 2000, 2001),
it is the doubling pronoun/DP that induces subject-verb agreement and displays
nominative Case. e doubling pronoun/DP has a complete set of Ф-features. It
therefore values the Ф-features of T and has its Case feature valued.
In the Agree based system of Chomsky (2000, 2001), the feature person
endows a given probe with Case-valuation properties; hence, nite T can value
a given Case feature (as nominative), but a participial head, for example, which
does not have a person feature, cannot. I tentatively propose that, conversely, a
nominal without a person feature may be devoid of Case-licensing requirements.
is hypothesis could be implemented by taking a nominal without person to
be an incomplete DP or, to be more precise, an NP articulated with just some
functional structure (e.g. NumP). Crucially, lack of person would signal lack of
13. Double subject sentences with SE are found with transitive, unergative and unaccusative
verbs as exemplified respectively in (i), (ii) and (iii).
(i) Aˉgente sega-se esses olhos todos e deixa-se
pron.1pl cuts.off-SE those eyes all and leaves.SE
este só. (CORDIAL-SIN.PST)
this.one only
“(When it is time for the pruning), we remove all these sprouts leaving only this one.”
(ii) Mas, enfim, cá se vai a gente andando. (CORDIAL-SIN. AAL)
but aer.all here SE goes pron.1pl walking
“But nevertheless we keep going.”
(iii) A gente chega-se à noite, eu faço o
pron.1pl arrives.SE at.the night I do the
meu trabalho… (CORDIAL-SIN. PIC)
my work
“When we arrive home at night-time, I do my work…”
Subject doubling in European Portuguese dialects 191
the higher functional categories of the DP domain, in particular lack of D0, the
locus of referentiality and, hypothetically, the locus of Case (cf. Longobardi 1994;
Chomsky 2001, 2004).
As for the contrast between the EP dialects that allow subject doubling and
standard EP, it can be simply derived under the assumption that in standard EP the
impersonal pronoun se is not a decient pronoun, but a full referential expression
with a complete set of Ф-features and Case (cf. D’Alessandro 2004).
e current proposal thus converges with Holmberg & Nikanne (2006) in the
observation that the grammars that possess the right kind of decient pronoun are
expected to allow subject doubling.14
In the next section independent additional evidence will be oered to support
the contention that the pronoun se has the feature plural and no person feature in
the EP dialects where the double subject impersonal se construction is attested.
6. Further evidence supporting the analysis
A cluster of contrasts between EP dialects and standard EP with respect to the
behaviour of se can be easily handled under the approach developed in this paper.
is section will briey consider: adjectival agreement with se in predicative con-
texts, anaphoric binding, and the alternation between null and overt subjects in
expletive constructions.
6.1 e pronoun se is plural in dialectal EP: Adjectival agreement
Adjectival agreement in predicative contexts shows that in standard EP the imper-
sonal pronoun se has the number feature ‘singular’. us the adjective novo “young”
14. e subject doubling impersonal se construction is a grammatical option in dialectal EP
because it satisfies (a), (b)-2. and (c)-2. of (i):
(i) On the availability of doubling (Holmberg & Nikanne 2006):
(a) A category a doubles a DP b iff they share a theta-role.
(b) Doubling is possible iff
1. a and b are copies (the case of A and A’-movement);
2. a is deficient.
3. [there is an extra source for case assignment/valuation
(cf. Schoorlemmer 2006). Addition mine]
(c) Deficient pronouns include
1. pronominal clitics (the case of clitic doubling);
2. personless pronouns.
192 Ana Maria Martins
in (19) is not allowed to display plural inection, as the grammaticality contrast
between (19a) and (19b) proves.
Standard European Portuguese:
(19) a. Quando se é novo…
when SE is young.masc.sg
“When one is young…”
b. *Quando se é novos…
when SE is young.masc.pl
“When one is young…”
In sharp contrast with the standard variety, European Portuguese dialects allow
plural agreement between se and an adjectival predicate as illustrated by (20a–b).
(20) a. Na idade é que é; uma pessoa quando
in.the age is that is; a person when
se é novo-s, poder (CORDIAL-SIN. ALV)
SE is young.masc-pl can.INFIN
“ere is a right time for everything; when one is young, one must have the
strength to work hard.”
b. Não se andava calçado-s (CORDIAL-SIN. CDR)
not SE would.go.3sg with.shoes.on.the.feet.adj.masc-pl
“We wouldn’t go around with shoes on our feet (we would go barefooted).”
ese data conrm that dialectal se may dier from standard se as for the value
of its number feature. is is certainly so in the dialects where the double subject
impersonal se construction is attested. Hence, in this construction the doubling
pronoun/DP is always plural, like se itself.15 e plural number of se has further
grammatical consequences as will be discussed in the next section.
15. e conclusion that dialectal se is plural and therefore can only be doubled by a plural
pronoun/DP seems to be at odds with the fact that in sentences like (i) below the subject
pronoun a gente doubles impersonal se but at the same time shows third person singular
agreement with the verb:
(i) a. A gente chama-se rãs a isto. (CORDIAL-SIN. FLF)
pron.1pl call.3sg-SE rãs to that
“We call these ones rãs (frogs).”
In fact the pronoun a gente displays (obligatorily in standard EP and optionally in dialectal
EP) a mismatch between its semantic and syntactic features, being semantically (1st person)
plural but syntactically (3rd person) singular (cf. Costa & Pereira 2005). Crucially, it is the
semantic features that matter with respect to agreement between se and its doubler. is is
Subject doubling in European Portuguese dialects 193
6.2 e pronoun se is person-less in dialectal EP: Anaphoric binding
In a sub-group of the dialects where the double subject impersonal se construction
is found, anaphoric se (reexive or inherent) can be bound by a 1st person plural
pronoun (null or overt), as exemplied in (21). Such sentences are excluded in
standard EP (as in many EP dialects).
(21) a. No mar ainda às.vezes se orientamos
in.the sea still sometimes SE orientate.1pl
pela vaga. (CORDIAL-SIN. MLD)
by.the wave
“In the sea we are able to orientate ourselves by watching the waves.”
b. A gente casámos-se novos. (CORDIAL-SIN. GRC)
pron.1pl married.1pl-SE young.masc.pl
“We were young when we married each other”
e availability of sentences like (21) is expected under the analysis put forth in
this paper. In the relevant dialects, defective se can be anaphorically bound by any
plural pronoun/DP because it has the feature plural and no person feature.
confirmed by clitic doubling structures in standard EP (where the same type of big DP is
at stake – cf. Uriagereka 1995; Kayne 2000). Although in standard EP a gente always sets
subject-verb agreement in 3rd person singular (see (ii) below), it can double 1st person
plural pronoun nos “us” (see (iii)). e accusative pronoun nos “us” is the clitic correlate of
the strong pronoun nós “we”.NOM which only permits 1st person plural agreement with the
verb (see (iv)). us nos “us ” a nd a gente can be part of the same big DP because they match
in semantic features, although they do not match in syntactic features.
(ii) a. A gente chega tarde. (Standard EP)
pron.1pl arrive.3sg late
b. *A gente chegamos tarde.
pron.1pl arrive.1pl late
“We arrive late.”
(iii) Fazer-nos isto à gente ?! É incrível.
do.INFIN-us this to pron.1pl is unbelievable
“It’s unbelievable that he/she would do this to us.”
(iv) a. Nós chegamos tarde.
we arrive.1pl late
b. *Nós chega tarde.
we arrive.3sg late
“We arrive late.”
194 Ana Maria Martins
e fact that the number feature of dialectal se is plural also explains why it
cannot take a 1st or 2nd person singular pronoun as antecedent, as the ungram-
maticality of the sentences in (22) illustrates.16 (By the same token, se is never dou-
bled by a 1st or 2nd person singular nominative pronoun in the subject doubling
impersonal se construction).
(22) *Eu casei-se cedo
I married-SE early
“I was young when I got myself married.”
Se cannot have a 2nd person plural pronoun as antecedent either, because the rel-
evant dialects, like standard EP, lack a 2nd person plural pronoun as well as 2nd
person plural verbal inection.
ere appears to be a problem for this approach though. e question that
comes to mind is why the clitic nos (“us”, “ourselves”), specied for number (plural)
and person (1st), does not block the insertion of underspecied se. e answer lies
in the fact that the relevant dialects virtually replaced nós (“we”) with the new 1st
person plural pronoun a gente (a former DP: “the people”) and presumably also
replaced nos (“us”, “ourselves”) – see (23).17
(23) Consequences of the evanescence of nós/nos-1pl in some EP dialects:
STAGE 1 STAGE 2
a. Nós vimos-te → A gente vimos-te
we saw-you pron.1
pl
saw-you
b. Eles viram-nos → Eles viram a gente
they saw-us they saw pron.1
pl
c. (Nós) vimo-nos → (A gente) vimos-se
we saw-ourselves pron.1pli saw-SEi
na televisão na televisão
on TV on TV
e innovation a gente replaced the former 1st person plural pronoun nós/nos
“we/us”. But the pronoun a gente obeys Principle B of Binding eory (see (24))
and is therefore unable to act as a substitute for the 1st person plural anaphora
16. Note that all dialects also have the non deficient 3rd person se of standard EP.
17. e evidence given by the CORDIAL-SIN dialectal corpus is not as clear-cut as we would
want because the speakers of EP dialects also know standard EP to a certain degree and may
produce forms of both varieties in spontaneous speech (especially when involved in conversation
with standard EP speakers).
Subject doubling in European Portuguese dialects 195
(i.e. nos “ourselves”). e person-less se thus became the universal plural anaphora
in the relevant dialects (see (23c)).18
(24) a. *[Eu e o Pedro]i vimos [a gente]i na televisão
[I and the Pedro]i saw [pron.1pl]i in.the television
“I and Pedro saw ourselves on TV.”
b. [Eu e o Pedro]i dissemos que [aˉgente]i adorámosˉ a festa
[I and the Pedro]i said that [pron.1pl]i loved the party
“I and Pedro said that we loved the party.”
c. [A gente]i sabemos que a Maria criticou [a gente]i
[pron.1pl]i know that the Maria criticized [pron.1pl]i
“We know that Maria criticized us.”
For anaphoric se to be bound by a 1st person plural pronoun (overt or null) two
conditions must be fullled: (i) se must be plural and person-less; (ii) the 1st per-
son plural anaphora nos “ourselves” must not be available. Otherwise, the more
specied form (i.e. nos-1PL) would have preference over the underspecied one
(i.e. se-PL). We thus expect that some dialects allow the double subject impersonal
se construction but do not allow binding of se by a 1st person plural pronoun. Such
dialects full condition (i), they have a plural person-less se, but not condition (ii),
they have not lost nos “ourselves”. e data oered by the CORDIAL-SIN dialectal
corpus conrm this prediction.19
1. We could hypothesize that sentences (21a–b) would instantiate the impersonal subject
doubling construction but either the impersonal or the reflexive clitic se would have been
deleted because the sequence *se se is not allowed in European Portuguese (similarly to
most Romance varieties), as illustrated by (i) below – cf. Wanner (1977), Bonet (1995),
among others.
(i) *Orienta-se-se bem.
orientates-SE(impersonal)-SE(reflexive) well
“One knows how to take care of one’s life.”
is hypothesis faces two problems. First, in the dialects that do not have the impersonal
subject doubling construction the option of deleting one of the clitics se is not available.
Second, sentences like (21a–b) are not found in all the dialects that have the impersonal subject
doubling construction but only in a sub-group of those dialects. is is expected under the
view that the dialects that have a person-less plural se but did not lose the 1st person plural
pronoun nós-NOM/nos-ACC will not display sentences like (21a–b).
19. e double subject se construction can be found in some regions of the North of Portugal
but binding of anaphoric se by a 1st person plural pronoun is not found anywhere in the North.
196 Ana Maria Martins
6.3 Pure expletive se with impersonal predicates
e syntactic deciency of dialectal se usually does not extend to its semantic con-
tent. us Ф-incomplete impersonal se still denotes a group of humans. Some dia-
lects, however, reveal that the weakening of se that changed it into a person-less
pronoun can go a step further depriving se from semantic content. At that point
the semantic [+ hum] feature of se is lost and the clitic starts to appear as the exple-
tive subject of impersonal predicates, as the examples in (25) illustrate. Sentences
(25a) and (25c) are particularly revealing as they show that the overt expletive se
can freely alternate with a null expletive.
(25) a. Às.vezes acontece; já se aconteceu. (CORDIAL-SIN. PIC)
sometimes happens; already SE happened
“Sometimes it happens; it has already happened.”
b. Nunca se me aconteceu isso. (CORDIAL-SIN. CRV)
never SE me.dat happened.3sg that
“It never happened to me a thing like that.”
c. Chega a pontos custa-se caro. Custa caro uma
arrives to a.point costs-SE expensive. Costs expensive a
parelha a arrancar uma charrua. (CORDIAL-SIN. CPT)
pair.of.animals to pull a plough
“At a certain point it is really hard for the animals to go on pulling the
plough.”
d. Vou também ao mar apanhar uns peixes (...),
go.1sg also to.the sea catch some sh
quando se calha. (CORDIAL-SIN. PIC)
when SE happens
“I also go sh occasionally.”
e appearance of the pure expletive se is again restricted to the dialects where
the double subject impersonal se construction is found. It thus constitutes one
more indicator that in the relevant EP dialects the pronoun se gradually lost its
referential content.
Finally, it is worth noting that expletive se does not occur with weather-
predicates. is is exactly what is expected if the subject of weather-predicates
is a quasi-argument with a complete set of ϕ-features, Case, and a theta-role (see
Chomsky 1981:325, 1995:288; Vikner 1995).
Moreover, only a sub-region of the South of Portugal has se as the universal plural anaphora
while the double subject se construction is pervasive in southern dialects. Cf. Footnote 3.
Subject doubling in European Portuguese dialects 197
7. Consequences of the proposed analysis
is paper focuses on the subject doubling impersonal se construction which is
found in certain EP dialects but disallowed in standard EP. e main proposal in
the paper is that a nominal that lacks a person feature may also be devoid of Case.
is is how the clitic se of the relevant EP dialects is like; it has number (plural),
but no person or Case. Its decient nature (in the spirit of Holmberg & Nikanne
2006) is what makes subject doubling a grammatical option in dialectal EP. In a
few dialects, the expletive-like se dried in the direction of a pure expletive.
If the analysis put forth in the paper is on the right track, the subject doubling
se construction of European Portuguese dialects brings new empirical evidence
that supports the contentions (i) to (iii) below.
i. “If all other conditions are right, a given argument can be linked to, and
coherently composed with, more than one expression”. (Chung & Ladusaw
2004:118, on the extra object in Chamorro).20
ii. e impersonal se/si of standard EP, Spanish and Italian is not decient with
respect to ϕ-features and Case (cf. Cinque 1988), but a full referential expres-
sion (D’Alessandro 2004). erefore, subject doubling with impersonal se/si is
not an option in standard EP, Spanish and Italian.21
2. “Multiple linking, as we have just characterized it, is obviously not the norm. Typically,
the syntax and the semantics conspire to make available exactly one expression to be com-
posed with a targeted argument, so that the mapping between arguments and expressions
appears to be one to one.” (Chung & Ladusaw 2004:76).
21. A reviewer notes that a type of subject doubling is found in Italian (especially Tuscan
Italian) in sentences like (i), where noi “we” doubles impersonal si, but agreement between 1st
person plural noi and the verb is not allowed:
(i) a. Noi si mangia sempre alle 12.
we.nom SI eat.3sg always at 12
b. *Noi si mangiamo sempre alle 12.
we.nom SI eat.1pl always at 12
“ We always eat at twelve o’clock.”
Burzio (1986:81, Fn. 47) briefly comments on the construction illustrated by (ia): “SI appears
in non impersonal uses only dialectally (Tuscan dialects, mostly). In such cases SI has the
force of a first person plural pronominal, which sometimes appears overtly in addition to
SI, as in (Noi) si vorrebe vedere i nostri amici “(We) si would like to see our friends” (notice
that whereas nostri agrees with noi, the verb still fails to agree: a fact for which we have no
account)”.
If si in Tuscan Italian is a rst person plural pronominal as suggested by Burzio, it might
be similar to EP a gente in bearing different values for its semantic and syntactic features,
respectively 1st person plural and 3rd person singular (cf. Fn. 15). It is very revealing in this
19 Ana Maria Martins
iii. Overt pure expletives are not necessarily blocked when in competition with null
pure expletives in a particular grammar (cf. Chomsky’s 1981 Avoid Pronoun
Principle).
References
Bonet, Eulàlia. 1995. “Feature Structure of Romance Clitics”. Natural Language and Linguistic
eory 13. 607–647.
Burzio, Luigi. 1986. Italian Syntax. Dordrecht: Reidel.
Cardinaletti, Anna. 1997. “Subjects and Clause Structure”. e New Comparative Syntax ed. by
Liliane Haegeman, 33–63. London & New York: Longman.
Cardinaletti, Anna. 2004. “Toward a Cartography of Subject Positions”. e Structure of CP and
IP. e Cartography of Syntactic Structures ed. by Luigi Rizzi, vol. 2, 115–165. Oxford &
New York: Oxford University Press.
Chierchia, Gennaro. 1995. “e Variability of Impersonal Subjects”. Quantication in Natural
Language ed. by Emmon Bach, Eloise Jelinek, Angelika Kratzer & Barbara Partee, 107–143.
Dordrecht: Kluwer.
Chomsky, Noam. 1981. Lectures on Government and Binding. e Pisa Lectures. Dordrecht: Foris.
Chomsky, Noam. 1995. e Minimalist Program. Cambridge, Mass.: MIT Press.
Chomsky, Noam. 2000. “Minimalist Inquiries: e framework”. Step by Step: Essays on Minimalist
Syntax in Honor of Howard Lasnik ed. by Roger Martin, David Michaels & Juan Uriagereka,
89–155. Cambridge & London: MIT Press.
Chomsky, Noam. 2001. “Derivation by Phase”. Ken Hale: A Life in Language ed. by Michael
Kenstowicz, 1–52. Cambridge & London: MIT Press.
Chomsky, Noam. 2004. “Beyond Explanatory Adequacy”. Structures and Beyond (= e Cartography
of Syntactic Structures, Volume 3.) ed. by Adriana Belletti, 104–131. Oxford & New York:
Oxford University Press.
Chung, Sandra & William A. Ladusaw. 2004. Restriction and Saturation. Cambridge, Mass.: MIT
Press.
Cinque, Guglielmo. 1988. “On si Constructions and the eory of arb”. Linguistic Inquiry 19.
521–582.
respect that: “In certain Tuscan dialects most typically Fiorentino, si has virtually replaced the
ordinary 1st pers. PL. ending of the verb (*andiamo → si va)” (cf. Cinque 1988:55, Fn. 34 and
references therein). Again the parallelism with “a gente+3SG verbal form” competing with
“nós+1PL verbal form” in (European and Brazilian) Portuguese is striking.
e Tuscan subject doubling construction exemplified in (ia) appears to differ from
the subject doubling construction discussed in this paper in two respects: (i) si in Tuscan
Italian does not behave like a deficient (person-less & Case-less) pronoun; (ii) as far as I know,
in Tuscan Italian the doubling pronoun noi necessarily precedes the doubled si. On these
grounds, I tentatively take (ia) and similar sentences to be instantiations of the Hanging Topic
Le Dislocation construction, with the hanging topic noi escaping the Case Filter (because it
is base-generated in an A-bar position) and eventually receiving Default Case, which in Italian
is Nominative (cf. Schütze 2001; Legate 2008 on Default Case).
Subject doubling in European Portuguese dialects 199
Corver, Norbert & Jairo Nunes, eds. 2007. e Copy eory of Movement. Amsterdam & Philadelphia:
John Benjamins.
Costa, João & Sandra Pereira. 2005. “Phases and Autonomous Features: A case of mixed agreement”.
Perspectives on Phases (= MIT Working Papers in Linguistics, 49.) ed. by Martha McGinnis &
Norvin Richards, 115–124. Cambridge, Mass.: MIT Press.
Craenenbroeck, Jeroen & Marjo van Koppen. 2002. “Subject Doubling in Dutch Dialects”. Pro-
ceedings of Console 9 ed. by Marjo van Koppen, Erica ri, Erik Jan van der Torre & Malte
Zimmermann, 54–67. (Also published at http://www.sole.leidenuniv.nl/).
D’Alessandro, Roberta. 2004. Impersonal si Constructions. Agreement and interpretation.
Ph.D. dissertation, University of Stuttgart (= Electronic Dissertation Series, Stuttgart
University Library), http://elib.uni-stuttgart.de/opus/volltexte/2004/1630.
Dobrovie-Sorin, Carmen. 1998. “Impersonal se Constructions in Romance and the Passivization
of Unergatives”. Linguistic Inquiry 29. 399–437.
Holmberg, Anders & Urpo Nikanne. 2006. “Subject Doubling in Finnish: e role of decient
pronouns”. Edisyn – Papers of Workshop on Doubling 2006. (= Electronic Publication),
http://www.meertens.knaw.nl/projecten/edisyn/.
Kayne, Richard. 2000. “A Note on Clitic Doubling”. Parameters and Universals, 163–184. Oxford &
New York: Oxford University Press.
Koppen, Marjo van. 2007. “Agreement with (the Internal Structure of) Copies of Movement”. e
Copy eory of Movement ed. by Norbert Corver & Jairo Nunes, 327–350. Amsterdam &
Philadelphia: John Benjamins.
Legate, Julie Anne. 2008. “Morphological and Abstract Case”. Linguistic Inquiry 39. 55–101.
Longobardi, Giuseppe. 1994. “Reference and Proper Names: A theory of N-movement in syntax
and Logical Form”. Linguistic Inquiry 25. 609–665.
Lopes, Célia. 2003. A Inserção de a gente no Quadro Pronominal do Português. Frankfurt:
Iberoamerica – Vervuert.
Martins, Ana Maria. 1994. Clíticos na História do Português. Ph.D. dissertation, University of Lisbon.
Menuzzi, Sérgio. 2000. “First Person Plural Anaphora in Brazilian Portuguese: Chains and con-
straint interaction in binding”. Portuguese Syntax: Comparative Studies ed. by João Costa,
191–240. Oxford & New York: Oxford University Press.
Miguel, Elena de. 1999. “El Aspecto Léxico”. Gramática Descriptiva de la Lengua Española ed. by
Ignacio Bosque & Violeta Demonte, 2977–3060. Madrid: Espasa.
Nunes, Jairo. 2004. Linearization of Chains and Sideward Movement. Cambridge & London:
MIT Press.
Otero, Carlos Peregrín. 1999. “Pronombres Reexivos y Recíprocos”. Gramática Descriptiva de la
Lengua Española ed. by Ignacio Bosque & Violeta Demonte, 1427–1517. Madrid: Espasa.
Raposo, Eduardo & Juan Uriagereka. 1996. “Indenite Se”. Linguistic Inquiry 14. 749–810.
Rivero, María-Luisa. 2000. “On Impersonal Reexives in Romance and Slavic and Semantic Vari-
ation”. Paper presented at 30th Linguistic Symposium on Romance Languages, University of
Florida, Gainesville, February 2000.
Schoorlemmer, Erik. 2006. “Double Agreement in Complex Inversion”. Romance Languages
and Linguistic eory 2004 ed. by Jenny Doetjes & Paz González, 275–295. Amsterdam &
Philadelphia: John Benjamins.
Schütze, Carson T. 2001. “On the Nature of Default Case”. Syntax 4. 205–238.
Strozer, Judith R. 1976. Clitics in Spanish. Ph.D. dissertation, University of California at Los Angeles.
Strozer, Judith R. 1978. “On the So-called “Dative of Interest”. Hispania 61. 117–123.
Uriagereka, Juan. 1995. “Aspects of the Syntax of Clitic Placement in Western Romance”. Lin-
guistic Inquiry 26. 79–123.
2 Ana Maria Martins
Vikner, Sten. 1995. Verb Movement and Expletive Subjects in the Germanic Languages. Oxford &
New York: Oxford University Press.
Wanner, Dieter. 1977. “On the Order of Clitics in Italian”, Lingua 43. 101–128.
Zagona, Karen. 1996. “Compositionality of Aspect: Evidence from Spanish Aspectual Se”. Aspects
of Romance Linguistics. Washington D.C.: Georgetown University Press. 475–488.
