Content uploaded by Victoria Mateu
Author content
All content in this area was uploaded by Victoria Mateu on Dec 08, 2015
Content may be subject to copyright.
Isabelle Charnavel* and Victoria Mateu*
The clitic binding restriction revisited:
Evidence for antilogophoricity
DOI 10.1515/tlr-2015-0007
Abstract: In some Romance languages, including French and Spanish, there is an
interesting asymmetry concerning the behavior of isolated clitics and clitic clusters
with respect to coreference. In the French example Anne croit qu’on va la lui
recommander pour la promotion ‘Anna thinks that they will recommend her to
him for the promotion’, the accusative clitic la ‘her’in the embedded clause cannot
corefer with ‘Anne’when a dative clitic, lui ‘to him’, co-occurs in the cluster. The
only previous account of this constraint (Bhatt and Šimík 2009) attributes this to a
binding restriction. Based on new data disentangling binding and logophoricity, we
show that the generalization capturing the distribution of clitics clusters in French
and Spanish is the following: an accusative clitic cannot be clustered with a dative
clitic if the accusative clitic refers to a logophoric center and is read de se. We derive
this antilogophoricity effect from perspective conflicts, which we represent as
intervention effects in the presence of a single logophoric operator in the relevant
domain. This analysis furthermore provides a semantic motivation for intervention
effects that have been postulated for the Person-Case Constraint (PCC), which we
hypothesize also derives from perspective conflicts.
Keywords: clitics, logophoricity, Person-Case Constraint, French, Spanish
Some languages such as French and Spanish exhibit an interesting coreference
restriction in clitic clusters. Consider the sentences in (1)
1
and (2): in (1), the
accusative clitic la ‘her’in the embedded clause can refer to the subject of the
matrix clause, Anna. However, when the accusative clitic is clustered with a
dative clitic as in (2), coreference with Anna is impossible
*Corresponding authors: Isabelle Charnavel, Harvard University, Cambridge, MA 02138, USA,
E-mail: icharnavel@fas.harvard.edu
Victoria Mateu, University of California, Los Angeles, CA 90095, USA,
E-mail: victoriam@ucla.edu
1Unless otherwise noted, (a) examples correspond to French and (b) examples to Spanish
throughout the paper. Note that in French, extraposition will be used to specify the reference of
dative clitics. Abbreviations are standard: ACC: accusative, COND: conditional, DAT: dative, EXPL:
expletive, FUT: future, GEN: genitive, LOC: locative, NOM: nominative, PL: plural, PRS: present, PST:
past, REFL: reflexive, s.o.: someone, SG: singular, SBJV: subjunctive.
The Linguistic Review 2015; 32(4): 671–701
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
(1) a. Anne
i
croit qu’on va la
i/j
recommander au patron
Anna thinks that s.o. will ACC.3FSG recommend to.the boss
pour la promotion.
for the promotion
b. Ana
i
cree que la
i/j
recomendarán al jefe para el
Anna thinks that ACC.3FSG recommend.FUT.3PL to.the boss for the
ascenso.
promotion
‘Anna
i
thinks that they will recommend her
i/j
to the boss for the promotion.’
(2) a. Anne
i
croit qu’on va la
*i/j
lui
k
recommander,[au
Anna thinks that s.o. will ACC.3FSG DAT.3SG recommend to.the
patron]
k
,pour la promotion.
boss for the promotion
b. Ana
i
cree que se
k
la
*i/j
recomendarán [al jefe]
k
Anna thinks that DAT.3 ACC.3FSG recommend.FUT.3PL to.the boss
para el ascenso.
for the promotion
‘Anna
i
thinks that they will recommend her
*i/j
to him
k
–[the boss]
k
–for
the promotion.’
The only previous account of this constraint (Bhatt and Šimík 2009) attri-
butes it to binding: when an accusative clitic is clustered with a dative clitic, the
accusative clitic cannot be bound. Based on new data experimentally controlled,
we will instead show that the constraint is due to antilogophoricity. That is, the
antecedent of an accusative clitic in a double object clitic construction may not
be a logophoric center, i.e. a perspective center, or the one whose speech,
thoughts, or feelings are being reported. We will derive this restriction from
perspective conflicts, which will be analyzed as intervention effects in the
presence of a single logophoric operator in the relevant domain. Specifically,
we will argue that dative clitics generally occupy a position encoding ‘empathy’,
and that when an accusative clitic refers to an attitude holder and co-occurs in
the same cluster, this results in a conflict of perspectives. This analysis further-
more provides a semantic motivation for intervention effects that have been
postulated for the Person-Case-Constraint (PCC, Bonet 1991), which we will
hypothesize also derives from conflicting centers of perspective.
After providing more details about the empirical facts of this clitic cluster
restriction and its previous account (Section 1), we will lay out our experimental
study (Section 2), which will lead us to our analysis based on antilogophoricity
(Section 3).
672 Isabelle Charnavel and Victoria Mateu
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
1 Background
1.1 A deeper look into the data
The constraint on accusative clitics in clusters observed in (2) is not only found
in French and Spanish –the target languages of this paper –but also in Catalan,
Czech, and Serbo-Croatian among others (see Bhatt and Šimík 2009). It emerges
in specific conditions that concern both the dative clitic and the antecedent
of the accusative clitic. First, the constraint arises whether the dative clitic is a
goal (3), a benefactor (4) or a possessor (5).
2
(3) a. * Pierre
i
dit qu’on le
i
lui
k
a présenté,à[la
Peter says that s.o. ACC.3MSG DAT.3SG has introduced to the
Reine]
k
.
Queen
b. * Pedro
i
dice que se
k
lo
i
presentaron a [ la Reina]
k
.
Peter says that DAT.3 ACC.3MSG introduce.PST.3PL to the Queen
‘Peter
i
says that they introduced him
i
to her
k
–[the Queen]
k
.’
(4) a. * Pierre
i
dit qu’on le
i
lui
k
a peint,
Peter says that s.o. ACC.3MSG DAT.3SG has painted
(pour [la Reine]
k
).
for the Queen
b. * Pedro
i
dice que se
k
lo
i
pintaron (a [la Reina]
k
).
Peter says that DAT.3 ACC.3MSG paint.PST.3PL to the Queen
‘Peter
i
says that they painted him
i
for her
k
([the Queen]
k
).’
(5) a. * Pierre
i
craint qu’on ne le
i
lui
k
mette dans les
Peter fears that s.o. EXPL ACC.3MSG DAT.3SG put in the
bras (de [la nounou]
k
).
arms of the nanny
b. * Pedro
i
teme que se
k
lo
i
pongan en los brazos
Peter fears that DAT.3 ACC.3MSG put.SBJV.3PL in the arms
de [la niñera]
k
.
of the nanny
‘Peter
i
is afraid that they put him
i
into her
k
arms ([the nanny]
k
’s).’
2In some dialects of French, clitic clusters display the order dative > accusative. Based on
preliminary results, this does not seem to have an effect on the constraints stated in this section.
The clitic binding restriction revisited 673
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
But there is no restriction on the accusative clitic when the dative is a full
DP, as in (1), further exemplified in (6), a strong pronoun, as in the French
example in (7),
3
a locative clitic, as in the French example in (8),
4
or an ethical
dative, as in the Spanish example in (9).
5
(6) a. Pierre
i
dit qu’on l’
i
a présenté à la Reine.
Peter says that. s.o ACC.3SG has introduced to the Queen
b. Pedro
i
dice que lo
i
presentaron a la Reina.
Peter says that ACC.3MSG introduce.PST.3PL to the Queen
‘Peter
i
says that they introduced him
i
to the Queen.’
(7) Pierre
i
dit qu’on l’
i
a présenté à elle ( pas à lui
k
).
Peter says that s.o. ACC.3SG has introduced to her not to him
‘Peter
i
says that they introduced him
i
to HER (not HIM
k
).’
(8) Pierre
i
dit qu’on l’
i
y
k
a emmené,[au château]
k
.
Peter says that s.o. ACC.3MSG LOC has taken to.the castle
‘Peter
i
says that they took him
i
there
k
–to [the castle]
k
.’
(9) Jesús
i
dijo a [su madre]
k
que se
k
lo
i
vendieron a los
Jesus said to her mother that DAT.3SG ACC.3MSG sell.PST.3PL to the
enemigos.
enemies
‘Jesus
i
said to [her mother]
k
that they sold him
i
to the enemies on her
k
.’
Furthermore, the constraint is only observed when the accusative clitic has an
antecedent, such as Peter, in the same sentence, but not if it is in the previous
sentence (10) or utterance. The clitic cluster is also acceptable if Peter antecedes
the dative clitic instead of the accusative clitic, as in (11), or if neither clitic
corefers with the subject, as in (12).
3Clitic doubling of dative pronouns is obligatory in Spanish.
4There is no locative clitic in Spanish.
5Ethical datives are generally restricted to 1st and 2nd persons in French (Jouitteau and Rezac
2008, a.o.) and Spanish (Ormazabal and Romero 2007), and as we will see in Section 2.4. Results
and Discussion, there is no coreference restriction when the dative clitic is a 1st/2nd person in
any case. However, there are cases where we can find 3rd person ethical datives co-occurring
with other clitics in Spanish as shown in (9) (see Franco and Huidobro 2008). Crucially, ethical
dative clitics do not give rise to the constraint whichever person they are, and as we will see in
Section 3 Proposal, this is so because they do not occupy the same position as regular datives,
which encodes point of view, but a higher position outside the logophoric domain.
674 Isabelle Charnavel and Victoria Mateu
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
(10) a. (Voici Pierre
i
).On va le
i
lui
k
présenter,[à la Reine]
k
.
here.is Peter S.o. will ACC.3MSG DAT.3SG introduce to the Queen
b. (Este es Pedro
i
).Se
k
lo
i
presentarán [a la Reina]
k
.
(this is Peter) DAT.3 ACC.3MSG introduce.FUT.3PL to the Queen
‘(This is Peter
i
). They will introduce him
i
to her
k
–[the Queen]
k
.’
(11) a. Pierre
i
dit qu’on la
k
lui
i
a présentée,[la Reine]
k
.
Peter says that s.o. ACC.3FSG DAT.3SG has introduced the Queen
b. Pedro
i
dice que se
i
la
k
presentaron - [la Reina]
k
.
Peter says that DAT.3 ACC.3FSG introduce.PST.3PL the Queen
‘Peter
i
says that they introduced her
k
to him
i
–[the Queen]
k
.’
(12) a. Pierre
i
dit qu’on le
j
lui
k
a présenté,Luc
j
,à
Peter says that s.o. ACC.3MSG DAT.3SG has introduced Luke to
[la Reine]
k
.
the Queen
b. Pedro
i
dice que se
k
lo
j
presentaron,Lucas
j
a
Peter says that DAT.3 ACC.3MSG introduce.PST.3PL Luke to
[la Reine]
k
.
the Queen
‘Peter
i
says that they introduced him
j
to her
k
–Luke
j
to [the Queen]
k
.’
In sum, it seems that an accusative clitic cannot be anteceded by a DP in the
same sentence when it is clustered with a dative clitic that is a goal, a benefactor
or a possessor. It remains to be seen in more detail which conditions the
antecedent must fulfill to trigger the constraint. This is our main departure
from the previous analysis of this phenomenon, i.e. Bhatt and Šimík’s (2009).
1.2 Previous account
First observed by Roca (1992) and Ormazabal and Romero (2007), the phenom-
enon described above has received very little attention in the literature. The only
attempt –to our knowledge –to account for this constraint is that of Bhatt and
Šimík (2009), who crucially attribute it to a binding restriction, as defined in (13).
(13) Clitic Binding Restriction (CBR)
When a [third person] indirect object (IO) clitic and a direct object (DO)
clitic co-occur in a cluster, the DO clitic cannot be bound.
The clitic binding restriction revisited 675
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
Bhatt and Šimík (2009) derive this constraint from the Person Case Constraint
(see Adger and Harbour 2003; Albizu 1997; Anagnostopoulou 2003, 2005; Béjar
and Rezac 2003; Bianchi 2003; Bonet 1991, 1994; Ormazabal and Romero 2002,
inter alia), defined in (14) and instantiated in (15) vs. (16).
(14) Person Case Constraint (PCC, Strong version)
In a combination of a weak direct object and an indirect object [clitic,
agreement marker, or weak pronoun], the direct object has to be a third
person (Bonet 1991: 182).
(15) a. * Pierre me lui a recommandé.
Peter ACC.1SG DAT.3SG has recommended
b. * Pedro se me recomendó.
Peter DAT.3SG ACC.1SG recommended
‘Peter recommended me to him.’
(16) a. Pierre me l’a recommandé.
Peter DAT.1SG ACC.3SG has recommended
b. Pedro me lo recomendó.
Peter DAT.1SG ACC.3SG recommended
‘Peter recommended him to me.’
Bhatt and Šimík (2009) propose that third person clitics acquire features as a
result of variable binding, triggering PCC effects. More precisely, they adopt the
core idea of the standard hypotheses accounting for PCC, namely that third
person pronouns come in two forms –a featurally more specified variant and a
featurally underspecified variant –and PCC requires the accusative clitic to be
featurally underspecified. Under feature-checking approaches, this is so because
the presence of a structural intervener (dative clitic) blocks certain agree rela-
tions between the Probe and the accusative clitic (see Anagnostopoulou 2003,
2005; Adger and Harbour 2007; Heck and Richards 2007). Under hierarchy-
based approaches, the ungrammaticality comes from failure of alignment
between two hierarchies: the thematic/argument structure hierarchy and the
person hierarchy (see Rosen 1990; Haspelmath 2004).
Drawing on the idea that variable binding involves feature transmission (see
Kratzer 2009) and that grammar favors binding over coreference (see Reinhart’s
[1983] Rule I and Roelofsen’s [2010] Rule S), Bhatt and Šimík (2009) claim that
accusative third person pronouns, which usually lack inherent features of the
sort that PCC cares about, acquire these features as a result of variable binding,
thus triggering PCC effects. Specifically, they implement this idea assuming the
676 Isabelle Charnavel and Victoria Mateu
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
existence of two hierarchies: the person hierarchy H: {1 > 3, 2 > 3, Bound 3 > 3}
and the argument structure hierarchy: indirect object > direct object. Clitics have
to respect both hierarchies, which fails in the case at hand and gives rise to the
ungrammaticality we observe. For instance in (3), the accusative clitic le/lo
acquires features because it is bound by Peter. It is thus higher than the dative
clitic lui/se on the person hierarchy, but lower than it on the argument structure
hierarchy; this discrepancy causes the ungrammaticality of the clitic cluster.
This analysis predicts that binding of the dative clitic rescues a bound
accusative clitic. This is what Bhatt and Šimík (2009: 7) call the ‘weak CBR effect’,
which they observe in French in (17) (Bhatt and Šimík 2009: 8) and in Czech.
(17) Marie
i
est persuadée que Charles
j
a demandé que tu la
i
Mary is convinced that Charles has asked that you ACC.3FSG
lui
j
présentes.
DAT.3SG introduce.
‘Mary
i
is convinced that Charles
j
asked that you introduce her
i
to him
j
.’
In sum, under Bhatt and Šimík’s approach, binding of the accusative clitic is
crucially responsible for CBR effects, which derive from PCC.
2 Experimental study
The goal of our experimental study was to question Bhatt and Šimík’s (2009)
conclusion. Using a grammaticality judgment task we investigated the condi-
tions under which the restriction on clitic clusters arises, and in particular,
whether binding of the accusative clitic is the crucial factor. In what follows
we provide the details of our experimental study, the results of which argue
against this view.
2.1 Hypothesis
On close scrutiny, it appears that there is a non-negligible confound in the
examples that instantiate the clitic cluster constraint in the literature: they all
involve psychverbs and verbs of saying whose subjects typically have perspec-
tive over the sentential complement. We thus hypothesized that binding by itself
is not the relevant factor, but rather, antilogophoricity constraints on the accu-
sative clitic antecedent. That is, the antecedent of the accusative clitic cannot be
logophoric.
The clitic binding restriction revisited 677
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
The notion of logophor was coined by Hagège (1974) to designate certain
pronominal forms in West-African languages that appear in specific environ-
ments such as indirect discourse as illustrated in (18) for Ewe.
(18) a. Kofi bey yè-dzo.
Kofi say LOG-leave
‘Kofi
i
said that he
i
left.’
b. Kofi bey e-dzo.
Kofi say 3SG-leave
‘Kofi
i
said that (s)he
k
left.’
According to Clements (1975: 130), the antecedent of logophors must be the center of
perspective, i.e. “the one whose speech, thoughts, feelings, or general state of con-
sciousness are reported”. More precisely, Sells (1987) proposes to distinguish between
threetypesoflogophoricantecedentsasin(19),whichhavebeenreportedtobe
crucial for licensing long distance anaphors such as Mandarin Chinese ziji (Huang
and Liu 2001, inter alia)orIcelandicsig (Maling 1984, inter alia)amongothers.
(19) a. Source: the one who is the intentional agent of the communication.
b. Self: the one whose mental state or attitude the proposition describes.
c. Pivot: the one with respect to whose (time-space) location the content of
the proposition is evaluated.
We conjectured that the relevant notion of logophoric center for the clitic cluster
restriction is that of attitude holder, corresponding to Sells’(1987) Source and Self, i.e.
we hypothesized that the antecedent of the accusative clitic cannot be an attitude
holder.
To test our hypothesis against Bhatt and Šimík’s (2009), we constructed an
experiment disentangling the two crucial variables, viz., binding and logophori-
city. In other words, for the constraint to arise, should the antecedent bind the
accusative clitic or be an attitude holder?
2.2 Participants
A total of 97 adult French native speakers participated in the French version of this
study. They were all born and raised in France and were aged between 23 and 76
(M= 40.1). Additionally, 35 adult Spanish native speakers participated in the
Spanish version of this study. They were all born and raised in Spain or Mexico,
and were aged between 23 and 59 (M= 28.9).
678 Isabelle Charnavel and Victoria Mateu
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
2.3 Materials and method
Participants had to provide grammaticality judgments on a continuous scale
online. They were asked to click towards the right edge of the bar if the sentence
sounded natural, or towards the left edge of the bar of it did not sound natural.
Subjects were also asked to pay close attention to the reference of the pronouns,
indicated in parenthesis after the sentence. There were three training items
involving clitics, and six control items with full DPs as the indirect object. We
employed a between-subjects design so that no participant was presented with
both the test item and the corresponding control sentence. The task included
27 test items with clitic clusters, as well as 6 control items with an accusative
clitic and a full DP indirect object.
6
The test sentences were constructed around
the following four variables (and the control sentences around the first two, see
Table 1):
i. whether the accusative clitic has a c-commanding antecedent or a non-
commanding antecedent;
ii. whether the accusative clitic antecedent is a logophoric center or a non-
logophoric center;
iii. whether the dative clitic is a local first/second person or a non-local third
person;
iv. whether the dative clitic has a c-commanding antecedent or a non-com-
manding antecedent.
With respect to the second variable, we ensured that the antecedent was an
attitude holder by using intensional predicates (e.g. ‘think’,‘believe’) or inten-
sional expressions (e.g. ‘according to’,‘someone’s letter’). Concerning the last
variable, aimed to test for Bhatt and Šimík’s (2009) weak CBR effect, we guar-
anteed binding of the dative clitic by using quantifiers as in (20). Note that to
avoid lengthening the experiment unnecessarily, we only manipulated that
variable when relevant, namely, when the accusative clitic was also bound (by
a logophoric center).
7
6Since it was virtually impossible for naïve subjects to infer the purpose of the study and the
character of the different variables no fillers were included. Moreover, adding the necessary
number of fillers to hide the pattern of manipulation, i.e. 2:1, would have lengthened the
experiment significantly and unnecessarily.
7It is thus understood that only condition 3 (see Table 1) had a bound dative clitic. In the other
conditions the dative clitic was free.
The clitic binding restriction revisited 679
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
(20) a. [L’actrice]
i
a promis à [chaque figurant]
k
que les producteurs
The actress has promised to each extra that the producers
la
i
lui
k
présenteraient.
ACC.3FSG DAT.3SG introduce.COND.3PL
b. [La actriz]
i
prometió a [cada extra]
k
que los productores
The actress promised to each extra that the producers
se
k
la
i
presentarían.
DAT.3SG ACC.3FSG introduce.COND.3PL
‘[The actress]
i
promised to [each extra]
k
that the producers would intro-
duce her
i
to him
k
.’
2.4 Results and discussion
First, the results obtained from the French and Spanish grammaticality judgment
tasks confirm the clitic cluster effect as shown in Table 2: participants gave lower
scores in conditions 1, 3 and 6 as compared to the control sentences with a full DP as
indirect object. Paired-sample t-tests confirmed that this difference was statistically
significant for both French (p< 0.001) and Spanish (p<0.001).
Furthermore, the results in conditions 4 and 6 crucially show that it is
logophoricity, not binding, that is relevant for this restriction on clitic clusters
(contra Bhatt and Šimík 2009). As we predicted, CBR effects arise when the
antecedent of the accusative clitic is a logophoric center, even if it does not
c-command the accusative clitic (condition 6), as we see in examples (21)–(22)
Table 1: Test conditions of the grammaticality judgment task.
Condition C-commanding
antecedent
Logophoric center as
antecedent
Person of dative
clitic
yes yes
yes yes /
yes yes bound
yes no
yes no /
no yes
no yes /
no no
no no /
Control yes yes N/A
Control no yes N/A
680 Isabelle Charnavel and Victoria Mateu
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
where the boy and the prisoner are attitude holders, but do not bind the
accusative clitic.
(21) a. *D’après [l’enfant]
i
,les maîtresses vont le
i
lui
k
according.to the child the teachers will ACC.3MSG DAT.3SG
confier,à[l’assistante]
k
.
entrust to the assistant
b. * Según [el niño]
i
,las maestras se
k
lo
i
encomendarán
according.to the boy the teachers DAT.3 ACC.3MSG entrust.FUT.3PL
a [la asistenta]
k
.
to the assistant
‘According to [the child]
i
, the teachers will entrust him
i
to her
k
,–[the
assistant]
k
.’
(22) a. * La lettre [du prisonnier]
i
explique qu’on le
i
lui
k
a
the letter of.the prisoner explains that s.o. ACC.3MSG DAT.3SG has
livré sans preuve,[au juge]
k
.
handed without evidence to.the judge
b. * La carta [del prisionero]
i
explica que se
k
lo
i
entregaron
the letter of.the prisoner explains that DAT.3 ACC.3MSG hand.PST.3PL
[al juez]
k
sin pruebas.
to.the judge without evidence
‘[The prisoner]
i
’s letter explains that they handed him
i
over to him
k
–
[the judge]
k
–without evidence.’
Table 2: Results of the grammaticality judgment task (the asterisk indicates that the sentences
in that condition received a significantly lower score than the control sentences).
Condition C-commanding
antecedent
Logophoric center as
antecedent
Person of dative
clitic
Grammaticality
yes yes *
yes yes /OK
yes yes bound *
yes no OK
yes no /OK
no yes *
no yes /OK
no no OK
no no /OK
The clitic binding restriction revisited 681
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
Conversely, we do not observe CBR effects when the antecedent c-com-
mands the accusative clitic but is not a logophoric center (condition 4), as we
see in (23) where the antecedent is inanimate and thus cannot be a perspective
center, or in (24) where the antecedent is animate but is not an attitude holder.
Note that this last case confirms that Sells’(1987) category Pivot is not relevant
here. That is, pivot antecedents do not give rise to the constraint as opposed to
Source and Self (attitude holders).
(23) a. [Le paquet]
i
spécifie qu’il faut le
i
lui
k
remettre, [au
the package indicates that it must ACC.3MSG DAT.3SG give to.the
concierge]
k
.
doorman
b. [El paquete]
i
especifica que se
k
lo
i
entreguen [al
the package indicates that DAT.3 ACC.3MSG give SBJV.3PL to.the
portero]
k
.
doorman
‘[The package]
i
specifies that they should hand it
i
over to him
k
–[the
doorman]
k
.’
(24) a. [Le criminel]
i
s’est échappé avant qu’on ne le
i
lui
k
the criminal REFL is escaped before that s.o. EXPL ACC.3MSG DAT.3SG
livre,[au directeur]
k
hand to.the director
b. [El criminal]
i
huyó antes de que se
k
lo
i
entregaran
the criminal escaped before of that DAT.3 ACC.3MSG hand.SBJV.3PL
[al director]
k
to.the director
‘[The criminal]
i
escaped before they handed him
i
over to him
k
–[the
director]
k
.’
Furthermore, the results of condition 3 reveal that binding of the dative clitic
does not rescue a bound accusative pronoun as claimed by Bhatt and Šimík’s
(2009). Thus, the weak CBR effect is invalidated in French and Spanish. We
suspect that the reduction of CBR effects in sentences such as (17) may be due to
the fact that Charles is an intervening perspective center. That is, in the relevant
domain, Charles, not Mary, is a logophoric center. This indirectly supports our
antilogophoricity hypothesis.
Finally, we also do not observe CBR effects when the antecedent of the
accusative clitic is a logophoric center but the dative clitic is a first or second
person pronoun (conditions 2 and 7) as shown in (25). These cases will be
discussed in more detail in Section 3.5.1.
682 Isabelle Charnavel and Victoria Mateu
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
(25) a. [La petite fille]
i
espère qu’on va te la
i
confier.
the little girl hopes that s.o. will DAT.2SG ACC.3FSG entrust
b. [La niña pequeña]
i
espera que te la
i
entreguen a ti
the girl little hopes that DAT.2SG ACC.3FSG entrust.SBJV.3PL to you
‘[The little girl]
i
hopes that they will entrust her
i
to you.’
3 Proposal
By disentangling binding and logophoricity, our experimental study provides
evidence in support of the idea that accusative clitics are antilogophoric when
clustered with third person dative clitics. In view of these results, we propose to
replace Bhatt and Šimík’s (2009) CBR with the following generalization:
(26) Clitic Logophoric Restriction (CLR)
When a third person dative clitic and an accusative clitic co-occur in a
cluster, the accusative clitic cannot corefer with a logophoric center.
Further evidence in support of this hypothesis comes from parallel constructions
with other antilogophors, such as epithets, as well as the fact that CLR effects
only emerge when the pronoun is read de se. These are discussed in the
following two sections. Subsequently, we will present our proposal.
3.1 Antilogophoricity effects
Antilogophoricity effects have also been observed for epithets (Dubinsky and
Hamilton 1998) or certain French pronouns like en/y (Ruwet 1990) as illustrated
in (27) and (28) respectively.
(27) a. John
i
ran over a man who was trying to give [the idiot]
i
directions.
b. * John
i
told us of a man who was trying to give [the idiot]
i
directions.
(28) a. Emile
i
mérite que Sophie en
i
tombe amoureuse.
Emile deserves that Sophie GEN falls in.love
‘Emile
i
deserves it –that Sophie falls in love with him
i
.’
b. * Emile
i
espère que Sophie en
i
tombera amoureuse.
Emile hopes that Sophie GEN fall.FUT.3SG in.love
‘Emile
i
hopes that Sophie will fall in love with him
i
.’
In (27a), the epithet the idiot can refer to John, which is not logophoric since
it is the subject of the non-intensional predicate run over. By contrast, in (27b),
The clitic binding restriction revisited 683
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
John, which is the subject of the intensional verb tell, cannot antecede the idiot.
The same contrast holds with the French pronoun en in (28), which can refer to
the subject of the non-attitude verb mérite ‘deserves’, but not to the subject of
the attitude verb espère ‘hopes’.
Thus, we can use these other cases of antilogophoricity effects as diagnos-
tics. By comparing a structure with a clitic cluster and a parallel one with an
epithet, we can test whether the antecedent is a logophoric center. For instance,
examples in (29), which contain epithets, are just as ungrammatical as their
counterparts in (21) (repeated below), which involve clitic clusters. On the other
hand, (30) with an epithet is as acceptable as its counterpart with a clitic cluster,
(24) (repeated below). Crucially, the former cases, unlike the latter, involve
coreference with a logophoric center.
(21) a. * D’après [l’enfant]
i
,les maîtresses vont le
i
lui
k
according.to the child the teachers will ACC.3MSG DAT.3SG
confier,à[l’assistante]
k
.
entrust to the assistant
b. * Según [el niño]
i
,las maestras se
k
lo
i
encomendarán
according.to the boy the teachers DAT.3 ACC.3MSG entrust.FUT.3PL
a [la asistenta]
k
.
to the assistant
‘According to [the child]
i
, the teachers will entrust him
i
to her
k
,–[the
assistant]
k
.’
(29) a. *D’après [l’enfant]
i
,les maîtresses vont confier [le coquin]
i
according.to the child the teachers will entrust the brat
àl’assistante.
to the assistant
b. * Según [el niño]
i
,las maestras encomendarán [el mocoso]
i
according.to the boy the teachers entrust.FUT.3PL the brat
a la asistenta.
to the assistant
‘*According to [the child]
i
, the teachers will entrust [the brat]
i
to the
assistant.’
(24) a. [Le criminel]
i
s’est échappé avant qu’on ne le
i
lui
k
the criminal REFL is escaped before that s.o. EXPL ACC.3MSG DAT.3SG
livre, [au directeur]
k
.
hand to.the director
684 Isabelle Charnavel and Victoria Mateu
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
b. [El criminal]
i
huyó antes de que se
k
lo
i
entregaran
the criminal escaped before of that DAT.3 ACC.3MSG hand.SBJV.3PL
[al director]
k.
.
to.the director
‘[The criminal]
i
escaped before they handed him
i
over to him
k
–[the
director]
k.
’
(30) a. [Le criminel]
i
s’est échappé avant qu’on ne livre [le
the criminal REFL is escaped before that s.o. EXPL hand the
crétin]
i
au directeur.
bastard to.the director
b. [El criminal]
i
huyó antes de que entregaran [el cabrón]
i
the criminal escaped before of that hand.SBJV.3PL the bastard
al director.
to.the director
‘[The criminal]
i
escaped before the guards handed over [the bastard]
i
to
the director.’
This diagnostic allows us to make the right predictions for sentences
like (31) with relative clauses (cf. Bhatt and Šimík 2009: 3), where the
logophoric status of the antecedent is not obvious in the absence of a
standard attitude verb. The unavailability of the epithet in (32) confirms
that the ungrammaticality of the clitic cluster in (31) comes from the logo-
phoricity of John.
(31) a. * Jean
i
a vu la fille qui le
i
lui
k
a présenté, [au
John has seen the girl who ACC.3MSG DAT.3SG has introduced, to.the
Pape]
k
.
Pope
b. * Juan
i
vio a la chica que se
k
lo
i
presentó [al
John saw to the girl who DAT.3 ACC.3MSG introduced, to.the
Papa]
k
.
Pope
‘John
i
saw the girl who introduced him
i
to him
k
–[the Pope]
k
.’
(32) a. * Jean
i
a vu la fille qui a présenté [l’idiot]
i
au Pape.
John has seen the girl who has introduced the idiot to.the Pope
b. * Juan
i
vio a la chica que presentó [el idiota]
i
al Papa.
John saw to the girl who introduced the idiot to.the Pope
‘John
i
saw the girl who introduced [the idiot]
i
to the Pope.’
The clitic binding restriction revisited 685
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
3.2 De se readings
We further observe that CLR effects do not simply arise when the antecedent of
the accusative clitic is an attitude holder. More specifically, the accusative clitic
also has to be read de se (see Chierchia 1989). If we force a non-de se interpreta-
tion of the antecedent, CLR effects disappear, as exemplified in (33).
(33) An intern is participating in the assignment of all the interns for the summer.
Instead of names, the list has numbers. When asked where to assign intern
#1234567, the intern, who does not know it is her, suggests assigning that
intern to Dr. Edmonds:
a. [L’interne]
i
a suggéré qu’on la
i
lui
k
assigne,[au
the intern has suggested that s.o. ACC.3FSG DAT.3SG assign to.the
Dr. Edmonds]
k
.
Dr. Edmonds
b. [La interna]
i
sugirió que se
k
la
i
asignaran
the intern suggested that DAT.3 ACC.3FSG assign.SBJV.3PL
[al Dr. Edmonds]
k
.
to.the Dr. Edmonds
‘[The intern]
i
suggested that they should assign her
i
to him
k
,
–[Dr. Edmonds]
k
.’
This provides further evidence for antilogophoricity, since logophors are often
characterized as de se elements (Anand 2006; Huang and Liu 2001; Schlenker
2003).
8
This also means that CLR effects more precisely correspond to anti de se
effects.
3.3 Hypothesis: antilogophoricity effects derive from
perspective conflicts
We propose that the antilogophoricity effects responsible for CLR derive from
conflicts of perspectives. This phenomenon is found in Mandarin with respect to
the logophoric long-distance anaphor ziji: two instances of long distance ziji in a
single clause must corefer (Huang and Liu 2001: 7) as shown in (34).
8Some African logophors have nevertheless been shown not to require a de se reading (see
Pearson 2015).
686 Isabelle Charnavel and Victoria Mateu
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
(34) a. Zhangsan
i
renwei Lisi
k
zhidao [Wangwu
n
ba ziji
i
de shu
Zhangsan think Lisi know Wangwu BA self’s books
song-gei-le ziji
i
de pengyou]
gave-to-PFV self’s friends
b. Zhangsan
i
renwei Lisi
k
zhidao [Wangwu
n
ba ziji
k
de shu
Zhangsan think Lisi know Wangwu BA self’s books
song-gei-le ziji
k
de pengyou]
gave-to-PFV self’s friends
c. * Zhangsan
i
renwei Lisi
k
zhidao [Wangwu
n
ba ziji
i
de shu
Zhangsan think Lisi know Wangwu BA self’s books
song-gei-le ziji
k
de pengyou]
gave-to-PFV self’s friends
d. * Zhangsan
i
renwei Lisi
k
zhidao [Wangwu
n
ba ziji
k
de shu
Zhangsan think Lisi know Wangwu BA self’s books
song-gei-le ziji
i
de pengyou]
gave-to-PFV self’s friends
‘Zhangsan thinks that Lisi knows that Wangwu gave self’s books to
self’s friends.’
In the case of French and Spanish clitic clusters, we hypothesize that
perspective conflicts arise between accusative clitics referring to perspective
centers and dative clitics because dative clitics are inherently logophoric. In
other words, our proposal consists of two hypotheses: (i) dative clitics always
encode perspective; (ii) accusative and dative clitics belong to the same logo-
phoric domain.
The hypothesis that dative clitics occupy a position encoding point of view
is supported by several facts. First, dative clitics, as opposed to indirect object
full DPs and locative clitics, generally have to be animate, which is a necessary
condition for being a logophor.
9
This is the case when dative clitics are goals as
9In Spanish there exists a type of inanimate dative clitics, namely, possessor datives. In these
cases the dative can be inanimate “in so far as the theme can be(come) part of it”(Cuervo 2003:
69), as exemplified by Demonte’s (1995: 12) sentences below:
i. Le puse el mantel a la mesa.
DAT.3SG put.PRS.1SG the tablecloth to the table
‘I put the tablecloth on the table’
ii. * Le puse los platos a la mesa.
DAT.3SG put.PRS.1SG the dishes to the table
‘I put the dishes on the table’
The clitic binding restriction revisited 687
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
in (35) and (36), inalienable possessors as in (37), or benefactors as in (38) from
Kayne (1975: 137), where the grandfather has to be alive. Note also that this
animacy condition on dative clitics is reflected in Ormazabal and Romero’s
(2007) analysis of PCC in Spanish, which assigns an Animacy/Gender feature
to dative clitics.
(35) a. Jean {lui
i
/y
i
} a envoyé une lettre,{à Marie
i
/* à Barcelone
i
}.
John DAT.3SG/LOC has sent a letter to Maria/ to Barcelona
b. Juan le
i
envió una carta {a María
i
/* a Barcelona
i
}.
John DAT.3SG sent a letter to Mary / to Barcelona
‘John sent her
i
–Mary
i
–a letter.’/‘*John sent it
i
–Barcelona
i
–a letter.’
(36) a. Sarah {* lui/ y} {préfère/ compare} cette maison,à cette
Sarah DAT.3SG/LOC prefers/ compares this house to that
maison-là.
house-there
b. * Sara le {prefiere/ compara} esta casa a esa otra.
Sarah DAT.3SG prefers/ compares this house to that other
‘Sarah {prefers/compares} this house to it –that other (house).’
(37) a. Je lui ai marché sur {le pied /* la branche}.
IDAT.3SG have stepped on the foot/ the branch
b. Le pisé {el pie / * la rama}.
DAT.3SG step.PST.1SG the foot/ the branch
‘I stepped on {his foot/its branch}.’
(38) a. Il lui
i
achète cette pierre.tombale (à [son grand-père]
i
).
he DAT.3SG buy that tombstone to his grandfather
b. Le
i
compra esta lápida (a [su abuelo]
i
).
DAT.3SG buy.PRS.3SG that tombstone to his grandfather
‘He is buying that tombstone for him
i
(his grandfather)
i
.’
Furthermore, certain facts support the idea that more specifically, dative
clitics encode point of view. When entering into constructions with verbs requir-
ing dative case marking on the semantic subject (e.g. quirky displacement or
However, Cuervo argues that these should be analyzed as (static) possessors and not arguments
affected by the verb. Thus, they involve a different low applicative (Appl-AT) than recipient
datives (APPL-TO), which are the ones under study here.
688 Isabelle Charnavel and Victoria Mateu
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
quirky subjects) such as French plaire or Spanish gustar ‘please’in (39), dative
clitics refer to experiencers, which are perspective centers.
(39) a. Jean lui plaît,à Marie.
John DAT.3SG pleases to Mary
b. A María le gusta él.
to Mary DAT.3SG likes NOM.3SG
‘Mary likes him –John.’
When alternating with a locative clitic in French as in (40), the dative clitic
yields a more perspectival interpretation than the locative clitic y(Rooryck p.c.):
in the case with lui, Luke must personally feel part of the team, while in the case
with y, Luke’s team membership is simply an objective description.
(40) Luc {lui/ y} appartient,à cette équipe.
Luke DAT.3SG/LOC belongs to this team
‘Luke belongs to it –this team.’
Based on these observations, it seems reasonable to suppose that dative
clitics in French and Spanish must be inherently logophoric. Note that PCC
analyses based on feature checking make similar assumptions in specifying
dative clitics for person (e.g. Adger and Harbour’s [2003] [ ± participant] and
[ ± empathy] features; Anagnostopoulou’s [2003, 2005] [ ± person/participant]
feature; Boeckx’s [2000] [ ± person] feature; Reinhart’s [2000] [ ± mental state]
feature, inter alia) as opposed to accusative 3rd person clitics, which are
assumed to lack a [person/participant] feature altogether (Anagnostopoulou
2003; Adger and Harbour 2007). This is outlined in (41).
(41) 1st person: [+person/+participant]
2nd person: [+person/+participant]
3rd person dative: [–person/–participant]
3rd person accusative: -------
Japanese may give us an insight into the kind of logophoric center that
dative clitics can correspond to, namely so-called empathy locus. It has been
reported that some Japanese verbs such as the transferring verbs yaru and kureru
‘give’alternate depending on the viewpoint from which the event is described.
In the case of yaru, the event is described from the point of view of the referent
of the subject or the neutral point of view, while in the case of kureru, the event
is described from the point of view of the referent of the dative object. Thus, in
The clitic binding restriction revisited 689
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
(42a), kureru, as opposed to yaru, is incompatible with a nominative first person:
the speaker, from whose perspective the event is described, must be expressed
by a dative pronoun in the case of kureru, as in (42a), but by a nominative
pronoun in the case of yaru, as in (42b).
(42) a. Boku ga Hanako ni okane o {*kure-ru/ya-ru}
INOM Hanako DAT money ACC give-PRS
‘I give money to Hanako.’
b. Taroo ga boku ni okane o {kure-ru/*ya-ru}.
Taroo NOM me DAT money ACC give-PRS
‘Taroo gives me money.’from Kuno (1987: 246)
In cases involving kureru, the dative is characterized as an empathy locus,
i.e. the event participant with whom the speaker empathizes/identifies (see
Kuno 1987; Oshima 2007). Similarly, we hypothesize that dative clitics in
French and Spanish can be empathy loci, i.e. they may refer to a specific type
of logophoric center different from attitude holders. This does not mean that
dative clitics cannot refer to attitude holders: empathic elements are in fact
compatible with attitudinal interpretations as illustrated in (11); this simply
means that attitude holders and empathy loci are two types of logophoric
centers (see Charnavel 2014), and dative clitics must be one of them.
Recall that crucially, CLR effects only arise with dative clitics, not with full
DPs as indirect objects. Under this approach, this implies that indirect object full
DPs are not necessarily logophoric as opposed to dative clitics. In fact, they are
not subject to the animacy requirement as illustrated in (43) and (44) to be
compared to (35) and (36) respectively. The same holds for the counterparts of
(37) and (38).
(43) a. Jean a envoyé une lettre à Barcelone.
John has sent a letter to Barcelona
b. Juan envió una carta a Barcelona.
John sent a letter to Barcelona
‘John sent a letter to Barcelona.’
(44) a. Sarah {préfère/ compare} cette maison-ci à cette maison-là.
Sarah prefers/ compares this house to that house-there
b. Sara {prefiere/ compara} esa casa a esa otra.
Sarah prefers/ compares this house to that other
‘Sarah {prefers/compares} this house to that other (house).’
690 Isabelle Charnavel and Victoria Mateu
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
This parallels the contrast between ditransitive constructions and double
object constructions in English (cf. Baker 1996; Stowell 1981). That is, inanimate
indirect objects can only participate in the former, not in the latter as exempli-
fied in (45).
(45) a. John sent a letter to Barcelona.
b. *John sent Barcelona a letter.
Based on this observation and drawing on Anagnostopoulou’s (2003) ana-
lysis, we hypothesize that only clitic constructions (as opposed to ditransitive
constructions with full DPs) qualify as double object constructions in Romance
languages, in the sense of including an applicative head (vAppl; Marantz 1993).
Arguments for this distinction –besides animacy –include possibilities of nomi-
nalization and possibilities of movement in raising and passive constructions
(see Anagnostopoulou 2003 for details; see also Demonte 1995 for arguments
specific to Spanish). In sum, our hypothesis is that dative clitics occupy a
position encoding empathy, because they enter into a double object construc-
tion, while indirect object full DPs appear in a different construction that does
not force an empathy interpretation.
The second component of our hypothesis that CLR effects derive
from perspective conflicts is that dative and accusative clitics belong to the
same logophoric domain. Assuming that there is only one logophoric center in
the relevant domain (Koopman and Sportiche 1989; Huang and Liu 2001;
Sundaresan 2012), the attitude holder (accusative clitic) and the empathy
locus (dative clitic) compete as logophoric centers, thus yielding ungrammati-
cality. However, note that this domain must exclude the subject, since a logo-
phoric accusative clitic clustered with a subject clitic does not trigger CLR as
shown in (46).
(46) a. Pierre
i
dit qu’il
i
lui
k
a présenté son fils,
Peter says that NOM.3SG DAT.3SG has introduced his son
à [ la Reine]
k
.
to the Queen.
b. Pedro
i
dice que él
i
le
k
presentó su hijo a [la Reina]
k
.
Peter says that NOM.3SG DAT.3SG introduced his son to the Queen.
‘Peter
i
says that he
i
introduced his son to her
k
–[the Queen]
k
.’
The domain we consider as relevant in French and Spanish is thus the one
represented in (47).
The clitic binding restriction revisited 691
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
(47) a. French (see Sportiche’s French clitic template [1996: 5])
NOM [3.ACC 3.DAT]
il/elle le/la lui
b. Spanish
NOM [3.DAT 3.ACC]
él/ella se/le lo/la
3.4 Person case constraint
This hypothesis provides us a way to derive PCC from a semantic constraint.
Instead of deriving CLR from PCC as Bhatt and Šimík (2009) do, we assume that
PCC, like CLR, derives from a ban on several conflicting centers of perspective in
the same domain. Strikingly, PCC directly arises if we follow Kuno’s (1987) direct
discourse representation hypothesis. If we transpose a sentence that violates
CLR into direct discourse, we observe PCC effects. Compare sentence (2),
repeated below, with sentence (48). In (48), a perspective conflict arises between
the inherently logophoric dative lui/se and the speaker me, a discourse partici-
pant, which is inherently a perspective center.
(2) a. * Anne
i
croit qu’on va la
i
lui
k
recommander,[au
Anna
i
thinks that s.o. will ACC.3FSG DAT.3SG recommend to.the
patron]
k
,pour la promotion.
boss for the promotion
b. * Ana
i
cree que se
k
la
j
recomendarán [al jefe]
k
Anna
i
thinks that DAT.3 ACC.3FSG recommend.FUT.3PL to.the boss
para el ascenso.
for the promotion
‘Anna
i
thinks that they will recommend her
*i/j
to him
k
–[the boss]
k
–
for the promotion.’
(48) a. Anne pense: “* on va me lui recommander […]”.
Anna thinks: s.o. will ACC.1SG DAT.3SG recommend
b. Ana piensa: “* se me recomendarán […]”.
Anna thinks: DAT.3SG ACC.1SG recommend.FUT.3PL
‘Anna thinks: “they will recommend me to him.”’
Further evidence for this hypothesis comes from the fact that PCC can be
overridden when the first person is not read de se. This is possible in the case of
dream reports, such as the one in example (49).
692 Isabelle Charnavel and Victoria Mateu
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
(49) a. ? J
i
’ai rêvé que j’étais Marilyn Monroe
m
,que j’étais
I have dreamed that I was Marilyn Monroe that I was
chez Kennedy
k
et que je
m
me
i
lui
k
présentais.
house Kennedy and that I ACC.1SG DAT.3SG introduced
b. ? Yo
i
soñé que era Marilyn Monroe
m
,que estaba en
I dreamed that be.PST.1SG Marilyn Monroe that be.PST.1SG in
casa de Kennedy
k
y que se
k
me
i
presentaba.
house of Kennedy and that DAT.3SG ACC.1SG introduce.PST.1SG
‘I
i
dreamed that I was M. Monroe
m
, that I was at Kennedy
k
’s house
and that I
m
introduced me
i
to him
k
.’
The relevant logophoric domain for clitics can thus be specified as follows.
(50) a. French
NOM [1/2.DAT 3.ACC 3.DAT]
il/elle/je me/te le/la lui
b. Spanish
NOM [1/2.DAT 3.DAT 3.ACC]
él/ella/yo me/te se/le lo/la
3.5 Analysis
3.5.1 First/second person dative clitics
The discussion above suggests that any first/second person clitic gives rise to a
perspective conflict when clustered with a logophoric clitic. But recall conditions
2 and 7 of our experimental study: there is no CLR effect when the antecedent of
the accusative clitic is a logophoric center and the dative clitic is a first or
second person pronoun as shown in (25), repeated below.
(25) a. [La petite fille]
i
espère qu’on va te la
i
confier.
the little girl hopes that s.o. will DAT.2SG ACC.3FSG entrust
b. [La niña pequeña]
i
espera que te la
i
entreguen a ti.
the girl little hopes that DAT.2SG ACC.3FSG entrust.SBJV.3PL to you
‘[The little girl]
i
hopes that they entrust her
i
to you.’
Furthermore, since French and Spanish have the strong version of PCC as
explained in (51), this means that the correlation between PCC and CLR is
The clitic binding restriction revisited 693
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
imperfect as observed by Bhatt and Šimík (2009): transposing (25) into a direct
discourse yields the ungrammatical sentence (52) showing a PCC effect.
(51) a. Strong version of PCC: *1/2.Accusative.Clitic Dative.Clitic
b. Weak version of PCC: *1/2.Accusative.Clitic 3.Dative.Clitic
(52) a. La petite fille pense: “*on va te me confier.”
the little girl thinks s.o. will DAT.2SG ACC.1SG entrust
b. La niña pequeña piensa: “*te me entregarán a ti.”
the girl little thinks DAT.2SG ACC.1SG entrust.FUT.3PL to you
‘The little girl thinks: “they will entrust me to you”.’
To account for this, we assume that all interactions between logophoric centers
are not equal. Namely, the different types of logophoric centers form a hierarchy
and only two adjacent elements on the hierarchy create perspective conflicts.
3.5.2 Hierarchy of logophoric centers
First, we suppose the classification of logophoric centers in (53) (cf. Charnavel 2014).
(53) a. Discourse participant: The speaker and addressee of the actual dis-
course, i.e. 1/2 person clitics.
b. Empathy locus: The event participant with whom the speaker
empathizes or identifies (see Kuno 1987; Oshima 2007), e.g. 3rd person
dative clitics. It involves direct integration of perspective.
c. Attitude holder: The event participant whose discourse or thoughts are
being reported, e.g. 3rd person accusative clitics read de se. It involves
indirect integration of perspective.
We moreover hypothesize a hierarchy of logophoric centers based on the degree
of perspective integration in the discourse as in (54).
(54) discourse participant >empathy locus >attitude holder
Discourse participants and empathy loci both involve the speaker (directly or by
identification) while empathy loci and attitude holders both involve a perspec-
tive center different from the speaker (implicitly or explicitly). Note that these
roles are not exclusive: elements that are intrinsically high logophoric centers on
the hierarchy can be used as lower logophoric centers. For instance, dative
694 Isabelle Charnavel and Victoria Mateu
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
clitics, which are inherently empathy loci, can also refer to attitude holders as
in (11).
Antilogophoricity effects emerge, we propose, when two identical or adja-
cent logophoric centers on this hierarchy co-occur in the domain represented in
(50). Specifically, CLR effects arise when an empathy locus (third person dative
clitic) and an attitude holder (third person accusative clitic read de se) appear in
the same domain, and PCC effects emerge when a discourse participant (first/
second person clitics) and an empathy locus (third person dative clitic) co-occur.
Table 3 summarizes the various possibilities of clitic combinations correctly
predicted by our analysis.
2.5.3 Further issue: reflexives
Reflexives pattern exactly like first and second person clitics (cf. Kayne 2000,
who shows that they belong to the same morphological class). In first place, they
trigger PCC effects when clustered with a dative clitic (see Bonet 1991) whether
third person as in (55) (weak PCC) or first/second person as in (56) (strong PCC).
(55) a. * L’interne se lui assigne.
the intern REFL DAT.3SG assigns
b. * La interna se le asigna.
the intern REFL DAT.3SG assigns
‘The intern assigns herself to him.’
(56) a. * L’interne se m’assigne.
the intern REFL DAT.1SG assigns
b. * La interna se me asigna.
the intern REFL DAT.1SG assigns
‘The intern assigns {herself to me/me to herself}.’
Table 3: Grammaticality of clitic combinations in French and Spanish.
Predictions Grammaticality Violation
Logophoric centers in the domain French Spanish
* discourse participant +discourse participant *me/te me/te *me/te me/te PCC
* discourse participant +empathy locus *me/te lui *se me/te PCC
* empathy locus +empathy locus *me/te lui *se me/te PCC
* attitude holder (read de se)+empathy locus *le lui *se lo CLR
* attitude holder (read de se)+attitude holder *le lui *se lo CLR
discourse participant +attitude holder (read de se) me/te le me/te lo
The clitic binding restriction revisited 695
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
Secondly, dative reflexive clitics do not trigger CLR effects as exemplified in (57).
(57) a. [L’interne]
i
a suggéré que le Dr. Edmonds
k
se
k
l’
i
the intern has suggested that the Dr. Edmonds REFL ACC.3FSG
assigne.
assign.SBJV.3SG
b. [La interna]
i
sugirió que el Dr. Edmonds
k
se
k
la
i
the intern suggested that the Dr. Edmonds REFL ACC.3FSG
asignara.
assign.SBJV.3PL
‘[The intern]
i
suggested that Dr. Edmonds
k
assigns her
i
to himself
k
.’
And lastly, they occur in the same position as first and second person clitics
relatively to other clitics.
(58) a. French
NOM [1/2.DAT/REFL 3.ACC 3.DAT]
il/elle/je me/te/se le/la lui
b. Spanish
NOM [1/2.DAT/REFL 3.DAT 3.ACC]
él/ella/yo me/te/se se/le lo/la
This leads us to hypothesize that reflexives behave like discourse participants
with respect to logophoricity. This is supported by certain facts suggesting that
reflexives have a first person component. Comparable to Mandarin ziji (see
Huang and Liu 2001), reflexives in French and Spanish are speaker-inclusive
when there is no antecedent as illustrated in (59).
(59) a. Se remettre en question est souvent une bonne idée.
REFL.3SG to.question in question is often a good idea
b. Cuestionarse-se es a.menudo una buena idea.
to.question-REFL.3SG is often a good idea
‘To question oneself is often a good idea.’
Similarly, the long distance reflexive French soi ‘oneself’must be speaker-
inclusive as shown in (60) (see Moltmann [2006]’s analysis of one(self) as a
first-person generic pronoun).
696 Isabelle Charnavel and Victoria Mateu
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
(60) a. On
i
pense souvent que les étrangers ont peur de soi
i
.
s.o. thinks often that the foreigners are scared of oneself
‘People
i
often think that foreigners are afraid of them
i
.’
b. * Là-bas,on
i
pense que les étrangers ont peur de soi
i
.
there s.o. thinks that the foreigners are scared of oneself
‘Over there, they
i
think that foreigners are afraid of them
i
.’
3.6 Implementation
We propose to represent perspective conflicts as intervention effects due to
Closest Attract/Agree (Chomsky 1995, 1998). A logophoric center intervenes
when another logophoric center occurs in the same domain, because both
enter a feature-checking relation with one and the same logophoric operator.
More specifically, intervention effects arise when two logophoric centers share
the same feature(s).
This implementation first requires the existence of logophoric operators
instantiating logophoric centers. As proposed by Koopman and Sportiche
(1989), Anand (2006) and Sundaresan (2012), logophoric operators are similar
to silent pronouns that are coreferential (or in a relation of non-obligatory
control) with the antecedent and bind logophoric elements as represented in
(61), thereby triggering de se readings. According to Anand (2006), this is the
case because the operator is in the immediate complement of a referential item
that denotes the de se center.
(61) Antecedent
i
OPLOG
i
XLOG
i
|
———————————————————
|
Moreover, like Koopman and Sportiche (1989), and Sundaresan (2012), we
hypothesize that there is at most one logophoric operator c-commanding the
relevant domain, as represented in (62).
(62) a. French
NOM OPLOG [1/2.DAT/REFL 3.ACC 3.DAT]
il/elle/je me/te/se le/la lui
b. Spanish
NOM OPLOG [1/2.DAT/REFL 3.DAT 3.ACC]
él/ella/yo me/te/se se/le lo/la
The clitic binding restriction revisited 697
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
Assuming a feature-checking mechanism between interpretable features on
logophoric elements and uninterpretable features on logophoric operators, two
logophoric elements sharing the same feature(s) give rise to an intervention
effect. To explain why two adjacent or identical logophoric centers on the
hierarchy share features, we assume the system of features in (63) to be linked
to the properties of logophoric centers described in (53).
(63) a. Discourse participant: [A,B]
b. Empathy locus: [B,C]
c. Attitude holder: [C]
The [B] feature shared by discourse participants and empathy loci expresses the
speaker component crucial to both cases. Discourse participants are directly
defined by the speaker, and empathy loci are participants with whom the
speaker identifies. The [C] feature common to empathy loci and attitude holders
corresponds to perspectival distance from the speaker: both –implicitly or
explicitly –involve a perspective center different from the speaker.
The combination of these ingredients –single logophoric operator in the
relevant domain, sharing of features by identical or adjacent logophoric centers
on the hierarchy –correctly predicts intervention effects only in the PCC and
CLR configurations.
(64) a. *OPLOG
[A,B,C]
[la
C
lui
[B,C]
] (CLR)
b. *OPLOG
[A,B,C]
[me
[A,B]
lui
[B,C]
] (PCC)
c. OPLOG
[A,B,C]
[me
[A,B]
la
C
]
4 Conclusion
Based on new data disentangling binding and logophoricity, we have shown
that the generalization capturing the distribution of clitics clusters in French and
Spanish is the following: an accusative clitic cannot be clustered with a dative
clitic if it refers to a logophoric center and is read de se. We derive this
antilogophoricity effect from perspective conflicts, which we represent as inter-
vention effects arising in a specific domain. This requires distinguishing
between different types of logophoric centers (i.e. discourse participant, empa-
thy locus, attitude holder) and ranking them on a hierarchy in which only two
identical or adjacent elements compete when co-occurring in the same domain.
This analysis furthermore provides a semantic motivation for intervention effects
698 Isabelle Charnavel and Victoria Mateu
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
that have been postulated for PCC –we hypothesize that PCC also derives from
perspective conflicts. More generally, this predicts that intervention effects can
arise when the relevant types of logophoric elements co-occur. Future work
should examine how this can shed light on the behavior of long distance
reflexives and other logophoric elements.
Acknowledgment: For their comments, suggestions and feedback, we are grate-
ful to two anonymous reviewers and the audiences of the UCLA Syntax/
Semantics Seminar, WCCFL 32 and GLOW 37; a special thanks to Dominique
Sportiche for very helpful discussion on this project. We are also greatly
indebted to all the participants who completed our linguistic survey.
Funding: This work was supported in part by the National Science Foundation
under grants 1424054 and 1424336.
References
Adger, David & Daniel Harbour. 2003. The syntax and syncretisms of the Person Case
Constraint. Ms. Queen Mary, University of London and MIT.
Adger, David & Daniel Harbour. 2007. Syntax and syncretisms of the Person Case Constraint.
Syntax 10. 2–37.
Albizu, Pablo. 1997. The syntax of person agreement. Los Angeles, CA: USC Dissertation.
Anagnostopoulou, Elena. 2002. Case 17. Clitic Doubling. SYNCOM project.
Anagnostopoulou, Elena. 2003. The syntax of ditransitives. Evidence from clitics. Berlin: Mouton
de Gruyter.
Anagnostopoulou, Elena. 2005. Strong and weak person restrictions. In Heggie, Lorie &
Francisco Ordóñez (eds.), Clitic and affix combinations –Theoretical perspectives, 199–
235. Amsterdam. John Benjamins.
Anand, Pranav. 2006. De de se. Cambridge, MA: MIT dissertation.
Baker, Mark. 1996. The polysynthesis parameter. Oxford: Oxford University Press.
Béjar, Susana & Milan Rezac, 2003. Person licensing and the derivation of PCC effects. In Ana
Teresa Pérez-Leroux & Yves Roberge (eds.), Romance linguistics,49–62. Amsterdam: John
Benjamins.
Bianchi, Valentina. 2003. On the syntax of personal arguments. Paper presented at the XXIX
Incontro di Grammatica Generativa, Urbino.
Bhatt, Rajesh & Radek Šimík. 2009. Variable binding and the Person-Case Constraint. 25th
Annual Meeting of the Israel Association for Theoretical Linguistics (IATL 25). Ben Gurion,
University of the Negev.
Blake, Barry J. 1990. Relational grammar. London: Routledge.
Boeckx, Cedric. 2000. Quirky agreement. Studia Linguistica 54. 354–380.
Bonet, Eulàlia. 1991. Morphology after syntax. Pronominal clitics in Romance. Cambridge, MA:
MIT dissertation.
The clitic binding restriction revisited 699
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
Bonet, Eulàlia. 1994. The Person-Case Constraint. A morphological approach. In Harley, Heidi &
Colin Phillips (eds.), MIT working papers in linguistics 22. The morphology-syntax con-
nection,33–52. Cambridge, MA: MITWPL.
Charnavel, Isabelle. 2014. Exempt anaphors and logophoricity in French. Ms. Harvard
University. lingbuzz/002683.
Chierchia, Gennaro. 1989. Anaphora and attitudes de se. In Bartsch, Renate, Johan van Benthem, &
Peter van Emde Boas (eds.), Semantics and contextual expressions,1–31. Dordrecht: Foris.
Chomsky, Noam. 1995. The minimalist program. Cambridge, MA: MIT Press.
Chomsky, Noam. 1998 [2000]. Minimalist inquiries, Ms., MIT. Published as ‘Minimalist
Inquiries: The framework’. In Martin, Roger, David Michaels, & Juan Uriagereka (eds.), Step
by Step. Cambridge, MA: MIT Press, 89–156.
Clements, George N. 1975. The Logophoric pronoun in Ewe. Its role in discourse. Journal of West
African Languages 10. 141–177.
Cuervo, María Cristina. 2003. Datives at large. Cambridge, MA: MIT dissertation.
Demonte, Violeta. 1995. Dative alternation in Spanish. Probus 7(1). 5–30.
Dubinsky, Stanley W. & Robert Hamilton. 1998. Epithets as antilogophoric pronouns. Linguistic
Inquiry 29(4). 685–693.
Franco, Jon & Susana Huidobro. 2008. Ethical datives, clitic doubling and the theory of pro.In
Joyce Bruhn de Garavito, & Elena Valenzuela (eds.) Proceedings of the 10th Hispanic
Linguistics Symposium. Somerville, MA: Cascadilla Proceedings.
Hagège, Claude. 1974. Les pronoms logophoriques [logophoric pronouns]. Bulletin de la Société
Linguistique de Paris 69. 287–310.
Haspelmath, Martin. 2004. Explaining the ditransitive person-role constraint: A usage-based
approach, elanguage.net/journals/constructions/article/download/3073/3052 (accessed
5 October 2013).
Heck, Fabian & Marc Richards, 2007. A probe-goal approach to agreement and incorporation
restrictions in Southern Tiwa. In Trommer, Jochen & Andreas Opitz (eds.), 1 2 many: One-
to-many relations in grammar, 205–239. University Leipzig.
Huang, C.-T. James & Cheng-Sheng Luther Liu. 2001. Logophoricity, attitudes and ziji at the
interface. In Cole, Peter, Gabriella Hermon & C.-T. James Huang (eds.), Long distance
reflexives,syntax and semantics 33. 141–195. New York: Academic Press.
Jouitteau, Mélanie & Milan Rezac, 2008. The French ethical dative, 13 syntactic tests. Bucharest
Working Papers in Linguistics IX(1). 97–108.
Kayne, Richard. 1975: French syntax: the transformational cycle. Cambridge, MA: MIT Press.
Kayne, Richard. 2000. Person morphemes and reflexives in Italian, French and related lan-
guages. Parameters and universals. 131–162. Oxford & New York: Oxford University Press
Koopman, Hilda & Dominique Sportiche. 1989. Pronouns, logical variables and logophoricity in
Abe. Linguistic Inquiry 20. 555–589.
Kratzer, Angelika. 2009. Making a pronoun. Fake indexicals as windows into the properties of
pronouns. Linguistic Inquiry 40. 187–237.
Kuno, Susumu. 1987. Functional syntax. anaphora, discourse and empathy. Chicago: University
of Chicago Press.
Maling, Joan. 1984. Non-clause bounded reflexives in modern Icelandic. Linguistics and
Philosophy 7. 211–41.
Marantz, Alec. 1993. Implications of asymetries in double object constructions. In Mchombo,
Sam (ed.), Theoretical aspects of Bantu grammar, 113–150. Stanford, CA: CSLI
Publications.
700 Isabelle Charnavel and Victoria Mateu
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
Moltmann, Friederike. 2006. Generic one, arbitrary PRO, and the first person. Natural Language
Semantics 14. 257–81.
Ormazabal, Javier & Juan Romero. 2002. Agreement restrictions. Ms., University of the Basque
Country and University of Alcalá, Spain.
Ormazabal, Javier & Juan Romero. 2007. The object agreement constraint. Natural Language
and Linguistic Theory 25. 315–347.
Oshima, David Y. 2007. On empathic and logophoric binding. Research on Language and
Computation 5(1). 19–35.
Pearson, Hazel. 2015. The interpretation of the logophoric pronoun in Ewe. Natural Language
Semantics 23. 77–118.
Perlmutter, David M. & Paul Postal. 1983 The relational succession law. In David Perlmutter
(ed.), Studies in relational grammar 1.30–80. Chicago: University of Chicago Press.
Reinhart, Tanya. 1983: Anaphora and semantic interpretation. Chicago, IL: University of Chicago
Press.
Reinhart, Tanya. 2000. The theta system. Syntactic realization of verbal concepts. Utrecht. UiL
OTS Working Papers.
Reinhart, Tanya & Eric Reuland. 1993. Reflexivity. Linguistic Inquiry 24(4). 657–720.
Rezac, Milan. 2003. The fine structure of cyclic Agree. Syntax 6. 156–182.
Rezac, Milan. 2006. Phi-agree and theta-related case. http://loargann.batcave.net/index.html
(accessed 18 September 2013).
Roca, Francesc. 1992. On the licensing of pronominal clitics. The properties of object clitics in
Spanish and Catalan. Barcelona, Spain: Universitat Autònoma de Barcelona MA thesis.
Roelofsen, Floris. 2010. Condition B effects in two simple steps. Natural Language Semantics
18. 115–140.
Rosen, Carol. 1990. Rethinking Southern Tiwa: the geometry of a triple agreement Language.
Language 66. 669–713.
Ruwet, Nicolas. 1990. En et y: deux clitiques pronominaux antilogophoriques [en and y: two
antilogophoric pronominal clitics]. Languages 97. 51–81.
Schlenker, Philippe. 2003. Indexicality, logophoricity, and plural pronouns. In Jacqueline
Lecarme (ed.), Research on Afroasiatic grammar, 409–428. Amsterdam: J. Benjamins.
Sells, Peter. 1987. Aspects of logophoricity. Linguistic Inquiry 18. 445–79.
Sportiche, Dominique. 1996. Clitic constructions. In Johan Rooryck & Laurie Zaring (eds.)
Phrase structure and the lexicon. 213–277. Dordrecht: Kluwer.
Stowell, Timothy. 1981. Origins of phrase structure. Cambridge, MA: MIT dissertation.
Sundaresan, Sandhya. 2012. Context and (co)reference in the syntax and its interfaces. Tromsø/
Stuttgart & Norway/Germany: University of Tromsø/University of Stuttgart dissertation.
The clitic binding restriction revisited 701
Brought to you by | University of California - Los Angeles - UCLA Library
Authenticated
Download Date | 12/7/15 2:11 AM
