PreprintPDF Available

Polite Pronouns and the PCC *

November 2024

Authors:

Leibniz-Centre General Linguistics

Preprints and early-stage research may not have been peer reviewed yet.

The Person-Case Constraint (PCC) has played a substantial role in the development of linguistic theory in recent decades, particularly with respect to agreement and the encoding of person (Anagnostopoulou 2017b and references therein). While much of the generative literature on the PCC advances or assumes a (mor-pho)syntactic agreement-based account of PCC effects, we provide novel evidence from polite pronouns that challenges this perspective. Polite pronouns have the useful property that they exhibit a striking "mismatch" between the features expressed in their agreement and in their forms on the one hand, and what is interpreted on the other. They therefore provide an ideal testing ground for the predictions of morphosyntactic analyses of the PCC, which predict that third-person polite pronouns used for addressees should behave like other third-person arguments and should therefore fail to give rise to PCC effects. We find that this prediction is falsified in Italian for the polite pronoun LEI, which is used for formal address but is grammatically third-person, and for which PCC effects obtain; we make related observations for the related person-hierarchy effect connected to the so-called Fancy Constraint (Postal 1989). We suggest that the PCC pattern with polite pronouns is more consistent with a syntacticosemantic, interpretation-based account of the PCC, such as that of Pancheva and Zubizarreta (2018), and sketch how this can be captured in their system. Lastly, we identify cross-linguistic expectations for ditransitives vs. other person-hierarchy effects with polite pronouns, and show how these expectations are borne out for Spanish USTED and German SIE.

Content uploaded by Luke Adamson

Content may be subject to copyright.

Polite Pronouns and the PCC*

Luke James Adamson (ZAS) and Stanislao Zompì (Universität Potsdam)

November 6, 2024

Abstract

The Person-Case Constraint (PCC) has played a substantial role in the devel-

opment of linguistic theory in recent decades, particularly with respect to agree-

ment and the encoding of person (Anagnostopoulou 2017b and references therein).

While much of the generative literature on the PCC advances or assumes a (mor-

pho)syntactic agreement-based account of PCC effects, we provide novel evidence

from polite pronouns that challenges this perspective. Polite pronouns have the

useful property that they exhibit a striking “mismatch” between the features ex-

pressed in their agreement and in their forms on the one hand, and what is in-

terpreted on the other. They therefore provide an ideal testing ground for the pre-

dictions of morphosyntactic analyses of the PCC, which predict that third-person

polite pronouns used for addressees should behave like other third-person argu-

ments and should therefore fail to give rise to PCC effects. We ﬁnd that this pre-

diction is falsiﬁed in Italian for the polite pronoun LEI, which is used for formal

address but is grammatically third-person, and for which PCC effects obtain; we

make related observations for the related person-hierarchy effect connected to the

so-called Fancy Constraint (Postal 1989). We suggest that the PCC pattern with po-

lite pronouns is more consistent with a syntacticosemantic, interpretation-based

account of the PCC, such as that of Pancheva and Zubizarreta (2018), and sketch

how this can be captured in their system. Lastly, we identify cross-linguistic expec-

tations for ditransitives vs. other person-hierarchy effects with polite pronouns, and

show how these expectations are borne out for Spanish USTED and German SIE. This

*We are grateful to Artemis Alexiadou, Elena Anagnostopoulou, Karlos Arregi, Nina Haslinger, Giovanni

Roversi, Andrés Saab, and audiences at ZAS, Universität Potsdam, and NELS 55 for insightful discussion.

Grammaticality judgments for Italian come from several native speaker consultants from different areas

of Italy, including one of the authors, Enrico Flor, and Jacopo Romoli. In addition to Leipzig Glossing Rules

conventions, we also use IMPRS = impersonal. Note that, unless otherwise stated, ﬁnite verbs appear in

present indicative forms.

study has important implications for the analysis of person-hierarchy effects and

afﬁrms the relevance of polite pronouns to the theory of agreement.

keywords: PCC, person hierarchy, agreement, interpretability

1 Introduction

Many languages exhibit person-based restrictions in their combination of object argu-

ments in ditransitive environments, which have been investigated extensively under the

label of Person-Case Constraint (PCC) effects (Bonet 1991, 1994; Anagnostopoulou 2003,

2005, 2017b; Béjar and Rezac 2003; Bianchi 2006; Ormazabal and Romero 2007; Béjar

and Rezac 2009; Adger and Harbour 2007; Nevins 2007; Preminger 2009, 2014; Rezac

2011; Pancheva and Zubizarreta 2018; Stegovec 2020; Coon and Keine 2021; Foley and

Toosarvandani 2022; Paparounas and Salzmann 2023; Deal 2024; among many others).

This type of restriction is found in Italian (e.g. Bianchi 2006; D’Alessandro and Pescarini

2016): while clitic combinations of a 3rd-person indirect object (IO) with a 3rd direct ob-

ject (DO) (henceforth 3>3) and 2>3 (1-a) are licit, clitic combinations of a 3rd-person IO

with a participant DO (such as 2nd) are illicit (1-b)-(1-c).

(1) a. {Glie=

3SG.DAT

/te}

2SG.DAT

3F.SG.ACC

has

afﬁdat-a.

entrusted-F.SG

‘(S)he entrusted her to him/her/you.’

b. *{Gli(e)

3SG.DAT

/le} {ti

2SG.ACC

/te} ha

have.3SG

afﬁdat{-o/-a}.

entrusted-{M.SG/F.SG}

Intended: ‘She entrusted you to him/her.’

c. *{Ti

2SG.ACC

/te} {gli(e)

3SG.DAT

/le} ha

have.3SG

afﬁdat{-o/-a}.

entrusted-{M.SG/F.SG}

Intended: ‘She entrusted you to him/her.’ (cf. Bianchi 2006)

PCC effects have played a substantial role in the theory of agreement and the en-

coding of person (Anagnostopoulou 2017b and references therein). While accounts vary

greatly, most in the generative literature converge on the idea that PCC effects arise in

the morphosyntax, e.g. via restrictions on Agree with multiple goals (Béjar and Rezac

2003; Anagnostopoulou 2005; Coon and Keine 2021; Deal 2024; among many others).

In the present work, we provide evidence from ‘polite’ pronouns that challenges the

received view that PCC effects with ditransitives should be attributed to morphosyn-

tax.1Polite pronouns have the useful property that they exhibit a mismatch between the

features expressed in their form and in their agreement on the one hand, and what is

interpreted on the other. They thus provide an ideal testing ground for morphosyntactic

analyses of the PCC, which predict that polite pronouns that are formally 3rd-person but

used for addressees should behave like 3rd-person arguments rather than 2nd, thereby

failing to give rise to PCC effects. We ﬁnd that this prediction is falsiﬁed in Italian for

the polite pronoun LEI, which, as we show, is used for formal address but is grammati-

cally 3rd-person, and for which PCC effects obtain (as also brieﬂy noted by D’Alessandro

and Pescarini 2016). We further show that LEI patterns with 2nd-person arguments for

at least one other person-hierarchy effect in Italian, namely Postal’s (1989) ‘Fancy Con-

straint’, an effect that has also been attributed to morphosyntactic mechanisms respon-

sible for PCC effects (e.g. Sheehan 2020; Deal 2024).

We argue instead that the PCC pattern with LEI is more consistent with a syntacti-

cosemantic analysis of the PCC, and we sketch how this can be captured in a system

such as that of Pancheva and Zubizarreta 2018. In identifying modularity-based dif-

ferences between distinct PCC theories, this study has important implications for the

theory of person restrictions and provides a novel empirical tool to probe such effects

cross-linguistically. While the relevance of polite pronouns to the theory of agreement

has been acknowledged in previous studies (see especially Wechsler and Hahm 2011;

Ackema and Neeleman 2018), this study is, to our knowledge, novel in situating them

within theories of person restrictions.

1We use the term morphosyntax loosely, to encompass both narrow syntax and ‘pure’ morphology, as

well as the interface between the two.

The organization of this work is as follows. In Section 2, we lay out the PCC pattern in

Italian, which has been observed previously in the literature. In Section 3, we describe

the morphosyntactic behavior of the polite pronoun LEI, which we demonstrate pat-

terns with third person rather than second person both morphologically and in terms

of verbal agreement. In Section 4, we show that PCC effects obtain with LEI, extend our

observations to another type of person restriction, and contrast LEI’s PCC behavior with

that of camouﬂage/imposter nominals. Section 5 then details how morphosyntactic ac-

counts of the PCC are undermined by the data from LEI, and sketches one alternative

syntacticosemantic account. Section 6 looks at some comparisons with other languages

that have also been reported to display person-hierarchy effects and have third-person

polite pronouns. Section 7 concludes.

2 The PCC in Italian

As mentioned in Section 1, PCC effects have been reported to obtain in Italian; see espe-

cially Bianchi 2006; D’Alessandro and Pescarini 2016 (on other Romance languages, see

Perlmutter 1971; Bonet 1991, 1994, among many others). In Italian, object clitics typi-

cally precede a ﬁnite verb and can occur in clusters that include both the indirect and

direct object, as in the combinations in (2) (adapted from Bianchi 2006), repeated from

above. In these examples, the indirect object is represented with a dative clitic and the

direct object is represented with an accusative clitic, with clitics reﬂecting person, num-

ber, case, and (for third person) gender. As evident from (2), 2>3 and 3>3 clitic combi-

nations are both grammatical.2

2For some clitics, elision of the clitic vowel is possible when a following adjacent verb begins with a

vowel, including with the third-person singular feminine la and masculine lo, though it is not obligatory

(see e.g. Maiden and Robustelli 2013:96-97). Because the unelided third-person forms more clearly reﬂect

gender, we opt to include them throughout; to be clear, changing the elision status for our ungrammatical

examples does not render them felicitous under the relevant interpretation in any of the cases discussed

(2) a. Te

2SG.DAT

3F.SG.ACC

have.3SG

afﬁdata.

entrusted.F.SG

‘He entrusted her to you.’

b. Glie

3SG.DAT

3F.SG.ACC

have.3SG

afﬁdata.

entrusted.F.SG

‘He entrusted her to him/her.’

All (or at least the vast majority of) Italian speakers reject combinations of two argu-

ment clitics in which the indirect object is third person and the direct object is ﬁrst or

second person. This is illustrated for second person in (3), ungrammatical regardless of

clitic ordering or clitic-allomorph selection.

(3) *{Gli(e)/le

3SG.DAT

2SG.ACC

/ti

2SG.ACC

gli}

3SG.DAT

have.3SG

afﬁdato/afﬁdata.

entrusted.M.SG/entrusted.F.SG

‘He entrusted you to her.’

As is well-known, PCC effects vanish when one of the arguments is a stressed (or

‘tonic’) pronoun rather than a clitic, as observed by Bianchi 2006 for Italian (Bonet 1994

on Catalan, Anagnostopoulou 2003 on Greek, and much subsequent literature). This is

shown in (4).

(4) a. Gli

3SG.DAT

have.3SG

afﬁdato

entrusted.M.SG

te.

2SG.STRESS

‘He entrusted you to him/her.’

b. Ti

2SG.ACC

have.3SG

afﬁdato

entrusted.M.SG

{lui

3M.SG.STRESS

/lei}.

/3.F.SG.STRESS

‘He entrusted you to him/her.’

While virtually all Italian native speakers reject clitic combinations of 3>1 and 3>2,

some speakers also reject combinations of 1>2 and 2>1 (see e.g. Bianchi 2006:2027), as

shown in (5).3Such speakers are said to have a Strong PCC grammar, with speakers who

accept these combinations – but not 3>1 or 3>2 – having a Weak PCC grammar. In the

in this manuscript.

3Note that ACC and D AT marking for ﬁrst and second are syncretic in this environment and the order

of the clitics necessarily places ﬁrst person before second person, regardless of which argument is the IO

and which is the DO.

present work, we focus chieﬂy on 3>2 combinations, though with some discussion of

1>2 combinations in both Sections 4 and 5.4

(5) %Mi

1SG

2SG

have.3SG

afﬁdato/afﬁdata.

entrusted.M.SG/entrusted.F.SG

%‘She entrusted you to me.’ / %‘She entrusted me to you.’

3 The polite pronoun LEI

The Italian ‘polite’ pronominal series identiﬁed here as LEI – in reference to its citation

form – is used as a politeness or courtesy form in reference to singular addressees. (For a

description of the sociopragmatic conditions under which LEI vs. other forms of address

is used, see Maiden and Robustelli 2013:460-465.) In this section, we show that LEI pat-

terns with the third-person feminine singular pronoun in terms of its morphology and

in terms of its verbal agreement behavior, including as an object clitic.

LEI is morphologically identical to the third-person feminine singular series, as can

be seen in Table 1. As reﬂected in the table, all of the elements in the LEI series are con-

ventionally capitalized in the written language except for the possessive pronoun, while

the third-person feminine forms are not (unless at the beginning of a sentence).

The third-person feminine behavior of LEI is not restricted to identity in pronomi-

nal forms. Italian distinguishes second person verbal agreement from third person, as

shown in (6). Despite referring to an addressee, LEI triggers third-person agreement

(7-a). Note further that, as a pro-drop language, Italian allows subjects to be non-overt;

LEI may also be dropped, still triggering third-person agreement in politeness contexts

(7-b).

4We set person-hierarchy effects with reﬂexives to the side. For discussion of such effects in Italian, see

Bianchi 2006; D’Alessandro and Pescarini 2016.

NOM ACC clitic DAT clitic stressed POSS

1SG io mi mi/me me mio

2SG tu ti ti/te te tuo

3SG.Mlui lo gli/glie lui suo

3SG.Flei la le/glie lei suo

LEI Lei La Le/Glie Lei suo

1PL noi ci ci/ce noi nostro

2PL voi vi vi/ve voi vostro

3PL.Mloro li %loro/%gli/%glie loro loro

3PL.Floro le %loro/%gli/%glie loro loro

Table 1: Personal Pronominal Series (cf. Maiden and Robustelli 2013:Ch. 6)

(6) a. Tu

2SG.NOM

{sei

be.2S G

/*è}

/be.3S G

qui.

here

‘You are here.’

b. Lei

3SG.F.NOM

{è

be.3S G

/*sei}

/be.2S G

qui.

here

‘She is here.’

(7) a. Lei

LEI.NOM

{è

be.3S G

/*sei}

/be.2S G

qui.

here

‘You (polite) are here.’

b. Dottor

Doctor

Biagi,

pro

PRO

{è

be.3S G

/#sei}

/be.PRS.2SG

qui!

here

‘Doctor Biagi, you’re here!’

Reﬂexive elements are also distinguished between second and third person: in the

second person, reﬂexive clitics are formally identical to non-reﬂexive clitics, whereas in

the third person, reﬂexive clitics are realized with si/se (8).5In the domain of reﬂexives,

LEI again patterns with the third person (9). (A parallel contrast, omitted here, is found

with stressed, non-clitic reﬂexives.)

(8) a. {Ti

2SG.ACC

/*si}

/3SG.REFL

vedi.

see.2S G

‘You see yourself.’

b. {Si

3SG.REFL

/#ti}

/2SG.ACC

vede.

see.3S G

‘She sees herself/he sees himself.’

(9) (Lei)

LEI.NOM

{si

LEI.REFL

/#ti}

/2SG.ACC

vede.

see.3S G

‘You see yourself.’

5Beyond reﬂexives, si/se is also found in a wider set of environments including impersonals; see

D’Alessandro 2007 for an overview and for discussion.

LEI also behaves as a third-person argument and unlike a second-person singular ar-

gument for the purposes of clitic-cluster ordering (on the ordering of clitics in Romance,

see Pescarini 2021 and references therein). In a combination of an accusative clitic with a

locative, the second-person clitic ti must precede the locative clitic, as in (10-a), whereas

the third person follows the locative, as in (10-b). LEI patterns again with the third person

(10-c).

(10) a. {Ti

2SG.ACC

ci/

LOC

*ci

LOC

ti}

2SG.ACC

hanno

have.3PL

portato.

brought.M.SG

‘They have brought you there.’

b. {Ce

LOC

la/

3F.SG.ACC

*la

3F.SG.ACC

ci}

LOC

hanno

have.3PL

portata.

brought.F.SG

‘They have brought her there.’

c. {Ce

LOC

La/

LEI.ACC

*La

LEI.ACC

ci}

LOC

hanno

have.3PL

portata.

brought.F.SG

‘They have brought you (formal) there.’

As an object clitic, LEI is also treated as if it were a third-person feminine argument

with respect to participle agreement. In Italian as in many Romance languages, verbal

past participles can display agreement with accusative object clitics (see Kayne 1989;

Belletti 2017; among many others). Crucially, though, such overt agreement is obligatory

with third-person object clitics but optional with participant object clitics (Burzio 1986;

Belletti 2017). This is illustrated in the contrast between (11) and (12). In the case of the

second person, the form of the clitic (ti) does not vary depending on the gender of the

referent, whereas agreement on the participle optionally varies: when the clitic refers

to a woman, a feminine form can be employed but ‘default’ masculine is also available;

when referring to a man, the clitic is necessarily masculine because both the agreeing

form and the default are masculine.6

6The optionality of participial gender agreement with participant clitics might at ﬁrst seem related

to the fact that gender contrasts are not overtly marked on such clitics, thus suggesting an account in

terms of gender impoverishment on the clitic optionally bleeding gender agreement with it. However, the

phenomenon is in fact more general: for example, participial number agreement is also optional with

plural participant clitics, despite the number contrast being overtly marked on the clitics themselves in

(11) a. Maria,

Maria,

2SG.ACC

have.1SG

{vist-a

seen-M.SG

/vist-o}

seen-F.SG

ieri

yesterday

TV.

‘Maria, I saw you on TV yesterday.’

b. Pietro,

Pietro,

2SG.ACC

have.1SG

{vist-o

seen-M.SG

/*vist-a}

seen-F.SG

ieri

yesterday

TV.

‘Pietro, I saw you on TV yesterday.’

As shown in (12), LEI as an accusative clitic patterns with third person accusative cl-

itics in that gender agreement is obligatory. Moreover, the agreement is for the formal

features of LEI and not for the conceptual gender of the referent, unlike the second per-

son. The gender agreement pattern thus suggests that the clitic LEI behaves as expected

if it were a third-person feminine clitic.

(12) a. La

3SG.F.ACC

have.1SG

{vist-a/*vist-o}

seen-F.SG/seen-M.SG

ieri

yesterday

TV.

‘I saw her on TV yesterday.’

b. Lo

3SG.M.ACC

have.1SG

{vist-o/*vist-a}

seen-M.SG/seen-F.SG

ieri

yesterday

TV.

‘I saw him on TV yesterday.’

(13) (Dottor

Doctor

Biagi,)

Biagi,

LEI.ACC

have.1SG

{vist-a

seen-F.SG

/*vist-o}

/seen-M.SG

ieri

yesterday

TV.

‘Doctor Biagi, I saw you on TV yesterday.’ (could also mean ‘I saw her on TV

yesterday.’) (adapted from Maiden and Robustelli 2013:459)

The evidence from the pronominal morphology and the verbal agreement behavior

of the polite series indicates that the formal features of LEI are that of the third-person

feminine singular.

4LEI and Person-Hierarchy Effects

In this section, we demonstrate that PCC environments yield ungrammaticality with LEI.

This is signiﬁcant because approaches that attribute PCC effects to formal agreement

this case (1SG mi vs 1PL ci, 2S G ti vs 2PL vi). For syntactic analyses of the pattern, cf. also Guasti and Rizzi

2002:191 and Belletti 2017:499.

rather than to the syntacticosemantic status of arguments predict that third-person po-

lite pronouns should not give rise to PCC effects, because for the purposes of verbal

agreement, these pronouns behave as if they were third person, despite being inter-

preted as referring to addressees. We contextualize this prediction of morphosyntactic

accounts more explicitly in Section 5.

We ﬁrst demonstrate the PCC effect with LEI and third-person clitics and show how

it cannot be attributed to a morphological restriction on accusative LEI in clitic clus-

ters (4.1). We extend our observations to another person-hierarchy effect found with

LEI in so-called ‘Fancy Constraint’ environments (4.2). Lastly, we contrast LEI with im-

posters/camouﬂage DPs, which do not exhibit PCC effects in Italian (4.3).

4.1 LEI and the PCC

The example in (14), repeated from (3), displays the PCC effect: the dative clitic is third-

person while the accusative clitic is second-person, and the result is ungrammatical.

(14) *Gli(e)/le

3SG.DAT

2SG.ACC

have.3SG

afﬁdato/afﬁdata.

entrusted.M.SG/entrusted.F.SG

‘He entrusted you to her.’

We now turn to the interaction between LEI and the PCC. First observe that LEI can

appear as the dative clitic with a third-person accusative, as this is not a PCC environ-

ment. Like third-person dative clitics (both masculine and feminine), a dative LEI clitic

is realized as glie in the context of a following clitic; the result is expectedly grammatical

(15).

(15) Glie=

{3SG.DAT /LEI.DAT}

3SG.F.ACC

have.3SG

afﬁdata.

entrusted.F.SG

‘He entrusted her to him/to you (formal).’

Strikingly, when LEI appears as an accusative clitic with a dative third person, the

result is ungrammatical (16), displaying a PCC effect comparable to the one observed for

the second-person clitic (14). Note that there is nothing inherently ill-formed about the

combination of the third-person dative with a third-person feminine accusative (which

is string-identical to (15)).7

(16) *Glie=

3SG.DAT

LEI.ACC

AUX .3SG

afﬁdata/afﬁdato.

entrusted.F.SG

‘He entrusted you (formal) to her.’

To ensure that the PCC effect with LEI is genuine, we offered the context in (17) to

facilitate the target interpretation. Our consultants still found the relevant clitic combi-

nation ungrammatical.

(17) ‘Oh avvocato, come sta? Non sa quanto mi è dispiaciuto che il mio medico L’abbia

trattata male. Quello lì è proprio un cretino sa?’

Oh, esteemed lawyer, how are you (formal)? You don’t know how sorry I was that

my doctor treated you badly. He’s a real idiot, isn’t he?

a. *Io

1SG.NOM

glie-L’

3DAT-LEI.ACC

avevo

HAD.1SG

afﬁdata

entrusted.F.SG

sperando

hoping

che

that

LEI.ACC

curasse

cure.SUBJ.PST.3SG

perbene

properly

and

invece...)

instead

‘I entrusted you to him hoping that he would take care of you properly (and

instead...)

While the sentences with (16) and (17) are formed with afﬁdare ‘entrust’, we observe

the same effect with the verb raccomandare ‘recommend’. When LEI is a dative clitic

occurring with a third-person accusative, no PCC effect is expected and indeed the sen-

tence is grammatical (18-a). When LEI is an accusative clitic occurring with a third per-

son dative, a PCC effect obtains and the result is ungrammatical (18-b).

(18) a. Glie=

LEI.DAT

3F.SG.ACC

have.3SG

raccomandata.

recommended.F.S G

‘He recommended her to you.’

7The contrast between (15) and (16) also helps assuage the suspicion that the effect might have to do

with an excess of markedness in the clitic cluster (on the assumption that DAT-ACC clitic clusters and polite

LEI are both inherently marked). If that were the issue, we would expect 3>LEI to be no worse than LEI>3,

contrary to fact.

b. *Glie=

3SG.DAT

LEI.ACC

have.3SG

raccomandata(/raccomandato).

recommended.F.S G

‘He recommended you (formal) to her.’

As with PCC effects with the second person, these types of expressions are ‘repaired’

if the ditransitive is not expressed with a combination of two clitics. In (19-a), the dative

argument is a clitic while the accusative LEI appears as a stressed postverbal pronoun;

the result is well-formed. In (19-b), LEI appears as an accusative clitic with the dative ar-

gument appearing postverbally in a prepositional phrase; the result is again grammati-

cal.8

(19) a. Gli

3SG.DAT

have.3SG

afﬁdato

entrusted.M.SG

Lei.

LEI.STRESS

‘(S)he entrusted you to him.’

b. La

LEI.ACC

have.3SG

afﬁdata

entrusted.F.SG

lui.

3M.SG

‘(S)he entrusted you to him.’

Further conﬁrmation for the PCC effect with LEI comes from corpus results from

the Paisà corpus (Lyding et al. 2014). Because of the prescriptive preference to capi-

talize LEI in the written language, we can ask whether the clitic combination glie-La

(3SG.DAT-LEI.ACC) is attested in this case-sensitive corpus, and compare it with the non-

capitalized glie-la (3SG.DAT-3SG.F.ACC). While it is possible that some instances of non-

8For completeness, note that PCC effects also hold for expressions in which the pronominal clitics are

encliticized to an inﬁnitive. As shown in (i-b), a combination of third-person dative with third-person

accusative is licit, as is LEI when it appears as the dative. We observe, however, that LEI as an accusative

clitic in this inﬁnitival environment is ungrammatical with a third-person dative clitic (i-c).

(i) a. *Intendo

intend.1S G

afﬁdar

entrust.INF

{gli

3SG.DAT

ti}/

2SG.ACC

{ti

2SG.ACC

gli}

3SG.DAT

‘I intend to entrust you to him.’

b. Intendo

intend.1S G

afﬁdar

entrust.INF

{glie=

3SG.DAT

/Glie=}

/LEI.DAT

la.

3SG.ACC

‘I intend to entrust her to him/you (formal).’

c. *Intendo

intend.1S G

afﬁdar

entrust.INF

glie=

3SG.DAT

La.

LEI.ACC

‘I intend to entrust you to him.’

capitalized gliela could be instances where the accusative is meant to be interpreted as

polite LEI, our expectation is at least that there should be no instances of capitalized LEI

in this position. A corpus search in 2024 yielded 580 hits for the string gliela tagged as a

‘clitic pronoun’ but zero hits for glieLa (even without tags).

The restriction on the distribution of LEI cannot be attributed to an inability for LEI to

participate in clitic clusters. LEI is shown above in (15) as a dative argument with a third-

person accusative; we also speciﬁcally observe that the accusative form La can appear

with a preceding locative clitic, as in (20-a), or with an impersonal clitic si (which follows

La), as in (20-b).9

(20) a. Dottore,

doctor

alla

in.the

ﬁne

end

LOC

LEI.ACC

hanno

have.3PL

portata,

brought.F.SG

tribunale?

court

‘Doctor, did they drag you to court in the end?’

b. Onorevole,

congressperson,

perché

why

LEI.ACC

IMPRS

soprannomina

refer.to.3SG

così?

this.way

‘Congressperson, how come they call you that?’

Based on the preceding evidence, we conclude that the third-person polite pronoun

LEI patterns with second person clitics in inducing PCC effects, rather than with third

person. Before proceeding, recall from Section 2 that some Italian speakers accept com-

binations of ﬁrst or second person accusative clitics when the dative is also ﬁrst or sec-

9Further evidence that the accusative LEI clitic is morphologically well-formed in clitic clusters comes

from ethical dative contexts, which have been observed not to give rise to PCC effects (see Bianchi 2006

on Italian). LEI may appear felicitously as an accusative clitic in an ethical dative context; we ﬁnd that

ﬁrst-person ethical datives are easiest to judge. Recall that Weak PCC Italian speakers do not accept com-

binations of participant clitics (see (21-a)). In contrast, all speakers, including Weak PCC speakers, accept

a combination of a dative ﬁrst-person and an accusative second-person clitic in an ethical dative con-

struction, like that seen in (i-a). Perhaps unsurprisingly, LEI in accusative position is well-formed in the

ethical dative construction (i-b), including for Weak PCC speakers.

(i) a. Come

how

1SG.ETH.DAT

2SG.ACC

hanno

have.3PL

conciato?

ill.treated.M.SG

‘How badly did they mess you up?’

b. Come

how

1SG.ETH.DAT

LEI.ACC

hanno

have.3PL

conciata?

ill.treated.F.SG

‘How badly did they mess you (formal) up?’

ond person, as shown above in (5) and repeated in (21-a). This is in contrast with the

combination of 1>3, which is acceptable for all speakers (21-b).

(21) a. %Mi

1SG

2SG

have.3SG

afﬁdato.

entrusted.M.SG

%‘He entrusted you to me.’ / %‘He entrusted me to you.’

(based on Bianchi 2006:2027)

b. Me

1SG

3SG.F.ACC

have.3SG

afﬁdata.

entrusted.F.SG

‘He entrusted her to me.’

Strikingly, even among speakers who accept (21-a), substituting the second person

accusative in (21-a) with an accusative LEI clitic makes the expression ungrammatical

(22). There is no morphological reason for this combination to be ill-formed; the use of

a third person accusative clitic in this position is grammatical, as seen in the preceding

example (21-b). Manipulating the order or the allomorphic exponence of the clitics for

(22) still results in ungrammaticality (*La me, *Le mi, *Le me, etc.).

(22) (Dottor

Doctor

Biagi,)

Biagi,

*me

1SG.DAT

LEI.ACC

have.3SG

afﬁdata.

entrusted.F.SG

Doctor Biagi, he entrusted you to me.

We mostly set 1>2 and 2>1 environments with LEI to the side, returning to this issue

later in Section 5.

4.2 LEI and the Fancy Constraint

We now extend our observations about person restrictions from PCC effects to the so-

called Fancy Constraint. Two analytic causative constructions have been identiﬁed in

the literature on Romance, often referred to as faire par and faire inﬁnitif causatives;

see Kayne 1975; Burzio 1986; Guasti 1996; Folli and Harley 2007; among many others. In

Italian, the two are distinguished by how a causee is introduced: within a PP headed by

da ‘from’ in the faire par construction but with a dative clitic or a PP headed by a‘to’ in

the faire inﬁnitif construction (23).

(23) a. Micol

Micol

make.3S G

pettinare

comb.INF

Giulia

Carlo.

Carlo

‘Micol is making Carlo comb Giulia’s hair.’

b. Micol

Micol

make.3S G

pettinare

comb.INF

Giulia

Carlo.

Carlo

‘Micol is making Carlo comb Giulia’s hair.’

(adapted, D’Alessandro and Pescarini 2016:275)

As noted by D’Alessandro and Pescarini (2016), a person-hierarchy effect can be

found with the faire inﬁnitif construction in Italian (but not for faire par)—an effect

that Postal (1989) ﬁrst discovered in French and termed the ‘Fancy Constraint’ (see also

Bonet 1991; Sheehan 2020). In particular, there is a contrast between the acceptability

of a third-person accusative clitic vs. a ﬁrst- or second-person clitic. The contrast is il-

lustrated in (24); notice that the effect obtains even though the causee argument is not a

clitic in the example, a point which has been taken to indicate that the PCC is not strictly

linked to clusters with clitics or weak pronouns (e.g. Sheehan 2020). (We note that our

consultants tend to ﬁnd the ‘baseline’-type case in (24) marked to begin with, though

there is nevertheless a sharp person-based contrast.)

(24) Micol

Micol

{la

3SG.F.ACC

/*ti}

2SG.ACC

make.3S G

pettinare

comb.INF

Carlo.

Carlo

‘Micol is making Carlo comb her/*your hair.’

(adapted, D’Alessandro and Pescarini 2016:276)

As with the PCC effect found with ditransitives, we ﬁnd that LEI patterns with the

second person and not the third person in triggering a person-hierarchy effect with the

faire inﬁnif, as shown in (25). As with the PCC, the hierarchy effect vanishes when the

argument LEI is expressed with a pronoun rather than a clitic (26).

(25) *Signor

Signor

Biagi,

Biagi

Micol

Michol

LEI.ACC

make.3S G

pettinare

comb.INF

Carlo.

Carlo

‘Signor Biagi, Michol is making Carlo comb your hair.’

(26) Micol

Micol

make.3S G

pettinare

comb.INF

te/Lei

2SG.ACC.STRESSED/LEI.STRESSED

Carlo.

Carlo

‘Micol makes Carlo comb your/your(polite) hair.’

4.3 Imposters and the PCC in Italian

We can also contrast the PCC effect from LEI with what is observed with non-pronominal

imposter and camouﬂage nominals, which, like third-person polite pronouns, are gram-

matically third-person for verbal agreement, but appear to represent their relationship

with participant pronouns differently from polite pronouns; see Collins and Postal 2012

for one view of imposter structure, and Servidio 2014 on imposters speciﬁcally in Italian.

We show that LEI behaves differently from non-pronominal imposters in Italian, both in

terms of PCC effects and in terms of other agreement behavior. (See also Rezac 2011:297

for the observation that French imposters do not induce PCC effects.)10 This evidence

suggests that the syntacticosemantic representation of LEI as (fundamentally) second

person is not shared with imposters, giving rise to divergent behavior in PCC contexts.

We ﬁrst show the absence of a PCC effect with the camouﬂage item le Vostre Maestà

‘the.F.PL 2PL.POSS majesties’ (with the possessive also agreeing in gender and number

with the noun Maestà). In order to test this item within a PCC context, we employ a left-

dislocated topic structure as in (27): the plural camouﬂage item le Vostre Maestà appears

as a left-dislocated topic, which is indexed by a third-person accusative clitic agreeing

with the feminine gender and plural number of the topic. (Because it is plural, it is clear

that the clitic is not polite LEI, which would appear as a singular clitic and can only be

used with singular addressees.) To make this a PCC context, a dative third-person clitic

is also present, such that, if the camouﬂage item were treated as second-person, a PCC

effect would be induced. In contrast, we ﬁnd that the camouﬂage item does not give rise

10A related observation for Georgian comes from Harris (1981:Ch. 3), who points out that Georgian

person-hierarchy effects in ditransitives vanish when a (reﬂexive-looking) ‘camouﬂage’ nominal is used

in place of a participant pronoun, which agrees as if it is grammatically third person.

to a PCC effect, unlike LEI.

(27) [Le

the.F.P L

Vostre

2PL.POSS.F.PL

Maestà]i,

majesty.PL

gliek

3SG.DAT

lei

3PL.ACC

hanno

have.3PL

già

already

presentate,

introduced.F.PL

al[l’ambasciatore]k?

to.the.ambassador.M.S G

‘Have they already introduced Your Majesties to the ambassador?’

We also ﬁnd that camouﬂage items do not give rise to violations of the Fancy Con-

straint in Italian. In (28), the camouﬂage item le Vostre Altezze ‘the.F.PL 2PL.POSS.F.PL

highnesses’ is again put into a left-dislocated topic structure, and is indexed by an ac-

cusative clitic in the faire inﬁnitif construction. The expression is grammatical.

(28) [Le

the.F.P L

Vostre

2PL.POSS.F.PL

Altezze]i,

highness.PL

the.M.SG .

professore

professor.M.SG

lei

3ACC.F.PL

farà

make.FUT.3SG

esaminare

examine.INF

all’

to.the.SG

assistente.

assistant

‘The professor will make the assistant examine Your Highnesses.’

We can replicate the same pattern with the non-camouﬂage imposter i signori ‘the

gentlemen’. (Once again, because it is plural, it is clear that the clitic is not polite LEI.)

In contrast to both second person and LEI, no person-hierarchy effect arises with the

imposter nominal.

(29) a. [I

the.M.PL

signori]i,

gentlemen

gliek

3SG.DAT

lii

3M.PL.ACC

hanno

have.3PL

già

already

presentati,

introduced.M.PL

al[l’ambasciatrice]k?

to.the.ambassador

‘Have they already introduced the gentlemen (you) to the ambassador?’

b. [I

the.M.PL

signori]i,

gentlemen

lii

3M.PL.ACC

faranno

make.3P L

esaminare

examine.INF

all’assistente.

to.the.assistant

‘They will make the assistant examine the gentlemen (you).’

That PCC and Fancy Constraint effects do not obtain with camouﬂage items or non-

pronominal imposters, unlike with LEI, suggests that their relationship to second-person

address is represented differently. While we take no ﬁrm stance on the representation of

camouﬂage or imposter DPs, we would like to highlight that they do indeed diverge in

their behavior from LEI in that they are treated as if they are third-person in at least

one other case where LEI is not. It has previously been noted in the literature on coor-

dination resolution that resolved agreement is sensitive to interpreted features of the

nominals, as evidenced especially by cases where formal features diverge from what is

interpreted (e.g. Wechsler 2008; Adamson and Anagnostopoulou 2024, among others).

We observe that this is true for coordination of LEI with a third-person nominal (30),

which yields obligatory second-plural resolved agreement (as is also observed when the

second-person informal pronoun is used instead of LEI). In contrast, the camouﬂage

nominal and the imposter are compatible with third-person resolved agreement when

coordinated with another third-person nominal, with the alternative second or ﬁrst per-

son agreement being ungrammatical or at least severely degraded (cf. Servidio 2014).

(30) Lei

LEI.NOM

and

l’ambasciatore

the.ambassador.M.SG

{vi

2PL.ACC

incontrerete

meet.FUT.2.PL

/*si

/REFL.3PL

incontreranno}

meet.FUT.3PL

domani.

tomorrow

‘You (formal) and the ambassador will meet tomorrow.’

(31) La

the.F.S G

Vostra

2PL.POSS.F.PL

Maestà

majesty

and

l’ambasciatore

the.ambassadorM.SG

Svezia

Sweden

{si

3.REFL

incontreranno

meet.FUT.3PL

/*vi

/2.PL.ACC

incontrerete}

meet.FUT.2PL

domani.

tomorrow

‘Your majesty and the ambassador from Switzerland will meet tomorrow.’

The contrast between LEI and imposters for the PCC suggests that PCC effects do not

arise pragmatically (in a broad sense), since both types of nominals are ultimately used

for reference to addressees, yet not all addressee-referring expressions give rise to PCC

effects. The contrast is instead in line with an account in which the syntacticosemantic

representation of polite pronouns is necessarily second person in a way that imposters

and camouﬂage nominals need not be.

5LEI and the Theory of the PCC

We argue that the evidence from Italian polite LEI is incompatible with morphosyntactic

analyses of the PCC, and instead favors a syntacticosemantic account. While space con-

siderations preclude a more comprehensive evaluation of PCC accounts, we consider

here a few representative examples from the recent literature.11 In essence, our view is

that for theories that derive PCC effects from the mechanics of Agree, because LEI be-

haves as a 3rd-person argument for purposes of agreement (as shown in Section 3), it is

expected to obviate PCC effects, contrary to fact (as shown in Section 4).

For concreteness, we assume that polite pronouns have two sets of φ-features, which

are both present in the narrow syntax, one of which is interpretable and the other of

which is uninterpretable. On this type of dual-feature system and related systems, see

Wurmbrand 2017; Smith 2015, 2017, 2021; Anagnostopoulou 2017a; Messick 2023a,b;

Adamson and Anagnostopoulou 2024; Adamson 2024; among others; on polite pronouns

in particular, Wurmbrand 2016 and relatedly Despi´c 2017; Puškar-Gallien 2019.12 We as-

sume LEI carries uninterpretable third-person, feminine, and singular features – consis-

tent with its verbal agreement behavior – and interpretable second-person and singular

features (with masculine and feminine genders both being viable depending on whether

the addressee identiﬁes as a man or as a woman) – consistent with the polite pronoun

being conﬁned to singular addressees. We postpone discussion of alternative analyses

11We brieﬂy note that the evidence from polite pronouns also poses a challenge to an analysis in which

PCC effects are fundamentally morphological in character, as in Perlmutter 1971; Bonet 1991, 1994. As

described in Section 3, LEI is formally identical to the third-person feminine singular pronoun across all

of its realizations, including as an object clitic, and verbal agreement also treats it as third-person. If LEI is

treated morphologically as though it is third person, any PCC-inducing morphological ﬁlter or constraint

that restricts the distribution of second person should not be applicable to LEI, contrary to fact.

12This dual-system is also similar to that of Wechsler and Zlati´c 2003; Hahm 2010; Wechsler and Hahm

2011, but with the division being between interpretable and uninterpretable values of φ-features, which

are sent to LF and PF respectively. See Kaur to appear; Puškar-Gallien 2019 on agreement with honoriﬁc

nominals for an alternative analytic possibility for having two distinct sets of features present in the syn-

tax.

of polite pronouns to Section 7.13

We ﬁrst critically evaluate several morphosyntactic accounts of the PCC in light of

the polite pronoun evidence (5.1) before turning to a sketch of a syntacticosemantic

account (5.2).

5.1 Morphosyntactic Analyses of the PCC

5.1.1 Deal 2024

As one representative example, consider Deal’s (2024) Interaction/Satisfaction model,

according to which a Probe may enter into an Agree relation with multiple Goals, inter-

acting with every Goal in its domain until its ‘satisfaction’ conditions are met. Deal ana-

lyzes PCC effects in this model as resulting from i) probes having a satisfaction condition

such as [PART (IC IPANT)], ii) an Agree relation being necessary for argument cliticization,

and iii) the relevant probe interacting with direct objects before indirect objects.

A derivation is schematized in the tree in (32): here, an Appl head is speciﬁed to inter-

act with elements bearing φfeatures, and to probe until it ﬁnds a feature [PART ]. In this

example, the direct object bears [PART ], which the Appl probe can agree with, thereby

licensing argument cliticization of the direct object. However, because the probe iden-

tiﬁed [PA RT ] on the direct object, its satisfaction condition has been met and it cannot

13That the polite pronoun’s interpretable features are represented in narrow syntax is supported by the

fact that some agreement targets, including predicative adjectives, agree with these interpreted values:

adjectives (as well as subject-agreeing participles) reﬂect gender agreement with the conceptual gender

of the addressee, rather than with the formal, feminine gender of LEI (see Wechsler and Hahm 2011 on this

pattern cross-linguistically). This can be observed in the examples in (i): the predicative adjective agrees

with the subject, yet despite the subject being (formally feminine) LEI, agreement is masculine when LEI

refers to a man, but feminine when LEI refers to a woman. See Wurmbrand 2016 for one view of how

agreement targets either uninterpretable or interpretable features of polite pronouns.

(i) a. Dottor

Doctor

Biagi,

Biagi

Lei

LEI.NOM

be.3S G

così

{buono/

good.M.SG/

#buona}.

good.F.SG

‘Doctor Biagi, you are so good.’ (male addressee)

b. Dottoressa

Doctor

Biagi,

Lei

LEI.NOM

be.3S G

così

{buona/

good.F.SG/

#buono}.

good.M.SG

‘Doctor Biagi, you are so good.’ (female addressee)

continue, thereby bleeding an Agree relation with the indirect object. This results in a

(Strong) PCC effect, and the indirect object cannot be cliticized.

(32)

Appl

[INT:ϕ

SAT:PART]

V DO

[PART]



Consider now how LEI is expected to behave in this account. Recall that LEI is treated

as 3rd-person for verbal agreement, including as an object clitic. Under this analysis, the

probe should agree with a direct object LEI as if it is third-person and it should therefore

fail to be satisﬁed, as it does not see a [PART ] feature. It should therefore be able to agree

with the indirect object (as in licit 1>3, 2>3, and 3>3 constructions) and no PCC effect

should obtain, contrary to fact.

5.1.2 Coon and Keine 2021

As another recent representative, consider the ‘Feature Gluttony’ analysis from Coon

and Keine 2021. In their analysis, PCC effects can arise when a probe enters into an Agree

relation with multiple goals, as this can lead to crash-inducing conﬂicts for subsequent

operations such as clitic movement and morphological realization. In their system, this

situation arises for an articulated probe (i.e. one with multiple feature segments) if it

ﬁrst encounters an argument with fewer person features, followed by an argument with

a richer set of person features. This is schematized in (33) for a 3>2 environment.

(33) vP







uPERS →1

uPART →2



π

B



uNUM

uPL



#





ApplP

DPIO

[3SG]

[[PE RS]1, [N UM]]

Appl0

Appl DPDP

[2SG]













PERS

PART

ADDR







,[NUM]







In Coon and Keine’s system, a ‘gluttonous’ probe bearing multiple feature sets is not

itself a problem, but rather, it can lead to derivational problems that refer to the feature

sets being copied onto the probe—such as cliticization, which they assume to be oblig-

atory in the case of an Agree relation with a clitic-doubling probe (p. 671). For the 3>2

situation schematized in (33), cliticization to the probing head would fail, because the

syntax can neither cliticize both arguments simultaneously (which is impossible for a

binary-operating Merge) nor can it do so sequentially, as this would produce a step in

the derivation where one argument is indeed clitized but the obligation to cliticize the

other argument is ﬂouted.

Since verbal agreement treats LEI as third-person, there is no reason to believe that its

interpretable participant features are visible to the Agree operation; thus it should pat-

tern with 3>3 environments, which are grammatical. For Coon and Keine, 3>3 conﬁgu-

rations ﬁrst involve a π-probe, which is not gluttonous because the lower direct object

does not bear more person features than the already identiﬁed indirect object. Subse-

quently, after cliticization has rendered the indirect object invisible to subsequent Agree

relations (following Anagnostopoulou 2003; Béjar and Rezac 2003; Preminger 2009), a

#-probe agrees with the direct object. In their system, LEI and other third-person po-

lite pronouns should pattern as third-person arguments, thereby allowing double clitic

conﬁgurations, contrary to what we ﬁnd with LEI.

5.1.3 The Person Licensing Condition

A distinct but related type of syntactic approach is pursued by Béjar and Rezac 2003; see

also Béjar and Rezac 2009; Preminger 2009; Rezac 2011; Preminger 2014, and see Coon

and Keine 2021 for criticism. According to such an account, PCC effects arise as a result

of a Person Licensing Condition, which requires 1st/2nd-person features to be licensed

via an Agree relation. In Béjar and Rezac’s analysis, 1st/2nd-person direct objects violate

the Person Licensing Condition when they occur with an indirect object, thereby yield-

ing a PCC effect: a π-probe merged higher than both objects ﬁrst sees the more local

indirect object, and consequently cannot agree in person with the direct object, thereby

failing to license it. (A second #-probe can Agree with the 1st/2nd-person direct object,

but not in person, and therefore cannot license it.)

In Béjar and Rezac’s system, the fact that LEI gives rise to PCC effects as a direct ob-

ject in ditransitives suggests that LEI behaves as a participant for the purposes of the

Person Licensing Condition. The problem, however, is that as we have seen, when ap-

pearing as a direct object without an indirect object, LEI is agreed with (and presumably

thereby licensed) as though it is third-person. In order to account for the behavior of

LEI with respect to both the PCC and ordinary verbal agreement, Béjar and Rezac would

thus have to assume that Agree licenses the interpretable φ-features on LEI even though

it copies its uninterpretable features. Notice, however, that this is effectively a syntac-

ticosemantic account: the interpretable person features factor into the PCC while the

features expressed in agreement do not.

A variant of this analysis comes from Rezac (2011), who formulates the Person Case

Constraint as being about (case-)licensing a feature [+PERSON] on the (would-be) ac-

cusative object (p. 98). This type of analysis can accommodate hierarchy effects that ex-

tend to other non-participant nominals, including (third-person) reﬂexives in French,

which also induce a PCC-type effect in accusative object position (Rezac 2011:2). The

idea is that [+PERSON] can be cross-linguistically variable to some degree in what types

of nominals it is assigned to, such that reﬂexives in one language carry [+PERSON] and

therefore induce PCC effects, while in another language, reﬂexives do not bear such fea-

tures and therefore fail to give rise to PCC effects (e.g. p. 299).

The same logic for [+PERSON] is applied to third-person animate nominals. As ob-

served by Ormazabal and Romero (2007) (also in their antecedent work), Spanish speak-

ers of so-called leísta varieties distinguish between accusative clitics that are used for

inanimates (lo/la – masculine singular and feminine singular, respectively) vs. those that

are used for animates (le, syncretic with the dative); a hierarchy effect obtains only with

the animate (34). This can be made sense of within Rezac’s system if third-person ani-

mates in Spanish (but not necessarily in other languages) are assigned the feature [+PER-

SON].

(34) Te

2SG.DAT

{lo

3SG.ACC

/*le}

3SG.ACC

di.

gave

‘I gave it/*him to you.’ Spanish leísta varieties (Ormazabal and Romero 2007:321)

In this type of analysis, the PCC effects observed with LEI could be attributed to

the obligatory presence of [+PERSON] on polite pronouns. This is in fact what Rezac

(2011:298-299) suggests for the third-person polite pronoun in Spanish USTED (includ-

ing in laísta varieties), which is also brieﬂy noted to give rise to PCC effects; see further

discussion of Spanish in Section 6.

While this type of analysis can be made to work mechanically, because Rezac’s [+PER-

SON] feature is not tied explicitly to exponence or to interpretation, its presence on spe-

ciﬁc kinds of elements outside of participant pronouns is essentially arbitrary (Rezac

2011:299); as far as we can tell, [+PERSON] is stipulated in order to induce PCC effects.

Absent other types of evidence for [+PE RS ON ] within a language, it is difﬁcult to see how

a learner would come to posit this feature, and if learners are in fact free to posit it or not,

we may may expect variation across speakers and across languages with third-person

polite pronouns, with only some speakers opting to assign polite pronouns [+PERSON],

contrary to what we ﬁnd with LEI.

5.2 Towards a Syntacticosemantic Account

We suggest instead that a syntacticosemantic account of the PCC is on the right track.

We view the fundamental property needed from such an account to be that these effects

are tied to conﬂicts arising from interpretable person features rather than being inher-

ently tied to the formal features that are overtly reﬂected in morphosyntactic agreement.

While it is not our objective to argue for a particular implementation, we sketch how one

recent account could contend with the PCC effects found with LEI.

We take a recent representative of the syntacticosemantic perspective to be the pro-

posal by Pancheva and Zubizarreta (2018) (henceforth “P&Z”; see also Charnavel and

Mateu 2015 and Stegovec 2020). P&Z attribute PCC effects to the encoding of point-

of-view centers within a phase deﬁned by some argument-introducing verbal head –

for them, Appl. In brief, this head enters into an Agree relation, necessarily with inter-

pretable person features, in order to establish its goal as a point-of-view center.

P&Z propose that PCC effects – including those referred to as ‘Strong’, ‘Weak’, ‘Strictly

Descending’, etc. – are captured by the P(erson)-Constraint in (35). Their P-Constraint

has several components, which are all subject to variation in parametric setting (see also

Zubizarreta and Pancheva 2017).

(35) P-Constraint on phases αheaded by an interpretable p(erson)-feature

a. The interpretable person feature is present on all heads of a certain func-

tional category (default), unless restricted. (Domain of application)

b. There must be an n-valued D located at the edge of αthat enters into an

agreement relation with the n-valued interpretable person feature on the

head of α.nis [+PROXIMATE] (default) or restricted to [+ PARTICIPANT ] or

[+AUTHOR]. (P-Prominence)

c. There can be at most one DP in αeligible to agree with the interpretable

p-feature on the head of α.(P-Uniqueness)

d. If there is more than one DP that can agree with the interpretable p- fea-

ture on the head of α, the DP marked [+AUTHOR] is the one that agrees.

(P-Primacy) (P&Z:1300)

In this system, [+PROXIMATE] is a feature motivated by proximate/obviative distinc-

tions made in some languages, and “proximate arguments are the ones suitable to be

perspectival centers” (p. 1300). It falls in an implicational hierarchy with the person

features [±PARTICIPANT] and [±AUTHOR], the plus values of which imply the presence

of [+PROXIMATE]. (In other words, all ﬁrst person and second person arguments are

[+PROXIMATE], being suitable perspective-holders, but only some third-person argu-

ments can be [+PROXIMATE], namely those that occur in the same domain as another

third-person argument.)

To derive the Strong PCC, whereby 1st-and 2nd-person arguments are altogether

banned from the direct object position in a double object construction, the setting in

(a) is active, (b) has its default setting of [+PROXI MAT E], and P-Uniqueness (c) is active

((d) is not). This rules in 3>3 conﬁgurations (as long as the IO counts as [+PROXIMATE])

but rules out 3>1 and 3>2 as follows: (b) requires that the IO be [+PROXIMATE], which is

not possible for third-person unless it occurs with another third-person argument, and

even if it did, P-Uniqueness (c) would be violated with 1st and 2nd direct objects, which

are also inherently [+PROXI MATE]. For the Weak PCC, P-Uniqueness is not active, and

thus 1>2 and 2>1 are possible for this grammar type, since there is no issue with having

multiple arguments with [+PROXIMATE] in the same domain.

We now turn to PCC effects with LEI. The polite pronoun should bear [+PROXIMATE]

by virtue of having interpretable 2nd person features (and as shown above in (15), LEI

can indeed serve as an IO as expected). Considering the Strong PCC grammar, LEI in

DO position should be ungrammatical with a third-person IO because i) the IO cannot

be [+PROXIMATE] in the context of (what is interpreted as) a non-third person argument,

and ii) even if the IO could be [+PROXIMATE], this would lead to the co-occurrence of two

[+PROXIMATE] arguments, thereby violating P-Uniqueness. For the Weak PCC grammar,

(i) still holds, thereby also giving rise to a PCC effect for 3>LEI.1415

The P&Z account shares with morphosyntactic analyses the idea that Agree-based

relations are complicit in the derivation of PCC effects. It is therefore able to capture

syntactic aspects of the phenomenon, such as the fact that clitics are subject to PCC

effects while stressed pronouns are not, under the widely held view that cliticization is

a consequence of an agreement relation (see Pancheva and Zubizarreta 2018:1321-1324

for discussion, and Anagnostopoulou 2003; Béjar and Rezac 2003; Preminger 2019; Coon

and Keine 2021; Deal 2024; among others).

A crucial difference between the type of syntacticosemantic analysis offered by P&Z

and a morphosyntactic analysis, however, is that the agreement relation has an interpre-

tive motivation (e.g. to identify point-of-view centers) and thus the person features that

are considered are naturally the interpretable ones rather than the formal features that

come to be expressed overtly in verbal agreement morphology.16 This is thus where an

14P&Z actually propose that P-Uniqueness is evaluated by having two person features on Appl: one that

is interpretable and agrees with the IO and one that is uninterpretable and agrees with the the DO. Since

this mechanism is meant to evaluate whether there are multiple potential point-of-view centers, we would

assume that the unvalued probe is instead sensitive to the interpretable features of the DO (thereby in-

ducing a PCC effect with LEI).

15Though P&Z do not discuss the Fancy Constraint, we suggest that their analysis can be extended to

this type of person hierarchy effect. Recall from (24) that in the faire inﬁnitif, a third-person causee could

co-occur with an accusative clitic that is third-person but not second-person, and that LEI patterned with

the second-person. In the detailed structural analysis of faire inﬁnitif s in Folli and Harley 2007 (following

insights also from antecedent literature), causative structure is introduced by a head vca us (not present

in the faire par construction, which also does not give rise to Fancy Constraint effects), and embeds a vP.

It could be that the vcaus head is like Appl in P&Z’s system, establishing another point-of-view domain.

Alternatively, we may follow Sheehan (2020) and Deal (2024) in taking faire inﬁnitif causatives to directly

involve Appl itself rather than vcaus. We leave the question open.

16A different syntacticosemantic agreement relation is proposed to account for PCC effects by Stegovec

(2020), who suggests that clitic pronouns enter the syntactic derivation as minimal pronouns (in Kratzer’s

2009 sense), whose interpretable person features must either be valued via Agree with a verbal head or be

analysis of the PCC as a syntax/semantics interface issue can make sense of the polite

pronoun data but where morphosyntactic accounts falter.

Before we conclude this section, we would like to return to the 1>2 and 2>1 patterns.

Recall that 1>LEI and LEI>1 appear to be unacceptable, even for speakers who otherwise

tolerate 1>2 and/or 2>1. While we are unsure of how this should be derived, we would

like to suggest that it seems consistent with the perspective that the Strong PCC is the

‘default’ (including in the P&Z system), with the Weak PCC being a more marked variant.

Ormazabal and Romero (2007:332-334) in fact suggest for Spanish that combinations

of ﬁrst and second pronouns appear to have a special status for speakers who tolerate

these combinations, only being possible with particular verbs and patterning with non-

argument clitics in various respects (see related discussion in Rezac 2011:150-151).

If this characterization is correct, this could mean within P&Z’s system that (35-c)

is invariably active, with ‘Weak’ PCC speakers actually just being Strong PCC speakers

who permit non-argument structure with some 1>2 and 2>1 conﬁgurations (which are

therefore not subject to PCC effects). In the case of LEI, this exceptional structure ap-

pears not to be permitted; both the ﬁrst person and LEI clitics are argument clitics and

therefore induce PCC effects for Strong and Weak speakers alike.

valued as third-person by default (see especially Stegovec 2020:273–276).

According to this account, PCC effects arise because “in a structure with two proDF [. ..] only the higher

proDF can receive a [π] value from [the verbal head]. In doing so, it blocks [π]-valuation of the lower proDF ,

which then gets a default value, that is 3P” (p. 274). In the case of DO polite pronouns, valuation by the

verbal head of the IO’s person features blocks valuation of the DO’s person features, resulting in inter-

pretable default third-person, inconsistent with the addressee interpretation. The account can therefore

correctly capture the PCC behavior observed with polite pronouns.

While the implementation is different, Stegovec’s view is similar to that of P&Z in that for him, valued

person features on the verbal head are present to encode “speech act participant perspective” (p. 276).

Stegovec follows Charnavel and Mateu 2015 in suggesting an interpretability-based view is supported by

PCC effects being ameliorated in non-de-se environments: Stegovec notes this for Slovenian, while Char-

navel and Mateau note this for Spanish and French. See brief discussion of non-de-se environments in

Italian in Section 7.

6 Other languages

On the assumption that PCC effects with ditransitives have the same grammatical prove-

nance cross-linguistically, the general prediction of a syntacticosemantic account of the

PCC is as in (36).

(36) PCC + Politeness Prediction: If a language displays PCC effects in ditransitives

for second-person arguments and has a third-person addressee-referring polite

pronoun, this pronoun should also give rise to PCC effects.

This is a strong and falsiﬁable prediction, which can be tested across languages that

both exhibit PCC effects with second person pronouns and have a third-person polite

pronoun. However, the prediction warrants two important caveats. First, it is possible

that a language has what might look like a third-person polite pronoun but is actually

a non-pronominal (third-person) imposter. We observed for Italian in Section 4.3 that

this distinction matters for the PCC, and expect this difference to carry over to other

languages, as well.

Second, while we take PCC effects with ditransitives to share a common grammatical

source across languages, we do not necessarily expect all other person-hierarchy effects

to be attributable to the same source.17 It has in fact been reported that some person-

hierarchy effects outside of the PCC are ameliorated by morphological syncretism be-

tween verbal inﬂectional forms, a phenomenon which has been argued to be more con-

sistent with an approach that attributes these person-hierarchy effects to morphological

conﬂicts that result in ineffability, as has been discussed for Icelandic dative-nominative

constructions in particular (e.g. Schütze 2003; Sigurðsson and Holmberg 2008; Coon and

Keine 2021; Hartmann and Heycock 2022).18 If the exponence-based view is on the right

17Alternatively, see Zubizarreta and Pancheva 2017 for the view that direct/inverse-type effects can be

accounted for with the same type of P-Constraint applying to other verbal heads.

18P&Z suggest (fn. 17) that their analysis can be extended to dative-nominative constructions if their

track for such effects, then we would like to suggest that, if a language has a person-

hierarchy effect that is ameliorated by syncretism, then no problem with exponence

should arise with a third-person polite pronoun; this is in contrast to the expectation

in (36).

In this rest of this section, we discuss two languages that have third-person polite

forms and have been shown to exhibit person-hierarchy effects: Spanish (6.1) and Ger-

man (6.2). We show ﬁrst that Spanish conforms to the prediction in (36) for ditransitive

PCC environments. For German, we ﬁrst demonstrate that the known ditransitive PCC

environment in German cannot be tested properly with a polite pronoun, for indepen-

dent reasons. We then show, however, that a distinct person-hierarchy effect, which is

known to be ameliorated by syncretism of verbal forms (Keine et al. 2019; Coon and

Keine 2021), does treat the third-person polite pronoun as if it were third person, and

thus no effect obtains with the polite pronoun, as is consistent with our expectation.

6.1 The PCC and Spanish USTED

Beyond Italian, it is well-known that PCC effects obtain across other Romance languages,

as well (e.g. Bonet 1991), though the interaction of person-hierarchy effects and third-

person polite forms are often not testable in other Romance languages because po-

liteness forms are frequently accomplished via second-person plural marking rather

than third person: this is true for French and Romanian (e.g. Wechsler and Hahm 2011).

One notable exception is Spanish, which exhibits PCC effects (see Bonet 1991, 1994; Or-

mazabal and Romero 2007; among many others) and has the third-person polite form

USTED.19

Like LEI,USTED behaves as a third-person argument despite referring to an addressee:

point-of-view principles can also apply to heads outside of Appl, such as to a speciﬁc ﬂavor of v. Here we

simply note that the syncretism data are not readily compatible with this extension, as far as we can tell.

19Another exception is Portuguese, which has a polite form você that takes third-person verbal agree-

ment (see e.g. Wang 2023:1307).

in addition to ﬁnite verbal agreement with USTED (see e.g. Wang 2023:1306), third-person

clitics are also employed for polite reference, including for the ACC clitic in laísta dialects

(as shown in (37)-(38)).20 (USTED, like LEI, also binds third-person reﬂexives.)

(37) a. Yo

1SG.NOM

2SG.ACC

respeto.

respect.1S G

‘I respect you.’

b. Yo

1SG.NOM

3SG.ACC

respeto.

respect.1S G

‘I respect her.’

(38) Yo

1SG.NOM

USTED.ACC

respeto

respect.1S G

(DOM

usted).

USTED)

‘I respect you.’

Rezac (2011:298) observes that a PCC effect obtains in Spanish with USTED; see also

Jambrovi´c to appear. The phenomenon can be seen in (39): as is well-known (see Perl-

mutter 1970; Bonet 1991; Nevins 2007, among many others), the third-person dative

clitic is realized as what is called ‘spurious’ se, a form syncretic with the reﬂexive-like

clitic se. In (39), the accusative clitic la is grammatical in a 3>3 conﬁguration if it refers

to a third-person feminine singular referent, but it cannot be used as a polite USTED

form in this expression. (The expression remains ungrammatical if se is changed to the

dative clitic le.)

(39) Sei

SE(=3PL.DAT)

{3F.SG.ACC /*USTED.ACC}

presentaré

present.FUT.1SG

los

the

estudiantesi).

students

‘I will introduce her/*you (formal) to the students.’ (Rezac 2011:298)

Three Spanish speakers we have consulted (Francesc Torres-Tamarit, p.c. and Cristina

Real, p.c.) also report a contrast for the sentences in (40). In the well-formed expression

in (40-a), USTED appears in the spurious se form of the dative with an accusative third-

person clitic (USTED>3), which is expectedly grammatical. (In the parenthetical contin-

20One difference between LEI and USTED is that the forms of the former are all identical to that of a

different pronominal series (the feminine singular third-person forms), whereas the latter includes the

citation form usted, which for many varieties is a dedicated politeness form not seen elsewhere among

Spanish personal pronominal forms.

uation, the accusative third-person clitic is coreferent with the direct object of the pre-

ceding clause, and a null USTED subject triggers third-person agreement on the verb.)

The infelicitous sentence in (40-b) is string-identical, though as reﬂected in the gloss-

ing, has an intended reading whereby USTED is the accusative clitic and the dative se is

a third-person argument (hence 3>USTED). This reading is unavailable.

(40) a. Yo

1SG.NOM

{3SG.DAT/*USTED.DAT}

3M.SG.ACC

encomendé

entrust.PST.1SG

(con

with

the

esperanza

hope

que

that

3M.SG.ACC

cuidara

take.care.SBJV.IMPF.3SG

bien).

well

‘I entrusted him to you (in the hope that you would take good care of him).’

b. #Yo

1SG.NOM

3SG.DAT

USTED.ACC

encomendé

entrust.PST.1SG

(con

with

the

esperanza

hope

que

that

USTED.ACC

cuidara

take.care.SBJV.IMPF.3SG

bien).

well

‘I entrusted you (formal) to him/her (in the hope that (s)he would take good

care of you).’

Italian and Spanish thus exhibit the same pattern with respect to their polite pronom-

inals, in that both give rise to PCC effects.21

6.2 Person-hierarchy effects and German SIE

We now turn to German, which shares with Italian both i) the use of third-person polite

pronouns for addressees; and ii) the appearance of person-hierarchy effects, including

the PCC. We ﬁrst consider PCC environments, which appear in a small corner of the

language (Anagnostopoulou 2008). Unfortunately, our prediction cannot be tested with

polite pronouns because, as we show, the conditions necessary to observe PCC effects

in the language cannot be satisﬁed with the pronouns in question. We then consider

a second effect reported by Keine et al. 2019; Coon and Keine 2021, in which copular

‘assumed-identity’ contexts exhibit person-based restrictions. This person-hierarchy ef-

21We also note that the PCC effect with USTED may challenge Collins and Ordóñez’s (2021) contention

that USTED is an imposter, if Spanish behaves like Italian in permitting imposters to appear in PCC envi-

ronments felicitously (as discussed for Italian in Section 4.3). It is alternatively possible that Collins and

Ordóñez’s characterization of USTED is only applicable to the form usted but not to the clitic forms.

fect is different from PCC effects in that there is an accompanying number-hierarchy

effect, and even more strikingly, these effects are ameliorated by syncretism. We as-

sume that amelioration by syncretism supports an exponence-based analysis of person-

hierarchy effects, and our expectation is therefore that a third-person polite pronoun

should not give rise to such effects. As we show, this is borne out.

German has a set of polite pronominal forms that uses the third-person plural series

across cases (nominative, accusative, dative, and possessive: Sie, Sie, Ihnen, Ihr, respec-

tively); we refer collectively to this set as SIE. As in Italian, polite subjects trigger agree-

ment as if they were third person, and are ungrammatical with second-person agree-

ment, as shown in (41). SIE also necessarily binds a third-person reﬂexive sich rather

than employing the second-person singular form dich. (Unlike Italian Lei, German Sie

can be used for both singular and plural addressees.)

(41) a. Sie

3PL.NOM

{sind/*bist}

be.3P L/be.2SG

nett.

nice

‘They are nice.’

b. Sie

SIE.NOM

{sind/*bist}

be.3P L/be.2SG

nett.

nice

‘You (SG/PL formal) are nice.’

Because Sie refers to (an) addressee(s), the expectation under the current view is that

any PCC effect in German that rules out combinations with a second person pronoun

should also obtain with Sie. While the PCC had previously been thought not to be oper-

ative in German, Anagnostopoulou (2008) ﬁnds that it is, albeit in a small corner of the

language. In particular, (weak) PCC effects only arise with pronominal clusters in the

so-called Wackernagel position when the cluster appears before the subject. As shown

in (42-a), it is possible to have a cluster of two weak pronouns precede a subject (in

ACC >DAT order; on which see Müller 1999 and references therein). However, certain

person combinations are ill-formed (at least for speakers Anagnostopoulou consulted),

including when the direct object is ﬁrst or second person and the indirect object is third

person; this is shown for a second-person object in (42-b). Anagnostopoulou observes

that the effect is very restricted; for example, it vanishes when the subject appears before

the pronominal cluster (Anagnostopoulou 2008:26).22

(42) a. dass

that

3SG.N.NOM

ihm

3SG.M.DAT

der

the.NOM

Fritz

gegeben

given

hat

has

‘that Fritz gave it to him’

b. *weil

because

dich

2SG.ACC

ihm

3SG.M.DAT

irgendwer

someone.NOM

vorgestellt

introduced

hat

has

‘because someone introduced you to him’ (Anagnostopoulou 2008:24-26)

The example in (42-a) shows that combinations of 3>3 are possible in this type of

expression. Unfortunately, it seems not to be possible to have this combination if the

direct object uses a 3P L pronoun sie (43) (and the same is true if sie picks out a third-

person singular feminine referent).23

(43) dass

that

{es

3N.SG.ACC

??/*sie}

3PL.ACC

ihm

3M.SG.DAT

dieser

the.NOM

Mann

man

gegeben

given

hat

has

‘that this man has given it/*them to him’ (e.g. the book, the books)

That the third-person use of sie is severely degraded in this construction means that

we cannot properly test whether polite SIE obviates PCC effects: the relevant (plural)

third person accusative pronoun cannot appear fully acceptably in this position in the

ﬁrst place. That being said, polite use of SIE is also bad in this position, consistent with

our prediction, though confounded by this problem.

(44) *dass

that

Sie

SIE.ACC

ihm

3M.SG.DAT

dieser

this

Mann

man

vorgestellt

introduced

hat.

has

‘that this maniintroduced you (polite) to himj.’

A different person-hierarchy effect in German is observed by Keine et al. (2019) and

22Judgments for such sentences seem unstable, though we were able to replicate Anagnostopoulou’s re-

ported generalization more reliably when preceding context licensed stress appearing on a deﬁnite nom-

inative subject.

23We do not understand the source of this restriction. Plausibly, the issue is related either to prosody or

to syncretism of sie across nominative and accusative cases, though the latter possibility is problematized

by the fact that the neuter singular es can go in this position despite also being syncretic across these two

cases.

is discussed further by Coon and Keine (2021), who ﬁnd that both person and number

hierarchies are implicated in copular ‘assumed-identity’ expressions (on which, see also

Heycock 2012). In particular, these authors present experimental evidence indicating

that German speakers permit 1/2 >3 sentences like (45-a), but do not allow 3 >1/2

(45-b). They also ﬁnd a number hierarchy effect, whereby PL >SG is possible (46-a), but

the reverse is not (46-b). (It is crucial that the copular sentences have the interpretation

that DP1 takes on the role of DP2 and not the other way around.)

(45) a. Du

2SG.NOM

bist

be.2S G

Martin.

Martin

‘You are Martin.’

b. *Martin

Martin

ist

be.2S G

du.

2SG.NOM

‘Martin is you.’ (C&K2021:685)

(46) a. Die

the.PL

Kinder

child.PL

sind

be.3P L

der

the.SG

Baum.

tree

‘The children are the tree.’

b. *Maria

Maria

ist

be.3S G

die

the.PL

Bäume.

tree.PL

‘Maria is the trees.’ (C&K2021:685)

While there is a clear resemblance to PCC environments, this context is nevertheless

different in four respects: i) there is no obvious sense in which applicative structure is

employed; ii) there is an accompanying number hierarchy effect; iii) the case of the two

nominals is nominative (rather than e.g. DAT-ACC) for both; and iv) it is reported that

syncretisms between the agreeing forms have an ameliorating effect on the hierarchy

effects.24 On (iv), Keine et al. (2019) report that judgments improve when the form of

the copula is compatible with agreement with either nominal, and this is the case for

1SG and 3SG forms of the copula in past tense forms (47) and in the subjunctive; this is

expected under an account that ties the effect (in part) to an issue of exponence.

(47) Er

3SG.NOM

{*ist

be.3S G

/?war}

/be.PST.3S G/1S G

ich.

1SG.NOM

‘He is/was me.’ (Keine et al. 2019:4,29)

24As far as we can tell, P&Z’s account does not predict this person hierarchy effect, though it does not

rule it out, either, if non-PCC person hierarchy effects may also arise in other parts of the grammar.

Because of the syncretism facts in particular, our expectation for this type of person-

hierarchy effect is that a third-person polite pronoun will behave like any third-person

pronoun in not giving rise to the effect. This is indeed borne out.

First observe the person-based contrast when the subject (DP1) is a third-plural

nominal (48): when the predicative nominal (DP2) is a third-person plural pronoun, the

expression is grammatical, but if DP2 is a second-person plural pronoun, the expression

is ungrammatical.25

(48) Die

the.NOM.PL

Zwillinge

twins

sind

be.3P L

{(?)sie

3PL.NOM

/*ihr}.

2PL.NOM

‘The twins are them/*you.’

In Coon and Keine’s (2021) analysis, the ungrammaticality in data like (48) arises be-

cause a probe on T agrees with both DPs, leading to a ‘gluttonous’ conﬁguration in which

the person feature speciﬁcations between the two goals clash. As Coon and Keine clar-

ify, this is itself not a problem; the issue is that the T probe cannot realize the clashing

feature sets with a single vocabulary item, thereby yielding ineffability.

A prediction of any such exponence-based account is that no such complication

should arise if DP2 is polite SIE rather than a second-person form. (For Coon and Keine’s

analysis, the probe on T should ﬁnd [P ER S] and [PL ] on both DPs, and should therefore

be able to be realized with third-plural morphology.) This is borne out, as shown in (49).

(49) Die

the.NOM.PL

Zwillinge

twins

sind

be.IN DC .3PL

Sie.

SIE.NOM

‘The twins are you (SG /PL).’ (also: ‘you are the twins’)

We have thus identiﬁed a person-hierarchy effect in German that conforms to the

predictions of a morphosyntactic analysis rather than a syntacticosemantic one. Given

25Some speakers we consulted report that seid ‘be.2SG’ is possible for them with the relevant inter-

pretation when DP2 is the second-person plural ihr, though this is ungrammatical for other speakers we

consulted. Experimental results from Keine et al. 2019 similarly indicate that speakers ﬁnd DP2 agreement

ungrammatical in contexts where person and number hierarchies are not respected.

the differences between PCC environments and the German assumed-identity context

with respect to their behavior with polite pronouns, we have suggested that PCC envi-

ronments have a syntacticosemantic source whereas other person-hierarchy effects may

have a morphosyntactic source. This also converges with Drummond and O’Hagan’s

(2020) recent conclusion that person-hierarchy effects across constructions and lan-

guages might have heterogeneous sources (in their proposal, syntactic sources as op-

posed to morphological sources). 26

7 Concluding Remarks

We found that the Italian polite pronoun LEI does not conform to the predictions of a

morphosyntactic account of PCC effects: despite LEI’s third-person verbal agreement

behavior, it patterns with second person in giving rise to PCC effects. We suggested that

these data are more compatible with a syntacticosemantic analysis of PCC effects such

as that of Pancheva and Zubizarreta 2018. We emphasize here that this is not intended

as an endorsement speciﬁcally of P&Z’s account; see especially Deal 2024 for criticism.27

26We know of at least one other language in which the prediction in (36) might be testable. According

to Wang (2023:1297), several languages make use of third-person in (certain types of) politeness contexts,

including the Pama-Nyungan language Warlpiri (see discussion in Wang 2023:1300, citing Laughren 2001),

and this language also has been reported to exhibit PCC effects (Haspelmath 2004:7, citing Hale 1973:334;

see theoretical discussion in Stegovec 2015).

The only other relevant prior claim that we are aware of comes from Preminger (2014:124-125), who

suggests that in the language K’ichee’, a pronoun used for formal address behaves more like a third-person

argument for the agreement purposes of the Agent Focus construction (citing a footnote from Stiebels

2006:526). However, no empirical support is provided for this, and moreover, this characterization is about

the choice of agreement controller in the construction, rather than with person restrictions on the co-

occurrence of arguments. Regardless, we assume that the Agent Focus construction should be subsumed

under the same umbrella of person-hierarchy effects as assumed-identity contexts, since there is likely no

applicative structure involved.

27One empirical concern about P&Z’s account is that their P-Prominence condition is meant to be pa-

rameterizable, such that the interpretable feature on the Appl head may be [+PARTIC IPA NT ] rather than

[+PROXIM ATE], which they make use of in the derivation of so-called ‘Super-Strong’ PCC effects which

ban combinations of 3>3 in some languages including Kambera. However, if we understand the system

correctly, this part of P&Z’s analysis seems to incorrectly predict that such languages do not tolerate third-

Rather, we submit that a syntacticosemantically based theory of the PCC is on the right

track, with P&Z’s point-of-view account being a recent representative example (see re-

latedly Charnavel and Mateu 2015 and Stegovec 2020).

While we evaluated morphosyntactic proposals of the PCC in Section 5 assuming a

particular dual-feature analysis of polite pronouns, alternative analyses of polite pro-

nouns can also be found in the literature; in 7.1, we show how these alternatives fare for

LEI and for PCC effects. We conclude with a discussion of outstanding issues pertaining

to the so-called Clitic Logophoric Restriction and to non-de-se environments (7.2).

7.1 The representation of polite pronouns

7.1.1 Wang 2023

Unlike our assumed dual-feature representation of polite pronouns, Wang’s (2023) pro-

posal for polite pronouns is semanticopragmatic in character. Considering a large sam-

ple of languages, Wang shows that the feature values ‘recruited’ for polite pronouns

across many languages are plural number, third person, and indeﬁniteness, which have

independently been shown to be semantically unmarked values with the weakest pre-

suppositions (see relatedly Sauerland 2008). Wang argues that polite pronouns use these

unmarked values in morphosyntax, motivated by a pragmatic maxim referred to as the

Taboo of Directness, according to which politeness contexts require the weakest presup-

position (for some feature(s)). This account links politeness to avoidance of speciﬁc ref-

erence, and correctly captures the values of person, number, and deﬁniteness that are

coopted in politeness contexts across languages.

The application of Wang’s account to LEI encounters (at least) two problems. First,

person IOs altogether, which the data they cite from Kambera already indicate is not true (P&Z2018:1321,

ex. 36b) (this data point involves a lack of DO cliticization but is grammatical with a third-person da-

tive clitic). Similar issues are applicable for the derivation of me-ﬁrst PCC effects, which they derive via

P-Prominence requiring a [+AU THOR] feature. In general, the P&Z system seems to make incorrect pre-

dictions for when there is an IO clitic but no agreeing/cliticizing DO.

it does not correctly predict that the marked feminine gender is employed in Italian for

LEI rather than the unmarked masculine: it is not plausible that the Taboo of Direct-

ness could dictate the use of a feminine form instead of a masculine one (on gender

markedness in Italian, see e.g. Percus 2011), so gender features would minimally have

to be morphosyntactically speciﬁed not to be interpreted. Second, because reference to

the addressee is not grammatically encoded directly, but rather, is implied by the prag-

matic system under this view, a morphosyntactic or a syntacticosemantic analysis of the

PCC would predict that no PCC effects should obtain with third-person polite pronouns,

contrary to fact. Alternatively, a semanticopragmatic account of PCC effects under this

view would incorrectly predict that imposters should give rise to PCC effects, contrary

to fact (as discussed in Section 4.3).

7.1.2 Wechsler and Hahm 2011

The proposal for polite pronouns from Wechsler and Hahm 2011 is framed within a dual-

feature system known from the LFG literature (especially Wechsler and Zlati´c 2003), with

CONCORD and INDEX features, which are respectively “grammaticalizations of the form

and the meaning of the [agreement] trigger” (p. 265), respectively. In their system, CON-

CORD features include lexically speciﬁc features such as declension and gender, whereas

INDEX features are tied to properties of discourse reference, including person features.

W&H’s analysis of polite pronouns focuses cross-linguistically on second-person plu-

rals, which display a striking ‘mixed’ pattern across many languages (e.g. in French),

whereby verbal agreement is necessarily plural – agreeing with the formal number of

the pronoun – whereas adjectival agreement is singular if the addressee is singular. W&Z

propose that this dichotomy arises as a result of i) a division between agreement targets

that express CONCORD features or INDEX features, with the former agreeing in features

such as gender, number, and case (e.g. adjectives) and the latter agreeing in features

such as person, number, and gender (e.g. verbs); ii) speciﬁcation of polite pronouns

only for INDEX features in these languages, and iii) allowance for ‘semantic agreement’ in

cases of feature underspeciﬁcation. Because verbal agreement requires INDEX features,

it necessarily refers to the speciﬁed features of polite pronouns (e.g. plural for second-

person plural polite forms).

It is not clear from W&H’s system how polite pronouns come to be interpreted as

denoting addressees: INDEX features are typically associated with referential proper-

ties yet they are the grammatically speciﬁed features for polite pronouns – hence third-

person plural pronouns will trigger third-person plural verb agreement (see Wurmbrand

2016:fn. 34 for a related criticism). More problematically for present purposes, third-

person polite pronouns in the W&H system would bear third-person INDEX features,

in which case they should behave like third-person arguments for person agreement.

Since person is an INDEX feature in their system, an agreement-based account of the

PCC would predict under this system that no effects should obtain with polite pronouns,

contrary to fact.

7.1.3 Ackema and Neeleman 2018

Ackema and Neeleman (2018:42-50) propose to derive formal identity between polite

pronouns and other pronouns within a language via impoverishment at PF or LF. In their

system, polite pronouns necessarily bear a feature HON (honoriﬁc) along with second

person features (for them, PROX and DIST) in the syntax, with these features being visi-

ble to syntactic processes. Subsequently, deletion triggered by the feature HON can occur

either on the PF branch or the LF branch, giving rise to discrepancies between the real-

ization of polite pronouns (and their agreement targets) at PF and the interpretation of

these elements at LF. For example, Ackema and Neeleman suggest that second-person

plural polite pronouns, such as those found in many languages (e.g. French), should

be speciﬁed with a PL feature in the syntax, which is deleted in the context of HON at

LF, resulting in number-neutrality at LF despite the plural form of the pronoun. Corre-

spondingly, for third-person politeness pronouns such as the one found in German, the

addressee-speciﬁc person features (for them, PROX) are deleted at PF in the context of

HON, yielding the less marked realization of the pronoun as third person.

This style of analysis could in principle be applied to LEI, in which case, syntax-

speciﬁc analyses of the PCC could correctly generate the badness of LEI in PCC environ-

ments, as LEI would be represented with second-person features in the narrow syntax.

As in German, PF impoverishment would apply to delete second person features yield-

ing third-person forms. To derive its feminine form and agreement, LEI would need to

have a feature FEM in the syntax and undergo impoverishment at LF, such that it was

interpreted gender-neutrally.

As Ackema and Neeleman discuss, such an account predicts that, for syntactic pur-

poses, the polite pronoun should behave as a second-person pronoun.28 We would like

to suggest that this is in fact not borne out for LEI. As noted above in Section 3, LEI as an

object clitic triggers obligatory feminine agreement with a participle, in contrast to par-

ticipant clitics including the second-singular ti, for which agreement is only optional. If

we take the optionality of participial agreement with participant clitics to be attributable

to the syntax (cf. Guasti and Rizzi 2002:191; Belletti 2017:499), then LEI appears to behave

syntactically as third-person rather than second, contrary to expectation.

28Ackema and Neeleman discuss this prediction for the German third-plural polite pronoun SIE, sug-

gesting that two types of expressions distinguish second from third person, with the polite pronoun be-

having in line with the former: i) adnominal pronoun constructions of the you linguists type – which are

possible for second person but not for third (as in English, you/*they linguists), and ii) relative clauses

headed by pronominals, which take resumptive subjects in German with second person but not with

third. We ﬁrst note that for both, the dual-feature analysis is compatible, so long as the interpretable

features can factor into the syntactic selection (which is poorly understood for both phenomena). Em-

pirically, regarding (i), these type of adnominal constructions are not possible with singular pronouns in

Italian for either second or third person, and therefore this cannot be tested; regarding (ii), there is no

known analogue in Italian, and we note that the judgments with second person are in fact more variable

than characterized by Ackema and Neeleman for German; see especially the experimental evidence from

Trutkowski and Weiß 2016.

7.2 Remaining Issues

One outstanding mystery internal to Italian, which we have heretofore not addressed,

concerns what Pancheva and Zubizarreta (2018) refer to as the the Clitic Logophoric

Restriction (CLR). As discussed by P&Z, following observations from Roca (1992) and

Charnavel and Mateu (2015) (see relatedly Ormazabal and Romero 2007 and Bhatt and

Šimík 2009), an accusative clitic referring to (but not necessarily c-commanded by) is

ungrammatical in the context of a third-person dative clitic in Spanish and in French,

but is reported to be grammatical when the dative is ﬁrst or second person. While the

CLR is discussed in the context of these other languages, we observe that the same gen-

eralization holds for at least some Italian speakers: in (50-a) a third-person dative with

a third-person accusative attitude holder induces a CLR effect, while for some speakers

(though not all we consulted), a second-person dative does not (50-a)-(50-b).

(50) a. #Mariai

Maria

credeva

believed

che

that

glie=k

3SG.DAT

lai

3F.SG.ACC

avremmo

have.CDL.1P L

afﬁdata

entrusted.F.SG

(alla

(to.the

reginak).

queen)

‘Mariaithought that we would entrust herito herk(the queenk).’

b. ?/#Mariai

Maria

credeva

believed

che

that

2SG.DAT

lai

3F.ACC

avremmo

have.CDL.1PL

afﬁdata.

entrusted.F.SG

‘Mariaithought that we would entrust herito you.’

P&Z suggest that the effect in (50-a), as well as its contrast with (50-b), can be cap-

tured under their account with the addition of a Point-of-View Principle (Pancheva and

Zubizarreta 2018:1328), which requires that “within a logophoric domain marking point

of view, if there are attitude holders among the event participants, one of them has to be

the point-of-view center”.

According to P&Z, a dative clitic like that in (50-a) is marked as a point-of-view cen-

ter because it enters into an Agree relation with Appl. The CLR effect is then induced

because the accusative clitic refers to an attitude holder while the dative clitic does not,

yet the dative is the point-of-view center, thereby violating the Point-of-View Principle:

the sole attitude holder in the domain is not a point-of-view center. In contrast, the idea

is that examples like (50-b) have two attitude holders (see their discussion on pp. 1326-

1327 on speech-act participants as attitude holders), and one of them, namely the dative

argument, is indeed the point-of-view center.

Regardless of the precise details for why the CLR should hold, we may ask whether

LEI patterns with third-person dative arguments in inducing the effect, or second-person

dative arguments in not giving rise to the effect. Under the view that LEI is interpretively

treated as a second-person argument, it should be eligible to be an attitude holder and

therefore, for speakers who accept (50-b), it should be possible for LEI to appear as a

dative alongside an accusative clitic attitude holder. However, this is not borne out; all

speakers, including those who accept (50-b), reject (51).

(51) #Dottoressa,

Doctor.F

Mariai

Maria

credeva

believed

che

that

Glie=

DAT.LEI=

lai

FEM.ACC.3

avremmo

have.CDL.1PL

afﬁdata.

entrusted.FEM.SG

‘Mariaithought that we would entrust herito you (formal).’

We are unsure of why (51) should pattern this way, and it is not expected under the

P&Z-type account, though we acknowledge that their analysis of the CLR is tentative.

A related issue comes from non-de-se environments. Charnavel and Mateu (2015)

suggest that their syntacticosemantic account of person-hierarchy effects found with

ditransitives is supported by amelioration of PCC effects in non-de-se environments

in both French and Spanish; amelioration is also observed for Slovenian by Stegovec

(2020). This is illustrated in the French example in (52), which has a 3>1 combination

that nevertheless does not induce (as strong of) a PCC effect.

(52) ?Ji’

have

rêvé

dreamed

que

that

j’

étais

was

Marilyn

Monroem,

Monroe

que

that

j’

étais

was

chez

house

Kennedyk

Kennedy

and

que

that

jem

mei

ACC.1SG

luik

DAT.3SG

présentais.

introduced

“Iidreamed that I was M. Monroem, that I was at Kennedyk’s house and that Im

introduced meito himk.’ (Charnavel and Mateu 2015:693)

While we agree that the amelioration effect is in line with a syntacticosemantic ac-

count, we were unable to replicate the amelioration pattern in Italian, including with

LEI, a topic that remains to be explored in future research.

Speaking more broadly, further research is needed in order to identify where in the

grammar person-hierarchy effects can and do reside, with the issue not being fully set-

tled in our view. The present article promotes the study of polite pronouns as a window

onto such phenomena, with the hope that it leads to further inquiry that can continue

to shed light on the PCC and other person-hierarchy effects.

References

Ackema, Peter, and Ad Neeleman. 2018. Features of person: From the inventory

of persons to their morphological realization. Cambridge, MA: MIT Press. URL

https://doi.org/10.7551/mitpress/11145.001.0001.

Adamson, Luke James. 2024. Split coordination with adjectives in Italian: An approach

with multidominance and semantic agreement. Natural Language & Linguistic The-

ory URL https://doi.org/10.1007/s11049-024-09617-5.

Adamson, Luke James, and Elena Anagnostopoulou. 2024. Gender features and coor-

dination resolution in Greek and other three-gendered languages. Linguistic Inquiry

1–76. URL https://doi.org/10.1162/ling_a_00543.

Adger, David, and Daniel Harbour. 2007. Syntax and syn-

cretisms of the person case constraint. Syntax 10:2–37. URL

http://dx.doi.org/10.1111/j.1467-9612.2007.00095.x.

Anagnostopoulou, Elena. 2003. The syntax of ditransitives: Evidence from clitics. Berlin:

Walter de Gruyter.

Anagnostopoulou, Elena. 2005. Strong and weak person restrictions: A feature check-

ing analysis. In Clitic and afﬁx combinations: Theoretical perspectives, ed. Lor-

rie Heggie and Francisco Ordoñez, 199–235. Amsterdam: John Benjamins. URL

https://doi.org/10.1075/la.74.08ana.

Anagnostopoulou, Elena. 2008. Notes on the Person Case Constraint in Germanic (with

special reference to German). In Agreement restrictions, ed. Roberta D’Alessandro,

Susan Fischer, and Gunnar Hrafn Hrafnbjargarson, 15–48. Berlin/New York: Mouton

de Gruyter. URL https://doi.org/10.1515/9783110207835.15.

Anagnostopoulou, Elena. 2017a. Gender and defaults. In A schrift to fest Kyle John-

son, ed. Nicholas LaCara, Keir Moulton, and Anne-Michelle Tessier, 23–31. Linguistics

Open Access Publications. URL https://doi.org/10.7275/R57D2S95.

Anagnostopoulou, Elena. 2017b. The Person Case Constraint. In The Wiley Blackwell

companion to syntax, ed. Martin Everaert and Henk van Riemsdijk, 3001–3047. Ox-

ford: Wiley-Blackwell, 2nd edition.

Béjar, Susana, and Milan Rezac. 2003. Person licensing and the derivation of

PCC effects. In Romance linguistics: Theory and acquisition, ed. Ana Teresa

Perez-Leroux and Yves Roberge, 49–62. Amsterdam: John Benjamins. URL

https://doi.org/10.1075/cilt.244.07bej.

Béjar, Susana, and Milan Rezac. 2009. Cyclic Agree. Linguistic Inquiry 40:35–73. URL

https://doi.org/10.1162/ling.2009.40.1.35.

Belletti, Adriana. 2017. (Past) participle agreement. In The Wiley Blackwell companion

to syntax, ed. Martin Everaert and Henk van Riemsdijk, 2972––3000. Oxford: Wiley-

Blackwell, 2nd edition.

Bhatt, Rajesh, and Radek Šimík. 2009. Variable binding and the Person-Case Constraint.

Paper presented at 25th annual meeting of the Israel Association for Theoretical Lin-

guistics (IATL), Ben Gurion University of the Negev.

Bianchi, Valentina. 2006. On the syntax of personal arguments. Lingua 116:2023–2067.

URL https://doi.org/10.1016/j.lingua.2005.05.002.

Bonet, Eulàlia. 1991. Morphology after syntax: Pronominal clitics in Romance. Doctoral

Dissertation, MIT, Cambridge, MA.

Bonet, Eulàlia. 1994. The Person-Case Constraint: A morphological approach. In

The morphology-syntax connection, ed. Heidi Harley and Colin Phillips, 33–52. Cam-

bridge, MA: MIT Working Papers in Linguistics 22.

Burzio, Luigi. 1986. Italian syntax: A Government and Binding approach. Reidel: Dor-

drecht.

Charnavel, Isabelle, and Victoria Mateu. 2015. The clitic binding restriction re-

visited: Evidence for antilogophoricity. The Linguistic Review 32:671–701. URL

https://doi.org/10.1515/tlr-2015-0007.

Collins, Chris, and Francisco Ordóñez. 2021. Spanish usted as an imposter. Probus

33:249–269. URL https://doi.org/10.1515/prbs-2021-0006.

Collins, Chris, and Paul Martin Postal. 2012. Imposters: A study

of pronominal agreement. Cambridge, MA: MIT Press. URL

https://doi.org/10.7551/mitpress/9780262016889.001.0001.

Coon, Jessica, and Stefan Keine. 2021. Feature gluttony. Linguistic Inquiry 52:655–710.

URL https://doi.org/10.1162/ling_a_00386.

D’Alessandro, Roberta. 2007. Impersonal si constructions: Agreement

and interpretation. Berlin/New York: Mouton de Gruyter. URL

https://doi.org/10.1515/9783110207514.

Deal, Amy Rose. 2024. Interaction, satisfaction, and the PCC. Linguistic Inquiry 55:39–

94. URL https://doi.org/10.1162/ling_a_00455.

Despi´c, Miloje. 2017. Investigations on mixed agreement: Polite plurals, hy-

brid nouns and coordinate structures. Morphology 27:253–310. URL

https://doi.org/10.1007/s11525-017-9301-3.

Drummond, Emily, and Zachary O’Hagan. 2020. Morphological and syntactic person

restrictions in Caquinte. In Proceedings of NELS 50, 197–206. Amherst, MA: University

of Massachusetts, GLSA.

D’Alessandro, Roberta, and Diego Pescarini. 2016. Agreement restrictions and agree-

ment oddities. In Manual of grammatical interfaces in Romance, ed. Susann

Fischer and Christoph Gabriel, 267–294. Berlin/New York: De Gruyter. URL

https://doi.org/10.1515/9783110311860-012.

Foley, Steven, and Maziar Toosarvandani. 2022. Extending the Person-Case Constraint

to gender: Agreement, locality, and the syntax of pronouns. Linguistic Inquiry 53:1–40.

URL https://doi.org/10.1162/ling_a_00395.

Folli, Raffaella, and Heidi Harley. 2007. Causation, obligation, and argument

structure: On the nature of little v. Linguistic Inquiry 38:197–238. URL

https://doi.org/10.1162/ling.2007.38.2.197.

Guasti, Maria Teresa. 1996. Semantic restrictions in Romance causatives and the incor-

poration approach. Linguistic Inquiry 27:294–313.

Guasti, Maria Teresa, and Luigi Rizzi. 2002. Agreement and tense as distinctive syntac-

tic positions: Evidence from acquisition. In Functional structure in DP and IP: The

cartography of syntactic structures, ed. Guglielmo Cinque, 167–194. New York: Oxford

University Press. URL https://doi.org/10.1093/oso/9780195148794.003.0006.

Hahm, Hyun-Jong. 2010. A cross-linguistic study of syntactic and semantic agreement:

polite plural pronouns and other issues. Doctoral Dissertation, University of Texas,

Austin.

Hale, Ken. 1973. Person marking in Walbiri. In A festschrift for Morris Halle, ed. Stephen

Anderson and Paul Kiparsky, 308–344. New York: Holt, Rinehart, and Winston.

Harris, Alice. 1981. Georgian syntax: A study in Relational Grammar. Cambridge: Cam-

bridge University Press.

Hartmann, Jutta M., and Caroline Heycock. 2022. Person effects in agreement with Ice-

landic low nominatives: An experimental investigation. Natural Language & Linguis-

tic Theory 41:1029–1090. URL http://dx.doi.org/10.1007/s11049-022-09564-z.

Haspelmath, Martin. 2004. Explaining the ditransitive Person-Role Constraint: A usage-

based approach. Constructions 2:1–71.

Heycock, Caroline. 2012. Speciﬁcation, equation, and agreement in copular sentences.

Canadian Journal of Linguistics/Revue canadienne de linguistique 57:209–240. URL

https://doi.org/10.1017/s0008413100004758.

Jambrovi´c, Samuel. to appear. Pronominal structure and the third-person gap in Span-

ish. In Proceedings of WCCFL 41.

Kaur, Gurmeet. to appear. Two φ-feature sets on honoriﬁc nouns. In Proceedings of

WCCFL 41.

Kayne, Richard. 1975. French syntax. Cambridge, MA: MIT Press.

Kayne, Richard. 1989. Facets of Romance past participle agreement. In Dialect vari-

ation and the theory of grammar, ed. Paola Benincá, 85–104. Dordrecht: Foris. URL

http://dx.doi.org/10.1515/9783110869255-005.

Keine, Stefan, Michael Wagner, and Jessica Coon. 2019. Hierarchy effects in copula con-

structions. Canadian Journal of Linguistics/Revue canadienne de linguistique 64:617–

648. URL https://doi.org/10.1017/cnj.2019.28.

Kratzer, Angelika. 2009. Making a pronoun: Fake indexicals as windows

into the properties of pronouns. Linguistic Inquiry 40:187–237. URL

https://doi.org/10.1162/ling.2009.40.2.187.

Laughren, Mary. 2001. What Warlpiri ‘avoidance’ registers do with grammar. In Forty

years on: Ken Hale and Australian languages, ed. Jane Simpson, David Nash, Mary

Laughren, Peter Austin, and Barry Alpher, 199–225. Canberra: Paciﬁc Linguistics.

Lyding, Verena, Egon Stemle, Claudia Borghetti, Marco Brunello, Sara Castagnoli,

Felice Dell’Orletta, Henrik Dittmann, Alessandro Lenci, and Vito Pirrelli. 2014.

The PAISÀ corpus of Italian web texts. In Proceedings of the 9th Web as

Corpus Workshop (WaC-9). Association for Computational Linguistics. URL

http://dx.doi.org/10.3115/v1/w14-0406.

Maiden, Martin, and Cecilia Robustelli. 2013. A reference grammar of

modern Italian. London/New York: Routledge, 2nd edition. URL

https://doi.org/10.4324/9780203783504.

Messick, Troy. 2023a. 3/4 of a monster: On mixed shifty agreement in Telugu. Linguistic

Inquiry 1–15. URL https://doi.org/10.1162/ling_a_00512.

Messick, Troy. 2023b. Hybrid agreement with English quantiﬁer partitives. Glossa: a

journal of general linguistics 8. URL https://doi.org/10.16995/glossa.10273.

Müller, Gereon. 1999. Optimality, markedness, and word order in German. Linguistics

37:777–818. URL https://doi.org/10.1515/ling.37.5.777.

Nevins, Andrew. 2007. The representation of third person and its consequences

for person-case effects. Natural Language & Linguistic Theory 25:273–313. URL

http://dx.doi.org/10.1007/s11049-006-9017-2.

Ormazabal, Javier, and Juan Romero. 2007. The Object Agreement Con-

straint. Natural Language & Linguistic Theory 25:315–347. URL

https://doi.org/10.1007/s11049-006-9010-9.

Pancheva, Roumyana, and Maria Luisa Zubizarreta. 2018. The Person Case Con-

straint: The syntactic encoding of perspective. Natural Language & Linguistic Theory

36:1291–1337. URL https://doi.org/10.1007/s11049-017-9395-7.

Paparounas, Lefteris, and Martin Salzmann. 2023. First conjunct clitic dou-

bling, the person case constraint, and ﬁrst conjunct agreement: Insights

from Modern Greek. Glossa: a journal of general linguistics 8. URL

https://doi.org/10.16995/glossa.8417.

Percus, Orin. 2011. Gender features and interpretation: a case study. Morphology

21:167–196. URL https://doi.org/10.1007/s11525-010-9157-2.

Perlmutter, David. 1971. Deep and surface structure constraints in syntax. Holt, Rinehart

and Winston.

Perlmutter, David M. 1970. Surface structure constraints in syntax. Linguistic Inquiry

1:187–255.

Pescarini, Diego. 2021. Romance object clitics: Microvariation

and linguistic change. Oxford: Oxford University Press. URL

https://doi.org/10.1093/oso/9780198864387.001.0001.

Postal, Paul M. 1989. Masked inversion in French. Chicago: University of Chicago Press.

Preminger, Omer. 2009. Failure to agree is not a failure: ϕ-agreement with post-

verbal subjects in Hebrew. Linguistic Variation Yearbook 9:241–278. URL

https://doi.org/10.1075/livy.9.07pre.

Preminger, Omer. 2014. Agreement and its failures. Cambridge, MA: MIT Press. URL

https://doi.org/10.7551/mitpress/9780262027403.001.0001.

Preminger, Omer. 2019. What the PCC tells us about “abstract” agreement, head

movement, and locality. Glossa: a journal of general linguistics 4. URL

https://doi.org/10.5334/gjgl.315.

Puškar-Gallien, Zorica. 2019. Resolving polite conﬂicts in predicate agreement. Glossa:

a journal of general linguistics 4:1–25. URL https://doi.org/10.5334/gjgl.587.

Rezac, Milan. 2011. Phi-features and the modular architecture of language. Dordrecht:

Springer. URL https://doi.org/10.1007/978-90-481-9698-2.

Roca, Francesc. 1992. On the licensing of pronominal clitics: The properties of object

clitics in Spanish and Catalan. Master’s thesis, Universitat Autònoma de Barcelona,

Barcelona.

Sauerland, Uli. 2008. The semantic markedness of φfeatures. In Phi the-

ory: Phi-features across modules and interfaces, ed. David Adger, Daniel

Harbour, and Susana Béjar, 57–82. Oxford: Oxford University Press. URL

https://doi.org/10.1093/oso/9780199213764.003.0003.

Schütze, Carson. 2003. Syncretism and double agreement with Icelandic nominative

objects. In Grammar in focus: Festschrift for Christer Platzack, ed. Lars-Olof Delsing,

Cecilia Falk, Gunlög Josefsson, and Halldór Ármann Sigurðsson, 295–303. Lund: De-

partment of Scandinavian Languages, Lund University.

Servidio, Emilio. 2014. Imposters and secondary sources in Ital-

ian. In Cross-linguistic studies of imposters and pronominal agree-

ment, ed. Chris Collins. Oxford: Oxford University Press. URL

https://doi.org/10.1093/acprof:oso/9780199336852.003.0006.

Sheehan, Michelle. 2020. The Romance Person Case Constraint is not about clitic clus-

ters. In Dative structures in Romance and beyond, ed. Anna Pineda and Jaume Mate,

143–171. Berlin: Language Science Press.

Sigurðsson, Halldór Ármann, and Anders Holmberg. 2008. Icelandic dative interven-

tion: Person and number are separate probes. In Agreement restrictions, ed. Roberta

D’Alessandro, Susann Fischer, and Gunnar Hrafn Hrafnbjargarson, 251–280. Berlin:

Mouton de Gruyter. URL https://doi.org/10.1515/9783110207835.251.

Smith, Peter William. 2015. Feature mismatches: Consequences for syntax, morphology,

and semantics. Doctoral Dissertation, University of Connecticut, Storrs.

Smith, Peter William. 2017. The syntax of semantic agreement in English. Journal of

Linguistics 53:823–863. URL https://doi.org/10.1017/s0022226716000360.

Smith, Peter William. 2021. Morphology–semantics mismatches and the

nature of grammatical features. Berlin: Mouton de Gruyter. URL

https://doi.org/10.1515/9781501511127.

Stegovec, Adrian. 2015. It’s not case, it’s personal! Reassessing the PCC and clitic restric-

tions in O’odham and Warlpiri. In Proceedings of the Poster Session of WCCFL 33, SFU

Working Papers in Linguistics, 131–140. Vancouver, BC: Simon Fraser University, SFU

Lin- guistics Graduate Student Association.

Stegovec, Adrian. 2020. Taking case out of the person-case con-

straint. Natural Language & Linguistic Theory 38:261–311. URL

http://dx.doi.org/10.1007/s11049-019-09443-0.

Stiebels, Barbara. 2006. Agent focus in Mayan languages. Natural Language & Linguistic

Theory 24:501–570. URL https://doi.org/10.1007/s11049-005-0539-9.

Trutkowski, Ewa, and Helmut Weiß. 2016. When personal pronouns compete with

relative pronouns. In The impact of pronominal form on interpretation, ed.

Patrick Grosz and Pritty Patel-Grosz, 135–166. Berlin: De Gruyter Mouton. URL

https://doi.org/10.1515/9781614517016-006.

Wang, Ruoan. 2023. Honoriﬁcs without [HON]. Natural Language & Linguistic Theory

41:1287–1347. URL https://doi.org/10.1007/s11049-022-09563-0.

Wechsler, Stephen. 2008. “Elsewhere” in gender resolution. In The na-

ture of the word: Essays in honor of Paul Kiparsky, ed. Kristin Han-

son and Sharon Inkelas, 567–586. Cambridge, MA: MIT Press. URL

https://doi.org/10.7551/mitpress/9780262083799.003.0024.

Wechsler, Stephen, and Hyun-Jong Hahm. 2011. Polite plurals and adjective agreement.

Morphology 21:247–281. URL https://doi.org/10.1007/s11525-010-9168-z.

Wechsler, Stephen, and Larisa Zlati´c. 2003. The many faces of agreement. Stanford, CA:

Center for the Study of Language and Information.

Wurmbrand, Susi. 2016. Formal and semantic agreement in syntax: A dual feature ap-

proach. In Language use and linguistic structure: Proceedings of the Olomouc Linguis-

tics Colloquium 2016, ed. Joseph Emonds and Markéta Janebová, 19–36. Olomouc:

Palacký University.

Wurmbrand, Susi. 2017. Feature sharing or how I value my son. In MIT Working Papers

in Linguistics, ed. Claire Halpert, Hadas Kotek, and Coppe van Urk, 173–182.

Zubizarreta, Maria Luisa, and Roumyana Pancheva. 2017. A formal

characterization of person-based alignment: The case of Paraguayan

Guaraní. Natural Language & Linguistic Theory 35:1161–1204. URL

https://doi.org/10.1007/s11049-016-9357-5.

ResearchGate has not been able to resolve any citations for this publication.

Pronominal structure and the third-person gap in Spanish

Conference Paper

Full-text available

Sep 2024

Samuel Jambrović

In English and many other languages, third-person pronouns cannot appear in pronoun-noun constructions: "we linguists" versus "*they linguists". Although Spanish uses the definite article in these constructions, it exhibits the same third-person gap: "nosotros los lingüistas" 'we the linguists' versus "*ellos los lingüistas" 'they the linguists'. Moreover, the ungrammaticality of "*ellos lingüistas" 'they linguists' suggests that the issue is not simply due to the "-llos" component of "ellos" 'they' competing for the same structural position as "los" 'the' as the definite article. In this paper, I argue that third-person pronouns realize the entirety of the noun phrase and that they can only do so when the following heads are structurally adjacent: π, D, Div, and n. If third-person pronouns themselves spell out Div and n, it follows that they cannot appear with an overt noun, ruling out both "*ellos los lingüistas" and "*ellos lingüistas". This analysis also captures the inability of third-person pronouns to be modified by restrictive relative clauses: "*ellos los que son altos" 'they the ones who are tall'. Since restrictive relative clauses enter the derivation below DP, they disrupt the adjacency of D and Div and thereby prevent the noun phrase from being realized as a third-person pronoun.

Split coordination with adjectives in Italian

Article

Full-text available

Sep 2024
NAT LANG LINGUIST TH

Luke Adamson

This work investigates the morphosyntax of nominal expressions in Standard Italian that have multiple adjectives in “split coordination,” which permits a plural noun to be modified by singular adjectives, for example le mani destra e sinistra (the hand.pl left.sg and right.sg). The proposal is (i) that these expressions are built from multidominant structures, with a constituent shared by the conjuncts, and (ii) that plural marking on the noun reflects “summative” feature resolution on the nP comparable to coordination resolution. This proposal captures various properties of split-coordinated expressions, including the availability of adjective stacking and of feature-mismatched conjuncts, as well as agreement with a class of nouns that “switch” gender in the plural. Taking agreement with resolving features to be a form of semantic agreement, which has been argued to be possible only in certain syntactic configurations (Smith 2015, 2017, 2021), the account captures prenominal-postnominal adjective asymmetries in split coordination. The work offers a coherent account of coordination and semantic agreement in the nominal domain, connects split coordination to related phenomena such as nominal right node raising and adjectival hydras, and, more broadly, evinces the unity of nominal and verbal agreement, pace analyses of nominal concord (Norris 2014).

Hybrid agreement with English quantifier partitives

Article

Full-text available

Nov 2023

Troy Messick

This paper presents a novel case study of a 3/4 agreement pattern typically found with hybrid nouns. This case study involves agreement and binding with Quantified NPs in English. I propose an analysis that relies on different classes of agreement targets agreeing at different times and couple this with a condition on the access to semantic agreement features. This new analysis can account for the novel data presented here as well as the data from the literature. This paper hence broadens both our empirical knowledge of 3/4 patterns as well as refines our theory of features and agreement that underlie such patterns.

3/4 of a Monster: On Mixed Shifty Agreement in Telugu

Article

Full-text available

Jun 2023

Troy Messick

Within the typology of embedded pronouns, there are languages that allow for non-first person pronouns to apparently control first person agreement morphology when in certain embedded contexts. This type of agreement displays some degree of optionality: it is also possible for the pronoun to control the expected agreement morphology given the pronoun’s own overt morphological features. This paper provides new data from the Dravidian language Telugu that shows when the embedded pronoun controls agreement on two separate targets, agreement may be uniform across the two targets or the two targets can mismatch in one direction, but crucially not the other. I show how we may account for this paradigm using the assumptions that the pronouns in question are similar to so-called hybrid nouns and that agreement features are restricted in principled ways.

First Conjunct Clitic Doubling, the Person Case Constraint, and First Conjunct Agreement: Insights from Modern Greek

Article

Full-text available

Feb 2023

In this paper, we discuss the implications of the phenomenon of first conjunct clitic doubling (FC CLD), based on data from Modern Greek. We first recap the arguments in Paparounas & Salzmann (to appear), where it is argued that FC CLD provides evidence for Agree-based approaches to clitic doubling (CLD) and against approaches involving movement, which would incorrectly rule out FC CLD as a violation of the Coordinate Structure Constraint. Extending this line of inquiry, we explore the implications of FC CLD for theories of the Person Case Constraint (PCC) and theories of first conjunct agreement (FCA). We firstly show that most theories of the PCC face serious challenges when confronted with our data. These theories i) involve movement in the generation of the clitic/the implementation of the PCC, thus being wholly incompatible with FC CLD; ii) assume Agree operations that provide insufficient φ-features to actually generate the clitics on the PCC probe; or iii) place this probe too low in the structure for languages like Greek, where clitics surface high. Secondly, we extend our consideration of the PCC to the domain of FCA, showing that PCC effects in coordination provide evidence for rule ordering and labeling-based theories of FCA when relativized probing is involved.

Person effects in agreement with Icelandic low nominatives: An experimental investigation

Article

Full-text available

Dec 2022
NAT LANG LINGUIST TH

This paper investigates agreement—in particular person agreement—in two configurations in Icelandic where there are two potential controllers of agreement and where at least in some cases agreement is with the lower of the two (a “low nominative”). One case is the Dative-Nominative construction, where there is a dative subject and a lower nominative argument. The other is the Specificational Copular Clause (SCC) construction, where there are two nominative arguments. A much-discussed aspect of agreement in the former case is that agreement in number with the low nominative is generally possible, but agreement in person is at best highly restricted, leading in some cases to ineffability. This person effect has been claimed to be ameliorated by syncretism in the agreement paradigm, but there is limited data available substantiating this effect, which is however crucial to deciding between two recent types of account. This paper reports on a pair of experimental rating studies on the Dat-Nom and SCC configurations in Icelandic. We show that, taken together, the two sets of data provide evidence against the Person Licensing Condition and in favour of an account of the Dat-Nom construction in terms of morphological conflict arising from double agreement, although we show that the ameliorating effect of morphological syncretism, while real, is limited. Further, we show that there is no evidence of double agreement in the copular clauses investigated. We argue that full agreement with the low nominative here arises if the first nominal can move out of the domain of agreement entirely. The possibility of agreement with the initial nominal we suggest indicates that nominatives, unlike datives, cause the search of the agreement probe to terminate.

Gender Features and Coordination Resolution in Greek and Other Three-Gendered Languages: Implications for the Cross-Linguistic Representation of Gender<