Double pronominalization and clitic doubling in Dialectal Brazilian Portuguese and Colloquial Standard Brazilian Portuguese

In this study, we propose a comparative analysis of the ditransitive constructions in Dialectal Brazilian Portuguese (DBP) and Colloquial Standard Brazilian Portuguese (BP), taking into consideration the phenomenon named Double Pronominalization


Introduction
In this study, we propose an analysis of ditransitive constructions in Dialectal Brazilian Portuguese (DBP) and Colloquial Standard Brazilian Portuguese (henceforth, BP) taking into consideration double pronominalization, in which both the theme and the goal arguments are pronominalized, with consequences for the occurrence of a type of Person Case Constraint.We further discuss a type of clitic doubling in Dialectal Brazilian Portuguese (DBP), which involves first and second person clitics.
Primarily we focus on DBP ditransitives, in which the grammatical licensing of the goal argument is found in a type of Double Object Construction (DOC), as illustrated in (1a-b). 1   (1) (from Rezende &Pádua 2004: 49;196) a. ua muié mandô pidi el'-Ø um remédio a woman made.3SGask.INF 3M-SG.ACC a medicine 'a woman made someone ask her a medicine' b. teve que dá água ele-Ø primeiro had.3SG that give.INF water 3M-SG.ACC first 'he had to give him water' In addition to DOC, DBP displays the prepositional ditransitive construction (PDC), in which the goal argument is introduced by the preposition para ('to'), which 1 Real data of DBP are collected in the following sources: Corpus 'Projeto Minerês', from Ramos (2010); Corpora dos Falares de Goiás, from Rezende;Pádua (2004); Nascimento (2007Nascimento ( , 2009)); Rezende (2008) (Anexo); Pereira (2019). is also found in BP, thus implying an alignment with benefactives and other oblique phrases, as illustrated in (2a-b). 2   (2) (from Rezende &Pádua 2004: 184;140) a. vô mostrá pa senhora um negocim go.PRS.1SGshow.INF to madam one small-thing 'I will show something to madam' b. eli-s dava cachaça... dava rapadura p'ra ele-s 3M-PL give.IPRF.3PLspirit... give.IPRF.3PLsweet to 3M-PL.OBL.'they used to give spirit... give sugar-cane-sweet sweet to them' (data In previous analyses, these facts have been attributed to the radical changes in the pronominal system of BP (both dialectal and colloquial standard), which involve the loss of third person dative and accusative clitics, namely lhe(s) and o(s), a(s), respectively, and the rise of the (full) forms você(s), for second person, and ele(s), ela(s), for third person, in different grammatical functions.
Assuming that DOC (as found in DBP) is a correlate of English DOC, we follow previous analyses proposing that it is licensed in an Applicative structure (Pylkännen 2002, 2008, for English;Torres Morais & Salles 2010, for DBP).We also consider the fact that (only) first and second person clitics me and te are found in both DBP and BP pronominal system.We further assume, following Torres Morais & Salles (2022), that the facts about the split system of clitics in BP point to a change from a bivalent to a monovalent value with respect to the feature [participant], which allows for the unification of the syntactic and semantic properties of the dative clitic as a high applicative head.
A related fact is that in double pronominalization, first and second person clitics (me/ te) are used either as accusative or dative arguments, along with the full pronoun (ele(s); ela(s)), or the prepositional ditransitive construction (PDC), as illustrated in (3) and (4), respectively.Interestingly, the example in (3) corresponds to the one involving a clitic cluster in European Portuguese, as illustrated in (3').In turn, the one in (4) has a direct correlate in EP, as it involves an oblique phrase (although the relevant preposition in EP is a, which occurs as a repair structure, given that clustering is blocked by the Person Case Constraint (PCC)).In BP the oblique configuration is uniformly found, as clitic clusters cannot be formed.
(5) (from Machado Rocha 2016: 23) deixa eu te=perguntá ocê um negócio let.IMP.3SGI 2SG=ask.INF you.ACC a thing 'let me ask you something' Assuming the split character of the pronominal system, we will propose that first and second person clitics in both BP and DBP are licensed in an Applicative projection by an interpretable Person feature.This configuration is taken to impose the restrictions on clitic clustering (thus implying that a type of PCC is at stake), further providing the grammatical conditions for the rise of clitic doubling in DBP.In the absence of third person (dative) clitic pronouns (as well as third person (accusative) pronouns), the dative argument is uniformly licensed as an oblique phrase.
The study is developed as follows, in section 2, we firstly present the facts about the innovative pronominal realization of the goal argument in ditransitive structures in Dialectal Brazilian Portuguese (DBP) and Colloquial Standard Brazilian Portuguese (BP), considering the facts about double pronominalization and clitic doubling, which display innovative strategies.In section 3, we provide a brief review of previous analyses of ditransitive constructions in EP, in terms of Pylkännen's (2008) applicative typology, in which the interpretable person feature on the applicative head is at stake.Given this approach, we provide an account of double pronominalization as found in DBP and BP, in which a type of PCC arises.We further discuss clitic doubling in DBP.In section 4, we present the final considerations.

Pronominalization in ditransitive constructions in DBP and BP
In this section, we look at the innovative facts concerning pronominalization of ditransitive constructions in DBP and BP in detail.In section 2.1, we present a brief note on the dialect spoken in the Central region, in which DOC is found.In the following section, we present the properties of DOC in DBP (in a contrastive perspective to English DOC), further considering the prepositional ditransitive construction (PDC), found in both DBP and BP.Given this, we discuss the facts about double pronominalization in DBP and Colloquial and Standard BP, in which clitic clusters are absent.Instead, the PDC configuration is uniformly found, thus differing from EP, in which this configuration only occurs as a repair strategy, that is, as an alternative to the restrictions on clitic clusters imposed by the PCC.

DBP in the Brazilian territory
In the present study, Dialectal Brazilian Portuguese is a designation for the language spoken in the rural area of the Central region of Brazil, which is delimited by Goiás and Minas Gerais states, as well as part of Bahia state, given their identification regarding the occurrence of the so-called Double Object Construction, and related properties of the pronominal system.Accordingly, this designation is not intended to refer to a dialect as defined in the scope of the dialectology discipline, which takes into consideration lexical-phonological variables that are not relevant for the present study.In this respect, we refer the reader to ALINGO (Atlas Linguístico de Goiás), in which relevant information of dialects from Goiás is provided (cf.Milani 2016).
Following previous studies, the occurrence of the above-mentioned Double Object Construction, and its related pronominal system allows for a distinction from other dialects of BP, including the Colloquial Standard variety, henceforth designated as BP.DOC in DBP has been widely investigated in the literature, from different theoretical perspectives (see Ramos 1992;Scher 1996;Salles 1997Salles , 2016;;Nascimento 2007Nascimento , 2009;;Torres Morais & Salles 2010;Ramos & Salles 2016;Rocha 2017;Pereira 2019;Salles & Torres Morais 2020; among others, regarding the occurrence of DOC in Minas and Goiás dialects). 3 We shall not go into the related social and historical facts that took place during the colonial period, particularly in the 17th and 18th centuries, giving rise to language contact between the language spoken by the colonizers, namely Portuguese, and the languages of the originary people (mostly from the Macro-Jê stock), who were confined under the so-called 'aldeamento' (tribalizing) policy, as well as of the African people (mostly from the Bantu family), who were brought to the region as slaves (on the facts about the occupation of the central region, see Chaim 1974;Chaul 1997;Brasil 1980, among others).
At this point we will focus on the analysis of the properties characterizing DOC in DBP (in a comparative perspective with English).

A type of Double Object Construction (DOC) in DBP in variation with the Prepositional Ditransitive Construction (PDC)
As mentioned above, Dialectal Brazilian Portuguese (DBP) has a type of DOC which alternates with the prepositional ditransitive construction (PDC), in which the goal 3 It should be noted that a similar type of DOC is also attested in dialects spoken in Bahia state (located in the Northeastern region), as pointed out in Lucchesi & Mello, 2006;Barros 2018, Barros & Calindro 2023, among others (cf. (i)).According to these analyses, the occurrence of DOC is due to language contact between Portuguese, spoken by the colonizers, and languages from the Bantu stock, spoken by the Africans brough to Brazil as slaves, during the colonial period (on the socio-historical facts, see, among many others, Mattos e Silva 2004).argument is introduced by the preposition para ('to'), and its reduced forms pra/ pa, as illustrated in (1) and (2), above, as well as in ( 6) and ( 7), below, respectively. 4As pointed out in Nascimento (2007Nascimento ( , 2009)), the preposition a ('to') is attested in the rural varieties spoken in Goiás.Adopting the Labovian framework, the author provides a detailed analysis of the distribution of the variants expressing the dative argument, showing that the preposition para ('to') is the preferred variant, as opposed to a, and the null variant (which corresponds to the type of double accusative construction).In particular, the preposition a is restricted to contexts in which the goal argument is a divinity, as illustrated in (i)for this reason we take the use of a in these contexts to be formulaic, implying that para ('to') is the innovative lexical variant.
(i) A gente pede muito a Deus (...) the-people ask a-lot to God 'we strongly ask God [for mercy]' The use of the preposition para ('to') allows for an alignment with benefactive constructions, as illustrated in (8a-c).In this respect, DBP differs from English, as the preposition introducing the benefactive construction is for, which is not the same as the one introducing the goal argument in ditransitive constructions (cf. he cooled the milk for the child/ he gave the milk to the child)we shall return to this matter.The structural properties of DOC in DBP have been examined in various studies, in terms of Barss and Lasnik's (1986) analysis of English DOC, pointing to a similar behavior (see Scher 1996;Salles 1997Salles , 2016;;Torres Morais & Salles 2010;Ramos & Salles 2017;Rocha 2017, among others).Given that asymmetric c-command is required for a pronoun to relate to a quantified NP (QNP) as variable, it is possible to show that the first object asymmetrically c-commands the second object in DOC from DBP, as the binding relation holds in (9a), but not in (9b), exactly as in the DOC examples from English (cf.(10a), as opposed to (10b)). 5As expected, the same structural conditions apply to the prepositional ditransitive construction (PDC), as the binding relation holds in (11a), but not in (11b), implying that the quantified NP in the first position asymmetrically c-commands the pronoun in the complement position of the preposition, but not conversely.
Recall that ditransitive predicates also occur in the theme-goal word order in DBP, the goal argument occurring without the preposition, as illustrated in (1b), repeated in (12a), below, including the examples in (12b) and (12c), thus differing from English, in which only the goal-theme word order is allowed (*Mary gave a book John).The examples in ( 9), ( 11) and ( 13) concerning quantifier binding are adapted from Ramos & Salles (2017), the judgements being provided by a native speaker of the DBP.b.Maria Lixande, vai dá mão el-a-Ø Maria Lixande, go.IMP give.INF hánd 3F-SG 'Maria Lixande, go to give her a hand' c. ê-s ficava dano consei nóis 3M-PL remain.IPRF giving advise 1PL 'they would remain giving us advise' Regarding the binding relations, it is noted that the structure in the theme-goal word order, without the preposition, patterns like the one with the preposition, as illustrated in (13a), as opposed to (13b), with examples adapted from Ramos & Salles (2017): while the binding relation between the QNP and the pronoun holds in (13a), it does not in (13b), implying that the QNP asymmetrically c-commands the pronoun in the former, but not in the latter.

Pedro showed hisi son each fatheri
A relevant property of DOC in DBP is that, differently from English DOC, passivization affecting the goal DP is not possible with verbs of transference, such as dar ('to give'), entregar ('to hand') (cf.14a).Crucially, passivization is possible with verbs of saying, as illustrated in (14b), suggesting that it is lexically restricted.
João was given a book b.João foi ensinado o caminho.

João was taught the address
We thus assume that passivization of the goal argument is not a conclusive test for determining its accusative status.We thus retain the view that DBP has a type of DOC. 6  Another fact concerning DOC in DBP is that it is not found with benefactive arguments (cf.(15a)), differently from English, which allows it in the transfer of possession interpretation (cf.( 15b)). 76The restriction on passivization may be seen as a challenge for the analysis of the goal argument as bearing structural Case in the DOC type from DBP (as pointed out by two anonymous reviewers).Given the contrastive facts in (14a) and (14b), we leave this matter open, referring the reader to Ramos & Salles (2017), for a more detailed discussion on the role of verb class in passivization.

7
The occurrence/ absence of benefactives in the DOC configuration can be analysed in terms of a requirement on a distinctive licensing of the non-argumental benefactive, as opposed to the goal argument of a ditransitive predicate, namely that at least one of them be assigned a non-ambiguous configuration.In English, the configuration with the preposition for is exclusive of the non-argumental benefactive (bake a cake for X), as opposed to DOC, in As already mentioned, the innovative licensing of the goal argument in DBP and BP interacts with the (innovative) pronominal system.These properties are summarized in the topics and tables 1 and 2, below (cf.Galves 2001, and reference therein).✓ a pronominal split regarding the distribution of clitics: (only) first and second person clitic pronouns, me and te, are foundas a corollary, the third person full pronouns ele(s)/ ela(s) substitutes for the third person clitics lhe(s)/ o(s); a(s) in the direct object position and in the complement of the preposition; ✓ the honorific pronoun você(s) (as found in EP) occurs as a second person pronoun both in BP and DBP (also found in subject position, triggering third person singular inflection), alternating with the second person clitic te; ✓ the collective DP a gente ('the people') includes the speaker, thus occurring as a first-person plural expression (triggering third person singular inflection, in subject position), alternating with the first person plural pronoun nós; ✓ the strong/ full pronominal series is found in the complement position of the directional preposition para ('to') encoding the goal argument in ditransitive and monotransitive (para mim/ você(s)/ nós/ a gente/ ele(s)/ ela(s)).
which both the non-argumental benefactive and the goal argument are found (to bake X a cake/ to give X a cake).In DBP, DOC is exclusive of the goal argument (dar X um bolo), while both the goal argument and the non-argumental benefactive are found in the prepositional configuration with para ('to') (dar um para X/ fazer um bolo para X) (cf.Salles 1997).
DBP and BP's pronominal systems are presented in Tables 1 and 2, respectively.The crucial difference between DBP and BP is that the former, but not the latter, uses full pronouns without the preposition in the realization of the goal argument (giving rise to DOC).8 Given these facts, we now turn to double pronominalization, found with ditransitive predicates in both DBP and BP, further considering a type of clitic doubling, found in DBP.

Double pronominalization in DBP and BP
As mentioned in the previous section, third person clitics (whether dative (lhe(s)) or accusative (o(s); a(s)) are not found in BP and DBP.A direct consequence is that clitic clusters of the EP type (cf.section 3.1, below) cannot be formed, as contrastively illustrated in ( 17 As already mentioned, first and second person clitics (me; te) are productively found in both DPB and BP.Accordingly, double pronominalization is found in two configurations given that the goal argument is realized either as the (dative) clitic pronoun or as full pronoun introduced by the lexical preposition, as illustrated in (18a) and (18b), respectively.If the clitic is the direct object, the goal argument is obligatorily realized as a full pronoun introduced by the lexical preposition, as illustrated in (18c), as opposed to (18d).
(18) a.  Bonet (1994Bonet ( , 1995)), exactly as in EPwe will return to the EP facts in section 3.1.Hence, in double pronominalization the pronominal clitic is either the direct object (DO) or the indirect object (IO).If it is the DO, the goal argument is (obligatorily) realized as a full pronoun introduced by the lexical preposition para ('to') (cf. (19a-b), as opposed to (19c)).If it is the IO, the direct object must be realized by a full pronoun, which is crucially the second person full form você (first-and second-person full pronouns eu and tu being marked forms) (cf. 19d-f) If the two arguments in ditransitive structures are third person (corresponding to lho(s)/lha(s)), the goal argument only occurs as an oblique pronoun in a configuration introduced by the lexical preposition.An independent restriction does not allow the realization of the (third person) goal argument as a full pronoun, as illustrated by the contrast in (20a) and (20b) (cf.Nascimento, 2009: 59).
(20) a. João apresentou ela-Ø pra ele-Ø.João introduced.3SG3F-SG.ACC to 3M-SG.OBL 'João introduced her to him.' b. *João apresentou ele-ØGOAL ela-Ø.João introduced.3SG3M-SG.ACC 3F-SG.ACC At this point, we can formulate the following generalization related to the goal argument in double pronominalization: the goal argument can be realized either as an oblique pronoun (introduced by the lexical preposition para ('to')), or as a first and second person (dative) clitic (me/ te) in both DPB and BP. 9 9 Pancheva & Zubizarreta (2018) present a very detailed typology and a formal analysis of PCC effects, mainly in Romance Languages.Based on Perlmutter (1971:21) and Bonet (1991:179), the authors documented that some speakers of Spanish and some speakers of Catalan accept a variety of PCC named Ultra-Strong PCC (Nevins, 2007), which ranks first person higher than second person in that it allows the second person but not the first person to Before we turn to the formal analysis for the above-mentioned facts, we would like to examine the occurrence of a type of clitic doubling in DBP, in which first and second person clitics are productively found.

4 Clitic Doubling in DBP
As already mentioned, another feature concerning the pronominal realization of the goal argument is that a type of clitic doubling is found in DBPa fact that has been originally noted in data from Minas dialect, also affecting direct objects (cf.be the direct argument in double object construction.For Spanish speakers who have this variety of PCC grammar, two participant clitics can co-occur without ambiguity, even though they do not show a distinction between dative and accusative case.As the example (i) shows, the first-person clitic (me) must be understood as the indirect object.

(i)
Él te=me=recomendó (a mí). he 2SG=1SG=recommended.3SG (to me) 'He recommended you to me.', not: 'He recommended me to you.' (Pancheva & Zubizarreta, 2018:1295) Also, the authors considered the fact that Spanish is a language with Strong PCC grammar, as it is well documented in the current literature.Consequently, for such speakers, this example is not acceptable.
Given the above-mentioned split in the pronominal system of DBP (as well as in BP), clitic doubling is restricted to first and second person, as third person clitics are not found in DBP.As noted in Pereira (2019), clitic doubling in DBP should be taken as additional evidence for the productivity of first and second person clitics in this dialect, along with the facts concerning double pronominalization, as presented in the previous section.
As noted in Machado Rocha's (2016: 88) analysis, a relevant property of clitic doubling in DBP is that in the former the doubled DP occurs without the preposition in both ditransitive and monotransitive predicates (cf.( 21b) and (27a-c), above)although the preposition para may be found with ditransitive predicates as well (21a).Machado Rocha points out that the role of the preposition is crucial in the analysis of the doubling clitic in DBP as agreement markers.We will return to this matter (cf.section 3.4).
In section 3 we will establish the theoretical assumptions for the present analysis.

A formal analysis for ditransitive constructions in Dialectal Brazilian Portuguese and Colloquial Standard BP
In this section, we will advance an analysis of the goal argument (indirect object) in BP and DBP in which it is syntactically expressed as either an applied or an oblique argument.In section 3.1, we will discuss the (high) applicative analysis of indirect objects in European Portuguese (EP) and the PCC effects, in which the role of the interpretable Person feature on the applicative head is at stake.In section 3.2, we discuss the corresponding facts in BP and DBP, considering double pronominalization.In section 3.3, we use this framework, taking into consideration clitic doubling in DBP.

The Applicative analysis of indirect objects in European Portuguese and the PCC effects
A well-known fact regarding ditransitive predicates European Portuguese (EP) is that the indirect object (IO) is a distinct structural class, morphologically identified by a dative morphological marker.As a full DP, it is preceded by the dative-morpheme a (a-DP), and as a pronominal argument, it is realized by the series of dative clitics.
As noted in Torres Morais (2006Morais ( , 2007)), the mutual implication between the dative marker a and the dative clitic is expressed under clitic doubling, in which not only the dative clitic is obligatory in the presence of the full (pronominal) DP (cf.(23a)), but also the dative marker a is obligatory, the lexical preposition being disallowed (cf.(23b)):  3, in a comparative perspective with BP and DBP, in which the DP complex is unavailable, as displayed in Table 1 and Table 2).Another remarkable aspect is the pervasive distribution of dative objects in EP, which is referred in traditional grammars as well as in modern descriptive ones (cf.Mateus et al. 2003;Raposo et al. 2013).It is found with activity verbs, including directional dynamic verbs of transference, creation verbs and non-directional dynamic verbs.It also appears with stative verbs.In all these contexts, the goal argument conveys different meanings, including recipient, source, benefactive, possessor, with respect to another argument, which is mostly the theme argument.
Based on Pylkkänen's (2008) theory of low and high applicatives, Torres Morais (2007) proposes that the morphological expression of the dative argument and the variety of meanings in EP ditransitive structures derive from the different positions in which the applicative head is licensed in the clause structure.In Pylkkänen's typology, the high applicative head denotes a relation between an event and one individual, while the low applicative head denotes a relation between two individuals.The DOC/ applicative analysis is also proposed in Cuervo (2003Cuervo ( , 2020)), for Spanish; Diaconescu & Rivero 2007, for Romanian;Fournier (2010), for French; Pineda (2013Pineda ( , 2020)), for Spanish and Catalan, just to mention a few studies in the same line of research on Romance languages.Accordingly, the applied argument is syntactically and semantically licensed in the projection of a low applicative head.Also, it is possible to assume three subtypes of low applicatives in EP: recipient (TOapplicative), source (FROM-Appl) and possessor (AT-Appl), as originally proposed in Cuervo, for Spanish (2003).
We further assume, following Torres Morais and Salles (2010, 2016, 2019), that the upper v head has uninterpretable phi-features that act as a Probe and enters an Agree relation with the (interpretable) phi-features of the DPTHEME (cf.Chomsky 1995Chomsky , 2001)).In turn, the applicative head bears an interpretable Person feature which enters an agreement relation with the inherently marked dative argument in the Applicative domain.The derivation of (24a) involving a low applicative head (Appl TO) is illustrated in (24b).
(24) a. O Pedro deu=lhe o colar (a el-a-Ø).the P. gave=3SG.DAT the necklace to 3F-SG.OBL 'Peter gave her the necklace.'b. [ Moreover, EP provides evidence for the two types of applicative heads, namely the low applicative and the high applicative.As contrastively shown in (25a-b), an ambiguous interpretation arises in (25a), as either a possessive or a benefactive interpretation is possible, but not in (25b), in which only the benefactive interpretation arises, as the predicate denotes that the argument benefits from the fact that the event was performed (not from being the goal of the transference).
(poss/ben) peel.PRS.1SG=3SG.DAT the orange 'I peeled his orange'/ 'I peeled the orange for him.' b. (adapted from Berlinck, 1996: 136) Descasquei=lhe uma laranja.(*poss/ben) peeled.1SG=3SG.DAT an orange 'I peeled him an orange.' In (25a), the possessive interpretation arises from the interaction between the definite interpretation of the DP a laranja (the orange) (as required in external possession) and the occurrence of the goal argument as a dative clitic, confirming their licensing as a high applicative.In turn, the realization of the theme argument as an indefinite DP, as in (25b), excludes the possessive interpretation, implying that only the high applicative is active.
Another related property is that clitic clusters are formed in EP under the Person Case Constraint (PCC), which is originally formulated in Bonet (1994Bonet ( , 1995) ) in the statement in (26) (see also Perlmutter 1971, Kayne 1975, on the PCC in French).
(26) If DAT then ACC-3 rd (Bonet 1994) In fact, EP manifests the Strong PCC type, only allowing the combinations involving a dative clitic and a third person accusative clitic, while excluding the realization of the goal argument in the prepositional phrase, as illustrated in (27a/a'c/c'): As can be inferred from ( 27), the PCC rules out the possibility of clitic clusters formed with a dative clitic (whether first, second or third person) and a first-and second-person accusative clitic -*me te; *te me; *lhe me; *lhe te.Given this, as a repair strategy, double pronominalization is expressed in a mixed structure, which is formed with the accusative clitic and the full pronoun introduced by the lexical preposition a ('to') (cf.28a-d). 10An analysis assuming the distinction between the true dative and the structure in which the full pronoun is introduced by the lexical preposition a ('to') is proposed in Sheehan's (2020) discussion of the contrastive facts in (ia-b).According to Sheehan (2020), clitic doubling (which is taken to involve the true dative marker) excludes differential object marking (DOM), this fact showing that the PCC is not restricted to clitic cluster formation.(i) a. Enviaron *(a) Mateo/ tu hijo a los doctores.sent.3PL*(a) Mateo/ your son to the doctors b.Les=enviaron (*a) Mateo/ tu hijo a los doctores.3PL.DAT=sent.3PL(*to) Mateo/ your son a the doctors 'They have sent Mateo/ your son to the doctors.' (from Ormazabal & Romero, 2013: 224, cited in Sheehan 2020: 150) We take the occurrence of the oblique pronoun in this particular context as a piece of evidence for the grammatical status of a as a lexical preposition, not as a dative marker (cf. (28a-d)).This is confirmed by the ungrammaticality of the construction with the full pronoun as a counterpart to clitic clusters.That is, where clusters are formed, the occurrence of the goal argument as a full pronoun introduced by the lexical preposition is excluded, as illustrated in (27a'-d'), above.
According to Pancheva & Zubizarreta's (2018) analysis of PCC, the interpretable Person-feature is also responsible for the person restrictions in clitic cluster formation.The authors assume a person feature specification that expresses the asymmetry between first, second, on a side, and third person, on the other side, in terms of the binary specification of the feature [participant].First and second person are distinguished by the [author] feature.They also adopt the distinction proximate/ obviative.The notion of proximity, particularly, is related to the speech situation, applying to first, second and third person.The difference is that first and second person arguments are inherently proximate, while the third person is marked [+/-proximate] only in the presence of another third person argument.The system proposed by the authors for the feature make-up of the Person feature on the applicative head is presented in ( 29 In what follows, we will discuss the facts concerning double pronominalization as well clitic doubling constructions in DBP and BP, in terms of the role of the (interpretable) person feature on the applicative head.

Going back to DBP and BP: the structural properties
In this subsection, we would like to assume that DOC in DBP is a high applicative, in which the applicative head excludes the interpretable person feature.We also show that the prepositional ditransitive construction (PDC) is uniformly found in both DBP and BP.Next, we will consider the implications of these analyses for double pronominalization and clitic doubling, as found in these varieties.
The crucial fact about DOC in DBP, as illustrated in (30a), is that, differently from EP, the interpretable person feature is absent in the feature make-up of the applicative head.In the absence of the interpretable person feature, structural/ accusative Case is activated, and the interpretable phi-features of the goal argument in the specifier position of the low applicative head enter an Agree relation with the uninterpretable phi-features of the v head.In turn, the uninterpretable phi-features of the low applicative head enter an Agree relation with the interpretable phi-features of the DP argument in its complement position, giving rise to a type of English DOC, as illustrated in the derivation in (30b) (cf.Torres Morais & Salles 2010).
(30) a. deu el-a-Ø chá gave.3SG3F-SG.ACC tea 'she gave her tea' b. [vP v [VP deu [ApplP [DP elaACC ] [ Appl [DP cháACC]]]]] The same configuration applies to the theme-goal word order, as the possession relation between two individuals is found (cf.(1b)), as proposed in Ramos & Salles (2017).The occurrence of the double accusative is thus a crucial property expressing the microparametric variation in DBP, as opposed to EP.
Regarding the prepositional ditransitive construction, found in both DBP and BP (PDC) (cf.section 2.2), we propose, in the spirit of previous analyses (cf . Kayne 1984;Hale & Keyser 1993;Pesetsky 1995;Salles 1997;Harley 2002;Manzini & Franco 2016, among others), that the goal argument is introduced by a lexical preposition, namely para ('to'), which occurs in the complement position of V, as illustrated in (31a-b), with data from DBP: (31) a.
Ele dava comida pa/po's caboco.he give.IPRF.3SGfood to-the guys 'He gave food to the guys.' b.

]ACC [ V [PP pa [DP os cabocos]]]]]
Recall that PDC is also found in EP, as a repair strategy in double pronominalization, given the effects of PCC, as illustrated in (28a) (cf.section 3.1), repeated in (32), below.
(32) O João apresentou=me a tiGOAL.the João introduced.3SG=1SG.ACC to you.OBL 'John introduced you to me.' We conclude that these facts should be taken as evidence for the distinction between clitic doubling and the so-called true dative, which is licensed in the applicative projection, on the one hand, and in PDC, on the other hand. 11In turn, in 11 See Pineda (2020), for a unified analysis of ditransitive constructions in Spanish (and other Romance languages) in terms of the applicative analysis (thus excluding a contrastive analysis involving the Prepositional Ditransitive Construction (PDC)).Brito (2008) in turn does not assume the applicative analysis, pointing out that ditransitive predicates in EP are projected in a prepositional configuration.In spite of the advantages of the uniform analysis, we will retain the distinction as a way to account for the above-mentioned morphosyntactic properties involving dative marking and the syntax of prepositions in EP, which support the distinctive status of datives and obliques.
BP and in DBP, PDC introduced by para ('to') is uniformly used.A related fact is the occurrence of first and second person clitics in both BP and in DBP, which allows for the above-mentioned (specific) facts about double pronominalization in both BP and in DBP, as well as the type of clitic doubling found in DBP.
Torres Morais & Salles (2022) have addressed the implications of the presence of first and second person clitics in BP (me/te), showing their pervasive distribution in dynamic and stative ditransitive predicates, in monotransitive predicates, inergative and unaccusative, including psychological verbs, and in predicates displaying ethical datives.In their analysis, they propose that the system of pronominal clitics in BP expresses a monovalent value with respect to the feature [participant], as only first and second person clitics are found, third person clitics being absent.They claim that the fact that first and second person clitics are marked as event participants allows for their unified occurrence in the specifier of the high applicative in BP, which is also consistent with the requirement that the argument in the high applicative enters a relation with the whole event.In this position, first and second person clitics are licensed under an agreement relation with the interpretable person feature on the high applicative head.
Under this approach, it is possible to account for double pronominalization, as well as to the type of PCC found in DBP and BP.A relevant point is that a difference arises regarding the structural licensing of the pronominal clitic.
If the clitic realizes the direct object (DO), as in (19b), repeated in (33a), it enters an Agree relation with the v head, being marked for structural Case, while the goal argument is licensed by the lexical preposition para ('to') (cf.(33b)).If the clitic realizes the goal argument, as in (19d), repeated in (34a), it is inherently licensed as a dative argument (under the thematic relation with the verb).In turn the DO, which is realized as a full pronoun, enters an Agree relation with the v head, being licensed under structural Case (cf.(34b)).
In the following subsection, we will discuss clitic doubling in DBP, which involves different structural and licensing properties as compared to clitic doubling in EP.

Clitic doubling in DBP
As mentioned above, DBP, and more restrictively BP, displays a type of clitic doubling, in which the doubled DP is introduced by the preposition para ('to'), not the dative marker a, as illustrated in (21a), from DBP, repeated in (35a), below.Moreover, in DBP, the doubled DP may occur without a preposition, as illustrated in (21b), repeated as (35b), below.In Machado Rocha's (2016) analysis, the absence of the preposition in (35b), above, provides support for the analysis of clitics in DBP as agreement markers.As such, they dispense with Case, while the full pronoun in the object position is licensed by the (little) v head in the VP shell projection.As pointed out by the author, "if clitics me and te were to receive case from v, there would be no source of Case to be assigned to the doubled DPs, unless they were always introduced by a preposition". 12 Galves (2020) notes that additional support for Machado Rocha's (2016) analysis could be the fact that this construction is also found in other dialects of BP, in relation to the innovative use of the second person clitic te.In Galves' analysis, an innovative property of the clitic te in BP is that it is deprived of the feature [familiarity], as confirmed by the example in (36), in which the clitic is doubled by the honorific DP a Senhora ('Madam').
(36) Eu te=dou esse livro para a Senhora I 2SG.DAT=give this book to the Madam 'I give this book to you [Madam]' (adapted from Galves 2020: 30) As already mentioned, Torres Morais & Salles (2022) provide an analysis for the presence of first and second person clitics in BP (as opposed to the absence of third person clitics), in which it is proposed that the system of clitics in BP displays a monovalent value for the feature [participant], which allows for their uniform occurrence as specifiers of a high applicative head.This property further interacts with the requirement that in this position the clitic enters a relation with the event (cf.section 3.2).
Following Torres Morais & Salles (2022), we propose that the clitics me and te in clitic doubling in DBP are licensed as a high applicative as well.However, differently from the contexts in which first and second person are not doubled (cf.( 16)), they occur as the spell out of the phi-features on the high Applicative head, given their non-argumental status in clitic doubling.In this respect, we follow Cuervo's (2003) analysis of ethical datives, which is crucially based the fact they are nonargumental, as well as on their occurrence in predicates with unergative verbs as well In the original: "Se os clíticos me e te recebessem caso de v, não haveria uma fonte atribuidora de Caso para os DPs redobrados, a menos que eles fossem sempre preposicionados."(Machado Rocha 2016: 88) as in configurations displaying a dative argument (involving a high applicative or an affected dative).Regarding the doubling DP, we propose that it is licensed either by the lexical preposition para ('to'), or in the DOC configuration, by the v head in the v-VP structure.In turn, in the former, the DO is structurally licensed by v (under Agree), while in the latter, it is licensed internally to the DOC configuration (see section 3.3), as shown in (37a) and (37b), respectively. 13(37) a. eu tô [vP ... v [ApplP te [VP falano [PP pra [DP você ]]]]] b. deixa eu [vP v [ApplP te [VP perguntar [ApplP ocê [Appl' [DP um negócio]]]]]] The idea of analyzing first and second person clitics in clitic doubling in DBP as the spell out of the high Applicative head, due to their non-argumental status, finds a correlate in Machado Rocha (2016) and Machado Rocha and Ramos' (2016) proposal of analyzing them as agreement markers.In present terms, first and second person clitics have two different syntactic expressions, occurring either in the specifier position of the high applicative projection, due to their argumental status in constructions without clitic doubling, or as proforms lexicalizing the high applicative head, due to their non-argumental status in clitic doubling constructions.

Final considerations
In this study, we have shown that ditransitive constructions in DBP and BP display innovative properties in the grammatical encoding of the indirect object¸ as compared to EP, which can be analyzed in terms of microparametric variation (cf.Roberts 2007;Biberauer & Roberts, 2017).We have proposed that the relevant property distinguishing them is the absence of the interpretable person feature on the high applicative head in both DBP and BP, implying the loss of the true dative category, and its related phenomena, namely clitic doubling and clitic clusters (of the EP type).In the absence of the interpretable person feature on the applicative head, the prepositional ditransitive construction (PDC) is uniformly found, thus implying an alignment with benefactive constructions.It was further argued that the absence of the interpretable person feature on the applicative head gave rise to a type of DOC in DBP, in which the goal argument is licensed structurally under an Agree relation with the upper v head.As a related fact, we have also discussed the occurrence of a type of clitic doubling in DBP (restrictively found in BP).Considering that only first and second person clitics are involved in these varieties of Portuguese, implying the monovalent expression of the feature [participant] on the clitic system, it was proposed that these pronouns are the spell out of a high applicative head, working as proforms grammaticalizing the universal expression of the interpretable person feature in the grammatical system.13 As expected, in the (restrictive) instantiation of clitic doubling in BP, the doubling DP is only introduced by the preposition para ('to'), given that DOC is not found in this variety.

(
35) a. eu to te=falano pra você I am 2SG.DAT=speaking to you 'I am talking to you' b. deixa eu te=perguntá ocê um negócio let I 2SG.DAT=ask.INF you.ACC a thing 'let me ask you something' (adapted from Machado Rocha 2016: 23)

Table 1 :
DBP pronominal system: clitics, full pronouns and obliques in the realization of the indirect object .

Table 3 :
The EP pronominal system realizing the indirect object