RRL, LXII, 4, p. 377–392, 2017
DOUBLING IN SPANISH AND CATALAN:
THE GRAMMATICALIZATION OF TOPICALIZATION DEVICES
ELISABETH AßMANN1
Abstract. I argue for a two-fold connection between DOM and CLD in Spanish and Catalan varieties which can be traced back to both phenomena being the result of the grammaticalization of topicalization structures. Under this view, the synchronic inter- and intralinguistic variation found in these varieties can be interpreted in a way that the varieties have not reached the same level of grammaticalization.
Keywords. Differential object marking, clitic doubling, Spanish, Catalan, grammaticalization, topicalization.
1. INTRODUCTION
Differential Object Marking (DOM) and Clitic Doubling (CLD) are widely attested and discussed phenomena of Romance languages. DOM (Bossong 1985, 1991) is the overt marking of a subset of direct objects by the preposition-like element a2, as exemplified in (1) for Spanish:
(1) a. Veo (*a) la mesa. b. Veo *(a) la mujer.
I.see A the table I.see A the woman ‘I see the table.’ ‘I see the woman.’
CLD, on the other hand, is a construction in which a DP in argument position co-occurs with a coreferential clitic in the same syntactic and prosodic domain, thus forming a discontinuous constituent with it (Jaeggli 1993, Anagnostopoulou 2005), exemplified in (2a.) for Spanish and (2b.) for Catalan:
1 Goethe-Universität Frankfurt/Main, Germany; [email protected].
2 This preposition-like element a is formally identical to the obligatory marker for indirect objects and to the directional preposition in both languages.
(2) a. Lei he preguntado a mi padrei. him I.have asked A my father
‘I asked my father.’
b. Emi van preguntar a mii. me they.go to.ask A to.me
‘They asked me.’
Both phenomena form part of the linguistic systems of Spanish and Catalan, which groups them together with other Romance languages such as Romanian and Sicilian, but sets them apart from others, most remarkably from Portuguese – a language closely related, but which accepts CLD only in very limited contexts and does not show DOM at all3.
However, if we have a closer look at Spanish and Catalan varieties, we also find great variation, and while some contexts seem to be possible for all varieties and both phenomena, in other contexts the languages and their varieties differ, as will be presented in detail in section 3. From this observation, I pose the following research questions:
1. How can we account for such differences, but also for the parallels, keeping in mind the closely related nature of the languages under consideration?
2. Is there a relation between the two phenomena and their distribution in the respective languages?
The proposal defended here is that the synchronic variation reflects different stages of the same diachronic development, viz., the grammaticalization of topicalization devices. It has been claimed that DOM as well as CLD origin from dislocation structures that serve to introduce or change topics. I propose that this is true for both Spanish and Catalan, though the two languages have grammaticalized the dislocation devices to different degrees4.
2. CONDITIONS FOR DOM AND CLD
Both DOM and CLD are known to depend on the feature make-up of the object DP (Suñer 1988, Franco 2000, Leonetti 2007, 2008, Rinke 2012), in particular semantic features such as animacy and specificity and morphosyntactic features such as definiteness, which are usually expressed in scales:
(3) Animacy scale (Croft 1988) human > animate > inanimate
3 At least in contemporary Portuguese. In earlier stages of the language, however, DOM can be found with personal pronouns and proper nouns, probably in consequence of convergence towards Spanish, which was considered as of high prestige (cf. Döhla 2014).
4 A similar idea has been proposed independently for Maltese and Old Sicilian, cf. Döhla (2016).
(4) Definiteness scale (Aissen 2003)
personal pronoun > proper noun > definite NP > indefinite specific NP >
non-specific NP
Aissen (2003:2) refers to DOM as being dependent on the prominence of the object. This is linked to the understanding of what a prototypical object is in contrast to a prototypical subject: “subjects tend to be definite, animate, and topic (thematic); while direct objects tend to be indefinite, inanimate, and rhematic”
(Comrie 1989:19). Therefore, whenever objects deviate from this definition, i.e., when they share properties with typical subjects, they tend to become overtly marked by the linguistic systems – this has been termed markedness reversal (Battistella 1990, Croft 1988, among others).
The combination of animacy and definiteness/specificity can also be understood as characteristics of topicality in the sense of given, activated information (Lambrecht 1994). Givón (1976) proposes the following hierarchies to capture the tendencies of topic-encoding elements:
(5) Topicality scale (Givón 1976) human > non-human definite > indefinite
agent > dative > accusative 1st person > 2nd person > 3rd person
The scale in (5) expresses a tendency for an NP to be topical ([+top]) if it corresponds more to the left-hand side features, e.g. Spanish mí, accusative or dative ‘me’, whose feature setup would be something like [+definite, +specific, +human, +1st person…]. With respect to the Topicality scale, we can say that this element is highly topical, i.e. very suitable to encode topic, salient information (in the sense of Givón 2001).
Both DOM and CLD are more likely to occur with objects that correspond to the left-hand side of each scale, and less likely with objects corresponding to the right-hand side of each scale. This relation of DOM and CLD to features connected with topicality has been previously noted and discussed, which has led scholars to propose that DOM and CLD originate in dislocated topicalization structures (Iemmolo 2010, Gabriel and Rinke 2010, among others). I propose that Spanish and Catalan undergo the same process with respect to the both phenomena described above: the two languages grammaticalize the means for CLD and DOM and therefore expand the possible contexts of applying those structures; however, the varieties are on different levels of this process, which explains the variation we find in distribution.
3. DIALECTAL VARIATION OF DOM AND CLD IN SPANISH AND CATALAN
In the present study, I consider three Spanish and three Catalan varieties:
Standard European Spanish, leísta Spanish from northern Spain and the Madrid region, rioplatense Spanish from the Buenos Aires region; Standard Catalonia Catalan, Spoken Valencian and Spoken Balearic Catalan. One condition is true for all varieties: CLD and DOM are obligatory with strong personal pronouns in object position5, irrespective of the object’s case, as exemplified in (6) for Spanish:
(6) Spanish
*(Te) he preguntado *(a) ti.
you I.have asked A you
‘I have asked YOU.’
The varieties under discussion can therefore only vary with respect to non- pronominal, lexical objects. For this reason they are the focus of the following discussion.
3.1. Dialectal variation of DOM in Spanish and Catalan varieties
If we consider that DOM depends on features such as animacy, definiteness, specificity, in short: on the degree of topicality (cf. section 2), the contexts for DOM in the discussed varieties can be summarized accordingly.
In Table 16, the rows show the varieties, ordered from more liberal to less liberal, while the columns represent the types of objects, ordered according to their morphosemantic make-up, i.e. their topicality. Table 1 shows that there is no variety accepting an inanimate, indefinite, unspecific direct object with DOM. At the same time, the opposite configuration, i.e. an animate, definite, specific direct object, is very likely a-marked.
5 According to Solà (1994), doubling of the third person pronouns in Catalan varieties is only optional. However, Todolí (1998) contradicts this statement and states that for many speakers a non- doubled strong third person pronoun would lead to a forced sentence.
6 Note for the following tables that indirect objects usually do not form part of the paradigm of DOM. However, since I take the a in the context of direct objects as a development from topic and dative marking (cf. chapter 4), I assume the obligatory marker for indirect objects part of the same process, thus marking the topical non-subject as such.
Table 1
Overview of the contexts of (D)OM in Spanish and some Catalan varieties
pron 1+2
pron 3 lex IO lex DO [+anim]
[±spec]
[+def]
lex DO [+anim]
[±spec]
[-def]
lex DO [-anim]
[+spec]
[+def]
all other configurat ions for lex DO [–anim]
rioplat. ❤ ❤ ❤ ❤ ( ) ( ) *
leísta + Stand
Span
❤ ❤ ❤ ❤ ( ) * *
Spok
Cent Cat ❤ ❤ ❤ ✓ * * *
Stand
Cat ❤ ❤ ❤ * * * *
pron = pronominal objects; lex = lexical object; IO = indirect object; DO = direct object;
anim = animate; spec = specific; def = definite; ❤ = obligatory; ✓ = preferred; ( ) = optional;
* = ungrammatical
For Standard Spanish, animacy is said to be the decisive feature for DOM as exemplified in example (1) above (cf. Torrego 2000, among others), but also definiteness plays a role:
(7) Vi (a) una mujer.
I.saw A a woman
‘I saw a (certain) woman. / I saw some (or other) woman.’
Una mujer, “a woman”, in (7) can be interpreted as specific or unspecific – but crucially, since the noun is indefinite, a-marking is optional (cf. Torrego 2000, Heusinger and Kaiser 2005). Leísta Spanish behaves like Standard Spanish in this respect. Rioplatense Spanish, on the other hand, seems to have relaxed the precondition that the direct object be animate in order for DOM to occur: even inanimate objects can be a-marked if they are [+specific] and [+definite], a configuration which is impossible in the other Spanish varieties, see example (8) (cf. also Zdrojewski and Sánchez 2014):
(8) Rioplatense Spanish (Suñer 1988: 399)
(Lo) quiero mucho (a) este arbolito.
him I.love a.lot A this little.tree
‘I love this little tree a lot.’
As for the Catalan varieties, we see that Spoken Central Catalan is one step behind the European Spanish varieties by only allowing DOM with direct objects that seem to be the most topical ones, i.e. animate, specific, definite (9a.), while indefinites are out (9b.).
(9) Spoken Central Catalan (Escandell Vidal 2009: 840) a. Les monges no estimen (a) les nenes.
the nuns not love A the girls ‘The nuns don’t love the girls.’
b. Vaig veure (*a) molts de policies al carrer.
I.go to.see A many of policemen at.the street ‘I saw many policemen in the street.’
Standard Catalan, on the other hand, is the most conservative variety and does not allow any marking for a direct object, the only exception being some universal quantifiers, some relative pronouns and reciprocal constructions7.
Spoken Balearic, finally, is an interesting case, as represented in Table 2:
Table 2
Overview of the contexts of (D)OM in Spoken Balearic pron
1+2 pron 3 lex IO lex DO [+dis]
[+spec]
[+anim]
[+def]
lex DO [+dis]
[+spec]
[–anim]
[+def]
lex DO [+dis]
[+spec]
[–anim]
[–def]
all other lex DO configurations
Spok Bal ❤ ❤ ❤ ✓ ( ) ? *
pron = pronominal object; lex = lexical object; IO = indirect object; DO = direct object;
dis = dislocated topic; spec = specific; anim = animate; def = definite; ❤ = obligatory;
✓ = preferred; ( ) = optional; ? = not as good; * = ungrammatical
On the one hand, it goes hand in hand with Standard Catalan in not allowing DOM except for pronominal objects, and is therefore more restricted than Spoken Central Catalan. On the other hand, it shows a-marking with a row of dislocated objects as long as they are specific: animate definite ones (10a.), inanimate definite ones (10b.) and, for less speakers, even with an inanimate, indefinite one (10c.)8:
7 Usually, DOM in Catalan is condemned as a sign of Spanish influence, such that its use is strongly discouraged by prescriptive grammars (e.g. Badia 1994).
8 The controversy surrounding examples like (10c.) likely relates to the fact that indefinite DPs are generally problematic when it comes to dislocation.
(10) Spoken Balearic
(ex. a.-c. from Escandell Vidal 2007, ex. d. from Escandell Vidal 2009) a. A ta mare, la vaig vore ahir.
A your mother her I.go to.see yesterday ‘Your mother, I saw her yesterday.’
b. (An)9es ganivets, els vaig ficar an es calaix.
A the knives them I.go to.put to the drawer ‘The knives, I put them into the drawer.’
c. ?(A) un gelat, me’l prendria amb A an ice cream me’it I.would.take with molt de gust.
much of pleasure
‘An ice cream, I would love to take that.’
As it seems, Spoken Balearic also marks topicality in the DO realm, however in a different context: the indispensable prerequisite for DOM in Spoken Balearic is that the object be dislocated. Within this dislocation paradigm then, the a-marking is still sensitive to the degree of topicality, just like in the other Spanish and Catalan varieties: while a specific, animate, definite dislocated DO is perfectly suitable for DOM, an unspecific, inanimate, indefinite one cannot be a-marked.
Tables 1 and 2 show that all Spanish and Catalan varieties have differential object marking – and that they all mark topical objects. For this reason, strong personal pronouns as well as indirect objects are obligatorily marked in all varieties under consideration. However, the linguistic systems differ with respect to the context in which topicality is differentiated.
Since it has been attested typologically that languages agree in using a configuration like DOM to mark topicality, but differ in terms of where to draw the line between topicality that needs to be marked and topicality that does not need to be marked, the inter- and intralinguistic differences with respect to Spanish and Catalan varieties therefore do not come as a surprise (Iemmolo 2010).
3.2. Dialectal variation of CLD in Spanish and Catalan varieties
In the same way as for DOM, the contexts for CLD can be represented according to the object’s degree of topicality: consider Table 3.
Parodi (1998) and Suñer (1988) argue that doubling with indirect objects is always possible and highly preferred10 in all Spanish varieties – Parodi (1998) even claims its obligatoriness for rioplatense Spanish, as exemplified in (11)11:
9 An is the Balearic allomorph of the usual a, used to avoid the merging of two unstressed vowels.
10 The only exception are bare plural nouns:
(i) (*Les) daré el dinero a personas pobres.
them I.will.give the money A persons poor
‘I will give the money to poor people.’
11 Apart from the linguistic variety, the semantic role of the object also plays a role regarding CLD with dative objects: while doubling is obligatory with benefactives and experiencers, it is mostly optional, though preferred, in ditransitive constructions (Fernández Soriano 1999; Cuervo 2003).
Table 3
Overview of the contexts of CLD in Spanish and Catalan varieties pron
1+2 pron 3
lex IO [+anim]
[+spec]
lex IO [–anim]
[+spec]
lex IO [+anim]
[–spec]
lex IO [–anim]
[–spec]
lex DO [+anim]
[+spec]
lex DO [–anim]
[+spec]
lex DO [±anim]
[–spec]
rioplat. ❤ ❤ ❤ ❤ ❤ ❤ ( ) ( ) *
leísta ❤ ❤ ✓ ✓ ✓ ✓ ( ) ( ) *
Stand
Span ❤ ❤ ✓ ✓ ✓ ✓ * * *
Spok Val ❤ ✓ ( ) ( ) ? ? * * *
Spok Cat ❤ ✓ ( ) ? ? ? * * *
Stand Cat ❤ ✓ * * * * * * *
pron = pronominal object; lex = lexical object; IO = indirect object; DO = direct object; anim = animate;
spec = specific; ❤ = obligatory; ✓ = preferred; ( ) = optional; ? = not as good; * = ungrammatical
(11) Rioplatense Spanish
??/*(Le) di un puñetazo a la puerta.
him I.gave a punch A the door ‘I hit the door.’
CLD with direct objects, on the other hand, is more restricted and therefore gives more room for dialectal variation. While in Standard Spanish, it is generally excluded, rioplatense Spanish (Suñer 1988) and Basque leísta Spanish allow optional doubling of [+human], [+specific] objects – the leísta varieties, however, only do so when the dative clitic le is employed12 (Bleam 1999) (cf. 12):
(12) Basque leísta Spanish (Bleam 1999: 52) (Le/ *la) vi a María.
her.DAT her.ACC I.saw A Maria
‘I saw Maria.’
Finally, in Madrid leísta Spanish (example 13) and also in rioplatense Spanish (Fernández Soriano 2000), it is even possible to double inanimate direct objects if they are [+specific]:
12 Hence the name leísta Spanish: accusative clitics lo, la are systematically replaced by the dative one, le, but only if they refer to [+human] entities.
(13) Madrid leísta Spanish (Suñer 1989: 384)
Pero no (la) voy a pisar la universidad.
but not her I.go to enter the university “But I will not enter the university.”
Catalan is generally more conservative with respect to CLD of lexical objects. Standard Catalan does not allow any kind of CLD with non-pronominal objects; this is usually condemned as Spanish influence, even with indirect objects (cf. Badia Margarit 1962). However, in all varieties of spoken Catalan, we find a high acceptability of CLD with indirect objects, preferably [+animate] ones (cf. 14):
(14) Spoken Catalan (Wheeler et al. 1999: 198)
(Li) hem d’entregar això al director.
him we.have to’hand.in that A.the director ‘We have to hand that in to the director.’
In more liberal varieties such as spoken Valencian, CLD is also possible with inanimate lexical indirect objects (ex. 15):
(15) spoken Valencian (Todolí 2002: 1399)
(Li) va pegar una patada a la porta.
him goes attach a kick A the door ‘He kicked the door.’
Interestingly, there seems to be a difference regarding the specificity of the indirect object: according to Todolí (1996), CLD with [–specific] indirect objects is not as good:
(16) Catalan (Todolí 1996: 451)
La beca (??li) donaran a un estudiant que the scholarship him they.will.give A a student that
haja aprovat totes les assignatures.
has.SUBJ passed all the subjects
‘They will give the scholarship to a student who has passed all of the subjects.’
CLD with direct non-pronominal objects, finally, is clearly ungrammatical in all Catalan varieties (Todolí 1996).
What the comparison of Spanish and Catalan varieties with respect to CLD, as represented in Table 3, shows, is again a sensitivity to topicality: the tendency to double highly topical objects on the left-hand side of the table, while objects that are less topical, i.e. on the right-hand side of the table, are never doubled. In between, we find an area of gradualness.
4. THE LINK OF VARIATION: GRAMMATICALIZING TOPICALIZATIONS
For both DOM and CLD, there have been proposals tracing the diachronic development back to dislocated topicalizing structures. Following this, I will propose how to account for the synchronic variation in Catalan and Spanish regarding DOM and CLD, and how these phenomena are related.
4.1. How topicalization structures are grammaticalized
Iemmolo (2010) argues that the usual approaches to DOM are insufficient.
The discriminatory approach, on the one hand, understands DOM and Case marking in general as a reflection of the marked status of objects with respect to unmarked ones (Comrie 1989, among others). According to Iemmolo (2010), it fails to explain why DOM in Romance languages starts from personal pronouns, the only category which still retains a case distinction. From the perspective of the indexing approach, on the other hand, the a-marker is thought to highlight the semantic role or properties of inherently salient NPs, or “affectedness” (Siewierska and Bakker 2008, among others). Iemmolo (2010: 243) claims that in some Romance languages, among them Catalan, the distribution of DOM does not seem to depend only on affectedness or the object’s semantics.
Therefore, Iemmolo (2010) proposes topicality as the triggering factor for DOM to appear and to extend, drawing on earlier ideas proposed by Pensado (1995), among others. According to this view, DOM is a means for coding the information structural role of (secondary) topics, which is a marked status for direct objects (cf. Lambrecht 1994). Since the a-marking appears first within left dislocations in Romance languages, Iemmolo (2010) proposes that DOM originates from these “marked syntactic constructions used to promote or re-introduce a topic referent” (Iemmolo 2010: 246). The a-marker used for marking this topicality stems from the preposition AD (basic meaning: motion towards a place), which already in Late Latin could be used to introduce dislocated topics. While Latin did not have anything similar to DOM, there is evidence that dislocated topics were the starting point for the spreading of the preposition to other DO-contexts in older stages of Romance, viz., personal pronouns and proper names (cf. also Aissen 2003; García and Putte 1995; Melis 1995). The grammaticalization path of the marker a proposed by Iemmolo (2010) is depicted in (17):
(17) Grammaticalization of the a-marker (Iemmolo 2010: 260)
allative marker > topic marker > dative marker > (differential) object marker Given this development, the assumption is that in earlier stages of the language, a only served to mark a dislocated topic. At some point, after reanalysing this lexeme first to a dative marker in base position, it spreads also to the direct
object paradigm. There, however, only the more topical objects are marked – which is the point of being a differential object marker. This would be the description of modern Peninsular Spanish and, to a lesser extent, also modern Spoken Catalan.
Consider (18) as an exemplification of this reanalysis:
(18) a. before reanalysis: a only for dislocated objects A María, la he visto.
A Maria her I.have seen
‘Maria, I have seen her.’
b. after reanalysis: a also for objects in argument position He visto a María.
I.have seen A Maria
‘I have seen Maria.’
Yet how does Spoken Balearic fit in the picture, given that DOM is only possible here with dislocated objects? Iemmolo (2010) proposes that grammaticalization can take different directions: either the spread or the narrowing of a construction. In Spanish and Spoken Catalan, the a-marking has been spreading to further contexts, which “involves the gradual loss of the link with information structure … the features relevant for marking to appear are merely referential properties such as animacy and/or definiteness.” (Iemmolo 2010: 259).
In Spoken Balearic, on the other hand, the construction has narrowed to the prototypical items of the category, i.e. dislocated direct objects, which are highly topical and are therefore the only ones marked. However, also in this paradigm of dislocated objects, topicality in the sense of the discussion in Section 2 plays a significant role: a-marking of highly topical objects is very acceptable, while this is not the case for non-topical objects.
For CLD, Gabriel and Rinke (2010) propose a fairly similar idea for the development in Spanish: the construction is the result of the grammaticalization of a dislocation structure, viz., clitic right dislocation (CLRD). While CLD and CLRD can be superficially similar in Spanish, they differ in various ways, e.g. with respect to information structure: whereas the dislocated object is necessarily a topic in CLRDs, the object can be (part of) the sentence focus in CLD structures. Given that clitic doubled objects are usually quite topical with respect to the Topicality Scale in section 2, this is an apparent clash: a topical object in a focal position. Basing their considerations on diachronic data that show that CLD starts with personal pronouns and experiencer arguments, Gabriel and Rinke (2010) argue that the doubling structure, i.e. the occurrence of a coreferential clitic and a (pro)nominal object in postnominal position, is the result of grammaticalizing a CLRD, where clitic and object occur in different clauses. Diachronic evidence comes from the fact that CLD occurs later than CLRD.
4.2. Putting together the pieces: depicting the synchronic parallels and differences regarding DOM and CLD in Spanish and Catalan Diachronic research (Fischer 2002, von Heusinger and Kaiser 2005, Dufter and Stark 2008, Gabriel and Rinke 2010, Fischer and Rinke 2013)13 gives reason to assume that the two languages pass through the same grammaticalization processes with respect to dislocation structures, which result in the phenomena of DOM and CLD, starting from strong personal pronouns – probably because these are inherently topical. Based on this assumption, we can easily explain the similarities and differences elaborated in section 2 by retaining the idea of representation in scales.
Regarding each phenomenon separately, it seems that the discussed Catalan and Spanish varieties have not reached the same level of grammaticalization of dislocation devices, since both DOM and CLD are still sensitive to the degree of topicality.
With respect to CLD, Parodi (1998) proposes to categorize Standard Spanish and rioplatense Spanish in an implicational hierarchy in order to capture the differences in doubling contexts:
(19) Implicational hierarchy (Parodi 1998: 91)
1 + 2 pronoun > 3 pronoun > [+animate DP] > [–animate DP]
[+definite DP] > [–definite DP]
[+specific DP] > [–specific DP]
Standard Spanish
rioplatense
Taking the hierarchy in (19) as a basis, it seems reasonable to understand the comparison between Spanish and Catalan varieties regarding the structures discussed in section 2 as a representation of an implicational hierarchy as well:
(20) implicational scale of CLD-structures in Spanish and Catalan varieties rioplatense ⊃ leísta ⊃ Stand Span ⊃ Spok Val ⊃ Spok Cat ⊃ Stand Cat
In (20), the varieties are ordered according to their acceptance of CLD:
rioplatense Spanish, which allows most contexts, is positioned on the far left side,
13 As far as DOM in Catalan is concerned, I am not aware of any published systematic diachronic study. It seems, however, that a-marking likewise starts with strong personal pronouns in the 13th century and becomes obligatory in the 14th century. Only after that do we find the marking with other strong pronouns (tots, el qual…) and certain proper names (Jorge Vega, p.c.). Meier (1947) gives examples of Old Catalan where proper names are a-marked. Also in the 19th century language, this was not uncommon (cf. Moll 1952).
while Standard Catalan, the most conservative one, is positioned on the far right side. The steps on the scale thus have to be interpreted not only as a description of the synchronic data, but also as a representation of the diachronic development.
The same logic can be employed for DOM. If we follow Iemmolo’s (2010) scale of the grammaticalization of the a-marker (cf. ex. 17), we can construe a developmental scale for the varieties under consideration that grammaticalize the a (cf. 21) and, building on that, create a developmental scale for the a-marking of structures (cf. 22):
(21) developmental scale of a-marking structures in Spanish and Catalan varieties rioplatense < other Spanish varieties < Spoken Catalan < Standard Catalan Spoken Balearic
(22) Grammaticalization of the a-marker in Spanish and Catalan
(differential) object marker < dative marker < topic marker < allative marker rioplatense
other Spanish varieties
Spoken Catalan
Standard Catalan
Spoken Balearic
(21) and (22) could be interpreted in a way that Spanish a has lost a major part of its topic-marking flavor and is therefore spreading to less topical objects – possibly, this grammaticalization path leads to the loss of all of the marker’s differential nature, which could turn it into a mere “object marker”. Catalan a, on the other hand, is a few steps behind and still ranges more on the “dative marker”
level as far as a more standardized variety is concerned. Spoken Balearic, on the other hand, is the odd one out: it does not stand exactly halfway between Standard and Spoken Catalan since its development with respect to grammaticalizing the a- marker took a different path. Therefore, I chose to put it under the actual scale in (21) and in between the “dative marker” and the “topic marker” steps in (22) so as to underline its peculiarity, which at the same time is not completely disjointed from the general picture.
5. CONCLUSION AND OUTLOOK
In this paper, I have argued for two things. First, DOM and CLD in Spanish and Catalan are connected synchronically as well as diachronically: in synchrony, both phenomena are sensitive to the object’s degree of topicality in the sense of animacy, definiteness and specificity. The more topical the object is, the more
likely it is to be doubled and/or a-marked. From a diachronic perspective, on the other hand, it has been argued that both phenomena are the result of the grammaticalization of dislocation structures, which served in older and modern stages of the languages to topicalize an object. Second, the synchronic variation found in Catalan and Spanish varieties can be interpreted in terms of this diachronic development. Some Spanish varieties, e.g. rioplatense, have developed the grammaticalizations furthest. Therefore, they a-mark and double even objects which are less topical in the sense of animacy, definiteness, and specificity, and which are excluded from these markings in other, more conservative varieties such as Standard Catalan. Finally, in Spoken Balearic Catalan, the grammaticalization of DOM has taken a different path as it mostly marks dislocation. However, also in this context, the a-marking is sensitive to the object’s degree of topicality.
With respect to future research, I believe it would be worthwhile to do a quantitative study for each variety focused on the reality of the two phenomena. In this way, it could be possible to underpin the affirmations made in the literature about the varieties’ behaviour regarding DOM and CLD. This might shed some light on open questions such as the seemingly unclear situation of Argentinian Spanish: is it really the case that DOM and CLD are much more acceptable in this variety? And if so, do the speakers also produce these structures spontaneously?
Two recent corpus studies that could be related with these questions come to interesting results: on the one hand, Döhla (2014:277) claims that “[t]he rumor … that DOM is fully grammaticalized in the Spanish of Argentina to the point where the preposition a always marks the direct object, independently of its properties, cannot be confirmed so far.” Unfortunately, the author does not give any numbers or statistics. On the other hand, Rinke, Wieprecht and Elsig (submitted) report two interesting facts: first, the rate of doubling with accusative DP-objects in rioplatense is significantly higher than the rate in Peninsular Spanish. Second, the effect of DOM (or animacy, respectively) on CLD is much more pronounced in rioplatense, where DOM is also exceptionally found with inanimate objects, than in Peninsular Spanish (ibd.). In my understanding, the two studies do not contradict each other, but point to the same conclusion: the marker a in Argentinian Spanish is still used for a differential object marking system, i.e. it is still sensitive to the object’s properties. However, in comparison to other varieties, primarily Peninsular Spanish, this DOM system is evidently more grammaticalized with respect to the factors discussed in section 2. This outcome should be taken as an argument in favour of the proposal of this paper. As for the other varieties discussed here, quantitatively analysed corpus studies would certainly close some gaps in the picture and would make the synchronic and diachronic link between DOM and CLD more evident.
REFERENCES
Aissen, J., 2003, “Differential object marking: iconicity vs. economy”, Natural language & linguistic theory, 21, 3, 435–483.
Anagnostopoulou, E., 2005, “Clitic doubling”, in: M. Everaert (ed), The Blackwell Companion to Syntax, Oxford, Blackwell, 519–581.
Badia Margarit, A., 1962, Gramática catalana, Madrid, Gredos.
Battistella, E., 1990, Markedness. The evaluative superstructure of language, Albany, State Univ. of New York Press.
Bleam, T., 1999, Leísta Spanish and the syntax of Clitic Doubling, University of Pennsylvania, PhD thesis.
Bossong, G., 1985, Empirische Universalienforschung. Differentielle Objektmarkierung in den neuiranischen Sprachen, Tübingen, Narr.
Bossong, G., 1991, “Differential Object Marking in Romance and beyond”, in: D. Wanner, D.H. Kibbee (eds), New analyses in Romance linguistics, Amsterdam/Philadelphia, Benjamins, 143–170.
Comrie, B., 1989, Language typology and linguistic universals: syntax and morphology, Chicago, University of Chicago Press.
Croft, W., 1988, “Agreement vs. case marking and direct objects”, in: M. Barlow, C. Ferguson (eds), Agreement in natural language: Approaches, theories, descriptions, Stanford, CSLI, 159–179.
Döhla, H.-J., 2014, “Diachronic convergence and divergence in differential object marking between Spanish and Portuguese”, in: K. Braunmüller, S. Höder, K. Kühl (eds), Stability and divergence in language contact. Factors and mechanisms, Amsterdam/Philadelphia, Benjamins, 265–289.
Döhla, H.-J., 2016, “The origin of differential object marking in Maltese”, in: G. Puech, B. Saade (eds), Shifts and patterns in Maltese, Berlin/Boston, de Gruyter, 149–174.
Dufter, A., E. Stark, 2008, “Double indirect object marking in Spanish and Italian”, in: E. Seoane, M.J. López-Couso (eds), Theoretical and empirical issues in grammaticalization, Amsterdam/
Philadelphia, Benjamins, 111–130.
Escandell Vidal, V., 2007, “Topics from Ibiza: differential object marking and clitic- dislocation”, in:
G.A Kaiser, M. Leonetti (eds), Proceedings of the workshop “Definiteness, specificity and animacy in Ibero-Romance languages”, Konstanz: Universität Konstanz, 23–43.
Escandell Vidal, V., 2009, “Differential object marking and topicality. The case of Balearic Catalan”, Studies in Language, 33, 4, 832–885.
Fernández Soriano, O., 2000, “El pronombre personal. Formas y distribuciones. Pronombres átonos y tónicos”, in: I. Bosque, V. Demonte (eds), Gramática descriptiva de la lengua española 1, Madrid, Espasa Calpe, 1317–1397.
Fischer, S,. 2002, The Catalan clitic system: a diachronic perspective on its syntax and phonology, Berlin, Mouton de Gruyter.
Fischer, S., E. Rinke, 2013, “Explaining the variability of clitic doubling across Romance: a diachronic account”, Linguistische Berichte, 236, 255–272.
Franco, J., 2000, “Agreement as a continuum. The case of Spanish pronominal clitics”, in:
F. Beukema (ed), Clitic phenomena in European languages, Amsterdam/Philadelphia, Benjamins, 147–190.
Gabriel, C., E. Rinke, 2010, “Information packaging and the rise of clitic doubling in the history of Spanish”, in: G. Ferraresi, R. Lühr (eds), Diachronic Studies on Information Structure:
Language Acquisition and Change (Language, Context, and Cognition 10), Berlin, Mouton de Gruyter, 55–77.
Givón, T., 1976, “Topic, pronoun, and grammatical agreement”, in: C. Li (ed), Subject and topic, New York, Academic Press, 149–188.
Givón, T., 2001, Syntax. An introduction, Amsterdam/Philadelphia, Benjamins.
Iemmolo, G., 2010, “Topicality and differential object marking. Evidence from Romance and beyond”, Studies in language, 43, 2, 239–272.
Jaeggli, O., 1993, “Tres cuestiones en el estudio de los clíticos: el caso, los sintagmas nominales reduplicados y las extracciones”, in: O. Fernández Soriano (ed.), Los pronombres átonos.
Madrid, Taurus, 141–173.
Lambrecht, K., 1994, Information structure and sentence form. Topic, focus and the mental representations of discourse referents, Cambridge, Cambridge University Press.
Leonetti, M., 2007, “Clitics do not encode specificity”, in: G. Kaiser, M. Leonetti (eds), Definiteness, specificity and animacy in Ibero-Romance languages, Konstanz, Universität Konstanz, 111–139.
Leonetti, M., 2008, “Specificity in Clitic Doubling and in Differential Object Marking”, Probus, 20, 1, 33–36.
Melis, C., 1995, “El objeto directo preposicional en El Cantar de Mio Cid. Estudio sintáctico-pragmático”, in C. Pensado (ed), El complemento directo preposicional, Madrid, Visor, 133–163.
Parodi, T., 1998, “Aspects of clitic doubling and clitic clusters in Spanish”, in: R. Fabri et al. (eds), Models of inflection, Tübingen, Niemeyer, 85–102.
Pensado, C., 1995, “El complemento directo preposicional: estado de la cuestión y bibliografía comentada”, in: C. Pensado (ed), El complemento directo preposicional, Madrid, Visor Libros, 11–59.
Rinke, E., 2012, “El doblado de clíticos en el español estándar y el argentino: variación lingüística y análisis sintáctica”, in: Á. di Tullio, R. Kailuweit (eds), El español rioplatense: lengua, literatura, expresiones culturales, Frankfurt am Main, Vervuert, 103–117.
Rinke, E., J. Wieprecht, M. Elsig (submitted), “Clitic doubling in Peninsular and Rioplatense Spanish:
a comparative corpus investigation”, Ms., Universität Frankfurt.
Siewierska, A., D. Bakker, 2008, “Case and alternative strategies”, in: A. Malchukov, A. Spencer (eds), The Oxford handbook of case, Oxford, Oxford University Press, 290–303.
Suñer, M., 1988, “The role of agreement in clitic-doubled constructions”, Natural language and linguistic theory, 6, 391–434.
Suñer, M., 1989, “Dialectal variation and clitic-doubled objects”, in: C. Kirschner, J. DeCesaris (eds), Studies in Romance linguistics, Amsterdam, Benjamins, 377–395.
Torrego, E., 2000, “El complemento directo preposicional”, in: I. Bosque, V. Demonte (eds), Gramática descriptiva de la lengua española 2, Madrid, Espasa Calpe, 1779–1805.
Todolí, J., 1996, “Doblament i represa pronominal en català”, in: A. Schönberger, T. Stegmann (eds), Actes del desè col·loqui internacional de llengua i literatura catalanes, Barcelona, Publicacions de l’Abadia de Montserrat, 441–464.
Todolí, J., 2002, “Els pronoms”, in: J. Solà et al. (eds), Gramàtica del català contemporani 2, Barcelona, Empúries, 1337–1433.
von Heusinger, K., G. Kaiser, 2005, “The evolution of differential object marking in Spanish”, in:
K. von Heusinger et al. (eds), Proceedings of the workshop “Specificity and the evolution/emergence of nominal determination systems in Romance”, Konstanz, Universität Konstanz, 33–70.
Wheeler, M., A. Yates, N. Dols, 1999, Catalan: a comprehensive grammar, London, Routledge.
Zdrojewski, P., L. Sánchez, 2014, “Variation in accusative clitic doubling across three Spanish dialects”, Lingua, 151, 162–176.
