THE GRAMMATICALIZATION AND PRAGMATICALIZATION OF THE ROMANIAN INDEFINITE PRO-FORMS : A CORPUS-BASED APPROACH

The present paper aims to explore, and address some issues concerning the Romance diachronic morphosyntax in the light of theoretical and methodological considerations on the grammaticalization and pragmaticalization phenomena, and the question of linguistic change. Building on the previous work concerning grammaticalization, we intend to reveal a pragma-linguistic scenario that accounts for the actual situation of the Romanian indefinite compounds. We consider this subject to be very particular, meaning that the historical development of these pro-forms is not a canonical case of grammaticalization or pragmaticalization.

Language change, with a focus on grammaticalization and pragmaticalization, is considered to be an essential and universal feature of human language, and, by investigating the laws of language change, we learn a great deal about language in general.As concerns grammaticalization, it has been underlined that the same set of processes and mechanisms are responsible for all aspects of grammar.Thus, all grammatical morphemes have developed out of lexical morphemes, principally nouns and verbs, and all grammatical structures have developed out of more loosely organized constituents (Bybee 2003), as in the case of indefinite pro-forms that developed out of verbs.Above all the controversies relating to the origin of these pro-forms, there are still many questions unanswered regarding the evolution steps and the relation existing between the forms that have coexisted for long periods in language use.
The next four sections of the article define the subject of the present study and describe the forms of these Romanian linguistic items (Section 2), present the methodological framework and some considerations on the data (Section 3), investigate the indefinite pro-forms in three stages of the Romanian language evolution and underline the grammaticalization/pragmaticalization phenomenon (Section 4), and summarize the final considerations (Section 5).
2. Definition, form and etymology of the indefinite pro-forms Indefinite pronouns are pronouns whose main function is to express indefinite reference.(Haspelmath 1997: 10-11) In our study we will use the cover term pro-forms proposed by Haspelmath, comprising pro-nouns, pro-adverbs andpro-adjectives. (Vater 1975 apud Haspelmath 1997: 10) Although the paradigm of these pro-forms in the Romanian language is large, we shall focus only on the indefinite pro-nouns and pro-adverbs constructed with 1 This paper is supported by the Sectoral Operational Programme Human Resources Development (SOP HRD), ID134378 financed from the European Social Fund and by the Romanian Government.
The pro-forms analysed here are structured as compounds: vare/oare/ori/veri ('any' -proclitic element) + care, cine, ce, cât, ('which', 'who', 'what', 'how much' -relative pronouns) vare/oare/ori/veri ('any' -proclitic element) + unde, când, cum, cât ('where', 'when', 'how', 'how much' -relative adverbs) From a large number of etymological debates on the roots of the indefinite particles we embrace the perspective that proposes the lat.volet (according to the alb.valle) as the historical origin of oare, while the term ori is the result of the transformation of vare/oare through some syntactic and phonetic criteria 2 (Dimitrescu 1974).As it shall be demonstrated in our research, this hypothesis is confirmed by a large number of contexts in which oare appears with the disjunctive meaning ori.
Analysed from a diachronic perspective, indefinite pronouns are considered to arise from a limited number of sources, the most common being phrases with original meanings such as whatever it may be, it does not matter which, or it is the same which, which come to acquire the 'free choice' function and then spread first to the functions adjacent to it and then to more distant ones (O Dahl 2005).

Grammaticalization
Taking the definitions, the characterizations, and the identified parameters (Lehmann 1995), principles (Hopper 1991) and characteristics (Brinton and Traugott 2005), as a starting point, Beijering (2012) proposes a new approach to grammaticalization in which formal reanalysis 3 (language change) and semantic reinterpretation are equally important.She distinguishes between two types of grammaticalization, viz.primary grammaticalization (Gzn1) (from lexical to grammatical status), and secondary grammaticalization (Gzn2) (from grammatical to (more) grammatical status) and proposes the following definition of grammaticalization: Grammaticalization is a composite type of language change whereby lexical or already grammaticalized items, in certain linguistic contexts, undergo both semantic reinterpretation and formal reanalysis.It is accompanied by a subset of correlated primitive changes and side effects.Grammaticalization leads to a grammatical item, i.e. a linguistic item belonging to a minor category, with 3 In this paper we take into consideration the following definition of reanalysis given by Beijering (2012: 37): Reanalysis is a covert operation that results in a new structural representation for a given linguistic string that is not immediately noticeable at its surface manifestation.There are different types of reanalysis that apply to different linguistic levels (hierarchical structure, constituent structure and category label), but there is no generally accepted terminology to refer to its subtypes.clac 65/2016, 223-256 Preda y Ardeleanu: romanian indefinite pro-forms 228 relational meaning, secondary status, the prime function of which is to regulate grammatical structure and grammatical relations.(Beijering 2012: 47) From this perspective, Gzn1 implies categorical reanalyses (formal reanalysis from major to minor category) and reinterpretation (semantic reinterpretation from referential to relational meaning), and it is accompanied by a set of primate changes, such as: loss of morphosyntactic properties (attrition) and loss of semantic substance (bleaching), and followed by side effects: increase in paradigmaticity, structural scope reduction, layering (synchronic variation of a given form), divergence (split), specialization, persistence, context expansion, increased type and token frequency and typological generality (cross-linguistic patterns).On the other hand, Gzn2 implies formal reanalysis from minor to minor category (categorial reanalyses), and semantic reinterpretation of relational meanings, being accompanied mandatorily only by loss of semantic substance (bleaching) and followed by side effects as: increase in paradigmaticity, decrease of paradigmatic variability, structural scope reduction, layering (synchronic variation of a given form), divergence (split), specialization, persistence, productivity increased type and token frequency.(Beijering 2012)

Pragmaticalization
Considering the proposed definitions (Erman andKotsinas 1993, Diewald 2011) and the identified characteristics (Aijmer 1997) as a departure point, Beijering (2012) proposes a new definition of pragmaticalization (Pgzn): Pragmaticalization can be thought of as the study of the origin and rise of discourse markers, as well as the gradual diachronic change leading to discourse markers.i.e. a linguistic item with conversational meaning, extrapropositional status, the prime function of which is to organize discourse structure.(2012: 56) Thus, Pgzn implies hierarchical reanalyses, meaning formal reanalysis from propositional to extra-propositional status and semantic reinterpretation from clac 65/2016, 223-256 Preda y Ardeleanu: romanian indefinite pro-forms 229 referential/relational meaning to conversational meaning (=(inter)subjectification 4 ), is accompanied by loss of semantic substance (bleaching) and gain of speaker's perspective (subjectification) and is followed by a set of side effects: layering (synchronic variation of a given form), divergence (split), specialization, persistence, productivity (context expansion) and increased type and token frequency (Beijering 2012).

Romanian indefinite pro-forms: results
4.1.The Old Romanian language (16 th -18 th c.) Starting with the first attested Romanian text -Psaltirea Hurmuzaki (1500) -we noticed the co-occurrence of the indefinite pro-forms, both composed with oare (vare) and ori.This fact demonstrates that all these compounds formed in an earlier stage of the Romanian language, meaning the Common Daco-Romanian (13 th -14 th c.).These were used in spoken language or in texts that have not been conserved until present time, and this explains the free variation of the forms at the beginning of the Old Romanian language.
( As far as the pro-adjectival use is concerned, the texts revealed two situations: one in which the pro-form is located in front or after the modified noun ( 7), and another one in which the pro-adjective takes a fixed position -in front of the noun with a preposition (de, la, în, cu, din, pre, supt, asupra)  According to SILR, the situation of the pro-forms in Modern Romanian was slightly different from the previous stage investigated.
4.2.1.Indefinite pro-nouns, also used in some contexts as pro-adjectives and connectors As far as the indefinite pro-nouns and pro-adjectives are concerned, the authors of SILR noticed that the use of the ori-series (oricine, orice, oricare, oricât) has been generalized in the 19 th century.The pro-noun oricare, identified in the previous stage as having variable uses, keeps its variability at the beginning of the century, but the unmarked forms prevail during the entire century.Thus, the forms oricarea, oricarii disappear and oricarele is rarely used.The 20 th century texts reveal the possibility of combining the pro-adjective orice/oricare with another indefinite form as: alt, altceva, in order to intensify the indefinite value.In the first two decades of the century the construction is rare (orice alt cetăţean ['any other citizen']), and it starts to increase in number after 1930 (orice altceva ['anything else']).Comparing this situation with the one found in OR, we can underline the difference in topic preference of the indefinite form alt, located in front of the indefinite pro-form, as in (10).

Pro-adverbs and connectors
Among the pro-adverbs, the oare-series is more frequently used than the ori-, oarecum being the form most used within the series.Adding to the forms vareîncotro/oareîncătruo specific to the old language, we can notice in this period the presence of the form ori încătro: ( The steps within the grammaticalization path cannot be clearly distinguished, as in the first texts attested for the Romanian language (16 th.c.) all these forms coexist with all the functions above mentioned.The verbal origin of the interrogative particle vare (VOLET) has already been established by previous research on grammaticalization (Dinică and Zamfir 2009), while the interrogative source of the indefinite proclitic element vare has been well-argumented (Haspelmath 1997).In addition, we consider that this grammaticalization step from the Old Romanian language has been possible due to the morphosyntactic association of the interrogative particle with an interrogative pro-noun (cine, ce).Later on, vare loses its interrogative meaning, starting the univerbation process of the indefinite pro-forms.The process had several stages of grammaticalization: in the first period, the two elements were unbound (vare ce, vare Semantically, there are contiguous, metonymic relations between the various interrogative (58), disjunctive (59), indefinite (60) and concessive (61) meanings of vare/oare/ori(ce), all being nuances in the domain of doubt/uncertainty and approximation.
( The primitive changes that accompany the formal reanalysis and semantic reinterpretation are represented by the features described in Section 3.2.At the phonetical level, there is some loss of phonological/phonetic substance in the development of the Romanian orice from voare/oarece (see 2.1).In addition, one may notice some changes in the morphological compositionality, since orice is a compositional form, and its internal structure was affected by reduction: oare + ce > orice.
We also notice the loss of morphosyntactic properties due to the shift form indefinite pro-noun to indefinite pro-adjective.The use in the Present-Day Romanian language of clac 65/2016, 223-256 Preda y Ardeleanu: romanian indefinite pro-forms 248 orice as pro-form and sentence connector results in a loss of syntactic variability (the fixation of the pro-adjective in prenominal position) and autonomy (does not allow interpositioning) of the indefinite orice.In extension, orice starts to be employed more frequently in quasi-phrasal structures (Chivu et al. 2012:517), such as: orice s-ar zice ('whatever they say'), orice s-ar întâmpla ('whatever it may happen') or it allows the combination with another indefinite pro-form: orice altceva aş fi făcut ('anything else I would have done').As far as the semantic changes are concerned, the data revealed a more bleached meaning of orice.It increasingly develops various nuances in the realm of indefiniteness.
As far as orice is concerned, the side effects are noticeable.Paradigmaticization 7 is generally related to productivity and frequency, all being correlated in the development of orice.In a general sense, orice enters the indefinite pro-noun paradigm, and is a frequent linguistic item.Obligatorification (decrease of paradigmatic variability (Beijering 2012: 48)) clearly does not apply to orice as it is a free choice element, exception being the concessive connector status.Likewise, condensation 8 does not occur because the structural scope of orice is not reduced.
With respect to layering, we see that the interrogative particle vare(ce) was reanalysed as the indefinite proclitic element ori(ce).The indefinite pro-form ori can be used to express various dimensions of indefiniteness.It may be used as a relative/concessive sentence connector, or it may occur in set phrases.As regards divergence 9 , it can be observed that the source of ori(ce), the interrogative form vare(ce) continued to exist in the 16 th c. along with the new indefinite form.Specialization does not apply to orice, as similar and simultaneously existing expressions are not reduced to one major expression.Persistence relates to the observation that a linguistic item or construction retains traces of the linguistic item or construction from which it emerged.In the second 7 Paradigmaticity is defined as The cohesion of a sign with other signs in a paradigm,(...) that is, the degree to which it enters a paradigm, is integrated into it and dependent on it.( Beijering 2012: 42)  8 As defined by Beijering 2012: 108, condensation represents a decrease in syntactic scope, but also an increased dependency. 9When a lexical form undergoes grammaticization to clitic or affix, the original lexical form may remain as an autonomous element and undergo the same changes as ordinary lexical items.( Beijering 2012: 44)  The primitive changes that accompany the formal reanalysis and semantic reinterpretation are explained below: at the phonetical level, there is some loss of phonological/phonetic substance in the development of the Romanian oricum from varecum/oareceum (see 2.1).In addition, one may notice changes in morphological compositionality, since oricum is a compositional form, its internal structure being affected by reduction: oare + cum > oricum.
We also notice the loss of morphosyntactic properties, due to the shift form interrogative adverb vare/oare to indefinite proclitic element oare/ori.The use of oricum as a pro-adverb and sentence connector in Present-Day Romanian results in loss of syntactic variability (the adjectival use of these forms in Old Romanian language is lost).In extension, oricum starts to appear more in quasi-phrasal structures, such as: oricum ar fi.
As far as the semantic changes are concerned, oricum has a more bleached meaning, developing various nuances in the realm of indefiniteness and modality: cumva With respect to layering, the data show instances of coexistence of both adverbial and subordinating oricum.The older oricum (with indefinite pro-adverbial properties) still exists along with the newer oricum that functions also as a sentence connector.As for the divergence, it can be noticed that the source of ori(cum) -the interrogative form vare(cum) -continued to exist in the OR besides the new indefinite form.Specialization does not apply to oricum, as similar and simultaneously existing expressions are not reduced to one major expression.Persistence relates to the fact that a linguistic item or construction retains traces of the linguistic item or construction from which it emerged.
In the second stage of the grammaticalization process, oare(cum) retains the properties of the interrogative particle vare, but these properties are lost in the case of the indefinite ori(cum).
The pragmaticalization of the indefinite pro-adverb oricum implies hierarchical reanalysis, meaning the shift from a propositional (67) to an extra-propositional (68) status, and reinterpretation from relational to communicative meaning.
( The primitive changes triggered by pragmaticalization manifest at semantic (bleaching) and discourse (increased speaker-perspective, attitude or judgment (subjectification) and attention to speaker-addressee interaction (intersubjectification)) levels.The side effects of the pragmaticalization of oricum result in layering and specialization, context expansion and increased frequency.
In conclusion, the patterns show that the development of oricum has most properties in common with secondary grammaticalization and pragmaticalization.
cine, vare care), allowing the interpositioning of a preposition (see examples 41 and 42 above), and it evolved through the stage of a bound indefinite pronoun (varece, varecine, varecare).clac 65/2016, 223-256 4.4.2.Case study -the grammaticalization scenario of the indefinite pro-noun oarece/orice vare (interrogative particle) > vare (indefinite proclitic element) + ce (relative pronoun) > (indefinite pro-form, vare ce/varece/vare prep.ce) > oare (interrogative particle/indefinite proclitic element) + ce (relative pronoun) (indefinite pro-form, oare ce/oarece/oare prep.ce) > ori + ce (indefinite pro-form, orice, ori prep.ce) First, we will focus on the essential mechanisms used in grammaticalization, on the primitive changes that occur and on the possible side effects of the process, and then we will comment on the linguistic status (i.e.lexical, grammatical or communicative) and stages of grammaticalization of the Romanian indefinite pro-form orice.The first step in the development of orice is represented by the reanalysis of the futuretense auxiliary vare into the indefinite element oare.Once the indefinite status has been established, oare agglutinated to ce and developed into orice, due to some phonetic reasons, cf. the examples in Section 4.1.The other gradual evolution of orice -from the subjunctive verbal form veri into the disjunctive connector ori, and then into the indefinite pro-noun orice -occurred probably simultaneously and closely influenced by these forms already used in the Old Romanian language.

(
'somehow'), într-o oarecare măsură ('to some extent'), în orice fel ('in any way'), măcar ('at least').clac65/2016,223-256   Paradigmaticization is generally related to productivity and frequency, all being correlated in the development of oricum.In a general sense, oricum enters the indefinite pro-adverb paradigm, and is a frequent linguistic item.Obligatorification clearly does not apply to oricum as it is a free choice element, exception being the concessive connector status.Likewise, condensation does not occur because the structural scope of oricum is not reduced, nor does it become (more) dependent upon other constituents in the clause, rather the opposite.That is, sentence adverbs are integrated into syntactic structures, but flexible with regard to their positions.
who offers quite a novel theoretical frame.The actual analysis focused on the semantic meanings, morphosyntactic and pragmatic functions taken by these forms during the grammaticalization /pragmaticalization paths: vare (interrogative particle) + pronoun/adverb > vare (indefinite proclitic element) + pro-noun/adverb > oare (interrogative particle/indefinite proclitic element) + pro-noun/adverb > ori (indefinite proclitic element) + pro-noun/adverb.The analysis of the data (comprising texts from 16 th -21 st centuries) has shown that the grammaticalization (that started from the Common Daco-Romanian stage of language formation as mentioned by different researchers) and pragmaticalization processes of the Romanian indefinite pro-forms continued and finalized in the Present-Day Romanian language.clac 65/2016, 223-256 interposed between the two elements of the And in any town you will get and you will be welcomed, do eat!' Certain writers prefer the second case, an example being Dosoftei, who, in Psaltirea în versuri, makes use of 23 interposed constructions out of 33 contexts with pro-forms.In some of the contexts in which the pro-forms function as connectors of headed relative clauses, the head is represented by the quantifier tot (9): Ardeleanu: romanian indefinite pro-forms 235 periods of language formation.This singular case can represent the first sign of the pragmaticalization process of such pragmatic markers 5 as the use of the marker implies that there is more to be said, but the author resorts to shared knowledge.Due to the fact that these forms were not yet tied, they appeared in free variation and there are contexts in which the pro-adverb oarecum (30) is used as periphrastic As much as you will pay, I will pay it back to you.'Among these pro-forms, there is the indefinite vareîncotro/oare-încătruo, specific only to the Old Romanian language, while its indefinite value is lost in the following language formation stages (35).
Considering the oare-series, it can be noticed that the form oarecare is widely employed during the entire century, while the forms oarecine, and oarece are rarely clac 65/2016, 223-256 used.'anykind of merchandise will he get' Although the authors of this volume state that the indefinite pro-forms compound with veri-are attested at the beginning of the 19 th century in Muntenia (vericare, verice), but also in Moldova (verice, verce), the analysis of the Old Romanian corpus revealed the presence of the veri-series as early as the 17th century (BB.1688).The grammars of the first half of the 20 th century record the forms vericare, vericine, verice as obsolete.A peculiar situation is the one in which orice is used with a noun in the plural, situation considered noncompliant with the Romanian grammar norms: chei pentru orice sertare ('keys for any drawers').
So, based on the comparison of the two major periods (OR and MR), we conclude that, beyond the variation of the formal aspects of these pro-forms, the morphosyntactical uses remain, mainly, the same.
Ardeleanu: romanian indefinite pro-forms 249 stage of the grammaticalization process, oare(ce) retains the properties of the interrogative particle vare, but these properties are lost in the case of the indefinite ori(ce).
clac 65/2016, 223-256 Preda y (indefinite pro-form, vare cum/varecum) > oare (interrogative particle / indefinite proclitic element) + cum (relative adverb) (indefinite pro-form, oarecum) > ori + cum (indefinite pro-form, oricum) We focus next on the essential mechanisms used in grammaticalization and pragmaticalization, on the primitive changes that occur and on the possible side effects of the process, and then we concentrate on the linguistic status and stages of grammaticalization and pragmaticalization of the Romanian indefinite pro-form oricum.The development of oricum has as its initial stage the reanalysis of the future-tense auxiliary vare into the indefinite element oare.Once the indefinite status has been established, oare agglutinated to cum and developed into oricum, due to phonetic reasons, cf. the examples in Section 4.1.