Partitive determiners in Piedmontese: A case of language variation and change in a contact setting

1 Introduction^[1]

The paper addresses the use of partitive determiners in Piedmontese, a Gallo-Italian dialetto spoken in the north-western Italian region of Piedmont. Following the conventions of Italian research, we use the label dialetti (singular dialetto) to refer to “primary dialects” (in Coseriu’s 1980 terminology) which – in most areas – are to be deemed, to a varying degree, Abstand languages, i. e., linguistic systems separate from Italian (see e. g., Maiden and Parry 1997: 2). Moreover, a dialetto is employed in limited areas (i. e., in a region, such as Piedmont, or in a part thereof) and within limited domains (i. e., mostly in oral and informal communication), its socio-cultural “roofing” (German Überdachung) being supplied by standard Italian. Gallo-Italian dialetti are generally considered by Italian scholars (see e. g., Pellegrini 1975 [1973]) as belonging to Italo-Romance; this will also be our starting assumption, while the whole question will be re-addressed in Section 6. The investigation of partitive determiners in Piedmontese provides us with an opportunity to elaborate on some key aspects of language variation and change, as well as language contact, in Italo-Romance. At the same time, it offers valuable insight into the interplay of quantification and classification in the grammar of indefinite determiners, with special reference to Romance languages.

Firstly, a brief overview will be provided of the sociolinguistic situation under scrutiny; particular emphasis will be given to the coexistence of Piedmontese, Italian, and French, which lasted until the end of the nineteenth century, and the superposition of both an official standard language (i. e., Italian) and a regional koine (based on the variety of Turin) over local varieties of Piedmontese, which is still relevant (Section 2). This will be followed by the description of partitive determiners in the regional koine, the latter having been codified by reference grammars (Section 3). Then, a diachronic account will be sketched of the development of partitive determiners in Piedmontese; such development will be claimed to reflect, on the one hand, how paradigmatic differences between different varieties of Piedmontese have been preserved, and on the other, how the influence of the regional koine on written texts has changed over time (Section 4). The attention will finally turn to the behavior of partitive determiners in contemporary Piedmontese. The analysis will help to shed light on the similarities and differences not only between different varieties of Piedmontese, but also between such varieties and Italian, and the main paradigmatic differences identified will be argued to relate to two different ways of categorizing the relationship between mass nouns and countable plurals (Section 5).

Our arguments will draw primarily on some unpublished materials collected within the ALEPO research program. ALEPO stands for Atlante linguistico ed etnografico del Piemonte Occidentale, ‘Linguistic and Ethnographic Atlas of Western Piedmont’ (http://www.alepo.eu): This atlas covers an area in which Italian is spoken alongside two Gallo-Romance varieties, i. e., Francoprovençal and Occitan, and a Gallo-Italian dialetto, i. e., Piedmontese, ^[2] and includes localities pertaining both to the Gallo-Romance and to the Gallo-Italian domains. For our purposes, we will focus in particular on some materials collected in the latter domain. Such materials are composed of both responses to a questionnaire survey and spontaneous speech data (see Section 5, fn. 11); the latter, which can be called ‘ethnotexts’ (cf. Canobbio 1989, 2002; see the notion of ethnotexte in Bouvier et al. 1980), consist mainly of conversations about traditional aspects of community life and culture. Moreover, we will rely on relevant data from other linguistic atlases (ASIS and AIS), as well as from previous studies conducted on written texts and spontaneous speech (cf. Sections 4 and 5).

We will follow Squartini (2017) in adopting a “paradigmatic approach”. From this perspective, the behavior of partitive determiners in each given variety of Piedmontese will be dealt with in relation to the overall organization of the paradigm of indefinite determiners in that same variety. Only the noun phrases displaying an argumental function will be considered. Mention will also be made of how internal, external, and extra-linguistic factors can condition the presence of partitive determiners (or the lack thereof) and the form they take.

2 Piedmontese: a socio-historical introduction

Using the label Piedmontese, we can refer to two slightly different linguistic objects: 1) a group of varieties spoken in the central part of Piedmont; or 2) a regional koine which, as of the eighteenth century, developed from the urban variety of the most important center of the area, Turin, and started acting as a reference code for the surrounding varieties of dialetto, also playing the role of a “lingua franca” among speakers of different varieties of Piedmontese. The mere existence of a regional koine sets Piedmontese apart from the bulk of Italo-Romance dialetti; the relationship between the variety of Turin and the Piedmontese koine, however, shares many similarities with that between the variety of Venice and the Venetan koine, the latter being based on the former. Such dialetti as Piedmontese and Venetan are defined by Muljačić (1997a, 1997b; 2011 [2000]) as middle languages, for they are High with respect to local varieties and Low with respect to the official standard language, and they display autonomy and heteronomy as well (see also the concept of Aubaudialekt/Kulturdialekt in Kloss 1978: 55–60). At first glance, this kind of repertoire resembles what has been termed double overlapping diglossia by Fasold (1984: 44–46), the latter implying, nevertheless, the existence of different communities of speakers within the same territory.

Piedmontese has been a dialetto with a certain degree of codification since the nineteenth century, its Turinese-based koine having been described in grammars and dictionaries; thus, it would not be wrong to maintain that Piedmontese had a reference variety of its own (an indication of autonomy), though the touchstone for its codification was offered by an external source (an indication of heteronomy). As for the possible external sources, it is worth recalling that two prestigious standard languages were traditionally present in Piedmont, Italian and French (see Regis 2013; Cerruti and Regis 2014: 94–97). In 1560 and 1561, Emmanuel Philibert issued two edicts, stating that Italian and French were to replace Latin in all of the administrative documents of the Duchy of Savoy: Italian in the cisalpine territories (except the Aosta Valley) and French in the transalpine areas. Even long after Italian became the official language in Piedmont, French was still the favored language of the aristocracy, which would converse alternating Piedmontese and French. Although this situation remained unchanged throughout the eighteenth century and for much of the following century (see Marazzini 1984), Italian was nevertheless selected as the model language for the codification of Piedmontese. ^[3] It is perhaps useful to remember that, out of five dictionaries published between 1783 and 1859, two were Piedmontese-Italian dictionaries (Ponza 1830; Sant’Albino 1859), two were Piedmontese-Italian-French-Latin dictionaries (Pipino 1783a; Zalli 1815), and only one was a Piedmontese-French dictionary (Capello 1814: it is worth underlining, however, that Capello’s book was printed during the Napoleonic annexation). We may wonder why Italian, instead of French, was assigned the delicate role of “guide language”, i. e., a language of cultural reference, for Piedmontese (see the notion of lingua guida in Pellegrini 1975 [1973]: 56–57). First of all, French was associated with the recurring political and military menace of France, whereas Italian was an institutional (and thus neutral) language, established as such by the duke of Savoy. Moreover, as Joseph (1985: 39) maintains, “the basic source of prestige is scarsity ” (bold and italics in the original); thus, it follows that Italian was probably reputed more prestigious than French, the former being far less common than the latter. All of these factors helped to promote Italian as the main reference for the codification of Piedmontese and also explain why many phonological, lexical, and morphological borrowings from Italian entered Piedmontese dictionaries in the 1800s, in spite of the limited diffusion of the source language. Yet at the same time, it would be inaccurate to argue that Piedmontese of past centuries was completely unaffected by French; for example, as suggested by Clivio (1976: 104), it is very likely that French helped Piedmontese contain the superposition of Italian: “as long as French remained a commonly used language of culture in Piedmont […], Piedmontese was able to retain its Western Romance character much better than after French died out and only Italian was left as a cultural language”. Clivio is here relying upon the classical (though disputed) division between Western Romance and Eastern Romance languages, which dates back to the early days of Romance linguistics (see Diez 1836: 3 and later on; Meyer-Lübke 1909: 17), his crucial statement being that after the decline of French, “Piedmontese was left entirely open to the Eastern Romance influence of Italian” (Clivio 1976: 105). Phonetically speaking, such “genuine” Piedmontese words as vel ‘calf’, [vɛl] (Lat. < vitĕllu), sercc ‘circle’, [sɛrʧ] (< Lat. cĭrcŭlu), and dman ‘tomorrow’, [dmaŋ] (< Lat. de mane ‘in the morning’) are closer to French (veau, [vo], cercle, [sɛːʀkl], and demain, [dǝˈmɛ̃], respectively) than to Italian (vitello, [vi'tɛlːo], cérchio, ['ʧerkjo], and domani, [do'maːni], respectively); at any rate, for each of these words, it is easy to find an “Italianized” correspondent: vitel, [vi'tel] (Capello 1814), cercc, [ʧerʧ] (Pipino 1783a), and doman, [du'maŋ] (Zalli 1815). While it is undeniable that the use of the “Italianized” variants became more and more widespread during the twentieth century (on the general topic of the Italianization of the dialetti, see e. g., Sobrero 1997; Scivoletto 2014; Cerruti 2016), it is worth underlining that they were first attested during a period in which French was still a cultural language in Piedmont, thereby reaffirming the importance of Italian in the codification of the Piedmontese koine, especially with regard to phonology and lexicon.

As of the eighteenth century, the overall relationship between Italian, French, the Piedmontese koine and local varieties of Piedmontese may be roughly depicted as follows. In the eighteenth and nineteenth centuries, two high languages were still coexisting in Piedmont, Italian and French. The Piedmontese koine mainly chose Italian as its guide language, while local varieties of Piedmontese oriented towards the regional koine. After the Unification of Italy (1861), French gradually disappeared from the repertoire, whereas Italian started to conquer new domains of usage, seeking to become what it had never been before: a language of common usage for the Italian people. In the first part of the twentieth century, the Piedmontese koine would continue to converge towards Italian, and the local varieties of Piedmontese would continue to look to the Piedmontese koine as their reference variety. Though the functional separation between Italian and Piedmontese remained rather strong in the early 1900s, the demotization (i. e., popularization) of the standard language was starting to take place. This situation, however, did not undergo significant modifications until the second half of the twentieth century, when Italian began to be used by an increasing portion of the population. The spreading of Italian as a medium for everyday communication had at least two effects. As a first consequence, the Piedmontese koine lost its traditional function as a “lingua franca” among speakers of different varieties of Piedmontese, as well as its mediating role between Italian and local varieties of Piedmontese. Second, the local varieties of Piedmontese began to draw directly on (regional) Italian. This overall process is still going on.

3 Partitive determiners according to the grammars of the Piedmontese koine

The definition of Piedmontese as a rather codified language requires some clarification. As is often the case with Italo-Romance dialetti, dictionaries of Piedmontese abound, while its grammars are very limited in number. ^[4] Moreover, we have to consider that dictionaries and grammars have been centered for centuries on the Turinese-based koine, with the regular codification of local varieties of Piedmontese taking place only as of the mid-twentieth century. It is important to clarify that this recent “microcodification” policy is rather different in scope from the “classic” codification of the Piedmontese koine, for whereas the former aims at documenting (and, to a certain degree, safeguarding) previously discarded varieties, the latter was traditionally used as a reference model for Piedmontese speakers/writers, regardless of their place of origin. All this considered, we will hereinafter examine how the subject of partitive determiners has been tackled (or not) by the grammars of the koine, a list of which appears in Table 1. The “+” and “–” signs indicate whether or not a section on partitive determiners is present in each grammar.

Pipino (1783b: 15–16) deals with both definite and indefinite determiners, while dismissing partitive determiners altogether (see, however, Section 4 for some examples of partitive determiners contemporary with Pipino’s grammar). Partitive determiners do not fit into Clivio (1964); they are also lacking in the earlier editions of Brero’s grammar (Brero 1967, Brero 1971 [1967]), being instead documented from the fourth edition onwards (Brero 1975 [1967]; see below). While the grammars displaying a “–” sign for the subject under scrutiny prove to be a minority, the “+” sign does not necessarily guarantee a thorough analysis of partitive determiners. Under the heading of prepositions, Aly-Belfàdel (1933: 278) deals with both partitive complements and partitive determiners (see examples below). The grammarian’s attention is thus focused on the different uses of the preposition dë, ^[5] [də] ‘of’, as becomes immediately clear from his introductory lines: “la preposizione dë è frequentemente usata nel partitivo, che in piemontese si usa assai, alla moda francese o non troppo diversamente” [‘The preposition dë is frequentely used in partitive constructions, which in Piedmontese are very common, following to a certain extent the French model’]. Aly-Belfàdel offers various examples of partitive complements (many of which, however, are pseudo-partitives, e. g., mutubèn d’üva ‘a lot of grapes’, bin ’d fracas ‘a lot of noise,’ and ëm poch ’d pan ‘a bit of bread’), whereas the partitive determiners are restricted to two cases: pié d’infurmasiun ‘to gather some information, lit. of information’ and dë dné ‘some money, lit. of money’. More accurately, Brero 1975 [1967]: 26) tackles partitive determiners in the paragraph on indefinite determiners, defining them as “la forma indefinìa plural ëd l’articol” [‘the plural indefinite form of the article’]; the example supplied by Brero includes two partitive determiners, ëd fiuej ‘some boys, lit. of boys’ and ’d fije ‘some girls, lit. of girls’: A-i son ëd fiuej che a canto e ’d fije ch’a balo ‘there are some boys who sing, and some girls who dance’ ^[6]. Both Griva (1980: 32) and Grosso (2000: 53) follow Brero 1975 [1967]), addressing partitive determiners along with indefinite determiners. Griva’s formulation and original examples are as follows: “Al plurale [per i determinanti indefiniti], così al maschile come al femminile, si usano le forme del partitivo: ëd, ’d, dë, d’ (vëdend ëd piante; l’hai ’d sòld; i serno dë strass; i mangio d’àmpole)” [‘For both the masculine and feminine plural [of indefinite determiners], the partitive forms used are ëd, ’d, dë, and d’ (vëdend ëd piante ‘seeing plants, lit. of plants’; l’hai ’d sòld ‘I’ve got some money, lit. of money’; i serno dë strass ‘I choose some rags, lit. of rags’; i mangio d’àmpole ‘I eat some raspberries, lit. of raspberries’)’]. Not unlike Griva and Grosso, Brero and Bertodatti (1988: 27) tackle partitive determiners in the same paragraph as indefinite determiners; with respect to Griva and Grosso, however, they add an important note: “In piemontese esiste la forma partitiva […] che usa sempre la preposizione «ëd (’d), dë (d’)», mentre spesso in italiano è sottaciuta” (our emphasis on the word sempre) [‘In Piedmontese the partitive form always requires the preposition ëd (’d) or dë (d’), whereas in Italian it is often omitted’]. We are thus informed that: 1) the use of ëd and its variants are mandatory in Piedmontese partitive constructions; and 2) in this respect, Piedmontese behaves differently from Italian, a language in which mass and plural countable nouns can occur as bare forms. Yet, Brero and Bertodatti’s statement is not completely accurate, since Italian allows the omission of partitive determiners only with mass nouns and non-specific countable nouns: C’è latte ‘there is milk’ and C’è del latte ‘there is some milk, lit. of the milk’ are both permitted (Renzi 2001: 392), as well as Mi ha regalato rose ‘He has given me roses’ and Mi ha regalato delle rose ‘He has given me some roses, lit. of the roses’ (Renzi 2001: 390). The partitive determiner is instead obligatory with specific countable nouns: I bambini hanno ritrovato dei cani che avevamo perduto ‘The children have found some dogs, lit. of the dogs, we had lost’) (Bonato 2004: 176). Villata (2009 [1997]: 22) insists on the fact that in the Piedmontese koine partitive determiners are mandatory, claiming, in addition, that this property is shared with French and conveniently highlighting the possible existence of geographical variation in the use of partitive determiners: “Vorremmo qui ricordare che in torinese il partitivo è costituito dalla sola preposizione semplice (ed /de), mentre in altre varietà di piemontese si preferisce usare la preposizione articolata (dla/dij)” [“Let us here recall that in Turinese the partitive is expressed by a simple preposition (ed/de ‘of’), whereas in other varieties of Piedmontese the articled preposition is preferred (dla ‘of the’ [feminine singular]/dij ‘of the’ [masculine plural]’]. In other words, the behavior of the Turinese-based koine does not reflect the behavior of all varieties of Piedmontese. This is an often overlooked matter concerning both nonstandard and standard languages, whereby one is inclined to consider such entities as Piedmontese, Lombard, or even Italian as perfectly homogeneous, when in fact they are far from it. ^[7]

The scattered images provided by the grammars of the Piedmontese koine may be summarized as follows:

1. partitive determiners are expressed by a simple preposition (i. e., an unarticled form), ëd, [əd] (’d), dë, [də] (d’), depending on the phonetic environment;

2. they are mandatory for both mass and countable nouns (see dë dné ‘some money’ and d’àmpole ‘some raspberries,’ respectively);

3. a. as for their being compulsory, they behave similarly to those of French and

4. b. differently from those of Italian, which, at least to some degree, can be omitted;

5. they can behave differently from those of local varieties of Piedmontese.

The table below indicates the overall paradigm of indefinite determiners in the Piedmontese koine. It includes the singular masculine and feminine forms for countable nouns, and the partitives are marked in bold.

It is striking that none of the grammarians above has underlined the peculiarity of the unarticled form as a unique partitive determiner, setting the Piedmontese koine apart from both Italian and French. Following Bossong’s (2016: 69) classification of partitive determiners, Italian and French basically belong to type (4), since “[t]he preposition de […] is combined with the definite article and forms with it an indissoluble unit”, while Piedmontese, along with most varieties of Occitan (Ronjat 1937: 128) and southern varieties of Francoprovençal (Kristol 2014, Kristol 2016: 358–359), can be assigned to type (3): “[t]he preposition de […] does not combine with the definitive [sic] article but remains unchanged.” At any rate, the unarticled form is not unknown to Old Italian (Migliorini 1994 [1960]: 210; Rohlfs 1968: 117; Renzi 2010: 346–347 and is well attested in formal styles of contemporary French (Grevisse and Goosse 2008: 747–748; see also Carlier and Melis 2006: 456), though in both cases it is limited to very restricted environments, namely before “attributive adjective + noun” phrases. As for Italian, further examples of this usage appeared during the Renaissance and between the eighteenth and nineteenth centuries (see again Migliorini 1994 [1960]: 636). Migliorini’s hypothesis that Italian modelled this construction on its French counterpart does not apply to all of the occurrences above, but rather only to the most recent ones. It must in fact be considered that, until the seventeenth century, French allowed both unarticled and articled forms before “attributive adjective + noun” phrases (Grevisse and Goosse 2008: 747); therefore, it would be rather peculiar for a language (Italian) to select an option (the unarticled form) from another language (French) which is not categorically employing it. Moreover, the unarticled form was widespread in Old Lombard, also with non-modified nouns, e. g., no ge lasón de roba in dosso ‘they do not leave him any clothes on’ (Pietro da Barsegapé, thirteenth century; Rohlfs 1968: 117).

4 Partitive determiners in Piedmontese:A diachronic perspective

In what follows, a brief diachronic account will be provided of the development of partitive determiners in Piedmontese, ranging from the twelfth and thirteenth centuries to the present. Such an overview draws upon the results of recent studies on this subject (see especially Bonato 2003-2004: 113–145 and 2004; cf. Ricca 2008; Miola 2017). ^[8]

The first occurrences of partitive determiners can be found in a collection of sermons, the so-called Sermoni subalpini, which dates back to the late twelfth to early thirteenth centuries. In these texts, mass nouns are sometimes marked by a partitive form derived from the Latin preposition de ‘of, from’, as is the case with d’ in utterance (1). On the contrary, plural indefinites always occur as bare nouns. However, it is worth noting that the Romance vernacular in which these sermons are written actually consists of both Italo-Romance and Gallo-Romance features ^[9]; there is indeed some question as to whether or not it can be considered as Piedmontese (see e. g., Clivio 2002: 21–30).

At any rate, it was not until the sixteenth century that partitive elements were regularly used as both singular and plural indefinite determiners. Moreover, in the sixteenth century indefinite determiners mostly displayed an articled form, deriving from the agglutination of a prepositional element (of the same type as the partitive marker d’ in Example (1)) and a definite article; a significant example is d’i, which is found in Example (2). However, both mass nouns and countable plurals from this period can be seen to occur as bare forms. In particular, partitive determiners often alternate with zero marking when used with non-specific plural indefinite nouns, and the determiner position is more frequently filled when reference is made to specific plural indefinite entities, as in (2).

As pointed out by Bonato (2004: 181–182), the use of partitive determiners increased considerably in the seventeenth century and became almost compulsory in the eighteenth century. However, the presence of bare nouns is still significant in the case of plural indefinite referents; for instance, in (3), zero marking occurs with a plural indefinite noun requiring a non-specific interpretation. In terms of form, unarticled determiners can still be found to alternate with articled ones, but the former (see e. g., ’d in sentence (4)) tend to be preferred.

The most radical changes appear to have taken place in the nineteenth century. In fact, partitive determiners became mandatory with both mass nouns and countable plurals, and the unarticled form of the partitive marker came to predominate over the articled forms, with the latter hardly being found to occur. A relevant example is found in (5):

These latter uses can be argued to reflect the influence of the regional koine on the texts in question. Such an influence became, in fact, particularly evident in the late eighteenth and nineteenth centuries, when Turinese established itself as the reference variety for the codification of Piedmontese (cf. Section 2). In particular, as regards partitive determiners, reference grammars such as Brero and Bertodatti (1988) consider the Turinese use of unarticled forms (ëd or dë) as mandatory with both mass nouns and countable plurals (cf. Section 3), and this is indeed the form which stood out in nineteenth-century written texts.

More broadly, the diachronic overview which has been sketched in this section does not portray the development of partitive determiners in a single variety of Piedmontese. Leaving aside the so-called Sermoni subalpini, which are subject to the controversial question mentioned above, it is important to point out that there is no evidence about Turinese usage until the seventeenth century. In fact, the texts of the previous centuries only describe the linguistic behavior of small urban centers and villages; it is significant that sentence (2), which contains an articled form, comes from a poem written by an Asti-born writer (Villata 2008). The differences identified between the texts prior to the seventeenth century and the texts of the following centuries can thus reflect differences in the paradigm of partitive determiners between different varieties of Piedmontese.

In this respect, it should be emphasized that we are dealing with written texts of an Italo-Romance dialetto. Texts of this kind, which are mostly the product of intellectual élites or (as of the twentieth century) activist groups, are typically learned and “refined”; in utterance (4), for example, we find the relative pronoun i quai (clearly corresponding to i quali in Italian), whose usage is mainly restricted to literary style (see e. g., Parry 2007). Most of these texts, by their very nature, can hence be deemed to some extent artificial. This may be particularly the case with those texts written from the late eighteenth century onwards by non-Turinese poets and novelists. Such texts, especially those from the nineteenth century, generally conform to the description of Piedmontese provided by grammars and dictionaries, and therefore, as argued, to the variety of Turin. However, it can be maintained that the codification of Piedmontese has exerted a greater influence on written texts than on everyday speech; in particular, the everyday speech of small urban centers and villages has presumably always differed to some extent from written texts and has been (and still is) more influenced by Turinese through the spontaneous diffusion of linguistic features (cf. Cerruti and Regis 2014: 94–95) than as a result of codification.

Finally, as far as partitive determiners are concerned, the twentieth century witnessed a significant presence of bare nouns and articled forms among non-Turinese poets and novelists, as in e. g., (6). This is consistent with the fact that by the second half of the twentieth century the influence of Turinese as a reference variety had weakened (cf. Section 2). In fact, as written texts ceased to conform strictly to the codified norm, linguistic features such as bare indefinite nouns and articled forms of indefinite determiners, which had always been present in local varieties of Piedmontese, could reemerge in writing.

Table 3 ^[10] summarizes the development of partitive determiners in Piedmontese from the sixteenth century to the twentieth century.

5 Partitive determiners in contemporary Piedmontese

A look into the behavior of partitive determiners in contemporary speech can provide us with a more comprehensive picture of the situation. To begin with, it can be argued that the use of indefinite determiners in spoken Turinese closely reflects the paradigm described by reference grammars of the Piedmontese koine (cf. Section 3). In fact, spoken Turinese displays an indefinite determiner for single bounded entities (un; e. g., un lìber ‘a book’) and an unarticled form of the partitive determiner for both mass nouns and countable plurals (ëd or dë, depending on the phonetic environment in which the form occurs). Evidence of this can be found in the relevant data provided by ALEPO, ^[11] e. g., (7), (8), and (9). A further example is (10), taken from the ASIS survey (see also AIS 1343).

However, contrary to the behavior described in reference grammars, the partitive determiner ëd/dë does not appear to be obligatory. ^[12] In fact, consistent with Bonato’s (2003–2004: 134–135; Bonato 2004: 185) results, spontaneous speech data reveal that countable plurals can sometimes occur as bare forms when referring to non-specific indefinite entities, as in utterance (11):

The absence of a partitive determiner is clearly conditioned by a complex array of factors, of which we can make just brief mention here. For instance, when commenting on the behavior of partitive forms in Piedmontese, Berizzi and Zanini (2011: 39–41) suggest that the absence of determiners can be favored in clauses higher in transitivity (consistent with the provisions of Hopper and Thompson’s 1980 transitivity continuum). From this perspective, it may be no coincidence that a partitive determiner occurs in an utterance such as (10), in which the predicate is non-actional, non-punctual, negative, and atelic as well (given that negated events are not completed), thereby reducing the degree of transitivity of the clause (cf. Hopper and Thompson 1980: 252). Moreover, the presence of a partitive determiner is obviously unaffected by the absence of inflectional morphology on the related noun, as the unarticled form of the determiner does not convey number marking.

In short, the paradigm of indefinite determiners in contemporary spoken Turinese can be broken down as follows ^[13]:

Furthermore, the geographic differentiation observed in earlier stages of development (cf. Section 4) is found in contemporary speech as well. In fact, some varieties of Piedmontese employ articled forms of partitive determiners for both mass nouns and countable plurals (see e. g., (12) and (13)), and such forms can alternate with zero marking (see e. g., (14) and (15)). As mentioned above, certain factors can be crucial in allowing the determiner to be omitted. It would seem, for instance, that partitive determiners tend to alternate more often with zero marking when reference is made to non-specific indefinite entities, as in (14) and (15). In fact, it has been argued in particular that the articled forms of plural partitive determiners are inclined to express specificity (Stark 2007: 52), while sharing the same ambiguity between a wide-scope and a narrow-scope reading as singular indefinites (cf. Squartini 2017: 12–13).

The presence of partitive determiners, or the lack thereof, can nevertheless result from different conflicting motivations. For example, in (13), where a non-specific plural occurs as well, the presence of the articled form [di] can be favored by the absence of inflectional markers of plurality on the related noun. ^[14] More broadly, the ALEPO data (see fn. 11) reveals that in all varieties displaying articled forms, the determiner position is consistently filled when an indefinite noun does not bear any (overt) inflectional marker of number, as in (12) and (13). This seems to corroborate the existence of a correlation between the mandatory presence of determiners and the unavailability of markers of number on nouns; the absence of inflectional endings could indeed be offset by the presence of a partitive determiner, which overtly marks number opposition (cf. Stark 2007: 52; Squartini 2017: 2, 7–8).

Such behavior seems to be common to most peripheral varieties of Piedmontese; in fact, it is found among speakers from the provinces of Biella and Alessandria, as well as among those from Canavese, Montferrat, and Langhe, to name but a few. This is apparent not only from questionnaire surveys, such as those carried out by AIS (see e. g., AIS 1343), ALEPO (see fn. 11 above), and ASIS (cf. Berizzi and Zanini 2011), but also from spontaneous speech data (see e. g., Bonato 2004: 184–187; Parry 2005: 142–144; cf. Ricca 2008: 122–124; Miola 2017: 154–157). In sum, the paradigm of indefinite determiners in most peripheral varieties of Piedmontese is similar to that mapped out in Table 5 (including the singular indefinite article); for the sake of example, all forms in the table refer to the variety of Campiglia Cervo (a village located about fourteen kilometers northwest of Biella), which is the source of utterances (12), (13), (14), and (15).

However, there is yet a third type of behavior which can be gleaned from ALEPO data. In fact, unarticled and articled forms of partitive determiners appear to coexist in some varieties of Piedmontese (see also Bonato 2004: 185; Ricca 2008: 123). Such is the case, in particular, with some of the so-called “rustic” and peri-Turinese varieties, that is, some of those varieties spoken in the plains situated northwest and southwest of Turin and in the hilly areas around the capital (cf. Telmon 2001: 70–72). Below are some examples collected in Bibiana:

The coexistence of unarticled and articled forms of partitive determiners in these varieties can be easily explained by contact, first and foremost with Turinese. The relationship between the variety of Turin and the surrounding varieties of small urban centers and villages reflects situations of “geographical diffusion” (Kerswill 2003: 223–224), in which linguistic features spread from a “high” variety to lower varieties (cf. Cerruti and Regis 2014: 95). When these features do not replace the original traits of lower varieties, the coexistence of features can result in intra-speaker variability, i. e., style variation. This phenomenon can also be observed in the area under scrutiny, on account of the functional differences between Turinese and non-Turinese varieties (cf. Section 2); indeed, cases of shifting between the former and the latter are found to occur mainly in response to the closeness, prestige, or provenance of the interlocutor, and, therefore, to the degree of formality of the situation (see e. g., Cerruti 2006: 230–237; Parry 1997: 237). Below is the case of a single speaker shifting from a “rustic” variety of Piedmontese (18a) to Turinese (18b) according to a change of addressee; it is worth noting that such a shifting involves the alternation between the articled form (18a) and the unarticled form (18b) of the partitive determiner.

Language-internal motivations can thus interplay with external and extra-linguistic factors in conditioning the alternation between unarticled and articled forms of partitive determiners. This may even be the case with peripheral varieties, for Turinese is ‒ or at least has been (cf. Section 2) ‒ superposed on all other varieties.

However, the alternation between unarticled and articled forms of partitive determiners is not always due to the contact with Turinese. In fact, both forms can also be found in those varieties of Piedmontese which are intensively in contact with Gallo-Romance, i. e., Francoprovençal and Occitan, varieties (some of which employ unarticled partitives as well; see e. g., ALEPO Q1323; cf. AIS 1343). Moreover, at least in some of these Piedmontese varieties, the distribution of unarticled and articled forms of partitive determiners appears to be related to the presence of inflectional markers of number on the related noun (or the lack thereof). The variety of Val della Torre, for instance, displays a distribution of this kind in the subset of feminine nouns; in fact, unarticled partitives are regularly found when indefinite nouns hold inflectional markers of number (e. g., [əd 'rɔba], [əd 'matʃe], in (19)), while articled forms consistently occur when such markers are lacking (e. g., [dla karn] and [del nus] in (19)).

At this point, a comparison can be made between the paradigm of indefinite determiners in spoken Turinese and that of peripheral varieties of Piedmontese. As regards the relationship between mass nouns and countable singulars, both paradigms oppose a partitive determiner to an indefinite article (cf. Tables 4 and 5), thus codifying a classificational distinction between unbounded entities (“substances”, e. g., some milk) and single bounded units (e. g., a book). However, whereas both singular indefinite determiners are obligatory in Turinese (e. g., [əd 'roba], un lìber), mass nouns are allowed to be bare in peripheral varieties (e. g., Ø ['roba] in utterance 14; except when they lack an inflectional marker of number, at least in some varieties; see e. g., [dal paŋ] in 12).

The main paradigmatic differences concern the relationship between mass nouns and countable plurals (cf. Tables 4 and 5). In peripheral varieties, the plural is conveyed at least by the articled form of the partitive determiner, which is found not to be omitted when the noun lacks inflectional markers of number (e. g., [di tur'nant] in 13). In Turinese, on the contrary, the plural is conveyed by the sole presence of inflectional endings on nouns, ^[15] as the unarticled form of the determiner does not express number marking (e. g., [d 'nivule] in 8). If inflectional endings are dismissed, then the distinction between mass nouns and countable plurals is guaranteed in only a few cases, namely, when partitive determiners (which are obligatorily required with mass interpretation) are opposed to bare plurals (which are found in some cases of non-specific indefinite referents). When such conditions are not met, the distinction between mass nouns and countable plurals is neutralized, for both mass nouns and countable plurals have a morphosyntactic structure consisting of an unarticled determiner plus an invariable noun (e. g., [əd kun'tʃim] in 9, ed liber in 10).

In this respect, as argued by Squartini (2017), two different ways of categorizing the relationship between mass nouns and countable plurals hence emerge. One way is to have quantification prevail over classification. The paradigm of peripheral varieties of Piedmontese appears to be consistently based on such a principle, as the classificational distinction between unbounded entities (“substances”, e. g., some milk) and unbounded sets of separable individuals (“aggregates”, e. g., some books; cf. Jackendoff 1991) ^[16] is subsumed under number opposition. The other way foregrounds classification, and the paradigm of spoken Turinese meets such characteristics under certain circumstances. In fact, when nouns dismiss inflectional endings, the use of an unarticled determiner with both mass nouns and countable plurals equates “substances” and “aggregates”, and therefore functions as a marker of classification.

Finally, it is quite clear that the paradigm of indefinite determiners in peripheral varieties of Piedmontese matches that of Italian (the latter is depicted, for instance, in Renzi 2001: 390–392), as shown in Table 6. In actual fact, the paradigmatic organization of indefinite determiners in these varieties may have been influenced by Italian, especially in the twentieth century, when the standard language extended its reach to encompass ordinary conversation, thereby weakening the influence of Turinese as a reference variety (cf. Section 2). However, both zero marking and articled forms of partitive determiners are inherent features of peripheral varieties of Piedmontese, and there are no compelling arguments to contend that the related paradigm has changed considerably over time, at least in everyday speech (cf. Section 4). All in all, the possible influence of Italian can be understood in terms of “contact-induced stability” (cf. Kühl and Braunmüller 2014: 30–31); in other words, changes may not have taken place because of inter-systemically equivalent elements, which are cost-saving in language processing and hence usually sought by bilingual speakers (see e. g., Matras 2009: 151, 235).

On the contrary, contact with Italian may have brought about changes in the paradigm of indefinite determiners of spoken Turinese. From this perspective, the availability of bare plurals with non-specific reading could be considered as an instance of convergence towards the standard language (as non-specific plural indefinites are allowed to be bare in Italian: cf. Section 3), albeit conditioned by language-internal factors (see above). Partitive determiners were indeed obligatory in earlier stages of Turinese (cf. Section 4). Nevertheless, as is the case with other varieties of Piedmontese, everyday speech may have always differed to some extent from written texts; therefore, the occurrence of actual changes cannot be taken for granted. Along these lines, it cannot be ruled out that the obligatoriness of partitive determiners has always been limited to written texts, on account of strict adherence to the codified norm.

6 Conclusions

At first sight, partitive determiners of the Piedmontese koine, i. e., the variety of Piedmontese traditionally codified by grammars, seem to behave similarly to those of French (cf. Section 3). Nevertheless, this affinity has to do with the compulsoriness of partitive determiners rather than with their formal realization, the Piedmontese koine allowing only unarticled forms (as opposed to the articled forms of French). Contemporary spoken Turinese displays a paradigm of partitive determiners certainly not far from that of the Piedmontese koine, yet permitting their omission with countable nouns (cf. Table 4). A different behavior occurs among peripheral varieties of Piedmontese: while some of them alternate articled forms and bare nouns (cf. Table 5), presumably depending on conditioning factors, others are characterized by a sharper functional distribution of the two (see the case of Val della Torre).

Combining historical and geographical data, we would have expected (i) the Piedmontese koine and Turinese to behave similarly to Italian, towards which both varieties are traditionally oriented, and (ii) western varieties of Piedmontese to behave similarly to their neighboring Gallo-Romance varieties (namely Occitan and Francoprovençal). However, these working hypotheses have been disconfirmed, even overturned, by our data. In fact, not only do the Piedmontese koine and Turinese diverge from Italian, but they also show a paradigm of partitive determiners which closely resembles that of the Gallo-Romance varieties of western Piedmont, while the paradigm of peripheral varieties of Piedmontese matches that of Italian. How can these puzzling behaviors be explained?

We have seen that the Piedmontese koine chose Italian as its “guide language” during the codification process. While Italian was, and still is, a touchstone for both phonetics/phonology and lexicon, the morphosyntactic level has been less affected by the official language, the codification process also contributing to preserve structural peculiarities – this is demonstrated by such Piedmontese features as the double-filled complementizer, postverbal negation, or the subject clitic pronouns, which are shared with other northern Italo-Romance dialetti but are unknown to standard Italian. The case of partitive determiners seems to fit well within this pattern, since the paradigm of the Piedmontese koine sharply contrasts with that of Italian. However, if we were to focus on the behavior of peripheral varieties and spoken Turinese, we would argue that the role played by the national language is not negligible. Italian has probably reinforced a tendency which had been attested ab antiquo in peripheral varieties, i. e., the use of articled forms and bare nouns; moreover, we cannot exclude the possibility that the omission of partitive determiners in spoken Turinese, though internally motivated, has been fostered by Italian.

It is clear that, when we take into account varieties of Piedmontese different from the koine, we are faced with a certain amount of variation. As is always the case with Italo-Romance dialetti, the label Piedmontese is a shortcut for a multi-faceted reality; it does not refer to a homogeneous linguistic entity, but rather to a group of varieties belonging to the same diasystem and behaving in a slightly different way from one another. The two main paradigms we have found are linked – at least partially – to different ways of categorizing the opposition between mass nouns and countable nouns, and, of course, pertain to geographical varieties of Piedmontese, i. e., Turinese and peripheral varieties. Yet we have to underline that features of either paradigm can occur within a given variety of Piedmontese as a byproduct of style (as well as interindividual) variation.

More generally, partitive determiners may become an important tessera in the much-debated collocation of Piedmontese (as well as of Gallo-Italian) within or without the Italo-Romance domain (for an overview, see Tamburelli and Brasca 2018: 442–444). The idea that Piedmontese is a sort of transitional variety between Italian and French, already expressed by Denina (1804: II, 61), has been then embraced by Tagliavini (1952: 314) and Vidos (1956: 292) and recently restated by Bossong (2016: 68), who maintains that “Piedmontese (and the other varieties of Gallo-Italian) […] forms a bridge between Gallo- and Italo-Romance” and that “[t]his alone would be sufficient to justify classifying Gallo-Italian as a language in its own right (‘language’ in the sense of ‘dialect group’ or ‘diasystem’)”. Unfortunately, as long as a comprehensive description of Piedmontese is lacking, this classificatory conundrum is doomed to remain unresolved.