Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Quantitative Approaches to Italian / Dialect Family Interactions: Considerations of Methodology and Language Transmission in a Contact Situation

Quantitative Approaches to Italian / Dialect Family Interactions: Considerations of Methodology... 1.IntroductionPsycholinguistic and sociolinguistic studies conducted in the Italoromance area have mostly taken two distinct directions. On the one hand, psycholinguistic studies have focused on Italian acquisition. These studies have investigated the acquisition of Italian vocabulary (Bates, Burani, D’Amico, and Barca, 2001; D’Odorico, Carubbi, Salerni, and Calvo, 2001) of Italian morphology (Pizzuto and Caselli, 1992; Noccetti, De Marco, Tonelli and Dressler, 2007) and of Italian syntax (Waxman and Guasti (2009) and Guasti (2009) for production and Arosio, Adani and Guasti (2009) for comprehension). Psycholinguistic research conducted in this area has generally adopted a quantitative approach. On the other hand, sociolinguistic studies have described the language contact phenomena observed in the multilingual social repertoires (Berruto, 1990; Cerruti, 2003; Regis, 2002). There have been few attempts to integrate psycholinguistic and sociolinguistic approaches when studying language development in the Italoromance area. For example, investigations on language acquisition have seldom included variation phenomena inherent in the social repertoires. Ignoring the elements of variation in the environment within which an Italian child constructs his/her language omits important information regarding a child’s language exposure. In fact, studying the variation phenomena (from phonological, lexical and morphosyntactic levels) present in the child’s language environment may help clarify the processes a child goes through when s/he acquires a language within spontaneous and informal contexts of interaction.We have adopted a usage-based approach as it has the advantage of being sufficiently flexible to integrate sociolinguistic variation generated by the Italian/dialect language contact with the study of child language. More generally, integrating methods and approaches from sociolinguistics and psycholinguistics could be particularly useful when analyzing acquisition data collected in contact situations, as in Italy. On the one hand, psycholinguistic studies could benefit from sociolinguistic approaches to define and understand the sources of variation in the child’s language environment. On the other hand, sociolinguistic approaches could benefit from psycholinguistic approaches in order to understand the cognitive foundations underpinning variation in child production.One of the main tenets of the usage-based account is that the child builds his/her language(s) from concrete pieces of language encountered in his/her environment (Behrens, 2009; Tomasello, 2003). Through linguistic exposure and experience, the child starts building constructions, the latter being “form and meaning pairings” (Goldberg, 2006, p. 3) from these concrete pieces of language. At the two-word stage (from approximately 18 – 24 months), the child will start forming his/her first constructions by combining two lexical items. Although these early constructions – called pivot schemas (Braine, 1976) – are not manifestations of the emergence of syntax, they nevertheless show the initial stages of categorisation and the expression of a child’s communicative intentions (Tomasello, 2003). These constructions are made up of two slots: one is filled by a pivot, which is the structuring unit of the construction in that it places it within a speech act (for example More in “More X” places this pivot schema within a request) and the other is a variable slot filled by an element pertaining to the immediate communicative setting (for instance cookie in more cookie, for further examples see Braine 1976 or Tomasello 2003, pp. 94-143). Once these constructions emerge, the child gradually starts making generalizations across these pivot-schemas, building more complex constructions. This process of generalization is an important step towards the production of abstract linguistic constructions (Tomasello, 2006) and, more specifically, to the development of syntax.The present study observes the production of Francesco, a child aged between 17 and 30 months, participating in dyadic and multiparty interactions with five members of his nuclear and extended family who live in Castelfranco Veneto, a city in the central part of Veneto, one of Italy’s northeastern regions. Here, Veneto dialect and Italian are spoken on a daily basis in both family and public spheres. The speakers taking part in the interactions belong to three generations. In this study, we observed language development and environment jointly. Our investigation was guided by three questions:(1)What languages are spoken in the child’s environment?(2)How are dialect and Italian used when speaking in dyadic and multiparty interactions, with speakers from different generations?(3)Is language usage in the child’s production related to the language usage observed in his environment?These three questions will guide the analyses of the data in the investigation of each speaker’s production observed in informal dyadic and multiparty interactional settings.2.A Sociolinguistic Approach to Italy’s MultilingualismSociolinguistic studies conducted in Italy have contributed to a better understanding of phenomena generated by the joint presence of Italian and the Italoromance dialects in social repertoires. As Berruto (2005) has pointed out, Italy’s dialetti do not reflect the same entity as the English dialects. According to Coseriu’s (1981) tripartite categorization of what is designated by the term “dialect,” Italoromance dialects figure among the primary dialects, i.e. dialects that have their own autonomous linguistic system whereas English dialects figure among the tertiary dialects because they are considered as varieties within one language system (see Sornicola (2002) for a review of Coseriu’s theory). Grassi (1993: 279) insists that Italoromance dialects – like any language – have a specific phonetic, phonological and morphosyntactic structure, yet they differ in the various functions that are attributed to each language. For instance, Italian is used in the public sphere and is both the medium of communication (in both formal and informal settings) and the object of study in the field of education. Clearly, dialects have a smaller speech community than Italian. Moreover, variation in usage can be observed within the same speech community. As Giacalone-Ramat (1995) notes, adults are the main dialect users in the Italoromance area. Her observation was confirmed by the results of a national survey on declared language usage conducted in Italy and published in 2007 in the Istituto Nazionale di Statistica (ISTAT) site.1 The results show a clear preference for Italian among children and young adults. Veneto’s dialect follows the national usage trends, being predominantly used by the adult population and within the restricted family environment. Despite the trends observed, statistics on declared language usages in Veneto show that the region’s dialect is one of the most vital in Italy. These findings however do not provide clear indications of speakers’ actual usage, and with whom and in which circumstances their code choices are effectuated.2.1.Issues of Variation: A Glimpse of the Processes of Dialect Italianization and Italian RegionalizationThe first investigations on diatopic variation in the Italoromance area were carried out by Rüegg (1956). His observations showed that, despite strong political movements favoring the process of Italianization, there was a considerable amount of variation in the Italoromance domain. Despite the adoption of Italian as the one and only official language, there was considerable interregional variation in daily language usage (Guerini, 2011; Marcato, 1991). Rüegg’s diatopic variation was later theorized by Pellegrini who introduced the concept of Regional Italian, which has subsequently become the object of numerous sociolinguistic studies (Cerruti, 2003, 2011; Cerruti and Regis, 2005; Dal Negro, 2005; Marcato, 1991). These studies underscore the significant amount of intralinguistic variation in the regional realizations of Italian given the influence the Italoromance dialects exert on Italian. Hence an important element contributing to Italian intralinguistic variation is the interlinguistic variation existing between Italian and dialect. The regional dialects “regionalize” Italian production, particularly in its oral form, on all linguistic levels (semantic, phonological, etc.). When analyzing the process of Italian regionalization, Telmon (1990, 1993) compares it to Selinker’s (1972) notion of interlanguage. This notion refers to a particular stage in second language acquisition where the learner’s production reveals the presence of a linguistic system “in progress” which neither belongs to the first language nor to the second language. Telmon shows numerous similarities between the contact phenomena observed in the formation of regional Italians and the phenomena observed in learning a second language. Similarly, observing the initial stages in the acquisition of French as a second language, Véronique (2001) compares these stages to the formation process of the French creoles’ grammatical systems. The fact that there are numerous similarities in different sociolinguistic contexts shows that processes of contact are in many respects generalizable.Since Italy’s unification in the late nineteenth century, and in the last three decades in particular, Italian has steadily gained substantial ground in the repertoires, being the lingua franca for interregional communication as Italy’s national and official language (Sobrero, 1992; Telmon, 1990, 1993; Tempesta, 2001). Moreover, as the language of instruction, the acquisition of Italian is a priority because it is essential for access to education, being the national and most prestigious language. Consequently, Italian dominates the majority of the repertoires and it has exerted substantial influence on regional languages, gaining ground in public (school, administration, etc.) and private (family) spheres. Italy’s unification in 1861 has accelerated the process of Italianization of the Italoromance dialects as the latter were seen as elements of division (Coveri et al. 1998). On the contrary, Italian was perceived as the language that could bring unity in the newly formed republic. Hence, Italy’s unification did not only entail a political union between the regions composing the newly formed country. The relationship between Italian and dialect has moved from diglossia to dilalia (Berruto, 1987a).2 Moreover, Italy’s unification created a zone of contact between the official and prestigious language and the regional and less prestigious languages.As Berruto (1983, 1997) points out, the formation of regional Italians shares numerous characteristics with the process of creolization. Both involve a prestigious variety that comes into contact with a less prestigious variety. In these two situations, the prestigious varieties (the superstrates) are Italian, in the Romance area, and the colonial language, in the case of creole situations. In the Romance area, the less prestigious languages (the substrates) are the Romance dialects and in the creole situation, these are local varieties spoken prior to colonization. When addressing the phenomenon of dialects’ Italianization, Berruto (1997) compares the formation process of hybrid forms – forms containing Italian and dialect morphemes – to the process of relexification within a creole situation. Scholars studying creole language practices have also dedicated much attention to the presence of hybrid forms. Prudent (1980) describes the intense variability in the community in his description of French Caribbean creoles. Prudent’s description gives a clear picture of the intricacy of French and Creole usage:La pratique linguistique de la communauté comprend une utilisation de systèmes hybrides extrêmement difficiles à décrire en l’état (Prudent 1980, pp. 79).Italoromance and creole situations have in common the presence of language varieties that promote the emergence of hybrid forms that draw from elements of both systems in contact.The majority of the Italoromance dialects are used only in their oral form. Certain dialects have also benefitted from a prestigious literary tradition, as is the case for the Veneto dialect which was the national language of the Venetian republic (Cortelazzo, 2001, 2004). As the national and official language, Italian is used in both oral and written contexts and is the most represented language in the media. Standard Italian has been formalized in grammars and exists only in its written form (Berruto, 1987a; Pellegrini, 1960; Sanga, 1981). In its oral form, standard Italian has rarely been documented (Berruto, 1987b; Stehl, 1995). Therefore, Italoromance scholars do not refer to standard Italian when dealing with oral production, because its heterogeneity is not only constrained by intralinguistic variation but also by interlinguistic variation, given the Italian/dialect language contact. The latter creates numerous zones of transfer and their joint usage forms a multidimensional linguistic continuum (Berruto, 1998). In sum, language contact in the Italian context may be seen as the presence of varieties, each of which exerts forces of attraction and repulsion on the other varieties, as in a magnetic field. As in a magnetic field, it is difficult to determine with precision where the pull of attraction and repulsion of one variety begins and the other stops.Given the high variability in language contact situations, it is then difficult to apply Bickerton’s (1975) definition of continuum to Italy’s dialetti and the Italian language contact situation. Berruto (1997) acknowledges the methodological interest in organizing linguistic data collected in the Romance area in an orderly fashion. Yet, a considerable impediment to such an ordering is the difficulty in neatly ranking the intermediate varieties (those along the Italian-dialect continuum, given the extensive interchangeability between Italianized forms of dialect and dialectalized forms of Italian) according to the traits that characterize them (Berruto 1985; 2005).Interestingly, Bickerton’s continuum model has also been criticized in creole studies conducted by Prudent (1980). Like Berruto’s, Prudent’s critique concerns the model’s rigidity, which makes it incompatible with the flexibility inherent in actual language usage. His description of the French Caribbean creole clearly conveys the intense variability in the community’s usages. This led him to introduce the notion of interlect that could be related to the notion of continuum as it has been described in the Italoromance area. In both situations, the grammatical structures of the intermediate varieties existing between two language poles cannot be predicted by the grammatical structure of the languages in contact. In both situations, there are no borders separating the structure of one variety from the structure of the other. Berruto (1985: 71) exemplifies these difficulties through the close scrutiny of language mixing found in recordings made in Bergamo. In his contribution, it is clear that the continuous switching from one language to another renders the task of language assignment particularly difficult, because at times it is not clear whether items are the fruit of a dialectalized Italian or, on the contrary, of an Italianized dialect. In order to circumvent this problem, Berruto (1985: 70) suggests it is best to assign these ambiguous items to the category of the co-occurring items within the utterance/clause level. We will come back to this point, in the methodology section.2.2.Focussing on the Sociolinguistic Characteristics of the Veneto Speech CommunityVeneto is one of Italy’s most vital dialects and shows resilience to the process of Italianization (Coveri et al., 1998; Sobrero, 1992), as its usage conveys strong regional identity (Cortelazzo and Paccagnella 1997; Marcato 2002). In daily social exchanges, dialect and Italian are used jointly, resulting in a profusion of mixed utterances (Trumper & Maddalon, 1982). The numerous descriptions of the dialect varieties (Zamboni, 1974; 2002; Marcato, 1981; Trumper et Vigolo, 1995; Cortelazzo et Paccagnella, 1997 ; Vanelli, 2003; inter alia) show that the Venetian dialect variety is the most prestigious and consequently it strongly influences other dialect varieties (Fray, 1962; Zamboni, 2002). The dialects belonging to the Veneto dialect group form a koiné, that is a common variety that arises from the structural convergence of the dialects in contact. In other terms, it could be viewed as “a spontaneous standardization” (Giovanni Depau, personal communication) that emerges through prolonged contact between dialect varieties, tending towards the variety spoken around Venice. In the case of the Veneto dialect group, the resulting koiné allows intercomprehension between speakers of different dialect varieties. All forms from the Veneto dialect koiné are labelled as dialect in the present research.The Veneto dialects are structurally similar to Italian and daily usage tends to move progressively towards Italian, the prestigious variety in the Italoromance repertoires.In sum, two important points can be drawn from the description of multilingualism in the Italoromance area. Firstly, the processes of language contact are observed in different sociolinguistic contexts. Hence the study of language acquisition in the Italoromance area could address issues pertaining to the construction of multilingual competence, particularly when dealing with issues of transmission and maintenance of minority/threatened languages. Secondly, the existence of regional Italians and Italianized dialects entails the contact between two or more language varieties within the same community. Variation is thus an intrinsic characteristic of the social repertoires and should be included especially when dealing with developmental data, as this variation is an important characteristic of the Italian child’s language environment.Considering these two points, our study aims to gain understanding on the acquisition processes observed in variable environments of informal contexts of interaction. In light of the results obtained, we suggest possible ways in which the dialect and the regional variety of Italian can be transmitted to children. This study also has a more practical aim: to improve methodology pertaining to the study of acquisition in a contact situation, by taking sociolinguistic characteristics into consideration within the protocol of investigation and in the data analyses. The methodological design of this case study takes into account the issues of variation inherent in the language environment observed in the Italoromance area.3.Method: Data collection and codingWe collected data in dyadic and multiparty family interactions. The group of speakers comprised Francesco (17 – 30 months) and five adults (the child’s parents, maternal grandparents and one aunt). Two types of dyadic interactions were taped: mother-to-child and mother-to-father. The multiparty interactions took place between the child and the five adults participating in the recordings. A total of 35 hours of interactions (= 90,000 word tokens) were taped and transcribed by a native Italian and Veneto dialect speaker.For the sake of simplicity, data were transcribed orthographically in an Excel database. However, we did consider phonetic aspects (e.g., the simplification of the geminate: /t:/ > [t]; consonantic reduction /ts/ > [s]; etc.): these were represented phonetically when they occurred in the linguistic data collected. In cases of doubt regarding the morphological segmentation of dialect words/phrases, we used two main reference books: Boerio (1856) and Marcato and Ursini (1998).3 Because we adopted a developmental perspective to the data collected in the Italoromance contact situation, we decided that the word was an important unit of analysis. All lexical items (excluding interjections) were coded in three categories:(1)Italian4;(2)Dialect;(3)Cognates.Items which were unequivocally Italian or dialect lexemes were placed in the “Italian” and “Dialect” categories respectively. Lexical items that share the same form and meaning in both Italian and dialect systems were placed in the “Cognates” category. As Sunderman and Schwartz (2008) have pointed out, there are two different conceptions of what a cognate is, depending on the perspective adopted. In a strict linguistic perspective, cognates are two words with a common etymological origin. However, from a psycholinguistic perspective, cognates are considered synchronically and are defined as any two words that have aspects of sound, meaning and spelling across two lexicons (cf. Dijkstra’s (2009) definition). In this paper, we have adapted the psycholinguistic interpretation of the term cognate to refer to any two words that share aspects of meaning and pronunciation. The establishment of a cognate category for the assignment of words that co-occurred in both Italian and dialect utterances avoided coding the same item in two different categories which could bias the coding procedure.In our study, all words placed in the cognate category have the signifier and signified in common. Hence, this definition does not correspond entirely to the term bilingual homophone (Clyne, 1991). Initially, we aimed to assign each word to either dialect or Italian categories. Yet, we soon realized that there were certain lexical items that co-occurred with either Italian or dialect lexemes. This rendered the word assignment task more difficult. Berruto’s (2005, p. 88) description illustrates this difficulty quite clearly:There is a link between interaction (discourse) and convergence in the system. This is so because in contexts of structural interference and mixing, there is also convergence in use, i.e. standard language and dialect are simultaneously present in discourse, via code-switching and code-mixing. […] Whether it is because they are genetically closely related, or because of convergence, the two systems have in their lexicon numerous homonyms.In sociolinguistic and dialectology research, words belonging to this category have often been called bilingual homophones (Berruto, 2005; Clyne, 1967, 1991; Sobrero, 1992). Clyne (1991, p. 194) defines these as “words that sound the same in the two languages, at least in the two language systems of the speaker”. Later on, Clyne (2003) refers to bilingual homophones when analysing the lexical transfers in the bilingual subject’s production. Although it has been a very important step towards acknowledging the special status of these words, especially when analysing code-switching, the term bilingual homophone presents some problems. Homophones share the same signifier but not necessarily the same signified. Calling the words we placed in the cognate category homophones could lead to misinterpretation.Naming the lexemes Italian and dialect share cognates circumvents the terminological problem bilingual homophones present. It should be made clear that cognates are not to be considered as a third language variety. Rather, cognates are words pertaining to two genetically related languages and co-occur with words from both language repertoires. In certain cases, nouns were considered cognates if they were used in both languages. In our corpus, nouns like barca ‘boat’, roba ‘stuff’, etc. were used in both Italian and dialect interactional contexts. Moreover, certain verb forms were considered to be cognates, like for instance parla (3ps) ‘he speaks’, canta (3ps) ‘he sings’, etc. Cognates also included other words like function words that are the same in both dialect and Italian (e.g. prep. da ‘from’, su ‘on’, fra ‘between’, etc., adv. talmente ‘so’, veramente ‘truly’, praticamente ‘practically, in truth’, etc.). We thought it best to categorize these “common” words in a specific category, as they may provide the child with important information regarding language usage in his environment. We will return to this point.Below, we reproduce the coding of an utterance from our corpus. The utterance has been represented in its orthographic form (first line) and its phonetic form (second line). We have included the English translation of the utterance (third line) and its gloss (fourth line).Words that were placed in the Italian category are in bold print (four out of the nine words composing this utterance), words figuring in the cognate category are in italics (two out of nine words) and words which were placed in the dialect category are in unmarked font (three out of nine words). From our coded data, we performed an inter-annotator agreement test – the Cohen’s Kappa coefficient test – asking a dialectologist to re-code an extract of our corpus (200 utterances, roughly 1,000 words). The rate of agreement between the two annotators is good (Cohen’s kappa = 0.8).Once each lexical item was placed in one of the three categories, we adopted a quantitative approach to calculate the average proportion of each language in the utterances produced by the six speakers participating in the interactions. Using this coding, utterances containing, for example, two dialect words and eight Italian words were not categorized as merely “mixed” utterances. Rather, we calculated the mean percentage of each category per utterance, so as to obtain a more precise picture of each speaker’s language usage. Hence, an utterance containing one dialect word and ten Italian words was not analyzed in the same way as an utterance containing ten dialect words and one Italian word.4.Analyses: Variation in Usage in Two Interactional ContextsIn this section, we focus on language usage observed in two types of family interactions: 1) dyadic – mother and child as well as mother and father – contexts and 2) multiparty exchanges between six members of the same family (nuclear and extended).4.1.Adult and Child Dyadic Interactions Versus Inter-adult Interactions: Differences in dialect and Italian usage rangeWe compared language usage observed in interactions involving the child directly (in mother and child dyads) to language in inter-adult interactions. The child was present in both types of interaction, yet in the former he participated directly whereas in the latter he was a third-party participant in the exchange. Henceforth, we refer to each speaker in relation to the family tie s/he has with the child.The percentages in Table 1 represent the average proportion of words per utterance in the three lexical categories. For example, the utterances the mother produces when speaking to her child are composed of an average 1.1% of dialect words, 35.3% of cognates and 63.6% Italian words.Starting from the mother and child dyad (top part in Table 1), production in the three categories differs significantly both for the mother (Friedman’s tests: chi2 = 2648.5, p < 0.0001) and for the child (chi2 = 1263.9, p = 0.0001). Italian is the language the mother uses the most when speaking to her child. Her Italian production is greater than her dialect production (Wilcoxon tests: z = -36.3, p < 0.0001) and her cognate production (z = -19.5, p < 0.0001). Moreover, her cognate production is higher than her dialect production (z = -33.7, p < 0.0001). With regards to the child’s production when speaking to his mother, Italian is the most frequently selected language. Its production rates are higher than his dialect production (z = -28.9, p < 0.0001) and his cognate production (z = -2.6, p = 0.0008). Finally, the child uses more cognates than dialect words (z = -27.9, p < 0.0001).Focusing on the inter-adult exchanges, production rates of Italian, cognates and dialect words differ significantly when the mother speaks to the father (Friedman’s tests: chi2 = 1828.7, p < 0.0001) and when the father speaks to the mother (chi2 = 113.9, p < 0.0001). The mother produces mainly dialect words when talking to the father and her dialect production is greater than her Italian production (Wilcoxon tests: z = -7.1, p < 0.0001) and her cognate production (z = -3.4, p < 0.0001). Moreover, Italian is used less frequently and these rates are lower than the cognate rates of production (z = -6.6, p < 0.0001).Concerning the father’s production when speaking to the mother, dialect is the language which he uses the most and these percentages are significantly higher than Italian (z = -6.3, p < 0.0001), which is the language he uses least frequently. The father’s dialect production tends to be higher than the cognate production (z = -1.8, p = 0.07). The latter is however higher than his Italian production (z = -6.1, p < 0.0001).The two types of dyadic interactions, one involving adult-and-child speech and the other involving inter-adult discourse, expose divergent usages of Italian and dialect and these differences seem to depend on the interlocutors involved. Consequently, Italian and dialect usage manifest a wide production range. Italian ranges from 10.3% to 63.6% and dialect ranges from 1.1% to 53%. Cognate production does not have the same wide range of usage. Regardless of the context of production, its usage remains within the 35% and 45% production bracket whether the child is involved directly in the exchange or not.These differences in the usage of Italian and dialect can be attributed to the age of the interlocutors involved in the exchanges: When interactions involve the child directly, Italian is the code more frequently selected, whereas in interactions involving only adults, dialect is more frequently selected in the utterances. Age seems to be an important factor influencing language usage.4.2.Age Factor in Variation: Analyses of language Usage in Intergenerational Multiparty InteractionsIn the next analyses, we examined the language choices during intergenerational multiparty interactions in a family context. The participants in these interactions belonged to three generations.Looking at the speakers’ productions, the order of dialect, cognates and Italian words is significant (Friedman’s tests: chi2 = 46.8 – 380.1, p < 0.0001).The child and his mother have similar general productions, where Italian is the most frequently chosen language, followed by cognates and dialect words. For both the child and his mother, Italian production is greater than their dialect production (respectively, Wilcoxon tests: z = -23.7, p < 0.0001; z = -12.4, p < 0.0001) and it is greater than cognates (respectively, z = -5, p < 0.0001; z = -6, p < 0.0001). Cognate production for these speakers is higher than their production of dialect words (respectively: z = -20.9, p < 0.0001; z = -9.3, p < 0.0001).The child’s maternal grandmother’s and the father’s productions reflect similar usages. Cognates are most frequently selected in their utterances. The grandmother produces more cognates than Italian words (z = -2.4, p < 0.02) and dialect words (z = -3.5, p = 0.0004). Her Italian production tends to be greater than her dialect production.However, the father’s cognate production does not differ significantly from his Italian production (z = -0.6, p = 0.5, N.S.), yet it is greater than his dialect production (z = -12.2, p < 0.0001).Like the grandmother and the father, the child’s aunt produces mostly cognates in her utterances. Her cognate production does not differ from her dialect production (z = -0.4, p = 0.7, N.S.), but it is higher than her Italian production (z = -6.8, p < 0.0001). Furthermore, she produces more dialect than Italian words (z = -4.4, p < 0.0001).The child’s grandfather is the speaker who uses dialect the most. He produces more dialect words than cognates (Wilcoxon tests: z = -2.7, p = 0.0008) and Italian words (Wilcoxon tests: z = -6.7, p < 0.0001). He produces more cognates than Italian words (Wilcoxon tests: z = -6.1, p < 0.0001).Like the results obtained in the previous analyses, the mean percentages reported in Table 2 show that dialect and Italian have a wider bracket of variation in usage. Dialect usage ranges from 5.1% to 41.9% and Italian ranges from 22.5% to 53.7%. Production in the cognate category reveals less variation, given its narrow usage bracket, ranging from 34.4% to 41.3%. The results provide a glimpse of the general usages found in the multiparty interactions: The child and his parents use more Italian than the other members of the family.4.3.Similarities and Differences in Language Usage across 3 GenerationsIn this analysis, we aimed to investigate the speakers’ productions by sorting them according to generation. The grandparents were placed in the first generation, the parents and the aunt in the second and the child was placed in the third generation.The three generations manifest different usages for dialect (Kruskal-Wallis tests: H = 800.3, p < 0.0001) and for Italian (H = 250.7, p < 0.0001). However, production of cognates does not vary according to generation (H = 1.6, p = 0.4, N.S.). As expected, dialect and Italian have a wider range (respectively, from 5.1% to 34.4% and from 30% to 53.7%) than cognates (from 35.6% to 41.2%). When comparing the first and third generations, these results demonstrate a sharp decrease in the use of dialect and a gradual increase in the use of Italian by the younger members of this family. The sharp decrease in dialect usage by the younger speakers led us to examine the child’s language environment more thoroughly, by comparing the adults’ lexical choices depending on interlocutors.4.4.Differences in Adult-to-Child and Inter-Adult SpeechAiming to investigate with more precision the child’s language exposure, we focused on the adults’ productions in two different discourse types: one in inter-adult speech and the other in adult-to-child speech.As reflected in the first column, all speakers reveal different amounts of dialect when speaking to another adult as opposed to the child (Mann-Whitney tests: U = 18341 – 58441.5, p < 0.0001). In inter-adult discourse, the child’s mother, grandmother and aunt produce the most dialect. Their production rates are all above 50%. The father and the grandfather’s inter-adult dialect production remains close to the other three speakers’ productions, 44.9% and 49.9% respectively. Looking at the dialect produced in child-addressed utterances, this decreases dramatically from the production observed in inter-adult discourse. With the exception of the grandfather, each speaker’s dialect production is below 5%.Regarding cognates, the grandfather, the aunt and the mother produce different amounts when speaking to another adult compared to when addressing the child (Mann-Whitney tests: U = 31675.5 – 293116, p < 0.0001). However, this is not the case for the grandmother and the child’s father, who do not vary their production when addressing another adult or the child (respectively, Mann-Whitney tests: U = 28470, p > 0.07; U = 71685.5, p > 0.9, N.S.). Therefore, their cognate production does not seem to be influenced by the change in interlocutors.Finally, in the Italian column, patterns of usage contrast with those in the dialect column. In inter-adult discourse, all adults produce utterances containing less than 15% Italian words. However, this percentage increases significantly when these same speakers talk to the child (Mann-Whitney tests: U = 9226 – 68399.5, p < 0.0001). With the exception of the grandfather who produces 46.6% of Italian words, all speakers produce over 55% Italian words per utterance addressed to the child.When considering inter-adult and child-addressed utterances, there appears to be greater variation in the usage of dialect and Italian words. Besides the fact that all differences in the adult- and child-addressed utterances are significant, the ranges in the usage of Italian and dialect are much wider than the cognates’ range. The usages in the Italian and dialect categories range from 9.4% to 67.2% and from 1.9% to 53.9% respectively. In the cognate category however, the usage bracket is narrower, ranging from 30.9% to 41.2%.Language usage varies according to the interlocutor. When engaging in conversations with the child, Italian is the most frequently selected language, whereas in inter-adult interactions dialect is more frequently used. Lexical choice also varies according to speaker. Generally, the parents produce the most Italian when speaking to their child. Furthermore, the grandfather is the one who produces the least Italian and the most dialect when he addresses Francesco. These results support the findings in the previous analysis: the grandparents (first generation speakers) generally produce more dialect than speakers from younger generations (second and third generations).4.5.Language Exposure and Interaction: Constructing Multilingual CompetenceIn this analysis, we aimed to investigate the child’s production and the production he is exposed to in relation to the amount of time the child spends with his interlocutors. We thus focused on child-addressed speech and on the child’s production when he engages in conversations with members from his nuclear family (his parents) and with members of his extended family (his grandparents and his aunt). We analyzed the lexical choices effectuated during five of eleven days in which the family members gathered together.In Table 5, we report the child’s and his interlocutors’ productions.Dialect is least frequently selected. When the parents and the child speak to one another, their dialect production fluctuates and does not follow a particular direction (it is not clearly on the increase nor on the decrease across the 5 days). Their dialect production alters significantly throughout the eleven days both for the parents (Kruskall-Wallis tests: H = 2, p = 0.02) and for the child (H = 5.6, p < 0.0001). Looking at the child’s production and the grandparents and aunt’s productions when they engage in conversation, we highlight two elements. First, the child produces the most dialect when he speaks to these interlocutors; his production reaches 12.7% (Day 10), which is the highest percentage of dialect usage attested in the child’s production. Second, both the child and his interlocutors increase their dialect production gradually but steadily, as if they were mutually influencing each other’s productions, showing a convergence in usage: The productions of the grandparents and aunt reveal a significant curb (H = 5.9, p = 0.006) although there is a tendency for the child’s to increase (H = 2.6, p = 0.07).Concerning cognates, their production does not change significantly throughout the eleven days. When addressing the child, the parents’ cognate production remains stable between 34.2% and 38.4% (H = 1.6, p = 0.8, N.S.). When addressing his parents, Francesco’s cognate production does not change significantly (H = 5.8, p = 0.2, N.S.) and he produces between 37.5% and 45.7% of cognates per utterance. In conversations between the grandparents, the aunt and Francesco, their cognate production ranges from 30% to 43.4% and does not fluctuate significantly for the grandparents and aunt (H = 6.2, p = 0.2, N.S.) or for Francesco (H = 1.5, p = 0.8, N.S.).Looking at the respective speakers’ Italian production, these generally seem to decrease in the eleven days, although not all change significantly. For instance, when Francesco speaks to his grandparents and aunt, his Italian production tends to decrease (H = 6.8, p = 0.1). Yet, when the grandparents and the aunt speak to Francesco, their Italian production decreases significantly, from 62% to 53% (H = 15.7, p = 0.0003). This is also the case when Francesco speaks to his parents. His Italian production decreases significantly from 60% to 48% (H = 13.5, p = 0.0005). However, when his parents speak to Francesco, their Italian production does not change significantly (H = 4.1, p = 0.4). It remains stable at 60%-65%.We have documented that, during the eleven days of observation, dialect fluctuates the most and tends to increase towards the end of this period. With the exception of the child’s parents, Italian production tends to decrease. The production of cognates remains stable for all speakers through the eleven-day period. The increase in the child’s dialect production may be due to the fact that during these days in which the family gathered together, Francesco was exposed to speakers who generally produce more dialect than his parents.These results show distinct Italian and dialect usages, depending on the interlocutors involved in the interactions. Cognates production manifests less variation through different contexts. To show how these language usages materialize in the interactions, we will now present two interactions (one involving adults and the other involving adults and the child) that illustrate how Italian and dialect are used in the spontaneous data collected.The first extract reproduces a multiparty interaction between the mother, the father, the maternal grandmother, the aunt and Francesco. The child was not directly involved in the exchange although he was present and “interrupted” the inter-adult exchange. The thematic orientation of this extract is language competence in Italian. The mother starts this conversation by saying that nobody in Italy is capable of speaking Italian properly. The father, in a playful tone replies that he is the only one who is capable of speaking Italian well.As can be seen this inter-adult exchange is mainly in dialect. It is important to note the way the different speakers use the languages in their repertoires to put forwards their arguments supporting their language views. The mother, for example, says in dialect that nobody knows how to speak Italian well. When she refers to ‘Italian’ she does not use the dialect form (/italjaɲ/) even though her speech turn is entirely in dialect. She does the same thing in speech turn 3. It is as if the mother wanted to metalinguistically underline the codes she is referring to.Interestingly, despite the use of dialect in the exchange, the aunt also “demarks” her dialect utterance by using the Italian adverb /antzi/. The contrast provoked by the change in codes attracts the attention of the speakers on what she is about to say. She supports the mother’s views but her utterance contrasts the common ideas on dialect’s so-called interference in language proficiency, explaining her use of the adverb /antzi/.In the successive turns, the grandmother’s /mamːa/ attracts the child’s attention and he produces /mama/ with the simplification of the geminate, a common phonetic trait in the dialectalized Italians and italianized dialects (dial. /mare/ versus Ita. /mamːa/ ‘mother’). The item /mama/ is both produced in Italian and dialect contexts, thus it has been classified as a cognate. The grandmother wraps up the exchange with the claim that Venetans, despite their use of dialect, have stronger language competence than speakers from other regions in Italy.In this extract the utterances are mainly in dialect. Cognates co-occur with the dialect words. When the speakers change codes, in this case by switching from dialect to Italian, they may do so to mark, and thus reinforce, their views that they put forward.In the exchange that follows, the change in code is linked not with the subject matter or the views the speaker tries to put forward. Rather, the change in code is in line with the change of interlocutor.These two utterances by the mother were produced in parallel with dinnertime conversations mainly among the adults. She is busy helping Francesco eat and tells him to chew his food properly. In the second turn, she then turns to the father and tells him to leave him alone and to let him eat. Although these two utterances are produced in isolation, it is nevertheless interesting to see the link binding language choice and interlocutor. In fact, the shift in interlocutors coincides with the shift in language choice, showing quite clearly the code choices in inter-adult discourse that were pointed out in the quantitative analyses. In these two utterances, the child is exposed to two variants of the verb ‘chew’, one is Italian and the other is dialect.5.DiscussionThree observations are highlighted. First, in interactions involving the child directly, Italian is the main language selected. Second, in interactions involving adults and not the child directly, utterances contain mainly dialect words. Third, dialect and Italian show greater variation in average production rates than cognates. Dialect and Italian have the widest ranges of usage. The dialect usage bracket stretches from 0.6% to 54.3% and the Italian bracket stretches from 8.5% to 67.2%. Cognates have a limited range, varying from 30.9% to 45.3%. The breadth of the ranges is an indication of the amount of variation observed within one category.The narrow ranges in the use of cognates suggests that their production is relatively stable: it does not vary from one interactional context to another. Three observations support this argument. First, the production of cognates does not vary significantly according to the generation of the speaker (Analysis 4.3). Second, in certain cases, it does not vary significantly between child to adult interlocutors (Analysis 4.4). Third, during the eleven-day period of observation, we noticed that the production of cognates did not vary significantly over time and remained within the 30%-40% usage bracket (Analysis 4.5). The consistent presence of cognates in both Italian and dialect contexts means that in Francesco’s language environment he can encounter cognates that co-occur with either dialect or Italian words. Developmentally, this is important for two reasons. Firstly, the child will be exposed to both dialect and Italian utterances containing the common content represented by cognates. He can then build pivotal schemas containing a pivot – a fixed element, a cognate in our case – and a variable element from either dialect or Italian lexicons. We will illustrate this point with two examples from our multiparty corpus.These two utterances were produced during mealtimes and both were child-addressed utterances. As in the previous linguistic examples, Italian words are in bold, cognates are in italics and dialect words are in unmarked font. We found the corresponding dialect utterances in inter-adult discourse (1b and 2b), the semantic content of which corresponds to the Italian utterances from 1(a) and 2(a).In Utterance (1) and (2), both the Italian and the dialect utterances fulfill the same communicative intention. For example, looking at (1a) and (1b), there is one common part questo, a cognate, and two contrasting elements, respectively cos’è, Italian, and cossa ze, dialect. The child will be exposed to cos’è in the child-directed discourse and to cossa ze in the inter-adult discourse to which he is exposed daily. Through linguistic experience and exposure to inter-adult and child-directed speech, he will gradually understand that cos’è questo and cossa zeo questo are produced to communicate the same pragmatic intentions, that is, the request of information regarding an object present in the context of the interaction. The child will thus match cos’è and cossa ze and use the one or the other after consideration of the appropriate language choices depending on the interactional context.Through the exposure to contrasting input (Italian in child-addressed speech and dialect in inter-adult discourse), the child will be able to associate Italian and dialect constructions that fulfill the same communicative intentions and to extract the single lexical items pertaining to each one. Gradually, Francesco should gain sufficient exposure to be able to recognize and memorize the dialect items found in inter-adult speech or in interactions that encourage his dialect comprehension and production, as was the case in the exchanges with extended family members. Moreover, the child will learn to generalize the construction (cognate/pivot + X) across the other constructions memorized by combining the cognate – the pivot – with any item that can fill the empty slot (X). In this way, cognates facilitate access to the form of dialect items by reducing the differences between the Italian and dialect systems. Boundaries separating the two codes become fuzzier, facilitating the possibility of change from one code to another within discourse. The presence of items that – like cognates – decrease the formal differences between two (or more) codes would seem to support the maintenance of dialect among the new generations. Ultimately, the child will be able to generalize the language constructions across dialect and Italian lexicons.6.ConclusionThese results may not seem new. Scholars working in the Italoromance area have observed that it is precisely through the production of mixed utterances6 that Italy’s multilingualism is maintained (Dal Negro, 2005; Rindler Schjerve, 2003) and that dialects have not fallen into disuse as suggested by statistics on declared usages. Moreover, mixed utterances provide evidence of the dynamic relationship in the Italian-dialect contact situation, where Italian and dialect exert influence on each other, through the processes of regionalization of the national language and Italianization of dialect. Yet, these studies do not explain in detail how mixed utterances support language maintenance. Our study suggests that it is not only due to speakers using both languages in their utterances that minority languages are maintained. Additionally, language maintenance may be supported by similarities in the input and output provided by cognates and the contrasting usages provided by Italian in the child-directed speech and dialect in inter-adult discourse. This may facilitate semantic access to dialect – the minority language – and could thus scaffold dialect acquisition. Psycholinguistic research on multilingual lexicons has yielded important results supporting our hypothesis on the potentially important role of cognates in language learning. A robust body of research which has shown that the semantic and formal proximity between cognates facilitates lexical access in reading, listening and production (Caramazza and Brones, 1979; Costa, Albareda, and Santesteban, 2008; Costa, Santesteban, and Cano, 2005; Cristoffanini, Kirsner, and Milech, 1986; Lemhöfer et al., 2008). In a recent article, Dijkstra and colleagues (2010: 297) conducted a study on the cognate processing in English-Dutch bilinguals. Interestingly, they found that “a facilitation effect did arise for lower frequency identical cognates relative to other items […]. This finding indicates that, as might be expected, identical English low-frequency cognates profit more from the co-activation of their Dutch counterparts than high-frequency cognates do”. Reading our results in light of the psycholinguistic studies mentioned above, we could hypothesize that cognates may help process low frequency dialect words. Hence, exposed to contrasting input in the multiparty interactions, Francesco’s dialect production may have been facilitated by the cognates co-occurring with the dialect produced in his language environment.The results and hypotheses discussed in this study should be tested with more quantitative and qualitative analyses conducted in similar – but larger scale – sociolinguistic contexts, for instance Creole or Spanish/Catalan contact situations. More importantly, future quantitative studies on development in contact situations should attempt to include variation phenomena in order to better represent the language environment and to gain a wider understanding of the stages of acquisition occurring in language contact situations. We have attempted to do so through the establishment of the cognate category. Much work remains to investigate the cognate category in more detail, for example from a morphosyntactic perspective. The outcomes of this research could also provide new methodological ideas for studies conducted on naturalistic data as well as elements for the debate on the transmission of minority languages and the pathways leading to the construction of multilingual competence in informal language learning contexts. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Language Contact Brill

Quantitative Approaches to Italian / Dialect Family Interactions: Considerations of Methodology and Language Transmission in a Contact Situation

Journal of Language Contact , Volume 6 (1): 28 – Jan 1, 2013

Loading next page...
 
/lp/brill/quantitative-approaches-to-italian-dialect-family-interactions-YjIaYmTwWy

References (87)

Publisher
Brill
Copyright
Copyright © Koninklijke Brill NV, Leiden, The Netherlands
Subject
Articles
ISSN
1877-4091
eISSN
1955-2629
DOI
10.1163/19552629-006001010
Publisher site
See Article on Publisher Site

Abstract

1.IntroductionPsycholinguistic and sociolinguistic studies conducted in the Italoromance area have mostly taken two distinct directions. On the one hand, psycholinguistic studies have focused on Italian acquisition. These studies have investigated the acquisition of Italian vocabulary (Bates, Burani, D’Amico, and Barca, 2001; D’Odorico, Carubbi, Salerni, and Calvo, 2001) of Italian morphology (Pizzuto and Caselli, 1992; Noccetti, De Marco, Tonelli and Dressler, 2007) and of Italian syntax (Waxman and Guasti (2009) and Guasti (2009) for production and Arosio, Adani and Guasti (2009) for comprehension). Psycholinguistic research conducted in this area has generally adopted a quantitative approach. On the other hand, sociolinguistic studies have described the language contact phenomena observed in the multilingual social repertoires (Berruto, 1990; Cerruti, 2003; Regis, 2002). There have been few attempts to integrate psycholinguistic and sociolinguistic approaches when studying language development in the Italoromance area. For example, investigations on language acquisition have seldom included variation phenomena inherent in the social repertoires. Ignoring the elements of variation in the environment within which an Italian child constructs his/her language omits important information regarding a child’s language exposure. In fact, studying the variation phenomena (from phonological, lexical and morphosyntactic levels) present in the child’s language environment may help clarify the processes a child goes through when s/he acquires a language within spontaneous and informal contexts of interaction.We have adopted a usage-based approach as it has the advantage of being sufficiently flexible to integrate sociolinguistic variation generated by the Italian/dialect language contact with the study of child language. More generally, integrating methods and approaches from sociolinguistics and psycholinguistics could be particularly useful when analyzing acquisition data collected in contact situations, as in Italy. On the one hand, psycholinguistic studies could benefit from sociolinguistic approaches to define and understand the sources of variation in the child’s language environment. On the other hand, sociolinguistic approaches could benefit from psycholinguistic approaches in order to understand the cognitive foundations underpinning variation in child production.One of the main tenets of the usage-based account is that the child builds his/her language(s) from concrete pieces of language encountered in his/her environment (Behrens, 2009; Tomasello, 2003). Through linguistic exposure and experience, the child starts building constructions, the latter being “form and meaning pairings” (Goldberg, 2006, p. 3) from these concrete pieces of language. At the two-word stage (from approximately 18 – 24 months), the child will start forming his/her first constructions by combining two lexical items. Although these early constructions – called pivot schemas (Braine, 1976) – are not manifestations of the emergence of syntax, they nevertheless show the initial stages of categorisation and the expression of a child’s communicative intentions (Tomasello, 2003). These constructions are made up of two slots: one is filled by a pivot, which is the structuring unit of the construction in that it places it within a speech act (for example More in “More X” places this pivot schema within a request) and the other is a variable slot filled by an element pertaining to the immediate communicative setting (for instance cookie in more cookie, for further examples see Braine 1976 or Tomasello 2003, pp. 94-143). Once these constructions emerge, the child gradually starts making generalizations across these pivot-schemas, building more complex constructions. This process of generalization is an important step towards the production of abstract linguistic constructions (Tomasello, 2006) and, more specifically, to the development of syntax.The present study observes the production of Francesco, a child aged between 17 and 30 months, participating in dyadic and multiparty interactions with five members of his nuclear and extended family who live in Castelfranco Veneto, a city in the central part of Veneto, one of Italy’s northeastern regions. Here, Veneto dialect and Italian are spoken on a daily basis in both family and public spheres. The speakers taking part in the interactions belong to three generations. In this study, we observed language development and environment jointly. Our investigation was guided by three questions:(1)What languages are spoken in the child’s environment?(2)How are dialect and Italian used when speaking in dyadic and multiparty interactions, with speakers from different generations?(3)Is language usage in the child’s production related to the language usage observed in his environment?These three questions will guide the analyses of the data in the investigation of each speaker’s production observed in informal dyadic and multiparty interactional settings.2.A Sociolinguistic Approach to Italy’s MultilingualismSociolinguistic studies conducted in Italy have contributed to a better understanding of phenomena generated by the joint presence of Italian and the Italoromance dialects in social repertoires. As Berruto (2005) has pointed out, Italy’s dialetti do not reflect the same entity as the English dialects. According to Coseriu’s (1981) tripartite categorization of what is designated by the term “dialect,” Italoromance dialects figure among the primary dialects, i.e. dialects that have their own autonomous linguistic system whereas English dialects figure among the tertiary dialects because they are considered as varieties within one language system (see Sornicola (2002) for a review of Coseriu’s theory). Grassi (1993: 279) insists that Italoromance dialects – like any language – have a specific phonetic, phonological and morphosyntactic structure, yet they differ in the various functions that are attributed to each language. For instance, Italian is used in the public sphere and is both the medium of communication (in both formal and informal settings) and the object of study in the field of education. Clearly, dialects have a smaller speech community than Italian. Moreover, variation in usage can be observed within the same speech community. As Giacalone-Ramat (1995) notes, adults are the main dialect users in the Italoromance area. Her observation was confirmed by the results of a national survey on declared language usage conducted in Italy and published in 2007 in the Istituto Nazionale di Statistica (ISTAT) site.1 The results show a clear preference for Italian among children and young adults. Veneto’s dialect follows the national usage trends, being predominantly used by the adult population and within the restricted family environment. Despite the trends observed, statistics on declared language usages in Veneto show that the region’s dialect is one of the most vital in Italy. These findings however do not provide clear indications of speakers’ actual usage, and with whom and in which circumstances their code choices are effectuated.2.1.Issues of Variation: A Glimpse of the Processes of Dialect Italianization and Italian RegionalizationThe first investigations on diatopic variation in the Italoromance area were carried out by Rüegg (1956). His observations showed that, despite strong political movements favoring the process of Italianization, there was a considerable amount of variation in the Italoromance domain. Despite the adoption of Italian as the one and only official language, there was considerable interregional variation in daily language usage (Guerini, 2011; Marcato, 1991). Rüegg’s diatopic variation was later theorized by Pellegrini who introduced the concept of Regional Italian, which has subsequently become the object of numerous sociolinguistic studies (Cerruti, 2003, 2011; Cerruti and Regis, 2005; Dal Negro, 2005; Marcato, 1991). These studies underscore the significant amount of intralinguistic variation in the regional realizations of Italian given the influence the Italoromance dialects exert on Italian. Hence an important element contributing to Italian intralinguistic variation is the interlinguistic variation existing between Italian and dialect. The regional dialects “regionalize” Italian production, particularly in its oral form, on all linguistic levels (semantic, phonological, etc.). When analyzing the process of Italian regionalization, Telmon (1990, 1993) compares it to Selinker’s (1972) notion of interlanguage. This notion refers to a particular stage in second language acquisition where the learner’s production reveals the presence of a linguistic system “in progress” which neither belongs to the first language nor to the second language. Telmon shows numerous similarities between the contact phenomena observed in the formation of regional Italians and the phenomena observed in learning a second language. Similarly, observing the initial stages in the acquisition of French as a second language, Véronique (2001) compares these stages to the formation process of the French creoles’ grammatical systems. The fact that there are numerous similarities in different sociolinguistic contexts shows that processes of contact are in many respects generalizable.Since Italy’s unification in the late nineteenth century, and in the last three decades in particular, Italian has steadily gained substantial ground in the repertoires, being the lingua franca for interregional communication as Italy’s national and official language (Sobrero, 1992; Telmon, 1990, 1993; Tempesta, 2001). Moreover, as the language of instruction, the acquisition of Italian is a priority because it is essential for access to education, being the national and most prestigious language. Consequently, Italian dominates the majority of the repertoires and it has exerted substantial influence on regional languages, gaining ground in public (school, administration, etc.) and private (family) spheres. Italy’s unification in 1861 has accelerated the process of Italianization of the Italoromance dialects as the latter were seen as elements of division (Coveri et al. 1998). On the contrary, Italian was perceived as the language that could bring unity in the newly formed republic. Hence, Italy’s unification did not only entail a political union between the regions composing the newly formed country. The relationship between Italian and dialect has moved from diglossia to dilalia (Berruto, 1987a).2 Moreover, Italy’s unification created a zone of contact between the official and prestigious language and the regional and less prestigious languages.As Berruto (1983, 1997) points out, the formation of regional Italians shares numerous characteristics with the process of creolization. Both involve a prestigious variety that comes into contact with a less prestigious variety. In these two situations, the prestigious varieties (the superstrates) are Italian, in the Romance area, and the colonial language, in the case of creole situations. In the Romance area, the less prestigious languages (the substrates) are the Romance dialects and in the creole situation, these are local varieties spoken prior to colonization. When addressing the phenomenon of dialects’ Italianization, Berruto (1997) compares the formation process of hybrid forms – forms containing Italian and dialect morphemes – to the process of relexification within a creole situation. Scholars studying creole language practices have also dedicated much attention to the presence of hybrid forms. Prudent (1980) describes the intense variability in the community in his description of French Caribbean creoles. Prudent’s description gives a clear picture of the intricacy of French and Creole usage:La pratique linguistique de la communauté comprend une utilisation de systèmes hybrides extrêmement difficiles à décrire en l’état (Prudent 1980, pp. 79).Italoromance and creole situations have in common the presence of language varieties that promote the emergence of hybrid forms that draw from elements of both systems in contact.The majority of the Italoromance dialects are used only in their oral form. Certain dialects have also benefitted from a prestigious literary tradition, as is the case for the Veneto dialect which was the national language of the Venetian republic (Cortelazzo, 2001, 2004). As the national and official language, Italian is used in both oral and written contexts and is the most represented language in the media. Standard Italian has been formalized in grammars and exists only in its written form (Berruto, 1987a; Pellegrini, 1960; Sanga, 1981). In its oral form, standard Italian has rarely been documented (Berruto, 1987b; Stehl, 1995). Therefore, Italoromance scholars do not refer to standard Italian when dealing with oral production, because its heterogeneity is not only constrained by intralinguistic variation but also by interlinguistic variation, given the Italian/dialect language contact. The latter creates numerous zones of transfer and their joint usage forms a multidimensional linguistic continuum (Berruto, 1998). In sum, language contact in the Italian context may be seen as the presence of varieties, each of which exerts forces of attraction and repulsion on the other varieties, as in a magnetic field. As in a magnetic field, it is difficult to determine with precision where the pull of attraction and repulsion of one variety begins and the other stops.Given the high variability in language contact situations, it is then difficult to apply Bickerton’s (1975) definition of continuum to Italy’s dialetti and the Italian language contact situation. Berruto (1997) acknowledges the methodological interest in organizing linguistic data collected in the Romance area in an orderly fashion. Yet, a considerable impediment to such an ordering is the difficulty in neatly ranking the intermediate varieties (those along the Italian-dialect continuum, given the extensive interchangeability between Italianized forms of dialect and dialectalized forms of Italian) according to the traits that characterize them (Berruto 1985; 2005).Interestingly, Bickerton’s continuum model has also been criticized in creole studies conducted by Prudent (1980). Like Berruto’s, Prudent’s critique concerns the model’s rigidity, which makes it incompatible with the flexibility inherent in actual language usage. His description of the French Caribbean creole clearly conveys the intense variability in the community’s usages. This led him to introduce the notion of interlect that could be related to the notion of continuum as it has been described in the Italoromance area. In both situations, the grammatical structures of the intermediate varieties existing between two language poles cannot be predicted by the grammatical structure of the languages in contact. In both situations, there are no borders separating the structure of one variety from the structure of the other. Berruto (1985: 71) exemplifies these difficulties through the close scrutiny of language mixing found in recordings made in Bergamo. In his contribution, it is clear that the continuous switching from one language to another renders the task of language assignment particularly difficult, because at times it is not clear whether items are the fruit of a dialectalized Italian or, on the contrary, of an Italianized dialect. In order to circumvent this problem, Berruto (1985: 70) suggests it is best to assign these ambiguous items to the category of the co-occurring items within the utterance/clause level. We will come back to this point, in the methodology section.2.2.Focussing on the Sociolinguistic Characteristics of the Veneto Speech CommunityVeneto is one of Italy’s most vital dialects and shows resilience to the process of Italianization (Coveri et al., 1998; Sobrero, 1992), as its usage conveys strong regional identity (Cortelazzo and Paccagnella 1997; Marcato 2002). In daily social exchanges, dialect and Italian are used jointly, resulting in a profusion of mixed utterances (Trumper & Maddalon, 1982). The numerous descriptions of the dialect varieties (Zamboni, 1974; 2002; Marcato, 1981; Trumper et Vigolo, 1995; Cortelazzo et Paccagnella, 1997 ; Vanelli, 2003; inter alia) show that the Venetian dialect variety is the most prestigious and consequently it strongly influences other dialect varieties (Fray, 1962; Zamboni, 2002). The dialects belonging to the Veneto dialect group form a koiné, that is a common variety that arises from the structural convergence of the dialects in contact. In other terms, it could be viewed as “a spontaneous standardization” (Giovanni Depau, personal communication) that emerges through prolonged contact between dialect varieties, tending towards the variety spoken around Venice. In the case of the Veneto dialect group, the resulting koiné allows intercomprehension between speakers of different dialect varieties. All forms from the Veneto dialect koiné are labelled as dialect in the present research.The Veneto dialects are structurally similar to Italian and daily usage tends to move progressively towards Italian, the prestigious variety in the Italoromance repertoires.In sum, two important points can be drawn from the description of multilingualism in the Italoromance area. Firstly, the processes of language contact are observed in different sociolinguistic contexts. Hence the study of language acquisition in the Italoromance area could address issues pertaining to the construction of multilingual competence, particularly when dealing with issues of transmission and maintenance of minority/threatened languages. Secondly, the existence of regional Italians and Italianized dialects entails the contact between two or more language varieties within the same community. Variation is thus an intrinsic characteristic of the social repertoires and should be included especially when dealing with developmental data, as this variation is an important characteristic of the Italian child’s language environment.Considering these two points, our study aims to gain understanding on the acquisition processes observed in variable environments of informal contexts of interaction. In light of the results obtained, we suggest possible ways in which the dialect and the regional variety of Italian can be transmitted to children. This study also has a more practical aim: to improve methodology pertaining to the study of acquisition in a contact situation, by taking sociolinguistic characteristics into consideration within the protocol of investigation and in the data analyses. The methodological design of this case study takes into account the issues of variation inherent in the language environment observed in the Italoromance area.3.Method: Data collection and codingWe collected data in dyadic and multiparty family interactions. The group of speakers comprised Francesco (17 – 30 months) and five adults (the child’s parents, maternal grandparents and one aunt). Two types of dyadic interactions were taped: mother-to-child and mother-to-father. The multiparty interactions took place between the child and the five adults participating in the recordings. A total of 35 hours of interactions (= 90,000 word tokens) were taped and transcribed by a native Italian and Veneto dialect speaker.For the sake of simplicity, data were transcribed orthographically in an Excel database. However, we did consider phonetic aspects (e.g., the simplification of the geminate: /t:/ > [t]; consonantic reduction /ts/ > [s]; etc.): these were represented phonetically when they occurred in the linguistic data collected. In cases of doubt regarding the morphological segmentation of dialect words/phrases, we used two main reference books: Boerio (1856) and Marcato and Ursini (1998).3 Because we adopted a developmental perspective to the data collected in the Italoromance contact situation, we decided that the word was an important unit of analysis. All lexical items (excluding interjections) were coded in three categories:(1)Italian4;(2)Dialect;(3)Cognates.Items which were unequivocally Italian or dialect lexemes were placed in the “Italian” and “Dialect” categories respectively. Lexical items that share the same form and meaning in both Italian and dialect systems were placed in the “Cognates” category. As Sunderman and Schwartz (2008) have pointed out, there are two different conceptions of what a cognate is, depending on the perspective adopted. In a strict linguistic perspective, cognates are two words with a common etymological origin. However, from a psycholinguistic perspective, cognates are considered synchronically and are defined as any two words that have aspects of sound, meaning and spelling across two lexicons (cf. Dijkstra’s (2009) definition). In this paper, we have adapted the psycholinguistic interpretation of the term cognate to refer to any two words that share aspects of meaning and pronunciation. The establishment of a cognate category for the assignment of words that co-occurred in both Italian and dialect utterances avoided coding the same item in two different categories which could bias the coding procedure.In our study, all words placed in the cognate category have the signifier and signified in common. Hence, this definition does not correspond entirely to the term bilingual homophone (Clyne, 1991). Initially, we aimed to assign each word to either dialect or Italian categories. Yet, we soon realized that there were certain lexical items that co-occurred with either Italian or dialect lexemes. This rendered the word assignment task more difficult. Berruto’s (2005, p. 88) description illustrates this difficulty quite clearly:There is a link between interaction (discourse) and convergence in the system. This is so because in contexts of structural interference and mixing, there is also convergence in use, i.e. standard language and dialect are simultaneously present in discourse, via code-switching and code-mixing. […] Whether it is because they are genetically closely related, or because of convergence, the two systems have in their lexicon numerous homonyms.In sociolinguistic and dialectology research, words belonging to this category have often been called bilingual homophones (Berruto, 2005; Clyne, 1967, 1991; Sobrero, 1992). Clyne (1991, p. 194) defines these as “words that sound the same in the two languages, at least in the two language systems of the speaker”. Later on, Clyne (2003) refers to bilingual homophones when analysing the lexical transfers in the bilingual subject’s production. Although it has been a very important step towards acknowledging the special status of these words, especially when analysing code-switching, the term bilingual homophone presents some problems. Homophones share the same signifier but not necessarily the same signified. Calling the words we placed in the cognate category homophones could lead to misinterpretation.Naming the lexemes Italian and dialect share cognates circumvents the terminological problem bilingual homophones present. It should be made clear that cognates are not to be considered as a third language variety. Rather, cognates are words pertaining to two genetically related languages and co-occur with words from both language repertoires. In certain cases, nouns were considered cognates if they were used in both languages. In our corpus, nouns like barca ‘boat’, roba ‘stuff’, etc. were used in both Italian and dialect interactional contexts. Moreover, certain verb forms were considered to be cognates, like for instance parla (3ps) ‘he speaks’, canta (3ps) ‘he sings’, etc. Cognates also included other words like function words that are the same in both dialect and Italian (e.g. prep. da ‘from’, su ‘on’, fra ‘between’, etc., adv. talmente ‘so’, veramente ‘truly’, praticamente ‘practically, in truth’, etc.). We thought it best to categorize these “common” words in a specific category, as they may provide the child with important information regarding language usage in his environment. We will return to this point.Below, we reproduce the coding of an utterance from our corpus. The utterance has been represented in its orthographic form (first line) and its phonetic form (second line). We have included the English translation of the utterance (third line) and its gloss (fourth line).Words that were placed in the Italian category are in bold print (four out of the nine words composing this utterance), words figuring in the cognate category are in italics (two out of nine words) and words which were placed in the dialect category are in unmarked font (three out of nine words). From our coded data, we performed an inter-annotator agreement test – the Cohen’s Kappa coefficient test – asking a dialectologist to re-code an extract of our corpus (200 utterances, roughly 1,000 words). The rate of agreement between the two annotators is good (Cohen’s kappa = 0.8).Once each lexical item was placed in one of the three categories, we adopted a quantitative approach to calculate the average proportion of each language in the utterances produced by the six speakers participating in the interactions. Using this coding, utterances containing, for example, two dialect words and eight Italian words were not categorized as merely “mixed” utterances. Rather, we calculated the mean percentage of each category per utterance, so as to obtain a more precise picture of each speaker’s language usage. Hence, an utterance containing one dialect word and ten Italian words was not analyzed in the same way as an utterance containing ten dialect words and one Italian word.4.Analyses: Variation in Usage in Two Interactional ContextsIn this section, we focus on language usage observed in two types of family interactions: 1) dyadic – mother and child as well as mother and father – contexts and 2) multiparty exchanges between six members of the same family (nuclear and extended).4.1.Adult and Child Dyadic Interactions Versus Inter-adult Interactions: Differences in dialect and Italian usage rangeWe compared language usage observed in interactions involving the child directly (in mother and child dyads) to language in inter-adult interactions. The child was present in both types of interaction, yet in the former he participated directly whereas in the latter he was a third-party participant in the exchange. Henceforth, we refer to each speaker in relation to the family tie s/he has with the child.The percentages in Table 1 represent the average proportion of words per utterance in the three lexical categories. For example, the utterances the mother produces when speaking to her child are composed of an average 1.1% of dialect words, 35.3% of cognates and 63.6% Italian words.Starting from the mother and child dyad (top part in Table 1), production in the three categories differs significantly both for the mother (Friedman’s tests: chi2 = 2648.5, p < 0.0001) and for the child (chi2 = 1263.9, p = 0.0001). Italian is the language the mother uses the most when speaking to her child. Her Italian production is greater than her dialect production (Wilcoxon tests: z = -36.3, p < 0.0001) and her cognate production (z = -19.5, p < 0.0001). Moreover, her cognate production is higher than her dialect production (z = -33.7, p < 0.0001). With regards to the child’s production when speaking to his mother, Italian is the most frequently selected language. Its production rates are higher than his dialect production (z = -28.9, p < 0.0001) and his cognate production (z = -2.6, p = 0.0008). Finally, the child uses more cognates than dialect words (z = -27.9, p < 0.0001).Focusing on the inter-adult exchanges, production rates of Italian, cognates and dialect words differ significantly when the mother speaks to the father (Friedman’s tests: chi2 = 1828.7, p < 0.0001) and when the father speaks to the mother (chi2 = 113.9, p < 0.0001). The mother produces mainly dialect words when talking to the father and her dialect production is greater than her Italian production (Wilcoxon tests: z = -7.1, p < 0.0001) and her cognate production (z = -3.4, p < 0.0001). Moreover, Italian is used less frequently and these rates are lower than the cognate rates of production (z = -6.6, p < 0.0001).Concerning the father’s production when speaking to the mother, dialect is the language which he uses the most and these percentages are significantly higher than Italian (z = -6.3, p < 0.0001), which is the language he uses least frequently. The father’s dialect production tends to be higher than the cognate production (z = -1.8, p = 0.07). The latter is however higher than his Italian production (z = -6.1, p < 0.0001).The two types of dyadic interactions, one involving adult-and-child speech and the other involving inter-adult discourse, expose divergent usages of Italian and dialect and these differences seem to depend on the interlocutors involved. Consequently, Italian and dialect usage manifest a wide production range. Italian ranges from 10.3% to 63.6% and dialect ranges from 1.1% to 53%. Cognate production does not have the same wide range of usage. Regardless of the context of production, its usage remains within the 35% and 45% production bracket whether the child is involved directly in the exchange or not.These differences in the usage of Italian and dialect can be attributed to the age of the interlocutors involved in the exchanges: When interactions involve the child directly, Italian is the code more frequently selected, whereas in interactions involving only adults, dialect is more frequently selected in the utterances. Age seems to be an important factor influencing language usage.4.2.Age Factor in Variation: Analyses of language Usage in Intergenerational Multiparty InteractionsIn the next analyses, we examined the language choices during intergenerational multiparty interactions in a family context. The participants in these interactions belonged to three generations.Looking at the speakers’ productions, the order of dialect, cognates and Italian words is significant (Friedman’s tests: chi2 = 46.8 – 380.1, p < 0.0001).The child and his mother have similar general productions, where Italian is the most frequently chosen language, followed by cognates and dialect words. For both the child and his mother, Italian production is greater than their dialect production (respectively, Wilcoxon tests: z = -23.7, p < 0.0001; z = -12.4, p < 0.0001) and it is greater than cognates (respectively, z = -5, p < 0.0001; z = -6, p < 0.0001). Cognate production for these speakers is higher than their production of dialect words (respectively: z = -20.9, p < 0.0001; z = -9.3, p < 0.0001).The child’s maternal grandmother’s and the father’s productions reflect similar usages. Cognates are most frequently selected in their utterances. The grandmother produces more cognates than Italian words (z = -2.4, p < 0.02) and dialect words (z = -3.5, p = 0.0004). Her Italian production tends to be greater than her dialect production.However, the father’s cognate production does not differ significantly from his Italian production (z = -0.6, p = 0.5, N.S.), yet it is greater than his dialect production (z = -12.2, p < 0.0001).Like the grandmother and the father, the child’s aunt produces mostly cognates in her utterances. Her cognate production does not differ from her dialect production (z = -0.4, p = 0.7, N.S.), but it is higher than her Italian production (z = -6.8, p < 0.0001). Furthermore, she produces more dialect than Italian words (z = -4.4, p < 0.0001).The child’s grandfather is the speaker who uses dialect the most. He produces more dialect words than cognates (Wilcoxon tests: z = -2.7, p = 0.0008) and Italian words (Wilcoxon tests: z = -6.7, p < 0.0001). He produces more cognates than Italian words (Wilcoxon tests: z = -6.1, p < 0.0001).Like the results obtained in the previous analyses, the mean percentages reported in Table 2 show that dialect and Italian have a wider bracket of variation in usage. Dialect usage ranges from 5.1% to 41.9% and Italian ranges from 22.5% to 53.7%. Production in the cognate category reveals less variation, given its narrow usage bracket, ranging from 34.4% to 41.3%. The results provide a glimpse of the general usages found in the multiparty interactions: The child and his parents use more Italian than the other members of the family.4.3.Similarities and Differences in Language Usage across 3 GenerationsIn this analysis, we aimed to investigate the speakers’ productions by sorting them according to generation. The grandparents were placed in the first generation, the parents and the aunt in the second and the child was placed in the third generation.The three generations manifest different usages for dialect (Kruskal-Wallis tests: H = 800.3, p < 0.0001) and for Italian (H = 250.7, p < 0.0001). However, production of cognates does not vary according to generation (H = 1.6, p = 0.4, N.S.). As expected, dialect and Italian have a wider range (respectively, from 5.1% to 34.4% and from 30% to 53.7%) than cognates (from 35.6% to 41.2%). When comparing the first and third generations, these results demonstrate a sharp decrease in the use of dialect and a gradual increase in the use of Italian by the younger members of this family. The sharp decrease in dialect usage by the younger speakers led us to examine the child’s language environment more thoroughly, by comparing the adults’ lexical choices depending on interlocutors.4.4.Differences in Adult-to-Child and Inter-Adult SpeechAiming to investigate with more precision the child’s language exposure, we focused on the adults’ productions in two different discourse types: one in inter-adult speech and the other in adult-to-child speech.As reflected in the first column, all speakers reveal different amounts of dialect when speaking to another adult as opposed to the child (Mann-Whitney tests: U = 18341 – 58441.5, p < 0.0001). In inter-adult discourse, the child’s mother, grandmother and aunt produce the most dialect. Their production rates are all above 50%. The father and the grandfather’s inter-adult dialect production remains close to the other three speakers’ productions, 44.9% and 49.9% respectively. Looking at the dialect produced in child-addressed utterances, this decreases dramatically from the production observed in inter-adult discourse. With the exception of the grandfather, each speaker’s dialect production is below 5%.Regarding cognates, the grandfather, the aunt and the mother produce different amounts when speaking to another adult compared to when addressing the child (Mann-Whitney tests: U = 31675.5 – 293116, p < 0.0001). However, this is not the case for the grandmother and the child’s father, who do not vary their production when addressing another adult or the child (respectively, Mann-Whitney tests: U = 28470, p > 0.07; U = 71685.5, p > 0.9, N.S.). Therefore, their cognate production does not seem to be influenced by the change in interlocutors.Finally, in the Italian column, patterns of usage contrast with those in the dialect column. In inter-adult discourse, all adults produce utterances containing less than 15% Italian words. However, this percentage increases significantly when these same speakers talk to the child (Mann-Whitney tests: U = 9226 – 68399.5, p < 0.0001). With the exception of the grandfather who produces 46.6% of Italian words, all speakers produce over 55% Italian words per utterance addressed to the child.When considering inter-adult and child-addressed utterances, there appears to be greater variation in the usage of dialect and Italian words. Besides the fact that all differences in the adult- and child-addressed utterances are significant, the ranges in the usage of Italian and dialect are much wider than the cognates’ range. The usages in the Italian and dialect categories range from 9.4% to 67.2% and from 1.9% to 53.9% respectively. In the cognate category however, the usage bracket is narrower, ranging from 30.9% to 41.2%.Language usage varies according to the interlocutor. When engaging in conversations with the child, Italian is the most frequently selected language, whereas in inter-adult interactions dialect is more frequently used. Lexical choice also varies according to speaker. Generally, the parents produce the most Italian when speaking to their child. Furthermore, the grandfather is the one who produces the least Italian and the most dialect when he addresses Francesco. These results support the findings in the previous analysis: the grandparents (first generation speakers) generally produce more dialect than speakers from younger generations (second and third generations).4.5.Language Exposure and Interaction: Constructing Multilingual CompetenceIn this analysis, we aimed to investigate the child’s production and the production he is exposed to in relation to the amount of time the child spends with his interlocutors. We thus focused on child-addressed speech and on the child’s production when he engages in conversations with members from his nuclear family (his parents) and with members of his extended family (his grandparents and his aunt). We analyzed the lexical choices effectuated during five of eleven days in which the family members gathered together.In Table 5, we report the child’s and his interlocutors’ productions.Dialect is least frequently selected. When the parents and the child speak to one another, their dialect production fluctuates and does not follow a particular direction (it is not clearly on the increase nor on the decrease across the 5 days). Their dialect production alters significantly throughout the eleven days both for the parents (Kruskall-Wallis tests: H = 2, p = 0.02) and for the child (H = 5.6, p < 0.0001). Looking at the child’s production and the grandparents and aunt’s productions when they engage in conversation, we highlight two elements. First, the child produces the most dialect when he speaks to these interlocutors; his production reaches 12.7% (Day 10), which is the highest percentage of dialect usage attested in the child’s production. Second, both the child and his interlocutors increase their dialect production gradually but steadily, as if they were mutually influencing each other’s productions, showing a convergence in usage: The productions of the grandparents and aunt reveal a significant curb (H = 5.9, p = 0.006) although there is a tendency for the child’s to increase (H = 2.6, p = 0.07).Concerning cognates, their production does not change significantly throughout the eleven days. When addressing the child, the parents’ cognate production remains stable between 34.2% and 38.4% (H = 1.6, p = 0.8, N.S.). When addressing his parents, Francesco’s cognate production does not change significantly (H = 5.8, p = 0.2, N.S.) and he produces between 37.5% and 45.7% of cognates per utterance. In conversations between the grandparents, the aunt and Francesco, their cognate production ranges from 30% to 43.4% and does not fluctuate significantly for the grandparents and aunt (H = 6.2, p = 0.2, N.S.) or for Francesco (H = 1.5, p = 0.8, N.S.).Looking at the respective speakers’ Italian production, these generally seem to decrease in the eleven days, although not all change significantly. For instance, when Francesco speaks to his grandparents and aunt, his Italian production tends to decrease (H = 6.8, p = 0.1). Yet, when the grandparents and the aunt speak to Francesco, their Italian production decreases significantly, from 62% to 53% (H = 15.7, p = 0.0003). This is also the case when Francesco speaks to his parents. His Italian production decreases significantly from 60% to 48% (H = 13.5, p = 0.0005). However, when his parents speak to Francesco, their Italian production does not change significantly (H = 4.1, p = 0.4). It remains stable at 60%-65%.We have documented that, during the eleven days of observation, dialect fluctuates the most and tends to increase towards the end of this period. With the exception of the child’s parents, Italian production tends to decrease. The production of cognates remains stable for all speakers through the eleven-day period. The increase in the child’s dialect production may be due to the fact that during these days in which the family gathered together, Francesco was exposed to speakers who generally produce more dialect than his parents.These results show distinct Italian and dialect usages, depending on the interlocutors involved in the interactions. Cognates production manifests less variation through different contexts. To show how these language usages materialize in the interactions, we will now present two interactions (one involving adults and the other involving adults and the child) that illustrate how Italian and dialect are used in the spontaneous data collected.The first extract reproduces a multiparty interaction between the mother, the father, the maternal grandmother, the aunt and Francesco. The child was not directly involved in the exchange although he was present and “interrupted” the inter-adult exchange. The thematic orientation of this extract is language competence in Italian. The mother starts this conversation by saying that nobody in Italy is capable of speaking Italian properly. The father, in a playful tone replies that he is the only one who is capable of speaking Italian well.As can be seen this inter-adult exchange is mainly in dialect. It is important to note the way the different speakers use the languages in their repertoires to put forwards their arguments supporting their language views. The mother, for example, says in dialect that nobody knows how to speak Italian well. When she refers to ‘Italian’ she does not use the dialect form (/italjaɲ/) even though her speech turn is entirely in dialect. She does the same thing in speech turn 3. It is as if the mother wanted to metalinguistically underline the codes she is referring to.Interestingly, despite the use of dialect in the exchange, the aunt also “demarks” her dialect utterance by using the Italian adverb /antzi/. The contrast provoked by the change in codes attracts the attention of the speakers on what she is about to say. She supports the mother’s views but her utterance contrasts the common ideas on dialect’s so-called interference in language proficiency, explaining her use of the adverb /antzi/.In the successive turns, the grandmother’s /mamːa/ attracts the child’s attention and he produces /mama/ with the simplification of the geminate, a common phonetic trait in the dialectalized Italians and italianized dialects (dial. /mare/ versus Ita. /mamːa/ ‘mother’). The item /mama/ is both produced in Italian and dialect contexts, thus it has been classified as a cognate. The grandmother wraps up the exchange with the claim that Venetans, despite their use of dialect, have stronger language competence than speakers from other regions in Italy.In this extract the utterances are mainly in dialect. Cognates co-occur with the dialect words. When the speakers change codes, in this case by switching from dialect to Italian, they may do so to mark, and thus reinforce, their views that they put forward.In the exchange that follows, the change in code is linked not with the subject matter or the views the speaker tries to put forward. Rather, the change in code is in line with the change of interlocutor.These two utterances by the mother were produced in parallel with dinnertime conversations mainly among the adults. She is busy helping Francesco eat and tells him to chew his food properly. In the second turn, she then turns to the father and tells him to leave him alone and to let him eat. Although these two utterances are produced in isolation, it is nevertheless interesting to see the link binding language choice and interlocutor. In fact, the shift in interlocutors coincides with the shift in language choice, showing quite clearly the code choices in inter-adult discourse that were pointed out in the quantitative analyses. In these two utterances, the child is exposed to two variants of the verb ‘chew’, one is Italian and the other is dialect.5.DiscussionThree observations are highlighted. First, in interactions involving the child directly, Italian is the main language selected. Second, in interactions involving adults and not the child directly, utterances contain mainly dialect words. Third, dialect and Italian show greater variation in average production rates than cognates. Dialect and Italian have the widest ranges of usage. The dialect usage bracket stretches from 0.6% to 54.3% and the Italian bracket stretches from 8.5% to 67.2%. Cognates have a limited range, varying from 30.9% to 45.3%. The breadth of the ranges is an indication of the amount of variation observed within one category.The narrow ranges in the use of cognates suggests that their production is relatively stable: it does not vary from one interactional context to another. Three observations support this argument. First, the production of cognates does not vary significantly according to the generation of the speaker (Analysis 4.3). Second, in certain cases, it does not vary significantly between child to adult interlocutors (Analysis 4.4). Third, during the eleven-day period of observation, we noticed that the production of cognates did not vary significantly over time and remained within the 30%-40% usage bracket (Analysis 4.5). The consistent presence of cognates in both Italian and dialect contexts means that in Francesco’s language environment he can encounter cognates that co-occur with either dialect or Italian words. Developmentally, this is important for two reasons. Firstly, the child will be exposed to both dialect and Italian utterances containing the common content represented by cognates. He can then build pivotal schemas containing a pivot – a fixed element, a cognate in our case – and a variable element from either dialect or Italian lexicons. We will illustrate this point with two examples from our multiparty corpus.These two utterances were produced during mealtimes and both were child-addressed utterances. As in the previous linguistic examples, Italian words are in bold, cognates are in italics and dialect words are in unmarked font. We found the corresponding dialect utterances in inter-adult discourse (1b and 2b), the semantic content of which corresponds to the Italian utterances from 1(a) and 2(a).In Utterance (1) and (2), both the Italian and the dialect utterances fulfill the same communicative intention. For example, looking at (1a) and (1b), there is one common part questo, a cognate, and two contrasting elements, respectively cos’è, Italian, and cossa ze, dialect. The child will be exposed to cos’è in the child-directed discourse and to cossa ze in the inter-adult discourse to which he is exposed daily. Through linguistic experience and exposure to inter-adult and child-directed speech, he will gradually understand that cos’è questo and cossa zeo questo are produced to communicate the same pragmatic intentions, that is, the request of information regarding an object present in the context of the interaction. The child will thus match cos’è and cossa ze and use the one or the other after consideration of the appropriate language choices depending on the interactional context.Through the exposure to contrasting input (Italian in child-addressed speech and dialect in inter-adult discourse), the child will be able to associate Italian and dialect constructions that fulfill the same communicative intentions and to extract the single lexical items pertaining to each one. Gradually, Francesco should gain sufficient exposure to be able to recognize and memorize the dialect items found in inter-adult speech or in interactions that encourage his dialect comprehension and production, as was the case in the exchanges with extended family members. Moreover, the child will learn to generalize the construction (cognate/pivot + X) across the other constructions memorized by combining the cognate – the pivot – with any item that can fill the empty slot (X). In this way, cognates facilitate access to the form of dialect items by reducing the differences between the Italian and dialect systems. Boundaries separating the two codes become fuzzier, facilitating the possibility of change from one code to another within discourse. The presence of items that – like cognates – decrease the formal differences between two (or more) codes would seem to support the maintenance of dialect among the new generations. Ultimately, the child will be able to generalize the language constructions across dialect and Italian lexicons.6.ConclusionThese results may not seem new. Scholars working in the Italoromance area have observed that it is precisely through the production of mixed utterances6 that Italy’s multilingualism is maintained (Dal Negro, 2005; Rindler Schjerve, 2003) and that dialects have not fallen into disuse as suggested by statistics on declared usages. Moreover, mixed utterances provide evidence of the dynamic relationship in the Italian-dialect contact situation, where Italian and dialect exert influence on each other, through the processes of regionalization of the national language and Italianization of dialect. Yet, these studies do not explain in detail how mixed utterances support language maintenance. Our study suggests that it is not only due to speakers using both languages in their utterances that minority languages are maintained. Additionally, language maintenance may be supported by similarities in the input and output provided by cognates and the contrasting usages provided by Italian in the child-directed speech and dialect in inter-adult discourse. This may facilitate semantic access to dialect – the minority language – and could thus scaffold dialect acquisition. Psycholinguistic research on multilingual lexicons has yielded important results supporting our hypothesis on the potentially important role of cognates in language learning. A robust body of research which has shown that the semantic and formal proximity between cognates facilitates lexical access in reading, listening and production (Caramazza and Brones, 1979; Costa, Albareda, and Santesteban, 2008; Costa, Santesteban, and Cano, 2005; Cristoffanini, Kirsner, and Milech, 1986; Lemhöfer et al., 2008). In a recent article, Dijkstra and colleagues (2010: 297) conducted a study on the cognate processing in English-Dutch bilinguals. Interestingly, they found that “a facilitation effect did arise for lower frequency identical cognates relative to other items […]. This finding indicates that, as might be expected, identical English low-frequency cognates profit more from the co-activation of their Dutch counterparts than high-frequency cognates do”. Reading our results in light of the psycholinguistic studies mentioned above, we could hypothesize that cognates may help process low frequency dialect words. Hence, exposed to contrasting input in the multiparty interactions, Francesco’s dialect production may have been facilitated by the cognates co-occurring with the dialect produced in his language environment.The results and hypotheses discussed in this study should be tested with more quantitative and qualitative analyses conducted in similar – but larger scale – sociolinguistic contexts, for instance Creole or Spanish/Catalan contact situations. More importantly, future quantitative studies on development in contact situations should attempt to include variation phenomena in order to better represent the language environment and to gain a wider understanding of the stages of acquisition occurring in language contact situations. We have attempted to do so through the establishment of the cognate category. Much work remains to investigate the cognate category in more detail, for example from a morphosyntactic perspective. The outcomes of this research could also provide new methodological ideas for studies conducted on naturalistic data as well as elements for the debate on the transmission of minority languages and the pathways leading to the construction of multilingual competence in informal language learning contexts.

Journal

Journal of Language ContactBrill

Published: Jan 1, 2013

Keywords: language contact; dialect; Italian; cognates; acquisition; transmission

There are no references for this article.