@inproceedings{8733937,
  abstract     = {{In the last two decades the literature has described the intonation of several varieties of spoken Italian (a.o., [1]), outlining an extremely complex picture characterized by a pronounced variation both within and across varieties. Besides, the relation among geographically close varieties has been scarcely investigated. Recently, [2] considered dialectological isogloss boundaries as a starting point to study geographically close varieties of Italian. Along these lines, we propose a first experimental investigation exploring closely related varieties of Italian spoken in Campania, namely Salerno (SI) and Cilento (CI) Italian, that will be compared to Neapolitan Italian (NI). On a dialectological basis, these points are distinguished by the Eboli-Lucera isogloss ([3],[4]), running south of Salerno; namely, the varieties have Campania (Naples and Salerno) and Lucania (Cilento) vernaculars as substrate. In this study we explore the realization of pitch accents in sentence-initial noun phrases (NP) in SI and CI. We consider statements under three pragmatic conditions: (1) Regular Topic (RT) in exhaustive answers, (2) Partial Topic (PT [5]) in non-exhaustive answers, (3) Contrastive Focus (CF) in corrective answers. Such conditions have been studied for NI in [6], [7] showing the presence of a phrase break between NP and the verb phrase (VP) in PT and CF, and differences in pitch accents’ alignment and span among the three conditions. Differences between RT and PT peak alignment were previously observed in a qualitative comparison between SI and NI. Here we carry out a replication study to enlarge the picture of Campania Italian to SI and CI to identify similarities and differences across diatopically close varieties. 
SI and CI speakers performed a reading task eliciting NP +VP target sentences (ex. 1), with NP (CV’CVCV) in 3 conditions: RT, PT, CF. Due to the covid-19 pandemic, a remote data collection was carried out via Zencastr (.wav, 44100 Hz,16 bit; [9]). The final dataset consists of 480 items: 10 target utterances * 3 conditions (RT, PT, CF) * 2 renditions * 4 spks * 2 varieties (SI, CI). The sample was segmented using Webmaus [10]; a set of Praat scripts was used to extract duration and f0 measurements (in Hz) to calculate peak latency (in ms), span (in ST) and slope. Data were analyzed using mixed-effect linear regression models in R [11].
Our results show that the three conditions are phonetically encoded in a similar way in the varieties of SI and CI (Figure 1) and in line with NI [7]. Both in CI and SI RT and PT are realized as a rise, reaching the peak late in the vowel. The difference between these two conditions lies in the span significantly wider in PT (Est: 2.9, p< .001), and in the slope of the rise, steeper in PT (Est: 9.9, p< .001). CF, on the other hand, is realized as a rise-fall movement, with the peak aligned earlier than both PT (Est:-140.9, p< .001) and RT (Est: -122.8, p< .001). For the other parameters, CF is less clearly differentiated from the two other conditions: a steeper slope groups together CF and PT viz RT (Est: 17.5, p< .001); however, along the dimension of span, CF is not significantly different from neither PT nor RT. As for duration, the NP-final syllable of PT is lengthened compared to RT (Est: 36.4, p< .05), suggesting the presence of a break in PT between NP and VP, although in [7] a lengthening of also the NP-final vowel was found for NI. A number of phonetic differences among the varieties (SI, CI and NI) can be noticed. Firstly, the difference in the early (CF) and the late (RT and PT) peak alignment is kept constant across SI and CI (Figure 2), though in CI the position of the peaks is overall earlier with reference to SI (Est: -33.2, p< .01). A second difference emerges from the comparison of SI and NI, and confirms the preliminary results of [8]. In SI PT peaks are aligned in the post-tonic syllable (73.2%), while RT ones are aligned within the stressed syllable. In contrast, NI data show this realization is reversed. Overall, the three conditions are distinguished within each variety by combinations of different phonetic features. However, differences in peak alignment indicate diatopic variation, in that SI, CI and NI differ in the phonetic implementation of the contrast among PT, RT and CF.

(1)	Milena (NP) lo vuole amaro (VP).
Milena (NP) takes it (the coffee) black (VP).
 
Figure 1. F0 curves of the topic NPs according to the three information conditions of RT (red line), PT (green line), CF (blue line) in CI (left panel) and SI (right panel).
(add figure)
 
Figure 2. Time in ms of pitch accent peak alignment taken from the stressed open syllable offset according to the topic condition (RT, PT, CF) and the variety (SI, CI).
(add figure)

[1] Gili Fivela, B., Avesani, C., Barone, M., Bocci, G., Crocco, C., D'Imperio, M., Giordano, R., Marotta, G., Savino, M. & Sorianello, P. 2015. Intonational phonology of the regional varieties of Italian. In Intonation in Romance. Oxford University Press, 140-197.
[2] Gili Fivela, B. & Nicora, F. 2018. Intonation in Liguria and Tuscany: checking for similarities across a traditional isogloss boundary. In Vietti, A., Spreafico, L., Mereu, D.  & Galatà, V. (Eds), Il parlato nel contesto naturale, Studi AISV, Milano: Officinaventuno, 131-156
[3] Avolio, F. 1989. Il limite occidentale dei dialetti lucani nel quadro del gruppo ‘altomeridionale’: considerazioni a proposito della linea Salerno-Lucera. L'Italia dialettale, 52, 1-22.
[4] Avolio, F. 2000. Ma nuje comme parlamme? Problemi di descrizione e classificazione dello spazio dialettale "campano". Romance Philology, 54(1), 1-28.
[5] Büring, D. 1997. The Meaning of Topic and Focus – the 59th Street Bridge Accent. London: Routledge.
[6] Brunetti, L., D'Imperio, M., & Cangemi, F. 2010. On the prosodic marking of contrast in Romance sentence topic evidence from Neapolitan Italian. In Proceedings of the 5th International Conference on Speech Prosody, Chicago.
[7] D’Imperio, M., & Cangemi, F. 2011. Phrasing register level downstep and partial topic constructions in Neapolitan Italian. In Gabriel, C. & Lleó, C. (Eds.), Intonational phrasing in Romance and Germanic: Crosslinguistic and bilingual studies, Amsterdam: John Benjamins, 75–94. 
[8] Orrico, R. 2020. Individual variability in intonational meaning identification: The role of cognitive and sociolinguistic variables. PhD thesis.
[9] Magistro G. (in prep.). Prosodic cues to syntactic reanalysis. Experimentally tracking Jespersen's cycle in progress. PhD Diss. Ghent University.
[10] Kisler, T.,Reichel, U.D. & Schiel, F. 2017. Multilingual processing of speech via web services. Computer Speech & Language, 45, 326–347.
[11] R Core Team 2019. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. URL https://www.R-project.org/}},
  author       = {{Cataldo, Violetta and Orrico, Riccardo and Crocco, Claudia}},
  booktitle    = {{Phonetics and Phonology in Europe (PaPE 2021), Abstracts}},
  keywords     = {{Italian,Campania,closely related varieties,Salerno Italian,Neapolitan Italian,Cilento Italian,topic accent,focus accent}},
  language     = {{eng}},
  location     = {{Online}},
  title        = {{Topic and focus accents in closely related varieties of Campania Italian}},
  year         = {{2021}},
}