PERCEPTION OF QUESTIONS AND STATEMENTS IN NEAPOLITAN ITALIAN Mariapaola D’Imperio* and David House** *Department of Linguistics, Ohio State University (USA) **Department of Linguistics and Phonetics, Lund University (Sweden)   Presented by: Maria del Mar Vanrell Bosch

Temporal alignment of the high level target with the syllable or presence of a rising or falling melodic movement within the syllable nucleus? Quina és la principal pista per a la distinció entre interrogatives i declaratives.

1. INTRODUCTION 2. METHOD 3. RESULTS 4. DISCUSSION 5. CONCLUSION

1.             INTRODUCTION In standard and in the Neapolitan variety of Italian, unmarked declarative sentences are characterized by a falling nuclear accent (H+L*). Results from a previous study: the accent peak is later in statements marked by narrow focus (H*). The shape of the canonical nuclear accent in Neapolitan yes-no question is L+H* (reached later in the stressed vowel than in statements).

INTRODUCTION It has questioned the assumption that melodic contours are exclusively encoded in terms of targets. In regions of high spectral stability tonal movement can be perceptually encoded as a contour (rise or fall).

INTRODUCTION Two possible hypothesis are tested about what determines the percept of a statement as opposed to a question: 1. The latency of the high target (peak) relative to the onset of the accented syllable is the main cue for a category shift. 2. The difference between a H* and a L+H* could not simply concern the temporal alignment of the high level target with the syllable but could also depend on the presence of a melodic rise for questions and a fall for statements. .

2. METHOD Two tokens of the sentence “Mamma andava a ballare da Lalla” (Mom used to go dancing at Lalla’s) produced with intended narrow focus on the last word by a native female speaker of Neapolitan Italian were selected from a corpus of read speech. Statement (peak occurs early in the accented vowel)/ yes-no question (peak at the end of the vowel).

METHOD Primer requadre: contorns originals. Segon requadre: contorns normalitzats. Línia vertical: onset de la vocal.

From the declarative base stimulus METHOD From the declarative base stimulus five stimuli shifting the peaks/ shifting the fall in four steps of equal duration (33 ms) forward through the vowel Es creen 20 estímuls.

From the interrogative base stimulus METHOD From the interrogative base stimulus five stimuli shifting the peaks/ shifting the fall in four steps of equal duration (35 ms) backwards through the vowel En la interrogativa es va de cul enrere!

METHOD 20 x 4: randomized in sequences of twenty and presented in blocks of ten (eight blocks). Participants: twenty native speakers of Neapolitan Italian (between 23 and 34). They were asked to indicate after each stimulus if it sounded most like a statement or a question.

3. RESULTS

RESULTS Decl-peak and inter-peak responses: stimulus set present a sharp category boundary at the middle stimulus. Decl-fall responses: largely declarative, ambiguous as the fall is moved late in the vowel. Inter-rise responses: majority of interrogative responses, more interrogative responses as the rise became more prominent through the vowel.

RESULTS

RESULTS The most consistent statement responses: decl-peak 1 and 2, and the decl-fall 1 and 2 (stimuli where there is a clear fall through the vowel). The most consistent question responses: decl-peak 4 and 5, the inter-peak 4 and 5 and the inter-rise 3, 4 and 5 (stimuli where there is a clear rise through the vowel). Mirar requadres diapo 9 (hi ha una baixada clara en els estímuls 1-2 (decl. Peak and fall). Mirar ara 9 i 10.

4. DISCUSSION The second hypothesis is confirmed: a rise in the vowel is the most important cue for the question and a fall in the vowel is the most important cue for the statement. Movent el pic de l’accent dins la vocal accentuada serveix per passar d’una categoria a altra. En canvi, eliminar la pujada o la baixada de la vocal, col·locant-la a la següent consonant incrementa l’ambigüitat. Mirar diapo 9 i 14.

DISCUSSION Since the duration and the timing of the rise in the inter-peak and the inter-rise stimuli of like number was the same, how can the different outcomes be explained? The timing of the peak is the same for equal stimulus number in the inter-peak and in the decl-peak series, why do we notice a bias for inter-peak to receive more questions responses? Subject responses were overall less consistent in the inter-peak results than in the decl-peak results. L’inter-rise, conté una caiguda, el començament i el final de la qual són realitzats sempre dins els límits de la coda sil·làbica. Això fa que sigui un mal marcador de declarativa. El decl-fall es caracteritza pel mateix. Decl-peak: ascens més empinat. Durada de l’onset en el decl-peak és més grossa. Potser per l’inter-peak, la regió d’inestabilitat espectral no és suficientment gran com per fer percebre l’ascens. Inter-peak: áscens i descens d’igual inclinació i mida (pistes conflictives). Decl-peak: ascens és sempre més ràpid i empinat que el descens (no hi ha conflicte entre les pistes).

DISCUSSION They propose a unified analysis of statement pitch accents as consisting of an HL sequence independent of breadth of focus: H+L* for broad focus utterances and H*+L for narrow focus ones. They propose an unique analysis of the question pitch accent as L+H*.  

5. CONCLUSION The primary cue for questions in Neapolitan Italian is a rise through the accented vowel, while the primary cue for statements is a fall through the accented vowel. These results add evidence in support of the perceptual importance of pitch movement through areas of spectral stability for identifying tonal categories. These results have implications for the labeling of tones in Neapolitan Italian and are seen as an important first step towards a more extensive perceptual study of pitch accent categories in Neapolitan Italian.