TL;DR
This paper investigates "post-aspiration" – a puff of air after a consonant sound – in voiceless geminate stops (/pp/, /tt/, /kk/) in Italian. Contrary to established linguistic descriptions, the study finds that post-aspiration is a surprisingly common acoustic feature across various Italian regions, not just a localized phenomenon in Calabria as previously thought. This challenges existing phonological models of Italian and suggests a broader acoustic reality that needs further investigation.
What Problem Does This Paper Address?
Traditional descriptions of Italian phonology state that voiceless stops (/p/, /t/, /k/) are unaspirated, meaning they are produced without a noticeable puff of air. The distinction between voiced (/b/, /d/, /g/) and voiceless stops is primarily based on voicing, not aspiration. This is in contrast to languages like English or German, where post-aspiration is a key feature of voiceless stops. While some recent acoustic studies have hinted at the presence of aspiration in Italian, particularly "preaspiration" (air before the consonant), the phenomenon of "post-aspiration" (air after the consonant) in voiceless geminate stops (/pp/, /tt/, /kk/) has been largely overlooked or considered absent in standard Italian. This paper aims to provide acoustic evidence for post-aspiration in these sounds across different Italian regions and to explore whether regional variations influence its occurrence and perception.
Methods
The researchers analyzed recordings of isolated Italian words containing voiceless geminate stops (/pp/, /tt/, /kk/). They used acoustic phonetic analysis software (Praat) to measure the duration of different parts of these sounds, specifically focusing on the "release" portion, which includes the burst of the stop and any subsequent aspiration (post-aspiration). They measured the duration from the start of the burst to the onset of voicing for the following vowel. The study also measured preaspiration (air before the consonant closure) and the closure duration itself. They compared these acoustic measurements across speakers from 15 different Italian cities to identify regional patterns.
Datasets
The study utilized a portion of the CLIPS corpus of spoken Italian. This corpus contains recordings of isolated words read by speakers from 15 Italian cities: Bari, Bergamo, Cagliari, Catanzaro, Firenze, Genova, Lecce, Milano, Napoli, Palermo, Parma, Perugia, Roma, Torino, and Venezia. The specific words analyzed were: bocca ('mouth'), bottoni ('buttons'), cappello ('hat'), macchina ('car'), occhi ('eyes'), specchietto ('little mirror'), and tetto ('roof'). Each word was read once by eight speakers from each city, resulting in a total of 935 analyzed tokens after some were excluded due to background noise.
Key Insights and Ideas
- Post-aspiration is widespread: Contrary to the expectation that post-aspiration is limited to specific regions like Calabria, the acoustic data reveals it's a frequent feature across many Italian cities, even in what is considered "standard" Italian.
- Challenging phonological descriptions: The presence of significant post-aspiration challenges existing phonological models that describe Italian voiceless stops as unaspirated. This suggests a mismatch between theoretical descriptions and actual speech production.
- Regional variation exists, but is nuanced: While Calabria shows distinct patterns, particularly regarding the interaction of pre- and post-aspiration, other cities also exhibit considerable post-aspiration. This suggests that the perception of post-aspiration might be more complex than just its acoustic presence.
- Preaspiration might influence perception: The study hints that the presence of preaspiration might hinder the perception of post-aspiration, especially in regions like Catanzaro where preaspiration is less frequent. This interaction warrants further investigation.
- Consonant place matters: As expected, velar stops (/kk/) consistently show longer aspiration durations than alveolar (/tt/) and bilabial (/pp/) stops, aligning with cross-linguistic tendencies.
Key Results
- The average release (burst + post-aspiration) duration for voiceless geminate stops across all 15 cities was 47.8ms, significantly longer than expected for unaspirated sounds.
- VOT durations were longest in Catanzaro (average 52.9ms), as anticipated, but surprisingly comparable to cities like Milan, Palermo, and Rome (averaging around 19-24ms in previous studies, but closer to 52ms in this study's benchmark comparison).
- VOT durations were consistently longer for /kk/ than for /tt/ and /pp/, with post-tonic /kk/ showing longer durations than pre-tonic /kk/.
- In Catanzaro, post-aspiration was found to be particularly prominent for /tt/ compared to other cities.
- Preaspiration was significantly less frequent in Catanzaro (3.2%) compared to the average across other varieties (30%).
Limitations and Future Work
- Perceptual studies needed: The study highlights a discrepancy between the acoustic presence of post-aspiration and its reported perception by native listeners (who generally only perceive it in Calabria). Future work should include perceptual tests to understand how listeners interpret these acoustic cues.
- Role of stress and vowel duration: While the study analyzed isolated words, the perception of consonant length and stress in Italian also depends on the preceding vowel duration. Future research should investigate these factors in connected speech.
- Broader phonological implications: The paper calls for acoustic phonetic evidence for voiced geminate stops (/bb/, /dd/, /gg/) to determine if they are pre-voiced or if the +/- [spread glottis] feature is sufficient to explain the voiced/voiceless distinction in Italian.
- Historical development: The authors suggest exploring whether post-aspiration is a recent development in Italian, potentially influenced by preaspiration, similar to theories proposed for Spanish.
Summary updated