| 1. |
|
|
| 2. |
- La Monica, Clelia
(författare)
-
Measuring perceived accent difference using oddball paradigm : acoustic distance vs. familiarity
- 2015
-
Konferensbidrag (refereegranskat)abstract
- This exploratory study investigates the relationship between acoustic distance and ERP amplitude: specifically, whether the distance between different pronunciations of the vowel /o/, investigated in terms of glide-weakening in American accents, is related to amplitude at the MMN and P3a time frames (here 150-250 and 250-350ms post-stimulus, respectively), as well as the possible “overriding” effect of accent exposure/familiarity. Using a multi-feature oddball paradigm, it was expected that the more acoustically different a pronunciation is from the standard, the higher the resulting amplitude will be for the ERP. Secondarily, this study aimed to see if this case would be especially clear in vowels marked as either dialectal or ‘non-native’.Following the oddball paradigm, five recordings of the word 'goat' were used as stimuli:General American served as standard, while Inland South, Inland North, and Swedish-accented were the main deviants, and New-England was added as a non-target. The vowel /o/ was used due to its amount of variation in pronunciation, and markedness as dialectal, and participants were instructed that they would hear an English word. Using Euclidean distance from the averaged 20% and 80% points of /o/'s F1/F2 coordinates (then normalized into Bark perceptual scale), acoustic distances were established between the recordings (which were synthesized from actual speech to match in duration, pitch, and volume). In terms of acoustic distance, Inland North was furthest from GenAm, Swedish slightly closer to GenAm but moresimilar to Inland North, and Southern was very similar to GenAm (but very distant from both Swedish and Inland North; see fig. 1).Participants in the preliminary study were American expats living in Sweden, from various regions in the USA who were all highly-mobile. Therefore, it was of interest whether ERP’s were affected more by acoustic distance of an accent to the standard, or the heightened exposure to Swedish-accented English. Initial results indicate a large difference between the perception of monophthongized /o/ and diphthongized /o/, as shown by a high P3a (see fig. 2).Inland North was the most different in terms of max amplitude at 250-350 ms, with Swedish second and Southern being closest to the standard. Most interestingly, the Swedish pronunciation yielded a less positive P3a than Inland North, even though it was of similar acoustic distance to the standard, and a more negative MMN, showing possible later processing of ‘familiar’ Swedish-accented English. While this study is only preliminary, it points in the direction of further research into the effect of language exposure on adaptation and perceptual learning vs. acoustic distance.
|
|
| 3. |
- La Monica, Clelia, 1987-
(författare)
-
The role of urbanization on dialect distribution in US English
- 2018
-
Konferensbidrag (refereegranskat)abstract
- The United States is often juxtaposed with nations such as the U.K. and Netherlands when discussingthe effects of its relative youth, rural expansiveness, and diversity on levels of dialectal variation. Thedistribution of urban centers, however, is not uniform, instead following a longitudinal pattern alongthe coasts and major river systems in the United States. This is in contrast to dialect isoglosses withinthe U.S., which tend to run latitudinally as a result of East-West expansion (Labov et al., 2006). Thisstudy therefore explores dialect variation in US English using dialectrometric methods (Nerbonne etal., 2009), comparing regions that have a higher concentration of urban centers (e.g. New-England)with those more sparsely populated (Midland). Using data from the Harvard dialect survey (Vaux &Golder, 2003), in which 30,788 participants from 50 states were presented with multiple choices forpronunciation, lexical, and syntactic preference, a factor analysis was made to determine whichlinguistic features relate to variation geographically (Grieve et al., 2011) and string-edit distanceswere computed between the most preferred variants to find the level of phonological differencebetween regions. Consideration was also given to the factors of ‘urbanization’ (population per sq.mile), regional ‘age’ (year of colonial establishment), and ‘linguistic diversity’ (U.S. Census, 2011).Preliminary results indicate that, as expected, regions with highest population densities contain morevariation with respect to pronunciation preference. However, despite the higher concentration ofcities on the West coast, variation was less than expected, likely due to its high diversity and morerecent establishment.
|
|
| 4. |
- La Monica, Clelia
(författare)
-
United States accents compared : the relation of acoustic distances and perceptual tasks
- 2015
-
Ingår i: The Journal of Acoustical Society of America. - : Acoustical Society of America (ASA). - 0001-4966.
-
Konferensbidrag (refereegranskat)abstract
- This project investigates correlations between accent production and perception by comparing two sets of data: acoustic distances of American dialectal speech samples, measured using Euclidean distances of each vocalic stimuli's first three formants across a trajectory, and perceptual data in the form of online survey material and EEG tests which judge the differences between the same varieties. The speech samples used consist of six sentences from eight regions of the United States, each sentence containing phonological features that may be marked as perceptually relevant for dialect classification.Here we examine the preliminary outcomes from such perceptual tests performed by naïve listeners, which include free classification, identification, perceived difference and ranking similarity, as well as attitude judgement tasks. Additionally, EEG tests were carried out to evaluate the relation between acoustic distance of accents and ERPs. Of special interest here is the relation between a non-regionally specified American accent and others. This “standardized” variety is judged in comparison with each of the regional accents in order to investigate listeners’ perceptions of non-regional vs. regional accents, and in turn the correlation between measured accent (dis)similarity and perception.
|
|
