Desk 2 presents the AusE and you may PS mix-words category investigation rates of the very most constant Dutch vowel group in order to an enthusiastic AusE term and PS vowel

If listeners’ L1 vowel inventory dimensions has an effect on non-native discrimination difficulties, AusE audience try predict so you’re able to surpass PS audience total within discrimination of Dutch vowel contrasts. All the latter training that used LDAs as an easy way away from testing the brand new predictive character off listeners’ L2 feeling simply looked at audience whoever L1 vowel list are smaller compared to regarding this new L2. Thus, i further utilized LDA models to check on whether or not acoustic similarities try predictive of categorization habits from the audience with a smaller and you may large vowel list than the address vocabulary (Strange et al., 2004, 2005; Gilichinskaya and Uncommon, 2010; Escudero and you can Vasiliev, 2011). Just like the revealed throughout the Methods point, this type of analyses model AusE and you can PS listeners’ likely classification designs out of Dutch vowels, and as a result predict its most likely discrimination troubles (Uncommon ainsi que al., 2004, 2005; Gilichinskaya and Unusual, 2010; Escudero and Vasiliev, 2011).

Desk dos. Percentage of Dutch vowel token classification in order to an enthusiastic AusE word and you can PS vowel predicated on total class designs of get across-words LDA.

In addition, this new L2LP design posits when distinguishing anywhere between L2 classes, listeners apply multiple sourced elements of acoustic-phonetic guidance inside their impact from phonological areas (Escudero, 2005). Earlier research has in reality displayed one attention is actually repaid so you’re able to the most salient acoustic cue regarding a certain sound (get a hold of Curtin et al., 2009; ). Continue reading →