First question session (after Chandan and Grant's presentations)

Mary Beckman asked Chandan whether he had considered an interaction with prosodic context, given that since IDS has shorter phrases, there are more boundaries and hence an interaction with other phenomena which makes pitch more important
Robert Daland seconds this question and points to the fact that more frequent boundaries should raise the number of stops at phrase-edges which may change the cue weighting (e.g. because of strengthening)
Amanda Seidl asked Chandan whether perhaps speech directed to younger babies would have cleaner categories given Sundberg's proposal of under/overspecification in younger infant-directed speech [I add here that this proposal suggests that vowels are overspecified - that is, have more marked acoustic categories - at younger ages, while consonants are underspecified then and overspecified later; and also that there is some counterevidence to these claims - see Englund & Behne, 2006]
* A member of the audience supports the idea that the f0-voicing association is dependent on the speaker's control, given that there are languages such as Spanish where listeners have no f0 sensitivity in this context, unlike English and Japanese listeners
* Another possibility arises from thinking English as a fortis/lenis rather than voicing language - note that the voiceless (marked) category was more coherent than the voiced category.
Peter Richtsmeier asks Grant whether attention to vocalic cues is a sort of 'default'. Grant points out that what is remarkable is that English listeners were ignoring fricative noise, given that they have a similar contrast in English where they do use cues in the fricative portion.
Mary Beckman points out that the Mandarin alveopalatal-retroflex contrast is the result of the phonologization of a diachronic sound change.
* Is there a block effect in the labeling task? Grant answers that English listeners are not integrating cues with this added experience, but just getting better at discriminating.
Matthew Goldrick asks whether integration of cues is the 'default' and we learn to separate cues or whether it is the other way around [On this topic, Tomiak, Mulenix and Sawusch (1987) show that adult listeners only show the 'integration' effect only when they hear the sounds as speech, but not when it's non-speech - which suggests that perhaps we integrate through over-learning. Although learning to integrate is probably only applicable to cues that are not 'inherently' processed together - e.g. integration should be the default for pitch and loudness.]