Intonational collocations, tropes & memes

  Hi all!
I'm trying to do more with the wiki again. It's just there's so much material. Do I start with Bolinger? Where do I start with the wiki? I wish I could get some sense of the overall structure. I think I will have to begin with ToBI's structure anyways. Arnold's notation is somewhat convenient but it isn't as interesting as it should be. Bolinger has 3 basic structures but lots of micro prosodic variation that doesn't show up again 'til we get back to IVIE or RAP e.g. pitch shifting.
  It's hard to get to a critical mass. I want to get enough going that other people can run with it and see it develop a life of its own...but other people won't do your work for you. I'm also concerned about making sure I don't borrow too much in fair use. Fortunately most of the ideas are easy to extend.
  What can other people do? A lot actually. Intonation should be intuitive; lots of people remember movie lines and can come up with more examples pretty easily when they hear one example. For instance listening to the 3rd column (HL%) here it is pretty easy to tell that the overall theme sounds very much like boredom or exasperation/irritation. It's not much stretch to hear a little kid saying some of those things, e.g. Ralphie from "A Christmas Story" or some of the sassy girls from "Mean Girls"; and from there to produce a specific quote someone can find via youtube or by renting a movie.
 Another concern is easily loading and hosting audio files from my device. SoundHost doesn't cut it because they want to play all the sounds in your album instead of just one at a time.
  I've found a very good book called "Prosody & Meaning" which has a fantastic article about lexicalization of pitch contours. As it turns out, they're essentially memes. This changes a lot of things, if micro prosodic variation is really so shifty. It means there's likely to be no underlying structure, and the rest is fashion. But, the authors point out that the same system that operates on focus and topicalization phenomena operates on the discourse functions, which include back channel responses, tags, etc. Which means that, very frequently, structures reappear just because without much meaning. They give the example of "meatball" and "school" showing up often with stereotyped patterns.
  They discuss other things like upstep and how it probably belongs in ToBI, and how pitch scaling may or may not matter. Other articles discuss how well-formed and discrete categories in intonation actually are. Turns out, not very. They're very slippery, and in fact TOBI categories appear to bleed into each other. Hence the usefulness of a exemplar-based theory of lexicalization, where lexemes come with a few available patterns. It also makes me wonder whether its worthwhile having a theory only so far as it helps organize the data. Too bad the data is so thorny, otherwise we might already have a TVtropes-styled wiki...Oh wait, I made one (it's just not useful yet.)
  Cheers!