Thoughts on Distributional Partitions There's two important things that profoundly effect the theory of DP --

1. Parducci's (1965) Range Frequency Theory
2. Huttenlocher, Hedges and Vevea's studies about relationships between estimation bias and distributions

Things that I have to do at this point --

1. Jessica Choplin: Given the understanding of Range Frequency Theory and her Comparison Distortion, convince her that the question of partition imposed by gradable adjectives is not yet understood, and we need to redo Parducci with symbolic labels instead of numeric scales. And of course, ask her for linguistic references
2. Jeff Loewenstein: Give him a summary of distributional/structural partitions, and try to determine the existince of both distributional and structural partitions in the domain of price -- expensive/cheap -- when are they distributional/structural (are there other symbolic partitions people will make up?) and their mappings. Manually, build a spreadsheet with some product domains and plot distributions for writeup
3. Janellen Huttenlocher, Larry Hedges: Check their earlier work and esp Hedges contribution to similar ideas. Then talk to them, maybe via Dedre, as Huttenlocher might make for a perfect person on the committee if that will work. Talk to Huttenlocher and try to convince her that how large/med/small depends upon distributions is a very important question that has to be answered.
4. Jose Quesada: Reply him and thank him for the emails. What could be a paper that me and him could write. Will he be interested in doing a LSA kind of analysis for providing some linguistic/statistical evidence for what gradable adjectives mean?
5. Check out the refs from Liz Spelke's paper: a lot of them should be online, get them.

Thesis proposal: Begin writing the proposal. Parts of the proposal that depend upon collaborative work might remain a little vague at this point, but write the proposal around them. Get a better picture of what could be accomplished in these collaborations. Method for finding people's partitions: How general is Huttenlocher's claim that people bias towards category centers? If I can assume that that is true, then we can find what categories people make. In the studies that she was doing, people were making only one category (that's her interpretation, and thats what the data suggests as well). Maybe we could instruct people to go through exactly the same experiment and make them make large/medium/small categories. In that case we will find that their bias curves will look much different as depending upon whether they judge a stimus to be l/m/s, they will be biasing towards the center of that category. First thing is to get this effect. The second thing will be then to go back from bias curves to finding where their category centers are located and where the boundaries are (do we have to worry about something else like boundary repulsion? check out literature. also send an email to Alicia Hund). And the second thing will be to see how much intra-subject agreement exists. Lookup Dedre's email about other Huttenlocher papers and then write her an email with a sneak preview of what I am thinking, the paper and ask for a meeting time. Library list: Hedges older work. Older refs from Spelke. Life list: Call home. Call Sprint. Make Apartment Ads.