What to cover, and how?

• History/'the big problems' ?
• Focus on current best-understanding?
  • or a small number of relatively detailed topics
    • E.g. Categorisation
• Understanding methodology as a primary goal
• Which hyphenation?
• Top-down vs. bottom-up
  • Start with e.g. cross-modal priming
    • McGurk effect
• Edinburgh perspective
• Extended mind/internal vs. external representation/cog sci and design (Don Norman, doors, affordances)
  • Cog Sci for CS -- designing and evaluating GUIs -- depth of menus
• Acquisition ?
• Corpora, gold standards, supervised vs. unsupervised (or a lab)
• Themes
  • Representation
  • Modelling
    • Propositions vs. Statistics
• Philosophical topics as (mostly) tutorial topics
  • Near the end
  • One lecture or two
  • Turing test (the real one), morality/humanity, consciousness (not)
• Abstraction/mind vs. brain at the end, if at all
• Broader brush in lectures, focus on two or three topics in more detail in tutorials
  • Students come to tutorial with different information
    • So they have to talk to each other
• Kumar's three levels [ref?]
• Cognitive Science as reverse engineering
• How much does embodiment matter?

"Like linguistics, but more useful
Like computing, but more interesting" -- Jon Oberlander

"There will be no interesting work on language generation until we have a computer that has something to say" -- Dave McDonald, maybe

-- HenrySThompson - 25 Aug 2010