Tuesday, Feb. 24, 2004
9:30-10:30 meeting with Steve Larson
------------------------------------

Notes:
1. Review of musical examples 3-7, with comments and alternate expectations.
(see notes on the papers).

     Come up with answers for Steve's expectations. (example 6)

   Ex. 3.  Continuation DCBA.

   Ex. 4.  Stop on the C; generate a stopping rule that predicts this.

   Ex. 5.  -Compound melody: continuation EF +1 DCBC.
           -Note: compound melody is an open challenge that needs to be
                 integrated into SeekWell.
           -Note2: My concept of "deceptive cadence" was wrong: a deceptive
                  cadence involves dominant->tonic harmony in the melody
                  but dominant -> non-tonic progression in the bass.

   Ex. 6.  -Try the continuation in m. 4 : GAGABDC, in lower octave.
           -My note: the last note of measure 2 is less constrained.  Key
           question: explain this and integrate into model.
           -Rule of "3 times": don't include this idea for now.  Note
            that it can be misleading; in this example the 4 measues can
            be perceived as 2 groups 2 measures long, and then we don't
            run into the "3 times" rule at all; the 4th measure sounds fine.

   Ex. 7.  - Introduces the concept of the harmonic rhythm speeding up in
             the penultimate measure (a common technique).  Note that the
	     harmonic pattern is I-V-I-V-I, one chord per measure, but to 
	     finish in m.4 it speeds up to: I-V-I-(V-I), with the chords
	     in parentheses both taking place in measure 4.
	   - This idea in turn leads to the idea of having multiple inertial
             descriptions in place at the same time, where these different
             descriptions exert conflicting pressures.  Consider the 
             preferences for 4-bar phrases as conflicting with the regular
             I-V alternating progression described above and the 
             pressure to change chords every 2(?) beats, leading to this  
             speeding up before the cadence.
           - Note that "following intertia" simply refers to 
	     "continuing to do the same thing", where this can apply to 
	     one level or to multiple levels of the hierarchy at once, 
	     and multiple descriptions may be active and in conflict.
  


2. A theme: for now, we'll ignore ideas of saturation (the expection of 
suprise).  We'll also ignore explicit mention of harmony, consistent with
the Schenkerian view of harmony deriving from musical forces.


3. To Do:

- Readings:  Start with Westergaard.  Then read GTTM, and finally Narmour.

- Musical Development:
  a) Play piano and sing every day for at least 20 minutes or so.
  b) Write down these activities in a notebook each day
  c) Improv. exercise: try playing a bass note / pattern of notes, and then 
		       find things to play above that bass.  Try singing
		       a pattern and then finding it on the piano.

- SeekWell:
  a) Perform clarification of some existing examples, working toward a 
technical report.  Include our previous discussions in the new descriptions.
Write an introduction briefly going over the theory of musical forces, 
and then move into an illustration of ~two examples of the forces at work.

  b) Pick simple, clear pieces from 18-19th centuries (Bach through Chopin),
and find places where the "ability to predict is high".  Describe the 
expectation at that point based on the theory.  This might go with the 
previous technical report.

  c) Develop focused examples relating to 1) compound melody and
					  2) interatction between different
					     inertial descriptions.
  d) Enumerate rules for the 1st "full" model
     1. System architecture
     2. Slipnet/concept graph
     3. Agents
     4. Grouping ideas