(From Samuel Saunders [950728:1650 EDT])
When I first saw Bruce's analysis of the cyclic ratio data, I thought of
variable ratios. After following the discussion and struggling with the
problem myself, I wonder if in some ways this schedule might be similar to
variable ratio rather than fixed ratio.
1. It seems unlikely that the rats are concerned with controlling rate of
reinforcement on a per-schedule basis. For one thing, I would expect the
global rate of reinforcement rather than the (abstracted) local rate on a
particular ratio, to be of interest to the rats. Even if the rats "wanted"
to control the rate of reinforcement on a per-schedule basis, it would
require a rather elaborate recording mechanism to keep this straight. The
rats would either have to remember the entire series, or use a compound of
direction of change and just-finished ratio, to keep track of a particular
schedule.
2. The local rate of reinforcement information comes with delivery of the
food pellet. After the food is consumed, the schedule is different. In the
increasing part of the cycle, rate of reinforcement may drop even if rate
of response increases. In the decreasing part of the cycle, rate of
reinforcement will increase, even if rate of responding is unchanged.
I suspect that the "molar" model based on control of integrated rate of
reinforcement is unlikely to work in this situation. The dynamic model
will be required. This may say little about the fixed ratio situation,
where repetition of the same response requirement may make a "molar"
control model viable.
I originally questioned my suspicion that the rats were not tracking the
individual ratios when Bruce quoted the result that post reinforcement
pause was appropriate to the upcoming schedule. I wonder now if this
doesn't result from the rat being aware of the phase of the cycle
(increasing or decreasing), without necessarily knowing the current ratio.
Anticipating that the next ratio will be longer (or shorter) than the one
just completed could influence the pause, and the fixed order would make it
appear that the pause was "appropriate to the following ratio."
//----------------------------------------------------------------------------
//Samuel Spence Saunders,Ph.D.