Friday, February 29, 2008

Evolution of Laddie's Whistle Sit

I've been thinking about our current work in the T-drill as the latest stage in the continuing evolution of Laddie's whistle sit (WS):
  • Laddie was trained to sit on a whistle from the first week I got him, when he was seven weeks old. He was willing to sit at distance from the beginning, and was continuously reinforced for every sit throughout his development, continuing into the present.
  • Laddie was introduced to the WS in the context of handling by means of the pinball drill I invented for Lumi. The intent was that the WS would come to predict a highly reinforced recall, or an angle back (which seems to be fun for Laddie in itself) that also predicted another WS, thus the cast being reinforced by the WS acting as a tertiary reinforcer.
  • When I added dummies to the pinball drill, the value of the WS initially increased because it predicted a retrieve, which is of even higher value to Laddie than the reinforced recall.
  • However, in time, the evolving quality of the WS changed direction and declined. I believe that that was because Laddie gradually realized that the WS delayed access to the retrieve, and experimenting showed him that he could get to the retrieve faster by ignoring the whistle sit. Removing extrinsic reinforcement for such retrieves did not improve matters because the retrieve itself remained paramount to Laddie.
  • When Laddie's group training resulted in added excitement and diversions such as dummies over duck scent, a cage of live pigeons at the flyer station, or even friendly humans throwing at other stations from the mark Laddie was running, the quality Laddie's WS declined even further.
  • In private training, drills such as the wagon wheel and the diversion drill, intended to repair the WS while working on other concepts such as lining in the presence of diversions, were apparently too complex and demanding for the WS to take on a clearly rewarding association. The WS became entirely unreliable.
  • Switching Laddie to a steady diet of T-drills (which I originally called "diamond drills"), the WS remained weak at first, and on a small diamond course, the articles needed to be pinned to avoid Laddie picking them up after ignoring whistles to sit.
  • Then, with the pinning and later with the introduction of a larger course, Laddie apparently came to view a correct response to the whistle as the only way to retrieve. While his reliability began to improve, he went through a day of apparent avoidance behavior suggesting a lowered level of motivation for the training.
  • At last, with ongoing sessions of high value extrinsic +R, combined with stretched out distances to give Laddie a chance to stretch his own legs, invigorating weather conditions, and a consistent pattern of the WS always predicting an opportunity to retrieve, Laddie seems to be delighting in the retrieves and other extrinsic +R for the WS. In addition, Laddie also seems to be discovering the mental stimulation of watching me to see which way I'm going to cast him. This discovered self-reinforcing quality of the WS itself, separate from any reward that may follow the behavior, is suggested by Laddie's hyper-alert and joyous expression as he spins on the whistle and lowers himself into a sit, legs splayed and muscles tensed. His body language seems to say, "Which way, Daddy, which way?"

No comments:

[Note that entries are displayed from newest to oldest.]