Andrew’s blog: Today began again with Ben and myself feeling very enthusiastic about what lay ahead of us today after yesterday’s success. That went beyond all expectations. We began by assessing Sundari’s and Unni’s behaviour whilst being bathed as it was requested yesterday for us to be involved in all the tasks that the elephants had to do.
One of the basic principles of training is that you should use opportunities and set up behaviours so it is easy for the animal to succeed. So bathing presents perfect opportunities to habituate the young elephant to having a human on his back as the dead skin, dirt and parasites are scrubbed away with brush and coconut shell. Of course our two had already been mounted the day before, but nonetheless having a mahout sprawled out on the back whilst relishing in bath time is a great way to deepen the habituation and the bond.
As a result we took advantage of the happy moments and decided to ride the elephants back to ‘the classroom’ a hundred metres away. This we did and on the way practiced transitions of go, stop and step-back and occasionally getting off and re-mounting. These transitions rapidly improved to the point where we could ask the mahout to give all the signals with his toes and heels while the assistant on the ground faded way. This was the real beginning of independent riding. Following this we took a short break and then returned to repeat the transitions we had done and to gradually give the mahout more independence in obtaining complete control. The only exception now was the reliance of a second person to deliver the food rewards.
So our next move was to train the elephant to take food from the mahout sitting above rather than from a person on the ground. Previously we had delivered food directly into the elephant’s mouth and this was for a specific purpose – to keep the trunk out of the road and thus prevent us from being marauded by an ever hungry and searching trunk! Neither elephant had ever been offered food from atop his head and it’s quite a trunk manoevre to achieve it as the elephant has to rely on feel and smell. I showed the mahouts how to signal the arrival of food from above by tapping with the fingers on the nasal planum and then we would gradually offer the coconut pith reward further and further up the elephant’s face until it was at a point where the mahout could reach it.
Unni was our first candidate for this and once he learned to take food from above (after about 12 repetitions) he was onto it! Now he could be ridden and rewarded for an increasing amount of time and distance without the need for an assistant on the ground. Next we trained Sundari and she took to the aerial food delivery in as many repetitions as Unni. We were now on a rapid learning curve as both elephants became increasingly under the control and reward of the mahouts riding them. More rewarding however for me was the spectacle of just how much progress we had made with both human and elephant learning. I was wary of making too much progress in too short a time for the young elephants but also aware of the fact that the elephant should be in charge of its own learning pace provided we only taught a couple of tasks per lesson.
After lunch I resisted the temptation to go the next step in training and just simply repeated and consolidated what had been learned previously. The day drew to a close with good memories and great feedback from the mahouts who were now seemingly unanimous in their support for this approach. Their timing and delivery had improved enormously.
As a finale to this day, a mahout and his tusker (adult male elephant) gave us a demonstration of the things the elephant is expected to learn. Because the trunk is such a versatile organ, the tasks that need to be trained are vast. By negative reinforcement they learn to place and maintain their trunks in every conceivable position and to pick up, hold and deliver a variety of objects, to hit and push things with their trunks, to drag chains with their trunks and to spray water with them also. The elephants also learn to not only assist the mahout in mounting the elephant by the raising of the foreleg, but also to do it by raising the hindleg and clamping the tail sideways.
Our job was not to train all of these, but to teach an alternative way of training them than by negative reinforcement alone: by inserting positive reinforcement and elements of timing and release that optimise learning and by eliminating punishment for non-compliance. The aim is for the mahouts to use this new methodology to train all of the tasks on the menu. Then we can prevent the high death rate of mahouts that occurs as a result of the PTSD-like condition in the elephants that propel them to sudden violence.