Killing Keiko
Page 10
Molecular Movement
Stepping up to the side of the med pool, I glanced to my left and waited. I wanted to make sure Keiko was moving in my direction before calling him over. It was an unusually nice day, sunny and with very little wind. The air was still enough that snow accumulated on every horizontal pocket in the cliffs and nearly blanketed the nearby town. In the easiness of the afternoon the birds overhead were loud, a ceaseless cacophony of chirps and whistles.
Recognizing my best opportunity, while Keiko was moving, I bent down and slapped the surface of the water. Keiko didn’t show any response to my slap. He casually continued his movement in my direction and eventually arrived right in front of me. He didn’t lift his head or make any motion to acknowledge that I was there; rather, he only came to a slow stop with his nose pressed lightly against the side of the med pool structure.
I waited. Keiko didn’t move. I could see he was watching me, rolling his eyes just so. But that was it. After a few moments I stepped back and walked away from the side of the pool. Behind me, Karen and Stephen had been watching.
“What was that?” Stephen asked with a smirk, finding the whole situation slightly humorous. After all that we had talked about during the last week, they probably anticipated that I would wield some magical power over Keiko and were expecting to be impressed. It couldn’t have been less impressive. I had to jump up to touch bottom.
“I’m not accepting that—if he wants my attention he’s going to have to sit up and give me his.” I didn’t mean for it to sound arrogant, but to Stephen and Karen it did. We stepped out of Keiko’s line of sight and continued debriefing the short session.
“If we’re going to get this boy moving we have to start looking for more energy … tough love,” I said. Karen didn’t like it.
“He came over the first time, though.” It was more a statement than a question (and a measure of what they were used to). She asked what was next, more to test me than out of genuine interest. We were leaning against the north side of the green research shack, staying out of Keiko’s view. Peering around the corner, I could see that he remained in the same position where I had left him.
“I’m going to wait until he moves away and is actively swimming, then I’ll call him over again and see if he gives me anything worth reinforcing.” I avoided a direct answer, hoping instead to illustrate my point. We weren’t even scratching the surface, and yet I could tell Karen was already uncomfortable with how I was pushing Keiko. A few minutes later, Keiko was swimming in the north pool again. I stepped back up and called him over, again waiting until he was facing in my direction before giving any signal.
He approached much the same way and stopped, nose pressed against the side of the pool. I made no reaction and waited.
Behind me Stephen asked, “What are you looking for?”
“Any sign of movement—even at the molecular level at this point,” I replied.
Finally Keiko lifted his head, and for the first time that morning his eyes appeared above the surface. I immediately reacted, giving the whistle bridge a short burst and tossing him a single herring. I moved quickly down to my right about thirty feet and slapped again. This time he came over slightly faster, not much, but arrived with his head up and looking at me intently. I made a huge difference in my energy and posture, reacting to Keiko’s improved attention. I reinforced him again, this time with a small handful of herring, then broke away from the session.
The rest of the afternoon I repeated the same sequence of interaction, asking for a little more each time. I kept my sessions with Keiko short, each four to five minutes at most, and put a lot of emphasis on a variety of differing reinforcements or rewards. He never knew when I would show up or how long I’d stay, but the result was always intriguing. Sometimes I gave him a quick rub down on his pectoral fins using a brush. Other times I used the water hose to massage his flukes or spray on his tongue. I made quick and novel changes matching the level of his effort. That day was all about how far and fast I could push him. I wanted to find his limits, see how far he would go for the new guy.
Over the first week or more I made few attempts to work with Keiko myself, instead working him through the existing staff. There was sound theory behind this approach. We needed to teach the staff solid foundations of behavioral modification, and the best way to do that was to make them actually do the training, with guidance of course. I was dead wrong.
Largely because of the gray areas that came with sloppy and inconsistent interactions, Keiko’s day-to-day life was rife with confusion. This did not lend itself to strong relationships with his caregivers. Also, most of the staff tried very hard to implement the instruction they were being given, but ultimately Keiko was learning faster than they were. I could no longer afford the delayed responses and second tries that came from working through the existing staff. More and more, I started working with Keiko directly.
It didn’t take long to get a solid grasp on my relationship with Keiko. In at least one way, it was easy for him, as I was very clear about what I expected. If he met the requirement, great things happened. If not, nothing happened. It was black and white. Keiko responded to this clarity and consistency and began to excel. He was a quick study and began discriminating when it was me working with him. He knew from the start of each session that I would ask a lot of him, but he also knew I would meet his efforts with like energy.
The most effective trainers put a lot of planning into their reinforcements ahead of time and are ready to respond when the right behavior happens. Contrasting this elusive principle, it seems human nature to focus on what the animal is doing wrong rather than what’s right. In the world I came from, the better an animal was doing, the harder the trainer worked in creating reinforcement. The best training sessions leave the trainer exhausted and dripping with sweat, no matter the frigid water or blustery chilled winds. There had never been much energy or planning put into providing exciting and diverse rewards in Keiko’s daily training interactions. This aspect alone allowed me to stand out from the rest of the crowd. Fortunately, the staff was learning also, and the results we were seeing with Keiko quickly became motivating to everyone.
“Dancing Queen”
Animals in a training environment tend to “mirror” their trainer’s energy level. Too many times I’ve witnessed a trainer plop down on his butt in front of an animal at the start of a training session only to have that animal take one look at Mr. or Ms. Boredom and promptly leave the scene for something more interesting, like watching paint dry. A tried and true method I had utilized in the past proved the easiest way to get trainers “off their butts” and infuse energy into their interactions with Keiko: music.
As tribute to one of our favorite Icelandic coworkers, Mr. Iceland, we blasted “Dancing Queen” by Abba across the bay pen during some exercise sessions (only in the early stages and weather permitting). The music was for the trainers, and their spike in energy produced night-and-day results in Keiko. Take a moment to picture the scene: “Dancing Queen” playing in the mind’s ear, trainers flamboyantly engaging their entire body presenting to Keiko what would otherwise be a mere “hand signal” moving in tune with “… young and sweet … having—the—time—of—your—life….” Music provided much levity but also produced some of Keiko’s most effective “workouts.” It was a pure, uncomplicated transition for the training staff to grasp and illustrated the fact that food alone was not a key motivator. By merely energizing the trainer’s posture and creativity to react to Keiko’s successes, we could extract some pretty spectacular energy from Keiko in return.
Bridging the Gap
In the beginning my time was dominated with sharpening the effectiveness of each and every tool in the training environment and eliminating “superstitious” beliefs about how learning occurs.
In every setting in which I had previously conditioned behavior with animals, we always “debriefed” after training sessions, discussing the various observations and identifyin
g areas for improvement. Not only was this an important part of any environment where success relied on consistency, but it was especially important when the subject of that discussion was a killer whale. Small mistakes can lead to menacing consequences with the ocean’s top predator. In this case, we couldn’t let small mistakes undermine Keiko’s needs.
I recall a conversation I had many times in just such debriefings, this time with Steve Sinelli and Karen McRea. We were sitting in the bay pen research shack having just finished a training session with Keiko.
Steve reminded me of one of Santa’s elves. He was not categorically short, but close. Balding and with a close cut and well-groomed beard, his black hair gave no signs of graying and lent to his elfin appearance. Steve had a youthful energy that contrasted slightly with his age. More often than not, he wore loose fitting water-resistant nylon and fleece lined pants and a black fleece vest over his white turtleneck shirt. It was Steve’s black and white uniform attire. He was confident, and like me, could be argumentative.
Steve sat in front of the computer but turned away from the screen, instead facing Karen and me. He had just finished writing up the session. Steve had worked with Keiko in that particular training session, rehearsing a behavior called a “fluke presentation.” When given the signal or Sd for the behavior, Keiko would turn on his back at the surface and present his tail flukes to Steve who sat on a water-level floating platform. In this ventral position, blood samples could be easily drawn from the larger veins that run through the flukes. Typically, Keiko would remain in this position for three to four minutes, and sometimes as long as ten minutes depending on what was required. In past research on his breath-holding capacity, Keiko had held this position (and his breathing) for over thirteen minutes. This and many other husbandry behaviors were a normal and important part of Keiko’s life. He knew them and performed them well. I described what I had seen during Steve’s session and suggested a few adjustments.
“Tell me why you were using your whistle bridge during the fluke presentation behavior?” I asked.
I had learned to sit during these types of discussions. At six–foot-two it was easy to inadvertently bully a shorter person, and I didn’t want to start off on the wrong foot. No matter, Steve was one of the older team members and was pretty comfortable in his own skin. Karen, not so much.
“We use a short whistle to let him know he’s doing good and then a longer whistle to end the behavior when he’s completed it.” Steve replied instructionally.
I’d seen this superstitious use of the whistle bridge before and debunked it just as many times. It was a critically important foundation in establishing how behavior is conditioned. As importantly, everything in Keiko’s world had to be refined, even the most seemingly innocuous training habits.
Karen didn’t say much, but I could tell she was listening intently. With a degree in psychology, she knew the language of behavior but had never used it in an applied setting. Nonetheless, I found it more productive to speak in layman’s terms, avoiding any confusion. “Okay, think of the whistle or bridge as a ‘secondary’ reinforcement. When Keiko first heard a whistle bridge it didn’t mean anything to him. But over time, someone taught him that the whistle meant ‘good’ by following it consistently and immediately with various rewards.”
They accepted this, but kept a poker face as if to say, Duh, tell me something I didn’t already know. This much was cliché and often part of educational spiels at various training facilities. I was being ultra-elementary on purpose; I didn’t know at what point along the line we would cross into new territory.
I continued, “The whistle bridges (I emphasized the word) the time between when Keiko completes the behavior correctly and receives his reward from you.” There was more. “It also takes a snapshot picture in his mind at the precise moment that he has committed the correct response.” I needed to drive the latter point home or they wouldn’t get it.
“Imagine that Keiko is learning to do a jump in a specific place in the pool that you choose. How do you teach him to jump in that precise spot every time?” I wanted them to engage in the discussion.
Steve replied with ease. He had seen this and was not fooled. “You use a target pole and slap in the position where you want him to exit the water.”
“Yes, but when you are fading that target, teaching him to ‘remember’ the spot without the help of the target, when do you bridge the correct response?”
Steve tested the water, “When he comes up in that spot?”
It was as much a question as a response. Steve knew I had a trick answer up my sleeve. He smiled …we were having fun with the discussion. Karen sat in the chair under the west window, happy that she was only indirectly involved. In my peripheral vision I could see that she was processing the question with a pensive look on her face.
“Not exactly … you want to use the bridge at the precise moment that he turns up from the bottom toward that spot. That’s the moment you want to grab his thought process and say YES! That’s it—you’ve got it!” I found myself standing to emphasize the importance of this precision tool, using my hands as if one were Keiko sweeping up to jump, and the other was the surface of the water. “That’s when you need the whistle to be sharp and powerful—grabbing his attention.”
But Steve was unsure of where this was going. “Okay, but what’s that got to do with bridging his fluke presentation while he’s doing it correctly?” he asked.
“Everything!” (I loved this stuff.) Continuing, I explained, “When you use the bridge and do not follow it immediately with a reinforcement or change, a consequence, some form of positive consequence—you are dulling a precision instrument—the whistle bridge. You are in fact desensitizing that bridge, reducing its value. After a while it is no longer a precision instrument but a blunt tool that has lost effectiveness. It no longer means anything when you most need it to.”
Steve was not convinced. “That’s why we use a long whistle during the behavior and a short whistle as the precision part.”
“The long whistle doesn’t mean anything to Keiko. It has no consequence, no change that gives it value. In effect, through generalization, you are only draining the value of the bridge as a whole—whether that’s a long or short whistle or a catchy melody.”
“Don’t you think he knows it though, like we’re saying good boy—keep going?” he asked.
“No I don’t. By allowing him to continue the behavior you are accomplishing the same thing. You don’t need the midterm whistle. Alternatively, you can rub his flukes, providing reinforcement while he’s holding the behavior, and achieve the intended result. But you need to think of that bridge as a vital learning tool, and protect its value by making sure that it has consistently positive and immediate consequences. The whistle bridge has a very specific application; it’s not a tool to be thrown around lightly and for convenience.”
We would have many and varied conversations of a similar nature, discussing everything from the whistle bridge to transferring learning and reinforcement history from one environment to another and beyond. Although the prospect of teaching behavioral modification was thoroughly enjoyable, added to the vast needs demanded by Keiko and the long road ahead, it was exhausting. We needed “top gun” trainers that knew this stuff intuitively.
Do-si-do
The first rotation team and I were just beginning to find our groove when it was time to alternate the entire team. The second rotation would be sashaying in and taking up residence in the hostel, the bay pen, and taking the reins on Keiko’s daily needs. The only holdover between the two rotations … me. This left no other option; I would have to pull out the most effective organizational secret weapon ever conceived by man—the “staff meeting” (in case there’s any doubt, that was indeed heavy sarcasm).
As chance would have it, Jeff’s alter ego, Peter Noah, was more of an organizational freak than I. Peter held an informal group meeting his second day, setting the record for 100 percent more meetin
gs than Jeff had tallied in two months. And just like that, it was a completely different atmosphere in our quaint but peculiar hostel.
E-mail: May 9, 1999
To: Alyssa
Subj: Good Morning My Sweet
New staff arriving and existing staff are on their way home. Quite an interesting exchange of issues. Lots of change happening … met the last supervisor level yesterday. I will be working with him through June. The weather finally laid down … sun is out today and the winds have dropped to about 19 mph. We have had this wind storm nonstop for the last several days straight. It is nice to finally have a little calm weather, not to mention seeing the sun. Speaking of which, sunrise is at 4:27 AM and sunset is at 10:03 PM, but it is never really dark. Between sunset and sunrise it just sorta stays twilight. Weird … I will take some more pictures today and try to send them your way by tonight. We had our first (though informal) staff meeting yesterday. They have not even had staff meetings on any regular basis. Given the nature (of the project) and safety issues involved with this operation it blows my mind that regular meetings and protocols are not in place. Soooooo much to do. I will work out my return date with Robin today and let you know my schedule.
Until then, love is in the air,
Marjke (Icelandic for “Mark”)
Peter and I hit it off immediately. We spent his first full day back in Heimaey sequestered on the bay pen—just the two of us. Peter was now the acting on-site project manager. He wanted to know who this guy was running things with Keiko. It just so happened that Peter was a very analytical, left-brain thinker. In my past, I had been accused of being a sneaky-deep-down “Vulcan,” favoring logic. We both enjoyed a good clean whiteboard.