Rio, a California sea lion, swims around
the training pool like an anxious athlete before a race. Shes preparing
for her starring role in an experiment designed to test her ability to
think. Eager to start, she voices her impatience with a call that sounds
like a cross between a lions roar and the croak of a toad
Rio can get antsy when shes about to do her experiment,
remarks Shannon Spillman, a trainer observing from inside the lab. A little
nervous, she takes a deep breath, puts on a microphone headset, studies
her notes for a moment, places her hands on a music keyboard, and checks
Rios activity on a video screen. She looks like an air-traffic controller.
Its a lot to keep track of, she says.
Ready? Spillman asks the student volunteers outside at the
pool. OK, lets go. And the show begins.
One of the volunteers calls to Rio. She leaps out of the water and gracefully
slides to a stop in front of a wooden poolside screen that looks a little
like the set of a T.V. game show. She rests her head on a stand facing
three windows. The center window opens to reveal a square card, a little
more than a foot tall, with a black and white picture of a ships
anchor on it. Rio studies the picture for a few seconds before the side
windows open, exposing pictures of a cactus on the left and a howling
wolf on the right. Rio quickly looks back and forth at the symbols and
then waits. Spillman hits a note on the keyboard to cue Rio, who immediately
touches her nose to the picture of the wolf. Spillman hits a higher note
signifying a correct answer, and Rio catches a fish that a volunteer flings
from behind the screen. Apparently, everyone agrees that the wolf is somehow
related to the anchor.
This mysterious experiment is actually a demonstration of Rios ability
to think. She has surprised researchers with feats of logic no animals
other than humans have accomplished in the laboratory. Rio and her cohort,
Rocky, have demonstrated an ability to draw logical conclusions that was
previously thought to require spoken language. Now it seems language requires
logic, rather than the opposite. This singularly important research comes
from the lab of Ronald Schusterman at the University of California, Santa
Cruz, who has spent years painstakingly training the sea lions.
Schusterman believes these abilities may have evolved to help animals
solve complex social problems in the wild, like the recognition of friends
and foes, by using many different cues such as sight, smell, and sound.
It really expands our notion of how an animal perceives its world,
Schustermans interest in the animal mind began in 1960 at Yerkes
Laboratories of Primate Biology in Orange Park, Florida, where he investigated
the cognitive and social behavior of chimpanzees, gibbons, and monkeys.
In 1963 he moved to the Stanford Research Institute to establish the first
laboratory in North America devoted to the behavior and physiology of
seals and sea lions. Schusterman and his sea lions spent a few years at
California State University at Hayward before they moved to the University
of Californias Long Marine Laboratory in Santa Cruz, where they
remain today. Whatever place would allow me to bring my sea lions,
thats where I went, he said.
Rocky and Rio are international stars in the scientific community and
in the public eye as well. They have been featured in Time magazine and
USA Today, and have appeared on a PBS television show called Animal
Schustermans prize pupil, Rio, was born in captivity 15 years ago
and raised by a human surrogate mother. She is an eager, playful student
and seems to enjoy her role in the research. Her motivation and focus
make her a perfect subject.
Rio may be the star of the show, but Rocky is the real veteran. Abandoned
in the wild as a pup, she has been with Schusterman since her rescue at
age one. Rockys 25 years put her well beyond the average 16-year
life span of a wild California sea lion. Despite cataracts that hamper
her ability to perform in experiments, she continues to be a willing participant.
At 150 pounds, she is smaller than the average female sea lion, but her
contribution to our understanding of the animal mind carries a lot of
Inspired by research with bottlenose dolphins at the University of Hawaii,
Schusterman taught Rocky sign language at age 10. She learned to perform
complex tasks involving multiple signs, such as touching a small white
ball with her flipper or bringing a big black cone to a small black ball.
Individual dolphins and primates had already become proficient with sign
language, but Rocky is all the more noteworthy for clearing the same mental
hurdle with a smaller brain. Her accomplishment led Schusterman to speculate
that he was dealing with a very basic learning process that could be at
the root of many complex behaviors. So he set out to reveal the underlying
Schustermans quest took him to the field of child psychology. He
modified methods used to elicit language comprehension in retarded children
and developed techniques to achieve the same results with sea lions. For
years, he has rigorously taught Rocky and Rio the skills and information
they need to succeed.
The sea lions learned to match many pairs of symbols through trial and
error, with fish as positive reinforcement. Once they had a firm grasp
of the relationships between the pairs, the animals were given new choices
that had not been directly trained to see how they would react. For example,
theyhad previously learned that the picture of a castle goes with the
golf club, and that the golf club goes with the pizza. But could they
make the logical connection that the pizza goes with the castle?
For Rio, the answer is a convincing yes. She knew the answer immediately,
and she responded correctly the first time she was given the same type
of choice with every group of symbols.
It was almost hard to believe when we got the results, Schusterman
said of Rios first demonstration. This intellectual leap proved
that Rio could form logical connections in her mind that werent
directly trained. Before Rios success, many scientists believed
that the ability to make these connections, known as equivalence,
was a uniquely human characteristic, impossible without spoken language.
Instead, it may be an essential building block for language.
Dr. Schusterman is the only one that has, I think, unequivocally
shown equivalence in nonhuman subjects, said Murray Sidman of the
New England Center for Children in Southboro, Massachusetts. Of
all the work with nonhumans, it most resembles what we see with human
beings. Research with human babies has shown that these logical
abilities are present before the ability to speak. Similar research with
retarded children also has revealed logical thinking in the absence of
Sidman was the first researcher to show this ability in a nonspeaking
human. He is considered the father of the equivalence theory often referred
to as Sidman equivalence. His 1971 study used equivalence
to explain the development of reading comprehension in a severely retarded
teenage boy. In this seminal experiment, the three symbols
to be matched were a picture, a printed word, and a spoken word. The boy
was taught to respond to the spoken word chair by selecting
a picture of a chair or the printed word, C-H-A-I-R. The boy was then
able to close the logical loop, responding to the written word by choosing
the picture and vice-versa.
In order to have language, humans need to be able to assign meanings to
words. One theory is that a set of equivalence associations between the
verbal and visual representations of an object is at the core of our ability
to speak. Human babies start making these connections before they are
able to verbalize them. Once they begin speaking, it is easy to watch
the set of associations grow as they connect the word mama
with the sight of their mothers, or the word cookie with their
Theres nothing magic about language, said William Dube
of the E.K. Shriver Center for Mental Retardation in Waltham, Massachusetts,
who has also done equivalence work with humans. One of the things
that underlies our language is this network of symbolic relationships.
You dont learn every relationship directly, but youre able
to behave as though all of them had been directly trained.
Schusterman believes Rio has shown us that equivalence might not be just
a step in human language development, but also a stage in the evolution
of language. I believe that equivalence is a precursor to language
in an evolutionary sense, Schusterman said. Sea lions havent
evolved their own language. But they may have laid the foundation, offering
a glimpse into the mysterious evolution of human language.
However, a few researchers who work with humans question the importance
of Rios demonstration. Steven Hayes, a psychologist at the University
of Nevada at Reno, is skeptical about the modifications Schusterman made
to the methods used with humans. Unlike Hayes and some of his colleagues,
Schusterman uses multiple examples of the same problem to help the sea
lions understand what is being asked of them. He also uses fish rewards
as reinforcement. Hayes feels the rewards and repetition are clouding
the results. Its very tricky to know if it is functionally
the same thing as we see in humans. I want to see it done exactly the
way we get it in 18-month-old babies, he said.
Schusterman stands by his technique. As in any kind of experiment
with animals in the laboratory where you are interested in an animal answering
a question, you have to make sure the animal understands the nature of
the question, he said. With humans, you can speak to them
and give them instructions. With animals, you can only provide lots of
examples. Once you provide those examples, then you can set up your testing
procedure. Thats what weve done.
Excited by their success, Schusterman and his graduate students continue
to test the limits of the sea lion psyche. Colleen Kastak, a graduate
student who has been working with Schusterman for nearly 10 years, has
been involved with the equivalence work almost from the start. During
that time, she has tirelessly worked with Rocky and Rio to expand on Rios
initial demonstration of equivalence. Her most recent work represents
a surprising and important step: proving that the sea lions can organize
symbols into functional categories.
Just as humans group things such as furniture, fruit, or college students
together, Kastak trained the sea lions to sort symbols mentally into separate
groups, coded as numbers and letters. To the sea lions, they are just
black marks on white cards. They had to gradually learn the associations
through years of trial and error. The sea lions learned to relate a group
of ten letters together and a group of ten numbers together by forming
connections between individual symbols within each category.
Rio excelled at keeping track of these groups. For this experiment, she
returned to the wooden poolside screen where she was given a choice between
a letter and a number in the two side windows (the center window remained
closed). If, for example, the researchers decided to begin with letters
as the correct choice, Rio was rewarded the first time she chose a letter.
In response to this reward, she pointed her nose at the letter in subsequent
trials. After earning ten fish in ten turns, the correct answer was switched
to numbers. Not knowing about the reversal, Rio incorrectly chose a letter.
But on the very next trial, she switched her strategy and began selecting
numbers and earning fish again. This couldnt be accomplished
through simple trial and error, Kastak said. Its truly
emergent knowledge: they are using old information in new ways.
Her work is in press with the Journal of the Experimental Analysis of
Another new dimension of the experiments involved varying the type of
fish used to reward correct answers. Appropriately picking a letter earned
Rio a piece of herring, while numbers were rewarded with a different type
of fish called capelin. The result was that the specific reward reinforced
the categories in her mind.
Recently, Rio was tested with this reversal procedure after a period of
one year with no testing. Amazingly, she performed without a flaw. She
came out perfect. I couldnt even believe it. I said she should have
had at least one error, because nobodys going to believe that,
Schusterman said with a chuckle.
To further test Rios command of the categories, Kastak and Schusterman
added new symbols into the existing sets of numbers and letters. For example,
they trained Rio that a new letter, such as K, went with an
old member of the letter group such as J. Through trial and
error with herring to reinforce the connection, the two symbols became
equivalent in Rios mind. They repeated the same procedure with new
numbers. The next step was to see if the new symbols were now equivalent
to the members of the appropriate original set in Rios mind. To
test this, Kastak returned to the three-window experiment. With one of
the original letters, such as A, in the center window, Rio
was given a choice of K or a new number. She chose the letter
immediately and received her herring. Without directly learning that K
was equivalent to each of the letters A through J,
Rio was able to make the correlation through a single connection.
The results shed light on how the ability to think logically might serve
a sea lion in nature. This suggests that animals can make inferences
about things that we had never realized they could. Really logical kinds
of inferences, Schusterman said.
For instance, Rio was able to relate different sensory signalstaste
and sightto each other. In the wild, this ability might allow a
sea lion to mentally group things such as the sight, smell and sound of
a friend or enemy, enabling her to respond to any of these signals in
the appropriate manner. For example, when a mother returns home after
a 2-day feeding bout to be reunited with her pup, they call to each other
and then use sight and smell to confirm the relationship. The pup can
recognize her sisters in the same way through their connections with the
Sea lions may also use this skill to find food. They can relate food to
environmental clues such as sights and sounds commonly associated with
the food. These clues may lead them to the food in the future. Lets
say a sea lion was to get salmon from a particular boat. Then it would
relate that salmon to the boat and all aspects of the boat: the configuration
of the boat, the sounds of the boat, the time of day that the feeding
episode occurs. So now the boat represents the salmon, Schusterman
Another significant result of this experiment was that it allowed the
aging Rocky to join Rio in her ability to close the logical loop. While
the first set of experiments often elicited a frustrated roar and toss
of the head from Rocky, she succeeded in the trials with specific fish
rewards. Her performance reinforces Schustermans theories about
the practicality of these skills. It also reduces the likelihood that
Rios earlier success was a fluke.
While Rocky and Rio are undoubtedly pioneers, they are still just two
sea lions. Skeptics have a hard time disputing Schustermans findings,
but more work with other animals needs to be done before the idea that
animals are capable of logic can be broadly applied and embraced by the
scientific community. Many researchers have attempted to reveal this ability
in other animals. To date, Schusterman and his sea lions are alone in
Sea lions are smart. But are they smarter than chimpanzees or dolphins?
A more likely explanation is the method behind the magic.
William MacIlvane of the E.K. Shriver Center for Mental Retardation has
worked with humans as well as capuchin monkeys. He believes that Schustermans
careful methodology may be one key to his success with the sea lions.
Ron thinks like a teacher, MacIlvane said. He has a
good sense of what to do to make the learning outcomes likely. The methods
are demanding technically. If the primate labs used Rons methods,
theyd get positive results.
While there are no experiments that prove primates have the capability
to think logically like Rocky and Rio, anecdotal evidence suggests that
MacIlvanes hunch is true. Herb Terrace, a psychologist at Columbia
University, reported teaching a chimpanzee sign language in 1986. The
chimp learned to sign the word dog when he saw a picture of a dog. Then
one day, the chimp heard a dog barking outside, and it signed the word
Schusterman is optimistic about the partial successes seen inspecies such
as primates, dolphins, and birds. Its likely that animals
like chimpanzees, if given the type of training weve used, with
lots of examples of matching and reversal, would do remarkably well.
Successful demonstrations of logical thinking by other species would lend
support to the idea that equivalence is an evolutionary predecessor of
The next step for Schusterman is to take what has been learned in the
laboratory and apply it to observations in the wild. Scientists have been
very resistant to the idea that sea lions are social animals. Schusterman
believes this may be due in part to the difficulty of observing sea lions
in the wild.
Along with graduate student Evelyn Hanggi, Schusterman documented friendly
social behavior among related female California sea lions in captivity
at Sea Life Park in Oahu. By learning to recognize all the sea lions at
the park, they documented the friendly and aggressive behavior. They found
that nearly all the positive interactions, such as swimming or napping
together, occurred between mothers and daughters, while aggressive behavior
was directed toward unrelated females. But similar work has yet to be
done in the wild. DNA is the key to the whole thing, Schusterman
said. The idea would be to make observations and predictions about
relatedness and then test those predictions with DNA.
has his eye on a particularly social group of California sea lions in
the Galapagos Islands, where he plans to test his ideas. Theres
not much known about the social organization of sea lions, he said.