Part 2 (1/2)

One day, the sun came out and the snow began to thaw. I went for a run and soaked up the rays and warmth with gusto. As I made my way along the trail up in the pastures next to the woods, my eyes caught a strange sight. A large pink earthworm was crawling very slowly across the snow, coming from an earth bank exposed to the sun. I stopped to observe it for a while. Like me, it had made its way out into the first warmth of the year and appeared to be going somewhere.

Chapter 5.

INSECT MINDS.

It was mid-May, and spring was turning to summer. I made a beeline for the South of France. Coming from the Swiss hills, I felt hot for the first time that year. I had an appointment at the University of Toulouse with Martin Giurfa, a scientist who had recently demonstrated that bees can handle abstract concepts.

The work by Giurfa and his colleagues had caught my attention in a scientific journal. They reported on an experiment in which they exposed honeybees to a simple Y-shaped maze. The entrance to the maze was marked with a particular symbol, such as the color blue. A bee flying through the entrance encountered a branching pathway, or ”decision chamber,” where it could choose between paths. One path was marked with the color blue, the other with the color yellow. Bees that followed the blue-marked path discovered at its end a vial filled with sugared solution. Bees that took the yellow path received no reward. The bees soon learned that sugar lay at the end of the route marked with the same symbol as the one marking the outside entrance. ”Same” equals ”sugar,” in other words. In a subsequent experiment, the opening to the maze was marked by a different symbol, such as horizontal dark lines. In that case, on entering the decision chamber, the bees reencountered the two pathways, which were marked this time not with colors but with lines”vertical lines on one path, horizontal lines on the other. The bees pa.s.sed with flying colors, heading straight for the pattern that matched what they saw at the entrance. Other experiments revealed that the bees could also transfer their knowledge across the senses: Bees that learned about sameness by matching odors were able to apply the concept to visual signs. Though bees have brains the size of pinheads, they can master abstract rules.

This research falsified the notion that ”brutes abstract not.” It also showed that small brains do not hinder thought. I felt moved to meet the person behind this research and hear his point of view.

The University of Toulouse has a sprawling campus. Despite the signs and pathways, it took me half an hour to find the Laboratoire d”Ethologie et de Cognition Animale. It was located in a four-story building that was being renovated. As I walked in, drills resonated from the floors above.

Martin Giurfa had recently been chosen by France”s National Center for Scientific Research (CNRS) to head their new center for the study of animal cognition. We had not met previously or spoken on the phone and had only communicated by electronic mail. As I knocked on his door, I considered the possibility that he might wear a white lab coat and speak with detachment.

Instead I found a youngish-looking man sitting in front of a computer at a comfortable desk, wearing a green-checked s.h.i.+rt with short sleeves. The room was filled with plants, and the blinds were down to fend off the sunlight. Giurfa wore wire-rimmed gla.s.ses, and his hair was dark and short. He smiled and invited me to sit down in the chair next to his desk. He spoke English with an indeterminate accent. I asked where he was from. He said that he was born in Argentina and that his family had come from Italy.

As a cultural hybrid, I felt at ease with Giurfa. I was curious to know how he had come to develop an interest in biology. ”Since I can remember, I have loved animals,” he replied. ”I was always fascinated by the observation and the magic of the living machine. But I have just made a big mistake: I used the term machine to describe living organisms. That is exactly the opposite of what I think. In fact they are not machines. But I was always fascinated by looking at the living organism, from the point of view of the exterior observer, seeing how it moves, takes decisions, and so on. It was always fascinating for me how a wasp decides to go here and not there, how a wasp finds its way home and identifies the nest, how a bee forages from flower to flower, always going from the flowers of one species to the same species.” As a child, Giurfa kept different animals as pets, including insects, water snakes, and a boa constrictor”much to his mother”s dismay.

Giurfa explained that he had referred to bees as ”machines” because that was how he used to think about them in the past. But the more he understood how animals make decisions and learn, the more he had to admit that they do not act mechanically. His view started to change in 1990, when he went to Berlin and began working in a leading neurobiology inst.i.tute, alongside sixty colleagues from different fields of science who were all studying memory and learning in honeybees. It soon dawned on him that bees learn in an intelligent way. For example, their capacity to navigate surpa.s.ses our own: ”If I take you to a distant part of the campus,” he said, ”and release you there, you won”t find your way easily back here. But bees can. How they do it is the question. This is why I started to think about cognition in invertebrates, which, of course, at the time, was considered a kind of contradiction in terms. People said, ”You are absolutely crazy for raising this kind of question. How could you think that invertebrates could have this kind of intelligent behavior?” That is what people were saying to me.”

”What did you make of that resistance?”

”I simply didn”t care about it. That was an advantage in Berlin; you had intellectual freedom to raise questions and perform research work.”

I wanted to know why some of his colleagues were so opposed to studying cognition in invertebrates. ”Basically,” he said, ”it was the dominating view, that you can find even now in some people, for instance in researchers working with vertebrates. They still think that invertebrates are small robots, that they are simple machines, reflex machines, you know, like Pavlov machines, or Skinner machines. Simply like hitting a hammer on your knee and having a jerk reaction. They say that invertebrates have to be simple like that.”

Though Giurfa was critical of the robotic view of insects, he admitted that it had helped advance the study of insect movements and behaviors. ”Considering insects as simple robots has, for instance, stimulated the creation of machines like the Mars explorer, which was inspired by how insects move their legs and so on. This point of view is of course short minded, if you will, but at least it had this positive aspect.”

Someone knocked on the door, interrupting the conversation. Giurfa had a brief exchange in French with a colleague, and I noticed that he spoke with greater fluency than in English. Once he was done, I renewed the conversation in French and we continued in that language. I asked whether there had been resistance to his recent work on the capacity of bees to handle abstract concepts. He said he was confident that the experiments were well conducted and that the results, which were published in the journal Nature, could not be attacked scientifically. But he did mention resistance from a group of researchers at a nearby center for the study of primate cognition. They contacted Giurfa to say that they had tested monkeys on the same task and found that certain species could not do it; therefore, they did not believe it was possible bees could. Giurfa said this kind of reaction occurred less and less frequently.

In his view, when animals are found not to accomplish a given task, this is not proof of their stupidity. ”In most cases, the problem lies with the person conducting the experiment and involves incapacity in the researcher to develop experiments that pose the problem correctly and allow one to answer it properly. If you will, a negative result shows nothing, in the final a.n.a.lysis. A positive result shows something. But when an animal cannot do something, the question remains: Is it incapable of doing it or have I not been clever enough in my research concepts and experimental design?”

”So, would you say that the problem for the moment is not that nature lacks intelligence but that researchers studying it do?” I asked.

”That is one of the problems, certainly. I think we are a long way from having made a kind of mental jump which would allow us to ask certain questions.”

I had read several recent books that discounted the intelligence of individual insects, referring instead to ”swarm intelligence.” The idea seemed to be that bees were mindless robots programmed according to a series of simple rules, and that intelligent behavior emerged from the interactions of the mindless parts. ”Emergence” was a concept that was used to explain how ”dimwitted” individuals could appear to act intelligently. I asked Giurfa what he thought about ”swarm intelligence” and ”emergence.”

He replied that these concepts could explain some behaviors, but not all, and that it was important to distinguish between group behavior and the intelligence of the individual. ”All these studies on emergent properties are certainly interesting, and they are a good challenge for me. I like these studies because they make me rethink my own research from another perspective.” He said it was important not to take his own point of view too far by claiming, for example, that bees are capable of the highest and most flexible forms of learning. In fact, bees sometimes behave stupidly. If placed in a maze with a gla.s.s cover, for example, they perform as well as rats up to the point of reaching the food reward, but they are incapable of turning around and going back to where they have come from. Once bees eat, they are rigidly programmed to fly upward. Bees in a gla.s.s-covered maze bang against the gla.s.s cover, trying to gain alt.i.tude, until they die. They are programmed according to a simple rule: To get back to the hive, first go upward, to where light intensity is greatest, toward the sky. So, Giurfa said, it is important to avoid exaggerating the plasticity of bee intelligence. Both principles operate: There are simple rules and emergent properties on the one hand, and plastic cognition on the other. ”That”s why it”s a challenge, because it obliges me to think about the problems in my system from another perspective as well.”

Scientists often use the concept of ”instinct” when explaining the capacities of animals. I asked Giurfa if he found it useful in his work. He said that he had started his work in Berlin by studying a question related to bee instinct, looking into whether or not bees have information encoded in their brains when they take their first exploratory flight. Giurfa built a large apiary containing a small beehive in which all external conditions are controlled and went on to demonstrate that bees spontaneously prefer certain colors, in particular very intense blue and yellow. These colors correspond to the flowers richest in nectar. So instinct exists, Giurfa said, and is a useful concept. But Giurfa also found that bees can modify their instinct according to what they learn about the world. In the controlled environment he constructed, Giurfa arranged for pollen to be a.s.sociated with other colors and found that bees can modify their color instincts. ”We see the incredible plasticity of the system,” he said. ”This means that they go into nature equipped with instinctive information, which is not rigid, and which they can forget or put aside on the basis of personal experience, meaning to say on the basis of learning.”

A loud hammering echoed through the ceiling. Upstairs, workers were bringing down a wall. Toulouse University was remodeling its Animal Behavior Department, turning it into a ”laboratory of animal cognition.” I took this as a sign of the times. Science is opening up to the intelligence of other species, and this is bringing down university walls, literally.

Giurfa turned to his computer and summoned up a full-screen image of the internal organization of the bee brain. He explained that a key part of their research involves looking into bee brains in search of the ”neuronal substrate” of a given behavior. For scientists, the great advantage of the bee brain is that it can handle complex mental tasks with less than a million neurons. This simplifies research. Working with brain-imaging techniques, Giurfa and his colleagues mapped which parts of the bee brain are active when the animal learns about the smell of the outside world. Their research revealed the existence of a sensory-integration center called the ”mushroom body,” which is made of 170,000 densely packed neurons. This central component of the bee brain receives sensory input and directs behaviors”such as when bees dance symbolically to communicate information about the location of pollen-laden flowers, or navigate over long distances according to the sun”s position in the sky, or estimate the quality of potential nest sites.

Giurfa explained that they looked into the bee brain using a technique known as calcium imaging. Given that active neurons exchange calcium, one can open the skull of a living bee and bathe its brain in fluorescent substances that latch onto calcium and reveal the active parts of the brain. ”This is another advantage of invertebrates,” he said. ”This process does not affect the animal. Invertebrates are enveloped in a capsule; their whole body is a capsule that is not innervated. It is very hard for us to imagine, but that is how it is. Imagine that instead of having skin, which is sensitive because it is innervated and filled with nerve endings, we had our internal system in armor.”

”So the nerves stop with the brain?”

”Exactly. If you open it, if you make a small hole in a bee”s head, it is just like taking a helmet off. You do not hurt it because it is not innervated. The outside part of the insect which you can see is like a protection sh.e.l.l.”

I viewed pain as an experience humans probably share with animals. I have pa.s.sed several gallstones, and know that pain has to do with the deep wiring of my body. I know just how paralyzing and excruciating it can be when raw nerves inside the body are sc.r.a.ped. Pain seems to be an undesirable experience one can have when one is equipped with a central nervous system. I knew nothing about pain among insects, but I figured that if their brains can handle abstraction, they can probably handle pain as well. I asked Giurfa if he thought bees feel pain. He said, ”If you hurt a muscle, then, yes, you hurt the animal, but if you just remove a bit of sh.e.l.l, you do not hurt it. So you can delicately expose the brain, by fastening the bee in a tube, and you can look at what is going on.”

Pointing at the map of the bee brain, he showed me the ”olfactory pathway.” On the tip of a bee”s antennas are olfactory receptors (corresponding to the mucus membrane inside the human nose), which feed chemo-electrical information into nerves leading to two small grapelike structures at the base of the brain (similar in shape to our own olfactory bulb). From there, wirelike neurons lead to the mushroom body, which processes the different inputs.

Giurfa and his colleague Randolf Menzel recently described the ”cognitive architecture of the honeybee minibrain” as a network of independent units, the ”modules of an insect mind.” Each module treats information from a specific input, such as smell. The different inputs are then combined in a central locus, the mushroom body, where ”context-dependent decisions” are reached. This enables honeybees to ”extract the logical structure of the world.”

Bees go out into the world equipped with a tiny brain and learn about their environment in next to no time. They have a lifespan of only two or three weeks. They seem ready to learn as soon as they hatch.

Some of Giurfa”s graduate students were running an experiment next door. He suggested we go and check their work. I followed him out of his office and found three students sitting at a white table gathered around an odd-looking device”a blue metal plate with a copper cartridge sticking out of one end. A bee was strapped into the cartridge. An array of small tubes was directed at its face. The graduate student conducting the experiment held a toothpick in his hand. He explained that when the antenna of a hungry bee is touched by a toothpick dipped in sugared solution, a reflex always occurs, causing the bee to stick out its tongue in a jerk reaction comparable to the reflex of a knee hit by a hammer. Giurfa explained that one could present an odor immediately before the sugar reward, and teach the bee to form an a.s.sociation that, in subsequent tests, causes the odor, rather than the sweetened toothpick, to release the tongue. This shows bees can learn about smell; it also reveals which parts of their brains are active when they do so. Bees, it turns out, can detect odors with greater sensitivity than dogs.

I looked closely at the bee in the cartridge. It was strapped in with blue tape. It could only move its antennas and tongue. Its head was glued to the back of the tube.

I chatted for a while with the graduate students. They were from Germany and said, speaking in English, that they loved Toulouse, which is near the Mediterranean, the Pyrenees, and the Atlantic, all at once. But they said it was more difficult to concentrate on science here; in Berlin, where it was ”gray and rainy all year,” they found it easier to work; here, they wanted to go on vacation all the time.

I focused once again on the bee. Spending an hour strapped in a bullet was a long time from a bee”s perspective. It did not seem very comfortable. I inquired about its fate after the experiment. Giurfa explained that the bees they tested in this fas.h.i.+on had to be killed, because otherwise they would return to the hive and falsify subsequent trials.

The bee I was observing had already experienced one nonrewarded odor and one rewarded one. Now it was about to receive the rewarded odor for the second time. We gathered around closely to see if it had learned something. The odor came out of the tube, and presto, the bee shot out its tongue. It had made the connection.

At that moment, I felt jubilation, and kins.h.i.+p with the bee. Like some humans, it was a fast learner. I asked Giurfa whether he thought finding intelligence in insects means they deserve better treatment. He said he agreed with the question and explained that there was research he would never do, in particular inserting electrodes into bee brains. He had not heard of animal-rights activists opposing research on invertebrates, though at the University in Berlin there had been much resistance to scientists studying vertebrate neurobiology. ”At that time we chose the invertebrates exactly because we did not want to offend the sensibilities of some students,” he said. ”If you want to study biology of the whole, and see all the possible fields it has, you have to see and try these experimental techniques and approaches. Being an experimental biologist, I could never approve of thinking that everything could be done with simulations and models.” He added that he would not perform the experiments he did on bees on cats, dogs, or apes, due to his ”particular personal sensibility,” which he knew was ”just an anthropocentric point of view.”

Though he refused to put electrodes into the brain of a living bee, he admitted that exposing the bees” brains and submitting them to calcium imaging was injurious to them, and would lead to their being killed. I returned to the question of whether bees feel pain. He laughed and called it a difficult question. In labs in South America, he said, scientists have shown that bee nervous systems produce opioids, presumably to induce a.n.a.lgesia. However, given that bees and humans are separated by hundreds of millions of years of evolution, he questioned whether the human concept of ”pain” applies to bees. In his view, no one knows the answer.

I asked about the overall implications of his work on bee cognition. He said it shows that brain size is irrelevant when it comes to the capacity of performing highly demanding cognitive tasks. He also said it is time to do away with the arbitrary barrier that scientists have erected between vertebrate ”learners,” such as apes, pigeons, dogs, cats, dolphins, and humans, and all other ”noncognitive organisms.”

We spent half an hour with the students, then left them to their painstaking research and went out to lunch at a nearby restaurant. We talked about several subjects. He asked me about the Peruvian Amazon, where he had traveled. I asked him about his intellectual influences. He spoke of his thesis advisor in Argentina, and of his love for bees.

At one point I asked for his view on plant intelligence. He said the problem with plants is that they do not move, which makes it difficult to perform scientific experiments on them. I mentioned the parasitic dodder plant, which roams about and correctly gauges the nutritional content of other plants. He immediately suggested research questions about dodder. Can it learn to avoid certain substrates through negative reinforcements? If it demonstrates a capacity to learn, at what level of its cellular structure does the learning establish itself? ”When you talk of learning, or cognition, the problem is that by definition you need a change of behavior resulting from individual experience,” he said. ”That is the only way to show that learning, or memory, has occurred. This means that all the approaches based on molecular biology”finding such-and-such a receptor or neuron X”are of no use whatsoever unless you demonstrate a change in behavior. When a given behavior changes, you can go and look into the box and find the molecules. But if you go looking for molecules without the change in behavior, you can say nothing. Learning manifests itself once the individual”s behavior changes. Changes at the cellular level are not necessarily relevant to this. And so, with plants, you need a visible change of behavior. That”s the challenge. But with the plant you mentioned, which moves, someone should be able to find something.”

I asked him to comment further on how other scientists had received his research. He said that when he travels and presents his work, it is not questioned much. Rather, it leads other scientists to ask themselves questions they have not previously considered. He regards this as a success, even if his work does not provide answers to his main questions.

When asked if he could suggest any other scientists doing work on intelligence in nature, he mentioned an Austrian biologist studying cognition in amoebas. He also suggested several j.a.panese research teams: one working on color vision among insects, another on cricket olfaction, and a third on b.u.t.terfly neurology. ”Go to j.a.pan,” he said.

I left Martin Giurfa in front of his laboratory in the early afternoon. We promised to stay in touch. I felt elated, but also a bit dazed. I had come half expecting to meet a cold scientist. Instead I found a fellow who encouraged me to keep looking into intelligence in nature. I felt as if he had given me a license to continue deeper into unknown territory.

Chapter 6.