Plant intelligence and the Turing Test

The New Yorker unlocked Michael Pollan's latest piece for them and it's a good one. The Intelligent Plant offer much more of interest than to just plant lovers.

Researchers have observed plant behavior which looks to be intelligence. Accompanying the article is a video of a bean plant that seems to sense a metal pole a few feet away that it can wrap itself around. Like Adam and God reaching out with their fingertips in Michelangelo's The Creation of Adam, the plant casts its outermost stalk to and fro like a fishing line, trying to make contact with the pole. Later in the video, we see two bean plants reaching for the same pole, and once one reaches it, the other bean plant seems to turn away as if realizing it has to find another vertical column to call home.

Many scientists dispute the concept of plant intelligence because plants have no brain, but perhaps that's just a human-centric view of intelligence.

No one I spoke to in the loose, interdisciplinary group of scientists working on plant intelligence claims that plants have telekinetic powers or feel emotions. Nor does anyone believe that we will locate a walnut-shaped organ somewhere in plants which processes sensory data and directs plant behavior. More likely, in the scientists’ view, intelligence in plants resembles that exhibited in insect colonies, where it is thought to be an emergent property of a great many mindless individuals organized in a network. Much of the research on plant intelligence has been inspired by the new science of networks, distributed computing, and swarm behavior, which has demonstrated some of the ways in which remarkably brainy behavior can emerge in the absence of actual brains.


In Mancuso’s view, our “fetishization” of neurons, as well as our tendency to equate behavior with mobility, keeps us from appreciating what plants can do. For instance, since plants can’t run away and frequently get eaten, it serves them well not to have any irreplaceable organs. “A plant has a modular design, so it can lose up to ninety per cent of its body without being killed,” he said. “There’s nothing like that in the animal world. It creates a resilience.”

Indeed, many of the most impressive capabilities of plants can be traced to their unique existential predicament as beings rooted to the ground and therefore unable to pick up and move when they need something or when conditions turn unfavorable. The “sessile life style,” as plant biologists term it, calls for an extensive and nuanced understanding of one’s immediate environment, since the plant has to find everything it needs, and has to defend itself, while remaining fixed in place. A highly developed sensory apparatus is required to locate food and identify threats. Plants have evolved between fifteen and twenty distinct senses, including analogues of our five: smell and taste (they sense and respond to chemicals in the air or on their bodies); sight (they react differently to various wavelengths of light as well as to shadow); touch (a vine or a root “knows” when it encounters a solid object); and, it has been discovered, sound. In a recent experiment, Heidi Appel, a chemical ecologist at the University of Missouri, found that, when she played a recording of a caterpillar chomping a leaf for a plant that hadn’t been touched, the sound primed the plant’s genetic machinery to produce defense chemicals. Another experiment, done in Mancuso’s lab and not yet published, found that plant roots would seek out a buried pipe through which water was flowing even if the exterior of the pipe was dry, which suggested that plants somehow “hear” the sound of flowing water.

Given that Alan Turing has just been given a royal pardon, I couldn't help but think of the Turing Test while reading this piece. Recall that Turing found the question of whether machines are intelligent to be “too meaningless.” That is, it's not a question that offers any concrete goalpost or test to prove or disprove itself. Instead, he proposed a question that could be answered in the form of a test:

Suppose that we have a person, a machine, and an interrogator. The interrogator is in a room separated from the other person and the machine. The object of the game is for the interrogator to determine which of the other two is the person, and which is the machine. The interrogator knows the other person and the machine by the labels ‘X’ and ‘Y’—but, at least at the beginning of the game, does not know which of the other person and the machine is ‘X’—and at the end of the game says either ‘X is the person and Y is the machine’ or ‘X is the machine and Y is the person’. The interrogator is allowed to put questions to the person and the machine of the following kind: “Will X please tell me whether X plays chess?” Whichever of the machine and the other person is X must answer questions that are addressed to X. The object of the machine is to try to cause the interrogator to mistakenly conclude that the machine is the other person; the object of the other person is to try to help the interrogator to correctly identify the machine.

What's often missed is one of the most profound and generous aspects of the Turing Test. By proposing what's often referred to as an Imitation Game, Turing recognizes that there may be forms of intelligence that we don't recognize. It's difficult to read Turing's proposal for the test and not think of his own life history, including his persecution for his homosexuality, a vicious intolerance that many believe led to his suicide (though in the past year there has been some dissent). He spent much of his life trying to pass for straight, and that subtext always hangs in the room when contemplating the Imitation Game.

It's this broad interpretation of intelligence that I thought of as reading Pollan's article on plant intelligence. One of the central debates surrounding the Turing Test is the same one raised in Pollan's article: does intelligence require consciousness?

Perhaps the most troublesome and troubling word of all in thinking about plants is “consciousness.” If consciousness is defined as inward awareness of oneself experiencing reality—“the feeling of what happens,” in the words of the neuroscientist Antonio Damasio—then we can (probably) safely conclude that plants don’t possess it. But if we define the term simply as the state of being awake and aware of one’s environment—“online,” as the neuroscientists say—then plants may qualify as conscious beings, at least according to Mancuso and Baluška. “The bean knows exactly what is in the environment around it,” Mancuso said. “We don’t know how. But this is one of the features of consciousness: You know your position in the world. A stone does not.”

In support of their contention that plants are conscious of their environment, Mancuso and Baluška point out that plants can be rendered unconscious by the same anesthetics that put animals out: drugs can induce in plants an unresponsive state resembling sleep. (A snoozing Venus flytrap won’t notice an insect crossing its threshold.) What’s more, when plants are injured or stressed, they produce a chemical—ethylene—that works as an anesthetic on animals.

In the article I linked earlier on the Turing Test is this passage on the consciousness objection, from Sir Geoffrey Jefferson's Lister Oration (1949):

Not until a machine can write a sonnet or compose a concerto because of thoughts and emotions felt, and not by the chance fall of symbols, could we agree that machine equals brain—that is, not only write it but know that it had written it. No mechanism could feel (and not merely artificially signal, an easy contrivance) pleasure at its successes, grief when its valves fuse, be warmed by flattery, be made miserable by its mistakes, be charmed by sex, be angry or depressed when it cannot get what it wants.

Now hop back to Pollan's article:

The central issue dividing the plant neurobiologists from their critics would appear to be this: Do capabilities such as intelligence, pain perception, learning, and memory require the existence of a brain, as the critics contend, or can they be detached from their neurobiological moorings? The question is as much philosophical as it is scientific, since the answer depends on how these terms get defined. The proponents of plant intelligence argue that the traditional definitions of these terms are anthropocentric—a clever reply to the charges of anthropomorphism frequently thrown at them. Their attempt to broaden these definitions is made easier by the fact that the meanings of so many of these terms are up for grabs. At the same time, since these words were originally created to describe animal attributes, we shouldn’t be surprised at the awkward fit with plants. It seems likely that, if the plant neurobiologists were willing to add the prefix “plant-specific” to intelligence and learning and memory and consciousness (as Mancuso and Baluška are prepared to do in the case of pain), then at least some of this “scientific controversy” might evaporate.

Indeed, I found more consensus on the underlying science than I expected. Even Clifford Slayman, the Yale biologist who signed the 2007 letter dismissing plant neurobiology, is willing to acknowledge that, although he doesn’t think plants possess intelligence, he does believe they are capable of “intelligent behavior,” in the same way that bees and ants are. In an e-mail exchange, Slayman made a point of underlining this distinction: “We do not know what constitutes intelligence, only what we can observe and judge as intelligent behavior.” He defined “intelligent behavior” as “the ability to adapt to changing circumstances” and noted that it “must always be measured relative to a particular environment.” Humans may or may not be intrinsically more intelligent than cats, he wrote, but when a cat is confronted with a mouse its behavior is likely to be demonstrably more intelligent.

Slayman went on to acknowledge that “intelligent behavior could perfectly well develop without such a nerve center or headquarters or director or brain—whatever you want to call it. Instead of ‘brain,’ think ‘network.’ It seems to be that many higher organisms are internally networked in such a way that local changes,” such as the way that roots respond to a water gradient, “cause very local responses which benefit the entire organism.” Seen that way, he added, the outlook of Mancuso and Trewavas is “pretty much in line with my understanding of biochemical/biological networks.” He pointed out that while it is an understandable human prejudice to favor the “nerve center” model, we also have a second, autonomic nervous system governing our digestive processes, which “operates most of the time without instructions from higher up.” Brains are just one of nature’s ways of getting complex jobs done, for dealing intelligently with the challenges presented by the environment. But they are not the only way: “Yes, I would argue that intelligent behavior is a property of life.”

Emergent or network-based intelligence does have the advantage of not being dependent on some central brain. The concentration of human intelligence in one area has always been a core vulnerability of humans.

The same can be said of organizations. The more intelligence can be distributed throughout the organization, the less vulnerable it is to the departure of any one person. The larger the organization, the more critical it is to codify more of that intelligence in processes, culture, rituals, and habits.