René Descartes was a very clever thinker. He proved his own existence, declaring that because he thought, he must exist: “I think, therefore I am.”
But the 17th century philosopher-mathematician-scientist committed a serious mental blunder when he decided that the mind doing the thinking was somehow separate from the brain it lived in. Descartes believed that thought was insubstantial, transmitted from the ether to the pineal gland, which played the role of something like a Wi-Fi receiver embedded deep in the brain. Thereafter mind-brain dualism became the prevailing prejudice. Nowadays, though, everybody with a properly working brain realizes that the mind and brain are coexistent. Thought processes and associated cognitive mental activity all reflect the physics and chemistry of cells and molecules inhabiting the brain’s biological tissue.
Many people today do not realize, though, that there’s a modern version of Descartes’ mistaken dichotomy. Just as he erroneously believed the mind was distinct from the brain, some scientists have mistakenly conceived of the brain as distinct from the body. Much of the early research in artificial intelligence, for instance, modeled the brain as a computer, seeking to replicate mental life as information processing, converting inputs to outputs by logical rules. But even if such a machine could duplicate the circuitry of the brain, it would be missing essential peripheral input from an attached body. Actual intelligence requires both body and brain, as the neurologist Antonio Damasio pointed out in his 1994 book, Descartes’ Error.
“Mental activity, from its simplest aspects to its most sublime, requires both brain and body proper,” Damasio wrote.
That is why the brain is wired to various body parts by a peripheral nervous system. It’s not just for the brain to send messages telling the body how to move; the body also sends messages back, e-mailing the brain directly via electrical signaling along nerve fibers. And slower supplemental communication (sort of like snail mail) consists of chemicals that travel from the body to the brain via the bloodstream.
“I am not saying that the mind is in the body,” Damasio emphasized. “I am saying that the body contributes more than life support and modulatory effects to the brain. It contributes a content that is part and parcel of the workings of the normal mind.”
Consequently, efforts to mimic human intelligence must keep in mind that the mind and brain require a body. What’s true for real brains should also constrain artificial ones. Otherwise efforts to create artificial intelligence in computers or robots are doomed.
“What we now call cognition or intelligence (and the brain, its physical ‘substrate’) has always evolved as part of a complete organism that had to survive and reproduce in the real world,” Rolf Pfeifer and colleagues write in the August issue of Trends in Cognitive Sciences.
Interactions of the brain, body and environment govern both the evolution of intelligent species and the development and growth of individual organisms. “We have to investigate how brain, body, and environment interact to understand brain-body coevolution,” argue Pfeifer, of the University of Zurich in Switzerland, and coauthors Fumiya Iida and Max Lungarella. But the details of these processes have been left largely unexplored.
“Although there seems to be increasing agreement that the body plays an essential role in cognition,” they write, “there has been relatively little work on detailing what the connection between brain and body looks like, how it shapes and drives our actions, and how it manifests itself in brain processing and behavior.”
Pfeifer and collaborators, all artificial intelligence researchers, believe research in robotics may prove helpful in unraveling the ways embodiment influences intelligence in real brains. But those robots need to recognize that it’s not just a matter of a brain communicating with limbs and organs. Shape, form, and the materials that a body is made of are crucial elements in understanding the embodiment-intelligence connection. In particular, real bodies, unlike computers, are largely made of “soft” materials. (Your skeleton provides only 15 percent or less of your body weight.) The deformability of muscles and skin control much of the brain’s input, enabling you to maintain balance while walking or judge the shape of objects you touch or grasp.
“It is as if the brain were outsourcing some of the control — or computation — to morphology and materials,” the researchers write.
In recent years, artificial intelligence experts have developed a field called “soft robotics” in which flexible materials replace some of the old-style rigidity of metallic robots. Such robots have demonstrated interesting skills, but not yet anything like human brainpower.
“What has been missing so far is an understanding of the relation of soft robotics to cognition and intelligence,” Pfeifer and coauthors suggest.
They believe that research focusing on material properties, shape and form can help transform the limited reach of pure computational power to systems that can evolve, learn and develop cognitive skills analogous to the body-brain team that gives people their mental prowess. This approach, the researchers say, will teach “very powerful lessons about how to learn from biology and how to build better robots.” Its success would reaffirm the importance of avoiding false divisions between brain and body as well as brain and mind.
So the next time you encounter people who try to defend Cartesian dualism, remind them of the old joke where Descartes walks into a bar and orders a beer. “Do you want nuts with that?” asked the bartender. “I think not,” Descartes replied, and disappeared.
Follow me on Twitter: @tom_siegfried