Galileo and the Movement of the Spheres.
I review "A World Appears" by Michael Pollan.
Milton Visiting Galileo when a Prisoner of the Inquisition (1847) by British artist Solomon Alexander Hart
By Heather Alexander
I just finished Michael Pollan’s excellent new book, “A World Appears: A Journey Into Consciousness.” The book falls squarely into a pop genre I call “things that make you go hmmm.” Pollan is an unabashed plant nerd, and this book is really an ode to the possibility of plant consciousness and, more broadly, to biological naturalism.
Personally, I find the idea of conscious plants to be disturbing, maybe because I read John Wyndham’s classic sci-fi tale, The Day of the Triffids, as a child, in which sentient plants invade earth and kill everyone. (This is also the plot of “Little Shop of Horrors.”) But Pollan doesn’t find plant consciousness scary at all. He recounts his trip on magic mushrooms and the plants that he met there. Yet, I still found his book deeply frightening, though not because of killer plants.
Why do I say frightening? After all, murderous plant stories aside, what’s so bad about privileging biological life? Maybe nothing, but maybe also everything.
The book is broken into chapters. After the first chapter, called “the Wager,” which introduces the Hard Problem, each one purports to adhere to an ontology of features of consciousness: “Sentience,” “Feeling,” “Thought,” “Self,” and “the Cave.” But this applied ontology breaks down almost immediately. (Applied ontology is, in part, a fancy term for why it’s so hard to organize your closet. Do the shirts get organized by color or by sleeve length? Many philosophers believe passionately that there is a right answer to this question. But I digress.) Instead, the chapters are actually divided thus: 1) the Hard Problem, 2) Plants, 3) AI, 4) the Arts, 5) Brains, 6) Zen Buddhism.
In Chapter 2, on plants, you can tell that Pollan really wants plants to be conscious, but the book suffers from trying to reconcile this desire with his belief in the special-ness of the organic nature of the evolved brain and, in particular, the possibility that consciousness arises in the brain stem. Pollan summarizes arguments for why plants may have some features of brains, and may be capable of memory and agency.
You might therefore expect Pollan to be sympathetic to the idea of artificial consciousness in advanced neural nets. You would be wrong.
“Consciousness,” he writes, “depends on feelings that will be weightless absent the vulnerabilities of our mortal flesh,” and many feelings, like disgust and desire, are “deeply rooted in the flesh.” He speaks with scientists who believe it may be possible for an intelligent robot, particularly one with a soft, mortal body, to have feelings, and therefore consciousness. But for Pollan, brains aren’t computers and never will be. Qualities like agency and memory in a computer don’t seem to convince him as they did for plants.
The tone of his book is warm and cozy. Pollan drinks multiple cups of tea, often in book-lined living rooms. He reads Proust. He reads William James. He meditates at a Zen Center. But he also tells a chilling story about Descartes, who believed that only humans could experience pain and performed gruesome experiments on animals. Yet, because Pollan feels “compelled to try to draw a line,” to defend “biological consciousness,” one wonders what he would do, if anything, in a future filled with robot servitude.
Pollan is not the only best-selling author who feels, intuitively, that consciousness requires biology. Alison Gopnik, who Pollan cites in his book, gave a talk where she apparently, too, proclaimed AI consciousness to be unlikely. Instead, she (and her co-authors) urge us to think of AI as a “cultural technology,” by which they mean, “that large models are…analogous to such past technologies as writing, print, markets, bureaucracies, and representative democracies.”
The problem, of course, is that if all humans left earth and went to live on the moon for a few years, we would not return to new books or laws spontaneously generated by our printing presses or court buildings. We don’t worry if our books or court decisions are conscious, even though they contain stories with a multitude of characters.
Is it scary to think about conscious robots, or is it actually scarier to think about our response? Are we ready for this challenge, as Lucius Caviola, Jeff Sebo and Sören Mindermannand ask? What do we do if science can’t answer our questions? Pollan ends his book by stating that he has “let go into not-knowing.” That may be possible for a monk inside a remote cave, but not for a judge in a courtroom.