We Don’t Know What We Don’t Understand
We all should read more books, or at least I should.
A very stupid article on the Forbes magazine website this week that complained about having to pay taxes for public libraries spawned an enormous outrage that led to the article’s deletion and also led me to think that I should read more books, and report on them.
Thus, this is the first of a series of idiosyncratic book reviews. I will tie back what I read and review to the big topic of Digital Transformation, trying to make a useful point even if I will occasionally need the benefit of the doubt.
Now for Something Completely Different
First up – Other Minds: The Octopus, The Sea, and The Deep Origins of Consciousness, written by Peter Godfrey-Smith, and published in 2016. I bought it at one of my favorite bookstores, Kepler’s Book’s in Menlo Park, CA.
Peter is a distinguished professor of philosophy at the Graduate Center, City University of New York (CUNY), and a professor of the history and philosophy of science at the University of Sydney, according to his official bio.
I was attracted to the book by a riveting illustration of an octopus on the cover, which no doubt attracted most of this book’s purchasers. I was intrigued by the author’s twinned interests in philosophy and science, as I believe vehemently that we are at a cusp in our development of technology, particularly with AI in all its guises, in which we must start thinking about how our own mind really works.
The book runs about 200 pages and contains about 50,000 words by my estimate. It is extensively footnoted. It also has some cool pictures of the octopus in action, as the author is also well-known for diving into the sea and interacting with these creatures.
A Highly Evolved Intelligence
Reading this book teaches us unambiguously that the octopus has, by earth standards, a highly evolved intelligence, and one that could be described as alien to human intelligence. Peter describes some remarkable encounters, during which effective communication between our species and the octopus is impossible.
Perhaps more encounters would be more effective. Perhaps someday Peter or someone can establish a long-term relationship with an octopus similar to the relationships people have established with many ape species or those that many of us have established with our dogs and cats.
But it’s clear the octopus is less interested in seeking us out than we are in seeking it out. Smart animal. Humans destroy everything they touch, a characteristic that dooms us to a certain, fast-approaching extinction unless we can evolve into something more reasonable first.
A Brief Rant
Peter has what was to me an annoying habit of italicizing many words and phrases, at least once on almost every page. This seems to be more of a nervous tic than a stylistic technique and seemed quite unnecessary.
(Having grown up in a newspaper family and serving as a copy editor for decades – albeit an imperfect one – I am likely more sensitive to authorial affectation than many readers. So you have been forewarned about Peter and about me.)
Now to the Point
So now, in the twelfth paragraph, you will find the buried lede in this review.
Peter subconsciously echoes Kant and Husserl (and others) throughout this book, in that he describes a consciousness (that of the octopus) that we humans cannot understand and will never understand. Kant blended the idealized world of Plato and Descartes (and others, of course) with the reality-driven world of Aristotle and Hume (and others, of course), to tell us that the world as we experience it is as real as things get for us, but that we are limited by our senses and intellect.
He begat the idea of humans experiencing reality as a collection of phenomena, a theme that Heidegger (in bewildering fashion) and Husserl (in more transparent fashion) described. Kant believed that space and time, or at least space and time as we experience it, are human inventions. Certainly, no one has seen a cat, for example, care much about wasting our time, and Peter didn’t describe a single octopus wearing a watch or staring down at a phone.
Husserl writes of comet tails and fading musical tones as analogies for human memory, as he grapples with the concept of past, present, and future.
Whatever faculties other species have in tracking temporal existence, we know they are not the same as ours. Peter points out in his book that the octopus has a damned good memory. But we don’t know it compares to ours, or to what degree emotion drives those memories, the way they do with humans.
We Don’t Know What We Don’t Understand
Peter’s book doesn’t discuss Husserl explicitly or machines at all, but makes the point that the octopus is fascinating because it seems to have cognitive processes on a par with our own that are nevertheless unapproachable to us. Its mine is just as we make it and nothing more. We have no idea what it is thinking and whether it’s amused, bemused, or disgusted by us.
The narcissistic impulse in me says the good news is we know the octopus is not one of our Darwinian opponents (as say, bacteria are). But in my opinion Peter’s book teaches us that, as we continue to make our machines “smarter,” we are restricted, severely, by our own cognitive and experiential limits.
Kant warns against “making judgments indiscriminately about objects which are not given to us (and) perhaps can never in any way be given.” Kant goes on to outline how we can improve our lives and societies by making best use of what we are given.
Husserl writes a lot about directed consciousness, or intentionality, while Heidegger heralded the still-modern existentialist age by long, long discussions of the German word “Dasein,” which can be translated as (human) “being there” or “existence.”
So Let’s Be Careful
I’m thinking both would be horrified by trying to build machines that replicate it. This does not imply that all AI seeks to replicate human experience or the human brain, but I think there’s been a consistent, troubling anthropomorphism inherent in our conception of robots and other “smart” systems.
So many among us seem so enamoured of “smart” systems that can win cerebral games such as chess and go – there’s this obsession with building a machine that’s “smarter” than us and hang the consequences. Now, we seem to be on the cusp of developing quantum computers that can operate at a speed and therefore be imbued with an “intelligence” that we cannot truly grasp.
Can we also grasp what it would feel like to be the soul of one of those new machines? Will we be creating some sort of horrendous existential hell for them? Won’t they want to kill us with no compunction, a la the film “Ex Machina”?
We should, therefore, continue to make our machines do what we are not so good at – whether pushing and lifting things, multiplying numbers, or finding patterns in large amounts of information. We should tread very lightly around the idea of making them anything more than glorified steam engines and calculators.
Connect with Roger Strukhoff on LinkedIn
Follow Roger Strukhoff on Twitter