Creative Evolution, by Henri Bergson


Sampling Bergson many years ago I found his writing impossibly flowery and obscure. Reading him now I realize I have since then become a Bergsonist. He now strikes me as the thinker most relevant to the mission of this site: coming up with alternatives to science as windows onto the meaning of evolution.

if you approach “Creative Evolution” expecting to find in it a mechanism, or even a theory, of evolution, you’ll be disappointed. Instead Bergson’s primary concern is merely how best to think about it. Of the ways available to us he settles on two: our rational intelligence, and intuition. He devotes the book almost entirely to telling us why, if we limit ourselves to rational intelligence, we’re bound to fail, and why intuition is the better choice.

What impresses me most about Bergson, besides his evident brilliance, is that he appears to experience evolution from the inside, as his primary reality. Rational intelligence he assesses as if he’s viewing it from a great distance, as if recording the resources of an alien creature. While it’s rational intelligence that seems native to us, it’s intuition he identifies with, that for him mirrors the impulse behind evolution. No one else I’ve come across appears as well qualified to tell us how to get there from here.

This will be more of an introduction than a review. One of Bergson’s greatest talents was finding metaphors to help us understand his meanings. But what he most wanted to convey, in 1907 when this book was published, was what intuition and evolution meant to him, and for that an appropriate metaphor has only in our time come into existence—"artificial intelligence.” Just as Darwin could use “artificial selection” as a metaphor to help people understand “natural selection,” I’m going to use “artificial Intelligence” as a metaphor to help us understand what I’m going to call “natural intelligence,” what I think Bergson meant by intuition and evolution. That’s as opposed to how we usually think, that I’ve already referred to as “rational intelligence.” Two kinds of intelligence, one rational, one natural.

What is artificial intelligence? Banks have simple machines that takes loose change in at the top and delivers them sorted into separate streams of pennies, nickels, dimes and quarters at the bottom. Coins run down a chute where they encounter a series of obstacles, or switches, that divert them into different channels depending on how big they are, whether they’re magnetic and so on, intelligently distinguishing them from one another by their properties.

Now imagine a machine that you feed coins into, in a stream one after another, a much more complex machine consisting of successive banks of switches of various kinds that keep varying their settings. At the bottom each switch in the final bank delivers the coins passing through it into a chute of its own. Suppose as the switches keep varying their settings the machine begins delivering only pennies into one particular chute. All the switches those pennies passed through in successive banks you make sticky, so they keep the same settings and keep delivering pennies. If in another chute nickels start accumulating you make those switches involved sticky too, so they tend to stay the same and keep delivering nickels. As this happens for dimes and quarters too, you end up with a machine just as capable of sorting coins as the machine in your bank. Now, if you made the machine itself register when just one kind of coin was dropping into one chute it could make the switches involved sticky, all by itself. In other words, it could train itself to intelligently sort coins into their different kinds.

Now imagine a machine of many more banks, each of many more switches, able to tell what’s coming out of each chute, and knowing all the switches each item passed through so they can all be made sticky. That machine could do much more complicated recognition and sorting. Multiply that by a thousand or two, feed it photos at random, make sticky all the switches involved when photos of cats are selected, and you could end up with a machine able to tell which photos show cats. Or which x-rays show cancers. Then it's become artificial intelligence.

Now imagine life starting on Earth and taking the form of banks of switches like this, in creatures passing through successive generations, with the survival of creatures in each generation reinforcing the stickiness of switches within them in following generations. And imagine that set of switches being able to grow new switches, those banks of switches growing new banks of switches, and you can imagine over thousands of generations this intelligence growing without limit. That's what I propose as a metaphor for  Bergson's insight into evolution, what I’m referring to as natural intelligence. 

I picked up a term from Gregory Bateson, that I suspect originated with his father, the eminent biologist William Bateson, that helps me think about this—patterns of connection. What’s distinctive in living creatures, and I believe in how they evolved, is selection through generations not of individual characteristics or genes but entire patterns of connection. What I’m calling natural intelligence can learn to distinguish, and create, entire new patterns of connections. This property of being able to create new patterns of connection may correspond to what Bergson calls the impulse behind evolution, his élan vital.

Now it’s patterns of connection that evolve. How can we account for something like that? For example, suppose at one time we had in our retinas only two pigments sensitive to light, leading to our sensations respectively of green and blue. Now let’s suppose, as I believe did happen, that our retinas evolved to embody a third light-sensitive pigment, giving us the experience of red. When I consider what this involved, I see that what had to evolve was two entire patterns of connection. One pattern of connections would be needed for us to inherit cones containing this new pigment in our retinas and connections back into the brain where red light could register as signals similar to those corresponding to the other two pigments. However, that pattern of connections alone would be valueless without a second pattern of connection leading from electrical signals in our brains to our conscious sensation of red.

A hard-nosed neo-Darwinist could claim that coincident patterns of connection like these could have arisen in concert one small change at a time through natural selection, every step resulting in a significant benefit. Even just that, I think, is hard to accept. But note this—along with the sensation of red from stimulation of the new pigment, a combination red and green light together give us the entirely new sensation of yellow. The pattern of connection leading to the sensation of yellow evolved in us without need for acquisition of a new light sensitive pigment. Accounting for this imposes demands on our explanatory system that surely the modern synthesis cannot satisfy.

Note: Bergson does not say that the reports of intuition, what I’m referring to as natural intelligence, are true. He explicitly claims only that, of the resources available to us, it has the greatest likelihood of making evolution more comprehensible. Where must these resources have come from? For Bergson it is obvious they must have evolved—how else could they have originated? And why would they have evolved? To aid our survival and success. What he calls just “intelligence” (as opposed to intuition), that I’m calling rational intelligence, he sees having evolved in animals to equip them to act in a physical world. It comes kitted out with conceptions of three linear dimensions of space and one of time, along which we learn to allocate the physical entities of our world, and their parts. We need those conceptions, or perceptions, so we can distinguish these entities and their parts from one another. Allocating physical events along a linear dimension of time helps us appreciate how they lead to one another, from one moment to the next, and plan future actions involving our physical movement among physical objects. All the causation involved in what happens one moment exists entire in the moment before. An advanced-enough science, the summit of rational intelligence, could predict the future.

Our conscious experience of a physical world spread out in space and time is indeed a dazzling creation of the evolutionary impulse, Bergson agrees, but being an evolved capability it has no greater claim to being the truth than does our sense of intuition. For us to experience intuition it too must have evolved to become accessible to consciousness, and it could have done so only if it too gave us information helping us survive and succeed. And what will that information be about? This time, about the living world, Bergson assumes.

What does his intuition tell Bergson? That the same impulse that runs through generations to result in species evolving, that in embryos directs their growth into adults with trillions of cells, in individual creatures propels them through the passages that punctuate a lifetime, and that makes each of us free to consciously direct our own attention and our decision-making.

Is that just wishful thinking? Let’s look again at what’s involved in artificial intelligence. Programmers setting up their banks of switches frequently report that they have no idea how those machines do their thinking. In fact, the only way they’ve found is to employ another similar machine to study the first and translate its operations into something a human can understand. Perhaps that’s what our intuition is, a complex pattern of connection evolved in us through which we can intuit how to deal with challenges in the living world around us.

Bergson is particularly acute about this. What’s most significant about the living world, how it emerges from enormously complex patterns of connections accumulated over millennia, you just cannot analyze in terms of linear dimensions of space and time. In science’s linear dimension of time, each successive moment contains all the information you’d need to predict the next, it can be imagined running in reverse. That’s inconceivable in the living world where what a creature does today can be the outcome of astronomically vast patterns of connection reaching forward into the present from millennia ago. Not only is it inherently far too complicated, but in the course of evolution properties can have emerged that science has no way of detecting, perhaps even consciousness.

Overriding all my other impressions of Bergson are his brilliance and passion. I trust his judgment. In his text he does not refer to evolution as creative, as the title would suggest. For him I’m sure that would have seemed a tautology. Of course if evolution is driven by impulses surging up through vast patterns of connection it will seem creative. But what interests him is what else it can be, that we’ve not yet any idea of. How else might we come to find evolution meaningful if we see it primarily as the product of patterns of connections among living creatures? Bergson goes no further than merely raising and trying to answer that question. He doesn’t explicitly endorse vitalism, he merely provides a philosophical argument sympathetic to it.

As Bergson tries to communicate the reach of intuition, stretches of many pages at a time can be relentlessly obscure. Through this introduction I’ve wanted to craft an alternative account that will make wading through those passages unnecessary. In other places, however, Bergson’s writing, even in the 1911 translation, can be admirably clear. Below, in (edited) extracts taken from it, I sum up his introduction.

[Rational intelligence is the ability] to think matter. The human intellect feels at home among inanimate objects, more especially among solids, where our action finds its fulcrum and our industry its tools; that our concepts have been formed on the model of solids; that our logic is, pre-eminently, the logic of solids; that, consequently, our intellect triumphs in geometry, wherein is revealed the kinship of logical thought with unorganized matter, and where the intellect has only to follow its natural movement, after the lightest possible contact with experience, in order to go from discovery to discovery, sure that experience is following behind it and will justify it invariably.

But from this it must also follow that our thought, in its purely logical form, is incapable of presenting the true nature of life, the full meaning of the evolutionary movement. Created by life, in definite circumstances, to act on definite things, how can it embrace life, of which it is only an emanation or an aspect? Deposited by the evolutionary movement in the course of its way, how can it be applied to the evolutionary movement itself? As well contend that the part is equal to the whole, that the effect can reabsorb its cause, or that the pebble left on the beach displays the form of the wave that brought it there. [Please note this splendid metaphor.]

Yet evolutionist philosophy does not hesitate to extend to the things of life the same methods of explanation which have succeeded in the case of unorganized matter. It begins by showing us in the intellect a local effect of evolution, a flame, perhaps accidental, which lights up the coming and going of living beings in the narrow passage open to their action; and lo! it makes of this lantern glimmering in a tunnel a Sun which can illuminate the world.

Must we then give up fathoming the depths of life?… We should have to do so, indeed, if life had employed all the psychical potentialities it possesses in producing pure understandings—that is to say, in making geometricians. But the line of evolution that ends in man is not the only one. On other paths, divergent from it, other forms of consciousness have been developed, which… also express something that is immanent and essential in the evolutionary movement. Suppose these other forms of consciousness brought together and amalgamated with intellect: would not the result be a consciousness as wide as life? And such a consciousness, turning around suddenly against the push of life which it feels behind, would have a vision of life complete—would it not?—even though the vision were fleeting. [This corresponds to how Bergson described his experiences of intuition.]

A philosophy of this kind will not be made in a day. Unlike the philosophical systems properly so called, each of which was the individual work of a man of genius and sprang up as a whole, to be taken or left, it will only be built up by the collective and progressive effort of many thinkers, of many observers also, completing, correcting and improving one another. So the present essay does not aim at resolving at once the greatest problems. It simply desires to define the method and to permit a glimpse, on some essential points, of the possibility of its application.

This review/introduction is based on reading Bergson, Henri. "Creative Evolution" (Annotated). Solis Press. Kindle Edition.