Storythinking with Dan & Tracy (Ch. 6)
The Brain Machinery of Storythinking
Storythinking: The New Science of Narrative Intelligence, Angus Fletcher, 2023
Table of Contents (evolving)
Chapter 1 (hosted by Dan)
Chapter 2 (hosted by Tracy)
Chapter 3 (hosted by Dan)
Chapter 4 (hosted by Tracy)
Chapter 5 (hosted by Dan)
Chapter 6 (this post)
Sparks and Soups. These are the two groups who were debating the nature of neural synapses, the gap between charged neurons. The Sparks were for continuity: “the charge itself hopped the gap, zapping like a lightning bolt across the synapse.” The Soups were for discontinuity: “the synapse was nonelectronic. It was traversed not by electrical brainwaves but by chemicals dumped into the gap.” (80) The story Fletcher tells about the Sparks and Soups and how they ultimately resolved who was right involves the characters John Eccles and Karl Popper, the latter being the famed philosopher of science. Fletcher traces the origins of the scientific method, which transcends Aristotelian deduction and Baconian induction to the necessary combination of logic and storytelling that marks the best human thinking.
Dan’s comments: The story of the "sparks" and the "soups" was one I hadn't heard before. Once again, I appreciate that Fletcher has uncovered interesting stories in the past and has used them to try to say something about the present. This is what keeps me reading. His perspective on Karl Popper is also interesting. I know a bit more about Popper than Eccles, so I suppose I could assess how accurately I think his depiction of Popper is, and then apply that to how I think he has portrayed the others I don't know.
As in the previous chapter, when we were introduced to some unfamiliar but influential characters, this one also does not disappoint in the historical storytelling. We can admire Eccles for his scientific integrity — when he finally implants an electrode in the poor cat and does the decisive experiment, he finds out the synaptic gap is chemical after all, and he abandons Sparkdom for Soupdom. But it’s really John Herschel and Karl Popper who steal the show in this chapter.
Dan: Amid some gobbledygook about "astral telescopes" and "atomic microscopes", Fletcher admits that induction (which he prefers to deduction) was also described in Aristotle's Organon. I think this sort-of kicks a leg out from under his stool. He goes out on a limb, I think, with his point about the "feudal universities" preferring deduction because it helped retard speculation and preserve the "old theories...cohesion and stability." I object to the implications of this train of thought specifically on the basis of storythinking: I think he misunderstands what motivated thinkers in the medieval world.
Herschel (1792-1871) was “born with a butterfly intelligence” and “flitted between scholarly flowers” across a vast range of subject areas. He also was motivated to out-achieve his father, a famous astronomer of global celebrity status (86).
Stricken by a filial sense of obligation to reach an equal zenith of achievement, young John procured his own telescope and chained himself to its eyepiece each twilight, determined to map every glowing object in the sky. This epic task required so much discipline that it turned young John machine. Ruthlessly restraining his nomadic curiosity, he cranked through automatic midnight after automatic midnight, surveying the sable constellations with such inflexible precision that he logged coordinates (checked in triplicate) for more than 70,000 stars. Until his dazzling persistence was rewarded with the Royal Society’s Copley Medal… twice. (86-87)
Dogged persistence was ultimately of no matter. Even 70,000 observational data points could not produce any brilliant science. Induction “simply didn’t work,” or rather it didn’t until John “released (at last) his butterfly brain.” He conjured a new hypothesis:
Instead of distilling millions of inductions into logical law, the scientist’s mind jumps from a few data points to hypothesize a provisional cause, which it uses to narrate bold predictions that can be verified via experiment. (87)
The modern scientific method is born. Jump from a few data points, speculate by letting the mind “leap” about causes, and then test them. This is how Copernicus achieved his revolution and Darwin his. Hershel, it turns out, is buried between the tombs of Newton and Darwin in Westminster Abbey. (88)
Dan: John Herschel's contribution, in Fletcher's narrative, seems to be that he theorized a sort-of short-circuiting of the inductive process, where "the scientist's mind jumps from a few data points to hypothesize a provisional cause" (Fletcher's italics). This dovetails nicely with Popper's focus on falsification, whom Fletcher jumps back to in his story. He's quite right, of course, when he says disconfirmation is much easier than confirmation and that in fact, no theory is ever proven. The failure of logic to provide certainty isn't necessarily a reason to abandon it, although I agree that people who have fallen into a scientism that believes this might benefit from being reminded.
Six decades later, Karl Popper (1902-1994), inspired by Einstein, whose “adventurous prediction was put to the test in 1919” by Eddington and Dyson (90), finished the job Herschel had begun. Karl Popper had begun his career smashing and debunking Marx and Freud (82-83), and he was ready to do the same in Einstein’s case. Popper, too, is haunted by his dual “split mind.”
Popper was enthralled—yet skeptical. Typically split of mind, he was convinced that Einstein’s theory differed from the pseudoscience of Marx and Freud—yet simultaneously convinced that it was logically incomplete. And as Popper bent his ferocious reason to scrutinize the data before him, he realized: both sides of his mind were right. Einstein’s theory wasn’t pseudoscience, and it also wasn’t a logical proof. Against what the world believed, relativity had not been verified by the 1919 solar eclipse. This was because verification exceeded the power of empirical experiments: experiments couldn’t confirm a scientific theory, any more than an anecdote (or indeed, a billion anecdotes) could confirm an absolute truth. Yet even so, Popper saw that Einstein had gone beyond Marx and Freud. His new theory of light had operated as a counternarrative that had successfully anticipated the major plot twist of the Eddington-Dyson expedition. And now that the twist had transpired, Einstein’s theory could be used to predict further plot twists that could themselves be tested via innovative experiments, creatively expanding the practice of science. With this, Popper proved what Herschel had intuited: science wasn’t built of logic. It was built of speculative narratives from outlying events. (90-91)
Science itself had now been rescued, by storythinking, from the ancient deductive and inductive logic of Aristotle. There is no hope for absolute logical verification. Data comes in two forms, first some “outliers” that challenge the reigning narrative, and later some carefully wrought “test data” that confirms or refutes (falsifies) a new narrative hypothesis about how nature works. Innovative experiments, like the one Eccles performs on his cat to decide between spark-like versus soup-like synaptic gaps, “creatively expands the practice of science.”
Plot twists are everything, and ongoing.
Notably, Eccles, who “had been totally, utterly mistaken” in adhering to the Spark theory never became a Soup. “Instead, hewing strictly to Popper’s method, he’d just stopped being a spark.” (92)
Dan: Fletcher admits that it's difficult to compare computer components to storythinking. He says, "'a symbolic language' is so different from 'self-powered wires'", so much so that "to compare them risks distorting more than it clarifies." But he insists on the comparison, claiming "our storythinking brain can rewire its hardware, while the computer can only rewrite its software." I think this comparison betrays a lack of understanding how computers actually work and how the boundaries between hardware and software have been porous and changing for the last several decades (FPGAs, Flash memory, all the way back to PROMs). Fletcher rightly says some AI enthusiasts have "fudged" (or, I'd agree, at least have misunderstood) the "mechanical distinction between computers and brains" and that "the distinction is important to enforce." I was struck by this language and wondered whether he really thinks it's a doctrine to be enforced rather than an observation to be elucidated?
Fletcher concludes his chapter by joining the Spark-Soup problem of neural structure to the "butterfly-brain of Herschel and the equally “dual intellectual nature” of Popper, who loved both logic and the material world and its living things and was “forever shattering one of his loves with the other.” (82)
Evolutionarily speaking, “the more-than-electric brain machinery glimpsed by Eccles emerged about a billion years ago in the tentacle-mouthed jellyfish known as archaic cnidarians.” (94) Peptide cells used by cnidarians to sting and paralyze their prey specialized and became internalized to found the animal nervous system. Or at least, “that’s our best unfalsified guess.” The 85 billion neurons of the human brain are in their deepest evolutionary origin, “part of a hugely involuted jellyfish that is incessantly stinging itself,” (!) says Fletcher. (94) Tellingly, these neurons and their synapses are able to link up to each other to create chains of activation. Unlike machines, brains can rewire themselves.
Okay, but why? Why so many neurons, why the peptides? Why has this neuronal machinery “become such a gigantic part of our headspace?” (94)
Dan: Fletcher's penultimate section, on "The Brain Machinery of Scientific Speculation" is one of the weakest in the chapter, I think. The story of cnidarians and peptides is evocative and interesting, but really doesn't prove what Fletcher hopes, I think. In a way, this is a demonstration of what he implied in the earlier sections, that deduction and even induction are less persuasive than narrative. I don't think the description of the cnidarian means what Fletcher wants us to think it means; in fact, I think it's a bit of a distraction. I don't know my Aristotle well enough to remember whether he described this, but the whole exercise seems like an elaborate analogy where the novelty and interest of the story distracts us from its explanatory inadequacy.
The logical answer, modeling the human brain on a computer, would say it’s storage and processing power. But that’s wrong.
If that’s what our synapses are providing [i.e. storage and processing], they’re extraordinarily bad at their job. Our brain’s memory is weak—and dodgy. It can record far less raw information than a comparatively sized hard drive and recall it with much less accuracy. Nor is our brain that good at crunching data. (94)
Scientific method to the rescue. The data doesn’t fit a machine-based hypothesis.
The alternative hypothesis from storythinking is that our synapses are contributing to the brain’s core behavioral function: plotting original actions… our synapses enlarge our neurons’ trial-and-error process into an experimental method for inventing and adapting plans. (95)
We humans thus have a “mixed architecture.”
This mixed architecture gives our neurons the benefit of both electronic and nonelectronic transmission. The benefit of the former is speed, empowering our thoughts to zip with light’s rapidity. The benefit of the latter is experimental action; by enabling our neurons to freely alter their connections to each other, our synapses allow us to improvise our mental architecture in ways that purely electronic machines cannot. (96)
It’s no wonder storythinking works to explain… storythinking. We’re human, all the way down. Storythinking, plus some logic, is how we humans do the world.
Dan: Fletcher's conclusion, that our brains operate via a couple of different mechanisms and that one favors logic while the other enables narrative thinking, is interesting but remains just a speculation. Where's the FMRI scan to indicate that something specific is happening in the brain when people think logically as opposed to imagining narratives? What other mental aptitudes do these modes of thinking correlate with, and can that tell us anything? I've reached the end of another chapter and I'm still prepared to agree that storythinking is valuable and important. But still not convinced about this dichotomy Fletcher is trying to draw or the neurology or evolutionary biology he's calling on to make his case. And finally, I wasn't impressed with his attempt to get me to experience storythinking by deliberately making his narrative choppy and difficult to follow.
I had some trouble with this chapter. First, as he has done before, Fletcher begins it with a few brief paragraphs, teasing the story of John Eccles which he tells later in more detail. I just don't like this technique, or maybe it's the way he uses it. Possibly because the story is a bit lacking in context, Fletcher seems to feel freer to extrapolate in a way that I'm not sure is an appropriate extension of Eccles' thinking. For example, because synapses "are nonelectric, they can operate free from the constraints of design, improvising narrative scripts beyond the laws of logic." What does this actually mean? Because neurons use electrical signaling, they're identical to transistors or semiconductors? I'm not sure this follows, either from Eccles' theories in 1951 or even in the present from the arguments Fletcher is making.
We’re halfway through Fletcher’s book, and he’s now wrapped his main idea up into the neatest of self-referential packages. Clearly he’s made some acceptable points and told some poignant stories along the way. It’s hard not to acknowledge that there are huge implications for education, for science, and for human adaptation and survival generally in whatever “fast-changing, contested domains where the only intel is anomalous” we happen to find ourselves in. (98)
In the remaining chapters, Fletcher will turn to some problems of how to improve and apply storythinking in a handful of those contested domains.
On to chapter 7.