Of Two Minds: The Problem with A.I.
We are caught between two narratives. Those engaged in promoting the technology keep telling us that major breakthroughs are coming soon. Some have gone so far as to publish detailed “scenarios” pointing to Artificial General Intelligence (A.G.I.) appearing as soon as 2027. Better than humans at everything. Meanwhile, veteran investors, such as Roger MacNamee of Elevation Partners, publicly asks “How much can you sell of a product that doesn’t do as promised?” In this note we hope to illustrate why there is a fundamental flaw in today’s approach to A.I., and why that approach – as reflected in today’s Large Language Models (LLMs) – will never be able to deliver on these inflated promises.
Both/And
The human mind has been examined, speculated about, poked, prodded, studied, and manipulated for thousands of years. The Greeks referred to it as the Psyche. The Latins called it the Anima. Etymologically, these terms refer to something closer to what is called “Soul” in English – which, in our secular age, would likely confuse many people. Aristotle began a rigorous examination in his 4th-century BC Peri Psyche. Psychology – the study of the Psyche – has been through many dark alleyways since then. What has been the best of all those millennia of effort?
We have “two minds” not one. But A.I. has only one of them. Even if A.I. could “emulate” that one, it would still be missing the other. In humans, these two minds must work together. Integrated. Half-a-mind is no mind-at-all. That’s the problem with A.I.
Many of our best thinkers have noted this basic reality. No, not all of them, but enough that we should consider this matter to be “settled.” We are “both/and” creatures. Of two minds. Angus Fletcher (1976--) -- a widely read author, Story Science professor , and neuroscientist at The Ohio State University -- has discussed this at length. Iain McGilchrist (1953--) -- author, psychiatrist, neuroscientist and Quondam Fellow at All Souls College, Oxford -- has also noted how this works. Marshall McLuhan (1911-1980) – still a popular author, who was a Media Ecologist and an English Professor at the University of Toronto – wrote his PhD dissertation, The Classical Trivium (1943), on related aspects of this phenomenon. As have others, like Mary Harrington with her Aristotelean distinction between intellectus and ratio, as regards work and leisure. So, this should hardly be a mystery.
Fletcher calls these two brains: Logic and Narrative. He bases his neuroscience on the fact that our neurons use two different mechanisms: electricity and neurotransmitters. Or “connections” (neurons) and “plugs” (synapses). Or “sparks” and “soup.”
Fletcher also calls it “Storythinking” (in his book with that title) and is expected to update his views in the forthcoming Primal Intelligence: You Are Smarter Than You Know (due August, 2025). This name was the result of Ohio State’s Project Narrative, who have been collaborators with NASA, Hollywood, and Silicon Valley, and their 2021 announcement about their research findings. This led U.S. Army Special Operations to develop Primal training for its most classified units – now considered a success. The Ohio State Project asserts that it has “scientific proof” that this is impossible for computers to accomplish. And they have the research experience, with many groups, to back up their claims.
He has written Primal Intelligence as a “training manual” for a “different way of using your brain.” It purports to center around a neuroscientific approach to Intuition, Imagination, Emotion, and Commonsense. All capabilities woefully lacking in A.I. All capabilities that are uniquely human. This is a gap that cannot be bridged today. As he notes, we have neither the science nor the engineering to accomplish this “next step” (p.98, Storytelling).
McGilchrist describes these two brains in neuroanatomical terms: Left and Right – roughly corresponding to Fletcher’s Logic and Narrative, respectively. Complementary but distinct approaches to the world. In his Master and His Emissary: The Divided Brain and the Making of the Western World (2009), he lays out his basic conclusions. His view is that we now live an “unbalanced” life because we have let the Left brain take precedence – whereas it should be the “Emissary” of the Right brain. As a result, we are often mistaken in how we “attend” to the world, since each hemisphere has a different way of “paying attention.” He continued this exploration is his The Matter with Things: Our Brains, Our Delusions, and the Unmasking of the World (2021, three-volumes-in-one), in which he becomes more polemical, castigating “scientific materialism.” He turns his attention to truth/judgement, intuition/imagination, and matter/consciousness. And in his interviews, he describes his own views as “panpsychic.”
McLuhan, on the other hand, described these two minds in terms of the components of the medieval Trivium – which was at the heart of Western education, explicitly from Charlemagne in the 8th-century until the 19th (sometimes summarized as the Three R’s (or “reading, writing and ‘rithmetic). Grammar corresponds to what McGilchrist called “Master” (Right) and Fletcher calls “Narrative.” Dialectics corresponds to what McGilchrist calls “Emissary” (Left) and what Fletcher calls Logic. The Trivium, of course, has a third element: Rhetoric, which could be considered the “battlefield” where Grammar and Dialectics fight it out. Whereas, both Fletcher and McGilchrist mostly base their arguments on “modern” observations, McLuhan takes a much longer view, detailing in his 1943 Cambridge dissertation how this conflict has developed over time – with particular attention paid to the Formal Causality involved in technological developments.
Artificial Intelligence
Where did today’s idea of Artificial Intelligence come from? The standard history begins at the Dartmouth Summer Workshop in 1956, while noting, in passing, various earlier notions. Dartmouth was not notional. It was meant to generate action. Those involved called themselves “Mathematical Psychologists.” With the emphasis on “Math” and with very little regard for “Psych.” They were aiming for a sweeping revision of social interactions. A new society. A really “Brave New World.”
These folks were hardcore Logic/Emissary/Dialectical. Not a touch of Narrative/Master/Grammarian in sight. This was a period in which social science was devoting a great deal of attention to “controlling” human behavior. This is how we got Noam Chomsky’s “Universal Grammar” at MIT (aka a “programming language” for humanity). For a while, Norbert Wiener’s Cybernetics: or Control and Communication in the Animal and the Machine (1948) was front-page news. With the emphasis on “control.” But Wiener wasn’t invited to Dartmouth. Home-schooled by his philologist father, Leo, partly on a diet of Russian literature, Wiener was too Narrative/Master/Grammar for this group. He would have spoiled their party.
As McLuhan emphasized, true Grammar was something not reducible to formulas or algorithms. Not “sentence diagramming.” That was for Dialectics. The other mind. And McLuhan wasn’t satisfied with limited application of the term. His parallel contribution to the 1956 Dartmouth/MIT modeling of humans as “computers” was the essay “The New Languages,” reporting on the early results of his 1953 Ford Foundation grant, “Changing Patterns of Language and Behavior and New Media of Communications,” which begins with –
“English is a mass medium. All languages are mass media. The new mass media -- film, radio, television -- are new languages, their grammars as yet unknown . . .”
McLuhan would go on to strongly champion the role of Analogy, while criticizing the application of Logic. His one-sided “grammarian” views could even have resulted in him over-looking important developments when the two approaches worked together. Two brains. Both are necessary. McLuhan plumed the depths of Analogy, often pointing to the “Analogy of Proper Proportionality,” expressed as A:B::C:D (or, A is to B, as C is to D). McLuhan had converted to Catholicism in his 20s while at Cambridge, and his sense of the importance of Medieval “theories” was powerful, ultimately resolving back to the metaphysics of Thomas Aquinas on this and other topics. Including the “proportionality” of Natural and Supernatural knowledge.
Fletcher’s emphasis on neurological “soups and sparks” is also quite important here – particularly as regards A.I. Or as he puts it –
“This storythinking process is lent enormous plasticity by the sheer number of synapses in the human neocortex. They’re estimated to be in the trillions, enabling our brain to fork and flex action scripts in almost endless new directions, innovating art, technology, business, and politics.
“Such innovation lies beyond the power of computer AI. AI can only reiterate symbolic logic’s three preprogrammed actions: AND, OR, NOT. Which is why AI is limited to deducing— based on the data flowing into it— what sport, science, or dance is. It cannot freestyle fresh actions that expand what sport, science, and dance can do.
“To change this situation, we’d need to create a computer capable of rearranging its core architecture. Which is to say, we’d need to exchange the etched transistors of modern CPUs with free- floating hardware that could riff itself into original circuitry. Beyond the obvious engineering challenges that this poses, it’s rendered impossible from the getgo by the fact that electronic systems are governed by mathematical equations of voltage and current. Step outside those equations, and the systems melt or shut down. This means that computer brains cannot be blindly improvised; they must be designed in advance to stick to the math. Such sticking is what their current logic gates are engineered, precisely to do. But if those gates were freed to shuffle about, they’d swiftly unglue, frying themselves or blinking out.” (pp. 96-7)
The human nervous systems (yes, there are many) evolved biologically. Flesh not semiconductors. Actions not logic-states. Despite claims by some that logic can “simulate” anything, that would require a complete characterization of all the “anythings” in logic-state terms. Alas, not possible. How Logic/Emissary/Dialectical of them. No, Narrative/Master/Grammar simply doesn’t work that way. We are of two minds.
Man and the Machines
The issue of humanity dealing with the effects of its own inventions has a long history. Today it is generally recognized that technology is not “neutral.” Or, as McLuhan’s colleague, John Culkin, remarked, “We shape our tools and thereafter they shape us . . .” The neurological “plasticity” described by Fletcher, is the main mechanism involved. It is “human nature” to have a very flexible “nature” – particularly as regards our engagement with our technological environments. In many ways, “We are what we eat . . .”
One of the more important “modern” commentors on all this was Romano Guardini (1885-1968). Guardini, a German Catholic theologian and pastoralist, wrote extensively on these topics. His post-WW II lectures, published as The End of the Modern World and Power and Responsibility, are essential for understanding the impact of A.I. on individuals and society. His final word, the 1959 lecture Man and the Machines, has been translated into multiple languages for publication by CSDL, the parent organization to this Substack.
He concludes that essay (addressed to what must have been one of the earliest “Computer Science” audiences at a Munich technical school) by saying --
“We might like to fantasize a little—after all, utopias have so often become real that dreaming would be legitimate. We could imagine a spiritual council of nations, in which the best from all political areas would consider these questions together. Human existence is so far advanced, the Human has taken himself so much into his own hands, the possibilities of both achievement and destruction have become so unforeseeable, that it is time for a new virtue: a spiritual statecraft, in which the Human, having become serious through so much experience, would step out from his bashfulness into the distinctive areas of thinking and living. That would be the best case. A living human consciousness would enable us to look out over the whole of our existence and, having reached a real, sovereign objectivity, consider the res hominis.”
Such a “spiritual council of nations” has not yet been formed. Instead, many organizations have spent their time pondering, debating, and releasing statements about “A.I. Ethics” – presumably out of concerns that the “robots” aren’t “aligned” with human goals. Terminator style. Autonomous weapons style. Modern “secular” society style. Issac Asimov’s “Three Laws of Robotics” style. But despite all the effort expended, one suspects this is nonetheless a futile PR waste of time. Since the robots have no understanding of “purpose” or “meaning,” how are we to expect them to follow a declaration of “ethics”?
One potential for some clarity and even positive action could be the just elected Catholic Pope, Leo XIV. Rather than attempting to tell the robots what to-do/not-do – noting that Asimov’s Laws were intended to sell pulp magazines, not inform robot designers – it seems far more productive to focus on the society-to-come, all occurring under Digital conditions. Leo XIV explicitly picked that name to indicate that he wants to continue the work of the 19th century Leo XIII (1810-1903). Leo XIII launched what is known as “Catholic Social Teaching” (CST) with his 1891 encyclical Rerum Novarum. Noted by Leo XIV, and then the current Leo mentioned A.I. – pointing to the need for CST to be upgraded to Digital Catholic Social Teaching (DCST, about which CSDL offered an 8-lecture series in 2022).
Digital Catholic Social Teaching uses the same three pillars as CST: Human Dignity, Subsidiarity, and Solidarity. But it recognizes that the effects of the new technology are different from those propelled by the earlier Electric Paradigm (c. 1850-2000, in the West). Instead of herding us into “mass audiences,” DCST fragments the world, as shown by the collapse of Globalism and the massive expansion of information sources. Instead of fragmenting the human personality, DCST reintegrates our Body and Soul – reversing the previous split, dating back to Descartes &al. Instead of promoting isolated (negative) “freedom,” DCST promotes a positive view of freedom, again orienting humans towards the Good.
Like all environmental technologies, A.I. changes us. We are already different from who we were before ChatGPT was released. Not just in the instrumental fashion of our use of the technology, we have changed at a more fundamental level. Indeed, a spiritual level. Serious attention to spirituality is clearly increasing worldwide. We believe that Digital has caused this shift. The new Paradigm is Christian, Daoist, Hindu and more – already. All together. Digital technology seeks to build “Artificial Humans.” That can only stimulate the actual humans to ask, “What Does It Mean to Be Human?” – a refrain that echoes around the world. We are now “retrieving” what was once lost. Humans are becoming “integrated” body-and-soul once again. What Max Weber bemoaned as the “dis-enchantment of the world” in his 1918 lecture, “Science as a Vocation,” is now undergoing a “re-enchantment.” Without the benefits provided by DCST, and its global counterparts, God help us . . . !!