The Radicalism of William James

Essays on Consciousness

William James in Brazil, 1865, Harvard University, Houghton Library.

The mind, in short, works on the data it receives very much as a sculptor works on his block of stone. In a sense the statue stood there from eternity. But there were a thousand different ones beside it, and the sculptor alone is to thank for having extricated this one from the rest. —William James

Although he credited his brilliant and somewhat flakey colleague Charles Peirce with coining the term, William James popularized using pragmatism to mean a constellation of ideas for approaching truth, inquiry, and knowledge. Pragmatism looks to the meaningful human consequences of a question to resolve differences about the answer. In his most succinct definition, James called pragmatism “a method for carrying on abstract discussion” and suggested that it allows us to “escape vain wrangling.” In On Techno-Pragmatism, I applied the pragmatic method to the current discourse about artificial intelligence in order to escape vain wrangling over whether we should be optimistic or pessimistic about its social impact. I aim instead, to find a direction of travel that looks to create a more democratic future. In this and a handful of future essays, I explore the implications of James's ideas about the human mind for how we approach generative AI.

As Alison Gopnik and

Henry Farrell

have suggested, generative AI is best understood as a cultural technology. Foundation models use data and energy, lots of it, to generate cultural artifacts—text, images, code, sounds, and now videos—indistinguishable from those created by humans. We call these cultural capabilities artificial intelligence, but it is not just their ability to apply knowledge to solving problems that provide value; rather, it is the extent to which, in Farrell’s words, “they provide new ways to access, order and remix human generated information.” Their capacity for sophisticated language games and their ability to create meaningful responses to human inputs give the appearance that foundation models are, in some sense, self-aware. More accurately, they respond to questions as if they were self-aware. As this is essentially the same limited evidence we have that other human beings are self-aware, it is unsurprising that questions about whether these tools are conscious have become ubiquitous.

Skeptics rightly point out how limited these new AI capabilities are, especially compared to humans. Artificial Intelligence as a non-human and powerful force is hyped by marketers looking to move product and researchers looking for status and funding. Exaggerations and speculative fear-mongering make evaluating the latest information about AI research and design difficult. Yet, we know that machines built using transformer architecture are drawing upon much larger datasets and have far greater processing power than models of even a few years ago. Based on public interest and the increased use of ChatGPT and other recently released foundation models, it is clear that some threshold has been reached. It appears something has changed. But what?

Improvements in natural language processing are perhaps the most underrated explanation for what seems different. We marvel at the size of the datasets and worry about their quality. We note how much more powerful their processing is compared to earlier models and worry about energy consumption. For most of the actual users of the technology, what matters is the way foundation models are able to understand and respond to human communication. How we relate to machines has changed in important ways. When we text them a question, ask them to draw a picture, or request they improve our code, foundation models understand us, truly understand what we say…or, at least it feels that way. To what extent is that feeling real or meaningful? What does it mean pragmatically to create technologies that engage in culture-making and have meaningful dialogue with humans?

Answering these questions is deeply connected to questions about consciousness. Can machines be conscious? What is consciousness? Is it something fundamental to the universe? William James answers these questions by reconstructing our understanding of reality. He abandons the dualistic framework of Descartes for one that understands reality as relations. James replaced mind/body and self/world with mind ←→ body and self ←→ world. He avoids Kant’s assumptions about preexisting categories by starting, as Hume does, with experiences and then identifying useful concepts to explain them. James insists on using the pragmatic method to consider how the words and concepts we use to explain data shape knowledge about our experience of those data. He named this approach to knowledge and consciousness radical empiricism.

If James sounds like a postmodernist, you may find reading his essays uncomfortable or thrilling depending on how committed you are to an unexamined belief in an external world containing truths verified by science or common sense.

James’s ideas drew directly from the physics and biology of his day and remain compatible with today’s scientific models for how the natural world works. Although he died in 1910 having never read Einstein and before the mathematical weirdness of quantum mechanics was first described, James had moved well beyond the notion of an ordered common-sense reality. He did not expect the universe to be rational; words like relativity, indeterminacy, and observer effects that are used by physicists to explain reality are all present in James’s writing about the human mind. When it comes to studying the human mind in a universe where common sense explanations fail the deeper we go, James provides insights that could lead to new areas of exploration for consciousness research. Pragmatism’s key insight is that truth is created through attempts to understand and change the world. Radical empiricism’s key insight is that there is no knower separate from what is known. Experience relates the two, and experience is all there is to reality.

His essay Does Consciousness Exist? states this idea plainly: 

My thesis is that if we start with the supposition that there is only one primal stuff or material in the world, a stuff of which everything is composed, and if we call that stuff ‘pure experience,’ the knowing can easily be explained as a particular sort of relation towards one another into which portions of pure experience may enter. The relation itself is a part of pure experience; one of its ‘terms’ becomes the subject or bearer of the knowledge, the knower, the other becomes the object known.

If you think James sounds like a postmodernist, you may find reading his essays uncomfortable or thrilling depending on how committed you are to an unexamined belief in an external world containing truths verified by science or common sense. In James’s view, any attempt to model a universe separate from our experience of it will fail, for assuming such a universe exists, it is completely unknowable. We live in a universe of experiences, full of messy and often contradictory data perceived through our senses and understood through functions shaped by evolution. James calls reality “apperception itself,” an idea that runs counter to how most of those living in the North Atlantic in the past three or four centuries have understood the universe. The feeling of perceiving an external object leads us–at least those of us socialized into a sense that there is a reality external to our subjective experience–to focus on the object and to believe that reality is the object itself. Our intuitive experience of that reality divides experience into two parts: a self that experiences internally and an external world to be experienced.

In The World Behind the World,

Erik Hoel

provides a concise and illuminating history of how modern humans have understood this division. Over the past 500 years, we have developed an intrinsic perspective, a language of the mind that narrates subjective experiences, and an extrinsic perspective, the methods of inquiry that developed into what we now call science. Hoel argues that “it is in the modern science of consciousness that the two perspectives come to a head.” By that definition, James is an early, perhaps the earliest, consciousness scientist. As important, he is an influential writer about consciousness, narrating his own subjective experience in essays and talks. James gave us the very language we use to describe our self-awareness. Literary critics, philosophers, and scientists use his metaphor of consciousness as a stream, often without knowing its origins in Principles of Psychology:

Consciousness, then, does not appear to itself chopped up in bits. Such words as 'chain' or 'train' do not describe it fitly as it presents itself in the first instance. It is nothing jointed; it flows. A 'river' or a 'stream' are the metaphors by which it is most naturally described. In talking of it hereafter, let us call it the stream of thought, of consciousness, or of subjective life.

Hoel says that “novels are the ultimate expression of the intrinsic perspective” and suggests that “there is an objective deepening in descriptions of minds over time” in the literary form, which starts with Cervantes and reaches maturity during the period from 1850 to 1950. He mentions Middlemarch as an example of a novel that explores the “depth of interiority,” but he could have just as easily named something by William’s brother Henry James. In my view, the essay form, from Montaigne on, deserves equal billing as a method for exploring subjective experience. Like his godfather Ralph Waldo Emerson, Willam James was a master of the essay and its sibling, the public lecture.

Yet, it is not so simple as this. While literature may, as Hoel says, “dramatize our inner lives” and teach us “how to make the mundane extraordinary” it also frustrates our efforts at clarity. Language is not simply a method of divining experience. Reading or listening to words is an experience itself, a way of connecting with others, to be sure, but not perfectly. As James puts it, “language works against our perception of the truth.” Words may describe what and how we think, but they cannot fully account for experience.

There is not a conjunction or a preposition, and hardly an adverbial phrase, syntactic form, or inflection of voice, in human speech, that does not express some shading or other of relation which we at some moment actually feel to exist between the larger objects of our thought. If we speak objectively, it is the real relations that appear revealed; if we speak subjectively, it is the stream of consciousness that matches each of them by an inward coloring of its own. In either case the relations are numberless, and no existing language is capable of doing justice to all their shades.

James was keenly aware of the inadequacy of language to do justice to the fullness of experience, particularly its limitations in making sense of data about our own minds. The tools we have to examine that data and the technical language we use to describe them have developed considerably in the century since he lived. However important to the history of that development, James belongs to a distant past, before computers, neural networks, and electronic instrumentation created the modern discipline of neuroscience. We have more and better data today, and better tools to shape information about the brain. What could a man who never used a computer or saw an MRI have to say that is relevant to consciousness research today?

The answer starts by revisiting the framework James provides for understanding the functions of the brain. As Hoel argues in the essay Neuroscience is pre-paradigmatic. Consciousness is why, James defines the stream of consciousness as the brain's primary function. Arguments against this view “must be weighted by the fact that, precisely because consciousness always seemed too ickily subjective for science, scholars have had strong non-scientific reasons to downplay its importance.” James not only faces the fact that empirical investigation of consciousness requires inquiry into subjective experiences, he argues that our usual notions of reality as divided are misleading in important and subtle ways. The intuition of our selves as subjects who perceive objects gives us the sense that the mind represents external reality, that thoughts duplicate things in the external world when we perceive those things. In other words, our minds recreate the world subjectively. Nope, says James, “Our sensations are not small inner duplications of things, they are the things themselves in so far as the things are presented to us." This radical explanation of subjectivity, that everything comes down to experiences, that the one primal stuff is perception, that thoughts and things are fundamentally the same, that the relation itself is a part of pure experience, has been mostly ignored. Instead, what James called rationalism and ordinary empiricism have shaped inquiry into consciousness including how we answer questions about machine consciousness.

As the name radical empiricism suggests, James rejected rationalist views and took the side of empiricists. Rationalists look to “trans-experiential agents of unification, substances, intellectual categories and powers, or Selves” for explanations. James rejected these concepts as starting points and instead looked to explanations based on inquiry into sensory experiences including what we think of as subjective experiences. The radicalism of this empirical approach is that where “ordinary empiricists” have always shown “a tendency to do away with the connections of things, and to insist most on the disjunctions”–that is they break things down into parts and study the parts to understand the whole–the radical empiricist looks to the relations instead, especially “conjunctive relations.” For James, the important question was how does the experience of this thing or thought relate to this thing or thought. 

Relations are of different degrees of intimacy. Merely to be with one another in a universe of discourse is the most external relation that terms can have, and seems to involve nothing whatever as to farther consequences. Simultaneity and time-interval come next, and then space-adjacency and distance. After them, similarity and difference, carrying the possibility of many inferences. Then relations of activity, tying terms into series involving change, tendency, resistance, and the causal order generally. 

When considering subjective experience, James looks to what he calls “cognitive relations,” the type of conjunctive relation where “Knowledge of sensible realities…comes to life inside the tissue of experience. It is made; and made by relations that unroll themselves in time.” This unrolling is a flow that has direction. “Whenever certain intermediaries are given, such that, as they develop towards their terminus, there is experience from point to point of one direction followed, and finally of one process fulfilled, the result is that their starting-point thereby becomes a knower and their terminus an object meant or known.” For James, a relation is cognitive if it is “experienced between terms that form states of mind, and are immediately conscious of continuing each other.” Cognitive relations are continuous and ever-changing.

Neuroscientists, philosophers, and AI researchers use terms like world models, self models, and global workspace to map the relational territory James described, and many agree with the functional definitions of cognition he gave in Principles. However, few have grappled with what it means to consider consciousness a continuous stream of experiences, as fully relational. They retain a dualistic framework that focuses on objects of study. Consciousness exists in brains, networks, and neurons. They are ordinary empiricists breaking down the problem into parts in order to understand the whole. Or, they are rationalists starting with categories and fitting the data to those pre-defined concepts. In both cases, they have methods and data, but no agreed-upon definition of what they are looking for.

James’s empiricism led him to search for data in places science had declared off limits. Experiences related to religion, spiritualism, and drugs are too subjective for scientific investigation. They rely on introspective reports and are tainted by irrationality. James wrote about all three, unapologetically and with a focus on the experiences. Religion may involve believing things that are empirically untrue, but the varieties of experiences associated with belief are deeply interesting. Spiritualism may be associated with con artists and hucksters, but if consciousness persists after death, empirical investigation of such phenomena is the only path to understanding. Drugs are morally and socially problematic, but studying altered states expands our knowledge of consciousness.

That these avenues of inquiry have been mostly ignored or maligned by scientists means that James has become a patron saint for seekers of truths outside the academy. This helps explain why he remains well-known today compared to other early researchers and scientists, yet it also means that our memory of James tends toward amusing anecdotes and popular history. He is often credited as the philosopher in the Turtles all the way down story, in which an exchange between a philosopher and a little old lady illustrates infinite regress. In books like Ghost Hunters: William James and the Search for Scientific Proof of Life After Death (2009), The Varieties of Spiritual Experience: 21st Century Research and Perspectives (2022), and Mind-Dust and White Crows: The Psychical Research of William James (2024), James appears as a freethinker who saw beyond the arbitrary boundaries of knowledge established by science at the turn of the twentieth century. True enough, but it was James’s failure to find evidence that consciousness extends beyond death or that paranormal phenomena exist that helped establish those boundaries. Of course, he hoped others would succeed where he failed. His experiments with psychedelics may even be considered an example of such success given the recent resurgence of scientific interest in using them to treat depression. For a good story about James’s self-experimentation with mind-altering drugs, check out this episode of Sean Carrol’s podcast featuring historian

Benjamin Breen

.

In an episode of Star Trek: TNG, Captain Picard gives young Wesley Crusher an unnamed book by William James to read before he heads off to take his Starfleet Academy exams. I believe it is Essays in Radical Empiricism. In the brief scene, Picard is frustrated because Wesley hasn’t read the book. Crusher explains that he doesn’t have much time and says “William James won't be on my Starfleet exams.” The scene concludes with a brief lecture from Picard on the importance of understanding “the past…art, history, philosophy” to explore—and make sense of— strange new worlds. Like Wesley, we know of James but don’t actually read him; like Picard, I urge you to do so.

Over the coming months, I will share essays that read passages from James in light of questions about consciousness in humans and machines. I believe his radical ideas will help us better understand how our relations with the technology we create, and how we might shape those relations.

𝑨𝑰 𝑳𝒐𝒈 © 2024 by Rob Nelson is licensed under CC BY-SA 4.0.