Skip to content

2026 08 17 Robert Epstein's arguments in his essay 'The Empty Brain' regarding the Information Processing metaphor Gemini Chat

created: 2026-08-17 source: https://share.gemini.google/ocKzSS5hv3SE tags: - artificial_intelligence

In his provocative 2016 Aeon essay, "The Empty Brain," psychologist and researcher Robert Epstein argues that the dominant paradigm in cognitive science—the Information Processing (IP) metaphor—is fundamentally flawed.

Epstein’s core claim is simple: the human brain is not a computer. It does not store memories, process data, encode representations, run algorithms, or retrieve information. Treating the brain as a computational device is not just an incomplete model; it is a category error that actively obstructs our understanding of human behavior.

1. The Core Argument: What the Brain Doesn't Do

Epstein begins by dismantling the foundational vocabulary of modern cognitive psychology:

"We don’t store words or the rules that tell us how to manipulate them. We don’t create representations of visual stimuli, store them in a short-term memory buffer, and then transfer the representation into a long-term memory device. We don’t retrieve information or images or words from memory registers. Computers do all of these things, but organisms do not."

According to Epstein, computers operate on physical symbols: they literally read, write, copy, store, and execute instructions using binary data stored in physical memory locations. Biological organisms, however, possess only tissue, blood flow, neurotransmitters, and complex neural connectivity. Nowhere in the brain has neuroscience ever located a byte of data, a retrieval index, or a discrete algorithm.

2. Key Demonstrations and Thought Experiments

To illustrate why the IP metaphor fails, Epstein uses two primary examples:

A. The "Dollar Bill" Drawing Demonstration

Epstein describes an exercise he frequently performed with his students:

  1. Drawing from Memory: A student is asked to draw a detailed picture of a one-dollar bill on a piece of paper from memory. The resulting drawing is inevitably crude, missing key engravings, text, serial numbers, and precise layout details.

  2. Drawing with a Specimen Present: The student is then handed an actual dollar bill and asked to draw it while looking directly at it. This second drawing is remarkably accurate and intricate.

                    Drawing from Memory           Drawing from Sight
                    ┌─────────────────┐           ┌─────────────────┐
                    │  [ 1 ]     [ 1 ]│           │ ┌─────────────┐ │
                    │       ($)       │           │ │(O)       (O)│ │
                    │  [ 1 ]     [ 1 ]│           │ └─────────────┘ │
                    └─────────────────┘           └─────────────────┘
                        Crude outline                Rich detail

  • The IP Explanation: The brain retrieved a stored "image file" or representation of a dollar bill from long-term memory, but the file was corrupted or lossy.

  • Epstein’s Explanation: There was no stored image at all. If the brain truly stored a representation, drawing it would simply be copying an internal file. Instead, past exposure to money simply changed the brain's physical structure, allowing the person to re-experience or re-enact aspects of seeing a dollar bill—not read an internal snapshot.

B. Catching a Fly Ball (Computation vs. Dynamic Action)

Epstein contrasts how an AI and a human baseball outfielder catch a fly ball:

  • The Computational View: The fielder’s brain calculates the ball’s initial velocity, launch angle, trajectory, air resistance, and wind speed, continuously solving differential equations to compute the coordinates where the ball will land.

  • The Ecological/Behavioral Reality: As demonstrated by researchers like Michael McBeath, the fielder simply runs in such a way as to keep the ball’s optical trajectory rising at a constant, linear rate relative to their field of vision (the Optical Acceleration Cancellation heuristic). The player maintains a continuous sensorimotor feedback loop with the environment—no computational trajectory modeling required.

3. "Changed, Not Filled"

Instead of treating the brain as a storage container, Epstein offers an alternative framing:

  • Organisms are changed by experience: When you listen to a song, practice an instrument, or read a book, your brain does not save a copy of the stimuli. Rather, your neural architecture undergoes structural and chemical alterations.

  • Behavior is re-enactment, not retrieval: When you sing a song you know by heart, you are not opening an audio file in a memory sector; you are executing an organized motor and perceptual sequence that your body has adapted over time to perform.

  • Radical Individual Uniqueness: In computers, a data file (photo.jpg) is identical whether stored on a Mac, a PC, or a flash drive. But because human brains change uniquely based on each person's specific life history and physical development, no two brains represent or respond to the same experience in the same physical way.

4. The Metaphor Trap

Epstein points out that the IP model is simply the latest entry in a long line of historical technological projections:

Era Dominant Technology Resulting Metaphor of Mind
Antiquity / 1600s Hydraulics & Pumps The body/brain as fluid pressure and "animal spirits" flowing through pipes.
1700s – 1800s Clockwork & Springs The mind as an intricate automaton operating with gears and levers.
Late 1800s – 1900s Telegraphs & Switchboards The brain as a telephone exchange routing electrical messages.
Mid-1900s – Present Digital Computers The brain as hardware running cognitive software and storing data.

Epstein argues that just as we look back on Descartes' hydraulic theory as a quaint historical artifact, future scientists will look back on the idea of human memory as "data storage" as an equally misguided conceptual trap.

5. Controversies and Critiques of Epstein's Essay

While "The Empty Brain" resonated with proponents of embodied cognition, behaviorism, and ecological psychology, it faced significant pushback from cognitive scientists and neurobiologists:

  • Overly Literal Definition of Computation: Critics argued that Epstein attacked a cartoonish, rigid 1970s "von Neumann" model of computing (discrete files, hard drives, linear CPUs) rather than modern computational neuroscience, which models distributed neural networks, dynamical attractors, and probabilistic inference.

  • Conflating Physical Implementations: Mainstream cognitive scientists pointed out that claiming the brain "processes information" does not require it to store literal JPEG files; information theory (Shannon information) and predictive coding describe how physical signals reduce uncertainty, which is biologically observable in neural firing rates.

Despite these pushbacks, Epstein’s essay remains one of the clearest, most accessible modern broadsides against computational functionalism, challenging researchers to discard the computer metaphor and study the nervous system on its own biological terms.