The long-standing scientific effort to unify the disparate fields of thermodynamics and information theory has reached a significant milestone with the formal recognition of a theorem that provides a universal link between physical entropy and the erasure of information. For decades, the association between these two concepts relied heavily on "plausibility arguments"—thought experiments involving imaginary beings and hypothetical engines, such as the famous Maxwell’s Demon. However, a new research paper by Penha Cardozo Dias, recently updated in late 2026, posits that a theorem by Charles Bennett on reversible computation serves as the definitive mathematical foundation for this relationship. By constructing a real, human-operated, non-thermal engine, the study demonstrates that information is not merely an abstract mathematical construct but a physical entity governed by the laws of thermodynamics.
The Theoretical Gap in Information Physics
Since the mid-19th century, physicists have grappled with the nature of entropy. Originally defined by Rudolf Clausius as a measure of a system’s thermal energy per unit temperature that is unavailable for doing useful work, the concept was later expanded by Ludwig Boltzmann and J. Willard Gibbs to describe the statistical distribution of microscopic states. In 1948, Claude Shannon introduced "information entropy," a mathematical measure of uncertainty or the "surprise" value of a message.
While the mathematical forms of Shannon entropy and thermodynamic entropy are strikingly similar, the physical connection between them remained a subject of intense debate. Critics argued that Shannon’s entropy was a metaphor for the lack of knowledge, whereas thermodynamic entropy was a tangible property of matter and energy. The bridge between these worlds was largely built on the Landauer Principle, proposed by Rolf Landauer in 1961, which suggests that the erasure of one bit of information results in the dissipation of a specific amount of heat ($kT ln 2$).
Despite the success of Landauer’s Principle in various experimental setups, the absence of a universal theorem hindered the full integration of information into the physical sciences. The work of Cardozo Dias addresses this by identifying Charles Bennett’s 1973 theorem on reversible computation as the "missing link." Bennett proved that a computer could, in principle, perform any calculation without consuming energy, provided it does not erase any information. The energy cost of computation, therefore, is strictly tied to the act of "forgetting" or resetting data.
A Real-World Non-Thermal Engine
To validate these theoretical claims, the research introduces a physical prototype: a non-thermal engine operated by humans. Unlike traditional heat engines that rely on temperature gradients to produce work, this engine operates on the manipulation and erasure of information.
The engine’s operation is structured in stages that mirror the three-tape model of reversible computation developed by Bennett. In this model, a computer uses an input tape, a work tape, and an output tape. To make the process reversible, the machine must keep a record of its intermediate steps. If it attempts to clear the work tape to prepare for a new task, it must expel entropy into its environment.
The research proves that this human-operated engine obeys two laws that are functionally identical to the First and Second Laws of Thermodynamics.
- The First Law (Conservation): Energy and information within the system are conserved through the operational cycles, accounting for the work performed by the human operator.
- The Second Law (Entropy Increase): Any process that involves the erasure of information results in a net increase in the entropy of the system-environment complex.
By demonstrating these laws in a non-thermal context, the paper establishes that entropy is not exclusive to heat-based systems. Instead, entropy is a measure of "erased or missing information" across all physical platforms.
Historical Chronology of Information-Entropy Research
The evolution of this field highlights a slow but steady transition from philosophical inquiry to hard physical proof:
- 1867: James Clerk Maxwell proposes "Maxwell’s Demon," a thought experiment where a tiny being sorts fast and slow molecules to decrease entropy, seemingly violating the Second Law of Thermodynamics.
- 1929: Leo Szilard publishes a paper suggesting that the Demon must "pay" for its sorting by consuming information, linking the acquisition of knowledge to physical entropy.
- 1948: Claude Shannon formalizes information theory, defining entropy as the average rate at which information is produced by a stochastic source of data.
- 1961: Rolf Landauer postulates that information is physical and that erasing one bit of information requires a minimum amount of energy dissipation.
- 1973/1982: Charles Bennett develops the theory of reversible computation, showing that only irreversible acts (like erasure) have a thermodynamic cost.
- 2023: Penha Cardozo Dias submits the initial version of "Information and entropy" to the arXiv preprint server, proposing a non-thermal engine to formalize the connection.
- 2026: The revised version of the paper (v2) is released, providing a refined proof and video evidence of a working prototype that operates on these principles.
Measuring Information in Physical Units
One of the most significant implications of the Cardozo Dias paper is the measurement of information in physical units. While computer scientists typically measure information in "bits" or "bytes," the paper aligns with Landauer’s Principle by expressing information in units consistent with the Boltzmann constant ($k_B$).
This shift has profound consequences for how we view the "cost" of the digital age. If information is a physical quantity, then every bit of data stored in global server farms has a literal "mass" and "energy" footprint that goes beyond the electricity used to power the hardware. It suggests that the "missing information" in a thermodynamic system is not just a lack of human knowledge but a physical vacancy that must be accounted for in the system’s energy balance.
Experimental Evidence: The Prototype at Work
Accompanying the research is a video demonstration of a prototype engine. This device serves as a "proof of concept" for the theoretical framework. In the video, the engine is shown undergoing a cycle where a human operator performs a task that involves the categorization and subsequent "erasure" of data states.
Sensors attached to the system monitor the energy input and the resulting changes in the system’s state. The data confirms that the "entropy" produced during the erasure phase matches the predictions of Bennett’s theorem and the modified laws of thermodynamics. This transition from "imaginary engines" to a "real engine" is what sets this research apart from previous attempts to unify the two fields. It provides a tangible laboratory setting where the abstract concepts of information theory can be measured with the same precision as temperature or pressure in a steam engine.
Broader Implications and Scientific Reaction
The scientific community has begun to weigh the implications of this formal association. If the Second Law of Thermodynamics is essentially a law about information loss, it could change our understanding of several fields:
1. Quantum Computing
Quantum computers rely on maintaining "coherence," a state where information is not lost to the environment. The formalization of entropy as information erasure provides a clearer roadmap for the thermodynamic limits of quantum error correction.
2. The Limits of Big Data
As the world produces more data, the energy required to "manage" (which includes deleting or overwriting) that data becomes a critical bottleneck. Understanding information as a physical entity could lead to more energy-efficient storage technologies that operate closer to the Landauer limit.
3. Cosmology and the Nature of Reality
Some physicists, such as the late John Wheeler, suggested a philosophy of "It from Bit"—the idea that the physical universe arises from information. By proving that thermal entropy is a measure of missing information via a universal theorem, this research lends weight to the idea that the universe’s fundamental "currency" is information.
4. Artificial Intelligence
As AI models require massive computational resources, the thermodynamic cost of training and resetting neural weights becomes a major concern. This research suggests that the "forgetting" processes in AI architectures are just as physically significant as the "learning" processes.
Conclusion
The recognition of Bennett’s theorem as the universal bridge between information and entropy marks a turning point in physics. By moving away from the "imaginary beings" of the 19th century and toward a human-operated, non-thermal engine, Penha Cardozo Dias has provided a framework where information is treated with the same physical rigor as matter and energy.
As the prototype continues to be studied, the focus will likely shift toward practical applications in nanotechnology and ultra-low-power computing. For now, the conclusion is clear: information is not a ghost in the machine; it is a physical component of the machine itself. The "missing information" that once served as a conceptual placeholder for entropy has finally been given a formal, measurable, and provable identity.
