Introduction
In the early 2010s, physicists David Deutsch and Chiara Marletto developed a new foundational framework for describing reality called **Constructor Theory (CT for short)**. In classical physics, reality is described by specifying an initial state and then applying dynamical laws to ask “what happens next”. Whether the resulting predictions are deterministic or probabilistic, the framework—as Deutsch and Marletto point out—is still based on trajectories, dynamical laws, and initial conditions.
Constructor Theory proposes a different mode of explanation: instead of predicting outcomes from initial conditions, it characterizes which physical transformations are possible or impossible, and why. But CT is not so much an alternative model to descriptions in classical physics but rather a much more foundational framework, in which questions that can be answered, or at least addressed, in classical physics are just a subset compared to questions addressable in CT.
Transformations, Tasks, Constructors, Counterfactuals
Instead of asking “What happens netxt?” CT asks, given a current state, “What is possible, what is impossible, and why?”—that is, which transformations of a system’s configuration can or cannot be brought about. In this framework, such a transformation is called a task.
A constructor is any physical system that can perform a task and retain the ability to perform it again (i.e., operate in a cycle), such as a catalyst in chemistry or a heat engine in thermodynamics. A task is possible if the laws of physics impose no limit, short of perfection, on how accurately it can be performed by some constructor, and on how well such constructors can retain that ability. A task is impossible if a law of physics forbids any constructor from performing it with arbitrarily high accuracy.
In this way, Constructor Theory has greater modelling capacity than classical theory because it treats counterfactual properties—what can and cannot happen—as fundamental, and it gives information an exact, basic, non-derivative status in physics. Constructor Theory also provides an objective, physical definition of knowledge: knowledge is information that can act as a constructor, i.e., information that can cause transformations and retain the capacity to do so again. In this framework, the basic structure of physical law is expressed in terms of which tasks are possible or impossible, with knowledge and its embodiments appearing as special kinds of information that play a causal, constructor-like role.
Probability and Prediction in Constructor Theory
Constructor Theory does not take probabilities or likelihoods as fundamental. A task being possible means that, in principle, a constructor could perform it with arbitrarily high accuracy and reliability; this does not imply that the task will occur with any particular probability. Instead, CT seeks to explain the appearance of probabilistic behavior—for example in quantum measurements—from deeper, non‑probabilistic laws about which tasks are possible or impossible and what kinds of information‑processing systems can exist.
Accordingly, CT does not abolish the usual “given initial conditions, what happens next?” predictions of physics; it reinterprets them as derived from deeper principles. The fundamental laws in CT are statements about which tasks are possible or impossible, not direct equations of motion. From these counterfactual constraints, the familiar dynamical laws and their predictions can often be recovered, but CT’s primary role is to constrain which laws and evolutions are admissible at all.
Classical Physics as a Subset of the CT Framework
In Constructor Theory, classical physics is not discarded but embedded: the familiar picture of initial states and dynamical laws appears as a special, derived case within the broader “space” of possible and impossible tasks that CT takes as fundamental. The CT framework defines a space of all admissible physical behaviors via possible/impossible tasks; classical physics corresponds to a particular region of that space where the allowed tasks and constraints happen to be representable by deterministic (or stochastic) differential equations with initial conditions.
In that sense, classical physics is not wrong; it is embedded in a wider explanatory structure where its laws are seen as special cases consistent with deeper possible/impossible‑task principles.
Historical Precursors: Scholasticism and Leibniz
Describing reality in terms of possibilities is, however, by no means a completely new approach. We can trace it back to a much older philosophical family of questions concerning possibility, necessity, contingency, and impossibility. Scholastic philosophy, mainly with Duns Scotus and, later, Leibniz’s theory of possible worlds asked which propositions or beings could be true or exist without contradiction; in the modern possible-worlds formulation, something is possible if it is true in at least one possible world, necessary if true in all, and impossible if true in none.
[Sep 04, 2026] – based on a discussion with perplexity.ai – to be continued