Schrödinger's Test for Artificial Intelligence

dong zhang

Published Mar 2, 2026• 0 citations

Abstract

Following the spirit of the Turing Test, we propose Schrödinger’s Test using the original text of Schrödinger’s Cat thought experiment as a test of logical reasoning and independent thinking in artificial intelligence. For a radioactive isotope, we introduce its Schrödinger’s Amount, the weight or quantity of its atoms equal to “a tiny amount” in the original text, and it is put in the steel box with the cat of the thought experiment. The test consists of a machine reading the text and answering several test questions. A higher level of the test is to have it write a report to see how many answers to the questions it contains. As a starting set, our questions are designed to dissect Schrödinger’s Cat part by part and step by step, a few of which are original, not yet available on the internet, unknown to current artificial intelligence programs, and can therefore serve as test questions. For example, why must the box be open? Are atoms identical? Is Schrödinger’s Cat repeatable? For each question, two answers are presented: a conventional or known answer reflecting standard interpretations in physics available on the internet, and an unconventional answer derived through facts and logic not yet available on the internet. Since Schrödinger’s Cat was introduced in 1935, with Einstein’s involvement, it has been extensively studied, taught, and wrote for nearly 100 years, Schrödinger’s Test represents an extremely high-level of human intelligence.

DOI: 10.5281/zenodo.18831847

Related papers

Schrödinger's Test for AI

dong zhang

This work revisits foundational thought experiments in physics—most notably Schrödinger’s Cat and Young’s double-slit experiment—from the perspective of logic, language, and discrete reasoning, rather than probability or continuous modeling. The book begins with a close reading of Schrödinger’s original text and treats the cat experiment as a logical and linguistic test rather than a physical paradox. By following the description step by step, the analysis shows that the experiment centers on first events, non-repeatability, amplification, and restricted access to information. The uncertainty lies in description and access, not in physical superposition. The discussion then broadens to related cases, including non-repeatable experiments, multifacet descriptions, sealed systems, and the distinction between reality and observation. Familiar examples such as the blind men and the elephant are used to clarify how superposition can arise from language and partial description rather than from objects themselves. A stripped-down “cat in a box” case shows that the logical structure of Schrödinger’s Cat exists independently of radioactive devices. The book also revisits Young’s double-slit experiment, focusing on the role of the slit as a physical boundary and interaction region rather than an abstract geometric opening. Light and electrons are treated as discrete entities whose cumulative outcomes arise from individual events and in-slit interactions with solid surfaces. The interference pattern is interpreted as a historical summary of discrete trajectories and interactions, not as evidence of continuous wave behavior passing through empty space. Taken together, these analyses lead to a unified view: many celebrated “paradoxes” arise from linguistic compression, continuous idealization, and the loss of discrete structure in description. In the final chapters, Schrödinger’s Cat is reframed as a reasoning benchmark, comparable in spirit to the Turing Test. Whereas the Turing Test evaluates imitation through conversation, Schrödinger’s Test evaluates careful reading, logical inference, and consistency when language, time, and access to information are intertwined. Artificial intelligence plays a natural role in this framework. AI systems operate through finite steps, substitution, and rule-based manipulation of language. These properties align closely with the logical structure required to analyze both Schrödinger’s Cat and the double-slit experiment. The book therefore proposes Schrödinger’s Test as a reading-comprehension and logical-reasoning benchmark for both humans and artificial intelligence systems. No technical background in physics is assumed. The discussion relies on logic, language, and close reading. The goal is not to resolve paradoxes, but to show that none are required when foundational experiments are read precisely.

Feb 23, 20260 citations

Mathematics as Equalities and Inequalities, Application to Schrödinger's Cat and AI

dong zhang

We define mathematics as the collection of equalities and inequalities. Therefore, mathematicians perform two tasks: identifying identity (equality) and identifying difference comparison (inequality). Applying this definition to Schrödinger’s Cat, we identify the thought experiment as equivalent to a first-decay experiment. A first-decay event reveals the differences among atoms and exposes an implicit identity assumption in physics. We propose an objective test, the Schrödinger Test, for artificial intelligence, since the Turing Test is subjective and human linguistic. The test is: given Schrödinger’s original 1935 text, can a system derive the associated logical consequences, such as the non-identity of atoms?

Dec 19, 20250 citations

The Cat's Schrödinger and the Determinate in Schrödinger's Cat

dong ZHANG

Working from the German original of Schrödinger’s well known thought experiment (here given in translation), this paper sets out several ideas that are known and determinate in Schrödinger’s cat. The first is the cat’s Schrödinger — that is, the cat putting Schrödinger into the steel box. The most important one is that atoms of the same isotope differ. And so on.

Oct 16, 20260 citations

The Original Schrödinger's Cat Revisited: Schrödinger's Schrödinger

dong ZHANG

Schrödinger's Cat has acquired many descriptions and interpretations since its introduction in 1935. This paper returns to Schrödinger's original text and separates what Schrödinger explicitly constructed from what is inferred from the construction. Schrödinger specified a steel chamber, “a tiny amount” of radioactive substance, a Geiger counter, relay, hammer, poison, a one-hour interval, equal possibilities that one atom decays or none does, and “the first atomic decay” as the decisive occurrence. He then described the wave function of the entire system as containing the living and dead cat “mixed or smeared in equal parts.” However, not knowing which of two mutually exclusive states is true, or being unable to rule out either, does not imply that both are physically true. The distinction involves not only Schrödinger's Cat but also Schrödinger's YOU—the observer outside the steel chamber who lacks information. Moreover, the causal chain from atomic decay to the cat's death contains additional physical and biological conditions. The original construction should therefore distinguish the physical cat, its quantum-mechanical description, and the information available to YOU.

Aug 25, 20260 citations

The Original Schrödinger's Cat Revisited: Shut Up and Calculate

dong ZHANG

“Shut up and calculate” is commonly associated with a pragmatic attitude toward quantum mechanics: instead of debating interpretation, calculate observable predictions. This paper applies that instruction directly to the original Schrödinger's Cat. Schrödinger specified “a tiny amount” of radioactive substance, an equal probability that one of its atoms decays or none does within one hour, and “the first atomic decay” as the decisive occurrence. These statements contain two quantitative questions: What is the amount? What is the probability of the first decay? Using Schrödinger's Quantity and the First Decay time t₁, these become: calculate the number of atoms n satisfying the one-half probability condition, and calculate the probability associated with the first decay occurrence. No interpretation of the cat is required for either question.

Aug 25, 20260 citations

The Original Schrödinger's Cat Revisited: Isolated Systems

dong ZHANG

Schrödinger's original Cat places a radioactive substance, Geiger counter, relay, hammer, poison, and cat inside a closed steel chamber and leaves “this entire system” to itself for one hour. The construction is commonly treated as an isolated system. This paper examines what isolation means in the thought experiment. Physical isolation, causal connection, and information available about individual members of a system are different conditions. A comparison with human mortality illustrates the distinction. An actuary may possess individual information such as birthdays and living habits, yet has no probability information about when the first death claim will occur; it may come at any time and at any probability. For radioactive atoms, the standard description supplies no comparable individual information and treats atoms of the same isotope as identical or dismisses their differences for radioactive decay. Schrödinger's Cat therefore raises a question beyond the physical isolation of the steel box: what information about individual radioactive atoms is available to determine the first decay?

Aug 25, 20260 citations