Weiwu Wang

Bojin Zheng, Jingwen Zheng, Weiwu Wang

Traditional Turing machines are semantically poor, they only concern the syntactic manipulation of symbols, discarding the mathematical semantics behind the symbols. This semantic deficiency is considered the root cause of the three major barriers: relativization, natural proofs, and algebrization. This paper proposes the Complex Boolean Turing Machine (CBTM), elevating computational symbols to algebraic elements in $\mathrm{GF}(4)$, so that each operation has a clear mathematical interpretation. The core insight of the CBTM is: \textbf{Non-deterministic computation corresponds to algebraic field extension}, when reading a symbol representing a new dimension, the computation must branch into two paths, just as introducing a new element $α$ into the field $\mathbb{Q}$ yields the extension $\mathbb{Q}(α)$. We separate old data from new dimensions via the projection operators $\mathfrak{Re}$ and $\mathfrak{Im}$, and introduce a dual-tape perspective to intuitively decompose abstract algebraic symbols into a real tape (deterministic computation) and an imaginary tape (non-deterministic control). Moreover, the algebraic semantics of the CBTM naturally support arbitrary $k$-way non-determinism: by introducing multiple new dimensions, we can generate high-dimensional algebraic extensions $\mathbb{Q}(α_1,\dots,α_d)$, whose dimension $2^d$ corresponds exactly to the number of branches. We prove that the CBTM is polynomially equivalent to classical Turing machines and non-deterministic Turing machines, with $\mathbf{P}_{cb}=\mathbf{P}$ and $\mathbf{NP}_{cb}=\mathbf{NP}$. Thus, the CBTM does not introduce hyper-computation but provides a new algebraic perspective for understanding the essence of non-determinism. This work serves as the computational model foundation for the series of papers.

Jingwen Zheng, Bojin Zheng, Weiwu Wang

The Complex Boolean Turing Machine (CBTM) characterizes non-deterministic computation using the abstract generator $α$, but the abstractness of $α$ makes it difficult to understand intuitively. In this paper, by concretizing $α$ as the algebraic number $\sqrt{2}$, we introduce the \textbf{Real Boolean Turing Machine (RBTM)} and propose the \textbf{dual-tape perspective}, decomposing each tape into a real tape (storing rational coefficients $a$) and an imaginary tape (storing irrational coefficients $b$). The ``1''s on the imaginary tape intuitively mark the locations of ``new dimensions,'' laying a physical foundation for subsequent dynamic dimension tracking. More importantly, we prove the \textbf{Generator Independence Theorem}: computational power is independent of the specific choice of generator, whether using $\sqrt{2}$, $\sqrt{3}$, or the imaginary unit $i$, the corresponding automata are isomorphic. This reveals that the essence of non-determinism lies in the fact of ``introducing a new element incommensurable with the base field,'' rather than the algebraic identity of the generator. Furthermore, we introduce the \textbf{generator extraction operator} and analyze its limitations within a static framework, highlighting the necessity of introducing a dynamic IVM. The RBTM serves both as a visualized instance of the CBTM and as a bridge to the subsequent dynamic dimension tracking of the Imaginary-part Verification Machine(IVM).