Introduction to the book

Introduction to the book

From Dialectical Thought to a Computable Regime of Representation and Revision

Living book / framework documentation Version: v0.1 — Publication Manuscript Date: September 15, 2026 Status: Research is ongoing; this text is not a final theory.

This book does not try to teach dialectics to artificial intelligence as a style, persona, or list of ready-made answers. The question is harder: can an artificial system construct representations of the world through relations, scopes, processes, and changing evidence; and when those representations become inadequate, can it reorganize not only its answer but, when necessary, the representation itself?

---

This text is both a book and documentation

A book was traditionally completed, printed, and later updated through new editions. For Computational Dialectics, that format alone is not enough. What we are describing is not only a system of ideas; it is also a research framework that runs, is tested, sometimes breaks, and is rebuilt through experiments.

For that reason, this text combines two forms in one source. Read from beginning to end, it progresses like a book: what is the problem, where does the idea come from, how do the concepts connect, what do we claim, and where might we be wrong? Used chapter by chapter, it behaves like framework documentation: what does a mechanism do, which part is actually implemented today, which experiment tests it, and what is the next open problem?

The same source text can later be rendered as web pages, PDF, EPUB, or a printed book. The format may change; the scientific story should not.

Four layers for the reader

Throughout the book, we will separate four layers when evaluating a claim:

Source / Idea. What conceptual structure can we actually defend from Hegel, Marx, Ilyenkov, or contemporary scientific work?

Computational Translation. What assumption do we add in order to make that structure processable by a machine?

Experimental Status. What have we actually built and tested so far, and what failed?

Open Problem. What do we still not know, have not established, or cannot yet measure correctly?

This distinction is one of the most important methodological rules of the book. If a computational translation fails, a philosopher has not been "falsified"; our translation has failed. Conversely, the fact that an idea is philosophically attractive does not show that it is a good AI architecture.

---