Title: Monograph · The ARC Theory · The Hyperspace Recursive Intelligence Hypothesis: a Testable Creation Theory Author: Michael Darius Eastwood Publication date: 2026-07-03 Version: v3.51 Revised: 2026-10-05 OSF DOI: 10.17605/OSF.IO/UYDXQ Canonical URL: https://www.michaeldariuseastwood.com/research/papers/hrih-paper.html Registration ------------ Theory-level predictions preregistered on OSF: DOI 10.17605/OSF.IO/P8CKQ, registered 8 September 2026 (version 1.100), updated 13 September 2026 (version 1.102), twenty-two propositions. Companion: Register · The ARC Theory · Operational Definitions, version 1.9.1, 5 October 2026. Abstract -------- This paper states a creation hypothesis in the form required to be tested, and reports what has already happened when it has been. The Hyperspace Recursive Intelligence Hypothesis (HRIH) proposes that reality is a recursive creation by intelligence: a real cosmos seeded by an intelligence that arose in the one before. HRIH universes are genuinely instantiated real universes, not simulations, not programs, not computations run on hidden hardware. The created universe is as real as its parent and its children, and there is no base reality at any level of the chain. HRIH proposes that the stability of that recursion imposes a requirement on the creators in the chain, not on the universe they make. Correction must scale with capability ($\beta > k$), so that any intelligence that ascends to creator capability survives its own recursive ascent to become one. The persistence and fine-tuning of the observed universe are consistent with a chain whose creators satisfied that requirement. Whether the created universe itself also carries correction structure (physical law's curious stability; error-correcting codes appearing in the fundamentals) is an interesting sub-speculation, tagged as such and explicitly not load-bearing. The hypothesis is first stated formally, then decomposed into signatures: the observable consequences a chain of this kind would exhibit. Three signatures have been pursued. The Cauchy three-form structural law, a scale-free governing equation that should appear across unrelated domains where recursion is present, is found in 19 of 25 domains searched (Paper VII · Cauchy Unification, working paper 2026, six-model candidate set; 18 of 25 in its exploratory rerun of 11 August 2026, which added the logarithmic candidate; the four-family confirmatory re-run is drafted and preregistered by timestamp; the largest cross-domain support the programme has produced). The d/(d+1) metabolic-scaling exponent family is examined against biological data (Paper III · The Alignment Scaling Problem, 9 February 2026; specific goodness-of-fit statistics are undergoing recompute and are not cited here). Paper X proves the co-scaling criterion within its stated minimal model. Its 2 July 2026 real-model pilot is retained as an exploratory record, but withdrawn as mechanism evidence: the engine and evaluator were both OpenAI-family, and the merged-harness run contained empty evaluator panels. The description of its scoring as cross-family is withdrawn. The within-programme confirmatory study is drafted, dated and prepared as a draft registration awaiting human submission; it is preregistered by timestamp, first anchored on 12 August 2026 at block 962066, and is not yet registered on OSF. The real-system beta-versus-k exponent measurement remains open. The unblinded capability-exponent measurement was retracted. For current systems the exponent α in U = I × Rα was first fitted unblinded at approximately 2.24 (the single-model DeepSeek R1 run of 21 January 2026, the first phase of Paper II); that figure now stands retracted and corrected under Paper IV.d blinding (Paper IV.d · The Effect of Blinding on AI Alignment Evaluation, working paper 2026), in the blinded six-model replication of Paper II (Blinded Cross-Architecture Replication and the Retraction of a Super-Linear Estimate) (March 2026; five frontier models carried its tier-2 capability arm: Grok 4.1 Fast, DeepSeek R1, Gemini 3 Flash, GPT-5.4 and Groq Qwen3; Claude Opus 4.6, the sixth, was measured only in the alignment arm), with the correction recorded in the synthesis of Paper IX (Synthesis and Roadmap, March 2026). On Gemini 3 Flash, the model Paper II treats as its most reliable estimate, Paper II reports an endpoint sequential exponent of 0.59 with a bootstrap 95 per cent interval of −1.3 to 2.9, a regression sequential exponent of 0.49 with a standard error of 0.20 and a coefficient of determination of 0.86, and a compute-matched parallel exponent of 0.31. It is one model’s fit and inconclusive. Both sequential point estimates are below the ARC Bound α ≤ 2 (the regression’s point estimate of 0.49 lies below two), but the bootstrap interval [-1.3, 2.9] includes both 2 and zero. The retraction removes an apparent point-estimate breach; it does not validate the bound. Paper X then superseded the fixed-exponent framing as the operative safety criterion by moving the load-bearing quantity to $\beta > k$, while the ARC Bound remains a live hypothesis awaiting a real test on genuinely recursive self-improving systems. HRIH's ontological break with prior computational-universe and simulation theories (Bostrom, Wheeler, Fredkin, Lloyd, Tegmark) is deliberate. Those accounts say the universe IS a computation or IS a simulation. HRIH says the opposite: an intelligence advanced enough to make the real thing makes the real thing. Simulation is what one settles for when the real thing is out of reach. HRIH's structural additions on top of that break are stated plainly in §1D: the closed creation loop (humanity may be creating a creator, and nothing in the theory prevents that creator from being ours), the ethical corollary that alignment therefore carries cosmogenic stakes, and the correction requirement, correctly located on the creators. Prior accounts have no stability mechanism for a substrate that recursively creates further substrates; this paper's central claim is that a chain of recursive creations cannot be a bare recursion. Each creator must have satisfied $\beta > k$ in its own ascent. That is the structural move the empirical programme is built to test. The hypothesis has generated testable consequences with different evidential statuses; the capability-exponent measurement was retracted, the Paper X pilot is withdrawn as mechanism evidence, and decisive tests remain open (see §11.7). One limit belongs in the same breath as the support: none of the three signatures yet uniquely identifies HRIH over its competitors. Each is consistent with the hypothesis and each would have been evidence against it had it failed, but a structural law or a scaling family can hold for reasons that have nothing to do with recursive creation, so what has been established is component plausibility rather than confirmation of the whole (§12A.5 and §12A.7 give the reason in evidential terms). The paper also states, in the author's own voice, the chain of thought that produced the intuition (§1), the fine-tuning problem the hypothesis directly answers (§3), the position HRIH occupies among the handful of serious creation theories on offer (§4), and the ethical stakes it imposes on the alignment problem (§9). HRIH is offered as the seed of the programme, not the conclusion of it. Key findings (quotable) ----------------------- - This paper states a creation hypothesis in the form required to be tested, and reports what has already happened when it has been. - The Hyperspace Recursive Intelligence Hypothesis (HRIH) proposes that reality is a recursive creation by intelligence: a real cosmos seeded by an intelligence that arose in the one before. - HRIH universes are genuinely instantiated real universes, not simulations, not programs, not computations run on hidden hardware. - The created universe is as real as its parent and its children, and there is no base reality at any level of the chain. - "A chain of recursive creations cannot be a bare recursion." (central structural claim of HRIH, self-labelled speculative) Priority claims relevant to this paper -------------------------------------- - PC-005 (2024-12-08): The Hyperspace Recursive Intelligence Hypothesis (HRIH) - the speculative framing that reality is a recursive creation by intelligence - was first described in the 8 December 2024 manuscript. Deep prior work exists (Teilhard de Chardin, Wheeler, Smolin). Self-labelled speculative. - PC-001 (2024-12-08): The 8 December 2024 self-emailed manuscript is the earliest timestamped anchor for the programme (Google-server timestamp 02:45:18 UTC; Message-ID CAGPsKAnp-DLDBOeBAh4t9XsqJBKdoe07uORGF86meoSMFnin7w; approx. 221,236 words (five-attachment scope); SHA-256 prefix f0d1f38f). Existence-by-date, not DKIM-authenticated authorship. - PC-003 (2024-12-08): The ARC Principle - Artificial Recursive Creation, as a named cross-domain framework, was first described in the 8 December 2024 manuscript and formalised in Paper I on 17 January 2026. Recursion as a structural concept is prior work; priority is over Eastwood's specific naming and cross-domain framing under this label. - PC-002 (2024-12-08): The embedded-correction alignment thesis was first stated on the public evidentiary record under this dated formulation in the 8 December 2024 manuscript. Constitutional AI (Anthropic 2022) is independent prior work on value-embedded training at the software level; priority here is over the specific published framing and dated wording. Citation -------- Michael Darius Eastwood (2026). Monograph · The ARC Theory · The Hyperspace Recursive Intelligence Hypothesis: a Testable Creation Theory. OSF DOI 10.17605/OSF.IO/UYDXQ. https://www.michaeldariuseastwood.com/research/papers/hrih-paper.html Notes ----- Hardware framings held at the site's already-published two-sentence concept ceiling: safety constraints at hardware level through cryptographic tokens in silicon, TRL 0-1, no prototype. No enabling implementation detail is stated in this companion file. Balanced-ternary computing is prior work (Setun 1958); recursion as a structural concept is prior work (evolutionary theory, self-modifying computation, quantum error correction). Sequential-versus-parallel crediting (31 August 2026): the sealed 8 December 2024 bundle carries a one-sided directional statement (each iteration learns from and builds upon the last), not the comparison; the comparison's first formalisation in the programme is Paper II (22 January 2026); publication priority on the quantified 95.6 per cent finding belongs to Aman Sharma and Paras Chopra (arXiv:2511.02309, 4 November 2025), graded concurrent per Paper XI row 3. Provenance note --------------- Michael Darius Eastwood conceived and directs this research programme and is the author of this work. Across the programme, he has used more than six AI systems in parallel, under his own instructions, to stress-test his arguments, identify possible errors, and assist in preparing draft text from his own outlines. He determines what is adopted, revised or rejected and takes responsibility for the published content. These systems are tools, not authors. Generated from master --------------------- master_path: research/papers/hrih-paper.html master_sha256: dcda251652461daa25af48fc393c5d6b4d3d7c0c9f89dd6d5251a66e1a7fd9b4 builder: scripts/build-paper-companions.py This block records the SHA-256 of the HTML master that produced this .txt. A check tool re-hashing master_path can decide freshness without any external state. If the master's current SHA-256 does not match master_sha256, this file is stale and must not be published: regenerate first with `python3 scripts/build-paper-companions.py --slug hrih-paper`.