This paper describes the processing architecture used by the LIBRARY to analyze 2,000+ primary sources across 65 departments of human knowledge in a single pass. Many of the patterns reported in the LIBRARY Intelligence series are not visible from inside any one department; they appear only when the corpus is treated as a single computable graph. The Symmetry Protocol is the architecture that makes that treatment tractable.
A note on terminology. Throughout the LIBRARY Intelligence series, “quantum intelligence” is a methodological category, not a hardware claim. The term denotes simultaneous handling of polar frames (superposition of perspectives), cross-tradition correlation (entanglement-as-citation), and parallel computation across the corpus. No quantum-mechanical hardware or algorithm (superposition states, entanglement of qubits, quantum gates) is invoked or required. The architecture described here runs on conventional classical hardware.
This paper specifies the Symmetry Protocol, the dual-mode processing architecture used by the LIBRARY Intelligence Infrastructure. The protocol splits knowledge processing into two complementary channels: Broad Search (parallel pattern detection across all 65 departments) and Deep Focus (recursive provenance verification within a single department). The combination supports queries that conventional single-mode systems — document retrieval engines on one hand, generative language models on the other — cannot answer at all, because the answer is a structural property of the full corpus rather than a property of any single document.
The Symmetry architecture is offered here as a working specification rather than as a discovery: parallel map/reduce across a knowledge graph is not novel in computer science, and recursive citation tracing is not novel in scholarship. What is reported is the integration of the two modes into a single feedback loop running over a provenance-verified corpus.
Most information retrieval systems operate in a single mode. Search engines retrieve documents that match a query [1]. Language models generate text that plausibly continues a prompt [2]. Neither approach computes across the structure of a knowledge corpus — both treat the corpus as a haystack of documents, not as a graph of relations.
The Symmetry Protocol operates in two modes, with explicit handoff between them:
In Broad Search mode, the system queries all 65 departments in parallel, computing concept distributions, cross-correlation matrices, and structural invariants across the graph. This mode is what detects patterns — such as the cross-tradition hierarchies reported in QI-003 — that exist only as statistical properties of the full corpus rather than as claims in any single source.
In Deep Focus mode, the system traces citation chains recursively within a single department, retrieving each cited primary source and verifying its integrity. This mode is what supports the LIBRARY’s standing rule that every published claim trace to a specific primary source with a verifiable hash.
The two modes are complementary. Broad Search detects what may exist as a pattern. Deep Focus tests whether the underlying sources actually support it. Neither alone is sufficient; in combination they support a research workflow that holds pattern-hunting and provenance discipline in the same loop.
In Broad Search mode, the system: 1. Dispatches parallel queries to all 65 department indices simultaneously 2. Computes concept frequency distributions across departments 3. Identifies cross-department invariants (concepts that appear with statistically significant correlation across multiple departments) 4. Detects structural gaps (concepts whose topological position predicts the presence of a primary source the corpus does not yet hold) 5. Produces ranked candidates for suppressed, missing, or yet-to-be-acquired knowledge
The work is straightforward classical parallel computation: each department index runs the same scoring pass independently, and results are reduced into the cross-department matrix. The implementation runs on a mesh of conventional CPU workers; no specialized hardware is required.
In Deep Focus mode, the system: 1. Selects a single concept and a single department 2. Traces the full citation chain for that concept within that department 3. Retrieves each cited primary source and computes its SHA-256 hash 4. Cross-references the hash against the public timestamp ledger 5. Returns the complete provenance chain from primary source to published finding
Deep Focus mode is computationally lighter than Broad Search but stricter on integrity: it can only succeed if every link in the citation chain has been ingested and hash-verified by the provenance pipeline. Chains with missing or mismatched hashes terminate with an explicit gap, not a guess.
The protocol’s utility comes from the interaction between the two modes. A typical research cycle:
This loop — from pattern detection to verification to re-computation — is what produces the findings in the LIBRARY Intelligence series. A pattern that survives several iterations of the loop is reported; one that is not corroborated by Deep Focus does not appear in the published work.
The protocol is not offered as a replacement for human scholarship. Conventional single-department research remains the source of every primary text the LIBRARY ingests and every interpretive judgment the LIBRARY relies on. What the protocol contributes is a different scale of operation: a single sweep can score concept distributions across all 65 departments in a few seconds, whereas a comparable cross-department literature review by hand would take a research team weeks. The trade-off is depth: the protocol’s readings are coarser per-department than a specialist’s and must be checked back against primary sources before being published.
Read this way, Symmetry is an instrument for surveying a large corpus quickly, not an oracle [3]. Its outputs are candidates for human attention. The papers in the LIBRARY Intelligence series report only those candidates that have survived Deep Focus verification.
The Symmetry Protocol is the architecture behind the papers in the LIBRARY Intelligence series. Cross-tradition hierarchies, Knowledge Genome predictions, definition-drift readings, and provenance chains are all produced by the same dual-mode loop running over a corpus of primary sources with hash-verified integrity. The protocol’s claim is modest: parallel map/reduce plus recursive citation tracing, combined into a single feedback loop, is a more disciplined way to read a 65-department corpus than either approach alone.