Parallel Scientific Expression
A Dual-Layer Method for Human–AI Research Communication
A Secretary Suite Project
John Swygert
November 14, 2026
DOI: To be assigned
Abstract
This paper proposes Parallel Scientific Expression (PSE), a dual-layer research communication method in which a fundamental-language account and a technically rigorous scientific account are authored, aligned, and maintained together. The first layer does not merely advertise or simplify the second: it provides an intellectually faithful entry point, a conceptual key, and a navigational roadmap that prepares readers to examine formal definitions, mathematics, evidence, and methodological constraints. The second layer preserves full scientific rigor. The proposed contribution is a traceable correspondence between the layers, including explicit marking of assumptions, qualifications, and inferential steps introduced during formalization. A principal aim is education—not remediation of any supposed deficiency in readers, but expansion of access to advanced ideas and development of scientific literacy across backgrounds and disciplines. The proposal is motivated in part by human–AI collaboration, which can accelerate production of sophisticated arguments faster than their reasoning can be independently understood or checked. PSE is offered as a testable communication and educational methodology, not as a demonstrated improvement.
1. The communication problem
Scientific papers often demand that readers understand specialized vocabulary, mathematical notation, modeling conventions, and the assumptions of a field before they can evaluate a central claim. Those requirements are sometimes necessary for precision. Yet they can also make the route from an intelligible idea to its formal development difficult to follow. A reader may understand the originating question but lose the thread as terminology and equations accumulate.
AI-assisted research adds a further challenge. A human may contribute a clear seed thought, while an AI system rapidly develops definitions, mathematical sketches, objections, and study designs. Such elaboration can be productive, but fluency and complexity are not substitutes for correctness. If the original idea disappears beneath technical expression, neither the human contributor nor outside reviewers may be able to identify where the argument became stronger, changed direction, or acquired unsupported assumptions.
PSE addresses this as a problem of intellectual traceability and education. It does not assume that specialist language is inherently bad or that all mathematics can be translated without loss. It requires the translation boundary itself to be visible.
2. Central proposal: two papers, one connected argument
Layer One: Fundamental Expression
The first paper states the seed thought, its significance, the underlying logic, the meaning of key terms, and the limits of the claim in language that an attentive nonspecialist can follow. It explains why each technical step will become necessary. Its purpose is not to replace technical learning, but to prepare readers for it—like a key and a roadmap supplied before entering a complex landscape.
Layer Two: Scientific Expression
The second paper develops the same research problem with discipline-appropriate definitions, equations, derivations, evidence, methodological designs, uncertainties, alternative models, and decision rules. It is a scientific paper in its own right, not a simplified version and not exempt from peer review.
The essential link
Both layers share a claim map and section-level correspondence. Where Layer Two introduces a stronger assertion than Layer One, that addition must be named and justified. Where a mathematical expression cannot be translated faithfully, Layer One must explain what the expression does and does not establish, rather than offering a misleading substitute.
3. Two hypotheses, explicitly connected
Each paired publication should present at least two versions of its principal hypothesis. The fundamental hypothesis (H-F) states the seed thought in direct, ordinary language. The scientific hypothesis (H-S) specifies variables, scope, predicted relationships, and possible observations. H-S must be connected to H-F by a documented chain of assumptions. It must not be portrayed as a mere restatement if it adds a causal mechanism, a measurement model, or stronger commitments.
Illustrative example: subjective experience
H-F: “A person might experience something that seems endless even when the physical event lasts only a limited time.”
H-S: “In a specified class of reports, perceived temporal boundlessness is hypothesized to covary with independently coded features of recursively structured representation after controlling for narrative complexity and elicitation conditions.”
These are not identical claims. H-S adds an association, an operational domain, measurements, and controls. The parallel presentation teaches the reader precisely how a broad intuition becomes a restricted scientific proposition, and why evidence for H-S would not establish every possible interpretation of H-F.
4. Education as a central scientific benefit
A major purpose of PSE is to make sophisticated scientific reasoning more learnable. The audience is not divided into “educated” and “uneducated” groups. Researchers themselves routinely encounter unfamiliar fields; students, engineers, practitioners, policymakers, and interested members of the public may each have expertise in some domains and little formal preparation in others. Accessible entry points serve all of them.
The first layer is intended to build a mental model before introducing the full symbolic system. It teaches the meaning of the question, the structure of the argument, the motivation for each variable, and the role of each assumption. Readers can then approach the second layer with landmarks already established. This resembles a guided route through a difficult map: the route does not remove the terrain, but makes it possible to navigate the terrain deliberately.
The educational value should be measured in more than perceived clarity. A reader who completes Layer One should be better prepared to identify the research question in Layer Two, explain the meaning of formal terms, distinguish assumptions from observations, recognize limits of inference, and detect when an equation is being asked to support more than it actually establishes. Genuine comprehension includes the ability to disagree intelligently.
5. Proposed structural standard
1. Seed thought / research question
Layer One presents the question in everyday terms; Layer Two formalizes the research domain and scope.
2. Fundamental hypothesis / scientific hypothesis
H-F and H-S appear side by side or in directly cross-referenced sections, with added assumptions identified.
3. Conceptual explanation / formal definitions
Every major symbol and construct has a plain-language meaning, including boundaries of the analogy.
4. Why it may work / mechanism and model
Layer One supplies the intuitive rationale; Layer Two distinguishes models from demonstrated mechanisms.
5. What is known / evidence review
Both layers separate established findings, interpretations, and open questions.
6. What could be wrong / alternatives
Competing explanations and possible failures appear in both layers, at different technical depth.
7. How to investigate / methods
Layer One explains the logic of a test; Layer Two specifies design, measurement, statistics, and feasibility.
8. What results mean / decision rules
Both layers distinguish evidence supporting a narrow claim from proof of a broader worldview.
Paired section identifiers, a shared glossary, and a claim-correspondence table can help readers move between the two texts without losing their place. These are proposed implementation conventions, not mandatory features of every discipline.
6. Safeguards against misleading simplification
A readable explanation can be wrong even when it sounds convincing. PSE therefore requires explicit disclosure of analogy, uncertainty, and omitted detail. A plain-language passage should never quietly convert correlation into causation, possibility into probability, or a model into an observed mechanism. Equations should not be used as decoration or as a substitute for empirical support.
The two papers should also preserve a visible record of divergence: (a) assumptions newly introduced in the technical formulation; (b) aspects of the seed thought not tested by the technical hypothesis; (c) uncertainties that remain unresolved; and (d) places where a faithful nontechnical translation is incomplete. If an important technical claim cannot be explained in Layer One, it must be flagged for review—not automatically declared invalid, but not concealed.
7. Human–AI collaboration and accountability
AI systems can assist with drafting, mathematical notation, literature organization, code, and the translation between registers. Human researchers remain responsible for verifying claims, validating citations, choosing justified methods, and deciding what is published. The PSE workflow may be especially useful when a human originates an idea and an AI expands it: the two layers preserve the conceptual starting point and make subsequent transformations inspectable.
A practical workflow is iterative rather than linear. Begin with H-F, develop H-S, record new assumptions, revise the fundamental explanation to reflect the actual scientific claim, and return to the technical paper when the explanation reveals a gap. Neither layer should be treated as an unquestionable authority over the other.
8. A testable research program for PSE
The claim that PSE improves understanding should itself be tested. A randomized study could assign readers to (A) a conventional technical paper, (B) the same technical paper plus a standard plain-language summary, or (C) a synchronized PSE pair. Participants could include students, researchers outside the target discipline, and domain specialists. Materials should be matched for factual content, with reading time and prior knowledge measured rather than assumed equal.
Primary outcomes might include accuracy on conceptual questions, correct interpretation of an equation or figure, ability to identify assumptions and unsupported inferences, and delayed retention. Secondary outcomes could include confidence calibration, transfer to a new example, time to comprehension, and perceived cognitive load. A particularly important test is whether readers can detect an intentionally inserted but ethically disclosed methodological flaw in a controlled evaluation exercise.
Analyses should preregister primary outcomes and distinguish the effects of additional reading time from the effects of paired structure. A null result would weaken the claim that synchronization offers benefits beyond conventional summaries; differential results might show that PSE is useful for certain audiences or topics but not universally.
9. Practical publication and maintenance
A PSE pair could be released as two separately citable papers with related titles, shared project identifiers, reciprocal links, and synchronized version dates. The fundamental paper must be substantial enough to stand alone; the technical paper must be complete enough for ordinary scientific scrutiny. Updates that alter a claim, assumption, or conclusion in one layer should trigger review of its counterpart. A correspondence table can record the mapping between claims across versions.
PSE need not be imposed on every publication. For some short reports a well-made abstract may suffice. Its strongest use cases may be interdisciplinary proposals, conceptually novel frameworks, mathematically dense explanations, public-interest science, and human–AI co-developed research where traceability matters.
10. Novelty, limits, and conclusion
Plain-language summaries, educational companion pieces, and layered exposition already exist. PSE does not claim to invent accessible scientific writing. Its proposed distinguishing feature is the disciplined, bidirectional correspondence of two full research expressions, with explicit mapping of hypotheses, assumptions, evidence, and limitations, maintained as a pair and evaluated for educational effectiveness.
The central principle is simple: scientific sophistication should not require the disappearance of the originating idea. The first layer gives readers a key and a roadmap; the second offers the full scientific terrain. Together, they aim to make advanced research more understandable, more teachable, and potentially more open to informed criticism. Whether the method delivers those benefits is an empirical question that should be tested.
Reference and positioning note
This document is a methodological proposal, not a systematic literature review or a claim of experimentally established educational efficacy. Related practices include structured abstracts, plain-language summaries, explanatory companion articles, and pedagogical scaffolding. A subsequent empirical study should review and compare these practices before claiming distinctive effectiveness.
No comments:
Post a Comment