# Foundations and Positioning of the Zemi Method **Companion to:** The Zemi Method, Version 1.0 **Status:** Published. Principal sources reviewed in full; maintained as the literature develops. **Published:** July 17, 2026 **Canonical home:** https://zemimethod.com **Copyright:** © 2026 Kevin V. Watson. All rights reserved. --- ## Purpose of this document The Zemi Method is a compact, normative doctrine. It states what the practice requires and stops there. This companion does the work the doctrine deliberately does not: it identifies the traditions the method relies on, acknowledges its predecessors, and states precisely what the method claims and what it does not. The method was not invented from nothing, and this document does not pretend otherwise. Most durable methodologies are built through synthesis, extension, and formalization of ideas that earned their standing elsewhere. The honest position, and the defensible one, is to name those ideas, show where they came from, and be exact about what the synthesis contributes. ## The traditions the method relies on **The epistemology of testing.** The method's premise, that every claim relevant to a decision is a hypothesis until independent evidence supports it, descends from the falsificationist account of scientific reasoning associated with Karl Popper. The requirement in the Frame phase that a working theory state its own breaking condition is that account applied to casework. The method makes no claim to this epistemology. It claims only its consistent operational use. **The anatomy of argument.** Stephen Toulmin's model of argumentation separates a claim from its supporting data, from the warrant connecting them, and from the qualifiers bounding them. The method's requirement that reports state the reasoning connecting evidence to conclusion, and its classification of findings as known, assumed, or undetermined, are working translations of that anatomy into report structure. **The mapping of evidence.** John Henry Wigmore's chart method, developed for legal evidence a century ago and carried forward in the modern evidence scholarship of Anderson, Schum, and Twining, established that every inferential step between evidence and conclusion can and should be made explicit and inspectable. The method's traceability requirement and its defensibility check are a lightweight, practice-ready expression of that tradition. **The competition of hypotheses.** Richards Heuer's Analysis of Competing Hypotheses, developed for intelligence analysis, requires that rival explanations be tested against the same evidence together, with disconfirmation privileged over confirmation. The Frame phase adopts this discipline directly: where the evidence admits more than one explanation, the surviving explanation is the one the evidence failed to break, not the one considered first. **The calculus of uncertainty.** Bayesian evaluative reporting, reflected in European forensic practice through likelihood-ratio approaches, represents the most developed probabilistic treatment of evidential strength. The method takes a different route for its reporting layer, categorical classification rather than probabilistic calculus, because categorical boundaries are auditable and communicable to decision-makers and triers of fact. Nothing in the method forbids probabilistic evaluation within analysis where the evidence type supports it. The choice concerns how conclusions are communicated and owned, not how strength of evidence may be assessed. **The precedent for graded vocabularies.** The GRADE framework in evidence-based medicine demonstrated that a spare, categorical grading language, published, versioned, and consistently applied, can become shared infrastructure for an entire field. The method's version discipline is modeled on that precedent deliberately. **The forensic process standards.** The established process models and consensus documents of the field, including the NIST forensic process, the ISO/IEC 27037 series, and the guidance of bodies such as SWGDE, govern the handling, acquisition, and examination of digital evidence. The method does not replace or compete with them. It governs the reasoning above them, and it expects engagements to comply with the applicable standards beneath it. **The tradition of investigative distrust.** The nearest published predecessor to the method's premise is Zero Trust Digital Forensics, proposed by Neale, Kennedy, Price, Yu, and Nuseibeh in 2022, which defined the strategy as one where each aspect of an investigation is treated as unreliable until verified, and introduced multifaceted verification of digital artefacts as its operating principle. The kinship is substantial and is acknowledged without reservation: both positions reject unexamined trust in investigative components, and both respond with verification. That paper also modeled the honest posture this document adopts, noting openly that the practices implementing its strategy were not all fundamentally new. The authors presented their strategy as theoretical, with its practical application deliberately left to future work, and the research program built on it has remained centered on the detection of artefact tampering. The Zemi Method stands in this tradition and extends it in kind as well as scope: from a verification strategy at the artifact level to an operating doctrine at the decision level, with a lifecycle, an adjudication gate, an automation boundary, and an assurance application the earlier work did not contemplate. **The emerging consensus on human accountability for automated assistance.** Across fields, a common position is forming: automated systems may assist evidence work, humans remain accountable for judgments, and consequential automated contributions must be disclosed. In evidence synthesis, the 2025 joint position of Cochrane, the Campbell Collaboration, JBI, and the Collaboration for Environmental Evidence requires human oversight and transparent reporting wherever automation makes or suggests judgments. In forensic practice, current vendor guidance holds that AI changes the review method while the evidentiary process and its human validation remain unchanged. The method's fourth principle, AI-assisted, never AI-decided, belongs to this emerging consensus and claims no priority over it. What the method adds is procedural form: a named adjudication phase that automation is barred from crossing, and a reporting requirement that keeps the automated contribution separable from human judgment. ## What the method inherits From these traditions the method inherits nearly all of its raw material: hypothesis testing, stated breaking conditions, explicit inference, competing explanations, bounded conclusions, verification before trust, chain of custody, and human ownership of judgment. None of these is claimed as original, and a reader who recognizes their sources is reading the method correctly. ## What the method contributes The contribution is integration, structure, and extension. Integration: the elements above are combined into a single doctrine with one epistemology, rather than existing as separate practices a skilled examiner may or may not apply. Structure: the doctrine is procedural. It has a lifecycle with a mandatory human adjudication gate, a categorical classification applied at that gate, a standing defensibility check, and a versioned, dated public form that an engagement can be held to. Extension: the same evidentiary standard is applied to a second subject. On the assurance side, the AI system itself becomes the matter under examination, its claims corroborated, its controls evidenced, its accountability located. The symmetry between investigating with AI assistance and investigating AI systems under one discipline is, to the author's present knowledge, not found elsewhere in a practitioner doctrine. That claim is bounded: it reflects the literature reviewed as of this document's date, and this companion will be revised as the review deepens. ## What is claimed, and what is not Claimed: the Zemi Method integrates established principles of forensic verification, evidentiary reasoning, bounded conclusions, and human accountability into a single, versioned methodology governing both digital investigation and AI assurance. Not claimed: that the method invented independent verification, human accountability for automated assistance, falsifiable framing, or any of its component principles. Those claims would be false, and the method's own third principle forbids making them. ## The gaps the method is designed to address The traditions above leave a practitioner in a specific position: the reasoning tools exist in scholarship, the handling standards exist in consensus documents, and the accountability principle is emerging in position statements, but no single, published, versioned doctrine binds them into an operating discipline that a client can read before an engagement and hold the practice to afterward, that names where automation stops, and that carries the same standard into the assurance of AI systems. That is the gap the Zemi Method occupies. Whether it occupies it well is a question its casework and its version history will answer, which is as it should be. --- ## References Anderson, T., Schum, D., & Twining, W. (2005). *Analysis of Evidence* (2nd ed.). Cambridge University Press. Champod, C., Biedermann, A., Vuille, J., Willis, S., & De Kinder, J. (2016). *ENFSI Guideline for Evaluative Reporting in Forensic Science: A Primer for Legal Practitioners*. Flemyng, E., Noel-Storr, A., Macura, B., Gartlehner, G., Thomas, J., Meerpohl, J. J., Jordan, Z., Minx, J., Eisele-Metzger, A., Hamel, C., Jemiolo, P., Porritt, K., & Grainger, M. (2025). Position statement on artificial intelligence (AI) use in evidence synthesis across Cochrane, the Campbell Collaboration, JBI and the Collaboration for Environmental Evidence 2025. *Environmental Evidence, 14*, 20. https://doi.org/10.1186/s13750-025-00374-5 Guyatt, G. H., Oxman, A. D., Vist, G. E., Kunz, R., Falck-Ytter, Y., Alonso-Coello, P., & Schunemann, H. J. (2008). GRADE: An emerging consensus on rating quality of evidence and strength of recommendations. *BMJ, 336*(7650), 924-926. Heuer, R. J. (1999). *Psychology of Intelligence Analysis*. Center for the Study of Intelligence, Central Intelligence Agency. Neale, C. (2023). Fool me once: A systematic review of techniques to authenticate digital artefacts. *Forensic Science International: Digital Investigation, 45*, 301516. https://doi.org/10.1016/j.fsidi.2023.301516 Neale, C., Kennedy, I., Price, B., Yu, Y., & Nuseibeh, B. (2022). The case for Zero Trust Digital Forensics. *Forensic Science International: Digital Investigation, 40*, 301352. https://doi.org/10.1016/j.fsidi.2022.301352 Popper, K. R. (1959). *The Logic of Scientific Discovery*. Hutchinson. Toulmin, S. E. (1958). *The Uses of Argument*. Cambridge University Press. Wigmore, J. H. (1913). *The Principles of Judicial Proof*. Little, Brown and Company. Willis, S. M., McKenna, L., McDermott, S., O'Donnell, G., Barrett, A., Rasmusson, B., Hoglund, T., Nordgaard, A., Berger, C. E. H., Sjerps, M. J., Molina, J. J. L., Zadora, G., Aitken, C. G. G., Lovelock, T., Lunt, L., Champod, C., Biedermann, A., Hicks, T. N., & Taroni, F. (2015). *ENFSI Guideline for Evaluative Reporting in Forensic Science*. European Network of Forensic Science Institutes. --- *This companion is maintained alongside the method and revised as the surrounding field develops. Its purpose is not to argue that the method is unique. It is to make sure the method's debts are on the record before its contributions are.*