The Convergent Diachronic Revitalisation System (CDRS) is a language-independent formal model class for the structured revitalisation of diachronically related linguistic stages. It provides an abstract computational and theoretical architecture — a directed acyclic graph of language states connected by deterministic transformation functions — within which any historically ordered sequence of linguistic stages can be systematically recovered and made operationally functional for communicative use.
The CDRS is currently instantiated as the Brittonic Convergent Diachronic Revitalisation System (BCDRS), applied to the Brittonic linguistic continuum from Modern Welsh through to Revitalised Cumbric. Within the BCDRS, the CDRS abstract pipeline is realised through three deterministic transformation frameworks: the Middle Welsh Revitalisation Framework (MWRF), the Old Welsh Revitalisation Framework (OWRF), and the Northern Brittonic Toponymic Revitalisation Framework (NBTRF).
Unlike purely reconstructive approaches to historical linguistics, each output layer of a CDRS implementation is immediately operational: it constitutes a fully functional language system capable of supporting human communication, not merely a scholarly artefact. The system's deterministic character makes every output auditable and every revision traceable, while its convergence property ensures that iterative refinement improves outputs without destabilising existing operational language systems.
The CDRS is governed by a formal peer-review pathway administered by Penrith Beacon Communications (PBC). All proposed modifications to the model — whether to its abstract specification, its Brittonic implementation, or any of the three transformation frameworks — must be supported by peer-reviewed academic documentation and evaluated by PBC before incorporation into the Polyglot™ dataset.
The Convergent Diachronic Revitalisation System (CDRS) is a language-independent formal model class for the structured revitalisation of diachronically related linguistic stages. It is language-independent in the most thorough sense: the formal structure of the CDRS — its directed acyclic graph architecture, its transformation function specification, its governance protocols — contains no reference to any particular language family, phonological system, or morphological paradigm. The same model governs any linguistic continuum for which the relevant historical linguistic evidence is available, whether the continuum in question is Brittonic, Goidelic, Germanic, Romance, or any other.
The model operates on diachronically ordered language states. A language state, in the CDRS framework, is the full linguistic system of a language or dialect at a particular historical period — its phonology, morphology, syntax, and lexicon as documented or reconstructible from available evidence. A diachronically ordered sequence of language states is a sequence {L₀, L₁, ..., Lₙ} in which each state is historically later than the preceding one: L₀ precedes L₁ in time, L₁ precedes L₂, and so on. The CDRS models the systematic transformation of each state into its successor through a deterministic transformation function.
The CDRS is formally characterised by five system properties, each of which has both formal and practical significance:
| Property | Formal content | Practical significance |
|---|---|---|
| Acyclic | The transformation graph is a DAG; no cycles permitted | Prevents circular reasoning in diachronic modelling |
| Diachronic | All transformations are ordered with respect to historical time | Grounds the model in documented history rather than invention |
| Deterministic | F(x) = y under fixed parameterisation; same input always yields same output | Makes outputs auditable and subject to scholarly challenge |
| Convergent | Iterative refinement produces progressively stabilised outputs | Allows improvement without disruption of existing operational outputs |
| Operational | All output layers are fully functional language systems | Outputs are learnable, teachable, and speakable — not merely archival |
The significance of language-independence cannot be overstated. The CDRS is not a methodology developed specifically for Brittonic and then generalised; it is an abstract formal structure within which any domain-specific implementation — called a Model Class Instance — operates. The BCDRS is the first such instance, but the model is explicitly designed to accommodate future instantiations for other language families. This places the CDRS within the broader tradition of formal language modelling, alongside frameworks such as Optimality Theory, which similarly provide abstract structures instantiated by domain-specific content.
The CDRS is governed by six core terms, each with a precise formal definition that differs in important ways from the term's colloquial or general scholarly usage. The definitions below are the operative definitions for all CDRS purposes; any ambiguity in application should be resolved by reference to these definitions and, where necessary, to the formal annexes maintained in the companion documentation.
Acyclic refers to the structural property of the CDRS transformation graph that prohibits directed cycles. Formally, a directed graph G = (V, E) is acyclic if and only if there exists no sequence of vertices v₁, v₂, ..., vₖ such that each consecutive pair forms a directed edge and vₖ = v₁. In the CDRS context, the vertices are language states and the directed edges are transformation functions; acyclicity means that no sequence of transformations can return to a previously visited language state.
The linguistic content of this requirement is a prohibition on circular reasoning. A reconstruction of Old Welsh cannot be derived from a form that was itself derived from Old Welsh; a Cumbric form cannot be used to revise its Old Welsh input after the fact. In practice, acyclicity is enforced by the pipeline structure: each transformation function takes as input only the output of the immediately preceding stage, and changes to any stage must propagate forward through the pipeline, never backward. In the Brittonic context, this means that a discovery about Cumbric phonology may prompt a revision to the xcb column, but it cannot directly revise the owl column — the owl column can only be revised by evidence relating to Old Welsh itself, and any resulting change to owl must then cascade forward to xcb through the NBTRF transformation.
Diachronic (and its adverbial form diachronically) refers to the relationship between linguistic phenomena across distinct historical periods, as opposed to synchronic, which concerns phenomena within a single period. The CDRS is a diachronic model: it models language change across time, not language structure at a single time. Every transformation function in the CDRS models the historical developments that occurred between two language states — phonological mergers, morphological simplifications, lexical borrowings, and the full range of documented linguistic changes.
The importance of diachronicity in the CDRS lies in its grounding function. A revitalisation project that is not diachronically constrained is, in effect, a language invention project — it creates forms by fiat rather than by principled derivation from historical evidence. The CDRS requires that every output form be derivable from its predecessor through documented or documentable historical processes. This is the mechanism by which CDRS outputs claim scholarly legitimacy: they are not invented but derived, and the derivation is explicit and auditable. In the Brittonic context, the diachronic ordering runs from Modern Welsh through Middle Welsh and Old Welsh to Cumbric — a sequence spanning roughly fifteen centuries of documented or partially documented linguistic development.
Revitalisation, within the CDRS, is the process of restoring a language system to operational communicative use through the application of deterministic diachronic transformations to documented or reconstructed ancestral forms. This definition has three components that deserve separate attention.
First, revitalisation is a restoration, not an invention. The CDRS does not create language forms from scratch; it derives them from documented predecessors through explicit rules. Second, revitalisation targets operational communicative use — the output must be usable for communication, not merely for scholarly analysis. Third, the transformations must be deterministic and diachronically grounded — not arbitrary or invented. These three requirements together distinguish CDRS revitalisation from both reconstruction (which may not target operational use) and from invented constructed languages (which are not diachronically grounded). In the Brittonic context, revitalisation means producing a Cumbric that speakers can learn and use, derived from Old Welsh by documented historical processes.
A revitalisation transformation is a single deterministic function Fₖ mapping one language state Lₖ₋₁ to the next Lₖ, parameterised by a complete set of rules Pₖ governing the transformation at stage k. A revitalisation transformation is not a single operation but a structured framework: it encompasses all the phonological, morphological, and lexical rules needed to derive every entry in Lₖ from the corresponding entry in Lₖ₋₁.
The parameterisation Pₖ is the intellectually substantive content of each transformation: it is the set of linguistic rules, historical evidence, and scholarly judgements that govern how the transformation operates. In the BCDRS, each of the three frameworks (MWRF, OWRF, NBTRF) is a revitalisation transformation with its own parameterisation. The parameterisation is not fixed forever; it can be refined as evidence improves. But any change to a parameterisation requires the peer-review governance process described in §7.
Convergence is the property by which iterative refinement of transformation parameterisations produces outputs that stabilise progressively rather than oscillating or diverging. Formally, the sequence of output states produced by successive refinements of Pₖ converges in the space of possible language states with respect to a linguistic distance metric. Practically, this means that each accepted revision to the framework rules moves the output closer to the ideal target state, and that the magnitude of revision decreases over time as the most significant gaps in evidence are addressed.
Convergence is importantly distinct from stasis: the CDRS does not require that the output be frozen. On the contrary, the explicit design of the system anticipates ongoing refinement as new evidence emerges from academic research. What convergence guarantees is that this refinement is bounded and progressive — each revision is an improvement, not a wholesale replacement. In the Brittonic context, convergence means that as new toponymic studies, revised manuscript analyses, or improved comparative reconstructions emerge, the BCDRS outputs can incorporate this evidence without the system becoming unstable or requiring complete reconstruction.
The CDRS is a model class: a parameterised family of formal models sharing the same abstract structure but differing in their domain-specific content. Every CDRS instantiation (called a Model Class Instance) inherits the five system properties and the DAG architecture, but instantiates them with the specific language states, transformation functions, and evidential bases appropriate to its linguistic domain. The BCDRS is the first Model Class Instance. The model class concept is the mechanism by which the CDRS achieves language-independence: the abstract structure is shared across all instances, but the linguistic content is domain-specific and independently governed.
The CDRS transformation graph G is a directed acyclic graph in which linguistic states are represented as nodes and transformation functions as directed edges. The direction of each edge reflects the direction of historical time: the edge points from the earlier language state to the later one. The acyclicity constraint — the prohibition of directed cycles — ensures that this historical directionality is preserved throughout the model.
The abstract structure of a three-stage CDRS instantiation is as follows:
Each node Lₖ represents a complete language state — the full phonological, morphological, and lexical system of a language at a particular historical period. Each edge Fₖ represents a deterministic transformation function that maps every entry in Lₖ₋₁ to a corresponding entry in Lₖ. The full set of entries mapped by Fₖ constitutes the complete output state Lₖ.
In the BCDRS instantiation, this abstract structure is populated as follows:
The DAG architecture has two critical consequences for the scholarly and intellectual character of CDRS outputs. First, it establishes unambiguous provenance for every output: any form in L₃ can be traced through L₂ and L₁ to its source in L₀. The transformation path is fully documented at each stage, and any gap or uncertainty in the transformation is explicitly recorded and confidence-graded. Second, the acyclicity constraint prevents the most common methodological error in linguistic reconstruction — the use of a later, partly speculative form as evidence for an earlier form that was itself used to derive the later form. In a CDRS implementation, the arrow of evidence always points forward: no output may be used as evidence for any of its inputs.
The graph-theoretic framework also provides the formal basis for cascade analysis. When a change is made to a node — whether by revision of a source entry or by refinement of a transformation rule — the cascade implications can be computed by traversing the graph forward from the changed node. All downstream nodes are potentially affected; all upstream nodes are unaffected. This provides the formal justification for the cascade protocols described in the BCDRS augmentation documentation.
Determinism is the property that, under fixed parameterisation, the transformation function F produces exactly one output for each input: F(x; P) = y, where x is the input entry, P is the fixed parameterisation (the framework rules), and y is the unique output entry. There is no branching, no probabilistic selection, and no non-deterministic choice in the transformation function itself. The uncertainty that characterises historical linguistic reconstruction — the genuine scholarly uncertainty about what a given Cumbric form might have been — is handled not by branching but by the confidence grade system: a single best output is produced, and that output is explicitly graded for confidence.
The formal properties of deterministic transformation within the CDRS are:
The significance of determinism for the CDRS's intellectual ambitions is profound. Determinism makes CDRS outputs auditable: any output can be checked by any competent scholar who has access to the rules and the source data. Determinism makes outputs subject to scholarly challenge: if a scholar disagrees with an output, they must identify either an error in the source data or an error in the rules — the challenge is specific and answerable. And determinism makes outputs amenable to formal governance: when a rule is revised, the implications of the revision can be computed precisely, and all affected entries can be identified and re-evaluated.
This last point connects determinism to the peer-review pathway. Because every output is a function of specific rules, a proposed change to a rule has determinate implications for the output. The peer-review process evaluates not only the evidential basis for the proposed rule change but also its downstream implications — which entries change, how their confidence grades are affected, and whether any cascade obligations arise at downstream stages of the pipeline.
Convergent stabilisation is the property by which iterative refinement of the transformation parameterisation produces outputs that progressively approach a stable limiting state, rather than oscillating arbitrarily or diverging from any useful approximation. It is the CDRS's answer to a fundamental challenge for any revitalisation project: how can the system be open to improvement as evidence accumulates, while still providing a stable operational output that learners can acquire and speakers can use?
The key distinction is between convergent refinement and static freezing. Static freezing — the approach that refuses to change outputs once established — provides stability at the cost of accuracy: the initial output, however good, cannot be improved even when better evidence becomes available. Unconstrained revision — the approach that allows any output to be changed at any time for any reason — provides accuracy at the cost of stability: no learner can rely on the system because it may change arbitrarily.
The CDRS convergence property navigates between these extremes through three mechanisms. First, all revisions must pass the peer-review governance process: changes cannot be made arbitrarily but only when evidential justification has been demonstrated and evaluated. Second, each revision is bounded: it affects only those entries for which the evidential basis has changed, leaving unaffected entries stable. Third, each revision is versioned: the full history of all changes is documented, so that any version of the output can be recovered and compared with any other. Together, these mechanisms ensure that the CDRS converges — each revision is an improvement, the improvements are bounded, and the overall trajectory of the system is toward progressively more accurate outputs without disruption of the operational language system at any stage.
Every output layer produced by a CDRS implementation is required to constitute a fully operational language system — one that can be learned, taught, and used for communication. This requirement, the operational output requirement, is the most distinctive feature of the CDRS and the feature that most sharply distinguishes it from purely reconstructive approaches to historical linguistics.
Purely reconstructive approaches are concerned with what the language was. They ask: given the available evidence, what phonological, morphological, and lexical forms did this language stage have? The output of reconstruction is a scholarly model — a best estimate of the historical form, typically accompanied by scholarly apparatus, caveats, and alternative possibilities. Reconstruction is invaluable for historical linguistics, but its output is not necessarily intended for use. A reconstructed Proto-Indo-European paradigm is a scholarly achievement; it is not, in itself, a language that anyone can learn or speak.
CDRS outputs, by contrast, are designed to be used. The operational requirement means that every output layer must provide: a complete entry for every grammatical and lexical domain in the dataset (no gaps allowed — gaps are filled with the best available approximation, explicitly graded); forms that are internally consistent (no contradictory forms within a single output layer); and forms that are learnable from documented rules (not merely listed without derivation). These requirements ensure that the output layer is, in the fullest sense, a language system — not merely a list of forms, but a structured system that a learner can acquire and a speaker can use productively.
The operational requirement does not preclude uncertainty. On the contrary, the confidence grade system is the mechanism by which the CDRS acknowledges uncertainty while still meeting the operational requirement: a Grade C form is uncertain, but it is still an operational approximation — it is the best available form for that entry, and it is presented honestly as uncertain rather than concealed or omitted. This is precisely the approach taken by successful revitalisation projects: the Cornish revitalisation did not wait until every form was certain before producing an operational language; it produced the best available system and refined it as evidence and scholarly consensus improved.
The CDRS is maintained by Penrith Beacon Communications (PBC), which exercises governance over the abstract model and, through the BCDRS, over its Brittonic implementation. Governance operates through a formal peer-review pathway that applies to all proposed changes at every level of the system: changes to individual dataset entries, changes to framework rules, changes to the BCDRS system architecture, and changes to the CDRS abstract model itself.
Submission. All proposed changes are submitted via www.penrithbeacon.com in the form of a CDRS Amendment Proposal (CAP) or, for framework-level changes, a Framework Amendment Proposal. The proposal must identify the specific change proposed, the evidential basis for the change, and a preliminary assessment of downstream implications. Proposals lacking these elements are returned without evaluation.
Evidential requirements. All proposals must be supported by peer-reviewed academic documentation. Pre-prints may be cited as supplementary support but are not acceptable as the primary evidential basis. The quality and scope of evidence required scales with the level of change proposed: a change to a single dataset entry requires strong evidence for that specific entry; a change to a framework rule requires evidence sufficient to justify the rule across all entries it governs; a change to the CDRS abstract model requires evidence of theoretical significance at the level of historical linguistics as a discipline.
Evaluation. PBC evaluates each proposal for: formal consistency with the CDRS architecture; adequacy of evidential support; internal consistency of the proposed change; and downstream implications across the BCDRS pipeline. Where specialist expertise is required, PBC may commission external review from Celtic Studies departments — at Aberystwyth, Bangor, Cardiff, Edinburgh, or Glasgow — or from other appropriate academic institutions.
Decisions and versioning. PBC issues one of three decisions: accepted, accepted with modification, or rejected. All decisions are documented with reasons. Accepted changes are versioned: the CDRS uses a MAJOR.MINOR versioning scheme in which MINOR increments reflect clarifications and refinements, while MAJOR increments reflect changes to the formal structure of the model. The BCDRS and its frameworks are versioned independently; a CDRS change triggers a BCDRS review to determine whether a BCDRS version increment is required.
Downstream impact assessment. Every accepted change triggers a formal assessment of downstream implications. For a change to the xcb column, the assessment is trivial (no downstream columns exist). For a change to owl, the assessment must evaluate all xcb entries derived from the changed owl entry. For a change to wlm, the assessment cascades through owl to xcb. For a change to a framework rule, the assessment must evaluate all entries governed by that rule. The downstream impact assessment is documented in the amendment record and forms part of the CDRS audit trail.
The peer-review requirement is, in the deepest sense, the embodiment of the CDRS's claim to intellectual seriousness. The model asserts that Revitalised Cumbric — and each intermediate layer — is not invented but derived: derived from evidence, by rules, through a process that any competent scholar can examine and challenge. The peer-review pathway is the institutional mechanism that maintains that claim across the lifetime of the project.
The CDRS rests on a foundational axiom that distinguishes revitalisation from reconstruction. This distinction is not merely terminological; it reflects a fundamental difference in intellectual purpose and methodological design that shapes every aspect of how the CDRS operates.
Reconstruction is the inferential historical modelling of what a language was. It asks the backward-looking question: given the evidence available, what forms did this language stage have? Reconstruction is the core activity of historical linguistics, and it is necessarily probabilistic and provisional. Different scholars may reach different conclusions from the same evidence; the output is a scholarly estimate, not a definitive account. The primary audience for reconstruction is the scholarly community; the primary product is academic knowledge.
Revitalisation, as the CDRS defines it, asks the forward-looking question: given what we know about this language, what operational system can we build? Revitalisation incorporates reconstruction as a necessary sub-process — each CDRS framework uses the full methodology of historical linguistics, including comparative reconstruction, internal reconstruction, and the evaluation of fragmentary evidence — but it does not treat reconstruction as the end goal. The end goal is an operational language system: one that learners can acquire, speakers can use, and communities can build on.
The practical consequence of this axiom is that the CDRS handles uncertainty differently from purely reconstructive approaches. A purely reconstructive approach may withhold a form entirely if the evidence is insufficient to determine it with scholarly confidence. The CDRS requires that every entry be populated with the best available operational approximation, explicitly graded by confidence. This is not a concession to imprecision; it is a methodological commitment. Operational revitalisation requires a complete system — a language without any of its basic forms is not a usable language — and the CDRS meets this requirement by being explicit about uncertainty rather than concealing it through omission.
The axiom also has important implications for how CDRS outputs are presented and interpreted. Revitalised Cumbric is not a claim about what Cumbric definitively was; it is the best available operational approximation of what Cumbric may have been, produced by documented and auditable methods. The confidence grades attached to each entry make this clear. The purpose is not to deceive learners into thinking they are speaking historical Cumbric; it is to provide the most intellectually honest and practically usable approximation that current scholarship can support.
The CDRS rests on a philosophical commitment that can be stated as a single proposition: language is a structured, recoverable system capable of restoration into operational communicative use through deterministic diachronic transformation. This proposition has four components, each of which carries both theoretical content and practical implications for how the CDRS is designed and operated.
Language as structured system. The CDRS treats language as a rule-governed system with describable phonological, morphological, and syntactic regularities. This is the standard assumption of modern linguistics. Without it, there would be no basis for diachronic transformation rules, and the CDRS would be impossible. The historical reality of language change — that languages change in regular, patterned ways, not randomly — is the empirical foundation on which the entire enterprise rests.
Linguistic discontinuity does not imply irrecoverability. A language that is no longer spoken is not necessarily beyond recovery. The history of scholarship on Proto-Indo-European, Gothic, Classical Hebrew, and numerous extinct languages demonstrates that the absence of living speakers is not an obstacle to productive scholarly engagement with a language system. The evidence may be fragmentary, mediated through other languages, or preserved only in proper names — but fragmentary evidence, evaluated carefully, can support principled inference about the full system from which it derives.
Partial evidence can support full functional systems. The CDRS does not require complete evidence for every form. It requires sufficient evidence to support a structured, graded, operationally functional output. This is not a compromise; it is a methodologically sound position supported by the history of successful revitalisation. Hebrew was revitalised from Biblical and Mishnaic sources alongside scholarly reconstruction, and is now the first language of millions of speakers [1]. Cornish was revitalised from manuscript sources by Henry Jenner and subsequent scholars [5] and is now taught in schools and spoken by a committed community. Māori revitalisation drew on the partial documentation of a living but severely endangered language to produce a thriving community of new speakers [3]. All three precedents demonstrate that operational revitalisation from imperfect evidence is academically established practice.
Continuity is achieved structurally, not only historically. A CDRS instantiation does not claim that the revitalised language was continuously spoken or transmitted without interruption. It claims that the structural relationship between language stages — the rule-governed diachronic transformations documented by historical linguistics — is sufficient to establish a principled connection between the living input and the revitalised output. Structural continuity, documented and auditable, is the basis for revitalisation. The absence of unbroken transmission chains, while historically significant, does not preclude the intellectual and practical achievement of principled diachronic recovery.
The Brittonic Convergent Diachronic Revitalisation System (BCDRS) is the first and currently the only Model Class Instance of the CDRS in operation. It applies the abstract CDRS architecture to the Brittonic linguistic continuum, populating the formal structure with the specific phonological, morphological, and historical evidence of the Welsh–Cornish–Breton–Cumbric family. The BCDRS inherits all five CDRS system properties and applies them through three concrete transformation frameworks.
The pipeline runs from Modern Welsh — a living, fully documented language — through Revitalised Middle Welsh (produced by the MWRF from the Modern Welsh input alongside the attested Middle Welsh corpus), through Revitalised Old Welsh (produced by the OWRF from the Revitalised Middle Welsh alongside the sparse attested Old Welsh corpus), to Revitalised Cumbric (produced by the NBTRF from the Revitalised Old Welsh alongside the northern Brittonic place-name record and the yan-tan-tethera counting tradition). The result is a 307-entry dataset covering 24 lexical and grammatical domains, maintained in the Polyglot™ system and available at polyglot.kingarthursroundtable.com.
The full specification of the BCDRS — its strata, dependency function, framework specifications, practical communicative implementation, and academic governance — is documented at bcdrs.html.
The CDRS is a formal model, and like all formal models, its reliability is bounded by the quality and completeness of its inputs. The following constraints apply to all CDRS implementations and should be understood by any user of CDRS outputs:
These constraints are not defects of the CDRS; they are honest reflections of the epistemic situation that any revitalisation project faces. The CDRS's contribution is to make these constraints explicit, systematic, and subject to governance — rather than leaving them implicit and ungoverned.
The CDRS is designed as an open and extensible framework. Its language-independence means that the same formal structure can be applied to any linguistic continuum with sufficient historical documentation, and the model class concept explicitly anticipates multiple instantiations. Several directions for future development are envisaged:
Additional CDRS instantiations. The most immediate extension would be the development of a Goidelic CDRS instance, applying the same formal architecture to the Irish–Scottish Gaelic–Manx continuum. A Germanic instance (addressing, for example, Old English and its descendants) or a Romance instance (addressing, for example, Latin and its descendant vernaculars) would follow the same principles. Each instantiation would require its own domain-specific frameworks and evidential bases, but would inherit the abstract CDRS architecture in full.
Computational implementation. The deterministic, rule-governed character of the CDRS makes it well suited to computational implementation. A software system that applies CDRS transformation rules automatically — producing output entries from source entries according to the parameterised rules, flagging cascade obligations when inputs change, and maintaining version control across the pipeline — would significantly enhance the scalability and auditability of the system. The formal structure of the CDRS is designed with this possibility in mind.
Relationship to formal grammar frameworks. The CDRS's formal structure invites comparison with established frameworks in theoretical linguistics, in particular Optimality Theory [8], which similarly provides an abstract constraint-satisfaction architecture instantiated by domain-specific rankings. A formal comparison of the CDRS with OT and other constraint-based frameworks could illuminate both the strengths of the CDRS and areas where its formal specification could be deepened. PBC welcomes scholarly engagement with this question.
Invitation to scholarly engagement. The CDRS is explicitly designed for scholarly engagement. Its peer-review pathway is not merely an administrative mechanism; it is an invitation to the academic community to contribute to the progressive refinement of the model and its instantiations. Proposals from qualified specialists in historical linguistics, Celtic studies, formal grammar, and allied fields are welcomed via www.penrithbeacon.com.