Towards Semantics and an Algebra of Governance

Just as accounting has formal semantics (assets, liabilities, equity, comprehensive income, investment by owner, distribution to owner, revenues, expenses, gains, losses) and an algebra (assets = liabilities + equity; beginning equity + comprehensive income + investment by owners - distribution to owners); governance can be viewed as having its own semantics and algebra.

This same perspective is emerging in fields such as policy-as-code, computational law, enterprise modeling, defensible compliance, and governance frameworks.

Semantic Model of Governance

A semantic model identifies the fundamental concepts of a domain and the meaning of those fundamental concepts. For governance, the core question is: How do rights, obligations, authority, and accountability exist, change, and interact within an enterprise?

Rather than viewing governance as a collection of isolated rules, a semantic approach views governance as a structured state of affairs consisting of:

  • Knowledge (what is known; transparency into rules, ability to query)
  • Actors (who participates; people, organizations, agents)
  • Resources (what is governed; things that can be controlled, owned, or used)
  • Rights (what actors are permitted to claim or do)
  • Obligations (what actors are required to do)
  • Authority (who has the power to establish, create, modify, or enforce rules)
  • Constraint (what is prohibited, limited, or otherwise restricted)
  • Accountability (who is responsible)
  • Event (what changes the state)
  • Governance (the system of constraints and accountability that gives rules legitimacy)

A particularly interesting approach to organizing the semantics of governance is KROG. KROG explicitly centers governance around the relationship between four key pillars:

  • Knowledge (K): Transparency and query-ability of all rules, states, and relationships; no hidden rules, no surprises; if it affects you then you can see it.
  • Rights (R): permissions; defines what actions are permitted; clear boundaries of allowed behavior
  • Obligations (O): duties; defines what actions are required; deadlines, compliance requirements, mandatory disclosures, noncompliance with obligation defines specific consequences
  • Governance (G): types and phases of rules; defines how rules themselves are created, modified, enforced; the rules about the rules

KROG proposes governance as the framework that constrains and coordinates the others: 

(K ∩ R ∩ O) ⊆ G (Every valid action satisfies Knowledge (K) intersection Rights (R), and intersection Obligations (O); and remains contained within Governance (G).

Algebra of Governance

What is a Governance Algebra? An algebra specifies the specific operations that can be performed on those concepts and the laws that govern their behavior in a specific domain. In governance, operations are things like:

  • Grant Right
  • Revoke Right
  • Delegate Authority
  • Create Obligation
  • Fulfill Obligation
  • Transfer Ownership
  • Enforce Policy
  • Appeal Decision

Governance as a State Space

A governance system can be viewed as maintaining an institutional state. Events change or transform state. The role of the rule engine is not merely to infer facts, but to maintain the integrity of this governance state. A financial statement reports state and changes in state.

Verification is different from verifying ordinary logical state such as if  some thing is true or false. Where as in traditional logic one might ask, "Is statement X true?" While governance does ask the traditional logic type questions; governance goes further and asks additional questions such as:

  • Who may act?
  • Who must act?
  • Who is accountable?
  • What happens if they do not?

This introduces concepts that ordinary predicate logic does not naturally capture such as:

  • Permissions
  • Prohibitions
  • Duties
  • Delegation
  • Accountability
  • Enforcement
  • Temporal obligations

These are often called normative semantics because they describe what ought to occur rather than merely what is true. For example, "Do not enter" is normative semantics; how the world ought to be; as contrast to the fact that you have, or have not, actually entered.

The interesting question is not: Is "Do not enter" true? but rather: Is it in force? Does it apply here? Who issued it? Was it violated? Those are governance questions.

Very succinctly: Facts tell us what is. Norms tell us what ought to be. Governance transforms facts into institutional rights and obligations. That transformation is the heart of normative semantics.

One of the most intriguing observations is that accounting appears to be a specialized form of governance semantics.  Accounting and governance appear to be tightly related.

Governance as a Pattern

Governance seems to be a reusable architectural or design pattern.  The thing that is governed changes (corporation, contract, project, data, finances), but the underlying mechanisms of authority, accountability, measurement, control, and enforcement remain consistent.

In ontology terms, the term "Governance" is a superclass  with different subclasses or types like "Corporate Governance", "Contract Governance", "Data Governance", "Financial Governance", and "Project Governance" as specializations that inherit the same core governance properties while applying them to different governed entities.

Governance is not limited to organizations. Governance can be applied to any important artifact, asset, relationship, or process that has stakeholders, accountability, risk, and state changes. That is a powerful abstraction because it means governance becomes a reusable pattern.

Every governed object seems to have: Governed Object + Identity + Roles + Risks + Lifecycle + Accountability + Controls + Metrics + Stakeholders + Audit Trail.

Governance Lifecycle

The thing that is being governed has a lifecycle. Governance is not merely a hierarchy of authority. It is the system that manages a governed object's lifecycle by assigning accountability for state transitions.

In that sense then; contract governance and corporate governance are instances of the same underlying governance pattern. The corporation is the governed object in one case; the contract is the governed object in the other. Both objects have owners, stakeholders, controls, policies, metrics, risks, audits, and lifecycle states. The object changes, but the governance architecture/design remains similar.

Guaranteed Safe Artificial Intelligence

The paper, Algebraic Semantics of Governed Execution: Monoidal Categories, Effect Algebras, and Coterminous Boundaries, proposes a mechanized formalization for guaranteed safe artificial intelligence based on a proposed framework for guaranteed safe AI comprising a world model, a safety specification, and a verifier. (page 16).  The paper goes on to conclude, "...capabilities such that the governed fragment coincides with the expressible fragment in the modeled substrate." This is exactly what is stated by the Law of Requisite Variety and my notion of "complete" (borrowed from Maria Keet, PhD) where reported facts and rules are matched.

Completeness

Completeness (a.k.a. fitness of purpose, pragmatic quality) is determined by the level of precision (high or low) and the level of coverage (high or low) relative to the specific goal or objective or range of goals/objective that a community of stakeholders has agreed to with respect to some system of interest. Completeness must match the needs.

Completeness is declared by the rules of governance.

Governance is a Second Order Conservation Law

A conservation law is an invariant principle that remains true across all legitimate transformations of a system and thereby preserves the system's coherence, traceability, and identity.

Accounting is the conservation law of value.  Accountability is the conservation law of traceable responsibility.  Governance, because it relies on accounting and accountability, is a second order conservation law of legitimate agency: ensuring that every power to act remains coupled to responsibility for outcomes.

Governance might be viewed as meta-accounting. Governance is the system that keeps the books not on money, but on the allocation, delegation, and justification of decision-making power itself. In the same way that financial accounting produces a balance sheet of resources and claims, governance produces a balance sheet of authority and accountability.

Governance Assets (Power to act):

  • decision rights
  • spending authority
  • hiring authority
  • policy authority
  • ownership rights

Governance Liabilities (Obligations to answer):

  • reporting duties
  • fiduciary duties
  • compliance obligations
  • performance accountability
  • legal responsibility

Then the governance invariant equation becomes:

  • Authority = Accountability or
  • Power = Responsibility

Determining Quality and Responsibility

There are many different definitions of quality. "Quality is conformance to a requirement." "Quality is fitness for purpose."  But if you think about those definitions, you will notice something interesting.  Each contains the term "quality".  One term uses "conformance" the other uses the term "fitness". One uses the term "requirement" the other uses the term "purpose".  Now, if you think about the notion of "meaning" and you think about those two definitions of quality, you will notice:

  • Who set the requirement or decided what the purpose is?  That is the role of governance.
  • Who decides if something is conforming or if something is actually fit? That is governance.
As I said above, "the governed fragment coincides with the expressible fragment in the modeled substrate".  If information is provided and you don't have a specification or stated purpose; how can you know if that information conforms to the requirement or satisfies the purpose?  If that information is not in conformance or is not fit; then who becomes responsible (e.g. who do you blame) if you don't have governance? This is part of accountability.

A Possible Invariant Equation of Governance

Accounting achieved remarkable success because it discovered a simple invariant conservation law referred to as the accounting equation:

  • Assets = Liabilities + Equity

Accounting preserves the traceability of value.  Accountability preserves the traceability of agency. Every state transition, caused by an event, must have an attributable actor. When attribution is lost, accountability is violated just as surely as a broken ledger violates accounting.

  • Agent >> Event >> Outcome or maybe
  • Actions ↔ Responsible agents

Might governance also ultimately have an analogous principle? It seems to me that there is a correspondence between and expressed fragment of information and the governance of that same information fragment per the notion of requisite variety and completeness.  Here are some possibilities:

  • Authority >> creates >> Rights and Obligations
  • Governance >> constrains >> Authority
  • Accountability >> links actions  >>  to consequences
  • Rights + Obligations = Governed Knowledge
  • Rights + Obligations = Governance(Knowledge)
  • (K ∩ R ∩ O) ⊆ G (Every valid action satisfies Knowledge (K) intersection Rights (R), and intersection Obligations (O); and remains contained within Governance (G).
  • Knowledge >> records  >> facts 
  • Governance >> interprets  >> facts
  • Rights >> represent >> enterprise claims
  • Obligations >> represent >> enterprise duties

Basically, the key idea is: Rights and/or obligations do not appear from nowhere. They must be grounded in: knowledge of facts, an authority structure, and a governance framework.

Just as assets and liabilities are created together in accounting, rights and obligations are often co-created by governance acting on knowledge.  It seems to be the case that accounting and governance share the notion of "duality".

Whether such a law can be expressed formally remains an open question. Frameworks like KROG are interesting because they suggest that governance may not merely be a set of rules, but a structured semantic system whose states, transformations, and invariants can be studied algebraically, much as accounting studies financial states and transactions algebraically. In that sense, a governance algebra is the study of how rights, obligations, authority, and accountability compose and transform, while a governance semantics is the study of what those concepts mean and how they define the state of an enterprise.

So, maybe I am seeing this wrong.  Maybe there is less of a link between accounting and governance; maybe the link is more about accountability and governance.

Applying Governance to a Financial Disclosure

Consider the following. If "corporate governance" exists and "contract governance" exists; it seems to me that "financial disclosure governance" might also exist.  From an accountability-framework perspective, a financial disclosure is arguably even closer to a contract than a corporation is. Both are discrete artifacts that move through states, accumulate approvals, carry obligations, are subject to audit, can be amended, and eventually become historical records. The corporation is the enduring entity; the contract and the disclosure are governed artifacts produced and managed within that entity.

A financial disclosure has a lifecycle. Who owns the disclosure? Who prepares it? Who reviews it? 
Who approves it? Who can change it? Who is liable for errors? Who monitors compliance to reporting rules the disclosure satisfies?

Financial reporting frameworks are actually governance systems disguised as accounting standards. They are governance frameworks that regulate the production of financial disclosures. IFRS and US GAAP can legitimately be viewed as specialized forms of governance frameworks for accountability artifacts referred to as financial disclosures.

Additional Information:

Comments

Popular posts from this blog

Internationalized Resource Identifier (IRI)

Example Financial Statement Holon

Inhabiting Babel, A Manifesto for Responsible Meaning Engineering