Semantics of Accountability

As I explained in another blog post, accounting is the universal technology of accountability. There is no "open world assumption". The essence of accounting is that it is a deliberately and consciously created to be a deterministic system. Accounting has solid boundaries. Accounting is a closed world.

People need to be aware of something.  Machines such as computers do not "understand" in the sense that humans understand.  Saying this another way, the notion of “understanding” in the human sense cannot be attributed to machines.  Computers are machines; they are not people.

Meaning is mediated, situated, and carried by human collectives.  Meaning is agreed to by human collectives. Meaning is an agreement; meaning is intersubjective. Meaning is the mutual cognitive agreed upon understanding by the members of some specific group.

Understanding is individual, dynamic, and a purely human characteristic. Machines simply have no capability to actually understand what they are doing.

A conceptual failure is where you think you have mutual cognitive agreement but you actually do not.  Conceptual failures are very dangerous because they don’t look like failures at first.

Machines can process meaning but cannot understand in the human sense what they are processing. As I have pointed out before; computers have certain specific capabilities and there are certain specific obstacles that must be overcome to use those capabilities reliably and effectively.

And so, to be crystal clear. Machines do not understand. Machines interpret. Machines manipulate representations. Machines model correlations. Machines simulate reasoning.  Machines do not grasp a subject. Machines don't set intention. Machines cannot be held responsible for making a mistake. Machines are not accountable.  Humans are accountable.

Meaning lives between people; understanding lives within people. Machines are not “people.” Machines do not “understand.”

Accounting is the world’s oldest, most mature system for turning human meaning and intention into verifiable, governed, explainable, provenance‑rich, error‑detecting action.  Accounting is the prototype of what modern artificial intelligence is trying to achieve through "purpose" and "alignment" and "governance". Accounting has been doing this successfully for 7,000 years.

  • Accounting is alignment.
  • Accounting is explainability.
  • Accounting is provenance.
  • Accounting is governance.
  • Accounting is accountability.

Fundamental to accountability is auditability. Provenance is the mechanism which provides auditability which leads to accountability.

Accountability: For every consequential outcome within a system, there exists a traceable chain of responsibility linking that outcome to one or more accountable agents.

Provenance is the traceability layer that makes the purpose clear, intention verifiable, alignment auditable, explainability trustworthy, governance enforceable, and therefore trust justifiable. Justifiable because it is provable. Provenance is the record that ensures what was meant, what was interpreted, what was justified, and what was authorized all match. In a court of law.

Meaning is the shared semantic structure. As pointed out by Nicholas Figay in his book Inhabiting Babel, "Meaning is always mediated, situated, and carried by human collectives."

Understanding is the ability to activate that structure, to take action.

Together, meaning and understanding are the foundation upon which intention, alignment, explainability, provenance, and governance are built. To summarize the semantics of accountability:

  • Meaning: Epistemic commitment; meaning enables understanding; meaning is mediated, situated, shared semantic content; intersubjective, mediated, contextual, exists within a collective.
  • Understanding: The skills, experience, human judgement to effectively to work with meaning as an individual system be that system a human, a machine, or a human and machine working together; individual, dynamic, competence; based on acquired skills and experience; understanding validates meaning.
  • Intention: Only humans have legitimate intention; intention means "What was meant?” within a system of governance.
  • Approval: Approval is a decision event in which an accountable agent, human or machine, applies criteria to a proposed action or artifact and produces an authorization outcome that changes its permissible state.
  • Provenance: semantic chain‑of‑custody; “How did meaning travel through the system?”; unbroken evidence chain.
  • Traceability: Traceability/trackability relates to identification of information, digitization of information, connecting that information into an unbroken chain to enable traceability and trackability, the actual tracing and tracking itself, and accounting/audit related aspects of traceability and trackability.
  • Explainability: Epistemic justification; epistemic transparency; “Why did the system act this way?”
  • Alignment: Semantic correctness; “Did the system act according to legitimate intention as specified within the system of governance?”
  • Governance: Authoritative validation; “Was the action authorized, constrained, and therefor trustworthy?”
  • Claim: A claim is a proposition, assertion, or statement that a person, organization, or agent presents as being true. A claim may be true, false, or uncertain until evaluated. Assertion about the state of the world.
  • Evidence: Evidence is an artifact or observation used to justify belief in a claim; making that claim a fact.
  • Fact: A fact (a.k.a. operational fact) is a claim (proposition, assertion, or statement) that corresponds to reality and has been verified using evidence or accepted as true. State of the world. A fact is a claim that has achieved a sufficient level of assurance through the evaluation of evidence. Fact is the claim that society, an auditor, a regulator, or a decision-maker accepts on the basis of that confidence.
  • Reasonable Assurance: Reasonable assurance is the level of confidence that results from applying judgment to available evidence.
  • Judgement: Judgment is the process of evaluating whether the evidence is sufficient, relevant, and reliable to support a claim. Judgment is about whether the evidence is sufficient to treat the claims as facts.
  • Assurance: Evidence creates the possibility of belief. Judgment converts evidence into confidence. Assurance measures that confidence. Reasonable assurance is having enough confidence to consider a claim as being a fact.
  • Element: An element is a type of thing that can exist in a financial statement; per the conceptual framework and reporting framework being used. Elements are used to define totals, subtotals, and line items.
  • Line Item: A line item is an actual thing that exists in a financial statement.
  • Classification: Classification relates to the inclusion of a thing in a set. Classification is asserting that a line item is a specific type of element or basic level category, superordinate level category, or subordinate level category.
  • Event: An event (a.k.a. business event) that affects an enterprise and is relevant to its operations, finances, obligations, rights, risks, or performance.

Meaning enables understanding. Understanding validates meaning. Without stable meaning, understanding collapses. Without understanding, meaning is inert. Provenance is the bridge that connects meaning, understanding, intention, explainability, alignment, and governance.

Using the metaphor of constructing a house.

  • Intention is the architect’s design.
  • Explainability is the engineering diagrams that deliver the intended architecture.
  • Governance is the permitting office's building code which approves the engineering diagrams.
  • Alignment is the construction contractor building the house consistent with the architects design, adhering to the engineering diagrams, and within the intention of the building code.
  • Provenance is the binder containing every blueprint revision, inspection note, and authorization signature; the audit evidence graph.

Without provenance, the building cannot be certified as safe. Without the physical chain of evidence, an engineer, architect, builder, or permitting office cannot be sued in a court of law to remediate issues related to the endeavor of constructing a house. 

Accounting is the only system in human history that: starts with meaning, requires human understanding, expresses intention, structures interpretation, enforces alignment, guarantees explainability, preserves provenance, operationalizes governance, produces accountability.

Accounting is a complete stack.

Accounting is a prototype or "blueprint" for artificial intelligence alignment.

Accounting is the prototype for responsible meaning engineering.

Accounting is the civilization‑scale solution to effective management of epistemic risk.

Provenance is the chain‑of‑custody of meaning. Accounting has had this for millennia:

  • journals
  • ledgers
  • audit trails
  • supporting documents
  • cross‑checks
  • reconciliations

Accounting is a logical system for telling a story about economic reality. A story is provenance. Every transaction is a provenance atom. Every ledger is a provenance chain. Every financial statement is a provenance graph.

Accounting is governance made operational. Governance means:

  • authority (e.g. FASB, IASB, SEC, GASB, and so forth)
  • constraints (e.g. reporting frameworks)
  • validation (e.g. accounting equation)
  • oversight (e.g. SEC, ESMA, other regulators)
  • trust (e.g. independent auditors, internal auditors)

Accounting is the world’s oldest alignment system.  It aligns human intention with structured representation, enforces explainability, preserves provenance, and embeds governance; producing accountability. Everything we are trying to build for artificial intelligence already exists in accounting. Accounting is the universal technology of accountability because it is the universal technology of meaning, intention, alignment, explainability, provenance, and governance.

If we mistakenly treat machine interpretation as machine understanding, we risk:

  • over‑trusting the system
  • misattributing agency
  • assuming semantic competence where it does not exist
  • granting unauthorized authority
  • collapsing the human–machine boundary
  • mistakes

Alignment depends on never confusing interpretation with understanding. Understanding is human. Machines only interpret. So the alignment stack should be read like this:

  • Meaning (human, shared)
  • Understanding (human, individual)
  • Interpretation (machine, computational)
  • Intention (human, purposive)
  • Alignment (matching machine interpretation to human intention)
  • Explainability (revealing machine interpretation)
  • Provenance (recording machine interpretation)
  • Governance (ensuring human intention is respected)

Understanding is not a machine capability. Understanding is the human anchor that keeps the entire alignment system grounded.

Not only do I see accounting as a prototype for harnessing the power and opportunities provided by artificial intelligence reliably, effectively, and safely; I also see accountants, auditors, and analysts as the custodians and stewards of knowledge of the enterprise.  Why? The accounting information system is commonly the most trusted information system in an enterprise.

Accountability is a conservation law.  Accountability says: "Every outcome must be attributable to some agent, decision, process, or event." Nothing happens organizationally without ownership somewhere in the system. Conceptually: Responsibility Assigned = Responsibility Traceable. Accounting preserves the traceability of value. Accountability preserves the traceability of agency. You could formalize accountability as: Every state transition must have an attributable actor.

Additional Information:

Comments

Post a Comment

Popular posts from this blog

Reference Reporting Frameworks

Image
A financial reporting framework is a purposeful arrangement of financial line items, totals, subtotals, aggregations, disaggregations, structures, metadata, categorization, classification, assertions, restrictions, constraints, other conditions, etc. A financial reporting framework represented in a digital format should not be semantically impoverished. Leaving information out, consciously or unconsciously, is not desirable. These reference reporting frameworks provide the minimum of what is necessary. They are modest yet not impoverished. A financial reporting framework provides the elements necessary for reporting organized by basic line item element, superordinate elements, and subordinate elements organized by type per the process of classification .  Financial reporting frameworks tend to start with the high level concepts of the basic accounting equation: assets, liabilities, equity; and then build out from that basic accounting equation. The following are a set of reference ...
Read more

Accounting & Audit by Design (A&AD) Framework

Image
The Accounting & Audit by Design (A&AD) Framework is about freedom from drudgery for accountants, auditors, and analysts. It is about streamlining processes. A&AD is about eliminating the waste of time and energy that goes into transaction chasing . It is about eliminating the need to resort to things like " the plug ".  A&AD is about fixing processes that allow mistakes rather than spending time fixing mistakes a process allowed or dealing with the consequences of a mistake escaping from the system undetected. The following sections will help you understand what the Accounting & Audit by Design (A&AD) Framework is and what it enables. Environment What is A&AD? Why A&AD Now? How does A&AD Work? Business Outcome Fifty to seventy five year ago it would have been impossible to contemplate something like A&AD; it just was not feasible. We did the best we could to implement accounting information systems using computers and software that we...
Read more

Overview

Image
( Executive summary ) Innovation doesn’t arrive. Innovation evolves. Imagine it.  Freedom from drudgery. Freedom from the toil of tedious, repetitive, mundane, monotonous tasks. Far less time needs to be spent fulfilling the role of "data janitor".  Time spent "transaction chasing" is significantly reduced. Accounting is a well understood system which has evolved over many thousands of years. Accounting, reporting, auditing, and analysis is being digitized. This digitization will enable processes to be industrialized  and productivity boosts to be realized. If you are an accountant, auditor, or analyst it is time to update your tradecraft. This overview is an attempt to take the work that I have done to understand XBRL-based reports and back into a framework for explaining those reports, accounting and audit working papers, analysis models to accountants and software engineers building software for accountants.  I am now referring to this as model-driven reporting ...
Read more