RSM Subject Model v1.0

#1. Purpose and Scope

#1.1 Purpose

The Regenerative Systems Map represents an ecology containing people, organizations, communities, ecological systems, places, facilities, materials, instruments, digital agents, and temporary formations. These entities differ fundamentally in identity, agency, lifecycle, authority, and the ways in which they can participate in regenerative systems.

At the same time, each may become meaningful within an observation, relationship, evaluation, activity, claim, commitment, outcome, or historical account. RSM therefore requires a subject model broad enough to represent heterogeneous reality without making every meaningful thing an Actor, Participant, or specialized canonical class.

The RSM Subject Model establishes that foundation. It defines Subject as a semantic role, specifies a deliberately small family of foundational identity-bearing kinds, establishes extensible domain typing through the governed RSM vocabulary, and defines the rules by which entities participate contextually as Subjects, Actors, Participants, Stewards, Observers, Committers, Resources, Beneficiaries, and other roles.

#1.2 Design Objective

The Subject Model is designed to keep the RSM kernel semantically small while allowing the represented ecology to become arbitrarily rich. Concepts such as farmer, buyer, processor, laboratory, watershed, soil system, grain lot, certification, transition loan, or conservation easement should not require changes to the canonical kernel merely because the regenerative system expands into another domain.

The resulting architecture separates four concerns that conventional models frequently collapse:

1. Identity establishes which enduring thing is being referenced.
2. Foundational Kind establishes what fundamentally different class of thing it is.
3. Domain Type establishes what that entity means within one or more subject-matter vocabularies.
4. Contextual Role establishes how that entity is participating in a particular relationship, activity, formation, observation, or outcome.

This separation allows RSM to represent ecological, institutional, economic, material, and computational reality without turning its core schema into an exhaustive ontology of the world.

#1.3 Governing Principle

The governing principle of the Subject Model is:

Subject is a semantic role, not an ontological superclass.

An entity becomes a Subject when it is the focus, target, bearer, affected entity, or feature of interest within an RSM expression. Subjecthood says nothing by itself about whether that entity can act, consent, commit, own something, represent another entity, or possess authority.

This distinction is foundational. A watershed can be a Subject of Observation and Outcome without being an Actor, while a Person can simultaneously be a Subject of evaluation, an Actor performing an Activity, a Participant in a Formation, and a Steward of a Community.

#2. Core Semantic Architecture

#2.1 Identity, Kind, Domain Type, and Role

RSM SHALL distinguish identity, foundational kind, domain type, and contextual role. These concepts answer different questions and SHALL NOT be collapsed into a single classification field.

The four questions are:

For example, a soil-testing company may be represented as:

Identity:
https://id.rsm.org/organization/01KXYZ

Foundational Kind:
rsm:Organization

Domain Types:
rsm:agriculture/soil-testing-provider
rsm:food/testing-laboratory

Contextual Roles:
rsm:role/observer
rsm:role/evidence-provider
rsm:role/participant

The organization does not cease to be an Organization when its role changes. Likewise, its classification as a laboratory does not require the creation of a new foundational Laboratory class in the RSM kernel.

#2.2 Semantic Layering

The architecture can be represented as three semantic layers above durable identity:

Foundational kinds are intentionally stable. Domain types are governed but extensible, while contextual roles are dynamic, contextual, and frequently temporal.

#2.3 Open-World Domain Semantics

RSM SHALL operate as an open-world semantic system. The kernel does not need prior knowledge of every possible domain concept in order to preserve identity, relationships, provenance, observations, outcomes, and other canonical semantics.

A newly introduced concept such as rsm:agriculture/anaerobic-digester-feedstock can therefore enter a governed domain vocabulary without requiring a new kernel release. Applications that understand the concept may use its richer meaning, while conformant applications that do not understand it can still preserve and exchange the term without semantic corruption.

#3. Subject as a Semantic Role

#3.1 Definition

A Subject is any addressable entity that is the focus, target, bearer, feature of interest, or affected entity of a meaningful RSM expression. The Subject role may be applied wherever a canonical or derived statement concerns something identifiable.

An entity may become the Subject of:

• a Description;
• an Observation;
• a Claim;
• Evidence;
• an Evaluation;
• an Assertion;
• a Relationship;
• an Activity;
• a Policy;
• an Outcome;
• a Biography; or
• another governed semantic expression.

Subjecthood does not create a new object or change the entity's foundational kind. It expresses the entity's semantic position within a particular statement or context.

#3.2 Subject Is Not Actor

RSM SHALL preserve the invariant:

Subject ≠ Actor

A soil system may be observed, restored, degraded, evaluated, or affected without possessing the ability to make a Commitment. A DigitalAgent may perform an Evaluation and simultaneously become the Subject of an audit concerning that evaluation.

The Subject role therefore describes what an expression is about. Actor describes the entity responsible for performing an action.

#3.3 Subject Is Not Participant

RSM SHALL also preserve:

Subject ≠ Participant

Participation is contextual. An Organization does not become a Participant merely because it exists in the network, and a watershed does not become a Participant merely because a Formation is intended to improve its condition.

A Participant is an entity participating within a bounded context such as a Formation, Community, Activity, or other governed arrangement. The same entity may participate in many contexts while retaining one durable identity.

#3.4 Subject Is Not a Stored Superclass

Implementations SHALL NOT require every canonical entity to inherit from, contain, or materialize a universal Subject superclass merely to participate in RSM semantics. Subject references SHOULD be expressed through canonical identity references.

For example:

observation:
  subjectRef: https://id.rsm.org/living-system/01KABC

and:

outcome:
  affectedSubjects:
    - https://id.rsm.org/living-system/01KABC

refer to the same durable entity in different contexts. No separate Subject object is required.

#4. Foundational Entity Kinds

#4.1 Foundational-Kind Principle

RSM SHALL maintain a deliberately small set of foundational entity kinds. A new foundational kind is justified only when it introduces materially different identity, lifecycle, agency, provenance, or structural semantics that cannot be represented adequately through domain typing and contextual roles.

The Subject Model defines ten foundational subject-capable kinds:

1. rsm:Person
2. rsm:Organization
3. rsm:Community
4. rsm:LivingSystem
5. rsm:Place
6. rsm:Facility
7. rsm:Material
8. rsm:Instrument
9. rsm:DigitalAgent
10. rsm:Formation

These kinds establish the stable semantic skeleton. Domain vocabularies provide specialization without expanding the kernel indiscriminately.

#4.2 Person

A Person represents a human individual whose identity persists independently of employer, organization, account, community, affiliation, or current role. A Person may possess capabilities, relationships, affiliations, authority, commitments, observations, activities, and biography according to applicable policy.

Examples such as farmer, agronomist, scientist, physician, consultant, buyer, investor, regulator, or community steward SHALL normally be represented as domain concepts, capabilities, affiliations, occupations, or contextual roles. They SHALL NOT require separate foundational Person subclasses in the RSM kernel.

Illustratively:

"@type": "rsm:Person"
"rsm:domainType":
  - "rsm:agriculture/farmer"
  - "rsm:agriculture/agronomist"

A Person is intrinsically capable of human agency. The ability to perform a particular consequential action nevertheless remains dependent upon authority, policy, consent, and context.

#4.3 Organization

An Organization represents a durable collective or institutional entity capable of possessing identity, governance, capabilities, relationships, assets, authority structures, commitments, and biography. Organizations may include commercial, nonprofit, governmental, academic, cooperative, civic, or other institutional forms.

Examples include farm enterprises, processors, laboratories, certification bodies, universities, conservation districts, retailers, banks, insurers, government agencies, and data providers. These distinctions SHALL normally be expressed through domain vocabulary rather than separate foundational kinds.

An Organization MAY hold multiple domain types simultaneously:

"@type": "rsm:Organization"
"rsm:domainType":
  - "rsm:agriculture/producer"
  - "rsm:food/processor"
  - "rsm:food/aggregator"

RSM SHALL NOT force an Organization into a permanent marketplace identity such as buyer, seller, producer, or supplier. Such meanings may arise from domain type, Capability, Offer, Seek, or contextual role depending upon what is actually being represented.

#4.4 Community

A Community represents a durable structure of belonging, association, shared context, mutual interest, or collective identity whose existence does not depend upon one transaction or Formation. Communities may have membership, stewardship, purpose, visibility rules, relationships, and Biography.

Examples include producer networks, watershed communities, learning communities, professional associations, regional food-system groups, research cohorts, and other place-based or interest-based collectives. A Community is not automatically an Organization, although a Community may create, be sponsored by, overlap with, or participate alongside Organizations.

Membership SHALL NOT imply authority. Stewardship, delegation, representation, membership, and affiliation remain semantically distinct.

#4.5 LivingSystem

A LivingSystem represents a biological or ecological system capable of possessing state, undergoing change, participating in ecological relationships, and experiencing consequences over time. LivingSystem allows ecological reality to exist directly inside RSM rather than appearing merely as metadata attached to human activity.

Examples include:

• soil systems;
• plants;
• crop populations;
• herds;
• microbial communities;
• agroecosystems;
• forests;
• wetlands;
• aquatic ecosystems;
• pollinator populations;
• habitat systems; and
• human biological systems where the modeled context requires them.

A LivingSystem may be observed, measured, affected, restored, degraded, connected, monitored, evaluated, or otherwise participate in system dynamics. These verbs do not imply contractual or institutional agency.

Illustratively:

"@type": "rsm:LivingSystem"
"rsm:domainType":
  - "rsm:agriculture/soil-system"
  - "rsm:agriculture/managed-agroecosystem"

This distinction is essential to the regenerative character of RSM. Soil, water, habitat, biodiversity, and other ecological systems become first-class subjects of system state and outcomes rather than external KPIs attached to transactions.

#4.6 Place

A Place represents spatially meaningful context. Places may correspond to points, parcels, polygons, fields, farms, watersheds, regions, foodsheds, jurisdictions, habitat boundaries, or other spatial constructs.

Place answers primarily:

Where does something exist, occur, apply, move, or belong?

LivingSystem answers a different question:

What living system exists, interacts, or changes there?

The distinction may be represented as:

Place:
Field 42

LivingSystem:
Field 42 Agroecosystem

Relationship:
Field 42 Agroecosystem located-at Field 42

Applications MAY project Place and LivingSystem together when doing so improves usability. The canonical semantic model SHALL retain the ability to distinguish them.

#4.7 Facility

A Facility is a bounded physical or operational system through which one or more capabilities can be exercised. A Facility may have location, equipment, certifications, operating characteristics, Capability, Capacity, constraints, observations, and Biography.

Examples include mills, processing plants, warehouses, cold-storage sites, laboratories, distribution centers, packhouses, composting facilities, irrigation infrastructure, and manufacturing sites. The Organization operating a Facility SHALL remain distinct from the Facility itself.

For example:

Organization:
Regional Milling Company

operates →

Facility:
Fresno Mill

Capability:
food-grade grain milling

Capacity:
120 tonnes/week

This distinction allows one Organization to operate multiple Facilities having different geographies, capabilities, available capacity, certifications, and constraints.

#4.8 Material

A Material represents physical matter whose identity, provenance, state, custody, properties, or transformation matters to the regenerative system. Material allows physical flows to be represented without turning RSM into a supply-chain-specific ontology.

Examples include seed lots, harvested crops, grain lots, ingredients, food products, animal feed, compost, fertilizer, biomass, soil samples, water samples, packaging, and other physical substances or lots.

Materials may participate in transformation chains:

Each transformation may create a new Material identity while preserving provenance and derived-from relationships to prior Materials. Observations, custody changes, certifications, claims, and evidence may attach to the appropriate Material identity.

#4.9 Instrument

An Instrument represents a durable formal, technical, legal, financial, policy, measurement, or representational construct that changes how the ecology can operate. Instruments may create rights, obligations, evidence, constraints, incentives, permissions, measurement interfaces, or transferable value.

Examples include:

• contracts;
• certifications;
• permits;
• environmental credits;
• financial instruments;
• insurance policies;
• grants;
• policy instruments;
• conservation easements;
• measurement instruments;
• standards; and
• protocols.

Domain types distinguish materially different Instruments:

rsm:finance/transition-loan
rsm:carbon/soil-carbon-credit
rsm:policy/conservation-easement
rsm:food/certification
rsm:measurement/soil-sensor

An Instrument SHALL NOT automatically be treated as an Actor. Where a technical Instrument such as a sensor can perform observations, the responsible Actor, software agent, device semantics, or provenance chain SHALL remain explicit.

#4.10 DigitalAgent

A DigitalAgent is a software entity capable of performing actions under defined computational, policy, and authority boundaries. DigitalAgent gives machine agency explicit identity rather than hiding machine behavior inside application services.

Examples include reasoning agents, coordination agents, monitoring agents, procurement agents, compliance agents, decision-support agents, and autonomous software services. DigitalAgent does not imply Personhood, principal status, or unrestricted authority.

The following distinctions are normative:

DigitalAgent ≠ Person
DigitalAgent ≠ Principal by default
DigitalAgent ≠ Authority
Authentication ≠ Authority
Recommendation ≠ Commitment
Inference ≠ Assertion

A DigitalAgent may perform operations only within applicable Capability and delegated Authority. A technically capable agent may therefore be allowed to discover, ask, evaluate, or propose while remaining prohibited from committing funds, entering contracts, disclosing protected information, or otherwise taking consequential action.

#4.11 Formation

A Formation is a bounded, identity-bearing arrangement of participants, commitments, capabilities, responsibilities, governance, and purpose that has actually come together. Formation represents emergent coordinated structure rather than merely a candidate combination of entities.

Examples include a regenerative grain transition coalition, watershed restoration coalition, procurement consortium, research collaboration, regional processing initiative, food-and-health intervention, or temporary investment consortium.

Once established, a Formation may itself become the Subject of observations, relationships, evaluations, activities, and outcomes. Where governance and authority support it, a Formation may also act through authorized Persons or DigitalAgents.

Formation SHALL remain distinct from Relationship and Configuration:

Relationship ≠ Formation
Configuration ≠ Formation

A Configuration represents a possible arrangement. A Formation represents an arrangement whose required commitments and governance have actually become sufficiently real to act.

#5. Concepts That Are Not Foundational Kinds

#5.1 Resource and Asset

Resource and Asset SHALL normally be contextual roles or domain concepts rather than foundational kinds. What constitutes a resource depends upon the purpose, constraints, and activity under consideration.

A tractor may be an Instrument and serve as a Resource in a Formation. Land may be represented as Place while functioning economically as an Asset, and available capital may be represented through financial Instruments while functioning as a Resource for a transition pathway.

The Subject Model therefore asks what something fundamentally is before describing how it is being used.

#5.2 Buyer, Seller, Producer, and Processor

Buyer, seller, producer, processor, aggregator, distributor, service provider, and similar commercial categories SHALL NOT be mandatory foundational kinds. They may be represented as domain types, capabilities, Offers, Seeks, or contextual roles depending upon the semantic requirement.

For example:

Organization:
Regional Grain Cooperative

Domain Types:
rsm:agriculture/producer
rsm:food/processor

Role in Formation A:
rsm:role/supplier

Role in Formation B:
rsm:role/buyer

This prevents marketplace assumptions from becoming permanent identity semantics.

#5.3 Laboratory and Certifier

A laboratory SHALL normally be represented as an Organization and, where appropriate, one or more Facilities. Its testing specialization, accreditation, methods, and services can be expressed through domain typing, Capability, Instrument, Evidence, and applicable relationships.

A certifier SHALL normally be represented as an Organization possessing relevant authority and Capability. The certification it issues may be represented as an Instrument with its own Subject, issuer, criteria, validity, provenance, and lifecycle.

#5.4 Program and Initiative

Program and initiative SHALL NOT automatically become foundational kinds. Their correct representation depends upon their actual semantics.

A government grant program may primarily be an Instrument administered by an Organization. A multi-party transition initiative may become a Formation, while an internal organizational initiative may be represented through Activity, Intent, policy, or an application-specific operational construct.

#6. RSM Semantic Namespace and Vocabulary Model

#6.1 One Governed Semantic Namespace

RSM SHALL use a single governed semantic namespace for core and domain vocabulary. Implementations SHALL NOT introduce independent prefixes such as rsm-agriculture:, rsm-food:, or rsm-water: merely to extend the RSM conceptual vocabulary.

The canonical compact prefix remains:

rsm:

Core classes and properties use stable RSM terms such as:

rsm:Person
rsm:Organization
rsm:LivingSystem
rsm:Capability
rsm:Capacity
rsm:Relationship
rsm:Observation
rsm:Formation
rsm:Outcome
rsm:domainType

Domain vocabulary is organized hierarchically inside the same RSM namespace:

rsm:agriculture/soil-system
rsm:food/ingredient
rsm:water/watershed
rsm:health/biomarker
rsm:finance/transition-loan
rsm:carbon/soil-carbon-credit
rsm:logistics/cold-storage
rsm:policy/conservation-easement

This produces one semantic system with many extensible domains rather than a collection of loosely related ontologies.

#6.2 Core Grammar and Domain Vocabulary

RSM SHALL distinguish core grammar from domain vocabulary.

Core grammar defines stable architectural concepts:

rsm:LivingSystem
rsm:Place
rsm:Material
rsm:Capability
rsm:Observation
rsm:Formation
rsm:Outcome

Domain vocabulary describes what those concepts mean in particular knowledge domains:

rsm:agriculture/soil-system
rsm:agriculture/crop-system
rsm:food/grain-lot
rsm:food/ingredient
rsm:water/aquifer
rsm:health/gut-microbiome

The distinction can be summarized as:

rsm:LivingSystem
    = foundational semantic grammar

rsm:agriculture/soil-system
    = governed extensible vocabulary

The domain term SHALL NOT silently become a second canonical class hierarchy unless a future normative specification explicitly promotes it.

#6.3 Domain Concepts

Domains themselves MAY be governed taxonomy concepts:

rsm:agriculture
rsm:food
rsm:water
rsm:health
rsm:finance
rsm:carbon
rsm:logistics
rsm:policy

More specific concepts MAY then participate in broader/narrower relationships:

rsm:food/grain-lot
    broader → rsm:food/material

rsm:agriculture/wheat
    broader → rsm:agriculture/grain

rsm:agriculture/soil-system
    broader → rsm:agriculture/living-system

The taxonomy mechanism SHALL remain governed and machine-readable. RSM Core SHALL NOT hard-code every domain term into programming-language enums.

#6.4 Cross-Domain Concepts

A concept or entity MAY participate in multiple domain vocabularies. RSM SHALL NOT force every entity into one mutually exclusive domain hierarchy.

For example, a grain lot may be described simultaneously as:

"@type": "rsm:Material"
"rsm:domainType":
  - "rsm:agriculture/grain"
  - "rsm:food/ingredient"

A conservation loan might be represented as:

"@type": "rsm:Instrument"
"rsm:domainType":
  - "rsm:finance/transition-loan"
  - "rsm:agriculture/regenerative-transition-finance"

Cross-domain typing reflects the actual nature of regenerative systems, where ecological, economic, food, policy, health, and infrastructure concerns routinely overlap.

#6.5 Naming Convention

Core RSM semantic classes SHOULD use PascalCase:

rsm:LivingSystem
rsm:DigitalAgent
rsm:Formation
rsm:Observation

Extensible taxonomy concepts SHOULD use lowercase kebab-case within hierarchical domain paths:

rsm:agriculture/soil-system
rsm:food/grain-lot
rsm:water/riparian-zone
rsm:health/gut-microbiome

The visual distinction makes canonical grammar recognizable while keeping taxonomy vocabulary human-readable and extensible.

#7. Schema Namespace and Entity Identity

#7.1 Vocabulary Identity and Object Identity Are Different

RSM SHALL distinguish semantic vocabulary IRIs from canonical entity identity IRIs. A schema term identifies a concept, while an instance identifier identifies an actual entity participating in the system.

For example:

rsm:LivingSystem

identifies a foundational semantic concept, while:

https://id.rsm.org/living-system/01KABC

identifies an actual LivingSystem.

Likewise:

rsm:agriculture/soil-system

identifies a taxonomy concept rather than a particular soil system.

#7.2 Illustrative JSON-LD Representation

An instance may therefore be represented conceptually as:

{
  "@context": "https://schema.rsm.org/context/v3",
  "@id": "https://id.rsm.org/living-system/01KABC",
  "@type": "rsm:LivingSystem",
  "rsm:domainType": [
    "rsm:agriculture/soil-system",
    "rsm:agriculture/managed-agroecosystem"
  ],
  "rsm:name": "Field 42 Soil System"
}

The vocabulary describes what the entity means. The entity identifier establishes which actual thing is being described.

#7.3 Context Governance

Canonical JSON-LD interchange SHALL reference the governed RSM context rather than redefining terms locally. Implementations SHALL NOT create competing meanings for governed RSM terms through arbitrary local context declarations.

Domain vocabulary extension SHALL use governed taxonomy mechanisms and absolute or properly compacted IRIs. Extensions may add vocabulary without redefining established RSM semantics.

#7.4 No Second Ontology

The RSM semantic projection SHALL remain governed by the RSM context. Applications SHALL NOT create a parallel ontology merely because they need to describe a new agricultural, health, financial, food, water, logistics, or policy concept.

The governing principle is:

One semantic namespace, one governed grammar, multiple extensible domain vocabularies.

#8. Domain Types and Facets

#8.1 Domain Type

rsm:domainType expresses one or more governed concepts that specialize the meaning of a foundational entity. Domain types SHALL remain concept references rather than redefining the entity's canonical foundational kind.

For example:

"@type": "rsm:Organization"
"rsm:domainType":
  - "rsm:food/processor"
  - "rsm:food/aggregator"
  - "rsm:agriculture/service-provider"

The Organization remains one canonical entity even as its domain meaning becomes richer.

#8.2 Multi-Typing

RSM SHALL support multiple simultaneous domain types. Real regenerative entities frequently span several functions, industries, ecological contexts, and market roles.

For example:

Community Grain Cooperative

rsm:Organization

domainTypes:
    rsm:agriculture/producer
    rsm:food/aggregator
    rsm:food/processor
    rsm:education/training-provider

Applications MAY select domain types appropriate to a particular projection. They SHALL NOT erase other valid types merely to fit a single-category user interface.

#8.3 Facets

A facet is a convenient projection or classification over one or more domain concepts. Facets MAY support discovery, filtering, navigation, or user experience without becoming canonical identity.

For example:

Facet:
producer

Organization:
Community Grain Cooperative

The facet does not transform the Organization into a Producer canonical class. It expresses a useful view of the Organization's domain semantics.

#9. Contextual Roles

#9.1 Definition

A Role describes how an entity participates within a particular context. Roles SHALL be modeled separately from foundational kind because participation can change across contexts and over time.

Illustrative roles include:

• Subject;
• Actor;
• Participant;
• Principal;
• Steward;
• Observer;
• Evidence Provider;
• Committer;
• Beneficiary;
• Affected Party;
• Provider;
• Consumer;
• Custodian;
• Issuer;
• Evaluator;
• Representative;
• Counterparty; and
• Resource.

These roles are not required to become canonical entity kinds.

#9.2 Context-Bound Roles

One Organization may hold different roles simultaneously:

Organization:
Regional Mill

Formation A:
processor

Formation B:
buyer

Research Context C:
evidence provider

Community D:
member

No semantic contradiction exists because the role is bound to context. RSM SHALL therefore avoid storing contextual roles as permanent immutable identity attributes.

#9.3 Temporal Roles

Roles MAY carry temporal validity.

For example:

subjectRef: https://id.rsm.org/person/01KPERSON
role: rsm:role/steward
contextRef: https://id.rsm.org/community/01KCOMMUNITY
validFrom: 2026-01-01
validUntil: 2027-01-01

Historical role state MAY become part of Biography. Expiration of a role does not erase the historical fact that it existed.

#10. Agency and Authority

#10.1 Agency Is Not Universal

RSM SHALL NOT attach a universal nullable agency=true|false field to every entity. Agency is applicable to particular kinds and contexts rather than a property every entity must possess.

A conceptual default can be represented as:

The diagram communicates ordinary semantics rather than introducing a universal agency property. Domain extensions may represent specialized systems carefully, but they SHALL NOT weaken the distinction between agency and mere subjecthood.

#10.2 Agency and Authority Are Different

Agency answers:

Can this type of entity perform an action?

Authority answers:

May this entity perform this specific action in this context?

A Person may possess agency but lack authority to commit an Organization. A DigitalAgent may possess technical Capability but lack authority to disclose private Evidence or commit funds.

RSM SHALL therefore preserve:

Agency ≠ Authority
Authentication ≠ Authority
Affiliation ≠ Authority
Membership ≠ Authority
Capability ≠ Authority

Consequential action requires applicable agency together with valid authority, policy, and any required consent.

#10.3 Representation

Where an Actor performs an action on behalf of another entity, the represented entity SHALL remain explicit. RSM SHALL not infer representation merely from employment, affiliation, membership, authentication, or technical access.

For example:

Actor:
Maria

actsFor:
Green Valley Farm

Operation:
COMMIT

AuthorityBasis:
verified delegated authority

The distinction allows RSM to explain not merely who performed an operation, but whose interests and authority were being exercised.

#11. Subject–Actor Interaction Grammar

#11.1 General Form

The simple subject–verb–object metaphor becomes more precise once Subject and Actor are separated. RSM interactions SHOULD be interpretable through the following general structure:

Actor
  + Authority
  + Context
       ↓
     Verb
       ↓
    Subject
       ↓
Object / State / Effect
       ↓
Exchange / Transformation
       ↓
Observation
       ↓
Outcome
       ↓
Biography

Actor identifies who performed the action. Subject identifies what the action or expression concerns.

#11.2 Observation Example

A laboratory observation may be represented conceptually as:

Actor:
Valley Soil Laboratory

Verb:
OBSERVE

Subject:
Field 42 Soil System

Observed Property:
soil organic carbon

Result:
2.7%

Context:
Regenerative Transition Formation

Provenance:
Laboratory Report 2026-10-04

The laboratory performs the Observation. The soil system is the Subject of that Observation.

#11.3 Commitment Example

A commitment may be represented as:

Actor:
Maria

actsFor:
Green Valley Farm

Verb:
COMMIT

Subject:
Green Valley Farm

Commitment:
500 acres of transition production capacity

Context:
Regional Grain Formation

This model separates the human Actor, represented Organization, commitment semantics, and Formation context. Those distinctions are necessary for explainable authority and provenance.

#12. Relationships Among Subjects

#12.1 Relationships Are First-Class

Foundational entities MAY participate in first-class Relationships. Relationship semantics SHALL be extensible through the governed RSM vocabulary rather than requiring every relationship type to become a schema property.

Illustrative relationships include:

Person        affiliated-with       Organization
Person        member-of             Community
Organization  operates              Facility
LivingSystem  located-at            Place
LivingSystem  part-of               LivingSystem
Facility      located-at            Place
Material      produced-by           Formation
Material      derived-from          Material
Instrument    issued-by             Organization
Formation     comprises             Participant
Formation     affects               LivingSystem
Community     stewards              Place

A Relationship is not merely a graph edge. It may carry lifecycle, temporal validity, provenance, authority, confidence, conditions, and Biography.

#12.2 Relationship Vocabulary

Relationship types SHOULD use governed vocabulary terms where interoperability matters. Domain Packs MAY add new relationship concepts without changing the foundational Relationship model.

For example:

rsm:relationship/located-at
rsm:relationship/operates
rsm:relationship/derived-from
rsm:relationship/stewards
rsm:relationship/affects

Domain-specific relationships may use hierarchical terms where appropriate:

rsm:agriculture/grows-on
rsm:food/processed-by
rsm:water/drains-into

The underlying canonical Relationship semantics remain stable.

#13. Subject and Observation

#13.1 Observation Semantics

Anything meaningfully observable may play the Subject role. The canonical Observation grammar should be capable of preserving the feature of interest, observed property, procedure, result, time, observer, provenance, and relevant context.

The conceptual form is:

Observer
    observes
Subject
    with respect to
Observed Property
    using
Procedure
    at
Time
    producing
Result

This grammar can describe ecological, material, institutional, technical, economic, and human observations without changing the core model.

#13.2 Examples

Examples include:

Subject: Soil System
Property: soil organic carbon

Subject: Grain Lot
Property: protein content

Subject: Facility
Property: available throughput

Subject: Organization
Property: certification status

Subject: Watershed Ecosystem
Property: nitrate concentration

Subject: Person
Property: applicable health measurement

The foundational kind of the Subject remains independent from the Observation. The Observation records something about the entity rather than redefining what the entity is.

#14. Subject and Outcome

#14.1 Outcome Must Identify What Changed

An Outcome SHOULD identify the Subjects whose state, condition, capacity, relationship, or other relevant property changed. This prevents Outcome from becoming an abstract success label detached from reality.

For example:

Outcome:
Reduced nitrogen runoff

Affected Subjects:
Field 42 Soil System
Mill Creek Watershed Ecosystem

A separate Outcome might state:

Outcome:
Improved operating margin

Affected Subject:
Green Valley Farm

The two Outcomes may arise from the same Formation while remaining semantically distinct.

#14.2 Multi-Subject Outcomes

Regenerative interventions frequently affect multiple Subjects at different levels. One Formation may influence soil condition, farm economics, water quality, worker wellbeing, processing resilience, food characteristics, community capability, and eventually human health.

RSM SHALL preserve these Outcome relationships rather than collapsing them into one universal regenerative score. Different consequences may have different evidence, time horizons, confidence, and affected Subjects.

#14.3 Outcome Is Not Intent

RSM SHALL preserve:

Intent ≠ Outcome
Activity ≠ Outcome
Observation ≠ Outcome

Intent describes what is desired. Activity describes what was done, while Observation and evaluation provide evidence concerning what actually changed.

#15. Subject Biography

#15.1 Biography as Accumulated History

A durable Subject MAY accumulate Biography where such history is meaningful. Biography may include relevant Relationships, Observations, Activities, Events, Commitments, Evidence, Assertions, transformations, and Outcomes.

Different kinds naturally accumulate Biography at different scales. A grain lot may have a short lifecycle measured in months, while a watershed may accumulate meaningful observations and interventions over decades.

#15.2 Biography Is Not Current State

RSM SHALL preserve the distinction:

Biography ≠ Current State

Biography answers what has happened to, around, or through a Subject. Current state answers what the system presently believes, observes, or is authorized to assert about that Subject.

Historical Evidence may inform future reasoning without automatically being treated as current truth.

#16. Domain Extension Architecture

#16.1 Agriculture Domain

An agriculture vocabulary may introduce concepts such as:

LivingSystem
    rsm:agriculture/soil-system
    rsm:agriculture/crop-system
    rsm:agriculture/agroecosystem
    rsm:agriculture/herd
    rsm:agriculture/pasture-ecosystem

Place
    rsm:agriculture/farm
    rsm:agriculture/field
    rsm:agriculture/parcel

Organization
    rsm:agriculture/producer
    rsm:agriculture/agronomy-provider

Material
    rsm:agriculture/seed-lot
    rsm:agriculture/grain
    rsm:agriculture/biomass

These terms extend the governed vocabulary. They do not introduce new foundational kinds.

#16.2 Food Domain

A food vocabulary may introduce:

Organization
    rsm:food/processor
    rsm:food/aggregator
    rsm:food/buyer

Facility
    rsm:food/mill
    rsm:food/packhouse
    rsm:food/cold-storage

Material
    rsm:food/grain-lot
    rsm:food/ingredient
    rsm:food/food-product

Instrument
    rsm:food/certification
    rsm:food/purchase-agreement

The same canonical grammar remains applicable across production, processing, distribution, and consumption.

#16.3 Water Domain

A water vocabulary may include:

LivingSystem
    rsm:water/river-ecosystem
    rsm:water/wetland
    rsm:water/aquatic-ecosystem

Place
    rsm:water/watershed
    rsm:water/aquifer
    rsm:water/riparian-zone

Instrument
    rsm:water/water-right
    rsm:water/discharge-permit

A watershed may sometimes be modeled as Place and sometimes as a LivingSystem depending upon what is being described. Where both meanings matter, separate related identities SHOULD be preserved rather than forcing one object to carry contradictory semantics.

#16.4 Health Domain

A health vocabulary may introduce:

Person
    rsm:health/patient
    rsm:health/clinician

Organization
    rsm:health/provider
    rsm:health/research-institution

Facility
    rsm:health/clinic
    rsm:health/laboratory

LivingSystem
    rsm:health/human-biological-system
    rsm:health/gut-microbiome

Material
    rsm:health/biological-sample

Instrument
    rsm:health/diagnostic-test
    rsm:health/care-protocol

The same RSM grammar can therefore connect ecological and food-system outcomes with human-health investigation without collapsing those domains into one causal model.

#16.5 Finance, Carbon, Logistics, and Policy Domains

Additional domains may introduce concepts such as:

rsm:finance/transition-loan
rsm:finance/investment
rsm:finance/insurance-policy

rsm:carbon/soil-carbon-credit
rsm:carbon/carbon-project

rsm:logistics/freight-service
rsm:logistics/cold-storage
rsm:logistics/aggregation-hub

rsm:policy/conservation-easement
rsm:policy/regulation
rsm:policy/grant-program

These terms remain inside the governed RSM semantic namespace. Domain expansion therefore enriches the ecology without fragmenting its semantic foundation.

#17. Complete Subject Ecology

#17.1 Structural View

The complete Subject ecology can be summarized as:

The diagram deliberately avoids making Subject a parent class. It illustrates which foundational kinds may play the semantic Subject role and how contextual participation overlays durable identity.

#17.2 Living Outcome-Forming Ecology

The Subject Model supports the wider RSM model of a living outcome-forming ecology. Persistent entities exist between outcomes, while relationships, signals, capabilities, capacity, commitments, materials, instruments, and ecological state provide the substrate from which new formations can emerge.

The ecology therefore contains more than actors. It includes everything that can materially shape, constrain, carry, experience, evidence, or preserve the consequences of regenerative activity.

#18. Normative Invariants

#18.1 Identity Invariants

Implementations SHALL preserve the following distinctions:

Person ≠ Organization
Organization ≠ Community
Organization ≠ Facility
LivingSystem ≠ Place
Formation ≠ Relationship
DigitalAgent ≠ Person
Material ≠ Facility
Instrument ≠ Organization

These concepts may participate in relationships with one another. Their identities and semantic meanings SHALL nevertheless remain distinct.

#18.2 Role Invariants

Implementations SHALL preserve:

Subject ≠ Actor
Subject ≠ Participant
Participant ≠ Organization
Membership ≠ Authority
Affiliation ≠ Authority
Resource ≠ Foundational Kind
Buyer ≠ Foundational Kind
Seller ≠ Foundational Kind

Contextual participation SHALL NOT silently rewrite durable identity.

#18.3 Agency Invariants

Implementations SHALL preserve:

Agency ≠ Authority
Authentication ≠ Authority
DigitalAgent ≠ Principal by default
Recommendation ≠ Commitment
Proposal ≠ Agreement
Inference ≠ Assertion

Machine or human ability to perform an action SHALL NOT by itself constitute permission to perform that action.

#18.4 Epistemic Invariants

Implementations SHALL preserve:

Observation ≠ Subject
Claim ≠ Evidence
Evidence ≠ Assertion
Activity ≠ Outcome
Intent ≠ Outcome
Biography ≠ Current State

These distinctions ensure that plans, actions, measurements, claims, and consequences remain independently inspectable.

#18.5 Vocabulary Invariants

Implementations SHALL preserve:

Domain Type ≠ Foundational Kind
Facet ≠ Identity
Domain Vocabulary ≠ Second Ontology
Schema IRI ≠ Entity Identity

A domain concept may enrich the meaning of a canonical entity. It SHALL NOT silently redefine the foundational RSM grammar.

#19. Illustrative End-to-End Ecology

#19.1 Scenario

Consider a farmer seeking to transition a field toward regenerative production while maintaining economic viability. The desired outcome cannot be created by the farmer alone because agronomic knowledge, testing, finance, processing capacity, certification, demand, ecological context, and market commitments all influence whether the transition can succeed.

The relevant ecology may include:

Person
    farmer
    agronomist
    researcher

Organization
    farm enterprise
    processor
    buyer
    laboratory
    certifier
    lender
    conservation district

Community
    regional producer network

LivingSystem
    field agroecosystem
    soil system
    watershed ecosystem

Place
    field
    farm
    watershed

Facility
    mill
    testing laboratory
    storage facility

Material
    seed lot
    crop
    grain lot
    flour
    food product

Instrument
    certification
    transition loan
    purchase agreement
    transition grant

DigitalAgent
    Wellzai coordination agent

Formation
    regenerative grain transition formation

These familiar terms remain domain concepts layered on top of stable foundational kinds.

#19.2 Semantic Representation

The ecology can express relationships such as:

Farmer
    acts-for → Farm Organization

Farm Organization
    operates-at → Farm Place

Field Agroecosystem
    located-at → Field

Laboratory
    observes → Soil System

Buyer
    expresses → Demand

Wellzai DigitalAgent
    discovers → relevant Capabilities

Lender
    commits → Capital

Processor
    commits → Processing Capacity

Farm Organization
    commits → Production Capacity

Formation
    coordinates → Participants

Formation
    performs → Activities

Activities
    affect → Soil System

Farm Organization
    produces → Grain Lot

Grain Lot
    transformed-into → Food Product

Observed Outcomes
    affect → Farm Economics
    affect → Soil System
    affect → Watershed Ecosystem
    affect → Community

The same model can preserve observations, evidence, commitments, transformations, outcomes, and biography across the entire sequence.

#19.3 JSON-LD Illustration

A particular soil system might appear as:

{
  "@context": "https://schema.rsm.org/context/v3",
  "@id": "https://id.rsm.org/living-system/01KSOIL42",
  "@type": "rsm:LivingSystem",
  "rsm:domainType": [
    "rsm:agriculture/soil-system",
    "rsm:agriculture/managed-agroecosystem"
  ],
  "rsm:name": "Field 42 Soil System"
}

A grain lot could be represented as:

{
  "@context": "https://schema.rsm.org/context/v3",
  "@id": "https://id.rsm.org/material/01KGRAIN",
  "@type": "rsm:Material",
  "rsm:domainType": [
    "rsm:agriculture/grain",
    "rsm:food/ingredient"
  ],
  "rsm:name": "2027 Field 42 Wheat Lot"
}

The examples use one governed RSM vocabulary while allowing domain semantics to overlap naturally.

#20. Consequences for RSM Architecture

#20.1 Small Kernel

The canonical RSM kernel SHOULD remain small and strongly governed. Foundational semantic differences belong in the kernel only when collapsing them would compromise identity, agency, provenance, lifecycle, or interoperability.

The kernel therefore does not need classes for every farmer type, crop, certification, test, ecosystem, product, financing structure, processing method, or regulatory program. These concepts belong primarily to extensible vocabulary and domain packs.

#20.2 Rich Ecology

A small kernel does not imply a simplistic domain model. The richness of RSM emerges from combinations of identity, domain types, roles, relationships, capabilities, observations, evidence, transformations, commitments, outcomes, and Biography.

A small number of stable primitives can therefore describe a highly heterogeneous ecosystem. The resulting model remains extensible without sacrificing semantic discipline.

#20.3 Governed Domain Packs

Domain Packs MAY provide governed collections of:

• domain concepts;
• relationship types;
• predicates;
• measurement semantics;
• units;
• constraints;
• evidence expectations;
• interoperability mappings; and
• domain-specific validation rules.

A Domain Pack SHALL extend RSM semantics rather than replace the RSM context or redefine its canonical terms. The architecture should allow agriculture, food, water, health, finance, carbon, logistics, and policy vocabularies to evolve at different rates while remaining mutually intelligible through RSM Core.

#20.4 Federation

Federated systems MAY expose only the semantic projections necessary for authorized interaction. Participation in RSM SHALL NOT require every organization, community, laboratory, farm, or ecological monitoring system to surrender its complete internal data model.

Shared identity, governed vocabulary, protocol semantics, provenance, and authorization boundaries provide interoperability. Local systems retain sovereignty over their authoritative state.

#21. Design Principle

#21.1 Small in Ontology, Rich in Ecology

The RSM Subject Model deliberately avoids the temptation to model the entire regenerative world as a growing hierarchy of canonical classes. Such an approach would eventually embed the assumptions of agriculture, food, finance, health, ecology, logistics, policy, and every future domain directly into the core platform.

RSM instead establishes a stable semantic grammar and allows vocabulary to evolve around it. Foundational kinds preserve the differences that matter structurally, while governed domain concepts describe the diversity of the world those kinds inhabit.

#21.2 The Resulting Grammar

With the Subject Model established, the RSM grammar can be stated as follows:

An Actor, acting under applicable Authority and within Context, performs a Verb concerning one or more Subjects. That interaction may create or transform state, materials, relationships, information, evidence, commitments, resources, or other objects, producing consequences that can be observed as Outcomes and preserved as Biography.

The Subjects themselves remain independently identifiable throughout this process. Their foundational kind explains what they fundamentally are, their domain types explain what they mean in relevant domains, and their contextual roles explain how they participate in a particular situation.

#21.3 Final Principle

The Subject Model therefore rests on a concise architectural principle:

RSM is one semantic language with a small stable grammar and an extensible vocabulary for a living ecology.

rsm: is the governed semantic namespace. rsm:LivingSystem is part of the grammar, rsm:agriculture/soil-system is vocabulary, the actual soil system is an independently identified entity, and Subject describes the role that entity plays in a particular expression.

This separation gives RSM the ability to grow across regenerative agriculture, food systems, ecology, water, finance, health, climate, logistics, policy, and domains not yet anticipated without losing semantic coherence. It allows the ontology to remain disciplined while the ecology it represents remains open, adaptive, and alive.