Guide 3 established the method. This guide addresses the material the method operates on. The grammar — the node taxonomy, the CHIP/ID tri-axis, the seed field, the domain-chain view — is the compression language through which the instrument renders a reconstructed case legible. This guide is expository: it states what the grammar is, what each layer of it is doing, and where the compression can deceive a reader who stops at the surface.
The grammar itself is authored in the factory. A target user reconstructing their own case applies it to their material according to the discipline described in the target-user corpus. The state user encounters the finished grammar — the chip field as it appears on the reviewer surface — and reads it. The task here is reading, not authoring.
Node markers and the lifecycle
Node markers describe the functional position of an event within a propagation chain. They answer a specific question: at this point in the sequence, what structural role is this event playing? A node marked as a Containment Structure is playing the role of absorbing pressure without producing change. A node marked as Actor Recurrence is playing the role of introducing a repeat actor into a process already in motion. The marker does not describe what the event is; it describes what the event does.
The marker set is distributed across six lifecycle stages of a propagation chain:
- Origin — the event at which a chain enters the record
- Distortion — events at which material is misframed, reframed, or selectively represented
- Linkage — events at which distorted material crosses from one actor or agency to another
- Structural — events at which the distorted material becomes embedded in formal process
- Suppression — events at which pressure to correct the chain is absorbed without correction occurring
- Outcome — events at which the chain produces a concrete consequence against the target
Each stage contains markers that indicate specific mechanisms of that stage's operation. A single node may carry multiple markers — a document that is both a Containment Structure (Suppression) and an Actor Recurrence (Linkage) is registering both functions simultaneously, and the chip field shows both.
The lifecycle is directional. A chain does not move backwards through it. A chain currently at Structural will produce Suppression events as pressure to correct accumulates; a chain at Suppression will produce Outcome events as the accumulated containment converts to material harm. This directionality is what makes the taxonomy predictive of chain behaviour within its own domain — not predictive of whether a chain will form, but predictive of how a chain already forming will continue. The practical consequence: a state user reading a case at mid-lifecycle can identify with reasonable confidence what categories of event are still to come if the chain is not interrupted.
The state user reading node markers sees, at a glance, where in the lifecycle a node sits. A reviewer who has absorbed the marker set can classify a chain's current position in seconds. No narrative reading of the events would surface this position as quickly.
The compression deceives when the reader stops at the marker and does not check the node content that produced it. The marker is the author's classification of the event, made against the author's primary evidence. It is not a guarantee that the classification is correct, that the evidence supports it, or that the event exists as characterised. The marker is a hypothesis the author has asserted; verification is the operator's.
The CHIP/ID tri-axis
Where node markers describe where in a chain a node sits, CHIP/IDs describe the structure of the cross-references between nodes — the jumps by which a chain moves from one node to another.
A CHIP/ID carries three axes of information at once:
- Colour: the domain at the jump target. Six domains: Origin, Process, Record, Institution, Suppression, Consequence. The colour tells the reader what terrain the chain enters at this jump — what kind of institutional space the chain is now occupying.
- Number: the class within that domain. The number tells the reader which specific pattern within the domain is engaged — within this terrain, which mechanism is operating.
- Seed field: the agency thread or seed-set. Up to twenty-six letters, A through Z. The seed field may contain no letter, one letter, or multiple letters. A multi-letter field is a compressed seed-set: the same domain/class function applies to each listed seed in that chip position.
A typical CHIP/ID may appear in compact form as origin:1:L: the target is in the Origin domain, class 1, and the seed field resolves locally to L. After seed-set compression, the seed field may carry multiple letters, for example record:3:ABCDEGMP. That does not name a single composite actor. It says that the Record / class 3 function is being carried by the seed-set A, B, C, D, E, G, M, and P at that jump position. A chip with an empty seed field indicates a class identified at the target but no seed actor currently assigned — the chain's origin is either unknown, not yet authored, or structurally anonymous.
The three axes are independent. A chip carries all three pieces of information in a single compact visual form. The seed field may be singular or compressed, but the domain and class still govern the chip. A reader must distinguish a compressed seed-set inside one chip from a stacked sequence of chips separated by a pipe character. For example, record:3:ABCDEGMP|institution:4:ABCDEGMP|suppression:6:ABMP contains three ordered chips, not one chip and not one actor.
A chip field allows structural patterns to become visible across hundreds of nodes in minutes. Same-colour convergences — the same domain appearing from multiple chains at one node — indicate intra-domain concentration. Different-colour convergences indicate cross-domain crossing points. Recurrence of a seed letter or seed-set across multiple nodes indicates agency persistence. Each of these is an observation the reader can form from the chip field alone, without reading any node's content.
The compression deceives in four specific ways. First, the chip is the author's classification; if the author has mis-assigned the domain, number, or seed field, the chip will still read cleanly, because the grammar does not validate itself against the evidence. Second, the legibility of the chip field can persuade the reader that they have understood the case when they have only understood the chip field; understanding requires descent to the nodes and the primary documents. Third, a compressed seed-set may be mistaken for a single composite actor, when it is actually several local seed threads sharing the same domain/class function. Fourth, a pipe-separated stack may be mistaken for a compressed seed-set, when it is actually an ordered sequence of distinct chip functions.
The seed field as causal-persistence instrument
Of the three CHIP/ID axes, the seed field is distinct in function. Colour and number describe the terrain and pattern at each jump target; both can change as a chain moves from node to node. The seed field preserves the agency thread or seed-set being carried through that jump.
This makes the seed field a causal-persistence instrument. Following a single seed letter through a chronology is following the propagation of that local seed thread through every institutional space it enters. Following a compressed seed-set is following a bundle of local seed threads that share the same domain/class function at that jump position. If actor L seeds a chain at Origin, L persists in downstream CHIP/IDs that belong to L's chain. If a chip carries ABMP, the reader is seeing A, B, M, and P carried together for that chip function, not a new actor called ABMP.
This is the property that makes grey-area reconstruction possible at all. Without a causal-persistence instrument, a chain that crosses four institutional domains would appear as four separate events distributed across those domains — with the origin invisible to any reviewer entering the material at any point past Origin. The seed field is the axis that preserves chain identity across propagation, whether that field contains one seed or a compressed set.
A state user tracing a seed field through the domain-chain view (addressed in the next section) sees the propagation path of a single seed thread or compressed seed-set in a single view. This is the instrument's deepest analytical capability at the single-case level. At the multi-case level it is powerful but dangerous: recurrence of a letter or seed-set across cases indicates a structural comparison point, not automatic identity of actors.
The compression deceives when the reader conflates seed-field identity with actor identity across cases. Within a case, letter L means the actor defined in that case's seed chain — the letter-to-actor map authored for that specific continuum. A compressed seed-set such as ABMP is also local to that case. Across cases, L in case-001 is not necessarily the same actor as L in case-002, and ABMP in one case is not automatically the same agency bundle in another. A state user cross-referencing seed fields across cases must check the seed-chain resolution in each case, not assume that matching letters denote matching actors. The seed field is a local identifier; the actor is an extra-grammatical reality that must be confirmed separately.
The domain-chain view
The primary chronology view organises nodes by date. This is the view most consumers of the record encounter first. It is also the view that obscures the propagation structure that the grammar most directly exposes.
The domain-chain view — the tier-two surface — organises the same nodes differently. Instead of by date, nodes are grouped by the CHIP/ID domain at their incoming jump target. All Origin targets appear together. All Process targets appear together. And so on across the six domains. Within each domain group, the seed field becomes an organising axis: targets may be grouped by single seed letters, compressed seed-sets, or both, depending on how the local case grammar has been authored.
What this view reveals that date-ordering hides: a chain's propagation path across domains appears as a sequence. If seed A's chain moves from Origin through Process through Institution through Suppression, that movement is a contiguous visible sequence in the domain-chain view. If a compressed set such as ABMP moves through the same path, the reader is seeing several agency threads sharing that path at those jump positions. In the primary chronology, the same events are scattered across whatever dates they occurred on — dates which may be years apart, interspersed with events from other chains, and unrecognisable as a chain's propagation without significant reading effort.
To state the effect concretely: in the primary chronology, a reader sees what happened when. In the domain-chain view, the same reader sees how a chain moved. The events are the same events. The view is different. The grammar makes both views possible from the same authored material without re-authoring.
The domain-chain view is where propagation becomes structurally visible rather than narratively reconstructed. A state user reading this view for a single case sees the chain's architecture directly, including whether a path is carried by one seed or by a compressed seed-set. At registry level, the view allows comparison of chain architectures across cases — a specific propagation path that recurs across multiple cases indicates a structural pattern that is invisible at single-case date-ordered resolution.
The compression deceives when the reader treats the domain-chain view as a substitute for the primary chronology rather than as its structural complement. The primary chronology is the temporal record; the domain-chain view is a structural re-organisation of that record for specific analytical purposes. Findings derived from the domain-chain view must still be traceable back through the primary chronology to the specific dates and evidence they rest on. The view compresses; it does not replace.
The grammar as a whole
The four layers — node markers, CHIP/ID tri-axis, seed field, domain-chain view — are not independent. They compose. Node markers classify what each event is doing; CHIP/IDs classify the jumps between events; the seed field preserves causal identity or seed-set identity across those jumps; the domain-chain view reorganises the whole to make propagation visible.
A state user reading a case in the primary chronology sees node markers and chips on individual events and their cross-references. Reading the same case in the domain-chain view, the same markers and chips are visible — but reorganised to foreground propagation structure. The grammar has not changed. The view has.
This compositional property is what makes the grammar analytically useful at different resolutions. A question about a single node is answered at the node with its markers. A question about a specific cross-reference is answered at the chip with its three axes. A question about a chain's propagation is answered in the domain-chain view with the seed field as organising axis. A question about a pattern across cases is answered in the registry, which reuses all four layers at cross-case scale and is the subject of guide 5.
Where the compression stands in the operator's work
The grammar is a compression. It is not the case. It is the author's compression of the case, against the author's primary evidence, applied with the discipline the authoring corpus requires. What the state user reads is the compression.
This does not make the grammar untrustworthy. It makes it testable. Every layer of the compression — markers at nodes, chips on jumps, seed fields across propagation, domains in the domain-chain view — rests on material the operator can descend to and verify. The compression is valid exactly to the degree that the descent has been done.
Axiom II from guide 2 governs this work: familiarity with the instrument breeds contempt for the instrument. The grammar is the surface where familiarity is most easily accumulated and contempt most easily developed. An operator who has read a thousand chip fields can classify the next one at a glance, and be wrong, and never discover the error because the glance was compatible with an answer that was never checked.
The grammar is a tool for reading. It is not the reading. What the operator reads is the evidence the grammar compresses. What the operator produces is a finding about that evidence. The grammar stands between them as a surface; it is not the substance of either.
Guide 5 — Reading a Registry — extends this reading to the multi-case surface, where the same grammar operates across cases and new analytical patterns become visible. The discipline of this guide applies throughout guide 5; the techniques of guide 5 are where the state user's characteristic work most directly happens.