RTR0: Morphogenesis
Chapter 4. RTR0: Morphogenesis
Atomic and composite units · Generation of nodes · Deterministic and autonomous synaptogenesis · Parallelism of phases · Transition to operation · Critical period
4. 1. Atomic unit: MORN
Each transformation of Gativus has an atomic unit — a basic structure that arises in the target space through the application of splice. The atomic units of each level are formally isomorphic: a triad of three components in which the middle one is a dynamic connecting vector shortening the distance between the first and the third.
- •
For GTR1, this is OPRN — the triad subject — b-vector — object. The b-vector shortens the distance in physical space MAP5: 'the object is there, the subject is here'.
- •
For GTR2, this is KLEN — the triad symbol — P-vector — symbol. The P-vector shortens the distance in the narrative space MAP7: 'state X, state Y is needed'.
- •
For GTR3, this is WILL — the triad Concept — W-vector — Concept. The W-vector shortens the distance in the axiological space MAP9: 'C-is exists, C-ought is required'.
- •
For GTR0, the atomic unit is MORN (MOrphogenesis Node). The triad of MORN is isomorphic to the others:
LOMN — the subject of morphogenesis, the executor of the m-vector. LOMN reads its part of MOVE, forms the local m-vector and executes it, turning the description into actual connections between nodes.
m-vector — the dynamic connecting vector of morphogenesis. It is a local connecting act with a unique identifier for each concrete functional group. The m-vector is a multidimensional object; its coordinates are distributed across all projections of MOLD. Each diagram — CLSS, COMM, COMP, RSRC, ACTD and the rest — gives the human perception one face of the m-vector. These projections jointly describe the m-vector completely: which nodes of which classes should be connected, by which rules, in which composite context, under which resource conditions, with what procedural semantics.
Group — the object of the m-vector's execution, the set of real NDDI nodes that are actually created and connected. The structure of the group is determined by the projection of the m-vector onto the relevant MOLD diagrams.
Isomorphism with the upper levels is complete: on every level there is an actor (executor of the vector), a vector (multidimensional with distributed coordinates), and a patient (with structure). At GTR0 the actor is especially prominent — it is a complete LOMN node, not a function inside a subject; yet the same triadic structure operates on all four levels.
a) Contradiction at the zero level
After NRGN, nodes exist in the network — cells of given classes, with loaded D-components and generated V-components. But this is not yet an organism. This is biomass: many independent nodes each on its own, without functional integrity. The network as a network is not working.
Between 'cells have been created' and 'the functional group is working' there lies a distance — the contradiction of the zero level. This contradiction has the same logical nature as the b/p/w vectors of the upper levels: there is an initial state, a target state, and a distance between them that must be shortened.
Shortening the distance is the connection of cells into a functional group according to a template. This act of connection is the m-vector. It is local, has a unique identifier, is formed at the moment of appearance of each concrete functional group, and is zeroed when the group is completed.
The m-vector works by the same logic as b/p/w:
- •
The b-vector of a concrete action is local and unique: 'this very motion towards this very goal'. The goal is reached — b is zeroed.
- •
The P-vector of a concrete logical step is local: 'from these premises this conclusion is to be made'. The conclusion is made — p is zeroed.
- •
The w-vector of a concrete contradiction is local: 'this contradiction here and now'. The contradiction is removed — w is zeroed.
- •
The m-vector of a concrete functional group is local: 'connect these very nodes into this very micro-column'. The connection is complete — m is zeroed, MORN exists as an actualized atomic unit.
b) Nature of the distance at the zero level
The distance shortened by the m-vector has the same logical nature as the distances shortened by the upper levels — it exists within a single space, between two states of that space. The b-vector shortens distance in MAP5, the p-vector in MAP7, the w-vector in MAP9. The m-vector shortens distance in the cellular space DOM1/DOM2, between its two states: the unorganized state (a set of cells without functional connections) and the organized state (the same cells connected by some template into one functional group).
Under a microscope, neural tissue looks similar in the two states — the cells are the same, the projections are the same. The difference is not in the material but in the structure of connectivity. Before the execution of the m-vector, connections between cells are either absent or random; after execution, connections are established according to a concrete template, and the set of cells works as one functional whole.
This aligns GTR0 with all the upper levels: at each level there is a space with two states (an initial state with distance and a target state without it) and a dynamic vector shortening the distance through a concrete local act.
c) Multidimensionality of the m-vector
The m-vector is a multidimensional object. Its coordinates are distributed across all projections of MOLD: one coordinate sets the node types (CLSS), another the rules of their connection (COMM), a third the composite context (COMP), a fourth the resource parameters (RSRC), a fifth the procedural semantics (ACTD), and so on.
The diagrams of MOLD are not parts of the m-vector but its projections onto planes of perception accessible to humans. The m-vector itself exists as a single multidimensional object; the diagrams are tools allowing finite human cognition to consider it from different sides.
This corresponds to the nature of MOVE as a whole: MOVE is a sparse vector in the library-class space, projected onto the MOLD diagrams. The m-vector is a local fragment of the same object, corresponding to a single connection act for one functional group. The relation between MOVE as a whole and the m-vector is the same as the relation between a subject's overall behaviour and the b-vector of one of his concrete actions.
d) Cardinality of the splice
A triad on any level is formally three-component, but the first and third components may be sets rather than individual elements. The RDF dogma 'one subject — one verb — one object' is a simplification that does not reflect the real nature of triads.
In OPRN, a set of subjects can act through one b-vector on a set of objects (collective perception, collective action). In KLEN, an n-ary logical relation connects a set of symbols. In WILL, real contradictions are rarely binary — they cover whole fields of Concepts.
In MORN, the cardinality of the third component can be very large. A group in one MORN may contain thousands of nodes of the same class (for example, a cortical column of 10000 nodes). This is not a violation of the triadic form but its natural manifestation at the zero level, where functional units are biologically large.
e) When MORN is considered actualized
MORN does not exist partially. LOMN — m-vector — group becomes MORN only when the m-vector is fully executed: the classes are defined in CLSS, the m-vector is formed in concrete form (from a template in COMM), the group is actually created and connected internally by the m-vector.
Before the execution of the m-vector, there are only classes and their potential forms — descriptive entities. This is pre-MORN. After the execution of the m-vector, MORN is complete: m is zeroed, the group exists as a functionally bound unit, MORN is fixed as an actualized atomic unit of the morphotransformation.
This explains why the NRGN and SYGN procedures are inseparably linked: NRGN generates nodes (the raw material of the m-vector), SYGN executes the m-vector (connecting nodes into groups). Only their joint completion gives a complete MORN.
4. 2. Composite unit: MLOM
Just as BLOM is assembled from OPRN, KLOM from KLEN, and WLOM from WILL, MLOM (Morphological LOM) is assembled from MORN. MLOM is a composite morphological construction in which several MORNs are joined by predefined internal connections and spatial relations.
a) Hierarchy of nesting
MLOM is recursive. At each level of nesting, an MLOM either directly contains MORNs or contains smaller-scale MLOMs. This yields a multi-level structure reflecting the hierarchical organization of real biological and engineering systems.
Examples of MLOM in the Gativus architecture:
A micro-column — an MLOM assembled from MORNs of different node types (for example, a MORN of receptor class, a MORN of processing class, a MORN of projection class). Inside the micro-column, groups are connected by strictly defined templates, and the spatial arrangement of the groups relative to one another is given by the template.
A column — an MLOM assembled from MLOMs of micro-columns. Several micro-columns of the same type are united into a column, and the templates of connections between micro-columns determine the functional role of the column.
An organ — an MLOM assembled from MLOMs of columns. The functional organs of Gativus — the spatial organ, the object organ, the symbolic organ, the narrative organ, the axiological organ — are MLOMs of the highest level within one of the components on the COMP diagram.
An organism — the highest-level MLOM, uniting all functional organs into a single network. Connections between organs are determined by the inter-organ interfaces in the COMP diagram.
b) Relation between MORN and MLOM
The difference between MORN and MLOM is the level of atomicity. MORN is atomic: it cannot be decomposed into simpler morphological units of the same kind. MLOM is composite: it consists of MORNs or smaller-scale MLOMs, and its structure is determined by the rules of composition of its parts.
The biological analogy is precise. A cell is an atomic unit (analogous to MORN on the lowest level). A tissue is a composite of cells (analogous to MLOM). An organ is a composite of tissues (an MLOM of MLOMs). An organism is a composite of organs (an MLOM of the highest level). On each level operates the same principle: a composite is formed from atomic units by certain rules of composition.
This recursiveness gives Gativus a fractal structure. At any level of observation, nodes and relations are seen; each node, on enlarging the scale, unfolds into a subnetwork in which nodes and relations are again present. The morphotransformation is applied recursively at each level — and the hierarchy MORN/MLOM directly reflects this recursion.
4. 3. NRGN: generation of nodes
NRGN (NeRoGeNesis) — the first phase of morphogenesis. Its task is to realize the classes described in the CLSS diagram of MOVE as groups of concrete NDDI nodes. NRGN is executed by the operational network OPNG (or, after the specialization of OPNG inside ANOD, by the LOMN of the M-section — see Chapter 5).
The NRGN procedure consists of three sequential steps: creation of classes, evaluation of resources, instantiation of objects.
a) Step 1: creation of classes
LOMN reads the CLSS diagram from MOVE. For each class described in the diagram, a special object is created — a class node. This is a full NDDI structure with the same sections as ordinary nodes (V, A, S, G, T, C, L, M depending on the level), with defined templates of connectors, and with established d-relations to D-components in the ROOT repository implementing the node's logic.
A class node has several key differences from an ordinary node. It has a local name in the GATE's name space but no global address UNON. It does not participate in network operation: it is not globally addressable, does not receive datagrams from outside, and does not form outgoing relations with other running nodes. D-components are not physically loaded into its M-section — only d-relations to the D-components in the ROOT repository or the local cache of the GATE are established.
A class node is a structural template plus a set of d-relations. Its function is to be a prototype for subsequent instantiation. One class node will be used to create thousands or millions of concrete nodes of the same class, all having the same structure and the same d-relations but different UNONs and different states of the V-section.
This gives an essential architectural economy. One D-component in ROOT serves thousands of class nodes in different GATEs through the network of d-relations. One class node in a GATE serves millions of concrete instances of the same class. The hierarchy ROOT → class → instance guarantees minimal duplication of executable code.
b) Step 2: evaluation of resources
Before proceeding to instantiation, LOMN evaluates the available AVEC resources along several axes (see GNET, Chapter 4). This evaluation determines how many concrete nodes of this class can be created within available quotas.
The evaluation does not reduce to a check of one numerical value and does not return a binary 'yes/no' answer. It checks:
Resource quotas. How many nodes are allowed to be created within the NRGN quota, how much memory and computation are allocated, what is the network traffic quota. The requested quantity is compared with the available.
Security. Does LOMN have the right to create nodes of this class. Does the creation violate the security policy of the GATE. Is the request beyond the limits of the delegated AVEC₀.
Compatibility. Are the D-components to which d-relations will be established compatible with the current version of the architecture on this platform.
The result of the evaluation is not 'permit or forbid' but a set of parameters for the next step. If resources are fewer than MOVE requires, the actual number of instances may be less than requested. This realizes the principle of resource constraints: the same MOVE on different platforms with different AVEC will yield organs functionally identical but quantitatively different.
If conditions cannot be satisfied at all — the requested class cannot be created in any quantity — LOMN awaits the restoration of resources or signals up the hierarchy (to ANOD, further to GATE) that morphogenesis is impossible under current conditions.
c) Step 3: instantiation of objects
After resource evaluation, LOMN starts the instantiation itself — the creation of concrete nodes from a class node. The GATE clones the class node N times, where N is determined by the result of the resource evaluation. Each copy passes through several sub-steps:
Inheritance of structure. The copy receives all sections of the class, all connector templates, all d-relations. Structurally it is identical to the class node, differing only in state.
Assignment of UNON. Each instance receives a unique global address UNON from the delegated name space. From this moment, it is addressable in the network.
Registration in routing. The UNON of the instance is entered into the GORP routing tables (see GNET, Chapter 4) — first into the local routing on the GATE, then if necessary up through the LRAI tree into global routing.
Start of execution. The instance begins to work: it loads its state from the M-section, activates its A-section, prepares to receive incident relations.
In parallel with instantiation, the next procedure starts — SYGD, the establishment of deterministic connections by the class's connector templates. NRGN and SYGD work almost simultaneously: new nodes are created and immediately included in the existing network of connections.
4. 4. SYGD: deterministic synaptogenesis
SYGD (SYnaptogenesis Deterministic) establishes connections between nodes by explicitly defined templates. It is the first of two mechanisms of synaptogenesis and is responsible for the skeleton connections — those that must be established immediately at morphogenesis and do not allow variability.
a) Source of deterministic connections
Templates of deterministic connections are contained both in the COMM diagram of MOVE and in the D-components of the classes themselves. The COMM diagram describes inter-class rules: which class connects with which, through which connectors, with what density. The D-component concretizes these rules at the level of an individual class: which connectors it has, which of them are deterministic, which are autonomous.
When NRGN creates an instance of a class, the templates of connectors have already been determined in this class through d-relations to the D-component. Some connectors are marked as deterministic — they immediately receive concrete addresses of partners upon creation. The remaining ones are marked as autonomous — for them the search-for-partner procedure starts later in SYGA.
b) What is established through SYGD
Deterministic connections in Gativus correspond to those that in biology are genetically predetermined and form by a strict procedure:
Templates inside a group. Connections inside a group of one MORN — for example, connections between the nodes of one micro-column. The nodes inside the group must be connected in a strictly defined way, and SYGD establishes these connections by the templates in the class's D-component.
Inter-layer projections. Connections between functional organs of different levels — for example, between the object domain and the symbolic domain. These connections set the trunk architecture of the network and are defined as inter-organ interfaces in the COMP diagram.
Administrative connections. Connections of each node to the ANOD of its organ (the administrative node of the organ), through which control signals and updates of AVEC quotas are transmitted. These connections are mandatory and are established automatically at the creation of nodes.
c) Properties of SYGD
SYGD has several important characteristics.
Determinism. SYGD gives the same result under the same initial conditions. Connections are established by explicit rules without variability.
Speed. Since the partner is known in advance, SYGD does not require a search procedure. Establishing a connection reduces to exchanging handshakes and recording the partner's name in the connectors. This is a fast operation, limited only by network latency.
Parallelism. SYGD between different pairs of nodes can proceed simultaneously. If an ANOD has created a thousand instances of class A and a thousand of class B and the template requires their pairwise connection, a thousand connections are established in parallel.
Start immediately after instantiation. SYGD does not wait for the completion of NRGN of all classes — it starts as soon as the first nodes whose partners are already known have been created. This yields the overlap of phases to which we return in section 4.6.
4. 5. SYGA: autonomous synaptogenesis
SYGA (SYnaptogenesis Autonomous) establishes those connections that are not explicitly defined in MOVE. It is the second mechanism of synaptogenesis and is responsible for the variable part of the architecture — connections whose concrete topology is determined at the moment of morphogenesis rather than being given in advance.
a) Connectors in the PEND state
After NRGN, not all connectors of new nodes are filled with partner addresses. Some of them are in the PEND (Pending) state — only the connector type CONT is known, but no partner name has been assigned. These connectors await resolution through SYGA.
Connectors in the PEND state are created for those connections that in the COMM diagram are marked as autonomous. For example: 'classes A and B are connected by connectors of CONT-X type with density 5 partners per node A'. The type of partner is specified here, but the concrete nodes of B to which each node of A will be connected are not yet defined — they will be found autonomously.
b) Hierarchical bulletin board
The SYGA mechanism is realized through a bulletin board of connectors in the ANOD. Each ANOD maintains in its G-section a data structure in which:
Nodes publish their PEND connectors with an indication of the CONT type they are seeking. This is equivalent to posting an advertisement 'seeking a partner of such-and-such a type'.
Nodes scan the bulletin board in search of complementary connectors — those whose partner type corresponds to their own. This is equivalent to searching for a matching advertisement.
When a match is found — two complementary-type connectors from different nodes — a connection is established. A handshake is exchanged, both connectors move from PEND to ACTV (Active), the partner names are recorded bilaterally.
The bulletin board is hierarchical. A local board inside one ANOD serves connections between nodes within one organ. If a partner is not found locally, the request is raised to the GATE level — search is performed among several ANODs of the same platform. If it is not found there either — it goes up to the inter-platform exchange level. This corresponds to the three layers of GORP routing (G1/G2/G3, see GNET).
c) Activity-dependent search
In its developed form, SYGA can take into account not only the type of connector but also the behaviour of the nodes — realizing what is called activity-dependent synaptogenesis. This means that, when searching for a partner, those nodes that exhibit some activity are preferred.
For example, a node of class A with a PEND connector for a connection to class B may prefer concrete B-nodes that are already actively exchanging data with other A-nodes. This corresponds to Hebb's rule in biology — connections between simultaneously active neurons are strengthened. Architecturally, it is realized as an additional criterion of choice among several possible partners.
Activity-dependent SYGA gives the network the capacity for self-organization on a functional principle: nodes that work together are connected with each other. This self-organization is an essential part of the formation of subjective reality and is unattainable through deterministic synaptogenesis.
d) Properties of SYGA
SYGA differs in principle from SYGD in several characteristics.
Emergence of the result. The concrete topology of the connections established by SYGA is not recorded anywhere before the procedure is completed. It is computed at the moment of morphogenesis and depends on the order of advertisements posted, on the availability of partners, and on the result of activity-dependent selection. Two instances of the same MOVE will give different concrete topologies in the SYGA-connected part.
Asynchrony. SYGA has no strict order of operations. The operations of publishing advertisements on the board and scanning the board are asynchronous, proceeding as availability permits. This yields robustness to temporary load fluctuations and to network latency.
Gradualness. SYGA does not complete in one pass. As new nodes are created (NRGN continues), new connectors appear on the board, and the search for partners continues. The overwhelming majority of connections are established in the initial period after NRGN, but some may remain to be resolved for a long time.
Possibility of remaining unresolved. Not every PEND connector necessarily finds a partner. If there are not enough complementary connectors in the network, some PEND connectors will remain unresolved. This is a normal state, not an error — it reflects the actual state of the network and can be corrected later through operational plasticity.
4. 6. Parallelism of phases
The phases of morphogenesis — parsing of MOVE, NRGN, SYGD, SYGA, maturation, verification — do not form a strictly sequential pipeline. They overlap in time: each subsequent phase begins before the previous one is fully completed.
a) Picture of overlap
In one ANOD, a complete cycle of morphogenesis proceeds approximately as follows. LOMN begins with parsing its part of MOVE — checking references to D-components, computing the need for AVEC. As soon as the parsing of the first class is completed, NRGN for that class starts — creation of the class node, resource evaluation, beginning of instantiation.
While instantiation of the first class proceeds, LOMN continues parsing the next class. When the first few instances of the first class are created, SYGD starts for them immediately — deterministic connections are established. If these connections require partners from another class not yet created, the corresponding connectors temporarily remain in PEND and are resolved as partners appear.
When several classes have been instantiated and part of the SYGD connections established, SYGA begins — nodes publish their PEND connectors on the board and search for autonomous partners. SYGA proceeds in parallel with the continued NRGN of the remaining classes.
When the main groups have been created and most connections established, maturation begins — nodes move to operation, continuing to resolve the remaining PEND on demand. SYGA does not stop immediately: it continues in background mode during maturation.
At this moment, the network is already working: the first groups of nodes exchange datagrams, perform their functions, while other groups are still forming. The network grows while working — as a biological embryo in the final stages of development.
b) Advantages of parallelism
The parallelism of phases has several essential advantages over the sequential model.
Reduction of total morphogenesis time. The complete unfolding cycle takes significantly less time than the sum of the individual phase durations. For large organisms this is important — their morphogenesis can take significant time even under parallel processing.
Gradual entry into operation. The network does not 'turn on' instantaneously — it gradually acquires functionality as groups are created and connected. The first organs to be working can begin to operate while others are still forming.
Activity-dependent correction. Since part of the network is already working while synaptogenesis is still proceeding, SYGA can use real activity to choose partners. This is impossible in the sequential model, where all connections are established before any function begins.
c) Biological parallel
The parallelism of phases is not an engineering optimization but an inheritance from biological morphogenesis. The development of a biological nervous system is organized exactly this way: neurogenesis, migration of neurons, formation of synapses, and the first network activity occur in overlapping waves, not sequentially.
The visual cortex begins to process signals from the first receptors long before all its neurons have been formed. This early activity participates in the formation of the cortex itself — through activity-dependent synaptogenesis. A strictly sequential model — 'first create everything, then connect everything, then start' — does not occur in biology at all.
Gativus inherits this principle. RTR0 is a parallel process, not a pipeline one, and exactly this parallelism guarantees the correspondence of the architecture with the biological model.
4. 7. Transformation OPNG → OPN
OPNG is the operational network of morphogenesis, specialized for morphogenesis itself. After morphogenesis is completed, it is no longer needed in this capacity: the organism has been built, and administrative functions are handled by OPN — the operational network corresponding to the running of the organism itself. The transition from OPNG to OPN is the fourth phase of germination (after parsing, NRGN/SYGD/SYGA, maturation) and concludes this cycle.
a) Gradual shift of balance
The transformation OPNG → OPN is not an instantaneous switch but the result of a gradual shift in the balance of activity. During maturation, LOMN engages less and less in neurogenesis (almost all classes have been instantiated) and more and more in coordinating the work of the organ: routing of datagrams, management of resources, response to requests from other ANODs.
At a certain moment, morphogenetic activity falls below a threshold: no new nodes are being created, no new SYGD connections appear, SYGA resolves only occasional remaining PEND. LOMN registers this transition and formally declares the completion of morphogenesis.
From this moment, LOMN continues to work but in another mode. It transforms from OPNG (operational network of morphogenesis) into OPN (operational network of operation) — from builder to administrator of the built organ.
b) What is preserved in the transformation
Not everything in OPNG disappears. A part of the functions is preserved:
Fragment of MOVE. LOMN continues to hold its part of MOVE as the ontogenetic dual of the organ it administers. This allows, if necessary, the reactivation of morphogenetic functions — for example, recovery after loss or induced neurogenesis.
Code of morphogenesis. The D-components implementing the NRGN, SYGD, SYGA procedures are preserved in the LOAI of the ANOD. They enter a dormant mode — not actively executed but accessible on demand.
Bulletin board. The SYGA data structures are preserved and continue to serve occasional new requests from existing nodes or from induced-creation nodes.
This stem-cell function — the ability to return to morphogenesis when necessary — is essential for the long-term sustainability of the system. Its biological analogue is the stem cells in the adult brain's hippocampus, capable of induced neurogenesis under certain conditions.
c) What changes in the transformation
The active functions of LOMN change essentially.
The queue of morphogenetic tasks is deactivated. LOMN no longer planfully schedules the creation of new nodes. The resolution of remaining PEND continues but in background mode, without active priority.
Routing of datagrams becomes the main task. The ANOD acts as the local gateway of its subordinate nodes: incoming datagrams from other organs pass through the ANOD and are distributed to the final nodes.
Management of AVEC enters operational mode. LOMN now manages not the distribution of NRGN quotas among classes but operational quotas: traffic limits, access policies, response to quota changes from the parent ANOD.
Security becomes the continuous concern of the G-section. During morphogenesis the G-section mainly validated creation requests; now its main task is control of the current traffic, detection of anomalies, immune response.
4. 8. Closure of the critical period
In biology there exist critical periods — time windows when morphogenesis proceeds actively, after which the formation of new structures is sharply curtailed. A visual cortex not stimulated during its critical period subsequently fails to develop fully. Gativus inherits this principle: the critical period of morphogenesis has a bounded duration, the closure of which transfers an organ into the operational mode with sharply lowered morphogenetic activity.
a) Mechanism of closure
The closure of the critical period in Gativus occurs through the zeroing of the NRGN and SYGN quotas in the AVEC₀ delegated to LOMN. After closure, LOMN can no longer create new nodes or establish new deterministic connections — the corresponding quotas are exhausted and are not replenished.
PEND connectors that remain unresolved after closure are not destroyed. They can be resolved later — but already not under LOMN's control, but through operational plasticity mechanisms which belong to the operation of the network rather than to its morphogenesis. Operational plasticity works with already existing nodes and does not require creating new ones; it is the subject of GNET, not MOGE.
b) Three causes of closure
The closure of the critical period may occur for one of three causes.
Internal cause. MOVE is exhausted. All classes have been instantiated, all deterministic connections established, all mandatory PEND resolved. The morphogenesis procedure has completed naturally. LOMN itself records the completion and enters STANDBY.
External cause. The parent ANOD or GATE sends a command on the completion of morphogenesis. This may be related to inter-organ redistribution of resources, planned completion of a certain phase of development, or response to a change in external conditions. LOMN receives the command and moves to STANDBY regardless of whether its own MOVE is exhausted.
Timeout cause. No requests for neurogenesis have arrived for a given period. The GATE scheduler records the stagnation and moves LOMN to the dormant mode to free resources. This is an insurance mechanism for cases of stalled morphogenesis for unforeseen reasons.
c) Reactivation
The closure of the critical period does not mean an irreversible end to the possibility of morphogenesis. LOMN can be reactivated upon the restoration of AVEC₀ quota and its re-delegation by the GATE. This is an analogue of induced neurogenesis in the hippocampus — a rare event requiring a decision at the level of the GATE or higher ANOD.
Reactivation may become necessary in several scenarios. Recovery after loss — if a substantial part of the organ's nodes is damaged, replacements need to be created. Adaptation to changed conditions — if functional requirements for the organ change significantly, partial reconstruction may be necessary. Evolution — if ROOT issues an updated version of D-components or an updated version of MOVE, reactivation allows applying that update to existing organs.
Under ordinary conditions, however, reactivation remains a rare event. Most organs after the closure of the critical period work for a long time in operational mode without essential morphogenetic changes, relying on operational plasticity for adaptive corrections.
4. 9. Transition to Chapter 5
This chapter has described RTR0 as the procedure of morphogenesis. The atomic and composite units of morphogenesis have been introduced: MORN as the triad class — template — group, isomorphic to the atomic units of the other transformations; MLOM as the recursive composite of MORNs, forming the hierarchy micro-column → column → organ → organism.
The three-step procedure of NRGN has been described: creation of the class node, evaluation of AVEC resources, instantiation of objects. It has been established that SYGD — deterministic synaptogenesis by the class's connector templates — starts in parallel with instantiation. The alternative mechanism of autonomous synaptogenesis SYGA has been described — through a hierarchical bulletin board of PEND connectors.
The principled parallelism of phases has been fixed — the overlap in time of parsing, neurogenesis, synaptogenesis and maturation. This overlap is not an optimization but an inheritance from biological morphogenesis, ensuring the activity-dependent self-organization of the network.
The transformation OPNG → OPN has been described — the gradual shift of balance from building to operation. The preservation of the stem-cell function in dormant mode ensures the possibility of reactivating morphogenesis when necessary. The closure of the critical period fixes the completion of the main morphogenesis and passes adaptive functions to operational plasticity.
Open remains the question of the architectural organization within which all this occurs. Where is GERM placed? Who manages the splitting of MOVE between functional organs? How is the hierarchy of nodes coordinating morphogenesis organized? These questions are the subject of the next chapter. Chapter 5 describes the architecture of morphogenesis: the three-level hierarchy ROOT → GATE → ANOD, the three-section structure of ANOD, the splitting of MOVE upon delegation, and fixes the boundary between MOGE and GNET.
Contents