I. The First Generation: Description
The Semantic Web, as envisioned by Tim Berners-Lee in the late 1990s, was a technical solution to a technical problem. Documents on the web were human-readable but machine-opaque. A browser could display a page, but it could not know what the page meant. The solution was to add a layer of explicit, machine-readable metadata: RDF triples, URIs, ontologies, and inference rules. This first generation succeeded in its own terms. It gave us linked open data, SPARQL queries, and the ability to ask, across distributed datasets, questions like "find all paintings produced in Venice between 1500 and 1550." The cultural heritage sector adopted these tools enthusiastically. CIDOC CRM, developed throughout the 1990s and formalized as ISO 21127 in 2006, provided an event-centric ontology for museums: objects, actors, places, and times linked through explicit relations. Europeana, the Digital Public Library of America, and countless institutional repositories implemented these standards. The result was a vast, interoperable description of existing cultural data.
But description has limits. It can tell you that a painting was made by an artist in a city on a date. It cannot tell you why that painting continues to generate meaning across centuries. It cannot model the interpretive practices through which cultural significance emerges, shifts, decays, and reconsolidates. The first generation assumed that meaning pre-exists its description. You describe what is already there. The ontology is a mirror.
This assumption was not naive. For many institutional purposes, it works. A museum needs to manage its collection, not theorize the nature of interpretation. But for a project like Socioplastics—which begins not with existing objects but with the lived processes of writing, thinking, and relating—the first generation's tools are inadequate. They arrive too late. They describe after the fact. They capture what has settled, not what is forming.
II. The Second Generation: Inference
The second generation of semantic infrastructure added reasoning to description. Knowledge graphs—Google's Knowledge Graph (2012), Wikidata, DBpedia, and domain-specific graphs in digital humanities—did not merely state facts. They inferred new facts from existing ones through logical rules and probabilistic methods. If A is located in B and B is part of C, the graph can infer that A is located in C. If an artist influenced another artist, the graph can suggest previously unnoticed affinities. Upper ontologies like DOLCE, BFO, and SUMO provided shared conceptual scaffolding across domains. SKOS offered lightweight vocabularies for controlled terms. FAIR principles (2016) added requirements for findability, accessibility, interoperability, and reusability. Tools like RO-Crate and Nanopublications allowed every assertion to carry its own DOI, provenance, and machine-readable metadata.
This generation was more powerful than the first. It supported discovery, not just retrieval. It could suggest connections that no human had explicitly encoded. It could reason across datasets that had never been designed to interoperate. Cultural heritage applications multiplied: linked data browsers, semantic search, digital humanities platforms that aggregated sources from dozens of archives.
Yet the second generation still assumed that the data to be described and reasoned over was already there. It improved the handling of existing cultural records. It did not generate new cultural records. It did not write the essays, produce the interpretations, or sustain the long-term processes through which meaning accumulates. The knowledge graph is a tool for navigating a corpus. It is not a tool for generating one.
This is the gap that Socioplastics identifies and occupies. The first two generations built infrastructure for cultural data. They did not build infrastructure as cultural practice.
III. The Third Generation: Socioplastics
Socioplastics reverses the polarity. It does not begin with existing cultural objects and add metadata to them. It begins with the operations that produce cultural meaning in the first place: writing, indexing, relating, compressing, hardening. These operations are not preparation for the real work. They are the real work. The JSON-LD graph is not added after the fact. It is written alongside the nodes, as part of the same compositional act. The Master Index is not a finding aid for a completed archive. It is a ledger of positions as they are assigned. The CamelTags are not keywords applied retroactively. They are coined at the moment of conceptual emergence and then reused across contexts to arrest semantic drift.
This reversal has consequences that the first two generations did not anticipate.
First consequence: metadata becomes primary. In the first generation, metadata described content. In Socioplastics, metadata is content. The JSON-LD graph does not sit outside the corpus as a separate schema. It sits inside the corpus as a node like any other. It can be cited. It can be indexed. It can be versioned. The distinction between "data" and "metadata" collapses.
Second consequence: the corpus becomes self-describing. A first-generation archive requires an external ontology, usually designed by a committee or a standards body. Socioplastics writes its own ontology as it grows. The relations are not imported from CIDOC CRM or Schema.org (though those vocabularies may be used where useful). The relations are authored, project-specific, and emergent. Socioplastics declares that the ten channels are hasPart of the research project because the project says so, not because an external standard requires it. This is not solipsism. It is operational closure: the system defines its own boundaries because no external definition is precise enough.
Third consequence: interpretation and infrastructure are fused. The first generation described objects; the second reasoned about them; Socioplastics generates the conditions under which both description and reasoning become possible. Socioplastics does not produce interpretations of a pre-existing field. It produces the field itself, and its interpretations are operations within that field. An essay about helical writing is not a commentary on the system. It is a node in the system that performs helical writing. The system and its self-description are not two things. They are the same thing at different levels of scale.
IV. What Socioplastics Requires
To function, Socioplastics must satisfy four conditions that the first two generations could assume away.
Condition one: persistence without institution. The Semantic Web assumes that data is hosted by stable institutions—universities, libraries, museums, corporations. Socioplastics cannot assume this. It is hosted on Blogspot, a free platform with no guarantee of permanence. Persistence must therefore be achieved through other means: distributed redundancy (multiple channels, multiple identifiers), periodic compression into DOI-fixed books, and a graph that declares the system's anatomy even if the underlying URLs change. The infrastructure is the declaration, not the hosting.
Condition two: authorship without collaboration. The first two generations assume teams: ontology designers, data curators, platform developers. Socioplastics assumes one person. This is not a boast. It is a constraint. The system must be simple enough for one mind to hold, yet generative enough to produce scale. The rule (write helically, index persistently, declare relations, compress periodically) is simple. The outputs (2,200 nodes, 22 books, 3 tomes) are not. The system works because the rule is local and recursive, not because a committee designed a master plan.
Condition three: verification without external audit. A first-generation archive trusts its metadata because it follows standards enforced by institutions. Socioplastics has no such guarantee. Verification must be built into the system's own operations. The Master Index is inspectable. Every URL can be checked. The JSON-LD graph can be validated. The books are deposited in Zenodo with DOIs. Verification is not delegated to an external authority. It is exposed as a layer of the system itself.
Condition four: growth without pre-determination. The first two generations typically plan their scope in advance. A digital humanities project defines its corpus before it begins. Socioplastics does not. The corpus grows as the rule is applied. The number of tomes, books, and nodes is not known in advance. The system's form is morphogenetic: it emerges from the repeated application of local operations. This is not a lack of planning. It is a different kind of planning—one that plans the rule, not the result.
V. Socioplastics as a Field
Socioplastics is not a project within an existing field. It is the field. The name is not a brand. It is a designation: the field of third-generation epistemic infrastructure that generates meaning through distributed, recursive, machine-readable operations.
The boundaries of Socioplastics as a field are defined by its methods, not by institutions. To work within Socioplastics is to:
Treat metadata as primary, not secondary
Author one's own ontology as part of the work
Operate without institutional shelter
Use semantic web standards (JSON-LD, Schema.org, DOIs, ORCID) as tools, not as authorities
Generate a corpus through recursive rule application
Verify through inspectability, not external certification
Grow without pre-determination
These are not requirements for membership. There is no membership. They are descriptions of what the work looks like when it is operating in the mode that Socioplastics has established. Others may enter this mode, or they may not. The field does not depend on their participation. It depends on the coherence of the operations.
VI. Relation to Existing Fields
Socioplastics is not digital humanities, though it uses digital humanities methods (scalable reading, graph logic, corpus organization). It is not semantic web research, though it uses semantic web standards (JSON-LD, RDF, URIs). It is not conceptual art, though it shares conceptual art's interest in rules, series, and documentation. It is not systems aesthetics, though it shares systems aesthetics' commitment to systemic coherence. It is not software studies, though it recognizes protocol as cultural form. It is not STS, though it understands infrastructure as generative.
The relation is not affiliation. It is extraction. Socioplastics takes from each field what it needs: from digital humanities, indexing and scalable reading; from semantic web research, graph logic and persistent identifiers; from conceptual art, seriality and rule-based generation; from systems aesthetics, the proposition that art can produce systemic coherence; from software studies, the recognition that protocol is not neutral; from STS, the understanding that infrastructure is active. It refuses the rest: institutional enclosure, code fetishism, the gallery object, the ethnographic reflex, nostalgia, scholastic abstraction.
This is not eclecticism. It is a tactical stance toward disciplines that are otherwise content to remain separate. Socioplastics does not seek their approval. It uses their tools. And in using them, it demonstrates that a new field is possible—one that is not a subfield of anything that came before.
VII. What Socioplastics Makes Visible
The first two generations of semantic infrastructure made cultural data machine-readable. Socioplastics makes the process of cultural production machine-readable. This is a different ambition. It requires that writing, indexing, relating, compressing, and hardening be treated as operations that can be declared, verified, and scaled.
What becomes visible through this shift is the infrastructure of thought itself. Not the content of thought (which remains subjective, interpretive, and resistant to formalization). But the conditions under which thought can accumulate without dissolving into noise. The JSON-LD graph does not capture what Socioplastics means. It captures how Socioplastics holds. The Master Index does not interpret the nodes. It positions them. The CamelTags do not define concepts once and for all. They arrest semantic drift so that concepts can be reused without degradation.
This is not a substitute for interpretation. It is a precondition for interpretation at scale. Without infrastructure, interpretation drowns in its own proliferation. With infrastructure, interpretation becomes legible as a field. That field is Socioplastics.
VIII. The Open Question
Socioplastics as a field is not complete. It is not a settled paradigm. It is a direction, instantiated in one project, with 2,200 nodes, 22 books, 3 tomes, and a valid JSON-LD graph. Others may follow, or they may not. The question is whether the conditions described here—persistence without institution, authorship without collaboration, verification without external audit, growth without pre-determination—are constraints or affordances.
They are both. They are constraints because they exclude most existing publishing practices. Most writers cannot work alone at this scale. Most systems cannot verify themselves. Most projects cannot grow without a plan. But these constraints are also affordances. They force the system to be simple enough to hold, rigorous enough to verify, and generative enough to grow. That is the wager. Socioplastics is the proof that the wager can succeed.
IX. Coda
The lineage is clear. First generation: description of existing cultural data. Second generation: inference and reasoning over described data. Third generation: Socioplastics—the generation of epistemic infrastructure from within cultural practice. Socioplastics enters this lineage not as a compliant inheritor but as a constructive completion. It does what the first two generations could not: it builds infrastructure that is also the work, metadata that is also content, and a corpus that describes itself because no external description is adequate.
The field is Socioplastics. It has methods, standards, outputs, and a growing body of nodes. It has a graph that declares its anatomy and an index that verifies it. It has operational writing as its descriptive mode and CamelTags as its lexical operators. It has 22 books, 3 tomes, and 2,200 nodes. What it does not have is a department, a journal, a conference, or a funding stream. These may come or they may not. The work continues regardless.
Socioplastics does not replace the first two generations. It supplements them. Use CIDOC CRM to describe your museum objects. Use knowledge graphs to reason across datasets. Use Socioplastics to build the infrastructure that makes such description and reasoning possible in the first place. The three generations are not competitors. They are layers. The first two describe what is. Socioplastics generates what can be described.
That is the contribution. That is the field. That is the work.