Structural and Framed Primes¶
A Typology for Cross-Domain Portability of Abstractions¶
A conceptual paper for the Encyclopedia of Abstractions project
Abstract¶
A prime abstraction is operationally defined as one applying across at least three domains of human knowledge. The threshold settles a binary question — is this prime? — but leaves untouched a further asymmetry: some primes slot into new domains as pure relational structure, while others import an interpretive context (vocabulary, evaluative commitments, institutional assumptions) along with them. This paper develops the refinement that names that asymmetry: structural primes were already domain-stripped at the moment they were named; framed primes carry an interpretive overlay that resists clean extraction. The distinction admits degree, correlates with the disciplinary character of a prime's home domain, and predicts that "super-primes" (those applying across nearly all domains) cluster on the structural end. A decomposition operation extracts a structural core from a framed prime, yielding one of three outcomes: unification with an existing structural prime, a new analytical tool, or loss of identity (a fate related to Bernard Williams's account of thick ethical terms). A bidirectional dynamic — structural primes acquiring frames when deployed in normative contexts — complicates the basic distinction productively. The paper proposes schema additions for the catalog and outlines a research agenda for testing the typology empirically.
1. Introduction¶
The Encyclopedia of Abstractions is built around an operational definition: a prime abstraction is an abstraction that applies meaningfully across at least three domains of human knowledge. The definition is practical and does real work. It draws a usable line between abstractions narrow enough to be discipline-bound and those broad enough to support cross-domain transfer, and it does so without committing to any particular metaphysics of abstraction. For the purposes of building a corpus, this is exactly what an operational definition should do.
But the definition is binary, and the phenomenon it picks out is not. Once an abstraction is classified as prime, the definition is silent about how it behaves when applied to a new domain. In practice, those behaviors differ substantially. Apply feedback to a new domain and what travels with the word is a relational pattern: a return signal modulating a controlling input. The pattern is either present in the new domain or it is not, and the answer doesn't depend on whether the new domain shares anything with cybernetics, the discipline that gave feedback its modern name. Apply sovereignty to a new domain and something quite different happens. The relational pattern of sovereignty — bounded authority over a defined region with power to exclude external interference — does travel, but it travels with companions. The political-legal frame in which sovereignty was developed comes along: questions of legitimacy, jurisdiction, the right to recognition. When we speak of "data sovereignty" or the "sovereignty of the user," we are not just identifying a relational pattern in a new domain; we are recasting that domain in terms whose interpretive weight was forged in political theory.
This paper is a conceptual response to that asymmetry. It develops a typology that distinguishes primes by how they travel, articulates a decomposition operation that exposes the structural patterns inside framed primes, and traces the consequences for both the Encyclopedia of Abstractions and the broader practice of cross-domain abstraction transfer.
The paper is conceptual rather than empirical. The framework is fresh — it emerged in dialogue with a large language model in approximately the form presented here, over the course of a single sustained conversation — and no decomposition study or inter-rater reliability test has yet been conducted. The goal is to develop the framework with sufficient precision to support that empirical work as a natural next step. The paper engages with several existing scholarly traditions, not to position itself as a contribution to any one of them, but to deepen the framework by tethering its parts to concepts that have already done analytical work in adjacent territory. The primary audience is the author's own continued thinking about the Encyclopedia of Abstractions project; secondary audiences are practitioners and theorists who want to apply or extend the encyclopedia.
The paper makes four contributions.
The first is the structural/framed distinction itself, with diagnostic criteria for classification and an account of how the distinction maps onto a spectrum rather than a binary partition. I argue that the placement of a prime on the spectrum correlates with the disciplinary character of its home domain — formal disciplines like mathematics, physics, information theory, and systems science tend to produce structural primes, while institutional and normative disciplines like law, governance, and ethics tend to produce framed primes. I also identify a corollary about super-primes — primes that apply across nearly all domains of human knowledge. Super-primes cluster heavily on the structural end of the spectrum, which is what the framework predicts: less context to shed means easier travel.
The second contribution is the decomposition operation and the three-fates typology. Given any framed prime, one can attempt to extract a structural core by abstracting away the institutional or normative overlay. The operation has three possible outcomes. It can unify the framed prime with an existing structural prime in the corpus, revealing the framed prime as a context-enriched specialization of a more general pattern. It can produce a new analytical tool — a structural pattern not previously named in the corpus, which can then project into domains where the original framed prime would have looked alien. Or it can result in loss of identity, where the extraction yields something so thin that the analytical weight of the original prime has been stripped out, indicating the prime is constitutively framed. The typology generates testable predictions about how decomposition outcomes correlate with the prime's properties.
The third contribution is the observation of bidirectional deployment dynamics. The structural-versus-framed character of a prime is not entirely intrinsic to the prime. Structural primes can acquire frames when deployed in normative or institutional contexts. Hierarchy, structural at definition, picks up frames about power and dominance in social deployment. Equilibrium, structural in physics, picks up frames about welfare and efficiency in economic deployment. This means the structural/framed distinction has both an intrinsic dimension (the prime's character at origin) and a deployment dimension (the frames it accretes in use). Tracking frame pickup matters for the corpus and for any practitioner trying to determine what assumptions are being smuggled across domain boundaries.
The fourth contribution is practical. The structural/framed distinction has direct consequences for how the Encyclopedia of Abstractions should be organized and for how LLM-assisted discovery of cross-domain applications should be evaluated. Structural prime applications to new domains are recognitional and admit relatively cheap, reliable validation. Framed prime applications are interpretive and require more careful judgment. The two require different workflows, different validation regimes, and different quality-control checks. I propose specific schema additions to the encyclopedia and a workflow for prime application that respects these differences.
The remainder of the paper proceeds as follows. Section 2 surveys five existing scholarly traditions that touch on adjacent territory: thick and thin ethical concepts, analogy and structure-mapping, essentially contested concepts, conceptual engineering, and boundary objects. Section 3 develops the core structural/framed distinction with diagnostic criteria and worked examples. Section 4 introduces the decomposition operation and the three-fates typology with detailed worked decompositions. Section 5 traces the bidirectional dynamic of frame pickup in structural prime deployment. Section 6 develops the epistemological perspective on what kinds of knowledge claims are being made when primes are applied across domains. Section 7 develops the pragmatic perspective, particularly for LLM-assisted abstraction discovery. Section 8 proposes schema and workflow changes for the Encyclopedia of Abstractions. Section 9 considers limitations and counter-arguments. Section 10 outlines an empirical research agenda. Section 11 concludes.
2. Background and Related Work¶
The structural/framed distinction has cousins in several existing scholarly traditions. Engaging with them serves two purposes. The first is honesty: the framework is not entirely new, and surfacing what it shares with prior work clarifies what it adds. The second is depth: the existing traditions have done analytical work that the framework can borrow from, particularly in cases where they have already addressed objections or developed concepts that map cleanly onto pieces of the typology developed here.
2.1 Thick and Thin Ethical Concepts
The closest philosophical cousin to the structural/framed distinction is the contrast between thick and thin ethical concepts, articulated most influentially by Bernard Williams in Ethics and the Limits of Philosophy (1985)[1] and elaborated by John McDowell[2], Hilary Putnam, and others. The distinction concerns the relationship between descriptive and evaluative content in ethical terms.
Thick concepts — cruel, brave, generous, kind, cowardly, just — combine descriptive and evaluative content in a way that resists clean separation. To call an act cruel is both to describe it (it involves the deliberate infliction of suffering) and to evaluate it (cruelty is bad). The evaluation is not a separate judgment we attach to a neutral description; it is constitutive of the concept's meaning. Thin concepts — good, right, ought, bad — are more abstract. They carry evaluative weight but very little descriptive content; what counts as good or right is highly context-dependent and varies across situations.
Williams's argument was that thick concepts are not eliminable in favor of thin concepts plus neutral descriptions. The descriptive and evaluative content cannot be peeled apart. McDowell pressed this point in connection with cognitive contact: only someone who has internalized the evaluative perspective associated with cruelty can reliably identify cruel acts, because the descriptive criteria themselves require evaluative discrimination.
The structural/framed distinction generalizes this picture beyond ethics. A framed prime is to a thick concept what a structural prime is to a thin concept, mutatis mutandis. The frame in a framed prime is doing analogous work to the evaluative content in a thick concept: it is partly constitutive of the prime's meaning, it cannot be cleanly stripped without changing what the concept is, and applying the prime requires internalizing the frame. The structural core of a framed prime, if extractable, plays a role analogous to the descriptive residue of a thick concept — and in the most extreme cases (fate three in the typology developed below), the residue is too thin to do the work the original concept was doing, exactly as Williams argued about thick ethical terms.
The generalization is not trivial. The thick/thin distinction in ethics has been heavily worked over, and a substantial literature exists on what makes ethical concepts thick, on whether the thickness can be partially decomposed in some cases (Pekka Väyrynen's recent work argues for a more pragmatic account[3]), and on cross-cultural variation in thick concepts. The structural/framed distinction can borrow heavily from this work, particularly when it comes to the third decomposition fate. But the framework also extends the thick/thin idea in ways that the ethics literature has not pursued. Thick concepts in ethics are about the relationship between description and evaluation. Framed primes can carry frames that are not primarily evaluative — institutional, procedural, vocabulary-based, evidential. The frame in a contract, for example, is partly evaluative (a contract is binding, and binding is normatively loaded) but also institutional (a contract presupposes a system of enforcement) and procedural (a contract has structural elements like offer, acceptance, consideration). Framed primes are thick in something like Williams's sense, but the thickness comes in multiple flavors, only one of which is the descriptive/evaluative entanglement that the ethics literature has focused on.
2.2 Analogy and Structure-Mapping
The cognitive science of analogy provides a second adjacent tradition. Dedre Gentner's structure-mapping theory (1983, 1989)[4][5] offers a precise account of how analogical transfer works empirically. In the theory, analogies map relational structure between a source domain and a target domain. Surface features — particular objects, contexts, vocabulary — may or may not transfer, but the analogical content is the system of relations. The famous Rutherford-Bohr analogy between the solar system and the atom maps the relational structure (a small massive object at the center with smaller objects orbiting it because of attractive force) while leaving behind the surface features (the planets, the sun, the colors of the heavens). Douglas Hofstadter and Emmanuel Sander, in Surfaces and Essences (2013)[6], develop a broader account of analogy as the core engine of cognition itself, with much the same emphasis on relational structure as the analogical payload.
Structure-mapping theory provides a vocabulary the structural/framed distinction can borrow from directly. A structural prime is precisely an abstraction whose entire content is relational structure — by construction, it has no surface features to leave behind because it was already domain-stripped at the moment it was named. A framed prime, by contrast, is an abstraction whose relational structure is wrapped in surface features (vocabulary, institutional context, evaluative connotations) that travel with it whether or not they fit the new domain. Applying a framed prime to a new domain therefore involves an implicit decision about what counts as surface and what counts as structure — a decision that the structural/framed typology makes more explicit.
This connection raises a question that Kurt has been thinking about and that deserves direct engagement: analogies break when pushed too far. The Rutherford-Bohr analogy fails as a description of the atom because electrons do not in fact orbit the nucleus like planets — the analogy maps surface features that turn out to mislead. Do prime abstractions break in the same way?
The honest answer is: it depends on the prime, and the structural/framed distinction predicts the dependency. Structural primes do not break in the analogy sense. A structural prime is the relational pattern itself, with no surface features to mislead. The pattern is either present in the target domain or it is not. There is no question of "pushing the analogy too far" because there is no analogy in the surface-feature sense; there is only a recognition of whether the pattern is instantiated. What structural primes do is fail by vacuity. A structural prime may be trivially present in a target domain in a way that does no analytical work — every system has "feedback" if you squint hard enough at any return relationship. Vacuity is a real failure mode but it is different from analogy breakdown. It is a failure of meaningful application, not a failure of fit.
Framed primes, on the other hand, can break in something closer to the classical analogy sense. The frame imposes assumptions, vocabulary, and evaluative commitments that may or may not fit the new domain. When the fit is poor, the framed prime fails not by being absent but by being misleading. Applying "sovereignty" to a multi-agent AI system might illuminate certain features (the boundary of the agent's authority, the question of who can intrude) but mislead about others (the AI does not have a legitimacy claim in the political-theoretic sense, and treating it as if it does imports assumptions that don't hold). This is closer to the analogy-breakdown picture: the frame transfers surface features that turn out not to fit.
A useful corollary follows. The more framed a prime is, the more its application behaves like analogy in the cognitive-science sense — and the more vulnerable it is to the failure modes that the analogy literature has analyzed (e.g., false confidence in mapped features, missed disanalogies, surface features that pose as structural). The more structural a prime is, the more its application is closer to pattern recognition or model checking — and the relevant failure mode is vacuity rather than breakdown. This corollary has implications for evaluation that the pragmatic section of this paper develops further.
2.3 Essentially Contested Concepts
W. B. Gallie's 1956 paper "Essentially Contested Concepts"[7] identifies a class of concepts whose proper application is constitutively disputed. Democracy, art, social justice, religion, and Christianity are Gallie's paradigm cases. These concepts share features: they are evaluative, internally complex, variously describable, openly evaluative across competing accounts, and the disputes about their application are not resolvable by purely descriptive or empirical means. The disputes are part of the concept's meaning rather than a sign that the concept is being misused.
Essentially contested concepts are relevant to the structural/framed distinction in two ways. First, they help characterize the third decomposition fate — loss of identity. When a framed prime is essentially contested in Gallie's sense, decomposing it into a structural core plus a frame is bound to fail in a particular way: the contestedness is part of what the prime is doing, and stripping it produces something that does not do that work. Justice in the strict Gallie sense is essentially contested; what counts as just is part of what people argue about when they use the term. The structural core "balanced allocation according to a rule" is a real pattern, and it shows up in many domains, but it has lost the contested-evaluation function that made justice (in the institutional and moral sense) the concept it is. The structural core is a structural prime, but it is not justice; it is something like equilibrium-with-correction, a different abstraction entirely.
Second, Gallie's framework provides a tool for distinguishing fate-3 framed primes from fate-2 framed primes. Not all framed primes are essentially contested. Sovereignty has a frame, but the frame is institutional rather than essentially contested — sovereignty's application can be disputed at the edges, but reasonable people can agree on core cases without that agreement undermining the concept. Sovereignty therefore admits structural extraction (fate 2). Justice, by contrast, is contested in a way that makes any structural extraction lose the analytical work the original was doing (fate 3). The distinction between fate 2 and fate 3 may track, in part, the distinction between framed primes whose frame is merely institutional and framed primes whose frame is essentially contested.
This connection between fate 3 and essentially contested concepts is a hypothesis worth testing empirically. If it holds up, the typology gains an external diagnostic: framed primes that are essentially contested in Gallie's sense should reliably produce fate-3 decompositions, while framed primes with merely institutional frames should produce fate-1 or fate-2 decompositions. If the correlation is weak, the typology and the contested-concepts literature are tracking different phenomena, and the relationship between them needs to be rethought.
2.4 Conceptual Engineering
Conceptual engineering is the most directly relevant contemporary philosophical movement. Herman Cappelen's Fixing Language (2018)[8], David Plunkett and Alexis Burgess's work in Conceptual Engineering and Conceptual Ethics (2020)[9], and the broader literature have developed an account of how concepts can be deliberately reshaped to serve new theoretical or practical purposes. The movement asks: what is the concept of X supposed to do, what is it currently doing, and what would it do if we modified it? It treats concepts as tools that can be designed, redesigned, and replaced rather than as natural kinds we discover.
The Encyclopedia of Abstractions is, in essence, a large-scale conceptual engineering project. It assembles a curated set of abstractions for the explicit purpose of supporting cross-domain transfer and discovery. The decomposition operation introduced in this paper is itself a conceptual engineering move: it deliberately reshapes a framed prime by separating out its structural core and its frame, in order to make the core available for new applications and to make the frame's assumptions explicit.
A central concern in conceptual engineering is topic continuity: when we re-engineer a concept, is the result still a version of the same concept, or have we replaced it with something else? Cappelen has argued that strict topic continuity is impossible to define in any way that aligns with how concept revision actually works in practice. The structural/framed distinction's three-fates typology can be read as offering a partial answer in the specific case of framed primes. Fate 1 (unification) is topic-preserving in the sense that the structural core and the frame can be reattached to recover the original concept. Fate 2 (new analytical tool) is topic-preserving in a weaker sense: the original framed prime can still be discussed as the structural core plus its frame, but the structural core has gained a life of its own. Fate 3 (loss of identity) is topic-breaking: the structural core does not, on its own, do the work of the original prime, and trying to substitute the core for the prime would change the topic of any discussion in which the prime was being used.
The conceptual engineering literature also raises a methodological point: the structural/framed distinction and the decomposition typology are themselves engineered concepts, and they should be evaluated for whether they do the work they are supposed to do. The pragmatic section of this paper takes that question seriously. The framework's value is not in being a true description of how primes secretly are; it is in being a useful organization of how primes behave when used.
2.5 Boundary Objects
Susan Leigh Star and James Griesemer introduced the concept of boundary objects in their 1989 paper on the Berkeley Museum of Vertebrate Zoology[10]. Star later returned to the concept in a reflective essay charting how it has been used (and misused) in the decades since[11]. A boundary object is an entity (a concept, a document, a specimen) that maintains enough coherence to be shared across different communities of practice while being interpreted differently within each. Boundary objects are productive precisely because of this dual character: they enable coordination across communities without requiring those communities to agree on a single interpretation.
Framed primes can function as boundary objects when they pass between domains. The framed prime "contract" can coordinate work between lawyers, software engineers, and biologists discussing mutualism — each community interprets contract through its own frame (legal enforcement, API specification, evolutionary fitness consequences), but the shared structural core supports productive communication. This is a productive use of framed primes, distinct from the failure mode where assumptions get smuggled across boundaries without being scrutinized. The difference is that in productive boundary-object use, the multiple interpretations are acknowledged and the structural core is doing the coordination work; in the smuggling failure mode, one interpretation is silently exported into a domain that doesn't warrant it.
The boundary objects literature has developed criteria for what makes a boundary object work well. The most important are interpretive flexibility (different communities can read different things into it) combined with structural stability (there is a shared core that all communities are responding to). These map almost directly onto the framed prime structure: the structural core provides the stability, the frame provides interpretive flexibility when communities can swap their own frames in. When the frame is rigid and tied to one community of practice — as it often is for framed primes whose home discipline has a strong institutional vocabulary — the prime functions less well as a boundary object and more like a one-way transfer from the home discipline to the receiving one.
2.6 What This Paper Adds
The five traditions just surveyed have done substantial analytical work on adjacent territory. The structural/framed distinction borrows from each. What does it add?
First, it generalizes the thick/thin distinction beyond ethics into all domains of human knowledge, in a form that respects the differences among kinds of frames (evaluative, institutional, procedural, vocabulary-based). Second, it connects the analogy and structure-mapping literature to a typology of abstractions, providing a vocabulary for distinguishing primes whose application is recognitional from those whose application is more like analogy projection. Third, it formalizes the decomposition operation and its three fates in a way that makes empirically testable predictions about how primes behave when their frames are stripped — a move that the conceptual engineering literature has not made systematically. Fourth, it identifies the bidirectional deployment dynamic, in which structural primes pick up frames in normative contexts, a phenomenon that has not, to my knowledge, been characterized as such in the existing literatures. Fifth, it makes the entire framework practical by tying it to a concrete project (the Encyclopedia of Abstractions) and developing implications for both schema design and discovery workflow.
The contribution is therefore both synthetic and extensive. The structural/framed distinction draws together threads from five literatures and adds extensions that none of them developed because none of them was working on the specific question of how abstractions travel between domains in a way that supports curated cross-domain transfer at scale.
3. The Core Distinction: Structural and Framed Primes¶
This section develops the structural/framed distinction with definitions, diagnostic criteria, and worked examples. The aim is to make the distinction precise enough to be applied to specific primes with reasonable inter-rater agreement.
3.1 Definitions
A structural prime is a prime abstraction whose meaning is constituted entirely by relational structure, independent of any particular substrate or interpretive frame. The prime can be defined and recognized without reference to the particular institutional, normative, or vocabulary context in which it was first named. Applying a structural prime to a new domain is therefore a recognition operation: the relational pattern is either instantiated in the domain or it is not.
A framed prime is a prime abstraction whose meaning is partly constituted by an institutional, normative, or contextual interpretive frame that cannot be cleanly stripped from the relational structure. The frame is not a coincidence of how the prime happened to be discovered; it is part of what the prime is. Applying a framed prime to a new domain is a projection operation: the prime brings its frame with it, recasting the new domain in vocabulary and assumptions borrowed from the original.
The distinction admits degree. Most primes fall somewhere between the pure poles. The spectrum is not a continuous measurable quantity but a useful gradient for classification.
3.2 Diagnostic Criteria
The following five questions form a diagnostic checklist for placing a prime on the structural/framed spectrum. They are not strict definitional criteria but indicators; a prime that scores "framed" on most of them is framed, one that scores "structural" on most is structural, and one that scores mixed sits in the middle.
The first question is whether the prime's vocabulary travels with it when applied to a new domain. Framed primes typically import a recognizable lexicon from their home discipline. Sovereignty, applied to data, brings jurisdiction, recognition, interference. Contract, applied to APIs, brings binding, enforcement, breach. Structural primes typically do not import their home discipline's vocabulary in the same way. Feedback, applied to ecology, does not import cybernetics-the-discipline; it just identifies a pattern.
The second question is whether the prime carries evaluative weight by default. Framed primes typically do: justice, dignity, due process, even sovereignty, have evaluative or normative connotations baked in. Structural primes typically do not: symmetry, recursion, hierarchy at definition are neutral.
The third question is whether the prime has an institutional or normative referent at its origin. Framed primes typically arose in disciplines whose work involves human institutions or normative practices. Structural primes typically arose in disciplines whose work is formal — mathematics, physics, information theory, systems science.
The fourth question is whether the prime can be defined without reference to human practices. Structural primes can — feedback is defined in terms of return signals modulating inputs, with no reference to humans. Framed primes typically cannot — contract is defined in terms of agreements, which presuppose agents capable of entering agreements, which is a human-practice-bound concept.
The fifth question is whether applying the prime in a new domain feels like identifying a pattern that was already there or like importing a perspective that recasts the domain. Structural primes identify; framed primes import. This question is partly phenomenological but tracks something real about the prime's mode of application.
Applied to a few primes:
Symmetry. Vocabulary doesn't travel (no math discipline-specific terms come along when symmetry is applied to a face or a piece of music). Evaluative weight is absent at definition. Origin in mathematics, no institutional referent. Definable without reference to human practices. Application is recognitional. All five indicators say structural. Symmetry sits near the structural pole.
Sovereignty. Vocabulary travels heavily. Evaluative weight is present. Origin in political theory and law, institutional referent throughout. Cannot be defined without reference to human practices (or analogues to them). Application is interpretive, importing a perspective. All five indicators say framed. Sovereignty sits near the framed pole.
Selection. Vocabulary travels partially — terms like "fitness" and "pressure" come along in some applications but not others. Evaluative weight is mixed (selection in evolutionary biology is neutral; "natural selection of ideas" sometimes accretes evaluative tone). Origin in evolutionary biology, which is partly formal and partly institutional in its conceptual structure. Definable in terms that don't require human practices. Application varies between recognitional and interpretive depending on context. Mixed indicators; selection sits in the middle of the spectrum, leaning structural.
The diagnostic checklist is not mechanical, and applying it requires judgment. But it is reproducible enough to support inter-rater agreement studies and to give the framework substance.
3.3 The Spectrum View
Most primes are not pure structural or pure framed. They sit along the spectrum. Examples by approximate position:
Pure structural end: symmetry, equilibrium, recursion, scale, conservation, invariance.
Strongly structural: feedback, hierarchy, network, modularity, emergence, system.
Mid-spectrum, leaning structural: selection, game, information, threshold, coupling.
Mid-spectrum, leaning framed: market, exchange, norm, role, identity.
Strongly framed: contract, sovereignty, property, agency, authority, due process.
Pure framed end: justice, dignity, legitimacy, sanctity.
The positions are approximate and depend on which version of a prime is in view. Game in pure game theory is more structural; game in everyday usage picks up frames from sports, gambling, and competition. The spectrum view is useful precisely because it accommodates these variations.
3.4 The Origin Correlation
The placement of a prime on the spectrum correlates with the disciplinary character of its home domain. Formal disciplines mint structural primes. Institutional and normative disciplines mint framed primes.
The correlation is not accidental. Formal disciplines work with concepts that have been deliberately stripped of substrate at the level of definition — that is what makes them formal. When such a concept is identified as a prime, it travels easily because it had no substrate to carry along in the first place. Institutional and normative disciplines work with concepts that are constituted partly by reference to human practices, evaluations, and institutional arrangements. Concepts arising from this work cannot be detached from those practices without changing what they are.
This correlation has predictive force. Without examining a specific prime, knowing that it originated in mathematics or physics tells you to expect a structural prime; knowing that it originated in law or ethics tells you to expect a framed prime. The correlation is imperfect — biology produces both structural primes (selection, hierarchy as taxonomy) and framed-leaning primes (organism, function). But it is robust enough to serve as a heuristic.
3.5 The Super-Primes Corollary
A super-prime, in the working vocabulary of the Encyclopedia of Abstractions project, is a prime that applies across nearly all domains of human knowledge — not merely the three-domain threshold for primality, but something approaching universality. Examples include symmetry, hierarchy, network, equilibrium, feedback, and a small handful of others.
The structural/framed framework predicts that super-primes should cluster heavily on the structural end of the spectrum. This prediction appears consistent with the primes I have examined informally, though the observation rests on impression rather than systematic measurement. Most primes that approach universal applicability across human knowledge seem to be structural at origin and at definition. If the impression survives empirical scrutiny, super-prime status would be partly a consequence of structural character: a prime that had no substrate to carry along can travel everywhere; a prime that hauls a substrate with it can only travel where the substrate can be accommodated.
There appears to be one important caveat. A small set of primes seem to achieve high portability not by being structural but by carrying frames that are themselves nearly pan-human. Candidates for this second route include value, exchange, and purpose — each appears to travel widely while still hauling recognizable interpretive content. If these are correctly characterized, they achieve quasi-super-prime status via the universality of their frames rather than the structural emptiness of their cores. These cases would account for apparent exceptions to the structural-equals-super-prime correlation and would suggest that the route to high portability has at least two paths: low substrate, or high-universality substrate. The conjecture is testable by systematic examination of which primes cross what number of domains, paired with classification of how each prime's travel is achieved.
3.6 Hierarchy as a Case Study
Hierarchy is worth dwelling on because it exhibits several of the dynamics this paper is developing.
At definition, hierarchy is a pure structural prime. It names a relational pattern: nested levels with asymmetric relations between adjacent levels. The pattern is present whenever you have a structure that decomposes into parts that themselves decompose, with directionality between the levels. Hierarchy appears in mathematics (the hierarchy of cardinalities), in biology (taxonomy, organ-tissue-cell-molecule), in computer science (memory hierarchies, file systems, abstraction layers), in linguistics (phrase structure), in chemistry (molecular structure), in physics (energy levels). Across all these domains, the structural pattern is the same.
But hierarchy in social and organizational deployment picks up a frame. In organizations, hierarchy connotes power, authority, dominance, sometimes oppression. In social-political discussion, hierarchy is rarely neutral. The structural prime — nested levels with asymmetric relations — is the same one used in biology, but social deployment has accreted normative connotations that biology does not impose.
Hierarchy therefore illustrates the bidirectional deployment dynamic that section 5 develops in detail. The prime is intrinsically structural but can be deployed in contexts that attach a frame. The same prime can be neutral in one context and frame-laden in another. The structural/framed character is partly the prime's and partly the deployment's.
This matters for the encyclopedia. An entry for hierarchy as a structural prime should note its frame-pickup history: the contexts where it accretes which kinds of frames. A practitioner querying the encyclopedia for whether hierarchy applies to a problem should be informed both about the structural pattern and about the frames the structural pattern tends to accrete in nearby deployments. The latter is not a defect of the prime but a piece of relevant practical information.
4. The Decomposition Operation and the Three Fates¶
This section introduces a procedure that the structural/framed distinction makes possible: given a framed prime, attempt to extract its structural core by abstracting away the institutional or normative overlay. The procedure has three possible outcomes — unification, new analytical tool, and loss of identity — each with distinct consequences for the corpus and for the prime's analytical use. The typology generates testable predictions about how decomposition outcomes vary with the prime's properties.
4.1 The Operation Defined
Given a framed prime P, the decomposition operation asks: is there a structural pattern S such that P can be characterized as S plus some institutional or normative frame F, where S is statable in purely structural terms (no human-practice referents) and F captures the institutional or normative content that distinguishes P from S?
The operation is a thought experiment, not a mechanical algorithm. Performing it requires judgment about what counts as structural, what counts as framework, and what counts as a successful extraction. But the operation is reproducible enough to be applied systematically across the corpus, and inter-rater agreement studies can be designed around it.
The output of the operation is one of three types of result.
4.2 Fate One: Unification
The decomposition can produce a structural core S that turns out to match a structural prime already in the corpus. The framed prime P is then revealed as a context-enriched specialization of an existing structural prime: P ≈ S(existing) + F(specific frame).
Worked example: justice. Apply the decomposition. Strip away the moral, legal, and institutional content. What remains is something like balanced allocation according to a rule, where deviations from the rule are corrected. This is recognizable as equilibrium-with-correction, a structural pattern already present in the corpus under other names. The structural core of justice is not a new prime; it is a known structural prime that has been wrapped in a frame about moral desert and institutional procedure.
Worked example: legitimacy. Strip away the political and institutional content. What remains is something like recognized warrant for authority, which decomposes further into recognition (a structural pattern about agreement among parties) plus authority (a structural pattern about asymmetric capacity to direct action). The structural core involves no new primes; it is a composition of recognized structural patterns.
Utility of fate-one outcomes: theoretical cleanup and explicit cross-referencing. The corpus does not gain a new node, but it gains an articulated relationship: a framed prime is now linked to the structural prime that constitutes its core, with the frame difference made explicit. This supports search and retrieval (a query on the structural pattern returns the framed prime via the link) and improves the corpus's truthfulness about how its concepts relate.
Risk of fate-one outcomes: confirmation bias. It is easy to declare unification too quickly, mapping the framed prime's structural core to whatever existing structural prime seems vaguely related. Discipline is required: the match has to be genuine, not approximate, and the framed prime's distinctive analytical work should be entirely accounted for by the combination of the structural prime and the frame.
4.3 Fate Two: New Analytical Tool
The decomposition can produce a structural core S that is a real pattern but not currently present in the corpus as a recognized prime. The framed prime P is then revealed as a domain-specific instantiation of a new structural prime, and the decomposition both refines P and adds S to the corpus.
Worked example: sovereignty. Strip away the political-legal content. What remains is something like bounded autonomous authority over a defined region or domain, with the power to exclude or regulate external interference. This pattern is not currently in the corpus under any recognized name. It is not the same as "boundary" (which is more passive), not the same as "encapsulation" (which lacks the authority component), not the same as "autonomy" (which lacks the exclusionary capacity). It is a real pattern, and once named, it can be recognized in many domains: cell biology (cell membranes plus selective transport, plus the cell's regulatory machinery), software (modules with encapsulation, internal state, and access control), ecological niches (a species's effective control over a portion of a habitat), cognitive selfhood (the self's authority over its mental contents and capacity to exclude or admit external influences). The structural pattern "encapsulated authority" is a candidate new prime in the corpus, and sovereignty is now characterized as encapsulated authority plus a political-legal frame.
Worked example: contract. Strip away the legal content. What remains is something like mutual binding commitment between parties, with specified terms and specified consequences for non-performance. This is a real pattern. It is present in biological mutualisms (two species commit to a mutually beneficial exchange, with consequences — usually loss of fitness — for failure to perform). It is present in network protocols (two endpoints commit to specified message exchanges, with failure modes specified). It is present in scientific peer review (the reviewer commits to confidentiality and timely review, with consequences for failure). The pattern "mutual binding with specified non-performance consequences" is not currently in the corpus under a recognized name. Naming it adds a new analytical tool to the corpus that projects into domains where the original framed prime "contract" would never have been considered.
Worked example: due process. Strip away the legal content. What remains is something like structured procedural pathway through which a consequential decision is reached, with mandatory checkpoints, safeguards against arbitrary judgment, and opportunities for the affected entity to contest or supplement the inputs. This pattern is present in scientific peer review, in clinical diagnosis pathways, in multi-stage software validation pipelines, in airline maintenance procedures, in regulatory approval processes. It is not currently in the corpus as a single named structural prime — the pieces (procedure, checkpoint, validation, contestability) exist separately but the integrated pattern is not consolidated. Naming it adds a structural tool that projects into many domains.
Utility of fate-two outcomes: the corpus gains a new structural prime with potentially broad projection power. These new primes are often surprising: they crystallize patterns that were sitting in the conceptual landscape without anyone having named them at the right level of abstraction. The naming itself is the contribution.
Risk of fate-two outcomes: over-eager pattern declaration. It is easy to convince oneself that an extracted structural core is a real pattern when it is actually a thin sketch that doesn't project meaningfully into other domains. The test for fate-two should include actual projection: can the new structural prime be applied to at least three additional domains where it does real analytical work, separate from any application of the original framed prime? If yes, the fate-two outcome is real; if no, it may be a fate-three outcome in disguise.
4.4 Fate Three: Loss of Identity
The decomposition can produce a structural core that, once extracted, is too thin to retain the analytical work the original framed prime was doing. The extraction is technically possible — every concept admits some structural description at some level of abstraction — but the result has lost what made the original prime interesting.
Worked example: dignity. Strip away the evaluative content. What remains is something like treated as inviolable, subject to non-fungible standing. This is technically a structural pattern. It involves a relational claim about how some entity is treated relative to other entities. But the structural description has stripped out exactly what dignity was doing analytically. The original prime did its work because of the normative weight: an act violates dignity because of what dignity claims about how the entity should be treated. The structural core, without the normative content, is not capable of generating the prescriptions and judgments that the original prime supports. The extraction has, in effect, replaced dignity with a different concept (something like inviolability of standing) that does not do dignity's work.
Worked example: sanctity. Strip away the religious or moral content. What remains is something like set apart for special treatment or protection. This is structurally identifiable but it is no longer sanctity. The original prime did its work through its connection to the sacred or the morally absolute; the structural residue is a pale description of any case of differential treatment.
Worked example: love (treating it as a candidate prime). Strip away the affective and evaluative content. What remains might be something like high-priority concern for another's welfare combined with strong preference for proximity. This is not love. The structural residue is something else, and substituting it for love in any analytical context would change the topic.
Fate-three outcomes correspond closely to Williams's thick concepts. The descriptive and evaluative content cannot be cleanly separated because the evaluative content is constitutive of the concept's analytical role. They also correspond, in some cases, to Gallie's essentially contested concepts: the framing context includes contested evaluative commitments that are part of what the concept is doing, not removable surface features.
Utility of fate-three outcomes: identifying them is useful. A prime classified as fate three is marked as constitutively framed; its analytical use should not be confused with the use of any structural prime that might superficially resemble its structural residue; and it cannot be straightforwardly added to a structural-prime catalog. The classification itself is the contribution.
Risk of fate-three outcomes: misclassification as fate two. Sometimes a structural residue looks like a real pattern at first glance but reveals itself on closer inspection to be too thin to do work. The discipline is to test the residue independently: can it project into domains where the original framed prime would not? If the residue does not project meaningfully, fate three is the correct classification even if the residue had the appearance of being a recognizable pattern.
4.5 Predicting Decomposition Fates
The typology makes predictions about which framed primes will produce which fates. The predictions are testable empirically; they are not yet tested. Stating them as hypotheses:
Hypothesis 1: Fate-one outcomes (unification) are more likely when the framed prime's home discipline historically borrowed a pattern from a structural discipline. Legal "balance" and "equilibrium" of obligations is a borrowed metaphor from physics. The structural pattern was there first and was wrapped in legal framing; decomposition rediscovers the structural pattern.
Hypothesis 2: Fate-two outcomes (new analytical tool) are more likely when the framed prime crystallizes a pattern that arose convergently across human institutions and natural systems via independent design pressures. Sovereignty/encapsulated-authority is a candidate: the pattern of bounded authority with exclusionary capacity arose independently in political theory, cell biology, software design, and ecological niches because the design pressures (the need for some entity to control a region against external interference) recur across substrates. Conceptual engineering in any one domain produces a partial articulation; the decomposition reveals the underlying convergent pattern.
Hypothesis 3: Fate-three outcomes (loss of identity) are more likely when the framed prime's analytical work depends on evaluative or normative content that no structural rendering can capture. Dignity, sanctity, and arguably justice fall here. Williams's thick concepts and Gallie's essentially contested concepts predict fate three.
Hypothesis 4: There is a correlation between the prime's age and its decomposition fate. Older framed primes (sovereignty, contract, property) are more likely to produce fate-two outcomes, because they have had time to crystallize patterns that natural and engineered systems have independently re-invented. Newer framed primes are more likely to produce fate-one outcomes (the structural pattern was already in the corpus when the framed concept was developed) or fate-three outcomes (the framing is freshly contested and the structural residue is thin). This is a weaker hypothesis but worth testing.
4.6 Empirical Tests
The decomposition typology can be tested empirically. A protocol might proceed as follows: curate a set of 30 framed primes drawn from law, governance, ethics, economics, sociology, and related disciplines. For each, have multiple decomposers (human or LLM-assisted) independently attempt the decomposition operation. Classify each outcome by fate. Calculate inter-rater agreement on classifications. Look for systematic patterns: discipline of origin, age of concept, evaluative load, frame complexity. Test the four hypotheses above against the distribution of outcomes.
Such a study would establish (or disconfirm) the typology's empirical traction. It would also produce a curated catalog of decompositions and a calibrated picture of which fates predominate in which subdomains. The Encyclopedia of Abstractions project could absorb the catalog directly: framed primes with fate-one decompositions would receive cross-references to their structural cores; fate-two decompositions would add new structural primes to the corpus; fate-three decompositions would flag the framed primes as constitutively framed.
4.7 Relation to Thick/Thin and Essentially Contested Concepts
The typology connects to the prior literatures discussed in section 2 in specific ways.
Fate-three framed primes are roughly Williams's thick concepts. The structural residue, once extracted, is too thin to do the work the original concept was doing — exactly the impossibility-of-clean-separation that Williams argued for in ethics.
Fate-one and fate-two framed primes admit decomposition. In thick/thin terms, they have descriptive content (the structural core) that can be separated from a layer of additional content (the frame). The thick/thin distinction in ethics has tended to treat all thick concepts as resisting decomposition, but the typology developed here suggests that "thickness" in the broader sense covers at least two phenomena: thickness that decomposes (fate one and two) and thickness that does not (fate three). Williams's argument may apply most forcefully to fate-three concepts and less forcefully to others.
Fate-three framed primes may also be essentially contested in Gallie's sense, when the constitutive frame includes contested evaluative commitments. The correlation is testable: fate-three classifications and essentially-contested classifications should overlap substantially.
The relationship to the analogy literature is more indirect. The decomposition operation is, in part, a procedure for distinguishing structural mapping from surface-feature transfer in the case of framed primes. A successful decomposition (fate one or fate two) means the structural pattern can be applied to new domains without bringing the frame along; a failed decomposition (fate three) means the frame and the pattern are inseparable in a way that defeats analogy projection. The decomposition typology thus extends structure-mapping theory's basic insight (relational structure is what analogies map) by characterizing when an abstraction's relational structure is cleanly separable from its surface features and when it is not.
5. Bidirectional Dynamics: Frame Pickup in Structural Prime Deployment¶
The basic structural/framed distinction treats the structural-versus-framed character of a prime as a property of the prime itself. This section complicates that picture. Primes can change their character based on where they are deployed. Specifically, structural primes can acquire frames when deployed in normative or institutional contexts, and tracking this frame pickup is important for both encyclopedia maintenance and for any practical use of the corpus.
5.1 The Reverse Phenomenon
A structural prime, by definition, is intrinsically free of any particular institutional or normative frame at definition. But intrinsic freedom does not prevent the prime from accreting a frame in use. When a structural prime is deployed in a normative or institutional domain, the domain's existing vocabulary and evaluative commitments tend to attach to the prime in ways that travel with it in further use.
Consider equilibrium. In physics, equilibrium is a pure structural prime: a state in which competing forces or processes balance, with the system tending to remain in that state absent external perturbation. The pattern is neutral; there is no evaluative content. In economics, however, "equilibrium" has acquired a frame. Economic equilibrium is not just the structural pattern; it has come to carry connotations of efficiency, Pareto optimality, sometimes desirability. When economists speak of restoring equilibrium, they often mean something more than restoring the structural pattern — they mean restoring a state that economic theory treats as analytically preferred. The structural prime equilibrium has picked up a welfare frame in economic deployment.
Consider feedback. In cybernetics and engineering, feedback is a pure structural prime. Negative feedback stabilizes; positive feedback amplifies; the pattern is neutral. In organizational management and governance, "regulatory feedback" has acquired a frame about institutional design and democratic responsiveness. The structural pattern is the same, but the deployment context attaches assumptions about accountability, transparency, and proper institutional function. The structural prime feedback has picked up a governance frame in management and policy deployment.
Consider hierarchy. As discussed in section 3, hierarchy is a pure structural prime at definition. In social and organizational deployment, it picks up frames about power, dominance, and sometimes oppression. In biological deployment, hierarchy is neutral. In organizational deployment, it is not. The structural prime is the same; the deployment context attaches a frame.
These examples illustrate a general phenomenon: deployment context can attach interpretive content to a structural prime, and the attached content can then travel with the prime in further use. The frame pickup is not arbitrary; it tends to draw from the receiving domain's existing institutional and normative vocabulary, and once attached, it functions much like an originally constitutive frame.
5.2 Intrinsic Character vs. Deployed Character
The phenomenon suggests a useful distinction: a prime's intrinsic character (the structural-versus-framed classification at definition) and its deployed character (the structural-versus-framed character it exhibits in a particular context of use).
For pure structural primes like symmetry, the intrinsic and deployed characters tend to align across most contexts: symmetry is structural in physics, in art, in mathematics, in biology, almost wherever it is applied. Frame pickup is rare and minor.
For primes like hierarchy or equilibrium, the intrinsic character is structural but the deployed character can shift substantially: structural in formal contexts, framed in normative or institutional deployment. The frame pickup is significant and consistent within each deployment domain.
For framed primes like sovereignty or contract, the intrinsic character is framed, but the deployed character is also framed almost everywhere (the frame travels with the prime by definition).
The intrinsic/deployed distinction matters because the failure modes and validation requirements depend on deployed character, not intrinsic character. A structural prime deployed without frame pickup admits recognitional validation; a structural prime deployed with frame pickup requires the additional check of whether the imported frame is warranted.
5.3 The Smuggled Frame Failure Mode
Frame pickup becomes dangerous when it happens unnoticed. The structural prime carries its newly accreted frame into subsequent uses, but the user may not be aware that a frame is now attached. The result is that assumptions get imported across domain boundaries without scrutiny — a phenomenon I will call smuggled framing.
A canonical example is the application of market to non-economic domains. Market in pure economic theory is, arguably, a structural prime: a coordination mechanism in which agents with preferences exchange resources via price signals. But market in cultural and political deployment has accreted a thick frame about agent rationality, preference primacy, exchange-as-paradigm, and sometimes the desirability of market-like coordination. When market is then applied to non-economic domains — the "marketplace of ideas," the "marriage market" — the structural pattern travels but so does the frame, often without the user's awareness. The receiving domain ends up importing assumptions about agents and preferences that may not warrant in that context.
Another example is the application of natural selection to non-biological domains. Natural selection in evolutionary biology is largely structural: differential reproduction under variation and heredity. In some popular and policy deployment, natural selection picks up an evaluative frame (survival of the fittest, with connotations of moral fitness and deserved outcomes). When this frame-laden version of natural selection is then applied to social or economic domains, the smuggled frame imports evaluative claims that no one has explicitly endorsed but that nonetheless shape the analysis.
A third example concerns information in computer science and complex systems theory. Information as Shannon defined it is structural: a measure of resolved uncertainty. In some recent applications to biology and consciousness studies, information has picked up frames about meaning, content, and representation that go well beyond Shannon's definition. The structural prime is being deployed with smuggled framing in ways that can muddle the analysis.
Detecting smuggled framing requires explicit attention to what content the prime is carrying in the current context. Workflow proposals for this attention are developed in the pragmatic section.
5.4 Productive Frame Pickup
Not all frame pickup is dangerous. Sometimes the receiving domain's frame is warranted, and the pickup imports useful interpretive structure along with the structural pattern. Section 2.5 discussed boundary objects: framed primes can coordinate work across communities precisely because the frame can be re-interpreted in each community. The same dynamic can occur in reverse with structural primes that pick up domain-appropriate frames.
When equilibrium picks up a welfare frame in economic deployment, the pickup is appropriate if the welfare interpretation is genuinely doing analytical work — for example, if welfare-relevant equilibria are precisely the ones economists care about and the frame is what makes the prime useful in that domain. Similarly, when hierarchy picks up a power frame in social deployment, the pickup is appropriate if the analysis is about social systems where power is in fact a relevant feature.
The line between productive and smuggled framing is not always sharp, and reasonable analysts can disagree. The framework's contribution is to make the question askable: what frame has this prime picked up in this deployment, and is the frame warranted by the receiving domain? The question may not have a clean answer, but asking it changes how the analysis is evaluated.
5.5 Implications for the Encyclopedia
The encyclopedia should track frame pickup as a phenomenon. Entries for structural primes can include a frame-pickup history — a list of deployment domains where the prime has been observed to accrete frames, with brief characterizations of what the accreted frames are. This is not a defect annotation; it is practically useful information. A practitioner thinking about applying hierarchy to organizational design should know in advance that hierarchy tends to accrete a power frame in such deployments, even though hierarchy at definition is neutral.
Frame-pickup history is also useful for distinguishing intrinsically framed primes from structural primes with persistent frame pickup. Sovereignty is intrinsically framed; its frame is constitutive. Hierarchy is intrinsically structural but its frame in social deployment is so consistent that, for practical purposes, the social deployment of hierarchy behaves almost like a framed prime. The distinction between these cases matters for analysis and for decomposition: hierarchy in social deployment can still be decomposed (the social-power frame is separable from the structural pattern), while sovereignty's frame is harder to separate cleanly.
6. Epistemological Perspective¶
This section steps back from the framework and asks what kinds of knowledge claims are being made when primes are applied across domains. The structural/framed distinction has direct epistemological consequences: the two prime types support different sorts of claims, admit different verification operations, and fail in different ways.
6.1 What Is Being Claimed?
When a structural prime is applied to a new domain, the claim is typically recognitional: the relational pattern is present in the domain. Applying feedback to a population dynamics model claims that the model has a return-signal-modulating-input relationship of a specifiable kind. Applying hierarchy to a taxonomy claims that the taxonomy has nested-levels-with-asymmetric-relations of a specifiable kind. The claims are about the presence or absence of a pattern, and the patterns are described in terms that can in principle be checked.
When a framed prime is applied to a new domain, the claim is typically interpretive: the frame is productive for understanding the domain. Applying due process to AI decision-making claims that the procedural frame of legal due process — structured pathways, checkpoints, contestability — is illuminating when imported into the AI context. The claim is not about whether AI decision-making has due process in the strict legal sense; it is about whether viewing AI decision-making through the frame of due process produces better analysis, better designs, better outcomes than other ways of viewing it. The claim is methodological, about productive perspective, more than propositional.
This distinction matters because the two kinds of claims support different kinds of inference and different kinds of error.
6.2 Verification Asymmetries
Structural prime claims admit relatively clean verification. The relational pattern is specified, the target domain is examined, and one can check whether the pattern is instantiated. Verification can be wrong — the analyst can miss a feature of the domain that contradicts the claimed instantiation, or can mistake a superficially similar pattern for the one in question — but the verification operation has a definite shape. There is a fact of the matter (assuming the structural prime is well-defined), and the analyst is trying to identify it.
Framed prime claims admit a softer verification. The question is whether the frame illuminates the domain. There is no single fact of the matter; there is a judgment about productive perspective. Verification proceeds by assessing whether the imported frame leads to insights, predictions, designs, or interventions that would not have been available without it. The judgment can be wrong, but wrongness here is about productivity rather than correspondence.
Failure modes follow from these asymmetries. Structural prime claims fail by misidentification (the pattern is not actually present in the form claimed) or by vacuity (the pattern is present but in a trivial way that does no analytical work). Framed prime claims fail by forced application (the frame does not illuminate the domain, or imports assumptions that distort it) or by failure to add value (the frame fits but doesn't say anything that wasn't already available without it).
6.3 The Status of Extracted Structural Cores
A specific epistemological question arises about the structural cores extracted in fate-one and fate-two decompositions: are they discoveries or constructions? The question is a special case of the broader debate Hacking surveyed in The Social Construction of What?[12], where he insists that "social construction" claims must be disambiguated by their target (an object, an idea, a kind) before they can be evaluated.
In fate-one outcomes, the extracted core matches an existing structural prime. The decomposition is a rediscovery: it reveals that the framed prime was, all along, an instance of a known structural pattern with an added frame. The core's existence is independent of the decomposition; the operation just makes the relationship visible.
In fate-two outcomes, the extracted core is a previously unnamed structural pattern. Here the question is sharper. Did the pattern exist before being named, awaiting discovery? Or did the act of naming bring it into existence as an analytical category?
The framework's answer is broadly realist about structural patterns. The pattern of encapsulated authority existed before anyone named it — cells, software modules, and political sovereignties all instantiate it, and they did so before the abstraction was articulated. The naming makes the pattern available as an analytical tool, but it does not bring the pattern into existence. This is a moderate version of joint-carving: structural primes are real features of the world, and the encyclopedia's job is to identify them.
But this realism is qualified. The choice of which structural patterns to elevate to prime status, the level of generality at which they are pitched, the boundaries between adjacent patterns — these are partly constructive choices. The framework can be realist about the patterns while acknowledging that the corpus of recognized primes is a curated product of theoretical decisions.
For framed primes, the situation is different. The frames are constituted at least partly by human practices, institutions, and conventions. There is no question of whether legal sovereignty existed before legal sovereignty was articulated; sovereignty in the political-legal sense came into being with the practices and institutions that constituted it. Framed primes are partly discovered (insofar as their structural cores correspond to real patterns) and partly constructed (insofar as their frames are products of human institutional history).
6.4 When Does Cross-Domain Application Produce Knowledge?
A practically important epistemological question: when does applying a prime to a new domain actually generate knowledge, as opposed to merely producing a plausible-sounding analogy?
For structural primes, knowledge is generated when the prime's presence in the new domain has non-trivial analytical consequences. Recognizing feedback in a population dynamics model is genuine knowledge if the recognition enables prediction, intervention design, or organization of further questions; it is empty if the recognition just labels a feature that was already understood by other means. The structural prime's value in the new domain depends on what its presence enables.
For framed primes, knowledge is generated when the frame's machinery applies to the new domain in non-trivial ways. Applying due process to AI decision-making generates knowledge if the procedural machinery imported from legal practice — structured pathways, checkpoints, contestability — produces design ideas, evaluation criteria, or system architecture that would not have been available without the frame. The framed prime's value depends on what its machinery enables in the target.
In both cases, the test is consequential rather than merely recognitional. A prime application that does not change what the analyst does, predicts, designs, or recommends is an application without epistemic payoff. The framework's value is not in finding many primes in many domains; it is in finding primes whose presence does work.
6.5 Vacuity as the Universal Risk
Vacuity is the central epistemic risk for structural prime applications, as discussed above. It is also a risk, though a smaller one, for framed prime applications: a framed prime can be applied in a way that imports frame vocabulary but adds nothing analytically.
The discipline of avoiding vacuity is the discipline of insisting that prime applications do work. For each application, the analyst should be able to state what the prime's presence (or the frame's application) is contributing — what predictions, organizing insights, or interventions become available. Without such a statement, the application is decorative rather than productive.
The encyclopedia can support this discipline by including, for each domain in which a prime has been applied, a brief characterization of what the application contributed. This makes the corpus more useful (a practitioner querying for primes applicable to their domain gets not just a list but a list with examples of payoffs) and more honest (applications that don't generate payoff get filtered out over time).
6.6 Theory-Ladenness
All observation is theory-laden in some sense — what we notice in a domain depends on the conceptual apparatus we bring to it. This is true for structural prime applications and for framed prime applications, but the theory-ladenness is more visible in the framed case. When applying sovereignty to data, it is obvious that the frame is shaping what the analyst sees; the frame is part of the analytical apparatus. When applying feedback to a population dynamics model, the theory-ladenness is less obvious, but it is still present: the analyst noticed the feedback structure because she had the concept available, and absent that concept she might have organized the same data differently.
Theory-ladenness does not disqualify either kind of prime application. It does mean that the analyst should be reflective about which theory is loading the observation. For framed primes, the relevant question is which frame; for structural primes, the relevant question is which structural concept is doing the organizing work. The encyclopedia can support both kinds of reflection by making the framing/concept choices explicit in its entries.
7. Pragmatic Perspective¶
This section addresses the practical consequences of the structural/framed distinction, with particular attention to LLM-assisted discovery of cross-domain prime applications. The Encyclopedia of Abstractions project was originally motivated by a desire to generate synthetic data — taking the Cartesian product of primes and domains and producing novel applications — and that motivation remains central. The framework developed in this paper directly affects how that generation should be done and how the results should be evaluated.
7.1 The Cartesian Product, Revisited
The basic generative move is: take the set of primes, take the set of domains, form the Cartesian product, and for each (prime, domain) pair, consider whether the prime applies to the domain in a productive way. The product is large — even a modest encyclopedia of 200 primes and 100 domains yields 20,000 pairs — and most pairs will not yield interesting applications. The practical question is how to allocate effort.
The structural/framed distinction changes the evaluation regime for each pair. For structural prime × domain pairs, the question is recognitional: is the structural pattern instantiated in the domain in a non-trivial way? The validation is relatively cheap and reliable. For framed prime × domain pairs, the question is interpretive: does the frame illuminate the domain productively? The validation is more expensive and noisier.
The yield rates differ accordingly. Structural prime applications produce many candidates with relatively reliable validation. Framed prime applications produce fewer candidates with more variable validation. Both yield useful results, but the workflow for each should be different.
7.2 Differential Workflows
For structural prime applications, the workflow can emphasize breadth: project the prime into many domains, identify candidate instantiations, validate by recognition. The validation operation can plausibly be partially automated, with LLM-assisted checks for whether the structural pattern is present in a domain's standard representations. Human review is needed primarily for ambiguous cases and for assessing vacuity (is the present pattern doing analytical work?).
For framed prime applications, the workflow should emphasize selectivity: identify target domains where the frame is plausibly relevant, project carefully, validate by interpretive judgment. The validation cannot be fully automated; it requires the kind of expert reading that establishes whether a frame is illuminating. LLM assistance can produce candidates and rough plausibility assessments, but high-confidence validation requires human judgment, especially for fate-three primes whose constitutive framing makes mechanical assessment unreliable.
For structural prime applications that may exhibit frame pickup in the target domain (recall section 5), an additional workflow step is needed: the analyst should explicitly identify any frames the structural prime has accreted in the target deployment domain, and assess whether those frames are warranted. This is the smuggled-frame check.
7.3 Quality-Control Checks
The framework suggests specific quality-control checks for prime applications. They differ for the two prime types.
For structural prime applications, the central checks are: First, is the structural pattern actually present in the target domain in the form claimed? Second, is the pattern doing analytical work — generating predictions, organizing insights, or suggesting interventions that were not available without it? Third, has the structural prime accreted any frames in this deployment context, and if so, are those frames warranted by the target domain or are they being smuggled in unnoticed?
For framed prime applications, the central checks are: First, is the frame's machinery actually applicable to the target domain — are the institutional, procedural, or evaluative components that constitute the frame meaningful in the new context? Second, is the application importing tested machinery from the home domain in a way that adds value, or is it merely importing vocabulary that decorates without illuminating? Third, has decomposition been considered — would the structural core of the framed prime do the work better in this domain, without the institutional baggage?
These checks can be operationalized as prompts for LLM-assisted evaluation, with human review for the harder cases.
7.4 LLM-Assisted Decomposition
The decomposition operation itself is amenable to LLM assistance. Given a framed prime, an LLM can produce a candidate decomposition into structural core and frame, can suggest the likely fate (unification, new tool, or loss of identity), and can offer projection targets where the structural core might apply. The conversation that led to this paper involved exactly this kind of work, and the results were substantive enough to warrant treating LLM-assisted decomposition as a practical research tool.
The risks are real. LLMs can produce plausible-looking decompositions that don't survive scrutiny. A structural core that sounds like a real pattern may turn out to be a thin sketch with no projection power. A claimed unification may match an existing prime only superficially. A fate-three classification may be a mistake (or a fate-three may be missed).
A reasonable workflow for LLM-assisted decomposition is: First, the LLM produces a candidate decomposition with proposed fate classification and projection targets. Second, the LLM is prompted to argue against its own decomposition — to identify reasons the structural core might be too thin, the unification might be spurious, or alternative decompositions might be better. Third, a human reviews the decomposition and the counter-arguments, makes a tentative classification, and (for fate-two cases) tests the proposed structural core by attempting to apply it to two or three domains independently of the original framed prime. Fourth, if the structural core projects meaningfully, the decomposition is accepted and the new structural prime is added to the corpus; if not, the classification is revised toward fate one or fate three.
To make the workflow concrete, consider two worked examples drawn from the conversational work that produced this paper.
Example one: sovereignty. An LLM was prompted to extract a structural core from sovereignty. The proposed core was bounded autonomous authority over a defined domain, with power to exclude or regulate external interference. The proposed fate was two — new analytical tool — with projection targets including cell biology (membranes plus selective transport plus regulatory machinery), software modules (encapsulation plus internal state plus access control), ecological niches (a species's effective control over a portion of habitat), and cognitive selfhood (the self's authority over its own mental contents and capacity to admit or exclude external influences). The self-critique step then asked whether the proposed core was meaningfully distinct from a simple composition of existing primes — could it just be boundary plus autonomy, for instance? The LLM responded that boundary captures only spatial demarcation and autonomy captures only agentic capacity, while encapsulated authority adds an exclusionary regulatory function that integrates both into a single pattern. Human validation tested the proposed core against the projection targets and found that each application surfaced features that boundary plus autonomy did not capture: cell membranes are not merely passive boundaries but actively regulate transport, software modules do not merely encapsulate state but enforce access policies. The decomposition was tentatively accepted as fate two, with encapsulated authority added as a candidate new structural prime to be tested through further projections.
Example two: justice. An LLM was prompted to extract a structural core from justice. The first-pass output proposed the core balanced allocation according to a rule, where deviations from the rule are corrected, and classified the decomposition as fate one — unification with equilibrium-with-correction, an existing structural pattern in the corpus. The self-critique step then asked whether the proposed core preserves justice's analytical work. The LLM noted that justice does substantial normative work — it generates ought-claims, supports moral argumentation, anchors institutional design, structures debates about desert — and that "balanced allocation with correction" does not preserve any of this. Human validation agreed: substituting equilibrium-with-correction for justice in any normative argument changes the topic of the argument rather than continuing it in different vocabulary. The classification was therefore revised. The case turns out to exhibit features of both fate one and fate three simultaneously: the structural pattern is already in the corpus (a narrow fate-one match), but the framed prime nonetheless loses analytical identity when reduced to its core (the fate-three phenomenon). This hybrid outcome was not anticipated by the typology in its initial formulation; the example shows that the three fates are not always cleanly disjoint, and that some framed primes may need to be tagged with both a structural-core cross-reference and a fate-three "constitutively framed" flag. The encyclopedia entry for justice under such cases would carry both pieces of information — a cross-reference to equilibrium-with-correction for navigation purposes, plus a constitutively-framed marker for analytical purposes. This case also revises a provisional classification offered earlier in this paper (section 4.2 treated justice as a clean fate-one example, which the more careful self-critique step here corrects).
The two examples show different decomposition outcomes and illustrate why an LLM's first-pass classification is unreliable on its own. In example one, the initial fate-two classification survived scrutiny, but only after the self-critique step forced explicit checking that the core was distinguishable from compositions of existing primes. In example two, the initial fate-one classification was overturned by the self-critique, revealing a hybrid case that purely mechanical workflow would have missed. The pattern generalizes: LLM-assisted decomposition is genuinely valuable for scaling the work, but only when the workflow includes adversarial self-critique and human validation against projection targets. Without the adversarial step, the LLM tends to converge on plausible-looking first-pass classifications that do not survive scrutiny; with the adversarial step, the workflow produces classifications that are calibrated to the actual analytical behavior of the primes.
This workflow can scale. With LLM assistance, decompositions can be produced and reviewed at a rate that would be impossible for unassisted human work, and the resulting catalog of decompositions becomes a substantial addition to the encyclopedia.
7.5 For Solution Architects
The framework has direct utility for solution architects who use abstractions to design systems and solve problems. The two prime types serve different design functions.
Structural primes are tools for identifying patterns in a problem domain. A solution architect facing a complex problem can ask: which structural primes apply? The answer surfaces invariants — features of the problem that have been studied in other domains where the same structural pattern is present, and from which design ideas can be adapted. The validation is recognitional: the architect checks whether the structural pattern is actually present, and if so, draws on the cross-domain knowledge associated with the pattern.
Framed primes are tools for importing tested machinery from analogous problem domains. A solution architect facing a problem that has institutional, normative, or procedural complexity can ask: which framed primes from related domains have tested machinery that might be imported? The answer surfaces frames — institutional designs, procedural pathways, evaluative criteria — that have been developed in other contexts to handle similar complexity, and which can be adapted to the new problem. The validation is interpretive: the architect assesses whether the imported frame is warranted and productive in the new context.
Both kinds of moves are valuable. The structural move is about discovering invariants; the framed move is about borrowing infrastructure. A mature use of the encyclopedia would treat both as part of the standard toolkit, with explicit awareness of which kind of move is being made and what kind of validation it requires.
7.6 Typed Cartesian Products
The original Cartesian product framing — prime × domain → candidate application — can be refined in light of the framework. A more useful formulation is:
(typed prime) × domain × deployment frame check × decomposition state → typed candidate application with appropriate validation regime
In words: each prime is typed (structural, mixed, framed). The application to a domain is checked for frame pickup. The framed prime's decomposition state (fate one, two, or three) is consulted. The resulting candidate application is assigned to the appropriate validation regime (recognitional, interpretive, or hybrid). The output is not just a candidate but a candidate-with-evaluation-protocol.
This refinement makes the Cartesian product generation more productive (it directs effort toward higher-yield pairs first) and more truthful (the validation regime matches the prime's actual character rather than treating all prime applications as analytically equivalent).
7.7 What the Pragmatic Perspective Asks the Encyclopedia to Do
The pragmatic perspective makes specific requests of the encyclopedia. The encyclopedia should: type each prime on the structural/framed spectrum, with diagnostic criteria documented; for framed primes, include the structural-core extraction (if it has been done) with the decomposition fate; for all primes, include frame-pickup history in deployment domains; for each prime × domain application in the corpus, document the validation operation and the analytical contribution. This makes the encyclopedia not just a catalog of primes but a catalog of typed primes with usage-ready evaluation guidance. The improvement in practical utility justifies the additional schema complexity.
8. Implications for the Encyclopedia of Abstractions¶
This section consolidates the schema and workflow implications for the Encyclopedia of Abstractions project. The framework's value depends on whether it can be operationalized in the corpus, and this section addresses that operationalization concretely.
8.1 Schema Additions
The encyclopedia's current entry for a prime can be extended with the following fields.
Type tag. A classification on the structural/framed spectrum, ideally with a categorical label (structural, mixed-structural, mixed-framed, framed) plus an optional rationale linking the classification to the diagnostic criteria in section 3.2.
Portability profile. A characterization of how the prime travels across domains. For structural primes, this typically means "broad" with a list of domains where the prime is known to instantiate. For framed primes, it typically means "targeted" with a list of domains where the frame applies productively. Mixed primes carry mixed profiles.
Structural core (for framed primes). The extracted structural pattern beneath the frame, when extraction has been attempted, along with the decomposition fate classification (unification with an existing structural prime, new analytical tool, or loss of identity). For fate-one cases, this is a cross-reference to the existing structural prime; for fate-two cases, this is an entry in its own right (the new structural prime is added to the corpus); for fate-three cases, this is a noted classification with brief rationale.
Frame-pickup history. For all primes, a list of deployment contexts where the prime is known to accrete frames or has historically picked up frames, with brief characterizations of the accreted frames. For pure structural primes, this list may be short or empty; for framed primes, it overlaps substantially with the portability profile.
Cross-prime relationships. Explicit links to related primes: the structural core of a framed prime (fate one) points to the existing prime; the framed prime's entry points back. Mixed primes link to both their structural and framed neighbors. Unifications across the corpus become navigable.
Application catalog. For each domain in which the prime has been applied, a brief entry documenting the application: what the prime contributed analytically, what kind of validation the application received, and any known instances of vacuity or smuggled framing.
8.2 Ontology Consequences
The schema additions have consequences for the corpus's structure.
Unifications reduce redundant nodes. Many framed primes may share structural cores, and these cores may already exist in the corpus or may be added as new primes. The mass of framed primes thus partially collapses into a smaller set of structural primes plus a layer of frames. The corpus becomes more economical without losing content.
Framed primes get linked entries. A framed prime is no longer a standalone node; it is a structural core plus a frame, with the relationship made explicit. This supports both navigation and analysis: a query on the structural pattern returns all framed primes whose cores match, and an analyst working with the framed prime knows immediately what structural pattern they are actually relying on.
Super-prime status becomes derived rather than declared. A prime's high portability is now a consequence of its structural character (or its carrying of a pan-human frame), not a separately stipulated property. The encyclopedia gains analytical leverage on what makes a super-prime a super-prime.
Cross-prime relationships become navigable. The encyclopedia is no longer just a list of primes but a graph of relationships, with structural cores at the center of many clusters and framed primes attached around them.
8.3 A Worked Example: Two Entries
To illustrate, consider sketches of two prime entries under the new schema.
Hierarchy (structural prime). Type tag: structural. Portability profile: broad. Applies in mathematics, biology (taxonomy, organ-tissue-cell-molecule), computer science (memory hierarchies, abstraction layers, file systems), linguistics (phrase structure), physics (energy levels), chemistry, organizational design, social systems. Frame-pickup history: in social and organizational deployment, hierarchy accretes a frame about power, authority, dominance, sometimes oppression; in biological deployment, the frame is largely absent. Cross-prime relationships: related to nesting, modularity, recursion. Application catalog: dozens of entries across the listed domains, with brief notes on what hierarchy contributed (e.g., in memory hierarchy design, hierarchy organizes the analysis of access patterns and latency trade-offs).
Sovereignty (framed prime). Type tag: framed. Portability profile: targeted; applies in domains where bounded authority over a defined region is analytically central. Structural core: encapsulated authority over a defined domain with power to exclude or regulate external interference; decomposition fate: new analytical tool (fate two); the structural core "encapsulated authority" is added to the corpus as a new structural prime. Frame-pickup history: not applicable in the same sense; sovereignty's frame is constitutive rather than accreted, though specific applications of sovereignty pick up additional contextual frames (e.g., "data sovereignty" picks up frames about user agency and information control). Cross-prime relationships: structural core points to encapsulated authority; related framed primes include jurisdiction, authority. Application catalog: applications in political theory, international law, data governance, AI agent design, cell biology (via the structural core), software encapsulation (via the structural core).
These entries are richer than current entries and provide the analyst with usage-ready guidance.
8.4 Migration Considerations
The existing corpus can be retro-typed under the new schema. The migration is a substantial project but tractable. LLM-assisted retro-typing of each prime would produce candidate classifications, candidate decompositions for framed primes, and candidate frame-pickup histories. Human review would establish reliability for each. The pace of migration could be set by available reviewer effort; the framework does not require all-or-nothing migration.
Some primes will resist clean typing. These are interesting cases. A prime that cannot be confidently placed on the spectrum may indicate either a fuzzy boundary or a deeper conceptual issue. Either way, the difficult cases are themselves a contribution: they map the frontier of the framework.
Some primes will produce contested decompositions. Different reviewers may disagree about whether a fate-one outcome is genuine, whether a fate-two structural core is a real pattern, or whether a fate-three classification is correct. Disagreements should be tracked and used as the basis for inter-rater reliability studies. The framework is more useful if it is calibrated against actual usage variation rather than treated as definitive.
8.5 The Encyclopedia as a Tool
With the schema additions, the encyclopedia becomes a tool that supports several specific kinds of work.
For abstraction discovery, the typed corpus supports the typed Cartesian product approach developed in section 7. Practitioners can query for primes of a given type applicable to their domain and get not just candidates but candidates with appropriate validation regimes.
For solution architecture, the typed corpus supports the structural-vs-framed move distinction developed in section 7.5. Architects can query for structural primes that identify invariants in their problem and for framed primes whose machinery might be imported.
For the underlying research program, the typed corpus supports the empirical study of decomposition fates and frame-pickup patterns. The encyclopedia becomes a research instrument as well as a reference.
9. Limitations, Counter-Arguments, and Open Questions¶
This section takes seriously some objections to the framework and identifies open questions that further work will need to address.
9.1 Objection: This Is Just Thick/Thin Concepts Dressed Up
The objection is that the structural/framed distinction is a restatement of Williams's thick/thin distinction with different vocabulary, and that the contribution is therefore mostly nominal.
The response has two parts. First, the structural/framed distinction generalizes thick/thin beyond ethics. Williams was working on ethical concepts; the framework here applies to all abstractions, including those whose frames are not primarily evaluative (procedural, vocabulary-based, institutional). The structural/framed distinction handles cases that thick/thin does not address. Second, the decomposition typology and the bidirectional deployment dynamic are not present in the thick/thin literature. The three-fates classification of decomposition outcomes is a genuinely new analytical tool, and the observation of frame pickup in structural prime deployment is, to my knowledge, not characterized in this form in any prior literature. The contribution is therefore more than nominal: it absorbs the thick/thin insight, generalizes it, and adds operational machinery the original distinction does not have.
9.2 Objection: The Distinction Is Artifactual of Naming
The objection is that whether a prime looks structural or framed depends on which name we happen to use. If we had named "encapsulated authority" before naming "sovereignty," sovereignty would look like a structural prime with social applications rather than a framed prime with a structural core.
The objection has force, and the framework absorbs it. The structural/framed distinction is partly a property of how a prime has been named and developed historically. A prime that was domain-stripped at the moment of naming reads as structural; a prime that was named within an institutional context reads as framed. But the framework does not depend on the framed/structural label being immutable. The decomposition operation is precisely the move that exposes the structural core of a framed prime regardless of how it was historically named. Over time, with systematic decomposition, the difference between "structural prime" and "framed prime with extracted structural core" becomes less significant — both are structural patterns plus possible frame overlays. The framed/structural distinction is a useful entry point for analysis, not a final partition of the conceptual landscape.
9.3 Objection: Decomposition Is Always Possible at Sufficient Abstraction
The objection is that any framed prime can be decomposed into a structural core if you abstract enough. The three-fates typology — and especially fate three — therefore misnames the phenomenon: there is no real "loss of identity," just a thin residue.
The response is that the framework does not deny the residue. Of course any concept admits a structural description at some level of abstraction. The question fate three names is whether the residue retains the analytical work the original concept was doing. Dignity admits a structural description (treated as inviolable), but the description does not do dignity's analytical work; substituting the description for the concept in any normative argument changes the topic. The phenomenon is real, and "loss of identity" is the appropriate name for it. The objection conflates the technical possibility of structural description with the practical question of whether the description preserves the concept's analytical function.
9.4 Objection: The Framework Is Folk Taxonomy
The objection is that the structural/framed distinction is a folk classification with no rigorous foundation, and that it should not be taken seriously as a theoretical contribution.
The response is that the framework includes explicit diagnostic criteria (section 3.2), an operational decomposition procedure (section 4), testable hypotheses about decomposition outcomes (section 4.5), and an empirical research agenda (sections 4.6 and 10). Folk taxonomies do not typically produce that level of operational specificity. The framework may turn out to be wrong in various ways, but it is not folk taxonomy; it is an explicit conceptual proposal with testable consequences.
9.5 Open Question: Inter-Rater Reliability
A critical open question is whether different analysts will agree on prime classifications, decomposition outcomes, and frame-pickup histories. The framework's value depends in part on its reproducibility. Without strong inter-rater agreement, the typology is more a personal lens than a shared analytical tool. The empirical research agenda (section 10) addresses this directly.
9.6 Open Question: Cross-Cultural Stability
The structural/framed distinction was developed in a single intellectual tradition. Whether framed primes' frames are stable across cultures is an open question. Sovereignty's frame components may be substantially different in different political traditions; contract may carry different institutional baggage in different legal systems. Cross-cultural variation could affect both classifications and decomposition outcomes.
9.7 Open Question: The Metaphysics of Structural Primes
The framework treats structural primes as real features of the world that admit recognitional application. This is a moderate realism that some philosophical traditions would contest. Whether the framework can be defended against more aggressive constructivist objections is an open question. The framework's pragmatic value does not depend on resolving the metaphysics — the typology can be useful as a working tool even if the underlying realism is open — but the philosophical question deserves further work.
9.8 Open Question: Relationship to Natural Kinds
Structural primes look like candidates for what some philosophers call natural kinds: categories that carve the world at its joints rather than reflecting human classification choices. Whether they actually function as natural kinds, or whether they are better thought of as engineering kinds (categories useful for thinking rather than reflecting deep structure of reality), is a question this paper does not resolve. The conceptual engineering literature would tend toward the second answer; some philosophy of science would tend toward the first. The framework is compatible with both positions but does not commit to either.
9.9 Open Question: Frame Pickup Detection
The framework asserts that structural primes can accrete frames in normative deployment. How to detect frame pickup systematically — especially when it is happening in real time rather than being identifiable retrospectively — is an open methodological question. Computational tools for analyzing the contextual usage of primes in large corpora might help, but the methodology is undeveloped.
10. Future Research Directions¶
The framework is conceptual and the empirical work has not been done. This section outlines specific empirical and methodological projects that would test, refine, or extend the framework.
10.1 The Decomposition Study
The most direct empirical test is a decomposition study. Procedure: curate a set of 30 to 50 framed primes drawn from law, governance, ethics, economics, sociology, and adjacent disciplines. For each, have multiple decomposers (humans, LLM-assisted humans, and pure LLM passes) independently attempt the decomposition operation. Classify each outcome by fate. Calculate inter-rater agreement on classifications. Analyze the distribution of fates against hypothesized correlates: discipline of origin, age of concept, evaluative load, frame complexity. Test the four hypotheses from section 4.5.
Expected contribution: empirical calibration of the typology; identification of stable versus unstable classifications; a curated catalog of decompositions for the encyclopedia.
10.2 Inter-Rater Reliability of Prime Classification
A complementary study: for a set of 100 primes (mix of structural, mixed, and framed), have multiple analysts classify each on the structural/framed spectrum using the diagnostic criteria from section 3.2. Measure inter-rater agreement. Identify primes where classifications consistently agree and primes where they do not. For the disagreements, conduct qualitative analysis of the sources of disagreement.
Expected contribution: a calibrated test of the framework's reproducibility; identification of which diagnostic criteria are most reliable and which generate disagreement.
10.3 Frame-Pickup Patterns
A more exploratory study: select 10 structural primes that are widely deployed in multiple disciplines. For each, analyze the prime's usage in three or four deployment domains, looking for frame pickup. Characterize the accreted frames. Identify cross-deployment patterns: are there frames that get picked up by many structural primes (a "power frame" picked up by hierarchy, by network, by authority)? Are there structural primes that pick up many frames (network in social science deployment picks up multiple frame components)?
Expected contribution: a map of the frame-pickup landscape; identification of dominant frames in normative deployment; tools for detecting and tracking frame pickup.
10.4 LLM-Assisted Decomposition Tools
A methodological project: develop and validate an LLM-assisted workflow for decomposition that includes self-critique steps, structural-core projection tests, and human review checkpoints. Compare the workflow's output to unassisted human decomposition on the same primes. Measure efficiency, agreement with expert classifications, and false-positive rates for fate-two claims (cases where the LLM proposes a structural core that turns out not to project meaningfully).
Expected contribution: a practical tool for scaling the decomposition work; calibration of LLM reliability for this specific task.
10.5 Cross-Cultural Validation
A cross-cultural study: take a set of framed primes (sovereignty, contract, property, justice, dignity) and analyze their decomposition outcomes across different intellectual traditions. Are the structural cores stable across traditions? Are the frames stable? Where do they differ?
Expected contribution: assessment of the framework's cultural specificity; identification of frames that vary culturally versus structural cores that may be cross-cultural invariants.
10.6 Application to the Encyclopedia Migration
A practical project: pilot the schema migration outlined in section 8 on a subset of the encyclopedia (say, 50 primes), using LLM-assisted retro-typing followed by human review. Measure the migration cost (analyst hours per prime), the ontology changes (new primes added via fate-two extractions, redundant primes consolidated via fate-one unifications, cross-prime links established), and the resulting corpus's usability.
Expected contribution: a working prototype of the migrated encyclopedia; calibration of the migration cost at scale; evidence about whether the framework actually improves the corpus's analytical utility.
10.7 Pre-Registration
For each of the empirical studies above, pre-registration of hypotheses and analytical protocols would strengthen the work substantially. The framework makes specific predictions; pre-registration would force commitment to those predictions in advance of seeing the data and would reduce the risk of post-hoc rationalization. Pre-registration is feasible for all of the studies sketched here.
11. Conclusion¶
The Encyclopedia of Abstractions began with a practical motivation: catalog the abstractions that travel across domains so that the catalog can support cross-domain transfer and discovery. The operational definition of prime abstraction — applicability across at least three domains — served that motivation well enough to get the project moving and to generate substantial early results. But the operational definition is silent about a phenomenon that anyone working with the corpus eventually notices: primes do not travel in the same way. Some show up in new domains as pure relational structure. Others bring an entire interpretive context along with them. The asymmetry has consequences for how the corpus should be organized, how prime applications should be evaluated, and what kind of analytical work the corpus can do.
This paper has developed a conceptual framework to address that asymmetry. The structural/framed distinction, with diagnostic criteria, captures the basic typology. The decomposition operation, with its three fates, extracts structural cores from framed primes when extraction is possible and identifies constitutively framed primes when it is not. The bidirectional deployment dynamic captures the observation that structural primes can accrete frames in normative deployment, complicating the basic structural/framed distinction in productive ways. Together, these constitute a refinement of the operational definition that does not replace it but enriches it.
The framework engages with several existing scholarly traditions. The closest cousin is the thick/thin concept distinction from Williams; the structural/framed distinction generalizes thick/thin beyond ethics. The analogy and structure-mapping literature provides vocabulary for the failure modes of prime application. Essentially contested concepts characterize a subset of fate-three primes. Conceptual engineering provides the broader methodological framing within which the encyclopedia project sits. Boundary objects describe productive uses of framed primes across communities. The framework absorbs contributions from each of these traditions and adds machinery that none of them has developed for the specific question of cross-domain abstraction transfer at scale.
The framework's value is pragmatic. It changes how the Encyclopedia of Abstractions can be organized, what the corpus contains, and how it is used. Schema additions — type tags, portability profiles, structural-core entries for framed primes, frame-pickup histories — make the corpus richer and more truthful about how its contents behave. Workflow refinements — typed Cartesian products, differential validation regimes, LLM-assisted decomposition — make the corpus more useful for both discovery and design. The framework is also a research instrument: the typology generates testable hypotheses, and the empirical work needed to test them produces calibrated knowledge that can feed back into the corpus.
The framework is conceptual and the empirical work has not been done. The ideas presented here emerged in a single sustained conversation and have been developed with care but without external validation. The next moves are empirical: the decomposition study to test the typology, the inter-rater reliability study to assess reproducibility, the frame-pickup analysis to map the deployment dynamics, the encyclopedia migration to operationalize the schema. Each of these is feasible. Each would either strengthen the framework or expose its weaknesses, both of which are useful outcomes.
The structural and framed primes distinction, in the end, is offered as a working tool rather than a final theory. It is an analytical lens that sharpens questions about how abstractions travel and what they bring with them. The lens may turn out to need adjustment, or it may turn out to need replacement. Either way, the questions it has helped articulate — how does an abstraction travel? what comes along for the ride? what survives extraction? what accretes in deployment? — are questions worth asking, and the Encyclopedia of Abstractions is a better project for having asked them.
References¶
[1] Williams, Bernard. Ethics and the Limits of Philosophy. Cambridge, MA: Harvard University Press, 1985. ↩
[2] McDowell, John. Mind, Value, and Reality. Cambridge, MA: Harvard University Press, 1998. ↩
[3] Väyrynen, Pekka. The Lewd, the Rude and the Nasty: A Study of Thick Concepts in Ethics. Oxford: Oxford University Press, 2013. ↩
[4] Gentner, Dedre. "Structure-Mapping: A Theoretical Framework for Analogy." Cognitive Science 7, no. 2 (1983): 155–170. ↩
[5] Gentner, Dedre. "The Mechanisms of Analogical Learning." In Similarity and Analogical Reasoning, edited by Stella Vosniadou and Andrew Ortony, 199–241. Cambridge: Cambridge University Press, 1989. ↩
[6] Hofstadter, Douglas, and Emmanuel Sander. Surfaces and Essences: Analogy as the Fuel and Fire of Thinking. New York: Basic Books, 2013. ↩
[7] Gallie, W. B. "Essentially Contested Concepts." Proceedings of the Aristotelian Society 56 (1955–1956): 167–198. ↩
[8] Cappelen, Herman. Fixing Language: An Essay on Conceptual Engineering. Oxford: Oxford University Press, 2018. ↩
[9] Burgess, Alexis, Herman Cappelen, and David Plunkett, eds. Conceptual Engineering and Conceptual Ethics. Oxford: Oxford University Press, 2020. ↩
[10] Star, Susan Leigh, and James R. Griesemer. "Institutional Ecology, 'Translations' and Boundary Objects: Amateurs and Professionals in Berkeley's Museum of Vertebrate Zoology, 1907–39." Social Studies of Science 19, no. 3 (1989): 387–420. ↩
[11] Star, Susan Leigh. "This Is Not a Boundary Object: Reflections on the Origin of a Concept." Science, Technology, & Human Values 35, no. 5 (2010): 601–617. ↩
[12] Hacking, Ian. The Social Construction of What? Cambridge, MA: Harvard University Press, 1999. ↩