Topic Facet

Isolate what a resource is about as a single independently selectable axis in a faceted classification, so topical queries combine with place, time, and form filters at query time instead of being fused in advance into precoordinated compound headings.

Core Idea

A topic facet is the design decision to isolate what a resource is about — its subject matter — as a single independently selectable axis within a faceted classification system, so that topical queries can be combined with place, time, form, or audience filters without those dimensions being fused in advance into compound headings. The mechanism is axis decomposition: instead of precoordinating "Cooking — Italy — 19^th century" into one indivisible string, the system assigns each resource a topical value from a controlled vocabulary (Cooking), a geographic value (Italy), and a chronological value (19^th century) as three separate facets. A user can then query any combination — "this topic, any place" or "any topic, this period" — by selecting values on independent axes, and the catalog constructs the conjunction at query time rather than at indexing time. OCLC FAST operationalizes this by decomposing Library of Congress Subject Headings into five facets, of which topic is one; e-commerce filter panels and digital library interfaces follow the same structural pattern. The design criterion is genuine orthogonality: the topic axis must vary independently of the other axes so that no topical value presupposes a particular place, form, or period, and no cross-axis pairing is foreclosed by the vocabulary structure.

Structural Signature

Sig role-phrases:

• the designated topic axis — what a resource is about isolated as one description dimension, separated from place, time, form, and audience
• the controlled topical vocabulary — the (often hierarchical) value list from which a resource's subject value is drawn
• the per-resource subject selection — each resource assigned one (sometimes several) topical values rather than fused into a precoordinated compound heading
• the designed orthogonality — the load-bearing precondition: the topic axis must vary independently, no topical value presupposing a place, period, or form
• the independent retrievability — queries constrain topic without constraining the other axes and vice versa, the conjunction formed at query time
• the sum-not-product economy — separating the axis keeps the enumerated vocabulary additive in the axes rather than multiplicative across the cross-product
• the cross-facet leak — the characteristic failure: a topic term that smuggles in a region or era, making its axis non-orthogonal and its pairings unreachable or doubly-counted

What It Is Not

• Not "the subject" as a whole. The topic facet is one axis among several — pure aboutness — not the indexing system's organizing trunk. A precoordinated string like "Cooking — Italy — 19^th century" reads as "the subject," but most of it is place and period riding along; the facet isolates the topical value (Cooking) from the descriptive machinery that merely co-occurs with it.
• Not any subject term whatever. A topical value that quietly presupposes a region or an era is a cross-facet leak, not a legitimate facet value — its axis is no longer orthogonal, and its pairings become unreachable or doubly-counted. Genuine orthogonality (no topical value presupposing a place, period, or form) is the load-bearing precondition, so a leaking term is a defect, not an accepted feature of the heading.
• Not the general form/content distinction. form_and_content is the broad expression-versus-aboutness relation; the topic facet is the narrow operationalization of the content half as one retrievable axis inside a controlled-vocabulary system. The facet presupposes a controlled topical vocabulary, per-resource subject selection, and a faceted interface — apparatus the general distinction does not carry.
• Not the whole faceted vocabulary. It is a single designated axis (subject/aboutness) within a faceted system, not the faceting design itself; its only distinctive content beyond the parent is the choice of which axis. Treating "topic facet" as the entire orthogonal-decomposition apparatus overstates it — the structure that gives it force is paradigmatic_vs_syntagmatic_relations, the parent.
• Not a catalog that pre-mints every topic-place-period combination. Giving topic its own axis collapses the cross-product into a sum: the catalog forms the conjunction at query time, so "this topic, any place" is a single selection with the other axes left free, and combinations need never have been anticipated at indexing time. Expecting a stored compound heading for each pairing mistakes the design for the precoordinated explosion it replaces.

Scope of Application

The topic facet lives across the classification-and-retrieval family — one substrate, retrieval-oriented descriptive systems; the substrate-spanning orthogonal-decomposition lesson (separation of variables, basis selection, parametric design) belongs to the paradigmatic_vs_syntagmatic_relations parent, not to the subject axis itself.

• Library cataloging — OCLC FAST and faceted-LCSH split traditional precoordinated headings into topic, place, time, and form facets, each with its own controlled value list.
• Faceted search interfaces — e-commerce filter panels and digital-library catalogs separate category/topic from brand, price, format, and the rest, forming the conjunction at query time.
• Web content taxonomies — news sites and Wikipedia categories separate subject topic from format (image, list, biography) and from geography.

Clarity

Naming the topic facet makes one design choice in subject access stand out as a choice rather than a given: subject matter is one axis among several, not the indexing system's organizing trunk. Without the label, a cataloger working in a precoordinated tradition treats "Cooking — Italy — 19^th century" as a single subject heading and reads the whole string as "the subject," so the question of whether topical access could be separated from place or period never surfaces — the dimensions are fused before anyone asks whether they should be. Isolating topic as a designated facet dissolves that fusion and lets a designer ask the sharper question: which of the dimensions a precoordinated heading welds together are genuinely about aboutness, and which are place, time, form, or audience riding along? It separates the topical value (Cooking) from the descriptive machinery that merely co-occurs with it.

It also sharpens the distinction between precoordination and postcoordination at exactly the point where it bites. Once topic is its own axis with its own controlled value list, the system can construct "this topic, any place" or "any topic, this period" as a conjunction at query time, and the combinatorial cost of enumerating compound headings in advance disappears as a cataloging burden. The crisp design test the facet supplies is orthogonality: does the topical vocabulary vary independently of the other axes, so that no topical value presupposes a particular place, form, or period and no cross-axis pairing is foreclosed? A topic term that quietly implies a region or an era is a leak across facets, and naming the topic facet is what makes that leak visible as a defect rather than an accepted feature of the heading.

Manages Complexity

A precoordinated subject vocabulary faces a combinatorial sprawl: every topic that could co-occur with every place, every period, and every form has to be anticipated and minted as its own compound heading, so the heading list grows as the product of the dimensions and a cataloger must either pre-enumerate "Cooking — Italy — 19^th century," "Cooking — France — 18^th century," "Architecture — Italy — 19^th century," and so on across the whole cross-product, or leave the un-minted combinations unreachable. Recognizing topic as one independently selectable facet collapses that product into a sum. Instead of tracking a heading space whose size is the multiplied cardinality of all the axes, the designer tracks each axis separately — a topical value list, a geographic value list, a chronological value list — and lets the catalog form the conjunction at query time. The quantities that determine the system's behavior shrink to a handful the designer can hold at once: how many axes there are, the size of each axis's controlled vocabulary, and one structural property per axis — does it vary independently of the others, or does some topical value smuggle in a place or a period? From those, the qualitative outcomes read straight off. If the axes are genuinely orthogonal, every cross-axis pairing is automatically reachable without being pre-minted, the catalog's enumerated vocabulary stays linear in the axes rather than multiplicative, and a query like "this topic, any place" is a single selection on one axis with the rest left free. If a facet leaks — a topic term that presupposes a region or an era — that one parameter flips and predicts the failure: the affected pairings become unreachable or doubly-counted, and the conjunction the user wanted cannot be constructed cleanly. The analyst no longer reasons case by case about which compound headings exist; they reason about a few axis-level properties and derive the retrieval behavior of the entire combination space.

Abstract Reasoning

Designating topic as a facet licenses a tight set of reasoning moves about a classification system's behavior, all turning on the orthogonality of the topic axis against the others.

Diagnostic — infer a cross-facet leak from a retrieval failure. When a user query like "this topic, any place" comes back with too few hits, or when certain topic-place pairings are systematically empty while the resources demonstrably exist, the cataloger reasons backward to a contaminated vocabulary: a topical value that silently presupposes a region or an era. The surface signature is an unreachable or doubly-counted pairing; the hidden cause inferred from it is that the topic term is not pure aboutness but carries place, period, or form riding along, so the axis it lives on is not actually independent. The classic instance is the precoordinated heading read as "the subject": from the observation that "Cooking — Italy — 19^th century" cannot be queried for "Cooking, any place" without enumerating every minted variant, one infers the dimensions were fused before indexing, not separable at query time. A topic value that never co-occurs with more than one value of another facet is the fingerprint of a leak.

Interventionist — predict the retrieval consequence of an axis edit. The facet design supplies a small repertoire of moves with forecastable effects. Splitting a fused heading into a topical value plus a geographic plus a chronological value predicts that every cross-axis pairing becomes reachable by conjunction at query time without being pre-minted, so previously unreachable combinations light up. Tightening a topic value that leaked place back into a pure-aboutness term predicts that the affected pairings stop collapsing and the "any place" query recovers its full result set. Coarsening or refining the granularity of the topical value list predicts a shift in recall versus precision on topical queries while leaving the other axes' behavior untouched — the orthogonality is precisely what guarantees the edit stays local to the one axis. Each move is a claim about the combination space that can be checked by re-running the conjunction.

Boundary-drawing — decide when the facet decomposition is the right design. The topic facet earns its keep only where the topical vocabulary genuinely varies independently of place, time, and form and where the cross-product of dimensions is large enough that precoordinating compound headings is infeasible. Where a topical value is inherently bound to one region or era — where the "aboutness" cannot be cleanly separated from where or when — the axis is not orthogonal and faceting it buys nothing the fused heading did not already provide. So the move is to test, before committing: does any topical value presuppose a value on another axis? If yes, the dimension is not a clean facet and the decomposition will leak; if no, postcoordination dominates precoordination and the facet should be split off.

Predictive — read the whole combination space off the axis properties. Given orthogonal axes, the system's retrieval behavior follows without case-by-case enumeration: every pairing of a topical value with any value on another axis is automatically constructable, the enumerated vocabulary stays additive in the axes rather than multiplicative, and any "this topic, free on the rest" query resolves to a single selection. Flip the independence property on one axis and the prediction inverts for exactly the pairings that axis touches. The order of operations the facet asserts is the load-bearing claim: the catalog forms the conjunction at query time, so combinations need never have been anticipated at indexing time — a topic value plus the freedom to leave the other axes unconstrained is sufficient to reach anything the orthogonal structure permits.

Knowledge Transfer

Within the classification-and-retrieval family the topic facet transfers as mechanism. The orthogonality test (does any topical value presuppose a place, period, or form?), the cross-facet-leak diagnostic, and the precoordination-versus-postcoordination reasoning carry intact across every subfield that gives subject matter its own axis. In library cataloging, OCLC FAST and faceted-LCSH split traditional precoordinated headings into topic, place, time, and form facets, each with its own value list. In faceted search interfaces, e-commerce filter panels and digital-library catalogs separate category/topic from brand, price, format, and the rest. In web content taxonomies, news sites and Wikipedia categories separate subject topic from format and geography. A designer who learns to read "this topic, any place" as a single selection on one orthogonal axis — and to diagnose a systematically empty topic-place pairing as a leaked topic term carrying place along — carries that reasoning without translation across all of these, because they are one substrate: retrieval-oriented descriptive systems.

Beyond that substrate the case is (B): the topic facet contributes no portable mechanism of its own — it is wholly a specialization, and the parent is what travels. The structure that gives the topic facet its force, decompose a description into orthogonal independently combinable axes so a description is a per-axis selection rather than one fused label, is exactly paradigmatic_vs_syntagmatic_relations (selection-axis versus combination-axis), the same parent that powers separation of variables in mathematics, basis selection in linear algebra, principal-component decomposition in statistics, and parametric design in engineering. The topic facet inherits everything from that parent and adds only the choice of which axis (subject/aboutness) and the catalog machinery around it: a controlled topical vocabulary, per-resource subject selection, and a faceted-retrieval interface. So what a reader could carry into physics or software architecture is the general orthogonality lesson, which belongs to the parent prime, not to "topic facet." Near-neighbor form_and_content is the broader expression-versus-aboutness distinction; the topic facet is the narrow operationalization of the content half as one retrievable axis inside a controlled-vocabulary system. Invoking "a topic facet" for any aboutness-versus-everything-else split outside cataloging renames the parts and borrows the parent's orthogonality while dropping the classification apparatus — the cross-domain lift is the parent's, and the named LIS axis should be reserved for the retrieval substrate where its vocabulary and interface do real work (see Structural Core vs. Domain Accent).

Examples

Canonical

OCLC's FAST (Faceted Application of Subject Terminology) is the defining construction. Traditional Library of Congress Subject Headings are precoordinated: a resource about nineteenth-century Italian cooking receives one indivisible string such as "Cooking — Italy — 19^th century," in which subject, place, and period are welded together in a fixed order. FAST takes that same intellectual content and decomposes it into separate faceted headings drawn from distinct value lists: a topical heading (Cooking), a geographic heading (Italy), and a chronological heading (19^th century), each assignable on its own. A catalog built this way can then answer "cooking, any place" by selecting one topical value and leaving the geographic axis unconstrained — a query the fused LCSH string cannot serve without a separately minted heading for every place-and-period combination. Topic is thus isolated as one independently selectable axis.

Mapped back: Extracting "Cooking" from the compound string is the designated topic axis drawn from the controlled topical vocabulary, assigned by the per-resource subject selection. That the same resource also carries an independent Italy value is the designed orthogonality, and answering "cooking, any place" as one selection is the independent retrievability — the conjunction formed at query time rather than pre-minted.

Applied / In Practice

Faceted navigation in online retail is the everyday field deployment of the same structure. An e-commerce catalog presents a shopper with a left-hand filter panel whose facets are separate axes — department/topic, brand, price range, size, color, customer rating — and each product is tagged with a value on each axis independently. The shopper builds a query by selecting values across axes ("running shoes, any brand, under $100") and the system constructs the conjunction at retrieval time, so the store never has to pre-enumerate a distinct category for every brand-price-color combination. The catalog's stored vocabulary stays additive across facets rather than exploding multiplicatively, and any facet left unselected simply remains free. When a topical category quietly presupposes another axis — a "department" that only ever contains one brand — that pairing collapses, the tell of a leaked axis.

Mapped back: The department/topic filter is the designated topic axis kept separate from brand and price; per-product tagging is the per-resource subject selection. Combining selections at retrieval time is the independent retrievability, and never pre-minting every combination is the sum-not-product economy — while a category bound to one brand is the cross-facet leak.

Structural Tensions

T1: Orthogonality as design ideal versus aboutness that is genuinely entangled (the leak may be the world, not the vocabulary). The topic facet's load-bearing precondition is that subject matter vary independently of place, time, and form. But real aboutness is often inseparable from where and when: "the Risorgimento" is intrinsically Italy-and-nineteenth-century, "Kabuki" carries Japan, and many topics simply do not factor into a pure content axis plus free geographic and chronological axes. The tension is that orthogonality is both the facet's enabling assumption and a condition the material frequently violates, so a cross-facet leak that the design reads as a vocabulary defect may instead be a faithful reflection of a topic that genuinely presupposes a region or era. Forcing such a topic into an artificially "pure" term to satisfy orthogonality can misrepresent the subject; accepting the leak forfeits the clean conjunction. Diagnostic: Is the topic value that presupposes a place or period a fixable vocabulary contamination, or a topic whose aboutness is genuinely bound to that dimension, so that orthogonalizing it would falsify the subject?

T2: Postcoordination reachability versus precoordination precision (the sum economy loses expressive power). Decomposing a fused heading into independent axes collapses the combinatorial cross-product into a sum and makes every pairing reachable at query time — the facet's central economy. But a query-time conjunction of "Cooking AND Italy AND 19^th century" is semantically coarser than a curated precoordinated heading, which can encode a relationship among the parts (aboutness in a specific role, a subject treated as influenced by a place) that a mere co-selection of independent values cannot express. The tension is that faceting buys combinatorial completeness at the cost of the relational precision precoordination could carry, so the postcoordinate system reaches more combinations while saying less about how the parts relate within any one of them. The pre/post-coordination trade is not a solved question but a standing choice between reachability and expressive specificity. Diagnostic: Does the retrieval need only the conjunction of independent values (faceting wins) or the specific relationship among topic, place, and period that only a precoordinated heading encodes (precision is lost in decomposition)?

T3: Topic as a clean axis versus aboutness as a contestable partition (which dimension is "the topic"?). The facet presupposes that "what a resource is about" can be cleanly separated from place, time, form, and audience — that subject is one natural axis among several. But the partition into facets is itself a design judgment, and "aboutness" is not a natural kind: a resource's genre can be its subject (a book about biography), its audience can be topical (children's science), and reasonable designers draw the topic/form/audience boundaries differently. The tension is that the topic facet's clarity depends on a facet partition that is itself contestable, so isolating "the topic" quietly imports decisions about where one axis ends and another begins that the material does not settle. The move that makes subject access tractable — treating topic as its own trunk-free axis — rests on a carving of the description space that is engineered, not given. Diagnostic: Is the topic/place/form/audience partition being treated as a fixed fact about the resource, or as a design carving whose boundaries a different classifier could reasonably draw elsewhere?

T4: Vocabulary granularity versus retrieval quality (a per-axis edit that trades recall against precision). Because the axes are orthogonal, coarsening or refining the topical value list is a local edit that leaves the other facets untouched — a genuine benefit of the decomposition. But that same locality means the granularity of the topic vocabulary directly sets a recall-versus-precision trade on topical queries: broad terms retrieve widely but imprecisely, narrow terms retrieve precisely but miss near-matches and burden the cataloger and the searcher with finer distinctions. The tension is that the facet gives the designer a clean knob on the topic axis while offering no principle for where to set it, so every choice of topical granularity buys recall at the cost of precision or the reverse, and the orthogonality that makes the edit safe does nothing to make it optimal. Diagnostic: Is the topical vocabulary's granularity tuned to the collection's actual query behavior, or set by convention in a way that silently sacrifices recall or precision on subject queries?

T5: Autonomy versus reduction (a named cataloging axis or a pure specialization of paradigmatic-versus-syntagmatic relations). The topic facet is a named library-and-information-science construct with real apparatus — a controlled topical vocabulary, per-resource subject selection, a faceted-retrieval interface, OCLC FAST's five-facet decomposition. But it is, unusually, wholly a specialization: the structure that gives it force — decompose a description into orthogonal independently combinable axes so a description is a per-axis selection rather than one fused label — is exactly paradigmatic_vs_syntagmatic_relations, the same parent behind separation of variables, basis selection, principal-component decomposition, and parametric design. The topic facet inherits everything from that parent and adds only the choice of which axis (subject/aboutness) and the catalog machinery around it, with form_and_content as the broader neighbor. The tension is sharper than usual: there is almost no portable mechanism proprietary to the topic facet, so the entire cross-domain orthogonality lesson belongs to the parent, and the named axis earns its keep only inside retrieval-oriented descriptive systems where its vocabulary and interface do real work. Diagnostic: Resolve toward paradigmatic_vs_syntagmatic_relations whenever the lesson is orthogonal-axis decomposition outside cataloging; toward the topic facet only where a controlled subject vocabulary is retrieved as one axis within a faceted classification system.

Structural–Framed Character

Topic facet sits at the framed-leaning position on the structural–framed spectrum: an evaluatively light but engineered design construct constituted by the practice of building classification-and-retrieval systems and codified within library-and-information science, resting on a portable orthogonal-decomposition skeleton that keeps it off the framed pole. The criteria mostly point framed. Its evaluative weight is low and structural-leaning: a topic facet is a design decision, not a verdict — it describes an axis of a classification, not a defect (the only evaluative note is the "leak," a design failure, which is a local diagnostic rather than the concept's core). Every other criterion points framed. It is strongly human-practice-bound: the topic facet exists only as a choice within a descriptive retrieval system built by catalogers for queriers — its whole reality is that "what a resource is about" has been engineered into an independently selectable axis with a controlled vocabulary and a faceted interface — so strip away the practice of classifying resources for retrieval and there is no topic facet, only resources; nothing here is a fact nature holds. Its institutional origin is pronounced: the concept is LIS apparatus — OCLC FAST's five-facet decomposition, faceted-LCSH, the precoordination/postcoordination distinction — an artifact of a cataloging tradition. On vocab_travels it is domain-pinned: the orthogonality test and leak diagnostic carry across retrieval-oriented descriptive systems but not off that substrate, where the parent pattern does the work under other names. And on import_vs_recognize it patterns as recognition within cataloging and as the parent doing the lifting beyond it.

The structural-looking feature is unusually total here, and the entry is candid that the topic facet is wholly a specialization contributing almost no proprietary mechanism: the structure that gives it force — decompose a description into orthogonal, independently combinable axes so a description is a per-axis selection rather than one fused label — is exactly paradigmatic_vs_syntagmatic_relations (selection-axis versus combination-axis), the same parent that powers separation of variables, basis selection, principal-component decomposition, and parametric design (with form_and_content as the broader expression-versus-aboutness neighbor). That orthogonal-decomposition skeleton is genuinely substrate-spanning. But it does not lift the topic facet off the framed side, because axis-decomposition is precisely what the topic facet instantiates from that umbrella prime, adding only the choice of which axis (subject/aboutness) and the catalog machinery around it: the cross-domain orthogonality lesson belongs entirely to the parent, while the controlled topical vocabulary, per-resource subject selection, faceted-retrieval interface, and precoordination/postcoordination framing are domain accent that stays home. Its character: an evaluatively light but wholly practice-constituted, LIS-codified design construct, structural only in the paradigmatic-axis-decomposition skeleton it inherits nearly whole from its parent and specializes to the subject axis of a faceted classification.

Structural Core vs. Domain Accent

This section decides why topic facet is a domain-specific abstraction and not a prime, and it carries the case for its domain-specificity in the same move — an unusually clean case, since the entry contributes almost no proprietary mechanism of its own.

What is skeletal (could lift toward a cross-domain prime). Strip the cataloging and a thin relational structure survives: decompose a description into orthogonal, independently combinable axes, so that a description is a per-axis selection whose conjunction is formed at use time rather than one fused label pre-minted in advance. The portable pieces are abstract — a set of independent axes, a per-axis value selection, an orthogonality precondition, and a combinatorial economy in which the enumerated vocabulary stays additive across axes rather than multiplicative over their cross-product. That skeleton is exactly paradigmatic_vs_syntagmatic_relations (selection-axis versus combination-axis), with form_and_content the broader expression-versus-aboutness neighbor. It is genuinely substrate-spanning — the same orthogonal-decomposition powers separation of variables in mathematics, basis selection in linear algebra, principal-component decomposition in statistics, and parametric design in engineering — which is why it is the core the topic facet instantiates, and, the entry is candid, nearly all it amounts to.

What is domain-bound. What makes the concept topic facet in particular is library-and-information-science furniture, and it is thin: the concept inherits everything from its parent and adds only the choice of which axis (subject/aboutness) plus the catalog machinery around it. That machinery is the domain accent — the controlled topical vocabulary the subject value is drawn from; the per-resource subject selection; the faceted-retrieval interface; the precoordination-versus-postcoordination framing; and the concrete apparatus (OCLC FAST's five-facet decomposition of LCSH, faceted-LCSH, the cross-facet-leak diagnostic). The decisive test is the constitutive practice: the topic facet exists only as a design decision within a descriptive retrieval system built by catalogers for queriers — "what a resource is about" has been engineered into an independently selectable axis — so strip away the practice of classifying resources for retrieval and there is no topic facet, only resources, with none of the vocabulary or interface that gives the axis its work.

Why this does not clear the prime bar. A prime is a relational structure whose vocabulary travels and whose cross-domain transfer is recognition of the same mechanism, not analogy. The topic facet's transfer is bimodal, and unusually lopsided. Within the classification-and-retrieval family it transfers as mechanism — the orthogonality test, the cross-facet-leak diagnostic, and the precoordination/postcoordination reasoning carry intact across library cataloging (OCLC FAST), faceted search interfaces (e-commerce filter panels, digital-library catalogs), and web content taxonomies, because they are one substrate of retrieval-oriented descriptive systems (recognition). Beyond that substrate the topic facet contributes no portable mechanism of its own: invoking "a topic facet" for any aboutness-versus-everything-else split outside cataloging renames the parts and borrows the parent's orthogonality while dropping the classification apparatus — analogy at the level of the named axis. And when the bare structural lesson is wanted cross-domain — decompose into orthogonal independently combinable axes — it is already carried, in fully general form, by the parent the topic facet specializes: paradigmatic_vs_syntagmatic_relations (with form_and_content the broader neighbor). The cross-domain reach belongs entirely to that parent; "topic facet," as named, earns its keep only inside retrieval systems, keeping the controlled subject vocabulary, per-resource selection, faceted interface, and precoordination/postcoordination framing as accent that stays home.

Relationships to Other Abstractions

Current abstraction Topic Facet Domain-specific

Parents (2) — more general patterns this builds on

• Topic Facet is a kind of Retrieval Facet Domain-specific

  Topic Facet is the aboutness species of Retrieval Facet.

• Topic Facet is a decomposition of Form and Content Prime

  Topic Facet operationalizes the content or aboutness side of the general form-content distinction as an independently retrievable catalog axis.

Children (1) — more specific cases that build on this

• Faceted Vocabulary Domain-specific is part of, conditional Topic Facet

  A Topic Facet is an internal axis of a faceted vocabulary when the scheme exposes aboutness independently from place, time, and form.

Hierarchy paths (7) — routes to 5 parentless roots

• Topic Facet → Retrieval Facet → Classification

Show alternative paths (6)

Not to Be Confused With

• Precoordinated subject heading (LCSH-style). A single fused compound string — "Cooking — Italy — 19^th century" — that welds topic, place, and period into one indivisible heading minted in advance. The topic facet is the postcoordinated alternative: it isolates the aboutness value (Cooking) as its own axis and forms the conjunction at query time. Tell: is subject access delivered as one pre-minted compound string (precoordinated heading), or as an independently selectable subject axis combined with others at query time (topic facet)?

• The sibling facets (place, time, form, audience). The other axes in the same faceted system — geographic, chronological, form, and audience facets. The topic facet is specifically the aboutness axis; its whole point is that place, time, and form are not part of it but ride on separate, orthogonal axes. Tell: is the value describing what the resource is about (topic facet), or where/when/in-what-form/for-whom it is (the place/time/form/audience facets)?

• Form/content distinction (form_and_content). The broad expression-versus-aboutness relation. The topic facet is the narrow operationalization of the content half as one retrievable axis inside a controlled-vocabulary system; the general distinction carries no controlled vocabulary, per-resource selection, or faceted interface. A near-neighbor, one level up. Tell: is the referent the broad aboutness-vs-expression relation (form_and_content), or its concrete realization as a retrievable subject axis in a faceted catalog (topic facet)?

• Folksonomy / free-text tagging. User-supplied, uncontrolled keywords with no fixed vocabulary or axis structure. The topic facet draws its value from a controlled topical vocabulary on a designated orthogonal axis; tags are flat, open, and not orthogonalized into independent dimensions. Tell: are subject terms drawn freely by users with no controlled list or axis discipline (folksonomy), or selected from a controlled vocabulary as one orthogonal facet (topic facet)?

• Enumerative taxonomy / single-hierarchy classification. A one-tree scheme (like a strict Dewey-style hierarchy) that places each resource at a single node in one dominant hierarchy, often fusing dimensions into that path. Faceting decomposes description into multiple independent axes instead of one enumerated tree; the topic facet is one such axis. Tell: is a resource slotted into one node of a single dominant hierarchy (enumerative taxonomy), or tagged independently across several orthogonal facets including subject (topic facet)?

• Paradigmatic vs. syntagmatic relations (the parent it instantiates). The substrate-neutral structure — decompose a description into orthogonal, independently combinable axes so a description is a per-axis selection rather than one fused label — that powers separation of variables, basis selection, PCA, and parametric design. Not a confusable peer but the umbrella, and the topic facet is wholly a specialization of it (adding only the subject axis and catalog machinery). Tell: outside cataloging, the portable content is this orthogonal-decomposition parent — treated more fully elsewhere — while the controlled subject vocabulary and faceted interface are the topic facet's home-bound accent.

Neighborhood in Abstraction Space

Topic Facet sits in a crowded region of the domain-specific corpus (12^th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Faceted Classification & Metadata (6 abstractions)

Nearest neighbors

• Form Facet — 0.90
• Method-as-Metadata — 0.89
• Geographic Facet — 0.89
• Faceted Vocabulary — 0.88
• Core Vocabulary — 0.85

Computed from structural-signature embeddings · 2026-07-12