Functional Generative Description

A stratified dependency-based linguistic framework that links surface form to a tectogrammatical dependency representation integrating valency, semantic roles, coreference, and topic–focus articulation.

Version

v3 · 2026-09-19 · History

Domain-specific #

1897

Domain group

Humanities

Origin domain

Linguistics & Semiotics

Subdomain

Functional Dependency Grammar → Linguistics & Semiotics

Aliases

FGD, Functional Generative Description framework

Core Idea

Functional Generative Description (FGD) is a formal linguistic framework developed in the Prague tradition. It represents an utterance through linked descriptive layers and treats the tectogrammatical layer as the underlying representation of sentence meaning. At that layer, lexical meaning units are nodes in a dependency tree, labeled with deep syntactic functions and enriched with valency, coreference, grammatical meaning, and information structure.

FGD is generative in the technical sense that its system of units and relations specifies how representations at one layer correspond to those at another. It is functional because communicative organization—especially topic–focus articulation—is part of the grammatical description rather than an optional discourse gloss.

How would you explain it like I'm…

The Meaning Tree

When you say a sentence, there are the sounds you make, and underneath there is what you mean. Some language scientists draw that hidden meaning as a tree, where each meaning word is a branch hooked onto the word it belongs to. They also mark which part is the old news you already know and which part is the new news.

Layers of Sentence Meaning

Functional Generative Description is a way of describing language that looks at a sentence in several connected layers, from the outside form down to the deepest meaning. At the deepest layer, the meaning is drawn as a tree: each important word is a node, and lines show which words depend on which, labeled with jobs like 'who did it' or 'what got done'. The rules of the system say how each layer matches up with the next one. It also counts 'what are we talking about' versus 'what's the new point' as a real part of grammar, not an extra.

Layered Dependency Meaning Grammar

Functional Generative Description (FGD) is a formal grammar framework from the Prague linguistic tradition. It describes an utterance on several linked layers. The deepest one, the tectogrammatical layer, represents sentence meaning as a dependency tree. In that tree, content words are nodes labeled with deep syntactic roles such as actor or patient, plus information on valency, coreference and grammatical meaning. It is 'generative' because it states exactly how a representation on one layer corresponds to one on the next. It is 'functional' because the topic–focus split counts as part of the grammar itself, not as an afterthought added for discourse.

 

Functional Generative Description is a Prague-school formal framework that models an utterance through stratified, linked levels of representation. The central level is the tectogrammatical layer, taken as the underlying representation of sentence meaning. At this level, content-bearing lexical units are nodes of a dependency tree (not a constituency tree), and the edges carry deep syntactic function labels. Nodes are further annotated with valency frames, coreference links, grammatical meanings such as tense or number, and information structure. 'Generative' is meant technically: the system of units and relations specifies the correspondences between representations on adjacent layers. 'Functional' reflects that communicative organization, especially topic–focus articulation, is a constituent of grammatical description rather than an optional pragmatic add-on. The framework is recognized by this combination: linked strata, dependency organization, a tectogrammatical semantic structure, and integrated topic–focus articulation.

Scope of Application

FGD supports theoretical syntax and semantics, valency lexicons, discourse and information-structure analysis, multilingual comparison, and richly annotated corpora. Its best-known operationalization is the Prague Dependency Treebank, where analytical and tectogrammatical annotations allow researchers and parsers to compare surface realization with deeper dependency structure.

The framework can guide computational analysis, but an implementation must declare which version of the annotation guidelines it follows because inventories and operational conventions evolve.

Clarity

An FGD analysis should state the layer, node inventory, dependency labels, valency source, treatment of ellipsis, coreference policy, and topic–focus ordering convention. A surface dependency tree must not be presented as if it were already tectogrammatical.

Manages Complexity

Layering prevents phonological, morphological, surface-syntactic, semantic, and discourse distinctions from being forced into one overloaded structure. Correspondence links preserve their relationship, while the tectogrammatical layer concentrates those features needed for meaning-oriented comparison across different surface realizations.

Abstract Reasoning

1. Segment and analyze the observed utterance at lower linguistic layers.
2. Construct the analytical dependency representation of surface syntax.
3. Identify lexical meaning units and restore licensed deletions where the theory requires them.
4. Project the structure into a tectogrammatical dependency tree.
5. Assign valency-based functors and grammatical-semantic attributes.
6. Resolve or mark coreference relations.
7. Encode contextual boundness and topic–focus articulation in deep word order.
8. Validate correspondence among layers and against the annotation manual.

Knowledge Transfer

The portable lesson is to separate representational levels while keeping explicit maps among them: surface evidence and semantic interpretation need not occupy the same structure. The proposed immediate parent is Dependency Grammar.

Relationships to Other Abstractions

Current abstraction Functional Generative Description Domain-specific

Parents (1) — outgoing typed relations

• Functional Generative Description is a kind of Dependency Grammar Domain-specific

  Dependency Grammar is the broader abstraction this entry instantiates.

Hierarchy path (1) — routes to 1 parentless root

• Functional Generative Description → is a kind of Dependency Grammar → presupposes Dependency

Neighborhood in Abstraction Space

Functional Generative Description sits in a sparse region of the domain-specific corpus (83^rd percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Language Structure & Communication Channels (15 abstractions)

Nearest neighbors

• Agglutination — 0.82
• Nanosyntax — 0.82
• Abstract Syntax Tree — 0.81
• Linguistic Bootstrapping — 0.81
• Lexical Converse Relation — 0.81

Computed from structural-signature embeddings · 2026-10-25