Abstract Meaning Representation¶

Abstract Meaning Representation encodes sentence meaning as a rooted, labeled, directed acyclic graph of concepts and semantic relations while abstracting away from surface syntax.

Version: v1 · 2026-09-28 · History
Domain-specific #: 7831
Domain group: Humanities
Origin domain: Linguistics & Semiotics
Subdomains: Computational Linguistics, Semantic Representation → Linguistics & Semiotics

Core Idea¶

Abstract Meaning Representation is treated here as the recurring socialscienceshumanitiesarts identity summarized by this source-grounded definition: Abstract Meaning Representation encodes sentence meaning as a rooted, labeled, directed acyclic graph of concepts and semantic relations while abstracting away from surface syntax. Abstract Meaning Representation (AMR) is a semantic representation language. AMR are rooted, labeled, directed, acyclic graphs (DAGs), comprising whole sentences. Abstract Meaning Representation encodes sentence meaning as a rooted, labeled, directed acyclic graph of concepts and semantic relations while abstracting away from surface syntax.

How would you explain it like I'm…

The Same-Story Map

'The boy ate the cake' and 'The cake was eaten by the boy' say the same thing in different words. Abstract Meaning Representation draws one picture of what's happening - who did what to what - with dots and arrows. Both sentences get the same picture because they mean the same thing.

Meaning Diagrams for Sentences

Abstract Meaning Representation, or AMR, is a way of writing down what a sentence means instead of how it's worded. It draws the meaning as a diagram: circles for ideas like 'chase', 'dog' and 'cat', and labeled arrows showing how they connect, like which one is doing the chasing. It starts from one main idea at the top. If two sentences mean the same thing but are worded differently, they should get the same diagram. It was built mostly around English, so it isn't meant to work the same for every language.

Meaning Graphs Beyond Grammar

Abstract Meaning Representation (AMR) is a semantic representation language that captures a whole sentence's meaning as a graph. The graph is rooted (it has a top concept), labeled (nodes are concepts and edges are named relations like agent or object), directed, and acyclic. It deliberately ignores surface syntax: sentences with similar meaning, like active and passive versions, should receive the same AMR even though their words and grammar differ. That is what separates it from a parse tree, which records sentence structure. AMR is biased toward English and grew out of work in natural language generation, not as a universal language.

Abstract Meaning Representation is a semantic representation language in which the meaning of a whole sentence is encoded as a rooted, labeled, directed acyclic graph: nodes are concepts, labeled edges are semantic relations, and a single root anchors the structure. Its defining design choice is to abstract away from surface syntax, so that sentences with similar meaning are assigned the same AMR even when they are not identically worded. This distinguishes it from syntactic parse trees, which track the grammatical structure of a particular wording. AMR is openly biased toward English and is not intended to serve as an international auxiliary language. It was introduced by Langkilde and Knight in 1998 as a derivation of the Penman Sentence Plan Language, placing it in the natural language generation tradition, its original application domain. It later regained attention through the work of Banarescu and colleagues.

Scope of Application¶

Documented setting. By nature, the AMR language is biased towards English – it is not meant to function as an international auxiliary language.
Documented setting. Abstract Meaning Representations have originally been introduced by Langkilde and Knight (1998) as a derivation from the Penman Sentence Plan Language, they are thus continuing a long tradition in Natural Language.
Example. As far as predicate semantics are concerned, the role inventory of PropBank is largely based on semantic role annotations in the style of PropBank.
Example. In addition, they claim that this representation makes the will of the boy more explicit, highlighting that the intention of the boy is that he himself goes away (because want-01 is.
Uniform Meaning Representations. In an extension of the original AMR formalism, Uniform Meaning Representations (UMR) have been proposed.

Clarity¶

A clear use of Abstract Meaning Representation names the carrier, the operative relation, and the conditions under which the source treats the identity as present. The minimal definition is Abstract Meaning Representation encodes sentence meaning as a rooted, labeled, directed acyclic graph of concepts and semantic relations while abstracting away from surface syntax.

Manages Complexity¶

Abstract Meaning Representation compresses multiple socialscienceshumanitiesarts details into a stable diagnostic relation. The source shows both the central mechanism—as far as predicate semantics are concerned, the role inventory of PropBank is largely based on semantic role annotations in the style of PropBank.—and the practical consequence—example sentence: The boy wants to go. This compression makes cases comparable while leaving parameters, conventions, exceptions, and evidential quality explicit.

Abstract Reasoning¶

Type the carrier. Identify the socialscienceshumanitiesarts entities to which the claim applies.
State the relation. Use the source-grounded identity: Abstract Meaning Representation encodes sentence meaning as a rooted, labeled, directed acyclic graph of concepts and semantic relations while abstracting away from surface syntax.
Check operation and conditions. In addition, they claim that this representation makes the will of the boy more explicit, highlighting that the intention of the boy is that he himself goes away (because want-01 is the type of the top-level predicate). 4.

Knowledge Transfer¶

Within the home domain. Knowledge about Abstract Meaning Representation transfers literally when a new case preserves the same carrier type, relation, and recognition test. By nature, the AMR language is biased towards English – it is not meant to function as an international auxiliary language. Abstract Meaning Representations have originally been introduced by Langkilde and Knight (1998) as a derivation from the Penman Sentence Plan Language, they are thus continuing a long tradition in Natural Language Generation and this has been their original domain of application. Beyond the home domain.

Relationships to Other Abstractions¶

Current abstraction Abstract Meaning Representation Domain-specific

Parents (1) — more general patterns this builds on

Abstract Meaning Representation is a kind of Representation Prime

Abstract Meaning Representation is a strict kind of Representation: Abstract Meaning Representation encodes sentence meaning as a rooted, labeled, directed acyclic graph of concepts and semantic relations while abstracting away from surface syntax.

Hierarchy path (1) — routes to 1 parentless root

Abstract Meaning Representation → Representation → Abstraction

Neighborhood in Abstraction Space¶

Abstract Meaning Representation sits in a sparse region of the domain-specific corpus (64^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Codes, Matrices & Combinatorial Problems (30 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08