Core arguments are the essential noun phrases that complete the inherent meaning of a predicate, typically a verb. Unlike adjuncts, which provide circumstantial details like time or location, core arguments are syntactically and semantically required. In the sentence 'The engineer deployed the model on the server,' the core arguments are 'The engineer' (Agent) and 'the model' (Patient); removing either renders the clause ungrammatical or semantically vacuous.
Glossary
Core Arguments

What is Core Arguments?
Core arguments are the semantically obligatory participants required by a verb to form a grammatically complete and logically coherent clause, distinguishing essential actants from optional peripheral adjuncts.
In Semantic Role Labeling (SRL) frameworks like PropBank and VerbNet, core arguments are assigned numbered roles (Arg0, Arg1) that map to verb-specific or generalized thematic roles. The distinction between core arguments and adjuncts is critical for semantic parsing and Abstract Meaning Representation (AMR), as it defines the mandatory nodes in a predicate's logical structure, enabling accurate extraction of 'who did what to whom' from unstructured text.
Key Characteristics of Core Arguments
Core arguments are the syntactically obligatory constituents required by a verb's semantic valency. Unlike optional adjuncts, these participants complete the predicate's meaning and define the clause's fundamental structure.
Syntactic Obligatoriness
Core arguments are syntactically required by the verb; their omission renders a clause ungrammatical or radically alters its meaning. This distinguishes them from adjuncts, which provide optional circumstantial information.
- Test: The did so pro-form test replaces the verb phrase but retains core arguments
- Example: In 'She placed the book on the shelf,' removing the book yields an ungrammatical clause
- Contrast: 'She read the book [in the garden]' — the bracketed adjunct can be freely omitted
- Valency: Verbs like give are trivalent, requiring three core arguments (Agent, Theme, Recipient)
Null Arguments and Implicit Participants
Core arguments may be syntactically unrealized yet semantically present. These null arguments are licensed by specific syntactic environments and language-specific parameters.
- Pro-drop: Languages like Spanish and Italian allow subject pronouns to be omitted ('pro hablo español')
- Object drop: 'John ate ∅' — the food argument is implicit but recoverable from context
- Passive demotion: 'The window was broken ∅' — the Agent is syntactically suppressed
- Control constructions: 'Mary promised PRO to leave' — the unpronounced subject of leave is controlled by Mary
- Imperatives: 'Close ∅ the door!' — the addressee Agent is implicit
Selectional Preferences
Predicates impose semantic type constraints on their arguments, restricting the ontological category of entities that can logically fill a role. These preferences are graded, not absolute.
- Violation detection: 'The rock ate the sandwich' — rock violates the animacy preference of eat
- Metonymy resolution: 'The White House issued a statement' — the location stands for the institution
- Coercion: 'She began the book' — book is coerced into an event reading (reading/writing)
- Modeling: Selectional preferences are learned from distributional statistics over large corpora
- Application: Used in SRL to prune impossible argument candidates and improve classification accuracy
Distinction from Adjuncts
Core arguments are licensed by the predicate's lexical entry, while adjuncts are freely attachable modifiers that add circumstantial detail without being subcategorized for.
- Licensing: Arguments are specified in the verb's lexical entry; adjuncts are not
- Iterability: Adjuncts can be stacked ('She sang [in the park] [on Tuesday] [with a microphone]')
- Mobility: Adjuncts move more freely in the clause than arguments
- VP-ellipsis: 'She read the book in the garden, and he did so in the kitchen' — adjuncts survive ellipsis
- Semantic scope: Arguments complete the predicate; adjuncts modify the entire event
Frequently Asked Questions
Clarifying the essential participants required by a verb to form a grammatically complete clause, as distinct from optional peripheral adjuncts.
A core argument is a noun phrase or complement that is semantically and syntactically required by the valency of a predicate to form a grammatically complete clause. Unlike optional adjuncts (which provide circumstantial details like time or location), core arguments are essential participants in the event or state described by the verb. For example, the verb 'devour' requires both an Agent (the eater) and a Patient (the thing eaten); omitting either results in an ungrammatical sentence. In PropBank annotation, these are labeled as Arg0 and Arg1, while in FrameNet, they correspond to Core Frame Elements. The distinction between arguments and adjuncts is fundamental to Semantic Role Labeling (SRL) and predicate-argument structure analysis.
Core Arguments vs. Adjuncts
A comparative analysis of the syntactic and semantic properties that distinguish obligatory core arguments from optional peripheral adjuncts in predicate-argument structure.
| Feature | Core Arguments | Adjuncts |
|---|---|---|
Obligatoriness | Required for grammatical completeness | Optional; can be omitted without ungrammaticality |
Selection by Predicate | Licensed and semantically selected by the specific verb | Not selected by the verb; compatible with any predicate |
Thematic Role Assignment | Receives a theta-role (Agent, Patient, Theme) | Does not receive a theta-role from the verb |
Iterability | Cannot be iterated (only one Agent per clause) | Can be iterated (multiple locatives or temporals) |
Syntactic Position | Occupies argument positions (subject, object) | Occupies peripheral positions (adverbial clauses) |
Semantic Contribution | Completes the core proposition | Adds circumstantial information (time, location, manner) |
Extraction from Wh-Islands | Strong island violations when extracted | Weaker island effects; easier to extract |
Passivization | Can be promoted to subject in passive voice | Cannot be promoted to subject via passivization |
Enabling Efficiency, Speed & Accuracy
Intelligent Analysis, Decision & Execution
We build AI systems for teams that need search across company data, workflow automation across tools, or AI features inside products and internal software.
Talk to Us
Search across company data
Give teams answers from docs, tickets, runbooks, and product data with sources and permissions.
Useful when people spend too long searching or get different answers from different systems.

Automate internal workflows
Use AI to route work, draft outputs, trigger actions, and keep approvals and logs in place.
Useful when repetitive work moves across multiple tools and teams.

Add AI to products and internal tools
Build assistants, guided actions, or decision support into the software your team or customers already use.
Useful when AI needs to be part of the product, not a separate tool.
Examples of Core Arguments by Verb Class
Core arguments are the essential participants required by a verb to form a grammatically complete clause. The number and type of arguments vary systematically by verb class, as defined in resources like VerbNet.
Intransitive Verbs (1 Argument)
These verbs require only a single core argument, typically the entity undergoing the action or state.
- Unergative verbs: The subject is an active, volitional Agent.
- Example: "The engine roared." (Agent)
- Unaccusative verbs: The subject is a Patient or Theme that undergoes a change of state or location.
- Example: "The window shattered." (Patient)
- Example: "The package arrived." (Theme)
Transitive Verbs (2 Arguments)
These verbs require two core arguments: a subject and a direct object. The semantic roles map to the Agent-Patient prototype.
- Agent + Patient: The subject acts upon the object.
- Example: "The technician calibrated the sensor." (Agent calibrates Patient)
- Agent + Theme: The subject moves or creates the object.
- Example: "The compiler generated the bytecode." (Agent generates Theme)
- Experiencer + Stimulus: The subject perceives or feels the object.
- Example: "The model detected an anomaly." (Experiencer detects Stimulus)
Ditransitive Verbs (3 Arguments)
These verbs require three core arguments, typically involving transfer of possession or communication. The structure maps to Agent, Theme, and Recipient.
- Dative Alternation: The Recipient can appear as an indirect object or a prepositional phrase.
- Example: "The admin assigned the user a new token." (Agent assigns Recipient Theme)
- Alternation: "The admin assigned a new token to the user." (Agent assigns Theme to Recipient)
- Benefactive Construction: An action is performed for someone's benefit.
- Example: "The script provisioned the developer a test environment." (Agent provisioned Beneficiary Theme)
Causative-Inchoative Alternation
A verb class where the transitive form (causative) and intransitive form (inchoative) share the same lexical root. The object of the transitive verb becomes the subject of the intransitive verb.
- Transitive (Causative): An Agent causes a change of state in a Patient.
- Example: "The heat melted the solder." (Agent melts Patient)
- Intransitive (Inchoative): The Patient undergoes the change of state spontaneously.
- Example: "The solder melted." (Patient melted)
- Other examples: break, open, freeze, reboot, sync.
Prepositional Object Verbs
These verbs require a prepositional phrase as a core argument, not an optional adjunct. The preposition is specified by the verb's semantic frame.
- Locative Prepositions: The argument specifies a location or path.
- Example: "The pointer points to the memory address." (Theme points to Goal)
- Oblique Arguments: The preposition is lexically determined by the verb.
- Example: "The outcome depends on the initial seed." (Theme depends on Stimulus)
- Example: "The engineer referred to the specification." (Agent referred to Theme)
Sentential Complement Verbs
These verbs take a full clause (finite or non-finite) as a core argument, typically the Theme or Proposition.
- Finite Clause (that-clause): The argument is a tensed proposition.
- Example: "The log indicated [that the connection timed out]." (Stimulus indicates Proposition)
- Non-finite Clause (to-infinitive): The argument is an untensed action.
- Example: "The scheduler attempted [to allocate more memory]." (Agent attempts Action)
- Interrogative Clause (wh-clause): The argument is an embedded question.
- Example: "The query determines [whether the index exists]." (Agent determines Question)

About the author
Prasad Kumkar
CEO & MD, Inference Systems
Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.
His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.
Partnered with leading AI, data, and software stack.
How We Work
Custom AI workflows for your Business
One-fit-all AI don't work for modern businesses. At Inferensys, we aim to understand your business & custom requirements; which we use to define most efficient agentic workflows, the data, and the tools for your business.
01
Review the use case
We understand the task, the users, and where AI can actually help.
Read more02
Pick the right approach
We define what needs search, automation, or product integration.
Read more03
Build the first useful version
We implement the part that proves the value first.
Read more04
Improve from there
We add the checks and visibility needed to keep it useful.
Read moreThe first call is a practical review of your use case and the right next step.
Talk to Us