5. Frame semantics and its application in DH

1. Frame semantics

Frame Semantics is a significant area within cognitive semantics, focusing on how language is connected to human experience and understanding. It suggests that understanding a concept requires understanding the system of concepts it is part of, meaning that introducing one concept makes the entire system available. This approach views words as categories of experience and aims to uncover the reasons a speech community has for categorizing experiences in particular ways. For example, Petruck (1996) outlines Frame Semantics as a research program that emphasizes the continuity between language and experience, aiming to understand words by exploring the categories of experience they represent and the reasons behind these categorizations (Petruck, 1996).

2. Definition of frame

A frame in semantic studies is a conceptual structure that organizes experience, predicts linguistic arguments, and is grounded in bodily, social, and cultural contexts, aiding in understanding lexical meanings and their applications in various fields such as cognitive science and financial analysis.

Let us study several definitions:

1) Semantic frames in semantic studies refer to how lexical fields are conceptually and pragmatically framed by or grounded in our bodily, social, and cultural experiences and practices.

Nerlich, B., & Clarke, D. (2000). Semantic fields and frames: Historical explorations of the interface between language, action, and cognition. Journal of Pragmatics, 32, 125-150. https://doi.org/10.1016/S0378-2166(99)00042-9.

2) Frames in semantic studies are embedded in wider narratives and react in systematic ways to the discursive environment, with a differentiated center-periphery structure.

Baden, C. (2010). Contextualizing frames in political discourse: using semantic network analysis to investigate political parties' framing strategies in the Dutch EU referendum campaign. Circulation Research.

3) A system of frames helps explain why seemingly unrelated expressions in different languages find a common denominator in higher-order frames, thus becoming semantic-pragmatic equivalents.

Pluwak, A. (2021). The frame system as an interlingual representation for parallel texts. Intercultural Pragmatics, 18, 657 - 685. https://doi.org/10.1515/ip-2021-5004.

4) A frame in semantic studies is a system of concepts related in such a way that understanding one concept requires understanding the entire system.

Petruck, M. (1996). FRAME SEMANTICS. https://doi.org/10.1075/hop.2.fra1.

3. Key concepts of the original theory

Frame semantics is a linguistic theory developed by Charles J. Fillmore in the late 1970s, building upon earlier work in semantics and cognitive linguistics. It focuses on how meaning is organized and structured in human language, particularly in terms of how individual words or phrases evoke mental "frames" or cognitive structures that help people understand and interpret language. Some key concepts of frame semantics include:

Frames: Frames are cognitive structures or mental representations that organize our understanding of the world. They consist of a set of interconnected concepts, roles, and relations that define a particular scenario or situation. Frames are often activated in language comprehension and production to help understand the meaning of words and sentences.

Frame Elements (Fillmorean Case Roles): Frame elements are the roles or participants associated with a particular frame. They represent the different semantic roles that are filled by the entities mentioned in a sentence. For example, in the frame of "eating," the frame elements might include the eater, the food, the instrument used for eating:

Lexical Units: Lexical units are words or phrases that evoke specific frames. Each word or phrase is associated with one or more frames, and understanding the meaning of the word involves activating the relevant frame(s). For example, the word "car" evokes the frame of transportation, which includes concepts like vehicles, roads, drivers, etc.

At the same time, there was another influential theory of Ch. J. Fillmore to contribute to the frame semantics. It is a theory of Construction Grammar.

Construction Grammar, a theory pioneered by Charles J. Fillmore along with Paul Kay and other linguists, offers a framework for understanding the structure and organization of language. Here's an overview of Construction Grammar as developed by Fillmore:

Constructional Approach: Construction Grammar posits that language is organized around constructions, which are form-meaning pairings that range from individual words to larger syntactic structures. Constructions capture regularities in language use, including both grammatical and semantic patterns.

Syntactic Constructions: In Construction Grammar, syntactic constructions are seen as pairings of form (surface structure) and function (meaning). These constructions can range from simple templates for word order to more complex patterns involving phrasal or clausal structures.

Semantic Constructions: Constructions also encode semantic information, capturing regularities in meaning beyond mere syntax. This allows Construction Grammar to account for how linguistic expressions convey not just grammatical relations but also nuanced semantic concepts.

Flexible Productivity: Unlike earlier generative grammatical frameworks, Construction Grammar allows for greater flexibility and productivity in language use. Constructions can be combined, modified, and extended creatively to convey new meanings and nuances.

Constructional Networks: Construction Grammar represents knowledge of language as a network of interconnected constructions. These networks capture the relationships between constructions, including inheritance (shared properties between related constructions) and constructional schemas (abstract templates that underlie multiple specific constructions).

Constructional Polysemy: Construction Grammar acknowledges that constructions can have multiple related meanings or functions, a phenomenon known as constructional polysemy. This allows for a more nuanced understanding of how constructions are used in different contexts.

Let us look at the example:

A lexical unit (LU) is a pairing of a word with a meaning.

A semantic frame is a script-like conceptual structure that describes a particular type of situation, object, or event along with its participants and props.

For example, the Apply _heat frame describes a common situation involving a Cook, some Food, and a Heating Instrument, and is evoked by words such as bake, blanch, boil, broil, brown, simmer, steam, etc. These are roles frame elements (FEs) and the frame-evoking words are LUs in the Apply _heat frame.

In the simplest case, the frame-evoking LU is a verb and the FEs are its syntactic dependents:

Case Grammar: In Fillmore's theory of Case Grammar, cases refer to the semantic roles that elements of a sentence play. Fillmore identified a set of basic semantic roles, often referred to as Fillmorean Case Roles. These roles capture the relationships between the elements of a sentence and the action or state described by the verb. They include:

Agent: The agent is the entity that performs or initiates the action described by the verb. It is typically animate and volitional. For example, in the sentence "The boy kicked the ball," "the boy" is the agent.

"The chef prepared the meal."

"The dog chased the cat."

"The teacher explained the lesson."

Patient: The patient is the entity that undergoes or is affected by the action described by the verb. It is typically affected by the action but does not necessarily initiate it. In the sentence "The boy kicked the ball," "the ball" is the patient.

"The storm damaged the roof."

"The virus infected many people."

"The earthquake destroyed several buildings."

Instrument: The instrument is the entity used by the agent to perform the action described by the verb. It is the means by which the action is carried out. For example, in the sentence "The boy kicked the ball with his foot," "his foot" is the instrument.

"She cut the paper with scissors."

"He wrote the letter with a pen."

"They built the house with bricks and mortar."

Experiencer: The experiencer is the entity that perceives or experiences the state or event described by the verb. It is typically sentient and capable of experiencing sensations or emotions. For example, in the sentence "Jane loves chocolate," "Jane" is the experiencer.

"He enjoys watching movies."

"She fears spiders."

"They hated the taste of the medicine."

Theme: The theme is the entity that undergoes movement or change as a result of the action described by the verb. It often corresponds to the patient but can also refer to the entity that undergoes a change of state or location. For example, in the sentence "She put the book on the table," "the book" is the theme.

"She read the book to the children."

"He ate the apple."

"They carried the bags."

Location: The location is the entity that specifies the place or spatial context in which the action described by the verb occurs. It provides information about where the action takes place.

"She lives in a small apartment."

"The party will be held at the park."

"He found his keys under the bed."

Goal/Recipient: The goal or recipient is the entity that receives the action or the result of the action described by the verb. It often involves a transfer of possession, information, or affection.

"She gave the book to her friend."

"He sent an email to his colleague."

"They donated money to the charity."

Source: The source is the entity from which the action originates or the point of departure for movement.

"He came from Canada."

"She received a gift from her grandmother."

"They fled from the burning building."

Beneficiary: The beneficiary is the entity for whose benefit the action is performed.

"He bought flowers for his wife."

"She made dinner for her family."

"They organized a fundraiser for the victims."

Attribute: The attribute case represents properties or qualities attributed to a noun.

"She is tall and athletic."

"He is intelligent and creative."

"The book is interesting and informative."

Time: The time case specifies the temporal context in which the action or event described by the verb occurs.

"She arrived at noon."

"He will finish the project by Friday."

"They met last summer."

Frame-to-Frame Relations: Frames are not isolated; they are interconnected through various relations. These relations capture the relationships between different frames, such as inheritance (one frame inherits properties from another), elaboration (one frame adds detail to another), and opposition (one frame contrasts with another).

Frame Evocation and Extension: Frame semantics recognizes that words can evoke frames directly but can also extend or modify existing frames. For example, the word "self-driving car" extends the frame of "car" by adding the concept of autonomy, leading to a modified understanding of the vehicle.

4. FrameNet project

The FrameNet project is a comprehensive online linguistic database aimed at documenting the range of semantic and syntactic frames within English. Initiated at the International Computer Science Institute in Berkeley, California, under the direction of Charles J. Fillmore and colleagues in 1997, FrameNet stands as a pivotal resource in the field of computational linguistics and natural language processing (NLP). Its development was motivated by principles of Frame Semantics, a theory that views meaning in terms of how words and phrases evoke certain types of situations, objects, or events, which are called "frames".

Key Features of FrameNet:

Lexical Database: FrameNet is built around the concept of "lexical units" (LUs), which are word senses linked to semantic frames. Each LU is associated with a particular frame that describes a type of event, entity, or relation, providing a rich description of its meaning.

Semantic Frames: At the heart of FrameNet are the semantic frames themselves. A frame is a conceptual structure that describes a particular type of situation, object, or event along with its participants and properties. For example, the "Commerce_buy" frame involves a buyer, a seller, goods, and money.

Frame Elements: Within each frame, participants and other conceptual roles are defined as frame elements (FEs). These elements capture the various roles that can be filled within the scenario described by the frame.

Annotations: FrameNet includes annotated examples of how LUs are used in actual sentences, illustrating the instantiation of frames and their elements in natural language. This corpus of annotated sentences is a valuable resource for training and testing NLP systems.

Applications: The structured semantic information in FrameNet has been applied in various areas of computational linguistics and NLP, including semantic role labeling, information extraction, machine translation, and question answering systems. It provides a foundational resource for understanding the meaning of text at a deeper level than simple word-based approaches.

Development and Impact:

The ongoing development of FrameNet involves the continuous addition of new frames, lexical units, and annotated examples. It employs a rigorous methodology for frame definition and annotation, involving both automated tools and manual review by linguists.

FrameNet's impact extends beyond computational applications; it also contributes to linguistic research and education by providing insights into the organization of lexical and conceptual knowledge in English. Its approach has inspired similar projects for other languages, contributing to a growing global network of frame-based linguistic resources.

Conclusion:

The FrameNet project represents a significant endeavor in the field of linguistics and NLP, offering a structured way to understand the semantics of language. By linking linguistic forms to underlying conceptual structures, FrameNet facilitates a deeper understanding of language meaning and supports advanced computational tasks that require semantic comprehension. Its continued development and application promise ongoing contributions to both theoretical linguistics and practical NLP solutions.