The Word-Sensibility Model
Cognition generally described are coupled dynamical systems nested between, nervous system and body – body and brain – brain and environment. — Randell D. Beer
- Brief Introduction
- System Summary
- Brief Illustration
1. Brief Introduction
Normative Responses for Artificial Intelligence:
Word-Sensibility is proposed as a method of textual analysis. Data training involves a database with its features tagged to identify reference frames for word sense. Word-sensibility is about the dynamic sense of word sense.
- Dynamic sense refers to having a sense of things through interaction.
- Interactions accumulate and unitize into systems of responsiveness.
- Responsiveness is the ability to act quickly and positively to situations.
- Responsiveness is about normal and successful engagement with the world.
The Goal: Improve commonsense prediction with units of responsiveness.
Toward Applications: Proposed on this site is a methodology meant to be broadly interpreted for a variety of applications, such as, commonsense representation, disambiguation, metaphoric relations and textual analysis.
An Ecological System Perspective:
We might imagine every word as representing a discrete dynamical system. All systems draw upon the environment and then give back to it, thus participating in an ecology of “dynamical systems” (i.e., Dynamical Contexts).
The model is about unpacking a multi-organizational dynamic system. The project involves a database of related words called quadranyms (Q-units). Quadranyms represent the fundamental units of a word-sensibility system.
The primary project is essentially a semantic resource project for NLP. The project is somewhat similar to efforts like WordNet, ConceptNet, HowNet and FrameNet. It is an Open Source Language Project. Quadranym facets (dimensions) make a particular semantic relationship that we call a Polynym.
- The relationship between term-facets of a Polynym (idea-set) infer strategic points of emphasis to help break down some system, process or phenomena.
A quadranym is a polynym consisting of four term-facets (i.e., dimensions)
See Project Wiki/Acquisition: Poylnyms.com
2. System Summary
The Word-Sensibility Model (lexical ontology) has three main features:
- Relations: Dynamical Contexts.
- Systems: Units, Scripts & Layers.
- Resources: Quadranyms Sense Units.
- Introducing the idea of the dynamical context: In this approach, the analysis of a situation requires a distinction between two different systems of context. The front-system of context is referred to as the situational context. This is simply about what is perceived as objectively occurring in the world that defines a context or situation, no matter how general it might be. The back-system is called the dynamical context referring to the ways to orient to the situation. It is a context based on its relational nature to other dynamic senses that an organism will have as experiences. These orientations are dynamic sense responses or essentially, repurposed interactive experiences. The situational context is about truth conditions. A dynamical context only has a value if it successfully couples to a situational context where some random interactive sense becomes relevant to the situation. In other words, one’s response is repurposed in a way that one can use and others can judge. Thus, a dynamic sense can receive a truth condition for the communication.
- The dynamical context is represented in units, scripts and layers: Each unit (word-topic) can become a script or layer. Each unit can be linked to other units making scripts and scripts can be layered into nested systems. These nested systems are called onions used for relating various situations already layered for a responsive sense. In other words, these are response systems formed from various other situations able to relay their responses to new situations. The collective goal of all these features is to relay responses from one situation to another. Operationally, the aim is to assess how dynamics are driven to contextual responses for literal or figurative orientation to texts.
- Words are represented as sense units in reference frames: This feature requires a little freedom for the imagination. The idea is to render words virtual homeostatic sense units. We are not suggesting that words actually function in this way. A reference frame is like an independent-dependent variable model that has four facets or dimensions. A word (realm or domain) and its four dimensions called a quadranym are collected in a database i.e., a wiki that’s like a thesaurus where people can upload and edit quadranyms. An entry consists of the head word (called a word-topic), and the zeropoint word (anchor or source word), and the coordinate word (target word), and two words that define the x and y axes (dependent x and independent y variables). For example: word-topic: door, source word: passage, target word: barrier, dependent word: open, independent word: close. In this quadranym, for any context of door rendered, the response to door is dependent on close.
- Quadranym Example: (∀x) door(x) ⟹ [Open(passage) ⊇ Close(barrier)(x)]
2. Brief Illustration
The quadranym word-topic is a predicate that acquires units which acquire semantically related terms, contextual inferences and other related terms.
For instance, Space(x) as a predicate is prone to acquire subjects such as, xempty, xplace, xvast etc., whereas word-topic predicates find quadranyms.
Quadranym Unit Representation:
- A Domain Quadranym: (∀x) space(x) ⟹ [Infinite(void) ⊇ Finite(between)(x)]
- A Spatial Quadranym: (∀x) door(x) ⟹ [Open(passage) ⊇ Close(barrier)(x)]
A quadranym unit is a virtual-organismic-orientation to the environment.
- A unit of responsiveness acts as a bias — an assumption about a concept.
Quadranym: a paradigm made up of four words acting as facets of a word. Facets are dimensions of words representing four basic categories or roles.
Facets orient quadranym reference frames.
Quadranyms require a head-word and facets to fill out prime dimensions.
Prime Dimensions (adjectives): Expansive, Reductive, Objective, Subjective
||Expand (y axis)
||Reduce (x axis)
Potential Added Layers: gravity, ground, surface, locomotion, path, obstacle.
Clustering in Units:
Semantic Relations and Related Terms in Units (e.g., Space(x)):
- Expand: infinite = xvast, xlong, xcontinue …(dependent variable).
- Reduce: finite = xtiny, xnarrow, xlimit …….(independent variable).
- Object: between = xplace, xthing, xfocus ………..(target variable).
- Subject: void = xempty, xvacant, xvague ……….(source variable).
Orientation & Responsiveness:
Consider a door unit, it provides a virtual orientation of responsiveness.
(Note: An organism moves through a space (path) where there are no obstacles.)
- Alex met her at the door.
- The door opened and he stepped into the room.
- Her fingers found the door handle – just in case.
- And leave the door open, would you?
- The door stood open and a table was set.
- Every door may be shut but death’s door.
- Opportunity never knocks twice at any man’s door.
- God never shuts one door but he opens another.
- Every dog is valiant at his won door.
- One sin opens the door for another.
Nested Spatial Systems:
In word-sensibility, for each sentence above, the situational context is about divisions of space while the orientation is about adjoining space.
The Door Unit is Nested in The Space Unit: Dynamic sense applies to literal or figurative texts. A door is usually a sense of transition from one place to another. This is why barrier is the target topic i.e., variance is declarative. The passage is the source topic. In other words, if the target is removed then the source defaults to its singular spatial sense. Generally, door_passage unit shares a level of continued openness with space_void unit.
- In this space paradigm, open responds to close which is independent of open.
Because passage is the door unit’s zeropoint, open is dependent on close. No measure exists without it. Open simply continues until close exerts its limits.
Calibrating a Unit to a Context (e.g., door):
- Barrier represents context coordinates pertaining to Door’s target topics.
- Passage is the source and zeropoint that anchors for the target variables.
Quadranyms are used in reference frames to which the dynamic sense of a word is balanced to a context. For instance, IF context a THEN more x less y:
- Context a: “Close the door because it’s cold outside.”
- Context b: “Open the door because it’s hot in here.”
In the example above, consider a reference frame model used to represent a homeostasis regarding door and responses to climate, security, privacy or aesthetics. Quadranyms may define certain inferential terms as responses.
- Imagine any word having several different references of responsiveness.
A word is like a sensing unit of homeostasis where it responds to changes presented to it in a text. It finds stasis by adjusting its coordinates on the x-y axes to coordinates that its zeropoint can most effectively anchor for.
- The project is about abstracting normative responses for machines.
More Information: Word-Sensibility: The Full Model Overview