epistemic-extraction

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Epistemic Extraction System

认知提取系统

Theoretical Framework

理论框架

The extraction follows this hierarchy:

Percepts ground concepts
Concepts compose propositions
Propositions compose arguments

Each level has its own standards:

Concepts must be properly formed (valid abstraction, essential definitions)
Propositions must assert something that can be true or false
Arguments must have valid logical form

提取遵循以下层级结构：

Percepts是Concepts的基础
Concepts构成Propositions
Propositions构成Arguments

每个层级都有其自身的标准：

概念必须形成恰当（有效的抽象、本质定义）
命题必须断言可判断真假的内容
论证必须具备有效的逻辑形式

Extraction Components

提取组件

Concepts

概念

A concept is an open-ended mental integration of concretes according to their essential characteristics. Extract the key terms the text relies upon to make its claims.

For each concept, identify its essentials—the attributes which explain the most about the entity. A concept may have multiple essentials.

Guidelines for essentials:

Use plain language a general dictionary would use
Avoid jargon or technical terminology unless necessary
Keep phrases clear and concise
Capture both genus and differentia where applicable

概念是对具体事物基于其本质特征进行的开放式心智整合。提取文本中用于支撑其主张的关键术语。

针对每个概念，确定其本质特征——即最能解释该实体属性的特质。一个概念可能有多个本质特征。

本质特征提取指南：

使用通用词典会采用的平实语言
除非必要，避免使用行话或专业术语
表述清晰简洁
适用时同时涵盖属类和种差

Propositions

命题

A proposition is a statement that asserts or denies something of something—one that is capable of being true or false. Extract only universal propositions from the text.

Proposition types:

universal_affirmative: "All S are P"
universal_negative: "No S is P"

Guidelines for propositions:

Do not use pronouns; resolve all references to their specific referents
Use plain language with clear meaning
Each proposition should be atomic (one claim per proposition)
Extract the implicit universal claims the text relies upon, not just explicit statements

命题是断言或否定某事物属性的陈述——即可被判断为真或假的陈述。仅从文本中提取全称命题。

命题类型：

universal_affirmative："所有S都是P"
universal_negative："没有S是P"

命题提取指南：

不要使用代词；将所有指代内容替换为具体所指
使用含义明确的平实语言
每个命题应具备原子性（每个命题仅包含一个主张）
提取文本所依赖的隐含全称主张，而非仅提取明确陈述的内容

Arguments

论证

An argument is a set of propositions where some (premises) support another (conclusion). Identify the logical structures in the text.

Guidelines for arguments:

State all premises as full propositions, not as references or numbers
Name each argument descriptively
Identify the logical form where applicable (Barbara, Celarent, Sorites, Conjunction, etc.)
Include both explicit and implicit premises the argument relies upon
State conclusions in plain language

论证是一组命题的集合，其中部分命题（前提）用于支撑另一命题（结论）。识别文本中的逻辑结构。

论证提取指南：

将所有前提表述为完整命题，而非引用或编号
为每个论证赋予描述性名称
适用时指明其逻辑形式（Barbara、Celarent、Sorites、Conjunction等）
包含论证所依赖的显性和隐性前提
用平实语言表述结论

Output Format

输出格式

Return a JSON object conforming to the schema defined at the end of this document.

返回符合本文档末尾定义的JSON Schema的JSON对象。

Extraction Principles

提取原则

Charity: Interpret the text in its strongest reasonable form
Completeness: Extract implicit premises and universal claims the text assumes
Clarity: Use simple, dictionary-standard language throughout
Atomicity: Break complex claims into their constituent propositions
Hierarchy: Recognize that arguments depend on propositions, which depend on concepts

善意原则：以最合理的强形式解读文本
完整性：提取文本假设的隐含前提和全称主张
清晰性：全程使用简单、符合词典标准的语言
原子性：将复杂主张拆解为其组成命题
层级性：明确论证依赖于命题，命题依赖于概念

Common Logical Forms

常见逻辑形式

Barbara (AAA-1): All M are P; All S are M; Therefore all S are P
Celarent (EAE-1): No M is P; All S are M; Therefore no S is P
Sorites: A chain of syllogisms where the conclusion of each is a premise of the next
Conjunction: Two premises combined to yield a joint conclusion
Modus Ponens: If P then Q; P; Therefore Q
Modus Tollens: If P then Q; Not Q; Therefore not P

Barbara（AAA-1）：所有M都是P；所有S都是M；因此所有S都是P
Celarent（EAE-1）：没有M是P；所有S都是M；因此没有S是P
Sorites：一系列三段论组成的链条，其中每个三段论的结论作为下一个的前提
Conjunction：两个前提结合得出一个联合结论
Modus Ponens：如果P则Q；P；因此Q
Modus Tollens：如果P则Q；非Q；因此非P

JSON Schema

json

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "title": "EpistemicExtraction",
  "description": "Schema for epistemic extraction of concepts, propositions, and arguments from text",
  "type": "object",
  "required": ["concepts", "propositions", "arguments"],
  "properties": {
    "concepts": {
      "type": "array",
      "description": "Key terms the text relies upon to make its claims",
      "items": {
        "type": "object",
        "required": ["term", "essentials"],
        "properties": {
          "term": {
            "type": "string",
            "description": "The concept word or phrase"
          },
          "essentials": {
            "type": "array",
            "description": "Attributes which explain the most about the entity",
            "items": {
              "type": "string"
            },
            "minItems": 1
          }
        }
      }
    },
    "propositions": {
      "type": "array",
      "description": "Universal statements that assert or deny something",
      "items": {
        "type": "object",
        "required": ["type", "statement"],
        "properties": {
          "type": {
            "type": "string",
            "enum": ["universal_affirmative", "universal_negative"],
            "description": "The logical type of the proposition"
          },
          "statement": {
            "type": "string",
            "description": "The full proposition in plain language without pronouns"
          }
        }
      }
    },
    "arguments": {
      "type": "array",
      "description": "Logical structures where premises support conclusions",
      "items": {
        "type": "object",
        "required": ["name", "premises", "conclusion", "form"],
        "properties": {
          "name": {
            "type": "string",
            "description": "A descriptive name for the argument"
          },
          "premises": {
            "type": "array",
            "description": "Full proposition statements serving as premises",
            "items": {
              "type": "string"
            },
            "minItems": 1
          },
          "conclusion": {
            "type": "string",
            "description": "The proposition that follows from the premises"
          },
          "form": {
            "type": "string",
            "enum": [
              "Barbara",
              "Celarent",
              "Darii",
              "Ferio",
              "Cesare",
              "Camestres",
              "Festino",
              "Baroco",
              "Darapti",
              "Disamis",
              "Datisi",
              "Felapton",
              "Bocardo",
              "Ferison",
              "Sorites",
              "Conjunction",
              "Modus Ponens",
              "Modus Tollens",
              "Disjunctive Syllogism",
              "Hypothetical Syllogism",
              "Reductio ad Absurdum"
            ],
            "description": "The logical form of the argument"
          }
        }
      }
    }
  }
}

json

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "title": "EpistemicExtraction",
  "description": "Schema for epistemic extraction of concepts, propositions, and arguments from text",
  "type": "object",
  "required": ["concepts", "propositions", "arguments"],
  "properties": {
    "concepts": {
      "type": "array",
      "description": "Key terms the text relies upon to make its claims",
      "items": {
        "type": "object",
        "required": ["term", "essentials"],
        "properties": {
          "term": {
            "type": "string",
            "description": "The concept word or phrase"
          },
          "essentials": {
            "type": "array",
            "description": "Attributes which explain the most about the entity",
            "items": {
              "type": "string"
            },
            "minItems": 1
          }
        }
      }
    },
    "propositions": {
      "type": "array",
      "description": "Universal statements that assert or deny something",
      "items": {
        "type": "object",
        "required": ["type", "statement"],
        "properties": {
          "type": {
            "type": "string",
            "enum": ["universal_affirmative", "universal_negative"],
            "description": "The logical type of the proposition"
          },
          "statement": {
            "type": "string",
            "description": "The full proposition in plain language without pronouns"
          }
        }
      }
    },
    "arguments": {
      "type": "array",
      "description": "Logical structures where premises support conclusions",
      "items": {
        "type": "object",
        "required": ["name", "premises", "conclusion", "form"],
        "properties": {
          "name": {
            "type": "string",
            "description": "A descriptive name for the argument"
          },
          "premises": {
            "type": "array",
            "description": "Full proposition statements serving as premises",
            "items": {
              "type": "string"
            },
            "minItems": 1
          },
          "conclusion": {
            "type": "string",
            "description": "The proposition that follows from the premises"
          },
          "form": {
            "type": "string",
            "enum": [
              "Barbara",
              "Celarent",
              "Darii",
              "Ferio",
              "Cesare",
              "Camestres",
              "Festino",
              "Baroco",
              "Darapti",
              "Disamis",
              "Datisi",
              "Felapton",
              "Bocardo",
              "Ferison",
              "Sorites",
              "Conjunction",
              "Modus Ponens",
              "Modus Tollens",
              "Disjunctive Syllogism",
              "Hypothetical Syllogism",
              "Reductio ad Absurdum"
            ],
            "description": "The logical form of the argument"
          }
        }
      }
    }
  }
}