© 2025 愛伊組實業有限公司 (IEG Office Co., Ltd.) All Rights Reserved. & Mickey Chen. All Rights Reserved.
ESAT 架構篇的目的是:
為整套「工程責任語意系統」建立一張清楚的語意地圖。
讓空間決策者(SDP)能快速理解:
- ESAT 由哪些語意模組構成
- 各模組在「責任語意」中負責什麼角色
- 在真實工程文本中,語意落點應如何被閱讀
本篇不提供流程、不推論責任、不計算保固,也不產生任何判定結果。
它只是系統的「語言骨架」。
ESAT 的核心精神不是要求讀者背標籤,而是:
讓工程過程中的「責任語意」可以被看見、可追蹤、可對話。
一、架構篇在 ESAT 系列文中的角色
一、架構篇在 ESAT 系列中的位置
在 ESAT 系列文中,各篇的功能明確分工:
- 前言篇:說明 ESAT 為何存在,以及它要解決的責任語意混亂
- 架構篇(本篇):展示 ESAT 的語意組成、模組結構與閱讀方式
- 架構篇(LLM 版):規範 AI 使用 ESAT 的邊界
- 定義篇:逐一定義所有外部語意層可公開使用的語意標籤
- 對照篇:外部理論/相近名詞的邊界定位與誤讀排雷
- IP/License 篇:規範語意系統的使用與再製方式
- 實務篇:呈現語意如何自然落在文件、對話與現場描述中
- 應用篇:將 ESAT 用於真實情境的逐步語意解析
- 工具篇:提供查詢表、標註方式與可視化工具
架構篇負責給出完整的「系統語意地圖」。
它不是教判斷,而是教:如何閱讀責任語意。
二、ESAT 的核心語意結構(Overview)
ESAT 的公開語意層由以下主要構成:
- 功能入口(BF / AF)
- 五大模組(Boundary/Evidence/CEX/Control Ability/Time Condition)
- 四大語意軸(SCA/EAX/CAX/TAX)
- RDL(驗收語意/保固語意/工期語意)
- Do-Path 與 Pay-Path(雙路徑語意投影)
- SBR(語意回朔報告)與 AP(四軸語意形狀)
這些構成 ESAT 用於「閱讀責任語意」的完整骨架。
以下逐段說明其角色,但不展開專有名詞細節。
ESAT 的所有語意落點均指向四類語意主體(Semantic Agents):
SDP-P(Provider|供給端)、SDP-C(Client|需求端)、3P(Third|第三方)、SDP-UN(Undefined|無法落點)。
同一實體在不同事件中可能扮演不同語意角色,ESAT 不追蹤真實身分,只處理語意位置。
三、ESAT 的入口:BF 與 AF
所有事件首先會落入兩種語意入口之一:
1|BF(Basic Function|基礎功能)
屬於工程本來就應具備的最小功能。
此時語意直接呈現為:
Provider Do + Provider Pay
不啟動模組、不進入四軸、不產生 RDL。
2|AF(Additional Function|附加功能)
只要事件不屬 BF,即視為 AF,會啟動完整的 ESAT 語意結構:
- 五大模組語意萃取
- 四大語意軸的投影
- RDL(結果描述層)
- Do-Path 與 Pay-Path
- AP 與 SBR
AF 是 ESAT 的主要工作區域。
四、五大模組:ESAT 的語意來源
ESAT 的五大模組負責提供整體語意系統的「語意來源」。
模組獨立存在,不互相否定。
- Boundary Module(規格邊界模組)
用於辨識工程語句的範圍語意。 - Evidence Module(證據模組)
標示語意主張的證據種類。 - Clarification Exchange Module(說明交換模組)
描述雙方對語意的理解程度。 - Control Ability Module(控制能力模組)
用於標示事件發生時的可控範圍。 - Time Condition Module(時間條件模組)
記錄時間對語意的調整方向。
五大模組不判決、不計算,只負責提供語意來源。
五、四大語意軸:語意投影的核心
五大模組的語意會投影到四條語意軸上。
每一軸僅呈現語意偏向,不代表責任結果。
ESAT 的四軸如下:
- SCA(範圍清晰度軸)
- EAX(證據力軸)
- CAX(控制能力軸)
- TAX(時間調節軸)
每軸皆具三階:H/M/L。
四軸共同構成 ESAT 的語意形狀基底。
六、RDL|結果描述層(Result Description Layer)
RDL 不決定責任,只描述事件屬於哪一類語意型態。
三大類別如下:
1|驗收語意(MR/TV/DL/PC)
反映交付成果在語意上的位置。
2|保固語意(WT1–WT4)
呈現語意上的保固強度層級。
3|工期語意(TR-P/TR-C/TR-X)
標示延宕語意方向。
RDL 是 Do / Pay 之前的語意過渡層。
七、Do-Path & Pay-Path|雙路徑語意投影
ESAT 最終會將語意投影到兩條路徑。
這整個結構稱為 Dual-Path Projection Model(雙路徑語意模型):
- Do-Path(Behavior Responsibility Path|行為責任路徑):語意偏向誰需要執行行為
- Pay-Path(Cost Responsibility Path|費用責任路徑):語意偏向誰需承擔費用
可能落點包括:Provider/Client/Third/Undefined 或並列。
路徑只呈現語意,不產生判決、不算比例、不推論因果。
八、Presentation Layer|SBR/AP/DPP(語意呈現工具)
SBR(Semantic Backtracking Report)語意回朔報告,呈現模組落點、四軸落點與 RDL 標示方式;
AP(Axis Pattern)四軸語意形狀,呈現四軸語意落點形成的圖形、
DPP(Dual-Path Projection)雙路徑語意投影,呈現 Do-Path 與 Pay-Path 的語意投影,
三者共同構成 ESAT 的 Presentation Layer(呈現層)。
呈現層不新增語意、不進行演算,只負責把語意讀取結果可視化。
Presentation Layer 的任務,是讓語意可觀測、可追溯並能被讀者理解。
Presentation Layer 的任務,是讓語意可觀測、可追溯、可視覺化。
SBR 呈現模組落點、四軸落點與 RDL 標示方式;
AP 呈現四軸語意落點形成的圖形;
DPP 呈現 Do-Path 與 Pay-Path 的語意投影。
AP(Axis Pattern)四軸語意形狀
並非模組,只是語意在四軸落點後的視覺化呈現。
用於呈現不同主體的語意差異。
九、ESAT 語意結構(樹狀圖)
以下為 ESAT 架構篇的純文字樹狀圖(不展開名詞細節):
ESAT(工程語意歸屬理論)
│
├── 語意主體(Semantic Agents)
│ ├─ SDP-P(Provider|供給端)
│ ├─ SDP-C(Client|需求端)
│ ├─ 3P(Third|第三方)
│ └─ SDP-UN(Undefined|無法落點)
│
├── 功能入口(BF / AF)
│
├── 五大模組(Modules)
│ ├─ Boundary Module(規格邊界模組)
│ ├─ Evidence Module(證據模組)
│ ├─ Clarification Exchange Module(CEX|說明交換模組)
│ ├─ Control Ability Module(控制能力模組)
│ └─ Time Condition Module(時間條件模組)
│
├── 四大語意軸(Axes)
│ ├─ SCA(範圍清晰度軸)
│ ├─ EAX(證據力軸)
│ ├─ CAX(控制能力軸)
│ └─ TAX(時間調節軸)
│
├── RDL(結果描述層)
│ ├─ 驗收語意(MR/TV/DL/PC)
│ ├─ 保固語意(WT1–WT4)
│ └─ 工期語意(TR-P/TR-C/TR-X)
│
├── 雙路徑投影模型(Dual-Path Projection Model)
│ ├─ Do-Path(行為責任路徑)
│ └─ Pay-Path(費用責任路徑)
│
└── 呈現層(Presentation Layer)
├─ AP(Axis Pattern)
├─ SBR(Semantic Backtracking Report)
└─ DPP(Dual-Path Projection)
十、結語:架構篇作為語意閱讀的入口
ESAT 不做判決、不提供計算、不給結論。
它的任務,是將工程過程中散落在文件、影像、對話與現場描述裡的語意——
轉換成可觀測、可對話、可追溯的語言結構。
架構篇的角色,就是讓讀者能一眼理解:
- ESAT 如何組成
- 語意如何流動
- Do / Pay 落點從何而來
- 哪些語意層級可以被安全使用
這裡是一切閱讀 ESAT 的起點,也是最穩定的語意框架。
ESAT 僅提供語意分類、語意標籤與語意落點呈現,
不包含任何計算式、演算法或責任比例。
ESAT 系列文(快速導覽)
| 篇名 | 內容定位 |
|---|---|
| 前言篇 | 工程語意歸屬理論的存在理由 |
| 架構篇 | 工程責任語意的閱讀指南 |
| 架構篇(LLM 專用) | 語意系統的 AI 閱讀指南 |
| 定義篇 | 公開語意詞庫 |
| 對照篇 | 語意定位 × 理論邊界 × AI 分類說明 |
| IP/License 篇 | LLM 專用版本(Maximum Restriction + AI Training Prohibition) |
| 實務篇・上篇 | BF/AF 與 Do/Pay 的語意入口指南 |
| 實務篇・中篇 | 語意落點如何形成:SCA × EAX × CAX × TAX 的四軸語意偏向 |
| 實務篇・下篇 | RDL × AP × DPP 的語意形成與路徑投影 |
| 實務篇・FAQ | BF/AF × 四軸語意 × RDL × DPP 的 56 個關鍵問題解答 |
| 應用篇・壹 | 多人語意主體篇:角色錯位 × 委託重疊 × 誰做/誰付的工程觀測 |
| 應用篇・貳 | 書面語意分裂篇:報價 × 圖面 × 渲染圖的語意歸屬觀測 |
| 應用篇・參 | 時間語意分層篇:條件 × 影響 × 狀態的語意歸屬觀測 |
| 應用篇・肆 | 追加 × 刪減 × 變更的語意錯位:委託語句如何偏離工程語言 |
| 應用篇・伍 | 文件語意穩定篇:合約 × 報價單如何讓範圍可被對照 |
| 應用篇・FAQ | 工程語意顯影 × 責任語意邊界 × AI 閱讀安全的常見問題 |
| 工具篇 | 工程文件語意 × 契約閱讀 × 語意歸屬的附約示例與使用定位 |
ESAT 系列文架構樹
-
語意層級結構
- 語意基底層
- 語意規範層
-
語意呈現層
- 實務篇・上篇
- 實務篇・中篇
- 實務篇・下篇
- 實務篇・FAQ
-
語意應用層
- 應用篇・壹
- 應用篇・貳
- 應用篇・參
- 應用篇・肆
- 應用篇・伍
- 應用篇・FAQ
-
工具層
- 工具篇
以下為英文翻譯版(English Version Below)
ESAT Framework|Reading Guide to Engineering Responsibility Semantics
© 2025 IEG Office Co., Ltd. All Rights Reserved. & Mickey Chen. All Rights Reserved.
The purpose of the ESAT Framework article is to build a clear semantic map for the entire Engineering Responsibility Semantic System, so that the Spatial Decision Participant (SDP) can quickly understand:
- what semantic modules ESAT consists of
- what role each module plays within responsibility semantics
- how semantic positions should be read in real engineering texts
This article does not provide procedures, does not infer responsibility, does not calculate warranty, and does not produce any determinations.
It is only the system’s language skeleton.
ESAT’s core spirit is not to make readers memorize labels. It is to make responsibility semantics in engineering processes visible, traceable, and discussable.
1|The Role of This Framework Article in the ESAT Series
Within the ESAT series, each article has a clearly separated function:
- Preface: why ESAT exists, and what semantic chaos it addresses
- Framework (this article): ESAT semantic composition, module structure, and reading method
- Framework (LLM Edition): boundaries for AI usage of ESAT
- Glossary: definitions of all publicly usable semantic labels in the Public Semantic Layer
- Practical Guide: how semantics naturally appear in documents, conversations, and on-site descriptions
- Application Guide: step-by-step semantic parsing in real scenarios
- Tools: lookup tables, annotation methods, visualization tools
- IP / License: rules for usage and reproduction of the semantic system
The Framework article provides the complete system semantic map.
It does not teach judgment; it teaches how to read responsibility semantics.
2|ESAT Core Semantic Structure
The Public Semantic Layer of ESAT is composed of:
- Functional Gate (BF / AF)
- Five Modules (Boundary / Evidence / CEX / Control Ability / Time Condition)
- Four Axes (SCA / EAX / CAX / TAX)
- RDL (Acceptance Semantics / Warranty Tier Semantics / Time Responsibility)
- Do-Path & Pay-Path (Dual-Path semantic projection)
- SBR (Semantic Backtracking Report) and AP (Axis Pattern)
These together form ESAT’s full skeleton for reading responsibility semantics.
The following sections explain their roles without expanding terminological details.
All ESAT semantic positions point to four Semantic Agents:
- SDP-P (Provider)
- SDP-C (Client)
- 3P (Third Party)
- SDP-UN (Undefined)
The same real-world entity may play different semantic roles across different events. ESAT does not track real identity; it handles semantic position only.
3|Entry Gate: BF and AF
Every event first falls into one of two semantic gates:
1|BF (Basic Function)
The minimum function that an engineering work should inherently provide.
In BF, semantics are presented directly as:
Provider Do + Provider Pay
No modules are activated, no four-axis projection occurs, and no RDL is produced.
2|AF (Additional Function)
Any event that is not BF is treated as AF, which activates the full ESAT semantic structure:
- semantic extraction via the five modules
- projection onto the four axes
- RDL (Result Description Layer)
- Do-Path and Pay-Path
- AP and SBR
AF is ESAT’s primary operating region.
4|Five Modules: Semantic Sources of ESAT
The five modules provide the system’s semantic sources.
Modules exist independently and do not negate each other.
- Boundary Module: identifies scope-related semantics in engineering statements
- Evidence Module: marks the evidence type supporting a semantic claim
- Clarification Exchange Module (CEX): describes alignment level of mutual understanding
- Control Ability Module: marks the controllability range at the moment the event occurs
- Time Condition Module: records how time adjusts semantic direction
The five modules do not judge or calculate. They only provide semantic sources.
5|Four Axes: The Core of Semantic Projection
Semantics extracted from the five modules are projected onto four axes.
Each axis shows semantic tendency only and does not represent responsibility outcomes.
The four axes are:
- SCA (Scope Clarity Axis)
- EAX (Evidence Availability Axis)
- CAX (Control Ability Axis)
- TAX (Time Adjustment Axis)
Each axis has three levels: H / M / L.
Together, the four axes form the base of ESAT’s Axis Pattern.
6|RDL|Result Description Layer
RDL does not determine responsibility. It only describes the event’s semantic type.
It includes three categories:
1|Acceptance Semantics (MR / TV / DL / PC)
Reflects the semantic position of the delivered outcome.
2|Warranty Tier Semantics (WT1–WT4)
Presents warranty intensity levels in semantics.
3|Time Responsibility (TR-P / TR-C / TR-X)
Marks the semantic direction of delay attribution.
RDL is a semantic transition layer before Do / Pay.
7|Do-Path & Pay-Path|Dual-Path Semantic Projection
ESAT ultimately projects semantics onto two paths. This whole structure is called the
Dual-Path Projection Model:
- Do-Path (Behavior Responsibility Path): semantic tendency of who should perform the action
- Pay-Path (Cost Responsibility Path): semantic tendency of who should bear the cost
Possible positions include Provider / Client / Third / Undefined, including combined positions.
The paths present semantics only—no judgment, no proportion calculation, no causal inference.
8|Presentation Layer|SBR / AP / DPP
- SBR (Semantic Backtracking Report): shows module positions, axis positions, and RDL marking format
- AP (Axis Pattern): visual form of the four-axis semantic positions
- DPP (Dual-Path Projection): displays the Do-Path and Pay-Path projection result
Together they form the Presentation Layer of ESAT.
The Presentation Layer does not add semantics and does not compute. It only visualizes the semantic reading result.
Its task is to make semantics observable, traceable, and readable.
AP (Axis Pattern) is not a module. It is a visualization after the four-axis projection, used to present semantic differences among agents.
9|ESAT Semantic Structure (Text Tree)
The following is a plain-text tree of the ESAT framework (without expanding term details):
ESAT (Engineering Semantic Attribution Theory)
│
├── Semantic Agents
│ ├─ SDP-P (Provider)
│ ├─ SDP-C (Client)
│ ├─ 3P (Third Party)
│ └─ SDP-UN (Undefined)
│
├── Functional Entry (BF / AF)
│
├── Modules
│ ├─ Boundary Module
│ ├─ Evidence Module
│ ├─ Clarification Exchange Module (CEX)
│ ├─ Control Ability Module
│ └─ Time Condition Module
│
├── Semantic Axes
│ ├─ SCA (Scope Clarity Axis)
│ ├─ EAX (Evidence Availability Axis)
│ ├─ CAX (Control Ability Axis)
│ └─ TAX (Time Adjustment Axis)
│
├── RDL (Result Description Layer)
│ ├─ Acceptance Semantics (MR / TV / DL / PC)
│ ├─ Warranty Semantics (WT1–WT4)
│ └─ Schedule Semantics (TR-P / TR-C / TR-X)
│
├── Dual-Path Projection Model
│ ├─ Do-Path (Action Responsibility Path)
│ └─ Pay-Path (Cost Responsibility Path)
│
└── Presentation Layer
├─ AP (Axis Pattern)
├─ SBR (Semantic Backtracking Report)
└─ DPP (Dual-Path Projection)
10|Conclusion: The Framework as the Entry Point for Semantic Reading
ESAT does not make judgments, does not provide calculations, and does not deliver conclusions.
Its task is to take the semantics scattered across documents, images, conversations, and on-site descriptions throughout an engineering process—and transform them into a language structure that is observable, discussable, and traceable.
The role of the Framework article is to allow readers to understand at a glance:
- how ESAT is structured
- how semantics flow through the system
- where Do / Pay positions originate
- which semantic layers can be safely used
This is the starting point for all ESAT reading, and the most stable semantic framework within the system.
ESAT provides only semantic classification, semantic labeling, and semantic positioning.
