模組組裝與模型整合：以現代免疫學的發現為案例

Assembling Modules and Integrating Models: A Case of Discovery from Modern Immunology

本文以模組組裝和模型整合觀點發展現代免疫學的發現模式。免疫學家為說明發炎反應現象，通常會從背景知識擷取可用的機制模組，將其與推測有因果關聯的機制成份拼裝出一個新的機制模型，然後用之設計干預實驗，產生多個資料模組進而組成資料模型，作為新發現的證據。過程中，免疫學家須在機制 、實驗與資料三種模型內建立模組間的因果連續性，並且使三種模型的結果達成一致。本文爭論，免疫學家是透過建立模型內因果連續性與整合異質模型間的一致性來層層構成發現證成。由於此觀點是將先前哲學重視以機制假說或實驗現象為主的研究對象拆解為單位更小的模組，故打破傳統以假說或現象為主的二分。本文以一篇當代免疫學論文為分析案例，為新觀點提供證據。

This paper proposes a module-based analysis of the discovery process in the current practice of immunologists, focusing on their published results. In their explanations of inflammatory response phenomena, immunologists usually start by scouring the available background knowledge on mechanistic modules and reassembling a select module(s) and its causally relevant components to construct a new mechanistic model. Then they employ their model to design interventional experiments, and, lastly, articulate the data modules with data models to illustrate the putative evidence. In the process, scientists not only establish the causal continuity of modules within mechanistic, experimental, and data models but achieve consistency among them. In this paper, I argue that in their research articles immunologists’ discovery pattern consists of forming layers of justification by assembling modules into modes and integrating heterogenous models. The module-based analysis of discovery patterns disrupts the traditional dichotomy between hypothesis-driven and phenomena-driven analyses. Compared to previous philosophical accounts, it better characterizes the generation of discoveries, as evidenced in a detailed analysis of the presentation of research results in an immunology article.