John Searle’s social ontology distinguishes between linguistic and non-linguistic institutional facts. He argues that every instance of the latter is created by declarative speech acts, while the former are exceptions to this far-reaching claim: linguistic phenomena are autonomous, their meaning is “built in,” and this is necessary, Searle argues, to avoid “infinite regress.” In this essay I analyze Searle’s arguments for this linguistic exceptionalism and reveal its flaws. My method is to follow Searle’s argument closely and comprehensively so as to avoid, insofar as is possible, a selective reading of his argument in my favor. Against Searle’s position, I argue that linguistic phenomena are not exceptions to his general theory of institutional facts, for they too always require supplementary representations in order to exist. Language itself is analogous to all other institutions and infinite regress turns out to be unavoidable.

约翰·塞尔（John Searle）的社会本体论将语言和非语言机构事实区分开来。他认为，后者的每一个实例都是由声明性言语行为造成的，而前者则是这一深远主张的例外：语言现象是自主的，其含义是“内在的”，塞尔认为，这是必要的，以避免“无限回归。” 在本文中，我将分析塞尔对这种语言例外论的论点，并揭示其缺陷。我的方法是密切和全面地遵循塞尔的论点，以便在可能的情况下，避免对他的论点有选择地阅读，以对我有利。反对塞尔的立场，我认为语言现象不是他的制度事实理论的例外，因为它们也总是需要补充性的表述才能存在。