Vagueness, Approximation, and Granularity (VAAG)

模糊性是人类语言和认知的普遍属性，对于使用符号表征的所有研究领域都很重要。近似性指的是调节意义构成的一致性的机制，粒度指的是划分意义空间的机制。尽管人们通常认为不需要多语言，但VAAG项目是基于越来越多的人认识到多语言实际上在许多方面是有用的。VAAG的目标是广泛的，跨学科的重新评估与一般认知科学，语言语义学，实验心理学，形式语用学和计算机科学的贡献。一个适合这些用法的“非同一性”理论不仅要解释非同一性是如何表现的，而且要解释它为什么存在，它的用法是什么，以及它是如何在句子内部和话语中受到约束的。VAAG项目的三个中心主题如下：a）语义性和认知效率的关系，B）语义性、粒度和意义的组合性，以及c）不同类型语义性之间的区别。第一个主题解决了为什么诚实有用的问题。我们探索的直觉是，在精确的信息可以提供太多的信息的情况下，准确性可以是有用的。第二个主题是关于近似，它通过像\n {exactly}这样的明确标记或作为概念组合的一般副作用来调节相似性。第三个主题与其他两个主题都有联系，因为不同的动机似乎会导致不同的类型，而这些需要不同类型的近似策略。通过这种方式，VAAG项目使用广泛的多学科视角针对智能交互模型的核心问题。

Vagueness is a pervasive property of human language and cognition that is important for all fields of research that make use of symbolic representations. Approximation refers mechanisms that regulate vagueness in meaning composition, and granularity to mechanisms that divide meaning spaces. While vagueness has often been regarded as undesirable, the VAAG project is based on a growing recognition that vagueness is actually in many respects useful. VAAG targets a broad, interdisciplinary reassessment of vagueness with contributions to general cognitive science, linguistic semantics, experimental psychology, formal pragmatics and computer science. An appropriate theory of vagueness for these uses must not only explain how vagueness is represented, but also why it exists, what its uses are, and how it is constrained sentence internally and within discourse. Three central themes of the VAAG-project are the following: a) the relation of vagueness and cognitive efficiency, b) vagueness, granularity and the compositionality of meaning, and c) the distinction between different kinds of vagueness. The first theme addresses the question why vagueness is useful. We explore the intuition that vagueness can be useful in scenario where precise information can provide too much information. The second theme concerns approximation which regulates vagueness either through explicit markers like \emph{exactly} or as a general side-effect of concept combination. The third theme connects to both of the other themes since different motivation of vagueness seem to lead to types of vagueness, and these require different types of approximation strategies. In this way, the VAAG project targets a central issue for models of intelligent interaction using a broad multi-disciplinary perspective.