Semiotics in Databases Bernhard Thalheim [0000−0002−7909−7786] Christian-Albrechts University at Kiel, Dept. of Computer Science, D-24098 Kiel, Germany thalheim@is.informatik.uni-kiel.de http://www.is.informatik.uni-kiel.de/∼thalheim Abstract. In database research and practice, syntax is commonly con- sidered a “firstness” property, while semantics is a “secondness” prop- erty (in the sense of Peirce); pragmatics is largely neglected. This paper discusses foundations in first-order predicate logic, highlights its useful- ness, but also point out its problematic issues. These cover in particular safe expressions in the relational tuple calculus, rigid normalisation em- phasising atomicity of attributes, and a large body of knowledge on database constraints. Database theory is still oriented on flat relational structures although systems became object-relational. We first survey the goodliness of classical database constraint theory and then develop an alternative approach to some constraints including (α) the handling of constraint sets instead of homogeneous classes of constraints, (β) vi- sual reasoning on constraints and structures, and (γ) calculi for robust reasoning, in particular for “exceptions” and object-relational reasoning. Keywords: constraints, dependencies, database theory, semiotics 1 Introduction 1.1 Semiotics: Syntax  Semantics  Pragmatics Semiotics is the study of signs and sign-using behaviour. According to Peirce, signs are icons, indexes, or symbols. The structuralist approach to semiotics distinguishes syntax as the arrangement of words and the study of formation of sentences, semantics as the study of meaning in natural and artificial languages, and pragmatics as the study of relations between languages and their users, e.g. the use of languages in communication. Morphology studies the internal construction of words. Natural languages use typically semantically fixed base elements from a vocabulary and have general purpose means for expressions. They use a manifoldness of expression forms and speech acts. They have a high error tolerance, use a metalanguage on its own, don’t distinguish abstraction levels, interpret connectives and quantifies in a variable way, widely use am- biguities and ellipses, apply indefiniteness as a concept for later focusing, and use tempus and modality as well as metaphoric elements with an embedment of context. Therefore, semiotics in linguistics is far more broad than in computer sci- ence and database research. Main elements of semiotics are, however, inherited keynote paper Proc. MEDI2019, LNCS 11815, 3-19 Springer