Database Semantics

Abstract

The Slim (As explained in the Introduction, Sect. XI, the acronym Slim abbreviates Surface compositional, time-Linear, Internal Matching, which are the methodological, empirical, ontological, and functional principles of the approach. When the acronym is viewed as a word, Slim indicates low mathematical complexity, resulting in the computational efficiency required for real-time processing.) theory of language outlined in Chaps. 3–6 serves as the theoretical foundation for a computational model of natural language communication, called Database Semantics (DBS). (As the name of a specific scientific theory, the term Database Semantics is capitalized. DBS is distinct from the generic use of the term in the sense of “semantics of databases.” For example, Bertossi et al. (eds.) (2003) discuss semantic constraints on databases.) Database Semantics models the transfer of information from the speaker to the hearer by representing the knowledge of speaker and hearer, respectively, in the form of agent-internal databases. Natural language communication is successful if a certain database content, encoded by the speaker into natural language signs, is reconstructed analogously in the database of the hearer.

Exercises

Section 22.1

• 1. Explain how an agent’s internal context of use may be viewed as a database.

• 2. Describe the storage and retrieval problem for modeling natural language communication.

• 3. How does the interaction between a user and a conventional database differ from the interaction between a speaker and a hearer?

• 4. Why is there no turn-taking in the user’s interaction with a conventional database?

• 5. Explain the basic procedures of the speak and the hear mode in natural language communication by using a simple example of a context.

Section 22.2

• 1. Is it possible to reinterpret model-theoretic semantics as a formal representation of the context of use? What would be required for this, and how would such a reinterpretation modify the original goal of logical semantics?

• 2. Expand the model structure defined in 22.2.1 by adding one more index and the predicate sleep such that the falling asleep of Felix and Fritz and the waking up of Zach and Eddie are being modeled. Write an evaluation of your experience.

• 3. Describe the basic principles of frame theory and compare it with model theory.

• 4. Is it possible to reinterpret frame-theoretic semantics as a computational representation of the context of use? How would such a reinterpretation modify the original purpose of frame-theoretic semantics?

• 5. Why does the use of a frame-theoretic semantics within the Slim theory of language necessitate a change of the original ontology?

• 6. Is the use of a uniform formalism for representing the levels of (i) the language meaning₁ and (ii) the context of interpretation a necessary or a sufficient condition for the successful definition of an internal matching pragmatics?

• 7. Is it a good idea to construct a [+sense] system by combining two instantiations of a [−sense] system? What kinds of problems are likely to occur?

• 8. Why is the format of 21.4.3 unsuitable for the storage and retrieval task of natural language communication.

Section 22.3

• 1. What are the requirements on a semantics in addition to being suitable for a modeling of reference as an agent-internal pattern matching?

• 2. What is the difference between the variables NP and NP′?

• 3. Explain the difference between the format of uninterpreted NEWCAT rules (e.g., 22.3.2, 22.3.5) and the format of unified DBS rules (e.g., 22.3.4, 22.3.7).

• 4. How are lexical meanings like field, triangle, or contain coded in DBS?

• 5. How are the lexical meanings composed into propositions?

• 6. How are propositions composed into complex content?

• 7. How are semantic relations between proplets coded by the unified DBS rules in the hear mode?

Section 22.4

• 1. Does the definition of content as a(n order-free) set of proplets help or hinder pattern matching between the language and the context level (reference)?

• 2. How does pattern matching between the language and the context level differ from the pattern matching between a rule pattern and a content?

• 3. Explain two ways for handling nonliteral use.

• 4. Why are the inferences for interpreting nonliteral uses of natural language expressions independently motivated?

Section 22.5

• 1. How does a sentence start in the unified rule format differ from a sentence start in the NEWCAT format?

• 2. How does this difference affect the role of the verb in time-linear addition of a postverbal phrasal noun?

• 3. Does a similar difference occur also in the combination of a preverbal phrasal noun with the finite verb?

• 4. What is the database aspect in the cycle of natural language communication and why is the unified rule format essential for this aspect?