Chinese character structure has often been described as representing a kind of grammar, but the notion of character grammar has hardly been explored. Patterns in character element reduplication are particularly grammar-like, displaying discrete combinatoriality, binarity, phonology-like final prominence, and potentially the need for symbolic rules (X→XX). To test knowledge of these patterns, Ch...

We study a variant, in the context of the Erdős–Ko–Rado Theorem for vector spaces [4], of an extremal problem of Erdős and Rothschild [3], who considered edge-colorings of graphs avoiding monochromatic triangles. For fixed positive integers r, k and ` with 1 ≤ ` < r, and a family F of linear r-dimensional subspaces in a linear n-dimensional vector space Vn over the fixed finite field GF (q), le...

Two research directions are to be explored in realizing our group’s TREC QA system in 2006. The first one is to investigate the possibilities of applying linguistically sophisticated grammatical framework in tackling the realworld natural language processing task such as question answering. The other is to exploit the possible world’s entities and relations as described in online encyclopedia i...

We present a system that integrates speech and image understanding in the domain of a construction task. Knowledge needed to interpret speech and images is stored in one homogeneous knowledge base using the semantic network formalism ERNEST. It provides a means for constraint propagation and therefore – using a uniform representation of the information gained from real input data – the speech a...

The present study explores two questions: What is the nature of older children’s syntactic knowledge; how is that knowledge used in an everyday speech situation? Six-, eight-, and ten-year-olds repeated grammatical sentences as read by the first experimenter. Half the sentences were syntactically clear, half slightly distorted. Clear versions displayed basic grammatical relations and constituen...

A number of recent studies have provided contradictory evidence on the question of whether grammatical class plays a role in the neural representation of lexical knowledge. Most of the previous studies comparing the processing of nouns and verbs, however, confounded word meaning and grammatical class by comparing verbs referring to actions with nouns referring to objects. Here, we recorded elec...

The framework we adopted for customizing linguistic knowledge to individual application domains is an integration of symbolic and statistical approaches. In order to acquire domain specific knowledge, we have previously proposed a rule-based mechanism to hypothesize missing knowledge from partial parsing results of unsuccessfully parsed sentences. In this paper, we focus on the statistical proc...

Starting from first principles, we have designed a rule-based firstorder logic (FOL) knowledge-based system (KBS), called EXPARS, and coupled it with an original inductive learning module for the purpose of grammatical inference. While we stress the flexibility of the KBS approach as applied to parsing, we also raise the issues imposed by such a coupling. The interaction between EXPARS and the ...