| Based on the theories of Neurocognitive Linguistics and Cognitive Functional Model, this present paper analyzes verbs in mental domain in order to construct structures reflecting the realizational relationship between conceptual structures and grammatical structures.Sydney Lamb, the originator of Neurocognitive Linguistics, emphasizes the close relationship between the language cognition and neural networks in human brain. He creates the relational network graphs to simulate the linguistic operation. Cognitive Functional Model, proposed by Cheng Qilong, aims to explain the linguistic system and further to interpret the operation of the system, and attempts to reflect the semiotic relationship between conceptual content and formal expression. In CFM, the linguistic system is a hierarchical organization that involves conceptual stratum, semantic stratum, grammatical stratum, phonological stratum and so forth.This paper, applying three methods, namely, analysis, catalysis and testing, aims at constructing structures of mental clauses. Firstly, it analyzes the conceptual structures and grammatical structures as well as their realizational relationships; Secondly, by means of catalysis, it generalizes the mental structures which contain the conceptual structures and grammatical structures with their realizational relationships; thereafter, in the light of Lamb's graphs, it draws the relational network graphs to manifest the mental structures. Lastly, it tests the operational validity and neurological plausibility of the structures through computer program compiled by language C++.With above study, I am able to summarize the findings. Firstly, in the mental domain, the basic conceptual types of mental clauses are space structure, action structure and causation structure, among which the action structure can be conflated with the space structure or the causation structure. Secondly, the mental verbs can be categorized into 10 types: KNOW, LOSE, ACCEPT, THINK, LOOK, SEE, LEARN, RECALL, CONSIDER and TEACH type. Thirdly, testing proves that the input and output of the structures are supported by the linguistic phenomena. The operationprocess of the structures is not incompatible with that of the neural network. Therefore, the structures are operational. |
