COMMUNICATION SYSTEM WITH TEXT MESSAGE RETRIEVAL BASED ON CONCEPTS INPUTTED VIA KEYBOARD ICONS

A Natural Language Pro- cessing System utilizes a symbol parsing layer (5) and an intelligent word parsing layer (5) to produce a syntactically or pragmatically cor- rect output. A plurality of polysem- ic symbol sequences are input through a keyboard with semantic, syntactic, or pragmatic segments in- cluding agent, action and patient segments, for example. One poly- semic sequence is input from each of the three segments of the key- board (1). A symbol parsing device, then parses each symbols in each polysemic symbol sequence to access a previously stored word, morpheme, or phrase. The accessed word, morpheme, or phrase (50) corresponds to the input polysemic symbol sequence and further corresponds to one of the designated agent, acti on or patient segments. Each accessed word, mor- pheme, or phrase further accesses corresponding and previously stored grammatical and semantic information. An intelligent word parsing layer (5) subsequently applies each of the plurality or words morphemes, or phrases to a predetermined hierarchy of rules based upon the accessed grammatical and semantic information. The intelligent word parsing device subsequently parses the received plurality of accessed words, morphemes, or phrases into a syntactically or pragmatically correct output. 2089177 9202890 PCTABS00010 A Natural Language Processing System utilizes a symbol parsing layer (5) and an intelligent word parsing layer (5) to produce a syntactically or pragmatically correct output. A plurality of polysemic symbol sequences are input through a keyboard with semantic, syntactic, or pragmatic segments including agent, action and patient segments, for example. One polysemic sequence is input from each of the three segments of the keyboard (1). A symbol parsing device, then parses each symbols in each polysemic symbol sequence to access a previously stored word, morpheme, or phrase. The accessed word, morpheme, or phrase (50) corresponds to the input polysemic symbol sequence and further corresponds to one of the designated agent, action or patient segments. Each accessed word, morpheme, or phrase further accesses corresponding and previously stored grammatical and semantic information. An intelligent word parsing layer (5) subsequently applies each of the plurality of words, morphemes, or phrases to a predetermined hierarchy of rules based upon the accessed grammatical and semantic information. The intelligent word parsing device subsequently parses the received plurality o