A LINGUISTIC MODEL OF HUMAN LANGUAGE MUST ACCOUNT FOR BOTH "SENTENCES" AND "MEANINGS" AS WELL AS EXPLAIN HOW SPEAKERS ARE ABLE TO TRANSLATE SENTENCES INTO MEANINGS AND VICE VERSA. THE AUTHOR DOUBTS THAT THE RULES OF A TRANSFORMATIONAL GENERATIVE GRAMMAR WILL BE ABLE TO CONSTRUCT A TRANSDUCER CAPABLE OF TRANSLATING MEANINGS INTO SENTENCES. "TRANSDUCTIVE" REFERS HERE TO A LINGUISTIC THEORY WHICH DEFINES A SEMANTIC AND A PHONETIC NOTATION, A "TRANSDUCTION RULE" FORM, AN "UPWARD TRANSDUCER" OR ALGORITHM WHICH TRANSLATES PHONETIC EXPRESSIONS INTO SEMANTIC EXPRESSIONS, AND A "DOWNWARD TRANSDUCER" WHICH TRANSLATES SEMANTIC INTO PHONETIC EXPRESSIONS. A LANGUAGE IS CHARACTERIZED BY A SET OF TRANSDUCTION RULES SPECIFYING THE OPERATION OF BOTH TRANSDUCERS. A FORM OF RULES IS DESCRIBED WHICH TRANSLATE SEMANTIC EXPRESSIONS IN A NOTATION AKIN TO THAT OF THE PREDICATE CALCULUS INTO P-MARKERS. THE SAME RULES, INTERPRETED BY A SLIGHTLY DIFFERENT TRANSDUCER WILL TRANSLATE P-MARKERS INTO SEMANTIC EXPRESSIONS. THE AUTHOR DOES NOT PROVIDE A SET OF SEMANTIC PRIMITIVES, BUT RATHER THE FORMAL SYSTEM WHICH WILL RELATE EXPRESSIONS INVOLVING SEMANTIC PRIMITIVES WITH SENTENCES IN A REAL LANGUAGE. OTHER ISSUES RAISED HERE INCLUDE--(1) THE RELATIONSHIP BETWEEN THE EXPRESSIONS WRITTEN IN THIS FRAMEWORK AND EXPRESSIONS IN A GENUINE LOGIC SUCH AS THE FIRST-ORDER PREDICATE CALCULUS, (2) THE ROLE OF RHETORICAL PRDICATES, AND (3) THE DISTINCTION BETWEEN "GIVEN" AND "NEW" IN THE STUDY OF SEMANTICS IN NATURAL LANGUAGES. THIS PAPER WAS PREPARED FOR INCLUSION IN "RECENT DEVELOPMENTS IN LINGUISTICS," EDITED BY M. BIERWISCH AND K. HEIDOLF, MOUTON, THE HAGUE (FORTHCOMING). A SHORTER VERSION WAS PRESENTED AT THE TENTH INTERNATIONAL CONGRESS OF LINGUISTS, BUCHAREST, 1967. IT IS AVAILABLE FOR $1.00 (CHECK OR MONEY ORDER) AS RAND PUBLICATION NO. P-3746 FROM THE REPORTS DEPARTMENT, THE RAND CORPORATION, 1700 MAIN STREET, SANTA MONICA, CALIFORNIA 90406. (JD)