THE OBJECTIVE OF THIS PROJECT WAS TO DEVISE AN ALGORITHM FOR A STEM RECOGNITION PROGRAM DESIGNED TO SEARCH FOR THE ROOT OF ANY HEBREW WORD AS WELL AS TO DETECT INNER CHANGES ON THE GIVEN ROOT. SUCH AN ALGORITHM COULD BE USED IN LIBRARY CATALOGING AND IN CREATING INDEXES AND CONCORDANCES OF TEXTS IN THE HEBREW LANGUAGE. IN THIS STUDY THE HEBREW WORD WAS CONCEIVED AS A CONSONANTAL, MORPHOLOGICAL UNIT. CENTRAL TO THE ENTIRE RESEARCH WERE TABLES AND LISTINGS ORGANIZED ON A GRAMMATICAL BASIS AND SO DEVISED AS TO PRESENT CERTAIN PERTINENT CORRELATIONS BETWEEN VERBALS AND NOMINALS AND THE AFFIXAL ELEMENTS. THIRTY-SIX GRAMMATICAL CATEGORIES WERE SET UP. FOUR TYPES OF AFFIXES WERE CORRELATED WITH THESE CATEGORIES. THE COMPUTER TECHNIQUE CONSISTED OF FRACTIONATING TEST WORDS INTO PREVIOUSLY DEFINED ELEMENTS AND FORMING VARIOUS COMBINATIONS WHICH WERE SUBSEQUENTLY SUBJECTED TO VALIDATION. THE RESULT OF THE COMPUTER FUNCTIONING REMAINED VALIDATED BY OBJECTIVE CRITERIA. A SPECIAL REFERENCE DICTIONARY IS NOW BEING COMPILED FOR THE PURPOSE OF TESTING THE RESIDUUM OF GRAMMATICALLY LEGITIMATE COMBINATIONS. (AUTHOR/DO)