DOCUMENT RESUME ED 054 702 48 FL 002 627 AUTHOR Price, James D. TITLE A Computerized Phrase-Structure Grammarof Modern Hebrew: Part T, Complex-Constituent Phrase-Structure Grammars. Final Report. INSTITUTION Franklin Inst. Research Labs., Philadelphia, Pa. SPONS AGENCY Office of Education (DHEW), Washington, D.C. Bureau of Research. BUREAU NO BR-9-7722 PUB DATE Jun 71 CONTRACT OEC-0-9-097722-4411 NOTE 63p. EDRS PRICE MF-$0.65 HC-$3.29 DESCRIPTORS Algorithms; *Computational Linguistics;.*Comput Programs; Deep Structure; *Grammar;*Hebrew; Instructional Materials; Language Patterns; Language Research; Language Universals; Phrase Structure; Semitic Languages; Sentence Structure; Syntax;Tables (Data) ; Teacher Education; TransformationGenerative Grammar; *Transformation Theory (Language) ABSTRACT This first part of a four-part report of research on the development of a computerized, phrase-structure grammarof modern Hebrew presents evidence to demonstrate the need formaterial to train teachers of Semitic languagues in the theory of grammar.It then provides a discussion of the research alreadydone on the application of computational grammars to artificialand natural languages. Research procedures are discussed.Following a section on computational grammars, there is discussion of grammartheories and of several grammars which might be suitable forgenerating and analyzing Hebrew sentences. The general requirements of complex-constituent-phrase structure grammar are outlined andmethods for applying it to Semitic languages are discussed.A list of references is provided. For related reports see FL002 628 FL 002629, and FL 002 630. (VM) Title VI, NDEA, Section 602 PA.48Sureau#9-7722 FINAL REPORT --Project-No.-097722 ==- Ipt1 CO Contract No. 0EC-0-9-0977224411 C:3 %Ai Franklin Institute Report No. F:-C2585-1 A COMPUTERIZED PHRASE-STRUCTURE GRAMMAR OF MODERN HEBREW PART I U.S.DEPARTMENT OF HEALTH, EDUCATION & WELFARE OFFICE OF EDUCATION THIS ,DOCUMENT HAS BEEN REPRODUCED EXACTLY AS RECEIVED FROM THE PErON OR ORGANIZATION ORIGINATING II.POINTS OF VIEW OR OPINIONS , STAiED DO KIT NECESSARILY REPRESENT OFFICIAL OFFICE OF EDUCATION POliITION OR POLICY. James D. Pric? T e Franklin Institute Reseerrch-laboratories '20th and Race-Sreets' Philadelphia, Penw. 19103 4 "PDWATME g 11C EgF isoot,,JJA,49,91) 'InstltuteofIn1!ernationa1 FINAL REPORT Project No. 097722 Contract No. OEC-0-9-097722-4411 Franklin Institute Report No. F-C2585-1 A COMPUTERIZED PHRASE-STRUCTURE GRAMMAR OF MODERN HEBREW PART I Complex-Constituent Phrase-Structure Grammar5 James D. Price The Franklin Institute Research Laboratories Philadelphia, Penn. 19103 June 197 The research reported herein was performed pursuant to a contract with the Office of Education, U.S. Department of Health, Education and Welfare. U. S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE Office of Education Institute of International Studies SUMMARY VOLUMES I-1V Over the past several years, The FranklinInstitute Research Laboratories has conducted research on the applicationof computational grammars to natural and artificiallanguages. Research in natural languages has been confined to the Semiticbranch, modern Hebrew in particular. This report describes the results of the most recentresearch to help meet the need for material totrain teachers of Semitic languages (especially Hebrew) in the theory of grammar and toprovide basic com- puterized tools for further 1Lnguistic research inSemitic languages. The material developed provides the foundation,framework, and some of the basic building blocks, bnt many additions,corrections, and improve- ments must yet be made. The basic computerized research toolsprovided, howeecr, will greatly facilitate the ultimatecompleti of the material. This report of the development of a ComputerizedPhrase-Structure Grammar of Modern Hebrew has been preparedin four parts. Part I presents evidence to demonstrate the need for material to trainteachers of Semitic languages in the theory of grammar. Transformational theory is shown to be the best for this purpose. The background of the present prolect is given together with a survey of relatedresearch and a description of the procedures involved in carrying out the research. A discussion of the theory of grammar follows in which variousother types of structural grammars are examined. It is concluded that each type uses adifferent property of sentences as a basis fordescribing a language; that the other preperties become restrictions onthe selected property; that, granted-sufficient restrictions, each type candescribe a language equally well; and that several of the most prominent grammars maybe viewed as highly restricted phrase-structure grammarswhich may be considered "transformational" grammars. This conclusion is verified byadding restraints to a simple phrase-structure grammar sufficient for it todescribe Semitic languages. The resultant grammar is called acomplex-constituent phrase-structure grammar because of the setof subscripts added to the symbols. This grammar has the power to explainthe common deep-structure relationships that exist between such forms as theactive and passive voices by showing that they originate from different optionsof the same symbol. With a few simple rules in phrase-structurenotation, it has the power to explain the universal patterns of a languagethat transcend the bounds of phrases. By the use of semantic subscripts,it has a type of contextsensitivity sufficient for explaining the semanticconcord found in natural languages. All of this is provided by a relativelysmall number of unordered rules without a second system of notation(i.e., without one system for.phrase- structure rules, and another fortransformational rules). Finally, the general requirements of this grammar areoutlined, and methods for apply- ing it to Semitic languages are discussed. Part II describes in detail the applicationof this generalized complex-constitLent phrase-structure grammar tomodern Hebrew. It was found to be suitable for accuratelydefining the syntax and orthography of a Semitic language and for mechanization on acomputer. This was demonstrated by the high degree of successachieved in producing a computerized algorithm for generatingHebrew sentences (Part III), in producing a computerized algorithm foranalyzing Hebrew sentences (Part IV), and in testing the rules of the Hebrew grammarby means of the computer. Of the 47 sentences generated, 42 weregrammatically correct, two were correct except for asuperfluous period, and three contained errors that require futuremodification of the rules. In the process of generating these sentences, alarge percentage of the rules were tested, and in numerous cases the rules weremodified to correct de- ficiencies and errors in their originalversion. Part III describes in detail acomputerized algorithm for generating Hebrew sentences, and Part IVpresents a computerized algorithm for analyzing Hebrew sentences. Parts III and IV include flowdiagrams, a listing of the computer programsin FORTRAN IV, and instructionsfor their use. The algorithms were used to testand demonstrate the Hebrew grammar, the results ofwhich indicate that the grammar ofHebrew is essentially correct, but that some ofthe rules are in need of further occurred, they were due to the development. In all cz-ses where errors content of the rules and not tothe form of the grammar. Although the results of further development is neededin some areas of the grammar, the research provide good reason tobelieve that the generalized grammar such as Arabic. can be successfullyapplied to other Semitic languages iv ABSTRACT This is Part I of a four-part report of researchfor the development of a Computerized Phrase-StructureGrammar of Modern Hebrew. This part of the report presents evidence todemonstrate the need for material to train teachers of Semitic languagesin the theory of grammar. The background of the present project iS giventogether with a survey of ralated ,:esearch and a description of theprocedures involved in carrying out the research. A discussion of the theory of grammarfollows in which it is shown that several of tieexisting computational grammars of natural languages may be viewed as highlyrestricted phrase-structure grammars and thus as of approximatelyequal merit. Finally, the general requirements of one of these grammars, acomplex-constituent phrase- structure grammar, are outlined, and methodsfor applying it to Semitic languages are discussec:. In subsequent parts, the generalized grammar is applied to modern Hebrew and demonstratedby computer tests to be suitable for accurately defining the syntax andorthography of a Semitic language and for implementation on a computer. ACKNOWLEDGMENTS The author would like to acknowledge the valuable assistance of Professor Esra Shereshevsky, Chairman, Departmentof Hebrew and Near Eastern Languages and Literatures, Temple University, who served as consultant. Prof. Shereshevsky carefully reviewed themanuscript of Part II, made many helpful suggestions regardingthe grammar of Hebrew syntax, and checked countless details of the grammarof Hebrew orthography. The author is also grateful to the members of the facultyof Dropsie University, especially Dr. Federico Corriente,Professor of Semitic Linguistics, who supervised the work performedunder subcontract with that institution, and Mr. Ezra Cohen who classifiedthe 1040 most commonly used Hebrew words contained in the computerizedHebrew-English Dictionary (Appendix A, Part II) and made numeroushelpful suggestions regarding the