Search

Looks for a method to formulate algorithms useful for the synthesis of Czech in machine translation where the point of departure for the synthesis can be more or less identical with the semantic representation. (Author/PF)

This annotated bibliography lists approximately 100 books, reports, Rand memoranda, and papers, written between 1961 and 1971, in the fields of linguistic theory, language research, computational linguistics, and formal linguistics. Many documents involve the use of computers in problems of language analysis; many concern the analysis of Russian. Several works on machine translation are also listed. (VM)

This report describes an optimization phase of the SYSTRAN (System Translation) machine translation technique. The most distinctive characteristic of SYSTRAN is the absence of pre-editing; the program reads tapes containing raw and unedited Russian texts, carries out dictionary and table lookups, performs all syntactic analysis procedures, and prints out translated texts without human direction. The report begins with a schematic introduction to SYSTRAN, followed by an extensive discussion of improvements effected during optimization in the dictionary, parsing system, and semantics. Improvements in other areas, i.e., the translation of homographs, adverbs, proper names, numerals, adnominal genitives, verbs plus objects, sequence of tenses, abbreviations, and articles, are also reported. A discussion of programing specifics in improving the efficiency of SYSTRAN processing is included, followed by remarks on SYSTRAN and the controversy concerning machine translation in general. Appendixes provide further details on the operation of the program. (Author/VM)

The translation of literature from one language into another is perhaps one of the most important problems in view of the fact that an inordinately large amount of human labor is expended on translation. A particularly complex problem arises before the scientists of all countries in connection with the need to become familiar with scientific literature and with scientific accomplishments described in the most varied languages of the world. This problem can be alleviated to a large degree by the mechanization of translation, the possibility of which has arisen with the development of automatic electronic computers. We already have some definite accomplishments in this domain, which demonstrate that machine translation has become a reality. The development of machine translation raises many legal questions, some of which are: the possibility of recognizing a machine translation as subject to copyright law, who should hold the copyright, and the legal protection of such a translation. (Author/SJ)

The ten papers in this collection are based on presentations given during a colloquium held November 14-15, 1969, at the University of Toronto. The papers, written in either English or French, cover various topics related to prosodic feature analysis in general; several concern the analysis of specific languages. General topics include discussions on physical correlates of some prosodic features, the expressive function of intonation, relative height, machine translation, the contribution of prosodic analysis to the study of phonology, and language acquisition. More specific language topics include remarks on French and Czech fixed accent, accent in the Eastern Slavic languages, rhythm in the language of French Canada, and the analysis of tone in the Tibeto-Burman languages of Nepal. Also included is a report of a round-table discussion concerning pitch extraction, recognition of intonational patterns, and speech synthesis. Questions and discussion on the various topics are included throughout the book. (VM)

A method is proposed in which nondefined Chinese characters may be uniquely classified thus making them compatible for machine translation. An optical-digital device is used to locate nondefined geometric shapes within Chinese characters via spatial filtering techniques and cyclic cross-correlation. Seventeen nondefined geometric shapes are found in a 98 character sample which is representative of a ten thousand Chinese character vocabulary. Preliminary results indicate that 60 percent of the geometric shapes are identified exactly and over 90 percent can be identified within specified tolerances. Geometric shape energy content is shown to be pertinent in the decision process. (Author)

A study was conducted to determine the feasibility of developing an automatic Chinese/Japanese input device for a machine translation complex. A specific scanning technique to digitize Chinese/Japanese ideographs is recommended as a result of the study. Results of the study include specification of the resolution necessary to successfully digitize the Chinese/Japanese ideographs, the minimum number of gray levels required for successful digitization, the recommended matrix size necessary for comparison/recognition analysis, and the logic required in the scanner to successfully digitize Chinese/Japanese ideographs at a reasonable rate. The study also determined that noncharacter recognition should be handled in a mini-computer rather than with built-in scanner logic. The conclusions reached with regard to resolution and gray scale requirements are supported by hard-copy displays of Chinese/Japanese ideographs. Scans of Chinese/Japanese test ideographs imaged onto film are made at the recommended resolution and displayed using the Mead Digital Graphics Generator. This unique display device made possible an examination of the effects on Chinese/Japanese ideographs of scanning with various resolutions and utilizing different gray scales. (Author)

The Multilingual Interpreter (MI), a language processor capable of executing more than one language, is being used at Western Washington State College to enhance interactive computing for computer-assisted instructional (CAI) users. The MI's two major functions are, first, that of scanning and translating a source language into internal interpretive code, accomplished by programs called front-end processors, and, second, that of interpreting, or executing the internal code, accomplished by a program called the generalized interpreter. As a result, unsophisticated users are equipped with an easy-to-learn language whose power can continue to satisfy them as their needs increase in complexity. APL is also provided despite the fact that the available core is limited to 84k. (PB)

This issue contains five articles dealing with several aspects of linguistic theory. The first, "Origine et evolution de la notion de 'situation' de 'l'Ecole linguistique de Londres': de Malinowski a Lyons," examines the origin and meaning of the concept of "situation" in articles by the London school of linguists. The second, "Some Aspects of a Semantic Taxonomy," discusses conditions required for structural description of linguistic content. The third defines "vigilant memory" and its application to the segmentation and error-correction of natural language input. A description of the mechanical translation system developed at the Universite de Montreal includes discussion of the theory, computer system, steps in translating, and future prospects. The last article, "Universal Quantifiers and Entailment," explores the relationship of reason and condition clauses. (RL)

The compilation of a dictionary for automatic translation from Russian to English is described in detail. Emphasis is placed on: (1) morphological and functional classifications of Russian words, (2) dictionary compilation, (3) the structure of signs, (4) automatic information-processing machines, (5) from the automatic dictionary to the automatic translator, and (6) dictionary operation. Less extensive chapters describe: (1) entry keys for an automatic dictionary, (2) problems in compilation and operation, (3) the problem of translation, and (4) flow charts and automatic coding. Many diagrams, tables, and graphs are included. (AF)

In this collection of articles by 25 linguistic experts, aspects of modern linguistics from language origins to computer programs are covered. Subjects discussed include phonology, morphology and syntax, lexicology and semantics, the origin of language, the English language family, English language history, British and American standard and nonstandard dialects, communication systems, national and international languages, dictionaries and usage, machine translation, computer linguistics, simulated speech, J. R. Firth and British linguistics, the Prague School of Linguistics, American schools of linguistics, glossematics, generative grammar, the nature and history of linguistics, and the relationships between linguistics and anthropology, psychology, literature, and instruction in native and foreign languages. (LH)

Since natural language may be assumed to have an underlying conceptual structure, it is desirable to have the machine structure its own experience, both linguistic and nonlinguistic, in a manner concomitant with the human method for doing so. This paper presents some attempts at organizing the machine's information conceptually. The different systems discussed include Deese's Categorical Dictionary, Salton's SMART, Wood's Question-Answering System, Quillian's Semantic Memory, the DEACON System, and Schank's Conceptual Dependency. The attempts are discussed and coordinated into a framework for what may be a practicable system. (Author/BB)

This report gives the results obtained in the semantic and syntactic analysis of the Japanese particles "de,""ni,""e," and "wo" in comparison to their equivalents in English, German, and Spanish. The study is based on the so-called "Correlational Analysis" as proposed by Ernst von Glaserfeld. The combinations of distinctive semantic features of the "correlata" with the relations are summarized in tables. A detailed and diagramed description is given of two examples showing the transition from any of the four languages to any of the others in the translation process concerning the particles involved. The model is based on the logical functions of conjunction, dejunction, and disjunction as suggested by A. Hoppe. The Japanese particles--particularly "de,""ni," and "wo"--have a wide semantic range. This will be a problem in machine translation but could be approached through coding the distinctive correlational features of the words in the dictionary and through adequate programming for the machine translation process. On the other hand, this lack of linguistic explicitness forces the speaker and hearer of Japanese to extract considerably more information from the context, both linguistic and social, than the speakers of the other three languages to achieve comprehension in the communication process. (Authors/DO)

The use of a system for on-line conversation in ordinary English grammatical patterns is illustrated by a number of examples showing how data bases of diverse content and structure may be described, interrogated, and modified. It is suggested that a system composed of a user extendable subset of English, a natural language compiler to translate the English sentences into storage and search procedures, and a data management machine to execute these procedures, can provide an on-line user with a natural, easy-to-use, and powerful means of man-to-machine communication. (Author)

Techniques of reversing English-Russian scientific and technical dictionaries into Russian-English versions through semi-automated compilation are described. Sections on manual and automatic processing discuss pre- and post-editing, the task program, updater (correction of errors and revision by specialist in a given field), the system employed for the rearrangement of entry pairs into a Russian alphabetization, and the final formatter program. Also included are explanations of arrangements made for the publication of the individual computer-reversed dictionaries. Appendixes include information on dictionaries selected for reversal, a system design flowchart, examples of the text in various stages, keypunch information, meeting agenda, and a letter to the dictionaries' reviewers. (AF)

A report on a series of computer programs being written to make frequency counts designed to indicate the learning problems to be encountered in any approach to a target literature discusses how these programs can be used to aid in foreign language vocabulary learning. (Author/AF)

This book, a revised and expanded English version of the author's Italian work "La Linguistica strutturale" (1961), is a survey of the main trends in structural linguistics intended not only for the linguist but for specialists in other fields and the general reader as well. The initial chapter, "Introductory Notions," discusses general linguistic problems. Subsequent chapters are devoted to: Ferdinand de Saussure, the Prague School (N.S. Trubetzkoi), the Copenhagen School (Louis Hjelmslev), the early development of American structuralism (Edward Sapir and Leonard Bloomfield), functional linguistics (Andre Martinet and Roman Jakobson), American structural linguistics (the Neo-Bloomfieldians), transformational grammar (Noam Chomsky), and mathematical linguistics and machine translation. (FWB)

CUM LAUDE (Computerized Understanding of Morphology--Language Acquisition under Development in Education) is a unique attempt to develop a system of computer programs enabling a machine to "understand" the grammatical relationships between the various parts of speech in a natural language. In response to an analysis of current computer-assisted language instruction (CALI) programs, it is argued that the full capability of a computer in drilling students on elementary language grammar will be attained when a greater understanding of the morphology of languages is built into the computer. Such an approach ultimately would enable a computer not only to instruct a language dynamically in an educational capacity, but also to make available its usefulness to those language teachers not well trained in machine languages. This project includes the exemplary development of a system of programs which teaches the verb formations of three languages: French, German, and English. The system is flexible in that the student has control over the selection of the verbs and tenses upon which he will be drilled. (Author)

In many cases machine-translation does not produce satisfactory results within the framework of purely formal (morphological and syntaxic) analysis, particularly, in the case of syntaxic and lexical homonomy. An algorithm for syntaxic-semantic analysis is proposed, and its principles of operation are described. The syntaxico-semantic structure is established in the course of this analysis. A special syntaxic analysis is absent, the rules of syntax playing a subordinate part and being used to remove the homonomy. The language of the RX codes is proposed as the intermediary language. Many of the machine-translation rules required can be obtained on a computer. Partial automation of the development of rules can be achieved with the aid of existing algorithms. (Author)

Five articles which were selected from a Russian language book on cybernetics and then translated are presented here. They deal with the topics of: computer-developed computers, heuristics and modern sciences, linguistics and practice, cybernetics and moral-ethical considerations, and computer chess programs. (Author/JY)

At Wayne State University research activity in the Department of Slavic Languages during the reporting period was concentrated on linguistic analysis, documentation, and on revising existing programs to run on new hardware. A survey was made of activities in the field of machine translation during the 1960's. Two studies were conducted by members of the staff: one on the classification and translation of certain Russian syntagmas and the other on the classification and translation of Russian sentence constituents. Research was also conducted on the automatic translation of telegrams from Japanese into English. A list of references is appended to the latter report. (JY)

The SIRIUS language is intended for solving both numerical and analytical problems. The new language is realized by means of a two-phase translator. The first phase is a translation into an intermediate (Polish-nonparenthetic) language; the second phase is an interpretation from this language. Two modes of program execution are envisaged in the SIRIUS language: automatic programing and man-machine "conversation" programing. Programs for problem solution include two types of information: initial nonalgorithmic and algorithmic. The first type includes formulas, equations, and relations written in the generally used linearized notation; the second type includes a sequence of instructions expressed by Russian verbs in the imperative form. The SIRIUS language uses Russian, Latin, and Greek characters, numerals, true-false symbols, operation symbols, operators, and housekeeping words and symbols. (Author/JY)

Cited are 573 U.S. and foreign articles, reports, and books particularly relevant to the field of computational linguistics with selective coverage in the fields of computation and programming, and social science uses of computers as language processors. In the area of linguistics, a fairly broad view of structural theory and semantics is taken without being exhaustive. Structural properties of some languages, especially English and Russian are also selectively covered. The coverage of psycholinguistics and sociolinguistics is also selective. (Author/MF)

The R2 Information Retrieval System, when complete, will be a question-answer information retrieval system in which a person, having questions over a pre-determined data base, will be able to receive answers to his questions. The paper concerns the initial processing of the data base for the R2 system and explores initially several simple strategies for retrieving relevant statements, rather than answers, for a given question over the data base. The paper is first concerned with the construction of a 'phrase dictionary' from the data base, and, secondly, with the development of several initial attempts for information retrieval from the data base. (Author)

In order to effect the unambiguous identification of particular members of sets of homonyms--homophones or homographs as the case might be--during the automated conversation of written language to spoken language and/or vice versa, additional linguistic information must become available regarding the primary thought-to-physical-manifestation transfer. The homoneme distinguishes among homonyms as does the phoneme among the members of a commutation set. The "existence" of the homoneme is proposed and discussed. Working definitions of certain relevant terms are offered, as well as a statement and demonstration of the problem(s). (Author/AMM)

Chapter I, an introduction to machine translation of languages, contains a simplified description of electronic computing machines, and a discussion of the advantages and disadvantages of machine translation research. A historical background of machine translation of languages is given, together with a description of various machine translation projects throughout the world, and the present state of the art is assessed. Chapter II presents a theory of translation that requires three structural grammars: a structural grammar of the source language (Hebrew), a structural grammar of the target language (English), and a transfer grammar that transforms the structural elements of the source grammar to those of the target grammar. Each grammar is discussed and illustrated together with its limitations and problems. An introduction to the subject of structural grammar is given, and the various types of structural grammars are described and illustrated. A context-sensitive, complex-constitutent, phrase-structure grammar is selected for the problems of machine translation from Hebrew to English; reasons for this selection are given. Methods for the evaluation of translation quality are described and a new simple method is developed and illustrated. Various problems of the transfer grammar are considered and solutions of the problems are presented and illustrated. (Author/AMM)

This survey of the field of Chinese language computational linguistics was prepared as a background study for the Chinese Linguistics Project at Princeton. Since the authors' main purpose was "critical reconnaissance," quantitative emphasis is on systems with which they are most familiar. The complexity of the Chinese writing system has presented a great challenge to computational linguists. Surveys by the Chinese Linguistics Project show that enough aspects of the problem have been tackeled and solved by various people so that computer analysis of Chinese texts may be considered feasible. In fact, it is being approached in varying ways at several places. This survey is organized into sections on--(1) Input (Form of Code, Mechanics of Coding); (2) In-Machine Manipulation (Phonological Analysis, Inventory and Statistics); (3) Output (Form and Purpose, Mechanics); (4) Notes on this survey; (5) Illustrations of Selected Computer-Used or -Generated Calligraphy; and (6) a Bibliography. More than 30 projects are discussed, ranging from very early attempts to deal with the problem (Standard Chinese Telegraph Code), to on-going research projects in the United States and abroad. (JD)

The conference proceedings include papers on sensory aids for visually handicapped mobility and reading. Two papers each treat mobility as a general problem, sociocultural surveys on mobility and reading, and echolocation in man and bats. Five papers concern reports and evaluations of practical trials of the sonic monaural aid; one deals with psychological aspects of mobility guidance displays; two report sensory researches in the Netherlands and Canada; and five on mobility aids consider current researches on sonic and optical systems. Other areas discussed are reading aids (four articles), rehabilitation and mobility (one article), and sensory evaluations and researches in the United States (three articles). Also provided are the opening and summary remarks for the conference and a panel discussion of future work. Appendixes outline workshop sessions, a visit to the Autonomics Division of the National Physical Laboratory, and aspects of tactile stimulation. (JD)

The author considers it essential that linguists develop the means to communicate their findings to those who may need them. The applications of linguistics have spread far beyond the traditional field of foreign language study to native language acquisition, social functions of speech differences, literacy, language standardization, interpretation of the learning process, cross cultural and cross national understanding, manipulation of natural language data for mass communications media, communication between man and machine, machine translation, and information storage and retrieval. Although the subject matter may be technical, linguistic information should be made available in comprehensible form to prospective users. Establishment of a clearer structuring of the field of linguistics would also make it easier for non-linguists to find materials on topics of concern to them. The author asks, "What are the limits of linguistics? What does linguistics include? How should the internal hierarchies be classified to provide for most effective transmission of linguistic knowledge for meaningful use by specialists in other disciplines?" (MK)

The proceedings include papers on computer translated braille (grade 2, display devices, recent developments), the new braille translation program and research at the American Printing House for the Blind, the braille electric typewriter, the Massachusetts Institute of Technology braille embosser, the braille reader, and braille transmission for the Instructional Materials Center for the Visually Handicapped in Chicago. Also considered are the professional activities of blind programers, higher level computer language for computer translations, the math translation program, braille out-puts from reading machines, IBM's role in research and development of aids for the blind, automatic braille translation in Germany, conversion of compositor's tape to braille, and planning for computerized braille. A list of participants and an appendix of the conference agenda are included. (LE)

THE PAPERS IN THIS VOLUME--CATEGORIZED AS EITHER (1) AUTOMATIC ANALYSES OF NATURAL LANGUAGES, (2) STATISTIC AND SEMANTIC ANALYSES OF LINGUISTIC DATA, OR (3) ALGEBRAIC THEORIES OF LANGUAGES--ARE THE COLLECTIVE PRODUCT OF THE INTERNATIONAL CONFERENCE ON COMPUTATIONAL LINGUISTICS HELD AT GRENOBLE, FRANCE, AUGUST 23-25, 1967. TOPICS RANGE FROM ALGORITHMS FOR DEPENDENCY GRAMMARS AND SLANT GRAMMAR CALCULUS TO AUTOMATED FIELD METHODS, AUTOMATIC TRANSLATION, AND CRITICAL COMMENTARY ON PREVIOUS LINGUISTIC WORK. THE THIRTY-NINE PAPERS REFLECT CURRENT THOUGHT AND TRENDS IN COMPUTATIONAL LINGUISTICS IN FRANCE, GERMANY, RUMANIA, THE SOVIET UNION, AND THE UNITED STATES. ALL ENTRIES ARE IN EITHER ENGLISH OR FRENCH. INFORMATION CONCERNING THE AVAILABILITY OF THIS DOCUMENT MAY BE OBTAINED FROM C.E.T.A., BOITE POSTAL NO. 8, 38 SAINT-MARTIN-D'HERES, FRANCE. (FB)

Various computer applications in linguistics, primarily speech synthesis and machine translation, are reviewed. Although the computer proves useful for statistics, dictionary building and programmed instruction, the promulgation of a world auxiliary language is considered a more human and practical solution to the international communication problem. (Text is in German.) Available from Humanities Press, Inc., Atlantic Highlands, NJ 07716. (TL)

Discusses the use of computers for research in the humanities, where their characteristics of "speed, flexibility and economy make them indispensable "to the production of literary concordances, dictionaries, and data stores. (Editor/WM)

Detailed version of a speech at the third Congress on Computational Linguistics in Stockholm, Sweden, September 1969. (VM)