Linguistic and
engineering studies in automatic language translation of scientific Russian
into English
TABLE OF CONTENTS
CONTRACT
STATEMENT vii
BIBLIOGRAPHICAL CONTROL INFORMATION ix
PREFACE Erwin Reifler xi
OUTLINE OF
THE RESEARCH Erwin Reifler 1
PART ONE: LINGUISTIC ANALYSIS
A. Machine Language Translation Erwin Reifler 11
II. The Fundamental Problems 19
III. MT Linguistic Bibliography 35
B. MT Linguistics and MT Lexicography at the University of Washington Erwin Reifler 57
C. Lexicography Lew R. Micklesen 67
D. MT Operational Analysis Lew R. Micklesen 89
E. Appendix
The Simulated Machine Translations 183
PART TWO: ENGINEERING ANALYSIS
A. The Digital Computer in Machine Translation David L. Johnson 349
B. The Digital
Data-Processing Problem of Machine Translation
of Russian to English Robert E. Wall, Jr. 355
C. A Possible Application of Electronic
Computers to the
Block Analysis of Greek Sentences Aristotelis D. Stathacopoulos 449
D. Appendices
I. Pattern Recognition in an Electronic Reader Philipp M. Pahl and
David L. Johnson 475
II. Logical Programming Research in the University
of
Washington Machine Translation Project Robert E. Wall, Jr. 483
PREFACE
Machine Translation (MT) research at
the University of Washington has since May, 1956 been sponsored by
the
Directorate of Intelligence and Electronic Warfare of the Rome Air Development Center, Air Research and
Development
Command, of the United States Air Force. The first phase of this research,
concluded on June 30,
1958, has
been described in our previous comprehensive report.1
In the present report we have
endeavoured to detail our MT research and its results during the period
from
June 30, 1958, till October 31, 1959. The report, to which five linguistic and
engineering members have
contributed papers, consists of two main sections entitled LINGUISTIC ANALYSIS
and ENGINEERING ANALYSIS, re-
spectively. Both sections are preceded by an OUTLINE OF THE
RESEARCH in which the present writer summarizes
the
history and problems of the project.
The linguistic section contains two
papers by the present writer and two by Dr. Lew R. Micklesen.
The
first paper, entitled MACHINE LANGUAGE TRANSLATION, outlines the history of MT
and its linguistic problems and
is accompanied by a comprehensive bibliography of MT-linguistic publications.
The second paper, entitled MT
LINGUISTICS AND MT LEXICOGRAPHY AT THE UNIVERSITY OF WASHINGTON, explains the
inadequacies of pre-MT linguis-
tics and describes our approach to and solution of the problem of idioms,
especially our most recent research
results in this field. The third paper, which carries the title LEXICOGRAPHY,
discusses four different word
counts conducted on our 31,403-word corpus. The fourth paper, entitled
MT-OPERATIONAL ANALYSIS deals with the
syntactic behavior of a representative number of
Russian verbs and its utilization for the resolution of
grammatical
and nongrammatical ambiguities. In the APPENDIX to
this section we have placed THE SIMULATED
MACHINE
TRANSLATIONS of the 111 selected Russian scientific texts. The latter were
included in our previous
report. These predictions of the expected automatic output have been worked out
on the basis of our Russian-
English
MT-operational lexicon.
The second section is introduced by a
paper by Dr. David L. Johnson, entitled THE DIGITAL COMPUTER IN
MACHINE TRANSLATION. The second paper, by Dr. Robert E. Wall, Jr., and entitled
THE DIGITAL DATA-PROCESSING
PROBLEM
OF MACHINE TRANSLATION OF RUSSIAN TO ENGLISH, presents a preliminary design of
a special purpose dig-
ital computer for language translation and describes three rounds of
grammatical processing. The third paper,
by Aristotelis D. Stathacopoulos
and entitled A POSSIBLE APPLICATION OF ELECTRONIC COMPUTERS TO THE BLOCK AN-
ALYSIS OF GREEK SENTENCES, has been included because the linguistic problems of
the Greek language are similar
to those of, and the suggested solutions are applicable to, Russian. The
section is accompanied by two appen-
dices.
The first, the reproduction of a published paper by Mr. Philipp M. Pahl and Dr. David L. Johnson, is en-
titled
PATTERN RECOGNITION IN AN ELECTRONIC READER. The second paper is entitled
LOGICAL PROGRAMMING RESEARCH
IN THE UNIVERSITY OF WASHINGTON MACHINE TRANSLATION PROJECT and has Drs. Micklesen and Wall as its authors.
ERWIN REIFLER
Project
Director
March 25, 1960
1 Linguistic and
Engineering Studies in the Automatic Translation of
Scientific Russian into English.