dth=100%>Yahoo! is not affiliated with the authors of this page or responsible for its content.
PANEL: THE PIONEERS OF MT PANEL: THE PIONEERS OF MT
Muriel Vasconcellos, Moderator

Present at Summit VI: Victor H. Yngve, Igor Meluk, Alexander Zholkovsky, Peter Toma
Christine Montgomery, A.F.R. Brown, Roger Heller, Loll Rolling

Additional contributors: Andrew D. Booth and Kathleen H.V. Booth,
Michael Zarechnak, Winfred P. Lehmann

The idea of using the computer to translate between human languages was first committed to paper half
a century ago. We celebrate the 50th anniversary of machine translation by bringing together a number
of its pioneers to reminisce about initiatives, early progress, and significant milestones in those first
decades. It is a unique and very special opportunity to have so many MT pioneers together in the same
room after such a long time. Indeed, several of the panel participants have not seen one another in more
than 30 years. Four of our invitees were unable to be present but have contributed written accounts to
help complete the story: Andrew and Kathleen Booth, Michael Zarechnak, and Winfred Lehmann. Four
others have preferred to make oral contributions only: Igor Meluk, Alexander Zholkovsky, and
Anthony Brown. The session is being audiotaped and videotaped for archival purposes.

The contributions below are arranged in chronological order of the authors entry on the MT scene.
Reading them sequentially, one gains an interesting picture of what the early days were like. We are
fortunate to be able to start out with an account by Andrew and Kathleen Booth that begins with the very
first musings about MT in 1946. The story unfolds as we follow the various contributors, ending with
an account by Loll Rolling of the introduction of MT at the European Commission, unarguably the largest
MT user in the world.





THE ORIGINS OF MT
Andrew D. Booth and Kathleen H.V. Booth
Autonetics Research Associates
Sooke, BC V0S 1N0 Canada

abooth@ios.bc.ca
Tel: (250) 642 5352

1. Origins
In his 1666 dissertation, The Art of Combinations, Leibniz had clearly indicated in his discussion of the
possible mechanization of both arithmetic and thought processes, that his logical processor could be used
to transform one language into another. He also considered the extraction of ideas from text and their
expression in terms of a metalanguage (Saw 1954). The culmination occurred in his design of the first
practicable mechanical multiplying machine (1694) and in a binary multiplier (Eriksson et al. 1996).

24 Another early claim, by the Russians, was to the effect that a Soviet engineer had produced a translating
machine in the 1930s. This appears unlikely because MT depend heavily on a large and complex
dictionarysomething which was not even on the horizon at that time.

We date realistic possibilities starting with two meetings held in 1946. The first was between Warren
Weaver, Director of the Natural Sciences Division of the Rockefeller Foundation, and Norbert Wiener.
The second was between Weaver and A.D. Booth in that same year. The Weaver-Wiener discussion
centered on the extensive code-breaking activities carried out during the World War II. The argument ran
as follows: decryption is simply the conversion of one set of wordsthe codeinto a second set, the
message. The discussion between Weaver and A.D. Booth on June 20, 1946, in New York identified the
fact that the code-breaking process in no way resembled language translation because it was known a
priori that the decrypting process must result in a unique output. The main purpose of this meeting,
however, was to interest the Rockefeller Foundation in supporting development of an electronic computer
at the University of London.

Things remained thus until the end of a visit to the Institute for Advanced Study, Princeton, in 1947. On
May 20 of that year another talk with Weaver showed that the Foundation was unlikely to fund a machine
at London for numerical work but that a submission indicating interest in MT would be well received.
This was duly submitted and was successful. The core of the proposal involved the use of the high-speed
storage organ as an electronic dictionary.

2. London,
1948-1962
On our return to London, work was at once started on two fronts: ADB designing and supervising the
construction of the new computer and KHVB directing the development of suitable programs. It must be
noted that in those days all coding was done in machine language (not even "assembler") and had to be
inserted via switches

Some code for a computerized dictionary was tested and found to be very effective. The lookup method
used from the first was binary partitioninga technique which we regarded as so obvious that we did not
deem it worth publication until we discovered that several major computer groups were using sequential
search strategies (A.D. Booth 1955). More extensive descriptions of our work are given in Automatic
Digital Calculators (A.D. Booth & K.H.V. Booth 1953) and in Programming for an Automatic Digital
Computer (K.H.V. Booth 1958). The latter contains a complete MT program of that era.

Other activities of our group included work on the translation of English to Braille by J.P. Cleave
(1956), language processing in general by M. Levison (1962), and the chronology of the Plato dialogues
by L. Brandwood. Much of the work is described in Mechanical Resolution of Linguistic Problems by
A.D. Booth, L. Brandwood, and J.P. Cleave (1958). Our thanks are due the Nuffield Foundation, which
provided major funding during this period.

3. Saskatchewan,
1962-1972
In 1962 we migrated to Canada and in due course became involved with the Queen's Printer in a most
interesting project. One problem in Canada is the presence of two official languages: English and
French. The parliamentary system requires that the Hansard, the daily transcript of the day's debates, be
available in both languages the next day. This was the problem which we were asked to address. Our own
view was, and still is, that output of good quality can be produced only with a post-editor. With the aid
of linguists Corinne Brown and Gloria Geitz and programmer Charles Stock (K.H.V. Booth & Stock

25 1971) we produced such a program and demonstrated it at the National Research Council Laboratories
in Ottawa in 1971. It is interesting to note that the only languages available on our University mainframe
were FORTRAN and COBOL. The program was written in the latter.

4. Lakehead
University,
1972-1978
When we migrated to Ontario in 1972, our work on MT was continued with emphasis on the construction
of a large dictionary and the use of statistical techniques in linguistic analysis (A.D. Booth et al.).

5. Retirement, 1978-
Our work on MT terminated with our retirement in 1978 except for one small but interesting experiment.
In 1979 the first personal computers became available and we were involved in a project to develop an
energy management package. The machine chosen was a Commodore PET 2000 and, having some time
to spare, ADB decided to write an MT program for this machine. Here the language was BASIC and a
French-to-English program was soon produced and demonstrated to the Canadian National Research
Council in 1980. The program incorporated most of the methods devised in our earlier work but, because
of storage limitations, it had only a small dictionary. The output quality was equal to or better than that
generated by currently available inexpensive translation programs. The main superiority of our program
lay in the use of our category count technique (K.H.V. Booth 1970) for the resolution of such ambiguities
as Chomskys The pen is in the box.

R
EFERENCES

Booth, A.D. 1955. On the Use of a Computing Machine as a Mechanical Dictionary. Nature, 176:565.
Booth, A.D., and Booth, K.H.V. 1953. Automatic Digital Calculators. 1st ed. London: Butterworths.

Booth, A.D., Brandwood, L., and Cleave, J.P. 1958. Mechanical Resolution of Linguistic Problems.
London: Butterworths.

Booth, K.H.V. 1958. Programming for an Automatic Digital Calculator. London: Butterworths.
Booth, K.H.V. 1960. Statistical Parsing by Computer. Pensiero e Linguaggio 1:107.

Booth, K.H.V., and Stock, C. 1971. English-French Translation on a Computer. Pensiero e Linguaggio,

2:231.

Cleave, J.P. 1956. The Application of an Automatic Digital Computing Machine to Problems of Data
Transformation and Language. Ph.D. thesis, University of London.

Eriksson, K., Estep, D., Hansbo, P., and Johnson, C. 1996. Computational Differential Equations.
Cambridge: Cambridge University Press. p. 20.

Levison, M. 1962. The Application of a Computing Machine to Linguistic Problems. Ph.D. thesis,
University of London.

Saw, Ruth Lydia. 1954. Leibniz. London: Penguin Books. p. 214.

26
SOMETHING OLD AND SOMETHING NEW
Victor H. Yngve

The University of Chicago

Home: 28 Crest Drive Dune Acres

Chesterton, IN 46304

v-yngve@chicago.edu

Tel: (219) 787-8340

As part of my research assistantship as a graduate student in physics at the University of Chicago in the
late 1940s, I was involved in preparing instrumentation for detecting incoming cosmic radiation in the
stratosphere. The electronic circuits we were using were similar to circuits found in the literature on the
new auto