A famous linguist gave lots of talks, talked almost all day long. And towards the end of the day he was talking about the relationships between linguistics and polythics---no, it is not the most famous linguist, it was another one. And he ended up at the end of the day sitting on a table in front of an audience of about 300 people, swinging his leg, talking about polythics and linguistics and all it really turned out to be /his/ polythics and no linguistics. 

I was sort of insulted and I thought 
I would never do that sort of... 

take the fact that I reach to certain agent 

traffic 
permitted just to
stand and say whatever comes into my head

but then I was called

depan to anai to

certain special event

it was a long way away so 
come

nobody will notice
let me try it and see what happens, but then I was actually encouraged by somebody, who probably wants to remain nameless. 

And so, I'm going to try, but I won't sit on the table and swing my legs. I will stand up here so that When I get really tired you will know and force me to sit down.

I'm going to sort of base it on my history on this field, but I hope I will be able to say and I 
have 
no little interesting thing 
appart of just all 
paisley 
that I have taken to get where I am. 

I'll make some nasty remarks about if you people 
stop by that.

I was walking through my college one day and I saw an announcement on a bulletin board saying that the department of engineering, or rather be, the engineering society, the ungraduated engineering society gave a prize of 25 pounds every year to the undergraduate judged to have given the best paper to the society. And I was always happy to collect 25 pounds or even any other smaller sum if I could find the way of doing it. On the other hand, I was studying medieval languages and I thought my chance of getting a 25 pounds prize out of the engineering society was rather low. So I thought one thing I could leave out with any possibility was a serious paper. But a humorous paper might stand some kind of a chance. So I thought what is the most sort of absurd topic that one may talk about and pretend that it was serious engineering and that had something to do with my field? And the answer was obvious: the idea was talking about something really clearly completely outrageous: a translating machine.

So I wrote a paper on this and I gave it to the society. Oh well, first of all I had to give a roll in the engineering society, like
colleagues ...
other events where I could go to talks about stuff that I had never even heard of in my life, and I thought I'll never go to any of those. But I did give a talk myself and I did get 25 pounds out of it. By the way, two people sitting in the audience came to see me afterwards and they were really annoyed by the talk that I had given, because they said they had been working on machine translation for the American Airforce for 4 years in Cambridge, and dammit it was no laughing matter! It had to be taken really serious. And if I doubted this, I should come by and see them sometime and they would explain all to me. Well I had exams to take and the year was moving on and I was not going to go and listen somebody talk about 
selling nuts like machine translation. So I finished up what I had to do and then I ran across this _______ before I was about to leave Cambridge, had taken all exams and everything but I decided to go to see them anyway. And at the end of that afternoon I had a job, which was something I had not had at the beginning of the afternoon so I was really rather happy with the upcome of it all. And I were there, working for this Cambridge Language Research Union and I was there for a couple of years working on some ideas for machine translation, based on what was called The Thesaurus (capital T, capital T). That was originally Roget's Thesaurus, but everybody realized that a better one would be nice if you had one. And the basic idea was---they didn't have any idea of building a whole machine translation system, but one idea they had was how to solve one of the lexical ambiguity problems that come along in machine translation. It had been a central part of that project for 4 years.

That idea consisted on

=== MOV00035.MPG 0:00
That you would let the words in a sentence, or the words in a text or the words in whatever happens to be, decide what they ment by sort of voting. If the words belonged to a economic text because most of the words simply are about economics or about this piece of resource over here and some this piece over there. Then they will vote on that as if you simply toss out the number of times it falls on this part of resource versus that one. And if more of them fell into this one, then you would excerpt meanings for all of the words that come along closest to that part of resource, a little bit as we have been just listening, actually. You will get a similarity measure amongst these things. And well, I  did that for another couple of years until I was hanging around the language Unit sometime in april, around easter. Everybody else in the place was up at easter being a good christian somewhere except me. Then suddenly somebody was at the door, from the RAND corporation in Santa Monica, California and said that he had been hoping to find oud what was going out at the Cambridge Language Research Unit and I said wow, I only have been here for a couple of years and I am not really quite sure by myself, but I'll do what I can since you are here, and I did that and he said why don't you come and spend 6 months in Santa Monica and you can tell us what you've been doing here, (that was hardly worth that money) and We can tell you what we are doing there. I thought that would be worth some real money, so I went to Santa Monica, California and 

found also how wonderful things happen. I discovered, for example, what they said about America was, or at least was then, true mainly was the land of opportunity. Where I came from it was a good idea to get a degree, it was a good idea to get a false degree, all of these things were good ideas, but the main way you would find success in life was to wait for the next personal relative to pull off. Then you moved up into his place if you were lucky. 

So I went from a computer which had 124 memory locations, to one that had 32768 just by crossing the atlantic and I still used punchcards for quite a long time, but not the way we had used them in Cambridge. In Cambridge we used punchcards with all the computer while in the States we used punchcards simply as a mean of getting things into the computer. It seems like the state of the things move forward.

I had spent two years working with Margaret Masterman, that is a name you may run across in this field, is a legend unfortunately no longer with us. Also one of the founder members of something called the epiphany philosophers. They were philosophers that were members of the philosophy department of Cambridge, Manchester, and other people... They were interested in the reconciliation of christianity and science; find the way these things concelliate one with another. 
...
to the extent
people were unable to reput you 
but it seems to me that it is a ____ problem
generate
not me 
...

=== MOV00035.MPG 4:30
My second employer was David Hayes from Rand Corporation, unfortunately also not longer with us. He was, I think, probably one of the people to whom we owe the existence of the field of computational linguistics. 
=== MOV00035.MPG 5:03
He invented the name. I remember being in his office one afternoon when he said let's call it "computational linguistics" and we did it and it's still called that way today. We needed a name badly, because, you see, we've been doing computational linguistics quite a while before it had a name. You can do that, you can do it without having a name for it, it was machine translation research. And then in 1964, a key period for that field, there was published a little black book. And I still believe we've actually written a stage in history I never thought we would reach there where people in this field who don't know immediatly when I mention the little black book what it must be I am talking about. It was called the ALPAC report, or the report of the Automatic Language Processing Advisory Comitee of the National Academy of Sciences who were called upon to looking to whether the goverment had been doing a good sign to support all these work in machine translation. Why did they have started in the first place? well, because the russians just put this sputnik into orbit and so there was a satellite going around the world all the time, something I recall of we thought we the americans (I shouldn't say "we") we should have done before they did. It was saw to be a rather something over inflection on american science that we blocked that before they did. So we needed to find out, rather urgently, whether russian science would be better than american science. And if it turned out to be will directed by this matter. The trouble with this turned out to be to require to read russian journals on things like physics and engineering and when we look into this, it turned out to be that russian journals have a great tendency to be written in russian and nobody could read russian, and as you probably know, requiring people in the United States to learn foreign languages is known in the constitution as cruel and unusual punishment. So they were never called upon to do it. Instead it was thought that we ought understand them using machines, and so people started working in all sort of places including Cambridge, England for the American Goverment on trying to design machine translating systems. All this happened since 1958, when the sputnik was launched. By 1964, it was felt that there wasn't a very urgent matter and you didn't had to worry too much about what the product was taken in modern area. By 1964 the sputnik hadn't been used for anything strictly notorious. There was a general feeling that the russians were doing as quiet as much damage the goverment expected. So they decided to look into wether this money on machine translation was well spent and unfortunately the answer was n"no". So, instead of working on machine translation as an engineering enterprise what got to happen was that we should try to lay down more solid scientific basis for this kind of work and may come back to engineering when we had where to place it all. Translating into scientific language, that means that if the goverment is going to science instead of engineering, let's move into science. So, if we all decided to move into science, we needed a field to be at. And, in order to be in a field, the main thing you need is a name. So, the name "computational linguistic" was invented. 
=== MOV00035.MPG 10:00
It was not the only one that was proposed, a name I can remember was "machinery linguistics". Nice not to have that name... Computational linguistics was a name proposed by David Hayes, and we stuck with that. David Hayes not only came up with a name, he also came out with the society of machine translation and computational linguistics which later became into the society of computational linguistics and he founded it also. He was the first president, or the second president. He also founded something called the international comitee on computational linguistics which has been organizing conferences on computational linguistics every two years. They later came to be known as COLING conferences, called sometimes "COOLING conferences" by people that find COLING not sweet. 
=== MOV00035.MPG 11:30

So I went to the RAND corporation for my 6th year and I did what many people do who go to the united states: I fulfilled the requirement that I stayed for 6 years, I didn't fulfilled the requirement to go back afterwards. So I stuck around for a while, went to Berkeley, and worked with Sid Lamper a little bit. Then I wind back to RAND corporation. I stayed there for another 10 years. The most crucial event that ever happened was the first piece of work, the first contribution that was purely a piece of computational linguistics, it could be interpreted as nothing but a piece of computational linguistics. It was an algorithm designed for no other purpose than to fulfill the requirement that come from computational linguistics was devised by Michael John Cop, a research fellow from the IBM corporation. John Cop came into the field of computational linguistics one wednesday afternoon around noon time, 4 o'clock in the afternoon. He changed funda