Interview with Robert (Bob) Kowalski

Robert Anthony Kowalski was born in Bridgeport, Connecticut, in 1941. His mother, who wanted some financial independence from her husband worked in one of the many factories in Bridgeport, while his father ran his own house renovation business. Bob and his two brothers were recruited to work with their father, which he says caused some tension, especially between himself and his father, who had very different goals and ways of thinking. One of his brothers remained in the family business while his other brother became an academic and worked in fluid dynamics.

Bob thought he was the first in the family to go on to higher education as neither of his parents had completed high school, until he discovered that his uncle in Poland, was a professor; he says: “I guess, given the opportunity, working-class origins can still lead to higher academic achievement.” The family however, valued education and encouraged their children to pursue it.

As a relatively shy boy, Bob says he did not play with many other children, but mostly because he was fascinated with learning about the world around him.

Bob’s primary school was a Polish parish school, run by Franciscan nuns where Bob was an altar boy and a patrol boy, helping other children to cross the street. He says that for a while he wanted to be a priest

He then went to a Jesuit all-boys high school where his favourite subject was Latin, but only because he was encouraged by his teacher to do extra work after school and train for Latin Sight Translation contests.  He and the team did very well, and he says the experience “made me feel that I was perhaps destined for higher academic achievement.  So that had a huge impact on me.”

However, at the same time, Bob started to feel that the education he was receiving was excluding other learning experiences; he started to read more widely than the courses he was given and adds: “After a while I began to resent that our education was so limited.”

University of Chicago

After high school, Bob went to the University of Chicago to study mathematics, it was while he was here that he first encountered logic.  He explains: “I failed the English placement course, so, I was forced to take a remedial course. I finished the year with a D in English writing skills, but an A, maybe even an A+, in mathematics, which had featured a small element of propositional logic.” This introduction to logic suggested to him that logic might be the key to understanding what was wrong with his English, and that it might also be the key to understanding the problems he experienced when reading Joad’s Guide to Philosophy, during his extracurricular reading in high school. He says “ The book exposed me to various philosophies which were very intriguing but equally puzzling, because they all made sense, and yet they all contradicted one another. I was a very troubled young man at that time, and I got the feeling that logic would help to find the ultimate way to truth in these matters.”

University of Bridgeport

In 1959, at the start of his second year, Bob chose to leave the University of Chicago for personal reasons. He spent the rest of the year working in a chemical factory before deciding to start studying again, this time at the University of Bridgeport. He says: “I was able to be myself at the University of Bridgeport and to work academically in a fulfilling way.  So, you don’t always have to go to the highest academic institution in order to receive a good education.”

Having spent time independently studying logic at Bridgeport, Bob was encouraged to pursue his interest by some of the professors at Yale University in New Haven, not far from Bridgeport, whom he spoke to and who suggested he should apply to Stanford, Berkeley, and Wisconsin.

Bob went to Stanford in 1963, and in 1964 he participated in an exchange programme with the University of Warsaw. Poland and Warsaw in particular were renowned for their work in mathematical logic. Bob says: “It was exciting for truly academic reasons, but at the same time, it meant that I would be able to meet my grandparents and uncles in Poland, who I had never met before, and learn a bit of Polish, which I had never managed to do as a child.  So, that was a very exciting year.” It was during that year that Bob met his wife. He adds: “Unfortunately I didn’t pay as much attention to my studies as I might have, because I ended up meeting my wife who was quite distracting, and is still rather distracting I have to say.” The couple married in Poland after just two months; they are still together today and have three daughters.

As well as getting married, Bob found that his trip to Poland influenced him upon his return to the US. He started to become very active in the protest against the war in Vietnam.  He adds: “I was not at the forefront of activities, but I was quite active behind the scenes. I dreamed up peaceful ways of protesting against the war. The scheme I helped to develop was one which involved bombing the streets of California with leaflets explaining what was happening in Vietnam.” Bob ended up resigning from the protest movement after becoming very distressed by the reactions from supporters of the war. He says: “I resigned from the protest movement simply because of the inability to deal with the kind of response that it evoked in those who felt threatened by protest against the war. I ended up leaving Stanford, abandoning my PhD, because I felt mathematical logic was not logic, and mathematical logic didn’t address the real human problems that were taking place at the time. I ended up for a year working in Puerto Rico.

In 1967, after nine months in Puerto Rico and the conviction that he would need a higher degree, Bob was accepted to study for a PhD at Edinburgh University at the ‘Metamathematics Unit, Department of Computer Science’.  He says: “When I saw “Computer Science”, I thought to myself, this is not what I had been expecting; and my heart dropped.  The last thing I wanted to do at any point in my life was to do computer science.”

Bob’s supervisor was Bernard Meltzer, an electronic engineer with an interest in artificial intelligence and automated theorem proving. In addition, Alan Robinson, who had developed the resolution principle for automated theorem-proving, was on sabbatical in Edinburgh. Bob adds: “Much of my work was inspired by the work that he had just done so I was able to dive directly into a PhD at the cutting edge of that particular subject.”

From his experience of the mathematical logic that he had studied at Stanford, Bob thought that automating mathematical logic didn’t make much sense. He adds: “I thought it was a crazy idea.  I didn’t think it was a goal that was particularly laudable, but nonetheless I thought I could give it a shot, but I didn’t think that the technical approach that was being proposed could really be taken very seriously compared with what I had learnt at Stanford.” However, after a couple of months, he realised that he was wrong and got into the swing of things. He completed his PhD in two years and during that period, met and worked with fellow PhD student Pat Hayes. Together they published a joint paper. Bob adds: “That was a great time really, despite the fact that I hated computer science.”

During his PhD, researchers at MIT were “attacking resolution” as Bob describes it.  They thought that logic alone would never solve any interesting problems and that what was needed instead was a controlled kind of logic, a procedural representation of knowledge. Having discovered this train of thought, Pat Hayes, who together with Bob was over 100 pages into writing a book on the subject, advised Bob to drop the book.  Bob says: “This was quite a difficult proposal to be put before me. It was obviously a huge challenge; how could resolution be so good and now be so bad, so much so that we had to abandon 100 pages of a book that we had been writing?” Bob explains that for the next two years, he and Pat, independently “tortured” themselves with trying to understand how procedural representations of knowledge could be good, and logic representations could be good, but both not be possible at the same time. He adds: “And, so, the development of logic programming which eventually came about, came about through an attempt to reconcile these two opposites really.  Pat had his own way of thinking about the reconciliation, and I began to develop my own way of thinking about that reconciliation.”

After completing his PhD in 1970, Bob applied for various academic positions, but was unsuccessful, and ended up staying at Edinburgh for a further four years on a research grant working on further research based on his PhD project plus other ideas that he says had “been troubling him since his first year at Chicago”. As a result, his work came to the attention of Alain Colmerauer, who had been developing a question and answering system and was beginning to use resolution systems of the kind Alan Robinson had developed. In 1971 Alain invited Bob to visit him in Marseille.  Bob says: “During that visit, we spent four days and nights in an almost marathon meeting during which we had wonderful discussions, and that resulted within the next year or so in the Prolog language that Alain developed, and the contributions to logic programming that I worked on and developed.”

In 1975, Bob was invited to apply for a role as a Reader in Theory of Computing at Imperial College, London by Manny Lehman, who was in charge of the Computing section of the Department of Computing and Control. Bob explains: “I applied, I was interviewed, and there was an extended period of time during which objections were raised to my lack of computer science credentials, which were very legitimate objections I have to agree with, but nonetheless I succeeded. I also had to teach courses in computing, in subjects I had never studied.  So that was quite a challenge.”

While at Imperial, Bob contributed to the development of a micro version of Prolog (Micro Prolog) which was used in the school environment.  In 1978 Bob started a course of logic lessons for twelve-year-old children in his daughter’s middle school. He explains: “It did seem to me that logic was something that children could understand and would benefit from understanding.  I started to develop examples of logic that would be understandable to children.  Little stories about dragons.  For example, a dragon is happy if all its children can fly, green dragons can fly.  Things of that nature. I talked with some of the teachers in the school, and we agreed that it might be of interest to try some of these lessons out using Prolog in school”.   The work went well and was given research funding to be continued and to be expanded to other schools.

In 1981, Bob found himself involved in both the Japanese Fifth Generation project announced by MITI and the British Alvey Programme, which was developed in response to the Japanese project.

He explains: “The Alvey Programme was the British response to the Japanese Fifth Generation project.  The Japanese Fifth Generation project, which was announced in 1981, had three main goals, one of which was to develop logic programming as the software for a highly parallel computer focused on AI, artificial intelligence, applications.” The Japanese had approached the British, French and German governments for collaboration, but they all refused. Bob continues: “The British government in particular set up a competing national research objective, without taking the Japanese proposal to work on logic programming very seriously.  So that created a huge difficulty for our logic programming group, because we could see that it was being used as the basis of the Fifth Generation project in Japan but being ignored in Britain. I was the most senior academic in the logic programming field in Britain, and the burden fell on me to argue the case for logic programming.”

He continues: “For purely technological reasons, there were good reasons why the British government wouldn’t want to collaborate with Japan.  Why didn’t they choose logic programming?  All the established academics who would be consulted, would have favoured other technologies, other approaches to computing than logic programming. That meant both that our own work was not promoted within the Alvey Programme to begin with, and that I in particular had to fight quite hard to try to convince the powers that be that it was at least worth a certain amount of support, along with the other support that might be given to other technologies in the UK.”

In 1982 Bob was promoted to professorship at Imperial, and he found himself having to make decisions about whether to support the Japanese project by accepting requests from Japanese institutions to visit the logic programming group. He adds: “So for a total of twelve years, during that period I had quite a bit of interaction with the Japanese researchers as well as with the European Research Computer Centre (ECRC).” The ECRC, headed by Hervé Gallaire, was set up in Munich by three computer companies, ICL in Britain, Bull in France, and Siemens in Germany, as a European counterpart of the Fifth Generation project..

Towards the end of the eighties, Bob was also involved with the European Basic Research Action, called Compulog. Bob explains: “ It involved the thirteen most active centres doing work in logic programming in Europe, two of which were in Britain, Edinburgh and ourselves at Imperial College in London, and then others throughout Europe, including Marseille and, several in Italy. I coordinated that project.”

In 1990, Fujitsu supported a five-year project at Imperial focused on abductive logic programming, which allowed Bob to work full-time on research.

Over the course of his career, Bob has worked with numerous PhD students who have gone on to contribute to the field, including David Warren, whom Bob credits with being responsible for many of the fine details of the Prolog language.  Keith Clark, of whom Bob says: “He probably taught me more about computer science than I taught him.  He was the guy who helped me to teach my first courses when I came to London.  He was at Queen Mary College, and he had already written a book about the course I was supposed to be teaching, so he was a great help.”  Marek Sergot who together with Bob developed the application of logic programming to law, legal reasoning, and to reasoning about time and causality.  Bob adds: “I suppose, if you ask what achievements I am proud of, I would say, obviously my work on logic programming, but also my work with Marek on the application of logic programming to law, and to causal reasoning to develop what’s called an event calculus, which has proved to be very influential in knowledge representation.”  He also points to his work with Fariba Sadri, who did much of the Prolog representation of the British Nationality Act, which was Bob and Marek’s main application of logic programming to legal reasoning. He also singles out his work with Francesca Toni, who is very prominent in the field of argumentation, which has now become an important area of artificial intelligence.

In 1997 Bob became head of the Department of Computing at Imperial College.  Bob says that at the time the department was in turmoil due, in part to, the Alvey Programme, which exacerbated competition between different groups with competing technologies, all seeking support at the expense of other groups. To try to solve the issues, the department had established an advisory committee to the head of department, of which Bob was a part. This did not solve the problems, the head of the department decided to resign, and Bob was invited to take over. He says: “I thought I would give it a try and see whether I couldn’t bring a bit of logic to reconcile the differences in our department.” When the logic did not succeed, Bob decided to retire from the role and from Imperial.

Bob adds: “The main reason for retiring was that I really wanted to be a logician not a computer scientist.  I wanted to be a logician who would help other people to think more clearly, and to express themselves more clearly, and to be more effective in their lives, with or without the aid of computers.”

In retirement Bob has run writer’s workshops with PhD students, he explains: “The idea was that the PhD students at Imperial, and later in Japan, would present the abstract either of their PhD or some research paper, and the rest of us, most of whom were not experts in that field, would try to understand it.  We would try to see if those abstracts could not be rewritten in a manner that would make them intelligible.  … That was very rewarding.  Of all my teaching, that’s the teaching that was most enjoyable, if only because I learned as much as I taught, and that’s always a very good way to be doing education.”

Bob was also invited to be part of a WHO/UNICEF project to help deal with problems giving the official estimates of infant immunisation coverage throughout the countries participating in the United Nations. After considering many different approaches, the project group settled for a logic programming representation of the estimates, which is still being used today.

In 2011, Bob’s book; Computational Logic and Human Thinking: How to Be Artificially Intelligent, was published. He says it was the fruition of his early retirement and shows how advances in symbolic approaches to artificial intelligence are usable by human beings in order to enhance and improve their own intelligence.

In parallel with his other projects, Bob has also been working primarily with Fariba Sadri on a system which attempts to reconcile the logic programming way of thinking about both logic and computing with the production system model of both human thinking and computation. He says: “We’ve developing an approach which combines production-like rules with logic programming rules, but gives them a logic which is coherent, and which is such that the logic programming rules are like the beliefs of an intelligent agent, and the production rules are like the goals of an agent.  And, one of them alone, either leaves you with an agent which has beliefs but no goals, or goals but no beliefs.” Bob regards this work as the most important of his contributions to computer science.  ”

Bob says: “I’ve been quite successful with logic programming, the event calculus, and to some extent legal reasoning, as an application of logic programming. I think they were genuine achievements. I haven’t been as successful as I would have liked in having the work on computational logic be taken up outside of the computer community.”

In looking to the future, Bob believes that ten years is too short a period, during which time we will see only a natural evolution of where we are today, however, a 50-year view, can offer greater insight. He says: “If you look 50 years ahead, I think you could begin to question whether the kind of languages that we have in computing can continue to be so low-level and so machine-oriented, and if they’re going to be more human-oriented, they’re going to be very different, I think, from the way they look today. I think the challenge will be for computer technologists who love machines to somehow think more about people and make their languages more congenial to people, more intelligible to people.  I see no reason why computer programs should require specialised education to be understood. They may be difficult to write 50 years from now, but they shouldn’t be so impossible to understand by stakeholders who are not computer scientists. I think that’s a challenge that will be not a technological challenge, it’ll be a psychological challenge.”

On the future of education, Bob says: “I think that, even today, formal lecture courses are missing some of the most important skills that people need to learn in order to think clearly, in order to be able to transfer their abilities from one domain to another domain, in this world which is changing so rapidly. There are some skills which are obvious but are untaught. The ability to think and express yourself clearly is not something which is taught.  If it’s taught at all, it’s taught by osmosis.”

In offering advice to young people today, Bob reflects on what he has learned, he says: “I have found it very difficult, but at the same time very rewarding, to try to reconcile my social obligations with my own personal goals. My personal goals involve logic for human affairs, my social obligations have required me to teach computing.  I have worked very hard to be a good citizen and fulfil my social obligations, but at the same time not to sell my soul and abandon my personal beliefs. I think that that’s something I would recommend to others.

Bob is a Distinguished Research Fellow and Emeritus Professor of Computational Logic in the Department of Computing at Imperial College, London.

He was elected a Fellow of the Association for the Advancement of Artificial Intelligence in 1991, a Fellow of the European Association for Artificial Intelligence  in 1999, and a Fellow of the Association for Computing Machinery in 2000.  He received the IJCAI Award for Research Excellence in 2011 and the Japan Society for the Promotion of Science Award for Eminent Scientists for 2012-2014.

Interviewed By: Elisabetta Mori at BCS London on the 10th January 2020

Transcribed By: Susan Nicholls

Abstracted By: Lynda Feeley