As an organisation, the ACL has clearly evolved over the forty years since its foundation. (Whether, or how, it has evolved in its area of intellectual interest over the same period is a tricky and interesting question.) The ACL's organisational development in the period up to Don Walker's death in 1993 was, however, rather different from its evolution since. This was partly because the ACL's community was not very large in its first two decades, partly because there were only three Secretary-Treasurers in the whole period from 1963 to 1993. Don Walker's term of office, from 1976-1993, was a particularly important period in the development of the ACL, and both for these reasons and because Don had been in post for so long, the two years 1993-1994 - 1993 when Don was ill and died, and 1994 when I was President - were a rough transition time.
There were very many things to do and, more importantly, to figure out how to do. The day-to-day operation of the ACL had to be kept going, and at the same time, administrative and policy changes that Don had initiated had to be carried through. Some idea of the challenges that the Executive, and especially Fernando Pereira as President in 1993 and me as President in 1994, had to face can be gotten from the fact that the ACL office was in the basement of Don and Betty Walker's house in Morristown (NJ), and that much of the information about ACL procedures, resources, decisions and strategies was carried in Don's head. Betty, as an ACL employee, was very fortunately familiar with the manifold details of ACL's office-driven activities, notably of the way membership, conferences, and publications were handled. Without Betty's dedicated support, at a distressing and demanding time for her, the ACL's day-to-day operations during 1993 and 1994 would never have been kept going. The ACL owes a great debt to her, and we were glad to be able to be able to present the fine volume of papers on Current Issues in Computational Linguistics in Honour of Don Walker to her at the 1994 Conference.
The main, very pressing, tasks for the Executive and especially the two Presidents in 1993-4 were first, to create a new office; second, to find a new Secretary-Treasurer; third, to ensure that the 1994 Conference ran smoothly; and fourth, to determine the state of the ACL's finances and establish them on a firm footing. The financial situation was very unclear, and the manifest need to be as prudent as possible about money meant that everything else had to be treated with extra care.
Don had already begun the process of creating an independent ACL office with an Office Manager: the Association's growth in size and complexity was making running it out of his basement impractical; after further negotiations we were very fortunately able to hire Priscilla Rasmussen as Manager and she began the administrative takeover during 1994. The search for a new Secretary-Treasurer was a major effort. Just as for the Office Manager, a whole new job specification had to be written: it was clear not only that having an Office Manager would mean that the nature of the Secretary-Treasurer's would change, but that we could not ask anyone of suitable standing as an active researcher and/or teacher in the field to take on too burdensome a job. The search was very time-consuming, and the Executive was delighted when Kathy McKeown, who was already familiar as a past President with the ACL, accepted our proposal.
The ACL Conference is a major financial commitment, as well as a logistic one, and Don had always paid close attention to the conference detail in all its ramifications from catering to proceedings: as well as being significant as an event for the participating community, it was an important showcase for the ACL as an organisation, so its image and style mattered. Jan Wiebe did admirably in the local arrangements at Las Cruces, but dealing with everything that the conference involved, which included, for example, preparations for concomitant meetings of the Executive, was a major effort for all the Officers and Executive: one consequence of my having to monitor absolutely everything in 1994 for its potential financial implications was being asked whether a conference participant could bring their spouse and 10 year old child to the banquet.
But the most serious task for the Executive, and especially for me, after Don's death, was understanding the ACL's finances. Though I had had one meeting with Don about them, finding out what was involved in detail after his death necessarily took months, which constrained current operations and made planning for the necessarily different future very difficult. It was clearly essential to do a very careful financial analysis, and to replace the invisible organisation carried in Don's head with a much more explicit structure with which the new Office Manager and Secretary-Treasurer could work. Membership fees were a major income source, but raising fees is not popular and does not bring money in quickly. In fact, the Executive was pleased to find it confirmed in 1994 that, while not rich, the ACL was perfectly viable financially.
There were very many other things to attend to during this period, like clearing up Don's files, deciding what to do about storage containers full of old Proceedings, and negotiating publishing contracts, to take three random examples. All the members of the Executive were assigned tasks ad hoc, for example Ed Hovy that of gathering all the annual reports to be presented to the Executive and General Meeting, and those with past experience, like Sandee Carberry of running conferences, were drafted in as mentors: the ACL could not have survived without all their work, and without that of Judith Klavans, the Associate Secretary-Treasurer. At the time when Don became Secretary-Treasurer, the ACL was a modest organisation, without all the apparatus of full-scale conference proceedings, chapters and SIGs, or a substantial journal, and without a significant international presence or world-wide membership, that it had achieved by 1993. He had fostered the ACL so that it could profit from the growth of the field, until it had reached the point where, as he himself had seen, a new adminstrative structure was needed.
The agenda for the Executive meeting in 1994 illustrates, very nicely, this growth in scale and range of the ACL's activities, with nearly forty items including the journal Editor's submission and one on CL reviews, reports on different aspects of the 94 conference and plans for future ones, statements about five SIGs and about related organisations and meetings, items on electronic resources, e.g. cmp-lg, and about student affairs.
As this implies, any transition from the old structure to a new one would have been a substantial undertaking, and that which Don had envisaged would certainly have required much to be made explicit or formal, that had hitherto been implicit or informal, for example guidelines for conference workshops, or for creating SIGs. But in the circumstances of 1993-4, a great deal of this had to be done, perhaps more than otherwise, and very quickly: I had to assemble, or write, many documents, ranging from how to cost ACL conferences via revised ACL constitutions to emergency ACL Meeting resolutions. Equally, Don had initiated policy changes, for example, about holding ACL conferences outside North America, which had to be kept moving.
All of this meant that the transition period was pretty bumpy. But the fact that the ACL has advanced successfully since the mid nineties suggests that Don had laid a good foundation for the latest phase of the ACL's life, and that the many people who had to start building walls on this foundation in 1993-4 also did a good job. There is one other point worth noting about the time too. Email had become an ever more important administrative tool during Don's time as Secretary-Treasurer; and there was certainly no way that all of the things that needed to be done during 1993-4, with the ACL's first non-North American President and other people spread around the world, could have been done without email.
Looking more broadly than at the 93-94 transition period, what has been happening in the last decade since, and what does ACL's future look like?
In the last decade, ACL has restructured itself formally and become more clearly an international organisation. Don worked towards this, and the formal reorganisation belatedly reflects the fact that CL and NLP have been international fields since the beginning (with significant early work in the USSR for example), and that activity is growing worldwide and can be expected to go on growing. In Europe, Japan and, increasingly, China people are able to build, to good effect, on long experience of the field.
The last decade has also seen Language Engineering come of age, so we now have three overlapping areas, CL, NLP and LE. Practical concerns, in the shape of MT, were begetters of the field as a whole, and recent developments have supplied much better tools for applications than before. We can expect the practical trend to continue, and not all as a one-way flow from CL through NLP to LE: it is arguable that shallow parsing, especially finite-state, has pushed back the other way, and tasks like summarising and question answering are stimulating new attacks on discourse structure and on sentence interpretation. It is much easier now than in the past to pull together the resources for experiments and prototypes: this will certainly continue, and for more complex functions and challenging tasks.
The third feature of the past decade has again been a revival of older ideas, on statistically-based language processing using corpus data, but now with volumes of material beyond the sixties' wildest dreams. With the benefit of lessons from speech recognition and from information processing, this has proved such a rewarding seam to mine that the digging will certainly continue and, when worked on by those with good ideas about the kind of jewellery one should aim at, will deliver some pretty impressive parures.
Speech processing and language processing have been coming together in the last decade, notably in dialogue work, though ACL still has some way to go to be as natural a home for speech researchers as for language ones. But the convergence is bound to get better in future. At the same time, research in the field has continued to grow away from linguistics as she is spoke in linguistics departments, and from artificial intelligence. This is a natural consequence of technical specialisation, though it would be a pity if we found ourselves cobbling up inferior versions of perfectly good wheels that those in linguistics or AI have already built.
The ACL has responded, as its conferences and journal show, to the trends of the last decade. Developments in IT have transformed research in the field. The growth of the Web has also moved the goal posts for applications. But though this has shaken up the game, it's not a completely different one, and we still need good models for it. The ACL has shown that it can successfully embrace both principles and practice, and can evolve as an organisation in keeping with changes in the zeitgeist: I trust it will continue to do so.