Originally an invited paper read to the FIPLV/Eurocentres seminar on: The use, application and limitations of new computer-based technology in foreign language learning, Goldsmiths College, University of London, London, September 1988. Subsequently read as a keynote address to the State Conference of the Computer Education Group of Queensland, 1989. Originally Published in On-Call, vol. 5, no. 4, 1991, pp. 2-8.

Introduction¹
In general, the notion of Computer-Aided Language Learning (CALL)² software has appeared to be non-problematic, at least if one examines such authoritative publications as CALICO, The Modern Language Journal, and System and the relatively small number of books which have appeared on the subject.

Yet it seems that if the potential of Computer-Aided or, more appropriately, Computer-Enhanced Language Learning³ is to be realised, then we have to look beyond the implicit definitions of CALL as inferred from reading these publications.

The discussion which follows is based principally upon experience at the University of Queensland, 4 Australia and at Bond University, also in Australia⁴.

This paper also focuses unashamedly and unrepentantly on problems of a fairly technical, systems-based nature. These problems are often forgotten or ignored but, unfortunately, they influence in a highly insidious manner our presuppositions about the nature and thus the development of CAL/ CALL.
 

CALL Software: Courseware
The first and most obvious category of CALL software is clearly courseware or computer-based lesson materials. (These have been described in various amounts of detail by most authors writing on the subject for it is their major preoccupation: what they see as the nitty gritty of CALL.)

One interesting recent description of CALL programs has been provided by John Higgins under the title of Types of Software (Higgins 1988).

Higgins specifies the following sorts of software:

1. "Do what 1 tell you": quizzes, tests, exercises, programmed learning.
2. "Guess what was there": text-reconstruction programs
3. "Can I help you?": programs to help the learner perform a job.
4. "How do we get out of this?": simulations, adventure games and other problem- solving games using language or stimulating language usage amongst learners

However, some additional points need to be made.

To begin with, there is some difficulty with the notion of the word computer. In particular, there is the popular assumption that a computer for the purpose of running CALL consists of a small single-user, single-tasking microprocessor-based machine, usually an IBM PC or clone running under the MSDOS operating system. On reading the literature, one is overwhelmed by the feeling that no other computer has ever existed or (ever will) although the Apple MACintosh is now beginning to make inroads into this state of affairs.

Inextricably linked to the definition of computer is the question of price. Education seems condemned to be poorly funded for the rest of eternity. Powerful computers, if teachers actually know of their existence, seem beyond reach. Thus, the notion of CALL software is limited even further through having to deal with the so-called realities of foreign language teaching and education generally. This is most unsatisfactory, and extraordinarily limiting, in a perspective where increasing pressure is being applied to the teaching professions to make their teaching more progressive, efficient, effective and accountable in a multitude of ways

That such limited notions of computing and CAL should have developed is hardly surprising in one sense as the people who make judgements about what to develop and how to go about doing so are essentially and primarily language teachers with little knowledge of the field of computing and who have been made aware of computing in a peripheral way through the incredible rise of the microprocessor chip and, especially, its use in the most popular microcomputers: first the APPLE II, then the BBC and now the IBM PC and the MACintosh. In other words, they are falling prey to commercial and other insidious pressures as well as to other popular but often unsound and limited myths as to the desirable shape of CALL.

The conceptualisation and development of CALL is therefore constrained significantly by the fact that in many cases relatively well-trained language teachers are trying to develop sophisticated software in the face of their own hardware and software ignorance, little experience of any kind of CAL as a medium in any form and at the same time, they face extraordinarily limiting commercial and/or financial pressures.

In the best of circumstances, this has resulted in some of the more aware and innovative teachers and researchers battling against the odds, and (at times even their CALL colleagues), to develop methodologically sound courseware materials.

In the worst of circumstances, the methodological demands and the limitations of the computing resources have so disappointed teachers that some have discarded educational computing as a worthless enterprise. This disillusionment is reflected in a statement at the 1987 World AILA Congress to the effect that the best days of CALL had already been and gone (Martha Pennington' opening remarks at the third session of the International CALL Symposium).

In the first instance, therefore, CALL software can be defined as methodologically sound lessons or programs capable of running on IBM PC style systems.

However, this description seems much too constraining. If learners are to benefit fully from the advantages of a CALL environment, then it seems important to broaden the concept of CALL software.

This is the subject which the rest of this paper will attempt to investigate briefly as a stimulus to further discussion.
 

A Question of Control
One of the significant developments in language learning methodology and hence in CALL is the growing belief that language learning ought to occur increasingly in what this author has called elsewhere .a very open, negotiable and negotiated, highly task-oriented, resource-based learning network where people will progress at their own pace in a relatively asynchronous fashion." (Lian, 1991; cf also Lian, 1987). Under these circumstances learners can, and indeed, must be given control over the ways in which they can select and interact with learning materials of all kinds, including computer-based materials.

In other words, the trend is to allow learners to have control over the when, where, how and why of their activities. Hopefully, the institution in which they are studying will allow them to break away from the teacher-centred model and to decide with whom (e.g. alone or in groups) they wish to learn, at what time (e.g. day or night, before dinner, after breakfast or simply when the fancy takes them), in what environment (i.e. under what physical and psychological conditions), how (i.e. using materials of their choice, organised in ways that they think fit) and the purpose for which they choose to use these facilities (e.g. to solve a real-life problem to prepare for an exam, to feel good about themselves). The flexibility of such a system and its responsiveness to students' demands is very great indeed.

Now, If students are to be given genuine control over their learning environment, then at least three things must occur.
 

1. They should be helped to develop autonomy in their learning,
2. a large array of resources needs to be made available. Indeed, in its purest form, modem CALL could be conceived as a highly sophisticated form of resource-based learning; and
3. they must be given the means for controlling their learning.

 
Such control is not accessory, a luxury. It is an integral part of what CALL methodology ought to be all about. It is imperative, therefore, that CALL methodologists participate actively in the development of appropriate control infrastructures in collaboration with other system developers. Unfortunately, the need for such collaboration is all too often not recognised by CALL developers who are often torn between their ideal of what CALL courseware might be like, the financial and methodological realities of their institutional environment, their relatively limited experience of computers and computing and, occasionally, the desire to reserve control for themselves.
 

For instance, at Bond University and the University of Queensland, we have developed or are now seeking to develop applications in four different categories.
 
 

1. The computer as a teacher/manager of learning. e.g. tutorials of all kinds (from listening comprehension to grammatical drill and practice).
2. The computer as a resource. e.g. dictionaries, databases and expert systems (such as dialogue generators and spelling helpers) to solve problems relating to language tasks.
3. The computer as a tool. e.g. word processing, spelling checker, text analysis, concordance, desktop publishing.
4. The computer as an aid to communication. e.g. international electronic mail with other students, teleconferencing, and national and international access to systems listed above (for further information re all four categories, cf. Lian, 1985; Cryle & Lian, 1985; Lian, 1986; Lian, Thornquist & Thornquist 1987; Lian 1988a; Lian 1988b).

In the long term, the sensible trend must be toward a centralised system of information storage and management, leading to the development of distributed systems of local, national and even international networks.

In such a structure, each network node and its administrators can be responsible for the development, upkeep and maintenance of its materials. Costs will be kept down to a manageable level, responsibility will be shared and software will be kept up-to-date quickly and efficiently by specialists in the field. Once installed or modified the materials instantly become available to all users. Cone will be the obligation to make and distribute tens or even hundreds of replacement floppies every time a minor alteration to software is made. .

Now, if this system is to work, a sophisticated high-speed, user-friendly (indeed totally transparent) communication system will have to link the various distribution and information nodes to one another.

Lest readers should feel that all of this is in the realm of pipedreams, they should be reassured that this is not the case. Two far-sighted projects already attest to this: MIT's ATHENA project and Camegie-Mellon University's (CMU) Andrew+ project. Both already allow transparent access to information on ETHERNET networks running at 10Mbits per second. Access is available through a range of machines, including MACs, IBM PCs, DEC and SUN workstations.

Thus, with the use of systems such as these, it will be possible to support very sophisticated and highly distributed language learning environments wherever appropriate delivery facilities are available even if the actual resources called upon happen to be hundreds or even thousands of kilometres away (Lian, 1988b).

Under these circumstances, it may be appropriate to modify our notion of computer to that popularised by SUN Microsystems and to say that: The Network is the Computer.

Thus the first and undoubtedly most important category of CALL software, i. e. courseware should perhaps now be redefined as: methodologically sound computer-based materials for language learners, which is available from many different sources, and which is capable of running on a broad range of machines.
 

CALL Software: Management Systems
The availability of large amounts of software for routine use by large numbers of people immediately raises at least two non-trivial problems.

1. The first relates to the making available of the software in as equitable a fashion as possible together with the reporting of software and other system problems. This is a user-management problem.
2. The second relates to the monitoring of students' interactions with courseware and to the definition of paths through and within the various software packages at their disposal. This is a management of learning problem.

1. specialised open access or public clusters of relatively powerful and expensive equipment owned by educational institutions. If necessary, this equipment is attached to specialised peripherals (e.g. videodisc), and
2. rather more general purpose, and relatively cheap, equipment owned privately by individuals to be used either on a stand-alone basis or in multiuser mode via remote links to a host computer. Learners could then work from their college room or wherever they have access to appropriate communication facilities. Although public clusters are likely to offer higher quality learning facilities there are many worthwhile activities which can still be performed on other equipment.

Scheduling
A not unrelated problem is that of ensuring that all students are given equal opportunity for access to the equipment of their needs and, as a consequence, that students are queued smoothly through the system without having to compete for access to equipment. For instance, it is not uncommon, where such systems do not exist, for a student to hog a machine for many hours and, as a result, for others to be deprived of its use. This has sometimes led to unpleasant scenes, and, in some cases, even violence.

A partial solution is to develop software incorporating a reservation system smart enough to know which peripherals are connected to which workstation and which knows the parts of the system to which each students is entitled to have access. 7bus a student could reserve a workstation ahead of time and be guaranteed allocation of the correct sort of equipment for a minimum period of time during which he/she will "own" that equipment. The system would also ensure that the programs available to the user at that time were guaranteed to operate correctly on the equipment reserved.
 

Problem-logging
A further important consideration is that of signalling and logging error conditions and system errors. It is most important to know which programs are running incorrectly (and why) so that steps can be taken to correct the problems. Very few programs are likely to be totally free of execution bugs, even those produced commercially. Students should be shielded from incomprehensible messages from the operating system and from workstation crashes wherever possible. It is unfair to expect CAL users to be experts in computing.
 

Security
Given the amount of time and effort invested in software development and the expense of purchasing programs, software should be protected from malevolent or naive attacks. It would be catastrophic if systems needed to be constantly reinstalled as a result of voluntary or involuntary removal or corruption of programs and records.
 

Communication
Another part of user-management which deserves attention is the potential by the learner to communicate with teachers or with other students. The nature of CAL activity is such that there are many times when students need to convey messages to others when these other people are unavailable. It is necessary, therefore, to find a way of allowing students to communicate with others at the time when they need to do so. The answer is relatively simple: provide an electronic mail ' facility containing a number of safeguards for protecting users from undesirable correspondents. Even there, however, short-cuts should be provided for such things as addressing mail so that students do not need to spend a significant amount of time learning system commands as opposed to engaging in CALL activities.
 

Profiling Activity
Finally, it is important to be able to profile students' activities on the system, to extrapolate work patterns and equipment usage. In turn, this information will enable better day-to-day and strategic management of the system and help to determine purchasing and maintenance policies. In a research perspective, it will allow courseware and system developers to examine the ways in which students are actually making use of the system, to infer learning patterns and to develop educational strategies.

A system meeting most of these requirements, CLAM (Deschamps 1986), has been in routine use for some years now at the University of Queensland and has proved invaluable as an aid to the development of students' autonomy and general self-management of CALL activities in a trouble-free environment.
 

Management of Learning
The second management aspect which needs to be addressed is that of management of learning.

Depending upon one's views of how learning environments ought to be constructed, it may be useful to provide learners with guidance mechanisms: programs to sequence students' CALL activities and to apply certain evaluation and checking procedures irrespective of the students' wishes.

Whatever one's philosophy of learning: e.g. directive or non-directive, the fact remains that all teachers are constantly called upon to make decisions about t e management of their students' learning and they constantly do so, even if it is only to permit students to do whatever they like. One suspects that the objections which CALL methodologists often express in respect of guiding learners through CALL materials lie rather in the fact that guidance systems are somewhat primitive and inflexible rather than in the fact itself that guidance is provided.

It is thus incumbent upon CALL developers not to ignore the issue, especially as systems grow in complexity. A concerted effort will need to be made to develop a system which will act as a kind of sophisticated surrogate expert teacher.

One such system is currently in the developmental stages in Australia at La Trobe, Melbourne and Deakin universities. It is called the EXCALIBUR system and is conceived as a general-purpose intelligent CAL system. In its finished form it is expected to provide automatic profiling mechanisms of students' performances leading to the elaboration of a learner model for each student. This model will then be analysed by an expert teacher program which will, in turn, make recommendations as to the path to be taken through the available software and as to the evaluation and procedures to be applied to each student's interactions (Richards, 1986)

One of the strengths of the system is that it will be able to provide teachers and planners with sets of tools for generating expert teacher models, for analysing performances and for constructing student profiles.

Although the project may appear too ambitious and, perhaps inevitably, less flexible than desirable it ought to be remembered

1. that the situation might change if one actually goes to the trouble of attempting to solve the problem instead of complaining about it; and
2. that the information obtained on students' interactions with the system is of considerable value to CALL research and to language learning research generally.

CALL software: Development Systems
The final aspect of CALL to be addressed here is that of a development system/environment.

CALL authors, in their enthusiasm to get on with the job of producing materials for their students tend to forget about this aspect of their work. Indeed they may simply not even be aware that choices are actually available. In turn, this has meant that far too many programs have been written in languages specific to a particular machine or under poor and idiosyncratic operating systems which provide no portability across hardware. i.e. the ability for the same program to run on a number of different machines

Forward-looking CAL designers have understood that the proliferation of idiosyncratic systems would ultimately be extremely damaging to CAL as a medium and CAL as an industry. It is therefore interesting to take a look at the emerging standards for CAL in the early 1990s. These standards are reflected in developments at MIT and CMU.

First is the emergence of high-powered workstations as a delivery medium (SUN, MicroVax) all connected to one another and to centralised software distribution facilities through transparent high-speed networks. Interestingly, prices of high-powered style workstations are dropping quite rapidly. For instance in Australia, some workstations costs about the same as a high-end PC. Prices of workstations will undoubtedly drop further. By the same token, the power of the PC is increasing all the time and it is likely that a de facto merger between workstations and PCs will soon occur.

Second is the emergence of UNIX as the operating system of choice.

Third is the emergence of a standard windowing and graphics environment. The likely standards in this area will probably be ATHENA's X-window system.

Fourth is the clear choice in favour of C as the preferred programming language although special authoring languages such as cT, Course of Action and Authorware Professional are gaining in importance.

Taken together, all of these changes add up to a major and dramatic improvement in flexibility and resource-sharing due to enormously enhanced communications and portability across hardware systems. Furthermore, developers of these systems have not forgotten the world of PCs and MACs. For instance, both ATHENA and Andrew+ support communications with these machines in one way or another and on Andrew+ at least provide cross-compilers so that courseware developed using authoring systems on SUNs can actually be run in stand-alone mode on PCs or MACs.
 

Conclusion
The notion of CALL software is not simple and covers a range of critically important systems-based requirements if computer-enhanced technology is to be exploited efficiently and effectively in language learning.

These requirements can be summarised as follows:

1. Methodologically valid lesson materials, courseware and other appropriate software: undoubtedly the core of any CALL system.
2. Effective distribution systems:
1. the ability to make available and upgrade software quickly and effectively in response to students' motivations and needs as they occur;
2. the ability to allow learners to communicate quickly and effectively with other users (including supervisors and friends or themselves: notes, diaries).

1. User-management systems capable of ensuring equitable self-managed access to CALL resources as well as reporting problems and difficulties in the day-today running of the system.
2. Learning management systems capable of providing help and guidance if and when learners require/request it.

4. High quality software development environments. These provide stable, proven, well-developed and machine-independent systems ensuring maximum portability and shareability from one system to the next.

Finally, it seems that if real progress in CALL is to occur then at least two things must happen:

1. It will be necessary to develop a new kind of language teacher: one who is highly proficient in the language being taught, highly competent in the methodology of language learning, highly aware of the technologies available and highly proficient in the use of such technologies. Of course, it is not suggested that all teachers should be educated in this way. However, some specialists to act as consultants are undoubtedly required. In addition, one ought to establish coherent working groups for CALL development where expertise would be shared constructively for the benefit of all.
2. As a consequence of the previous point, one ought to encourage the coordination of resources and activities in such a way as to focus both national and international efforts toward the effective pursuit of common objectives while, at the same time, not discouraging personal initiatives. The solution to both of these problems may well lie in the formation of new national or international bodies committed to active and continuing research in the area but operating under the umbrella of an established, active and highly-respected parent body.

1. The research reported here has been partially supported by grants from the Australian Research Grants Scheme (ARCS) and various University of Queensland Special Projects grants. This support is gratefully acknowledged.
2. Acronyms and abbreviations used in this article are:
CAL Computer-Aided Learning;
CALL Computer-Aided Language-Learning;
CELL Computer-Enhanced Language-Learning.

3. Computer-Enhanced Language Learning is simply language-learning made better through the use of computers. No judgement is made as to the kind of programs or materials used. Computer-Aided (or Assisted) Learning often appears to imply interaction between learners and programs conceived primarily as lessons.
4. From 1984 to 1988, the author was Director of the University of Queensland's Humanities Computing Resource Unit (formerly known as the French/German/ Philosophy CAL Unit). In that time, it has provided about 20,000 hours of CALL and has produced about 300 separate lessons. The delivery system consisted of 13 terminals/workstations.

Cryle, P. M. and Lian, A-P. (1985) 'Sorry, I'll Play That Again', in Boarder, J. A. and Lichtenstein, S. (eds): Student Control of Learning: Computers in Tertiary Education, Centre for the Study of Higher Education, University of Melbourne, December, pp. 204-213.

Deschamps, G. (1986) Controlled Line Access Monitor (CLAM), (User administration and management package), Brisbane, FGPCAL Unit. University of Queensland,

Higgins, J. J. (1988) Handout accompanying 'Power to the Pupils', Plenary lecture delivered at the 1988 RELC congress, Singapore, April.

Lian, A-P. (1985) 'An Experimental Computer-Assisted Listening Comprehension System', in Revue de Phonétique Appliquée, nos. 73-74-75, pp. 167-184.

Lian, A-P. (1986) 'Generative Computer-Aided Language Learning: The University of Queensland CALL Project and EXCALIBUR', in (Girle, R. A. ed.): Report of the First Round Table Conference.. Australian Educational Expert System Project, Project EXCALIBUR Publications, University of Queensland, Brisbane, pp. 25-136.

Lian, A-P. (1987) 'Awareness, Autonomy and Achievement in Foreign Language Learning', in Revue de Phonétique Appliquée, vols. 82-84, pp. 167-184, and pre-published in The SGAV Review.

Lian, A-P. (1988) 'Distributed Learning Environments and Computer-Enhanced Language Learning', in Dekkers, J., Griffin, H. and Kempf, N. (eds): Computer Technology serves Distance Education, Rockhampton, Capricornia Institute, 1988, pp. 83-88.

Lian, A-P. (1991)'Intelligence in CALL, in Pennington, M. C. and Stevens, V. (eds): Computers in Applied Linguistics: An International Perspective, Multilingual Matters, Clevedon, Avon, U.K., (1991) (in press).

Lian, A-P. and Thomquist, L. and Thornquist, L. (1987) 'Computer-Based Technology in Language Learning: Beyond the Walls of the Traditional Classroom', in Journal of Educational Techniques and Technologies, vol. 20, no. 2, Premiere, pp. 24-31.

Richards, T. J. (1986) A Description Of the Educational Expert Systems Project, Department of Computer Science, La Trobe University, Bundoora, Victoria, Australia, (Educational Expert Systems Project, Technical Report 1).

Andrew Lian is Professor of Computer-Enhanced Language Learning and Director of the Language Centre at Bond University, Cold Coast, Queensland 4229.