Chapter Two:  Background and Definitions

           The  purpose  of this chapter is to define the  terms  
     computer-assisted  language learning (CALL) and  "authoring  
     system" and to discuss the importance of the latter to  the  
     future development of CALL.

               Computer-Assisted Learning (CAL) is discussed and  
     CALL  is defined as a special subset of CAL which  presents  
     unique  difficulties which the those designing software for  
     the support of authors have not sufficiently addressed.

          The  various  approaches that have been used  for  the  
     production of language learning materials,  or  courseware,  
     are  briefly  described.    The term "authoring system"  is  
     defined and contrasted with other mechanisms for courseware  
     production,   such as authoring languages.    The nature of  
     authoring  systems,  as well as some of the  features  that  
     serve to differentiate different types of authoring systems  
     are also briefly examined. 

          Finally,  the current state of CALL as a teaching tool  
     is  discussed  and the major remaining impediments  to  its  
     widespread   reviewed.     These  include  the  dearth   of  
     acceptable  courseware and the lack of teacher  involvement  
     in the creation of CALL materials.   Key factors  governing  
     the  quantity  of software produced,  the quality  of  that  
     software,   and   the  relationship  between  the  two  are  
     examined.    Reasons  are advanced for arguing that use  of  
     authoring  system  by teachers provides the best  means  of  
     satisfying the demand for both quality and quantity.

     2.1  Computer-assisted language learning (CALL)

          To  arrive  at an understanding  of  computer-assisted  
     language learning (CALL),  the more global term,   computer  
     assisted  learning (CAL) is examined first.    The specific  
     type  of  CAL  dealt with in this  study  is  defined,  the  
     confusion  over  terminology  existing  in  the  field   is  
     discussed, and examples of the types of activities that are  
     most often called CAL are given.    CALL is then defined as  
     a  subset of CAL and the special factors which set it apart  
     are considered.

     2.1.1 Computer-assisted learning (CAL)

               For the purposes of this study a specific type of  
     computer-assisted  learning (CAL) which might be refered to  
     as "practice-oriented CAL" is  defined.   Practice-oriented  
     CAL  may  be  thought of as  learners'  practice  involving  
     interactions  with  computers in a structured  environment.   
     This   structured  environment  is  provided  by   computer  
     programs called "courseware", the content of which has been  
     established prior to the learners' interaction.  The course  
     of  the  interaction is mediated,  or  controlled,  by  the  
     courseware.   (Henceforth, references made to actions taken  
     by  the  "computer" should be interpreted to  mean  actions  
     foreseen  in  the  computer  programs  that  make  up   the  

          "Practice",  in the context of learning, is defined as  
     the act of doing something repeatedly in order to acquire a  
     skill.   It can also be interpreted as instruction received  
     through the repetition of exercises or lessons.  "Rehearse"  
     is offered as a synonym (Funk & Wagnall's Dictionary, 1976,  

          The general application of the term CAL involves  more  
     than  simply  the "practice"  of a given concept or  skill.  
     Learning via the computer can,  after all,  certainly  take  
     place in many situations in which there is no engagement in  

     practice  activities,  per  se.   These   applications  are  
     excluded from the current,  working definition of CAL.  The  
     use  of  the  computer as a tool of inquiry when  making  a  
     data-base search, for example,  is not practice, unless the  
     subject  is  data-base  retrieval.   Learning  will  often,  
     nevertheless,  result from this activity.   Similarly,  the  
     use  of  the computer as a word processor is  not  strictly  
     practice  unless  one is practicing  typing  and  redaction  
     skills.   Many experts maintain,  however,  that use of the  
     computer for word-processing can facilitate learning higher  
     level composition skills (Hopper,1984; Bloch, 1986, p.4)

          There  is general agreement that CAL is interactive in  
     the  sense that the learners participate in  an  immediate,  
     back-and-forth  interaction  with the computer  during  the  
     learning experience.   This activity can be contrasted with  
     situations where the user deals with the computer in a more  
     remote, detached manner, as when submitting a batch program  
     for  processing  and returning hours later to  receive  the  

          Interactive  practice with the computer may take place  
     within the structured environment provided by the framework  
     of  a  pre-established  computer  program.   Alternatively,  
     learners may interact with the computer  directly,  through  
     some high-level computer language.  Boyd (1982, p.307) sees  
     the  need to consider programmers as learners,  and  Bagley  
     (1974,  p.5)  even  considers such non-structured  practice  
     (where  students  must define the  problem  themselves  and  
     write  a program to solve it) as a second category of  CAL.    
     The more general trend,  however,  is to adopt the narrower  
     interpretation  of  CAL  as  "learners'  interactions  with  
     computers   within   the   structured  framework   of   CAL  

          Higgins (1982,  p.  28) illustrates the wide range  of  
     forms  that interaction within a structured environment can  
     take,  from the restrictiveness of courseware built on  the  
     model   of  programmed  instruction,   to  the  variability  
     exhibited by modern `adventure' games.   The common element  
     is  that all variation follows paths pre-determined by  the  
     courseware  author.   Higgins (1982,  p.28) contrasts  this  
     with  `synthetic'  programs where the computer is  able  to  
     explore  variations not previously foreseen by the  author.   
     He  cites,  as  possible examples  of  synthetic  programs,  
     Weizenbaum's  Eliza,    Winograd's  Shrdlu,   and  his  own  
     Grammarland,  none  of  which  could be considered  as  CAL  
     according to the above definition.

          The   final   component  of  the  CAL  definition   is  
     mediation,  or evaluation of students' progress through the  
     material,  on  the  basis of some  set  of  pre-established  
     criteria.   The courseware can act as a surrogate moderator  
     on the basis of its built-in `knowledge' of correct answers  
     or  paths.   Mediation might involve matching the student's  
     response against a stored `correct' response and responding  
     accordingly, as in a simple drill and practice exchange. It  
     could also,  as in a sophisticated adventure game,  involve  
     making  learners  accountable for their  actions,  some  of  
     which are more appropriate than others to the situation  at  
     hand.   The Eliza program,  where there are no "correct" or  
     "incorrect"  responses,  would represent an example of non- 
     mediated   interaction   and   thus   be   excluded    from  
     consideration as an example of CAL.

          It  is difficult to find any definition of CAL   which  
     is acceptable across-the-board,  hence the qualification of  
     the current definition as "practice-oriented" CAL.  Several  
     historical   factors   make  it  difficult  to  achieve   a  
     definition which would receive universal acceptance.

          Initially  there existed a confusion between  CAL  and  
     Programmed  Instruction (PI) because CAL was simply seen as  
     an  outgrowth of print-based PI materials  (Orstein,  1970,  
     p.216). Scanlan (1971, p.85) distinguished the two terms by  
     pointing  out that PI did not allow for  student-controlled  
     learning  or  "unprogrammed  instruction".    Kenning   and  
     Kenning  clearly  differentiate PI and CAL by showing  that  
     any attempt to match the flexibility of the computer by the  
     use  of  traditional print materials would  be  technically  
     infeasible,  requiring  "huge `scrambled books' with  pages  
     and pages of mostly unnecessary explanations, together with  
     an  extremely  complicated  system  of   cross-references."  
     (Kenning and Kenning, 1983, p.2)

          The  difficulty  in defining CAL is increased  by  the  
     large  number of acronyms and abbreviations that have  come  
     into  use.   The greatest recent confusion has been between  
     CAL  and  Computer Assisted  Instruction  (CAI).   CAL  has  
     historically   been   seen  as  the  more   generic   term,   
     encompassing  all manifestations of computer-aided learning  
     (Bagley, 1974, p.4).  CAI has been seen as a subset of CAL,  
     including  only  those applications of the  computer  which  
     most resembled traditional programmed instruction  (Hooper,  
     1975, p.13). 

           Recently,  however,  several  writers have been  more  
     conservative in their application of the term CAL.    In an  
     article  on the subject of CAI/CAL acronyms,  Goldes (1984,  
     pp.353-357) views CAL as only slightly wider in scope  than  
     CAI,  while in his book on the subject,  Davies (1982, p.3)  
     sees the two terms as more or less synonymous.  

          At  the 1983 annual convention of TESOL  (Teachers  of  
     English  to Speakers of Other Languages) in Toronto it  was  
     found  that a major source of difference in the application  
     of the two terms was geographical.  British writers  tended  
     to  use  CAL  while American writers  used  CAI.   While  a  
     perceived  difference  in philosophy between the two  terms  
     was  recognized,  "instruction"  being  seen  as  "teacher- 
     centred"  and "learning" as "student-centred",   a  general  
     consensus developed that CAL should be standardized as  the  
     acronym to use in the field of language learning.  (Sanders  
     and Kenner, 1983-b). 

          Besides CAL and CAI,   there is an abundance of  other  
     acronyms  and abbreviations,  some of which fall within the  
     scope  of  the definition of CAL given above  and  some  of  
     which do not.   Usually roughly synonymous with CAL/CAI are  
     computer-based education (CBE) (Bagley,  1974,  p.4; Avner,  
     1978, p.24), computer-based instruction (CBI) and computer- 
     based  learning (CBL) (Davies,  1982,  p.3),  and computer- 
     assisted teaching (CAT) (Kenning &  Kenning,  1983,  p. x).

          Examples   of  non-CAL  abbreviations  are   computer- 
     generated  materials (CGM),   computer-managed  instruction  
     (CMI),   and computer-based testing (CBT).  CGM can be seen  
     as  a  teacher's aid,  and not as CAL,  because it  is  the  
     teacher who interacts with the computer,  not the  student.   
     The  student  works with the materials that  are  produced.   
     CMI,  by  itself,  need not include interactive practice on  
     the computer.  It can be seen, as well, as a teacher's aid,  
     used for keeping track of a student's progress.   CMI "does  
     not instruct but rather guides instruction." (Goldes, 1984,  
     p.355).  CBT does not meet the requirement that CAL involve  

          Lists  of typical CAL activities generally include the  
     following basic classes: 

               drill and Practice programs, which drill learners  
     on their knowledge of a particular subject,

               simulation  and  modelling  programs,  where  the  
     student engages in a true-to-life situation,

               gaming,  where there is a competitive interaction  
     between participants (one of whom might be the computer) to  
     achieve a goal, and

               problem solving,  which requires the synthesis of  
     higher  order  rules  and  concepts.   (Bagley, 1974, p. 5;  
     Crawford, 1981, pp. 23-37; Cohen, 1983, pp. 10-11).

          Drill  and Practice is often further broken down  into  
     Drill and Practice and Tutorial CAL,  the distinction being  
     that  Drill  and  Practice  involves  only  practice  while  
     Tutorial CAL involves the presentation of the concept,  its  
     practice,   and  subsequent  evaluation  of  the  students'  
     understanding. (Fitzgibbon & Grate, 1970, p. 916; Crawford,  
     1981, p. 30; Paul, 1982, pp. 6-7; Cohen, 1983, p. 11)

     2.1.2  CALL as a special subset of CAL 

          Before  proceeding with the examination of what  makes  
     CALL a unique subset of CAL,  it is useful to simplify  the  
     question   by  excluding  from  the  discussion  the   many  
     applications of the computer to language study which do not  
     fit  within  the definition of CAL offered  above.   Davies  
     (1982,  pp. 1-3) lists a series of language-related uses of  
     the   computer  which  are  not  CAL;   such   as   machine  
     translation,  automatic dictionaries, literary & linguistic  
     processing of texts,  word processing, and speech synthesis  
     and analysis.   To this list,   Higgins (1982, p. 25). adds  
     style editor software  such as The Writer's Workbench. 

          CAL has been applied to the learning of language since  
     the  earliest days of its introduction into the  domain  of  
     teaching.  As early examples, Suppes and Mackey (1978 p. 9)  
     cite  a  first-year Russian program at  Stanford  in  1967;  
     Scanlan  cites  (1971, p. 84)   the use of  the  PLATO  CAL  
     system  for  Latin  since  1968;  and   Suppes  and  Mackey  
     (1978, p. 10) cite the use of the TICCIT system for English   
     at Brigham Young University  in 1972. 

           Applications   of  CAL  to  language  learning   have  
     consistently  presented  course designers  and  programmers  
     with  "severe technical problems in relation to  strategies  
     for  coping with the complexity of natural language  within  
     the confines of a computing context" (Last,  1983, p.Ć 84).      
     Orstein  (1970,  p. 215)  notes that the  ease  with  which  
     material  can be adapted to CAL is in direct proportion  to  
     the concreteness and specificity of the material.  However,  
     linguists  and educators have so far been unable to provide  
     a  concrete,  specific  and  workable  description  of  how  
     language works, let alone how people learn it.  

          Many  of the programming approaches popular in CAL are  
     difficult to apply to CALL.  For example, "Generative CAL",  
     where  the courseware,  based on an established  algorithm,  
     generates  new,  unique  problems for each  student,  is  a  
     popular tool in scientific disciplines.   It is often  used  
     "to  clarify  a particularly difficult concept by means  of  
     models  and simulations,  the performance of  calculations,  
     the  plotting  of  graphs,  etc."  (Davies,  1982,  p. 10).    
     Application  of this technique to CALL,  however,  is  made  
     difficult,   except   in  carefully  constructed  contexts,  
     because  generation  of original language by  the  computer  
     assumes  the  computer's ability to analyze  the  learner's  
     original, possibly mal-formed, open-ended responses.

          The   parsing  of  open-ended  natural   language   is  
     considered  briefly  in  section 3.10.4  but  lies  largely  
     outside  the  scope  of this thesis as its  realization  is  
     beyond  the  capabilities  of  the  current  generation  of  
     special-purpose   computer  languages  and  systems   under  
     discussion.   Hart (1981, p. 9), for example, believes that  
     the  application  of  the  very  popular  TUTOR   authoring  
     language  to CALL is hindered by the language's poor string  
     (word)  manipulation  and  limited  recursive  capabilities  
     which make the creation of parsing and  language-generation  
     routines  difficult.    Contrary to the situation in  other  
     forms of CAL where developmental work on generative CAL has  
     proceeded  much  more  rapidly,   the  production  of  CALL  
     courseware  will  continue  to require  the  provision,  in  
     advance, of data for each discrete interaction.

          Another  complication  in  CALL   programming  is  the  
     frequent   necessity  to provide for situations that  often  
     yield   multiple  correct  answers,   varying  degrees   of  
     acceptability of an answer,  and the placement of  emphasis  
     on either the content or the form of the answer.

           The  uniqueness of CALL involves more than the simple  
     enhancement of computer programming techniques.  It extends  
     to  the  very  nature of  the  learning  experience,  which  
     differs  radically from that of many of the disciplines  to  
     which  CAL has traditionally been applied so  successfully.   
     Paramskas   (1983,  p. 6)  explains that language  learning  
     cannot be measured in terms of the quantity of  information  
     acquired,  in  terms of black and white or right and wrong.   
     The  fact  that students can sucessfully master  a  set  of  
     exercises  does not,  in any way,  indicate that they  have  
     improved their ability to use the language. Crawford (1981,  
     p. 10)  proposes  that  CAL materials  must  lead  students  
     beyond  the  simple  "knowledge" of  material  towards  its  
     integration   and   the   synthesis  of  new   systems   of  
     understanding.   Nowhere does this seem more important than  
     in  CALL,  where students can only acquire language  skills  
     through the ongoing synthesis of new language systems based  
     on   the  integration  of  new,   and  often   conflicting,  
     information.     The  traditional  programmed   instruction  
     approach, which still provides the basis of much CAL, where  
     students cumulatively add to their repertoire of "mastered"  
     skills,  will  not  lead  inexorably  towarda  the  desired  
     result,  the mastery of all aspects of language use, in the  
     same way as it might in, say, biochemistry or physics. 

               The  problem  posed by  evaluation  of  learners'  
     performance  offers  a  good  example of  the  CALL  design  
     difficulties  which   stem  from the  very  nature  of  the  
     language   learning  experience.     In  many  disciplines,   
     improvements  in  performance  can be  measured  through  a  
     simple comparison of pre- and post-test results and can  be  
     linked  almost  directly  to students'  competence  in  the  
     field.     Students can proceed step-by-step, mastering one  
     unit  before moving on to the next,  and,  at the end,  the  
     computer  can  assure  them that they  have  "learned"  the  
     content of the activity.   Language is a much more holistic  
     enterprise,  where  performance  in  any  one  activity  is  
     coloured by a host of factors and where mastery of a single  
     "unit"  in  a  language  course can in  no  way  be  linked  
     directly to improved linguistic competence.   


               The  analysis  to  be developed  in  this  thesis  
     focuses  on  the unique programming problems posed by  CALL  
     and  on  the  programming implications  of  the  particular  
     nature of language learning. 

     2.2  Authoring systems

          "Authoring"  is defined and the historical  approaches  
     to  authoring  CALL courseware are  briefly  examined.  The  
     historical  evolution  of  the term "authoring  system"  is  
     discussed  and a working definition of authoring system  is  
     established.   Finally, the nature of authoring systems and  
     some  of  the major factors that   differentiate  one  from  
     another are considered.

    2.2.1   "Authoring":  A working definition

          Authoring,  in the context of CALL, is defined here as  
     the process of producing courseware.   It involves not only  
     the design and writing of the educational content, but also  
     its  subsequent  realization in a form deliverable  by  the  
     computer.   The  need  to  produce both a  lesson  and  the   
     computer  program  that embodies it  means  that,  ideally,  
     potential  writers of lessons have to have an ability,  not  
     only to communicate their subject matter well,  but also to  
     work  well in a computer programming  environment  (Dowsey,  
     1974,  p. 401).  It has long been recognized that some type  
     of   support  is  required to bridge the  gap  between  the  
     existing  level of computer programming expertise exhibited  
     by  most potential "authors" (usually teachers  with  ideas  
     for  learning  activities)   and  the high  level  of  such  
     expertise required to produce the desired final product  in  
     the form of computer programs (McCambridge, 1982).

          This problem was addressed early in the development of  
     CAL,  through the creation of a family of "special-purpose"  
     computer-programming  languages  designed to  simplify  the  
     authoring process and referred to as "authoring languages".   
     They  first  made  their  appearance  around   1960.    IBM  
     introduced  T.I.P.  (Translator for Interactive  Programs),  
     the   forerunner  of  the  Coursewriter  system,   and  the  
     University  of Illinois was working on  C.A.T.O.  (Compiler  
     for Automatic Teaching Operations),  the forerunner of  the  
     TUTOR  authoring  language used in today's  PLATO  teaching  
     system (Kearsley, 1982, p. 430).

     2.2.2   Approaches for authoring CALL 

          The  many  options  available for  the  production  of  
     computer  programs range from the "low-level" or  "machine"  
     end of the spectrum to the "high-level" or "human" end.  As  
     prospective authors move towards the high-level end of this  
     continuum,  the  increased complexity and sophistication of  
     the  software they are using,  the "instrument" with  which  
     they  actually interact with the computer,  allows them  to  
     describe in ever more abstract,  human terms what they want  
     the   computer   to  do  (Boyd,   1970).   This   increased  
     sophistication is purchased at the cost of pre-defining and  
     limiting  ever  more the available  options.   The  avenues  
     explored in the production of CALL courseware range,  then,  
     from the use of low-level "machine" language,  through  the  
     use  of  "general-purpose  high-level"  languages  such  as  
     BASIC,  through  the  high-level authoring languages  noted  
     above,  to the most sophisticated instruments for producing  
     CALL courseware which are called "authoring systems".   Machine language

          Programming in low-level computer languages, "machine"  
     language or "assembly" language, remains the true domain of  
     computer  professionals.   Here  alone  is  the  programmer  
     actually  controlling the computer directly,  providing  in  
     the   instructions   a  code  for  each  discrete  computer  
     operation.   "Low-level" programmers must have an  intimate  
     knowledge  of  the structure of the computer.   There  are,  
     however,  no pre-established formats that prevent them from  
     taking full advantage of its capabilities.   Little use has  
     been made of this approach in CALL beyond the production of  
     over-the-counter  devices  such  as "Speak  'n'  Spell"  or  
     pocket translators (Jensen, 1982, p. 50).   Original programming using general-purpose, high- 
     level programming  languages

          Original  programming  of educational  material  using  
     general-purpose,  high-level,  programming languages  (GPL)  
     allows   programmers  to  describe  the  operations  to  be  
     performed in terms which begin to reflect human rather than  
     computer realities.   One can,  for example,  can issue the  
     instruction   to  "print"  something,   rather   than   the  
     instruction  to  "move contents of register C  into  memory  
     location  HB8000".   The object of a GPL remains,  however,  
     the discrete,  step-by-step  control of the computer.   The  
     syntax  and organization of a GPL mirrors the structure  of  
     the  computer,    and  not  the  needs  of  any  particular  
     application or group of computer users,  such as  educators  
     (Shuyler, 1979, p. 29).

          GPL  programming represents by far the most widespread  
     approach used for CALL.    A 1976 survey showed that six of  
     the  ten most commonly used programming  languages for  CAI  
     were GPLs,  the most popular being BASIC,  APL, and FORTRAN  
     (Kearsley,  1976).   Olsen (1980,  p. 345),  citing another  
     survey,  notes that BASIC is the leading language for CALL,  
     being used 51 times as compared to 21 for TUTOR and 10  for  
     Coursewriter,   the   latter  being  "authoring"  languages  
     (Olsen, 1980).     Shortcuts  to  the  original   programming   of  

         In  order  to reduce the time required for  programming  
     and  to  improve  the quality  of  the  resultant  computer  
     programs,   various  "shortcuts" to the completely original  
     programming of courseware have been adopted by programmers. 

          "Structured programming" is a method of programming in  
     which   programmers  develop  independent,   modular   sub- 
     procedures  to  accomplish various  fundamental  operations  
     common to all of a set of programs.   Holmes (1983,  p. 29)  
     describes  the use of this method to streamline programming  
     operations  during  the development of the CLEF  series  of  
     French lessons.   A series of programming routines specific  
     to  language  teaching requirements,  such as  analysis  of  
     input, spelling error checks, feedback strategies, and help  
     procedures were modularized and standardized so as to  make  
     those  aspects  of producing the computer programs for  the  
     lessons a degree  easier and less time-consuming.

          "Programming  aids" or "authoring utilities" are  sets  
     of such procedures,  specially designed for CALL, which are  
     available in the marketplace.   They are usually programmed  
     in low-level computer languages so as to be very  efficient  
     and rapid during execution,  and they often exhibit a level  
     of  sophistication  beyond  that  of  many  would-be   CALL  
     programmers.    Jensen   (1882,   p. 52)  points  out  that  
     "Utilities reduce the drudgery of some kinds of programming  
     by turning over the tedious,  error-prone aspects of coding  
     to  the  computer".   He cites the example of  Apple  PILOT  
     utilities  for  graphics,   sound  effects,   and   foreign  
     character  sets.    EnBASIC  (Enhanced  BASIC)  is  another  
     example,  being a set of machine language routines that the  
     BASIC program can call upon for sophisticated and otherwise  
     time  consuming operations such as the mechanical  analysis  
     of  a  student's answer and the noting,  via  proofreader's  
     symbols,  of  the  discrepancies from  the  correct  answer  
     (Chapelle & Jamieson, 1983).

          "Template" programming is a time-saving technique that  
     involves the separation of the "logic" of the program  (the  
     type  of presentation and the logical structures,  such  as  
     answer-handling  routines)  from  the "data"  (the  course- 
     specific content) (Boyd,  Keller &  Kenner,  1982,  p.110).    
     Authors can create new lessons by fitting new data into old  
     programs, so long as the new content can be accommodated to  
     the logic of the original program  (Dowsey,  1974, p. 403).  
     Teachers  can  thus insert their own pedagogical  materials  
     into pre-determined lesson formats such as multiple-choice,  
     fill-in-the-blank,   transformation,   cloze  tests,   etc.  
     (Holmes,  1982,  p. 12).  A template program is, in effect,  
     an "editable CALL lesson".   The programmer simply replaces   
     the  data  and content statements of one  lesson  with  new  
     content, yielding a different but algorithmically identical  
     clone.  To be effective, the original template program must  
     have   been  designed  so  as  to  offer  some  degree   of  
     flexibility  and  so as to facilitate the  editing  process  
     (Jensen, 1982, p. 51).

          "Driver" programs are still more sophisticated.  These  
     programs are neutral, and do not have to be edited in order  
     to  produce a new lesson.   All lesson-specific content  is  
     stored  on  an external data-base and is presented  to  the  
     student in a pre-established format (Shuyler, 1979, p. 33).   
     This approach is not new.   As long ago as 1970, Fitzgibbon  
     and  Grate (1970,  p. 92l) described what was in essence  a  
     cloze-generator,  a program to delete every nth word from a  
     text  externally in data then allow the student to fill  it  
     in.  Many driver-based CALL systems provide an option which  
     allows  for  the  creation of the external  data  bases  by  
     teachers and other non-programmers.    Authoring languages

          "Authoring   languages"  (AL)  can  be  described   as  
     "special-purpose"  programming  languages because they  are  
     designed specifically for one application:   the coding  of  
     CAL programs (Ashmore,  1983,  p. 4).   Authoring languages  
     are  sometimes  called "very-high level" languages  because  
     several GPL instructions may be required to accomplish what  
     can be covered by a single AL instruction  (Shuyler,  1979,  
     p. 30).  "They  represent an attempt by their designers  to  
     predict  the  types of commands and capabilities  educators  
     will need" (Wyatt, 1983a,  p. 36).  Essentially, the ready- 
     made,   CAL-specific   routines   described   above   under  
     "authoring  aids" have been incorporated into the  language  
     and  can  be  invoked with a  single  program  instruction.   
     Programmers are thus offered features which they would have  
     to  design  from scratch if they were using a  GPL  (Wyatt,  
     1983-b,  p. 7).   The  convenience and simplicity  of  ALs,  
     however,  is achieved at great cost in versatility (Jensen,  
     1982,  p. 51; Higgins & Johns, 1984, p. 102; Wyatt, 1983-b,  
     p. 7).

          Wyatt (1983-a,  p. 37) offers the following example of  
     the  convenience vs.  versatility  problem:    Apple  PILOT  
     contains  a  powerful  feature which allows  for  the  easy  
     creation  of special characters,  such as accented letters.    
     The drawback is that each new letter must be allocated  its  
     own  key  on the keyboard,  and the result is  a  confusing  
     cross-reference  of individual keys.  Designing the routine  
     from scratch in a GPL,  while more difficult,  would  allow  
     for  the  possibility  of combining keystrokes  so  that  a  
     simple  rule like "press letter-key followed by accent-key"  
     could be taught to the student. 

          Merrill  (1982,   p. 70)  offers  the  framework   for  
     summarizing  the key features of ALs.   1) They reduce  the  
     domain  of  possible  commands and  structures.  (In  other  
     words,  they  make the language do less.)  2) They  provide  
     commands  and structures which meet the specific  needs  of  
     instructional applications. (That is, they second-guess the  
     needs of educators.).  Finally, 3) they provide commands or  
     routines  which  perform  high  level  tasks.  (Thus,  each  
     command does more.)

    2.2.3    The definition of "authoring system"

          An "authoring system" (AS), in the context of CALL, is  
     a  computer  software  package through which  teachers  can  
     create  language-learning activities without having  to  do  
     any  actual  computer  programming  (Pogue,   1980,  p. 58;  
     Merrill, 1982, p. 77; Kearsley, 1982, p. 429; Wyatt, 1983a,  
     p. 37; Higgins & Johns, 1984, p. 102).  This removal of the  
     need  to  do  any actual programming  of  the  computer  is  
     central  to the definition and is the primary factor  which  
     distinguishes  ASs  from other related software (Higgins  &  
     Johns, 1984, p. 10).

          There  remains,   nevertheless,    a  great  deal   of  
     confusion in CALL literature over the nature of ASs (Wyatt,  
     1983-a,  p. 37).   There  is  confusion over the family  of  
     software  to which ASs should be ascribed.   For  observers   
     ASs  are an extension,  somewhat more sophisticated but  of  
     the  same  species,  as  authoring  languages  (Villeneuve,  
     1984),   while,  for  others,  they  are  an  extension  of  
     template- and  driver- programs   (Jensen,   1982,   p. 51)    
     Paramskas  (1983,  p. 5) states that the only consideration  
     setting templates apart from true authoring systems is that  
     the generation of new lessons from existing templates still  
     requires the intervention of programmers.    Indeed, Holmes  
     (1983-b, p 22) points out that some of the simplest ASs are  
     little  more than template programs in which the  entry  of  
     new lesson material has been made interactive.      A rigid  
     application  of the no-need-to-program criterion can  serve  
     to clearly separate the two types of software.

          The  distinction  between  ASs and  other  avenues  of  
     courseware  production  is  often  further blurred  by  the  
     practice  of  many  AS designers  to  resort  to  exterior,  
     independently- designed  procedures  programmed in GPL  for  
     the  handling of  sophisticated,  specialized  applications  
     (Pogue,   1980,  p. 65;  Holmes,  1983).   The  ability  to  
     reference  external,   GPL  procedures  renders  much  more  
     difficult the determination of the range of procedures that  
     a particular AS is inherently capable of.    The AS becomes  
     capable of anything that is programmable.   The no-need-to- 
     program  criterion of the above defintion of AS removes the  
     need to consider what might be possible using external  GPL  

          A  lingering  confusion  over terminology  stems  from  
     changes  in the application of the term "Authoring  system"  
     over the years.   When first used, it referred to an entire  
     educational  process which started with an analysis of  the  
     needs  of  the target  population,  proceeded  through  the  
     design   and  programming  stages,   and  ended  with   the  
     evaluation  of  the material against stated goals.  By  the  
     early 1970's,  the term was being used mainly to refer only  
     to the process of translating lesson-content into computer- 
     program.    1974  finally  saw it used as a label  for  the  
     computer  software that facilitated the  process.   (Pogue,  
     1980,  pp. 57-58;   Zinn, l974, pp. 381-384)  By that time,  
     there  were at least 64 different examples of this type  of  
     software in use (Braun, 1973, p. 1).

    2.2.4   The nature of authoring systems

         The prime criteria of ASs are that they should be easy to  
    learn,   simple to use, and should demand of the user no more  
    than a fundamental level of computer literacy  (Pogue,  1980,  
    p. 6;  Wyatt,  1983-a, p. 37).  Pogue (1980, p. 58) lists, in  
    addition to this,  four other desirable characteristics of an  
    authoring system:

          1.    a wide variety of instructional strategies should  
          be  available  to authors.   Authors should be able  to  
          create new strategies and modify existing ones;

          2.   an AS should be easy to use with all instructional  
          strategies,  whether  simple  or complex,  and  by  all  
          authors, whether novices or highly experienced;

          3.    an AS should allow for the use of the full  range  
          of   possibilities  offered  by  the   computer,   from  
          sophisticated  answer-analysis and complex branching to  
          the use of audio-visual media;

          4.    authors  should be able to test and modify  their  
          material easily.

          A typical AS could be described as follows.   Teachers  
     work in an interactive mode to create CALL material.   They  
     sit  before  the  computer and respond  to  the  software's  
     prompts as these appear on the screen.    Teachers indicate  
     the number of items in the exercise,  their sequencing, the  
     explanations to be offered,  the number of tries the  pupil  
     is  to be allowed,  etc.    They then provide the material,   
     stimulus,  correct answers,  predictable wrong answers, and  
     so forth,   for each of the items.   When finished,  the AS  
     software  makes  up  the exercise in  accordance  with  the  
     material  and  instructions that were  originally  entered.   
     (Barker &Singh, 1982, pp. 167-196; Edwards & Tillman, 1982,  
     p. 19; Jensen, 1982, p. 51; Kearsley, 1982, p. 431; Kenning  
     & Kenning, 1983, p. 11.)

    2.2.5   Factors differentiating authoring systems

         Authoring  systems  differ from one  another  in  three  
     ways:   the  manner  in which the user interacts  with  the  
     software  during the entry of course material,   the manner  
     in  which  the  software formats  the  material  for  later  
     delivery by the computer, and, most importantly, the degree  
     to which the user can alter the pre-established pedagogical  
     structures inherent in the software.   The entry of data

          Data  entry can be "on-line" (entered directly by  the  
     teacher seated before a computer) or "off-line" (written on  
     paper  by  the teacher and entered into the computer  by  a  
     computer professional).   During the data-entry phase,  the  
     AS   software  can  proceed  in  an  "interrogative  mode",  
     prompting the teacher to provide each piece of  infomation,  
     or  it  can allow the user to simply type in  the  required  
     data, without offering prompts. 

          Today, most authoring systems are interactive computer  
     programs,  but  this  has not always been the  case.   Many  
     early  ASs  used  elaborate coding  sheets  which  required  
     authors  to prepare their material "off-line" in  a  format  
     which  facilitated data entry.   The data which represented  
     lesson  material  was then entered by  computer  operators.  
     (Dowsey,  1974,  pp. 404,410).   It  was only as access  to  
     computer   terminals  became  more  widespread   that   the  
     convenience  of  "on-line",  interactive  entry  of  course  
     material began to be appreciated (Zinn, 1974, p. 388).

          Zinn   (1974,    p. 391)   finds   that,   while   the  
     interrogative mode seems well suited to the beginning user,  
     it  often frustrates the more experienced author.   What is  
     easy  for  the  beginner  may  be  constraining  and  time- 
     consuming to the professional.   He submits that,  ideally,  
     an  AS  should  provide for  both  interrogative  entry  of  
     material  by beginners,  as well as off-line preparation by  
     more experienced users.   Paloian (1974,  p. 442), as well,  
     opts for two levels of interaction,  interrogation for  the  
     inexperienced  author  and straight entry of data  for  the  
     experienced author.

          The  manner in which lesson data is entered is part of  
     the larger question of "user-friendliness".   Pogue  (1980,  
     p. 62)  believes  that  the level of assistance  should  be  
     determined   by  the  author.     When   interacting   with  
     beginners,  the  AS should guide authors through the entire  
     entry process,  prompting  them for  information,  checking  
     for structure or format mistakes,  and allowing them to ask  
     the  system for additional "help" if they do not understand  
     the next step.  

    2.5.2   The formatting of lesson material

          Once authors have provided the course content, ASs can  
     realize  the  final product by one of  two  mechanisms,  as  
     "drivers"  or  as  "code  generators".  With  the  "driver"  
     system,  a part of the AS software system elicits the  data  
     from  the author  and places it onto a data file.   Another  
     part of the software,  the driver itself,  "reads" the file  
     and   presents  the  material  to   the   students.    Code  
     generators,  on  the  other  hand,  actually  "write"  new,  
     separate  computer  programs  based  on  the  content  data  
     elicited.  Once written, the lessons are independent of the  
     AS that created them.  (Pogue, 1980, p. 65).

          The  literature does not clearly indicate a preference  
     for   either   approach,   insofar   as   the   pedagogical  
     effectiveness of the product is  concerned.    Rather,  the  
     mechanism  chosen affects factors such as  transportability  
     of  courseware  across  different types  of  computer,  and  
     memory  storage  requirements,   which  only  impinge  upon  
     pedagogical effectiveness in extreme cases.   Design options offered to the author

               Most  important to the purpose of this thesis  is  
     the range of options an AS offers to authors during  design  
     of the course material.   Again,  there are two fundamental  
     approaches,  the  "paradigm"  approach  and  the  "toolbox"  
     approach  (Hart, 1981, p. 7).   

          An  AS which uses the paradigm approach calls upon the  
     author   to   fit  lesson  content   into   pre-established  
     instructional formats, such as matching or multiple choice.   
     This  approach  is  resembles  the  interactive   templates  
     described  earlier.   Jensen (1982,  p. 51) finds that  the  
     convenience  and  ease of use offered by this  approach  is  
     achieved at great cost to flexibility,  because of the fact  
     that  the  underlying  format and  strategy  remain  fixed.    
     Merrill (1982,  p. 77), however, considers that the lack of  
     flexibility  of  any one template can be offset if  the  AS  
     provides  a wide range of paradigms for different types  of  

          Kearsley (1982, p. 434)  believes that AS should allow  
     authors  to specify  "instructional logic and strategy"  as  
     well  as content.   The "toolbox" approach offers this more  
     flexible possibility.  In this system, the author begins by  
     describing  to the AS the structure of the lesson  and  the  
     pedagogical strategy to be used.   Such description is made  
     through  the use of high-level "verbs".    These verbs call  
     into play a range of previously programmed basic strategies  
     called  "macros",  which  the  author  can  further  modify  
     through the specification of "arguments".   Once the  logic  
     of  the  lesson has been established,  the AS  elicits  the  
     actual  content,  based  on the requirements of the  chosen  
     strategy (Paloian,  1974; Dowsey, l974, p. 405; Dean, 1978,  
     pp. 21-23).

          As  already  incicated  (Pogue,  1980,  p.58),  it  is  
     desirable  that  an  AS be easy to use by  beginners  while  
     providing for the realization of the full range of computer  
     possibilities.   These  two vectors often move in  opposing  
     directions.  Using an AS can sometimes be even more complex  
     than using programming languages.  Indeed,  as the user  is  
     forced  to  provide "verbs" in order to design the  lesson,  
     there  is  a  danger of blurring  the  distinction  between  
     programming  language  and authoring  system  .   In  other  
     cases,  the AS is so unsophisticated that it forces authors  
     to  answer an endless series of  questions,  without  being  
     able  to  produce  anything  worthwhile  (Villenueve  1984,  
     p. 5).

          Jensen (1982,  p. 50)  and Dean (1978, p. 23) echo the  
     warning that sophisticated authoring systems, although very  
     versatile,  may become too cumbersome to use and the degree  
     of  flexibility  may be more than some  users  can  handle.   
     Dean opts for a system  more closely following the paradigm  
     approach  in  which "pedagogical  structure,  instructional  
     logic,  and  the nature of the  machine-student  interfaces  
     (are)  defined separately from the course content (and are)  
     provided by professionals" (Dean, 1978, p.23).

          Merrill  (1982,  p. 77)  opts for an AS  which  allows  
     authors  to begin with a simple subset of instructions  and  
     capabilities.    As  authors  become more adept,  they  can  
     learn  additional commands which will give them  additional  
     capabilities  and allow them to develop more  sophisticated  

          Kearsley (1982,  pp. 432-434) proposes,  as an optimum  
     situation, four levels of interaction with the AS software.   
     The  lowest  level  would involve authors  in  a  paradigm-  
     authoring environment, where they would be required only to  
     provide content for pre-established,  fixed templates.   In  
     the  next  level,  authors  would  be  able  to  alter  the  
     parameters  of the established template in relation  to  l)  
     how  information is to be displayed on the screen,  2)  how  
     answers  are  to  be analyzed,  and  3)  how  instructional  
     segments are to be linked together (branching).   The third  
     level  would allow authors to create new templates  through  
     the  use  of high-level verbs.   A final level would  allow  
     authors to control the nature of the AS interaction itself,  
     by opting, for example, to turn off the interrogative input  

     2.3 The Importance of authoring systems

          The current state of CALL is examined and the  serious  
     problem  posed  by  the  continued  lack  of  a  sufficient  
     quantity  of  quality  material  is  explored.   The  terms  
     "quality",  "quantity",  and  "flexibility" are defined and  
     the   various  approaches  to  producing  CALL   courseware  
     outlined  in section 2.2 are then re-examined  with  regard   
     to  these  three factors.  Finally,  authoring systems  are  
     shown  to provide a mechanism for overcoming  the  shortage  
     while offering the potential for quality materials.  

     2.3.1     The current state of CALL

          CALL  has developed slowly since the earliest research  
     in  the  mid-sixties.    The  literature  is  unanimous  in  
     outlining the three major factors which have acted  against  
     the  widespread  use of CALL,  or CAL,  or any use of  high  
     technology in education:  the high cost and contrary nature  
     of  the equipment,   negative teacher  attitudes,  and  the  
     shortage of teaching material.


          The  early epoch of CALL research has been  called  by  
     Paramskas (1984,  p. 14) the period of "monster machines" .  
     CALL   experimentation  was  limited  to  those  few  large  
     institutions  which could absorb the high cost of  computer  
     access.   The equipment of the day,   consisting,  for  the  
     most part, of noisy teletype terminals connected to remote,  
     seemingly  mysterious,  and rather undependable  main-frame  
     computers,  was  considered to be "unfriendly"  (Paramskas,  
     1984,  p. 14).   Language  teachers  were found to  have  a  
     generally negative attitude towards the use of computers in  
     language  teaching  (Fitzgibbon  &  Grate,  1970,  p. 921).      
     Those teachers who decide to investigate the  possibilities  
     of  CALL,  and  whose parent institutions could  afford  to  
     offer  the computing resources,  were faced with a  drastic  
     shortage  of useful and usable  materials  (Orstein,  1970,  
     p. 216). 

           The  complicated  procedure of passwords and  account  
     numbers  of the earlier main-frame epoch has given  way  to  
     the  much more democratic and non-threatening procedure  of  
     simply  switching  on  the  inexpensive  microcomputer  and  
     seeing  it  instantly respond,  "READY"  (Higgins &  Johns,  
     1984, p. 10).  


          With  the  adverse factors of cost and the  unfriendly  
     nature  of the equipment eliminated,  the current  euphoria  
     among  educators would seem to indicate that teachers  have  
     developed a positive attitude, which should lead to a rapid  
     increase  in  the use of CALL.   Such a conclusion  may  be  
     premature  and  illusory,   however,   as  the  course  and  
     direction  of CALL,  indeed,  its very future  in  language  
     teaching,  remains  far from settled.   The desire to  move  
     from   the   experimentation   phase  of  CALL   into   the  
     implementation  phase,  now that many of the other  factors  
     which  inhibited  the use of computers have been  resolved,  
     faces  a major remaining obstacle.

          Experts   are   almost  unanimous  in   warning   that  
     widespread  implementation cannot proceed until the serious  
     shortage of quality teaching materials,  or courseware, for  
     the medium is resolved.   The supply of sound,  substantial  
     CALL courseware is not much greater than existed before the  
     "computer revolution"  and this lack of material represents  
     the  major  barrier  to widespread use  and  acceptance  of  
     computers in language  teaching.   (Crawford,  1981,  p. 3;  
     Davies, 1982, p. vii; Barker & Singh, 1983, p. 174; Holmes,  
     1983, p. 21; Stevens, 1983-a, p 293).

          Centres  wishing  to implement CALL  have  essentially  
     only  two  options:  to  make use  of  existing  commercial  
     materials  or to create new material.   The first option is  
     inadequate,  however.  Commercial sources offer only a very  
     limited  number of computer programs for language  learning  
     (Kenning and Kenning,  1983,  p. 10).  What is available is  
     often  of  mediocre quality and even what  "good"  material  
     there  is  frequently requires  adaptation  (Wright,  1980,  
     p. 11).  The  list  of available material is  even  further  
     reduced  by  the inability of most programs to run  on  any  
     computer  except  the one on which they were  designed  and  
     programmed on,  a situation likened to the "Tower of Babel"  
     (Wyatt,   1983-b,   p. 7).   Wyatt  thus  concludes,  "With  
     microcomputer-based  CAI,   it  is  very  likely  that   an  
     institution   will  have  to  spend  at  least  some   time  
     developing its own courseware"  (Wyatt, 1983-b, p. 7)

          The  local creation of courseware by teachers is  also  
     seen  as  a  necessary  factor in  maintaining  a  positive  
     attitude  towards  CALL  among  language   teachers.   Many  
     language  teachers  have  felt  left out  by  the  computer  
     revolution,  a  situation that will be put right only  when  
     teachers  are  given  the  tools  with  which  to   exploit  
     computers on their own (Higgins & Johns, 1984, p. 10).

          Often  cited  among  CALL experts  as  a  "worst-case"  
     scenario  is  what can be referred to as the "Language  Lab  
     Analogy" (Davies,  1982,  p. vii;  Alatis, 1983, pp. 11-12;  
     Kenning & Kenning,  1983,  p. 1;  Holmes, 1983-b; Sanders &  
     Kenner,  1983-a,  p. 33; Higgins & Johns, 1984, pp. 11-12).   
     It is feared that,  as with the language lab of the  1950's  
     and  1960's,   the  current  wave  of  optimism  about  the  
     possibilities  of CALL will result in educators  attempting  
     to  make  heavy use of computers in the  language  teaching  
     curriculum.   Results will be disappointing, not because of  
     any  basic shortcoming in CALL as a medium,  but because of  
     the  lack of quality courseware to use with  the  computers  
     and  because of teachers' inability to participate actively  
     in the creation of "home-grown" material.    Weible  (1983,  
     p. 64)  points out that language labs began to experience a  
     rebirth  in  popularity  once  cassette  recorders   became  
     familiar  everyday  conveniences  and teachers  could  make  
     their own recordings.  Unless a similar development  occurs  
     in  CALL,   disillusionment  may  set in and CALL  will  be  
     discredited as a force in language teaching. 

     2.3.2   Factors affecting the quantity and quality of  CALL  

          There is a measure of agreement that the resolution of  
     the  coureware shortage problem involves two  factors,  the  
     quantity  of courseware and the quality of that courseware.   
     These  terms,  "quantity"  and  "quality",  are  used  very  
     loosely  in  the literature and  need to  be  more  closely  
     examined,  and  their  role in the present discussion  more  
     rigorously defined. 

          The  quantity  of production of CALL courseware  is  a  
     function of the cost,  in terms of time and  resources,  of  
     designing  and  programming CALL material (Barker &  Singh,  
     1983, p. 174).  Figures of anywhere from 50 to 500 hours of  
     development-time  per hour of finished product using a  GPL  
     approach are cited.  (Braun,  1973,  p. 2;  Kearsley, 1982,  
     p. 430;  Ashmore,  1983,  p. 6).   It  would seem that  any  
     method  of  speeding up the development  process  would  be  
     laudable, providing quality can be maintained.

         Speeding-up the production of material,  however, often  
     affects  quality adversely.   Gillespie (1983,  p. 6)  sees  
     this  relationship as a direct one,  arguing that the  more  
     time and and money available for production,  the better is  
     the quality of the product. Orstein (1970, p. 215) counters  
     this view by citing the millions of dollars that government  
     and private foundations poured into massive CAI projects in  
     the late sixties,  not always yielding a positive effect on  

          The relationship between quantity and quality must  be  
     more complex than Gillespie indicates above.   The question  
     of  speeding-up or facilitating the development process can  
     relate  to  either speeding-up the design  process  or  the  
     coding  process  or both.   Each expedient will  inevitably  
     affect the final quality of the product.

          It  was  pointed out earlier that CALL  courseware  is  
     really a combination of two separate factors:  the logic of  
     the  program,  which can be called the "vehicle",  and  the  
     content of the lesson.   The two are often confused.   Dean  
     (1978,  p. 20) points out that,  despite the fact that over  
     90%  of  a  CAI author's time and  effort  is  expended  in  
     activities   which   are  unrelated  to   computer   coding  
     procedures, the question "How do I write a CAI lesson ?" is  
     often  interpreted to mean "How do I code a lesson into the  
     computer ?"  


          Discussions of quality must take into account both the  
     quality  of  the  pedagogical content and the  quality  and  
     sophistication of the vehicle.  (Holmes,  1982, p.12).  The  
     interrelationship of these factors is addressed below. 

          Two important questions arise in any discussion of the  
     quality of pedagogical content:   1) Who should design  the  
     material,  regular  teachers or those specially trained  in  
     CALL  techniques?  and  2) To what degree can CALL  can  be  
     adapted to current language teaching philosophies.

          One  school of thought holds that effective  materials  
     cannot  be created by teachers alone,  that people  trained  
     in CALL techniques must be  involved in the design  process  
     (Edwards & Tillman,  1982, p. 20; Stevens, 1983-a, p. 294).    
     The  skills  required to generate effective courseware  are  
     said  to be different from those required of  an  effective  
     teacher (Joiner,  1982, p. 13) and   failure to provide the  
     necessary  expertise  results in the sort of  unsound  CALL  
     material widely found today (Dimas, 1978, p. 27).

          Teachers'  difficulties  stem from their inability  to  
     fully  appreciate  the  possibilities of  the  new  medium.   
     Computers represent a level of complexity far greater  than  
     that  found  in devices such as tape recorders,  for  which  
     teachers  can  often prepare a good lesson  in  single  day  
     (Marti,  1981,  p. 26).   Many teachers, as authors,  still  
     adhere  to  the approaches of these  other,  more  familiar  
     media,  producing  lessons  that tend to look like  lessons  
     prepared  for presentation in textbooks or  on  blackboards  
     (Stevens, 1983-b, p. 28). Teachers fail to fully appreciate  
     or  exploit the power of the computer:  they  are unable to  
     see beyond a linear sequencing of material and tend to want  
     students  to complete all of the material  provided  (Hart,  
     1981, p. 14).   Stevens (1983-b, p. 28) finds the situation  
     similar  to  the phenomenon exhibited in early  educational  
     films,  where  teachers  were shown standing  in  front  of  

          Adherents of the opposing viewpoint dispute the notion  
     that  professionally  produced  commercial  materials   are  
     necessarily  any  better than  teacher-produced  materials,  
     even  going  so far as to claim that the latter  are  often  
     superior in quality (Crawford,  1981,  p. 16).    According  
     to  this thesis,  only experienced language teachers can be  
     relied  upon  not  to make pedagogical errors  as  well  as  
     errors in actual language usage (Allen, 1972, p. 349).   It  
     is  therefore  probably easier to teach  language  teachers  
     about  CALL  than  it  is to  teach  CALL  "experts"  about  
     language teaching (Wilson, 1982).

          The  training  of  teachers  to  understand  and  take  
     advantage  of  the full range of possibilities  offered  by  
     CALL   does not relate directly to the present  discussion.   
     Nor  does  the  ability of an authoring  system   to  guide  
     teachers  through  the  authoring process  and  to  provide  
     assistance as to the various CALL options available at each  
     point.   Both of these factors are important, however.  The  
     former  is clearly one of the most important  variables  in  
     any  discussion of the preparation of  quality  courseware,  
     while  the  latter  may be an important  criterion  in  the  
     selection  of an authoring system for teachers not  already  
     familiar with CALL.

          Kenning and Kenning (1983,  p. 5) do not consider that  
     there is any incompatibilty between CALL,  in general,  and  
     the present emphasis on meaningful,  communicative practice  
     which   forms  the  basis  of  current  ideas  of   quality  
     courseware.    If the ability of teachers,  in general,  to  
     fill  the  role  of  qualified,  knowledgeable  authors  is  
     assumed, then the remaining variable is the degree to which  
     the quality and sophistication of the software vehicle used  
     to  create CALL lesson material,   the authoring mechanism,  
     affects  and  colours  the  nature  and  quality  of   that  

           The  term  most often used in describing the  quality  
     and   sophistication   of   a   authoring   mechanisms   is  
     "flexibility",  which  can  be  defined as "the  number  of  
     choices available to the instructor and the learner" (Paul,  
     1982, p. 11).

          Again,  there  are alternate schools of thought as  to  
     the  effect  of flexibility.   Merrill  (1982,  p. 77)  and  
     Kearsley (1982, p. 435) believe that the flexibility of the  
     vehicle determines the quality of the material it presents,  
     while  Holmes (1982,  p. 49) and Wyatt (1983-b, p. 10) hold  
     that  flexibility need only have a minor effect on quality.   
     The  latter hold that valid,  effective courseware  can  be  
     produced  using  very simple,  unsophisticated  mechanisms;  
     that  it  is  not necessary to exploit the  full  range  of  
     features available on the computer. 


          Both  arguments have some degree of merit,  but  there  
     are  subtle,  confounding factors which must be taken  into  
     account.  Authors with a deal of imagination and skill will  
     be  able  to  create excellent  activities  with  even  the  
     simplest,  least flexible vehicle while other,  less gifted  
     authors  may  allow a poorly designed vehicle to impose  or  
     encourage  poor lesson design  (Joiner,  1982,  p. 52).   A  
     highly  flexible  authoring  mechanism  allows  for  higher  
     quality  material,  without  being  able  to  guarantee  it  
     (Joiner,  1982,  p. 52;  Paul,  1982,  p. 11).   The highly  
     sophisticated system may provide options which many authors  
     may  not  know  how to take  advantage  of  (Joiner,  1982,  
     p. 52). 

     2.3.3     The quality and quantity of courseware production  
     as a function of the approach to authoring

          Of  the  three major approaches to the programming  of  
     CALL  courseware,   1)  using general  purpose  programming  
     languages (GPLs),  2)  using authoring languages (ALs), and   
     3)   using authoring systems (ASs),  only the latter  shows  
     promise  in providing a mechanism for the production  of  a  
     sufficient  quantity  of  courseware of  sufficiently  high   
     quality  to  be accepted in the  field.    The  two  former  
     approaches,  which  have  been in use since the  beginning,  
     have failed on one or both counts.   Production using general purpose languages (GPLs)

          While original programming in a GPL is the most  time-  
     consuming  approach  to the coding of educational  content,  
     the flexibility offered by this approach makes it the  most  
     popular (Stevens,  1983-a,  p. 294;  Wyatt,  1983-a, p. 38)  
     and,  as  has been shown,  it accounts for most of the CALL  
     material produced to date. 

           Productions  using this approach fall into any one of  
     four main  categories:    1) Courseware conceived by  those  
     trained  in  programming,  but without a language  teaching  
     background;   2) Courseware  conceived  and  programmed  by  
     language  teachers  newly  trained  in  the  rudiments   of  
     computer  programming techniques;   3) Courseware conceived  
     by  language  teaching  professionals  and  programmed   by  
     experienced   programmers  (the  "team"   approach);    and   
     4) Courseware   conceived  and  programmed  by  a  language  
     teaching  professional  who  was,  at  the  same  time,  an  
     experienced programmer (the "great man" approach).

          There  is little debate among CALL professionals about  
     the  need  to  exclude materials  designers  of  the  first  
     category,  that is,  those who have a knowledge of computer  
     programming   alone  and  expertise  in  neither   language  
     teaching  nor  CALL  methodology  (Davies,  1982,  p. vii).   
     Productions  realized  in  this  category,   unfortunately,  
     account  for a good deal of what is commercially  available  
     and  contributes to the shortage of "quality" courseware by  
     crowding  out better  material  (Stevens,  1983-a,  p. 294;  
     Gillespie,  1983, p. 7).  On the basis of the discussion of  
     the  unique nature of CALL (in section 2.1.2 above) and the  
     discussion of who should design CALL courseware (in section  
     2.3.2  above),  those prospective courseware designers  who  
     are  experts  in the field of CAL alone,  but who  are  not  
     experienced  in language teaching must be included  in  the  
     present category.

          The  recent  exaggerated  claims  for  the   "computer  
     revolution"   in  language  teaching  has  awakened   among  
     teachers  the  desire  to become involved in CALL  and  has  
     given great impetus to the second approach, the programming  
     of  courseware  by language teachers newly trained  in  the  
     rudiments  of computer programming.   Proponents   of  this  
     approach  argue  convincingly at  teaching  workshops  that  
     teachers  can  learn  in  a few  hours  all  the  computer- 
     programming  skills  they need in order to  produce  simple  
     language  activities.   Many  of  those  who  venture  into  
     programming language activities find,  however,  that it is  
     much more difficult to produce acceptable results than they  
     had been led to believe.  

          There  is  no doubt that GPL  programming  offers,  in  
     theory at least,  the greatest flexiblity in designing CALL  
     materials  (Wyatt,  1983-a,  p. 35).   The only constraints  
     limiting  the  possible  design  of  activities  are  those  
     imposed  by the computer,  the computer language,  and  the  
     programming  abilities  of  the  designers  (Holmes,  1983,  
     p. 27).   While the constraints imposed by the computer are  
     most  often  blamed  by amateur authors  for  the  ultimate  
     shortcomings   of  their  material,   their  own  lack   of  
     programming expertise is usually the prime limiting  factor  
     (Hazen, 1982, p. 22). 

          Davies  (1982,   p. 44)  notes  that  programming   is  
     essentially  a  creative art in which programmers can  give  
     free  rein to their imagination.   The key word  is  "art".   
     Knowing  how  to paint or how to play an instrument is  not  
     the  same  as being a gifted  painter  or  musician.   Many  
     experts  agree that GPL programming is only as flexible  as  
     the  creativity  and  skill possessed by  the  practitioner  
     (Kearsley,  1982,  p. 429;  Otto &  Pusack,  1983,  p. 26).   
     Language  teaching  is  sufficiently demanding that  it  is  
     impractical  for teachers,  en masse,  to seek  to   become  
     accomplished  computer  programmers as well  (Pogue,  1980,  
     p. 64; Merrill, 1982. p. 70). 

          The second approach to GPL programming has thus failed  
     to  produce  either  the quantity or the  quality  of  CALL  
     courseware  required.   Not  enough  teachers  have  become  
     involved  in  authoring materials,  because of the  immense  
     learning  burden  of GPL programming.   Of  those  who  did  
     succeed   in   learning   to   program,   their   lack   of  
     sophistication  in programming has had a negative effect on  
     the quality of their materials.

          The  approach  to  GPL programming  favoured  by  most  
     experts  is  the   "team"  approach,   where  there  is   a  
     collaboration  among specialists,  each providing input  in  
     their respective area of expertise (Crawford,  1981, p. 15;  
     Marty,  1981,  pp 43-44,  Mydlarski,  1981, p. 119; Davies,  
     1982,  p. 44).  The team approach to courseware development  
     can involve as little as a faculty member designing content  
     material  and having it programmed by a student  programmer  
     (McCambridge,  1982).    The  process  may include a  third  
     party,  the educational technologist or CALL expert (McGee,  

          Among  the  courseware  accepted as  being  of  higher  
     quality  is  much  that  has been  produced  using  a  team  
     approach.    The  Clef series,  a set of beginner's  French  
     grammar  lessons,  was  produced  through  a  collaboration  
     between  teachers  and  professional  programmers  (Holmes,  
     1981),  as  was the Grammar and Vocabulary  Mastery  series  
     produced  by the American Language Academy  (Wyatt,  1982).  
     Similarly,   the  production  of  courseware  at  Concordia  
     University's  Learning Laboratories has been made  possible  
     through a collaboration of interested teachers, who develop  
     the content, and the Learning Laboratory staff, who design,  
     code,  and  test  the computer program (Kenner &  Richards,  

          The   team   approach,   while  responsible  for   the  
     production  of  some  quality  courseware,  has  failed  to  
     satisfy  the demand for the delivery of quality  courseware  
     in massive quantities.  The most severe disadvantage is the  
     high cost of hiring professional programmers (Pogue,  1980,  
     p. 64;  Kenning & Kenning,  1983,  p. 145).   In situations  
     where programming resources are minimal,  faculty may  have  
     to  concentrate  on less sophisticated CALL  (Dimas,  1978,  
     p.27).   In the case of Concordia  University,  instructors  
     may  have  to  resign  themselves to a very  slow  rate  of  
     production  and  long  delays between  the  design  of  the  
     content  and  its  subsequent  realization  as  a  computer  

          Complications  in the "team" approach arise out of the  
     need  for the various parties to communicate when each  may  
     know little or nothing of each other's field of  competence  
     (Kenning and Kenning,  1983, p. 145).  Elliot (1982, pp. 6- 
     8)  does not believe such a situation can produce effective  
     results,  unless  each party is willing and able  to  learn  
     about the discipline of the other. For teacher - programmer  
     collaboration  to succeed,  teachers must know enough about  
     programming  to understand what level of performance is  to  
     be  expected  from   programmers  and  to  communicate   to  
     programmers  their  requirements  on  content  and  format.  
     Without  the added expertise of a CALL specialist to act as  
     intermediary,  the team approach begins to suffer from some  
     of  the same constraints as those noted above for  teachers  
     learning to program on their own.

          A  few  institutions are extremely fortunate  in  that  
     they have individuals who are,  at the same time,  language  
     teaching  experts,   experts  in  CALL   methodology,   and  
     accomplished programmers.   Dimas (1978,  p. 27)  refers to  
     this  as  the  "great man" approach and it  represents  the  
     fourth option to GPL programming cited at the beginning  of  
     this  section.   Barker and Singh (1982,  p. 169) note that  
     this  enviable combination of cross-disciplinary skills  is  
     extremely rare.   Again, there is no guarantee that what is  
     produced  will be of the quality  desired.  The  courseware  
     will  reflect the particular author's personal  preferences  
     and  it  may very well be found that other members  of  the  
     teaching  faculty  are reluctant to use  it  (Dimas,  1978,  
     p. 27).   Often,  when the "great man" moves on,  or  loses  
     interest  or  motivation,  an institution's CALL  programme  
     collapses and the material passes out of circulation.

          GPL  programming,  then,  despite  the  fact  that  it  
     accounts  for  a majority of what has  been  produced,  has  
     failed  to  satisfy  the double  demand  for  quantity  and  
     quality in courseware.  The first and the second approaches  
     have  led to the production of a large amount of  material,  
     but  often  of  inferior quality.   The  third  and  fourth  
     approaches  have  had limited impact in terms  of  quantity  
     because  of the high cost and/or lack of available experts,  
     still with no guarantee of quality.   Production using authoring languages (ALs)
          Authoring  Languages  (ALs) have been advanced as  the  
     final  solution to the courseware shortage  situation.   It  
     was   suggested  that  teachers  who  knew  nothing   about  
     programming  and little about computers could sit  down  at  
     the  computer and dash off computer-based exercises quickly  
     and with great ease.    It was supposedly no more difficult  
     to  learn to type than it was to learn to use  ALs  (Allen,  
     1972,  p. 348),  yet  they were  to offer a  sophistication  
     equal   to  that  of  material  produced  by   accomplished  
     programmers (Merrill, 1982, p. 76).  ALs were designed with  
     non-computing,   liberal  arts  people  (such  as  language  
     teachers)  in mind (Otto & Pusack,  1983).  ALs were to  be  
     languages   oriented   towards   the  specific   needs   of  
     instructors,  and  were  to  create  an  environment  where  
     individual   instructors   were   freed   from   technical,  
     programming concerns.  (Hazen, 1982, p. 21). 

          The question inevitably arises as to why ALs have  had  
     so little impact on the production of CALL materials, given  
     these   apparent  advantages,   the  disadvantages  of  GPL  
     programming,  and the fact that ALs have been around almost  
     since the inception of CAL. 

              There are several key criticisms of ALs which must  
     be examined.   Universal to all ALs is the observation that  
     they  do  not live up to their rationale of being  easy  to  
     learn  and  use.   Possibly a function of the  specific  AL  
     under discussion is the criticism that ALs place limits  on  
     flexibility,  with  little  apparent gain in  other  areas;   
     that  they  tend to steer authors in the direction of  less  
     effective pedagogical strategies;  and, finally, that there  
     are technical aspects of their operation which limit  their  

          Most  non-programming teachers who have endeavored  to  
     use  ALs  would  not agree with the claim,  often  made  by  
     proponents    of ALs,  that only an hour or two of  initial  
     training  is  needed  before  courseware  can  be  produced  
     (Pogue, 1980, p. 64; Davies, 1982, p. 39).  The claim often  
     refers only to learning the essential operation of the  AL.   
     A  great  deal more learning and practice-time is  required  
     before the user is able to exploit the capabilities of  the  
     language at a moderately sophisticaed level (Wyatt, 1983-a,  
     p. 36).   The  simplicity  and ease of learning  which  was  
     initially  proposed  as  representing one of  the  foremost  
     arguments for using an AL over a GPL (Merrill, 1982, p. 76)  
     has  been traded for continued increases in sophistication.  
     This  has resulted in ALs which are as complex as GPLs  and  
     take  as  long to learn to use  effectively  (Hazen,  1982,  
     p. 22;  Jensen,  1982,  p. 51; Merrill, 1982, p. 76; Wyatt,  
     1983-a, p. 36).

          It  is  beyond  dispute  that it takes  less  time  to  
     produce  CALL  materials  using an AL rather  than  a  GPL.   
     Wyatt (1983-a,  p. 36) warns,  however, against overstating  
     the  speed advantage.   The programming time will be  less,  
     but  of  the same order  of  magnitude.   Stevens  (1983-a,  
     p. 294)  cites 30 hours of programming per hour of finished  
     product  using  an AL,  which can be compared with  the  50  
     hours and upwards cited in section 2.3.2.

          The  claim  often made by proponents that ALs are  for  
     those  who do not know how to program and that  programming  
     using an AL is, somehow, not really computer programming is  
     also  disputed  (Jensen,   1982,  p. 51).   Paloian  (1974)  
     reminds us that, while ALs may facilitate the actual coding  
     of  a  CAL program,  they do not  contribute  much  towards  
     shortening the program-planning stage or the debugging of a  
     program once it has been entered into the computer.   Hardy  
     & Elliot (1982) say of MicroPilot that it "demands the same  
     level  of  ability and skill [as that] required of a  BASIC  
     programmer".   The  difficulty  lies  not so  much  in  the  
     programming language itself,  as in the need to  understand  
     programming  concepts  like  variables,   counters,  flags,  
     matching,  conditional  and unconditional  branching,  etc.  
     (Davies, 1982, p. 39). 

          Indeed, the problem is often hightened by the need, in  
     any case,  to program certain sections of a CALL program in  
     a  GPL.   Many  ALs  allow for the  inclusion  of  routines  
     written  in  the  host language (the computer  language  in  
     which AL was originally written).  The author is thus given  
     powers of operation unforeseen by the designers of the  AL.  
     Merril  finds,  in reviewing a PILOT program,  that  almost  
     half  of the statements were in the host language (Merrill,  
     1982, p. 76).

          The  second major cristicism of ALs is their  lack  of  
     flexibility.   An AL attempts to simplify the author's task  
     by reducing the number and complexity of commands, but "the  
     restriction in the domain of commands creates a restriction  
     in   the  range  of  possible  outcomes  or   applications"  
     (Merrill,  1982,  p. 76).   Teachers,  through  the lack of  
     flexibility  of the given AL in fostering   development  of  
     truly  innovative  activities,  will  either  produce  only  
     simple  exercises  (Wyatt,  1983-a,  p. 37)  or  they  will  
     recognize  the limitations of the AL quickly and will  wish  
     to  move on to GPL programming (Kenning and Kenning,  1983,  
     p. 13).   Davies  (1982,  p. 44)  sees an  analogy  in  the  
     "paint-by-number"  kits  which please the newcomer to  oil- 
     painting  but cease to satisfy the creative needs of  those  
     who  wish  to carry their artistic endeavours to  a  higher  

          The inflexibility builds upon itself.   Teachers  will  
     be  limited by their lack of programming experience and may  
     decide  that  certain  things  that they  wish  to  do  are  
     impossible  on the computer.   This situation will then  be  
     compounded  by  the inability of ALs to offer  avenues  for  
     innovative  teaching  techniques and by their  tendency  to  
     push  teachers towards those pedagogical  strategies  which  
     are  the easiest to  program.  Otto & Pusack (1983,  p. 27)  
     have concluded, by virtue of this  inflexibility,  that ALs  
     are not suitable to most forms of language practice.

          The   tendency  of  the  software  to  influence   the  
     selection   of  pedagogical  strategies   is   inescapable.   
     Inherent  in  any very high level software such as ALs  are  
     the   assumptions  and  decisions  made  by  the   original  
     designers within which (or against which,  as the case  may  
     be)  the user has to work.   Dean (1978,  p. 23) makes  the  
     point   that  even  a  basic  sequence  like  "presentation  
     followed by decision" implies an instructional logic.   ALs  
     are criticized because they lead novice authors to follow a  
     particular,  restrictive  instructional strategy since  the  
     restriced domain of commands makes this strategy easiest to  
     implement (Hazen, 1982, p. 21; Merrill, 1982, p. 76).  

          There  are  technical considerations which  limit  the  
     pedagogical utility of some ALs for CALL applications. Some  
     ALs  have  been observed to have an immense "overhead"  (in  
     terms  of  computer memory and resources)   which  acts  on  
     microcomputers  to trivialize the type of exercise that can  
     be designed for the little remaining room in memory.   This  
     memory waste is compounded by the tendency of ALs to  guide  
     would-be   programmers  away  from  strategies  of   sound,  
     economic  programming such as structured programming or the  
     re-use of generic procedures.   Often only parts of the AL,  
     or  the  program,   or  both,   can  be  loaded  into   the  
     microcomputer  at one time,  leaving students to face  long  
     delays  as  the computer loads data from  external  storage  
     media.  "Students are lulled by the endless whirring of the  
     floppy disk" (Gillespie, 1983, p. 7).

          Another  technical consideration is  portability,  the  
     ability to use material on computers other than the one for  
     which they were designed.  When good courseware is produced  
     using  an  AL,  it  is often not available  to  the  entire  
     language  teaching community because the AL in which it was  
     written  is supported only by certain computers or must  be  
     purchased separately.    This is also, of course, a problem  
     for  GPL programs.    It becomes more acute when an  AL  is  
     involved, because there is less chance of finding a version  
     of  an AL compatible with a given machine than there is  of  
     finding GPL versions (Hazen,  1982,  p. 21;  Otto & Pusack,  
     1983, p. 27).    Production using authoring systems (ASs)

          Many  of the criticisms levelled against ALs can  also  
     be  made  against many  authoring systems (ASs),  with  the  
     notable  exceptions  of  1) the  initial  learning  burden,    
     2) the time taken to produce material,  3) and the need  to  
     understand computer programming. There is general agreement  
     that  ASs offer a dramatic reduction in the time  necessary  
     for  CALL  courseware  development,  both  in  the  initial  
     training of prospective authors (Wyatt,  1983-a, p. 37) and  
     in  the  time needed to produce an exercise  (Pogue,  1980,  
     p. 62).   By definition,  use of an AS does not require  an  
     understanding of  computer programming.

           Merrill  complains  that  ASs "reduce  the  cost  and  
     effort  by reducing variety in much the same way that  cost  
     and  effort  are reduced in  fast-food  restaurants,  tract  
     homes, and formula television shows." (Merrill, 1982, p.77)     
     To  carry  this  analogy further and to relate  it  to  the  
     present discussion,  it can be argued 1) that the above did  
     lead  to  a rapid expansion in the number  of  restuarants,  
     homes,  and television shows and 2) that not all fast  food  
     restuarants,  tract homes,  or formula television shows are  
     uniformly bad.

          While  Merrill  (1982,  p. 77) concedes that  ASs  can  
     enhance  the  quality of courseware produced by novices  to  
     CALL,  he,  and  others (Edwards &  Tillman,  1982,  p. 20;  
     Davies,  1982,  p. 35) nevertheless  hold that ASs restrict  
     the  quality  of the courseware that could be  produced  by  
     truly  creative authors.  Few would dispute Davies'  (1982,  
     p. 35)  observation  that ASs can never  produce  the  same  
     standard  of  CALL software as a team  of expert  teachers,  
     CALL experts, and expert programmers using a GPL.

          The  reasons  why the team approach has not  satisfied  
     the  demand  for  material have  been  explored  above  (in  
     section     Without  such  a  team,   Merrill's  
     "creative  authors"  would  probably  find  the  range   of  
     possibilities in a GPL or AL environment no more satisfying  
     or  unique  than an AS.   The details in  programming  CALL  
     materials tend to overwhelm print-oriented authors.  Rarely  
     do  they have to concern themselves,  when writing a  book,  
     with such issues as paper choice, page design, typesetting,  
     binding,  and  a myriad of other activities involved in the  
     production  of  their work.   The details of  converting  a  
     manuscript  into  a book are left to  the  publishing  firm  
     (Edwards & Tillman, 1982, p. 20).

          After  a  fashion,  an AS provides  analogous  support  
     during  CALL  courseware design.   The degree to which  the  
     support  becomes  a  strait-jacket depends  on  the  AS  in  
     question.   "If  the  system is  reasonably  flexible,  the  
     teacher  will  have a fair degree of latitude on any  given  
     point" (Kenning and Kenning, 1983, p.12).

           Wyatt  (1983-a,  p. 38)  claims  that  ASs  will  not  
     respond  to the demand for quality CALL courseware  because  
     it is impossible to use them for the production of tutorial  
     exercises  or "open-ended or communicative activities of  a  
     collaborative  sort."    This is one of the  key  questions  
     considered in this thesis. 

          Despite  misgivings on the subject,  Wyatt reluctantly  
     concedes that ASs offer the only short-term solution to the  
     courseware shortage dilemma:

          "Although  there  is  no  general  agreement  on  which  
          approach  to  take (to address the courseware  shortage  
          problem),  there does seem to be a significant movement  
          towards  the  adoption  of  authoring  systems...    It  
          appears likely that sound and worthwhile courseware can  
          be  developed in this way at far greater speed and  far  
          less expense."  (Wyatt, 1983-b, p.5)

Converted from the original Wordstar format and placed in Html package
Roger Kenner, 2006
[Return to Top Level]