.    Tha Li Vishu Sl ish Vixan Sl Xwe.







     The Effects of the Formal Features of Language Learning  
        Activities on the Selection of Authoring Systems




              Thesis Presentation By:  Roger Kenner


                    Wednesday, April 9, 1986



     1.   The General Problem:    

          Despite   over  20  years  of  research  into  the  
     applications of computers to language learning and...

          Despite  the  recent advent  of  inexpensive,  yet  
     powerful, microcomputers and...

          Despite  the  widespread  interest  in   computers  
     expressed by language teachers and educators ...

          
          Computer-Assisted Language Learning (CALL) has yet  
     -    to  be  adopted  on  a  widespread  scale  by  language  
     teachers.


          The  absence  of  an adequate  supply  of  QUALITY  
     computer-based   language   learning   materials,    or  
     COURSEWARE,  is  cited  by most authorities as a  major  
     contributing factor.


          The use by language teachers of software  packages  
     called AUTHORING SYSTEMS,  by which teachers can create  
     CALL  materials  quickly and without the need to  delve  
     into  computer programming,  is frequently cited  as  a  
     possible solution to the courseware shortage problem.



     2.   The Specific Problem:


          Before teachers can begin to make use of authoring  
     systems,  they  must  have a means of  SELECTING  those  
     authoring  systems  most  suited to  realising  quality  
     language learning activities.


          Authoring  systems  have  been  examined  in   the  
     literature according to a range of criteria:
          a.   cost
          b.   ease of learning
          c.   ease of use
          d    graphics and sound capabilities
          e.   screen design
          f.   support documentation
          g.   time required to author a package
          h.   etc.

          Such  criteria  are  all  very  important  to  the  
     overall  selection  process,  but they have  no  direct  
     bearing  on  the  utility of the  authoring  system  in  
     producing effective language learning activities.


          Those  features  of authoring systems which  might  
     have  a direct impact on their ability to be  used  for  
     the production of quality materials have not adequately  
     been identified.


          Where  writers and designers of authoring  systems  
     have  specifically  addressed language  learning,  they  
     have  tended,  first,   to focus on the powers  of  the  
     computer,  then  worked  backwards  in  applying  these  
     powers to language learning activities.


          In  this thesis,  I propose the opposite approach:   
     To  begin with existing language  learning  activities,  
     from  more  traditional media,  and to determine  those  
     powers  of  the computer that are necessary  for  their  
     realization.



     1.     Can  a  model  of  existing  language   learning  
     activities  be  constructed  which  will  reveal  those  
     inherent  features  of  the activities  which  must  be  
     considered when planning their computerisation?


     2.      What  is  the range of computer-based  language  
     learning  activities that can theoretically be  created  
     through the use of an authoring system?


     3.       What  special features inherent to  meaningful  
     and  communicative language learning activities  affect  
     their  realisation  through  the use  of  an  authoring  
     system?

          The    writers    hypothesis    is    that     the  
          "meaningful/communicative"   dimension   of   such  
          activities is entirely a function of their content  
          and,    therefore,   will   present   no   special  
          requirements. 



     1.     ...to   establish  those  variable  features  of  
     authoring  systems which are important  in  determining  
     the  range and type of language learning activity  that  
     can be realised.

               Teachers  may  then  use  these  features  as  
     criteria in selecting particular authoring systems


     2.   Given a particular authoring system, teachers will  
     be able to determine types of activities to which it is  
     best suited.


     3.   Teachers will become aware of the "formal" aspects  
     of an activity and of how activities can be modified in  
     order  to  make their realisation on the computer  more  
     practical.


     4.   Developers of authoring systems for CALL will have  
     a  guide  as to the features which must  be  made  more  
     flexible is authoring systems are to be applicable to a  
     wider range of activities.



     1.     The   range  of  possible  present  and   future  
     applications  of the computer to human language and  to  
     language teaching is almost limitless.

            This  range  has  been arbitrarily  narrowed  by  
     rigidly  defining CALL so as to include only  the  more  
     traditional  applications  of the computer to  language  
     teaching.

            The   range   has  been  further   narrowed   by  
     considering only those types of activities which  could  
     theoretically  be  accomplished through the use  of  an  
     authoring system.


     2.       The  question  of  what  particular  types  of  
     activites should,  or should not,  be presented via the  
     computer  has  not been addressed.    Nor has the  more  
     global question of whether certain types of  activities  
     should be done at all been considered.   All activities  
     are considered in a neutral fashion,  without reference  
     to their pedagogical effectiveness.


     3.          In   this  theoretical  study,   no  actual  
     authoring   systems   are  considered   or   evaluated.   
     Reference is made to the features and limitations of  a  
     theoretical, general authoring system.



     1.   Computer-Assisted Language Learning (CALL), in its  
     narrow  sense,  is  defined as a subset  of  "practice- 
     oriented computer-assisted learning (CAL).

          Practice-oriented CAL involves:

     a.   Practice or rehearsal of a given concept or skill

          (Non-practice   activities  include  use  of   the  
          computer as a "tool", such as in word-processing.)

     b.   Interactive  practice:   the computer is an active  
          participant   in   an   immediate   back-and-forth  
          interchange.

     c.   Structured  practice:    the practice takes  place  
          within a pre-established framework.

          (An "adventure" game is structured, ELIZA is not.)

     d.   Mediated  practice:   the computer  evaluates  the  
          learner's performance according to pre-established  
          criteria.  There is a "correct" and an "incorrect"  
          outcome.

          (ELIZA is not mediated.)



     2.   CALL is a unique subset of CAL:

     a.   Evaluating  human-language  presents   programmers  
          severe technical problems.

     b.   Popular CAL approaches,  such as "Generative CAL",  
          are  difficult to accomplish.    All data specific  
          to an activity must be provided in advance by  the  
          author. 

     c.   There   is  freqently  the  need  to  provide  for  
          situations  that yield multiple  correct  answers,  
          varying degrees of acceptability of an answer,  or  
          the  placement  of emphasis on either the form  or  
          the content of the answer.

     d.   Mal-formed  responses  must  be  dealt  with   and  
          analyzed.

     e.   The nature of language learning itself is unique



     3.     "Authoring"    is   the  process  of   producing  
     courseware and involves,  not only the original  design  
     of  the  educational content,  but also its  subsequent  
     realization on the computer.


     4.     Approaches to Authoring:

     a.    creation  of original programmes  using  general- 
     purpose programming languages, such as BASIC.

     b.       programmer's   shortcuts   to   the   original  
     programming of each piece of courseware:

          i.   modular programming
         ii.   programming aids
        iii.   template programs: 

                    The   separation   of    lesson-specific  
                    content  from  the general program  that  
                    delivers it.

         iv.   driver programs:

                    (Almost an authoring system).

     c.    creation of original programmes using  "authoring  
     languages",      programming     languages     designed  
     specifically for the purpose of writing CAL courseware.

     d.    creation of coursware using authoring systems.


     5.   The nature of authoring systems:

     a.   No computer programming is required on the part of  
     the user.

     Typically:

          Teachers  work in an interactive mode.   They  sit  
     before  the  computer  and respond  to  the  software's  
     prompts  as  they  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 supply the material:   stimulus, correct  
     answers,  predictable wrong answers,  etc.,   for  each  
     item.
          The software then makes up the lesson according to  
     the instructions.


     What   can  be  produced  using  an  authoring  system:   
     Anything that can be described formulaicly, independent  
     of content.



     6.   On QUALITY and QUANTITY.


     -    Traditional approaches to authoring have tended to  
     yield two extreme situations:

     a.    Courseware  of good QUALITY,  but in insufficient  
     QUANTITY

     b.    A sufficient QUANTITY of poor QUALITY material.

          Authoring systems seem to provide a mechanism  for  
     increasing  the  QUANTITY  of  production  while  still  
     providing for some measure of QUALITY.


     -    There  is an argument over WHO should produce CALL  
     courseware:

     a.    One  school holds that teachers cannot be  relied  
     upon to produce good materials as they have no training  
     in instructional design,  CAL  methodology,  etc.,  and  
     they  lack the imagination to take advantage of the new  
     medium.
     b.     The other school holds that only teachers can be  
     relied  upon  to  produce  QUALITY  language   teaching  
     materials,    arguing  that  it  is  easier  to   teach  
     computers  to  teachers  than it is to  teach  language  
     teaching to computer experts.
          (Whether or not ...)


     -    There  is an argument over how the quality of  the  
     software  vehicle  will  affect  the  quality  of   the  
     courseware that it can deliver.
          (One of the objects of this thesis).


     
     General Overview:

     
     1.   Drills and drill-type activities pertaining to the  
     practice-phase   of  language  learning  were  seen  to  
     satisfy   the   definition   of   CALL-type   activity:   
     Interactive, structured, and mediated practice.

          ("Drill"  or  "exercise" is seen here  as  a  term  
          referring  to a particular format for practice and  
          divorced from the pedagogical connotations of  the  
          terms)


     2.   The nature of a drill:  

     a.     a   drill   consists  of  discrete   interactive  
     interchanges, or "items"

     b.      each item can be broken down into the following  
     essential stages:

     1.   the presentation of a "stimulus" or "prompt"
     2.   the "response" to the stimulus by the learner
     3.   the "analysis" of that response
     4.   "feedback" to the response
     5.   a decision as to the next item to present.


          In  the classroom,  numbers 3-5 are  normally  the  
     domain  of the teacher.    In CALL,  the computer  must  
     assume these three tasks.

          the  prompt or stimulus can be further broken down  
     into the "model",  the piece of language upon which the  
     learner will be working,  and the "cue", the element or  
     instruction which tells the learner how the model is to  
     be used or modified.

          The basic pattern of  MODEL, CUE, and RESPONSE was  
     used  as  a common framework into which  to  place  the  
     original material considered.



     3.   Examples were taken from drill typologies produced  
     by a number of accepted specialists, spanning a l4 year  
     period  and  representing a range of langauge  teaching  
     philosophies  (Overhead).    127 original classroom and  
     language  laboratory  activities were included  in  the  
     sample.  


     4.     The   drill   typologies   consulted   organized  
     activities   according   to  two   principal   systems:   
     categorization  by  "operational" distinctions  and  by  
     "pedagogical" distinctions. 

            The former group was used in the construction of  
     the  model.   The model was then applied to the  latter  
     group.


     5.      Traditional   categorizations   according    to  
     operational  type focus on the relationship between the  
     stimulus and the response:  How the stimulus yields the  
     response  i.e   the action of the learner   (Overhead).   
     The  categories thus established are fairly  universal,  
     but  are  of  little immediate use  in  discussing  the  
     computerisation of the activities.


     6.      A new categorization was established,  based on  
     Cook's  analysis of drills in terms of the relationship  
     between  successive stimuli and  successive  responses.   
     This  approach  focuses on the formulaic  nature  of  a  
     drill,  and more closely describes the presenter's task  
     rather than on the student's task.


     7.      "model"  is now re-defined as "response-model",  
     the basic framework for the response.  
             Four   fundamental  types  of  operation   were  
     identified:

          Insertion of an element into the model.
          Removal of an element from the model.  
          Re-order of model elements.
          Production of a new response model,  based on  the  
     cue alone.

          "Simple"  drills were those which involved only  a  
     single  operation.   "Complex" drills were those  which  
     involved more than one single operation.

          The "operational" sample was categorized according  
     to this scheme.     8           Two  other  factors  considered   were   1)   
     variations in the manner in which the cue was presented  
     and  2) whether insertions were made into a fixed or  a  
     variable  "slot" in the model.    The pattern in  these  
     factors    was    consistent    with    the    original  
     categorization.


     9.         The  manner  in which the cue  and  response  
     model  might be presented on the computer was  examined  
     and  formulas  were  developed to describe  each  drill  
     example.    The   mechanisms  described   were   fairly  
     consistent with the categorization.


     9a.        In  designing  the formulas,  a new  factor,  
     response-elicitation mechanism, was discussed:

          a.   fill-in
          b.   fill-in = multiple choice
          c.   simple multiple choice
          d.   re-order multiple choice
          e.   holistic multiple choice
          f.   free response

          The   REM   serves   as  the  link   between   the  
     presentation of a drill and the analysis of its answer.


     10.         Factors  affecting  the  analysis  of   the  
     learner's response were next considered:

          a.   Response Set:   The set of expectable correct  
     and incorrect answers.
          b.   Answer Set:    The set of correct answers
          c.    Response  Possibilities,  a function of  the  
     REM.   (Yes/No example).

     11.         Formulas  describing  the  answer  analysis  
     required  for  all  the  drills  in  the  sample   were  
     postulated.  


     12.        The  effect of the presentation and  answer- 
     analysis    phases   of   the   drills   on   Branching  
     possibilities,  the  decision  as to the next  item  to  
     present to the learner,  was considered.  "interlocked"  
     drills  were  seen to pose a problem.   The  degree  of  
     sublety  in branching possibilities was constrained  by  
     the sublety of the answer analysis.



     13.    The   pedagogical  sample  was  then   examined.   
     Attempts  to  arrive  at  a  universal  definition   of  
     "meaningful"    and    "communicative"   drills    were  
     unsuccessful.

     Nine   discrete  factors  contributing  to  a   drill's  
     "meaningfullness" were postulated:

     1.   Realistic Interchange
     2.   Extended Context
     3.   Truth Value
     4.   Information Gap
     5.   Function practice
     6.   More than one response
     7.   Lexical understanding
     8.    Elicitation  of new,  creative  elements,  within  
     constraints
     9.    Elicitation of new,  creative  elements,  without  
     constraints.

          In applying these factors,  it was often necessary  
     to  use the examples offered rather than the discussion  
     provided by the experts.

          Various  tables  compare the various  factors  and  
     they  appear,  for the most part,  to  be  independent.   
     There  is  indication,  however,  that Truth Value  and  
     Information  Gap represent the same factor  and  depend  
     only on the direction of information flow.

          It   appears   from  the  findings  that   Lexical  
     Understanding   or   Realistic  Interchange   are   not  
     sufficient to make an activity "meaningful".

     14.    With the non-meaningful drills removed from  the  
     pedagogical  sample,   it was analyzed according to the  
     model.

     a.    No  new operational categories or templates  were  
     required  to  describe  all the drills  in  the  second  
     sample.

     b.    Removal and Re-order operations, however, did not  
     figure prominently.   Operations were either insertions  
     or production of new response-models.

     c.    Other than that,  there was no consistent pattern  
     based  on  pedagogical factors,  even when each  factor  
     was examined indepedently.


     

     1.     Can  a  model  of  existing  language   learning  
     activities  be  constructed  which  will  reveal  those  
     inherent  features  of  the activities  which  must  be  
     considered when planning their computerisation?


          Yes,  such  a model has been inductively  created.   
     It  is internally consistent and has been applied to  a  
     second sample without modification.

     2.      What  is  the range of computer-based  language  
     learning  activities that can theoretically be  created  
     through the use of an authoring system?

          The  model describes formulaicly the  presentation  
     of an activity.    What can be formulaicly described as  
     a standard template,  can theoretically be produced  by  
     an  authoring  system.    The  model  identifies  those  
     activities which cannot be so accomplished.


     3.       What  special features inherent to  meaningful  
     and  communicative language learning activities  affect  
     their  realisation  through  the use  of  an  authoring  
     system?

          "Meaningful"   activities   do  not  differ   from  
     "meaningless"   ones  insofar  as   their   operational  
     characteristics are concerned.
          The  Quality  of  the software vehicle  is  not  a  
     significant  negative  factor affecting the quality  of  
     CALL courseware.
          The  quality of CALL courseware is a  function  of  
     the imagination and training of the teacher/author.


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