Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An intelligent management information system as a test of a formal accounting theory Clarke, Bevan J. 1975

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1975_A1 C54_1.pdf [ 22.02MB ]
Metadata
JSON: 831-1.0100050.json
JSON-LD: 831-1.0100050-ld.json
RDF/XML (Pretty): 831-1.0100050-rdf.xml
RDF/JSON: 831-1.0100050-rdf.json
Turtle: 831-1.0100050-turtle.txt
N-Triples: 831-1.0100050-rdf-ntriples.txt
Original Record: 831-1.0100050-source.json
Full Text
831-1.0100050-fulltext.txt
Citation
831-1.0100050.ris

Full Text

AN  INTELLIGENT  MANAGEMENT I N F O R M A T I O N SYS TE E FORMAL ACCOUNTING TEEOBY.  AS  A TEST  By  M.Coro.,  A  thesis the  Bevan  J .  University  cf  submitted  in  reguireuerts Doctor  in  Clarke Canterbury,  p a r t i a l fcr  of  the  1964.  fulfilment degree  cf  cf  Philosophy  the  division of  Accounting  and  in  We  The  University  cf  Information  Faculty  Commerce  the  accept to  Management  this  the  British  of  thesis  required  Columbia  as  Systems  conforming  standard,  May  1975.  OF  A  In  presenting  an  advanced  the I  Library  further  for  degree shall  agree  scholarly  by  his  of  this  written  thesis  in  at  University  make that  it  thesis  for  partial  freely  permission  purposes  may It  is  fulfilment  of  of  Columbia,  British  available for  by  the  understood  gain  for  extensive  be g r a n t e d  financial  shall  Head  be  requirements  reference copying  that  not  the  of  of my  of  University  Compare* of  British  fcColumbia  I  agree  and this  copying  or  allowed  without  /^YMAiAah^  for that  study. thesis  Department  permission.  2075 Wesbrook P l a c e V a n c o u v e r , Canada V6T 1W5  Date  the  representatives.  Depa r t m e n t The  this  or  publication my  i  ABSTRACT An  'automated  implemented. and  updating  as  input  into the  so-called  that  i f  r e l i a b l y  MIS  in  design  very of  circumstances  £ e r se:  and  complex  such  i s  a  focus  around  an  an  their  system,  computer  itself..  i t  will  i t s  and  i t s ultimate  the IRIS  of  A r t i f i c i a l  objectives  nethodclc_g2  the iflstrumental  of  the  formalisms  axiomatic  construction.  Instrumental  appear  f o r extensions  the  to c a l l  direction  of  delegated  method  employed  necessary t c  to s c i e n t i f i c logics.  via the by,  the  use the  Intelligence  tc i t .  we  s c i e n t i f i c such  for  changing  would  sciences.:  of  disciplines,  ncn-declarative  be  largely  fcrm  t o p management  merits  become  directed  to  cf  are tc  cf  consider  the organization  and inductive  the  of  research  We  adaptation  cf  cf  System.  deductive  The  we  'instrumental  t o t h e development  software  systems  a l l to be i n t e g r a t e d  (fc)  avenues  application  Information  computer  pursue  which  Management  three the  given  i d i c i a t i c  'Intelligent  calls  i s  entries  i s  ordinary  heading  and t h e management  In  a  It  in  Intelligence  Management  computers  accounting  the world.  dialogue  this  bookkeeper'  They a r e  A r t i f i c i a l  computerized  cf  Mattessich  are u n i f i e d .  The I n t e l l i g e n t  'robct  deducing  an IMIS,  Under  what  of  project  of  System'. into  sciences'  the  empirical  cr  model  or  the possibility  research  a  capable  narrative  This  Information  ( )  i s  an a c c o u n t i n g  a  English. study  This  accountant'  as  logic,  review thecry  accountancy, nctably  in  i i  We  (c) cf  a  e^ajliijjjl formal  accounting the  the  theory,  the cry.  enspiritiny  that  when  the  to  At  deduce  theory  interesting  i n i t s cv>n  accounting  a  that  tc  impleinentatle Systeirs.  nay  as  ar  te  theoretic  fce  \ia a  resent t l a n c e  tc  theoretical cuidirg in a  fcurdaticn  his a>icns  i t iray  attune ted  te  usee  computer  W a t t e s s i c 1' s  the  c f ar. PIS  acccuntinc krcvledge  nay  c f an  style  f c r Wacagenert  i s  nanec€ff€ct  a cere  similar  as  find  iefitaticn  since  test  exicnatized  ccrclusicns  as  cf a  icbct.  where  regarded  the  fait  v-i e n p l c j  Fcrsover,  elaboration  as  Ijiri's  point  closer  i f  this  the  ticr  acccirtirg  the  right.  implemented  lead  cur  Such  successfully may  cf  tine  acccurtitc  inferroatier  i t i s suggested  sane  tc  bears  theory.  ,  the  knowledge  axiomatized  of  cf i l i i i j e r t a specifically  elaborated  successfully prcgrair  vajue  ci  te IRIS ar  Information  i i i  TABLE  OF  I N T R O D U C T I O N AND O B J E C T I V E S . OTT SpecifIc~objectives 0.2 Comments upon t h e s e 0.3 Four caveats. 0.4 Outline.  CONTENTS.  1 5 6 12 15  of the r e s e a r c h . objectives  C H A P T E R ONE:. I n t e l l i g e n c e knowledge and a r t i f i c i a l in t e l licence.. 1.1 1.2 1.3 1.3.1 1.3.2 1.4  18  1.4.4 1.4.5 1.4.6  Intelligence. A r t i f i c i a l intelligence. Knowledge i n e p i s t e m c l o g y and i n A l . . . . . . . . . . . What i s "knowledge"? The e p i s t e m o l c g i s t s " k n o w l e d g e " A r t i f i c i a l Intelligence commentaries upon knowledge. Relational descriptions: networks and g r a p h s . . Procedural descriptions The value of semantic models in aiding deductive inferences Actions as t r u t h s . Inductive l o g i c as p r a c t i c e d ty a machine. . . . . Structures for organizing knowledge. . . . . . . . . . .  1.5 1.6  Examples of Al r e s e a r c h . The languages of A r t i f i c i a l  1.6.1 1.6.2 1.6.3  An i n t r o d u c t i o n t c L I S P Very high l e v e l languages for The s i g n i f i c a n c e of very high  1.4.1 1.4.2 1.4.3  1  C H A P T E R TWO.: T h e m e t h o d o l o j g v c f t h e and a x i o f a t i z e d a c c o u n t i n g t h e o r y  Intelligence  40 43 50 58 64 67 72  81 .......102  theorem-proving. level languages.  instrumental  21 24 30 30 34  102 111 132  sciences 134  2.1 2.2  The p o v e r t y Theory: the  2.2.1  Testing theories for truth and instrumental theories for u t i l i t y . . . . . . . . . . . 99 A t h e o r y as t h e d e f i n i t i o n o f i t s label .165 A n a l y t i c d e f i n i t i o n s of "accounting" and "MIS" ..170  2.3 2.3.1 2.4 2.4.1 2.4.2  of theory in accounting a n d MIS h y p o t h e t i c o - d e d u c t i v e method.  Implementation as a t e s t of an i n s t r u m e n t a l t h e o r y . Nested t h e o r i e s and domains T e s t s i g n a l s as s u r r o g a t e domains for I j i r i ' s theory.  ....136 .....139  ...193 . . . . . . . . . . . 1 9 7 200  iv CHAPTER  THREE:,  The i n t e l l i g e n t  s j s t e i i 3.1 3.2 3.3 3.4 3.4.1 3.4.2 3.4.3 3.4.4 3.5 CHAPTER  Iiii~Iis~to  management  information  IMIS  210  T h e " t r a d i t i o n a l " . MIS A new m o d e l o f t h e M I S S i x f a c u l t i e s s u f f i c i e n t f o r an i n t e l l i g e n t , goal-seeking entity F o u r l e v e l s o f e x p e r i m e n t a l IMIS KIWI ONE - t h e a c c o u n t i n g d e d u c t i o n  ...210 .220 223 ......234 system. ...236  K I W I TWO - a d d i n g natural language understanding. KIWI THREE - a s i m p l e m a n a g e r i a l model. K I W I FOUR - a g o a l - d i r e c t e d s y s t e m Paradigm shift or incremental growth? FOUR.:  Implementing  a simple  ........241 .......248 254 . . . . . . . . . 2 5 4  IMIS  261  4.1 4.2 4.3 4.4 4.5 4.5.1 4. 5.2  The narrow o b j e c t i v e : accounting deduction. ...263 Preference f o r a procedural notation .....274 T h e b r o a d e r o b j e c t i v e : a modest IMIS 277 Further sets cf test signals. . . . . . . . . . . . . . . . . . 278 Overview o f methods. ...283 Three stages of analysis. 284 E x t e n s i b i l i t y 286  4.5.3 4.6 4.6.1  The flow of c o n t r o l . Data s t r u c t u r e s , Notations, f o r p r i m i t i v e s  4.6.2 4.6.3 4.6.4 4.7 4.7.1 4.7.2  Overall data structures. A p r i o r i knowledge Expressing accounting relationships. Natural language analysis of the signal Syntax: parsing English w i t h a n ATN grammar. Adding the notations of bookkeeping & mathematics. Procedural semantics: canonical verbs & case-frames.  4.7.3  8 relationships  287 292 292 .297 ....301 .304 314 ..314 324 326  4.7.4 4.8  The semantic lexicon Applying the semantic  4.8.1  Asserting  4.8.2  signal i n t o SIGFACTS 348 The problems of implied p r i o r assertions, look-ahead S concept recognition. . . . . . . . . . . . 358  4.9 4.9.1 4.9.2  U p d a t i n g t h e model o f t h e w o r l d . The updating method. TIMELINE: running time forward and backward.  4.9.3 4.9.4 4.10  SWG a f t e r t h e f i r s t t h r e e The "Accounts". Some v a r i a n t signals.  functions.  the propositions  implied  signals.  339 . . . . . . . . . . . . . . 3 4 8 by t h e  373 ....373 ..377 380 383 387  V  CHAPTER  FIVEj. C o n c l u s i o n s §L M § theories^_  testing 3 93  5.1 5.1.1 5.1.2  Implementation as a test The method o f t e s t i n g The results obtained.  5.1.3  The r e v i s e d theory of d o u b l e - c l a s s i f i c a t i o n a l  5.1.4  Comparison  5.2  T h e IMIS  CHAPTER 6.1 6.2 6.3 6.4  SIX:  with  a  theory  accounting.  Hattessich's  theory.  concept.  Directions  Methodologic A r t i f i c i a l i Problems of Problems of  BIBLIOGRAPHY^  of  f o r  394 397 ....398 ...407 . . . . . . . . . . 423 4 26  future  research.  al problems. n t e l l i g e n c e problems. accountancy. h i g h e r - o r d e r IMIS.  427 427 . . . . . . . . . . . . . 433 ...434 437  4 49  APPENDICES ~ ~I"l Glossary. ....469 A2 Natural language understanding: W i n o g r a d ' s SHRDLU. 475 A3 Epistemclogy as an e x p e r i m e n t a l s c i e n c e : E x t r a c t from " T h e understanding o f t h e t r a i n " by Eccles. . . . . . . . . . . 3 3 5 A4 A d e f i n i t i o n of syntax: A Conversational Accounting language. 493 A5 T h e f o r m a t o f s i g n a l s i n p u t t c K I W I TWO 496 A6 L i s t i n g o f K I W I TWO c o m p u t e r c o d e 504 A7 The Semantic lexicon: the semantics of commercial verbs. 512  vi  1 1  1 1  1 1  LIST  Figure  1. 1:  A  network  Figure  1. 2:  Descriptive  Figure  1. 3:  A  Figure  1, 4 :  task  The  Figure  network  plane  Evans*  the  representing  ANALOGY  defining  Q u i l l i a n ' s  Example of by G u z m a n ' s  traced  Figure  1. 7 :  Figure  1. 8 :  Figure  1. 9:  Figure  1. 10 :  The  Figure  1. 11 :  Declarative  Figure  4. 1  :  Overall  Figure  4. 2  :  A  Figure  4. 3  :  Two  Figure  4. 4  :  Sentence  Figure  4. 5  :  4. 6 Figure s i g n a 1 s.  :  SIGFACTS  4. 7 Figure heuris tic •  :  Decision-table  4. 8  :  The  SWG  after  the  4 .9 :  The  SWG  after  updating  word  6  through  a scene analysed program.  1. 6 :  Figure only) .  in  memory.  Figure  Figure  ARCH.  program.  'food'  pathways  semantic  Algebra  | 1  Memory.  Intersecting  1. 5 :  FIG ORIS.  description.  for  Semantic  OF  problems  for  STUDENT  CARPS.  S  Training seguence for Winston's concept learning program. The  binary  tree  effects  of  audit  of  a  reports  control  knowledge  alternative  -  in  the  fcr  by  prices.  IMIS SWG.  ATU  the  from  the  f i r s t  share  readings.  after  resulting  l i s t .  deductions.  i n i t i a l  output  immediately  U S E upon  an  syntactic  structures  SIGFACTS  cf  non-declarative  8  flow  p r i o r i  structure  grammar.  f i r s t  the  2nd  signal. and  3rd  concept-recognizing  signal. by  the  third  signal  (part  v i i  Figure  4 . 1 0 :  The  combined  Figure  4 . 1 1 :  The  accounts,  Figure  A3.1:  comprise Figure  a l l  A3.2:  c cnscicusne Figure  A3.3:  Tabular  SWG  after  after  the  the  representation  third  t h i r d cf  signal.  signal.  the  three  worlds  that  existence. The  three  postulated  components  in  the  world  ss. Information  flew  between  the  three  worlds.  of  v i i i  ACKNOWLEDGEMENT Nc  formal,  sufficient have  typist,  my  and  My  completion We  must  and  are  listened lives  Mcudgill,  than  past  their  to  and  Eduardo  S S  to  big the  my  many  tc  wife  Jeffrey  their  le who  Vancouver  their  many  is  as  members  have  have  permanently  Gloria  of  favcrite  Julian. the  at  great cf  s h a l l  Drummond,  and  as  my  me,  have  enriched  our  never  Eravin  Schwartz  the  faculty  stimulated  We  t*y  strains  r e l i e f  who  Judi  and  and  share  book'  friendships.  Peter  the  journey  and  "daddy's  students  others,  f i r s t  Brendon,  grateful  to  cculd  rewarding.  more  patiently  with  gratitude  go  tc  months  of  fellow  among  deeply  thanks  many  my  acknowledgement  i n t e l l e c t u a l  have'borne  these  own.  sc  Velma,  family  and  express  made  possible  of  to  printed  John  forget, 8  Indra  S  Barbie  Jiagen. The University University study  leave  shouldered  generous cf of  B r i t i s h Canterbury  and my  acknowledged. Centre special  and  the  the  remark.  Columbia, by  support  teaching The  f i n a n c i a l  way cf  of my  support the an  helpfulness  cf of  of  the  extension  of  my  there  who  colleagues  the its  the  support  r e s p o n s i b i l i t i e s  excellence  cf  is UEC staff  gratefully Computer deserve  Above f i r s t (as  Dr.  a l l  G.  I  A.  Chairman)  wculd  like  Feltham  and  and  B.  Commerce  and  Dr.  Computer  Science  -  hundreds their  of  debt  to  encouragement ways  and  his  own  my  own,  c l a r i t y  to  Dr.  has  thank  later  Mattessich Eeiter  for  their  of  I  Dr.  Dexter  his  his  practical  Hattessich and  tried  his to  s c i e n t i f i c  teach  cf  endurance  of  a  to  c a r e f u l  a  in  the  meaning  and  for  particularly  thoughtful  through  the  research  Department  memoranda  ewe  patient,  Dexter of  assistance  who,  S.  A.  Faculty  patient  counsel.  committee  the  the  preliminary  for  my  Drs.  of  of  and  and  writings  in  R.  pages  c r i t i c i s m  great  V.  tc  and  warm  so  many  example  cf  c r i t i c i s m  of  of  rigcur  d i f f i c u l t  Bevan  and  pupil.  J .  Clarke  Christchurch, Hew  Zealand  June  1975.  IMISj.  |  Chapter  I n t e l l i g e n t  0:  Introduction  Computer-based of  rapidly  response  growing  of  organizations  and  has  as  progressed  f r a c t i o n design, of  on  of  parts systems  their  net  neither of have  the  very  art,  proven  by  rather  methods  knowledge  nor  changing  rapidly  at  dismayed  f a i l u r e  to  use  than  f i e l d  a  MIS  such  a  is few  cf  by  its  lack  the  tools  we  research  young  i t s  of  and  to  and  of  any  unifying  self-analysis  cf in  auj the  with  precision  the  standards judged.  We  use  the  advance  cur  Perhaps  practitioners  such  pragmatic  cannct  its  tc  systems  are  be  to  a  inextricable  we  nor is  as  However,  define  system  the  possible  expression  creatures.  s t i l l  that  to  as  crudely  i t s e l f  science  own  serve.  result  able  an  our  new  outward  design  increased  become  but  a  s c i e n t i f i c  c r i t i c i s e  the  the As  of  they  and  information  which  any  being  which  of  because  a l l  of  concept  performance an  net  is  component  system  able  has  management  tc  t h e o r y  position  It  its  cf  more  processing  processes  . are  testable  become  art  ('MIS')  technological  managing  The  competence,  subject  e v a l u a t i o n T h e y  unfortunate  have  Systems  important  of  resources.  management  are  one  problems  maintain  I  information  expenditure.  and  Theory.  Objectives.  are  information  p r a c t i c a l  the  rigorous^  the  computers  implement  of  and  Accounting  Management  to  national  impressive  and  complexity  society  expenditure  Mis  i t  is  technology  is  appear  tc  be  theory  and  its  which  science  affords.  Introduction.  IMIS;.  Intelligent  Moreover, Systems  are  at  error-prone systems, o v e r a l l in of  a  a  building  those must,  who in  writing make more  for  attempt  their  of  adaptive than  svstem  manner. in i t .  the  are  previous  no nor  of  the  and  along  a b i l i t y  systems  these  proposals  to cf  tc in  such the  suggestions lines for  although  large of  the  manage the  frequently  solution  and  intc  guarantee  complexity  tec  the  Information  modules  its  2  laborious  inherent  the  computer 1  the  small  manner  Clearly  software, MIS  by  Indeed,  this  Theory..  Management  of  carries  the  involve own  Accounting  created  assembly which  systems  part,  bold  present  process  integrated  and  computer-based  manual  i n t e g r i t y  MIS  task  defeats problems task we  shall  they  'automated  of  are  system  design'.  1  "At present the largest m a c h i n e s come w i t h more o r l e s s a million words of software. In a n o t h e r d e c a d e c c u l d we h a v e 100 m i l l i o n ? . . . If this is tc be achieved, I think i t w i l l h a v e t o b e d o n e by t h e m a c h i n e s themselves generating most o f their own software because humans simply cannot do things c o r r e c t l y at that r a t e . " B. W. H a m m i n g , B e l l Telephone Laboratories, 1969.  Introduction.  IMISj^  Intelligent  is££2iIJQtanc^ profession. d i f f i c u l t y subject But too  In in  from  strained offers lacks  to many  an  i t s  which  they  accountancy the  core  c r i t i c a l such  of  growing  more  respond models  varied  have  theories  may  rigorous  their  We  similarly  proposed. and  tc  framed be  of  matter  of  and  of  their  them.  therein  society  with  formal  Eeeper  It  tec  And  like  MIS  framework, and  in  attempt  within terms  analytic  of  tc  cf  identify  propositions  for  systems  methodological  open  to  theories, and  i t  theories  their  legic,  a  purposes.  i t s  are  it  been  However,  for  its  also  place  These  had  has  systems.  premises  deductive  rigorous  and  hypotheticc-deductive  tested  these  among  competence.  organize  as  factual  commend  methods  have  evaluated.  been  MIS  theoretical  responses  processing  sometimes  Accountancy  basic  unifying  accountancy  and  i t s  3  distinguishing  complexity  diversity  of  has  in  technical  a  be  -  subject  disturbing.  to  When  refutation.  created  the  study.  consistency  the  to  in  i t  disciplines,  proper  are  may  boundaries  i t s  growth  Theory.  information  identity  emerging  embracing,  which  of  i t s  ficcountinc[  older  has  responded  corresponding  and  far  crises  Accountancy  s i m i l a r i t i e s  a  defining  matter  has  is  HIS  which wish  'theories  cf  possible employ tc  see  MIS.  Introduction.  IMIS.:  I n t e l l i g e n t  A r t i f i c i a l intersection study we  of  draw  are  as  yet  modest  Management is  f i e l d  Accounting seen  as  models  than to  in  purposeful  the  and  for  between  A r t i f i c i a l short.  n  are use  of  a l l  area  that  while  its  its  hopes  its  Information  Systems.  Management  structure  theories cf  and  cf  have  is  the  much  faced  mere  results  in  that  when  Systems  are  fundamental and  of  representation  cr  MIS  and  representation to  by  that  Intelligence,  with  and  Eowever,  theories the  which  attainments  argue  there  such for  A r t i f i c i a l  knowledge  shall  the  insights  Information  Intelligence  concerned  We  and  frcm  problems  'application'  another.  s c i e n t i f i c  the  third  to  in  logic  the  applicable  incidental  problems  theories  I  Accountancy  and  to  i n t e r d i s c i p l i n a r y  mathematical is  4  T h e o r y  an  x  believe  relation  already  subjects  liffiii3£iii  We  Accounting  suggested  one  in  Accounting  science,  problem-solving  proposals,  are  and  lJQt€llia,ejce  computer  human  our  methods  of  MIS  lend  each  and  ether.  Intrcd uction.  I M I S I n t e l l i g e n t  0^1  SPECIFIC  OBJECTIVES  (1) is  to  We  be  i n t e l l i g e n t Information review which  CF  shall  seen  THIS  researches  model as  In  significance  of  some  We  of  the  Management  work  and  (3)  We  forraal  as  and  to  1969,  formal, (1972c)  a  System,  and  we  shall  automatic  in of  (Al)  devising knowledge  in  behaviour. svstem  cf  an  in  the  displaying Intelligent domain  cf  management. shall  novel  axiomatized (1964,  implement  characteristics  employ  the  methodological  axiomatized  x  (1967)  a  present  in  and  i n t e l l i g e n t  Information  accounting  system  shall  MIS  displaying  particular  representations  (2)  an  Intelligence  machine-accessible  guide  5  Management  shall  A r t i f i c i a l  proposal.  to  which  entity  we  robotics  suitable  Theorv..  Intelligent  theorem-proving,  forms  in  ar  •IMIS'),  in  this  the  a  an  (or  Accounting  BESEJJRCH..  primarily  behaviour,  support  and  propose  System^  consider  WIS  theory  provide  a  theory  of  1972a,  axiomatized  of  resulting device  accountancy  theory  1973a, of  tc  test  accounting  foundation  1972b,  computer  MIS  for by  by  I j i r i  testing  an  Mattessich  1973b) by  a  and  a  Mattessich  .  Introduction.  IMIS,: I n t e l l i g e n t  MIS  ana  Accounting  Thecrj,.  6  Q.j.2 COMMENTS UPON THESE OBJECTIVES. Our  r e s e a r c h has  closely  related  proposal  f o r the  (A)  the  possibility System  may  manner  that  exploratory Any  the  a  instrumental system  hypothesis fulfil may easy  call  to  system,  evaluation  Management as  theory. cf  the  Information  intelligent - is  in  the  necessarily  such  as  a  we  'traditional' test  i t s purpose  management by  and  a judgment  then  upon  the  system  is  i t s purpose.  If  the  have, of c o u r s e ,  the  overriding  out  IMIS, as  purposes  unequivocal  regarded  i t fulfils  find  t h a t an  the  the  i s t o be t e s t e d p r i n c i p a l l y  as a t h e o r y  attempt  s t a t u s : the  conjectural.  d e g r e e t o which i t f u l f i l s  to  -  human manager i s i n t e l l i g e n t  t o whether o r n o t  expressed  although  o f an a c c o u n t i n g  computer-based t o be  which  methodological  test  component  designed  and  major components  differing  IMIS and first  a  information as  have  that be  two  whether  i t i s true.  we  propose  o f an MIS  shall  MIS  of s u c h  an  Eut  we  hypothesis  I t i s cur  i t , promises  more a d e g u a t e l y  designs.  2  duty  tc  than  what  we  recognise  that  nc  is  possible.  SSftaintv.  that a factual theory i s true i s , s t r i c t l y speaking, unobtainable. See C h a p t e r s 1 and 2, When we r e f e r i n t h i s study to the t r u t h of a f a c t u a l p r o p o s i t i o n without qualification we mean t c i m p l y t h a t the p r o p o s i t i o n h a s t h u s f a r s u r v i v e d a l l c u r a t t e m p t s t o r e f u t e i t and is, i n P o p p e r ' s t e r m s (19nn), h i g h l y c o r r o b o r a t e d . When we r e f e r t o ' t e s t i n g ' a t h e o r y we r e f e r t o such a t t e m p t s . 2  Intrcduction.  IMIS:.  We  are  considerable  Intelligent  HIS  and  encouraged  tc  posit  advances  behaviour  in  in  which  Al  A r t i f i c i a l is  theorem-proving mobile,  a  environments s h a l l  as  the  into  attainments  and  i n t e l l i g e n c e computer  so  that  i t  i n t e l l i g e n t  entity  the  appeared  as  conventional be  seen  cf  the  Bover  d i s t i l  means  to  the  MIS  is  research that  intc  tc  the  The and  essential  the  Project  goals  l i e  model  is  most  relevant  in  of  cn  the  represent  knowledge  of the  interpreted.  viewpoints  as  as  Without  such  a  a  single  unifying an  We  these  nature  value  handbooks of  instructive  a  of  to  a  potentially  theory  l i t e r a t u r e and  unifying  and  and  framework  o v e r a l l  data-processing  technical The  i n t e l l i g e n t  The  3  the  i n t e l l i g e n t l y .  offered  and  of  by  applied  proposal  the  cf  MIS  wisdom.  to  as  7  system  model  robot.  general  behave  is  mere  in  the  may  research.  mainstream  Mars  on  model  and  SBI  attempt  This  design  present  a  research.  p a r t i c u l a r l y  the  Theory^  simulation  principal  discuss  and  the  goal-seeking  which  therefore  such  Intelligence  systems  between  project  in  our  self-managing,  differences such  made  Accounting  fcr  MIS  texts 4  in  have  collections framework  relevant  cf can  example  3  For example, see Sutrc 6 Kilmer (1969); Winston Nagata et a l . (1973); see also sessions 11,15 and the IJCAI Proceedings 1973; see Chapter Jackson (1974).  *  For example: Glans e t . (1963), Murdick 8 Boss Strater (1974), Davis  cf  (1972a); 24 of 9 of  a l (1968), Kelly (1970), Van ficrne (1971), Blumenthal (1969), Burch 8 (1974).  Introduction.  IMISp  Intelligent  Information  Economics.  s t a t i s t i c a l  decision  Schaifer,  1968)  of  value  the  systems  has  Radner  a  been  (1972),  Feltham  8  Given  functions,  the  timing,  the  they  lacked  solutions i S ! 3 S £  of and  employed  by  we  s h a l l work  are  (1972), many  o  cite  of  Krasncw  only  the  referenced and  to  in  for  example,  Eaiffa  and  fcr  Feltham  8  (1972)  the  analysis  Marshak  Eemski  and  so  the  guide  signals  because  come of  the_y.  the  information than  these  cn.  choice  the  pay-off  have  a  the  metamcdel  are  grounded  Information system  is  s  under  decisions,  tc  8  (1970),  state-dependent  Specializations  certain  point  involving  tc  example,  r  starting  information-delivering  Demski  MIS  8  a  (1963,1971),  having  optimal  8  and  issues  of  Theorv^  framework  (1968,1972),  credible  more Boyd  (see,  Marshak  aggregation  the  references,  by  choice  F  from  general  actions  i i § 2 U i  elegant,  The  Feltham  before.  which  selection  s  evolved  of  Accounting  information  metamodel  uncertainty  and  Herein,  and  Butterworth  this  and  theory  new of  MIS  the unity yield  in  the  Economics simpler,  simulation  mere  procedure  (1963).  most  usually  relevant points  lengthy  to  texts many  mere  bibliographies  throughout. pertinent in  several  cases.  Introduction.  IMIS:.  The to  present  supply  paradigm:  MIS *  entity.  I n t e l l i g e n t  of  the  particular the a  device  robotic  the  of  states IE  of  of  with  a  in a  9  T h e o r y  (1)  atcve,  similarly  knowing  component  is Tc of  and  seeks  unifying perceiving  general  an  these  a  proposal  theories theories,  'automatic  research  of  tc  is test  accountancy  in  the  manner  bookkeeper'.  in for  Kuhn the  complementary tc the Information c o m p e t i t i v e with i t . Ecth models system's r e a c t i o n tc the information  perceived  and  with  the  system's  modelling  world. limits  distributions  expresses  of  the  used here i n the sense given i t describe the new models knowledge w h i c h we d i s c u s s l a t e r .  signals the  as  of  rigorous:  implementing  basically  probability u t i l i t y  as  more  This viewpoint is Economics mcdel, not are concerned with a  content  and  indicated  design  MIS  mechanism,  "Paradigm" (1970). representation  of  the  interpretations  6  7  and  second  methodologically  of  as  Recounting  7  (B)  by  and  proposal,  research  that  MIS  the  the over  reaction  modelling a  via  to  pre-defined generalised  revision set  cf  cf  states,  outcome  and  functions.  The Al we are concerned with eschews generalized p a r a m e t e r s and f u n c t i o n s and a t t e m p t s to mcdel the state of t h e world and the b e h a v i o u r a l r e a c t i o n s i n d i s c r e t e d e t a i l .  Introduction.  IMISj.  Intelligent  Although  i t  in  systems.  economy  evaluating  that  in  recent  they  and  are  contradictions From  cf  of  i n s u f f i c i e n t  Intelligence  believe  that  we  putative deductive  theories  to  aid  those  seek  accountants tc  conclusions or  tc  elegance,  practices  deduced  in be from  a l t e r n a t i v e l y ,  because  believe are  rigour  aspects in  theory they  with  many  they  are  premises.  research  A r t i f i c i a l  we  showing  validly  to  i t s  general  seek  invalid  attention  for  descriptions  l o g i c a l l y  proposed  a  its  hypotheses,  Although  of  by  Accountancy  two  furnished  auxiliary  these  formal  foundations  general  They  are  any  mathematico-lcgical such  10  Theory^  lacked  been  giving  experience  discussions  test  of  £ £ a c t i c e s they  axioms  The  of  generality.  because  has  years  form  authors  their  Accounting  underlying  accountancy,  and  accepted  i t s  the  The  systematise  valid  cf has  foundations  and  Accountancy  consideration centuries  MIS  the —  are  often  and has  -  the  IMIS may  that  theoretical  tco  precise an  casual,  d e f i n i t i o n .  essential  proposal and  appear  intimately  the to  with  f c r  u n i ty.. applied  methodological  span  related  a  broad for  area three  reasons:  Introduction.  IMISj.  (a) function  Intelligent  Accounting and  its  incontestably  a  System.  seek  If  Management  we  Under  accounting  actions as  The  fact  logically  has  a  animating  conclusions systems. accountancy  where  prove  to  systems.  interpretation  and  the  for  an  tc  accountancy are from  given  minimizes of  error  be  set  cf  is  in  derive  such  computer axioms  the the  being  tc  Intelligence  the  into  a  entry  tc  extent  is  begin?  entries)  a  1  Intelligent  of  attempts  d i r e c t l y  risk  1  r  recording  theorem-proving  This  c  Information  A r t i f i c i a l  some  encoded  its  conclusions that  with  e  Management  better  conclusion.  that  be  a  theories  mechanical,  find may  cf  inferred to  h  decision-making)  accounting  experience  theorem-proving  the  is,  is  valid  with we  to  axiomatized (that  T  principles,  knowledge  System  challenge  considerable  its  component  deductively  axioms.  Accounting  assistance  core  (b)  and  (with  Information  regarded  MIS  of  notations  cf  problems  of  introduced  intc  IMIS. (c)  We  believe  i n t e l l i g e n c e ,  that  accounting  management  information  essential  problem;  EiilEOse-oriented  the  (b)  with  systems  machines  which  instrumentally  efficient,.  s h a l l  the  c a l l  science'.  this We  that  That i t s  are  problem  believe  are  areas:  with  is,  with  For of that  and  (a)  of  a  in  which  is  truly  as  use  of  notations, sound  better  'epistemology  same  normative  epistemically want  and  the  representation  representation both  a r t i f i c i a l  processing  concerned  effective  knowledge^  and  three  information  systems  knowledge or  the  an  and  phrase  we  applied  characteristic  Introduction.  IMISJ.  of  an  MIS,  commercial the  (and  machine'  the 'IMIS'  an  MIS  i s  to  i t  l i e s  f i r s t  concern  not i t s  an attempt  a  of  hardware,  i s to  theory  pose  net be  T h e MIS  i s  net usually  knowledge  in sc  f i l e s  i n  an  regarded. nature i f  of  there  information  not  f o r  essence  C e r t a i n l y ,  and  to a  concern  see the epistemic  management  software,  12  r e s t r i c t e d  but i t s  second.  of  Theory..  systems  the  current  and o p e r a t i n g  systems.  CAVEATS.. Although w i l l t o  an  a  future  problems  or  the  f o r which  any c f  which  forecasts  see  proposals  i t  mere  of  cn  Fcr  'Daedalus'  i s  easy  technology  enumeration  unsubtle will  the  (1969),  less  not  fastest  prcblem-solvina  see Eernstein  (1971,1972).  our  conventional,  mega-centuries methods  computing  engineering  Moreover,  exhaustive  Better  i n  purely  8  problems  as  forecasts  K r i e b e l  relevant  Weiner  such  f o r  computer.  technology  of  advances  technologies.  combinatorial  solution  (1970)  many  that  we d o n o t c o n d i t i c n  solutions  conceivable  assume  remove  solve;  algorithmic  we  IMIS  simple,  never  8  i t  i n  unknown,  but  although  theory  obstacles  For  knowledge.  general  with  reach  of  need  technology  be a  technology  can  label  technology  (1)  upon  i s  and Accounting  and i t s  0.. 3 F O U R  8  i s  representation  Thus  MIS  that  system)  'epistemic  to  Intelligent  A.  are  Clarke  t e c h n c l c g i c a l  (Summer  1967),  Kahn  (1967).  Introduction.  I M I S I n t e l l i g e n t  required,  not  technology the  most  faster  i s  the  rapidly  MIS  use  least  f a l l i n g  and Accounting  of  of  e l d  T  h  e  c  methods.  the obstacles  tc  r  1  3  Hardware  an  IMIS  and  i s  obstacle.  "In 25 y e a r s i t i s predicted that a single c h i p c o m p u t e r w i l l b e a v a i l a b l e c a p a b l e o f 20 m i l l i o n instructions a second with about 65K c f i n t e r n a l memory s e l l i n g f o r about $1.00! Even i f this e s t i m a t e were o f f by an o r d e r c f m a g n i t u d e the social significance i s enormous." Firschein, e t a l . , 1973.  Core  sizes,  relevant,  cost  but  development  of  (2)  represented  review  a  to to  on  (3) ^ i i § i S l f i i i §  be a  interest.  to  A l  processing  of  than  are the  knowledge.  primarily  and Methodology. to Philosophy.  It  as i s  At best  science  should  locate  seme  be s e e n problems  a not the  as  a i n  A l a n d MIS.  their  t c  a r t , although  without  of  our research  examining  previous  offerred  theory  attempt  by both  A l  extend  i s  cf  important  the organization  philosophy  We d o n o t s e e k  Intelligence  less  contribution  an  In  f a r  thesis  faced  n  c f  Accounting  as  i  are  theory  the  and  methodology  IMIS.  they  This  contribution  sections  per b i t , and parallelism  draws  enly  possible  advance  researches  into  that a  the  application  the state  we b e l i e v e  upon  of  existing tc  the  the A r t i f i c i a l  cur  attempts  new d o m a i n  t c  may n e t b e  Introduction.  IHISi.  (4) that  in  Intelligent  To  avoid  our  view  MIS  a  than  to  t o t a l  set  better  described  we  be  seek  states  to  of  a  position  means  for  the  exercise  recognise However  must MIS  that  In  Management  Decision  we  are  System"  a r t i f a c t ' s  a  portion  human's  portion  defines  the  term  and  -  a  is  be  Action  movement of  the  mcvement  'Management  System  people^  set  We  cf  are  by  and  the  possible  perhaps  Information  very  hcwever, proposing that  is  people.  we  inherently  a  "Intelligent Management this  boundary  the  MIS  understood.  is  management  an  which  The  manage;  In  is  eguivalent  "Intelligent  whole in  a  what  appropriate. in  not  exercised  for  or  more  tc  established is  System"  would  envisaging  MIS  System*  objectives,  a b i l i t y  formerly  cutset  System. '  is  our  14  the  and  g u a l i f i e d .  its  at  is  'Management  of  proposal  Information  Support  view  increase  Management  It  procedures  c a r e f u l l y  term  the  a  state  i n t e l l i g e n c e  a b i l i t i e s  the  since  as  Theory^  Information  system.  machines,  be  to  some  s o c i a l  with  machine.  t h i s  to  a of  imbue  we  'Management  rather  might  Accounting  misunderstanding  a r t i f a c t .the  and  research  between system  boundary  the  and  the  which  System'.  Introduction.  IWISp  (K4  Intelligent  certain  order  of  some  a the  in  is  terms  of  and  arguments  -  concerning  is  read  in  Theory..  be  15  included  where  d i f f i c u l t  understood  d i f f e r e n t l y two  for  the  because  we  have  concept  i t  is a  (or  cf  are  of  is  the  with  key  the  and  the  Systems  and  explained fcr  may  and an  that order  most The  concepts  cf  interwoven  find  explained  but  seme It  two  to  different  The  therefore  accounting  introduced.  few  tc  science.  not  f a i l e d  topic  Computer  d i s c i p l i n e s .  axiomatized -  in  appendices  other  ftgain  every  p a r t i c u l a r l y  the  the  l i e  Accountancy,  used  in  the  regard  Information  with  jargon  cf  terminology  philosophy  familiar  science  which  with  backgrounds  the  from  nature  the  Management  concerning  point  in  6  f i r s t .  computer  reader,  The  accounting  symbols  that  text  the  rigour)  the  in  to  Glossary  be  at  degree  terms  short  the  p a r t i c u l a r l y  of  Accounting  presumed  usefully  for  Accounting  i n t e r d i s c i p l i n a r y  arise,  presumed  small  few  the  presentation.  Science,  reader  of  problems  backgrounds  the  and  OUTLINE.. Because  to  BIS  f u l l y  task which  is are  attention  f i e l d s .  Introduction.  tc  IMISj,  Two the  introductory  prerequisite  •  Chapter  Intelligence i t s  Intelligent  1  the  chapters  reviews  with  on  a and  a  of  i n t e l l i g e n t  reasoning.  and  epistemology  use  •knowledge'  •  Chapter  axiomizations theories  w i l l  be  t h i s  i n  and  that  the  h  e  c  r  j  1  l i t e r a t u r e  b r i e f of  term  the  6  and  tc  A r t i f i c i a l  doubts  deductive  cf  logic  study  of  in  Al  as  a  mcde  Intelligence  with  1  sense  cast  A r t i f i c i a l  'knowledge  taken  cf  presentation  Because  the is  topics  unaided  introduces  their  attempts theory.  readiness  to  construction.  MIS  the  the  accounting  theory  methodology  believe formal  2  unfamiliar  Accounting  T  the  different  meanings  of  ' i n t e l l i g e n c e ' .  of  l i e s  s c i e n t i f i c  care  and  review  discussion  adequacy  some  Accounting  the  necessarily  of  emphases  ana  concepts.  achievements  research  MIS  we and  IMIS  to  those  attempt, then may  to be  tc The  employ Because  educated in  Chapter  describe  the  employed  as  create virtue  the  2,  such cf  methodology  the  f i e l d s  tc  manner a  of  methcdolcgy  this in  formal  cf  summarise in  test  which of  we  such  theories.  Introduction.  I n t e l l i g e n t  IMIS.:  •  On  these  proposal  for  the  with  IMIS  •  the  IMIS  detailed  describes  s u f f i c i e n t  to  •  f i n a l  study  and  and  in  the  the  chapters  Chapter  establishes  the  d e t a i l  of  appendices  system  to  present  p o s s i b i l i t i e s  be  3  inherent  the  Although  simulations  supporting permit  Theory,  conventionally  implemented.  tests  Accounting  contrasts  4  the  the  and  present  presented  The  IMIS  was  and  foundations  the  Chapter  modest  MIS  methods  the to  a  cur  i n t e g r i t y  designed  only  17  MIS.  by  which  portion  a  cf  the  could  be  chapter  are  findings  of  system that  cf  r e p l i c a t e d .  and  assess  the  for  future  research.  Introduction  Chapter  1.1 1.2 1.3  1.4  1.5 1.6  1: Intelligence, knowledge a r t i f i c i a l intelligence.  and  intelligence. A r t i f i c i a l intelligence. Knowledge i n epistemology and knowledge i n Al. 1.3.1 what i s knowledge? 1.3.2 T h e e p i s t e m o l o g i s t •s 'knowledge'. A r t i f i c i a l Intelligence insights concerning 'knowledge'. 1.4.1 R e l a t i o n a l d e s c r i p t i o n s : n e t w o r k s and graphs. 1.4.2 Procedural descriptions. 1.4.3 The i n a d e q u a c y c f p u r e l y d e d u c t i v e i n f e r e n c e . 1.4.4 'Actions* as 'truths*. 1.4.5 Inductive l o g i c a s p r a c t i c e d by a m a c h i n e . 1.4.6 Structures for organizing knowledge. Examples of Al r e s e a r c h . The languages of A r t i f i c i a l Intelligence. 1.6.1 A b r i e f i n t r o d u c t i o n to U S E 1.6.2 Very h i g h - l e v e l languages for theorem-proving. 1.6.2.1 Theorem-proving languages. 1 . 6 . 2 . 2 The MICBC-PIANNEB notation. 1.6.2.3 Backtracking. 1.6.2.4 A s s e r t i n g and e r a s i n g truths. 1 . 6 . 2 . 5 C o n s t r u c t i v e use of f a i l u r e . 1 . 6 . 2 . 6 a d d i n g a x i o m s and procedures. 1.6.2.7 N o n - d e c l a r a t i v e logics. 1.6.3 The s i g n i f i c a n c e of v e r y h i g h l e v e l languages  Intelligence,  knowledge  and  A l .  IMIS^  We  have  information should  I n t e l l i g e n t  be  common the  considered  and  philosophic  there  of  c a r e f u l l y the  their  desirable  of  problems  such  and  by  'applied  we  attempts  coin in  v a l i d i t y  1  Such  refer  an  the  A r t i f i c i a l which  of  and  1  See  2  Echoing Hewitt's ' p r o c e d u r a l embedding  Let  us  of  of  of  applied a  mcst  knowledge  and  peculiar  decision-making  sciences,  epistemology  (1975).  Intelligence  expressing  the  arguments  seems  argument  understand  study  to  epistemclcgy  accommodate  reservations.  are  we  are  regard  areas.  'an  phrase  to  for  in  there  of  science  interested  Mattessich  with  between  established  epistemology'  i t  knowledge  former  reasoning  the  we  problems  phrase  for  as  three  the  'a  administrative  premises.  elaborated  and  the  in  grounds  with  forms  methods  and  the  their  systems  (i)  areas  common  therefore  of  of  be  accounting  common  aspects:  three  and  19  information  their  two  the  emerging  the  extension  problems  i t s  By  in  to  sciences'  methodology  deontic  us  distinguish  epistemology'.  the  to  by  have  knowledge  knowledge  applied  related  grounds  T h e o r y  i n t e l l i g e n c e ,  management  problems  appear  instrumental  be  Accounting  that  and  to  These  v a l i d i t y  and  suggested  processing  epistemology.  MIS  tc  is  reader On  almost 2  with  We  the  not tc  yet the  other  or  f u l l y  study hand  of the  s e l f - c c n t r a d i c t c r y use  devise  epistem ically_  imperative  it  tc  refer  tc  representations  interesting  (for  Glossary. 'procedural cf knowledge'  epistemclcgy' and (Hewitt, 1972,1972).  Intelligence,  knowledge  and  Al.  I WIS.:  example,  Intelligent  notations  injst rumen t a l l y or that  they  stating  In to  in  this  to  the  and  research  which  •Intelligence' proposal  and  for  •intelligence' understanding  IMIS.  ordinary-language MIS.  In  devising  to  represent  the  representation  regarding  the  Solutions  to  cf data  knowledge of  this  and  fcr  such  interpretations,  so-called  3  'very  "After  high  to  work  Ezra  Pound.  in  a  so  far  level  to  examples  have  new  was  laboratory."  Al  knowledge. in  our  meaning  of  technical  and  of lead  guy  Intelligence,  structures  concern  has  been  is,  statements  other  statements.  tc  unaided  We  who  Ycrick  the  Accountancy  program  computer  a  cf  illuminate  languages'.  philosopher a  Computer  concepts  that  deductions,  and  cf  the  and  use  in  introduce  Computing,  problem  e f f i c i e n t  analytic  Leibniz  lazy  major  problem  respect as  a  true  attempts  We  the  to  structures  0  notation  philosopher's  in  2  e f f i c i e n t  these  3  consider  word  y  tc  the  with  key  'meta-knowledge',  v a l i d i t y  purely  are  r  calculus),  representing  the  the  o  solution.  of  'knowledge' use  and  for  We  use  of  from  i t s  some  methods  and  gap for  'knowledge'  the  e f f i c i e n t l y  sub-discipline  overview  display  lead  concerned  a  e  predicate  epistemology'. as  h  the  the  are  T  computationally  program  'applied  b r i e f l y  and  we  Intelligence  Science  they  step  chapter  an  A r t i f i c i a l  that  one  problem  devise  in  Accounting  upon  conclusions  bridge  the  and  based  useful  reasonable  MIS  a  diminishing  by  heuristics  languages discuss  was  too  Wilks,  knowledge  and  damned quoting  and  Al.  IMIS^  i l l u s t r a t e  Intelligent  both  these  MIS  and  important  Recounting  21  T h e o r y  issues.  1A1 IINTELLIGENCE^i Intelligence International  is  defined  Dictionary  by  Webster's  Third  New  as:  "The available a b i l i t y to use o n e ' s existing k n o w l e d g e t o m e e t new s i t u a t i o n s and to solve new problems; a b i l i t y to perceive one's environment, tc deal with i t symbolically . . . Tc work towards a goal; . . . " But many  ' i n t e l l i g e n c e *  senses  view  is  (1967),  to  convey  supported for  example,  by  is  seen  by  of  the  many  components  MIS  an  one  of  the  usage  presents  i t  In  some  any  is  an  psychologists of  l i t e r a t u r e  overworked them of  with  the  a c t i v i t i e s  overview and  a  i n t e l l i g e n c e , by  moronic  taxonomy ' l e v e l s '  many  such  systems. organizing  of  generality  which  This Guilfcrd  intelligence  f a c t o r - a n a l y t i c  as  model  such in  as: t y e, n ve  of  with (such  Zannetcs  i n t e l l i g e n c e  tco  c l a r i t y .  cf  definitions  concomitant cf  in  i n t e l l i g e n c e .  excessive  merely  used  psychologists.  "Intelligence is d e f i n e d as the a b i l i measure, c l a s s i f y , store, retriev knowledge purposefully i . e . to gai knowledge i n a s u b j e c t i v e and o b j e c t i 1970)  r e f l e c t  word,  as  the  or  Management  word  and  include  prereguisite  storage)  (1968)  a c t i v i t i e s  tc i d e n t i f y , and process gcal-related sense" (Will,  are  presented as  the  lower  Information  Intelligence,  knowledge  to  performed a  useful of  seven  Systems  and  A  I.  IMISj,  but  his  lead  away  I n t e l l i g e n t  higher from  The Moreover, d i f f e r  levels  for  of  each  • i n t e l l i g e n t '  length.  and  only  Accounting  to  attitude  which  l e v e l  is  is  revision  the We  definitions adopted  dependent  intended  of  begins.  is  22  and  research.  that  discuss  Thecry^  probability  ' i n t e l l i g e n c e '  behaviour  to  What  Al  attribute  regarding  profitable  refer  present  meaning  MIS  cn  context.  opinions  attribute  do of  net  will whereat  think  intelligence  within  the  i t  at  any  l i t e r a t u r e  of  Al? The problem  f i r s t  altogether  paraphrase  Henry  does."  Since  of  has  what  defined  come  machine  Despite  i t s  problem  is  adopting  time to  of  be  the  a i f is  machine performed equally  offer  a  attitude  Turing's a  That  by  a  is  human  tetter  is  a  avcid  i f  taken  then  i t  in  this  (tc  (1950)  intelligence  say,  be  the  intelligence  suggestion  to  poor  tc  viewpoint  what  t e s t '  would  i n t e l l i g e n t this  is  o r i g i n a l  'Turing  which  is  pragmatic  "Intelligence  called  popularity to  common  ostensively.  by  i n t e l l i g e n c e  by  most  Hatfield)  the  only  performed  the  and  is  a  task  tc  is  imply  inferred  problem  is  that  domain.  d e f i n i t i o n .  The  one.  Intelligence,  knowledge  and  A l .  IMIS^  A p.447)  I n t e l l i g e n t  more  when  HIS  illuminating  he  and  Accounting  attitude  is  Theory.  taken  by  23  Minsky  (1961,  says  "To me intelligence seems tc denote l i t t l e more t h a n t h e c o m p l e x c f p e r f o r m a n c e s w h i c h we happen to respect, b u t do n e t u n d e r s t a n d . " "My cwn v i e w is that this is more c f an a e s t h e t i c q u e s t i o n , c r ene cf a sense of dignity, than a t e c h n i c a l matter." "Ycu r e g a r d an a c t i o n as i n t e l l i g e n t u n t i l ycu understand i t . . . It may b e s o w i t h man a s with machine that, w h e n we u n d e r s t a n d f i n a l l y the s t r u c t u r e and program, the f e e l i n g of mystery (and self-apprcbaticn) will weaken."  This is  is  any  more  one  thing.  i n t e l l i g e n c e of  constructive  since  i t  offers  the  hope  emulating,  one  by  It  by  performances .  It  1  mysterious, d e f i n i t i o n course,  a  unitary may  always  also  cf  elude  In  us.  possible  f a l l a c y  emulate  intelligence  that  to  valid  theory  of  that  that  one,  denies  'fountain  of  denies  we  each  that pure  this  not  may  emulate  the  'ccmplex  there  is  any  intelligence*  whose  view  i d e n t i f i c a t i o n . is  cf  intelligence  there We  is,  cf  acknowledge  necessarily  tc  have  i n t e l l i g e n c e .  Intelligence,  knowledge  and  Al.  a  IMIS.: I n t e l l i g e n t  MIS and /Accounting Theory.,  24  1*2 ARTIFICIAL INTELLIGENCE. The  goal  sub-discipline  of  of  Artificial  Computer  Intelligence  as  a  Science i s described ty Nilsscn  (1971) i n these words: "The goal of work i n a r t i f i c i a l intelligence is to build machines that perform tasks normally requiring human i n t e l l i g e n c e . " Artificial  Intelligence  interdisciplinary  study  f o r more  boundaries  are i l l - d e f i n e d  and  disciplines,.  allied  researchers either  projects  Cognitive  l£i§IIi3§££S• Psychology  According to the Al  or  including  Intelligence, to  problems  regard  to  or  K. M. Colby  the  decades.  Its  interests  simply  at  as  cf the  allied  Artificial  to  Stanford,  Cognitive  is  processes as they  of human i n t e l l i g e n c e  reguiring  'humanity'  directed The  Artificial  intelligence  of the p a r t i c u l a r  at  with a l l i t s  seeks computer  human  4  are.  i t s pathological conditions.  as normally regarded, usually  an  by the work cf Simon and Newell  i s an understanding  failings,  two  been  be loosely categorised as  Simulation,  towards modelling human thought goal  may  Simulation^  Cognitive  Carnegie-Mellon  than  has  and shade o f f into i t s many parent  in  and t y p i f i e d  research  solutions without  solution.  The  See, for example, Newell, Shaw and Simcn (1958) and Newell and Simon (1972). One early management-directed model i n t h i s f i e l d was Clarkson's study (1961) cf the decision-making of an investment t r u s t executive.  I n t e l l i g e n c e , knowledge and A l .  IMIS^  immediate the  Intelligent  goal  is  ultimate  the  goal  the  of  A r t i f i c i a l however,  o v e r a l l  •  easily  •  c l e a r l y  because one  the  o v e r a l l  A r t i f i c i a l •general  thinking  or  of  that  to  has  Al  no-one are  models  cf  the  work  For  three the  achievements  the  geed  are  approached  ether.  to  brain  create with  and  even  not  modest  confidence  paths  model because  a  models  either  of  tc  be  Because  mechanistic  the  say  yet  attainments  possible CS  of  present  their  its of  to  net  MIT.  has  for  of  can  and  other.  a  yet  by  tc  and  those  lacks  task  way.  subfield  functioning  the  this  the  from  s t i l l  Intelligence  models  in  the  25  intelligence  at  distortion  s u f f i c i e n t l y  i n t e l l i g e n c e *  the  a  cf  t y p i f i e d  Laboratory  neither  goals  is  researchers  f i e l d  because  Psychology,  be  Thecrv.  mechanical  f i e l d  divided  distinguishable  Cognitive for  would  though  ultimate  of  Intelligence i t  from  second,  own  that  as  f i r s t l y ,  cross  i t s  f i e l d  This  Accounting  performance  creation  appplicajbilitv..  reasons,  and  successful  general the  MIS  a  human general  i n t e l l i g e n c e .  •  t h i r d ,  regard  many  workers  the  d e f i n i t i o n  of  In  elaborating  the  narrow.  in  A r t i f i c i a l  the  f i e l d  gcal  strongly  'engineering-oriented'  there  ever  view  can is  not  be  any  universal.  'general We  given  Nilsson  the  cf  by  goes  approach  theory  share  Intelligence Nilsson on  and  to tc  would as  adept doubt  i n t e l l i g e n c e ' .  concern  Intelligence,  too  expressed  knowledge  and  a that  This by  Al.  IMIS.:  Papert the  and  Intelligent  by  Reiter  development  viewpoint  in  the  the  evolution  methods such  of  notion  of of  to  'an  which  performances  planning,  the  applied to  to  that  represent  we  in  in  cne  deduction, recognition,  is  Al  with  from  draw for  26,  cf  cur  we  this  support  are  knowledge  implementable  as  generalizing,  It  epistemology'  embrace  Theory.,  engineering  epistemology.  readily  promise  accounting  Intelligence  methods  be  and  balancing  its  A r t i f i c i a l  for  designed  for  MIS  seeing  which  computer  are  systems,  overall  framework  induction,  learning,  memory,  imagination  and  expectation.  "These r e j e c t s many t r philosophy; i f seriously, cne Seeing Machine! Given  this  problem  domains  address  the  ideas pass a fundamental test that a d i t i o n a l notions in psychology and a theory of Vision is t c be taken s h o u l d be a b l e to use it to make a " (Minsky and P a p e r t , 1972,p2) viewpoint,  of  domain-specific  manner,  approaches  pattern  (1963))  -  are  i n t e l l i g e n c e Conversely,  •  of and  of  game  relevance tc  the  and  that  to  in  (see the cf  are  search  such  the  an  cr Uhr  the  cf  those  playing  concept  interest  deductive  us  knowledge  recognition  l i t t l e hence  as  to  Intelligence  of  such  particular  creating  seems  A r t i f i c i a l  representation  to  i t  which  extremely s t a t i s t i c a l  and for  Vcssler a  general  i n t e l l i g e n t subtopics  inductive  many  MIS.  as  problem-solving  systems, • modelling control knowledge, particularly •  modelling  •  robotics  of  human  research  systems when t h e  belief and  for using organized data base i s very large,  systems,  hand-eye  systems,  Intelligence,  knowledge  and  Al.  IMIS.:  Intelligent  MIS  and  Accounting  •  the  analysis  and  understanding  cf  visual  •  the  analysis  and  understanding  of  written  natural •  27  Theory  scenes, and  spoken  language(NL).  children)  learn,  the  world.  been  especially  r e t r i e v e In  5  this  and  use  respect  i n f l u e n t i a l  descriptive the  (Eiaget,  work  cf  models Piaget  1954;  cf has  Inhelder  S  Piaget,1964).  The  evolution We  stages,  of  see  roughly In  the  A r t i f i c i a l AT  research  as  in  sequence,  as  f i r s t  demand  i n t e l l i g e n c e  work.  That  encouraging the  success  the  were  machine  those  checker-player  See  period  enough. in  Charniak  Intelligence having  isolated solved cculd  terms 1967)  of is  three  evolutionary  follows: ££cblems  by  any  solve  Game-playing  (Samuel,  had  was  such well  method them  a  thought  programs  which  cculd  a l l  was  problem  and  at  t y p i c a l  to  as  Samuel's  known.  (1972,1973)  Intelligence,  knowledge  and  A l .  IMISi I n t e l l i g e n t  In  the second  credibility behaviour The to  HIS and  Accounting T hecry...  stage more a t t e n t i o n  problem  interest  domains remained  limit  excessive  in  game-playing  intelligent  time  or  storage  became  in  the  behaviour).  may  play;  cften  may  it  still  i t s cwn perceive  unsolved.  Authors f r e q u e n t l y expressed  hope f o r the g e n e r a l i t y of  algorithms  and  and  heuristics  enhancements might b r i n g such In  the  third  The  sophisticated,  the machine l e a r n  problems  said  demands.  more  h e u r i s t i c s f o r e v a l u a t i n g p o s i t i o n s or how patterns  intelligent  narrow but t h i s was  a d d r e s s i n g problems such as how  overall  focussed on the  of the methods as p o s s i b l e models of  (or p a r t i c u l a r a s p e c t s of  be to  was  28  indicated  that  their  'miner'  generality.  stage, which continues today, there i s  t h i s g r e a t e r i n t e r e s t i n r e p r e s e n t a t i o n s of knowledge and procedures  and  control  knowledge i n t e l l i g e n t l y .  structures  which  To t h i s extent A l has  p o i n t where some g e n e r a l i t y i s emerging. both  o r g a n i z e d combinations  6  are  6  See,  'artificial*  vision  parsed moreover with and  similarly pragmatic  •heterarchica1• c c n t r c l  the viewpoint of t h i s t h i r d  the term  robot  of s y n t a c t i c , semantic  knowledge guided by s i m i l a r From  a  Where, f c r example,  of  and  by  the  l i n g u i s t i c u t t e r a n c e s heard by the computer can fce spoken 'parsed'  viewed  reached  that  and  being  scenes  use  system  as  visual  might  in  stage i t i s r e g r e t a b l e that  i n t e l l i g e n c e came t c be a p p l i e d  f o r example, Huffman  systems.  tc  the  (1971).  I n t e l l i g e n c e , knowledge and A l .  I M I S I n t e l l i g e n t  f i e l d . of  The  7  knowledge  are  in  in  i f  neurophysiology least)  i t  into  is  are  gross,  humans.  tc  reasonable  i n t e l l i g e n c e ,  in  Al  mere  models the be  to  Acco untincj  the  Since  seem  and  emerging  machines  c r e d i b l e ,  knowledge  models  MIS  for  the  models less  regard  this  rather  the  in  that  they  representaticn  cf  psychclcgy  tentative research  than  29  representation  appealing  for  no  Theory..  (to as  say  cf and the  research  into  ' a r t i f i c i a l '  cf  researchers,  i n t e l l i g e n c e . Of as is  the a  course  quotations  meaningful  opinion  on  mechanical may  be  ancient  By  i t  and  is  a  basic  indicate, possible  such  that goal.  premature  self-consciousness  creative  or  'mind-body'  McCarthy,  in  conviction  a r t i f i c i a l  But  there  guestiens is  whether  an  intelligence  is  as,  possible,  such  Al  nc say,  whether  research  w i l l  received whether machines  c l a r i f y  the  problem.  the  mid  '50s.  Intelligence,  knowledge  and  A l .  IMIS:.  Al have  wish  to  of  these  KNOWLEDGE  l i J i l  What  IN  is  and  been  Accounting  1972;  space  and  L i g h t h i l l ,  here  to  a  EPISTEMOLOGY  AND  c r i t i c i s m  We  cf  Al  to  use  the  'knowledge'  word  following  "The r e s u l t s Intelligence new p h a s e o f  dc ncr  IN  epistemology  examples  of a decade have brought knowledge-based  from  Al  in  writers a  Ibid,  »o  Minsky,  an  et  a l ,  1973,  p.1.  Al  manner.  l i t e r a t u r e :  of work on A r t i f i c i a l us t o t h e t h r e s h o l d c f a p r o g r a m mi n g ^ 8  1  9  to  in  casual  " . . . the proposal of S. Papert and myself §^^~§tructure knowledge i n t o "micro-worlds"  Minsky  net  A l i  notion of p r o c e d u r a l embedding of knowledge p o p u l a r f o r some t i m e i n t h i s laboratory."  8  and  knowledge? of  "The been  197U).  KNOWLEDGE  viewpoint  the  times  attacks.  the  Consider  at  polemic  'defence'  From appear  even  30  T h e o r y  cvercptimistic  uninformed  Dreyfuss,  devote  MIS  have  informed,  1961;  analysis  l i l  researchers  drawn  (Taube,  Intelligent  Emphasis  has 9  tc  0  added.  p.78 (draft  Nov.73,  p1)  Intelligence,  knowledge  and  A l .  IMIS:. I n t e l l i g e n t  MIS and A c c o u n t i n g  " F i n a l l y , there i s a context world and t h e way that understanding of language."  31  Theory..  c f knowledge about knowledge affects  x  1  the cur  1  "... The l e v e l o f a l a n g u a g e i s t h e amount c f knowledge £1235 I l S i M S t y l e i m p l i c i t i n t h e l a n g u a g e p r o c e s s o r ..." "But i t i s p r e m a t u r e , we t h i n k , between any o f t h e s e : ~ -  t o propose a sharp  H a v i n g knowledge a b o u t how t o s o l v e a p r o b l e m , Having a procedure t h a t can s o l v e the problem, Knowing, a p r o c e d u r e t h a t c a n s o l v e t h e p r o b l e m ! " It  informal  would a p p e a r t h a t common-sense  dictionary  notion  of  'knowledge',  2  3  using as  1  boundary  1  A l w r i t e r s e i t h e r are  of  in  the the  entry:  " 1. Any f a c t o r t r u t h , o r t h e a g g r e g a t e of facts, t r u t h s and p r i n c i p l e s , known, a c q u i r e d c r r e t a i n e d by the mind",  or  they  are sharing  an  understanding  cf  'knowledge'  s p e c i f i c t o A r t i f i c i a l I n t e l l i g e n c e and need nc g l c s s t c make t h e i r meaning c l e a r t o o t h e r workers i n t h a t f i e l d .  Which i s  t h e case and what i s being denoted  i n such  c o n t e x t s as t h e s e ?  by  'knowledge'  The d i c t i o n a r y e n t r y quoted c o n t i n u e s :  "2. P r a c t i c a l u n d e r s t a n d i n g or s k i l l i n a n y t h i n g ; f a m i l i a r a c g u a i n t a n c e d e r i v e d from practise cr experience,"  11  Winograd  1  Sussman, Winograd, C h a r n i a k  2  13  (1972), p33  Minsky & P a p e r t ,  (1972,p.23)  (1972) I n t e l l i g e n c e , knowledge and A l .  IMIS.:  Intelligent  MIS  and  Accounting  Theory.  32  "3. In a s t r i c t sense (knowledge is) the c l e a r and c e r t a i n apprehension of t r u t h , or the agreement cf thought with thing, the conviction cr assurance, arising from proper evidence, that a mental apprehension corresponds with r e a l i t y or that which i t represents; assured r a t i o n a l c o n v i c t i o n . " Of  course,  sub-concepts. designed, use  of  we  The  net  so  to  'knowledge'. Bell of  as  to  have  definitions  f i t  the  purposes  They  may  not  (1973)  characterizations  poor  philosophers'  naturally,  example,  are  of  serve  offers  'knowledge'  cf the  every  enly knowledge  are  philosophers• purpose.  some  starting  three  For  alternative  with  the  statement  "I shall define knowledge as a set cf organized statements of f a c t s or i d e a s , presenting a reasoned judgment o r an e x p e r i m e n t a l r e s u l t , which is transmitted to others through some communication medium i n some s y s t e m a t i c form." He c c n t i n u e s "Thus I distinguish knowledge from news and from entertainment. . . . This d e f i n i t i o n is broader than some established philosophical e f f o r t s . . . . my d e f i n i t i o n is narrower, however, than Machlup's comprehensive c l a s s i f i c a t i o n (1962) . . . Machlup  1.  P r a c t i c a l decisions, his  P o l i t i c a l p r a c t i c a l  and  types  useful  actions;  can  into:  (a)  knowledge;  (c)  knowledge;  be  of  in  a  man's  subdivided,  Professional Wcrkman's  (e)  knowledge:  Household  work,  his  according  to  knowledge;  (b)  knowledge;  knowledge;  (f)  (d) Other  knowledge.  Intellectual c u r i o s i t y ,  five  knowledge:  a c t i v i t i e s  Business  2.  distinguishes  knowledge: regarded  as  satisfying a  part  a cf  Intelligence,  man's  i n t e l l e c t u a l  l i b e r a l  knowledge  education,  and  Al.  IfilSj.  I n t e l l i g e n t  humanistic  and  acquired  cultural  3.  a of  entertainment  5.  learning,  active  existence  general  culture,  concentration of  33  T h e c r v.,.  open  with  an  problems  and  and  pastime curiosity emotional  knowledge:  satisfying  the  or  his  desire  stimulation,  . . .  fcr apt  light to  dull  sensitivness.  S p i r i t u a l of  the  and  non-intellectual  4.  rule  accounting  values.  Small-talk  his  ana  s c i e n t i f i c  as  appreciation  MIS  God  and  knowledge: of  Unwanted accidentally  the  related  ways  knowledge: acquired,  to  the  to  his  religious  salvation  outside aimlessly  his  cf  the  interests,  knowledge soul.  usually  retained."  Intelligence,  knowledge  and  A l .  IMIS:.  J[_.  2  The  Intelligent  s £ i s t e m o l o g i s t ^ s  S£i§temology concerned than  with  with  with  "How  is  in  the  may  this  34  Theory..  knowledge^  philosophic  items  of  "Do  study  d e f i n i t i o n  we  philosophic  cf  of  when  our  know  some  sense  of  knowledge,  "knowledge"  knowledge.  discover with  Accounting  It  is  beliefs  rather  concerned are  particular  knowledge  is  that  valid  datum?". Quintcn  says  "According knowledge  to is  the  as  to  look  ' t r u t h ' ,  most  j u s t i f i e d  being  opinion at  ana  rigorous  we  than  (1967,p.315)  Philosophy  the  x  particular  knowledge?" It  MIS  what  widely true  i t  this  d e f i n i t i o n and  d e f i n i t i o n ,  belief  what  ' b e l i e f  is  accepted  there  is  room  i t s e l f  fcr  may  difference  mean.  Let  us  ' j u s t i f i c a t i o n ' .  Intelligence,  knowledge  and  of  Al  I M I S j.  I n t e l l i g e n t  MIS  and  Accounting  Thecr  35  ITruth^i (a) which  is  It  true  can is  which  are  true?  If  there  would  deny  one  word  would  reduction candidates  not  of  be  knowledge. merely are  the  then  a b i l i t y  geometry  to  empty. and  in  But  thought  none  become  for  argued  the  extreme  are  to  be  there true  holding  but  this  of  Since  the  time  to  logic  s e l f - c e r t i f y i n g  having  statements  only  that  any  statements  are  necessarily  extreme  speak  analysis  that  viewpoint  knowledge. cf  the the  have  The  e f f e c t i v e p r i n c i p a l  teen  reduced  t o r -  Intelligence,  knowledge  and  Al  IMS:.  (1)  a  then  which  says  about  "If...  A",  verbal which  £ l i o r i  HIS  so-called  something  (2)  I n t e l l i g e n t  logical  definitions c l e a r l y  add  existing  data  examples  below  define  a  concept  denote  that  and  truths  nothing then"  about and  such  as  nothing  but  structures. of  Accounting  labelling with  the  such any  1  "Kittens new See  the  "If  Quine  A,  but  young  tckens (1953)  'semantic  A  V  A  only  4  are  label  word-symbol  as  thing.  " c r " .  36  Theory..  cats."  tc and  cne's the  networks' tc  be  used  Al  which tc  concept.  It is d i f f i c u l t t o s e e by what t e s t s the t r u t h s of lcgic and mathematics labelled *a p r i o r i a n a l y t i c ' by Kant can be d i s t i n g u i s h e d from e m p i r i c a l . These truths are indeed both true of a l l that we know cf this universe and alsc compelling to the mind. But that they are 'true fcr a l l the u n i v e r s e t h a t we k n o w ' is evidence guite as much for their empirical status as for their analytic status. Nc 2££21iillliil f o r r e f u t a t i o n o f t h e i r a l l e g e d l y a n a l y t i c s t a t u s available.. Thus t h e i r apparent logical truth r e f l e c t s only their highly corroborated empirical truth. This view would r e g u i r e us to regard some sort of "uniformity of nature" or "Law o f l i m i t e d v a r i e t y " axiom as true - but only of this universe, i . e . empirically. Despite their privileged, apparently analytic position, the meanings of "and", " o r " , "not", and " i f . . . t h e n " are as much abstracted symbolic descriptions which our minds have c o n s t r u c t e d the better to grasp r e a l i t y as i s , say, "chair". The' significance of this viewpoint to the a p p l i c a t i o n c f Al is that i f i t is a tenable viewpoint then 'Semantics' the mechanisms by which we r e l a t e d e s c r i p t i o n s cf knowledge tc each other to y i e l d meanings is more fundamental than logic. See Quine, (1953, 1960). 1  4  Intelligence,  knowledge  and  A l .  IMIS.:  (3)  I n t e l l i g e n t  i n c o r r i g i b l e pain!"  or  light."  can  "I  s a me  want  the  such  time  to  go  truth  upon  report  and  introspective  Their  contingent  MIS  to  Accounting  Tbecrj,.  statements  such  sleep!"  values  are  experience  believe  them  wholly  they  experiences  or  see  report.  be  as  "I  am  a  green  c e r t i f i e d  f a l s e l y ,  to  "I  37  we  in  by  and  Although  we  cap.net  true..  at  the  (Quinton,  1 9 6 7 , p . 347) .  (4)  certain Two  attempts  appealing  extended" neatly  empirical,  If,  word  believed  to  no  true  longer  by  be  -  of  and the  We  have  must  tc  (1967)  can  thereby  grounds We no  for  have access  no to  are  be  that  which  later  be  in  fact  and  raises  b e l i e f ,  direct  is  have  weaker 'an  Intelligence,  merely tc  meaning  aggregation the  that  truth  issue i s ,  if  a  be  formerly  cf  the  apprehension  absolute  as  believed  tc  This usage  the  i t s e l f "  'knowledge'  ceasing  popular  is  synthetic'.  include which  to  truths,  coloured  precedes  accept  p r i n c i p l e s ' ,  l j u s t i f i c a t i o n ^ i  world.  are  synthetic  "Anything  temporarily  'knowledge'. the  p r i o r i  these:  Salmon  that  without  a  posteriori  we  true  resembles  sufficiency  'a  i t  knowledge  of  is  label  truths  event  then  the  of  Wesley  that  deserved  facts,  "No  however,  meaningful  finding  examples  and  shown  at  i t  knowledge  cf  of of the  issue the  exists.  and  A l .  IMIS.:  Intelligent  A l l  that  we  have  and  our  internal  approximation knowledge  support  a i t  be  in  for  watershed from  cf  may  calibrated  a  at,  conjecture, grounds contrary  (cf.  models  r e a l i t y .  (See  or  an  of  opinion  'faith*)  be  *degrees -  to  ' b e l i e f form  to  or  in  a  be is  scale  even  on  3). a  of  b e l i e f '  cr  cn  which  ordinary -  spite  of  cf  b e l i e f s '  to  describe to  a  label  promising  grounds, a  and  subjective  language perhaps  seme  'degree  'mere  i n s u f f i c i e n t in  cf  continuum  some  Thus  organs  Eelief  cn  separates  38  sensory brain  areas  knowledge' held  ty  the  knowledge'. in  Theory..  Appendix  confidence  considered  'weak  mediated  symbolic  99.9$  a  Accounting  sensations  proposition  as  and  considered  say  proposition a  our  'confidence*,  ' b e l i e f s  as  is  MIS  cn  wealth  nc of  evidence.  "It should be n o t e d t h a t where knowledge and belief overlap, the kind of knowledge involved is propositional knowledge, or what Ryle (1949) has c a l l e d "knowledge that". There is also what Ryle called "knowing how" (tc s k a t e , tie a reef knet, do long division) w h e r e t h e r e a r e no p r o p o s i t i o n s to be true or f a l s e . " (Quintcn,1967, p.346)  Intelligence,  knowledge  and  Al.  IHISl  Intelligent  MIS  and  accounting  39  Theory.  " J u s t i f i c a t i o n ^ " and strongly not  more  held.  qualify  A  as  gives  i t  strong  belief  is  required  lucky  knowledge  the  highest held  on  other  or  b e l i e f s  " j u s t i f i c a t i o n "  leads  i n t u i t i v e usually  can  offer  'basic  but not  Since  empty,  The  a  to  only  the  the  contingent  be  which  knowledge,  ignored; on  truth  set  of  'truths  urging  cf  a l l  a  statements'  but p r i o r i  this of  a l l  can  have  in  be  the  of  factual  on  discussed  way  is  by not  j u s t i f y i n g appears  epistemclogy  mathematics' above.  into  s c i e n t i f i c  truths  c l a s s i c a l  and  this  regress  mutually  synthetic  are seme  broken  Indeed  a  ty  knowledge  only  does  teen  what  supported  cf  is  evidence  And  Eut  our  b e l i e f  fanatic  must  be  true.  regress logic  a  grounds.. must  the  later  It  d e f i n i t i o n  regress  terminate  that  support.  to  the  just  because  b e l i e f  known  uninferred  rest  b e l i e f s ^ be  or  into  broken  theories  to  and  merely  the  s u f f i c i e n t  grounds"?  c i r c u l a r i t y  cr  possible  " s u f f i c i e n t belief  guess  than  and  the  Neither  seem  adeguate.  Intelligence,  knowledge  and  Al.  IMIS.:  iiii  Intelligent  ARTIFICIAL  We  INTELLIGENCE  believe  generated  by  than  a l l  However, the  (we Al  neurophysiology know  MIS  we  reveal  past  must  pragmatic,  that  •knowledgeable ,  i . e . or  "the  is  s k i l l " .  ^.assured  the  at  for  them  Waltz,  5  in  f i r s t cf  is  which  when  present  of  simulating  and  imputed  when,  say,  fund  'vision model and  to  Al  cf  children  analysis*  i t s  Fcr  which  goal in  p r e v a i l :  net  d i r e c t l y  itruthj^ everyday  1973) about  stcries  Guzman,  and sense  'knowledge  'commonsense  written  Winston  knowledge'  and about  shading.  Appendix  3.  Intelligence,  5  this  C h a r n i a k (1972,  by  1  understanding  for the  we  i n t e l l i g e n t ,  are  concern in  as  human.  knowledge  and  which  behaviour  " p r a c t i c a l  Thus  by  has  subcencepts  to  is  see  Al  a  twe  data  philosophy.  by  used  solids  of  performed  the  connection  processes  judged  with  within  the  time  psychology  be  facts"  being  in  would  c o n v i c t i o n *.  surfaces,  this  task  i f  (1968,1970,1972)  vertices,  1  cr  that  references  length  •PIGGY-BANKS'  of  40  JlKllOjjLEDGEM .  cognitive  epistemclcgist•s  rational  knowledge  deals  the  Al  that  know!)  behaviour  aggregation  addressing  of  such  1  i t  net  recognise  where  reason  UPON  more  instrumental  Theory..  COMMENTARIES  introspections  domains is,  Accounting  research,  w i l l  the  do  and  knowledge  and  Al.  IMIS.:  It  is  understood  knowledge Their we  I n t e l l i g e n t  may  be  concern  believe  which  we  turn  only  a  l i t t l e  i t s  know?"  next. to  What  pursuit  with  i t  the  of  acguire,  what  true  is  the  answers  we  to  and tc  this  b e l i e f .  use  1  "How  that  this  which  question  arises  cf  the  tc has  applied  incidentally  have  called  'a  science  this  l a t t e r  area  cf  6  may  Intelligence  epistemclogy say  that  question:  A r t i f i c i a l  'the  have  modify  41  Thecry..  judged  non-recursive  about  does  already  ' j u s t i f i e d  short  say  Accounting  have  It  In  and  as  structure,  we  sciences*.  they  regarded is  e f f i c i e n t l y  we  that  MIS  in  applied  epistemology'. We a l l the  its  turn  richness  a r t i f i c i a l  interesting.  •  therefore  represented  •  dynamic  Static or  (or,  are  which  as  either  seem  research  to  and  us  which  select  to is  from  characterise epistemically  follows:  eguivalently, by  sets  s t a t i c  of  beliefs)  r e l a t i o n a l  may  descriptions  be cr  procedures..  r e l a t i o n a l  graphs,  points  intelligence  These  Knowledge  by  four  to  may  descriptions, encode  in  the  form  axiomatizaticns  of in  networks  f i r s t - c r d e r  I*  It should be noted that t h e r e have been A l p r o j e c t s in which every p r o p o s i t i o n i n the data base has associated with i t a 'degree cf b e l i e f datum, w h i c h may b e a measure s i m i l a r to a p r o b a b i l i t y measure. Kling (1973) among others, has proposed an implementable lcgic to infer conclusions (with their 'degrees of b e l i e f ) from such assertions.  Intelligence,  knowledge  and  Al.  IMISj,  Intelligent  predicate  but  knowledge  in  subsume  preferred  means  a b i l i t y  to  an  Minsky is  to  are  actions  others  It  deductive  mere  powerful..  and and  are  the  strategies  inference  machine.  that  instrumentally  with  I § £ i e s § n t a t i o n s  descriptions  encoding  practice  i n t e l l i g e n t and  programs)  42  Theory,.  for  knowledge.  ifhayiour^  (i)  for  Accounting  procedural  r e l a t i o n a l  manipulating  An  and  calculus  (embedding They  MIS  pure,  However,  needs  tc  higper^crder  be  i t  f i r s t - o r d e r  i n s u f f i c i e n t  to  is  necessary  is  argued  deductive  simulate  by  logic  i n t e l l i g e n t  augmented  propositions  about  prcblem-sclving  procedures,  (ii)  with  analytic  (iii)  semantic  propositions  with  heuristic  non-logical.  These  Theorem-proving whose the  goals  e f f e c t i b l e This  c l a s s i c a l  systems  actions  would  w i l l  may  appear  deductive  which be  may  be  the  satisfaction  though to  be  are  i l l u s t r a t e d  (especially  operate as  (interpretations)  cf  the  and/or  aids  include  real-world)  models  reasonable  in  due  those cf  successfully they an  were  r c t c t i c s purpose  by  true  marked  course.  in  seme  but  in  regarding  prepositions.  departure  from  logic.  Intelligence,  knowledge  and  Al.  IMIS;.  We  shall  elaborate  •classic' from  they  a  our  small  knowledge'. supports l o g i c a l  of  HIS  of  and  these  the  Accounting  points  and  A r t i f i c i a l  The c r y ,  then  43  present  Intelligence  seme  research  arose.  ESlSticnal For  mind  each  examples  which  1±BL±1  I n t e l l i g e n t  descriptiensj,  networks  i l l u s t r a t i o n s  in  set  The  of  l i s t  curious  will  i n t e l l i g e n t  remind  and  graphs..  this  chapter  but  selected  us  that  performances  is  the  not  let  us  keep  'items  of  knowledge  limited  tc  in  which simple  predicates.  Consider: 1.  "The  meaning  of  the  symbol,  2.  "The  meaning  of  the  symbol,  3.  "The  meaning  of  the  r e l a t i o n ,  4.  The  his  aunt  assertion:  confused  had  "Charlie  never  when  said  driving  in  "The  s k i l l  of  flexing  6.  "The  s k i l l  of  riding  "The  ' a ' , ' b ' , ' c  s k i l l =  a  +  b«  his  one's a  ' . "  'CN  TCP  that  he  mother  C F . ' " believed  had  seldom  yields  c  when  'find  c ' "  that become  the  finger."  b i c y c l e " .  that and  'CHAIR  fog."  5.  7.  said  that  'EICCK.'"  goal  given  8 . "The suspicion w h i c h may s u d d e n l y be acquired (that is, an 'attention interrupt') when c n e s w i n g s a d e e r shut at time t1 and has not heard a c l i c k by time t2=f ( t 1 , c h a r a c t e r i s t i c s of a c o r ) , that the action i n i t i a t e d was n o t c o m p l e t e d successfully." 9. is  "The certitude that i f ' t h e day has become true then ' t h e day i s no l o n g e r T u e s d a y * is  Intelligence,  Wednesday' true.  knowledge  and  A l .  IMIS.:  I n t e l l i g e n t  How may  test  whether  requiring in  may  some  response  on  the  an  entity  a  cf  such  masterly  micro-fiche  r e t r i e v a l  •knew would i t  would  i t ' . be  would  to  to  is  theory.  a  It  complexities  or  the  served  aunt)  could  digit,  thereby:  inexpensively.  transferring of  the  by  be  a l l  signal  of  Item  minimally  knowledge by  somewhere  are  few,  #4  l i s t  for in  cur as  a  the cf  the  If  the  simple  and  represented (Charlie's  single  capturing the  a  information  symbol.  be  in  in  between  probably  represented  r e s p o n s i b i l i t y  .to  cede  which  knowledge.  complexity  that  the can  the  by  of  analysis  and  data  knowledge  means  of  MIS  the  for  item  keyed  some  storing  trade-offs  symbol  purposes  and  to  to  item  the  back  consider  for  devise  quantum  amenable  of  simply  to  f a l l  say,  test  by  knowledge  must  as,  Turing  We  knowledge  example,  entity  refute  s p e c i f i c  the  we  44  knows?  the  suffices  entity  subject  then  fcr  the  environment  predetermined  not,  powerful  is  be  cf  again  that  encoding-decoding to  certain  Once  (which  the  problem  entity  performances  verify a  Theory,  another  possesses  would  very  Encoding computer  we  require  seek  that  re-presentation  indicate  A  Accounting  stimulus.  test:  purposes)  (a)  judge  Turing  that  and  we  form  to  MIS  the  binary  meaning  decoder-recipient  environment,  (b)  losing  a l l  mutability  in  the  representation  Intelligence,  knowledge  (except  and  to  A l .  IMISj.  (c)  Intelligent  negate  it) ,  losing  a l l  unexpected  as  signal  #4,  that  "Did  as  early  Al  added  "indexing"  1968),  subtle  complex  #4  SAD  related denoting Charlie,  by  knowledge  in  example,  45  strings  cf  adopted to  SAM  may  (and  predicates. car,  not-ever-said,  fog;  In the  mother-cf,  in  Item  as  it  text.  Some  representation  with  cf  answers  to  (Green  et  EASEEALL STUDENT  (Bobrow,  etc.) Al  practice  the  sense  being  encoded  in  some  in  which  nodes  are  fundamental  edges  network  cr  graph  fundamental  a r b i t r a r i l y )  this  is  l i t e r a l  1963),  t y p i c a l  arbitrarily) (and  extreme  r e t r i e v a l  a  such  another  allow  by  tc  Such  a  captured  r e l a t i v e l y  almost  and  drive?"  such  1966),  structure:  At  o r i g i n a l  (Lindsay,  following be  us.  (See  further  respond  mother  represented  devices  to  interest  (Keizenbaum,  data  r e l a t i v e l y  the  guesticns.  ELIZA  Item  use  not  be  research  F i n a l l y , of  Theory..  Charlie's  do  could is,  a l . , 1963),  Accounting  (For  representations  more  tc  ways.  this  and  and  a b i l i t y  questions  MIS  example  predicates s i s t e r - o f ,  Intelligence,  the  nodes  might drives,  be:  might  be:  believes,  etc.  knowledge  and  Al.  FIGURE 1.1:  A NETWORK DESCRIPTION.  IMISj.  Intelligent  There  are  representations example, belief or  such  as  systems  or  a  that  of  'training  Accounting  Cognitive  Semantic that  S Smith  of  by  (see  Tesler,  (1965),  (1969)  such  Psychology.  Memory  and  and  network  Consider  below)  Fnea  47  Theory..  and Cclby  work  fcr in  (1968)  Carroll,  Abelscn,  Reinfeld  work  learning  systems  in  Winston. the  following  final  learnt  by  Winston's  sequence'  of  arches  illustration  descriptions  and  and  examples  Carroll  example  AN ARCH i s  the  Al  such as  and  Colby  For of  many  Quillian's  Abelson  (1963)  in  MIS  from  is  from  examples",  his  descriptive  program  and  'near-miss'  thesis  quoted  during  "Learning  in  Minsky  definition exposure  tc  ncn-arches; structural and  Papert  (1972).  Intelligence,  knowledge  and  Al.  FIGURE 1.2:  DESCRIPTIVE NETWORK REPRESENTING ARCH.  IfilSjL  A  set  c i r c u l a r .  (a)  ( ) b  of  The  semantic by  Intelligent  its  networks  relations  of  we before  i f  network by  empirical  truth  example:  A  (i)  of  1  7  the  i t  on  for  'red  need  the  words  same  be  node  to  is  be  a  given  and  defined  dictionary.  1  in  7  i n f i n i t e  regress  computer-implemented  resolved axiom,  hand-eye by  a  are  or  in  some  tends  nodes,  the  same  contingent  procedure.  system  ultimate  in  seme  truth-returning  cube'  of  edges  as  defined  truth-returning  example  might  Fcr  establish  reference  " s k i l l  knowledge"  as  learning to or  of use  s k i l l " s k i l l s one's  by  -  our with  tc  a twe  quoted by  Sussmann  programs  knowledge  case  for  such  d e f i n i t i o n  dictionary by  this  and  (tests)  dictionary the  in  contrasts  procedures  compare  as  execution  cn  some  •'knowledge  a b i l i t y  to  cf  other  s§£se-imjDressions_:  i n c o r r i g i b l e  For  the  resolved  terminating cr  meaning  to)  49  between  c i r c u l a r i t y  be  w i l l  descriptions  which  of  Theory..  truths:  impressions  (ii)  in  this tc  the  (edges  vision-analysis  meaning  classes  is  Acccunting  apparent  which  vocabulary  i t  manner:  true  in  is  with  encountered  and  r e l a t i o n a l  d e f i n i t i o n s  the  and  semantic  such  s i m i l a r i t y  dictionary terms  MIS  above  d e f i n i t i o n in  (1973),  e f f e c t i v e l y  his  cf  thesis  namely: and  of on  "The  readily  in  performance."  Intelligence,  knowledge  and  A l .  IMIS:.  Intelligent  attributes  as  frequency.  It  then  these  l^HiZ  to  an  world' of  atomic  exercises.  i  8  is  a  few  cf  i f  and It  can  be  set  (The  by A  a  is  r e l a t i v e l y  When  I  the  heard me,  when  was  I  the  proofs,  before  cube'  items  50  and  lightwave  s i g n i f i e s  means.  shown  CT. c u r  s t a t i c  term  1  simple in  i n t u i t i v e l y  most  f e l t  recognition outside  more  reader  the  anything  8  1  9  learn'd  as  EIOCK  cf  the  that,  of  a  entity  invited  tc  would  appear  descriptions  at  BLOCK  system  complex is  l i s t  network  as  primitive  It is said t h a t A. K c r z y b s k i this?" Receiving the reply "Matchbox i s a noise — this the bcx at the respondent.  when  "means"  'red  the  well  desired^  description.  Theory..  r e c t i l i n e a r i t y  that  what  CHAIR.  symbol  description  CHAIR of  a  and  and  us  are  tolerably  BLOCK  demanding  to  Accounting  descriptions..  only  served  perhaps i s  seems  Procedural  be  and  egual-sidedness,  r e a l i t i e s  But  HIS  cr  Al  ER1CK 'blocks  the is  level net  toe  boundary. at  the  same  define  a  But level chair  would ask a c l a s s "What is "A m a t c h b o x " he w o u l d answer is what i t i s ! " and hurl  astronomer,  the  figures,  the  charts  were  and  ranged  in  columns  diagrams,  tc  add,  he  lectured  d i v i d e , and measure them, when I s i t t i n g heard the astronomer  where  w i t h much a p p l a u s e i n t h e l e c t u r e - r o c m , How s o o n u n a c c o u n t a b l e I became t i r e d and sick, t i l l rising and g l i d i n g cut I wandered o f f by myself, in the mystical moist night-air, and from time tc time, look'd  up  in  perfect  silence  at  the  Intelligence,  stars. (Walt  knowledge  Whitman.)  and  A l .  IMISj. I n t e l l i g e n t  with  an  analytic  HIS  definition  alpha and  beta p r o b a b i l i t i e s  matching  a  long  not-chairs). static  and  series  such of  of  Accounting  that there are acceptable  the  description  presented  C HAIR-RE COG N IS EE ,  a  CHAIB-CONSTRUCTCR which can build  would  correctly  images cf chairs  However, i t would most l i k e l y  description  51  Theory.  be agreed  suffice  fcr,  that say,  NIAR-CHAIR-EISCRIMINATOR,  (something  assemblies  and a a  or  a  analogous to the extant systems  of blocks from  a  description  cr  copy an assembly of blocks with another). But  #3  for  primitive r e l a t i o n , and  for  no other  "ON  TOP  OF",  was  regarded  the guestion seems to have  teen  'items of knowledge' i n cur l i s t  d e s c r i p t i v e networks seem wholly to  which  adeguate.  We  as a  begged, dc  static  therefore turn  procedures^ In the case where the knowledge  itself  procedural  as  with  "flexing  tc  be  one's  impounded finger"  procedural model seems t c be the most appropriate. "bicycle-riding",  which  i s also procedural, was  i l l u s t r a t e an item of "knowing enter  the  actions. as  with  of  time,  into  causality  whose  or  first  inelegant and such  predicate  inefficient  concepts.  three with r e l a t i v e  order  i f they  Item  #6,  selected  to  conditional descriptions  calculus are  then a  definition  and  I t i s well known that such r e l a t i o n a l  networks  extremely deal  concepts  how"  is  become  required  Procedural statements  tc  handle a l l  ease.  I n t e l l i g e n c e , knowledge and A l .  IMIS.:  Let  Sussman  I n t e l l i g e n t  (1973,p.12)  HIS  and  summarise  Accounting  their  52  Theory.,  advantages:  "The expert program offers performance. I claim that the performance is due t o t h e procedural representation of knowledge in the expert. For each kind of problem i t k n o w s how t o s o l v e , a procedure is available which, when interpreted, performs the required manipulations in a directed way. It i m p l i c i t l y "knows" (embedded i n t h e s t r u c t u r e o f the p r o c e d u r e by t h e programmer) the c o r r e c t seguence tc perform, including the order of steps a n d how tc avoid or f i x destructive i n t e r a c t i o n s between them. It k n o w s how t o s e t up f o r s t e p s that will later need some c o n d i t i o n t o be t r u e by p r i o r e x e c u t i o n c f steps which achieve the r e q u i r e d prerequisites".  This  increasing  representation quite  general  of in  preference  knowledge  over  fcr  static  procedural  representations  is  Al.  "If we were to e l a b o r a t e our t h e s i s in f u l l d e t a i l we w o u l d p u t m u c h m e r e e m p h a s i s cn procedural (program-like) descriptions because we b e l i e v e that these are the most useful 8 versatile in mental processes." (Minsky 8 Papert, 1972, p.5)  In  similar  entire we  vein,  viewpoint  believe,  a  alternatives  to  that  procedures  systems judges domain that  the to  be  in  domain.  acknowledge  Hewitt, as  the  "procedural  preference static  choice e f f i c i e n t  See the  light  of  which  that  but  rather  As  representation  within of  what  what  Churchman,  one  (1971).  superiority  describes  of  This  holds  descriptions.  cf  PLANNER  epistemology".  descriptions,  subsume  the  author  one judges  net,  procedures  usual  a  are  recognition  in  normative  depends  cn  what  ene  judges  tc  be  the  to  be  Therefore, procedural  Intelligence,  is  his  the  goals  of  although  we  embedding  cf  knowledge  and  A l .  IMISj.  knowledge  Intellicignt  i t  may  MIS  be  and  Recounting  foolish  to  53  Theory..  polarise  the  two  representations:  •  filSi  heavily i s  to  be  • the  on  because  the  the  domain  used,  choice  and  cn  the  way  in  which  will  the  depend  knowledge  because  at  some  level  the  one  may  have  need  of  other.  the  example,  relation  we  have  previously  ON-TOP-GF  more  e f f i c i e n t  attribute  than  some  LISP  The  very  domain-dependent.  subsumes about  a  great  such  concepts  extension,  domains  In  far and  procedure  simplicity  of  as:  The  -  always  a  d i f f e r e n t  from  symbol  for in  in  which  example a  in  and and  context  cf  imagine  to  circumstances  be in  primitively which  is  for an  teen  usage  is  word  best in  Intelligence,  is fact  s o l i d i t y , frames  cf  r e l a t i o n  which  envisage  i t s  the  as  a use  meaning  we  procedural about  in  modelled  Again,  knowledge  elusive  may  with  the  knowledge  this  one  and  used  balance, spatial  action,  'PLACE-ON-TOP-OF-(IF-PGSSIBLE)". " f i n g e r - f l e x i n g "  test  has  Similarly  'blocks'  simpler  common-sense  "ON-TOP-OF" a  r e l a t i o n  "ON-TOP-OF"  support,  context  primitive.  usages  which  physical,  surfaces,  to  simplicity  gravitaticn-and-mass,  reference. be  deal  seems  function But  attribute. much  of  The  c e r t a i n l y  "ON-TOP-OF-NESS". notion  called  •primitive".  'B10CK3- (ON-TOP-OF)-XBLOCK78'  i s  representation  and  §§£2ii5  For  may  of  tcck but  we  may  f i n g e r - f l e x i n g  knowledge  and  A l .  IMIS.:  would our  well  I n t e l l i g e n t  suit  'obvious'  our  choice  we  selected  p.123)  of  uses  a  net-work  finger"  cerebellum packaged  l e v e l  at  the  may to  which  in  then of  d e t a i l s "  (1973,  the of  p.3)  muscle the  the  volition  "I  passed  replay  of for  performance  networks  and  of fact  (1972,  interesting  is  commands  In  Eccles  describes  entire  compares  action.  He  a  wcrds,  side-effect  because  cerebrum  i n i t i a t e s  other  an  the  54  Theory^  In  merely  only  that  the  be  observe  example.  evidence  arising  c h a r a c t e r i s t i c  this  which  same  Accounting  representation.  ' f i n g e r - f l e x i n g '  description  Moore  and  c l a s s i f i c a t i o n  neurophysiological my  MIS  shall  flex  tc a  the  complete  the  "fine  requested.  procedures  in  way:-  "Methods advocated for representing knowledge in a r t i f i c i a l intelligence programs have included l o g i c a l statements..., semantic networks..., and procedures (Hewitt, Sussman S McEerraott). A l l these approaches share one f u n d a m e n t a l c o n c e p t , the notion of p r e d i c a t i o n , . . . In this respect the various s y s t e m s a r e more o r l e s s e g u i v a l e n t . But this basic idea must be extended to handle problems cf q u a n t i f i c a t i o n and knowledge about knowledge. Here the s y s t e m s do d i f f e r . We w i l l a r g u e , however, that these differences result from the descriptive apparatus used,,, rather than from an inherent advantage of, say, procedures over declaratives cr vice versa.  Intelligence,  knowledge  and  Al.  IMIS.:  Intelligent  MIS  and  Accounting  55  Theory..  Advocates of PLANNER (e.g. Wincgrad, 1972, p.215) have argued t h a t the predicate calculus cannct r e p r e s e n t how a p i e c e o f k n o w l e d g e should be used. But this is true only cf the f i r s t - o r d e r predicate calculus. In a higher-ordered declarative language statements could b e made w h i c h w o u l d t e l l a theorem p r o v e r hew o t h e r s t a t e m e n t s a r e t c be u s e d . PLANNER, o n t h e o t h e r h a n d , h a s n o way o f d i r e c t l y s t a t i n g an e x i s t e n t i a l q u a n t i f i c a t i o n , but this dees not mean that procedural languages are necessarily incapable of handling that problem. Our b e l i e f , then, is that the type used to r e p r e s e n t knowledge is unimportant, as i t has s u f f i c i e n t e x p r e s s i v e p o w e r . " 2  2  °  2  1  we cannot reading into this "Only i f the language has the language w i l l i t then s u f f i c e ; i t such"!  0  of system sc long 2  1  e n t i r e passage the sense |cser cf a procedural although we n e e d n e t c a l l  Speaking c r i t i c a l l y cf overuse of the c l i c h e s "Meaning procedures" and " H e t e r a r c h y not h i e r a r c h y " W i l k s says:  is  ' T h e r e may b e s o m e t h i n g i n b o t h c f t h e s e theses, but I h a v e n e v e r s e e n e i t h e r o f t h e m s t a t e d i n s u c h a way a s tc make clear that they don't mean t h e t h i n g s philosophers interpreted them as meaning at v a r i o u s times in the past. For in those senses the statements are pretty straightforwardly false. . . . The f i r s t . . . would m e a n t h a t we c o u l d n c l o n g e r usefully distinguish between words whose meanings plausibly are procedures (like "unscrew") and words whese meanings c l e a r l y aren't like "mud".' Y.  Wilks  in  Anderson  Intelligence,  et  al.  knowledge  (1974)  and  Al.  IMI§1  Intellicjent  Illustrating and  procedure  transition fact,  network  the  t r a n s i t i o n is  a  same  (1972,  "If  the  could] For i t into a  seen  cf  flowchart  is  p.51)  refer  of  Winston's  to  this  program  PBOGRAMMAR?  also  and  procedures  that  a  easily  a  recognizer  semantic  network  say  objects  see  to  procedural At  the  procedural,  required,  for  to which  was  more  the l i k e  and  ether..  might  Intelligence,  We  employ  relations  a  a tut  recognize system  executive  red  network  to  such  [we  described. description  each  the  network a  Minsky  that  cube'  store  pattern-matching  compare  augmented  present,  for  'red  level  doing  procedure  i t .  primitives  high  In  correspondence:  were  uses  an  any  i l l u s t r a t e  have  about  just  of  also  the  [augmented  talking  re-representation  "egual-sidedness".  to  our  with  56  description  do  r e l a t i o n a l  need  need  ways  Indeed  "How  of  may  Quillian-type  also  says compare  two  thing...  r e l a t i o n a l  make an a r c h by t h e p r o c e s s just is net hard to convert the . . . procedure for building arches."  descriptions  would  just  of  Theory..  flowchart  use  We  have  Accounting  (p.44)  grammars]  network".  s t a t i c ,  Papert  are  and  equivalence  Winograd  programs  exactly  the  MIS  if  wculd it  descriptions  cf  were two  cube.  knowledge  and  A l ,  IMIS:.  It  seems  embedding' important  (i)  I n t e l l i g e n t  of  are  (as  they  2  2  i t  and  causes  can  seems  These  for  'procedural  strength  from  several  or  causal  are:  with  a  select,  with  the  contain  far  a  temporal  or  with  as  a  easier  1.4.4  world become  in  and  other  or (as  knowledge  modify when  true.  a 2  lower primitive  2  descriptions  (plans-  subset.  use  section  to  order  r e l a t i o n a l  about  tc  real  proposition  knowledge  See  its  dynamically  conditional how  case  57  Theory,.  plans-withzactioiu  procedures  withgut^acticn]_  (v)  draws  interface  procedures)  (iv)  the  descriptions  are  can  when  action  Accounting  added..  procedures  they  and  that  knowledge  procedures  (iii)  us  c h a r a c t e r i s t i c s .  dimension  (ii)  to  MIS  procedural and  notations  meta-knowledge  structure  to  model  i . e .  both  knowledge  knowledge.  below.  Intelligence,  knowledge  and  Al.  IMIS.:  ! i i J i l  The  Intelligent  yalue  of  MIS  and  semantic  Accounting  models  58  Theory,.  in  aiding  deductive  inf® .§ . J ..s. r  n  c  s  Deductive evolution  of  A l .  logic  has  played  a  It  have  played  an  may  major  role  excessive  in  the  role.  "Insistence on c l a r i t y at a l l c o s t s is based on sheer s u p e r s t i t i o n as t o t h e mode i n w h i c h human intelligence functions. Our reasonings grasp at straws for premises and float on gossamers fcr deductions." A.  It domain  has  of  been  an  object  (allegedly)  i n t e l l i g e n t  systems.  been  to  used  study,  i n t e l l i g e n t Forms  express  Sandewall,  of  of  the  natural  1970)  and  In  themselves  home  at  theorem  provers  domains  in  bases  may  2  3  It  in  axiomatic then  describing problem,  and  the 2  seems  major,  be  a  to  circumstantial Hempel's  us  *permanent'  (Darlington,  1964; complex  (Quillian,1968;  to  have  devise  of  problem  division  corresponds  statements'  and  of  the in  these  premises particular  general  axioms  prccf  intc  the  group  meaningful  ' i n i t i a l  Intelligence,  cr  lesser a  To  some a  found  automatic  given  axioms  system  and  of  have  encode  facts  the  tool  notation  logicians  auxiliary  axioms  a  (Sandewall, 1970).  l o g i c a l  axioms  'law-like  tc  ephemeral  that  and  problem  express  attempting  bases.  purely  predicate  symbolic  in  added  exemplary  networks  attempting  data  more  In  3  Al  Whitehead.  behaviour  to  consequence in  an  language  e n t i t y - a t t r i b u t e - r e l a t i o n s h i p Simmons,1973).  N.  of  way  tc  conditions'.  knowledge  and  Al,  Intelligent  IMIS.:  method valid may  is  then  inference be  domain  a or  v a l i d  t h i s  prove  and  case  almost notes  a  noticed  the  we  that  model  posited  w i l l  by  (a  either  theorem, when  that  appear  mechanical  be  been  is  developed  by  which  in  in  the  fact  a  deduced  i n t e l l i g e n t . Such  writers.  procedure,  is)  validly  ' l o g i c a l ' .  management  arrive  interpreted  the  tc  59  to  that  conclusion  is  means  has  is  problem  do  not  find  that  a n a l y t i c a l 1965)  recent  is  pure  models,  that  has  proven  MIS prccf  have  permeates  methods  Eut  using so  In  King a  have (1973)  technigue  that  the  system  been  researchers methods  found  research.  logic,  i n s u f f i c i e n t  such  wanting.  B r i e f l y  without as  have  the a  as This  stated,  addition medium  of for  i n t e l l i g e n c e :  purely  deductive  limitations; of  crippled  4  hope  (Robinscn,  4  viewpoint  simulating  2  The  purely  judgment  semantic  •  2  goal  The cry..  proof".  that  resolution  the a  ' r e s o l u t i o n ' ,  perform  The  that  purely  However,  the  to  notice  Accounting  conclusion  *i n t e l l i g e n t *  "that  recent  any  actions  escaped  called can  at  theorem.  statements  and  applied.  solution to  MIS  is  human by  not  logic,  capable  cf  i n t e l l i g e n c e . i l l - s t r u c t u r e d  although serving It or  is  not  useful as  within  the  See J a c k s o n (1974,Chapter 6) for a clear theorem-proving which includes a brief •resolution' as a r u l e of inference.  Intelligence,  p r i n c i p a l  ampliative,  incomplete  its  i t  data-tases  review cf exposition  knowledge  is  and  Al of  A l .  IMIS:.  (indeed, is  an  that  i t  is  to  since  be  meta-knowledge  gained i t  is  are  be  as be  to  obtained  Accounting  merely  resort.  a  last  lacks  for  an  any  cr  abortive  such  cf  fcr  the  order  how  tc  path  mechanisms  reasonable  use  in i f  and  reasoning  means  example,  60  data-bases)  as  attempted  in  large mode  ,  add  Theory..  tortuous  logic  exploring  necessary  to  such  should  while  tc  order  and  by  and  turned  f i r s t  inferences  HIS  crippled  unnatural  humans,  •  I n t e l l i g e n t  a  expressing in  which  information search  relevant  time  for  fcr  tree, proofs  r e a l i s t i c  domains.  Thereupon, incorporate or  interpreted  knowledge  f i r s t - o r d e r  puts  about  the  extent  models  of  systems  representations  and of  that  the  problem-solving  deductive  conjectural p.41)  to  inference  system  strategy move  knowledge.  (i.e. they  semantics) cease  towards As  systems  to  the  Beiter  be  new, (1973,  i t :  "It  quickly  became a p p a r e n t t h a t these prccf i m p r a c t i c a l on a n y interesting One approach towards  procedures alone were mathematical theory.  a l l e v i a t i n g these d i f f i c u l t i e s was to develop completeness preserving refinements of the rules of inference. . . Experimental evidence indicates that this approach alone f a i l s on even mildly serious theorems. V i r t u a l l y everyone is new agreed that knowledge about the problem domain must be used in the logic. The question is how. T h e r e seem t c be two a p p r o a c h e s . . . (1) Semantics as domain dependent heuristics. , . (2) Semantics as t h e representations of models. . ."  Intelligence,  knowledge  and  Al.  IMISj.  Both  Intelligent  these  below.  ISemantic_i  related  senses.  1968)  a  former by  i t s  In  of  the  such  Figure  1.2  a  one  above  propositions  concept Since  is  interpretation  assumptions  meaning aid  to  which  that  a in  Therefore an  can  i f  untrue  the  a  a  axioms. the  proposition general  v a l i d l y  invaluable  is The path is  a  a  form  of  In  the  of  aid  cf  untrue  to  a  in 2  give  are  second^  given  are  It a  that gives  h e u r i s t i c the i t  Reiter  f c r  an  under  in  cf  the  given  mcdel  mcdel  a  fact  must  be  (1972). theory  proofs  in  is  that  theory.  Thus: i f  (The  i f  we  had  controls  converses  an  on  do  analytic the  not  theory  growth  of  of  our  management firm's  and  data-base  hold.)  Intelligence,  x  adjoining  as  l i e s  See  l c g i c a l  by  model  the  If  model'.  search  the  graph.  exists  the  those  including a  See  5  semantic any  turn  twc  Raphael,  which  domain  reasoning  theory.  in  capture  calculus  mcdel  value  Al  uninterpreted.  'semantic  constructed  heuristic  tc  in  reviewed  1968;  concepts,  with  i f  in  concepts  particular  model  d i r e c t  the  the  examples  used  meanings  uninterpreted  in  this  other  61  T h e o r y  (Quillian,  other  contrasted  an  is  example.  interpretation  interpretation  tc  to  the  attempts  their  an  in  sense  takes  for  in  Accounting  seen  memory'  memory  'semantic'  are  f i r s t  relationships  f i r s t ,  and  i n c i d e n t a l l y ,  x  meaning^  their  sense  devices  'semantic  relationships  MIS  knowledge  and  Al.  IMIS.:  Intelligent  guaranteed i f  we  its  the  seeking  we  could  database  as  search-tree. involving  If  reasoning  both  semantic need  gain a  could  some  then  in  a b i l i t y  at  and  contradictions  l o g i c a l .  What  no  extensive  of  (which  contradictions). and  'Part  stored 'Some  near  2  6  27  #370 the  batteries  devices  See  thousand  modes  are  is  true  in  part  may  rechargeable' does  1971,  i f  not  to  ycu  same  do  not  be  a  yet  make  "I've  the  reasoning  shrewd  and  'Some  imply  'Seme  logical  as  logical  the  'Part place  two  makes  involve  near  not  cf  compare  imply  it  hyperboles,  thing is  with  idiomatic  -  human  #38  emulate  knowledge)  meaningful  which  'Part  to  normal  machine  the  paths  (augmented  statements  #38'  i t  any  reguired  In  everyday  7  using  deductive  'steering*  that  of  data-base.  logic  are  theorem  by  our  the  62  then  new  hand  is  inference  But  in  cf  to  times  not  expensive'  alas,  out  l o g i c a l  example,  door'.  Churchman,  Overheard,  of  near  are  prune  which  2  is  For  to  use  s t r i c t l y  comparison!"?  use  non-truths  to  a  e f f i c i e n c y  ngn-logic^  metaphors  there's  given  hundred  that of  procedural  are  a  great  not  we  you  theory  r e j e c t  language  told  under  addition  and  a  Theory..  intelligence  models  is  ac c o u n t i n g  proof  model  general  human  may  We  *  2  the  prepositions  our  and  consistency  were  management  MIS  door' #370  tc  is  lcok.  rechargeable batteries  are  p.31. a  Computing  Center!  Intelligence,  knowledge  and  Al.  I n t e l l i g e n t  expensive*.  i s  to  purely  add  more to  nearness  proximity  system  can  to  be  in  be  an  standards This in  a  former  implies manner  •induction'  the to i f  the  plus  knowledge  this  abcut  obtained  for  that  'while  quantitative  then  abcut  relevant.  nature  system;  machine  data  sentence  of  deductive  be  • consistent  mode are  problems  can  mode  of  well-structured  particular  to  a  measure  when As  i t  cf  would  usual,  the  data.  abnormal  unusually  tc  t r a n s i t i v e  computed* the  Formally w i l l  not  63  Theory..  are.  knowledge  direction  make  drowns  the  ftcccuntipg  solution  sentences  is  and  and  certainly  l o g i c a l  give  i t s e l f  proximity  wise  some  The  example,  be  But  MIS  of  deductive operation,  similar  performance  below  i n t e l l i g e n t to  to  be  human  strengthens  of  a  It system  rather  than  machines  will  error. this  Cur  in  a  machine  employed  circumstances.  'reasonableness* that  reasoning  will whose  cnly  in  be  a  normal  'consistency*. be  prone  tc  err  consideration  of  b e l i e f .  Intelligence,  knowledge  and  Al.  IHISl  Intelligent  l-a.ii2.ii l A c t i o n s ^ In  since to  ordinary  they  alter  and  in  are  the  the  as  in  "He  the  the  did  our  proposed  64  Theory..  means  by  which  world.  deductive  IMIS  iactions^  the  system  Actions  dc  except  as  logic  are may  net  enter  subject  inference:  sold  He  Accounting  real  declarative  matter,  and  2t ruths.*...  robotics  important attempt  as  WIS  not  car  cr  s e l l  he  the  sold  the  truck.  car.  Therefore: He  sold  However,  the  truck,"  consider  " S e l l the car! To s e l l a c a r  the  imperative  v i s i t  a  car  inference:  dealer.  Therefore: V i s i t  a  car  dealer!  1 The inference. In  this  argument But  inference  dealer"  is  l e t  the  speaker  since  the  We  discuss  2  8  2  8  we  an  is  require the  have  i t  to  sentence  instrumental the  premises  section  we  imperative  be  "To  authority are  1.6.2.7, logic  true;  s e l l  tc  the  car  Let  it  command  empirically  and only  logics section to  it  empirically  hypothesis.  non-declarative  1.4.5, use  analytically  true  true  introduce  Intelligence,  a  valid  as  s e l l .  visit be  car Also  the  this  At  a  true.  " S e l l  themselves 6.1.  is  is  car!" tc  say  in  section  the  mcment  actiens.  knowledge  and  Al.  IMIS;.  that In  we  require  short,  system w i l l  Intelligent  act..  require  For  a  A  prove The  off  a  Opa"  the  onto  in  system  EDGE  which  SHELF Aea  6  Pa  this  —>  gap  that  very  goal  the  factory  cf  getting  goal  te  PALLET  A).  data-base  Pa  analytic  robot  (ON  the  in  finds  where  purposes  between  simple let  i t s  management  such  our  notation  next  The  true.  actions.  Aea,  Opa,  i s ,  the  65  empirically  dealer.  taking  the  in  It  be  Tbecrj.  f u l f i l l i n g  pallet.  has  A).  car  suppose  MICRO-PLANNER  already  MICRO-PLANNER (THCONSE (THCOND  by  achieve a  to  bridge  proposition  OF  says  example  shelf  or  to  world  to  a  general  system real  Accounting  also  v i s i t  in  wants  declarative (AT  And  second  storekeeper part  must  the  and  and  conclusion  someone  must  deduction  the  MIS  "tc 2  9  ordinary  MICRO-PLANNER  an  axiom  means  "Push  cr  theorem  A!".  In  say:  ON! (X Y ) (ON $ ? Y $?X) ( (THAND (AT E D G E - O F - S H E L F (PUSH  $?X)  $?X))  (TH S U C C E E D ) ...  ))  The  theorem-prover  true  the  A  would  E f f e c t i v e l y ,  29  For  the  be  (PUSH  rightly on A)  notation  the is  concludes pallet.  that  So  i f  (PUSH  A)  were  the  robot  pushes!  see  section  1.6.2.2  ' t r u e ' .  of  MICRO-PLANNER  below.  Intelligence,  knowledge  and  A l .  IMISj.  Intelligent  Alongside D.  S.  Clarke  may  t h i s a the  which  longer  have  once  "Push"  may  beam  f a l l  impeccable  may  our  no  roof  simply  guarantee  not to  i t  is may  have  truth.  appropriate  to  f i t  ' P r a c t i c a l  well.  Clearly  conclusions:  pushing same  inferences  of  very  our  arm  is  above;  wrong!  may  none: turn  may  just  break  actually for  cr  true  of  example,  Analytic  tc  be  effected  set  of  actions  that  we  know  about  constraints  that  nearer into  we  can  (such to the  a  truth  value  but  is  i f  an  truth  we  is  no  is  attempt  is  "'Push*  won't  the  edge")  then  If  empirical  are  we  may  (almost) We  must  actions  sc  basic  work  inference  an  a  made.  we  we  whereby  that  which  them.  action?  means  actions  as  cf  an  into  ' p r i m i t i v e s ' ;  perform  chain  the  condition  only  to  i t  of  proposition  boundary  be  propositions  class  in  the  truth  empirical  the  is  the  argument  an  r e s t r i c t  heavy  But  the  system  certain  to  faith  facts  management  as  A.  declarative  then  A n a l y t i c a l l y ,  treat  work  a  66  e m p i r i c a l ' t r u t h .  What  The  seems  T h e o r y  inductive  describes mcdel  our  Accounting  and  absolute  on  have  of  and  deductive  (1973,p.2)  Inferences' we  MIS  have  knowledge  i f  something  must  build  which  leads  knowledge  and  such to  "Push".  Intelligence,  Al.  IMIS.:  We which  labour  actions  logic,  Intelligent  the  this  f i t  A r t i f i c i a l  of  to a  actions  1_^4 5  Inductive  A  Although necessary inference may  be  f o r i s  that  a  3°  an a b i l i t y  also  actions  to i t  Pcple,  f o r deductive  machine. logic  net s u f f i c i e n t . doubt  i s  i s  inductive  inductive  logic  (Sclomonoff,1957,1964:  Interestingly.,,  logic  i n  importance  deductive  machine 30  deductive  truth-values  and t h e  by a  Without  a  1973)  i s  cf  with  IMIS.  practice  necessary. by  systems  practiced  67  t h e awkwardness  receive  instrumental  intelligence  f a c i l i t y  inductive  as  x  cf  Theory..  presentations  Intelligence  practiced  Becker,1969;  which  real-world  l c g i c  because  standard  with  procedural  and Accounting  point  into  ease  HIS  i t  necessary  would  tc  seem  practice  logic..  In t h e Proceedings of the third International Jcint Conference on A r t i f i c i a l Intelligence Pople (1973) describes " . . . the mechanization cf abductiye logic". This term of C . S. P i e r c e describes a particular s y l l o g i s t i c form i n inductive inference i n which a 'law-like* statement i s induced. An e x a m p l e m i g h t b e This  plant  has  This  plant  has a  labour  troubles.  l o w wage  Inductive inference: [Perhaps] a l l plants with have l a b o u r troubles. This  reference  generalizations (1974)  to i s  *abductive* also  quoted  rate. l o w wage  logic in  as  a  relation  rates  possible t c  IMIS  source by  cf King  .  Intelligence,  knowledge  and A l  IMISj,  When machine often  i t  we  to  presented  so  as  of  l e t An  us  C  proposition  no-one.  described  as  Inductive  (b)  Any  tool  rule to  propositions conclusions  a in  are  whatever  to  and  be  is  by  which  contrasted  j u s t i f i a b l y  is  the  one  a i s  with  called  an  be  computer  an  unconventional  notation  is  in is  by  the  (in  This  system  which  might  prove)  that:  logic  any  be  deductive may  net  be  cf  an  specification  a  statement  of  and  therefore  the  premises.  Intelligence,  should  universe  fact  employed  additional  using  necessarily  the  'new'  instrument.  material  deductive  i m p l i c i t  cannot  i t s  the  true.  only  as  machine  machine!)  that  deductive  logic  the  buys  programs  (but  modelled  i n  some  i s  becomes  whereas  a  which  cost  suggest  inductive  s p e c i f i c a t i o n  induction  inference,  This  68  further.  we  the  Theory..  of  to  logic.  Observing  using  practice  untrue'  may  anything  case  be  logic  i n f e r e n t i a l of  i t  inductive  as  modelled  cculd  cost:  may  Accounting  inductive  inference  a  surprise  logic  that  explain  'C  the  deductive  at  and  complementary  ampliative  knowledge  (a)  us  inference,  a££2i£Jtion  WIS  consider  seems  deductive  view  Intelligent  and  is  the  regardless  of  3  1  this  knowledge  and A l .  115152 I n t e l l i g e n t HIS and  Accounting  Theory..  69  whether the r u l e i s expressed i n hardware c r software i n the b r a i n , which we r e g a r d as a machine.  (c)  An  inductive  conclusions  inference  whose  worth  some 'degree of s u p p o r t ' TBUE and  FALSE.  3 3  or  3 2  is  recognized  as  yielding  or  c r e d i b i l i t y i s measured by  r a t h e r than by the b i n a r y  this i s  itself  the  outcome  values of  an  analytical rule.  That i s t o say, an i n d u c t i v e system, I ,  is  system which c o n t a i n s c e r t a i n sentences  a  deductive  whose meaning resembles  C  is  reached i n I i t i s a l s o t r u e t h a t C i s a m p l i a t i v e and  the  support f o r C i s 0 =<  "If  a  valid  f (I,C,...) =<  conclusion  1,"  >-  The f a m i l i a r example i n commerce of the l a r g e s c a l e s i m u l a t i o n i s an example i n p o i n t . That which i s presented i n the o u t p u t , however ccmplex, was i m p l i c i t i n the i n p u t . Thus a s i m u l a t i o n , l i k e any program, i s an aS§Ilii£ i.e. deductive system. Tc say t h a t we are s u r p r i s e d by the c o n c l u s i o n s i s not the same t h i n g as s a y i n g t h a t a c o n c l u s i o n i s a m p l i a t i v e . To say the fcrmer i s merely t o acknowledge t h a t the completed i n f e r e n c e outstripped cur personal i n f e r e n t i a l a b i l i t i e s or the r e s o u r c e s we had devoted t o p r e d i c t i n g the cutcome. 3  3 2  One may ask: "where d i d the i n f e r e n t i a l r u l e i t s e l f come from?". With the p r e s e n t knowledge c f the b r a i n this guestion i s s i m p l y unanswerable and s u p p o r t s nc p o i n t c f view.  3 3  Here 'degree of s u p p o r t ' intended to suggest, confirmation'.  i s used l o c s e l y say, Carnap's  and is 'degree  net of  I n t e l l i g e n c e , knowledge and A l .  IMIS.:  Intelligent  As machine  an  example  perform  to  enumeration . count  examples  instances  cf  research  the  machine  to  build  is  not  The  machine  may  "Product (Px1 S  2.  N.  n  be  number  of  the  S  familiar  process the  general  case..  i t  is  model and  new Hx1)  is new S Hx3)  n Nxn  is &  number  recognize  and  and  •i n d u c t i o n  general  find  is  highly  is  we  see  by  in  to  case  the  program if  this  instances).  that,  say:  profitable"  highly  profitable"  highly  is  a  recognise  its  may  is  tc  possible  to and  instructing  (As  the  and  new  of  of  and  is new Hx2)  cf  machine  already  70  Theory,.  Bethod  program  proceed  is S  3 Nx3  "Product (Pxn  rule  then  x1 Nx1  "Product (Px3 S  a  given,  "Product x2 (Px2 S Nx2 S  3.  a  Accounting  consider  a  Al  already  and  f i r s t  of  1.  Let  the  We  1  and  His  profitable"  highly  profitable"  Hxn)  of  instances  counter-instances  observed  observed.  Then  and we  m  the  assert  the  I: "If  n>25  and  m=0  then  record  as  true  prepositions  the  two  sentences: (i) (ii)  3  4  We  have  rather  (x) ( P x & Nx) --> Hx Statement (i) may b e wrong; i t s degree of support is f(25,0,...).  given than  here  the  an  rule  interpretation  cf  the  3  4  "  rule,  I,  i t s e l f .  Intelligence,  knowledge  and  Al.  I n t e l l i g e n t  The served a  is  we  both  by  and  by  the  sense have  sweeping  the  recording  how  awkward This  predicate  calculus  deductive  inference  to  n  but  conclusion another  have  in  fact of and  for  the  but not  be just the  using  a  the  is  some  statement 3  tc  6  not  cf  of  the  the  be  tc  a  rule  the the  cf  requires  separate 1  easy  of  a  procedural  5  a  instances  tc  how  laboriously,  3  f i r s t - o r d e r  also  i l l u s t r a t e d ,  argument  predicate  It  programming  wrong.  conclusion  in  truth  a  in  conclusion.  state  series  I  premise  the  in  advantages  a  cf  te  the  about  Notice state  as  purpose  rule  may cf  7 1  the  cur  higher-order.  from  form  in  statement  would  merely the  guality  statement  i t  j u s t i f y  conclusion  induction  rule  Theory.  deductive,  cannot  second  of  is  analytic  f i r s t  from  includes  we of  the  because  also  would  discussion recording  is  statement.  argument  (i)  the  the  Accounting  argument  rule in  cf  and  Because  sentence  calculus.  1  the  preserved  f a l l i b i l i t y  Notice  of  inductive^  logical  But  form  MIS  n  and  i t s  value  of  the  whole  language. our  We  e a r l i e r  language  for  metaknowledge.  3  5  A n d we d o n o t c o m m i t p e t i t i o use the inductive rule conclusion.  3  6  A second and seen in the least-squares  p r i n c i p i i since we i t s e l f tc arrive  dc at  not cur  familiar example of mechanical induction is operation of a program for s t a t i s t i c a l regression. The r e g r e s s i o n line f i t t e d tc  the data is a generalization and may be regarded as an inductive inference. The program will normally compute f (I,C,...) whose v a l u e i s the f a m i l i a r set cf measures cf f i t and r e l i a b i l i t y .  Intelligence,  knowledge  and  A l .  IMIS,: Intellicjent  It should minimise nor  the  the  be  psychological and  inference  observed that a  STRUCTURES x  particular  dc  pragmatic value between  faculty not  knowledge about knowledge.  achieved,  f o r implemented the  fact  that  One  aspect  knowledge  may  be  clustering  in the quotation  have only  clustered  for  frames. i s facing  the  structures which impound more resembles objective  of knowledge tc which the  is  a  KNCWLEEGij^ d e s c r i p t i o n s S  Whether the r e s u l t  structures are knowledge  We  fcr  highway.  a network or a procedure i s unimportant i f the be  tc  of induction  precondition  cf Al research  building  mean  support  micrc-worlds and  •generation' of  a  a parallel  ORGANIZING x  is  the  72  net  that for an i n d u c t i v e .  actors  challenge  we  and  FOR  ££2£^^ME§Sx demons  that  logic  deductive  present  Accounting Theory..  in  implementing induction and  The  and  appreciated  distinction  deductive  MIS  can  designs  paying p a r t i c u l a r a t t e n t i o n i s not  homogeneous.  in many ways.  from Machlup.  The  We  Units saw  MIT  ene  cf such  Laboratory  o f f e r s another: "Our knowledge about the r e a l wcrld, cr atcut those subjects in which we are particularly competent, i s not a bland, uniform structure cf simply-interconnected "facts" ... Our knowledge i s made up not of s i m i l a r knowledge about many d i f f e r e n t things, but of elaborate constructions about a few things." (Minsky et al,,1973, p.101)  I n t e l l i g e n c e , knowledge and  Al.  IMIS;.  For  example,  one  may  Intelligent  for  speak  genera 1  "simple"  Many  represented  "  one  knowledge:,  because in  any  cannot  McCarthy  domain  factories insolvent  the  knowledge^ with firms  things  push  back  basic  in  Accounting  and  for  a  73  Theory,.  particular  raw with  "the  stuff  any  simplifying experiments  obvious  programs.  of  John  logic,  arise  Recognition to  system  physics,  problems  intelligence  but  given  and  domain  o f : -  cgmmgnsense about  any  MIS  with  need  for  in  1958.  "What  Cne  can  any pull  one.)"  and  so  cn  a  are  p.77).  knowledge  monkeys  with  shadows  payable,  etc.  with (See  not  string,  (I^id.,  do  cn.  machine  person with  knows  common-sense  materials, accounts  to  child  gees  bananas, blocks, Sussman  (1973))  Intelligence,  knowledge  and  Al.  IMIS.:  -  Intelligent  problem-solving knowledge, modify also the  how  to  and  Accounting  seta-knowledge^  structure  i t ;  when  MIS  i t ,  "Hew  simplify  tc  i t ,  3  74  The cry..  eveke  relevant  doubt  7  i t ,  to  make  an  inductive  generalisation  suggest  that  i t  might  unwise.  d i f f i c u l t  be  3  8  and  This  is  task!  "At f i r s t , many w o r k e r s h o p e d i t would he possible to separate the problems of reasoning and deductions into: • A basis cf factual knowledge. •  A  (possibly  complicated)  procedure  for  using  the  knowledge. But the s e p a r a t i o n leads to serious problems. Too much o f o n e ' s knowledge i s concerned p r e c i s e l y with which other k n o w l e d g e s h o u l d be a p p l i e d , " (Sussman, 1973)  3  7  The o r i g i n a l PLANNER thesis provided fcr theorems, ANTECEDENT t h e o r e m s and SIMPLIFYING  (Hewitt,1972) s i m p l i f i c a t i o n be s u b s t i t u t e d 3  8  See  an example of a simple but essential w o u l d be t h e k n o w l e d g e that " T c p r c v e A" may for "To prove A V A V A".  McDermott's  Minsky  et  "How piece  a l . , tc  doubt  of to  current (1973,  work  f c r  remove  on  belief  systems,  gucted  in  p.71)  things.  knowledge  responsible needed  CONSEQUENT theorems.  to  saving i t  later  Each  routine  the  world  with  i t  should  Intelligence,  adds  mcdel  the i t  that  a is  information  c o n f l i c t . "  knowledge  and  A l .  IMIS:.  I n t e l l i g e n t  Consider door's  the  f a i l u r e  to  Wednesdays". which  situation  arises..  twc  we  Since scan  a l l  i t  i t s  currently  these  "facts  in  our  about  are  commonsense,  to  apply  wish  ' so  let  us  look  at  75  Theory,.  items  not  is  net  knowledge  applicable of  forth  appropriate  the  the  theorem-proving drown  in  pool  'active  of  relevant  in  l i s t :  "the  Tuesdays  and  'background'  u n t i l  a  relevant  . . .  only  the  solution.  Thus  Charniak,  Sussman  guickly  relevant  axioms'. datum  and  is  structured  units  triggered  by  of  hand  The name demon,' and  a  (its  include  now  run  many  be  tossed  the  of  speaks  knowledge own  otherwise  in  control  the of  lie  cf  it  tc is the  cf  f a i l i n g  tc  have  ensured  a  Following  -  school  procedurally  dormant  the  is deliberately reminiscent cf Selfridge's 'Eandemonium* model.  Intelligence,  c a l l  intc  preconditions.  structure  the  tend  MICRO-PLANNER 'demons'  is,  axioms  device.  which  special  risk  is  axiomatic  devices)  tec  to  item  should  since  robct  given  That  pattern)  Horecver,  i f  constantly  desirable.  heuristic  their the  is  would  others  Intelligence of  We  entity  particular  knowledge  which  i t  an  any  world  (which  A r t i f i c i a l  sleight  reverse the  for  i f  knowledge.  very  that  see  the  systems  data  feasible  to  event  desirable  9  remaining  and  do  structure  3  Accounting  DEMONS:.  IM.  make  and  items  close"  Both  knowledge  HIS  knowledge  u n t i l Through  system  a  'Maxwell's  and  Al.  IMIS2  demon  3  does  9  called  Intelligent  MIS  net  be  have  tc  automatically  particular  for  a  at  f i r s t  do  not  "The  sight, know  i l l  >NOT  door's  of  any  i t  is  to  the  such  close"  on  problem  a  always  occurs  seme  a  is  by  The  demcn.  is  triggered  work  there  is  typical more  a  task  complex  non-event!  demons  We  called  certainly  f i t s  by  into  the  class.  ACTORSj. In  the  (1971,1972,1973) evolved  into  progression  demcn  a  proposal  fcr  universal  knowledge.  the  notion  of  event)  nothing  to  i n i t i a t i o n  •world'.  In  a  This  ultimate  (Hewitt,1973)  of  of  a the  The  ultimate,  procedure less  procedure  ACTOR sole  is  by  than  entire  a  has  exemplifying  'pattern-directed  invocation  pattern  knowledge,  'ACTOR'.  of  writings  knowledge  embedding' called  Hewitt's  procedural  for  of  of  the  representation  the  just  it  76  conditions.  NOW-TDESDAY  Al  But  when  and  is  f a i l u r e  because  non-happenings. 'demon'  THEOREM  Theory..  Instead  only  events  ANTECEDEKT  demon.  and  of  NOW-W EDN E S D A Y  Accounting  ••called".  when  pattern  MICRO-PLANNER  and  with  current  grandly building  of  the  its  block  cf by  cf  is  a  This  data  Hewitt's knowledge  the  is,  pattern  relevance  state  conceived  know-all  (that  'world-invocation '.  because  cf  generalises  invocation'  matching  form  'procedural  modular he  gradually  the  of  an  is  the  tc  the  base  -  a  group  as  and  of  computation:  Intelligence,  knowledge  and  Al.  I M I S I n t e l l i g e n t  MIS  and  Accounting  Theory..  77  " I n t u i t i v e l y , an ACTOR is an active agent which p l a y s a r o l e on cue according tc a script. Data structures, functions, semaphores, monitors, ports, descriptions, Q u i l l i a n nets, logical formulae, numbers, i d e n t i f i e r s , demons, processes, contexts, and data-bases c a n a l l be shown t c be s p e c i a l cases of actors."  Actors  respond  has  INTENTION  an  context  of  t h i s  theorem  been  met..  see  is The  and  from  his  how  the  solves  being  sent  the  message  same  thing  theorem,  is  system.  invention more  prerequisites  applicable  not  basic  interactions  Each  the  ACTOR  is  messages.  that  the  We of  and  which that  the  to  the  dc  not  the  f i r s t  i t s  understand  lines  check  solid from  and  of that have  heterarchical  modular  learning  the  s a t i s f i e d .  preconditions  Hewitt's his  Actor  and  few  invariably  universal, of  are  ultimately  separate  problems of  in  Unfortunately  easy  conjectures.  the  manifold  i t  appropriate  checks  THCONSE  knowledge-based obscure  own  actor  we  MICRO-PLANNER  their which  the  Functionally, a  to  writings  are  attainments his  writings  ACTOR  per  describing  se the  knowledge.  Intelligence,  knowledge  and  A l .  IMIS:.  Intelligent  HIS  and accounting  Theory..  78  WICRO-WORLDS^  1£1  The three  label  senses.  §  It  H22l  as  meant  and so  what  i n  from  serious every  i n  of  i t s  Al  in  a  at  least  such  as t h e  says:  SHRELU,  "Chess  i s  'micro-world •  in  adult  s i x  properties,  Wincgrad's  Greenblatt  distinguished  do  used  one o f : -  world'  on.  with  i s  simplified  'blocks  competing to  has  J2Sai£  ubiquitous HACKER  "Micro-world"  persons  year  o l d  rather can,  Sussman's a  'world'  that  we a r e  than as  trying  in  natural  language."  -  a  small  domain  in  which  whose  inherently  Our  KIWI  r e a l i s t i c  small  domain  size  detail  avoids  explored  i s  modelled  the problems  i n  Chapter  4  i s  fork  the  of  but  scale. such  a  micro-world.  -  a ^contingents of  some  made.  universe  higher In  LISP  (GETWORLD...) , generate retain  and  proposal global  universe  i n  at a  a  i n  particular  and  branch  hypothetical  In  into  back the  to  o f f  (REALWORLD...)  PLANNER  p a r a l l e l  and f o r allowing  and there  micrc-wcrlds  communication  Intelligence,  was  functions which  universes  alternate  o r i g i n a l  world-line  decision  the extraordinary  (NEWWORLD...)  f o r s p l i t t i n g  world  when  we s e e t h i s  ' t i c k e t s '  r e a l i t i e s .  spawned  and  ultimate was  the  under  between  knowledge  can  a  them.  and A l .  IHISj.  Each  Intelligerit  micro-world  information for  and  i f  such  to  be  states  most seems  is  tc  Accounting  wholly in  harbour  in  knowledge separate  and  exist  to  Minsky  of  is  ty  quite  possible  which  contradictions; then  we  the  taken indeed  would  wish  them  for  organizing  micro-worlds.  cf  that  It  propositions  comprehensive be  79  Theory..  characterised  i t .  inconsistencies of  harboured  The knowledge  imply  and  knowledge  micro-worlds  globally  MIS  these  models  who  .  .  .  "has formulated a new t h e o r y , c a l l e d Frame-Systems, which, i t i s h o p e d , w i l l s h o w q u i c k l y how t c develop common-sense reasoning in systems that are net c o n f u s e d by c o n t a i n i n g large amounts of non-relevant information." (Minsky et al.,1973) Frame-systems s h a l l  let  a  draft  paper  is  few  too  brief  (1974)  for  big  a  topic  quotations the  tc  discuss  present  the  f u l l y theme  here. -  see  We the  d e t a i l s : -  "Here is the essence of our theory. When cne e n c o u n t e r s a new s i t u a t i o n (cr makes a substantial change in his view of the present problem) cne selects f r o m memory a s u b s t a n t i a l structure called a •Frame . *° This is an elaborate stereotype 'scenario - a remembered framework t c be a d a p t e d tc f i t r e a l i t y by c h a n g i n g details as necessary... 1  cr  1  *°  Recall  again  structure  of  the  'Paradigm'  s c i e n t i f i c  concepts  reygluticns^i  Intelligence,  cf  Kuhn's  The  (1970)  knowledge  and  Al.  IBIS.:  Intelligent  Attached of  tc  this  information.  what  is  these  to  MIS  happen  Accounting  structure  Seme  of  next.  expectations  and  are  this  Some not  are  several  information  is  about  80  Theory.  what  kinds  is  about  tc  do  i f  confirmed...  The top levels of a frame are fixed, and represent things that are always true about the supposed situation. The lower levels have many terminals - •slots' t h a t c a n be f i l l e d w i t h specific instances or data. Each t e r m i n a l has markers that specify conditions the assigned instance must meet... Collections together This  into w i l l  Frame-Systems  of  related  s u f f i c e . are  It  the  descriptive/procedural  ideas  and  an  a  synthesis.  general  solutions  to  a l l  under  this  label  model  for  the  mentioned and  It  presents  such  before.  r a t i o n a l i t y  to  is  and  this as  It  in  'actors'  problems  data-base  i n t e l l i g e n c e . performances  the  he  be  networks,  'demons'  at  are  should  invocations, attempt  frames  linked  Frame-systems."  a  learning recognizes  which  any  ' p a t t e r n - d i r e c t e d ' a l l  the  most  gathered  which and the  together  dees  'Frames'  not  processes embraces  of  the  Intelligence,  is  tut  human  range  generalizing  intelligence  claim  conjectural  the  limitations  in  approach  appealing  r e t r i e v a l  model  that  'micro-wcrlds',  Minsky of  clear  cn  of  which  we  storage  subject.  knowledge  and  Al.  IfiJS:.  1  A  5  EXAMPLES For  and  by  OF  Al  reader  texts  by  and  For  a  variety  of  topics  of  International  Intelligence We  detailed  structuring  have  of  of  perception  contention worth  serious  in  A r t i f i c i a l  Intelligence  is  1969, to  and  Nilsson of  81  examine  1971  of  work the  (1968),  and  Jackson  on  a  on  A r t i f i c i a l  1973.  present  only  seven  some  way  demonstrates tc  commercially  a  one  degree  useful  cr  which  wide  proceedings  and  in  the  theses  Minsky  (1971)  current  should  concerned and  c o l l e c t i o n s  Conferences  understanding a  the  (1963)  (1971),  Joint  space  notable with  more  the  aspects  supports  application  of  cur  Al  is  consideration.  Sy.ans.1 A N A L O G Y  program:  the  of  exercising  Theory..  consult  reader  knowledge  that  Accounting  Feldman  for  Each  or  and  appraisal  the  (IJCAI)  accomplishments.  date  S^agle  (1974).  the  to  should  Feigenbaum  the  MIS  RESEARCH  attainments  interested edited  Intelligent  pure  intelligence..  Intelligence,  knowledge  and  A l .  IMIS  X  Intelligent  Consider to  resolve  the  the  task  question:  MIS  set "A  and  in is  Accounting  Figure to  B  as  Theory..  1.3.  The  C  t c . . . ? "  is  Intelligence,  82  problem  knowledge  and  is  Al.  B  1  FIGURE 1.3:  A T A S K FOR EVAN'S A N A L O G Y PROGRAM.  IMISj.  Is  Intelligent  this  Before problem  in  a  test  he  Figure  processes.  State  examination  intelligence  solve  almost  f a i l i n g  only  disputable  a  s p e c i f i c  implemented construct Next  the  of  the  the to  this  sought  to  1  advices  progressively  intelligence? tc  from part  (1964) type  solve  a of  was  procedures  in  the  his  New a  own York  battery able  presented  were  of  problem  tc  tc  i t ,  d i f f i c u l t  cr  was  applied made  to  5.  on  the  various  generalizing  picture  is  The  to  program  rule  the  if  they  sought  Figure a  tc 1.3.  rule  of  A  a  second-crder  into  the  description result  contained cn  the  elements no  were  cf  the  process,  descriptive the  within  thus  match  matching  methods  (induce)  then  tc  program  program  description  rule  was  the  the  Evans'  construct the  This  rule  attempt  of  as  84  observe  solutions  each  of  time  drawn  this  of  The  importance  whose  of  description  B.  h e u r i s t i c  tests  b r i e f ,  of  descriptions  invited  one  In  description  an  is  program  class  mapping  and  called  administered  detailed  correspondence  C  is  be  same  general.  program  description.  the  Theory.,  human.  were a  at  Evans'  problems  Although highly  was  to  reader  problem  tests.  on  the  and  and  a l l  for  1.3 This  Accounting  something  continues  mental  of  of  and  MIS  immediate  with  cf the  various relative and match  on was  found.  Intelligence,  knowledge  and  Al.  IMIS.:  Intelligent  Evans* reasons. capable  early  Most of  i t  exercising when  demonstrates  an  the  is  an  to  of  potentially  a b i l i t y  some  induce  general  Understanding  is  of  to  a  and  test  rules  twc  machine human  by  major tc  is  be  level  of it  ccnstructing  simplifying  which 4  85  Secondly  generalise  a p p l i c a b i l i t y .  for  small.  phenomena  and  the  at  is  performance plans  The cry..  interesting  demonstrates  domain  descriptions u n t i l  Accounting  intelligence  a b i l i t y  descriptions  ana  example  strikingly  performance,  higher-order  MIS  those  satisfied. is  modest  This but  1  c o n c e p t S E M A N T I C  MEMGBY.  The author, in a simple unpublished exercise, applied methods r e s e m b l i n g Evans* to problems cf the fcrm EDNA IS TO J I M AS HAROLD IS TO WHAT? o r WATER I S TO ICE AS M I L K I S TO WHAT? cr EICTURE-1 IS TC FICTURI-2 AS P I C T U R E - 3 IS TO W H A T ? In each case a l l concepts were defined a n d i n t e r r e l a t e d by a network o f d e s c r i p t i v e p r o p e r t y and set relationships. The machine s o l v e d simple problems with ease, r e t u r n i n g an o r d e r e d l i s t cf probable answers. When t h e data base for a problem grew beyond about 100 relationships i t f a i l e d to reach even a f i r s t solution in an economically feasible time. 4  1  Intelligence,  knowledge  and  Al.  IMISj.  Ross which  an  Intelligent  Quillian  for  relationships  At  terms some  label  'planes'  " 1 .  That  notation  appreciated planes  in  the  his  set  a  memory  are  in  any a  organized  in  Q u i l l i a n ' s each  is  concept  the  l i v i n g  plane  being  -  has  1.4  dotted where  is  arrows the  to  its  defined system, may  intersect therefore  of  has  semantic  become  the  way. shows  cne  cf  comprised food  take  other  clear are  may  which  defining  of  the  was  this  thesis,  growth. meals e s p e c i a l l y  terms  network  in  as:  tc  than  keep  drink."  i t  should  paths  to  the  Intelligence,  terms  the the  but  referenced  by  concept  concepts  concept  a  process a  in  two  ('memory')  86  turn  concepts  bases  Figure the  in  given  memory*  This  which  is  'semantic  for  and f o r forming  of  of  as  the  meaning  of  of  Theory^  model  term  -  one  to  These  manner  from  that  use  data-base  1.4,  in  the  d e f i n i t i o n  data  memory.  l i v i n g Things  of  novel  and  for  Figure  semantic  a  the  definitions  the  in  of  concepts.  Q u i l l i a n s ' s  relationships  The  the  or  organized  i t 2.  other  Accounting  sought  i n t e l l i g e n t  elements  remove  i t s e l f .  many  to  of  intersect  usual  i t s  and  (1968)  'understanding'  s u f f i c i e n t  in  MIS  are  knowledge  be  ether  defined.  and  Al.  FOOD:  1. That which living being has to take in to keep It living and for growth. Things forming meals, especially other than drink  (FOOD) OR  ,,,  _  FIGURE 1.4:  $  7  /  :  ^  :  THE PLANE (NETWORK) DEFINING 'FOOD' IN QUILLIAN'S SEMANTIC MEMORY.  IMISIntelligent  The semantic the  relationships  system  sentences for  performance  i t s  was  also  embodying  pioneering  operate  at  the  general  methods  MIS  and  asked  of  between capable  i t s  cf  findings  and  Figure  1.5  shows  during  the  system's  two  of an  a  this  two  demonstration level  Accounting  given  it  was that  easily  expanding  was  as  primarily an  Al  d e f i n i t i o n  search  the  trace  semantic  through  Intelligence,  its  the  Although  output  extensible of  tc  88  concepts.  generating  semantic  examples  system  Theory,.  cf  English interest  machine  can  with  simple,  data  base.  paths  found  memory.  knowledge  and  Al.  IMISj,  Intelligent  MIS  Example 1. C o m p a r e C R Y , A. I n t e r s e c t : SAD (1) CRY2 IS AMONG O T H E R (2)  TO  C0MF0RT3  CAN  E x a m p l e 2. C o m p a r e A. 1st i n t e r s e c t .  BE  and  Accounting  Tbeer  COMFORT THINGS TO  MAKE  PLANT, LIVE  LIVE  TO  MAKE  A  SOMETHING  (1) PLANT IS A LIVE STRUCTURE. B. 2 n d i n t e r s e c t : LIVE (1) PLANT IS S T R U C T U R E WHICH GET3 FOOD T H I S FOOD I S T H I N G WHICH B E I N G 2 HAS TO T A K E I N T O I T S E L F T07 KEEP LIVE.  SAD  SOUND.  LESS2  FROM  SAD.  AIR.  • • • Example 7. Compare F I R E , BURN A. 1st I n t e r s e c t : BURN (1) F I R E IS C O N D I T I O N WHICH BURN. E. 2 n d i n t e r s e c t : ARE (1) TO BURN2 CAN BE TO D E S T R C Y 2 C . 3 r d i n t e r s e c t : BURN (1) F I R E IS A FLAME CONDITION. A GAS T O N G U E 4 . T H I S GAS I S GAS  Figure  1.5:  SOMETHING  EYU  T H I S F L A M E CAN WHICH B U R N .  Intersecting pathways traced 'Semantic Memory'.  Intelligence,  FIEI. EE  through  the  knowledge  and  Al.  90;  FIGURE 1.6:  EXAMPLE OF A SCENE ANALYSED BY GUZMAN'S PROGRAM.  IMIS:.  SEE -  a  is  one-eyed  might  Intelligerjt  able  view  of  determines, probably  the  a  solid  on  information.  governing  the  strong  a  view  program  of  is  a  of  The  pointing  a  X,  object.  v i s i b l e  away  frcm  c h i e f l y  T,  etc.) cn  relies  As cf  i t the  say.  base  in  cf  viewer. that  it  abcut  the  of  local  the  local  vertices  intc  certain  rules  and  the  to  the  cn  cf  are  3  although  needless  the  information)  description  and  processing  regions  blocks.  same  a  c o r r e c t l y  conjectures  its  into  21  interesting  system  (semantic  r e c t i l i n e a r  building  be  r e l i e s  arrow,  evidence.  the  proven,  could  It  that  c l a s s i f i c a t i o n  consolidation  solid  be  of  program by  of  i n t e l l i g e n t , basis  (fork,  presumption  part  cannot seen  regions  9,10,11,12  and  drawing  factcry-hand  pclyhedra. regions  91  line  rcbct  rectangular  p a r t i c u l a r l y  the  a  object,  'pyramid'  provided  types  that  region  The  as  the  net  Theory..  two-dimensional  solid  is  plausible,  scene  p r i o r i  22,23,24  conclusions  establishes  several  cf  Accounting  such  same  is  a  analyse  the  program  information  scene  example,  irregular  Guzman's  a  a  and  every  and  analyse  number  of  worst,  global  a  with  Such At  -  for  part  contiguous  and  of  encounter  description  to  MIS  basis cn  that  the the  proceeds world  cf  it  weak global  scene  is  Guzman's thinks  it  sees.  Intelligence,  knowledge  and  Al.  IMIS,:  Intelligent  Guzman's  work  directions  by  emphasises  the  tentative  reading  analysis)  may  o b j e c t ' . the  a  produced  shadow  data  data  base.  deductive  But  approach  did  the  analysis  the  extra  actual  added  many  we  so  of  scene  make  have to  to  that  in  had  some been  the  visicn  as  a 4  It  2  and  eliminated  is  fu22y, The in  semantic  Although  the  competing shadows  well  'ncnsense  effect,  tc  cf  tc  the  tendency cf  and  as  true the  fcr  axioms  algorithmic  early  acted  before  edges  illuminated  stages  semantic  one  false  addition  number  the  the  or  a  when  but  semantic  i n d e f i n i t e  of  A  syntactic  meaningless  the  the a l l  analysis.  elsewhere  many  further  unacceptable  scrt  Waltz's  from cases  visicn  hardware.  data.  yield  interesting  Huffman  'propositions',  noted  in  (an  cracks  collapse  drown  began  sense  with  twc  acceptable  rejected.  i n t e r e s t i n g l y . not  in  in  92  Theory.  Waltz.  (an  just  be  mere  information  f i l t e r the  by  scene  shadows,  dealing  systems  r i s e s .  may  and  semantics  rejected  added  thus  a  not  sentence  Waltz  scene,  image  be  of  Accounting  extended  (1971)  cf  does  and  been  importance  which  of  has  Huffman  analysis)  reading  MIS  of  readings a  powerful  perception analysis  was  complete.  *  Both s e m a n t i c and p r a g m a t i c knowledge may s c r e e n o u t some of the excess syntactic readings, (Consider Noam Chomsky's famous "Colourless green ideas sleep furiously", which is s y n t a c t i c a l l y faultless but cenveys almost no meaning, or Roger Schank's "I saw t h e G r a n d C a n y o n f l y i n g tc Chicago", which has two e q u a l l y valid syntactic readings but o n l y one s e m a n t i c reading 2  Intelligence,  knowledge  and  A l .  IMIS.:  Intelligent  Solving  mathematical  Mathematical general faced  symbolic Evans  challenge  was  acquire  i n t e l l i g e n c e  d i f f i c u l t  for  solution. However, found  humans  It for  was  thereby  ceased,  a  most  domain,  post,  tc  an  any have  1971)  success.  area  for Like  thought  to  symbolic  integration  is  to  no  have test  unlike  Evans',  fcr  solutions. an  fcr  which  (Slagle,  Turing  be  a b i l i t y  projects  in  that  a  English..  interesting  appeared  thus  in  system  working  in  Al  93  Theory.  necessary  SAINT  algorithmic ex  stated  several  a  was  and  S l a g l e ' s  e f f i c i e n t  is  Slagle's  Slagle's  Accounting  problems  Among  integration  program  1  and  reasoning  i n t e l l i g e n c e .  this  HIS  algorithmic i n t e l l i g e n c e .  later Eas  i l l u s t r a t i o n  work  has  Slagle's  work  of  i n t e l l i g e n t  behaviour? Bobrow's program,  CAEPS(1969),  English.  In  interest  to  competence the  analyse  and  the  S T U D E N T (19 6 4 ) , accepted  retrospect  was  input  consult  program,  neither the  output  originals  of  Figure  because  computation  as  printed  employed,  including  idiomatic  substitutions  employed  both.  out  in  1.7  cf  their  quotes  the  intermediate  the  theses  reveal  the  Intelligence,  to  lasting  example  reader stages the  ask  of  should cf  the  processes  a b i l i t i e s and  in  mathematical  cne  interested  assumptions  problems  methods  but  The  exercising and  Charniak's  mathematical  English  impressive.  and  tc for  knowledge  make hints.  and  A l .  IMIS.:  A  problem  (THE  Intelligent  for  PROBLEM  MIS  Bobrcwj^s  TO  BE  and  Accounting  Thecry..  94  STUDENT.:  SOLVED  IS)  ( T H E R U S S I A N ARMY. H A S 6 T I M E S A S MANY R E S E R V E S IN A UNIT AS IT HAS UNIFORMED SOLDIERS. THE PAY FOR RESERVES E A C H MONTH I S 50 DOLLARS T I M E S TRE NUEEER OF RESERVES IN T H E U N I T , AND T H E AMOUNT SPENT ON THE REGULAR ARMY EACH MONTH IS 150 DOLLARS T I M E S THE NUMBER OF U N I F O R M E D SOLDIERS. T H E SUM O F T H I S L A T T E R AMOUNT AND T H E P A Y F O R RESERVES EACH MONTE EQUALS $45000. FIND T H E NUMBER O F R E S E R V E S IN A UNIT THE R U S S I A N A R M Y H A S AND T H E N U M B E R O F U N I F O R M E D SCLDIERS IT HAS.)  ( T H E NUMBER HAS I S 600) (THE  1  NUMBER  ElSilem  OF  RESERVES  OF  UNIFORMED  for  IN  Charniakj^s  (WATER IS FLOWING INTO OF 1 5 . 0 C U B I C I N C H E S PER THE BASE OF THE F I L T E R IS IS 10.0 INCHES, FIND L E V E L IS R I S I N G WHEN T H E  (THE  ANSWER  (TIMES  A  UNIT  SOLDIERS  THE  IT  RUSSIAN  HAS  IS  ARMY  100)  CARPS:. A CONICAL F I L T E R AT THE RATE SECOND. IF THE RADIUS CE 5.0 I N C H E S AND T H E A L T I T U D E THE R A T E AT W H I C E T H E W A T E R VOLUME IS 100 C U B I C INCHES.)  IS)  .531 32943  IN  (EXPT  SEC  -  1.0)  (EXPT  P1  -  0.3333332))  Figure  1.7:  Algebra  and  calculus CARPS.  problems  Intelligence,  fcr  STUDENT  knowledge  and  S  Al.  IMIS  I n t e l l i g e n t  X  MIS  Learning The learning f a i l u r e of  e a r l i e s t  in  the  was  thousands  •learning' used  to  phase Alpha  of  of  a  consistently 'learned' new  the  was  Herein,  supplemented fcr  value  of  a  a  played  randomly  created  game  player  function  'Beta' and  to  machine  success  a  process  i t s e l f .  In  If  function  was  to  cne  'player'  evaluation Beta  of  function  move.  tried  or  rote-learning  polynomial  then  Alpha  cf  by  given  machine  outplayed  95  given  the  outplay  that  was  a  most  ' l e a r n '  to  write  function.  programs  by  the  patterns  work.  learning  f a i l u r e .  and  The  chance  • i n t e l l i g e n t '  This  device  which  instructions  pure  response  player.  strategic  evaluation  i n s t r u c t i v e  not  in  the  Friedberg's  b i t  i l l u s t r a t i o n  adaptation  checker  Theory..  experience^  c o - e f f i c i e n t s  learning  using  cf  'book-moves*  optimal  assess  from  accounting  impressive  sense  Samuel's  and  were  program  tested  steering  sought  and  tc  random as  whether  performed  predict  (1958,1959)  generating  discovering  program  would  of  machine  potential the  results  s i g n i f i c a n t l y  demonstrated  knowledge  modifications  somewhere  Intelligence,  the in  program did  or  worse necessity  the  to  did than fcr  system.  knowledge  and  Al.  IMIS.:  I n t e l l i g e n t  Winston^s But a *  3  and  Patrick  world  can  induction.  of  of the  were  exchanged  the  program  'a-kind-of, and a  PPly  the  a  above, and ether  parameter  'bricks'  a  scene  learnt for  structural  programs  build  adjustment  class  as  recognize  say,  The  network  supports, relations  a  a l l  major  as  concepts  primitive than a  new  bricks  in  in  as  the  an  scene  built  by  relations:  right-of,  'must'  cf  l i t e r a l  scene  primitive  a  the  form  a  descriptions  i n - f r o n t - c f , such  structure.  Moreover,  rather  i f ,  their  learn  practice  could  96  presupposes  supporting  'wedges'.  structural and  Theory..  descrj.rtic.ns..  (1970)  and  and  wedges.  employ  marries* to  of  Accounting  domain-specific  generalize  They  description example  probably  and  of  through  Winston's  blocks-type programs  learning  learning  complex  MIS  l e f t - c f  'roust-not*  may  relations.  This is the reason for cur disinterest in 'learning' m o d e l l e d as B a y e s i a n r e v i s i o n of p r o b a b i l i t i e s . A system may i n d e e d a d a p t to a changing environment in this manner but t h e mechanism i s i n s u f f i c i e n t l y primitive for the l e v e l cf Al p r o b l e m we a r e considering.  •3  Intelligence,  knowledge  and  A l .  IMISj.  Intelligent  Winston's by  exposure  the  concept  Figure  1.8  exposure  to  descriptive p a r a l l e l human  to and  a  which  examples a  a  its  cf  of  the  descriptions  examples:  instructive cf  enriches  programs  Winston's  system  well-constructed  examples  cf  'near-misses'.  ABCEES and  97  Theory..  structural  sequence  which  teaching  Accounting  sequence  progressively  network  via  and  learns  training  i l l u s t r a t e s  between  being  program  MIS  5  NCN-ABCHFS,  generalizes  are  the  building.  The  and  teaching  a  training  sequence  is  appealing.  Intelligence,  knowledge  and  A l .  98  FIGURE 1.8:  TRAINING SEQUENCE FOR WINSTON'S CONCEPT LEARNING MACHINE.  IMISj.  Intelligent  MIS  Perception  yia  g u a r v. a n d Man concepts speech  is  a  in  the  may  be  regarded  intelligence.  Can  attainments  so  impressive  and  that  the  w i l l  be  basic based  far  a in  as  machine  todays  An tc  do  **  is  transmit  as  *  Not  'written 5  There  language  current work understanding et a l . , (1973).  tc  is  and  perceive  concepts  sign  well?  established  natural  a b i l i t y  principal  understanding  methodology on  and  a  99  systems..  animal.  others  Theory..  languagej  natural  encouraging.  For a review cf principles to the language see Newell *  of  Accounting  command  communicating speech  and  yet,  via  cf  his  but  speech'  the are  l i t t l e  doubt  future  work  systems.  in of  applying spoken  similar natural  Although most Al conversational systems have not c o n c e n t r a t e d on g e n e r a t i n g idiomatic English seme useful results have been obtained on this closely related problem. Woods (1968) stated that his ATN grammar for recognizing English syntax was e a s i l y r e v e r s i b l e a n d c c u l d be used as a g e n e r a t o r as w e l l as a r e c o g n i z e r , but we h a v e not seen this done i n such a d i r e c t manner. Simmons and Slccum (1972) discuss "Generating English disccurse frcm semantic networks". T h e common s o l u t i o n , when NI generation is not the research objective, is t o use s e m a n t i c knowledge tc ' f i l l i n the blanks' in s t a n d a r d message forms. We u s e t h i s method i n Chapter 4 but i t is inadequate for a useful IMIS. 5  Intelligence,  knowledge  and  A l .  IMIS:.  To  understand  knowledge of  Intelligent  of  a  semantic  pragmatic  needs  a  such  i t  system. of  commonly  For  tied  and  pragmatic steering  SHRDLU  is  simulated a  a  a  a or  cf  the  theorem-proving  major  role  cf  prove  contribution has  interplay  the  Since  course  syntactic,  and  It  l i n g u i s t i c s  i n t r i c a t e the  possible.  power  the  of is  to  procedural,  general  body what  the  a b i l i t y  be:  a  body  together.  direct  must  to  in  recent  been  which  is  semantic of  and  inference  SHRDLU  language  created  in  of  Fcdor  the  simulated  blocks.  (program-oriented)  theory  in  is  an  approach  fcr  expressing  S  the  notion  and  created  excellent  Intelligence,  Sussman)  an  theorem-preving  SHRDLU  tc  implement systematic  (1964)  robot,  this  of  MICBC-PLANNFB Katz  in  Tc  Charniak  Hewitt's  and  SHRDLU  rebct.  PROGRAMMAR  (with cf  implemented  which  accomplishment  Halliday's  (Micro-planner)  physical  procedural  of  Winograd's  special  markers'  micro-world  to  they  the  implemented  PLANNER,  •semantic  both  knowledge  grammars,  implementation language,  type  between  of  Terry  created  procedural  some  of  needs  order  impressive  Winograd  grammar,  in  these  needs  a  only  and/cr  a l l  the  needs  net  true  system  essential  knew  be  t i e  find  it  100  analysis.  most  direction  may to  a  grammar,  may  also  l i n g u i s t i c s  bodies  The  tc  and  necessary  the  known  together  demonstration  possible  it  Theory..  language,  i t  structure  hypotheses  computational  the  that  language  is  and  is  Accounting  syntactic  structure  control  programming  parse  what  control  disprove  cf  so  ana  natural  knowledge  but  a  a  theory  knowledge  meaningful  MIS  cf a  manipulated example  epistemology.  knowledge  and  of In  A l .  IMIS:.  Intelligent  Winograd's  system  Chapter  4)  the  defined  by p r o g r a m s  called  these  implemented SHRDLU note  i s  Winograd's  i s  language  his  i t s e l f .  Thus  an X  order  to  the for  (that  to i s ,  data-base  that a  natural  A  words  powers.  Wincgrad and  conversation  careful  in are  s p e c i a l i s t s '  The i n t e r e s t e d  and h i s  described  language  'semantic  2.  this  r e t r i e v a l  with  reader  should  analysis  a  r e t r i e v a l  P a r t (X)  whether Given  be i n v o k e d  to  a b i l i t y  'by this  of the  of  f a c i l i t y  being),  in  i t s e l f  modifies i t  a  i s  a  level  part  tc  by t h e  "Does  their  the verb  simple cf  managerial  a  system  a n d By ( X , D o o r ) "  the top as  i±S...  guesticns  "Put the spare  the door'  understand  The  addressing  enquire  at  to  a b i l i t y  a n d S p a r e (X)  d i r e c t l y ,  human  of  subsystem^  f o r  beginning  t h e system  that  order  guestion-answering  f a c i l i t y  phrase.  by  project  and a r b i t r a r y  claims  sentence  determine  noun i t  such  KIWI  101  Theory..  system.  a  requires  exist  f u l l  information  employs  of  Appendix  evident  a  our  MICRO-PLANNER.  modest  already  "  i n  and Accounting  programs  i n  subsumes  managerial  door  with  l i t t l e  of  i n  semantics  quoted  It  (as  them  l i m i t a t i o n s  HIS  i n cr  matter command  query  or  system.  Intelligence,  knowledge  and A l .  IMIS:.  1±§  LANGUAGES  1HI  The be  a  Intelligerit  choice  significant  posed  i n  and  Accounting  AETIFICIAL  of  notation  for  representing  towards  i t s  solution.  step  designed clear  INTEIIIGENCI  Theory..  OF  A r t i f i c i a l  languages  MIS  Intelligence to  the  manipulate  maintain  a  d i s t i n c t i o n  between  program-creation  £  102  A  a  problem  Fcr  the  conventional  numerical  data  between  program  and  program  use  are  may  problems computer and data  to and  guite  inappropriate.  1*6*1  brief  1  LISP in high  i s  l e v e l  Chapter  4,  very  al.,1962)  a r b i t r a r i l y non-numeric  o r i g i n a l  a  We  and  LISP  as  for  complex  lambda-calculus through  many  structure  be  of  will  was  and as  an  intent  have  by  and  John  cf  Intelligence,  the  test  knowledge  the  of fcr  Alcnzc  expressions. then  as  McCarthy  structures  since  in to  such  describe  embodiment  stood  the  foreign  manipulation  data  functional  implementations and  tc  created  creation  and  languages  digress  'very  described  guite  commercial  employed  most  embedded  power  the  f o r  are  which  study  1,5  and  in  empirical  program  environments  freguently  language  therefore  b r i e f l y . tool  most  languages  knowledge  COBOL.  a  the  the  structure  with  language  for  LISP..  language  Intelligence,  Its  or  evolved  computer  employed  readers  Church's  to  thecrem-prcving  1  FORTRAN  (et  the  A r t i f i c i a l  language  most  introduction  It but of  has i t s time.  and A l .  IMIS  X  Intelligent  MIS  and  Accounting  103  T h e o r y  46  Structure^ l i k e  P/L  1 or  programs  LISP  Algol.  i s (Tt  net  have  an  Algol-like  languages').  very  level  high up  his  structural  own  or  ordered  n-tuples F c r  of  atom  *  6  or  • (  and  an  A LIST  )»  is  empty  i n  atoms  desirable  and  the even  and  user  (which COEOL) cr  cwn  LISP  l i s t s  Such called  very  lew  extend  and  i t  syntax has  resemble and  a  two  variables  in  which  contained  by  from  only  i l i s t s ^  x  create  are  may  his  units.  tc  statements.  simultaneously  since  language  are  within  example:  THESE  'NIL'  is  ^atoms^  i d e n t i f i e r s  and  format  predefined  Fortran  IS  It  prcgramming  GO T O - l i k e  functions  components:  parentheses.  use  language  primitive,  (THIS  possible  do  'structured  the  is  structured  which  languages  building  a  ARE  (WITH  a  ATOMS  5  ((SEVERAL)  distinguished  ATOMS  NESTED  and  (SUBLISTS) ))) .  structure  which  is  both  an  l i s t .  See McCarthy e t a l . ( 1 9 6 2 ) , Magadin and Segcvia (1974), and Weissman (1967). The most p o w e r f u l e x t e n s i o n a t t h e moment w o u l d a p p e a r t o b e t h e I f l T I B L I S P of Teitleman et al. (Teitleman, 1974).  Intelligence,  knowledge  and  A l .  IMIS_:  Intelligent  Lists special  are  storage  successor  MIS  constructed  c e l l s  branches  of  be  the  With possible even  the  node.  Thus  does  not  l i s t s  of  f i e l d s ,  'programs )  i d e n t i f i e s identify return to  the  For  the  the a  value  current  example:  i f  function  binary  the  against  the  is and  which One  structures  may  may,  of  of  ether  f i l e s . in  LISP  remaining  that  those  it  interpreting  length).  data  the  of  two  l a t t i c e s ,  cycles,  and  values  structure  LISP  the  the  the  1.9.  tree  wherein  is  to  with  l i s t  Figure  i n f i n i t e  and  of  has  the  104  structures^  structures,  user  which  VAL8  in  functions  l i s t s ,  arguments  a  simple  records A H  are  1  the  tree  pointing  (although  the  create  of tree  cycles  functions.,, or  n-ary  Theory..  (ATOM) ) )  shown  protect  as  easily  languages:  (AH  simplicity  processed  course,  IS  the  with  binary node  structure  construct  as  each  tree  graphs  program be  to  Accounting  for  ((3)  would  and  f i r s t  (i.e. or  of  only  elements  function.  result  •procedures' element (if  Function  applying  that  any) forms  function  arguments.  value,  8  then (ADD  3  VftL8)==>11,  that is, t h e f u n c t i o n ADD i s applied 3 and t h e v a l u e c f t h e atom VAL8. (TIMES VAL8 (SUB (TIMES This last equivalent  64)==>512, and VAL8 64) (ADD VAL8 form is evaluated Fortran expression,  tc  the  the  atom  in the same order as ( (VAL8*64)- (VAL8+VAI8)).  the  VAI8))==>  Intelligence,  value  cf  496.  knowledge  and  A l .  105  IMISj. I n t e l l i g e n t MIS and Accounting Theory..  106  Again i f VAL#OF#L has the value (3 ONE TWO THREE) then (LENGTH VAL#OF#L)==>4, and (SETQ RESULT (LENGTH VALfOF#L))==>4 side-effect of setting the value (LENGTH VAL#OF#L). Functions exist for creating (LIST VAL8) ==>  also, cf  with RESULT  the to  lists:-  (8) ,  (LIST (LIST (LIST VAL8)))==> (((8))), for adding to them:(CONS 32 VAL#OF#L)==> (32 8 ONE TWO THREE), (APPEND VAL#OF#L (LIST VAL-8))==> (3 CNE TWO TEBEE 8), and f o r dismembering  lists:-  (CAR VAL#0F#L))==> 3, (CDR VAL#OF#L)==>(ONE TWO THREE), (CAR (CDR (CDR VAL#OF#I)))==> TWC, and so on. * 7  *  7  See glossary for CAR, CDR.  Intelligence, knowledge and A l .  IMISj.  Many control.  I n t e l l i g f n t  functions  This  (COND  ( (EQ (  w i l l  l i s t ,  the  value the  user  the  system  one  value. by  which  is  remaining arguments' (to  any  outer  4  8  A  conditional  function,  COND:-  VAI-8))  and  the  item  arguments.  depth)  functions  These  themselves  be  he  an  be  completed.  the value l i s t .  that  may  even  He  when  it  atcm  passed  is  the  EVAL  the  value  Thus  the  the  cf  the  in  cf  atom the  functions  evaluation  of  s>  function  Intelligence,  the  remaining  Innermost the  i t s  obtained  values  on  LISF  returns  *  8  tc  ' l i s t e n i n g '  with  before  of  is  normally  functions.  can  l i s t ,  associated  'values  evaluated  atcm,  wishes.  returns  l i s t  be  with the  is  EVAL  must  Or a s s o c i a t e d f i r s t item of  i f  function  the  an  element  functions.  d e f i n i t i o n on  4  is  i n c i d e n t a l l y ,  form  l i s t  a  evaluated  the  the  function  f i r s t  own  is  value  ).  functions  thus  a  is  (  (which,  If is  that  VAL-8  his  form  EVAL  i f  l i s t ,  system  form  may  of  create  on).  the  the  and  4))VAL-8)  VAL-8  may  i t s e l f  executing  of  empty  basic  signs If  (LIST  the  function  interpreter when  conditional  VAL-8  value  Evaluation.. the  the  control  107  NIL))  otherwise  redefine  for  VAL-8) (LENGTH  i f  The  of  A c c o u n t i no, T h e o r y . .  T  return  otherwise  and  exist  example  ( (ATOM  MIS  which  knowledge  is  and  the  A l .  IMIS:. I n t e l l i g e n t HIS and Accounting T h e o r y  108  (LIST (ADD1 VAL8) (APPEND VAL#OF#L (CONS (LIST VAL8) (CDR VAL#0F#L))))  the  forms  (LIST  VAL-8)  The  function  evaluated.  S  (CDR QUOTE  VAIfOE#L) are f i r s t to be suppresses  any  further  evaluation; thus (QUOTE VAL8)==> VAL8. Property lists.: although most functions treat atoms as indivisible  there  exist  t h e i r fine structure. value,  atoms  may  some  which  can access and modify  In addition tc a  'print-name  have a 'property-list'.  1  and  In every respect  property l i s t s are normal l i s t s and their content may be LISP  structure.  The  normal  usage,  a  however,  is  any  fcr the  property l i s t to consist of alternate property-lNDICATORS and property-VALUES; for example the atom  PART86  may  have  the  property-list  (COLOUR RED  COST 28.50 DIHENSICNS (10 131 2.5))  it  i s by  are  associated with  entire  body  means of property l i s t s that particular functions  of  the  the  names  program,  cf  particular  the  stored as a property-VALUE of the atom  atoms.  The  function expression, i s under  the  prcperty-  INDICATOR •EXPR • . Intelligence, knowledge and A l .  I M I S I n t e l l i g e n t  Binding^  Binding  value  with  a  LISP  are  r e l a t i v e l y  the  value  function  LISP  "AL"  MIS  is  form.  at  and  As  the  most  the  in  l o c a l  X  28)  (PROG  EXECUTED THE AT  ENTRY  TO  (SETQ  the  X  A l l  other  modify data  or  t h e i r  in  l e v e l ,  X  and  and  Y  if  the  in  both  have  nested  THOSE  between  in  LISP  things  this  themselves own  program  GLOBAL  ...)))  THE  LOCAL  EVALUATION VALUES  VALUE OF  ABE  THEM  is  an  uncompiled are  is  form  AN  Y)  AND  Y  28  AND  RETURNS  STACKEI REAPPEAR  ON WHEN  FUNCTIONS.  in  a  ACQUIRES  LISP  form  (PRINT  RE-BIND  core  a  'UEC) X)  WHEREAS  WHICH  forms  when  X  MORE  prograajSj.  remains  point  programs  THE  FROM  and  d i s t i n c t i o n Among  UBC,  FUNCTIONS  mutable. at  VARIABLE  VALUE  EMERGES  code  If  bindings  a  . . . . (X)  LEVELS.  Data a l l  THE  LOCAL  ALL  CONTROL  associating  Algol  global.  global  (PRINT  "AL"  and  of  109  (X) (SETQ  LATER  operation  P/I-1  or  Theory..  c a l l s  (PROG  ARE  Accounting  that  to and  this -  net  cr  EVAL a  programs  with  structure  and  atoms  'program'  means (and  form  simply  passed  as  i n t e r p r e t i v e is  completely  l i s t s . is  form may  Only  there as  the  a  ' d a t a ' .  modify  f a c i l i t y merely  systemj  ether  they  may  values  of  switches).  I n t e l l i g e n c e ,  knowledge  and  A l .  lUIiHi I n t e l l i g e n t MIS and Recounting Theory..  Recursion. itself  110  I f t h e d e f i n i t i o n o f a f u n c t i o n c a l l s cn  d i r e c t l y o r i n d i r e c t l y we have r e c u r s i o n .  Fcr example  the f o l l o w i n g f u n c t i o n ATOMS (shown here as i t i s d e f i n e d v i a the f u n c t i o n DEFUN) w i l l r e t u r n a l i s t c f a l l atoms at  any depth  (sublist)  t h a t ATOMS c a l l s  occuring  i n an argument passed t o i t .  Note  itself:  (DEFUN ATOMS (ARG) (COND ( (NULL ARG) ARG) ( (ATOM ARG) (LIST ARG) ) ( T (APPEND (ATOMS (CAR ARG)) (ATOMS (CDR ARG) ) ) ) ) )  A l t h o u g h a r e c u r s i o n can i n p r i n c i p l e be ' f l a t t e n e d an i t e r a t i o n t h i s f a c i l i t y an  elegance  and freedom  1  cut into  f o r u n l i m i t e d r e c u r s i o n g i v e s LISP net  shared  by  non-recursive  languages.  Intelligence,  knowledge and A l .  IMISj.  Ii.64.2  Very  I n t e l l i g e n t  high  l e v e l  A r t i f i c i a l exciting  in  for  languages  commercial lay  Theorem^proying of  in  are  the  which  proofs  rules  of  character-strings and to  languages increase  higher-order blind  rules  to  of  cr  we  believe  Much  the  created  of  111  implementing  at  by is  for  (ii)  which a  be  a  valid  specialized f a c i l i t a t e d  the  purpose,  of  our  logic.  least  and  sought  cf  the  the  a  system of  theorems  be  may  in  matching  computer  programs  mere  so  proof-finding  extra-logical  (i)  representation  problem by  find  cf  deductive  require  will  and  inspiration  problems  for  novel  the  be  e f f i c i e n c y  l o g i c a l  has  knowledge-base  may  designed  the  Theory.,  axiomatic  task and  in  systems  the  The  which  systems  inference  constructed.  research  systems.  theorem-proving  representation  Accounting  languages^  languages  application  e f f i c i e n t  and  Intelligence  computer  these  MIS  f a c i l i t i e s  i f by  tc  we  are  adding  steer  cur  inference.  Intelligence,  knowledge  and  Al.  IMIS.: I n t e l l i g e n t MIS and Accounting  Theory..  112  lJ2i=o£iiJBz£Ioving languages^. To us  i l l u s t r a t e specialized theorem-proving systems l e t  consider  Imagine  that  the propositions  presented  in  Figure 1 . 1 0 .  many others exist i n the system.  x,y S 2 are  variables, p, r , n and a are constants. r~ 1. 2. 3. 4. 5. 6. 7.  8. 9.  10.  11.  12. 13. 14. 15.  = X i s unexpected Ux = X w i l l upset the Market Mx = X i s an Audit Report Ax = X i s a loss Lx = X i s a new product Nx = X makes reference to z Rxz Cyx = y created x = X i s the President of the company Px (x y) LX V (Rxy 8 Ly) -> Mx = a l l losses or anything which refers to losses w i l l upset the market (x y) Cyx 8 Py 8 -Lx -> -Mx = everthing that the president made (except losses) w i l l not upset the market (x z) Ax 8 Ux 6 -Rxz 8 N Z -> Mx = everything which i s an Audit Report and unexpected and doesnt refer to a new product w i l l upset the market (El) 11 = 1 i s seme losses (Er) Ar = r i s an Audit Report (1) Ul = a l l losses are unexpected (Er)(Ea) Car = the auditor created the Audit Report  Figure 1 . 1 0 : The effects c f Audit Reports on share prices.  Intelligence, knowledge and A l .  IMIS:.  Of  this  predicate  data  calculus  such  Will  •  Is  the  Audit  •  Is  the  auditor  (fr  anything  If  b  (Ab  —>-  we  demands  is  will  MIS  base  •  •  But  Intelligent  we  the  Report  an  might  Accounting  ask  and  market?  (\  Theory..  answer  upsetting? also  the  (|-  (Ex)  questions  in  Mr)  Kx)  *  9  president?  Audit  Report  is  trouble  with  such  b  5  (|  ° -Pa)  not  a  loss?  Lb)  have  reasonable  i t e r a t i v e  as  • Find a l l upsetting things. c o n s t r u c t x = { x: (Ex) Mx }.  This  requires  • If the P r e s i d e n t w r o t e an A u d i t R e p o r t the a l t e r n a t i v e sets of c o n d i t i o n s i t would satisfy.  And  113  as:  upset  not  and  our  simple  propositions  make  no  us  tc  enumerate need tc  reference  to  time  or  causality.  *  9  The  symbol  " \"  is  explained  in  the  glossary.  See  ' e n t a i l s ' . so  (The  symbol  (Ex)  and  is  employed  this  type-font  means  "there  instead  cf  exists  the  an  x  standard  such  that  inverted  . . . "  E  which  and  A l .  lacks.)  Intelligence,  knowledge  IMIS_2 I n t e l l i g e n t  which  HIS  A c c o u n t i n g Theory..  We  introduce  this  data  to  discuss  the  complexities  suggest  the  particular  theorem-proving  languages  of  Such  inference.  (Hewitt,1971a  base  as  apart  languages  McDermott,1972a,1972b),  interesting  a  immediate  directly  t o our c o n c e p t c f t h e IMIS  discuss  in  examples  t o MICRO-PLANNER  Ecr  and t c  (Sussman,  Wincgrad,  (Sussman Derkson si  on  which  reasons  rules  Planner  application  in  of  example  (Rulifscn,  philosophy  C h a p t e r 3.  against  problems  Conniver  They  we  a  and  are  not  tc deductive  would  manner  o f s p a c e we  but t h e i n t e n t i o n s ,  S  transfer which  we  limit  cur  structure  and  of a l l these languages o v e r l a p .  • Notation primitive The  for  MICRO-PLANNER  logic.  capabilities  and  S A I L , POELER and so  for their  embody  background  which c a n a r i s e  -  QA4  111  from t h e c o n v e n t i o n a l  Baumgart,1972)  W a l d i n g e r , 1 9 7 2 ) , QLISP,  a  strategies  1971b,1972),  Charniak,1972;  only  and  usual  is a first entities device  we  problem. need  is  tc  With a  a larger  rich  store  use L I S P - l i k e  of  set  cf  symbols.  n-tuples  as  follows:  (UNEXPECTED AUDIT-REPORT) (CREATED AUDITOR  si  (LOSSES L)  AUDIT-REPORT)  in a r e c e n t a r t i c l e Bobrow and R a p h a e l (1974) p r e s e n t a most u s e f u l a n a l y s i s and c o m p a r i s o n c f t h e s e l a n g u a g e s .  Intelligence,  knowledge  and A l .  IMISJ.  An  axiom  I n t e l l i g e n t  such  as  conveniently will  see  its  the This  proven:  Mx.  disprove f i r s t  •  the  finds  that  Try  •  the  we  data  once  the  relevant  Al  languages  can  is  in  the  to  they  free  where  or  which  data  f a i l  'is  as  prcver to  in  the  f u l l y be  the to  cr  quoted any  or  While  wants,  large  a  seguential  the by  form  cf  instances  i t s  devices  from  data.  cf  He  the  n-tuples procedures  such  are need  need  most"  associative  r e t r i e v a l  these  base.  search  Therefore  tc a l l  It  f i r s t * .  instantiated  the  data  i t .  information:  partly  programmer find  x  propositions  some  or  any  prohibitive.  implement  prove  disprove  cf  grows  be  cr  base  is  either  tc  prove  steering  base  We  anything  conclusion  may  some  true  less  there  a  used  base.  he  or  is  shortly.  have  axioms  how  guestion  theorem  115  l i s t - s t r u c t u r e .  fcr  find  data  a  true  themselves order  as  Theory..  1 . 1 0  Figure  net  Herein  in  in  presented  The  data  for  memory.  f i r s t is  Mx  Accounting  representation  proposition  Already  9  number  t y p i c a l  upsetting?". (Ex)  and  represented  Consider  |-  HIS  records  s t i l l to  only  slow  specify specify  'pattern*.  Intelligence,  knowledge  and  A l .  IMISi  Next directed axiom  (x  9,  f i r s t  •  let  search  ' s p l i t t i n g a  us  y)  Lx to  MIS  say  guided  r u l e '  ana  that by  V  via  Mx  (Hxy  this  Accounting  S  the  the  Ly)  —>  disjunction  116  Theory..  aevice  of  pattern  theorem-prover Mx.  and  It  finds  applies  a  <J L x  as  establishes  subgoal.  Problems:  which  should  s p l i t  into  subgoals  In  Intelligent  this  succeeds true;  1  upset  the  If be  which  several tried  f i r s t ?  several should  serendipiticus immediately instantiates  in x;  axioms If  match an  conjunctive be  t r i e d  example the  axicm  pattern may  be  or  disjunctive  subgoal  \  f i r s t ?  the  data  therefore  this  Ml  base: is  L l  now  is  (Ix)  known  true;  Lx to  be  flosses  market'  Intelligence,  knowledge  and  Al.  Intelligent  IMISJ.  MIS ana A c c o u n t i n g  li£.2.2 The MICBO-PLANNER Let style (x  notation^  us now encode t h i s  of  y) Lx V  these  i n MICRO-PLANNER  languages.  (Rxy £ Ly) — >  (THASSERT  Mx w i t h  we t h e n  (THPROG  •  (X) (THGOAL  from  embedded •  initiate  A l l MICRO-PLANNER them  Ll  1  in  2  Y) (UPSETTING $?X) $?X) $T) (REFERS $?X $?Y) $T) (LOSSES $?Y) $T) ) ) )  our s e a r c h  with  3 4 5  the g c a l : 6 7  LISP  with  'TH' t o  functions.  most  distinguish  MICRO-ELANNER  b u t h a s i t s own c o n t r o l  like  and  the f u n c t i o n c a l l s :  f u n c t i o n s begin  i n LISP  operates  and  assert  {UPSETTING $?X) $ T ) ) )  similar  MICRO-PLANNER,  That  We  to capture the  (LOSSES I ) )  (THCONSE WHATS-UPSETTING (X (THOR (THGOAL (LOSSES (THAND (THGOAL (THGOAL And  117  Theory.  structure.  theorem-proving  a 'reverse-guote•  is  languages,  notation f c r convenience.  i s (CREATED AUDITOR AUDIT-REPORT) means what i t s a y s net 'apply  AUDITOR  &  the f u n c t i o n  AUDIT-REPORT',  CREATED which  to  the  are  values  intended  tc  of be  constants. •  Therefore  the  explicitly. function,  value This  of  a  variable  i s done w i t h  (THVALUEOF...)  and  must  be  accessed  an ' a c c e s s - t h e - v a l u e - c f "  $?x  is  a  shorthand  for  (THVALUEOF X ) . •  Matching  variables  with  constants  is  Intelligence,  governed  by seme  knowledge and A l .  l£i§lIA3.§rit MIS I D J A c c o u n t i n g  precedence r u l e s f o r i n s t a n t i a t i o n we  say  (THGOAL  pattern  i s fully  Taxes  »  upsetting?") then  line  2 with  $?x  undefined.  from  the  axiom  in  lines  is  a  code 'try  the  thereby  i  i  Eut On  as our  2 t o 5 $?x  fcr  a  line line  2 $?x  be  The that will  7 stands  successful  will  anything  recommendation -  data  (THGOAL expedite  existence  of  (UPSETTING $?X) the  axiom  bound  we  exit tc L  cr  (meta-kncwledge)  theorems - t h a t  9,  we  could  say  (THUSE WHATS-UPSETTING))  may  in and  search.  T h r e e more major f a c i l i t i e s  Ii6 2 3  (i.e.  goal!'  know o f t h e 7  enter  If  "losses".  attain  we  v a l u e TAXES.  7  means " i s i t t r u e  i f we  enter  "$T"  line  and  h e l d the  meaning  If  instantiated  118  a i  variables.  (UPSETTING TAXES)) i n l i n e  assume and  i.e. %  are  c f the  Thecr  remain  t c be  demonstrated.  Back-tracking^  Intelligence,  knowledge and  Al.  IJ3I§1  Intelligent  Consider Auait  Report?"  instead fina and  of  (-  axiom  9,  to  discover  in  prove  B.3£K~tracks  1)  backward!  runs The  prove  that  reference "-Mx'J  which  is  to  MICRO-PLANNER  other ' f a c t s  an new  could  contradiction  predicate,  of  of  subgcal.  1  net  base  say,  ana  which that  and  the  other  no  axiom  we  may  changes  then  audit  products.  It  be  equally  the  Cx  goal. by  the  <==>  such  a r b i t r a r i l y  as  But  clumsy  QLISP,  complex  Axiom led  cannot  device ('C  -Mx.  f a i l . i t  fcr  permit  Mx  The  to  makes  nc  implies a  guick  denials  inventing  a  'cemfcrting). the  was  starts  10  to  assert  cf  therefore  and  which  i f  one  11  The  undone..  report  will  useful  axiom  an  f a i l e d .  (where  are  x  assist  has  control  made  tc  grateful!).  point  finds  is to  chesen  languages  decision  a l l  be  Ax  theorem-pro ver  exists  strategy  unexpected  except  languages, 1  our  theorem-prcver r  6  A l "  last  program  Mx  1  as  the  (Ix)  bind  these  with  let  f-  upsetting,  concluaes  instantiated  again  as  are  (For  the  this  upsetting  losses  data  to  frame  an  that  the  and  exist  Once  have  MICRO-PLANHER  there  119  Theory..  Mx.  we  Al.  theorem-prover In  S  Accounting  "Does  carelessly  Ax  Now  Report  guestion  We  (Ex)  return.  Audit us  the  ana  MIS  assertion  structures,  including  Intelligence,  knowledge  in new Seme as  denials.  and  Al.  IMIS.:  This  process  d e p t h - f i r s t disagree  tree  only  exercisable i s  Intelligent  i l l - s t r u c t u r e d thrashing  cash  is  the  equal  express  the  this  in  But  this  a  Bxa  is  true  Thus  Funt  function explore QLISP IS)  other has  Bc3,  paths  the  &  for  i f  waste  of  i t  or  an  much  time  backtracking  is  less  as  "the  any  Dxb  (a-b)  (for  Bx  non-2ero  we  b). tc  be  HICBC-ELABHFF  a l t e r n a t i v e backs  functions  IS  $2  to  GOAL,  up and  to BIS  balance  from  balance  f i r s t  implication  f a i l u r e  new  calculus.  "the  deducticn". At  —>  as  "deduct  calculus?  the  predicate  interpreted  modified an  twin  i t  backtracking  misfortune  may  bane  hypothesis  c o n t r a d i c t i o n ,  (1973)  a  interpreted  predicate  "IS*,  against  a  truths.  are  balance  then,  is  MICRO-PLANNER  Conniver  erasing  Dc2,  old  In  It  automatic.  Exa  52  In  causality  and  over  with  program  used.  directed  control  120  Theory..  widely  programmer.  the  52 not  is  programmer:  instrumental  to  is  and  propositions,  $3"  accounting  c r i t i c i s m s  r e l a t i v e  the  and  and  the  Consider i s  is  Asserting  Consider  The  database  and  Time  of  by  b l i n d l y .  discouraged  JU6i2iii  the  program  uncontrollable  and  back-tracking  search.  over by  cf  MIS  true,  Ex  by  we  write  i s ,  which  cash  may  That  i t . (=  cf  How may  cash".  i f (a-b)  adding does  a not  Similarly  'Backtracking'  .  Intelligence,  kncwledge  and  A l .  IMISj.  must is  also  true  which  true.  The  a  later  time  we  i t  theorem  attempt  is  true  grows  match every  be  at If  I n t e l l i g e n t  time.  in  Thus  the i t  than  also  solution  t1  a  becomes  true  near-chacs.  added  the  to  "assert J ,  ends  x  time  w i l l  then  base  Bxa  fi 6  5  3  £  in  which to  is  to  to  a l l  other  and  recorded  —>  A s s e r t (Bx  the  (a-b)  it  time  at  axicm  cr  must  wffs  and  true  at  both  re-assert the  new  propositions.  The  M I C R O - E L AN NER regards  the  true  upon  data  i t  The  as  base.  is  is  If  as  predicates  f i r s t - o r d e r  emits  true  erasedj  effect  always-true  accomplished  Dxb  in  asserted.. two  be  Bt1x  propositions  is  Ex  Every  since  —>  in  be  predicate  other  Hewitt  to  that  write  Dtlxb  regard  operate  be  complexity  values  by  calculus  which  in  121  3  is  ceases  it  5  result  currently  proposition  Bxa.  the  employed  propositions  is  course,  every  misleading  from  of  in  having  time-symbols  Theory..  record  TOxa  without  Accounting  to  data  is  and  solution,  appallingly  appropriate wff  MIS  a l l  at  time  i f  one  thcugh  cne  ^erase^  and  predicates.  Our  by:  E r a s e (Exa)  S  Erase  (Ext)  (a-b))  Non-logician programmers (material implication) in after a".  must a p p r e c i a t e t h a t " a — > no way i m p l i e s the sequence  Intelligence,  knowledge  and  b" "b  A l .  IMISj,  which  i s  a  Intelligent  true  implication  2^.64.2^.5 C o n s t r u c t i v e Not are  a l l  discards  terminates from  deep  f a c i l i t y  i n  have on  these  a b i l i t y  to  of  languages abandon l e v e l ,  may  i n  a  of  recursion  of  a  exists  tc  the tcp  tc  i s  the set  t c  cause  pure  deductive  pass  which  a  message  5  *  A  of  the  a b i l i t i e s  newer  flow 5  some cr  process  (at  when  t h e new p r o c e s s  of One  5  in  processes  and l a t e r  crude  but  deductions.  another  included)  f a i l u r e  levels.  vary  p a r a l l e l  invoke  in  subsearch  MICRO-PLANNER  invoke  i t ,  in  intermediate  process,  warrant  true.  end  search  a b i l i t i e s  these to  means  up towards  this  which  gained  and no  basis  tree  the blind  general  variant  hierarchical evidence  very  122  Theory..  Bxa and Exb a r e  search  search  f o r  Accounting  failure^  insights  a  the  interesting  a  f a i l u r e  exists  control  s*  in  and  i f  of  Yet  a l l  within  languages  of  paths  use  uninformative,  logic  MIS  the any  tetter or  any  we f o u n d t h a t t h e (THFAIL THMESSAGE) f a c i l i t y had c e a s e d t o work on t h e i m p l e m e n t a t i o n o f P l a n n e r a v a i l a b l e tc us.  If our example were more c o m p l i c a t e d so t h a t the gcal f- ( E x ) A x m i g h t be attempted frequently i t wculd be useful t o p a s s upward a message t h a t A l i n p a r t i c u l a r is false. Or c o n s i d e r t h e f a i l u r e to discover that Audit Report r i s upsetting: with blind search the system will f a i l only after f a i l i n g to prove the sub-goals I- ( E z ) (Ar & R r z SNz) and J- P a . After the f a i l u r e the higher, calling s y s t e m o r t h e human m a n a g e r should be a b l e tc be t o l d " I f only I h a d known t h a t 'The auditor is not the President' I could have saved a l o t o f t i m e . " 5  5  Intelligence,  knowledge  and A l .  IMIS.:  Intelligent  other  to  also  possible  which  was n o t i t s  1 . . 6 2 ..6  resume  Adding  A  Because in  the  which  to  a  resume o r i g i n a l  axioms  or  knowledge  to  some  Assert  f o r i t .  IMIS  may  sophisticated available  to  between  bases  a  and i t  in  this  managerial be a s s e r t e d  It  stored  dees  i s  context  with  knowledge  maintained  n o t have  to  using  t h e model  te  there  more as  new tied  programmer.  wherever  scon  ty the  subsystems  the p o s s i b i l i t y  manner, as  programs  of  b y t h e human  available  tools into  items  memory  system i s  ^as-needed^  are  system  We a r e i n t r i g u e d grow  123  abandoned.  o n some  related  associative  control  t h e new d a t u m  The c r y .  context.  such  sequential  c a l l  process  procedure and  added  had been  and procedures.:  the links  axiom  and Accounting  the process  pattern-matching  into  MIS  they  that and  i s  a an  mere become  base.  Intelligence,  knowledge  and A l .  IMIS 2  It  is  also  procedures  possible which  occurrence  demons  X±6±2±J  MIS  most tc  their  the  are  in  are  of  MICRO-PLANNER such  Intelligent  only  the  and  Accounting  theorem-proving  be  invoked  pattern two  Theory..  and  conditions  assertion  or  Non-declarative logics  by  and  languages the  some  of  tc  add  simultaneous  condition..  allowed  erasure  124  fcr  activating  assertions.  theorem-proving  In  5  6  languages..  Curiously MICRO-PLANNER does not provide a b u i l t - i n d e v i c e t o e n a b l e ANTECEDENT t h e o r e m s (the ' c n - a s s e r t i c n ' demons) or ERASING theorems tc f a i l (inhibit) the c o n d i t i o n which invoked them. This limits the u t i l i t y cf the  S  6  antecedent  theorem  in  particular.  Intelligence,  knowledge  and  Al.  IHHSj.  Intelligent  Consider propositional  the  and  MIS  and  following  deontic  logic,  1:  N  is  2:  A l l  a B  2:  N  2:  Dc  B  A i f B is t o be attained,  2:  Do  Dc  A  tc  is A  dc  i f E t o be  logic  E.  is attained,  to  do  (one  A.  example)  necessary i f  E  is be  tc  necessary attained.  Therefore Do A!  Therefore: (N:)  ought  Modal  logic  1:  Do  in  logic  Therefore: N ought  B!  (N:) Do  A.  1:  A  Imperative  arguments  7  Deontic  logic  are  cf  ARGUMENTS.  B.  Therefore: N is an  1:  5  125  Theory.  juxtaposition  DEDUCTIVE  Declarative  Accounting  a!  i  i Figure  s  7  1.11:  After light  Declarative  Mattessich of  D.  S.  and  non-declarative  (1972e,  Clarke  p . 14  (1973)  and  p . 16)  deductions.  expanded  in  the  ,  Intelligence,  knowledge  and  A l .  IMISi  In the  such  c a r ! " ,  the  car",  truth (if  special  deontic  and  value i t  Intelligent  even  of  even  deductive  logic  the  and  modal  sold  as  the  are  philosopher's  forms:  is  car".  problem  is  of  conditions of  to  from  each  each  have  other.  kind  be  to  so  is  a  that  and  a  simplest  of  the  hardest  proposition years"  i t  interwoven  "This be  a  w i l l  reason. in  an  the  the  means  that  s e l l  become  hypothesis  hypothetico-deductive  a  which  is  the  its  nova is  Intelligence,  ncne,  well  on  'event' car!" of  seme  the  car"  why  this  that  there  car"  truth  within  s c i e n t i f i c  s e l l  true  s e l l  have  so  authority  reason  i t  which  the  moral  may  sets  seme  s u f f i c i e n t  It  the  ether  case  tc  yet  the  such  and  fact  will  or  preposition.  ought  "He  and  for  legic  each  i f  of  determine  main  "Jcnes:  true  The  prepositions  Thus  be  car."  extensions  certain  in  appearance  w i l l  above  in  s e l l  declarative  "He  determine. star  past  Notice  has  tc  discoverable  may  that  Jones.  the  " S e l l  x  consideration  values,  speaker  exists  three  to  may  occur.  statement  s u f f i c i e n t  of  we  on  126  ought  such  imperative  also  upen  the  define  example  logic,  the  goal  there  the  to  i f  easily  which  truth  Notice  yet  true  impose  implies i s  has  as  worthy  comments  declarative  'action'  might  the  for  must  s e l l  propositions.  meta-linguistic  Demarcating  or  exist  propositions  make  the  in  "Ee  will  with  to  Theory.  injieratiyes  Nevertheless as  truth  as  not  case  regarded  circumstances  "He  is  the  i n s t i t u t i o n a l value  Accounting  prcpositicns_:  proposition  exists)  "He  and  sentence-forms  future  the  sentence,  MIS  cr one  is  the  value  is  like  the  million  corroborated  by  b e l i e f s .  knowledge  and  A l .  IMIS.:  In or  an  o u r IMIS  I n t e l l i g e n t  instrumental or  a  Communicating one the or  of  "In  this  these  Sheet!",  forms  point  languages s u f f i c i e n t l y modification  device,  forms  tc  f o r  an a c c o u n t i n g  purpose  should  be i n  i s i s  to  to  this  valid  produce  that  involve  10  "Erint  mere  units!"  logic  whether  arguments.  the procedures  At and  which  work  reguire  minimal  non-declarative  logics.  deductions i n  served..  e.g.  i n  theorem-proving  conclusions  tc  system  room".  deductive  observe  te  l i k e l y  disagreement  on d e c l a r a t i v e  deduce  as  the system  be some  automatic  well  a  "Y.ou  merely  such  127  Theory..  we a r e d i s c u s s i n g ,  do r e p r e s e n t  we w i s h  to  to  you w i l l  appears  and Accounting  system  the purpose  minutes  10  There a l l  robot  the proposition  Balance  HIS  Intelligence,  knowledge  and A l .  IMIS.:  For  example  following  Do If If  A! A B  Intelligent  consider  argument  do do  B E  the  in  8 8  MIS  and  MICRO-PLANNER  imperative  C F  8 8  E M  do do  M! 0!  Therefore Do 0 a n d  F  N  and  In MICRO-PLANNER ('THAND * c o r r e s p o n d s tc our example): notice " (ATTAIN-GOAL A ) " is ass have been a s s e r t e d , eith  S  8  This an  is  a  free  i l l u s t r a t i o n  G  do  C  do  H! If D If K If P  N!  H and  coding logical ' particula erted as er declar  interpretation in  128  logic.  cf 5  the  8  D G!  If  and  T h e o r y  interpretation  instrumental  If If If  Accounting  Mattessich  R  and  Q  and  do do do  K 8 L! P 8 Q! R !  L!  this appears as shewn belcw: and' and 'THGOAL to 'Do* in rly that where the pattern a goal any pattern might ative cr imperative:  into  a  (1972e,  commercial p . 16)  demain  of  129  (THCONSE IF-A! ( ) (THAND (THGOAI ( A T T A I N - G C A L (THGOAL (ATTAIN-GOAL (THGOAL (ATTAIN-GOAL (THCONSE IF-B! ( ) (THAND (THGOAL (ATTAIN-GOAL (THGOAL (ATTAIN-GOAL (THGOAL (ATTAIN-GOAL (THCONSE  (ATTAIN-GOAL  A)  B) $T) C) $T) E) $T) ) ) (ATTAIN-GOAL  E)  E) $T) F) $T) G) $T) ) )  IF-C! ( ) (THGOAL (ATTAIN-GOAL  (ATTAIN-GOAL H) $T) )  C)  (THCONSE IF-D! ( ) (THAND (THGOAL (ATTAIN-GOAL (THGOAL (ATTAIN-GCAL  (ATTAIN-GOAL  D)  K) $T) I ) $T) ) )  (THCONSE IF-K! ( ) (ATTAIN-GOAL (THAND (THGOAL ( A T T A I N - G C A L P) $T) (THGOAL ( A T T A I N - G O A L Q) $T) ) )  K)  (THCONSE  IF-P! ( ) (THGOAL (ATTAIN-GCAL  (ATTAIN-GOAL B) $ T ) )  P)  (THCONSE  IF-E! ( ) (THGOAL (ATTAIN-GOAL  (ATTAIN-GOAL M) $T) )  E)  (THCONSE  IF-H! ( ) (THGOAL (ATTAIN-GOAL  (ATTAIN-GOAL 0 ) $T) )  M)  (THCONSE  IF-G! ( ) (THGOAL (ATTAIN-GOAL  (ATTAIN-GOAL N) $T) )  G)  (THCONSE IF- 0 ! (THDO (PB (THCONSE I F-F! (THDO (PB (THCONSE I F -H! (THDO (PB (THCONSE I F -L ! (THDO (PB (THCONSE I F -B! (THDO (PB (THCONSE I F -Q! (THDO (PB (THCONSE I F -N!  (  )  "BASIC (  )  "BASIC (  )  "EASIC (  )  "EASIC (  )  "BASIC (  )  "BASIC (  )  (ATTAIN-GOAL IS NOW BEING (ATTAIN-GOAL BEING GOAL F IS NCW (ATTAIN-GOAL GOAL fl IS NCW E E I N G (ATTAIN-GOAL GOAL L IS NCW E E I N G (ATTAIN-GOAL GOAL B IS NCW E E I N G (ATTAIN-GOAL GOAL Q IS NCW E E I N G (ATTAIN-GOAL GOAL  0  0)  ACHIEVED!) F) ACHIEVED!) H) ACHIEVED!) L) ACHIEVED!) R) ACHIEVED! ) Q) ACHIEVED! ) N)  130  The statements line  by  i s  a trace  included that  under  here  a  which  because  the goal  the  output  of the goal  trace  printout  the process  i s  t h e MICRO-PLANNER  i s our a s s e r t i o n  remainder Such  following  listing  would  t o be a t t a i n e d of the prcof  not normally  c a n be  may made  running  interpreter.  of the s t r i k i n g  of prcof  from  also  The  these first  (Ec A ! ) , t h e  as i t i s found.  be r e q u e s t e d .  Iti s  way  i n which  i t reveals  return  the plan  of  actions  true.  I  T  0  > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >  (THGOAL  ( A T T A I N - G O A L A)  $T)  ; <==  ENTER THM I F - A ! ENTER THM I F - B ! ENTER THM I F - E ! ENTER THM I F - M ! ENTER THM I F - 0 ! B A S I C GOAL 0 I S NOW B E I N G A C H I E V E D ! I F - 0 ! S U C C E E D E D THNOVAL I F - M ! S U C C E E D E D THNOVAL I F - E ! S U C C E E D E D THNOVAL ENTER THM I F - F ! B A S I C GOAL F I S NOW B E I N G A C H I E V E D ! I F - F ! S U C C E E D E D THNOVAL ENTER THM I F - G ! ENTER THM I F - B ! B A S I C GOAL N I S NOW E E I N G A C H I E V E D ! I F - B ! S U C C E E D E D THNOVAL I F - G ! S U C C E E D E D THNOVAL I F - B ! S U C C E E D E D THNOVAL ENTER THM I E - C ! ENTER THM I F - H ! B A S I C GOAL H I S NOW B E I N G A C H I E V E D ! I P - H ! S U C C E E D E D THNOVAL I F - C ! S U C C E E D E D THNOVAL ENTER THM I F - D ! ENTER THM I F - K ! ENTER THM I F - P ! ENTER THM I F - R ! B A S I C GOAL R I S NOW B E I N G A C H I E V E D ! I F - R ! S U C C E E D E D THNOVAL I F - P ! S U C C E E D E D THNOVAL ENTER THM I F - Q ! B A S I C GOAL Q I S NOW B E I N G A C H I E V E D ! I F - Q ! S U C C E E D E D THNOVAL I F - K ! S U C C E E D E D THNOVAL  OUR  COMMAND  |> |> |> |> |> |>  10 10  E N T E R THM I F - 1 ! E A S I C GOAL L IS NOW B E I N G A C H I E V E D ! IF-L! SUCCEEDED THNCVAL IF-D! SUCCEEDED THNOVAL IF-A! SUCCEEDED THNCVAL ( A T T A I N - G O A L A) ; <== THE  We  may  interpret  this  notation  GOAL  IS  with  I I I I I |  ATTAINED  the  I I  following  model: Let Make  A B C D E  = =  F G  =  N H  = =  Advertise! Establish a distribution system! Maintain a f l e e t cf trucks! Survey ruling exit market prices! •= U s e c o s t - e f f e c t i v e machinery! Survey r u l i n g entry market prices Employ cheap labour! Buy c o s t - e f f i c i e n t machinery! Service eguipment regularly!  K  = =  Q  If  A!")  hand!  =  R L P  are  on  Manufacture for stock! Possess functioning eguipment!  M 0  this  p r o f i t s !  S e l l goods! Set prices high! Keep c o s t s low! Have s u f f i c i e n t i n v e n t o r i e s  the  0,  F,  inference primitively is  returned  true a  are  H,  in  R,  the  e f f e c t i b l e  (or,  plan  N,  by  in  this  which  Q  and  data  base  actions case,  the  I  which as  then  we  true the  may  Intelligence,  reached  by  prepositions  or  theorem  attainable).  imperative  have  The be  A  (cr  "Be  theorem  has  achieved.  knowledge  and  Al.  132  The  interpreted  conclusion  |  "Possess  maintain  |  | | |  a f l e e t of trucks 8 survey ruling e x i t market prices 8 and survey ruling entry market p r i c e s 8 s e r v i c e eguipment r e g u l a r l y 8 and employ cheap l a b c u r ! "  | | |  Of terms.  functioning  i s :  course, the There a r e nc  conjunction  is  i l l u s t r a t e  how  into  1J.6J.3  such  Al  THE  used.  i t s e l f ,  languages  on  commercial  systems.  f r u i t f u l  in  this  advertise  example  gcal  8  BICBG-PLANNEB.  OF  VERY  HIGH  over  computing a  LEVEL the  Al  serve  may  years  search  be  has  tc  encoded  so  had  (list  fcr  and  language  logical and only  L A N GU A GJES^  languages  expressions  Current  w i l l  structures  as  processing,  conditional  this  such  research  general  string  recursion,  very  However,  SIGNIFICANCE  influences  S  i l l u s t r a t i o n is t r i v i a l in variables, f r e e or o t h e r w i s e ,  readily  AI-inspired  machinery  many  processing parallelism,  on)  research  and  thus  will  on  prcve  regard.  Intelligence,  knowledge  and  A l .  133  Bobrow importance high  of  level  S  Raphael  these  describe  languages  when  exactly  they  c u r views  describe  them  cn the as  very  languages.  "The e x t e n t t c which a programmer may specify what he wants accomplished without d e t a i l i n g how i t is t o be done i s o n e way o f d e f i n i n g t h e ' l e v e l ' c r 'power' o f a programming language. F c r example, the lowest level computer languages, assembly codes, require an e x p l i c i t statement f c r each wired-in instruction to be executed. Algebraic languages, such as FORTRAN, permit the user t c describe a desired result by a combination of mathematical relations, e.g. X=A+B-C, and l e a v e t c t h e c o m p i l e r decisions about the order in which elementary commands are performed. LISP programs may b e recursive i n which case the system has added responsibility f o r stack manipulation—and possibly translating t h e r e c u r s i o n i n t o an i t e r a t i o n .  permi a c t i v decid const each In  T h e new l a n g u a g e s g o a step further. They t the programming system tc carry out certain i t i e s , including modifying the data base and ing which subroutines to run next, using only raints and guidelines t h e programmer sets up f c r programmed a c t i v i t y . " Chapter  analogously. Al  languages  3  we s h a l l  We s u g g e s t are to  that  compare t h e IMIS  conventional  t h e MIS i s  computer  Intelligence,  to  and t h e MIS  the  IMIS  as  these  languages.  knowledge  and A l .  134  |  Chapter  |  2:  and  2.1  The  2.2  'Theory':  2.2.1  2.3.1  2.4  poverty  A  methodology  of  the  Testing for  2.3  The  axiomatized  of  the  accounting  theory  in  instrumental  accountancy  hypothetico-deductive  theories  for  truth  sciences  | I  theory.  and  MIS  method.  and  instrumental  theories  u t i l i t y .  theory  as  Analytic  the  definition  definitions  2.3.1.1  Sprouse  2.3.1.2  Churchman  2.3.1.3  Chambers  2.3.1.4  I j i r i  2.3.1.5  Mattessich  2.3.1.6  Mattessich's  Implementation 2.4.1  Nested  2.4.2  Test  and  of  of  its  label.  'accounting'  Moonitz  and  'MIS'.  (1961,1962)  (1961) (1966)  (1967)  as  (1964  a  MIS test  theories  signals  as  et  seg.)  theory of  and  an  (1972) instrumental  theory.  domains.  surrogate  domains  fcr  I j i r i ' s  theory.  accounting  theory  S  methodology.  IMISj. I n t e l l i g e n t  Our  objective  MIS and Accounting  in  this  chapter  T h e o r y 1 3 5  is  S S t h o d o l o g i c a l l y sound tests cf theories of MIS.  However,  because  an  appreciation  method and i t s terminology i s omitted traditional design formal method. and  of  In section  these matters  should  and  whom  rule  MIS  value f o r testing  statements  the  courses i n system outlining  the  we  Mattessich  quote.  tc o r i g i n a l i t y and  to  Eunge  Readers f a m i l i a r  section  theories  as  instrumental theories and  of t h i s concept.  2.3.  We  We  with  to  shall  regard  stress  an instrumental theory cf i n c l u d i n g i n  presentation e x p l i c i t and  from  proceed  explain our understanding  domain  a  2,1 below we make no claim  acknowledge our heavy debt to  accounting  and  'theory* and the value cf the s c i e n t i f i c  from  its  as  we must lay some groundwork by b r i e f l y  (1967, vol. 1) ,  the  accountancy  of the s c i e n t i f i c  accounting curriculum and from  meaning  tc construct  statements  cf  what  we  call i t s  range and i n addition f c r an instrumental theory  of the purpose of the theory and the standards  which i t s e f f i c i e n c y  by  i s to be judged.  Accounting  fhecry 6 methodology.  IMIS! I n t e l l i g e n t  2il  HIS  and  Accounting  2UI 'POVERTY OF THEORY IN ACCOUNTING AND  Theory..  136  MIS..  For centuries accountancy has been a i r c f e s s i o n i n i t s service  role  so-called  and  "arts"  an  art  in  accountancy  science^ that i s , a technology method  as  pure  practical.  accounting become than  science  However,  consciousness  is  for  ends  then and  which have  do  our  present  understanding  nineteenth century but " p r i o r  as a well-defined body of l o g i c a l precede accounting p r a c t i c e " accountancy  is  still  store of conventional  SiS£2£Stign even  more  of  a  information  to  would  system management.  the  as  the  accountant's  any  necessarily  we  of  fifteenth  tc  tc 1930, accounting reasoning  may  foundations  did  net  theory usually  1965, p.15).  Today  more as the a p p l i c a t i o n wisdom  as  of a the  This c h a r a c t e r i z a t i o n i s  embryonic MIS  than  profession  design and  a d d i t i o n , lacks the t r a d i t i o n s of centuries and narrow  if  subject rather  from the  traditional  the  ultimately  With hindsight of  body of theory..  applicable  applied  not appreciate that i f  (Hendrikscn,  practiced or  are  skeleton cf t h e i r  accountancy into the works of authors the  an  little  'theory*  a d e s c r i p t i v e derivation of i t .  read  Like most  which employs the same general  cf t h i s category and  foundation  methodology.  potentially  accountants  were a science  the  its  focus  upon  of  operation, i n a  focus  income and  as  wealth  measurement.  Accounting  theory S methodology.  IHISj.  Yet  the  accountancy: which these  are  there  rather  (that  been  a  i s ,  p r a c t i t i o n e r  practice, rather  which  than  is  One  of  in  theory'.  each  of  boundaries each  boundaries  -  a  i l l - d e f i n e d  sense  set and  which  not  to  tc  to  least  this  weak  At  a  the  statements  Accounting  cut  of  practice l e v e l  we  management',  f e l t  to  instrumental  seen  practice,  same  of  than  have  is  is  theory',  is  to  a  accounting  a  f e l t  is  within  'theories  set  and  areas  is  is  it  structure  there  which  as  be  there  'theory'  c r i t i c i s e  ' f i n a n c i a l  i t  that  about  categorize  ,  might  theory  fcrm  about  from  profession  disccurses  or  Since  induced  'Theory'  Presumably  each  to  the  the  describe  of  theory'.  surprising  upen  in  prepositions  practice  systematic  finance'  accounting  related  In  which  speak  'cost  closely  is  theory  attempts may  by  137  used  of  'accounting  practice.  business  these  i t  some  applied  body  codifications  At  that  Theory..  currently  'non-practice'.  is  applied  accountancy^.  of  induce  semi-formal  'theories  to  to  as  conjectures  concepts^  by  is  mainly  science. of  Accounting  considerable  taught  impact  eguivalent  seeks  a  hypotheses),  often  have  and are  minimal  Systematication  which  and  'theory'  than  potential  the  MIS  exists  propositions  guided  the  term  regarded  practice,  has  I n t e l l i g e n t  of  that  within  statements  be  within ' t r u e '  more ether  in  seme  value.  theory  8  methodclcgy.  IMIS.:  I n t e l l i g e n t  However, accountancy not  (i)  been  the  (and  lack  of  the  method Western  the top  the  and  speak  of  are  defined  conscious  s c i e n t i f i c  which  has  aware in  Accounting  a  that  our  poverty so  method^ the  theory  in  •accountancy'  has  discussion)  application  become  far  to  the  138  Theory.,  cf  we  mean:  accounting  practice  hypotheticc-deductive  dominating  mental  set  of  from  the  c i v i l i z a t i o n ,  f a i l u r e down,  subordinate  (iii)  we  we  formally  of  (ii)  when  MIS  lack  to to  apply the  areas  of  this core  within  an  method of  the  rigorously  accountancy subject  embracing,  and  rather  than  tc  and  unifying  d e f i n i t i o n  of  accountancy^  Accounting  thecry  S  methodology.  IHIsLi  ls.2  1THEORY2.1 He  to  wholly  our  In  see  see  are  the  present  construction  and  Accounting  that  these  are  the  level  cf  short,  attained  prove  MIS  M£OTHETICO-DEDDCTIVE  in  characteristic  should what  THE  shall  shortcoming. wish  Intelligent  to  be  accountancy  of  the  possible.  characteristics context and  we  of  w i l l  not  testing'  as  three  go the  faces  139  A  cf  theory  which  and  is  Why  this  J  METHOEJ.  MIS  s c i e n t i f i c 1  T h e c r  is  method,  ' s c i e n t i f i c  wrong  we  that  this  i f  we  essence  cf  the  same would  which if  is  that  desirable  and  method'?  In  take the  'theory s c i e n t i f i c  method,  The  GOALS  of  s c i e n t i f i c  theory  construction  " (i)  to systematize knowledge by establishing logical relations among previously disconnected items. In particular, to explain empirical generalizations by d e r i v i n g them from h i g h e r - l e v e l hypotheses. facts by means of propositions that  are:  (ii)  to explain entailing the concerned.  systems express  cf  hypotheses the facts  (iii)  to increase knowledge (e.g. p r e d i c t i o n s ! from the relevant information,  (iv)  to enhance the t e s t a b i l i t y of the hypotheses, subjecting them t o t h e c o n t r o l o f e t h e r h y p o t h e s e s of system," ( B u n g e , p. 383)  by d e r i v i n g new prepositions premises in conjunction with  by the  If i t be not possible, we w i l l s t i l l h a v e g a i n e d the knowledge o f the manner and d e g r e e to which Accountancy or Information System Managing i s net a s c i e n c e .  Accounting  theory  S  methodology.  I^IS!  Bunge  I n t e l l i g e n t  contrasts  MIS  the  and  goals  accounting  of  science  Theory,.  with  140  these  of  psuedo-science:  "By  contrast  the  n o n - s c i e n t i f i c  speculations  abcut  r e a l i t y (i)  do n o t u s u a l l y ask proper guestions, but r a t h e r problems with false or untestable presuppositions, s u c h as 'How a n d when was t h e u n i v e r s e created?';  (ii)  t h e y do not propose hypotheses and procedures both grounded and checkable but o f f e r groundless and usually untestable theses (e.g. revelation);  (ii)  t h e y do n o t d e s i g n o b j e c t i v e t e s t s cf of their alleged sources of knowledge authority;  (iv)  they accordingly i l l u s t r a t i o n s of purposes r a t h e r than eagerness with which away;  (v)  they give r i s e t o no new p r o b l e m s their whole point being t o put an end t o e n g u i r y by p r o v i d i n g a r e a d y made set of answers to every possible cr permitted question."  their theses but r e s o r t tc  and seme  . . . remain content with finding their conceptions fcr persuasion f o r the sake of t e s t a s s h o w n by the they explain every negative evidence  (Ibid,p.29)  accounting  theory  8  methodology.  IMISj.  The  NATURE  (i)  A  Intelligent  of  a  is  a  system  hypotheses  interrelationships understanding  of  propositions formal  (ii)  A  Accounting  Theory..  1 4 1  are  of  propositions  (among  are  conspicuous)  and  interpretations  some  domain  commonly  whose  cf  expressed  which  law  logical  encode  our  knowledge. as  The  sentences  in  some  formal  (or  language.  theory  may  'analytic')  ha  formal  theory  determined the  and  theory:  theory  formula  Mis  by  and  symbols  at  value  which  the  only  factual.. truth  by  the  employed,  interpretation. theory  or  In  least  one  is  a  In  value  cf  logical  without  factual  a  the  in  recourse  determinable  the  only  is  relationships  ('synthetic'  proposition  theory  or  their  'empirical)  thecry by  tc  of  has  a  truth  recourse  to  factual  are  experience.  (iii)  A l l  theories,  whether  hy22thetico^deductive  formal  systems"  cr  p.226)  (Ibid,  that  is  tc  say  (a) must  be  At an  reference  one  hypothesis^,  assumption.. statement  least  In (if  to  a  a a  of  a  fact  statements  statement  formal then  the  theory  which  this  b);  in  a  which  is  not  may  factual yet  Accounting  in is  be  a  a  theory  the  6  logical  conditional it  experienced  thecry  thecry  may and  be  a is  methodclcgy.  112152  I n t e l l i g e n t  therefore  capable  (b)  The  well-formed sought the of  The  by  other  FORM  of  applying  a  •statements',  and  Accounting  Theory,.  142  refutation.  truth  value  cf  every  in  the  the  i n f e r e n t i a l  statement  statements  The  themselves  of  formula  unproven  MIS  in  statement  theoretical  and  the  the  system,  which  system  rules  cf  established  may  logic, truth  including  is  its  a be  given values  axioms,  theory:  components 'axioms', be  of  the  'rules  systematized  theoretical  of  inference'  (Ibid,  system, and  pp,226,  so  these cn  487,490)  may as  follows: 1.  Foundations. 1.1 1.2  1.3 1.4 2.  presuppositions, Primitive Frame. 1.2.1 Primitives. 1.2.2 Formation Rules, 1.2.3 Transformation rules. axioms. Interpretation assumptions.  Superstructure. 2.1 2.2 2.3  Definitions. Theorems. Pragmatic  Conventions.  accounting  theory  6  methodology.  IMIS,:  I n t e l l i g e n t  MIS  and  Accounting  143  Theory,.  £ £ § § u £ £ 2 § i t ionsj. "The exhibited  background in  assumptions  of  presupposes for  the  meaning  is  used  or  a  background order  (Ibid,  are only  theory  The  t a c i t  its  and  i f  or  verisimilitude  set  A of  is  a  E  and  theory  logic  and  usually  unguestioned  logical  . . .  necessary A  will  B  condition  guesticned  with  This  the  net  while  presuppose  (propositional  calculus  theory. others,  a  is  presuppositions.  (1)  Commonly  "elementary  among  . . .  i f  functional  elementary provides,  idea  tested." as  a  entirety.  an  A  B  f i r s t  i t s  of  such  calculus  and  i d e n t i t y ) ,  body  of  signs  cf  and  presuppositions the  t h e c r y . . . "  p.483)  £ £ i l i l i v e s 2 The stock  of  a l l  already  i s the  special  contained  undefined remarks  primitives  may  theory  theory  is  provided  in  but  they  of  do  desiderata  are  " ( i ) . . .  Their  not  (in  range to  of  thecry,  ether  meaning  below)  and  correspondence that  maximum  of  number  Accounting  and  the cf  these  They  are  pedagogical meanings  primitives proper  u n t i l  r e f e r i t i o n s i . e .  is  the  tc  possible  primitives,  is  the  addition  parenthetic  acquire  (see  the  primitives  used  relationship  symbol-to-factual-object choice  be  cf  presuppositions).  them.  their  alphabet  though  i n t e r f e r e d for  tc  the  even  accompany by  the  symbols  symbols,  constrained  are  in the are  u n t i l  a  established.  The  t h e c r i z e r  the  possible  thecry  6  but  relations  methodology.  IMISj_  with  other  from  Intelligent  concepts..;  immediate  with  a  high  greatest source number  that  derivative  These  primitives  respectively ' (Dr)  Cash  greatest  order  that  generality of  Theory..  possible  high  level  144  distance formulas  c a n be c o n s t r u c t e d ;  reference  may a c c o u n t  tc  f c r the (Ibid,  ( i i i )  fundamental largest  or  possible  p.490)  Rules^ rules  Formation  Accounting  properties."  prescribe which  well-formed-formulas the  i n  the sense  properties  Formatron  of  in  cf  and  ( i i ) . . .  experience, degree  depth,  of  HIS  the 1000,  are  (wffs)  Rules  of  the syntax  of  to  algebra,  1000  these  be  Cash'  For example,  Fortran '  a=b++ are  Accounting  not  compositions  regarded  the theory.  statements (Dr)  of  and  ' ,  ,  as by  Accountancy  4=AIPHA*  and  admissible.  theory  6  methodclcgy.  IMIS;.  Intelligent  Transformation These may  validly  rules be  from  ^IS  and  Accounting  145  Theory..  Rules.: prescribe  derived  •a-b=a-c  from we  1  the other  may  syntax  of  given  derive  these  wffs.  wffs  For  both  '-b=-c'  Cash  1000 '  which  example,  and  'fc=c*.  the  wffs  From •IF we  may  ( A ) 2 1 ,31 , 2 1 »  derive  • I F (A.EQ.O) GO and  TO  from •(Dr)  we  G C T O 31  21 •  may  possibly  These  rules  independent Many w i l l  of be  derive  are of  the  Cash  the  1000, •  '.  dependent  cn  meaning  values  Formation  i m p l i c i t ,  (Cr)  in  the  or  and  the  syntax of  the  Transformation  cf  symbols Rules  in  and  in  them.  a  theory  presuppositions.  Accounting  thecry  S  methodology.  IMIS!  I n t e l l i g e n t  HIS  and  Accounting  Theory^  146  Axioms! Axioms  are  Il2£2theses  which  to  of  members  truth  w i l l  theorizer  the are  the not  Far  basic  class be  should  formulas.  unproven to  of  i n i t i a l the  theory  from is  but  they  presuppositions  questioned.  seek  assumptions.  In  precise being  and  as  commonly  h i g h - l e v e l  generalizations.  that  the  supposed,  their  axioms  They  truths  axioms may  are  similar  the  comparatively  ' s e l f - e v i d e n t '  experience,  are  in  choosing  They  rich  close  should  be  tc be  far  from  self-evident.  Interpretation If given the i t s  a  no  meaning  thecry  is  s p e c i f i c given  l e f t  of  therefore theory  them  the  theory.  be  a  have  s t r i c t l y by  formal a  uninterpreted  real-world  presuppositions)  symbols  that  assumptions!  truth  A  cr  l i n g u i s t i c  within  t h e i r  the  the  primitives meanings.  thecry  relationships  uninterpreted  theory.  Conclusions but,  of  Only  (including  with  s t r i c t l y  value  are  the  ether  theory  must  deduced  course,  within  s t i l l  nc  interpretation.  Accounting  thecry  S  methodology.  IMIS:.  An symbol,  w i l l  or  f a l s i t y  depend  to  the  on  give  which  acceptance  <account1>  a  a  way  147  meaning  that  the  containing  or  (Ibid,  <v>,  f c r  a  factual  the  rejection  p. 151)  upon  symbol  of  the  example,  i f  we  (Cr)  <acccunt2>  is  as  meaning  to  the  f i r s t  formula  a  truth  value.  i t s e l f  a  measure  assumptions  give  meaning.  it  one  another.  Of  uninterpreted  <v>»  assumption  interpretation  might  well  have  single  symbol,  ordinary  language,  such been  the  few  theories:  contains  to  of  monetary  primitives  'economic implies  symbol  own  second  another  proposition  as  of  as  will  Accounting  give  cf will it as  accounting,  fcr  rescurce'.  It  'x*.  being  But  a  interpretation  This i n t e r p r e t a t i o n assumption may established or r e j e c t e d i n another related.  set  expressed  'economic  such  one theory  set  are  thecry  resource', its  a  Adjoining  2  theories  I j i r i ' s  a  via  value'  uninterpreted  Adjoining  course,  example,  2  such  <v>  has  as  in  Theory..  "confers  expressions  postulate."  interpretive  '  we  but  the  the  Accounting  formula  ' (Dr)  the  of  and  assumption  conventionally  interpretation add  HIS  interpretation  not  truth  I n t e l l i g e n t  phrase  this cf  assumption  be a p r o p o s i t i o n t c be thecry, separate but  thecry  8  methodology.  113112  (a  I n t e l l i g e n t  presupposition,  possible  future  poorly  chosen  uninterpreted the  in  reinterpretation symbol.  uninterpreted  primitive  Thecry..  and  cculd  of  but  and  a  the  whatever  wider  we  theory  the  is  unobvious  if  fewer  i t  order  theory,  axiom, range  of  148  envisage would  theories  like  as  to  the  bcth  a  invention  pricr  has  a  te  end  of  l o g i c a l l y  uninterpreted  the  benefit  Accounting  Even  theory  The  and  potential  effect)  theories;  interpreted.  and  MIS  the  the deep  greater  opportunities  tc  be  tested.  Definitions: A i t of  definition  commonly a  is  new  written  ordinary  symbol  correspondence example,  in  to in  within  in  with  afford Fortran  appended  established  to  correspondence  will  is  not  language,  i t  is  by  a  theory  symbols  upward II,  already  compatibility  the  nominal  l i s t '  Fortran  symbols  theory  the  " ' W R I T E (6,N)  i s  a  IV.  =df.  If  the  give  meaning  to  the  statement  for  as  establishment cf  established.  'PRIST  form  description,  computer  its For  programs  definition  meaning  then  a  K,  has  of  the  the  new  Accounting  l i s t " '  been  given  tc  the  symbol-to-symbol symbol  thecry  S  also.  methodclcgy.  IMISj.  I n t e l l i g e n t  Symbols are  new  hypotheses.  be  derived  the  theory  so-called we  w i l l  than  They  are  would  "It  is  cf  the  a  they  A  3  i e f e r i t i g n  we  perform is  the more  confusing  understanding, (Ibid,  such  p.151).  ' d e f i n i t i o n '  of  are  may  convenient  d e f i n i t i o n '  philosophically  ostensive  basic;  a  Theory.  d e f i n i t i o n  symbcl-tc-object  operations:..."  muddle".  by  Accounting  unwieldy.  from a  i f  interpretation, single  be  'operational  measure  accounting  for  rather  distinguished  establishment  and  established  they  which  MIS  149  dispensible;  net  introduce  shorthand  without  d e f i n i t i o n  should  which  is  the  correspondence.  A  x  as  "'Income'  following  what  seguence  properly  a  because  of  r e f e r i t i c n . it  d e f i n i t i o n  and  Hence  e a r l i e r  our  is  test  drags to  a  disdain  i n t e l l i g e n c e .  Any m e t a l i n g u i s t i c statement in Eackus-Naur Eorm cf the allowable syntax of a programming language is an e x c e l l e n t example of a cascading set cf definitions (usually recursive d e f i n i t i o n s ) . A relevant example is p r e s e n t e d i n A p p e n d i x 4. Its own t y p i c a l u n w i e l d i n e s s shews what the reader is spared in the use c f a t h e c r y by the device of ' d e f i n i t i o n s ' . 3  Accounting  thecry  8  methodology.  IMIS.:  I n t e l l i g e n t  MIS  ana  Accounting  Theory..  150  Theorems_: Theorems are  aerivea  theories by  set,  the  the  under  propositions  A,  fina  T  that  in  A.  be  one  entailment become  the  members A  given  least  some  whose  truth  value  truth  we is  more  thecry  value  may  come  tc  false  in  fact  may  be  one  missing  or  tc  and  cf  to of  thecry  of in in  the  may in  theory  we  must  false  A  of  which  be  theorems  that  proven  a  is, that  is  thus  theorem  another.  r e a l i t y .  believe then  set,  they  That  members  that  the  that  deduction.  the  axiom  in  previously  sentence an  more  a  ana  rules  operator  at  This or  of  air  axioms  a  their  sentences  cf  cf  theory  against  aynamic  set  deduction..  formulation  f a c t u a l  a  application  conclusions  closed one  from  by  applying  lena  in  In  a  will  te  factual  can  be  testea  Given  a  entails  T.  find  the  propositions  set  of  If  we  deficiency cr  i t  may  propositions.  Accounting  thecry  S  methodology.  1115!  Intelligent  Pragmatic The  propositions  propositions  New  Zealand  of  Accounting  151  Theory..  which  ensure  consistency  Testing  a  in  in  are  as  a  sat  true  of  in  units  truth  cf  and  the  system.  Canadian  as  conventions and  their  sentences  for  true  Foct-Peund-Seccnd  (through  the  as  pragmatic  theory  theories  are  the  The  conventions  in or  l i n g u i s t i c  presuppositions)  the  theory.  instrumental  theories  utility..  So i s  in  dollars. to  for  physics  accountancy  adjoin  2.2.1  cf as  assumptions  semantic  and  Conventions!  Centimeter-Gram-Second The  HIS  far  possible  then  may  we  seek  f i r s t  Chapter  1  whether  a  includes attempt  to  we to  have cast  judge  described nonsense  the  to  know  i t  is  worth  whether not  the  in  of  a  they  given  the  are to  is  true;  rules  -  methodological the  truth  form  theory?  b e l i e f  ensure  form  our  theories.  of  a  For  true.  us  therefore  of  cf  to  thecry.  -  How  theories  we  As  we  in  with  saw  certainty  s c i e n t i f i c  advices  *  it  a l l  know  the  Eut  by  which  method we  may  theories.  See f o r e x a m p l e , 'the axiomatic thecry of i t presented by Eunge, (op.cit.,p.477).  Accounting  thecry  phantoms',  8  as  methodology.  IMISj.  "What (i)  to  (ii)  be  to  (iii)  science  truer  be  to  Intelligent  to  able  to  and  claims  than, any  able  be  MIS  test  accounting  Theory^.  152  is  n c n s c i e n t i f i c such  a  discover  truth  i t s  model  cf  the  world,  claim,  own  shortcomings  and,  (iv) t o be a b l e t o c o r r e c t i t s own s h o r t c o m i n g s ^ i.e. be able to b u i l d more and more a d e g u a t e p a r t i a l mappings the p a t t e r n s of the world. No e x t r a s c i e n t i f i c s p e c u l a t i o n as modest and y i e l d s as much." (Ibid,p.29)  tc cf is  How  choice  of  Theories  are  is  this  accomplished?  hypotheses constructs  and of  by  theory  may  other  related  may  arbitrary  be  i t s e l f  and  custom  rather  acceptable " (i)  as the  correct) some (ii) on  usually  be  restated  the  that  hypotheses in  the  theories  have  or  of  the  a  hypothesis  H of  thecry  thecry  T'  -  matters  axioms  terminology  that  cf  dividing T,  a cf  lines  thecry  convenience  T  and  necessity. mere  fancy  hypothesis  previous  them.  appreciate  hypothesis.  It  must  meaningful  s c i e n t i f i c  the  we  between  no  and  r e c a l l  that  than  an  we  also  larger  However,  If  testing  and  theories  some  controlling  rigorously  hypotheses.  •presuppositions* given  By  cr  wild  guess  is  required  be  w e l l - f o r med  (semantically  will  te  that (formally  non-empty)  in  context;  hypothesis knowledge;  must i f  be  grounded  e n t i r e l y  Accounting  tc  novel  thecry  some it  5  extent  must  be  methodology.  153  IMISj. I n t e l l i g e n t MIS ana A c c o u n t i n g Thecry.  compatible with the bulk of s c i e n t i f i c knowledge.  s  ( i i i ) the hypothesis must be empirically testa tie by the objective procedures of science, i . e . by confrontation with empirical data controlled i n turn by s c i e n t i f i c techniques and theories." (Ibid, p.229)  Thus our hypotheses are guesses. guesses  hope  they  are gccd  but in the course of our s c i e n t i f i c method we s t r i v e  to refute them  x  d i r e c t l y i f possible, i n d i r e c t l y i f not.  Because of the importance recommends  that  cf exposure  at  provisional sufficiently  refutation (that  to  refutation  Pepper  a good hypothesis i s also a * t c l d ' one, cne  which risks much to gain much. attempts  We  we  If hypotheses may  come  i s , contingent)  general  we  may  survive our  tc ' regard  truths.  the tr as  If they  c a l l them laws.  are  Although the  s c i e n t i f i c method offers no guarantee against error  i t dees  describe a process by which a closer and closer approximation to r e a l i t y may be attained.  Although the formal axioms of I j i r i ' s thecry cf accounting measurement do not include the axioms of utility theory he c l e a r l y regards that thecry as l o g i c a l l y prior to his own and as j u s t i f i c a t i o n for imputing to a resource gained the value of i t s cause that i s , the value of the u t i l i t y sacrificed. The argument the measurement theory i s e x p l i c i t l y grounded in the sociology of exchanges of economic resources which i s in turn grounded in u t i l i t y theory. 5  x  Accounting theory 8 methodology.  IMISj.  INSTRUMENTAL  I n t e l l i g e n t  hypotheses  Cognitive goal  of  making  HIS  and  theories  a  and  are  those to  knowledge.  Their  truth-value  instrumental  theory  may  e f f i c i e n c y ,  on  standard,  i t  Instrumental  theories^,  toward  ends,  but  for  our  are  also  is  that  such f i t  a  true they  of  may  is  a l l  154  the  relative cf  of  An  cn  its  tc  some  some  goal.  serve  as  of  means  knowledge  Hopefully,  concern may  cf  important.  furtherance  theory  the  p r i n c i p a l  structure  suggests,  primary  in  a  additionally  which,  theory T2  growing  name  Any  theory  a  u t i l i t a r i a n .  the  a  serve  attainment  be  more  example  the  to  their  but  which  judged  the  effective,.  For  role  as  are  theories be  in  which  purposes  role.  the  ends  be  extent  assists  Thecry..  theories:.  contribution  the  Accounting  cf  come  following  their  to  measurement  they  te  users  used  would  in  amply  quotation  from  Bunge: "There a evidence (usually relevant cases a t h e o r y T1 somehow planning  r e no neutral experiences in science: every is produced in the light of some theory a set cf fragments of theories) and is t o some t h e c r y o r o t h e r . In the best cf datum can be n e u t r a l with r e s p e c t to the that is being tested, but then it will be related to a second t h e o r y , T2, used in the observation, designing the instruments,  or interpreting the results of observation. T2, a substantive theory in ether ccntexts, plays new an Instrumental r o l e ^ . . (Ibid, p. 502)  Here  the  secondary ' c o g n i t i v e  emphasis theory, 1  instrumental contrast,  is T2,  role. theory theories  instrumental  whereas  That is  the  OD  is  regarded of  the to as  primary  say, an  in  this  MIS  theory  cf  thecry  and  8  the  has  quotation  a n c i l l i a r y  accounting,  Accounting  role  theory.  a the In  Operations  methodology.  IMIS.:  I n t e l l i g e n t  Research,  among  situation  may  to  a  small  ethers, arise  set  manipulation  of  user.  case  In  our  manager  -  (1972e,1975  make  T2 the  and  others)  profits!")  many  are  become  formal  premises  Science'  that  a  to  the  are  theory  of  In  tc  cf  is  accountant  or  term  applied  Mattessich 'instrumental  "Hold  inventories  to  policy.  If  and  sciences  the the  coined  practice  and of  theory.  (e.g.  This shrinks  attention  the  the  T1, T2  instrumental  such  rules  must  theories.  methodological  controls  focus  user  rules  become  thecry,  relative  the  155  Theory..  instrumental..  dominating  administrative  areas  of  an  has  the cf  of  the  immediate of  these  Most  primarily  becomes  user  for  Accounting  propositions  the  hypotheses*  are  when  of  and  HI§  employed  i s  true.  this  an  instrumental  canons to  of  the  improve  connection  the  we  'Philosophy probability  wculd  observe  that:  •  •  In  its  use  relation  in  that  response  in  its  The  purpose  clearly  •  The  domain  should without  with  (in be  of  it  maps  thecry a  resembles  signal  in  its  a  mathematical  Jdomain^  into  a  J.rangej.i  an  or  in  this made  instrumental the  thecry  should  be  presented  theory.  sense)  of  e x p l i c i t  i f  any  instrumental  the  thecry  is  Accounting  thecry  8  tc  be  theory tested  confusion.  methodology.  IMIS,:  The  I n t e l l i g e n t  relation  between  HIS  the before  major  the  support  of  for  previous  an  the  and  theory  thecry  may  existing  unformalized  Accounting  domain  well-established function  and  practices  the is  be  set  Theory,.  of  156  range  may  conceived.  Thus  tc  logical  offer  conclusions^  into  be  a  formal,  a  tying logical  structure.  Because  an  there  should  (or the  means  instrumental be  f c r  presented  assessing  effectiveness  Bringing  of  their from  symbolic  i f  instrumental rules  of  we  in  error)  by  which  theory  in  our  in  the  real  wish  to  Mattessich  s t i l l  (1975),  and  not  assess goal.  into  the  reguire  systems  validly  must  imperative  are  that  other  goal  standards  attaining  tc  a  and  that  we  determine  way.,  may  we  mcve  from  normative  world.  reason we  thecry may  seems  no  attain  one  i n f e r e n t i a l  since  to  the  to  theories  framework  premises  modal,  unfortunately  See  into  truth-values,  Similarly  or  instrumental  actions  the  attempts  with  the  hypothetico-deductive admit  theory  employ deontic  f u l l y  Bescher  the l c g i c  i n f e r e n t i a l (rules  determined).  (1966)  and  and  E.  which  6  S.  Clarke  (1973).  Accounting  theory  S  methodology.  IMIS.:  The consider  I n t e l l i g e n t  two  the  (a)  l a t t e r  others  In  i t s  mathematical range  in  applied  ammeter "the  (via  a  longer  are  sets  in  common  of  data  in  cf  star  $4000  car"  an  The  although and  This  looking  of  of  viewpoint;  primarily p.416)  as  for  157  Chapter  as  1.  We  i t i t s  an  gain  (which  data the  now  as  range  a  on  have is  sets  usefulness  we  no  set  of  and  domain  no  members  closed of  under  at  least  domain.  that as  In  transaction  range  instrumental  recognizes  &  another  the  the  an  x°K".  $5000  into  in  a  proposition  is  cwn  its  along  flowing  say,  and  in  a  example,  distance  whcle  are  value For  need  a  resemble  a  crucis  theory  the  at  a  into,  "our  set  from  y i e l d  mapped  the  may  current  "we  be  formulae  wffs  way  map a  instrumental  value  in  domain..  Alpha  may  such  well-formed even  its  or  the  For  Theory..  thecry  will  hypotheses)  of  statements  i t  may  number  instantiated  matter  (1939,  that  signals  deduction).  valued  in  input  defined $1000".  a  system  discussed  instrumental  such  own  i s  were  theory  real  temperature  partly  an  to  a  Accounting  turn.  r e l a t i o n  into  accountancy  well  in  measurement  meter-rule  and  points  use  response  MIS  theory the  a  is  only  thecry  tool.  As  is Nagel  says:  Theories. . . "function primarily as a means for effecting transitions from one s e t o f s t a t e m e n t s tc other sets, with the i n t e n t of controlling natural changes and of s u p p l y i n g predictions capable of being checked through manipulating d i r e c t l y experienceatie subject matter. Accordingly in their actual use in science, t h e o r i e s serve as instruments in p a r t i c u l a r conte  xts  A  i  i  2!  Accounting  thecry  8  methodology.  IMIS.: I n t e l l i g e n t MIS and Accounting  Theory  158  But any theory i s suspect i f i t i s not constructed and used with sound methodology which i n turn i s valued it  i s our only guarantor  cf truth..  because  Fcr this reascn i t seems  to us that Nagel goes too far when he concludes: "...and i n t h i s capacity they are to be characterised as good or bad, e f f e c t i v e or i n e f f e c t i v e , rather than as true or false or p r o b a b l e ^ {emphasis added) (b) It i s most desirable instrumental which systems  theory  should  the instrument  that  the proponent  of an  specify the goal cr purpose for  i s designed.  In  discussing  robot  based on mechanical theorem prcvers we noted that i t  i s possible to present a desired goal to a theorem-prover in the form of a conclusion proposition (e.g. "Current maximized" wealth  may  proposition it  7  and be  have  the computer  maximized  to be true.  find a plan ty which  i n the ccurse  cf finding  the  In the s p e c i f i c a t i o n of the theory  should be made clear which set of goals i s admissible.  It should be made c l e a r , f c r example, that for an  7  wealth i s  Or the eguivalent imperative  8  accounting  "Maximize current wealth!"  We note that we find little d i s t i n c t i o n between our SSSSlSSiP.n propositions (the formulae in the so-called range) and the propositions which express the simple goals used i n our i l l u s t r a t i o n s . For example, the difference between (a) the goal Pu1! ("Produce 1 finished unit!" Or " I t i s desirable that 1 finished unit be produced") and (b) the proven conclusion, Pul ("1 finished unit has been produced") i s tco subtle for our machine, which reads into Pu1 i t s own i n b u i l t imperative "Find or make the conclusion to be true." Only an action-taking (i.e. instrumental) system could exercise such an i n b u i l t imperative. 8  Accounting  thecry S methodology.  °  I M S i  Intelligent  measurement not  theory  meaningful  system  (which  "Measure  must  current  However, included our  in  a  i t s as  a  own  garbage  thecry  comprise  Although  i t s  of  limits  location  of  in  Is  action the  gcal  the  be  is  expression the  i t  of  may the  passage  present which  as  theory  Far a  an from  thecry  purpose  in tc  purposes  be  input  are  also  and  may  in be  gucted.  not  success. is  cf  serve.  i t  be  d i f f i c u l t  set  to  may  thecry  very  r i c h  proposition  in  hypothesis  signals in  its  desirable the  in,  that  set  an  domain  "Garbage  9  specify  to  of  the wffs  domain.  learn  H.  the  and  system)  Mattessich's  may  instrumental  may  As  in  propositions  that  wealth!"  decision  purpose  theory  instrumental  we  theory. example  of  T1  current  measurement  i t  Therefore  an  is  ignorantly  expect  of  i t  the  theory  we  out".  proponent which  the  i s ,  a  token  imperative  If  cannot  as  "Maximize  159  Thecry.  acceptable.  I j i r i ' s ,  that  as  instrumental  a  Accounting  management  is  instrumental  right;  (c)  we  although  single  complex  "  expression  implementable being  a  subsume  wealth  of  an  gcal  fcr  theory,  revision  construct  the  whereas  and  MIS  A.  general order  those  something  Simon to  device find  abcut  (1969) of  cut  the  limits  observes driving more  a  this system  about  the  cf is  the an  beyond  nature  and  limits.  Accounting  thecry  S  methodology.  IMIS;,  Let  Example  "The  us  1:  employing  the  of  a  market  Example  Similarly,  change  in  Example  3:  economic been  the  is  money's  The  to  value  axioms  o r i g i n a l  s a c r i f i c e  in to  accounting  system  conventions  owned  system  w i l l  of  accounting value  is  $4000"  produce  a is  wff not  a  no  response  I j i r i ' s  i s ,  Section  accept  which  assumes  proposition  a  theory a  2.5  the  and  describing  Chapter  a for  Accounting  4  fcr  has  is below  input  reference that  that  measure  value  language  when value  assume  value  that  natural  f a i l s  quoting  no  system  a  relevant.  s a c r i f i c e d  theory  without  in is  That  when  system  any  invariant  assigned^  I j i r i ' s  The  car  160  examples;  conservative  the  Theory..  i t .  resources  Therefore  three  accounting  cf  domain.  Accounting  with  cost  value  to  money  this  h i s t o r i c a l  whatsoever  that  and  t r a d i t i o n a l l y  the  2:  Mis  i l l u s t r a t e  In  current  member  Intelligent  we is  a l l  already defined..  we  extend  find made  that tc  s a c r i f i c e .  thecry  S  methcdclcgy.  a  IMISj.  (d)  In  Sentences the  seme  In  a  generally  a l l  system  e x i s t s  of  accepted  pairs  by  theory (b)  the  i n f e r  (b)  as  that  focus  of  not  a  but  a  of  search  consistency accepted  by  information to  the  its  the  one  axioms,  wealth" true  we  of  with would  i t  tc  be  tc  Where the  cf  be  such  more  exists  an  is  tc  e a r l i e s t is  tested  (the  generally  that  here or  not  the  'conclusion  certainly  IMIS  gcal  "Current  there with  the  11,  wealth  thecry  which the  costs  cf  relevant reached  1  incorporating  is  than  8  be  tc  "maximize  successful  Accounting  of  foundations.  refer  i f  body  p a r t i c u l a r ,  are  the  made  in  these  Thus  the  conclusions  effectiveness  We  for  theories  closure;  desirable  the  cr  factually  is  data.  thecry  immediate  (b)  from  the there  tc the  and  conclusion  the  s a t i s f i e d  to  conclusions,  which  which  a  theorem-prover  desired.  actions  evaluate  of  and  goals..  to  goal  judge  may  although  extent  is  and  Judge  mechanical  input  (a)  The  tc  those  unsuspected  presumed  may  prior  to  as  role  which  i t  antecedent  data  each  and  is  conclusion  input  signals.  responses)  we  approaches set  the  analysis  the  (b)  consistency  meets of  new  cf  161  (a)  thecry  construction  addition  which  theory  the  set  exists, of  conclusion.  theory  for  In  e f f i c i e n c y  but  the  accounting  standards  prior  for  between  (e)  the  in  search  each  T h e c r y  set  responses  instrumental the  attention  implications  pairs  adjoining and  unequivocal  of  a  this  accounting  v a l i d i t y  obtained  accounting  an  sciences  there  ordered  of  Accounting  inaeea  x  accounting  role  the  aria  Accountancy  set  cognitive  MIS  instrumental  already  theory.  theory,  The  Intelligent  current $120,000" 12  which  methodology.  IMIS.:  made are  Intelligent  "Current so  obvious  wealth they  presuppositions.  MIS  is  may It  be is  that  standards  not  well  for  so  s a t i s f i e d wealth  is  wealth A2  the  a  incorporate This as  modify  suggests  theories  real  such  communicate conclusion,  or  which  that  as  record)  "The  theories  truly  premises  capable  us  a  although  both  A1  and  normative  x!"  11,  l a t t e r ' s in  cf  as  x  fact is Are  instigating  that  this  If  A2  A1  value  tc  11  A1  "Current  between  we  may  whese  and for  could  yield  premises  theories and  dees  "Current  standard cr  such  the  with  ' t r u e r '  such  instrumental  cf  Thecry.  chcose a  here  presented  A1  instrumental? of  be  intc  The  value  consider  axioms  world,  part  ccnclusicn  not  had  with  truly  A1.  could  we  -  wealth!"  the  told  i f  tc  standards  Measurement  current  we  Our  evaluation  proved  then  theories  between the  for  measurement  instrumental  distinguish may  its  interesting  A2  But  true.  162  Theory..  i m p l i c i t  Accounting  and  standard  wealth.  l e f t  "Measure  $110,000"  without  current  goal  $120,000"  is  an  Accounting  $110,000"  reguirement hold  and  A3.  be  cast  one  may  actions  guasi-instrumental  imperative, disguises  a  y".  Are  enly  only  imperative  "Measure  (or  declarative action-taking or  decntic  action?  Accounting  thecry  6  methodology.  IMISj.  Testing  a  theoryj:  The of  approaching  We  •  ever  The  theory  and  method to  closer one's  review  KIS  Accounting  163  Theory..  su^mary^  purports  constructs shall  a  s c i e n t i f i c  Science')  one  I n t e l l i g e n t  them  should  be to  (or a  the  be  broadly  in 1  'a  s e l f - c o r r e c t i n g unknowable  theories b r i e f l y :  more  truth  accordance  Philosophy device  for  provided  that  with  its  canons.  0  constructed  as  a  hypotheticc-deductive  system.  •  Where  possible*  the  distinguished  from  interpretation  •  Uninterpreted rich  in  their  possible  •  The  theory  basic  assumptions  the  assumptions  primitives potential  and  should  be  c l e a r l y  presuppositions,  and  the  axioms  the  pragmatic  conventions.  should  be  chosen  of  the  relationships  and  tc  be  highest  generality.  should  have  'conceptual  unity'.  This  is  tc  say  that: (a)  * o  that  a l l  formulas  of  the  theory  should  be  The f i r s t five of these canons are general in l i t e r a t u r e on the s c i e n t i f i c method and r e f l e c t f o r m u l a t i o n s of Bunge. The l a s t three are our own r e f l e c t the i n f l u e n c e cf H a t t e s s i c h ' s work.  Accounting  thecry  S  tied  the the and  methodology.  IMIS!  together (b)  a l l  the  Intelligent  by  chains  concepts  universe  (c)  a l l  group  may  axiom the  •  A  only more  be  as  and  •  An  to  should  the  same  belong only  may  appear  set  cf  objects,  the  same  semantic  the  theory's  axioms,  and  in  the  tc  via  definitions  ensures  is as  stringent  i t  cnly  cne  conceptual  axicm  cf  the  'connectedness'  expanded  to  standards  by  one  that  of the  of  theory which  of  presented  with  as  its  is  to  the be  true  refute  and i t .  precision  possible the  to  Ever  tc  test  theory  that  a  the  theory  plus  the  experiment.  theories  in  logic  the  of  commands,  the  administrative  s c i e n t i f i c  advices  and  method rules,  actions.  has i t s  tc  greater  tests  the  accommodate  be  truth  seldom  of  truth-values  instrumental  is  nature  reguires  attempts  ever  it  hypotheses  provisionally  give  instead  instrumental  only  survives  tests  i s o l a t i o n ,  the  always  However,  sciences be  this  long  auxiliary  The  relations  refer  introduced and  thecry  theory.  •  164  Theory,.  theory.  factual  in  Accounting  discourse  primitive set;  and  l o g i c a l  predicates  and  no  of  should  of  presuppositions (d)  HIS  a  purpose.  attainment  theory  i f  its  This is  tc  purpose be  and  judged  effectiveness  the  should as  well  assessed.  Accounting  theory  8  methodology.  IMIS2  •  The  I n t e l l i g e n t  particular  body  instrumental of  'input*  V.  It  is  and  a  wrongly  attribute  1  was  THEORY What  not  AS is  Information Grady's  T1  THE 'an  (1965)  approval  the  from  194 1  the  relation  'output*  testing  that  we  the  whole  a  signal  was  set  wffs,  and  cf  fact  a  I  failure in  given  value  done  V  be may  theory  presented  I.  OF  ITS  LABEL.  system'?  What  us  turn  compendium  cf  accounting  Bulletin  a  between  or  T1  Let  AICPA  of  net  DEFINITION  System'?  T1,  165  is  of  accounting  Theory..  this  when  member  a  of  If  the  alone, a  set f c r  s u f f i c i e n t l y .  which  propositions,  important  to  accounting  establishes  specified  (T1+I+V)  2i.l  I  and  of  theory  wffs,  MIS  No.  43  to  (which  is  a  the  authorities.  lore  is  'Management  guotes  quoting  in  with turn  B u l l e t i n ) :  "Accounting is the art of r e c o r d i n g , classifying and summarizing in a significant manner and i n terms cf money, transactions and e v e n t s which a r e , in part at least, of a f i n a n c i a l c h a r a c t e r , and i n t e r p r e t i n g the results thereof. Public by  one  whose  accounting services  is  are  the  practice  available  to  cf  the  this  art  public  fcr  compensation... If accounting were c a l l e d a s c i e n c e attention would be directed (and perhaps limited) tc the ordered c l a s s i f i c a t i o n s used as the accountant's framework, and t o t h e known body c f f a c t s which i n a given case are f i t t e d into this framework. These aspects of accounting cannot be i g n o r e d , but i t is more important tc emphasize the c r e a t i v e s k i l l and a b i l i t y with which the accountant applies his knowledge tc a given problem."  Accounting  thecry  5  methodclcgy.  115152 I n t e l l i g e n t  And  Grady  himself  "Accounting concerned  HIS  and  Accounting  Theory..  166  says: is  the with  body of knowledge systematic  and functions originating,  authenticating, recording, c l a s s i f y i n g , processing, summarizing, analyzing, interpreting, and. supplying of dependable and s i g n i f i c a n t information covering transactions and e v e n t s which a r e , i n p a r t at least, of a f i n a n c i a l character, reguired for the management and operation o f an e n t i t y and f o r t h e r e p o r t s that h a v e t o be s u b m i t t e d t h e r e o n t o meet fiduciary and other r e s p o n s i b i l i t i e s . "  In  view  may  be  turn a  of  the to  a  best  an  we  can  of  of  'overview'  accountancy  dc;  but  before  well-received  Information  d i f f i c u l t y  c a l l s  state  recent,  Management  the  the  Systenu  defining as  an  such we  textbook Davis MIS;  a  characterisation  evaluate fcr  a  (1974) he  offers  i t  let  d e f i n i t i o n  us of  acknowledges something  he  follows:  "There is no agreement on the term 'management information system." (p.3)... "The computer is useful for these c l e r i c a l data processing tasks, but a management i n f o r m a t i o n system performs other tasks as w e l l and i s more t h a n a data processing system. It is an i n f o r m a t i o n p r o c e s s i n g system applying the power o f t h e computer tc provide information fcr management and decision-making." (p.4)..."fl d e f i n i t i o n of a management information system, as the term is generally understood, is an integrated man/machine system for providing information tc support the operations, management and decision-making functions in an o r g a n i z a t i o n . The system u t i l i z e s computer hardware and software, manual p r o c e d u r e s , management and d e c i s i o n m o d e l s and a data base." (p.5)  Accounting  theory  8  methodology.  IMIS.:  Intelligent  These are  descriptions  undeniably  may  be  useful  valuable  phrases  as  to  "events  f i n a n c i a l  character"  or  read  who  can  manner"  and  already  "the  f o c i  •Accounting' a l l ,  they  which  do  which  one  to  exclude  information  review  and  of  a  words  any  prediction  and  software  'Operating  such  that  Davis's  as  that  an  of  given they  lack  a  in  not  to  exact  s t e e l - r o l l i n g  Accounting  a  system"  ene  do  who  is  usefully not  define  Systems'.  Above  conditions or  by  management  deficiencies  does  not  precision  satisfy  because or  exclude by  would on-line  by the they  management  Grady's  offered  an  cf  may  they  such  significant To  accounting  d e f i n i t i o n and  a  they  example,  does  with  man/machine  But  system  They  least,  accounting  operations  System*  at  s u f f i c i e n t  the  definitions)  familiar  computer".  and  167  fashion.  Information  For  MIS  them  part  interest.  putative  accounting  analyst  system  the  identify  systems.  d e f i n i t i o n  in  subject-matters  recognize  other  already  integrated  necessary  or  c a l l  "summarizing  of  the  give  judge  In  into  is  'Management  system  d e f i n i t i o n . f a i l  not  may  "an  power  may  information  are,  and  net  Theory..  ordinary-language  which  particular  and  one  an  Accounting  do  who  with  of  (we  anyone  meaning  familiar  describe  in  and  MIS  improved  the  h i s t o r i c a l  a  investment  apply  both  process  tc  control  mill.  thecry  S  a  methodology.  IMIS^  We  Intelligent  have  seen  complex  set  of  essential  relationships  primitive  semantic  in  concept ARCH,  the  close  methods  are  of  way  the  by  concept.  For  accounting statement  of  symbol  in  be  which  the an  the  at  is  of  i s  d e f i n i t i o n X  is  whether  entity  of  a  cr X  This is  semantic  system'  or  cf  of  label.  is  not  meets  cf set  tc  cf  the  stand It  the .  of  with  name  say,  cf  a the  symbol fcr  is  the  because  we  described  is  as  the  an a l l  then,  theory  and  structure labelled  cur  best  a  network,  whese  arcs  defines  the  and  our  d e f i n i t i o n  such  as  a  d e f i n i t i o n ^  best  of  thecry  cur  that of  clearest  means.  'theory that  tc  that  the  theory,  'accounting'  body  universe  the  definiens. these  specified  ncticn  be  a  which  labelled  to is  in  complex  and  theory,  also  what  entity  as  a  is  whole  a  is  epistemology*.  i t  'Accounting  which  introduced  such  interpreted a  the  That  theory  theory  met  was  relationships. the  consistent  learnt  prepositions  which of  we  168  thecry  between  1  is  a  established  x  semantically  network,  looking  that  network  network.  'applied  of  are  formulation  the  the  the  as  logical  symbol  The  of  Al  nodes  i t s  by  correspondences  One whose  'ARCH'  Theory..  a  A l ,  definiendum,  relationships of  a  could  as  defined  are  of  interrelationships such  Accounting  chapter  Chapter  network  symbol,  this  from  And  and  i n t e r r e l a t i o n s h i p s  terms  discourse.  in  HIS  'MIS'  is  accounting' If  we  Accounting the  a  or  label a  seek  Accounting  theory  the  cf  MIS'  'theory tc  System  necessary  and  know we  and  S  whether  test  tc  see  s u f f i c i e n t  methodology.  I MIS j .  conditions true a  in  of  the  at  •model'  for  theory is  at  a  for  symbolic  those  -  as  are  general,  for  and  emitting  output  we  denotes. theory..  Equally, We  s h a l l  the  return  of  and  level  thing tc  this  tc  the the  actual  choice  cr  Interestingly discourse -  if  we  accounting  it  implementation  Accounting  a  thecry  entries  which  is  in  add  MIS  i t  thought  that  but  is  and  thing; the  of  description  recording  is  then  may  cf  theories  is  is  thecry  level.  input,  which  entity  cf  universe MIS  169  The  interpretation  produce the  Theory..  proposition  According  perceiving  becomes the  every entity.  whose  accounting  theory  that  interpreted  theories  actions  Accounting  abstract  highly  primitive  The  in  rules  the  systems^  such  and  theory.  and  low-level,  enough  Mis  true  the  axioms  presented a  the  theory  primitives, be  Intelligent  cf  the  moment.  8  methodology.  IMIS:.  Intelligent  2±2±1 a n a l y t i c  d e f i n i t i o n s  i2l2IJI!§tion  In in  our  Chapter  we  note  systems'  rather  defined.  The  (1964),  I j i r i  degree  interest other few  n  For  with  almost  They  (1967),  Chambers  (1961)  and  and  brought  without  170  and  Jmanage went  just  quoted,  b i o l o g i c a l although  l i t t l e  example  more  see  theory  Ruse of  exception,  described s c i e n t i f i c  accounting  ty  Mattessich  (See  Prince  (1963),  Euckley,  tc  seme  Kircher  and  administrative to 1  reasons  quoting  (1973) of  'management  the  similar  fcr  and  in  sciences.  than  analyses  on the  thecry  and  grounded  (1966),  so  to  passages  'accounting*  of  methodological axiomatic  the  axiomatizations  exceptions,  content  in  that  are,  formalised  s o c i a l  Theory..  ^accountings  not  1968) in  accounting  are  Churchman  Mathews,  cf  Introduction,  3  method.  and  system.*...  information than  MIS  that We  1  cf  their  and  the  sciences  which  arose  turn  now  tc  space  we  must  an in the be  'axioms'.  Gcudge p o s s i b i l i t y  (1961)  for  of  an  biology.  Accounting  thecry  6  methodclcgy.  IMIS.:  2^3^2.s.l  Semi-axiomatic "The  and  Intelligent  "A  Basic  of  interpret  their  the  The while  AICPA these  generally time.  of  Accounting  the  of  (Sprouse  and  approach  postulates  inferences  without  from  studies  a  they  are  accepted 1  too  stating  valuable radically  accounting  Eoard  a  a  authors'  proceeded sufficient  tc  tc trace  interpretation.  later  contribution different  principles  fcr  reflected the  they  postulates  Principles  Principles  Mconitz,1962)  giving  §££reach  (Mocnitz,1961)  Accounting  However,  Accounting are  pcstulaticnal  in  171  Theory..  Accounting"  Broad  'postulates'.  logical  thinking,  this  Set  axiomatic  selection  of  theories^  Enterprises"  consciously  and  Postulates  Tentative  Business  MIS  said tc  "...that accounting  frcm  fcr  present  acceptance  at  2  we do not suggest that the use of a semi-fcrmal theoretical style was the cause cf this rejection although its unfamiliarity may h a v e b e e n a contributing factor. In retrospect i t appears to us that these twe studies were n e i t h e r f i s h nor fowl: they were a radical departure from t r a d i t i o n a l t h e o r y i n t h a t t h e y were based cn axiom-like premises yet they f a i l e d tc meet the methodological canons cf geed theory. Fcr example, it is d i f f i c u l t to find any practices of accountancy which an inference from the postulates would exclude^ Since they t h e r e f o r e admit c o n t r a d i c t i o n s t h e y o f f e r no c c u n s e 1 ^ 1  2  Accounting  theory  8  methodclcgy.  IMIS! I n t e l l i g e n t  The The  fourteen  basic  BIS a n d  Accounting  postulates  172  Thecry..  were:  Environment:  Postulate A-Ji Quantification. Quantitative data are helpful in making r a t i o n a l economic decisions, i . e . , i n m a k i n g c h o i c e s among alternatives so that actions are c o r r e c t l y r e l a t e d to consequences. Postulate A-2.. Exchange Most cf t h e g c c d s and s e r v i c e s that a r e produced are distributed through exchange, and are net directly consumed by the producers. x  Postulate c a r r i e d on  A-3.. Entities.. through s p e c i f i c  Postulate A-4.. c a r r i e d on d u r i n g  Economic activity units or e n t i t i e s .  Time period economic specifiable periods cf  i s  activity i s time.  P o s t u l a t e A_25.. U n i t o f measure.. Honey i s t h e common denominator i n terms of which the e x c h a n g e a b i l i t y c f goods and services, including labor, natural r e s o u r c e s , and c a p i t a l , a r e measured. The  field  of  Accounting!  Postulate B-J F i n a n c i a l statements.. The r e s u l t s c f the accounting process a r e expressed in a set cf fundamentally related financial statements which a r t i c u l a t e w i t h each o t h e r and rest upon the same underlying data. x  Postulate B;2 b a s e d on p r i c e s exchanges which expected t o .  X  Karket prices.. Accounting data are g e n e