An English-Like Language For Qualitative Scientific Knowledge

KAL (Knowledge A c q u i s i t i o n Languaqe) is a language intended not f o r programming but f o r the expression o f q u a l i t a t i v e s c i e n t i f i c knowledge. I t is designed to accommodate both the semantic cons t r u c t s most f r e q u e n t l y found in s c i e n t i f i c t e x t ( a t present, excluding the d e s c r i p t i o n of change) and the knowledge storage and r e t r i e v a l mechanisms of languages l i k e C0NNlVER, in which i t may be implemented. Emphasis is placed on using " r e a l " t e x t , and i n i n t e r a c t i n g with s c i e n t i s t s ( p a r t i c u l a r l y b i o l o g i s t s ) t o discover t h e i r p a r t i c u l a r needs. Though KAL is E n g l i s h l i k e , most natural languages could be s u b s t i t u t e d w i t h no d i f f i c u l t y . I n t r o d u c t i o n I t i s generally agreed t h a t mathematics i s not well suited to describe most b i o l o g i c a l knowledge, though whether t h i s is an (unfortunate) h i s t o r i c a l accident or a r e f l e c t i o n of how our minds r e l a t e to the world is an open question. What appears to be more appropriate, however, are the formalisms which are becoming increasingly well developed through AI research i n t o representing non-quantit a t i v e knowledqe. For many reasons, however, some of which are j u s t i f i a b l e , these languages (e.g., CONNIVER) tend to remain w i t h i n AI la b o r a t o r i e s . I wish to suggest t h a t t h i s tendency of "growing outward from the AI lab to someday i n t e r s e c t the real world", evidenced by the comparative lack of attempts to apply any developments " w i t h i n A I " to problems not invented by AI researchers may, in f a c t , be detrimental to both AI and the rest of the world. With t h i s in mind I have t r i e d to place myself in an " i n t e r f a c e " p o s i t i o n , to t r y to bring together the types of knowledge that b i o l o g i s t s need and the tools developed w i t h i n the AI community f o r expressing q u a l i t a t i v e knowledge. I am therefore in the process of d e f i n i n g a language, c a l l e d KAL (Knowledge A c q u i s i t i o n Language), which is not a programming language but a l a n guage intended to be used by b i o l o g i s t s or physicians to express as much as possible of the q u a l i t a t i v e knowledge they use in t h e i r everyday work. The development is at the point (December, 1974) of making i n i t i a l attempts t o expose some b i o l o g i s t s to the language, by means of a number of hours of explanation plus a 50 page "user's manual", w r i t t e n f o r persons u n f a m i l i a r with computers. To p a r a l l e l KAL's evolu t i o n through i n t e r a c t i o n w i t h b i o l o g i s t s , I have begun implement i n g some of i t , using UCI LISP and CONNIVER, so tha t it can be developed both "on paper" and as a program, in the question answering system t r a d i t i o n . I t i s not a n t i c i p a t e d t h a t a l l o f what can be defined on paper w i l l be implemented. Design Criteria for KAL The f o l l o w i n g desiderata were foremost in the design of the language: 1) It must be suited p r i m a r i l y f o r the expression of that p o r t i o n of b i o l o g i c a l ( i n c l u d i n g medical) knowledge which is now expressed in natural language t e x t . This excludes a l l forms of graphical and mathematical expressions, since the problem of representing graphical knowledge (pictures and diagrams) is a f l o u r i s h i n g f i e l d in A l , and representing mathematical knowledge is the bread and b u t t e r of computers. 2) It must be comprehensible to persons having no experience with computers. Hence, I have chosen t o make KAL look as much l i k e a natural language (English) as possible, so that most KAL sentences can be immediately understood by anyone f a m i l i a r w i t h KAL who knows the technical vocabulary. Most motivated b i o l o g i s t s or physicians should be able to learn to w r i t e KAL in a reasonable amount of time. 3) The technical vocabulary must be determined by the user, in an open ended ( i . e . c o n t i n u a l l y ext e n s i b l e ) manner. Thus, the major parts of speech, used in an appropriate context, should be automat i c a l l y added to the machine's lexicon. This forces some r i g i d i t y on the syntax, but t h i s i s not considered detrimental to content, only s t y l e . The meaning of any word is given by the t o t a l i t y of sentences in which it has appeared. The skeleton of the syntax and semantics are a reserved group of English* words and phrases (keywords) around which the user may invent a r b i t r a r y f l e s h . 4) If it is to be LISP/CONNIVER based, KAL should i n t e r f a c e as simply as possible to those aspects of the host languages which are u s e f u l , without becoming subjugated by them. One must r e s i s t the temptation to the seduced by the so-called "power" of such languages i n t o doing only what they make " n a t u r a l " , and thereby avoiding what would be more appropriate to the task. 5) Deduction, used to answer questions and to part i a l l y check the consistency of new sentences, should r e l y p r i m a r i l y on a v a r i e t y of special mechanisms (e.g., set hi e r a r c h i e s , property l i s t s and f u n c t i o n a l correspondences), and secondly, on the "problem reduction" paradigm a la PLANNER. Question posing is to be in the form of normal d e c l a r a t i v e sentences having "blanks" to be f i l l e d i n . 6) Constant a t t e n t i o n is maintained while encoding a t e x t f o r those sentences which cannot be properly handled in the current ( s t a t i c ) version of KAL, p a r t i c u l a r l y those which describe actions or change. These are c a r e f u l l y noted in a n t i c i p a t i o n of a f u t u r e "dynamic" version. *They could be j u s t as well chosen from many other languages, suggesting the a t t r a c t i v e p o s s i b i l i t y of easy t r a n s l a t i o n f o r KAL-based knowleoye. Douglas Skuce Dept. of E l e c t r i c a l Engineering and Montreal Neurological I n s t i t u t e McGill U n i v e r s i t y 3801 U n i v e r s i t y St., Montreal, P.Q., Canada