American Journal of Computational Linguistics 
~icrofiche 21 
NEWSUTTER OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS 
VOLUME 12 - NUMBER 3 
JULY 1975 
Recent  compute^ science research 
in natural languaqe process.ing 
by Allen Klinger - 2 
Current bibliography - 26 
AMERICA@ JOURlUAL OF COMPUTATIONAL tTNGUISTICS is published by 
the Center for Applied Linguistics for the Association for 
Computational Linguistics 
EDITOR David G . Hays Professor of Linguistics and of Computer Science 
State University of New York, ~uffalo 
EDITORIAL STAFF Jeff F. Lesinski Assistant Pro Tern 
Jacquln Brendle secretary 
EDITORIAL ADDRESS Twin Willows, Wanakah, New Pork 14075 
MANAGING EDITOR A. Hood Roberts Deputy Director, Center for AppJied 
Linguistics 
MANAGEMENT STAFF Nancy Jokovich and David Hoffman 
PRODUCTION AND SUBSCRIPTIOW ADDRESS 1611 North Kent Street, 
Arlington, Virginia 22209 
Copyright 1975 
Association for Computational Linguistics 
Americaa Journal of Computational Linguistics Microfiche 21 : 2 
RECENT COMPUTER SCIENCE RESEARCH IN 
NATURAL LANGUAGE PROCESSING 
ALLEN KLINGER 
Computer Science Department 
School of Enqineering and Applied Science 
University of California, Los Angeles 
ABSTRACT 
Ihe maohine translation problem has recently been replaced 
by much narrower goals and computer processing of language has 
become part df artificial intelligence (AI), speech recognition, 
and structural pattern recognition. These are each specialized 
computer science research fields with distinct objectives and 
assumptions. The narrower goals involve making it possible for 
a computer user to employ a near natural-language mode for 
problem-salving, information retrieval, and other applications. 
Naturdl computer responses have also been created and a special 
term, "understanding", has been used to describe the resulting 
computek-human dialogues. Phe purpose of this paper is to 
survey these recent developments to make the A1 literature ac- 
cessible to researchers mainly interested in computation on 
written text or spoken language. 
1, INTRODUCTION 
The computer literature discussed in this paper uses 
several linguistic terms in special ways, when there is a 
possibility sf congusion, quotation marks will be used to 
identify technical terms in computer science. The term 
"understanding" is frequently used as a synonym for "the addi- 
tion of logical relationships or semantics to syntactic pro- 
cessing". This use is substantially qarrower than the word's 
implicit association with "human behavior implemented by 
computer'' the narrower use is introduced as a neutral refe- 
rence point, The question of whether a computer porgram can 
operate in a human-like way is central to artificial intelli- 
gence. "Do current 'understanding' program systems show how 
extended human-like capability can be implemented using com- 
puters?" is a related pragmatic questton Initially this 
investigation sought to examine whether programs which "under- 
stand" language in the stipulated narrow sense are protatypes 
which could lead to expanded capability. Unfortunately, 
"language understanding" and its special subtopic "speeeh 
understanding'' are insufficiently developed to permit profi- 
table discussion of the original question Hence an opera- 
tional approach to the recent literature is taken here. This 
paper outlines how "language understanding" research has evolved 
and identifies key elements of program organization used to 
achieve limited computer "understanding". 
2, LEVEL AND DOMAIN 
Current A1 programs for lankuage processing are organized 
by level and restricted to specified domains. This section 
presents those ideas and comments on the limitations that 
they entail. 
Three principal levels of language-processing software are 
1. "Lexical" (allowed vocabulary) 
2. "Syntactic" (allowed phrases or sentences) 
3 "Semantic" (allowed meanings) 
ln practice all these levels must operate many times for the 
computer to interpret even a small portion, say two words, of 
restricted natural-language input. Programs that perform 
operations on each level are, respectively, 
1. Word in a table? 
2. Word string acceptable grammatically? 
3. Word string acceptable logically? 
A program to detect "meaning" (logical consequences of word 
interpretations) must also perform grammatical operations for 
certain words to determine a part of speech (noun, verb, adjec- 
tive, etc.) One method makes a tentative assigrlment, parses, 
then tests for plausibility via consistency with known facts. 
To reduce the complexity of this task, the designer limits 
the subset of language allower or the "world" (i.e. the subject) 
discussed. The word "domain" sums up this concept, other terms 
for "restricted domain" are "limited scope of discourse", 
"narrow problem domain", and "restricted English framework" 
The limitation of vocabulary or context constrains the 
lexicon and semantics of the "language". 
The trend in the 
design of software for "natural-language understanding" is 
to deal with (a) a specialized vocabulary, and (b) a parti- 
cular context or set of allowed interpretations in order to 
reduce processing time. 
Although computing results for several 
highly specialized problems Le,g. 7, 231 are impressive exam- 
ples of language processing in restricted domains, they do 
not answer several key concerns. 
1. Do specialized vocabularies have sufficient 
complexity to warrant comparison with true 
natural language? 
2. Are current "understanding" programs, orga- 
nized by level and using domain reatrictidn, 
extendable to true natural language? 
The realities are severe. Syntactic processing is interdependent 
with meaning and involves the allowed logical relationships 
among words %n the lexicon. Most natural-language software is 
highly developed at the "syntactic" level Howwer, the number 
of times the "syntactic" level must be ent'ered can grow explo- 
sively as the "naturalness" of the language to be processed 
increases. Success on artificial domains cannot imply a great 
deal about processing truly natural language. 
3, PROGRAM SYSTEMS 
The systems cited in this section answer questions, per- 
form commands, or conduct dialogues. 
Programs that enable a user to execute a task via computer 
in an on-line mode are generally called "interactive" Some 
systems are so rich in their language-processing capability 
that they are called "conversational" Systems that have 
complicated capabilities and can reply with a sophisticated 
tesponse to an inquiry are called "question answering". The 
survey [I] discusses two "conversational" pkograms ELIZA 
[2, 31 and STUDENT [41, which answers questions regarding 
algebraic word problems. SIR [51 answers questions about 
logic. Both [41 and [51 appear in [61 , the introd,rction 
there provides a general discussion of "semantic information 
and computer programs involving "semantics" 
The "question-answering" program systems described in 
[2-51 were sophisticated mainly in methods of solving a prob- 
lem or determining a response to a statement. Other systems 
have emphasized the retrieval of facts encoded in English. 
The "blocks-world" system described in [71 contrasts with 
these in that it has sophisticated language-processing capa- 
bility It infers antecedents of pronouns and resolves ambi- 
guities in input word strings regarding blocks on a table. 
The distinction between "interactive", "conversational", and 
"question-answering" is less important when the blocks-world 
is the. domain. The computer-science contribution is a program 
to interaet ,wfth the domain as if it could "underktand" the 
input, in the sense that it takes the proper action even when 
the input is somewhat ambiguous. To resolve ambiguities the 
program refers to existing relationships among the blocks. 
The effect of [71 was to provide a sophisticated example of 
computer "understanding" which led to attempts to apply simi- 
lar principLes to speech inputs. 
(More detail on parallel de- 
velopments in speech processing is presented later.) 
The early "language-understanding" systems, BASEBALL [9], 
ELLZA, and STUDENT, were based on two special formats: one to 
represent the knowledge they store and one to find meaning in 
the English input. 
They discard all input information which 
cannot be transformed for internal storage. The comparison 
of ELIZA and STUDENT in [I] is with regard to the degree of 
"understanding" ELIZA responds either by transfoiming the 
input sentence (essentially mimicry) following isolation of a 
key word or by using a prestored content-free remark. STUDENT 
translates natural-language "descriptions of algebraic equations, 
... proceeds to identify the unknowns involved and the relation- 
ships which hold between them, and (obtains and solves) a set 
of equations" [I, p 851. Hence ELIZA munderstands" only a few 
key words; it transf6rms these words via a seatence-reassembly 
rule, discards other parts of the sentence, and adds stock 
phrases to create the response. STUDENT solves the underlying 
algebraic problem-- i t "unders'tands" in that it "answers questions 
based on information contained in the input" [4, p. 1351 . ELIZA 
responds but does not "understand", since the reply has little 
to do with the information in the input sentence, but rather 
serves to keep the person in a dialogue. 
Programs with an ability to spout back similar to ELIZA's 
usually store a body of text and an indexing scheme to it. 
This 
approach has obvious limitations and was replaced by systems 
that use a formal representation to store limited logical 
concepts associated with the text. One of them is SIR, which 
can deduce set relationships anong objects described by natural 
language. SIR is designed to meet the requirement that "in 
addition tu echoing, upon request, the facts it has been 
given, a machine which 'understands' must be able to recog- 
nize the logical implications of those facts. It alqo must 
be able to identify (from a large data store) facts which are 
relevant to 8 particular question'' [51 . 
Limited-logic systems are important because they provide 
methods to represent complex facts encoded in English-language 
statements so that the facts can be used by computer programs 
or accessed by a person who did not input the original textual 
statement of the fact. Such a second user may employ a com- 
pletely different form of language encoding. Programs of this 
sort include DEACON [lo, 111 and the early version of CONVERSE 
[121. The former could "handle time questions" and used 
a bottom-up analysis method which allowed questions 
to be nested. For example, the question "Who is 
the commander of the battalion at Fort Fubar?" was 
handled by first internally answering the questian 
"What battalion is at Fort Fubar?" The answer was 
then substituted directly into the original question 
to make ic "Who is the commander of the 69th batta- 
lion?" which the system then answered. [7, p. 371 
CONVERSE contained provisions for allwing even more complex 
forms of input questions (Recent versions are described in 
113-151 .) 
Deductive systems can be divided into general systems 
which add a flrst-order predicate-calculus theorem-proving 
capability to limited-loglc systems to improve the complexity 
oE the facts they can "infer", and proccdurnl systcms which 
enable other computations to obtain complex information The 
theorem-proving capability is designed to work Erom a group 
of logical statements given as input (or statements consistent 
with the'se input s-tements) However, facts INCONSISTENT 
with the original statements cannot always be detected and 
deductive systems quickly become impractical as the number of 
input statements (elementary facts, axioms) becomes large 
[b, 7, 161, since the time to obtain a proof grows to an im- 
practical length. Special programming languages (e.g. QA4 
[17, 181 , PLANNER [20, 211 ) , have added strategy capabilities 
and better methods of problem representation to reduce computing 
time to practical values 
QA4 (seeks) to develop natural, intuitive represen- 
tations of problems and problem-solving programs. 
(The user can) blend ... procedural and declarative 
information that includes explicit instructions, 
intuitive advice, and semantic definitions. €171 
However, there is currently no body of evidence regarding the 
effectiveness of the programs written in this programming 
language or related ones on problem-solving tasks in general 
or "lapguage understanding1' in particular. There is a need 
for experimental evaluation of the strategies that the pro- 
gsahing language permits for "language understanding" problems. 
Procedural deductive systems facilitate the augmentation 
of an existing store of complex information. Usually systems 
require a new set of subprograms to deal with new data: 
each change in a subprogram may affect more of the 
other subprograms. The structure grows more awkward 
and difficult to generalize. ... Finally, the. system 
may become too unwieldy for further expkrimentation. 
15, p. 911 
In procedural systems the software is somewhat modular In 
19 "semantic primitives" were assumed to exist as LISP sub- 
routines. PLANNER 1201 allows complex information to be 
expressed as procedures without requiring user involvement 
with the details of interaction anong procedures (but [21] 
reports some second thoughts). 
The work of many other groups could be added to this 
survey. Recent work on REL, building on on [lo, 111 is 
reported in [36, 371; [24, 251 are relevant collections; and 
[26] is a survey paper. 
4, DEDUCTION 
In all of the program systems described thus far, "language 
understanding" depends on the "deductive capabilities" of the 
*Some experiments on problem-solving effectiveness of 
special programing languages in another context appear in [22]. 
program, that is, its ability to "infer" facts and rela.tionships 
from given statements. In some cases deduction involves dis- 
cerning structure in a set of facts and relatimships. 
This 
section describes how "imderstanding" prOgrAmS qhemselves are 
structured and how that structure limits tfheir capability for 
general deduction. 
Theorem-proving programs use an inference rule illus- 
trated in [23 p. 611 to deduce new knowledge. A formal suc- 
cession of 1ogi.cal steps called resolutions leads to the new 
fact. The example there begins with P1 - P4 given: 
P1 if x is part of v, and if v is part of y, then 
x is part of y; 
P2 a finger is part of a hand; 
P3 a hand is part of an arm; 
P4 an arm is part of a man 
A pr'oof that 
P9 a finger is part of a man 
is derived by steps, such as combining P1 and P2 to get 
P6 if a hand is part of y, then a finger is part of y 
Unfortunately, it is easy to move outside the domain where 
the computer can make useful deductions, and the formal reso- 
lution process is extremely lengthy and thus prohibitively 
costly in computer time. In [31, 321 it is shown that some 
statements ("whol did not write ---?It) are unanswerable and 
that there is no algorithm which can detect whether a question 
stated in a zero-one logical form can beb answered. 
Henc.e 
theorem proving is not: e-sential to "deduction" and "under- 
standing" systems, natural or artificial, must rely on other 
techniques, e.g., outside information such as knowledge about 
Lhe domain. 
In most "understanding" programs, information on a primi 
tive level of processing can be inaccurate; for example, the 
identification of a sound string "blew" can be inaccurately 
"blue" Subsequent processing levels combine identified pri- 
mitives. If parts of speech are concerned, the level is syn- 
tactic; if meaning is involved, "semantic"; if domain is in- 
volved, the lave1 is that of the "world". Sach level can be 
an aid in a deductive process, leading to "understanding" an 
input segment of language. Programs NOW EXIST which opera- 
tionally satisfy most of the following points concerning 
"understanding" in narrow domains (emphasis has been added) 
Perhaps tha most importaht criterion for undersvand- 
ing a language is the ability TO RELATE THE INFORMA- 
TION CONTAINED in a sentence TO KNOWLEDGE PREVIOUSLY 
ACQUIRED. This IMPLIES HAVING SOME KIND OF MEMORY 
STRUCTURE IN WHICH THE INTERRELATIONSHIPS OF VARIOUS 
PIECES OF KNOWLEDGE ARE STORED AND INTO WHICH NEW 
INFORMATION MAY BE FJTTED ... The memory structure 
in these programs may be regarded as 3emantic, cog- 
nitive, or conceptual structures.,.these programs can 
make statements or answer questions based not only 
an the individual statemegts they were previausly 
told, but also On THOSE INTERRELATIONSHIPS BETWEEN 
CONCEPTS that were built up from separate sentences 
as information was incorporated into the structure ... 
THE MEANINGS OF THE TERMS STORED IN MEMORY ARE PRE- 
CISELY THE TOTALITY OF THE RELATIONSHIPS THEY HAVE 
WITH OTHER TERMS IN THE MEMORY. [28 pp. 3-4) 
This has been accomplished through clever (and lengthy) com- 
puter programming, and by taking advantage of structure inhe- 
rent in special proklem domains such as stacking blocks on 
a table, moving chess pieces, and retrieving facts about a 
large naval organization. 
Program systems for understanding begin with a "front 
end": a portion designed to transform language input into a 
computer representatiort. The representation may be as simple 
as a character-by-character encoding of alphabetic, space 
marker, and punctuation elements. However, a complex "front 
end" could involve word and phrase detection anti encoding. 
The usual computer science term fol a computer representation 
is "data structure'' [271 and there are many types. The language 
processimg program DEACON used ring sqructures 1111, a repre- 
sentation frequently used to store queues. In principle a 
data structure can represent involved associations, but in 
practice simple order or ancestor relationships predominate 
Completely different and far more complex types of structure 
are inherent in natural language. For example, from 1281 
"The-professors signed a petition." is not true. 
has for valid interpretations: 
(a) 
The professors DIDN'T sign a petition. 
(b) 
THE PROFESSORS didn't sign a petition. 
(c) 
The professors didn't sign a PETITION. 
(d) 
The professors di'dn't SIGN a petition: 
Iterative substitution of alternatives to deduce overall mean- 
ing yields cumbersome processing, especially when there are 
nested uncertainties. The recursive properties associated 
with the data structure term "list" [271 are not easily 
adapted to multiple meanings. Hence, representing linguistic 
data for computation is ah opwr and fundamental researrh 
problem. Nevertheless, the programs which de~uce facts from 
language do so withnut a clear best technique for computer 
representation. To do this, restrictions on the language 
implicit in the input domain are used, and repeated process- 
ing by level (lexical, syntactic, semantic) is used in the 
absence of an efficient representation language. Data struc- 
tures that facilitate following the language structure are 
needed Existing programs provide special solutions to the 
problems of deductive processing in narrow language domains 
While these programs are not a general breaktht-ough in reure- 
senting language data for computation, they demonstrate that 
current programming t.echniques enable a us.eful "understanding" 
capability Furthermore, tbere is a reql potential for use 
ot the "understanding" in an interactive node to facilitate 
use of computers by nonspecialists and to tap fhe more sophis- 
ticated human understanding capabilities 
5 INTERACT I ON 
Research and computer program development desrgned to 
store multitudes of facts so that they can be accessed [29] 
qx combined [301 and "unders,tood (see pp. 3-10 in [301') in 
linguistic form (see pp. 11-17 of [30]) is highly relevant to 
recent research programs in text and speech understanding. 
When such a system is used a user might fail to get a fact or 
relataonship because the natural-language subset chosen to 
represent his question was too righ--i.e., it includes a com- 
plex set of logical relationships not in the computer. Thos 
a block could result in a human-computer dialogue if the 
program has no logical connection between "garage" and "car" 
but only between "garage" and "house" (the program replies 
"OK" or "??'!I1 to user input sentences) 
I LIKE CHEVROLETS. 
OK 
CHEVROLETS ARE Eco~oMICAL. 
OK 
M.Y HOUSE HAS A LARGE GARAGE. 
OK 
I CAN GET TWO IN 
??? 
The computer failed to "understand" that there was no change 
of discourse subject. This is an example of a "semantfo" 
failure whi~h could be overcome by interaction. That is; the 
human user would need to dnput one more meaning or associa- 
tion of a valid word so that computer "understanding" may be 
achieved. Syntactic blocks may also occur. M. Denicoff 
pointed out that in [7] 172 different syntactic features were 
used fox a situation where there are no statements with psy- 
chological content and no use of simile. 
If the psychological 
meanings are added as in 1381, these features would not be 
enough to describe all the possible meanings of a text drawn 
from a less artificial soprce. Indeed, a key problem which 
formal granhars seem ill-suited for is the reality that many 
contexts may-be sifiultaneousLy valid: multiple meanings give 
natural-language communication the richness of ovextones, ana 
subtleties--poetry carries this ta an extreme. 
The above dialogue on "Chevroletsl' is an example of what 
Carbonell [39, p. 1941 called "mixed-initiative discourse". 
This important aspect of interaction is considered in the LISP 
program DWIM ("Do What I .Mean1'), which is a useful working tool 
for text-input error correction precisely because it "under- 
stands" the user's character~stics . (For example, typical 
spelling errors .) This is discussed by Teitelman 140, 41 ,, 421 
A great deal of effort has been put into making DWI-N 
'-'smartn. Expkrience with perhaps a dazen different 
users indichtes we have been very successful: DWIM 
seldom fails to correct an error the user feels it 
should have, and almost never mistakenly corrects an 
error. [40, p. 111 
Another limited-discourse interactive program 1431 facilitates 
introduction of expert knowledge on 7hess. The program uses 
search with a maxitnum look-ahead depth of 20 and has back- 
tracking capability; both s,yntactic and semantic knowledge is 
incorporated. By grouping similar board positions (i.e., all 
involving a piece on cell 1, all involving a queen moue), it 
imposes semantic organization on the vast files to be searched 
and improves syntactic processing speed 
Bublication of 1443, which coined the term "speech 
understanding", initiated,the natural next ,step toward use of 
the computer's "understandfng" capability. The goal of easy 
intqraction with the computer becomes more exciting with 
speech as input medium. Systems to recognize both text and 
speech have used syntax and context [45, 461, but [471 added 
a comprehensive approach using multiple processing levels to 
resoLve ambiguities., In the dkrect successors of this work 
[ 8, 
491, the same Frocess of parriaL acceptance of primitive 
elements (phonemic candidates from digitized acoustic data) 
followed by lexical, syntactic, and semantic processing ko 
rank alternatives has shown significant success. Reddy (in a 
,Carnegie-Mellon University film on the Hearsay System) sta-tee 
hat on 144 connected utterances, involving 676 words, obtained 
from 5 speakers, performing 4 tasks (chess, news retrieval, 
medical diagnosis, and desk calculator use), req.uiring 28 to 
76-word vocabul~ries, t,he computer program recognition, in 
terms of words spotted and identified correctly, was 
a. 89% with all sources of kmm1edge 
. 67% without use of semantic bowledge 
c. 44% without use of sptactic or semantic knowledge 
These results were obtained in October 1973, and have been im- 
proved since [501. However, a key limitation of this form of 
computer speech "understandkng" is response rate. Reddy 
estimated that the third w~rd-accuracy figure (without use of 
syntactic and semantic knowledge) would have to be in excess 
of 90% to allow the program to achieve a near-human response 
speed. 
The nature of computer "understanding" programs leads to 
problems of combinatoric explosion in number of alternatives 
and this lessens the usefulness of multilevel program organi- 
zation (acoustic-phonetic, lexical, syntactic, semantic, 
domain, and user interactions) as much in speech processing as 
in text pro.cessing. Prototype speech "understanding" systems 
have been build 149, 501 and newer acoustic-phonetic and 
syntactic techniques have been incorporated into this work 
[49, 51, 521, yet it seems clear that the development of theory 
in prosody and grammar cannot provide a breakthrough to escape 
the combinatoric explosion. The reason is that the search of 
parse rrees and the use of semantics (look up related words) 
depend on a single context--both take geometrica'lly increasing 
amounts of computing time as the number of contexts grows. 
Furthermore, this increase in time is added onto that which 
occurs when the size of lexicon is expanded. ks words are 
added, the number of trees that can Be-produced by the gram- 
mar's rewriting rules in an attempt to "recognize" a string 
expqnds rapidly. Hence in speech as in text processing, 
"under,standihgn1 exists via computer yet it is not likely to 
lead to rhachine processing of truly natural language. Indeed 
the artificiality of speech "understanding" by computer is 
even greater than that of text input. The "moon rocks" text 
system [33, 351 used a vocabulary of 3500 words, while the 
speech "understanding" version based on it [5M used only 
250 words. 
The COMMERCIAL AVAILABILITY of systems that recognize 
isolated words with 98.5% accuracy [531* and the need for 
a rapid human-computer input interface [54] promise that the 
last word has not been spoken on "understanding". Research 
and development on language handling systems is continuing in 
the hope of achieving useful "understanding". Indeed, Stan- 
ford Research ,Institute's Artificial Intelligence Center is 
basing its current work on the just-mentioned isolated-word 
recognizer. It is likeap that useful developments will occur 
where language, and probably spoken-language, "understanding" 
will be exhibtred. These developments will occur through 
careful design of tasks and use of advances in computer 
technology However, the general problem of machine' "under- 
standing" of natural language- -whether text Dr speech- - is not 
likely to be aided by these developments. 
7 CONCLUS IONS 
A large body of research in computer science is devoted 
to language processing. A survey of the program systems that 
*Threshold Technology Inc. has sold such a system to se- 
veral users. Their VIP-100 includes a miriicamputer dedicated 
to the recognition task; there are otker isolated-word systems 
[541 
have been reported shows that two main goals have emerged: 
1 To enable "intelligent" processing by the 
computer ("hrtificial intelligence") 
2. To produce a more useful way to access 
daa and solve problems ("man-machine 
interaction") 
Techniques in artificial intelligence and speech recognition 
have been developed to the extent that prototype computer 
program systems which exhibit "understanding" have been de- 
veloped for highly limited conrexts. To extend these pro- 
grams to larger subsets of natural language poses problens, 
it is unlikely that any of the yesearch directivns currently 
being explored will of thewselves "solve" the I1na.tural lan 
guage problem". (The techniques include, but are not limited 
to, further developments in artificial intelligence program- 
ming languages [17, 18 20, 21, 551.; refinements in theories 
of grammar; improved deductive ability, possibly by better 
theorem-proving techniques; and the introduction of stress- 
related features in the ehcoding of speech [52]. A useful 
collection of language models appears in [56].) Nevertheless, 
prorotype systems for "understanding" both text and speech are 
useful achievements of engineering, and spoken entry of data 
by humans to computers is beginning to be established by 
isolated-word re-cognizers which use a minicomputer dedicated 
to the task. A multiplicity of purposes beyond this simple 
but practical task of data entry are mentioned briefly in the 
foregoing discussion of "interaction". 
Developments along, 
the many diverse paths indicated under that heading are 
likely to be rapid in the future as practical "understanding" 
of subsets of language becomes part of computer technology 
Far another view of the evolution of that process, see [57]. 

American Journal of Complltational Linguistics 
CURRENT BIBLIOGRAPHY 
The selection of material through the current second yeas of 
AJCL's existence remains tentative. A survey of subscribes- 
members will be included in the last packet mailed during 1975 
to establish patterns of coverage for fdture years. 
Categorization Bf entries deepens as the field defines itself 
and the collection of literature against which new items can. 
be matched increases. The advice of members is welcome. 
Many summaries are authors' abstracts, sometimes edited for 
clarity, brevity, or completeness. Where possible, an infor- 
mative summary i.s provided. 
The Linguistic Documentation Centre of the University of 
Ottawa provides many entries; by editorial accident, some 
of the entries recently received from that source remain to 
be included in the next issue. AJCL gratefully adknowledges 
the assistance of Brian Harris and his colleagues. 
Some entries are reprinted with permission from Computer 
Abstracts. 
See the following frames for a kist of subject headings and 
items with extended presentation or review. 
SUBJECT HEADINGS 
..................... 
General 30 
Phonetics ~honology 
................. 
Recognition 34 
Writing 
................ Recognition 35 
Lexicography . ~exlcology 
Dictiodary ..................... 37 
................ Statistics 3.8 
Grammar 
Parser ...................... 38 
........... Semantics . Discaurse 40 
............... Comprehension 45 
Expression ................ 47 
Memory .................... 51 
REPRESENTATION AND UNDERSTANDING 
Edited by ~aniel G . Bobrm and A.7.lan Collins 
Linguistics 
.................. 
Methods 61 
STRING AND LIST PROCESSZNG IN SNOBOL41 
TECHNIQUES AND APPLICATIDNS 
By Ralph E. Griswold 
Reviewed by Norman Badler 
FORTRAN TEC-HNIQUES WITH SPECIAL REFERENCE 
TO NON-NUMEeICAt APPLICATIONS 
By A. Colin Day 
Reviewed by Richard J. Miller 
........... Information structures 71 
Pictorial systems ............. 72 
Documentation ............... 74 
Indexing ................. 7 8 
Retrieval- ..........,...... 7 9 
.................. Thesauri 8 0 
................... Management: dl 
Robotics ................... 82. 
......... 
Social-Behavioral Science 83 
................. 
Humanities 84 
XNDEX THOMISTICUS. SANCTI THOMAE AQUINATIS 
Compilkd by Roberto BuSa, S. J. 
A.reView of the first ten v~hmes by Ford Lewis Battle* 
................ Cohcordance 90 
........... ..... Analysis ; 90 
Instruction . 
General 
THEORETICAL ISSUES 
I I4 
NATURAL LANGUAGE PROCESSING 
AN INTERDISCIPLINARY WORKSHOP IN 
Cambridge, Massachusetts 
June 10-13, 1975 
EDITORS : 
_ -_ _. -- 
Professor R. Schank 
.Department of Camputar science 
Yale University 
10 Hillhome Avenue 
and 
New Haven, Connecticut 06520 
B, L. Nash-Webber 
Bolt Eeranek and Newman Inc 
50 Moulton Street 
Cambridge, Massachusetts 
02138 
AVAILABLE ~OK: Center for Applf ed Linguistics 
1611 North Kent Street 
Arlington, Virginia 22209 
ABSTRACTS FOUND.ELSEWHERE ON THE MICROFICHE 
General. 
- 
THE PRAGUE BuL~ETIN OF MATHEMATICAL 
LINGUISTICS 
Unive'rsi ta Karlova 
Praha 1974 
TABLE OF CONTENTS 
ON VERBAL FRAMES IN FUNCTIONAL GENERATIVE DESCRIPTION 
................... PARTI. J.Panevova 3 
STELLUNG UND AUFGABEN DER UEBRAISCHEN LTNGUISTIK I 
(EINFUHRWGSSTUDIE). P. Sgall ............. 41 
REVIEWS 
ALGEBRAIC LINGUISTICS IN SOME FRENCH SPEAKING COUNTRIES 
(S. Machova) .................... 5 3 
METODIKA PODGOTOVKI INF~RMATSIONMYKH TEZAURUSOV PEREV s 
VENGERSKOGO POD RED I PREDISLOVIEM JU. A. SHREJDERA V 
SB. PERSVODOV "NAUCHNO-TEKHNICHESWA INFORMATSIJA" VY;P 
17, 1971 (T. Ja. Kazavchinskaja4 ............. 74 
FORMAL LOGIC AND LXNGUISTICS, Mouton, The Hague, 1972 
. ............... (0. Prochazka) E. Zierer ,. 74 
AUTOMATIC ANALYSIS OF DUTCH COMPOUND WOKDS, Amsterdam 1972 
W. A. Verloren van Themaat; EXERCISES IN COMPUTATIONAL 
LINGUISTICS, Amsterdam 1970, I+. Brandt Cars-tius 
(M. Platek, L Vomacka) ...... ............A 
77 
General 
COMP.UTAT1 ONAL ANALYSES 
OF ASIAN & AFRICAN LANGUAGES 
A new j,ournal 
Mantaro J. Hashimoto, Editor 
Project on Computational Analysis 
National Inter-Universi ty Reseaxch Institute 
of Asian & African Languages S; CuSl.tqres 
4-51-21 Ni'shigahara, Ki taku, Tokyo 
11 4 Japan 
March, 1975 
TABLE OF CONTENTS 
A STATISTLC STUDY OF NAMES IN TAMIL INSCRIPTIONS 
....... Noboru Karashima and Y. Sabbarayalu 3 
IMPLICATIONS OF ANCIENT CHINESE WITROFLEY ENDINGS 
Mantaro 5. Hashimoto .............. 17 
THE SINO-KOREAN READING OF KENG-SLIE RIMES 
MantaroJ,Hashimoto. .......'....... 25 
"TO", "YUAN" AND "TE"-.. .A COMPARISON WITH JAPANESE 
............... Masayuki Nakagawa 31 
LARYNGEAL GESTUMS AND THE ACOUSTIC CHARACTERISTICS 
IN HINDI STOPS--PRELIMINARY REPORT. 
......... Ryonei Kagaga and Hajime Hirose 47 
General 
William A. Woods 
Bolt Beranek and Newman Inc. 
Cambridge, Mass 02138 
Report No. BBN 3067 April 1975 
Acquaints speech researchers in the state of the art in 
the conceptual development of, and the new perspectives they place 
on, parsing, syntax and semantic interpretation. 
Includes the 
Chomsky hierarchy of grammar models, n0.n-determinism in parsing 
and its implemen&ation in either backtrdcking or multiple in4epen- 
dent alternatives, predictive vs. non-predictive parsing, word 
lattices and chart parsing, Early's algorithm, transition network 
grammars, transfornational grammars and augmented transition net- 
works, procedural semantics, selectional restrictions and semantic 
association, 
General 
IMPROVING METHODOLOGY IN NATURAL LANGUAGE PROCESS I NG 
William C. Mann 
USC Information Sciences Institute 
Marina Del Rey, California 
In: R. Schank ahd B.L Nash-Webber, eds., Theoretical Issues in Natural Language 
Processing, 1975, 126-129; 
Process models are rigorous, process specifications are made 
very explicit, and complexity ts handled by use of computera. 
A 
methodology should be reliable, efficient and have integrative 
power. The distinctive strengths of the currenl computer oriented 
methodology are (a) the complexity of ,data and tzheory is easy to 
accommodate, (b) time sequence and dependencies are preserved, and 
(c) a diversity of hypotheses can be tested. 
Weaknesses are (a) 
experbents often take years to perform, (b) the activity is treated 
as a programming exercise with the status of data and program un- 
clearly defined and (c) in attempting to be general on a particular 
phenomenon, significant others are missed. 
As whole systems are 
produced, they are difficult to disseminate and jud e. 
A systemmay 
process its examples, but ir is hard to determine i 
it is ad-hoc 
and tuned to the examples. 
t 
General 
W. A. Woods 
Bolt Beranek and Newrnan, Inc 
Cambridge, Mass 
In: R. Schank and B.L. Nash-Webbar eds. Theoreticel fssues in Natural Language 
Processing, 1975, 134-139. 
There are two tasks for which methodologies are used, (a) 
building intelligent: machines, and (b) under standing human language 
performance, Both depend on the development of a 'device-indepen- 
dent ' language, understanding theory. Fof theoretical studies, a 
methodology should be cognitively efficient and should deal effect- 
ively with the problem of scale--having a large number of facts 
embodied in the theory. Studies should be performed in the context 
of total language understanding; isolation of qomponents limits 
scope. Intuition on human language performance is a good guide to 
computational linguistics. 
Phonetics -.Phonology : .Recognition 
SPEECH RECOGNITION BY COMPUTER : A BIBLIOGRAPHY WITH ABSTRACTS 
D. W. Grooms 
National Technical Information Service 
5285 Port Royal Rd. 
Springfield, Virginia 22161 
Report No. Corn-74-21435/6, September 1974. Price: $20.00 
Contains 142 abstracts covering recognition, synthesis, and 
the acoustical, phonological and linguistic processes necessary in 
conversion of various waveforms. Retrieved using the National 
Tech~iical Information Service on-line search system. 
Writing - Recognition 
- 
FUZZY LOG l C FOR HANDWRITTEN NUMERICAL CHARACf ER RECOGNITION 
P. Siy and C. S. Chen 
Akron University 
IEEE Transactions on Systems, Man and Cybernetics, $Me-#! 570-575, 1974 
considers characters as a directed abstract graph, of which 
the node set consists of tips, corners, and junctions, and the 
branch set ~onsists of line segments connecting pairs of adjacent 
nodes. Classificat-ion of branch fypes produces features which are 
treated as fuzzy variables. A character is represented by a fuzzly 
function which relates its fuzzy variables, and by the node pair 
involved in each fuzzy variable. After producing a representation 
of an unknown character recognition occurs when a previously learned 
character's representation is isomorphic to the unknown. 
Writing : Recognition 
A MEANS OF ACHIEVING A HIGH DEGREE OF COMPACTlON 
ON SCANDIGITIZED PRINTED TEXT 
R. N. Ascher and G. Nagy 
IBM Corporation 
IEEE Transactions on Computers, C-23) 1174-1179, 1974 
A 16:l compaction ratio was achieved by storing only the first 
instance of each pattern class and thereafter substituting this exem- 
plar for every subsequent occurrence of the symbol. Proposed are 
refinements to yield a 40:l ratio. 
Writing : Reco-gnitian 
A SURVEY OF MODELS AND APPLICATIONS 
William Stallin 6 
Center for Nava f Analyses 
Arlington, Virginia 
Computers and the Humanities 9, 1: 13-24, 1975 
Various proposals are discussed, principally (1) Rankin, 
who has a two-level grammar, the first gives the strokes and 
rules for combination and the second explicates the order, with 
a recursive definition of subframes . 
(2) Fuj imara has an inven- 
tory of strokes and operators. 
For each stroke 3 functional 
points are isolated and operators define the linking by reference 
to these points. Applications include keyboard input, storage 
and retrieval or characters, and automatic recognition. 
There 
are two different approaches. 
One seeks a logically efficient 
system; the other one that seems natural to a user of the language. 
Writing : Recognition 
36 
LHINESE CHARACTER RECOGNlTION BY A STOCHASTIC SECTIONALGRAM METHOD 
Y-L. Ma 
National Taiwan University 
IE6E Transactions on Systems, Man and Cybernetics, SEIG-4; 575-584, 2974 
An approach to recognition of a block picture by comparing 
it with stochastic sectionalgrams obtained by grouping many samples. 
TO calculate the risk, the absolute values of the differences be- 
tween the stroke-occurrence probabilities of corresponding quanta 
in the two sectionalgrams are summed one of these two sectional- 
grams being derived from the input pattern and the other from the 
prototype pattern. The smaller the sum of these differences is, 
the more accurate the input pattern tecognition. 
Wrf ting. : Recognition 
COMPUTER IDENTIFICATION OF CONSTRAINED HAND PRINTED CHARACTERS 
WITH A HIGH RECOGNITION RATE 
W. C. Lin and T. L, Scully 
case western Reserve University 
Cleveland, Ohio 
IEEE Transactions on Systems, Man end Cybernetics, SMC-4, 497-504, 1974 
Hand printed on a standardizing grid made of twenty line seg- 
ments, yielding twenty features, and input using a television camera, 
49 character classes were recognized at a greater than 99.4% rate. 
Feature values calculated utilizing a Gaussian point-to-line distance 
concept were used in a weighted minimum distance classifier. All 
character-dependent data are obtained through training techniques 
Both statistical linear regression and averaging methods are used 
to obtain the parameters defining each character class in feature 
space. 
Lexicoqraphy - Lexicology. : Dictionary 
Joseph D. Becker 
Zn: R. Schank and B.5. flash-Webber, eds., Theoretical Issui?s in Natural Language 
Processing, 1975, 611-63. 
We speak mostly by conjoining remembered phrases. Productive 
processes have secondary roles of adapting old phrases to new situ- 
ations and of gap filling. 
Lexico~raphy - Lexicolopy : Statistics 
PROGRAMS FOR LINGUISTIC STATISTICS 
PART 1: WORD ROOTS IN SCIENTIFIC AND TECHNICAL RUSSIAN 
[Programme zur Sprachstatistlk. Teil 1: 
fVortst3mme in russf-schen natunrissenschaftlichen und technischen.Fachsprachen] 
S. Halbauer 
Angewandte Informatlk, 16: 469-470, 1974 
Description of a program, written in machine language, that 
searches for words containing a fixed stem from Russian mathematical 
texts. 
Grammar : Parser 
George E. Heidorn 
Computer Sciences Dept. 
IBM Watson Research Center 
Yorktown Heights, 'NY 
In R. Schank and B.L. Nash-Webber, was., Theoretical Issues in Natural Language 
Processing, 1-5, 1975. 
Augmented phrase structure grammars consist of phrase struc- 
ture rules with embedded conditions and structure building actions 
Data structures are records consisting of attribute-value pairs. 
Records can be actions, words, verb phrases, etc. There are three 
kinds of attributes: relations, whose value is a pointer to other 
records; properties, with values either numbers or character 
strings; and indicators, whose values have a role similar to lin- 
guistic featires. Structure building rules have a left part indi- 
cating the contiguous segments that must be present for a structure 
building operation, given in a right part, to apply. 
Gramar : Parser. 
DIAGNOSPS AS .A NOTION OF GRAMMAR 
Mitchell Marcus 
Artificial Intelligence Laboratory 
Massachusetts Institute of Technology 
Cambridge 
In R. Schank and B.L. Nash-Webber, eds., Theoretical Issues in Natural Language 
Processing, 1975, 6-14. 
The hypothesis is that every language user knows as part of 
his recognitAon grammar, a set of highly specific diagnostics that 
he uses to decide deterministically what structure to build next at 
each point in the process of parsin a sentepce. This theory re- 
k jects 'backup as a standard contro mechamsm for parsing. A 
grammar is a set of modules. The parser works on two levels, a 
group level and a clause level. Group level modules work on a word 
buffer and build group level structures. Modules have a pattern, 
a pretest procedure and a body to be executed if the pattern matches 
and the pretest succeeds. If the parser fails, it keeps the struc- 
ture constructed to date, and makes whatever substructures it can 
from the remaining part. 
Grammar : Parser 
William White 
National Institutes of Health 
Divislon of Computer Research and Technology 
Bethesda, Maryland 
Journal of Clinical Computing, 3, 180-102, 1973 
A morphological analyzer is written in PL/~ using a recursive 
macro actuated generator. Called with a word as argument 
it returns 
a skem, part of speech, possible transformations, and semantic infor- 
mation. 
Semantics - Discourse 
Brian Phillips 
Department of Information Engineering 
University of Illinois at Chicago Circle 
Doctoral dissertation, State University of New York, Buffalo, 1975 
A theory for the structure of discourse is developed. It 
is sh~wn that proposi~ions of a coherent discourse must be logic- 
ally cbnnected and exhibit a hierarchic thematic structure that 
has a single root. An eqample of a logical connective is 'Cause'; 
a theme is a generalized pattern that is associated with a single 
word, e.g., 'poison' is describable as 'Someone ingests something 
that causes him to become ill'. A theme applies to a discourse 
if 2ts definiens matches part of the discourse. The topic of a 
coherent discourse is its matrix theme; an illformed discourse 
has no topic. 
Not all discourse structure is expressed. If omitted, it 
must be inferrable. The process of inference requires a store of 
world knowledge - encyclopedic knowledge. An encyclopedia is des- 
cribed that contains all the devices reqursed by the discourse 
analysis problem. In fact, the encyclopedia is s general model 
for human cognition and is applicable to inany diverse cognStive 
tasks. The encyclopedia is a directed graph. Categories of nodes 
agd arcs, and of processes, are presented in detail 
Semmtics - Discourse 
ON "FUZZY~ ADJECTIVES 
Fred J. Damerau 
IBM Watson Res~arch Center 
Yorktown Heights, N .Y . 
Repqrt No. RC 5340 
March 27, 1975 
Discusses some of the problems that arise when the concept 
of a linguistic vatiable is combined with the concept of a fuzzy 
set: the range of the numerical base variable, in ordering usagR, 
is not fixed for a given linguistic variable. 
Does not explain the 
computation of values of compound expressions from the values of 
their components. 
Not all adjectives can be related LO an under- 
lying numerical base. 
Other features involved in a complete anal- 
ysis are: average value, typical value, observed value, standard 
deviation of values and polarity. 
[Dl'stribution limited prior to publication 6 ] 
Semantics - Discburse 
USER'S GUIDE TO THE SOLAR THEORETICAL BACKGROUNDS FILE 
Timothy Diller and Tom Bye 
System Development Corporation 
Santa Monica, California 90406 
Report No. TM-5292/002/00 April 1975 
For each analysis in the semantic analysis file the author's 
theoretical orientation, his assumptions, and his notational conven- 
tions are entered on this file. 
The data fields are: idenrifyhg 
number, document source, related sources, words analyzed, conventions, 
theoretlcal basis including - acknowledgements, assumptions, stated 
purpose, and limits, a SOLAR critique, and the name of the person 
responsible for the entry. 
This file is available via on-line 
queries or in a listing format. 
The file can be searched using the 
identifying- number on document source fields. 
Other fields can be 
searched using a string-matching facility. 
Semantics - Discourse 
USER S GUIDE TO THE SOUR SEMANTIC ANALYSIS FILE 
Tom Bye, Timothy Diller, and John Olney 
System Development Corporation 
Santa Monica, California 90406 
Rdgort No. ~M-.5292/001/01) April 1975 
This file contains formal descriptions of word meanings, 
including q~alificatfons, informal explanations, and criticisms of 
descriptions. The wards used are found in the lexicons of the 
Speech Understanding Research groups being sponsored by ARPA. The 
semantic analysis produces 23 data fields for each word, of which 
the following are searchable: word, domain analysis number, source 
part of speech and components Other fields can be searched using 
a string matchin facility. This file is available via on-line 
f queries or in a isting format. 
Semantics - Qiscourse. 
USER'S GUIDE TO THE SOLAR BIBLIOGRAPHY FILE 
Timothy Diller 
System Development Corparation 
Santa Nonica, California 90406 
Report No. TM-5292/000/02 December 1974 
This file provides the citations to the documents raer- 
enced in other SOLAR files. Thirty data tFelds are used, of which 
Che following are searchable: author, year, index term, document 
type, subject ID, document number, and Bell ID, Other fields can 
be searched using a string-matching facility. This file available 
via on-line queries or in a listing format including an mthor. 
keyword and 'sequence number index. 
Semantics - Discour- 
PRIMITIVES AND WORDS 
Yorick Wilks 
Tstituto per Gli Studi 
S-emantici e Cognitivli 
Castagnola, Switzerland 
3n R. Schank and B.L. Nash-Webber, ds., Thearetical Issues in Natural Language 
Processing, 1975, 38-41. 
If semantic primitives are seen as essentially djtfferent 
from words, this leads to attempts to justify them directly, 
usually psychological~y. Otherwise the justification is merely 
that they work. Primitives can be taken as a small natural 
language, with no essential difference betQeen primitives and 
mrde. But the set of primitives cannot be extended indefinitely. 
ptherwise the distinction between the representation end the 
nntural language will be lost. If it is not possible to escape 
frw natural language into another realm, one cannot separate 
semantic representation from reasoning as is attempted. It is 
probably more sensible to say that natural. language understanding 
depends on reasoning rather than vice-versa. 
Semant,ics - Discourse 
META-COMPILING TEXT GRAMMARS AS A MODEL FOR HUMAN BEHAVIOR 
Sheldon Klein 
Computer Sciences Department 
University of Wisconsin 
Madison 
In: R. Schank and B.L. Nash-Webb-, eds., Theoretical Zssues in Natural Language 
Processing, 1975, 84-88. 
A key feature of the system is that the semantic deep struc- 
ture of the non-verbal, behavioral, rules may be represented in the 
same network as the semantics for natural language grammars, and, 
as a consequence, provide non-verbal context for linguistic rules. 
The total system has the power of at least the 2nd Order predicate 
calculus. 
Semantics - Discourse 
RQger C . Schank 
Yale University 
New Haven, Connecticut 
Irl R. Schank and B.L. Nash-Webber, eds., Theoretical Issues in Natural Language 
Processing, 1975, 34-37. 
Canonical reprebentations of cqnceptualisations are composed 
of an ACTOR, an ACTION and a set of ACTION dependent cases. The 12 
primitive actions are ATRANS, transfer of possession; PTRANS, trans- 
fer of physical location; MTRANS, transfer of information; PROPEL, 
application of physical force; M13UILD construction of new contep- 
tual information; INGEST, taking in of an object by an animal; 
GRASP, to grasp; ATTEND, to focus sense organ on an object; SPEAK, 
to make a noise; MOVE, to move a b.ody part; &WEL, to push something 
out of the body; and PLAN, which characterizes the ability to form 
a course of action that leads to a goal. 
Semantics - Discourse 
George A. Miller 
In R. Schank and B.L. Nash-Webber, eds., Theoretical Issues in Natural mguage 
Processing, 1975, 30-33. 
An analysis of the verb 'hand' is paraphrased as: 'S had 
Y prior to some the t at which X used his band to do something 
that caused Y to travel to 2, after which Z had Y' The analysis 
includes a dYscussion of the subsumed Concepts HAPPEN, USE, ACT, 
CAUSE, ALMW, BEFORE, TRAVEL, and AT. 
Semantics - Discours.@ 
A SYSTEM OF SEMANTIC PRIMITIVES 
by Jackendoff 
Department of English 
Brandeis Univefsity 
In R. Schank and 8.1;. Nasb-Webber, eds. Theoretical Issues in Natural Language 
Processing, 1975, 24-29. 
Primitive functions GO, BE an8 STAY can be extended from a 
positional interpretation to possessional and identificational in- 
terprecations. Two kinds of muse are distinguished, CAUSATIVE 
and PERMISSIVE. Inference rules based on the form of semantic 
representations derive logical entailments. e.g. CAUSE, (X,E)-- E. 
Semantics - Piscourse . Comprehension 
Christopher K. RIesbeck 
IA R. Schank: ,and. B;L; Nash-webber, eds., Theoretical Issues in Natural Language 
Processing, 1975, 11-16. 
Comprehension is a memory process; breaking computational 
under standing into sccbyrob-leem af parsirig and semantic iktetpre- 
tation has hindered progress with much effort wasted on the con- 
struction of parsers. A system is described in which a monitor 
takes words from a sentence one at a time, from left to right. 
From a lexicon expectatisns. of the word (or its root) are added 
to a master list of expectations. If an element of the master 
list evaluates m true', programs associated with the element 
are executed. The final structure built by the triggered expecta- 
tions is the meaning ~f the sentence. 
Semantics - Discourse : Comprehension 
Robert P. Abelson 
Yale University 
New Haven Connecticut 
In: R. Schank and B.L. Hash-Webber, Eds., Theoretical Issues in Natural Language 
Processing, 1975, 140-143. 
Reasoning may be propositional ar by mental simulation wing 
visual imagery. In the latter situation, do people include acts 
and objects not present in a given story, but necessary to carry 
out the simulation. This has not yet been experimentally tested. 
Experiments have shown that a listener may simulate a story from 
the point of view of an observer or of a participaht in the story. 
One problem that this raises for AI, if a program can construct an 
interconnected structure from the text, is the non-uniqueness of 
this meaning representation, Another problem is that programs should 
not be designed to preserve all details, but then, what should be 
forgotten; point of view may be useful here 
Semantics - Discourse : Compreha46on 
Her'bert H. Clark 
Stanford University 
Stanford, California 
In; A. Schank and B.L. Nash-Webber, Eds., Theoretical Issues in Nataz-a1 Language 
Processing, 1975, 169-174. 
Listeners draw inferences from what they hear, but different 
listeners can make different inferences. One kind of inference in 
comprehension is in the context of given-new information: the speaker 
tries to construct the given and new information of each utterance, 
so that the listener is able to compute unique antecedents for the 
given information, and so that he will not already have the new in- 
formation attached to the antecedent. Inference mechanisms include 
direct reference, identity, pronominalization, epithets, set member- 
ship, indirect reference by association, indirect reference by 
characterization, reaaons, causes, conseqQences, and concurrences. 
Bridging inferences need not be determinate, but in discourse they 
seemingly are, and further, are the inferences with fewest assump- 
tions. Both backward and forward inferences are possible, but only 
the former are determinate. 
Semantics - Discourse : Comprehension 
COMPUTERS AND NATURAL LANGUAGE 
A. Y. Pratt, M. G. Pacak, M. Epstein and G. Dunham 
National Institutes of Health 
Division of Computer Research and Technology 
Bethesda, Maryland 
Journal of Clinical Conlputihq, 3, 85-99, 1973 
The Systematized Nomenclature of Pathology (SNOP), in use at 
NIH, consists of about 15,000 entries in four lists: topography, 
morphology, etiology, and function. Only a few binary relations on 
terms are needed; e.g., location of morphology, (lesion) at topog- 
raphy (body site). Numerous relations on the primary relational 
triples evidently have to be defined. 
Semantics - Discourse : Expression 
Bertram C. Bruce 
Bolt Beranek & Newman 
Cambridge, Mass 02138 
In: R. Schank and B.L. Nash-webber, eds., Theoretical Issues ih iVatural. tanguage 
Processing, 1975, 64-67. 
Generation is a two stage process. The first formulates a 
lan and the second expresses these intentions; 
there is feedback 
getween the stages. Intentions can be encoded by (i) establish- 
ing presuppositions, (ii) by linguistic conventions, and (iii) by 
discourse structure. kSoc*al Actioii Paradigm is a model of the 
flow of social actions. 
Semantics - Discourse : Expression 
Neil M. Goldman 
Information Sciences Institute 
University of Southern California 
Tn: R. Schank and B.L. Nash-Webber, eds., Theoretical Issues in Natural Language 
Processing, 1975, 74-78. 
In generating natural language from a conceptual structure 
words and syntactic structure must be deduced from the information 
content of the message. Words are accounted for by a pattern mat- 
ching mechanism, a discrimination net. The case framework of verbs 
is one source of knowledge for choice of syntactic structure. 
Semantics - Discourse : Expression 
SPEAKING WITH MANY TONGUES: SOME PROBLEMS IN MODELING SPEAKERS 
OF ACTUAL DISCOURSE 
John H. Clippinger, Jr, 
Teleos 
Cambridge, Mass 02138 
In: R. Schank and B.L. Nash-Webber, eds., Theoretical Issues in Natural Language 
Processing, 1975, 68-73 
In therapeutic discourse the subject is not so much gener- 
ating discourse as regulating it. Statements are made, retracted, 
qualified and restated. The ERMA model simulates this. It has 
five stages, represented as CONNIVER contexts. The discourse 
stream has its source in a special program and then flows back 
and forth between the contexts before achieving its final expres- 
sion. Each context determines suitability for expression; whether 
it should be censored or passed on with suggestions for modifica- 
tion. Concepts are represented by means similar to Minsky's 
frames. 
Semantics - Discourse : Expression 
CONSIDERATIOP~S FOR COMPUTATIONAL THEORIES OF SP6AKING : 
SEVEN THINGS SPEAKERS DO 
John H . Clippinger , Jr . 
Teleos 
Cambridge, Mass 02138 
In: R. Schank and B.L. Nash-Webber, eds., Theoretical Issues.in Natural bnguage 
Processing, 3975, 122-125. 
Technological computational linguistics is primarily con- 
cerned with software technology whereby computers can use and pro- 
cess natural language. Descriptive computational linguistics uses 
the computer a6 a means of developing an accurate and empirically 
valid model of linguistic and cognitive behaviors of human speakers. 
There is no inherent representation of intentions in the former, and 
experience is that it cannot easlly be generalized to.the latter. 
One problem of modeling is that important things are often hidden 
by their familiarity. 
Semantics - Discourse : Expression 
CREATIVITY IN VERBAL1 ZATlON AS EVIDENCE FOR ANALOGIC KNCIWLEDGE 
Wallace L. Chafe 
Department of Linguistics 
University of California 
Berkeley 
Iq: R. Schank and B.L. dash-Webber, Eds., Theoretical IsSues in Natural Language 
Processing, 1975, 144-145. 
Both propositional and non-propositional knowledge must exist. 
Interpretive processes during perception individuate and categorize 
objects. If an object cannot: be categorized then the object will 
be stored with analogic information. During verbalization analogic 
images will be compared with available category prototypes to decide 
on the best match for use in the utterance. 
Semantics - Discourse : Memory 
REPRESEiITATION AND UfjDERSTAliD1!'dG 
STUDIES IN COGNITIVE SCIENCE 
edited by 
Daniel G. Bobrow Allan Collins 
Xerox Palo Alto Research Center Bolt Beranek and Newan 
Pala Alto, California Cambridge, Massachusetts 
Academic Press 
1975 
Dedicated to the memory of 
JAIME CARBONELL, 1928-1973 
1. ~imensions of representation Daniel G. Bobrow ... 1 
2. What's in a link: Foundations for semantic networks 
William A. Woods .................. 35 
3. Reflections on the formal description of behavior 
Joseph D. Becker .................. 83 
4. Systematic understanding: Synthesis, Analysis, and 
contingent knowledge in specialized understanding 
systems Robert J. Babrow and John ~eely Brown ... 103 
11 !~EW MEMORY MODELS 
5. Some principles of memory schemata Daniel G. ~obrow 
and Donaldd. Norman ................ 13i 
6. A frame for frames: representing knowledge for 
recognition Benjamin J. Kuipers .......... 151 
Semantics - Discourse : Memory 
7. Frame representations and the declarative-procedural 
............ contrOVersy Terry Winograd 185 
I 11, HIGHER LEVEL STRUCTURES 
8. Notes on a schema for stories David E. Rumelhart . 211 
9. Tho structure of episodes in memory Roger C. schank-237 
la. Concepts for representing mundane reality in plans 
Robert P. Abelson ................. 273 
IV, SEMANTI c KNOWLEDGE IN UNDERSTANDER SYSTEMS 
11. Multiple representations of knawledge for tutorial 
reasoning John Seely Brown and Richard R. Burton . 311 
12. The role of semantics in automatic! speech understand- 
ing Bonnie Nash-Webber ............... 351 
13. Reasoning from incomplete knowledge Allan Co:lins, 
Eleanor H. Warnack, Nelleke Aiello, and Mark L. 
Miller. ...................... 383 
The preface is reprinted an the following frames 
by permission. 
REPRESENTATION AND UNDERSTANDING 
Preface 
Jaime Carbonell was our friend and colleague. For 
many years he worked with 11s on problems in Artificial 
Intelligence, especially on the developrnerit of RII intelligent 
instructionnl system. Jainle directed the Artificinl 
Intelligence group et Bolt, Deranek, and Ncwman (in 
Cambridge, Massnchusetts) until his death in 1973. Some 
of us who hod worlred with Jaime decided to hold a 
conference in his nlemory, a confcrerlce whose guiding 
principle wonld be that Jaime would have enjoyed it. This 
book is the result of that conferencb. 
Jain~e Carbonell's isnportant contribution to cognitive 
science is best sumnlarized in the title of one of his 
publicatio~ls: A7 irt CAI. Jaime wanted to put principles oE 
Aritificial Intelligence into Computer-Assisted Instruction 
(CAI) systems. He dreamed of a system which had a data 
base of knowledge about a topic matter and general 
information about language and the principles of tutorial 
instruction. The system could then pursue a natural 
tutorial dialog with a student; sometimes following the 
student's initiative, sometimes taking its own intiative, but 
always generating its statements rind responses in a natural 
way from its general knowledge. This system contrasts 
sharply with existing systems for Computer-Assisted 
Instruction in which a relatively fixed sequence of questions 
and possible reponses have to be determined for each topic. 
Jaime did construct working versions of his &ream--in a 
system which he called SCHOLAR. But he died befoi-e 
SCHOLAR reached the full realization of the dream. 
It was a pleasure to work with Jaime. His kindness and 
his enthusiasm were infectious, and the discussions we had 
with him over the years were a great stimulus to our own 
thinking. Both as a friend and a colleague we miss him 
greatly. 
Cognitive Science. This book contains studies in a new 
field we call cognitive science. Cognitive science includes 
elements of psychology, computer science, linguistics, 
philosophy, and education, but it is more than the 
intersection of these disciplines. Their integration has 
produced a new set of tools for dealing with a broad range 
REPRESENTATION AND UNDERSTANDING 
of questions. In recent years, the interdctions among the 
workers in these fields has led to exciting new 
developments in our undemtanding of intelligent systems 
and the development of a science of cognition. The group 
of workers has pursued problem$ that did not appear to be 
solvable from within any single discipline. It is too early 
to predict the future course of this new interaction, but 
the work to date has been stimulating and inspiring. It is 
our hope that this book can serve as an illustrntion of the 
type of problems Chat can be qq~roached through 
interdisciplinary cooperation. The participants in this book 
(and at the conference) represent the fields sf Artificial 
Intelligence, Linguistics, and Psychology, all of whom work 
on similar problems but with different viewpoints. The 
book focuses on the common problems, hopefully acting as a 
way of bringing these issues to the attention of all workers 
in those fields related to cognitive science. 
Subject Matter. The book contains four sections. In the 
first section, Theory of Representotion, general issues 
involved in building representations of knowledge are 
explored. Daniiel G. Bobrow proposes that solutions to a set 
of design issues be used as dimensions for comparing 
different represetltations, and he examines different forms 
such solutions might take. William A. Woods explores 
problems in representing natural-language statements in 
semantic networks, illustrating difficult theoretical issues 
by examples. Joseph D. Becker is concerned with the 
representation one can infer for behavioral systems whose 
internal workings cap not be observed directly, and he 
considers the interconnection of useful concepts such as 
hierarchical organization, system gaals, and resource 
conflicts Robert J. Bobrow and John Seely Brown present 
a model for an expert understander which can take a 
collection of data &scribing some situation, synthesize a 
contingent knoultedge structure which places the input drrta in 
the context of a larger structural organization, and which 
answers questions about the situation based only on the 
contingent knowledge structure. 
Section two, New Memory Models, discusses the 
implications of the assumption that input information is 
always interpreted in terms of large structural units derived 
REPFLESENTATION AND UNDERSTANDING 
from esporionce. Daniel G. Bobrow and Donnld A. Norman 
postulate active sciicrnnla in memory which r~fer to each 
other through use of car~tuxl-deper~derlt descriptions, and 
whir11 respo~irl both to input data and to hypothcs~s about 
structure. Benjamin J. Kuipers describes thc conc~l)t of a 
frame us a structural organizing unit for data ele~nents, nnd 
he discusses the use of these units in the context of n 
recognition system. Terry Winogrod explores issues 
i~lvolved in the controversy on rcl~rcsc~lting knowledge iu 
declarative vcrsus procedural form. Winogr~tl uses the 
concept of a frame ns a basis for the synthesis of the 
declarative and proccdural tipproaches. The frame provides 
an organizing structure on which to attach both declarative 
and procedural informhtion. 
The third section, lligher Level Structures, focuses on 
the representation of plans, episodes, and stories within 
memory. David E. Rumelhart proposes a grammar for well- 
formed stories. 'His summarization rules for stories based 
on this grammar seem to provide reasonable predictions of 
human behavior. Roger C. Schank postulates that in 
understanding paragraphs, the reader fills in causal 
connections between propositions, and that such causally 
linked chains are the basis for most human memory 
organization. Robert P Abelson defines a notation in 
which to describe the intended effects of plans, and to 
express the conditions necessary for achieving desired states. 
The fourth section, Semantic Knowledge in 
Understander Systems, describes how knowledge has been 
used in existing systems. John Seely.Brown and Richard R. 
Burton describe a system which uses multiple 
representations to achieve expertise in teachiog a student 
about debugging electronic circuits. Bonnie Nash-Webber 
describes the role played by semantics in the understanding 
of continuous speech in a limited domain of discourse. 
Allan Collins, Eleanor H. Warnock, Nelleke Aiello, and 
Mark L. Viller describe a continuation of work on Jaime 
Carbonell's SCHOLAR system. They examine how liumans 
use strategies to find reasonable answers to questions for 
which they do net have the knowledge to answer with 
certainty, and how people can be taught to reason this way. 
iii 
REPRESENTATION AND UNDERSTANDING 
Acknowledgments. We are grateful for the help of o largo 
number of people who made the conference and this book 
possible. The conference partici.pants, not all of whom are 
represented in this book, created an atmosphere in which 
interdisciplinary exploration became a joy. The people 
attending were: 
From Bolt Beranek and Newman--Joe Backer, Rusty 
Bobrow. John Brown. Allan Collins. Bill Merriam. Bonnie -. 
 ash-~ebber,   lea no; Warnock, and-Rill Woods. 
. 
From Xerox Palo Alto Research Center--Dan Bobrow, 
Ron Kaplan, Sharon Kaufn~an, Julie Lwtig, and Terry 
Winograd from Stanford University). 
From the University of Cnlifowia, Sen Diego--Don 
Norman and Dave Rumelhart. From the University of 
Texas--Bob Simmons. From Yale University--Bob Abelson. 
From Uppsala University--Eric Sandewall. 
Julie Lustig made all the arrangements for the 
conference at Pajaro Dunes, and was largely responsible for 
making it a comfortable atmosphere in which to discuss 
some very difficult tecllnical issues. Carol Van Jepmond 
was responsible for typing, editing, and formatting the 
manuscripts to meet the specifications of the pysterns used 
in the production of this book. It is thanks to her skill 
and effort that the book looks as beautiful as it does. 
June Stein did the final copy editing, made general 
corrections, and gave many valuable suggestions on format 
and layout. 
Photo-ready copy was produced with the aid of 
experimentd formatting, illustration, and printing systems 
built at the Xerox Palo Alto Research Center. We would 
like to thank Matt Heile-r, Ron Kaplan, Ben Kuipers, 
William Newman, Ron Rider, Bob Sproull, and Larry Tesler 
for their help in making photo-ready production of this 
book possible. We are grateful to the Computer Science 
Laboratory of the Xerox Palo Alto Research Center for 
making available the experimental facilities and for its 
continuing support. 
Daniel G. Bobraw 
Allan M. Collins 
March 1975 
iv 
Semantics - Discourse : Memory 
KNOWLEDGE 
Eugene Charntak 
Institute for Semantic and Cognitive Studies 
Castagnola, Switzerland 
In R. Schank, and B.L. Mash-Webber, eds., Theoretical Issues in Natural Language 
Processing, 1975, 62-51. 
Frames are static structures about one stereotyped topic 
Each frame has many statements about the topic, each expressed 
in a suitable semantic representation. The primary goal in under- 
standing is to find instances of frame statements in the discourse 
Questions about a source statement can be answered by reference to 
the frame of which it is en instance. 
Semantics - Discourse : Memory 
COGNITIVE NETWORKS AND ABSTRACT TERM1 NOLOGY 
N avid G. Hays 
Department of Linguistics 
State University of New York at Buffalo 
Journal of Clinical Computing, 3, 110-118, 1973 
By systematic application of a cognitive network or similar 
theory of knowledge the internal structure of a (medical) code can 
be improved and tools developed for different purposes. Hays's 
theory uses paradigmatic, syntagmatic, discursive, attitudinal, and 
metalingual (MTL) arcs. The MTL arcs shift level of abstraction; 
e.g., anemia is neither a fewness nor an erythrocyte but an abstract 
condition. An abstract definition can include several syntagmatic 
propositions, linked discursively. A medical term can be linked by 
MTL to definitions in different languages (clinical, ~athophysio- 
logical, etc.) 
Semantics - Discourse : Memory 
Ronald~J. Brackman 
Center for Research in Computing Technology 
Harvard University 
Cambridge, Mass 02138 
Report No. TR 6-75. 
A data structure scheme for creating structured concept 
nodes tn a semantic network is presented, with structuring tech- 
niques based op a set of primitive link types including: defined 
as attribute part, modality, role, structural/condition, valuelre- 
striction, subconcept and superconcept. This structure will atore 
descriptions of bibliographic references in a way that will facili- 
tate the important processes of inference, paraphrase and analogy. 
Semantics - Discourse : Memory 
Allan Collins 
Bolt Beranek & Newman 
Cambridge, Mass. 02138 
Iai R. Schnk and B.L. Nash-Webber, eds., Theoretical Issues in Natural Language 
Processing, 1975, 52-54. 
Tulving's episodic memory is seen as a record of experiences 
and their context. However, both episodic and semantic memories 
must have similar power of representation, so their structures are 
not disttnguishable. Similarly, a lexical memory must have the 
power to represent propositional information about words. 
Thus, 
the fabric of knowledge is merely cut into different shapes. 
Semantics - Discourse : Memory 
A FORMALISM FOR RELATING LEXICAL AND PRAGMATIC INFORMATI ON: 
ITS RELEVANCE TO RECOGNITION AND GENERATION 
Aravind K. Joshi and Stanley J. Rosenschein 
The Moore School of Electrical Engineering 
University of Pennsylvania 
Philadelphia, 19174 
In: R. Schank and B.L. Nash-Webbet, eds., Theoretical Issues in Natural Language 
Processing, 1975, 79-83. 
A uniform formal structure for the interpretation of events, 
initiation of actions, understanding language, and using language 
The components of the system are CONTROL --the pro- 
cedura is component; SCHEMATA --a lattice whose points are lexical 
decompoeitions; LEXICON --non-definitional infomation; BELIEFS -- 
a closed and consistent set of statements in a predicate calculus; 
and GOALS. 
Semantics - Discourse : pemory 
Andrew Ortony 
University of Illinois at Urbana-Champaign 
In: R. Sclaank and B.L. dlash-Webber, Bds., Theoretical Issues in Natural Language 
Processing, 1975, 55-60. 
The distinction between semantic and episodic memory is not 
so much one between different kinds of memo , but one between dif- 
'K ferent kinds of knowledge. The dbtinction as been rejected, be- 
cause it is said that since we know everything Erom experience, 
there is no room for the distinction. The error lies in confusing 
knowledge from knowledge, and knowledge of knowledge. Semantic 
knowledge is knowledge that has been reorganized around concepts 
from knowledge originally encoded around events; it fs stripped of 
personal experience. One question raised by the distinction is how 
does information get into semantic memory, and how and when does it 
get 10s t from episodic memory. 
Semantics - Discourse : Memory 
~CTD-MOUTH ING FRAMES 
Jerry Feldman 
University of Rochester 
New York 
In: R. Schank and B.L. Nash-Webber, eds., Theoretical Issues in Natural knguage 
Processing, 1975, 92-93. 
There is evidence that people use three-dimensional models 
and that they integrate several views into a single model. This 
is counter to the claim that we symbolically store a large number 
of separate views. Another problem is with the assumption of de- 
fault values for slots in frames. In the extreme, this gives 
visual perception without vision. The evidence is that people can 
understand totally unexpected images presented for quite short 
pexiods. A rhird point concerns the telatively static nature of 
frames. A better model is to condtruct a goal oriented subsystem 
making use of context specific knowledge. 
Semantics - Discourse : Memory 
SOME THDUGHTS ON SCHEMATA 
Wallace L. Chafe 
Untvers it!y of ~aligornia 
Berkeley 
In: R. Scharik and B.L. Bash-Webber, eds., Theoretical Issues in Natural Language 
Processing, 1975, 89-91. 
Stories are broken down into schemata, e.g., plot plus moral. 
Questions about schemata are: what are the essential ingredients 
of a schema; are some more abstract thah others ; and how are they 
to be discovered--by imagination and intuition? 
Semantics - Discourse : Memory 
STEREOTYPES AS AN ACTOR APPROACH TOWARDS SOLVING THE PROBLEM OF 
PROCEDURAL ATTACHMENT IN FRAME THEORIES 
Carl Hewitt 
In: R.SchalJt and B.L. Nash-Webber, eds., Theoretical Issues in Natural Language 
Processing, 1975, 94-103. 
Stereotypes are actor vereions of frames. A stereotype 
has the following parts: a collection of characteristic objects, 
characteristic relations for these obj,ects and invocable plans 
for transforming the objects and relations. 
Semantics - Discourse : Mew- 
Marvin Minskv 
Artificial 16telli~ence Laboratory 
M.I.T. - --- 
Cambridge, Mas a 
In: 8. &hank and B.L. Nash-Webber, ds., Theoretical Issues in Natural Language 
Processing, 1975, J04-116. 
Frames are data 8 tructures for representing Stere~typed 
situations. Each frame contains information about how to use the 
frame, what to expect to happen next, and what to do if the expec- 
tations are not fulfilled. Lower levels of a frame have termfnals 
that can be filled by specific instances from source statements. 
Frames are linked together into a frame system and the action to 
go from one to another indicated. 
Different fraples can share the 
same terminals. Unfilled slots in instances of frames are filled 
by default optiorls from the general frame. 
Semantics - Discourse : Memory 
Roger C. Schank 
Yale University 
New Haven, Connecticut 
Int R. Schank and B.L. Nash-Webber, eds., Theoretical Issues in Natural Language 
Processing, 1975, 11 7-121. 
A SCRIPT is a structure consisting of slots and requirements 
on what can fill the slots. It is defined as a predetermined 
causal chain of conceptualizations that describe the normal sequence 
of things in a familiar situation. A SmPT header defines the 
circumstances under which a SCRIPT is called into play. 
Linguistics : Methods 
K. S. Fu and T. L. Booth 
Pufdue University Connecticut University 
IEEE Pransactions on Systems, Man and Cybernetics, SMC-5: 95-111, 1975 
Potential engineering applications. Inference algorithms for 
finite-state and context-free grammars. Application of some of the 
algorithms to the inference of pattern grammars in syntactic pattern 
recognition illustrated by examples. 
Linguistics : Dialectology 
AN ANTHROPOLOGICAL LINGUISTIC VIEW OF TECHNZCAL TERM1 NOLOGY 
Paul L. Garvin 
Department of Linguistics 
State University of New York at Buffalo 
Journal of Clinical Computing, 3, 103-109, 1973 
The health-care community has a functional dialect, with 
subdialects fhr physicians, nurses, etc. 
Anthropological study 
of the naming behavior of the community is a suitable preliminary 
step in thesaurus building. It would determine what are terms to 
be entered, how they are related, and what theoretical differences 
require alternative definitions of the same term. 
kinguistics - Methods 
SYNTACTIC RECOGNITION OF INPERFECTLY SPEC IF1 ED PATTERNS 
M. 6. Thomason and R. C. Gonzalez 
Tennessee University 
IEEE Transactions on Computers, C-24: 93-95, 1975 
Using for illustration a recognition system for chromosome 
structures, methods are developed which basically consist of apply- 
ing error transformations to the productions of context-free gram- 
mars in order to generate new context-free grammars capable of des- 
cribing not only the original error free patterns, but also patterns 
containing specific types of errors such as deleted, added, and inter- 
changed syinbofs which often arise in the pattern-scanning process. 
computation 
Yotick Wilks 
Istituto per Gli Studi 
Semantici e Cognitivi 
Castagnola, Switzerland 
Ih: R , Schank and B . L . Nash- Webber , eds . , Theore tical Issues in Natural Language 
Processing, 1975, 130-133. 
Artificial Intelligence has had at least four benefits for 
the study of natural language: (a) emphasis on complex stored struc- 
tures, (b) emphasis pn the importance of real world knowledge, (c) 
em hasis on the communicative function of sentences in context, and 
(dy emphasis on the expression of rules, structure and information 
within the operational environment. The only test of a natural 
language system is its success on a task, any demand for more theory 
must bear this in mind. Neither can recent work in A1 be regarded 
as theoretical; it is the semi-formal expression of intuition. A1 
is engineering, not a science, and as such there is no boundary to 
natural language; one counter example does not overthrow a rule sys- 
tem. Further, talk of theory distracts from heuristics. 
Cornpubation : Inference 
Raymond Rei t er 
Depattment of Computer Science 
University of BPitish Columbia 
In: R.. Schank bnd B.L. Nash-Webber, Eds., Theoretical Issues in Natural Language 
Processing, 1975, 175-179. 
There are two mechanisms for formal reasoning: (a) resolution 
pxinciple, a campetence model, b virtue of its completeness, and 
(b) natural deductive systems, w g ich are attempts to define a per- 
forrnance model for logical reasoning. A system could be designed 
that interfaces the two systems, each doing what it does best. 
Natural deductive systems have not considered fuzzy kinds of reas- 
onfng. Future questions concern other quanrifiers, concexrs for 
representing wanting, needing, etc., and the balance between com- 
putation and deduction. 
Computation : Inference 
THE COMMONSENCE ALGORITHM AS A BASIS FOR COMPUTER PIODELS OF HUMAN 
MEMORY, INFERENCE* BELIEF AND CONTEXTUAL LANGUAGE COMPREHENSION 
Ck~ck Rieger 
Department of Computer Science 
University of Maryland 
In: A. Schank and B.L. Nash-Webber, Eds., Theoretical Issues In Natural Language 
Processing, 1975, 180-195. 
Commonsense algorithms are basic structures for modeling 
human cognition. The structure is defined by specifying a set of 
links which build up large structures of nodes of five types: Wants, 
Actions, States, Statechanges and Tendencies. There are 25 primitive 
links, e.g., one-shot causality, action concurrency, inducement 
Various applications are active problem solving, basis for conceptual 
representation of language, basis of self model, etc. 
Computation : Inference 
Charles F. Schmidt 
Rutgers University 
New Brunswick, New Jersey 
In: R. Schank and B.L. Nash-WebBer, Eds., Theoretical Issues in Na tufa1 Language 
Processing, 1975, 196-200. 
A model of reasoning about human action must include (1) how 
people arrive at a plan, (2) what can count as a reason for choosing 
to perform the plan, and (3) discovering plans and motivations from 
observation or linguistic report of actions. A plan is the internal 
representation or set of beliefs about how a particular goal may be 
achieved. The belief by an observer that an actor performed one 
act to enable a second to be performed can follow neither from deduc- 
tive nor inductive reasoning. An observer may have other propositions 
that are reasons for believing or nut believing that a plan correctly 
characterizes the beliefs of the actor. An act name organizes a set 
of beliefs about how a move of this type might relate to other moves, 
and the cognitive and motivational statea of the actors. 
Proqsamming 
STRING AND LIST PROCESS I NG IN SNOBOL4: TECHNIQUES AND APPLICATIONS 
Ralph E. Griswold 
Prentice-Hall, Inc. 
Englewood Cliffs 
New Jersey 
1975 
Reviewed by Norman Badler 
Department of Computer and Informatian Science 
The Moore School of Electrical Engineering 
University of Pennsylvania 
Among popular computer progrdg languages, SNOBOL4 stands out as the only 
one offering complex pattern definition and mtching capabilities. It also has 
a flexible function definition facility and programmer-defined data types. While 
not unique, these two features encourage problem-dependent extensions of the 
language. . All ?bee aspects of SNOBOLA fom the basic tools in Griswold's new 
book 
Intended as a text for the SNC1BON user (it is not an "introductoryt' text), 
it presents techniques for the representation and manipulation of data in string, 
list, or otherwise "structured" fm. The text includes my programed examples, 
problems with a wide range of difficulty, and answers to my of these problems. 
The first Shree chapters develop pattern matching, function definition, and data 
s-tructures. The last four chapteFs examine particular application domiins: mthe- 
mtics, cryptography, document ~rpar-~or , Plus a few more specialized problems. 
Although this my seem tb ignore cqutational linguistics, the greatest imnediate 
benefit for the pmgranurter lies in the fFrst the chapters anyway. 
Within Chap 1, the section on gramws and patterns can be used for the 
inIplemntation of simple syntactic analysis. For exaqle, there is a straight- 
forward mpping of a BNf g~mmar into SNOBON patterns, but there are pitfalls 
(as well as scsne more efficient representations in the balance) that the programer 
ought to know. 
These are c'arefully explained. 
A topic that I felt was inadequately covered in Chapter 1 was the definition 
of the pattern mtching mechanism itself. The immediate presentation of examples 
using pattern rratchhg (page 2) calls for a brief overview of pattern mtchhg 
syntax and seman-tics. Surely a progrwnnw? muld appreciate not having to refer 
back to his in.troductory text should some patten function or construct bebhazy 
in his memory. Even an appendix wuld be satisfactory. In addition, this would 
support the section on ptterns as procedures by providing the underlying semantics 
for such tlprocedures. EUrther incentive for its inclusion is provided by. the 
excellent review of progt-annner-defined data types in Chapter 3. Why leave pattern 
mtching to the userJ.s recollection? 
The function definition facility discussed in Chapter 2 eraables the constructic 
of generic functions. Since there are no data type declarations for function 
argunents or parwters, often only one function is required for the execution of 
related op&ations.on various data types. The proliferation of functions in a 
complex system might therefore be systemtically reduced., The burdeg falls on the 
progrmer, of course, to sort out +he admissible combinatians or appropriate 
actions. An addition function for real and cmplex numbers is discussed, where 
the fom is a SNOBOL4 primitive and the latter is constructed from pmgrmer- 
defined data types. Although not in the realm of cqutatiowl linguistics, it 
does have a parallel, for exmple, in a function whi& inserts data into a semantic 
network and is expected to handle various chunks of netwmk as well as atomic data. 
The data type might only be determined. during prop execution; using a generic 
function avoids distracting logic within the user's primary function. 
The section on functions as generators is a little weak from the point of 
view of computational linguistic requirements fa procedures which generate suc- 
cessive alterna?ives frcmn a complex s-tructure, for example, sentence parsing or 
referent resolution. 
The use of sjmple global variables is tm lMted in these 
contexts; one often heeds to become involved with saving the values of ~e~eral 
lwal variables in special data blocks or stacking the decision pints associated 
with alternatives. The first is a well-known compiler-design technique, while 
he second involves a backbxicking control struc'hxe. In fact, an excellent 
illustration of these ideas would be an implementation of the SfJOBOL4 pattern 
matohing system in SNOBON. 
Chapter 3 is the mst useful because it desmibes how propammer-defined 
data types can be used to build "sh.u~tures~~: stacks, queues, lhked lists, 
binary trees, am3 trees. The skillful user of such representations will find a 
reduced role for complicated patterm mtchhg expressions because the implicit 
structure &coded into a string becomes dfest in the explicit links of the 
shcture. Not only is there often an economic advantage, but the semmtics of 
SNOBOL4-are easier to use than the implicit itcMracking semantics of pattern 
mtching. (Grimold himself phts this out in the section on patterns as 
procedures. The prog~mner is encouraged to consider economic trade-offs in 
the hplemehtation of stcuchuws. Often overlooked questions are addressed: 
for example, the relative merits of implementing stacks using strings, arrays, 
tables, or defined data types. Programs for the use or traversal of structures 
are also provided. 
Although exercise 3.40 requests a representation for directed graphs, neither 
hint nor answer is pvided. The canputatid linguist having an hterest in 
smtic networks or similar associative structures is thus lef-t to his own exper- 
tise. ?he basic -tree representation mst be significantly dified to incorporate 
labelled edges, a nears of eaversal (search) thmqh the edge set, and, of course, 
non-txee structures. 
Gritwold apologizes for not covering every application, 
6 8 
but the generality and current popularity of networks for the representation of 
knowledge calls for expanded treatment of the topic. 
Among the appSications covered in detail, the ones most relevant to ccsnpu- 
tational linguistics include a Mndom sentence generator (fm a pra~~~nar), an 
mcm processor, and (perhaps) a context editor. The input and output of textual 
mterial is covered in depth undw d6cument prepmtion (Chapter 6). Since the 
text does not delve into computational linguisfics pw se, the reader (or instruc- 
tor) will dften be called upon to map techniques described in the text onto his 
own problem. I think that a gkd programer muld be able to perform this 'trans- 
formation since solutions are provided for my of the basic problems in handling 
input text, setting up data structures, and traversing these structmes . 
Before you begin programming your next computational linguistics project, a 
glance thr"ough this book my save you considerable programing time and reward 
you with usable and flexible data structures. Even if you do not program in 
SNOBOL4, the techniques presented here might guide you to more efficient usage of 
other languages. On the other hand, it might convince you to try SNOBOLU. 
Proqramming 
FORTRAN TECHNIQUES 
WITH SPECIAL REFERENCE TO NON-NUMERI CAL APPLICATIONS 
A. Colin Day 
Cambridge Universf ty Press 
~ew York 
1972 
Reviewed by Richard J. Miller 
St. Olaf College 
A practical guide for the occasional Fortran TV programmer to 
the basic "tricks" and vocabulary used by the systems programmers. 
 his  boo^ ranges over topics from plotting on a line printer to hash- 
ing and basic storage structures (stacks, queues, etc.) using a con- 
cise, to-the-point writing style. This style reinforces the stated 
intention of the book, which is to heLp a programmer with a problem 
by providing descriptions of non-mathematical techniques. The style 
and intention do limit the usefulness of this book, as some of the 
topics would be well known to advanced programmers and are not cov- 
ered in sufficient depth for such a person. It is then the area be- 
tween these two extr'emes to which this book is aimed, and there it 
can be of great service. 
The only important assumption made of the reader is that be know 
the variable types of Fortran (integer, real, Hollerith, etc.) and 
their attendant foymat specifications. A good knowledge of character 
formats is especially useful, although the major use for them is in 
output statements used in the examples given in the book. It is also 
assumed that the reader knows the basic Fortran statements, but this 
is simple matter as opposed to the format and variable type problems 
which confront a Fortran programmer. 
The bodk also includes several exercises at the end of each 
chapter (answers not supplied unfortunately) and a short but very 
complete bibliography which includes several sources for each chapter. 
The book's primary value is as a source for Hints to problems encoun- 
tered during programming, providing an introduction to the'more 
sophisticated literature which can be found by starting with the 
bibliography. This book is therefore a starting point for picking 
up a basic vocabulary, techniques, and references for someone who has 
just completed a programming course or who needs a quick introduction 
to some technique which he may want to look at later in more detail. 
Computation : Information Structures 
INFORMATION SYSTEMS 
VOLUME 1 NUMBER 2 
APRI L 1975 
Hans-JochenSchneider, Editor-ln-Chief 
Institute fur Informatik 
Universitdt Stuttgart 
Herdweg 51 
D-7000 Stuttgaft 1fFedrep. Germany 
Pergam Press 
Oxford, England 
1975 
TABLE OF CONTENTS_ 
IMFORMATION AND INFOmmOPJ PROCESSTNG STRUCTURE 
Isamu Kpbayashi . . . . . . . . . . . . . . . . - . . 39 
A PARAMETRIC MODEL OF AL:TEIWATIVE FILE STRUCTURES 
Dennis G Severance .. ...,+...-..... 51 
MUDELXBG AND ANALYSXS OF DATA BASE ORGANIZATION. 
THE DOUBLY CHAINED TREE STRUCTURE 
Alfonso F. Ewdenes and James f. Sagamang .*-.. 57 
XNFORMATXQN ABOUT COMPUTER-ASSISTED INFORMATION SYSTEMS 
SPIRES - Stanford Public Information Retrieval System 
Stanford University, StanEord, California 94305 7 5 
GDLEM - Grosspeicher Orientierte, Listenorganisierte 
Ermittlungs Methode. SIEmNS A. G., M€lnehenfGennany 7 6 
SESAM - System for the Electronic Storage of AlpRa- 
n~teric Material. SIEMENS A. G., NunchenfGesmanp 
Computation : Pictorial systems 
ON RETRIEVING INFORMATION FROM VISUAL IMAGES 
Stephen Michael Kosslyn 
The Johns Hopkins University 
Baltimore, MD 
In: R. Schank and B.L. Nash-Webber, Ed$., Theoretical Issues in Natural language 
Processing, 1975, 146-150. 
A computer graphics metaphor is useful for human visual ima- 
gery. Analogous oroperties are found: as objects become smaller 
their constituent parts become more difficult to discern perceptual- 
ly; as more parts are added to an image it becomes more degraded 
due to capacity limitations; image6 displaying more identifiable 
details take longer to construct; images cannot be indefinitely ex- 
panded before overflowing; and the existence of decay time for an 
image which affects the time taken to construct a riep Image. 
Computation : Pictorial systems 
Zenon W. Pylyshyn 
Department of Psychology 
University of Western Ontario 
London, Canada 
In: R. SchalUr and B.L. Nash-Webber, Eds., Theoretical Issues in Natural Language 
Processing, 1975, 160-163. 
Semantic structure is relative to the process that constructs 
and uses the representatton. By positing analogue representations 
it is suggested that a process does noc need to know the rules of 
transformation, e.g., rotation, but this is impossible unless the 
analogical modelling medium intrinsically follows khe laws of 
physics, i.e., ascribing these laws to brain tissue. 
Cornputatian : Pictorial systems 
THE NATURE OF PERCEPTUAL REPRESENTATION: AN EXAMINATION OF THE 
ANALOG/PROPOS ITIONAL CONTROVERSY 
Stephen E. Palmer 
Department of Psychology 
University of California 
Berkeley 
In: R. Schank and B.L. Nash-Webber, Eds., Theoretical Issues ia Natural Language 
ProcPssi ng, 1975, 151-159- 
Sensory data is considered as having several levels of inter- 
pretation. At the sensory end, the representation is analog, and 
propositional at the cognitive end. Analog images are incorrectly 
seen as having all details of the stimulus whereas quasi-linguistic 
representations are only partial. 
The important issue is not the 
partiality but the selection, possibly information that discrimin- 
ates the object&n_c~texJ, Fokstru_c_ral information there needs 
to be a mechanism for both parts ana wholes. 
Parametric informa- 
tion can be coded componentially and explicitly, but some seems to 
function integrally. It is claimed that structural perception is 
qualitative whereas parametric perception is quantitative, but 
structural elements may have quantitative aspects--its 
strength of 
association with different groups. Although both structure and par- 
ameters are encoded relative to other information, thexe is evidence 
of preferred orientation and perspectives for parameters 
Computation : Pictorial systems 
Aaron Sloman 
Cognitive Studies Programme 
School of Social Sciences 
University of Sussex 
Brighton. England 
In: R. Schank and B.L. Nash-Webber, Eds., Theoretical Issuesoin Natural Language 
Processing, 1975, 164-168. 
The distinction between Fregean (symbolic) and analogical 
representations is that in the latter both representation and 
thing must be complex and there must be correspondence between the 
structures, whereas in the farmer case there is no need for a 
correspondence. 
Attempts to subsume either representation under 
the other have not succeeded. 
There is a mistaken belief that only 
proofs in Fregean symbolism are rigorous. 
Although analogical 
representations can sometimes be implemented using Fregean ones, 
this does not imply that they are not used. 
MEDICAL VOCABULARY 
PROCEEDINGS OF THE FIFTH BUFFALO CONFERENCE ON COMPUTERS IN MEDICINE 
October 29-31, 1973 
Published as the Journal of Clinical Computing 
Volume 3, Number 2, September 1973 
Editor-In-Chief: 
E. R. Gabrieli 
TABLE OF CONTENTS 
EDITORIAL: COMPUTER-COMPATIBLE, STABLE AND CONTROLLED 
MEDICALVOCABULARY. E. R. Gabrieli ............. 82 
CONFERENCE OPENING. Robert L. Ketter ' ' ' ' ' ' ' ' ' ' ' ' ' 8 3 
COMPUTERS AND NATURAL LANGUAGE. A. W. Pratt, M. G. Pacak, ... 
M. Epstein, andG. Durham ................. 85 
SOME PROGRAMMING ASPECTS OF NATURAL LANGUAGE DATA 
PROCESSING. William White ................ 100 
AN ANTHROPOLOGICAL LINGUISTIC VIEW OF TECHNICAL 
TERMINOLOGY. Paul L, Garvin ............... 103 
COGNITIVE NETWORKS AND ABSTRACT TERMINOLOGY. David G. Hays . . 110 
THE EVOLUTION OF A MEDICAL VOCABULARY. William D. Sharpe . . 119 
CODING DIAGNOSES OF MEDICAL RECORDS: A CHALLENGE. 
J.v~n$gmond,andR.Wiem.e ................ 130 
RETRIEVAL-ORIENTED STORAGE OF MEDICAL DATA: OPERATIONAL 
ASPECTS. Charles W. Conaway ,and Edward T. O'Neill ..... 136 
PROPOSED USE IN CANADA OF SNOMED IN A MEDICAL INFORMATION 
MANAGEMENT SYSTEM. Roger A. Cote .............. 142 
SECONDARY USERS OF CLINICAL RECORDS: AN OVERVIEW 
WilliamH. Kirby, Jr. .................... 153 
THE BUREAU OF DRUGS FOOD AND DRUG ADMINISTRATION, 
SCIENTIFIC INFORMATION SYSTEMS. Alan Gelberg ....... 155 
DRUG PRODUCTS INFORMATION FILE. Frederick M. Frankenfeld . . 163 
DATA MANAGEMENT SYSTEMS AT THE SOCIAL AND REHABILITATION 
SERVICES. WebsterA. Rogers ................ 164 
PROCEEDINGS OF THE FIFTH BUFFALO CONFERENCE ON COMPUTERS IN MEDICINE 
(CONT'D). 
PSRO - A GENERAL OVERVIEW. James S. Roberts . . . . . . . . , 172 
AUTOMATED REVIEW OF PROFESSIONAL SERVICES AND THE 
PROBLEMS OF MEDICAL RECORDS. Paul Y. Ertel . . . . . . . . 177 
USES OF CLINICAL DATA IN THE NATIONAL CENTERFOR HEALTH 
STATISTICS AND POSSIBLE APPLICATION OF SNOMED. 
1waoM.Moriyama. . . . . . . . . . . , . . . . . . . . . . 185 
RADIATION EPIDEMIOLOGIC SURVEILLANCE USING THE SYSTEMATIZED 
NOMENCLATURE OF PATHOLOGY. Margaret S. Littman, 
Henry F. L'ucas Jr., William D. Sharpe, and Andrew F. Stehney 191 
SOME RELATIONSHIPS BETWEEN THE MEDICAL THESAURUS AND 
COMPUTER OPERATIONS IN A WLRGE BIBL'IOGRAPHIC CITATION 
RETRIEVAL SYSTEM. Clifford A. Bachrach . . . . . . . . . 198 
A PROGRESS REPORT. William H. Kirby, Jr. . . . . . . . . . 202 
INFORMATIOIi STORAGE AND RETRIEVAL 
Gerard Salton, Proj ect Director 
Department of Computer Science 
Cornell University 
Ithaca, New York 14853 
scientific Report No. ISR-22 
to 
The National Science Foundation 
TABLE OF CONTENTS 
4 VECTOR SPACE MODEL FOR AUTONATIC INDEXING 
G. Salton, A. Wong, and C. S. Yang 
AN INVESTIGATION ON THE EFFECTS OF DIFFERENT INDEXING METHODS ON THE 
DOCUMENT SPACE CONFIGURATION. A. Wong 
A THEORY OF TERM INPORTANCE IN AUTOMATIC TEXT ANALYSIS 
G. Salton, C. S. Yang, and C. T Yu. 
NEGATIVE DICTIONARY CONSTRUCTION. R. Crawford 
DYNAMICALLY VERSUS STATICALLY OBTAINED INFORMATION VALUES 
A. van der Meulen 
AUTOMATIC THESAURUS CONSTRUCTION THROUGH THE USE OF PRE-DEFINED 
RELEVANCE JUDGMENTS. K. Welles 
CONTENT ANALYSIS AND RELEVANCE FEEDBACK ABSTRACT 
A, Wong, R. Peck and A. van der Meulen 
ON CONTROLLING THE LENGTH OF THE FEEDBACK QUERY VECTOR 
Karamvir Sardana 
INFORMATION STORAGE AND RETRIEVAL 
TABLE OF CONTENTS (Cont ' d . ) 
THE SHORTENING OF PROFILES ON THE BASIS OF DISCRIMINATION VALUES OF 
TERMS AND PROFILE SPACE DENSITY, M. Kaplan 
ON DYNAMIC DOCUMENT SPACE MODIFICATION USING TERM DISCRLMINATION 
VALUES. C. S. Yang 
THE USE OF DOCUMENT VALUES FOR DYNAMIG QUERY PROCESSLNG 
A. Wong and A. van der Meulen 
AUTOMATIC DOCUMENT RETIREMENT ALGORITHMS. K. Sardana 
~ocumentation : Indexing 
A THEORY OF TERM IMPORTANCE IN AUTOMATLC TEXT ANALYSIS 
G. Salton, C. S. Yang and C. T, Yu 
Department of Computer Science Dept. of Computer Sciehce 
Cornell University University of Alberta 
fthaca, NY 14853 Edmonton, Ata, Canada 
In: ~nformatlon Storage and Retrieval, Gerard Salton, Edi tot Report No. ISR-22 
November 1974 
Discrimination value analysis ranks text words in accord.. 
ance with how well they are able to discriminare the documents of 
a collection from each other. The value of a term depends on how 
much the average separation between individual documents changes 
when the given term is assigned for content identification. The 
best words are those which achieve the greatest separation. Effec- 
tive criteria are given for assigning each term to either single 
word, phrase or word group categories and for constructing optimal 
indexing vocabularies. The theory is validated by citing experi- 
mental results. 
Documentation 
R. Turn, A. S. Hoffman, T. F. Lippiatt 
Rand Corporation 
Santa Monica, California 
Report No. R-1434-ARPA, June 1974 
The general military environment. Possible uses: avionics 
equipment control, field data entry, tactical command systems, aqd 
data base management in tactical and administrative systems. New 
operational capabilities may arise from spoken language translation, 
biomedical monitoring, and speech-operated writing machines. 
Appli- 
cations areas for further research. Methodology for transferring 
this technology into operational systems. 
Docmentation : Retrieval 
AUTOMATED REVIEW OF PROFESSTONAL SERVICES 
AND THE PROBLEMS OF MEDICAL RECORDS 
Paul Y. Ertel 
Ohio State University 
Columbus 
Journal of Clinical Computing, 3, 177-184, 1973 
The Medical Advances Institute developed a system to keep 
records, select cases for review, and information about individuals 
and categories, for the quality control system now established by 
law. Over 200 quality criteria packages have been developed. They 
concern the process and result of medical care. The system screens 
each case within a day of hospital admission and frequently there- 
after. It provides a review of use of facilities and conformity to 
standards of care. 
Documentation : Retrieval 
A. Wong, R. Peck, and A. van der Meulen 
Cornell University 
Ithaca, New York 
In; Information Storage and Retrieval, Gerard Salton, Editor. Reprt No. 
ISR-22, November 1974 
Experimental results indfcate that final retrieval system 
performance, after user feedback is applied using Rocchio's al- 
gorithm, is highly dependent on the system performance of the 
initial indexing process. Thkrefore every tool which imgroves 
thc indexing performance as an outcome of the content analysis 
of natural language is beneficial because initial differences 
in a system performance are retained after user feedback is 
applied. 
THE BUREAU OF DRUGS, FOOD AND DRUG ADMINISTRATION 
SCIENTIFIC INFORMATION SYSTEMS 
Alan Gelberq 
Bureau of Drugs 
Food and Dkug Administration 
Rockville, Maryland 
Journal of Clinical Computing, 3, 355-162, 1973 
ASTRO-4 is a file of new drug applications. The Ingredients 
File lists 36,000 chemicals believed to have biological effects. 
The National Drug Code is a list of manufacturers and products. A 
file of Clinical Investigators and a file of Facilities are kept. 
Also Drug Experience and Adverse Drug Reaction, Poison control 
Center file of incidents and a Drug Product Defect file. A diction- 
ary- of adverse reaction terms is in progress. Sophisticated hard- 
ware, software, and terminological controls are in use or develop- 
ment. 
Documentation : Thesauri 
PROPOSED USE IN CANADA OF SNOMED 
IN A MEDICAL INFORMATION MANAGEMENT SYSTEM 
Roger A. Cote 
University of Sherbrooke Faculty of Medicine 
Department of Pathology 
Sherbrooke, Quebec 
Journal of Clinical Computing, 3, 142-152, 1973 
SNOP, published in 1965, is not rich enough to code problems, 
signs, symptoms, disease entities, administrative, diagnostic, and 
therapeutic procedures. SNOMed is to cover the whole. The code is 
hierarchical: Topography is organized by system or tract, Morphology 
by such categories as traumatic, neoplasm, etc., Etiology by cate- 
gories of organisms and chemicals, Normal function by metabolism, 
enzyme, etc., Abnormal function correspondingly, and Procedure by 
medical discipline. Qualifiers such as history of, laboratory diag- 
nosis, etc., are included, and terms aan be linked. 
Documen~tation : Thesauri 
J. van Egmond and R.. Wieme 
Medische Informatics Gent 
Gent; Belgium 
Journal of Clinical Computlng, 3, 130-135, 1973 
The authors' codification System splits compound diagnoses 
into units, interrelated if relevant by the grammatical operator 
l'complication of1'. The content of a unit is described with three 
sets of codes: disturbance, localization, and etiology. Represen- 
tation is mnemonic for the coder, npneric for the processor. 
Management 
Charles W. Conaway and Edward T. O'Neill 
School of Information and Library Studies 
State University of New York at Buffalo 
Journal of Clinical Computing, 3, 136-141, 1973 
Records are encoded by a clerk. The system is to give a 
physician at a terminal the current synopsis of a patient record, 
the complete record (delay of a few minutes), any facts selected 
for periodic determination in the pool of Clinical Experience; 
input is to be interactive, with verification of single statements. 
Management 
DATA MANAGEMENT SYSTEMS AT THE SOCIAL AND REHAB1 LITATION SERVICES 
Webster A. Rogers 
Division of Management Systems 
Sobial and Rehabilitation Services 
Department of Health,   ducat ion and Welfare 
Washington, D.C. 
Journal of Clinical Computihg, 3, 164-171, 1973 
To improve the management of Medicaid, which spends (predicted) 
$9 billion for 27 million persons in 1974, an information system was 
designed and installed in a pilot state. It maintains data about 
eligibility of persons, qualification (administrative) of providers, 
claims, background (e.g. normal prices); it delivers statistical 
dummaries and exception reports for managers in addition to proces- 
sing claims. 
THE CLQWNS Mi CROWORLD 
Robert F . Simmons 
Department of Computer Science 
University of Texas 
Austin 
In R. Schank and B.L. Nash-W&ber, eds., Theoretical Issues in Natural Carrguage 
Processing, 1975, 17-19. 
Sentences describing scenes centred around a clown who can 
balance and move are analyzed by an ATN parser. 
The parser pro- 
duces prop6rty list semantic structures which are adequate to 
transmit data to a package that generates the scene on a display 
screen. 
Social-~ehavioral Science 
AN HISTORICAL NOTE ON THE USE OF WORD-FREQUENCY CONTIGUITIES 
IN CONTENT ANALYSIS 
H. P.. Iker 
Rochester University 
computers and the Humanities, 8: 93-98, 1974 
Discusses the development of this form of content analysis 
in information retrieval, the social sciences, and literary analysis. 
Social-Behavioral Science 
BIBLIOGRAPHY ON SOCIAL SCI ENCLCOMPUTING 
R. E. Anderson 
Minnesota University 
Minneapolis 
Campuber Reviews, 351 247-261, 1974 
Contains 591 references in the period 1960 to 1923 
covering 
statistical analysis, simulation, text processing, and laboratory 
automation. 
Literature 
ASSOCIATION FOR 
LITERARY AND 
LINGUISTIC 
COMPUT I NG BULLETIN 
Volume 3 Number 1 
Lent Term 1975 Editor 
Joan M. Smith 
6 Sevenoaks Ave 
Hearon t1oor Stockport 
Cheshire SK4 4AW ENGLAND 
CONTENTS 
GUEST EDITORIAL: QUANTIFIZIERBARE STRUKTUREN DER SPRACHE 
I. T. Piirainen ..................... 1 
A MODEL OF A DICTIONARY INFORMATION BANK 
Lidia N. Zasorina and P. V. Silvestrov ......... 3 
THE STRUCTURE OF LEXICON. M. Alinei ............ 10 
COCOA AS A TOOL FOR THE ANALYSIS OF POETRY Wendy Rosslyn . 15 
THE AVAILABILITY OF TEXTS IN MACHINE-READABLE FORM: 
PRACTICAL CONSIDERATTONS. Joan M. Smith and L- M. Smith 19 
LITERARY STATISTICS V: ON CORRELATION AND REGRESSION 
N. D. Thomson ..................... 29 
REPORT ON THE NORDIC SUMMER SCHOOL IN COMPUTATIONAL 
LINGUISTICS: Copenhagen, 29 July - 10 August 1974 
...................... Bente Maegaard 36 
AMERICAN PHILOLOGICAL ASSOCIATION MEETINGS: 
Chicago, Illinois 28-30 December 1974. S. V. F. Waite 3 8 
............. THE STATE OF SOFTWAR3 A. C. Day. 42 
TEACHING ANCIENT Gm (WITH THE HELP OF A COMPUTER) 
D. W. Packard ..................... 45 
HOW TO BRING THE DEAD LANGUAGE TO LIFE (REPORT ON THE 
.... ALLC INTERNATIONAL MEETING, 1974) Stacey Tanner 52 
ADDRESS: ALLC 1974 Annual General Meeting. R. Busa, S.3. . 55 
Humanities :. Concordance 
INDEX THOMISTICUS 
SANCTI THOMAE AQUINATIS OPERUM OMNIUM INDICES ET CONCORDANTIAE 
Roberto Busa, S. J. 
Friedrich Fromann Verlag / Gunther Hol?boog KG, Stuttgart, 1974 - 
DM 370 per volume, half-leather 
A review of volumes 1-10 of the 26-volume Concordantia Prima 
Ford Lewis Battles 
Pittsburgh Theological Seminary 
It is appmpriate indeed that, on the seven hundredth anniver- 
sary of the death of St. Thomas Aquinas, Father Roberto Busa with 
his co-workers of the Faculty of Philosophy at the Aloisianum, 
Gallarate, Italy has begun to publish the long awaited massive con- 
cordance to his writings and to texts by other authors long assoc- 
iated with his circle. For many of us who have done lesser work in 
computerized humanistic studies, rumors and reports of Busa's enter- 
prise aroused our curiosity and, in some cases, led us also to grapple 
with the manifold problems of producing a concordance by computer. 
In studying the specifications and sampling the first ten vol- 
umes of Index Thomisticus (Seatio 11, Concordantia Prima (A-Initor)) , 
this reviewer has been reminded of his own struggle to produce a con- 
cordance to the Institutes of ~ohn Calvin (Pittsburgh, 1972) . The r . T. 
provides a hierarchically organized concordance to a literary corpus 
of 10,600,000 words of Latin Texts; by comparison, the CaLvin concor- 
dance contains 405,338 words of Latin text in a single sequence. 
Thus, the vastly greater literary task of Busa called for a series 
of basic literary and philological and logical declsions not only 
to make the enormous work of processing possible, but also to pro- 
duce a final instrument for the use of scholars that would ration- 
ally encompass the vaat corpus. 
At the outset the character of the Latin language and espec- 
ially its morphological peculiarities had to be translated into com- 
puterizable routines, so that something other than a sea of raw al- 
phabetical sdrting would result. Lematization by hand sorting 
after the basic concordancing (feasible for a small corpus), prepar- 
ation of an interlined ("glossed") lemmatized machine readable rext 
(also suitable for smaller texts), even the elaborated encoding of 
the text developed by De Latte at the Liege Centre - none of these 
methods was chosen by Busa and his associates. They turned rather 
to Forcellini ' s Lexicon tot ius ~atinlhtis and encarded the 90,000 
Forcellini lemmata (in all possible forms) plus additional ones in 
the Thornistic corpus to a total of 10,000,000 codes, put this on 
magnetic tape, and worked out procedures to apply this Latin Machine 
Dictionary (LEL) to the machine-readable text. This instrument is 
now available fox the use of others working on Latin texts. To any- 
one knowing the homographs of Latin, the limitations of any mechan- 
ical routine are apparent: the T., however, handles these problems 
in a clear and workable manner. 
The size of the literary corpus also called for basic decisions 
by successively sequestering different fractions of the corpus (in a 
way that would have doubtless intrigued Thomas himselfl), the com- 
pilers reduced the mass to manageable proportions. Their decisions 
may be set down serially. 
(1) LITERARY. First, divide the authentic works of 
Thomas (100 + 181) from those of other authorship (61): 
treat each in separate series. Secondly, extract literal 
quotations, citations in references to other authors, and 
cross-citations to other Aquinian treatises; treat these 
sepaiately. This leaves distinct layers of mat'erial for 
concor dancing.. 
(2) PHILOLOGICAL. First, separat.e out indeclinable6 
like prepositions, conjunctions, adverbs, f oms of esse , 
helping verbs, etc.. pronouns, numerals, etc., etc ; these 
will occupy a pareiaular concordance. Second, put the re- 
maining nouns, adjectives and verbs in a primary concordance: 
nbuns and adjectives arranged alphabetically by termination; 
verbs by a standard order. 
(3) LOGICAL. TO reduce the bulk of the concordance 
and to make the vacabdlary and context more rationally acces- 
sible: First, analyze out frequently used phrases (.e.g., 
liberum arbitrim, acceptio personarum, caelum et terra - 
there are about 500 of these), concordancing them under one 
word only; for example, acceptia personarum would be listed 
after all other instances of acceptio. Secondly, distinguish 
words which are either proper names or so commonly found 
that only a brief context (1% lines) need be quoted; the 
rest can then be set in a context of a lines - and the whole 
interfiled in a single series. 
These decisions have determined both the cnaraccer ana con- 
tent of Sectio 11, comprising two series (Aquinian and non-Aquin- 
ian works) of five concordances each. So much for the Cotlcordance 
proper. There remains a further instrument for the use of schol- 
ars, Sectio I. In this are included indices of diseribution, sum- 
rnarhs of the lexicon, and indLces of frequency: Through these 
lists linguistic and literary studies of all sorts can be made. 
By reverse alphabetical ordering of lemmata and forms additional 
'kinds of analysis are facilitated. Since these volumes have not 
yet appeared, they can only be briefly mentioned here. 
A massive concordance of this type must carry a concise yet: 
precise location code for each item. The editors have determined 
the proper modern edition to be used, have set a precise order of 
works to be followed In the concordancing of each type under its 
appropriate lemma, and have summarized this on a separate 4-page 
insert, to whidh the user will doubtless make frequent reference 
as he learns to use this grand instrument of research. 
A short review cannot do justice to the immense detail and 
the intelligence with which this detail has, with human and com- 
puter help, been marshalled in the I.T. Ea'ther Busa and his 
associates are to be commended not only for their achievement, 
but: also for the example they have set for other laborers. 
Future Concordances to comparable literary corpora will obviously 
have their own special features; yet Busa's method of attacking 
problems - linguistic, philological, logical, quantitative - will 
suggest analogous modes of approach to others. While an index to 
St. Thomas by no means exhausts even the whole body of Christian 
Latlnity, it provides a key to the heart of Roman theology during 
8 9 
its period of greatest fruitfulness and to classical and patristic 
thought that passed through the schoolmen's filter. Space elso 
precludes discussion of the physical aspects of concordancing and 
printing, fo~ which Father Busa had the assistance of IBM. 
In a work of such vast proportiohs, the care of men cannot 
obviate error. Some 53 errors have been noted by the compilers in 
Conoordantia Prima. But even the correction of at least one of 
these contains a minor error: B.O12(QDV)23 13.ag8/8 should read 
B.012(QDV)22.13.ag8/8. See Sectio 2 vol- 1, p. 526, co1.2. Also 
12.Tabula Syntagmatum, the list of phrases concordanced under only 
one member of the phrase (pp.xiv-xvi) has at least two errors: bona 
esteriora should read bona exteriora (p. xiv): drosperitas terrena 
should read prosperitas terrena (p. xvi) The most useful pentaglot 
descriptive booklet of 46pp. is somewhat marred in the English ver- 
sion by misprints and verbal infeliaities. But these small matters 
are quite eclipsed by the enormous accomplishment of Father Busa 
and his co-workers. 
Humanities : Concordance 
[Mbylicbkeiten der maschinellen Verarbeitung spatmittelhochdeutscher Texte. 
Berichht 
Uber eln Forschungsunternehmenl 
T. Baumgarten 
Institute for ~ommunication Resegrch and Phofietics 
Bonn University 
Computrsrs and the Humanities, 8: 85-91, 1974 
Lemmatized and classifying index, verse concordance, rhyming 
index, reverse morphological index frequency list, computer-readable 
"MHG Working Dictionary" "Syntactical Rule System' making possible a 
mechanical text description and expandable to a "descriptive grammar". 
Humanities : Analysis 
ON UNDERSTANDlNG POETRY 
D. L. Waltz 
University of Illinois 
Urbana 
In I?. Schank and B.L. Nash-Webber, eds., Theoretiad1 Issues in Natural Language 
Processing, 1975,. 20-23. 
The processing of discourse is generally organized around 
verbs. However the structure at a topical or thematic level may 
not be so organized, bearing little resemblance to a deep struc- 
ture. Analogies are used, not only in poetry, to transfer large 
amounts of information from one domain to another; to enable com- 
munications of the otherwise. inexplicable; to make distinctions 
vivid; and to understand new concepts by analogy to old ones. 
Cora Angier Sowa and John E. Sowa 
21 Palmer Avenue 
Croton-on-Hudson, New York 10520 
Computers and the Hpmunities, 8, 3: 131-146, 1974 
Analyzing associations in a literary text is analogous to 
the problem of computing term associations in document retrieval. 
This paper describes haw the theory of clumps was used to find 
clusters of closely associated words in the Homeric Hymns. For 
each cluster, the program printed a mini-concordance to the lines 
of text containing each word in the cluster. The results showed 
two types of patterns in the poet's use of words: localized word 
plays extendinq over a few lines, and global interactions between 
a cluster of words and the overall thematic structure of the text. 
Instruction 
COMPUTER AIDED INSTRUCTION 
A BIBLIOGRAPHY WITH ABSTRACTS 
E. J. Lehmann 
National Technical Information Service 
Springfield, Virginia 
Seport No. Com-74-11376/2, August 1974. Price $20.00 
Contains 252 abstracts dating from 1970 to June 1974 cover- 
ing use in education, computer system requirements, motivation, 
technical training, learning factors and human factors engineering. 
Retrieved using the National Technical Information Service on-line 
search system. 

REFERENCES 
1. Nievergelt, J., and J. C. Farrar, "\&at tlachines Can and Cannot 
DO,'' Co,~uting Surveys, 4, June 1972, 81-96. 
2. Weizenbam, J., "ELIZA--A Computer Program for the Study of Nstural 
Language Comunication Eetween Man and EIachine," Corn. ACbt 9, 
January 1966, 36-45. 
3. Weizenbaum, J. , '!Contextual Understanding by Computers ," Corn. 
AC14 10,' August 19fi7, 474-480. 
4. Bobrow, D. G., "Natural Language Input for a Computer Problem 
Solvtng System," in M. Minsky (ed.), Senantic I.r.fomation Proc- 
essicg, &11T Press, Cambridge, Flass., 1968, 135-215 
5. Raphael, B. , "SIR: Semantic Information Retrieval," in M. Ninsky 
[e& ) , Senimtic Infortxt{on Procsssi~g, NIT, Press, Cambridge, 
aass., 1968, 33-134, 256-266. 
6. Hnsky, N, (ed.) , Semantric Irrfomation Processing, >!IT Press, 
Czabridke , Nass. , 196 8. 
7 Winograd, T. , Understaxding ?Jahral Lunguage , Academic Press, 
New York, 197i. 
8. Plath, W., "Rest~icted English as a User Language," IBM T. 3. 
Watson Research Center, Yorktown Heights, New York, 3 972, 
9. Green, P. F., A. K. IJolf, C. Clonsky, and K. Laugherty, "BASEB-ALL. 
An Automatk Ques tion-Answer," in E. A. Feigenbgum and J. Feldman 
(eds. ) , Cor/rpv.ters and Thought, IIcGraw-Hill, New York, 1963. 
10. Thompson, F. B., "English for the Conputer," Proc. FJCC, Spartan, 
Kew York, 1968, 349-356. 
11. Craig, J. A., S. Berezner, H. Carney, and C. Longyear, "DE8CON: 
Direct English Access 2nd Control ," Pmc. FJCC, Spartan, New York, 
1968, 365-380. 
12. Kellogg, C. , "A Natural Language Coicpilc?r for On-Line DBCa ~M3nag.- 
ment," Proc. F~cC, Spartan, New York, ,1968, 473-492. 
13. 
Travis, L., C. Kellopg, P. Klal~r, , In'Ji.r~?~tiaZ @~aotion-Anm~l<ng; 
EZ~JL<~?!$ Cu?:I~'2>r"o, qy~ien Dsvclopnei.: Corpara::up, SP-3579, 
J-anuary 31, 1973. 
14. Kellogg, C, I&, , J. Butpi, 51. DiLler, and K, Fogt; "The COhTVEEE 
Natural Language Data Nmagexnt. System: Current Status and 
Plcms ,I1 in' J 1I;llikc.r and S. 'Rouenfcld (eils .X, Pmc. S~$F. b:fox 
rnatidn Sco~ag$ ~nd i?e tb<evaZ, University of 'P!qrylr?nd, College 
Park, April 1271, 33-16. 
15. Ullogg, e. A., Qdes5io~-Anqwsring irt ths Co~uerss S,usten, System 
Development Corporation, TPi 5015, October. 1971. 
16. Nilsson, N, J. , ~rob~en4bZirin~ I'2 thods in Arti~iciaZ IvteZZigzncs 
McGraw-Hill, Nbw York , 19 71. 
17. Rulifsoul., 3. F; , 8. J. Waldinger, and J. A. Derksen, "A Lansuage 
for, Writing Problem-Solving F'ragrans ,'I hoe. I7I.P Congr. .I972 
(presented at Ljubljana, Yugoslav$a, August 1971). 
18. Ru13.feonB Y. F.,-Qj4 Progrming Conozpts, Stanford Research 
Institute, Artificial Xntelligence Group, Technical Note 60, 
Apgust 1'971. 
19. Woods, W. A. , "Procedural Semantics for a Qttestion-Answering. 
Machine," Proc. fJCC, Spartan, New York, 1968, 457,-471. 
20.. Hsvitt, C., 
"A .Language for Theorems in Robots," Prac. Int. J0ir.t 
Coif. A~tZfici~Z In-teZZigence , Washin~to.ii, D. C., 1969, 295-301. 
21. 
Suss-, C. J., and D. V. McDerzm tt, "From PLU3TFIZ to COhTIVEX-- 
A Genetic Approach" 
Conniving is Better Than Planning"), 
Proc. 1972 FJCC, MIPS, Vol. 41, Part 11, 1171-1179. 
22. Fikes. X. E. "Monitored Executibn of Robot Plans Produced by 
STIJPS," Proc. DIP Congr. 1971 (presented at Ljubljana, 
Yugoslavia, August;1971). 
Also sea R. E. Fikes and h'. J. Nilsson, 
"STkIPS: 
A New ApProach to the AppHcatiob of Theorem.Protdn~ 
to Brpblem Solviqg," Artificial In&eZZiger.cs, 2, 1971, 189-208. 
23. Slagle, J. R., Artif{ciaZ InteZZigmcz: The ileynistiz Progrmir.; 
Asproach, McGrm-Hill, New York, 1971. 
24. 
Garvtn, P. L. (ed.) , Vatma2 Language arx? tht? Conputer, McGrac- 
Hill. New York, 1903. 
25. 
Sass, H. A., and W. 11. Wilkinson (eds .) , Co,nputar Gtcpsntation of 
Humcm Reasoning, Spartan, Washington, D.C., 1965. 
26. 
Martiris, G. R., "Tlineasions of Text Processing," Proo. 1872. F.TCC, 
AFIPS, Yo:. 41, PZ:~ .II, 831-SIO. 
27. Knuth, D. Tlze Art of Gomuter A*ogrmirzg: Vol. I: Fiartdmer,taZ 
AZgori tFns, Chap. 2 "Informtion Structure 
" Addison-Vesley , 
Reading, )lass. 1968. 
28. 
Shqpiro, S.C., T!LC !.IT?Pi-Sydkom: C\ t~tl&l %%~tdUne 6o"L Smnt~~c 
Sn~04rrati'd~t Prrocobibirlg, The Rand CQrporation, R-337-PI<, August 
1971. 
29. Levlep, R. E., and N. C, Naron, ".4 Computer System for lnf aronce 
Execution and Data-Eetxieval, L'omfn. ACI'I, 10, 11, tJ,ovei;lber 1967, 
715-721. 
30; Kochen, TI., D. PI. ?hcI;ay, M. E. Won, M. Seriven, and L. Uhr, 
Con~puL&u and Com,me~~eutdbiol~, The Rand Corporation, IUI-4065-PX, 
April 1964. 
31. Kuhns, J. L., AnawurAzg 2uob;tionb by C0rnptLteh.K: A Logk2 SXcrdy, 
The Rand' Corporation, ~1-5428-PRs December 1967. 
32, DiPaala, R., "The Solvability of the Decfsion Problem for Classes 
of Proper Formulas and. Related Results," 3 ACI1, 20, January 
1973, 112-126. 
33. Woods, PI. A,, and R, M, Kaplan, The Lui~Nr Science~ NqX~rctl Language 
~11$0lun~t311 Sr(hkrn, Bfln Report 2265, Cambridge, Hass., Septeiiber 
1971. 
34. wpods, d. A., Avr Expenimazltcr-e Pamh~g 3t~h.t~ dat Tha~aLtiabz !~e~m)t/i 
Gh2innah6, BBl1 Report: 2362, Ca~bridge, IIass., Hay 1972, 
35. 
Woods, W. h., R. Pi. %plan, and B. gash-tlebber, Tlze LUI& Scie~ccb 
NM Larzguage In jo&naR;ion Sgbim: Fit& Repa&, BBX Report 
2378, CamSriilge, :.fass., June 1972. 
36, DosteEt, B. H., and F. B, Thompson, Tke St)AtPm. 06 ?EL EbzgLbh, 
California Institute of Technology, REL Eepott 1, September 1971. 
37. Dostert, B. H. "PGL-Ah Infornation Systea for a Dynamic Environ- 
nent," REL. Repo~L2 No. 3, ~alifornia Institute of Technology, 
December 1971. 
38. Charniak, E., "Jack and Janet in Search of a Theory of lhor~ledge," 
Phoc, I1tt. 30.Dt.t Coni. Atcti~iciaG Krzte&Xges~cc, Stanford , Calif, , 
1973, 
Carbontlll, J. TI., "AT in CAI: An Artificial IntBll igence lZ~proach 
to Conputer-Assisted Instruct ion," l EEE T~116. AINL-! !(IC!L&LC Sy.5 5672 
/.!!IS- 1 la Dacember 1370, 110-202, 
Teitelmun, W. , "DO That I ileaA: The ~rogram~er's ~ssis tant ,'I 
Comp(~te'rh nrtd Ai~tamntio~z, April 1972, 8-11. 
---i- 
'C ' 
, "~oward a Progrmiiing Laboratory," P,'LI?c.  TI:^, jif(~lt C~ll$, 
I\&ti{ic iz& I;!t~PLi,lc;1ce, !!ashington, !I. C., 1969, 5-11 
Zobrist, A. I,. and P, R. Carlson, Jr., "~n Advice-Tnkin: ~he'ss 
Conputer," Scic?i~ti;Sic ~\n?,Gw~, 228, June 1973, Bf-la5. 
ITewell, A. , et . al. , Speccfi-Unde~~ +al:tit~g Sp!stemb : Fi~d !?~!:o&t 
05 a SXud~r Gkorrp, ~at~onal Technical Infornation Service, Sprins- 
field, Virginia. 
Duda, R. O., and P. Z, Hart, "Experincnts in the r,ecognitipn oE 
Hand-Printed Text: Part TI-Context Analysis," n~oc, FJCC, 
Spartan, 3ew Yolk, 1965, 1139-1145. 
Alter, R. , "Utilization of Contextuql Constraints in Au~onatic 
Speech Recognition," 1EEE T~~zI~J. Au& ~~~&co~~~cJ, AU-16, 
Elarch 6-11, 1968. 
Vicens, P., "Aspects of Speech Recognition by Conouterj'l Ph.D, 
Dissertation, Stanford University, April 1369. (Also airailable 
U, S. Dept. of Cometce Clearinnllor~se for Federal Scientific 
and n~eclinical Information, AD687720) 
D. R. Reddy,, L. D. Eman$ and P. B. ilcely, "11 bdel and a System 
for ?lachina Recognition of Speech," TEE€ Tkccr~6 . Audio U~C.~?O~CCL~S tich, 
SpeW TAJU~ ail 1972 Coa6errcnce art Speech Com(nccj?kntio~t altd 
Phoc~~~zg, Val, AU-21, pp. 229-238, June 1'973. 
V. R. Lesser, R. D. Fennell, L. D. Eman, aad D. R. Reddy, "Or,qanizatioit 
of Dearsay IT Speech Understanding System," lEEE Tri:wnctio& air 
Acouj&cb, Spceclt, ~rzd Sgrd P&ocadhril, (Spechat ?&5i(c on IEEE 
Sflnnpo~h on Speccil ?ccogr~n], t'ol. .ASSF-23, pp. 11-24 February, 
1975. 
XEEE Tm~anctioa on Acouj.tiu, S~ecc!~, and Si z.6 Rocu~h~~, 
(Special   hue on IFEE Smpobium ort Sl~zcck ccoa~i~on) Vat. ASS?-23. 
February 1975. 
f? 
51. VJ. A, Ikods, "?lotivation and Overview of SPEECI'ILIS: An Experimental 
Prototype for Speech Understanding Research," SEE€ T?m/~act-io~d OM 
Acoun.f;ics, Syecclt, and S.ig~*d P8wcus&~g, (Speciat T~siie on IEEE 
Sppo5ium on Speech Aeqognitio~) , Val. A SSP-23, pp. 2-LO, February, 
1975. 
52. W, A. Len, M. F. Xedress, and T. E. Skinner, "A Prosoclically Guided 
Speech Understanding Strategy, " TEEE 7mi~acCicrnb or: ,4l\c0~4,ttcd, Spccot'r 
mid Sigrd Pf~oce~dig (S$eccrL2 TAA~CQ. air IEEE Sylj304 LLUII OH S;~CCCJ~ 
Rcco_onLtio~l), Vot. ASSF-23, pp. JO233, Pebrtlg-lry 1375, 
53. T. B. Plartin, "~~~lications of Limited Vocabulary Recogfdtion Systems," 
TEE€ Tiransactionb oil Acoustic,$, Speech, cuzd Sgtd Rocc~~i~~g, 
(Sl~ecine Ibdue on 1EEE Sgrlpobh orL ~yic?c\l RecogLttl.i41~), VoL. ASSP-25 
February 1975. 
54. Turn, R. A,, S. iioffman, T. Lippiatt, Po.tentide Cliug AppLGwCior~ 
od Speech Undmaiuzdivtg Stj4t~4, Tile ~Lnd Corporation, R-1434, 
June 1974. 
55. 
Feldman, J. A., J. R. Lokl, D. C. SwFnehart, and l?. H. Taylor, 
"~ecent Developnents in SAIL--An Algol-Based Language for 
Art'ificial Intelligence," PXQC. 1972 FJCC, MXPS, Vol. 41, Part -11, 
1193-1202, 
56. Sci~~k, R. L. ana K. M. Colby, edb., 
hiod& 04 TJ~ougi& and 
Language, We H. Fr-eeman and Company, San Francisco, 1973. 
57. !*!ilks, 'l., "Qo llachines Understand Nore Than They Did?", 
Ntzttrvre, 
Val. 252, 22 Iiovenber, 1974, pp. 275-278. 
