IR-NLI : AN EXPERT NATURAL LANGUAGE INTERFACE TO ONLINE DATA BASES 
(o) 
Giovanni Guida , Carlo Tasso (*) 
Istituto di Matematica, Informatica e Sistemistica 
Universit~ di Udine 
Via Antonini,8 - 33100 Udine, Italy 
ABSTRACT 
Constructing natural language interfaces to 
computer systems often requires achievment of ad- 
vanced reasoning and expert capabilities in addition 
to basic natural language understanding. In this pa- 
per the above issues are faced in the frame of an 
actual application concerning the design of a natu- 
ral language interface for the access to online in- 
formation retrieval systems. After a short discus - 
sion of the peculiarities of this application, which 
requires both natural language understanding and 
reasoning capabilities, the general architecture and 
fundamental design criteria of a system presently 
being developed at the University of Udine are then 
presented, The system, named IR-NLi, is aimed at al- 
lowing non-technical users to directly access through 
natural language the services offered by online da- 
ta bases. Attention is later focused on the basic 
functions of IR-NLI, namely, understanding and dia- 
logue, strategy generation, and reasoning. Knowledge 
represenetation methods and "~igorithms adopted are 
also illustar~ed. A short example of interaction 
with IR-LNT~I is presented. Perspectives and direcZions 
for future research are also discussed. 
I INTRODUCTION 
Natural language processing has developed in 
the last years in several directions that often go 
far beyond the original goal of mapping natural laua- 
~uage expressions into formal internal representa - 
zions. ?roblems concerned with discourse modeling, 
reasoning about beliefs, knowledge and wants of 
speaker and hearer, expiicitely dealing with goals, 
plans, and speech acts are only a few of the topics 
~f current interest in the field. This paper is con- 
:erned with one ~spect of natural language processing 
that we name here reasoning, it is intended as a ba- 
sic • ctivity in natural language comprehension that 
is aimed at capturing spes/~er's goals and inten - 
tions that often lie behind the mere literal mea- 
ning of the utterance. In this work we explore the 
main implications of reasoning in the frame of an 
acttu&l application which is concerned with the na- 
tttral language acees to online information retrie- 
val services (Politt,1981; Waterman,1978). 
In particular, we shall present the detailed 
design of a system, named IR-NLI (Information Re- 
trieval Natural Language Interface), that is being 
developed at the University of Udine and we shall 
discuss its main original features. The topic of 
natural language reasoning is first shortly i!lu- 
strated from a conceptual point of view and compa- 
red to related proposals. The main features of the 
chosen application domain are then described and 
the specifications of ~R-NLI are stated. We later 
turn to the architecture of the system and we go 
fturther into a detailed account of the structure of 
its knowledge bases and of its mode of operation. 
Particular attention id devoted to the three funda- 
mental modules STARTEGY GENERATOR, REASONING, ~nd 
UNDERSATNDING AND DIALOGUE. A sample search session 
with IR-NLI concludes the illustarion of the pro- 
ject. A critical evaluation of the work is zhen pre- 
sented, and main lines of the future development ~f 
the project are outlined with particular ~tention 
to original research issues. 
II NATURAL LANGUAGE UIi~ERST?S4DING .~ND REASONI:~G 
Research in natural langage precessin 6 has re- 
come in the last years a highly multiiisciplinzry 
topic, in which artificial inteliigence, computatio- 
nal linguistic, cognitive science, psycholo~- , ~nz 
logic share a wide set of common intrests, in tnia 
frame, reasoning i~ not a new issue. The meaning 
that we attach to this term in the zontex~ ~f this 
\[~ R±~o with : Milan Polytechnic Artificial Intelligence Project, Milano, Italy. 
' ~ ~iso with : CISM, International Center for Mechanical Sciences, Udine, Italy. 
31 
work is, nevertheless original. We distinguish in 
the natural language comprehension activity between 
a surface comprehension that only aims at represen- 
ting the literal content of a natural language ex - 
pression into a formal internal representation, and 
a leap comprehension that moves beyond surface mea- 
ning to capture the goals ~nd intentions which lie 
behind the utterance (Grosz,1979; Hobbs,1979; Allen, 
?errault,1980). The process that brings from surfa- 
ce to deep comprehension is just what we name here 
reasonin~ activity. Differntly from Winograd (1980), 
reasoning is not , in our model, something that ta- 
kes place after understanding is completed and aims 
at developing deductions on facts ~nd concepts 
%cquired. Reasoning is a basic paz~c of deep compre- 
hension and involves not only linguistic capabili- 
ties ~understanding and dialogue) but also deduc - 
tion, induction, analogy, generalization, etc., on 
common sense and domain specific knowledge. Figure 
presents a graphic representation of the basic 
relationships betwe=n understanding and reasoning, 
and clearly shows how reasoning moves the internal 
representation of an utterance from a first point, 
zorresponding to surface eoprehension, to a second 
Dne that represents ~eep comprehension. 
in the application of online information retrie- 
val that we face in this work,the above concepts are 
:~nsi!ered in the fr%me of man-machine commLLnica - 
:ion, and reasoning will mostly be concerned with 
terrsinolo~j, as we shall further explore in the next 
section. 
SPEAi(!'R ' S GOAL 
/ ~'~IATU RAL LANGUAGE EXPRESS IONS 
• . OERSTANO ING 
'~ REASON I NG ..~ 
~ur"r~lCe ~lees 
INTERNAL ~EPRESE.NTAT ION 
Figure I. Capturing speaker's goals through natu- 
ral lang,/age reasoning. 
!if A SA~MPLE APPLICATION : 
3NLIN£ INFO2MATION RETRIEVAL 
in :his section we present an application domain 
;here the topic of natural language reasoning plays 
fundamental role, namely, natural language access 
to online information retrieval services. As it is 
well known, online services allow interested users 
zo solve information problems by selecting and re- 
~riev~ng reievant io~uments stored in very large bi- 
bilogr%phic sr f~ctual data bases. 3enerail~" end-users 
ire unwilling or unable\] ~o serach ~ersonai!y and 
iirectly access these large files, but they Dften 
rely 3n the ~ssistanee ~f e skilful information pro- 
fessional, the intermediary, who h.ws how tc select 
e~prtpriate data bases end hr.. to design good search 
~%artegies for the retrieval of the desired informa- 
tion, and how ~o impiement them in e suitable formal 
luery !an~/age. Usually, the interaction between end- 
~=er %n~ intermediary begins ~ith ~ presearch inter- 
vlew aimed e~ precisely clarifying the content and 
t .e Db.jecti'zes of nhe information need. On tha base 
zf the information gathered, the inzeremdiary chooses 
32 
the most suitable data bases and, with the nei~ :f 
seraching referr~l aids such ~s thesauri, iirecnc- 
ties, etc., he devises the search ~trate~# no 0e 
executed by the information retrieval system. The 
output of the search is then evaluated by the end- 
user, who may propose K refinemen~ ~nd an interac- 
tion of the search for better matching hi3 requests. 
We claim that the intermediary's task represents 
a good example of the issues of natur~i ian~aage rea- 
soning, part~icularly for what concerns the ebliity ~f 
understanding natural language user'3 :'equest~ an! ;:' 
reasoning on their linguistic and aemanzlz z'..nn:e~ 
in order :o fully :~pture user's nears ~nd gczl~. 
Besides, it has to be stressed that ~he intermedizrv 
should also posses other important skills, nh..: i_ 
expertise ~nd precise knowledge ~bout ia~a ba~e can- 
tent, organization, and indexing criteria, abcu~ 
availability and use of searching referral ai!s,abBut 
system query languages and ~coess procedures, znd 
last about how %o plot ~nd construct en adequate 
search strategy. The above illustrated :hrzcteriszlcs 
motivate the design of a natural Language expert ~y- 
stem for interfacing online ~ata bases. 3n fact, the 
!R-NLI project has among its long term goals the i~ 
plementation of a system to be interposed between the 
end-user and the information retrieval system, capa 
ble of fully substituting the intermediary's role. 
IV THE IR-NLI SYSTEM 
IR-NLI is conceived as an interactive interface 
to online information retrieval system suppoz~ing 
English language interaction. It should be able to 
manage a dialogue with the user on his information 
needs and to construct an appropiate search strategy. 
More precisely, IR-NLI is aimed at meeting the needs 
of non-technical users who are not acqua/nted with on 
line searching. For this purpose three different ca- 
pabilities are requested. First, the system has to 
be an expert of online searching,i.e, it must embed 
knowledge of the intermediary's professional skill. 
Second, it must be able of understanding natural lan- 
guage and of carrying on the dialogue with the user. 
Third, it has to be capable of reasoning on langua- 
ge in order to capture the information needs of the 
user and to formulate them with appropriate terms in 
a given formal query language. 
In the current first phase of the project we 
have considered a set of working hypotheses for IR- 
NLI : 
- it operates on just one data base; 
- it utilizes only one query language; 
- it refers to only one subject domain; 
- it is conceived only for off-line use without in- 
teraction with the data base during the search 
session. 
fire suitable sequences of understanding, dialogue, 
and reasoning functions until the internal repre.- 
sentation of the user's requests is completely ex- 
panded and validated. 
The UNDERSTANDING AND DIALOGUE module is devo- 
ted to perform activities mostly of linguistic con- 
cern. First, it has to translate the natural langu- 
age user's requests into a basic formal internal re- 
presentation (IR). Second, it manages the dialogue 
with the user by generating appropriate queries and 
by translating his replays,thus expanding the IR with 
new information. The UNDERSTANDING AND DIAILGUE ~o- 
dule utilizes for its operation a base of lin~uiszic 
kno.led~e (LK). 
The REASONING module is aimed at reasoning on 
IR in order to enlarge its content with all the in- 
formation required to generate an appropiate search 
strate~\[. It utilizes for this task a base of domain 
specific knowledge (DSK). 
The FORMALIZER module, after the STRATEGY GEf~- 
RATOR has completed its activity, constructs from 
the IR the output search stra~e~ to be executed for 
accessing the online data base. The FORMALIZER utili 
zes for its operation knowledge about the formal fan 
guage needed to interrogate the online data base and 
operates through a simple syntax-directed schema, it 
is conceived as a parametric translator capable 
of producing search strategies in several languages 
for accessing online services, such as SDC ORBIT, 
Euronet DIANE, Lockheed DIALOG, etc. 
"! SYSTSM ARCHITECTURE 
The general architecture designed for the IR- 
NLI system is shown in Figure 2. The kernel of the 
system is constituted by the 3~dEI'ZGY GENERATOR , 
which is devoted to devise the top-level choices con- 
zerning the overall operation of the syszem and to 
:cntroi their execution. It utilizes for it~ acti- 
"'ity a base 3f expert knowledge (E~K) which concerns 
the evaluation of user's requests, the managament 
3f the presearch interview, the selection of a sui- 
table approach for ~eneration of the search strate- 
gy, and ~uheduling of the activities of the lower 
Level modules :~ERSTAINDING AND DIALOGUE, REASONING, 
and FORMALIZER. The operation of the STRATEGY GENERA- 
TOR is organized around a basic sequence of steps, 
each taking into ~ccount a differnt subset of ex- 
pert rules, that r%ppiy tO different situations and 
33 
.|t~rll ~angua$1 ) userrs ~UeStS 
t 
'-°'°"°'"° J r ..... --ml A,~O ~IAt.QGUE ~nowildqQlinqulstic ! 
0". .... 
Figure 2. Functions and knowledge bases ~:" \[R-~\[L2. 
Vl KNOWLEDGE BASES 
In this section we shall illustrate the main 
features of the three knowledge bases utilized by 
the IE-NLI system. 
Let us begin with DSK. The purpose of this kno K 
ledge base is to store information about the domain 
covered by the online data base to which IR-NLI re- 
fers. This information presents two ~spects : a se- 
mantic facet concerning what concepts are in the da- 
ta base and how they relate to each other, and a lin- 
guistic one concerning how the concepts are curren- 
tly eta'pressed through appropriate termS. The struc- 
ture of DSK proposed reflects and generalizes to 
~ome extent that of classical searching referral 
aids (in particular, thesauri and subject classi- 
fications). At a logical level, it is constituted 
by a labelled directed network in which nodes re- 
present concepts and directed arcs represent rela- 
tions between concepts. Each node contains a ~erm., 
a fla~ denoting whether the term is controlled or 
not, a field that stores the post in~ count, i.e. 
the number of items in the data base in which the 
term appear, and a level number which represents 
the degree of specificity of ~he term in a hirar- 
:hical subject classification. Arcs g'n~erai!y denote 
the usual cross-reference relationships utilized 
for struc=uring thesauri; e.g., BT (broader term), 
(narrower term), RT (related term), UF (used for). 
In addition, arcs of type ne~ are provided that al- 
Low, in connection with the level numbers of nodes, 
sequential scanning of the knowledge base accor- 
iing to the currently utilized hierarchical subject 
~iassification. This s~ructure is conceived to be 
ilrec~ly obtained (possibly in a partially automa- 
tic way ~hrough appropriate data conversion programs) 
from ~vailable searching referral aids and online 
thesauri. 
Le~ us turn now to LK. This knowledge base is 
~imed at supplying all information concerning natu- 
ral language ~ha~ is needed to understand user's re- 
luests. According to the mode of operation of =he 
'XDERSTANDING A~D DIALOGUE module (see section IX), 
it 2on%ains the lexicon of the application domain 
which is currently considered. Each record of the 
lexicon 2ontains ~ word of =he language, its sem:~n- 
t~2 ~2~.~e concept, ~onnec~ive, f'anction~, ~nd its 
~e%nin~. The semantic type denotes ~he role ~f % 
~ord in a sentence; namely: 
- ienoting ~ term of zhe da=a base; 
- iefining z parz icuiar relation between different 
2.\]nc~ta in user's requests; 
- specifying ~ particular function that the user ie- 
~ires ~o obtain from the information re.ri_,al~ =z 
~ystem. 
The meaning 9f R word may be expressed zs ~ pointer 
34 
to a term of the DSK in the case of a word of type 
concept, as a special purpose procedure in ~he ca- 
se of a connective or a f~nc:ion. 
Let us note that,in order to avoid ~nuseful 
duplication of informa=ion in =he DSK and LK, a sha- 
red directory of en~r y words may be u~iiized f~r 
boLh bases. 
The purpose of EK is to contain information 
tha= concerns the professional expertise of the in- 
termediary on how to manage a search session in 0r- 
der to appropriately satisfy the information needs 
of the end-user. Its contort= is made up sf several 
classes of rules concerning the different kinds 3f 
activities performed during a search session. 3he 
general s~ructure of the rules is of the ciassical 
type !F-THEN. 
Vll STRATEGY GENERATOR 
The task of the STRATEGY JENEEATCR can be con- 
sidered from two differen~ poinzs of view : 
- an external one, tha= concerns performing in~er- 
media~j's activity; 
- an in=ernal one, that rela=es to management and 
control of REASONING and U~DERSTANDiI;G ~ 31ALC- 
GUE modules. 
On ~he base of these specifications, it mus~ embed 
exper~c capabilities and behave Rs a consultation 
system for information retrieval \[?oli:t,~9@: ..... 
basic mode of operation of this =odtule is ~rganized 
around the following four main steps tha~ reflect 
the usual practice of online information searching 
(Lancaster,~979; Meadow, Cochrane,~Od~. 
~. perform presearch interview 
2. select approach 
3. devise search startegy 
h. construct search s~ar~egy. 
The IN adopted is unique throughou: the whole 
operation of the system and it is :onszitu:e/ by z 
frame, initialized by the UNDERSTAS~ING ~ :IALSOUE 
modu_le,and then further refined and expanded cy the 
reasoning module. This fr~-me i~ &ErucEured in%z ~uc- 
frames in such % way no :ontain, :!zssi:'~ei un!er 
different headings, any information ~ha% is rele- 
vant for searching an online data base, and ~3 zii;w 
an effective pattern-matching for the seleczizn cf 
search approaches and tactics. More ~ecifizziiy, l? 
encompasses terminology about zoncepts and facetz 
present in user's requests, c~-ifi=~tizns about 
search constraints and output forma~, ~nd fi~lres 
about search objectives such &s recall and ~recisi~n 
~Meadow, Cochr~ne,1981~. 
To go further in our description, let us introduce 
precise definitions of two technical terms above 
used in an informal way : 
search approach : the abstract way of facing a search 
problem, reasoning on it, analyzing its facets, and 
devising a general mode of opera~ion for having ac- 
cess to desired informazion stored in an online da- 
ta base; 
search tactic : a move, a single step or action, in 
the execution of a search approach. 
Let us recall that a search strafe@D, is a program, 
written in an appropriate formal query lan~aa~e, for 
obtaining desired information from an online system; 
taking into account the two above definitions a 
search strate~j can be viewed as the result ~£ the 
execution of a search approach through application 
of appropriate ~earch tactics. 
Within IR-NLI, a search approach is represen- 
ted as an al~orithm that defines which tactics to 
utilize, among the available ones, and how to use 
them in the construction of a s~rate~. An approach 
is not however a fixed procedure, since it does not 
~pecify at each step which paz~cicular tactic to 
execute, but only suggests a set of candidate tac- 
tics, whose execution may or may not be fired. 
Tactics are represented at two different levels 
of abstraction : 
- an high-level representation <name, objectives> 
provided for ~se by the STRATEGY GENERATOR; 
- a low-level representation <name,reasoning actions> 
supplied for use by the REASONING modLtle. 
About ~5 tactics ~re considered, taken from the ve- 
ry rich discussion by Bates (\]979) : CORRECT, CUT, 
SELECT, EXHAUST, .REDUCE, PARALLEL, PINPOINT, BLOCK, 
$b~ER, SUB, RELATE, NEIGHBOUR, FIX, RESPELL, RESPA- 
2Z, etc. 
The operation of the STRATEGY GENERATOR is ba- 
sically pattern-directed; namely, the particular 
activities to be performed and the way in which UN- 
DERSTANDING ~D DIALOGUE and REASONING modules are 
activated are determined by the content of the cur- 
rent IR *or of some par~cs of i~), which is matched 
with zn appropriate subset of the exper~ rules. In 
:his way !~3 mode of ~peration is not strictly de- 
terminate : ~ome %ctivities may or may not be fired 
.r may be perfDrmed in !ifferent ways according ~o 
the results 3f :he pattern-matching algorit~hm. 
module STRATEGY GENERATOR 
initialize search session 
per~o~ presearch interview 
actiuage UNDERSTANDING AND DIALOGUE 
<generation of IR from first user's requests> 
<pattern-cna~ching with the cu.rr~nt It? and se- 
lection of aubfr~nes tha~ could be appropria- 
tel~ filled up with new info~naeion> 
a~iv~e UNDERSTANDING AND DIAbgGUE 
<engag~nent of suitabZe dialogue for gather~ing 
additional information about search convent 
and objectives : concepts, limitations, con- 
attaints, e~uaions, decider precision ~nd 
reca~Z> 
<aa~ansion of IR> 
seZect approach 
<selection of the approach which best fits 
search objeccive8 through pattern-marching 
bemaeen IR and high-level represencation of 
~a~ics involved in each approach> 
devise search strategy 
<pa~tarn-ma~ohing bevween vhe current IR and 
tactics involved ~n the selected approach> 
<firing of appropriate tactics> 
activate REASONING 
<as:passion of IR through execuvion of reaso- 
ning a~ions> 
activuZe UNDERSTANDING AND DIALOGUE 
<uaZidation of eurrencZy expanded IR> 
oonaCru~C search strategy 
activate FORMALIZE2 
<generation of search aCraCegy f..om fully ~x- 
pc:dad IR> 
close search session 
endmodule 
As already mentioned in section IV, in the 
first version of IR-NLI the off-line operation :f 
the system lead us go consider only the buiidin~ 
block approach; future versions of the system viii 
encompass also other classical and ~ommoniy u~ili- 
zed approaches such as successive fraction, zita - 
:ion pearl growing, most speclfic facet flrsn, ~nc. 
{Meadow, Cochr~e,1981), ~hac are more ~uizab!e fJr 
an on-line operation of the ~ystem in which iicezt 
interaction with the data base luring the ae~rcn 
session is allowed. 
VIII REASONING 
The .ctivity 3f the STRATEGY GENERATOR can now 
he repr:sented in % more ~etaiied way through ~he 
fsllowing high-lave! program : 
The REASONING module operates on the IR ani za 
aimed at precisely capturing user's ~cals ~nd neela 
through deductive and inferenniai processes. More 
specifically, the REASONING uodule has the main 
~a~k of executing the tactics fired by the S.-?..ATE- 
GY GENERATOR. Hence, from this point cf view, it 
35 
represents the actuator of the reasoning process 
devised by the STRATEGY GENERATOR. 
It utilizes the low-level representation of 
the tactics, which specifies suitable reasoning ac- 
tions expressed in terms of : 
- accessing DSK; 
- updating the I~ with the new information. 
Among the basic capabilitis of the ~ESONING mo- 
dule we consider generalization to broader terms, 
extension to related concepts, particularization 
to narrower terms, analysis of synonymi and homo- 
nymi, etc. its operation is based on special-pur- 
pose procedures that correspond to the reasoning 
actions involved in the tactics. Furthermore, 
when an action has to be performed on IR for ex- 
tending its content, validation may be requested 
from the user in order to ensure a correct mat- 
zhing betvwen his wants and system proposals. This 
is done through the U~DERSTANDING AND DIALOGUE mo- 
lule which has to gather user's agreement about 
the new terms to be introduced in the IR. 
ZX U~DERSTA/~DING .~D DIALOGUE 
Zdr.e purpose of the UNDERSTANDING AND DIALOGUE 
module is twofold : 
- =o ~ranslate user~ requests into IR; 
- to generate ~ueries to the user and to understand 
his ~nswers, i.e. to manage a bounded scope dia- 
Logue. 
The conception of this module strongly relies, 
fDr what ~oncerns ~he ,anderstanding function, un 
~he experience previously developed by the authors 
vith \]\[LI project, and is organized around the con- 
:epts of semantics-directed and goal-oriented par- 
sing ~Guida,Tasso,IQ82a).its mode of operation is 
ma~niy rule-based : ~ main parsing algorithm per- 
for=Is the most elementa~/ steps of the ~nalysis 
search in the lexicon, construction of a basic 
ten~ative internal representation, validation of 
• :'.: basic internal ~epresentation~, an~ manages a 
7%ttprn-iirected invocation of heuristic rules for 
:-eso~uti~n 9f critical ~vent~ 'e.g., ambi~ai~ies, 
9ilizpes , %naphorlc references, indirect ~9eech, 
:\[2..~ important feature 0f the understanding fun- 
~%lon iz the ability to solve critical situations 
by engaging the user in a clarification iizlDgue 
±ccivated by some of the above mentioned heuristic 
rules, to gather additional information which is 
zecessar}" to correctly ~inderstand the input natu- 
r%i \[~nguage requests. 
For what concerns the dialogue function, it 
relies on two strictly connected activities : 
- generation of a lue~j, according to some requests 
from the STRATEGY GENERATOR ~r HEASONIIIG modules, 
through assembly and completing of parametric =e:~ 
fragments stored in the UNDERST.~ND!NG '~D DIALOGb~ 
module% 
- understanding of the user's answer and refinement, 
i.e. validation, updating or completing, of the 
current IR. 
Let as s~ress that, according to the basic goal- 
oriented conception of the parsing mechanism of 
"~TDERSTANDING AND DIALOGUE module, the ur.ders~u- 
ding activity performed in the frame of :he diaio- 
gue function is strongly directed by knowledge of 
the query tha~ the system has asked the user ~nd, 
therefore, of expected information to be zap~ured 
in the answer. 
X A SAMPLE SEARCH SESSION 
In this section we present % short example of 
the basic mode of operation of IE-ZLI. Figure 
shows a sample ~ession in which, in ~ddition to 
the user-system dialog/e, parts of the .'R and the 
search strategy generated (in Euronet DI.L\'~ EL~O- 
LANGUAGE) ~re reported. The -~xample refers to the 
domain of computer science. 
iCi CONCLUSIOn; 
In the paper the main features of =he ZR-~;LI 
system have been presented. The projec~ is now en- 
tering the experimental phase =hat will be carried 
on on a VAX 11,'7~0 syszem. 
The design activity so far !evelc~ed ,3uiia, 
Tasso,1982b,1982c) has reached, 2n cur mind, a qui- 
te assessed point so that fuzure work :n -his norA-" 
will oe mainly concerned wi-h r~mo':~i -f -he restric 
=ions and working hypotheses -'onsiaered in the cur- 
rent first ~nase and with refinemenr. ::' \[z~/emen=~- 
%ion de=aLia. ~he authors also g~±n %: \[aziememt _n 
the next future a-'omple~.e prototype versicr. ~f =::z 
system to be -'onnected to a real ;nlane S~:s=e~ in 
"-he fro-me .~f a strictly application ~rle..ted inte- 
rest . 
The research activity will be focused, .'n -he 
ocher hand, on several issues c'::z= \[i,~-r'--~ f-~her 
36 
> PLEASE SPECIFY YOUR REQUEST ! 
* I would like to have information about intelligent terminals for online information 
retrieval 
IR : concept-1 : INTELLIGENT TERMINALS 
concept-2 : ONLINE INFORMATION RETRIEVAL 
> DO YOU PREFER AN HIGH RECALL OR AN HIGH PRECISION SEARCH ? 
* an high recall one 
> DO YOU HAVE ANY PARTICULAR POINT OF VIEW TO CONSIDER FOR THIS RESEARCH ? 
* yes, cost considerations 
> ANY OTHER ONE ? 
* no 
> DO YOU LIKE ONLINE OR OFFLINE PRINTING OF RESULTS ? 
* please supply offline prxnting 
> PLEASE SPECIFY THE FORMAT : WITH OR WITHOUT ABSTRACT ? 
with abstract 
IR : 
concept-3 : cost 
recall : HIGH 
output : OFFLINE, FULL FORMAT 
> BUILDING BLOCK APPROACH SELECTED 
> TRYING TACTICS : 
SELECT 
PARALLEL 
RELATE 
RESPELL 
RESPACE 
> THE CONCEPT "ONLINE INFORMATION RETRIEVAL" COULD BE SPLITTED INTO : 
- ONLINE 
- INFORMATION RETRIEVAL 
PLEASE VALIDATE 
yes 
> THE FOLLOWING ADDITIONAL TERMINOLOGY IS PROPOSED FOR THE CONCEPT "ONLINE" 
L~FfERACTIVE, REALTD\[E 
PLEASE '/ALIDATE ! 
* ok 
> THE CONCEPT "ONLINE" COULD BE EXPRESSED THROUGH THE FOLLOWING TERMS : 
ONLINE, ON LINE, ON-LINE 
PLEASE VALIDATE ! 
* ok 
> SEARCH STRATEGY \]ENERATED : 
I. FI?D I:~ELLIGENT.TERMINALS 
2. FInD ONLINE OR ON.LI~ O__RR INTERACTIVE O__RR REALTIME OR REAL.TIME 
FInD INFORMATION OR REFERENCE OR BIBLIOGRAPHICS OR RETRIEVS OR SEARCH$ 
2. FInD COSTS OR PRICE 
5. FI~ S=~ A~ S=2 ;~ND S=3 A~ S=4 
6. ?RI2~ FI 
> END OF SEARCH SESSION, BYE 
Figure 3. Sample search session with IR-NLI. 
37 
investigation. Among these we mention : 
- the development of more flexible and robust dia- 
logue capabilities, including limited justifica- 
tion of the mode of operation of ~he system (Web- 
bet,1982); 
- the study of advanced representations of ~actics 
through generalized rule structures that will al- 
low more refined matching and firing mechanisms 
(Winston,198E); 
- the design of new tactics (e.g., PATTERN, ~ECORD, 
BIBBLE (Bates,~979)) ~nd reasoning actions, that 
enable the system to keep track of previous search 
sessions ~nd to ~nalogize from experience in de- 
vising and executing a search approach. 

REFERENCES 

J. F. Allen, C. R. Perrault. 1980. Analysing intention 
in utterances. Artificial Intelligence 15,3. 

A. J. Bates. 1979. Information search tactics. J. of 
the American Society for Information Science 30, 
205-2' ". 

J. D. Grosz. 1979. Utterance and objective : issues 
in za-ur~i lan%,lage ~-ommunication. Proc. 6th int. 
Joint 7cnf. on Artificial intelligence, Tokio, 
Japan, '~67-TO76. 

.3uida, 2.7zsso. 1982a. NLI : a robust interface 
for natural l~nguage person-machine ccmmunica - 
tion. int. 7. ~f Mam.-Machine Studies 17. 

.\]uida, ".Tasa=. '982b. Natural language access to 
'.~ ine iaua ba-~_a, in R.Trappl (Ed.), Cybernetics 
~.~-~ystems ?.ese~_r~h ' North-Halland Amsterdam, 

~.~ul.~,C.:~sso. '}c2:. )nline information retrieval 
~hrou~h na.~ra I 2 n%_aa~e" Proc. Conf. on Analyzis, 
"'- ~j~%lua-lzn~-:- Man-Machine Systems, 
Baden-B~den, .;est Germ. an7, Pergamon Press, Oxford. 

J.R. Hobbs. ~gz~" l°ncversatlon as olanned behaFiour. Proc. 6th "-- --= 
............. r.z \]onf. :n A.-tificia! lnte!ii- 
~, Tokio, j=_~. 390-\]96-? 

F.:I. Lancaster. '}-}. i~fc_z..a_i:.i retrieval ~ystems. 
./=an Wiley ~ $cna, ie',,' Ycr-. 

7-~-" )!ezdo'w,P.A. is~.::r%n~ :~-~! -=asizs ~" ::.line 
-~. ucP.r. ,~l-9~; k oorls "'e ",- 

A.S.P'~Ii<t. ~98\]..-'m exne.-- system as ~ online 
search, inter.mediary. ~Sth .'n~. :nline !nf~r- 
~at-C= )\[e~r : n - ~,~._ _ ......... , =-'.gland, 25-32. 

L.C.Smi=h. "~,~0. imoli.-~:ic.~ ~f ~trifi~ial " 1~; .... in% e--. - 
gence f=r end user uae of \]nline sys-ems. Online 
Review; ~, 3~3-39~. 

D.A. Waterman. 1978. Exemplery programming in RITA. 
In D.A.Waterman, F.Hayes-Roth (Eds.), Patzern ±i- 
rected inference ~stems, Academic Press, ::~w f~rk, 
261-279. 

B. Webber, A. Jozhi. 1982. Taking the initiative in 
natural language data base interactions : justifying
why. Proc. COLING 82, Prague, Czechoslovakia, 
North-Holland Linguistic Series 47. 

T.Winograd. 1980. Wh~rt does it mean to understand 
natural language? Co~nuitive Science k, ~09-2~. 

P.H.Winston. 1982. Learnin~ by au~mentin~ rules an! 
accumulatin~ censors. Artificial intelligence La- 
boratory, Massachusetts Institute of Technol~%7, 
AIM 678, Cambridge, Massachuse~zs. 
