Interactive Problem Solving and Dialogue 
in the ATIS Domain 1 
Stephanie Seneff, Lynette Hirschman, and Victor W. Zue 
Spoken Language Systems Group 
Laboratory for Computer Science 
Massachusetts Institute of Technology 
Cambridge, Massachusetts 02139 
ABSTRACT 
This paper describes the present status of the discourse and 
dialogue models within the MIT ATIS system, extended to sup- 
port the notion of booking a flight. The discourse model includes 
not only the resolution of explicit anaphoric references, but also 
indirect and direct references to information mentioned earlier in 
the conversation, such as a direct reference to an entry in a previ- 
ously displayed table or an indirect reference to a date, as in "the 
following Thursday." The system keeps a history table containing 
objects such as flights and dates, represented as semantic frames, 
as well as the active ticket, previously booked tickets, and previ- 
ously displayed tables. During flight reservations scenarios, the 
system monitors the state of the ticket (which is displayed to the 
user), making sure that all information is complete (by querying 
the user) before allowing a booking. It may even initiate calls to 
the database to provide additional unsolicited information as ap- 
propriate. We have collected several dialogues of subjects using 
the system to make reservations, and from these, we are learning 
how to design better dialogue models. 
INTRODUCTION 
A key goal of spoken language systems is to provide sup- 
port for interactive problem solving. To be effective in this 
role, a system should permit the user to build up a solution 
incrementally, and therefore to make implicit and explicit ref- 
erence to information from earlier parts of the system/user 
dialogue. In addition, in certain scenarios, e.g., making flight 
reservations, it is natural to have the system play an ac- 
tive role in the dialogue. It is in fact our belief that spoken 
language systems will never reach a practical level of perfor- 
mance unless special attention is paid to issues of discourse 
and dialogue. To this end, we spent considerable effort over 
the last several months augmenting our system's discourse 
capabilities, and developing a much more active role for the 
computer during the dialogue with a user. 
We can view interactive problem solving systems on a 
continuum of system/user control. At one extreme would be 
a system that answers user's questions without at any point 
1This research was supported by DARPA under Contract N00014- 
89-J-1332, monitored through the Office of Naval Research. 
asking questions or making suggestions. At the other extreme 
would be a menu system that forces the user into a very 
limited range of choices. This paper will describe an attempt 
at MIT to explore a middle ground in this continuum, in the 
context of allowing users to make flight reservations. 
There are a number of advantages to an approach which 
explores a mixed initiative system for flight reservations. First, 
it provides the user with a more focused way to interact: 
making a reservation provides a natural notion of task com- 
pletion. Second , a system that can take the initiative on 
occasion not only can provide a more flexible and natural 
interaction (e.g., asking for clarification, eliciting missing in- 
formation) but also can reduce the search space for spoken 
input, thus potentially improving performance. 
There are also a number of potential disadvantages to 
such an approach. One problem with making reservations 
is the difficulty in eliciting clean data that does not con- 
tain extraneous information ('names of travelers, VISA card 
numbers, etc.) A problem that arises when the system takes 
the initiative is that it must have substantial problem solving 
expertise; it crosses the line into an expert system, and there- 
{ore may extend beyond the scope of the Spoken Language 
Systems effort as currently defined. Third, such systems are 
much more difficult to evaluate automatically. Finally, by 
making the system more natural, we run the risk of raising 
user expectations too high. 
This paper describes the present status of the discourse 
and dialogue models within the MIT AWlS system. After de- 
scribing the models, we will illustrate some of the system's 
capabilities by way of an example. We then describe our pre- 
liminary attempts at collecting data in a booking mode, for 
which we hive included a complete dialogue elicited from one 
of our subjects. Finally, potential implications for improve- 
meats in speech recognition are discussed. 
MODELLING METHOD OLOGY 
The back-end component of the MIT ATm system has 
been completely redesigned since last June \[5\]. The main 
system is described in detail in \[4\], and will only be briefly 
354 
Frame format : 
\[name ~ype 
key1: va3.uel 
key2:value2 
°.,\] 
Fr a,me : 
\[veri~y clause 
~opic: \[flight qset number: 22\] 
predicate: \[serve predicate 
theme': \["dinner" reference 
ref~ype: meal\]\]\] 
Figure 1: Frame representation resulting from analysis of the 
parse tree for the sentence, "Does flight twenty two serve dinner?" 
mentioned here. The parser delivers a semantic frame to the 
back-end, which processes it to produce a verbal response and 
a database table for display. Processing includes a step to 
incorporate previous elements from thehistory that may still 
hold, even though they were not explicitly mentioned in the 
immediate sentence. An example frame is given in Figure 1 
for the sentence "Does flight twenty two serve dinner?" 
The system can be operated in both a non-booking and 
a booking mode. 2 In the former, when a user tries to make 
a reservation, he/she is simply informed that such a capabil- 
ity does not yet exist. In the latter, the system launches a 
reservations plan upon user request, which includes a num- 
ber of subgoals initiated by either the system or the user. 
Once a user initiates a booking, a complex series of events 
transpires, in which the computer is actively interpreting the 
state of the ticket and initiating both explicit requests to the 
user and calls to the database to provide relevant additional 
information. It also displays a facsimile of the ticket (see 
Figure 3 below), and slots get filled in as they become spec- 
ified. The computer can carry the user all the way through 
a round trip ticket, being sure to get unique flight/fare/date 
specifications for both legs, and making sure that the dates 
are not violated by fare restrictions. It also warns the user 
about the date limits for the return flight when they try to 
book a restricted fare. 
In our current system, we keep at most two distinct flight 
events in the history table. One of these refers roughly to 
the most recently mentioned set of flights requested by the 
user, appearing as a new object for reference. The other 
flight event in the history records the most recently refer- 
enced unique flight or itemized flight set, typically introduced 
when the user specifically asks for more information about 
a flight that was previously presented in a table. 3 Flight 
events are not inherited wholesale except through specific 
~We define booking as the process of acquiring all relevant informa- 
tion for an itinerary, including the fare. At the moment, no further 
action, such as seat assignment and purchase/issue of the ticket, is per- 
formed, although it could presumably be simulated at a later time. 
3This is a more limited approach than a general stack of available 
anaphoric reference such as "it" or "these flights." Instead, 
individual modifiers are inherited unless new modifiers over- 
ride their inheritance. History elements are stored in the 
standard frame format, and inheritance of a modifier usually 
amounts to simply inserting it into the appropriate frame of 
the new sentence. 
A flight that is incompletely specified in a new sentence 
inherits modifiers that are consistent with its current state. 
Defining consistency is tricky and requires knowledge of how 
information is structured in the domain. Within the ATIS 
domain, the explicit mention of a flight number is taken to 
mark a change of focus, and therefore blocks almost all in- 
heritance modifiers except source and destination. Similarly, 
the modifier "cheapest" would block an inheritance of a spe- 
cific flight number, since it implies taking a subset of a pre- 
viously mentioned set. Whenever both a new source and a 
new destination are present, all inheritance is blocked, unless 
the new sentence was a clarifier, such as "How about from 
Boston to Denver?" Of course, a modifier always blocks in- 
heritance of an entry under the same key in the history ta- 
ble. Exactly which modifiers should block which others was 
determined empirically from subject data through the data 
collection episodes. 
The history table contains not only the frames associated 
with previously mentioned noun phrases and their modifiers, 
but also the previously displayed table, the previous state of 
the ticket under development, and previously booked tickets 
or first legs of a round trip ticket. The system frequently con- 
sults the tickets, as well as other elements from the history, 
to decide what directed questions to ask the user. 
Occasionally, the history elements have to be reinterpreted 
after being inherited. This is particularly true for "return" 
flights, which can be mentioned in a number of different ways: 
using an adjective or a verb phrase modifier, with or without 
explicit mention of a source, destination, or date, and with 
or without a mention of a forward leg in the same sentence. 
It turned out to be quite difficult to make all conditions work 
out, inheriting source and destination when appropriate, and 
reversing them only if they came fl'om a history flight that 
was not also marked as a return flight. Return flights also 
inherit fare class, fare restrictions, and airline. In addition, 
a restricted weekday fare is generalized to include a com- 
patible weekend fare, and some restrictions require a min- 
imum/maximum stay restricting the return date. Finally, 
users often mention the return date early in the dialogue, in 
which case the system stores it and recalls it later, when the 
topic turns to the return leg. 
Figure 2 gives a block diagram of the control flow for 
managing discourse and dialogue. As shown in the figure, 
both the user and the computer may issue questions to the 
back-end component. These questions are processed the same 
discourse references \[I\]. However it has been sufficient for the nTIS 
domain to date. 
355 
~--,TA--~=,~ Subject Input ~ Update History ~ Computer Query L •. I ~1 ~ =a,ogue StsW---\]C..O."S,ow,.res" V- 
• g Round trip, ~- ~;~y/ .... I 
I Generate Response } 
No Pop Dialogue , State Stac,k 
Figure 2: Block diagram of control flow for managing dialogue 
and discourse. 
way, updating both the discourse and dialogue components 
accordingly. For instance, when the user has booked a partic- 
ular flight but has said nothing about fares, the system can 
simply issue the request "Show fares" to the back-end. The 
discourse history will incorporate automatically the relevant 
flight information. If a user query is ambiguous, the system 
defers calling the database until it has queried the user for res- 
olution of the ambiguity. After the computer has answered 
the user's question, it assesses the dialogue state, which is 
maintained as a stack. When the state stack is popped, the 
system may update the information contained in the ticket. 
The computer may decide at this point to take the initiative, 
anticipating the user's needs. 
Consider the example in which the user says, "Book the 
cheapest flight." The system does not immediately know 
whether it should find the cheapest one-way fare or the cheap- 
est round-trip fare. It also must remember, however, that a 
booking has been requested. The system pushes <booking> 
onto the dialogue state, followed by <resolve flight cycle>. 
This is similar to the stack-based approach described in \[2\]. 
The database query function examines the top of the dialogue 
state and finds that more information is needed before a table 
can be displayed, so it does nothing. Control now passes to 
the computer, which asks the directed question, "One way or 
round trip?" and pops the top of the dialogue stack. After 
the question has been answered by the user, the user's an- 
swer is incorporated into the flight-event object, and a table 
is displayed. Now it consults the dialogue state once again 
and finds <booking>, so it calls up the booking routine to 
fill in all the ticket information and make the next decision 
about what to ask. 
Before deciding to query the user about the one-way/round- 
trip ambiguity, the computer tries hard to infer the answer 
from history information. Of course, if this had been spec- 
ified in a previous sentence, then it would be available as a 
flight-modifier frame in the history. It is also possible that 
the table previously displayed contained only round-trip (or 
only one-way) fares, in which case it could decide based on 
the table. Finally, if the user had previously specified a re- 
turn date (as in the example below), the computer would 
assume a round-trip fare was wanted. Only when all these 
conditions fail does it resort to asking the user. 
AN EXAMPLE 
The easiest way to explain some of the discourse/dialogue 
aspects of the system is through an example of a simulated 
dialogue, as-shown in Table 14 . The dialogue involves making 
a round trip reservation fl-om Boston to San Francisco. The 
subject's first sentence indicates only a destination. Rather 
than displaying all flights to San Francisco from anywhere, 
the system asks instead a directed question for clarification of 
the source location. The subject responds with only the word 
"Boston," but the system infers "from Boston" based on the 
dialogue state. The system then immediately asks for a date, 
since this will dictate whether the flight is available on that 
day and also which fares apply. The next sentence from the 
subject, "I'll be leaving Boston next Sunday and returning 
the following Tuesday," provides two dates, the first relative 
to today's date and the second relative to the first one. The 
system updates the reference date "on the fly" so that the 
appropriate date is available for reference when the phrase 
"the following Tuesday" is processed. The system now an- 
swers by focusing on the forward leg, holding the return date 
for later reference. It also provides verbal feedback verify- 
ing its understanding of the flight conditions: "These are the 
flights from Boston to San Francisco on Sunday, January 6," 
where it has interpreted "next Sunday" as "Sunday of next 
week." Finally, it fills in the source, destination, and date on 
the ticket form being displayed continuously to the subject. 
The subject now asks a question referring to a specific 
entry in the list, "the third one." The system extracts airline, 
flight number, source and destination from the table entry, 
and constructs the appropriate semantic frame, inheriting the 
date from the history. Again it confirms understanding with 
a verbal response. 
Question 5 from the subject is a complicated one, con- 
taining two superlative adjectives. The system knows to first 
extract all non-stop flights that serve dinner 5, then to pick 
the earliest among them; and finally to find the cheapest fare 
for that flight. If these modifiers were processed in a different 
order, it would not get the right answer. Usually the system 
would ask for a one-way/round-trip disambiguation at this 
point, but since the subject has already specified a return 
date, the system assumes they want a round-trip ticket. The 
cheapest fare is an unrestricted one because the restrictions 
are failing on advance-purchase requirements, given the date. 
In Question 6, the subject says, "Book it," where "it" 
4Tables have been abbreviated somewhat so that they will fit within 
a single column on the page. 
5It inherits source, destination, and date from the flight set history 
created by Question 3. 
356 
Subject: I WANT TO GO TO SAN FRANCISCO (i) 
ATIS: Where from7 
Subject: Boston. 
ATIS: What date will you be travelling on? 
Subject: I'LL BE LEAVING BOSTON NEXT SUNDAY AND 
RETURNING THE FOLLOWING TUESDAy 
ATIS: These are the flights from Boston to 
San Franoisco on Sunday January 6. 
AIRLINE NUMBER FROM TO DEPARTURE ARRIVAL STOPS 
UNITED 21 BOS SFO 8:00 A.M. 11:31A.M. 0 
UNITED 343 BOS SFO 8:40 A.M. 1:36 P.M. 1 
AA 813 BOS SFO 11:49 A.M. 5:58 P.M. 1 
CONT 1765 BOS SFO 5:40 P.M. 11:10 P.M. 1 
UNITED 93 BOS SFO 5:45 P.M. 9:27 P.M. 0 
TWA 61 BOS SFO 6:I0 P.M. 9:37 P.M. 0 
Subject: WHERE DOES THE THIRD ONE STOP 
ATIS: American flight 813 from Boston to San Francisco 
on Sunday January 6 stops in the Tollowing places. 
(2) 
(3) 
(4) 
AIRLINE NUMBER FROM TO DEPARTURE ARRIVAL STOPS STOP AIRPORT 
AA 813 BOS SFO II:49A.M 5:58P.M. I DFW 
Subject: WHAT IS THE CHEAPEST FARE FOR THE EARLIEST 
NONSTOP FLIGHT THAT SERVES DINNER 
ATIS: This is the cheapest round-trip fare for the 
earliest non-stop flights from Boston to San Francisco 
serving dinner on Sunday January 6. 
AIRLINE NUMBER RESTRICTION ROUND TRIP FARE CLASS 
UNITED 93 NONE $1210.00 Y:COACH CLASS 
Subject: BOOK IT 
ATIS: I'll show it to you again. 
(s) 
(6) 
AIRLINE NUMBER RESTRICTION ROUNDTRIP FARE CLASS 
UNITED 93 NONE $1210.00 Y:COACH CLASS 
I'll book United flight 93 from Boston to San Francisco 
for you. Can I help you with the return flight on 
Tuesday January 87 
(7) Subject: YES PLEASE 
ATIS: These are the United flights from San Francisco 
to Boston class Y on Tuesday January B. 
AIRLINE NUMBER FROM TO DEPARTURE ARRIVAL ROUND TRIP FARE 
UNITED 92 SFO BOS 8:00A.M. 4:27P.M. $1210.00 Y:COACH 
UNITED 20 SFO BOS 1:40P.M. IO:02P.M. $1210.00 Y:COACN 
Please select a booking. 
(8) 
Subject: I'D LIKE TO BOOK UNITED FLIGHT NINETY TWO PLEASE 
ATIS: This is United flight 92 from San Francisco to Boston 
class Y on Tuesday January 8. 
AIRLINE NUMBER FROM TO DEPARTURE ARRIVAL ROUND TRIP FARE 
UNITED 92 SFO BOS 8:00A.M. 4:27P.M. $1210.00 Y:COACN 
I'll book united flight 92 from San Francisco to Boston 
for you. Can I help you with something else7 
Subject: NO THANK YOU (9) 
ATIS: Have a nice day. 
Table 1: Simulated dialogue involving a round trip flight reser- 
vation. 
Figure 3: The ticket facsimile displayed to the subject, as it 
appears after Question 6 in the simulated dialogue from Table 1. 
could be the flight or the fare. The system assumes "it" 
means the entire noun phrase in the system's answer ("the 
cheapest round-trip fare for the earliest non-stop flights"), 
i.e., the fare along with the flight restrictions implied by the 
for-phrase. The system then fills in the appropriate slots 
in the displayed ticket, including airline, flight number, de- 
parture and arrival times, far e category, and dollar amount. 
When the system displays the table this time, it says, "I'll 
show it to you again," rather than the usual "This is the <fare 
with appropriate description>." This represents an attempt 
to reduce the verbose nature of the computer responses, done 
only on the condition that the sentence about to be spoken is 
identical to the one the computer just said. Figure 3 shows a 
reproduction of the ticket as it appears after Question 6 has 
been processed. 
By examining the ticket, the system determines that there 
are both a unique flight and a unique fare available for book- 
ing. Had there only been flight information specified, the 
system would have taken the initiative to display the fare 
Options for that flight on that date, and to ask the subject to 
"narrow down the fare to a single choice." The system now 
says, "I'll book United flight 93 from Boston to San Francisco 
for you," thus renaming "the earliest non-stop flight serving 
dinner." 
At this point, the system reminds the subject that there 
is a return leg, and also that the date, "Tuesday January 8" 
had been previously specified. Even if the subject had not 
mentioned a return date, the system would still ask whether 
it could help with the return flight. Furthermore, when the 
subject selects a restricted fare, the system warns them about 
the earliest and latest dates they are allowed to return on, 
and rejects a return date whenever it is outside this range. 
The subject has two general options for a "yes" answer to 
the question, "Can I help you with the return flight on Tues- 
day January 87" One is a simple, "Yes please," in which case 
the system Constructs the appropriate restrictions based on 
the ticket slots. The other is a direct statement explicitly ask- 
ing for return flights, such as "Show me the return flights," or 
"I'd like to see flights returning on January ninth," in which 
case it inherits appropriate information from the semantic 
357 
frame in the history table, reversing source and destination. 
The system now shows the subject two United flights from 
San Francisco to Boston, and asks the subject to select one. 
The subject selects flight 92, and the system is now ready to 
form a complete booking consisting of two flights tied to a 
single fare, adding it to a list of previous bookings. The sys- 
tem finally asks the subject, "Can I help you with something 
else?" and the subject concludes the dialogue. 
COLLECTING DIALOGUE DATA 
We have collected several thousand sentences from sub- 
jects using our system \[3\], but only about ten of the subjects 
were allowed to use the system in booking mode. Even for 
these ten, we only asked them to do one booking scenario, in 
addition to several non-booking scenarios. In part, this was 
done because data collection at TI is done in non-booking 
mode, and we wanted our data to be better matched to the 
likely TI test data. In addition, we were not confident that 
our booking dialogue was sufficiently robust to be ready for 
data collection until the last month or so. We are encour- 
aged, however, by preliminary results of the booking scenar- 
ios. While a subject almost never gets through a booking 
"without a hitch," we do find that subjects are able to spec- 
ify flights to be booked and successfully complete scenarios. 
With each data collection episode, we gain new insights on 
faulty assumptions in the system. A close coupling between 
data collection and system development should ultimately 
yield a robust dialogue model, which could possibly be viewed 
as an "artist's conception" of a useful system. 
In the appendix can be found an example of an actual di- 
alogue between one of our subjects and the computer. Twice 
during the dialogue the computer made faulty assumptions, 
but the subject was able to recover from the error and ulti- 
mately achieve all the goals of the scenario. Obviously, we 
are improving the system so that, next time around, these 
errors will not reccur. 
HELPING THE RECOGNIZER 
There are potentially many ways to use discourse and di- 
alogue to make the speech recognition task more successful. 
A simple step that we are taking is to restrict flight num- 
bers to be only those that have previously been displayed in 
a table. Hypotheses with unlicensed flight numbers would 
be pruned away. This can be effective, since numbers are 
relatively difficult to recognize correctly. 
Another feature that we have implemented has to do with 
ease of recovery from a recognition error. A system which re- 
members the history can become quite confused if it remem- 
bers false information from sentences that were incorrectly 
recognized. We have therefore implemented a "scratch that" 
or "erase that" command which allows the system to com- 
pletely "forget" all information that was newly introduced in 
the most recent sentence. This includes erasing entries from 
the ticket display if the sentence requested a booking. This 
command is distinct from a "clear history" command, which 
has a more global effect of erasing all records in the history. 
Finally, we hope to be able to use the dialogue state to dy- 
namically modify probabilities on arcs in the grammar. Once 
we have a fully integrated recognizer, with parse probabili- 
ties incorporated into the scores of partial theories, perplex- 
ity can be reduced by rewarding paths that are supported by 
the dialogue state. For example, we could introduce a bonus 
on the <date> node; whenever the system asks the question, 
"What date will you be travelling on?" We hope to be able 
to explore some of tlhese ideas in the near future. 
SUMMARY 
This paper describes our current model for discourse and 
dialogue within the ATIS system. Heavy emphasis has been 
placed on the specific goal-directed task of making flight 
reservations. This is in part because we believe that this par- 
ticular capability represents the most useful possible outcome 
of our efforts in developing a system based on the Official Air- 
line Guide, potentially reaching a large population of users 
with no prior expert training. But in part this choice was 
also made because it offers a very well-defined goal with a 
number of well-defined subgoals, ideally suited for exploring 
issues in interactive human/computer dialogues. The task is 
nontrivial, but the potential pay-off is large, both in terms of 
reducing the complexity of the recognition task and improv- 
ing user satisfaction. We believe that it would be particularly 
worthwhile to narrow down the ATIS system to be focused 
only on making flight reservations, discarding, for example, 
all references to ground transportation and questions probing 
the database for other types of information, such as capacity 
of aircraft, wingspan, etc. If we can demonstrate a system 
that actually works with speech input and speech/text/tables 
as output, then such a system should stand as a convincing 
demonstration of the utility of spoken language systems. 
REFERENCES 
\[1\] Dahl, D. A. and C. N. Ball, "Reference Resolution in PUN- 
DIT," Chapter 8 in Logic and Logic Grammars for Language 
Processing, P. Saint-Dizier and S. Szpakowicz, Ed., Ellis Hor- 
wood, New York, 1990. 
\[2\] Grosz, B. J. and C. L. Sidner, "Attention, Intentions, and 
the Structure of Discourse," Computational Linguistics, Vol. 
12, No. 3,, PP. 175-204, July-September, 1986. 
\[3\] Polifroni, J., S. Seneff, V. W. Zue, "Collection of Sponta- 
neous Speech for the nTIS Domain and Comparative Analy- 
ses of Data Collected at MIT and TI," These Proceedings. 
\[4\] Zue, V., J. Glass, D. Goddeau, D. Goodine, L. Hirschman, H. 
Leung, M. Phillips, J. Pofifroni, and S. Seneff, "Development 
and Preliminary Evaluation of the MIT ATIS System," These 
Proceedings. 
\[5\] Zue, V. J. Glass, D. Goodine, H. Leung, M. Phillips, J. Po- 
lifroni, and S. Seneff, "Preliminary ATIS Development at 
MIT," Proc. Third DARPA Speech and Natural Language 
Workshop, Hidden Valley, PA, June 1990. 
358 
APPENDIX: A dialogue between a subject and the computer, from our data collection files. 
SCEHARIO: You live in Philadelphia. You need to make a business trip to San Francisco next week. You have 
an old friend in Dallas and you'd therefore like to spend the afternoon in Dallas on your way to San 
Francisco. You'd prefer to fly first class on American. Find out what kind of aircraft you'll be flying on. 
SUBJECT: I'D LIKE TO FLY FROM PHILADELPHIA TO SAN FRANCISCO 
ATIS: What date will you be travelling on? 
SUBJECT: NEXT TUESDAY 
ATIS: These are the flishts from philadelphia to san francisco on Tuesday January 16 
(four flights) 
SUBJECT: DO ANY STOP IN DALLAS 
ATIS: I'm sorry I got stuck just before the words "STOP" "IN" 
SUBJECT: IS THERE A FLIGHT WHICH STOPS IN DALLAS 
ATIS: This is the with-stop flight from philadelphia to san francisco stopping in dallas on Tuesday January 16 
(one flight) 
SUBJECT: WHICH FLIGHTS GO FROM PHILADELPHIA TO SAN FRANCISCO ON MONDAY 
ATIS: These are the flights from philadelphia to san francisco on Monday January 15 
(four flights) 
SUBJECT: WHICH FLIGHTS GO FROM PHILADELPHIA TO DALLAS ON TUESDAY 
ATIS: These are the flights from philadelphia to dallas on Tuesday January 16 
(seven flights) 
SUBJECT: ARE THERE FLIGHTS FROM DALLAS TO SAN FRANCISCO OH TUESDAY AFTERNOON. 
ATIS: These are the flights from dallas to san francisco in the afternoon on Tuesday January 16 
(four flights) 
SUBJECT: I'D LIKE TO TAKE AMERICAN AIRLINES FLIGHT FOUR EIGHT SEVEN. (MISUHDERSTANDING) 
ATIS: There are no american flights 487 from dallas to san francisco on Tuesday January 16 
I'm sorry, you'll have to make another selection. 
SUBJECT: I'D LIKE TO TAKE AMERICAN AIRLINES FLIGHT FOUR EIGHTY SEVEN FROM PHILADELPHIA TO DALLAS ON TUESDAY. 
ATIS: This is american flight 487 from philadelphia to dallas on Tuesday January 16 
(one flight) 
These are the fares for american flight 487 from philadelphia to dallas on Tuesday January 16 
(two fares) 
Please narrow down the fare Co a single choice. 
SUBJECT: FIRST CLASS ONE WAY 
ATIS: This is the one-way fare for first class american flight 487 from philadelphia to dallas on Tuesday January 16 
(one fare) 
I'll book american flight 487 from philadelph~ ~ to dallas fort-worth for you. 
Can I help you with something else? 
SUBJECT: I'D LIKE TO TAKE AMERICAN AIRLINES FLIGHT EIGHT THIRTEEN FROM DALLAS TO SAN FRANCISCO ON TUESDAY. 
ATIS: This is american flight 813 from dallas to san francisco on Tuesday January 16 
(one flight) 
These are the fares for american flight 813 from dallas to san francisco on Tuesday January 16 
(tWo fares) 
Please narrow down the fare to a single choice. 
SUBJECT: FIRST CLASS ONE WAY 
ATIS: This is the one-way fare for first class american flight 813 from dallas to san francisco on Tuesday January 16 
(one fare) 
I'll book american flight 813 from dallas fort-worth to san-francisco for you. 
Can I help you with something else? \ 
SUBJECT: I'D LIKE TO FLY FROM SAN FRANCISCO BACK TO PHILADELPHIA OH THURSDAY. 
ATIS: These are the flights from san francisco to philadelphia on Thursday January 18 
(four flishts) 
SUBJECT: COULD YOU PLEASE BOOK AMERICAH AIRLINES ONE OH SEVEN FOUR FIRST CLASS 
ATIS: This is first class american flight 1074 from san francisco to philadelphia on Thursday January 18 
(one flight) 
One-way or round-trip? 
SUBJECT: ONE WAY 
ATIS: This is first class one-way american flight 1074 from san francisco to philadelphia on Thursday January 18 
(one flight) 
I'll book american flight 1074 from oakland to philadelphia for you. 
You have a roand trip fare. (MISUNDERSTANDING) 
C~n I help you with the return flight? 
SUBJECT: WHAT KIND OF AIRCRAFT IS THAT 
ATIS: This is aircraft for first class one-way american flight I074 from san francisco to philadelphia on Thursday January 18 
(one entry) 
SUBJECT: END SCENARIO EIGHT 
359 
