The Genesis of a Script for Bankruptcy in Ontological Semantics
Victor
Raskin
1,3
Sergei
Nirenburg
2,3
Christian F.
Hempelmann
1
Inna
Nirenburg
3
Katrina E.
Triezenberg
1
1
CERIAS and Linguistics Program
Purdue University
West Lafayette, IN
vraskin, hempelma, kattriez@purdue.edu
2
ILIT
University of Maryland, Baltimore County
Baltimore, MD
sergei@cs.umbc.edu
3
Onyx Consulting, Inc.
iniren_99@yahoo.com
Abstract
This paper describes the creation of a script in
the framework of ontological semantics as the
formal representation of the complex event
BANKRUPTCY. This script for BANKRUPTCY
serves as the exemplary basis for a discussion
of the general motivations for including
scripts in NLP, as well as the discovery proc-
ess for, and format of, scripts for the purposes
of processing coreference and inferencing
which are required, for example, in high-end
Q&A and IE applications.
1 Introduction
A spate of advanced new applications has called for a
massive effort in script acquisition. Conceptualized as
complex events, they have been provided for in the on-
tology since its inception (see Carlson and Nirenburg,
1990) and their format has always been reasonably well-
defined as well as constantly adjusted to the consecutive
releases (see Nirenburg and Raskin, 2003, Section 7.1.5;
cf. Moreno Ortiz et al. 2002). Throughout the early and
mid-1990s, however, lower-end NLP applications, such
as knowledge- and meaning-based MT, did not neces-
sitate a heavy use of scripts. The new generation of
higher-end Q&A and similar IE applications make it
necessary to recognize individual events and their ef-
fects as part of scripts, both because humans do and
because such recognition is necessary for establishing
(co)reference relations. Thus, in the following text, only
the availability of the BANKRUPTCY script can relate (i)
and (ii) (and thus determine whose bankruptcy it is in
the latter), which may be immediately adjacent in a text:
(i) ACME, Inc., was actually doomed the moment Jorge
Jimenez and 52 other employees were laid off without a
warning.
(ii) That bankruptcy was not, however, the last blow.
As an example, we will sketch out the creation proc-
ess of the BANKRUPTCY script. In Section 2, we will
describe the status of scripts in ontological semantics,
and in Section 3 the format of their representation. Sec-
tion 4 deals with the heuristics/discovery of the infor-
mation that goes into a script, a sort of knowledge
engineering, if you will. Section 5 presents the resulting
script BANKRUPTCY, formatted to a certain grain size of
the information discovered in Section 4. Section 6
touches briefly upon just a few of the problems script
acquisitions poses.
2 Scripts in Ontological Semantics
In order to represent the meaning of connected text, not
simply that of a sequence of ostensibly independent
sentences, several things must happen. One of the most
obvious connections across sentence boundaries is co-
reference. The TMR in ontological semantics allows for
the specification of co-reference, and special procedures
exist for treating at least facets of this phenomenon in
extant applications of ontological semantics (see Niren-
burg and Raskin, 2003, Section 8.6.1). Discourse rela-
tions among propositions can also hold across sentence
boundaries, and ontological semantics includes facilities
for both detecting and representing them as well (ibid,
Section 8.6.3).
There are, however, additional strong connections
among elements of many texts. These have to do with
the understanding that individual propositions may hold
well-defined places in “routine,” “typical” sequences of
events (often called complex events, scripts or scenar-
ios) that happen in the world, with a well-specified set
of object-like entities that appear in different roles
throughout that sequence. For example, if the sequence
of events describes a state visit, the “actors” may, under
various circumstances, include the people who meet (the
“principals”), their handlers, security personnel and
journalists, possibly, a guard of honor; the “props” may
include airplanes, airports, meeting spaces, documents,
etc. All these actors and props will fill case roles and
other properties in the typical component events of the
standard event sequence for a state visit, such as travel,
arrival, greetings, discussions, negotiations, press con-
ferences, joint statements, etc. The component events
are often optional; alternatively, some component
events stand in a disjunctive relation with some others
(that is, of several components only one may actually be
realized in a particular instantiation of the overall com-
plex event), and their relative temporal ordering may be
fuzzy.
Such typical scripts can be expressed in natural lan-
guage using expository texts or narratives, sets of the
above (indeed, one conceptual story can be “gathered”
from several textual sources), plus text in tables, pic-
tures, TV and movie captions, etc. The notion of script
is clearly recursive, as every component event can itself
be considered a script, at a different level of granularity.
The notion of script, under a variety of monikers, was
popularized in computer science by Minsky (1975),
Schank and Abelson (1977), Charniak (1972), and their
colleagues in the 1970s. However, at that time, no real-
istic-size implementation of natural language processing
using scripts could be undertaken, in part, because there
was no clear idea about the required inventory of
knowledge sources, their relations and content. Script-
based theories of semantics were proposed in theoretical
linguistics (Fillmore 1985, Raskin 1986) but were over-
shadowed by the fashion for formal semantics, which is
not descriptive in nature, focusing instead on the ele-
ments of semantics capturable by such logical devices
as quantifiers (see Raskin 1994). Moreover, the size of
the task of creating the ontological semantic knowledge
sources, which are a sine qua non of script representa-
tion, was at the time underestimated by the practitioners
and overestimated by critics. It can be said that onto-
logical semantics is a descendant of the script-oriented
approach to natural language processing, especially in
the strategic sense of accentuating semantic content, that
is, the quantity and quality of stored knowledge required
for descriptions and applications. Ontological semantics
certainly transcends the purview and the granularity
levels of the older approach as well as offering an en-
tirely different take on coverage of world and language
knowledge and on its applicability. Ontological seman-
tics has also the advantage of having been implemented,
tested, and (constantly) improved in actual applications.
In the script-based approach to processing text in-
puts, the scripts in the ontology that get instantiated
from the text input provide expectations for processing
further sentences in a text. Indeed, if a sentence in a text
can be seen as instantiating a script in the nascent TMR,
the analysis and disambiguation of subsequent sentences
can be aided by the expectation that propositions con-
tained in them are instantiations of event types that are
listed as components of the activated script. Obviously,
the task of activating the appropriate script from the
input is far from straightforward. Also, not all sentences
and clauses in the input text necessarily fit a given
script—there can be deviations and fleeting extraneous
meanings that must be recognized as such and con-
nected to other elements of the TMR through regular
discourse relations, that is, through a weaker connection
than that among the elements of a complex event.
Scripts usually describe situations with multiple
agents. Each of these agents can be said, in some sense,
to carry out their own plans that are made manifest
through the reported component events in a script. Plans
are special kinds of scripts that describe the process of
attaining a goal by an agent or its proxies. Goals are
represented in ontological semantics as postconditions
(effects) of events (namely, steps in plans or compo-
nents of general scripts). For example, if an agent’s goal
is to own a TV set, this goal would be attained on a suc-
cessful completion of one of a number of possible plans.
In other words, it will be listed in the ontology as the
postcondition (effect) of such events as BUY, BORROW,
LEASE, STEAL, MANUFACTURE. Note that the plans can
be activated only if all the necessary preconditions for
their triggering hold. Thus, the ontology, in the precon-
dition property of BUY, for example, will list the re-
quirement that the agent must have enough money (see
McDonough 2000).
Manipulating plans and goals is especially important
in some applications of ontological semantics, for in-
stance, in advice giving applications where the system is
entrusted with recognizing the intentions (goals) of an
agent or a group of agents based on processing texts
about their behavior. Goal- and plan-directed processing
relies on the results of the analysis of textual input, as
recorded in the basic TMR, as well as the complemen-
tary knowledge about relevant (complex) events and
objects and their instances, stored in the ontology and
the Fact Database (see Nirenburg and Raskin, 2003,
Section 7.2), and instantiated in the extended TMR. It is
clear that reasoning based on the entire amount of
knowledge in the extended TMR can be much richer
than if only those facts mentioned in the input texts
were used for inference making. Richer possibilities for
reasoning would yield better results for any NLP appli-
cation, provided it is supplied with the requisite infer-
ence making programs, for instance, for resolving
translation mismatches. The reason we are making a
distinction among NLP applications is the extent to
which an application depends on such capabilities. For
example, MT practitioners have typically assumed that
this application does not really need machinery for in-
ference making. This belief is clearly based on the per-
ception that acquiring the knowledge necessary to
support reasoning is prohibitively expensive or even
outright infeasible, and therefore one must make do
with simpler approaches. Of course, should MT devel-
opers be able to obtain such resources, they would use
them. Ontological semantics has among its goals that of
supplying application builders with exactly this kind of
knowledge.
A good if unusual example for a family of applica-
tions for which knowledge representation at the level of
complex events is crucial but where its integration has
so far been carefully avoided is computational humor
(see, for instance, Raskin 1996; Stock and Strapparava
2002). Here, the analysis and generation of ambigu-
ity—generally the key issue for semantics, and, ac-
cordingly NLP—is a key requirement for a text to be
funny. Computational humor aims to increase the ac-
ceptability of natural language interaction between hu-
man and machine by injecting relevant humor into
natural language interfaces. The most developed theory
of humor—not only in the framework of script-based
semantics—and the formalized model based on it (At-
tardo and Raskin 1991) have at their core the notion of
incongruity conceptualized as two partially overlapping
scripts in a relation of opposition. Earlier attempts at
computational humor have simply hardwired two com-
plex events in such a relation into templates (Raskin and
Attardo 1994; Binsted and Ritchie 1997) instead of
aiming at true generation based on an operational model
of humor. And true humor generation is impossible
without the identification of scripts in an appropriate
relation of opposition, e.g., sexual vs. religious. On this
basis, the overlap between the two scripts can be ana-
lyzed and generated, for example, in puns, where one
lexical item from each of the two scripts are brought to
overlap in one surface form through sound similarity
(Hempelmann 2003). This is just an example—esoteric,
as it may seem to some—of how the set of script-
requiring NLP applications may expand with time and
need.
Obviously, as mentioned above, in addition to the
knowledge, efficient reasoning procedures must be de-
veloped. Such procedures must conform to a number of
constraints, an example of which is the following. It is
common knowledge that, unless a limit is imposed on
making inferences from knowledge units in rich knowl-
edge bases, the inferencing process can go too far or
even not halt at all. In advanced applications, for exam-
ple, advice giving again, a good candidate for such a
limit is deriving the active goals and plans of all rele-
vant agents in the world. However, even applications
that involve more or less direct treatment of basic text
meaning, such as MT, will benefit from making fewer
inferences. There will always be difficult cases, such as
the need to understand the causal relation in The sol-
diers fired at the women and I saw some of them fall to
select the correct reference for them—in Hebrew, for
example, the choice of the pronoun (the masculine otam
or the feminine otan will depend on the gender of the
antecedent). Such cases are not overly widespread, and
a prudent system would deliberately trigger the neces-
sary inferences when it recognizes that there is a need
for them. In general, any event is, in fact, complex, that
is, one can almost always find subevents of an event;
whether and to what extent it is necessary to represent it
as a script is a matter of grain size dictated by whether
an application needs this information for reasoning, and
that, in turn, is largely determined by the nature of the
corpora in the domain served by the application.
3 Format of Scripts
Scripts are represented in ontological semantics using
the ontological property HAS-PARTS. It has temporal
semantics if it appears in events, and spatial semantics if
it appears in physical objects, e.g., to indicate that an
automobile consists of an engine, wheels, the chassis,
etc. The properties PRECONDITION and EFFECT also carry
information necessary for various kinds of reasoning
and apply to any event, complex or otherwise. Scripts
require an extension to the specification format. The
reason for that is the need to bind the case roles and
other property values in component events to establish
co-reference. Also, the HAS-PARTS slot of scripts should
allow for the specification of rather advanced combina-
tions of component events. Therefore, the format of the
filler of HAS-PARTS in scripts should allow a) Boolean
operators and, or and not, as well as IF, THEN/ELSE and
b) loop statements. Scripts also need statements about
partial temporal ordering of their components. For this
purpose, a special new property, COMPONENT-
RELATIONS is introduced.
Component events in a script have a peculiar status.
They are not regular instances of concepts, as no instan-
tiation occurs in the ontology—instantiation is one of
the two main operations in generating TMRs, the other
being matching selectional restrictions in order to com-
bine individual concept instances—but their meaning is
different from that of the general concepts to which they
are related. Thus, asking questions in the context of a
class at school is clearly different from the general idea
of asking questions. In order to represent this difference,
the notion of ontological instance is introduced. In an
ontological instance, some properties are constrained
further as compared to their “parent” concept. The con-
straints typically take the form of cross-reference to the
filler of another component event in the same script.
For reasons of clarity and convenience, instead of
describing the component events and component rela-
tions directly in the fillers of corresponding slots in the
concept specification for the complex event, we use the
device of reification by just naming them in a unique
way in that location (we identify ontological instances
by appending letters, not numbers as in the case of real
instances) and describe their content separately, at the
same level as the main script. As a result, the format of
the ontological description of a script is a set of onto-
logical concept frames.
Reification in ontological semantics is a mechanism
for allowing the definition of properties on properties by
elevating properties from the status of slots in frames to
the level of a free-standing concept frame. It is desirable
from the point of view of nonproliferation of elements
of metalanguage to avoid introducing a concept of, say
DRIVER if it could always be referred to as
DRIVE.AGENT. However, this brings about certain diffi-
culties. For example, if we want to state that somebody
is a driver of trucks, we would have to say that there is
an instance of DRIVE in which the THEME is TRUCK and
the AGENT is the person in question. There is no direct
relationship between THEME and AGENT, and it would
take a longer inference chain to realize that TRUCK is, in
fact, the value of a property of DRIVER, too, not only of
DRIVE. The more properties one would want to add to
DRIVER and not to DRIVE, the more enticing it would be
to reify the property DRIVE.AGENT and treat it as a sepa-
rate concept. In principle, we can use reification on the
fly, while building a TMR, when we need to add a
property to a property, which is prohibited in the static
knowledge sources such as the ontology and the lexi-
con. As we will see in the example below, reification
also facilitates the specification of scripts.
In the example below, we present a simplified view
of the script/complex event TEACH. As illustrated,
TEACH has as PRECONDITION two events—that the
teacher knows the material and the students do not; as
EFFECT, it has the event that the students (now) know
the material. The process of teaching is presented as
follows: the teacher presents the material to the stu-
dents, the students ask the teacher questions about this
material, and the teacher answers these questions. The
above is admittedly a gross simplification of the actual
state of affairs but will serve well for the purposes of
illustration.
The ontological instances introduced in the process
are: TEACH-KNOW-A, -B AND -C, TEACH-DESCRIBE,
TEACH-REQUEST-INFO, TEACH-ANSWER, TEACH-AFTER-A
AND -B. The constraints in these instances are all refer-
ences to fillers of slots in other components of the script
or the script itself. Reference is expressed using the tra-
ditional dot notation (m.s[.f] is read as ‘the filler of the
[facet f of the] slot s of the frame m’). Ontological in-
stances are not indexed in the Fact Repository. They
appear in appropriate slots of scripts and their fillers are
all references to fillers of other ontological instances
within the same script or the script itself. They are
PART-OF (inverse of HAS-PARTS) of the script in which
they are listed but instance-of their corresponding basic
concept, that is, TEACH-DESCRIBE-A is the first ontologi-
cal instance of DESCRIBE that is at the same time PART-
OF TEACH.
teach
is-a value communicative-event
agent sem human
default teacher
themesem knowledge
destinationsem human
default student
precondition default (teach-know-a teach-know-b)
effectdefault teach-know-c
has-partsvalue (teach-describe
repeat (teach-request-information
teach-answer)
until teach-know-c)
component-relations 
value(teach-after-a teach-after-b)
component-modalities
value(teach-modality-a)
teach-know-a
instance-of value know
patient value teach.agent.sem
theme     value teach.theme.sem
teach-know-b
instance-of value know
patient value teach.destination.sem
theme      value teach.theme.sem
teach-modality-a
typevalueepistemic
scopevalueteach-know-b
valuevalue0
teach-know-c
instance-of value know
patient value teach.destination.sem
theme      value teach.theme.sem
teach-describe
instance-of value describe
agent       value teach.agent.sem
theme       value teach.theme.sem
destination value teach.destination.sem
teach-request-information
instance-of value request-information
agent       value teach.destination.sem
theme       valueteach.theme.sem
destination value teach.agent.sem
teach-answer
instance-of value answer
agent      value teach.agent.sem
theme       value teach.request-information.theme.sem
destination value teach.destination.sem
teach-after-a
domainvalueteach-describe
rangevalueteach-request-information
teach-after-b
domainvalueteach-request-information
rangevalueteach-answer
4 Heuristics
As massive research on expert systems in the 1980s
abundantly demonstrated, knowledge engineering re-
quires human intelligence and stubbornly resists auto-
mation. Moreover, human intelligence must be devoted
to heuristics, another highly non-trivial intellectual
process. Deciding what goes into a script requires
knowledge engineering and heuristics. Part of the prob-
lem is similar to the task of extending the ontology to a
new domain—something the ontological semantics
community has had to face a number of times for vari-
ous applications, most recently for information security
applications (Raskin et al., 2002). There are three main
sources for obtaining and structuring the required in-
formation to chart out a new domain or to fill out a new
script:
• dictionaries, encyclopedias, thesauri;
• textbooks and reference books;
• pertinent corpora, most conveniently websites.
General common sense or a small sample of perti-
nent texts brings up a small number of apparently basic
terms. These terms are looked up in the first source, and
that leads to the second. A selection of key terms forms
the basis of an Internet search that brings up the cor-
pora. Thus, in the case of bankruptcy, the term itself
brings up an informative entry from Barron’s Finance
and Investment Handbook (1995).
“BANKRUPTCY State of insolvency of an individ-
ual or an organization—in other words, an inability to
pay debts. There are two kinds of legal bankruptcy un-
der U S. law: involuntary, when one or more creditors
petition to have a debtor judged insolvent by a court;
and voluntary, when the debtor brings the petition. In
both cases, the objective is an orderly and equitable
settlement of obligations.
The 1978 Bankruptcy Reform Act removed some of
the rigidities of the old law and permitted more flexibil-
ity in procedures. The Bankruptcy Reform, Act of 1984
curtailed some of the more liberal provisions (mainly
affecting consumer bankruptcy) of the 1978 act.
Chapter 7 of the 1978 act, dealing with
LIQUIDATION, provides for a court appointed interim
trustee with broad powers and discretion to make man-
agement changes, arrange unsecured financing, and
generally operate the debtor business in such a way as to
prevent loss. Only by filing an appropriate bond is the
debtor able to regain possession from the trustee.
Chapter 11, which deals with REORGANIZATION,
provides that, unless the court rules otherwise, the
debtor remains in possession of the business and in
control of its operation. Debtor and creditors are al-
lowed considerable flexibility in working together. The
1978 law relaxes the old absolute priority rule, which
gave creditor claims categorical precedence over owner-
ship claims. It also makes possible the negotiation of
payment schedules, the restructuring of debt, and even
the granting of loans by the creditor to the debtor.”
The entry, while somewhat cryptical, offers a pretty
good guide for what to look for in a textbook or in legal
sources. It can easily lead to a number of sources of the
textbook category, such as Summers (1989), Caplan
(1992), Davidson (1992). The pertinent information,
thus acquired, helps to identify the corpora, which
should be both essential for the domain and crucial for
the application(s), such as the various bankruptcy-
related pages at http://www.uslaw.com/. Just as in field
linguistics (cf. Samarin 1967), the corpora should be
varied, multi-sourced, and as representative/exhaustive
as possible. The corpora give us a good sense of the
grain size of the information to be included in the
script—see more on this in Section 6.
The most important step is to structure this informa-
tion in the script. Models of the script’s series of events
obtained at the previous stage and their key concepts
need to be checked against the ontology before the
scripting takes place, to avoid later, costly adaptation of
newly introduced concepts to the existing inventory.
The models will also tend to pay too much attention to
details from the field to which they belong. These de-
tails have to be weeded out and the parts of the models
to be united into the script have to be translated into
ontological concepts, existing and, if necessary, newly
acquired.
The methods for doing this are not easy to formulate
as recommendations, let alone rules. A similar situation
in lexical and ontological acquisition leads, with experi-
ence, to pretty well-established routines and, as a result
of adhering to them, quite good uniformity among dif-
ferent acquirers. Our work on routine acquisition of full-
fledged scripts has only been going on for slightly over
a year and has included only two domains so far, the
financial domain and the domain of meetings. We hope
to be able to make enough useful generalizations in the
style of Chapter 9 on acquisition of ontological concepts
and lexical entries in Nirenburg and Raskin (2003) as
we acquire more practical experience. The discovery of
heuristic rules remains a major challenge, possibly un-
attainable.
The following are the factors to be identified in the
script as concepts:
• the candidates for component events;
• the concepts involved in/created by the series of
events;
• the goals of the component events;
• their temporal and causal relations leading to
their groupings into subscripts;
• decision forks, such as whether to file Chapter 7
or Chapter 11 bankruptcies.
5 Formatted Script
The results of the operations described in Section 4 are
incorporated in the script for BANKRUPTCY below (for
legibility the FACET types SEM and VALUE are omitted in
this example):
BANKRUPTCY
is-afinancial-event
agentowe.agent
owe.beneficiary
preconditionapproach-bankruptcy
has-parts(IF modality.pay.value = 0
THEN bankrupt-chapter-7
ELSE bankrupt-chapter-11)
APPROACH-BANKRUPTCY
is-afinancial-event
agentcorporation-a
has-parts
(IF
AND
owe
agentcorporation-a
beneficiaryhuman-a
employed-bycorporation-a
lending-institution-a
corporation-b
thememoney
pay
agentcorporation-a
beneficiaryhuman-a
lending-institution-a
corporation-b
thememoney
THENbankruptcy
agentcorporation-a
beneficiaryhuman-a
lending-institution-a
corporation-b)
PAY.MODALITY
typepotential
scopepay
value<1
BANKRUPT.MODALITY
typeepistemic
scopebankruptcy
value0.6
BANKRUPT-CHAPTER-7
is-afinancial-event
agentowe.agent
owe.beneficiary
preconditionbankruptcy
has-parts(ANDbankrupt.declare
bankrupt.business-activity
bankrupt.appoint
bankrupt.change-event
bankrupt.pay)
BANKRUPT-CHAPTER-11
is-afinancial-event
agentowe.agent
owe.beneficiary
preconditionbankruptcy
has-parts(ANDbankrupt.declare
bankrupt.legal-case
bankrupt.audit
(IF modality.bankrupt.audit.value = 0
THEN bankrupt-appoint)
bankrupt.planning-event
bankrupt.follow-plan
(IF modality.pay.value = 0
THEN bankrupt-chapter-7
ELSE modality.pay.value = 1))
BANKRUPT.DECLARE
instance-ofdeclare
agentowe.agent
owe.beneficiary
destinationowe.agent
pathjudicial-branch
BANKRUPT.BUSINESS-ACTIVITY
instance-ofbusiness-activity
beneficiaryowe.agent
BANKRUPT.BUSINESS-ACTIVITY.MODALITY
typeepistemic
scopebankrupt.business-activity
value0
BANKRUPT.AUDIT.MODALITY
typeepiteuctic
scopebankrupt-audit
value?
BANKRUPT.APPOINT
instance-ofappoint
thememanager-corporation
beneficiaryowe.agent
agentjudicial-branch
BANKRUPT.CHANGE-EVENT
instance-ofchange-event
themeasset
destinationmoney
agentmanager-corporation
beneficiaryowe.agent
BANKRUPT.PAY
instance-ofpay
thememoney
agentowe.agent
beneficiaryowe.beneficiary
BANKRUPT.LEGAL-CASE
instance-oflegal-case
thememoney
agenthuman
beneficiaryowe.agent
BANKRUPT.LEGAL-CASE.MODALITY
typeepistemic
scopebankrupt.legal-case
value0
BANKRUPT.AUDIT
instance-ofaudit
agentjudicial-branch
beneficiaryowe.agent
themecash-flow
asset
debt
BANKRUPT.PLANNING-EVENT
instance-ofplanning-event
agentowe.agent
judicial-branch
BANKRUPT.FOLLOW-PLAN
instance-offollow-plan
agentowe.agent
6 Grain Size Issues
The script above, even though much more complex than
the script for TEACH is presented in its simplest and
probably coarsest form. The gain is parsimony, in the
sense of minimizing the need to acquire new lexical
entries or concepts. Are there losses? A text may men-
tion, for instance, a supplier’s refusal to ship stuff to the
bankrupt corporation. It does that because the corpora-
tion cannot pay it for the supplies. Can we consider it
covered in the script? What if a text mentions the in-
ability to meet the payroll? Meeting the payroll may
deserve a script of its own. It may be seen to be covered
sufficiently in the script above, but laying off employees
may not. To owe a loan is actually to owe an installment
payment on a certain date, and to be unable to pay the
loan means, actually, the inability to pay an installment
payment of the loan on a certain date. The script above
also omits the entire credit ratings game.
The rationale for having the scripts is, not surpris-
ingly, to do what Schank declared his group would do a
quarter of a century ago (Schank, 1975; Schank and
Abelson, 1977) and, unlike them, to deliver a workable
non-toy product, in which the whole script is evoked
when any element of it at any level of the script hierar-
chy occurs lexically in the text. The simplistic repre-
sentation above obligates our analyzer to reduce any
such pertinent lexical material to the level of owing and
paying. Is it possible? The alternative is to develop
much more elaborate scripts, involving a great deal
more of ontological acquisition and change.
A more complex and more accurate level of repre-
sentation, with all the intermediate subsidiary scripts
embedded in other scripts as well as component simple
events enriched with precondition and effect (and, we
increasingly believe, goal values), will be much costlier,
so the question is whether the gain in analysis makes it
worthwhile. We expect this to be dictated by the needs
of the current and future applications as manifested in
their goals and the nature of the texts in the pertinent
corpora. But much more effort will have to be devoted
to developing more specific grain-size recommenda-
tions, rules of thumb and repair/recovery procedures for
cases when the grain size of the script is not sufficient to
handle a text.

References
Attardo, S., and V. Raskin 1991. Script theory re-
vis(it)ed. Joke similarity and joke representation
model. HUMOR 4:3-4, 293-347.
Barron’s Finance and Investment Handbook, 1995. 4th
ed., ed. by J. Downes and J. E. Goodman. New York:
Barron’s.
Binsted, K. and G. Ritchie 1997. Computational rules
for generating punning riddles. HUMOR 10:1, 25-76.
Caplan, S. 1992. Saving Your Business: How to Survive
Chapter 11 Bankruptcy and Successfully Reorganize
Your Company. Englewood Cliffs, NJ: Prentice Hall
Carlson, L., and S. Nirenburg 1990. World Modeling
for NLP. Technical Report CMU-CMT-90-121,
Center for Machine Translation, Carnegie Mellon
University, Pittsburgh, PA. A short version appeared
in: Proceedings of the 3rd Conference on Applied
Natural Language Processing, Trento, Italy, April.
Charniak, E. 1972. Toward a Model of Children's Story
Comprehension. Artificial Intelligence Technical Re-
port Number 266, Department of Computer Science,
Massachusetts Institute of Technology, Cambridge,
MA, December.
Davidson, R. L. 1992. The Small Business Bankruptcy
Kit. New York: Wiley.
Fillmore, C. J. 1985. Frames and the Semantics of Un-
derstanding. In: V. Raskin (ed.), Round Table Dis-
cussion on Frame/Script Semantics, Part I, Quaderni
di Semantica VI: 2, 222-254.
Hempelmann, C. F. 2003. Paronomasic Puns: Target
Recoverability towards Automatic Generation. Un-
published Ph.D. thesis, Interdepartmental Program in
Linguistics, Purdue University, West Lafayette, IN.
McDonough, C. J. 2000. Complex Events in an On-
tologic-Semantic Natural Language Processing Sys-
tem. Unpublished Ph.D. thesis, Department of
English, Purdue University, West Lafayette, IN.
Minsky, M. 1975. A Framework for Representing
Knowledge. In: P. H. Winston (ed.), The Psychology
of Computer Vision. New York: McGraw Hill, 211-
77.
Moreno Ortiz, A., V. Raskin, and S. Nirenburg (2002)
New Developments in Ontological Semantics. In:
Proceedings of the Third International Conference
on Language Resources and Evaluation (LREC
2002). Las Palmas de Gran Canaria, Spain. May 29-
31, pp. 1196-1202.
Nirenburg, S., and V. Raskin 2003. Ontological Seman-
tics. Cambridge, MA: MIT Press (forthcoming).
Raskin, V. 1986. Script-Based Semantic Theory. In: D.
G. Ellis and W. A. Donohue (eds.), Contemporary Is-
sues in Language and Discourse Processes,
Hillsdale, NJ: Erlbaum, 23-61.
Raskin, V. 1994. Frawley: Linguistic Semantics. A Re-
view Article. Language 70: 3, 552-556.
Raskin, V. 1996. Computer Implementation of the Gen-
eral Theory of Humor. In: J. Hulstijn and A, Nijholt
(eds.), Automatic Interpretation and Generation of
Verbal Humor. Twente Workshop on Language
Technology TWLT 12, The Hague: CIP, 9-19.
Raskin, V., and S. Attardo 1994. Non-Literalness and
Non-Bona-Fide in Language: Approaches to Formal
and Computational Treatments of Humor. Cognition
and Pragmatics 2:1.
Raskin, V., C. F. Hempelmann, K. E. Triezenberg, and
S. Nirenburg 2002. Ontology in Information Secu-
rity: A Useful Theoretical Foundation and Methodo-
logical Tool. In: V. Raskin and C. F. Hempelmann
(eds.), Proceedings. New Security Paradigms Work-
shop 2001. September 10th-13th, Cloudcroft, NM,
USA, New York: ACM Press, 53-59.
Samarin, W. J. 1967. Field Linguistics. New York: Holt,
Rinehart and Winston.
Schank, R. 1975. Conceptual Information Processing.
Amsterdam: North-Holland.
Schank, R., and R. Abelson 1977. Scripts, Plans, Goals,
and Understanding. Hillsdale, NJ: Erlbaum.
Stock, O., and C. Strapparava 2002. Humorous Agents
for Humorous Acronyms: The HAHAcronym Pro-
ject. In: O. Stock, C. Strapparava, and A. Nijholt
(eds.) 2002. The April Fools’ Day Workshop on
Computational Humor. April 2002, ITC-irst, Trento.
Twente Workshop on Language Technology TWLT
20. Trento: ITC-irst, 125-135.
Summers, M. 1989. Bankruptcy Explained, a Guide for
Businesses. New York: Wiley.
