the theory makes use of some standard abbreviatory devices that enable nodes and or paths to be omitted in certain cases
global inheritance for example has a dynamic aspect which is difficult to represent in terms of static links
the hierarchy provides a useful means of visualising the overall structure of the lexical knowledge encoded by the datr theory
this picture of datr as a formalism for defining partial functions is complicated by two features of the language however
evaluable paths provide a powerful means of capturing generalizations about the structure of lexical information
where v vlv2 and vx is the longest prefix of v such that f vl is defined
in this case the global context will initially consist of the node path pair walk mor pres
for example consider again the definition of the node walk from the theory of figure NUM and repeated below
green solves the problem by extending the boundaries of the analysis to discourse units
we present a set of examples that covers a representative group of pragmatic inferences
let uttered u be the logical translation of a given utterance or set of utterances
this paper provides a theoretical framework called stratified logic that can accommodate defeasible pragmatic inferences
model m0 explains why mary came to the party is a presupposition for utterance NUM
the formalism yields an algorithm that has been implemented in common lisp with screamer
having a satisfaction definition associated with each level of strength provides a high degree of flexibility
figure NUM stratified tableau for john does not regret that mary came to the party
the achilles heel for most theories of presupposition has been their vulnerability to the projection problem
mary knows that john smith is a married male so the utterance makes sense for her
insert heel first as shown then push top in firmly
the battery low indicator will light when the battery is low
the testing set is exclusively procedural and is included in full
the accuracy of imagene s realizations for purpose and precondition expressions
the accuracy of imagene s realizations for result and sequence expressions
these are the forms that were present in our corpus
goal status determines whether the use of goal metonymy is warranted
it typically places the purpose expression in the final position
the second task that of determining the functional context in which each of the forms is used is more difficult
the current study has made use of five such relations purpose precondition result sequence and concurrent
the first task is that of determining the range of lexical and grammatical forms used to express each particular rhetorical relation
because of the common use of rst the results can be more easily applied to other work in this area
here the activity of removing the phone is realized by the execution of the sub steps of grasping and pulling the phone
the match was judged on four separate lexical and grammatical issues linker form slot and clause combining
i3 NUM this analysis of nominalizations is an example of the descriptive nature of the current study of instructional text
here are some representative forms as generated by imagene 15a if light flashes insert credit card
the first type is non reader actions that are not the result of an explicit command to monitor a particular device state
an extended definition of satisfaction and a notion of optimism with respect to different interpretations yield the preferred interpretations for an utterance or sequences of utterances
at this point jane realizes that mary misunderstands her all the time jane was talking about john pevler the five year old boy
these interpretations contain the pragmatic inferences that have not been canceled by context or conversant s knowledge plans or intentions
the utterances in NUM constitute a possible answer that jane may give to mary in order to clarify the problem
the mgorithm builds the set of all possible interpretations for a given utterance using a generalization of the semantic tableau technique
stratified logic supports one type of indefeasible information and two types of defeasible information namely infelicitously defeasible and felicitously defeasible
the framework yields an algorithm that computes the conversational conventional scalar clausal and normal state implicatures and the presuppositions that are associated with utterances
assume now that after a moment of thought the same person utters NUM john says that some of the boys went to the theater
lilt was not demoimtratext that boris can be scaled up to over a signiticant part of the english lexicon
it has been designed froln the start to deal with a comprehensiw range ot seman
significant scale up can be accomplished under such a constraint without jeopardizing the high values on the depth and breadth scales
depth and breadth are of course of little use if the system cmmot bc scaled up to a signilicant size
we have chosen representative systems fl om the different approaches for lack of precise terms to name tlle approaches
for example snch a system may have a large lexicon with only word frequency and collocation information in each entry
the coverage of phenomena was strictly utilitarian which we believe is quite appropriate
its knowledge bases and text meaning representations are rather deep and of nontrivial sizes
boris was able to process a very smml number of texts sutticiently for its goms
focusing on size to the exclusion of other criteria has biased the field away from semantic solutions to nlp problems
the notion of a model presented in the preceding section is too liberal in that it takes no account of information implicit in a theory
now by default the value associated with mor pres at walk is inherited locally from mor pres at verb
finally the semantics presented in this paper provides a sound basis for subsequent investigations into the mathematical and computational properties of datr
gate plans to support conversions between tipster annotations and multext sgml markup allowing data produced by the tipster compliant creole modules to be written out in sgml for processing by multext tools and the sgml results then converted back into native format
an interpretation m0 is preferred to or is more optimistic than an interpretation ml m0 ml if it contains more information and that information can be more easily updated in the future
there is no contradiction between these two presuppositions so one would expect a conversant to infer both of them if she hears an utterance such as NUM
the stratified semantic tableau that corresponds to its logical theory yields NUM models but only chris is an adult satisfies definition NUM NUM and is projected as presupposition for the utterance
unlike most research in pragmatics that focuses on certain types of presuppositions or implicatures we provide a global framework in which one can express all these types of pragmatic inferences
an appropriate formalization is given in NUM where the second formula captures the defeasible scalar implicatures and the third formula reflects the relevant semantic information for all
suppose for example that jane has two friends john smith and john pevler and that her roommate mary has met only john smith a married fellow
this algorithm computes uniformly pragmatic inferences that are associated with simple and complex utterances and sequences of utterances and allows cancellations of pragmatic inferences to occur at any time in the discourse
the theory provides one optimistic model schema figure NUM that reflects the expected pragmatic inferences i.e. not most not many not all of the boys went to the theater
at this point the hearer mary starts to believe that one of her previous utterances has been elaborated on a false assumption but she does not know which one
the definitions given there deal with a subset of datr that includes core features of the language such as the notions of local and global inheritance and datr s default mechanism
what distinguishes liquids appears to be their ability to drip or drain over a period of time
the current study is critically interested in discerning principled reasons for choosing between these sorts of expressions
of the NUM purpose expressions approximately NUM are fronted and NUM are not fronted
although this greatly improves the text it still tends toward text that is difficult to read
it will detail what rhetorical relations in instructional text are and how they were collected and represented
in addition trl specifies the textual order and clause combining using additional new sentence and continue sentence links
our conclusion is that the network metaphor is of primary value to the datr user
each of these approaches formalizes a notion of default inheritance by defining appropriate operations e.g.
the denotation function is extended to sequences of value descriptors in the obvious way
figure NUM a lexical inheritance hierarchy associated with mor past at walk is obtained as
the formal model provides a transparent treatment of datr s notion of global context
second it is necessary to provide an account of datr s default mechanism
this in turn inherits globally from the path mor root
in fact all of the information given explicitly in a datr theory is strict
the node provided by the current global context which is just walk
when such a set of conditions is met that inference is drawn but it is assigned a defeasible status
being a bachelor presupposes that chris is a male while being a spinster presupposes that chris is a female
our solution for the projection problem does not differ from a solution for individual utterances
the lattice shows a partial order that is defined over the different levels of truth
a partial ordering determines the set of optimistic interpretations for a theory
green s algorithm makes use of discourse expectations discourse plans and discourse relations
we now study how stratified logic and the model ordering relation capture one s intuitions
NUM not lcb many most all rcb of the boys went to the theater
NUM john says that some of the boys went to the theater
a uniform treatment of pragmatic inferences in simple and complex utterances and sequences of utterances
ld pull out sharply for removing the phone
le for the phone pull out sharply
rst distinguishes between what are called nucleus satellite and multi nuclear schemata
lb to remove phone pull out sharply
penman s networks are specifically designed to construct english sentences
a segment of the text structure for the remove phone example
this issue of choice is central to the current study
proof of either of these issues would require psycholinguistic testing
the choice of clause combining strategies is made here as well
figure NUM contains a purpose relation and a sequential relation
the current study addresses these four issues of choice in the context of instructional text
example 3d uses a for preposition with a gerund phrase as the complement
the issues of slot and form of purpose expressions are treated largely independently by imagene
consider the following examples of these situations 6a when instructed approx
the results can also serve as a source of preliminary hypotheses with respect to the analysis of other related genres
this characterization is directly applicable to current work on instructional text particularly in the context of natural language generation
in our corpus this is the fronted to infinitive which occurred in NUM of the purpose expressions
to judge these results more fully consider an alternative system that always generates the single most common purpose form
possible interpretations of a given set of utterances with respect to a knowledge base are computed using an extension of the semantic tableau method
normally we would expect a conversant to notice this contradiction and to drop each of these elementary presuppositions when she interprets NUM
therefore an appropriate formalization of utterance NUM and the req fisite pragmatic knowledge will be as shown in NUM
drawing appropriate defeasible inferences has been proven to be one of the most pervasive puzzles of natural language processing and a recurrent problem in pragmatics
how do the approaches illustrated in figure NUM rate with respect to the metrics suggested above
null how many word senses from standard human oriented dictionaries are covered in the nlp oriented lexicon entry
the proposed framework is illustrated using sever fl prominent nlp systems
what is the size of the semantic zones of the entry
figure NUM l imensions of semantic coverage cur
but size alone is not a sufficient metric for evaluating semantic coverage
what types of information are included
figure NUM denotation function for datr descriptors
on the other hand the value of
nodes are denoted by n and atoms by a
default unification for combining strict and default information
a collection of equations can be used to specify the properties of different nodes in terms of one another and a finite set of datr sentences NUM is called a datr theory
figure NUM revised denotation for global node path pairs
as a result the default mechanism is effectively over ridden
in order to understand the use of global i.e.
this specification is achieved either explicitly or implicitly
implicit specification is achieved via datr s default mechanism
in fact all of them went to the theater
the warning is conveyed by implicature
this new empirical focus is supported by several recent advances an increasing theoretical consensus on discourse models a large amount of on line dialogue and textual corpora available and improvements in component technologies and tools for building and testing discourse and dialogue testbeds
it is shown that da i r s default mechanism can be accounted for by interpreting value descriptors as families of values indexed by paths
there is a standard view of datr as a language for representing a certain class of non monotonic inheritance networks semantic nets
note that there is nothing inherent in datr to ensure that theories correspond to simple isa hierarchies of the kind shown in the figure
values given explicitly are specified either directly by exhibiting a particular value or indirectly in terms of local and or global inheritance
by contrast the denotation of an inheritance descriptor is in general sensitive to the global context c in which it appears
in order to provide a satisfactory formal model of how the language works it is necessary to adopt a different perspective
as in the previous example the one dealing with a sequence of utterances we obtain a different interpretation after each step
moreover it is by no means clear that the approach can be generalized appropriately to cover these features
we associate with the interpretation i u t f a partial denotation function d desc c u and write d to denote the meaning value of descriptor d in the global context c
definition NUM NUM a datr interpretation is a triple i u ic f where NUM u is a set NUM is a function assigning to each element of the set u x v a partial funclion from u x u
in the same way the value of syn cat at mow is inherited locally from enverb which in turn inherits locally from verb and the value of syn cat at can is inherited locally from modal which ultimately gets its value from verb via aux
second it serves as a standard against which other operational definitions of the formalism can be judged
first it provides the datr user with a concise implementationindependent account of the meaning of datr theories
a primary objective in the development of datr has been the provision of an explicit mathematically rigorous semantics
although this is true for both johns it is more appropriate for the married fellow than for the five year old boy
therefore the implicature at the end of the dialogue is that the conversant who answered went shopping
answer NUM a makes the necessary conditions for implicating no true and the implication is computed
an appropriate formalization for utterance NUM and the necessary semantic and pragmatic knowledge is given in NUM
at the syntactic level we allow atomic formulas to be labeled according to the same underlying lattice
in this context it is natural for mary to become confused and to come to wrong conclusions
note that there is a gap of one statement between the generation and the cancellation of this presupposition
those that are not canceled are labeled as pragmatic inferences for the given utterance or set of utterances
this research was supported in part by a grant from the natural sciences and engineering research council of canada
quite often these substances are liquids but may also include other inanimates
these determinations are made by the scope system that is not repeated here
grasp the top of the handset and pull it
this training set constitutes approximately NUM of our corpus
NUM sec firmly grasp top of handset and pull out
it is thus fronted to set the appropriate context for the prescribed action
NUM curly braces indicate that all sub networks on the right should be entered
7a return handset to wall unit from which it was taken
in both cases the to infinitive form was preferred
the handset must be returned to the base for recharging
there is one table for each of the major rhetorical relations purpose precondition result and sequence
when the phone is installed and the battery is charged move the off stby talk switch to the stby position
it is not handled particularly well by imagene and the results given in this section will not include it
keith vander linden and james h martin expressing rhetorical relations spans of text that express a sequence of two actions
the rst representation of the remove phone text is a small portion of the full hierarchy that represents the entire manual
because of the prevalence of this use of the term however it will be retained in this paper
the complexity of procedural relations typically expressed in such text has given rise to complex variations of expression in language
the representation of the grammatical form of the clauses and phrases is based on traditional principles of syntax and semantics
we will present a detailed discussion of the purpose sub network below which is representative of the other grammatical form sub networks
a second input shown in figure NUM the process structure is a representation of the process to be expressed
it is currently built by hand which allows us to focus on the problem of expression rather than on planning
this provides a good starting hypothesis for determining the slot of purposes namely that global purpose clauses are fronted
the purpose slot analysis just discussed for example identifies the existence of a covariation between purpose scope and slot
it transforms inputs shown on the left in figure NUM into instructional text shown on the right
the system network is basically a decision network in which each choice point distinguishes between alternative features of the communicative context
more specifically datr s default mechanism ensures that any path that is not explicitly specified for a given node will take its definition from the longest prefix of that path that is specified
a definitional sentence n p c specifies a property of the node n namely that the path p is associated with the value defined by the sequence of value descriptors c
for a given datr theory NUM and node n of NUM we write NUM n to denote that subset of the sentences in NUM that relate to the node n
thus it is not possible in practice to combine definitional and extensional sentences within a theory NUM functionality for datr theories as defined above is really a syntactic notion
for c a possibly empty sequence of value descriptors an equation of the form n p c is called a definitional sentence
a model of a datr theory is then an assignment of func null tions to node symbols that is consistent with the definitions of those nodes within the theory
it follows by default from the statements made about walk that the path mor inherits locally from verb and that the value associated with any extension of mor root is walk
for this reason the work described in this paper reflects a rather different perspective on datr as a language for defining certain kinds of partial functions by cases
NUM NUM pragmatic inferences in sequences of utterances
pragmatic inferences are triggered by utterances
he is only five years old
the behavior described is mirrored both by our theory and our program
the model ordering relation filters the optimistic interpretations
i regret that you have misunderstood me
our approach does not exhibit these constraints
pd in any optimistic model schema of the theory
the third utterance NUM c comes to clarify the issue
however it approximates a deeper semantic requirement that the nodes should correspond to partial functions from paths to values
however they exclude some important and widely used constructs most notably string or list values and evaluable paths
in order to re interpret node definitions in the manner suggested above it is necessary to modify the interpretation of value descriptors
third the process of formalisation itself aids our understanding of the language and its relationship to other non monotonic attribute value formalisms
the set desc of datr value descriptors or simply descriptors is built up from the atoms and nodes as shown below
intuitively an extensional sentence n p a states that the value associated with the path p at node n is a
in fact there are a number of constructs available in datr that are impossible to visualize in terms of simple inheritance hierarchies
the treatment of implicit information is covered in section NUM NUM which provides a definition of a default model for a datr theory
the function tc can be thought of as mapping global contexts onto partial functions from local contexts to values
in the case of datr however the formalism does not draw an explicit distinction between strict and default values for paths
there is just one kind of local descriptor node path but three kinds of global descriptor node path path and node NUM a path al an is a possibly empty sequence of atoms enclosed in angle brackets
walk NUM verb mor root walk according to the definition of a model given previously any model of the theory of figure NUM will associate with the node walk a function from paths to values which respects the above definition
however these additional forms can be viewed as conventional abbreviations in the appropriate syntactic context for node path pairs n p b and n p c e NUM implies b c there is a pragmatic distinction between definitional and extensional sentences akin to that drawn between the language used to define a database and that used to query it
this means that the statement walk NUM verb NUM actually corresponds to a class of implicit definitional sentences each obtained by extending paths on the left and the right hand sides of the equation in the same manner
thus in the preceding example after the quoted descriptor mor root is encountered the global context effectively becomes walk mor root i.e. the path component of the global context is altered
first the meaning of a given node depends in general on the global context of interpretation so that nodes do not correspond directly to mappings from paths to values but rather to functions from contexts to such mappings
rather than interpreting node definitions in a given global context as partial functions from paths to values i.e. of type u u we choose instead to interpret them as partial functions from explicit paths to functions from extensions of those paths to values i.e. of type u u u
then i models 7just in case the following containment holds for each node n and context c n d t n that is an interpretation is a model of a datr theory just in case for each global context the function it associates with each node respects the definition of that node within the theory
we present a view of datr as a language for defining certain kinds of partial functions by cases
section NUM NUM introduces datr interepretations and describes the semantics of a restricted version of the language without defaults
thus the node can inherits from modal which inherits from aux which in turn is a verb
elements of the set c u x u are called contexts and denoted by c
suppose then that we wish to determine the value associated with the path mor pres at the node walk
although a number of implementations of datr exist the full language has until now lacked an explicit declarative semantics
however the semantic network metaphor is of far less value as a way of thinking about the datr language itself
section NUM NUM then shows how implicit information can be modelled by treating value descriptors as families of values indexed by paths
we will write t n c to denote that partial function from u to u given by
u is stronger in a sense to be defined below than something that is infelicitously defeasibly true t i or felicitously defeasibly false l a formally we say that the u level is stronger than the i level which is stronger than the d level u i d
we say that utterance u pragmatically implicates p if and only if p d or p i is derived using pragmatic inferences in at least one optimistic model of the theory u uttered u and if p is not canceled by any stronger information p pi
schema indefeasible in elicitously felicitously de easible de feasible mo we t some boy theater oe t most boys theater went many boys theater went all boys theater d went
come i y z besides the translation of the utterance the initial theory contains a formalization of the defeasible implicature that natural disjunction is used as an exclusive or the knowledge that mary is not a name for males the lexical semantics for the word bachelor and the lexical pragmatics for bachelor and regret
utterance NUM b warns mary that is very likely she misunderstood a previous utterance NUM
therefore it cancels the early presupposition NUM NUM john is an adult
a factive such as the verb regret presupposes its complement but as we have seen in positive environments the presupposition is stronger it is acceptable to defeat a presupposition triggered in a negative environment NUM but is infelicitous to defeat one that belongs to a positive environment NUM
in contrast with most other approaches we provide a consistent methodology for computing these inferences and for determining whether they are canceled or not for all possible configurations simple and complex utterances and sequences of utterances
this theory yields the following definition definition NUM NUM let b be a theory described in terms of stratified first order logic that appropriately formalizes the semantics of lezical items and the necessary conditions that trigger pragmatic inferences
expressing rhetorical relations in instructional text a case study of the purpose relation
NUM in the system network notation vertical lines indicate decision points
to the last action alone rather than to the sequence of actions
it is assumed that it could be constructed using artificial intelligence planning methodologies
determining the precise communicative context in which each of these forms is used
examples crucial to the formalization of the results were reviewed by both authors
the spans used in our analysis are propositional units that express single actions
li pulling out sharply achieves the purpose of removing the phone
mor form mor syn form mor pres mor root mor past mor root ed mor pres part mor root ing mor pres sing three mor root
the approach provides a transparent treatment of datr s notion of local and global context and accounts for datr s default mechanism by regarding value descriptors semantically as families of values indexed by paths
note the way in which the path argument v is used to extend vl vn in order to define the new local and in this case also global context c on the other hand the meaning of each of the di is obtained with respect to the em pty path e i.e.
while this perspective provides an intuitive and appealing way of thinking about the structure and representation of lexical knowledge it is less clear that it provides an accurate or particularly helpful picture of the datr language itself
the second aspect of this structure is the nature of the rhetorical relations themselves
the second problem is one of relating these segments in the appropriate rhetorical structure
15b the battery low indicator will light when the battery is low
here holding the mouse button and moving the mouse must be done simultaneously
statistics were kept on how well its realizations matched the expressions in the corpus
15d return the off stby talk switch to the stby position after your call
the form in example 8a is more commonly used in our corpus in this context
tm such a system would score NUM under the verification criteria used here
text generation systems must know which forms to produce and when to produce them
in principle a datr theory NUM may consist of any combination of datr sentences either definitional or extensional but in practice datr theories are more restricted than this
more generally the global context is used to fill in the missing node path when a global path node is encountered
this approach is possible only on the overly restrictive assumption that darr statements involve either local or global inheritance relations but never both
in the remainder of this paper we will use the term datr theory always in the sense functional definitional datr theory
datr theories can be viewed semantically as collections of definitions of partial functions nodes in datr parlance that map paths onto values
at the present time datr is probably the most widely used formalism for representing natural language lexicons in the natural language processing nlp community
intuitively value descriptors denote elements of u as we shall see this will need to be revised later in order to account for datr s default mechanism
the notion of infelicitously defeasible information is meant to capture inferences that are anomalous to cancel as in NUM john regrets that mary came to the party but she did not come
assume now that jane has a conversation with mary in which jane mentions only the name john because she is not aware that mary does not know about the other john who is a five year old boy
compound formulas are obtained in the usual way
the stratified semantic tableau generates NUM model schemata
answer NUM b reinforces a previous condition
it explicitly expresses that john is not an adult
the set tin is referred to as the definition of n in NUM
an example of a fragment of a datr theory is shown in figure NUM
quoted inheritance descriptors it is necessary to introduce datr s notion of a global context
the former has no effect on the global context while the latter effectively overwrites it
the definition of the interpretation of node definitions can be taken over unchanged from the previous section
figure NUM shows the revised clause for global node path pairs the other definitions being very similar
strict information is allowed to over ride default information where the combination would otherwise lead to inconsistency i.e.
section NUM presents the syntax of the datr language and introduces the notion of a datr theory
each node is associated with a collection of definitional sentences that specify values associated with different paths
similarly the definition of walk gives the value of mor root directly as walk
c so i had to take the bus
NUM i john is a male adult
the algorithm applies equally to simple and complex utterances and sequences of utterances
only four of them are kept as optimistic models for the utterance
for example mary may reply that john is not a bachelor
consider utterance NUM and its implicatures NUM
NUM NUM chris is a male adult
the first utterance in the sequence presupposes NUM
NUM chris is a bachelor or a spinster
the discussion of peculiarities of the various approaches should be expanded in at least two directions greater detail of description and analysis of the relative ditficulty of reaching the set goal of attaining an optimum value on each of the three measurement scales
the cyc large scme knowledge base as significant amounts of deep knowledge llowever it is not clear whether the knowledge is apl licm le in a straightff rward manner to deal with a range of linguistic phenomena
in this article we argue that depth of semantic representation is essential for covering a broad range of phenomena in the computational treatment of language and propose lepth as an important additional dimension for measuring the semantic coverage of nlp systems
we hope that this paper will elicit interest in continned discussion of the issues of coverage measurement which in turn will lead to better quantitative as well as qualitative measures including a methodology for comparing lexicons and ontologies
r eference and coreference attitude modality stylistics quantitative comparative and other mathematical relations textual relations and other discourse relations null multiple ambiguous interpretations propositional and story dialog structure NUM measuring semantic coverage
tic phenomena including the linldng of syntax attd semantics ore semantic analysis sense disaml iguation i rocessing non literm expressions slit lcb so on althongh not all of them have yet been im plemented
it may i e considered an extreme example of a system with deep rich knowl edge of its rather narrow worhl in which cowwing language phenomena is nee ted only inasmuch as it supports general reasoning
poor coverage of language phenomena i.e. poor brea dth indicates that the acquired knowh dge even when it is deep and large in size may not be applicable to other phenomena and may not transfer to other applications
we have presented a formal model of datr as a language for defining partial functions and this model has been contrasted with an informal view of datr as a language for representing inheritance hierarchies
an interpretation i u x f is a default model for theory t just in case for every context c and node n we have in a it ni as an example consider the default interpretation of the definition of the node walk given above
note that there is a real distinction between a local inheritance descriptor of the form n p and it s global counterpart n p
the author would like to thank roger evans gerald gazdar bill rounds and david weir for helpful discussions on the work described in this paper
datr interpreters conventionally treat all extensional sentences as goal statements and evaluate them as soon as they are encountered
the empty path at walk is given indirectly by local inheritance as the value of the empty path at verb
also the values of mor and mor form amongst many others are inherited from verb
in addition however the evaluation of a global descriptor results in the global context being set to the new node path pair
thus the default mechanism defines a class of implicit definitional sentences with paths on the left that extend paths found on the left of explicit sentences
NUM f is a valuation function assigning to each node n and atom a an element of u such that distinct atoms are assigned distinct elements
what is more the datr language includes constructs that can not be visualized in terms of simple networks of nodes connected by local inheritance links
now suppose that f u u u is the function associated with the node definition t n in a given datr interpretation
the non monotonic nature of datr theories arises from a general default mechanism which fills in the gaps by supplying values for paths not explicitly specified in a theory
imagene s accuracy for results and sequence expressions is similar to that presented for purposes and preconditions
imagene expresses these as present tense relational expressions as seen in example 16b
we have the following picture the default models of a theory NUM constitute a proper subset of the models oft just those that respect the default interpretations of each of the nodes defined within the theory
in effect the function a f makes explicit that information about paths and values that is only implicit in f but just in so far as it does not conflict with explicit information provided by f
that is verb mor pres mor root consequently the required value is that associated with mor root at the global node walk i.e.
in a similar fashion the value walk ed i.e. the string of atoms formed by evaluating the specification mor root ed in the global context walk mor past
in a given global context c a value descriptor d now corresponds to a total function d u u intuitively a function from path extensions to values
however datr s default mechanism permits any definitional sentence to be applicable not only to the path specified in its left hand side but also for any rightward extension of that path for which no more specific definitional sentences exist
for example atoms now denote constant functions a c v f a for all v g u more generally value descriptors will denote different values for different paths
paths are denoted by p for n a node p a path and a atom a possibly empty sequence of atoms an equation of the form n p a is called an extensional sentence
one spl command is produced for each sentence in the text structure
all of these types of features have proven relevant in the analysis
examples of these expressions were taken from the remainder of the corpus
we leave a more complete treatment of this distinction to future work
given this structural representation of a sequence of actions the content and
rhetorical status selection system sub network can be viewed as using the inquiry
example 3a uses a to infinitive form tnf
there are other natural language generation projects that have addressed similar issues
these details are provided for our study in this paper
determining the range of expressional forms commonly used by instructional text writers
this relation is represented graphically with a directed arrow from the satellite to the nucleus
the handset must be returned to the base to recharge the battery
lh the purpose of pulling out sharply is to remove the phone
lk the method for removing the phone is to pull out sharply
the translation process is performed by a recursive descent of the text structure hierarchy
as with reorder unlink is used to un structure a default structuring
the fronted form in example 3a makes no such implication
the form is determined by the sub networks shown later in figures NUM and NUM
they have been removed here to remain consistent with the statistics shown in this section
there are two fundamental contributions of the current study to the field of computational linguistics
the total development time required to adapt satz to german including building the lexicon and constructing training texts was less than one day
elements of the set u are denoted by u and elements of u are denoted by v intuitively u is the domain of semantic values paths
it is easy to verify that t n does indeed denote a partial function it follows from the functionality of the theory NUM
for example the definition of the verb node gives the values of the paths syn cat and syn type directly as verb and main respectively
note that in itself this might not appear to be particularly useful since the theory does not provide an explicit value for the empty path in the definition of verb
in effect each explicit path is associated not just with a single value specification but with a whole family of specifications indexed by extensions of those paths
null the theory defines the properties of seven nodes an abstract verb node nodes enverb aux and modal and three abstract lexemes walk mow and can
figure NUM the lattice that underlies stratified logic
the provision of a formal semantics for datr is important for several reasons
this paper rectifies the situation by providing a mathematical semantics for datr
this section considers a restricted version of datr without the default mechanism
this suggests the following approach to the semantics of defaults in datr
furthermore this extension of paths is also carried over to paths occurring on the right
similar problems are presented by both string values and evaluable paths
path extension does not apply to subterms of inheritance descriptors
in the following sections this viewpoint is made more precise
let node and atom be disjoint sets of symbols the nodes and atoms respectively
it will be convenient to present our account of the semantics of datr in two stages
each pragmatic inference is associated with a set of necessary conditions that may trigger that inference
NUM a no john is not a bachelor
when the question is asked there is no conversational implicature
symmetrically one can define what a negative pragmatic inference is
in both cases w uttered u is u consistent
we can offer here only a brief overview of stratified logic
a recta logical construct uttered applies to the logical translation of utterances
for example something that is indefeasibly false l
the lattice in figure NUM underlies the semantics of stratified logic
these sorts of context setting purposes are not demoted to phrase status
such an account is beyond the scope of the current study
l j removing the phone involves pulling out sharply
consider for example the problem of expressing purpose relations
remove phone by firmly grasping top of handset and pulling out
the slot is determined by the sub network shown in figure NUM
two related issues must be addressed in the corpus analysis
NUM this section will now provide specific definitions of these relations
purpose precondition and result are examples of such relations
implement the results of this analysis in the text generation system
example 6b seems to make the incorrect implication that the prescribed actions work only when instructed
this context usually leads to the use of the so that purpose as shown in example 3g
in the test phase the analyst attempts to validate the hypothesis by querying the database for the relevant information
these two phases are repeated until a good match is achieved or until a relevant hypothesis can not be found
her constraints are taken at least in part from a study of text revisions made by expert editors
example 3e uses a simple imperative for the purpose with by conjoining participial forms of the intended actions
as previously mentioned the process structure will eventually become a fundamental source of procedural information for the text inquiries
the boldfaced names are systems the normal font names are features and the italicized names are realization statements
this yields strong support for the hypothesis and allows us to go on and discern what factors motivated the counter examples
the iterative process of hypothesis generation and testing can then be conducted on these other cases in a similar manner
a direct comparison to the unix style program is not possible for german texts as the style program is only effective for english texts
decision trees from a NUM item german training set from the sz corpus resulted in a tree utilizing NUM of the NUM attributes
the heuristics used within the system to classify unknown words can compensate for inadequacies in the lexicon and these heuristics can be easily adjusted
i or exatnple there is no hfl ortnal ion in its knowledge sources to make judgements about constr fint relaxal ion to process non literal expressions snch as metonymies and metal hors
moreover as already noted cow rage in depth breadth and size must all be achievetl in conjnnction with maintaining good me asnres of correctness et iciency and robustness
lb what extent do entries share semantic primitives or concepts to represent word meanings
what is the relation between the number of semantic primitives defined and the number of word senses covered
we have made a first step in formulating a more appropriate and complete set of measures of semantic coverage
our experience over the years has led us to the following sets of criteria for measuring semantic coverage
the limited depth constrains the ultimate potentia NUM of the systetn as a sentatd ic and pragmatic processor
tile number of different types of meaning elements included fl om the following set provides a reasonable measure of coverage argument structure only template filling only events and participants thematic role assignments time and temporal relations aspect properties attributes of events and objects relations between events and objects
the same theory can be interpreted from a perspective that allows more freedom u satisfaction or from a perspective that is tighter and that signals when some defeasible information has been canceled i and d satisfaction
the model ordering relation establishes m0 as the optimistic model for the theory because it contains as much information as ml and is easier to defeat
however when one examines utterance NUM one observes immediately that there is a contradiction between the presuppositions carried by the individual components
chris is not a bachelor presupposes that chris is a male adult chris regrets that mary came to the party presupposes that mary came to the party
among them the defeasible inferences that have been triggered on pragmatic grounds are checked to see whether or not they are canceled in any optimistic interpretation
corned mary party cornea mary party figure NUM model schemata for john does not regret that mary came to the party
wentd all boys theater figure NUM model schema for john says that some of thc boys went to the theater
answer NUM conveys a yes but a reply consisting only of NUM a would implicate a no
the semantics of lezical terms is formalized using the quantifier v while the necessary conditions that pertain to pragmatic inferences are captured using v trt
the models yield mary came to the party chris is a male and chris is an adult as pragmatic inferences of utterance NUM
answer NUM c makes the preconditions for implicating a no false and the preconditions for implicating a yes true
note also that in the case of a global inheritance descriptor the global context is effectively altered to reflect the new local context c
we can define a partial function a f u u the default interpretation of t n as follows
an informal introduction to the datr language is provided by example in section NUM the semantics of datr is then covered in two stages
value descriptors are either atoms or inheritance descriptors where an inheritance descriptor is further distinguished as either local unquoted or global quoted
the denotation function is defined as shown in figure NUM note that an atom always denotes the same element of u regardless of the context
the work described in this paper fulfils one of the objectives of the datr program to provide the language with an explicit declarative semantics
also we write see NUM verb to abbreviate the definitional sentence see NUM verb NUM and similarly elsewhere
for example sets of sentences relating to the same node are written with the node name implicit in all but the first given sentence in the set
to a first level of approximation the datr theory of figure NUM can be understood as a representation of an inheritance hierarchy a semantic network as shown in figure NUM in the diagram nodes are written as labeled boxes and arcs correspond to local inheritance or isa links
the language lacks many of the constructs found in general purpose knowledge representation formalisms yet it has sufficient expressive power to capture concisely the structure of lexical information at a variety of levels of linguistic description
rhetorical status selection sub network would realize if it were executed in isolation
it is rather an illustrative view of what the content and
determines the grammatical form of purpose expressions
14g tilt the pan so that the fluid drains out
order of expression of the action nodes
corpus it does not indicate ungrammaticality
this reflects the fact that global purposes are not expressed in phrasal form in our corpus
this can be addressed by querying all the fronted non global purpose clauses in the corpus
at the text level there is no definitive list of what might be called text functions
this section discusses the systems that have been included in imagene to distinguish among these contexts
square brackets indicate that all inputs must be true before entering the system on the right
finally we compare the output of the text generator with the text in the corpus
the deep semantic and pragmatic knowledge that was required for the current study has necessitated this
the slot of NUM this study has temporarily characterized these so that expressions as action sub action expressions
seleetional restrictions constraint relaxation information syntax semantics linking collocations procedural attachments for contextual processing stylistic parameters aspectual temporal modal and attitudinal meanings other idiosyncratic information about the word and finally the total number of entries in the lexicon
the semantics engine of this system is equipped with a large size ontology of over NUM NUM entries knight and link t99d which is nsed essentially as an am hot for mapl ing lexicm traits front the 8otlrce to the tar gel language
the notion of felicitously defeasible information is meant to capture the inferences that can be canceled without any abnormality as in NUM john does not regret that mary came to the party because she did not come
null by adding the extra utterance to the initial theory NUM uttered went ail boys theater one would obtain one optimistic model schema in which the conventional implicatures have been canceled see figure NUM
consider the following utterances and some of their associated presuppositions NUM the symbol t precedes an inference drawn on pragmatic grounds NUM either chris is not a bachelor or he regrets that mary came to the party
the lattice depicts the three levels of strength that seem to account for the inferences that pertain to natural language semantics and pragmatics indefeasible information belongs to the u layer infelicitously defeasible information belongs to the i layer and felicitously defeasible information belongs to the d layer
most boys theater d went many boys theater wentd all boys theater figure NUM model schema for john says that some of the boys went to the theater
that means that if an interpretation m0 makes an utterance true by assigning to a relation r a defensible status while another interpretation ml makes the same utterance true by assigning the same relation r a stronger status m0 will be the preferred or optimistic one because it is as informative as mi and it allows more options in the future r can be defeated
maled y z regret come y z cored y vv y z regret ome y
i male x a adultu z a married x vutx NUM bacheloru marriedi x vut x bachelor x adulta x vu x bacheloru x
since our translation strategy chooses the best match rule in the context if the condition of NUM is satisfied the rule NUM is chosen although the condition of NUM is satisfied too
