<?xml version="1.0" standalone="yes"?>
<Paper uid="C94-2212">
  <Title>NL Understanding with a Grammar of Constructions</Title>
  <Section position="2" start_page="0" end_page="0" type="metho">
    <SectionTitle>
IBM Research l)ivision
</SectionTitle>
    <Paragraph position="0"/>
  </Section>
  <Section position="3" start_page="0" end_page="0" type="metho">
    <SectionTitle>
Abstract
</SectionTitle>
    <Paragraph position="0"> We present an approach to natural language understanding based on a computable grammar of const~ctions. A construetionconsists of a set of features of form and a description of meaning in a context. A grammar is a set of constructions. This kind of grammar is the key element of MINCAL, an implemented natural language speech-enabled interface to an on-line calendar system. Tile architecture has two key aspects: (a) the use of constructions, integrating descriptions of form, meaning an(t context into one whole; and (b) the separation of domain knowledge (about calendars) from application kno'wledgt; (about the particular on-line calendar).</Paragraph>
  </Section>
  <Section position="4" start_page="0" end_page="1290" type="metho">
    <SectionTitle>
1 Introduction: an overview
</SectionTitle>
    <Paragraph position="0"> of the system We present an approach to natural language understanding based on a computablc gTummar of constructions. A construction consists of a set of features of form and a description of meaning in a context. A grammar is a set of constructions. This kind of grammar is the key clement of MINCAI,, an imt)lemented natural language speech-enabled interface t() an on-line calendar system.</Paragraph>
    <Paragraph position="1"> The system consists of a NL grammar, a parser, an on-line calendar, a domain knowledge base (about dates, times and meetings), an application knowledge base (about the calendar), a speech recognizer, a speech generator.</Paragraph>
    <Paragraph position="2"> In this paper we describe two key aspects of the system architecture: (a) the use of constructions, where instead of separating NL processing into the phases of syntax, semantics and pragmatics, we integrate descriptions of form, meaning and context into one whole, and use a parser that take, s into account all this information (see \[10\] fin' details); (b) the separation of the domain knowledge (about calendars) and the application knowledge (about tile particular on-line calendar).</Paragraph>
    <Paragraph position="3"> *M. Szummer and S. Jarecki ~re also from MYF.</Paragraph>
    <Paragraph position="4"> The dialogs The system allows users to engage in dialogs like: Schedule a meeting with Bob/ At what time and date? On August 30th.</Paragraph>
    <Paragraph position="5"> At what time.? At8.</Paragraph>
    <Paragraph position="6"> Mm~ing or afte~won? In the evening.</Paragraph>
    <Paragraph position="7"> TILe parser recognizes Schedule a meeting with Bob as an instance of sent(imp), the imperative construction consisting of a verb and an NP, here up(event). TILe context is used to prevent another reading in which with Bob modifies schedule, as in l)ance a tango with Bob/. That is, a contextual rule is used which says that for calen(lar applications, peotlle do not modify actions or places. Context also plays an important role in understanding answers, e.g. At 8. This is understood as a time expression (and not t)lace or rate or something else) only because of the eontcxt.</Paragraph>
    <Paragraph position="8"> The tlarameters of a meeting can be given in many ways, e.g. synonyms or differe, nt constructions can be used, users Call in(hide as many parameters in a senten(:e as they wish, and the parameters can be given in any order. As a result there are about 10,000 ways of scheduling meetings (with a given set of parameters). null How are the dialogs understood With respect to parsing, grammars of constructions can be parsed like &amp;quot;standard&amp;quot; grammars, except that tile set of features is richer. Civen a string (representing a sentence, a fragment of a discourse or a paragraph), the parser assigns it a construction. From this viewpoint, the situation is similar to &amp;quot;regular&amp;quot; parsing, and the t)ossible algorithms arc, similar. We have implementect a prototype chart parser for construction grammars, disensse, d fllrther in Section 3. But, c, learly, having understood tile sentence as a linguistic entity in isolation is not the ultimate goal. Here the message of all utterance mils( be understood in tile context of an intended action. This is clone in two steps. First, the system determines the intended  action and its parameters, using domain knowledge (meetings+time+places). Second, once all the parameters have been extracted from tile dialog, the system executes the action. To do this, the program uses application-specific knowledge to translate the action and its parameters lute a form that can be executed by the application (Xdiary).</Paragraph>
    <Paragraph position="9"> 2 Constructions as data structures null A construction is given by the matrix:</Paragraph>
    <Paragraph position="11"> The vehicle V consists of formulas describing presence (or perhaps absence) of certain taxemes, or features of form, within the structure of the ('onstruction. Such a structure is given by a list of subconstructions and the way they have been put together (in all our examples this is concatenation, but there are other possibilities, e.g. wrapping). The context, C , consists of a set of semantic and pragmatic constraints limiting the application of the construction.</Paragraph>
    <Paragraph position="12"> It can be viewed as a set of preconditions that must be satisfied in order for a construction to be used in parsing. The message, M , describes the meaning of the construction, via a set of syntactic, semantic and pragmatic constraints.</Paragraph>
    <Paragraph position="13"> To make this concrete, let us consider a few examples. We begin with a simple &amp;quot;command construction&amp;quot; consisting of an action verb followed by its argument.</Paragraph>
    <Paragraph position="15"> The context of the construction describes all situations in which the the hearer hr (hmnan or machine) is paying attention to the speaker sr (a &amp;quot;ready&amp;quot; state). The feature struc is a list of variables and/or words/tokens; it is used to describe the structure of a construction, and its role is similar to a rule in a generative grammar. (We will write names of variables in capital letters, e.g. NP, inside matrices of constructions). The attribute cons_n gives the name of a construction that could be assigned to a string. We use it here to say that the form of the construction can be described as a concatenation of two strings, of which one is a verb ((;onstruction) and the other an nt) (construction). l,hlrthermore, the verb type &lt; V M vAype &gt; is &amp;quot;action_verb&amp;quot;. (The expression &lt; V M v_type &gt; should be read &amp;quot;tile v_.type of the message of V&amp;quot;).</Paragraph>
    <Paragraph position="16"> The message M describes the meaning of tile construction as that of a command in which tile type of action is described by tile nmaning of the verb, and the object of the action is given by the meaning of the noun phrase. The attribute sex_type stands for the &amp;quot;semantic type&amp;quot; and we identify it currently with tile word sense. Thus &amp;quot;erase the file&amp;quot; is understood as a command to delete the file, if &lt; erase M sex_type &gt;= delete, but &amp;quot;erase the picture&amp;quot; might refer to the type of action associated with rub_out. In both cases the hearer hr is supposed to be tile agent of the action.</Paragraph>
    <Paragraph position="17"> Constructions: from words to discourse Words~ phrases, and fragments of discourse can be analyzed as constructions. We view languages as collections of constructions which range fi'om words to discourse. We claim that the same representation scheme can be used for all constructions.</Paragraph>
    <Paragraph position="18"> The examples we are going to present; have been developed with a specific purpose in mind, namely for scheduling calendar events. In other papers (\[10\] and \[6\]), we have presented examples showing that we (:an give a good descriptions of non-standard constructions. However, in either case descriptions of meanings and contexts are general, and hence applicable to other tasks.</Paragraph>
    <Paragraph position="19"> We uow turn our attention to words. Tile verb &amp;quot;cancel&amp;quot; can be represented as follows:</Paragraph>
    <Paragraph position="21"> Notice that even simple words (prop'orly) interpreted. In C require context to be we say that English text is expected (but in other cases it could also be l,Y=ench text, or 1,Y=ench speech, etc.). Some aspects of context do not have to be explicitly specified and cart be replaced by defaults.</Paragraph>
    <Paragraph position="22"> Although the vehicle and the message are both very simple in this example, the simplicity of tile message is a result of deliberate simplification. We have restricted it to the specification of the scream tie type, identified with one sense of tile word, and to describing the xmrb type of &amp;quot;cancel&amp;quot; as a verb of action. Notice that the other sense of&amp;quot; cancel&amp;quot; -&amp;quot;offset, balance out&amp;quot; wouht appear ill another entry. Of course, in reality, tile lexical meaning of any word is a much more complicated matter \[1\]. For instance, in our lexicon tile messages of words may contain many of the attrilmtes that appear in the  explanatory combinatorial dictionary of Meleuk \[7\]. Discourse constructions: qb illustrate discourse constructions, we consider the folh)wing dialog: Have you arranged the worn yet? No, but I'll do it right away.</Paragraph>
    <Paragraph position="23"> We view the pattern of the answer no.but.,5' as a discourse construction. It can represented by the following array of features:</Paragraph>
    <Paragraph position="25"> At we can sue, the construction applies only in the context of a previously asked question, and its message says that the answer to the question is negative, after which it elaborates tile answer with a sentence S.</Paragraph>
  </Section>
  <Section position="5" start_page="1290" end_page="1291" type="metho">
    <SectionTitle>
3 System Architecture
</SectionTitle>
    <Paragraph position="0"> The parts MINCAL consists of a NI, grarmnal', a t)arser, a tit)main knowledge base (about dates, times and meetings), an on:line calendar (Xdiary), an application knowledge base (about Xdiary), a continuous speech recognizer (IBM, ICSS), a speech generator (Sl)eech Plus, Text to Speech Converter), and the interfat:es. At present, the grammar consists of a few hun(\[red lexi('al constructions, and about 120 &amp;quot;productions&amp;quot;, i.e. constructions describing combinations of ether constructions. ~ it (:overs tim basic fornm of assertive sentences, but it emphasizes (}ontntan(ts.</Paragraph>
    <Paragraph position="1"> Thus a comman(t can, for exanq)h:, be given either by v.np (also with &amp;quot;please&amp;quot;, or &amp;quot;kindly&amp;quot;), or' by art assertive sentence re('ognized as art indirect si)ee(:h act (&amp;quot;I'd like to ...&amp;quot;, &amp;quot;Leora wants you to ...&amp;quot;, etc.). The next large group of constructions (:overs PPs, with particular emphasis (m time and places. Finally, it covers a few discourse construetions, sin('c it is important to deal with sentence hagnmnts in dialogs, e.g. understanding &amp;quot;evening&amp;quot; as &amp;quot;in the e, vening&amp;quot;, when it is art answer to the question &amp;quot;when?&amp;quot;.</Paragraph>
    <Paragraph position="2"> The interaction of the modules &amp;quot;\]Phe calendar and the application knowledge base: Xdiary is an on-line calendar for which we have not written a complete interface, lint have focused on the three most iml)ortant funct;ions: appointment, moving, and canceling appointments.</Paragraph>
    <Paragraph position="3"> Other functions, su('h as &amp;quot;to dC' lists, window mana.gement, listing somebody's aptn)intments, etc., (',art  tend the interface to deal with them. At this point tile apl)lieation knowledge t)ase in very simple. \]t consists of rules that say how to interpret the data given by the semantic interpreter, for instance the rules for tormatting paranmters and re, naming slots (e.g. event_duration - ~ duration). Such rnles are ne(:essary, if the distinction between at)plication att(t domain knowledge ix to be, maintained.</Paragraph>
    <Paragraph position="4"> The (tomain knowledge base: This has two kinds of facts: (I) background ontology, i.e., is, basle facts about time and places, and (2) linguistic knowledge associated with tim domain. The former includes slteh obvious facts as the number of clays ill a inonth, which month folh)ws the ottmr, that oitlcos are places etc. The latter includes tiu'ts about how tile language is use(l, l,br examt)le , the filters saying that places do not modify people, so that I want to meet my manager in th.c cafeteria can be, unambiguously parsed, with &amp;quot;cafeteria&amp;quot; 1)eing a meeting place, and not an attribute of the nlanager.</Paragraph>
    <Paragraph position="5"> The organization of knowledge: The issue of the organization of knowledge has been discussed at h:ngth in \[8\] and \[9\] and the formal model tleveh)pe, d tiler(: ix applicabh; in the present context. At this t)oint, however, this tbrmal model has only been implenmntcd very crudely. Still till; model is worth briefly discussing, because the concet)tual (lis.~ tinetions made guide our work and have important t)rat'tiea\] eonsequene, es. The most important thing about it is that we discard the model of t)aekground knowledge as a logical theory, and replace it by a model consisting of collection of theories and nmehanisms for putting thenl together depending on eireantst.:~nces. T\]UlS, the )lSllal, two-1)art h)gical structures, consisting of a metalevel and art object level , are augmented by a third level a referential level.</Paragraph>
    <Paragraph position="6"> The referential level is a t)artially ordered collection of theories; it encodes background knowledge in a way resembling a dictionary t)r art encyelopedia. ~ Parser, construction grammar and linguistic knowledge Parser: The parser does not produce (syntactic) struetural descriptions of sente, nces. Instead, it computes meaning representations. For example, it converts adjuncts directly into attributes of place, time, t)articipant etc., once they can be computed, and thus tit(; message of the sentonee does not contain any infornmtion about how these attributes where expressed or about tim attachnmnt of Pl's that appear in it. 1,'or example, the sentence I want you to arrange a conference in my office at 5 is analyzed as .sent(a,ssert, svoe), an assertive sentence consisting of a su\[)jeet, a vert), an object and a eomplement.</Paragraph>
    <Paragraph position="7"> :?As llStl~tl, cllrr(!llt situations are described on the object level, and the metalevel is a t)lace for rules that can elilniil~ite some of t;he models permitted by the (~bjet:t level and the retkwential level.</Paragraph>
    <Paragraph position="8">  The latter and the message of the imperative that is passed to sent(assert, svoc) does not contain any structural information about the attachment of the PPs. This message is combined with the messages of the verb and the noun, yielding  Application-specific defaults then produce yet another interpretation where, in addition to filling the slots of Xdiary, \[ hour \[ 5 am or_pro\] \] is interpreted as \[ hour \[ 17l\].</Paragraph>
    <Paragraph position="9"> The parser is a chart parser, working left to right, with no lookahead. The grammar is L-attributed, i.e., has has both synthesized and inherited attributes, but each inherited attribute depends only on inherited attributes of the parent or attributes of the sisters to the left. Hence, although the parser does not have a lookahead step at present,, such a step can be added following \[2\].</Paragraph>
  </Section>
  <Section position="6" start_page="1291" end_page="1291" type="metho">
    <SectionTitle>
4 Comparisons with related
</SectionTitle>
    <Paragraph position="0"> work Linguistic arguments for constructions-based grammars has been worked out chiefly by Ch. Fillmore and his colleagues (ef. \[31). Their motivation for advocating such an approach comes from the fact that typical generative theories of grammar cannot deal tu:operly with open idioms illustrated by constructions such as: The morc carefully you work, the easier it will get.</Paragraph>
    <Paragraph position="1"> Why not fix it yourself? Much as I like fgonnie, I don't approve of anything he does.</Paragraph>
    <Paragraph position="2"> It's time you brushal your teeth.</Paragraph>
    <Paragraph position="3"> Him be. a doctor? The same is true about even so-called robust parsers of English. The reason for this failure can be attributed to the fact that expressions like these &amp;quot;exhibit properties that are not fully predictable from independently known properties of its lexical make-up and its grammatical structure&amp;quot; \[3\], p.511. However we do not need a list; of &amp;quot;strange&amp;quot; constructions m conclude that ~horoughly integrating syntax with semantics and pragmatics could provide us with a better handle on natnral language understanding. On a closer examination &amp;quot;normal&amp;quot; construetions exhibit enough complexity to warrant the new approach (see \[10\] for details).</Paragraph>
    <Paragraph position="4"> Jurafsky \[4\] has independently come up with a proposal for a computable grammar of constructions. We compare our work with his in \[10\]. Here, we limit ourselves to a few remarks. What is common in both approaches is the centrality of the concept of grammatical construction as a data structure that represents lexical, semantic and syntactic knowledge. However, there are important differences between the two formalisms. First, the actual data structures used to represent constructions are different. The most important differenee has to do with the presence of the context field in our version of the construction grammar. This allows us to account for the importance of pragmatics in representing many constructions, and to deal with discourse construetions. null Secondly, while Jnrafsky acknowledges the need for abstract constructions (pp.43-51), his abstract constructions (weak constructions) are not first class citizens they are defined only extensionally, by specifying the set of constructions they abstract over, and their abstract meaning (e.g. entity for NOUN).</Paragraph>
    <Paragraph position="5"> They are used to simplify descriptions of constituents of other constructions. However, because they do not have a separate vehicle part, they cannot be used to assign default meanings. For instance, since verb is defined as a collection of all verbs is + read + cancel 4- know + look-up + ..., it cannot be assigned a feature action_verb without introducing a contra/292. null diction its semantics is therefore given as RELA-TION/I'I~OCESS. For us the important feature of &amp;quot;abstract&amp;quot; constructions is not that they simplify (tescriptions of other constructions, trot that they have default meanings. (A similar critique of \[5\] can be found in \[10\]).</Paragraph>
  </Section>
class="xml-element"></Paper>