<?xml version="1.0" standalone="yes"?>
<Paper uid="C94-1058">
  <Title>DEFAULT HANDLING IN INCREMENTAL GENERATION</Title>
  <Section position="3" start_page="356" end_page="359" type="intro">
    <SectionTitle>
2 GENERAL DISCUSSION OF
DEFAULTS
</SectionTitle>
    <Paragraph position="0"> In the literature of norl-incremental generation, the need for defaults is hardly ever taken into account. The conunon point of view restricts the iulmt to be sulIicient for generation (see, e.g., the Te:ct Slructure by (Meteer, 1990) for a syntactic generator). In incremental gm,eration, most authors agree on the necessity of using defaults (see, e.g., (l)e Smedt, 1990; Kitano, 1990; Ward, 1991)). Nevertheless, they do not in sufficient depth answer the question of how to guide the processes of default handling an(l repair wil;hin a generator. This I~roblem is the starting--point tbr the following considerations.</Paragraph>
    <Paragraph position="1"> We assume tlm.t generation is a decision-making process witll the aim o\[' producing it phmsiMe ul:t(wance 1)ased on given information. As mentioned in section 1, there are cases where this I)rocess stops (caused by underspccifical.ioll of the input) before finishing its output.</Paragraph>
    <Paragraph position="2"> We define a module named d@tv.l! handh:r which tries to resume the process by giving advice to il;, i.e., by maldng assumptions about tile missing input specification. With respect to this task it is discussed  1. in whicll situations de.faults are applied (see section 2.1 ), 2. how default handling is integrated into a sb, st:cm (see sectiou 2.2), 3. Ilow 1.he knowh'dg:e for de\['ault handling is  ~h!scrihed (see. s&lt;'ctioll 2.3), and d. how assumptious arc cancelled when they I.urll out 1.o Im inconsistent wil,\], newly arriving input; (see section 2.4).</Paragraph>
    <Paragraph position="3"> hi incremental generation, as mentioned in section 1, interleaved input consumption and OUtlml. l)rodllcl.ion causes spc'cific default situations. An incremental processing scheme allows for an increase of elficiency and flexibility, e.g., by making tlm analysis and general.ion l~rocesses of a system \['or simultane,ous interl~ret, al:ion ow~rlap in Lime. There are two COmlml.ing goals of incremental generai;i(m for spoken oul, imt , thai, mnsl; be tal,:en  into account when estimating the usefulness of defaults: Fluency: Long hesitations should be a.voided during the production of an utl;erance, in order to be aeceptalde to the hearer 2.</Paragraph>
    <Paragraph position="4"> Reliabillty:Errors in an utterance may cause misunderstanding. In most cases, errors should be recovered by appropriate self-corrections a. Excessive use of self-corrections or erroneous expressions should be avoided because they decrease intdligibility of the utterance.</Paragraph>
    <Paragraph position="5"> Obviously there is a trade-off between fluency and reliability: maximal reliability requires 'secure' decisions and theretbre leads to output delay. On the other hand, maximal fluency necessitates the use of assuml)tions and repair, respectively.</Paragraph>
    <Section position="1" start_page="357" end_page="357" type="sub_section">
      <SectionTitle>
2.1 When to Trigger Default Handling
</SectionTitle>
      <Paragraph position="0"> We define as de.fa.~dt sit'ltatio~z the situation where a generation system has not yet finished the utterance but at the same t;ime has consumed all given iul)ut and is not ahle to continue processing. In non-incremental gel&gt; eration, this corresponds to the fact that the input lacks necessary informatiou, because the entire input is assumed to be given at one time (e.g., the undecidable number wdue of the example described in section 1 ). Thus, default handling should be triggered inunediately. null In incremental generation, however, tile system may gel; a new piece of infBrmation lai:er on that enables it to continue processing (e.g., the specification of a negation wdue + as outlined in the example in section 1). Tllerefore, possible alternatives are either to wait for the next inpnt or to trigger default handling. The former violates the fluency goal, the latter may violate the reliability goal. We propose the explicit use of time-limits .lot delay intervals 4 .</Paragraph>
      <Paragraph position="1"> =Ilumans often fill such pauses with filler.s like %r&amp;quot; or &amp;quot;what shall I say&amp;quot;.</Paragraph>
      <Paragraph position="2"> aSometimes, correct, ion is mmeeessary il' (I.he speaker believes that) the hearer ean inl&gt;r t.he intended utterance from errol~eous speech.</Paragraph>
      <Paragraph position="3"> 4An explicit parameter expressing tile desired degree of fluency influences the time--limit.s.</Paragraph>
      <Paragraph position="4"> Furthermore, the certainty of a &amp;fault is described by a value. As soon aM a default situation is identified, the certainty of the default is checked to see whether it exceeds a predefined threshold that: determines the degree of fluency/relia.bility r'.</Paragraph>
      <Paragraph position="5"> l'.;ach application o1&amp;quot; a default decreases the global certainty o\[' the system's state. Consequently, there should be a limit Jbr the mnzireal n'~,mbcr of dcfoults apl)licable to the same sentence.</Paragraph>
    </Section>
    <Section position="2" start_page="357" end_page="358" type="sub_section">
      <SectionTitle>
2.2 How to Integrate Default Handling
</SectionTitle>
      <Paragraph position="0"> Basically, there are two strategies to integrate default handling into ongoing processing.</Paragraph>
      <Paragraph position="1"> \])efanlts may be handled in a way that dif\[&gt;rs from the 'normal' processing of the system, e.g., as short-cuts. One. advantage can be an eflicient haudling o\[' defaults, lVurther more, the. designm' o\[' the de\['a.ult component is completely free in decidiug about the realization of defaults in the system. A disadwmtage is the di\[\[icull;y of providing consistency between del'aull;-caused and inputliceused processhlg.</Paragraph>
      <Paragraph position="2"> Alternal:iw~ly, l:l~e ongoing processing can deal with the de('aull: values in a.n ordinary manner (processing conlbrming default handling). This may he less efficient but guaran_tees consistency during processing, especially in case of a replacement by an input-licensed wdue. got incremental generation, tile system has to provide repair fa.cilil;ies in any case. So, they can also be used \['or nmt-monoloT/c nlodi\[ical,iOllS {~\[&amp;quot; d~q'aull, caused results. We lake this opl;iou in order to make the own'all system \]J.(~mogc,~co'tzs.</Paragraph>
      <Paragraph position="3"> 2.a How to Describe Defaults The knowledge source that is used for defaulg handling should provide the most plausible actions \[Br a dei'ault: situation. We represenl; the Iwlowledge as a sel: o\[' heuristic rules called dq/'a'tdl dc.~'(&amp;quot;~'ip/io~.s. A default description defines a. set el' ol)eratious t\]lal; should be carried '&gt;Pile I~asis for assigning cel'l,aill|,y wflues 1.o defaults should be a corl311s sl.udy l.hal, can be used 1,o find sl.al.isl.ical evidem'e for various Datures with all.ernal.iw' values (like Immb(% w)ice, ..., see, e.g., (Beck aim Warr~m, 1!~85)).</Paragraph>
      <Paragraph position="4">  out in a certain situation where the generation process can not be continued. A def'ault description has the following form: \[default default eerl&lt;ti',zt.y\] preconditions ~ body ; val,.c The set of defaull prceondilions defines tests that are applied to the given situal`ion in order to find out whether the corresponding default body can be acl,ival,ed. They include tests for the existence of parl,icular information, tests for the strucl`ure undc'r creation and tests for the sl,ate of processing.</Paragraph>
      <Paragraph position="5"> The default body describes how to continue 1)rocessing with de\['aull`s in an adequat.e way. For incremenl`al systems, we propose to express the body as a specification of input increments. An impotl,anl, prerequisil:e is that the size of inct'ements is defin(~d \[h'xibly enough to cope With varying amount;s o\[&amp;quot; information. Obviously, an import;ant; a(lvangage of this approach is homogeneity of l`\]te overall system. Especially, l;he homogeneous represenl,ation of default-caused and input-licensed sl`ructures is the easiest a.lld mosl` direct way l,o tesl` coincidences or conl`radicl~ious between defaull,-specified and hlpul; caused values. Ill section 3, this a.ppvoach is outlhled by differenl, examples. For non-h~cremenl`al systems, an operal,ional approach is pre\['ezable since l,here is no way to consume ad(til,ional input increments, presul)lmshlg l`hal; l,he input has been considered as a whole hefore a defaull` situation occurs';.</Paragraph>
      <Paragraph position="6"> If several defaull` precon(lil:ions are al)ldi-cable, the certainty vrtlucs for (lel'aull: d('scvil)tions are examined to find which provides l.he system wil`h the most l)lausil)le aci.hm.</Paragraph>
      <Paragraph position="7"> The individual default (lescril~tions should take into account the aloha\] constraints \['~)r processing stated in the knowledge sources of the system. For example, the assuml)l`iol~ ()f nominative case for a, German NI' comph'menl, can regularly be made only once \['or the aThe difl'erence between incremental aml non-incremental generation I)econles small(w, if we assume that defaults ira a non-incremental system can he triggered alter the system has only considered parts of its inlmt information, hr this case, tim input cousidel'ed after default han&lt;lling I)ec&lt;&gt;m~s COml&gt;aral)le to later increments.</Paragraph>
      <Paragraph position="8"> same verb. For reasons of homogeneity, the defaull, description should at least be compatible wil`h the specificat;ions of tile knowledge used \[Br basic processing. Ill order to guaranl,ee consisl:ency, def'aull` descriptions should merely contain whal: is orl,hogonal to tire basic 1,:nowledge sources.</Paragraph>
    </Section>
    <Section position="3" start_page="358" end_page="359" type="sub_section">
      <SectionTitle>
2.4 How to Cancel Defaults
</SectionTitle>
      <Paragraph position="0"> The repair of false assumptions is a crucial point of defaull, llandling in l`he contexl, of incremenl`al processing because the defaull, informal`ion does not remain locally bul` cap.</Paragraph>
      <Paragraph position="1"> cause \['m'lahe.r decisions of the system. Contrarily, for non--ilwremental input l`here will l)e no wdue gixq!n l.\]lal: can conl`radict default, Vgllrr(?s.</Paragraph>
      <Paragraph position="2"> As a \[h'st: step o\[' rc'.pah', inconsistencies l)el, ween \[lllml~--provlded and default:- camsed va.1ues are hle.nl, ified by simply real, citing the values. Tllen e\[\['ecl`s of l.he respecl`iw~ defaults are wil`hdrawn hll,roducing the inpul`-l)rovidcA values iul.o the sysl`eln. (\[enerally, a decision during general.ion il~fluences other decisions all over the system. Thus l`he ell'eel` o\[' a deI'aull, body may tw. propagal`t;d through I;\]te enl:ire sysl,en: (e.g., choosing a consl`ruc-I:ion o\[' ma.iu clause, wil;h causal subordinate clause influences l.he choice o\[&amp;quot; synl`acl`ic rca.liza.l`ions). null Roughly speaking, wil:hdrawlng a defaull` assuml~t.iou can be l'ea.lized l)y baekh'acking I`o the earlier sl`at,e o\[' the sysl`em where the d(d'aull~ had beeu inl`roduced or 1)y ~mu-mono/o,dc c/iruUics I.o t.he current, sl,ai.e o\[' the sys-I.em. The disa(lvalll;age of 1)ackln'acking is thai; imrt.ia\] resull`s are t:llrown away which could be reused duriug fm'I:her processhtg. Nonmonol.onic changes preserw; l,hese resull`s. Ill this framework, cancelling defaull`s requires l`lle sysl;em t:o idenl`i\[~, which resull;s are caused by del'a.ult, handling. 1)cpcnde.ney ling:s be-I.ween l`he immedial`e r'esull, of a defaulI; body and result:s of lshe influenced decisions allow for l`his ideul`ifica.l`iou. The disadvanl`age of non--monol:onic changes is I.he complexity o\[&amp;quot; COmlml.at:ion , e.g., SUl)lrorl`ed by a I;rul;h maint:enance system. When designing an increment.el sysl`em, simple lm.ckl,racking is ruled  out because the part of the sentence uttered cannot be withdrawn after it has been perceived by the addressee of the message r.</Paragraph>
      <Paragraph position="3"> So, we end up with a processing-conforming default handler for generation realizing repair by non-monotonic changes.</Paragraph>
    </Section>
  </Section>
class="xml-element"></Paper>