<?xml version="1.0" standalone="yes"?>
<Paper uid="C00-1017">
  <Title>Probabilistic Parsing and Psychological Plausibility</Title>
  <Section position="2" start_page="0" end_page="111" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Language engineering and coml)ut~tional psycholinguistics are often viewed as (list|net research progrmnmes: engineering sohttions aim at practical methods which ('an achieve good 1)erformance, typically paying little attention to linguistic or cognitive modelling. Comlmtational i)sycholing,fistics, on the other hand, is often focussed on detailed mo(lelling of human lmhaviour tbr a relatively small number of well-studied constructions. In this paper we suggest that, broadly, the human sentence processing mechanism (HSPM) and current statisti(:al parsing technology can be viewed as having similar ol)jectives: to optimally (i.e. ral)idly and accurately) understand l;he text and utl;erances they encounter.</Paragraph>
    <Paragraph position="1"> Our aim is to show that large scale t)robabilistic t)arsers, when subjected to basic cognitive constraints, can still achieve high levels of parsing accuracy. If successful, this will contribute to a t)lausil)h; explanation of the fact th~tt I)(;() \])lc, in general, are also extremely accurate and rol)llS(;. Sllch a 1'o81111; Wollld also strellgthclt existing results showing that related l)robal)ilistic lne('hanisms can exl)lain specific psycholinguistic phenomena.</Paragraph>
    <Paragraph position="2"> To investigate this issue, we construct a standard 'l)aseline' stochastic parser, which mirrors t;he pertbrmance of a similar systems (e.g.</Paragraph>
    <Paragraph position="3"> (,lohnson, 1998)). We then consider an increre(total version of th(', parser, and (;v~,htat(; tim etf'c(:ts of several l)rol)al)ilistic filtering strategies which m'e us(,(l to 1)rune the l)arser's search space, and ther(;l)y r('(lu('(', memory load.</Paragraph>
    <Paragraph position="4"> rio &amp;,,-;sess th(; generMity of oltr resnll;s for more Sol)histi(;ate(t prot)al)ilistic models, we also conduct experiments using a model in which parent-node intbrmation is encoded on the (laughters. This increase in contextual information has t)(;(;11 shown 1;o improve t)erforlnance (.Johnson, 1998), and the model is also shown to be rolmst to the inerementality and memory constraints investigated here.</Paragraph>
    <Paragraph position="5"> We present the results of parsing pertbrmance ext)eriments , showing the accuracy of these systems with respect to l)oth a parsed corpus and the 1)aseline parser. Our experiments suggest that a strictly incremental model, in which memory resources are substantially reduced through filtering, can achieve l)recision and recall which equals that of 'unrestricted' systems. Furthermore, implementation of these restrictions leads to substantially faster 1)(;rtbrmance. In (:onchlsion, we argue that such 1)road-coverage probabilistic parsing  models provide a valuable framework tbr explaining the human capacity to rapidly, accurately, and robustly understand &amp;quot;garden variety&amp;quot; language. This lends further supt)ort to psycholinguistic accounts which posit probabilistic ambiguity resolution mechanisms to explain &amp;quot;garden path&amp;quot; phenomena.</Paragraph>
    <Paragraph position="6"> It is important to reiterate that our intention here is only to investigate the performance of probabilistic parsers under psycholinguistically motivated constraints. We do not argue for the psychological plausibility of SCFG parsers (or the parent-encoded variant) per se. Our investigation of these models was motivated rather by our desire to obtain a generalizable result for these simple and well-understood models, since obtaining similar results for more sophisticated models (e.g. (Collins, 1996; Ratnaparkhi, 199711 might have been attributed to special properties of these models. Rather, the current result should be taken as support tbr the potential scaleability and performance of probabilistic I)sychological models such as those proposed by (aurafsky, 1996) and (Crocker and Brants, to appear).</Paragraph>
  </Section>
class="xml-element"></Paper>