<?xml version="1.0" standalone="yes"?>
<Paper uid="W94-0319">
  <Title>Has a Consensus NL Generation Architecture Appeared, and is it Psycholinguistically Plausible?</Title>
  <Section position="8" start_page="168" end_page="168" type="ackno">
    <SectionTitle>
Acknowledgements
</SectionTitle>
    <Paragraph position="0"> I would like to thank Jean Carletta, Robert Dale, Michael Elhadad, David McDonald, Richard Kittredge, Tanya Korelsky, Chris Mellish, Owen Rainbow, and Graeme Ritchie for their very helpful comments on earlier versions of this work. It goes without saying, of course, that the views represented are my own, and that any factual errors are entirely my fault. This re4 Gould's various popular books on evolutionary biology, such as \[Gould, 1983\], give an excellent feel for evolution as an engineer-cum-hackers; see also the interesting discussion of language and evolution in \[Pinker, 1994\].</Paragraph>
    <Paragraph position="1"> 5Of course, the best way to do something on a machine is often not the best way to do it in nature; e.g., birds and airplanes use different mechanisms to fly. On the other hand, there does seem to be a remarkable congruence between effective vision processing strategies in animals and computers \[Marr, 1982J. One could also argue that since language (unlike flying) is purely a product of the human mind, any effective language processor is probably going to have to share some of the mind's processing strategies.</Paragraph>
    <Paragraph position="2"> search was mostly done while the author was at the University of Edinburgh, where he was supported by SERC grant GR/F/36750.</Paragraph>
  </Section>
class="xml-element"></Paper>