Pullum (1982), in refuting claims that there are languages that are nonconfigurational and hence radically different from languages such as English, presents an analysis that accounts for the data under consideration using mechanisms already in place for GPSG. He uses ID/LP format to account for free word order within a constituent, and shows how metarules can account for cases of apparent discontinuity of the elements of a constituent, all while maintaining a grammar that is equivalent to finite state phrase structure rules. On the way, however, Pullum takes the curious tack of casting doubt on the existence of the phenomena he ends up explaining. In particular he argues (p. 6) that Ross's (1967) theory of scrambling `had no acceptable empirical motivation' because his example was taken from Horace, who `is noted for stretching tendencies in the living Latin language beyond all grammatical limits', and was not even typical of Latin poets (p. 5). In this squib I refute the claim that discontinuity was an extragrammatical device restricted to a few radical poets, but show that it was a phenomenon widespread even in prose. Furthermore, I will show that the discontinuity was of such an order that Pullum's metarule cannot account for it. Thus data that has always been readily available to Western scholars will be used to counter claims that Pullum and others have made about the limits of discontinuity.
Available as a Postscript file (98 Kbytes, 12 pages).