One of the central arguments for a level-ordering of English morphology (Siegel 1974) is based on the nonoccurrence of certain pairs of affixes, e.g. *-ness-ic, *-er-ian, etc. The descriptive claim which has been made on the basis of this is that English affixes are divided into two sets, sometimes called LEVEL I AFFIXES and LEVEL 2 AFFIXES, and that level 1 affixes are not able to attach to a word to which a level 2 affix has already attached. This has been incorporated into Lexical Phonology/Morphology by having a word pass through a sequence of levels of representation as it is derived. Level 1 suffixes are first attached to the word, with each suffixation followed by the application of a subset of the English phonological rules, e.g. the English Stress Rule (Halle and Mohanan 1985); then level 2 suffixes are attached to the word, and other phonological rules are allowed to apply. This system encodes not only the ordering of suffixes, preventing level 1 suffixes from attaching to a word to which a level 2 suffix has attached, but also accounts for the fact that certain phonological rules (such as the English Stress Rule) do not take as their input words derived with level 2 suffixes. In fact, this is usually taken as the test of a level 2 as opposed to a level 1 suffix.The English data provide two well-known types of counterexample, which suggest that the generalisation about nonoccurring affix-pairs (from which the level-ordering theory draws much support) may be incorrect. On the one hand, there are BRACKETING PARADOXES, where a level 1 suffix appears to attach to the output of a level 2 prefixation; an example is un-grammatical-ity (where level 1 -ity must attach after level 2 unin order to satisfy categorial selectional restrictions). On the other hand, there are a few specific pairs of suffixes which involve a level 2 suffix preceding a level 1 suffix; for example -abil-ity, -ist-ic, -ment-al.These problems have sometimes been accommodated within systems which nevertheless retain level-ordering for the morphology. For example, Strauss (1982) suggests that suffixes and prefixes are not levelordered with respect to each other (thus dealing with the bracketing paradoxes). This paper looks again at how it can be predicted that certain pairs of Natural Language and Linguistic Theory 6 (1988) 527-539. O 1988 by Kluwer Academic Publishers.' This calculation is derived as follows. Working through the 43 suffixes in Table A, we find that 9 suffixes select for adjectives, and 16 form adjectives (hence 9x 16 = 144 combinations are allowed); 21 suffixes select for nouns and 21 form nouns (hence 21 x 21 = 441 allowed); and 13 suffixes select for verbs and 6 form verbs (hence 6x 13=78 combinations allowed). The total allowed is 144 + 441 + 78 = 663.
Link to this article: http://journals.cambridge.org/ abstract_S0022226700014420How to cite this article: Nigel Fabb (1990). The difference between English restrictive and nonrestrictive relative clauses.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.