E-Book Overview
Phonological Studies 10. Research Paper. Based on a comprehensive investigation, the author shows that some differences in English stress assignment are categorical, while others are non-categorical. The author argues that Partial Ordering Theory can properly accommodate such variation. Keywords: categoricality, stress patterns, English suffixes, Partial Ordering Theory.
E-Book Content
To appear in Phonological Studies 10, ed. by the Phonological Society of Japan, Tokyo: Kaitakusha.
Categorical and Non-categorical Variation in English Stress Assignment* Hideki Zamma Kobe City University of Foreign Studies ABSTRACT.
The purpose of this paper is twofold. First, based on a comprehensive investigation utilizing the
SOED, I will show that some differences in English stress assignment are categorical, while others are non-categorical -- a difference which has not been studied in the literature. I will further argue that Partial Ordering Theory (proposed by Anttila (2002)) can properly accommodate such variation. This analysis predicts another pattern of variation among stress patterns, which is actually observed in English, suggesting appropriateness of the analysis. Keywords: categoricality, stress patterns, English suffixes, Partial Ordering Theory
1. Introduction As summarized in Zamma (2003, 2005a), English has the following five major stress patterns; extrametrical (where stress falls on the antepenult if the penult is light as in (1a)); non-extrametrical (where stress falls on the light penult as in (1b)); non-retracting (where stress falls on the last syllable as in (2a-d)); strongly-retracting (where stress falls on the antepenult as in (3a)); and weakly-retracting (where stress falls on the penult if it is heavy as in (3b)). Representative suffixes of each pattern are summarized in (4). (1) a. (nátu) (húmo) (dómi) (áddi) b. alco(hóli) a(tómi) ti(táni) sym(phóni)1 (2) a. Jàpanése, Chìnése, Viètnamése, Pòrtuguése, jòurnalése b. ènginéer, vòluntéer, pìonéer, mòuntainéer, àuctionéer, pùppetéer c. àrabésque, Ròmanésque, pìcarésque, pìcturésque, gròtésque d. nòvelétte, kìtchenétte, màrionétte, màisonétte, cìgarétte (3) a. désignàte, démonstràte, cónfiscàte; sátisfỳ, récognìze, ánecdòte, ásymptòte b. ellípsòid, mollúscòid, stalágmìte, gelígnìte, eleméntary, perfúnctory, reféctory2 (4) a. extrametrical suffixes: -ity, -ion, -(i)an, -al, -ous, -ive, etc. b. non-extrametrical suffixes: -ic, -id, etc. c. non-retracting suffixes: -ese, -eer, -esque, -ette, etc. d. strongly retracted suffixes: -ate, -(i)fy, -ize, etc. e. weakly retracted suffixes: -oid, -ite, -ary, -ory, etc. These facts are not new, having been studied by many researchers such as Chomsky and Halle (1968), Liberman and Prince (1977) and Hayes (1980). In the literature, however, it is generally assumed that a suffix categorically shows one of the possible stress patterns.
Liberman and Prince (1977), for example, give an analysis in which -oid is assigned a Weak Retraction rule while -ate gets a Strong Retraction rule. Words which do not conform to these observations are simply treated as 'exceptions.' A closer investigation shows that this is not an adequate way of dealing with the stress patterns of suffixes. Surveying a large corpus of the Shorter Oxford English Dictionary (SOED),3 I found that while some suffixes do show such categorical stress patterns, others do not. In the latter cases, quite a few suffixes show variant stress patterns. We will see this difference in the degree of consistency in the next section. 2. Categorical and Non-categorical Patterns 2.1. Categorical Suffixes In (5), I counted the frequencies of words in the SOED having a particular stress pattern for several suffixes. The suffixes all show more than 90% consistency in their stress patterns. In other words, we can categorically regard -ity as