On a New Conception of the Metrical Structure of Greece

Recent descriptions of the metrical structure of Greek postulate the existence of rhythmic stresses which are added at the surface level to preserve an even rhythm. No phonological evidence is presented in support of this claim and sometimes there is disagreement between the analyses about the location of the rhythmic stresses. It is claimed here that Greek does not exhibit rhythmic stress, a possible exception being cases that violate the Stress Well Formedness Condition (SWFC) of Greek i.e., words stressed on the (ante)penultimate followed by one or two clitics. For example, /simetoxitu/ his co-participants. Results of a perceptual experiment and acoustical analyses of natural speech support this claim as: (a) words like /simetoxitu/ his co-participants and /simetoxitu/ his participation, heard in segmentally identical contexts, are correctly disambiguated by Greek native speakers in 99.5% of the cases, although proponents of rhythmic stress claim that such pairs are metrically as well as segmentally identical and thus indistinguishable; (b) such word pairs are also acoustically distinct as far as the presence of stress on the host’s antepenult is concerned (/me/ in our example). It is concluded that Greek speakers do not impose stresses for purely rhythmic purposes; secondary stress is triggered only by the SWFC. Since Greek does not have rhythmic stress, it follows that, unlike English, it does not normally exhibit an even alternation of strong and weak syllables. It is suggested that Greek is best described metrically without reference to the foot level and using n-ary instead of binary branching. “Enclitic stress” is explained by the SWFC only. Clitics are metrically presented as forming compound Stress Groups with their host