On the phonological status of Spanish compound words

Primarily inspired by Bustos de Gisbert (1986), Hualde (2006/2007), and Moyna (2011), specifically by their comments on stress deletion in the left members of some Spanish compounds, the current investigation fills a gap in this field by conducting an acoustic analysis of fundamental frequency evidence of stress in Spanish compounds and phonologically classifying them through the Autosegmental-Metrical model. Using a data elicitation task of 30 noun + noun compounds demonstrating syntactic, semantic, orthographic, and phonological variation, eight speakers of Mexican Spanish recorded two iterations of each compound, which is embedded at the beginning or at the end of a declarative carrier phrase. The acoustic analysis reveals that, as expected, right members categorically exhibit F0 evidence of stress. However, only the 15 compounds written as two orthographic units show strong evidence of left stress. This is especially so in utterance-initial position. The number of unstressed syllables between left and right stresses determines the tonal sequences produced in left members. In compounds spelled as two orthographic units, the phonological targets of each member resemble those of two simple content words in broad focus declaratives. Evidence of left stress in orthographically united compounds occurs in less than 20% of cases, and these are viewed as carrying postlexical secondary stress. The tonal sequences of this group are more wide-ranging than those of the first group. The results have implications for language processing and raise questions for the study of compound stress in Spanish.