Cascading activation in spoken word production drives incomplete neutralization: An internet-based study of Mandarin 3rd tone sandhi

Incomplete neutralization occurs when two underlying contrastive sounds are phonologically neutralized but remain phonetically distinct. Its proper understanding is foundational to phonology and speech production. Using the incomplete neutralization of the Mandarin 3rd tone sandhi as a test case (T3 + T3 → T2 + T3), we confirmed the presence of this incomplete neutralization with generalized additive modeling and growth curve analysis. Crucially, we found that the degree of this incomplete neutralization was reduced when speakers were additionally required to perform a concurrent verbal memorization task while speaking. We also found that the degree of incomplete neutralization was smaller in words than in non-words, and that words were more impacted by the secondary task than non-words. As the concurrent verbal memorization task is expected to add processing load and decrease cascading activation in spoken word production, our results suggest that cascading activation in spoken word production, which permits upstream distinctions to surface in downstream acoustics, drives incomplete neutralization. Implications for the phonological analysis of incomplete neutralization, the understanding of systematic phonetic variability, and theories of speech production are also discussed.