Date of Award


Document Type

Open Access Dissertation


Communication Sciences and Disorders

First Advisor

Dirk den Ouden


The Covert Repair Hypothesis (CRH) is an account for speech errors in normally fluent speakers, and also hypothesizes errors in the phonological encoding stage in people who stutter (PWS). Previous research has shown that PWS exhibit poorer performance compared to typically fluent adults (TFA) on linguistic tasks designed to tap into the level of phonological encoding, such as phoneme monitoring. Stuttering and cluttering often co-occur, thus the field can benefit from extending this methodology to study people who clutter (PWC). Experiment 1 in Chapter 2 used phoneme monitoring to study phonological encoding in PWS and PWC, with three conclusions: (1) slower performance by PWS; (2) increased errors by PWS compared to TFA; and (3) similar performance by PWC compared to TFA, suggesting that PWC do not exhibit difficultly with phonological encoding at the single word level. One criticism of the CRH is that the cause of errors in the speech plan has not been accounted for. Chapter 3 proposed the Near Neighbor Interference Hypothesis (NNIH) as an account for errors in the speech plan in PWS, which hypothesizes that due to a lifetime of word-substitution behavior to avoid stuttering, semantic neighborhoods of PWS may be organized differently than TFA, with more neighbors and/or stronger connections between neighbors. Chapter 3 tested the NNIH by investigating the effects of the number of associates (NoA) and the degree of relatedness on performance during lexical decision. Previous research shows TFA respond faster to words with a high vs. low NoA, and words preceded by a picture with a high vs. a low degree of relatedness. Following from the NNIH, it was hypothesized that the magnitude of these effects would be greater in PWS. In Experiment 2, both groups responded faster to words with higher NoA, but PWS were slower to respond than TFA overall, regardless of NoA. In Experiment 3, PWS were not overall slower than TFA, and the effect of degree of relatedness was actually stronger for TFA than PWS. Together, these results suggest that rather than experiencing a benefit from more semantic neighbors, it appears PWS may experience interference from these additional neighbors. Overall, results suggest that PWS may have errors in their speech plan that originate prearticulatoraily, potentially at the lexical-semantic level, and are passed down to the phonological encoding level.


© 2015, Emily O’Dell Garnett