This study examined how phonemic mergers affect word processing. A previous study found that exposure to regional sounds resulted in quicker recognition times (Dufour, Brunellière & Nguyen, 2013). We explored if listeners can recognize the difference of a merged sound in speakers with and without mergers. Data was collected via online survey, where listeners responded to audio from speakers with different accents. Participants were asked to select one word from a pair of homophones. Results show listeners selected correct answers 20% more often when hearing the speaker without the merger, which can help develop virtual assistants to understand more sounds.