I'd like to see this theory put to the test. Get a number of differently-sized individuals record "see" or "saw" and test the ability of native speakers to identify which they heard.