I'm a little late to the party, but here goes -- While it's true that people who stutter generally stutter less when talking with animals, very young children and to themselves, stuttering doesn't totally disappear. If you pay attention you'll likely notice that you will continue to stutter, but your disfluencies will be less frequent and less intense. The reason for this is not yet empirically known, but it likely has to do with a number of factors a) there is a reduction in social pressure when talking to self / infant / pet b) the propositionally of your message (fancy term for how important the words are to your message) is lower than when you're engaged in conversation c) when talking to pets and infants people often start talking in a sing-songy way (something that's going to reduce stuttering regardless of whom you're talking to). You mentioned singing in a reply, but singing is a different story -- there are, again, a variety of reasons why there is a remittance of stuttering during singing 1) uses different brain regions 2) music has a rhythm 3) vocal production of singing is slow and connected. Each of these are true, and it's probably a combination of all three that results in the remittance of stuttering during singing.