Anthropomorphizing aside :rolleyes: , the point of positive (or reward-based) training is that you don’t need a negative stimulus reinforcement – you can simply withhold the positive stimulus (reward). Properly conducted, you can show the dog what you want and reinforce when you get it…and when you don’t, waiting a moment for the dog to figure it out will often produce the desired behavior, at which point you reward normally.
It may be that a given dog might not respond to positive training methods quickly enough to suit your taste. Even so, it seems to me that starting with positive training (and making a serious good-faith effort at it) gives a chance for an all-positive relationship to develop; if you find you must fall back onto aversive, punishment-based training, you can always do that afterward.
I have had very good results staying all-positive with our rescued dogs. A possible caveat is that we don’t know what else they were exposed to before they lived with us. They do seem to have had hard lives before living with us; maybe a dog who’s never suffered wouldn’t be as urgently inclined to please, for all I know.
Here is a website with a comprehensive article surveying the positives and negatives of the choice between aversive and reward-based (positive) training. Although the author favors positive training, she makes considerable effort to show both sides in a factual manner and to be honest about some of the difficulties of all-positive techniques, and she emphasizes that you should consult multiple sources and make up your own mind.
A big source of confusion about how the techniques differ is that reward-based techniques involve positive reinforcement but negative punishment (you withhold praise or reward or attention to correct a dog, thus negative in the sense that you are not supplying something) and aversive-based techniques involve negative reinforcement (you stop supplying an unpleasant stimulus to reward the dog) and positive punishment (you apply an unpleasant stimulus to correct misbehavior). Perhaps I should be using the terms “reward-based/aversive” since positive/negative is used so confusingly here.
She also mentions that some trainers mix and match, and specifically that Victoria Stillwell of It’s Me or the Dog uses some aversive techniques that do not involve dominance or pain (she’s probably talking about the sharp “ah-ah!” warning sound Victoria uses on problem dogs).
Despite the Shiba Shake website’s concern that aversive techniques may produce faster and more obedient responses, our dogs have done especially well with the critical, potentially life-saving commands: the recall (come here and don’t run into the road) and “leave it” (don’t eat that, drop it!). Difficult as it may be to believe, we used entirely reward-based methods for “leave it” and it’s their best response. They can ignore a tasty treat in front of them quite reliably.
Bottom line: the more you learn, from a variety of sources, the better-equipped you will be. I advocate reward-based, non-aversive techniques out of personal philosophy, because I learned from a great trainer, who works with rewards and the dogs’ natural body mechanics, and because they’ve worked well for me. I see no need to punish if one gets everything one wants without punishment. Other people may bring a different initial psychology, or a different set of less-successful experiences, to this question.
Very little basis in reality, lots of basis in ideology 
