And allow me to be the nitpicking twerp. In a transmission, one doesn't use oil or antifreeze. One uses hydraulic fluid with a specific viscosity that gets traction from the turbines. I also happen to think the simile is a bit misplaced.
The original point is that there are many things that happen or that we do that, in behavioral terms, are a punishment. The event or thing causes an interruption or decrease in the undesired behavior. What separates the "positive" trainer from the "balanced" trainer is that the positive trainer, while not against punishment, isn't going to use it just to be "balanced" or as a first resort.
To use my metaphor of tools I should mention that I have a 16 oz claw hammer. I also have a 15 lb sledge hammer. And I will use neither to install a switch cover plate, which requires a trim (small flat blade) screwdriver. And I don't have to justify the occasional need to drive in a ground rod to have the 15 lb "heavy artillery." Nor am I ashamed of having that big a hammer. It's there for just such the odd times when it is needed and only for the short time that it is needed. Otherwise, it sits in the garage, while I carry the precision tools in my normal tool bag.
I do believe some dogs require more of a phyical touch. I also happen to think that when a human is using a "correction" on the dog, it's more of a cue to the dog than a punishment. By original definition, if the punishment is needed more than a few times, it is not punishing or stopping anything and if the dog responds to its use every time, it is a cue to the dog. No matter how macho we think we are.
A dog does what it does because it is rewarding to the dog. If dropping a behavior in lieu of something more rewarding (+R, with a side dish of extinction) happens, that will happen. If doing a new behavior is accompanied by the cessation of nuisance (-R), then so be it.
In the same breath, a physical barrier is not always a punishment. Such as dogs that are escape artists. Or dogs that finally get away from an owner who is gung ho for collar pops and other physical means.
Also, I will venture to say that it is possible that some use of punishment might be similar in frequency to reward. To stop a behavior with punishment in one scenario might be successful. It doesn't mean that the punishment has been generalized. The dog may misbehave in a new setting just as easily as the first. So, using punishment to stop a dog from jumping on people might work at the house, behind closed doors, but that doesn't mean that the dog knows not to jump on people in public around other rambunctious dogs. That is a whole new world and a new set of behavior to learn or unlearn for the dog. Dogs don't always generalize the same way that humans do.
I trained "off" with treats. And Shadow will off, regardless of circumstance, and I have rewarded with highly valued rewards in various levels of engagement. Does that make the use of that command a correction, since I am "stopping" a behavior? Or does it mean that disengaging from whatever is more rewarding than continuing the behavior? I tend to think the latter. The "rules of engagement" are that breaking off and listening to me is always more rewarding than anything else, ever, amen.
Whether we rely heavily on corrections or rewards, we are always having to compete with the dog's environment. That's just a fact of life and if it's too much to bear, perhaps a person is better off owning fish.
That being said, a correction can be used effectively if timed with offense as it is happening and if the dog sees it as a punishment. If Shadow is barking at something I can bump his hindquarters or fingerpoke his neck and it doesn't hurt him and doesn't stop anything. It's a minor distraction and could lead to an accidental tooth graze. When a dog is keyed up, physical intervention is often not the way to go and not because we are afraid of the dog or that such "corrections" are "bad or unethical" but simply because they are ineffective. Getting the dog to an emotional state of everything being okay and in order is what will work. For many of us, the surest way to do that is to lead them with the rewards that they value most.
We also have to be in charge of their environment, to some extent. From the beginning. That is why Dunbar stresses training from the earliest moment possible. If all the dog has ever known is listening to you is always better than anything else, much of these problems can be avoided, such as never allowing the dog to think he has to guard you. Otherwise, he will guard for you. But if the dog was raised with the laws of the universe, the sun rises, the sun sets, the seasons change, and listening to Ron is the gateway to Heaven (or the dog equivalent) on Earth, then chances are likely that is what will happen.
Which requires diligence on the part of the human. Does that make us control freaks? Sure. And I'm really tall, too. And blonde and blue-eyed. So what? Certainly not anymore control freakish than someone who thinks that dogs are always attempting social dominance and view it as a never ending political contest. I think the difference is that a positive trainer is a control freak by getting the dog to follow them because it's in the dog's best personal interest to do so, as opposed to a dog following us to avoid pain, physical or emotional.