Comparing dogs to humans doesn't really work.
While sometimes it gets the point accross, the "would you work for free?" analogy... the people who use compulsion to train will say "do you give your 3 year old kid a cookie everytime he does what you ask?". Well obviously the answer is "no".
When it comes to dogs, the only factor to consider is "what makes a dog learn?". Well (drumroll please).... REPETITION. Dogs are pretty stupid, simultaneously to being smart. But the intelligence of a dog is usually judged by "how many times does a dog have to repeat a behavior before he realizes that is what he is supposed to do?".
You can force a dog to repeat a behavior, or you can motivate the dog to repeat a behavior. If you force the dog to do so (like pushing their butt down to teach a sit for example), you are creating a certain level of conflict, because the dog starts out resisting you. If you hold a treat over the dogs nose and they sit down, there is no conflict because the action of sitting is simply a side effect of following the motivator over the top of their head.
At the end of the day, both techniques will put the dog into a sit, but the real question is: Which method makes the dog enjoy it more? And therefore more likely to perform.
Well this is easily witnessed by looking at compulsion trainers in the working dog world. How many dogs have titled in sport successfully without food, balls, tugs etc, AND without a prong collar, e-collar or other physical correction method? I dont know of any. But I do know of dogs who have titled without ever receiving a collar correction or worn a prong.
So what does this say about the way a dog thinks? Simple: The dog who does not have a strong incentive to listen to you (like food or a ball) are more likely to give you the finger when they figure out there's no consequence for doing something more interesting. The dog who is trained motivationally is constantly thinking about his reward, and how he will not get it if he doesn't listen. It is a form of correction, but one that carries far less conflict between dog and handler than compulsion.
Then there is this whole "the dog wants to please me" stuff. Usually this "will to please" I see from dogs that are compulsion trained, are simply dogs that are trying to find a way to please their handlers to avoid another correction. There are dogs out there who genuinely want to do what you say, without food and without corrections, but this is hardly "the norm". What many people consider "will to please" is actually submissive behavior that usually follows punishment or the threat of punishment. The "will to please" in dogs trained motivationally is more like "will to please themselves". Dogs are selfish animals (although not as bad as cats). If I am in the kitchen making a cheese sandwich and my dog sits next to me all cute wagging his tail, do you think he is saying "daddy I love you, you're awesome", or is he thinking, "he always tells me to sit before I get cheese, so I'll sit, maybe I'll get some, yummy, cheese cheese cheese, mmmm.... cheese, ooh look at that cheese. hmmm hes not giving it to me, maybe if i shake paw he'll give me some CHEEEEEEESE." Yes, the latter, exactly! As I mentioned, there are dogs who genuinely thrive on praise. Heck, I own one. But when not using corrections and only using praise, it is difficult to get consistent obedience. To make this work you need to "turn on" the dog, for example, by acting very goofy and animated. Works wonders for a dog that responds to it, but when you're out in public and you tell your dog to heel, you want the dog to get in heel position and heel, you don't want to have to act like a complete goofball to get the dog "in the mood" first. It works for teaching a dog, but it is difficult to maintain consistency and reliability, unless you are in a unique situation where your dog only has to listen when you can act like a dork
When you're not acting excited and you tell the dog to sit, the sit will be slow and unmotivated, if the dog sits at all (unless the dog has learnt there is a correction coming for disobeying). The dog is not thinking "oooh if I sit, maybe he'll act like a dork as soon as I've sat down!!".
Manipulating a dogs selfishness into obedience is a far less stressful way to train the dog than to force them to do something that maybe they don't feel like doing at that particular point in time. So the dog ignores you, and you have nothing to offer the dog that is more interesting than his distraction, so what do you do? You correct the dog, which puts him into submissive avoidance. He is avoiding a correction, a softer dog will duck his head down, a very soft dog will just lie down completely, a hard dog will either bite you, or say "is that it?" and continue to ignore you.
So what do you do if the dog is incredibly hard and incredibly stubborn because his motivation is something that you aren't controlling? (i.e. his distraction, perhaps a piece of cheese...)
Either you control their motivation, or you find a way to correct them that registers. This is where people with over the top dogs get into the realm of sharpened prong collars, high level e-collars, and various other creative ways to punish a dog that doesn't respond to a regular prong collar.
Perhaps all this could have been avoided if you were the one who controlled their motivation, rather than having to compulsion the dog into ignoring the motivation they chose for themselves when ignoring you?
Used correctly, food and toys are hands down the best way to train a dog. The problem is when people use the food and toys as a bribe, rather than a reward.