I think that what your sister is getting at is the distinction between several kinds of discipline. Let's say that I want my child (let's call him Tommy) to eat a healthy dinner. How can I do that?
- "Tommy, if you eat (all of) your spinach, I'll let you have a cookie."
- "Tommy, if you don't eat your spinach, you can't have a cookie."
These are inverses of each other, but they both work the same way: if A then B, or if not A then not B. They're also what's at the root of this question, I think.
These solutions have a fundamental simplicity which appeals greatly to both parents and children. Mommy simply conveys the information to Tommy (what he needs to eat to get a reward), and Tommy knows how to get his reward. This will tend to work fairly often - Tommy might whine some about the need to eat spinach, but Mommy can reinforce this by simply restating the if/then exchange, and ultimately Tommy does or does not eat his spinach, and does or does not get his cookie.
The fundamental weakness here, however, is that Tommy's not learning to eat healthy foods: he's rather learning the above-mentioned Pavlovian response. This isn't entirely bad; it is one of the tools in the toolkit for parents, and hopefully ultimately Tommy will learn to like Spinach by eating it frequently enough that he gets used to the taste (and the often bitter tastes that make it objectionable to some).
However, Tommy isn't learning why he needs to eat his spinach, and he isn't equipped with intelligent tools for future decisions. He's not learning the real consequences of eating cookies without eating spinach; he's learning to follow your rules. Again, this isn't entirely a bad thing: learning to follow your rules is good, to some extent; and if Tommy is two or three years old, realistically he can't learn how to make good food choices yet, so the reward paradigm is reasonable.
It also turns it into a game: learn what the minimum necessary is to achieve the reward, and 'game' that further. Your child stops thinking about what he'd want to do, or even what's intelligent to do; he instead looks for what is specifically going to earn him the reward. Maybe he learns he likes spinach - but he still eats just the amount you tell him he must, and whines about it some, because he knows that you'll offer him the cookie. He's not going to get good habits from this - instead, he's going to alter his eating habits in a negative way to ensure the reward continues.
How you solve this, is to adjust this paradigm over time, and reduce the action/reward when possible as the child develops. The more reasonable way to put what your sister is saying is, "Avoid bribing/extortion when a superior alternative exists". This is important to learn as a parent, in large part because we get so used to the reward/extortion paradigm as early parents that we tend to stick with them later in life. Explaining the whys is important here; and as they get older, more and more often earning their buy-in on decisions is important.
One of those alternatives is to define rules, rather than rewards. Rules don't have an exchange or if/then; it's simply "then". "You must eat your dinner." Rules can be very helpful in establishing the boundaries of healthy decisionmaking: "Any decision you make is fine, so long as it has these limitations". For example, "You may feel free to eat anything you want for dinner, as long as it's healthy and doesn't require additional work on my part." That's a rule that enables the child to then make decisions - deciding what to eat, and how much - within the boundaries of what's necessary (eating something healthy).
Ultimately, teaching children to make good decisions on their own is more important than having them make the right decision every time, and this is why it's important to avoid the reward/extortion paradigm when you can - at least, think about it actively: is this something I can begin to teach as an intelligent decision? Children learn much faster than we think, most of the time, and just like you're amazed at how quickly they learn to read, or climb a ladder, you'll be amazed at how much they're able to understand about decisionmaking.
Here's an example of this kind of thinking; it's not intended to elicit a discussion of how to get a child to eat, but rather to show the different approaches.
In the specific example above, we often told our son when he was 2 that he had to eat a certain amount of his dinner before he could have dessert; basically method one. However, this quickly turned into the game above: every day he'd ask how much of his dinner he had to eat in order to get dessert. We realized this was a bad thing (as he wasn't learning to eat the right amount), so we changed a few elements of what we were doing.
First, we usually stopped telling him a specific amount. I still often had a specific amount, but didn't tell him up front; instead, we told him he needed to eat until he was full, and didn't directly link dessert. This might sound like a bad thing (and initially this is why we would tell him specific amounts - trying to be more open and clear), but in this particular case it backfired, so being less clear was a good thing.
Second, more importantly, we began teaching him with choices. Rather than "you must eat your dinner", it became that he must eat something healthy, and a reasonable amount; but if he doesn't like the food served, we would let him choose any leftover that was reasonable (ie, a complete meal itself). If he doesn't like meatloaf and peas, he's able to choose some lasagna and broccoli from the fridge. Also, we ask him to choose his portion size: defining the rules as he must eat what he takes to his plate, but he doesn't necessarily have to take a lot to his plate, for example - he's free to have seconds if he wants/needs more. Here we use the 'boundaries' approach; he's allowed to make choices within a framework we provide. He does still understand that he can't have dessert without passing this stage, so there's still some sense of a reward paradigm here, but it's as much as possible separated from this process.
Finally, we don't force him to eat a full meal if he's not hungry. Sometimes he might want to avoid dinner because he wants to play, and while he has to sit down for the duration, he will resist eating; sometimes he might actually not be hungry. That's fine - another teachable choice. Rather than reward/extortion ("you don't get dessert" or "you'll go to bed hungry"), he's free to not eat: but if he is hungry later on, his only food option is the option he had for dinner, no snack or dessert. This again encourages healthy decisions, as he knows he's not able to get more snack-type foods: it's not a reward or an extortion, it's simply a rule. It enables him to make healthy decisions by eating the correct amount - if he's not hungry, that amount might be very little, and then when he's hungry later he sees the consequence of not eating at dinner. Again, the boundaries approach here: basically, "You must eat some healthy food for dinner. You may choose what (from whatever is available to choose) and you may choose when (so long as you sit at the dinner table)." The reward (Eating dessert) still exists at the back end, on days where dessert exists, but most days it doesn't, and on most days that it does it's not explicitly stated.