The distinction between logical and punitive consequences is one of those parenting hinges where small wording differences carry large structural ones. A logical consequence is connected to the action by a chain of reasoning the child can follow: you spilled the juice, you wipe the juice. A punitive consequence is connected to the action by parental will alone: you spilled the juice, you lose television tonight. Both produce a negative outcome after the unwanted behavior, and from a behaviorist standpoint they might look similar. From the child's standpoint, and from the standpoint of what is actually being taught, they are different operations.
Jane Nelsen's positive discipline tradition codifies the difference through what she calls the three Rs: related, respectful, reasonable. A logical consequence is related to the action, delivered with respect for the child as a person, and reasonable in scope. A punitive consequence is typically unrelated, often delivered with disrespect (sarcasm, contempt, or cold withdrawal), and frequently disproportionate. The three Rs are not a stylistic preference; they are a structural test. Strip the test, and most parental sanctions fail it. The child who refuses to wear a coat and then loses screen time is being punished. The child who refuses to wear a coat and then is cold is experiencing a logical or natural consequence. The behaviors look identical from the outside; what differs is whether the parent's response is part of the structure of the situation or an addition imposed on it.
The Law 4 frame is about which kind of consequence the household has actually planned. Most households default to punitive consequences not because the parents prefer them but because logical consequences require more thought in advance. You have to know what a related consequence would be before the situation arises, which means thinking through the household's recurring friction points and pre-deciding the structure of response. Without that work, the parent in the moment reaches for whatever lever is closest, usually screen time or dessert, neither of which is related to the actual behavior. The punitive default is not chosen; it is what happens when planning is absent.
The deeper case against punitive consequences is not that they don't work; in the short term they often do, in the same shallow way that reward charts do. The case is what they teach. A punitive consequence teaches that adults with power can inflict pain at will, that the connection between action and outcome is the will of the powerful rather than the structure of reality, and that the appropriate response to authority is calculation about likely punishment rather than reasoning about consequences. These are lessons children learn very efficiently and carry into adulthood, where they manifest as either compliance with authority or chronic adversarial stance toward it. Neither is what most parents want for their children. Logical consequences teach a different lesson: that actions have meaningful relationships to outcomes, that the world has structure the child can learn and work with, and that being held accountable does not require being demeaned.
The hardest part of the distinction is that punitive consequences feel satisfying to parents in the moment, and logical ones often don't. A child who has spilled juice and is being asked to wipe it up does not feel "punished" in a way that resolves the parent's anger; the child is simply doing a reasonable cleanup. The parent's anger has nowhere to go, which is uncomfortable. Punitive consequences, by contrast, discharge the anger into the child's experience, which feels like justice. This is why the punitive impulse is so persistent: it serves the parent's emotional needs more than the child's developmental ones. Recognizing this is the entry point to actually using logical consequences. The shift requires the parent to handle their own affect somewhere other than in the consequence delivered to the child. Once that's possible, the choice between logical and punitive becomes a real choice rather than a foregone conclusion.
A workable practical heuristic: if you cannot explain the consequence to the child in a sentence they can understand, where the connection between action and consequence is plain on its face, the consequence is probably punitive. If you can, and the sentence has the form "because you did X, the natural response is Y," and Y is proportionate, you are probably in logical territory. The test is not perfect, but it screens out most of the worst defaults.