I think that new askers are more likely to get their questions answered, and new answerers are more likely to have their answers actually help people, if they post questions and answers rather than comments. Here are some reasons why.
- Instead of commenting they are tempted to post "answers" which are nothing more than comments. These have to be reviewed and/or flagged for deletion.
While that's true, the mechanisms for alerting reviewers and other users to new answers are much more robust than for comments. A not-useful comment, for instance, can't be downvoted --- only flagged for moderator attention. It's much more efficient to have the entire community empowered to evaluate low-effort contributions from new or low-rep users than to ask the moderators to shoulder that entire burden.
I personally think that asking people to write paragraph-complexity answers and questions that five or ten other users find voteworthy to be a pretty low bar for the privilege of leaving a sentence-complexity comment.
Moderators can (and sometimes do) convert brief non-answers into comments, but I mostly encourage people to meet the participation threshold and come back.
- Instead of being able to ask for clarification of posted answers, they post a duplicate question explaining that they did not understand the answers already posted. The new question is flagged for review as a duplicate.
Consider the alternatives here.
Comment from a new user to an old question: "I don't understand any of the answers here, please explain more." Depending on the age of the question, its asker may never get notified about the comment. Commenting doesn't bump a question to the front page, so other users don't notice the new comment either. New user does not get help. The end.
Duplicate question from a new user: "I don't understand the answers over there, please explain more." Question appears on the front page and the in the First Posts Review Queue, where many people see it. Those people will vote and ask clarifying questions in the comments, possibly while the new user is still hanging around. Maybe the question gets closed, and maybe it gets answered, and maybe both.
"Duplicate" question from new user: "I don't understand the answers over there for reasons A, B, C." That's potentially a good new question.
The thing is that there's a path from (2) to (3), via editing of the new question, but the path from (1) to (3) is much more tortuous.
(Furthermore, as ACuriousMind says in a comment, I'm pretty sure that individual SE sites don't have the ability to muck around with the privilege hierarchy. So this isn't something that could be implemented only on Physics --- you'd have to drag the entire network along.)