I do not think there is any dispute about resource recommendations being on topic. But there is a difference of opinion about what this means.
Your interpretation of the policy appears to be that voting to close any resource recommendations questions (RRQs) because they are primarily opinion based (POB) or too broad (TB) is a breach of the policy. Routinely voting to close all RRQs as POB (or for any other reason) would indicate a misapplication of the policy, but there is no evidence that I or anyone else has been doing that. You have identified only one instance - I do not recall voting to close the Fractional Statistics question as POB.
ACuriousMind comments :
I see no consistent pattern of RRQs being flagged to close as primarily opinion-based (your two recent examples are the only ones among the 20 most recent RRQs), so I'm not sure there is something to understand here - I'd presume the flaggers thought that this particular question was too opinion-based and were not making some general statement about RRQs.
Your logic appears to be :
Premise 1 : RRQs are necessarily too broad and opinion based, as stated in your comment.
Premise 2 : The community has voted that RRQs are on topic - sentence #2 of your question.
Deduction : The community has decided that the criteria primarily opinion based and too broad do not apply to RRQs.
The error of logic is the false assumption in Premise 1. I think there is also confusion about the meaning of Premise 2.
Premise 1
I do not see Premise 1 expressed or implied anywhere in the policy Are resources recommendations allowed? In fact, it seems to me that the policy clearly states the opposite : that valid RRQs should not be opinion-based nor too broad.
Questions are required to provide adequate details of the kind of book which is required and to specify the level of understanding of the asker. Answers are required to be descriptive, specifying objective qualities of the book such as contents, style and pre-requisites. The objective is clearly that answers should provide recommendations suitable to the asker's needs, rather than promote the answerer's preferences.
The rationale given in Good list, bad list stresses that answers must be based on facts not opinions and that questions should be sufficiently limited in scope :
If these policies are in place, the major problems with books questions are reduced. They're much more relevant and useful now. They also are less "all answers are equally valid". They're still not conceptual, though, but this isn't so important if the answers are explanatory.
The intention of the policy is clearly to exlcude opinion and broadness as much as possible. Therefore it is entirely appropriate to vote to close RRQs (and to delete answers) which the reviewer considers to be primarily opinion-based or too broad.
Premise 2
Your interpretation of the policy seems to be that the community has voted to make RRQs inherently on topic. If this is correct, voting to close them for any reason is a breach of site policy.
My interpretation of the policy is that RRQs are not inherently off topic as a category, just as homework-like questions are not inherently off topic. Special requirements apply to both categories, in addition to the usual criteria which apply to all questions regardless of the tag, including primarily opinion based and too broad.
Many high rep users (including yourself) have voted to close RRQs for these two reasons. This shows that they do not agree with your interpretation of the policy.
Examples no more than 6 months old in which you (among others) voted to close as RRQs as too broad :
* Where should I start?
* On the State of Quantum Electrodynamics
* Good physics book to discover stuff by yourself
What are the open problems in the field of quantum thermodynamics?
Examples no more than 2 years old (in which you did not vote) which were closed as primarily opinion based :
Landau and Lifshitz or Goldstein classical mechanics
The best book for learning the renormalization group
* Improvement of Von Neumann's measurement scheme
Studying Quantum mechanics from Schwinger's book
Feynman Lectures for a newcomer
* Any good statistics books under the context of astronomy?
* Books on entropy
* Recommended reading about antennas?
* Learning physics from a strong math background?.
(Questions marked with an asterisk ask for suggestions of books or articles; others ask for comparisons between books or opinions of a particular book.)
Whether or not a particular question is too broad or primarily opinion based is a matter of judgement, on which any two reviewers may disagree. That is why there is voting : to try to average out such differences of opinion.
Moreover, questions are closed when 5 votes are cast for any reason. Reviewers might find multiple reasons for closing, but they can only select one. They might disagree on any particular reason, but the key issue is that they agree that the question should be closed.
Just because the opinions of other reviewers disagree with your own it does not follow that they are voting contrary to site policy.
Justification of my VTC Dependent Motion: Motion of Several Particles
The question of what is the "best book" on any topic can only be answered by expressing an opinion - although I admit it is ungenerous of me to take this part of the request literally. The question also specifies that the book should deal with "motor systems in real applications" which I interpreted as meaning that the OP is studying mechanical engineering. In common with many such requests, the OP displays no prior research effort. And the level of study is not specified.
So I can see 4 valid reasons for voting to close this question : (1) primarily opinion-based; (2) the request ought to be on another SE site; (3) insufficient research effort; (4) unclear what you are asking. The 3rd is not an official close reason, but has been used before for non-homework questions and is covered under "some other reason" why the question is considered "off topic".
Only one close reason is allowed per reviewer. I chose the one which came to mind first, and mentioned 2 others in a comment.
I did not select the strongest close reason in this case. Sometimes that happens by mistake or haste. The decision was then whether to retract my vote altogether, because I cannot change the reason, or to leave the matter for other reviewers to consider. I decided it was more important to put the question in review than to ensure that I only voted for the most convincing reason.
Insufficient research effort could also apply to Resource recommendation for fractional statistics : a google search using "abelian and non-abelian statistics" gives several promising leads on the 1st page, many of which are used in the accepted answer. The OP could easily have found the same resources himself.
I appreciate that you disagreed with my decision, but that is what voting is for. If others had agreed with me and voted to close the question, for whichever reason, I think that is a proper exercise of the democratic principle on which this site is founded.