I'll try to break down your concern.
Are the questions with the tag low quality?
I have skimmed through the question list in the tag, and I see most of them are highly upvoted. The lowest-positive-in-number-of-vote questions (namely some questions with 3 or 4 votes) have enough detail to safely say that they have good quality. If we can keep this attitude when the site comes to public beta, I don't think we have to worry anything.Should personal experience be allowed in answers?
This has been discussed in a separate question: http://meta.languagelearning.stackexchange.com/q/87/168Should I use my own experience as answer?. In short, yes.If non-peer-reviewed sources are allowed, will questions only allowed for answers with reference hurt the process of transferring knowledge?
I think yes. Peer-reviewed studies obviously have more quality in personal experience. As in the answer of mine in the above question, when we talk about knowledge, any opinion should be allowed to speak. After all, what is said in a highly cited and peer-reviewed paper from a reputable in-field researchers is opinion.But we don't want to have low quality post, nor the OP only wants to have answer with reference!
That's fine. If the OP only want to have referenced answers, let they show their choice by their accept tick and votes. They can even say this in the question that "answers with scientific evidence are preferred". If the community wants to get rid the low quality answers, let they show their choice by their votes.So, do we need to keep the tag?
I don't have a strong objection to the tag, so either ways are fine for me. I even think that having a tag is convenience, we can click on it to find questions that really have peer-reviewed resources, quite handy for linguists I guess. However, I don't think any restriction in answer is good in any how.
It's the question that the OP asks, so they have the right to choose the answers they want. However, don't forget that once it shows up in the internet, it also belongs to the knowledge, and we should not prohibit the flow of transferring knowledge.