Enjoying reading this debate.
St John's Wort does contain an active compound within in it (hypericum). Some herbal remedies at the right dosage can and do work although the caveat is that a person may not seek appropriate medical advice, as arabella said, and risk interaction with other medications.
My understanding with homeopathy is that the "active" ingredient is diluted down so much that it's the treatment is simply water or a sugar pill with no active compound. To my mind that's why no good quality study has ever shown homeopathy to be effective beyond the placebo effect.
Sorry, just making the distinction between herbal (which may contain an unrefined active ingredient and consequently work) and homeopathy where there is no active compound.
Some herbal remedies have been shown to work in good quality clinical trials. St John's Wort is one of them (The Cochrane Colllaboration, which seeks to look for the highest level of evidence looked at this), but the problem is getting the dosage right. The positive effect was found to be greater than placebo, but less than other anti-depressants. However, on the other side, there were less side effects. I guess though the benefits aren't quite great enough to make it worth Big Pharma developing it, and do remember it does reduce the effectiveness of some hormonal contraception, which is a bad major side effect for many women and, being cynical, massively reduces the market for the product. The reason I am cynical is until I gave up work recently I spent the last four years as a marketing manager for a medical company. Although I worked in devices rather than pharma, I saw us scrap a number of products at the R&D stage because, although they worked and would have really benefitted a patient, the potential market wasn't large enough to repay the costs and make a profit.
Anyway, I digressed. As far as I can see there is no good evidence for homeopathy.
But that is my personal opinion and I certainly don't want to upset anyone.