The evolving world of semantics

I promised come back to this.
First of all I have decided to move all tags to lower case, but that is appro pro nothing apart from the last post.
The thread of thought is in the notion of insight into the evolving world of semantics.
First of all that is not my professional world, so I only know a little about it. Still there are ideas to explore which I shall come to over time.
Further I will be unselfconscious in the statements I make and the questions I ask.
I don't mind seeming foolish, no doubt I am foolish.
This applies to all of my ramblings. Call me a fool if you will, it's OK.
An example of my foolishness would be in the use of tags.
What, actually, are tags for and what do they have to do with web 2.0/3.0 - the semantic web?
You do not have to be in this industry for too long to come up against received opinion, the twist is that received opinion here has been formed about things that have existed for hardly any time, e.g. tags, but the way of forming those opinions is the same as ever.
So someone may argue the benefits of tags and tagging as if there is something in them more than what is quite apparent about them.
Tags will be sold, and I use the word deliberately, as a semantic idiom, something that expands meaning on the item that has been tagged, that is to say, enriches our understanding of that item.
Tags will be sold as a facility that fits in with web 3.0 and complements it.
But I think it is just that, a sales pitch given in a work situation where advocates don't want to lose face and want, if they can, to sell a facility for the benefit to themselves of having that to their credit, without thinking too hard about what, really, the benefit is.
I am speaking of a real work situation I have witnessed here.
But my rule of thumb comes in here, that is where there is one instance there is at least 2 * one order of magnitude other examples that lie beneath the surface.
So here I am talking about the pitch for tag enablement.
- My rule of thumb is not for the situation where the phenomenon is absolutely ubiquitous. An example of the latter would be where Ben Goldacre, in Is this a joke? noticed a report about the connection between crime, imprisonment and re-offending.
His conclusion was, to paraphrase, that the authors had not understood, or possibly even read, the reports from which they formed their own report and that the only people who could be interested in this report would be those who had no interest in the underlying arguments.
That behaviour, I would say, is ubiquitous and not subject to my 2 * one order of magnitude rule.
Coming back to tags.
Really this depends on how they are used. In the context of a blog they aford a useful way of sorting information and filtering, and of displaying the basis for filters.
But I don't think that is more than adding a single flat association to each item in the first place. I don't think that tags expand the meaning through description of the tagged item.
I also can't see how they can be reasoned over as they have little of no relationship to each other.
Still there is a frustration here. Why is this? delicious does a very good job of suggesting tags now. MT (MovableType) does an even better job, more later.
What with AdWords and so, I would have thought google would offer something in that direction. But I think that is a bit of the problem. The sort of tag suggestions I might want wouldn't be the same as what might appear from AdWords, and the priority would be entirely different.
Something to think about.