?page= produces fictional pages in Drupal

How to make these pages as clean URL's ?
i.e. : domain/page/22

I have found similar problems, like domain/nodename?page=5 - as I noted here: http://drupal.org/node/475256

One can remove the ? from robots.txt - but this doesn't prevent situations where one gets

domain/HEADLINEOFPOST?page=5
&
domain/NAMEOFTOPIC?page=7890009

The top one being the issue. It's possible that the topic pages may have 7890009 pages returned as result (or at least had in one point in time) but a single post itself never ever ever has more pages. So why does the "?page=" thing even go there? How can i remove the "?page=" part (301 redirect so that searchengines quit indexing them)? How it ended up in searchengines such as google - which has collected many many "?page=1 through ?page=16" on the same exact post/node - is a mystery to me. Nothing points to these URL's, is googlebot guessing?

Disallowing "?" or "?page" from robots.txt may prevent search engines from linking content further down in the site, which is a little bad (though sitemap may help them find the individual pages anyway, which is probably better for search results rather than a bunch of "topicname?page=7" results whose content changes as you post more)

I'm using the following modules: Pathauto and Global Redirect as well as clean URL's. I have yet to figure out which modules is causing this behaviour, any clues welcome.

Nobody has any idea why this is happening?

yes, this is HTTP.
Not sure where the links are COMING from though.
And if you worry about Google-juice because of amateur SEO speculation, A suggested solution could be 301 redirect or cannonical url that are intended to flatten these multiple hits.

Let's raise the priority of this issue a little bit.
I will explain, why this is important.

For example we have a node with 30 comments and 20 comments per page setting.
let node be our first page
node?page=1 is our second comments page, and we want it to be indexed, therefore we can't block it in robots.txt
node?page=2 ... what's that? actually, a copy of node?page=1
where did this link come from? I do not know.
But the search engines keep finding the pages with ?page=3 and so on
And as I said earlier, I do not want to use robots.txt page blocking because that will block existing comment pages.

In fact, I've found how i happens that Drupal returns those pages as existing. Please have a look at this code from in;cudes/pager.inc in function pager_query

$pager_page_array[$element] = max(0, min((int)$pager_page_array[$element], ((int)$pager_total[$element]) - 1));

Here $pager_page_array[$element] is the page being requested, $pager_total[$element] is the number of pages we have.
As a result, requesting for pages with greater numbers results in the last page as well.

My suggestion is to modify the Drupal core and to insert before this line:
if ($pager_page_array[$element]>=((int)$pager_total[$element])) drupal_not_found();

What do you think?