Most wikis store their data in a database. These are not simple pages that you can download off the web server, they are dynamically created at the time you request them, using a number of queries to that database.
Finding out the size would be tricky.... You would need the total size of the database, plus any supporting files in the web accessible directory.
I suppose if you wanted to download all 2000 articles as they stand today, you could write a script that would query the database for each article, and download it to your machine. But to get to the revisions of each article, and to access the possibly deleted articles, you would need to understand the URL scheme of the wiki software in question. Then you could measure the size of all of those files.... But that may not give you an accurate idea of the size when it is all stored on the web and database servers.