I feel that creating a sitemap for a site of that scale may impede on the whole purpose for the sitemap which is to orient the user.
You can definitely bi-pass this with your search function by placing it in a primary location.
Here is an example applied to a knowledge-base. In this case, the designer has included some commonly used links as well.
If you need an example with scale, Wikipedia uses this method as well.
EDIT: After some thought, here is another idea with a case study. You may want to consider a Card Sorting exercise for the architecture. It can't all be under 1 site if the content can clearly be segregated.
Universities are notorious for having different web portals for each of its academic subjects, the same way they have separate buildings for them. You may want to consider splitting up the site itself into separate segregated sections each with its own image and unique messaging.
An example of this is the U of T website: http://www.utoronto.ca
University of Toronto is made up of a collection of Colleges each with its own web portal. There are also separate areas for Student Life, and Registrars information on top of each program.
Here are examples of their sub-sections with their own sites:
Their meta data contains the quick links and search in google:
Other industries that do this include the government with subsection sites for Ministry of Education, Ministry of Health, and Ministry of Transportation.
Personally, I just don't see Language students interested in being exposed to Math/Science related topics unnecessarily. My term for this is "Collateral Exposure".