I think LrC would do it very nicely. Especially as keywording can be applied minimally at first but more intensively where and when it is required. It's not like there are only a few photos involved, so the expense of buying LrC would be minor compared with the time that will be devoted to the task; unless he realizes the task is probably huge and he quickly abandons it.
I would suggest that the very first step is for the drives to be clearly numbered or named so they can be quickly found again later on. I mean the actual drives as well as the boxes they might be in. Folders and files from each one should go into folders of the same names within the LrC library with no attempt to reorganize anything this early. This is important, or will become so later on.
Potentially many files will have matching names but that will not matter if the original folder structure is maintained with unique folder names for each source drive.
It is important to keep all of the drives and files within the one LrC catalog. That's the only way to be sure of correctly finding and grouping related files later on without extreme effort.
If intending to develop a new folder structure, the trick is to not try to create the right final structure in advance - leave that until after he has found where the important files are and where the big folders are. Indexing helps greatly with this, as will simple browsing of thumbnails within LrC.
I recommend using keywords extensively but selectively. Do not put keywords into the files - let LrC manage them in its catalog, especially at this early stage. That's what it is so good at. Until there are more than about 2 or 3 screenfuls of thumbnails per keyword then a new keyword is not required to help isolate the better pics. This approach makes it easy to start very high-level and apply detail only where it is needed. If he goes the other way around then the job will never be finished because he'll burn out while creating the ultimate keyword structure before he's loaded the pictures, and the super complex keyword hierarchy will be so much harder to navigate.
Keep in mind that the indexing gradually lets him find stuff by ever more specific topic.
Also keep in mind that keywords can be repeated with different meanings / contexts in different parts of the LrC keyword hierarchy. That, for me, is a very powerful advantage of LrC over a lot of other photo software.
Absolutely do not overwrite or rename or relocate anything on the source drives if it can be avoided. Let the LrC keywords do all the work.
Almost a separate issue; keep image editing to an absolute minimum, and only in LrC.
1. Within LrC, only do the most basic edits - like enough to see an underexposed but precious image, or show that it was a dud.
2. Do all editing in LrC parametrically - without changing the original files. LrC is great at this. That will save tons of backing up edited image files because the files are not being changed.
3. Set up LrC to automatically compress and backup the catalog at the end of every session.
4. Backup those compressed backups to other drives that are not always connected. This is not overkill. If he loses some edits and cannot retrieve them then it is natural to try and recreate them, but that usually takes far longer than the original edits did. If he loses a heap of edits then depression could set in and/or he could lose interest in a worthwhile project.
5. Eventually thin out the older backups because they lose merit after enough changes have been made; they'll simply be too far out of date except in a dire emergency. In short, the time and effort needed to bring an old catalog up to date is greater than start over in the most recent catalog.
6. Do not keep all the backups on the same drive as the master catalog because that's far too risky.
7. Avoid editing photos too early because there could be a bunch of better photos on the next drive, but also avoid deleting poor photos too early because there may no others available.
I reckon that if he takes this top-down approach without bogging down in the fine details of editing or keywording too early, then he'll maintain interest for long enough to see some great progress. That in itself is encouraging and rewarding. Eventually, he'll learn enough about editing that he'll be able to fix even complex photo problems - but by then he'll already know which photos are worth editing.