For the former, it's obvious.
Let's look at the starter station. (The shoestring go RER station I'm talking about). We don't currently (For the purposes of this paragraph) care yet about what it will become. Just that it'll be a station closer to the airport than Weston, that serves all rail transit users coming from either direction (Kitchener, Toronto). Let's call this "GO Woodbine" or "GO Pearson". For the moment, we're temporarily ignoring HSR, we're temporarily ignoring development (existing or future), for the purposes of just getting a RER infill station that services Pearson.
Brand new GO stations are typically in the the neighborhood of very roughly $50 million (Examples: $44M for James North). Sometimes it's lower like the $35M station portion of the $150M Stoney Creek rail extension (extra parallel track, etc). Numerous budgets show GO stations typically cost well under $50M each, with only a few larger complex station projects breaking through, or massive addons like Oakville's beached cruseship for cars (aka parking garage) but the government won't be providing free parking for airport travellers. Even the Weston unmanned UPX platform, smaller than a GO platform, is just a line-item in a ~$50M station budget including the slightly fancier Weston UPX waiting room.
Any basic starter GO station that gains the standard station name "GO Pearson" would obviously need to have a way to service Pearson quickly, and the best candidate would be the Woodbine racetrack simply because UPX spurs off to Pearson right after Woodbine.
Let's, for the moment, focus on where a GO (RER) station that contains "Pearson" in its name, and costs only $50M to build, will not be able to afford the cost of relocation of the UPX. No spur at Malton, no Pearson terminal relocation.
I now turn the tables -- how do you propose to install an equally inexpensive starter Pearson station that's not located near Woodbine racetrack?
This question is relevant because once "GO Pearson" station exists between "GO Weston" and "GO Malton", is relevant because, once this happens, it becomes one of the many possible candidates to be upgraded to a HSR station. They need the RER station much sooner than HSR, possibly two decades sooner. Many considerations occur, and by the time HSR arrives, it becomes more attractive (space for adjacent station without shutting down existing 15-min-RER service, some kind of developments already occurring, etc) it is likely whatever is named "GO Pearson" becomes the primary candidate if no UPX/LINK replacement occurs in a sequence before HSR (that connects to Malton, etc).
Metrolinx has a habit of replacing placeholder names. GO Woodbine is probably likely to be renamed GO Pearson if it interchanges with the UPX spur. See GO James North renamed to GO West Harbour (a geographical feature located further away). Metrolinx, also, happens to own and run the corridor and the UPX. So they call the shots on what cheap GO station receives the name "GO Pearson" -- that name becomes a neodynium magnet for future station expansions.
Sure, the Pearson Hub could relocate to Malton but before the HSR Pearson Hub, there may be 5 years, 10 years, 20 years of some kind of a Pearson-servicing GO station before HSR finally arrives to it. By then, it'll be a much stronger candidate than today, especially if the HSR station can be built adjacent to it without shutting down the existing station. I'm not betting HSR will be on time, so I feel there's quite a time window between GO Pearson and the HSR Pearson Hub.
Another scenario is if everyone can suddenly afford to build a $1bn station all at once, then yes, bulldoze Malton for I care and also build the new spur for whatever (UPX or LINK II or another link). But if we do that now, it will be empty, with so much redundant infrastructure. Best to start small, start a GO RER station, and go from there. While GTAA and Metrolinx (who brings the rail services) is hashing out details for any future megastation master plan, a RER infill is far more likely to occur many, many years before HSR service.
The starter station may be a completely independent location of what becomes the big HSR Pearson Hub (separate locations) but once a starter station exists, becomes named "GO Pearson", and starts being a magnet, it is already a much stronger candidate to be selected as the HSR station. And when we've gone that far, it might as well be the Pearson Hub.
But the question here, is specifically addressed to "which station becomes the starter Pearson station of some sort" (Tiny GO station? HSR station? Big Pearson Hub?), and I think the answer is obvious -- both from a financial and infrastructural sense from the Pearson perspective.
For this latter part, a correction:
"in 30 years, we'll have perhaps already redeveloped half of Woodbine"
Which of course, implies it could be a little, could be a lot. Similiar statements I've said in that post here (if I didn't word them correctly), should be interpreted as such. It's meant to help you understand. A progression, an incremental station expansion, since there's a gap between RER and HSR. If you built Pearson Hub all at once, sure, raze Malton -- as part of a megaproject to do RER+HSR+new spur. But if you're doing it incrementally, in the most inexpensive progression possible at first, that pretty much strongly limits station location choices.
Everything you speak of seems to be able to be accommodated at Malton saving the $50 million cost of a new proto-pearson station at Woodbine.
You assume that UPX would take passengers from Woodbine the rest of the way to the terminal (otherwise how will they get from pearson station to the airport). That would degrade the UPX service as it would have to stop at yet another station, part of their marketing is the speed between pearson and union. A station at Malton would require some way to move passengers to/from the airport, would building a second airport mover (or replacing the current link train) cost more than a new GO station?