Probably but I'm sure the technology would cost a lot more. Maybe if the miners were paid more.
Not really. The difficulty doesn't necessarily make Bitcoin any more or less secure. Difficulty doesn't make Bitcoin "work" - it could run perfectly fine if the network were made up of 80kh/s so long as it was very expensive to purchase/lease 1kh/s of hashpower.
Put another way - right now, it costs maybe $125 for 1GH/s with the algorithms BTC uses, and the total network hashpower is ~225TH/s. So, @ $125/1GHps the network "costs" around $28.13M.
Let's say it's the 90s and everyone's hashing on an AMD K6, and (out of my ass) let's say it provided 100h/s for $125. The NV1, RAGE and whatever was out, but my point works out whether GPU mining would've been viable or not. Anyway, the Bitcoin network could still do everything securely if, in total, it still cost $28.13M for the active hardware - it's just the total network hashpower would be something like 225KH/s (no guarantees on my math, there!). This wouldn't be a problem because difficulty would be EXTREMELY low (relative to now). The network would still be secure because it'd cost everyone else the same amount of cash for the same % of total network hashrate. The amount miners earm would basically scale, too - so 1kh/s in 199x's theoretical Bitcoin would earn the same as 1GH/s in 201x's Bitcoin.
Bitcoin doesn't become more secure based on hashpower, but based on how expensive it is to buy a % of the network hashpower. At its heart, though, Bitcoin's really secured by technological advancements in solving BTC's problems being made available publicly. BTC's been really lucky so far to see a relatively smooth transition from CPUs to GPUs to FPGAs to true ASICs without someone quietly developing the hardware and software behind closed doors and suddenly annihilating the Bitcoin network. That kind of scorched Earth attack would require a lot of remedial work on the devs' parts, rapid response by the network's infrastructure facilitators, and strong public awareness of a HUGE failure in Bitcoin. The network may be unusable for days or weeks, and it'd probably involve a ton of reversed transactions. In the same way the Bitcoin network is secured by honest nodes and miners making up the vast majority of service, it's also essential the network be secured by the vast majority of involved designers and developers being honest and open.
uhh... and on-topic @ OP Q (sorry - drinking night): I agree with D&T. It's kind of like asking if the ICBM could've been invented in the 20's. I can't think of a reason it absolutely couldn't (but there probably are some unsolvable problems not obvious), but it'd take many more stretches of the imagination to go from what they had to what they could do. The infrastructure was mostly there back then. DSL and cable Internet connections were becoming commercially available about halfway through. There were encryption schemes which probably would've worked just as well (for their time). The low hard drive capacity may've crippled growth without some type of hardcore pruning system, though (if BTC used MD5, how much smaller would current blockchain be?). Hard drive reliability back then would've been a much more significant problem than now, too. Maybe a 1997-1999 launch would theoretically work if "Satoshi" was about 10x as intelligent and made all the necessary foundational advancements himself. Much earlier, though, and there's probably too much missing to make it work.