"Free money", I'm not so sure about it. For any token to have value, there needs to be buyers. If those buyers are buying the new token spending old tokens, the two new tokens will in sum be worth exactly the same as the old one. For the "free money"-effect to take place, new fiat money needs to be poured into the market.

With Bitcoin Cash, I think there was new optimism at both sides of the schism. The Big-Blockers finally got their own token "following the spirit of Satoshi", they could transact between themselves with low fees just like in the old days. The Small-Blockers was happy that SegWit got launched, and also hoped that the Bitcoin Cash project would keep the "big-block-trolls" busy, easing tensions and conflicts.

In general, any split is bad. The major selling point of Bitcoin, as compared to any altcoin, is the network effect. I believe every Bitcoin Cash-supporter also can accept or send bitcoins. Hence the network effect of Bitcoin was not that much lowered with the introduction of Bitcoin Cash.

With Bitcoin Core vs SegWit, that's an entirely different beast - if there will be a lasting split, we may end up with a community split, with parts of the Bitcoin community totally ignoring the SegWit2X-chain, while many businesses and ordinary folks may be unaware of the Bitcoin Core chain. In the very worst case those two chains will have the same amount of users. In reality, I believe one chain will become dominant with the other one either being completely dead, or almost worthless and only used by enthusiasts.

I like to compare Bitcoin with IRC. IRC was the protocol for internet chat back in the 90s. Virtually everyone on the Internet that was doing chatting would be connected to the same IRC-network. Then, due to disagreements and arguments around some obscure protocol details, the community was split into two. The split pretty much followed geographic borders; as with the Bitcoin schism what matters probably wasn't the technical arguments as such, but whom to trust (and what other nodes one would already be peering with).