Agreed. I have faith that the markets will find an equilibrium, transaction processing has a cost which can be approximated for the network at any point in time, and a diff that sets it, diff can lower or raise to change the cost of securing the network, since it's all variable, human forces will work to strike a balance over the very long time they have to do so
I'm not sure how the difficulty figures into this. Difficulty doesn't set the tx fees, miners (at least in theory) do. It costs ~10%/yr to "keep the network secure" now. This level of security will continue to cost 10% of Bitcoin's market cap, regardless of Bitcoin price or mining difficulty.
Tangent: It's hard to say the network is really secured when the majority of hashrate is controlled by a handful of people (megamines/pool operators).
Difficulty determines the target hashrate to produce blocks at an average of 1 every 10 minutes. It sets the cost in hashes per second to keep the network running. If the difficulty goes up, the cost rises, if difficulty goes down, the cost lowers.
For every given combination of market price tag and transaction volume, there is a corresponding desired hashrate (and therefore difficulty) which keeps the network costs at a rough equilibrium. If BTC goes up in price-tag, the hashrate and difficulty rise to find a new balance, conversely if price-tag lowers then hashrate and difficulty reduce to find a balance.
The difficulty/target is perhaps the most important metric in many respects.