Are you saying that you don't understand leaving everything as it is and let the free fee market do its job?
We should probably take this to a different thread, I'm already feeling guilty about going off on this tangent here. But quickly:
We already didn't 'leave everything as it is'. This is an entirely artificial system with a number of arbitary design choices. The 1MB block limit is an obvious example, and wasn't even originally intended to serve an economic function beyond limiting spam. It's not needed for a fee market: as block subsidy decreases fees must increase or miners will shut off. Thus, when the block subsidy is insufficient to pay for mining costs, miners are incentivized to require fees in order to accept transactions.
If you say X is the amount of transactions the network should, at most, be able to handle, you're making a policy decision. If you say there should be no limit and miners should decide whether they want to include 0-fee transactions, you're also making a policy decision. You can't pretend one is more of a "free market" solution than the other.
The network is considered roughly secure against economically sensible attacks when the cost of an attack is greater than the profit gained.In rough terms, the greater [network value] / [cost to gain 51% of hashrate] is, the less secure the network becomes. To increase security, you need to either decrease network value or increase the cost of gaining a percentage of hashrate. Both the small-blocks and large-blocks solutions essentially hope that a balance is naturally found where the cost of running the network doesn't hamper usage to the point that fees and subsidies aren't sufficient to provide security.
I don't have a solution here, but I think we're turning a blind eye to this issue of security vs. cost when that's the real question that needs addressing.