Block size, if handled correctly, will end up having nothing to do with scaling and everything to do with regulating compensation for miners. If miners are not being given enough incentive to provide sufficient security to the blockchain than blocks are too big, if the opposite is the case than blocks are too small. Scaling will have to come from second layer solutions. Some transactions will always need to be on chain, the value that people place on putting transactions on chain needs to be enough to compensate the miners, and in the absence of the ability to regulate demand, the only alternative is to regulate the supply of on chain space inorder to increase its price to the level necessary to accomplish this.
Yet we have known since Adam Smith that the hidden hand of the marketplace will ensure that the optimal solution is arrived at by letting prices find their own equilibrium. This has been more formalized in more recent times in that the price and quantity of a good will be set at the point where the demand/price curve and the supply/price curve intersect. Further still, dead losses are incurred any time production quotas are enforced. Given this, why do you assert the utility of a production quota on transaction throughput? Why not trust miners to set the tx throughput supply to maximize profit under the demand and supply curves?