We have a capacity of around half a million transactions per day no matter how high the fees get. We are closer to that limit than we were before blocks filled up. Fees do not add capacity and only mitigate the problem without solving it. With less room to grow, we will grow less. How is that even controversial?
Your underlying assumption is that spam=legit tx volume=growth.
I do not assume the same thing:
https://blockchain.info/charts/n-transactions-excluding-chains-longer-than-10
https://blockchain.info/charts/n-transactions-excluding-chains-longer-than-100
This is the number of txs when you exclude coins that have been sequentially sent over and over again for >10 or >100 times.
This is like saying, I gave you 5$ in paypal, you instantly gave it to someone else, and he instantly gave it to someone else, and this happened over 100 times in a single day (with the same money). Sounds legit

Well, if you have people that do chains of >10, >100, >1000, by sending coins between their addresses or something, they are occupying the space of >100 or >1000 ordinary txs by doing so. This is only made possible by the extremely low fee structure. If fee competition is introduced in a meaningful way, for one such chain that is eliminated, you get +1000tx capacity for normal uses.
Legacy systems won't really allow that kind of abuse. By the time you've done that a few times with a bank wire or visa or paypal, the 5$ will be gone in fees.
There is no way to prevent that type of abuse without also preventing legitimate marginal use cases. It's an open protocol. A transaction limit is a ham-fisted, draconian anti-spam measure that costs more than it benefits. You either need to accept that spam is a cost of doing business, or you need to find a better anti-spam tool.
It's only a matter of time before the next spam attack, and that will doubtless trigger another sell-off. If I was an amoral opportunist, I'd do it myself and make a pile of money. Regardless, it's likely somebody will do it and smallblockers are giving that spammer the tool to use.