On a practical level, it is doubtful that the user community would trust a closed source implementation with their money, when so many open source implementations exist.
If you are talking about wallet ripoffs, then you are probably right. But what about POS systems, ATM firmware, ERP applications, all which could use some bitcoin integration but are unable to open source for one reason or another. Maybe the codebase uses other third-party closed source code... maybe there are corporate license restrictions to conform to a specific industry law... maybe there are security implications for open sourcing the code.
The point is that there are so many untold reasons why someone may want to use the cbitcoin library in their code to further bitcoin adoption. Pigeonholing it with GPL unnecessarily limits it, especially considering that the LGPL was designed specifically with libraries in mind to fix this issue.
I think the main question is what is the end goal. To expand the uptake of bitcoin? Or make sure that developers of open source software have a competitive advantage over the corporations? If we were further along with bitcoin, I think you could make an argument for a GPL library. Unfortunately, unless those with closed source interests have an easy and cheap way into bitcoin, they aren't going to spend their money. Bitcoin isn't big enough.
If LGPL is going to be a no go... let me make a suggestion. How about a dual license. GPL and a closed source license where the closed source licensee pays some fee to use the library how they see fit.