Lets start with a small explanation about engineering and science. There is engineering and there is engineering science. Engineering science is like normal science only focused on engineering thing. For example, at the moment there is a lot of research in the structure of the internet. This research can be used by others to make recommendations and to create systems based on that idea, just like with research in every other scientific discipline. Engineering uses science in order to create something. Various things can be created, like cars, computers etc. But not all these things are for the public to know, like how to build a atomic bomb or how to genetically engineer a super influenza virus. These are good things to actually research, but you'll want to keep access to this kind of information restricted. Another point to note is that engineering mainly takes place in companies. Companies don't want everybody to know how their products work, while the things they used may be very interesting research topics. This also means that there was no public funding, so the company can do with the info what they want. I know that a pay wall is probably not the best solution, but engineering knowledge is not always good knowledge to give to just anybody. EDIT: added a small bit.
Exactly, if your work with a company is available to everybody then you are now worth much less to that company. It's also why you have to (generally) sign a non-disclosure agreement for these types of jobs, and I would say that these things help to foster competition between companies. That, and you don't want to make it even easier for other countries to reverse engineer your product.
Reverse engineering is exactly why some people don't apply for a patent anymore in the Electrical/Computer Engineering world. When there is a patent, reverse engineering a product and building a copy suddenly becomes a whole lot easier. And if you change the product just a little bit, the patent doesn't even apply.
I just looked it up for my field (Electrical Engineering). IEEE is the go to association for publishing in the EE field. When I probed a little deeper, I saw that one needs a subscription to read any of the articles. You even need to be a member of IEEE before you can get a subscription. So essentially, pretty much everything is paywalled. Even the standards, which is ridiculous. It's like having to pay in order to read the law.
IEEE membership is hardly exclusive. I was a mechanical engineering major once upon a time, and they try to hit me up for membership every year. I could then buy any IEEE article I wanted to. I'm pretty much "just anybody" as far as the engineering community is concerned.
IEEE does more than just EE. Software engineering and aerospace engineering are also subjects they publish about. And they need members. Otherwise there won't be enough funding for them to continue existing I guess.