There are times when that applies and times when it definitely doesn't. It's almost impossible to construct a legal system which recognises that. I'm not sure if quoting people who lived over a century ago is useful when discussing modern patent needs though. In the time since then we've established a much more defined engineering process informed by scientific research. Inventors typically aren't dabbling in new fields, but have spent years understanding the background of the field in which they work.

There are many industrial processes, mechanical and electrical designs and drugs which would have been almost impossible to arrive at without extensive testing and trial and error. It's worth noting that hindsight is key. It's easy to look back at almost any patent and say "it's obvious that that nozzle design or molecule works like it does, because it fits these theories.

Fundamentally, legal protection of innovation is required for investment. The current patent system is certainly flawed, but to argue for the complete abolition of it doesn't recognise the significant investment required to advance many technologies.