In the first instance, evolution does not require 'into a new species'. Why? Because the very notion of what is a species
is contested. Secondly, evolution merely requires that lovely mixture of mutation-adaptation-competitive advancement-reproductive benefit-advancement. There are a number of examples, though, of artificial speciation
that create two populations that do not interbreed.
The distinction between hard and soft science as a binary opposition is meaningless. Each science covers so broad a subject matter that it will have both hard and soft elements contained therein. Biology, for example, contains such diverse disciplines as biomechanics and evolutionary pscyhology.
Inherent in the distinction is the belief that one is somehow more genuine, better than the other. What we have here is, as Darklingalice pointed out, a problem of definitions. I cannot subscribe to Sure's understanding of science as allowing us to predict or control an outcome of a phenomenon is cripplingly limited. By this definition, statistics and logic aren't sciences as they don't directly address a predictable or controllable phenomenon. I should at this point like to point out that Newtonian physics is not ironclad. In fact, it fundementaly fails to predict the outcome of certain events (things going really fast and/or being really small). It means that Newtonian physics is, according to Sure, no longer science when things are very small and / or fast.
I think it's incorrect to determine what is science purely by its outcomes. There are certain non-scientific modes of conduct and analysis that allow for understanding or control over a phenomenon (I wouldn't have said that the ancient Egyptian practice of trepanning people with cerebral overpressure was based on an understanding of pressure, but rather that 'it just worked' or perhaps related to the relase of demons).
To this end, I'm with Popper
that what makes something scientific is the process. You can think scientifically about anything, be it atoms, people, wombats, or moods.
The idea that something is only science because it requires a lab is a fetishisation of empiricism.