Fair enough. I would suggest that a good working definition of 'works' in this context to be: "Produces things that are widely used and would not be available without it." That said, I didn't mean for this thread to be a debate on the merits of the scientific method. But I won't be upset or offended if anyone else wants to continue along those lines.
I am unsure if I would agree with your definition of works, as this would exclude a large set of systems which can be said to "work". Though ultimately my statement was poorly worded. The scientific method much like many systems of inquiry is more of a thought experiment than a practical tool. While many researchers and thinkers have applied the methodology successfully to create great advances, it become questionable if the method itself is to be commended for these advances or if the raw imaginative power of the human mind would have made the discoveries outside of such rigor.
So do I. Lay discussions of linguistic theories--well, I'm nerd enough to think that's almost as fun as picking flowers in someone else's garden.
I was seriously having fun when I wondered about just what, if any, 'structural' limits a lack of recursiveness in a language would place on people who spoke that language. That's not least because I've not been pointed to a solid riposte to Stephen Gould's The Mismeasure of Man. I'd very much like to know whether a lack of recursion in language meant that a group speaking only that language was at a 'cognitive disadvantage' to surrounding groups with recursive grammar systems.
The lack of recursion means that a language is semantically finite as long as it is conceptually finite. Let's look at language L with some grammar G. Now, we shall define a non-recursive grammer in the following way, Grammer G contains rules R. We shall denote A->B to mean that A contains a reference to rule B. For G to be non-recursive, two things must be true. First, that for all A and B in R, there is no such A and B such that A->B and B->A. Secondly, there is no such sets {X1,X2,X3,...Xn} and {Y1,Y2,...,Ym} in R such that A->X1->X2->X3->...->Xn->B and B->Y1->Y2->...->Ym->A. The first rule being a special case of the second wherein our two sets are both {empty}. Under this non-recursive grammar, our language L will always terminate unless G contains a rule set R which is both infinite and contains an infinitely long set of references. A language without semantic recursion would mean that concepts could not be circularly referenced. Thus like our syntactic recursion, one could only have a finite set of meaningful reference. In a larger sense, I doubt any structural limits would cause a significant 'cognitive disadvantage'.