Newton's laws are testable. An example of a modern, high-precision solar system test of the first law is Battat 2007. The Eot-Wash group has done various high-precision laboratory tests, such as searches for coupling of spin to a preferred frame. The third law has been tested to high precision in Kreuzer 1968 and Bartlett 1986. For modern tests of the first law, the keyword you want to search for is "Lorentz violation." For the third law, try "local position invariance."
It's particularly absurd to claim that Newton's laws are only definitions and not testable, since they have already been falsified by experiment. Special relativity contradicts Newtonian mechanics, so all the experiments that verified SR demonstrate that Newton's laws constitute a real, falsifiable theory. Similarly, quantum mechanics falsifies Newton's laws.
Some textbooks may, e.g., present the second law as a definition of force or as a definition of mass, but this has no effect on the validity of experiments that test Newton's laws. This is because the concepts of force and mass carry other baggage, such as conservation of mass and vector addition of forces. Someone who prefers to think of the second law as a definition of force will consider such an experiment to be a valid test of some other claim or set of claims of Newtonian mechanics, such as vector addition of forces.
With the first law, you have to be careful because Newton formulated it in one way, but later people have formulated it differently: History of interpretation of Newton's first law . In a modern context, we think of it as having to do with Lorentz invariance and the lack of a preferred frame.
Bartlett and van Buren, Phys. Rev. Lett. 57 (1986) 21, summarized in Will, http://relativity.livingreviews.org/Articles/lrr-2006-3/
Battat 2007, http://arxiv.org/abs/0710.0702
Kreuzer, "Experimental measurement of the equivalence of active and passive gravitational mass," Phys. Rev. 169 (1968) 1007, http://bit.ly/13Z6XAm