Canonically, ‘classic’ tests of general relativity (GR) include perihelion precession, the bending of light around stars, and gravitational redshift; ‘modern’ tests have to do with, inter alia, relativistic time delay, equivalence principle tests, gravitational lensing, strong field gravity, and gravitational waves. The orthodoxy is that both classic and modern tests of GR afford experimental confirmation of that theory in particular. In this article, we question this orthodoxy, by showing there are classes of both relativistic theories (with spatiotemporal geometrical properties different from those of GR) and non-relativistic theories (in which the lightcones of a relativistic spacetime are ‘widened’) which would also pass such tests. Thus, (a) issues of underdetermination in the context of GR loom much larger than one might have thought, and (b) given this, one has to think more carefully about what exactly such tests in fact are testing.

从规范的角度来看，广义相对论 （GR） 的“经典”测试包括近日点岁差、光围绕恒星的弯曲和引力红移;“现代”测试主要涉及相对论时间延迟、等效原理测试、引力透镜、强场引力和引力波。正统观念是，GR 的经典和现代测试都特别为该理论提供了实验证实。在本文中，我们质疑了这种正统观念，通过证明相对论理论（具有与 GR 不同的时空几何特性）和非相对论理论（其中相对论时空的光锥被“拓宽”）的类别也可以通过此类测试。因此，（a） GR 背景下的不确定问题比人们想象的要大得多，并且 （b） 鉴于此，人们必须更仔细地考虑这些测试实际上究竟在测试什么。