The classical emission (ballistic) theory of light predicted that the speed of light reflected from a moving mirror is c + 2v, where v is a component of the mirror velocity. In 1913, A. Michelson carried out an experiment to test this hypothesis and concluded that the speed of light reflected from a moving mirror is constant c independent of mirror velocity, to a high degree of precision. With the advent of Albert Einstein’s special relativity theory, and with additional experimental counter evidences such as moving source experiments, the classical emission theory was finally abandoned. Many years later, in 1967, an experiment was being carried out to test Einstein’s gravitational time dilation by bouncing radar pulses grazing the sun off the planet Venus. As analyzed and disclosed by Bryan G Wallace, large ‘anomalous’ first order variations in the round-trip time were found in the raw data, in complete disagreement with Einstein’s light postulate, but conforming to the long forgotten classical emission/ ballistic theory. In this paper, I present a new model of the speed of light reflected from a moving mirror that resolves these contradictions. Although the model can make correct predictions, its physical meaning is inexplicable. Light behaves as if it is reflected from the point in space where the mirror is/was at the instant of emission, and the speed of the reflected light is the sum of the speed of light c and twice a component of the mirror velocity, i.e. c + 2v. Logically, one would have to take into account the motion of the mirror during the transit time of light to determine the point in space where light is reflected. This paper shows that this logical and conventional thinking is possibly wrong in the case of light.