Over the last two decades, philosophical discussions of special relativity have repeatedly focused on whether relativistic effects—most notably length contraction and time dilation—require a dynamical explanation or a kinematical one, or whether this binary choice should be abandoned altogether. Although this debate is primarily theoretical, advocates of the dynamical approach often appeal to historical considerations. In particular, it has been claimed that, while Einstein initially described length contraction and time dilation as merely apparent coordinate effects, he ultimately aimed to show that they reflect real physical changes in the equilibrium states of moving atomic systems. This paper challenges that historical narrative and argues that clarifying Einstein’s position bears directly on contemporary disputes concerning the dynamical–kinematical distinction.

As I have argued elsewhere, in the early 1910s, the categories ‘real’ and ‘apparent,’ as well as ‘kinematical’ and ‘dynamical,’ were already under scrutiny in discussions of length contraction. I argue that an analogous and even clearer dialectic emerges in the case of time dilation. Like length contraction, time dilation is ‘apparent,’ since it is a coordinate effect that vanishes for a suitably chosen co-moving observer; yet it is ‘real,’ since it cannot be eliminated for all non-co-moving observers simultaneously, provided that the new kinematics hold. In this case, no comparable philosophical controversy arose, largely because there was no genuine Lorentzian counterpart to time dilation. Nevertheless, in contrast to length contraction, the empirical testability of time dilation appeared reasonably feasible.

As early as 1907, Einstein invoked Johannes Stark’s experiments on fast-moving ions in canal rays, which seemed to indicate a second-order (transverse) Doppler effect—i.e., a frequency shift that persists even when the classical Doppler effect vanishes. This provided a concrete opportunity to test relativistic kinematics. From the outset, Einstein rejected Stark’s dynamical interpretation, according to which moving ions undergo a ‘real’ contraction of their intrinsic frequency. Instead, he characterized the observed frequency shift as ‘apparent,’ emphasizing that the intrinsic frequency of the ions remains invariant. Nevertheless, the transverse Doppler effect is ‘real’ in that it must occur for non-co-moving observers if relativistic kinematics is correct. In principle, it thus constitutes a direct test of time dilation, provided that atomic spectral emitters function as reliable clocks.

From the 1920s onward, Einstein stressed that a complete microscopic theory of atomic clocks was still lacking. Such a theory would provide a dynamical account of why all atoms are spectrally identical in their rest frame. Yet this very assumption leads to a striking conclusion: given spectral identity and the absence of motion along the line of sight, the transverse frequency shift admits a purely kinematical explanation as a manifestation of time dilation. Paradoxically, the full dynamical program reinforces the kinematical status of time-dilation effects. In this sense, the transverse Doppler effect offers a particularly illuminating case for reassessing the long-standing debate between kinematical and dynamical interpretations of special relativity. The paper concludes that much of the misunderstanding arises from reading Einstein’s demand for a dynamical account of rods and clocks as a demand for an explanation of the new kinematics, whereas it was primarily motivated by the problem of their confirmation.