I tend to agree (surprise! :-) But some scholars do attempt this by extrapolating from known languages to hypothetical common ancestors, which is a somewhat speculative enterprise IMO. However, even if we don't employ that method of classification, we still have to call these groups something if we're going to talk about them, in which case if we're going to use the name "Scythian" or "Celt" we need to agree upon who we're talking about, which is why I prefer to stick with Herodotus' terminology. Incidentally, though, the Scythians and Celts did leave written artifacts:
THE PRELIMINARY REPORT ON THE DECIPHERMENT OF THE PROTO-SLAVONIC WRITING SYSTEM
The Hungarian Rune Writing (During those times Sekler runic was referred to as the "Scythian letters", because the Seklers believed themselves to be the successors of the Scythians.)