It's a matter of where the distortion is. Transducers distort much more than other hi-fi components. There is some logic in source-first especially for record players because the record player is also relatively high distortion. Amplifiers are low distortion as are digital sources making the loudspeakers relatively much more important. There is also the question of rooms and room effects that are discussed I would say too little by hi-fi enthusiasts as that too makes a big difference to the sound.
Distortion, frequency response accuracy, dynamic range, noise floor, impulse response accuracy, phase accuracy ..... All are measurably worse for a turntable front end, and all are as near as matters to perfect* in digital front ends and modern solid state amplifiers.
Turntables/arm/cartridge combinations of even the best achievable had several of those performance characteristics that were well within the domain of audibility, thus turntable front end "quality" mattered, and it mattered a lot! The gulf just in frequency response accuracy alone between a bottom of the range cheap all in one turntable and a high end turntable combination was huge and easily audible to anyone. The differences were of the same order as can be heard between speakers. So the source first mantra made perfect sense because there was a massive difference in the amount of information retrieval possible between a low end and high end turntable.
That is just simply not true for digital sources, where digital sources that are unable to retrieve 10, 20, 30% or more of the information stored in the format just don't exist. Similarly such amps don't exist. Speakers however, despite being far less variable than they were 40 years ago, still have the potential to do exactly that.
On top of that speakers are (just like turntables) physical engineering devices, and that is what cost the most money to develop and manufacture at a high level. So absolutely in todays world, if you're using digital sources the majority of your money (or at least a good proportion of it) is best spent on speakers, followed by the amplification (especially if it's traditional class A or A/B stuff), and finally the digital source. IF your goal is accuracy of retrieval of the information in the source format. IF you have different aims for your system, do what sounds best to your ears.
*Perfect with respect to human ability to detect any of the above listed characteristics of audio reproduction.
?
