While I tend to agree with your hypothesis in general, I'll point out that when I was coming up (in the 70s-80s), there were damn few pieces you COULD chain as they are today without degradation. Yes, Sontec's have been around forever, and are very quiet, but how many LA-series compressors would YOU willingly put in line before squealing "uncle!"? For me that number was 2. And that was at the high end!
My point is that every single stage of the process, including those you cite, Brad (pres, consoles, effects), was much noisier than contemporary equivalents. Many modern Pultec clones are cleaner than the originals. Further, most balancing in "classic" gear was done with transformers, and most manufacturers weren't Neve, so the signal got duller and less defined at every interface. How many studios had speakers as good as todays B&W or Lipinskis? Amps as quiet as a Pass or Cello to hear what was going on. Conservatism in processing wasn't a choice so much as a necessity. In short, a modern workflow would rapidly fall apart in a vintage setting. Albums like Sandinista! and Scary Monsters pushed those limits, and fairly creak under the processing of the day.
Furthermore, most studios only HAD one or two "money" pieces in the day. We used our LA2's and 4's on nearly EVERYTHING at the tracking level when overdubbing. We used our best EQs when needed, but mostly got it right in tracking because there was no automation like we have today. We simply couldn't kick every can down the road for mixdown. This forced a degree of focus that ultimately helped the records. You had to be in the zone when you hit record. Everything we did was lossy and destructive to a much greater degree than is the case with 24 bit PCM, so you had to think ahead and plan for those losses at the outset. I would argue that focus and mindset accounts for much more of the difference than "over processing" ever could. I see this in classes I teach as well as records I master and projects I consult on.
If I had the kind of automation we take for granted then, and an unlimited processor budget, I might well have pushed this envelope long before plug-ins let me put my best stuff on every track at will. But we didn't. I still apply the focus I mentioned, and try to move things as if every step mattered, just as before. Only now I can undo. I am certain my current work sounds better, not worse, as a result. And while I've gained a lot of knowledge and experience, if I had a time machine I STILL wouldn't over-process past sessions with "classic" gear and tape formats. The signal and its path were simply too fragile to support this.
Finally, I often feel that ProTools TDM is every bit as fragile as those older platforms, and it's dominance has more to do with contemporary sound than the plugs and processors people use inside and outside of it. A string of 24 bit TDM plugs is a just a different wringer, trading quantization and filter distortion for tape hiss and low dynamic range. I know PT users will disagree, but this is an easily measured effect (string some plugs and do simple gain changes, watch the s/n plunge!). In that sense I'm completely with you, Brad. The ability to over-process leads folks to do it, and the dominant tools are no more friendly to that approach than old-time analog was. Still there are cleaner platforms and plugs that are amenable to this (Sonar, Sequioa, Sonic all pass data between stages at much higher resolution, and Algorithmix and other plugs can maintain the precision through the chain), so it's no so much a question of method, but means.\
Playin' devil's advocate today...
-d-