The use of separates came about in the late 1940s, blossomed in the 50s and 60s, then exploded in the 70s. Stereo lovers had contracted mix-and-match fever. Before separates, of course, stereo manufacturing companies produced integrated products that just worked: no interconnects, multiple AC plugs, or collection of boxes common to separates.
We moved away from all-in-one systems because of one essential idea. An individually powered, shielded, and purpose-designed component has a better chance at perfection than an integrated. And that is true save for one detail, the need for interconnection.
The tying together of individual components from disparate companies is problematic. There are few standards in consumer electronics and we hope our chosen products play nice with each other.
The unofficial connection agreement between consumer equipment is what I like to think of as the low/high dictum. The output of an electronic component should be low impedance and its input should be high. Thus, when we connect two pieces of kit together low feeds high. While imperfect from the viewpoint of interconnecting cables happier with matched impedances—a practice common in pro gear—following the low/high standard helps everything connect without losses.
If we are really vigilant we will always go balanced. True balanced cables and equipment take advantage of something called common mode rejection which lowers distortion and reduces the sonic impacts of impedance mismatches.
If you’re curious enough to want more information on the interconnection of separates, check out my video here.