In hindsight I think it's partly the dynamics that the mix prevails, which is a wise decision for everything with dense and steady-state signals, and chimes and other high-frequency signals on the album like tambourine would cause severy distortion otherwise. But it's also partly the sonic structure of the mix. The vocals have their own space (quite dominant), kick & bass aren't that overly loud (in fact it's quite shy in the sub-bass region, except for 'Prayers for Rain' probably), (pick) bass is more 'click' than 'oomph', and the whole mix resembles the typical pink-noise curve (-3dB/oct) that was common at that time (these days a -4.5 to 5dB slope is more common). The highs are a bit rounded off, hihat is barely audible, making it quite mid-centric. Although the instrumentation and arrangement prevents the mids from becoming clogged. It was easier to achieve headroom with all this.
We didn't have streaming at that time, so we could never expect music to be equally 'loud' (except for radio), that's why we used a volume knob. But we also needed to memorize maps and phone numbers ourselves, so turning up volume probably wasn't a cognitive stunt
