comment: I'm not sure how to answer this question is trying to accomplish, because it seems to assume that the distinctive use in written or oral speech were strictly proscribed and followed--which is not the case. Then it asked more broadly for the origin of the pasts, while only concerned about the false dichotomy, but going further, potentially expecting a clue from this for that.
Today, It is true that preterite is rarely used colloquially. This can have several underlying reasons. A reason not mentioned yet in other answers is that strong verbs inflect irregularly. This is difficult. The relation between ich mag and möchte is not even recognized (took me only twenty years) and they are thus treated as separate verbs. For the most part we just immitate what we hear, so we follow the past. Reasons that facilitate the prevalence of distinctive preferences also counted in the past.
In the past, as @DavidVogt points out, several developments happened, which were perhaps not semantic, but phonetically motivated. I criticized that very good answer for it did not touch upon strong verbs (admittedly, I read not the article that it has linked), so consider this: They suffered from sound change much worse than simple regular preterite forms. It's so bad, that [Grundfragen der Umlautphonemisierung] are not fully understood yet, for example. Forms like ?ich dächte, ?bräuchte (cp En ?thunk, ?brung), seem substandard. I'd say that after vowel shift, the strong verbs were perhaps not euphonic anymore. English even completely lost subjunctive inflections except in a closed class of verbs. It's hard to say which loss dragged the other ones down with them.[1] There really is no answer I am aware of, I'm affraid. Obviously we can hardly tell what was spoken, only what was written! As paper, ink and time is costly, scribes went to extreme ends to make the best use of it. Incidently, they would also be predestined to be taught in the proper forms of the grammar, and phonetics would hardly even play a role, except for speech transcripts. Obviously, the shorter preterite saves costs.
In much earlier past had be come these tenses to form. Indo-Germanic --the hypothetical last common ancestor of German and Indic, so to speak, today widely called Proto-Indo-European-- supposedly had a more powerful inflection system, as Sanskrit, Greek, etc attest to. This is mostly out of scope, so suffice to say that the origin of the participle prefix ge- in German is apparently not well understood, and might be a multifacetted problem, including tangents to geh- (gehen, confer *gana in Guus Kroonen's dictionary). The weak inflections supposedly developed later than strong inflections --which stem from Grammatischer Wechsel, Ablaut patterns in Proto-Indo-European-- perhaps in response to these loosing their charm not only on phonetic grounds, but also driven by semantic shift.
Afterthought: The Konjunktiv moods present a similar situation. This stuation is still more difficult to describe, so suffice to say that many forms are either obsolete (ich würf) or indistinct (ich schalte) and thus often replaced by an auxilary construction, i.e. in reported speech (Er sagt ich würde werfen, ich würde schalten) or irrealis (er sagt ich hätte geworfen--that en.wiktionary names plusquam perfect subjunctive II), which can get kind of cumbersome (ich hätte Geld gehabt, ich würde geworfen haben) and error prone (substandard hab ich dich getroffen [?gehabt]?). So it's become common concensus to simplify (ich würde Geld haben) when a report is neither decidedly potential nor irrealias and the tempus inappropriate.
[Grundfragen der Umlautphonemisierung]: Eine strukturelle Analyse des nordgermanischen i/j-Umlauts unter Berücksichtigung der älteren Runeninschriften. (Michael Schulte)--Review by Anatoly Liberman in Alvismal is free to read online
[1]: A similar topic might be strong inflection with Schwebelaut i/e as in ich werfe, du wirfst, wir werfen, from Middle High German, but Upper German i/e in singular/plural respectively, from uncertain origin (see Watts, West, Solms 2013, Zur Verbmorphologie germanischer Sprachen, pp. 40-45).