In the phonology of a series of languages, /r/ exists as a trill, and is reduced into a flap in informal speeches or in a syllable-final position. Why is it happened to be a flap, not a fricative?
I tought it could be a fricative appearing as the reduced version of /r/. Besides, there should be a plenty of examples of sound change from [r] to [ʐ] diachronically. Below is my assumption.
The trill [r] does not exist in my mother tongue's phonology. In order to pronounce the sound, I need to send airflow from the lung with extra force, so that the force can "blow" my tongue to be vibrating, otherwise the sound will result in a fricative. Another thing I just tried is to pronounce an [r], and then reduce the airflow level from my lung, then the continuous pronunciation of the trill is over, replaced with a continuous pronunciation of a fricative. This shows, with other factors remaining unchanged, a weaker airflow would change the trill into a fricative.
Meanwhile, informal speeches and syllable-final positions are both characterized by a weaker airflow and lazier pronunciation. It seems thus plausible that /r/ would be realized as [ʐ] or some other fricatives.
As for the plausibility of [ɾ] being a realization of /r/, at least for me, the mechanisms of pronouncing [r] and [ɾ] are two matters. The former is about to blow the tongue with the airflow from the lung, the latter is just about to raise the tongue to hit the palatal quickly. Put in another word, I would never mix up the trill and the flap if they appear together in a tongue twist.
Based on the above mentioned, I would say that 1. Reduced pronunciation from [r] to a fricative would be normal. 2. Diachronic sound change from [r] to a fricative would be normal.
So, my question are, is my assumption reasonable? to what extent are these two correct? and does the realization pattern of different allophones of a trill-phoneme vary by native and non-native "prouncer" of /r/?