Historically, melody existed long before harmony. Before the modern theories of chords were developed, people were writing music that we can analyze harmonically today without any harmonic theory: they analyzed and probably composed this music by concentrating on the intervals between each voice and the melody. That is, you'd start with a melody and write other voices to go with it, sometimes one other voice, frequently two or three, and sometimes even more.
With large numbers of voices especially, it became increasingly common during the renaissance for the lowest voice to move by leaps of a fourth or fifth. By the baroque period, organists and players of other chordal instruments were playing from parts that included only the bass part (originally it included whatever part was lowest at any given time, hence basso continuo -- "continuous bass"). Perhaps because of this, the study of composition began to include writing harmonic accompaniments to a bass line without a melody.
The need to realize a chordal accompaniment naturally led to a theory of chords with which a musician could decide what to play, and that in turn influenced the theory of music composition. In the middle of the baroque period, Rameau published his theory analyzing chords as stacked thirds and identifying for example a chord consisting of E-G-C as an inversion of the chord C-E-G.
When you apply this theory to music that existed at the time you find that the implied bass note -- the root in the modern sense -- moves even more frequently by fourth or fifth. For example, a bass line C-D-E-F-G-C could be harmonized (in modern notation) C G7/D C/E Dm/F G7 C -- one move by step and the others by fourth/fifth.
So, historically, the answer to your question "do the roots actually serve as the basis for what is being done melodically" is no or at least not directly -- a melody that might go with that progression is E-G-E-D-D-C, with only two of six being the same as the bass note and only one being the root in the modern sense.
But in fact it's not clear to me what you mean by "the note that stands out is always the root" -- maybe you are noticing a certain effect when the melody note has the root that is absent when it has the third or fifth. Or maybe you're noticing a tendency for the melody to be harmonized with chords built on the melody note at certain points in the melody's course.
To elaborate: melodies typically do have a certain "pull" toward the tonic note, and there is a sense of rest or arrival when they are on that note, and this is typically harmonized with the tonic chord. So in the sense of trying to understand tonality, you may be heading in the right direction. Chords and melody tones that appear in the middle of the melodic phrase are less likely to be the tonic chord and the melody is less likely to be on the root tone of the chord -- it's more likely to be the third or fifth or seventh or even something more dissonant.
(A broken chord that isn't in ascending or descending order is unlikely to be described as an arpeggio.)