From the Western perspective, stylistic development has largely revolved around counterpoint, with melody frequently constrained to fit harmonic priorities. Perhaps more to the point, the West has based new musical forms on shorter melodic patterns, particularly in the imitative polyphony arising from the height of the Franco-Flemish age, which so strongly conditioned later practice. While earlier medieval music, whether plainchant or Ars Subtilior songs, contains some longer or even free-wheeling melodies, the Western sense of melody has been historically proscribed by some rather rough notational bounds. Even today, separating ornamental tones from the more "structural" part of the melody in e.g. Machaut's music has taken some effort. Grace notes were not readily written, and in any case, Western theorists were preoccupied with note-against-note writing and resulting chord relations. In the case of Indian classical music, melody has been emphasized, yielding perhaps the world's most sophisticated melodic system.
The melodic complexity of today took time to develop in India. There is no reason to believe, for instance, that there was much difference in sophistication between classical modes from Greece to India, although many tentative histories do place the flow of musical ideas from East to West. There is speculation that Gregorian chant was inspired by Indian styles of scriptural singing, as exemplified by dhrupad today. Such a transmission would have been indirect, and presumably via Christian centers in Syria or farther East. What is known of Carnatic music in the fourteenth century shows some correspondence with French music in terms of melodic complexity, and it was really only during subsequent generations that priorities diverged more substantially. Staying with Carnatic music, where I am more familiar with the history (although there is little doubt that Hindustani music parallels it in many ways), the late eighteenth century was as stylistically decisive as it was in the West. Probably not coincidentally, the new world order was established during this period, British hegemony, and with it a more self-conscious distinction between what it meant to be Indian and what it meant to be European.
Whatever the influence of colonialism on Indian musical style, there is little to suggest that Indian music would have moved away from a primary melody as the focus of classical music. Not only did such a disposition have the weight of historical primacy throughout the Middle East, but raga development had already reached such a stage that melodic nuances could not be preserved under harmonic constraints. Such a blanket statement is likely unwarranted, especially as some Middle Eastern traditions have transferred microtonal melodic structures to orchestral contexts, and some current Indian composers have attempted the same. However, the degree of microtonality & phrase convolution needed to differentiate ragas is much larger than in traditions which have successfully created Western-style harmonic idioms. While a straightforward harmonic accompaniment could certainly be developed, one must immediately wonder whether it would yield any benefit to Indian classical music. The drone & other traditional forms of accompaniment have served well to highlight & enhance the principal melody. A more interesting attempt, combining multiple melodies and/or ragas simultaneously would immediately demand a level of complexity in tuning & rhythm commensurate only with the most difficult Western contemporary styles.
The issue of modulation, or more generally connecting one mode to another, has helped to forge distinctions between styles of classical performance. Modulation has dictated the rather indistinct tuning associated with mainstream Western music, and Indian music has steadfastly refused to go in that direction. However, music combining multiple ragas is far from uncommon, and more technical processes such as shifting the tonic to a note of the previous raga have been devised. In any case, attention must be given to relationships with the drone, although the advent of electronic drone boxes make a push-button switch to another drone relationship easy in principle. The musical merit of modulating while changing the drone seems, at best, questionable, though. Moreover, let us recall that the "tonic" in Indian music does more to define the frequency ratios of the raga than it does to mark cadential motion. There is no special priority for Sa at the end of phrases in ragas in general, so that multiple ragas can fit into the same acoustic space. The idea of "dissonance" — or tension & release — when moving to more distant relations to the drone & back again could certainly be applied, but would require even more attention to intonation. Ideas on "verticality" and acoustic space are already played out against a drone and melodic registers in these and other ways.
To many Western ears, the principal characteristic of Indian music, or most any world tradition, is rhythm. For one, many musicians raised on equal temperament claim that it is physically impossible to hear some of the intonational precision required for raga separation. Of course, such claims are both arrogant & self-limiting. On the other hand, Indian rhythm is indeed sophisticated, and the basis for the sort of sequential exposition I have described as "non-simultaneous polyphony" (a description which is surely a personal quirk of my own orientation). Again, for rhythmic displacement of this sort, Western classical music must look back to the late medieval Ars Subtilior. It is rhythmic sophistication which lets raga grammar become as complicated as it has, as it is clearly defined temporal positioning which serves to frame ornament and define articulation. The idea of ornament as articulation can almost be broached, and such effects certainly characterize individuality within broader styles. It is ornament & improvisation which ensure that straightforward melodies will not seem boring, and that dynamic may give a hint toward a more authentic artistic context for Western plainchant. (Note, however, that the relationship between virtuosity and tala position is rather different in Hindustani music, and the khayal practice of increased animation before the beat does not yield for me a similar "polyphonic" impression.)
Beyond religion, improvisation as it connects song to raga is one of the most important expressions of absolute in Indian music. Involved, melismatic melodies have been able to articulate the absolute in their small inflections & displacements, as a clear primacy of line has left them unconstrained by any push toward blocked units & simpler phrases. Consequently, purely melodic richness continues to develop not only via practical elaboration, but via theory as well. New ragas have been created in systematic fashion, giving a dynamic context for melodic exploration which is as rich in its own way as the harmonic experiments of Western music. There are "dissonant" ragas and corresponding controversies of style, and ultimately a pallet of musical possibilities which is far from exhausted. While arbitrary melodic construction could be undertaken via any number of algorithms, the ability of Indian musicians to find a real context for expression in new ragas serves to invest them with a melodic significance. This cannot happen overnight. Once invested, however, a melody is a very human element which can & will retain its significance — not to mention a critical ability to shift contexts. It can take on a life & relevance of its own. The melodic basis of Indian music has consequently given it a very firm footing heading into the twenty-first century, with little crisis of conscience.