Philological and linguistic evidence indicate that the Rigveda was composed in the north-western region of the Indian subcontinent, roughly between 1700–1100 BC (the early Vedic period).
The surviving form of the Rigveda is based on an early Iron Age (c. 10th c. BC) collection that established the core ‘family books’ (mandalas 2–7, ordered by author, deity and meter ) and a later redaction, co-eval with the redaction of the other Vedas, dating several centuries after the hymns were composed. This redaction also included some additions (contradicting the strict ordering scheme) and orthoepic changes to the Vedic Sanskrit such as the regularization of sandhi (termed orthoepische Diaskeuase by Oldenberg, 1888).
As with the other Vedas, the redacted text has been handed down in several versions, most importantly the Padapatha that has each word isolated in pausa form and is used for just one way of memorization; and the Samhitapatha that combines words according to the rules of sandhi (the process being described in the Pratisakhya) and is the memorized text used for recitation.
In order to achieve this the oral tradition prescribed very structured enunciation, involving breaking down the Sanskrit compounds into stems and inflections, as well as certain permutations. This interplay with sounds gave rise to a scholarly tradition of morphology and phonetics.
The Rigveda was probably not written down until the Gupta period (4th to 6th century AD), by which time the Brahmi script had become widespread (the oldest surviving manuscripts date to the Late Middle Ages). The oral tradition still continued into recent times.
The original text (as authored by the Rishis) is close to but not identical to the extant Samhitapatha, but metrical and other observations allow to reconstruct (in part at least) the original text from the extant one, as printed in the Harvard Oriental Series, vol. 50 (1994).