Reconstruction:Proto-Indo-Iranian/Áryas

Etymology
There have been many attempts to qualify the ar- verbal root of 🇨🇬 *arya- (with short -a, found in Old Persian as ariya-, and Avestan as airiia-, etc), and Old Indic ārya- (with long -a, vriddhi-formed Sanskrit ā́rya-). No Nuristani cognate has yet been identified, however. The most influential (and in the case of Pictet, notorious) of these include:

For a review of these and many other considerations, see Szemerényi 1977, pp. 103–147.
 * Before 1957 (these assume that various Sanskrit (near-)homonyms derive from a single historical unity):
 * Franz Bopp (1830): ar- "to go, to move", read as "one who roams" (like a nomad)
 * Adolphe Pictet (1858): ar- "to plough", read as "cultivator of the land"
 * Hermann Güntert (1924): ar- "to fit", read as "allied, friendly"
 * Paul Thieme (1938): ar- "to give, allot, share", read as "hospitable, friendly"
 * 1957 and later (these differentiate between the Sanskrit gentilic ā́rya- and the (near-)homonyms árya-, aryá-, aryà-, ā́rīḥa-, etc.):
 * Emmanuel Laroche (1957): ara- "to fit", read as "fitting, proper"
 * Georges Dumézil (1958): ar- "to share", read as a uniting property of "belonging to the Aryan world" ("appartenant au monde aryen")
 * Harold Walter Bailey (1959): ar- "to beget", read as "born, nurturing"
 * Émil Benveniste (1969): ar- "to fit", read as "companionable"

A derivation from Proto-Indo-European cannot be obtained with certainty either. This is because For a review of the etymological problems involved, see Szemerényi 1977, pp. 103–147; Mayrhofer 1956, p. 49, 52, 79. and Siegert 1941/1942, pp. 73-99. Watkins/IE Roots (2000) treats the Indo-Iranian autonym as an isolate and derives it tentatively from "perhaps ... ar- [to fit]" (cf., "to fit, put together"), giving "allied compatriot" or the like. This reading derives ultimately from an analysis by H. Güntert (1924, reiterated by Laroche and in numerous other later etymologies). Etymologies that consider the Indo-Iranian term to be a loanword include Oswald Szemerényi suggestion that *arya- is a loanword from an word meaning "kinsmen", but this hypothesis has today generally been discarded (EIEC: "hardly compelling").
 * the ā/a in ārya- have a morphological value unique to Indo-Iranian languages. Indo-European ā, ē, ō merge as Indo-Iranian ā (a similar merger also occurs for short vowels).
 * the rules governing ablauts are poorly understood and it is not certain whether PIE had an a-vowel at all; in principle ārya- could simply reflect zero-grade n̥ryo-.
 * the a priori assumption that ārya- is Indo-European is not assured. A comparable word does not exist in any other Indo-European language (i.e. other than the Indo-Iranian ones). 18th/19th-century assumptions of a relationship to 🇨🇬, 🇨🇬, etc. have long since been dismissed. It is possible that the autonym was originally a name given to the Indo-Iranians by another (non-Indo-European) people.
 * the relationship between various Sanskrit (near-)homonyms has not been established. In addition to the vriddhi-formed ā́rya- that corresponds to Old Iranian ariya/airiia etc., Sanskrit also has árya-, aryá-, aryà-, ā́rīḥa-, etc. Prior to the 1950s, these were all assumed to be variants of the same word (i.e. assumed to have a historical unity), but since 1957 (Laroche), this approach is no longer considered tenable. (The relationship of these terms with Sanskrit ari- "attached to, faithful, trustworthy; faithful, devoted, pious man, kinsman" is also not clear, but it has frequently been suspected as a derivational base.)

Etymologies that insist on a PIE root for Indo-Iranian *arya, and also read it as a substantive rather than a conjugation of the verbal root ar(y)-, include Mallory and Adams (EIEC, 1997) who derive *arya from, and adduce comparanda from Anatolian and Celtic, partly reiterating Laroche's analysis.

Noun

 * 1) autonymic (self-identifying) ethnonym of the respective Indo-Iranian peoples, i.e. the Indo-Aryans and the Iranians.