ārš-^(ti)

The root underlying PIE *h₁érs‑/*h₁r̥s‑´ ‘to flow’ is said to be found in different formations, none of which, however, is beyond any doubts.What is cited by IEW:337 as “toch. A yär‑s‑, B yar‑s- ‘baden’” are probably ghost words (not found in Hackstein 1995a, Malzahn 2010a, DTB²). Pokorny’s source for this word is not clear, since it is not found in WP I:149–151, WH I:16f., and Trautmann 1923a:237f., either, which are the works cited at the end of his lemma. According to Couvreur 1937a:97, Toch. A, B yärs- ‘to show respect, affection’ (Malzahn 2010a:801f.) is a causative ske/o‑formation of Toch. A yär- ‘to bathe (intr.); to purge’ (cf. Pal. arr-(i) ‘to wash’) and thus might be translated as ‘to bathe (tr.)’, but cf. DTB²:540. 
One group of possible cognates is represented by Lat. rōs, rōris ‘dew’, etc.Cf., e.g., EDLIL:527: “The resemblance in form and meaning is too great to reject this etymology.”However, the Italic and Balto-Slavic evidence (with the meaning ‘dew, drizzle’) does not point to flowing but rather to drizzling or dripping water, if the semantic source of ‘dew, drizzle’ is moving water at all, cf. also below on the phonological problems.
Another assumed and already PIE cognate is the stem *h₁r̥s‑én‑/*h₁r̥s‑n‑´ ‘male’, which is oftentimes connected with PIE *h₁ers- ‘to flow’, cf. the refs. in NIL:585 n. 1. However, it should be noted that there seem to be no reflexes of an assumed meaning ‘flowing’ (> ‘ejaculating’ > ‘male’) vel sim., and independent parallels of a development ‘flowing’ (> ‘ejaculating’?) > ‘male’ are lacking, too.Cf. the equally challenging parallels PIE *Hu̯ŕ̥s‑en- ‘male’ (NIL:722–724), if indeed connected with PIE *h₁u̯ers- ‘to rain’ (LIV²:291f., LIV²⁺ s.v. *h₂u̯ers‑). Cf. Kümmel 2000a:476f. on the possible connection with PIE 2.*u̯ers- ‘to rise’ (LIV²:691). It might be possible to connect (at least) the word for the Western capercaillie with the root PIE 2.*u̯ers- ‘to rise’ (LIV²:691), based on a certain behaviour during courting. Furthermore, Neri in EWAhd VI:1208 suggests PIE *h₂u̯r̥h₁‑én- (> Gk. ἀρήν ‘lamb’) ← PIE *h₂u̯erh_2/3- ‘to splash (intr.)’.Neither the CLICS³ database nor Buck 1949a:84f., 139f. provide comparable developments, which leads to the conculsion that, at least for now, the etymological affiliation of PIE *h₁r̥s‑én‑/*h₁r̥s‑n‑´ ‘male’ with PIE *h₁ers- ‘to flow’ cannot be maintained.Due to the sparse evidence for the root underlying Ved. arṣ- ‘to push, thrust’ (LIV²:224, EWAia I:123f., KEWA I:125), the proposal to explain PIE *h₁r̥s‑én‑/*h₁r̥s‑n‑´ ‘male’ as ‘pushing, thrusting’ is only hypothetical.
If, however, the third group of lexemes containing Lat. errāre ‘to roam, go astray’, etc. is connected (as per EIEC:206f.), the original root meaning should be reconstructed as ‘to move without following a certain direction’Cf., e.g., Klingenschmitt 1982a:96, GED:20, and also EDPG:119.or ‘to move quickly’.Cf., e.g., Grassmann 1996a:119 on Vedic.This may be a plausible point of departure for the meaning ‘to flow’, on the one hand (cf. Ved. ARṢ ‘to run, flow’ with both meanings), and for denoting a male (PIE *h₁r̥s‑én‑) in coital or courting context. However, since this solution implies that the meaning ‘to flow’ in IIr. and Anat. is secondary and the further evidence would have to be checked against this proposal, the conclusion that PIE *h₁ers- could have meant ‘to move hastily’ vel sim. has yet to be regarded as tentative.Furthermore, IEW:336 suggests to connect the root in question with PIE *h₁er- ‘to attain, reach’.

The root underlying the PIE root present is reconstructable as PIE *h₁ers- ‘to flow’.On further considerations regarding its semantics, cf. above sub Semantic Reconstruction.However, it is disputable, which of the lexemes adduced sub Transmission above (cf. even more in, e.g., IEW:336f. and WH I:416f.) belong to the root in question.
Lat. rōs ‘dew’, etc. (< PIE *h₁rós‑/*h₁rés-), beside the semantic mismatch (‘to flow’ vs. ‘to drizzle’, cf. above), can only be compared if a motivation for the Schwebeablaut is found (cf. NIL:574 n. 1), which is, however, not the case.
The stem PIE *h₁r̥s‑én‑/*h₁r̥s‑n‑´ ‘male’ can only be connected if the semantic development is explained. Since parallels for ‘to flow’ → ‘male’ are lacking, the only plausible possibility is a different original root meaning (not ‘to flow’ but ‘to move hastily’ or ‘to thrust’), cf. above. It must be refrained from this connection for the time being, because the alternative root semantics had to be established before.
The above-mentioned uncertainties regarding the possible cognates, while important in terms of the original root meaning, do not influence the phonological reconstruction of the root as PIE *h₁ers- ‘to flow’ (LIV²:241).

[A.O.]