ONlwSG

arspag f. [ˈɑʂpaɡ̊] [ˈɑrs̪paɡ̊], (Lewis) [ˈaṣbag] (Oftedal 1972, 120), (Uig, Lewis) [ɑʴsbɑɡ], (Sutherland) [ɑʴspɑɡ] (Faclan bhon t-Sluagh^˄, which also indicates use in Scalpay), gen. arspaig [ˈɑʂpæɡ̊^j], [ˈɑʂpɛɡ̊^j] etc. ‘great black-backed gull, Larus marinus’, is derived from ON svart-bak, acc. of svart-bakr m. ‘idem’ (Oftedal 1972, 120; 1980, 172: arspag or arsbag 

There is no phonemic distinction between sb and sp in Scottish Gaelic orthography, although sp is recommended for most words.

; Cox 1991, 493; 1992, 138: [ɑʂpak]; 2022, 108, 112, 136, 186, 196, 210, 236, 280, 906) 

Although Christiansen (1938, 15) is doubtful.

, although the nominative is sometimes cited (Lockwood 1961, 45; Oftedal 1962a, 120; Stewart 2004, 408; Ó Muirithe 2010), or even the Icelandic nominative svart-bakur (Henderson 1910, 122–23; McDonald 2009). Several reflexes of ON svart-bak are found in Scottish Gaelic:

1. arspag (as above), and

2. farspag (MacBain 1896; Mackenzie 1905, 77: (St Kilda) farspock; Henderson 1910; McDonald 1972; Cunningham 1990, 126; Wentworth 2003, s.v. gull: (Gairloch) [fɑʂpɑk], although arspag [ɑʂpɑk] is occasionally found; Faclan bhon t-Sluagh^˄: South Uist or Skye), with prothetic f-, cf. falc < alc, q.v.

For the development from ON svart-bak, Henderson (1910, 123) suggests the intermediate forms *sfarst-bak, *sfars-pac, *farspac, farspag, arspag. This assumes that ON s + semivowel yields EG sf, an initial cluster that does not otherwise appear to occur in Early Gaelic; indeed, unlenited Primitive Gaelic sw- developed into OG s-, e.g. OG (dissyllabic) sïur f. ‘sister’, Goth. swistar (Thurneysen 1975, 84–85), while lenited Primitive Gaelic sw- developed into OG f-, e.g. OG sïur > (lenited) fïur, hence SG (dissyllabic) piuthar [ˈpʰju-əɾ] > (lenited) phiuthar [ˈfju-əɾ], via back-formation, as /f/ is the lenited form of initial /p/. It may be that the Old Norse initial was simply lost in some areas, but rendered /f/ via back-formation in others, while in yet others f- is prothetic.

Although it may be that a Gaelic reflex *sarspag survives in a lenited form in the Lewis place-name Rubha Sharspag -[ˈhɑʂpaɡ̊] ‘the promontory of (the) black-backed gulls’ NB207352, with lenited genitive plural of the specific (Cox 2002a, 358; 2022, 905) – though note (nMO) -[ˈɑʂpɑɈ].

The second element in the original Old Norse cluster rt evidently survived long enough in Gaelic for the development /rt/ > /rsd/ to take place in at least some areas, which was in turn reduced to /rs/, before being borrowed perhaps into other areas (that might otherwise have retained original /rt/) as /rs/.

For the development of rt, and of rtl and rtn, see Robertson 1908a, 345–47.

ON final ‑ak would normally be expected to yield SG -[əɡ̊] but instead has been remodelled on Gaelic bird names in final ‑ag ‑[aɡ̊] (EG ‑óc), e.g. curracag ‘lapwing’ (< currac f. ‘cap’).

3. arspach (Glengarry) [arsˈbɐK], gen. arspaich (Dieckhoff 1932), and

4. farspach (MacBain 1896; Faclan bhon t-Sluagh^˄: 

Written both farspach and farsbach.

Harris, Scalpay, Tiree, Skye).

Reflexes (3–4) have been formed on the alternative Gaelic suffix -ach (presumably the adjectival suffix used nominally), again perhaps modelled on other bird names, e.g. fachach, q.v.

5. farpach (Defayes (1927-^˄, s.v. Laurus marinus) seems to list both farspag and farbach (presumably for farpach 

As /rp/ and /rb/ are phonemically distinct from each other, the latter producing an epenthetic vowel.

) for St Kilda.).

Assuming farpach is not in error for farspach, we can compare the St Kilda place-names Creagan Fharspeig 

Formally, Creagan Farspaig with unlenited gen. sg. or Creagan Fharspaig with lenited gen. pl. of farpag would be expected.

and Geo Creag an Arpaid 

Perhaps for *Geò Creagan Arpaid, if arpaid is feminine like farspag.

cited by Coates (1988; and 1990, 69 and 90), who sees St Kilda dialect behind palatal /d’/ for /ɡ’/ and the lack of a developed sibilant in arpaid (cf. SGDS, Item 172: ceart: Points 14–15 [k’ar’t], Point 16 [k’æʀ̥t]); if this is correct, farpach might be similarly explained.

6. fairspreig (Forbes 1905, 28: Argyll), and

7. *farsaibreac (Faclan bhon t-Sluagh^˄, s.v. farsabreac: (Skye) [fɑʴsəbɾɛxk]).

Reflex (6) seems to have acquired an epenthetic r, cf. briosgaid < Eng. biscuit (Calder 1972, 70) and may be a normalised dative form (hence the final palatal velar), while (7) seems to be based on a folk etymology that the word contains final -breac ‘speckled, variegated’ etc., shown by the irregular pre-aspiration (here /x/) before a final stop in a polysyllabic word.

8. While Forbes (1905, 28) cites fairspreig for Argyll under farspach, he separately lists fairsleag ‘the large gull’; this may represent an alternative strategy to that under (7) and be a remodelling based on SG faoileag f. ‘gull’, a common variant of faoileann (EG faílenn).

9. Reference is made to supposedly related forms in Irish: fáspág (McDonald 2009; but corrected to faspróg in 2015a, 112) and fárspróg (Ó Muirithe 2010, 5, 48), but it seems likely that these are miscopyings for Ir. fáspróg (Dinneen 1947; so Henderson 1910, 122), leg. faspróg, a variant form of ospróg f., a word based on Eng. osprey but with the Ir. suffix -óg, with prothetic f- (Akerbeltz: Prosthetic f, backformation or eagal and feagal^˄).