¹ Substantially as presented in a paper read before the annual meeting of the Linguistic Society of America at New York, December 27th, 1935.

² In words like Ital. slargare the sl is not strictly initial.

³ See Walde-Pokorny, 2.706–16; 603–51. I have not noticed anything relevant to this topic in the available data of Hittite and Tokharian.

⁴ As Sanskrit acquired the cacuminal dentals from the Dravidian aborigines. See Meillet, Introd. ⁶ 11.

⁵ Who these Mediterranean people were and what language they spoke is unknown, and no supposition is advanced. However, we are, I think, justified in calling the stl combination ‘Mediterranean’.

⁶ Oscan slagím (Acc.) ‘regionem, finis’, slaagid (Abl.), beside Lat. stlocus, stloppus, stlat(t)a, stlis, stlembus; Oscan proper name Slabiis, beside Lat. Stlabius, Labius. Buck, Osc.-Umbr. Gram., s.vv.—Celto-Ligurian inscr. No. 269 slaniai (Prae-Ital. Dial. 2.86), Dat. ‘to Slania'?, beside Stlania (Venetia). So, generally in the coastal dialects : Stlaccia (Lucania, Calabria, Campania, Latium, Venetia); Stloga (Latium); Stlabia, Stlaboria (Campania); Stlatta (Volsci); Stlar∗∗∗ (Daunia); Stla∗∗∗∗ (Picenum); and the Illyrian town Stlupi (Ptolemy), with ethnicon Stulpini, on the coast of Liburnia. For the Italic names see Conway, Italic Dialects (1897), and Conway, Whatmough, and Johnson, Prae-Italic Dialects (1933), passim.—Umbrian Tlatie ‘Latii’ (Buck, op. cit.) does not, I think, represent a local stl, and thus separate Umbrian from Oscan; but is from an older name ∗Stlatium of the Latins themselves, i.e. ∗Slatium ‘the flat country’. Cf. Campania; stlat(t) a.

⁷ And, it seems to me, in contact with early Latin. The most archaic stl forms are in Breton.

⁸ Prov., Sp., Pg. scl beside escl; OFr. escl (Fr. écl); Ital. and Rum. sch. The statement of Brugmann (Grundr.² 1.585) that Lat. stl, ‘nach gewissen lauten’, became sl > l, does not apply to initial stl.—Although we are not concerned with the general history of -sl- in interior position, it may be observed that the tendency of the dialects of the south Italian coasts to intercalate t into sl was so strong, that when Lat. insula became ∗isla it was pronounced ∗istla, and became the name of the island Iscla, now Ischia, off the Bay of Naples. Grandgent, Vulg. Lat. §284. In Calabria, Sicily, and Sardinia the common noun appears as iska, and in Provence as iscla. Meyer-Lübke, REW ² 4475.

⁹ Although this loan-epoch seems to have been primarily European the eastern languages may also be represented. We may venture to call it the Prehistoric European Loan-period, waiving, for the moment, any participation of the eastern group that may be shown.

¹⁰ ‘Toutefois, il importe de ne jamais l'oublier, le terme de mots indo-européens recouvre deux choses hétérogènes et qui ne restent confondues que par suite de l'absence d'un critère donnant le moyen de les distinguer; et la part des emprunts préhistoriques d'un dialecte indo-européen à un autre ou de plusieurs dialectes indo-européens à des langues d'autres familles est certainement immense’. Meillet, Introd.⁶ 339, cf. 343.

¹¹ I do not discuss details of etymologies since those heretofore advanced are usually treated on an Indo-European basis, and must necessarily clash with those on the interloan theory. I have sought to use for the exposé of the principles announced illustrative forms whose semantic relations are simple and obvious. There are, however, other groups which similarly demonstrate the general principles, and which I hope to discuss hereafter.