Between The Hobbit and Indo-European linguistics

The Dwarf Name Bǫfurr

In the Old Icelandic text known as the Vǫluspá (the vision of the seeress) we find a þyla[1] known as the Dvergatal (Vǫluspá 9-17). This list gives the names of mythical dwarfs, many of which were used by Tolkien in his The Hobbit (1937). One of them is named for example Gandálfr “conjuring elf.” In general however the names of the dwarfs are quite diverse. Some refer to natural phenomena such as the winds and the wind directions (cf. Austri, Vestri) and may hint to the original function of these supernatural beings as nature spirits. Other names refer to attributes of the dwarfs in questions. For example, the name Alþjofr may refer to the dwarfs capacity as a thief or raider, a trait praised in Old Germanic society judging from the popularity of the onomastic element *þeuƀa- “thief” (Kershaw 2000: 56). Among these names we also find the name bǫfurr which does not have an etymological explanation.

I want to propose we connect this name with the Proto-Indo-European root *gwhobhmeaning “smith” which we find in Old Irish gobae, gobenn “smith” < PCelt. *gobasn, *gobasnos and Latin faber  “smith” < PIE *gwhobhro-. The Latin a for older *o might be explained from its phonetic environment, cf. PIE *ghwok(w)> Lat. fax “torch” (see Blažek 2008: 74). The connection of the Latin word faber “smith” to Armenian darbin “id.” is not compelling since the Armenian word might also be derived from PIE *dhrbh-ino- (PIE *d(h)erbh– “to prepare, to work” cf. Lith. dìrbu, dìrbti “id.”, see Blažek l.c.).

Old Icelandic bǫfurr is a normalized spelling popularized by 20th century editors for the actual form bavör in the Codex Regius and bafur in the Hauksbók (see Bugge 1867: 2). Therefore we cannot establish the operation of u-umlaut which would have secured the Proto-Norse reconstruction *bafur. Nevertheless, an interpretation báförr which is often done to justify an etymology “trembler” (cf. OIc. bifa “to tremble”) is also not compelling, since the Codex Regius and the Hauksbók generally use both <ö> and <u> for the unaccented u in suffixes (Syrett 1994). Both manuscripts also vacillate in the spelling of the u-umlauted a which might be represented by both <a> <o> and <ö> (Harðarson 2001: 83). This means that the normalization Bǫfur is still legitimate, although not exclusively so. Assuming Bǫfurr is the underlying form reflected in the manuscript spellings, we may depart from Proto-Norse *bafur.

Seebold (1967) has established that initial PIE *gwh– could yield *b- in Germanic, e.g. *gwhonōn  (cf. PIE *gwheni̯o– “to kill, to hit” > Gk. θείνω “id.”) > PGmc. *banōn “murderer” (cf. OHG bano, OE bana “id.”), PIE *gwhedi̯ono- > PGmc. *bidjan- (cf. OHG bitten, OE biddan).  Furthermore it is relevant to know that PIE *-bh– > PGmc. *-ƀ- yielded *-f- in Proto-Norse, as in PGmc. *weƀan- “to weave” > OIc. wefa “id.” Proto-Norse *bafur would then reflect Proto-Germanic *baƀur-. This PGmc. *baƀur might very well have evolved out of PIE *gwhobh“smith”. The thematization to a-stem would then have occurred later in the prehistory of Norse.

In Germanic mythology the dwarfs are the smiths par excellence. If a mythological weapon was deemed exceptionally good, it was surely made by the dwarfs (West 2007: 296). The most famous of these mythological smiths was Weland (cf. OIc. vǫlund) whom we know from the Anglo-Saxon Deor poem and the Old Icelandic Vǫlundarkviða. We also encounter Weland in some remarks in Anglo-Latin texts and in the continental Waltharius poem.  It stands to reason that Weland, who is known in Norse mythology as the prince of the elves (i.e. Old Icelandic vísi álfa) was originally also a dwarf (West 2007: 296-97). In Grimm’s Deutsche Sagen (1818, tale 29-44) we also encounter the dwarfs in their capacity of artisans of extraordinary quality. Many encounters with dwarfs in 18th century German folk tales ended with the protagonist acquiring a dwarvish artifact of exceptional beauty. Since the dwarfs were known for their metallurgical skills in Germanic mythology, it would only be logical if one of the dwarfs would be named with an extremely archaic word for smith, namely PIE *gwhobhr-.

Peter Alexander Kerkhof


Blažek, Václav

2008        “Celtic “smith” and his colleagues,” in: Evidence and Counter-Evidence; Essays in honour of Frederik Kortlandt, Volume 1:Balto-Slavic and Indo-European Linguistics, Alexander Lubotsky, Jos Schaeken, Jeroen Wiedenhof eds., Rodopi, New York, Amsterdam, 67-86.

Bugge, Sophus

1867        Sæmundar Edda hins Froða, Norrœn Fornkvæði, Christiana.

Grimm, Jacob et Wilhelm Grimm

1818        Deutsche Sagen, revised edition 1865, Kassel.

Harðarson, Jón Axel

2001        Das Präteritum der schwachen Verba auf -ýia im Altisländischen und verwandte Probleme der altnordischen und germanischen Sprachwissenschaft, Innsbrucker Beiträge zur Sprachwissenschaft, Innsbruck.

Kershaw, Kris

2000      The one-eyed god; Odin and the (Indo-)Germanic Männerbunde, Journal of Indo-European Studies Monograph36 (Washington).

Seebold, E.

1967        “Die Vertretung idg. guh im Germanischen” Zeitschrift für vergleichende Sprachforschung auf dem Gebiete der indogermanische Sprachen, Vandenhoeck & Ruprecht, Göttingen.

Syrett, Martin

2012        The unaccented vowels of Proto-Norse, NOWELE supplement series 11, John Benjamins, Amsterdam, Philadelphia.

Tolkien, J.R.R.

1937        The Hobbit or There and Back Again, George Allen & Unwin.

West,  M. L.

2007  Indo-European Poetry and Myth, Oxford University Press, Oxford.

[1] Special thanks to Seán Vrieland (University of Copenhagen) for pointing me to the spelling variants in the manuscripts. A þyla is a list of names containing mythological information

One response to “Between The Hobbit and Indo-European linguistics

  1. Thank you for an interesting and reasonable etymology. Coincidentally, I wrote my own piece on Tolkien’s dwarf names four days later, repeating the ‘trembler’ explanation, not knowing about yours.

    The Etymologisch Woordenboek van het Nederlands is not convinced by Seebold’s arguments for PIE *gwh- yielding PGmc *b-, as evidenced by their entry voor bidden, yet unfortunately they don’t state their reasons. Similarly they don’t even mention the possibility of PIE *ǵhw- yielding PGmc *b-, which would link English bear to Greek θήρ ‘wild beast, beast of prey’ and Latin ferus ‘wild’, something which Ringe (2006) is certainly willing to accept.

    (Note: I hope my formatting comes out right; I cannot preview my message.)

Leave a Reply

Your email address will not be published. Required fields are marked *