By Peter Alexander Kerkhof
Etymologies live long lives, whether they turn out to be false or not. In Old Germanic studies it is common practice to cite earlier etymological suggestions going back as far as the nineteenth century. In most cases where Old Germanic etymology is involved, this is not a big problem because a lot of the earlier scholarship has been replaced by better more up-to-date suggestions. Not so much in the field of Dutch toponymy where scholarship has largely lain dormant for at least 20 years and almost every publication since World War II has been notoriously conservative in etymologizing place-name elements.
This article is about the Old Dutch word harag which is known from its occurrence in Early Medieval place-names, such as Haragon (1083 CE), the oldest form for the locality of Hargen in the province of North Holland. The Old Dutch word harg is related to the Old English word hearh (also found in Modern English place-names as harrow), Old High German harug and Old Norse hǫrgr. In Early Medieval English and German the vernacular word translates Latin terminology relating to the pre-Christian religion, namely a holy grove, a cultic building and even the idols themselves. In the Old Scandinavian languages however, the cognate word hǫrgr is found in both the meaning ‘heathen temple’ and the meaning ‘rocky ground, mountain or cliff side’. In the 1960s, Scandinavists such as Rostvik preferred to derive the meaning ‘heathen temple’ from an earlier meaning ‘mountain or cliff side’, assuming that Old Norse preserved the older semantic range that West Germanic must have lost. The Germanic ancestral form *hargu- could then be linked to words for rocks in the Celtic languages such as Old Irish carrag and Old Welsh creic. This reasoning was quickly accepted in the wider field of Old Germanic studies, since post-war scholarship had grown weary of the fascination of older generations of scholars with Germanic pre-Christian religion.
Rostvik tried to fortify this etymology by suggesting that Old Norse hǫrgr could by way of Verner’s Law be connected to Middle Dutch and Middle Low German haar, a toponymical element interpreted as ‘hill, ridge’. This connection needs a Proto-Germanic variant *harhu– to have existed alongside *hargu– which is the ancestor of Old Norse hǫrgr and Old Dutch harag. The existence of the Verner variant has been defended as recently as 2013 by Guus Kroonen in his Etymological Dictionary of Proto-Germanic (Kroonen 2013: 211), but it does not hold up to scrutiny. Middle Dutch and Middle Low German haar is continued further south into High German territory where the reflex is also haar and not the expected **harch (compare German durch against Dutch door), indicating that we are dealing with a different etymon altogether.
As a consequence of Rostvik’s argument for a primary meaning of ‘rocky area, mountain, cliff side’, Dutch toponymists have doubted whether the place-name Hargen can be interpreted as a sacral toponym that signals in its name the former existence of a pre-Christian sanctuary. Quak (2002) and most recently Berkel en Samplonius (2006) are all skeptical in this regard. It is possible, they argue, that Old Dutch harag referred to the soil as being ‘rocky’ or the landscape being ‘hilly’, although they admit that in reference to Dutch geography that would not make a lot of sense. Alternatively, both Quak and Berkel and Samplonius suggest that the word might have referred to another kind of building which did not necessarily fulfill a religious function. This reasoning clearly tries to avoid any association with pre-Christian religion and ignores the fact that the cognates in the Early Medieval West Germanic languages all solely mean ‘pagan sanctuary’.
In a recent article, Eldar Heide (2014) has found fault with the traditional etymological argument for Old Norse hǫrgr and argued that the occurrence of hǫrgr in Scandinavian toponyms in relation to their place in the landscape actually shows that the meaning ‘mountain or cliff side’ is secondary to an older meaning ‘enclosure’ which could refer to the enclosed cultic area of a pagan sanctuary or to a geographical enclosure in the landscape. He therefore proposes to rehabilitate an older etymology for Old Norse hǫrgr that was coined by Adolf Noreen (1894) who connected it to the Latin word carcer meaning ‘prison, barrier’. To be fair, this etymology was also considered by Quak and Berkel and Samplonius but according to Heide, the sacral meaning and the geographical meaning were interconnected and would have existed side by side in Germanic prehistory. If Heide’s argument holds true this would provide an interesting explanation for the polysemy of the Scandinavian words.
Although one could argue that the occurrence of harag in Dutch toponyms also reflects this more mundane and older meaning of ‘enclosure’, to my mind, there are no compelling reasons to do so. There are however three solid facts that argue in favor of a sacral interpretation of the toponym. 1) The meaning of all the cognates in the West Germanic languages refer to a cultic building or a cultic area. 2) the settlement of Haragon was located in the same wooded areas behind the dunes where just a few kilometers away a locality called Heligelo ‘holy forest’ (present-day Heilo) was situated. 3) the Old Dutch word harag was also attached to a river near present-day Kethel (province of South Holland) as an alternative name. This would not make sense if the word Harag referred to any mundane building or enclosure as suggested by earlier scholarship but it would make perfect sense if it referred to a significant religious building or cultic site that could serve as a widely recognized location marker. Those three facts make a sacral interpretation of the toponym Harag the best-fitting solution for the West Germanic language area.
Finally, I would like to suggest that the time has come for scholars to revisit the issue of sacral toponyms in the Germanic language areas. Recently a new and exciting dialogue is taking place between historians and philologists, and sacral toponyms may play an important role in defining some of the problems that separate the disciplines. For the past two years, the Utrecht Centre for Medieval Studies has organized a multidisciplinary workshop on the Christianization of Northwestern Europe where several approaches to assessing the Pre-Christian culture of northwestern Europe are debated, whether they be historical, theological, archeological or linguistic. Since sacral toponyms convey valuable information about the religious topography of prehistoric societies, their occurrence in the Dutch language area constitute some of the most uncontroversial evidence for the former existence of pre-Christian cultic sites in these regions, especially because pre-Christian cultic buildings prove notoriously hard to identify in the archaeological record. We may therefore add the place-name Hargen to the small but telling list of place-names (place-names such as Heilo, Elst and Woensdrecht) in the Low Countries whose etymology can be reliably connected to the mythical Pre-Christian past of the Germanic speaking peoples.
Berkel, G. Van & K. Samplonius (2006; 3de herziene druk). Nederlandse plaatsnamen herkomst en historie. Het Spectrum.
Heide, Eldar (2014). “Hǫrgr in Norwegian names of mountains and other natural features,” Namne og Nemne 31, 7-51.
Kroonen, Guus (2013). Etymological Dictionary of Proto-Germanic. Leiden Indo-European Etymological Dictionary Series. Brill.
Noreen, Adlof (1984). Abriss der urgermanischen Lautlehre. Strassburg; Verlag von Karl J. Trübner.
Quak, Arend (2004). “Germaanse sacrale plaatsnamen in de Nederlanden,” Naamkunde 34.
Rostvik, Allan (1967). Har och harg. Uppsala: Almqvist & Wiksell.
Vries, Jan de (1962). Altnordisches Etymologisches Wörterbuch. Brill.
Zeiler, Frist David (1994). “’Harga, dat ook Ketel heet,’ Voor en Vroegchristelijke sporen in het Maasmondgebied.” In: Heidenen, Papen, Libertijnen en Fijnen. Artikelen over de kerkgeschiedenis van het zuidwestelijk gedeelte van Zuid-Holland van de voorchristelijke tijd tot heden. J.C. Okkema e.a. eds. 25-32.