The Concept of Halvemål in Norwegian Linguistics – a Historiographical Account

Introduction^[1]

The Norwegian field of dialectology prospered from the last decades of the 1800s onwards, and in a 1962 review article, Oskar Bandle, the Swiss Nordist, concludes that the discipline had by then brought forward “rather considerable achievements.”^[2] In addition to paying tribute to the inventor of dialect research in Norway, the autodidact Ivar Aasen (1813–1896), Bandle highlights the contributions of Johan Storm (1836–1920), Marius Hægstad (1850–1927), and Amund B. Larsen (1849–1928), all of whom are considered representatives of a predominantly German Neogrammarian tradition (Bandle 1962, 301). How strong the influences from the German academic milieus actually were has since been debated (cf. Venås, 1986; 2001; Jahr 1996; Hovdhaugen et al. 2000; Sandøy 2013). In any case, there is little doubt that many of the fundamental Neogrammarian concepts, such as “sound law” and “analogy,” found their way to Norway, more or less directly from the German equivalents, and rapidly became part of the common linguistic terminology.

A particularly interesting item within the nomenclature of this period, which was not imported from the Continent, however, is halvemål.^[3] This paper attempts to trace the history of the notion and its use in order to clarify how a term possibly based on a folk explanation (cf. Malle 2004) can become part of the scientific vocabulary – and eventually enter the common educational discourse. By pursuing this concept in detail, we also gain insight into the sociology of science within this particular field: A notion’s impact and penetration depend in general on its propagators’ authority, and the inquiry into the use of halvemål will subsequently indicate the relative influential force of different academic figures throughout history.

Method and materials

This account is based on an extensive search of a wide range of texts on Norwegian language, folklore and cultural history in general. In principle, the relevant material is everything that has ever been published on Norwegian linguistics and philology in a broad sense. Needless to say, an exhaustive hunt for one single term in such a vast corpus is impracticable. Undoubtedly, there are occurrences of the term that I have not managed to track down, and which are therefore not considered in this paper. Nonetheless, I would claim that the selection is fairly representative, comprising data from more than 650 unique publications over a period of more than 200 years.

The digital archives of the National Library of Norway^[4] have been the essential starting point for my inquiries. Findings there have led me to further analogue sources, and additionally I have consulted general digital databases (Google Scholar, Hathi Trust, Internet Archive, Jurn, etc.). Thus, my analyses mainly draw on digitized material, but I have also conferred with a large number of texts that are not yet available digitally. As the primary focus has been on the scholarly context, I have more or less ignored instances from newspapers and general magazines. An analysis concentrating more specifically on the folk discourse would certainly be an interesting complement to this study, but here I have focused on texts from renowned academics. I have not, however, included material from the private archives of the scholars in question.

The term and the linguistic phenomenon in question

Before we turn to the historical use of this signifier and its differing denotations, I will briefly look into the morphological and semantic structure of the term: As a compound it is rather opaque, although most users of Norwegian would comprehend the latter element mål as a reference to a variety of speech (similar to “tongue” or “idiom”). The first element and the compound itself, on the other hand, are less transparent: As a simplex, halve means “half”/“halve” and can be used as an adjective, a verb and a noun (more or less as in English, although the verbal use is rather uncommon in modern Norwegian). Thus, the literal meaning of the compound halvemål would be “half speech” or “halving speech.” However, the form halve- as a prefix is atypical, as most compounds would require the monosyllabic variant halv-, for example as in halvtime (“half hour”), halvgjort (“half done”), halvveis (“halfway”).^[5] This anomalous structure suggests another origin than the common and very productive composition with halv-. I will return to the etymological question below.

Disregarding its varied application, the term consistently involves a (socio-)phonetic process that has parallels in a wide range of languages, namely the omission – and subsequently inverse insertion – of a syllable-initial and pre-vocalic aspirate, commonly represented by the grapheme ‹h› (cf. e. g. Eikel 1953 on German; Häcker 1998 on Scots; Horrocks 2010 on Greek; Milroy 1983 on (modern) English; Scragg 1970 on Old English). Strictly speaking, we are dealing with two distinct phenomena both typologically and diachronically, but as we shall see, most sources present these processes as largely intermingled. The h-deletion is the primary process and can be conceived of as a typical neogrammarian change: a systematic sound alteration without exceptions. The (erratic) h-insertion is a secondary process, which is mainly dependent on sociolinguistic conditions such as social stigma. We can assume that the speakers eventually lost their intuition as to which vowel-initial words historically had had an initial [h], and impelled by social sanctions related to the indexicality of the “lost h,” they began the process of hypercorrection.

Linguistic literature has assigned a variety of labels to these processes. The deletion process is most often referred to as “h-dropping”, “aitch-dropping”, “h-elision”, “h-fall”, “psilosis” and “H-Schwund”. For the opposite (hypercorrection) process the terms “h-restitution”, “h-insertion” and “h-adding” are the most frequent. This paper will not deliberate on the phonetic details or try to delimit the articulatory circumstances involved.

The historical lineage and diffusion of the term

Not surprisingly, the oldest written record of the term appears in Ivar Aasen’s pioneering dictionary Ordbog over det norske Folkesprog published in 1850. He provides the following definition: “Halvemaal kaldes i Sdm. den Egenhed ved et Bygdemaal, at Lyden H bliver udeladt, f. Ex. i Av for Hav, Est for Hest” (Aasen 1850, 155). Most of Aasen’s works build on his extensive compilations of empirical data from local dialects (cf. e. g. Knudsen 1963), and the entry in the 1850 dictionary indicates that the term at that time was well known in some parts of Western Norway.^[6] Aasen also included the term in a glossary where he collocated lexemes according to their semantics (published posthumously as Norsk maalbunad, Aasen 1925). Here Halvemaal is listed under the hypernym “Speech” together with a number of other compounds ending with -maal. This substantiates the impression that the common user would comprehend the term as a denomination of a variety of speech (see above). Aasen already mentions the phenomenon in his book Det norske Folkesprogs Grammatik (1848), but without applying the term Halvemaal. Interestingly, Aasen here distinctly emphasizes that the h-omission is not consistent, but varies even in the core areas (Western Norway):

Det er dog ikke Tilfældet, at H-Lyden ganske mangler; men Brugen av samme er skjødeløs, saa at den deels bortfalder, deels bruges og tildeels endog tilføies saadanne Ord, som den ikke tilhører, [...]. (At den, som nogle have paastaaet, altid bortfalder, hvor den skulde være, og tillægges, hvor den ikke skulde være, er derimod urigtigt.) (Aasen 1848, 53–54)

This point is retained, although rephrased, in the revised version of the same book, Norsk Grammatik from 1864: “[I]midlertid er der neppe noget Sted, hvor det ganske mangler; det er kun en Uorden, som er indkommen i Brugen af denne Lyd, saa at den deels mangler og deels indskydes paa urette Sted” (Aasen 1864, 113).

Several historical sources show that there has been a considerable awareness of the occasional h-omission at least since the 1700s in some parts of Western Norway: The clergyman Hans Strøm (1726–1797) makes notes of it in his pioneering naturalist descriptions from the 1760s, Physisk og Oeconomisk Beskrivelse over Fogderiet Søndmør (1762–66, 537). Later, Bishop Jacob Neumann of Bergen (1772–1848) addresses this “serious error” in his 1825 handbook for teachers:

Der er Egne paa Vestlandet af Norge, hvor den grove Feil følger fra Slægt til Slægt, at den i visse Ords Udtale lægge et h til, hvor intet h skulde høres, og kaste h’et bort, hvor det netop skulde høres […]. (Neumann 1825, 24)

Neumann further instructs the teachers to counteract the local pronunciation: “Denne Feil maa af Lærerne i Skolerne udryddes og vil da efterhaanden falde bort”. Similarly, the folklorist Eilert Sundt (1817–1875), known as Norway’s first sociologist, mentions the h-omission in the west — “the so-called Halvemaal” — in the popular journal Folkevennen in 1863. The poet Magdalene Thoresen (1819–1903) refers to “Southern Sunnmøre’s faulty pronunciation of H”^[7] in her book Billeder fra Vestkysten af Norge from 1872, without however using the term halvemål. Kristofer Randers (1851–1917), on the other hand, uses halvemål in his highly popular travelogue Søndmøre (1890), labeling it “the most conspicuous peculiarity of the Sunnmøre dialect”.^[8]

Several of the early records tend to localize both the linguistic phenomenon and the term to western Norway, predominantly the regions of Sunnmøre, Nordfjord, Sunnfjord and Nordhordland. For instance, Aasen (1848, 54) mentions “Søndmør, Søndfjord og Nordhordlehn”^[9] as the core areas for the omission of h. The same regions are listed in his Søndmørsk Grammatik (Aasen 1851, 13), though somewhat more specified: “in some districts”, “in some places”, “in a large part of the region”.^[10] Furthermore, a pivotal work like Rygh’s Norske Gaardnavne^[11] on one occasion mentions “[…] the ‘halvemål’, which originally belongs to parts of Nordfjord and Sunnmøre”^[12] (Rygh 1901, 287), thus inferring that this is the hotbed of the phenomenon. However, numerous sources demonstrate that h-dropping is known in several parts of Norway — and about the Scandinavian speaking area in general: e. g. Bandle (1973, 104); Björklund (1958, 167): Christiansen (1948, 181); Freudenthal (1870, 56); Hægstad (1900, 53; 1907, 23); Jakobsen (1897, 136); Kristensen (1909, 120); Levander (1928, 30f); Lundell (1878, 81); Matras (1957); Minugh (1985); Pettersen (1975, 299f); Seip (1958, 244); Storm (1908, 149f); Vendell (1894, 122); Widmark (1994, 189). However, only very few of these scholars employ the term halvemål.^[13] Nevertheless, the main impression from the Norwegian body of literature is that halvemål is a characteristic of northwestern Norway.

In 1873, Aasen published an extended and revised edition of his 1850 dictionary, titled Norsk Ordbog med dansk Forklaring. There he altered his definition of Halvemaal slightly to read: “Benævnelse på en Dialekt med meget forkortede Ord; saaledes paa Sdm. om den Brug, at Lyden ‘H’ bliver udeladt; f. Ex. i ‘Av’ for ‘Hav’, ‘Est’ for ‘Hest’” (Aasen 1873, 259). Interestingly, Aasen widens the scope of the term from the h-omission alone to a more general truncation, and he includes both a reference to a variety (“a dialect”) and to the phonetic phenomenon. However, he does not mention any additional types of truncation or clipping (like apocope or syncope). As we shall return to below, this notional vacillation recurs throughout the material.

Alongside Ivar Aasen, two scholars stand out as the most significant disseminators of the term in the 1800s: Hans Ross and the above mentioned Oluf Rygh. Ross (1833–1914) was Aasen’s primary adherent, and in his 1895 dictionary Norsk Ordbog, which was greatly inspired by Aasen’s methods, Ross exhibits this relationship by adding the subtitle Addendum to ‘Norsk Ordbog’ by Ivar Aasen.^[14] He makes a lengthy reference to the h-omission under the lemma “H, letter” (Ross 1895, 289). Ross also touches upon the sociolinguistic aspect of the phenomenon. He starts out by locating the trait in Sunnmøre, Nordfjord and Nordhordland, but later on he draws parallels to Sweden and England. He also mentions the h-insertion in the Oslofjord area. Lastly, Ross refers to Aasen’s publication on the grammar of the dialect of Sunnmøre, Søndmørsk Grammatik (1851), where Aasen mentions the “most peculiar omission” of the sound h (see above).^[15]

Almost concurrently, Oluf Rygh makes his first mention of “halvemål.”^[16] In a paper on farm names in the Trøndelag region, he comments:

Der forekommer jo af og til, ogsaa i Stedsnavne, omkring i Landet spredte Eksempler paa ‘Halvemaal,’ eller den i en Del af Nordfjord og Søndmør temmelig konsekvent gjennemførte Udtaleeiendommelighed, at h tilføies foran Ord, der begynde med Vokal, og omvendt sløifes i Ord, der begynde med h med efterfølgende Vokal. (Rygh 1892, 219f)

From the following year, 1893, there is documentation of Amund B. Larsen’s knowledge and use of the term in an unpublished report from one of his study trips submitted to the academic college.^[17]

A few years later, in 1898, Rygh elaborates on the notion in his introductory volume to the momentous work on Norwegian farm names (Norske Gaardnavne, see footnote 14 above). Rygh’s Indledning has been immensely influential in Scandinavian onomastics (cf. Schmidt 1998), and numerous scholars (and laypersons) have subsequently referred to the explanation of halvemål that he provides there. Contrary to Aasen’s 1873 definition, Rygh describes halvemål as a “pronunciation peculiarity,” i. e. a phonetic feature (or variable) and not a variety or dialect. Furthermore, Rygh points out that halvemål is a “popular name,”^[18] and he adds a footnote with some reflections on the possible etymology of the term. He hypothesizes that the term is a paronomasia involving alvemål, i. e. “the speech of elves”: “a name created in jest or to scorn”.^[19] This aspect of Rygh’s explanation also reappears in multiple sources. It should be noted, that Rygh mainly applies the term in the context of scattered onomasiologic examples of h-deletion or -insertion, e. g. in his discussion of *Afdalir as the possible origin of Havdøl in Trøndelag (Rygh 1901, 287), and likewise *Atlaruð > Haslerud (Rygh 1902, 142) and *Haahaugen > Aahaugen (Rygh 1908, 10).

Interestingly, Hans Ross refers straightforwardly to Rygh’s explanation in his “New addendum” to Norsk Ordbog published in 1902. Under the lemma Halvemaal he states: “may derive from Alvemaal: the unifying speech of ‘the elves’ or ‘the fairies’ (Oluf Rygh)”.^[20] He makes the same point in a footnote in the first volume of his Norske bygdemaal (“Norwegian rural dialects,” 17 volumes published 1905–1909) from 1905: “Er kanskje istadenfor ’alvemaal’; jf. alv, alvskoten og fleire ord med ’alv’. O. Rygh i si ’indledning til Norske gaardnavne’ nemner dette som ikkje heiltupp urimelegt” (Ross 1905, 29, note 3). In the same publication, Ross elaborates on the motivation of the hypercorrection aspect of the phenomenon:

Men dei [folki i desse bygder] hev av boki og av velviljuge grannar fenge vita, at det vantar ein slik ljod som h i deira uttale av mange ord. So strævar dei med aa faa fram ein h: dei som er so vidt boklærde fær sett han til i dei ordi som i dansk byrjar med h, oftast med nokor møda; dei andre set han til paa ei von, og mest der som dei ikkje hev sin eigen vernande veike andeljod, um dei elles hev denne enno. (Ross 1905, 29)

Ross also draws parallels with Sweden and England where similar “confusion” occurs alongside “the same witless general belief that people are so intractable that they insist on speaking the direct opposite of what they know to be right.”^[21] Ross’ commentaries clearly suggest that halvemål is a stigmatized feature, and that the aberrant h-distribution is commonly considered willful and recalcitrant behavior.

The following year, in 1906, three other prominent linguists employ the term halvemål in their publications: The professors Hjalmar Falk and Alf Torp in the second volume of their pivotal Etymologisk ordbok over det norske og det danske sprog (“Etymological dictionary of the Norwegian and the Danish language”), and Olai Skulerud, later to become a professor at the University of Bergen, in a treatise on toponyms in southeastern Norway. Falk & Torp (1906, 481) use it in their article on Ankel (“ankle”) where the transition ǫkull > hǫkull is in question. In his introduction, Skulerud (1906, 1) states that his work is largely based on Rygh’s data, and we can assume that Skulerud has taken over the term halvemål from Rygh’s writings. However, Skulerud’s teacher Amund B. Larsen, who originally presented the treatise to the academy’s plenary meeting, might also be the source, as he demonstrably was familiar with the term already in the 1890s (see above). None of the entries, like the discussion of *Eitrárdalr as the possible origin of Hiterdal (Skulerud 1906, 11), provides a definition or explanation of the term. One might take this as an indication of how well established the term halvemål was in philological circles at that time. Another leading Norwegian linguist at the time, Johan Storm (1836–1920), professor of philology at the university in Christiania (Oslo), refers to Ross’ writing on halvemål from 1905 (cf. Storm 1908, 150), and thus he was evidently familiar with the term. However, as far as I can see, Storm never applied it in his own theorizations. Similarly, Marius Hægstad (1850–1927), professor of Scandinavian linguistics (initially “professor of ‘Landsmaalet’ and its dialects”), addresses the phenomenon in several works (e. g. Hægstad 1907, 23; 1915, 65f; 1917, 126), but without labelling it halvemål.

1901 is the first time that halvemål appears in a collegiate dictionary, namely Matias Skard’s Landsmaals-ordlista med rettleiding um skrivemaaten. His very brief definition is “rural dialect with ‘est’ for ‘hest’ etc.”.^[22] Halvemål is similarly an entry in Lars Eskeland’s Norsk rettskrivingsordbok from 1906: “rural dialect that does not use h, or uses it wrongly”.^[23] Interestingly, both of these definitions echo Aasen’s 1873 entry by referring to an entire variety and not the phonetic phenomenon itself. Eskeland’s definition is reiterated in several later editions.^[24] He also adds halvemål to the 1919 fourth edition (first edition 1897) of his Norsk ordlista (“authorized for school use by the ministry of church and educational affairs”): “a rural dialect with omitted h”.^[25] In 1912 Matias Skard publishes another widely used school dictionary, Nynorsk ordbok for the high school (“den høgre skulen”). There, the lemma Halvemaal is defined as a “rural dialect with peculiar use of h”.^[26] Again the variety and not the phonetic trait is the main focus. Steinar Schjøtt includes the term in his 1914 Norsk ordbok “for school and literary use,” defining it as the “name for a dialect that has lost h,” followed by the exact same examples as in Aasen’s 1850 dictionary (Schjøtt 1914, 164). That same year, the term appears for the first time in a schoolbook on grammar, Leiv Heggstad’s Norsk grammatikk for skule-ungdom. Under a short clarification of the sound (and letter) h, Heggstad adds information about the h-omission “in some coastal districts west of the mountains” (1914, 26), followed by halvemål — in quotation marks and parentheses. It is not clear from the context whether Heggstad uses halvemål to denote the phonetic phenomenon or a variety. This entry was retained in the 1931 revised edition.

Predictably, all of these scholars – Skard, Eskeland, Schjøtt and Heggstad – were students at the university in Christiania (Oslo) in different time periods between ca. 1865 and 1905. That means that they were all affiliated with a philological milieu where halvemål circulated as a technical term, and they subsequently contributed to its wider circulation.

Aschehougs konversationsleksikon (1920–1925) is the first encyclopedia to enter the term halvemål. It is not a separate lemma, but it appears in the article on the region of Nordfjord (Krogvig 1923, column 707), where the local dialects are described as “partly halvemål, i. e. they lack h or often use it wrongly”.^[27] The article is unsigned, but the most likely author is the above-mentioned Hjalmar Falk, who is listed as a main contributor on linguistic matters.

Halvemål finds its way into a bilingual dictionary for the first time in 1927, in Ola Raknes’ English – Norwegian Dictionary. Raknes employs halvemål in two translations of the idiom “to drop one’s h’s” (“ho/han talar ‘halvemål’”, p. 8 and 337). The second time (under the lemma ‘H, h’) he adds an interesting amplification to the translation: “he forgets his h’s (i. e. he has not learnt how to talk properly)”.^[28] This parenthetical comment indicates the general stigmatization that pervaded halvemål. Bjarne Berulfsen, later professor of Scandinavian linguistics at the University of Oslo, published his English–Norwegian dictionary in 1933, where a similar use of halvemål can be found: “speaking halvemål” is listed as an equivalent to “dropping one’s aitches” and “dropping one’s h’s” (1933, 7 and 125).^[29]

In Gyldendals konversasjonsleksikon (Schieldrop 1933), halvemål makes its debut as a separate lemma in a general encyclopedia. The short article reiterates comments of several of the sources mentioned above, including the possible association with ‘alvemål’; and that the definition focuses on “a language with strongly truncated words”,^[30] which is then expanded on by referring to the h-omission in certain dialects, followed by the exact same examples as in Aasen (1850) and Ross (1895). Furthermore, the article states that the phenomenon is widely known, but nevertheless points to the Sunnmøre region as the most prominent area for the phenomenon. The article is unsigned, but it is most probably authored by Didrik Arup Seip (1884–1963), professor of Scandinavian linguistics at the University of Oslo for almost four decades. Seip is listed as a member of the main editorial committee, and in the slightly revised edition from 1959, the signature “D.A.S.” is added to the article. Moreover, the 1959 version makes explicit reference to Ross. Seip’s wording from Gyldendals konversasjonsleksikon later finds its way into several other encyclopedias – with minor revisions. Seip probably also provided the article halvemål to the competitor Aschehougs konversasjonsleksikon (1939–1952). Admittedly, the entry in volume VII (1941) is unsigned, but it resembles the Gyldendal version significantly, and in the revised edition from 1957, the signature “D.A.S” is added.

Similarly, halvemål has its own article in the one-volume Nyco konversasjonsleksikon (Anker and Haffner 1935, col. 508), but this text is considerably shorter and most probably not based on Seip’s original. The same article is retained in the extended Cappelens Leksikon from 1939 (Anker and Haffner 1939, col. 630). No author is specified, and in the preface the chief editors merely state that “expert authors” have been engaged from numerous disciplines.

This review demonstrates that by the early 1940s, the notion of halvemål had entered the public domain by means of a wide range of authoritative texts: mono- and bilingual dictionaries, school grammars and textbooks, and encyclopedias. Although frequently listed in brackets, the concept is employed as a full-fledged scientific term with a seemingly indisputable place in Norwegian linguistics.

Selmer (1948) – the sole academic exploration of halvemål

As demonstrated above, the Aasen, Ross and Rygh threesome were the main propagators of the term in the 1800s. Several prominent scholars rapidly included halvemål in their linguistic vocabulary, and in the first half of the 1900s, the term occurs in numerous dialectological and other philological works. Nevertheless, there is no substantial scientific exploration of the phenomenon itself before Ernst Selmer’s work from 1948. Until then, the notion is basically employed to label any type of h-omission (or -insertion). The previously mentioned Amund B. Larsen and his former student Olai Skulerud continue to use the term throughout their careers (cf. e. g. Larsen and Stoltz 1912, 237; Larsen 1926, 54; Skulerud 1922, 375; 1924, 7). Other examples are the philologists Magnus Olsen (1912, 36), Gustav Indrebø (1917, 119; 1931, 41) and Trygve Knudsen (1924, 138), and the dialectologists Eilert Mo (1922), Lars E. Søreide (1930, 36), Halvor Dalene (1947, 106) and Hallfrid Christiansen (1948, 181).

Ernst W. Selmer (1876–1968), professor of Germanic philology and general phonetics at the University in Oslo, in 1948 published the first – and as yet only – scientific study of halvemål as an articulatory phenomenon. Selmer was a prominent phonetician and served as head of the department of phonetics for more than 40 years (Jahr 2009). His work on halvemål is part of a 120-page treatise called Sunnmøre-Studier (lit. “Studies from the region of Sunnmøre”) containing the subsections “Musical and dynamic accent”, “Halvemål” and “The crossed d (ð)”.^[31] Selmer primarily undertakes experimental phonetic analyses of the phenomenon (kymographic measuring), but he also delivers an interesting theoretical discussion of h-omission and “false h-prosthesis”^[32] in general. Moreover, he carries out what we today would call a folk linguistic survey on halvemål in the region of Sunnmøre. By questioning pupils, teachers and “other interested parties”, Selmer contributed some interesting insights into the status of the halvemål spoken at that time, although he judged the results to be “rather meager, both quantitatively and qualitatively” (Selmer 1948, 87). Nevertheless, he shows that the notion is still well known, although the actual phonetic trait is reported to be on the decline in most areas. The informants also report substantial differences between the generations, and there are several signs that the feature carries a certain stigma.

Post-war linguists and halvemål

It is difficult to estimate the impact of Selmer’s work on fellow linguists in Norway, but the references made to his halvemål-treatise in the subsequent decades are scant. One of the few can be found in the relatively lengthy encyclopedia article halvemål in volume 5 of Norsk allkunnebok (Sudmann 1948–1964). The author was Magne Oftedal (1921–1985), at that time a university fellow in general linguistics, later professor of Celtology at the University of Oslo. Oftedal’s definition starts out rather conventionally by mentioning both halvemål and alvemål and focusing on the variety: “halvemål el. alvemål er nemning på målføre der lyden h anten ikkje vert nytta el. vert nytta på annan måte enn i normaltalemål” (Sudmann 1953, col. 734). In the following wording, however, Oftedal appears considerably more nuanced than many of his predecessors. He points out that the h-omission may be gradual and not regular, and that the speech rate and intra-sentential placement may impinge on the h-quality. Interestingly, he emphasizes, much like Aasen (1848, see above), that the omission and insertion of h do not occur systematically or invariably: “[D]en vanlege trua at mange konsekvent seier øy for høy og høy for øy, [er ikkje rett]”. Lastly, Oftedal refers to Selmer (1948) for further reading.

The term halvemål continued to be used among distinguished post-war Scandinavian philologists, notably by Per Nyquist Grøtvedt (1954, 179; 1969, 125; 1970, passim), Ragnvald Iversen (1954, 58; 1957, 170), Per Hovda (1961, passim), Einar Lundeby (1981, 73; 1995, 28), Egil Pettersen (1975, 299f) and Kjell Venås (2002, 194). The dialectologists Olav T. Beito (1963, passim), Gunnvor Rundhovde (1964, 28), Kåre Elstad (1982, 76), Arnold Dalen (1997, 198), Audun Fitje (1995, passim) also used the term, as did sociolinguists like Helge Sandøy (1987, 104; 1990, 69; 2011, 218) and Randi Solheim (2010, 277) and onomasiologists including Tom Schmidt (1981, 414), Eivind Vågslid (1974, passim; 1978, 112f), Peter Hallaråker (2000, 148), Inge Særheim (2001, 368), Berit Sandnes (2003, 49), Kristoffer Kruken (2006), Frode Korslund (2006), Margit Harsson (2008, passim) and Kåre Hoel (2014). Although he was not primarily a linguist, the literacy historian Jostein Fet (2003, passim; 2010, 138) could be added to this list.

The fact that several sub-disciplines within Norwegian linguistics and philology have employed the term indicates the prevalence of the concept, and this also clarifies why it is found in many textbooks and general reference books throughout the latter part of the 20th century.

The introductory chapter of the primary school dictionary Ordliste for grunnskolen (Bakke and Dalene 1968) includes halvemål among the dialectal characteristics (“målmerker”) that the pupils are supposed to be able identify in their own dialect. The short explanation provided (“omission or incorrect insertion of the h-sound”^[33]) is omitted in the 1975 edition. This and subsequent editions (1982 and 1987) give the following instruction: “De få fagorda vi bruker, får dere slå opp i andre handbøker”. Evidently, halvemål is among the “technical terms” that can be readily found elsewhere.

In his 1969 textbook on “Nynorsk” in upper secondary school (Nynorsk i gymnaset), Magne Thingelstad uses halvemål as a term. He briefly explains the phenomenon (“loss of the h-sound”) and further points to parallels in the Romance languages and English Cockney. This explanation is retained in the 1977 revised edition.

Apparently, the most recent textbook that includes halvemål is Språkboka for den videregående skolen (Nøklestad and Tveterås 1999).^[34] Here halvemål is listed among several other dialectal characteristics in a chapter on Norwegian dialects: “I nokre vestnorske dialektar og elles i landet fell h-en av og til bort i framlyd” (1999, 171).

Didrik A. Seip’s article on halvemål is replicated in several encyclopedias throughout the post-war period. It appears for the last time in Gyldendals store konversasjonsleksikon in 1972 (Bull et al. 1972). For the joint project Aschehoug og Gyldendals Store norske leksikon (1978–1981) the lexicographer Dag Gundersen (1928–2016) authored a new article based on Seip’s text. His article was retained in the second and third editions (1988 and 1997), and most recently it went online in the Internet versions from 2000 onwards. The current entry in the online version of Store Norske Leksikon^[35] still has Gundersen’s signature:

Halvemål er dialekter som utelater h i fremlyd foran vokal, for eksempel av for hav og est for hest, og omvendt setter inn h i ord der den ikke hører hjemme, for eksempel høy for øy. Fenomenet er kjent fra flere engelsk og svenske dialekter. I Norge finnes dette mest på Sunnmøre og i litt av Nordfjord, Sunnfjord og Nordhordland. Etymologi: kanskje egentlig alvemål, det vil si det mål som alvene brukte.

Although the wording differs slightly from that of Seip, we can see that the authoritative definition of this concept has been almost identical for more than 80 years. It highlights the variety (and not the phonetic trait itself); provides examples that date back to Aasen; draws parallels to other languages; localizes the phenomenon to parts of northwestern Norway, and it includes a note on the possible etymology involving elves.

Similar definitions can be found in several dictionaries from the 1970s onwards. Tanums store rettskrivingsordbok denotes halvemål as a linguistic term for a “dialect where h is not pronounced before a vowel”^[36] (Sverdrup et al. 1985, 166). A very similar definition occurs in the widely used Escolas ordbok (Taule 1988). The editions from 1996 and 1998 (Bokmål and Nynorsk respectively) retain the entry, as do the 2005 editions.

The widely used – and highly authoritative – dictionaries Nynorskordboka (Hovdenak et al. 1986) and Bokmålsordboka (Landrø and Wangensteen 1986) both include the lemma halvemål with a definition that echoes Seip’s and Gundersen’s texts^[37]: “possibly originally ‘alvemål’. Linguistics: dialect that omits initial pre-vocalic h, and conversely often inserts it where it does not belong”.^[38] These definitions have been maintained in the digital versions of both dictionaries (http://ordbok.uib.no). An almost identical definition is listed in the general dictionary Norsk ordbok (1996) and reiterated in the 2005 edition.

The halvemål entry in the monumental Norsk ordbok: ordbok over det norske folkemålet og det nynorske skriftmålet^[39] strongly resembles the definitions mentioned above. The entry draws on Aasen, Ross and Rygh and further presents halvemål as a linguistic term for “language use where the words are truncated through the elimination of initial pre-vocalic h.”^[40] However, it is worth noticing that this definition employs the notion “language use”^[41] instead of “dialect” or “variety of speech.”

Interestingly, this aspect has been at the core of the definition of halvemål in most of the dictionaries published by the Norwegian Academy for Language and Literature (Norw. “Det Norske Akademi for Språk og Litteratur”) from the 1977 Riksmålsordboken onwards. There halvemål is defined as a “dialect phenomenon” and not a variety of speech (Guttu 1977, 256). The exact same wording is maintained in Aschehoug og Gyldendals store norske ordbok (Guttu 1991, 231), which to a large extent is based on the 1977 publication. Furthermore, Norsk illustrert ordbok (Guttu 1993), Norsk ordbok med 1000 illustrasjoner (Guttu 2005) and Kunnskapsforlagets store norske ordbok (Guttu 2017) all include this definition.^[42] The digitalized dictionary Det Norske Akademis ordbok (NAOB) adopted the same article with some minor changes.^[43]

The etymological background of halvemål

As indicated above, some scholars have touched upon the etymology of the term, but no one has discussed it extensively. Oluf Rygh’s (1898) brief comment on the possible link to the (unattested) Norse alfamál (“elves’ speech”) in his seminal introduction to Norske Gaardnavne is frequently passed on in the literature and appears to be the general recognized explanation. Norsk ordbok additionally mentions halve as a possible primary signification of the first element (Vikør 2002, col. 1357).^[44] However, most sources seem to favor the association with alfr (“elf”), and as we have seen, this connection is incorporated in several digital dictionaries and the most used Norwegian web-based encyclopedia.

The prominent philologist Magnus Olsen (1878–1963) is among the proponents of Rygh’s idea (cf. Olsen 1912, 36; 1951; 175). Olsen elaborates on the possible connection to the mythological creature alfr (“elf”, “fairy”), which is known from both Old Norse poetry (Poetic Edda) and later folklore. In Norse mythology the alfr was an anthropomorphic figure that belonged to the underworld (cf. e. g. Holtsmark 1990, 77). As Ármann Jakobsson (2015, 216) rightly points out, alfr in medieval textual sources is a fairly broad term that denotes not only a specific race or species or even categories of elves, but is used for all paranormal figures that are superior to humans. The Prose Edda (Snorri Sturluson’s work) distinguishes between the ljósalfar (“light elves”) and the døkkalfar (“dark elves”, also svartalfar “black elves”), where the first type is characterized by its benevolence and beauty, and the latter by its grimness and animosity towards humans. However, it seems unlikely that this categorization was generally accepted in medieval Scandinavia, and there are many indications of a prevalent ambiguity about whether alfar were good or evil (Ármann Jakobsson 2015, 216). The later ballad tradition (ca. 1350–1550) maintains this vagueness regarding the elves’ relationship with humans, but nevertheless consolidates the belief that these creatures have the power to intervene in the earthly realm, be it with a helping, bewildering or hindering purpose.^[45] The best-known ballad dealing with elves is arguably “Olav Liljekrans” (Eng. “Olaf lily-lei”), which has its parallels in several parts of Europe (Jonsson 1992). A central motif in this text is precisely how elves might inflict injury on mankind by causing illnesses and evils (Bø 1977, 250f).^[46]

Given this background we might assume that the term (h)alvemål is meant to suggest a type of speech that is being meddled with by supernatural creatures, the elves. A somewhat ingenious aspect is that the term itself can be considered a product of the tampering: The insertion of an h-sound thus materializes the denotation of the allegedly paranormal phenomenon. Conceivably, this explanation was intended as a joke, as Rygh proposes.

An additional link to the elves’ speech is based on the insight into the alfamál from the poem Alvíssmál (“Talk of Alvíss”) in the Poetic Edda. Here the dwarf Alvíss comes to Thor and accounts for how diverse mythological beings (among them the alfar) express themselves in their surroundings. The crucial point is that these beings use different types of heiti or kenning — poetic synonyms — for different objects, earthly as well as celestial (cf. e. g. Grundtvig 1870, 235). The euphemistic circumlocution that characterizes “classical” elves’ speech has also been considered a distinguishing trait of “modern” halvemål: The folklorist Svale Solheim (1940, 126) suggests that halvemål has been used by fishermen to conceal local names and thereby make it harder for outsiders to obtain the tricks of the trade.^[47] Older sources reveal that this type of taboo language has been employed in the Norwegian Sea area for a long time (cf. e. g. Strøm 1762–66, 536; Jakobsen 1897, 82f.), but Selmer (1948, 102) does not support the idea that the h-omission is a part of this. Olsen (1951, 175, footnote) further makes a brief comment on how the physical or phonetic nature of halvemål relates to elves by assuming that alfamál “has been spoken in such a way (with a whisper or inhaled speech or mumbling voice) that h-sound had to fall away”.^[48] We might view Olsen’s bold theory in relation to the folklore notions of elveblest and alvegust (lit. “elves’ puff ” and “elves’ gust”), which were widely used appellations for different illnesses (mainly skin eruptions) that were supposedly instigated by the breath of elves (or other creatures from the underworld).^[49] The folk literature contains some accounts of the whispering noises expressed by elves (cf. e. g. Feilberg 1910, 112f.), but even if a feeble link between elves and aspirate sounds does exist, further support of h-omission or h-insertion as a traditional characteristic of the mythological elves’ speech is difficult to find.

The other imaginable etymological strand, halve, is grounded on the concrete truncation of words. Several sources suggest that the term halvemål is commonly conceived of as a variety where words are “halved” (e. g. Støyva 1994). Even though the elision only concerns one initial phoneme, the folk perspective is that the result is incompleteness or defectiveness. Selmer (1948, 102) reasons that a concept grounded on truncation is illogical considering that h-insertion is at least as frequent as the omission. However, we might assume that the main aspect of the “halving” is the rendering corrupt or imperfect rather than the concrete curtailing.

In a humorous verse on the Norwegian language dispute, Idar Handagaard (1874–1959) waggishly collocates halve maal and halvemaal to allude to the contamination of the Danish-Norwegian language (1924, 29).^[50] Similarly, the author Anders Hovden (1860–1943) refers to halvemål with the denotation “incorrect and bad language” in a private letter to Ivar Aasen in 1893.^[51] This disapproving aspect is also reflected in the traditional notion det rånge målet (lit. “the wrong speech”), which has been used pejoratively for the h-omitting varieties (Støyva 1994; Dvergsdal 1996, 13). Likewise, the term kråkemålet (lit. “crows’ speech”) occurs as a synonym for halvemål, accentuating its incompleteness and inappropriateness (cf. e. g. Ølmheim 1980, 406; Sandøy 1990, 69).

Similar notions exist in Swedish, where halftal (“half-speech”, Vendell 1894, 122), alvä ård (h-less variant of halfva ord, “half words”), halvórdä and halvördt (“half-worded”) have been used in reference to h-omission (cf. Rietz 1862–67, 238; Gustavson 1991, 312 and 314).^[52]

To summarize, we can argue that the provenience of halvemål is rather muddled, and to establish an undisputable etymology is impossible. An original construction where “alv-” (alfr) is the root of the prefix is plausible both in view of the many other compounds with “alve-” and the parallel notion dverg(a)mål (dvergmáa/-máli, lit. “dwarfs’ speech”, i. e. “echo”). The motivation for this signification would be the folk belief that elves may interfere in peoples’ lives, as well as in human language. The variant with h would quite simply be the halvemål speakers’ pronunciation of alvemål (*alfamál). However, there is little evidence that the variant ‘alvemål’ has been used orally to any great extent: A search in the vast material that Norsk Ordbok is based on, renders only one hit, which is a written source from 1984. Even though the connection to the alfr-element is possible, it is, admittedly, rather far-fetched. Thus, the alvemål hypothesis might just as well be the product of an academic flight of fancy on Oluf Rygh’s part, subsequently reinforced by Hans Ross and Magnus Olsen, and disseminated by a large number of scholars.

The explanation involving the “halving” (i. e. truncation) might appear far less exotic, but considering the parallels from Swedish, which indicate that the signification “truncated, deficient, curtailed language” has been widespread, makes this hypothesis more plausible. Gun Widmark (1994, 190f) proposes that h-omission and “false restitution” historically was prevalent in large parts of the Scandinavian-speaking area, an assumption that is further corroborated by the fact that similar notions have been circulating in different regions.

Moreover, theco-occurrence of similar terms also suggests that the provenience of halvemål might be a scholarly context and not primarily based on a folk etymology. One could presume that terms involving the half-element were exchanged in oral communication (and even in written texts which have later gone missing or remained unexplored) between (pre-)dialectologists and other academics. However, it would be very hard to establish the facts in this matter, considering that the relevant sources are very scant.

The application of the term – epistemic aspects

Lastly, I will turn to the scientific application of the term, with regards to its differing epistemic status. As we have seen above, the notion halvemål might originate from a folk explanation, i. e. a picturesque term (cf. Harré 1970, 47). We can postulate that this folk notion once entered the common language because it was conceived of as useful, even if it was also considered a misnomer (assuming that elves have never existed). In the 1850s, the term entered the academic context, and already around the turn of the century it was established as a scientific term in the linguistic milieu at the only Norwegian university at that time, the Royal Frederick University in Christiania (Oslo). This does not mean that all early Norwegian linguists employed it, Johan Storm and Marius Hægstad are examples of prominent professors who did not, but for many others the notion seems to have been the natural choice. However, as mentioned earlier, the signification has varied noticeably since Aasen’s days, and the term’s employment as a scientific concept is equally inconsistent.

The most frequent occurrence of the term in the academic literature is as a parenthetic synonym for “h-omission” and similar notions^[53] — often in quotes. Amund B. Larsen, for instance, is consistent in his use of quotation marks, as is Gustav Indrebø.^[54] Hans Ross and Olai Skulerud mostly use quotes, but there are a few exceptions to this. In several cases an additional “so-called” is included, possibly to emphasize the folk character of the term: e. g. “det såkalte ‘halvemaal’” (Falk and Torp 1906, 481), “det sokalla ‘halvemål’” (Bjørlykke and Liset 1935, 10), “de såkalte halvemål” (D. A. Seip in Werenskiold and Winsnes 1941, col. 651), and “såkalla ‘halvemål’” (Skjerpe 1979, 89). Interestingly, this dissociating conduct can be found throughout the entire period. A small number of scholars use the term more straightforwardly and mostly without any such reservation. Most notably the onomasiologist Eivind Vågslid who applies the term throughout his writings.

Another recurring pattern is that halvemål is used as a denominator of the phonetic substitution phenomenon (the actual h-omission and h-insertion) far more often than as the name of a variety. As shown above, the latter use is predominant in texts meant for a broader audience (dictionaries, textbooks and encyclopedias). Thus, the many dictionaries that explicitly label halvemål as a linguistic term for a variety are — strictly speaking — inaccurate since the prevailing scientific use primarily denotes the phonetic phenomenon per se.

Furthermore, the literature shows that halvemål has been employed both as a descriptive term and as an explanatory concept. Thus, the notion has been applied on two different epistemic levels. The first application is the prevailing one, and examples can be found in texts from several different linguistic disciplines and scholars. Instances of the second type occur most often in academic texts by authors who are not primarily linguists, even though there are also examples from linguistics studies. As already mentioned, the multi-volume work Norske Gaardnavne contains several instances of onomastic details that are explained as “due to halvemål” or “caused by halvemål” (e. g. Rygh 1908, 39 and Rygh 1910, 111). This idea has been passed on to several works on local and social history (mainly authored by historians) where toponyms are reviewed, often with explicit reference to Rygh (and his successors). Some illustrative examples are: “Denne forma må ha sin grunn i halvemål” (Slyngstad 1951, 39); “H-en er bortfalt ved det såkalte ’halvemål’” (Rynning 1968, 172); “At h er borte i dialekta […], er utslag av ’halvemål’” (Rød 1994, 332); “[…] her måtte halvemålet ha gjort namnet utydeleg” (Krogsæter 1995, 147). In these cases, the concept halvemål is ascribed an agency that in itself instigates linguistic change. Consequently, a taxonomic principle appears in the guise of a causal principle (cf. Harré 1970, 203) and thus provides the concept with added explanatory power. We can observe the same mode of thought in Korslund (2006, 53): “Vekslingen mellom formene med og uten h skyldes fenomenet halvemål”. With explicit reference to Rygh (1898) the writer here appoints halvemål the active vehicle in a linguistic process rather than employing the notion as mere description. Several of these scholars touch upon (more or less overtly) the sociolinguistic conditions that may explain the motivation behind the h-substitution. As we have seen, Ross mentioned this already in his 1895 publication, and aspects of the social stigma associated with the erroneous distribution of h — and the general awareness of it — can be found in several sources.

Concluding remarks

What I have been examining here, is the history of a linguistic term that most probably has its origins in a language-related folk explanation. Even though the etymology is blurred, we can regard the notion as a popular attempt at describing a linguistic conduct that has been known for hundreds of years.

As prominent scholars – chiefly affiliates of the philological environments in Oslo – applied the term over time, it earned the status of a scientific term and eventually spread to the public educational discourse. Moreover, the term was included in several seminal textbooks related to Scandinavian linguistics (e. g. Christiansen 1948 and Beito 1963) and hence appears as a key term in the dialectology and onomastics fields.

At the present time halvemål features in most digital dictionaries and encyclopedias, commonly categorized as linguistic terminology. What is striking with the scholarly application of the term, however, is the recurring effort to keep the term at a certain distance by framing it with different modifiers, such as “so-called,” inverted commas, or brackets. This conduct indicates that many scholars have disavowed the term, at least partially, even though halvemål has belonged to the linguistic terminology for 150 years. Nevertheless, it has subsisted as a quasi-scientific concept, occurring in a wide range of papers and theses. Most recently, it has been included in the new four-volume language history Norsk språkhistorie (2016–2018),^[55] and then chances are good that halvemål will endure well into the 21st century, albeit tucked away in brackets.