Available via license: CC BY 4.0
Content may be subject to copyright.
Citation: Eide, Kristin Melum, and
Arnstein Hjelde. 2023. Linguistic
Repertoires: Modeling Variation in
Input and Production: A Case Study
on American Speakers of Heritage
Norwegian. Languages 8: 49.
https://doi.org/10.3390/
languages8010049
Academic Editors: Cristina
Guardiano and Giuseppina
Silvestri
Received: 2 September 2022
Revised: 15 January 2023
Accepted: 18 January 2023
Published: 6 February 2023
Copyright: © 2023 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
languages
Article
Linguistic Repertoires: Modeling Variation in Input and
Production: A Case Study on American Speakers of
Heritage Norwegian
Kristin Melum Eide 1, * and Arnstein Hjelde 2
1
Department of Language and Literature, Faculty of Humanities, NTNU: Norwegian University of Science and
Technology, 7491 Trondheim, Norway
2Department of Languages, Literature and Culture, Faculty of Teacher Education and Languages,
Østfold University College, 1757 Halden, Norway
*Correspondence: kristin.eide@ntnu.no
Abstract:
Heritage Norwegian in the American Midwest is documented through a corpus of record-
ings collected and compiled over a time span of 80 years, from Einar Haugen’s recordings in the
1940s via the CANS corpus up to the present-day in the authors’ own recordings. This gives an un-
precedented opportunity to study how a minority language changes in a language contact situation,
over several generations and under gradually changing circumstances. Since we also have thorough
historical knowledge of the institutions and societal texture of these communities, this privileged
situation allows us to trace the various sources of input available to the heritage speakers in these
communities in different relevant time slots. We investigate how the quality and quantity of input at
different times are reflected in the syntactic production of heritage speakers of the corresponding
generational cohorts, focusing on relative ratios of specific word orders (topicalization and verb
second, prenominal and postnominal possessive noun phrases) and productive morphosyntactic
paradigms (tense suffixes of loan verbs). Utilizing a model of relations between input and output,
receptive and productive competence, to show how input–output effects will accumulate throughout
the cohorts, we explain the observed linguistic change in individuals and society.
Keywords:
Heritage Norwegian; quality versus quantity of input; linguistic change; language contact;
linguistic repertoires; receptive versus productive competence; syntactic variation; accommodation;
intergenerational transmission
1. Introduction
The recent massive growth spurt of heritage language research provides a rich body
of new types of evidence challenging fundamental questions and assumptions in the field
of linguistics. This paper provides new data to bear on the well-attested observation
that the linguistic production of heritage language speakers differs from monolingual
homeland speakers in a non-arbitrary fashion. Wiese et al. (2022, p. 3) suggest that
this difference is often described in terms implying “a deficit-oriented view of heritage
speakers»; employing notions such as “incomplete acquisition, attrition and loss”. The
authors instead advocate terms such as “development, variation, and innovation” signaling
a turn toward “a research program that does not take monolingual standard norms as
a yardstick to identify what is missing or incorrect in heritage speakers’ language use”.
Wiese et al. (2022) argue that heritage speakers are like any other type of speaker, and
should be expected to exhibit interesting patterns of variation deviating from standard
norms in particular ways, just like monolinguals and other native speakers. This is clearly
reminiscent of the view on heritage languages in Amaral and Roeper (2014, p. 29), the
framework of Multiple Grammars (MGs) Theory, where any human grammar consists of
sets of sub-grammars, sometimes contradictory.
Languages 2023,8, 49. https://doi.org/10.3390/languages8010049 https://www.mdpi.com/journal/languages
Languages 2023,8, 49 2 of 36
[H]eritage and L2 grammars are frequently described as ill-conceived versions of
[
. . .
] artificial gold standards based on idealized L1 [
. . .
] monolingual grammars
[
. . .
]. From an MG perspective [however] all grammars are created equal. Each
individual will have a grammar with a unique configuration of rules, and these
individual configurations may converge or diverge from what is considered
standard in a given language by different social groups. However, an individual
grammar is never deficient in any way.
We applaud this turn, dovetailing with an emerging wave within multilingualism
research at large. The negative take on multilingual societies, where hardly any multilingual
speaker can ever be assumed to reach “ultimate attainment” as “non-monolingual speakers”
are inherently incapable of sufficient proficiency in any language, is being replaced with
a view where multilingualism, bidialectism and heritage language maintenance creates a
varied repository of potent linguistic resources, intrinsically valuable to both individuals
and society.
Heritage languages undoubtedly differ from the idealized monolingual homeland
variety, an observation often tied to the quantity and quality of input provided in the
heritage language; cf. Polinsky and Scontras (2020, p. 14):
[R]educed input quality–in addition to reduced quantity–appears to play a central
role in the unique outcomes of heritage speakers. The causes behind this effect
remain to be explored, but Gollan et al. (2015) suggest that richer variety in the
input could lead to more robust encoding of the relevant representations.
Meisel (2020, p. 33) questions the conjecture (adopted by “virtually all heritage
language scholars”) that the quality of input received by heritage language learners is
lessened due to the reduced number of potential interlocutors. “In comparable situations
of bilingualism, children still develop two native competences,” and Meisel argues instead
that quantitative differences are the cause of heritage speakers’ divergent development.
The challenge is to quantify the necessary minimum, an issue where “heritage language
research has not yet provided valuable information”.
Although we are in no position to pin-point the necessary minimum for “successful
acquisition of a (heritage) language”, this paper adds to the discussion on the intertwining
aspects of quantity and quality of input in a heritage language context. We give an overview
over the main sources of (Norwegian) input available to a typical speaker of Heritage
Norwegian in a Norwegian settlement in the American Midwest at different times in a time
span of roughly 80 years of history of this diaspora. We further trace how a selected set
of morphosyntactic features in the available data, within the same time period, gradually
responds to this changed availability of input sources through five idealized generations–
cohorts distinguished by their birth year.
The linguistic production of heritage speakers has been shown to display simplifi-
cation on all linguistic levels; the morphosyntactic features we study are also simplified
under crosslinguistic influence. We frame our observations in the presence vs. absence
of productive distinctions acquired and maintained by language users. This allows us to
refrain from otherwise required separation between features at different linguistic levels
(e.g., vocabulary, morphosyntax, phonology) and to talk about linguistic repertoires
1
instead.
These features cluster and covary, and in many cases a separation is neither possible nor
fruitful (cf. discussion in Section 5).
Ideally, the process of language acquisition and maintenance for all types of speakers
should be captured by the same type of model, and in Section 5we introduce a model which
is suitable for all types of language contact situations and has the advantage that differ-
ences between speakers of Heritage Languages and other native speakers are minimized
2
(cf. Amaral and Roeper 2014, p. 29).
Moreover, language acquisition and maintenance may possibly always take place in
language contact situations; “multilingualism is the normal condition for human language”
(Wiese et al. 2022, p. 3) and every language user is multilingual in a particular sense of the
Languages 2023,8, 49 3 of 36
word, namely that “any person (no matter how “monolingual”) has numerous grammars”
(Roeper 1999, p. 170); cf. also Cook and Newson (2007); Eide and Sollid (2011); Roeper
(2016); Eide and Åfarli (2020), inter alia. In this case, the model should apply even in what
is usually seen as purely “monolingual” situations.
In this model, heritage speakers are no different from other speakers, as the linguistic
biography for each language user will be unique by necessity, no matter the acquisition
situation. This unique linguistic biography will be determined not only by each and every
source of input available to the speaker and the speaker’s continuous intake, but also by
the temporary and more permanent selections and choices made by the speaker regarding
production, determined in part by the speaker’s preferences, leniency and personality.
Not all speakers are equally prone to accommodate their language to an interlocutor, and
speakers are put in situations where such choices are necessary to very different degrees.
This affects the speaker’s production, the memory strength of specific linguistic features,
and correspondingly the possibility for maintenance viz language attrition in the individual
speaker. As the production of one generation constitutes substantial parts of the relevant
input for the next generation, the individual choices of each speaker will accumulate and
eventually manifest as linguistic changes in the society; more so if the available input
sources are few.3
2. Studying Norwegian Immigrants in the American Midwest: Language and Society
Norwegian immigration into the American Midwest started with the arrival of a group
of 53 Norwegians, (mostly) confessed quakers, in New York in 1825. By 1930, more than
800,000 Norwegians had arrived in the US and 40,000 in Canada (Johannessen and Salmons
2015, p. 10). Only Ireland had a higher emigration rate than Norway; in both countries,
a huge population growth prompted this emigration surge—the number of Norwegian
immigrants in the US matched the remaining population in Norway in 1800 (Haugen 1953,
p. 29), and the number of people reporting Norwegian descent in the US (about five million;
US 2000 census) matches the current population of Norway.
The Norwegian North American Heritage Language (henceforth NorNAHL
4
) has been
subject to linguistic studies since the early 1900s. The pioneer of the field is George T. Flom,
professor of linguistics at the University of Iowa (1900–1909) and the University of Illinois
(1909–1927), who published a range of papers between 1901 and 1931 on various aspects
of NorNAHL. His main focus was on the vocabulary and especially on the integration
of English loanwords into NorNAHL vernacular (e.g., Flom 1929), a trend continued by
many linguists studying NorNAHL in the 20th century (e.g., Hjelde 1992,1996;Åfarli 2012;
Riksem 2018;Johannessen and Laake 2017;Hjelde et al. 2019).
Professor Flom’s investigations laid the important groundwork for his student Einar
Haugen, who started collecting data in 1936 for what was to become The Norwegian Language
in America (1953). Like Flom’s papers, Haugens’s study—often referred to as the starting
point of contact linguistics as a scientific field—also had vocabulary integration as one
main area of interest, but Haugen’s approach was broader and methodologically more
advanced than that of his predecessor. Haugen had 98 informants complete a very extensive
questionnaire of (up to) 795 items (later reduced to about 550), usually requiring 8–12 h of
test time for each informant, and he recorded more than 50 h of interviews with altogether
207 speakers of NorNAHL. This allowed Haugen to address a wider range of topics
(including folklore and onomastics), explicitly separating language issues from societal
issues. Thereby, he established core premises to scholars within the field, not only in
language contact and bilingualism, but in the field of linguistics at large.
Around 30 years after the publication of Haugen’s seminal and groundbreaking work,
the University of Wisconsin-La Crosse launched The Norwegian Language and Culture Project
prompting a renewed exploration of NorNAHL in which Arnstein Hjelde was invited to
conduct fieldwork on the Trønder dialect in the US. This led him to further explore and
document the Norwegian spoken in the area surrounding the two small Wisconsin towns
Languages 2023,8, 49 4 of 36
Coon Valley and Westby, where he collected about one hundred hours of recordings of
contemporary NorNAHL.
Apart from Hjelde’s work, the studies and publications on NorNAHL in this time were
still few and far apart, but a real game changer was to appear on the scene: The Norwegian
American Dialect Syntax Project (NorAmDiaSyn) led by the very industrious professor Janne
Bondi Johannessen at the University of Oslo.
5
In September 2010, she brought a group of
Norwegian trained linguists to conduct thorough fieldwork on NorNAHL to the Midwest,
and for the next decade she arranged numerous field trips, organized and co-organized
many conferences and co-edited a number of books and special volumes on NorNAHL
and other heritage languages (cf. Johannessen and Salmons 2015;Johannessen and Putnam
2020;Johannessen 2021). She even managed to track down (with the aid of her colleague
Kristin Hagen) a long-forgotten chest filled with wax cylinder recordings of NorNAHL
speakers made by Norwegian professors Didrik Arup Seip and Ernst W. Selmer, still in the
“America-trunk” where they were stored for the Atlantic crossing to Norway in 1931.
Janne Bondi Johannessen’s ever-present objective was to make research data available
to the entire research community, resulting (among many other things) in the searchable
corpus CANS—Corpus of American Nordic Speech (Johannessen 2015). This corpus rests
heavily on fieldwork conducted in 2010 and onward under Johannessen’s leadership,
where the heritage speakers were recorded on video, typically in conversation with another
heritage speaker and/or a fieldworker. These recordings normally lasted between one and
two hours. A selection of these recordings is transcribed and included into the corpus.
CANS has been extended continuously and currently its Norwegian part consists of about
750,000 individual words. At present, it includes some of Haugen’s and Hjelde’s original
recordings and also some examples of American Swedish. CANS has provided materials
for over 100 scientific papers by more than forty scholars since its inception, with more
to come.
2.1. Rural Settlements and Close-Knit Communities
Leaving their farming communities behind, most Norwegian immigrants sought the
means to continue a life as farmers in the new environments, and the Midwest promised the
land necessary to build new, close-knit farming communities. Most Norwegian settlements
resulted from chain immigration, as new groups of immigrants clustered where family
members and neighbors had landed and settled before. The rather impressive long-lasting
preservation of NorNAHL in the US was clearly facilitated by this strong clustering ten-
dency. Here, the Norwegians resemble the Swedish, the latter amounted to one million
immigrants in the relevant time slot. The Danish immigrants came in relatively smaller
numbers (275,000 in 1840–1914), settled in urban areas, and often married non-Danish
partners. Hence, unlike Swedes and Norwegians, Danish immigrants, to a lesser degree,
dominated entire communities (cf. Kjær and Larsen 1972;Johannessen 2018). Norwegian
sociologist P. A. Munch, studying Norwegian settlements in the late 1940s, observed that
some of these communities could be rather self-absorbed, unforgiving and isolated from
the greater society (Munch 1949, p. 784):
There is a strong loyalty to the community and a correspondingly strong social
pressure against any deviation from the accepted local pattern. What foreign
elements have come in have either been assimilated completely to the cultural
pattern of the community or they have been isolated socially until they preferred
to leave.
The Norwegian settlers rapidly refurbished their societies with a number of key soci-
etal institutions promoting “horizontal ties” (Salmons 2005a,2005b) between the Norwegian
immigrants, such as churches, hospitals, old people’s homes, schools for Norwegian lan-
guage learning, newspapers in Norwegian and even colleges, e.g., Luther College (Decorah,
Iowa) and St. Olaf College (Northfield, Minnesota). The Norwegian churches, almost exclu-
sively Lutheran, imported pastors from the homeland from the outset, regularly providing
services and confirmation ceremonies in Norwegian up until WWII in some settlements.
Languages 2023,8, 49 5 of 36
This early installment of key institutions led to rather self-contained communities where
the inhabitants could cope without speaking much English.
Even way beyond first-generation immigrants, individuals are attested to prefer to
conduct their daily communications in Norwegian instead of English. “Some speakers,
and not only the first generation, still [in 1986] found it easier to speak Norwegian”; Moen
(1991, p. 113). An informant quoted by Moen stated that “Speaking English makes me
sweat”, and our informant “Lars”,
6
born in 1961, claims that his parents (third-generation
immigrants) and his family spoke only Norwegian at home (Hjelde 2012, p. 183). This does
not signal a community of a monolingual Norwegian-speaking majority—this preference
for Norwegian applies to a proportion of the immigrants only. Similar cases are attested in
the Midwest for German.
7
The normal speaker in these settlements, like heritage speakers
around the world (cf., e.g., Rothman 2009;Benmamoun et al. 2013) would acquire a minority
language (Norwegian) as their first language in the family home, and proceed to learn the
dominant language of the larger society (English).
Though the Norwegian settlers could dominate an area to the extent of active exclusion
of other nationalities, they were not in a position to isolate themselves totally and avoid
contact with the majority of the English language. Connections between local and central
authorities had to be maintained (“vertical ties”; cf. Salmons 2005a,2005b), and for many
Norwegian immigrants, their sudden and sometimes reluctant immersion into English
happened when they started school at six. In most cases, this encounter did not lead to a
rapid language shift to English within the community; instead, most speakers ended up as
balanced (or unbalanced) bilinguals with English as their dominant language.
The impact of this bilingual state on NorNAHL is evidenced in extensive lexical
borrowing of English words, subject to early linguistic investigations (cf. above). Here is
Haugen’s (1938, p. 39) description:
The shell is still Norwegian, but the inward pattern, the spirit of the thing,
is American. This is Norwegian-American, the language of the Norwegian
immigrant.
The impact from English, omnipresent in the greater society, constitutes one of the two
important axes of crosslinguistic influence (CLI) relevant to the context of NorNAHL in the
American Midwest. The second axis is the influence from other Norwegian dialects and
standards in the immigrant’s daily input. Certain Norwegian settlements were somewhat
heterogeneous at the outset, with immigrants from different parts of Norway, but many
were linguistically rather homogeneous communities, each characterized by the dominance
of a single Norwegian dialect, usually determined by the origin of the majority of the
immigrants. The relevant communities in Haugen’s study mostly belong to the latter;
with one or two dominating dialects, with others also present in the community. In
communities with one type of prevailing Norwegian input, maintaining distinctive features
of the original Norwegian dialect may seem easier. However, even in this context the key
institutions (churches, newspapers), together with letters from home and other written
sources, provided a rich environment for input in Norwegian from various varieties in
different proportions. For a typical Norwegian immigrant in these communities at this time,
different dialects, standards and varieties would be present in these input sources, although
people would avail themselves to this input to different degrees (cf. below for details).
2.2. Selecting the Communities
Our study includes data from three different locations in southern and western Wis-
consin and eastern Minnesota, namely the settlement encircling the two towns Westby and
Coon Valley in Vernon County, Wisconsin; the small-town Blair, Trempealeau County, also
Wisconsin; and the settlement around the two small neighbor towns Wanamingo and Zum-
brota in Goodhue County, Minnesota (cf. Figure 1. Although Coon Valley/Westby were
dominated by immigrants from the area around Mjøsa (especially Gudbrandsdalen), the
immigrants in Blair more often originated from Solør, near the Swedish border (cf. Figure 2).
In Wanamingo/Zumbrota many trace their roots to Stjørdalen, northeast of Trondheim,
Languages 2023,8, 49 6 of 36
and this settlement served as a major abutment for other Stjørdalen settlements further
west.
1
Figure 1. Map with the location of Wanamingo, Zumbrota, Blair, Westby and Coon Valley.
Languages 2023, 8, x FOR PEER REVIEW 7 of 38
Figure 2. Map indicating the localization of some of the Norwegian places mentioned in this study:
1 Gudbrandsdalen, 2 Stjørdal (Trøndelag), 3 Hallingdal, 4 Hardanger, 5 Numedal, 6 Sogn, 7 Solør,
8 Telemark, 9 Valdres, 10 Voss.
The area around Coon Valley and Westby seemingly maintained NorNAHL for the
longest time as a widespread home language as well as in official settings. Here, Nor-
NAHL was the dominating church language until the Second World War, gradually re-
placed by English, with a few occasional services in Norwegian even in the 1990s. Norwe-
gian was the only foreign language offered at the local high school until the late 1980s,
and the strong position of NorNAHL is still evidenced in the continued use of Norwegian,
in the names of Norwegian congregations and churches (Haugen 1953, pp. 65–67) but also
in place names. Coon Valley—perhaps helped by its more Norway-like landscape, dis-
plays a plethora of Norwegian names, such as Øyumsdalen, Sørryggen, Bredebygda, Skogda-
len and Indihaugen (Kruse 1991).
3. Main Sources of Norwegian Input
In these settlements, constituting a rather long-lasting Norwegian diaspora, the num-
ber and presence of societal institutions would provide a heritage speaker with a variety
of potential sources for Norwegian input. Although, as is the typical situation for a herit-
age speaker who (by definition, cf. above) acquires the heritage language in the family
home, the family’s spoken language and hence the local dialect is always fundamental,
while a range of other spoken and written varieties would be available as relevant and
regular Norwegian input to a heritage speaker in these communities, although to differing
degrees throughout the history of this contact situation.
3.1. The Local Spoken Dialect(s): Family, Neighbors and Settlement
The linguistic situation in the family home is evidently key to almost any heritage
language acquisition. The communities relevant to this study were characterized by a
strong tendency among the Norwegian immigrants toward endogamy. A total of 75% of
Figure 2.
Map indicating the localization of some of the Norwegian places mentioned in this study:
1 Gudbrandsdalen, 2 Stjørdal (Trøndelag), 3 Hallingdal, 4 Hardanger, 5 Numedal, 6 Sogn, 7 Solør,
8 Telemark, 9 Valdres, 10 Voss.
Languages 2023,8, 49 7 of 36
These are all rural, traditional farming communities on the Mississippi river near
Midwestern towns, established around 1850 and subject to the chain immigration and
clustering tendencies discussed above. They belong to a belt of old Norwegian settlements
in counties where the density of Norwegian descendants is among the highest in the US. In
the US 2000 census, in Wanamingo 38% claimed a Norwegian descent, in Vernon County
(Westby/Coon Valley) 36%. A total of 3–5% of the population in these areas report in 2000
that they still speak Norwegian at home, “sometimes”. As witnessed by Einar Haugen who
visited in the early 1940s, Norwegian was widely spoken even by the younger generations.
The area around Coon Valley and Westby seemingly maintained NorNAHL for the
longest time as a widespread home language as well as in official settings. Here, NorNAHL
was the dominating church language until the Second World War, gradually replaced by
English, with a few occasional services in Norwegian even in the 1990s. Norwegian was
the only foreign language offered at the local high school until the late 1980s, and the strong
position of NorNAHL is still evidenced in the continued use of Norwegian, in the names of
Norwegian congregations and churches (Haugen 1953, pp. 65–67) but also in place names.
Coon Valley—perhaps helped by its more Norway-like landscape, displays a plethora of
Norwegian names, such as Øyumsdalen,Sørryggen, Bredebygda, Skogdalen and Indihaugen
(Kruse 1991).
3. Main Sources of Norwegian Input
In these settlements, constituting a rather long-lasting Norwegian diaspora, the num-
ber and presence of societal institutions would provide a heritage speaker with a variety of
potential sources for Norwegian input. Although, as is the typical situation for a heritage
speaker who (by definition, cf. above) acquires the heritage language in the family home,
the family’s spoken language and hence the local dialect is always fundamental, while a
range of other spoken and written varieties would be available as relevant and regular
Norwegian input to a heritage speaker in these communities, although to differing degrees
throughout the history of this contact situation.
3.1. The Local Spoken Dialect(s): Family, Neighbors and Settlement
The linguistic situation in the family home is evidently key to almost any heritage
language acquisition. The communities relevant to this study were characterized by a
strong tendency among the Norwegian immigrants toward endogamy. A total of 75%
of all Norwegian-Americans in the Coon Valley and Westby area reported to have single
Norwegian ancestry (1990 US census). Ethnicity played a significant role in finding a spouse,
enforced by religion, as most Norwegian-Americans were Lutherans. Unlike Swedish
immigrants, who spread across different denominations, Norwegian immigrants remained
strictly Lutheran, to the extent that Lutheranism became an expression of Norwegian
ethnicity in the Midwest (Joranger 2010;Aschim 2019). Aschim further notes that “the
correct doctrinal Lutheranism was a matter of contention” and “by 1876, Norwegian
immigrants could choose between five different competing Norwegian-language Lutheran
church bodies” (cf. also Granquist 2015, p. 186). Marrying a fellow Norwegian often did
not suffice, as there was pressure to marry within the same church, and according to one of
our informants (born in 1929) “dating a boy from a different church would be considered
a sin”. Marrying an Irishman or a German, the Catholic neighbors, would definitely be
frowned upon.
Even if the importance of ethnic background in marriage faded over the years, the
strict preference throughout several generations toward endogamy and the resulting all-
Norwegian household clearly facilitated Norwegian heritage language transmission. For
the early generations in these communities, the typical Norwegian family would work
together at the farm, with parents and older siblings who were bilingual, but also including
grandparents who were unable or reluctant to speak English. Hence, the family would
speak Norwegian in the home, providing a young NorNAHL learner with ample spoken
input daily. Most children were Norwegian monolinguals before starting school, enforcing
Languages 2023,8, 49 8 of 36
an abrupt language-shift to English which could be quite traumatic for many six-year-olds,
as speaking Norwegian at school may elicit disciplinary measures. Certain informants
in later generations told similar stories, for instance “Lars”, born in 1961, by his own
admission spoke hardly any English before he started school in 1967. Unlike earlier
generations, however, by then Lars’s lack of English skills made him a strange animal,
evoking ridicule from his classmates (Hjelde et al. 2019, p. 189).
Although the chain immigration tendency caused many dialectically rather homoge-
neous settlements in the Midwest, the dialectal input in the community could still be varied.
Homeland Norwegian has great dialectal diversity, but prior to the mass emigration, most
people in Norway had little exposure to dialects other than their own, tending to live their
lives in the area where they were born. Haugen (1953, p. 341) notes that for many early
Norwegian immigrants, emigration caused their first exposure to dialectal variation. His
study on settlers in Koshkonong, one of the first successful Norwegian settlements, lists
the largest group as originating from Upper Telemark, the second largest from Sogn, but
also many from Voss, Lower Telemark and Numedal, Hallingdal, Valdres, Hardanger and
Eastern Norway (cf. Figure 2). Even if these different groups tried to cluster and form
groupings on their own, “any kind of strict segregation was out of the question” (op.cit.).
The communities in our study were each characterized by the prevalence of one
dominating dialect, but with other dialects represented. Coon Valley and Westby had the
Gudbrandsdalen dialect as its main variety, with some competition from the Telemark
dialect. There was even a substantial group of so-called Flekkefjordings, and a large
settlement of people from Sogn a bit further south. In Wanamingo and Zumbrota, people
spoke the Stjørdalen dialect, but had other dialects around the neighborhood, evidenced
by place names such as Toten, Eidsvold, Dovre, Sogn, Henning and Vang (Valdres). Blair,
though dominated by immigrants from Solør, also had several variants, as “(m)any different
dialect areas are found within this region” (Haugen 1953, p. 612).
Several NorNAHL settlements over time developed a regional spoken variety, a koiné;
the Coon Valley and Westby community serving as a case in point (Hjelde 2012).
8
No
koineization can take place unless the different dialects forming the input of the process
are regularly used within the community, which entails that the NorNAHL dialects are
not confined to being strict home languages (op.cit). Haugen’s recordings in Coon Valley
provide an interesting testimony to the massive presence of Norwegian dialects within the
community from a man of German heritage background (born in 1894) recorded speaking
the local Norwegian dialect of the community in a native-like manner. He explains that
since all of his playmates spoke Norwegian, he had no choice but to acquire Norwegian
himself, as his third language. This further suggests that English was not voluntarily
spoken at any range among his Norwegian friends. The recordings also contain narratives
from bidialectal children growing up with one dialect at home and another one spoken
by their peers. The daily walk to school was spent becoming mentally and linguistically
prepared to speak the community dialectal variant, and the walk back home to switch back
to thinking and speaking in the dialect of the family home.
After a koinéhad formed in the community, the earlier dialectal variation in the
community was reduced. At later stages, as the community shifted their main language
from the heritage language to English, the family was left as the last stronghold where
Norwegian was, and still is, spoken at some rate. Eventually (especially after WWII),
Norwegian was reduced to being a home language, and from then on, the heritage speakers
would hardly be exposed to any varieties different from their own.
3.2. Other Spoken Input: Church and Confirmation
At one point there were more than seven thousand Norwegian-American churches and
congregations according to Haugen (1939, pp. 65–67). The Norwegian-American Lutheran
church was probably the most important institution for maintaining NorNAHL in the
immigrant community and constituted a pivotal arena for many Norwegian immigrant
families, of course for matters concerning their faith, but clearly also socially. Not only
Languages 2023,8, 49 9 of 36
did the immigrants convene and talk before and after service, the church hosted a range
of social gatherings and groups, such as Ladies Aid, Bible study groups and parochial
schools. Already in 1844 Norwegian settlements in Wisconsin wrote to their mother church
in Norway asking for a Norwegian minister to be sent to the Midwest to work among the
immigrants (Blegen [1940] 1969, p. 141). This would provide another source of spoken
Norwegian input, as these pastors were university-educated and would speak a version of
“educated Norwegian”. Maintaining NorNAHL in the liturgy and including the Norwegian
Lutheran hymns (from Landstads salmebok) as unequivocal parts of the service retained
Norwegian as the lingua sacra for early Norwegian immigrants. Young Lutherans attended
confirmation lessons, which, like the official confirmation ceremony (a public profession of
faith and a rite in which adolescents become full members of the church and adult members
of the general society) was offered in Norwegian in many congregations until the 1920s;
cf. Gilbertson (2009, p. 5).
If the person was confirmed in the Norwegian language, she had to read and
memorize entire religious texts. Therefore, she would be able to easily recite
[
. . .
] Bible verses, passages from sermons, and other religious phrases. The
Norwegian language persisted longest when it came to matters of faith. Even if
a person could attend school, conduct business, or read newspapers in English,
their entire religious vocabulary was in Norwegian.
Norwegian remained the primary lingua sacra throughout the 1800s and early 1900s, until
the dawn of 1917, which featured two separate developments that taken together would
affect the status of the Norwegian church language to a devastating extent (Haugen 1953,
p. 255). In the early years, a series of controversies led to the Norwegian-American Lutheran
church splitting into several competing synods. In 1917, however, the majority of these
synods remerged into one church fostering an ambition not to confine itself to being a
church for Norwegian-Americans only, but to become a Pan-American Lutheran church.
This of course motivated a switch to English services.
This development within the church co-occurred with the US participation in World
War I, causing a wave of general xenophobia accompanied by a self-imposed urge to
downplay the heritage background (Lovoll 2006, p. 227). One extreme example of this
trend was the so-called Babel proclamation in Iowa, issued by governor Harding, where the
use of any other language than English in public was banned. This ban included schooling
and religious gatherings, as well as private telephone conversations (as the telephone line
was public). Even outside of such extreme precautions, the young Norwegian-American
bilingual generation gradually identified more with English than with their (old-fashioned)
heritage language and culture. These developments yielded an incremental shift to English
church services, first offering English services alongside Norwegian ones, before eventually
fading out Norwegian as the lingua sacra. The process was completed in a couple of
decades; shortly after WWI, about two-thirds of the services were given in Norwegian; two
decades later, as the US entered WW2, about one-tenth; cf. the graph in Figure 3.
Languages 2023,8, 49 10 of 36
Languages 2023, 8, x FOR PEER REVIEW 10 of 38
schooling and religious gatherings, as well as private telephone conversations (as the tel-
ephone line was public). Even outside of such extreme precautions, the young Norwe-
gian-American bilingual generation gradually identified more with English than with
their (old-fashioned) heritage language and culture. These developments yielded an in-
cremental shift to English church services, first offering English services alongside Nor-
wegian ones, before eventually fading out Norwegian as the lingua sacra. The process was
completed in a couple of decades; shortly after WWI, about two-thirds of the services were
given in Norwegian; two decades later, as the US entered WW2, about one-tenth; cf. the
graph in Figure 3.
Figure 3. Percentage of Norwegian services; Norwegian Lutheran church (based on Haugen 1953,
pp. 262–7).
3.3. The Written (Standardized) Language: Lingering Conservatism
In the early days of Norwegian emigration, public education in Norway was rather
sparse. Obligatory lessons were limited to religion and some training in reading, with the
main purpose to prepare children for confirmation. Through new legislation in 1860, the
Norwegian school system was thoroughly reformed and the curriculum was extended
with several topics and skills, including writing. Thus, the immigrants who arrived in the
US during the last decades of the 1800s were fairly well acquainted with the written lan-
guage. Lovoll (1983, p. 136) describes the Norwegian-Americans at the end of the century
as “a people of readers” (et lesende folk). In 1838, the first Norwegians settled in Wiscon-
sin, and their first newspaper in Norwegian, Nordlyset, was first printed in Muskego in
1847. Over the next decades, several hundreds of NorNAHL publications were printed,
ranging from religious texts and political pamphlets via newspapers and cook books to
novels and biographies. Immigrants also wrote significant amounts of letters to their rel-
atives who had remained in the homeland. These “America-letters” were clearly heavily
influenced by religious writings in the linguistic style chosen, turns of phrases and in other
modes of expression (Gulliksen 2002; Mathiesen 2014).
Most writings by the Norwegian Americans were characterized by a conservative
style, adhering to old orthography and word form standards, which many of the emi-
grants were familiar with prior to their emigration, and perhaps supported by the turbu-
lent language situation in the homeland around 1900. Due to the four-hundred-year-long
union with Denmark (dissolved in 1814), Danish was the written standard in Norway
during most of the 19th century. Around 1850, a new Norwegian standard was developed
(today’s Nynorsk), eventually entering into competition with a norwegianized version of
Danish spelling (today’s Bokmål); a competition which still remains (cf. Haugen 1966).
During the first decades of the 1900s, both standards Nynorsk and Bokmål went through
a series of reforms, triggering wide-spread skepticism from the Norwegian American
community, who responded by blatantly ignoring most of the “unnecessary” homeland
Figure 3.
Percentage of Norwegian services; Norwegian Lutheran church (based on Haugen 1953,
pp. 262–67).
3.3. The Written (Standardized) Language: Lingering Conservatism
In the early days of Norwegian emigration, public education in Norway was rather
sparse. Obligatory lessons were limited to religion and some training in reading, with the
main purpose to prepare children for confirmation. Through new legislation in 1860, the
Norwegian school system was thoroughly reformed and the curriculum was extended with
several topics and skills, including writing. Thus, the immigrants who arrived in the US
during the last decades of the 1800s were fairly well acquainted with the written language.
Lovoll (1983, p. 136) describes the Norwegian-Americans at the end of the century as “a
people of readers” (et lesende folk). In 1838, the first Norwegians settled in Wisconsin,
and their first newspaper in Norwegian, Nordlyset, was first printed in Muskego in 1847.
Over the next decades, several hundreds of NorNAHL publications were printed, ranging
from religious texts and political pamphlets via newspapers and cook books to novels and
biographies. Immigrants also wrote significant amounts of letters to their relatives who
had remained in the homeland. These “America-letters” were clearly heavily influenced
by religious writings in the linguistic style chosen, turns of phrases and in other modes of
expression (Gulliksen 2002;Mathiesen 2014).
Most writings by the Norwegian Americans were characterized by a conservative
style, adhering to old orthography and word form standards, which many of the emigrants
were familiar with prior to their emigration, and perhaps supported by the turbulent
language situation in the homeland around 1900. Due to the four-hundred-year-long union
with Denmark (dissolved in 1814), Danish was the written standard in Norway during
most of the 19th century. Around 1850, a new Norwegian standard was developed (today’s
Nynorsk), eventually entering into competition with a norwegianized version of Danish
spelling (today’s Bokmål); a competition which still remains (cf. Haugen 1966). During
the first decades of the 1900s, both standards Nynorsk and Bokmål went through a series
of reforms, triggering wide-spread skepticism from the Norwegian American community,
who responded by blatantly ignoring most of the “unnecessary” homeland reforms, a
protest evident in the secular press. The writing traditions in the immigrant community
and homeland Norway thus parted ways, and written Norwegian in the US became distinct
from the spoken language, much more so than in Norway.
The Bible and The Small Catechism in Norwegian were printed in Gothic letters (fraktur
fonts) and children of the Norwegian immigrants attending parochial schools had to read
these religious texts to get through their confirmation rites. However, these two books
featured Gothic letters in Norway too, long after 1900 (Danbolt 2021). For other types of
written publications though, the Norwegian-Americans kept the gothic print for many
decades after Latin print (antigua) had become predominant in Norway.
Languages 2023,8, 49 11 of 36
One illustrative example is the prominent NorNAHL newspaper, Decorah-Posten,
where Gothic print remained until 1952, half a century after most homeland Norwegian
newspapers. Capitalized nouns appeared in Decorah-Posten even in 1961, 60 years after
Norway, and more generally, this newspaper adopted the standard of the Norwegian 1907
spelling reform only in the late 1930s. Erling Innvik, the later editor for Decorah-Posten,
explained that immigrants are traditionally very conservative in their culture preferences,
and this nostalgic conservatism, in his view, was the reason why Decorah-Posten survived its
NorNAHL competitors. One side-effect to this conservatism of course is that the newspaper
ends up being “a newspaper written by old people for old people” (Lovoll 1977, p. 98). This
being said, the readers of Decorah-Posten would be regularly exposed to texts deviating from
this conservative and obsolete standard. We find examples of short texts in written dialects
and Nynorsk, and the most popular content in Decorah-Posten was the comic strip Han Ola
og han Per, where the Midwestern characters speak several different dialects and display
bilingual varieties characteristic of NorNAHL. Some NorNAHL publications, especially
those associated with the Norwegian-American church, followed homeland linguistic
reforms. Nynorsk was also used, but to a much more limited degree than Bokmål.9
3.4. Sources of Input for Written Norwegian: Language Learning and Maintenance
During our visits in contemporary Norwegian-American homes, the host would often
show us the household’s collection of old Norwegian books, suggesting that Lovoll’s (1983)
observation of Norwegian immigrants as “a people of readers” has merit. The collections
range from The Bible and the occasional hymn book to comprehensive book collections,
religious pamphlets and manuals or novels usually from around 1900.
As detailed above, an eager reader of American-Norwegian publications during the
first decades of the 1900s would be exposed to a number of different standards of writing,
though the dominant one would be a very conservative Dano-Norwegian standard in
Gothic print. The Bible,The Small Catechism and other religious texts and pamphlets would
be important sources for input of written Norwegian, and the church as a religious and
social arena would be a constant motivation to learn how to read Norwegian, acting as an
important source of many different varieties of spoken Norwegian (cf. Section 3.2 above).
Equally important to NorNAHL maintenance is the church’s role in providing formal
education for the children in written Norwegian. The organization of this education was
continuously under harsh debate in the Norwegian-American communities. Due to the
lack of proper Lutheran teachings in American common schools, several congregations
wanted—and some tried—to establish full-fledged Norwegian-American schools to re-
place or at least compete with the local American common schools. Running a full-time
Norwegian-American Lutheran school proved too costly for most congregations, and a
compromise was found where the children were enrolled in and attended the secular
common school (often referred to as the “Yankee school” or “English school”), whereas
a Norwegian parochial school was established and offered as a supplement (often called
“Summer school”, “Norwegian school” or “Religion school”). Besides being educated in the
Lutheran faith, the children were also trained in reading and comprehending standardized
Norwegian (i.e., Dano-Norwegian). Even in these schools, NorNAHL came under pressure
in the aftermath of WWI, and we see a rapid shift from Norwegian to English. In 1918,
more than 80% of the parochial school instruction was given in Norwegian. Only ten years
later, attesting to a rapid and rather dramatic language shift, English ruled the ground
(cf. Figure 4).
Languages 2023,8, 49 12 of 36
Languages 2023, 8, x FOR PEER REVIEW 12 of 38
1918, more than 80% of the parochial school instruction was given in Norwegian. Only
ten years later, attesting to a rapid and rather dramatic language shift, English ruled the
ground (cf. Figure 4).
Figure 4. Percentage of the Norwegian Lutheran church’s parochial schools given in Norwegian
1917–1928 (after Haugen 1953, p. 262).
There is ample evidence that the ability of young Norwegian-Americans to read Nor-
wegian underwent a rapid decline during the first decades of the 1900s. Hjelde and Bjørke
(2022) compare two dictionaries compiled for Nordahl Rolfsen’s reader, published in 1909
and 1930, respectively.
10
The 1909 edition comments on special, uncommon words—
words beyond the everyday vernacular, or with a spelling pattern unfamiliar to a reader
with a predominant English literacy, and the words are explained in both Norwegian and
English. The 1930 edition glosses and explains all words in the reader and only offers a
translation into English, no explanations in Norwegian. The conclusion drawn is that the
1909 edition addresses an audience who is proficient and partly literate in Norwegian,
and the 1930 edition addresses an audience who is neither.
Parochial schools were never mandatory, and only a fraction of Norwegian-Ameri-
cans born one hundred years ago had a formal education with the skill to read Norwegian.
Importantly, however, almost none of their successors born around 1940 were given any
such education.
The impact of this change on the reading ability of the entire Norwegian-American
community and the position of the Norwegian written language in the immigrant com-
munity at large is evidenced by the graph in Figure 5, displaying the number of Norwe-
gian newspapers issued throughout the 20th century.
Figure 4.
Percentage of the Norwegian Lutheran church’s parochial schools given in Norwegian
1917–1928 (after Haugen 1953, p. 262).
There is ample evidence that the ability of young Norwegian-Americans to read
Norwegian underwent a rapid decline during the first decades of the 1900s. Hjelde and
Bjørke (2022) compare two dictionaries compiled for Nordahl Rolfsen’s reader, published
in 1909 and 1930, respectively.
10
The 1909 edition comments on special, uncommon words—
words beyond the everyday vernacular, or with a spelling pattern unfamiliar to a reader
with a predominant English literacy, and the words are explained in both Norwegian and
English. The 1930 edition glosses and explains all words in the reader and only offers a
translation into English, no explanations in Norwegian. The conclusion drawn is that the
1909 edition addresses an audience who is proficient and partly literate in Norwegian, and
the 1930 edition addresses an audience who is neither.
Parochial schools were never mandatory, and only a fraction of Norwegian-Americans
born one hundred years ago had a formal education with the skill to read Norwegian.
Importantly, however, almost none of their successors born around 1940 were given any
such education.
The impact of this change on the reading ability of the entire Norwegian-American
community and the position of the Norwegian written language in the immigrant commu-
nity at large is evidenced by the graph in Figure 5, displaying the number of Norwegian
newspapers issued throughout the 20th century.
Languages 2023, 8, x FOR PEER REVIEW 12 of 38
1918, more than 80% of the parochial school instruction was given in Norwegian. Only
ten years later, attesting to a rapid and rather dramatic language shift, English ruled the
ground (cf. Figure 4).
Figure 4. Percentage of the Norwegian Lutheran church’s parochial schools given in Norwegian
1917–1928 (after Haugen 1953, p. 262).
There is ample evidence that the ability of young Norwegian-Americans to read Nor-
wegian underwent a rapid decline during the first decades of the 1900s. Hjelde and Bjørke
(2022) compare two dictionaries compiled for Nordahl Rolfsen’s reader, published in 1909
and 1930, respectively.
10
The 1909 edition comments on special, uncommon words—
words beyond the everyday vernacular, or with a spelling pattern unfamiliar to a reader
with a predominant English literacy, and the words are explained in both Norwegian and
English. The 1930 edition glosses and explains all words in the reader and only offers a
translation into English, no explanations in Norwegian. The conclusion drawn is that the
1909 edition addresses an audience who is proficient and partly literate in Norwegian,
and the 1930 edition addresses an audience who is neither.
Parochial schools were never mandatory, and only a fraction of Norwegian-Ameri-
cans born one hundred years ago had a formal education with the skill to read Norwegian.
Importantly, however, almost none of their successors born around 1940 were given any
such education.
The impact of this change on the reading ability of the entire Norwegian-American
community and the position of the Norwegian written language in the immigrant com-
munity at large is evidenced by the graph in Figure 5, displaying the number of Norwe-
gian newspapers issued throughout the 20th century.
Figure 5.
Number of secular Norwegian-American newspapers issued every year, in the period of
1900–1990 (based on numbers from Lovoll 2010).
Languages 2023,8, 49 13 of 36
After the modest first attempts, the last decades of the 1800s witnessed the arrival
of groups of well-educated immigrants, gradually turning these papers into important
tools to form a Norwegian-American identity. Norwegian newspapers likewise provided
helpful aid for newcomers struggling to cope with a new way of living, gaining many
immigrant readers. An estimated 400 different Norwegian-American newspapers were
established, many rather short-lived—about one third existed less than a year. Others
gained an impressive impact, especially the so-called big three, Skandinaven (Chicago,
1866–1941), Minneapolis Tidende (Minneapolis, 1888–1935) and Decorah-Posten (Decorah,
1874–1972). In the early 1900s, Skandinaven had a larger circulation than any newspaper
in the homeland Norway.11 However, soon the number of Norwegian ethnic publications
plunged, from close to sixty in 1900, down to about ten in the early 1940s. From then on,
the number of newspapers decreased, but at a slower rate.
The “English only” ideology gained dominance once the US entered WWI, urging
immigrants to become loyal, monolingual Americans. Additionally, the number of US-born
Norwegian-Americans was steadily growing, surpassing the number of first-generation
immigrants. This process accelerated during the 1920s as immigration restrictions were
imposed, reducing the number of “newcomers”. The American public school made English
the dominant written language for US-born Norwegian-Americans.
For the early emigrants, leaving Norway mostly meant never seeing their homeland
again, and the only way of keeping in contact with the family back home was through
letters. Thus, writing skills in Norwegian were key and presumably served as a motive
for educating subsequent generations in writing Norwegian, long after the decline of
Norwegian as a language for church and public press. The existing letter collections reveal
that many American-Norwegians were rather able in writing their heritage language.
These letters have been subject to few systematic studies, but we know, also from our
interviews with informants over four decades, that such letters continued to be sent for
several generations, down to our times. Additionally, visiting Norway was a wide-spread
life-long dream for many. As flights became affordable, many fulfilled this goal, though
family ties faded over the years and the goal of meeting relatives was gradually replaced
by revisiting the roots, and seeing the country of their ancestors. While Norwegian may be
useful when communicating with older Norwegians, Norwegians today are rather fluent
in English, so this exposure to contemporary homeland Norwegian is very much reduced.
One may have expected the Norwegian-Americans to take advantage of the opportu-
nities provided by modern technology, but this has hardly been the case. It is reported that
the theatre in Westby at rare occasions offered Norwegian movies with English subtitles up
to the early 1960s. Internet potentially represents an opportunity to keep in contact with
Norway and its language, but hardly any of the contemporary speakers access Norwegian
media, written or spoken, via the Internet, presumably due to their age group which
typically possesses a limited digital literacy.
The immigrants also had their own literature, though no Norwegian-American authors
could make a living from writing. After 1920, the NorNAHL audience was too small to
guarantee a profit for any publisher of Norwegian literature, which made it difficult to
publish in Norwegian in the US. Still, this “immigrant literature” played an important role
in the community, helping individuals to interpret their role as an immigrant. Few authors,
with the exception of Ole E. Rølvaag, are remembered and read today, and Rølvaag’s
tetralogy I de dage (Giants in the Earth) from the 1920s figures among the last literature
published in Norwegian in America. The more recent Norwegian-American authors were
published for Norwegian readers in Norway, not in the US.
4. The Data
4.1. Recordings and Idealized Cohorts
Featuring a high density of Heritage Norwegian speakers (cf. Section 2.2), the commu-
nities in this study have been subject to numerous studies, providing us with a wide range
of data over an eighty-year time span, from the early 1940s and up to recent years. We
Languages 2023,8, 49 14 of 36
have Einar Haugen’s recordings and field notes from Westby and Coon Valley (1942/1948);
Arnstein Hjelde’s recordings from Wanamingo, Zumbrota, Coon Valley and Westby (1987
and 1992); and CANS recordings from all these areas made during the 2010s.
Our data are not confined to the transcribed recordings available in CANS; we have
also included recordings and field notes that are not yet a part of the searchable online
corpus (cf. Table 1for an overview of the data in our study). Based on these sets of
recordings, we idealized five different age groups or cohorts: those born around 1870, those
born in the 1900s, 1920s, 1940s and those born after 1950. We will mainly focus on the last
four cohorts, i.e., speakers born after 1900, since this is where we expect to find the most
interesting changes. The presumed and attested absence of linguistic changes is also the
reason why the first cohort covers a time span of 30 years, whereas the four latter cohorts
are a bit more fine-grained; cf. Table 2for details.
In addition to these cohorts, which are “abstracted generations” of varying time spans,
we have identified and included in this study actual pairs of parent and child. Such pairs
provide more direct insight into the language transmission between actual generations.
Table 3gives an overview over the parent–child pairs in this study. Informants are placed in
the various cohorts based on their birth year, and in what follows, we use the information
on available sources of Norwegian input discussed in Section 3and discuss the relation
between this and the observed language production of these different cohorts.
Our many interviews with American-Norwegian informants during the last four
decades (with an overweight of interviews conducted during the 1980s and 1990s) allow
us to contextualize these cohorts further with respect to the available input. The typical
informant belonging to cohort II (and the older segment of cohort III) would remember a
childhood with Norwegian newspapers, although they usually did not read these them-
selves. The exception was Han Ola og han Per, the comic appearing in Decorah-Posten
for decades. Many of these older speakers had attended Norwegian parochial summer
school and were confirmed in Norwegian, thus they also had some familiarity with the
written language. This amount and range of input would be much narrower for most of
the younger speakers, those belonging to cohorts IV and later.
Table 1. The data sets used in this study.
Location/Source Blair Coon Valley/Westby Wanamingo/Zumbrota Hours Recorded
1942: Haugen X X 8
1930–40s: Haugen: CANS X 3
1987: Hjelde X 6
1992: Hjelde X 75
2010: Eide/Hjelde
(NorAmDiaSyn) X X 10
2010–2014: CANS X X 12
2015–2018: Hjelde
(NorAmDiaSyn) X1512
Field notes:
Haugen, Hjelde X X X
Table 2. Generational cohorts and data.
Cohort I II III IV V
Born Around 1870 1900–1920 1920–1930 1940–1950 after 1950
Data sets Haugen 1940s Haugen 1940s Hjelde 1980s and 1990s CANS Eide/Hjelde 2010 CANS Hjelde 2010–2018
Languages 2023,8, 49 15 of 36
Table 3.
Parent and child pairs placed according to cohort (Data: CANS/Hjelde). (Wa = Wanamingo;
Z = Zumbrota; CV = Coon Valley; W = Westby).
Cohort I II III IV V
Born Around 1870 1900–1920 1920–1930 1940–1950 after 1950
Parent and child Thor (Wa) Iris (Z)
Mason (CV) Liam (CV)
Ted (CV) Beth (W)
Laura (W) Lars (W)
The information gained through these interviews, as well as the graphs and the
discussions in Section 3on available sources for different varieties of Norwegian, leads
us to assume that there will be a noticeable change between cohorts II and IV as regards
acquired productive linguistic distinctions in Heritage Norwegian. If the quantity and
quality of input is visible in the production of heritage speakers, this is where we ought
to see the biggest changes. Our expectations are based on the fact that the bulk of natural
language acquisition usually happens when the learner is between 0 and 12 years, and
many of the input sources start to recede around 1920. This is when cohort III children
are born.
4.2. Selecting Linguistic Features for the Study
As is well-attested, “[i]f two or more languages are in contact, with speakers of one
language having some knowledge of the other, they come to borrow, or copy [
. . .
] linguistic
features and forms of all kinds” (Aikhenvald 2006, p. 2). This phenomenon, more recently
referred to as crosslinguistic influence, trivially applies to NorNAHL, a minority language
in a continuous contact situation. However, the impact from English, omnipresent in
the greater community and society, constitutes only one of the two important sources
of crosslinguistic influence (CLI) relevant in the context of NorNAHL in the American
Midwest. The second source is the influence from other Norwegian dialects and standards
present in the immigrant’s daily input.
Language contact research has sought to identify linguistic domains (universally)
more and less susceptible to contact-induced linguistic change across various types of
language contact and particular languages, and although there is little consensus also
in this field, the disintegration and simplification of morphological systems triggered by
language contact is widely attested. Correspondingly, Benmamoun et al. (2013, p. 141)
identify inflectional morphology as “particularly vulnerable in heritage languages”. On the
other hand, word order phenomena are usually seen as rather robust in language contact
(e.g., Winford 2003, p. 97); this has been attested for heritage languages in general and for
NorNAHL specifically. Johannessen and Laake (2015) find that many syntactic features
remain rather unchanged from the earliest Heritage Norwegian up to the present day.
However, although the morphosyntactic features and operations may well still exist in
the heritage language, the frequency with which certain constructions and features occur
gradually shifts. In this study, we scrutinize such ratios within one morphological domain,
tense affixes, and in two syntactic domains, complex noun phrases and topicalized verb
second declaratives.
The selected morphosyntactic features exhibit productive distinctions subject to vari-
ation and change in NorNAHL throughout the 80-year time span, and these changes in
the distributional pattern reflect the reduction in types of input available over time. Firstly,
the morphological paradigms we address in Section 4.3 are different from one variant to
the next of spoken and written NorNAHL available in the community. The internal word
order of complex noun phrases, specifically, in prenominal and postnominal possessives,
distribute differently across spoken variants and across spoken and written input, and it
is intriguing to track the potential impact of the written input in the spoken production
especially (Section 4.4). Finally, the topicalization of non-subjects and the corresponding
Languages 2023,8, 49 16 of 36
subject–verb inversion (i.e., verb second) addressed in Section 4.5 are features where Nor-
wegian and English are very different. Firstly, English is not a V2 language, and secondly,
topicalization is rather frequent in Norwegian and rather rare in English. Roughly 35% of
Norwegian main clauses feature topicalization, compared to about 7–10% of corresponding
English structures.
4.3. Morphological Paradigms: Productive Tense Suffixes of Loan Verbs
Eide and Hjelde (2015a,2018) report on the tense paradigms in NorNAHL and focus on
loan verbs, as these verbs, being newly introduced to the Norwegian varieties, necessarily
inflect according to the (most) productive morphological paradigm.
13
When Haugen (1953)
discusses verb paradigms of borrowed verbs (exemplified here by the borrowed verb claim,
adapted as kleima in NorNAHL), the paradigm from standard Nynorsk is the baseline.
In line with the analysis of Norwegian (and other Germanic) tense suffixes in Eide (2002,
2005,2009a,2009b,2016,2019 and subsequent work) and Eide and Hjelde (2015a,2018), we
enter these verb forms into a table featuring matrices of
±
past and
±
finite. According to
Haugen’s discussion of verb forms, the distinction
±
finite is retained in the non-past forms
(infinitive kleima versus present kleimar), but absent from the +past forms: the preterit kleima
shares the form with the past participle (cf. Table 4a). Haugen’s paradigm for Nynorsk
shows syncretism between all forms except the present. We also show the contemporary
bokmål paradigm (1907–standard), with distinct forms for infinitive and present, but non-
distinct forms for the +past cells; cf. Table 4b. This latter standard was prevalent in the
written input of NorNAHL learners at the time, and the same system is documented in
Haugen’s transcriptions of the Westby minister’s “educated speech”.14
Haugen’s paradigm in (Table 4a) is clearly an idealization; thus; we revisited Haugen’s
recordings to attest for the paradigms for the spoken varieties. Haugen’s informants were
born between 1849 and 1927, thus spread across cohorts I, II and III. The productive class
of verbs for the dialects in Coon Valley/Westby displayed two competing systems in these
three cohorts, mirroring the Norwegian dialectal origin of the immigrants.
Table 4. Nynorsk (a) and Bokmål (b) in the 1940s.
a. +finite -finite b. +finite -finite
+past Preterit:
kleima
Participle:
kleima +past Preterit:
kleimet
Participle:
kleimet
-past Present:
kleimar
Infinitive:
kleima -past Present:
kleimer
Infinitive:
kleime
Most of Haugen’s informants in the 1940s traced their roots from around Lake Mjøsa
and the lower part of Gudbrandsdalen Valley. The paradigm of the dialect in this area
maintains the finiteness distinction for the non-past forms (expressed by the suffix -rin the
present). The distinction
±
past is neutralized in this dialect, but only for non-finite forms
(infinitive versus past participle), cf. Table 5a. We dub this the Southern system, noting that
this expresses the same distinctions as the written Norwegian standard Nynorsk in Table 4a
above, although the exponents are different. The second system in the data we dub the
Northern system, consistent with the dialects of most of the (northern) Gudbrandsdalen area
(cf. Venås 1974). Here, the finite distinction is neutralized in the +past and the non-past
forms (Table 5b). What remains is simply the distinction past/non-past.
Table 5. Southern system (a) and Northern system (b) in Coon Valley/Westby in the 1940s.
a. +finite -finite b. +finite -finite
+past Preterit:
kleime
Participle:
kleime +past Preterit:
kleima
Participle:
kleima
-past Present:
kleimer
Infinitive:
kleime -past Present:
kleime
Infinitive:
kleime
Languages 2023,8, 49 17 of 36
The paradigm for this most productive class of verbs varies greatly in the relevant
spoken varieties; still, the overall picture is clear: there are two competing, rather stable
paradigms in use, depending on the origin of the immigrant community. In the Trønder
dialect of Wanamingo and Zumbrota (cf. Table 6b below), the present tense lacks the present
tense -r feature, just like the Northern system of Coon Valley and Westby.
15
This dialectal
micro-variation is maintained for certain speakers until well into the 2010 recordings
(Hjelde 2015). However, we see a propensity among younger speakers already in the
1940s recordings toward a third co-existing system, described by Eide and Hjelde (2015a)
as a hybrid system. The paradigm features a compromise between the two paradigms in
(Table 5a,b), cf. the paradigm in (Table 6a).
Table 6. Hybrid paradigm in Coon Valley/Westby later cohorts (a) trønder system (b).
a. +finite -finite b. +finite -finite
+past Preterit:
kleima
Participle:
kleima +past Preterit:
kleima
Participle:
kleima
-past Present:
kleimer
Infinitive:
kleime -past Present:
kleime
Infinitive:
kleim
We strongly suspect that the reason this “hybrid paradigm” emerges is that it converges
on the written standard Bokmål, the standard of written input, in the distinctions made.
Compare the hybrid system in (Table 6a) to the Bokmål system in (Table 4b). If the prevalence
of this paradigm is a result of the steady written input, this entails a contact-induced change
where the written standard is the donor language, and we may predict a steady rise and
prevalence of the Bokmål paradigm. However, this is not what we find. The prevalence
of the hybrid paradigm is clearly temporary, when we study later recordings of younger
cohorts.
16
Hjelde’s recordings from the 1980s and 1990s feature informants born between
1905 and 1932, thus belonging to cohorts II and III. The previous prevalent “hybrid system”
in (Table 6a) is now clearly receding, and perhaps somewhat unexpectedly, the Northern
system in (Table 5b) is the system now on the rise. This is a paradigm with no finiteness
distinction, mirroring exactly the distinctions made in the productive English system; again,
not in the actual exponents of tense forms, but in the distinctions made; cf. Table 7.
Table 7. Final Northern system in Coon Valley/Westby (a); English system (b).17
a. +finite -finite b. +finite -finite
+past Preterit:
kleima
Participle:
kleima +past Preterit:
claimed
Participle:
claimed
-past Present:
kleime
Infinitive:
kleime -past Present:
claim
Infinitive:
claim
Our 2010 recordings feature informants born between 1940 and 1961, i.e., cohorts IV
and V. The tendencies from previous recordings are now almost categorical: The North-
ern system, converging on the English system, is the one preferred by younger speakers,
although several parallel systems co-exist. Children of the Norwegian immigrants had
to cope with a great deal of dialectal micro-variation in their acquisition of Heritage Nor-
wegian. In a confusing linguistic landscape, the written input, mostly in Bokmål, would
provide a haven of regularity and authority and a compromise paradigm for the dialectal
micro-variation in the spoken input. Hence, cohorts II and III, with their access to ample
written Norwegian input, parochial schools, confirmation lessons and church services in
Norwegian, clearly show this tendency to favor the paradigm that converges on the Bokmål
paradigm, although refurbished with exponents from the spoken input.18
For cohorts IV and V, the access to written input, and Norwegian input in general,
is dramatically reduced. There is no longer enough input to maintain this paradigm,
and the prevailing productive paradigm for these younger cohorts is the one converging
Languages 2023,8, 49 18 of 36
on English—which incidentally makes the same distinctions as the previously receding
Northern system of Coon Valley/Westby. When there is no longer competition from the
written standard, this system belonging to one of the spoken dialects prevails, as it happens
to converge on the English productive paradigm for regular weak verbs.
Assuming this to be a reasonable explanation for the trends in our data and relating
this to “quantity or quality of input” discussed in Section 1, it seems that both play a
role—as the quantity of input is incrementally reduced for each cohort. However, the type
of input is also mirrored in the production of the different cohorts: Written input supports
a morphological paradigm of tense suffixes absent from the dominating spoken variants
for cohorts II and III. When the supply of written input stops, as for cohorts IV and V, the
system of one of the spoken variants wins out. The prevailing paradigm converges on the
paradigm of English, the dominating language for a NorNAHL bilingual.
4.4. Intra-Phrasal Syntax: Prenominal and Postnominal Possessives
Norwegian allows two orderings of the possessive and the noun phrase; the pronomi-
nal (1) and the postnominal (2), differing pragmatically, distributionally and morphologi-
cally. The postnominal possessive is morphologically more complex with two instances of
definiteness, the possessive and the definite suffix (her: -en). The prenominal possessive in
contrast has one definite marker, notably the possessive itself.
(1) min hest
me.poss.def horse
‘My horse’
(2) hest-en min
horse.def me.poss.def
‘My horse’
In spoken homeland Norwegian the postnominal possessive in (2) is considered
the unmarked and the one with the widest distribution. The prenominal in (1) has a
contrastive implication and is unlikely to appear in non-contrastive contexts without a
stressed possessive, cf. (3a) vs. (3b). In a contrastive context the prenominal possessive is
felicitous, but only with a stressed possessive; cf. (4a) and (4b).
Q1 Hvordan skal vi dra til byen?
How shall we go to town
‘How will we go to town?’
(3) a. *Vi kan ta vår/VÅR hest.
we can take our horse
‘We can take our horse.’
b. Vi kan ta hasten vår.
We can take horse.def
our.poss.def.
‘We can take our horse.’
Q2
Skal vi ta hesten vår til byen?
Shall we take horse.def
our.poss.def
to town
‘Should we take our horse to town?’
(4) a. Nei vi tar VÅR/*vår hest
no we take our horse
‘No, let’s take OUR horse.’
b. Nei vi tar hesten VÅR/*vår
no we take horse.def
our.poss.def
‘No, let’s take OUR horse.’
The two constructions have different distributions across spoken and written registers;
this is not surprising, given the dependence of stress for felicitous use in spoken registers.
According to Lødrup (2012), in written texts 78% of possessive noun phrases are prenominal,
and only 22% are postnominal. In spoken corpora, the numbers are reversed, with 25%
prenominal possessives and 75% postnominal possessives.
Our initial investigation of possessives in NorNAHL of the informants from cohorts
I–II in the 1940 material reveals proportions of pre- and postnominal possessives similar to
spoken homeland Norwegian (i.e., Lødrup’s findings); cf. Table 8. Over the next cohorts,
Languages 2023,8, 49 19 of 36
the percentage of prenominal possessives drops substantially, down to 7 percent in the
1990s recordings and then with an unexpected rise again in the recordings from 2010.
However, it turns out that one informant is responsible for this rise—an informant we call
outlier Per. Excluding Per from the count sends the percentage of prenominal possessives
down to 5 percent.
Table 8. Possessives in production of the different generational cohorts (cf. Table 3above).
Recordings Generational
Cohort
Prenominal
Possessive
Postnominal
Possessive
1940s I–II 23% 77%
1990s III 7% 93%
2010s Including
outlier Per III–V 13% 87%
2010s Excluding
outlier Per III–V 5% 95%
2010s outlier Per III–V 89% 11%
Westergaard and Anderssen (2015) find that the proportion of postnominal possessives
in the CANS corpus is slightly higher in NorNAHL than in homeland Norwegian, at 80%
vs. 75% (op. cit. 25–26). Here, outliers too make a big impact. Most prenominal possessives
are produced by three speakers. To increase the reliability of our numbers, we hence
included more informants, revisiting Haugen’s 1942 recordings from Coon Valley and
Westby and Hjelde’s 1980s–1990s transcribed recordings from Coon Valley and Westby,
Zumbrota and Wanamingo. We studied all CANS material from 2010–2016 from Blair,
Westby and Coon Valley, Zumbrota and Wanamingo, and lastly, we examined all Haugen’s
speakers in the CANS material, regardless of where they came from. For this larger study,
we used a search function [Pos + Noun] and [Noun + Pos], before manually verifying all
hits to remove irrelevant constructions like, e.g., “boka hennes Jane” (book the her Jane,
‘the book of Jane’s’).19
Haugen’s NorNAHL recordings taken together show the expected proportion of
prenominal possessives, mirroring that of homeland Norwegian (Lødrup 2012); cf. Table 9.
When we separate Haugen’s recordings in Coon Valley and Westby from the rest, however,
we find that at this stage the data are already different from the rest of the NorNAHL
material in CANS. The Coon Valley/Westby material features a proportion of 13 percent
for prenominal possessives. This is similar to what we find in the data from this area still in
the 1980s, 1990s and 2010s, suggesting that this may be a dialectal feature from early on.
Preliminary investigations of older Norwegian dialect recordings from Gudbrandsdalen
in the database LIA reveal data to support this assumption: the Gudbrandsdalen dialect
features a proportion of prenominal possessives closer to 10 percent instead of the 25 percent
typical of other Norwegian dialects and standards.20
Table 9. Prenominal possessives in the CANS material.
Recordings Prenominal Possessive
Percentage
Prenominal Possessive
Numbers
1930s and 1940s (Haugen) 23% 91/404
1942 Coon Valley/Westby
(Haugen) 13% 16/115
1980s and 1990s (Hjelde) 12% 15/126
2010s (CANS-project) 5% 40/739 *
*: These are the numbers excluding outlier “Per”, cp. Table 8.
Languages 2023,8, 49 20 of 36
This levels out in the overall CANS material from the 1930s and 1940s because the
proportion is evidently higher in other spoken variants of NorNAHL. Again, we believe that
this is the impact of written Norwegian, which would draw the proportion of prenominal
possessives higher. Let us recall that in written Norwegian (Lødrup 2012), the proportion
of prenominal possessives is 78% instead of 25%, the latter being the average proportion
for spoken Norwegian. This conjecture would still not explain the numbers for Coon
Valley/Westby, unless we assume these speakers to be exempt from the “people of readers”,
which is implausible. The proportion of prenominal possessives may definitely be affected
by written input in certain cases, and we have convincing evidence of this effect in the
case of “outlier Per”. This informant produces prenominal possessives in 89 percent of
the instances, in sharp contrast to his peers whose production features 5% prenominal
possessives on average. Per volunteers the information that he obtains almost all his
Norwegian input from written texts, which is mirrored in his oral production. Clearly,
the behavior of one single informant does not prove the effect of written input for other
NorNAHL speakers, but the numbers for this particular informant are certainly very
striking, and the trends in the material are clearly compatible with this explanation.
CLI seemingly plays a lesser role here, as the postnominal structure spreads in Nor-
NAHL in spite of the prenominal structure being the only option in English. In spoken
Norwegian, the postnominal possessive is the unmarked option. Three-quarters of posses-
sive construction contexts feature a postnominal possessive, while the use of prenominal
possessives is limited to contrastive contexts. The fact that the prenominal possessive
is dependent on stress and intonational features for its felicitous use only adds to the
complexity of restrictions for this construction. Acquiring the contextual restrictions for
each construction requires rich input to observe and internalize the specific triggering cues,
and in this NorNAHL contact situation such rich input is not available. When input from
written Norwegian decreases, the default construction hence extends its functional domain
into the domain of both variants, as compared to the homeland varieties.21
4.5. Inter-Phrasal Syntax: Topicalization and Verb Movement
In an average corpus of spoken and written texts, about two-thirds of all Norwegian
main clause declaratives are subject-initial, although the type of subject differs somewhat
from spoken to written register. Spoken declaratives more often start with a pronoun or an
expletive (Eide and Sollid 2011;Eide 2011), written declarative main clauses show a stronger
propensity toward lexical, more complex subjects. When a main clause declarative starts
with a non-subject, this is referred to as topicalization; depending on your theoretical stance,
the topicalization position can also host subjects (cf. van Craenenbroeck and Haegeman
(2007) for a very useful overview of this debate). Topicalization, fundamentally constrained
by a set of rather complex contextual and pragmatic cues, belongs to the syntax–pragmatic
interface, attested as an area rather vulnerable to attrition in heritage languages, and
challenging for bilinguals generally (Sorace 2011;Benmamoun et al. 2013;Kroch and Taylor
1996). Typically, the pragmatic factors governing and licensing topicalization constructions
in a monolingual context are known to be changed and sometimes relaxed in a bilingual
population, resulting in what is perceived as “overuse” of the construction (cf. Prince
1981 on “Yiddish movement”). Pulling in the opposite direction is the robust observation
that bilinguals are prone to avoid constructions calling for a fine-tuned set of complex
pragmatic rules due to the amount of input required to acquire these rules, which may
not be available in a bilingual context. Hence, we may also expect avoidance in otherwise
typical topicalization contexts (e.g., Kleinmann 1977).
In Norwegian, whatever type of element fills the first position, the finite verb occurs in
second position, so when a non-subject is topicalized, the verb moves, ending up to the left
of the subject in Norwegian declarative main clauses. English is different; firstly, English’s
main clause declaratives are subject-initial in 85–90% of the cases; secondly, even when the
declarative starts with a non-subject, the verb does not move and remains to the right of
the subject no matter what fills the first position.
Languages 2023,8, 49 21 of 36
Topicalization and verb second (V2) in NorNAHL occurs at a proportion mirroring
homeland Norwegian (cf. Eide and Hjelde 2015a;Eide 2019). For the youngest speakers,
however, the proportion of topicalizations in production drops to a level converging on
the English ratio, correlating with a rise in placements of the verb differing from that of
homeland Norwegian.
22
There is presumably an underlying relation between topicalization
and V2: as heritage speakers gradually decrease their number of topicalizations, the V2
rule starts to recede. A speaker of younger cohorts typically topicalizes less than previous
cohorts and puts the verb either in the V2 position (as in 5) or in the V3 position (as in 6), in
a seemingly random fashion.23
(5)
Nå går vi ferbi hår je vaks opp
now walk we past where I grew up
‘Now let’s walk past the place where I grew up.’
(6)
Der dem lager vin
Now walk we past
’There they make wine’
Examining our own data from parent-and-child pairs, we find a development similar
to that attested in previous work. The older generation clearly exhibits a stronger tendency
to topicalize a non-subject than the younger generation; cf. Table 10.
The older generation topicalizes in 23 percent of the declaratives, similar to spoken
homeland Norwegian. The child generation topicalizes non-subjects at about half this rate
(10–15 percent).24
The dent in Figure 6is due to Liam, born 1943, who uses topicalization almost as often
as his father, born 1915, and in a limited data set, this distorts an otherwise clear picture:
25
The proportion of topicalizations in the informant’s production decreases and the number
of non-V2 placements of the finite verb rises for the younger generation as compared to
parents and earlier cohorts.
Table 10. V2 and topicalizations in parent–child pairs.
Speaker Thor Iris Mason Liam Ted Beth Laura Lars
Cohort II IV II IV III V III V
Birth year 1908 1939 1915 1943 1932 1957 1922 1961
Declaratives (N) 235 274 87 164 779 68 254 318
Topicalizations 23% 13% 28% 25% 29% 10% 27% 10%
Non-V2 5% 11% 6% 0% 1% 14% 0% 20%
Languages 2023, 8, x FOR PEER REVIEW 22 of 38
Topicalizations 23% 13% 28% 25% 29% 10% 27% 10%
Non-V2 5% 11% 6% 0% 1% 14% 0% 20%
The older generation topicalizes in 23 percent of the declaratives, similar to spoken
homeland Norwegian. The child generation topicalizes non-subjects at about half this rate
(10–15 percent).
24
The dent in Figure 6 is due to Liam, born 1943, who uses topicalization almost as
often as his father, born 1915, and in a limited data set, this distorts an otherwise clear
picture
25
: The proportion of topicalizations in the informant’s production decreases and
the number of non-V2 placements of the finite verb rises for the younger generation as
compared to parents and earlier cohorts.
Figure 6. Topicalization and non-V2 according to informant’s year of birth.
This development is likely tied to the quantity of input. The ratio of topicalizations
remains similar across written and spoken Norwegian (ca. 35%) and nothing suggests that
topicalization ratios vary across Norwegian traditional dialects.
26
“Exceptional non-V2”,
found with certain adverbs (cf. note 18) occurs in a small percentage of homeland main
clause declaratives only; thus, the non-V2 emerging in youngest cohorts is not likely to be
an effect of positive evidence in specific types of input. A plausible explanation of either
of the trends seen in Figure 6 should therefore not rest on different input sources (i.e.,
input quality), but rather on the absence of relevant input (i.e., input quantity).
5. Modeling Multilingualism and Linguistic Change
Until recently, (idealized) monolingualism was held as the type of language use to
be explained, a widespread view across different frameworks (e.g., Chomsky 1965; Schil-
ling-Estes 2004). Contemporary linguistics embraces bilingualism and multilingualism as
the norm in a global view on language use (Vaid 2002, p. 1), and multiple languages in
one mind requires a different approach to language description and explanation. The abil-
ity for multilingualism “should in effect form the basis of the description from the begin-
ning”; according to Cook and Newson (2007).
T. Roeper (1999); T. W. Roeper (2016) advocates that every language is in effect a
superset of many parallel, often conflicting grammars, entailing that any language user is
in fact multilingual, however “monolingual”: any speaker knows a range of registers, di-
alects, child talk, etc., from exposure to different speakers of the “same” language. Some
arrived at this insight earlier than the mainstream; cf. Martinet’s preface to Weinreich
([1953] 2011):
It remains to be emphasized that linguistic diversity begins next door, nay, at
home and within one and the same man [...]. What we heedlessly and somewhat
rashly call ‘a language’ is the aggregate of millions of such microcosms many of
which evince such aberrant linguistic comportment that the question arises
whether they should not be grouped into other ‘languages’.
Figure 6. Topicalization and non-V2 according to informant’s year of birth.
This development is likely tied to the quantity of input. The ratio of topicalizations
remains similar across written and spoken Norwegian (ca. 35%) and nothing suggests that
topicalization ratios vary across Norwegian traditional dialects.
26
“Exceptional non-V2”,
found with certain adverbs (cf. note 18) occurs in a small percentage of homeland main
clause declaratives only; thus, the non-V2 emerging in youngest cohorts is not likely to be
Languages 2023,8, 49 22 of 36
an effect of positive evidence in specific types of input. A plausible explanation of either of
the trends seen in Figure 6should therefore not rest on different input sources (i.e., input
quality), but rather on the absence of relevant input (i.e., input quantity).
5. Modeling Multilingualism and Linguistic Change
Until recently, (idealized) monolingualism was held as the type of language use
to be explained, a widespread view across different frameworks (e.g., Chomsky 1965;
Schilling-Estes 2004, pp. 163–64). Contemporary linguistics embraces bilingualism and
multilingualism as the norm in a global view on language use (Vaid 2002, p. 1), and
multiple languages in one mind requires a different approach to language description
and explanation. The ability for multilingualism “should in effect form the basis of the
description from the beginning”; according to Cook and Newson (2007, pp. 223–24).
Roeper (1999); Roeper (2016) advocates that every language is in effect a superset
of many parallel, often conflicting grammars, entailing that any language user is in fact
multilingual, however “monolingual”: any speaker knows a range of registers, dialects,
child talk, etc., from exposure to different speakers of the “same” language. Some arrived at
this insight earlier than the mainstream; cf. Martinet’s preface to Weinreich ([1951] 2011):
It remains to be emphasized that linguistic diversity begins next door, nay, at
home and within one and the same man [
. . .
]. What we heedlessly and some-
what rashly call ‘a language’ is the aggregate of millions of such microcosms
many of which evince such aberrant linguistic comportment that the question
arises whether they should not be grouped into other ‘languages’.
Chomsky (2000, p. 187) suggests how different styles link to different contexts:
[E]very human being speaks a variety of languages. We sometimes call them
different styles or dialects, but they are really different languages, and somehow
we know when to use them, one in one place and another in another place. Now
each of these different languages involves a different [“grammar”].27
What is referred to here as styles, dialects or “grammars”, is referred to in other works
as varieties, lects or registers, typically depending on the specific linguistic feature variants
one wants to address. While grammar usually refers to morphosyntactic features and register
refers to the choice of vocabulary,
28
the term styles often refers to phonological variants and
pronunciation, but also to the choice of rhetorical devices. Standard often regards distance to
some official norm, and while lects are “languages” confined, e.g., by geography (dialects)
or social class (sociolect), a “language is a dialect with a navy and an army” (proverb
attributed to Max Weinreich).
However, the features targeted through these specific terms, although they clearly
belong to different linguistic levels and systems, strongly tend to covary within certain
lects or languages. For instance, the occurrence of a particular morphological variant
is usually accompanied by a particular phonological variant (Blom and Gumperz 2000,
pp. 118–19). The choice of a specific grammatical morpheme variant often affects the
accompanying syntax, e.g., Reite (2011) shows how the choice of a particular wh-word
allows for “exceptional V3” if it belongs to the set of older dialectal wh-words, as opposed
to the set of “new” dialectal wh-words that happen to converge on the standard Bokmål.
To a bilingual speaker of Norwegian and English, a string consisting of all English
words is usually accompanied by English pronunciation and grammar, and Norwegian
features on all levels are selected for a string consisting of only Norwegian words (but
cf. below for exceptions to this typical situation; cf. also Sunde 2019). This covariation or
clustering of feature variants on different levels is what constitutes varieties, sociolects,
dialects, languages, and styles, according to Ulbricht and Werth (2021). In the remainder
of this paper, we use the term repertoire for all these different types of lects. Although
these specialized terms involve specific variants at different linguistic levels occurring in
particular linguistic context types, they do in effect all imply a speaker’s linguistic choices
Languages 2023,8, 49 23 of 36
determined by contextual adaptation, where the synthetizations of selected features are
what constitutes the specific lect, or in our terms, repertoire.
Language learners from early on come to associate specific registers (including national
languages) with specific contexts. A specific input with specific traits is always given
in a specific context, and learners evidently associate the features synthesized from a
given linguistic input source with its contextual features.
29
Thus, we may assume that a
vocabulary element such as a word comes with a feature matrix indicating not only the
meaning (including connotations), morphology, distribution (syntax) and pronunciation,
but also with features encoding appropriate context, depending on the input source(s) for
this word. The same will be true for the separate linguistic features.
A Heritage Norwegian speaker in the 1930s’ American Midwest would have access to
many sources of continuous input (cf. discussion above), and this input would contribute
to the pool of linguistic resources of the language user. Figure 7depicts the input landscape
of such a NorNAHL speaker. Note that since every language user has their own unique
linguistic biography, there would be a lot of individual variation, e.g., some NorNAHL
speakers had access to Irish or German input as well.
Languages 2023, 8, x FOR PEER REVIEW 24 of 38
Figure 7. Norwegian receptive repertoire for a NorNAHL speaker in the 1930s, adapted from Eide
(2022). Norwegian input: blue arrows; English input: green arrows.
5.1. Receptive and Productive Repertoire
The necessity to distinguish between receptive and productive competence was
noted already by Troike (1970), who noted that whereas productive competence “can be
inferred from free performance”, receptive competence “is not so directly attested”.
Troike proposed elicitation experiments as a means to access and assess receptive compe-
tences of language users, and many such techniques have later been developed. Whereas
receptive and productive language competences, governing perception and production
could in principle adhere to completely different principles, it seems instead that the
“laws” applying across linguistic reception and production systems are strikingly similar
(Pater 2004, p. 1), although production usually lags behind in language acquisition. In
language attrition (e.g., for heritage speakers), the opposite relation holds: receptive com-
petence lingers on well after the productive competence is long gone in the individual
language user (Montrul 2008, p. 81).
Eide and Sollid (2011, p. 335) discuss the relation between receptive and productive
repertoires and note that the line between these two is sometimes fuzzy. Many features of
the receptive repertoire may become productive and any feature from the productive rep-
ertoire is potentially relegated to the receptive repertoire (cf. passive and active vocabu-
laries). Trivially true is that there is no acquisition without input, cf. Meisel (2020, p. 33);
thus, nothing can become part of the receptive competence which is not first in the input.
30
In turn, for a feature to be acquired, it must first be a part of the receptive repertoire before
it can be transferred to the productive repertoire. This is trivially true. Likewise, for almost
all individual language users, the productive repertoire must be a subset of the usually
substantially larger receptive repertoire (although exceptions exist, cf. below). Language
users are sometimes reluctant to produce languages and variants where they have sub-
stantial receptive competence, and acculturation may be one factor. Schumann (1978) ob-
served that an affinity to identify with the cultural group speaking a given language trans-
lates into a propensity to speak that language, and this is evident even for the children of
Norwegian immigrants in the Midwest. The Norwegian-American immigrant and teacher
Carl Raugland, MI, wrote a letter to his brother in Norway in 1900 (Soleng 2007, p. 44):
Figure 7.
Norwegian receptive repertoire for a NorNAHL speaker in the 1930s, adapted from Eide
(2022). Norwegian input: blue arrows; English input: green arrows.
5.1. Receptive and Productive Repertoire
The necessity to distinguish between receptive and productive competence was noted
already by Troike (1970), who noted that whereas productive competence “can be inferred
from free performance”, receptive competence “is not so directly attested”. Troike pro-
posed elicitation experiments as a means to access and assess receptive competences of
language users, and many such techniques have later been developed. Whereas receptive
and productive language competences, governing perception and production could in prin-
ciple adhere to completely different principles, it seems instead that the “laws” applying
across linguistic reception and production systems are strikingly similar (Pater 2004, p. 1),
although production usually lags behind in language acquisition. In language attrition
(e.g., for heritage speakers), the opposite relation holds: receptive competence lingers on
well after the productive competence is long gone in the individual language user (Montrul
2008, p. 81).
Eide and Sollid (2011, p. 335) discuss the relation between receptive and productive
repertoires and note that the line between these two is sometimes fuzzy. Many features
Languages 2023,8, 49 24 of 36
of the receptive repertoire may become productive and any feature from the productive
repertoire is potentially relegated to the receptive repertoire (cf. passive and active vo-
cabularies). Trivially true is that there is no acquisition without input, cf. Meisel (2020,
p. 33); thus, nothing can become part of the receptive competence which is not first in
the input.
30
In turn, for a feature to be acquired, it must first be a part of the receptive
repertoire before it can be transferred to the productive repertoire. This is too trivially true.
Likewise, for almost all individual language users, the productive repertoire must be a
subset of the usually substantially larger receptive repertoire (although exceptions exist,
cf. below). Language users are sometimes reluctant to produce languages and variants
where they have substantial receptive competence, and acculturation may be one factor.
Schumann (1978) observed that an affinity to identify with the cultural group speaking
a given language translates into a propensity to speak that language, and this is evident
even for the children of Norwegian immigrants in the Midwest. The Norwegian-American
immigrant and teacher Carl Raugland, MI, wrote a letter to his brother in Norway in 1900
(Soleng 2007, p. 44):
The language of our children is almost exclusively English, but they understand
every word spoken to them in Norwegian; it is hard to teach the children Norwe-
gian in this country, as English comes more naturally to them.31
NorNAHL speakers could be anywhere on the native speaker continuum from mono-
lingual Norwegian to unbalanced bilinguals with English as their almost exclusive language.
Whatever their productive proficiency, not all variants present in the input (cf. Figure 7)
would be part of their respective productive repertoires. In fact, several of these sources
of input provide linguistic features that would never make it past the border between
receptive and productive repertoire. For instance, the church services, the parochial schools
and specifically language-learning settings would include Bible studies and Norwegian
hymns with linguistic traits never used in ordinary spoken production (cf. Figure 8). The
Lord’s Prayer has several instances of the archaic Norwegian subjunctive, cf. examples (7)
and (8). Likewise, Norwegian Lutheran hymns were often translated from German and
therefore often feature verb-final clausal structures that simply never appear in the spoken
production of NorNAHL speakers, cf. examples (9)–(11) from Indgangssalme in Landstad’s
book of hymns for Lutheran Christians in America, 1909.
Languages 2023, 8, x FOR PEER REVIEW 26 of 38
Figure 8. Receptive and productive repertoire for a NorNAHL speaker in the 1930s, adapted from
Eide (2022). Norwegian input: blue arrows; English input: green arrows.
Likewise, the cartoons in Decorah-Posten portrayed characters speaking different di-
alects, who would also contribute to the receptive competence of the readers. The in-
stances of the non-local dialect would usually not however make it into the speakers’ daily
productive repertoire. When prompted, many speakers would nevertheless be able to re-
cite, reproduce and hence produce various words and phrases from this contextually
much more confined input, normally relegated to their receptive repertoire.
The flexible, fuzzy border between receptive and productive repertoires allows the
productive repertoire of a language user to vary and change based on all kinds of con-
scious or subconscious choices. Likewise, changes in the productive repertoire may be
short-term (accommodation) or long-term (e.g., age-grading or language attrition). Although
this can easily be depicted in this model, we will not go into the details of age-grading
here, although we will discuss accommodation and briefly also attrition.
5.2. Selecting the Productive Repertoire: “Somehow We Know When to Use Them…”
Acculturation, as mentioned above, is one factor influencing a speaker’s propensity
to activate features from her receptive repertoire and include these in her production.
However, there are a plethora of possible ways and potential reasons why a speaker se-
lects the specific set of features constituting her daily productive repertoire. Here,
Matsuda (1991, p. 1329), discussing a person’s accent (although in effect this pertains
equally well to what we refer to as productive repertoire):
Your accent carries the story of who you are–who first held you and talked to
you when you were a child, where you have lived, your age, the schools you
attended, the languages you know, your ethnicity, whom you admire, your loy-
alties, your profession, your class position: traces of your life and identity are
woven into your pronunciation, your phrasing, your choice of words. Your self
is inseparable from your accent.
The “default settings” of a person’s productive repertoire is likely to be overridden
almost on a daily basis, since all speakers, consciously or unconsciously, adjust their lan-
guage to their interlocutors to enhance the potential for successful communication. This
fundamental observation gave rise to seminal approaches such as accommodation theory
(Giles 1973) and theories of audience design (Bell 1984). Speakers select, usually from a wide
spectrum of possible resources, the sub-repertoire presenting itself as most appropriate to
Figure 8.
Receptive and productive repertoire for a NorNAHL speaker in the 1930s, adapted from
Eide (2022). Norwegian input: blue arrows; English input: green arrows.
Languages 2023,8, 49 25 of 36
Examples of subjunctive (The Lord’s Prayer)
(7)
Helliget vorde Dit Navn
holy.made be.SUBJ Your name
‘Hallowed be Thy name.’
(8)
Komme Dit Rike
come.SUBJ Your kingdom
‘Thy kingdom come.’
Examples of verb final clause structures (Indgangssalme)
(9)
i Evighet Vær hos oss og vår Sjæl bered
In eternity be with us and our soul prepare
‘Be with us in eternity and prepare our soul’
(10)
At Gud vi søke Nåde få
That God we seek
mercy receive
‘that we seek God, receive mercy’
(11)
I våre
Synder
la oss ei
forgå!
In our sins let us not
perish
‘Do not let us perish in our sins’
Likewise, the cartoons in Decorah-Posten portrayed characters speaking different di-
alects, who would also contribute to the receptive competence of the readers. The instances
of the non-local dialect would usually not however make it into the speakers’ daily pro-
ductive repertoire. When prompted, many speakers would nevertheless be able to recite,
reproduce and hence produce various words and phrases from this contextually much
more confined input, normally relegated to their receptive repertoire.
The flexible, fuzzy border between receptive and productive repertoires allows the
productive repertoire of a language user to vary and change based on all kinds of conscious
or subconscious choices. Likewise, changes in the productive repertoire may be short-term
(accommodation) or long-term (e.g., age-grading or language attrition). Although this can easily
be depicted in this model, we will not go into the details of age-grading here, although we
will discuss accommodation and briefly also attrition.
5.2. Selecting the Productive Repertoire: “Somehow We Know When to Use Them . . . ”
Acculturation, as mentioned above, is one factor influencing a speaker’s propensity
to activate features from her receptive repertoire and include these in her production.
However, there are a plethora of possible ways and potential reasons why a speaker selects
the specific set of features constituting her daily productive repertoire. Here, Matsuda
(1991, p. 1329), discussing a person’s accent (although in effect this pertains equally well to
what we refer to as productive repertoire):
Your accent carries the story of who you are–who first held you and talked to you
when you were a child, where you have lived, your age, the schools you attended,
the languages you know, your ethnicity, whom you admire, your loyalties, your
profession, your class position: traces of your life and identity are woven into
your pronunciation, your phrasing, your choice of words. Your self is inseparable
from your accent.
The “default settings” of a person’s productive repertoire is likely to be overridden
almost on a daily basis, since all speakers, consciously or unconsciously, adjust their
language to their interlocutors to enhance the potential for successful communication.
This fundamental observation gave rise to seminal approaches such as accommodation
theory (Giles 1973) and theories of audience design (Bell 1984). Speakers select, usually
from a wide spectrum of possible resources, the sub-repertoire presenting itself as most
appropriate to a given linguistic context, varying their set of covariant features synthesized
into specific varieties. Additionally, whereas the specific linguistic features are usually
subject to co-occurrence restrictions, the selection of varieties (i.e., a synchronized set of
features) is clearly subject to extra-linguistic factors. Any knowledge the speaker has about
the interlocutor, including his level of comprehension, is likely to affect the productive
choices. This has previously been attested in foreigner-directed speech, infant- and child-
Languages 2023,8, 49 26 of 36
directed speech, and computer-directed speech, to mention but a few, where this knowledge
is likely to affect lexical selection, selection of phonological, morphological, syntactical
features, as well as the specific rhetorical devices.32
The productive repertoire of a NorNAHL speaker would be influenced by the same
general mechanisms; the specific context, interlocutor or audience and the speaker’s sym-
pathy or antipathy toward the different dialects, homeland Norwegian and the language of
the dominating society, English, would drive conscious and unconscious decisions in the
direction of convergence or divergence, respectively. The propensity for language mixing,
either between Norwegian and English, or between the different dialects available in the
input, would also depend to a certain degree on the given speaker’s level of metalinguistic
awareness, and their inclination for purism or alignment.
Any language contact situation can be described as a push and pull between these
two conflicting forces, where purism applies to the conservative, protectionist language
forces, where the ideal is the “pure language” (cf., e.g., Thomas 1991, p. 33). Alignment
refers to the opposite forces pushing linguistic variants toward convergence; thus, previous
distinctions between two language systems are reduced. This implies the tendency to
borrow linguistic elements, to code-switch and to relax restrictions on the use of linguistic
elements, allowing previously restricted elements to appear in new contexts. Alignment
can hence be explained as a relaxation on the contextual restrictions on specific forms.
Matras (2009, pp. 151, 237) explains how any bilingual mind is always struggling to keep
the different languages apart and to make sure that contextual restrictions on different
languages are respected. The language user constantly scans the linguistic environments to
discover an overlap between his two grammars, allowing him or her to exploit his linguistic
repertoire to the maximum.
[T]here is pressure on the bilingual to simplify the selection procedure by reduc-
ing the degree of separation between the two subsets of the repertoire [. W]e
might view the replication of patterns as a kind of compromise strategy that
[
. . .
] reduce[s] the load on the selection [
. . .
] mechanism by allowing patterns
to converge, thus maximizing the efficiency of speech production in a bilingual sit-
uation.
This process shows itself in NorNAHL i.a. through a tendency toward a certain degree
of koinéization and dialect leveling among the youngest cohorts (cf. Hjelde 2012 for a
detailed discussion). When previous micro-distinctions are discarded and instead aligned
in a common paradigm (see the “hybrid” tense affix paradigm illustrated in Table 6) this can
be seen as instances of the same process (cf. also Eide and Hjelde 2015a;Eide 2019).33 The
same type of alignment is attested at later stages as the most recent productive paradigms
converge on the exact same distinctions made in the corresponding English paradigms.
The incrementally decreasing ratios of topicalizations in declaratives in NorNAHL and the
gradual increase in non-V2 structures can also be seen as alignment with and convergence
toward English information structure.
Morphosyntactic features are hard to keep track of and consciousness about these
require a metalinguistic level of awareness which is rather rare among the general public.
However, in bilingual and bilectal contact situations, tendencies toward a kind of hyper-
dialectism are sometimes found with speakers of the purist inclination. Jamieson (2020)
describes this tendency among younger speakers of Shetland Scots, and Cornips and Po-
letto (2005) discuss several studies where speakers who take an unusually strong pride
in their language (teachers, local poets or other people with a special interest in the local
dialect) are rather unreliable as informants due to their inclination toward normativity and
hypercorrection.
In a previous study (Eide and Hjelde 2015a), we found that one informant “Lena”
obviously associated topicalization with “norwegianness”, since in her planned speech,
almost half of the declaratives featured topicalization (against an average of 35% in home-
land Norwegian). In her online production, however, this ratio could not be upheld, and
Languages 2023,8, 49 27 of 36
her proportion of topicalizations in a dialogue with another NorNAHL speaker plunged to
about 15%.
Words and phrases are easier to relate to metalinguistically. Hence, the purists mostly
make their mark in the lexical domain, actively trying to avoid all words they suspect to
be of English origin. Eide and Hjelde (2015a) report on an informant who evidently had
consciously banned the heritage Norwegian loan word leik (from English lake) from his
production. As he also did not produce the corresponding homeland Norwegian entry,
innsjø, and he ended up using the word ei mjøs ‘a mjøs’. It seems probable that the informant
has analyzed the proper name Mjøsa as a feminine common noun, cf. ei bok–bok-a (‘a
book-book.DEF’); thus, Mjøsa would be the definite singular of ei mjøs:mjøs+a. Even in the
2010 fieldwork, we were repeatedly confronted with the request from many informants to
help them rid themselves of their production of English “contamination”, i.e., English loan
words, since we as Norwegian scholars surely would be able to pinpoint what was “real
Norwegian” and what was not. Additionally, even if there was a general acceptance for the
inclusion of English words into Norwegian in the speech community, this was only to a
certain limit; those who used many such English words were derogatorily referred to as
Yankee-sprengt (Yankee-bursting) or engelsk-sprengt (English-bursting).
Although NorNAHL speakers would repeatedly reject the idea that they may use
a word such as spost (‘supposed’) in their own production, the recordings revealed a
different story (Eide and Hjelde 2015b). As is well known (e.g., Trudgill 1972), linguistic
self-reporting is notoriously misleading, and even the most faithful purist will come to use
constructions and phrases which are banned from his productive repertoire—in principle.
In online production, however, these purist intentions often cannot be upheld.
5.3. How Individual Choices Accumulate and Affect the Input for the Next Generation
Purism may seem to be a more viable route than alignment toward long-lasting mainte-
nance of a heritage language. However, this may be something of a misconception. Natural
languages develop to fit the society where the language is used and include or invent new
words and elements to cover ever shifting circumstances of the modern society. Purists
tend to expel any attempts at modernizing the language, cf. the development described in
Section 3.3 for the newspaper Decorah-Posten, where this type of purism was an advantage
in the competition against other newspapers for some time, but eventually yielded “a news-
paper written by old people for old people” (cf. above). Likewise, Norwegian continued as
a NorNAHL lingua sacra, the language of faith, well after English had taken over all other
domains. However, without a normal renewal and adjustment of the language toward new
domains, the number of contexts for NorNAHL as a useful communication tool recedes.
For young NorNAHL learners, English will present itself as more convenient even if all
your peers also know Norwegian, simply because your NorNAHL does not include the
registers necessary to talk to your friends about modern matters. This effect is likely to be
propelled as the number of potential interlocutors decreases, since this too contributes to
the rapidly shrinking number of NorNAHL input sources. Additionally, this development
probably affects the overall robustness of your NorNAHL representations; cf. Gollan et al.
(2015, p. 154):
Contextual diversity leads to greater ease of retrieval in single-word recognition
(and reading aloud) to a greater extent than does frequency of use [
. . .
], and
similarly, use of an HL in different places, with different people at different times
of day, and in different locations may be more important to achieving proficient
HL word production than the number of exposures to HL words.
Alignment may be more fruitful in the long run, but in its extreme consequence
it will likely relegate the heritage language to some characteristic words, phrases and
constructions, where the overall frame of the language (cf. Myers-Scotton (1993)), is now
the dominating language, English—a state that Haugen claimed to exist already in 1938
(Haugen 1939, p. 46). If either purism or alignment is the chosen path for the individual
NorNAHL speaker, it is crucial to note that the choices made by the individual language
Languages 2023,8, 49 28 of 36
user may accumulate into linguistic change for the next generation, since particular features
or constructions not selected, and hence not produced by the parent generation, will
be absent from the input for the next generation of speakers. This vulnerability will be
counterbalanced in a situation where the language acquirers have access to many different
sources of input, because the construction or feature missing from the production of the
parent may be present in the other sources, and hence still available for acquisition. Gollan
et al. (2015, p. 154):
Different speakers talk in different ways, and interaction with a variety of speak-
ers will likely result in exposure to a broader variety of accents, speech rates, and
syntactic structures and a larger number of HL words. Alternatively, using the
HL with a greater number of speakers may lead to a greater diversity of contexts
with which the HL words are associated.
For NorNAHL speakers in the 2000s, the previous steady supply of sources for Nor-
wegian input has now dried up; cf. Figure 9, where the blue arrows depicting sources for
Norwegian input are now replaced by green arrows (Figure 9), depicting the English input
replacing the Norwegian sources (cp. Figures 7and 8). The receptive repertoire in Norwe-
gian has shrunk to a size where selection between different styles and synonyms are no
longer possible, the domains and contexts for which appropriate words and phrases exist
are likewise reduced to almost none and the number of possible speakers for interaction in
Norwegian are not only few but also far apart.
Languages 2023, 8, x FOR PEER REVIEW 29 of 38
Alignment may be more fruitful in the long run, but in its extreme consequence it
will likely relegate the heritage language to some characteristic words, phrases and con-
structions, where the overall frame of the language (cf. Myers-Scotton (1993)), is now the
dominating language, English—a state that Haugen claimed to exist already in 1938
(Haugen 1938a). If either purism or alignment is the chosen path for the individual Nor-
NAHL speaker, it is crucial to note that the choices made by the individual language user
may accumulate into linguistic change for the next generation, since particular features or
constructions not selected, and hence not produced by the parent generation, will be ab-
sent from the input for the next generation of speakers. This vulnerability will be counter-
balanced in a situation where the language acquirers have access to many different
sources of input, because the construction or feature missing from the production of the
parent may be present in the other sources, and hence still available for acquisition. Gollan
et al. (2015, p. 154):
Different speakers talk in different ways, and interaction with a variety of speak-
ers will likely result in exposure to a broader variety of accents, speech rates,
and syntactic structures and a larger number of HL words. Alternatively, using
the HL with a greater number of speakers may lead to a greater diversity of
contexts with which the HL words are associated.
For NorNAHL speakers in the 2000s, the previous steady supply of sources for Nor-
wegian input has now dried up; cf. Figure 9, where the blue arrows depicting sources for
Norwegian input are now replaced by green arrows (Figure 9), depicting the English in-
put replacing the Norwegian sources (cp. Figures 7 and 8 ). The receptive repertoire in
Norwegian has shrunk to a size where selection between different styles and synonyms
are no longer possible, the domains and contexts for which appropriate words and
phrases exist are likewise reduced to almost none and the number of possible speakers for
interaction in Norwegian are not only few but also far apart.
Figure 9. Receptive and productive repertoire for a NorNAHL speaker in the 2000s, adapted from
Eide (2022). Norwegian input: blue arrows; English input: green arrows.
Furthermore, since the receptive repertoire is now too small to accept different ac-
cents, dialects and variants of NorNAHL, understanding new speakers will be hard. In
our 2010 field work, the diversity of Norwegian dialects and standards posed a real prob-
lem in the interviews, since the go-to-mechanism of most Norwegians when talking to
foreigners is accommodating toward standard spoken Bokmål, which was of no use
Figure 9.
Receptive and productive repertoire for a NorNAHL speaker in the 2000s, adapted from
Eide (2022). Norwegian input: blue arrows; English input: green arrows.
Furthermore, since the receptive repertoire is now too small to accept different accents,
dialects and variants of NorNAHL, understanding new speakers will be hard. In our 2010
field work, the diversity of Norwegian dialects and standards posed a real problem in the
interviews, since the go-to-mechanism of most Norwegians when talking to foreigners is
accommodating toward standard spoken Bokmål, which was of no use whatsoever. Instead,
all field workers had to improvise their best imitation of a Gudbrandsdalen dialect, and
this strategy only worked in some of the conversations.
As language users are exposed to different variants of the variety they acquire, their
receptive channel grows more robust, and their receptive competence increases. Continuous
investigations of mutual intelligibility between the Scandinavian countries clearly reveal
that the more exposure to different variants, different dialects and neighboring languages,
Languages 2023,8, 49 29 of 36
the more robust and tolerant are your receptive abilities. Delsing and Lundin-Åkesson
(2005) (cf. also Frøshaug and Stende 2021) found that Norwegian adolescents are far better
at understanding the other Scandinavian languages than their Swedish and Danish peers,
and even Norwegian adolescents with foreign-born parents are better at decoding the
neighboring languages than non-immigrant youngsters in Sweden and Denmark. One
widely accepted explanation offered for these facts is that Norwegians are used to language
variation, with the rather high status of dialects, their two written standards and their
no-dubbing policy of Scandinavian movies and TV shows.
6. Summary
In our study, we investigated Heritage Norwegian in the American Midwest through
a corpus of recordings collected and compiled over a time span of 80 years, from Einar
Haugen’s recordings in the 1940s via the CANS corpus and up to the present day in the
authors’ own recordings. Historical accounts of the institutions and societal texture of these
communities allow us to trace the various sources of input available to the heritage speakers
in these communities in different relevant time slots. Utilizing a model of relations between
input and output, receptive and productive competence, to show how input–output effects
will accumulate throughout the cohorts, we discuss and explain the observed linguistic
change in individuals and society.
Specifically, we investigated how the quality and quantity of input at different times
are reflected in the linguistic production of heritage speakers of the corresponding genera-
tional cohorts, focusing on relative ratios of specific word orders (topicalization and verb
second, prenominal and postnominal possessive noun phrases) and productive morphosyn-
tactic paradigms (tense suffixes of loan verbs). The latter are susceptible to cross-linguistic
influence both from neighboring dialects (koinéization), from written standards of Nor-
wegian and from the dominating language of the major society (English). This is in line
with earlier studies according to which morphology is vulnerable to change in heritage
languages (cf. Polinsky and Scontras 2020). For word order in possessives, we also find
rather clear effects of written input—in this case, the impact from written input seems
rather unequivocal, since the word order ratios for written and spoken Norwegian are
very different. Nevertheless, hypercorrection and the connotation of norwegianness of the
postnominal possessive may also play a part in the explanation of the trends we observe.
For the topicalization/verb second construction, there is little or no difference between the
ratios in spoken and written Norwegian; hence, the decline in this construction is better
attributed to the receding quantity of input, not the diversity of input sources.
This study reveals effects on the heritage language production of different types of
input, evidenced most clearly in the impact of differing availability of written sources,
but also of other dialects in the heritage community. As such, this study corroborates
the hypothesis according to which quality of input is rather important, since the access
to different types of Norwegian, including several different written standards, clearly
supports and maintains the Heritage Norwegian in these communities.
However, this does not mean that quantity is less important than diversity (cf. the
quote from Meisel 2020 in the introduction). Recent investigations of NorNAHL speakers,
e.g., Lykke and Hjelde (2022), in a sense support Meisel’s stance according to which certain
heritage speakers may have access to only one or two Norwegian speakers (typically their
parents) and still achieve a very high level of proficiency in the heritage language. As
discussed above, diversity is probably even more important when input sources are few—
when registers and domains recede, so inevitably does the potential receptive competence
of the speaker. If the input sources reduce to, e.g., two speakers, the potential for these two
speakers alone to exemplify to the learner all types of potential repertoires in a language
seems rather unrealistic and, even if it could be maintained, would also seem rather forced.
We tend to use a specific repertoire with specific people, and unless the context changes
substantially, randomly altering the communication would seem rather confusing to the
Languages 2023,8, 49 30 of 36
learner. This can potentially be counterbalanced with a steady supply of written sources in
the heritage language.
As the present study demonstrates, written input may potentially have a substantial
impact on the acquisition and maintenance of a heritage language, and the access to written
input sources may play a much greater role than hitherto assumed. Thus, the various
sources of written input available to a given heritage language learner probably deserves
to be investigated to a much greater extent than is currently the case, even beyond the
mapping of writing skills and high-level literacy.
Besides written input, the number of available sources is codependent on the number
of speakers and the contexts where the Norwegian heritage language is used. A high
number of available sources will lead to more diversity; hence, this will affect quality,
but is also likely to provide more input in sheer quantity. Not unexpectedly, then, the
question of whether quality or quantity of sources in heritage language acquisition is more
important to maintain a heritage language is theoretically important, but methodologically
and empirically rather hard to tease out.
The historical and linguistic information on a heritage language and its speakers on
the level attested in the case of Heritage Norwegian in the Midwest is probably rather
unprecedented, and the documentable relation between input types and speaker production
very rarely exists at this level of accuracy. Exactly for this reason, the present study seemed
imperative to us.
Author Contributions:
Conceptualization, K.M.E.; methodology, K.M.E. and A.H.; software, K.M.E.
and A.H.; validation, A.H.; formal analysis, K.M.E.; investigation, A.H. and K.M.E.; resources, A.H.
and K.M.E.; data curation, A.H.; writing—original draft preparation, K.M.E.; writing—review and
editing, A.H.; visualization, K.M.E. and A.H.; project administration, A.H. and K.M.E. All authors
have read and agreed to the published version of the manuscript.
Funding:
This paper specifically was written without any additional external funding although it
builds on previously conducted fieldworks, transcriptions and preparations of various kinds to make
the data accessible and usable as linguistic data. Cf. http://tekstlab.uio.no/norskiamerika/korpus.
html (accessed on 15 January 2023) for a list of contributors who have provided funding to this
overall project.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement:
Most of the data used in this study can be found at this url: http:
//tekstlab.uio.no/norskiamerika/english/corpus.html (accessed on 1 August 2022). CANS-Corpus
of American Nordic Speech v.3.1.
Conflicts of Interest: The authors declare no conflict of interest.
Notes
1We use the term repertoires instead of register to avoid the latter’s association with vocabulary.
2The model has previously been described in Norwegian, see Eide (2022).
3
In most normal acquisition situations, the preferred linguistic role model for a child will be its playmates and slightly older peers;
hence, a child usually acquires the dialect or language of the local children. In the case of heritage language, this type of input is
often lacking; hence, the input from parents becomes much more important.
4
We are specifying North American here since there are currently ongoing projects and fieldworks of Norwegian South-American
Heritage language (NorSAHL), and Norwegian Heritage Language in Spain (NorSpaHL). Cf. the home pages of the NorAmDiaSyn–
Norwegian in America project that was led by Janne Bondi Johannessen at the University of Oslo http://www.tekstlab.uio.no/
nota/NorAmDiaSyn/english/index.html (accessed on 3 August 2022) and the new project Norwegian across the Americas led by
Kari Kinn, University of Bergen. https://www.uib.no/en/lle/134611/norwegian-across-americas (accessed on 03 August 2022).
5
Janne Bondi Johannessen died of cancer in June 2020, only 59 years old. She will be remembered as one of the great figures in
Scandinavian linguistics, and as we are writing this, we are still in mourning and coping with this loss. For an overview of Janne’s
contribution to our field, cf. the English preface of Hagen et al. (2020). https://journals.uio.no/osla/article/view/8486/7896
(accessed on 2 August 2022).
Languages 2023,8, 49 31 of 36
6The same speaker has been referred to under the pseudonym “Johan” in previous studies, but here we call him “Lars”.
7
Cf. Putnam and Salmons (2015) “In the village of Hustisford, in eastern Wisconsin, twenty-four percent of residents reported
being monolingual in German in 1910, well over a half century after the main immigration to the community. Over a third of
the reported monolinguals were born in the U.S. This included numerous third generation monolinguals—grandchildren of
European immigrants who had not learned English—in 1910. Farther to the north, one scholar found that in a census district in
New Holstein, twenty-eight percent reported being monolingual, with forty-nine of those born in the U.S.”
8
Akoinéis a stabilized contact variety resulting from the mixing and leveling of features of varieties which are similar enough to
be mutually intelligible, such as regional or social dialects; (cf. Siegel 2001, p. 175).
9
The terms Nynorsk and Bokmål were introduced in 1929. Before that, Bokmål, a standard originating from Danish, was called
Rigsmaal/Riksmål. Nynorsk, based on Norwegian dialects, was called Landsmaal/Landsmål. We use the terms Nynorsk and Bokmål
consistently to simplify; but note that the terms are slightly anachronistic.
10
Nordahl Rolfsen played a very central role in Norway’s school history as he published a very ground-breaking reader for the
public school (Rolfsen 1892), which dominated the Norwegian school from the 1890s and well into the 1950s. This reader was
also used in the US, and Rolfsen even compiled a reader intended for Norwegian American readers, Boken om Norge.
11
Decorah-Posten reached a circulation of 37,000 and Skandinaven approached 50,000 issues at a time when the largest newspaper in
Norway, Aftenposten, had a circulation of 14,000 issues. It has to be mentioned though that while Decorah-Posten and Skandinaven
came twice a week, Aftenposten had two daily issues.
12
12 and 15 are the number of hours transcribed and selected for this study, out of many more hours of recordings that exist, also
among our own recordings. We also use selected examples and informants from our earlier studies from Blair.
13 Haugen (1953) points out that 93% of all borrowed verbs are assigned to this class.
14
Cf. http://tekstlab.uio.no/media/lydogbilde/amerikanorsk/einar_haugen/Transkripsjoner/ (accessed on 9 June 2022). A
couple of comments are in order. Firstly, papers such as Decorah-Posten held on to the 1907 standard up until 1938, and it is more
than likely that both written and spoken input would contain a lot of variation in the paradigms, depending on how conservative
the writer would be. In our earlier investigations we also found that speakers would produce different inflections on different
verbs, depending on whether they were of Norwegian or English origin, or even different inflections on the same verb, seemingly
in a random fashion (e.g., our informant “Lena” who would produce /like/ or /liker/ as a present tense of the Norwegian
verb ‘to like’). As the realizations of the inflections are more reliable in writing, we consulted a corpus of 500 “America-letters”
and found that the same type of variation, e.g., between the more modern preterit kastet and the more archaic variant kastede,
exists in variation for speakers of the same cohort. https://www.nb.no/emigrasjon/brev_oversikt_forfatter.php (accessed on 5
August 2022).
15
The infinitive in the old trønder dialect featured circumflex intonation, where monosyllable words are distinguished by tonality,
like /kast/ ‘throw’ (noun) without circumflex and /kâst/ ‘throw’ (infinitive) with circumflex. This intonation is clearly in place
for many of the older Trønder speakers, but no longer exists in recordings from the 2010s.
16 Other explanations are possible, but not as plausible. For instance, there was a substantial presence of speakers of the Telemark
dialect (cf. Section 3.1 above) who would employ a paradigm featuring the same distinctions as Bokmål (but with other exponents).
However, we cannot think of one single reason why the Telemark dialect would suddenly gain a much stronger impact in the
community; furthermore, there is no reason why this impact, hypothetically, once established, would suddenly drop to a level
explaining the developments we see in our data. The only causal feature of any substance that fits the time frame is written input.
17
We abstract away from the present tense third person singular marker -s, assuming that this is an agreement marker outside the
tense system proper. Cf. Eide (2009a,2009b,2016) for a detailed discussion.
18
According to adult second-language learners of Norwegian, this is also how they approach the verbal paradigms of modern
Norwegian in the homeland today. The written standard provides the regularity and authority of a norm, whereas the exponents,
the actual forms realizing the slots in the paradigms, are recruited from the dialect they hear spoken around them. Thanks to
Adrew Weir and Yvonne van Baal for volunteering their intuitions on this question.
19
Possessive constructions such as “boka hennes Janne” do not have a prenominal alternative and are as such ignored in this study.
20
The search yielded 25 POSS-N tokens versus 69 N-POSS ones. However, only about six of these twenty-five POSS-N structures
have an actual N-POSS alternative, indicating that the proportion of prenominal possessives in the Gudbrandsdalen dialect is
about 10%. These are small numbers and are only used to corroborate what we find from other sources.
21
One anonymous reviewer would like to see a lengthier discussion on why CLI is not at stake here. A proper answer to this
question would require its own paper, but we mention some main points. Whenever a homeland variety contains two or more
variants, i.e., exponents of the same morphosyntactic feature, rather sophisticated contextual restrictions regulate the selection
between the two. This is the case for, e.g., null vs. overt pronouns (Sorace 2011), the presence or absence of resumptive pronouns
in topicalizations (Bousquette et al. 2021) and the presence or absence of topicalizations themselves. In many language contact
situations, simplification will occur such that one of the variants resides. Even if both variants are present in the input to some
extent, there is too little input to provide the language learner with sufficient cues to observe the specific contextual restrictions
relevant to the two variants. Thus, the language learner resorts to acquiring one of the variants, extending the functional domain
Languages 2023,8, 49 32 of 36
of this variant to cover the domain of both variants. Now there is only one variant, performing the tasks split between two
variants in the homeland variety. This is what constitutes “simplification” in language contact situations; an aspect of the debate
on whether registers are possible in the “simplified grammars” we find in various contact situations. The acquisition of both
variants hence depends on whether input is rich enough to provide the contextual cues. Moreover, which variant is chosen also
depends on the input. In our data, we observe that “outlier Per”, whose input is almost exclusively written input, selects the
prenominal possessive as his productive “fits-all” exponent, whereas his peers, whose inputs are exclusively spoken input, select
the postnominal one as their single productive variant. The fact that the prenominal possessive is also present in the dominant
language of the language learner (English) may be very significant to some learners and trigger a “close enough” response (those
prone to alignment, cf. Section 5below). For other learners, who tend to keep their two languages as separate as possible, the
reaction may be exactly the opposite (those with an affinity to purism). To the latter learners, “norwegianness” will always trump
the possibility for convergence between their two languages.
22
Cf. Bousquette et al. (2021) for a recent investigation into various topicalization variants in Norwegian and German heritage
languages in the American Midwest.
23 Data from informant “Lena” born in 1929.
24
The rate of topicalizations varies for the same speaker across contexts; cf. Eide and Hjelde (2015a), e.g., in online production vs.
planned speech. Additionally, more complex topics typically trigger less V2. Note that certain adverbs, such as bare and kanskje
trigger non-V2 even in homeland Norwegian, triggering a small percentage of non-V2 declaratives in production.
25
Liam is atypical compared to most of his cohort peers, as he speaks Norwegian frequently and seeks any opportunity to talk
Norwegian. Hence, Liam’s continuous Norwegian input allows him to maintain his heritage language at a level different from
most of his contemporary heritage speakers.
26
However, cf. Nistov and Opsahl (2014) for a study on new urban ethnolects in Norway where topicalization recedes, cf. also
Freywald et al. (2014) for a comparative study of the phenomenon in new European urban vernaculars.
27 Chomsky uses the term “switch setting”, which we have substituted with “grammar” as it is more general.
28
Cf. Bakhtin (1981, p. 293): “Each word tastes of the context and contexts in which it has lived its socially charged life; all words
and forms are populated by intentions. Contextual overtones (generic, tendentious, individualistic) are inevitable in the word.”
29
The term register in its classical sense has been confined to vocabulary, e.g., Trudgill (1983, p. 101) “Registers are usually
characterized solely by vocabulary differences; either by the use of particular words, or by the use of words in a particular sense.”
In more recent works, register is however extended to other linguistic features, as discussed, e.g., by Eide and Weir (2020, p. 229).
30
Or rather intake. Input refers to all the target language that the learner reads and hears; intake refers to the part of input which the
learner comprehends and uses to develop his or her internal grammar of the target language. Van Patten (1996) attributes this
term to Corder (1967). We are glossing over the input–intake debate here.
31
“Vore Børns Sprog er nesten udelukket Engelsk mend de forstaar hvert Ord, der tales til dem i norsk, det er haardt at faa lære
Børnene norsk her i landet, da Engelsk falder dem lettere.”
32
Foreigner-directed, child-directed and computer-directed speech feature traits facilitating disambiguation, e.g., “less vowel
reduction, a larger vowel space, a higher pitch, fewer idiomatic expressions, more high frequency words, simple syntactic
constructions, and [
. . .
] more repetitions as well as clarifications when compared to casual speech”; cf. Rothermich et al. (2019,
p. 22).
33
Note that when we talk about the youngest speakers, these are the speakers born around 1940 and later; hence, some of these
“youngsters” are in fact in their 80s, and the youngest of these “youngest” were around 50 when recorded (in their early 60s today).
References
Åfarli, Tor A. 2012. Hybride verbformer i amerikanorsk og analyse av tempus. [Hybrid verb forms in American Norwegian and
analysis of tense]. Norsk Lingvistisk Tidsskrift 30: 381–98.
Aikhenvald, Alexandra Y. 2006. Grammars in Contact: A Cross-Linguistic Typology. In Alexandra Aikhenvald and R M W Dixon
Grammars in Contact: A Cross-Linguistic Typology. Oxford: Oxford University Press, pp. 1–66.
Amaral, Luiz, and Tom Roeper. 2014. Multiple grammars and second language representation. Second Language Research 30: 3–36.
[CrossRef]
Aschim, Anders. 2019. Luther and Norwegian Nation-Building. Nordlit 43: 127–41. [CrossRef]
Bakhtin, Michail M. 1981. The Dialogic Imagination: Four Essays. Edited by Michael Holquist. Translated by Caryl Emerson, and Michael
Holquist. Austin: University of Texas Press.
Bell, Allan. 1984. Language style as audience design. Language in Society 13: 145–204. [CrossRef]
Benmamoun, Elabbas, Silvina Montrul, and Maria Polinsky. 2013. Heritage languages and their speakers: Opportunities and challenges
for linguistics. Theoretical Linguistics 39: 129–81. [CrossRef]
Blegen, Theodore Christian. 1969. Norwegian Migration to America: The American Transition. Lake Geneva: Ardent Media. First
published 1940.
Blom, Jan-Petter, and John J. Gumperz. 2000. Social meaning in linguistic structure: Code-switching in Norway. In The Bilingualism
Reader. London: Routledge, pp. 111–36.
Languages 2023,8, 49 33 of 36
Bousquette, Joshua, Kristin Melum Eide, Arnstein Hjelde, and Michael T. Putnam. 2021. Competition at the Left Edge: Left-Dislocation
vs. Topicalization in Heritage Germanic. In Selected Proceedings of the 10th Workshop on Immigrant Languages in the Americas (WILA
10). Edited by Arnstein Hjelde and Åshild Søfteland. Somerville: Cascadilla Proceedings Project.
Chomsky, Noam. 1965. Aspects of the Theory of Syntax. Cambridge: The MIT Press.
Chomsky, Noam. 2000. The Architecture of Language. Oxford: Oxford University Press.
Cook, Vivian Jones, and Mark Newson. 2007. Chomsky’s Universal Grammar. An introduction, 3rd ed. Hoboken: Blackwell Publishing.
Corder, Pit. 1967. The significance of learner’s errors. International Review of Applied Linguistics in Language Teaching 5: 161–70. [CrossRef]
Cornips, Leonie, and Cecilia Poletto. 2005. On standardising syntactic elicitation techniques (part 1). Lingua 115: 939–57. [CrossRef]
Danbolt, Alf. 2021. Book Review of Biografi om C. F. W. Walther. [Biography of C. F. W. Walther]. Available online: https:
//www.fbb.nu/artikkel/biografi-om-cfw-walther/ (accessed on 8 August 2022).
Delsing, Lars-Olof, and Katarina Lundin-Åkesson. 2005. Håller språket ihop i Norden?: En forskningsrapport om ungdomars förståelse av
danska, svenska och norska. [Does the Language Hold the Nordic Countries Together? A Research Report on Young People’s Understanding
of Danish, Swedish and Norwegian? Copenhagen: Nordic Council Report.
Eide, Kristin Melum. 2002. Norwegian Modals. Ph.D. thesis, Norwegian University of Science and Technology (NTNU), Trondheim,
Norway.
Eide, Kristin Melum. 2005. Norwegian Modals. Studies in Generative Grammar 74. Berlin: Mouton de Gruyter.
Eide, Kristin Melum. 2009a. Finiteness: The haves and the have-nots. In Advances in Comparative Germanic Syntax. Edited by Artemis
Alexiadou, Jorge Hankamer, Thomas McFadden, Justin Nuger and Florian Schäfer. Linguistik Aktuell/Linguistics Today 141.
Philadelphia: John Benjamins, pp. 357–90.
Eide, Kristin Melum. 2009b. Tense, finiteness and the survive principle: Temporal chains in a crash-proof grammar. In Towards a
Derivational Syntax: Survive-Minimalism. Edited by Michael Putnam. Linguistik Aktuell/Linguistic Today 144. Philadelphia: John
Benjamins, pp. 91–132.
Eide, Kristin Melum. 2011. Norwegian non-V2 declaratives and the Wackernagel position. Nordic Journal of Linguistics 34: 179–213.
[CrossRef]
Eide, Kristin Melum. 2016. Finiteness: Finiteness, inflection, and the syntax your morphology can afford. In Finiteness Matters: On
Finiteness-Related Phenomena in Natural Languages. Edited by Kristin Melum Eide. Linguistik Aktuell/Linguistic Today 2231.
Philadelphia: John Benjamins, pp. 121–67.
Eide, Kristin Melum. 2019. Convergence and Hybrid Rules: Verb Movement in Heritage Norwegian of the American Midwest. In
Selected Proceedings of the 9th Annual Workshop on Immigrant Languages in the Americas (WILA 9). Edited by Kelly Biers and Joshua
R. Brown. Somerville: Cascadilla Proceedings Project, pp. 45–53.
Eide, Kristin Melum. 2022. Språket som superkraft. [Language Is Your Superpower]. Bergen: Fagbokforlaget.
Eide, Kristin Melum, and Andrew Weir. 2020. Introduction to special issue on morphosyntactic variation within the individual
language user. Nordic Journal of Linguistics 43: 229–31. [CrossRef]
Eide, Kristin Melum, and Arnstein Hjelde. 2015a. Verb Second and Finiteness Morphology in Norwegian Heritage Language of the
American Midwest. In Moribund Germanic Heritage Languages in North America. Edited by Page B. Richard and Michael T. Putman.
Leiden: Brill, pp. 64–101.
Eide, Kristin Melum, and Arnstein Hjelde. 2015b. Borrowing modal elements into American Norwegian. In Germanic Heritage Languages
in North America. Edited by Janne Bondi Johannessen and Joseph C. Salmons. Amsterdam: John Benjamins Publishing Company,
pp. 256–80.
Eide, Kristin Melum, and Arnstein Hjelde. 2018. Om verbplassering og verbmorfologi i amerikanorsk. [On verb placement and verb
morphology in American Norwegian]. Maal og Minne 110: 25–70.
Eide, Kristin Melum, and Hilde Sollid. 2011. Norwegian main clause declaratives: Variation within and across grammars. In Linguistic
Universals and Language Variation. Edited by Peter Siemund. Berlin: De Gruyter Mouton, pp. 327–60.
Eide, Kristin Melum, and Tor A. Åfarli. 2020. Dialects, registers and intraindividual variation: Outside the scope of generative
frameworks? Nordic Journal of Linguistics 43: 233–48. [CrossRef]
Flom, George T. 1929. On the Phonology of English Loanwords in the Norwegian Dialects of Koshkonong in Wisconsin. Arkiv for
Nordisk Filologi–Tilläggsband till bd, 178–89.
Freywald, Ulrike, Natalia Ganuza, Ingvild Nistov, and Toril Opsahl. 2014. Urban vernaculars in contemporary northern Europe:
Innovative variants of V2 in Germany, Norway, and Sweden. In Language Youth & Identity in the 21st Century. Edited by Jacomine
Nortier and Bente Ailin Svensen. Cambridge: Cambridge University Press.
Frøshaug, Andrea Skjold, and Truls Stende. 2021. Har Norden et språkfellesskap? [Does the Nordic Region Have a Language Community?].
Copenhagen: Nordic Council Report.
Gilbertson, Laurann. 2009. Religion and Norwegian-American Quilts. University of Nebraska: Proceedings of the 4th Biennial Symposium of
the International Quilt Museum: 3. Available online: https://digitalcommons.unl.edu/ (accessed on 20 August 2022).
Giles, Howard. 1973. Accent mobility: A model and some data. Anthropological Linguistics 15: 87–105.
Gollan, Tamar H., Jennie Starr, and Victor S. Ferreira. 2015. More than use it or lose it: The number-of-speakers effect on heritage
language proficiency. Psychonomic Bulletin & Review 22: 147–55.
Granquist, Mark. 2015. Lutherans in America: A New History. Minneapolis: Fortress Press.
Languages 2023,8, 49 34 of 36
Gulliksen, Øyvind T. 2002. Interdisciplinary Approaches in American Immigration Studies: Possibilities and Pitfalls. In Interpreting the
Promise of America: Essays in Honor of Odd Sverre Lovoll. Edited by Odd S. Lovoll and Todd W. Nichol. Northfield: Norwegian-
American Historical Association, pp. 31–51.
Hagen, Kristin, Arnstein Hjelde, Karine Stjernholm, and Øystein A. Vangsnes, eds. 2020. Bauta: Janne Bondi Johannessen in Memoriam.
OSLa. 11.2. Available online: https://journals.uio.no/osla/article/view/8486/7896 (accessed on 4 August 2022).
Haugen, Einar. 1938. Language and Immigration. Norwegian-American Studies 10: 1–43.
Haugen, Einar. 1939. Norsk i Amerika. [Norwegian in America]. Oslo: Cappelen.
Haugen, Einar. 1953. The Norwegian Language in America: A Study in Bilingual Behavior. Vol. I, The Bilingual Community. Vol. II, The
American Dialects of Norwegian. Philadelphia: University of Pennsylvania Press.
Haugen, Einar. 1966. Language Conflict and Language Planning: The Case of Modern Norwegian. Cambridge: Harvard University Press.
Hjelde, Arnstein. 1992. Trøndsk talemål i Amerika. [The Trønder Dialect in America]. Trondheim: Tapir.
Hjelde, Arnstein. 1996. Some phonological changes in a Norwegian dialect in America. In Language Contact across the North Atlantic.
Edited by Ureland P. Sture and Iain Clarkson. Berlin: De Gruyter, pp. 283–95.
Hjelde, Arnstein. 2012. “Folkan mine, dæm bære snakka norsk”–norsk i Wisconsin frå 1940-talet og fram til i dag. [«My people, they
only spoke Norwegian»: Norwegian in Wisconsin from the 1940s until today]. Norsk Lingvistisk Tidsskrift 30: 183–203.
Hjelde, Arnstein. 2015. Changes in a Norwegian dialect in America. In Germanic Heritage Languages in North America Germanic-
Acquisition, Attrition and Change. Edited by Janne Bondi Johannessen and Joseph C. Salmons. Amsterdam: John Benjamins
Publishing Company, pp. 283–98.
Hjelde, Arnstein, and Camilla Bjørke. 2022. Reading Skills among Young Norwegian-American Heritage Speakers in the Early 1900s.
In Selected Proceedings of the 11th Workshop on Immigrant Languages in the Americas (WILA 11). Edited by Kelly Biers and Joshua R.
Brown. Somerville: Cascadilla Proceedings Project, pp. 18–28.
Hjelde, Arnstein, Bjørn Harald Kvifte, Linda Evenstad Emilsen, and Ragnar Arntzen. 2019. Processability Theory as a Tool in the Study
of a Heritage Speaker of Norwegian. In Widening Contexts for Processability Theory: Theories and Issues. Edited by Anke Lenzing,
Howard Nicolas and Jana Roos. Amsterdam: John Benjamins Publishing Company, pp. 185–206.
Jamieson, E. 2020. Viewing dialect change through acceptability judgments: A case study in Shetland dialect. Glossa 5: 1–28. [CrossRef]
Johannessen, Janne Bondi. 2015. The Corpus of American Norwegian Speech (CANS). Paper present at the 20th Nordic Conference of
Computational Linguistics, NODALIDA 2015, Vilnius, Lithuania, May 11–13. NEALT Proceedings Series 23.
Johannessen, Janne Bondi. 2018. Factors of variation, maintenance and change in Scandinavian heritage languages. International Journal
of Bilingualism 22: 447–65. [CrossRef]
Johannessen, Janne Bondi. 2021. From Fieldwork to Speech Corpus: The American Norwegian Heritage Language and CANS. In
Proceedings of the 10th Workshop on Immigrant Languages in the Americas (WILA 10). Edited by Arnstein Hjelde and Åshild Søfteland.
Somerville: Cascadilla Proceedings Project, pp. 1–10. Available online: http://www.lingref.com/cpp/wila/10/paper3535.pdf
(accessed on 5 August 2022).
Johannessen, Janne Bondi, and Joseph C. Salmons. 2015. The study of Germanic heritage languages in the Americas. In Germanic
Heritage Languages in North America: Acquisition, Attrition and Change. Edited by Janne Bondi Johannessen and Joseph C. Salmons.
Amsterdam: John Benjamins Publishing Company, vol. 18.
Johannessen, Janne Bondi, and Michael T. Putnam. 2020. Heritage Germanic languages in North America. In The Cambridge Handbook
of Germanic Linguistics. Edited by Michael T. Putnam and Page B. Richard. Cambridge: Cambridge University Press, pp. 783–806.
Johannessen, Janne Bondi, and Signe Laake. 2015. On two myths of the Norwegian language in America: Is it old-fashioned? Is it
approaching a written Bokmål standard? In The Study of Germanic Heritage Languages in the Americas. Edited by Janne Bondi
Johannessen and Joe Salmons. Amsterdam: John Benjamins, pp. 299–322. Available online: http://www.jbe-platform.com/
content/books/9789027268198 (accessed on 4 August 2022).
Johannessen, Janne Bondi, and Signe Laake. 2017. Norwegian in the American Midwest: A Common Dialect? Journal of Language
Contact 10: 5–21. [CrossRef]
Joranger, Terje Mikael. 2010. Lokale eller nasjonale kollektive identiteter? Etnifisering og identitetsbygging blant norske immigranter i
Amerika. [Local or national collective identities? Ethnification and identity building among Norwegian immigrants in America].
Historisk Tidsskrift 89: 223–47. [CrossRef]
Kjær, Iver, and M. Baumann Larsen. 1972. Dansk i Amerika: Status og perspektiv. [Danish in America: Status and perspectives]. Sprog
i Norden 3.
Kleinmann, Howard H. 1977. Avoidance behavior in adult second language acquisition 1. Language Learning 27: 93–107. [CrossRef]
Kroch, Anthony, and Ann Taylor. 1996. Verb movement in Old and Middle English: Dialect Variation and Language Contact. Philadelphia:
Department of Linguistics, University of Pennsylvania [Host].
Kruse, Arne. 1991. Norske stadnamn i Coon Valley, Wisconsin: Møte mellom to tradisjonar. [Norwegian place names in Coon Valley,
Wisconsin: A meeting between traditions]. In Norsk språk i Amerika. [Norwegian Language in America]. Edited by Botolf Helleland.
Oslo: Novus Press, pp. 135–71.
Lovoll, Odd Sverre. 1977. Decorah-Posten: The Story of an Immigrant Newspaper. Norwegian-American Studies 27: 77–100.
Lovoll, Odd Sverre. 1983. Det løfterike landet. [The Land of Promise]. Oslo: Universitetsforlaget.
Lovoll, Odd Sverre. 2006. Norwegians on the Prairie: Ethnicity and the Development of the Country Town. Saint Paul: Minnesota Historical
Society Press.
Languages 2023,8, 49 35 of 36
Lovoll, Odd Sverre. 2010. Norwegian Newspapers in America: Connecting Norway and the New Land. Saint Paul: Minnesota Historical
Society Press.
Lykke, Alexander K., and Arnstein Hjelde. 2022. New perspectives on grammatical change in heritage Norwegian: Introducing the
adult speaker and adolescent relearner. University of Bergen: Bergen Language and Linguistics Studies 12: 63–76. [CrossRef]
Lødrup, Helge. 2012. Forholdet mellom prenominale og postnominale possessive uttrykk. In Grammatikk, bruk og norm. Festskrift til
Svein Lie på 70-årsdagen, 15 April 2012. Edited by Hans-Olav Enger, Jan Terje Faarlund and Kjell Ivar Vannebo. Oslo: Novus,
pp. 189–203.
Mathiesen, Henrik Olav. 2014. Belonging in the Midwest: Norwegian Americans and the Process of Attachment, ca. 1830–60. American
Nineteenth Century History 15: 119–46. [CrossRef]
Matras, Yaron. 2009. Language Contact. Cambridge: Cambridge University Press.
Matsuda, Mari J. 1991. Voices of America: Accent, Antidiscrimination Law, and a Jurisprudence for the Last Reconstruction. The Yale
Law Journal 100: 1329–407. [CrossRef]
Meisel, Jürgen M. 2020. Shrinking structures in heritage languages: Triggered by reduced quantity of input? Bilingualism: Language and
Cognition 23: 33–34. [CrossRef]
Moen, Per. 1991. The influence of a Norwegian substratum on the pronunciation of Norwegian-Americans in Upper Midwest. In Norsk
språk i Amerika. [Norwegian Language in America]. Edited by Botolf Helleland. Oslo: Novus Press, pp. 97–115.
Montrul, Silvina. 2008. Incomplete Acquisition in Bilingualism: Re-Examining the Age Factor. Amsterdam and Philadelphia: John Benjamins.
Munch, Peter A. 1949. Social Adjustment among Wisconsin Norwegians. American Sociological Review 14: 780–87. [CrossRef]
Myers-Scotton, Carol. 1993. Duelling Languages: Grammatical Structure in Codeswitching. Oxford: Oxford University Press.
Nistov, Ingvild, and Toril Opsahl. 2014. The social side of syntax in multilingual Oslo. In The Sociolinguistics of Grammar. Edited by Tor
A. Åfarli and Brit Mæhlum. Amsterdam: John Benjamins Publishing, pp. 91–116.
Pater, Joe. 2004. Bridging the gap between receptive and productive development with minimally violable constraints. In Constraints
in Phonological Acquisition. Edited by RenéKager, Joe Pater and Wim Zonneveld. Cambridge: Cambridge University Press,
pp. 2019–244.
Polinsky, Maria, and Gregory Scontras. 2020. Understanding heritage languages. Bilingualism: Language and Cognition 23: 4–20.
[CrossRef]
Prince, Ellen F. 1981. Topicalization, focus-movement, and Yiddish-movement: A pragmatic differentiation. Annual Meeting of The
Berkeley Linguistics Society 7: 249–64. [CrossRef]
Putnam, Michael T., and Joseph Salmons. 2015. Multilingualism in the Midwest: How German has shaped (and still shapes) the
Midwest. Middle West Review 1: 29–52. [CrossRef]
Reite, AndréMidtskogseter. 2011. Spørjing i skedsmokorsmålet: Ein generativ analyse av leddstelling og spørjeord i interrogative
hovudsetningar i talemålet på Skedsmokorset. [Asking in the Skedsmokorset Dialect: A Generative Analysis of Sentence Structure
and Interrogatives in Interrogative Main Clauses in the Skedsmokorset Dialect]. Master’s thesis, Norwegian University of Science
and Technology, Trondheim, Norway.
Riksem, Brita Ramsevik. 2018. Language Mixing in American Norwegian Noun Phrases. An Exoskeletal Analysis of Synchronic and
Diachronic Patterns. Ph.D. thesis, Norwegian University of Science and Technology, Trondheim, Norway.
Roeper, Thomas. 1999. Universal bilingualism. Bilingualism: Language and Cognition 2: 169–86. [CrossRef]
Roeper, Tom W. 2016. Multiple grammars and the logic of learnability in second language acquisition. Frontiers in Psychology 7: 14.
[CrossRef]
Rolfsen, Nordahl. 1892. Læsebog for Folkeskolen: Med Tegninger af Norske Kunstnere. [Reader for the Primary School: With Drawings by
Norwegian Artists]. Kristiania: Jacob Dybwads forlag.
Rothermich, Kathrin, Havan L. Harris, Kerry Sewell, and Susan C. Bobb. 2019. Listener impressions of foreigner-directed speech:
A systematic review. Speech Communication 112: 22–29. [CrossRef]
Rothman, Jason. 2009. Understanding the nature and outcomes of early bilingualism: Romance languages as heritage languages.
International Journal of Bilingualism 13: 155–63. [CrossRef]
Salmons, Joseph. 2005a. The role of community and regional structure in language shift. In Regionalism in the Age of Globalism: Volume 1:
Concepts of Regionalism. Edited by Lothar Hönnighausen, Marc Frey, James Peacock and Niklaus Steiner. Madison: Center for the
Study of Upper Midwestern Cultures, pp. 129–38.
Salmons, Joseph. 2005b. Community, Region and Language Shift in German-speaking Wisconsin. In Regionalism in the Age of Globalism:
Volume 2: Forms of Regionalism. Edited by Lothar Hönnighausen, Anke Ortlepp, James Peacock, Niklaus Steiner and Carrie
Matthews. Madison: Center for the Study of Upper Midwestern, pp. 133–44.
Schilling-Estes, Natalie. 2004. Constructing ethnicity in interaction. Journal of Sociolinguistics 8: 163–95. [CrossRef]
Schumann, John H. 1978. The Pidginization Process: A Model for Second Language Acquisition. New York: Newbury House Publishers.
Siegel, Jeff. 2001. Koine formation and creole genesis. Creole Language Library 23: 175–98.
Soleng, Kine Dahlen. 2007. Their Minneapolis: Novels of Norwegian-American Life in Minneapolis. Master’s thesis, University of
Oslo, Oslo, Norway.
Sorace, Antonella. 2011. Pinning down the concept of “interface” in bilingualism. Linguistic Approaches to Bilingualism 1: 1–33.
[CrossRef]
Languages 2023,8, 49 36 of 36
Sunde, Anne Mette. 2019. Skjult påvirkning: Tre studier av engelskpåvirkning i norsk. [Hidden Influence: Three Studies of English
Influence in Norwegian]. Ph.D. thesis, Norwegian University of Science and Technology, Trondheim, Norway.
Thomas, Georg. 1991. Linguistic Purism. London and New York: Longman.
Troike, Rudolph C. 1970. Receptive competence, productive competence, and performance. In Linguistics and the Teaching of Standard
English to Speakers of Other Languages or Dialects. Edited by James E. Alatis. Washington: Georgetown University Press, pp. 63–74.
Trudgill, Peter. 1972. Sex, Covert Prestige and Linguistic Change in the Urban British English of Norwich. Language in Society 1: 179–95.
[CrossRef]
Trudgill, Peter. 1983. On Dialect: Social and Geographical Perspectives. New York: New York University Press.
Ulbricht, Christiane, and Alexander Werth. 2021. What is Intra-individual Variation in Language? In Intra-Individual Variation in
Language. Edited by Alexander Werth, Lars Bülow, Simone E. Pfenninger and Markus Schiegg. Berlin and Boston: De Gruyter
Mouton (Trends in Linguistics).
Vaid, Jyotsna. 2002. Bilingualism. In Encyclopedia of the Human Brain. Edited by Vilayanur S. Ramachandran. Amsterdam: Elsevier,
pp. 417–44.
van Craenenbroeck, Jeroen, and Liliane Haegeman. 2007. The Derivation of Subject-Initial V2. Linguistic Inquiry 38: 167–78. [CrossRef]
Van Patten, Bill. 1996. Input Processing and Grammar Instruction in Second Language Acquisition. Norwood: AbLex Publishing Corperation.
Venås, Kjell. 1974. Linne verb i norske målføre: Morfologiske studiar. [Weak Verbs in Norwegian Dialects: Morphological Studies]. Oslo:
Universitetsforlaget.
Weinreich, Uriel. 2011. Languages in Contact: French, German, and Romansh in Twentieth-Century Switzerland. Introduction and Notes by
Ronald I. Kim and William Labov. Amsterdam and Philadelphia: John Benjamins. First Publication of Weinreich’s Ph.D. thesis,
Research Problems in Bilingualism with Special Reference to Switzerland. Columbia University, New York, NY, USA, 1951.
Westergaard, Marit, and Merete Anderssen. 2015. Word order variation in Norwegian possessive constructions. In Germanic Heritage
Languages in North America: Acquisition, Attrition and Change. Edited by Janne Bondi Johannessen and Joseph C. Salmons.
Amsterdam: John Benjamins Publishing Company, vol. 18.
Wiese, Heike, Artemis Alexiadou, Shanley Allen, Oliver Bunk, Natalia Gagarina, Kateryna Iefremenko, Maria Martynova, Tatiana
Pashkova, Vicky Rizou, Christoph Schroeder, and et al. 2022. Heritage speakers as part of the native language continuum.
Frontiers in Psychology 12: 717973. [CrossRef] [PubMed]
Winford, Donald. 2003. An Introduction to Language Contact. New York: Wiley Blackwell.
Disclaimer/Publisher’s Note:
The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.
